i

THE ROLE OF LACTATE MEASUREMENT

IN THE PREDICTION OF

FETAL HYPOXIC-ISCHAEMIC BRAIN INJURY

DURING LABOUR

Craig Edward Pennell

MBBS(Hons) FRANZCOG

A thesis submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy

School of Women’s and Infants’ Health

Faculty of Medicine and Dentistry

The University of Western Australia

June 2003

ii

This thesis is dedicated to my darling wife Nicola,

for without her my achievements would not be possible.

iii

Science

To the physician particularly, a scientific discipline is an incalculable gift,

which leavens his whole life, giving exactness to habits of thought and

tempering the mind with that judicious faculty of distrust which can alone,

amid the uncertainties of practice, make him wise unto salvation.

For perdition inevitably awaits the mind of the practitioner

who has never had the full inoculation with the leaven,

who has never grasped clearly the relations of science to his art,

and who knows nothing and perhaps cares less, for the limitations of either.

Sir William Osler, 1894

iv

ABSTRACT

In this thesis the role of lactate measurement has been evaluated in intrapartum assessment

of fetal wellbeing. Specifically, I have addressed the question of whether fetal lactate

measurement is better than the assessment of fetal heart rate patterns or the measurement of

pH at predicting fetal brain injury after intrapartum asphyxia. Using an ovine model of

repeated umbilical cord occlusion designed to mimic events which may occur during human

labour, I have shown that the measurement of fetal lactate levels after repeated cord

occlusion is significantly associated with the severity of brain injury after the asphyxial insult.

No significant associations were identified with fetal pH measurements or with the duration

of decelerative or compound fetal heart rate patterns; however, this is the first study to

describe an association between the duration of both increased fetal heart rate variability and

fetal heart rate overshoot with the severity of subsequent brain injury. Although no

significant association was identified between fetal arterial pressure measured between

umbilical cord occlusions and the grade of brain injury, the studies performed in this thesis

are the first to show a strong correlation between the duration of specific arterial pressure

responses during cord occlusions and the grade of brain injury, accounting for approximately

90% of the variability seen in the severity of injury.

The mechanism responsible for the improved ability of lactate measurement to predict fetal

brain injury is unknown. It may be because fetal lactate levels are a more stable marker of

anaerobic metabolism of glucose than fetal pH levels, which are influenced by both

increasing levels of carbon dioxide and anaerobic metabolism of amino-acids and fatty acids.

In addition fetal pH levels can be rapidly normalised through placental exchange of carbon

dioxide whereas fetal lactate levels are slow to normalise across the placenta as they rely on

facilitated diffusion.

In this thesis I have provided evidence for two additional potential mechanisms explaining

why fetal lactate levels may be better predictors of fetal brain injury than fetal pH assessment.

First, evidence has been provided which suggests that fetal lactate levels are a more reliable

marker of cardiovascular instability during umbilical cord occlusion than other markers of

fetal acid-base status, with high sensitivity, specificity and accuracy. The second potential

mechanism may involve the association between fetal lactate levels and oxygen free radical

activity in the fetal brain. There is a growing body of evidence suggesting that oxygen free

v

radical activity in the fetal brain is critically involved in the mechanism and timing of fetal

hypoxic-ischaemic brain injury. In this thesis I have provided evidence which suggests that

fetal lactate levels are significantly better than other markers of fetal acid-base status at

predicting increases in oxygen free radical activity in the blood perfusing the fetal brain.

In human labour the fetal scalp is the window for biochemical assessment of fetal wellbeing.

In this thesis I have thoroughly investigated the relationship between fetal scalp blood lactate

levels and central blood lactate levels, considering both the effects of fetal asphyxia on this

relationship and also the effects of the adverse intrauterine environment associated with fetal

growth restriction. The relationship between fetal scalp blood and carotid arterial lactate

levels was influenced by asphyxia but not by fetal growth restriction. In contrast, the

accuracy of fetal scalp pH assessment was impaired in growth restricted fetuses with scalp

pH levels significantly underestimating the severity of changes in central blood pH. Despite

the complex relationships between fetal scalp blood lactate levels and central blood lactate

levels, scalp lactate levels were significantly better at predicting lipid peroxidation in the fetal

brain than central markers of fetal acid-base status. In addition, fetal growth restriction did

not impair the ability of fetal scalp blood lactate levels to predict cardiovascular instability

during cord occlusion.

One of the major limitations with the current clinical technique for biochemical assessment

of fetal wellbeing during labour is the restricted availability of the expensive equipment

required to perform acid-base measurement on micro volume blood samples, thus limiting

the technique to major obstetric centres. In this thesis I have assessed the accuracy and

precision of a handheld lactate meter which is commercially available in Australia, costing

approximately AUS$900 with analysis of each 15µL sample costing $1, and found it to be

both accurate and precise in lactate measurement in human blood micro samples across the

gestational age range of 26 to 42 weeks.

Together the results of this combined clinical and laboratory study suggests that the

intrapartum assessment of fetal lactate levels may be a superior method of fetal assessment,

warranting consideration for widespread clinical practice.

vi

DECLARATION

I, Craig Edward Pennell, declare that this thesis is less than 100,000 words in length,

exclusive of references. Except where due acknowledgement has been made, this thesis

comprises original work that I have performed. To the best of my knowledge and belief, this

thesis contains no material that has been previously written or published by any other person,

except where due reference is given in the text. No part of this thesis has been submitted to

any institution for the award of any other degree or diploma.

Craig Pennell

vii

ACKNOWLEDGEMENTS

I take this opportunity to express my sincere gratitude to all those who, throughout my

doctoral candidature, have contributed to its completion.

I would like to begin by acknowledging my wife Nicola, who has supported me tirelessly

throughout my doctoral candidature which has been performed simultaneously with my

specialty training in obstetrics and gynaecology and my subspecialty training in maternal-fetal

medicine. Few women are capable of caring for a young family whilst providing practical,

psychological and emotional support during the thousands of hours which have dissolved

into this thesis.

This thesis began when I was a first year obstetric and gynaecology registrar in Sydney in

1997. I performed the human studies and many of the animal studies presented in this thesis

whilst I was completing my specialty training in obstetrics and gynaecology in Sydney during

1997 to 2001. I would like to sincerely thank Dr Henry Murray, my supervisor in Sydney for

encouraging me to begin the studies presented in this thesis. I would also like to thank him

for teaching me the principles of ovine surgery and for providing the practical and financial

support for the studies which I performed whilst in Sydney. In addition, I would like to

thank him for his assistance during the final stages of preparation of this thesis.

In January 2001, my family and I moved to Perth so I could accept the Forrest Fellowship in

Fetal Medicine to undertake subspecialty training in maternal-fetal medicine. It was in Perth

that I completed the animal studies and the final preparation of this thesis. I would sincerely

like to thank Professor John Newnham and Dr Timothy Moss, my supervisors at The

University of Western Australia, for allowing me to perform the last year of ovine studies

presented in this thesis and, more importantly, for their scientific guidance during the

preparation of this manuscript. Professor John Newnham has instilled in me the principles

of thorough scientific evaluation and taught me the importance of taking small steps

forwards during scientific investigation. In addition, he has been both a friend and mentor

during my time in Perth and his enthusiasm, encouragement, support and patience have

acted to shape my approach to scientific research. Dr Timothy Moss has provided both

practical and technical assistance during the final series of animal studies performed in this

thesis. More importantly, he has shaped the way that I present my research, both written and

viii

oral, and he has provided hours of invaluable support, guidance and friendship during the

preparation of this manuscript.

I would like to thank Mrs Anita Turner, Dr Heather Coughtrey, Dr John Smyth, Dr Ping

Yan, Dr Ilias Nitsos and Mr Adrian Jonker for their assistance in performing the animal

studies presented in this thesis. I would particularly like to thank Mrs Anita Turner for

teaching me the techniques required to chronically instrument fetuses and for her presence

and support during the many early mornings, late nights and weekends that were invested in

the animal studies in this thesis.

Dr Dorota Doherty, from the Women and Infants Research Foundation Biostatistics and

Research Design Unit, Perth, has provided invaluable assistance both during my postgraduate

statistics training at The University of Western Australia and during the preparation of my

thesis. Thank you for your time and your critical appraisal of the data and the analyses

presented in this thesis.

I take this opportunity to thank Dr Thomas Ng from the Institute of Chemical Pathology

and Medical Research, Westmead Hospital, The University of Sydney, for his assistance in

performing the histopathological assessment of the fetal brains in this thesis. In addition, I

would like to thank Dr CC (Ronald) Wang from the Chinese University of Hong Kong for

performing the measurements of lipid peroxidation presented in this thesis and Miss Vittoria

Lazzaro from the Renal Laboratory at King George the Fifth Hospital, Sydney, for

performing the catecholamine assays presented in this thesis.

Finally I would like to take this opportunity to thank my father Roy and my late mother

Wendy for instilling in me the belief that anything is possible if you try hard enough.

ix

SPECIAL ACKNOWLEDGEMENT

I wish to give special acknowledgment to Mr Andrew and Mrs Nicola Forrest who provided,

through donation, the financial resources to support my research fellowship. This fellowship

position is known as the Forrest Fellowship in Fetal Medicine and I was the inaugural

recipient of this award to undertake subspecialty training in maternal-fetal medicine.

The Forrest family has descended from a distinguished list of prominent Western Australians

acclaimed for pioneering the early development of this State. This pioneering tradition lives

on through Andrew and Nicola Forrest in the context of the modern mining industry. It has

been their wish to support advancement through research in fetal medicine which represents

the pioneering edge of contemporary science, with a promise to improve lives of people

through interventions at early times in life.

It is my hope this thesis has contributed to their vision.

x

MANUSCRIPTS, ABSTRACTS AND AWARDS ARISING FROM

THIS THESIS

Manuscripts

PENNELL, C.E., TRACY, M.B. (1999). A new method for rapid measurement of lactate in fetal and neonatal blood. Aust N Z J Obstet Gynaecol 39, 227-33. ROGERS, M.S., MURRAY, H.G., WANG, C.C., PENNELL, C.E., TURNER, A., YAN, P., PANG, C.C., CHANG, A.M. (2001). Oxidative stress in the fetal lamb brain following intermittent umbilical cord occlusion: a path analysis. British Journal of Obstetrics and Gynaecology 108, 1283-90.

Published Abstracts

PENNELL, C.E., TRACY, M.B. A new method for rapid measurement of lactate in fetal and neonatal blood. Canberra Centre for Perinatal Care Fetal Medicine Meeting. Canberra, Australia. September, 1997. PENNELL, C.E. The role of intrapartum lactate analysis in fetal distress. Taking Midwifery to the Future. Sydney, Australia. March, 1998 PENNELL, C.E., MURRAY, H.G., TURNER, A., PING, Y. Fetal lactic acid changes during cord occlusion in the healthy ovine fetus. Twelfth Annual Workshop on Fetal and Neonatal Physiology. Alice Springs, Australia. March, 1998. PENNELL, C.E., MURRAY, H.G., TURNER, A., PING, Y. Fetal lactic acid changes during cord occlusion in the unhealthy ovine fetus. Twelfth Annual Workshop on Fetal and Neonatal Physiology. Alice Springs, Australia. March, 1998 PENNELL, C.E., MURRAY, H.G., TURNER, A., PING, Y. Fetal lactic acid changes during cord occlusion in the healthy and unhealthy ovine fetus. Royal Australian College of Obstetricians and Gynaecologists Annual Scientific Meeting. Canberra, Australia. April, 1998. PENNELL, C.E., TURNER, A., PING, Y., WANG, C.C., ROGERS, M.S., MURRAY, H.G. Lipid peroxidation and lactic acid production during cord occlusion in the ovine fetus. Fetal and Neonatal Physiological Society 26th Annual Meeting. Vlieland, Netherlands. September, 1999. PENNELL, C.E., TURNER, A., PING, Y., MURRAY, H.G. Central and peripheral lactic acid changes during cord occlusion in healthy and compromised ovine fetuses. Fetal and Neonatal Physiological Society 26th Annual Meeting. Vlieland, Netherlands. September, 1999.

xi

PENNELL, C.E., TURNER, A., PING, Y., WANG, C.C., ROGERS, M.S., MURRAY, H.G. Lipid peroxidation and lactic acid production during cord occlusion in the ovine fetus. Perinatal Society of Australia and New Zealand 4th Annual Congress. Brisbane, Australia. March, 2000. PENNELL, C.E., TURNER, A., PING, Y., MURRAY, H.G. Central and peripheral lactic acid changes during cord occlusion in healthy and compromised ovine fetuses. Perinatal Society of Australia and New Zealand 4th Annual Congress. Brisbane, Australia. March, 2000. PENNELL, C.E., NG, T., SMYTH, J.P., TURNER, A.J., COUGHTREY, H., YAN, P., MURRAY, H.G. Predictors of physiologic and histologic responses to repeated cord occlusion in the near term ovine fetus. Fetal and Neonatal Physiologic Society 28th Annual Meeting. Auckland, New Zealand. August, 2001. PENNELL, C.E., SMYTH, J.P., TURNER, A.J., COUGHTREY, H., YAN, P., DOHERTY, D.A., MURRAY, H.G., NEWNHAM, J.P. Fetal growth restriction produces discordance between peripheral and central acid base measurements in the ovine fetus. Perinatal Society of Australia and New Zealand 6th Annual Congress. Christchurch, New Zealand. March, 2002. PENNELL, C.E., NG, T., SMYTH, J.P., TURNER, A.J., COUGHTREY, H., YAN, P., MURRAY, H.G. Predictors of physiologic and histologic responses to repeated cord occlusion in the near term ovine fetus. Perinatal Society of Australia and New Zealand 6th Annual Congress. Christchurch, New Zealand. March, 2002. PENNELL, C.E., TRACY, M.B., DOHERTY, D.A., MURRAY, H.G. Fetal scalp lactate for monitoring fetal wellbeing in labour. Perinatal Society of Australia and New Zealand 6th Annual Congress. Christchurch, New Zealand. March, 2002. PENNELL, C.E., SMYTH, J.P., TURNER, A.J., COUGHTREY, H., YAN, P., DOHERTY, D.A., MURRAY, H.G., NEWNHAM, J.P. The effects of growth restriction on the fetal heart rate patterns and blood pressure responses to repeated cord occlusion in the ovine fetus. Perinatal Society of Australia and New Zealand 6th Annual Congress. Christchurch, New Zealand. March, 2002. SMYTH, J.P., PENNELL, C.E., COUGHTREY, H., TURNER, A.J., PING, Y., MURRAY, H.G., NEWNHAM, J.P. The effect of intrauterine growth restriction on lactate response during neonatal resuscitation following an asphyxial insult. Perinatal Society of Australia and New Zealand 6th Annual Congress. Christchurch, New Zealand. March, 2002. SMYTH, J.P., PENNELL, C.E., COUGHTREY, H., TURNER, A.J., PING, Y., MURRAY, H.G., NEWNHAM, J.P. The effect of intrauterine growth restriction on lactate response during neonatal resuscitation following an asphyxial insult. Society for Gynecologic Investigation 49th Annual Meeting. Los Angeles, California, United States of America. March, 2002. PENNELL, C.E., SMYTH, J.P., TURNER, A.J., COUGHTREY, H., YAN, P., DOHERTY, D.A., MURRAY, H.G., NEWNHAM, J.P. Fetal growth restriction produces discordance between peripheral and central acid base measurements in the ovine fetus. Society for Gynecologic Investigation 49th Annual Meeting. Los Angeles, California, United States of America. March, 2002.

xii

PENNELL, C.E., SMYTH, J.P., TURNER, A.J., COUGHTREY, H., YAN, P., DOHERTY, D.A., MURRAY, H.G., NEWNHAM, J.P. The effect of umbilico-placental embolisation on the fetal response to acute asphyxia. Federation of Placenta Associations 8th Meeting, Melbourne, Australia. October, 2002.

Awards

Perinatal Society of Australia and New Zealand New Investigator Award in the discipline of obstetrics. Perinatal Society of Australia and New Zealand 4th Annual Congress. Brisbane, Australia. March, 2000. Society for Gynecologic Investigation President’s Presenter Award and Abstract Number One in the Opening Plenary Session; Society for Gynecologic Investigation 49th Annual Meeting. Los Angeles, California, United States of America. March, 2002.

xiii

ABBREVIATIONS

AACB Australasian Association of Clinical Biochemists AAP American Academy of Pediatrics ACOG American College of Obstetricians and Gynecologists ACTH Adrenocorticotropic hormone AMPA a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid ANOVA Analysis of variance ATP Adenosine triphosphate BP Blood Pressure BPM Beats per minute C Control Ca2+ Calcium CI Confidence intervals CO2 Carbon dioxide CTG Cardiotocography CV Coefficient of Variation DNA Deoxyribonucleic acid E Embolised EFM Electronic fetal monitoring ECG Electrocardiogram EEG Electroencephalogram FADH Flavin adenine dinucleotide FBS Fetal blood sample FHR Fetal heart rate H:- Hydride ion H+ Hydrogen ion HIE Hypoxic-ischaemic encephalopathy IL-1β Interleukin-1beta IM Intramuscular IQR Inter quartile range IR Ischaemia-reperfusion IV Intravenous kg Kilogram LTV Long term variation MAC Minimal alveolar concentration MAP Mean arterial pressure MDA Malondialdehyde mmHg Millimetres of mercury mmol Millimole MMR Mean minute R-R interval MRI Magnetic resonance imaging MRS Magnetic resonance spectroscopy msec Millisecond n Number N Normal fetuses NADH Nicotine adenine dinucleotide NE Neonatal encephalopathy ng Nanograms NIRS Near infrared spectroscopy NMDA N-methyl-D-aspartate NNT Number needed to treat NO Nitric oxide NS Not significant O2 Oxygen OFR Oxygen free radical OHP Organic hydroperoxide

xiv

OR Odds Ratio PAF Platelet activating factor PB Passing Bablok PCO2 Partial pressure of carbon dioxide pg picograms PGE2 Prostaglandin E2 PO2 Partial pressure of oxygen PSANZ Perinatal Society of Australia and New Zealand PVL Periventricular leukomalacia RANZCOG Royal Australian and New Zealand College of Obstetricians and Gynaecologists RCOG Royal College of Obstetricians and Gynaecologists RCPA Royal College of Pathologists RCT Randomised clinical trial RIR Repetitive ischaemia-reperfusion ROC Receiver Operator Curve RR Relative risk SA Spontaneously acidaemic fetuses SC Subcutaneous SD Standard deviation SEM Standard error of the mean SOGC Society of Obstetricians and Gynaecologists of Canada SSS Superior sagittal sinus STV Short term variation TNF-α Tumour necrosis factor alpha UCO Umbilical cord occlusion XDH Xanthine dehydrogenase XO Xanthine oxidase

xv

TABLE OF CONTENTS

ABSTRACT IV DECLARATION VI ACKNOWLEDGEMENTS VII SPECIAL ACKNOWLEDGEMENT IX MANUSCRIPTS, ABSTRACTS AND AWARDS ARISING FROM THIS THESIS X

Manuscripts x Published Abstracts x

ABBREVIATIONS XIII TABLE OF CONTENTS XV TABLE OF FIGURES XXI TABLE OF TABLES XXIII CHAPTER 1: LITERATURE REVIEW AND THESIS OUTLINE 1

1.1 Introduction 1 1.1.1 The clinical problem of birth asphyxia.......................................................................................1 1.1.2 Chapter overview............................................................................................................................2

1.2 Birth asphyxia –aetiology and implications 3 1.2.1 Exploration of definitions.............................................................................................................3

1.2.1.1 Fetal distress........................................................................................................................3 1.2.1.2 Birth asphyxia and fetal asphyxia....................................................................................4 1.2.1.3 Neonatal encephalopathy.................................................................................................6 1.2.1.4 Hypoxic-ischaemic encephalopathy...............................................................................6 1.2.1.5 Cerebral palsy......................................................................................................................7

1.2.2 Historical perspective of birth asphyxia.....................................................................................8 1.2.3 Current view of fetal brain injury during labour ......................................................................9 1.2.4 Conclusion .................................................................................................................................... 15

1.3 Mechanisms of brain injury 16 1.3.1 The fetal response to asphyxia.................................................................................................. 16 1.3.2 Pattern and frequency of brain injury after asphyxia ........................................................... 19

1.3.2.1 Cerebral ischaemia models............................................................................................ 20 1.3.2.2 Global asphyxia models................................................................................................. 21 1.3.2.3 Partial asphyxia models.................................................................................................. 25 1.3.2.4 Summary of pattern and frequency of brain injury after asphyxia ....................... 27

1.3.3 Molecular mechanisms of perinatal brain injury ................................................................... 28 1.3.3.1 Biphasic energy failure ................................................................................................... 30 1.3.3.2 Pathogenesis of neuronal injury................................................................................... 30 1.3.3.3 The role of inflammatory mechanisms ...................................................................... 35

1.3.4 Summary of mechanisms of brain injury................................................................................ 37 1.4 Current techniques for intrapartum fetal surveillance 38

1.4.1 What are we aiming to prevent with intrapartum monitoring?.......................................... 38 1.4.1.1 Perinatal death, cerebral palsy and neurodevelopmental disability....................... 38 1.4.1.2 Neonatal encephalopathy.............................................................................................. 39 1.4.1.3 Neonatal seizures ............................................................................................................ 40

xvi

1.4.1.4 Severe metabolic acidosis .............................................................................................. 40 1.4.1.5 Summary........................................................................................................................... 42

1.4.2 Current monitoring techniques and their limitations........................................................... 43 1.4.2.1 Fetal heart rate monitoring ........................................................................................... 44

1.4.2.1.1 Historical perspective.......................................................................................... 44 1.4.2.1.2 Fetal heart rate monitoring and adverse clinical outcomes ......................... 45 1.4.2.1.3 Limitations of fetal heart rate monitoring....................................................... 48

1.4.2.2 Fetal scalp blood gas analysis ....................................................................................... 50 1.4.2.2.1 Physiologic basis for fetal scalp blood gas analysis ....................................... 51 1.4.2.2.2 Fetal blood sampling and adverse clinical outcomes.................................... 54 1.4.2.2.3 Limitations of fetal blood sampling ................................................................. 55 1.4.2.2.4 Scalp stimulation as an alternative to fetal blood sampling......................... 57

1.4.2.3 Umbilical cord blood gas analysis................................................................................ 58 1.4.2.3.1 Physiological basis................................................................................................ 58 1.4.2.3.2 Prediction of adverse outcomes........................................................................ 58 1.4.2.3.3 Limitations of umbilical cord blood gas analysis........................................... 59

1.4.3 Summary........................................................................................................................................ 59 1.5 Lactate and fetal surveillance 61

1.5.1 Lactate physiology ....................................................................................................................... 61 1.5.1.1 Lactate production.......................................................................................................... 61

1.5.1.1.1 Hypoxic conditions ............................................................................................. 61 1.5.1.1.2 Non-hypoxic conditions..................................................................................... 62

1.5.1.2 Fetal lactate levels during pregnancy and delivery ................................................... 63 1.5.1.3 Negative effects of lactate ............................................................................................. 64 1.5.1.4 Positive effects of lactate............................................................................................... 64

1.5.2 Fetal scalp lactate ......................................................................................................................... 65 1.5.3 Umbilical cord lactate levels at delivery .................................................................................. 68 1.5.4 Lactate levels in the newborn.................................................................................................... 68

1.6 Thesis outline 69 1.6.1 Aim ................................................................................................................................................. 69 1.6.2 Studies performed ....................................................................................................................... 69

1.6.2.1 Evaluation of a new method for rapid assessment of lactate in fetal and neonatal blood................................................................................................................. 69

1.6.2.2 The relationship between neuropathologic damage and biochemical and physiologic markers of asphyxia induced by graded cord occlusion in chronically instrumented ovine fetuses. ..................................................................... 69

1.6.2.3 The relationship between fetal scalp lactate levels, central lactate levels and oxygen free radical activity in the fetal brain during asphyxia induced by graded cord occlusion in the ovine fetus ................................................................... 70

1.6.2.4 Fetal lactate responses in growth restricted ovine fetuses during asphyxia induced by graded cord occlusion............................................................................... 70

CHAPTER 2: EVALUATION OF A NEW METHOD FOR RAPID ASSESSMENT OF LACTATE IN FETAL AND NEONATAL BLOOD 72

2.1 Introduction 72 2.1.1 Aim ................................................................................................................................................. 74

2.2 Methods 75 2.2.1 Measuring principle ..................................................................................................................... 75

2.2.1.1 Accutrend® meter.......................................................................................................... 75 2.2.1.2 Reference method........................................................................................................... 75

2.2.2 Assessment of precision............................................................................................................. 75 2.2.3 Assessment of accuracy.............................................................................................................. 79

2.2.3.1 Neonatal plasma samples .............................................................................................. 79 2.2.3.2 Umbilical cord blood samples...................................................................................... 79 2.2.3.3 Neonatal arterial blood samples................................................................................... 79 2.2.3.4 Effect of temperature on accuracy.............................................................................. 79

2.2.4 Statistical methods ....................................................................................................................... 80 2.3 Results 82

2.3.1 Analysis of precision ................................................................................................................... 82

xvii

2.3.2 Analysis of accuracy using neonatal plasma samples ........................................................... 83 2.3.3 Analysis of accuracy using umbilical cord blood samples................................................... 83 2.3.4 Analysis of accuracy using neonatal arterial samples............................................................ 85 2.3.5 Effect of temperature on accuracy........................................................................................... 89

2.4 Discussion 91 2.4.1 Reference method precision and accuracy ............................................................................. 91 2.4.2 Accutrend® handheld lactate meter precision and accuracy ............................................. 91 2.4.3 Temperature and lactate measurement ................................................................................... 92 2.4.4 Haematocrit and lactate measurement .................................................................................... 94 2.4.5 Lactate measurement bias.......................................................................................................... 95

2.5 Conclusion 96 CHAPTER 3: THE RELATIONSHIP BETWEEN NEUROPATHOLOGIC DAMAGE AND BIOCHEMICAL AND PHYSIOLOGIC MARKERS OF ASPHYXIA INDUCED BY GRADED CORD OCCLUSION IN CHRONICALLY CANNULATED OVINE FETUSES 97

3.1 Introduction 97 3.1.1 Aim ............................................................................................................................................... 100 3.1.2 Hypothesis .................................................................................................................................. 100

3.2 Methods 101 3.2.1 Surgical procedure ..................................................................................................................... 101 3.2.2 Cord occlusion protocol .......................................................................................................... 103 3.2.3 Fetal heart rate assessment ...................................................................................................... 103 3.2.4 Fetal arterial pressure assessment........................................................................................... 104 3.2.5 Umbilical blood flow ................................................................................................................ 104 3.2.6 Biochemical assessment ........................................................................................................... 105 3.2.7 Delivery and neuropathologic examination ......................................................................... 105 3.2.8 Statistical methods ..................................................................................................................... 106

3.3 Results 108 3.3.1 Biochemical responses to repeated cord occlusion ............................................................ 108

3.3.1.1 Fetal arterial blood gas and lactate responses during umbilical cord occlusion......................................................................................................................... 108

3.3.1.2 Fetal glucose, cortisol and catecholamine responses during umbilical cord occlusion......................................................................................................................... 110

3.3.1.3 Fetal haemoglobin response to experimental protocol......................................... 110 3.3.1.4 Biochemical prediction of fetal response to umbilical cord occlusion .............. 113

3.3.2 Fetal heart rate responses to repeated cord occlusion ....................................................... 113 3.3.2.1 Baseline FHR response to repeated cord occlusion.............................................. 113 3.3.2.2 Fetal heart rate variability response to repeated cord occlusion ......................... 113 3.3.2.3 Fetal heart rate patterns during repeated cord occlusion ..................................... 117

3.3.2.3.1 Variable fetal heart rate decelerations ............................................................ 117 3.3.2.3.2 Overshoot after variable fetal heart rate decelerations............................... 117 3.3.2.3.3 Late components to variable fetal heart rate decelerations ....................... 117 3.3.2.3.4 Increased fetal heart rate variability between cord occlusions.................. 123 3.3.2.3.5 Complex variable fetal heart rate decelerations ........................................... 123

3.3.3 Fetal arterial pressure responses to repeated cord occlusion............................................ 123 3.3.3.1 Fetal arterial pressure between cord occlusions ..................................................... 123 3.3.3.2 Fetal arterial pressure during cord occlusion .......................................................... 125

3.3.4 Neuropathologic findings after repeated cord occlusion .................................................. 125 3.3.5 Relationship between markers of fetal well-being during repeated cord occlusion

and fetal neuropathologic damage ......................................................................................... 127 3.3.5.1 Correlation of biochemical markers with fetal neuropathologic damage ......... 127 3.3.5.2 Correlation of fetal heart rate patterns with fetal neuropathologic damage..... 131 3.3.5.3 Correlation of fetal arterial pressure with fetal neuropathologic damage ......... 132

3.3.5.3.1 Fetal arterial pressure between the umbilical cord occlusions.................. 132 3.3.5.3.2 Fetal arterial pressure during the final three occlusions............................. 132 3.3.5.3.3 Duration of fetal arterial pressure responses during umbilical cord

occlusion .............................................................................................................. 133

xviii

3.3.6 Prediction of fetal arterial pressure responses during cord occlusion ............................ 133 3.3.6.1 Prediction of arterial pressure responses during umbilical cord occlusion

from fetal heart rate patterns ...................................................................................... 134 3.3.6.2 Prediction of arterial pressure responses during umbilical cord occlusion

from fetal biochemical measures ............................................................................... 134 3.4 Discussion 138

3.4.1 Cord occlusion model............................................................................................................... 138 3.4.2 Histopathological changes induced by cord occlusion model ......................................... 138 3.4.3 Fetal acid-base status and neuropathologic damage........................................................... 139 3.4.4 Fetal lactate and neuropathologic damage ........................................................................... 142 3.4.5 Fetal glucose and neuropathologic damage ......................................................................... 144 3.4.6 Fetal heart rate patterns and neuropathologic damage...................................................... 145

3.4.6.1 Increased fetal heart variability and neuropathologic damage............................. 145 3.4.6.2 Fetal heart rate overshoot and neuropathologic damage ..................................... 146 3.4.6.3 Other fetal heart rate patterns and neuropathologic damage .............................. 148

3.4.7 Fetal arterial pressure and neuropathologic damage .......................................................... 149 3.4.8 Prediction of rate of response to repeated umbilical cord occlusion ............................. 151 3.4.9 Integration of results into clinical practice ........................................................................... 152 3.4.10 Future studies ............................................................................................................................. 153

3.5 Conclusion 155 CHAPTER 4: THE RELATIONSHIPS BETWEEN FETAL SCALP LACTATE LEVELS, CENTRAL LACTATE LEVELS AND OXYGEN FREE RADICAL ACTIVITY IN THE FETAL BRAIN DURING ASPHYXIA INDUCED BY GRADED CORD OCCLUSION IN THE OVINE FETUS 156

4.1 Introduction 156 4.1.1 Development of an experimental model .............................................................................. 157 4.1.2 The relationship between fetal scalp lactate levels and central lactate levels................. 158 4.1.3 The relationship between fetal scalp lactate levels and fetal brain injury....................... 160

4.1.3.1 Oxygen free radical activity and fetal brain injury.................................................. 160 4.1.4 Aim ............................................................................................................................................... 163 4.1.5 Hypothesis .................................................................................................................................. 164

4.2 Methods 165 4.2.1 Surgical Procedure..................................................................................................................... 165 4.2.2 Cord occlusion protocol .......................................................................................................... 166 4.2.3 Fetal heart rate and arterial pressure assessment................................................................. 167 4.2.4 Umbilical blood flow ................................................................................................................ 167 4.2.5 Biochemical assessment ........................................................................................................... 167 4.2.6 Study group classification......................................................................................................... 168 4.2.7 Statistical methods ..................................................................................................................... 169

4.3 Results 171 4.3.1 Baseline fetal biochemical status prior to the experimental phase .................................. 171 4.3.2 Investigation of the validity of the acutely catheterised model for the assessment

of fetal scalp blood samples..................................................................................................... 172 4.3.2.1 Fetal acid-base status and lactate responses ............................................................ 172 4.3.2.2 Fetal endocrine responses ........................................................................................... 172 4.3.2.3 Fetal heart rate responses............................................................................................ 175 4.3.2.4 Fetal arterial pressure responses ................................................................................ 175

4.3.3 Biochemical responses of normal (non-acidaemic) and spontaneously acidaemic fetuses to repeated umbilical cord occlusion ....................................................................... 178

4.3.3.1 Controls .......................................................................................................................... 178 4.3.3.2 Fetal acid-base status and lactate responses ............................................................ 178 4.3.3.3 Fetal endocrine responses ........................................................................................... 178

4.3.4 The relationship between fetal scalp blood lactate levels and central lactate levels..... 181 4.3.4.1 Fetal scalp blood lactate levels and carotid arterial lactate levels ........................ 181 4.3.4.2 Fetal scalp blood lactate levels and jugular venous lactate levels ........................ 186 4.3.4.3 Fetal carotid artery and jugular vein blood lactate levels ...................................... 186

4.3.5 The relationship between fetal scalp blood lactate and lipid peroxidation in the fetal brain..................................................................................................................................... 187

xix

4.3.5.1 Controls .......................................................................................................................... 187 4.3.5.2 Normal and spontaneously acidaemic fetuses ........................................................ 187 4.3.5.3 Prediction of the lipid peroxidation in fetal brain perfusate ................................ 190

4.4 Discussion 194 4.4.1 Comparison of the acute and chronic instrumented models ........................................... 194

4.4.1.1 Selection of maternal anaesthesia .............................................................................. 195 4.4.1.2 Similarities and differences between the acute and chronically

instrumented models.................................................................................................... 199 4.4.1.2.1 Control fetuses ................................................................................................... 199 4.4.1.2.2 Fetuses subjected to repeated cord occlusion.............................................. 199

4.4.1.3 What is the optimum model for assessing fetal scalp blood during asphyxia? ......................................................................................................................... 201

4.4.2 The relationship between fetal scalp blood, carotid arterial and jugular venous blood lactate levels..................................................................................................................... 202

4.4.2.1 Scalp lactate and central lactate levels in control and normal fetuses during asphyxia........................................................................................................................... 202

4.4.2.2 Information gained from spontaneously acidaemic fetuses................................. 206 4.4.3 Fetal scalp blood lactate and lipid peroxidation in the fetal brain ................................... 208

4.4.3.1 Markers of lipid peroxidation..................................................................................... 208 4.4.3.2 Patterns of lipid peroxidation during asphyxia ....................................................... 209 4.4.3.3 The relationship between fetal scalp lactate and lipid peroxidation in the

fetal brain........................................................................................................................ 212 4.4.4 Future studies ............................................................................................................................. 213

4.5 Conclusion 215 CHAPTER 5: FETAL LACTATE RESPONSES IN THE GROWTH RESTRICTED OVINE FETUS DURING ASPHYXIA INDUCED BY GRADED CORD OCCLUSION 217

5.1 Introduction 217 5.1.1 Aim ............................................................................................................................................... 220 5.1.2 Hypothesis .................................................................................................................................. 220

5.2 Methods 221 5.2.1 First surgical preparation.......................................................................................................... 221 5.2.2 Embolisation protocol.............................................................................................................. 223 5.2.3 Second surgical preparation..................................................................................................... 223 5.2.4 Cord occlusion protocol .......................................................................................................... 224 5.2.5 Fetal heart rate, arterial pressure and umbilical blood flow assessment......................... 225 5.2.6 Biochemical methods................................................................................................................ 225 5.2.7 Statistical methods ..................................................................................................................... 225

5.3 Results 228 5.3.1 Physiologic responses of the growth restricted fetus during umbilical cord

occlusion...................................................................................................................................... 228 5.3.1.1 Umbilicoplacental embolisation................................................................................. 228 5.3.1.2 Fetal acid-base responses to repeated cord occlusion........................................... 230 5.3.1.3 Fetal arterial lactate response to repeated cord occlusion .................................... 233 5.3.1.4 Fetal endocrine responses to repeated cord occlusion ......................................... 233 5.3.1.5 Fetal heart rate responses to repeated cord occlusion .......................................... 236

5.3.1.5.1 Baseline fetal heart rate response to repeated cord occlusion .................. 236 5.3.1.5.2 Fetal heart rate variability response to repeated cord occlusion .............. 236 5.3.1.5.3 Fetal heart rate patterns during repeated cord occlusion........................... 236

5.3.1.6 Fetal arterial pressure responses to repeated cord occlusion............................... 239 5.3.1.6.1 Fetal arterial pressure between cord occlusions .......................................... 239 5.3.1.6.2 Fetal arterial pressure during cord occlusions.............................................. 243

5.3.1.7 Physiologic associations with fetal heart rate patterns in embolised and non-embolised fetuses ................................................................................................. 243

5.3.1.7.1 Fetal heart rate accelerations............................................................................ 248 5.3.1.7.2 Late component to decelerations ................................................................... 249 5.3.1.7.3 Severe variable decelerations. .......................................................................... 249 5.3.1.7.4 Increased fetal heart rate variability................................................................ 250

xx

5.3.2 The effect of umbilicoplacental embolisation on the relationship between fetal scalp blood levels and central levels of lactate and pH...................................................... 251

5.3.2.1 Fetal scalp blood lactate levels ................................................................................... 251 5.3.2.1.1 Fetal scalp blood lactate levels and carotid arterial lactate levels ............. 251 5.3.2.1.2 Fetal scalp blood lactate levels and jugular venous lactate levels ............. 255 5.3.2.1.3 Fetal carotid artery and jugular venous blood lactate levels...................... 255

5.3.2.2 Fetal scalp blood pH levels ......................................................................................... 258 5.3.2.2.1 Fetal scalp blood pH levels and carotid arterial pH levels ........................ 258 5.3.2.2.2 Fetal scalp blood pH levels and jugular venous blood pH levels ............ 262 5.3.2.2.3 Fetal carotid artery and jugular venous blood pH levels ........................... 262

5.3.3 Prediction of fetal arterial pressure changes during cord occlusion................................ 263 5.3.3.1 Prediction of fetal arterial pressure responses during cord occlusion from

fetal heart rate patterns ................................................................................................ 263 5.3.3.2 Prediction of fetal arterial pressure responses during cord occlusion from

fetal scalp blood biochemistry.................................................................................... 265 5.3.4 Summary of similarities and differences in fetal responses to asphyxia in

embolised and non-embolised fetuses .................................................................................. 267 5.4 Discussion 268

5.4.1 Stability of fetal lactate responses during asphyxia ............................................................. 268 5.4.1.1 Arterial lactate responses............................................................................................. 269 5.4.1.2 Fetal scalp blood lactate responses............................................................................ 272 5.4.1.3 Lactate and the prediction of arterial pressure responses during cord

occlusion......................................................................................................................... 273 5.4.2 Variation in cardiovascular and endocrine responses during asphyxia .......................... 274

5.4.2.1 Fetal heart rate patterns ............................................................................................... 274 5.4.2.1.1 Fetal heart rate variability ................................................................................. 275 5.4.2.1.2 Fetal heart rate decelerations ........................................................................... 277 5.4.2.1.3 Fetal heart rate accelerations and overshoot ................................................ 278

5.4.2.2 Fetal arterial pressure responses ................................................................................ 279 5.4.2.3 Fetal endocrine responses ........................................................................................... 280

5.4.3 Influence of the experimental model of fetal scalp blood biochemistry........................ 281 5.4.4 Integration of results into clinical practice ........................................................................... 282 5.4.5 Future studies ............................................................................................................................. 283

5.5 Conclusion 285 CHAPTER 6: CLINICAL APPLICATION: IS THERE EVIDENCE FOR CHANGE? 286 REFERENCES 295

xxi

TABLE OF FIGURES

Number Page 1.1 Fetal pathophysiology after an asphyxial exposure ................................................................................ 29 1.2 Hypoxia-ischaemia induced energy depletion and cascade to neuronal death ................................. 33 1.3 Pathways of hypoxia-mediated neuronal cell death................................................................................ 34 1.4 Fetal acidosis with abnormal fetal heart rate patterns............................................................................ 52 1.5 Fetal aerobic and anaerobic metabolism................................................................................................... 53 2.1 Accutrend® handheld lactate meter.......................................................................................................... 76 2.2 Design of lactate test strip for the Accutrend® handheld lactate meter............................................ 76 2.3 Reaction principles used for lactate measurement by the Accutrend® and ABL625®................. 77 2.4 Analysis technique for the Accutrend® handheld lactate meter ......................................................... 78 2.5 Accuracy of the Accutrend® lactate meter for measurement of lactate in neonatal plasma......... 84 2.6 Accuracy of the Accutrend® lactate meter for measurement of lactate in umbilical arterial

blood ................................................................................................................................................................ 86 2.7 The effect of haematocrit on accuracy of the Accutrend lactate meter ............................................. 87 2.8 Accuracy of the Accutrend® lactate meter in neonatal arterial blood ............................................... 88 2.9 The effect of temperature on the accuracy of the Accutrend® lactate meter .................................. 90 3.1 Experimental timeline and cord occlusion protocol ............................................................................ 102 3.2 Fetal arterial gas tensions, pH, base excess, lactate, and glucose over the seven day study

period ............................................................................................................................................................. 109 3.3 Fetal cortisol, noradrenaline, adrenaline and dopamine concentrations over the seven day

study period .................................................................................................................................................. 111 3.4 Fetal haemoglobin responses over seven day study period................................................................ 112 3.5 Biochemical prediction of response to repeated umbilical cord occlusion ..................................... 114 3.6 Baseline fetal heart rate during the experimental phase....................................................................... 115 3.7 Fetal heart rate variability during the experimental phase................................................................... 116 3.8 Variable fetal heart rate decelerations...................................................................................................... 118 3.9 Overshoot after fetal heart rate decelerations........................................................................................ 119 3.10 Late components to fetal heart rate decelerations ................................................................................ 120 3.11 Increased FHR variability between cord occlusions ............................................................................ 121 3.12 Complex variable fetal heart rate decelerations..................................................................................... 122 3.13 Fetal arterial pressure responses between repeated cord occlusion .................................................. 124 3.14 Fetal arterial pressure responses during repeated cord occlusion ..................................................... 126 3.15 Range of histopathologic damage to fetal brains .................................................................................. 129 3.16 Receiver operator curves for biochemical prediction of blood pressure patterns during

cord occlusion .............................................................................................................................................. 137 4.1 Adverse effects of oxidative stress on neuronal lipids, protein and, deoxyribonucleic acid........ 161 4.2 Comparison of arterial gas tensions, pH, base excess and lactate levels in acute and

chronically catheterized ovine fetuses during repeated umbilical cord occlusion.......................... 174 4.3 Comparison of endocrine responses in acute and chronically catheterized ovine fetuses

during repeated umbilical cord occlusion............................................................................................... 173 4.4 Comparison of FHR and FHR variability responses in acute and chronically instrumented

fetuses during repeated umbilical cord occlusion ................................................................................. 176 4.5 Comparison of inter-occlusion fetal blood pressure responses in acute and chronically

instrumented fetuses during repeated cord occlusion.......................................................................... 177 4.6 Fetal arterial gas tensions, pH, base excess and lactate in normal and spontaneously

acidaemic fetuses during experimental phase ........................................................................................ 179 4.7 Fetal endocrine responses to repeated cord occlusion in normal and spontaneously

acidaemic fetuses.......................................................................................................................................... 180 4.8 The relationship between fetal scalp blood, carotid arterial blood and jugular vein blood

lactate levels in controls.............................................................................................................................. 182 4.9 The relationship between fetal scalp blood, carotid arterial blood and jugular vein blood

lactate levels in normal fetuses during repeated cord occlusion ........................................................ 183

xxii

4.10 The relationship between fetal scalp blood, carotid arterial blood and jugular vein blood lactate levels in spontaneously acidaemic fetuses during repeated cord occlusion ........................ 184

4.11 Comparison of three lactate sampling sites in control, normal and spontaneously acidaemic fetuses.......................................................................................................................................... 185

4.12 Lipid peroxidation in carotid arterial and jugular venous blood of control fetuses....................... 188 4.13 Lipid peroxidation in carotid arterial and jugular venous blood of fetuses that experienced

umbilical cord occlusion............................................................................................................................. 189 4.14 Range of carotid arterial pH and scalp blood lactate associated with lipid peroxidation in

jugular venous blood................................................................................................................................... 191 4.15 Receiver operator curves for the prediction of OHP production in jugular venous blood ........ 192 4.16 Control of jugular venous and scalp capillary blood content............................................................. 204 4.17 Mechanism of oxygen free radical production with umbilical cord occlusion ............................... 210 5.1 Outline of experimental protocol............................................................................................................. 222 5.2 Fetal acid-base responses during umbilicoplacental embolisation .................................................... 229 5.3 Fetal arterial lactate responses during repeated umbilical cord occlusion ...................................... 231 5.4 Fetal arterial lactate responses during repeated umbilical cord occlusion ...................................... 232 5.5 Fetal endocrine responses during umbilical cord occlusion ............................................................... 234 5.6 Fetal ACTH and cortisol responses during umbilical cord occlusion.............................................. 235 5.7 Fetal heart rate responses during umbilical cord occlusion ................................................................ 237 5.8 Fetal heart rate responses in two groups of fetuses exposed to repeated cord occlusion............ 238 5.9 Fetal heart rate decelerative patterns during umbilical cord occlusion............................................. 240 5.10 Fetal systolic pressure responses between umbilical cord occlusions .............................................. 241 5.11 Fetal diastolic pressure responses between umbilical cord occlusions............................................. 242 5.12 Fetal arterial pressure responses during umbilical cord occlusions................................................... 244 5.13 The relationship between fetal scalp blood, carotid arterial blood and jugular vein blood

lactate levels in control fetuses.................................................................................................................. 252 5.14 The relationship between fetal scalp blood, carotid arterial blood and jugular vein blood

lactate levels in occluded fetuses .............................................................................................................. 253 5.15 Comparison of three lactate sampling sites in control and occluded non-embolised and

embolised fetuses......................................................................................................................................... 254 5.16 Lactate and oxygen differences across the fetal cerebral circulation during cord occlusion ....... 257 5.17 The relationship between fetal scalp blood, carotid arterial blood and jugular vein blood

pH levels in control fetuses ....................................................................................................................... 259 5.18 The relationship between fetal scalp blood, carotid arterial blood and jugular vein blood

pH levels in occluded fetuses.................................................................................................................... 260 5.19 Comparison of three pH sampling sites in control and occluded non-embolised and

embolised fetuses......................................................................................................................................... 261 5.20 Receiver operator curves for the prediction of any blood pressure instability during cord

occlusion in non-embolised and embolised fetuses ..................................................................... 266

xxiii

TABLE OF TABLES

Number Page 1.1 Classification of intrapartum fetal asphyxia ................................................................................................5 1.2 Criteria to define an acute intrapartum hypoxic event........................................................................... 11 1.3 Cardiovascular responses to an asphyxial exposure ............................................................................... 17 1.4 Four patterns of perinatal brain damage................................................................................................... 20 1.5 Brain damage after cerebral ischaemia ...................................................................................................... 21 1.6 Brain damage after global asphyxia............................................................................................................ 22 1.7 Brain damage after partial fetal asphyxia .................................................................................................. 25 1.8 Important adverse clinical outcomes......................................................................................................... 38 1.9 Short term neonatal complications of severe metabolic acidosis at birth.......................................... 41 1.10 Scalp pH and the prediction of neonatal outcome................................................................................. 55 1.11 Normal ranges for scalp lactate in first stage of labour ......................................................................... 66 1.12 Scalp lactate and the prediction of neonatal outcome ........................................................................... 68 2.1 Precision of the Accutrend® Reflectance Photometer ......................................................................... 82 2.2 Precision of the Radiometer ABL 625® blood-gas machine............................................................... 82 2.3 Characteristics of 160 cord samples utilized in accuracy analysis........................................................ 83 2.4 Normal reference range for umbilical arterial lactate after labour....................................................... 95 3.1 Outcome after term neonatal encephalopathy ........................................................................................ 99 3.2 Grading of histopathologic changes in fetal brains .............................................................................. 106 3.3 Histopathologic damage in right side and left side of fetal brain ...................................................... 128 3.4 Spearman’s Rho for fetal biochemical markers and the grade of fetal brain injury....................... 130 3.5 Spearman’s Rho for the duration of FHR patterns and the grade of fetal brain injury................ 131 3.6 Spearman’s Rho for the duration of compound FHR patterns and the grade of fetal

brain injury .................................................................................................................................................... 131 3.7 Spearman’s Rho for arterial pressure at the completion of the cord occlusion protocol

and the grade of fetal brain injury ............................................................................................................ 132 3.8 Spearman’s Rho for the duration of arterial pressure instability during UCO and the

grade of fetal brain injury ........................................................................................................................... 133 3.9 Fetal heart rate prediction of arterial pressure responses during cord occlusion........................... 134 3.10 Biochemical prediction of arterial pressure responses during cord occlusion................................ 135 3.11 Accuracy of biochemical prediction of arterial pressure responses during cord occlusion ......... 136 4.1 Baseline fetal biochemical condition prior experimental phase......................................................... 171 4.2 Accuracy of biochemical prediction of OHP production in brain perfusate ................................. 193 4.3 Comparison of maternal anaesthesia techniques between experimental models........................... 195 5.1 Effect of 23 days of umbilicoplacental embolisation on fetal biochemical and endocrine

status............................................................................................................................................................... 230 5.2 Univariate associations of FHR patterns with blood gas, blood pressure and

catecholamine variables .............................................................................................................................. 245 5.3 Multivariate associations of FHR patterns considering all blood gas variables and fetal

arterial pressure variables simultaneously ............................................................................................... 246 5.4 Multivariate associations of FHR patterns considering all blood gas, fetal blood pressure

and catecholamine variables simultaneously .......................................................................................... 247 5.5 Receiver operator curve analysis for the prediction of FHR patterns from fetal

physiologic parameters............................................................................................................... 248 5.6 Summary of differences between superior sagittal sinus blood and jugular

venous blood ............................................................................................................................... 256 5.7 Fetal heart rate prediction of arterial pressure responses during cord occlusion ........... 263 5.8 Odds Ratios of abnormal FHR pattern if fetal arterial pressure response is

present........................................................................................................................................... 264 5.9 Biochemical prediction of arterial pressure responses during cord occlusion ................ 265 5.10 Accuracy of biochemical prediction of arterial pressure responses during UCO .......... 267 5.11 Summary of differences and similarities in physiologic responses to asphyxia in

embolised and non-embolised fetuses .................................................................................... 267

1

C h a p t e r 1

LITERATURE REVIEW AND THESIS OUTLINE

At the outset do not be worried about the big question – Truth.

No human being is constituted to know the truth, the whole truth, and nothing but the truth;

And even the best of men must be content with fragments, with partial glimpses, never the full fruition.

Sir William Osler (1892)

1.1 Introduction

1.1.1 The clinical problem of birth asphyxia

Worldwide there are approximately 130 million births annually. Of these infants, it has been

estimated that four million babies suffer from birth asphyxia, one million die and a similar

number develop sequelae including cerebral palsy, seizures, or developmental delay (WHO,

1991, 1995). The incidence of birth asphyxia is higher in developing than in so-called

developed countries yet in the latter there are still between two and six neonates per

thousand who exhibit hypoxic-ischaemic encephalopathy (Levene et al., 1985). This amounts

to between 8,000 and 25,000 infants each year in both the United States of America and the

European Union and between 500 and 1,500 infants each year in Australia. Birth asphyxia

therefore represents an important, potentially preventable condition that has major personal,

social and financial implications for both families and the community worldwide.

The ultimate fear of every obstetrician, midwife and parent is a poor neonatal outcome. Fetal

distress is a widely used but poorly defined term (Steer, 1982; Parer & Livingston, 1990). Its

use commonly indicates anxiety about the fetal condition which then leads to a caesarean

section or instrumental delivery. The aim of intrapartum fetal surveillance is to more

accurately diagnose fetal distress to reduce the incidence of birth asphyxia.

The three techniques that are currently widely used for the assessment of fetal wellbeing in

labour are electronic fetal heart rate monitoring, scalp blood pH measurement and umbilical

cord blood gas analysis. Fetal heart rate (FHR) monitoring was introduced into clinical use

and patient management more than three decades ago. It continues to be the predominant

2

method for intrapartum fetal surveillance despite questions about its efficacy and outcomes

associated with its use (Freeman, 2002). It is generally accepted that both continuous

electronic FHR monitoring and intensive intermittent auscultatory monitoring are better than

non-intensive auscultatory monitoring in the prevention of intrapartum fetal death.

Nonetheless, despite three decades of immense changes in obstetric management, including

the widespread use of fetal surveillance tests and increased caesarean deliveries for fetal

indications, the incidence of cerebral palsy has remained constant (Scheller & Nelson, 1994;

MacLennan, 1999).

The principal biochemical markers of perinatal asphyxia are intrapartum fetal scalp acid-base

assessment and the assessment of umbilical cord acid-base status at delivery. Both these

techniques have a sound physiologic basis and umbilical cord acid-base assessment at delivery

provides the most sensitive reflection of birth asphyxia based on current published literature

(MacLennan, 1999). Unfortunately the use of these traditional markers of fetal acid-base

status is limited because they reflect fetal metabolic adjustments that are not directly related

to fetal damage unless the values are extreme (Gilstrap et al., 1989; Goldaber et al., 1991;

Winkler et al., 1991). Even a pH of less than 7.0 is poorly predictive of neurological deficit

because a severely acidotic infant will usually have a normal outcome (Winkler et al., 1991;

Socol et al., 1994; Yudkin et al., 1994).

In order to reduce the burden of birth asphyxia on families, health care systems and the

community, new techniques that are inexpensive and widely available are required to allow

recognition of early intrapartum asphyxia so that timely obstetric intervention can avoid

asphyxia-induced brain damage or death. Furthermore, new techniques are required to

identify which fetuses exposed to an asphyxial insult have acute brain injury so that

appropriate cases can be selected for therapies designed to limit the extension of this injury

that can occur after birth (Edwards & Azzopardi, 2000).

1.1.2 Chapter overview

In this chapter I will review birth asphyxia, its aetiology and its implications for the newborn

and child. This will be followed by a review of the animal studies of fetal asphyxia addressing

the fetal response to asphyxia, the pattern and frequency of brain injury after asphyxial

exposure, and the molecular mechanisms of brain injury. The third part of this chapter will

review our current markers of fetal wellbeing during labour with particular emphasis on their

ability to predict important outcomes. The final part of this chapter will review the role of

lactate measurement during intrapartum surveillance and outline the studies that I have

performed to expand knowledge in this field.

3

1.2 Birth asphyxia –aetiology and implications

1.2.1 Exploration of definitions

There are a number of important definitions that should be established at this point as they

have specific meanings and should not be interchanged.

1.2.1.1 Fetal distress

Fetal distress is a widely used but poorly defined term (Steer, 1982; Parer & Livingston,

1990). It is often prefixed by the term acute or chronic and it is commonly used as an

indication for caesarean section or as a justification for instrumental delivery. Some authors

have defined it based on FHR patterns such as “late decelerations”, “severe variable

decelerations”, or “tachycardia with total loss of short-term variability” (Haverkamp et al.,

1979; Haesslein & Niswander, 1980). Others believe that fetal acidosis or depressed Apgar

scores are necessary to define fetal distress (Sykes et al., 1982). Some authors suggest that the

presence of meconium-stained amniotic fluid and any abnormalities in the fetal heart rate

define fetal distress warranting delivery (Steer, 1999).

In 1893, Von Winckel described the criteria for fetal distress that were to remain essentially

unchanged until the arrival of fetal scalp blood sampling for pH measurement and electronic

heart rate monitoring (Von Winckel, 1893). These criteria included tachycardia (FHR more

than 160bpm), bradycardia (FHR less than 100bpm), irregular fetal heart rate, passage of

meconium and gross alteration of fetal movement. Few studies either challenged or

supported the validity of these auscultative and clinical criteria for fetal distress. It was not

until 1968, when the results of the Collaborative Project commissioned by the National

Institute of Neurologic Diseases and Blindness were published, that these criteria came under

serious scrutiny (Benson et al., 1968). This study reviewed the benefits of FHR monitoring in

the management of intrapartum fetal distress in 24,683 deliveries and concluded that there

was no reliable indicator of fetal distress in terms of fetal heart rate save in extreme degree

(Benson et al., 1968).

Parer et al. (1990) attempted to define fetal distress in terms of the fetal physiologic

adaptations that occur with decreased oxygenation. They defined fetal distress as persistent

fetal asphyxia that if not corrected or circumvented, will result in decompensation of the

physiologic responses (primarily redistribution of blood flow to preserve oxygenation of vital

organs) and cause permanent central nervous system damage or death. This definition,

4

although physiologically accurate, is difficult to apply clinically. The first half of the

definition (persistent fetal asphyxia) may possibly be determined from continuous FHR

monitoring through the identification of late, variable or prolonged decelerations; however,

the second half of the definition implies identifying fetal decompensation which is more

difficult (Quilligan, 1990). Some authors have proposed that this can be done through the

assessment of FHR variability and say that as long as variability is present, there is no need to

intervene (Parer & Livingston, 1990). Others have contended that if one waits until

variability is lost it may be too late (Goodlin, 1990; Freeman, 2002).

A more appropriate terminology for describing fetal condition during labour is “reassuring”

and “non-reassuring” fetal status as clinical signs often poorly predict a compromised fetus.

This terminology has been endorsed by the Perinatal Society of Australia and New Zealand

(PSANZ, 1995; MacLennan, 1999), the American College of Obstetricians and

Gynecologists (ACOG, 1998) and the Society of Obstetricians and Gynaecologists of Canada

(SOGC, 1996b) and it is stated by these three professional bodies that the continued use of

the term fetal distress may encourage incorrect assumptions or inappropriate management.

1.2.1.2 Birth asphyxia and fetal asphyxia

Asphyxia can literally be translated from the Greek as meaning absence of pulse. It is more

usually applied to situations where there has been anoxia or hypoxia. A task force set up by

the World Federation of Neurology defined asphyxia as a condition of impaired gas exchange

leading, if it persists, to progressive hypoxaemia and hypercapnia (Bax & Nelson, 1993). This

is a good definition of asphyxia, however, in labour asphyxia may occur intermittently.

Therefore, for fetal asphyxia during labour, a more appropriate definition is “a condition of

impaired blood gas exchange leading to progressive fetal hypoxaemia and hypercapnia with a

significant metabolic acidosis” (Low, 1998; MacLennan, 1999).

In the clinical context, Thorp et al. (1990) suggested that birth asphyxia should be defined to

include not only acidaemia (pH less than two standard deviations below the mean; <7.10) but

also neonatal depression and evidence of hypoxic end-organ damage such as early neonatal

seizures and renal or cardiac dysfunction (Thorp, 1990).

5

The American Academy of Pediatrics and the American College of Obstetrics and

Gynecology (AAP & ACOG, 1996) recommended that the diagnosis of birth asphyxia

should be based on the finding of three of the four following criteria:

1. Umbilical arterial pH<7.00

2. Apgar score <4 for more than 5 minutes

3. Multiple organ damage

4. Hypoxic-ischaemic encephalopathy

It is clear that the diagnosis of birth asphyxia can only be made retrospectively as it is the

sequence of symptoms and signs and how the brain and other organs react over time that

indicates the diagnosis (Nelson & Emery, 1993).

A classification system for fetal asphyxia has been developed to aid in the prediction of long

term outcome for the child (Low, 1998). This classification, summarised in Table 1.1,

includes metabolic acidosis, neonatal encephalopathy and multiple organ system dysfunction.

Table 1.1 Classification of intrapartum fetal asphyxia

Asphyxia* Metabolic Acidosis at Delivery

Encephalopathy Cardiovascular,

Respiratory and Renal Complications

Mild Moderate Severe Minor Moderate/Severe

Mild + +/- +/-

Moderate + + +/-

Severe + + + *Umbilical artery base deficit ≥12mmol/l, +/- present in some but not all fetuses (Adapted from Low, 1998) Although our current understanding of the physiologic mechanisms of intrapartum fetal

asphyxial brain injury has suggested a strong association with multiple organ system injury,

severe fetal brain injury can occur in response to severe acute insults in the absence of

multiple organ system dysfunction (Phelan et al., 1998). This finding was illustrated by Phelan

et al. (1998) who reported a series of cases of hypoxic-ischaemic encephalopathy in the

6

neonatal period, all due to acute catastrophic events, in infants who did not have multiple

organ system dysfunction but went on to have permanent neurologic injury. The acute

insults responsible for the encephalopathy included uterine rupture (43%), prolonged FHR

deceleration (36%), fetal exsanguination (7%), cord prolapse (7%), and maternal cardio-

pulmonary arrest (7%). These acute injuries, associated with a prolonged FHR deceleration,

may be linked to a severe decrease in cardiac output and hypotension that cause vulnerable

portions of the fetal brain to be injured before other organs.

1.2.1.3 Neonatal encephalopathy

Neonatal encephalopathy (NE) is a clinically defined syndrome of disturbed neurological

function in the infant at or near term during the first week after birth, manifest as difficulty

with initiating and maintaining respiration, depression of tone and reflexes, altered level of

consciousness, and often seizures (Leviton & Nelson, 1992).

In addition to asphyxia and hypoxic-ischaemic injury, a diverse group of disorders can result

in this syndrome, many of which begin antenatally. These disorders include maternal and

fetal infections; maternal thyroid and nutritional deficiencies; placental disorders; and genetic

diseases including Prader-Willi and Angelman’s syndromes, Rett’s syndrome, myotonic

dystrophy, non-ketotic hyperglycinaemia, thrombophilic disorders, brain malformations, and

a family history of epilepsy (Smith et al., 1996; Grether & Nelson, 1997; Badawi et al., 1998a;

Ellis et al., 2000; Impey et al., 2001; Richer et al., 2001; Geerdink et al., 2002; Johnston, 2003).

1.2.1.4 Hypoxic-ischaemic encephalopathy

Hypoxic-ischaemic encephalopathy (HIE) is a subtype of neonatal encephalopathy where the

origin of the condition is thought to be hypoxia and ischaemia.

The condition was first described in 1976 by Sarnat and Sarnat who described a syndrome of

abnormal neonatal behaviour that followed asphyxia. Hypoxic-ischaemic encephalopathy is

traditionally sub-classified into mild, moderate and severe (or Stage I, II or III) using a

classification system relying on the assessment of consciousness, neuromuscular control,

complex neonatal reflexes, autonomic function and seizures (Sarnat & Sarnat, 1976). Briefly,

mild encephalopathy is characterised by hyper-alertness, hyper-reflexia, dilated pupils,

tachycardia, and the absence of seizures. Moderate encephalopathy includes the presence of

lethargy, hyper-reflexia, miosis, bradycardia, seizures, hypotonia and weak suck and Moro

responses. Severe encephalopathy is characterised by stupor, flaccidity, small to mid-position

pupils which react poorly to light, decreased stretch reflexes, hypothermia and absent Moro

and suck responses (Sarnat & Sarnat, 1976; Finer et al., 1981). Various attempts have been

7

made to improve on these descriptions(Levene et al., 1985; Amiel-Tison & Ellison, 1986), but

the original Sarnat grading appears robust and clear in its categorisation.

There is debate in the literature regarding whether NE or HIE is the most appropriate term

to describe encephalopathy in the neonate. Multiple causes of brain injury have been

identified in both premature and term infants which include hypoxia-ischaemia,

haemorrhage, infection, and metabolic derangement. From this list, hypoxia-ischaemia is

recognised to predominate (Volpe, 2001b). The argument for using the term NE rather

than HIE is that hypoxia and ischaemia have often not been proven and have been assumed

from a variety of clinical markers that do not accurately reflect hypoxia or ischaemia of either

acute or chronic origin (Leviton & Nelson, 1992). This argument is supported by several

epidemiological studies which have described over 75% of cases of NE as having no clinical

signs of intrapartum hypoxia (Nelson & Leviton, 1991; Adamson et al., 1995; Badawi et al.,

1998b, a). In contrast a number of recent studies suggest that hypoxia-ischaemia, rather than

antenatal events, is the major cause of NE (Evans et al., 2001; Cowan et al., 2003). This

debate will be further discussed in section 1.2.3.

In this thesis, when referring to encephalopathy, I will use the term neonatal encephalopathy

unless citing a reference that specifically reports hypoxic-ischaemic encephalopathy.

1.2.1.5 Cerebral palsy

Cerebral palsy is an umbrella term covering a group of non-progressive, but often changing

motor impairment syndromes secondary to lesions or abnormalities of the brain arising in the

early stages of its development (Mutch et al., 1992; Stanley et al., 2000). There are a number

of different types of cerebral palsy. Spastic quadriplegia and, less commonly, dyskinetic

cerebral palsy are the only two subtypes of cerebral palsy associated with acute hypoxic

intrapartum events (Stanley et al., 1993; Rosenbloom, 1994; MacLennan, 1999). Spastic

quadriplegia is not specific to intrapartum hypoxia (Stanley et al., 2000). Only 24% of a

population base series of children with moderate or severe spastic quadriplegia were thought

to have been affected by intrapartum events (Stanley et al., 1993). Hemiplegic cerebral palsy,

spastic diplegia and ataxia have not been associated with acute intrapartum hypoxia

(MacLennan, 1999; Stanley et al., 2000).

8

1.2.2 Historical perspective of birth asphyxia

Fetal asphyxia was first described by the ancient Greeks who described severe depression at

birth as ‘asphyxia neonatorum’ meaning without pulse. This description was subdivided into

‘asphyxia livida’ meaning asphyxia with cyanosis, and ‘asphyxia pallida’ which was asphyxia

with shock and vasoconstriction of the periphery (Claye, 1963).

The first description of an association between abnormal neurologic outcome in children and

complications during birth was made by the English surgeon William John Little during the

19th century (Little, 1862). In a series of nine lectures beginning in 1843 entitled a "Course of

Lectures on Deformity of the Human Frame," Little presented the first medical descriptions

of cerebral palsy. He described cerebral palsy as "a peculiar distortion, never previously

described, that affects newborn children with spasmodic tetanus-like rigidity and distortion of

the limbs" (Accardo, 1989; Schifrin & Longo, 2000). By the end of the 19th century, the

disorder was commonly known as "Little's Disease". William Little began correlating the

relationship of Little's Disease to both premature and complicated births. By 1861, after

studying the disorder for 20 years, Little had tabulated over 200 cases of Little's Disease. In a

thesis before the Obstetrical Society of London, Little proposed that the causes of this

previously overlooked disorder included asphyxia at birth, premature birth, and direct

mechanical injury (Little, 1862; Schifrin & Longo, 2000). Little further noted that the

susceptibility of the fetal nervous system to damage depended on its stage of development.

Consistent pathological features associated with birth asphyxia have been recognised since

the 19th century. Anton described status marmoratus or état marbre involving the basal ganglia in

1893 (Levene, 1995) and Bressler first described ulegyria or gyral scarring in 1899.

Sir William Osler (1849-1919) described by many as the most influential physician in history,

was the first to use the name cerebral palsy, in a series of five lectures entitled, "The cerebral

palsies of children". Osler also described associations between cerebral palsy and difficult

delivery, asphyxia requiring prolonged resuscitation, neonatal seizures and arachnoid or

subarachnoid bleeding.

Sigmund Freud (1856-1939) was the next physician to write on cerebral palsy and he devoted

several volumes to the study of this condition (Freud & Rie, 1891; Freud, 1893, 1897). He

agreed with William Little that difficult birth was a cause of cerebral palsy, but went on to

suggest that difficult birth might more often be a symptom rather than a cause of the

condition. Freud, who had originally trained as a neurologist with a strong background in

9

neuroanatomy, stressed the lack of correlation of clinical findings with either purported

aetiology or neuropathological findings. It was Freud who first proposed that antenatal fetal

exposure to asphyxia could also result in cerebral palsy (Freud, 1897). He also described the

relationship between lesions in the brain and the degree of spasticity and paralysis in the

body, noting that more superficial brain lesions were more likely to cause problems in the

lower extremities.

Even though there were no reliable clinical data to support a diagnosis of “fetal distress”,

from 1862 until the late 1980’s it was commonly believed that the major cause of cerebral

palsy was asphyxia occurring during the intrapartum period (Little, 1862; Lilienfeld &

Pasamanick, 1955; Eastman et al., 1962; Steer & Bonney, 1962). Even in the mid-1970’s it

was believed that more than half of the cases of mental retardation were due to intrapartum

asphyxia and that electronic FHR monitoring could potentially prevent it (Quilligan & Paul,

1975).

1.2.3 Current view of fetal brain injury during labour

The theories of the last 150 years were challenged during the last two decades by

epidemiological studies suggesting that intrapartum hypoxia was not the cause of the majority

of cases of cerebral palsy. In 1988 both a cohort study from North America (Nelson, 1988)

and case-control study from Western Australia (Blair & Stanley, 1988) reported that in 90%

of cases of cerebral palsy, intrapartum hypoxia could not be the cause of cerebral palsy and in

the remaining 10% intrapartum signs compatible with damaging hypoxia may have had

antenatal or intrapartum origins. Similar results were also reported in another

geographically-based cohort study from Oxford, in the United Kingdom (Yudkin et al., 1995).

These three studies and a number of other epidemiological studies suggested that a large

proportion of cases of cerebral palsy were associated with maternal and antenatal factors

such as prematurity, intrauterine growth restriction, intrauterine infection, fetal coagulation

disorders, multiple pregnancy, antepartum haemorrhage, breech presentation, and

chromosomal or congenital anomalies (Nelson & Ellenberg, 1984; Stanley & Alberman,

1984; Nelson & Ellenberg, 1986; Blair & Stanley, 1988; Naeye et al., 1989; Nelson, 1989; Blair

& Stanley, 1990, 1993; PSANZ, 1995; Volpe, 1995b; Evrard et al., 1996; Grether & Nelson,

1997; Nelson et al., 1998; Nelson & Grether, 1998).

This body of epidemiological evidence led the American College of Obstetricians and

Gynecologists (ACOG), the Society of Obstetricians and Gynaecologists of Canada (SCOG)

and the Perinatal Society of Australia and New Zealand (PSANZ) to release policy

10

statements stating that a number of criteria were required to be met to link perinatal asphyxia

to neurological deficit. As early as 1992, ACOG issued a technical bulletin concluding that

for perinatal asphyxia to be linked to a neurological deficit in the child, all of the following

criteria must be present (ACOG, 1992):

1. Profound umbilical artery metabolic or mixed acidaemia (pH<7.00),

2. Persistence of an Apgar score of 0-3 for longer than 5 minutes,

3. Neonatal neurological sequelae including seizures, coma or hypotonia, and

4. Multiple organ system dysfunction (eg. cardiovascular, gastrointestinal, haematologic,

pulmonary or renal).

In 1995, the Task Force on Cerebral Palsy and Neonatal Asphyxia of the Society of

Obstetricians and Gynaecologists of Canada issued a policy statement (SOGC, 1996a) which

stated that the same criteria as those described by the ACOG should be used for defining a

potential link between intrapartum events and neurologic outcome. In addition to the four

ACOG criteria, they added a fifth criterion that the umbilical artery base deficit had to be

greater than 16mmol/L. The Canadian task force stated that if all these criteria were not

present, one could not conclude that hypoxic acidaemia existed nor that it had the potential

to cause neurological deficits during the intrapartum period.

In 1999, the Consensus Statement of the Perinatal Society of Australia and New Zealand

regarding the intrapartum origin of cerebral palsy (MacLennan, 1999) was published and has

now been endorsed by 21 obstetric and paediatric societies from around the world. This

statement defined criteria required to suggest the occurrence of damaging intrapartum

hypoxia sufficient to cause permanent neurological impairment. These criteria are presented

in Table 1.2 and included three essential criteria and five suggestive criteria which could be

used to help time the hypoxic event to the intrapartum period.

Cerebral palsy is not the only form of brain injury which may result from intrapartum

hypoxia. Early onset neonatal encephalopathy can also result from intrapartum hypoxia and

ischaemia and has a prevalence that varies widely in the literature according to the definition

used and the population to which that definition is applied. As a result the prevalence quoted

in the literature ranges from 1.8 to 7.7 per 1000 term live births (Finer et al., 1981; Ergander et

al., 1983; Levene et al., 1985; Nelson & Leviton, 1991; Adamson et al., 1995; Thornberg et al.,

1995). The United States Collaborative Perinatal Project studied 39,000 infants born at birth

weights greater than 2500g and found that 70% of those who showed early onset neonatal

encephalopathy died or were disabled (Ellenberg & Nelson, 1988). Mortality from neonatal

11

encephalopathy varies in different populations from 7% to 30.5% (Brown et al., 1974; Ellis et

al., 1999) and the spectra of impairment which can result from neonatal encephalopathy

include cerebral palsy, cerebral blindness, cognitive impairment, developmental delay,

convulsive disorders and severe hearing loss (Robertson et al., 1989; Dixon et al., 2002).

Table 1.2 Criteria to define an acute intrapartum hypoxic event

Essential Criteria

1

Metabolic acidaemia in intrapartum fetal scalp, umbilical cord arterial or very early neonatal blood samples (pH<7.00 and base deficit >12mmol/l)

2 Early onset moderate or severe neonatal encephalopathy in neonates ≥ 34 weeks

3 Cerebral palsy of the spastic quadriplegic or dyskinetic type

Non-specific criteria to time the hypoxic event to the intrapartum period

4 A sentinel hypoxic event occurring immediately before or during labour

5

A sudden, rapid, and sustained deterioration of the fetal heart rate pattern usually after the sentinel hypoxic event where the pattern was previously normal

6 Apgar scores of 0-6 for longer than 5 minutes

7 Early evidence of multisystemic involvement

8 Early imaging evidence of acute cerebral abnormality (Adapted from MacLennan, 1999) Similar to cerebral palsy, for many years it was accepted that fetal asphyxia during labour was

the major cause of neonatal encephalopathy despite the fact that the evidence for this was

sparse (Edwards & Nelson, 1998). Again, during the last decade, epidemiological studies

suggested that the primary aetiology of neonatal encephalopathy was antenatal rather than

intrapartum events (Nelson & Leviton, 1991; Yudkin et al., 1994; Badawi et al., 1998a). The

independent risk factors for neonatal encephalopathy occurring before conception or during

the antepartum period included socioeconomic status, family history of seizure or other

neurologic disease, conception after infertility treatment, maternal thyroid disease, severe pre-

eclampsia, bleeding in pregnancy, viral illness, having an abnormal placenta, intrauterine

growth restriction, and postmaturity (Edwards & Nelson, 1998).

Some caution is required in the interpretation of the epidemiological studies of neonatal

encephalopathy. First, the definition of encephalopathy varies between publications. In the

two studies by Badawi et al. (1998a,b) the diagnosis of encephalopathy is wider and less

restrictive than other studies and the long term prognosis of this wider definition is uncertain.

12

For comparison, in a consecutive series of 56 infants at the Hammersmith Hospital in

London diagnosed as encephalopathic on criteria similar to those used by Badawi et al.,

almost half would have been classified as clinically normal at 18-24 months (Edwards &

Nelson, 1998). Second, Badawi et al.(1998a,b) included infants with obvious neuro-

developmental and chromosome abnormalities, and those who presented with the disorder

late in the first week of life. The inclusion of 37 infants in the cohort with birth defects (23%

of the cases) clearly has implications for the subsequent analysis of the likely time of the

insult (Ellis & de Costello, 1999). Furthermore, infants with problems such as low tone and

feeding difficulties, which are common to many disorders, were included as having

encephalopathy. Therefore, the population selected in Badawi’s studies would have reduced

the proportion of infants in whom difficulties during birth were the most important factor,

and increased the chance that earlier antenatal factors would be significant. It is worthy to

note that while Badawi et al. (1998b) attribute the cause of neonatal encephalopathy to

antenatal factors rather than any intrapartum abnormality, they reported significant

differences between cases and controls in the frequency of post maturity beyond 42 weeks,

acute intrapartum events, Apgar score, onset of respiration, need for resuscitation, and birth

trauma. Furthermore, they identified an inverse association between elective caesarean

section and neonatal encephalopathy (Badawi et al., 1998b), a relation also suggested by

Gaffney and colleagues (Gaffney et al., 1994a). Together, these observations suggest that

intrapartum events may have a greater contribution to neonatal encephalopathy than

reported by the authors.

Other authors have presented epidemiological evidence to suggest that events during labour

are associated with neonatal outcome (Gaffney et al., 1994a; Nelson & Grether, 1998; Heller

et al., 2000; Draper et al., 2002). Gaffney et al. reported a cohort of children with cerebral

palsy with and without neonatal encephalopathy at birth. The maternal and antenatal

characteristics were similar in the two groups of children; however, they did find that the

seriousness of intrapartum complications and the pattern and severity of motor and

intellectual deficit were significantly worse in the neonatal encephalopathy group, strongly

suggesting that events during birth do affect infant outcome (Gaffney et al., 1994a). Nelson et

al. (1998) have noted that a tight nuchal cord at delivery was seen significantly more often in

infants with neonatal encephalopathy than in controls, which lends further support to the

explanation that events late in labour are important in neonatal encephalopathy (Nelson &

Grether, 1998). There are also data which suggest that infants born at night are more than

twice as likely to die from asphyxia than those born during the day, a finding that might be

attributable to tiredness and inexperience of night time staff, or a delay in appropriate

13

treatment (Heller et al., 2000). Evidence from confidential inquiries into neonatal

encephalopathy and neonatal deaths in term infants in the United Kingdom documents a

high proportion of suboptimal care during labour (Draper et al., 2002).

Over the last three years a number of studies have been published which lend support to the

original hypothesis that hypoxic-ischaemic events in the perinatal period are the main cause

of early neonatal encephalopathy, seizures or both (Biagioni et al., 2001; Maalouf et al., 2001;

Cowan et al., 2003) and perhaps play a more important role in the origin of cerebral palsy

than postulated over the last two decades (Hagberg et al., 2001). The most recent publication

by Cowan et al. (2003) was a review of 351 neonates referred to two tertiary centres for the

assessment of neonatal encephalopathy, seizures or both. In this study magnetic resonance

imaging (MRI) found evidence of an acute insult without established injury or atrophy in

80% of infants with neonatal encephalopathy and 69% of infants with seizures. Similar

results have been reported in other studies using MRI to assess neonates with neonatal

encephalopathy (Biagioni et al., 2001; Maalouf et al., 2001). The series reported by Cowan et

al. (2003) was not population-based, however this disadvantage is somewhat offset by the

detailed investigations and specific diagnosis for all of the 351 infants, ensuring that there was

no ambiguity about their inclusion in a study of term infants with encephalopathy. The

authors commented that while it is accepted that a few normal term infants will have

established lesions on MRI at birth, the low rate of established lesions in their study suggests

that there was no relation between these lesions and presentation with encephalopathy. This

observation is in accordance with data from Hagberg and colleagues' who noted that term

infants later presenting with cerebral palsy of apparent antenatal origin rarely had symptoms

as a neonate (Hagberg et al., 2001).

There have been concerns expressed about the prognostic significance of abnormalities

identified on neonatal MRI, particularly in neonates with encephalopathy, where there is a

high incidence of abnormalities on MRI. In the study by Cowan et al. (2003) all of the babies

with hypoxic-ischaemic patterns of lesions on MRI have been followed up for a minimum of

18 months; 34% died and 43% developed cerebral palsy. The location of injury in those who

died was largely central grey matter and the brain stem. No infant with only cortical or white

matter injury on MRI scan died, and only 2 with focal infarction died. The severity of

cerebral palsy in the surviving infants was consistent with published data that relates to

pattern of injury with imaging findings (Rutherford et al., 1996; Mercuri et al., 1999; Cowan,

2000). One long-term follow-up study has been performed on children with neonatal

encephalopathy who underwent MRI within their first week of life (Barnett et al., 2002). In

14

that study 68 term infants with NE were assessed between 5 and 6 years of age. Fifteen of

the 68 infants (22%) died in the neonatal period. Of the 53 surviving infants, 19 (36%)

developed cerebral palsy. The remaining 34 infants were considered normal at 2 years of age

but when assessed at school age, 8 (15%) had minor neurological dysfunction and/or

perceptual-motor difficulties, 1 (2%) had only cognitive impairment and 25 (47%) were

normal. Again in this study, the outcome largely reflected the pattern of lesions on MRI

imaging; 83% of those with a normal outcome had normal scans or minimal white matter

lesions and 80% of those with minor neurological dysfunction and/or perceptual motor

difficulties had mild or moderate basal ganglia lesions or more marked white matter lesions.

Taken together, recent studies utilizing MRI imaging suggest that MRI findings assist in the

prediction of outcome better than reliance on clinical assessment and Apgar score, which are

less specific (Barnett et al., 2002; Mercuri et al., 2002).

There are two recent population-based cohort studies that add useful information to the

debate regarding the potential intrapartum origin of cerebral palsy. The first study is a

population-based cohort of children with cerebral palsy from Sweden (Hagberg et al., 2001)

which encompassed the western health care region of Sweden, a total population of 2.1

million. Unlike many of the previous population-based cohort studies of cerebral palsy,

postnatal computer tomography and/or MRI was available for 90% of the 241 cases of

cerebral palsy identified in the population over a 4 year period (1991-1994). Of the cases of

cerebral palsy, 34% were assessed to have an antenatal aetiology, 46% had a perinatal

aetiology (intrapartum + neonatal) and 19% were unclassifiable. In term children who

developed cerebral palsy, hypoxic-ischaemic encephalopathy (HIE) was the single most

common aetiology present in 79% of cases and the birth asphyxia responsible for the HIE

began in labour in 86% of these cases. The outcomes in this study were graded; normal with

no seizures, seizures and not disabled, seizures and disabled, four-limb cerebral palsy and

death. Across this range of outcomes a gradient of severity was apparent in Apgar scores,

acidosis, renal dysfunction and the need for ventilation. Overall in this cohort, birth asphyxia

severe enough to cause cerebral palsy was recorded and documented in 28% of cases.

The second recent study suggesting an intrapartum origin of cerebral palsy is a population

cohort of children with neonatal encephalopathy from the South West Thames Region in the

United Kingdom (Evans et al., 2001). In their cohort of 143 cases of neonatal

encephalopathy (NE), which deliberately used an over-inclusive case definition similar to that

of Badawi et al. (1998), there were only seven cases (5%) with well-defined conditions that

were clearly unrelated to intrapartum asphyxia. Of the 15 NE deaths, 13 (87%) were

15

attributed to HIE; there was one dysmorphic baby and one with a probable metabolic

disorder. There were 16 cases of cerebral palsy in this cohort, giving a prevalence of cerebral

palsy of 12%, and antenatal complications or risk factors were identified in only 3 cases

(19%). Of the 16 cases of cerebral palsy, 13 had four-limb cerebral palsy of which 12 had

evidence of HIE based on their neonatal illness, the occurrence of acidosis and renal

dysfunction. The authors concluded that in term infants, neonatal encephalopathy with

seizures is commonly the result of an acute insult, rather than a direct reflection of

antepartum events, and it carries a high risk of death or cerebral palsy (Evans et al., 2001).

1.2.4 Conclusion

In summary, there is a vigorous debate about how important birth asphyxia really is as a

cause of either neonatal encephalopathy or cerebral palsy. The question is not whether birth

asphyxia happens or whether it can cause brain injury; both of these occurrences are real.

The issues are; first, what proportion of neurological impairment in children is due to

perinatal hypoxia-ischaemia and how much is due to entirely different causes? Second, how

many infants with asphyxia also have significant predisposing factors? Third, what are the

non-asphyxial factors which contribute to or determine unfavourable outcome? Considering

all of the published literature over the last three decades, between 5% and 95% of neonatal

encephalopathy is solely due to perinatal hypoxia-ischaemia and between 5% and 28% of

cerebral palsy has an intrapartum origin.

Although as an obstetrician I would like to believe that the great majority of adverse events

occur antenatally and few occur during the intrapartum period, there is growing evidence that

intrapartum events may be more important with regard to brain injury. The answers to the

questions posed above are yet to be established but they probably lie somewhere between

two divergent opinions. It is only with further large prospective population-based studies

utilising both advanced neonatal imaging with MRI and detailed follow-up with blinded

assessors, that the truth will be established.

16

1.3 Mechanisms of brain injury

During the 1950’s, it was recognised that animal models were required to permit systematic

examination of the effects of asphyxia and ischaemia on the fetus. At a symposium marking

the beginning of studies of the fetus in the monkey colony in Puerto Rico, Dr Baillie, the

Director of the National Institute of Neurological Disease, stated “Medical research in regard

to cerebral palsy and other neurological disorders has been relatively neglected. The focus

should be on those adverse biological factors, which operate in the perinatal period. The

proper point of departure is through controlled animal observations” (Windle, 1956).

Since 1956, there have been a number of research models utilised to examine the effect of

asphyxial exposure in the fetal monkey and fetal lamb. More recently, other animals such as

the rat, guinea pig and gerbil have also been utilised. Asphyxia has been induced by maternal

hypoxaemia, reduced uteroplacental blood flow, and umbilical cord occlusion whilst cerebral

ischaemia has been induced by carotid artery occlusion.

Animal models of asphyxia have been used to address two issues; first, to determine the fetal

response to asphyxial exposure, and second, to investigate the pattern and frequency of brain

injury after an asphyxial exposure.

1.3.1 The fetal response to asphyxia

Fetal survival after perinatal hypoxia-ischaemia depends in part on maintaining a balance

between oxygen supply and demand. It is well established that at extreme levels of

hypoxaemia the fall in oxygen delivery inhibits aerobic metabolism and eventually leads to a

fall in the concentration of high-energy phosphates in all tissues. This however represents

the final stage. Before this, during more moderate reductions in oxygen delivery, there is

evidence that the fetus is capable of mounting a complex response involving compensatory

vascular and metabolic mechanisms, which, at least in the short term, might preserve

oxygenation and function in essential organs such as the heart and brain (Jensen & Lang,

1992). The fetal cardiovascular responses to asphyxia have been thoroughly investigated in

the fetal lamb using a variety of different models of asphyxia including maternal hypoxaemia

(Cohn et al., 1974; Peeters et al., 1979; Richardson et al., 1989; Richardson et al., 1993; Asano et

al., 1994), reduced utero-placental blood flow (Field et al., 1990; Jensen et al., 1991; Reid et al.,

1991; Ball et al., 1994), and cord occlusion (Itskovitz et al., 1987; Ball et al., 1994; Kaneko et al.,

17

2003). There has been a remarkable consistency in the fetal cardiovascular response to

asphyxia seen in these different models as demonstrated in Table 1.3.

Table 1.3 Cardiovascular responses to an asphyxial exposure

Blood Flow

Reference:

Mod

el*

Dur

atio

n (m

in)

Mea

n A

rteria

l Pr

essu

re

Brai

n

Hea

rt

Adr

enal

Kid

ney

Car

cass

Cohn et al, 1974 1 4-44 ↑ ↑ ↑ ↑ ↓ ↓

Peeters et al, 1979 1 60 ↑ ↑ ↑ ↑ ↓ ↓

Richardson et al, 1989 1 420 ↑ ↑

Field et al, 1990 2 90 ↑ ↑ ↑ ↑ ↓ ↓

Reid et al, 1991 2 20 ↑ ↑ ↑ ↑ ↓ ↓

Jensen et al, 1991 2 35 ↑ ↑ ↑ ↑ φ ↓

Ball et al, 1994 2 90 φ φ ↑ ↑ ↓ ↓

Newman et al, 2000 2 60 φ φ ↓

Itskovitz et al,1987 3 5 ↑ ↑ ↑ ↑ φ φ

Ball et al, 1994 3 90 ↑ ↑ φ ↑ ↓ ↓ *Model 1: maternal hypoxaemia; Model 2: reduced utero-placental blood flow; Model 3: umbilical cord occlusion. The up arrow (↑) indicates a significant increase, the down arrow (↓) a significant decrease, and the φ represents a change that was not statistically significant. (Adapted from Low, 1998)

The late gestation fetal sheep mounts a rapid response to an episode of acute hypoxia with

bradycardia and redistribution of the combined ventricular output to the heart, brain and

adrenal glands at the expense of the fetal carcass (Campbell et al., 1967; Cohn et al., 1974;

Peeters et al., 1979; Giussani et al., 1994). The initial response is an increase in arterial blood

pressure due to increased systemic vascular resistance mediated by the autonomic nervous

system (Jensen & Lang, 1992). This response is initiated by arterial chemoreceptors (Bartelds

et al., 1993) and then augmented by an increase in plasma catecholamines, ACTH, cortisol

and possibly by arginine vasopressin, angiotensin and circulating endogenous opiates (Rurak,

1978; LaGamma et al., 1982; Jones et al., 1988; Challis et al., 1989; Giussani et al., 1994). The

18

fetus compensates by increasing blood flow through the ductus venosus, thus reducing blood

flow through the liver. Therefore, a fall in arterial oxygen content results in a response that

reduces the delivery of oxygen to organs such as the gut and carcass, while protecting the

heart, adrenals and brain.

During hypoxic insults in late gestation fetal sheep there is an increase in cerebral blood flow

to all areas of the fetal brain. After both two and four minutes of asphyxia there is a 2.5 fold

increase in cortical blood flow. The increases in blood flow to the brainstem and subcortex

are relatively more pronounced (Kaneko et al., 2003). With more prolonged insults, the

increases in cerebral blood flow are of a smaller magnitude (Richardson et al., 1989). Of note,

the lowest increase in blood flow in the subcortical regions occurs in the corpus striatum

(Kaneko et al., 2003) and it is this structure that shows neuronal loss in the near-term ovine

fetus after repetitive cord occlusion of a severe degree (Mallard et al., 1995b). The increase in

cerebral blood flow is thought to result from both a significant increase in perfusion pressure

(Kaneko et al., 2003) and cerebral vasodilation from local, circulating and neurogenic factors

that affect cerebral vasculature (Licata et al., 1975; Kontos, 1981).

During the first 60 seconds of an acute hypoxic episode in the late gestation fetal sheep the

oxygen uptake in the fetal brain is well maintained (Richardson et al., 1996). After this time

there is a progressive reduction in oxygen delivery and fractional oxygen extraction despite

the marked increase in cerebral blood flow (Kaneko et al., 2003). It has been shown that the

cortex and cerebellum reduce their oxygen delivery by one third after two minutes of acute

hypoxia whereas the subcortical regions and the brainstem are able to maintain oxygen

delivery for longer periods of time (Kaneko et al., 2003).

Despite maintaining or partially maintaining oxygen delivery to the near-term fetal brain

during hypoxia, the fractional oxygen extraction has been shown to be reduced toward zero

after two minutes of an acute hypoxic episode whereas the glucose fractional extraction is

doubled over the same time period (Chao et al., 1989; Kaneko et al., 2003). This suggests that

oxygen availability is the rate limiting step for oxygen consumption and in the absence of

adequate oxygen there is a rapid switch to anaerobic metabolism in the fetal brain. Lactate is

the end product of anaerobic metabolism, and under normoxic conditions the fractional

extraction of lactate is close to zero. During hypoxia a small but significant efflux of lactate

has been demonstrated from the ovine fetal brain (Kaneko et al., 2003). The limited size of

the lactate efflux may relate to constraints of the blood brain barrier transport systems (Chao

et al., 1989).

19

Overall, during an acute hypoxic event it has been calculated that there is an 80% decrease in

total energy production in the fetal brain after 2 minutes of cord occlusion (Kaneko et al.,

2003). This is countered by a decrease in synaptic transmission in the fetal brain, reflected by

a flat isoelectric electroencephalogram which has been suggested to reduce the total energy

requirements of the fetal brain by 50% (Astrup, 1982; Jones & Traystman, 1984). The twin

strategies of reducing energy turnover and maximising the efficiency of ATP production

mean that supply and demand are balanced, and ATP levels remain stable (Connett et al.,

1990; Newman et al., 2000).

These cardiovascular and metabolic responses to asphyxia are important to the integrity of

the central nervous system during the compensatory phase of an asphyxial exposure. If the

asphyxial exposure persists, the fetus will cross a threshold after which decompensation

begins to occur. When fetal oxygen saturation falls below 30%, a progressive metabolic

acidosis will develop. The threshold of cerebral dysfunction, characterised by suppression of

electrocortical activity and seizures in the fetal lamb, occurs when metabolic acidosis is severe

(base deficit greater than 16mmol/L) (Low, 1998). Decompensation with severe metabolic

acidosis is associated with a fall of arterial blood pressure due either to a decrease in cardiac

output or a failure to maintain peripheral vascular resistance. The combination of asphyxia

and ischaemia due to hypotension results in a decrease of cerebral oxygen consumption and,

if sustained, brain damage (Maulik, 1998).

1.3.2 Pattern and frequency of brain injury after asphyxia

Since the first animal studies of asphyxia were performed in the 1950’s, it has been

recognised that specific patterns of perinatal brain damage are associated with different types

of asphyxial insult. This is illustrated by the work of Myers et al. who described four patterns

of perinatal brain damage in primates associated with different categories of insults (Myers,

1975). These are summarised in Table 1.4. Anoxia or total asphyxia, whether in the

newborn or the adult, resulted in injury in the brainstem. In those animals that survived,

cerebral structures outside the thalamus and brainstem were spared. In contrast, situations

leading to hypoxia associated with severe acidosis (usually of a mixed respiratory and

metabolic type), cause brain oedema. When this oedema is limited in distribution, the

damage is restricted to specific cortical loci. However, when oedema becomes more

generalised, more extensive regions of the hemispheres are damaged until the entire

cerebrum may become necrotic. When the experimental model led to hypoxia without

acidosis, damage was restricted to the white matter, characterised by periventricular white

20

matter haemorrhage and/or focal areas of periventricular leukomalacia. The fourth pattern

described by Myers was that which occurred with combined episodes of hypoxia plus anoxia

with acidosis; these circumstances favoured a predominance of lesions that affected the basal

ganglia.

Table 1.4 Four patterns of perinatal brain damage

Neuropathology

Pattern

Corte

x

Whi

te

Mat

ter

Basa

l G

angl

ia,

Thala

mus

Brain

stem

Anoxia or total asphyxia + ++

Hypoxia with severe acidosis ++

Hypoxia without acidosis ++

Combined episodes of hypoxia, anoxia and acidosis ++ (Adapted from Myers et al., 1975)

Many different animal models have been utilised to investigate the occurrence of cerebral

dysfunction and brain damage after an asphyxial exposure. These studies can be broadly

categorised as cerebral ischaemia models, global asphyxia models and partial asphyxia

models.

1.3.2.1 Cerebral ischaemia models

Cerebral ischaemia has been examined using a carotid artery occlusion model in chronically

instrumented fetal lambs and rats (Williams et al., 1992; Mallard et al., 1993; Uehara et al.,

1999). The results of these studies are summarised in Table 1.5. Ten to twenty minutes of

ischaemia resulted in brief depression of the electroencephalogram (EEG) which then

progressively recovered. Mild selective neuronal necrosis was seen on post mortem

examination. After 30-40 minutes of ischaemia, EEG activity remained depressed for 8

hours, followed by a transition to low-frequency epileptiform activity that reached maximum

intensity at 10 hours. This sequence was associated with laminar necrosis of the parasagittal

cortex, and with necrosis in the hippocampus, basal ganglia and thalamus (Williams et al.,

1992).

The electrophysiologic and histologic consequence of isolated brief episodes of reversible

cerebral ischaemia of 10 minutes at 1- and 5-hour intervals has also been examined. The

21

effect of frequent episodes of brief cerebral ischaemia alters the pattern of brain injury with

basal ganglia damage a major feature (Mallard et al., 1993).

Table 1.5 Brain damage after cerebral ischaemia

Neuropathology

Reference Model* Duration Cerebrum Basal Ganglia Thalamus

Williams et al, 1992 1 10 minutes +

Williams et al, 1992 1 40 minutes +++ +

Mallard et al , 1993 1 10 minutes x 3 ++ +++

Uehara et al, 1999 2 2 days ++ * Model 1: carotid artery occlusion (lamb); Model 2: carotid artery occlusion (rat). (Adapted from Low, 1998)

Prolonged carotid artery occlusion has also been examined in a rat model where both carotid

arteries were ligated for 2 days. When the brains were assessed, 91% had evidence of white

matter injury including coagulation necrosis and cystic lesions in and around the internal

capsule while only 9% had evidence of cerebral infarction and 23% had evidence of

ischaemic neurons in the cortex (Uehara et al., 1999).

Therefore, considering these studies it appears that both the duration and pattern of

ischaemia are important factors regarding the resultant neuropathology. Repeated brief

episodes of ischaemia appear to sensitise the fetal brain to neuronal injury, particularly if the

episodes were frequent.

1.3.2.2 Global asphyxia models

A number of series of experiments in the monkey and lamb have been performed, using a

variety of models, to examine the effect of global asphyxia on the fetal brain (Ranck &

Windle, 1959; Mallard et al., 1992; Mallard et al., 1995a; de Haan et al., 1996; de Haan et al.,

1997b; Ohyu et al., 1999; Loeliger et al., 2003). These studies are summarised in Table 1.6.

Ranck and Windle performed the initial studies with the Mucacca mulatta monkey where they

induced global asphyxia by placental separation. They found that global asphyxia for more

than 10 minutes resulted in profound hypoxaemia with severe metabolic acidosis and

hypotension. The resultant neuropathology included neuronal necrosis principally in the

brain stem, basal ganglia, and thalamus (Ranck & Windle, 1959). On the basis of their

22

Table 1.6 Brain damage after global asphyxia

Fetal Response Neuropathology

Reference Model* Duration (minutes) Acidosis Hypotension Cerebrum Basal Ganglia Thalamus Brain Stem

Ranck & Windle, 1959 1 8-30 + + + ++ +++

Mallard et al., 1992 2 10 + + ++ +

Mallard et al. 1995a 2 Recurrent 5 min every 30 min + ++

de Haan et al., 1996 2 Recurrent 5 min every 30 min + + + ++

de Haan et al., 1997b 2 Recurrent 1 min out of 2.5 min or 2 min out of 5 min

+ + + +

Ohyu et al., 1999 2 Recurrent 3 min every 8 min + +

Marumo et al., 2001 2 Recurrent 3 min every 8 min + +/- + +

Loeliger et al., 2003 2 Recurrent 2 min every 30 min day 1, 4 min every 30 min day 2

+ - ++ +

* Model 1: placental separation (monkey); Model 2:– cord occlusion (lamb); +/- present in some but not all fetuses (Adapted from Low, 1998)

23

studies, it was concluded that global asphyxia for less than 8 minutes may not cause brain

damage, while global asphyxia for more than 10 minutes invariably resulted in

neuropathology. This pattern of brain damage after 10 minutes of global asphyxia was

confirmed by 10 minutes of cord occlusion in chronically instrumented fetal sheep (Mallard et

al., 1992). Post mortem examination of these fetuses demonstrated neuronal necrosis in the

hippocampus and to a lesser degree, in the basal ganglia and thalamus. There was a

correlation between hypotension and the magnitude of neuronal necrosis.

Repeated short episodes of global asphyxia have been investigated by a number of groups,

and are associated with different patterns of neuronal loss when compared with prolonged

global asphyxia. Mallard et al., using five minute complete cord occlusion repeated four times

at 30-minute intervals, found the areas principally affected were the basal ganglia and

specifically the striatal projection neurons (Mallard et al., 1995b). De Haan et al. utilised the

same occlusion protocol and found selective neuronal loss throughout the brain, with the

most extensive damage in the putamen. The second most damaged region was the caudate

nucleus of the striatum, followed by the hippocampus, cortex, thalamus and cerebellum (de

Haan et al., 1997c). Fetal EEG activity decreased and cerebral impedance increased during

the occlusions and the maximum spike and seizure activity occurred 5 to 10 hours after

asphyxia (de Haan et al., 1997c).

In further studies, De Haan and Gunn studied 2 groups of lambs; group one experienced

cord occlusion for 1 minute out of every 2.5 minutes and group two underwent occlusion for

2 out of every 5 minutes over several hours (de Haan et al., 1996). Both protocols resulted in

severe acidosis and hypotension and intermittent cord occlusion was continued until mean

arterial pressure was less than 20mmHg or it failed to recover to baseline before the next

occlusion. Ongoing asphyxia was associated with progressive suppression of the EEG,

which occurred faster and with more epileptiform activity in group 2. The predominant

neuropathology was selective mild neuronal necrosis throughout the fetal brain (de Haan et

al., 1996); however, in six of the 14 surviving asphyxiated fetuses, there were focal infarcts in

the parasagittal cortex, thalamus, and cerebellum. Infarction was associated with a longer

period of arterial pressure below baseline levels, with more epileptiform activity, and a slower

normalisation of the EEG (de Haan et al., 1997b). The authors felt that the histologic

damage that was found mimicked that of birth asphyxia in humans, and that their data

suggested brief repeated insults interact, leading to cardiac compromise and cumulative cell

membrane damage in the fetal cerebrum.

24

Ohyu et al. (1999) utilised yet another intermittent umbilical cord occlusion protocol during

which the cord was occluded for 3 minutes every 8 minutes. Interestingly, unlike the

majority of other intermittent cord occlusion models which show injury in the vast majority

of cord occluded fetuses, brain injury was not found in all of the fetuses exposed to this

pattern of cord occlusion. Seven of the 11 (63%) occluded fetuses had white matter and

cortical injury, six (55%) had evidence of thalamic injury and 4 (36%) had no histologic

evidence of damage (Ohyu et al., 1999).

Similar results were reported by Marumo et al. (2000) who used the same cord occlusion

protocol as Ohyu’s group. They found that one third of the occluded fetuses had

periventricular white matter injury, one third had cortical and thalamic injury, and one third

had no or minimal brain lesions. The pattern of neuronal injury was associated with fetal

arterial pressure responses and lipid peroxide concentrations. The fetuses with white matter

injury had higher arterial pressures and lipid peroxide levels prior to the onset of cord

occlusion and the fetuses with thalamic injury had the most severe hypotension during

episodes of asphyxia. Of note, both Marumo et al. and Ohyu et al. delivered most of their

fetuses 1 day after the asphyxial insult rather than waiting three days, the time used by the

majority of other investigators. This may partially explain why they found no neuronal injury

in some of their occluded fetuses.

Loeliger et al. specifically designed their recurrent brief global asphyxia study to induce white

matter injury (Loeliger et al., 2003). Their model was successful in inducing injury to white

matter, layer six of the cerebral cortex, and the caudate-putamen complex. Brain damage was

observed in all fetuses with the most extensive damage occurring in fetuses with the greatest

increase in glutamate levels. No association was found with fetal hypotension.

The preterm fetus appears to have a different response to global asphyxia than near-term

fetuses. At mid-gestation, before myelination has begun, 10 minutes (Mallard et al., 1994) and

20 minutes (Keunen et al., 1997) of cord occlusions produced no brain damage in fetal lambs.

It appears that the mid-gestation fetus may be more resistant to global asphyxia. On the

basis of these studies, the effect of global asphyxia on the fetal brain appears to be dependent

on the nature (prolonged or intermittent) and duration of the asphyxia, and the gestational

age of the fetus.

25

1.3.2.3 Partial asphyxia models

The effects of sustained partial asphyxia have been examined in fetal monkey and lamb

models and the results of these studies are summarised in Table 1.7 (Myers, 1972; Ting et al.,

1983; Clapp et al., 1988; Gunn et al., 1992; de Haan et al., 1993; Ikeda et al., 1998a).

Table 1.7 Brain damage after partial fetal asphyxia

Fetal Response NeuropathologyRe

fere

nce

Mod

el*

Dur

atio

n (m

inut

es)

Acid

osis

Hyp

oten

sion

Cere

brum

Basa

l Gan

glia

Thala

mus

Myers, 1972 1 60-180 + + +++ ++

Ting et al., 1983 2 120 + + ++ ++

Gunn et al., 1992 3 60 + + ++ ++

de Haan et al., 1993 3 60 + φ ++

Clapp et al., 1988 4 1 min out of 3 min for 120 min φ ++

Ikeda et al., 1998a 4 60 + + ++ ++

* Model 1: reduced uteroplacental blood flow (monkey); Model 2: maternal hypoxaemia-fetal hypovolemia (lamb); Model 3: reduced uteroplacental blood flow (lamb); Model 4: partial cord occlusion (lamb); φ, change not statistically significant. (Adapted from Low, 1998) Myers in 1972 exposed fetal monkeys to partial asphyxia by impairing utero-placental blood

flow for one to three hours. All of the monkeys experienced severe acidosis, but only some

had evidence of neuropathology. Those monkeys with neuropathology had experienced

severe acidosis with concurrent hypotension. The neuropathology ranged from generalised

neuronal necrosis to focal neuronal necrosis principally involving the parasagittal regions and

the junctional zones between the occipital and parietal lobes. Neuronal necrosis also

occurred in the basal ganglia and thalamus (Myers, 1972).

26

Ting et al. utilised a different model to investigate partial asphyxia in mid-gestational fetal

lambs; they exposed the fetuses to two hours of maternal hypoxaemia and fetal hypovolemia

(Ting et al., 1983). Subsequent post-mortem examination demonstrated that some lambs had

brain damage; the fetuses with brain damage had experienced moderate acidosis with

concurrent hypotension. The extent of neuropathology in the white matter, cerebral cortex

and basal ganglia correlated with the severity of hypotension.

Gunn et al. investigated partial asphyxia in the late gestation fetal lamb and found similar

results to both the monkey and mid-gestational fetal lamb models (Gunn et al., 1992). They

exposed 14 fetuses to partial asphyxia by reducing uterine blood flow such that the fetal

arterial oxygen content was <1.5mmol/L for 60 minutes (Gunn et al., 1992). Only eight

fetuses had evidence of brain damage, and these had severe acidosis (pH<7.0) with

concurrent hypotension. The predominant neuronal necrosis was in the parasagittal cortex,

and to a lesser extent in the basal ganglia; there was a strong correlation between the severity

of the hypotension and the extent of the neuropathology. They also found an association

between the reduction in EEG intensity during the insult and the severity of neuronal injury

(Gunn et al., 1992). Gunn et al. also compared neuropathology induced by 60 minutes of

reduced uterine artery blood flow with that induced by 30 minutes of cerebral ischaemia

from carotid artery occlusion (Williams et al., 1990; Gunn et al., 1992). They found greater,

but patchy damage to the striatum, and although the overall level of damage in the thalamus

was not significantly different, severe focal damage was seen in one fetus that was never

found in the ischaemic model. These differences may reflect greater susceptibility of the

basal ganglia to acidotic insults. Alternatively, asphyxia may be accompanied by differential

effects on oxygen delivery to cerebral regions. During transient cerebral ischaemia, there is

also likely to be residual anastomotic flow to the mid and hindbrains (Williams et al., 1990).

De Haan et al. performed an interesting project where severe acidosis was induced, without

hypotension, using a partial asphyxia model. Partial asphyxia was induced by reducing

uterine blood flow for 60 minutes to expose the fetuses to severe acidosis (pH<7.0), but in

their model, there was no change in arterial pressure compared with the control group (de

Haan et al., 1993). Without concurrent hypotension, only a few fetuses showed neuronal

necrosis that was principally in the cerebellum and to a lesser extent in the cerebral cortex.

Several different partial cord occlusion models have been employed to induce neuronal

damage in the late gestation fetal lamb. Clapp et al. repeated 1-minute partial cord occlusions

every three minutes for two hours. They found that this model did not induce acidosis, but it

did alter fetal heart rate patterns and fetal EEG. Eight of the nine fetuses subjected to

27

repeated cord occlusion were found to have histological damage in their cerebral white

matter (Clapp et al., 1988).

Ikeda et al. utilised a very different model in which partial umbilical cord occlusion lasted for

60 minutes. This induced severe acidosis (pH<6.9, base excess >20mmol/L) and there was a

wide variety of neuropathology ranging from almost total infarction of cortical and

subcortical structures to extremely subtle and patchy white matter alterations characterised by

slight vacuolisation of the white matter or slight to moderate increases in cellularity confined

to the junction of cerebral cortex and white matter. Even fetuses that showed full recovery

of all physiologic parameters, including EEG activity, demonstrated subtle but distinct white

matter lesions. The gray matter, including the hippocampal neurons, was generally spared in

these cases. EEG parameters, duration of hypotension during asphyxia, and delayed

recovery of blood lactate concentrations correlated well with the histologic grading of brain

damage (Ikeda et al., 1998a).

Most of the animal research with partial asphyxia has been performed on near-term fetuses.

Rees at al. studied the mid-gestation fetal sheep using a model which reduced uteroplacental

blood flow by 50% for 6 or 12 hours at 84 days of gestation (term 147 days; Rees et al, 1997).

In contrast to global asphyxia, which appeared to induce no damage in the preterm fetus

after 10 or 20 minutes of asphyxia, partial asphyxia resulted in white matter damage and

neuronal death in the hippocampus and to a lesser extent in the cerebral cortex and

cerebellum. Partial asphyxia also retarded neuronal migration and the growth of neural

processes in the hippocampus where development is well established at this age.

1.3.2.4 Summary of pattern and frequency of brain injury after asphyxia

Animal models of hypoxia and ischaemia confirm that cerebral ischaemia or global asphyxia

of ten minutes duration is consistently associated with brain injury. The studies using

repeated short episodes of ischaemia or global asphyxia have found that the pattern of brain

injury is dependent on the nature and duration of asphyxia, and the gestational age of the

fetus. Sustained partial asphyxia of 1 to 3 hours duration may or may not be associated with

brain damage, with the severity of the brain injury correlating with the severity of fetal

hypotension.

The sequence of events during global asphyxia or cerebrovascular occlusion provides the

fetus with limited compensation prior to the threshold of brain damage. However, although

global asphyxia or cerebrovascular occlusion may occur, they are an infrequent cause of brain

damage in the surviving human newborn (Low, 1998). With partial asphyxia, the fetal brain

28

is protected for variable periods (hours) before the threshold of metabolic acidosis and

hypotension leading to brain damage is reached. An explanation for the variability of

outcome seen after fetal asphyxial exposure is summarised in Figure 1.1. The fetus may be

exposed to asphyxia with a developing metabolic acidosis; however, if the fetal compensatory

cardiovascular response is maintained with increased blood flow to the brain, heart and

adrenals, then oxygen supply to these organ systems will be sustained and cerebral and other

organ system dysfunction or damage will not occur. If the asphyxia persists with progression

of the metabolic acidosis, then fetal decompensation with hypotension, decreased blood flow

and decreased oxygen supply to the brain will develop. The resulting tissue oxygen debt and

anaerobic metabolism will initiate a cascade of biochemical events leading to brain cell

dysfunction and death (Low, 1998). It is likely that this pathophysiological sequence of

events in response to asphyxia operates in the human fetus; however unlike in the animal

studies, in the human fetus the duration of the exposure or whether the exposure is

continuous or intermittent is rarely known.

1.3.3 Molecular mechanisms of perinatal brain injury

Perinatal brain injury has multiple causes including hypoxia-ischaemia, haemorrhage,

infection, and metabolic derangement. Of these, hypoxia-ischaemia predominates and is

generally accepted as the single most important cause of brain injury in the newborn with

consequences that are potentially devastating and lifelong (du Plessis & Volpe, 2002).

Our understanding of the fundamental mechanisms underlying hypoxic-ischaemic brain

injury has advanced rapidly in recent years. A number of the features of the pathophysiology

and clinical consequences of cerebral hypoxia-ischaemia are unique to the immature brain.

Particularly important in this regard is the critical interface between mechanisms of normal

development and injury to the immature brain, as well as the poorly understood but

important role of developmental plasticity in long-term outcome (du Plessis & Volpe, 2002).

Hypoxia-ischaemia leads to different neuropathological manifestations in term and preterm

infants. In term infants, neuronal injury is the primary manifestation of hypoxia-ischaemia

whereas in the preterm infant, oligodendroglial white matter injury predominates (Volpe,

2001b). The biochemical mechanisms leading from the hypoxic-ischaemic insult to neuronal

injury in the term infant are better understood than those resulting in oligodendroglial injury

in the premature infant.

FETAL

EXPOSUREASPHYXIA

Anoxia

+

Minutes

Hypoxia

+

Hours

HypercapniaRespiratory

Acidosis

HypoxemiaMetabolic Acidosis

FETAL

RESPONSECompensation Decompensation

Blood Pressure

Cerebral Blood Flow

Cerebral O2Consumption

OUTCOME

Cerebral Dysfunction None / MinorModerate /

Severe

Brain Damage Nil +

Mortality Nil +

Figure 1.1 Fetal pathophysiology after an asphyxial exposure. (Adapted from Low, 1998)

29

30

The most recent advances in the understanding of hypoxic-ischaemic brain injury involve the

delayed mechanisms of brain injury, the role of inflammatory processes in this injury and the

pathogenesis of injury to the immature oligodendrocyte.

1.3.3.1 Biphasic energy failure

It is now well established that the cerebral energy failure that results from hypoxic- ischaemic

insults occurs in two phases (Lorek et al., 1994; Roth et al., 1997; Inder & Volpe, 2000).

During the hypoxic-ischaemic insult there is a primary phase of energy failure, which is

followed after reperfusion by at least partial energy recovery. After a latent period, typically

between 8 and 48 hours, a secondary phase of energy failure may develop despite apparently

adequate cerebral perfusion and oxygenation. During the latent phase, processes of delayed

cellular injury become apparent and present attractive targets for clinical interventions to

interrupt this deleterious cascade of events even after termination of the insult.

During the primary energy failure phase, cell loss is related directly to hypoxia with

exhaustion of energy metabolism. Three mechanisms are thought to be involved in cell

death during this phase: (1) depolarization due to hypoxia, causing an influx of sodium,

chloride and water into the cell which if severe enough will result in lysis and cell death; (2)

intracellular calcium accumulation; (3) damage of cell membranes by oxygen free radicals in

the immediate reoxygenation-reperfusion phase (Saugstad, 2001).

Secondary or delayed neuronal death is triggered by events occurring in the primary phase

and is associated with hyperexcitability and cytotoxic oedema for between six and 100 hours

after the insult. The mechanisms involved in delayed neuronal injury are thought to include:

(1) excito-toxic amino acids; (2) apoptosis; and (3) cytotoxic activation of activated microglia

(Lorek et al., 1994; Inder & Volpe, 2000; Saugstad, 2001).

1.3.3.2 Pathogenesis of neuronal injury

Neurons have a high energy demand driven by their continuous activities essential for cellular

function and survival, but they possess inadequate reservoirs of oxygen and substrates for

energy production (Siesjo, 1992). Because neurons do not contain any glycogen or fat, they

depend exclusively on a continuous supply of glucose to fulfil their energy needs. By limiting

the availability of oxygen and glucose, the two obligatory substrates for brain energy

metabolism, hypoxia-ischaemia has a greater potential for neuronal injury than hypoxia alone.

There appears to be two perfusion thresholds for neuronal compromise, one for the

cessation of electrical activity and the other at a lower level for the loss of ionic homeostasis

31

which also serves as a threshold at which neuronal energy failure occurs (Branston et al.,

1974; Heiss et al., 1976; Siesjo, 1992).

Adenosine triphosphate (ATP) is the principal source of energy in a cell. The other high-

energy phosphate compound important as a source of energy in the mammalian brain is

phosphocreatinine. Oxidative phosphorylation is the primary ATP-generating mechanism of

the brain and is responsible for 95% of ATP production. This process occurs in the

neuronal cytosol where glycolysis of glucose generates pyruvate, which is further metabolised

in the Krebs cycle in the mitochondrial matrix to yield carbon dioxide and high-energy

electrons. These electrons are then transported by a complex system of electron carriers in

the respiratory chain in the mitochondrial inner membrane where they are eventually

combined with molecular oxygen. At the end of the respiratory chain an enormous amount

of energy is unleashed which is chemically stored by the formation of ATP. The Krebs cycle

and the mitochondrial respiratory chain require oxygen; from the combustion of one

molecule of glucose, oxidative phosphorylation generates 38 molecules of ATP whereas

anaerobic glycolysis, the alternative pathway utilised during hypoxia, yields only two

molecules of ATP. Thus oxidative phosphorylation is 16-17 times more efficient in

replenishing cellular energy supply than anaerobic metabolism.

Depletion of high-energy phosphates almost certainly initiates a cascade of events leading to

neuronal death after ischaemia and deprivation of oxygen and glucose (Volpe, 2001b). At a

cellular level, cerebral hypoxia-ischaemia sets in motion a cascade of biochemical events

commencing with a shift from oxidative to anaerobic metabolism, which leads to an

accumulation of nicotine adenine dinucleotide (NADH), flavin adenine dinucleotide

(FADH), lactic acid and hydrogen ions (Palmer et al., 1990). If the asphyxial insult persists,

the fetus is unable to maintain circulatory centralisation, and cardiac output and cerebral

perfusion fall. Owing to the acute reduction in oxygen supply, oxidative phosphorylation

and ATP production in the brain are diminished (Berger et al., 1992; Yager et al., 1992; Berger

et al., 1994). As a result, the sodium-potassium pumps in the cell membranes are deprived of

energy needed to maintain ionic gradients. With a reduced membrane potential, large

numbers of calcium ions flow through voltage-dependent ion channels into the cell down an

extreme extra- to intracellular concentration gradient. Intracellular accumulation of sodium

and chloride ions leads to swelling of the cells as water enters by osmosis leading to cytotoxic

cell oedema (Vannucci et al., 1993).

32

There are many consequences of the excessive increase in levels of intracellular calcium

which are summarised in Figures 1.2 and 1.3. (Rehncrona et al., 1982; de Courten-Myers et al.,

1989; Huang & Gibson, 1989; Mattson & Cheng, 1993; Berger & Garnier, 2000; Volpe,

2001b). One damaging effect is the activation of phospholipases A2 and C. These reactions

lead to membrane phospholipid hydrolysis, induction of free radicals, disruption of cell and

organelle membranes, increased membrane permeability and alterations in ionic distribution

(de Courten-Myers et al., 1989). Phospholipase C also catalyses reactions leading to the

production of inositol triphosphate, a second messenger that releases calcium from the

endoplasmic reticulum, and diacyl glycerol, which decreases calcium-sodium exchange

(Rehncrona et al., 1982; Huang & Gibson, 1989). Both these reactions further augment

calcium concentrations in the cell and amplify its deleterious effects, creating a vicious cycle

that eventually destroys the cell.

During ischaemia, besides an influx of calcium ions into the cells via voltage-dependent

calcium channels, additional calcium enters the cell through glutamate-regulated ion channels.

Glutamate, an excitatory neurotransmitter, is released from presynaptic vesicles during

ischaemia following anoxic cell depolarization. Extracellular glutamate accumulates during

hypoxia-ischaemia not only because of enhanced release but also impaired uptake both by

astrocytes and by presynaptic nerve endings (Rosenberg, 1997). The presence of increased

levels of glutamate in the synaptic cleft causes a further uptake of calcium in the postsynaptic

cells which leads to additional damage (Figure 1.2 A and Figure 1.3; Berger & Garnier, 2000).

Glutamate also leads to toxicity of immature oligodendroglia by both receptor and non-

receptor mediated mechanisms (Volpe, 2001b).

The acute lack of cellular energy during ischaemia induces almost complete inhibition of

cerebral protein biosynthesis. Once the ischaemic period is over, protein biosynthesis returns

to pre-ischaemic levels in non-vulnerable regions of the brain, while in more susceptible areas

it remains inhibited (Gitto et al., 2002). The inhibition of protein synthesis appears to be an

early indicator of subsequent neuronal cell death.

A second wave of neuronal cell damage occurs during the reperfusion phase after

termination of the hypoxic-ischaemic insult (Buonocore et al., 2001). It is during this phase

that the principal mechanisms leading to cell death operate. The cascade of events during the

reperfusion phase involves accumulation of cytosolic calcium and the production of oxygen

radicals, synthesis of nitric oxide, inflammatory reactions and an imbalance between the

excitatory and inhibitory neurotransmitter systems (Figure 1.3). Part of the secondary

neuronal cell damage may be caused by induction of apoptosis (Berger & Garnier, 2000).

33

GlutamateRelease

MembraneDepolarization

GlutamateUptake

ExtracellularGlutamate

ATP

IntracellularCa2+

A

ATP

Ca2+

Nitric OxideSynthase

Xanthineoxidase

Lipases Proteases MicrotubuleDisassembly

Nucleases UncoupleOxidative

Phosphorylation

Adenosine

Hypoxanthine

Xanthine

Oxygen FreeRadicals

Arachidonate

Eicosanoids

Degradation ofNeurofilaments

MembraneInjury

CytoskeletonDisruption

Cell Death

B

Figure 1.2 Hypoxia-ischaemia induced energy depletion and the cascade to neuronal death.

(A) Relationship between energy depletion, activation of glutamate receptors and

accumulation of intracellular calcium. (B) Deleterious effects of increased intracellular

calcium; note the critical role of oxygen free radicals. (Adapted from Volpe, 2001b)

33

Figure 1.3 Pathways of hypoxic mediated neuronal cell death.

During hypoxia ATP stores become depleted resulting in an increase in pre-synaptic release

of excitatory amino acids, primarily glutamate which acts on N-methyl-D-aspartate (NMDA),

α-amino-3-hydroxy-5-methyl-4-isoxazoleproprionic acid (AMPA) and kainite (KA) receptors

causing an influx of calcium ions (Ca2+). High intracellular Ca2+ levels trigger a cascade of

intracellular reactions resulting in excess activation of kinases, proteases, mitochondrial

dysfunction and consequently additional energy failure, ultimately resulting in cell death of a

necrotic nature. During this process lactic acid, oxygen free radicals (OFR) and nitric oxide

(NO) may be produced and accumulate within the cell. Factors such as these can contribute

to damage of not only the compromised neuron but also neighbouring cells and the blood-

brain barrier, and are thought to play an important role in reperfusion injury..

Hypoxic insults may also result in apoptotic cell death by either the death receptor or

mitochondrial pathways. Binding of ligands to membrane bound death receptors (Fas)

results in the activation of pro-caspases 8 and 3 to initiate cell death and may interact with the

mitochondrial pathway by the pro-apoptotic protein BID. The mitochondrial pathway

responds to stress during which changes in the expression of the pro-apoptotic protein BAX

and the anti-apoptotic protein Bcl-2 permits cytochrome c to leave the mitochondria,

interacting and activating Apaf-1 and pro-caspase 9. Caspase 9 then activates caspase 3

which initiates apoptosis. It is thought that the mitochondrial pathway plays the largest role

in hypoxic apoptotic cell death.

(Adapted from Mattson and Cheng, 1993; Herr and Debatin, 2001; and Duncan 2002)

Bloodbrain

barrier

Glutamate

AMPA

Ca2+

Ca2+

NMDA

Ca2+ Ca2+

KACa2+

Ca2+

Hypoxia+

↓Glucose

ATP

Ca2+

Energy

Necrotic Cell Death

SynapticCleft

Neuron

NeighbouringCell

CellDeath

OFR

OFRNO

NO

Ca2+

MitochondrialDysfunction

KinasesProteases

LactateAxon

Fas

Apoptotic Cell Death

Fas

Cytochrome c

Bax

Bcl-2

Caspase 8

Caspase 3

BID

Apaf-1c

Caspase 9

pro

pro

pro

pro

34

35

There is a large amount of data to suggest that oxidative stress plays a role in the initiation of

apoptosis (MacManus et al., 1993; Fernandez et al., 1995; Polyak et al., 1997). Studies have

shown also that free radicals are important in initiating delayed cell death via apoptosis after

cerebral ischaemia (Buttke & Sandstrom, 1994; Saikumar et al., 1998). Free radical mediated

cell death also appears to be the final common pathway for oligodendroglial cell death

(Volpe, 2001b).

The notion of early necrosis and more delayed apoptosis has recently been challenged by

evidence that suggests that the distinction and interaction between these two modes of cell

death is more complex than previously thought. Instead of the two temporally distinct

modes of cell death, an apoptotic-necrotic continuum has been described with an

intermediate group of hybrid cells having features of both necrosis and apoptosis (Nakajima

et al., 2000).

Recent studies have provided additional information regarding the timing of apoptosis in

brain injury. Caspase-3 activation, the pivotal apoptotic event, occurs between 24 hours and

7 days post-insult suggesting a prolonged role for apoptosis in neonatal brain injury. The rate

of apoptosis also differs across various brain regions being maximal in the thalamus and

hippocampus between 24 and 72 hours but delayed for up to seven days in the basal ganglia

and neocortex (Nakajima et al., 2000). Although it has been suggested that apoptosis is

delayed by up to 24 hours after an acute insult, critical apoptotic processes are in progress as

early as six hours post-insult. These include elevation of the Fas death receptor protein,

accumulation of cytosolic cytochrome c, caspase-8 activation and peri-nuclear clustering of

mitochondria (Figure 1.3; Northington et al, 2001).

1.3.3.3 The role of inflammatory mechanisms

The fetal inflammatory response appears to be important in the genesis of hypoxic-ischaemic

brain injury (Volpe, 2001b). Recent epidemiological and experimental studies have

demonstrated a strong association between inflammation and the development of

periventricular leukomalacia (PVL) and cerebral palsy, especially in the preterm infant

(Nelson et al., 1998). It is known that maternal, neonatal and fetal infections are associated

with increases in circulating cytokines, but this can also be triggered by ischaemia.

Pro-inflammatory cytokines may cause brain injury through several different mechanisms.

Inflammatory cytokines have direct cerebral cytotoxicity, inhibiting the differentiation of

oligodendrocyte precursors (Cammer & Zhang, 1999), stimulating oligodendrocyte apoptosis

(Selmaj et al., 1991; Louis et al., 1993; Selmaj et al., 1998), and causing vacuolar myelin

36

degeneration (Probert et al., 1995; Taupin et al., 1997). Cytokines may also have potent

vasomotor and vaso-occlusive effects (Brian et al., 1995; Brian & Faraci, 1998; Eklind et al.,

2001). Recent data suggest that cytokine toxicity may also be mediated through disturbances

in glutamate transport (Liao & Chen, 2001).

Studies suggest that the relationship between circulating inflammatory substances and

hypoxia-ischaemia in white matter injury may operate through a ‘two-hit’ mechanism. In

rats, the presence of an inflammatory stimulus prior to an hypoxic-ischaemic insult markedly

increased the vulnerability of the brain to hypoxic-ischaemic injury (Eklind et al., 2001).

Similarly, in an autopsy study of human neonates with PVL, infants with a combination of

infection and PVL had higher levels of cytokines than those with PVL without infection.

Together these data suggest that preceding infection with cytokine production amplifies the

effect of insults such as hypoxia-ischaemia (Kadhim et al., 2001).

Further evidence that inflammatory mechanisms are involved in the genesis of hypoxic-

ischaemic brain injury is provided by the growing body of evidence that manipulation of

inflammatory processes may offer protective effects in hypoxic-ischaemic brain injury. A

beneficial effect of induced neutropenia prior to hypoxia-ischaemia has been shown in an

animal model of perinatal hypoxic-ischaemic brain injury (Hudome et al., 1997). Platelet-

activating factor (PAF) levels increase with ischaemia reperfusion and PAF is important in

induction of leucocyte adhesion molecules and thereby subsequent events in the

inflammatory cascade. In the immature rat, administration of a PAF antagonist has been

shown to decrease infarct size both before the insult and during the reperfusion phase (Liu et

al., 1996). Studies using adult animal models of cerebral ischaemia suggest that the use of

specific antibodies or drugs directed at leukocytic adhesion molecules or various cytokines

may make an important contribution to neuro-protection (Zhang et al., 1994; Jiang et al.,

1995; Eun et al., 2000). A critical product of microglia/macrophages is interleukin-1β (IL-1β)

which in turn induces formation of other pro-inflammatory cytokines including tumour

necrosis factor-α (TNF-α). The use of an antagonist of the IL-1 receptor both pre-insult and

upon reperfusion has had protective effects in hypoxic-ischaemic brain injury in the neonatal

rat (Hagberg et al., 1996). Similarly, pentoxifylline, a phosphodiesterase inhibitor that inhibits

production of TNF-α, has been shown to ameliorate hypoxic-ischaemic brain injury when

administered immediately before hypoxic-ischaemic brain injury (Eun et al., 2000).

37

1.3.4 Summary of mechanisms of brain injury

Fifty years of research have consistently produced evidence that hypoxia and ischaemia can

produce prenatal brain injury. The pattern of damage is dependent on many factors

including the duration of the insult, the gestational age at exposure and the nature of the

insult (single or repeated exposure, partial or global asphyxia). Molecular mechanisms

involved in brain injury are complex and include cascades of events inducing the release of

glutamate, an increase in intracellular calcium, the production of oxygen free radicals, and

both necrotic and apoptotic cell death. Recent evidence also suggests that the fetal

inflammatory response is important in the genesis of hypoxic-ischaemic brain injury.

It is now well established that the cerebral energy failure that results from hypoxic-ischaemic

insults occurs in two phases in both animal and human newborns. The secondary phase can

result in neuronal injury and cell death that continues for up to seven days after the insult.

Both the delayed mechanisms of brain injury and the fetal inflammatory response are

attractive targets for clinical intervention to minimise the implications of hypoxic-ischaemic

brain injury.

38

1.4 Current techniques for intrapartum fetal surveillance

1.4.1 What are we aiming to prevent with intrapartum monitoring?

When reviewing our current techniques for intrapartum surveillance it is important to be

clear about what clinicians are attempting to prevent with intrapartum fetal monitoring.

1.4.1.1 Perinatal death, cerebral palsy and neurodevelopmental disability

The aim of intrapartum fetal monitoring is to reduce the incidence of three important

adverse clinical outcomes of fetal hypoxia; perinatal death, cerebral palsy and

neurodevelopmental disability. All of these outcomes are rare and both cerebral palsy and

neurodevelopmental disability only become apparent with the passage of time. Approximate

incidence figures for these conditions in the developed world are summarised in Table 1.8.

Table 1.8 Important adverse clinical outcomes

Condition Prevalence Source

Perinatal mortality 8 per 1000a MCHRC, 2000

Cerebral palsy 1.1 per 1000c Nelson et al., 1996

Perinatal mortality with intrapartum origin 0.8 per 1000a MCHRC, 2000

Cerebral palsy with potential intrapartum origin

0.1 per 1000c

0.2 per 1000a

Nelson et al., 1996 Watson et al., 1999

Neonatal encephalopathy 7 per 1000b Spencer et al., 1997

Neonatal seizures 11 per 1000a Thacker et al., 2001

Severe metabolic acidosis 20 per 1000d Low, 1998 a per 1000 live births b includes all grades of encephalopathy c per 1000 children who survived to three years of age (includes all birth weights) d pH<7.0, base deficit >12 mmol/L (Adapted from RCOG, 2001) It is important to note that only 10% of all perinatal mortality occurs during labour

(prevalence 0.8 per 1000 live births) and similarly it is thought only 10% of cases of cerebral

palsy have an intrapartum origin (prevalence 0.1-0.2 per 1000 children surviving to 3 years of

age) (Nelson et al., 1996; MacLennan, 1999; Watson et al., 1999; MCHRC, 2000; RCOG,

2001).

39

Other important neonatal outcomes which warrant prevention include neonatal

encephalopathy, neonatal convulsions and severe metabolic acidosis. The prevalence of

these outcomes is much more common than death and cerebral palsy, occurring in 2-3% of

all live births (Table 1.8)

1.4.1.2 Neonatal encephalopathy

In addition to asphyxia and hypoxic-ischaemic injury, a diverse group of disorders can result

in neonatal encephalopathy including maternal and fetal infections, maternal thyroid and

nutritional deficiencies, placental disorders and genetic diseases (Smith et al., 1996; Grether &

Nelson, 1997; Badawi et al., 1998b, a; Ellis et al., 2000; Impey et al., 2001; Richer et al., 2001;

Geerdink et al., 2002; Johnston, 2003). Although as discussed in section 1.1.4 there is debate

about the exact contribution of hypoxic-ischaemic events in labour to the incidence of

neonatal encephalopathy, there is general agreement that this condition is worthy of

prevention if possible.

Early onset neonatal encephalopathy is a good predictor of early developmental outcomes

(Dixon et al., 2002) and one of the better predictors of long-term outcome (Robertson et al.,

1989; Low, 1998). In a meta-analysis of five studies of term newborns with encephalopathy,

good outcomes were reported for newborns with minor encephalopathy, whereas the risk of

death or severe handicap in moderate encephalopathy was 5.6% and 20% and in severe

neonatal encephalopathy it was 60% and 72% respectively (Pelowski & Finer, 1992).

When neonates with neonatal encephalopathy were reviewed between one and two years of

age, they were found to have both statistically and clinically significant deficits in all areas of

development with the largest deficits seen in speech and hearing, which are crucial areas for

all aspects of development and learning (Dixon et al., 2002). These deficits persisted even

when children with cerebral palsy were excluded from the analysis.

The prevalence of cerebral palsy amongst children who experience neonatal encephalopathy

varies both between and within cohorts, based on the severity of the encephalopathy. The

overall prevalence of cerebral palsy in children with encephalopathy at birth ranges from 10%

in a Western Australian cohort to 22% in a cohort in Kathmandu (Ellis et al., 1999; Badawi et

al., 2001). Children with severe neonatal encephalopathy have the worst prognosis with

reported rates of cerebral palsy of 23% and abnormal development in 22% of those who

survive the neonatal period. This is to be compared with those who survive moderate

encephalopathy at birth, who have reported rates of cerebral palsy of 10%; abnormal

40

development occurs in 12% of these cases and 77% have normal developmental assessment

at two years of age (Dixon et al., 2002).

The deficits identified in short term follow-up of children with neonatal encephalopathy

continue to be evident when these children are reviewed later in childhood. Long term

follow-up studies at eight years of age suggest that intellectual, visual-motor integration,

receptive vocabulary scores, reading, spelling, and arithmetic grade levels for those with

moderate or severe encephalopathy, were significantly below (p<0.01) those in the mild

encephalopathy or peer comparison groups (Robertson et al., 1989).

1.4.1.3 Neonatal seizures

Intrapartum asphyxia is considered a major contributor to the occurrence of early neonatal

seizures (Derham et al., 1985; Low et al., 1985; Minchom et al., 1987; Patterson et al., 1989;

Lien et al., 1995) which occur in approximately 11 per 1000 live births (Thacker et al., 2001).

A 1985 consensus statement of the National Institutes of Child Health and Human

Development concluded that seizures were the best predictor of later neurologic damage in

the term infant (Freeman, 1985). More recent studies of children with neonatal

encephalopathy have shown that children with a history of encephalopathy and seizures were

more than three times more likely to develop cerebral palsy (16%) compared with those

without a history of seizures (OR 3.4; 95%CI 1.14-10.2; Dixon et al, 2002). Even seemingly

normal survivors of neonatal seizures may have a higher incidence of specific learning

disabilities and poor social adjustment when assessed during the late teenage years (Temple et

al., 1995).

1.4.1.4 Severe metabolic acidosis

Severe metabolic acidosis (pH<7.0, base excess<-12mmol/L) in umbilical arterial blood is a

sensitive reflection of birth asphyxia (MacLennan, 1999) and, ideally, intrapartum monitoring

should reduce the incidence of this condition. A number of studies have examined the

relationship between acidaemia at birth and both short- (Gilstrap et al., 1989; Low et al., 1995;

van den Berg et al., 1996) and long-term complications (Low et al., 1983, 1988; Ruth & Raivio,

1988; Low et al., 1994; Socol et al., 1994; Nagel et al., 1995; Ingemarsson et al., 1997). In short

term studies, babies with severe acidosis (pH<7.0) were significantly more likely to suffer

neonatal complications (Gilstrap et al., 1989; Low et al., 1995; van den Berg et al., 1996), and

in one study this relationship was found for those babies with metabolic acidosis rather than

severe respiratory acidosis (Low et al., 1994). The incidence of severe metabolic acidosis at

birth is consistently reported at approximately 2% of live births (Nagel et al., 1995; Low,

41

1998; MacLennan, 1999). In the series reported by Nagel et al. 77% of babies born with

severe metabolic acidosis were admitted to the neonatal intensive care unit, 27% required

ventilation and 7% died. Low et al. (1994) examined in detail the neonatal complications of

babies with severe metabolic acidosis and found that 78% of those with severe acidosis had

neonatal complications, compared with 27% in the babies with normal cord blood gas

analysis (p<0.001). More importantly they found that babies with severe metabolic acidosis

at birth were more likely to have complications in three or more organ systems (severe

acidosis 38% vs. normal blood gas analysis 6%, p<0.001). The specific neonatal

complications assessed in this study are summarised in Table 1.9. Of note, 61% of babies

with severe acidosis at birth had neonatal encephalopathy, with 42% having moderate or

severe encephalopathy.

Table 1.9 Short term neonatal complications of severe metabolic acidosis at birth

Umbilical Cord Blood Gas Analysis Neonatal Complication

Severe acidosisa Normalb p-value

Encephalopathy 61% 15% <0.001

Moderate or severe encephalopathy 42% 6% <0.001

Cardiovascular complications 59% 18% <0.001

Renal complications 24% 3% <0.010

Moderate or severe renal complications 16% 0% <0.010 aSevere acidosis, metabolic acidosis with base deficit >16mmol/L, mean pH 6.97, SD 0.13 bNormal, base deficit <16mmol/L, mean pH 7.25, SD 0.05 (Adapted from Low et al, 1994) In addition to an increased incidence of neonatal complications at birth, babies with severe

metabolic acidosis at birth have an increased incidence of neurodevelopmental deficits during

early infancy. When babies with severe metabolic acidosis at delivery were reviewed at one

year of age, the incidence of both major and minor motor and cognitive deficits were

significantly greater in the group with severe acidosis at birth than the control group, which

had normal acid-base status at birth (Low et al., 1988). The incidence of major deficits in the

group with severe acidosis at birth was 14% which was significantly greater than the 1%

incidence in the control group (p<0.01). Similarly the incidence of minor deficits in the

acidosis group was 27% which was significantly greater than the 6% in the control group

(p<0.01).

42

In long term follow-up studies, the association between acidaemia and neurodevelopmental

outcome is poor, with all but one study reporting no significant association between

acidaemia and neurodevelopmental disability (Low et al., 1983, 1988; Ruth & Raivio, 1988;

Dennis et al., 1989; Socol et al., 1994). The only study which did identify a significant

association between acidaemia at birth and neurodevelopmental disability found that babies

with umbilical arterial pH less than the first centile (umbilical arterial pH<7.05), when

compared to a control group with umbilical arterial pH>7.1, were more likely to have

abnormal developmental assessment at four years of age (24% vs. 13%, p=0.05; Ingemarsson

et al, 1997). Considering the acidotic babies, there was a significant increase in the number of

children with speech problems on developmental screening at 4 years (19% vs. 8%, p=0.03).

The risks for babies with severe acidosis at birth were worse in the subgroup that had a

pathological intrapartum FHR pattern which lasted longer than 60 minutes. In this

subgroup, both the incidence of speech problems and attention deficits were increased

compared to the acidotic babies with pathologic FHR patterns with durations less than 60

minutes (Attention deficit 18% vs. 6%, p=0.06; speech problems 32% vs. 14%, p=0.03;

Ingemarsson et al, 1997).

1.4.1.5 Summary

Intrapartum fetal surveillance is designed to reduce the incidence of perinatal death, cerebral

palsy, neurodevelopmental disability, neonatal encephalopathy, neonatal seizures and severe

metabolic acidosis at birth. Although it is universally accepted that perinatal death and

cerebral palsy are outcomes worthy of prevention, their combined prevalence relating to

intrapartum events is uncommon, occurring in approximately 0.1% of live births. Neonatal

encephalopathy and severe metabolic acidosis at birth are significantly more common than

cerebral palsy and perinatal death, occurring in 2-3% of live births, and have both short-,

medium- and long-term implications worthy of prevention.

43

1.4.2 Current monitoring techniques and their limitations

The three techniques which are currently widely used for the assessment of fetal wellbeing in

labour are FHR monitoring, fetal scalp blood gas analysis and umbilical cord blood gas

analysis.

The majority of women in labour in the developed world have FHR monitoring as the

principal method of intrapartum fetal surveillance. Intrapartum FHR monitoring is

performed as either intermittent auscultation or continuous electronic fetal heart rate

monitoring (EFM) with the selection of monitoring technique dependent on many issues. In

most obstetric centres, women with pregnancies deemed to be “high risk” have continuous

electronic FHR monitoring and those deemed to be “low risk” undergo intermittent FHR

auscultation. The specific criteria required for a pregnancy to be classified as “high risk” or

“low risk” varies between obstetric units, however there is remarkable consistency between

the recommendations of professional obstetric societies of developed countries (ACOG,

1995; RCOG, 2001; Liston et al., 2002a; Liston et al., 2002b; RANZCOG, 2002). The choice

of FHR monitoring technique is also influenced by issues such as hospital staffing, midwifery

experience, patient request, availability of equipment, and the unit’s philosophical opinion

regarding fetal heart rate monitoring.

When abnormalities are identified with intrapartum FHR monitoring a cascade of events is

initiated. The options for further fetal assessment vary significantly between obstetric

centres. In some centres delivery is expedited with either an instrumental vaginal delivery or

caesarean section. In other centres fetal scalp blood sampling is performed to assess the fetal

acid-base status. Once the baby is delivered, the fetal response to labour and birth is assessed

with clinical assessment, including Apgar scores. Umbilical cord acid-base assessment is

performed in some centres to determine if birth asphyxia has occurred.

In the next three subsections I will expand on FHR monitoring, fetal scalp blood sampling

for pH assessment, and umbilical cord acid-base assessment at birth. In particular I will

discuss their ability to predict or prevent fetal compromise with specific attention to the

prediction and prevention of fetal neurologic injury. Each subsection will conclude with a

discussion of the limitations of each assessment modality.

44

1.4.2.1 Fetal heart rate monitoring

Fetal heart rate monitoring was introduced into clinical practice and patient management

over three decades ago. It continues to be the predominant method of intrapartum fetal

surveillance despite questions about its efficacy and outcomes associated with its use

(Freeman, 2002).

1.4.2.1.1 Historical perspective

The era of modern obstetrics began in 1821 with the discovery of the ability to auscultate the

fetal heart by Jean Alexandre Lejumeau Vicomte de Kergaradec (Kergaradec, 1822). He was

genius enough not only to recognize the origin of the fetal heart tones, but also more

significantly, to anticipate the practical possibilities that could derive from its auscultation.

He stated, “From the changes occurring in strength and rate of fetal heart beats, wouldn’t it

be possible to know about the status of health or sickness of the fetus?” (Kergaradec, 1822)

Over the next thirty years, the fetal stethoscope allowed many discoveries including the

relationship between FHR accelerations and fetal movement, the association between FHR

and maternal temperature and the association between FHR and maternal blood pressure. It

also made possible the diagnosis of fetal life or death, twins, and the determination of fetal

presentation and position. It was during the 1840’s that the adverse fetal and neonatal

consequences of both late decelerations and progressive and long-lasting bradycardia were

described during labour (Sureau, 1996). In 1847, Depaul claimed that auscultation should be

used for every delivery and not be restricted to “high risk cases” (Depaul, 1847). In 1893,

Von Winckel described the criteria for fetal distress (Von Winckel, 1893). These criteria

included tachycardia (FHR more than 160bpm), bradycardia (FHR less than 100bpm),

irregular fetal heart rate, passage of meconium and gross alteration of fetal movement. By

the beginning of the 20th century, auscultation of the fetal heart was an established practice in

Europe (Sureau, 1996). The criteria set for the normal FHR in the latter part of the 19th

century remained virtually unchanged until the 1950’s.

The technology for electronic fetal heart rate monitoring was first developed in the 1950’s

and became commercially available in the 1960’s. During the 1960’s theories regarding the

cause of FHR patterns were formulated and a limited amount of experimental and empirical

evidence was produced to support these theories (Hon, 1968). While the original aim of

intrapartum EFM was to prevent harm, it was introduced into delivery suites in the 1950’s

and 60’s with the emphasis on improving fetal birth outcomes by detecting fetal hypoxia

before it led to death or disability. Like intermittent auscultation in the 19th century,

45

continuous EFM was introduced into clinical practice before its effectiveness had been fully

evaluated scientifically (RCOG, 2001).

The initial studies supporting the use of intrapartum fetal heart rate monitoring were

retrospective observational studies published between 1972 and 1976 which reported a

decrease in perinatal mortality in those women who had continuous EFM as opposed to

those who had selective EFM or no EFM at all (Koh et al., 1975; Shenker et al., 1975; Tipton

& Lewis, 1975; Hochuli, 1976; Lee & Baggish, 1976; Beard et al., 1977; Johnstone et al., 1978;

Weinraub et al., 1978). Similarly, studies comparing fetal outcomes at two serial time periods

covering the introduction of EFM into the United States reported a virtual disappearance of

intrapartum stillbirths once EFM was instituted (Parer, 1979; Yeh et al., 1982). Before the

introduction of FHR monitoring, approximately one third of all stillbirths, or three per 1000

births, occurred in the intrapartum period (Lilien, 1970). While these studies were

encouraging, the methodological biases of observational studies prompted the need for

randomised controlled trial evidence to more rigorously evaluate the use of intrapartum EFM

for preventing perinatal mortality and morbidity.

Thirteen randomized clinical trials have assessed the efficacy and safety of EFM, comparing

EFM to intermittent auscultation (Haverkamp et al., 1976; Renou et al., 1976; Kelso et al.,

1978; Haverkamp et al., 1979; Langendoerfer et al., 1980; Wood et al., 1981; MacDonald et al.,

1985; Neldam et al., 1986; Luthy et al., 1987; Shy et al., 1990; Vintzileos et al., 1993; Herbst &

Ingemarsson, 1994; Mahomed et al., 1994). Nine of these trials have been utilised in the three

systematic reviews comparing EFM to intermittent auscultation and include over 18,000

high- or low-risk pregnancies from seven clinical centres in the United States, Europe, and

Australia.

1.4.2.1.2 Fetal heart rate monitoring and adverse clinical outcomes

Three systematic reviews have examined the effect of EFM, in comparison with intermittent

auscultation, on perinatal death rates (Grant, 1989; Vintzileos et al., 1995b; Thacker et al.,

2001) and none found a significant reduction in the perinatal death rates with EFM. When

analysed separately, none of the trials included in these reviews demonstrated a reduction in

either intrapartum or neonatal deaths. The addition of fetal blood sampling to EFM did not

improve the ability of EFM to reduce perinatal deaths.

In the systematic review of Vintzileos et al. (1995b) a subgroup analysis for perinatal death

was undertaken. Deaths in each trial were classified as due to either hypoxia or other causes.

A significant reduction in the odds of perinatal death due to hypoxia with the use of EFM

46

was found; however, this was a post-hoc analysis and is therefore prone to subjective

selection bias and thus the subgroup analysis result should be treated with caution. It should

be noted that the prevalence of perinatal mortality of intrapartum origin has been reported as

0.8 per 1000 live births (MCHRC, 2000) and for a clinical trial to show a 50% reduction in

the prevalence of this end point with 80% power it would require 64,000 women in each

arm.

In addition to reduction in perinatal death, the three systematic reviews comparing EFM to

intermittent auscultation found no significant differences in 1-minute Apgar scores below

four or seven, rates of admission to neonatal intensive care unit, or cerebral palsy. The use of

EFM was associated with an increase in the rate of caesarean delivery (RR 1.41, 95% CI 1.23-

1.61) and operative vaginal delivery (RR 1.20, 95% CI 1.11-1.30).

There have been three studies investigating cerebral palsy rates in the cohorts of babies

included in three of the randomised controlled trials comparing electronic FHR monitoring

with intermittent auscultation (Langendoerfer et al., 1980; Grant et al., 1989; Shy et al., 1990).

One of these studies found an increase in rates of cerebral palsy in babies monitored by EFM

compared with intermittent auscultation (19.5% versus 7.7%; relative risk (RR) 2.54; 95%

confidence interval (CI) 1.10-5.86; Shy et al, 1990). These results should be interpreted with

caution as the cohort studied included only preterm babies who weighted less than 1750gm

at birth. Furthermore, in that study, management subsequent to the detection of FHR

abnormalities was not consistent in both arms of the study, with a significantly longer delay

between onset of FHR abnormality and delivery in the EFM group (105min vs. 45 minutes).

The remaining two cohort studies found no significant difference in cerebral palsy rates

between the two groups at the end of the followup period (Langendoerfer et al., 1980; Grant

et al., 1989).

In two of the larger case-control studies investigating the relationship between EFM and

cerebral palsy, a significant association has been identified between abnormal

cardiotocograph (CTG) findings and cerebral palsy (Gaffney et al., 1994b; Nelson et al., 1996).

In the study of Nelson et al. (1996) the two fetal heart rate patterns found to be associated

with an increased risk of cerebral palsy were multiple late decelerations (OR 3.9; 95%CI 1.7-

9.3) and decreased FHR variability (OR 2.7; 95%CI 1.1-5.8); however, only 0.19% of

singletons with these patterns develop cerebral palsy, giving a false positive rate of 99.8%.

No other FHR patterns were found to have significant associations with the diagnosis of

cerebral palsy.

47

Three case control studies have examined whether abnormal EFM traces predict subsequent

development of neonatal encephalopathy (Gaffney et al., 1994a; Socol et al., 1994; Adamson et

al., 1995). In all three studies there were significant increases in the odds of developing

neonatal encephalopathy in the presence of abnormal intrapartum FHR patterns (odds ratios

ranging from 1.98 to 10.2). Unfortunately all of the studies used different FHR pattern

classification systems and none investigated individual patterns of FHR abnormality. In

addition, in one of the studies, if a different FHR pattern scoring system was used, no

association with neonatal encephalopathy was seen (Spencer et al., 1997). As all these studies

were case-control studies, caution is needed in ascribing a causal relationship to the observed

effect.

The use of continuous EFM rather than intermittent auscultation has been found to be

associated with a significant reduction in neonatal seizures in two systematic reviews (0.24%

vs. 0.5%; RR 0.51; 95%CI 0.32-0.82; number needed to treat (NNT) 384; Grant, 1989;

Thacker et al, 2001). The protective effect for neonatal seizures was only evident in studies

with high-quality scores in the systematic review by Thacker et al. (2001) and it was only

present in the studies which combined EFM with fetal blood sampling in the systematic

review by Grant et al. (1989).

In only one of the thirteen randomised clinical trials of EFM versus intermittent auscultation

has the relationship between convulsions and subsequent neurodevelopmental disability been

examined (MacDonald et al., 1985). In this study, the reduced convulsion rate seen in the

EFM arm in the original trial did not translate into a significant reduction in the rate of

cerebral palsy in the group on follow-up. Of the six babies from this cohort who

subsequently developed cerebral palsy, five were thought to be attributable to antepartum

factors (Grant et al., 1989). These results are dissimilar to other studies investigating the

outcome of children with neonatal seizures, which suggest that neonatal seizures are a good

predictor of later neurologic damage, including higher rates of cerebral palsy, specific learning

disabilities and poor social adjustment (Freeman, 1985; Temple et al., 1995; Dixon et al.,

2002). It is important to note that in the original Dublin trial of EFM there were only nine

neonates in the EFM arm and 21 neonates in the intermittent auscultation arm of the study

who had neonatal seizures (MacDonald et al., 1985). Given that the highest rate of

subsequent cerebral palsy that has been reported in neonates with encephalopathy and

seizures is 16% (Dixon et al., 2002), the followup study by Grant et al. had only 14% power to

detect a difference in cerebral palsy rates in neonates with seizures.

48

Only one randomised controlled trial has found that EFM was significantly more sensitive

than intermittent auscultation for detecting both respiratory and metabolic acidosis

(Vintzileos et al., 1995a). Despite having better sensitivity for detecting acidosis at birth, the

specificity for EFM was poor (detection of all acidosis: EFM: sensitivity 97%, specificity

84%; intermittent auscultation: sensitivity 34%, specificity 94%).

A number of recent studies have revisited EFM and the prediction of outcome. Most of

these case-control studies have investigated the ability of EFM to predict intervention at

delivery, umbilical artery acidaemia or neonatal intensive care admission (Spong et al., 1998;

Dellinger et al., 2000; Hadar et al., 2001; Low et al., 2001; Sheiner et al., 2001; Low et al., 2002).

None has investigated the ability to directly predict fetal neurologic injury. All of these recent

studies suggest that abnormalities in EFM are associated with an increased risk of adverse

clinical outcomes at birth; however, as all these studies were case-control studies, caution is

needed in ascribing a causal relationship to the observed effect

The last two decades of research into FHR monitoring have been best summarised by Low

(1998) who commented that EFM is a useful screening test for the prediction of intrapartum

fetal asphyxia in preterm and term pregnancy. Although fetal heart rate patterns will not

discriminate all asphyxial exposures, continuous fetal heart rate monitoring supplemented by

fetal blood gas and acid-base assessment can be a useful fetal assessment paradigm for

intrapartum fetal asphyxia. Such an assessment paradigm will not prevent all cases of

moderate or severe fetal asphyxia. However, prediction and diagnosis with intervention and

delivery during the first or second stage of labour may prevent the progression of mild

asphyxia to moderate or severe asphyxia in some cases (Low et al., 2001; Low et al., 2002).

1.4.2.1.3 Limitations of fetal heart rate monitoring

There are a number of important limitations in FHR monitoring which may partially explain

why its widespread use has not been matched by improvements in neonatal outcomes. First,

there is a lack of standardisation of interpretation of FHR patterns and there is disagreement

regarding algorithms for intervention for specific FHR patterns (NICH, 1997; Parer & King,

2000). This is particularly important when reviewing the results of the systematic reviews of

EFM versus intermittent auscultation where the definitions for those FHR patterns requiring

obstetric intervention were not the same in each study (Thacker et al., 2001). Recent

publications by professional societies have addressed some of the issues of standardisation of

definitions (ACOG, 1995; NICH, 1997; RCOG, 2001; Liston et al., 2002a; Liston et al.,

2002b; RANZCOG, 2002).

49

Second, amplifying the problem of varying definitions is the fact that inter-observer and

intra-observer reliability for interpreting FHR patterns is relatively poor. All the studies

assessing inter-observer agreement in interpretation of FHR patterns have found marked

variability in interpretation of both normal and abnormal tracings by obstetricians,

experienced perinatologists, midwives and residents with inter-observer agreement

commonly reported to be approximately 30% (Peck, 1980; Beaulieu et al., 1982; Helfand et al.,

1985; Nielsen et al., 1987; Bernardes et al., 1997; Blix et al., 2003). The proportion of

agreement is generally better for FHR patterns with normal baseline FHR, variability and

accelerations (range 56% to 71%) whereas for abnormal FHR patterns, the proportion of

agreement has been reported to be as low as 14% (Bernardes et al., 1997). Intra-observer

agreement is equally poor. This is illustrated by three studies (Beaulieu et al., 1982; Nielsen et

al., 1987; Borgatta et al., 1988) which repeatedly showed the same FHR records to

experienced clinicians on a number of occasions. The changed interpretation rate in these

studies varied from 16% to 28%.

Third, intrapartum FHR monitoring performs poorly as a screening test. It was originally

anticipated that EFM would solve two problems; first, that it would be a screening test for

severe asphyxia, and second, it would allow recognition of early asphyxia so timely

intervention could avoid asphyxia-induced brain damage or death. Unfortunately, if all

women had continuous EFM during labour, between 30% and 79% of them would have

some abnormality in their FHR patterns which may in some clinicians’ minds be ominous or

at best suspicious and therefore trigger intervention (Nielsen et al., 1987; Parer & King, 2000).

A screening test for uncommon outcomes, which is considered abnormal in such a high

percentage of the tested population, is destined to have poor specificity and a very high false

positive rate.

The fourth limitation of intrapartum FHR monitoring relates to fetal physiology. It has been

thought for many years that abnormal FHR patterns would progressively evolve as the fetal

biochemical and cardiovascular systems decompensate with progressive asphyxia. Both

human (Low et al., 2001) and animal studies of asphyxia (de Haan et al., 1997a) have shown

that this does not always occur. This is illustrated in the study by de Haan et al. (1997a) which

investigated FHR changes during repetitive brief umbilical cord occlusions in the sheep fetus.

They found that the changes in FHR pattern with ongoing hypoxia were variable and, even

though fetal arterial pressure fell after approximately 15 minutes of repeated brief cord

occlusion, the pattern of the FHR decelerations remained relatively consistent throughout the

two hours of the experiment. Despite the fact that fetuses had a severe metabolic acidosis

50

and hypotension, fewer than 30% of them had a lengthening of the time needed for the fetal

heart rate to recover to baseline after the final occlusion (de Haan et al., 1997a). It would

therefore appear that there may also be physiologic limitations in the ability to identify all

cases of moderate or severe asphyxia from FHR patterns alone.

1.4.2.2 Fetal scalp blood gas analysis

Intrapartum fetal scalp blood sampling (FBS) was introduced into clinical practice by Saling

in 1962 as the first invasive fetal monitoring tool (Saling, 1962). It allowed greater

understanding of the effect of labour on blood gas homeostasis in the human fetus and it

aided in defining the importance of FHR patterns revealed with continuous EFM which was

introduced soon after (Kubli et al., 1969; Beard et al., 1971).

Fetal blood sampling is limited to the skin covering the fetal presenting part and it requires

ruptured membranes and the cervix to be at least 2 to 3-cm dilated. Initially only fetal pH

was measured but with improvements in micro sample analysis technology it became

possible to measure pH, and oxygen and carbon dioxide tensions from the fetal scalp in

labour. The addition of carbon dioxide measurement allowed the calculation of base deficit

to quantify the metabolic component to an acidosis.

It is widely accepted that a normal FHR pattern in labour is associated with a good outcome

and, conversely, a persistent bradycardia is associated with a poor outcome. Unfortunately,

between these extremes are various patterns that do not provide reassurance, and as

described previously, there is poor inter- and intra-observer variability in the description of

these patterns. Even once a particular pattern is identified, there is difficulty in obtaining

agreement about the most appropriate management algorithm for all but a few patterns

(NICH, 1997). Heart rate decelerations occur in 50% of babies monitored with continuous

EFM (Ingemarsson, 1981). As at least 50% of labours are monitored by EFM in the

developed world, the high incidence of FHR decelerations has the potential to initiate a

cascade of events frequently resulting in intervention in a large number of women.

Intrapartum fetal scalp blood sampling is traditionally used to clarify fetal acid-base status

when a non-reassuring FHR pattern is present in labour. Although there is an association

between low cord pH (Krebs et al., 1979) and increased base deficit (van den Berg et al., 1987)

with increasingly abnormal FHR patterns in labour and their duration (Fleischer et al., 1982),

an ominous FHR pattern such as one which is tachycardic with late decelerations and

51

decreased baseline variability has only a 50% incidence of low pH (Beard et al., 1971). The

percentage of fetuses with fetal scalp pH less than 7.25 for particular FHR patterns is

presented in Figure 1.4.

1.4.2.2.1 Physiologic basis for fetal scalp blood gas analysis

Prior to labour, fetal acid-base status is largely determined by the acid-base status of the

mother. During labour, fetal acid-base status is determined by both the mother’s acid-base

status and placental function. When intra-uterine pressure exceeds 30mmHg, the arteries

supplying the intervillous space of the placenta are constricted so that there is temporarily

little blood flow and little exchange of oxygen (O2) and carbon dioxide (CO2). This can be

accommodated by a healthy feto-placental unit; however, if the placenta is compromised or

contractions occur too frequently or last too long, there will be a progressive accumulation of

CO2 leading to hypercapnia and hypoxaemia. Elevated levels of CO2 in the fetus lead to a

respiratory acidosis as the CO2 combines with water to form carbonic acid. Prolonged

hypoxia can lead to a decrease in oxygen delivery to fetal tissues which prompts a switch to

anaerobic metabolism and produces a metabolic acidosis (Figure 1.5A). The fetus usually

develops a respiratory acidosis before a metabolic acidosis with the former being less

significant as it may not be associated with tissue oxygen lack. With prolonged impaired

placental blood flow the fetus develops a mixed acidosis.

Any factor that affects placental blood flow and gas exchange in labour affects fetal acid-base

status. These factors include cord compression, hypertonic uterine contractions, maternal

supine hypotension, cord prolapse or placental abruption. Any decrease in FHR results in a

decreased cardiac output, decrease CO2 transfer, and CO2 accumulation so that prolonged or

repetitive bradycardias result in a respiratory acidosis; this sequence commonly occurs in the

second stage of labour (Greene, 1999). The rate of development of metabolic acidosis

depends on the particular insult and the individual fetal reserve and ability to compensate.

The fetus normally derives its cellular energy primarily from glucose. Under normal

conditions with sufficient oxygen supply to the fetus, aerobic metabolism occurs when

glucose (or glycogen) is broken down via the glycolytic pathway and citric acid cycle to

produce water, carbon dioxide and 38 units of ATP for each glucose molecule (Figure 1.5B).

52

Normal

FHR patte

rn

Accele

ration

s

Early D

ecele

ration

s

Baseli

ne Tac

hyca

rdia

Baseli

ne B

radyc

ardia

Variab

le Dec

elerat

ions

Compli

cated

Vari

able

Decel

Decrea

sed F

HR varia

bility

Decrea

sed v

ariab

ility +

other

0

10

20

30

40

50

60Fe

tal s

calp

blo

odpH

<7.2

5 (%

)

Figure 1.4 Fetal acidosis with abnormal fetal heart rate patterns. Fetal scalp samples were

collected during the first stage of labour in 281 fetuses and their FHR patterns were

classified. A normal FHR pattern was defined as one with a baseline between 120 and

160bpm with baseline variability between 5 and 15bpm. Tachycardia was defined as a FHR

greater than 160bpm and bradycardia less than 110bpm. Complicated variable decelerations

were defined as FHR patterns with variable decelerations and either late components or a

biphasic pattern. Decreased FHR variability was defined as less than 5bpm. Decreased FHR

variability + other included FHR patterns with both decreased FHR variability and either

tachycardia and/or FHR decelerations.

(Adapted from Beard et al. 1971)

52

Figure 1.5 Fetal aerobic and anaerobic metabolism. Figure A illustrates the pathophysiology

of respiratory and metabolic acidosis both of which increase the fetal concentration of

hydrogen ions. It is only by assessing base deficit that the type of acidosis can be determined.

Figure B schematically represents fetal cellular glucose metabolism. Glucose is broken down

along the glycolytic pathway to form pyruvate. Under aerobic conditions large amounts of

energy are produced in the citric acid cycle whereas anaerobic glycolysis renders less energy

and lactic acidosis. When there is adequate oxygen supply to the fetal tissues nicotinamide

adenine dinucleotide (NAD+) accepts a hydrogen ion to produce NADH. With the catabolic

reaction of NADH, oxygen is directly consumed producing NAD+ and water. This reaction

occurs at the inner membrane of the mitochondrion with the release of large amounts of

energy to drive oxidative phosphorylation producing ATP and water. H:- = hydride ion,

containing an extra unpaired electron.

(Adapted from Nordstom and Arulkumaran, review, 1998)

DecreasedPlacental Blood

Flow

OxygenSaturation

OxygenDelivery

AnaerobicMetabolism

MetabolicAcidosis

(Lactate + H+)

H+

PCO2

RespiratoryAcidosis

CO2 + H20 H2CO3 H+ + HCO3

A

B Glucose

2 Pyruvate 2 Lactate- + 2H+

CitricAcidCycle

CO2

2 NAD+

O2

2 NADH

4H+ + O2 + 4e-

2 ADP

2 ATP

4NAD+4NADH36 ADP

36 ATP

36 ATP

CO2

2 H20

Cytosol

2H:-

4e-

Mitochondrion

54

The fetus is able to adjust to short episodes of hypoxaemia and maintain aerobic metabolism

by initiating a synchronised response that involves behavioural, cardiovascular, metabolic and

hormonal adjustments (Greene & Rosen, 1995). When these compensatory mechanisms fail

to maintain an adequate oxygen supply to the central organs, aerobic metabolism is

supplemented by anaerobic metabolism of glucose and glycogen as far as lactic acid in the

Kreb cycle, with the generation of two units of ATP for each molecule of glucose.

Anaerobic metabolism is much less efficient than aerobic metabolism but it is an important

survival mechanism particularly to maintain cardiac and brain activity during hypoxaemia and

asphyxia. This form of anaerobic metabolism is largely dependent on the pre-asphyxial

glycogen content of the myocardium and liver which are depleted in this process (Dawes et

al., 1959).

1.4.2.2.2 Fetal blood sampling and adverse clinical outcomes

As previously described, there have been 13 randomised trials comparing EFM with

intermittent auscultation. In three of these trials intermittent auscultation was compared with

EFM without the use of FBS for pH (Haverkamp et al., 1976; Kelso et al., 1978; Haverkamp

et al., 1979) and there was a four fold increase in the odds of caesarean section for fetal

distress (OR 4.14, 95%CI 2.29-7.51) without improvement in neonatal outcome (neonatal

seizures OR 0.79, 95%CI 0.21-2.97). In contrast, EFM with the option of FBS (Renou et al.,

1976; Haverkamp et al., 1979; MacDonald et al., 1985; Neldam et al., 1986; Luthy et al., 1987)

showed a less marked increase in caesarean section for fetal distress (OR 1.98, 95%CI 1.33-

2.94) and a decrease in neonatal seizures (OR 0.49, 95%CI 0.29-0.82). It should be noted

that the reduction in caesarean section for fetal distress with the addition of FBS was

primarily through postponing intervention to the second stage of labour.

Although FBS has been used in clinical practice for 40 years there is limited information

about the predictive value of the results. A review of seven studies including 1547 patients

published in 1979 concluded that scalp blood pH was a poor predictor of Apgar score of less

than seven at one minute (Banta & Thacker, 1979). They report sensitivities ranging from

28% to 62% and specificities from 70-92%. No long term follow-up data were available. A

more recent prospective cohort of 1709 patients who underwent scalp sampling for fetal pH

assessment reported similar findings with scalp pH a poor predictor of cord pH<7.0, base

deficit >16mmol/L, low Apgar scores at 1 and 5 minutes and hypoxic-ischaemic

encephalopathy (Kruger et al., 1999). The area under the receiver operator curves (ROC) and

the sensitivities and specificities for pH at predicting these poor outcomes is summarised in

55

Table 1.10. It can be seen from this table that the areas under the ROC are only marginally

better than 0.5 which reflects chance alone and the sensitivities are universally poor.

Table 1.10 Scalp pH and the prediction of neonatal outcome

Outcome Variable Area under

ROC (95%CI)Sensitivity* Specificity*

Umbilical artery

pH < 7.0 0.64 (0.62-0.70) 35.7% 83.8%

Base deficit > 16mmol/L 0.69 (0.66-0.71) 54.2% 85.4%

Apgar score

<7 at 1 min 0.63 (0.60-0.65) 32.1% 85.4%

<7 at 5 min 0.63 (0.61-0.66) 23.5% 82.9%

<4 at 6 min 0.55 (0.52-0.57) 22.2% 82.4%

Hypoxic-ischaemic encephalopathy

Mild 0.57 (0.54-0.60) 33.3% 82.7%

Moderate or severe 0.66 (0.63-0.69) 50.0% 80.2% *Sensitivity and specificity are presented for scalp pH values at the 25th percentile (<7.21) (Adapted from Kruger et al. 1999)

1.4.2.2.3 Limitations of fetal blood sampling

There are many limitations associated with fetal scalp blood sampling. Although the use of

FBS has been widely advocated to improve the specificity of FHR monitoring, it remains an

unpopular procedure (Clark & Paul, 1985) because it is inconvenient for both the clinician

and mother and invasive for the mother and fetus. Consequently, the frequency of its use

varies dramatically between Europe, North America and Australia, from unit to unit, and

even within the same unit, depending on the attitude of individual clinicians (Greene, 1999).

In addition to being unpopular amongst some clinicians there are a number of important

practical limitations to the use of FBS. First, to perform the testing each obstetric unit

requires a blood gas analyser capable of analysing micro samples, preferably located in the

delivery suite. These machines cost from AUS$20,000 to AUS$80,000 depending on

whether they measure pH alone or pH, oxygen and carbon dioxide tensions, and they have

ongoing maintenance costs. This alone limits the utilisation of this method of testing to

larger units.

56

Second, to determine pH and base deficit, which is required to identify metabolic acidosis,

requires between 35 and 130µL of blood depending on the analysis technique. Obtaining

these volumes of blood can be technically challenging from a distant fetal scalp as was found

in one study evaluating scalp pH, where adequate samples (35µL) could be collected only in

80% of cases where assessment was attempted (Westgren et al., 1998). If inadequate sample

volumes are collected one is limited to either measuring pH alone or opting for intervention

to expedite delivery which may be unnecessary.

The third practical limitation to FBS for pH assessment relates to specimen contamination.

If the sample is contaminated with air bubbles, meconium or amniotic fluid, invalid results

will be obtained during analysis. Air bubbles can confound the analysis of gas tensions which

are required for the calculation of base deficit. Amniotic fluid has a pH ranging from 7.0 to

7.5 at term (Brace, 1999) and therefore contamination can yield both false positive and

negative results.

The final limitation to FBS for pH assessment relates to the relationship between scalp pH

and central arterial pH. A number of studies have investigated this relationship in animals

(Adamsons et al., 1970; Shier & Dilts, 1972) or humans (Stephens et al., 1969; Bowe et al.,

1970; Yoshioka & Roux, 1970; Sturbois et al., 1977) and the conclusion reported by most is

that there is a good correlation between fetal scalp pH values and central pH status (Greene,

1999). Detailed review of these studies demonstrates that the relationship between central

and peripheral acid-base status is not constant and is affected by many things including

hypoxia and local tissue conditions.

The original studies performed in monkeys during labour reported no difference between

exteriorised scalp and carotid arterial pH, whereas there was a significant difference between

transvaginal acquired scalp pH and central pH (Adamsons et al., 1970). However, when

samples were collected transvaginally, scalp pH was lower than carotid arterial pH by an

average of 0.047 pH units, with the difference becoming greater at lower pH values

(Adamsons et al., 1970). A similar discrepancy was reported in an ovine model of asphyxia

where scalp pH levels and carotid arterial and jugular venous pH levels were assessed under

both normoxic and hypoxic conditions in partially exteriorised fetuses (Shier & Dilts, 1972).

During normoxia scalp pH was, on average, 0.02 pH units lower than arterial pH whereas

during hypoxia this difference became greater until the mean difference was 0.07 pH units.

Furthermore, in this study scalp carbon dioxide tensions (PCO2) did not correlate with either

arterial or venous levels, which were consistently higher (Shier & Dilts, 1972). Similarly, scalp

base deficit levels were not representative of values in either arterial or venous blood. This

57

inaccuracy is probably explained by the poor PCO2 correlation, as base deficit is calculated

from pH and PCO2. Human studies comparing scalp pH and central pH reported no

significant differences between these two variables; however, the difference between them

became greater in “distressed fetuses” (Bowe et al., 1970; Yoshioka & Roux, 1970). The lack

of statistical significance of the difference identified in the study of Yoshioka et al. (1970) may

relate to the small sample size (non-distressed fetuses n=9, distressed fetuses n=5).

It has been noted by a number of authors that there are several situations in which scalp pH

measurements are inaccurate (Bowe et al., 1970; Parer & Livingston, 1990; Greene, 1999) .

These include scalp oedema and caput succedaneum which can result in scalp pH levels

being 0.026 to 0.06 pH units lower than carotid arterial or umbilical arterial pH. This can

confound interpretation of scalp pH results in as many as 10-20% of samples (Parer &

Livingston, 1990). Apparent inconsistencies between peripheral and central pH that may

occur with hypoxia and the presence of scalp oedema or caput succedaneum in labour may

partially explain the poor predictive ability of scalp pH with regard to the important neonatal

outcomes of severe metabolic acidosis and hypoxic-ischaemic encephalopathy.

1.4.2.2.4 Scalp stimulation as an alternative to fetal blood sampling

There is good evidence to suggest that stimulating the fetal scalp during a vaginal

examination during labour will reduce the need for fetal scalp sampling (Clark et al., 1982,

1984; Elimian et al., 1997; Porter & Clark, 1999; Skupski & Eglinton, 2002). A total of 512

studies have been performed in this area and a recent meta-analysis of the 11 high quality

randomized controlled trials reported 97% sensitivity and 50% specificity for a stimulated

FHR acceleration predicting normal fetal acid-base status in the fetus (Skupski & Eglinton,

2002). Acceleration in the FHR after scalp stimulation is reassuring and negates the need to

perform FBS. The absence of FHR acceleration is less helpful in clinical decision-making,

with 27% of these fetuses being acidotic on FBS. The utilisation of scalp stimulation is also

not assisted by FHR patterns with frequent decelerations or fluctuating baseline FHR where

identification of the required stimulated acceleration is sometimes difficult (Dixon et al,

1999).

58

1.4.2.3 Umbilical cord blood gas analysis

1.4.2.3.1 Physiological basis

As discussed in previous sections of this chapter, umbilical cord acid-base assessment at

delivery has a sound physiologic basis and provides a sensitive reflection of birth asphyxia

(Low, 1998; MacLennan, 1999). Eleven studies have reported umbilical cord arterial blood

gas values encompassing more than 35,000 deliveries (Huisjes & Aarnoudse, 1979; Sykes et

al., 1982; Eskes et al., 1983; Yeomans et al., 1985; Low, 1988; Ruth & Raivio, 1988; Ramin et

al., 1989; Thorp et al., 1989; Riley & Johnson, 1993; Nagel et al., 1995; Helwig et al., 1996).

These studies suggest that the reference ranges for all of the components of blood gas

analysis are wide and the lower limit for normal cord arterial pH should be between 7.04 and

7.10 and the lower limit of venous pH should be 7.14 to 7.20. Ideally paired arterial and

venous samples should be collected from the umbilical cord immediately after delivery

because a difference in the partial pressure of carbon dioxide of more than 25mmHg

suggests acute rather than chronic acidosis (Belai et al., 1998).

1.4.2.3.2 Prediction of adverse outcomes

Metabolic acidaemia at birth (pH<7.0, base deficit >12mmol/L) is relatively common

occurring in 2% of all births (Ruth & Raivio, 1988; Low et al., 1994). Severe metabolic

acidaemia with a base deficit greater than 16mmol/L is less common occurring in 0.05% of

all births (Low et al., 1994) and it has been suggested that this is a realistic cut off point for

defining pathological fetal acidaemia that correlates with an increasing risk of neurological

deficit (Goldaber et al., 1991; Sehdev et al., 1997). There is growing evidence that base deficit

values in cord arterial blood are significantly more useful than umbilical cord pH values

because base excess does not change significantly with respiratory acidosis, and demonstrates

a linear, rather than logarithmic correlation to the degree of metabolic acidosis (Low, 1998;

Ross & Gala, 2002).

As described in detail in Section 1.4.1.4, babies born with severe metabolic acidosis in cord

arterial blood are significantly more likely to suffer short term neonatal complications than

babies with normal acid-base status (Gilstrap et al., 1989; Low et al., 1995; van den Berg et al.,

1996). The outcomes for babies born with severe respiratory acidosis are no different from

babies born with normal umbilical arterial blood gas results (Low et al., 1994). In long term

follow-up studies, the association between metabolic acidaemia at birth and

neurodevelopmental outcome is poor with all but one study reporting no significant

59

association between acidaemia and neurodevelopmental disability (Low et al., 1983, 1988;

Ruth & Raivio, 1988; Socol et al., 1994; Ingemarsson et al., 1997).

1.4.2.3.3 Limitations of umbilical cord blood gas analysis

Umbilical cord blood gas assessment at birth is a sensitive marker of asphyxia, but its use is

somewhat limited because traditional markers of fetal acid-base status reflect fetal metabolic

adjustments that are not directly related to fetal damage unless the values are extreme

(Gilstrap et al., 1989; Goldaber et al., 1991; Winkler et al., 1991). Even a pH less than 7.0 is

poorly predictive of neurological deficit because a severely acidotic infant will usually have a

normal outcome (Winkler et al., 1991; Socol et al., 1994; Yudkin et al., 1994).

Other limitations of this technique relate to the cost of machines which perform blood gas

analysis and the issues of accurate collection of samples. To reliably perform blood gas

analysis on all deliveries, each delivery suite would require a blood gas machine on site. The

cost of this equipment precludes many smaller units from establishing this technique.

Second, for these results to be clinically useful the collection technique is critical and despite

staff eduction this has proven difficult in a number of units which report that 25% of their

results are invalid (Westgate et al., 1994).

1.4.3 Summary

Each of the three techniques that are widely used for the assessment of fetal wellbeing in

labour have a number of limitations.

Electronic fetal heart rate monitoring is associated with a significant increase in obstetric

intervention. It is only with the addition of fetal scalp blood sampling for acid-base

assessment that there is any improvement in neonatal outcome. Such combined monitoring

results in a 50% reduction in the number of babies with neonatal seizures. No studies of

EFM or FBS have shown a reduction in neonatal encephalopathy or cerebral palsy and even

though multiple late decelerations and decreased FHR variability have been shown to be

associated with the subsequent development of cerebral palsy, only 0.19% of singletons with

these patterns develop cerebral palsy giving a false positive rate of 99.8%.

Intrapartum fetal blood sampling, the one technique that has been shown to limit the

increase in intervention associated with EFM and reduce neonatal seizures, is unpopular and

the cost of the equipment required to utilise this technique limits its use primarily to tertiary

obstetric centres. Similarly, although umbilical cord blood gas assessment at birth is a

60

sensitive marker of asphyxia, the machines to perform this routinely are expensive. Babies

born with severe metabolic acidosis, as identified by umbilical cord gases, usually have

normal outcomes.

In order to reduce the burden of birth asphyxia on families, health care systems and the

community, new techniques that are inexpensive and widely available are required to allow

recognition of early intrapartum asphyxia so that timely obstetric intervention can avoid

asphyxia-induced brain damage or death in the newborn.

61

1.5 Lactate and fetal surveillance

My thesis describes studies designed to explore the role of lactate measurement in

intrapartum fetal surveillance. The preceding sections of this chapter have outlined the

importance of anaerobic metabolism and metabolic acidosis in perinatal asphyxia and brain

injury. Lactate is the end-product of anaerobic metabolism and it is the main fixed organic

acid in metabolic acidosis and the main contributor to base deficit (Nordstrom, 2001).

The first study investigating lactate measurement for intrapartum assessment of wellbeing

was performed in 1970 when it was noted that scalp lactate levels were higher in “distressed”

babies than those who were “not distressed” even though there was no difference in pH

(Yoshioka & Roux, 1970). The authors of this study concluded that scalp lactate may be an

earlier indicator of actual or impending fetal distress than scalp pH alone and the

measurement of scalp lactate was proposed as a test to discriminate when in doubt about the

fetal condition in labour. Despite this and other studies in the 1970’s and 1980’s which

suggested that lactate was at least as effective as pH at identifying fetal distress (Yoshioka &

Roux, 1970; Gardmark, 1974; Eguiluz et al., 1983; Smith et al., 1983; Schmidt, 1988), lactate

assessment did not become established in clinical practice. In all likelihood, this is because of

the amount of blood required for analysis. A blood sample volume of 150µL or more was

required, followed by haemolysing or spinning the blood sample and immediate

transportation of the sample to a biochemistry department for analysis (Soutter et al., 1978).

It has only been with the development of micro-volume test strip lactate meters that

sampling and analysis have been adapted to clinical obstetric practice (Gambke & Muller,

1993; Shimojo et al., 1993; Nordstrom et al., 1998a; Cobbaert et al., 1999; Pennell & Tracy,

1999; Sinn et al., 2001).

1.5.1 Lactate physiology

1.5.1.1 Lactate production

Lactate production and its effects on the mother and fetus are more complex than originally

thought. Lactate can be produced both under hypoxic and non-hypoxic conditions.

1.5.1.1.1 Hypoxic conditions

Lactate is the end-product of anaerobic metabolism. During anaerobic glycolysis, glucose is

broken down to pyruvate and converted to lactate and H+ ions. As described previously and

illustrated in Figure 1.5, the glycolytic process is dependent on the intracellular concentration

62

of NADH. NAD+ is the most important hydrogen carrier in catabolic reactions. It accepts a

hydrogen atom together with an additional electron, a hydride ion (H:-), forming NADH.

The conversion of NAD+ to NADH occurs during the reactions converting glucose to

pyruvate. The NADH molecule is unstable and the electron is easily transferred to other

molecules. To maintain glycolysis, NADH is oxidised to produce NAD+ which is required

for the further conversion of glucose to pyruvate. The hydride ion is used in the conversion

of pyruvate to lactate and H+. Energy in the form of 2 ATP molecules is produced, as

compared with 36 ATP molecules when pyruvate enters the citric acid cycle.

Fetal lactate concentrations increase during hypoxic conditions and it is likely that this lactate

is derived from the fetus (Milley, 1988); however, the exact source of lactate production is

difficult to identify. As previously discussed, the fetal response to hypoxia is to prioritise its

circulation to the heart, brain and adrenals whereas splanchnic and soft tissue circulations are

reduced. This hypothesis is supported by recent animal studies which showed that ATP

depletion was most severe in the kidneys compared with the brain and heart during graded

perinatal asphyxia (Seidl et al., 2000).

There is a steady state relationship between pyruvate and lactate under normal conditions

(1:10); however, the ratio increases with hypoxia (Marx & Greene, 1964; Low et al., 1979).

The value of this ratio does not appear to predict oxygen deprivation better that determining

the lactate level (Daniels, 1966).

1.5.1.1.2 Non-hypoxic conditions

Lactate production usually takes place in hypoxic tissue, but it can also be produced under

aerobic conditions. Lactate production can be induced by glycogenolysis and thus lactate

levels correlated with that of glucose. There is a steady state relationship between pyruvate

and lactate levels meaning that hyperglycaemia can produce lactic acidaemia even if oxygen is

available.

The experimental induction of fetal hyperglycaemia by high dose glucose infusions

(Gardmark, 1974; Kenepp et al., 1982; Philipson et al., 1987) or beta-mimetic drugs (Lunell et

al., 1977) induces glycogenolysis and may increase lactate levels even under aerobic

conditions. Maternal glucose infusions which are sometimes used in labour (5% Dextrose,

180mL/hr) do increase maternal and fetal glucose concentrations but do not alter maternal

or fetal lactate levels or acid-base status (Nordstrom et al., 1995a).

63

The “stress of being born” induces high levels of catecholamines in the course of normal

labour and delivery (Hagnevik et al., 1984; Lagercrantz & Slotkin, 1986). This catecholamine

surge leads to glycogenolysis in the fetal liver with concomitant hyperglycaemia; however, no

related lactate increase has been found, either in normal (Nordstrom et al., 1996b) or

complicated deliveries (Nordstrom et al., 1998b).

Maternal hyperventilation can also influence lactate levels. Decreasing maternal CO2 results

in bicarbonate depletion through the carbonic anhydrase buffering system. Maternal pH will

then increase as respiratory alkalosis develops. To maintain electroneutrality during the

depletion of the negatively charged bicarbonate ions, lactate serves as a supplementary source

of anions. Clinically, maternal hyperventilation can result in both maternal and fetal lactate

levels increasing by up to 1mmol/L (Huch, 1986).

1.5.1.2 Fetal lactate levels during pregnancy and delivery

Unlike pH, PO2, PCO2 and base excess, which all change with advancing gestation, the

concentration of lactate in umbilical blood samples obtained by cordocentesis in

normoxaemic healthy human fetuses does not change over the gestational age range of 17 to

38 weeks (Nicolaides et al., 1989). The levels obtained at cordocentesis are also not different

from the levels obtained at elective caesarean section at term (Nicolaides et al., 1989).

Animal and human studies suggest that there is a fetal consumption of lactate that probably

originates from uteroplacental metabolism of glucose (Burd et al., 1975; Char & Creasy, 1976;

Schneider et al., 1981; Nicolaides et al., 1989). In humans before labour, lactate in umbilical

venous blood is 7% higher than in umbilical arterial blood. In fetal lambs, lactate is

aerobically produced in the placenta and there is net transfer to the maternal and fetal

circulations (Battaglia & Meschia, 1986).

Lactate is likely to serve as a source of fuel for the fetus. In the human fetus, oxidative

metabolism of lactate represents 10-25% of the overall metabolic activity (Fowden, 1994). In

the fetal lamb, the heart uses lactate as a substrate for one third of myocardial oxygen

consumption (Bartelds et al., 2000).

Lactate equilibrates slowly over the placenta and the maternal lactate level has limited effect

on the fetal level, except during severe maternal hyperventilation or prolonged second stage

of labour (Kalinkov et al., 1980). Transfer of lactate between the maternal and fetal

circulations occurs in both directions although at a much slower rate than for transfer of

carbon dioxide (Daniels, 1966). A lactate carrier mechanism has been demonstrated on both

64

the maternal and fetal surfaces of the trophoblast (Carstensen et al., 1983) and the transfer

mechanism is probably through facilitated diffusion coupled with proton transfer (Piquard et

al., 1990).

Several early studies suggested that fetal lactate levels increased during labour (Jacobsen &

Rooth, 1971; Suidan et al., 1984b), but recent studies have found no correlation between

cervical dilatation and fetal lactate levels, nor have they found any change in fetal lactate

levels during the first stage of labour (Smith et al., 1983; Nordstrom et al., 1994; Nordstrom et

al., 1995b). There are increases in both maternal and fetal lactate levels during the active

component of second stage labour (Kalinkov et al., 1980; Nordstrom et al., 2001) which may

account for the results found by the earlier investigators who may not have differentiated

first and second stages of labour. The increase in lactate during the second stage of labour is

partly explained by transplacental injection of maternal lactate and partly due to increasing

duration of intermittent fetal hypoxia during maternal pushing (Nordstrom et al., 1996a;

Nordstrom et al., 2001).

It is well established that during fetal distress, the increase in fetal lactate level is derived from

the fetus (Nordstrom et al., 1996a). Higher scalp blood lactate and umbilical arterial lactate

levels were found in babies with non-reassuring FHR patterns, when compared with babies

with reassuring FHR patterns (Nordstrom et al., 1995b). Fetal hypoxia may also result in an

increase of the umbilical cord arterio-venous lactate gradient (Suidan et al., 1984a). In animal

studies, the net transfer of lactate to the fetus during experimental hypoxia is reduced to zero

(Milley, 1988). Thus endogenous fetal lactate production occurs in response to hypoxia.

1.5.1.3 Negative effects of lactate

From studies in hypoxic monkeys, Myers suggested that lactate was toxic for brain tissue

causing oedema and necrosis in the late gestation fetus (Myers, 1977). In-vivo magnetic

resonance spectroscopic (MRS) studies of fetal lambs undergoing experimental hypoxia have

shown that lactate concentrations in the fetal brain begin to increase, due to anaerobic

metabolism, at an arterial pH of less than 7.28 or a base deficit over 8mmol/L (van

Cappellen et al., 1999).

1.5.1.4 Positive effects of lactate

Lactate has traditionally been thought of as the “dead end” product of anaerobic metabolism

but recent studies suggest that it can also have positive effects. Lactate can dilate human

placental arteries in the chorionic plate, a mechanism that might be of importance in fetal

adaptation to hypoxia (Omar et al., 1993). Lactate may also have a role in attenuation of

65

brain injury. When L-lactate was infused into rats prior to experimentally-induced brain

injury, the cortical injury was smaller in size (Ros et al., 2001) and there were fewer cognitive

deficits two weeks later (Rice et al., 2002). The positive effects of lactate on the brain may be

explained by the finding that neurons are capable of utilising lactate as an energy source. In

vitro studies have demonstrated the ability of neurons to carry out synaptic function with

lactate as the only available carbon source (Schurr et al., 1997).

1.5.2 Fetal scalp lactate

There are few studies which have investigated the relationship between lactate concentrations

in blood from the fetal scalp lactate and central circulation. This paucity of data is probably

due to the volume of blood that was required for lactate measurement and the wet chemistry

techniques which were necessary prior to the development of micro volume test strip lactate

meters.

The relationship between fetal scalp lactate concentrations and umbilical artery and vein

lactate concentrations have been investigated in two human studies (Yoshioka & Roux, 1970;

Kruger et al., 1998). The first of these studies compared lactate concentrations from these

three sampling sites in 14 human fetuses, six who were delivered with forceps vaginally and 8

who were delivered by caesarean section (Yoshioka & Roux, 1970). The scalp blood samples

were collected one to three minutes prior to delivery of the fetus and collection of the

samples from the umbilical cord. There were no significant differences in the mean lactate

concentrations from the three sampling sites in either the fetuses delivered vaginally or those

delivered by caesarean section. Of note, the lactate concentrations in blood from the fetal

scalp and umbilical artery were identical in samples collected at caesarean section whereas

scalp lactate concentration was lower than arterial lactate concentration for fetuses delivered

vaginally. These differences were not statistically significant; however, only six blood samples

were analysed. In a subsequent study, 103 scalp lactate concentrations were compared with

umbilical artery and vein lactate concentrations sampled within 60 minutes of delivery.

Lactate concentration in fetal scalp blood before delivery showed a significant correlation

with lactate levels in the umbilical artery (r=0.65, p<0.001) and venous blood (r=0.62,

p<0.001) (Kruger et al., 1998). In both these human studies no paired statistical analysis was

performed for blood collected from the same individual but different sampling sites. Unlike

scalp pH assessment, there have been no studies in animal models comparing fetal scalp

lactate concentration with carotid arterial and jugular venous blood lactate concentration.

There have also not been any studies which have investigated the relationship between

66

different sampling sites for lactate concentration and fetal hypoxaemia or acidaemia, two

known confounders of scalp pH assessment.

Normal lactate levels derived for human fetuses in the first stage of labour with reactive FHR

patterns prior to their FBS are presented in Table 1.11. The variation between studies almost

certainly relates to the different sample types and measuring techniques used in the each

study; some analysed plasma whilst others analysed whole blood or haemolysed blood. The

highest levels were measured in plasma, which is to be expected because lactate is unevenly

distributed in plasma and erythrocytes with the concentrations in erythrocytes only 50-60%

of that in plasma (Devadatta, 1934; Decker & Rosenbaum, 1942; Sahla et al., 1964). The

higher levels in the haemolysed blood may relate to production of lactate by erythrocytes,

which can occur during storage of blood in a test tube prior to analysis. Issues relating to

delay in sample analysis are abolished by the new bedside micro volume test strip methods.

Table 1.11 Normal ranges for scalp lactate in first stage of labour

Authors Mean lactate

(mmol/L)

95th percentile

Sample size (N) Sample type

Jacobsen & Rooth, 1971 3.0 5.0 45 Plasma

Schmidt, 1973 2.9 3.2 20 Haemolysed blood

Eguiluz et al., 1983 1.0 NA 52 Whole blood

Smith et al., 1983 1.8 2.9 97 Haemolysed blood

Nordstrom et al., 1994 1.7 3.3 66 Whole blood

Nordstrom et al., 1995b 1.2 2.9 64 Whole blood NA = not available There have been a number of observational studies which have shown that scalp lactate

performs as well as pH in predicting low Apgar scores at delivery (Eguiluz et al., 1983; Smith

et al., 1983; Nordstrom et al., 1995b). There has been one randomised clinical trial of 341

parturients, comparing fetal scalp blood lactate and pH in the diagnosis of intrapartum fetal

distress (Westgren et al., 1998). No difference was identified between the two groups with

respect to mode of delivery, intervention rates, neonatal outcome or umbilical artery acid-

base status or lactate. The positive findings of the trial were that scalp lactate assessment was

quicker, required significantly fewer scalp incisions, and had a lower blood sampling failure

rate than scalp pH assessment. Lactate analysis required 5µL of blood whereas pH required

35µL. Sampling failure rates were 1.2% and 20.6% respectively (OR 16.1, 95% CI 5.8-44.7).

67

There has been one large retrospective observational study investigating the predictive values

of 1709 scalp lactate and pH samples as markers of neurological disability (Kruger et al.,

1999). The pH results have previously been discussed in section 1.4.2.2.2 and are presented

in Table 1.10. Scalp lactate was a significantly better predictor than scalp pH of Apgar score

less than 4 at 5 minutes (area under ROC: lactate 0.79, pH 0.55, p=0.033) and, more

importantly, moderate and severe hypoxic-ischaemic encephalopathy (area under ROC:

lactate 0.95, pH 0.64, p=0.015). The sensitivity and specificity for lactate to predict severe

metabolic acidaemia in cord arterial blood, low Apgar scores and hypoxic-ischaemic

encephalopathy are summarised in Table 1.12. Comparing Tables 1.10 and 1.12 it can be

seen that scalp lactate assessment has higher sensitivity for predicting these adverse outcomes

than scalp pH with no significant loss in specificity. There were six cases in this series that

had hypoxic brain injuries and five had both scalp lactate and pH measurements performed.

Four of the five had elevated levels of scalp lactate; only two of the five had low levels of

scalp blood pH. There are a number of possible explanations for the discrepancy between

the hyperlactaemic and acidaemic fetuses. First, it has been suggested by animal studies of

asphyxia that subcutaneous lactate increases earlier than the pH decrease, and precedes the

brain tissue lactate increase (Engidawork et al., 1997). Second, the difference may relate to

the slower time to normalise lactate compared to pH after hypoxic insults (Ikeda et al., 1998a;

Keunen et al., 1999; Kaneko et al., 2003). This has been illustrated by a recent study, in which

repeated cord occlusion of four minutes every 90 minutes for a total duration of six hours

was performed in fetal sheep. Significant reductions in pH occurred in response to each cord

occlusion; however, by the time of next cord occlusion pH had normalised. The pH before

the first cord occlusion was 7.33±0.01 and before the last occlusion it was 7.30±0.02. The

pH response was very different from the lactate response in this study; there was a

progressive rise in lactate with each cord occlusion such that before the first occlusion lactate

was 1.21±0.21mmol/L and before the last occlusion it had increased to 6.1mmol/L (Kaneko

et al., 2003). It appears from these data that lactate is a better marker than pH for

intermittent acute insults because pH can be normalised by both the placenta and buffering

systems whereas the slow clearance of lactate by the placenta affords it a greater ability to act

as an indirect marker of these insults.

68

Table 1.12 Scalp lactate and prediction of neonatal outcome

Outcome Variable Area under ROC (95%CI) Sensitivity* Specificity*

Umbilical artery

pH < 7.0 0.72 (0.69-0.75) 78.6% 74.4%

Base deficit > 16mmol/L 0.78 (0.75-0.81) 70.8% 75.5%

Apgar score

< 7 at 1 min 0.68 (0.65-0.71) 49.1% 76.2%

< 7 at 5 min 0.68 (0.65-0.71) 52.9% 73.5%

< 4 at 6 min 0.71 (0.68-0.74) 77.8% 73.5%

Hypoxic-ischaemic encephalopathy

Mild 0.70 (0.66-0.72) 66.7% 73.0%

Moderate or severe 0.83 (0.80-0.85) 100% 72.8% *Sensitivity and specificity for scalp lactate values at the 75th percentile (>4.8mmol/L). (Adapted from Kruger et al. 1999)

1.5.3 Umbilical cord lactate levels at delivery

There are several reasons for performing routine cord blood acid-base balance or lactate

measurements at delivery: a quality assessment of obstetric practice; a diagnostic test for

hypoxia in cases of depressed newborns; and for the prediction of neonatal complications

and long-term outcome. There have been several reports of routine cord blood lactate

measurements (Westgren et al., 1995; Chanrachakul et al., 1999; Pennell & Tracy, 1999).

Umbilical artery lactate measurement has been shown to be equally as good as pH and base

excess at predicting low Apgar scores at 1 and 5 minutes (Westgren et al., 1995). Umbilical

artery lactate levels are significantly better predictors of hypoxic-ischaemic encephalopathy

than pH (Chou et al., 1998; Kruger et al., 1999).

1.5.4 Lactate levels in the newborn

There have been a number of studies that have investigated lactate levels in the newborn.

Early postnatal magnetic resonance spectroscopy (MRS) examinations of lactate levels have

shown a significant correlation with poor outcome (death or neurologic deficit) (Leth et al.,

1996). Urinary lactate-creatinine ratio at 6 hours of age assessed by proton nuclear MRS have

also been shown to have both high sensitivity (94%) and specificity (100%) for predicting

hypoxic-ischaemic encephalopathy in a group of babies with perinatal asphyxia (Huang et al.,

1999).

69

1.6 Thesis outline

1.6.1 Aim

The aim of this thesis is to further evaluate the role of lactate in intrapartum fetal

surveillance.

1.6.2 Studies performed

1.6.2.1 Evaluation of a new method for rapid assessment of lactate in fetal and

neonatal blood

It has only been with the development of micro-volume test strip lactate meters that

sampling and analysis of lactate have been adapted to clinical obstetric practice. When

commencing my candidature no micro volume lactate meter that had been evaluated in fetal

and neonatal blood was available in Australia. Therefore the aim of my first series of studies

was to determine the accuracy and precision of the commercially available Accutrend®

handheld reflectance photometer for measuring whole blood lactate levels in umbilical cord

blood and neonatal blood micro samples across the gestational age range of 26 to 42 weeks.

1.6.2.2 The relationship between neuropathologic damage and biochemical and

physiologic markers of asphyxia induced by graded cord occlusion in

chronically instrumented ovine fetuses.

The aim of my second series of experiments was to use an ovine model of repeated umbilical

cord occlusion (UCO) to evaluate the relationship between fetal lactate responses and fetal

brain injury. To address this aim I have utilised an UCO protocol, designed to mimic events

which may occur in labour, to compare the relationship between:

1. fetal acid-base status and neuropathologic damage,

2. fetal lactate and neuropathologic damage,

3. fetal heart rate patterns (FHR) and neuropathologic damage, and

4. fetal arterial pressure and neuropathologic damage.

70

1.6.2.3 The relationship between fetal scalp lactate levels, central lactate levels and

oxygen free radical activity in the fetal brain during asphyxia induced by

graded cord occlusion in the ovine fetus

Fetal scalp blood sampling is the principal technique utilised to collect blood for biochemical

assessment of fetal wellbeing during labour. Prior to the potential introduction of a new

biochemical marker of fetal wellbeing into clinical practice, two important questions require

investigation. First, what is the relationship between fetal scalp blood levels and central

blood levels of the new biochemical marker and is this relationship altered by asphyxia?

Second, do fetal scalp levels rather than central levels of the new marker predict fetal brain

injury? The purpose of the series of studies presented in this chapter was to address these

two important questions. Prior to exploring the peripheral-central lactate relationships, I

have compared and contrasted the acutely instrumented model used in this study with the

chronically instrumented model described in the previous chapter. To investigate the

relationship between fetal scalp lactate and brain injury I have investigated the relationship

between fetal lactate levels and oxygen free radical activity in the fetal brain. This approach

was chosen because there is growing evidence at a molecular level that oxygen free radicals

are critically involved in the mechanisms of hypoxic-ischaemic brain injury (Maulik, 1998;

Volpe, 2001b; du Plessis & Volpe, 2002; Gitto et al., 2002; Inder et al., 2002) and because of

the terminal nature of the experiments. Oxygen free radical activity was assessed both

during and after the asphyxial insult and related to scalp lactate and other biochemical

markers of fetal wellbeing.

1.6.2.4 Fetal lactate responses in growth restricted ovine fetuses during asphyxia

induced by graded cord occlusion

There is growing evidence that an adverse intrauterine environment alters fetal physiologic

responses during asphyxia (Gardner et al., 2002; Gardner et al., 2003). Fetal growth restriction

secondary to placental insufficiency is associated with an adverse intrauterine environment

which includes hypoxaemia, hypercapnia, hyperlacticaemia, and acidaemia (Nicolaides et al.,

1989; Marconi et al., 1990). Growth restricted fetuses have also been shown to have altered

lactate metabolism (Marconi et al., 1990), glucose metabolism (Jansson et al., 1993; Marconi et

al., 1996; Pardi et al., 2002), amino acid metabolism (Cetin et al., 1996; Marconi et al., 1999)

and fatty acid metabolism (Cetin et al., 2002) when compared to appropriately grown

controls. Human growth restricted fetuses are at increased risk of non-reassuring FHR

patterns during labour (Low et al., 1976; Dashe et al., 2000), fetal acidosis in fetal scalp blood

samples collected during labour (Nieto et al., 1994), metabolic acidosis in umbilical arterial

blood samples collected at delivery (Low et al., 1981) and neonatal encephalopathy secondary

71

to acute intrapartum hypoxic events (Bukowski et al., 2003). The purpose of the series of

studies presented in this chapter was to explore fetal lactate responses and central-peripheral

lactate relationships in growth restricted ovine fetuses during asphyxia.

72

C h a p t e r 2

EVALUATION OF A NEW METHOD FOR RAPID ASSESSMENT

OF LACTATE IN FETAL AND NEONATAL BLOOD

The best preparation for tomorrow is to do today’s work superbly well.

Sir William Osler (1892)

2.1 Introduction

Fetal scalp blood lactate measurement for the assessment of fetal wellbeing has been

investigated since the 1970’s (Yoshioka & Roux, 1970; Gardmark, 1974; Eguiluz et al., 1983;

Smith et al., 1983; Schmidt, 1988). Despite initial studies suggesting that lactate was at least as

effective as pH at identifying fetal distress as reflected by low Apgar scores (Smith et al.,

1983), lactate assessment has not become established in clinical practice. In all likelihood, the

reluctance to adopt this technique probably related to the amount of blood required for

analysis. Until the recent development of micro-volume test strip methods (Gambke &

Muller, 1993; Shimojo et al., 1993; Nordstrom et al., 1998a; Cobbaert et al., 1999; Sinn et al.,

2001), at least 150µl of blood has been required for analysis (Soutter et al., 1978) which is far

greater than the 20 to 50µl volume routinely available when collecting fetal scalp blood

samples.

Lactate concentrations vary in different body compartments. This is important because

different methods of lactate measurement sample different body compartments. Lactate is

unevenly distributed between plasma and erythrocytes with the concentration of lactate in

erythrocytes being 50-60% of the plasma concentration (Devadatta, 1934; Decker &

Rosenbaum, 1942; Sahla et al., 1964). Thus within an individual, plasma has the highest

lactate concentration with lower concentrations in haemolysed blood and the lowest

concentration in whole blood with intact erythrocytes.

Until 1986, whenever blood was to be analysed for lactate levels, samples required

stabilisation for transport to the laboratory and wet chemistry techniques involving

precipitation with perchloric acid in order to obtain lactate concentrations from both

erythrocytes and plasma. In recent years micro volume (5-25µl) test strip based analysis

73

methods have been developed which use capillary whole blood with intact erythrocytes (Weil

et al., 1986; Shimojo et al., 1989; Wandrup et al., 1989; Gambke & Muller, 1993; Shimojo et al.,

1993). These newer meters are easy to use and provide both accurate and precise

measurement in adult, fetal and neonatal blood. Unfortunately, when commencing my

candidature, none of these meters were commercially available in Australia.

During the 1990’s lactate electrodes were developed for micro sample blood-gas analysers.

The electrodes measure lactate with amperometric techniques where the current generated

from a series of biochemical reactions is proportional to the lactate concentration. The lactate

electrodes are both accurate and precise (Toffaletti et al., 1992); however, these machines are

expensive, ranging in price from AUS $50,000 -$80,000. These machines require biomedical

staff for maintenance and sample stabilisation if delays are encountered between sampling

and testing.

For lactate to be a practical alternative to the current biochemical techniques for the

assessment of fetal wellbeing during labour, the measurement system needs to meet a

number of criteria. First, the system should be inexpensive, easy to use, and suitable for

primary, secondary and tertiary obstetric units. Second, the measurement system should

analyse small samples of fetal blood (approximately 20µl), preferably at the bedside, and

require minimal ongoing maintenance. Third, the measurement system should be accurate

and precise across a range of ages from 24 to 42 weeks gestation for samples from fetal scalp,

umbilical cord and neonatal sampling sites. Finally, the measurement system should be

accurate across a range of haematocrit concentrations. This is particularly important as the

haematocrit of fetal and neonatal blood is influenced by gestational age, fetal growth

restriction, sampling location and labour (Jacobsen & Rooth, 1971; Soothil et al., 1987;

Economides & Nicolaides, 1990; Weiner et al., 1991; Brace, 1992). Errors due to varying

haematocrit levels are a flaw of many lactate analysers including those which use

amperometric techniques (Toffaletti et al., 1992) and test strip methods (Nordstrom et al.,

1998a).

The Accutrend® lactate meter (Roche Diagnostics, Mannheim, Germany) is an inexpensive

handheld unit that costs approximately AUS$900 and analysis of each sample costs AUS$1.

The unit measures 115mm x 62mm x 18.5mm and analyses sample volumes as small as 15µl

with a measurement time of 60 seconds. The meter uses reflectance photometry to measure

lactate from 0.8 to 22mmol/L for blood and 0.7 to 26mmol/L for plasma. It performs at a

temperature range of 5-35°C for blood lactate values less than 8mmol/L and 15-35°C for

blood lactate values greater than 8mmol/L

74

The Accutrend® meter has been compared with many laboratory lactate assessment

techniques and found to be accurate when measuring lactate in adult capillary blood samples

from 0.8 to 22mmol/L (Gambke & Muller, 1993). Comparisons with wet chemistry

techniques for lactate determination in horse plasma have also been performed and the

results from the two techniques were highly correlated (r=0.928; Williamson et al, 1996).

Based on its specifications, the Accutrend® lactate meter is ideally suited to point of sample

analysis in the labour ward and the nursery; however, prior to using this meter for both

research and clinical management, it requires evaluation for accuracy and precision in

neonatal blood across the gestational ages from 24 to 42 weeks with a range of haematocrit

concentrations.

2.1.1 Aim

To determine the accuracy and precision of the commercially available Accutrend® handheld

reflectance photometer in measuring whole blood lactate levels in umbilical cord blood and

neonatal blood micro samples across the gestational age range of 26 to 42 weeks.

75

2.2 Methods

All experimental procedures were approved by the Wentworth Area Health Service Scientific

Advisory Panel and Ethics Committee, New South Wales, Australia. Informed maternal

consent was obtained prior to the collection of cord and neonatal blood.

2.2.1 Measuring principle

2.2.1.1 Accutrend® meter

The Accutrend® meter is a handheld reflectance photometer (Figure 2.1) based on dry

chemistry and requires 15-50µL of blood for analysis. Whole blood is applied to a multilayer

test strip and a thin glass fibre carpet separates red blood cells from whole blood (Figure 2.2).

Plasma is drawn into a plasma reservoir via capillary force to reach a reaction layer where

enzymatic determination of lactate commences using lactate dehydrogenase (Figure 2.3A). A

colour reaction is measured using an optical system through a transparent film, with the lactate

concentration being calculated from the intensity of reflected light (Figure 2.4). The

Accutrend® meter is accurate in adult blood for lactate concentration ranging from 0.7-

26mmol/L (Gambke & Muller, 1993).

2.2.1.2 Reference method

The reference method for this study was the Radiometer ABL 625® blood gas machine

(Radiometer Medical Pty Ltd, Copenhagen, Denmark) which measures lactate with an

amperometric electrode and an enzyme membrane utilising lactate oxidase. A two-step

reaction occurs and the current generated is proportional to the lactate concentration in the

sample (Figure 2.3B). The ABL625® lactate electrode has been validated over a range of 0 to

30mmol/L (Buch-Rasmussen, 1990; Cobbaert et al., 1999; Sinn et al., 2001).

2.2.2 Assessment of precision

The precision of a clinical test is the degree of refinement with which an operation is

performed or a measurement stated (Gennaro et al., 1984). To assess the precision of the

Accutrend® handheld lactate meter, repeated analysis of neonatal plasma samples (n=20) at

six known lactate concentrations were performed on both the Accutrend® meter and the

reference method by a single operator. Plasma was harvested from a single neonate during an

exchange transfusion for septic shock. The baseline serum lactate concentration was

76

Figure 2.1 Accutrend® handheld lactate meter

Figure 2.2 Design of lactate test strip for Accutrend® handheld lactate meter

Glass Fibre Reaction Layer Detection Film

Side View

Bottom View

Calibration Strip

77

A

Lactate + Mediator form1 → Pyruvate + Mediator reduced

Mediatorreduced + 2,18-Phosphomolybdate → Molybdenum Blue + Mediatorform 2

B

Lactate + O2 + H2O → Pyruvate + H2O2

H2O2 → 2H+ + O2 + 2e-

Figure 2.3 Reaction principles used for lactate measurement by the Accutrend® and

ABL625®. Figure A is the reaction principle of the Accutrend® handheld lactate meter using

lactate dehydrogenase for step one of the reaction. Figure B is the measuring principle of the

comparison method, the Radiometer ABL625® blood gas machine. Step one of this reaction

uses lactate oxidase and the generated current from step 2 is proportional to the lactate

concentration in the sample.

78

Figure 2.4 Analysis technique for Accutrend® handheld lactate meter. Plasma is drawn into a

plasma reservoir by capillary force to reach a reaction layer where the enzymatic determination

of lactate commences using lactate dehydrogenase. A colour reaction is measured using an

optical system with the lactate concentration being calculated from the intensity of reflected

light.

LED(660nm)

Detector

Plasma Reservoir

5°

40°

79

5.3mmol/L. D-lactate was added to produce higher concentrations of 9.6mmol/L,

15.4mmol/L and 19.5mmol/L and dilution of the baseline serum with saline produced lower

concentrations of 2.0mmol/L and 1.0mmol/L. The coefficients of variation were calculated

for both techniques at each of the six concentrations assessed.

2.2.3 Assessment of accuracy

The accuracy of a test is the degree of conformity of a measure to a standard or true value

(Gennaro et al., 1984). To assess the accuracy of Accutrend® handheld lactate meter, neonatal

plasma and blood, and umbilical arterial blood were tested simultaneously on the Accutrend®

handheld lactate meter and the Radiometer ABL625 blood gas machine (reference method).

2.2.3.1 Neonatal plasma samples

The initial studies of accuracy were performed on the 120 neonatal plasma samples used for

the analysis of precision (see Section 2.2.2).

2.2.3.2 Umbilical cord blood samples

Umbilical artery blood was collected at the Nepean Hospital Delivery Suite from 160

consecutive deliveries with gestational ages between 26 and 43 weeks. Each umbilical cord

was clamped at delivery, before the first breath by the neonate, and arterial samples were

collected in heparinised syringes which were capped and kept on ice until analysis could be

performed simultaneously using the Accutrend® meter and the reference method in the

neonatal intensive care unit. Clinical data from the mother and baby were recorded together

with the full blood-gas analysis performed on the Radiometer ABL 625®.

2.2.3.3 Neonatal arterial blood samples

Point-of-sample testing is where blood samples are analysed at the same location as their

collection. To test the Accutrend® meter as a true point-of-sample test, 110 neonatal arterial

blood samples were collected from umbilical artery catheters from babies in the Nepean

Hospital Neonatal Intensive Care Unit (gestational age from 26-41 weeks) and immediately

analysed simultaneously with both techniques.

2.2.3.4 Effect of temperature on accuracy

To assess the effect of temperature on the accuracy of the Accutrend® meter, we undertook

repeated measurement of neonatal blood at temperatures from 0.5°C to 37°C performed at

5°C increments on both the Accutrend® meter and the reference method. Blood was

80

collected from an exchange transfusion and samples were stored in capped heparinised

syringes (Radiometer Medical Pty Ltd, Nunawading, Victoria, Australia). The samples were

maintained at the appropriate temperature in a water bath while repeated measurements were

made on each machine. To confirm the discrepancy that became apparent in chilled samples,

blood at two known lactate concentrations (4.8mmol/L and 9.4mmol/L) was held in a water

bath at 0.5°C and measurement was repeated 20 times using both methods.

2.2.4 Statistical methods

Coefficients of variation (CV), defined as the standard deviation divided by the mean and

expressed as a percent, were calculated for the repeated sampling at known concentrations.

Accuracy between the two techniques was assessed using Passing Bablok (PB) regression

analysis (Passing & Bablok, 1983; Bablok et al., 1988) and modified Altman-Bland plots (Bland

& Altman, 1986; Pollock et al., 1992; NCCLS, 1995).

Passing Bablok regression is a technique widely used in clinical chemistry for formally testing

for agreement between two clinical methods. This regression technique is useful when

imprecision occurs in both variables and unlike other techniques, the imprecision does not

need to be normally distributed. Passing Bablok regression does not assume the variance to

be constant over the range of measurement and the regression line is not excessively biased by

extreme values (Passing & Bablok, 1983).

Altman-Bland plots illustrate graphically the agreement (or disagreement) between two clinical

methods. Two methods can be compared for both bias and agreement, and the relationship

between imprecision and concentration is presented graphically. Various options are available

for comparing the methods, depending on the best estimation of the true values. In Altman-

Bland plots, the difference between the methods is plotted against the mean of the methods.

The mean of the methods is treated as the best estimation of the true values (Bland & Altman,

1986). In the NCCLS EP9-A modification of Altman-Bland plots, the difference between

methods is plotted against the reference method or gold standard which is known to have

greater precision and therefore is the best estimate of the true value (NCCLS, 1995). In this

study NCCLS EP9-A plots have been used.

Descriptive data regarding the cord samples used in the analysis are presented as median, inter

quartile range (IQR) and range. The relationships between haematocrit, operator, and delay in

analysis and the accuracy of the Accutrend® meter were investigated using regression

modelling. Changes in accuracy (test method – reference method) at different temperatures

81

were analysed using analysis of variance (ANOVA) and paired t tests for the repeated sampling

at 4.8 and 9.4mmol/L. Statistical analyses were carried out using the Statistical Package for

Social Sciences (SPSS) for Windows, version 11.0 (SPSS Inc., Chicago, IL, USA) and Analyse-

It for Excel, version 1.63 (Leeds, England, UK).

82

2.3 Results

2.3.1 Analysis of precision

The precision of plasma lactate analysis for the Accutrend® handheld lactate meter and the

reference method are summarised in Tables 2.1 and 2.2. The coefficients of variation were

between 1.23% and 5.53% over the clinically significant lactate range of 2.2 to 19.3mmol/L

for the Accutrend® handheld lactate meter and between 1.93% and 6.74% for the reference

method. The Accutrend® meter had a higher coefficient of variation for lactate

concentrations of 1.1mmol/L; however, the range was only 1.0 to 1.2mmol/L.

Table 2.1 Precision of the Accutrend® Reflectance Photometer

Test Sample No. 1 2 3 4 5 6

Replicates (n) 20 20 20 20 20 20

Mean 1.01 2.23 6.16 9.50 16.39 19.05

SD* 0.079 0.091 0.11 0.35 0.91 0.23

CV† (%) 7.32 4.09 1.78 3.64 5.53 1.23

Range 1.0-1.2 2.1-2.4 6.0-6.4 8.8-10.0 15.3-18.3 18.5-19.7

*SD = standard deviation; † CV = coefficient of variation

Table 2.2 Precision of the Radiometer ABL 625® blood-gas machine (reference method)

Test Sample No. 1 2 3 4 5 6

Replicates (n) 20 20 20 20 20 20

Mean 1.05 2.04 5.48 9.57 15.39 19.49

SD* 0.079 .091 0.11 0.49 1.04 0.78

CV† (%) 4.89 2.93 1.93 5.12 6.74 4.03

Range 1.0-1.1 2.0-2.2 5.4-5.7 8.9-10.4 14.3-17.0 18.3-21.2

*SD = standard deviation; † CV = coefficient of variation

83

2.3.2 Analysis of accuracy using neonatal plasma samples

The accuracy of the Accutrend® lactate meter for measuring plasma lactate was assessed by

comparing the results of simultaneous analysis with the Accutrend® meter and the reference

method. This assessment was performed over the clinically significant range of lactate

concentrations of 1-20mmol/L. The Passing Bablok regression equation for this comparison

was y = 0.200 + 1.000x with 95% confidence intervals for slope of 0.989 to 1.018 and for y

intercept of 0.080 to 0.222 indicating a high degree of accuracy. Both the Passing Bablok

regression and the Altman-Bland plots demonstrate a constant bias, with the Accutrend®

meter overestimating the lactate concentration as measured by the reference method by

0.228mmol/L (Figure 2.5).

2.3.3 Analysis of accuracy using umbilical cord blood samples

The characteristics of umbilical arterial samples from the 160 consecutive deliveries used for

this analysis are summarised in Table 2.3. Gestational ages ranged from 26 to 43 weeks and

haematocrits ranged from 41.5 to 63.4%.

Table 2.3 Characteristics of 160 cord samples utilised in accuracy analysis.

Median Inter Quartile Range (IQR) Range

Gestation (weeks) 40 39 – 40 26 - 43

Haematocrit (%) 51.3 48.8 – 55.1 41.5 – 63.4

Haemoglobin (g/L) 16.8 15.9 – 18.1 13.3 – 20.8

pH 7.28 7.23 – 7.34 7.00 – 7.48

Base Excess -4.2 -7.0 to -2.2 -14.4 to - 0.50

PO2 (mmHg) 18.5 14.1 – 25.3 3.5 – 48.7

PCO2 (mmHg) 49.3 41.2 – 59.0 27.2 – 91.6

Lactate (mmol/L) 3.7 2.7 – 5.1 1.3 – 9.8

Analysis delay* (min) 11 5 - 25 2 - 132

*time between collection and analysis

83

Figure 2.5. Accuracy of the Accutrend® lactate meter for measurement of lactate in neonatal

plasma. Figure A is an X-Y plot of lactate levels obtained from the Accutrend® lactate

meter and the reference method for 120 neonatal plasma samples, showing the Passing

Bablok regression line (y = 0.200 + 1.000x) with 95% confidence intervals, slope 0.989 to

1.018 and intercept 0.080 to 0.222. The identity line, which demonstrates the relatinship if

the two test methods were to report identical lactate levels, is plotted in blue in Figure A.

Figures B, C and D are Altman-Bland plots displaying the difference in each pair of analyses

against the reference method in this study. Figure B is a scatterplot of the difference between

the two methods plotted against the reference method. Figure C is a frequency distribution

of the data presented in Figure B. Figure D is a scatterplot of the difference between the two

methods as a percentage of the reference lactate concentration plotted against the reference

method.

84

Identity lineY = X

y = x + 0.2

0

5

10

15

20

25

0 5 10 15 20 25Radiometer ABL625 Lactate (mmol/L)

Accu

tren

d La

ctat

e (m

mol

/L)

Zero bias

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

0 5 10 15 20 25

Radiometer ABL625 Lactate (mmol/L)

Diff

eren

ce b

etw

een

met

hods

(m

mol

/L)

0 20 40 60 80

-2.0

-1.4

-0.7

-0.1

0.6

1.2

1.9

Diff

eren

ce b

etw

een

met

hods

(m

mol

/L)

Frequency (%)

Zero bias

-10

-5

0

5

10

15

20

25

0 5 10 15 20 25

Radiometer ABL625 Lactate (mmol/L)

Diff

eren

ce b

etw

een

met

hods

(%)

A

D

B C

85

The accuracy of the Accutrend® handheld lactate meter for measurement of lactate in

umbilical artery blood from 160 consecutive deliveries is presented in Figure 2.6. The

Passing Bablok regression line was very similar to that obtained with plasma; however, the

95% confidence intervals were slightly wider (y = 0.008 + 0.923x with 95% confidence

intervals for slope of 0.857 to 1.000 and intercept of -0.300 to 0.214). The Altman-Bland

plots suggest that the Accutrend® meter has a bias of -0.380mmol/L to underestimate the

reference method lactate concentration when analysing umbilical arterial blood. This bias

applied consistently across the test range of 1.3mmol/L to 9.8mmol/L.

The accuracy of the Accutrend® meter was not affected by gestational age (p=0.237) or

operator (p=0.399). Similarly, haematocrit from 41.5 to 62% did not affect the accuracy of

the meter (r=0.09, p= 0.220; Figure 2.7). Four of the 160 cord samples had haematocrits

higher than 62%, and for these samples, the Accutrend® meter tended to underestimate the

lactate level by 1.3 to 1.8mmol/L. A weak correlation was identified between haemoglobin

concentration in the blood sample and the accuracy of the meter (r=0.199, 95% CI 0.044 to

0.346, p=0.012). For each increase in haemoglobin of 3g/L, the Accutrend® meter

underestimates lactate level by 0.2mmol/L. A weak correlation was also identified between

time delays in analysis and accuracy of the meter (r= 0.185, 95% CI 0.027 to 0.335, p=0.019),

with greater delays in analysis being associated with a small reduction in the accuracy.

2.3.4 Analysis of accuracy using neonatal arterial samples

The analysis of umbilical arterial blood samples suggested that the Accutrend® lactate meter

was less accurate with blood samples than with plasma samples. The reduced accuracy may

have been due to the presence of red cells or the delay in analysis of the cord samples. The

reference method blood gas machine was located in the neonatal intensive care unit that was

one floor above the delivery suite, and the Accutrend® meter was located in delivery suite.

Samples were kept in foam cups filled with crushed ice whilst awaiting analysis.

To further investigate these issues, 110 neonatal arterial blood samples were collected from

umbilical artery catheters of babies (gestational age from 26 to 41 weeks) in the nursery and

were immediately analysed simultaneously with both techniques (point-of-care analysis). The

accuracy of the Accutrend® meter measuring lactate from these neonatal blood samples is

presented in Figure 2.8. The Passing Bablok regression equation was y=0.011 + 0.916x with

85

Figure 2.6. Accuracy of the Accutrend® lactate meter for measurement of lactate in umbilical

arterial blood. Figure A is an X-Y plot of lactate levels obtained from the Accutrend® lactate

meter and the reference method for 160 umbilical arterial samples, showing the Passing

Bablok regression line (y = 0.008 + 0.923x) with 95% confidence intervals, slope 0.857 to

1.000 and intercept -0.300 to 0.214. The identity line, which demonstrates the relationship if

the two test methods were to report identical lactate levels, is plotted in blue in Figure A.

Figures B, C and D are Altman-Bland plots displaying the difference in each pair of analyses

against the reference method. Figure B is a scatterplot of the difference between the two

methods plotted against the reference method. Figure C is a frequency distribution of the

data presented in Figure B. Figure D is a scatterplot of the difference between the two

methods as a percentage of the reference lactate concentration plotted against the reference

method.

86

Identity lineY = X

y = 0.9231x + 0.0077

0

2

4

6

8

10

0 2 4 6 8 10Radiometer ABL625 Lactate (mmol/L)

Accu

tren

d La

ctat

e (m

mol

/L)

Zero bias

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

0 2 4 6 8 10

Radiometer ABL625 Lactate (mmol/L)

Diff

eren

ce b

etw

een

met

hods

(m

mol

/L)

0 20 40 60

-2.5

-2.1

-1.6

-1.2

-0.8

-0.3

0.1

0.6

Diff

eren

ce b

etw

een

met

hods

(m

mol

/L)

Frequency (%)

Zero bias

-60

-50

-40

-30

-20

-10

0

10

20

30

0 2 4 6 8 10

Radiometer ABL625 Lactate (mmol/L)

Diff

eren

ce b

etw

een

met

hods

(%)

B C

A

D

87

40 45 50 55 60-1

0

1

2

3

Haematocrit (%)

∆La

ctat

e (m

mol

/L)

Figure 2.7 The effect of haematocrit on accuracy of the Accutrend lactate meter. No

relationship was identified between haematocrit and the accuracy of the meter for

haematocrits between 41.5 and 62%. The Pearson correlation coefficient r = 0.09 (95% CI

-0.06 - 0.25, p=0.219).

87

Figure 2.8 Accuracy of the Accutrend® lactate meter for measurement of lactate in neonatal

arterial blood. Figure A is an X-Y plot of lactate levels obtained from the Accutrend® lactate

meter and the reference method for 110 umbilical arterial samples collected in the nursery

showing the Passing Bablok regression line (y = 0.011 + 0.916x) with 95% confidence

intervals, slope 0.889 to 0.949 and intercept -0.146 to 0.111. Figures B, C and D are Altman-

Bland plots displaying the difference in each pair of analyses against the reference method.

Figure B is a scatterplot of the difference between the two methods plotted against the

reference method. Figure C is a frequency distribution of the data presented in Figure B.

Figure D is a scatterplot of the difference between the two methods as a percentage of the

reference lactate concentration plotted against the reference method.

88

y = 0.9157x + 0.0107

0

2

4

6

8

10

12

14

16

18

20

0 5 10 15 20Radiometer ABL625 Lactate (mmol/L)

Acc

utre

nd L

acta

te (m

mol

/L)

Zero bias

-3

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

0 5 10 15 20 25

Radiometer ABL625 Lactate (mmol/L)

Diff

eren

ce b

etw

een

met

hods

(m

mol

/L)

0 20 40 60

-3.0

-2.4

-1.9

-1.3

-0.7

-0.1

0.4

Diffe

renc

e be

twee

n m

etho

ds

(mm

ol/L

)

Frequency (%)

Zero bias

-60

-40

-20

0

20

40

0 5 10 15 20 25

Radiometer ABL625 Lactate (mmol/L)

Diff

eren

ce b

etw

een

met

hods

(%)

B C

A

D

89

95% confidence intervals for the slope of 0.889 to 0.949 and for the intercept from -0.146 to

0.111. The confidence intervals were narrower than those found for the analysis of umbilical

arterial samples. A fixed bias was identified of -0.367mmol/L with lactate concentrations

reported by the Accutrend® meter being consistently lower that those reported by the

reference method.

2.3.5 Effect of temperature on accuracy

The accuracy in the Accutrend® meter improved when it was used for point-of-care

assessment rather than delayed assessment. The absence of any interaction with haematocrit

suggests that the inaccuracies that occurred in the cord blood analyses (Section 2.3.3) were

related to the delays in analysis that occurred (median delay 11 minutes, inter quartile range

(IQR) 5-25, range 2-132 minutes). During these delays, the samples were kept in foam cups

filled with crushed ice.

To assess the effect of temperature on accuracy, blood with a known lactate concentration

(4.8mmol/L) was assessed simultaneously on both meters, 10 times each, over a temperature

range of 37°C to 0.5°C. The differences between each pair of results are plotted in Figure

2.9A. Over the duration of the repeated analysis (160 minutes), the lactate concentration (as

assessed by the reference methodology) slowly increased from 4.84±0.02 (mean±SEM)

during the 37°C sampling to 5.26±0.02 during the 0.5°C sampling (p<0.001). Over the range

of temperature studies, there was a significant change in the accuracy of the Accutrend®

meter (test method minus reference method; p<0.001). When the sample temperature was

30°C or 37°C, the Accutrend® meter significantly overestimated the lactate concentration

(p<0.004). Over the temperature range 15°C to 25°C inclusive, both methods were in

agreement (p=0.07 to 0.371). When the temperature was decreased to 10°C or below, the

Accutrend® meter significantly underestimated the lactate concentration (p<0.016).

To complete the assessment of the effect of temperature, blood with known lactate

concentrations of 4.8mmol/L and 9.4mmol/L was stored in heparinised syringes in ice slurry

with a temperature of 0.5°C and 20 repeated measurements were performed on both the

Accutrend® meter and the reference method (Figure 2.9B). The chilled samples had

significantly lower lactate levels with the Accutrend® meter than the reference method at

both 4.8mmol/L (mean difference in lactate 1.10mmol/L, p<0.001) and 9.4mmol/L (mean

difference in lactate 2.12mmol/L, p<0.001). The variance in the Accutrend® meter was

significantly greater at 0.5°C than that of the reference method (p<0.001 for both

concentrations).

90

0.5 5 10 15 20 25 30 37-2

-1

0

1

2

* * *

* *

Temperature (°C)

Diff

eren

ce in

Lac

tate

†

(mm

ol/l)

AB

L625

Acc

utre

nd

AB

L625

Acc

utre

nd

0.0

2.5

5.0

7.5

10.04.8 mmol/L 9.4 mmol/L

*

*

Lact

ate

(mm

ol/L

)

A

B

Figure 2.9 The effect of temperature on the accuracy of the Accutrend® lactate meter.

Figure A demonstrates the differences in lactate concentration measured by the Accutrend®

lactate meter and the reference method as a box and whisker plot over the 8 temperatures

assessed († test method minus reference method). Figure B shows the 20 paired analyses at

4.8mmol/L and 9.4mmol/L for both test methods. The asterisk represent p-values <0.05

for repeated measures ANOVA in Figure A and paired t tests for Figure B.

91

2.4 Discussion

The Accutrend® handheld lactate meter measures lactate accurately and precisely in neonatal

plasma and neonatal blood micro samples, providing the sample is analysed at room

temperature. It is significantly cheaper than the reference method (AUS$900 vs.

AUS$80,000) is portable, and requires minimal maintenance; it is ideally suited to point-of-

care analysis of fetal and neonatal blood in both the delivery suite and in the neonatal

intensive care unit.

2.4.1 Reference method precision and accuracy

The reference method used in this study, the Radiometer ABL625® has been found to be

accurate to within ±0.8mmol/L in 95% of samples (Sinn et al., 2001) and it has published

coefficients of variation (CV) that range from 12.2% at 0.9mmol/L, 2.6% at 4.3mmol/L and

2.3% at 12.8mmol/L (Radiometer, 1997). Similar CV’s were found for the Radiometer

ABL625® in this study, however it performed less well at lactate concentrations

approximating 10mmol/L where the CV was found to be 5.1%.

2.4.2 Accutrend® handheld lactate meter precision and accuracy

The Accutrend® meter’s precision compares favourably with the more expensive reference

method and is well within the guidelines defined by the Royal College of Pathologists of

Australia (RCPA) and the Australasian Association of Clinical Biochemists (AACB). The

Altman-Bland difference plots in my study show that 95% of the Accutrend® lactate results

were within ±0.9mmol/L when blood was analysed at the point-of-care. This finding

compares very favourably with the AUS$80,000 Radiometer ABL 625® which when

compared to another gold standard “reference” methods was found to be accurate to within

±0.8mmol/L in 95% of samples (Sinn et al., 2001). The RCPA-AACB quality assurance

program quotes acceptable limits for performance for lactate assessment by any method as

±1.0mmol/L if less than 10mmol/L and 10% if greater than 10mmol/L (RCPA-AACB,

2002).

The Accutrend® handheld lactate meter is significantly more precise than other handheld

lactate meters previously reported which have coefficients of variation ranging from 5.5% to

45.9% over the same range of lactate concentrations used in this study (Nordstrom, 1995). A

new handheld lactate meter (Lactate Pro™) has become commercially available in Australia

92

in 2003. This meter has been compared with the Accutrend® meter (Nordstrom et al.,

1998a) and has coefficients of variation similar to those obtained in this study, however it

analyses samples as small as 5µL. In comparison, the Accutrend® meter is most accurate

with samples of 15-50µL and when presented with smaller samples it provides a result which

significantly underestimates the true lactate levels (Nordstrom et al., 1998a).

Under ideal circumstances, whole blood should be used to assess the precision of the two

lactate testing methods. In this study precision was assessed using plasma samples as the

additives required to inhibit lactate production can damage the electrode used in the

reference technique. Both in this study and the study by Sinn et al. (2001), lactate

concentrations increase when samples are stored at room temperature without additives to

inhibit lactate production. In my study during the temperature analysis, the lactate

concentration increased from 4.84mmol/L to 5.26mmol/L over 160 minutes in partially

chilled samples. Sinn et al. (2001) found that lactate concentrations increased at

0.012mmol/min at room temperature and were stable for 20 minutes with samples kept in

ice slurry. After 20 minutes, lactate levels of samples kept in ice slurry also increased (Sinn et

al., 2001).

2.4.3 Temperature and lactate measurement

The Accutrend® lactate meter provided accurate lactate measurement in neonatal plasma

and neonatal blood providing the sample was analysed at room temperature and the

haematocrit was less than 62%. Similar accuracy was reported by Nordstrom et al. when they

compared the Accutrend® meter to another test-strip based lactate meter (Nordstrom et al.,

1998a).

The reduced accuracy identified in the cord blood study when compared to the neonatal

blood point-of-care study is probably secondary to the effect of temperature as all samples in

the cord study were kept in ice for between 2 to 132 minutes before analysis. Ice was

required because the reference method was in the neonatal intensive care unit, not labour

ward or the operating theatre. The Radiometer ABL 625® blood-gas machine is not subject

to inaccuracies associated with temperature because all samples are warmed within the

machine to 37°C prior to analysis.

The issues of sample temperature is critical for the assessment of both fetal acid-base status

and lactate because deliveries often occur in operating theatres rather than delivery suites, or

the events that are occurring at delivery preclude immediate analysis of cord blood. In both

93

of these situations, samples are typically maintained on ice to minimise changes in gas

tensions, pH and lactate. If one is to use the Accutrend® lactate meter both clinically or for

research, it is optimal that the samples are warmed to at least 10°C prior to analysis if they

have been kept on ice for greater than 5 minutes. This can be easily achieved by warming the

sample in one’s hands for one minute prior to analysis.

There are a number of reasons why cooling of blood may impair the accuracy of the

Accutrend® meter. Cooling of blood increases its viscosity which may interfere with the

process of plasma separation in the lactate test strip. Cooling may also slow diffusion into

the reaction layer of the test strip or slow the enzymatic process that occurs in the reaction

layer of the strip. These possibilities are supported by evidence that the meter tends to

overestimate the lactate concentration if the sample is analysed at temperatures greater than

35°C where the elevated temperature may accelerate the analysis processes in the multilayer

strip.

My study has demonstrated that the ideal utilisation of the Accutrend® lactate meter is at the

point of sample. Such utilisation avoids the effects of temperature and clotting and it affords

optimal performance of the meter. This is supported by the 95% confidence intervals of the

Passing Bablok regression lines assessing the accuracy of the Accutrend® lactate meter. The

confidence intervals were narrower when blood samples were analysed at the point of

sampling rather than when the analysis was delayed. Wider confidence intervals imply greater

inaccuracy for each measurement.

A number of different methods have been used in clinical chemistry to compare a new

method with an established one prior to changing to a new analysis technique (Pollock et al.,

1992). These methods include simple least squares regression analysis which assumes that

the reference method (x variable) has no error, Deming’s regression method which allows for

variance in both variables (Payne, 1985) and graphical difference plots (Bland & Altman,

1986). In this study, in addition to the difference plots, the Passing Bablok regression

method has been utilised as it is a more reliable method of comparability than simple linear

regression or Deming’s regression. Passing Bablok regression is similar to Deming’s

regression where variance is present in both variables. Unlike Deming’s regression, Passing

Bablok regression assumes constant variance over the range of measurements and the

regression line is not excessively biased by extreme values (Passing & Bablok, 1983; Bablok et

al., 1988).

94

2.4.4 Haematocrit and lactate measurement

There are a number of factors which alter fetal haematocrit during pregnancy. First, with

advancing gestational age, fetal red cell count increases whereas white cell count and platelet

count do not change (Forrestier, 1987). This results in the fetal haematocrit increasing with

advancing gestational age from 34% (95%CI 31-37%) at 20 weeks gestation to 41% (95%CI

38-44%) at 40 weeks (Weiner et al., 1991). Second, growth restricted fetuses, which are at

increased risk of adverse events during labour (Low et al., 1976; Nieto et al., 1994; Dashe et al.,

2000; Bukowski et al., 2003), may experience chronic hypoxaemia in utero and they often

develop haemoconcentration with a resulting increase in their haematocrit (Soothil et al.,

1987; Economides & Nicolaides, 1990). Third, in the course of normal labour, haemo-

concentration has been described in fetal scalp blood (Jacobsen & Rooth, 1971). This effect

is probably due to loss of water from the fetal scalp with the average haemoglobin

concentration measured in scalp blood being 24g/L (Jacobsen & Rooth, 1971). Owing to

plasma reduction, fetal blood volume is reduced by 14% during labour when compared with

pre-labour values (Brace, 1992). It is unknown if this effect is more pronounced in

prolonged labour although it is doubtful whether the increase would be great enough to

significantly affect lactate measurement. Therefore, as seen in my study, there can be a wide

range of haematocrits in blood samples collected around delivery and when considering a

device to measure fetal lactate levels during the perinatal period, it is important to confirm

that accuracy is not affected by changes in haematocrit.

In this study the accuracy of the Accutrend® lactate meter was maintained across a

haematocrit range of 41.5 to 62% in umbilical cord blood. In contrast, other authors have

previously reported a deleterious effect of haematocrit on the accuracy of different handheld

lactate meters to the one assessed in this study (Shimojo et al., 1993; Nordstrom et al., 1998a).

They found that the meters they assessed reported lower lactate concentrations than the

reference method when haematocrit values were greater than 50%. This would limit the

clinical value of these meters as a significant proportion of blood samples in fetal and

neonatal blood (63% in this series) have haematocrit values greater than 50%. As with the

effect of temperature, it has been suggested the reduced accuracy at high haematocrits may

be due to slower diffusion into the reaction layer of the test strip (Nordstrom, 1995). The

Accutrend® lactate meter may be less affected by haematocrit than other meters due to the

different structure of the test strip that has a thin glass fibre carpet to separate plasma from

red cells. In this study, the accuracy of the Accutrend® lactate meter did not change until

95

the haematocrit was greater than 62%, which occurred in 2.5% of the consecutive arterial

samples collected from umbilical cords at delivery across a gestational age from 26 to 43

weeks.

2.4.5 Lactate measurement bias

In this study the Accutrend® meter consistently underestimated lactate levels by

0.367mmol/L relative to the reference method. This bias however was constant across the

range of lactate concentrations assessed. Similar bias has been identified for other methods

of lactate measurement (Nordstrom et al., 1998a; Sinn et al., 2001). It is imperative that if

using the Accutrend® meter clinically, appropriate reference ranges are utilised, preferably

determined with the same testing methodology. The published normal reference ranges for

umbilical arterial lactate after labour vary widely depending on the type of sample analysed,

the method used to measure each sample and the number of cases in each series (range 14 to

3932; Table 2.4). Many of the published reference ranges are for whole blood, which is

expected to yield values 40-50% lower than those obtained by analysing plasma from the

same specimen.

Table 2.4 Normal reference range for umbilical arterial lactate after labour

Lactate (mmol/L)

(mean)

+2SD (95th centile)

Sample type Reference

2.8 N.A. N.A. Rooth & Nilsson, 1964

2.7 5.1 N.A. Jouppila & Hollmen, 1976

3.3 6.1 Haemolysed blood Low et al., 1974

3.5 7.3 Haemolysed blood Smith et al., 1983

2.6 5.0 Haemolysed blood Suidan & Young, 1984

2.9 5.4 Haemolysed blood Ruth & Raivio, 1988

3.7 6.1 Whole blood Nordstrom et al. 1994

1.9 3.8 Whole blood Westgren et al., 1995

96

2.5 Conclusion

The Accutrend® handheld lactate meter provides an affordable, commercially available

means of measuring plasma lactate levels in only 60 seconds at the point-of-care. It requires

only 15µl of blood and is significantly cheaper than other assay methods.

The precision of the Accutrend® handheld lactate meter compares favourably with more

expensive machines and is well within the guidelines prescribed by the Royal College of

Pathologists of Australasia and the Australasian Association of Clinical Biochemists. I have

shown that the accuracy of the Accutrend® handheld lactate is acceptable for clinical and

research use providing that samples are analysed at room temperature and that the

haematocrit is less than 62%.

The Accutrend® handheld lactate meter is well suited to assess lactic acidaemia in small

samples to help quantify hypoxic stress in the perinatal period.

97

C h a p t e r 3

THE RELATIONSHIP BETWEEN NEUROPATHOLOGIC

DAMAGE AND BIOCHEMICAL AND PHYSIOLOGIC MARKERS

OF ASPHYXIA INDUCED BY GRADED CORD OCCLUSION IN

CHRONICALLY CANNULATED OVINE FETUSES

In seeking absolute truth we aim at the unattainable,

and must be content with finding broken portions.

Sir William Osler (1889)

3.1 Introduction

Until recently, hypoxic-ischaemic events in the perinatal period were assumed to be the main

cause of early neonatal encephalopathy and seizures. They were also thought to be the major

cause of brain damage in infants with a clinical diagnosis of cerebral palsy. These theories

were challenged during the last decade by epidemiological studies showing that there was

often an absence of evidence of severe intrapartum asphyxia in infants with neonatal

encephalopathy, and conversely, many infants who do have signs of fetal distress and birth

asphyxia do not develop neurologic sequelae (Blair & Stanley, 1988; Nelson & Leviton, 1991;

Yudkin et al., 1994; Badawi et al., 1998a; Nelson & Grether, 1998). These observations,

together with epidemiological evidence that there is a higher incidence of maternal illness,

antenatal complications, and adverse social factors in infants with neonatal encephalopathy,

seizures or both, have led to the view that the main causes of such conditions occur before

birth (Badawi et al., 1998b). These epidemiological studies suggest that in about 90% of cases

intrapartum hypoxia could not be the cause of cerebral palsy and that in the remaining 10%,

intrapartum signs compatible with damaging hypoxia may have had antenatal or intrapartum

origins (Nelson, 1988; Yudkin et al., 1995).

A number of recent studies however, provide support to the original hypothesis that events

in the immediate perinatal period are the most important in neonatal brain injury in the term

infant. Hagberg et al. studied the prevalence, cause and timing of cerebral palsy in a

population based study in Sweden and reported a perinatal origin for cerebral palsy in 36% of

98

affected term infants (Hagberg et al., 2001). Similarly in a study of 351 term infants with

neonatal encephalopathy, seizures or both, magnetic resonance imaging found evidence of an

acute insult without established injury or atrophy in 80% of infants with neonatal

encephalopathy and 69% of infants with seizures (Cowan et al., 2003). When these infants

were followed up for a minimum of 18 months, 19% (66/351) and 24% (85/351) had

evidence of cerebral palsy (Cowan et al., 2003).

The goal of intrapartum fetal surveillance is to reduce the incidence of fetal asphyxial

exposure and to prevent moderate and severe asphyxia. The three techniques which are

currently widely used for this surveillance are electronic fetal heart rate monitoring (EFM),

scalp blood gas analysis and umbilical cord blood gas analysis after delivery.

Continuous EFM has been adopted widely since its development in 1968. Thirteen

randomized clinical trials have addressed the efficacy and safety of EFM comparing EFM to

intermittent auscultation (Haverkamp et al., 1976; Renou et al., 1976; Kelso et al., 1978;

Haverkamp et al., 1979; Langendoerfer et al., 1980; Wood et al., 1981; MacDonald et al., 1985;

Neldam et al., 1986; Luthy et al., 1987; Shy et al., 1990; Vintzileos et al., 1993; Herbst &

Ingemarsson, 1994; Mahomed et al., 1994). Meta-analysis of nine of these trials found a 50%

reduction in neonatal seizures in the continuous EFM group but no significant difference in

rate of admission to neonatal intensive care units, cerebral palsy or perinatal death (Thacker et

al., 2001). Late decelerations are the only fetal heart rate (FHR) pattern that has been noted

to correlate with neonatal neurologic examination in the first week of life (Ellison et al., 1991).

Fetal blood sampling for pH measurement was introduced into clinical practice by Saling in

1962 as the first invasive fetal monitoring tool (Saling, 1962). Although it has been widely

advocated to improve the specificity of FHR monitoring (Clark & Paul, 1985; Low et al.,

2001; Low et al., 2002) its utilisation varies widely, partly due to the restricted availability of

the equipment required to perform the test and also because of its invasive nature.

Based on current published literature, umbilical cord acid-base assessment at delivery

provides the most sensitive reflection of birth asphyxia (MacLennan, 1999). There are

several reasons for performing routine cord blood acid-base balance at delivery: a quality

assessment of obstetric practice; a diagnostic test for hypoxia in cases with depressed

newborns; and for the prediction of neonatal complications and long-term outcome.

Although the presence of metabolic acidaemia does not define the timing of onset of birth

asphyxia, its absence excludes an acute intrapartum hypoxic event (MacLennan, 1999).

99

Two percent of fetuses experience an asphyxial exposure during labour and delivery as

demonstrated by an umbilical artery base deficit greater than 12 mmol/L (Low et al., 1994).

Most of these events are mild with little short or long-term significance (Gilstrap et al., 1989;

Low, 1997). More severe asphyxial exposures are associated with a range of outcomes.

Pelowski and Finer (1992), in a review of five studies of term newborns, reported a good

outcome for newborns with minor encephalopathy. The risk of death and severe handicap

in moderate encephalopathy was 5.6% and 20% respectively, and in severe neonatal

encephalopathy, 60% and 70% respectively. A recent study of 276 term newborns with

moderate or severe encephalopathy reported similar findings with 25% and 62% respectively

having poor outcomes (Table 3.1, adapted from Dixon et al., 2002).

Table 3.1 Outcome after term neonatal encephalopathy

Outcome Moderate Neonatal Encephalopathy

Severe Neonatal Encephalopathy

Died 3% 31%

Cerebral palsy then died 3% 2%

Cerebral palsy surviving 7% 14%

Abnormal cognitive impairment or developmental delay 12% 15%

Normal 75% 38%

Adapted from Dixon et al, 2002

Our principal biochemical markers of perinatal asphyxia, fetal scalp and umbilical cord acid-

base status, are limited because they reflect fetal metabolic adjustments that are not directly

related to fetal damage except if values are extreme (Gilstrap et al., 1989; Goldaber et al., 1991;

Winkler et al., 1991). Even a pH of less than 7.0 is poorly predictive of neurological deficit

because a severely acidotic infant will usually have a normal outcome (Winkler et al., 1991;

Socol et al., 1994; Yudkin et al., 1994). The presence of acidosis, reflected by a fall in pH, may

reflect the physiological switch to anaerobic metabolism which is a healthy response to

hypoxia, and paradoxically it may be the fetuses that are unable to mount this response and

100

therefore have “normal” umbilical gases in the face of hypoxia who suffer long-term

neurological sequelae (Dennis et al., 1989; Westgren et al., 1999).

There is growing evidence that lactate, the end product of anaerobic metabolism of glucose,

has a role in the assessment of the fetus in labour and the neonatal period. In human studies,

intrapartum scalp lactate assessment has been shown to be a more sensitive diagnostic tool

than the measurement of scalp pH values for predicting either an Apgar score <4 at five

minutes of age or moderate to severe neonatal encephalopathy (Kruger et al., 1999).

Similarly, urinary lactate-creatinine ratios at six hours of age assessed by proton nuclear

magnetic resonance spectroscopy (MRS) have both high sensitivity (94%) and specificity

(100%) for predicting hypoxic-ischaemic encephalopathy in a group of babies with perinatal

asphyxia (Huang et al., 1999).

3.1.1 Aim

The aim of this study was to use an ovine model of repeated umbilical cord occlusion (UCO)

to evaluate the relationship between fetal lactate responses and fetal brain injury. To address

this aim I have utilized an UCO protocol designed to mimic events which may occur in

labour to compare the relationship between:

1. fetal acid-base status and neuropathologic damage,

2. fetal lactate and neuropathologic damage,

3. fetal heart rate patterns (FHR) and neuropathologic damage, and

4. fetal arterial pressure and neuropathologic damage.

3.1.2 Hypothesis

Fetal lactate is a better predictor of neuropathologic injury after repeated UCO than fetal

heart rate patterns and fetal acid-base status.

101

3.2 Methods

All experimental procedures were approved by the Wentworth Area Health Service Scientific

Advisory Panel and Ethics Committee and the Westmead Hospital Animal Ethics

Committee, New South Wales, Australia. Merino ewes with singleton pregnancies were

transported to the research facility at 120 days of pregnancy to allow for acclimatization.

The timeline for the protocol utilised in this study is presented in Figure 3.1

3.2.1 Surgical procedure

Sixteen ewes underwent surgery at 126±3 days (mean±SEM) of pregnancy. Anaesthesia was

induced with Thiopentone 4mg/kg in 15mL saline solution (Abbott Australasia Pty Ltd,

Cronulla, NSW, Australia), ewes were intubated, and anaesthesia was maintained with 0.75-

1.5% Isoflurane in 100% oxygen delivered by positive pressure ventilation with the ventilator

settings selected to maintain maternal blood gases in the normal range.

The uterus was exposed through a paramedian abdominal incision and a hysterotomy was

performed to deliver the fetal head, neck and upper thorax through a small incision with care

to minimize amniotic fluid leakage. Vascular catheters were inserted into the left fetal carotid

artery and jugular vein and an amniotic catheter was inserted. Fetal electrocardiograph

(ECG) leads were sutured to the fetal chest and a vascular snare was sutured around the

umbilical cord to enable intermittent complete cord occlusion. A circular three crystal

ultrasound cuff (10MHz Triton Technology Inc., USA) was loosely applied around the

umbilical vessels. After replacement of the fetus into the uterus, Benacillin (1mL procaine

penicillin, 150mg/mL; benzathine penicillin, 150mg/mL; procaine hydrochloride, 20mg/mL;

Troy Laboratories, Smithfield, Australia) was injected into the amniotic fluid. The uterus was

closed and catheters were tunnelled subcutaneously to the ewe’s flank. The maternal

abdomen was closed and 3-mL intramuscular Benacillin was given to the ewe at completion

of the surgery. After surgery, sheep were held in individual pens with access to food and

water. Two hours after surgery, fetal blood samples (3.5mL) were collected for baseline

measurement of arterial gas tensions and pH (Radiometer ABL 300, Radiometer Pty Ltd,

Copenhagen, Denmark), lactate (Accutrend® Lactate, Roche Diagnostics GmbH,

Mannheim, Germany), glucose, cortisol, haemoglobin, and catecholamine concentrations.

Two days after surgery, fetal blood was also collected (0.3mL) for measurement of arterial gas

tensions, pH and lactate measurement. After all fetal blood samples were withdrawn sterile

heparinised saline was infused to replace the volume removed for analysis.

102

Figure 3.1 Experimental timeline for the entire study period and below, the cord occlusion protocol

during the experimental phase of the study. Red bars during the experimental phase indicate both the

timing and the duration of the cord occlusions. The first nine occlusions were for 0.5 minutes

duration every 3 minutes; the second nine occlusions were 1.0 minutes duration every 3 minutes; and

the third nine occlusions were for 1.5 minutes duration every 3 minutes. This figure demonstrates the

progressive shortening of the interval between occlusions.

ExperimentalPhase

0 1 2 3 4 5 6 7

SurgeryCord

Occlusion Delivery

SampleSamplesSample Sample

Days

0 3 6 9 12 15 18 21 24 27

Sample SampleSample

27 30 33 36 39 42 45 48 51 54

Sample SampleSample

54 57 60 63 66 69 72 75 78 81

Sample SampleSample

Minutes

Study Period

103

3.2.2 Cord occlusion protocol

After a four day recovery period (130±3 days gestation) the experimental phase of the study

commenced. After a 60-minute control period, baseline fetal blood samples (3.5mL) were

collected for measurement of gas tensions, pH, lactate, glucose, cortisol and catecholamine

concentrations. Twelve fetuses were then randomly allocated to the cord occlusion protocol

which was designed to mimic events that may occur in labour (Figure 3.1). The remaining 4

fetuses served as controls with blood samples and other measurements taken as scheduled

for the experimental group but without UCO. All studies were performed with the ewe

unrestrained with free access to food and water.

For those fetuses that underwent the UCO protocol the umbilical cord was completely

occluded every 3 minutes for a total of 81 minutes. The first nine occlusions each lasted for

0.5 minutes every three minutes; the second series of nine occlusions were for 1.0 minutes

every 3 minutes; and the final series of nine occlusions were for 1.5 minutes every three

minutes. Fetal carotid arterial blood samples (0.3mL) were collected for measurement of gas

tensions, pH, and lactate levels every nine minutes over the 81 minute experimental phase.

Additional arterial samples (1.0mL) were collected at 27 and 54 minutes to measure fetal

glucose. Arterial blood samples (3.5mL) were collected from all fetuses at the end of the

occlusion phase for the assessment of fetal catecholamine, cortisol and glucose

concentrations. The cord occlusion protocol was terminated early, prior to 81 minutes, if any

of the following were observed; arterial pH levels less than 6.9; base excess levels less than

-20mmol/L; persistent fetal hypotension; or prolonged fetal bradycardia.

3.2.3 Fetal heart rate assessment

Fetal ECG was recorded continuously onto a computer at 1 KHz after amplification (Bio

amp, PowerLab 8e, ADInstruments, Sydney, Australia) during the cord occlusion protocol.

Fetal heart rate (FHR) was calculated by measuring every RR interval on the fetal ECG.

Over the 60 second period immediately before each cord occlusion, three measures of FHR

variation were assessed; mean minute range (MMR), short term variation (STV) and long

term variability (LTV). Both MMR and STV were calculated as previously described (Dawes

et al., 1991; Street et al., 1991). Mean minute range was calculated as the difference between

the maximum and minimum R-R intervals for each minute and short term variation was

calculated as the average of successive epochal R-R interval differences (Epoch = 1/16th-

minute {3.75 second}). Long term FHR variation was defined as the difference in the

maximum and minimum fetal heart rate during each 60 second block (NICH, 1997; RCOG,

104

2001). I also visually confirmed the FHR variability, a method which has previously been

shown to relate directly to mathematical quantitation (Knopf et al., 1991). FHR traces were

visually assessed to classify the fetal heart rate pattern during each of the 27 cord occlusions.

All measurements were made blind to the treatment groups. The baseline FHR during the

experimental phase was the mean FHR over the one minute prior to the cord occlusion.

Accelerations were defined as an abrupt increase in FHR above the baseline of more than 15

beats and lasting longer than 15 seconds (NICH, 1997). Variable decelerations were defined

as an abrupt decrease in FHR below the baseline, which was ≥15 beats/min, lasting ≥15

seconds, followed by an abrupt return to baseline (NICH, 1997; RCOG, 2001). Severe

variable decelerations were defined as variable decelerations with depth greater than 60 beats

per min below the baseline FHR. Given that there were no uterine contractions in this study,

the traditional definition of a late deceleration was not appropriate in this context. A late

component to the decelerations was therefore defined as FHR decelerations that had not

returned to baseline within 15 seconds of the cessation of cord occlusion. Overshoot

instability was defined as a rebound increase in FHR immediately after the release of the cord

occlusion. These were at least 15 beats above the baseline FHR and their duration was >15

seconds, after which the FHR returned to the baseline or another deceleration. Increased

FHR variability was defined as long term FHR variability which was greater than 25 beats per

min.

3.2.4 Fetal arterial pressure assessment

Fetal arterial pressure was recorded continuously onto a computer at 1 KHz after

amplification (BP transducer MLT380, bridge amplifier ML110, PowerLab 8e,

ADInstruments, Sydney, Australia) during the cord occlusion protocol. Fetal arterial pressure

was normalised for amniotic fluid pressure and responses to repeated UCO were assessed

during and between cord occlusions. For assessment during the resting phase between

occlusions, systolic, diastolic and mean arterial pressure were measured for each heart beat

and averaged for the 1 minute preceding each cord occlusion.

3.2.5 Umbilical blood flow

The ultrasound cuff around the umbilical vessels was linked to a vascular spectrum analyser

(Model SP 25 by Vasculab, Meda Sonics, USA) which enabled continuous monitoring of the

umbilical artery Doppler waveforms to confirm normal umbilical blood flow between

occlusions and complete obstruction of umbilical blood flow during cord occlusion.

105

3.2.6 Biochemical assessment

Plasma samples for glucose, cortisol and catecholamine measurement were recovered after

centrifugation at 1500g for 10 minutes at 4°C and were stored at -80°C until analysed.

Plasma adrenaline, noradrenaline and dopamine were measured by electrochemical detection

after high performance liquid chromatography (HPLC) fractionation. Fetal cortisol

concentrations were measured by solid-phase chemi-luminescent immunoassay (DPC

Immulite System; Diagnostic Products Corporation, Los Angeles, USA). The intra-assay

coefficient of variation was 6.8% and inter-assay 9.9%. This immunoassay has been validated

in human, dog, cat, horse, sheep and dolphin plasma (Singh et al., 1997; Ma, 1999; Beetson,

2001). Catecholamine measurements were performed by the renal research laboratory of

King George V Hospital, Camperdown, Sydney. The HPLC technique utilised was the same

as that previously described and validated in fetal sheep plasma by others (Murotsuki et al.,

1997b).

3.2.7 Delivery and neuropathologic examination

Seventy two hours after the experimental phase of the study, fetal blood samples (3.5mL)

were collected for measurement of arterial gas tensions, pH and lactate, glucose, cortisol,

catecholamine and haemoglobin concentrations. Ewes were then anaesthetised (as above)

and the fetus was delivered and injected with a lethal dose of phenobarbitone. Brains were

perfused with ice-cold heparinized saline solution through newly inserted catheters in both

carotid arteries, and then with 10% buffered formalin. After perfusion the fetal brain was

carefully removed, sectioned and blocked at multiple levels including representative blocks of

all cortical regions of the cerebrum, basal ganglia, thalamus, brain stem and cerebellum.

Blocks were then embedded in paraffin and 6 to 8µm sections were cut and stained with

hematoxylin and eosin, Luxol fast blue-cresyl violet, Bielschowsky, glial fibrillary acidic

protein, myelin basic protein, HAM-56, or vimentin. The severity of histologic changes was

classified into five grades for 9 areas on each side of the fetal brain using a classification

system previously described (Ikeda et al., 1998a) and summarized in Table 3.2.

All neuroanatomical assessments were performed by a neuropathologist (Dr Thomas Ng)

who was blinded to experimental group and biochemical data. The grade attributed to each

hemisphere was the highest grade found in any of the nine areas studied.

106

Table 3.2 Grading of histopathologic changes in fetal brains Grade 0 Intact white matter

Intact gray matter, including that of hippocampus

Grade 1 Random or focal vacuolisation and loss of myelin or increase in cellularity (macrophages, astrocytes) primarily in white matter at its junction with cerebral cortex

Intact gray matter, including that of hippocampus

Grade 2 More widespread involvement of subcortical white matter with increase in cellularity with loss of myelin sheaths

Intact gray matter, including that of hippocampus

Grade 3 Focal necrosis of subependymal white matter with or without cystic change (random)

Random or focal scattered ischaemic neuronal change in cerebral cortex (minimal)

Grade 4 More extensive change within white matter, including that of basal ganglia, thalamus, and other gray matter structures (increase in cellularity, loss of myelin)

Well-defined infarcts of gray and white matter (random)

Grade 5 Extensive necrosis involving both gray and white matter with marked macrophagic and astroglial reaction

(Adapted from Ikeda et al., 1998a)

3.2.8 Statistical methods

Descriptive data are presented as mean and standard error of the mean (SEM) unless

otherwise stated. Univariate analyses of categorical data were performed with Fisher’s exact

test and continuous data were analysed with t-tests or the non-parametric equivalent as

appropriate.

Correlations between the three measures of FHR variability were assessed using regression

modelling considering experimental group and time as fixed effects, and individual animals as

random effects repeated over time. Correlations between the grade of neuronal injury and

the duration of FHR patterns, arterial pressure responses and biochemical markers were

performed using Spearman rank correlation test.

Changes in fetal biochemical variables and cardiovascular parameters were analysed using

multivariate regression modelling with occlusion and time modelled as a fixed effects and

individual animals as random effects repeated over time.

107

Receiver operator curve (ROC) analysis has been used, both in this chapter and subsequent

chapters, to assess the ability of biochemical measures to predict categorical outcomes. The

ROC is a technique to visualise and quantify the effectiveness of a procedure by graphing the

true positive rate (sensitivity) against the false positive rate (1-specificity). The ROC curve

depicts the relationship between sensitivity and specificity for all allowable cut-off points

(Harding et al., 1995). The incorrectly classified non-diseased cases (1-specificity) are plotted

as the x-coordinate and sensitivity as the y-coordinate. The point (0,1) is called an ideal point

of the test as it reflects perfect classification while the line through points (0,0) and (1,1)

reflects a random classification rule (no discrimination line) (Chmura Kraemer, 1992). A

diagnostic test no better than chance alone would have a ROC curve positioned close to the

no discrimination line. Curves above and to the left of this random line reflect better

discrimination ability that is usually assess by the area under the curve. Generally areas under

the curve of 90-100% are considered excellent; those between 80-90% are considered good,

while those of approximately 50% indicate chance alone.

In this chapter, sensitivity and specificity were calculated for the ability of specific FHR

patterns to predict each pattern of fetal arterial pressure response during UCO. Receiver

operator curve analysis was used to assess the ability of biochemical measures to predict fetal

arterial pressure responses during UCO (Hanley & McNeil, 1983; Beck & Shultz, 1986;

DeLong et al., 1988). Optimum sensitivity and specificity were chosen from the ROC tables

to maximize the total of these two measures of diagnostic accuracy (Hanley & McNeil, 1983).

Accuracy is reported for biochemical prediction of arterial pressure responses during UCO

and is defined as the proportion of correctly classified cases (true positives + true negatives)

using the biochemical cut-off point selected to maximise the sensitivity and specificity (Zweig

& Campbell, 1993).

SAS statistical software was used for data analysis (Statistical Analysis Systems Institute Inc.

1999, Version 8.2., Cary, NC). P-values <0.05 were considered statistically significant, and p-

values <0.10 were considered marginally significant. A conservative significance level of 0.10

is often used while performing tests on small sample sizes which are often associated with

low power to detect differences. This allows the identification of potential effects which may

be worthwhile investigating in larger studies (Woodward, 1999).

108

3.3 Results

For consistency in the description of results, the total seven day period of this study will be

referred to as the “study period” and the two hours of intensive fetal assessment, when 12 of

the fetuses underwent UCO and four fetuses were studied as controls, will be referred to as

the “experimental phase” (Figure 3.1).

3.3.1 Biochemical responses to repeated cord occlusion

3.3.1.1 Fetal arterial blood gas and lactate responses during umbilical cord

occlusion

In controls there were no significant changes in fetal arterial pH, PCO2, or base excess over

the 7 days of the experimental protocol (all p>0.778; Figure 3.2). Fetal PO2 did not change

during the 81 minutes of the experimental phase in controls (p=0.369); however there was a

significant fall in PO2 in controls three days after the experiment (experimental phase PO2

22.9±0.7mmHg; day 7, 11.7±2.8mmHg, p=0.026).

Fetal lactate levels were stable during the experimental protocol in controls (1.5±0.1mmol/L,

p=0.108); however, at the time of catheter implantation (day 0) and at delivery (day 7), lactate

concentrations were slightly increased compared with the levels during the experimental

phase (day 0, 4.4±0.4mmol/L; day 7, 3.4±0.9mmol/L; p<0.001 for both). A significant fall

in fetal glucose concentrations occurred between the time of catheter implantation and

commencement of the experimental phase (day 0, 1.72±0.30mmol/L; day 4, 0.95±0.13

mmol/L, p<0.001) but after this time there were no significant changes in fetal glucose

concentrations (p=0.875).

Nine of the twelve fetuses that underwent the repetitive UCO protocol completed the 81

minutes of intermittent occlusion and these have arbitrarily been called regular responders.

In three of the twelve fetuses that underwent UCO protocol was terminated at 54 (n=1) or

57 minutes (n=2) respectively when physiologic variables crossed the predetermined end-

points for terminating the protocol. These fetuses have arbitrarily been classified as fast

responders. Despite ceasing the cord occlusion protocol early, none of the fast responders

was alive at the time of delivery 72 hours later whereas all but one of the regular responders

and all controls were alive at delivery.

109

-4-26.8

7.0

7.2

7.4

7.6

Control

0 27 54 81 100

Regular(p=0.004)

Fast(p<0.001)

3

pH

Days Minutes Days

Arte

rial p

H

-4-20

10

20

30

40

0 27 54 81 100

Control

Regular(p=0.142)

Fast(p<0.001)

3

PO2

Days Minutes Days

Arte

rial P

O2

(mm

Hg)

-4-20

25

50

75

100

125

0 27 54 81 100

Control

Regular(p=0.171)

Fast(p<0.001)

3

PCO2

Days Minutes Days

Arte

rial P

CO

2 (m

mH

g)

-4-2-30

-20

-10

0

10

0 27 54 81 100

Control

Regular(p<0.001)

Fast(p<0.001)

3

BE

Days Minutes DaysAr

teria

l Bas

e Ex

cess

-4-20

5

10

15

20

0 27 54 81 100

Control

Regular(p<0.001)

Fast(p=0.016)

3

Lact

Days Minutes Days

Arte

rial L

acta

te(m

mol

/L)

-4-20

1

2

3

0 27 54 81 100

Control

Regular(p=0.003)

Fast(p=0.002)

3

Gluc

Days Minutes Days

Arte

rial G

luco

se(m

mol

/L)

Figure 3.2 Fetal arterial gas tensions, pH, base excess, lactate and glucose (mean± SEM) for

control (open circles, n=4), regular responders (closed circles, n=9) and fast responders

(closed triangles, dashed line, n=3) over the seven day study period. P values were calculated

using repeated measures ANOVA comparing the occlusion groups to the control group.

Significant changes were identified for changes in fetal arterial gas tensions, pH, base excess,

lactate, and glucose for the fast responders and for pH, base excess, lactate and glucose for

the regular responders.

110

Regular responders had significant falls in fetal pH (7.40±0.01 to 7.08±0.03, p<0.001) and

base excess (-0.59±1.35 to -14.73±1.30, p<0.001) and significant increases in fetal lactate

(1.48±0.25 to 10.40±1.47mmol/L, p<0.001) and glucose (0.76±0.09 to 1.34±0.29mmol/L,

p=0.003) during repeated cord occlusion relative to controls (Figure 3.2). Fetal PO2 fell and

PCO2 increased during the experimental phase but these differences did not reach statistical

significance (p=0.142 and p=0.171 respectively).

Fast responders had significant changes in pH, PO2, PCO2, base excess, lactate and glucose

when compared to controls (all p<0.016). When compared with regular responders, fast

responders had significantly greater falls in fetal pH (p<0.001) and base excess (p=0.011).

Lactate, glucose and PCO2 were higher in fast responders than regular responders but these

differences did not reach statistical significance (all p>0.200).

3.3.1.2 Fetal glucose, cortisol and catecholamine responses during umbilical cord

occlusion

There were no significant changes in fetal cortisol, noradrenaline, adrenaline or dopamine

levels in the control fetuses over the seven day study period (all p >0.442; Figure 3.3).

Regular responders had significantly higher cortisol concentrations than controls (Figure 3.3)

after repeated cord occlusion (regular responders cortisol 24.79±4.472ng/mL; vs. control

12.75±1.58ng/mL, p=0.001). Fetal noradrenaline, adrenaline and dopamine concentrations

at the end of the experimental phase were higher than in controls but these differences did

not reach statistical significance (noradrenaline regular responders 114.21±31.39ng/L vs.

control 1.80±0.90ng/L, p=0.055; adrenaline regular responders 88.39±42.71ng/L vs. control

1.10±0.35ng/L, p=0.082; dopamine regular responders 1.03±0.92ng/L vs. 0.00±0ng/L,

p=0.170). In contrast, fast responders had significantly higher cortisol, noradrenaline,

adrenaline and dopamine concentrations than controls by 57 minutes when their UCO

protocol was discontinued (all p<0.030).

3.3.1.3 Fetal haemoglobin response to experimental protocol

During the seven days of the experimental protocol, a total of 19.0mL of fetal blood was

sampled for analysis. Each time the fetus was sampled, an equal volume of normal saline was

replaced. No significant difference was found between baseline fetal haemoglobin (day 0)

and haemoglobin collected prior to delivery (day 7; mean difference 0.27 g/L, 95%CI -0.57

to 1.10, p=0.495). Paired haemoglobin concentrations are presented in Figure 3.4.

111

-4-20

10

20

30

40

Control

0 27 54 81 100

Regular(p=0.001)

3

Fast(p<0.001)

Days Minutes Days

Cortisol

Cor

tisol

(ng/

mL)

-4-20

100

200

300

Control

0 27 54 81 100

Regular(p=0.055)

Fast(p=0.033)

3

Days Minutes Days

Noradrenaline

Nor

adre

nalin

e (n

g/L)

-4-20

25

50

75

100

125

Fast(p<0.001)

0 27 54 81 100

Regular(p=0.082)

Control

3

Days Minutes Days

Adrenaline

Adre

nalin

e (n

g/L)

-4-20

1

2

3

4

5

Fast(p<0.001)

0 27 54 81 100

Regular(p=0.170)

Control

3

Days Minutes Days

Dopamine

Dop

amin

e (n

g/L)

Figure 3.3 Fetal cortisol, noradrenaline, adrenaline and dopamine concentrations (mean±

SEM) for control (open circles, n=4), regular responders (closed circles, n=9) and fast

responders (closed triangles, dashed line, n=3) over the seven day study period. P values

were calculated using repeated measures ANOVA comparing the occlusion groups to the

control group. Regular responders had significantly increased cortisol concentrations

(p=0.001) and non-significant increases in noradrenaline (p=0.055), adrenaline (p=0.082) and

dopamine (p=0.170) after 81 minutes of cord occlusion. The fast responders had significant

increases in cortisol and all catecholamines after 54 minutes of cord occlusion (p<0.033).

112

0.0

2.5

5.0

7.5

10.0

12.5

Day 0 Day 7

p=0.495

Hae

mog

lobi

n (g

/L)

Figure 3.4 Fetal haemoglobin responses to experimental protocol. Paired haemoglobin

concentrations are presented on day 0 of the experimental protocol (collected after

instrumentation) and day 7 of the experimental protocol (collected prior to delivery). Paired

analysis found no significant decrease in haemoglobin over the duration of the protocol.

113

3.3.1.4 Biochemical prediction of fetal response to umbilical cord occlusion

Arterial gas tensions, pH, cortisol and catecholamine concentrations collected over the four

days prior to the experimental phase were not different between regular and fast responders

(p=0.233 to p=0.960); however, there were differences in fetal lactate and glucose levels

between groups (Figure 3.5). Fast responders had significantly higher lactate concentrations

(fast responders lactate 6.81±1.19mmol/L vs. regular 2.29±0.82mmol/L, p=0.011), and

higher glucose concentrations (fast responder glucose 1.38±0.19mmol/L vs. regular

0.78±0.14mmol/L, group effect p=0.094) than regular responders over the four days prior to

the experimental phase.

3.3.2 Fetal heart rate responses to repeated cord occlusion

3.3.2.1 Baseline FHR response to repeated cord occlusion

Baseline FHR did not change over the duration of the experimental phase in controls

(182.5±9.7 beats per minute (bpm), p=0.816; Figure 3.6A). In the fetuses that underwent

UCO the baseline FHR varied during the occlusion protocol (p=0.055) and was significantly

different from controls (p=0.003). There was a significant decrease in baseline FHR after

commencement of UCO (p=0.043) followed by a progressive increase in baseline FHR to

levels higher than before the commencement of the UCO (Figure 3.6B). Unlike the

biochemical markers of fetal condition, there was no difference in the baseline FHR response

to UCO between regular and fast responders (p=0.943; Figure 3.6C)

3.3.2.2 Fetal heart rate variability response to repeated cord occlusion

The three measures of FHR variability did not change significantly during the experimental

phase in controls (MMR 46.67±1.67 msec, p=0.999; STV 3.19±1.17msec, p=0.999; LTV

21.67±2.19bpm, p=0.287; Figure 3.7). In fetuses that underwent UCO there was a

significant increase in all three measures of FHR variability during the experimental phase

when compared to controls (MMR, p=0.028; STV, p=0.050; and LTV p<0.001). Statistically

significant increases in MMR occurred after 42 minutes of UCO (p<0.05) and significant

increases in STV and LTV occurred after 45 minutes (p<0.05) and 36 minutes respectively

(p<0.05). Fetal heart rate variability as indicated by MMR and LTV appeared to decrease at

the end of the experimental phase in the fetuses that were subject to UCO with the

differences between occluded and control fetuses not reaching statistical significance.

114

Pre-occlusion Lactate

Control Regular Fast0.0

2.5

5.0

7.5

10.0 p=0.011

Lact

ate

(mm

ol/L

)

Pre-occlusion Glucose

Control Regular Fast0.0

0.5

1.0

1.5

2.0 p=0.064

Glu

cose

(mm

ol/L

)

Figure 3.5 Biochemical prediction of response to repeated cord occlusion. Baseline fetal

lactate and glucose concentrations for the controls (n=4), regular responders (n=9) and the

fast responders (n=3) are presented as the adjusted mean ± standard error for lactate and

glucose concentration over the 4 days prior to the experimental phase considering both

sheep and study group in repeated measures ANOVA model. Fast responders had

significantly higher lactate concentrations (p=0.011) and glucose concentrations (p=0.064)

prior to repeated cord occlusion protocol than regular responders.

115

100

125

150

175

200

225

Control

p=0.816FH

R (b

pm)

0 9 18 27 36 45 54 63 72 81100

125

150

175

200

225

Fast

Regular

p=0.943

Time (minutes)

FHR

(bpm

)

100

125

150

175

200

225

Occluded

Control

p=0.003

FHR

(bpm

)

A

B

C

Figure 3.6 Baseline fetal heart rate during the experimental phase. FHR (mean± SEM) over

the one minute prior to each cord occlusion for controls (open circles, n=4), occluded

fetuses (regular and fast responders combined; closed diamonds, n=12), regular responders

(closed circles, n=9), and fast responders (closed triangles, dashed line, n=3). There was no

change in the baseline FHR in the control fetuses over the duration of the experimental

phase (p=0.816) and in the fetuses that underwent UCO the baseline FHR was significantly

different from the controls (p=0.003). Baseline FHR was not significantly different between

fast and regular responders (p=0.943).

116

0 9 18 27 36 45 54 63 72 810

20406080

100120140160

D

Control

Time (minutes)

FHR

MM

R (m

sec)

Fast

Regular

0 9 18 27 36 45 54 63 72 810

20406080

100120140160

A

Control

Occluded

p=0.028

Time (minutes)

FHR

MM

R (m

sec)

0 9 18 27 36 45 54 63 72 810

3

6

9

12

15E

Fast

Control

Regular

Time (minutes)FH

R S

TV (m

sec)

0 9 18 27 36 45 54 63 72 810

3

6

9

12

15B

Control

Occluded

Time (minutes)

FHR

STV

(mse

c) p=0.050

0 9 18 27 36 45 54 63 72 810

10203040506070

F

Control

Regular

Fast

Time (Minutes)

FHR

LTV

(bpm

)

0 9 18 27 36 45 54 63 72 810

10203040506070

C p<0.001

Occluded

Control

Time (Minutes)

FHR

LTV

(bpm

)

Figure 3.7 Fetal heart rate variability during the experimental phase. Figures A, B and C

present data from control and occluded fetuses (both the regular and fast responders). P

values are for the comparison between groups, controlling for sheep and time. A significant

increase in MMR (p=0.028), STV (p=0.050) and LTV (p<0.001) occurred in the occluded

fetuses. Figures D, E and F present data from the three study groups. MMR, STV and LTV

were not significantly different between regular and fast responders (MMR, p=0.167; STV,

p=0.925; LTV, p=0.436).

117

Fetal heart rate variability was not different between fast and regular responders (MMR,

p=0.167; STV, p=0.925; and LTV, p=0.436; Figure 3.7D-F). After controlling for

experimental group, sheep and time, there was a significant increase in all three measures of

FHR variability (LTV vs. STV r=0.838, p<0.001; LTV vs. MMR r=0.858, p<0.001; MMR vs.

STV r=0.862, p<0.001).

3.3.2.3 Fetal heart rate patterns during repeated cord occlusion

No FHR accelerations or decelerations were present during the 81 minutes of the

experimental phase in controls (Figures 3.8 – 3.12). One control fetus had a single episode of

increased FHR variability after fetal blood sampling. No fetal heart rate information is

available for one regular responder due to difficulties collecting the fetal ECG. Due to the

small sample size, no formal statistical comparison was made between the FHR patterns of

the regular and fast responders during the experimental phase. Descriptive data relating to

these patterns is presented as median and range.

3.3.2.3.1 Variable fetal heart rate decelerations

Variable FHR decelerations occurred in all but two of the 275 umbilical cord occlusions for

which data is available. All fetuses that underwent UCO had severe variable decelerations

with their first UCO. During the experimental phase, seven of the 11 occluded fetuses had

intermittent variable decelerations which did not meet the criteria for severe variable

decelerations (decrease in FHR by >60 bpm during cord occlusion). The pattern of variable

decelerations is presented graphically in Figure 3.8 which demonstrates that there is no

obvious difference in the pattern of variable decelerations between regular and fast

responders.

3.3.2.3.2 Overshoot after variable fetal heart rate decelerations

Fetal heart rate overshoot began at a median of 30 minutes (range 9-57 minutes) into the

experimental phase in nine of the 11 fetuses subjected to UCO. The timing of FHR

overshoot for each fetus is presented graphically in Figure 3.9. No overshoot was present in

two fetuses who were both fast responders.

3.3.2.3.3 Late components to variable fetal heart rate decelerations

With their first UCO, all fetuses subjected to UCO had late components associated with their

severe variable decelerations (Figure 3.10). These continued throughout the experimental

phase in two of the three fast responders and one of the eight regular responders. In the

remainder of the fetuses subjected to UCO, late components to variable decelerations were

117

Figure 3.8 Variable fetal heart rate decelerations. The blue segments of the bar chart

represent no variable decelerations during the cord occlusion. The yellow segments

represent variable FHR decelerations and the orange segments represent the severe variable

FHR decelerations. No decelerations were present in control fetuses. All of the fetuses that

underwent UCO had severe variable decelerations in response to their first umbilical cord

occlusion.

118

9908

9918

9922

9925

9917

9913

9912

9911

9906

9920

9919

9905

0 9 18 27 36 45 54 63 72 81Time (minutes)

9915

9914

9921

9918

9908

9922

9925

9919

9906

9905

9920

9913

9917

9912

9911

9921

9914

9915

Con

trols

Reg

ular

Res

pond

ers

Fast

Res

pond

ers

Variable Decelerations

SheepNumber

118

118

Figure 3.9 Overshoot after fetal heart rate decelerations. The blue segments of the bar chart

represent no FHR overshoot after the decelerations during the cord occlusion and the purple

segments represent the presence of overshoot.

119

9908

9918

9922

9925

9917

9913

9912

9911

9906

9920

9919

9905

0 9 18 27 36 45 54 63 72 81Time (minutes)

9915

9914

9921

9918

9908

9922

9925

9919

9906

9905

9920

9913

9917

9912

9911

9921

9914

9915

Con

trols

Reg

ular

Res

pond

ers

Fast

Res

pond

ers

Overshoot

SheepNumber

119

119

Figure 3.10 Late components to fetal heart rate decelerations. The blue segments of the bar

chart represent no late component to the variable decelerations during the cord occlusion

and the red segments represent the presence of late components.

120

9908

9918

9922

9925

9917

9913

9912

9911

9906

9920

9919

9905

0 9 18 27 36 45 54 63 72 81Time (minutes)

9915

9914

9921

9918

9908

9922

9925

9919

9906

9905

9920

9913

9917

9912

9911

9921

9914

9915

Con

trols

Reg

ular

Res

pond

ers

Fast

Res

pond

ers

Late Components to Decelerations

SheepNumber

120

120

Figure 3.11 Increased FHR variability between cord occlusions. The blue segments of the bar

chart represent no increased FHR variability and the green segments represent the presence

of increased variability between the cord occlusions.

121

9908

9918

9922

9925

9917

9913

9912

9911

9906

9920

9919

9905

0 9 18 27 36 45 54 63 72 81Time (minutes)

9915

9914

9921

9918

9908

9922

9925

9919

9906

9905

9920

9913

9917

9912

9911

9921

9914

9915

Con

trols

Reg

ular

Res

pond

ers

Fast

Res

pond

ers

Increased Variability

SheepNumber

121

121

Figure 3.12 Complex variable fetal heart rate decelerations. The blue segments of the bar

chart represent no FHR deceleration. The yellow segments represent simple variable

decelerations and the orange segments represent simple severe variable decelerations. The

red segments represent variable decelerations with late components to their recovery and the

purple segments represent variable decelerations with overshoot. The dull cherry segments

of the bar chart represent variable decelerations with both late components to their recovery

and overshoot after the deceleration.

122

9908

9918

9922

9925

9917

9913

9912

9911

9906

9920

9919

9905

0 9 18 27 36 45 54 63 72 81Time (minutes)

9915

9914

9921

9918

9908

9922

9925

9919

9906

9905

9920

9913

9917

9912

9911

9921

9914

9915

Con

trols

Reg

ular

Res

pond

ers

Fast

Res

pond

ers

Complex Variable Decelerations

SheepNumber

122

123

not present in the latter half of the experimental phase (median of 25.5 min, range 6-57 min)

usually having been replaced by FHR overshoot. Comparing the figures illustrating late

components to variable FHR decelerations and overshoot after variable FHR decelerations

(Figures 3.9 and 3.10) it appears that in the majority of cases, fetuses have only one of these

patterns at each time point, with the combined pattern occurring in only 30 of the 275 (11%)

cord occlusions for which data are available for analysis.

3.3.2.3.4 Increased fetal heart rate variability between cord occlusions

Increased FHR variability was present in all fetuses that underwent UCO at some period of

the experimental phase (Figure 3.11). It was present intermittently during the first 36

minutes of cord occlusions after which it was present for the majority of the time (85%) in 10

of the 11 occluded fetuses. Regular and fast responders displayed similar patterns of

increased FHR variability.

3.3.2.3.5 Complex variable fetal heart rate decelerations

Figure 3.12 graphically presents the combinations of variable FHR decelerations with both

late components to the decelerations and FHR overshoot. In two of the three fast

responders fetal heart rate overshoot did not occur and late components to the variable FHR

decelerations were present with each UCO. In all of the regular responders, combined

patterns were present.

3.3.3 Fetal arterial pressure responses to repeated cord occlusion

3.3.3.1 Fetal arterial pressure between cord occlusions

There were no significant changes in systolic pressure (76.0±5.7mmHg, p=0.967) or diastolic

pressure (53.5±3.9mmHg, p=0.967) during the experimental phase in controls (Figure 3.13).

Systolic pressure increased significantly during the experimental phase in fetuses that

underwent UCO (regular + fast responders) compared to controls (p=0.049). The systolic

pressure progressively increased during the first 36 minutes of UCO and remained

significantly elevated between 39 and 51 minutes (p<0.05) after which it decreased again

towards baseline levels. Diastolic pressures were not significantly different between fetuses

subjected to UCO and controls (p=0.999).

124

0 9 18 27 36 45 54 63 72 8160708090

100110120

Occluded

Control

p=0.049

Time (minutes)

Syst

olic

Pre

ssur

e(m

mH

g)0 9 18 27 36 45 54 63 72 81

30405060708090

Control

Occluded

p=0.967

Time (minutes)

Dia

stol

ic P

ress

ure

(mm

Hg)

0 9 18 27 36 45 54 63 72 8160708090

100110120

Control

Regular

p=0.244

Time (minutes)

Syst

olic

Pre

ssur

e(m

mH

g)

0 9 18 27 36 45 54 63 72 8130405060708090

Control

Regular

p=0.653

Time (minutes)

Dia

stol

ic P

ress

ure

(mm

Hg)

0 9 18 27 36 45 54 63 72 8160708090

100110120

Fast

Control

p=0.024

Time (minutes)

Syst

olic

Pre

ssur

e(m

mH

g)

0 9 18 27 36 45 54 63 72 8130405060708090

Fast

Control

p=0.024

Time (minutes)

Dia

stol

ic P

ress

ure

(mm

Hg)

A

B

C

D

E

F

Figure 3.13 Fetal arterial pressure responses between repeated cord occlusion. Fetal systolic

pressures are presented in Figures A, B and C and diastolic pressures in D, E and F. The p

values on each graph represent the comparison of the occluded group or subgroup to the

control group. There were no significant changes in the systolic pressure (p=0.967) or the

diastolic pressure (p=0.999) in controls.

125

Systolic (p=0.244) and diastolic (p=0.159) pressures were not significantly different between

regular and fast responders; however, the magnitude of the increase in systolic and diastolic

pressure was greater in the fast than the regular responders with only fast responders having

significantly higher pressures than controls (p=0.024).

3.3.3.2 Fetal arterial pressure during cord occlusion

Fetal arterial pressure responses to UCO progressed through several distinct patterns during

the experimental phase. These have been arbitrarily classified as the “hypertensive response”,

the “decreasing hypertensive response”, the “biphasic systolic response” and the “biphasic

diastolic response”. Representative arterial pressure and heart rate recordings, illustrating

each of these different patterns of response, are shown in Figure 3.14.

Initially during UCO the fetal arterial pressure was elevated and this was accompanied by a

deceleration in fetal heart rate. This has been arbitrarily classified as the “hypertensive

response”. The first sign of fetal cardiovascular decompensation was when the fetus could

no longer maintain hypertension during the UCO. This has been arbitrarily classified as the

“decreasing hypertensive response” and is illustrated by the second column of panels in

Figure 3.14. As progressive cardiovascular deterioration continued, the fetus first developed

systolic hypotension during cord occlusion (biphasic systolic response) and finally systolic

and diastolic hypotension during occlusion as the pre-terminal event (biphasic diastolic

response). The fall in systolic and diastolic pressures during cord occlusion represents the

most severe form of deterioration of fetal wellbeing preceding prolonged fetal bradycardia,

hypotension and death.

All regular responders progressed through these four stages of cardiovascular decompression

albeit at different rates. The time of onset of the biphasic diastolic response phase varied

from 12 to 81 minutes into the protocol with a median time of 46.5 minutes. One of the fast

responders displayed the preterminal biphasic diastolic response during UCO for the entire

protocol and another developed this response 12 minutes into the protocol.

3.3.4 Neuropathologic findings after repeated cord occlusion

Fetal brains were examined from all four control fetuses and from eight of the nine regular

responders to repeated cord occlusion. Considerable variation was present on macroscopic

brain examination. Some brains appeared normal with good topographic preservation and

clear gray-white matter demarcation. In other cases the brain had extensive white and gray

matter damage with pronounced softening and cavitation.

125

Figure 3.14 Fetal arterial pressure responses during repeated cord occlusion. The four

distinct patterns of arterial pressure response during cord occlusion are presented in this

figure which is taken directly from the digital recording of fetus 9921. The top row of panels

shows the arterial pressure signal in red. The systolic and diastolic arterial pressures are

shown in the middle two rows, and the fetal heart rate patterns are in blue in the lowest row

of panels with the red bars representing the cord occlusion.

126

FHR

BP Signal

Systolic Pressure

Diastolic Pressure

Biphasic Diastolic

Response

Hypertensive Response

Decreasing Hypertensive

Response

Biphasic Systolic

Response

126

127

The grades of histopathologic injury in right and left sides of the brain are summarised in

Table 3.3. Injury was present in the left side of all fetal brains, corresponding with the side of

vascular catheterization. There was no correlation between damage on the right and left side

of the fetal brains (r=0.032, p=0.922). All quantification of neurologic damage relating to

biochemical markers has been performed on the right side of the fetal brain.

There was no histopathologic damage in the right side of the brain in any of the controls. In

the fetuses exposed to UCO, four were found to have mild histopathologic damage (grade 1)

and the other four had damage ranging from widespread white matter injury (grade 2)

through to areas of necrosis and infarction in the gray and white matter (grade 4 and 5).

In controls the cortical neurons had abundant cytoplasm and vesicular nuclei (Figure 3.15A).

In the brains of four of the fetuses exposed to UCO there were subtle changes, with

vacuolisation and loss of myelin or an increase in cellularity with macrophages and astrocytes.

This occurred primarily in the parietal and frontal subcortical white matter at its junction with

the cerebral cortex. Four of the eight fetuses exposed to repeated cord occlusion had more

substantial neuropathologic damage (Figure 3.15B and C) varying from more widespread

white matter injury through to well defined infarcts and extensive necrosis of gray and white

matter with marked macrophagic and astroglial reaction.

3.3.5 Relationship between markers of fetal well-being during repeated cord

occlusion and fetal neuropathologic damage

3.3.5.1 Correlation of biochemical markers with fetal neuropathologic damage

Fetal arterial gas tensions, pH, base excess and lactate, glucose, cortisol, and catecholamine

concentrations at three time points were compared to the grade of fetal brain injury.

Correlation coefficients for these comparisons are presented in Table 3.4. Fetal base excess

(r=-0.743, p=0.009) was the only biochemical measure prior to repeated UCO with a

significant correlation with the grade of fetal brain injury assessed 72 hours after UCO.

Fetuses with lower base excess levels prior to umbilical cord occlusion had higher grades of

fetal brain injury.

Of the twelve biochemical markers collected immediately after repeated UCO, significant

correlations were found between the grade of fetal brain injury and lactate (r=0.759,

p=0.007), noradrenaline (r=0.640, p=0.046) and arterial PO2 (r=-0.637, p=0.035). The only

biochemical marker 72 hours after UCO with a significant correlation with the grade of fetal

128

Table 3.3 Histopathologic damage in right side (no vascular ligation) and left side (vascular ligation) of fetal brain

Sheep Number

Overall Right Grade

Parietal Frontal White Occipital Frontal

Gray Thalamus Striatum

(basal ganglia)

Hippo-campus Pons Cerebellum

9918* 0 9908* 0 9922* 0 9925* 0 9901 1 + 9912 1 + 9911 1 + + 9917 1 + + + 9920 2 + ++ 9905 2 ++ ++ ++ + ++ 9919 4 ++++ ++++ ++++ ++++ ++++ ++++ ++++

Right side of fetal brain

9906 5 +++++ ++++ +++++ ++++ ++++ +++ ++++ ++++

9918* 5 +++++ ++++ +++++ ++++ 9908* 5 +++++ +++ +++ +++ ++++ ++ 9922* 4 ++++ ++++ ++++ ++++ 9925* 4 ++++ ++++ ++++ ++++ 9901 2 + ++ ++ 9912 4 ++++ ++ ++++ ++ 9911 1 + + + + 9917 1 + + + 9920 4 ++ ++++ ++++ 9905 2 ++ ++ ++ ++ ++ 9919 4 ++++ ++++ ++++ ++++ ++++ ++++ ++++

Left side of fetal brain

9906 5 +++++ +++++ ++++ +++++ ++++ ++++ +++++ ++++ ++ * Denotes Control Fetuses

128

128

Figure 3.15 Range of histopathologic damage to fetal brains. A, B and C are photo-

micrographs (haematoxylin-eosin stain) of ovine fetal brain under high power showing the

range of neuronal injury produced by the intermittent cord occlusion protocol utilized in this

study. Figure A is a section of cortex (100x magnification) showing neurons with abundant

cytoplasm and vesicular nuclei (grade 0 changes). Figure B is a section of the frontal cortex

(40x magnification) showing scattered ischaemic neuronal changes in the cerebral cortex.

The neurons are shrunken, triangular, and eosinophilic with dark nuclei (grade 3 changes).

Figure C is a section of the parietal lobe (40x magnification) showing extensive necrosis

involving both grey and white matter with marked macrophagic and astroglial response

(grade 5 changes).

129

A

B

C

130

Table 3.4 Spearman’s Rho for fetal biochemical markers and the grade of fetal brain injury in the right side of brain

pH

PO2

PCO

2

Base

Exc

ess

Lact

ate

Glu

cose

Corti

sol

Nor

-adr

enali

ne

Adr

enali

ne

Dop

amin

e

r -0.421 -0.249 -0.508 -0.743 -0.481 -0.291 -0.356 -0.334 0.072 0.216 Pre-Occlusion

p-value 0.197 0.460 0.111 0.009 0.134 0.386 0.313 0.380 0.854 0.681

r -0.432 -0.637 0.388 -0.478 0.759 0.560 0.234 0.640 0.611 0.390 Post-Occlusion

p-value 0.184 0.035 0.238 0.137 0.007 0.136 0.515 0.046 0.060 0.265

r -0.350 -0.541 0.363 -0.226 0.322 -0.847 -0.046 0.178 0.352 0.306 Delivery (3 days later)

p-value 0.322 0.160 0.302 0.530 0.364 0.008 0.907 0.647 0.353 0.390

130

131

brain injury was fetal glucose levels (r=-0.847, p=0.008) with the highest grades of injury

occurring in the fetuses with the lowest glucose concentrations.

3.3.5.2 Correlation of fetal heart rate patterns with fetal neuropathologic damage

The total duration of the UCO protocol in which specific FHR decelerative patterns were

present was calculated for each fetus and correlation coefficients for the comparison of these

durations with the grade of neuronal injury are presented in Table 3.5. The only two patterns

in which there was a significant correlation between the duration of the pattern and the grade

of fetal brain injury was increased FHR variability (r=0.671, p=0.039) and overshoot after

variable decelerations (r=0.705, p=0.027).

The correlation coefficients between the duration of compound FHR abnormalities (similar

to those assessed in the clinical environment) and the grade of fetal brain injury are presented

in Table 3.6. None of these coefficients reached statistical significance.

Table 3.5 Spearman’s Rho for the duration of specific FHR patterns and grade of fetal brain injury

Duration of Pattern (n=11) R p-value

Late component 0.470 0.166

Severe variable decelerations 0.566 0.096

Increased FHR variability 0.671 0.039

Overshoot 0.705 0.027

Table 3.6 Spearman’s Rho for the duration of compound FHR patterns and grade of fetal brain injury

Duration of Pattern (n=11) R p-value

Late component and overshoot 0.325 0.349

Increased variability and late component 0.336 0.349

Severe variable and late component 0.510 0.133

Severe variables and increased variability 0.568 0.088

Increased variability and overshoot 0.592 0.073

132

3.3.5.3 Correlation of fetal arterial pressure with fetal neuropathologic damage

3.3.5.3.1 Fetal arterial pressure between the umbilical cord occlusions

The mean fetal arterial pressures over the 10 minutes following the final cord occlusion of

the protocol were compared with the grade of neurologic injury. There was no significant

correlation between the grade of fetal brain injury and absolute arterial pressure or changes in

arterial pressure over the duration of the experimental phase (Table 3.7). Similarly, there was

no significant correlation between the grade of fetal brain injury and the percentage of the

experimental phase where the diastolic pressure was less than 30mmHg (r=0.142, p=0.802).

Table 3.7 Spearman’s Rho for arterial pressure at the completion of the cord occlusion protocol and grade of neuronal injury

(n=12) R p-value

Systolic pressure 0.104 0.840

Diastolic pressure 0.169 0.713

Change in systolic pressure 0.660 0.110

Change in systolic pressure 0.374 0.396

3.3.5.3.2 Fetal arterial pressure during the final three occlusions

Fetal arterial pressure during the final three occlusions in each fetus was assessed as an

estimate of the severity of arterial pressure decompensation during the experimental phase.

The following parameters were not significantly correlated with the grade of fetal brain injury

(p all >0.268):

- Mean decrease in systolic, diastolic and mean arterial pressure

- Maximum decrease in systolic, diastolic and mean arterial pressure

- Percentage of the UCO with hypotension

- Total duration of mean arterial pressure hypotension

133

3.3.5.3.3 Duration of fetal arterial pressure responses during umbilical cord

occlusion

The duration of each of the fetal arterial pressure responses illustrated in Figure 3.14 were

compared to the grade of fetal brain injury (Table 3.8). There was a strong correlation with

the duration of the biphasic diastolic response during UCO and the grade of fetal brain injury

(r=0.808, p=0.022) and an even stronger correlation for the duration of any arterial pressure

instability (total of the three patterns of cardiovascular instability) during UCO (r=0.955,

p=0.001). The earlier the fetus lost the ability to maintain fetal hypertension during the

UCO, the greater the degree of neuronal injury with damage only occurring after more than

39 minutes of cardiovascular instability. Duration rather than the severity of cardiovascular

instability during the UCO correlated with the grade of fetal brain injury after repeated UCO

in the sheep fetus.

Table 3.8 Spearman’s Rho for the duration of arterial pressure instability during cord occlusion and grade of neuronal injury

Duration of Pattern (n=12) R p-value

Biphasic Diastolic Response 0.808 0.022

Biphasic Systolic Response* 0.602 0.120

Any arterial pressure instability** 0.955 0.001

* Sum of biphasic systolic response and biphasic diastolic response as both patterns had biphasic systolic pressure patterns

** Sum of decreasing hypertensive response, biphasic systolic response and biphasic diastolic response phases

3.3.6 Prediction of fetal arterial pressure responses during cord occlusion

Given the strong correlation between fetal responses during UCO and neuronal injury it

would be clinically useful to predict the presence of these patterns from the two techniques

available for intrapartum fetal assessment; fetal heart rate patterns and fetal scalp blood

biochemical assessment.

134

3.3.6.1 Prediction of arterial pressure responses during umbilical cord occlusion

from fetal heart rate patterns

The sensitivity and specificity for fetal heart rate patterns to predict fetal arterial pressure

responses during cord occlusion are presented in Table 3.9. None of the FHR patterns had

both good sensitivity and good specificity for the prediction of fetal arterial pressure

responses during cord occlusion. The four patterns assessed had poor specificity for

detecting any arterial pressure instability and poor sensitivity for the biphasic diastolic

response during cord occlusion.

Table 3.9 Fetal heart rate prediction of arterial pressure responses during cord occlusion

Any Arterial Pressure Instability*

Biphasic Systolic Response**

Biphasic Diastolic Response FHR Pattern

Sensitivity Specificity Sensitivity Specificity Sensitivity Specificity

Severe Variable Decelerations 82% 66% 70% 84% 59% 89%

Late component to decelerations 72% 20% 61% 39% 52% 51%

Increased variability 89% 53% 74% 63% 63% 72%

Overshoot 98% 39% 83% 51% 66% 57%

* Sum of decreasing hypertensive response, biphasic systolic response and biphasic diastolic response phases ** Sum of biphasic systolic response and biphasic diastolic response as both patterns had biphasic systolic pressure patterns

3.3.6.2 Prediction of arterial pressure responses during umbilical cord occlusion

from fetal biochemical measures

Receiver operator curves (ROC) were used to assess the diagnostic utility of fetal biochemical

measures to assess fetal responses during cord occlusion. The area under the ROC and the

optimum sensitivity and specificity are summarised in Table 3.10.

For all three patterns of arterial pressure response during UCO, fetal lactate levels had

significantly greater area under the receiver operator curves than pH (any arterial pressure

instability, p=0.011; biphasic systolic response, p=0.037; biphasic diastolic response,

p=0.049; Figure 3.16). The area under the lactate ROC for any arterial pressure instability

was also significantly greater than the area under the PO2 and PCO2 receiver operator curves

(p<0.001) but was not significantly different from the area under the base excess curve

135

(p=0.730). Similarly, the areas under the lactate ROC for both the biphasic systolic response

and the biphasic diastolic response were significantly greater than those for base excess, PO2

and PCO2 (all p<0.031).

Table 3.10 Biochemical prediction of arterial pressure responses during cord occlusion

Any Arterial Pressure Instability*

Biphasic Systolic Response**

Biphasic Diastolic Response Measure

Area§ Sens† Spec‡ Area§ Sens† Spec‡ Area§ Sens† Spec‡

Lactate 0.931 86% 91% 0.950 86% 91% 0.924 86% 83%

pH 0.891 82% 95% 0.923 73% 95% 0.902 87% 84%

Base excess 0.925 85% 86% 0.919 68% 89% 0.887 77% 74%

PO2 0.781 87% 27% 0.780 88% 69% 0.769 91% 60%

PCO2 0.662 48% 89% 0.757 59% 91% 0.786 62% 88% * Sum of decreasing hypertensive response, biphasic systolic response and biphasic diastolic response phases ** Sum of biphasic systolic response and biphasic diastolic response as both patterns had biphasic systolic pressure patterns §Area under receiver operator curve, † sensitivity, ‡ specificity

The accuracy for each of the biochemical measures in predicting arterial pressure patterns

during cord occlusion was calculated for optimum cut off points that maximised sensitivity

and specificity (Table 3.11). There was little or no difference in optimum cut off values for

the three patterns of arterial pressure during cord occlusion for pH, base excess, PO2 and

PCO2 whereas for lactate, the level progressively increased as the arterial pressure responses

during cord occlusion deteriorated. For all three measures of arterial pressure response

during UCO, lactate had greater accuracy detecting these patterns than the other markers of

acid-base status.

The ROC analysis suggests that lactate is a better indicator of arterial pressure changes during

UCO than other markers of fetal acid-base status. A progressive rise in lactate is associated

with deterioration in control of fetal arterial pressure during cord occlusion.

136

Table 3.11 Accuracy of biochemical prediction of arterial pressure responses during UCO

Any Arterial Pressure Instability*

Biphasic Systolic Response**

Biphasic Diastolic Response Measure

Accuracy§ Optimum Cut Off Accuracy§ Optimum

Cut Off Accuracy§ Optimum Cut Off

Lactate (mmol/L) 90% 2.2 90% 2.7 87% 3.6

pH 88% 7.365 88% 7.360 85% 7.360

Base excess 85% -3.0 83% -3.0 79% -3.0

PO2 (mmHg) 79% 17.0 78% 17.0 73% 17.9

PCO2 (mmHg) 69% 44.6 79% 44.6 79% 44.3

* Sum of decreasing hypertensive response, biphasic systolic response and biphasic diastolic response phases ** Sum of biphasic systolic response and biphasic diastolic response as both patterns had biphasic systolic pressure patterns § Accuracy is the proportion of correctly classified cases using the biochemical cut off point selected to maximise the sensitivity and specificity

137

Any Blood Pressure Instability

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 11 - Specificity

Sens

itivi

ty

No discriminationLactpH

Biphasic SystolicResponse

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

1 - Specificity

Sens

itivi

ty

No discriminationLactpH

Diastolic BiphasicResponse

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 11 - Specificity

Sens

itivi

ty

No discriminationLactpH

Figure 3.16 Receiver operator curves for biochemical prediction of arterial pressure patterns

during cord occlusion. Sensitivity is the proportion of true positives and 1-specificity is the

proportion of false positives. The area under each of these curves was significantly greater for

lactate than pH (p=0.011, p=0.037 and p=0.049).

138

3.4 Discussion

The purpose of this study was to relate a number of biochemical and physiologic parameters,

measured before, during or after the experimental period of UCO, to histopathologic

changes that occur after asphyxia. Recent evidence suggests that 80% of infants with

neonatal encephalopathy and 69% of infants with seizures at birth have evidence of an acute

brain insult (Cowan et al., 2003). Despite advances in perinatal care over the past three

decades, the incidence of cerebral palsy attributed to birth asphyxia has not changed (Volpe,

2001b). The ability to predict the fetal response to labour and subsequent neuronal injury

may allow both intrapartum and new postnatal interventions to prevent or reduce hypoxic-

ischaemic brain injury (Perlman, 1999).

3.4.1 Cord occlusion model

The cord occlusion model in this study was designed to replicate events which may occur in

labour. Other studies have focused on relatively prolonged episodes of asphyxia (10 minutes

to 4 hours) which, if they occur clinically, are easily detected and seldom present a therapeutic

dilemma. In labour, contractions typically begin at duration of 20-30 seconds and

progressively increase to be 1 or at most 2 minutes long. When uterine contractions

compromise placental blood flow or compress the umbilical cord, a likely outcome will be

one of brief repetitive asphyxia, rather than prolonged periods of reduced placental gas

exchange.

3.4.2 Histopathological changes induced by cord occlusion model

In chronic preparation models where fetal sheep are used, different protocols of umbilical

cord occlusion lead to various patterns of fetal damage. Chronic partial umbilical cord

occlusion leads to severe acidaemia and diffuse brain damage mainly in the white matter

(Ikeda et al., 1998a). Repeated short occlusions lead to striatal or white matter damage

(Penning et al., 1994; Mallard et al., 1995a; de Haan et al., 1996; Ohyu et al., 1999; Marumo et

al., 2001). Ten minutes of complete umbilical cord occlusion results in predominant

hippocampal damage (Mallard et al., 1992; Mallard et al., 1994; Fujii et al., 2003). The reason

for the dissimilar results is unclear but may relate to the different patterns of haemodynamic

response or the different distributions of regional blood flow in the brain during the hypoxic-

ischemic insults (Ball et al., 1994; Bennet et al., 1998; Kaneko et al., 2003).

139

The neuropathologic changes identified in my study are consistent with changes identified in

other studies using similar cord occlusion protocols in fetal sheep (Clapp et al., 1988; de Haan

et al., 1996; Marumo et al., 2001). Using recurrent cord occlusion for 1 out of every 2.5

minutes and 2 out of every 5 minutes, De Haan et al. (1996) found the predominant

neuropathology was selective neuronal necrosis throughout the fetal brain; however, six of

the surviving 14 asphyxiated fetuses had focal infarcts in the parasaggital cortex, thalamus

and cerebellum. Similarly, Marumo et al. (2001) found three patterns of brain injury utilising

five three minute total cord occlusions at five minute intervals. One third of the fetuses had

no or minimal pathologic lesions, one third had dominant lesions in periventricular white

matter and one third had lesions in the cortex and thalamus.

The neuropathologic changes identified in this study are also consistent with those identified

using MRI in term human infants with neonatal encephalopathy (Cowan, 2000; Biagioni et al.,

2001; Volpe, 2001a; Cowan et al., 2003). In a study of 351 full term human infants with

neonatal encephalopathy, the lesions identified were mostly bilateral abnormalities in the

basal ganglia, thalami, cortex or white matter, although three percent had focal infarction

(Cowan et al., 2003).

3.4.3 Fetal acid-base status and neuropathologic damage

In this study I have demonstrated that traditional biochemical markers of fetal acid-base

status measured prior to cord occlusion did not predict the rate of cardiovascular or

biochemical deterioration in response to repeated cord occlusion. In addition, similar to

previous animal studies (Gunn et al., 1992; Ikeda et al., 1998a), fetal acid-base status did not

correlate with the grade of neuronal injury identified histologically 72 hours after cord

occlusion. A similar poor association between fetal acid-base status and neurologic outcome

has been identified in human studies where fetal umbilical arterial acidosis at delivery is a

relatively good measure of the severity or duration of intrauterine asphyxia (Winkler et al.,

1991). A realistic cut off point for defining pathological fetal acidaemia that correlates with

an increased risk of neurological deficit has been found to be a pH of less than 7.00 and a

base deficit of more than 16mmol/L (Goldaber et al., 1991; Sehdev et al., 1997; MacLennan,

1999). Despite defining a group of neonates at increased risk, fetal umbilical arterial acidosis

at delivery has a poor correlation with outcome as most term infants with severe acidaemia at

delivery have an uncomplicated neonatal course (Winkler et al., 1991). Even in newborns

with birth asphyxia, no differences have been shown in the first postnatal arterial pH, base

140

deficit or blood glucose in those newborns with encephalopathy compared with those

without this condition (Huang et al., 1999).

A lack of association between fetal acid-base status and the severity of neuronal injury has

consistently been reported in animal studies since 1972 in both models of global asphyxia and

partial asphyxia (Myers, 1972; Ting et al., 1983; Gunn et al., 1992; Mallard et al., 1992; de Haan

et al., 1993; Ikeda et al., 1998a). Using a global asphyxia model with three 10 minute cord

occlusions, Mallard et al. (1992) induced severe acidosis (pH 6.9, lactate 6mmol/L) and

hypotension in fetal sheep. At post-mortem, neuronal necrosis was identified in the

hippocampus, basal ganglia, and thalamus and the magnitude of neuronal necrosis correlated

with fetal hypotension but not with the severity of acidosis.

Similar results have also been identified using sheep and monkey models of partial asphyxia.

Myers (1972) exposed fetal monkeys to partial asphyxia for one to three hours and at post-

mortem found that some had evidence of neuropathology while others were normal. The

one to three hours of partial asphyxia induced focal neuronal necrosis principally in

parasaggital regions and the junctional zones between the occipital and parietal lobes. In

some fetuses there was also neuronal damage in the basal ganglia and thalamus. All of the

fetal monkeys had severe acidosis; however, it was only those with concurrent hypotension

that had neuropathology. Similarly, both mid- (Ting et al., 1983) and late- (Gunn et al., 1992)

gestation fetal lambs exposed to prolonged partial asphyxia for 60 to 120 minutes exhibited

neuronal damage only if moderate acidosis and hypotension occurred concurrently. Fetuses

that were acidaemic without hypotension showed no neuronal damage. In addition to the

concurrent requirement for acidosis and hypotension to produce neuronal injury, the severity

of hypotension has also been shown to correlate with the extent of neuropathology (Gunn et

al., 1992; Ikeda et al., 1998a).

Severe acidosis in the absence of fetal hypotension appears to induce minimal or no injury to

the fetal brain. This was illustrated in fetal sheep in a study by de Haan et al. (1993) in which

partial asphyxia for 120 minutes induced severe acidosis (pH<7.0) without hypotension.

Neuronal injury, principally in the cerebellum and to a lesser extent the cerebral cortex,

occurred in only a few fetuses. No injury was identified in the striatum or hippocampus of

any of the asphyxiated fetuses. Coagulative cell changes were found in the cortex of one

fetus, and the thalamus or medulla of two other asphyxiated fetuses. The remaining 27 of

the 34 areas of the fetal brains that were assessed had no coagulative cell changes.

141

The lack of association between fetal acid-base status and the severity of neuronal injury

probably relates to the fetal response to the insult. The variability of outcome after a fetal

exposure to either partial asphyxia or repeated short episodes of global asphyxia can be

attributed to the concept emerging from laboratory studies (Figure 1.1; Low, 1998):

Fetal Exposure + Fetal Response = Outcome

Fetal exposure to asphyxia is associated with the development of an acidosis which provides

some information about the timing and severity of the exposure (Belai et al., 1998;

MacLennan, 1999). The fetal response to the exposure is variable with the fetal brain

protected for differing periods of time before the threshold of decompensation is reached.

Initially the fetal response to the insult is compensatory, consisting of increased blood flow to

the fetal brain, heart and adrenals (Cohn et al., 1974; Peeters et al., 1979; Field et al., 1990;

Jensen et al., 1991; Reid et al., 1991) with oxygen supply to these organs being sustained and

no cerebral dysfunction or damage occurring. This response is initiated by arterial

chemoreflexes and then augmented by an increase in plasma catecholamines, ACTH, cortisol

and possibly by arginine vasopressin and angiotensin (Rurak, 1978; Jones et al., 1988; Challis

et al., 1989; Bennet et al., 1994). With prolonged or severe insults the fetus may

decompensate with hypotension, decreased cerebral blood flow and decreased oxygen supply

to the fetal brain (Low, 1998). The resulting oxygen debt and anaerobic metabolism will

initiate a cascade of biochemical events leading to brain cell dysfunction and death.

The lack of association between fetal acid-base status and the severity of neuronal injury may

also relate to the mechanisms of fetal brain injury. There is a growing body of evidence to

suggest that changes in pH, PCO2 and base excess are not the main cause of fetal brain

damage, but instead are contributing factors. The primary initiator appears to be neuronal

energy failure which initiates a complex cascade of interacting events which include

generation of reactive oxygen species and its consequences (Maulik, 1998). This oxidative

stress is paradoxically prominent during reoxygenation after the initial insult (McCord, 1985).

The importance of the reoxygenation-reperfusion phase after injury has recently been

demonstrated in fetal sheep; the severity of ischaemia and also reperfusion had significant

correlations with the severity of neuronal damage after a ten minute asphyxial episode. No

correlation with the severity of neuronal injury was found with acidemic or hypoxic

parameters (Fujii et al., 2003).

142

3.4.4 Fetal lactate and neuropathologic damage

In this study fetal lactate levels were found to be associated with both the pattern of fetal

response to repeated cord occlusion and with the severity of neuronal injury associated with

the asphyxial insult. The fetal lactate concentration at delivery accounted for 58% (r=0.759,

r2=0.576) of the variation in the grade of neuronal injury assessed 72 hours after repeated

UCO. This compares very favourably with the current markers of fetal wellbeing, namely,

umbilical artery pH and base excess, collected at delivery, which accounted for 19% and 23%

of the variability in neuronal injury respectively. Although accounting for more of the

variability in outcome than pH and base excess, fetal lactate levels after cord occlusion did

not perform as well as the duration of fetal arterial pressure instability during cord occlusion,

which accounted for 91% of the variability in outcome. Of all of the biochemical and

physiologic markers of fetal wellbeing assessed in this study, fetal arterial pressure was the

best predictor of the severity of neuronal injury. Unfortunately the assessment of fetal

cardiovascular responses expressed by cardiac output, systemic arterial pressure and blood

flow to the brain is not available in the clinical setting (Low, 1998).

There is now a growing body of evidence from animal and human studies that lactate levels

can predict neuronal injury. In this study a significant association was identified between

lactate collected immediately after cord occlusion and brain injury 72 hours later. Ikeda et al

found a similar significant correlation in fetal sheep between lactate concentrations 24 hours

after a 60 minute partial cord occlusion and brain injury assessed 72 hours later (Ikeda et al.,

1998a). In human studies, intrapartum scalp lactate assessment has been shown to be a more

sensitive diagnostic tool than the measurement of scalp pH values for predicting either an

Apgar score <4 at 5 minutes or moderate to severe neonatal encephalopathy (Kruger et al.,

1999). Urinary lactate-creatinine ratio at 6 hours of age assessed by proton nuclear magnetic

resonance spectroscopy (MRS) has been shown to have both high sensitivity (94%) and

specificity (100%) for predicting hypoxic-ischaemic encephalopathy in a group of babies with

perinatal asphyxia (Huang et al., 1999).

There have been a number of studies which have found elevated levels of lactate with proton

MRS in the fetal brain after birth asphyxia (Hanrahan et al., 1996; Hanrahan et al., 1998;

Amess et al., 1999; Barkovich et al., 1999; Malik et al., 2002). Elevated lactate has been shown

to be associated with poor outcome in the neonatal intensive care unit (Hanrahan et al.,

1996), at two months (Malik et al., 2002) and at 12 months of age (Amess et al., 1999). In

infants with birth asphyxia who had neurologic and developmental abnormalities present at

12 months of age, the most common findings on MRS performed at birth were elevated

143

lactate and diminished N-acetylaspartate peaks (Amess et al., 1999). The maximum cerebral

peak-area ratio lactate:N-acetylaspartate accurately predicted outcome at 12 months of age

with a specificity of 93% and a positive predictive value of 92% whereas neurological

assessment had a tendency for false-positive predictions (Amess et al., 1999). Magnetic

resonance spectroscopy studies have also shown that infants exposed to birth asphyxia with

elevated levels of lactate on MRS at one month of age were more likely to have an abnormal

neurodevelopmental outcome than those whose lactate levels had returned to normal

(Hanrahan et al., 1998).

Lactate levels in the fetus and neonate may be an indirect marker of the severity of hypoxic-

ischaemic insults. Mitochondrial aerobic respiration can be disturbed by severe hypoxic-

ischaemic insults leading to heightened conversion to lactate, and increased lactate levels may

enhance the severity of damage to brain tissue (Wagner & Myers, 1979). Similarly, severe

asphyxial insults may result in a decrease in cerebral blood flow. It has been suggested that

hypotension plays a major role in the development of hypoxic-ischaemic encephalopathy

(Gunn et al., 1992; Volpe, 1995a; Volpe, 2001b) and it is likely that the hypoxia-related

increase in cerebral blood flow that occurs initially as a protective response to hypoxia cannot

be maintained to the same extent during associated hypotension (Yaffe et al., 1987).

Fetal lactate levels may be an indirect marker of fetal arterial pressure. High levels of lactate

have previously been shown to be associated with progressive compromise in cardiac

function and thus output (Richardson et al., 1996). This study provides evidence that fetal

lactate is an indirect measure of fetal arterial pressure responses during UCO. For all three

patterns of arterial pressure response during UCO, fetal lactate levels had significantly greater

area under the receiver operator curves than traditional markers of fetal acid-base status. In

addition, the fetal lactate levels identified as optimum cut off values for prediction of the

three stages of cardiovascular deterioration during UCO progressively increased as the blood

pressure control during UCO deteriorated. There was no change in optimum cut off values

for pH, base excess, PO2 or PCO2 with progressive cardiovascular deterioration during UCO.

It may be through the association with arterial pressure responses during asphyxia that lactate

collected at the end of the experimental phase predicted the severity of neuronal injury.

The rise in fetal lactate in response to repeated short episodes of global asphyxia may have a

role in attenuation of brain injury. In rats, when L-lactate is infused into the cortex with

glutamate, a significant reduction in the size of the cortical lesion was noted compared to

glutamate infusion alone. When L-lactate was replaced with D-lactate, the lesion size was

greater than when glutamate was infused alone (Ros et al., 2001). Similarly in rats, an

144

intravenous infusion of lactate 30 minutes after a stimulus designed to produce a moderate

brain injury, resulted in significantly less cognitive deficits assessed 2 weeks later (Rice et al.,

2002). These results may be explained by the finding that neurons are capable of utilizing

lactate as an energy source (Schurr et al., 1997).

Rather than being a marker of the insult producing the injury, lactate may be a marker of the

endogenous response to protect the fetal brain from injury. Lactate has traditionally been

thought of as the “dead end” product of anaerobic metabolism. Recent in vitro studies have

demonstrated the ability of neurons to carry out synaptic function with lactate as the only

available carbon source (Schurr et al., 1997). When administered intravenously after neuronal

injury, lactate has been shown to be taken up at the injury site, which led to a preservation of

extracellular glucose levels after injury (Chen et al., 2000b). After brain injury, lactate levels

rise while glucose levels decline in the extracellular fluid, suggesting that lactate may be

preferentially produced in the injured tissue, as opposed to being a by-product of anaerobic

metabolism, as traditionally thought (Chen et al., 2000a). The beneficial effect of lactate may

be to provide neurons with an easily metabolised carbon source to more rapidly replenish

energy stores after injury when there is a large energy demand to re-establish ionic

homeostasis (Schousboe et al., 1997). Normally, a cell catabolises glucose to pyruvate using

two ATP molecules in turn generating four ATP molecules. If ATP levels are depleted, then

glucose may not be as readily metabolised. Thus, providing the neuron with lactate bypasses

this initial ATP requirement and may help it to recover more quickly.

3.4.5 Fetal glucose and neuropathologic damage

In this study fetal glucose concentrations prior to UCO were identified as predictors of the

fetal response to cord occlusion with high glucose levels predicting rapid biochemical and

cardiovascular deterioration. Glucose levels 72 hours after the cord occlusion protocol were

also found to strongly correlate with the grade of neuronal injury present at that time;

however, after the injury it was low glucose levels predicting higher grades of neuronal injury.

These results suggest that the fetuses with brain injury which survived the experimental phase

of this study utilised glucose as a source of energy during their recovery phase after the

repeated hypoxic-ischaemic insult.

The importance of glucose in hypoxic-ischaemic brain injury has been noted since the earliest

studies of asphyxia in monkeys (Myers et al., 1981). In that study it was noted that

supranormal levels of glucose may be associated with dangerously high levels of cerebral

lactate which may potentiate injury. Similarly, in piglets up to three days old, glucose worsens

145

hypoxic-ischaemic brain injury (LeBlanc et al., 1997). The mechanism by which supranormal

levels of glucose increase brain injury is unknown. In the piglet model of hypoxic-ischaemic

brain injury, the provision of additional glucose has been shown to preserve ATP during the

hypoxic injury and did not change cerebrospinal fluid glutamate, suggesting that glucose

worsens outcome by elevating brain L-lactate levels above the threshold for cellular injury

(LeBlanc et al., 1997). The importance of this issue warrants further direct assessment in

human infants using techniques such as proton magnetic resonance spectroscopy (Volpe,

2001a).

3.4.6 Fetal heart rate patterns and neuropathologic damage

In this study the duration of increased FHR variability and FHR overshoot correlated with

the grade of subsequent neuronal injury. The duration of severe variable decelerations, late

components to decelerations and compound FHR patterns did not have significant

correlations with the grade of neuronal injury 72 hours after the occlusion protocol.

3.4.6.1 Increased fetal heart variability and neuropathologic damage

An increase in FHR variability has been shown previously in studies of inhalational hypoxia

in fetal sheep (Dalton et al., 1977; Parer et al., 1980; Lilja et al., 1984; Field et al., 1991; Kozuma

et al., 1997) and monkeys (Ikenoue et al., 1981), placental embolisation in fetal sheep (Gagnon

et al., 1996; Murotsuki et al., 1997a) and cord occlusion in the ovine fetus (Westgate et al.,

1999) and rat (Syutkina, 1988). It is also consistent with the increase in baseline variability

seen in humans in some cases of hypoxia (Pello et al., 1991). The mechanisms which underlie

the increase in FHR variability and indeed control normal FHR variation are not fully

understood but are likely to involve a combination of activation of the carotid

chemoreceptor reflex (Kozuma et al., 1997) and sympathetic stimulation (Dalton et al., 1983;

Rosen et al., 1986; Lindecrantz et al., 1993; Segar et al., 1994).

To my knowledge, mine is the first study to report the association between the duration of

increased FHR variability during repeated short episodes of global asphyxia and the severity

of the resultant neuronal injury. A relationship between FHR variability and neuronal injury

has previously been reported in near-term fetal lambs, however in that study it was only the

FHR variability after the 60 minute episode of partial asphyxia that related to the severity of

subsequent neuronal injury (Ikeda et al., 1998b). No differences in LTV during the

prolonged period of asphyxia were found between the fetuses with mild, moderate and

severe neuronal injury with all fetuses progressively decreasing their long term variability after

60 minutes of partial cord occlusion (pre-occlusion LTV 32.0±17.3bpm; 60 minutes

146

occlusion LTV 4.0±12.6bpm) (Ikeda et al., 1998a). There were, however, differences in LTV

during the recovery phase with the fetuses with severe neuronal injury having slower recovery

in FHR variability to baseline levels than those with mild or moderate injury. The fetuses

with severe neuronal injury had a further secondary reduction in LTV 12 hours after the

insult with no FHR variability present at 24, 48 and 72 hours of recovery. In fetuses with

mild or moderate injury, the delayed secondary fall in LTV did not occur (Ikeda et al., 1998b).

The mechanism responsible for the association between the duration of increased FHR

variability and the severity of neuronal injury is unknown. There is reasonably good evidence

to suggest that brain injury after fetal asphyxia is associated with fetal hypotension (Ting et al.,

1983; Gunn et al., 1992; Ikeda et al., 1998a) however in the study by Ikeda et al. (1998b) fetal

blood pressure could not be reliably predicted from FHR variability with fetal arterial

hypotension occurring in only 34% (17/50) of the occasions when FHR variability was

reduced. Similarly in this study, increased FHR variability had both poor sensitivity and

specificity for predicting the three stages of arterial pressure decompensation during repeated

UCO (Table 3.9).

The association between increased FHR variability and neuronal injury may be indirect and

through the association of this FHR pattern with stimulation of the sympathetic nervous

system (Dalton et al., 1983; Rosen et al., 1986; Lindecrantz et al., 1993; Segar et al., 1994).

Stimulation of the sympathetic nervous system results in increases in FHR, peripheral

vascular resistance and catecholamine concentrations (Jensen & Lang, 1992). These changes

lead to a progressive increase in arterial blood pressure and rapid circulatory centralisation in

favour of the brain stem and heart at the expense of most of the peripheral organs. This

response is important for short-term adaptation and intact survival of asphyxia (Jensen &

Lang, 1992). It may be that the earlier the fetus requires this sympathetic response during

repeated UCO (which may also be responsible for the increase in FHR variability seen in this

study and that of Westgate et al, 1999), the earlier the fetus will cross the critical threshold for

cardiovascular decompensation. Once crossed, there is a fall of arterial pressure due to either

a decrease in cardiac output or a failure to maintain peripheral vascular resistance. The

combination of asphyxia and ischaemia due to hypotension results in a decrease in cerebral

oxygen consumption and, if sustained, brain damage (Low, 1998).

3.4.6.2 Fetal heart rate overshoot and neuropathologic damage

In this study and two retrospective human studies, FHR overshoot has been associated with

neurologically compromised fetuses (Goodlin & Lowe, 1974; Schifrin et al., 1994). This FHR

147

pattern, occurring when there is an acceleration immediately following a variable deceleration,

has been reported previously in ovine cord occlusion models (Westgate et al., 1999; Westgate

et al., 2001). A number of authors have attempted to ascribe clinical significance to this

pattern. Goodlin and Lowe (1974) reported that a deceleration-overshoot pattern was

associated with newborns requiring resuscitation and suggested that the pattern might be

caused by an acute hypoxic insult. Overshoot has also been described as one component of

a chronic fetal distress pattern which was associated with subsequent cerebral palsy (Shields

& Schifrin, 1988).

A number of mechanisms have been proposed for FHR overshoot. Goodlin and Lowe

(1974) suggested that the FHR overshoot pattern was a marker of fetal hypoxia. It has also

been suggested that overshoot is a marker for fetal acidaemia (Akagi et al., 1988; Saito et al.,

1988) or previous injury (Shields & Schifrin, 1988). In experimental studies in fetal sheep, the

overshoot pattern has been related to development of fetal acidosis and a decrease in cerebral

glucose metabolic rate during recurrent umbilical cord occlusions (Okamura et al., 1989).

Recently, in an ovine model of repeated umbilical cord occlusion, FHR overshoot was

reported to occur under two distinct conditions: long cord occlusions (2 minutes) and during

developing fetal acidosis and hypotension during 1-minute cord occlusions (Westgate et al.,

2001). In my study, overshoot was only seen in 2% of the 0.5-minute occlusions, 40% of the

1-minute cord occlusions, and all of the 1.5-minute cord occlusions in the regular responders.

Two of the fast responders, despite having significant acidosis and hypotension did not

display FHR overshoot but instead had recurrent late decelerations for the duration of the

experimental phase.

The aetiology of FHR overshoot is unknown but potential mechanisms include a

combination of high levels of circulating catecholamines (Goodlin & Lowe, 1974) and

impaired neural control of fetal heart rate as suggested by the progressive impairment of

EEG activity seen with frequent repeated cord occlusions (Westgate et al., 1999). It has been

suggested that it is the interplay between reduced vagal stimulation during the occlusion and

beta-adrenergic myocardial stimulation immediately after the occlusion which results in the

FHR overshoot pattern (Shields & Schifrin, 1988; Westgate et al., 2001). The initial

component of FHR deceleration caused by umbilical cord occlusion is vagal, mediated by the

carotid chemoreflex (Jensen & Hanson, 1995). As the occlusion is continued, the

deceleration is maintained by direct hypoxic myocardial depression. The longer the occlusion

continues, the less vagal tone contributes to bradycardia (Harris et al., 1982; Jensen &

Hanson, 1995). Consistent with a role of reduced myocardial vagal tone in the overshoot

148

pattern, atropine produces overshoot tachycardia in humans (Hon & Lee, 1961; Caldeyro-

Barcia et al., 1966) and fetal sheep (Harris et al., 1982; Itskovitz et al., 1982). Catecholamine

stimulation is also required because overshoot induced by atropine can be abolished by

concurrent administration of propanolol (Saito et al., 1988). These experimental findings

suggest that FHR overshoot is caused by beta-adrenergic stimulation that is unopposed

because vagal tone has been inhibited.

In this study, similar to Westgate et al. (2001), FHR overshoot was associated with

hypotension during umbilical cord occlusion. The presence of FHR overshoot in my study

was not associated with systolic or diastolic hypotension between the UCOs, but it had

sensitivity of 84% and 66% for predicting systolic hypotension and diastolic hypotension,

respectively, during the cord occlusion. The mean arterial pH associated with FHR

overshoot was 7.19±0.13 which is similar to 7.17 (Westgate et al., 2001) and 7.15 (Saito et al.,

1988) previously reported; however, overshoot was not a sensitive marker of arterial pH less

than 7.19 in my study (sensitivity 47%, specificity 93%). The findings of my study taken in

conjunction with previous studies investigating FHR overshoot suggest that the association

between the duration of FHR overshoot and the severity of neuronal injury is mediated

through hypotension during UCO and catecholamine release rather than acidosis or hypoxia.

3.4.6.3 Other fetal heart rate patterns and neuropathologic damage

The lack of association between decelerative FHR patterns and the severity of neuronal

injury seen in this study is consistent with decades of clinical research on electronic fetal heart

rate monitoring (EFM). The two studies which followed up cohorts of babies included in

the two randomised controlled trials comparing electronic FHR monitoring with intermittent

auscultation found no significant difference in cerebral palsy rates between the two groups at

the end of the follow-up period (Langendoerfer et al., 1980; Grant et al., 1989). In two of the

larger case-control studies investigating the relationship between EFM and cerebral palsy, a

significant association has been identified between abnormal cardiotocograph (CTG) findings

and cerebral palsy (Gaffney et al., 1994b; Nelson et al., 1996). In the study of Nelson et al.

(1996) the two fetal heart rate patterns found to be associated with an increase risk of

cerebral palsy were multiple late decelerations in the FHR (OR 3.9; 95%CI 1.7-9.3) and

decreased FHR variability (OR 2.7; 95%CI 1.1-5.8); however, only 0.19% of singletons with

these patterns develop cerebral palsy giving a false positive rate of 99.8%. No other FHR

patterns were found to have significant associations with the diagnosis of cerebral palsy.

149

3.4.7 Fetal arterial pressure and neuropathologic damage

In my study, the best predictor of the severity of neuronal injury after repeated UCO was the

duration of the arterial pressure changes which occurred during the cord occlusions. The

duration of fetal arterial pressure instability during cord occlusion accounted for 91% of the

variability in grading of neuronal injury 72 hours after repeated UCO. No significant

associations were identified between inter-occlusion arterial pressure and the severity of

subsequent injury.

In animal models of either global asphyxia or partial asphyxia, fetal arterial pressure responses

are associated with the severity of resultant neuronal injury (Myers, 1972; Ting et al., 1983;

Gunn et al., 1992; Mallard et al., 1992; de Haan et al., 1993; de Haan et al., 1997b; Ikeda et al.,

1998a; Fujii et al., 2003). All of these studies have suggested that the key factor precipitating

cerebral damage is hypotension and the consequent compromise of cerebral oxygen delivery.

The pattern of association between hypotension and neuronal injury varies between studies

and experimental models. In each of the aforementioned studies, both metabolic acidosis

and hypotension were required for neuronal injury. In some of the studies it was the severity

of hypotension during asphyxia which correlated with the severity of injury (Ting et al., 1983;

Gunn et al., 1992; Mallard et al., 1992). In other studies the duration of hypotension had

significant correlations with neurologic outcome, with fetuses with the most sustained

hypotension having the most severe injuries (Gunn et al., 1992; de Haan et al., 1997b; Ikeda et

al., 1998a). In one study, the duration of hypotension after the asphyxial insult was

associated with the grade of neuronal injury (Ikeda et al., 1998a).

The lack of association between inter-occlusion arterial pressure and neurologic outcome in

this study may relate to the cord occlusion model used in this study. The majority of

previous studies where a relationship has been found between arterial pressure and

neurologic outcome have used prolonged periods of asphyxia (10-180 minutes) (Myers, 1972;

Ting et al., 1983; Gunn et al., 1992; Mallard et al., 1992; de Haan et al., 1993; de Haan et al.,

1997b; Ikeda et al., 1998a; Fujii et al., 2003). In these studies, the arterial pressure responses

during asphyxia were biphasic. During the asphyxial episode there was an initial increase in

mean arterial pressure which was then followed by hypotension and it was the severity or

duration of this hypotension which correlated with outcome. In my study and in others

using brief repetitive umbilical cord occlusions in near-term fetal sheep (de Haan et al., 1997a;

de Haan et al., 1997b; Westgate et al., 1999), the pattern of arterial pressure responses is

somewhat different to those described in models with prolonged periods of asphyxia. Short

150

repetitive cord occlusions of one minute every 2.5 minutes are associated with a sustained

increase in inter-occlusion mean arterial pressure (MAP) (Westgate et al., 1999). A different

pattern is seen in short repetitive cord occlusions of 2 minutes duration every 5 minutes

where a biphasic pattern occurs with the MAP initially increased only to be followed by a

progressive fall in MAP to return to baseline levels after 90 minutes of occlusions (de Haan et

al., 1997a; Westgate et al., 1999). In my study there was a significant increase in the systolic

but not diastolic inter-occlusion pressures in the fetuses that underwent UCO during the 0.5-

minute 1.0-minute cord occlusions. During the 1.5-minute cord occlusions, the systolic

pressures returned to baseline levels. There were no episodes of inter-occlusion hypotension

in my study and this may account for the absence of association between inter-occlusion

arterial pressure and neurologic outcome.

In this study, during the 0.5-, 1.0- and 1.5-minute occlusions there were four distinct patterns

of arterial pressure response. Similar patterns of arterial pressure response during brief

repetitive UCO have been previously reported (de Haan et al., 1997a; de Haan et al., 1997b;

Marumo et al., 2001) but the association of the duration of these patterns with the severity of

neuronal injury has not been previously described. De Haan et al. (1997b) reported an

association between the total duration of arterial pressure below baseline during the entire

experimental protocol and the severity of neuronal injury. Marumo et al. (2001) reported

different patterns of neuronal injury associated with different arterial pressure responses

during UCO. Fetuses with highest arterial pressure both between occlusions and during

occlusions had their dominant neuropathologic lesions in their periventricular white matter.

Fetuses which had both a smaller increase in arterial pressure in response to UCO and those

with systemic hypotension during the final UCO had their dominant lesions in the cerebral

cortex and thalamus (Marumo et al., 2001). In that study there were no differences in the

FHR patterns or acid-base status between the fetuses with lesions in their periventricular

white matter and those with lesions in their cerebral cortex and thalamus. The only

difference between the two groups of fetuses was their arterial pressure response to the brief

repetitive asphyxia.

The mechanism behind the association between the duration of cardiovascular instability

during UCO and the severity of brain injury may relate to changes in cerebral blood flow that

occur during the UCO. Recent published data suggest that the severity of both cerebral

ischaemia during UCO and cerebral reperfusion after UCO correlates with the severity of

fetal neuronal injury (Fujii et al., 2003). The arterial pressure patterns during UCO that have

151

been described in my study may be indirect markers of changes in cerebral blood flow that

occur with cord occlusion.

3.4.8 Prediction of rate of response to repeated umbilical cord occlusion

In this study, there were clearly two different patterns of response to repeated UCO which

have arbitrarily been classified as regular or fast. All of the regular responders completed the

UCO and all but one was alive at delivery three days after the occlusion protocol. None of

the fast responders completed the UCO protocol and none was alive at delivery. Fast

responders had more rapid changes in acid-base status, glucose, cortisol, and catecholamines

in response to UCO than regular responders. They also had greater increases in their systolic

and diastolic pressures during the experimental phase of the protocol.

The only variables measured in this study which predicted whether a fetus would have a

regular or fast response to repeated UCO were lactate and glucose levels over the 4 days

prior to the experiment. There were no differences in any other biochemical markers, FHR,

FHR variability, FHR patterns in response to UCO or baseline fetal arterial pressures.

Fetuses with high baseline levels of lactate and glucose may be different from those with low

levels of these analytes in ways which have not been assessed in this study. The small sample

size and the inadvertent creation of this subgroup do not allow detailed or statistical

evaluation of the differences between fast and slow responders. All fetuses in this study were

from singleton pregnancies and there was no apparent difference in maternal condition,

surgical procedure, post-operative recovery, meconium passage or fetal weights between

these two subgroups. I speculate that the fast responder subgroup may have experienced an

insult prior to the experimental phase of this study which increased fetal lactate levels and

mobilized fetal glucose as part of the fetal stress response. Given the small sample size of the

subgroups and the fact that my study was not designed to investigate predictors of rate of

response to repeated UCO, interpretation of these results and extrapolating them to clinical

situations requires much caution. The results of my study do however support the common

clinical practice of careful fetal monitoring of growth restricted fetuses and fetuses of diabetic

mothers in labour.

Growth restricted fetuses, when compared with appropriately grown fetuses, are more

frequently hypoxaemic, hypercapnic and hyperlacticacidaemic (Pardi et al., 1987; Nicolaides et

al., 1988; Nicolaides et al., 1989; Marconi et al., 1990) and human studies have found the

incidence of intrapartum fetal distress approximates 25-50% in growth restricted fetuses

152

(Low et al., 1972; Lin et al., 1991; Spinillo et al., 1995). In addition there is evidence that

growth restricted fetuses not only begin labour with lower pH values, but they also have a

faster rate of decrease in pH per hour in labour than normally grown fetuses (Nieto et al.,

1994).

3.4.9 Integration of results into clinical practice

In my study I have found that both lactate and glucose were associated with the pattern of

fetal response to repeated cord occlusion and the severity of neuronal injury associated with

the asphyxial insult whereas no association was identified with traditional markers of fetal

acid-base status. Although one must be careful in interpretation and extrapolation of animal

models of asphyxia with small sample sizes, as used in this study, the consistency of the

results obtained with other animal studies and human data suggests that lactate warrants

further evaluation in the assessment of fetal well-being in labour and its ability to predict

neuronal injury in the newborn.

The lack of association between fetal acid-base status and the severity of neuronal injury after

repeated UCO does not negate the value of umbilical arterial and venous acid-base

assessment at birth. The presence of a significant metabolic acidosis in umbilical arterial

blood at birth confirms that an asphyxial exposure has occurred and provides an indication

of the severity and time of the insult (Belai et al., 1998; MacLennan, 1999). Perhaps more

importantly, the absence of metabolic acidaemia in umbilical arterial blood at birth excludes

the diagnosis. In addition to identifying neonates at risk of birth asphyxia, the presence of

severe metabolic acidosis at birth also identifies a group of neonates at higher risk of

encephalopathy, cardiovascular and renal complications (Nagel et al., 1995; Ingemarsson et al.,

1997).

This study, like the majority of animal studies of asphyxia, suggests that changes in fetal

arterial pressure are the key factors precipitating cerebral damage resulting from asphyxia.

Unfortunately the assessment of the fetal cardiovascular responses as expressed by cardiac

output, systemic arterial pressure and blood flow to the brain and other organ systems are, in

general, not available in labour. Thus in the clinical situation there is little information of

fetal cardiac compensation or decompensation in response to asphyxial exposure. Future

research in this area should include new techniques which can indirectly assess fetal arterial

pressure during asphyxia.

153

The ability of lactate to predict which fetuses will undergo rapid physiologic deterioration in

response to UCO and subsequent neuronal injury may aid in the selection of fetuses for early

delivery and for new rescue therapies such as hypothermia (Yager et al., 1993; Edwards et al.,

1995; Laptook et al., 1995a; Laptook et al., 1995b; Thoresen et al., 1995; Thoresen et al., 1996a;

Thoresen et al., 1996b; Gunn et al., 1997; Gunn et al., 1998a; Gunn & Gunn, 1998; Gunn et

al., 1998b; Gunn et al., 1999; Gunn, 2000; Battin et al., 2001; Battin et al., 2003), erythropoietin

(Dame et al., 2001; Chong et al., 2002; Eid & Brines, 2002; Juul, 2002; Kumral et al., 2003;

Solaroglu et al., 2003), and antioxidants (Halks-Miller et al., 1986; Mishra & Delivoria-

Papadopoulos, 1999), especially as early patient selection is important to gain the maximum

benefit from these new experimental techniques.

3.4.10 Future studies

The inadvertent creation of the fast responder subgroup reduced the number of fetuses with

complete biochemical, physiological and histopathological data and it complicated statistical

analysis. If future studies were to utilise the same experimental protocol as that used in this

study, the criteria for cessation of repeated cord occlusion should be modified so that more

fetuses survive 72 hours in utero after the cord occlusion protocol. Catheterisation of blood

vessels to the left side of the fetal brain is a potential confounder in my study. These vessels

were catheterised to enable direct sampling of the blood supply to and from the brain. Even

though no neuronal injury was identified in the right side of the fetal brain of any of the

control fetuses, it would have been better to sample fetal blood and assess arterial pressure

from another vessel, such as the brachial artery or aorta.

The experimental model used in this study involved a repeated cord occlusion protocol on

day four of a seven day study period. There were no significant changes in the fetal pH,

PCO2 or lactate levels in the the control fetuses during the seven day study period suggesting

that the experimental preparation was stable. The “regular responders” who underwent the

cord occlusion protocol continued to have a mild combined respiratory and metabolic

acidosis on day seven of the study period. The etiology of the mild acidosis in these fetuses

three days after the cord occlusion protocol is unknown. The elevated lactate levels may

relate to cellular damage from the brain injury which was known to occur in the “regular

responders”. Alternatively, as discussed in Section 3.4.4, the persistent elevation in lactate

levels may have a role in attenuation of brain injury. The elevation in fetal PCO2 levels in the

regular responders on day seven of the study period suggests that there may have been some

ongoing impairment in fetal gas exchange after the cord occlusion protocol in these fetuses

154

despite cessation of umbilical cord occlusion and confirmation of normal umbilical artery

Doppler wave forms at the completion of the experimental phase. The variable biochemical

and physiologic instability over the three days after the uniform cord occlusion protocol has

the potential to confound the histopathologic results of this study; however, as

histopathologic brain damage usually takes three to seven days to become apparent after the

insult (Nakajima et al., 2000) the variation in fetal biochemical status on days 5-7 of the study

period are unlikely to have had a significant impact on the final histopathologic grading. If

future studies were to utilise this experimental protocol more detailed biochemical and

physiologic assessment during the three day recovery phase would be valuable to enable the

integration of data regarding the fetal condition during the recovery phase into the

interpretation of the results.

The model utilised in my study induces neuronal injury in a pattern similar to that which

occurs in humans with intrapartum birth asphyxia (Cowan et al., 2003) and the cord

occlusion protocol mimics events which may occur in labour. Further exploration with this

protocol could be enhanced by the direct assessment of cerebral blood flow, ischaemia and

reperfusion. Cerebral blood flow could be assessed with ultrasonic flow probes around a

carotid artery, flow probes in the superior sagittal sinus or with laser Doppler flow probes on

the superior sagittal sinus. Additional information could also be obtained by assessing the

relationship between oxygen availability and tissue metabolism with near-infrared

spectroscopy in the fetal brain. Near infrared spectroscopy (NIRS) provides continuous,

quantifiable measurements of changes in the concentrations of the oxygen-sensitive

chromophores, oxyhaemoglobin, deoxyhaemoglobin, and the redox state of cytochrome

oxidase from intact tissues such as the fetal brain (Peebles et al., 1992; Marks et al., 1996;

Bennet et al., 1998). Cytochrome oxidase is the terminal enzyme in the oxidative

phosphorylation pathway and changes in the redox state of this enzyme are associated with

changes in oxygen availability as well as with the flow of reducing equivalents down the

phosphorylation pathway and the rate of ATP turnover (Cooper et al., 1997; Cooper &

Springett, 1997). The assessment of FHR and BP after the cord occlusion protocol would

also provide valuable information in future studies.

155

3.5 Conclusion

A brief repetitive UCO protocol in near term fetal sheep was used to induce neuropathologic

changes consistent with those identified in human studies in term infants with neonatal

encephalopathy.

Fetal lactate and glucose were associated with the pattern of fetal response to repeated cord

occlusion and with the severity of neuronal injury associated with the asphyxial insult.

Traditional biochemical markers of fetal acid-base status measured prior to cord occlusion

did not predict the rate of cardiovascular or biochemical deterioration in response to repeated

cord occlusion. When these traditional markers of acid-base status were collected after cord

occlusion no significant associations with the severity of neuronal injury were identified.

No significant associations were identified between the severity of neuronal injury and the

duration of decelerative FHR patterns or compound FHR patterns; however, this is the first

study to describe an association between the duration of both increased FHR variability and

FHR overshoot with the severity of subsequent neuronal injury.

In this study fetal arterial pressure between UCO was not associated with the grade of

neuronal injury. This study is the first to describe the strong correlation between the

duration of specific arterial pressure responses during cord occlusions and the grade of

neuronal injury.

Fetal lactate assessment may predict fetal neuronal injury through its association with fetal

arterial pressure responses during UCO. In this study this association was illustrated by fetal

lactate levels having significantly greater area under receiver operator curves predicting

arterial pressure responses during UCO than traditional markers of acid-base status. The

association was also illustrated by the progressive increase in optimum cut off values for

lactate predicting the stages of progression in cardiovascular decompensation during

progressive asphyxia. The traditional markers of acid-base status had stable optimum cut off

scores as the stages of cardiovascular deterioration progressed, suggesting that acid-base

status may function as a threshold for injury, whereas lactate optimum cut off scores

progressively increased parallel to cardiovascular decompensation with asphyxia.

156

C h a p t e r 4

THE RELATIONSHIPS BETWEEN FETAL SCALP LACTATE

LEVELS, CENTRAL LACTATE LEVELS AND OXYGEN FREE

RADICAL ACTIVITY IN THE FETAL BRAIN DURING

ASPHYXIA INDUCED BY GRADED CORD OCCLUSION IN THE

OVINE FETUS

The truth is the best you can get with your best endeavour,

the best that the best men accept –

with this you must learn to be satisfied,

retaining at the time with due humility an earnest desire for an ever larger portion.

Sir William Osler (1892)

4.1 Introduction

Fetal scalp blood sampling is the principal technique utilised to collect blood for biochemical

assessment of fetal wellbeing during labour; however, all experimental studies investigating

the ability of biochemical markers to predict fetal brain injury after asphyxia have assessed

blood samples collected from central blood vessels rather than the fetal scalp, due to

experimental design limitations.

In the previous chapter I presented results from my fetal sheep study demonstrating a

significant correlation between fetal carotid arterial lactate levels and brain injury. Similar

associations between central arterial lactate levels and fetal brain injury have been

demonstrated by others in sheep studies of asphyxia (Ikeda et al., 1998a). Prior to the

potential introduction of fetal scalp lactate measurement into clinical practice, two important

issues require investigation. First, what is the relationship between fetal scalp blood and

central blood lactate levels and how does asphyxia alter this relationship? Second, and most

important, do fetal scalp lactate levels, rather than central lactate levels, predict fetal brain

injury? The studies presented in this chapter aim to address these two important questions.

157

4.1.1 Development of an experimental model

In both animal and human studies investigating fetal scalp blood samples, a variety of

experimental designs have been utilised. All of these study designs have limitations which

require consideration prior to selection of the experimental model used in this chapter.

Human studies investigating the relationship between central and peripheral biochemical

markers of fetal wellbeing are limited by time delays between scalp blood sample collection

and cord blood collection. In some studies the delay between scalp and cord blood sampling

has been short (1-10 minutes); however, in these studies either the fetal head was on view at

the maternal perineum or the fetus was delivered by forceps or caesarean section under

maternal anaesthesia (Bowe et al., 1970; Yoshioka & Roux, 1970). In other studies the delays

between scalp and cord sampling have been as great as 60 minutes making interpretation of

the results more challenging (Kruger et al., 1998). Human studies investigating the

relationship between fetal scalp blood biochemical markers and brain injury are also limited

by the low incidence of fetal brain injury resulting from intrapartum events.

Animal experimental models can overcome the issue of time delays between sampling from

different fetal sites by allowing simultaneous blood sampling from the fetal scalp and central

blood vessels during asphyxia. Previous studies investigating the relationship between fetal

scalp blood pH and central blood pH have all used acutely instrumented fetuses. These

studies have primarily utilised partially exteriorised fetuses in either monkeys (Adamsons et

al., 1970) or sheep (Gare et al., 1967; Shier & Dilts, 1972; Aarnoudse et al., 1981). There has

been one small study in monkeys where fetuses were instrumented, returned to the uterus

and fetal scalp samples were collected vaginally during induced labour while the mother

remained under anaesthesia. In this study a median of 1 scalp sample could be collected on

each of the 11 monkeys in the study (Adamsons et al., 1970). There have been no

experimental studies sampling both fetal scalp blood and central blood in chronically

instrumented fetuses.

No animal studies have investigated the relationship between brain injury and biochemical

markers measured in fetal scalp blood. The lack of studies in this area probably relates to

experimental design limitations. Although asphyxia can be induced and scalp samples can be

reliably collected in acutely instrumented fetuses, the experimental model is not conducive to

analysis of histopathologic brain damage which can take three to seven days to become

apparent after the insult (Nakajima et al., 2000). It is this significant experimental limitation

which is likely to have limited study in this important area.

158

In this study I have chosen to use an acutely catheterised, partially exteriorised, sheep model

to investigate both the relationship between scalp lactate levels and central lactate levels and

also to investigate the relationship between fetal scalp lactate and precursors of brain injury.

This model was chosen because it allowed regular, reliable, repeated collection of adequate

volumes of fetal scalp blood during asphyxia to explore the relationships between fetal scalp

lactate and central lactate levels both at rest and during asphyxia. Concerns have been raised

previously in the interpretation of acutely instrumented models because maternal anaesthesia

is thought to adversely affect autonomic control of the fetal heart (Levinson et al., 1973; Biehl

et al., 1983a; Biehl et al., 1983b; Bachman et al., 1986; Alon et al., 1993) and the responsiveness

of the fetal heart rate to hypoxaemia (Yarnell et al., 1983; Brett et al., 1989). In some studies,

such as those which I have presented in this chapter, the collection of fetal scalp blood

necessitates an acutely instrumented model. In the acutely instrumented model that I have

used for the studies in this chapter the anaesthetic agents were selected to minimise any

adverse affects on the maternal and fetal cardiovascular and autonomic systems. Prior to

exploration of the peripheral-central lactate relationships, I have investigated the merit of the

acute model used in this study by comparing and contrasting it with the chronic model

described in Chapter Three. This evaluation was performed to aid in the interpretation of

the data obtained from the acute model.

4.1.2 The relationship between fetal scalp lactate levels and central lactate levels

Despite the growing evidence that fetal lactate levels may predict fetal brain injury in animal

models of asphyxia (Ikeda et al., 1998a) and observational studies of human labour (Kruger et

al., 1999), little is known about the relationship between scalp blood lactate levels and central

arterial lactate levels. To my knowledge, there have been only two studies that have

investigated the relationship between peripheral and central fetal lactate levels (Yoshioka &

Roux, 1970; Kruger et al., 1998). The first study compared lactate levels from three sampling

sites (fetal scalp, umbilical artery and umbilical vein) in 14 human fetuses, six who were

delivered vaginally with forceps and eight who were delivered by caesarean section (Yoshioka

& Roux, 1970). The scalp blood samples were collected one to three minutes prior to

delivery and collection of umbilical cord samples. There were no significant differences in

the mean lactate concentrations from the three sampling sites in either those fetuses delivered

vaginally or those delivered by caesarean section. On detailed inspection of the results of this

study, lactate levels in scalp blood and umbilical artery blood were identical in babies

delivered by caesarean section whereas scalp lactate levels were lower than umbilical arterial

levels if the fetus was delivered vaginally. Furthermore, if the fetus was considered

159

“distressed” by the authors, scalp lactate levels were less than umbilical artery lactate levels

irrespective of the mode of delivery (Yoshioka & Roux, 1970). These differences, although

potentially clinically significant, did not reach statistical significance; however, only six of the

14 babies studied were delivered vaginally and five of the 14 fetuses studied were considered

“distressed”. The second study addressing the relationship between central and peripheral

lactate levels compared 103 human scalp lactate samples with umbilical artery and vein lactate

samples where the scalp blood was collected within 60 minutes of delivery. Lactate

concentration in fetal scalp blood before delivery showed a modest correlation with umbilical

artery lactate and umbilical vein lactate levels (Kruger et al., 1998). In the study by Yoshioka

et al. (1970) the group means for different sampling sites were compared without considering

the paired nature of the samples and in the study by Kruger et al. (1970) simple univariate

correlation was used to assess the relationship between central and peripheral lactate levels

without considering the wide variation in time between collection of the paired samples. No

published studies have used appropriate paired statistical analysis techniques to evaluate the

differences in lactate levels in blood collected simultaneously from different sites.

Although there have been two studies which have partially investigated the relationship

between fetal scalp lactate levels and umbilical vessel lactate levels, there have been no studies

investigating the relationships between fetal scalp blood lactate, carotid arterial lactate and

jugular venous lactate levels. In addition, there have been no studies investigating the effect

of asphyxia on the relationship between central and peripheral lactate levels. The significance

of investigating the effect of asphyxia on central-peripheral biochemical relationships is

illustrated by studies validating fetal scalp pH measurement, our current biochemical

technique for intrapartum assessment of fetal wellbeing. Although it is generally accepted

that fetal scalp pH levels approximate central fetal pH levels, previous studies have

demonstrated that the relationship between scalp blood pH and fetal carotid and umbilical

arterial pH is influenced by fetal acidaemia, hypoxia, skin oedema and caput succedaneum

(Adamsons et al., 1970; Bowe et al., 1970; Yoshioka & Roux, 1970; Shier & Dilts, 1972;

Greene, 1999).

In the series of studies described in this chapter, I have investigated the relationship between

fetal scalp lactate and carotid arterial and jugular venous lactate levels both at rest and during

asphyxia induced by a cord occlusion protocol designed to mimic events which may occur in

labour.

160

4.1.3 The relationship between fetal scalp lactate levels and fetal brain injury

The experiments performed in this chapter were terminal and the passage of several days to

evaluate the evolution of neurological injury was not possible. As an alternative I have

measured oxygen free radical activity in the fetal brain, an indirect marker of potential

neuronal injury. Oxygen free radical activity was assessed both during and after the asphyxial

insult and was related to scalp lactate and other biochemical markers of fetal wellbeing.

4.1.3.1 Oxygen free radical activity and fetal brain injury

Since the first suggestion 25 years ago that oxygen free radicals may be an important

mediator of ischaemic brain injury (Flamm et al., 1978), a vast body of evidence has

developed which suggests that at a molecular level oxygen free radicals are critically involved

in the mechanisms of hypoxic-ischaemic brain injury (Maulik, 1998; Volpe, 2001b; du Plessis

& Volpe, 2002; Gitto et al., 2002; Grow & Barks, 2002; Inder et al., 2002). Substantial

experimental data are available that links oxidative stress with other pathogenic mechanisms

such as excitotoxicity, calcium overload, mitochondrial cytochrome C release, and caspase

activation in the central nervous system (Choi, 1988; Siesjo, 1988; Hockenbery et al., 1993;

Lewen et al., 2000). There is also a large amount of data to suggest that oxygen free radicals

are important in initiating apoptosis (MacManus et al., 1993; Buttke & Sandstrom, 1994;

Fernandez et al., 1995; Polyak et al., 1997; Saikumar et al., 1998), the well known cellular

suicide program which is thought to be an important part of the second wave of neuronal

cell damage that occurs during reperfusion after termination of an hypoxic-ischaemic insult

(Buonocore et al., 2001). Free radical mediated cell death also appears to be the final

common pathway for oligodendroglial cell death after hypoxic-ischaemic insults (Volpe,

2001b).

Oxygen free radicals avidly interact with the essential organic molecules of a cell including

lipids, proteins, and DNA (Figure 4.1). Consequent to these reactions, free radicals affect

cellular and extracellular constituents of neural tissue, profoundly alter its function, and may

lead to immediate necrotic or delayed programmed death of neurons (See Chapter One,

Figure 1.3; Maulik, 1998).

In this study I have focused on lipid peroxidation as a footprint of free radical activity in the

fetal brain. Lipid peroxidation was selected because it has an important role in both

mediating and extending oxidative neuronal injury (Vanucci, 1994). In addition, lipid

peroxidation is one of the earliest signs of oxidative stress, with elevated levels being

detectable within minutes of brain injury (Vagnozzi et al., 1999). Previous ovine studies using

161

ReactiveOxygenSpecies

Lipids Proteins DNA

Lipid PeroxidationLipid Radical Formation

DegredationAggregation

Protein Radical Formation

Strand BreaksBase Alterations

Membrane injury and dysfunctionTransmembrane ionic imbalanceAlteration of membrane fluidity

Alteration of membrane transport proteinsImpaired repair of DNAAltered gene expression

Mitochondrial dysfunction - decoupling of respirationMitochondrial permeability transition formation

Altered nuclear pore permeabilityInitiation of programmed cell death

Figure 4.1 Adverse effects of oxidative stress on neuronal lipids, protein and,

deoxyribonucleic acid. (DNA).

(Adapted from Maulik, 1998)

162

umbilical cord occlusion to induce asphyxia have shown evidence of lipid peroxidation in the

fetal circulation and brain perfusate (Masaoka et al., 1998; Ikeda et al., 1999; Rogers et al.,

2001).

Lipid peroxidation is the process by which unsaturated lipids react with molecular oxygen to

undergo ‘autoxidation’ or ‘peroxidation’ as it is now more widely known (Gutteridge &

Halliwell, 1990). The brain has a large lipid content: 75% of myelin, 55% of white matter,

and 35% of grey matter (Siegel et al., 1989) and the phospholipid bilayer in the neuronal cell

membrane is rich in polyunsaturated fatty acids, thereby increasing the susceptibility of

neurons to peroxidative damage. The presence of unsaturated carbon-to-carbon bonds in

the lipids enhances their roles as electron donors and the lower the saturation of these fatty

acids, the greater the propensity of oxidative reactions with free radicals (Maulik, 1998). The

occurrence of lipid peroxidation in biological membranes has been shown to cause

impairment of membrane function, decreased membrane fluidity, inactivation of membrane-

bound receptors and enzymes, and to increase non-specific permeability to ions such as

calcium (Slater & Sawyer, 1971; Eklow-Lastbom et al., 1986; Comporti, 1987; Gutteridge,

1988; Halliwell & Gutteridge, 1989). In addition, lipid hydroperoxides decompose upon

exposure to transitional metals, metal complexes and some iron proteins including

haemoglobin and myoglobin (Gutteridge, 1988; Halliwell & Gutteridge, 1989) which results

in an ever-expanding chain reaction that leads to the production of more free radicals, lipid

peroxyl and alkoxyl radicals. These attack more lipids and proteins locally, and also diffuse

out of their site of generation and produce distant effects (Halliwell & Chirico, 1993).

There is evidence from both animal and human studies showing that levels of lipid

peroxidation measured in both brain and plasma correlate with both the timing and the

severity of brain injury after hypoxic-ischaemic insults (Russell et al., 1992; Tan et al., 1998;

Ikeda et al., 1999; Candelario-Jalil et al., 2001; Marumo et al., 2001; Yu et al., 2003). When near

term rabbit fetuses were exposed to repetitive ischaemia-reperfusion (RIR) or sustained

uterine ischaemia-reperfusion (IR), the fetuses subjected to RIR were found to have

significantly more brain injury and lower cortical cell viability than IR fetuses and controls

(Tan et al., 1998). In addition the fetuses exposed to RIR had higher levels of lipid

peroxidation, 3-nitrosine, and nitrogen oxides and lower total antioxidant capacity than IR or

control fetuses. It was proposed that the higher free radical production in the fetuses

exposed to repeated ischaemia-reperfusion may be responsible for the greater fetal brain

injury in this experimental group (Tan et al., 1998). In a series of studies of transient cerebral

ischaemia in gerbils, the time course of oxidative damage in different brain regions was

163

assessed. It was found that the timing of maximal disturbance in oxidant-antioxidant balance

in the hippocampus corresponded with the time course of neuronal loss in the hippocampal

CA1 sector (Candelario-Jalil et al., 2001). In preterm human neonates, a fourfold increase in

levels of hypoxanthine (which is broken down with the aid of xanthine oxidase and oxygen

to produce oxygen free radicals) has been noted in neonates who develop cavitating

periventricular leukomalacia compared to age- and weight-matched controls who did not

develop the condition (Russell et al., 1992). Term human infants with asphyxia and hypoxic-

ischaemic encephalopathy (HIE) have been shown to have higher plasma levels of markers

of lipid peroxidation at birth than both asphyxiated neonates without HIE and healthy term

newborn infants (Yu et al., 2003). In addition when the infants were followed up at 5 months

of age, all of the babies with asphyxia without HIE and the healthy term newborn infants

were normal whereas one third of the term neonates with asphyxia, HIE and elevated lipid

peroxidation at birth were neurologically impaired.

In addition to the evidence that elevated levels of markers of lipid peroxidation are associated

with the severity of fetal brain injury, there is also evidence that treatment with novel

inhibitors of lipid peroxidation can protect neurons against damage (Andrus et al., 1997; Tan

et al., 1998, 1999; Wada et al., 1999; Ercan et al., 2001; Chang et al., 2003). When these drugs

are administered after both hypoxic-ischaemic injury (Tan et al., 1998; Chabrier et al., 1999;

Tan et al., 1999) and traumatic brain injury (Inci et al., 1998; Chabrier et al., 1999; Wada et al.,

1999; Ercan et al., 2001; Chang et al., 2003) there is less histopathologic brain damage than in

non-treated animals.

4.1.4 Aim

There are three specific aims for the studies presented in this chapter:

1. to evaluate the acutely instrumented model of repeated cord occlusion to assess

fetal scalp lactate levels during asphyxia,

2. to investigate the relationship between fetal scalp, carotid arterial and jugular

venous blood lactate levels both at rest and during asphyxia, and

3. to investigate the relationship between fetal scalp blood lactate levels and lipid

peroxidation in the fetal brain.

164

4.1.5 Hypothesis

There are two hypothesis for the experiments performed in this chapter:

1. The relationship between fetal scalp blood lactate and central blood lactate levels

are influenced by fetal asphyxia

2. Fetal scalp blood lactate will be a better predictor of lipid peroxidation in the fetal

brain than arterial pH.

165

4.2 Methods

All experimental procedures were approved by the Wentworth Area Health Service Scientific

Advisory Panel and Ethics Committee and the Westmead Hospital Animal Ethics

Committee, New South Wales, Australia. Merino ewes with singleton pregnancies were

transported to the research facility at 125 days of pregnancy to allow for acclimatisation.

4.2.1 Surgical Procedure

Thirty two ewes underwent surgery at 135±4 days (mean±SEM) of pregnancy. Anaesthesia

was induced with Thiopentone 4mg/kg in 15mL saline solution (Abbott Australasia Pty Ltd,

Cronulla, NSW, Australia), intubation performed, and anaesthesia maintained with 0.75-

1.25% Isoflurane in 100% oxygen delivered by positive pressure ventilation with the

ventilator settings selected to maintain maternal blood gases in the normal range. A maternal

femoral arterial catheter was inserted to enable continuous monitoring of maternal blood

pressure and blood gas status. Maternal blood pressure was recorded continuously onto a

computer at 1 KHz after amplification (BP transducer MLT380, bridge amplifier ML110,

PowerLab 8e, ADInstruments, Sydney, Australia). Maternal blood gas status was assessed

every 15 minutes and ventilator settings were selected to maintain maternal blood gases in the

normal range.

The uterus was exposed through a paramedian abdominal incision and a hysterotomy was

performed to deliver the fetal head, neck and upper thorax with care to minimize amniotic

fluid leakage. Vascular catheters were inserted into the left fetal carotid artery and jugular

vein. Fetal electrocardiograph (ECG) leads were sutured to the fetal chest and a vascular

snare was sutured around the umbilical cord to enable intermittent complete cord occlusion.

A circular three crystal ultrasound cuff (10MHz Triton Technology Inc., USA) was loosely

applied around the umbilical vessels to enable confirmation of normal umbilical blood flow

and complete obstruction of umbilical blood flow during cord occlusion. The fetal scalp was

shaved and the fetus was returned to the uterus with its head externalised to facilitate

repeated collection of fetal scalp blood samples. Care was taken to prevent loss of amniotic

fluid and thermal stability was maintained with overhead radiant warmers and bubble wrap

covering the fetal head between scalp blood samples to minimize heat and water loss.

166

After a thirty minute recovery phase, baseline fetal carotid arterial blood samples (0.3mL)

were collected for baseline measurement of arterial gas tensions and pH (Radiometer ABL

300, Radiometer Pty Ltd, Copenhagen, Denmark) and lactate (Accutrend® Lactate, Roche

Diagnostics GmbH, Mannheim, Germany). Repeat fetal arterial samples (0.3mL) were

collected after a further 30 minute rest phase and if the fetal preparation was stable the cord

occlusion protocol was commenced. In three of the 32 fetuses instrumented in this study,

the second arterial pH was higher than the first and a further 30 minute rest phase and

arterial blood sample were required to confirm stability of the fetal preparation.

4.2.2 Cord occlusion protocol

After confirmation that each fetal preparation was stable, each animal entered the

experimental phase of the study. Fetal arterial blood samples (4.0mL) were collected to

measure baseline arterial gas tensions, pH, lactate, glucose, cortisol, organic hydroperoxide

(OHP), malondialdehyde (MDA) and catecholamine concentrations. Fetal blood (0.4mL)

was also collected from the jugular venous catheter to measure lactate levels, OHP and

MDA. Silicone jelly was applied to the fetal scalp and a scalp blood sample was collected in a

heparinised capillary tube (20µL) from a stab wound for lactate measurement.

Twenty-four of the 32 instrumented fetuses were randomly allocated to the same umbilical

cord occlusion (UCO) protocol utilised in the previous chapter and described in detail in

Section 3.2.2. In brief, 27 complete umbilical cord occlusions were performed every three

minutes for a total of 81 minutes; the first nine occlusions were for 0.5-minutes duration

every 3 minutes; the second nine occlusions were 1.0-minute duration every 3 minutes; and

the third nine occlusions were for 1.5-minutes duration every 3 minutes. The remaining eight

fetuses served as controls with blood samples and other measurements taken as scheduled

for the experimental group but without UCO. After each fetal blood sample was withdrawn,

sterile heparinised saline was infused to replace the volume removed for analysis.

Fetal blood samples were collected every nine minutes throughout the experimental phase.

Carotid arterial blood (0.6mL) was sampled to measure arterial gas tensions, pH, lactate,

OHP and MDA. Jugular venous blood (0.5mL) was sampled to measure lactate, OHP and

MDA levels. Fetal scalp blood samples (20µL) were collected to measure lactate levels.

Additional arterial samples (1.0mL) were collected at 27 and 54 minutes to measure fetal

glucose levels. At the end of the experimental phase, a final arterial blood sample (4.0mL)

was collected to measure glucose, cortisol, organic hydroperoxide (OHP), malondialdehyde

167

(MDA) and catecholamine concentrations. The ewes and fetuses were then humanely killed

using a lethal dose of sodium phenobarbitone and the positions of the implanted catheters

were confirmed and fetal weights measured.

4.2.3 Fetal heart rate and arterial pressure assessment

Fetal heart rate and arterial pressure responses to repeated cord occlusion were recorded

throughout the experimental phase to allow validation of the acutely instrumented fetal

preparation. Fetal ECG and blood pressure were recorded continuously onto a computer at

1KHz after amplification (Bioamp, BP transducer MLT380, bridge amplifier ML110,

PowerLab 8e, ADInstruments, Sydney, Australia). FHR and FHR variability were assessed

over the 1 minute prior to each cord occlusion. Mean fetal heart rate, mean minute range

(MMR), short term variation (STV) and long term variability (LTV) were calculated using the

methods described in Section 3.2.3. Fetal arterial pressure was measured for each heart beat

and averaged for the 1 minute preceding each cord occlusion.

4.2.4 Umbilical blood flow

The ultrasound cuff around the umbilical vessels was linked to a vascular spectrum analyser

(Model SP 25 by Vasculab, Meda Sonics, USA) which enabled continuous monitoring of the

umbilical artery Doppler waveforms to confirm normal umbilical blood flow between

occlusions and complete obstruction of umbilical blood flow during cord occlusion.

4.2.5 Biochemical assessment

Plasma samples for glucose, cortisol, OHP, MDA and catecholamine measurement were

recovered after centrifugation at 1500g for 10 minutes at 4°C and were stored at -80°C until

analysed. Plasma cortisol, adrenaline, and noradrenaline were measured using the same

techniques as described in Section 3.2.6. Fetal lipid peroxide measurements were performed

by Dr CC Wang in the Chinese University of Hong Kong. The method used for

determination of organic hydroperoxides was the enzymatic technique described by Heath

and Tapel (1976) with modification as previously published (Wang et al., 1996; Rogers et al.,

2001). Briefly, 50µL plasma and 350µmol/L Tris-ethylenediamine-tetraacetic acid, pH 7.6,

were incubated with 5µmol/L catalase (17,500kU/L) at room temperature for ten minutes.

Then 25µmol/L reduced nicotinamide adenine dinucleotide phosphate (2mmol/L), 5µmol/L

glutathione peroxidase (25,000kU/L) and 50µmol/L glutathione (4.25mmol/L) were added.

168

After further incubation at 33°C for 10 minutes the absorbance of the reaction mixture at

340nm was measured on a response spectrophotometer (Gilford Instruments Laboratory,

Oberlin, Ohio). A 2.5µL aliquot of glutathione reductase (100kU/L) was added immediately,

and the reaction mixture was allowed to proceed for another ten minutes at 33°C. Its

absorbance at 340nm was again measured to determine the net amount of reduced

nicotinamide adenine dinucleotide phosphate oxidised. Values were expressed by

comparison with a standard of butyl hydroperoxide (0 to 600umol/L). The coefficient of

variation (CV) within assays was 5.8% and the between assay CV was 11.4%.

Malondialdehyde levels were estimated as reactive substances by a thiobarbituric acid

adduction method described by Richard et al. (1992) with modification. The primary reagents

were thiobarbituric acid/perchloric acid (2:1, vol/vol) mix, a calibration solution of 1,1,3,3-

tetraethoxypropane (10µmol/L) and butylated hydroxy-toluene. Each analytic run included

the assay of a reagent blank, tetraethoxypropane working standard solutions, plasma

specimens and quality control samples. The continuation of lipid peroxidation during the

reaction was prevented by addition of 10µmol/L butylated hydroxy-toluene (20gm/L). In

each reaction tube 100µL plasma and 750 µmol/L thiobarbituric acid-perchloric acid mix

were combined, vortex mixed, tightly capped and placed in a 95°C water bath. After 60

minutes the samples were chilled on ice to stop the reaction. Reagent blanks, standard

blanks, and assay blanks were left at room temperature. Two millilitres of n-butanol was

added to each tube to extract the thiobarbituric acid-malondialdehyde complex by vigorous

vortex mixing and centrifugation. The thiobarbituric acid-malondialdehyde adduct in the

butanol extract was determined by fluorometry at 532nm excitation and 553nm emission.

The CV was 1.7% within assays and 8.2% between assays.

4.2.6 Study group classification

Recent evidence suggests that spontaneously acidaemic ovine fetuses have altered

cardiovascular, endocrine and metabolic responses to acute hypoxaemia (Gardner et al.,

2002). It has been suggested by these authors that fetuses with mild spontaneous acidaemia

should either be considered as a subgroup or excluded from studies of hypoxia due to their

altered physiology. For this reason I have retrospectively classified the fetuses in this study

into two groups using a classification system previously described in ovine models of

hypoxaemia (Gardner et al., 2002): non-acidaemic (arterial baseline pH>7.25, n=18) and

spontaneously acidaemic (arterial baseline pH<7.25, n=14). As described in Section 4.2.2,

twenty-four fetuses were randomly allocated to the UCO protocol and eight served as

169

controls. Six of the eight controls were non-acidaemic and two were spontaneously

acidaemic. For clarity in presentation of results, the non-acidaemic fetuses will be referred to

as “Normal” (N), the fetuses with a spontaneous baseline metabolic acidosis will be referred

to as “Spontaneously Acidaemic” (SA) and the instrumented fetuses that were not subjected

to UCO will be referred to as “Controls” (C).

4.2.7 Statistical methods

Descriptive data are presented as mean and standard error of the mean (SEM) unless

otherwise stated. Univariate analyses of categorical data were performed with Chi-square test

and continuous data were analysed with t-tests or the non-parametric equivalent as

appropriate. Changes in fetal biochemical variables and cardiovascular parameters were

analysed using multivariate regression modelling with occlusion, time and preparation type

(acute or chronic) modelled as fixed effects and individual animals as random effects repeated

over time. Regression modelling was also used to analyse differences between samples

collected simultaneously from the three sites compared in this study matched by time and

sheep. Variable selection in multivariate modelling was based on stepwise procedures

(forwards and backwards selection) including the assessment of interaction terms to arrive at

the most parsimonious model. Comparisons of effect magnitude were based on evaluation

of changes in each outcome for one standard deviation of each independent variable

considered.

To assess the relationship between fetal scalp lactate and lipid peroxidation in the fetal brain

perfusate two types of analyses were performed. First, fetal arterial and venous OHP levels

were plotted for each individual fetus. The first time point when the venous OHP levels

began to rapidly rise (which exceeded the mean+SEM for venous OHP in controls) was

identified for each fetus. The scalp lactate and carotid artery pH were extracted for these

time points and used to calculate the range in pH and lactate associated with the

commencement of lipid peroxidation in fetal brain perfusate. Comparisons between the

proportions of pH and lactate values which were within this range were performed using

Chi-squared analysis of contingency tables. Second, receiver operator curve (ROC) analysis

was performed to evaluate the ability of biochemical markers to differentiate lipid

peroxidation in the fetal brain perfusate (jugular venous OHP - carotid artery OHP > 0).

(Hanley & McNeil, 1983; Beck & Shultz, 1986; DeLong et al., 1988). Optimum sensitivity

and specificity were chosen from the ROC tables to maximise the total of these two

170

measures of diagnostic accuracy (Hanley & McNeil, 1983). Accuracy is reported for

biochemical prediction of blood pressure responses during UCO and is defined as the

proportion of correctly classified cases (true positives + true negatives) using the biochemical

cut off point selected to maximise the sensitivity and specificity (Zweig & Campbell, 1993).

P-values<0.05 were considered statistically significant and p-values<0.10 were considered

marginally significant. SAS statistical software was used for data analysis (Statistical Analysis

Systems Institute Inc. 1999, Version 8.2., Cary, NC).

171

4.3 Results

4.3.1 Baseline fetal biochemical status prior to the experimental phase

The demographic and baseline fetal biochemical data prior to the experimental phase of this

study are presented in Table 4.1. Fetal body weights and gestational ages were similar

between the two study groups. Meconium-stained amniotic fluid was more common in the

spontaneously acidaemic fetuses (p=0.064). As a result of the retrospective classification,

fetal arterial pH was significantly lower in the spontaneously acidaemic fetuses (7.16±0.02)

when compared to normal fetuses (7.28±0.01; p<0.001). Similarly, arterial base excess was

significantly less in acidaemic fetuses (-10.19±0.84mmol/L) than normal fetuses (-6.51±1.01;

p=0.010). Fetal arterial gas tensions, lactate, glucose, cortisol and catecholamine levels were

not significantly different between the two study groups (Table 4.1).

Table 4.1 Baseline fetal biochemical condition prior to experimental phase.

Normal Acidaemic p-value

Number 18 14

Gestational age (days) 136 ± 1 134 ± 2 0.238

Weight (kg)* 3.40 ± 0.39 3.42 ± 0.37 0.286

Amniotic fluid meconium 1/18 5/14 0.064

Arterial pH 7.28 ± 0.01 7.16 ± 0.02 < 0.001

Arterial PO2 (mmHg) 19.91 ± 1.87 22.13 ± 3.00 0.517

Arterial PCO2 (mmHg) 45.28 ± 3.45 52.93 ± 3.60 0.140

Arterial Base Excess (mmol/L) -6.51 ± 1.01 -10.19 ± 0.84 0.010

Arterial Lactate (mmol/L) 4.01 ± 0.40 4.94 ± 0.80 0.307

Glucose (mmol/L) 1.21 ± 0.15 1.22 ± 0.14 0.970

Cortisol (ng/mL) 12.48 ± 1.98 17.15 ± 3.34 0.263

Adrenaline (ng/L) 0.49 ± 0.28 0.78 ± 0.36 0.532

Noradrenaline (ng/L) 22.58 ± 6.76 20.80 ± 6.38 0.862 * Measured at post-mortem

172

4.3.2 Investigation of the validity of the acutely catheterised model for the

assessment of fetal scalp blood samples

I have compared control and normal fetuses that underwent UCO in this study with fetuses

(regular, fast responders and controls) in the chronically catheterised model described in

Chapter Three. Fetal acid-base status, lactate, glucose, cortisol and catecholamine levels as

well as FHR, measures of FHR variability and inter-occlusion arterial pressure were

compared between the two models.

4.3.2.1 Fetal acid-base status and lactate responses

Fetal arterial PO2, PCO2, and base excess levels in controls were not significantly different

between the acute and chronically instrumented models (p>0.05; Figure 4.2). Acutely

instrumented controls had lower arterial pH (acutely instrumented pH 7.28±0.01 vs.

chronically instrumented pH 7.39±0.02, p=0.002) and higher arterial lactate levels (acutely

instrumented lactate 4.3±0.5mmol/L vs. chronically instrumented lactate 1.4±0.6mmol/L,

p=0.003) than chronically instrumented controls. In fetuses that underwent UCO, the

patterns of change in all measures of fetal acid-base status were similar in both acute and

chronically instrumented fetuses; however, the magnitude of the changes was different for

pH, PO2 and PCO2. Acutely instrumented fetuses that experienced UCO had significantly

greater changes in arterial pH (p=0.002), PO2 (p=0.013) and PCO2 (p<0.001) than chronically

instrumented fetuses exposed to the same occlusion protocol. Lactate and base excess

responses during UCO were not significantly different between experimental models

(p>0.05).

4.3.2.2 Fetal endocrine responses

Fetal arterial glucose, cortisol, and adrenaline levels in controls were not different between

the two experimental models (all p>0.200; Figure 4.3). Fetal noradrenaline levels in acutely

instrumented controls (25.9±6.8ng/L) were higher than chronically instrumented controls

(4.3±1.1ng/L; p=0.002). Cortisol, noradrenaline and adrenaline levels were significantly

increased in fetuses that underwent UCO in both experimental models (all p<0.05; Figure

4.3). Fetal glucose levels in fetuses that underwent UCO were increased compared to

controls in both experimental models; however, p-values associated with these increases were

only marginally significant (p-values 0.05-0.10). The magnitude of the increases in glucose,

cortisol and adrenaline levels in fetuses that underwent UCO were not significantly different

172

Figure 4.2 Comparison of arterial gas tensions, pH, base excess and lactate levels in acute and

chronically instrumented ovine fetuses during repeated umbilical cord occlusion. Data are

presented as mean±SEM. The p-values listed compare the experimental groups between

surgical models (control vs. control; occluded vs. occluded). Non-significant (ns)

comparisons are defined as p-values>0.10. A similar pattern of changes occurred in arterial

gas tensions, pH, base excess and lactate levels in both experimental models. The acidaemia

which developed in acutely instrumented fetuses had a greater respiratory component than

chronically instrumented fetuses, with acutely instrumented fetuses having greater decreases

in pH (p=0.034) and PO2 (p=0.013), and a significantly greater increase in PCO2 (p>0.001)

than chronically instrumented fetuses.

173

0 27 54 816.7

6.9

7.1

7.3

7.5 pH

Time (minutes)

Arte

rial p

H

0 27 54 810

10

20

30

40PO2

Time (minutes)

Arte

rial P

O2

(mm

Hg)

0 27 54 810

50

100

150

200PCO2

Time (minutes)

Arte

rial P

CO

2 (m

mH

g)

0 27 54 816.7

6.9

7.1

7.3

7.5 pH

Time (minutes)

Arte

rial p

H

0 27 54 810

10

20

30

40PO2

Time (minutes)

Arte

rial P

O2

(mm

Hg)

0 27 54 810

50

100

150

200PCO2

Time (minutes)

Arte

rial P

CO

2 (m

mH

g)

0 27 54 81-20

-10

0

10Base Excess

Time (minutes)

Arte

rial B

ase

Exce

ss(m

mol

/L)

0 27 54 810

5

10

15Lactate

Control Occluded

Time (minutes)

Arte

rial L

acta

te(m

mol

/L)

0 27 54 81-20

-10

0

10Base Excess

Time (minutes)

Arte

rial B

ase

Exce

ss(m

mol

/L)

Chronically InstrumentedModel

0 27 54 810

5

10

15Lactate

Control Occluded

Time (minutes)

Arte

rial L

acta

te(m

mol

/L)

Acutely InstrumentedModel

Control(p=0.002)

Occluded(p=0.002)

Control(ns)

Occluded(p=0.013)

Occluded(p<0.001)

Control(ns)

Control(ns)

Occluded(ns)

Occluded(ns)

Control(p=0.003)

173

Figure 4.3. Comparison of endocrine responses in acute and chronically instrumented ovine

fetuses during repeated umbilical cord occlusion. The p-values compare similar experimental

groups between models and non-significant (ns) comparisons are defined as p-values>0.10.

Glucose, cortisol and adrenaline levels in both controls and fetuses that underwent UCO

were not significantly different between the two experimental models. The acutely

instrumented controls had higher baseline noradrenaline levels than chronically instrumented

controls (p=0.002) and the magnitude of the increase in noradrenaline in acutely

instrumented fetuses that underwent UCO was less than the increase seen in chronically

instrumented fetuses that underwent UCO (p=0.024).

174

0 27 54 810

10

20

30

40Cortisol

Time (minutes)C

ortis

ol (n

g/m

L)

0 27 54 810

50

100

150Noradrenaline

Time (minutes)

Nor

adre

nalin

e (n

g/L)

0 27 54 810

50

100

150Noradrenaline

Time (minutes)

Nor

adre

nalin

e (n

g/L)

0 27 54 810

10

20

30

40Cortisol

Time (minutes)

Cor

tisol

(ng/

mL)

0 27 54 810

20

40

60

80Adrenaline

Control Occluded

Time (minutes)

Adre

nalin

e (n

g/L)

0 27 54 810

1

2

3

4

5Glucose

Time (minutes)

Glu

cose

(mm

ol/L

)

0 27 54 810

1

2

3

4

5Glucose

Time (minutes)

Glu

cose

(mm

ol/L

)

0 27 54 810

20

40

60

80Adrenaline

Control Occluded

Time (minutes)

Adre

nalin

e (n

g/L)

Control(ns)

Occluded(ns)

Control(ns)

Occluded(ns)

Occluded(p=0.024)

Control(p=0.002)

Control(ns)

Occluded(ns)

Chronically InstrumentedModel

Acutely InstrumentedModel

175

between the two models (p>0.05) whereas the magnitude of the increase in noradrenaline

levels in acutely instrumented fetuses (22.5±4.9ng/L to 45.4±6.4ng/L, p=0.009) was

significantly less than in chronically instrumented fetuses (5.2±1.3ng/L to 114.2±31.4ng/L,

p<0.001; group difference p=0.024).

4.3.2.3 Fetal heart rate responses

The pattern of FHR responses to the experimental phase of the study was similar in both the

acute and chronically instrumented fetuses (Figure 4.4). FHR in acutely instrumented

controls (141.3±4.3bpm) was significantly lower than chronically instrumented controls

(173.0±4.4bpm; p<0.001). Similarly, the three measures of FHR variability in acutely

instrumented controls were lower than in chronically instrumented controls (acute MMR

27.5±2.3msec vs. chronic MMR 36.7±2.4msec, p<0.001; acute STV 2.00±0.33msec vs.

chronic STV 2.87±0.37msec, p<0.001; acute LTV 12.4±1.0bpm vs. chronic LTV 16.4±1.0

bpm, P=0.006).

In both acutely instrumented fetuses that underwent UCO (p<0.001) and chronically

instrumented fetuses that underwent UCO (p=0.003) there was a significant increase in inter-

occlusion FHR over the duration of the study. The magnitude of the increase in inter-

occlusion FHR in acutely instrumented fetuses was significantly greater than chronically

instrumented fetuses (p=0.007). MMR, STV and LTV were significantly increased in fetuses

that experienced UCO compared to controls in both experimental models (p>0.05; Figure

4.4) and the magnitude of these increases were not significantly different between models

(MMR p=0.267, STV p=0.241, LTV p<0.967).

4.3.2.4 Fetal arterial pressure responses

The inter-occlusion arterial pressure for the acutely instrumented fetuses in this study and the

chronically instrumented fetuses described in the previous chapter are presented in Figure

4.5. Fetal arterial systolic and diastolic pressures in both controls and fetuses that underwent

UCO were not significantly different between the two experimental models (p>0.10).

175

Figure 4.4 Comparison of FHR and FHR variability responses in acute and chronically

instrumented fetuses during repeated umbilical cord occlusion. The p-values shown compare

similar experimental groups between models. The magnitude of increase in inter-occlusion

FHR in acutely instrumented fetuses in response to UCO was significantly greater than

chronically instrumented fetuses (p=0.007). The MMR, STV and LTV responses to UCO

were not significant different (ns) between the two experimental models (p>0.241).

176

0 9 18 27 36 45 54 63 72 810

40

80

120

160

Time (minutes)

FHR

MM

R (m

sec)

0 9 18 27 36 45 54 63 72 810

3

6

9

12

15

Time (minutes)

FHR

STV

(mse

c)

0 9 18 27 36 45 54 63 72 810

10203040506070

OccludedControl

Time (Minutes)

FHR

LTV

(bpm

)

0 9 18 27 36 45 54 63 72 81100

150

200

250

Time (min)

FHR

(bpm

)

0 9 18 27 36 45 54 63 72 81100

150

200

250

Time (minutes)

FHR

(bpm

)

0 9 18 27 36 45 54 63 72 810

40

80

120

160

Time (minutes)

FHR

MM

R (m

sec)

0 9 18 27 36 45 54 63 72 810

3

6

9

12

15

Time (minutes)

FHR

STV

(mse

c)

0 9 18 27 36 45 54 63 72 810

10203040506070

Control Occluded

Time (Minutes)

FHR

LTV

(bpm

)

Acutely InstrumentedModel

Chronically InstrumentedModel

Control(p<0.001)

Occluded(p=0.007)

Control(p<0.001)

Occluded(ns)

Occluded(ns)

Control(p<0.001)

Control(p=0.006)

Occluded(ns)

177

70

80

90

100

110

Syst

olic

Pre

ssur

e(m

mH

g)

70

80

90

100

110

Syst

olic

Pre

ssur

e(m

mH

g)

0 9 18 27 36 45 54 63 72 8120

30

40

50

60

Control Occluded

Time (minutes)

Dia

stol

ic P

ress

ure

(mm

Hg)

0 9 18 27 36 45 54 63 72 8120

30

40

50

60

Control Occluded

Time (minutes)

Dia

stol

ic P

ress

ure

(mm

Hg)

Acutely Instrumented Model Chronically Instrumented Model

Control(ns)

Occluded(ns)

Control(ns)

Occluded(ns)

Figure 4.5 Comparison of inter-occlusion fetal arterial pressure responses in acute and

chronically instrumented fetuses during repeated cord occlusion. Fetal arterial pressure

responses in acutely instrumented fetuses were not significantly different from those of

chronically instrumented fetuses (p>0.10).

178

4.3.3 Biochemical responses of normal (non-acidaemic) and spontaneously

acidaemic fetuses to repeated umbilical cord occlusion

4.3.3.1 Controls

Fetal arterial pH, gas tensions, base excess and lactate levels in controls did not significantly

change over the duration of the experimental phase (p>0.913 for all; Figure 4.2). Similarly,

fetal arterial glucose, cortisol, noradrenaline and adrenaline levels in controls did not change

over the duration of the experimental phase (p>0.324; Figure 4.3).

4.3.3.2 Fetal acid-base status and lactate responses

Fetal arterial pH (p=0.002) and PO2 (p<0.001) levels in normal fetuses decreased during the

experimental protocol whilst arterial PCO2 (p=0.046) and lactate (p<0.001) increased

compared to controls (Figure 4.2). Arterial base excess in normal fetuses decreased in

response to repeated UCO compared to controls but the difference did not reach statistical

significance (p=0.451).

The changes in pH, PO2 and base excess levels in spontaneously acidaemic fetuses in

response to the experimental protocol were smaller in magnitude than similar changes in

normal fetuses exposed to the identical experimental protocol (pH, p=0.045; PO2, p=0.013;

base excess, p=0.065; Figure 4.6). Similarly, arterial PCO2 in spontaneously acidaemic fetuses

did not significantly increase over the duration of the experimental protocol (p=0.394)

whereas in normal fetuses PCO2 levels significantly increased during the UCO protocol

(p=0.007). Arterial lactate levels increased significantly in spontaneously acidaemic fetuses

that experienced UCO (p=0.046) and there was no difference in the lactate response to UCO

between normal and spontaneously acidaemic fetuses (p=0.929).

4.3.3.3 Fetal endocrine responses

Arterial glucose, cortisol, noradrenaline and adrenaline levels increased in normal fetuses

exposed to UCO compared to controls (glucose, p=0.07; cortisol, p=0.050; noradrenaline,

p=0.009; adrenaline, p=0.001; Figure 4.7). Similar increases occurred in noradrenaline

(p<0.001) and adrenaline (p<0.001) levels in spontaneously acidaemic fetuses exposed to

UCO and there was no difference in the magnitude of these responses between normal and

spontaneously acidaemic fetuses (noradrenaline, p=0.414; adrenaline, p=0.269; Figure 4.7).

The fetal arterial glucose and cortisol levels in spontaneously acidaemic fetuses that

experienced UCO were not significantly different from controls (glucose, p=0.583; cortisol,

p=0.370).

179

-30

6.8

7.0

7.2

7.4

SA(p=0.021)

0 27 54 81

N(p<0.001)

pH

N vs. SA p=0.045

Time (minutes)

pH

-30

-16

-12

-8

-4

0

SA(p=0.855)

0 27 54 81

N(p=0.579)

Base Excess

N vs. SA p=0.065

Time (minutes)

Base

Exc

ess

(mm

ol/L

)

-300

50

100

150

200

SA(p=0.394)

0 27 54 81

N(p=0.007)

PCO2

Time (minutes)P

CO

2(m

mH

g)

N vs. SA p=0.839

-300.02.55.07.5

10.012.515.0

SA(p=0.046)

0 27 54 81

N(p=0.012)

Lactate

N vs. SA p=0.929

Time (minutes)

Lact

ate

(mm

ol/L

)

-300

10

20

30

SA(p=0.141)

0 27 54 81

N(p<0.001)

PO2

N vs. SA p=0.013

Time (minutes)

PO

2(m

mH

g)

Figure 4.6 Fetal arterial gas tensions, pH, base excess and lactate in normal (N) and

spontaneously acidaemic (SA) fetuses during experimental phase. The p-values in the legend

for each graph relate to the change in the specific variable in that subgroup over time. The p-

values in boxes compare the normal and spontaneously acidaemic fetuses over the duration

of the experimental phase considering group and time as fixed effects and individual animals

as random effects repeated over time. Despite being exposed to the same UCO protocol,

the magnitude of the reduction in arterial pH, PO2 and base excess in response to UCO in

spontaneously acidaemic fetuses were smaller than in normal fetuses (p=0.045, p=0.013 and

p=0.065 respectively). The increase in arterial lactate levels in response to repeated UCO was

not significantly different between groups (p=0.929).

180

0 810

10203040506070

Noradrenalinep=0.414

C

N(p=0.001)

SA(p<0.001)

Time (minutes)

Nor

adre

nalin

e (n

g/L)

0 810

10

20

30

40Cortisol p=0.526

C

N(p=0.050)

SA(p=0.370)

Time (minutes)

Cor

tisol

(ng/

mL)

0 810

10

20

30

40Adrenaline

p=0.269

C

N(p=0.001)

SA(p<0.001)

Time (minutes)

Adre

nalin

e (n

g/L)

0 27 54 810

1

2

3

4Glucose

p=0.09

C

N(p=0.07)

SA(p=0.583)

Time (minutes)

Glu

cose

(mm

ol/L

)

Figure 4.7 Fetal endocrine responses to repeated cord occlusion in normal (N), control (C)

and spontaneously acidaemic (SA) fetuses. P-values attached to the legends compare the

experimental group to controls over the duration of the protocol and were calculated using

regression modelling considering group and time as fixed effects and individual animals as

random effects repeated over time. In both normal and spontaneously acidaemic fetuses,

catecholamine levels were significantly increased after repeated UCO compared to controls

(p≤0.001). Arterial glucose and cortisol levels in normal fetuses exposed to UCO were

increased compared to controls (glucose, p=0.07; cortisol p=0.05); however glucose and

cortisol levels in spontaneously acidaemic fetuses that underwent UCO were not significantly

different from controls.

181

4.3.4 The relationship between fetal scalp blood lactate levels and central lactate

levels

Fetal scalp, carotid arterial and jugular venous blood lactate levels for controls, normal and

spontaneously acidaemic fetuses during the experimental phase are presented in Figures 4.8

to 4.10. These figures also present graphically the differences between fetal scalp and jugular

venous lactate levels compared to carotid arterial lactate levels (∆lactate) based on paired

analysis within the experimental group. Comparisons between the three sampling sites in the

three experimental groups are presented in Figure 4.11.

4.3.4.1 Fetal scalp blood lactate levels and carotid arterial lactate levels

The relationship between fetal scalp blood and carotid arterial lactate levels was modified by

experimental group (p=0.015). In controls (scalp-artery, -0.37± 0.46mmol/L, p=0.496;

Figure 4.8A) and spontaneously acidaemic fetuses during UCO (0.05±0.25mmol/L,

p=0.852; Figure 4.10A) fetal scalp blood lactate levels were not significantly different from

carotid arterial lactate levels. In contrast, in normal fetuses prior to UCO (time -30 minutes to

0 minutes), scalp blood lactate levels were not significantly different from carotid arterial

levels (0.10±0.44mmol/L, p=0.748; Figure 4.9A); however, during UCO fetal scalp blood

levels underestimated carotid arterial lactate levels by 0.62±0.28mmol/L in normal fetuses

(p=0.034).

The relationship between fetal scalp blood lactate levels and carotid arterial lactate levels was

not influenced by fetal pH (p=0.687), PO2 (p=0.260), PCO2 (p=0.975), base excess (p=0.957),

glucose (p=0.261), noradrenaline level (p=0.410) or time during the experimental phase

(p=0.920); however, the scalp-carotid arterial relationship was modified by adrenaline levels

(p=0.048) in all experimental groups and carotid arterial lactate levels in normal fetuses

(p=0.002). As fetal adrenaline levels increased, the difference between scalp and carotid

arterial blood lactate levels became smaller. Conversely, as arterial lactate levels increased

during cord occlusion the difference between scalp and carotid arterial lactate levels became

larger with scalp blood lactate levels tending to underestimate carotid arterial lactate levels.

The analysis of interactions suggested that the effect of arterial lactate levels on scalp lactate

levels was greater in normal than spontaneously acidaemic fetuses (p=0.002). The net result

of these combined influences on the relationship between the two sampling sites was such

that in normal fetuses scalp blood lactate levels underestimated carotid arterial lactate levels

during cord occlusion whereas in the spontaneously acidaemic fetuses scalp lactate levels

were not significantly different from carotid arterial lactate levels.

181

Figure 4.8 The relationships between fetal scalp blood, carotid arterial blood and jugular

venous blood lactate levels in controls. Fetal scalp blood lactate was not significantly

different from either carotid arterial lactate (Figure A, p=0.496) or jugular vein lactate (Figure

B, p=0.472). ∆lactate in Figure C compares fetal scalp blood lactate and jugular venous

lactate to carotid arterial lactate levels; scalp-artery and vein-artery respectively.

182

Scalp Vein-1.0

-0.5

0.0

0.5

1.0

p=0.496 p=0.893

∆La

ctat

e (m

mol

/L)

Control Fetuses

-300

3

6

9

12

15

0 27 54 81

CarotidArtery

FetalScalp

Time (min)

Lact

ate

(mm

ol/L

)

-300

3

6

9

12

15

0 27 54 81

JugularVein

FetalScalp

Time (min)

Lact

ate

(mm

ol/L

)

A

B

C

182

Figure 4.9 The relationships between fetal scalp blood, carotid arterial blood and jugular

venous blood lactate levels in normal fetuses during repeated cord occlusion. Fetal scalp

blood lactate was significantly lower than carotid arterial lactate by 0.62±0.28mmol/L (Figure

A, p=0.034) but not significantly different from jugular venous lactate (Figure B, p=0.529).

∆lactate in Figure C compares fetal scalp blood lactate and jugular venous lactate to carotid

arterial lactate levels; scalp-artery and vein-artery respectively. This figure illustrates that

during repeated cord occlusion in normal fetuses scalp blood lactate levels were less than

jugular venous lactate levels which were less than carotid arterial lactate levels.

183

Scalp Vein-1.0

-0.5

0.0

0.5

1.0

p=0.034 p=0.061

∆La

ctat

e (m

mol

/L)

Normal Fetuses

-300

3

6

9

12

15

0 27 54 81

FetalScalp

CarotidArtery

Time (min)

Lact

ate

(mm

ol/L

)

-300

3

6

9

12

15

0 27 54 81

FetalScalp

JugularVein

Time (min)

Lact

ate

(mm

ol/L

)

A

B

C

183

Figure 4.10 The relationships between fetal scalp blood, carotid arterial blood and jugular

venous blood lactate levels in spontaneously acidaemic fetuses during repeated cord

occlusion. Fetal scalp blood lactate was not significantly different from either carotid arterial

lactate (Figure A, p=0.852) or jugular venous lactate levels (Figure B, p=0.615). ∆lactate in

Figure C compares fetal scalp blood lactate and jugular venous lactate to carotid arterial

lactate levels; scalp-artery and vein-artery respectively.

184

Spontaneously AcidaemicFetuses

-300

3

6

9

12

15

0 27 54 81

FetalScalp

CarotidArtery

Time (min)

Lact

ate

(mm

ol/L

)

-300

3

6

9

12

15

0 27 54 81

FetalScalp

JugularVein

Time (min)

Lact

ate

(mm

ol/L

)

Scalp Vein-1.0

-0.5

0.0

0.5

1.0

∆La

ctat

e (m

mol

/L)

p=0.852 p=0.240

A

B

C

184

Figure 4.11 Comparison of three lactate sampling sites in control (C), normal (N) and

spontaneously acidaemic (SA) fetuses. Figure A illustrates the relationship between fetal

scalp blood lactate and carotid arterial lactate in the three experimental groups with normal

fetuses having a greater difference between fetal scalp lactate and carotid arterial lactate levels

than spontaneously acidaemic fetuses (p=0.082). Figure B illustrates the relationship

between fetal scalp lactate and jugular venous lactate levels which was constant and

independent of experimental group. Figure C illustrates the relationship between jugular

venous and carotid arterial lactate levels in the three experimental groups with divergent

relationships present between normal and spontaneously acidaemic fetuses (p=0.026).

185

∆ Lactate (Scalp - Artery)

-1.0

-0.5

0.0

0.5

1.0

p=0.082

C N SA

∆La

ctat

e (m

mol

/L)

∆ Lactate (Scalp-Vein)

-1.0

-0.5

0.0

0.5

1.0

C N SA

∆La

ctat

e (m

mol

/L)

∆ Lactate (Vein - Artery)

-1.0

-0.5

0.0

0.5

1.0

p=0.026

C N SA

∆La

ctat

e (m

mol

/L)

A

B

C

186

4.3.4.2 Fetal scalp blood lactate levels and jugular venous lactate levels

Fetal scalp lactate levels were not significantly different from jugular venous lactate levels in

controls (scalp-vein, -0.25±0.34mmol/L, p=0.472), normal fetuses that experienced UCO

(-0.18± 0.28 mmol/L, p=0.529) and spontaneously acidaemic fetuses that experienced UCO

(0.12±0.2, p=0.615; Figure 4.11).

4.3.4.3 Fetal carotid artery and jugular vein blood lactate levels

Fetal jugular venous lactate levels were not significantly different from carotid arterial levels

in either controls (vein-artery, -0.03±0.21mmol/L, p=0.893) or spontaneously acidaemic

fetuses (0.18±0.16mmol/L, p=0.240). In normal fetuses that experienced UCO fetal jugular

venous lactate levels were less than carotid arterial lactate levels (-0.36±0.18mmol/L,

p=0.061).

In all groups of fetuses in this study, the relationship between venous and arterial lactate

levels was not influenced by fetal arterial pH (p=0.200), PCO2 (p=0.369), base excess

(p=0.849), noradrenaline (p=0.829), adrenaline (p=0.297), glucose levels (p=0.827) or time

during the experimental phase (p=0.218). However, the relationship between venous and

arterial lactate levels was modified by arterial lactate (p=0.021) and PO2 (p=0.047). As arterial

lactate levels increased the venous-arterial lactate difference became significantly less

(p=0.021). The relationship between the venous-arterial lactate difference and PO2 is more

complex. In normal fetuses, as PO2 levels fell during cord occlusion, the venous-arterial

lactate difference tended to decrease in size (i.e. venous and arterial levels become more

similar). In contradistinction, when PO2 levels fell during cord occlusion in spontaneously

acidaemic fetuses, the venous-arterial lactate difference tended to increase in size with jugular

venous lactate levels tending to be higher than carotid arterial lactate levels.

187

4.3.5 The relationship between fetal scalp blood lactate and lipid peroxidation in

the fetal brain

4.3.5.1 Controls

Neither fetal carotid arterial organic hydroperoxide (OHP) levels nor jugular venous OHP

levels changed significantly over the duration of the experimental phase in controls (arterial

OHP 0.060±0.004µmol/L, p=0.128; venous OHP 0.058±0.004µmol/L, p=0.871; Figure

4.12). Fetal carotid arterial OHP was 3.4% higher than jugular venous OHP levels in

controls (artery-vein 0.002±0.0005, p<0.001). An identical pattern was observed for

malondialdehyde (MDA) levels where both carotid arterial MDA and jugular venous MDA

levels did not change significantly over time (arterial MDA 1.317±0.038µmol/L, p=0.680;

venous OHP 1.297±0.047µmol/L, p=0.848; Figure 4.12). Arterial MDA levels were

significantly greater (1.8%) than venous MDA levels in controls (artery-vein

0.023±0.002µmol/L, p<0.001).

4.3.5.2 Normal and spontaneously acidaemic fetuses

Carotid artery and jugular vein OHP and MDA levels in normal and spontaneously

acidaemic fetuses during UCO are presented in Figure 4.13. Carotid artery OHP levels were

not significantly different between normal and spontaneously acidaemic fetuses experiencing

UCO (p=0.836). Similarly jugular venous OHP levels were not significantly different

between the two groups (p=0.670). An identical response was present for MDA levels;

carotid arterial levels were not significantly different between the two groups (p=0.294) and

jugular venous levels were also not significantly different between the two groups (p=0.262).

Therefore, for all further discussions relating to lipid peroxidation in the fetal brain, the two

groups of fetuses that underwent UCO will be considered as one group.

In fetuses that underwent UCO carotid arterial OHP levels were not significantly different

from arterial OHP levels in controls (p=0.152); however, carotid arterial OHP varied over

the duration of the experimental phase of the study (p<0.001). There was an initial increase

in arterial OHP levels between nine and 27 minutes of UCO compared to baseline OHP

levels (p<0.008). This increase was followed by a decrease toward baseline levels between 36

and 54 minutes of UCO and a secondary increase from 63-81 minutes (p<0.027) compared

to OHP levels collected prior to UCO.

188

Control OHP

-300.00

0.05

0.10

0.15

JugularVein(p=0.871)

0 27 54 81

CarotidArtery(p=0.128)

Time (min)

OH

P (u

mol

/L)

Control MDA

-301.00

1.25

1.50

1.75

2.00

JugularVein(p=0.848)

0 27 54 81

CarotidArtery(p=0.680)

Time (min)

MD

A (u

mol

/L)

Figure 4.12 Lipid peroxidation in carotid arterial and jugular venous blood of control fetuses.

Data are presented as mean±SEM for carotid arterial and jugular venous blood OHP and

MDA levels in control fetuses. P-values represent change in each variable over the duration

of the experimental protocol. Both measures of lipid peroxidation did not change

significantly over the duration of this study (P>0.128).

189

Normal and SpontaneouslyAcidaemic OHP

-300.00

0.05

0.10

0.15

0 27 54 81

N ArterySA Artery

N VeinSA Vein

Time (min)

OH

P (u

mol

/L)

Normal and SpontaneouslyAcidaemic MDA

-301.00

1.25

1.50

1.75

2.00

0 27 54 81

N ArterySA Artery

N VeinSA Vein

Time (min)

MD

A (u

mol

/L)

Figure 4.13 Lipid peroxidation in carotid arterial and jugular venous blood of fetuses that

experienced umbilical cord occlusion. Carotid artery OHP (p=0.836) and MDA (p=0.294)

levels were not significantly different between the normal (N) and spontaneously acidaemic

(SA) fetuses. Similarly jugular vein OHP (p=0.670) and MDA (p=0.262) were not

significantly different between groups. Umbilical cord occlusion produced a 260% increase

in venous OHP levels (p<0.001) and a 125% increase in venous MDA levels (p<0.001)

compared to controls.

190

Jugular venous OHP levels in fetuses that underwent UCO were significantly different from

controls (p<0.001). In the first 27 minutes of the experimental phase jugular venous OHP

levels were not significantly different from controls (p>0.232). This was followed by a

massive increase in venous OHP levels during both the 1.0-minute occlusions (36-54

minutes, p<0.01) and the 1.5-minute occlusions (61-81minutes, p<0.001) such that levels

during the 1.5-minute occlusions were 260% higher than controls.

Carotid arterial MDA levels in fetuses that underwent UCO were not significantly different

from controls (p=0.595). Repeated UCO was associated with a 125% increase in venous

MDA levels compared to controls (p<0.001); no difference between the UCO and control

groups was identified until after 54 minutes of UCO (P<0.015).

4.3.5.3 Prediction of the lipid peroxidation in fetal brain perfusate

The time point at which the rapid increase in jugular venous OHP levels commenced was

determined for each fetus as described in Section 4.2.7. The rapid increase in venous OHP

levels occurred over a range of scalp lactate values from 6.2 to 8.0mmol/L and a range of

carotid arterial pH values from 7.04 to 7.19. These ranges are plotted graphically in Figure

4.14 against the arterial pH and scalp blood lactate changes seen in the two groups of fetuses

that experienced UCO. This figure illustrates the observation that the rapid increase in

venous OHP levels occurred over a wider range of arterial pH levels than scalp lactate levels.

If one were to use either arterial pH levels or scalp lactate levels as a technique of predicting

when the rapid increase in venous OHP levels began, a greater proportion of arterial pH

values (117/307; 38%) than fetal scalp blood lactate levels (52/274; 25) were within the

window where OHP production commenced (OR 2.63, 95%CI 1.80-3.94; p<0.001).

Therefore if one were to predict the rapid increase in venous OHP levels from scalp lactate

values, it could occur anywhere between 18 and 54 minutes of the experimental phase

(Figure 4.14).

In order to reflect oxidative stress in an organ, significantly higher levels of lipid peroxidation

must be detected in the venous drainage than found in the blood perfusing the organ (i.e. a

positive venous-arterial difference must exist; Rogers et al., 2001). Receiver operator curves

(ROC) were calculated for biochemical markers of fetal well-being to predict OHP

production (jugular OHP-carotid artery OHP > 0) and these are presented in Figure 4.15.

191

-30

6.8

7.0

7.2

7.4

SA

0 27 54 81

N

pH

StartOHP

Production

Time (minutes)

Car

otid

Arte

rial p

H

-300

5

10

15

SA

0 27 54 81

NLactate

StartOHP

Production

Time (minutes)

Scal

p La

ctat

e(m

mol

/L)

Figure 4.14 Range of carotid arterial pH and scalp blood lactate associated with lipid

peroxidation in jugular venous blood. The rapid increase in jugular venous OHP levels

occurred over a larger range of pH values than lactate levels during repeated UCO (OR 2.63,

95%CI 1.80-3.94; p<0.001).

192

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 11 - Specificity (false positives)

Sen

sitiv

ity (t

rue

posi

tives

)

No discriminationScalp LactateArterial LactateArterial pH

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 11 - Specificity (false positives)

Sen

sitiv

ity (t

rue

posi

tives

)

No discriminationScalp LactateArterial PO2Arterial PCO2

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 11 - Specificity (false positives)

Sen

sitiv

ity (t

rue

posi

tives

)

No discrimination

Scalp Lactate

Arterial BaseExcess

Figure 4.15 Receiver operator curves for the prediction of OHP production (jugular venous

OHP–carotid arterial OHP >0) in jugular venous blood.

Scalp lactate vs.

Arterial pH p=0.056

Scalp lactate vs.

Arterial PO2 p<0.001

Scalp lactate vs. Arterial Base Excess

p<0.001

193

Three hundred and forty nine data points were used for the generation of these curves. Fetal

scalp blood lactate (area under ROC 0.782) had significantly greater area under the ROC than

carotid arterial pH (area 0.717, p=0.056), base excess (area 0.640, p<0.001), and PO2 (0.647,

p<0.001). The area under the carotid arterial lactate ROC was not significantly different

from the fetal scalp blood lactate ROC (p>0.05). The sensitivity and specificity for optimum

cut off points and accuracy associated with these points are presented in Table 4.2. Even

though scalp blood lactate levels had significantly greater area under the ROC than routine

blood gas measures, the optimum sensitivity was still only 62.7% and specificity was 83.3%

for the prediction of OHP production in the fetal jugular venous blood.

Table 4.2 Accuracy of biochemical prediction of OHP production in brain perfusate

Area Under ROC Optimum Cut Off Sensitivity Specificity Accuracy§

Scalp Lactate (mmol/L)

0.782 (0.726-0.839) 7.0 62% 83% 72%

Arterial Lactate** (mmol/L)

0.759 (0.699-0.819) 8.7 63% 81% 72%

Arterial pH** 0.717 (0.653-0.781) 7.125 41% 92% 66%

Arterial PCO2** (mmHg)

0.724 (0.663-0.786) 65.3 59% 77% 62%

Arterial PO2 ** (mmHg)

0.647 (0.580-0.714) 14.0 62% 62% 62%

Arterial Base Excess**

0.640 (0.571-0.708) -7.3 68% 55% 61%

* OHP production ** All arterial samples collected from carotid artery catheter § Accuracy is the proportion of correctly classified cases using the biochemical cut off point selected to maximise the sensitivity and specificity

Receiver operator curves were also calculated for biochemical markers of fetal well-being to

predict MDA production (jugular MDA-carotid artery MDA > 0). Similar to OHP

production, scalp lactate had the largest area under the curve (0.725) for predicting MDA

production in the fetal brain perfusate, followed by arterial PCO2 (0.682), pH (0.614), base

excess (0.580), and PO2 (0.543). The area under the scalp lactate curve was significantly

greater than the area under the oxygen partial pressure curve (p=0.041). No significant

differences were identified between the areas under the ROC for the other biochemical

markers assessed.

194

4.4 Discussion

In labour the fetal scalp is the window for biochemical assessment of fetal wellbeing.

Despite the many technological advances in obstetrics over the last 40 years, the technique

for acquiring fetal blood during labour has not changed since its original description in 1962

(Saling, 1962). Intrapartum fetal blood sampling is limited to the skin covering the fetal

presenting part. It requires ruptured membranes and the cervix to be at least two to three

centimetres dilated. Sample collection is further complicated by the fact that the presenting

part can vary from one to ten centimetres away from the perineum. Although other

techniques such as ultrasound guided cordocentesis can provide access to fetal blood, these

techniques are invasive and are generally considered not suitable to perform on women in

labour.

The purpose of this chapter was to investigate the relationships between fetal scalp lactate

levels, central lactate levels, and oxygen free radical activity in the fetal brain during asphyxia.

Prior to exploration of these relationships I have compared and contrasted the acutely

instrumented experimental model used in this study which allowed regular, reliable access to

fetal scalp blood during asphyxia with responses from chronically instrumented fetuses.

4.4.1 Comparison of the acute and chronic instrumented models

As detailed in Section 4.1.1, all study designs to investigate fetal scalp blood in humans and

animals have important limitations. Human experimental models are limited by the

infrequent nature of birth asphyxia in human labour and the delays between the collection of

fetal scalp blood and the collection of umbilical cord blood which can vary from 1 to 60

minutes (Bowe et al., 1970; Yoshioka & Roux, 1970; Kruger et al., 1998). Animal

experimental models, whilst allowing both experimental exposure to asphyxia and

simultaneous collection of fetal scalp blood and central blood, have all been performed in

acutely instrumented fetuses with mothers under anaesthesia (Gare et al., 1967; Adamsons et

al., 1970; Shier & Dilts, 1972; Aarnoudse et al., 1981).

It is almost universally accepted that chronically instrumented fetal preparations rather than

acutely instrumented preparations should be used when investigating fetal responses to

asphyxia. However, although chronic preparations are considered optimal, there are some

circumstances, such as the studies presented in this chapter, where acute preparations are

necessary. One area of concern is the effect of maternal anaesthesia on the fetus (Rudolph &

Heymann, 1967; Levinson et al., 1973; Palahniuk & Shnider, 1974; Palahniuk & Cumming,

195

1977; Biehl et al., 1983a; Alon et al., 1993) and fetal physiologic responses to asphyxia (Smith

et al., 1975; Yarnell et al., 1983; Cheek et al., 1987; Swartz et al., 1987; Morishima et al., 1989).

The pattern and magnitude of the effects of maternal anaesthesia on the fetus and the fetal

response to asphyxia vary widely in previously published literature depending on both the

anaesthetic agents used and their doses.

4.4.1.1 Selection of maternal anaesthesia

To optimise the experimental preparation used in this study, careful consideration was given

to maternal anaesthesia. Anaesthetic agents were selected specifically to minimise the fetal

biochemical, endocrine, cardiovascular and autonomic responses to maternal anaesthesia. In

addition, the specific agents were selected to produce minimal or no change to the fetal

responses to asphyxia. A single low dose of thiopentone (4mg/kg) was chosen as the

induction agent and isoflurane at an inspired concentration of 0.75-1.25% in oxygen was

chosen as the volatile agent to maintain anaesthesia. It can be seen from Table 4.3 that the

maternal anaesthesia used in this study is significantly different from that used in the original

studies using acutely instrumented fetuses to validate fetal scalp pH assessment.

Table 4.3 Comparison of maternal anaesthesia techniques between experimental models

Author Induction of Anaesthesia Maintenance of Anaesthesia

Gare et al., 1967 2% Lidocaine local anaesthesia

2% Lidocaine local anaesthesia

Adamsons et al., 1970 Pentobarbital 10mg/kg IM* Nitrous oxide Halothane Further pentobarbital if required

Shier & Dilts, 1972 Pentobarbital 15.2mg/kg IV† Supplemental pentobarbital if required

Aarnoudse et al., 1981 Pentobarbital 30mg/kg IV† Nitrous oxide Piritramide 7.5mg SC‡ 2 hourly Pancuronium 0.08mg/kg IV† hourly

Chapter 4, Pennell, 2003 Thiopentone§ 4mg/kg IV† Isoflurane 0.75-1.25% *IM intramuscular injection, IV† intravenous injection, SC‡ subcutaneously § Conversion of pentobarbital:thiopentone is 3:1(Fragen & Avram, 1999)

196

A single low dose of thiopentone was chosen as the anaesthetic induction agent for the

studies in this chapter. In pregnant ewes (125 days gestational age), low dose thiopentone

has been shown to have transient effects on both the mother and fetus (Alon et al., 1993). A

single intravenous bolus of 5mg/kg produced a transient decrease in maternal mean arterial

pressure (MAP) and uterine blood flow. This was associated with a transient increase in

FHR, MAP and PCO2 and a decrease in fetal pH. The duration of these effects was

approximately five minutes after which all the physiological variables returned to their pre-

anaesthetic levels (Alon et al., 1993). In humans, when considering Apgar scores and the

need for neonatal resuscitation, low doses of thiopentone have been shown to not depress

the fetus whereas larger doses, such as 6mg/kg and 8mg/kg, are associated with increased

risk of neonatal depression at birth (Kosaka et al., 1969). In women given 8mg/kg

thiopentone for induction of anaesthesia prior to caesarean section, 43% of the babies had

depressed Apgar scores at birth (Kosaka et al., 1969). The single low dose of thiopentone

was chosen because both repeated and high doses of this drug, although not altering the

maternal or fetal peak concentrations, have been shown to significantly increase the time

taken to clear the drug from both the maternal and fetal circulation (Kosaka et al., 1969).

There is a swift decline in thiopentone concentration in maternal blood after a single

intravenous administration with levels falling to 20% of the initial peak level within three

minutes of administration. Although the drug is highly lipid soluble and maximum fetal

levels are reached within three minutes of administration, the maximum fetal blood levels are

only 15% of the initial maternal peak level and within ten minutes of maternal administration,

the fetal arterial levels have fallen to 5% of the peak maternal concentration (Kosaka et al.,

1969).

Induction of anaesthesia can be induced with other agents such as high dose volatile agents

or ketamine; these were avoided in this study due to their adverse fetal effects. High dose

volatile agents have been shown to induce greater depression of myocardial contractility than

low dose barbiturates (Frankl & Poole-Wilson, 1981) by altering calcium uptake by cardiac

sarcoplasmic reticulum (Blanck & Stevenson, 1988). In ovine fetuses, ketamine abolished the

fetal hypertension and bradycardia produced by partial cord occlusion (Swartz et al., 1987).

In this study, isoflurane was chosen as the volatile agent to maintain anaesthesia because it

has significantly less effect on baroreceptor reflexes than other inhaled anaesthetics (Kotrly et

al., 1984). In general, all volatile anaesthetics depress baroreceptor reflexes, but not equally

(Fragen & Avram, 1999). When comparing volatile anaesthetic agents it is typical to compare

effects at specific minimum alveolar concentrations (MAC). The MAC of an anaesthetic is

197

the concentration in the alveolus which is just capable of blocking response to a noxious

stimulus in 50% of a population. As the alveolar concentration is proportional to blood and

brain concentration at equilibrium, MAC’s can be used to compare therapeutic ratios of

various drugs. Halothane depresses baroreceptor reflexes such that by 2.5 MAC there is

virtually no tachycardic response to changes in blood pressure (Duke et al., 1977). Nitrous

oxide combined with halothane produced less depression in baroreceptor reflexes than

halothane alone at the same MAC levels (Duke & Trosky, 1980). Nitrous oxide combined

with enflurane produced virtually the same degree of depression as nitrous oxide and

halothane (Morton et al., 1980). At one MAC, isoflurane has been shown to reduce the

sensitivity of the baroreceptors controlling heart rate by 10-15% whereas halothane

combined with nitrous oxide or enflurane alone reduced baroreceptor sensitivity by 70% and

halothane alone reduced it by 85% (Kotrly et al., 1984).

A variety of agents to maintain anaesthesia have been used previously to investigate the

effects of maternal anaesthesia on the ovine fetus. Isoflurane at an inspired concentration of

2% in oxygen has been shown to not significantly alter fetal arterial pressure or FHR over 90

minutes of anaesthesia; however, a 25% decrease in fetal cardiac index and small decreases

in pH and base excess were noted (Biehl et al., 1983b). When isoflurane was used at lower

concentrations it did not result in fetal acidosis or significant falls in cardiac output (Bachman

et al., 1986) Halothane, whilst not altering uterine artery blood flow (Cheek et al., 1987), has

been shown to be associated with a 9-33% fall in fetal mean arterial pressure. Despite the fall

in fetal MAP, no change was demonstrated in cardiac output, placental blood flow or

regional blood flow to vital organs (Biehl et al., 1983a; Yarnell et al., 1983; Cheek et al., 1987).

Medroxyflurane at one MAC has been shown to be associated with moderate falls in

maternal arterial pressures, cardiac output and uterine blood flow whilst no significant effects

on fetal heart rate, MAP or acid-base status were noted after 90 minutes of anaesthesia

(Smith et al., 1975). It was only when the dose of medroxyflurane was increased to two MAC

that marked changes in maternal arterial pressures, cardiac output and uterine blood flow

became apparent and resulted in a mixed respiratory and metabolic acidosis in the fetus

(Smith et al., 1975).

A variety of anaesthetic agents have also been used in studies investigating the effects of

anaesthesia on the fetal responses to asphyxia. When the ewe was anaesthetised with

halothane during fetal asphyxia it was noted that the increase in arterial pressure in response

to asphyxia was slightly smaller (9%) and the increase in FHR was greater than with asphyxia

alone (Yarnell et al., 1983; Cheek et al., 1987). Despite these differences, the addition of

198

halothane did not alter the ability of the fetus to increase cardiac output or cerebral blood

flow in response to asphyxia, nor were vital organ blood flows altered by (Yarnell et al., 1983;

Cheek et al., 1987). Both ketamine and lidocaine adversely affected the fetal response to

asphyxia. Ketamine has been shown to abolish fetal hypertension and bradycardia produced

by partial cord occlusion (Swartz et al., 1987). Lidocaine has been shown to be associated with

loss of fetal cardiovascular adaptations to asphyxia with significant increase in fetal PCO2, and

decreases in pH, MAP, and blood flow to the brain, heart and adrenals (Morishima et al.,

1989).

The anaesthetic agents used in this study have previously been used in studies involving fetal

sheep at similar doses (Baker et al., 1990; Alon et al., 1993). I chose to use 0.75% to 1.25%

inspired concentration of isoflurane in oxygen, as the MAC for isoflurane previously

determined for both pregnant ewes and fetuses were 0.86% and 0.34% respectively

(Bachman et al., 1986). In pregnant ewes, isoflurane at 1% inspired concentration has been

shown to be associated with a 28% increase in uterine blood flow, a 25% increase in maternal

heart rate and a 20% decrease in maternal blood pressure (Baker et al., 1990; Alon et al.,

1993). Similar responses are seen in pregnant humans under isoflurane anaesthesia (Jouppila

et al., 1979). Despite these cardiovascular changes in pregnant ewes, fetal arterial pressures,

heart rate and acid-base status were not altered by isoflurane anaesthesia compared with non-

anaesthetised controls (Alon et al., 1993). Fetal heart rate variability (long term variability)

was however reduced in the presence of 1% isoflurane (Alon et al., 1993).

The anaesthetic agents used in this study have also been used previously in an ovine model of

asphyxia and were shown to not alter fetal responses during asphyxia (Baker et al., 1990).

The same pattern of changes occurred in fetal acid-base status, fetal arterial pressure, cerebral

oxygen consumption and oxygen delivery in the asphyxia-alone and the isoflurane-asphyxia

groups. In addition, during asphyxia similar responses were seen in the two groups with total

brain, heart and adrenal blood flows increasing while flow to the spleen and carcass

decreased. The only difference between the asphyxia-alone and the isoflurane-asphyxia

groups was that the magnitude of the changes in pH and PCO2 were greater in the isoflurane-

asphyxia group (Baker et al., 1990). Similar results have also been shown in young lambs

exposed to asphyxia with and without isoflurane (Brett et al., 1989).

In summary, based on previous studies, the agents used to anaesthetise pregnant ewes in this

study have been chosen to have minimal adverse effects on fetal responses to asphyxia and

were unlikely to have any significant effects on fetal scalp blood collection or interpretation.

199

4.4.1.2 Similarities and differences between the acute and chronically instrumented

models

4.4.1.2.1 Control fetuses

The acutely instrumented model used in this chapter has been compared and contrasted with

the chronically instrumented model described in the previous chapter. In control fetuses, no

significant differences were found in fetal PO2, PCO2, base excess, glucose, cortisol or

adrenaline levels between the two models. However, the acutely instrumented fetuses had

lower pH levels, and higher lactate and noradrenaline levels. In addition, compared to

chronically instrumented fetuses, the acutely instrumented fetuses in this study had lower

baseline FHR and a 25% reduction in FHR variability. Similar reductions in FHR variability

have previously been reported in fetuses under the influence of maternal isoflurane

anaesthesia (Alon et al., 1993). Differences between the acute and chronically instrumented

control are likely to have resulted from the combination of fetal manipulation,

instrumentation, and the effect of maternal anaesthesia (Biehl et al., 1983b; Taylor et al., 1992).

The higher baseline FHR in the chronically instrumented fetuses may suggest subclinical

inflammation or infection; however, I have provided no direct evidence for this in my study.

In addition, previous studies that have measured white blood cell counts in chronically

catheterised fetuses (Nitsos et al., 2002) found levels that were not different from levels

reported by others in non-catheterised fetuses (Kallapur et al., 2001) suggesting that chronic

catheterisation of fetal sheep did not elicit a systemic inflammatory response.

4.4.1.2.2 Fetuses subjected to repeated cord occlusion

In this study a similar pattern of changes in fetal acid-base status, fetal arterial pressures and

fetal heart rate occurred in both the chronically instrumented fetuses exposed to asphyxia and

acutely instrumented fetuses exposed to the same insult whilst under anaesthesia. These

observations have previously been reported by other authors (Baker et al., 1990). In this

study I have demonstrated for the first time that there were no differences in the pattern of

endocrine responses seen in acute and chronically instrumented fetuses exposed to the same

asphyxial challenge. In fetuses subjected to UCO there were no significant differences in

glucose, cortisol or adrenaline levels between the two models. Despite the 25% reduction in

FHR variability in acutely instrumented controls, when the acutely instrumented fetuses were

exposed to an asphyxial insult, all three measures of FHR variability increased significantly

and were no longer significantly different from values obtained in chronically instrumented

fetuses exposed to the identical asphyxial insult. This similarity between the acute and

chronically instrumented fetal preparations has not previously been reported.

200

The acutely instrumented fetuses subjected to UCO had the same lactate and base excess

responses as chronically instrumented fetuses exposed to the same asphyxial insult. Despite

the same degree of metabolic acidosis, the acutely instrumented fetuses exposed to asphyxia

developed a more severe respiratory acidosis than those in the chronic preparation. This was

demonstrated in the acutely instrumented fetuses by a more rapid fall in arterial pH and more

dramatic changes in PO2 and PCO2 than seen in the chronically instrumented fetuses exposed

to the identical asphyxial insult. This response has previously been reported in ovine fetuses

exposed to asphyxia under the influence of isoflurane anaesthesia (Baker et al., 1990) and may

be related to impaired placental function under anaesthesia (Taylor et al., 1992). Fetal PO2

and PCO2 equilibrate rapidly across the placenta (Daniels, 1966), unlike lactate, which is slow

to equilibrate across the placenta as it is reliant on facilitated diffusion coupled with proton

transfer (Piquard et al., 1990). Fetal respiratory gas levels are reliant on both umbilical blood

flow and uterine blood flow and the additional respiratory acidosis which developed in the

acute fetal preparation during asphyxia may have resulted from the maternal left lateral

position, the mechanical ventilation used in the acute preparation, or from alterations in the

fetus’ ability to direct blood toward the placenta during asphyxia. It is less likely that changes

in fetal cardiac output or uterine blood flow during asphyxia were responsible for the

additional acidosis as both have been shown to be preserved during isoflurane anaesthesia

and asphyxia (Baker et al., 1990; Alon et al., 1993).

Another difference between the two models was the magnitude of increase in FHR during

asphyxia. The inter-occlusion FHR in the acutely instrumented controls was less than the

chronically instrumented controls; however, when exposed to asphyxia the FHR increased to

a higher level in the acute than the chronic fetal preparation. This does not suggest blunting

of the hypoxaemia-induced tachycardia as previously suggested in lambs exposed to

isoflurane (Brett et al., 1989), but rather it suggests enhanced responsiveness of the FHR to

hypoxia. This may be secondary to the elevated noradrenaline levels seen in acutely

instrumented fetuses prior to asphyxia rather than enhanced sensitivity of the fetal heart to

catecholamines as both thiopentone (Skovsted et al., 1970) and isoflurane (Joas & Stevens,

1971) have previously been shown to not sensitise the heart to catecholamines. The elevated

levels of noradrenaline in controls in the acute compared to chronic preparation may relate to

the maternal and fetal stresses associated with the surgery (Palahniuk & Cumming, 1977). An

alternative more likely hypothesis is that the elevated levels of noradrenaline are secondary to

the isoflurane anaesthesia because there were no differences between the adrenaline and

cortisol levels between models and isoflurane anaesthesia has previously been shown to

201

increase noradrenaline levels in animals (Hellyer et al., 1989) and humans (Tanaka et al., 1996;

Segawa et al., 1998).

4.4.1.3 What is the optimum model for assessing fetal scalp blood during asphyxia?

There is little doubt that when exploring fetal responses to asphyxia it is ideal for the

experimental design to replicate the natural setting in as many ways as practically possible.

Acutely instrumented fetal preparations are at risk of maternal and fetal effects from

anaesthesia; however, as described both in this chapter and by others (Brett et al., 1989; Baker

et al., 1990) this can be minimised by appropriate selection of anaesthetic agents and their

doses. Chronic fetal preparations, although not influenced by maternal anaesthesia, are

influenced by other potential confounders which include surgical recovery, confinement

during experiments, multiple fetal catheters, and potential infection. There is also evidence

that placental function is temporarily disturbed after fetal surgery as illustrated by impaired

fetal growth after surgery to implant fetal catheters (Newnham et al., 1986) and the 15-30

days taken for amniotic fluid volume to return to normal after simple fetal surgery to implant

two catheters into the amniotic sac (Pennell et al., 2003).

After considering the fetal biochemical, endocrine and cardiovascular responses to asphyxia,

I have found the acute preparation used in this study does not appear to impair the ability of

the fetal lamb to coordinate neural, endocrine and local tissue responses to asphyxia.

Specifically, during asphyxia there was no evidence identified in this study that maternal

anaesthesia adversely affected autonomic control of the fetal heart or the responsiveness of

the FHR to hypoxemia. Although there were some differences between the acutely

instrumented model used in this chapter and the chronically instrumented model used in the

previous chapter, the differences identified are unlikely to be detrimental to the investigation

of relationships between fetal scalp blood lactate levels, central blood lactate levels and fetal

brain injury.

I propose that the model that I have chosen to use in this study is currently the best that is

available to assess fetal scalp blood in a regular, reliable, repeated manner during asphyxia,

allowing for collection of adequate volumes of fetal scalp blood for biochemical assessment.

However, in the absence of the requirement to collect fetal scalp blood, chronic fetal

catheterisation remains the model of choice for other experiments.

202

4.4.2 The relationship between fetal scalp blood, carotid arterial and jugular venous

blood lactate levels

Despite growing evidence that fetal lactate levels may predict fetal brain injury, both in

animal models of asphyxia (Ikeda et al., 1998a) and observational studies of human labour

(Kruger et al., 1999), little is known about the relationship between scalp blood lactate levels

and central arterial lactate levels. My study is the first to explore the relationships between

fetal scalp lactate levels and carotid arterial and jugular venous lactate levels. Two previous

studies have partially investigated the relationship between fetal scalp blood lactate and

umbilical artery and vein lactate levels (Yoshioka & Roux, 1970; Kruger et al., 1998); however

the statistical analyses performed in these studies were limited. Unlike previous studies which

have compared group mean lactate levels between different sampling sites (Yoshioka &

Roux, 1970) or used simple univariate correlation to compare blood from different sampling

sites at different times (Kruger et al., 1998), I have used statistical techniques which evaluate

differences between matched samples collected simultaneously whilst taking into account the

repeated nature of the sampling on each of the fetuses.

4.4.2.1 Scalp lactate and central lactate levels in control and normal fetuses during

asphyxia

In this chapter I have shown that in the absence of asphyxia, scalp lactate levels are not

significantly different from carotid arterial and jugular venous lactate levels; however, when

fetuses are exposed to asphyxia, the relationship between fetal scalp blood lactate and carotid

arterial lactate is altered. During asphyxia fetal scalp blood lactate levels were found to

underestimate carotid arterial lactate levels by a mean of 0.6mmol/L (lactates<lactatea).

These observations are similar to those which were suggested by Yoshioka et al. (1970) where

mean scalp blood lactate levels and mean umbilical artery lactate levels were not significantly

different when collected from “non-distressed” fetuses at elective caesarean section;

however, when samples were collected from “distressed” fetuses, scalp blood lactate levels

were less than umbilical artery lactate levels. The differences identified by Yoshioka et al.

(1970) were clinically significant but failed to reach statistical significance which probably

relates to their small sample size (five “distressed” and nine “non-distressed” fetuses) and

their statistical analysis.

The observation in my study that during asphyxia fetal scalp lactate tends to underestimate

carotid arterial lactate levels is the opposite of the relationship previously described for pH

assessment from the same two sampling sites. During induced labour in monkeys, fetal scalp

203

pH overestimated carotid arterial pH by an average of 0.047 pH units (pHs<pHa) (Adamsons

et al., 1970). Similarly, in human fetuses scalp blood pH can be lower than umbilical arterial

pH by as much as 0.06 pH units (Bretscher & Saling, 1967; Bowe et al., 1970). These

observations considered together suggest that the additional acidosis that develops in the

fetal scalp during asphyxia is respiratory rather than metabolic in origin. This is supported by

evidence that during hypoxia fetal scalp PO2 is not significantly different from arterial PO2

(difference 0.2mmHg) whereas fetal scalp PCO2 is significantly higher than arterial PCO2

(difference 8mmHg; Shier & Dilts, 1972). The additional carbon dioxide can combine with

water to form carbonic acid and result in a respiratory acidosis and an additional fall in fetal

scalp blood pH.

In my study fetal scalp blood lactate levels were similar to jugular vein lactate levels and this

relationship was not influenced by asphyxia or fetal condition as demonstrated by the

preservation of this relationship in spontaneously acidaemic fetuses during asphyxia. A

similar response has previously been described in fetal monkeys and fetal sheep where pH,

PO2 and PCO2 in scalp capillary blood closely resembled jugular venous blood levels even

when pH levels were as low as 6.75 (Adamsons et al., 1970; Shier & Dilts, 1972). The

observation that fetal scalp blood more closely resembles jugular venous blood than carotid

arterial blood may be advantageous. The composition of jugular venous blood is determined

not only by the composition of arterial blood but also by the relationship between flow rate

and metabolic needs of the brain (Figure 4.16). Therefore, jugular venous blood and fetal

scalp blood lactate levels may be more informative than arterial lactate levels regarding the

operational conditions of the fetal brain.

In this study I have used multivariate regression techniques to explore factors which

influence the relationship between fetal scalp blood lactate levels and carotid arterial lactate

levels in an attempt to elucidate potential mechanisms for why fetal scalp blood

underestimates carotid arterial lactate levels during asphyxia. Through this exploratory

analysis I have found that the scalp-arterial lactate difference was not influenced by fetal

arterial pH, PO2, PCO2, base excess, glucose levels or noradrenaline levels; however, it was

influenced by arterial lactate levels and adrenaline levels. In all fetuses in this study, as arterial

lactate levels increased the scalp-arterial lactate difference became greater. Conversely, as

fetal adrenaline levels increased, the scalp-arterial lactate difference became smaller with the

magnitude of the adrenaline effect being approximately half the size of the lactate effect. The

association between arterial lactate levels, adrenaline levels and the difference between scalp

204

Figure 4.16 Control of jugular venous and scalp capillary blood content. The composition of jugular venous blood and scalp capillary blood is determined not

only by the composition of carotid arterial blood, but also by the relationship between flow rate and metabolic needs of the fetal brain and scalp tissues

respectively. This figure illustrates the multiple factors which directly influence carotid arterial blood content, blood flow and tissue metabolism and therefore

indirectly influence the content of jugular venous or scalp capillary blood.

BLOOD

FLOW

CAROTID

ARTERIAL BLOOD

TISSUE

METABOLISM

JUGULAR VENOUS BLOOD OR

SCALP CAPILLARY BLOOD

Placental Function

Cardiovascular and Neuroendocrine

Reflexes

Nutrient Supply

Previous Stresses on Fetus

Metabolic Processes to Facilitate Energy

Supply to Fetal Brain

205

and arterial lactate levels suggests that the difference between these sites during asphyxia may

be secondary to changes in fetal scalp perfusion.

During asphyxia fetal scalp blood flow rates are altered as part of the fetal cardiovascular

responses to redistribute cardiac output to vital organs (Cohn et al., 1974; Itskovitz et al.,

1987; Jensen et al., 1987; Jensen & Berger, 1991; Jensen et al., 1991; Reid et al., 1991). The

changes in skin blood flow which occur during asphyxia appear to be influenced by the

mechanism producing the asphyxia (Jensen et al., 1991; Bobby & Santos, 2000) and also may

vary with the severity of the resulting acidaemia (Cohn et al., 1974). During asphyxia induced

by a reduction in uterine artery blood flow in fetal lambs, fetal skin, carcass and lung blood

flow were decreased whilst blood flow to the myocardium, cerebral cortex and adrenals were

increased (Jensen et al., 1987; Jensen et al., 1991). In contrast, when asphyxia was induced by

partial umbilical cord occlusion in fetal lambs, blood flow to the skin overlying the head and

hindquarters increased three fold and there also was an increase in blood flow to the fetal

carcass which did not occur at the expense of other cardiovascular adaptations to asphyxia

(Itskovitz et al., 1987; Bobby & Santos, 2000). It has been postulated that this seemingly

paradoxical increase in blood flow to the fetal carcass during hypoxaemia induced by reduced

umbilical perfusion may be because the rapid blood flow through the carcass may act as a

fast time-constant vascular bed thus maintaining venous return under conditions where

placental blood flow is restricted, such as may occur with umbilical cord occlusion (Itskovitz

et al., 1987).

In my study fetal scalp blood flow was not measured; however, it can be postulated that the

widening of the scalp-arterial lactate level difference during asphyxia may be secondary to an

increase in scalp blood flow similar to the three-fold increase in scalp blood flow previously

described with 90 minutes of partial cord occlusion (Bobby & Santos, 2000). To understand

how improved scalp blood flow reduces scalp lactate levels during asphyxia it is important to

remember that the fetal scalp blood collected in this study (and also in human fetuses in

labour) is actually capillary blood. Therefore, similar to jugular venous lactate levels which

are influenced by carotid arterial lactate levels, cerebral blood flow and brain metabolism,

fetal scalp capillary lactate levels are influenced by carotid arterial lactate levels, scalp blood

flow rate and scalp metabolism (Figure 4.16). As local lactate levels are determined by the

balance between tissue lactate production and the overall capacity to clear lactate from the

tissue (Perret & Enrico, 1978) it becomes apparent that increased scalp blood flow during

asphyxia may be responsible for the lower levels obtained in scalp capillary blood than arterial

blood.

206

The widening of the scalp-arterial lactate difference with increasing arterial lactate levels may

be because arterial lactate levels are an indirect marker of the severity of asphyxia. As the

severity of the asphyxial insult increases, the magnitude of the cardiovascular responses to

maintain cardiac output and redirect blood flow to vital organs have been shown to increase

(Jensen et al., 1991). Therefore, in an experimental model using repeated UCO to induce

asphyxia, as the severity of the asphyxia increases the fetal scalp blood flow may continue to

increase which potentially reduces fetal scalp capillary blood lactate levels.

The narrowing of the scalp-arterial lactate level difference in response to increasing

adrenaline levels is not surprising if one considers the effect of adrenaline on fetal circulation.

Adrenaline is released from the adrenal medulla, partly as a result of increased sympathetic

outflow, but also because the fetal adrenal gland is directly responsive to falls in PO2 (Hanson

& Kisured, 2001; Wood, 2001) Adrenaline is a potent vasoconstrictor which has been shown

to decrease scalp blood flow in sheep during asphyxia (Jensen et al., 1987) and therefore may

reduce the clearance of lactate from fetal scalp tissues resulting in higher capillary lactate

levels.

4.4.2.2 Information gained from spontaneously acidaemic fetuses

Recent evidence suggests that ovine fetuses which are spontaneously acidaemic have altered

cardiovascular, endocrine and metabolic responses to acute hypoxaemia (Gardner et al.,

2002). In the studies performed in this chapter I have confirmed that spontaneously

acidaemic fetuses have altered pH, base excess, PO2 and glucose responses to asphyxia whilst

their lactate, cortisol and catecholamine responses were not different from non-acidaemic

fetuses exposed to an acute asphyxial challenge. Rather than exclude these fetuses from

analysis, I have considered them as a subgroup because they have the ability to provide

potentially useful information about the fetal response to asphyxia.

The spontaneously acidaemic fetuses provided evidence to suggest that fetal scalp lactate

levels in previously compromised fetuses accurately reflect carotid arterial lactate levels rather

than underestimating lactate levels as demonstrated in the healthy fetuses exposed to

asphyxia. The spontaneously acidaemic fetuses also provided evidence to suggest that scalp

blood lactate levels reliably represents jugular venous lactate levels at rest and during asphyxia

in fetuses and previously compromised fetuses.

It has been suggested that fetal metabolism of glucose and lactate during acute hypoxaemia is

altered in fetuses exposed to an adverse intrauterine environment prior to the acute challenge

(Gardner et al., 2003). My study provided further evidence in support of this hypothesis

207

because the spontaneously acidaemic fetuses displayed different cranial metabolism of lactate

and different glucose responses than non-acidaemic fetuses exposed to the same asphyxial

challenge. Considering arterio-venous lactate differences across the fetal cranial structures,

the normal fetus exposed to an asphyxial insult appeared to metabolise lactate as a source of

fuel during asphyxia whereas the spontaneously acidaemic fetus exposed to the same

challenge produced lactate. Both human and sheep fetuses are capable of using lactate as a

source of metabolic fuel (Fowden, 1994). In fact, in vitro studies have demonstrated that

neurones can carry out synaptic function with lactate as the only available carbon source

(Schurr et al., 1997). The altered lactate metabolism in the cranial structures of the

spontaneously acidaemic fetuses exposed to asphyxia may reflect attempts by the fetus to

mobilise an additional source of energy for the fetal brain during asphyxia because these

fetuses did not mount a hyperglycaemic response to the asphyxial insult which would have

provided energy to the fetal brain during asphyxia. There is evidence from rat models of

hypoxic-ischaemic brain injury that an increase in lactate output by brain tissue during

hypoxia-ischaemia serves to meet the energy needs of glutamate-activated neurons and may

be important to neuronal survival after asphyxia (Schurr & Rigor, 1998; Schurr et al., 1999).

Care is required in interpreting data derived from spontaneously acidaemic fetuses because

the cause of the specific reduction in fetal arterial pH and base excess despite appropriate

lactate concentrations is not known. No differences between the spontaneously acidaemic

fetuses and normal fetuses were observed in maternal condition, maternal cardiovascular

stability during anaesthesia, surgical duration or complications, or fetal biochemical status

prior to the experimental phase. Although the spontaneously acidaemic subgroup may have

developed during the instrumentation phase of the fetal preparation used in this study, the

significant alterations observed in fetal biochemical, endocrine and metabolic responses

during asphyxia seen in this subgroup suggest that these fetuses may have been inherently

different prior to the commencement of the fetal instrumentation. Arterial pH has been

shown to change independently of the concentration of circulating metabolic acids (Hooper

et al., 1990; Gardner et al., 2001a, b), and it has been postulated that this response may be

related to changes in specific metabolically-active hormone concentrations, in particular

catecholamines and prostaglandin E2 (PGE2; Hooper et al, 1990). Although my study

suggests that there were no catecholamine differences between the two groups, a role for

changes in PGE2 concentration or the placental and/or renal handling of acid cannot be

ruled out.

208

4.4.3 Fetal scalp blood lactate and lipid peroxidation in the fetal brain

Although understanding the relationship between fetal scalp blood and central blood lactate

levels is important prior to the potential introduction of fetal scalp lactate measurement into

widespread clinical practice, the most important relationship that requires investigation is that

between fetal scalp blood lactate levels and fetal brain injury. No previous studies have

investigated the relationship between fetal scalp blood biochemical measurements and fetal

brain injury. This is probably because experimental models which afford regular, reliable

collection of adequate volumes of fetal scalp blood are not conducive to waiting three to

seven days after the insult for histopathologic brain injury to develop.

This study is the first to attempt to investigate the relationship between fetal scalp blood

biochemical measurements and potential brain injury through investigating the relationship

between fetal scalp blood lactate levels and oxygen free radical activity in the fetal brain. I

have shown during repeated umbilical cord occlusion there is evidence of increased lipid

peroxidation in the brain perfusate and that lactate levels are better predictors of the presence

of lipid peroxidation in the brain perfusate than routine acid-base analysis. Of particular

note, scalp lactate predicts oxygen free radical activity in the fetal brain at least as well as

carotid arterial lactate levels; perhaps slightly better.

4.4.3.1 Markers of lipid peroxidation

The two markers of lipid peroxidation measured in this study were organic hydroperoxides

(OHP) and malondialdehyde (MDA). Organic hydroperoxides are relatively stable (half-life

approximately three hours) products of free radical attack on lipids contained in cell

membranes (Ward & Peters, 1995; Wang et al., 1996) which appear to have limited transfer

across the human placenta (Rogers et al., 1999). Malondialdehyde is a by-product of lipid

peroxidation due to di-ene conjugation, which is itself responsible for some of the damaging

effects of free radicals on DNA and cell membranes (Ward & Peters, 1995). Unlike OHP,

MDA is thought to be actively transported across the placenta from the fetus to the mother

(Rogers et al., 1999) which would provide one mechanism to protect the fetus from the

known toxic effects of MDA on cellular metabolism (Gutteridge & Halliwell, 1990).

Malondialdehyde, as measured in this study by thiobarbituric acid adduction, includes not

only MDA produced by metabolism of OHP in vivo, but also that produced from

breakdown of OHP and endoperoxides in vitro (Gutteridge & Halliwell, 1990). Therefore,

of the two techniques to measure lipid peroxidation, OHP is the more robust measurement

for assessing fetal response to asphyxia (Rogers et al., 1997; Rogers et al., 1999).

209

4.4.3.2 Patterns of lipid peroxidation during asphyxia

In this study I have shown a 260% increase in jugular venous levels of OHP and a 125%

increase in jugular venous levels of MDA in the fetuses exposed to repeated umbilical cord

occlusion. These results are similar to previous studies using similar cord occlusion protocols

in instrumented ovine fetuses (Rogers et al., 2001). It is likely that the increased lipid

peroxidation that appears evident in my study is secondary to the repeated hypoxia-

ischaemia-reperfusion that occurs with repeated umbilical cord occlusions. Similar dramatic

elevations of markers of lipid peroxide have previously been described in repeated ischaemia-

reperfusion models in rabbits and rats (Tan et al., 1998; Al Nita et al., 2001). During umbilical

cord occlusion, fetal hypoxia leads to depletion of the adenosine triphosphate (ATP) and an

increase in its degradation products (Harkness et al., 1982). Fetal hypoxia also results in the

accumulation of lactate and the conversion of xanthine dehydrogenase to xanthine oxidase

(Figure 4.17). During reoxygenation and reperfusion between umbilical cord occlusions,

xanthine oxidase catalysed by calcium causes the breakdown of hypoxanthine and its

derivatives generating oxygen free radicals. It is thought that with re-oxygenation,

metabolism of polyunsaturated fatty acids via the cyclooxygenase pathway and adenine

nucleotides via the xanthine oxidase pathway (Figure 4.17), are the most likely sources of

oxygen radicals responsible for subsequent lipid peroxidation (Fridovich, 1978).

It is likely that the OHP measured in my experiments is derived from vascular endothelial

cells in the brain and some from lipids in the plasma rather than directly from neuronal cells

themselves. However, neuronal and vascular endothelial cells are almost certainly subject to

similar degrees of oxidative stress to the blood perfusing the organ. In controls and during

the 0.5-minute cord occlusions, jugular venous OHP levels were consistently lower than the

matched carotid arterial samples, suggesting somatic rather than neural production of OHP

with scavenging and metabolism of peroxides by the brain, possibly by copper-zinc-

superoxide dismutase and anti-apoptotic protein Bcl-2 (Chan, 1996). The absence of an

increase in venous OHP levels during the 0.5-minute cord occlusions may be because the

occlusions were of insufficient duration to produce significant changes in oxygenation to the

fetal brain because of cardiovascular shunting of oxygenated blood to the fetal brain in

response to hypoxia. This hypothesis is supported by the receiver operator curve analysis of

lipid peroxidation in the brain perfusate which suggested that lipid peroxidation does not

reliably start in the fetal brain until oxygen tensions fall below 14mmHg. It has previously

been shown that arterial PO2 levels fall approximately 30% during 1.0-minute umbilical cord

occlusions to levels in the order of 14mmHg (Richardson et al., 1996). Therefore, it is

probable that during 0.5-minute cord occlusions, small reductions in carotid arterial PO2

210

Oxygen FreeRadicals

Oxygen FreeRadicals

CordOcclusion

ONHypoxia

XDH XO LactateAccumulationATP Hydrolysis

Hypoxanthine

Xanthine

CordOcclusion

OFF

Urate

O2

XO

Figure 4.17 Mechanism of oxygen free radical production with umbilical cord occlusion.

Cord occlusion leads to fetal hypoxia which results in the accumulation of lactate, the

conversion of xanthine dehydrogenase (XDH) to xanthine oxidase (XO) and accumulation

of hypoxanthine due to depletion of adenosine triphosphate (ATP). When oxygen is re-

introduced during intervals between cord occlusions, XO catalysed by calcium causes the

breakdown of hypoxanthine in two steps, each generating oxygen free radicals.

211

occur, but oxygen levels in the fetal brain are not low enough to stimulate lipid peroxidation.

Following from this, during 1.0-minute cord occlusions one would expect that oxygen levels

in the fetal brain would reach levels that would stimulate lipid peroxidation at least during

cord occlusion. This in fact is what occurred in this study. The absence of an increase in

venous OHP during the 0.5-minute cord occlusions may also have occurred because the

antioxidant defences in the fetal brain were not overwhelmed by the free radical production

in the early phase of the experimental protocol. Somatic OHP production during the 0.5-

minute cord occlusions may be secondary to hypoperfusion in peripheral, renal and

gastrointestinal circulations that occurs in response to cardiovascular adaptation to asphyxia

induced by umbilical cord occlusion (Itskovitz et al., 1987; Bobby & Santos, 2000).On

increasing the duration of cord occlusion to 1.0-minute and decreasing the inter-occlusion

recovery periods, the jugular venous OHP levels increased 260%. It is reasonable to assume

that by this stage of the experiment plasma low density lipoproteins were fully oxidised and

the predominate source of OHP release would be from vascular endothelial cells. During the

1.0-minute cord occlusions fetal carotid arterial OHP concentrations started to decrease. This

may have been due to severe restriction of oxygen supply to somatic tissues causing either a

decrease in peroxidation or a shutdown of these somatic vascular beds, or both. On further

increasing the duration of cord occlusions to 1.5-minutes, OHP production started to

plateau. The decrease in peroxidation may have been due to severe restriction of oxygen

supply in the fetal neural circulation by this stage of the experiment. Alternatively, there may

be a maximum rate of release of peroxides from endothelial cell membranes into the

circulation which was reached during the 1.0-minute occlusions.

A similar pattern of changes occurred in the MDA levels during this experiment; however,

the increase in jugular venous MDA levels did not occur until during the 1.5-minute cord

occlusions. The delayed response in MDA compared to OHP is not surprising as OHPs are

the primary molecular products of oxidative stress while MDA is one of the end products of

peroxidative cleavage of fatty acids. Malondialdehyde production involves the formation of

hydroperoxides, cleavage to yield hydro-peroxy-aldehydes, and finally, di-ene conjugation via

a second scission and dismutation of lipid peroxyl radicals (Gutteridge, 1988).

Malondialdehyde is also less stable than OHP, and given its particularly toxic effects on

cellular metabolism, the fetal brain may have better MDA scavenging mechanisms than exists

for OHP.

212

4.4.3.3 The relationship between fetal scalp lactate and lipid peroxidation in the

fetal brain

In order to reflect oxidative stress in an organ, significantly higher levels of lipid peroxidation

must be detected in the venous drainage than found in the blood perfusing the organ (i.e. a

positive venous-arterial difference must exist). In this study I have shown that both fetal

scalp lactate levels and carotid arterial lactate levels are significantly better predictors of lipid

peroxidation in the fetal brain than acid-base assessment. This observation has not

previously been reported; however, it is consistent with the experimental findings presented

in chapter three and also with the current understanding of the mechanisms of fetal hypoxic-

ischaemic brain injury.

The fetal biochemical responses observed in the normal and spontaneously acidaemic fetuses

in response to asphyxia in this study may aid in unravelling the explanation of the potential

mechanisms responsible for the ability of fetal lactate levels to predict lipid peroxidation in

the fetal brain. The normal and spontaneously acidaemic fetuses in this study had the same

changes in lipid peroxidation in response to repeated UCO despite their different responses

in arterial pH, PO2, base excess and glucose. The two groups of fetuses also had the same

responses in arterial lactate levels, catecholamines and cortisol. These observations suggest

that changes in lactate, catecholamine and lipid peroxidation relate more to the insult than to

fetal status prior to acute asphyxia. In addition, similar findings in the two study groups

suggest that the common link may be perfusion and control of fetal blood pressure.

I propose that the ability of fetal lactate levels to predict lipid peroxidation in the fetal brain

during asphyxia is secondary to the ability of lactate to predict blood pressure instability

during umbilical cord occlusion and cerebral hypo-perfusion during asphyxia. In Chapter

Three evidence was presented that fetal lactate levels were better predictors of fetal blood

pressure instability during cord occlusion than other markers of fetal acid-base status. This is

particularly relevant because the duration of fetal blood pressure instability during umbilical

cord occlusion accounted for 91% of the variability in grading of histopathologic brain injury

72 hours after the same cord occlusion protocol as used in this study. In addition, there is

substantial evidence from adult critical care medicine that lactate levels are better predictors

of tissue hypoperfusion than pH and base excess (Perret & Enrico, 1978; Bakker et al., 1996;

Husain et al., 2003). As our understanding of the mechanisms responsible for hypoxic-

ischaemic brain injury has improved over the last 30 years, it has become increasingly

apparent that it is the severity of the changes in fetal blood pressure and cerebral blood flow

during asphyxia which predict fetal brain injury rather than acid-base and blood gas

213

alterations (Myers, 1972; Ting et al., 1983; Gunn et al., 1992; Mallard et al., 1992; de Haan et al.,

1993; de Haan et al., 1997b; Ikeda et al., 1998a; Fujii et al., 2003).

The limited ability of routine acid-base assessment to predict lipid peroxidation in the fetal

brain is consistent with previous studies showing that the process of raised lipid peroxidation

products derived from reactive oxygen species may occur independently of changes in acid-

base status (Mongelli et al., 1997; Rogers et al., 1999). It is also consistent with animal (Myers,

1972; Ting et al., 1983; Gunn et al., 1992; Mallard et al., 1992; Ikeda et al., 1998a) and human

studies (Winkler et al., 1991; Low et al., 1994; Kruger et al., 1999) which suggest that fetal acid-

base status is a poor predictor of the severity of neuronal injury after asphyxia.

In human labour, the fetal scalp is the window for biochemical assessment of fetal wellbeing.

It was reassuring to find in this study that fetal scalp lactate was at least as good at predicting

lipid peroxidation in the fetal brain perfusate as carotid arterial lactate levels; perhaps even

better.

4.4.4 Future studies

The studies presented in this chapter I have significantly increased the understanding of the

relationship between fetal scalp blood lactate levels and central blood lactate levels during

asphyxia. I have not addressed the issue of the effect of fetal scalp oedema or caput

succedaneum on the reliability of fetal scalp lactate assessment, nor have I assessed the effect

of fetal head compression or moulding. The effect of caput on fetal scalp blood pH

assessment varies between studies. Some studies suggest that scalp oedema and caput can

result in scalp blood pH levels being up to 0.06 pH units lower than carotid arterial or

umbilical arterial blood levels in as many as 10-20% of samples (Bowe et al., 1970; Parer &

Livingston, 1990; Greene, 1999). Others authors have found that caput does not alter the

accuracy of fetal scalp blood pH assessment with no difference in the haematocrit between

scalp blood samples collected vaginally and those collected from the carotid artery and

jugular vein via fetal catheters (Adamsons et al., 1970). Fetal head compression and moulding

have not been shown to alter the accuracy of fetal scalp blood pH assessment (Adamsons et

al., 1970). Although it is important to understand the effect of caput succedaneum, scalp

oedema and moulding on the accuracy of fetal lactate assessment, these factors did not

appear to adversely affect the ability of fetal scalp blood lactate levels to predict low Apgar

scores at five minutes or moderate-severe hypoxic-ischaemic encephalopathy in humans

(Kruger et al., 1999).

214

Fetal scalp pH measurements were not collected in this study to compare directly with fetal

scalp blood lactate levels because the blood gas machine in the research laboratory at the time

these studies were performed required 80µL of blood to perform a pH assessment in

contrast to the 15µL required to analyse lactate on the Accutrend® handheld lactate meter.

Although it is possible to collect large volumes of blood from the fetal scalp, I chose not to

attempt this to optimise the quality of the results obtained in this study. The collection of

large volumes of blood from the fetal scalp takes time and it has previously been shown that

the longer scalp blood is exposed to air, the less accurate the result (Zernickow, 1966). In

addition, the repeated nature of the sampling required in this study protocol would have

compromised the stability of the preparation as a substantial amount of time would have

been devoted to attempting to collect adequate volumes of blood rather than maintaining

thermal stability of the preparation and minimising fetal stimulation and manipulation. These

issues are particularly important given that one of the primary aims of the studies in this

chapter was to validate the model utilised for collecting fetal scalp samples.

This study raised the issue that lactate metabolism may be altered in the brains of fetuses in

an adverse uterine environment. This is potentially important given that many complications

of human pregnancy are associated with an adverse uterine environment. Limited

conclusions regarding cerebral metabolism can be made using only carotid artery and jugular

venous blood. More detailed evaluation of this area would require blood sampling from the

superior sagittal sinus and assessment of cerebral blood flow.

215

4.5 Conclusion

In this chapter, three important issues relating to the collection of fetal scalp blood lactate

levels rather than central lactate levels during intrapartum asphyxia have been addressed.

The first aim was to compare the experimental model used to collect fetal scalp blood for

lactate assessment during asphyxia with the chronically instrumented model described in

Chapter Three. After considering the fetal biochemical, endocrine and cardiovascular

responses to asphyxia, the acutely instrumented fetal preparation used in this study did not

appear to impair the ability of the fetal lamb to coordinate neural, endocrine and local tissue

responses to asphyxia. Specifically, there was no evidence identified that the selected

maternal anaesthesia adversely affected autonomic control of the fetal heart during asphyxia

or the responsiveness of the FHR to hypoxaemia. I propose that the experimental model

chosen in this study is currently the best that is available to assess fetal scalp blood in a

regular, reliable, repeated manner during asphyxia, allowing collection of adequate volumes of

fetal scalp blood for biochemical assessment.

The second aim of this study was to evaluate the relationship between fetal scalp blood

lactate and central blood lactate levels during asphyxia. I have shown that in the absence of

asphyxia, scalp blood lactate levels are not significantly different from carotid arterial lactate

levels; however, when the fetus is exposed to asphyxia, fetal scalp blood lactate levels were

found to underestimate the carotid arterial lactate levels by a mean of 0.6mmol/L. Although

the mechanism responsible for scalp lactate underestimating central lactate during asphyxia is

unknown, I have provided evidence to suggest the underestimate during asphyxia may be due

to the cardiovascular changes which occur in response to umbilical cord occlusion and the

influence of adrenaline on the fetal scalp. In this study fetal scalp blood lactate levels were

found to be not significantly different from jugular venous lactate levels and this relationship

was not influenced by fetal asphyxia or fetal condition prior to asphyxia. This observation

may be advantageous because the composition of jugular venous blood is determined not

only by the composition of arterial blood, but also by the relationship between flow rate and

the metabolic needs of the fetal brain.

The final aim of this chapter was to investigate the relationship between fetal scalp lactate

levels and an indirect marker of fetal brain injury, lipid peroxidation. I have shown that fetal

scalp lactate levels are significantly better than carotid arterial pH, PO2, and base excess at

predicting lipid peroxidation in the fetal brain during asphyxia. The potential mechanism

responsible for the improved ability of lactate to predict lipid peroxidation may be secondary

216

to the ability of lactate levels to predict fetal blood pressure instability during umbilical cord

occlusion and tissue hypoperfusion during asphyxia.

217

C h a p t e r 5

FETAL LACTATE RESPONSES IN THE GROWTH RESTRICTED

OVINE FETUS DURING ASPHYXIA INDUCED BY GRADED

CORD OCCLUSION

After all, concentration is the price the modern student pays for success.

Thoroughness is the most difficult habit to acquire,

but it is the pearl of great price,

worth all the worry and trouble of the search.

Sir William Osler (1892)

5.1 Introduction

The motivation to perform the studies presented in this chapter arose from the observation

in both Chapters Three and Four that an adverse intrauterine environment, as demonstrated

by high fetal lactate levels or a spontaneous mild acidaemia prior to the experimental

protocol, altered cardiovascular, endocrine and metabolic responses to acute hypoxaemia in

the ovine fetus. The study subgroups in Chapters Three and Four occurred spontaneously

and their aetiology is uncertain. This limits the interpretation of the results relating to these

subgroups; however, it does suggest that formal investigation of lactate responses during

asphyxia in fetuses with an adverse intrauterine environment should be considered as part of

the evaluation of the role of lactate in intrapartum assessment of fetal wellbeing.

There are many potential causes of an adverse intrauterine environment in human pregnancy.

These include uteroplacental insufficiency, diabetes mellitus, preterm premature rupture of

the fetal membranes, intrauterine infection, maternal medical conditions, and exposure to

maternal medications. In the studies presented in this chapter, I have specifically chosen to

investigate the effect of the adverse intrauterine environment associated with growth

restriction due to uteroplacental insufficiency.

The intrauterine environment of the growth restricted fetus is significantly different from that

of appropriately grown fetuses. Growth restricted human fetuses have been shown to be

218

hypoxaemic, hypercapnic, hyperlactaemic, and acidaemic when compared to appropriately

grown gestational age-matched controls (Nicolaides et al., 1989; Marconi et al., 1990). It has

also been shown that growth restricted fetuses have altered lactate metabolism (Marconi et al.,

1990), glucose metabolism (Jansson et al., 1993; Marconi et al., 1996; Pardi et al., 2002), amino

acid metabolism (Cetin et al., 1996; Marconi et al., 1999) and fatty acid metabolism (Cetin et

al., 2002) when compared to appropriately grown controls.

Fetal growth restriction occurs in between three and five percent of pregnancies. These

fetuses are over-represented in morbidity and mortality data, with infants born between 38

and 42 weeks gestation, whose weights are between 1500g and 2500g, having perinatal

morbidity and mortality rates 3-30 times those of infants whose weights are between the 10th

and 90th centile (Williams et al., 1982; Dashe et al., 2000).

There is a growing body of evidence to suggest that growth restricted fetuses are at increased

risk of adverse events during labour (Low et al., 1976; Low et al., 1981; Nieto et al., 1994;

Dashe et al., 2000) and, more importantly, they are at increased risk of adverse neurologic

outcomes related to acute intrapartum hypoxic events than appropriately grown fetuses

(Bukowski et al., 2003). Observational studies in humans have shown a significant inverse

relationship between birthweight percentile and frequency of abnormal fetal heart rate

patterns such as total and late decelerations (Low et al., 1976; Dashe et al., 2000). Similarly,

growth restricted fetuses have been shown to commence labour with slightly lower scalp pH

levels and also to have more rapid decreases in scalp pH levels during labour than normally

grown fetuses (Nieto et al., 1994). As the birthweight centile falls, the incidence of metabolic

acidaemia in umbilical arterial blood samples increases. Infants with birthweight less than the

3rd centile are three times more likely to have a metabolic acidaemia than those born with

weights in the 50th centile (Low et al., 1981).

Growth restricted human fetuses are at increased risk of neonatal encephalopathy and

cerebral palsy when compared with fetuses demonstrating normal growth (Blair & Stanley,

1990, 1992; Badawi et al., 1998a; Bukowski et al., 2003). Although there are data to suggest

that a proportion of cases of neonatal encephalopathy have an antenatal origin (Badawi et al.,

1998a) recent evidence suggests that fetuses with reduced growth potential are 20 times more

likely to develop neonatal encephalopathy secondary to an acute intrapartum hypoxic event

than normally grown fetuses (Bukowski et al., 2003). A consistent relationship has been

shown between the risk of cerebral palsy and birth weight centile (Stanley et al., 2000) with

22% of cases of spastic cerebral palsy occurring in babies born below the tenth centile (Blair

& Stanley, 1990). Fifty percent of cases of cerebral palsy in infants born between 34 and 37

219

weeks occur in infants with birth weights below the 10th centile (Rantakallio, 1985). Although

the majority of cases of cerebral palsy are thought to have an antenatal origin, between 5 and

28% of cases of cerebral palsy can be attributed to intrapartum events (Nelson, 1988;

MacLennan, 1999; Hagberg et al., 2001). It is not known whether fetuses with reduced

growth potential are more likely to develop cerebral palsy secondary to acute intrapartum

hypoxic events than fetuses with normal growth potential; however, the relationship between

growth restriction, acute intrapartum hypoxic events, and neonatal encephalopathy would

suggest that growth restricted fetuses may also be at an increased risk of cerebral palsy

attributable to intrapartum events. Therefore, considering the evidence that growth restricted

fetuses are more likely to experience adverse events in labour than normally grown fetuses, I

believe that it is important to explore any new technique proposed to assess fetal wellbeing in

labour in this important group of fetuses prior to its potential introduction into clinical

practice.

Despite the known importance of fetal growth disorders, there have been no published

experimental studies using animal models in which fetal heart rate patterns, acid-base, lactate

and endocrine responses to asphyxia have been investigated using a standardised intervention

in fetuses in which growth has previously been restricted. The studies presented in this

chapter were designed to investigate the lactate responses during asphyxia in growth

restricted fetuses. The experimental model was specifically designed to investigate central

and scalp lactate levels during asphyxia. In addition to characterising the biochemical and

physiologic responses of the growth restricted fetus during repeated UCO, I have also

investigated the ability of FHR patterns, pH and lactate levels to predict fetal arterial pressure

responses during cord occlusion, the factor previously identified (Section 3.3.5.3.3) as the

single best predictor of severity of fetal brain injury after repeated umbilical cord occlusion.

In the studies presented in this chapter, I used repeated umbilicoplacental embolisation to

induce fetal growth restriction because it reproduces many features seen in human growth

restricted fetuses (Nicolaides et al., 1989) and it has been widely used to study fetal endocrine,

renal, metabolic and cardiovascular responses to chronic placental insufficiency (Clapp et al.,

1981; Trudinger et al., 1987; Gagnon et al., 1995; Gagnon et al., 1996; Cock & Harding, 1997).

Umbilicoplacental embolisation has been shown to induce asymmetric growth restriction

and it reproduces the fetal acid-base and lactate changes seen in growth restricted human

fetuses (Nicolaides et al., 1989). In addition, this model of growth restriction has been shown

to result in high systolic-diastolic ratios in umbilical artery Doppler flow velocity waveform

analysis (Trudinger et al., 1987; Morrow et al., 1989; Gagnon et al., 1994) similar to those

220

which occur in growth restricted human fetuses (Giles et al., 1985; Trudinger et al., 1985;

McCowan et al., 1987). Umbilicoplacental embolisation has also been shown to result in fetal

cerebral white matter injury (Mallard et al., 1998; Duncan et al., 2000) similar to that which can

occur in severely growth restricted human fetuses (Leviton & Paneth, 1990; Gaffney et al.,

1994c). After asymmetric growth restriction had been induced, fetuses were acutely

instrumented using the protocol described in Chapter Four to enable the assessment of both

carotid arterial and scalp blood lactate levels during asphyxia induced by repeated umbilical

cord occlusion.

5.1.1 Aim

There are four specific aims for the studies presented in this chapter:

1. to characterise the physiologic responses of growth restricted fetuses to repeated

UCO using a pattern designed to mimic events which may occur in labour,

2. to investigate the fetal lactate responses during UCO in the growth restricted

fetus,

3. to investigate the relationship between fetal scalp, carotid arterial and jugular

venous blood lactate levels both at rest and during asphyxia in the growth

restricted fetus, and

4. to investigate the ability of fetal scalp blood lactate levels to predict fetal arterial

pressure changes during asphyxia in the growth restricted fetus.

5.1.2 Hypothesis

I hypothesise that:

1. growth restricted ovine fetuses exposed to UCO will have more rapid

deterioration in their cardiovascular, biochemical and endocrine responses during

asphyxia than normally grown fetuses,

2. lactate responses in growth restricted fetuses will be similar to those in normally

grown fetuses exposed to the same asphyxial challenge,

3. fetal scalp lactate levels will reliably predict jugular venous lactate levels in growth

restricted fetuses and growth restriction will alter the relationship between fetal

scalp blood lactate levels and carotid arterial lactate levels, and

4. the ability of fetal lactate levels to predict fetal arterial pressure responses during

UCO will not be altered in growth restricted fetuses.

221

5.2 Methods

All experimental procedures were approved by the Animal Experimentation Ethics

Committee of The University of Western Australia. Merino ewes with singleton pregnancies

were transported to the research facility at 96 days gestational age to allow for acclimatisation

to the facility. The experimental protocol utilised in this study had three consecutive stages

which are outlined in Figure 5.1.

5.2.1 First surgical preparation

Twenty four ewes underwent surgery at 101±1 days (mean±SEM) of pregnancy.

Anaesthesia was induced with Thiopentone 4mg/kg in 15mL saline solution (Abbott

Australasia Pty Ltd, Cronulla, NSW, Australia), ewes were intubated, and anaesthesia was

maintained with 0.75-1.25% Isoflurane (Abbott Australasia Pty Ltd, Cronulla, NSW,

Australia) in 100% oxygen delivered by positive pressure ventilation with ventilator settings

selected to maintain maternal blood gases in the normal range.

The uterus was exposed through a paramedian incision in the anterior abdominal wall and a

fetal hind limb was manipulated to an avascular region of a uterine horn. A small

hysterotomy was performed with care taken to avoid loss of amniotic fluid. Using an

established procedure (Trudinger et al., 1987; Gagnon et al., 1995; Cock & Harding, 1997;

Murotsuki et al., 1997a; Mallard et al., 1998; Duncan et al., 2000), a 1.5mm polyvinylchloride

catheter was implanted into the fetal descending abdominal aorta via a femoral artery, such

that the tip lay 1-2cm below the renal arteries and above the common umbilical artery

(approximately 7cm). This catheter subsequently enabled fetal blood sampling and injection

of microspheres into the umbilicoplacental circulation. After replacement of the hind limb, a

1-mL injection of Benacillin (procaine penicillin, 150mg/mL; benzathine penicillin,

150mg/mL; procaine hydrochloride, 20mg/mL; Troy Laboratories, Smithfield, Australia)

was injected into the amniotic fluid. The uterus was closed and the catheter was tunneled

subcutaneously to the ewe’s flank. The maternal abdomen was closed and 3mL

intramuscular Benacillin was given to the ewe at completion of the surgery. After surgery,

sheep were held in individual pens with access to food and water and catheter patency was

maintained by daily flushing with heparinised saline (2IU/1mL 0.9% NaCl).

222

Figure 5.1 Outline of experimental protocol. Twenty four ovine fetuses were instrumented

at 101 days to allow umbilicoplacental embolisation for 23 days. At 129 days the fetuses were

acutely instrumented to allow the collection of both central and peripheral blood samples

during asphyxia induced by repeated umbilical cord occlusion. Emb = embolised, Non-Emb

= non-embolised, Occl = cord occlusion protocol, Non-Occl = instrumented but no cord

occlusion.

Instrumented101 days

n = 24

UmbilicoplacentalEmbolisation

23 daysn = 12

Non-Embolised23 daysn = 12

CordOcclusion129 days

n = 7

Control129 days

n = 5

CordOcclusion129 days

n = 7

Control129 days

n = 5

Emb / Occl Emb / Non-Occl

Non-Emb /Occl

Non-Emb /Non-Occl

5 dayRecovery

Em

bolis

atio

nP

hase

Exp

erim

enta

lP

hase

223

5.2.2 Embolisation protocol

After a five day recovery period (106±1days gestational age), each ewe was allocated at

random to one of two treatment groups: embolised (n=12) or non-embolised (n=12; Figure

5.1). In both groups, fetal blood was sampled daily (0.3mL) for measurement of arterial gas

tensions and pH (Rapidlab 228, Bayer Diagnostics, Tarrytown NY) and lactate (Accutrend®

Lactate, Roche Diagnostics GmbH, Mannheim, Germany). For 23±1 days, the fetuses

allocated to the embolisation group were given daily aortic injections of nonsoluble,

nonradiolabeled, mucosaccharide microspheres (Sephadex Superfine G-25, 40-70µm,

Pharmacia, Sweden) suspended in a 1%wt/vol solution of sterile isotonic saline with 0.02%

Tween 80 (≈1 million spheres/mL). Once per day repeated injections of microspheres (0.05-

0.20 million) were made every 20 minutes until the fetal arterial PO2 levels were ≈8mmHg or

more below pretreatment levels (daily total dose 0.05 to 2.00X106 spheres). The catheters of

fetuses in the non-embolised group were sampled and flushed daily, but no microspheres

were injected. Arterial blood samples (4.0mL) were collected from both groups at the

beginning and end of the embolisation phase for assessment of fetal catecholamine,

adrenocorticotropic hormone (ACTH), cortisol and glucose concentrations. Fetal blood was

centrifuged at 1500g for 10 minutes at 4°C. Plasma was removed and stored at -80°C until

analysed.

5.2.3 Second surgical preparation

At 129±1 days gestational age, all ewes were acutely instrumented using the experimental

protocol described in Chapter Four. Maternal anaesthesia, arterial pressure monitoring and

biochemical monitoring were the same as described in Section 4.1. As described previously,

great care was taken to minimise the use of anaesthetic agents and to maintain stability of

maternal arterial pressure and acid-base status during the procedure.

The fetuses were instrumented as described in Section 4.2.1 to allow collection of carotid

arterial blood samples, jugular venous blood samples, fetal electrocardiogram (ECG) and

fetal arterial pressure. To assess lactate metabolism in the fetal brain an additional catheter

was inserted into the superior sagittal sinus to allow collection of blood directly flowing from

the fetal brain in three fetuses of each of the four experimental groups. A vascular snare and

a circular three-crystal ultrasound cuff (10MHz Triton Technology Inc., USA) were applied

around the umbilical vessels (Section 4.2.1) to enable confirmation of normal umbilical blood

flow and complete obstruction of umbilical blood flow during cord occlusion. The fetal

scalp was shaved and the fetus was returned to the uterus with its head externalised to

224

facilitate repeated collection of fetal scalp blood samples. Care was taken to prevent loss of

amniotic fluid and thermal stability was maintained with overhead radiant warmers and

bubble wrap covering the fetal head between scalp blood samples to minimise heat and water

loss.

After a thirty minute recovery phase, baseline fetal carotid blood samples (0.3mL) were

collected for baseline measurement of arterial gas tensions and pH. Repeated fetal arterial

samples (0.3mL) were collected after a further 30 minute rest phase and if the fetal

preparation was stable the cord occlusion protocol was commenced.

5.2.4 Cord occlusion protocol

After confirmation that each fetal preparation was stable, the fetuses entered the

experimental phase of the study. Fetal arterial blood samples (4.0mL) were collected to

measure baseline arterial gas tensions, pH, lactate, glucose, ACTH, cortisol and

catecholamine concentrations. Fetal blood (0.2mL) was collected from the jugular vein to

measure venous pH, gas tensions and lactate levels. Silicone jelly was applied to the fetal

scalp and a scalp blood sample was collected in a heparinised capillary tube (40µL) from a

stab wound for lactate, pH and gas tension measurements. In the subgroup of fetuses

instrumented with a superior sagittal sinus catheter, an additional blood sample (0.2mL) was

collected to measure pH, gas tensions and lactate levels.

Seven fetuses from the embolised group and a further seven fetuses from the non-embolised

group were randomly allocated to the cord occlusion protocol. The remaining five fetuses in

each group acted as controls (Figure 5.1). The umbilical cord occlusion (UCO) protocol

used in this study is identical to that used in Chapters Three and Four (see Section 3.2.2).

Control fetuses underwent identical physiological monitoring and blood sampling as the

fetuses exposed to UCO. After each blood sample was withdrawn, sterile saline was infused

to replace the volume of blood removed for analysis.

Throughout the experimental phase, fetal blood samples were collected every nine minutes.

Carotid arterial blood (0.2mL), jugular venous blood (0.2mL) and fetal scalp blood (40µL)

were collected to measure gas tensions, pH, and lactate levels. In fetuses with superior

sagittal sinus catheters, an additional blood sample (0.2mL) was collected to measure gas

tensions, pH, and lactate levels. Additional arterial samples (1.0mL) were collected at 27 and

54 minutes to measure fetal glucose levels. At the end of the experimental phase, a final

arterial blood sample (4.0mL) was collected to measure glucose, ACTH, cortisol, and

225

catecholamine concentrations. The ewes and fetuses were then humanely killed using a lethal

dose of sodium phenobarbitone and the positions of the implanted catheters were confirmed

and fetal weights measured.

5.2.5 Fetal heart rate, arterial pressure and umbilical blood flow assessment

Fetal ECG and arterial pressure were recorded continuously onto computer at 1 KHz after

amplification (Bio amp, BP transducer MLT380, bridge amplifier ML110, PowerLab 8e,

ADInstruments, Sydney, Australia). Over the 1 minute prior to each cord occlusion, FHR

and FHR variability were assessed to calculate mean fetal heart rate, mean minute range

(MMR), short term variation (STV) and long term variability (LTV) as previously described in

Section 3.2.3. FHR traces were visually assessed to classify the fetal heart rate pattern during

each of the 27 cord occlusions using the classification system described in Section 3.2.3.

Fetal arterial pressure was measured for each heart beat and averaged for the 1 minute

preceding each cord occlusion. The fetal arterial pressure responses during UCO were also

assessed and classified using the arbitrary classification system described in Section 3.3.3.2

and illustrated in Figure 3.14. The ultrasound cuff around the umbilical vessels was linked to

a vascular spectrum analyser (Model SP 25 by Vasculab, Meda Sonics, USA) which enabled

continuous monitoring of the umbilical artery Doppler waveforms to confirm normal

umbilical blood flow between occlusions and complete obstruction of umbilical blood flow

during cord occlusion.

5.2.6 Biochemical methods

Plasma glucose, cortisol, and catecholamine concentrations were measured using the same

techniques as described in Section 3.2.6. Plasma immunoreactive ACTH concentrations

were measured using a commercial radioimmunoassay (Incstar, Stillwater, MN, USA)

validated previously for use in fetal sheep (Norman et al., 1985). The intra-assay coefficient

of variation was 4.5% and the inter assay 4%. The mean assay sensitivity was 15pg/mL.

5.2.7 Statistical methods

Descriptive data are presented as mean and standard error of the mean (SEM) unless

otherwise stated. Univariate analyses of categorical data were performed with Chi-square test

and continuous data were analysed with t-tests or the non-parametric equivalent as

appropriate. Levine’s test of equality of variance was used to compare group differences in

variation about the mean. Time to development of abnormal fetal heart rate patterns was

226

assessed using t-tests and Wilcoxon log rank tests and exact inference due to small sample

sizes.

Changes in fetal biochemical variables and cardiovascular parameters were analysed using

multivariate regression modeling with occlusion, time and experimental group modeled as

fixed effects and individual animals as random effects repeated over time. Regression

modeling was also used to analyse differences between samples collected simultaneously

from the four sites compared in this study matched by time and sheep. Variable selection in

multivariate modeling was based on stepwise procedures (forwards and backwards selection)

including the assessment of interaction terms to arrive at the most parsimonious model.

Comparisons of effect magnitude were based on evaluation of changes in each outcome for

one standard deviation of each independent variable considered.

When investigating the different physiological associations with FHR patterns in both the

embolised and non-embolised fetuses, analyses of FHR patterns consisted either of all 27

occlusion periods (648 observations in total) or were limited to occlusions 1, 2, 26 and 27 in

order to incorporate catecholamine concentrations as candidate covariates (96 observations

in total). Final multivariate models were obtained via backward selection methods following

a careful examination of all two-way interaction effects. Covariate effects for each FHR

pattern considered were presented as odds ratios and 95% confidence intervals (CI). Odds

ratios for continuous measurements were reported for a change of 0.5 standard deviation for

each covariate, relative to no change in this covariate. A positive effect was reported when

the odds ratio for a particular FHR pattern increased as the value of the considered variable

increased. An inverse effect was reported when the odds ratio for a FHR pattern increased

as the value of the considered variable decreased. In cases of wide confidence intervals,

conservative comparisons of effect magnitudes were based on the confidence interval limit

closest to one. Receiver operator curve (ROC) analysis was used to identify optimum cut off

values that maximised sensitivity and specificity for variables found to be associated with

abnormal FHR patterns in both groups of fetuses (Hanley & McNeil, 1983; Beck & Shultz,

1986; DeLong et al., 1988).

Sensitivity and specificity were calculated for the ability of specific FHR patterns to predict

each pattern of fetal arterial pressure response during UCO. ROC analysis was also used to

assess the ability of biochemical measures to predict fetal arterial pressure responses during

UCO and also to identify optimum cut off values that maximised sensitivity and specificity.

Accuracy is reported for the prediction of arterial pressure responses during UCO and is

defined as the proportion of correctly classified cases (true positives + true negatives) using

227

either the specific abnormal FHR pattern or the biochemical cut off point selected to

maximise the sensitivity and specificity (Zweig & Campbell, 1993).

P-values <0.05 were considered statistically significant and p-values <0.10 were considered

marginally significant. SAS statistical software was used for data analysis (Statistical Analysis

Systems Institute Inc. 1999, Version 8.2., Cary, NC).

228

5.3 Results

Results are presented in three sections, each addressing one of the following specific aims:

1. assessment of physiologic responses of the growth restricted fetus during

umbilical cord occlusion,

2. assessment of the relationship between fetal scalp blood lactate levels and central

lactate levels, and

3. the ability of FHR patterns and biochemical markers to predict arterial pressure

responses during UCO.

To aid with comparison between this and previous chapters, similar analyses and graphic

presentations are used. For all graphical representations fetuses exposed to umbilicoplacental

embolisation are represented as solid circles or bars and the non-embolised fetuses are

represented as open circles or bars.

5.3.1 Physiologic responses of the growth restricted fetus during umbilical cord

occlusion

5.3.1.1 Umbilicoplacental embolisation

Umbilicoplacental embolisation produced a 30 percent reduction in fetal arterial PO2 (Figure

5.2). The PO2 levels in the embolised fetuses were significantly lower than in the non-

embolised fetuses by day 5 of the 23±1 day embolisation protocol and remained significantly

lower thereafter (embolised 12.29±0.15mmHg; vs. non-embolised 17.31±0.15mmHg,

p<0.001). Fetal pH was also slightly lower in the embolised fetuses than in the non-

embolised fetuses during this phase (embolised 7.371±0.002; vs. non-embolised

7.392±0.002, p=0.003). Fetal arterial lactate (p=0.370), PCO2 (p=0.487), and base excess

levels (p=0.361) were not significantly different between groups.

At completion of the embolisation phase, the fetuses subjected to repeated umbilicoplacental

embolisation had lower PO2 levels than non-embolised fetuses (12.09±0.64mmHg vs.

17.64±0.73mmHg in non-embolised; p<0.001) and a 2-3 fold increase in adrenaline and

noradrenaline levels (Table 5.1). All other blood gas variables were not significantly different

between groups. Fetal glucose levels were lower in the embolised fetuses

(0.90±0.09mmol/L) than non-embolised fetuses (1.11±0.07mmol/L; p=0.084) but the

differences did not reach statistical significance. ACTH and cortisol levels were not

significantly different between groups. Umbilicoplacental embolisation produced a 29%

229

Figure 5.2 Acid-base changes during 23 days of umbilicoplacental embolisation from 106 to

129 days gestational age. Arterial PO2, pH and lactate concentration (mean ± SEM) for

embolised (E, solid circles, n=12) and non-embolised fetuses (NE, open circles, n=12). Fetal

PO2 was significantly lower in the embolised fetuses by day 5 of the experimental protocol

(p<0.001) and fetal pH was also marginally lower in the embolised fetuses during the

embolisation phase (p=0.003). Fetal lactate levels were not significantly different between

groups (p=0.370).

5

10

15

20

25

p<0.001

NE

E

PO2

PO

2 (m

mH

g)

7.30

7.35

7.40

7.45

NE

E

pH

pH

0 2 4 6 8 10 12 14 16 18 20 220

1

2

3

NE

E

Lactate

Days

Lact

ate

(mm

ol/L

)

230

reduction in body weight (embolised, 2.30±0.15kg vs. non-embolised, 3.22±0.07kg; p<0.001)

at the time of UCO phase.

Table 5.1 Effect of 23 days of umbilicoplacental embolisation on fetal status

Non-Embolised Embolised P-value

pH 7.33 ± 0.01 7.34 ± 0.01 0.650

PO2 (mmHg) 17.64 ± 0.73 12.09 ± 0.64 <0.001

Lactate (mmol/L) 2.77 ± 0.23 2.92 ± 0.40 0.748

Glucose (mmol/L) 1.11 ± 0.07 0.90 ± 0.09 0.084

ACTH (pg/mL) 23.44±2.66 20.80±1.81 0.529

Cortisol (ng/mL) 4.57 ± 0.64 6.20 ± 0.84 0.165

Adrenaline (ng/L) 0.23 ± 0.05 0.60 ± 0.16 0.025

Noradrenaline (ng/L) 4.48 ± 1.08 9.16 ± 1.57 0.021

Weight* (kg) 3.22 ± 0.07 2.30 ± 0.07 <0.001 *measured at completion of experimental phase

5.3.1.2 Fetal acid-base responses to repeated cord occlusion

Fetal arterial pH, PO2, PCO2, and base excess levels were not significantly different between

non-embolised controls and embolised controls. In addition, there were no changes in these

variables over the duration of the experimental phase in both control groups (Figure 5.3).

Arterial pH levels fell significantly during repeated UCO in both non-embolised fetuses

(p<0.001) and embolised fetuses (p<0.001) when compared to controls. The rate of fall of

arterial pH during UCO was significantly greater in embolised fetuses (p=0.002), and

considering the entire experimental phase, the embolised fetuses were significantly more

acidaemic than the non-embolised fetuses (mean group difference: -0.08 pH units [95% C.I.:-

0.14, -0.02], p=0.013).

Fetal arterial PO2 fell significantly during repeated UCO in both non-embolised fetuses

(p=0.003) and embolised fetuses (p<0.001) when compared to controls; however, the

decrease in PO2 levels occurred earlier and was more severe in the embolised fetuses

(p=0.027). The embolised fetuses had a significantly lower arterial PO2 than non-embolised

fetuses over the duration of the cord occlusion protocol (mean group difference: 5.17mmHg

[95% C.I. 3.67, 6.66], p<0.002).

230

Figure 5.3 Fetal acid-base responses during repeated umbilical cord occlusion. P-values on

each graph compare the fetuses that experienced UCO (occluded fetuses) with controls

whereas p-values between boxes compare the two groups of controls (non-embolised and

embolised) or the two groups of fetuses that experienced UCO. All measures of fetal acid-

base balance presented in this figure were not significantly different between the two groups

of controls. All fetuses exposed to repeated UCO developed hypoxia and respiratory

acidaemia. Although base excess appeared to fall in both groups of fetuses experiencing

UCO, these differences did not reach statistical significance. The magnitude of the hypoxia,

hypercapnia and acidaemia were significantly greater in the embolised fetuses than the non-

embolised fetuses.

231

Occluded(p=0.027)

0 27 54 81

7.0

7.1

7.2

7.3

7.4 pH

Time (minutes)

pH

p<0.001

0 27 54 81

7.0

7.1

7.2

7.3

7.4 pH

p<0.001

Time (minutes)

pH0 27 54 81

10

15

20

25

30

35 PO2

Time (minutes)

PO

2(m

mH

g)

p=0.003

0 27 54 81

10

15

20

25

30

35 PO2

p<0.001

Time (minutes)P

O2

(mm

Hg)

0 27 54 81

30

50

70

90PCO2

p=0.003

Time (minutes)

PC

O2

(mm

Hg)

0 27 54 81

30

50

70

90PCO2

p<0.001

Time (minutes)

PC

O2

(mm

Hg)

0 27 54 81

-10

-6

-2

2

Control Occluded

Base Excess

Time (minutes)

Base

Exc

ess

p=393

0 27 54 81

-10

-6

-2

2Base Excess

Control Occluded

Time (minutes)

Base

Exc

ess p=0.226

NON-EMBOLISED EMBOLISED

Control(ns)

Occluded(p=0.002)

Control(ns)

Occluded(p=0.131)

Control(ns)

Occluded(p=0.027)

Control(ns)

231

Figure 5.4 Fetal arterial lactate responses during repeated umbilical cord occlusion. P-values

on each graph compare fetuses experiencing UCO (occluded fetuses) with controls whereas

the statistical comparisons between the boxes compare the two groups of controls (non-

embolised and embolised) or the two groups of fetuses experiencing UCO. Arterial lactate

levels in controls did not vary during the experimental phase and were not significantly

different between groups. Repeated UCO was associated with an increase in arterial lactate

levels in non-embolised and embolised fetuses (p<0.001). The lactate response in the

embolised fetuses exposed to UCO was not significantly different from the lactate response

in the non-embolised fetuses experiencing UCO (p=0.300).

232

0 27 54 81

0

2

4

6

8

10 Lactate

Control

Occluded(p<0.001)

Time (minutes)

Lact

ate

(mm

ol/L

)

0 27 54 81

0

2

4

6

8

10 Lactate

Control

Occluded(p=0.002)

Time (minutes)

Lact

ate

(mm

ol/L

)

NON-EMBOLISED

EMBOLISED

Control(ns)

Occluded(ns)

233

Fetal arterial PCO2 levels increased significantly during repeated UCO in both non-embolised

fetuses (p=0.003) and embolised fetuses (p<0.001) when compared to controls. Similar to

arterial pH and PO2 responses, the magnitude of the change in arterial PCO2 levels in the

embolised fetuses was significantly greater than non-embolised fetuses (Figure 5.3).

5.3.1.3 Fetal arterial lactate response to repeated cord occlusion

Fetal arterial lactate levels were not significantly different between non-embolised controls

and embolised controls and there were no significant changes in these levels during the

experimental phase (Figure 5.4). Repeated UCO significantly increases arterial lactate levels

in both non-embolised fetuses (p<0.001) and embolised fetuses (p=0.002) when compared

to controls and there were no differences in lactate responses between groups (p=0.300).

5.3.1.4 Fetal endocrine responses to repeated cord occlusion

Fetal arterial glucose, cortisol, noradrenaline and adrenaline levels in both non-embolised and

embolised controls did not change significantly over the duration of the experimental phase

(Figure 5.5). Glucose levels were significantly lower in embolised controls (0.87±

0.08mmol/L) than non-embolised controls (1.13±0.08mmol/L; p=0.024); however cortisol,

noradrenaline and adrenaline levels were not significantly different between the two control

groups.

Fetal arterial glucose levels were not significantly different between non-embolised controls

and non-embolised fetuses that underwent UCO (p=0.101) In embolised fetuses UCO was

associated with a significant increase in fetal glucose levels (p=0.041). The opposite pattern

of responses occurred in fetal cortisol levels; non-embolised fetuses subjected to UCO

significantly increased their cortisol levels compared to controls (p=0.048) whilst embolised

fetuses exposed to the same UCO protocol did not significantly increase their cortisol levels

(p=0.156) when compared to appropriate controls. The different cortisol responses between

the two groups exposed to UCO occurred despite fetal ACTH levels increasing by the same

amount in both groups in response to asphyxia (Figure 5.6).

In both non-embolised and embolised fetuses, repeated UCO was associated with significant

increases in both noradrenaline (non-embolised p=0.022, embolised p=0.006) and adrenaline

levels (non-embolised p=0.008, embolised p=0.001) when compared to controls (Figure 5.5).

The magnitude of the increase in catecholamines was significantly greater in the embolised

fetuses (noradrenaline p=0.024, adrenaline p<0.001).

233

Figure 5.5 Fetal endocrine responses during umbilical cord occlusion. P-values on each

graph compare fetuses experiencing UCO (occluded fetuses) with controls whereas p-values

between the boxes compare the two groups of controls (non-embolised and embolised) or

the two groups of fetuses experiencing UCO. Repeated UCO was associated with an

increase in blood glucose levels in embolised occluded fetuses (p=0.041) which did not occur

in the non-embolised fetuses exposed to the same asphyxial insults (p=0.078). Conversely,

an increase in arterial cortisol levels in response to asphyxia only occurred in the non-

embolised fetuses (p=0.048) and not in the embolised fetuses during UCO. Both groups of

fetuses that experienced UCO increased their catecholamine levels in response to asphyxia;

however, the magnitude of this response was greater in the embolised fetuses (p<0.024).

234

0

1

2

3

0 27 54 81

Glucose

Time (minutes)

Glu

cose

(mm

ol/L

) p=0.041

0

10

20

30

40

50

0 81

Cortisol p=0.048

Time (minutes)

Cor

tisol

(ng/

mL)

0

20

40

60

80

0 81

Noradrenaline p=0.022

Time (minutes)

Nor

adre

nalin

e (n

g/L)

0

20

40

60

80

0 81

Noradrenaline

Time (minutes)

Nor

adre

nalin

e (n

g/L) p=0.006

0

20

40

60

80

0 81

Adrenaline

Control Occluded

p=0.018

Time (minutes)

Adre

nalin

e (n

g/L)

0

20

40

60

80

0 81

Adrenaline

Control Occluded

Time (minutes)

Adre

nalin

e (n

g/L)

p=0.001

NON-EMBOLISED EMBOLISED

Occluded(p=0.078)

Control(p=0.024)

Occluded(p<0.001)

Control(ns)

Occluded(p=0.018)

Control(ns)

Occluded(0.024)

Control(ns)

0

1

2

3

0 27 54 81

Glucose p=0.101

Time (minutes)

Glu

cose

(mm

ol/L

)

0

10

20

30

40

50

0 81

Cortisol p=0.156

Time (minutes)C

ortis

ol (n

g/m

L)

235

Non-em

bolis

ed 0

min

Contro

l 81 m

in

Occlud

ed 81

min

0

200

400

600

800

ACTHp=0.008

AC

TH (p

g/m

L)

Embolis

ed 0

min

Contro

l 81 m

in

Occlud

ed 81

min

0

200

400

600

800

ACTHp=0.001

AC

TH (p

g/m

L)

Non-em

bolis

ed 0

min

Contro

l 81 m

in

Occlud

ed 81

min

0

10

20

30

40

50

Cortisol

p=0.038

Cor

tisol

(ng/

mL)

Embolis

ed 0

min

Contro

l 81 m

in

Occlud

ed 81

min

0

10

20

30

40

50

Cortisol

p=0.160

Cor

tisol

(ng/

mL)

NON-EMBOLISED EMBOLISED

Occluded(ns)

Control(p=0.074)

Occluded(p=0.018)

Control(ns)

Figure 5.6 Fetal ACTH and cortisol responses during umbilical cord occlusion. P-values on

each graph compare fetuses experiencing UCO with baseline levels collected at the beginning

of the experimental phase. P-values between the boxes compare the two groups of controls

(non-embolised and embolised) or the two groups of fetuses experiencing UCO. Fetal

ACTH levels were significantly increased in both non-embolised fetuses (p=0.008) and

embolised fetuses (p=0.001) experiencing UCO when compared to baseline levels and post-

occlusion ACTH levels were not significantly different between groups of fetuses exposed to

UCO (p=0.768). Despite similar ACTH responses, the cortisol responses were different in

the two groups of occluded fetuses with a significant increase in cortisol occurring only in the

non-embolised fetuses subjected to UCO (p=0.038)

236

5.3.1.5 Fetal heart rate responses to repeated cord occlusion

5.3.1.5.1 Baseline fetal heart rate response to repeated cord occlusion

Baseline FHR, assessed during the one minute prior to each UCO, was not significantly

different between non-embolised controls (137.6±5.1bpm) and embolised controls

(138.0±5.9bpm; p=0.955; Figure 5.7) and there was no significant change in baseline FHR

over the duration of the experimental phase in both control groups (p>0.893). Fetal heart

rate increased significantly in both non-embolised fetuses exposed to UCO (p=0.008) and

embolised fetuses exposed to UCO (p=0.044) when compared to appropriate controls.

Visually the two occluded groups had different patterns of response in baseline FHR during

the experimental phase (Figure 5.8); however, these differences were not statistically

significant (p=0.125). Fetal heart rate in the embolised fetuses increased earlier in response

to UCO than occurred in non-embolised fetuses (Figure 5.7). In addition, the maximum

increase in FHR in embolised fetuses experiencing UCO was less than non-embolised fetuses

during UCO (p=0.042).

5.3.1.5.2 Fetal heart rate variability response to repeated cord occlusion

Fetal heart rate variability (MMR, STV and LTV) was lower in embolised controls than in

non-embolised controls; however, these differences did not reach statistical significance

(non-embolised vs. embolised; MMR 22.49±3.39msec vs. 20.72±3.11msec, p=0.695; STV

1.60±0.20msec vs. 1.53±0.19msec, p=0.804; LTV 6.15±1.54bpm vs. 3.69±1.42bpm,

p=0.383). All measures of FHR variability increased in both non-embolised and embolised

fetuses subjected to UCO (Figure 5.7). The magnitude of the increase in STV and LTV was

significantly greater in non-embolised than embolised fetuses experiencing UCO (STV,

p=0.001; LTV, p=0.012). Fetal heart rate variability responses of the two groups of fetuses

subjected to UCO are presented graphically in Figure 5.8.

5.3.1.5.3 Fetal heart rate patterns during repeated cord occlusion

The ultrasound cuff around the umbilical vessels confirmed normal umbilical blood flow in

controls and between cord occlusions in the occluded fetuses. Doppler assessment also

confirmed complete obstruction of umbilical blood flow during each cord occlusion during

the experimental phase.

In controls, no accelerations or decelerations were present during the 81 minute experimental

phase. In five of the seven non-embolised fetuses subjected to UCO, FHR accelerations

were present primarily during the 0.5-minute cord occlusions; in three fetuses these were

236

Figure 5.7 Fetal heart rate responses during umbilical cord occlusion. P-values on each graph

compare fetuses experiencing UCO (occluded fetuses) with controls whereas p-values

between the boxes compare the two groups of controls (non-embolised and embolised) or

the two groups of fetuses experiencing UCO. All fetuses subjected to UCO increased their

inter-occlusion FHR and FHR variability; however, the magnitude of these increases were

less in the embolised fetuses compared to the non-embolised fetuses. The maximum

increase in FHR in embolised fetuses was less than in non-embolised fetuses during UCO

(p=0.042). Similarly, the magnitude of the increase in STV and LTV in embolised fetuses

was less than in non-embolised fetuses during UCO (p=0.001 and p=0.012 respectively).

237

-27 0 27 54 81100

150

200

250

p=0.008

Time (minutes)

FHR

(bpm

)

-27 0 27 54 81100

150

200

250

p=0.044

Time (minutes)

FHR

(bpm

)

-27 0 27 54 810

20

40

60

80

100

Time (minutes)

FHR

MM

R (m

sec)

p<0.001

-27 0 27 54 810

20

40

60

80

100

Time (minutes)

FHR

MM

R (m

sec)

p=0.057

-27 0 27 54 810

3

6

9

Time (minutes)

FHR

STV

(mse

c)

p<0.001

-27 0 27 54 810

3

6

9

Time (minutes)

FHR

STV

(mse

c)

p=0.717

-27 0 27 54 810

20

40

60

p<0.001

Control Occluded

Time (minutes)

FHR

LTV

(bpm

)

-27 0 27 54 810

20

40

60

p=0.018

Control Occluded

Time (minutes)

FHR

LTV

(bpm

)

NON-EMBOLISED EMBOLISED

Occluded(ns)

Control(ns)

Occluded(0.012)

Control(ns)

Occluded(ns)

Control(ns)

Occluded(p=0.001)

Control(ns)

237

5.8 Fetal heart rate responses in the two groups of fetuses exposed to repeated cord

occlusion. P-values on each graph compare the non-embolised fetuses subjected to UCO

with the embolised fetuses subjected to UCO. Although both groups of fetuses experiencing

UCO increased their inter-occlusion FHR and FHR variability, the pattern of these responses

was altered by umbilicoplacental embolisation.

238

-27 0 27 54 81100

150

200

250

Embolised

Non-embolised

Time (minutes)

FHR

(bpm

)

p=0.125

-27 0 27 54 810

20

40

60

80

100

Non-embolised

Embolised

Time (minutes)

FHR

MM

R (m

sec) p=0.129

-27 0 27 54 810

3

6

9

Non-embolised

Embolised

Time (minutes)

FHR

STV

(mse

c) p=0.549

-27 0 27 54 810

20

40

60

Non-embolised

Embolised

Time (minutes)

FHR

LTV

(bpm

) p=0.447

239

repeated events and in two they were isolated events. A similar pattern was observed in the

embolised fetuses subjected to UCO with FHR accelerations present in four of the seven

fetuses; in two fetuses they were repeated during the 0.5-minute occlusions and in two they

were isolated events. Variable FHR decelerations were seen in all fetuses subjected to UCO.

Severe variable decelerations occurred in five of the seven non-embolised fetuses and six of

the seven embolised fetuses during UCO. Similarly late components to decelerations

occurred in five of 7 and six of 7 fetuses in these groups, respectively. Overshoot instability

was present in all of the non-embolised fetuses at some time during the experimental phase

but was only present in four of the seven embolised fetuses subjected to UCO. Both severe

variable decelerations (Figure 5.9A and 5.9B) and late components to decelerations (Figure

5.9C and 5.9D) occurred significantly earlier during repeated cord occlusions in embolised

occluded fetuses when compared to non-embolised occluded fetuses (severe variable

decelerations, p=0.030; late component, p=0.042). Embolised fetuses also had more

variability in the time of onset of both of these decelerative patterns than non-embolised

fetuses (severe variable deceleration, p=0.025; late component, p=0.013). Conversely,

overshoot instability after severe variable decelerations (Figure 5.9E and 5.9F) occurred later

in embolised occluded fetuses than in non-embolised occluded fetuses (p=0.031).

5.3.1.6 Fetal arterial pressure responses to repeated cord occlusion

5.3.1.6.1 Fetal arterial pressure between cord occlusions

Fetal arterial pressures did not change significantly in either control group over the duration

of the experimental phase (p=0.576; Figure 5.10 and 5.11). Both systolic and diastolic

pressures in embolised controls were significantly higher than non-embolised controls

(systolic pressure 62.6±2.4mmHg vs. 55.6±2.4mmHg, p<0.001; diastolic pressure

46.7±2.6mmHg vs. 36.8±2.6mmHg, p<0.001).

In non-embolised fetuses both systolic and diastolic pressures increased significantly in

response to UCO when compared with controls (Figures 5.10 and 5.11; both p<0.001). In

contrast, fetal systolic (p=0.236) and diastolic pressures (p=0.753) in embolised fetuses that

underwent UCO were not significantly different to controls. Over the duration of the

experimental phase, the systolic and diastolic pressure responses in the non-embolised

fetuses subjected to UCO were significantly different from embolised fetuses subjected to

UCO (systolic pressures, p=0.002; diastolic pressures, p=0.012).

240

Severe VariableDecelerations

0

27

54

81

Embolised Occluded

Non-embolised Occluded

p = 0.030

Tim

e of

Ons

et (m

in)

Severe VariableDecelerations

0 9 18 27 36 45 54 63 72 810

25

50

75

100

p = 0.138

Time (min)

Patte

rn P

rese

nt (%

)

Late Components toDecelerations

0 9 18 27 36 45 54 63 72 810

25

50

75

100

p = 0.042

Time (min)

Patte

rn P

rese

nt (%

)

Late Components toDecelerations

0

27

54

81

Embolised Occluded

Non-embolised Occluded

p = 0.029

Tim

e of

Ons

et (m

in)

Overshoot

0 9 18 27 36 45 54 63 72 810

25

50

75

100

p = 0.031

Non-embolised / OccludedEmbolised / Occluded

Time (min)

Patte

rn P

rese

nt (%

)

Overshoot

0

27

54

81

Embolised Occluded

Non-embolised Occluded

p = 0.083

Tim

e of

Ons

et (m

in)

A B

C D

E F

Figure 5.9 Fetal heart rate decelerative patterns during umbilical cord occlusion. Figures A, C

and E show the percentage of fetuses at each time point with particular FHR patterns.

Figures B, D and F show the median time of onset of each FHR pattern. Both severe

variable decelerations and late components to decelerations occurred earlier and with more

variability in time of onset in embolised fetuses that underwent repeated UCO than non-

embolised fetuses exposed to the same asphyxial insult. In contrast, overshoot instability

occurred later, and only in 57% of embolised fetuses experiencing UCO.

241

-27 -18 -9 0 9 18 27 36 45 54 63 72 8140

50

60

70

80

90

Control

Occluded

p<0.001

Time (min)

Feta

l Sys

tolic

Pres

sure

(mm

Hg)

-27 -18 -9 0 9 18 27 36 45 54 63 72 8140

50

60

70

80

90

Control

Occluded

Time (min)

Feta

l Sys

tolic

Pres

sure

(mm

Hg) p=0.236

NON-EMBOLISED

EMBOLISED

Control(p<0.001)

Occluded(p=0.002)

Figure 5.10 Fetal systolic pressure responses between umbilical cord occlusions. P-values on

each graph compare fetuses experiencing UCO with controls whereas p-values between the

boxes compare the two groups of controls (non-embolised and embolised) or the two groups

of fetuses that experienced UCO during the experimental phase. The systolic pressures in

the embolised controls were significantly higher than in the non-embolised controls

(p<0.001) and, when subjected to UCO the embolised fetuses did not significantly increase

their inter-occlusion systolic pressures (p=0.236). The non-embolised fetuses increased their

systolic pressures between the 1.0-minute and 1.5-minute cord occlusions (p<0.001).

242

-27 -18 -9 0 9 18 27 36 45 54 63 72 8120

30

40

50

60

70

Control

Occluded

Time (min)

Fet

al D

iast

olic

Pres

sure

(mm

Hg) p<0.001

-27 -18 -9 0 9 18 27 36 45 54 63 72 8120

30

40

50

60

70

Occluded

Control

Time (min)

Fet

al D

iast

olic

Pres

sure

(mm

Hg) p=0.753

NON-EMBOLISED

EMBOLISED

Control(p<0.001)

Occluded(p=0.002)

Figure 5.11 Fetal diastolic pressure responses between umbilical cord occlusions. P-values

on each graph compare fetuses experiencing UCO with controls whereas p-values between

the boxes compare the two groups of controls or the two groups of fetuses that experienced

UCO during the experimental phase. The diastolic pressures in embolised controls were

significantly higher than non-embolised controls (p<0.001). In response to UCO the non-

embolised fetuses significantly increased their diastolic pressures compared to non-embolised

controls (p<0.001) whereas the diastolic pressures in the embolised fetuses experiencing

UCO were not significantly different from embolised controls (p=0.753).

243

5.3.1.6.2 Fetal arterial pressure during cord occlusions

Both non-embolised and embolised fetuses subjected to UCO displayed the four distinct

patterns of fetal arterial pressure responses during cord occlusion (hypertensive, decreasing

hypertensive, biphasic systolic and biphasic diastolic) that were described in detail in Section

3.3.3.2 and illustrated in Figure 3.14. The rate of progression through the four distinct

patterns was similar in embolised and non-embolised fetuses with no significant differences

between groups in the timing of decreasing hypertensive response and biphasic systolic

response (Figure 5.12). The embolised fetuses that experienced repeated UCO developed

the biphasic diastolic response earlier (p=0.056) and in a more variable manner (p=0.046)

than the non-embolised fetuses exposed to the same UCO protocol.

5.3.1.7 Physiologic associations with fetal heart rate patterns in embolised and non-

embolised fetuses

To investigate the potential mechanisms responsible in embolised fetuses for the earlier

decelerative FHR patterns when compared to non-embolised fetuses, univariate and

multivariate regression modeling was performed.

Summaries of the univariate analyses of risk factors associated with particular heart rate

patterns, including fetal blood gas variables, fetal arterial pressure between cord occlusions

and fetal catecholamine concentrations, are presented in Table 5.2. A number of significant

associations were identified with each FHR pattern except overshoot after variable

decelerations which did not appear to be associated with any of the variables considered in

the analysis.

Summaries of the multivariate analyses considering all of the acid-base measures, lactate

levels and arterial pressures simultaneously as risk factors for particular FHR patterns are

presented in Table 5.3. A subset analysis of occlusions 1, 2, 26 and 27 was also performed to

assess fetal catecholamine concentrations in the multivariate analysis together with fetal blood

gases and fetal arterial pressures. Significant associations with FHR patterns are presented in

Table 5.4. Optimal cut off values for significant predictors of FHR patterns were estimated

using ROC analysis and are summarised in Table 5.5.

243

Figure 5.12 Fetal arterial pressure responses during umbilical cord occlusions. The rate of

progression through the four distinct patterns was similar in embolised and non-embolised

fetuses with no significant differences between groups in the timing of decreasing

hypertensive response and biphasic systolic response (Figure 5.12). The embolised fetuses

that experienced repeated UCO developed the biphasic diastolic response earlier (p=0.056)

and in a more variable manner (p=0.046) than the non-embolised fetuses exposed to the

same UCO protocol

244

0 27 54 810

25

50

75

100

p=0.173

Time (min)

Patte

rn P

rese

nt (%

)

0

27

54

81p=0.171

EmbolisedOccluded

Non-embolisedOccluded

Tim

e (m

in)

0 27 54 810

25

50

75

100

Time (min)

Patte

rn P

rese

nt (%

)

p=0.219

0

25

50

75

100p=0.175

EmbolisedOccluded

Non-embolisedOccluded

Tim

e (m

in)

0 27 54 810

25

50

75

100

p=0.086

Non-embolised / Occluded

Embolised / Occluded

Time (min)

Patte

rn P

rese

nt (%

)

0

25

50

75

100p=0.056

EmbolisedOccluded

Non-embolisedOccluded

Tim

e (m

in)

Decreasing Hypertensive Response

Biphasic Systolic Response

Biphasic Diastolic Response

245

Table 5.2 Univariate associations of FHR patterns with blood gas, arterial pressure and catecholamine values

Univariate Associations FHR Pattern Association a

Blood Gases Arterial Pressure Catecholamines

Positive Base excess‡ pH*

Acceleration

Inverse Lactate‡ PCO2

* Diastolic pressure† Adrenaline* Noradrenaline*

Positive Lactate‡ PCO2

‡ Diastolic pressure‡ Systolic pressure†

Adrenaline* Noradrenaline†

Late Deceleration

Inverse pH‡ Base excess‡ PO2

†

Positive Lactate‡ PCO2

* Adrenaline*

Noradrenaline* Severe Variable Deceleration Inverse

Base excess‡ pH† PO2

†

Positive Lactate† PCO2

* Diastolic pressure* Adrenaline† Increased Variability

Inverse Base excess‡ pH† PO2

†

Positive Overshoot

Inverse

a A positive association indicates that the odds of the FHR pattern occurring increased as the value of the variable increased. An inverse association indicates the odds of the FHR pattern occurring increased as the value of the variable decreased. *p≤0.05, † p<0.01, ‡ p<0.001

246

Table 5.3 Multivariate associations of FHR patterns considering all blood gas variables and fetal arterial pressure variables simultaneously

Group

Embolised Non-embolised

FHR Pattern Associationa Variable Variable P-value

Positive pH Base excess

0.027 0.022

Acceleration

Inverse PO2

Systolic pressure -

PO2 Systolic pressure

PCO2

<0.001 0.020 0.041

Positive Lactate Systolic pressure*

Lactate Systolic pressure

<0.001 0.025 Late

Deceleration Inverse Base excess

pH Base excess

- 0.017

<0.001

Positive

Systolic pressure Lactate†

- PCO2

Systolic pressure Lactate

Base excess -

<0.001 0.003 0.001 0.026 Severe

Variable Deceleration

Inverse - Base excess

pH -

0.011 0.001

Positive -

PCO2 Systolic pressure

Diastolic pressure - -

<0.001 <0.001 <0.001

Increased Variability

Inverse

- pH

Base excess Lactate

Diastolic pressure

Systolic pressure - - - -

<0.001 <0.001 <0.001 <0.001 <0.001

Positive Overshoot

Inverse

a A positive association indicates that the odds of the FHR pattern occurring increased as the value of the variable increased. An inverse association indicates the odds of the FHR pattern occurring increased as the value of the variable decreased. *Effect significantly greater in embolised fetuses, p=0.003 † Effect significantly greater in non-embolised fetuses, p=0.031

247

Table 5.4 Multivariate associations of FHR patterns considering all blood gas, fetal arterial pressure and catecholamine variables simultaneously

FHR Pattern Associationa Multivariate Associations

(blood gas, arterial pressure, catecholamines)

P-value

Positive

Acceleration Inverse PO2

Noradrenaline 0.016 0.006

Positive Adrenaline 0.007 Late Deceleration

Inverse Base excess <0.001*

Positive Systolic pressure 0.006 Severe Variable Deceleration

Inverse pH 0.043

Positive Diastolic PCO2

0.009 0.011 Increased Variability

Inverse Systolic pressure Base excess

0.016 <0.001

Positive Overshoot

Inverse

a A positive association indicates that the odds of the FHR pattern occurring increased as the value of the variable increased. An inverse association indicates that the odds of the FHR pattern occurring increased as the value of the variable decreased. *Greater effect in embolised fetuses (p=0.021)

248

Table 5.5. Receiver operator curve analysis for the prediction of FHR patterns from fetal physiologic parameters

FHR Pattern Physiologic Variable

Umbilicoplacental Circulation

Optimum Cut Off Sensitivity Specificity

Non-embolised 22.1 88.9 61.2 Arterial PO2 (mmHg) Embolised 21.1 100.0 50.0

Embolised 65 100.0 43.9 Acceleration

Systolic pressure (mmHg) Non-embolised 50 26.7 95.4

Non-embolised 3.1 85.7 67.3 Lactate (mmol/L) Embolised 3.1 97.0 77.7

Non-embolised -1.8 80.0 60.5 Base Excess

Embolised -4.1 69.7 79.8

Non-embolised 61 74.3 64.1

Late Deceleration

Systolic pressure (mmHg) Embolised 55 95.5 55.0

Non-embolised 61 97.2 67.0 Systolic pressure (mmHg) Embolised 63 87.1 68.7

Non-embolised 4.5 77.8 93.5

Severe Variable Deceleration Lactate

(mmol/L) Embolised 3.5 92.9 78.5

Non-embolised 53 92.7 29.1 Systolic pressure (mmHg) Embolised 67 89.5 72.1

Non-embolised 35 97.6 33.3 Increased Variability Diastolic

pressure (mmHg) Embolised 41 100.0 47.8

5.3.1.7.1 Fetal heart rate accelerations

Multivariate analyses indicated that the only variables associated with FHR accelerations in

both groups of fetuses were fetal arterial PO2 (p<0.001) and fetal systolic pressure (p=0.020).

The magnitudes of these effects were similar with the odds ratio for accelerations increasing

by 2.73 (95%CI 2.06-3.71) for each fall in PO2 of 5 mmHg and decreasing by 1.40 (95%CI

1.05-1.86) for each increase in systolic pressure of 5mmHg. After inclusion of catecholamine

data into the multivariate analysis, a significant inverse association was also found between

noradrenaline (p=0.006). ROC analysis indicated the very close optimum cut offs for PO2 in

both groups with high sensitivity and modest specificity (Table 5.5). The optimum cut off

values for systolic pressure was different in the embolised and non-embolised fetuses with

249

very poor sensitivity in the non-embolised fetuses. In the non-embolised fetuses, additional

associations were found between pH, PCO2, and base excess and the presence of FHR

accelerations. The magnitudes of these effects were at least five times greater than the effect

of PO2 or systolic pressure. Arterial pH, PO2 and base excess were not associated with FHR

accelerations in fetuses that underwent embolisation.

5.3.1.7.2 Late component to decelerations

Multivariate analyses indicated that the only variables associated with a late component to

decelerations in both groups of fetuses were arterial lactate concentrations (p<0.001), systolic

pressure (p=0.025) and base excess (p=0.017; Table 5.3). In both groups of fetuses the

magnitude of effect of lactate and base excess were similar with the odds ratio of late

decelerations increasing by 1.62 (95%CI 1.09-2.41) for each fall of 2 base excess units and

increasing by 2.90 (95%CI 1.60-5.24) for each 2mmol/L increase in lactate. The effect of

systolic pressure on late decelerations was significantly greater in the embolised fetuses

(p=0.003) with each 5 mmHg increase in fetal systolic pressure increasing the odds ratio of

late decelerations by 6.44 (95%CI 2.36-17.55) in the embolised fetuses, in contrast to 1.33

(95%CI 1.01-1.76) in the non-embolised fetuses. Fetal arterial pH was only associated with

late decelerations in the embolised group, with each fall in pH of 0.05 pH units increasing the

odds ratio of late decelerations by 4.63 (95%CI 2.37-9.06, p<0.001). Inclusion of

catecholamine data into the model identified an additional significant positive association

with adrenaline (p=0.007; Table 5.4).

ROC analysis for the presence of late decelerations found the same lactate cut off

(3.1mmol/L) for both groups of fetuses (Table 5.5). Embolised fetuses developed late

decelerations at lower systolic pressure (embolised 55mmHg vs. non-embolised 61mmHg)

and lower base excess values than the non-embolised fetuses (embolised base excess -4.1 vs.

non-embolised -1.8).

5.3.1.7.3 Severe variable decelerations.

Multivariate analysis identified different associations between severe variable decelerations

and fetal physiological variables in embolised and non-embolised fetuses (Table 5.3). Fetal

lactate (p=0.003) and systolic pressure (p<0.001) were found to be significantly associated

with severe variable decelerations in both groups of fetuses; however, changes in fetal lactate

concentration had a significantly greater effect on the odds ratio of severe variable

decelerations in the non-embolised fetuses (p=0.031). For each increase of 1mmol/L of

lactate, the odds ratio of severe variable decelerations was increased by 2.30 (95%CI 1.33-

250

3.97) in the non-embolised fetuses compared to 1.22 (95%CI 1.02-1.45) in the embolised

fetuses. For each increase in systolic pressure of 5mmHg, the odds ratio of severe variable

decelerations increased by 2.14 (95%CI 1.45-3.17).

With multivariate analysis, arterial pH was associated only with severe variable decelerations

in the non-embolised group of fetuses (Table 5.3). The magnitude of this inverse association

was greater than that with systolic pressure or lactate, with each decrease in pH of 0.05 units

increasing the odds ratio of severe variable decelerations by 8.67 (95%CI 2.41-31.1, p=0.011).

Conversely, arterial PCO2 was associated with severe variable decelerations only in the

embolised group of fetuses and the magnitude of this positive association was at least twice

that of systolic pressure and lactate. Each increase in PCO2 of 5mmHg increased the odds

ratio of severe variable decelerations by 4.64 (95%CI 1.04-20.82, p=0.026).

The covariate adjusted association between base excess and severe variable decelerations was

different between embolised and non-embolised fetuses. In non-embolised fetuses, a

decrease in base excess by 2 units resulted in a reduction of the odds ratio for severe variable

decelerations of 1.83 (95%CI 1.45-2.31, p=0.001), while in the embolised fetuses the same

change resulted in a large increase in the odds ratio for severe variable decelerations of 10.5

(95%CI 1.79-61.45, p=0.001).

No association was found between severe variable decelerations and concentrations of

catecholamines. ROC analysis identified the same optimum cut offs in systolic pressure for

the prediction of severe variable decelerations in both groups of fetuses, however they

occurred at lower lactate concentrations in the embolised fetuses (embolised 3.5mmol/L,

non-embolised 4.5mmol/L; Table 5.5).

5.3.1.7.4 Increased fetal heart rate variability

Multivariate analysis showed entirely different variables associated with increased FHR

variability in the non-embolised and embolised fetuses. In the non-embolised fetuses a

positive association was found between increased FHR variability and diastolic pressure

(p<0.001) and an inverse relationship was found with systolic pressure (p<0.001; Table 5.3).

For each rise in diastolic pressure of 5 mmHg the odds ratio for increased FHR variability

increased by 2.28 (95%CI 1.31-3.95). Conversely, for each rise in systolic pressure of 5

mmHg, the odds ratio for increased FHR variability was reduced by at least 9.44 (95%CI

9.44-7361.80). In embolised fetuses, the opposite effect was apparent with a positive

association between systolic pressure (p<0.001) and an inverse association between diastolic

pressure (p<0.001). The magnitude of these effects was similar to those for non-embolised

251

fetuses except their direction was reversed. ROC analysis demonstrated very different

optimum cut off values for systolic and diastolic pressures for non-embolised and embolised

fetuses (Table 5.5).

In the embolised fetuses multivariate analysis showed a positive association between PCO2

and increased FHR variability and an inverse association with pH, base excess, and lactate

concentration (all p<0.001). The magnitudes of these associations were all greater than those

associated with the arterial pressure changes. No association was found between increased

FHR variability and concentrations of the catecholamines that were measured.

5.3.2 The effect of umbilicoplacental embolisation on the relationship between fetal

scalp blood levels and central levels of lactate and pH

5.3.2.1 Fetal scalp blood lactate levels

Fetal scalp, carotid arterial and jugular venous blood lactate levels for both non-embolised

and embolised fetuses during the experimental phase are presented in Figures 5.13 and 5.14.

In addition to lactate levels from the three sampling sites, these figures present graphically the

differences between fetal scalp and jugular venous lactate levels compared to carotid arterial

lactate levels (∆lactate) based on paired analysis within the experimental group. Comparisons

between the three sampling sites in controls, non-embolised and embolised groups are

presented in Figure 5.15.

5.3.2.1.1 Fetal scalp blood lactate levels and carotid arterial lactate levels

Fetal scalp blood lactate levels were significantly higher than carotid arterial lactate levels in

all four experimental groups (scalp-artery lactate: non-embolised control 1.4±0.5mmol/L,

p=0.017; non-embolised occluded 1.4±0.4mmol/L, p=0.005; embolised control

1.7±0.5mmol/L, p=0.005; embolised occluded 1.7±0.5mmol/L, p=0.002, Figure 5.15A).

The relationship between fetal scalp blood and carotid arterial lactate levels was not

influenced by umbilicoplacental embolisation (p=0.242), cord occlusion (p=0.114),

noradrenaline (p=0.829), adrenaline (p=0.137), glucose levels (p=0.919) or time during the

experimental phase (p=0.161); however, it was modified by arterial lactate levels (p<0.001)

and all of the measures of fetal acid-base status (pH p<0.001; PO2 p<0.001; PCO2 p<0.001).

The effect of arterial lactate and PCO2 levels on the scalp-carotid arterial lactate relationship

was that fetal scalp blood lactate underestimated arterial lactate levels as lactate and PCO2

252

Control Fetuses

-300

2

4

6

8

10

0 27 54 81

FetalScalp

CarotidArtery

Time (min)

Lact

ate

(mm

ol/L

)

Control Fetuses

-300

2

4

6

8

10

0 27 54 81

FetalScalp

CarotidArtery

Time (min)

Lact

ate

(mm

ol/L

)

-300

2

4

6

8

10

0 27 54 81

FetalScalp

JugularVein

Time (min)

Lact

ate

(mm

ol/L

)

-300

2

4

6

8

10

0 27 54 81

FetalScalp

JugularVein

Time (min)La

ctat

e (m

mol

/L)

Scalp Vein

-2

-1

0

1

2p=0.017 p=0.882

∆La

ctat

e (m

mol

/L)

Scalp Vein

-2

-1

0

1

2p=0.005 p=0.575

∆La

ctat

e (m

mol

/L)

NON-EMBOLISED EMBOLISEDA

B

C

D

E

F

Figure 5.13 The relationship between fetal scalp blood, carotid arterial blood and jugular

venous blood lactate levels in control fetuses. Fetal scalp blood lactate levels were

significantly higher than carotid arterial lactate levels in both non-embolised (Figure A,

p=0.017) and embolised controls (Figure D, p=0.005). Similarly fetal scalp blood lactate

levels were significantly higher than jugular venous lactate levels in both non-embolised

(Figure B, p=0.024) and embolised controls (Figure E, p=0.007). ∆lactate in Figures C and F

compares fetal scalp blood lactate and jugular venous lactate to carotid arterial lactate levels;

scalp-artery and vein-artery respectively. In the absence of cord occlusion fetal scalp blood

lactate levels are higher than both carotid arterial and jugular venous lactate levels, which are

similar.

253

Occluded Fetuses

-300

2

4

6

8

10

0 27 54 81

FetalScalp

CarotidArtery

Time (min)

Lact

ate

(mm

ol/L

)

Occluded Fetuses

-300

2

4

6

8

10

0 27 54 81

FetalScalp

CarotidArtery

Time (min)

Lact

ate

(mm

ol/L

)

-300

2

4

6

8

10

0 27 54 81

FetalScalp

JugularVein

Time (min)

Lact

ate

(mm

ol/L

)

-300

2

4

6

8

10

0 27 54 81

JugularVein

FetalScalp

Time (min)La

ctat

e (m

mol

/L)

Scalp Vein

-2

-1

0

1

2p=0.005 p=0.214

∆La

ctat

e (m

mol

/L)

Scalp Vein

-2

-1

0

1

2p=0.002 p=0.066

∆La

ctat

e (m

mol

/L)

NON-EMBOLISED EMBOLISEDA

B

C

D

E

F

Figure 5.14 The relationship between fetal scalp blood, carotid arterial blood and jugular

venous blood lactate levels in occluded fetuses. Fetal scalp blood lactate levels were

significantly higher than carotid arterial lactate levels in both non-embolised (Figure A,

p=0.005) and embolised occluded fetuses (Figure D, p=0.002). Similarly fetal scalp blood

lactate levels were significantly higher than jugular venous lactate levels in both non-

embolised (Figure B, p=0.002) and embolised occluded fetuses (Figure E, p<0.001). ∆lactate

in Figures C and F compares fetal scalp blood lactate and jugular venous lactate to carotid

arterial lactate levels; scalp-artery and vein-artery respectively. This figure suggests that

during cord occlusion scalp lactate levels are higher than carotid arterial levels which

themselves may be higher than jugular venous lactate levels.

253

Figure 5.15 Comparison of three lactate sampling sites in control and occluded non-

embolised (NE) and embolised (E) fetuses. Figure A illustrates that the relationship between

fetal scalp blood lactate and carotid arterial lactate levels is independent of both cord

occlusion (p=0.114) and umbilicoplacental embolisation (p=0.242) with fetal scalp lactate

overestimating carotid arterial lactate levels by 1.6mmol/L. Figure B illustrates that the

relationship between fetal scalp lactate and jugular venous lactate levels is independent of

cord occlusion (p=0.518) but modified by umbilicoplacental embolisation (p=0.006). The

scalp-vein lactate difference was greater in fetuses exposed to embolisation (1.9±0.2mmol/L)

than non-embolised fetuses (1.4±0.2mmol/L, p=0.006). Figure C illustrates the relationship

between jugular venous lactate levels and carotid arterial lactate levels was independent of

embolisation (p=0.385) but modified by cord occlusion with jugular venous lactate levels

significantly lower than carotid arterial lactate levels in occluded fetuses (p<0.001).

254

∆ Lactate (Scalp - Artery)

-2

-1

0

1

2

NEcontrol

NEoccluded

Econtrol

Eoccluded

∆La

ctat

e (m

mol

/L)

∆ Lactate (Scalp-Vein)

-2

-1

0

1

2

NEcontrol

NEoccluded

Econtrol

Eoccluded

∆La

ctat

e (m

mol

/L)

∆ Lactate (Vein - Artery)

-2

-1

0

1

2

NEcontrol

NEoccluded

Econtrol

Eoccluded

∆La

ctat

e (m

mol

/L)

A

B

C

255

levels increased with progressive asphyxia. In contrast, the effect of decreasing pH, PO2 and

base excess levels during asphyxia on the scalp-carotid lactate relationship was for scalpblood

lactate levels to overestimate arterial lactate levels. The analysis of interactions suggested that

the effect of arterial lactate levels on scalp lactate levels was less in the embolised than non-

embolised fetuses and, conversely, the influence of PCO2 levels was greater in embolised than

non-embolised fetuses. The net result of these combined effects were such that scalp lactate

levels reliably overestimated carotid arterial lactate levels by the same amount in all four

experimental groups.

5.3.2.1.2 Fetal scalp blood lactate levels and jugular venous lactate levels

Fetal scalp lactate levels were significantly higher than jugular venous lactate levels in all four

experimental groups (scalp-vein lactate: non-embolised control 1.4±0.6mmol/L, p=0.024;

non-embolised occluded 1.7±0.5mmol/L, p=0.002; embolised control 1.7±0.6mmol/L,

p=0.007; embolised occluded 2.2±0.5mmol/L, p<0.001, Figure 5.15B).

The relationship between fetal scalp blood and jugular venous lactate levels was not

influenced by cord occlusion (p=0.518), time during the experimental phase (p=0.152),

noradrenaline (p=0.374), adrenaline (p=0.361) or glucose levels (p=0.611) whereas it was

significantly modified by umbilicoplacental embolisation (p=0.006). The scalp-jugular lactate

difference in embolised fetuses (1.9±0.2mmol/L) was significantly greater than in non-

embolised fetuses (1.4±0.2mmol/L, p=0.006). Similar to carotid arterial blood, the

relationship between fetal scalp blood and jugular venous lactate levels was influenced by pH

(p=0.002) and lactate levels (p<0.001). The effect of decreasing fetal pH levels during

asphyxia was for fetal scalp lactate levels to overestimate jugular venous lactate levels and,

conversely, as lactate levels increased during UCO, scalp lactate levels underestimated jugular

lactate levels. The net result of these two combined effects was for fetal scalp lactate levels to

reliably overestimate jugular venous lactate levels. The analysis of interactions suggested that

the increased scalp-lactate difference in embolised fetuses compared to non-embolised

fetuses was secondary to a greater effect of pH (p=0.002) and a lesser effect of lactate

(p<0.001) in the embolised fetuses compared to the non-embolised fetuses.

5.3.2.1.3 Fetal carotid artery and jugular venous blood lactate levels

Fetal jugular venous lactate levels were not significantly different from carotid arterial lactate

levels in both groups of non-embolised fetuses and embolised controls (vein-artery lactate:

non-embolised control 0.0±0.2mmol/L,p=0.882; non-embolised occluded -0.3±0.2mmol/L,

p=0.214; embolised control -0.1 ±0.2mmol/L, p=0.575); however, in embolised fetuses

256

exposed to UCO, jugular venous lactate levels were lower than carotid arterial lactate levels

(embolised occluded -0.4± 0.2mmol/L, p=0.066 Figure 5.15C).

The relationship between jugular venous blood and carotid arterial blood lactate levels was

not influenced by embolisation (p=0.385) or time during the protocol (p=0.567); it was

significantly modified by cord occlusion with the jugular venous lactate levels significantly

less than carotid arterial lactate levels in fetuses exposed to UCO (p<0.001). These results

suggest that the fetal brain may be metabolising lactate as a source of fuel during asphyxia.

To investigate this hypothesis, further analysis using the blood sampled from the superior

sagittal sinus (SSS) was performed.

Lactate levels were not significantly different between the SSS and the jugular vein (SSS-

jugular vein, -0.10±0.10mmol/L, p=0.3020) nor were there any significant differences

between PCO2 (p=0.234) or base excess (p=0.274) between the two sampling sites (Table

5.6). Blood from the SSS was significantly more acidaemic (pH -0.010±0.002, p<0.001) and

had slightly higher oxygen levels (PO2 +0.104±0.41mmHg) than blood from the jugular vein.

There was no evidence that umbilicoplacental embolisation or cord occlusion modified the

relationship between lactate levels in blood collected from the SSS or jugular vein.

Table 5.6 Summary of differences between superior sagittal sinus blood and jugular venous blood

∆ SSS-jugular vein Group Effect Occlusion Effect

(mean±SEM) p-value

Lactate -0.10±0.10 0.302 0.212 0.558

pH -0.010±0.002 <0.001 0.559 0.991

PO2 1.04±0.41 0.015 0.351 0.883

PCO2 -2.27±1.88 0.234 0.232 0.933

Base excess -0.70±0.63 0.274 0.489 0.980

To assess fetal brain metabolism during asphyxia, the lactate and PO2 levels in blood collected

from the SSS and the carotid artery have been compared (Figure 5.16). Lactate levels in the

SSS were not significantly different from carotid arterial lactate levels in both control and

occluded non-embolised fetuses (SSS-carotid artery: non-embolised control 0.1±0.2mmol/L,

p=0.540; non-embolised occluded 0.1±0.2mmol/L, p=0.714; Figure 5.16A). However, in

256

Figure 5.16 Lactate and oxygen differences across the fetal cerebral circulation during cord

occlusion. Figure A is a box and whisker plot of ∆lactate levels* across the fetal cerebral

circulation. There was no evidence of a net change in lactate levels across the cerebral

circulation of the non-embolised fetuses; however, in the embolised fetuses SSS lactate levels

were significantly less than carotid arterial lactate levels (p<0.001) suggesting a net uptake of

lactate by the fetal brain. Figure B plots ∆PO2 across the fetal cerebral circulation in absolute

values and Figure C plots ∆PO2 across the fetal cerebral circulation as a fraction of the

available oxygen.

*∆lactate = carotid arterial lactate - SSS lactate

**∆PO2 = carotid arterial PO2- SSS PO2

***Fractional oxygen extraction = (carotid artery PO2 - SSS PO2) / carotid artery PO2

257

-4

-3

-2

-1

0

1

2

Control Occl Control Occl

p<0.001

EmbolisedNon-Embolised

p<0.01∆

Lac

tate

* (m

mol

/L)

0

2

4

6

8 p<0.001

Control Occl Control Occl

EmbolisedNon-Embolised

p<0.001

∆ P

O2*

* (m

mH

g)

Control Occl Control Occl

EmbolisedNon-Embolised

0

10

20

30p=0.003

p=0.071

Frac

tiona

l Oxy

gen

Extra

ctio

n***

(%)

A

B

C

258

fetuses that experienced umbilicoplacental embolisation, SSS lactate levels were significantly

less than carotid arterial lactate levels (SSS-carotid artery: embolised control -

1.4±0.4mmol/L, p=0.011; embolised occluded -1.8± 0.4mmol/L, p=0.002) suggesting that

the fetal brain may be using lactate as a source of fuel in these fetuses (Figure 5.16A). The

oxygen utilisation by the fetal brain was also altered by umbilicoplacental embolisation.

When non-embolised fetuses were exposed to hypoxia during UCO, the difference between

SSS PO2 levels and carotid arterial PO2 levels was significantly reduced in both absolute

(p<0.001) and fractional measures compared to controls (p=0.003; Figure 5.16 B and C). In

contrast, embolised fetuses had reduced oxygen extraction by the fetal brain (p<0.001) in

both control and occluded embolised fetuses when compared to non-embolised controls and

no further reduction in oxygen extraction occurred with the addition of UCO.

5.3.2.2 Fetal scalp blood pH levels

Fetal scalp, carotid arterial and jugular venous blood pH levels for both non-embolised and

embolised fetuses during the experimental phase are presented in an identical series of figures

to those presented for scalp lactate levels (Figures 5.17 – 5.19). In all four experimental

groups it can be seen that scalp blood pH was significantly higher than carotid arterial pH,

which itself was significantly higher than jugular venous blood pH levels (all p<0.001).

5.3.2.2.1 Fetal scalp blood pH levels and carotid arterial pH levels

Fetal scalp blood pH levels were significantly higher than carotid arterial pH levels in all four

experimental groups (scalp-artery pH: non-embolised control 0.058±0.011, p<0.001; non-

embolised occluded 0.048±0.010, p<0.001; embolised control 0.048±0.011, p<0.001;

embolised occluded 0.077±0.011, p<0.001, Figure 5.19A).

The relationship between fetal scalp blood pH and carotid arterial pH was not modified by

embolisation (p=0.895), or time during the protocol (p=0.184) whereas it was modified by

cord occlusion, but only in the embolised fetuses (p=0.034). This resulted in scalp pH levels

underestimating the severity of the acidaemia in the carotid artery by significantly more in

embolised occluded fetuses than non-embolised occluded fetuses. In both non-embolised

and embolised fetuses the relationship between fetal scalp blood pH and carotid arterial pH

was modified by arterial pH (p=0.005), lactate (p=0.003) and adrenaline levels (p=0.036).

The scalp-carotid pH relationship was not modified by any other blood gas variable,

noradrenaline (p=0.934) or glucose level (p=0.934) and no interactions between variables and

groups were identified. The effect of arterial pH and adrenaline levels on the scalp-carotid

259

Control Fetuses

-307.0

7.1

7.2

7.3

7.4

7.5

0 27 54 81

CarotidArtery

FetalScalp

Time (min)

pH

Control Fetuses

-307.0

7.1

7.2

7.3

7.4

7.5

0 27 54 81

CarotidArtery

FetalScalp

Time (min)

pH

-307.0

7.1

7.2

7.3

7.4

7.5

0 27 54 81

FetalScalp

JugularVein

Time (min)

pH

-307.0

7.1

7.2

7.3

7.4

7.5

0 27 54 81

FetalScalp

JugularVein

Time (min)pH

Scalp Vein-0.10

-0.05

0.00

0.05

0.10 p<0.001 p<0.001

∆pH

Scalp Vein-0.10

-0.05

0.00

0.05

0.10 p<0.001 p<0.001

∆pH

NON-EMBOLISED EMBOLISEDA

B

C

D

E

F

Figure 5.17 The relationship between fetal scalp blood, carotid arterial blood and jugular

venous blood pH levels in control fetuses. Fetal scalp blood pH levels were significantly

higher than carotid arterial lactate levels in both non-embolised (Figure A, p<0.001) and

embolised controls (Figure D, p<0.001). Similarly fetal scalp blood pH levels were

significantly higher than jugular venous pH levels in both non-embolised (Figure B, p<0.001)

and embolised controls (Figure E, p<0.001). ∆pH in Figures C and F compares fetal scalp

blood pH and jugular venous pH to carotid arterial pH levels; scalp-artery and vein-artery

respectively. This figure suggests that without cord occlusion fetal scalp blood pH levels are

significantly higher than carotid arterial pH levels which are significantly higher than jugular

venous pH levels.

260

Occluded Fetuses

-307.0

7.1

7.2

7.3

7.4

7.5

0 27 54 81

FetalScalp

CarotidArtery

Time (min)

pH

Occluded Fetuses

-307.0

7.1

7.2

7.3

7.4

7.5

0 27 54 81

CarotidArtery

FetalScalp

Time (min)

pH

-307.0

7.1

7.2

7.3

7.4

7.5

0 27 54 81

FetalScalp

JugularVein

Time (min)

pH

-307.0

7.1

7.2

7.3

7.4

7.5

0 27 54 81

FetalScalp

JugularVein

Time (min)

pH

Scalp Vein-0.10

-0.05

0.00

0.05

0.10 p<0.001 p<0.001

∆pH

Scalp Vein-0.10

-0.05

0.00

0.05

0.10 p<0.001 p<0.001

∆pH

NON-EMBOLISED EMBOLISEDA

B

C

D

E

F

Figure 5.18 The relationship between fetal scalp blood, carotid arterial blood and jugular

venous blood pH levels in occluded fetuses. Fetal scalp blood pH levels were significantly

higher than carotid arterial pH levels in both non-embolised (Figure A, p<0.001) and

embolised occluded fetuses (Figure D, p<0.001). Similarly fetal scalp blood pH levels were

significantly higher than jugular venous pH levels in both non-embolised (Figure B, p<0.001)

and embolised occluded fetuses (Figure E, p=0.011). ∆pH in Figures C and F compares fetal

scalp blood pH and jugular venous pH to carotid arterial pH levels; scalp-artery and vein-

artery respectively. The difference between scalp blood pH levels and carotid artery pH

levels is greater in the embolised occluded fetuses than the non-embolised occluded fetuses

(p<0.049).

260

Figure 5.19 Comparison of three pH sampling sites in control and occluded non-embolised

and embolised fetuses. Figure A illustrates the relationship between fetal scalp blood pH and

carotid arterial pH levels is independent of umbilicoplacental embolisation (p=0.895) but

altered by cord occlusion in the embolised fetuses (p=0.034). In fetuses exposed to

embolisation and cord occlusion, scalp pH underestimates carotid arterial pH by 60% more

than non-embolised fetuses exposed to UCO (p=0.049). Figure B illustrates the relationship

between fetal scalp pH and jugular venous pH levels is independent of umbilicoplacental

embolisation (p=0.895) but modified by cord occlusion in the embolised fetuses with

embolised fetuses having a significantly greater difference between scalp and venous pH

levels than non-embolised fetuses (p=0.034). Figure C illustrates the relationship between

jugular venous pH levels and carotid arterial pH levels which was independent of cord

occlusion (p=0.553) but modified by embolisation with the difference between jugular

venous and carotid arterial blood pH levels significantly less in embolised than non-

embolised fetuses (p=0.015).

261

∆ pH (Scalp - Artery)

-0.10

-0.05

0.00

0.05

0.10

p=0.049

NEcontrol

NEoccluded

Econtrol

Eoccluded

∆ p

H

∆ pH (Scalp-Vein)

-0.10

-0.05

0.00

0.05

0.10

NEcontrol

NEoccluded

Econtrol

Eoccluded

∆pH

∆ pH (Vein - Artery)

-0.10

-0.05

0.00

0.05

0.10

NEcontrol

NEoccluded

Econtrol

Eoccluded

∆ p

H

A

B

C

262

pH relationship was to increase the difference between fetal scalp pH and central pH during

asphyxia. Conversely the effect of arterial lactate levels on the scalp-carotid pH relationship

was to decrease the pH difference between the two sampling sites. The net result of these

combined effects was for scalp pH levels to be higher than carotid arterial pH levels (ie. to

underestimate the severity of the acidosis). As the embolised occluded fetuses had greater

falls in pH, larger increases in adrenaline, and the same lactate responses as non-embolised

fetuses during UCO, the greater difference in scalp and carotid arterial pH levels in

embolised fetuses is consistent with this explanatory model.

5.3.2.2.2 Fetal scalp blood pH levels and jugular venous blood pH levels

Fetal scalp blood pH levels were significantly higher than jugular venous pH levels in all four

experimental groups (scalp-vein pH: non-embolised control 0.080±0.011, p<0.001; non-

embolised occluded 0.078±0.010, p<0.001; embolised control 0.073±0.012, p<0.001;

embolised occluded 0.095±0.011, p<0.001, Figure 5.19B). The relationship between fetal

scalp blood and jugular venous blood pH levels was not influenced by embolisation

(p=0.895) or time during the protocol (p=0.184) but it was modified by cord occlusion;

although only in the embolised fetuses (p=0.034). Embolised fetuses exposed to UCO had a

significantly greater difference between scalp and venous pH levels than non-embolised

fetuses exposed to UCO.

5.3.2.2.3 Fetal carotid artery and jugular venous blood pH levels

Fetal jugular venous pH levels were significantly lower than carotid arterial pH levels in all

groups of fetuses (vein-artery pH: non-embolised control -0.022±0.002, p<0.001; non-

embolised occluded -0.029±0.002, p<0.001; embolised control -0.026±0.02, p<0.001;

embolised occluded -0.019 ±0.002, p<0.001, Figure 5.19C). The relationship between jugular

venous blood and carotid arterial blood pH levels was not influenced by occlusion (p=0.553)

or time during the protocol (p=0.154); however it was significantly modified by embolisation

with the difference between jugular venous and carotid arterial pH levels significantly smaller

in embolised than non-embolised fetuses (p=0.015).

263

5.3.3 Prediction of fetal arterial pressure changes during cord occlusion

Given the strong correlation between fetal arterial pressure responses during UCO and

neuronal injury demonstrated in Chapter Three, it may be clinically useful to predict these

arterial pressure responses during cord occlusion from either FHR patterns or fetal scalp

blood biochemistry.

5.3.3.1 Prediction of fetal arterial pressure responses during cord occlusion from

fetal heart rate patterns

The sensitivity, specificity and accuracy for FHR patterns to predict fetal arterial pressure

responses during cord occlusion are presented in Table 5.7 for both non-embolised and

embolised fetuses.

Table 5.7 Fetal heart rate prediction of arterial pressure responses during cord occlusion

Any Arterial pressure Instability †

Biphasic Systolic Response ‡

Biphasic Diastolic Response

FHR Pattern

Sens

itivi

ty

Spec

ificit

y

Acc

urac

y

Sens

itivi

ty

Spec

ificit

y

Acc

urac

y

Sens

itivi

ty

Spec

ificit

y

Acc

urac

y

NEδ 78% 51% 62% 96% 51% 62% 97% 47% 56%Severe Variable Decelerations E§ 81% 45% 66% 85% 38% 55% 98% 43% 62%

NE 36% 94% 71% 44% 91% 80% 45% 88% 81%Late component to decelerations E 65% 95% 78% 84% 87% 86% 98% 88% 92%

NE 34% 89% 68% 38% 85% 74% 30% 82% 73%Increased variability E 18% 98% 51% 14% 91% 62% 16% 91% 67%

NE 88% 29% 51% 89% 26% 41% 85% 24% 35%Overshoot

E 47% 64% 54% 54% 64% 60% 47% 60% 55%† Sum of decreasing hypertensive response, biphasic systolic response and biphasic diastolic response phases ‡ Sum of biphasic systolic response and biphasic diastolic response as both patterns had biphasic systolic pressure patterns δ NE=non-embolised § E=embolised

264

Severe variable decelerations, increased FHR variability between cord occlusions and FHR

overshoot after decelerations were all poor predictors of fetal arterial pressure responses

during UCO in both embolised and non-embolised fetuses. In embolised fetuses, increased

FHR variability and overshoot after decelerations were less predictive of arterial pressure

instability during UCO than in non-embolised fetuses (p<0.05; Table 5.8).

In contrast, late decelerations were moderate predictors of fetal arterial pressure responses

during UCO in non-embolised fetuses and excellent predictors of systolic and diastolic

hypotension during UCO in embolised fetuses with accuracy of 86% and 92% respectively

(Table 5.7). Embolised fetuses with late components to FHR decelerations were 6.81 times

(95%CI 2.71-17.12, p<0.001) more likely to have systolic hypotension during UCO and 60

times (95%CI 7.36-487.6, p<0.001) more likely to have diastolic hypotension during UCO

than non-embolised fetuses with the same FHR pattern (Table 5.8). The improved accuracy

of late decelerations at predicting hypotension during UCO in embolised fetuses is secondary

to the improved sensitivity in this group which was two fold higher than non-embolised

fetuses (Table 5.7).

Table 5.8 Odds Ratios of abnormal FHR pattern if fetal arterial pressure response is present

FHR Pattern Any Arterial

pressure Instability†

Biphasic Systolic Response‡

Biphasic Diastolic Response

Severe Variable Decelerations

OR§ (95%CI)

1.22 (0.57-2.63)

0.25 (0.05-1.24)

1.56 (0.09-25.90)

Late components to decelerations

OR (95% CI)

3.33*** (1.75-6.35)

6.81*** (2.71-17.12)

60.00*** (7.36-487.6)

Increased variability

OR (95% CI)

0.41* (0.20-0.85)

0.26** (0.10-0.69)

0.43 (0.15-1.23)

Overshoot OR

(95% CI)0.13***

(0.06-0.28) 0.14***

(0.05-0.42) 0.16***

(0.05-0.48)

§ All odds rations calculated embolised: non-embolised † Sum of decreasing hypertensive response, biphasic systolic response and biphasic diastolic response phases ‡ Sum of biphasic systolic response and biphasic diastolic response as both patterns had biphasic systolic pressure patterns * p <0.05, ** p <0.01, *** p<0.001

265

5.3.3.2 Prediction of fetal arterial pressure responses during cord occlusion from

fetal scalp blood biochemistry

Receiver operator curves (ROC) were used to assess the diagnostic utility of fetal scalp blood

biochemical measures to predict fetal arterial pressure responses during cord occlusion in

both non-embolised and embolised fetuses (Table 5.9). The areas under the ROCs for scalp

blood lactate were marginally higher than for scalp blood pH levels for all three patterns of

arterial pressure response during UCO for both non-embolised and embolised fetuses;

however, the magnitude of these differences did not reach statistical significance (Figure

5.20). Umbilicoplacental embolisation did not alter the areas under the ROCs for both scalp

blood lactate and pH levels for the three arterial pressure response patterns. In both non-

embolised and embolised fetuses the optimum cut off levels for scalp blood lactate increased

as the fetuses progressed through the three stages of cardiovascular decompensation whereas

the optimum cut off values for scalp pH were similar for the three patterns (Table 5.10).

Table 5.9 Biochemical prediction of arterial pressure responses during cord occlusion

Any Arterial pressure Instability †

Biphasic Systolic Response ‡

Biphasic Diastolic Response Biochemical

measure (fetal scalp blood)

Are

a

Sens

itivi

ty

Spec

ificit

y

Are

a

Sens

itivi

ty

Spec

ificit

y

Are

a

Sens

itivi

ty

Spec

ificit

y

NEδ 0.839 85% 69% 0.966 93% 88% 0.983 100% 92% Lactate

E§ 0.874 91% 70% 0.891 83% 83% 0.891 83% 83%

NE 0.721 46% 94% 0.909 93% 76% 0.965 100% 82% pH

E 0.735 71% 66% 0.865 83% 74% 0.856 83% 72% † Sum of decreasing hypertensive response, biphasic systolic response and biphasic diastolic response phases ‡ Sum of biphasic systolic response and biphasic diastolic response as both patterns had biphasic systolic pressure patterns δ NE=non-embolised § E=embolised

266

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 11 - Specificity (false positives)

Sens

itivi

ty (t

rue

posi

tives

)NodiscriminationScalp lactate

Scalp pH

0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 11 - Specificity (false positives)

Sens

itivi

ty (t

rue

posi

tives

)

NodiscriminationScalp Lactate

Scalp pH

Figure 5.20 Receiver operator curves for the prediction of any arterial pressure instability

during cord occlusion in non-embolised and embolised fetuses. Umbilicoplacental

embolisation had no effect on the ability of either scalp blood lactate or pH to predict fetal

arterial pressure responses during cord occlusion.

NON-EMBOLISED

EMBOLISED

267

Table 5.10 Accuracy of biochemical prediction of arterial pressure responses during UCO

Any Arterial pressure Instability†

Biphasic Systolic Response‡

Biphasic Diastolic Response Biochemical

measure

(fetal scalp blood) Accuracy Optimum Cut Off Accuracy Optimum

Cut Off Accuracy Optimum Cut Off

NEδ 77% 4.1 88% 4.5 93% 4.9 Lactate (mmol/L) E§ 75% 4.3 77% 4.7 77% 4.9

NE 84% 7.301 78% 7.308 93% 7.295 pH

E 67% 7.329 76% 7.316 76% 7.316 † Sum of decreasing hypertensive response, biphasic systolic response and biphasic diastolic response phases ‡ Sum of biphasic systolic response and biphasic diastolic response as both patterns had biphasic systolic pressure patterns δ NE=non-embolised § E=embolised

5.3.4 Summary of similarities and differences in fetal responses to asphyxia in

embolised and non-embolised fetuses

A summary of the similarities and differences between embolised and non-embolised fetuses

in biochemical, endocrine, cardiovascular and metabolic responses to asphyxia are presented

in Table 5.11.

Table 5.11 Summary of differences and similarities in physiologic responses to asphyxia in embolised and non-embolised fetuses

Similarities Differences

Arterial lactate and base excess responses pH, PCO2, PO2 responses

Relationship between scalp blood lactate and carotid arterial lactate levels

Relationship between fetal scalp blood pH and carotid arterial pH

Arterial pressure responses during cord occlusions

Arterial pressure responses between cord occlusions

Ability of lactate to predict arterial pressure responses during UCO

Ability of FHR patterns to predict arterial pressure responses during UCO

FHR variability and decelerative FHR patterns

Glucose, cortisol and catecholamine responses

Cerebral lactate and oxygen metabolism

268

5.4 Discussion

The purpose of the studies performed in this chapter was to investigate fetal lactate

responses during asphyxia after growth restriction induced by uteroplacental insufficiency.

These studies were performed because of the evidence obtained in experiments presented in

Chapters Three and Four of this thesis, and recently by others (Gardner et al., 2002; Gardner

et al., 2003), that previous exposure to an adverse intrauterine environment alters

cardiovascular, endocrine and metabolic responses to acute hypoxaemia in the ovine fetus.

In addition, growth restricted human fetuses are more likely than appropriately grown fetuses

to have non-reassuring FHR patterns (Low et al., 1976; Dashe et al., 2000) and fetal acidosis in

labour (Low et al., 1981; Nieto et al., 1994) and are at increased risk of adverse outcomes

secondary to acute hypoxic intrapartum events (Bukowski et al., 2003). Given that growth

restricted fetuses are more likely than normally grown fetuses to have adverse events during

labour and that they have altered physiologic responses to asphyxia, it seemed important to

specifically investigate the effect of growth restriction on the response to UCO. Such

investigation is necessary to evaluate the potential of fetal scalp lactate monitoring in this

important group of fetuses.

As in previous studies where umbilicoplacental embolisation has been used to induce

placental insufficiency (Clapp et al., 1981; Trudinger et al., 1987; Gagnon et al., 1994; Cock &

Harding, 1997; Murotsuki et al., 1997a; Duncan et al., 2000), I have found significant

reductions in fetal PO2 and pH throughout the duration of the embolisation period and there

was a tendency for the fetuses to be hypoglycaemic. Similar changes have been observed in

blood samples collected by cordocentesis in human growth restricted fetuses (Nicolaides et

al., 1989). Twenty three days of umbilicoplacental embolisation induced a 29% reduction in

fetal weight at 129 days gestational age in my study which is similar to previous reports using

similar embolisation techniques over a period of 25 to 30 days during gestation (Duncan et

al., 2000).

5.4.1 Stability of fetal lactate responses during asphyxia

This is the first experimental study using an animal model to investigate fetal lactate

responses during asphyxia induced by repetitive cord occlusion in growth restricted fetuses.

In this chapter, I have shown that umbilicoplacental embolisation for 23 days alters fetal acid-

base, endocrine, and cardiovascular responses to subsequent asphyxia in the ovine fetus. In

addition, evidence has been provided that suggests that lactate and oxygen metabolism in the

269

fetal brain during asphyxia is altered by previous umbilicoplacental embolisation. Despite the

multiple effects of umbilicoplacental embolisation on fetal responses to asphyxia,

embolisation did not alter fetal arterial lactate responses to asphyxia, the relationship between

fetal scalp blood and carotid arterial lactate levels or the ability of fetal scalp lactate levels to

predict fetal arterial pressure responses during cord occlusion.

5.4.1.1 Arterial lactate responses

The consistent increase in arterial lactate levels seen in this study in response to repeated cord

occlusion in both embolised and non-embolised fetuses suggests that the increase in lactate

levels during asphyxia more closely relate to the severity of the fetal insult than the condition

of the fetuses prior to asphyxia. Similar findings were reported in Chapters Three and Four

of this thesis. In Chapter Four (Section 4.3.3.2, Figure 4.6) both “normal” and

“spontaneously acidaemic” fetuses had the same lactate responses to repeated cord occlusion

whilst having different responses in the other markers of fetal acid-base status. In chronically

instrumented sheep fetuses in Chapter Three, “fast responding” fetuses were noted to have

higher baseline lactate levels than the other fetuses in the study; however, in response to

asphyxia they had lactate levels that increased in parallel to those of fetuses that commenced

the cord occlusion protocol with normal baseline lactate levels (Section 3.3.1.1, Figure 3.2).

The consistent increases in arterial lactate and base excess levels during asphyxia in both

embolised and non-embolised fetuses suggest the same degree of metabolic acidaemia is

present in both groups of fetuses. Conversely the two groups of fetuses had different

patterns of response in pH, PO2 and PCO2 during asphyxia. The more rapid fall in arterial pH

in the embolised fetuses during cord occlusion is likely to be due to the respiratory acidosis

associated with the greater increase in PCO2 in these fetuses. Both the more rapid rise in PCO2

and the greater fall in PO2 with repeated cord occlusion in embolised fetuses may have

resulted from a decrease in placental exchange of respiratory gases associated with the

placental damage induced by prolonged repeated embolisation. Umbilicoplacental

embolisation for 30 days from 110 days has been shown to result in a reduction in

placentome weight and cross sectional area with 20% of the fetal tissue in the placentomes

showing evidence of necrosis (Duncan et al., 2000).

The similar lactate levels during cord occlusion in the two groups of fetuses, despite more

rapid falls in PO2 in the embolised compared with non-embolised fetuses, suggests that the

fetuses that experienced umbilicoplacental embolisation either 1) increased oxygen extraction

to maintain tissue oxygen consumption, and/or 2) increased utilisation of lactate as a

270

significant metabolic substrate in embolised fetuses and/or 3) had altered placental handling

of lactate.

It is unlikely that the embolised fetuses in this study increased oxygen extraction when faced

with an asphyxial insult. In the short term (<24hours), the fetus has a considerable buffer to

reduced delivery of oxygen by increasing the efficiency of oxygen extraction from the blood.

Oxygen delivery can be reduced by up to 40-50% for durations between five minutes and 24

hours and a concomitant increase in oxygen extraction from 30% to 50-60% enables

maintenance of tissue oxygen consumption (Itskovitz et al., 1983; Edelstone, 1984; Edelstone

et al., 1985; Bocking et al., 1992), as long as fetal acidaemia does not develop (Rurak et al.,

1990). This ability of the fetus to compensate metabolically for reduced delivery of oxygen

has led to the concept of the margin of safety for fetal oxygenation (Meschia, 1985).

However, in studies in which the period of adverse intrauterine conditions has been

prolonged beyond 24 hours, there is evidence that long term reductions in the delivery of

oxygen to the fetus leads to a breakdown of the oxygen margin of safety and leads to down-

regulation of fetal oxygen-consuming processes (Anderson et al., 1986). In the study by

Anderson et al. (1986) fetal oxygen consumption was assessed in chronically instrumented

ovine fetuses during 10 days of reduced aortic blood flow. A prolonged reduction (≈40%) in

oxygen delivery to the fetus resulted in a strong correlation between total oxygen delivery to

the fetus and oxygen consumption by the fetus (Anderson et al., 1986). In this study fetal

oxygen consumption was not formally assessed; however, a limited assessment of oxygen

consumption in the fetal brain was performed. This suggested that fractional oxygen

extraction was reduced by approximately 30% in embolised fetuses when compared to non-

embolised fetuses (Figure 5.16C) and when challenged with an acute asphyxial insult,

embolised fetuses did not increase fractional oxygen extraction across the fetal brain.

Interestingly, fractional oxygen extraction did not fall when embolised fetuses were exposed

to repeated cord occlusion. At the completion of the embolisation phase PO2 levels in the

embolised fetuses were approximately 12mmHg. When these fetuses were acutely

instrumented to enable the collection of fetal scalp blood samples, fetal PO2 levels increased

to approximately 23mmHg, probably in response to the maternal ventilation. The

subsequent repeated cord occlusion resulted in fetal PO2 levels falling to approximately the

same levels (≈12-13mmHg) as those that were present during the embolisation phase. This

reduction may not have been adequate to stimulate a change in oxygen extraction as the

fetuses had been accustomed to these oxygen levels for at least 20 days prior to the insult.

Alternatively, it may not be physiologically possible for the fetus to decrease oxygen

271

extraction to levels below 15%, the lowest level seen in this study in both non-embolised and

embolised fetuses during cord occlusion.

It is possible that the similar arterial lactate responses in embolised and non-embolised

fetuses during UCO, despite more severe hypoxia in the embolised fetuses, are due to

increased utilisation of lactate as a metabolic substrate in embolised fetuses. In the sheep

fetus lactate metabolism accounts for 40-50% of the total fetal oxygen consumption and it is

the main carbohydrate taken up by the liver and heart for glycogen and lipid synthesis

(Fowden, 2001). Fetuses exposed to umbilicoplacental embolisation have a trend to

hypoglycaemia and it is known that when fetal glucose availability is limited, glucose

utilisation decreases and fetal oxygen consumption is maintained by oxidising more amino

acids resulting in urea production and a reduced rate of protein synthesis by the fetus

(Fowden, 2001). Little is known about lactate metabolism in fetuses exposed to chronic

hypoxia; however I have provided evidence to suggest increased utilisation of lactate by the

fetal brain as a source of metabolic fuel in fetuses exposed to umbilicoplacental embolisation

(Figure 5.16A). Lactate utilisation by other fetal tissues was not formally assessed in this

study.

The similar arterial lactate responses between the two groups of fetuses may be secondary to

altered placental handling of lactate, which was not assessed in this study. It has been shown

that during periods of 24 hours of adverse intrauterine conditions, the placenta acts as a

major site of lactate clearance from the fetal circulation (Hooper et al., 1995). It is not known

if the placenta continues to clear fetal lactate during periods of adverse intrauterine

conditions lasting beyond 24 hours. Placental weight is reduced by umbilicoplacental

embolisation (Duncan et al., 2000) and I have previously suggested that this may impair

oxygen and carbon dioxide transport across the placenta. Lactate transport across the

placenta is probably through facilitated diffusion (Carstensen et al., 1983; Piquard et al., 1990)

and the effect of umbilicoplacental embolisation on placental lactate handling is unknown.

Of the three potential mechanisms responsible for the uniform response in fetal arterial

lactate levels during asphyxia in the two groups of fetuses experiencing cord occlusion in this

study, I favour the theory that embolised fetuses have increased utilisation of lactate as a

metabolic substrate. It is likely that the embolised fetuses have down-regulation of oxygen-

consuming processes in the whole fetal body which results in an increase in tissue lactate

production. The effect of this on circulating lactate levels may be partially minimised by the

30% reduction in fetal body weight in embolised fetuses; however, increased utilisation of

lactate by the fetal heart, brain, liver or other tissues may result in a net lactate response

272

which is similar to normally grown fetuses. To evaluate the mechanisms responsible for the

arterial lactate responses in the growth restricted fetus, further studies using appropriately

instrumented chronically catheterised fetuses are required to formally assess fetal oxygen

utilisation during cord occlusion, fetal lactate metabolism and clearance of lactate by the

placenta.

5.4.1.2 Fetal scalp blood lactate responses

In the studies presented in this chapter, the relationships between fetal scalp blood lactate

levels and carotid arterial lactate levels were independent of embolisation and cord occlusion.

This observation suggests that fetal scalp lactate levels reliably reflect central arterial lactate

levels across a range of fetal conditions, both at rest and during asphyxia. Conversely,

exposure to an adverse intrauterine environment of chronic hypoxia impaired the ability of

fetal scalp blood pH levels to accurately reflect central arterial pH levels during an acute

asphyxial challenge. The addition of asphyxia to prolonged umbilicoplacental embolisation

resulted in the difference in pH levels between the two sampling sites increasing by 60% with

fetal scalp blood pH levels underestimating the severity of changes in central pH levels by a

greater amount in fetuses exposed to embolisation than controls. The effect of pre-asphyxial

umbilicoplacental embolisation on the accuracy of fetal scalp blood pH assessment has not

been reported previously; however, others have reported inaccuracies in fetal scalp blood pH

assessment in the presence of acute hypoxaemia (Shier & Dilts, 1972) and acute acidaemia

induced by hypercarbia (Aarnoudse et al., 1981). Considering the results of this study and

those by other authors (Shier & Dilts, 1972; Aarnoudse et al., 1981) it appears that the

accuracy of scalp pH assessment may be impaired in pregnancies complicated by acute

hypoxaemia, acidaemia and growth restriction. This is a significant limitation to the use of

scalp pH measurement to assess fetal wellbeing as it appears that the ability of scalp pH

measurements to accurately reflect central fetal acid base status is poor at the times when the

most accurate results are required.

The mechanism responsible for the divergent responses in scalp lactate and pH during

asphyxia in the two groups of fetuses in this study is unknown and not immediately apparent;

however, a possible explanation can be obtained from a review of acid-base physiology.

Hydrogen ions are highly reactive and small changes in pH of 0.1 to 0.2 pH units can have

profound effects on metabolic activity and on cardiovascular and central nervous system

function (Gillstrap, 1999). Dramatic changes in fetal pH are minimised via the action of

buffers. The two major buffers are plasma bicarbonate and haemoglobin; other

quantitatively less important buffers include erythrocyte bicarbonate and inorganic

273

phosphates (Blechner, 1993). In the fetus the placental bicarbonate pool may also play a role

in buffering the fetus against changes in maternal pH. In this study, fetal scalp blood lactate

levels were higher than in paired carotid arterial samples suggesting local production of

lactate by tissues in the fetal scalp. Conversely pH levels in the fetal scalp were higher than

carotid arterial pH levels, underestimating the falls in arterial pH by 0.04 to 0.08 pH units.

The difference in pH measures between the two sampling sites may reflect improved tissue

pH buffering in the embolised fetuses in this study. The embolised fetuses were hypoxaemic

and acidaemic for at least 20 days prior to their acute asphyxial insult. It is likely that during

this time the ability of the fetus to buffer acute changes in tissue pH would be improved as

part of the cascade of physiological events designed to protect the fetus in an adverse

intrauterine environment. Therefore, when the embolised fetuses were exposed to an acute

asphyxial insult associated with more dramatic changes in fetal pH (7.323 to 7.111) than that

experienced during the embolisation phase, improved tissue buffering of locally produced

hydrogen ions would result in scalp blood samples underestimating the pH levels in arterial

blood. Other evidence to support the hypothesis that embolised fetuses have improved pH

buffering during acute asphyxia is provided by the comparisons between carotid arterial,

jugular venous and superior sagittal sinus pH levels during asphyxia. Despite the more

dramatic fall in arterial oxygen content in the embolised fetuses, the net output of hydrogen

ions from the fetal brains of embolised fetuses was significantly less than in non-embolised

fetuses again suggesting improved tissue pH buffering in fetuses exposed to prolonged

embolisation.

Therefore, considering fetal scalp blood lactate and pH assessment, scalp blood lactate levels

appear to be the more reliable marker of central arterial acid-base status than scalp blood pH

assessment with inaccuracies in scalp blood pH assessment occurring in fetuses exposed to

prolonged periods of hypoxaemia prior to asphyxia.

5.4.1.3 Lactate and the prediction of arterial pressure responses during cord

occlusion

In this study I have shown that the ability of fetal scalp blood lactate levels to predict arterial

pressure instability during umbilical cord occlusion is not altered by umbilicoplacental

embolisation prior to the insult, suggesting that fetal scalp lactate levels are a relatively robust

marker of fetal brain perfusion during cord occlusion. This observation has not previously

been reported but is consistent with observations made in adult critical care medicine where

lactate levels have been shown to be reliable markers of tissue perfusion across a wide range

274

of conditions including acute trauma, major surgery and septic shock (Perret & Enrico, 1978;

Bakker et al., 1996; Husain et al., 2003).

In Chapter Three, I provided evidence which suggested that the duration of cardiovascular

instability during cord occlusion was the strongest predictor of fetal brain injury 72 hours

after repeated UCO, accounting for 91% of the variability in grading of neuronal injury.

Therefore the ability of scalp lactate levels to reliably predict arterial pressure instability

during cord occlusion across a range of fetal conditions is reassuring if scalp lactate is to be

used as a tool for the assessment of fetal wellbeing during labour where the fetal growth

status prior to delivery is often unknown. The study described in this chapter was not aimed

at determining the link between fetal arterial pressure changes during cord occlusion and the

severity of brain injury in the growth restricted fetus. Further studies are required to address

this issue.

5.4.2 Variation in cardiovascular and endocrine responses during asphyxia

In the studies presented in this chapter I have shown that placental insufficiency induced by

umbilicoplacental embolisation alters cardiovascular, endocrine and metabolic responses to

asphyxia. Similar observations have been made by others using different techniques to

induce an adverse uterine environment (Gardner et al., 2002; Gardner et al., 2003). When

considering the differences identified in cardiovascular and endocrine responses during

asphyxia reported in this chapter, one must remember that the repeated cord occlusions were

performed in an acute fetal preparation. In Chapter Four I demonstrated that the specific

acute preparation used in this study resulted in fetal responses which were similar to those

seen in chronic fetal preparations but differences were noted between the two models in

acid-base and noradrenaline responses to asphyxia and in baseline measures of FHR

variability. Despite these limitations, the alterations in cardiovascular and endocrine

responses described in this chapter warrant discussion and strongly suggest that further

appropriately designed studies are required to investigate the cardiovascular and endocrine

responses of growth restricted fetuses during repeated cord occlusion.

5.4.2.1 Fetal heart rate patterns

I have shown that fetuses exposed to umbilicoplacental embolisation exhibit different FHR

responses during repeated umbilical cord occlusion than non-embolised fetuses. The

differences observed between the two groups occurred both between cord occlusions and

during cord occlusions. Despite the clinical importance of fetal growth disorders and the

increased perinatal morbidity and mortality rates in such fetuses, this is the first experimental

275

study that used an animal model to investigate FHR patterns during asphyxia induced by

repetitive cord occlusion in growth restricted fetuses.

The fetal heart rate patterns in growth restricted fetal sheep during umbilicoplacental

embolisation have been described previously in detail (Murotsuki et al., 1997a). During the

first 48 hours of hypoxaemia secondary to embolisation, the number of accelerations and

decelerations increased compared to controls. In addition, both short- and long-term FHR

variability increased initially which was followed by a return to baseline after 20 hours of

hypoxaemia. After 21 days of embolisation fetuses had a 30% reduction in the number of

FHR accelerations and lower FHR variability than controls (Murotsuki et al., 1997a).

5.4.2.1.1 Fetal heart rate variability

In this study, fetuses exposed to embolisation had reduced FHR variability when compared

to non-embolised fetuses; however, when subjected to repeated UCO, embolised fetuses

were still capable of increasing their FHR variability, although the magnitude of this response

was significantly smaller than seen in non-embolised fetuses during asphyxia. Reduced FHR

variability has previously been reported in both growth restricted sheep (Gagnon et al., 1996;

Murotsuki et al., 1997a) and humans (Gagnon et al., 1988; Snijders et al., 1992) but the

responses in FHR variability during cord occlusion in growth restricted fetuses has not been

described previously.

Fetal sheep normally increase their FHR variability with advancing gestation from 106 to 129

days gestational age (Murotsuki et al., 1997a). Ovine fetuses that were subjected to placental

embolisation over the same gestation did not increase their FHR variability leading the

authors to suggest that the “inherently low” FHR variability in the growth restricted fetuses

may reflect a delay in maturation of the autonomic control of the fetal heart as opposed to a

reduction in FHR variability in response to chronic hypoxaemia (Murotsuki et al., 1997a). A

similar response has been shown in healthy human fetuses, in which there is an increase in

FHR variability between 28 and 30 weeks gestation (Gagnon et al., 1987), suggesting there is

functional maturation of the autonomic control of the fetal heart occurring early in the third

trimester. Longitudinal studies of FHR variation in growth restricted fetuses have shown

that in this group, FHR variability does not change over time and it can be reduced for up to

five weeks before the occurrence of “fetal distress” (Snijders et al., 1992). Theories for the

reduced heart rate variability seen in fetuses that have experienced embolisation include: 1)

alterations in the central control of heart rate variability associated with neuronal hypoxia or

damage and/or 2) decreased number or sensitivity of ß-adrenoreceptors in the fetal heart

276

during chronic placental insufficiency possibly associated with chronically elevated levels of

catecholamines.

The mechanisms which underlie the control of normal FHR variation are not fully

understood (Kozuma et al., 1997). In primates, variability is primarily transmitted via

parasympathetic nerves (Parer, 1999). Variability is likely due to numerous sporadic inputs

from various areas of the cerebral cortex and lower centres to the cardiac integratory centres

in the medulla oblongata which are then transmitted down the vagus nerve (Parer et al.,

1981). It has been hypothesized that hypoxic-tolerant cells such as neurons respond to

oxygen lack by orchestrating a series of molecular defence processes to reduce ATP demand

for ion pumping. These responses include a 90% or greater global decline in protein

biosynthesis (translational arrest), and a generalised decline in membrane permeability

(“channel arrest”) or in firing frequency (“spike arrest”) in nervous tissue (Hochachka et al.,

1996). When these defence mechanisms are overcome tissue injury may occur. The

potentially reversible decrease in firing frequency of the cortical and subcortical neurons, or

the irreversible brain injury associated with prolonged hypoxia may reduce sporadic inputs to

the medulla resulting in decreased fetal heart rate variability. Although I did not examine the

brains of the fetuses in this study, others using the same umbilicoplacental embolisation

protocol as used in this study have demonstrated fetal brain injury associated with the

technique (Mallard et al., 1998; Duncan et al., 2000). Similarly, in humans, severe growth

restriction is associated with fetal brain injury (Gaffney et al., 1994c; Grafe, 1994).

In this study there is evidence that the reduced heart rate variability seen after

umbilicoplacental embolisation may be secondary to alterations in the number or sensitivity

of ß-adrenoreceptors in the fetal heart, possibly associated with chronically elevated levels of

catecholamines. When compared to controls, fetuses exposed to embolisation had reduced

FHR variability prior to UCO, a smaller increase in FHR variability in response to UCO, and

a smaller and less sustained increase in inter-occlusion heart rate with progressive UCO.

These responses occurred despite a two-fold greater increase in noradrenaline levels and a

nine-fold greater increase in adrenaline levels in response to cord occlusion in the embolised

fetuses than in non-embolised fetuses (Figure 5.5).

Both embolised and non-embolised fetuses in this study increased their FHR variability when

exposed to repeated cord occlusion. This response has not been shown before in fetuses

exposed to embolisation. An increase in FHR variability has been shown previously in

studies of inhalational hypoxia in fetal sheep (Dalton et al., 1977; Parer et al., 1980; Lilja et al.,

1984; Field et al., 1991; Kozuma et al., 1997) and monkeys (Ikenoue et al., 1981), placental

277

embolisation in fetal sheep (Gagnon et al., 1996; Murotsuki et al., 1997a) and umbilical cord

occlusion in the ovine fetus (Westgate et al., 1999) and rat (Syutkina, 1988). It is also

consistent with the increase in baseline variability seen in humans in some cases of hypoxia

(Pello et al., 1991). The mechanisms which underlie the increase in FHR variability are not

fully understood but are likely to involve a combination of activation of the carotid

chemoreceptor reflex (Kozuma et al., 1997) and sympathetic stimulation (Dalton et al., 1983;

Rosen et al., 1986; Lindecrantz et al., 1993; Segar et al., 1994). The increase in FHR variability

seen in the embolised fetuses in this study suggests that the carotid body chemoreceptor

reflexes and/or sympathetic stimulation are still capable of responding to a more severe

hypoxic challenge as indicated by increased FHR variability. It is also possible that the

mechanisms responsible for the increase in FHR variability in response to hypoxia may be

different in embolised and non-embolised fetuses. This tenet is supported by the

multivariate logistic regression performed in this chapter which found complex and divergent

associations with increased variability in the two groups of fetuses. The only significant

association identified in the non-embolised fetuses was with inter-occlusion arterial pressure

whereas in the embolised fetuses changes in FHR variability were also associated with fetal

acid-base status.

5.4.2.1.2 Fetal heart rate decelerations

In this study fetuses exposed to embolisation developed late components to their FHR

decelerations and severe variable decelerations earlier and in a more variable manner than

non-embolised fetuses. Human growth restricted fetuses have also been shown to have

twice the incidence of FHR abnormalities in labour (Low et al., 1976; Low et al., 1981). There

are a number of potential mechanisms that may account for the earlier decelerative patterns

in growth restricted fetuses undergoing repetitive cord occlusion. Fetal umbilicoplacental

embolisation results in occlusion of small arterioles of the fetal stem villi and is associated

with a reduction in total placental flow, elevated placental vascular resistance and fetal

hypoxaemia (Gagnon et al., 1996). During umbilical cord occlusion transient falls in fetal

arterial PO2 of approximately 7mmHg are known to occur (Richardson et al., 1996). Given

that the embolised fetuses in this study had lower PO2 levels than the non-embolised fetuses

between the cord occlusions, it is likely that their PO2 levels were also lower during the

occlusions, which may account for the earlier decelerative patterns in the growth restricted

fetuses. Both hypertension and myocardial hypertrophy have also been shown to be induced

by embolisation of the utero-placental vascular bed (Murotsuki et al., 1997b). Increased

cardiac muscle mass may be less tolerant of hypoxaemia, possibly contributing to the

presence of late decelerations earlier in the cord occlusion protocol.

278

Late FHR decelerations, with their delay in onset and resolution in relation to uterine

contractions, are one of the sentinel features that clinicians aim to identify during continuous

intrapartum electronic FHR monitoring. Late decelerations are primarily a reflex change that

occurs during non-acidaemic hypoxia (Martin et al., 1979), but when hypoxia is severe enough

to result in acidaemia, the mechanism of late deceleration appears to be non-reflex

myocardial depression (Freeman et al., 1991). The results of the multivariate analysis in this

study suggest that the presence of late components to FHR decelerations was associated with

increases in lactate concentration and systolic pressure, and decreases in base excess in both

normally grown and growth restricted fetuses. The analysis of interactions also found an

additional association between late components to decelerations and arterial pH in the

embolised fetuses suggesting that late decelerations may also be associated with respiratory

acidosis in embolised fetuses.

Both the ROC analyses investigating the optimum cut off points for the significant

associations in the multivariate analysis (Table 5.5) and the analyses comparing late

decelerations and fetal arterial pressure responses during cord occlusion (Table 5.8) suggest

that late components to decelerations in the embolised fetus have more sinister implications

than when they occur in non-embolised fetuses. The results of the ROC analysis suggest that

late decelerations do not reliably begin in embolised fetuses until lower base excess levels

than those in non-embolised fetuses have been reached. In addition, when late decelerations

are present in fetuses exposed to embolisation they are six times more likely than non-

embolised fetuses to have systolic hypotension during umbilical cord occlusion and 60 times

more likely to have systolic and diastolic hypotension during the cord occlusion.

5.4.2.1.3 Fetal heart rate accelerations and overshoot

Accelerations in FHR during intrapartum monitoring have been found to be an indicator of

good perinatal outcome in two cohort studies where the presence of more than two

accelerations in 20 minutes had a sensitivity of 97% for an Apgar score of greater than seven

at five minutes (Powell et al., 1979; Krebs et al., 1982). There are thought to be at least two

physiologic mechanisms responsible for FHR accelerations. The first is the response to fetal

movement and the second mechanism is the response to partial umbilical cord occlusion.

During the 0.5-minute cord occlusions in my study, FHR accelerations were seen in

approximately half the fetuses experiencing cord occlusion (both embolised and non-

embolised) during total cord occlusion confirmed by simultaneous umbilical artery Doppler

assessment. No accelerations were seen in the control fetuses. The multivariate analysis used

in this study suggests that FHR accelerations were present during total cord occlusion when

279

there were falls in fetal PO2 prior to the systolic pressure increase and changes in pH, PCO2

and base excess that occur with repeated cord occlusion. Once the latter changes occurred,

they exerted a significant negative effect on the presence of accelerations with cord occlusion.

The net result was that accelerations were no longer present and were replaced by

decelerative FHR patterns. The results of my analysis suggest that recurrent accelerations

during total umbilical cord occlusion may not be as reassuring as the more common type of

acceleration which occurs in response to fetal movement.

In this study, both non-embolised and embolised occluded fetuses displayed FHR overshoot

patterns with the FHR pattern occurring significantly later in embolised fetuses. As discussed

at length in Chapter 3 (Section 3.4.6.2), the aetiology of FHR overshoot is unknown.

Goodlin et al. have suggested that the potential mechanisms for this heart rate pattern may

include a combination of high levels of circulating catecholamines and impaired neural

control of fetal heart rate (Goodlin & Lowe, 1974). No significant associations were

identified in the multivariate analysis between FHR overshoot and fetal biochemical acid-

base status, catecholamines or arterial pressure in this study. The delayed appearance of this

FHR pattern in fetuses exposed to embolisation may suggest that decelerations associated

with hypoxia and acidosis may have a greater impact on the fetal heart rate than the

overshoot in fetal heart rate associated with catecholamine release. Alternatively it may be

another reflection of decreased sensitivity to circulating catecholamines in fetuses exposed to

embolisation.

5.4.2.2 Fetal arterial pressure responses

In this study fetuses exposed to 23 days of umbilicoplacental embolisation had higher arterial

pressures than controls. Elevations in fetal arterial pressure have been reported previously

after as little as one day of umbilicoplacental embolisation at 109 days gestational age

(Murotsuki et al., 1997a) and fetal arterial pressures remain elevated throughout daily

umbilicoplacental embolisation until approximately 130 days gestational age (Cock &

Harding, 1997; Murotsuki et al., 1997a). When umbilicoplacental embolisation has been

continued to 140 days gestational age, arterial pressures in embolised fetuses were lower than

controls (Cock & Harding, 1997). This pattern is probably secondary to the progressive rise

in arterial pressures in control fetuses with advancing gestational age whereas in embolised

fetuses arterial pressure remains relatively constant with advancing gestation after its initial

elevation secondary to embolisation.

280

The additional hypoxic challenge of repeated cord occlusion produced an increase in arterial

pressures in non-embolised fetuses whereas in embolised fetuses there were no significant

changes in inter-occlusion arterial pressure during 81 minutes of repeated cord occlusion

despite a significantly larger increase in circulating catecholamines in these fetuses. These

results are different to those reported after a sub-acute hypoxic stress (acute severe

umbilicoplacental embolisation to reduce fetal pH to 7.0) was applied to fetuses that had

undergone daily umbilicoplacental embolisation between 120 and 130 days (Gagnon et al.,

1997). In the study by Gagnon et al. (1997), fetal arterial pressures increased in chronically

embolised fetuses that were exposed to acute severe embolisation. The different results

between this previous study and my own may be secondary to the different durations of

embolisation or the different insults used to induce the acute on chronic hypoxic stress.

It has been suggested that the elevation in fetal arterial pressure during prolonged

umbilicoplacental embolisation results from α-adrenoreceptors remaining responsive to

chronic fetal hypoxaemia and to chronic elevations in fetal plasma noradrenaline (Murotsuki

et al., 1997a). In my study and in others (Gagnon et al., 1994; Murotsuki et al., 1997a;

Murotsuki et al., 1997b), umbilicoplacental embolisation was associated with small increases

in plasma catecholamine concentrations. However, in my study when the fetuses subjected

to umbilicoplacental embolisation were exposed to large increases in noradrenaline levels and

very large increases in adrenaline levels during repeated cord occlusion (Figure 5.5), no

further increase in arterial pressure occurred. This may suggest that with chronic hypoxia the

function of adrenergic nerves in fetal arteries may decline as has been shown to occur with

chronic high-altitude hypoxia (Buchholz & Duckles, 2001). Alternatively it may suggest that

fetal arterial pressure is controlled by different mechanisms in growth restricted fetuses such

as the renin-angiotensin system, which has been found to have an important role in

controlling arterial pressure in fetuses in which growth restriction was induced by

carunclectomy of ewes (Edwards et al., 1999).

5.4.2.3 Fetal endocrine responses

The effects of umbilicoplacental embolisation on fetal glucose, cortisol and catecholamine

concentrations described in this study are similar to those reported previously using similar

embolisation techniques (Gagnon et al., 1994; Cock & Harding, 1997; Murotsuki et al., 1997a;

Murotsuki et al., 1997b). Similarly, the increases in catecholamines and glucose in response to

an acute asphyxial insult in embolised and control fetuses are similar to those previously

described (Gagnon et al., 1994). In my study, when fetuses subjected to 23 days of

umbilicoplacental embolisation were exposed to asphyxia induced by repeated cord

281

occlusion, no increase in fetal cortisol concentrations occurred despite a three fold increase in

ACTH. In contrast, non-embolised fetuses exposed to cord occlusion had the same increase

in ACTH but fetal cortisol levels increased two fold (Figure 5.6). This observation suggests

that the adrenal cortex is less responsive to ACTH after prolonged umbilicoplacental

embolisation. Short-term umbilicoplacental embolisation for eight days has been shown to

not alter the responsiveness of the adrenal cortex to ACTH (Carter et al., 1996) and an acute

hypoxic challenge after ten days of umbilicoplacental embolisation resulted in the same

increase in cortisol as controls exposed to acute hypoxia (Gagnon et al., 1994). Fetal adrenal

responses after more prolonged periods of embolisation have not been studied and require

evaluation in future studies using this technique to induce fetal growth restriction.

5.4.3 Influence of the experimental model of fetal scalp blood biochemistry

In this study the relationship between fetal scalp blood lactate levels and carotid arterial

lactate levels in non-embolised controls was different from that observed in controls

described in Chapter Four. In this chapter, fetal scalp blood lactate levels were higher than

carotid arterial lactate levels (≈1.4mmol/L), whereas in Chapter Four there was no significant

difference in lactate levels between the two sampling sites. These results suggest that the

addition of an episode of fetal surgery at 101 days gestational age, the presence of a fetal

arterial catheter for 28 days, and 28 days in a metabolic cage influenced the relationship

between fetal scalp blood and carotid arterial lactate levels.

On careful review of the data from the controls in this study it can be seen that the fetal PO2

levels in controls during the embolisation phase were between 17 and 20mmHg, which is

slightly lower than reported in controls in other studies using umbilicoplacental embolisation

(Gagnon et al., 1994; Cock & Harding, 1997; Murotsuki et al., 1997a). Mild fetal hypoxaemia

of similar levels to those seen in controls in this study (≈17mmHg) has been shown to alter

fetal pressor responses and peripheral vascular resistance (Gardner et al., 2002). In addition,

the adrenaline and noradrenaline levels reported at the end of the embolisation period in

controls in this study, although similar to controls in other similar studies (Murotsuki et al.,

1997a), were higher than those reported in fetuses after five (Gardner et al., 2002) and ten

days of instrumentation (Gagnon et al., 1997). The combined effects of increased peripheral

vascular resistance, increased fetal pressor responses and mildly elevated catecholamine levels

are likely to result in vasoconstriction in the fetal scalp and increased local tissue lactate

production. Similar increases in hind limb lactate production have been described after

transient hypoxaemia (Gardner et al., 2003) and the multivariate regression analyses

282

investigating factors influencing the relationship between scalp and carotid arterial blood

lactate levels performed in both this chapter and Chapter Four support the hypothesis that

local scalp tissue production of lactate occurred in this study, but not in the studies presented

in Chapter 4.

Considering the data from both chapters, the control data from Chapter Four where the

fetuses were only instrumented at the time of cord occlusion, is probably the most reliable.

In addition, the differences between fetal scalp blood and carotid arterial lactate identified in

Chapter Four are consistent with the limited human data that are available relating to

sampling from the two sites. The differences identified between fetal scalp biochemistry in

this chapter and Chapter Four highlight the difficulties involved in studies collecting fetal

scalp blood. They also provide strength to studies investigating the relationship between fetal

scalp blood biochemical levels and outcomes directly as the relationships between scalp and

central blood are complex and can be altered by experimental design.

5.4.4 Integration of results into clinical practice

In attempting to integrate the results of the studies performed in this chapter into clinical

practice the question arises as to what extent the events detailed in laboratory animals are

applicable to humans. Although it would be optimal to obtain the data presented in this

chapter from humans, it is clearly not possible to obtain repeated invasive samples in utero in

humans nor is it ethical to induce asphyxia.

The studies performed in this chapter suggest that the relationship between fetal scalp blood

lactate levels and carotid arterial lactate levels is robust and not influenced by fetal growth

status or asphyxia in the ovine fetus. Conversely the accuracy of fetal scalp pH assessment,

our current marker of fetal biochemical status during labour, is adversely affected by asphyxia

in growth restricted ovine fetuses. In addition, the studies presented in this chapter suggest

that fetal growth status does not adversely affect the ability of scalp blood lactate levels to

predict cardiovascular instability during cord occlusion, a marker that I have identified as

being strongly correlated with the severity of neuronal injury after repeated cord occlusion

(Section 3.3.5.3.3). Taken together these observations suggest that fetal scalp blood lactate

levels may be a more reliable predictor of anaerobic metabolism in human fetuses during

asphyxia across a range of fetal conditions and therefore should be a better predictor of fetal

outcome. This conclusion is consistent with a large observational study in humans which

suggested that fetal scalp lactate assessment was a better predictor of moderate-severe

283

neonatal encephalopathy and low Apgar scores at five minutes than scalp pH assessment

(Kruger et al., 1999).

The analyses in this study suggest that late FHR decelerations should be considered

particularly ominous in growth restricted fetuses. The multivariate analyses investigating

associations with FHR patterns and the ROC analysis of the significant covariates suggested

that late decelerations were more strongly associated with metabolic acidosis in growth

restricted fetuses than in normally grown fetuses. In addition, in growth restricted fetuses

late FHR decelerations were more closely associated with cerebral hypotension during cord

occlusion than in normally grown fetuses. In humans, late FHR decelerations are one of the

sentinel features that clinicians aim to identify during continuous intrapartum electronic FHR

monitoring and they occur more frequently in growth restricted human fetuses during labour

(Low et al., 1976; Dashe et al., 2000). The results of this study relating to sinister associations

of late decelerations require replication in chronically instrumented growth restricted fetuses

prior to translation into human practice.

5.4.5 Future studies

The studies presented in this chapter suggest that fetal growth restriction alters the fetal

cardiovascular, endocrine and metabolic responses to repeated cord occlusion and it supports

the need for further studies in this area. As described previously, FHR, arterial pressure and

endocrine responses of growth restricted fetuses during repeated umbilical cord occlusion

require evaluation in a chronically instrumented fetal preparation. Although

umbilicoplacental embolisation produces many of the features of human growth restriction,

this may not be the ideal model to use for future studies relating to asphyxia and fetal growth

restriction as the generation of growth restriction requires fetal surgery and prolonged

catheterisation. A more appropriate model may be maternal heat exposure which has been

shown to induce asymmetric growth restriction (Galan et al., 1999a; Regnault et al., 1999),

abnormal fetal umbilical artery Doppler signals (Galan et al., 1998; Galan et al., 1999b), and

impaired placental transport of oxygen and amino acids similar to that seen in human

pregnancy (Battaglia & Regnault, 2001; Regnault et al., 2002; Regnault et al., 2003). This

model of growth restriction does not require surgery or prolonged instrumentation to

generate growth restriction which would allow the growth restricted fetuses to be chronically

instrumented three to five days prior to the cord occlusion studies.

In addition, further studies are required to investigate the ability of FHR patterns and fetal

biochemistry to predict fetal brain injury in growth restricted fetuses exposed to repeated

284

cord occlusion. These studies are particularly important given the data presented in this

chapter suggesting that late declarations may be particularly sinister in growth restricted

fetuses, but also because of the recent evidence which suggests that human fetuses with

reduced growth potential are 20 times more likely to develop neonatal encephalopathy

secondary to an acute intrapartum hypoxic event than fetuses with normal growth potential

(Bukowski et al., 2003).

The results presented in this chapter suggest that further studies are required to investigate

fetal oxygen utilisation, lactate metabolism and placental lactate clearance in growth restricted

fetuses and additional studies investigating the hypothalamic-pituitary-adrenal axis are

required given the evidence in this study that the fetal adrenal responsiveness to ACTH is

impaired by prolonged umbilicoplacental embolisation. These studies would need to assess

paraventricular hypothalamic corticotropin-releasing hormone and arginine vasopressin

mRNA levels as well as anterior pituitary pro-opiomelanocortin mRNA levels. Analysis of

ACTH receptor and steroidogenic enzyme levels from the adrenal would provide insight into

adrenal sensitivity to circulating ACTH. Analysis of adrenal medulla phenyl-ethanolamine-

N-methyltransferase levels may provide insight into whether the observed increase in

cathecholamine levels in this study was due to increased production secondary to endocrine

responsiveness or increased sympathetic nervous stimulation or due to decreased clearance.

285

5.5 Conclusion

In this chapter, I have shown that umbilicoplacental embolisation for 23 days alters fetal acid-

base, endocrine, and cardiovascular responses to subsequent asphyxia in the ovine fetus. In

addition, evidence has been provided that suggests that lactate and oxygen metabolism in the

fetal brain during asphyxia are altered by umbilicoplacental embolisation. Despite the

multiple effects of umbilicoplacental embolisation on fetal responses to asphyxia,

embolisation did not alter fetal arterial lactate responses to asphyxia, the relationship between

fetal scalp blood lactate levels and carotid arterial lactate levels or the ability of fetal scalp

lactate levels to predict fetal arterial pressure responses during cord occlusion. These results

suggest that scalp lactate levels are a robust marker of fetal anaerobic metabolism and a

reliable predictor of arterial perfusion during umbilical cord occlusion in both appropriately

grown and growth restricted fetuses.

286

C h a p t e r 6

CLINICAL APPLICATION:

IS THERE EVIDENCE FOR CHANGE?

Science warns me to be careful how I adopt a view which jumps with my pre-conceptions

and to require stronger evidence for such belief.

My business is to teach my aspirations to conform with fact,

not to try and make facts harmonise with my aspirations.

Thomas Huxley (1860)

Worldwide there are approximately 130 million births annually and of these it has been

estimated that four million suffer from birth asphyxia, one million die and a similar number

develop sequelae including cerebral palsy, seizures, or developmental delay (WHO, 1991,

1995). Although it is widely accepted that a significant proportion of cases of neonatal

encephalopathy (Nelson & Leviton, 1991; Adamson et al., 1995; Badawi et al., 1998b, a) and

cerebral palsy (Blair & Stanley, 1988; Nelson, 1988; Yudkin et al., 1994; MacLennan, 1999)

have an antenatal origin, recent studies suggest that hypoxic-ischaemic events in the perinatal

period are the main cause of early neonatal encephalopathy, seizures or both (Biagioni et al.,

2001; Maalouf et al., 2001; Cowan et al., 2003) and perhaps play a more important role in the

origin of cerebral palsy than postulated over the last two decades (Hagberg et al., 2001).

The two techniques that currently are used widely for the assessment of fetal wellbeing

during labour in the developed world are continuous electronic fetal heart rate monitoring

(EFM) and fetal scalp blood pH measurement. Electronic fetal heart rate monitoring has

never been shown to be associated with a reduction in neonatal encephalopathy, cerebral

palsy or perinatal death (Grant, 1989; Vintzileos et al., 1995b; Thacker et al., 2001). The last

three decades of research into FHR monitoring suggest that EFM should be considered as a

screening test for the prediction of intrapartum fetal asphyxia in preterm and term pregnancy

(Low, 1998; Low et al., 2001; Low et al., 2002) and that EFM should be supplemented by fetal

scalp blood pH measurement. Assessment with the combined techniques has been shown to

287

limit the increase in intervention associated with EFM alone and are associated with a 50%

reduction in neonatal seizures (Grant, 1989; Thacker et al., 2001). Although fetal scalp blood

pH measurement has been advocated to improve the specificity of FHR monitoring (Clark &

Paul, 1985; Low et al., 2001; Low et al., 2002), its utilisation varies widely, partly due to the

invasive nature of the testing but also due to the restricted availability of the expensive

equipment required to perform acid-base measurement on micro volume blood samples

(Greene, 1999). In addition, scalp blood pH measurement, although offering additional

benefit to EFM, is a poor predictor of low Apgar scores (Banta & Thacker, 1979; Kruger et

al., 1999) and moderate to severe hypoxic-ischaemic encephalopathy (Kruger et al., 1999).

Therefore, there are significant limitations with our current techniques for the intrapartum

assessment of fetal wellbeing. In order to reduce the burden of birth asphyxia on families,

health care systems and the community, new techniques that are inexpensive, widely available

and accurate are required to allow recognition of early intrapartum asphyxia so that timely

obstetric intervention can avoid asphyxia-induced brain damage or death.

In this thesis I have evaluated the role of lactate measurement in intrapartum assessment of

fetal wellbeing. Specifically, I have addressed the question, “Is fetal lactate measurement

better than the assessment of FHR patterns or the measurement of fetal pH at predicting

fetal brain injury after intrapartum asphyxia?” Using an ovine model of repeated umbilical

cord occlusion designed to mimic events which may occur during human labour, I have

shown that the measurement of fetal lactate levels after repeated cord occlusion is

significantly associated with the severity of brain injury after the asphyxial insult. No

significant associations were identified with fetal pH or with the duration of decelerative or

compound FHR patterns; however, mine is the first study to describe an association between

the duration of both increased FHR variability and FHR overshoot with the severity of

subsequent brain injury. Although no significant association was identified between fetal

arterial pressure between umbilical cord occlusions and the grade of brain injury, the studies

performed in this thesis are the first to show a strong correlation between the duration of

specific arterial pressure responses during cord occlusions and the grade of brain injury,

accounting for approximately 90% of the variability seen in the severity of injury.

The mechanism responsible for the improved ability of lactate measurement to predict fetal

brain injury compared to other measures of fetal acid-base status is unknown. It may be

because lactate is a more stable marker of anaerobic metabolism in the fetus than fetal pH.

During hypoxia, lactate is the end product of anaerobic metabolism of glucose and it

equilibrates slowly over the placenta (Daniels, 1966) probably through facilitated diffusion

288

(Carstensen et al., 1983; Piquard et al., 1990) making it a stable marker of anaerobic

metabolism in the fetus. Conversely, changes in fetal pH can result from anaerobic

metabolism of glucose, elevated carbon dioxide levels resulting in respiratory acidosis, and

also from anaerobic metabolism of non-sulphur containing amino acids and fatty acids

(Gillstrap, 1999). The multiple potential sources of changes in fetal pH, and the fact that pH

can normalise rapidly via placental exchange of carbon dioxide, make this traditional marker

of fetal asphyxia less stable than lactate levels.

In this thesis I have provided evidence for two potential mechanisms that may explain why

lactate is a better predictor of fetal brain injury than fetal pH assessment. First, fetal lactate

may predict fetal brain injury through its association with fetal blood pressure responses

during umbilical cord occlusion. This was illustrated by the significantly greater area under

receiver operator curves for lactate than the traditional markers of acid-base status for

predicting arterial pressure responses during cord occlusion. The association was also

illustrated by the progressive increase in optimum cut off values for lactate predicting the

stages of cardiovascular decompensation during progressive asphyxia. The traditional

markers of acid-base status had stable optimum cut off scores as the stages of cardiovascular

deterioration progressed, suggesting that acid-base status may function as a threshold for

injury, whereas optimum lactate cut off scores progressively increased parallel to

cardiovascular decompensation with asphyxia. The second potential mechanism responsible

for why lactate is a better predictor of fetal brain injury than traditional acid-base assessment

may be through the association between fetal lactate levels and oxygen free radical activity in

the fetal brain. There is evidence from animal and human studies showing that levels of lipid

peroxidation measured in both brain and plasma, correlate with both the timing and the

severity of brain injury after hypoxic-ischaemic insults (Russell et al., 1992; Tan et al., 1998;

Ikeda et al., 1999; Candelario-Jalil et al., 2001; Marumo et al., 2001; Yu et al., 2003). In this

thesis, evidence has been provided which suggests that fetal blood lactate levels are

significantly better than pH, PO2 and base excess levels at predicting lipid peroxidation in the

fetal brain.

In human labour the fetal scalp is the window for biochemical assessment of fetal wellbeing.

In this thesis I have thoroughly investigated the relationship between fetal scalp blood lactate

levels and central blood lactate levels, considering both the effects of fetal asphyxia on this

relationship and also the effects of fetal growth restriction secondary to placental

insufficiency. The relationship between fetal scalp blood and carotid arterial lactate levels

was influenced by asphyxia but not by fetal growth restriction whereas the accuracy of fetal

289

scalp pH assessment was impaired in growth restricted fetuses with scalp pH levels

underestimating the severity of changes in central blood pH. Despite the complex

relationships between fetal scalp blood lactate levels and central blood lactate levels, scalp

lactate levels were significantly better at predicting lipid peroxidation in the fetal brain than

central markers of fetal acid-base status. In addition, fetal growth restriction did not impair

the ability of fetal scalp blood lactate levels to predict cardiovascular instability during cord

occlusion.

One of the major limitations with our current technique of biochemical assessment of fetal

wellbeing during labour is the restricted availability of expensive equipment required to

perform acid-base measurement on small volumes of blood, which limits this technique to

tertiary obstetric centres. In this thesis I have assessed the accuracy and precision of a

handheld lactate meter which is commercially available in Australia. The meter costs

approximately AUS$900 with analysis of each 15µL sample costing $1. I have found this

meter to be both accurate and precise in lactate measurement in human blood micro samples

across the gestational age range of 26 to 42 weeks. Therefore if the assessment of fetal

lactate levels were to be introduced into clinical practice, unlike fetal scalp blood pH

measurement, lactate measurement would not necessarily be limited to large obstetric centres

based on the cost of purchase and maintenance of measuring equipment.

Is it time to change? What is the optimum method of fetal monitoring in labour to prevent

hypoxic-ischaemic brain injury? Published human data (Low et al., 1976; Low et al., 1981;

Nieto et al., 1994; Dashe et al., 2000; Grant & Glazener, 2001; Bukowski et al., 2003) and data

presented in this thesis relating to growth restricted ovine fetuses suggests that growth

restricted fetuses respond differently than normally grown fetuses to asphyxial stress that can

occur in labour. Therefore, when considering the optimum method of fetal monitoring

during labour, I will address fetuses with appropriate growth and those with growth

restriction separately.

For fetuses with appropriate growth for gestational age, based on the body of currently

published literature, there is little doubt that the minimum standard of care is good quality

intermittent FHR auscultation, every 15 minutes during the first stage of labour and every

five minutes during the second stage of labour (ACOG, 1995; NICH, 1997; Parer & King,

2000; RCOG, 2001; Thacker et al., 2001; Freeman, 2002; Liston et al., 2002a; Liston et al.,

2002b; RANZCOG, 2002). In addition, it is considered by most clinicians and professional

obstetric societies that if a pregnancy is considered high risk then the woman should be

offered continuous EFM for her fetus during labour (NICH, 1997; RCOG, 2001; Liston et

290

al., 2002a; Liston et al., 2002b; RANZCOG, 2002), although this opinion is not based on any

data from randomised clinical trials in humans.

If non-reassuring FHR patterns are identified by intermittent auscultation then it is

considered optimal that continuous EFM is undertaken. If the non-reassuring FHR pattern

continues, randomised clinical trial evidence suggests that fetal scalp blood sampling for acid-

base assessment should be performed (Grant, 1989; Vintzileos et al., 1993; Thacker et al.,

2001) and further obstetric management should be based on the pH result and the overall

clinical picture including parity, progress in labour, proximity to delivery and the presence of

meconium stained amniotic fluid. Electronic fetal heart rate monitoring in the absence of

fetal blood sampling should not be advocated as it has never been shown to be better than

intermittent auscultation in preventing adverse neonatal outcomes. Furthermore, it is

associated with increased rates of instrumental vaginal delivery and caesarean section (Grant,

1989; Vintzileos et al., 1993; Thacker et al., 2001), both of which have small but real risks to

the mother and fetus .

Is there currently enough evidence to recommend introduction of measurement of lactate

levels in fetal scalp blood samples rather than, or in addition to, pH assessment? In addition

to the data presented in this thesis there have been a number of observational studies in

humans (Eguiluz et al., 1983; Smith et al., 1983; Nordstrom, 1995; Nordstrom et al., 1995b;

Kruger et al., 1999; Pennell et al., 2002) and one randomised clinical trial (Westgren et al.,

1998) addressing this question. In the published human data comparing the two techniques,

no differences were identified between the two techniques in rates of instrumental vaginal

delivery or caesarean section, nor was there any difference in neonatal outcome, umbilical

artery acid-base status or umbilical artery lactate levels (Westgren et al., 1998). Four

differences were identified between the two techniques. First, the process of fetal blood

sampling to assess fetal biochemical status was 16 times less likely to fail if blood was

collected for lactate measurement rather than pH measurement, probably due to the smaller

samples required for lactate measurement (Westgren et al., 1998). Second, fewer scalp

lacerations were performed if the fetuses were assigned to lactate measurement rather than

pH measurement (Westgren et al., 1998). Third, fetal scalp blood sampling for lactate

measurement was associated with a shorter testing time than scalp pH measurement and

fourth, scalp lactate measurement was significantly cheaper than scalp pH measurement

(Westgren et al., 1998). In an observational study performed in my obstetric unit whilst I was

undertaking my doctoral candidature, in addition to confirming the observations previously

described, we found that fetuses with both an abnormal scalp pH measurement (pH<7.20)

291

and an abnormal scalp lactate measurement (>4.8mmol/L) were more likely to have an

umbilical artery pH<7.1 than fetuses with abnormal scalp pH measurement but normal

lactate measurement (Pennell et al., 2002). In addition, fetuses with both an abnormal scalp

lactate level and an abnormal scalp pH level were more likely to be admitted to the neonatal

intensive care unit for longer than three days than babies who only had an abnormal scalp

pH measurement (Pennell et al., 2002). The only study which has compared scalp lactate

measurement with scalp pH measurement in humans with an adequate sample size (n=1709)

to investigate neonatal encephalopathy reported that scalp lactate measurement was

significantly better than scalp pH measurement at predicting moderate to severe neonatal

encephalopathy with a scalp lactate level of 4.8mmol/L having a sensitivity of 100% and

specificity of 73%. Therefore, considering the fetal physiological studies presented in this

thesis and the published literature regarding fetal scalp pH and scalp lactate assessment, there

probably is adequate data to change from scalp pH assessment to scalp lactate assessment, or

to change from EFM with no biochemical assessment to EFM combined with scalp lactate

assessment; however, it would be ideal if another large observational study such as the one

reported by Kruger et al. (1999) was performed to further evaluate the two techniques with

regard to the prevention of important adverse outcomes of neonatal encephalopathy and

cerebral palsy.

Is a randomised clinical trial (RCT) comparing scalp pH assessment to scalp lactate

assessment required prior to the introduction of scalp lactate assessment into clinical

practice? Although it is widely accepted that randomised clinical trials are the gold standard

for the assessment of clinical interventions (Woodward, 1999), they are particularly difficult

in the situation where one is attempting to prevent rare outcomes such as neonatal

encephalopathy and cerebral palsy. To perform a RCT comparing scalp pH assessment to

scalp lactate assessment aiming to identify a 25% reduction in intrapartum acquired neonatal

encephalopathy would require a total sample size of approximately 11,500 women (80%

power, α=5%, calculated on rates presented in Table 1.8, assuming 50% of cases of neonatal

encephalopathy are acquired during labour). If the RCT was designed to investigate similar

changes in cerebral palsy rates a total sample size of 200,000 women would be required. In

addition to the sample size issues, performing RCTs on women in labour with non-

reassuring FHR patterns has important ethical and recruiting issues. If the measurement of

fetal scalp blood pH and fetal scalp blood lactate were assessing two different biochemical

pathways, I believe that it would be optimal to perform a RCT at least addressing the issue of

reducing the risk of neonatal encephalopathy. However, scalp lactate assessment and scalp

pH assessment are both measures of fetal metabolic acidaemia and lactate measurement is a

292

more accurate way of assessing metabolic acidaemia than pH alone. In addition, fetuses born

with a severe respiratory acidosis (pH<7.0, base excess 4mmol/L to -8mmol/L) have been

shown to have neonatal outcomes that are no different from babies with normal acid-base

status at birth (Low et al., 1994). Therefore, as both techniques are assessing the same

biochemical pathway and lactate measurements appears a more accurate way of measuring

metabolic acidaemia than pH measurement, I propose that there is no need to perform a

RCT comparing scalp lactate to scalp pH measurement.

Once a baby is born, it is considered optimal for umbilical artery and vein acid-base

measurement to be performed (MacLennan, 1999). This procedure is not universally

performed, partly because of blood-gas equipment issues but also due to unfounded fears of

medico-legal retribution. Should we be performing umbilical artery lactate assessment at

birth instead of or in addition to acid-base assessment? A lack of association between fetal

acid-base status and the severity of fetal brain injury has consistently been reported in animal

studies since 1972 (Myers, 1972; Ting et al., 1983; Gunn et al., 1992; Mallard et al., 1992; de

Haan et al., 1993; Ikeda et al., 1998a; Fujii et al., 2003). In addition, human studies have

suggested that despite defining a group of neonates at risk, fetal umbilical artery acidaemia at

delivery has a poor correlation with outcome as most term infants with severe acidaemia at

delivery have an uncomplicated course (Winkler et al., 1991; Farrell et al., 1998; Huang et al.,

1999). Taken collectively, the animal and human data suggest that pH assessment identifies a

threshold for injury. If the pH level is higher than this threshold, fetal brain injury is unlikely

and if is lower than the threshold, fetal brain injury is possible. In this thesis, I have shown

that in sheep, fetal lactate levels after repeated cord occlusion are associated with the grade of

resultant fetal brain injury, accounting for approximately 60% of the variation in the severity

of brain injury. Similar findings have been reported by others in fetal sheep (Ikeda et al.,

1998a) and there is evidence from humans suggesting that measurement of lactate-creatinine

ratios in neonates at 6 hours of age (Huang et al., 1999) and the measurement of fetal brain

lactate levels using proton magnetic resonance spectroscopy (MRS) are predictive of poor

outcome in the neonatal intensive care unit (Hanrahan et al., 1996; Hanrahan et al., 1998), at

two months of age (Malik et al., 2002) and at 12 months of age (Amess et al., 1999).

Therefore, there are sufficient data to suggest that lactate assessment in umbilical arterial

blood may be beneficial; however, I believe further large human studies are required because

umbilical arterial lactate assessment at birth, although potentially valuable, may not be as

useful as lactate assessment performed in other ways such as MRS. Umbilical arterial lactate

assessment may define a subgroup of babies that could be evaluated by MRS prior to the

initiation of new experimental rescue therapies such as hypothermia (Yager et al., 1993;

293

Edwards et al., 1995; Laptook et al., 1995a; Laptook et al., 1995b; Thoresen et al., 1995;

Thoresen et al., 1996a; Thoresen et al., 1996b; Gunn et al., 1997; Gunn et al., 1998a; Gunn &

Gunn, 1998; Gunn et al., 1998b; Gunn et al., 1999; Gunn, 2000; Battin et al., 2001; Battin et al.,

2003), erythropoietin (Dame et al., 2001; Chong et al., 2002; Eid & Brines, 2002; Juul, 2002;

Kumral et al., 2003; Solaroglu et al., 2003), and antioxidants (Halks-Miller et al., 1986; Mishra

& Delivoria-Papadopoulos, 1999). This may reduce the number of babies requiring MRS,

which is expensive, and it would allow early identification of babies at risk of adverse

outcomes because early patient selection is important to gain the maximum benefit from

these new rescue techniques.

Growth restricted fetuses pose many challenges when considering labour. These fetuses are

more likely than appropriately grown fetuses to have non-reassuring FHR patterns (Low et

al., 1976; Dashe et al., 2000) and fetal acidosis in labour (Low et al., 1981; Nieto et al., 1994)

and are at increased risk of adverse outcomes secondary to acute hypoxic intrapartum events

(Bukowski et al., 2003). One of the first issues which require addressing is whether fetuses

with growth restriction should undergo labour or be delivered by an elective caesarean

section? A recent meta-analysis of published studies has addressed this issue and concluded

that currently there is not enough evidence to evaluate the use of elective caesarean section

for growth restricted fetuses (Grant & Glazener, 2001). The authors did note that growth

restricted fetuses exposed to labour were more likely to have neonatal seizures and death

than those delivered by elective caesarean section, but these differences did not reach

statistical significance (Grant & Glazener, 2001).

In this thesis I have presented the first experimental study using an animal model to

investigate fetal lactate responses during asphyxia induced by repetitive cord occlusion in

growth restricted fetuses. I have shown that growth restricted ovine fetuses have different

acid-base, endocrine, and cardiovascular responses to subsequent asphyxia than appropriately

grown fetuses. In addition, evidence has been provided which suggests that lactate and

oxygen metabolism in the fetal brain during asphyxia is altered by fetal growth restriction.

Despite the multiple effects of growth restriction on fetal responses to asphyxia, it did not

alter fetal arterial lactate responses to asphyxia, the relationship between fetal scalp blood and

carotid arterial lactate levels, arterial pressure responses during cord occlusion or the ability of

fetal scalp lactate levels to predict fetal arterial pressure responses during cord occlusion. The

data that I have presented suggesting that abnormal FHR patterns may have altered

significance in growth restricted fetuses are clinically vitally important but require replication

in appropriately instrumented fetuses. The data presented suggest that in growth restricted

294

fetuses late components to FHR decelerations are particularly sinister. In growth restricted

fetuses late components to decelerations did not reliably begin until lower base excess levels

than in appropriately grown fetuses and they were associated with cardiovascular instability

during umbilical cord occlusion with sensitivity of 98% and specificity of 88% for predicting

systolic and diastolic hypotension during cord occlusion. In growth restricted fetuses, FHR

overshoot after variable decelerations or changes in FHR variability had weaker associations

with changes in arterial pressure during cord occlusion than in normally grown fetuses. If

these observations are replicated in future studies, it may be that clinical decisions during

labour of growth restricted human fetuses can be made based on the duration of repeated

late decelerations whilst changes in FHR variability and overshoot may still require

biochemical assessment of fetal wellbeing. The responses of the growth restricted fetus to

labour is a challenging area which to date has undergone little scientific evaluation and in the

future requires detailed and thorough assessment both in animal models of asphyxia and in

human studies.

The suggestion that lactate levels are better predictors of outcome than traditional acid-base

status, although relatively recent in perinatal medicine, has been known for a number of years

in adult critical care medicine (Perret & Enrico, 1978; Bakker et al., 1996; Husain et al., 2003).

In adults, lactate level has been shown to be a reliable marker of tissue perfusion across a

wide range of conditions including acute trauma, major surgery and septic shock (Perret &

Enrico, 1978; Bakker et al., 1996; Husain et al., 2003) and lactate assessment is now widely

used in accident and emergency departments and in intensive care units around the world.

The assessment of fetal lactate levels both during and after labour is a simple, inexpensive

technique that can easily be integrated into clinical obstetric practice. It is probably superior

to our current techniques for the assessment of fetal wellbeing during labour; however, it is

not the final solution to intrapartum fetal monitoring. Abnormal lactate levels, whilst having

high sensitivity and specificity for perinatal brain injury, have poor positive predictive value

because the incidence of severe adverse outcomes is rare. Data presented both in this thesis

and by others (Myers, 1972; Ting et al., 1983; de Haan et al., 1993; de Haan et al., 1997b; Ikeda

et al., 1998a; Fujii et al., 2003) suggest that the key factor precipitating cerebral damage is

hypotension and the consequent compromise in cerebral oxygen delivery. I propose that the

next challenge for scientists, engineers and clinicians is to develop a technique to assess fetal

blood pressure non-invasively during labour because this measurement appears to be one of

the fundamental observations that we as clinicians are currently lacking when attempting to

make decisions relating to care in the hours before birth.

295

REFERENCES

AAP & ACOG. (1996). Use and abuse of the Apgar score. Committee on Fetus and Newborn, American Academy of Pediatrics, and Committee on Obstetric Practice, American College of Obstetricians and Gynecologists. Pediatrics 98, 141-142. AARNOUDSE, J. G., OESEBURG, B., KWANT, G., ZWART, A., ZIJLSTRA, W. G. & HUISJES, H. J.

(1981). Influence of variations in pH and PCO2 on scalp tissue oxygen tension and carotid arterial oxygen tension in the fetal lamb. Biol Neonate 40, 252-263.

ACCARDO, P. (1989). William John Little and cerebral palsy in the nineteenth century. J Hist

Med Allied Sci 44, 56-71. ACOG. (1992). Fetal and neonatal neurologic injury. ACOG technical bulletin no. 163.

American College of Obstetricians and Gynecologists, Washington, DC. ACOG. (1995). ACOG technical bulletin. Fetal heart rate patterns: monitoring,

interpretation, and management. Number 207--July 1995 (replaces No. 132, September 1989). Int J Gynaecol Obstet 51, 65-74.

ACOG. (1998). Inappropriate use of the terms fetal distress and birth asphyxia [Opinion

197]. American College of Obstetricians and Gynecologists committee on obstetric practice, Washington, DC.

ADAMSON, S. J., ALESSANDRI, L. M., BADAWI, N., BURTON, P. R., PEMBERTON, P. J. &

STANLEY, F. (1995). Predictors of neonatal encephalopathy in full term infants. Bmj 311, 598-602.

ADAMSONS, K., BEARD, R. W. & MYERS, R. E. (1970). Comparison of the composition of

arterial, venous, and capillary blood of the fetal monkey during labor. Am J Obstet Gynecol 107, 435-440.

AKAGI, K., OKAMURA, K., ENDO, C., SAITO, J., TANIGAWARA, S., SHINTAKU, Y.,

WATANABE, T., SATO, A. & YAJIMA, A. (1988). The slope of fetal heart rate deceleration is predictive of fetal condition during repeated umbilical cord compression in sheep. Am J Obstet Gynecol 159, 516-522.

AL NITA, D., NITA, V., SPULBER, S., MOLDOVAN, M., POPA, D. P., ZAGREAN, A. M. &

ZAGREAN, L. (2001). Oxidative damage following cerebral ischemia depends on reperfusion - a biochemical study in rat. J Cell Mol Med 5, 163-170.

ALON, E., BALL, R. H., GILLIE, M. H., PARER, J. T., ROSEN, M. A. & SHNIDER, S. M. (1993).

Effects of propofol and thiopental on maternal and fetal cardiovascular and acid-base variables in the pregnant ewe. Anesthesiology 78, 562-576.

AMESS, P. N., PENRICE, J., WYLEZINSKA, M., LOREK, A., TOWNSEND, J., WYATT, J. S.,

AMIEL-TISON, C., CADY, E. B. & STEWART, A. (1999). Early brain proton magnetic resonance spectroscopy and neonatal neurology related to neurodevelopmental outcome at 1 year in term infants after presumed hypoxic-ischaemic brain injury. Dev Med Child Neurol 41, 436-445.

296

AMIEL-TISON, C. & ELLISON, P. (1986). Birth asphyxia in the fullterm newborn: early

assessment and outcome. Dev Med Child Neurol 28, 671-682. ANDERSON, D. F., PARKS, C. M. & FABER, J. J. (1986). Fetal O2 consumption in sheep during

controlled long-term reductions in umbilical blood flow. Am J Physiol 250, H1037-1042.

ANDRUS, P. K., FLECK, T. J., OOSTVEEN, J. A. & HALL, E. D. (1997). Neuroprotective effects

of the novel brain-penetrating pyrrolopyrimidine antioxidants U-101033E and U-104067F against post-ischemic degeneration of nigrostriatal neurons. J Neurosci Res 47, 650-654.

ASANO, H., HOMAN, J., CARMICHAEL, L., KORKOLA, S. & RICHARDSON, B. (1994). Cerebral

metabolism during sustained hypoxemia in preterm fetal sheep. 170, 939-944. ASTRUP, J. (1982). Energy-requiring cell functions in the ischemic brain. Their critical supply

and possible inhibition in protective therapy. J Neurosurg 56, 482-497. BABLOK, W., PASSING, H., BENDER, R. & SCHNEIDER, B. (1988). A general regression

procedure for method transformation. Application of linear regression procedures for method comparison studies in clinical chemistry, Part III. J Clin Chem Clin Biochem 26, 783-790.

BACHMAN, C. R., BIEHL, D. R., SITAR, D., CUMMING, M. & PUCCI, W. (1986). Isoflurane

potency and cardiovascular effects during short exposures in the foetal lamb. Can Anaesth Soc J 33, 41-47.

BADAWI, N., KEOGH, J. M., DIXON, G. & KURINCZUK, J. J. (2001). Developmental outcomes

of newborn encephalopathy in the term infant. Indian Journal of Pediatrics. 68, 527-530. BADAWI, N., KURINCZUK, J. J., KEOGH, J. M., ALESSANDRI, L. M., O'SULLIVAN, F., BURTON,

P. R., PEMBERTON, P. J. & STANLEY, F. J. (1998a). Antepartum risk factors for newborn encephalopathy: the Western Australian case-control study. Bmj 317, 1549-1553.

BADAWI, N., KURINCZUK, J. J., KEOGH, J. M., ALESSANDRI, L. M., O'SULLIVAN, F., BURTON,

P. R., PEMBERTON, P. J. & STANLEY, F. J. (1998b). Intrapartum risk factors for newborn encephalopathy: the Western Australian case-control study. Bmj 317, 1554-1558.

BAKER, B. W., HUGHES, S. C., SHNIDER, S. M., FIELD, D. R. & ROSEN, M. A. (1990).

Maternal anesthesia and the stressed fetus: effects of isoflurane on the asphyxiated fetal lamb. Anesthesiology 72, 65-70.

BAKKER, J., GRIS, P., COFFERNILS, M., KAHN, R. J. & VINCENT, J. L. (1996). Serial blood

lactate levels can predict the development of multiple organ failure following septic shock. Am J Surg 171, 221-226.

BALL, R. H., PARER, J. T., CALDWELL, L. E. & JOHNSON, J. (1994). Regional blood flow and

metabolism in ovine fetuses during severe cord occlusion. Am J Obstet Gynecol 171, 1549-1555.

297

BANTA, H. D. & THACKER, S. B. (1979). Assessing the costs and benefits of electronic fetal monitoring. Obstet Gynecol Surv 34, 627-642.

BARKOVICH, A. J., BARANSKI, K., VIGNERON, D., PARTRIDGE, J. C., HALLAM, D. K.,

HAJNAL, B. L. & FERRIERO, D. M. (1999). Proton MR spectroscopy for the evaluation of brain injury in asphyxiated, term neonates. AJNR Am J Neuroradiol 20, 1399-1405.

BARNETT, A., MERCURI, E., RUTHERFORD, M., HAATAJA, L., FRISONE, M. F., HENDERSON,

S., COWAN, F. & DUBOWITZ, L. (2002). Neurological and perceptual-motor outcome at 5 - 6 years of age in children with neonatal encephalopathy: relationship with neonatal brain MRI. Neuropediatrics 33, 242-248.

BARTELDS, B., KNOESTER, H., SMID, G. B., TAKENS, J., VISSER, G. H., PENNINGA, L., VAN

DER LEIJ, F. R., BEAUFORT-KROL, G. C., ZIJLSTRA, W. G., HEYMANS, H. S. & KUIPERS, J. R. (2000). Perinatal changes in myocardial metabolism in lambs. Circulation 102, 926-931.

BARTELDS, B., VANBEL, F., TEETEL, D. & RUDOLPH, A. M. (1993). Carotid not aortic

chemoreceptors mediate the fetal cardiovascular response to acute hypoxemia in lambs. 34, 51-55.

BATTAGLIA, F. C. & MESCHIA, G. (1986). An introduction to fetal physiology, pp. 88-95.

Academic Press, London. BATTAGLIA, F. C. & REGNAULT, T. R. (2001). Placental transport and metabolism of amino

acids. Placenta 22, 145-161. BATTIN, M. R., DEZOETE, J. A., GUNN, T. R., GLUCKMAN, P. D. & GUNN, A. J. (2001).

Neurodevelopmental outcome of infants treated with head cooling and mild hypothermia after perinatal asphyxia. Pediatrics 107, 480-484.

BATTIN, M. R., PENRICE, J., GUNN, T. R. & GUNN, A. J. (2003). Treatment of term infants

with head cooling and mild systemic hypothermia (35.0 degrees C and 34.5 degrees C) after perinatal asphyxia. Pediatrics 111, 244-251.

BAX, M. & NELSON, K. B. (1993). Birth asphyxia: a statement. World Federation of

Neurology Group. Dev Med Child Neurol 35, 1022-1024. BEARD, R. W., EDINGTON, P. T. & SIBANDA, J. (1977). The effects of routine intrapartum

monitoring on clinical practice. Contrib Gynecol Obstet 3, 14-21. BEARD, R. W., FILSHIE, G. M., KNIGHT, C. A. & ROBERTS, G. M. (1971). The significance of

the changes in the continuous fetal heart rate in the first stage of labour. J Obstet Gynaecol Br Commonw 78, 865-881.

BEAULIEU, M. D., FABIA, J., LEDUC, B., BRISSON, J., BASTIDE, A., BLOUIN, D., GAUTHIER, R.

J. & LALONDE, A. (1982). The reproducibility of intrapartum cardiotocogram assessments. Can Med Assoc J 127, 214-216.

BECK, J. R. & SHULTZ, E. K. (1986). The use of relative operating characteristic (ROC) curves

in test performance evaluation. Arch Pathol Lab Med 110, 13-20.

298

BEETSON, S. (2001). Validation of DPC immunolite immunoassay for measurement of cortisol in blood samples obtained from sheep - Vetpath WA. Personal Communication, Perth.

BELAI, Y., GOODWIN, T. M., DURAND, M., GREENSPOON, J. S., PAUL, R. H. & WALTHER, F.

J. (1998). Umbilical arteriovenous PO2 and PCO2 differences and neonatal morbidity in term infants with severe acidosis. Am J Obstet Gynecol 178, 13-19.

BENNET, L., PEEBLES, D. M., EDWARDS, A. D., RIOS, A. & HANSON, M. A. (1998). The

cerebral hemodynamic response to asphyxia and hypoxia in the near-term fetal sheep as measured by near infrared spectroscopy. Pediatr Res 44, 951-957.

BENNET, L., WATANABE, T., SPENCER, J. & HANSON, M. A. (1994). Fetal reflexes in chronic

hypoxaemia. Adv Exp Med Biol 360, 337-339. BENSON, R. C., SHUBECK, F., DEUTSCHBERGER, J., WEISS, W. & BERENDES, H. (1968). Fetal

heart rate as a predictor of fetal distress. A report from the collaborative project. Obstet Gynecol 32, 259-266.

BERGER, R. & GARNIER, Y. (2000). Perinatal brain injury. J Perinat Med 28, 261-285. BERGER, R., GJEDDE, A., HECK, J., MULLER, E., KRIEGLSTEIN, J. & JENSEN, A. (1994).

Extension of the 2-deoxyglucose method to the fetus in utero: theory and normal values for the cerebral glucose consumption in fetal guinea pigs. J Neurochem 63, 271-279.

BERGER, R., JENSEN, A., KRIEGLSTEIN, J. & STEIGELMANN, J. P. (1992). Cerebral energy

metabolism in immature and mature guinea pig fetuses during acute asphyxia. J Dev Physiol 18, 125-128.

BERNARDES, J., COSTA-PEREIRA, A., AYRES-DE-CAMPOS, D., VAN GEIJN, H. P. & PEREIRA-

LEITE, L. (1997). Evaluation of interobserver agreement of cardiotocograms. Int J Gynaecol Obstet 57, 33-37.

BIAGIONI, E., MERCURI, E., RUTHERFORD, M., COWAN, F., AZZOPARDI, D., FRISONE, M. F.,

CIONI, G. & DUBOWITZ, L. (2001). Combined use of electroencephalogram and magnetic resonance imaging in full-term neonates with acute encephalopathy. Pediatrics 107, 461-468.

BIEHL, D. R., TWEED, W. A., COTE, J., WADE, J. G. & SITAR, D. (1983a). Effect of halothane

on cardiac output and regional flow in the fetal lamb in utero. Anesth Analg 62, 489-492.

BIEHL, D. R., YARNELL, R., WADE, J. G. & SITAR, D. (1983b). The uptake of isoflurane by

the foetal lamb in utero: effect on regional blood flow. Can Anaesth Soc J 30, 581-586. BLAIR, E. & STANLEY, F. (1990). Intrauterine growth and spastic cerebral palsy. I.

Association with birth weight for gestational age. Am J Obstet Gynecol 162, 229-237. BLAIR, E. & STANLEY, F. (1992). Intrauterine growth and spastic cerebral palsy II. The

association with morphology at birth. Early Hum Dev 28, 91-103.

299

BLAIR, E. & STANLEY, F. (1993). When can cerebral palsy be prevented? The generation of causal hypotheses by multivariate analysis of a case-control study. Paediatr Perinat Epidemiol 7, 272-301.

BLAIR, E. & STANLEY, F. J. (1988). Intrapartum asphyxia: a rare cause of cerebral palsy. J

Pediatr 112, 515-519. BLANCK, T. J. & STEVENSON, R. L. (1988). Thiopental does not alter Ca2+ uptake by cardiac

sarcoplasmic reticulum. Anesth Analg 67, 346-348. BLAND, J. M. & ALTMAN, D. G. (1986). Statistical methods for assessing agreement between

two methods of clinical measurement. Lancet i, 307-310. BLECHNER, J. N. (1993). Maternal-fetal acid base physiology. Clin Obstet Gynecol 36, 3-12. BLIX, E., SVIGGUM, O., KOSS, K. S. & OIAN, P. (2003). Inter-observer variation in assessment

of 845 labour admission tests: comparison between midwives and obstetricians in the clinical setting and two experts. Bjog 110, 1-5.

BOBBY, P. D. & SANTOS, A. C. (2000). Effect of prolonged asphyxia on skin blood flow in

fetal lambs. Obstet Gynecol 95, 607-611. BOCKING, A. D., WHITE, S. E., HOMAN, J. & RICHARDSON, B. S. (1992). Oxygen

consumption is maintained in fetal sheep during prolonged hypoxaemia. J Dev Physiol 17, 169-174.

BORGATTA, L., SHROUT, P. E. & DIVON, M. Y. (1988). Reliability and reproducibility of

nonstress test readings. Am J Obstet Gynecol 159, 554-558. BOWE, E. T., BEARD, R. W., FINSTER, M., POPPERS, P. J., ADAMSONS, K. & JAMES, L. S.

(1970). Reliability of fetal blood sampling. Maternal-fetal relationships. Am J Obstet Gynecol 107, 279-287.

BRACE, R. A. (1992). Fluid distribution in the fetus and neonate. In Fetal and Neonatal

Physiology. ed. POLIN, R. A. & FOX, W. W. WB Saunders Company, Harcourt Brace Jovanovich, Philadelphia, Pensylvania 19106.

BRACE, R. A. (1999). Dynamics and Disorders of Amniotic Fluid. In Maternal-fetal medicine. ed.

CREASY, R. K. & RESNIK, R., pp. 632-643. W.B. Saunders, Philadelphia. BRANSTON, N. M., SYMON, L., CROCKARD, H. A. & PASZTOR, E. (1974). Relationship

between the cortical evoked potential and local cortical blood flow following acute middle cerebral artery occlusion in the baboon. Exp Neurol 45, 195-208.

BRETSCHER, J. & SALING, E. (1967). pH values in the human fetus during labor. Am J Obstet

Gynecol 97, 906-911. BRETT, C. M., TEITEL, D. F., HEYMANN, M. A. & RUDOLPH, A. M. (1989). The young lamb

can increase cardiovascular performance during isoflurane anesthesia. Anesthesiology 71, 751-756.

BRIAN, J. E., JR. & FARACI, F. M. (1998). Tumor necrosis factor-alpha-induced dilatation of

cerebral arterioles. Stroke 29, 509-515.

300

BRIAN, J. E., JR., HEISTAD, D. D. & FARACI, F. M. (1995). Mechanisms of endotoxin-induced

dilatation of cerebral arterioles. Am J Physiol 269, H783-788. BROWN, J. K., PURVIS, R. J., FORFAR, J. O. & COCKBURN, F. (1974). Neurological aspects of

perinatal asphyxia. Dev Med Child Neurol 16, 567-580. BUCHHOLZ, J. & DUCKLES, S. P. (2001). Chronic hypoxia alters prejunctional alpha(2)-

receptor function in vascular adrenergic nerves of adult and fetal sheep. Am J Physiol Regul Integr Comp Physiol 281, R926-934.

BUCH-RASMUSSEN, T. (1990). Flow system for direct determination of enzyme substrate in

undiluted whole blood. Anal Chem 62, 932-936. BUKOWSKI, R., BURGETT, A. D., GEI, A., SAADE, G. R. & HANKINS, G. D. (2003).

Impairment of fetal growth potential and neonatal encephalopathy. Am J Obstet Gynecol 188, 1011-1015.

BUONOCORE, G., PERRONE, S. & BRACCI, R. (2001). Free radicals and brain damage in the

newborn. Biol Neonate 79, 180-186. BURD, L. I., JONES, M. D., JR., SIMMONS, M. A., MAKOWSKI, E. L., MESCHIA, G. &

BATTAGLIA, F. C. (1975). Placental production and foetal utilisation of lactate and pyruvate. Nature 254, 710-711.

BUTTKE, T. M. & SANDSTROM, P. A. (1994). Oxidative stress as a mediator of apoptosis.

Immunol Today 15, 7-10. CALDEYRO-BARCIA, R., MEDEZ-BAUER, C., POSEIRO, J. J., ESCARCENA, L., POSE, S. &

BIENIARZ, J. (1966). Control of the human fetal heart rate during labour. In The heart and circulation in the newborn and infant. ed. CASSELS, D., pp. 7-36. Grune & Stratton, New York.

CAMMER, W. & ZHANG, H. (1999). Maturation of oligodendrocytes is more sensitive to TNF

alpha than is survival of precursors and immature oligodendrocytes. J Neuroimmunol 97, 37-42.

CAMPBELL, A. G., DAWES, G. S., FISHMAN, A. P. & HYMAN, A. I. (1967). Regional

redistribution of blood flow in the mature fetal lamb. Circ Res 21, 229-235. CANDELARIO-JALIL, E., MHADU, N. H., AL-DALAIN, S. M., MARTINEZ, G. & LEON, O. S.

(2001). Time course of oxidative damage in different brain regions following transient cerebral ischemia in gerbils. Neurosci Res 41, 233-241.

CARSTENSEN, M. H., LEICHTWEISS, H. P. & SCHRODER, H. (1983). Lactate carriers in the

artificially perfused human term placenta. Placenta 4, 165-174. CARTER, A. M., CHALLIS, J. R. & SVENDSEN, P. (1996). Regional adrenal blood flow

responses to adrenocorticotropic hormone after chronic embolization of the fetal placental circulation in sheep. J Endocrinol 148, 517-522.

301

CETIN, I., GIOVANNINI, N., ALVINO, G., AGOSTONI, C., RIVA, E., GIOVANNINI, M. & PARDI, G. (2002). Intrauterine growth restriction is associated with changes in polyunsaturated fatty acid fetal-maternal relationships. Pediatr Res 52, 750-755.

CETIN, I., RONZONI, S., MARCONI, A. M., PERUGINO, G., CORBETTA, C., BATTAGLIA, F. C.

& PARDI, G. (1996). Maternal concentrations and fetal-maternal concentration differences of plasma amino acids in normal and intrauterine growth-restricted pregnancies. Am J Obstet Gynecol 174, 1575-1583.

CHABRIER, P. E., AUGUET, M., SPINNEWYN, B., AUVIN, S., CORNET, S., DEMERLE-

PALLARDY, C., GUILMARD-FAVRE, C., MARIN, J. G., PIGNOL, B., GILLARD-ROUBERT, V., ROUSSILLOT-CHARNET, C., SCHULZ, J., VIOSSAT, I., BIGG, D. & MONCADA, S. (1999). BN 80933, a dual inhibitor of neuronal nitric oxide synthase and lipid peroxidation: a promising neuroprotective strategy. Proc Natl Acad Sci U S A 96, 10824-10829.

CHALLIS, J. R., FRAHER, L., OOSTERHUIS, J., WHITE, S. E. & BOCKING, A. D. (1989). Fetal

and maternal endocrine responses to prolonged reductions in uterine blood flow in pregnant sheep. Am J Obstet Gynecol 160, 926-932.

CHAN, P. H. (1996). Role of oxidants in ischemic brain damage. Stroke 27, 1124-1129. CHANG, E. F., WONG, R. J., VREMAN, H. J., IGARASHI, T., GALO, E., SHARP, F. R.,

STEVENSON, D. K. & NOBLE-HAEUSSLEIN, L. J. (2003). Heme oxygenase-2 protects against lipid peroxidation-mediated cell loss and impaired motor recovery after traumatic brain injury. J Neurosci 23, 3689-3696.

CHANRACHAKUL, B., CHUA, S., NORDSTROM, L., YAM, J. & ARULKUMARAN, S. (1999).

Umbilical artery blood gas and lactate in healthy newborns. J Med Assoc Thai 82, 388-393.

CHAO, C. R., HOHIMER, A. R. & BISSONNETTE, J. M. (1989). The effect of elevated blood

glucose on the electroencephalogram and cerebral metabolism during short-term brain ischemia in fetal sheep. Am J Obstet Gynecol 161, 221-228.

CHAR, V. C. & CREASY, R. K. (1976). Lactate and pyruvate as fetal metabolic substrates.

Pediatr Res 10, 231-234. CHEEK, D. B., HUGHES, S. C., DAILEY, P. A., FIELD, D. R., PYTKA, S., ROSEN, M. A., PARER,

J. T. & SHNIDER, S. M. (1987). Effect of halothane on regional cerebral blood flow and cerebral metabolic oxygen consumption in the fetal lamb in utero. Anesthesiology 67, 361-366.

CHEN, T., QIAN, Y. Z., DI, X., RICE, A., ZHU, J. P. & BULLOCK, R. (2000a). Lactate/glucose

dynamics after rat fluid percussion brain injury. J Neurotrauma 17, 135-142. CHEN, T., QIAN, Y. Z., RICE, A., ZHU, J. P., DI, X. & BULLOCK, R. (2000b). Brain lactate

uptake increases at the site of impact after traumatic brain injury. Brain Res 861, 281-287.

CHMURA KRAEMER, H. (1992). Evaluating medical tests. SAGE Publications, Newbury Park.

302

CHOI, D. W. (1988). Calcium-mediated neurotoxicity: relationship to specific channel types and role in ischemic damage. Trends Neurosci 11, 465-469.

CHONG, Z. Z., KANG, J. Q. & MAIESE, K. (2002). Erythropoietin is a novel vascular

protectant through activation of Akt1 and mitochondrial modulation of cysteine proteases. Circulation 106, 2973-2979.

CHOU, Y. H., TSOU YAU, K. I. & WANG, P. J. (1998). Clinical application of the measurement

of cord plasma lactate and pyruvate in the assessment of high-risk neonates. Acta Paediatr 87, 764-768.

CLAPP, J. F., 3RD, SZETO, H. H., LARROW, R., HEWITT, J. & MANN, L. I. (1981). Fetal

metabolic response to experimental placental vascular damage. Am J Obstet Gynecol 140, 446-451.

CLAPP, J. F., PERESS, N. S., WESLEY, M. & MANN, L. I. (1988). Brain damage after

intermittent partial cord occlusion in the chronically instrumented fetal lamb. Am J Obstet Gynecol 159, 504-509.

CLARK, S. L., GIMOVSKY, M. L. & MILLER, F. C. (1982). Fetal heart rate response to scalp

blood sampling. Am J Obstet Gynecol 144, 706-708. CLARK, S. L., GIMOVSKY, M. L. & MILLER, F. C. (1984). The scalp stimulation test: a clinical

alternative to fetal scalp blood sampling. Am J Obstet Gynecol 148, 274-277. CLARK, S. L. & PAUL, R. H. (1985). Intrapartum fetal surveillance: the role of fetal scalp

blood sampling. Am J Obstet Gynecol 153, 717-720. CLAYE, A. (1963). Management of labour. In British Obstetric and Gynaecologic Practice. ed.

CLAYE, A. & BOURNE, A., pp. 184-204. William Heinemann Medical Books, London. COBBAERT, C., MORALES, C., VAN FESSEM, M. & KEMPERMAN, H. (1999). Precision,

accuracy and linearity of radiometer EML 105 whole blood metabolite biosensors. Ann Clin Biochem 36 ( Pt 6), 730-738.

COCK, M. L. & HARDING, R. (1997). Renal and amniotic fluid responses to umbilicoplacental

embolization for 20 days in fetal sheep. Am J Physiol 273, R1094-1102. COHN, H. E., SACKS, E. J., HEYMANN, M. A. & RUDOLPH, A. M. (1974). Cardiovascular

responses to hypoxemia and acidemia in fetal lambs. Am J Obstet Gynecol 120, 817-824.

COMPORTI, M. (1987). Glutathione depleting agents and lipid peroxidation. Chem Phys Lipids

45, 143-169. CONNETT, R. J., HONIG, C. R., GAYESKI, T. E. & BROOKS, G. A. (1990). Defining hypoxia: a

systems view of VO2, glycolysis, energetics, and intracellular PO2. J Appl Physiol 68, 833-842.

COOPER, C. E., COPE, M., QUARESIMA, V., FERRARI, M., NEMOTO, E., SPRINGETT, R.,

MATCHER, S., AMESS, P., PENRICE, J., TYSZCZUK, L., WYATT, J. & DELPY, D. T. (1997). Measurement of cytochrome oxidase redox state by near infrared spectroscopy. Adv Exp Med Biol 413, 63-73.

303

COOPER, C. E. & SPRINGETT, R. (1997). Measurement of cytochrome oxidase and

mitochondrial energetics by near-infrared spectroscopy. Philos Trans R Soc Lond B Biol Sci 352, 669-676.

COWAN, F. (2000). Outcome after intrapartum asphyxia in term infants. Semin Neonatol 5,

127-140. COWAN, F., RUTHERFORD, M., GROENENDAAL, F., EKEN, P., MERCURI, E., BYDDER, G. M.,

MEINERS, L. C., DUBOWITZ, L. M. & DE VRIES, L. S. (2003). Origin and timing of brain lesions in term infants with neonatal encephalopathy. Lancet 361, 736-742.

DALTON, K. J., DAWES, G. S. & PATRICK, J. E. (1977). Diurnal, respiratory, and other

rhythms of fetal heart rate in lambs. Am J Obstet Gynecol 127, 414-424. DALTON, K. J., DAWES, G. S. & PATRICK, J. E. (1983). The autonomic nervous system and

fetal heart rate variability. Am J Obstet Gynecol 146, 456-462. DAME, C., JUUL, S. E. & CHRISTENSEN, R. D. (2001). The biology of erythropoietin in the

central nervous system and its neurotrophic and neuroprotective potential. Biol Neonate 79, 228-235.

DANIELS, S. S. (1966). Lactate and pyruvate as an index of prenatal oxygen deprivation.

Pediatrics 37, 942-953. DASHE, J. S., MCINTIRE, D. D., LUCAS, M. J. & LEVENO, K. J. (2000). Effects of symmetric

and asymmetric fetal growth on pregnancy outcomes. Obstet Gynecol 96, 321-327. DAWES, G. S., MOTT, J. C. & SHELLEY, H. J. (1959). The importance of cardiac glycogen for

the maintenance of life in fetal lambs and newborn animal during anoxia. J Physiol 146, 516-538.

DAWES, G. S., MOULDEN, M. & REDMAN, C. W. (1991). The advantages of computerized

fetal heart rate analysis. J Perinat Med 19, 39-45. DE COURTEN-MYERS, G. M., FOGELSON, H. M., KLEINHOLZ, M. & MYERS, R. E. (1989).

Hypoxic brain and heart injury thresholds in piglets. Biomed Biochim Acta 48, S143-148.

DE HAAN, H. H., GUNN, A. J. & GLUCKMAN, P. D. (1996). Experiments in perinatal brain

injury: what have we learnt? Prenatal Neonatal Medicine 1, 16-25. DE HAAN, H. H., GUNN, A. J. & GLUCKMAN, P. D. (1997a). Fetal heart rate changes do not

reflect cardiovascular deterioration during brief repeated umbilical cord occlusions in near-term fetal lambs [see comments]. 176, 8-17.

DE HAAN, H. H., GUNN, A. J., WILLIAMS, C. E. & GLUCKMAN, P. D. (1997b). Brief repeated

umbilical cord occlusions cause sustained cytotoxic cerebral edema and focal infarcts in near-term fetal lambs. Pediatr Res 41, 96-104.

DE HAAN, H. H., GUNN, A. J., WILLIAMS, C. E., HEYMANN, M. A. & GLUCKMAN, P. D.

(1997c). Magnesium sulfate therapy during asphyxia in near-term fetal lambs does not compromise the fetus but does not reduce cerebral injury. 176, 18-27.

304

DE HAAN, H. H., VAN REEMPTS, J. L., VLES, J. S., DE HAAN, J. & HASAART, T. H. (1993).

Effects of asphyxia on the fetal lamb brain. Am J Obstet Gynecol 169, 1493-1501. DECKER, D. G. & ROSENBAUM, J. D. (1942). The distribution of lactic acid in human blood.

Am J Physiol 138, 7-11. DELLINGER, E. H., BOEHM, F. H. & CRANE, M. M. (2000). Electronic fetal heart rate

monitoring: early neonatal outcomes associated with normal rate, fetal stress, and fetal distress. Am J Obstet Gynecol 182, 214-220.

DELONG, E. R., DELONG, D. M. & CLARKE-PEARSON, D. L. (1988). Comparing the areas

under two or more correlated receiver operating characteristic curves: a nonparametric approach. Biometrics 44, 837-845.

DENNIS, J., JOHNSON, A., MUTCH, L., YUDKIN, P. & JOHNSON, P. (1989). Acid-base status at

birth and neurodevelopmental outcome at four and one-half years. Am J Obstet Gynecol 161, 213-220.

DEPAUL, J. (1847). Traite theorique et pratique d'auscultation obstetricale. Labe, Paris. DERHAM, R. J., MATTHEWS, T. G. & CLARKE, T. A. (1985). Early seizures indicate quality of

perinatal care. Arch Dis Child 60, 809-813. DEVADATTA, S. C. (1934). Distribution of lactate between corpuscles and the plasma in

blood. Q J Exp Physiol 24, 295. DIXON, G., BADAWI, N., KURINCZUK, J. J., KEOGH, J. M., SILBURN, S. R., ZUBRICK, S. R. &

STANLEY, F. J. (2002). Early developmental outcomes after newborn encephalopathy. Pediatrics 109, 26-33.

DRAPER, E. S., KURINCZUK, J. J., LAMMING, C. R., CLARKE, M., JAMES, D. & FIELD, D.

(2002). A confidential enquiry into cases of neonatal encephalopathy. Arch Dis Child Fetal Neonatal Ed 87, F176-180.

DU PLESSIS, A. J. & VOLPE, J. J. (2002). Perinatal brain injury in the preterm and term

newborn. Curr Opin Neurol 15, 151-157. DUKE, P. C., FOWNES, D. & WADE, J. G. (1977). Halothane depresses baroreflex control of

heart rate in man. Anesthesiology 46, 184-187. DUKE, P. C. & TROSKY, S. (1980). The effect of halothane with nitrous oxide on baroreflex

control of heart rate in man. Can Anaesth Soc J 27, 531-534. DUNCAN, J. R. (2002). Adverse perinatal conditions: their effects on physiological responses,

placental morphology and brain development. In Department of Anatomy and Cell Biology, Faculty of Medicine, Dentistry and Health Sciences, pp. 251. The University of Melbourne, Melbourne.

DUNCAN, J. R., COCK, M. L., HARDING, R. & REES, S. M. (2000). Relation between damage

to the placenta and the fetal brain after late-gestation placental embolization and fetal growth restriction in sheep. Am J Obstet Gynecol 183, 1013-1022.

305

EASTMAN, N., KOHL, S., MAISEL, J. & KAVALER, F. (1962). The obstetrical background of 753 cases of cerebral palsy. Obstet Gynecol Surv 17, 459-500.

ECONOMIDES, D. L. & NICOLAIDES, K. H. (1990). Metabolic findings in small-for-gestational

age fetuses. Contemp Rev in Obstet and Gynecol 2, 75-79. EDELSTONE, D. I. (1984). Fetal compensatory responses to reduced oxygen delivery. Semin

Perinatol 8, 184-191. EDELSTONE, D. I., CAINE, M. E. & FUMIA, F. D. (1985). Relationship of fetal oxygen

consumption and acid-base balance to fetal hematocrit. Am J Obstet Gynecol 151, 844-851.

EDWARDS, A. D. & AZZOPARDI, D. V. (2000). Perinatal hypoxia-ischemia and brain injury.

Pediatr Res 47, 431-432. EDWARDS, A. D. & NELSON, K. B. (1998). Neonatal encephalopathies. Time to reconsider

the cause of encephalopathies. Bmj 317, 1537-1538. EDWARDS, A. D., YUE, X., SQUIER, M. V., THORESEN, M., CADY, E. B., PENRICE, J.,

COOPER, C. E., WYATT, J. S., REYNOLDS, E. O. & MEHMET, H. (1995). Specific inhibition of apoptosis after cerebral hypoxia-ischaemia by moderate post-insult hypothermia. Biochem Biophys Res Commun 217, 1193-1199.

EDWARDS, L. J., SIMONETTA, G., OWENS, J. A., ROBINSON, J. S. & MCMILLEN, I. C. (1999).

Restriction of placental and fetal growth in sheep alters fetal blood pressure responses to angiotensin II and captopril. J Physiol 515 ( Pt 3), 897-904.

EGUILUZ, A., LOPEZ BERNAL, A., MCPHERSON, K., PARRILLA, J. J. & ABAD, L. (1983). The

use of intrapartum fetal blood lactate measurements for the early diagnosis of fetal distress. Am J Obstet Gynecol 147, 949-954.

EID, T. & BRINES, M. (2002). Recombinant human erythropoietin for neuroprotection: what

is the evidence? Clin Breast Cancer 3 Suppl 3, S109-115. EKLIND, S., MALLARD, C., LEVERIN, A. L., GILLAND, E., BLOMGREN, K., MATTSBY-

BALTZER, I. & HAGBERG, H. (2001). Bacterial endotoxin sensitizes the immature brain to hypoxic--ischaemic injury. Eur J Neurosci 13, 1101-1106.

EKLOW-LASTBOM, L., ROSSI, L., THOR, H. & ORRENIUS, S. (1986). Effects of oxidative stress

caused by hyperoxia and diquat. A study in isolated hepatocytes. Free Radic Res Commun 2, 57-68.

ELIMIAN, A., FIGUEROA, R. & TEJANI, N. (1997). Intrapartum assessment of fetal well-being:

a comparison of scalp stimulation with scalp blood pH sampling. Obstet Gynecol 89, 373-376.

ELLENBERG, J. H. & NELSON, K. B. (1988). Cluster of perinatal events identifying infants at

high risk for death or disability. J Pediatr 113, 546-552. ELLIS, M. & DE COSTELLO, A. M. (1999). Antepartum risk factors for newborn

encephalopathy. Intrapartum risk factors are important in developing world. Bmj 318, 1414; author reply 1415.

306

ELLIS, M., MANANDHAR, N., MANANDHAR, D. S. & COSTELLO, A. M. (2000). Risk factors

for neonatal encephalopathy in Kathmandu, Nepal, a developing country: unmatched case-control study. Bmj 320, 1229-1236.

ELLIS, M., MANANDHAR, N., SHRESTHA, P. S., SHRESTHA, L., MANANDHAR, D. S. &

COSTELLO, A. M. (1999). Outcome at 1 year of neonatal encephalopathy in Kathmandu, Nepal. Dev Med Child Neurol 41, 689-695.

ELLISON, P. H., FOSTER, M., SHERIDAN-PEREIRA, M. & MACDONALD, D. (1991). Electronic

fetal heart monitoring, auscultation, and neonatal outcome. Am J Obstet Gynecol 164, 1281-1289.

ENGIDAWORK, E., CHEN, Y., DELL'ANNA, E., GOINY, M., LUBEC, G., UNGERSTEDT, U.,

ANDERSSON, K. & HERRERA-MARSCHITZ, M. (1997). Effect of perinatal asphyxia on systemic and intracerebral pH and glycolysis metabolism in the rat. Exp Neurol 145, 390-396.

ERCAN, M., INCI, S., KILINC, K., PALAOGLU, S. & AYPAR, U. (2001). Nimodipine attenuates

lipid peroxidation during the acute phase of head trauma in rats. Neurosurg Rev 24, 127-130.

ERGANDER, U., ERIKSSON, M. & ZETTERSTROM, R. (1983). Severe neonatal asphyxia.

Incidence and prediction of outcome in the Stockholm area. Acta Paediatr Scand 72, 321-325.

ESKES, T. K., JONGSMA, H. W. & HOUX, P. C. (1983). Percentiles for gas values in human

umbilical cord blood. Eur J Obstet Gynecol Reprod Biol 14, 341-346. EUN, B. L., LIU, X. H. & BARKS, J. D. (2000). Pentoxifylline attenuates hypoxic-ischemic

brain injury in immature rats. Pediatr Res 47, 73-78. EVANS, K., RIGBY, A. S., HAMILTON, P., TITCHINER, N. & HALL, D. M. B. (2001). The

relationships between neonatal encephalopathy and cerebral palsy: a cohort study. J Obstet Gynaecol 21, 114-120.

EVRARD, P., GADISSEUX, J. & GRESSENS, P. (1996). Hypoxia opportunism during brain

development. In Acute brain asphyxia in term infants. ed. WRIGHT, L. L., MERENSTEIN, G. B. & HIRTZ, D., pp. 43-50. National Institiute of Child Health and Development, Bethesda, Maryland.

FARRELL, T., MIRES, G. J., OWEN, P. & PATEL, N. B. (1998). Antecedents of long-term

handicap. In Asphyxia and Fetal Brain Damage. ed. MAULIK, D., pp. 1-21. Wiley-Liss, New York.

FERNANDEZ, A., KIEFER, J., FOSDICK, L. & MCCONKEY, D. J. (1995). Oxygen radical

production and thiol depletion are required for Ca(2+)-mediated endogenous endonuclease activation in apoptotic thymocytes. J Immunol 155, 5133-5139.

FIELD, D. R., PARER, J. T., AUSLENDER, R., BAKER, B. W., ROSS, B. K. & LEICHT, C. H.

(1991). Fetal heart rate variability and cerebral oxygen consumption in fetal sheep during asphyxia. Eur J Obstet Gynecol Reprod Biol 42, 145-153.

307

FIELD, D. R., PARER, J. T., AUSLENDER, R. A., CHEEK, D. B., BAKER, W. & JOHNSON, J. (1990). Cerebral oxygen consumption during asphyxia in fetal sheep. J Dev Physiol 14, 131-137.

FINER, N. N., ROBERTSON, C. M., RICHARDS, R. T., PINNELL, L. E. & PETERS, K. L. (1981).

Hypoxic-ischemic encephalopathy in term neonates: perinatal factors and outcome. J Pediatr 98, 112-117.

FLAMM, E. S., DEMOPOULOS, H. B., SELIGMAN, M. L., POSER, R. G. & RANSOHOFF, J.

(1978). Free radicals in cerebral ischemia. Stroke 9, 445-447. FLEISCHER, A., SCHULMAN, H., JAGANI, N., MITCHELL, J. & RANDOLPH, G. (1982). The

development of fetal acidosis in the presence of an abnormal fetal heart rate tracing. I. The average for gestational age fetus. Am J Obstet Gynecol 144, 55-60.

FORRESTIER, F. (1987). Some aspects of fetal biology. Fetal Ther 2, 181-187. FOWDEN, A. L. (1994). Fetal metabolism and energy balance. In Textbook of fetal physiology. ed.

THORNBURN, G. D. & HARDING, R., pp. 74-75. Oxford University Press, Oxford. FOWDEN, A. L. (2001). Growth and metabolism. In Fetal growth and development. ed. HARDING,

R. & BOCKING, A. D., pp. 44-70. Cambridge University Press, Cambridge. FRAGEN, R. J. & AVRAM, M. J. (1999). Barbiturates. In Anesthesia, vol. 1. ed. MILLER, D. M.,

pp. 229-246. Churchill Livingstone, New York. FRANKL, W. S. & POOLE-WILSON, P. A. (1981). Effects of thiopental on tension

development, action potential, and exchange of calcium and potassium in rabbit ventricular myocardium. J Cardiovasc Pharmacol 3, 554-565.

FREEMAN, R., GARITE, T. & NAGEOTTE, M. (1991). Fetal Heart Rate Monitoring. Williams and

Wilkins, Baltimore, Maryland. FREEMAN, R. K. (1985). Prenatal and perinatal factors associated with brain disorders.

National Institutes of Health NIH Publ. No. 85-1149. FREEMAN, R. K. (2002). Problems with intrapartum fetal heart rate monitoring interpretation

and patient management. Obstet Gynecol 100, 813-826. FREUD, S. (1893). Zur kenntnis der cerbralen diplegien des kindesalters (im anschluss an die

little'sche krankheit). ed. DEUTICKE, F., Leipzig. FREUD, S., ed. (1897). Die infantile cerebrallahmung. Nothnagel's specielle pathologie und therapie, vol. 9,

part 2, section 2, Wein. FREUD, S. & RIE, O. (1891). Klinische studie ueber die halbseitige zerebrallahmungen des

kindes. Kassowitz Beitrage, old series: 3. FRIDOVICH, I. (1978). The biology of oxygen radicals. Science 201, 875-880. FUJII, E. Y., TAKAHASHI, N., KODAMA, Y., ROMAN, C., FERRIERO, D. M. & PARER, J. T.

(2003). Hemodynamic changes during complete umbilical cord occlusion in fetal sheep related to hippocampal neuronal damage. Am J Obstet Gynecol 188, 413-418.

308

GAFFNEY, G., FLAVELL, V., JOHNSON, A., SQUIER, M. & SELLERS, S. (1994a). Cerebral palsy

and neonatal encephalopathy. Arch Dis Child Fetal Neonatal Ed 70, F195-200. GAFFNEY, G., SELLERS, S., FLAVELL, V., SQUIER, M. & JOHNSON, A. (1994b). Case-control

study of intrapartum care, cerebral palsy, and perinatal death. Bmj 308, 743-750. GAFFNEY, G., SQUIER, M. V., JOHNSON, A., FLAVELL, V. & SELLERS, S. (1994c). Clinical

associations of prenatal ischaemic white matter injury. Arch Dis Child Fetal Neonatal Ed 70, F101-106.

GAGNON, R., CAMPBELL, K., HUNSE, C. & PATRICK, J. (1987). Patterns of human fetal heart

rate accelerations from 26 weeks to term. Am J Obstet Gynecol 157, 743-748. GAGNON, R., CHALLIS, J., JOHNSTON, L. & FRAHER, L. (1994). Fetal endocrine responses to

chronic placental embolization in the late-gestation ovine fetus. Am J Obstet Gynecol 170, 929-938.

GAGNON, R., HUNSE, C., FELLOWS, F., CARMICHAEL, L. & PATRICK, J. (1988). Fetal heart

rate and activity patterns in growth-retarded fetuses: changes after vibratory acoustic stimulation. Am J Obstet Gynecol 158, 265-271.

GAGNON, R., JOHNSTON, L. & MUROTSUKI, J. (1996). Fetal placental embolization in the

late-gestation ovine fetus: alterations in umbilical blood flow and fetal heart rate patterns. Am J Obstet Gynecol 175, 63-72.

GAGNON, R., MUROTSUKI, J., CHALLIS, J. R., FRAHER, L. & RICHARDSON, B. S. (1997). Fetal

sheep endocrine responses to sustained hypoxemic stress after chronic fetal placental embolization. Am J Physiol 272, E817-823.

GAGNON, R., RUNDLE, H., JOHNSTON, L. & HAN, V. K. (1995). Alterations in fetal and

placental deoxyribonucleic acid synthesis rates after chronic fetal placental embolization. Am J Obstet Gynecol 172, 1451-1458.

GALAN, H. L., HUSSEY, M. J., BARBERA, A., FERRAZZI, E., CHUNG, M., HOBBINS, J. C. &

BATTAGLIA, F. C. (1999a). Relationship of fetal growth to duration of heat stress in an ovine model of placental insufficiency. Am J Obstet Gynecol 180, 1278-1282.

GALAN, H. L., HUSSEY, M. J., CHUNG, M., CHYU, J. K., HOBBINS, J. C. & BATTAGLIA, F. C.

(1998). Doppler velocimetry of growth-restricted fetuses in an ovine model of placental insufficiency. Am J Obstet Gynecol 178, 451-456.

GALAN, H. L., JOZWIK, M., RIGANO, S., REGNAULT, T. R., HOBBINS, J. C., BATTAGLIA, F. C.

& FERRAZZI, E. (1999b). Umbilical vein blood flow determination in the ovine fetus: comparison of Doppler ultrasonographic and steady-state diffusion techniques. Am J Obstet Gynecol 181, 1149-1153.

GAMBKE, B. & MULLER, C. (1993). Accusport: evaluation of a new system for determining

lactate in capillary blood. Poster presented at the 33rd German Congress of Sports Medicine Paderborn, October 14-17, 1993.

309

GARDMARK, S. (1974). Studies on acid-base balance, carbohydrate and lipid metabolism in human fetal and maternal blood, in clinical and experimetnal conditions during labour. In [Thesis]. Student litteratur, Lund, Sweden.

GARDNER, D. S., FLETCHER, A. J., BLOOMFIELD, M. R., FOWDEN, A. L. & GIUSSANI, D. A.

(2002). Effects of prevailing hypoxaemia, acidaemia or hypoglycaemia upon the cardiovascular, endocrine and metabolic responses to acute hypoxaemia in the ovine fetus. J Physiol 540, 351-366.

GARDNER, D. S., FLETCHER, A. J., FOWDEN, A. L. & GIUSSANI, D. A. (2001a). A novel

method for controlled and reversible long term compression of the umbilical cord in fetal sheep. J Physiol 535, 217-229.

GARDNER, D. S., FLETCHER, A. J., FOWDEN, A. L. & GIUSSANI, D. A. (2001b). Plasma

adrenocorticotropin and cortisol concentrations during acute hypoxemia after a reversible period of adverse intrauterine conditions in the ovine fetus during late gestation. Endocrinology 142, 589-598.

GARDNER, D. S., GIUSSANI, D. A. & FOWDEN, A. L. (2003). Hindlimb glucose and lactate

metabolism during umbilical cord compression and acute hypoxemia in the late-gestation ovine fetus. Am J Physiol Regul Integr Comp Physiol 284, R954-964.

GARE, D. J., WHETHAM, J. C. & HENRY, J. D. (1967). The validity of scalp sampling. Am J

Obstet Gynecol 99, 722-724. GEERDINK, N., ROTTEVEEL, J. J., LAMMENS, M., SISTERMANS, E. A., HEIKENS, G. T.,

GABREELS, F. J., MULLAART, R. A. & HAMEL, B. C. (2002). MECP2 mutation in a boy with severe neonatal encephalopathy: clinical, neuropathological and molecular findings. Neuropediatrics 33, 33-36.

GENNARO, A. R., NORA, A. H., NORA, J. J., STANDER, R. W. & WEISS, L., ed. (1984). Gould

Medical Dictionary. McGraw-Hill, New York. GILES, W. B., TRUDINGER, B. J. & BAIRD, P. J. (1985). Fetal umbilical artery flow velocity

waveforms and placental resistance: pathological correlation. Br J Obstet Gynaecol 92, 31-38.

GILLSTRAP, L. C. (1999). Fetal acid-base balance. In Maternal-Fetal Medicine, vol. 1. ed.

CREASY, R. K. & RESNIK, R., pp. 331-340. W.B. Saunders Co., Philadelphia. GILSTRAP, L. C., 3RD, LEVENO, K. J., BURRIS, J., WILLIAMS, M. L. & LITTLE, B. B. (1989).

Diagnosis of birth asphyxia on the basis of fetal pH, Apgar score, and newborn cerebral dysfunction. Am J Obstet Gynecol 161, 825-830.

GITTO, E., REITER, R. J., KARBOWNIK, M., TAN, D. X., GITTO, P., BARBERI, S. & BARBERI, I.

(2002). Causes of oxidative stress in the pre- and perinatal period. Biol Neonate 81, 146-157.

GIUSSANI, D. A., SPENCER, J. A. D. & HANSON, M. A. (1994). Fetal cardiovascular reflex

responses to hypoxaemia. Fet Mat Med Rev 6, 17-37. GOLDABER, K. G., GILSTRAP, L. C., 3RD, LEVENO, K. J., DAX, J. S. & MCINTIRE, D. D.

(1991). Pathologic fetal acidemia. Obstet Gynecol 78, 1103-1107.

310

GOODLIN, R. C. (1990). Discussion to "What is fetal distress". Am J Obstet Gynecol 162, 1426. GOODLIN, R. C. & LOWE, E. W. (1974). A functional umbilical cord occlusion heart rate

pattern. The significance of overshoot. Obstet Gynecol 43, 22-30. GRAFE, M. R. (1994). The correlation of prenatal brain damage with placental pathology. J

Neuropathol Exp Neurol 53, 407-415. GRANT, A. (1989). Monitoring the fetus during labour. In Effective care in pregnancy and childbirth.

ed. CHALMERS, I., ENKIN, M. & KEIRSE, M. J., pp. 846-882. Oxford University Press, Oxford.

GRANT, A. & GLAZENER, C. M. (2001). Elective caesarean section versus expectant

management for delivery of the small baby. Cochrane Database Syst Rev, CD000078. GRANT, A., O'BRIEN, N., JOY, M. T., HENNESSY, E. & MACDONALD, D. (1989). Cerebral

palsy among children born during the Dublin randomised trial of intrapartum monitoring. Lancet 2, 1233-1236.

GREENE, K. R. (1999). Scalp blood gas analysis. Obstet Gynecol Clin North Am 26, 641-656, vii. GREENE, K. R. & ROSEN, K. G. (1995). Intrapartum asphyxia. In Fetal and neonatal neurology

and neurosurgery. ed. LEVENE, M. L. & LILFORD, R. J., pp. 389-404. Churchill Livingstone, New York.

GRETHER, J. K. & NELSON, K. B. (1997). Maternal infection and cerebral palsy in infants of

normal birth weight. Jama 278, 207-211. GROW, J. & BARKS, J. D. (2002). Pathogenesis of hypoxic-ischemic cerebral injury in the term

infant: current concepts. Clin Perinatol 29, 585-602, v. GUNN, A. J. (2000). Cerebral hypothermia for prevention of brain injury following perinatal

asphyxia. Curr Opin Pediatr 12, 111-115. GUNN, A. J., BENNET, L., GUNNING, M. I., GLUCKMAN, P. D. & GUNN, T. R. (1999).

Cerebral hypothermia is not neuroprotective when started after postischemic seizures in fetal sheep. Pediatr Res 46, 274-280.

GUNN, A. J., GLUCKMAN, P. D. & GUNN, T. R. (1998a). Selective head cooling in newborn

infants after perinatal asphyxia: a safety study. Pediatrics 102, 885-892. GUNN, A. J. & GUNN, T. R. (1998). The 'pharmacology' of neuronal rescue with cerebral

hypothermia. Early Hum Dev 53, 19-35. GUNN, A. J., GUNN, T. R., DE HAAN, H. H., WILLIAMS, C. E. & GLUCKMAN, P. D. (1997).

Dramatic neuronal rescue with prolonged selective head cooling after ischemia in fetal lambs. J Clin Invest 99, 248-256.

GUNN, A. J., GUNN, T. R., GUNNING, M. I., WILLIAMS, C. E. & GLUCKMAN, P. D. (1998b).

Neuroprotection with prolonged head cooling started before postischemic seizures in fetal sheep. Pediatrics 102, 1098-1106.

311

GUNN, A. J., PARER, J. T., MALLARD, E. C., WILLIAMS, C. E. & GLUCKMAN, P. D. (1992). Cerebral histologic and electrocorticographic changes after asphyxia in fetal sheep. Pediatr Res 31, 486-491.

GUTTERIDGE, J. M. & HALLIWELL, B. (1990). The measurement and mechanism of lipid

peroxidation in biological systems. Trends Biochem Sci 15, 129-135. GUTTERIDGE, J. M. C. (1988). Oxygen radicals and tissue injury. ed. HALLIWELL, B., pp. 9-19.

The Federation of American Societies for Experimental Biology, Bethesda, Maryland. HADAR, A., SHEINER, E., HALLAK, M., KATZ, M., MAZOR, M. & SHOHAM-VARDI, I. (2001).

Abnormal fetal heart rate tracing patterns during the first stage of labor: effect on perinatal outcome. Am J Obstet Gynecol 185, 863-868.

HAESSLEIN, H. C. & NISWANDER, K. R. (1980). Fetal distress in term pregnancies. Am J

Obstet Gynecol 137, 245-253. HAGBERG, B., HAGBERG, G., BECKUNG, E. & UVEBRANT, P. (2001). Changing panorama of

cerebral palsy in Sweden. VIII. Prevalence and origin in the birth year period 1991-94. Acta Paediatr 90, 271-277.

HAGBERG, H., GILLAND, E., BONA, E., HANSON, L. A., HAHIN-ZORIC, M., BLENNOW, M.,

HOLST, M., MCRAE, A. & SODER, O. (1996). Enhanced expression of interleukin (IL)-1 and IL-6 messenger RNA and bioactive protein after hypoxia-ischemia in neonatal rats. Pediatr Res 40, 603-609.

HAGNEVIK, K., FAXELIUS, G., IRESTEDT, L., LAGERCRANTZ, H., LUNDELL, B. & PERSSON,

B. (1984). Catecholamine surge and metabolic adaptation in the newborn after vaginal delivery and caesarean section. Acta Paediatr Scand 73, 602-609.

HALKS-MILLER, M., HENDERSON, M. & ENG, L. F. (1986). Alpha tocopherol decreases lipid

peroxidation, neuronal necrosis, and reactive gliosis in reaggregate cultures of fetal rat brain. J Neuropathol Exp Neurol 45, 471-484.

HALLIWELL, B. & CHIRICO, S. (1993). Lipid peroxidation: its mechanism, measurement, and

significance. Am J Clin Nutr 57, 715S-724S; discussion 724S-725S. HALLIWELL, B. & GUTTERIDGE, J. M. C. (1989). Free radicals in biology and medicine.

Clarendon Press. HANLEY, J. A. & MCNEIL, B. J. (1983). A method of comparing the areas under receiver

operating characteristic curves derived from the same cases. Radiology 148, 839-843. HANRAHAN, J. D., COX, I. J., EDWARDS, A. D., COWAN, F. M., SARGENTONI, J., BELL, J. D.,

BRYANT, D. J., RUTHERFORD, M. A. & AZZOPARDI, D. (1998). Persistent increases in cerebral lactate concentration after birth asphyxia. Pediatr Res 44, 304-311.

HANRAHAN, J. D., SARGENTONI, J., AZZOPARDI, D., MANJI, K., COWAN, F. M.,

RUTHERFORD, M. A., COX, I. J., BELL, J. D., BRYANT, D. J. & EDWARDS, A. D. (1996). Cerebral metabolism within 18 hours of birth asphyxia: a proton magnetic resonance spectroscopy study. Pediatr Res 39, 584-590.

312

HANSON, M. A. & KISURED, T. (2001). Cardiovascular system. In Fetal growth and development. ed. HARDING, R. & BOCKING, A. D., pp. 70-93. Cambridge University Press, Cambridge.

HARDING, K., EVANS, S. & NEWNHAM, J. (1995). Screening for the small fetus: a study of the

relative efficacies of ultrasound biometry and symphysiofundal height. Aust N Z J Obstet Gynaecol 35, 160-164.

HARKNESS, R. A., WHITELAW, A. G. & SIMMONDS, R. J. (1982). Intrapartum hypoxia: the

association between neurological assessment of damage and abnormal excretion of ATP metabolites. J Clin Pathol 35, 999-1007.

HARRIS, J. L., KRUEGER, T. R. & PARER, J. T. (1982). Mechanisms of late decelerations of the

fetal heart rate during hypoxia. Am J Obstet Gynecol 144, 491-496. HAVERKAMP, A. D., ORLEANS, M., LANGENDOERFER, S., MCFEE, J., MURPHY, J. &

THOMPSON, H. E. (1979). A controlled trial of the differential effects of intrapartum fetal monitoring. Am J Obstet Gynecol 134, 399-412.

HAVERKAMP, A. D., THOMPSON, H. E., MCFEE, J. G. & CETRULO, C. (1976). The evaluation

of continuous fetal heart rate monitoring in high-risk pregnancy. Am J Obstet Gynecol 125, 310-320.

HEATH, R. L. & TAPPEL, A. L. (1976). A new sensitive assay for the measurement of

hydroperoxides. Anal Biochem 76, 184-191. HEISS, W. D., HAYAKAWA, T. & WALTZ, A. G. (1976). Cortical neuronal function during

ischemia. Effects of occlusion of one middle cerebral artery on single-unit activity in cats. Arch Neurol 33, 813-820.

HELFAND, M., MARTON, K. & UELAND, K. (1985). Factors involved in the interpretation of

fetal monitor tracings. Am J Obstet Gynecol 151, 737-744. HELLER, G., MISSELWITZ, B. & SCHMIDT, S. (2000). Early neonatal mortality, asphyxia

related deaths, and timing of low risk births in Hesse, Germany, 1990-8: observational study. Bmj 321, 274-275.

HELLYER, P. W., BEDNARSKI, R. M., HUBBELL, J. A. & MUIR, W. W., 3RD. (1989). Effects of

halothane and isoflurane on baroreflex sensitivity in horses. Am J Vet Res 50, 2127-2134.

HELWIG, J. T., PARER, J. T., KILPATRICK, S. J. & LAROS, R. K., JR. (1996). Umbilical cord

blood acid-base state: what is normal? Am J Obstet Gynecol 174, 1807-1812; discussion 1812-1804.

HERBST, A. & INGEMARSSON, I. (1994). Intermittent versus continuous electronic

monitoring in labour: a randomised study. Br J Obstet Gynaecol 101, 663-668. HERR, I. & DEBATIN, K. M. (2001). Cellular stress response and apoptosis in cancer therapy.

Blood 98, 2603-2614.

313

HOCHACHKA, P. W., CLARK, C. M., MONGE, C., STANLEY, C., BROWN, W. D., STONE, C. K., NICKLES, R. J. & HOLDEN, J. E. (1996). Sherpa brain glucose metabolism and defense adaptations against chronic hypoxia. J Appl Physiol 81, 1355-1361.

HOCHULI, E. (1976). ["Fetal monitoring" within the framework of actual obstetric

monitoring]. Schweiz Med Wochenschr 106, 841-846. HOCKENBERY, D. M., OLTVAI, Z. N., YIN, X. M., MILLIMAN, C. L. & KORSMEYER, S. J.

(1993). Bcl-2 functions in an antioxidant pathway to prevent apoptosis. Cell 75, 241-251.

HON, E. (1968). An atlas of fetal heart rate patterns. Harty Press, New Haven. HON, E. H. & LEE, S. T. (1961). The electronic evaluation of the fetal heart rate. VIII.

Patterns preceding fetal death; further observation. Am J Obstet Gynecol 87, 1055-1063. HOOPER, S. B., COULTER, C. L., DEAYTON, J. M., HARDING, R. & THORBURN, G. D. (1990).

Fetal endocrine responses to prolonged hypoxemia in sheep. Am J Physiol 259, R703-708.

HOOPER, S. B., WALKER, D. W. & HARDING, R. (1995). Oxygen, glucose, and lactate uptake

by fetus and placenta during prolonged hypoxemia. Am J Physiol 268, R303-309. HUANG, C. C., WANG, S. T., CHANG, Y. C., LIN, K. P. & WU, P. L. (1999). Measurement of

the urinary lactate:creatinine ratio for the early identification of newborn infants at risk for hypoxic-ischemic encephalopathy. N Engl J Med 341, 328-335.

HUANG, H. M. & GIBSON, G. E. (1989). Phosphatidylinositol metabolism during in vitro

hypoxia. J Neurochem 52, 830-835. HUCH, R. (1986). Maternal hyperventilation and the fetus. J Perinat Med 14, 3-17. HUDOME, S., PALMER, C., ROBERTS, R. L., MAUGER, D., HOUSMAN, C. & TOWFIGHI, J.

(1997). The role of neutrophils in the production of hypoxic-ischemic brain injury in the neonatal rat. Pediatr Res 41, 607-616.

HUISJES, H. J. & AARNOUDSE, J. G. (1979). Arterial or venous umbilical pH as a measure of

neonatal morbidity? Early Hum Dev 3, 155-161. HUSAIN, F. A., MARTIN, M. J., MULLENIX, P. S., STEELE, S. R. & ELLIOTT, D. C. (2003).

Serum lactate and base deficit as predictors of mortality and morbidity. Am J Surg 185, 485-491.

IKEDA, T., CHOI, B. H., YEE, S., MURATA, Y. & QUILLIGAN, E. J. (1999). Oxidative stress,

brain white matter damage and intrauterine asphyxia in fetal lambs. Int J Dev Neurosci 17, 1-14.

IKEDA, T., MURATA, Y., QUILLIGAN, E. J., CHOI, B. H., PARER, J. T., DOI, S. & PARK, S. D.

(1998a). Physiologic and histologic changes in near-term fetal lambs exposed to asphyxia by partial umbilical cord occlusion. Am J Obstet Gynecol 178, 24-32.

314

IKEDA, T., MURATA, Y., QUILLIGAN, E. J., PARER, J. T., THEUNISSEN, I. M., CIFUENTES, P., DOI, S. & PARK, S. D. (1998b). Fetal heart rate patterns in postasphyxiated fetal lambs with brain damage. Am J Obstet Gynecol 179, 1329-1337.

IKENOUE, T., MARTIN, C. B., JR., MURATA, Y., ETTINGER, B. B. & LU, P. S. (1981). Effect of

acute hypoxemia and respiratory acidosis on the fetal heart rate in monkeys. Am J Obstet Gynecol 141, 797-806.

IMPEY, L., GREENWOOD, C., MACQUILLAN, K., REYNOLDS, M. & SHEIL, O. (2001). Fever in

labour and neonatal encephalopathy: a prospective cohort study. BJOG: an International Journal of Obstetrics & Gynaecology. 108, 594-597.

INCI, S., OZCAN, O. E. & KILINC, K. (1998). Time-level relationship for lipid peroxidation

and the protective effect of alpha-tocopherol in experimental mild and severe brain injury. Neurosurgery 43, 330-335; discussion 335-336.

INDER, T., MOCATTA, T., DARLOW, B., SPENCER, C., VOLPE, J. J. & WINTERBOURN, C.

(2002). Elevated free radical products in the cerebrospinal fluid of VLBW infants with cerebral white matter injury. Pediatr Res 52, 213-218.

INDER, T. E. & VOLPE, J. J. (2000). Mechanisms of perinatal brain injury. Semin Neonatol 5, 3-

16. INGEMARSSON, E. (1981). Routine electronic fetal monitoring during labor. Acta Obstet

Gynecol Scand Suppl 99, 1-29. INGEMARSSON, I., HERBST, A. & THORNGREN-JERNECK, K. (1997). Long term outcome

after umbilical artery acidaemia at term birth: influence of gender and duration of fetal heart rate abnormalities. Br J Obstet Gynaecol 104, 1123-1127.

ITSKOVITZ, J., GOETZMAN, B. W. & RUDOLPH, A. M. (1982). The mechanism of late

deceleration of the heart rate and its relationship to oxygenation in normoxemic and chronically hypoxemic fetal lambs. Am J Obstet Gynecol 142, 66-73.

ITSKOVITZ, J., LAGAMMA, E. F. & RUDOLPH, A. M. (1983). The effect of reducing umbilical

blood flow on fetal oxygenation. Am J Obstet Gynecol 145, 813-818. ITSKOVITZ, J., LAGAMMA, E. F. & RUDOLPH, A. M. (1987). Effects of cord compression on

fetal blood flow distribution and O2 delivery. Am J Physiol 252, H100-109. JACOBSEN, L. & ROOTH, G. (1971). Interpretive aspects on the acid-base composition and its

variation in fetal scalp blood and maternal blood during labour. J Obstet Gynaecol Br Commonw 78, 971-980.

JANSSON, T., WENNERGREN, M. & ILLSLEY, N. P. (1993). Glucose transporter protein

expression in human placenta throughout gestation and in intrauterine growth retardation. J Clin Endocrinol Metab 77, 1554-1562.

JENSEN, A. & BERGER, R. (1991). Fetal circulatory responses to oxygen lack. J Dev Physiol 16,

181-207. JENSEN, A. & HANSON, M. A. (1995). Circulatory responses to acute asphyxia in intact and

chemodenervated fetal sheep near term. Reprod Fertil Dev 7, 1351-1359.

315

JENSEN, A., HOHMANN, M. & KUNZEL, W. (1987). Redistribution of fetal circulation during

repeated asphyxia in sheep: effects on skin blood flow, transcutaneous PO2, and plasma catecholamines. J Dev Physiol 9, 41-55.

JENSEN, A. & LANG, U. (1992). Foetal circulatory responses to arrest of uterine blood flow in

sheep: effects of chemical sympathectomy. 17, 75-86. JENSEN, A., ROMAN, C. & RUDOLPH, A. M. (1991). Effects of reducing uterine blood flow on

fetal blood flow distribution and oxygen delivery. J Dev Physiol 15, 309-323. JIANG, N., MOYLE, M., SOULE, H. R., ROTE, W. E. & CHOPP, M. (1995). Neutrophil

inhibitory factor is neuroprotective after focal ischemia in rats. Ann Neurol 38, 935-942.

JOAS, T. A. & STEVENS, W. C. (1971). Comparison of the arrhythmic doses of epinephrine

during Forane, halothane, and fluroxene anesthesia in dogs. Anesthesiology 35, 48-53. JOHNSTON, M. V. (2003). MRI for neonatal encephalopathy in full-term infants. Lancet 361,

713-714. JOHNSTONE, F. D., CAMPBELL, D. M. & HUGHES, G. J. (1978). Has continuous intrapartum

monitoring made any impact on fetal outcome? Lancet 1, 1298-1300. JONES, C. T., ROEBUCK, M. M., WALKER, D. W. & JOHNSTON, B. M. (1988). The role of the

adrenal medulla and peripheral sympathetic nerves in the physiological responses of the fetal sheep to hypoxia. J Dev Physiol 10, 17-36.

JONES, M. D., JR. & TRAYSTMAN, R. J. (1984). Cerebral oxygenation of the fetus, newborn,

and adult. Semin Perinatol 8, 205-216. JOUPPILA, P., KUIKKA, J., JOUPPILA, R. & HOLLMEN, A. (1979). Effect of induction of

general anesthesia for cesarean section on intervillous blood flow. Acta Obstet Gynecol Scand 58, 249-253.

JOUPPILA, R. & HOLLMEN, A. (1976). The effect of segmental epidural analgesia on maternal

and foetal acid-base balance, lactate, serum potassium and creatine phosphokinase during labour. Acta Anaesthesiol Scand 20, 259-268.

JUUL, S. (2002). Erythropoietin in the central nervous system, and its use to prevent hypoxic-

ischemic brain damage. Acta Paediatr Suppl 91, 36-42. KADHIM, H., TABARKI, B., VERELLEN, G., DE PREZ, C., RONA, A. M. & SEBIRE, G. (2001).

Inflammatory cytokines in the pathogenesis of periventricular leukomalacia. Neurology 56, 1278-1284.

KALINKOV, D., SCHACHINGER, H. & BUCHHOLZ, R. (1980). Theoretical and practical

significance of parallel assays of serum lactic acid, pH and blood gases in mothers and neonates at birth. J Perinat Med 8, 134-141.

KALLAPUR, S. G., WILLET, K. E., JOBE, A. H., IKEGAMI, M. & BACHURSKI, C. J. (2001). Intra-

amniotic endotoxin: chorioamnionitis precedes lung maturation in preterm lambs. Am J Physiol Lung Cell Mol Physiol 280, L527-536.

316

KANEKO, M., WHITE, S., HOMAN, J. & RICHARDSON, B. (2003). Cerebral blood flow and

metabolism in relation to electrocortical activity with severe umbilical cord occlusion in the near-term ovine fetus. Am J Obstet Gynecol 188, 961-972.

KELSO, I. M., PARSONS, R. J., LAWRENCE, G. F., ARORA, S. S., EDMONDS, D. K. & COOKE, I.

D. (1978). An assessment of continuous fetal heart rate monitoring in labor. A randomized trial. Am J Obstet Gynecol 131, 526-532.

KENEPP, N. B., KUMAR, S., SHELLEY, W. C., STANLEY, C. A., GABBE, S. G. & GUTSCHE, B.

B. (1982). Fetal and neonatal hazards of maternal hydration with 5% dextrose before caesarean section. Lancet 1, 1150-1152.

KERGARADEC, J. L. D. (1822). Memoire sur l'auscultation appliquee a L'etude de la grossesse.

Communication Academie Medecine Paris, 1-37. KEUNEN, H., BLANCO, C. E., VAN REEMPTS, J. L. & HASAART, T. H. (1997). Absence of

neuronal damage after umbilical cord occlusion of 10, 15, and 20 minutes in midgestation fetal sheep. 176, 515-520.

KEUNEN, H., DEUTZ, N. E., VAN REEMPTS, J. L. & HASAART, T. H. (1999). Transient

umbilical cord occlusion in late-gestation fetal sheep results in hippocampal damage but not in cerebral arteriovenous difference for nitrite, a stable end product of nitric oxide. J Soc Gynecol Investig 6, 120-126.

KNOPF, K., PARER, J. T., ESPINOZA, M. I., HORTON, J. I., GUNN, A. J., WILLIAMS, C. E.,

GLUCKMAN, P. D. & PARER, W. J. (1991). Comparison of mathematical indices of fetal heart rate variability with visual assessment in the human and sheep. J Dev Physiol 16, 367-372.

KOH, K. S., GREVES, D., YUNG, S. & PEDDLE, L. J. (1975). Experience with fetal monitoring

in a university teaching hospital. Can Med Assoc J 112, 455-456, 459-460. KONTOS, H. A. (1981). Regulation of the cerebral circulation. Annu Rev Physiol 43, 397-407. KOSAKA, Y., TAKAHASHI, T. & MARK, L. C. (1969). Intravenous thiobarbiturate anesthesia

for cesarean section. Anesthesiology 31, 489-506. KOTRLY, K. J., EBERT, T. J., VUCINS, E., IGLER, F. O., BARNEY, J. A. & KAMPINE, J. P.

(1984). Baroreceptor reflex control of heart rate during isoflurane anesthesia in humans. Anesthesiology 60, 173-179.

KOZUMA, S., WATANABE, T., BENNET, L., GREEN, L. R. & HANSON, M. A. (1997). The

effect of carotid sinus denervation on fetal heart rate variation in normoxia, hypoxia and post-hypoxia in fetal sheep. Br J Obstet Gynaecol 104, 460-465.

KREBS, H. B., PETRES, R. E., DUNN, L. J., JORDAAN, H. V. & SEGRETI, A. (1979).

Intrapartum fetal heart rate monitoring. I. Classification and prognosis of fetal heart rate patterns. Am J Obstet Gynecol 133, 762-772.

KREBS, H. B., PETRES, R. E., DUNN, L. J. & SMITH, P. J. (1982). Intrapartum fetal heart rate

monitoring. VI. Prognostic significance of accelerations. Am J Obstet Gynecol 142, 297-305.

317

KRUGER, K., HALLBERG, B., BLENNOW, M., KUBLICKAS, M. & WESTGREN, M. (1999).

Predictive value of fetal scalp blood lactate concentration and pH as markers of neurologic disability. Am J Obstet Gynecol 181, 1072-1078.

KRUGER, K., KUBLICKAS, M. & WESTGREN, M. (1998). Lactate in scalp and cord blood from

fetuses with ominous fetal heart rate patterns. Obstet Gynecol 92, 918-922. KUBLI, F. W., HON, E. H., KHAZIN, A. F. & TAKEMURA, H. (1969). Observations on heart

rate and pH in the human fetus during labor. Am J Obstet Gynecol 104, 1190-1206. KUMRAL, A., OZER, E., YILMAZ, O., AKHISAROGLU, M., GOKMEN, N., DUMAN, N.,

ULUKUS, C., GENC, S. & OZKAN, H. (2003). Neuroprotective effect of erythropoietin on hypoxic-ischemic brain injury in neonatal rats. Biol Neonate 83, 224-228.

LAGAMMA, E., ITSKOVITZ, J. & RUDOLPH, A. M. (1982). Effect of naloxone on fetal

circulatory responses to hypoxemia. 143, 933-940. LAGERCRANTZ, H. & SLOTKIN, T. A. (1986). The "stress" of being born. Sci Am 254, 100-

107. LANGENDOERFER, S., HAVERKAMP, A. D., MURPHY, J., NOWICK, K. D., ORLEANS, M.,

PACOSA, F. & VAN DOORNINCK, W. (1980). Pediatric follow-up of a randomized controlled trial of intrapartum fetal monitoring techniques. J Pediatr 97, 103-107.

LAPTOOK, A. R., CORBETT, R. J., BURNS, D. & STERETT, R. (1995a). Neonatal ischemic

neuroprotection by modest hypothermia is associated with attenuated brain acidosis. Stroke 26, 1240-1246.

LAPTOOK, A. R., CORBETT, R. J., STERETT, R., GARCIA, D. & TOLLEFSBOL, G. (1995b).

Quantitative relationship between brain temperature and energy utilization rate measured in vivo using 31P and 1H magnetic resonance spectroscopy. Pediatr Res 38, 919-925.

LEBLANC, M. H., QIAN, X. B. & CAI, Z. W. (1997). The effect of glucose during ischemia on

brain ATP, lactate, and glutamate in piglets. Biol Neonate 72, 243-254. LEE, W. K. & BAGGISH, M. S. (1976). The effect of unselected intrapartum fetal monitoring.

Obstet Gynecol 47, 516-520. LETH, H., TOFT, P. B., PEITERSEN, B., LOU, H. C. & HENRIKSEN, O. (1996). Use of brain

lactate levels to predict outcome after perinatal asphyxia. Acta Paediatr 85, 859-864. LEVENE, M. L. (1995). The asphyxiated newborn infant. In Fetal and neonatal neurology and

neurosurgery. ed. LEVENE, M. L., pp. 405-425. Churchill Livingstone, Edinburgh. LEVENE, M. L., KORNBERG, J. & WILLIAMS, T. H. (1985). The incidence and severity of post-

asphyxial encephalopathy in full-term infants. Early Hum Dev 11, 21-26. LEVINSON, G., SHNIDER, S. M., GILDEA, J. E. & DELORIMIER, A. A. (1973). Maternal and

foetal cardiovascular and acid-base changes during ketamine anaesthesia in pregnant ewes. Br J Anaesth 45, 1111-1115.

318

LEVITON, A. & NELSON, K. B. (1992). Problems with definitions and classifications of newborn encephalopathy. Pediatr Neurol 8, 85-90.

LEVITON, A. & PANETH, N. (1990). White matter damage in preterm newborns--an

epidemiologic perspective. Early Hum Dev 24, 1-22. LEWEN, A., MATZ, P. & CHAN, P. H. (2000). Free radical pathways in CNS injury. J

Neurotrauma 17, 871-890. LIAO, S. L. & CHEN, C. J. (2001). Differential effects of cytokines and redox potential on

glutamate uptake in rat cortical glial cultures. Neurosci Lett 299, 113-116. LICATA, R. H., OLSON, D. R. & MACK, E. W. (1975). Cholinergic and adrenergic innervation

of cerebral vesels. In Cerebral circulation and metabolism. ed. LAGFITT, T. W., MCHENRY, L. C., REIIC, M. & WOLLMANN, H., pp. 466-469. Springer-Verlag, Berlin.

LIEN, J. M., TOWERS, C. V., QUILLIGAN, E. J., DE VECIANA, M., TOOHEY, J. S. & MORGAN,

M. A. (1995). Term early-onset neonatal seizures: obstetric characteristics, etiologic classifications, and perinatal care. Obstet Gynecol 85, 163-169.

LILIEN, A. (1970). Term intrapartum fetal death. Am J Obstet Gynecol 107, 595-603. LILIENFELD, A. & PASAMANICK, B. (1955). The association of maternal and fetal factors with

the development of cerebral palsy and epilepsy. Am J Obstet Gynecol 70, 93. LILJA, H., KARLSSON, K., KJELLMER, I., LINDECRANTZ, K., OLSSON, T. & ROSEN, K. G.

(1984). Heart rate variability and electrocardiogram changes in the fetal lamb during hypoxia and beta-adrenoceptor stimulation. J Perinat Med 12, 115-125.

LIN, C. C., SU, S. J. & RIVER, L. P. (1991). Comparison of associated high-risk factors and

perinatal outcome between symmetric and asymmetric fetal intrauterine growth retardation. Am J Obstet Gynecol 164, 1535-1541; discussion 1541-1532.

LINDECRANTZ, K., CERUTTI, S., CIVARDI, S., HOKEGARD, K. H., LILJA, H., ROSEN, K. G.,

SIGNORINI, M. G. & WIDMARK, C. (1993). Power spectrum analysis of the fetal heart rate during noradrenaline infusion and acute hypoxemia in the chronic fetal lamb preparation. Int J Biomed Comput 33, 199-207.

LISTON, R., CRANE, J., HAMILTON, E., HUGHES, O., KULING, S., MACKINNON, C.,

MCNAMARA, H., MILNE, K., RICHARDSON, B. & TREPANIE, M. J. (2002a). Fetal health surveillance in labour. J Obstet Gynaecol Can 24, 250-276.

LISTON, R., CRANE, J., HUGHES, O., KULING, S., MACKINNON, C., MILNE, K.,

RICHARDSON, B. & TREPANIER, M. J. (2002b). Fetal health surveillance in labour. J Obstet Gynaecol Can 24, 342-355.

LITTLE, W. J. (1862). On the influence of abnormal parturition, difficult labours, premature

birth, and asphyxia neonatorum, on the mental and physical condition of the child, especially in relation to deformities. Trans London Obstet Soc 3, 293-325.

LIU, X. H., EUN, B. L., SILVERSTEIN, F. S. & BARKS, J. D. (1996). The platelet-activating

factor antagonist BN 52021 attenuates hypoxic-ischemic brain injury in the immature rat. Pediatr Res 40, 797-803.

319

LOELIGER, M., WATSON, C. S., REYNOLDS, J. D., PENNING, D. H., HARDING, R., BOCKING,

A. D. & REES, S. M. (2003). Extracellular glutamate levels and neuropathology in cerebral white matter following repeated umbilical cord occlusion in the near term fetal sheep. Neuroscience 116, 705-714.

LOREK, A., TAKEI, Y., CADY, E. B., WYATT, J. S., PENRICE, J., EDWARDS, A. D., PEEBLES, D.,

WYLEZINSKA, M., OWEN-REECE, H., KIRKBRIDE, V. & ET AL. (1994). Delayed ("secondary") cerebral energy failure after acute hypoxia-ischemia in the newborn piglet: continuous 48-hour studies by phosphorus magnetic resonance spectroscopy. Pediatr Res 36, 699-706.

LOUIS, J. C., MAGAL, E., TAKAYAMA, S. & VARON, S. (1993). CNTF protection of

oligodendrocytes against natural and tumor necrosis factor-induced death. Science 259, 689-692.

LOW, J. A. (1988). The role of blood gas and acid-base assessment in the diagnosis of

intrapartum fetal asphyxia. Am J Obstet Gynecol 159, 1235-1240. LOW, J. A. (1997). Intrapartum fetal asphyxia: definition, diagnosis, and classification. Am J

Obstet Gynecol 176, 957-959. LOW, J. A. (1998). Fetal asphyxia and outcome. In Asphyxia and fetal brain damage. ed. MAULIK,

D., pp. 21-37. Wiley-Liss, New York. LOW, J. A., BOSTON, R. W. & PANCHAM, S. R. (1972). Fetal asphyxia during the antepartum

perid in intrauterine growth retarded infants. Am J Obstet Gynecol 113, 351-356. LOW, J. A., GALBRAITH, R. S., MUIR, D. W., KILLEN, H. L., PATER, E. A. & KARCHMAR, E. J.

(1983). Intrapartum fetal hypoxia: a study of long-term morbidity. Am J Obstet Gynecol 145, 129-134.

LOW, J. A., GALBRAITH, R. S., MUIR, D. W., KILLEN, H. L., PATER, E. A. & KARCHMAR, E. J.

(1985). The relationship between perinatal hypoxia and newborn encephalopathy. Am J Obstet Gynecol 152, 256-260.

LOW, J. A., GALBRAITH, R. S., MUIR, D. W., KILLEN, H. L., PATER, E. A. & KARCHMAR, E. J.

(1988). Motor and cognitive deficits after intrapartum asphyxia in the mature fetus. Am J Obstet Gynecol 158, 356-361.

LOW, J. A., KARCHMAR, J., BROEKHOVEN, L., LEONARD, T., MCGRATH, M. J., PANCHAM, S.

R. & PIERCY, W. N. (1981). The probability of fetal metabolic acidosis during labor in a population at risk as determined by clinical factors. Am J Obstet Gynecol 141, 941-951.

LOW, J. A., KILLEN, H. & DERRICK, E. J. (2002). The prediction and prevention of

intrapartum fetal asphyxia in preterm pregnancies. Am J Obstet Gynecol 186, 279-282. LOW, J. A., PANAGIOTOPOULOS, C. & DERRICK, E. J. (1994). Newborn complications after

intrapartum asphyxia with metabolic acidosis in the term fetus. Am J Obstet Gynecol 170, 1081-1087.

320

LOW, J. A., PANAGIOTOPOULOS, C. & DERRICK, E. J. (1995). Newborn complications after intrapartum asphyxia with metabolic acidosis in the preterm fetus. Am J Obstet Gynecol 172, 805-810.

LOW, J. A., PANCHAM, S. R., PIERCY, W. N., WORTHINGTON, D. & KARCHMAR, E. J. (1979).

Maternal and fetal lactate characteristics during labour and delivery. In Lactate in acute conditions. ed. BOSSART, H. & PERRET, C., pp. 29-47. Karger, Basel.

LOW, J. A., PANCHAM, S. R. & WORTHINGTON, D. (1976). Fetal heart deceleration patterns in

relation to asphyxia and weight-gestational age percentile of the fetus. Obstet Gynecol 47, 14-20.

LOW, J. A., PANCHAM, S. R., WORTHINGTON, D. & BOSTON, R. W. (1974). Acid-base, lactate,

and pyruvate characteristics of the normal obstetric patient and fetus during the intrapartum period. Am J Obstet Gynecol 120, 862-867.

LOW, J. A., PICKERSGILL, H., KILLEN, H. & DERRICK, E. J. (2001). The prediction and

prevention of intrapartum fetal asphyxia in term pregnancies. Am J Obstet Gynecol 184, 724-730.

LUNELL, N. O., JOELSSON, I., LARSSON, A. & PERSSON, B. (1977). The immediate effect of a

beta-adrenergic agonist (salbutamol) on carbohydrate and lipid metabolism during the third trimester of pregnancy. Acta Obstet Gynecol Scand 56, 475-478.

LUTHY, D. A., SHY, K. K., VAN BELLE, G., LARSON, E. B., HUGHES, J. P., BENEDETTI, T. J.,

BROWN, Z. A., EFFER, S., KING, J. F. & STENCHEVER, M. A. (1987). A randomized trial of electronic fetal monitoring in preterm labor. Obstet Gynecol 69, 687-695.

MA, G. (1999). Institute of Clinical Pathology and Medical Research validation of DPC

immunolite assay for analysis of cortisol in blood samples obtained from sheep and dolphins. Personal Communication, Sydney.

MAALOUF, E. F., DUGGAN, P. J., COUNSELL, S. J., RUTHERFORD, M. A., COWAN, F.,

AZZOPARDI, D. & EDWARDS, A. D. (2001). Comparison of findings on cranial ultrasound and magnetic resonance imaging in preterm infants. Pediatrics 107, 719-727.

MACDONALD, D., GRANT, A., SHERIDAN-PEREIRA, M., BOYLAN, P. & CHALMERS, I. (1985).

The Dublin randomized controlled trial of intrapartum fetal heart rate monitoring. Am J Obstet Gynecol 152, 524-539.

MACLENNAN, A. (1999). A template for defining a causal relation between acute intrapartum

events and cerebral palsy: international consensus statement. Bmj 319, 1054-1059. MACMANUS, J. P., BUCHAN, A. M., HILL, I. E., RASQUINHA, I. & PRESTON, E. (1993). Global

ischemia can cause DNA fragmentation indicative of apoptosis in rat brain. Neurosci Lett 164, 89-92.

MAHOMED, K., NYONI, R., MULAMBO, T., KASULE, J. & JACOBUS, E. (1994). Randomised

controlled trial of intrapartum fetal heart rate monitoring. Bmj 308, 497-500.

321

MALIK, G. K., PANDEY, M., KUMAR, R., CHAWLA, S., RATHI, B. & GUPTA, R. K. (2002). MR imaging and in vivo proton spectroscopy of the brain in neonates with hypoxic ischemic encephalopathy. Eur J Radiol 43, 6-13.

MALLARD, E. C., GUNN, A. J., WILLIAMS, C. E., JOHNSTON, B. M. & GLUCKMAN, P. D.

(1992). Transient umbilical cord occlusion causes hippocampal damage in the fetal sheep. Am J Obstet Gynecol 167, 1423-1430.

MALLARD, E. C., REES, S., STRINGER, M., COCK, M. L. & HARDING, R. (1998). Effects of

chronic placental insufficiency on brain development in fetal sheep. Pediatr Res 43, 262-270.

MALLARD, E. C., WALDVOGEL, H. J., WILLIAMS, C. E., FAULL, R. L. & GLUCKMAN, P. D.

(1995a). Repeated asphyxia causes loss of striatal projection neurons in the fetal sheep brain. Neuroscience 65, 827-836.

MALLARD, E. C., WILLIAMS, C. E., GUNN, A. J., GUNNING, M. I. & GLUCKMAN, P. D. (1993).

Frequent episodes of brief ischemia sensitize the fetal sheep brain to neuronal loss and induce striatal injury. Pediatr Res 33, 61-65.

MALLARD, E. C., WILLIAMS, C. E., JOHNSTON, B. M. & GLUCKMAN, P. D. (1994). Increased

vulnerability to neuronal damage after umbilical cord occlusion in fetal sheep with advancing gestation. Am J Obstet Gynecol 170, 206-214.

MALLARD, E. C., WILLIAMS, C. E., JOHNSTON, B. M., GUNNING, M. I., DAVIS, S. &

GLUCKMAN, P. D. (1995b). Repeated episodes of umbilical cord occlusion in fetal sheep lead to preferential damage to the striatum and sensitize the heart to further insults. Pediatr Res 37, 707-713.

MARCONI, A. M., CETIN, I., FERRAZZI, E., FERRARI, M. M., PARDI, G. & BATTAGLIA, F. C.

(1990). Lactate metabolism in normal and growth-retarded human fetuses. Pediatr Res 28, 652-656.

MARCONI, A. M., PAOLINI, C., BUSCAGLIA, M., ZERBE, G., BATTAGLIA, F. C. & PARDI, G.

(1996). The impact of gestational age and fetal growth on the maternal-fetal glucose concentration difference. Obstet Gynecol 87, 937-942.

MARCONI, A. M., PAOLINI, C. L., STRAMARE, L., CETIN, I., FENNESSEY, P. V., PARDI, G. &

BATTAGLIA, F. C. (1999). Steady state maternal-fetal leucine enrichments in normal and intrauterine growth-restricted pregnancies. Pediatr Res 46, 114-119.

MARKS, K. A., MALLARD, E. C., ROBERTS, I., WILLIAMS, C. E., SIRIMANNE, E. S., JOHNSTON,

B., GLUCKMAN, P. D. & EDWARDS, A. D. (1996). Delayed vasodilation and altered oxygenation after cerebral ischemia in fetal sheep. Pediatr Res 39, 48-54.

MARTIN, C. B., JR., DE HAAN, J., VAN DER WILDT, B., JONGSMA, H. W., DIELEMAN, A. &

ARTS, T. H. (1979). Mechanisms of late decelerations in the fetal heart rate. A study with autonomic blocking agents in fetal lambs. Eur J Obstet Gynecol Reprod Biol 9, 361-373.

MARUMO, G., KOZUMA, S., OHYU, J., HAMAI, Y., MACHIDA, Y., KOBAYASHI, K., RYO, E.,

UNNO, N., FUJII, T., BABA, K., OKAI, T., TAKASHIMA, S. & TAKETANI, Y. (2001). Generation of periventricular leukomalacia by repeated umbilical cord occlusion in

322

near-term fetal sheep and its possible pathogenetical mechanisms. Biol Neonate 79, 39-45.

MARX, G. F. & GREENE, N. M. (1964). Maternal lactate, pyruvate and excess lactate

production during labour and delivery. Am J Obstet Gynecol 90, 786-793. MASAOKA, N., HAYAKAWA, Y., OHGAME, S., SAKATA, H., SATOH, K. & TAKAHASHI, H.

(1998). Changes in purine metabolism and production of oxygen free radicals by intermittent partial umbilical cord occlusion in chronically instrumented fetal lambs. J Obstet Gynaecol Res 24, 63-71.

MATTSON, M. P. & CHENG, B. (1993). Growth factors protect neurons against

excitotoxic/ischemic damage by stabilizing calcium homeostasis. Stroke 24, I136-140; discussion I144-135.

MAULIK, D. (1998). The molecular mechanisms of fetal brain injury: the role of oxygen free

radicals. In Asphyxia and fetal brain damage. ed. MAULIK, D., pp. 53-78. Wiley-Liss, New York.

MCCORD, J. M. (1985). Oxygen-derived free radicals in postischemic tissue injury. N Engl J

Med 312, 159-163. MCCOWAN, L. M., ERSKINE, L. A. & RITCHIE, K. (1987). Umbilical artery Doppler blood

flow studies in the preterm, small for gestational age fetus. Am J Obstet Gynecol 156, 655-659.

MCHRC. (2000). Confidential enquiry into stillbirths and deaths in infancy. Seventh annual

report, 1 January - 31 December 1998. Maternal and Child Health Research Consortium, London.

MERCURI, E., RUTHERFORD, M., BARNETT, A., FOGLIA, C., HAATAJA, L., COUNSELL, S.,

COWAN, F. & DUBOWITZ, L. (2002). MRI lesions and infants with neonatal encephalopathy. Is the Apgar score predictive? Neuropediatrics 33, 150-156.

MERCURI, E., RUTHERFORD, M., COWAN, F., PENNOCK, J., COUNSELL, S., PAPADIMITRIOU,

M., AZZOPARDI, D., BYDDER, G. & DUBOWITZ, L. (1999). Early prognostic indicators of outcome in infants with neonatal cerebral infarction: a clinical, electroencephalogram, and magnetic resonance imaging study. Pediatrics 103, 39-46.

MESCHIA, G. (1985). Safety margin of fetal oxygenation. J Reprod Med 30, 308-311. MILLEY, J. R. (1988). Uptake of exogenous substrates during hypoxia in fetal lambs. Am J

Physiol 254, E572-578. MINCHOM, P., NISWANDER, K., CHALMERS, I., DAUNCEY, M., NEWCOMBE, R., ELBOURNE,

D., MUTCH, L., ANDREWS, J. & WILLIAMS, G. (1987). Antecedents and outcome of very early neonatal seizures in infants born at or after term. Br J Obstet Gynaecol 94, 431-439.

MISHRA, O. P. & DELIVORIA-PAPADOPOULOS, M. (1999). Cellular mechanisms of hypoxic

injury in the developing brain. Brain Res Bull 48, 233-238.

323

MONGELLI, M., WANG, C. C., WANG, W., PANG, C. C. & ROGERS, M. S. (1997). Oxygen free radical activity in the second stage of labor. Acta Obstet Gynecol Scand 76, 765-768.

MORISHIMA, H. O., PEDERSEN, H., SANTOS, A. C., SCHAPIRO, H. M., FINSTER, M., ARTHUR,

G. R. & COVINO, B. G. (1989). Adverse effects of maternally administered lidocaine on the asphyxiated preterm fetal lamb. Anesthesiology 71, 110-115.

MORROW, R. J., ADAMSON, S. L., BULL, S. B. & RITCHIE, J. W. (1989). Effect of placental

embolization on the umbilical arterial velocity waveform in fetal sheep. Am J Obstet Gynecol 161, 1055-1060.

MORTON, M., DUKE, P. C. & ONG, B. (1980). Baroreflex control of heart rate in man awake

and during enflurane and enflurane--nitrous oxide anesthesia. Anesthesiology 52, 221-223.

MUROTSUKI, J., BOCKING, A. D. & GAGNON, R. (1997a). Fetal heart rate patterns in growth-

restricted fetal sheep induced by chronic fetal placental embolization. Am J Obstet Gynecol 176, 282-290.

MUROTSUKI, J., CHALLIS, J. R., HAN, V. K., FRAHER, L. J. & GAGNON, R. (1997b). Chronic

fetal placental embolization and hypoxemia cause hypertension and myocardial hypertrophy in fetal sheep. Am J Physiol 272, R201-207.

MUTCH, L., ALBERMAN, E., HAGBERG, B., KODAMA, K. & PERAT, M. V. (1992). Cerebral

palsy epidemiology: where are we now and where are we going? Dev Med Child Neurol 34, 547-551.

MYERS, R., WAGNER, K. & COURTEN, G., ed. (1981). Lactic acid accumulation in tissues as cause of

brain injury and death in cardiogenic shock from asphyxia. In: Clinical perinatal biochemical monitoring. Williams & Wilkins, Baltimore.

MYERS, R. E. (1972). Two patterns of perinatal brain damage and their conditions of

occurrence. Am J Obstet Gynecol 112, 246-276. MYERS, R. E. (1975). Four patterns of perinatal brain damage and their conditions of

occurrence in primates. 10, 223-234. MYERS, R. E. (1977). Experimental models of perinatal brain damage: relevance to human

pathology. In Intrauterine asphyxia and the developing brain. ed. GLUCK, L., pp. 37-97. Year Book Medical Publishers Inc, Chicago.

NAEYE, R. L., PETERS, E. C., BARTHOLOMEW, M. & LANDIS, J. R. (1989). Origins of cerebral

palsy. Am J Dis Child 143, 1154-1161. NAGEL, H. T., VANDENBUSSCHE, F. P., OEPKES, D., JENNEKENS-SCHINKEL, A., LAAN, L. A.

& GRAVENHORST, J. B. (1995). Follow-up of children born with an umbilical arterial blood pH < 7. Am J Obstet Gynecol 173, 1758-1764.

NAKAJIMA, W., ISHIDA, A., LANGE, M. S., GABRIELSON, K. L., WILSON, M. A., MARTIN, L. J.,

BLUE, M. E. & JOHNSTON, M. V. (2000). Apoptosis has a prolonged role in the neurodegeneration after hypoxic ischemia in the newborn rat. J Neurosci 20, 7994-8004.

324

NCCLS. (1995). Method comparison and bias estimation using patient samples; Approved guideline, NCCLS.

NELDAM, S., OSLER, M., HANSEN, P. K., NIM, J., SMITH, S. F. & HERTEL, J. (1986).

Intrapartum fetal heart rate monitoring in a combined low- and high-risk population: a controlled clinical trial. Eur J Obstet Gynecol Reprod Biol 23, 1-11.

NELSON, K. B. (1988). What proportion of cerebral palsy is related to birth asphyxia? J Pediatr

112, 572-574. NELSON, K. B. (1989). Relationship of intrapartum and delivery room events to long-term

neurologic outcome. Clin Perinatol 16, 995-1007. NELSON, K. B., DAMBROSIA, J. M., GRETHER, J. K. & PHILLIPS, T. M. (1998). Neonatal

cytokines and coagulation factors in children with cerebral palsy. Ann Neurol 44, 665-675.

NELSON, K. B., DAMBROSIA, J. M., TING, T. Y. & GRETHER, J. K. (1996). Uncertain value of

electronic fetal monitoring in predicting cerebral palsy. N Engl J Med 334, 613-618. NELSON, K. B. & ELLENBERG, J. H. (1984). Obstetric complications as risk factors for

cerebral palsy or seizure disorders. Jama 251, 1843-1848. NELSON, K. B. & ELLENBERG, J. H. (1986). Antecedents of cerebral palsy. Multivariate

analysis of risk. N Engl J Med 315, 81-86. NELSON, K. B. & EMERY, E. S., 3RD. (1993). Birth asphyxia and the neonatal brain: what do

we know and when do we know it? Clin Perinatol 20, 327-344. NELSON, K. B. & GRETHER, J. K. (1998). Potentially asphyxiating conditions and spastic

cerebral palsy in infants of normal birth weight. Am J Obstet Gynecol 179, 507-513. NELSON, K. B. & LEVITON, A. (1991). How much of neonatal encephalopathy is due to birth

asphyxia? Am J Dis Child 145, 1325-1331. NEWMAN, J. P., PEEBLES, D. M., HARDING, S. R., SPRINGETT, R. & HANSON, M. A. (2000).

Hemodynamic and metabolic responses to moderate asphyxia in brain and skeletal muscle of late-gestation fetal sheep. J Appl Physiol 88, 82-90.

NEWNHAM, J. P., LAM, R. W., HOBEL, C. J., PADBURY, J. F., POLK, D. H. & FISHER, D. A.

(1986). Differential response of ovine placental lactogen levels in maternal and fetal circulations following single umbilical artery ligation in fetal sheep. Placenta 7, 51-64.

NICH. (1997). Electronic fetal heart rate monitoring: research guidelines for interpretation.

National Institute of Child Health and Human Development Research Planning Workshop. Am J Obstet Gynecol 177, 1385-1390.

NICOLAIDES, K. H., BILARDO, C. M., SOOTHILL, P. W. & CAMPBELL, S. (1988). Absence of

end diastolic frequencies in umbilical artery: a sign of fetal hypoxia and acidosis. Bmj 297, 1026-1027.

325

NICOLAIDES, K. H., ECONOMIDES, D. L. & SOOTHILL, P. W. (1989). Blood gases, pH, and lactate in appropriate- and small-for-gestational-age fetuses. Am J Obstet Gynecol 161, 996-1001.

NIELSEN, P. V., STIGSBY, B., NICKELSEN, C. & NIM, J. (1987). Intra- and inter-observer

variability in the assessment of intrapartum cardiotocograms. Acta Obstet Gynecol Scand 66, 421-424.

NIETO, A., VILLAR, J., MATORRAS, R., SERRA, M., VALENZUELA, P. & KELLER, J. (1994).

Intrauterine growth retardation: fluctuation of fetal pH measured between beginning and at the completion of labor. J Perinat Med 22, 329-335.

NITSOS, I., MOSS, T. J., COCK, M. L., HARDING, R. & NEWNHAM, J. P. (2002). Fetal responses

to intra-amniotic endotoxin in sheep. J Soc Gynecol Investig 9, 80-85. NORDSTROM, L. (1995). Fetal lactate levles during albour and at delivery determined with test

strip methods (Thesis). In Department of Obstetrics & Gynaecology, Karolinska Institute, Huddinge University Hospital, Stockholm, Sweden.

NORDSTROM, L. (2001). Lactate measurements in scalp and cord arterial blood. Curr Opin

Obstet Gynecol 13, 141-145. NORDSTROM, L., ACHANNA, S., NAKA, K. & ARULKUMARAN, S. (2001). Fetal and maternal

lactate increase during active second stage of labour. Bjog 108, 263-268. NORDSTROM, L. & ARULKUMARAN, S. (1998). Intrapartum fetal hypoxia and biochemical

markers: a review. Obstet Gynecol Surv 53, 645-657. NORDSTROM, L., ARULKUMARAN, S., CHUA, S., RATNAM, S., INGEMARSSON, I., KUBLICKAS,

M., PERSSON, B., SHIMOJO, N. & WESTGREN, M. (1995a). Continuous maternal glucose infusion during labor: effects on maternal and fetal glucose and lactate levels. Am J Perinatol 12, 357-362.

NORDSTROM, L., CHUA, S., ROY, A. & ARULKUMARAN, S. (1998a). Quality assessment of two

lactate test strip methods suitable for obstetric use. J Perinat Med 26, 83-88. NORDSTROM, L., CHUA, S., ROY, A., NAKA, K., PERSSON, B. & ARULKUMARAN, S. (1998b).

Lactate, lactate/pyruvate ratio and catecholamine interrelations in cord blood at delivery in complicated pregnancies. Early Hum Dev 52, 87-94.

NORDSTROM, L., INGEMARSSON, I., KUBLICKAS, M., PERSSON, B., SHIMOJO, N. &

WESTGREN, M. (1995b). Scalp blood lactate: a new test strip method for monitoring fetal wellbeing in labour. Br J Obstet Gynaecol 102, 894-899.

NORDSTROM, L., INGEMARSSON, I., PERSSON, B., SHIMOJO, N. & WESTGREN, M. (1994).

Lactate in fetal scalp blood and umbilical artery blood measured during normal labor with a test strip method. Acta Obstet Gynecol Scand 73, 250-254.

NORDSTROM, L., INGEMARSSON, I. & WESTGREN, M. (1996a). Fetal monitoring with lactate.

Baillieres Clin Obstet Gynaecol 10, 225-242.

326

NORDSTROM, L., MARCUS, C., PERSSON, B., SHIMOJO, N. & WESTGREN, M. (1996b). Lactate in cord blood and its relationship to pH and catecholamines in spontaneous vaginal deliveries. Early Hum Dev 46, 97-104.

NORMAN, L. J., LYE, S. J., WLODEK, M. E. & CHALLIS, J. R. (1985). Changes in pituitary

responses to synthetic ovine corticotrophin releasing factor in fetal sheep. Can J Physiol Pharmacol 63, 1398-1403.

NORTHINGTON, F. J., FERRIERO, D. M., FLOCK, D. L. & MARTIN, L. J. (2001). Delayed

neurodegeneration in neonatal rat thalamus after hypoxia-ischemia is apoptosis. J Neurosci 21, 1931-1938.

OHYU, J., MARUMO, G., OZAWA, H., TAKASHIMA, S., NAKAJIMA, K., KOHSAKA, S., HAMAI,

Y., MACHIDA, Y., KOBAYASHI, K., RYO, E., BABA, K., KOZUMA, S., OKAI, T. & TAKETANI, Y. (1999). Early axonal and glial pathology in fetal sheep brains with leukomalacia induced by repeated umbilical cord occlusion. Brain Dev 21, 248-252.

OKAMURA, K., TANIGAWARA, S., SHINTAKU, Y., WATANABE, T., ENDOH, H., AKIYAMA, N.,

IWAMOTO, M., SAITO, J., AKAGI, K., YAJIMA, A. & ET AL. (1989). Alteration of FHR pattern and cerebral metabolic rate of glucose of the fetus measured by positron emission tomography during progress of acidemia. The significance of overshoot acceleration in FHR. J Perinat Med 17, 289-295.

OMAR, H. A., FIGUEROA, R., TEJANI, N. & WOLIN, M. S. (1993). Properties of a lactate-

induced relaxation in human placental arteries and veins. Am J Obstet Gynecol 169, 912-918.

OSLER, W. (1889). Aequanimitas. In The Collected Essays of Sir William Osler, vol. 1. ed.

MCGOVERN, J. P. & ROLAND, C. G., pp. 7. Classics of Medicine Library, 1985, Birmingham.

OSLER, W. (1892). The Student Life [Farewell address given to American and Canadian

Medical students, 1892]. In Aequanimitas, with other addresses to medical students, nurses, and practitioners of medicine, pp. 400-401. Blakiston's Son & Co., Philadelphia, 1925.

OSLER, W. (1899). After Twenty-five Years [Address given at the McGill University,

September 21, 1899]. In Aequanimitas, with other addresses to medical students, nurses, and practitioners of medicine, pp. 213. Blakiston's Son & Co., Philadephia, 1925.

PALAHNIUK, R. J. & CUMMING, M. (1977). Foetal deterioration following thiopentone-nitrous

oxide anaesthesia in the pregnant ewe. Can Anaesth Soc J 24, 361-370. PALAHNIUK, R. J. & SHNIDER, S. M. (1974). Maternal and fetal cardiovascular and acid-base

changes during halothane and isoflurane anesthesia in the pregnant ewe. Anesthesiology 41, 462-472.

PALMER, C., BRUCKLACHER, R. M., CHRISTENSEN, M. A. & VANNUCCI, R. C. (1990).

Carbohydrate and energy metabolism during the evolution of hypoxic-ischemic brain damage in the immature rat. J Cereb Blood Flow Metab 10, 227-235.

PARDI, G., BUSCAGLIA, M., FERRAZZI, E., BOZZETTI, P., MARCONI, A. M., CETIN, I.,

BATTAGLIA, F. C. & MAKOWSKI, E. L. (1987). Cord sampling for the evaluation of

327

oxygenation and acid-base balance in growth-retarded human fetuses. Am J Obstet Gynecol 157, 1221-1228.

PARDI, G., MARCONI, A. M. & CETIN, I. (2002). Placental-fetal interrelationship in IUGR

fetuses--a review. Placenta 23 Suppl A, S136-141. PARER, J. (1999). Fetal heart rate. In Maternal Fetal Medicine. ed. CREASY, R. K. & RESNIK, R.,

pp. 270-299. W.B. Saunders Co., Philadelphia. PARER, J., LAROS, R., HEILBRON, D. & KREUGER, T. (1981). The roles of parasympathetic

and beta-adrenergic activity in beat-to-beat fetal heart rate variability. In Cardiovascular Physiology-Heart, Peripheral Circulation, and Methodology: Proceedings of the 28th International Congress of Physiological Sciences, Budapest, 1980. (Advances in Physiological Science, Vol 8, pp 327-329). ed. KOVACH, A., MONOS, E. & RUBANYI, G. Pergamon, New York.

PARER, J. T. (1979). Fetal heart-rate monitoring. Lancet 2, 632-633. PARER, J. T., DIJKSTRA, H. R., VREDEBREGT, P. P., HARRIS, J. L., KRUEGER, T. R. & REUSS,

M. L. (1980). Increased fetal heart rate variability with acute hypoxia in chronically instrumented sheep. Eur J Obstet Gynecol Reprod Biol 10, 393-399.

PARER, J. T. & KING, T. (2000). Fetal heart rate monitoring: is it salvageable? Am J Obstet

Gynecol 182, 982-987. PARER, J. T. & LIVINGSTON, E. G. (1990). What is fetal distress? Am J Obstet Gynecol 162,

1421-1425; discussion 1425-1427. PASSING, H. & BABLOK. (1983). A new biometrical procedure for testing the equality of

measurements from two different analytical methods. Application of linear regression procedures for method comparison studies in clinical chemistry, Part I. J Clin Chem Clin Biochem 21, 709-720.

PATTERSON, C. A., GRAVES, W. L., BUGG, G., SASSO, S. C. & BRANN, A. W., JR. (1989).

Antenatal and intrapartum factors associated with the occurrence of seizures in term infant. Obstet Gynecol 74, 361-365.

PAYNE, R. B. (1985). Deming's regression analysis in method comparison studies. Ann Clin

Biochem 22, 430. PECK, T. M. (1980). Physicians' subjectivity in evaluating oxytocin challenge tests. Obstet

Gynecol 56, 13-16. PEEBLES, D. M., EDWARDS, A. D., WYATT, J. S., BISHOP, A. P., COPE, M., DELPY, D. T. &

REYNOLDS, E. O. (1992). Changes in human fetal cerebral hemoglobin concentration and oxygenation during labor measured by near-infrared spectroscopy. Am J Obstet Gynecol 166, 1369-1373.

PEETERS, L. L., SHELDON, R. E., JONES, M. D., JR., MAKOWSKI, E. L. & MESCHIA, G. (1979).

Blood flow to fetal organs as a function of arterial oxygen content. Am J Obstet Gynecol 135, 637-646.

PELLO, L. C., ROSEVEAR, S. K., DAWES, G. S., MOULDEN, M. & REDMAN, C. W. (1991).

Computerized fetal heart rate analysis in labor. Obstet Gynecol 78, 602-610.

328

PELOWSKI, A. & FINER, N. (1992). Birth asphyxia in the term infant. In Effective care of the

newborn infant. ed. SINCLAIR, J. & LUCEY, J., pp. 263-266. Oxford University Press, Oxford.

PENNELL, C. E., NITSOS, I., MOSS, T. J. M. & NEWNHAM, J. P. (2003). Factors influencing

postnatal lung function after oligohydramnios in sheep. J Soc Gynecol Investig 10(Suppl 1), 352A.

PENNELL, C. E. & TRACY, M. B. (1999). A new method for rapid measurement of lactate in

fetal and neonatal blood. Aust N Z J Obstet Gynaecol 39, 227-233. PENNELL, C. E., TRACY, M. B., DOHERTY, D. A. & MURRAY, H. G. (2002). Fetal scalp lactate

for monitoring fetal wellbeing in labor. In Perinatal Society of Australia and New Zealand 6th Annual Congress, Christchurch, New Zealand.

PENNING, D. H., GRAFE, M. R., HAMMOND, R., MATSUDA, Y., PATRICK, J. & RICHARDSON,

B. (1994). Neuropathology of the near-term and midgestation ovine fetal brain after sustained in utero hypoxemia. Am J Obstet Gynecol 170, 1425-1432.

PERLMAN, J. M. (1999). Markers of asphyxia and neonatal brain injury. N Engl J Med 341,

364-365. PERRET, C. & ENRICO, J. F. (1978). Lactate in acute circulatory failure. In Lactate in Acute

Conditions. International Symposium, pp. 69-82, Basel. PHELAN, J. P., AHN, M. O., KORST, L., MARTIN, G. I. & WANG, Y. M. (1998). Intrapartum

fetal asphyxial brain injury with absent multiorgan system dysfunction. J Matern Fetal Med 7, 19-22.

PHILIPSON, E. H., KALHAN, S. C., RIHA, M. M. & PIMENTEL, R. (1987). Effects of maternal

glucose infusion on fetal acid-base status in human pregnancy. Am J Obstet Gynecol 157, 866-873.

PIQUARD, F., SCHAEFER, A., DELLENBACH, P. & HABEREY, P. (1990). Lactate movements in

the term human placenta in situ. Biol Neonate 58, 61-68. POLLOCK, M. A., JEFFERSON, S. G., KANE, J. W., LOMAX, K., MACKINNON, G. & WINNARD,

C., B,. (1992). Method comparison - a different approach. Ann Clin Biochem 29, 556-560.

POLYAK, K., XIA, Y., ZWEIER, J. L., KINZLER, K. W. & VOGELSTEIN, B. (1997). A model for

p53-induced apoptosis. Nature 389, 300-305. PORTER, T. F. & CLARK, S. L. (1999). Vibroacoustic and scalp stimulation. Obstet Gynecol Clin

North Am 26, 657-669. POWELL, O. H., MELVILLE, A. & MACKENNA, J. (1979). Fetal heart rate acceleration in labor:

excellent prognostic indicator. Am J Obstet Gynecol 134, 36-38. PROBERT, L., AKASSOGLOU, K., PASPARAKIS, M., KONTOGEORGOS, G. & KOLLIAS, G.

(1995). Spontaneous inflammatory demyelinating disease in transgenic mice showing

329

central nervous system-specific expression of tumor necrosis factor alpha. Proc Natl Acad Sci U S A 92, 11294-11298.

PSANZ. (1995). The origins of cerebral palsy--a consensus statement. The Australian and

New Zealand Perinatal Societies. Med J Aust 162, 85-90. QUILLIGAN, E. J. (1990). Discussion to "What is fetal distress?" Am J Obstet Gynecol 162,

1425-1426. QUILLIGAN, E. J. & PAUL, R. H. (1975). Fetal monitoring: is it worth it? Obstet Gynecol 45, 96-

100. RADIOMETER. (1997). Blood gas, electrolyte and metabolite systems reference manual; EML 100/105

specifications. Radiometer Medical Pty Ltd, Nunawading, Victoria, Australia. RAMIN, S. M., GILSTRAP, L. C., 3RD, LEVENO, K. J., BURRIS, J. & LITTLE, B. B. (1989).

Umbilical artery acid-base status in the preterm infant. Obstet Gynecol 74, 256-258. RANCK, J. & WINDLE, W. F. (1959). Brain Damage in the monkey Mucacca mulatta by

asphyxia neonatorum. 1, 130-154. RANTAKALLIO, P. (1985). A 14-year follow-up of children with normal and abnormal birth

weight for their gestational age. A population study. Acta Paediatr Scand 74, 62-69. RANZCOG. (2002). Royal Australian and New Zealand College of Obstetricians and

Gynaecologists. Clinical Guidelines - Intrapartum Fetal Surveillance. RCOG. (2001). Royal College of Obstetricians and Gynecologists. Evidence-based Clinical

Guidelines Number 8. May 2001. The use of electronic fetal monitoring. RCPA-AACB. (2002). Handbook for the RCPA-AACB Quality Assurance Program. Royal College

of Pathologists of Australasia and the Austalian Association of Clinical Biochemists, Sydney.

REGNAULT, T. R., DE VRIJER, B. & BATTAGLIA, F. C. (2002). Transport and metabolism of

amino acids in placenta. Endocrine 19, 23-41. REGNAULT, T. R., DE VRIJER, B., GALAN, H. L., DAVIDSEN, M. L., TREMBLER, K. A.,

BATTAGLIA, F. C., WILKENING, R. B. & ANTHONY, R. V. (2003). The relationship between transplacental O2 diffusion and placental expression of PlGF, VEGF and their receptors in a placental insufficiency model of fetal growth restriction. J Physiol.

REGNAULT, T. R., ORBUS, R. J., BATTAGLIA, F. C., WILKENING, R. B. & ANTHONY, R. V.

(1999). Altered arterial concentrations of placental hormones during maximal placental growth in a model of placental insufficiency. J Endocrinol 162, 433-442.

REHNCRONA, S., WESTERBERG, E., AKESSON, B. & SIESJO, B. K. (1982). Brain cortical fatty

acids and phospholipids during and following complete and severe incomplete ischemia. J Neurochem 38, 84-93.

REID, D. L., PARER, J. T., WILLIAMS, K., DARR, D., PHERNETTON, T. M. & RANKIN, J. H.

(1991). Effects of severe reduction in maternal placental blood flow on blood flow distribution in the sheep fetus. J Dev Physiol 15, 183-188.

330

RENOU, P., CHANG, A., ANDERSON, I. & WOOD, C. (1976). Controlled trial of fetal intensive

care. Am J Obstet Gynecol 126, 470-476. RICE, A. C., ZSOLDOS, R., CHEN, T., WILSON, M. S., ALESSANDRI, B., HAMM, R. J. &

BULLOCK, M. R. (2002). Lactate administration attenuates cognitive deficits following traumatic brain injury. Brain Res 928, 156-159.

RICHARD, M. J., PORTAL, B., MEO, J., COUDRAY, C., HADJIAN, A. & FAVIER, A. (1992).

Malondialdehyde kit evaluated for determining plasma and lipoprotein fractions that react with thiobarbituric acid. Clin Chem 38, 704-709.

RICHARDSON, B. S., CARMICHAEL, L., HOMAN, J., JOHNSTON, L. & GAGNON, R. (1996).

Fetal cerebral, circulatory, and metabolic responses during heart rate decelerations with umbilical cord compression. Am J Obstet Gynecol 175, 929-936.

RICHARDSON, B. S., CARMICHAEL, L., HOMAN, J. & PATRICK, J. E. (1993). Cerebral oxidative

metabolism in fetal sheep with prolonged and graded hypoxemia. 10, 77-83. RICHARDSON, B. S., RURAK, D., PATRICK, J. E., HOMAN, J. & CARMICHAEL, L. (1989).

Cerebral oxidative metabolism during sustained hypoxaemia in fetal sheep. J Dev Physiol 11, 37-43.

RICHER, L. P., SHEVELL, M. I. & MILLER, S. P. (2001). Diagnostic profile of neonatal

hypotonia: an 11-year study. Pediatr Neurol 25, 32-37. RILEY, R. J. & JOHNSON, J. W. (1993). Collecting and analyzing cord blood gases. Clin Obstet

Gynecol 36, 13-23. ROBERTSON, C. M., FINER, N. N. & GRACE, M. G. (1989). School performance of survivors

of neonatal encephalopathy associated with birth asphyxia at term. J Pediatr 114, 753-760.

ROGERS, M. S., MONGELLI, M., TSANG, K. H. & WANG, C. C. (1999). Fetal and maternal

levels of lipid peroxides in term pregnancies. Acta Obstet Gynecol Scand 78, 120-124. ROGERS, M. S., MURRAY, H. G., WANG, C. C., PENNELL, C. E., TURNER, A., YAN, P., PANG,

C. C. & CHANG, A. M. (2001). Oxidative stress in the fetal lamb brain following intermittent umbilical cord occlusion: a path analysis. Bjog 108, 1283-1290.

ROGERS, M. S., WANG, W., MONGELLI, M., PANG, C. P., DULEY, J. A. & CHANG, A. M.

(1997). Lipid peroxidation in cord blood at birth: a marker of fetal hypoxia during labour. Gynecol Obstet Invest 44, 229-233.

ROOTH, G. & NILSSON, I. (1964). Studies of foetal and maternal metabolic acidosis. Clin Sci

26, 121-132. ROS, J., PECINSKA, N., ALESSANDRI, B., LANDOLT, H. & FILLENZ, M. (2001). Lactate reduces

glutamate-induced neurotoxicity in rat cortex. J Neurosci Res 66, 790-794. ROSEN, K. G., HRBEK, A., KARLSSON, K. & KJELLMER, I. (1986). Fetal cerebral,

cardiovascular and metabolic reactions to intermittent occlusion of ovine maternal placental blood flow. Acta Physiol Scand 126, 209-216.

331

ROSENBERG, P. A. (1997). Potential therapeutic intervention following hypoxic-ischemic

insult. MRDD Res Rev 3, 1-9. ROSENBLOOM, L. (1994). Dyskinetic cerebral palsy and birth asphyxia. Dev Med Child Neurol

36, 285-289. ROSS, M. G. & GALA, R. (2002). Use of umbilical artery base excess: algorithm for the timing

of hypoxic injury. Am J Obstet Gynecol 187, 1-9. ROTH, S. C., BAUDIN, J., CADY, E., JOHAL, K., TOWNSEND, J. P., WYATT, J. S., REYNOLDS, E.

O. & STEWART, A. L. (1997). Relation of deranged neonatal cerebral oxidative metabolism with neurodevelopmental outcome and head circumference at 4 years. Dev Med Child Neurol 39, 718-725.

RUDOLPH, A. M. & HEYMANN, M. A. (1967). The circulation of the fetus in utero. Methods

for studying distribution of blood flow, cardiac output and organ blood flow. Circ Res 21, 163-184.

RURAK, D. W. (1978). Plasma vasopressin levels during hypoxaemia and the cardiovascular

effects of exogenous vasopressin in foetal and adult sheep. J Physiol 277, 341-357. RURAK, D. W., RICHARDSON, B. S., PATRICK, J. E., CARMICHAEL, L. & HOMAN, J. (1990).

Blood flow and oxygen delivery to fetal organs and tissues during sustained hypoxemia. Am J Physiol 258, R1116-1122.

RUSSELL, G. A., JEFFERS, G. & COOKE, R. W. (1992). Plasma hypoxanthine: a marker for

hypoxic-ischaemic induced periventricular leucomalacia? Arch Dis Child 67, 388-392. RUTH, V. J. & RAIVIO, K. O. (1988). Perinatal brain damage: predictive value of metabolic

acidosis and the Apgar score. Bmj 297, 24-27. RUTHERFORD, M., PENNOCK, J., SCHWIESO, J., COWAN, F. & DUBOWITZ, L. (1996).

Hypoxic-ischaemic encephalopathy: early and late magnetic resonance imaging findings in relation to outcome. Arch Dis Child Fetal Neonatal Ed 75, F145-151.

SAHLA, S. D., MORISHIMA, H. O., JAMES, L. S. & ADAMSON, K. J. (1964). Lactate and

pyruvate gradients between red blood cells and plasma during acute asphyxia. J Appl Physiol 19, 1100-1104.

SAIKUMAR, P., DONG, Z., WEINBERG, J. M. & VENKATACHALAM, M. A. (1998). Mechanisms

of cell death in hypoxia/reoxygenation injury. Oncogene 17, 3341-3349. SAITO, J., OKAMURA, K., AKAGI, K., TANIGAWARA, S., SHINTAKU, Y., WATANABE, T.,

AKIYAMA, N., ENDO, C., SATO, A. & YAJIMA, A. (1988). Alteration of FHR pattern associated with progressively advanced fetal acidemia caused by cord compression. Nippon Sanka Fujinka Gakkai Zasshi 40, 775-780.

SALING, E. (1962). Neues Vorgehen zur Untertesuchung des Kindes unter der Geburt-

Einfuhrung, Grundlagen. Arch Gynak 197, 108-122. SARNAT, H. B. & SARNAT, M. S. (1976). Neonatal encephalopathy following fetal distress. A

clinical and electroencephalographic study. Arch Neurol 33, 696-705.

332

SAUGSTAD, O. D. (2001). Resuscitation of the asphyxic newborn infant: new insight leads to

new therapeutic possibilities. Biol Neonate 79, 258-260. SCHELLER, J. M. & NELSON, K. B. (1994). Does cesarean delivery prevent cerebral palsy or

other neurologic problems of childhood? Obstet Gynecol 83, 624-630. SCHIFRIN, B. S., HAMILTON-RUBINSTEIN, T. & SHIELDS, J. R. (1994). Fetal heart rate patterns

and the timing of fetal injury. J Perinatol 14, 174-181. SCHIFRIN, B. S. & LONGO, L. D. (2000). William John Little and cerebral palsy. A reappraisal.

Eur J Obstet Gynecol Reprod Biol 90, 139-144. SCHMIDT, S. (1973). Glukose, laktat und pyruvat in der schwangerschaft und unter der

geburt. Forschritte der Geburtshilfe und Gynaekologie 50, 1-99. SCHMIDT, S. (1988). Human research on lactate. J Perinat Med 16 Suppl 1, 131-137. SCHNEIDER, H., CHALLIER, J. C. & DANCIS, J. (1981). Transfer and metaboism of glucose

and lactate in the human placenta studied by a perfusion system in vitro. Placenta 2(Suppl), 129-134.

SCHOUSBOE, A., WESTERGAARD, N., WAAGEPETERSEN, H. S., LARSSON, O. M., BAKKEN, I.

J. & SONNEWALD, U. (1997). Trafficking between glia and neurons of TCA cycle intermediates and related metabolites. Glia 21, 99-105.

SCHURR, A., MILLER, J. J., PAYNE, R. S. & RIGOR, B. M. (1999). An increase in lactate output

by brain tissue serves to meet the energy needs of glutamate-activated neurons. J Neurosci 19, 34-39.

SCHURR, A., PAYNE, R. S., MILLER, J. J. & RIGOR, B. M. (1997). Brain lactate, not glucose,

fuels the recovery of synaptic function from hypoxia upon reoxygenation: an in vitro study. Brain Res 744, 105-111.

SCHURR, A. & RIGOR, B. M. (1998). Brain anaerobic lactate production: a suicide note or a

survival kit? Dev Neurosci 20, 348-357. SEGAR, J. L., MERRILL, D. C., SMITH, B. A. & ROBILLARD, J. E. (1994). Role of sympathetic

activity in the generation of heart rate and arterial pressure variability in fetal sheep. Pediatr Res 35, 250-254.

SEGAWA, H., MORI, K., MURAKAWA, M., KASAI, K., SHIRAKAMI, G., ADACHI, T. & ARAI, T.

(1998). Isoflurane and sevoflurane augment norepinephrine responses to surgical noxious stimulation in humans. Anesthesiology 89, 1407-1413.

SEHDEV, H. M., STAMILIO, D. M., MACONES, G. A., GRAHAM, E. & MORGAN, M. A. (1997).

Predictive factors for neonatal morbidity in neonates with an umbilical arterial cord pH less than 7.00. Am J Obstet Gynecol 177, 1030-1034.

SEIDL, R., STOCKLER-IPSIROGLU, S., ROLINSKI, B., KOHLHAUSER, C., HERKNER, K. R.,

LUBEC, B. & LUBEC, G. (2000). Energy metabolism in graded perinatal asphyxia of the rat. Life Sci 67, 421-435.

333

SELMAJ, K., RAINE, C. S., FAROOQ, M., NORTON, W. T. & BROSNAN, C. F. (1991). Cytokine cytotoxicity against oligodendrocytes. Apoptosis induced by lymphotoxin. J Immunol 147, 1522-1529.

SELMAJ, K., WALCZAK, A., MYCKO, M., BERKOWICZ, T., KOHNO, T. & RAINE, C. S. (1998).

Suppression of experimental autoimmune encephalomyelitis with a TNF binding protein (TNFbp) correlates with down-regulation of VCAM-1/VLA-4. Eur J Immunol 28, 2035-2044.

SHEINER, E., HADAR, A., HALLAK, M., KATZ, M., MAZOR, M. & SHOHAM-VARDI, I. (2001).

Clinical significance of fetal heart rate tracings during the second stage of labor. Obstet Gynecol 97, 747-752.

SHENKER, L., POST, R. C. & SEILER, J. S. (1975). Routine electronic monitoring of fetal heart

rate and uterine activity during labor. Obstet Gynecol 46, 185-189. SHIELDS, J. R. & SCHIFRIN, B. S. (1988). Perinatal antecedents of cerebral palsy. Obstet Gynecol

71, 899-905. SHIER, R. W. & DILTS, P. V., JR. (1972). Comparison of fetal scalp, carotid, and jugular blood

gas values and oxygen consumption in the lamb. Am J Obstet Gynecol 112, 397-403. SHIMOJO, N., NAKA, K., NAKAJIMA, C., YOSHIKAWA, C., OKUDA, K. & OKADA, K. (1989).

Test-strip method for measuring lactate in whole blood. Clin Chem 35, 1992-1994. SHIMOJO, N., NAKA, K., UENOYAMA, H., HAMAMOTO, K., YOSHIOKA, K. & OKUDA, K.

(1993). Electrochemical assay system with single-use electrode strip for measuring lactate in whole blood. Clin Chem 39, 2312-2314.

SHY, K. K., LUTHY, D. A., BENNETT, F. C., WHITFIELD, M., LARSON, E. B., VAN BELLE, G.,

HUGHES, J. P., WILSON, J. A. & STENCHEVER, M. A. (1990). Effects of electronic fetal-heart-rate monitoring, as compared with periodic auscultation, on the neurologic development of premature infants. N Engl J Med 322, 588-593.

SIEGEL, G. J., AGRANOFF, B. W., ALBERS, R. W. & MOLINOFF, P. (1989). Basic Neurochemistry.

Raven Press, New York. SIESJO, B. K. (1988). Historical overview. Calcium, ischemia, and death of brain cells. Ann N

Y Acad Sci 522, 638-661. SIESJO, B. K. (1992). Pathophysiology and treatment of focal cerebral ischemia. Part I:

Pathophysiology. J Neurosurg 77, 169-184. SINGH, A. K., JIANG, Y., WHITE, T. & SPASSOVA, D. (1997). Validation of nonradioactive

chemiluminescent immunoassay methods for the analysis of thyroxine and cortisol in blood samples obtained from dogs, cats, and horses. J Vet Diagn Invest 9, 261-268.

SINN, J. K., LLOYD, J., TODD, D. A., LAZARUS, R., MAESEL, A. & JOHN, E. (2001). Umbilical

cord blood lactate in normal infants: comparison between two methods of measurement. J Paediatr Child Health 37, 24-27.

334

SKOVSTED, P., PRICE, M. L. & PRICE, H. L. (1970). The effects of short-acting barbiturates on arterial pressure, preganglionic sympathetic activity and barostatic reflexes. Anesthesiology 33, 10-18.

SKUPSKI, D. W. & EGLINTON, G. S. (2002). Intrapartum fetal stimulation tests: a meta-

analysis. Obstet Gynecol 100, 830. SLATER, T. F. & SAWYER, B. C. (1971). The stimulatory effects of carbon tetrachloride and

other halogenoalkanes on peroxidative reactions in rat liver fractions in vitro. General features of the systems used. Biochem J 123, 805-814.

SMITH, A., WILES, C., HAAN, E., MCGILL, J., WALLACE, G., DIXON, J., SELBY, R., COLLEY,

A., MARKS, R. & TRENT, R. J. (1996). Clinical features in 27 patients with Angelman syndrome resulting from DNA deletion. J Med Genet 33, 107-112.

SMITH, J. B., MANNING, F. A. & PALAHNIUK, R. J. (1975). Maternal and foetal effects of

methoxyflurane anaesthesia in the pregnant ewe. Can Anaesth Soc J 22, 449-459. SMITH, N. C., SOUTTER, W. P., SHARP, F., MCCOLL, J. & FORD, I. (1983). Fetal scalp blood

lactate as an indicator of intrapartum hypoxia. Br J Obstet Gynaecol 90, 821-831. SNIJDERS, R. J., RIBBERT, L. S., VISSER, G. H. & MULDER, E. J. (1992). Numeric analysis of

heart rate variation in intrauterine growth-retarded fetuses: a longitudinal study. Am J Obstet Gynecol 166, 22-27.

SOCOL, M. L., GARCIA, P. M. & RITER, S. (1994). Depressed Apgar scores, acid-base status,

and neurologic outcome. Am J Obstet Gynecol 170, 991-998; discussion 998-999. SOGC. (1996a). Policy statement of the Task Force on Cerebral Palsy and Neonatal

Asphyxia of the Society of Obstetricians and Gynaecologists of Canada (Part I). J Soc Obstet Gynecol Can, 1267-1279.

SOGC. (1996b). Society of Obstetricians and Gynaecologists of Canada Task Force on

cerebral palsy and feal asphyxia. J Soc Obstet Gynaecol Can 18, 1267-1279. SOLAROGLU, I., SOLAROGLU, A., KAPTANOGLU, E., DEDE, S., HABERAL, A., BESKONAKLI,

E. & KILINC, K. (2003). Erythropoietin prevents ischemia-reperfusion from inducing oxidative damage in fetal rat brain. Childs Nerv Syst 19, 19-22.

SOOTHIL, P. W., NICOLAIDES, K. H. & CAMPBELL, S. (1987). Prenatal asphyxia,

hyperlactemia, hypoglycaemia, and erythroblastosis in growth retarded fetuses. Bmj 294, 1051-1056.

SOUTTER, W. P., SHARP, F. & CLARK, D. M. (1978). Bedside estimation of whole blood

lactate. Br J Anaesth 50, 445-450. SPENCER, J. A., BADAWI, N., BURTON, P., KEOGH, J., PEMBERTON, P. & STANLEY, F. (1997).

The intrapartum CTG prior to neonatal encephalopathy at term: a case-control study. Br J Obstet Gynaecol 104, 25-28.

SPINILLO, A., CAPUZZO, E., EGBE, T. O., FAZZI, E., COLONNA, L. & NICOLA, S. (1995).

Pregnancies complicated by idiopathic intrauterine growth retardation. Severity of

335

growth failure, neonatal morbidity and two-year infant neurodevelopmental outcome. J Reprod Med 40, 209-215.

SPONG, C. Y., RASUL, C., COLLEA, J. V., EGLINTON, G. S. & GHIDINI, A. (1998).

Characterization and prognostic significance of variable decelerations in the second stage of labor. Am J Perinatol 15, 369-374.

STANLEY, F. J. & ALBERMAN, E. (1984). The epidemiology of the cerebral palsies. Lippincott,

Philadelphia. STANLEY, F. J., BLAIR, E. & ALBERMAN, E. (2000). Cerebral palsies: epidemiology and causal

pathways. In Clinics in Developmental Medicine, vol. 151, pp. 83-98. Mac Keith Press, London.

STANLEY, F. J., BLAIR, E., HOCKEY, A., PETTERSON, B. & WATSON, L. (1993). Spastic

quadriplegia in Western Australia: a genetic epidemiological study. I: Case population and perinatal risk factors. Dev Med Child Neurol 35, 191-201.

STEER, C. & BONNEY, W. (1962). Obstetric factors in cerebral palsy. Am J Obstet Gynecol 83,

526. STEER, P. (1982). Has the expression 'fetal distress' outlived its usefulness? Br J Obstet Gynaecol

89, 690-693. STEER, P. (1999). ABC of labour care: assessment of mother and fetus in labour. Bmj 318,

858-861. STEPHENS, S. R., CLARK, R. B., BEARD, A. G. & BROWN, W. E. (1969). Validity of scalp

blood sampling. Comparison of scalp and arterial blood in the newborn. Am J Obstet Gynecol 105, 616-619.

STREET, P., DAWES, G. S., MOULDEN, M. & REDMAN, C. W. (1991). Short-term variation in

abnormal antenatal fetal heart rate records. Am J Obstet Gynecol 165, 515-523. STURBOIS, G., UZAN, S., ROTTEN, D., BREART, G. & SUREAU, C. (1977). Continuous

subcutaneous pH measurement in human fetuses: correlations with scalp and umbilical blood pH. Am J Obstet Gynecol 128, 901-903.

SUIDAN, J. S., ANTOINE, C., SILVERMAN, F., LUSTIG, I. D., WASSERMAN, J. F. & YOUNG, B.

K. (1984a). Human maternal-fetal lactate relationships. J Perinat Med 12, 211-217. SUIDAN, J. S., WASSERMAN, J. F. & YOUNG, B. K. (1984b). Placental contribution to lactate

production by the human fetoplacental unit. Am J Perinatol 1, 306-309. SUIDAN, J. S. & YOUNG, B. K. (1984). Outcome of fetuses with lactic acidemia. Am J Obstet

Gynecol 150, 33-37. SUREAU, C. (1996). Historical perspectives: forgotten past, unpredictable future. Baillieres Clin

Obstet Gynaecol 10, 167-184. SWARTZ, J., CUMMING, M. & BIEHL, D. (1987). The effect of ketamine anaesthesia on the

acidotic fetal lamb. Can J Anaesth 34, 233-237.

336

SYKES, G. S., MOLLOY, P. M., JOHNSON, P., GU, W., ASHWORTH, F., STIRRAT, G. M. & TURNBULL, A. C. (1982). Do Apgar scores indicate asphyxia? Lancet 1, 494-496.

SYUTKINA, E. V. (1988). Effect of autonomic nervous system blockade and acute asphyxia

on heart rate variability in the fetal rat. Gynecol Obstet Invest 25, 249-257. TAN, S., ZHOU, F., NIELSEN, V. G., WANG, Z., GLADSON, C. L. & PARKS, D. A. (1998).

Sustained hypoxia-ischemia results in reactive nitrogen and oxygen species production and injury in the premature fetal rabbit brain. J Neuropathol Exp Neurol 57, 544-553.

TAN, S., ZHOU, F., NIELSEN, V. G., WANG, Z., GLADSON, C. L. & PARKS, D. A. (1999).

Increased injury following intermittent fetal hypoxia-reoxygenation is associated with increased free radical production in fetal rabbit brain. J Neuropathol Exp Neurol 58, 972-981.

TANAKA, S., TSUCHIDA, H., NAKABAYASHI, K., SEKI, S. & NAMIKI, A. (1996). The effects of

sevoflurane, isoflurane, halothane, and enflurane on hemodynamic responses during an inhaled induction of anesthesia via a mask in humans. Anesth Analg 82, 821-826.

TAUPIN, V., RENNO, T., BOURBONNIERE, L., PETERSON, A. C., RODRIGUEZ, M. & OWENS,

T. (1997). Increased severity of experimental autoimmune encephalomyelitis, chronic macrophage/microglial reactivity, and demyelination in transgenic mice producing tumor necrosis factor-alpha in the central nervous system. Eur J Immunol 27, 905-913.

TAYLOR, P. M., SILVER, M. & FOWDEN, A. L. (1992). Intravenous catheterisation of foetus

and mare in late pregnancy: management and respiratory, circulatory and metabolic effects. Equine Vet J 24, 391-396.

TEMPLE, C. M., DENNIS, J., CARNEY, R. & SHARICH, J. (1995). Neonatal seizures: long-term

outcome and cognitive development among 'normal' survivors. Dev Med Child Neurol 37, 109-118.

THACKER, S. B., STROUP, D. & CHANG, M. (2001). Continuous electronic heart rate

monitoring for fetal assessment during labor. Cochrane Database Syst Rev, CD000063. THORESEN, M., BAGENHOLM, R., LOBERG, E. M., APRICENA, F. & KJELLMER, I. (1996a).

Posthypoxic cooling of neonatal rats provides protection against brain injury. Arch Dis Child Fetal Neonatal Ed 74, F3-9.

THORESEN, M., HAALAND, K., LOBERG, E. M., WHITELAW, A., APRICENA, F., HANKO, E. &

STEEN, P. A. (1996b). A piglet survival model of posthypoxic encephalopathy. Pediatr Res 40, 738-748.

THORESEN, M., PENRICE, J., LOREK, A., CADY, E. B., WYLEZINSKA, M., KIRKBRIDE, V.,

COOPER, C. E., BROWN, G. C., EDWARDS, A. D., WYATT, J. S. & ET AL. (1995). Mild hypothermia after severe transient hypoxia-ischemia ameliorates delayed cerebral energy failure in the newborn piglet. Pediatr Res 37, 667-670.

THORNBERG, E., THIRINGER, K., ODEBACK, A. & MILSOM, I. (1995). Birth asphyxia:

incidence, clinical course and outcome in a Swedish population. Acta Paediatr 84, 927-932.

337

THORP, J. (1990). What is birth asphyxia? Reply. Am J Obstet Gynecol 163, 1368. THORP, J. A., SAMPSON, J. E., PARISI, V. M. & CREASY, R. K. (1989). Routine umbilical cord

blood gas determinations? Am J Obstet Gynecol 161, 600-605. TING, P., YAMAGUCHI, S., BACHER, J. D., KILLENS, R. H. & MYERS, R. E. (1983). Hypoxic-

ischemic cerebral necrosis in midgestational sheep fetuses: physiopathologic correlations. Exp Neurol 80, 227-245.

TIPTON, R. H. & LEWIS, B. V. (1975). Letter: Induction of labour and perinatal mortality. Br

Med J 1, 391. TOFFALETTI, J., HAMMES, M. E., GRAY, R., LINEBERRY, B. & ABRAMS, B. (1992). Lactate

measured in diluted and undiluted whole blood and plasma: comparison of methods and effect of hematocrit. Clin Chem 38, 2430-2434.

TRUDINGER, B. J., GILES, W. B., COOK, C. M., BOMBARDIERI, J. & COLLINS, L. (1985). Fetal

umbilical artery flow velocity waveforms and placental resistance: clinical significance. Br J Obstet Gynaecol 92, 23-30.

TRUDINGER, B. J., STEVENS, D., CONNELLY, A., HALES, J. R., ALEXANDER, G., BRADLEY, L.,

FAWCETT, A. & THOMPSON, R. S. (1987). Umbilical artery flow velocity waveforms and placental resistance: the effects of embolization of the umbilical circulation. Am J Obstet Gynecol 157, 1443-1448.

UEHARA, H., YOSHIOKA, H., KAWASE, S., NAGAI, H., OHMAE, T., HASEGAWA, K. &

SAWADA, T. (1999). A new model of white matter injury in neonatal rats with bilateral carotid artery occlusion. Brain Res 837, 213-220.

VAGNOZZI, R., MARMAROU, A., TAVAZZI, B., SIGNORETTI, S., DI PIERRO, D., DEL BOLGIA,

F., AMORINI, A. M., FAZZINA, G., SHERKAT, S. & LAZZARINO, G. (1999). Changes of cerebral energy metabolism and lipid peroxidation in rats leading to mitochondrial dysfunction after diffuse brain injury. J Neurotrauma 16, 903-913.

VAN CAPPELLEN, A. M., HEERSCHAP, A., NIJHUIS, J. G., OESEBURG, B. & JONGSMA, H. W.

(1999). Hypoxia, the subsequent systemic metabolic acidosis, and their relationship with cerebral metabolite concentrations: An in vivo study in fetal lambs with proton magnetic resonance spectroscopy. Am J Obstet Gynecol 181, 1537-1545.

VAN DEN BERG, P., SCHMIDT, S., GESCHE, J. & SALING, E. (1987). Fetal distress and the

condition of the newborn using cardiotocography and fetal blood analysis during labour. Br J Obstet Gynaecol 94, 72-75.

VAN DEN BERG, P. P., NELEN, W. L., JONGSMA, H. W., NIJLAND, R., KOLLEE, L. A., NIJHUIS,

J. G. & ESKES, T. K. (1996). Neonatal complications in newborns with an umbilical artery pH < 7.00. Am J Obstet Gynecol 175, 1152-1157.

VANNUCCI, R. C., CHRISTENSEN, M. A. & YAGER, J. Y. (1993). Nature, time-course, and

extent of cerebral edema in perinatal hypoxic-ischemic brain damage. Pediatr Neurol 9, 29-34.

338

VANUCCI, R. (1994). Heterogeneity of hypoxic-ischaemic thresholds in experiememtal animals. In Brain lesions in the newborn: hypoxic and haemodynamic pathogenesis. ed. LOU, H. C., GREISEN, G. & LARSEN, J. F., pp. 192-204. Munksgaard, Copenhagen.

VINTZILEOS, A. M., ANTSAKLIS, A., VARVARIGOS, I., PAPAS, C., SOFATZIS, I. &

MONTGOMERY, J. T. (1993). A randomized trial of intrapartum electronic fetal heart rate monitoring versus intermittent auscultation. Obstet Gynecol 81, 899-907.

VINTZILEOS, A. M., NOCHIMSON, D. J., ANTSAKLIS, A., VARVARIGOS, I., GUZMAN, E. R. &

KNUPPEL, R. A. (1995a). Comparison of intrapartum electronic fetal heart rate monitoring versus intermittent auscultation in detecting fetal acidemia at birth. Am J Obstet Gynecol 173, 1021-1024.

VINTZILEOS, A. M., NOCHIMSON, D. J., GUZMAN, E. R., KNUPPEL, R. A., LAKE, M. &

SCHIFRIN, B. S. (1995b). Intrapartum electronic fetal heart rate monitoring versus intermittent auscultation: a meta-analysis. Obstet Gynecol 85, 149-155.

VOLPE, J. (1995a). Hypoxic-ischemic encephalopathy. In: Neurology of the newborn. W.B. Saunders,

Philadelphia. VOLPE, J. (2001a). Neurology of the newborn. W.B. Saunders, Philadelphia. VOLPE, J. J. (1995b). Metabolic encephalopathies. In Neurology of the newborn, pp. 467-582.

Saunders, Philadelphia. VOLPE, J. J. (2001b). Perinatal brain injury: from pathogenesis to neuroprotection. Ment Retard

Dev Disabil Res Rev 7, 56-64. VON WINCKEL, F. (1893). Lehrbuch der geburtshilfe, Weisaben. 634. WADA, K., ALONSO, O. F., BUSTO, R., PANETTA, J., CLEMENS, J. A., GINSBERG, M. D. &

DIETRICH, W. D. (1999). Early treatment with a novel inhibitor of lipid peroxidation (LY341122) improves histopathological outcome after moderate fluid percussion brain injury in rats. Neurosurgery 45, 601-608.

WAGNER, K. & MYERS, R. (1979). Topographic aspects of lactic acid accumulation in brain

tissue during circulatory arrest. Neurology 29, 546. WANDRUP, J., TVEDE, K., GRINSTED, J. & JORDENING, H. (1989). "Stat" measurements of L-

lactate in whole blood and cerebrospinal fluid assessed. Clin Chem 35, 1740-1743. WANG, W., PANG, C. C., ROGERS, M. S. & CHANG, A. M. (1996). Lipid peroxidation in cord

blood at birth. Am J Obstet Gynecol 174, 62-65. WARD, R. J. & PETERS, T. J. (1995). Free radicals. In Clinical biochemistry: metabolic and clinical

aspects. ed. MARSHALL, W. J. & BANGERT, S. K., pp. 765-777. Churchill Livingstone, London.

WATSON, L., STANLEY, F. & BLAIR, E. (1999). Report of the Western Australian cerebral

palsy register to birth year 1994, pp. 77. TVW Telethon Institute for Child Health Research, Perth.

339

WEIL, M. H., LEAVY, J. A., RACKOW, E. C., HALFMAN, C. J. & BRUNO, S. J. (1986). Validation of a semi-automated technique for measuring lactate in whole blood. Clin Chem 32, 2175-2177.

WEINER, C. P., WILLIAMSON, R. A., WENSTROM, K. D., SIPES, S. L., GRANT, S. S. &

WIDNESS, J. A. (1991). Management of fetal hemolytic disease by cordocentesis. I. Prediction of fetal anemia. Am J Obstet Gynecol 165, 546-553.

WEINRAUB, Z., CASPI, E., BROOK, I., RAHMANI, P., BUKOVSKY, I. & SCHREYER, P. (1978).

Perinatal outcome in monitored and unmonitored high-risk deliveries. Isr J Med Sci 14, 249-255.

WESTGATE, J., GARIBALDI, J. M. & GREENE, K. R. (1994). Umbilical cord blood gas analysis

at delivery: a time for quality data. Br J Obstet Gynaecol 101, 1054-1063. WESTGATE, J. A., BENNET, L., DE HAAN, H. H. & GUNN, A. J. (2001). Fetal heart rate

overshoot during repeated umbilical cord occlusion in sheep. Obstet Gynecol 97, 454-459.

WESTGATE, J. A., BENNET, L. & GUNN, A. J. (1999). Fetal heart rate variability changes

during brief repeated umbilical cord occlusion in near term fetal sheep. Br J Obstet Gynaecol 106, 664-671.

WESTGREN, M., DIVON, M., HORAL, M., INGEMARSSON, I., KUBLICKAS, M., SHIMOJO, N. &

NORDSTROM, L. (1995). Routine measurements of umbilical artery lactate levels in the prediction of perinatal outcome. Am J Obstet Gynecol 173, 1416-1422.

WESTGREN, M., KRUGER, K., EK, S., GRUNEVALD, C., KUBLICKAS, M., NAKA, K., WOLFF,

K. & PERSSON, B. (1998). Lactate compared with pH analysis at fetal scalp blood sampling: a prospective randomised study. Br J Obstet Gynaecol 105, 29-33.

WESTGREN, M., KUBLICKAS, M. & KRUGER, K. (1999). Role of lactate measurements during

labor. Obstet Gynecol Surv 54, 43-48. WHO. (1991). World Health Organisation: Child health and development, health of the

newborn, pp. 180-186. World Health Organisation, Geneva. WHO. (1995). World Health Report. World Health Organisation, Geneva. WILLIAMS, C. E., GUNN, A. J., MALLARD, C. & GLUCKMAN, P. D. (1992). Outcome after

ischemia in the developing sheep brain: an electroencephalographic and histological study. 31, 14-21.

WILLIAMS, C. E., GUNN, A. J., SYNEK, B. & GLUCKMAN, P. D. (1990). Delayed seizures

occurring with hypoxic-ischemic encephalopathy in the fetal sheep. 27, 561-565. WILLIAMS, R. L., CREASY, R. K., CUNNINGHAM, G. C., HAWES, W. E., NORRIS, F. D. &

TASHIRO, M. (1982). Fetal growth and perinatal viability in California. Obstet Gynecol 59, 624-632.

WINDLE, W. F. (1956). Neurological and psychological deficits of asphyxia neonatorum: with consideration

of the use of primates for experimental investigations. By twenty-eight contributors. Charles C Thomas, Springfield, IL.

340

WINKLER, H., HOUTH, J. & TUCHER, J. (1991). Neonatal complications at term as related to

the degree of umbilical artery acidaemia. Am J Obstet Gynecol 164, 637. WOOD, C., RENOU, P., OATS, J., FARRELL, E., BEISCHER, N. & ANDERSON, I. (1981). A

controlled trial of fetal heart rate monitoring in a low-risk obstetric population. Am J Obstet Gynecol 141, 527-534.

WOOD, C. E. (2001). Endocrine functions. In Fetal growth and development. ed. HARDING, R. &

BOCKING, A. D., pp. 186-206. Cambridge University Press, Cambridge. WOODWARD, M. (1999). Epidemiology: Study Design and Data Analysis. Chapman & Hall/CRC,

Florida. YAFFE, H., PARER, J. T., BLOCK, B. S. & LLANOS, A. J. (1987). Cardiorespiratory responses to

graded reductions of uterine blood flow in the sheep fetus. J Dev Physiol 9, 325-336. YAGER, J., TOWFIGHI, J. & VANNUCCI, R. C. (1993). Influence of mild hypothermia on

hypoxic-ischemic brain damage in the immature rat. Pediatr Res 34, 525-529. YAGER, J. Y., BRUCKLACHER, R. M. & VANNUCCI, R. C. (1992). Cerebral energy metabolism

during hypoxia-ischemia and early recovery in immature rats. Am J Physiol 262, H672-677.

YARNELL, R., BIEHL, D. R., TWEED, W. A., GREGORY, G. A. & SITAR, D. (1983). The effect

of halothane anaesthesia on the asphyxiated foetal lamb in utero. Can Anaesth Soc J 30, 474-479.

YEH, S. Y., DIAZ, F. & PAUL, R. H. (1982). Ten-year experience of intrapartum fetal

monitoring in Los Angeles County/University of Southern California Medical Center. Am J Obstet Gynecol 143, 496-500.

YEOMANS, E. R., HAUTH, J. C., GILSTRAP, L. C., 3RD & STRICKLAND, D. M. (1985). Umbilical

cord pH, PCO2, and bicarbonate following uncomplicated term vaginal deliveries. Am J Obstet Gynecol 151, 798-800.

YOSHIOKA, T. & ROUX, J. F. (1970). Correlation of fetal scalp blood pH, glucose, lactate and

pyruvate concentrations with cord blood determinations at time of delivery and cesarean section. J Reprod Med 5, 209-214.

YU, T., KUI, L. Q. & MING, Q. Z. (2003). Effect of asphyxia on non-protein-bound iron and

lipid peroxidation in newborn infants. Dev Med Child Neurol 45, 24-27. YUDKIN, P. L., JOHNSON, A., CLOVER, L. M. & MURPHY, K. W. (1994). Clustering of

perinatal markers of birth asphyxia and outcome at age five years. Br J Obstet Gynaecol 101, 774-781.

YUDKIN, P. L., JOHNSON, A., CLOVER, L. M. & MURPHY, K. W. (1995). Assessing the

contribution of birth asphyxia to cerebral palsy in term singletons. Paediatr Perinat Epidemiol 9, 156-170.

ZERNICKOW, K. (1966). [The effect of air contact on microblood samples of fetuses].

Gynaecologia 161, 277-287.

341

ZHANG, R. L., CHOPP, M., LI, Y., ZALOGA, C., JIANG, N., JONES, M. L., MIYASAKA, M. &

WARD, P. A. (1994). Anti-ICAM-1 antibody reduces ischemic cell damage after transient middle cerebral artery occlusion in the rat. Neurology 44, 1747-1751.

ZWEIG, M. H. & CAMPBELL, G. (1993). Receiver-operating characteristic (ROC) plots: a

fundamental evaluation tool in clinical medicine. Clin Chem 39, 561-577.