UvA-DARE (Digital Academic Repository) · Sabine Ensing op 18 december 2015 om 10.00 uur...

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Risk factors and prognostic models for perinatal asphyxia at term

Ensing, S.

Publication date2015Document VersionFinal published version

Link to publication

Citation for published version (APA):Ensing, S. (2015). Risk factors and prognostic models for perinatal asphyxia at term.

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an opencontent license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, pleaselet the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the materialinaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letterto: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. Youwill be contacted as soon as possible.

Download date:07 Sep 2021

https://dare.uva.nl/personal/pure/en/publications/risk-factors-and-prognostic-models-for-perinatal-asphyxia-at-term(effb8480-01f4-4b4e-bd8c-a40c2f632292).html

500599-L-sub01-os-Ensing500599-L-sub01-os-Ensing500599-L-sub01-os-Ensing500599-L-sub01-os-Ensing Processed on: 11_6_2015Processed on: 11_6_2015Processed on: 11_6_2015Processed on: 11_6_2015

Sabine enSing

RiSk factoRS and pRognoStic modelS foR peRinatal

aSphyxia at teRm

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Uitnodiging

voor het bijwonen van deopenbare verdediging

van het proefschrift

Risk factors and prognostic models for perinatal

asphyxia at term

door Sabine ensing

op 18 december 2015om 10.00 uur

agnietenkapelOudezijds Voorburgwal 229 - 231,

1012 EZ Amsterdam

Receptie na afloop

Paranimfen:

Rianne de Vries06-55 38 01 89

Loes Thomaes06-16 13 18 19

[email protected]

Auteur:Sabine Ensing

Uiterwaardenstraat 157-1 1079 CK, Amsterdam.

[email protected]

Risk fac

toRs a

nd

pRog

no

stic m

od

els foR peRin

atal a

Sph

yx

ia at teRm

sabin

e ensin

g

apgaRSabine enSing

RiSk factoRS and pRognoStic modelS foR peRinatal

aSphyxia at teRm

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Uitnodiging

voor het bijwonen van deopenbare verdediging

van het proefschrift

Risk factors and prognostic models for perinatal

asphyxia at term

door Sabine ensing

op 18 december 2015om 10.00 uur

agnietenkapelOudezijds Voorburgwal 229 - 231,

1012 EZ Amsterdam

Receptie na afloop

Paranimfen:

Rianne de Vries06-55 38 01 89

Loes Thomaes06-16 13 18 19

[email protected]

Auteur:Sabine Ensing

Uiterwaardenstraat 157-1 1079 CK, Amsterdam.

[email protected]

Risk fac

toRs a

nd

pRog

no

stic m

od

els foR peRin

atal a

Sph

yx

ia at teRm

sabin

e ensin

g

apgaR

500599-sub01-os-Ensing.indd 1 06-11-15 08:56

500599-L-sub01-bw-Ensing500599-L-sub01-bw-Ensing500599-L-sub01-bw-Ensing500599-L-sub01-bw-Ensing

Risk factors and prognostic models for perinatal asphyxia at term.

S. Ensing


Colofon

Cover and design by: Alex Wesselink - www.persoonlijkproefschrift.nlPrinted by: Ipskamp Drukkers BVISBN: 978-94-6259-926-0

© Sabine Ensing, Amsterdam, the Netherlands 2015All rights reserved. No part of this thesis may be reproduced, stored or transmitted in any form or by any means, without prior permission of the referenced journal or the author.

Financial support for printing of this thesis was kindly supported by: Universiteit van Amsterdam, Chipsoft, BMA BV, Memidis Pharma BV, Academic Medical Center.


Risk factors and prognostic models for perinatal asphyxia at term.

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Universiteit van Amsterdam op gezag van de Rector Magnificus

prof. dr. D.C. van den Boom ten overstaan van een door het college voor promoties

ingestelde commissie, in het openbaar te verdedigen in de Agnietenkapel op vrijdag 18 december 2015, te 10.00 uur

door Sabine Ensinggeboren te Nijmegen


Promotiecommissie

Promotores: Prof. dr. B.W.J. Mol Universiteit van Amsterdam Prof. dr. A. Abu-Hanna Universiteit van Amsterdam

Co–promotores: Dr. A.C.J. Ravelli Universiteit van Amsterdam

Overige leden: Prof. dr. G.J. Bonsel Erasmus Universiteit Rotterdam Prof. dr. A.H.L.C. van Kaam Universiteit van Amsterdam Dr. A. Kwee Universiteit Utrecht Prof. dr. T.J. Roseboom Universiteit van Amsterdam Prof. dr. E.W. Steyerberg Erasmus Universiteit Rotterda Prof. dr. F. van der Veen Universiteit van Amsterdam

Faculteit der Geneeskunde


Table of ConTenTs

Chapter 1 General introduction

Part I: Trends of perinatal asphyxia and obstetricinterventions among termsChapter 2 Trends in birth asphyxia, obstetric interventions and perinatal

mortality among term singletons: a nationwide cohort study

Chapter 3 Are higher rates of obstetric interventions associated with a better neonatal outcome? A population wide cohort study

Part II: Risk factors for perinatal asphyxia relatedmortality and morbidityChapter 4 Recurrence risk of low Apgar score among term singletons:

a population-based cohort study

Chapter 5 Risk of poor neonatal outcome at term after medically assisted reproduction: a propensity score-matched study

Chapter 6 Maternal and neonatal risk factors for asphyxia related perinatal mortality at term

Part III: Prognostic modelling of adverse pregnancyoutcomes at termChapter 7 Predictive capacity of the five individual components of the 5

minute Apgar score on neonatal mortality and morbidity in term infants

Chapter 8 The development of a clinical prediction model to support clinicians in the simultaneous assessment of the risks of failure to progress and fetal asphyxia in term pregnancies

Chapter 9 Summary and general discussion

Appendices Samenvatting in het Nederlands

List of co-authors and their contribution to the manuscript

List of publications

PhD portfolio

Curriculum Vitae

9

21

39

61

77

97

115

135

161

172

178

182

184

188


Chapter 1General introduction

Chapter 1

10

General inTroduCTion

Perinatal asphyxiaPerinatal asphyxia is a rare yet serious complication during labor with immediate consequences and possible long-term neurological impairment. 1-3 It is suggested that normal labor, expressed per centimeter travelled, is one of the most dangerous journeys in life. In western countries the rates of perinatal asphyxia have decreased considerably in the late nineties. 4 However results from the National Vital Statistics Reports, which contain data on the U.S. births, and data from The European PERISTAT project, which monitors perinatal health in European countries, showed that nowadays still one to two percent of children born suffering perinatal asphyxia. 5, 6 This would suggest that in the Netherlands, with a prevalence of perinatal asphyxia of 1.3% and 175 000 births annually, each year 2200 infants suffer perinatal asphyxia. Among term infants the prevalence of perinatal asphyxia is around 0.88% (1300 infants each year, of which 250 infants suffering severe asphyxia).

Perinatal asphyxia is a clinical condition of impaired oxygen supply or blood flow to the fetus and can occur before the onset of labor (antepartum) or during labor (intrapartum). In order to create a worldwide accepted definition the American Academy of Paediatrics (AAP) and the American College of Obstetrics and Gynecology (ACOG) suggested the following criteria in diagnosing asphyxia: (1) umbilical cord pH less than 7, (2) persistence of an Apgar score of 0 – 3 for longer than 5 minutes, (3) neonatal neurological manifestations (e.g., seizures, coma, hypotonia), and (4) multiple organ involvement (e.g., kidney, lungs, liver, heart). 7

The Apgar score was proposed by Dr. Virginia Apgar in 1952 as an objective tool to assess the condition of newborns immediately after birth. The score consists of five characteristics of the newborn infant: skin color (pink, blue or pale), heart rate (≥100/minute, <100/minute or absent), respiratory effort (normal rate and effort, irregular gasping or absent), muscle tone (active, arms and legs flexed or ‘floppy tone’) and reflex irritability (pulls away, grimace or absent). All contribute equally to the total score each with rating of 0, 1 or 2 points. 8

Acronym Score of 0 Score of 1 Score of 2Appearance (skin color) Blue or pale Blue extremities / body pink Totally pinkPulse Absent < 100 bpm ≥ 100 bpmGrimace (reflex irritability) No response Grimace/ Feeble cry when stimulated Pulls awayActivity (muscle tone) None / ‘Floppy tone’ Some flexion ActiveRespiration Absent Weak /

Irregular gaspingStrong cry

General introduction

1

11

Because of the very small number of term newborns with a five-minute Apgar score < 3, most studies examine perinatal asphyxia as a five minute Apgar score < 7 and/or umbilical cord pH ≤ 7.0 “severe acidosis”. Indeed, in term infants without severe congenital malformations, a low five-minute Apgar score most likely reflects intrauterine asphyxia. 3

Perinatal asphyxia and perinatal mortality Perinatal mortality (stillbirth and neonatal mortality) at term is often unexpected and not entirely explicable, however it is indisputable that asphyxia plays a major role in this. The European PERISTAT project showed that the perinatal mortality rates in the Netherlands are among the highest in Europe. 6 85% of these perinatal deaths can be contributed to one of the so called BIG 4 determinants: preterm birth, intra-uterine growth restriction, congenital anomalies and perinatal asphyxia. 9 Although term neonates (born between 37.0-41.6 weeks) have the lowest risk for perinatal mortality, they account for 25% of the total number of perinatal mortality in the Netherlands and perinatal asphyxia plays a major role (>75%) in this. 10 Except the high risk of mortality, the surviving infants could experience long-term consequences on their physical and mental development. Most infants with sever asphyxia are admitted to an neonatal intensive care unit (NICU) and receive specialized care during the first years of their lives. So besides the individual consequences for the neonates suffering perinatal asphyxia, the related morbidity and mortality also exacts a high toll on their families and the communities in which they live and generates are considerable costs for the society. Therefor scientific research to predict and prevent perinatal asphyxia and asphyxia related health problems in general and specific for term infants are needed.

Perinatal asphyxia and obstetric interventionsThe obstetric caregiver constantly needs to weight the risk of an adverse pregnancy outcome when deciding between continuing labor versus obstetric interventions. Obstetric interventions consist of caesarean delivery or assisted vaginal delivery with a vacuum or forceps. The World Health Organization (WHO) stated “a population-level caesarean section rate above 15% is hardly justified from the medical perspective”. 11 However the caesarean delivery rates are continuing on the rise worldwide. Various reasons have been suggested for this increase, including advanced maternal age at first pregnancy, improved surgical techniques and obstetricians’ as well as patients’ perception on the safety of the procedure as compared to vaginal delivery. Higher rates of obstetric interventions can be justified if they improve neonatal outcome, without causing unnecessary maternal harm. As fetal distress (or suspected asphyxia) if often an indication for an obstetric intervention during labor, one may conjecture that the rising trend in obstetric interventions observed in developed countries would influence the incidence of perinatal asphyxia.


Chapter 1

12

Risk factors and prognostic modeling of perinatal asphyxia and the need of an intervention for non-progressive laborA lot of research has been done on the risk factors of perinatal asphyxia. These risk factors pertain to basis demographic characteristics, medical diseases, obstetric history and current pregnancy characteristics. 2, 12, 13 Examples of antepartum risk factors related to the pregnant woman are hypertension, advanced maternal age, (gestational) diabetes, medically assisted reproduction (MAR). Some of the neonatal characteristics are congenital anomalies and severe growth retardation. Despite the identification of maternal and fetal risk factors their prognostic interaction is not well understood. Prognostic models are used in various settings and for various reasons. The main reasons are to inform individuals about the future course of their illness (or their risk of developing illness) and to guide doctors and patients in joint decisions on further treatment. 14, 15 In the Netherlands, where women with low risk for pregnancy complications receive care in a separate midwifery led care system as compared to high risk women, who receive care in an obstetric led care system. Thus, optimal antepartum risk assessment is important to identify pregnant women with a high(er) risk of developing adverse neonatal outcome, so obstetric management can be individualized. 16 In literature, few prognostic tools have been presented for assigning the risk of perinatal asphyxia, but none of them were generalizable for the whole term population as they were based on small datasets or did not include some important antepartum and intrapartum variables. 17, 18 Furthermore, in labor there is continuous weighing of the risk of adverse outcome for either the neonate (asphyxia) or the mother (obstructed labor). An integrated model assessing both adverse pregnancy outcomes simultaneously could assist obstetric caregivers and guide subsequent interventions during labor and delivery.

This thesis will focus on the risk factors and prognostic models for adverse perinatal outcome at term, with a special focus on perinatal asphyxia and obstetric interventions during labor to reduce adverse pregnancy outcomes. For the majority of the studies in this thesis we were allowed to use data of the Netherlands Perinatal Registry (PRN). The PRN consists of national population-based data containing information on pregnancies, deliveries and (re)admissions until 28 days after birth. For one study the database of the Foundation Perinatal Audit in the Netherlands (PAN) was used



1

13

aim of This Thesis

- To identify risk factors for asphyxia related morbidity and mortality at term singleton pregnancies.

- To investigate the practice variance according to obstetric interventions and explore the association between variation in caesarean delivery rates with maternal and neonatal outcomes.

- To develop prognostic models to predict adverse neonatal and pregnancy outcomes at term.

ouTline of This Thesis

Part I: Trends of perinatal asphyxia and obstetric interventions among terms. In chapter 2 we describe trends in perinatal asphyxia, perinatal mortality and obstetric interventions for fetal distress among term singletons in the Netherlands.

In chapter 3 we describe the variation in planned and unplanned caesarean rates among general hospitals in the Netherlands. Furthermore we studied the possible association between caesarean delivery rates and maternal and neonatal outcomes.

Part II: Risk factors for perinatal asphyxia related mortality and morbidity.In chapter 4 we studied the influence of a previous pregnancy complicated by low Apgar score. We assessed the recurrence risk of perinatal asphyxia in a subsequent pregnancy among term singletons.

Since the introduction of medically assisted reproduction (MAR) over 35 years ago, the number of children born after MAR has rapidly increased. Chapter 5 studies the risk of a poor neonatal outcome at term after MAR.

In chapter 6 we explored maternal and neonatal risk factors for asphyxia related mortality at term in order to improve the identification of low-risk and high-risk pregnant women. Optimal antepartum risk assessment is crucial in reducing adverse pregnancy outcomes.


Chapter 1

14

Part III: Prognostic modelling of adverse pregnancy outcomes at term.The Apgar score consists of five components all contributing equally to the total score. In chapter 7 we studied the association of these five individual components with neonatal mortality and morbidity at term.

The risk of perinatal asphyxia as well as obstructed labor can be predicted. Chapter 8 describes the development of a two-dimensional decision tool to predict the risk of a poor neonatal outcome (asphyxia) and the need of an intervention for non-progressive labor simultaneously. Such a tool with combined predictions could facilitate in decision – making during labor.



1

15

referenCes

1. Ehrenstein V, Pedersen L, Grijota M, Nielsen GL, Rothman KJ, Sorensen HT. Association of Apgar score at five minutes with long-term neurologic disability and cognitive function in a prevalence study of Danish conscripts. BMC pregnancy and childbirth 2009;9:14.

2. Thornberg E, Thiringer K, Odeback A, Milsom I. Birth asphyxia: incidence, clinical course and outcome in a Swedish population. Acta paediatrica 1995;84:927-32.

3. Hogan L, Ingemarsson I, Thorngren-Jerneck K, Herbst A. How often is a low 5-min Apgar score in term newborns due to asphyxia? European journal of obstetrics, gynecology, and reproductive biology 2007;130:169-75.

4. Dzakpasu S, Joseph KS, Huang L, et al. Decreasing diagnoses of birth asphyxia in Canada: fact or artifact. Pediatrics 2009;123:e668-72.

5. Martin JA, Hamilton BE, Ventura SJ, Osterman MJ, Mathews TJ. Births: final data for 2011. National vital statistics reports : from the Centers for Disease Control and Prevention, National Center for Health Statistics, National Vital Statistics System 2013;62:1-69, 72.

6. Zeitlin J, Mohangoo AD, Delnord M, Cuttini M, Committee E-PS. The second European Perinatal Health Report: documenting changes over 6 years in the health of mothers and babies in Europe. Journal of epidemiology and community health 2013;67:983-5.

7. Committee on Obstetric Practice A, American Academy of P, Committee on F, Newborn A. ACOG Committee Opinion. Number 333, May 2006 (replaces No. 174, July 1996): The Apgar score. Obstetrics and gynecology 2006;107:1209-12.

8. Apgar V. A proposal for a new method of evaluation of the newborn infant. Current researches in anesthesia & analgesia 1953;32:260-7.

9. van der Kooy J, Poeran J, de Graaf JP, et al. Planned home compared with planned hospital births in the Netherlands: intrapartum and early neonatal death in low-risk pregnancies. Obstetrics and gynecology 2011;118:1037-46.

10. Ravelli AC, Eskes M, Tromp M, et al. [Perinatal mortality in The Netherlands 2000-2006; risk factors and risk selection]. Nederlands tijdschrift voor geneeskunde 2008;152:2728-33.

11. Organization WH. Appropriate technology for birth. Lancet 1985;2:436-7.

12. Milsom I, Ladfors L, Thiringer K, Niklasson A, Odeback A, Thornberg E. Influence of maternal, obstetric and fetal risk factors on the prevalence of birth asphyxia at term in a Swedish urban population. Acta obstetricia et gynecologica Scandinavica 2002;81:909-17.

13. Thorngren-Jerneck K, Herbst A. Low 5-minute Apgar score: a population-based register study of 1 million term births. Obstetrics and gynecology 2001;98:65-70.


Chapter 1

16

14. Moons KG, Royston P, Vergouwe Y, Grobbee DE, Altman DG. Prognosis and prognostic research: what, why, and how? Bmj 2009;338:b375.

15. Moons KG, Altman DG, Vergouwe Y, Royston P. Prognosis and prognostic research: application and impact of prognostic models in clinical practice. Bmj 2009;338:b606.

16. Amelink-Verburg MP, Verloove-Vanhorick SP, Hakkenberg RM, Veldhuijzen IM, Bennebroek Gravenhorst J, Buitendijk SE. Evaluation of 280,000 cases in Dutch midwifery practices: a descriptive study. BJOG : an international journal of obstetrics and gynaecology 2008;115:570-8.

17. Westerhuis ME, Schuit E, Kwee A, et al. Prediction of neonatal metabolic acidosis in women with a singleton term pregnancy in cephalic presentation. American journal of perinatology 2012;29:167-74.

18. Low JA, Pickersgill H, Killen H, Derrick EJ. The prediction and prevention of intrapartum fetal asphyxia in term pregnancies. American journal of obstetrics and gynecology 2001;184:724-30.



1

17


Part I:Trends of perinatal asphyxia and

obstetric interventions among terms


Chapter 2 Trends in birth asphyxia, obstetric interventions and perinatal mortality among term singletons:

a nationwide cohort study

S. EnsingA. Abu-Hanna

J.M. SchaafB.W.J. Mol

A.C.J. Ravelli

J Matern Fetal Neonatal Med. 2015 Apr;28(6):632-7

Chapter 2

22

absTraCT

ObjectiveThe objective of the present study is to investigate trends in birth asphyxia and perinatal mortality in the Netherlands over the last decade.

MethodsA nationwide cohort study among women with a term singleton pregnancy.We assessed trends in birth asphyxia in relation to obstetric interventions for fetal distress. Birth asphyxia was defined as a 5-minute Apgar score <7 (any asphyxia) or 5-minute Apgar score <4 (severe asphyxia). Perinatal mortality was defined as mortality during delivery or within 7 days after birth. Multivariable analyses were used to adjust for confounding factors.

ResultsThe prevalence of birth asphyxia was 0.85% and severe asphyxia 0.16%. Between 1999 and 2010 birth asphyxia decreased significantly with approximately 6% (p = 0.03) and severe asphyxia with 11% (p = 0.03). There was no significant change in perinatal mortality rate (0.98 per 1000 live births). Simultaneously the referral rate from primary to secondary care during labor increased from 20% to 24% (p<0.0001) and the intervention rate for fetal distress from 5.9% to 7.7% (p<0.0001).

ConclusionIn the Netherlands, the risk of birth asphyxia among term singletons has slightly decreased over the last decade; without a significant change in perinatal mortality.

Trends in birth asphyxia

2

23

inTroduCTion

Perinatal asphyxia is a severe complication within obstetric healthcare, which can lead to neonatal mortality and to other adverse severe neonatal outcomes with long-term consequences. Most studies examine birth asphyxia as a 5-minute Apgar score <7 and/or umbilical cord pH < 7.0. In term infants without severe congenital malformations, a low 5-minute Apgar score most likely reflects intrauterine asphyxia. 1 Indeed, the value of a low 5-minute Apgar score provides the best predictive value for neonatal mortality and is highly associated with subsequent neurologic disability or low cognitive function. 1-6

Several studies have been performed to describe the incidence of birth asphyxia in western countries and the rates of birth asphyxia seem to have decreased considerably in the late nineties. 7-11 More recent data of the National Vital Statistics Reports, which contain data on the U.S. births, showed a slight increase in the prevalence of birth asphyxia from 1.6% in 2007 to 1.9% in 2011, respectively, among all live births from 22 weeks of gestation and more. 12 In Europe one of the most recent data sources has been provided by the European PERISTAT project, which monitors perinatal health in European countries. In the most recent report based on 2010, one to two percent of children born alive have difficulties at birth that require resuscitation. 13 The PERISTAT reports are however insufficient for (adjusted) trend analysis. Glinaiaia et al. studied trends in asphyxia in a region in England but this was based on the years before 1994.14

Fetal distress is often an indication for an obstetric intervention during labor, so one may conjecture that a trend in obstetric interventions could influence the incidence of birth asphyxia. In many western countries, the Cesarean delivery rates have risen beyond the recommended maximum level of 15%. 13, 15, 16 Besides the rise in Cesarean deliveries, there is also discussion about the Dutch maternal care system. In the Netherlands, the intrapartum referral rate from primary midwife-led care to obstetrician-led care is still rising, without clear evidence that perinatal safety improved among low-risk pregnancies which started labor in primary care. 17 Although Evers et al. observed in a regional study a doubled risk of perinatal mortality in term births that started in primary midwife-led care compared with births that started in obstetrician-led care. 18 Another recent study in the Netherlands reported that birth asphyxia is the main cause of death in term infants who were admitted to Neonatal Intensive Care Units (NICU). 19 The aim of the present study is to describe trends in birth asphyxia, perinatal mortality and obstetric interventions for fetal distress among term singletons in the Netherlands. This study can add to the international body of knowledge on trends in obstetric care systems and the influence of possible changes on the perinatal outcomes among term singletons.

Chapter 2

24

meThods

Data for this study were extracted from a national cohort using the Netherlands Perinatal Registry (PRN). PRN consists of population-based data containing information on pregnancies, deliveries and (re)admissions until 28 days after birth. The PRN database is obtained by a validated linkage of three different registries: the midwifery registry (LVR1), the obstetrics registry (LVR2) and the neonatology registry (LNR) of hospital admissions of newborns. 20, 21 The coverage of the PRN registry is about 96% of all deliveries in the Netherlands and currently includes over 1.9 million records derived from deliveries in the last decade. 22 All data contained in PRN are voluntarily recorded by the caregiver during prenatal care, delivery and the neonatal period. For this study, all births between 1 January 1999 and 31 December 2009 were included. We selected all singletons born between 37 + 0 and 42 + 6 weeks of gestation. Women who had multiple pregnancy, elective Cesarean section, breech presentation or fetuses with any congenital anomalies were excluded. Our main outcomes for birth asphyxia were defined as a 5-minute Apgar score <7 (asphyxia) or as a 5-minute Apgar score < 4 (severe asphyxia). In addition, we investigated perinatal mortality which was defined as intrapartum mortality or mortality within 7 days after live birth. We also assessed obstetric interventions for fetal distress as well as interventions for non-progressive labor. We considered instrumental vaginal delivery (i.e. forceps or vacuum extraction) and emergency Cesarean sections. We analyzed the incidence and trends in all cases of a low Apgar score, as well as perinatal mortality. In addition we reviewed trends in obstetric interventions for fetal distress for in the same period. Because the Dutch maternity care model is based on risk selection and division between low-risk pregnancies in primary care and high-risk pregnancies in secondary care, additional analyses according to the level of care (primary and secondary care) at onset of labor and at delivery were performed. Finally, we assessed the perinatal outcome measures separately for the different subgroups of way of delivery.

Statistical analysisFirst the prevalence of total and severe asphyxia, perinatal mortality and obstetric interventions was calculated for the study population. To investigate whether there was a trend among the outcome measures we fitted a multivariable logistic regression model with year as the independent variable and each of low Apgar score risk, perinatal mortality risk or obstetric intervention risk as the dependent variable. We adjusted for factors that impact birth asphyxia and perinatal mortality. The odds ratios for al outcome measures are adjusted for maternal age, parity, ethnicity, socio-economic status, and SGA (birth weight below the 5th percentile) and LGA (birth weight above the 95th percentile). Ethnicity (Western versus Non-Western) was ascribed by the obstetric

2

25

care giver. We defined SGA and LGA according to the Dutch reference curves for birth weight by gestational age by parity, sex and ethnicity. 23 Statistical significance was set at the 0.05 level. In addition we performed a Cochran Armitage trend test for birth asphyxia with year as the independent variable. Furthermore, we described to what extent birth asphyxia and its subtype contribute to the overall incidence of perinatal mortality in the Netherlands. All statistical analyses were carried out with SAS 9.2 (SAS Institute Cary, NC). Permission for record use and analysis of data for the purpose of this study was obtained from PRN registry (registered as data petition 11.91).

resulTs

After exclusion of multiple pregnancies, fetuses with congenital anomalies, breech presentation, death prior to labor (stillbirth), preterm (< 37 weeks) and severe post term (> 42 + 6 weeks) and women who had elective Cesarean section we retained 1 603 340 term singleton pregnancies who constituting our study population, of which 13 625 infants had some form of birth asphyxia (0.85%) and 2605 had severe asphyxia (0.16%; Table 1).

Table 1. Trend in birth asphyxia & perinatal mortality among term singletons in the Netherlands 1999 –

2009.

Year No. births Apgar score < 7

Apgar score < 4

Perinatal mortality *

Obstetric interventionfetal distress

n % n % n ‰ n %

1999 146 375 1 321 0.90 259 0.18 125 0.85 8 615 5.9

2000 151 591 1 377 0.91 297 0.20 170 1.12 9 494 6.3

2001 149 862 1 253 0.84 228 0.15 146 0.97 10 003 6.7

2002 149 448 1 217 0.81 221 0.15 146 0.98 9 585 6.4

2003 151 680 1 285 0.85 262 0.17 160 1.05 9 910 6.5

2004 144 449 1 237 0.86 224 0.16 143 0.99 10 019 6.9

2005 141 514 1 209 0.85 237 0.17 155 1.10 9 681 6.8

2006 141 211 1 174 0.83 197 0.14 129 0.91 9 956 7.1

2007 139 756 1 123 0.80 221 0.16 135 0.97 10 049 7.2

2008 143 351 1 202 0.84 234 0.16 156 1.09 10 529 7.4

2009 144 103 1 227 0.85 225 0.16 107 0.74 11 102 7.7

Total 1 603 340 13 625 0.85 2 605 0.16 1 572 0.98 108 943 6.8

* intrapartum mortality or within 7 days after live birth

Chapter 2

26

The overall risk for birth asphyxia in the term singletons decreased slightly from 0.90% in 1999 to 0.85% in 2009 (Cochran-Armitage trend test p = 0.03). The decreasing risk of birth asphyxia in term singletons amounted to a decrease of 80 term neonates who suffer from asphyxia per year (adjusted odds ratio (aOR) 0.99, 95% CI 0.990–1.000). Risk of severe asphyxia also decreased from 0.18% to 0.16% (Cochran Armitage trend test p = 0.03) over the years (a decrease of 32 neonates per year, aOR 0.98, 95% CI 0.975–0.999).

The perinatal mortality rate among term singletons within the first week after birth did not change significantly over the years (aOR 0.99, 95% CI 0.98–1.01). Overall perinatal mortality within the first week after birth was 0.98 per 1000 births (n = 1572/1 603 340) as shown in Table 1. For infants with birth asphyxia the perinatal mortality rate was 100 per 1000 births and in case of severe birth asphyxia this increased to 454 per 1000 births. The contribution of birth asphyxia to the overall perinatal mortality did not change significantly in the last decade, approximately 87%. There was no significant change of mortality risk of 10% for neonates suffering of asphyxia during the study period. (See Table A in Appendix A).

The overall obstetric intervention rate (of 20%) in this term population was stable in the study period while the proportion of interventions for fetal distress significantly increased from 5.9% to 7.7% (p50.0001, aOR 1.03 95% CI 1.02–1.03; Table 1). In 304 125 deliveries (19%) an obstetric intervention was performed, in 6.8% (n = 108 943) because of suspected fetal distress. Interventions for suspected fetal distress considered 38 345 emergency Cesarean deliveries (2.4% of the overall population) and 70 598 instrumental vaginal deliveries (4.4%).

Table 2 shows level of care specific trends for the two main perinatal outcome measures. The proportion of women who started labor in secondary care increased from 35% in 1999 to 39% of all births in 2009 (p < 0.0001). In total 654 868 term women (41%) gave birth in primary care and 948 472 women (59%) gave birth in secondary care. A total of 361 198 women (23%) were referred to secondary care during labor, in 0.81% (n = 13 066) because of suspected urgent fetal problems. The referral rate of women increased significantly from 20% in 1999 to 24% in 2009 (p < 0.0001). This increase was mainly a result of a significantly increased referral rate for non-urgent reasons. Birth asphyxia risk decreased for both infants who were born in primary care setting (aOR 0.96 95% CI 0.94–0.97) and in secondary care setting (aOR 0.98 95% CI 0.97–0.99). Overall neonates born after a delivery which started in midwife-led primary care have a lower risk of an Apgar score <7 than women who started labor in obstetrician-led secondary care, 0.68% versus 1.1%, respectively, (aOR 0.60 95% CI 0.57–0.62). In the last decade perinatal mortality did not change significantly in

2

27

Table

2.

Tren

d in

per

inat

al o

utco

me

by le

vel o

f car

e du

ring

labo

r am

ong

term

sin

glet

ons.

Onse

t Pri

mary

care

Onse

t Se

condary

care

Del

iver

y P

rim

ary

care

Del

iver

y S

econdary

care

No b

irth

sBir

ths

AS

<

7*

Mort

alit

y

^Bir

ths

AS

<

7*

Mort

alit

y

^Bir

ths

AS

<

7*

Mort

alit

y

^Bir

ths

AS

<

7*

Mort

alit

y

^

%%

‰

%%

‰

%%

‰

%%

‰

19

99

146

375

650.

760.

7635

1.18

1.02

450.

430.

3255

1.29

1.30

20

00

151

591

640.

781.

1336

1.14

1.10

420.

460.

5058

1.23

1.57

20

01

149

862

640.

680.

8836

1.11

1.13

410.

340.

2959

1.18

1.46

20

02

149

448

640.

700.

8836

1.02

1.16

410.

380.

3259

1.12

1.45

20

03

151

680

640.

680.

8736

1.14

1.38

410.

340.

2259

1.20

1.61

20

04

144

449

630.

730.

8637

1.08

1.21

400.

390.

3659

1.17

1.42

20

05

141

514

640.

690.

8436

1.14

1.55

420.

350.

3159

1.21

1.64

20

06

141

211

630.

630.

6936

1.18

1.29

410.

290.

2860

1.20

1.34

20

07

139

756

630.

610.

8537

1.12

1.17

390.

270.

3361

1.14

1.38

20

08

143

351

620.

650.

8538

1.15

1.48

380.

350.

3362

1.15

1.57

20

09

144

103

610.

610.

4939

1.23

1.15

370.

280.

1163

1.18

1.11

Tota

l1

60

3 3

40

6

30

.68

0.8

3

37

1.1

41

.24

41

0.3

60.3

2

59

1.1

91.4

4

^ in

trapa

rtum

mor

talit

y or

with

in 7

day

s af

ter l

ive

birth

*AS

< 7:

5 m

inut

e A

pgar

sco

re <

7

Chapter 2

28

one of the subgroups of level of care at onset of labor or delivery. The risk of adverse neonatal outcomes according to the referral indication can be found in more detail in Appendix B.

Additionally, we performed separate analyses of birth asphyxia and perinatal mortality according to the way of delivery (Table 3). The increase in interventions for fetal distress was mainly a result of a significant increased risk for emergency Cesarean section for fetal distress, which increased from 1.9% in 1999 to 2.9% in 2009 (p < 0.0001). A 5-minute Apgar score <7 and perinatal mortality were both most seen among neonates delivered by obstetric interventions because of suspected fetal distress. The incidence rates presented in Table 3 showed a slightly decreasing trend for Apgar score <7, except for the neonates delivered for non-progressive labor. Perinatal mortality decreased significantly among neonates born after an obstetric intervention for suspected fetal distress from 4.4 in 1999 to 2.8 per 1000 births in 2009 (p < 0.0001; Table 3).

CommenT

Our study describes a significant slightly decrease in birth asphyxia in term singleton pregnancies in the Netherlands, however without a change in perinatal mortality. This decline in asphyxia was accompanied by a significant increase of obstetric interventions for fetal distress. The contribution of birth asphyxia (87%) to the overall perinatal mortality did not change in the last decade and there was no significant change of mortality risk for neonates suffering of asphyxia during the study period.

This study comprises data from a large, well-maintained, population-based national registry, and therefore provides a reliable overview in the situation of birth asphyxia in the Netherlands. Our sample size was large, as the PRN database consists of about 96% of all pregnancy and neonatal characteristics in the Netherlands. Furthermore, considering also trends in obstetric interventions for fetal distress and perinatal mortality contribute to the value of this study. A low 5-minute Apgar score alone as definition for birth asphyxia has been questioned in the literature; and the umbilical cord arterial pH less than 7.0 was suggested as being more objective as the Apgar score includes subjective components. 24, 25 However, information on the umbilical cord pH is not routinely measured and was therefore not available in the PRN database. As the strategy to estimate a new born infant for the Apgar score did not change over the study period, and the 5-minute Apgar score is accepted worldwide, we believe that our definition does not hinder the implementation of a trend analysis. In this subgroup of term infants, a 5-minute Apgar score < 7 is caused generally by intrauterine asphyxia, especially as we have a cohort of term singletons without congenital anomalies. 1

2

29

Table

3.

Tren

d in

per

inat

al o

utco

me

by w

ay o

f del

iver

y am

ong

term

sin

glet

ons.

Sponta

neo

us

Obst

etri

c in

terv

entions

Obst

etri

c in

terv

entions

Feta

l dis

tres

sN

on p

rogre

ssiv

e la

bor

No b

irth

sBir

ths

AS

< 7

* M

ort

alit

y ^

Bir

ths

AS

< 7

* M

ort

alit

y ^

Bir

ths

AS

< 7

* M

ort

alit

y^

%%

‰%

%‰

%%

‰

19

99

146

375

810.

550.

575.

94.

674.

4113

1.38

1.00

20

00

151

591

800.

610.

796.

34.

155.

5813

1.20

1.03

20

01

149

862

800.

550.

696.

73.

694.

7014

1.15

0.84

20

02

149

448

810.

520.

676.

43.

744.

8013

1.19

0.99

20

03

151

680

810.

520.

646.

54.

236.

2612

1.23

1.02

20

04

144

449

810.

550.

806.

94.

124.

0912

1.01

0.51

20

05

141

514

820.

540.

806.

84.

055.

1611

1.20

0.74

20

06

141

211

820.

510.

687.

13.

953.

7211

1.23

0.82

20

07

139

756

820.

460.

767.

24.

013.

9811

1.22

0.51

20

08

143

351

810.

510.

787.

33.

814.

8411

1.30

0.87

20

09

144

103

810.

450.

587.

74.

162.

7911

1.44

0.49

Tota

l1

60

3 3

40

81

0.5

00

.70

6.8

4.0

44

.612

1.2

30.8

1

^ in

trapa

rtum

mor

talit

y or

with

in 7

day

s af

ter l

ive

birth

*A

S <

7: 5

min

ute

Apg

ar s

core

< 7

Chapter 2

30

Relationship to other studies Many studies on birth asphyxia, as also the PERISTAT report and U.S. Births reports, included preterm births. We explicitly excluded this subgroup, as it makes a comparison not appropriate. 12, 13 There are several studies which have reported on low 5-minute Apgar score among terms. Birth asphyxia rates of 0.69% and 0.76%, respectively, are comparable with the overall risk of 0.85% in our study. 7, 8 However, the neonatal mortality rate among infants suffering from severe birth asphyxia in their population (86 per 1000) is strikingly lower than the neonatal mortality rate in our study. After performing a sensitivity analysis without the intrapartum deaths in our study cohort, the neonatal mortality rate was 197 deaths per 1000 within the 1st week after live birth.

Trends in birth asphyxia showed conflicting results. Wu et al. did report a consistent decline in birth asphyxia between 1991 and 2000 while a Canadian study reported that the decreasing trend of birth asphyxia between 1991 and 2005 was rather an artefact of changes in the use of the International Classification of Diseased (ICD) coding. 9, 11 In the Netherlands the ICD codes are not introduced in the obstetric charts, so our data are not influenced by differences in coding over the time. A recent study by Offerhaus et al. on a low-risk (midwifery) pregnancy population, i.e. women who started labor in midwife-led primary care, reported a stable risk of a 5-minute Apgar score < 7 between 2000 and 2008 for both nulliparous and parous women, despite a significant rise in referrals during labor from primary to secondary care. 17 Our study confirms the rise in referrals. New in our study is that among women who started labor in primary care and among women who actually delivered in primary care there was a significant decreasing trend in low Apgar score. This might implicate that the ongoing risk selection process during labor by midwifes generally succeed.

There are no reports on perinatal mortality rates among term infants to compare with our study. Two studies described a rapid decrease perinatal mortality between 1999 and 2009 in Turkey and between 2000 and 2006 in the Netherlands, respectively. As they both also included premature births these data are hard to compare with our findings among term births. 22, 26 Offerhaus et al. showed that according to our results, perinatal mortality rates among terms did not significantly change between 2000 and 2008, although this was in a part of our population: the low-risk midwifery population.17

In obstetric care, the caregiver needs constantly to weigh the risk of an adverse pregnancy outcome, i.e. neonatal and maternal morbidity or even mortality, when continuing labor versus obstetric interventions. In 1985 the World Health Organization stated that there was ‘‘no justification for any region to have Cesarean section rates higher than 10–15%’’. 27 Two decades later, however, the Cesarean section rates



2

31

increased to an average of 21.1% in developed countries. Our rate of approximately 10% of Cesarean deliveries among term singletons is low and in contrast with most other studies. 16, 28, 29 We even observed stable rates for obstetric interventions in total, although there was a significant increasing number of interventions for suspected fetal distress mainly due to an increase of Cesarean deliveries, a possible effect of improved fetal monitoring. Poignant et al. reported also an increase in emergency Cesarean sections among term singletons, but they found prolonged labor as major indication for this intervention. 30 Lack of an increasing trend in obstetric interventions overall may be attributable to differences in sociocultural and organizational factors. Dutch physicians have traditionally been reluctant to perform Cesarean sections, without strong medical indication. This relatively reserved position towards elective Cesarean section was underlined by the PERISTAT report. 13

One would expect that a rising trend in obstetric interventions for suspected fetal distress together with a decrease in birth asphyxia at term would contribute to a decrease in perinatal mortality for this cause. We, however, could not confirm this from our observations among term singletons; the perinatal mortality rate did not decrease. So our findings suggest that birth asphyxia remains one of the major causes for perinatal mortality in term births.

The main aim of this study was to describe trends in birth asphyxia among term infants, which is a clinically important outcome of pregnancy. The reported decline in birth asphyxia, in combination with an unchanged and relatively high risk of perinatal mortality should be further investigated in future research.


Chapter 2

32

referenCes



3. Ehrenstein V. Association of Apgar scores with death and neurologic disability. Clinical epidemiology 2009;1:45-53.

4. Skolnick AA. Apgar quartet plays perinatologist’s instruments. Jama 1996;276:1939-40.


6. Casey BM, Mcintire DD, Leveno KJ. The continuing value of the Apgar score for the assessment of newborn infants. The New England journal of medicine 2001;344:467-71.



9. Wu YW, Backstrand KH, Zhao S, Fullerton HJ, Johnston SC. Declining diagnosis of birth asphyxia in California: 1991-2000. Pediatrics 2004;114:1584-90.

10. Smith J, Wells L, Dodd K. The continuing fall in incidence of hypoxic-ischaemic encephalopathy in term infants. BJOG : an international journal of obstetrics and gynaecology 2000;107:461-6.

11. Dzakpasu S, Joseph KS, Huang L, et al. Decreasing diagnoses of birth asphyxia in Canada: fact or artifact. Pediatrics 2009;123:e668-72.



14. Glinianaia SV, Pharoah P, Sturgiss SN. Comparative trends in cause-specific fetal and neonatal mortality in twin and singleton births in the North of England, 1982-1994. BJOG : an international journal of obstetrics and gynaecology 2000;107:452-60.



2

33

15. Althabe F, Sosa C, Belizan JM, Gibbons L, Jacquerioz F, Bergel E. Cesarean section rates and maternal and neonatal mortality in low-, medium-, and high-income countries: an ecological study. Birth 2006;33:270-7.

16. Betran AP, Merialdi M, Lauer JA, et al. Rates of caesarean section: analysis of global, regional and national estimates. Paediatric and perinatal epidemiology 2007;21:98-113.

17. Offerhaus PM, Hukkelhoven CW, De Jonge A, Van Der Pal-De Bruin KM, Scheepers PL, Lagro-Janssen AL. Persisting rise in referrals during labor in primary midwife-led care in the Netherlands. Birth 2013;40:192-201.

18. Evers AC, Brouwers HA, Hukkelhoven CW, et al. Perinatal mortality and severe morbidity in low and high risk term pregnancies in the Netherlands: prospective cohort study. Bmj 2010;341:c5639.

19. Evers AC, Brouwers HA, Nikkels PG, et al. Substandard care in delivery-related asphyxia among term infants: prospective cohort study. Acta obstetricia et gynecologica Scandinavica 2013;92:85-93.

20. Tromp M, Ravelli AC, Meray N, Reitsma JB, Bonsel GJ. An efficient validation method of probabilistic record linkage including readmissions and twins. Methods of information in medicine 2008;47:356-63.

21. Meray N, Reitsma JB, Ravelli AC, Bonsel GJ. Probabilistic record linkage is a valid and transparent tool to combine databases without a patient identification number. Journal of clinical epidemiology 2007;60:883-91.

22. Ravelli AC, Tromp M, Van Huis M, et al. Decreasing perinatal mortality in The Netherlands, 2000-2006: a record linkage study. Journal of epidemiology and community health 2009;63:761-5.

23. Visser GH, Eilers PH, Elferink-Stinkens PM, Merkus HM, Wit JM. New Dutch reference curves for birthweight by gestational age. Early human development 2009;85:737-44.

24. Marrin M, Paes BA. Birth asphyxia: does the Apgar score have diagnostic value? Obstetrics and gynecology 1988;72:120-3.

25. Committee on Obstetric Practice A, American Academy Of P, Committee On F, Newborn A. ACOG Committee Opinion. Number 333, May 2006 (replaces No. 174, July 1996): The Apgar score. Obstetrics and gynecology 2006;107:1209-12.

26. Demirel G, Tezel B, Ozbas S, et al. Rapid decrease of neonatal mortality in Turkey. Maternal and child health journal 2013;17:1215-21.

27. Organization WH. Appropriate technology for birth. Lancet 1985;2:436-7.

28. Klemetti R, Che X, Gao Y, et al. Cesarean section delivery among primiparous women in rural China: an emerging epidemic. American journal of obstetrics and gynecology 2010;202:65 e1-6.


Chapter 2

34

29. Loudon JA, Groom KM, Hinkson L, Harrington D, Paterson-Brown S. Changing trends in operative delivery performed at full dilatation over a 10-year period. Journal of obstetrics and gynaecology : the journal of the Institute of Obstetrics and Gynaecology 2010;30:370-5.

30. Poignant M, Hjelmstedt A, Ekeus C. Indications for operative delivery between 1999-2010 and induction of labor and epidural analgesia on the risk of operative delivery--a population based Swedish register study. Sexual & reproductive healthcare : official journal of the Swedish Association of Midwives 2012;3:129-34.

2

35

Appendix A. Mortality risk among neonates suffering from birth asphyxia and the contribution of birth

asphyxia to the overall neonatal mortality in the Netherlands between 1999 – 2009.

Year No. Births Perinatal mortality

Apgar score <7

Contribution Apgar score < 7 to mortality ^

Mortality ^ riskin case of Apgar score < 7

n n % %

1999 146 375 125 1 321 93 8.8

2000 151 591 170 1 377 88 11

2001 149 862 146 1 253 86 10

2002 149 448 146 1 217 85 10

2003 151 680 160 1 285 84 11

2004 144 449 143 1 237 90 10

2005 141 514 155 1 209 88 11

2006 141 211 129 1 174 85 9.3

2007 139 756 135 1 123 90 11

2008 143 351 156 1 202 87 11

2009 144 103 107 1 227 82 7.2

Total 1 603 340 1 572 13 625 87 10

^ intrapartum mortality or within 7 days after live birth

Chapter 2

36

Appendix B. Trend in perinatal outcome by indication of referral during labor among term singletons.

Delivery in primary care

Home Hospital

No births Births AS < 7 Mortality^ Births AS < 7 Mortality^

% % ‰ % % ‰

1999 146 375 31 0.36 0.30 14 0.58 0.35

2000 151 591 30 0.39 0.37 12 0.68 0.86

2001 149 862 29 0.31 0.18 12 0.41 0.54

2002 149 448 29 0.35 0.30 13 0.46 0.36

2003 151 680 29 0.31 0.27 12 0.45 0.26

2004 144 449 28 0.32 0.32 12 0.57 0.46

2005 141 514 29 0.31 0.25 13 0.46 0.51

2006 141 211 28 0.25 0.31 13 0.37 0.23

2007 139 756 26 0.23 0.38 13 0.37 0.22

2008 143 351 25 0.31 0.33 13 0.43 0.31

2009 144 103 23 0.25 0.18 14 0.36 0.00

Total 1 603 340 28 0.31 0.29 13 0.47 0.37

^ intrapartum mortality or within 7 days after live birth

AS < 7: 5 minute Apgar score < 7

2

37

Delivery in secondary care

Urgent referral Non urgent referral Start of labor in

to secondary care to secondary care secondary care

Births AS < 7 Mortality^ Births AS < 7 Mortality^ Births AS < 7 Mortality^

% % ‰ % % ‰ % % ‰

1.04 5.1 15 19 1.3 1.0 35 1.2 1.0

1.02 4.6 17 21 1.2 1.6 36 1.1 1.1

0.96 4.5 17 21 1.2 1.3 36 1.1 1.1

0.88 3.6 14 21 1.2 1.4 36 1.0 1.2

0.70 4.7 14 22 1.2 1.6 36 1.1 1.4

0.76 4.5 10 22 1.2 1.5 37 1.1 1.2

0.72 4.3 18 22 1.2 1.3 36 1.1 1.5

0.74 3.8 11 22 1.1 1.1 37 1.2 1.3

0.69 3.3 9 23 1.1 1.5 37 1.1 1.2

0.72 3.4 15 22 1.1 1.3 38 1.1 1.5

0.71 3.5 9 24 1.0 0.8 39 1.2 1.1

0.81 4.2 14 22 1.2 1.3 36 1.1 1.2


Chapter 3 Are higher rates of obstetric interventions

associated with a better neonatal outcome? A population wide cohort study

S. EnsingJ.J. Kaandorp

J.B. DerksA. Abu-Hanna

B.W.J. MolA.C.J. Ravelli

Submitted, under embargo

Chapter 3

40

absTraCT

Objectives To assess the variation in both elective and emergency caesarean deliveries and to assess its association with neonatal and maternal outcome.

Design Population cohort study. We used multiple logistic regressions to calculate the predicted caesarean delivery rates for each hospital. Subsequently we categorized hospitals in quartiles according to their risk adjusted elective and emergency caesarean delivery rates.

Setting All 87 non-tertiary hospitals in the Netherlands.

Participants 458, 712 primiparous women with term, singleton pregnancies between 1999 and 2011.

Main Outcome Measures Primary outcome was defined as 5 minute Apgar score < 4, including intrapartum mortality. Secondary outcomes included adverse neonatal and maternal outcome. Results The adjusted mean elective caesarean rate was 1.8% (range 0.72% to 5.2%) while the emergency caesarean delivery rate was 15% (range 8.2% to 28%). Infants born at hospitals performing more than average elective caesareans have a decreased risk of Apgar score < 4 (aOR 0.79, 95%CI 0.86 – 0.96). Hospitals with low rates of emergency caesarean sections had an increased risk for a 5 minute Apgar score < 4 (aOR 1.22, 95%CI 1.06 – 1.41) and adverse neonatal outcomes (aOR1.08, 95CI 1.03-1.15). High rates of emergency caesarean deliveries was associated with a decreased risk of adverse neonatal and adverse maternal outcome (aOR 0.91, 95%CI 0.86 – 0.96 and aOR 0.87, 95%CI 0.84 – 0.91).

Conclusion Caesarean delivery rates vary substantially among Dutch hospitals. Hospitals with low rates of emergency caesarean deliveries have more adverse neonatal outcomes, however a caesarean delivery does not guarantee an optimal neonatal outcome. This emphasizes the need for more general management in obstetric healthcare.


Obstetric interventions and neonatal outcomes

3

41

inTroduCTion

Caesarean delivery rates are continuing on the rise worldwide. Various reasons have been suggested for this increase, including advanced maternal age at first pregnancy, improved surgical techniques, and obstetricians’ as well as patients’ perception on the safety of the procedure as compared to vaginal delivery. 1-3 Despite the widespread presence of guidelines and structured care pathways - multidisciplinary tools that improve the efficiency of evidence-based care while improving the quality of care- 4 obstetric medical practice varies considerably in and between countries. 2, 5-7

Although most research reports remarkable variations in elective surgery rates, there is evidence indicating that also variety in numbers is seen in emergency care. In 2010, The Dutch Healthcare Performance Report notes that the percentage of unplanned caesarean sections carried out in low risk pregnancies varies from 7.3% to approximately 30% across hospitals. 8 One would hypothesize that high intervention rates during labour would result in optimal fetal outcome. Higher rates of obstetric interventions can be justified if they improve neonatal outcome, without causing unnecessary maternal harm. Literature is not unequivocal on this, although most research demonstrates that infants born to women in hospitals with both very high or very low caesarean rates experience greater risk of adverse neonatal outcomes. 1, 5, 9, 10 The former studies did not distinguish between elective and emergency caesarean deliveries and included premature and breech deliveries. Differences between hospitals in incidences of these characteristics influence both the caesarean delivery rates and neonatal outcomes. Compared with other countries, caesarean rates of 15% in the Netherlands are among the lowest worldwide, and elective caesarean deliveries on maternal request are rarely performed. 11 In this paper, we aim to describe the variation in elective and emergency caesarean rates separately and to assess the possible association between variation in Caesarean section rates with maternal and neonatal outcomes. We studied a large cohort of primiparous women with a singleton fetus in cephalic position at term who delivered in in non-tertiary (general) hospitals.

meThods

We used data from the Dutch Perinatal Registry (PRN). The PRN registry is obtained by a validated linkage of three different registries: the midwifery registry (LVR1), the obstetrics registry (LVR2) and the neonatology registry (LNR) of hospital admissions of newborns. 12, 13 The PRN registry contains anonymous data and covers about 96% of all deliveries in The Netherlands.

Chapter 3

42

We studied all deliveries in secondary referral hospitals of primiparous women with term, singleton, and cephalic pregnancies between 1999 and 2011 in the Netherlands. Tertiary perinatal centres were excluded. Fetuses with congenital abnormalities and cases with antepartum fetal death were excluded. For each non tertiary hospital we investigated the association between each of elective (i.e. before the onset of labour) and emergency caesarean delivery (after intended vaginal delivery) rates and adverse neonatal or maternal outcomes. Adverse neonatal outcomes assessed were a 5 minute Apgar score < 4 (primary outcome) and secondary outcomes: adverse neonatal outcome is a composite outcome of a 5 minute Apgar score < 4 and/or admission to a tertiary perinatal care centre after birth and/or neonatal mortality and/or hypoxic ischemic encephalopathy (HIE) and/or neonatal convulsions and/or respiratory problems. Maternal outcomes considered post-partum haemorrhage ≥ 1000ml (PPH) with or without blood transfusion and manual placental removal in case of vaginal delivery.

StatisticsAs an elective caesarean delivery precludes an intended vaginal delivery (competing risk) we have formed two groups for the analyses. Group 1 comprised all term primiparous women (n = 458 712) and group 2 comprised all women with intended vaginal delivery (n = 450 319), because emergency caesarean delivery can only occur in women who undertake trial of labour. First, using the appropriate base population as described above, actual elective and emergency caesarean delivery rates were calculated for each hospital. Second, multiple logistic regression was used to calculate the predicted probability of a woman having an elective caesarean delivery and of an emergency caesarean delivery. Based on her age, ethnicity, socio-economic status and clinical risk factors for caesarean section. Clinical risk factors for elective caesarean section were identified by literature search and clinical reasoning. We also tested for interaction between (a) the predicted score of caesarean delivery and (b) the actual mode of delivery on (c) adverse neonatal outcome, while adjusting for (a) and (b). The probabilities of caesarean delivery for women who give birth in the same hospital were summed to give the hospitals’ predicted mean probability of elective caesarean deliveries. Risk adjusted rates of elective caesarean deliveries for each hospital was then calculated by dividing the hospitals’ actual rate by its predicted rate and multiplying this ratio by the national elective caesarean delivery rate. 2 This procedure was repeated for the emergency caesarean delivery rates. A multinomial logistic regression model was used to estimate the probability of unplanned emergency caesarean delivery as there were three outcomes for each woman: emergency caesarean delivery, assisted vaginal delivery or spontaneous vaginal delivery. 14 Subsequently we categorized hospitals in quartiles according to their risk adjusted elective and emergency caesarean delivery rates. Hospitals with caesarean delivery

3

43

rates in the two mid quartiles were used as reference category (25-75%). We then used multilevel logistic regression analysis, with a random intercept for each hospital, to predict risk for adverse neonatal and maternal outcome for each hospital. Using multilevel logistic regression takes into account similarities of births within hospitals. 7, 15 We adjusted for maternal age (years), ethnicity (western versus non-western), socio-economic status (high/middle versus low), hypertensive disease (pre-existent and pregnancy induces hypertension, pre-eclampsia), diabetes (pre-existent or gestational), female fetal gender, small for gestational age ( <5th percentile), large for gestational age (> 95th percentile), gestational age (weeks) and chronical maternal diseases ( endocrinological and coagulation disorders, congenital uterus anomalies and uterus myomatosis).

Data were analysed using the R statistical software environment version 2.15.1 (R Foundation for Statistical Computing, Vienna, Austria). All statistical tests had a P – value of 0.05 as the threshold to indicate statistical significance. Permission for record use and analysis of data for the purpose of this study was obtained from PRN (registered as data petition 12.57).

resulTs

We studied 458 712 primiparous women with a singleton term pregnancy who delivered in a non-tertiary (general) hospital. Baseline characteristics of the study population are shown in Table 1. Our logistic regression model to estimate the caesarean rates for each hospital by maternal and clinical factors showed that maternal age, gestational age, ethnicity, socio-economic status, hypertension, diabetes, other pre-existing maternal diseases and birth weight were all associated with caesarean delivery (See Appendix). After adjusting for these variables the mean elective caesarean rate was 1.8% (range 0.72% – 5.2%). The mean emergency caesarean rate varied widely among the Dutch secondary hospitals with a mean of 15% (range 8.2% - 28%). Table 2 shows that 56% (49/87) of hospitals had elective caesarean rates far below or above the range as predicted from their patient characteristics. For the emergency caesarean delivery rates 55% (48/87) of the hospitals had rates below or above the predicted rate. The predicted rates for both the elective and emergency caesarean sections were much more comparable for the three groups than the actual adjusted rates.


Chapter 3

44

Table 1. Baseline characteristics of all 458 712 primiparous women who delivered a term, cephalic

singleton in a general hospital.

Elective caesarean deliveryN = 8393

Vaginal deliveryN = 383 023

Emergency caesarean delivery N = 67 296

N (%) N (%) N (%)

Maternal age (years) * 29.8 (± 5.2) 31.0 (± 5.2) 29.8 (± 4.7)

Gestational age (days) * 272 (± 8.4) 279 (± 8.3) 280 (± 9.0)

Non - Western ethnicity 1276 (15%) 52 832 (14%) 9980 (15%)

Low socioeconomic status 2063 (25%) 99 167 (26%) 17 143 (25%)

Hypertensive diseases 1740 (21%) 59 282 (15%) 12 897 (20%)

Diabetes 201 (2.4%) 4079 (1.1%) 1393 (2.1%)

Artificial reproduction technologies 1528 (18%) 53 766 (14%) 10 389 (15%)

Small gestational age P5 851 (10%) 19 612 (5.1%) 4423 (6.6%)

Large gestational age P95 917 (11%) 20 917 (5.5%) 7996 (12%)

Induction of labour NA 69 465 (18%) 18 756 (28%)

Epidural anaesthesia at intended vaginal delivery

NA 59 773 (16%) 19 144 (28%)

Chronical maternal disease 562 (6.7%) 11 562 (3.0%) 2426 (3.6%)

*Mean ± SD



3

45

Table

2.

Pred

icte

d an

d ac

tual

ele

ctiv

e an

d em

erge

ncy

caes

area

n de

liver

y ra

tes,

by

caes

area

n de

liver

y ra

ting

grou

p.

Hosp

itals

(N

)D

eliv

erie

s (N

)Range

of

act

ual c

aes

are

an

del

iver

y r

ate

s (%

)Range

of

pre

dic

ted c

aes

are

an

del

iver

y r

ate

s (%

)*1

Elec

tive

caes

area

n de

liver

yLo

w24

114

134

1.2%

(0.7

2% to

1.4

%)

1.8%

(1.5

% to

2.0

%)

Mid

dle

rang

e38

228

059

1.8%

(1.4

% to

2.1

%)

1.8%

(1.6

% to

2.3

%)

Hig

h25

116

519

2.6%

(1.8

% to

5.2

%)

1.8%

(1.5

% to

2.2

%)

Tota

l 1.8

% (r

ange

0.7

2% to

5.2

%)

Emer

gen

cy c

aes

are

an d

eliv

ery

Onl

y fo

r wom

en w

ith in

tend

ed v

agin

al d

eliv

ery

Hosp

itals

(N

)D

eliv

erie

s (N

)Range

of

act

ual c

aes

are

an

del

iver

y r

ate

s (%

)Range

of

pre

dic

ted c

aes

are

an

del

iver

y r

ate

s (%

)*2

Low

1810

1 32

511

% (9

% to

13%

)15

% (1

4% to

16%

)M

iddl

e ra

nge

3923

4 08

214

% (1

3% –

17%

)15

% (1

4% to

17%

)H

igh

3011

4 91

219

% (1

6% to

30%

)15

% (1

3% to

16%

)To

tal 1

5% (r

ange

8.2

% to

28%

)

low

: hos

pita

ls w

ith th

e lo

wes

t 25%

of c

aesa

rean

del

iver

ies

high

: hos

pita

ls w

ith th

e hi

ghes

t 25%

of c

aesa

rean

del

iver

ies

Mod

el *

1: m

ater

nal a

ge (y

ears

), ge

statio

nal a

ges

(wee

ks),

ethn

icity

, so

cio-

econ

omic

sta

tus,

hyp

erte

nsio

n, d

iabe

tes,

ute

rus

myo

mat

osis,

con

geni

tal u

teru

s an

omal

ies,

co

agul

atio

n di

sord

ers,

arti

ficia

l rep

rodu

ctio

n te

chno

logi

es a

nd b

irth

wei

ght.

Mod

el *

2:

mat

erna

l age

(yea

rs),

gesta

tiona

l age

s (w

eeks

), et

hnic

ity,

soci

o-ec

onom

ic s

tatu

s, h

yper

tens

ion,

dia

bete

s, u

teru

s m

yom

atos

is, c

onge

nita

l ute

rus

anom

alie

s,

coag

ulat

ion

diso

rder

s, a

rtific

ial r

epro

duct

ion

tech

nolo

gies

, birt

h w

eigh

t and

≥ 2

4h ru

ptur

ed m

embr

anes

.

Chapter 3

46

Neonatal and maternal outcomes for elective caesarean delivery ratesOf the 458 712 infants born, 8393 (1.8%) were born by elective caesarean delivery. Adverse neonatal outcomes were decreased for hospitals with elective caesarean delivery rates in the highest quartile (aOR 0.79, 95%CI 0.68 to 0.91 and aOR 0.91, 95%CI 0.86 to 0.96 respectively). For infants born at hospitals performing few elective caesarean sections no difference in risk for low Apgar score was detected compared to infants born at hospitals within the mid part (aOR 0.94, 95%CI 0.82 to 1.08) The decreased risk of adverse neonatal outcome among infants born in hospitals with more than average elective caesarean deliveries was seen both after elective caesarean and intended vaginal delivery. (Table 3 and supplemental table 1). Women who deliver in hospitals with higher or lower elective caesarean rates has no better or worse outcome than women who deliver in hospitals with average caesarean delivery rates. Although after caesarean delivery in hospitals with fewer elective caesarean sections women are at increased risk of adverse maternal outcomes (aOR 1.53, 95%CI 1.19 to 1.96).

Neonatal and maternal outcomes for emergency caesarean section Hospitals performing few emergency caesarean sections have worse neonatal outcomes compared to hospitals in the midrange. For infants born after emergency caesarean delivery in hospitals with low emergency caesarean delivery rates risk for an Apgar score <4 increased to 8.2 per 1000 (aOR 1.63, 95%CI 1.31 to 2.12) and the risk for any adverse neonatal outcome increased to 45 per 1000 (aOR 1.30, 95%CI 1.17 to 1.44). (Table 4). The neonatal mortality risk was 4.1 per 1000 (aOR 1.69, 95%CI 1.18 to 2.41) among neonates born after emergency caesarean delivery in hospital with low emergency caesarean delivery rates. (Supplemental table). Performing more than average emergency caesarean sections does not significantly decrease the risk of an adverse neonatal outcome for the remaining infants delivered vaginally (aOR 0.99, 95%CI 0.93 – 1.06). (Supplemental table 2) In hospitals performing relatively many emergency caesareans the risk for maternal morbidity decreases for both women who delivered by caesarean section as women who delivered vaginally (aOR 0.83, 95%CI 0.76 to 0.91 and aOR 0.91, 95%CI 0.88 – 0.95).

3

47

Table

3. R

isk o

f adv

erse

neo

nata

l and

mat

erna

l out

com

e co

mpa

red

betw

een

hosp

itals

with

hig

h or

low

adj

uste

d el

ectiv

e ca

esar

ean

deliv

ery

rate

s ve

rsus

hos

pita

ls w

ith

elec

tive

caes

area

n de

liver

y ra

tes

with

in th

e m

iddl

e ra

nge,

stra

tified

for m

ode

of d

eliv

ery.

Tota

l popula

tion,

N =

45

8 7

12

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Adve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

14 1

340.

260.

94 (0

.82

– 1.

08)

2.0

1.12

(1.0

6 –

1.18

)3.

91.

00 (0

.97

– 1.

04)

Mid

par

tN

= 22

8 05

90.

27Re

fere

nce

1.8

Refe

renc

e3.

9Re

fere

nce

Hig

hest

quar

ter

N =

116

519

0.22

0.79

(0.6

8 –

0.91

)1.

60.

91 (0

.86

– 0.

96)

3.8

0.98

(0.9

5 –

1.02

)

Elec

tive

Caes

are

an d

eliv

ery,

N =

839

3

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Adve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

313

1.4

1.40

(0.8

0-2.

46)

4.5

1.03

(0.7

6 –

1.39

)7.

31.

53 (1

.19

– 1.

96)

Mid

par

tN

= 4

011

0.97

Re

fere

nce

4.4

Refe

renc

e4.

9Re

fere

nce

Hig

hest

quar

ter

N =

306

90.

690.

71 (0

.41-

1.20

)3.

4 0.

76 (0

.60

– 0.

96)

5.0

1.01

(0.8

2 –

1.26

)

adve

rse

neon

atal

out

com

e is

a co

mpo

site

outc

ome

of a

5 m

inut

e A

pgar

sco

re <

4 a

nd/

or a

dmiss

ion

to a

terti

ary

perin

atal

car

e ce

ntre

afte

r birt

h an

d/or

neo

nata

l m

orta

lity

and/

or h

ypox

ic is

chem

ic e

ncep

halo

path

y (H

IE) a

nd/

or n

eona

tal c

onvu

lsion

s an

d/or

resp

irato

ry p

robl

ems.

Chapter 3

48

Table

4. R

isk o

f adv

erse

neo

nata

l and

mat

erna

l out

com

e fo

r wom

en w

ith in

tend

ed v

agin

al d

eliv

ery

com

pare

d be

twee

n ho

spita

ls w

ith h

igh

or lo

w a

djus

ted

emer

genc

y ca

esar

ean

deliv

ery

rate

s ve

rsus

hos

pita

ls w

ith e

mer

genc

y ca

esar

ean

deliv

ery

rate

s w

ithin

the

mid

dle

rang

e, s

tratifi

ed fo

r mod

e of

del

iver

y.

Inte

nded

Vagin

al D

eliv

ery,

N =

45

0 3

19

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Adve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

01 3

250.

281.

22 (1

.06

– 1.

41)

1.9

1.08

(1.0

3 –

1.15

)3.

90.

98 (0

.94

– 1.

02)

Mid

par

tN

= 2

34 0

820.

23Re

fere

nce

1.7

Refe

renc

e3.

9Re

fere

nce

Hig

hest

quar

ter

N =

114

912

0.22

0.95

(0.8

2 –

1.10

)1.

60.

92 (0

.87

– 0.

97)

3.5

0.87

(0.8

4 –

0.91

)

Emer

gen

cy c

aes

are

an d

eliv

ery a

fter

inte

nded

vagin

al d

eliv

ery,

N =

67

296

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Adve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

1 58

70.

82

1.63

(1.3

1 –

2.12

)4.

5 1.

30 (1

.17

– 1.

44)

3.9

0.94

(0.8

4 –

1.05

)

Mid

par

tN

= 3

4 02

90.

51

Refe

renc

e3.

5 Re

fere

nce

4.1

Refe

renc

e

Hig

hest

quar

ter

N =

21

680

0.36

0.

72 (0

.55

– 0.

93)

2.4

0.67

(0.6

0 –

0.74

)3.

50.

83 (0

.76

– 0.

91)

adve

rse

neon

atal

out

com

e is

a co

mpo

site

outc

ome

of a

5 m

inut

e A

pgar

sco

re <

4 a

nd/

or a

dmiss

ion

to a

terti

ary

perin

atal

car

e ce

ntre

afte

r bi

rth a

nd/

or n

eona

tal

mor

talit

y an

d/or

hyp

oxic

isch

emic

enc

epha

lopa

thy

(HIE

) and

/or

neo

nata

l con

vulsi

ons

and/

or re

spira

tory

pro

blem

s.

3

49

Table

4. R

isk o

f adv

erse

neo

nata

l and

mat

erna

l out

com

e fo

r wom

en w

ith in

tend

ed v

agin

al d

eliv

ery

com

pare

d be

twee

n ho

spita

ls w

ith h

igh

or lo

w a

djus

ted

emer

genc

y ca

esar

ean

deliv

ery

rate

s ve

rsus

hos

pita

ls w

ith e

mer

genc

y ca

esar

ean

deliv

ery

rate

s w

ithin

the

mid

dle

rang

e, s

tratifi

ed fo

r mod

e of

del

iver

y.

Inte

nded

Vagin

al D

eliv

ery,

N =

45

0 3

19

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Adve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

01 3

250.

281.

22 (1

.06

– 1.

41)

1.9

1.08

(1.0

3 –

1.15

)3.

90.

98 (0

.94

– 1.

02)

Mid

par

tN

= 2

34 0

820.

23Re

fere

nce

1.7

Refe

renc

e3.

9Re

fere

nce

Hig

hest

quar

ter

N =

114

912

0.22

0.95

(0.8

2 –

1.10

)1.

60.

92 (0

.87

– 0.

97)

3.5

0.87

(0.8

4 –

0.91

)

Emer

gen

cy c

aes

are

an d

eliv

ery a

fter

inte

nded

vagin

al d

eliv

ery,

N =

67

296

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Adve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

1 58

70.

82

1.63

(1.3

1 –

2.12

)4.

5 1.

30 (1

.17

– 1.

44)

3.9

0.94

(0.8

4 –

1.05

)

Mid

par

tN

= 3

4 02

90.

51

Refe

renc

e3.

5 Re

fere

nce

4.1

Refe

renc

e

Hig

hest

quar

ter

N =

21

680

0.36

0.

72 (0

.55

– 0.

93)

2.4

0.67

(0.6

0 –

0.74

)3.

50.

83 (0

.76

– 0.

91)

adve

rse

neon

atal

out

com

e is

a co

mpo

site

outc

ome

of a

5 m

inut

e A

pgar

sco

re <

4 a

nd/

or a

dmiss

ion

to a

terti

ary

perin

atal

car

e ce

ntre

afte

r bi

rth a

nd/

or n

eona

tal

mor

talit

y an

d/or

hyp

oxic

isch

emic

enc

epha

lopa

thy

(HIE

) and

/or

neo

nata

l con

vulsi

ons

and/

or re

spira

tory

pro

blem

s.

disCussion

Caesarean delivery rates vary substantially between hospitals in the Netherlands. Higher rates of elective caesarean sections could result in better neonatal outcomes, although hospitals performing few elective caesarean sections have no significant increased risk of adverse neonatal outcome. Caesarean delivery after trial of labour in hospitals with restraint in emergency caesarean deliveries results in increased neonatal morbidity and mortality. Women who undergo elective caesarean delivery are at increased risk of adverse maternal outcomes.

Strengths and limitationsThis study comprises data from a large, well-maintained, population based national registry (PRN) including all 87 general hospitals, and therefore provides a reliable overview in the situation of variation in obstetric care among hospital births in the Netherlands. 12, 13

Several methods for comparing intervention rates have been proposed, all with respect to control for differences in patient case mix. 5, 10, 15-18 From the national PRN database we selected a clinically homogenous population consisting of primiparous women with term, cephalic singleton pregnancies who delivered in general hospitals. In the Netherlands obstetric care is highly standardized and accessible to everyone; so influence of socio-economic and demographic factors is limited. Aside from this homogenous study cohort, we built logistic regression models that included covariates pertaining to maternal and clinical variables to further control for differences in our patient population. 5, 19 Furthermore, the use of multilevel modelling for risk adjustment allowed inclusion of hospitals with a small number of births and taking into account similarities of births within hospitals. 7, 15, 18 A limitation of our analysis is the possible inaccuracy of the information on clinical risk factors used in the regression models to calculate the predicted rates. Data on some maternal characteristics of obstetric conditions, like body-mass index and abruption placenta, were not reliable available in the PRN registry so we could not include these in the regression models. It is however not very likely that these factors were to differ widely between the hospitals we studied.Using a low 5-minute Apgar score as a surrogate for adverse neonatal outcome has been questioned as the diagnosis is not based on an exact definition, such as umbilical pH-values. 20 Data on umbilical pH-values were however not available in the PRN. As we have a cohort of term singletons without congenital anomalies and in this subgroup of infants a low five minute Apgar score is caused generally by intrauterine asphyxia, we do not believe that our definition hinder our analysis. 21 A recent article by Iliodromiti et al. showed a very strong association between low five minute Apgar score and neonatal mortality due to intrapartum asphyxia. 22


Chapter 3

50

Other literatureLiterature on variation in caesarean rates shows, that there is a large variation in caesarean rates between hospitals, which can partly be explained by case-mix, labour and hospitals factors. 15, 17-19, 23 Results on the association between hospital caesarean rates and neonatal and maternal outcomes are conflicting. The most recent study from Nippita et al. found no significant differences in rates of adverse outcomes, whereas others have shown significant increase in neonatal asphyxia in hospitals performing little or many caesarean sections compared to average. 5, 9, 10, 15, 18 We only found increased asphyxia related adverse outcomes for infants delivered at hospitals performing few emergency caesareans. A possible explanation is that we stratified our analyses according to the actual mode of delivery. This stratified approach highlighted a subgroup of infants vaginally delivered in hospitals with high emergency caesarean delivery rates. One would expect that performing more than average emergency caesarean sections, would result in decreased morbidity among the remaining pool of vaginally delivered infants. We could not confirm this hypothesis, we found that vaginally delivered infants among hospital with high emergency caesarean rates have risk of adverse outcomes comparable to hospitals with average or even low emergency caesarean rates. In other words, with the selection of additional infants for caesarean delivery during the delivery process, some infants were missed who could have benefited from caesarean delivery. 10 Furthermore, there was an overall increased morbidity and mortality of infants born by caesarean delivery after intended vaginal delivery again suggesting there may be deficiencies in the timing or selection of cases during the delivery process. The increased morbidity of infants born after caesarean delivery in hospitals with low emergency-delivery rates could be a reflection of a more severe indication for the caesarean section (for example fetal blood sampling demonstrating acidosis instead of reassuring heart pattern on the CTG alone). However, it is beyond the scope of this article to illustrate that. In the PRN we could not distinguish between elective caesarean planned early in pregnancy or emergency pre-labour interventions. This might contribute to the overall high adverse neonatal outcome rates among the elective caesarean deliveries. Furthermore, 15% of the infants born after elective caesarean had a birth weight below the 10th percentile. As low birth weight is known to be highly correlated to adverse neonatal outcomes, this substantial subgroup probably influenced the outcomes in this group. 24

Obviously, caesarean delivery is an intervention that improves neonatal outcome in some women, thus justifying a risk of maternal complications. However, rates of elective caesarean deliveries with no clear medical of obstetrical indication are rising dramatically in Western countries. 25, 26 In the Netherlands, elective caesarean delivery in primiparous women with a singleton pregnancy is rarely performed, except in cases of breech presentation or the presence of placenta praevia. Exclusion of breech



3

51

pregnancies in our study ruled out the known great impact of maternal decision making on the caesarean delivery rate variance. Planned caesarean deliveries are associated with significantly increased risks of specific severe postpartum complications relative to planned vaginal deliveries. 26, 27 One could hypothesize that the increased occurrence of adverse maternal outcomes after elective caesarean delivery in our study may have been due to the clinical indications for caesarean delivery rather than to the procedure itself. A suggestion underlined by the decreased risk of maternal complications among women with an emergency caesarean delivery after trial of labour. Indications for emergency caesarean delivery are fetal distress or non-progressive labour, both conditions with small prior risk for adverse maternal outcomes. The best way to avoid confounding by indication would be a randomized controlled trial to evaluate the relative risks and benefits of planned elective caesarean versus planned vaginal delivery — although such a design would raise ethical concerns. Maternal complications after caesarean delivery do not only affect the post-partum period, they also cause risks for the course of the subsequent pregnancy. This consequence of a caesarean delivery could not be assessed in our study as we only included primiparous women.

Clinical implicationsHistorically, in the Netherlands there is conservative attitude towards caesarean delivery, with average caesarean delivery rates among the lowest of Western countries. Although many Western countries focused on lowering their caesarean rates, our study implicates that too few caesarean deliveries may affect neonatal outcomes adversely. The reported variation in both elective and emergency caesarean delivery rates causes concern and should be debated. Reasons for variance in obstetrical care and how the consistency can be improved need further study. Based on this study a general judgement on the quality of obstetric care of the individual Dutch hospitals cannot be made. Quality measurement is a very complex process and although adjustment for case-mix and hospital characteristics is necessary, a recent study by Grobman et al. demonstrated that the differences in frequency of adverse obstetric outcomes were for 50-100% unexplained by patient or hospital characteristics and the obstetric care provided. 23

Future studies exploring the risk-adjusted obstetric care variations could help improve the obstetric care by identifying modifiable characteristics of healthcare pathways and hospitals that contribute to the quality of care.

After adjustment for maternal and clinical factors the management of term cephalic deliveries still varies substantially among hospitals in the Netherlands. Performing less caesarean sections has negative effect on the neonatal outcomes; however, a caesarean delivery does not guarantee an improved neonatal outcome. This underlines the delicate balance surrounding care during labour and delivery and emphasizes the need for more general management in obstetric healthcare.


Chapter 3

52

referenCes

1. Villar J, Valladares E, Wojdyla D, et al. Caesarean delivery rates and pregnancy outcomes: the 2005 WHO global survey on maternal and perinatal health in Latin America. Lancet 2006;367:1819-29.

2. Bragg F, Cromwell DA, Edozien LC, et al. Variation in rates of caesarean section among English NHS trusts after accounting for maternal and clinical risk: cross sectional study. Bmj 2010;341:c5065.

3. Mikolajczyk RT, Schmedt N, Zhang J, Lindemann C, Langner I, Garbe E. Regional variation in caesarean deliveries in Germany and its causes. BMC pregnancy and childbirth 2013;13:99.

4. Sarrechia M, Van Gerven E, Hermans L, et al. Variation in 17 obstetric care pathways: potential danger for health professionals and patient safety? Journal of advanced nursing 2013;69:278-85.

5. Bailit JL, Garrett JM, Miller WC, Mcmahon MJ, Cefalo RC. Hospital primary cesarean delivery rates and the risk of poor neonatal outcomes. American journal of obstetrics and gynecology 2002;187:721-7.

6. Lutomski JE, Morrison JJ, Lydon-Rochelle MT. Regional variation in obstetrical intervention for hospital birth in the Republic of Ireland, 2005-2009. BMC pregnancy and childbirth 2012;12:123.

7. Lee YY, Roberts CL, Patterson JA, et al. Unexplained variation in hospital caesarean section rates. The Medical journal of Australia 2013;199:348-53.

8. G.P. W, M.J. VDB, S.L.N. Z, R. H, J.D. DJ, H. V. Dutch health care performance report 2010. National Institute for Public health and Environment 2010.

9. Pallasmaa N, Alanen A, Ekblad U, et al. Variation in cesarean section rates is not related to maternal and neonatal outcomes. Acta obstetricia et gynecologica Scandinavica 2013;92:1168-74.

10. Gould JB, Danielsen B, Korst LM, et al. Cesarean delivery rates and neonatal morbidity in a low-risk population. Obstetrics and gynecology 2004;104:11-9.





3

53


14. Schuit E, Kwee A, Westerhuis ME, et al. A clinical prediction model to assess the risk of operative delivery. BJOG : an international journal of obstetrics and gynaecology 2012;119:915-23.

15. Nippita TA, Lee YY, Patterson JA, et al. Variation in hospital caesarean section rates and obstetric outcomes among nulliparae at term: a population-based cohort study. BJOG : an international journal of obstetrics and gynaecology 2015;122:702-11.

16. Glantz JC. Cesarean delivery risk adjustment for regional interhospital comparisons. American journal of obstetrics and gynecology 1999;181:1425-31.

17. Howell EA, Zeitlin J, Hebert PL, Balbierz A, Egorova N. Association between hospital-level obstetric quality indicators and maternal and neonatal morbidity. Jama 2014;312:1531-41.

18. Ye J, Betran AP, Guerrero Vela M, Souza JP, Zhang J. Searching for the optimal rate of medically necessary cesarean delivery. Birth 2014;41:237-44.

19. Bailit JL, Dooley SL, Peaceman AN. Risk adjustment for interhospital comparison of primary cesarean rates. Obstetrics and gynecology 1999;93:1025-30.



22. Iliodromiti S, Mackay DF, Smith GC, Pell JP, Nelson SM. Apgar score and the risk of cause-specific infant mortality: a population-based cohort study. Lancet 2014;384:1749-55.

23. Grobman WA, Bailit JL, Rice MM, et al. Can differences in obstetric outcomes be explained by differences in the care provided? The MFMU Network APEX study. American journal of obstetrics and gynecology 2014.

24. Visentin S, Londero AP, Grumolato F, et al. Timing of delivery and neonatal outcomes for small-for-gestational-age fetuses. Journal of ultrasound in medicine : official journal of the American Institute of Ultrasound in Medicine 2014;33:1721-8.

25. NIH State-of-the-Science Conference Statement on cesarean delivery on maternal request. NIH consensus and state-of-the-science statements 2006;23:1-29.


Chapter 3

54

26. Liu S, Liston RM, Joseph KS, et al. Maternal mortality and severe morbidity associated with low-risk planned cesarean delivery versus planned vaginal delivery at term. CMAJ : Canadian Medical Association journal = journal de l’Association medicale canadienne 2007;176:455-60.

27. Pallasmaa N, Ekblad U, Gissler M. Severe maternal morbidity and the mode of delivery. Acta obstetricia et gynecologica Scandinavica 2008;87:662-8.

3

55

Appendix. Predictive model for elective or emergency caesarean delivery.

Variable Odds Ratio (95%CI) Elective CS

Odds Ratio (95%CI) Emergency CS

Maternal characteristics

Maternal age (years) 1.09 (1.08 – 1.10) 1.09 (1.05 – 1.05)

Non – western ethnicity 1.25 (1.18 – 1.34) 1.30 (1.27 – 1.33)

Low socio-economic status 1.00 (0.95 – 1.06) 1.03 (1.00 – 1.05)

Uterus myomatosis 5.15 (4.38 – 6.05) 1.04 (0.90 – 1.20)

Congenital uterus anomalies 2.60 (1.43 – 4.70) 1.86 (1.30 – 2.67)

Coagulation disorders 1.42 (1.02 – 1.98) 1.25 (1.08 – 1.45)

Artificial reproduction technologies 1.15 (1.09 – 1.22) 1.08 (1.05 – 1.10)

Clinical risk factors

Gestational age

37+0 – 37+6 weeks 8.64 (7.93 – 9.42) 0.78 (0.75 – 0.81)

38+0 – 38+6 weeks 8.37 (7.75 – 9.03) 0.73 (0.71 – 0.75)

39+0 – 39+6 weeks 3.21 (2.96 – 3.48) 0.81 (0.79 – 0.83)

40+0 – 40+6 weeks Reference Reference

41+0 – 41+6 weeks 1.24 (1.13 – 1.37) 1.39 (1.36 – 1.42)

Hypertensive diseases 0.95 (0.90 – 1.00) 1.38 (1.35 – 1.41)

Diabetes 1.06 (0.91 – 1.22) 1.84 (1.73 – 1.96)

SGAP5 2.02 (1.88 – 2.18) 1.41 (1.36 – 1.46)

LGAP95 1.88 (1.75 – 2.02) 2.37 (2.30 – 2.43)

SGAP5: Small for Gestational age, birth weight < 5th percentile

LGAP5: Large for Gestational age, birth weight > 95th percentile

Chapter 3

56

Supple

men

tal

table

1.

Risk

of a

dver

se n

eona

tal a

nd m

ater

nal o

utco

me

com

pare

d be

twee

n ho

spita

ls w

ith h

igh

or lo

w a

djus

ted

elec

tive

caes

area

n de

liver

y ra

tes

vers

us h

ospi

tals

with

ele

ctiv

e ca

esar

ean

deliv

ery

rate

s w

ithin

the

mid

dle

rang

e, s

tratifi

ed fo

r mod

e of

del

iver

y.

Tota

l popula

tion,

N =

45

8 7

12

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Neo

nata

l Mort

alit

yA

dve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

14 1

340.

260.

94 (0

.82

– 1.

08)

0.15

1.03

(0.8

6 –

1.24

)2.

01.

12 (1

.06

– 1.

18)

3.9

1.00

(0.9

7 –

1.04

)

Mid

par

tN

= 22

8 05

90.

27Re

fere

nce

0.15

Refe

renc

e1.

8Re

fere

nce

3.9

Refe

renc

e

Hig

hest

quar

ter

N =

116

519

0.22

0.79

(0.6

8 –

0.91

)0.

015

0.99

(0.8

3 –

1.19

)1.

60.

91 (0

.86

– 0.

96)

3.8

0.98

(0.9

5 –

1.02

)

Elec

tive

Caes

are

an d

eliv

ery,

N =

839

3Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Neo

nata

l Mort

alit

yA

dve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

313

1.4

1.40

(0.8

0-2.

46)

0.84

1.

28 (0

.63

– 2.

58)

4.5

1.03

(0.7

6 –

1.39

)7.

31.

53 (1

.19

– 1.

96)

Mid

par

tN

= 4

011

0.97

Re

fere

nce

0.66

Re

fere

nce

4.4

Refe

renc

e4.

9Re

fere

nce

Hig

hest

quar

ter

N =

306

90.

690.

71 (0

.41-

1.20

)0.

600.

91 (0

.50

– 1.

65)

3.4

0.76

(0.6

0 –

0.96

)5.

0 1.

01 (0

.82

– 1.

26)

Inte

nded

vagin

al d

eliv

ery N

= 4

50

319

Apgar

< 4

Neo

nata

l Mort

alit

yA

dve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

eRate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

12 8

210.

25

0.97

(0.8

4 –

1.12

)0.

15

1.04

(0.8

7 –

1.26

)2.

0 1.

13 (1

.08

– 1.

19)

4.3

1.05

(1.0

2 –

1.09

)

Mid

par

tN

= 2

24 0

480.

26

Refe

renc

e0.

14

Refe

renc

e1.

7 Re

fere

nce

4.1

Refe

renc

e

Hig

hest

quar

ter

N =

113

450

0.95

0.

82 (0

.70

– 0.

95)

0.14

0.

99 (0

.82

– 1.

20)

1.6

0.91

(0.8

6 –

0.96

)4.

00.

96 (0

.92

– 0

.99)

adve

rse

neon

atal

out

com

e is

a co

mpo

site

outc

ome

of a

5 m

inut

e A

pgar

sco

re <

4 a

nd/

or a

dmiss

ion

to a

terti

ary

perin

atal

car

e ce

ntre

afte

r bi

rth a

nd/

or n

eona

tal

mor

talit

y an

d/or

hyp

oxic

isch

emic

enc

epha

lopa

thy

(HIE

) and

/or

neo

nata

l con

vulsi

ons

and/

or re

spira

tory

pro

blem

s.

3

57

Supple

men

tal t

able

2. R

isk o

f adv

erse

neo

nata

l and

mat

erna

l out

com

e co

mpa

red

betw

een

hosp

itals

with

hig

h or

low

adj

uste

d em

erge

ncy

caes

area

n de

liver

y ra

tes

vers

us h

ospi

tals

with

ele

ctiv

e ca

esar

ean

deliv

ery

rate

s w

ithin

the

mid

dle

rang

e, s

tratifi

ed fo

r mod

e of

del

iver

y.To

tal P

opula

tion,

N =

45

0 3

19

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Neo

nata

l Mort

alit

yA

dve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

01 3

250.

281.

22 (1

.06

– 1.

41)

0.13

1.29

(1.0

7 –

1.55

)1.

91.

09 (1

.03

– 1.

15)

3.9

0.98

(0.9

4 –

1.02

)

Mid

par

tN

= 2

34 0

820.

23Re

fere

nce

0.17

Refe

renc

e1.

7Re

fere

nce

3.9

Refe

renc

e

Hig

hest

quar

ter

N =

114

912

0.22

0.95

(0.8

2 –

1.10

)0.

151.

11 (0

.92

– 1.

34)

1.6

0.92

(0.8

7 –

0.97

)3.

50.

87 (0

.84

– 0.

91)

Unpla

nned

caes

are

an d

eliv

ery a

fter

inte

nded

vagin

al d

eliv

ery,

N =

67 2

96

Caes

are

an r

ate

ca

tegori

esA

pgar

< 4

Neo

nata

l Mort

alit

yA

dve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

e

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 1

1 58

70.

82

1.63

(1.3

1 –

2.12

)0.

411.

69 (1

.18

– 2.

41)

4.5

1.30

(1.1

7 –

1.44

)3.

90.

94 (0

.84

– 1.

05)

Mid

par

tN

= 3

4 02

90.

51

Refe

renc

e0.

24Re

fere

nce

3.5

Refe

renc

e4.

1Re

fere

nce

Hig

hest

quar

ter

N =

21

680

0.36

0.

72 (0

.55

– 0.

93)

0.23

0.

93 (0

.65

– 1.

32)

2.4

0.67

(0.6

0 –

0.74

)3.

50.

83 (0

.76

– 0.

91)

Vagin

al d

eliv

ery,

N =

383 0

23

Apgar

< 4

Neo

nata

l Mort

alit

yA

dve

rse

neo

nata

l outc

om

eA

dve

rse

Mate

rnal o

utc

om

eRate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Rate

(%

)O

R (

95

%CI)

Low

est q

uarte

rN

= 8

9 73

80.

20

1.06

(0.8

8 –

1.27

)0.

14

1.20

(0.9

6 –

1.49

)1.

6 1.

08 (1

.02

– 1.

16)

4.4

1.03

(0.9

9 –

1.07

)

Mid

par

tN

= 2

00 0

530.

19

Refe

renc

e0.

11

Refe

renc

e1.

4Re

fere

nce

4.2

Refe

renc

e

Hig

hest

quar

ter

N =

93

232

0.19

1.

02 (0

.85

– 1.

22)

0.13

1.

09 (0

.88

– 1.

37)

1.4

0.99

(0.9

3 –

1.06

)3.

9 0.

91 (0

.88

– 0.

95)

adve

rse

neon

atal

out

com

e is

a co

mpo

site

outc

ome

of a

5 m

inut

e A

pgar

sco

re <

4 a

nd/

or a

dmiss

ion

to a

terti

ary

perin

atal

car

e ce

ntre

afte

r bi

rth a

nd/

or n

eona

tal

mor

talit

y an

d/or

hyp

oxic

isch

emic

enc

epha

lopa

thy

(HIE

) and

/or

neo

nata

l con

vulsi

ons

and/

or re

spira

tory

pro

blem

s.


Part II:Risk factors for perinatal asphyxia

related mortality and morbidity


Chapter 4 Recurrence risk of low Apgar score among term singletons:

a population-based cohort study

S. EnsingJ.M. Schaaf

A. Abu-HannaB.W.J. Mol

A.C.J. Ravelli

Acta Obstet Gynecol Scand. 2014 Sep;93(9):897-904

Chapter 4

62

absTraCT

ObjectiveTo examine the risk of recurrence of low Apgar score in a subsequent term singleton pregnancy.

DesignPopulation-based cohort study.

SettingThe Netherlands. Population. A total of 190 725 women with two subsequent singleton term live births between 1999 and 2007.

MethodsWe calculated the recurrence risk of low Apgar score after adjustment for possible confounders. Women with an elective cesarean delivery, fetus in breech presentation or a fetus with congenital anomalies were excluded. Results were reported separately for women with a vaginal delivery or a cesarean delivery at first pregnancy.

Main outcome measuresPrevalence of birth asphyxia, a 5-min Apgar score <7.

ResultsThe risk for an Apgar score of <7 in the first pregnancy was 0.99% and overall halved in the subsequent pregnancies (0.50%). For those with asphyxia in the first pregnancy, the risk of recurrence of a low Apgar score in the subsequent pregnancy was 1.1% (odds ratio 2.1, 95% confidence interval 1.4–3.3). This recurrence risk was present in women with a previous vaginal delivery (odds ratio 2.1, 95% confidence interval 1.2–3.5) and in women with a previous cesarean delivery (odds ratio 3.8, 95% confidence interval 1.7–8.5). Among women with a small-for-gestational-age infant in the subsequent pregnancy and a previous vaginal delivery, the recurrence risk was 4.8% (adjusted odds ratio 5.8, 95% confidence interval 2.0–16.5).

ConclusionWomen with birth asphyxia of the first born have twice the risk of renewed asphyxia at the next birth compared to women without birth asphyxia of the first born. This should be incorporated in the risk assessment of pregnant women.

Recurrence risk of low Apgar score

4

63

inTroduCTion

Birth asphyxia is one of the main causes of death as well as of short- and long-term adverse outcomes among new born infants. 1 Because of the very small number of term newborns with an Apgar score <7 and/or an umbilical cord pH ≤ 7.0. In term infants without severe congenital malformations, a low 5-min Apgar score most likely reflects intrauterine asphyxia. 2, 3 Indeed, a 5-min Apgar score <7 has been associated with neonatal mortality 4, with subsequent neurologic disability or low cognitive function. 5, 6

Data of the National Vital Statistics Reports, which contain data on US births, reported that 1.9% of all deliveries in 2011 were complicated by birth asphyxia. 7 According to the European PERISTAT project 8, which monitors perinatal health in Europe, 1–2% of the pregnancies are complicated by low Apgar scores. This would suggest that in the Netherlands, with 175 000 deliveries annually and a prevalence of birth asphyxia of 1.3%, approximately 2200 new born infants will suffer from perinatal asphyxia annually. To improve obstetric health care and prevent adverse neonatal outcomes, many studies have examined potential risk factors for birth asphyxia, such as intra-uterine growth retardation, prematurity and maternal hypertensive diseases. 9

In current practice, obstetric history is used as an important determinant of risk in a subsequent pregnancy and influences clinical management. After a previous complicated pregnancy or delivery, the information from the midwife or gynecologist on the estimated risk of recurrence of an adverse pregnancy outcome is very important for most parents. The risk of recurrence of adverse pregnancy outcomes such as preterm birth 10, 11, stillbirth 12 or small-for-gestational-age (SGA) 13 has been described in the literature. Data on the recurrence risk of birth asphyxia are lacking and literature on the risk of birth asphyxia after a previous cesarean delivery is limited. Our aim was to assess the recurrence risk of birth asphyxia in a subsequent pregnancy among term singletons.

maTerial and meThods

For this study we used a longitudinal linked dataset of 272 551 women with first and second deliveries, previously described by Schaaf et al. 10, 11 This linked dataset of first and second born children is based on a 9-year Netherlands Perinatal Registry (PRN) birth cohort in the period from 1999 until 2007. These data were obtained from the PRN, which consists of population-based data containing information on pregnancies, deliveries and (re)admissions until 28 days after birth. The PRN registry is a validated linkage of three registries: the midwifery registry (LVR1), the obstetrics registry (LVR2),

Chapter 4

64

and the neonatology registry (LNR) of hospital admissions of newborns. 14, 15 Permission for record use and analysis of data for the purpose of this study was obtained from PRN (registered as data petition 11.91). The study was submitted to the Medical Ethics Review Committee and it was confirmed that the Medical Research Involving Human Subjects (act WMO) did not apply to our research project (reference number W14_029 # 14.170045).

From this database, we selected all women who delivered two singleton live-born infants between 37+0 and 42+6 gestational age. Women who had elective cesarean section, breech presentation or a fetus with congenital anomalies in their first and/or second pregnancy were excluded to estimate the incidence of “pure” asphyxia related causes of low Apgar score. Birth asphyxia was defined as a 5-min Apgar score of <7 (any asphyxia).

Previous cesarean delivery increases the risk of birth asphyxia in a subsequent pregnancy 16-18; to eliminate this possible confounding factor we stratified the subsequent pregnancies into previous vaginal delivery and previous cesarean delivery. We analyzed the outcome of the subsequent pregnancy in (i) women with a previous vaginal delivery and a history of birth asphyxia, (ii) women with a previous cesarean delivery with and (iii) women without a history of birth asphyxia. Women with a vaginal delivery of the firstborn without a history of birth asphyxia served as reference population. For these three groups we compared demographic and obstetric baseline characteristics recorded in the PRN database: maternal age, gestational age, ethnicity, socio-economic status, use of in vitro fertilization, any hypertensive disease, birthweight below the 5th percentile (SGA), birthweight above the 95th percentile (large-for-gestational-age, LGA), late onset of perinatal care (≥18 weeks gestational age) and induction of labor.

Ethnicity was ascribed by the obstetric care giver; we differentiated between Western (native Dutch women and women from other Western nations) and non-Western (including different ethnicities such as African/Surinamese Creole, Hindustani, Mediterranean and Asian). The continuous socio-economic status score, which is based on mean income level, was categorized into a high/middle and low group based on percentiles. SGA and LGA were defined by the Dutch reference curves for birthweight by gestational age, sex and ethnicity.19

Statistical analysis Univariate analysis of the baseline characteristics was performed with Student’s t-test for normally distributed continuous variables and the chi-square test for categorical variables. All statistical tests had a p-value of 0.05 as the threshold for statistical significance.

4

65

The association between history of birth asphyxia and the effect of other risk factors for birth asphyxia in the subsequent pregnancy were calculated by multivariate logistic regression analysis. In addition, for each risk factor, we calculated the population attributable risk (PAR) percentage 20, which was based on the prevalence and relative risk (RR):

PAR% = [prevalence x (RR - 1)] / [prevalence x (RR - 1) + 1] x 100

The 95% confidence intervals (CI) were estimated to determine the precision of each odds ratio (OR). We first analyzed the total population and later stratified according to the mode of delivery in the first pregnancy. We adjusted for maternal age (years), gestational age (weeks), non-Western ethnicity, low socio-economic status, assisted reproduction technology (ART), hypertensive disease, intra-uterine growth retardation, late onset of perinatal care, and induction of labor. The confounders were tested for multicollinearity, to guarantee independency of the observations. We also assessed the interaction between (i) a low Apgar score of the previous delivery and (ii) SGA in the subsequent pregnancy on (iii) low Apgar score of the subsequent delivery, while adjusting for (i) and (ii). All statistical analyses were performed with SAS 9.2 (SAS Institute, Cary, NC, USA).

resulTs

There were 272 551 women with two subsequent deliveries between 1999 and 2007. After excluding women who in their first or second pregnancy had fetal mortality (0.49%), preterm delivery (11%), twin pregnancy (1.4%), breech presentation (8.9%), a fetus with congenital anomalies (3.3%) or elective cesarean section (7.2%), 190 725 women remained with complete follow-up data. Figure 1 shows that after a previous infant with a 5-min Apgar score of <7, the elective cesarean section rate was 13% compared with 3.5% for women with a previous healthy infant.

Frequencies of demographic, medical and obstetrical factors are shown in Table 1 for all second births. Women with an infant suffering from birth asphyxia in their subsequent pregnancy were more likely to be older and more frequently had a lower socio-economic status. Late onset of perinatal care, hypertensive diseases, previous cesarean delivery, intra-uterine growth restriction or macrosomia, and induction of labor were also more observed among women with a neonate with birth asphyxia.

Chapter 4

66

Table 1. Baseline characteristics of the subsequent delivery by asphyxia outcome second pregnancy.

Characteristics subsequent delivery

Apgar score ≥ 7N = 189 757

Apgar score < 7N = 968

P-value

N % N %Mean maternal age ª 31.0 ± 4.2 31.4 ± 4.3 < 0.0001Maternal age ≥ 36 years 25 119 13 148 15 0.06Mean gestational age ª 39.7 ± 1.3 39.7 ± 1.2 0.05Non – Western ethnicity 25 014 13 148 15 0.05Low socio-economic status 39 445 21 231 24 0.02Late onset perinatal care 11 410 6.0 82 8.5 0.001Hypertensive disease 10 351 5.5 83 8.6 < 0.0001IVF 798 0.42 9 0.93 0.01Induction of labor 42 579 22 354 37 < 0.0001Previous Cesarean delivery 14 768 7.8 169 17 < 0.0001SGA 7 647 4.0 72 7.4 < 0.0001LGA 10 826 5.7 107 11 < 0.0001First born Apgar score < 7 1 862 0.99 20 2.1 0.0007

ª presented as mean ± SD

Apgar score ≥ 7: a 5 minute Apgar score of seven or more.

Apgar score < 7: a 5 minute Apgar score of less than seven.

IVF: In Vitro Fertilization

SGA: Small-for-Gestational-age (birthweight < 5th percentile)

LGA: Large-for-Gestational-age (birthweight > 95th percentile)

The overall risk for an Apgar score of <7 in the first pregnancy was 0.99% (n = 1882) and the overall rate in the subsequent delivery was 0.50% (n = 968). Among women with a first born with birth asphyxia, the risk of renewed birth asphyxia in the subsequent pregnancy was unchanged; in 1.1% of the women (20/1882) there was recurrence of a low Apgar score (OR 2.1, 95% CI 1.4–3.3). Table 2 also shows a stratified analysis according to the mode of delivery in the first pregnancy. After vaginal delivery complicated by birth asphyxia, the recurrence risk was 0.92% (n = 14) (OR 2.1, 95% CI 1.2–3.5). If the preceding cesarean delivery was complicated by birth asphyxia, then risk of an Apgar score of <7 for the second born increased to 1.7% (n = 6) (OR 3.8, 95% CI 1.7–8.5).

The PAR of previous birth asphyxia was 2.0%, which is comparable to the PAR of maternal age, ethnicity, and late onset perinatal care. When the first born was delivered by cesarean section, the PAR of birth asphyxia increased to 10–13% (Table 2).

4

67

Table 2. Risk of low Apgar score and the population attributable percentage of risk factors for birth

asphyxia in subsequent live born term singletons in the Netherlands.

Subsequent delivery (N = 190 725)Crude Risk factors Apgar score

< 7Crude Odds Ratio (95%CI)

PAR*

N % %Previous asphyxia overallFirst born Apgar score ≥ 7 948 0.50 ReferenceFirst born Apgar score < 7 20 1.1 2.1 (1.4 – 3.3) 2.0

Previous asphyxia according to way of deliveryPrevious vaginal delivery and Apgar score first born ≥ 7 785 0.45 ReferencePrevious vaginal delivery and Apgar score first born < 7 14 0.92 2.1 (1.2 – 3.5) 2.2Previous cesarean delivery and Apgar score first born ≥ 7 163 1.1 2.5 (1.2 – 3.5) 10Previous cesarean delivery and Apgar score first born < 7 6 1.7 3.8 (1.7 – 8.5) 13

Other risk factorsMaternal age ≥ 36 years 148 0.59 1.2 (0.99 – 1.4) 2.4Non – Western Ethnicity 148 0.59 1.2 (0.99 – 1.4) 2.4Low Socio-Economic status 231 0.58 1.2 (1.1 – 1.4) 3.9IVF 9 1.12 2.2 (1.2 – 4.3) 0.51Late onset perinatal care 82 0.71 1.4 (1.2 – 1.8) 2.6Hypertensive disease 83 0.80 1.6 (1.3 – 2.0) 3.3SGA 72 0.93 1.9 (1.5 – 2.4) 3.5LGA 107 0.98 2.1 (1.7 – 2.5) 5.6Induction of labor 354 0.82 2.0 (1.7 – 2.3) 18

*PAR: Population – Attributive Risk.

Apgar score ≥ 7: a 5 minute Apgar score of seven or more.

Apgar score < 7: a 5 minute Apgar score of less than seven.

IVF: In Vitro Fertilization

SGA: Small-for-Gestational-age (birthweight < 5th percentile)

LGA: Large-for-Gestational-age (birthweight > 95th percentile)

Chapter 4

68

Table 3 presents unadjusted and adjusted odds ratios (aOR) for risk factors of an Apgar score of <7 in a subsequent pregnancy. After adjustment for possible confounders,the OR for recurrence of birth asphyxia after previous vaginal delivery was significantly increased in a subsequent pregnancy (aOR 1.8, 95% CI 1.1–3.0). This repeated risk was the highest when the first born was delivered by assisted vaginal delivery for suspected fetal distress (1.7%, aOR 3.1, 95% CI 1.4–7.1). Multivariate analysis showed that, besides the recurrence risk of a low Apgar score for newborns from women with a previous vaginal delivery, in the subsequent delivery, induction of labor, variables indicating that an infant is SGA or LGA, and late onset of perinatal care were all associated with an increased risk of birth asphyxia.

For women with a previous vaginal delivery there was a significant interaction between low Apgar score of the first born and SGA of the subsequent neonate. Therefore we carried out an additional stratified analysis for SGA of the second birth. Among SGA of the second births there was an increased recurrence risk for birth asphyxia (aOR 5.8, 95% CI 2.0–17). Separate analysis of women with a previous cesarean delivery showed that the repeated risk of birth asphyxia had an aOR of 1.3 (95% CI 1.2–1.4). When the previous cesarean delivery was performed for suspected fetal distress, the risk of low Apgar score in the subsequent delivery increased to 2.1% (aOR 3.8, 95% CI 1.6–9.3). Other factors associated with an increased risk of birth asphyxia in the subsequent pregnancy were the previous cesarean section itself (aOR 1.2, 95% CI 1.1–1.4), induction of labor and non-Western ethnicity (Table 3).

4

69

Table

3. U

nadj

uste

d an

d ad

juste

d od

ds ra

tios

(OR)

of b

irth

asph

yxia

(Apg

ar <

7) i

n a

subs

eque

nt te

rm s

ingl

eton

pre

gnan

cy, s

tratifi

ed a

ccor

ding

pre

viou

s m

ode

deliv

ery.

Mode

of

del

iver

yPre

vious

vagin

al d

eliv

ery (

N =

175

788

)Pre

vious

cesa

rean d

eliv

ery (

N =

14 9

37)

Ris

k f

act

ors

Unadju

sted

OR

(95

%CI)

Adju

sted

OR (

95

%CI)

*U

nadju

sted

OR

(95

%CI)

Adju

sted

OR (

95

%CI)

*

Obst

etri

c in

terv

ention f

etal d

istr

ess

pre

vious

pre

gnancy

Vagi

nal d

eliv

ery

and

Apg

ar s

core

fir

st bo

rn ≥

7Re

fere

nce

Refe

renc

eRe

fere

nce

Refe

renc

e

Ass

isted

vag

inal

del

iver

y fo

r sus

pect

ed fe

tal

distr

ess

and

Apg

ar s

core

firs

t bor

n <

73.

7 (1

.6 –

8.3

)3.

1 (1

.4 –

7.1

)X

X

Ces

area

n de

liver

y fo

r sus

pect

ed fe

tal d

istre

ss

and

Apg

ar s

core

firs

t bor

n <

7X

X4.

8 (2

.0 –

11.

7)3.

8 (1

.6 –

9.3

)

Mate

rnal a

ge

< 36

yea

rsRe

fere

nce

Refe

renc

eRe

fere

nce

Refe

renc

e≥

36 y

ears

1.1

(0.9

0 –

1.3)

1.1

(0.8

9 –

1.3)

1.3

(0.8

8 –

1.9)

1.3

(0.9

1 –

1.9)

Ethnic

ity

Wes

tern

Refe

renc

eRe

fere

nce

Refe

renc

eRe

fere

nce

Non

– W

este

rn1.

1 (0

.87

– 1.

3)0.

98 (0

.79

– 1.

2)1.

8 (1

.2 –

1.6

)1.

7 (1

.2 –

1.5

)So

cio –

Eco

nom

ic s

tatu

sH

igh/

mid

dle

Refe

renc

eRe

fere

nce

Refe

renc

eRe

fere

nce

Low

1.2

(0.9

9 –

1.4)

1.1

(0.9

6 –

1.4)

1.3

(0.9

0 –

1.8)

1.1

(0.7

6 –

1.6)

IVF

N

oRe

fere

nce

Refe

renc

eRe

fere

nce

Refe

renc

eYe

s1.

3 (0

.46

– 3.

5)1.

3 (0

.46

– 3.

5)3.

7 (1

.2 -

11)

9.0

(0.9

2 –

88)

Late

onse

t per

inata

l care

^N

oRe

fere

nce

Refe

renc

eRe

fere

nce

Refe

renc

eYe

s1.

6 (1

.3 –

2.1

)1.

6 (1

.3 –

2.0

)0.

58 (0

.29

– 1.

2)0.

58 (0

.29

– 1.

2)

Chapter 4

70

Mode

of

del

iver

yPre

vious

vagin

al d

eliv

ery (

N =

175

788

)Pre

vious

cesa

rean d

eliv

ery (

N =

14 9

37)

Ris

k f

act

ors

Unadju

sted

OR

(95

%CI)

Adju

sted

OR (

95

%CI)

*U

nadju

sted

OR

(95

%CI)

Adju

sted

OR (

95

%CI)

*H

yper

tensi

ve d

isea

seN

oRe

fere

nce

Refe

renc

eRe

fere

nce

Refe

renc

eYe

s1.

7 (1

.3 –

2.2

)1.

3 (0

.98

– 1.

6)1.

04 (0

.60

– 1.

8)1.

02 (0

.58

– 1.

8)SG

AN

oRe

fere

nce

Refe

renc

eRe

fere

nce

Refe

renc

eYe

s2.

0 (1

.5 –

2.6

)1.

8 (1

.3 –

2.1

) $1.

5 (0

.82

– 2.

7)1.

3 (0

.74

– 2.

4)LG

AN

oRe

fere

nce

Refe

renc

eRe

fere

nce

Refe

renc

eYe

s2.

4 (1

.9 –

3.0

)2.

3 (1

.9 –

2.9

)0.

63 (0

.29

– 1.

3)0.

63 (0

.29

– 1.

2)In

duct

ion o

f la

bor

No

Refe

renc

eRe

fere

nce

Refe

renc

eRe

fere

nce

Yes

1.9

(1.6

– 2

.2)

1.5

(1.3

– 1

.8)

1.6

(1.2

– 2

.1)

1.4

(1.2

– 1

.6)

SGA

* A

pgar

score

firs

t born

< 7

(inte

ract

ion t

erm

)N

on –

SG

AX

1.4

(0.7

7 –

2.7)

XX

SGA

X5.

8 (2

.0 –

17)

XX

*Adj

uste

d fo

r: m

ater

nal a

ge (

year

s), g

esta

tiona

l age

(w

eeks

), no

n –

wes

tern

eth

nici

ty, lo

w s

ocio

– e

cono

mic

sta

tus,

ass

isted

rep

rodu

ctio

n te

chno

logy

, hy

perte

nsiv

e di

seas

es, i

ntra

ute

rine

grow

th re

tard

atio

n, la

te o

nset

per

inat

al c

are

and

indu

ctio

n of

labo

r.^

Late

ons

et p

erin

atal

car

e: b

ooki

ng v

isit ≥

18

wee

ks g

esta

tion.

$

Ther

e w

as a

sig

nific

ant i

nter

actio

n in

the

grou

p w

ith a

pre

viou

s va

gina

l del

iver

y be

twee

n SG

A o

f the

sec

ond

born

and

low

Apg

ar o

f the

firs

t bor

n.

Apg

ar s

core

≥ 7

: a 5

min

ute

Apg

ar s

core

of s

even

or m

ore.

Apg

ar s

core

< 7

: a 5

min

ute

Apg

ar s

core

of l

ess

than

sev

en.

IVF:

In V

itro

Ferti

lizat

ion

SGA

: Sm

all-fo

r-Ges

tatio

nal-a

ge (b

irthw

eigh

t < 5

th p

erce

ntile

)LG

A: L

arge

–for

-Ges

tatio

nal-a

ge (b

irthw

eigh

t > 9

5th

perc

entil

e)

Table

3.

Con

tinue

d

4

71

disCussion

In this large longitudinal cohort study we investigated the recurrence risk of an Apgar score of <7. For all births the risk of an Apgar score of <7 was halved during the subsequent delivery. Except for newborns of women with a previous delivery complicated by a low Apgar score, their risk of a low Apgar score in the second birth was similar to that in the first birth. This was independent of the mode of delivery of the first born.

In the literature, the use of a low 5-min Apgar score alone as a definition for birth asphyxia has been questioned and an umbilical cord arterial pH of less than 7.0 has been suggested to be a better measure. 21, 22 However, the PRN dataset lacks information on the umbilical cord pH. As the strategy of determining the Apgar score did not change over the study period and because the 5-min Apgar score is accepted worldwide, we believe that our definition did not hinder our analysis. In particular, as we have a cohort of term singletons without congenital anomalies, in this subgroup of infants a 5-min Apgar score of <7 is caused generally by intra-uterine asphyxia. 2 A recent review by Ehrenstein et al. 5, 6 showed an association between a 5-min Apgar score of <7 and neonatal mortality and with subsequent neurologic disability or low cognitive function.

Our study was based on data from a large population based national registry. Since birth asphyxia and perinatal mortality are important but rare events, large amounts of data are necessary to detect any differences in populations. Although the PRN registry contains a lot of information, data on possible confounders such as tobacco use, body mass index, intra-partum fever and abruption placenta are not available in this database. Another drawback of this study might be the longitudinal linked dataset; non-linkage due to missing values of the linkage variables might have influenced the composition of our database. As “postal code of mother” was one of the linkage variables, changes of home address might have led to non-linkage. However, if the birth date of both the mother and child were available, linkage could still be done. 10, 11 Furthermore, comparison of our dataset with national figures of demographic characteristics and pregnancy outcomes did not show any major differences. In our database there is only one record for each specific woman, which consists of the first and second pregnancy and outcomes. For the multivariate analysis we only used variables of the second pregnancy and we found no multicollinearity between the variables.


Chapter 4

72

Risk factors for birth asphyxia are frequently described in the literature 9, 23, 24 and we found, according to that literature, that induction of labor, SGA and LGA were all significantly associated with an increased risk for a low Apgar score in a subsequent pregnancy. Studies addressing aspects of birth asphyxia in subsequent pregnancies focus on maternal and neonatal outcomes mainly after a previous cesarean delivery and present conflicting results. 16-18 None of the published studies reported on the recurrence risk of a low Apgar score (e.g. birth asphyxia) after a previous cesarean or vaginal delivery. For SGA there is substantial evidence of an increased recurrence risk in a subsequent pregnancy. 13, 25 As SGA infants are more vulnerable than their larger counterparts, SGA might contribute to the recurrence risk of birth asphyxia. In a stratified analysis for SGA of the second birth, we found that after previous vaginal delivery, SGA infants have the highest recurrence risk for birth asphyxia (aOR 5.8, 95% CI 2.0–17).

Assisted reproduction technology has been associated with an elevated risk of adverse outcomes for both mother and child. 26 We also found an increased risk of birth asphyxia for second births after in vitro fertilization (OR 2.2, 95% CI 1.2–4.3). As ART fetuses are at reach increased risk of intrauterine growth retardation, there might be a combined causal pathway in the increased risk of low Apgar score. Future research on this topic is needed to obtain a better understanding of the possible association between SGA, ART and birth asphyxia.

Along with prematurity, congenital anomalies and intrauterine growth retardation, birth asphyxia is one of the four major causes of an adverse neonatal outcome. 27 In the Netherlands obstetric health care is subdivided into primary health care, which includes the midwives and general practices who attend low-risk pregnancies, and secondary health care, where gynecologists attend high-risk pregnancies. After a previous cesarean delivery, women will receive secondary obstetric care in their future pregnancies because they are considered high-risk patients. On the other hand, women with a previous vaginal delivery which was complicated by birth asphyxia receive primary health care (i.e. referred to midwife’s practices) in their next pregnancy, as their subsequent pregnancy is considered to be a low-risk pregnancy. According to our findings, in which women with birth asphyxia of the first born had twice as a high a risk of renewed asphyxia at the next birth independent of the mode of delivery of the first birth, this division is remarkable. This is particularly the case where other risk factors for birth asphyxia such as SGA, LGA, hypertensive diseases and advanced maternal age classify a pregnancy as high-risk. This study shows that these risk factors have a PAR and OR for birth asphyxia in a subsequent pregnancy that is comparable to the PAR and OR of the asphyxia of the first born. Despite the low prevalence, an obstetric



4

73

history of birth asphyxia and the interaction of all these risk factors together should be incorporated in the risk estimation for the subsequent pregnancy. These results can be used to support obstetric caregivers in their patient counseling.

In conclusion, the findings of this study show that women with birth asphyxia of the first born have twice the risk of renewed asphyxia at the next birth compared with women without birth asphyxia of their first child. This finding was independent of the mode of delivery of the first birth.


Chapter 4

74

referenCes

1. Kaandorp JJ, Benders MJ, Rademaker CM, et al. Antenatal allopurinol for reduction of birth asphyxia induced brain damage (ALLO-Trial); a randomized double blind placebo controlled multicenter study. BMC pregnancy and childbirth 2010;10:8.









10. Schaaf JM, Hof MH, Mol BW, Abu-Hanna A, Ravelli AC. Recurrence risk of preterm birth in subsequent twin pregnancy after preterm singleton delivery. BJOG : an international journal of obstetrics and gynaecology 2012;119:1624-9.

11. Schaaf JM, Hof MH, Mol BW, Abu-Hanna A, Ravelli AC. Recurrence risk of preterm birth in subsequent singleton pregnancy after preterm twin delivery. American journal of obstetrics and gynecology 2012;207:279 e1-7.

12. Bhattacharya S, Prescott GJ, Black M, Shetty A. Recurrence risk of stillbirth in a second pregnancy. BJOG : an international journal of obstetrics and gynaecology 2010;117:1243-7.

13. Voskamp BJ, Kazemier BM, Ravelli AC, Schaaf J, Mol BW, Pajkrt E. Recurrence of small-for-gestational-age pregnancy: analysis of first and subsequent singleton pregnancies in The Netherlands. American journal of obstetrics and gynecology 2013;208:374 e1-6.



4

75



16. Carlsson Wallin M, Ekstrom P, Marsal K, Kallen K. Apgar score and perinatal death after one previous caesarean delivery. BJOG : an international journal of obstetrics and gynaecology 2010;117:1088-97.

17. Huang X, Lei J, Tan H, Walker M, Zhou J, Wen SW. Cesarean delivery for first pregnancy and neonatal morbidity and mortality in second pregnancy. European journal of obstetrics, gynecology, and reproductive biology 2011;158:204-8.

18. Hemminki E, Shelley J, Gissler M. Mode of delivery and problems in subsequent births: a register-based study from Finland. American journal of obstetrics and gynecology 2005;193:169-77.


20. Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. American journal of epidemiology 1974;99:325-32.





25. Kinzler WL, Kaminsky L. Fetal growth restriction and subsequent pregnancy risks. Seminars in perinatology 2007;31:126-34.

26. Pinborg A, Wennerholm UB, Romundstad LB, et al. Why do singletons conceived after assisted reproduction technology have adverse perinatal outcome? Systematic review and meta-analysis. Human reproduction update 2013;19:87-104.

27. Van Der Kooy J, Poeran J, De Graaf JP, et al. Planned home compared with planned hospital births in the Netherlands: intrapartum and early neonatal death in low-risk pregnancies. Obstetrics and gynecology 2011;118:1037-46.


Chapter 5 Risk of poor neonatal outcome at term after medically assisted reproduction:

a propensity score-matched study

S. EnsingA. Abu-HannaT.J. Roseboom

S. ReppingF. van der Veen

B.W.J. MolA.C.J. Ravelli

Fertil Steril. 2015 Aug;104(2):384-390

Chapter 5

78

absTraCT

ObjectiveTo study risk of birth asphyxia and related morbidity among term singletons born after medically assisted reproduction (MAR).

DesignPopulation cohort study.

SettingNot applicable.

PatientA total of 1,953,932 term singleton pregnancies selected from a national registry for 1999–2011.

Main Outcome MeasurePrimary outcome Apgar score <4; secondary outcomes Apgar score <7, intrauterine fetal death, perinatal mortality, congenital anomalies, small for gestational age, asphyxia related morbidity, and cesarean delivery.

ResultThe risks of birth asphyxia and related morbidity were calculated in women who conceived either through MAR or spontaneously (SC), with a subgroup analysis for in vitro fertilization (IVF). An additional propensity score matching analysis was performed with matching on multiple maternal baseline covariates (maternal age, ethnicity, socioeconomic status, parity, year of birth, and preexistent diseases). Each MAR pregnancy was matched to three SC controls. Relative to SC, the MAR singletons had an increased risk of adverse neonatal outcomes including Apgar score <4 (adjusted odds ratio [OR] 1.29; 95% CI, 1.14–1.46) and intrauterine fetal death (adjusted OR 1.61; 95% CI, 1.35–1.91). After propensity score matching, the risk of an Apgar score <4 was comparable between MAR and SC singletons (OR 0.99; 95% CI, 0.87–1.14). Cesarean delivery for both fetal distress and nonprogressive labor occurred more among MAR pregnancies compared with SC pregnancies.

ConclusionTerm singletons conceived after MAR have an increased risk of morbidity related to birth asphyxia. Because this is mainly due to maternal characteristics, obstetric caregivers should be aware that the increased rates of cesareans reflect the behavior of women and physicians rather than increased perinatal complications.

Risk of poor neonatal outcome after MAR

5

79

inTroduCTion

Since the introduction of medically assisted reproduction (MAR) over 35 years ago, the number of children born after MAR has rapidly increased. In Europe, more than half a million treatment cycles have been reported, according to the most recent report from the European Society of Human Reproduction and Embryology covering 2009. 1 Despite the expanding use of MAR, there is concern regarding the safety of these treatments for both mother and child. 2 Although the elevated risks of adverse pregnancy outcomes have mostly been attributed to the occurrence of multiple gestations, there is growing evidence that MAR singletons also have increased risks of perinatal mortality and morbidity, such as low birth weight, prematurity, congenital malformations, and cerebral palsy as well as increased levels of cardiometabolic risk factors. 2-4 These potential safety concerns are becoming an increasingly important topics in the medical, social, and political debates pertaining to MAR. Consequently, more current studies focus on evaluating the effects on perinatal outcomes and long-term health among these infants. 5, 6

Among the poor neonatal outcomes, asphyxia-related problems are rarely described in the MAR literature, although peripartum asphyxia is one of the main causes of perinatal morbidity and mortality. 7 Risks for an Apgar score <7 among IVF singletons compared with the total population have both been reported to be both increased and not increased. 8-11 None of these studies have distinguished between different gestational ages or used matched populations. Therefore, the results of these studies cannot easily be attributed to parental factors associated with subfertility and/or MAR treatment or to the other main risk factors for birth asphyxia such as prematurity.

Our study assessed the risk of poor neonatal outcomes with a focus on birth asphyxia in a large retrospective cohort of liveborn term infants. Additionally we used propensity score matching analysis to minimize the selection bias and maternal confounding factors. The propensity score is a single variable that represents a collection of maternal characteristics and is defined as the probability of getting treatment (e.g., MAR) or not. Estimating a propensity score for MAR treatment and matching based on this score may help to assess whether MAR has a causal effect on birth asphyxia. 12-15

meThods

We studied a nationwide cohort using data from the Netherlands Perinatal Registry (PRN). The PRN consists of population-based data containing information on pregnancies, deliveries, and (re)admissions until 28 days after birth. The PRN registry

Chapter 5

80

is obtained by a validated linkage of three different registries: the midwifery registry (LVR1), the obstetricsregistry (LVR2), and the neonatology registry (LNR) of hospital admissions of newborns. 16, 17 The PRN registry contains anonymous data and covers about 96% of all deliveries in the Netherlands.

For this study all singletons between 37 + 0 and 41 + 6 weeks of gestation, born between January 1, 1999, and December 31, 2010, were included. Gestational age was based on the last menstrual period, crown-rump length (CRL) measured during early pregnancy ultrasound, or both. We defined MAR as any form of nonspontaneous conception filed by the obstetric caregiver in the PRN registry at a prenatal visit, including in vitro fertilization (IVF), intrauterine insemination with or without ovarian hyperstimulation, or nonspecified procedures. To assess the influence of laboratory techniques, we performed a subgroup analysis on IVF pregnancies.

Our primary outcome was birth asphyxia defined as a 5-minute Apgar score <4. Secondary outcomes included intrauterine fetal death (IUFD), perinatal mortality (intrapartum mortality or within 7 days after live birth), Apgar score <7, congenital anomalies, small for gestational age (birth weight <5th percentile), poor neonatal outcome, and elective cesarean delivery or cesarean delivery for suspected fetal distress or nonprogressive labor. Poor neonatal outcome was defined as 5-minute Apgar score <7, IUFD, perinatal mortality, and asphyxia-related morbidity (composite of hypoxic ischemic encephalopathy, convulsions, and ischemic cerebral anomalies).

Statistics The incidence of the main outcome measures was compared between MAR and spontaneously conceived (SC) pregnancies and between the subgroup of IVF and SC pregnancies. First we described and compared the maternal baseline characteristics of the women by conception categories. We used Student’s ttest for normally distributed continuous variables and the chisquare test for categorical variables. To estimate the influence of MAR pregnancies on birth asphyxia, univariate and multivariate logistic regression analyses were performed, and we obtained the odds ratios (OR) along with their 95% confidence intervals (CI). Potential confounders were selected based on the literature and on clinical reasoning. 2, 4, 8, 9, 18-20 For each variable we decided whether it is a possible confounder by applying the traditional criteria for identifying confounders by three conditions. 21 We adjusted for maternal age (years), maternal ethnicity (Caucasian versus other), socioeconomic status (low vs. high/middle), diabetes mellitus (preexistent or gestational vs. none), hypertensive disease (essential, pregnancy induced, or preeclampsia vs. none), gestational age (days), birth weight (grams), congenital anomalies, fetal presentation (cephalic vs. other), parity, and mode of delivery.

5

81

Finally, because infertility patients may systematically differ from fertile patients on numerous factors contributing to birth asphyxia, we implemented a propensity score matching approach to identify fertile women who were most similar to the infertile women. 12-14 For each woman with a term singleton pregnancy, the probability of receiving MAR (i.e., propensity score) was estimated from a logistic regression model that considered multiple maternal baseline covariates (maternal age at delivery, maternal ethnicity, socioeconomic status, parity, year of birth, and preexistent diseases correlated to subfertility such as diabetes, hypertension, endocrinologic disorders, congenital uterus anomalies, endometriosis, uterus myomatosis, or coagulation disorders). We then implemented one-to-one matching without replacement on the closest propensity score of the MAR women and SC women (nearest neighbor matching). Furthermore, the matching required equality on parity, previous cesarean delivery, and level of care at onset of labor (primary care midwifery led care or secondary hospital-led obstetric care). By first checking all matching variables as possible confounders, we reduced the risk for overmatching to a minimum. 22 The process was repeated until each case (women pregnant after MAR) was matched with three controls (SC women). Having more than one control per case increased the power by reducing biasing of the test statistic and increasing the effective sample size. 23 Additionally we repeated this matching procedure for the subgroup IVF pregnancies and SC controls separately. In assessing and testing the association between treatment and outcome, we accounted for correlation within the matched women. Data were analyzed using the R statistical software environment version 2.15.1 (R Foundation for Statistical Computing, Vienna, Austria) with the MatchIt Library for the matched analysis. All statistical tests had a P value of .05 as the threshold to indicate statistical significance. Permission for record use and analysis of data for the purpose of this study was obtained from PRN (registered as data petition 11.91).

resulTs

We studied 1,953,932 women with singleton term pregnancies of whom 48,921 (2.5%) were pregnant after MAR. Among the MAR pregnancies 16,177 (33%) conceived by IVF. Comparison of the baseline characteristics of the total eligible study population are shown in Table 1. Women in the MAR groups tended to be older, Caucasian, and nulliparous, and more had a preexisting disease. The MAR women with a singleton term pregnancy more often received secondary care than those who had conceived spontaneously (P<.0001). The results in terms of risk estimations after logistic regression were performed for the total group of MAR singletons versus SC singletons and for the subgroup of IVF versus SC. The analyses were done on the total eligible study population (shown first) and on the propensity score matched sample.

Chapter 5

82

Table

1.

Base

line

char

acte

ristic

s of

the

elig

ible

stu

dy s

ampl

e an

d af

ter p

rope

nsity

sco

re m

atch

ing

for s

ingl

eton

s fo

llow

ing

MA

R ve

rsus

spo

ntan

eous

con

cept

ion.

Elig

ible

stu

dy S

am

ple

N=

1 9

53

932

Aft

er p

ropen

sity

sco

re m

atc

hin

gN

= 1

95

684

MA

RN

= 4

8 9

21

Sponta

neo

us

N =

1 9

05

01

1M

AR

N =

48

921

Sponta

neo

us

N =

146 7

63

N (

%)

N (

%)

P-v

alu

eN

(%

)N

(%

)P-v

alu

e

Mea

n ge

statio

nal a

ge *

(wee

ks)

39.4

± 1

.339

.6 ±

1.3

< 0.

0001

39.4

± 1

.439

.5 ±

1.3

< 0.

0001

Mat

erna

l age

* (y

ears

)32

.7 ±

4.5

30.4

± 4

.9<

0.00

0132

.7 ±

4.6

32.7

± 4

.50.

55

Cau

casia

n Et

hnic

ity43

803

(90%

)1

609

264

(84%

)<

0.00

0143

803

(90%

)13

2 08

6 (9

0%)

0.45

Low

soc

io-e

cono

mic

sta

tus

9959

(20%

)47

9 47

7 (2

5%)

< 0.

0001

9959

(20%

)30

191

(21%

)0.

31

Pree

xiste

nt d

iseas

e ^

7420

(15%

)34

706

(1.8

%)

< 0.

0001

7420

(15%

)21

105

(14%

)<

0.00

01

Nul

lipar

ous

32 9

04 (6

7%)

857

971

(45%

)<

0.00

0132

904

(67%

)98

712

(67%

)1.

0

Prev

ious

ces

area

n de

liver

y30

42 (6

.2%

)11

7 94

9 (6

.2%

)0.

8130

42 (6

.2%

)91

26 (6

.2%

)1.

0

Seco

ndar

y ca

re a

t ons

et la

bor

37 0

92 (7

6%)

775

623

(41%

)<

0.00

0137

092

(76%

)11

1 27

6 (7

6%)

1.0

*Pre

sent

ed a

s M

ean

± SD

MA

R: M

edic

ally

Ass

isted

Rep

rodu

ctio

n (In

tra u

terin

e in

sem

inat

ion,

In v

itro

ferti

lizat

ion,

intra

cyto

plas

mic

spe

rm in

ject

ion)

^Pre

exist

ent d

iseas

e in

clud

es o

ne o

f the

follo

win

g: p

reex

isten

t hyp

erte

nsio

n, p

reex

isten

t dia

bete

s, e

ndoc

rinol

ogic

al d

isord

ers,

coa

gula

tion

diso

rder

s, e

ndom

etrio

sis, u

teru

s m

yom

atos

is an

d co

ngen

ital u

teru

s an

omal

ies.

5

83

MAR Versus SCThe MAR singletons had an increased risk of an Apgar score <4 compared with the SC singletons (adjusted OR [aOR] 1.29; 95% CI, 1.14–1.46). The perinatal mortality risk did not statistically significantly differ between these groups, although MAR singletons had an increased risk of IUFD (aOR 1.61; 95% CI, 1.35–1.91). Congenital anomalies (aOR 1.33; 95% CI, 1.26–1.40) and small for gestational age (aOR 1.14; 95% CI, 1.09–1.18) occurred more among MAR infants. Overall MAR singletons had a 2.3% risk of a poor neonatal outcome compared with 1.6% among SC singletons (aOR 1.05; 95% CI, 0.99–1.12). The risk of delivery by cesarean section was higher for MAR singletons compared with SC singletons: 4.8% of the MAR singletons received a cesarean section for suspected fetal distress (aOR 1.44; 95% CI, 1.38–1.50) (Table 2).

IVF Versus SCTable 3 shows that for the subgroup of IVF there was a borderline significant increased risk of an Apgar score <4 compared with SC singletons (aOR 1.23; 95% CI, 1.00–1.53). There was also an increased risk of IUFD: 2.6 per 1,000 MAR pregnancies versus 1.4 per 1,000 SC pregnancies (aOR 1.35; 95% CI, 0.99– 1.82). The overall risk of a poor neonatal outcome was slightly higher for IVF infants (2.1%) compared with SC infants (1.6%), although after adjustment for confounders this turned the other way around and was no longer statistically significant (aOR 0.97; 95% CI, 0.87–1.09). The IVF singletons were more often small for gestational age compared with their SC counterparts — 6.0% versus 4.9% (aOR 1.25; 95% CI, 1.17– 1.34) — and had more often congenital anomalies (aOR 1.31; 95% CI, 1.19–1.43). After adjustment for confounders such as breech presentation, the risk for elective cesarean delivery was doubled for IVF singletons: 10% versus 5.4%. Risk of cesarean delivery for fetal distress was also increased for IVF singletons compared with SC singletons (aOR 1.40; 95% CI, 1.30–1.50), which is comparable to the total singleton MAR population.

Chapter 5

84

Table

2.

Preg

nanc

y an

d ne

onat

al o

utco

mes

am

ong

term

sin

glet

ons

follo

win

g M

AR

vers

us s

pont

aneo

us c

once

ptio

n.

MA

RN

= 4

8 9

21

Sponta

neo

us

N =

1 9

05

01

1M

AR v

ersu

s Sp

onta

neo

us

MA

R v

ersu

s Sp

onta

neo

us

N (

% o

r ‰

)N

(%

or

‰)

Cru

de

Odds

ratio (

95

%CI)

Adju

sted

Odds

ratio (

95

%CI)

Neo

nata

l outc

om

esA

pgar

sco

re <

4 *

260

(5.3

‰)

6401

(3.4

‰)

1.59

(1.4

0 –

1.80

)1.

29 (1

.14

– 1.

46)

Apg

ar s

core

< 7

*76

0 (1

6‰)

20 7

21 (1

1‰)

1.44

( 1.

33 –

1.5

4)1.

08 (1

.04

– 1.

17)

Intra

uter

ine

feta

l dea

d **

140

(2.9

‰)

2743

(1.4

‰)

1.99

(1.6

8 –

2.36

)1.

61 (1

.35

– 1.

91)

Perin

atal

mor

talit

y *

67 (1

.4‰

)19

54 (1

.0‰

)1.

33 (1

.05

– 1.

71)

1.01

(0.7

8 –

1.28

)C

onge

nita

l ano

mal

ies

***

1420

(29‰

)39

840

(21‰

)1.

40 (1

.33

– 1.

48)

1.33

(1.2

6 –

1.40

)SG

A P

5 **

**27

19 (5

6‰)

94 1

40 (4

9‰)

1.13

(1.0

9 –

1.18

)1.

14 (1

.09

– 1.

18)

Poor

neo

nata

l out

com

e *

1101

(23‰

)30

475

(16‰

)1.

42 (1

.33

– 1.

51)

1.05

(0.9

9 –

1.12

)Pre

gnancy

outc

om

esEl

ectiv

e C

S **

***

4572

(9.3

%)

102

712

(5.4

%)

1.81

(1.7

5 –

1.87

)1.

55 (1

.49

– 1.

61)

CS

feta

l dist

ress

***

**23

34 (4

.8%

)44

939

(2.4

%)

2.08

(1.9

9 –

2.17

)1.

44 (1

.38

– 1.

50)

CS

non-

prog

ress

ive

labo

r ***

**46

80 (9

.6%

)10

0 50

5 (5

.3%

)1.

90 (1

.84

– 1.

96)

1.34

(1.3

0 –

1.38

)

Perin

atal

mor

talit

y: in

trapa

rtum

mor

talit

y or

mor

talit

y w

ithin

28

days

afte

r liv

e bi

rth.

Poor

neo

nata

l out

com

e is

a co

mpo

site

of: A

pgar

sco

re <

7, i

ntra

uter

ine

feta

l dea

d, p

erin

atal

mor

talit

y, N

ICU

adm

issio

n, h

ypox

ic is

chem

ic e

ncep

halo

path

y, c

onvu

lsion

s,

and

resp

irato

ry s

uppo

rt.

MA

R: M

edic

ally

Ass

isted

Rep

rodu

ctio

nC

S: C

esar

ean

sect

ion

* A

djus

ted

for:

mat

erna

l age

(yea

rs),

low

soc

io-e

cono

mic

sta

tus,

eth

nici

ty, d

iabe

tes,

any

hyp

erte

nsiv

e di

seas

e, p

arity

, ge

statio

nal a

ge (d

ays),

feta

l pr

esen

tatio

n, b

irth

wei

ght (

gram

s), c

onge

nita

l ano

mal

ies

and

mod

e of

del

iver

y.

** A

djus

ted

for:

mat

erna

l age

(yea

rs),

low

soc

io-e

cono

mic

sta

tus,

eth

nici

ty, d

iabe

tes,

any

hyp

erte

nsiv

e di

seas

e, p

arity

, ge

statio

nal a

ge (d

ays),

birt

h w

eigh

t (gr

ams)

and

cong

enita

l ano

mal

ies.

***

Adj

uste

d fo

r: m

ater

nal a

ge (y

ears

), lo

w s

ocio

-eco

nom

ic s

tatu

s, e

thni

city,

dia

bete

s, a

ny h

yper

tens

ive

dise

ase

and

parit

y.**

** A

djus

ted

for:

mat

erna

l age

(yea

rs),

low

soc

io-e

cono

mic

sta

tus,

eth

nici

ty, d

iabe

tes,

any

hyp

erte

nsiv

e di

seas

e, p

arity

and

con

geni

tal a

nom

alie

s.

****

* A

djus

ted

for:

mat

erna

l age

(yea

rs),

low

soc

io-e

cono

mic

sta

tus,

eth

nici

ty, d

iabe

tes,

any

hyp

erte

nsiv

e di

seas

e, p

arity

, ge

statio

nal a

ge (d

ays),

fe

tal p

rese

ntat

ion,

birt

h w

eigh

t (gr

ams)

and

cong

enita

l ano

mal

ies.

5

85

Table

3.

Preg

nanc

y an

d ne

onat

al o

utco

mes

am

ong

term

sin

glet

ons

follo

win

g IV

F ve

rsus

spo

ntan

eous

con

cept

ion.

IVF

N =

16

17

7Sp

onta

neo

us

N =

1 9

05

011

IVF

vers

us

Sponta

neo

us

IVF

vers

us

Sponta

neo

us

N (

% o

r ‰

)N

(%

or

‰)

Cru

de

Odds

ratio (

95

%CI)

Adju

sted

Odds

ratio (

95

%CI)

Neo

nata

l outc

om

esA

pgar

sco

re <

4 *

86 (5

.3‰

)64

01 (3

.4‰

)1.

59 (1

.28

– 1.

96)

1.23

(1.0

0 –

1.53

)A

pgar

sco

re <

7 *

238

(15‰

)20

721

(11‰

)1.

36 (1

.20

– 1.

55)

1.01

(0.8

8 –

1.15

)In

traut

erin

e fe

tal d

ead

**42

(2.6

‰)

2743

(1.4

‰)

1.81

(1.3

3 –

2.45

)1.

35 (0

.99

– 1.

82)

Perin

atal

mor

talit

y *

23 (1

.4‰

)19

54 (1

.0‰

)1.

39 (0

.92

– 2.

10)

1.06

(0.6

9 –

1.61

)C

onge

nita

l ano

mal

ies

***

466

(29‰

)39

840

(21‰

)1.

39 (1

.27

– 1.

53)

1.31

(1.1

9 –

1.43

)SG

A P

5 **

**97

0 (6

0‰)

94 1

40 (4

9‰)

1.23

(1.1

5 –

1.31

)1.

25 (1

.17

– 1.

34)

Poor

neo

nata

l out

com

e *

344

(21‰

)30

475

(16‰

)1.

34 (1

.20

– 1.

49)

0.97

(0.8

7 –

1.09

)Pre

gnancy

outc

om

esEl

ectiv

e C

S **

***

1682

(10%

)10

2712

(5.4

%)

2.04

(1.9

4 –

2.14

)1.

69 (1

.58

– 1.

80)

CS

feta

l dist

ress

***

**80

6 (5

.0%

)44

939

(2.4

%)

2.17

(2.0

2 –

2.33

)1.

40 (1

.30

– 1.

50)

CS

non-

prog

ress

ive

labo

r ***

**14

50 (9

.0%

)10

0 50

5 (5

.3%

)1.

78 (1

.67

– 1.

87)

1.19

(1.1

3 –

1.26

)Pe

rinat

al m

orta

lity:

intra

partu

m m

orta

lity

or m

orta

lity

with

in 2

8 da

ys a

fter l

ive

birth

.Po

or n

eona

tal o

utco

me

is a

com

posit

e of

: Apg

ar s

core

< 7

, int

raut

erin

e de

ad, p

erin

atal

mor

talit

y, N

ICU

adm

issio

n, h

ypox

ic is

chem

ic e

ncep

halo

path

y, c

onvu

lsion

s, a

nd

resp

irato

ry s

uppo

rt.

IVF:

In V

itro

Ferti

lizat

ion

CS:

Ces

area

n se

ctio

n

* A

djus

ted

for:

mat

erna

l age

(yea

rs),

low

soc

io-e

cono

mic

sta

tus,

eth

nici

ty, d

iabe

tes,

any

hyp

erte

nsiv

e di

seas

e, p

arity

, ge

statio

nal a

ge (d

ays),

feta

l pr

esen

tatio

n, b

irth

wei

ght (

gram

s), c

onge

nita

l ano

mal

ies

and

mod

e of

del

iver

y.

** A

djus

ted

for:

mat

erna

l age

(yea

rs),

low

soc

io-e

cono

mic

sta

tus,

eth

nici

ty, d

iabe

tes,

any

hyp

erte

nsiv

e di

seas

e, p

arity

, ge

statio

nal a

ge (d

ays),

birt

h w

eigh

t (gr

ams)

and

cong

enita

l ano

mal

ies.

***

Adj

uste

d fo

r: m

ater

nal a

ge (y

ears

), lo

w s

ocio

-eco

nom

ic s

tatu

s, e

thni

city,

dia

bete

s, a

ny h

yper

tens

ive

dise

ase

and

parit

y.**

** A

djus

ted

for:

mat

erna

l age

(yea

rs),

low

soc

io-e

cono

mic

sta

tus,

eth

nici

ty, d

iabe

tes,

any

hyp

erte

nsiv

e di

seas

e, p

arity

and

con

geni

tal a

nom

alie

s.

****

* A

djus

ted

for:

mat

erna

l age

(yea

rs),

low

soc

io-e

cono

mic

sta

tus,

eth

nici

ty, d

iabe

tes,

any

hyp

erte

nsiv

e di

seas

e, p

arity

, ge

statio

nal a

ge (d

ays),

fe

tal p

rese

ntat

ion,

birt

h w

eigh

t (gr

ams)

and

cong

enita

l ano

mal

ies.

Chapter 5

86

Propensity Score AnalysisFor 48,921 singleton MAR pregnancies, a total of 146,763 SC women were selected as controls (3 SC patients for each MAR case). The baseline characteristics after propensity score matching are shown in Table 1 and were much more comparable, except for a slight difference in the appearance of preexistent disease. Repeating the matching procedures for only IVF and spontaneous pregnancies resulted in similar baseline results (Supplemental Table 1, available online). After propensity score matching, the risk of poor neonatal outcomes for MAR singletons was similar to the risk among SC singletons, except for the risk of congenital anomalies which was slightly higher among MAR infants: 29 per 1,000 births versus 26 per 1,000 births (aOR 1.11; 95% CI, 1.04–1.18). The results are shown in Table 4. Overall the risk for an Apgar score <4 was 5.3–5.2 per 1,000 births, and the perinatal mortality risk was 1.4–1.3 per 1,000 births. The risk of a cesarean delivery were more comparable after the propensity score matching, although the increased cesarean delivery rate for nonprogressive labor among MAR singletons still remained (OR 1.14; 95% CI, 1.10–1.18). Table 4 also presents the results after propensity score matching for IVF pregnancies. Again neonatal outcomes of IVF singletons were comparable with the neonatal outcomes of SC singletons, but it seems that IVF infants on some outcomes have slightly decreased risks compared with SC. Overall, 21 per 1,000 IVF infants have a poor neonatal outcome compared with 24 per 1,000 SC infants (aOR 0.86; 95% CI, 0.76–0.97). In this matched population the elective cesarean delivery rate was 10% of all pregnancies.

Table 4. Pregnancy and neonatal outcomes among term singletons following MAR or IVF versus spontaneous

conception. Matched for propensity score.

MARN = 48 921

SpontaneousN = 146 763

MAR versus Spontaneous

N (% or ‰) N (% or ‰) Odds ratio (95%CI)

Neonatal outcomes

Apgar score < 4 260 (5.3‰) 759 (5.2‰) 0.99 (0.87 – 1.14)

Apgar score < 7 760 (16‰) 2294 (16‰) 0.99 (0.92 – 1.08)

Intrauterine fetal dead 140 (2.9‰) 389 (2.7‰) 1.08 (0.89 – 1.31)

Perinatal mortality 67 (1.4‰) 184 (1.3‰) 1.09 (0.83 – 1.45)

Congenital anomalies 1420 (29‰) 3852 (26‰) 1.11 (1.04 – 1.18)

SGA P5 2719 (56‰) 8591 (56‰) 0.95 (0.91 – 0.99)

Poor neonatal outcome 1101 (23‰) 3381 (23‰) 0.98 (0.91 – 1.05)

5

87

Pregnancy outcomes

Elective CS 4572 (9.4%) 13 752 (9.4%) 0.99 (0.96 – 1.03)

CS fetal distress 2334 (4.8%) 6546 (4.5%) 1.07 (1.02 – 1.13)

CS non-progressive labor 4680 (9.6%) 12 889 (8.8%) 1.10 (1.06 – 1.14)

IVFN = 16 177

SpontaneousN = 48 531

IVF versus Spontaneous

N (% or ‰) N (% or ‰) Odds ratio (95%CI)

Neonatal outcomes

Apgar score < 4 86 (5.3‰) 278 (5.7‰) 0.93 (0.73 – 1.18)

Apgar score < 7 238 (15‰) 805 (17‰) 0.89 (0.77 – 1.02)

Intrauterine dead 42 (2.6‰) 125 (2.6‰) 1.01 (0.71 – 1.43)

Perinatal mortality 23 (1.4‰) 71 (1.5‰) 0.97 (0.61 – 1.56)

Congenital anomalies 466 (29‰) 1358 (28‰) 1.03 (0.96 – 1.11)

SGA P5 970 (60‰) 2829 (58‰) 1.03 (0.93 – 1.15)

Poor neonatal outcome 344 (21‰) 1173 (24‰) 0.88 (0.78 – 0.99)

Pregnancy outcomes

Elective CS 1682 (10%) 5206 (11%) 0.97 (0.91 – 1.02)

CS fetal distress 806 (5.0%) 2274 (4.7%) 1.07 (0.98 – 1.16)

CS non-progressive labor 1450 (9.0%) 4451 (9.2%) 0.98 (0.92 – 1.04)

Perinatal mortality: intrapartum mortality or mortality within 28 days after live birth.

Poor neonatal outcome is a composite of: Apgar score < 7, intrauterine dead, perinatal mortality, NICU

admission, hypoxic ischemic encephalopathy, convulsions, and respiratory support.

MAR: Medically Assisted Reproduction

IVF: In vitro fertilization

CS: cesarean section

disCussion

In this observational national cohort study of over 45,000 children conceived after MAR, we found that singletons conceived through MAR had higher rates of asphyxia related poor neonatal outcomes than SC singletons. After propensity score matching on maternal characteristics, the risk of adverse neonatal outcomes was similar between MAR infants and SC infants. Cesarean deliveries occurred more often in MAR than in SC pregnancies.

Chapter 5

88

A low 5-minute Apgar score alone as definition for birth asphyxia has been questioned in the literature, and an umbilical cord arterial pH <7.0 was suggested as being more objective as the Apgar score includes subjective components. 24, 25 However, information on the umbilical cord pH is not routinely measured and was therefore not available in the PRN database. We believe that our definition does not hinder our analysis, as in our study group of term infants a 5-minute Apgar score <4 is generally caused by intrauterine asphyxia and research has shown that a low Apgar score is correlated to long-term outcomes. 7, 26 Furthermore, we defined birth asphyxia on the 5-minute Apgar score and clinical diagnosis of asphyxia-related morbidity. Propensity score techniques can balance observed variables but, unlike randomized controlled trials, cannot control for unmeasured variables. The PRN database contains no reliable data on the variables of body mass index or smoking, which can influence both the fertility of women and the perinatal outcome. Nevertheless, the propensity score analysis presented here generated a less biased comparison of outcomes.

Another limitation might be that we used register-based data; in the PRN database, mode of conception is not an obligatory field to complete so there is probably an underreporting of intrauterine insemination, IVF, and other fertility treatments. Another drawback of this register-based data is that we do not have specific information on the different MAR treatments, so we could not distinguish between IVF and intracytoplasmic sperm injection. The extent of the underreporting is not precisely known; it could result in an overestimation of the adverse effects of MAR when MAR is only registered in cases of maternal or neonatal complications, or in an underestimation of the adverse effects in cases of troublesome MAR pregnancies that are not denoted as MAR. Checking the reporting of MAR in the PRN with a clinical annual report tells us that there are around 5% missing in the database. 27 We simulated the effect of MAR registry on the estimate of the OR for a 5-minute Apgar score <4. From 8% missing due to underreporting, there was an effect on the study results. Therefore, we believe that the possibly marginally biased reporting of the estimate of the adverse outcome will not substantially influence our study results.

In concordance with other studies, we found an increased risk of birth asphyxia among MAR infants compared with SC infants. 8, 9, 18, 20, 28 Earlier studies suggested that possible explanations such as associated morbidity in MAR pregnancies (prematurity, intrauterine growth retardation, and congenital anomalies) contribute to the higher rates of (asphyxia-related) morbidity and perinatal mortality. 2, 4, 19 Furthermore, maternal factors associated with subfertility may contribute to the adverse outcomes in MAR pregnancies, rather than the MAR treatments themselves. 2, 8, 9 Without distinguishing between different gestational ages and analysis in unmatched general populations, the



5

89

exact causal pathways are still poorly understood. By using propensity score techniques and excluding prematurely born neonates, we generated a less biased comparison of the outcome measures. Our results after propensity score matching underline the hypothesis that the increased risk of impaired pregnancy and neonatal outcomes of term singletons conceived by MAR is at least partly explained by the maternal and parental characteristics of subfertile couples. However, the remaining increased risk of congenital anomalies in term MAR singletons compared with their SC controls after propensity score matching and the lack of specific information on the different MAR treatments makes it difficult to completely rule out the influence of the (laboratory) procedures performed in fertility treatments by the results of this study. A recent cohort study by Marino et al. 8 demonstrated that neonatal outcomes varied by the type of assisted conception. However, they did not adjust for maternal risk factors, and the data were restricted until 2002. Perinatal outcome in MAR pregnancies improved over time and might reflect the present milder ovarian stimulation, improvements in the laboratory techniques, and better culture media. 3, 29-31 On the other hand, with the expanding use of MAR in Western countries, this could also be attributable to a shift toward less severe subfertility in the couples receiving MAR. 29

Increased cesarean delivery rates in MAR pregnancies have been noted previously. 32-34 We found that the majority of the MAR pregnancies receive specialized obstetric care during labor. The increased intervention rates may be attributed to the knowledge that MAR pregnancies are at increased risk of adverse neonatal outcomes; they thereby receive more monitoring via specialized obstetric care and have a lowered threshold for obstetrician intervention (such as cesarean delivery) in addition to the anxiety surrounding these pregnancies. The latter may be the reason why women who conceive with MAR more oftenreceive specialized obstetric care.

In conclusion, we found that characteristics of subfertile women seem to be associated with the prevalence of health problems in term MAR newborns. With the increasing rates of MAR pregnancies, information concerning obstetric and perinatal outcomes after subfertility treatment is essential. Unfortunately, because it is not possible to identify women at risk for adverse neonatal outcome at term, the situation encourages obstetric caregivers to see every MAR pregnancy as potentially high risk. This contributes to the high rates of obstetric interventions and their potential negative consequences for both mother and child. Obstetric caregivers should be aware that the increased rates of cesarean deliveries reflect the choice of patients and their physicians rather than any inherent biological abnormality in MAR patients. Due to the lack of specific information on the different MAR treatments (intrauterine insemination, IVF, or intracytoplasmic sperm


Chapter 5

90

injection) and the indication for the MAR, we could not consider these as causative factors. Future research must focus on the different MAR treatments and ways for better prediction of high-risk patients for adverse neonatal outcomes at term. This is needed for better understanding of the causality between all the different aspects surrounding MAR and adverse neonatal outcomes so that more individualized care can be provided.



5

91

referenCes

1. Ferraretti AP, Goossens V, Kupka M, et al. Assisted reproductive technology in Europe, 2009: results generated from European registers by ESHRE. Human reproduction 2013;28:2318-31.

2. Pinborg A, Wennerholm UB, Romundstad LB, et al. Why do singletons conceived after assisted reproduction technology have adverse perinatal outcome? Systematic review and meta-analysis. Human reproduction update 2013;19:87-104.

3. Henningsen AK, Pinborg A, Lidegaard O, Vestergaard C, Forman JL, Andersen AN. Perinatal outcome of singleton siblings born after assisted reproductive technology and spontaneous conception: Danish national sibling-cohort study. Fertility and sterility 2011;95:959-63.

4. Helmerhorst FM, Perquin DA, Donker D, Keirse MJ. Perinatal outcome of singletons and twins after assisted conception: a systematic review of controlled studies. Bmj 2004;328:261.

5. Hart R, Norman RJ. The longer-term health outcomes for children born as a result of IVF treatment: Part I--General health outcomes. Human reproduction update 2013;19:232-43.

6. Hart R, Norman RJ. The longer-term health outcomes for children born as a result of IVF treatment. Part II--Mental health and development outcomes. Human reproduction update 2013;19:244-50.


8. Marino JL, Moore VM, Willson KJ, et al. Perinatal outcomes by mode of assisted conception and sub-fertility in an Australian data linkage cohort. PloS one 2014;9:e80398.

9. Hayashi M, Nakai A, Satoh S, Matsuda Y. Adverse obstetric and perinatal outcomes of singleton pregnancies may be related to maternal factors associated with infertility rather than the type of assisted reproductive technology procedure used. Fertility and sterility 2012;98:922-8.

10. Kallen B, Finnstrom O, Nygren KG, Olausson PO. In vitro fertilization (IVF) in Sweden: risk for congenital malformations after different IVF methods. Birth defects research Part A, Clinical and molecular teratology 2005;73:162-9.

11. Silberstein T, Levy A, Harlev A, Saphier O, Sheiner E. Perinatal outcome of pregnancies following in vitro fertilization and ovulation induction. The journal of maternal-fetal & neonatal medicine : the official journal of the European Association of Perinatal Medicine, the Federation of Asia and Oceania Perinatal Societies, the International Society of Perinatal Obstet 2014;27:1316-9.


Chapter 5

92

12. Rosenbaum Pr RD. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41-55.

13. Luo Z, Gardiner JC, Bradley CJ. Applying propensity score methods in medical research: pitfalls and prospects. Medical care research and review : MCRR 2010;67:528-54.

14. Austin PC. An Introduction to Propensity Score Methods for Reducing the Effects of Confounding in Observational Studies. Multivariate behavioral research 2011;46:399-424.

15. Braitman LE, Rosenbaum PR. Rare outcomes, common treatments: analytic strategies using propensity scores. Annals of internal medicine 2002;137:693-5.



18. Raatikainen K, Kuivasaari-Pirinen P, Hippelainen M, Heinonen S. Comparison of the pregnancy outcomes of subfertile women after infertility treatment and in naturally conceived pregnancies. Human reproduction 2012;27:1162-9.

19. Pandey S, Shetty A, Hamilton M, Bhattacharya S, Maheshwari A. Obstetric and perinatal outcomes in singleton pregnancies resulting from IVF/ICSI: a systematic review and meta-analysis. Human reproduction update 2012;18:485-503.

20. Romundstad LB, Romundstad PR, Sunde A, et al. Effects of technology or maternal factors on perinatal outcome after assisted fertilisation: a population-based cohort study. Lancet 2008;372:737-43.

21. Mcnamee R. Confounding and confounders. Occupational and environmental medicine 2003;60:227-34; quiz 164, 234.

22. Rothman KJ, S. G, T.L. L. Modern epidemiology 3rd ed. Philadelphia: Lippincott Williams & Wilkins; Number of pages.

23. Rosenbaum PR. Impact of multiple matched controls on design sensitivity in observational studies. Biometrics 2013;69:118-27.






5

93

27. Steures P, Van Der Steeg JW, Hompes PG, Van Der Veen F, Mol BW. [Results of intrauterine insemination in the Netherlands]. Nederlands tijdschrift voor geneeskunde 2006;150:1127-33.

28. Ombelet W, Martens G, De Sutter P, et al. Perinatal outcome of 12,021 singleton and 3108 twin births after non-IVF-assisted reproduction: a cohort study. Human reproduction 2006;21:1025-32.

29. Kallen B, Finnstrom O, Lindam A, Nilsson E, Nygren KG, Otterblad Olausson P. Trends in delivery and neonatal outcome after in vitro fertilization in Sweden: data for 25 years. Human reproduction 2010;25:1026-34.

30. Sazonova A, Kallen K, Thurin-Kjellberg A, Wennerholm UB, Bergh C. Obstetric outcome after in vitro fertilization with single or double embryo transfer. Human reproduction 2011;26:442-50.

31. Nelissen EC, Van Montfoort AP, Coonen E, et al. Further evidence that culture media affect perinatal outcome: findings after transfer of fresh and cryopreserved embryos. Human reproduction 2012;27:1966-76.

32. Jaques AM, Amor DJ, Baker HW, et al. Adverse obstetric and perinatal outcomes in subfertile women conceiving without assisted reproductive technologies. Fertility and sterility 2010;94:2674-9.

33. Thomson F, Shanbhag S, Templeton A, Bhattacharya S. Obstetric outcome in women with subfertility. BJOG : an international journal of obstetrics and gynaecology 2005;112:632-7.

34. Basso O, Baird DD. Infertility and preterm delivery, birthweight, and Caesarean section: a study within the Danish National Birth Cohort. Human reproduction 2003;18:2478-84.

Chapter 5

94

Supple

men

tal T

able

1.

Base

line

char

acte

ristic

s of

the

elig

ible

stu

dy s

ampl

e of

IVF

singl

eton

s an

d af

ter p

rope

nsity

sco

re m

atch

ing.

Elig

ible

stu

dy s

am

ple

N =

1 9

21

18

8A

fter

pro

pen

sity

sco

re m

atc

hin

gN

= 6

4 7

08

IVF

N =

16

17

7Sp

onta

neo

us

N =

1 9

05

01

1IV

FN

=1

6 1

77

Sponta

neo

us

N =

48 5

31

N (

%)

N (

%)

P-v

alu

eN

(%

)N

(%

)P-v

alu

e

Mea

n ge

statio

nal a

ge *

(wee

ks)

39.4

± 1

.339

.6 ±

1.3

<0.0

001

39.4

± 1

.339

.5 ±

1.4

<0.0

001

Mat

erna

l age

* (y

ears

)34

.1 ±

4.2

30.4

± 4

.9<0

.000

134

.1 ±

4.2

34.1

± 4

.20.

88

Cau

casia

n Et

hnic

ity14

315

(88%

)1

609

264

(84%

)<0

.000

114

315

(88%

)42

707

(88%

)0.

12

Low

soc

io-e

cono

mic

sta

tus

3247

(20%

)47

9 47

7 (2

5%)

<0.0

001

3247

(20%

)97

23 (2

0%)

0.92

Pree

xiste

nt d

iseas

e ^

2975

(18%

)34

706

(1.8

%)

<0.0

001

2975

(18%

)88

56 (1

8%)

0.69

Nul

lipar

ous

10 9

75 (6

8%)

857

971

(45%

)<0

.000

110

975

(68%

)32

925

(68%

)1.

0

Prev

ious

ces

area

n de

liver

y89

6 (5

.5%

)11

7 94

9 (6

.2%

)0.

0001

896

(5.5

%)

2688

(5.5

%)

1.0

Seco

ndar

y ca

re a

t ons

et la

bor

12 7

13 (7

9%)

775

623

(41%

)<0

.000

112

713

(79%

)38

139

(79%

)1.

0

*Pre

sent

ed a

s M

ean

± SD

IVF:

In V

itro

Ferti

lizat

ion.

^Pre

exist

ent d

iseas

e in

clud

es o

ne o

f the

follo

win

g: p

reex

isten

t hyp

erte

nsio

n, p

reex

isten

t dia

bete

s, e

ndoc

rinol

ogic

al d

isord

ers,

coa

gula

tion

diso

rder

s, e

ndom

etrio

sis, u

teru

s m

yom

atos

is an

d co

ngen

ital u

teru

s an

omal

ies.

5

95

Supple

men

tal T

able

1.

Base

line

char

acte

ristic

s of

the

elig

ible

stu

dy s

ampl

e of

IVF

singl

eton

s an

d af

ter p

rope

nsity

sco

re m

atch

ing.

Elig

ible

stu

dy s

am

ple

N =

1 9

21

18

8A

fter

pro

pen

sity

sco

re m

atc

hin

gN

= 6

4 7

08

IVF

N =

16

17

7Sp

onta

neo

us

N =

1 9

05

01

1IV

FN

=1

6 1

77

Sponta

neo

us

N =

48 5

31

N (

%)

N (

%)

P-v

alu

eN

(%

)N

(%

)P-v

alu

e

Mea

n ge

statio

nal a

ge *

(wee

ks)

39.4

± 1

.339

.6 ±

1.3

<0.0

001

39.4

± 1

.339

.5 ±

1.4

<0.0

001

Mat

erna

l age

* (y

ears

)34

.1 ±

4.2

30.4

± 4

.9<0

.000

134

.1 ±

4.2

34.1

± 4

.20.

88

Cau

casia

n Et

hnic

ity14

315

(88%

)1

609

264

(84%

)<0

.000

114

315

(88%

)42

707

(88%

)0.

12

Low

soc

io-e

cono

mic

sta

tus

3247

(20%

)47

9 47

7 (2

5%)

<0.0

001

3247

(20%

)97

23 (2

0%)

0.92

Pree

xiste

nt d

iseas

e ^

2975

(18%

)34

706

(1.8

%)

<0.0

001

2975

(18%

)88

56 (1

8%)

0.69

Nul

lipar

ous

10 9

75 (6

8%)

857

971

(45%

)<0

.000

110

975

(68%

)32

925

(68%

)1.

0

Prev

ious

ces

area

n de

liver

y89

6 (5

.5%

)11

7 94

9 (6

.2%

)0.

0001

896

(5.5

%)

2688

(5.5

%)

1.0

Seco

ndar

y ca

re a

t ons

et la

bor

12 7

13 (7

9%)

775

623

(41%

)<0

.000

112

713

(79%

)38

139

(79%

)1.

0

*Pre

sent

ed a

s M

ean

± SD

IVF:

In V

itro

Ferti

lizat

ion.

^Pre

exist

ent d

iseas

e in

clud

es o

ne o

f the

follo

win

g: p

reex

isten

t hyp

erte

nsio

n, p

reex

isten

t dia

bete

s, e

ndoc

rinol

ogic

al d

isord

ers,

coa

gula

tion

diso

rder

s, e

ndom

etrio

sis, u

teru

s m

yom

atos

is an

d co

ngen

ital u

teru

s an

omal

ies.


Chapter 6 Maternal, neonatal and care risk factors for asphyxia related perinatal mortality at term

S. Ensing F. Groenendaal

M. Eskes A. Abu – Hanna

B.W.J. Mol A.C.J. Ravelli



Chapter 6

98

absTraCT

ObjectiveTo assess maternal, neonatal and care characteristics associated with asphyxia-related mortality at term.

DesignWe performed a case-control study. Relative risks as well as Population Attributable Risk (PAR) percentages were calculated.

SettingWe compared babies that died from asphyxia related causes in the perinatal period selected from the database of the Dutch Perinatal Audit Foundation (PAN), and compared them to live born singleton infants born selected from the Dutch Perinatal Registry (PRN).

ParticipantsTerm singletons without congenital anomalies born between 2010 and 2012 in the Netherlands.

Main Outcome MeasuresThe primary outcome was asphyxia related mortality which included intrauterine fetal death at term, intrapartum death or neonatal death within 28 days after live birth.

ResultsWe compared 214 infants who died from asphyxia at term with 451,066 infants born alive. Mothers from asphyctic infants that died were more often smoking (RR 4.1, 95%CI 2.6 to 4.4), conceived true medically assisted reproduction (RR 3.1, 95%CI 1.9 to 4.4) and of non – western ethnicity (2.1, 95%CI 1.5 to 2. 6). Their infants were more frequently SGA (RR 5.8, 95%CI 3.7 to 6.3). The highest PAR was found for SGA below 10th percentile (17%), nulliparity (20%) and smoking (16%). Also, prelabor classification of high-risk pregnancy based on medical features had a high PAR (39%). In women who started labor in a primary care setting, the PAR of SGA below the 10th percentile was 24% as compared to 13% for women who started labor in secondary care.

ConclusionNulliparity, non-western ethnicity, smoking, being SGA and medical conditions indicating high-risk pregnancy are the most important risk factors for asphyxia related mortality at term. The high risk of being SGA in general and its increased risk in women with otherwise low risk pregnancies indicate the potential improvement that adequate detection of SGA can bring to reduce asphyctic related death.


Asphyxia related mortality at term

6

99

inTroduCTion

Perinatal asphyxia plays a major role in perinatal mortality at term 1-5. Mortality of term neonates is much lower than that of preterm neonates, but it still accounts for 25% of the total number of perinatal deaths in the Netherlands. 6 In 2010, the most recent Peristat Report showed that perinatal mortality rates in the Netherlands are among the highest in Europe. 7, 8 This resulted in the implementation of a nationwide perinatal audit. A multidisciplinary team of healthcare professionals audits term mortality cases on local level, in order to improve quality of obstetric care and decrease perinatal mortality rates. 9 In this audit, within 2 years after start all hospitals and their surrounding midwife practices contribute. 10

There are known maternal and fetal risk factors for perinatal asphyxia such as advanced maternal age, nulliparity and growth restriction. 11-13 Evers et al. pointed also at the association between asphyxia related mortality and morbidity and substandard care factors. 14, 15 In The Netherlands women with low risk for pregnancy complications receive care in a separate midwifery led care system, whereas high risk women (based on medical features) receive care in an obstetric led care system. Thus, optimal antepartum risk assessment is of crucial importance. 16, 17 On one hand, low risk women laboring in a high-risk system are at risk for unnecessary interventions, while high-risk women in a low-risk system are at risk of underdiagnoses and poor outcomes as a consequence of that. This dilemma also occurs in other health systems. We aimed to assess maternal and neonatal characteristics and care factors to explore possible risk assessment of asphyxia related mortality at term. Better risk assessment of factors associated with asphyxia related deaths could contribute to improvement of the identification of low-risk and high-risk pregnant women.

meThods

DatabasesFor this case-control study we used two databases. First, we used the database of the Dutch Perinatal Audit Foundation (PAN) to select perinatal mortality cases at term. For each case of perinatal mortality a systematically audit is performed at the local level. Data on the perinatal mortality cases are gathered by the health care professionals form the medical records and registered with specific details regarding health care processes in the PAN chronological database. The PAN database consists of >86% of all term perinatal mortality cases registered by healthcare professionals. (9, 10 From the PAN database we selected all singleton pregnancies between 37+0 and 41+6 weeks of gestation of neonates who died in the perinatal period due to asphyxia related causes between 2010 and 2012. Pregnancies with congenital anomalies

Chapter 6

100

were excluded. Second, we used the database of the Dutch Perinatal Registry (PRN) to select our control group of live born infants. The PRN database contains population based information on all pregnancies, deliveries from 22 weeks onwards and (re)admissions until 28 days after birth. The PRN registry is the result of a validated linkage of three registries (midwifery registry, obstetrics registry and neonatology registry) and covers around 98% of all births in the Netherlands. 18-20 From the PRN database we selected all singletons born between 37+0 and 41+6 weeks of gestation who were alive 28 days after live birth. Infants with a 5 minute Apgar score < 7 or with any admission to the Neonatal Care Unit (NCU) in the post-partum period were excluded in order to get a healthy control group.

Outcome measuresThe primary outcome was asphyxia related mortality. In the PAN database the cause of death is assigned by the perinatal audit team. First the classification of Gardosi et al. is used to assign the dominant clinical condition at time of death.21, 22 Furthermore the underlying pathophysiology of the clinical condition causing death is defined following the Tulip classification.23 For example, a neonate suffering intra-uterine Group- B streptococcus (GBS) infection who died could be diagnosed with asphyxia as the dominant clinical condition (and therefore included in our study cohort) which was caused by the GBS infection as the underlying pathophysiology. Perinatal mortality included intrauterine fetal death at term, intrapartum death or neonatal death within 28 days after live birth.

Risk factorsFor all term singletons we compared demographic and obstetric baseline characteristics which were recorded in the PRN and PAN database, i.e. maternal age, gestational age, ethnicity, socio-economic status (SES), use of medically assisted reproduction technologies (MAR), any hypertensive disease (pre-existing and pregnancy induced), fetal gender, birth weight below the 10th,5th and 2.3th percentile (SGA), birth weight above the 95th percentile (LGA), mode of delivery and induction of labor. Ethnicity was ascribed by the obstetric care giver and we differentiated between Western (native Dutch women and women from other Western nations) and non- Western (including different ethnicities as African/Surinamese Creole, Hindustani, Mediterranean and Asian).The continuous socioeconomic status score, which is based on mean income level, was categorized into a high/middle and low group based on percentiles. SGA and LGA were defined by the Dutch reference curves for birth weight by gestational age, parity, sex and ethnicity. 24 Information on risk factors as body mass index (BMI) and smoking was only available in the PAN database. For analyses with these potential risk factors we incorporate incidence rates from the Statistics Netherlands (Centraal Bureau voor Statistiek, CBS). 25, 26 The CBS contains information on the



6

101

prevalence of increased BMI and smoking among fertile women. By incorporating the prevalence numbers to the control group selected from the PRN we were able to calculate Relative Risks. In the Netherlands, there is a set of criteria that is used for classification between low- and high-risk pregnancies. 27 For example, women suffering from any disease as hypertension, diabetes, coagulation disorders but also those women who delivered previous by caesarean section are classified as high risk. Women with prelabor classified as having a low risk pregnancy received primary care under the responsibility of midwives or general practitioners. Women with prelabor classified as having a high risk pregnancy received secondary care, i.e. obstetric led care in the hospital at onset of labor.

StatisticsUnivariate analysis of the baseline characteristics was performed with the Student’s t-test for normally distributed continuous variables and the chi-square test for categorical variables. The association between the maternal, pregnancy and fetal related risk indicators and asphyxia related perinatal mortality was classified in two by two tables. Subsequently, relative risks and 95% confidence intervals were calculated. Multivariate analyses could not be performed as we used incorporated prevalence numbers for smoking and BMI. In addition, for each risk indicator, we calculated the population attributable risk (PAR) percentage, which is based on the Relative Risk and Prevalence of a certain risk indicator:

PAR % = [prevalence x (RR-1)] / [prevalence x (RR-1)+1] x 100 28

Combined PAR were calculated: 1-(1-PAR1)*(1-PAR2)

We also assessed the interaction between (a) all the significant risk indicators and (b) level of care at onset of labor (i.e. prelabor classified high or low risk pregnancy) on (c) perinatal mortality, while adjusting for (a) and (b). For all statistical tests, we used a p-value of 0.05 as the threshold to indicate statistical significance. All statistical analyses were performed with SAS 9.3 (SAS Institute, Cary, NC, USA).

resulTs

We studied 214 (32%) term women who suffered perinatal deaths caused by asphyxia. They were selected from 679 term infants that died between 2010 and 2012. We compared these pregnancies with 451,066 term pregnancies of infants who were alive 28 days after birth and not admitted to a NICU. Baseline characteristics of the two groups are presented in table 1. Mothers from infants who died due asphyxia in the perinatal period were likely to be older, more often nulliparous and more often of non-western ethnicity. Also, smoking and conception due medically assisted reproduction

Chapter 6

102

were more observed among women with perinatal asphyxia related mortality. Their infants were more often delivered by 37 weeks of gestation and frequently small for gestational age (SGA).

Table 1. Baseline characteristics all live born and asphyxia related perinatal mortality at term in the

Netherlands 2010 – 2012.

Asphyxia Related MortalityN = 214

Live-born

N = 451 066 p-valueMaternal ageMean (years)≥ 35 years

31.3 (±4.9)53 (25%)

30.8 (± 4.9)85 512 (19%) P = 0.03

Non - Western ethnicity 44 (21%) 47 468 (10%) P < 0.0001

Low socio-economic status 47 (22%) 118 625 (26%) P = 0.15

Nulliparous 119 (56%) 201 801 (45%) P = 0.001

Gestational ageMean (days)37+0 – 37+6 weeks38+0 – 38+6 weeks39+0 – 39+6 weeks40+0 – 40+6 weeks41+0 – 41+6 weeks

279 (±9.3)26 (12%)29 (14%)47 (22%)62 (29%)50 (23%)

279 (±8.1)32 256 (7.2%)73 655 (16%)121 132 (27%)137 789 (31%)86 234 (19%)

P = 0.005NSNSNSNS

Level of care at onset of laborPrimary care; prelabor classified low risk pregnancySecondary care; prelabor classified high risk pregnancy

71 (33%)

143 (69%)

246 682 (55%)

204 384 (45%) P= 0

Place of birthHomeHospital midwife ledHospital gynecologist led

11 (5.1%)9 (4.2%)194 (91%)

77 007 (17%)63 972 (14%)310 087 (69%)

P < 0.0001P < 0.0001P = 0.03

Induction of labor 78 (36%) 91 600 (20%) P = 0.001

Mode of deliverySpontaneous vaginalPlanned caesareanAssisted vaginalUnplanned caesarean

118 (55%)13 (6.1%)29 (14%)54 (25%)

342 124 (76%)28 250 (6.3%)44 004 (9.8%)36 688 (8.1%)

P = 0.02NSNSP = 0.01

Primary care: prelabor classified low risk pregnancies, midwifery care at onset of labor

Secondary care: prelabor classified high risk pregnancies, specialized obstetric care at onset of labor

6

103

The majority of term infants was born by spontaneous vaginal delivery (76%) under supervision of gynecologists’ (69%) and there were more inductions of labor in the asphyxia mortality group. In the control group, 55% started labor in primary care compared to 33% of all women who suffered perinatal mortality. From the 214 asphyxia related perinatal mortality cases, 37% (n=79) died before the onset of labor, 13% (n=27) were intrapartum deaths and 108 babies (50%) died in the first 28 days after live birth. Placental problems were the underlying pathophysiology leading to asphyxia in 88 (41%) pregnancies. Of the 88 cases with placental problems, 40 fetuses (45%) died in utero. Other reasons for asphyxia related mortality were birth trauma (n =35, 16%) or intra-uterine infection (n = 9, 4.2%). In 82 babies who died (38%) the underlying cause leading to asphyxia was unknown.

In our study cohort smoking (RR 4.06, 95%CI 2.64 – 4.41), medically assisted reproduction (RR 3.11, 95%CI 1.93 to4.36) and non-western ethnicity (RR 2.12, 95%CI 1.51 – 2.56) were the strongest risk indicators for asphyxia related mortality at term (Table 2). Prelabor classified high-risk pregnancies (PAR 39%), SGAP10 (birth weight < 10th percentile), which PAR was 17% and smoking (PAR 16%), contributed most to the increased risk of asphyxia related mortality at term. The combined PAR of smoking and SGAP10 was 30%.

As we found a significant interaction between small for gestational age (SGA) and level of care at onset of labor, we performed a stratified analyses for the association between SGA and mortality according to the level of care at onset of labor. The results are presented in Table 3. For infants born from low risk mothers, being small for gestational age was the far most important risk factor (RR 9.6, 95%CI 5.0 to 13). The PAR of SGA below the 10th percentile was 24% for women otherwise classified as low risk versus 13% for women with other risk factors (high risk) women. A similar pattern was seen for SGA below the 5th and 2.3th percentile. Of all 45 babies with a birth weight below the 10th percentile who died due asphyxia related causes, only 5 (11%) were suspected before birth. All 5 fetuses with suspected intra-uterine growth retardation were from high risk mothers.

Chapter 6

104

Table

2.

Cha

ract

erist

ics,

Rel

ativ

e Ri

sk’s

(RR)

and

Pop

ulat

ion

Attr

ibut

ive

Risk

(PA

R) o

f asp

hyxi

a re

late

d m

orta

lity

at te

rm.

Asp

hyx

ia r

elate

d m

ort

alit

yN

= 2

14

Live

born

N =

45

1 0

66

RR (

95

%CI)

PAR %

Mate

rnal

Mat

erna

l age

≥ 3

5 ye

ars

53 (2

5%)

85 5

12 (1

9%)

1.41

(1.0

3 –

1.65

)7.

2%N

on- W

este

rn e

thni

city

44 (2

1%)

4717

1 (1

0%)

2.12

(1.5

1 –

2.56

)11

%N

ullip

arou

s11

9 (5

6%)

201

801

(45%

)1.

55 (1

.10

– 1.

40)

20%

BMI ≥

30

23 (1

1%)

45 1

07 (1

0%)

1.08

(0.7

3 –

1.58

)0.

83%

Smok

ing

46 (2

2%)

28 4

17 (6

.3%

)4.

06 (2

.64

– 4.

41)

16%

Hyp

erte

nsiv

e di

seas

e16

(7.5

%)

37 7

95 (8

.4%

)0.

88 (0

.56

– 1.

43)

NA

Med

ical

ly a

ssist

ed re

prod

uctio

n21

(9.8

%)

15 2

42 (3

.4%

)3.

11 (1

.93

– 4.

36)

6.7%

Seco

ndar

y ca

re a

t ons

et o

f lab

or *

143

(69%

)20

4 38

4 (4

5%)

2.43

(1.3

4 –

2.92

)39

%N

eonate

Fem

ale

feta

l gen

der

111

(52%

)22

2 29

7 (4

9%)

1.11

(0.9

3 –

1.20

)5.

1%Br

eech

pre

sent

atio

n11

(5.1

%)

14 3

34 (3

.2%

)1.

65 (0

.91

– 2.

88)

2.0%

Larg

e fo

r ges

tatio

nal a

ge >

P95

9 (4

.2%

)24

559

(5.4

%)

0.76

(0.4

1 –

1.46

)N

ASm

all f

or g

esta

tiona

l age

SGA

P10

45 (2

1%)

19 6

81 (4

.5%

)5.

83 (3

.72

– 6.

25)

17%

SGA

P532

(15%

)18

297

(4.1

%)

4.15

(2.6

8 –

5.08

)11

%SG

AP2

.320

(9%

)45

10 (1

.0%

)10

.2 (6

.15

– 14

.2)

8.4%

BMI:

Body

Mas

s In

dex

Med

ical

ly a

ssist

ed c

once

ptio

n co

ntai

ns: I

ntra

-ute

rine

inse

min

atio

n, In

vitr

o fe

rtiliz

atio

n or

intra

cyto

plas

mic

spe

rm in

ject

ion

Seco

ndar

y ca

re: p

rela

bor c

lass

ified

hig

h ris

k pr

egna

ncie

s, s

peci

aliz

ed o

bste

tric

care

at o

nset

of l

abor

SGA

P10:

Sm

all G

esta

tiona

l age

(birt

h w

eigh

t < 1

0th

perc

entil

e)SG

AP5

: Sm

all G

esta

tiona

l age

(birt

h w

eigh

t < 5

th p

erce

ntile

)SG

AP2

.3: S

mal

l Ges

tatio

nal a

ge (b

irth

wei

ght <

2.3

th p

erce

ntile

)

6

105

Table

2.

Cha

ract

erist

ics,

Rel

ativ

e Ri

sk’s

(RR)

and

Pop

ulat

ion

Attr

ibut

ive

Risk

(PA

R) o

f asp

hyxi

a re

late

d m

orta

lity

at te

rm.

Asp

hyx

ia r

elate

d m

ort

alit

yN

= 2

14

Live

born

N =

45

1 0

66

RR (

95

%CI)

PAR %

Mate

rnal

Mat

erna

l age

≥ 3

5 ye

ars

53 (2

5%)

85 5

12 (1

9%)

1.41

(1.0

3 –

1.65

)7.

2%N

on- W

este

rn e

thni

city

44 (2

1%)

4717

1 (1

0%)

2.12

(1.5

1 –

2.56

)11

%N

ullip

arou

s11

9 (5

6%)

201

801

(45%

)1.

55 (1

.10

– 1.

40)

20%

BMI ≥

30

23 (1

1%)

45 1

07 (1

0%)

1.08

(0.7

3 –

1.58

)0.

83%

Smok

ing

46 (2

2%)

28 4

17 (6

.3%

)4.

06 (2

.64

– 4.

41)

16%

Hyp

erte

nsiv

e di

seas

e16

(7.5

%)

37 7

95 (8

.4%

)0.

88 (0

.56

– 1.

43)

NA

Med

ical

ly a

ssist

ed re

prod

uctio

n21

(9.8

%)

15 2

42 (3

.4%

)3.

11 (1

.93

– 4.

36)

6.7%

Seco

ndar

y ca

re a

t ons

et o

f lab

or *

143

(69%

)20

4 38

4 (4

5%)

2.43

(1.3

4 –

2.92

)39

%N

eonate

Fem

ale

feta

l gen

der

111

(52%

)22

2 29

7 (4

9%)

1.11

(0.9

3 –

1.20

)5.

1%Br

eech

pre

sent

atio

n11

(5.1

%)

14 3

34 (3

.2%

)1.

65 (0

.91

– 2.

88)

2.0%

Larg

e fo

r ges

tatio

nal a

ge >

P95

9 (4

.2%

)24

559

(5.4

%)

0.76

(0.4

1 –

1.46

)N

ASm

all f

or g

esta

tiona

l age

SGA

P10

45 (2

1%)

19 6

81 (4

.5%

)5.

83 (3

.72

– 6.

25)

17%

SGA

P532

(15%

)18

297

(4.1

%)

4.15

(2.6

8 –

5.08

)11

%SG

AP2

.320

(9%

)45

10 (1

.0%

)10

.2 (6

.15

– 14

.2)

8.4%

BMI:

Body

Mas

s In

dex

Med

ical

ly a

ssist

ed c

once

ptio

n co

ntai

ns: I

ntra

-ute

rine

inse

min

atio

n, In

vitr

o fe

rtiliz

atio

n or

intra

cyto

plas

mic

spe

rm in

ject

ion

Seco

ndar

y ca

re: p

rela

bor c

lass

ified

hig

h ris

k pr

egna

ncie

s, s

peci

aliz

ed o

bste

tric

care

at o

nset

of l

abor

SGA

P10:

Sm

all G

esta

tiona

l age

(birt

h w

eigh

t < 1

0th

perc

entil

e)SG

AP5

: Sm

all G

esta

tiona

l age

(birt

h w

eigh

t < 5

th p

erce

ntile

)SG

AP2

.3: S

mal

l Ges

tatio

nal a

ge (b

irth

wei

ght <

2.3

th p

erce

ntile

)

Table

3.

The

asso

ciat

ion

betw

een

smal

l for

ges

tatio

nal a

ge a

nd th

e ris

k of

asp

hyxi

a re

late

d m

orta

lity

at te

rm, s

tratifi

ed fo

r lev

el o

f car

e at

ons

et o

f lab

or.

Pre

labor

class

ified

low

ris

k p

regnancy

Asp

hyx

ia r

elate

d m

ort

alit

yN

= 7

4Li

ve b

orn

N =

24

6 6

82

RR (

95

%CI)

PAR %

Small

for

ges

tational a

ge

SGA

P10

20 (2

7%)

9127

(3.7

%)

9.62

(5.0

2 –

12.6

)24

%

SGA

P516

(22%

)74

97 (3

.0%

)8.

78 (4

.60

– 11

.0)

19%

SGA

P2.3

10 (1

4%)

1762

(0.7

1%)

21.6

(10.

6 –

33.7

)13

%

Pre

labor

class

ified

hig

h r

isk

pre

gnancy

Asp

hyx

ia r

elate

d m

ort

alit

yN

= 1

40

Live

born

N =

204

384

RR (

95

%CI)

PAR %

Small

for

ges

tational a

ge

SGA

P10

25 (1

8%)

10 5

54 (5

.2%

)3.

98 (2

.42

– 4.

94)

13%

SGA

P516

(11%

)10

800

(5.3

%)

2.31

(1.3

6 –

3.43

)6.

5%

SGA

P2.3

10 (7

.1%

)27

48 (1

.3%

)5.

63 (2

.92

– 9.

66)

5.9%

SGA

P10:

Sm

all G

esta

tiona

l age

(birt

h w

eigh

t < 1

0th

perc

entil

e)SG

AP5

: Sm

all G

esta

tiona

l age

(birt

h w

eigh

t < 5

th p

erce

ntile

)SG

AP2

.3: S

mal

l Ges

tatio

nal a

ge (b

irth

wei

ght <

2.3

th p

erce

ntile

)


Chapter 6

106

disCussion

In this case-control study, we found nulliparity, non-western ethnicity, smoking, being SGA and prelabor classification of high risk pregnancy based on medical features, to be the most important risk factors for asphyxia related mortality at term. The contribution of SGA, particularly in women who started labor in primary care show the potential improvement of adequate antenatal detection of SGA to limit the chance for asphyxia related mortality at term.

Strengths and limitationsOur study was based on data from two different large databases. The PAN database is primarily developed for the audit of perinatal mortality in The Netherlands, and therefore very detailed and specific. After an extended audit of all factors contributing to the death of the baby, data are filled by health care professionals using the Tulip classification to define the underlying cause of death. 21-23 In the PRN database the mortality cases were also available however the amount of detailed information on these records in the PAN database is much more accurate and detailed, so we decided to combine these two data resources. The PRN database is a population-based database, missing data on important risk factors such as smoking and body mass index (BMI). Incorporating population prevalence numbers of BMI and smoking to the healthy controls selected from the PRN might have led to overestimation of the number smoking pregnant women, as it is known that women quit smoking before or during pregnancy. Furthermore, trend analysis of smoking habits of pregnant women showed a decreased prevalence from 2001 onwards and the 6.3% used in our study was the prevalence of the most recent year; 2010. This possible overestimation in the control group implicates that the relative risk and PAR of smoking would even be more increased. Therefore, we believe that it will not substantially influence our conclusions. Also, due to missing information on smoking in the PRN database we could not test for interaction between smoking en SGA. Following the mortality cases, 36% of the mothers with a SGA baby who died smoked during pregnancy, suggesting there might be an interaction between these risk indicators.We included the mortality cases from the PAN database and did not use the cases from the national PRN registration. The strength is that we had more detailed information in the PAN database and all the information was verified. A limitation might be that not all term mortality cases were included in the PAN cohort, however the characteristics were in general not different between the included and non-included cases. 9 Another strength is that we are not only presenting the associations with confidence intervals between identified risk factors and asphyxia related mortality. The addition of population attributable risk percentages (PAR) contributes to the understanding of the potential improvement that could be reached when eliminating a certain risk factor.



6

107

Attention for characteristics which have a high individual risk that are actually very scare will contribute less to better health outcome on the population level than those with a high prevalence in population. 28

Other literatureExisting literature on maternal and fetal risk factors for perinatal mortality is mainly non-cause specific, not limited to term neonates or studies performed on small databases. 1, 5 With the unique PAN database, including 214 term infants who died due to asphyxia related causes, we were able to assess asphyxia related perinatal mortality in these children more in detail. In agreement with previous literature, we found that being SGA was the most important fetal risk factor for perinatal mortality at term. 1, 5, 13,

29 Our results underline that the early detection of SGA infants could improve perinatal outcome, as we specifically found an increased risk for perinatal mortality among SGA infants who start labor in primary care.

In primary care settings, fetal surveillance is done intermittent, and sometimes not from the start of labor, thus indicating that prelabor diagnosis of SGA could improve neonatal outcomes. Recent research by Kazemier et al. showed that for SGA neonates delivery around 38 weeks of gestation generates optimal outcomes, thus suggesting induction of labor in those women. 30 However, before introducing such clinical guidelines, the prenatal detection of SGA must be improved. In our study, only 11% of the SGA babies had been diagnosed as such before birth. Literature on the accuracy of detecting intrauterine growth restriction (IUGR) is inconclusive. A population based study in France showed that antenatal suspicion of IUGR among SGA infants was low and one-half of infants suspected IUGR were not SGA. 31 Whereas others found that improved training of obstetric caregivers has led to increased antenatal detection of IUGR, and this in turn has resulted in significant reductions in stillbirths. 32 Maternal smoking is associated with both increased perinatal mortality and long-term morbidity. The combined PAR of smoking and SGA of 30% shows that around one third of all asphyxia related mortality is associated with these indicators. Smoking not only directly increases the risk for perinatal mortality, it also works as a mediator due increased risk for intrauterine growth retardation among smoking women. 33, 34 Non – western ethnicity was previously described as a risk factor for adverse neonatal outcomes at term. 35 We confirmed this association for asphyxia related mortality, however like the other study group, we cannot distinguish whether the observed differences can be explained by genetic differences or by (un)known other confounding factors. The same problem of confounding factors applies to medically assisted reproduction (MAR) which is shown to be an important risk factor for asphyxia related mortality. A difficulty here is to assess whether this risk is due to the MAR treatment, or due to the underlying infertility. We previously showed in an observational cohort study among over 45,000 children


Chapter 6

108

conceived after MAR using propensity score analysis that characteristics of women receiving MAR contribute to the increased of health problems in MAR newborns. 36

Risk of severe asphyxia-related morbidity was found to increase with maternal BMI in a Swedish population of term infants. 37 We did not find that increased BMI was a risk factor for asphyxia related mortality. Increased risks of asphyxia related morbidity, as low Apgar score, may be caused by traumatic labor complications like shoulder dystocia which are seen more frequently in obese women. Fortunately this asphyxia is transitory and mostly not leading to neonatal death. Being identified as high-risk prior to start of labor, and thus starting delivery in secondary care, was the most important ‘risk factor’ that we identified. Starting delivery in secondary care obviously is a decision based on multiple other features, including (pregnancy induced) hypertension, (gestational) diabetes, previous preterm birth or underlying maternal disease.

Clinical implicationsPerinatal asphyxia is along with prematurity, congenital anomalies and intrauterine growth retardation one of the four major factors that contributes adverse neonatal outcome. 2 To improve these poor outcomes, risk estimation and individual prediction of adverse health outcomes is the needed to allow targeted interventions. Before one can apply individualized risk estimation for a certain health outcome such asphyxia related mortality, first the association between potential risk factors and the proposed outcome needs to be understand completely. Results of our study contribute to the understanding of asphyxia related mortality at term, thus allowing more targeted intervention, bot form the perspective of the mechanism leading to asphyxia, as well from the perspective of risk selection and antenatal identification of low-risk and high-risk pregnant women. Identification of high-risk women with improved detection of SGA may lower perinatal mortality rates, however if every pregnant women receives routine ultrasound screening this leads to great rise in health care costs. So before this implementation the clinical - and cost-effectiveness should be assessed.

In conclusion, one third of the term perinatal mortality cases is caused by perinatal asphyxia. Since being SGA and maternal smoking are the most important risk factors for asphyxia related mortality, improved detection of SGA fetuses and reduction of the number of pregnant women smoking might reduce perinatal mortality at term.



6

109

referenCes


2. Van Der Kooy J, Poeran J, De Graaf JP, et al. Planned home compared with planned hospital births in the Netherlands: intrapartum and early neonatal death in low-risk pregnancies. Obstetrics and gynecology 2011;118:1037-46.

3. Evers AC, Brouwers HA, Hukkelhoven CW, et al. Perinatal mortality and severe morbidity in low and high risk term pregnancies in the Netherlands: prospective cohort study. Bmj 2010;341:c5639.

4. Gonzalez FF, Miller SP. Does perinatal asphyxia impair cognitive function without cerebral palsy? Archives of disease in childhood Fetal and neonatal edition 2006;91:F454-9.

5. Winbo I, Serenius F, Dahlquist G, Kallen B. Maternal risk factors for cause-specific stillbirth and neonatal death. Acta obstetricia et gynecologica Scandinavica 2001;80:235-44.

6. Ravelli AC, Eskes M, Tromp M, et al. [Perinatal mortality in The Netherlands 2000-2006; risk factors and risk selection]. Nederlands tijdschrift voor geneeskunde 2008;152:2728-33.


8. Mohangoo AD, Buitendijk SE, Szamotulska K, et al. Gestational age patterns of fetal and neonatal mortality in Europe: results from the Euro-Peristat project. PloS one 2011;6:e24727.

9. De Reu P, Van Diem M, Eskes M, et al. The Dutch Perinatal Audit Project: a feasibility study for nationwide perinatal audit in the Netherlands. Acta obstetricia et gynecologica Scandinavica 2009;88:1201-8.

10. Eskes M, Waelput AJ, Erwich JJ, et al. Term perinatal mortality audit in the Netherlands 2010-2012: a population-based cohort study. BMJ open 2014;4:e005652.

11. Miltenburg AS, Van Elburg RM, Kostense PJ, Van Geijn HP, Bolte AC. Neonatal morbidity in term neonates is related to gestational age at birth and level of care. Journal of perinatal medicine 2011;39:605-10.

12. Fretts RC, Schmittdiel J, Mclean FH, Usher RH, Goldman MB. Increased maternal age and the risk of fetal death. The New England journal of medicine 1995;333:953-7.

13. Leijon I. The prognostic significance of antenatal diagnosis of fetal growth retardation. International journal of technology assessment in health care 1992;8 Suppl 1:176-81.


Chapter 6

110

14. Evers AC, Brouwers HA, Nikkels PG, et al. Substandard care in delivery-related asphyxia among term infants: prospective cohort study. Acta obstetricia et gynecologica Scandinavica 2013;92:85-93.

15. Evers AC, Nikkels PG, Brouwers HA, et al. Substandard care in antepartum term stillbirths: prospective cohort study. Acta obstetricia et gynecologica Scandinavica 2011;90:1416-22.

16. Van Alten D, Eskes M, Treffers PE. Midwifery in The Netherlands. The Wormerveer study; selection, mode of delivery, perinatal mortality and infant morbidity. British journal of obstetrics and gynaecology 1989;96:656-62.

17. Amelink-Verburg MP, Verloove-Vanhorick SP, Hakkenberg RM, Veldhuijzen IM, Bennebroek Gravenhorst J, Buitendijk SE. Evaluation of 280,000 cases in Dutch midwifery practices: a descriptive study. BJOG : an international journal of obstetrics and gynaecology 2008;115:570-8.




21. Gardosi J, Kady SM, Mcgeown P, Francis A, Tonks A. Classification of stillbirth by relevant condition at death (ReCoDe): population based cohort study. Bmj 2005;331:1113-7.

22. Chan A, King JF, Flenady V, Haslam RH, Tudehope DI. Classification of perinatal deaths: development of the Australian and New Zealand classifications. Journal of paediatrics and child health 2004;40:340-7.

23. Korteweg FJ, Gordijn SJ, Timmer A, et al. The Tulip classification of perinatal mortality: introduction and multidisciplinary inter-rater agreement. BJOG : an international journal of obstetrics and gynaecology 2006;113:393-401.


25. http://www.cbs.nl/nl-NL/menu/themas/gezondheid-welzijn, 2012.

26. https://www.volksgezondheidenzorg.info/onderwerp/overgewicht/cijfers-context/huidige-situatie#node-overgewicht-bij-volwassenen, 2012.



6

111

27. Bleker Op HVL, Eskes M, Bonsel Gj. Place of birth: evidence for best practice. Recent advances in Obstetrics and Gynaecology. London: Royal Society of Medicine Press Ltd, 2005.

28. Miettinen OS. Proportion of disease caused or prevented by a given exposure, trait or intervention. American journal of epidemiology 1974;99:325-32.

29. Soudee S, Vuillemin L, Alberti C, et al. Fetal growth restriction is worse than extreme prematurity for the developing lung. Neonatology 2014;106:304-10.

30. Kazemier BM, Voskamp BJ, Ravelli AC, Pajkrt E, Groot CJ, Mol BW. Optimal timing of delivery in small for gestational age fetuses near term: a national cohort study. American journal of perinatology 2015;30:177-86.

31. Monier I, Blondel B, Ego A, Kaminiski M, Goffinet F, Zeitlin J. Poor effectiveness of antenatal detection of fetal growth restriction and consequences for obstetric management and neonatal outcomes: a French national study. BJOG : an international journal of obstetrics and gynaecology 2015;122:518-27.

32. Gardosi J, Giddings S, Buller S, Southam M, Williams M. Preventing stillbirths through improved antenatal recognition of pregnancies at risk due to fetal growth restriction. Public health 2014;128:698-702.

33. Mund M, Louwen F, Klingelhoefer D, Gerber A. Smoking and pregnancy--a review on the first major environmental risk factor of the unborn. International journal of environmental research and public health 2013;10:6485-99.

34. Been JV, Mackay DF, Millett C, Pell JP, Van Schayck OC, Sheikh A. Impact of smoke-free legislation on perinatal and infant mortality: a national quasi-experimental study. Scientific reports 2015;5:13020.

35. Kazemier BM, Ravelli AC, De Groot CJ, Mol BW. Optimal timing of near-term delivery in different ethnicities: a national cohort study. BJOG : an international journal of obstetrics and gynaecology 2014;121:1274-82; discussion 83.

36. Ensing S, Abu-Hanna A, Roseboom TJ, et al. Risk of poor neonatal outcome at term after medically assisted reproduction: a propensity score-matched study. Fertility and sterility 2015;104:384-90 e1.

37. Persson M, Johansson S, Villamor E, Cnattingius S. Maternal overweight and obesity and risks of severe birth-asphyxia-related complications in term infants: a population-based cohort study in Sweden. PLoS medicine 2014;11:e1001648.


Part III:Prognostic modelling of adverse

pregnancy outcomes at term


Chapter 7 Predictive capacity of the five individual

components of the 5 minute Apgar score on neonatal mortality and morbidity in term infants

S. EnsingM.C. EngelenP. Tamminga

A.C.J. RavelliB.W.J. Mol

A. Abu-Hanna


Chapter 7

116

absTraCT

ObjectiveTo assess the predictive capacity of the five individual components of the 5-minute Apgar score on neonatal mortality and morbidity outcomes.

Study designWe performed a retrospective cohort study of term singletons born with a 5-minute Apgar score <7 submitted to the Neonatal Intensive Care Unit (NICU) between 2000-2010. Logistic regression and penalized variable selection were used to assess the association of the five Apgar score’s components to (1) neonatal mortality within 28 days and (2) combined poor neonatal outcome (neonatal mortality or morbidity defined as clinical asphyxia, hypoxic ischemic encephalopathy, respiratory problems, meconium aspiration syndrome, infection or convulsions). In addition we developed a reduced model based on the most predictive components, and compared its AUC to that of the Apgar-based model.

ResultsWe studied 276 term neonates with an Apgar score <7, of whom 55 (20%) died within 28 days and another 33 suffered serious morbidity, resulting in a total of 86 children (31%) suffering poor outcome. Heart rate score of zero (aOR 6.0, 95%CI 1.9 – 19) was the only predictor independently associated with neonatal mortality within 28 days. Heart rate score of zero (aOR 18, 95% 3.7–86) and respiratory effort score of zero (aOR 6.0, 95%CI 1.3 – 28) were both independently associated with a combined poor neonatal outcome and formed together the best predictive variables. The difference between the AUC’s reduced model of 0.67 was not statistically significantly different than that of the Apgar-based model (CI -0.07 – 0.017).

ConclusionAlthough the Apgar score is the most used prognostic test in obstetrics, only its components heart rate and respiratory effort have strong predictive value for neonatal mortality and morbidity. The reduced model showed comparable discrimination to the Apgar-based model.

Individual components of the Apgar score

7

117

inTroduCTion

In current perinatal care, the Apgar score is used widely to assess the condition of newborns immediately after birth. The score was proposed by Virginia Apgar in 1952, and consists of five characteristics of the new born infant: skin color (pink, blue or pale), heart rate (≥100/ min, <100/min or absent), respiratory effort (normal rate and effort, irregular gasping or absent), muscle tone (active, arms and legs flexed or ‘floppy tone’) and reflex irritability (pulls away, grimace or absent) that contribute equally to the total score each with rating of 0, 1, or 2. 1-3 An Apgar score of seven or higher is thought to indicate a good neonatal condition, while a score below 4 is strongly associated with the risk of neonatal and infant death. 4-7

The initial objectives of the Apgar score are four-fold. The score provides an assessment of the general condition of the newborn as well as its prognosis. Moreover, it provides a basis for the comparison of obstetric care among different centers. Finally it helps midwives and clinicians to focus on the neonate in the first minutes of it’s live. 8, 9

Despite its wide acceptance and use, the value of the Apgar score is questionable. 8-13 The Apgar score fails to predict specific neurologic outcomes, as it was never developed for this goal. Although recent studies have shown a strong association between Apgar score and neonatal mortality and morbidity, all these studies focused on total Apgar score as predictor of neonatal outcome, rather than on its components 4, 14-17.

While speculation on the arbitrary point system is noted, the relative importance of the individual factors on the total Apgar score at 5 minutes has never been investigated systematically. Previous research was limited to 1-minute Apgar scores, which are known to be depressed by anesthesia drugs or laryngeal spasm. 18

In this study, we aimed to assess the association of the five individual components of the 5-minute Apgar score with neonatal mortality and morbidity among singleton babies born at term. In addition we developed a reduced model based on the most predictive components, assessed its predictive performance and compared it to the Apgar-based model.

Chapter 7

118

meThods

We performed a retrospective cohort study among neonates admitted to the Neonatal Intensive Care Unit (NICU) of the Academic Medical Center in Amsterdam between 2000 and 2010. The Academic Medical Center is one of the ten 3rd level care hospitals in the Netherlands and its NICU had in the study period 18 intensive care beds.

We studied singleton babies that were born with a 5-minute Apgar score <7, and were admitted to the NICU. To do so, we identified all singleton babies born between 37+0 and 42+0 weeks of gestation. Stillborn babies (“antepartum and intrapartum deaths”) were not included. Information on pregnancy and neonatal characteristics were extracted directly from the medical files. Gestational age was determined based on first trimester sonography of crown–rump length according to our national guidelines. Children could have been born in the Academic Medical Centre, in 1 of the 20 referring hospitals, or at home. The Apgar score contained heart rate, respiratory effort, color, tonus and reflex irritability. The components were addressed by the attending midwife, obstetrician, pediatrician or neonatologist.

Our primary outcome was neonatal mortality defined as mortality within 28 days after live birth. Secondary outcomes were early neonatal mortality, defined as mortality within 7 days after live birth. Neonatal morbidity was defined by the neonatologists as clinical asphyxia, hypoxic ischemic encephalopathy, respiratory problems, meconium aspiration syndrome, infection or convulsions. Adverse neonatal outcome was a composite outcome of neonatal mortality and / or neonatal morbidity.

StatisticsFirst, we compared baseline characteristics of the neonates with complete information of the individual components of the 5-minutes Apgar score to the neonates without complete information on the individual components of the 5-minutes Apgar score. We used Chi-square tests or the Student’s t-test where appropriate. We then used univariate and multivariate logistic regression to assess the association of the five individual components of the Apgar score with the outcome measures. In the logistic regression analysis, the scores for each component were coded as 0, 1 and 2, with a score of two points serving as the reference category. Associations were expressed as odds ratios (OR) with their 95% confidence interval (CI). The combination of components with the best predictive value was identified by a stepwise selection in both the forward and backward directions using the Akaike’s Information Criterion. 19 The area under the Receiver Operating Characteristic curve (AUC) for the reduced and the Apgar-based



7

119

model were obtained by the standard bootstrap method that corrects for optimism due to the risk of overfitting. The confidence interval around the difference between the AUC of the two models was based on the bootstrap percentile method at the 0.05 level. The AUC assesses the ability of a prediction model to discriminate between neonates who died in the neonatal period versus those who were still alive after 28 days. 20 The predicted probabilities for neonatal mortality were compared between the different prediction models by plotting the prediction of the new (reduced) model as function of the complete model (conventional Apgar score). To improve the clinical visualization, we developed nomograms to calculate the probability to neonatal mortality within 28 days and adverse neonatal outcome according to the components of the 5-minute Apgar score.

Congenital anomalies increase the risk of a low Apgar score. To eliminate this possible confounding factor we performed a sensitivity analysis on a group of term infants without any congenital anomaly. As selection bias might play a role, we also studied the Apgar scores of all children born in the catchment area of the NICU of the Academic Medical Centre in the study period, and compared the score to neonatal outcome. Data were analyzed using the R statistical software environment version 2.15.1 (R Foundation for Statistical Computing, Vienna, Austria) and SAS statistical software package version 9.3 (SAS Institute Inc, Cary, NC, USA).

resulTs

Between 2000 and 2010, there were 2042 term neonates admitted to the Neonatal Intensive Care Unit (NICU), of whom 433 (21%) who had a 5-minute Apgar score below 7. In 276 neonates (62%), there was information on all five individual components of the Apgar score.

Baseline characteristics are presented in Table 1. Baseline characteristics were comparable for neonates with complete data compared to neonates with incomplete data. The majority of infants (42%) were born by unplanned caesarean delivery, mostly performed for suspected fetal distress. 55 (12%) neonates had some congenital anomaly. Mean 1-minute and 5-minute Apgar scores were 2.7 and 4.4, respectively. From the 276 neonates with complete data on the 5 minute Apgar score, 55 (20%) died within 28 days after live birth and another 32 suffered serious neonatal morbidity, resulting in 87 children (31%) suffering poor outcome. Among the 87 neonates that suffered adverse neonatal outcome, respiratory problems were among the most frequent with 46% (Table 1).

Chapter 7

120

Table 1. Baseline characteristics of all neonates admitted to the NICU with a 5 minute Apgar score < 7.

Apgar score < 7N = 443

Information for all componentsN = 276

With missing informationN = 167

P - value

N (%) N (%) N (%)Mode of deliveryVaginalElective CSUnplanned CSFetal distressAssisted vaginalFetal distress

168 (38%)17 (3.8%)

187 (42%)158 (36%)

71 (16%)36 (8.0%)

94 (35%)15 (5.4%)

127 (46%)103 (37%)40 (14%)22 (7.9%)

74 (44%)2 (1.2%)

60 (36%)55 (33%)31 (19%)14 (8.4%)

p = 0.03p = 0.04p = 0.04nsnsns

Delivery AMC 207 (47%) 139 (50%) 68 (41%) p = 0.08Mean GA (days) 279 (± 10) 279 (± 11) 279 (± 10) nsMean birth weight (grams) 3391 (± 644) 3380 (± 623) 3395 (± 624) nsCongenital anomalies 55 (12%) 26 (9.4%) 29 (17%) p = 0.02Mean Apgar score at 1 minute 2.7 (±2.0) 2.7 (± 2.1) 2.7 (± 2.0) nsMean Apgar score at 5 minutes 4.4 (±1.7) 4.4 (± 1.8) 4.4 (± 1.9) nsNeonatal mortality< 24 hours< 7 days< 28 days

21 (4.7%)74 (17%)92 (21%)

11 (4%)46 (17%)55 (20%)

10 (6.0%)28 (17%)37 (22%)

nsnsns

Neonatal morbidityTotalClinical AsphyxiaHIERespiratory problemsMeconium aspiration syndromeInfectionNeonatal convulsions

151 (34%)19 (4%)61 (14%)

191 (43%)50 (11%)23 (5.2%)24 (5.4%)

87 (31%)8 (2.9%)

41 (15%)127 (46%)24 (8.7%)19 (6.9%)6 (2.2%)

65 (39%)11 (6.6%)20 (12%)64 (38%)26 (16%)4 (2.4%)

10 (6.0%)

nsnsnsnsp = 0.03p = 0.05p = 0.04

HIE: hypoxic ischemic encephalopathy

Among 276 neonates, 27 (9.8%) had a value of zero for heart rate after 5 minutes, while each of the other four components more often had zero points for color (25%), reflex irritability (47%), muscle tone (51%) or respiratory effort (38%). Neonatal mortality within 28 days occurred in 55 neonates. Compared to the maximal score of two points, heart rate with score zero (aOR 6.0, 95%CI 1.9 – 19) was the only item of the Apgar score independently associated in the multivariable analysis, while respiratory effort reached borderline statistical significance (aOR 7.1, 95%CI 0.88 – 58). In total, 46 (17%) neonates died within 7 days after birth. Heart rate with a score of zero was the only item significantly associated with 7-day mortality (aOR 7.8, 95%CI 2.4 – 25). (Table 2) The three other variables color, reflex irritability and tone did not reach statistical significance and had odds ratios varying between 0.73 and 2.7.

7

121

Table

2.

Ass

ocia

tion

betw

een

indi

vidu

al c

ompo

nent

s of

5-m

inut

e A

pgar

sco

re <

7 a

nd n

eona

tal m

orta

lity.

N =

276

.

Mort

alit

y <

7 d

ays

N =

46

Mort

alit

y <

28

days

N =

55

Indiv

idual c

om

ponen

ts o

f th

e 5

-min

ute

Apgar

score

N (

%)

Univ

ari

ate

analy

sis

Multiv

ari

ate

analy

sis

N (

%)

Univ

ari

ate

analy

sis

Multiv

ari

ate

analy

sis

Zero

poi

nts

for C

olor

(N =

68/

276)

20 (4

3%)

3.3

(1.3

– 8

.4)

1.1

(0.3

4 –

3.5)

22 (4

0%)

2.0

(0.8

9 –

4.5)

0.73

(0.2

6 –

2.1)

Zero

poi

nts

for H

eart

Rate

(N =

27/

276)

16 (3

5%)

12 (5

.1 –

30)

7.8

(2.4

– 2

5)17

(31%

)10

(4.3

– 2

5)6.

0 (1

.9 –

19)

Zero

poi

nts

for R

eflex

irrit

abili

ty(N

= 1

31/

276)

28 (6

1%)

2.2

(0.4

7 –

10)

1.4

(0.2

8 –

7.3)

35 (6

4%)

2.9

(0.6

4 –

13)

2.4

(0.5

0 –

12)

Zero

poi

nts

for T

one

(N =

142

/27

6)27

(59%

)2.

3 (0

.29

– 19

)1.

7 (0

.18

– 16

)32

(58%

)2.

9 (0

.35

– 24

)1.

9 (0

.19

– 18

)

Zero

poi

nts

for R

espi

rato

ry e

ffort

(N =

106

/27

6)30

(65%

)4.

0 (0

89 –

18)

2.1

(0.4

2 –

10)

36 (6

5%)

10 (1

.4 –

85)

7.1

(0.8

8 –

58)

1. A

val

ue o

f tw

o po

ints

for a

n in

divi

dual

com

pone

nt s

erve

s as

the

refe

renc

e ca

tego

ry fo

r tha

t com

pone

nt.

2. M

ultiv

aria

te m

eans

adj

ustm

ent f

or th

e re

mai

ning

com

pone

nts

of th

e A

pgar

sco

re o

ther

than

the

com

pone

nt o

f int

eres

t.


Chapter 7

122

Figure 1 provides a nomogram to calculate the probability to neonatal mortality within 28 days according to the five individual components of the Apgar score. The score for each of the five individual components corresponds with a certain value on the point scale of the nomogram. The sum of all points gives the total score for each neonate. The total score correlates with a probability for mortality within 28 days after live birth. For example, a zero score of hearth rate and respiratory effort corresponds with a 35% chance of neonatal mortality even if the neonate had a score of two for the remaining components color, reflex irritability and muscle tone.

Figure 1. Nomogram of the individual components of the 5 minute Apgar score and perinatal mortality

within 28 days after live birth.

The composite adverse neonatal outcome occurred in 87 of the neonates, and consisted of neonatal mortality within 28 days after live birth or neonatal morbidity. Table 3 shows the association between the Apgar score and adverse neonatal outcome. None of the individual components of the 5-minute Apgar score was independently associated with neonatal morbidity. For the composite outcome again only heart rate with score zero (aOR 18, 95%CI 3.7 – 86) and respiratory effort with score zero (aOR 6.0, 95%CI 1.3 – 28) were independently associated with the combined poor neonatal outcome. Figure 2 shows the nomogram to calculate the probability for adverse neonatal outcome. It is shown that color has no predictive value on poor neonatal outcomes and that mediate hearth rate or reflex irritability also has limited value in the prediction of poor neonatal outcome.

A sensitivity analysis on a subgroup of neonates without any congenital anomaly showed similar results. Again, only heart rate (aOR 21, 95%CI 4.1 – 110) and respiratory effort (aOR 8.5, 9%CI 1.1 – 69) were independently associated with poor neonatal outcome. See Table in appendix.

7

123

Table

3.

Ass

ocia

tion

betw

een

indi

vidu

al c

ompo

nent

s of

5-m

inut

e A

pgar

sco

re <

7 a

nd a

dver

se n

eona

tal o

utco

me.

N =

276

.

Neo

nata

l morb

idity

N =

32

Poor

neo

nata

l outc

om

eN

= 8

7In

div

idual c

om

ponen

ts o

f th

e 5

-min

ute

Apgar

score

N (

%)

Univ

ari

ate

analy

sis

Multiv

ari

ate

analy

sis

N (

%)

Univ

ari

ate

analy

sis

Multiv

ari

ate

analy

sis

Zero

poi

nts

for C

olor

(N =

68/

276)

15 (4

7%)

3.2

(1.1

– 9

.5)

2.0

(059

– 7

.0)

37 (4

3%)

3.2

(1.5

– 6

.6)

1.3

(0.5

3 –

3.3)

Zero

poi

nts

for H

eart

Rate

(N =

27/

276)

8 (2

5%)

4.6

(1.7

– 1

2)1.

9 (0

.55

– 6.

8)25

(30%

)44

(10

– 19

4)18

(3.7

– 8

6)

Zero

poi

nts

for R

eflex

irrit

abili

ty(N

= 1

31/

276)

20 (6

3%)

1.4

(0.3

1 –

6.8)

1.1

(0.2

0 –

6.1)

55 (6

4%)

2.5

(0.7

9 –

8.1)

2.0

(0.5

7 –

7.0)

Zero

poi

nts

for T

one

(N =

142

/27

6)20

(63%

)0.

43 (0

.10

– 1.

8)0.

38 (0

.08

– 1.

8)52

(60%

)1.

0 (0

.28

– 3.

6)0.

67 (0

.15

– 3.

0)

Zero

poi

nts

for R

espi

rato

ry e

ffort

(N =

106

/27

6)18

(56%

)4.

4 (0

.56

– 35

)2.

7 (0

.32

– 23

)54

(62%

)11

(2.3

– 4

7)6.

0 (1

.3 –

28)

1. A

val

ue o

f tw

o po

ints

for a

n in

divi

dual

com

pone

nt s

erve

s as

the

refe

renc

e ca

tego

ry fo

r tha

t com

pone

nt.

2. M

ultiv

aria

te m

eans

adj

ustm

ent f

or th

e re

mai

ning

com

pone

nts

of th

e A

pgar

sco

re o

ther

than

the

com

pone

nt o

f int

eres

t.3.

Com

bine

d po

or n

eona

tal o

utco

me:

neo

nata

l mor

talit

y w

ithin

28

days

afte

r liv

e bi

rth o

r ne

onat

al m

orbi

dity

con

sistin

g of

any

of

the

follo

win

g di

agno

sis;

Asp

hyxi

a (u

mbi

lical

cor

d pH

< 7

.0),

Hyp

oxic

Isch

emic

Enc

epha

lopa

thy,

Res

pira

tory

pro

blem

s, M

econ

ium

asp

iratio

n sy

ndro

me,

Infe

ctio

n an

d N

eona

tal c

onvu

lsion

s.

Chapter 7

124

Figure 2. Nomogram of the individual components of the 5 minute Apgar score and poor neonatal

outcome.

Combined poor neonatal outcome: neonatal mortality within 28 days after live birth or neonatal morbidity

consisting of any of the following diagnosis; Asphyxia (umbilical cord pH < 7.0), Hypoxic Ischemic

Encephalopathy, Respiratory problems, Meconium aspiration syndrome, Infection and Neonatal convulsions.

After penalized variable selection based on the Akaike Information Criterion to identify the components with the best predictive capacity, heart rate and respiratory effort were selected. The AUC of the prediction model for neonatal mortality within 28 days based on those two variables (model 1) was 0.75 (95%CI 0.59 -0.89). Bootstrapping the whole selection procedure in 1000 bootstrap samples and correcting for optimism resulted in an AUC of 0.67 (CI 0.55 – 0.70) for the reduced model. The AUC of the Apgar-based model was 0.69 (CI 0.57 – 0.71). The difference between the AUCs was however not statistically significant (CI -0.07 – 0.017). Figure 3 shows plots of predicted probabilities for perinatal mortality of the Apgar score at 5 minutes (model 2) versus those of the reduced model.

For this study, we needed detailed information on the components of the Apgar score and therefore we created a cohort of term neonates admitted to the NICU of a tertiary care center. As selection bias might play a role, we compared the Apgar scores of all children born in the referral area of the NICU of the Academic Medical Centre with our study cohort. In the referral area of the Amsterdam NICU, 83 909 term singletons were born alive in between 2005 -2007, of whom 61 died within 28 days after birth. Of all babies who died 54 (88.5%) had a 5-minute Apgar score < 7. These rates are comparable to the prevalence rates of our study cohort; of the children admitted to the NICU between 2005 - 2007, 37 died and 23 (62%) of them had a 5 minute Apgar score < 7. The remaining 24 neonates (61-37) are probably admitted to the NICU of the VU medical center, another 3rd level care hospital in the referral area of the Amsterdam.

7

125

Figure 3. Plots of predicted probabilities for perinatal mortality within 28 days of total Apgar score at 5

minutes (model 2) versus selected group of individual components (model 1).

disCussion

In this cohort study we investigated the association between the five individual components of the Apgar score and neonatal mortality and morbidity. Among term infants with a 5-minute Apgar score < 7, we found that for adverse neonatal outcome only heart rate and respiratory effort were independently associated and had the best prognostic value for short-term neonatal outcome. Color was not predictive while reflex irritability and tone had limited predictive capacity. In addition we showed that the reduced model based on these two components improved predictions when compared to the Apgar-based model.

Strengths and limitations Although complete information on the individual components of the 5 minute Apgar score was only available in 62% of the term neonates in our study population, we found no differences in characteristics between the neonates with or without detailed information on the Apgar score and therefore we have no indication that this hinder our results. Due the roughly equal number of neonates born in the AMC and neonates referred form other hospitals directly after birth, the influence of the different resuscitation


Chapter 7

126

care in first minutes of a newborns life is limited. We studied 5-minutes Apgar score and not 1-minute scores, since 5-minutes Apgar score shows stronger associations with long-term outcomes than at 1 minute. 4, 18 Indeed, 5-minute Apgar is more used in clinical practice and more often used as prognosticators. A number of factors, including prematurity and congenital anomalies, influence the Apgar score and its prognostic value. Since we restricted our analysis to a cohort of term neonates and performed a sensitivity analysis on a subgroup without congenital anomalies we created a homogenous study cohort and limited these effects.

Another limitation might be that we have very restricted information on maternal and pregnancy characteristics corresponding to the neonates in our study cohort. Maternal characteristics, for example increased maternal age and pregnancy induced hypertension, are known to increase the risk of a low Apgar score. However, we did not study the risk of low Apgar score, but aimed to study the impact of the Apgar score on neonatal outcome. It is unlikely that maternal or pregnancy characteristics influence the association between the individual components of the Apgar score and neonatal outcomes. As we wanted to assess the association between components of the Apgar score on neonatal outcome, and this specific information on the individual components is lacking in population based data, we think the choice for our study cohort is justifiable.

Other literatureMost literature on the assessment of the Apgar score is focused on the total score, although it consists of five individual components. The Apgar score, despite being the most used prognostic test in obstetrics, is frequently criticized as it would be subject to strong inter observer variability. A recent study by Liodromiti et al. 4 showed that although the Apgar score was first described more than 60 years ago, a low 5-minute Apgar score is strongly associated with the risk of neonatal and infant death. Dalili et al, compared different Apgar scoring systems in predicting birth asphyxia. 17 They found that the combined Apgar score, which is a consolidation of the Specified and Expanded Apgar score has the best predictive value for birth asphyxia. Although the AUC of 0.99 suggest an almost perfect discrimination tool, the feasibility of the combined score is questionable. The combined score ranges from 0 to 17 and consist of 12 items that need to be evaluated. 21-23 Our study results show that with the evaluation of heart rate and respiratory effort obstetric caregivers can achieve a quite good prognosis for the newborn baby. However, the results are not completely comparable as we used other outcome measurements and restricted our analyses to term singletons. Virginia Apgar herself also noticed that not all components are of equal importance, but she did not want to differentially weight the individual items as the score would become too complicated. 24 Several other investigators have tried to



7

127

modify the Apgar score. 8, 18 In 1973 Crawford et al. already investigated the relative importance of the individual components. 18 They suggested an “Apgar-minus-Color score”, as the inclusion of color to the total score weakened the performance of the Apgar score. Besides the enormous changes in perinatal care with increased survival rates in the last decades, their findings were mostly based on the 1-minute Apgar score. In contrast, we used the 5-minute Apgar scores and the results are not fully comparable, we also found color do not have any additional predictive capacity. Besides color we also found that muscle tone both have no additional prognostic value for the short-term neonatal outcome. This difference can be explained by the fact that we used 5-minute Apgar scores, where the first transition problems have been solved and excluded premature births. Color and muscle tone are all three factors partially dependent on the physiologic maturity of newborns. 12, 13

Clinical relevanceThe main goal of our study was not to change the Apgar score. We think that a “health index” like the Apgar score which is integrated in the Millennium Development Goal of reducing child mortality has proven its value. We only wanted to assess if after more than 50 years of worldwide use the individual components remains pertinent. Prognostic models become more and more integrated in (obstetric) medicine and the Apgar score is a good example of how a prognostic tool can be misinterpreted. Critics of using the Apgar score as a prognostic tool for short or long-term infant outcomes ignore the fact that prediction of the condition of newborns was indeed one of the initial objectives. 1, 24 Our study underlines the hypotheses that the Apgar score still has prognostic value, however an assessment limited to the heart rate and breathing effort could be more appropriate for this purpose. Furthermore, score of zero for heart rate is only given in the absence of heart beats, one could wonder if the assessment of heart rate as a continuous scale, i.e. the actual beats per minute, is not more appropriate to give prognostic information than the categorical division used now. On the other hand, we believe that the total Apgar score still is very valuable, not only for its simplicity, but also in the assessment of the need for resuscitation. The five individual components of the Apgar score are strongly correlated to each other and zero points for color and tone imply problems on oxygenation. As for the obstetric caregivers a pale ‘floppy’ infant is easy to recognize, this will contribute to an accelerated course of action immediately after birth.

ConClusionAlthough the Apgar score is probably the most used prognostic test in obstetrics, only heart rate and respiratory effort have marked predictive value for short-term neonatal outcome in term infants. A reduced model based on both of these variables had comparable discrimination to the current Apgar-based model.


Chapter 7

128

referenCes


2. Apgar V, Kreiselmen J. Studies on resuscitation; an experimental evaluation of the Bloxsom air lock. American journal of obstetrics and gynecology 1953;65:45-52.

3. Van Aarnhem Ams VBM, Crone-Kraaijeveld E, Merkx Jam, Renckens Aljm, Somford Rg, Flikweert S, Pijnenborg L. NHG-Standaard Onderzoek van de pasgeborene. Huisarts Wet 2001;44:609-14.





8. Chamberlain G, Banks J. Assessment of the Apgar score. Lancet 1974;2:1225-8.

9. Schmidt B, Kirpalani H, Rosenbaum P, Cadman D. Strengths and limitations of the Apgar score: a critical appraisal. Journal of clinical epidemiology 1988;41:843-50.

10. Auld PA, Rudolph AJ, Avery ME, et al. Responsiveness and resuscitation of the newborn. The use of the Apgar score. American journal of diseases of children 1961;101:713-24.


12. Hegyi T, Carbone T, Anwar M, et al. The apgar score and its components in the preterm infant. Pediatrics 1998;101:77-81.

13. Li F, Wu T, Lei X, Zhang H, Mao M, Zhang J. The apgar score and infant mortality. PloS one 2013;8:e69072.




7

129

15. Razaz N, Boyce WT, Brownell M, et al. Five-minute Apgar score as a marker for developmental vulnerability at 5 years of age. Archives of disease in childhood Fetal and neonatal edition 2015.

16. Anyaegbunam A, Fleischer A, Whitty J, Brustman L, Randolph G, Langer O. Association between umbilical artery cord pH, five-minute Apgar scores and neonatal outcome. Gynecologic and obstetric investigation 1991;32:220-3.

17. Dalili H, Nili F, Sheikh M, Hardani AK, Shariat M, Nayeri F. Comparison of the four proposed Apgar scoring systems in the assessment of birth asphyxia and adverse early neurologic outcomes. PloS one 2015;10:e0122116.

18. Crawford JS, Davies P, Pearson JF. Significance of the individual components of the Apgar score. British journal of anaesthesia 1973;45:148-58.

19. Atkinson A. A note on the generilzed information on criterion for choice of a model. Biometrika 2010;67:413-8.

20. Harrell FE, Jr., Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in medicine 1996;15:361-87.

21. Rudiger M, Wauer RR, Schmidt K, Kuster H. The Apgar score. Pediatrics 2006;118:1314-5; author reply 15-6.

22. Rudiger M, Braun N, Gurth H, Bergert R, Dinger J. Preterm resuscitation I: clinical approaches to improve management in delivery room. Early human development 2011;87:749-53.

23. Rudiger M, Aguar, M. Newborn assessment in the delivery room. Neoreviews 2012;13:336 -42.

24. Apgar V, Holaday DA, James LS, Weisbrot IM, Berrien C. Evaluation of the newborn infant; second report. Journal of the American Medical Association 1958;168:1985-8.

Chapter 7

130

Appen

dix

. A

ssoc

iatio

n be

twee

n in

divi

dual

com

pone

nts

of 5

-min

ute

Apg

ar s

core

< 7

and

adv

erse

neo

nata

l out

com

e am

ong

infa

nts

with

out c

onge

nita

l ano

mal

ies.

N

= 25

0.

Mort

alit

y <

28 d

ays

N =

43

Poor

neo

nata

l outc

om

eN

= 6

6In

div

idual c

om

ponen

ts o

f th

e 5

-min

ute

Apgar

score

N (

%)

Univ

ari

ate

analy

sis

Multiv

ari

ate

analy

sis

N (

%)

Univ

ari

ate

analy

sis

Multiv

ari

ate

analy

sis

Zero

poi

nts

for C

olor

(N =

68/

276)

19 (4

4%)

2.6

(1.0

3 –

6.5)

0.48

(0.1

3 –

1.8)

33 (5

0%)

5.1

(2.2

– 1

2)1.

7 (0

58 –

4.9

)

Zero

poi

nts

for H

eart

Rate

(N =

27/

276)

17 (4

0%)

18 (6

.7 –

47)

14 (3

.5 –

52)

24 (3

6%)

66 (1

5 –

298)

21 (4

.1 –

110

)

Zero

poi

nts

for R

eflex

irrit

abili

ty(N

= 1

31/

276)

29 (6

7%)

2.3

(1.2

– 4

.7)

1.5

(0.6

2 –

3.6)

45 (6

8%)

8.5

(1.1

– 6

7)6.

0 (0

.40

– 27

)

Zero

poi

nts

for T

one

(N =

142

/27

6)26

(60%

)1.

5 (0

.17

– 13

)0.

61 (0

.06

– 6.

6)43

(65%

)3.

0 (0

.35

– 25

)1.

7 (0

.17

– 16

)

Zero

poi

nts

for R

espi

rato

ry e

ffort

(N =

106

/27

6)29

(67%

)8.

8 (1

.1 –

69)

4.3

(0.5

1 –

37)

45 (6

8%)

18 (2

.3 –

137

)8.

5 (1

.1 –

69)

1. A

val

ue o

f tw

o po

ints

for a

n in

divi

dual

com

pone

nt s

erve

s as

the

refe

renc

e ca

tego

ry fo

r tha

t com

pone

nt.

2. M

ultiv

aria

te m

eans

adj

ustm

ent f

or th

e re

mai

ning

com

pone

nts

of th

e A

pgar

sco

re o

ther

than

the

com

pone

nt o

f int

eres

t.3.

Com

bine

d po

or n

eona

tal o

utco

me:

neo

nata

l mor

talit

y w

ithin

28

days

afte

r liv

e bi

rth o

r ne

onat

al m

orbi

dity

con

sistin

g of

any

of t

he fo

llow

ing

diag

nosis

; A

sphy

xia

(um

bilic

al c

ord

pH <

7.0

), H

ypox

ic Is

chem

ic E

ncep

halo

path

y, R

espi

rato

ry p

robl

ems,

Mec

oniu

m a

spira

tion

synd

rom

e, In

fect

ion

and

Neo

nata

l con

vulsi

ons.

7

131

Appen

dix

. A

ssoc

iatio

n be

twee

n in

divi

dual

com

pone

nts

of 5

-min

ute

Apg

ar s

core

< 7

and

adv

erse

neo

nata

l out

com

e am

ong

infa

nts

with

out c

onge

nita

l ano

mal

ies.

N

= 25

0.

Mort

alit

y <

28

days

N =

43

Poor

neo

nata

l outc

om

eN

= 6

6In

div

idual c

om

ponen

ts o

f th

e 5

-min

ute

Apgar

score

N (

%)

Univ

ari

ate

analy

sis

Multiv

ari

ate

analy

sis

N (

%)

Univ

ari

ate

analy

sis

Multiv

ari

ate

analy

sis

Zero

poi

nts

for C

olor

(N =

68/

276)

19 (4

4%)

2.6

(1.0

3 –

6.5)

0.48

(0.1

3 –

1.8)

33 (5

0%)

5.1

(2.2

– 1

2)1.

7 (0

58 –

4.9

)

Zero

poi

nts

for H

eart

Rate

(N =

27/

276)

17 (4

0%)

18 (6

.7 –

47)

14 (3

.5 –

52)

24 (3

6%)

66 (1

5 –

298)

21 (4

.1 –

110

)

Zero

poi

nts

for R

eflex

irrit

abili

ty(N

= 1

31/

276)

29 (6

7%)

2.3

(1.2

– 4

.7)

1.5

(0.6

2 –

3.6)

45 (6

8%)

8.5

(1.1

– 6

7)6.

0 (0

.40

– 27

)

Zero

poi

nts

for T

one

(N =

142

/27

6)26

(60%

)1.

5 (0

.17

– 13

)0.

61 (0

.06

– 6.

6)43

(65%

)3.

0 (0

.35

– 25

)1.

7 (0

.17

– 16

)

Zero

poi

nts

for R

espi

rato

ry e

ffort

(N =

106

/27

6)29

(67%

)8.

8 (1

.1 –

69)

4.3

(0.5

1 –

37)

45 (6

8%)

18 (2

.3 –

137

)8.

5 (1

.1 –

69)

1. A

val

ue o

f tw

o po

ints

for a

n in

divi

dual

com

pone

nt s

erve

s as

the

refe

renc

e ca

tego

ry fo

r tha

t com

pone

nt.

2. M

ultiv

aria

te m

eans

adj

ustm

ent f

or th

e re

mai

ning

com

pone

nts

of th

e A

pgar

sco

re o

ther

than

the

com

pone

nt o

f int

eres

t.3.

Com

bine

d po

or n

eona

tal o

utco

me:

neo

nata

l mor

talit

y w

ithin

28

days

afte

r liv

e bi

rth o

r ne

onat

al m

orbi

dity

con

sistin

g of

any

of t

he fo

llow

ing

diag

nosis

; A

sphy

xia

(um

bilic

al c

ord

pH <

7.0

), H

ypox

ic Is

chem

ic E

ncep

halo

path

y, R

espi

rato

ry p

robl

ems,

Mec

oniu

m a

spira

tion

synd

rom

e, In

fect

ion

and

Neo

nata

l con

vulsi

ons.


Chapter 8 The development of a clinical prediction model to support clinicians in the simultaneous assessment

of the risks of failure to progress and fetal asphyxia in term pregnancies

S. EnsingU. OgbaE. Schuit

A.C.J. RavelliA. Abu-Hanna

B.W.J. Mol

Under embargo

Chapter 8

134

absTraCT

ObjectiveDuring labor, problems can occur due to e.g. failure to progress or perinatal asphyxia. Obviously, both problems can occur simultaneously. We therefore developed a clinical tool to assess the risk of failure to progress (FTP) and fetal distress simultaneously.

Study designWe used data on term singleton pregnancies in cephalic presentation from the Netherlands Perinatal Registry (PRN) between 2000-2010. We developed prediction models for perinatal asphyxia (Apgar score at 5 minutes < 7) and for obstetric intervention due to failure to progress (FTP), using first only antepartum variables and secondly both antepartum and intrapartum variables. Predictive performance was assessed by calibration and discrimination (Area Under the Receiver Operating Characteristic-curve (AUC)). Two-dimensional graphs were constructed to show the predicted probabilities in the antepartum en intrapartum period.

ResultsOut of 1,517,419 women, 175,708 (12%) had an intervention due to FTP while 12,571 (0.83%) delivered a neonate with asphyxia. The antepartum models had AUCs of 0.72 (95%CI 0.71 - 0.74) and 0.65 (0.65 – 0.66) for asphyxia and failure to progress, respectively. Prognostic performance of the intrapartum models was good as indicated by AUCs of 0.78 (95%CI 0.77-0.79) and 0.67 (95%CI 0.66 – 0.68), respectively. The predicted probability of an obstetric intervention due to FTP varied between 0.3% to over 90% and for perinatal asphyxia the probabilities varied between 0.1% and 13%. Overall, the probability of an intervention due to FTP increased with an increased risk of perinatal asphyxia.

ConclusionIn women with a term singleton pregnancy models predicting risks of fetal asphyxia and an obstetric intervention due to FTP showed good performance. After external validation of the models the graph combining the predicted probabilities could aid clinicians in the choice of interventions during labor and delivery in both high and low risk pregnancies.


8

Two dimensional clinical tool

135

inTroduCTion

Normal labor is, when expressed per centimeter travelled, one of the most dangerous journeys in life. Risks include, amongst other, perinatal asphyxia during labor which can result in cerebral palsy or even neonatal death, and obstructed labor. 1, 2 Failure to progress (FTP) has been linked to an increase in risk of maternal obstetric trauma, postpartum hemorrhage, puerperal febrile morbidity and composite maternal morbidity, but also to adverse neonatal outcomes like a low 5-minute Apgar scores, admission to the neonatal intensive care unit and composite perinatal morbidity. 3-7

In 1 in 5 deliveries the baby is not born spontaneously, thus requiring vaginal instrumental delivery or Caesarean Section. 8, 9 In labor, there is a continuous trade-off of adverse outcomes on the neonatal (asphyxia) or maternal level (instrumental delivery), eventually leading to the decision to start or continue natural labor versus delivery of the baby instrumentally.

The risk of perinatal asphyxia as well as obstructed labor can be predicted. Several prognostic models have been developed to predict either of these interventions. 10-

16 Since during labor there is a risk for both events, ideally a combination of two such models could facilitate decision-making. In case there is a high risk of obstructed labor, an accompanying high risk of fetal distress might lead to the decision to deliver the baby, while in case of a low risk of fetal distress one might allow more time to attempt vaginal delivery. Congruently, a low risk of obstructed labor and anticipated fast spontaneous delivery might allow some more expectant management. However, at present clinicians base their clinical decisions typically on one of these problems, rather than integrating both dimensions.

At present there is no clinical model that combines the predictions of both outcomes to provide clinicians with an objective assessment of the risks of both outcomes simultaneously. Recently, in the Netherlands two prognostic models have been developed in a selected group of pregnant women with high-risk pregnancies using both antepartum and intrapartum variables to independently predict the risk of perinatal asphyxia as well as mode of delivery, including interventions for failure to progress. 10,

11 However, studies have indicated that a significant proportion of intrapartum perinatal asphyxia occurs in pregnancies classified as low-risk pregnancies. 13, 17 The two recently developed prediction models for perinatal asphyxia and failure to progress are insufficient to be used in the primary care setting where women receive obstetric care who are pre-labor classified as low – risk for pregnancy complications. Therefore we developed new models to predict perinatal asphyxia and an intervention due to FTP


Chapter 8

136

both before and during labor, using the whole obstetric population of the Netherlands, and combined the predicted probabilities into a two-dimensional graphical tool. This tool could assist clinicians in the assessment of the risks of perinatal asphyxia and failure to progress and guide subsequent interventions during labor and delivery.

meThods

For this study we used data from the Netherlands Perinatal Registry (PRN). The PRN is a registry combining data on pregnancies, births and neonatal outcomes of births in the Netherlands. Records include information on mothers (e.g. maternal age, obstetric history, and parity), pregnancy (e.g. mode of conception, mode of delivery) and children (e.g. birth weight, gestational age, Apgar score) and the coverage of the PRN registry is about 96% of all deliveries in the Netherlands. 18 As this was an anonymous database, IRB approval was not needed under Dutch law.

Study populationWe included women with a singleton pregnancy between 37+0 to 41+6 completed weeks of gestation who delivered between 2000 and 2010, and had the intention to deliver vaginally. Antenatal deaths, non-cephalic presentation, elective caesarean section and fetuses with congenital anomalies were excluded.

Outcome measurementsOur primary outcomes were obstetric intervention for failure to progress and perinatal asphyxia. Obstetric intervention for failure to progress was defined as instrumental vaginal delivery or caesarean delivery for non-progressive labor, which is defined by the obstetric caregiver. Following international guidelines failure to progress is defined as no cervical dilatation during the active phase of labor for at least 2 hours or no descent of the fetus’s head during the second stage of labor for at least 1 hour despite adequate uterine contraction. 19 Perinatal asphyxia is defined as an infant having an Apgar score of less than 7 at 5 minutes after birth.

Prediction modelsFirst we developed four prediction models, two for antepartum prediction and two for integrated antepartum and intrapartum prediction. To predict the risk of obstetric intervention for failure to progress (model 1 and 2), women were assigned to: obstetric intervention for failure to progress (FTP) versus other women with spontaneous vaginal delivery (reference) or obstetric intervention for suspected fetal distress (FD). Model 3 (antepartum only) and 4 (ante and intrapartum) were prediction models to estimate the risk of perinatal asphyxia.

8


137

Candidate predictorsIn development of the prediction models, we selected candidate predictors based on literature and clinical reasoning. 10-12, 14, 15 The antepartum variables included maternal age (years), gestational age (days), multiparous, non-Caucasian ethnicity, low social-economic status, female fetal gender, previous cesarean section, hypertensive disease (pre-existing and pregnancy induced), diabetes mellitus (pre-existing and gestational diabetes), use of artificial reproduction techniques, birth weight below 5th percentile and above 95th percentile according to birth weight versus P5-P95 as a reference 20, vaginal bleeding in second and/or third trimester, coagulation disorder, obstetric history of intrauterine fetal death (IUFD), and previous preterm birth <37 weeks of gestation.

The second set of candidate predictors contained variables obtained early during labor such as induction of labor, onset of labor in secondary care (i.e. prelabor specified high-risk pregnancies based on medical features), oxytocin augmentation, use of epidural anesthesia, ruptured membranes > 24 hours and meconium-stained amniotic fluid. These intrapartum predictors were added to the antepartum predictors to determine their added predictive value.

Statistics We described the baseline characteristics of the study population, using the mean and standard deviation for continuous variables and proportions of the total population for categorical variables. Univariable associations between the candidate predictors and the outcome categories were estimated with logistic regression analysis (for perinatal asphyxia) and multinomial logistic regression analysis (for mode of delivery). 21 Selection based on univariable statistics might result in unstable prediction models, so we chose not to perform any preselection and to include all candidate predictors in the multivariable analysis. 22, 23

The final predictors for all four prediction models were identified by a backward stepwise selection using Akaike’s Information Criterion. 24 We evaluated the discriminative performance of the models by the area under the Receiver Operating Characteristic curve (AUC). For the multinomial model we calculated three AUCs, each time relating one outcome versus the reference category. The predicted probabilities were compared with the observed frequencies of the different outcome categories using calibration plots to assess the calibration of the four prediction models. The models were internally validated using bootstrapping to account for optimism and over fitting. 22, 25 One hundred bootstrap samples were drawn from the original dataset with replacement,

Chapter 8

138

allowing for multiple sampling of the same individual. Within each bootstrap sample the statistical process described earlier was repeated. Finally, we created two-dimensional graphs by plotting the individual risks for non-progressive labor and perinatal asphyxia calculated for each individual in the study population using the developed antepartum and integrated antepartum and intrapartum prediction models. Ordinarily, neonates who do not have an Apgar score < 7, but who have been delivered by emergency caesarean because of suspected fetal distress, would not be captured with our outcome of Apgar < 7 (perinatal asphyxia). So a sensitivity analysis was performed to predict combined adverse neonatal outcome. This combined adverse neonatal outcome was defined as obstetric intervention for suspected fetal distress and / or an Apgar score < 7 at 5 minutes. A second pair of graphs was created by plotting the probabilities of an intervention for non – progressive labor against the probabilities for adverse neonatal outcome. The study population was selected in SAS version 9.3 and all analyses were performed in R version 0.98.1091 (The R Foundation for Statistical Computing). We adhered to the TRIPOD statement to ensure transparent reporting of the prediction models included in this study. 26 Permission for record use and analysis of data for the purpose of this study was obtained from PRN (registered as data petition 11.91).

resulTs

Between 2000 and 2010 there were 1,517,419 women who gave birth to a baby with cephalic presentation at term. 276,359 (19%) had an obstetric intervention of which 175,708 (12%) was due to non-progressive labor and 100,651 (7%) for FD. Of the 1,517,419 term infants alive at the onset of labor, 12,571 (0.83%) had a 5-minute Apgar score less than 7. Additional characteristics of the study population are presented in table 1. Women with an intervention for non-progressive labor were more often nulliparous, suffering hypertensive diseases or pregnant after artificial reproduction technologies. Their labor was often induced, they received frequently oxytocin augmentation and their neonates were large for gestational age. For women who gave birth to a baby with perinatal asphyxia the baseline characteristics were comparable to women with an intervention for suspected fetal distress.

8


139

Table

1.

Cha

ract

erist

ics

of th

e stu

dy p

opul

atio

n.

Del

iver

yTo

tal

Sponta

neo

us

FTP

FDPer

inata

l asp

hyx

iaChara

cter

istic

N =

1 5

17

419

N =

1 2

41

060

(82

)N

= 1

75

70

8 (

12)

N =

100

65

1 (

7)

N =

12 5

71 (

0.8

3)

Mate

rnal a

ge

(yea

rs)

30.3

±4.

930

.4 ±

4.9

30.3

±4.

630

.3 ±

5.1

30.2

±5.

1G

esta

tional a

ge

(days)

278

±827

8 ±7

.627

9 ±9

278

±927

8 ±7

Non-C

auca

sian e

thnic

ity

248

297

(16%

)20

8 28

5 (1

7%)

23 1

30 (1

3%)

16 8

82 (1

7%)

2614

(21%

)Lo

w s

oci

o-e

conom

ic s

tatu

s38

2 07

3 (2

5%)

313

854

(25%

)41

124

(23%

)27

095

(27%

)37

18 (3

0%)

Nulli

paro

us

682

166

(45%

)47

2 19

7 (3

8%)

136

369

(78%

)73

600

(73%

)76

85 (6

1%)

Pre

vious

caes

are

an s

ection

65 2

51 (4

%)

33 4

60 (2

.7%

)20

760

(12%

)11

031

(11%

)87

6 (7

.0%

)H

yper

tensi

ve d

isea

se11

9 30

9 (8

%)

81 5

37 (7

%)

21 6

03 (1

2%)

16 1

69 (1

6%)

1664

(13%

)D

iabet

es m

ellit

us

14 4

78 (0

.95%

)10

545

(0.8

5%)

2507

(1.4

%)

1426

(1.4

%)

234

(1.9

%))

Art

ifici

al r

epro

duct

ion t

echniq

ue

124

965

(8.2

%)

87 0

14 (7

.0%

)22

269

(13%

)15

682

(16%

)15

50 (1

2%)

Bir

th w

eight

cate

gori

esBir

th w

eight

P5

- P

95

1 36

0 99

8 (9

0%)

1 12

3 19

8 (9

0%)

151

962

(86%

)85

838

(85%

)10

285

(82%

)SG

A <

P5

71 0

87 (4

.7%

)56

455

(4.5

%)

3740

(2.1

%)

10 8

92 (1

1%)

1282

(10%

)LG

A >

P9

585

344

(5.6

%)

61 4

07 (4

.9%

)20

006

(11%

)39

21 (3

.9%

)10

04 (8

.0%

)Coagula

tion d

isord

er23

93 (0

.16%

)18

91 (0

.15%

)30

1 (0

.17%

)20

1 (0

.20%

)30

(0.2

4%)

Neo

nata

l fem

ale

gen

der

774

462

(51%

)61

8 67

7 (5

0%)

97 4

38 (5

5%)

58 3

47 (5

8%)

7204

(57%

)V

agin

al b

leed

ing

4286

(0.2

8%)

3130

(0.2

5%)

677

(0.3

9%)

479

(0.4

8%)

61 (0

.48%

)Pre

vious

IUFD

3325

(0.2

2%)

2922

(0.2

4%)

188

(0.1

1%)

215

(0.2

1%)

32 (0

.25%

)Pre

vious

pre

term

bir

th45

83 (0

.30%

)40

86 (0

.33%

)22

9 (0

.13%

)26

8 (0

.27%

)33

(0.2

5%)

Labor

that

start

s in

sec

ondary

care

527

862

(35%

)39

0 65

2 (3

1%)

80 3

77 (4

6%)

56 8

33 (5

6%)

5860

(47%

)In

duct

ion o

f la

bor

209

781

(14%

)15

4 16

3 (1

2%)

30 1

36 (1

7%)

25 4

82 (2

5%)

2612

(21%

)Ruptu

red m

embra

nes

> 2

4h

107

671

(7.1

%)

76 1

68 (6

.1%

)21

993

(13%

)95

10 (9

.4%

)12

09 (1

0%)

Ox

yto

cin a

ugm

enta

tion

347

845

(23%

)21

4 93

3 (1

7%)

90 1

42 (5

1%)

42 7

70 (4

2%)

4599

(37%

)Ep

idura

l anes

thes

ia96

111

(6.3

%)

69 4

84 (5

.6%

)14

923

(8.5

%)

11 7

04 (1

2%)

1499

(12%

)M

econiu

m s

tain

ed a

mnio

tic

fluid

143

573

(9.5

%)

105

116

(8.5

%)

20 6

93 (1

2%)

17 7

64 (1

8%)

2051

(16%

)

Dat

a ar

e pr

esen

ted

as m

ean

± sta

ndar

d de

viat

ion

or n

(%)


Chapter 8

140

Prediction models for failure to progressIn Table 2 A and B the multivariate associations between the variables and an obstetric intervention for non-progressive labor (model 1 and 2) are presented. 16 antepartum variables were identified to predict obstetric intervention due to failure to progress: maternal age, gestational age, nulliparity, non-Caucasian ethnicity, low socioeconomic status, female fetal gender, previous cesarean section, hypertensive disease, diabetes, use of artificial reproduction techniques, birth weight below 5th percentile and above 95th percentile according to birth weight reference, vaginal bleeding, coagulation disorder, previous IUFD and previous preterm birth. The AUC for model 1 (antepartum only) was 0.72 (95%CI0.71 -0.74) showing moderate discrimination. The addition of intrapartum characteristics (model 2) yielded that induction of labor, oxytocin augmentation, secondary care at onset of labor, prolonged rupture of membranes, meconium stained amniotic fluid and epidural anesthesia were additional predictors for obstetric intervention due to failure to progress (model 2). The AUC for model 2 was 0.78 (95%CI 0.77 – 0.79). The calibration graphs of model 1 and 2 are presented in figure 1, showing good calibration. The probability of an obstetric intervention due to FTP varied between 0.3% to > 90%.

Prediction models for perinatal asphyxiaIn model 3, 15 antepartum variables were identified to predict perinatal asphyxia: maternal age, gestational age, nulliparity, non-Caucasian ethnicity, female fetal gender, previous cesarean section, hypertensive disease, diabetes, use of artificial reproduction techniques, low socioeconomic status, birth weight below 5th percentile and above 95th percentile, vaginal bleeding, coagulation disorder and previous IUFD (Table 3). The AUC for model 3 (antepartum only) was 0.65 (95%CI 0.65 – 0.66), showing moderate discrimination. In model 4, the addition of intrapartum characteristics to model 3 yielded that induction of labor, oxytocin augmentation, secondary care at onset of labor, prolonged rupture of membranes, meconium stained amniotic fluid, epidural anesthesia and sedation were additional predictors for perinatal asphyxia (Table 3). The AUC for model 4 was 0.67 (95%CI 0.66 – 0.68,) also showing moderate discrimination. The predicted probability of perinatal asphyxia varied between 0.1% and 13%. The calibration plots for model 3 and 4 are shown in figure 2 and some overprediction on the higher probabilities.

8


141

Table

2A

. M

ultiv

aria

ble

asso

ciat

ions

bet

wee

n id

entifi

ed a

ntep

artu

m p

redi

ctor

s an

d in

terv

entio

n fo

r fai

lure

to p

rogr

ess.

FTP v

s sp

onta

neo

us

del

iver

yFD

vs

sponta

neo

us

del

iver

y

Chara

cter

istic

Reg

ress

ion c

oef

fici

ent

Odds

Ratio (

95

%CI)

Reg

ress

ion c

oef

fici

ent

Odds

Ratio (

95

%CI)

Inte

rcep

t-1

0.52

314

-11.

0200

0

Mate

rnal a

ge

(yea

rs)

0.04

9881

941.

05 (1

.04

- 1.0

6)0.

0454

2156

1.05

(1.0

2 - 1

.05)

Ges

tational a

ge

(wee

ks

days)

0.01

8813

851.

02 (1

.02

- 1.0

2)0.

0183

2474

1.02

(1.0

2 - 1

.04)

Non-

Cauca

sian e

thnic

ity

0.00

6047

311

1.01

(0.9

9 - 1

.02)

0.18

6949

442

1.20

(1.1

8 - 1

.23)

Low

soci

o-e

conom

ic s

tatu

s-0

.035

8889

00.

96 (0

.95

- 0.9

8)0.

0782

0327

1.08

(1.0

6 - 1

.10)

Nulli

paro

us

2.20

8184

9.10

(8.9

8 - 9

.22)

1.85

8210

6.41

(6.3

1 - 6

.52)

Pre

vious

caes

are

an s

ection

2.54

1239

12.7

(12.

4 - 1

3.0)

2.31

4297

10.1

(9.8

7 - 1

0.4)

Hyper

tensi

ve d

isea

se0.

4502

235

1.57

(1.5

4 - 1

.60)

0.76

5149

52.

15 (2

.11

- 219

)

Dia

bet

es M

ellit

us

0.53

5030

61.

71 (1

.63

- 1.7

9)0.

6178

490

1.85

(1.7

5 - 1

.95)

Art

ifici

al r

epro

duct

ion t

echniq

ue

0.52

9853

61.

70 (1

.67

- 1.7

2)0.

7835

514

2.19

(2.1

5 - 2

.23)

SGA

 P

95

0.91

3873

82.

49 (2

.45

- 2.5

4)-0

.151

9161

0.86

(0.8

3 - 0

.89)

Coagula

tion d

isord

er0.

1356

297

1.13

(1.1

5 - 1

.15)

0.28

4855

91.

33 (1

.32

- 1.3

4

Vagin

al b

leed

ing

0.45

8941

71.

58 (1

.47

- 1.7

0)0.

6411

474

1.90

(1.8

2 - 1

.99)

Neo

nata

l fem

ale

gen

der

0.27

1028

11.

31 (1

.30

- 1.3

3)0.

3740

285

1.45

(1.4

3 - 1

.47)

Pre

vious

IUFD

0.31

1502

11.

36 (1

.36

- 1.3

6)0.

7453

951

2.11

(2.1

0 - 2

.12)

Pre

vious

pre

term

bir

th-0

.158

9169

0.85

(0.6

2 - 1

.17)

0.30

6947

31.

36 (1

.36

- 1.3

6)

Chapter 8

142

Table

2B.

Mul

tivar

iabl

e as

soci

atio

ns b

etw

een

iden

tified

ant

epar

tum

and

intra

partu

m p

redi

ctor

s an

d in

terv

entio

n fo

r fai

lure

to p

rogr

ess

and

susp

ecte

d fe

tal d

istre

ss.

FTP v

s sp

onta

neo

us

del

iver

yFD

vs

sponta

neo

us

del

iver

y

Chara

cter

istic

Reg

ress

ion c

oef

fici

ent

Odds

Ratio (

95

%CI)

Reg

ress

ion c

oef

fici

ent

Odds

Ratio (

95

%CI)

Inte

rcep

t-1

0.34

096

-10.

5998

4

Mate

rnal a

ge

(yea

rs)

0.04

2420

301.

04 (1

.04

- 1.0

4)0.

0392

9343

1.04

(1.0

3 - 1

.05)

Ges

tational a

ge

(wee

ks

days)

0.01

7665

361.

02 (1

.02

- 1.0

2)0.

0162

3423

1.02

(1.0

2 - 1

.02)

Non-

Cauca

sian e

thnic

ity

-0.1

0123

775

0.91

(0.9

0 - 0

.93)

0.08

4614

731.

09 (1

.06

- 1.1

2)

Low

soci

o-e

conom

ic s

tatu

s-0

.054

4378

50.

95 (0

.94

- 0.9

6)0.

0448

3646

1.05

(1.0

3 - 1

.06)

Nulli

paro

us

2.09

1583

8.10

(7.9

8 - 8

.22)

1.82

4647

6.29

(6.1

8 - 6

.40)

Pre

vious

caes

are

an s

ection

2.19

3282

9.05

(8.8

5 - 9

.26)

1.92

5900

6.98

(6.8

0 - 7

.16)

Hyper

tensi

ve d

isea

se0.

0846

234

1.08

(1.0

6 - 1

.10)

0.18

6484

21.

20 (1

.18

- 1.2

3)

Dia

bet

es M

ellit

us

0.21

6373

11.

24 (1

.17

- 1.3

0)0.

1267

452

1.13

(1.0

6 - 1

.20)

Art

ifici

al r

epro

duct

ion t

echniq

ue

0.22

9224

1.28

(1.2

6 - 1

.30)

0.42

1376

1.54

(1.5

1 - 1

.57)

SGA

 P

95

0.83

8186

52.

30 (2

.25

- 2.3

4)-0

.231

8634

0.79

(0.7

7 - 0

.82)

Coagula

tion d

isord

er-0

.137

7361

50.

87 (0

.86

- 0.8

8)-0

.062

0246

90.

94 (0

.93

- 0.9

5)

Neo

nata

l fem

ale

gen

der

0.26

7423

61.

31 (1

.29

- 1.3

2)0.

3850

771

1.47

(1.4

5 - 1

.49)

Vagin

al b

leed

ing

0.20

9885

31.

24 (1

.16

- 1.3

3)0.

2707

882

1.31

(1.2

5 - 1

.38)

Pre

vious

IUFD

0.02

8116

051.

02 (1

.01

- 1.0

3)0.

3761

2915

1.45

(1.4

5 - 1

.45)

Pre

vious

pre

term

bir

th-0

.201

5564

0.81

(0.8

0 - 0

.81)

0.26

9684

21.

30 (1

.30

- 1.3

1)

Seco

ndary

care

at

onse

t of

labor

0.44

5148

61.

61 (1

.59

- 1.6

3)0.

8060

317

2.26

(2.2

2 - 2

.30)

Induct

ion o

f la

bor

0.07

5556

991.

10 (0

.92

- 1.3

1)0.

3791

2044

1.48

(1.4

5 - 1

.50)

Ruptu

red m

embra

nes

> 2

4h

0.18

3716

031.

23 (1

.20

- 1.2

5)0.

0950

4392

1.11

(1.0

8 - 1

.14)

Ox

yto

cin a

ugm

enta

tion

1.23

2697

03.

74 (3

.69

- 3.7

8)0.

7648

116

2.21

(2.1

8 - 2

.25)

Epid

ura

l anes

thes

ia-0

.718

2507

0.42

(0.4

1 - 0

.43)

-0.2

3363

280.

75 (0

.74

- 0.7

7)

Mec

oniu

m s

tain

ed a

mnio

tic

fluid

0.11

8591

91.

14 (1

.12

- 1.1

6)0.

8275

009

2.30

(2.2

5 - 2

.34)


8


143

Figure 1. Calibration graphs of the antepartum model (model 1 (A)) and the ante- plus intrapartum model

(model 2 (B)) for an obstetric intervention for failure to progress.

Chapter 8

144

Table

3.

Mul

tivar

iabl

e as

soci

atio

ns b

etw

een

iden

tified

pre

dict

ors

and

perin

atal

asp

hyxi

a; a

ntep

artu

m a

nd a

ntep

artu

m/

intra

partu

m m

odel

.

AN

TEPA

RTU

MA

NTE

AN

D IN

TRA

PARTU

M

Chara

cter

istic

Reg

ress

ion c

oef

fici

ent

Odds

ratio (

95

%CI)

Reg

ress

ion c

oef

fici

ent

Odds

ratio (

95

%CI)

Inte

rcep

t-7

.971

5-7

.466

2

Mate

rnal a

ge

(yea

rs)

0.01

921.

14 (1

.11

- 1.1

7)0.

0162

1.13

(1.1

0 - 1

.16)

Ges

tational a

ge

(wee

ks

days)

0.00

631.

10 (1

.07

- 1.1

2)0.

0044

1.07

(1.0

4 - 1

.09)

Non-

Cauca

sian e

thnic

ity

0.33

311.

40 (1

.33

- 1.4

6)0.

2835

1.32

(1.2

6 - 1

.39)

Low

soci

o-e

conom

ic s

tatu

s0.

1447

1.16

(1.1

1 - 1

.20)

0.12

941.

13 (1

.09

- 1.1

8)

Nulli

paro

us

0.72

792.

07 (1

.99

- 2.1

5)0.

6249

1.84

(1.7

6 - 1

.91)

Pre

vious

caes

are

an s

ection

0.68

321.

98 (1

.85

- 2.1

2)0.

5171

1.69

(1.5

7 - 1

.85)

Hyper

tensi

ve d

isea

se0.

4285

1.53

(1.4

6 - 1

.62)

0.21

641.

25 (1

.18

- 1.3

2)

Dia

bet

es m

ellit

us

0.55

211.

74 (1

.52

- 1.9

8)0.

3651

1.45

(1.2

7 - 1

.66)

Art

ifici

al r

epro

duct

ion t

echnolo

gy

0.33

801.

40 (1

.33

- 1.4

8)0.

1540

1.15

(1.0

9 - 1

.22)

SGA

 P

95

0.40

621.

50 (1

.41

- 1.6

0)0.

3639

1.45

(1.3

6 - 1

.55)

Coagula

tion d

isord

er0.

4253

1.53

(1.0

7 - 2

.20)

0.32

321.

39 (0

.96

- 1.9

9)

Neo

nata

l fem

ale

gen

der

0.25

390.

78 (0

.75

- 0.8

0)0.

2543

0.77

(0.7

5 - 0

.80)

Vagin

al b

leed

ing

0.50

321.

65 (1

.28

- 2.1

3)0.

3708

1.45

(1.1

2 - 1

.86)

Pre

vious

IUFD

0.47

001.

60 (1

.13

- 2.2

7)0.

3151

1.37

(0.9

7 - 1

.95)

Seco

ndary

care

at

onse

t of

labor

XX

0.27

211.

28 (

1.22

- 1.

34)

Induct

ion o

f la

bor

XX

0.18

561.

19 (1

.12

- 1.2

5)

Ruptu

red m

embra

nes

> 2

4h

XX

0.14

891.

14 (1

.08

- 1.2

2)

Ox

yto

cin a

ugm

enta

tion

XX

0.27

921.

25 (1

.20

- 1.3

0)

Epid

ura

l anes

thes

iaX

X0.

2882

1.48

(1.3

9 - 1

.57)

Mec

oniu

m s

tain

ed a

mnio

tic

fluid

XX

0.59

091.

79 (1

.70

- 1.8

8)

8


145

Table

3.

Mul

tivar

iabl

e as

soci

atio

ns b

etw

een

iden

tified

pre

dict

ors

and

perin

atal

asp

hyxi

a; a

ntep

artu

m a

nd a

ntep

artu

m/

intra

partu

m m

odel

.

AN

TEPA

RTU

MA

NTE

AN

D IN

TRA

PARTU

M

Chara

cter

istic

Reg

ress

ion c

oef

fici

ent

Odds

ratio (

95

%CI)

Reg

ress

ion c

oef

fici

ent

Odds

ratio (

95

%CI)

Inte

rcep

t-7

.971

5-7

.466

2

Mate

rnal a

ge

(yea

rs)

0.01

921.

14 (1

.11

- 1.1

7)0.

0162

1.13

(1.1

0 - 1

.16)

Ges

tational a

ge

(wee

ks

days)

0.00

631.

10 (1

.07

- 1.1

2)0.

0044

1.07

(1.0

4 - 1

.09)

Non-

Cauca

sian e

thnic

ity

0.33

311.

40 (1

.33

- 1.4

6)0.

2835

1.32

(1.2

6 - 1

.39)

Low

soci

o-e

conom

ic s

tatu

s0.

1447

1.16

(1.1

1 - 1

.20)

0.12

941.

13 (1

.09

- 1.1

8)

Nulli

paro

us

0.72

792.

07 (1

.99

- 2.1

5)0.

6249

1.84

(1.7

6 - 1

.91)

Pre

vious

caes

are

an s

ection

0.68

321.

98 (1

.85

- 2.1

2)0.

5171

1.69

(1.5

7 - 1

.85)

Hyper

tensi

ve d

isea

se0.

4285

1.53

(1.4

6 - 1

.62)

0.21

641.

25 (1

.18

- 1.3

2)

Dia

bet

es m

ellit

us

0.55

211.

74 (1

.52

- 1.9

8)0.

3651

1.45

(1.2

7 - 1

.66)

Art

ifici

al r

epro

duct

ion t

echnolo

gy

0.33

801.

40 (1

.33

- 1.4

8)0.

1540

1.15

(1.0

9 - 1

.22)

SGA

 P

95

0.40

621.

50 (1

.41

- 1.6

0)0.

3639

1.45

(1.3

6 - 1

.55)

Coagula

tion d

isord

er0.

4253

1.53

(1.0

7 - 2

.20)

0.32

321.

39 (0

.96

- 1.9

9)

Neo

nata

l fem

ale

gen

der

0.25

390.

78 (0

.75

- 0.8

0)0.

2543

0.77

(0.7

5 - 0

.80)

Vagin

al b

leed

ing

0.50

321.

65 (1

.28

- 2.1

3)0.

3708

1.45

(1.1

2 - 1

.86)

Pre

vious

IUFD

0.47

001.

60 (1

.13

- 2.2

7)0.

3151

1.37

(0.9

7 - 1

.95)

Seco

ndary

care

at

onse

t of

labor

XX

0.27

211.

28 (

1.22

- 1.

34)

Induct

ion o

f la

bor

XX

0.18

561.

19 (1

.12

- 1.2

5)

Ruptu

red m

embra

nes

> 2

4h

XX

0.14

891.

14 (1

.08

- 1.2

2)

Ox

yto

cin a

ugm

enta

tion

XX

0.27

921.

25 (1

.20

- 1.3

0)

Epid

ura

l anes

thes

iaX

X0.

2882

1.48

(1.3

9 - 1

.57)

Mec

oniu

m s

tain

ed a

mnio

tic

fluid

XX

0.59

091.

79 (1

.70

- 1.8

8)

(A)

(B)

Figure 2. Calibration graphs of the antepartum model (model 1 (A)) and intrapartum model (model 2 (B))

for perinatal asphyxia.


Chapter 8

146

Two-dimensional decision toolCombining the predicted probabilities of the antepartum prediction models for FTP (model 1) and perinatal asphyxia (model 3) into a two-dimensional prediction graph showed that the probability of an intervention due to FTP increased with an increased risk of perinatal asphyxia. However, in some cases the risk of FTP is high while the risk of asphyxia is low and vice versa (Figure 3). The graph was constructed to show the predicted probabilities of suspected perinatal asphyxia versus the predicted probabilities of an intervention due to FTP and the number of women within each 10th percentile combination of both outcomes. Clinicians could use the two-dimensional graph in counseling pregnant women before the onset of labor. In Figure 3 we marked the mean observed proportion of an intervention for FTP and for perinatal asphyxia for both nulliparous and multiparous women. For example the upper right quadrant is showing pregnant women with a antepartum predicted probability of perinatal asphyxia between 1.5% and 5.0% and a predicted probability of failure to progress between 10% and 90%. With this predicted high risk for both outcomes is it arguable for multiparous women to not undergo a trial of labor.

In Figure 4 we present the predicted probabilities of the integrated antepartum and intrapartum models for intervention for FTP (model 2) and perinatal asphyxia (model 4). As could be concluded from the more widespread dots in the graph, intrapartum risks are more divers.


8


147

Figure 3. Predicted antepartum probability of perinatal asphyxia versus obstetric intervention due to failure

to progress.

M: mean observed proportion for multiparous women.

Intervention failure to progress: 3.4%, Perinatal asphyxia 0.55%

N: mean observed proportion for nulliparous women.

Intervention failure to progress: 20%, Perinatal asphyxia 1.2%

Figure 4. Predicted antepartum and intrapartum probability of perinatal asphyxia versus obstetric

intervention due to failure to progress.





Chapter 8

148

In Figure 5 and Figure 6 we present the predicted probabilities of the sensitivity analyses of adverse neonatal outcome (child) versus the predicted probabilities of an obstetric intervention due to failure to progress (mother). 109, 245 (7%) woman had adverse neonatal outcome, i.e. emergency delivery due to suspected fetal distress and/or a 5-minute Apgar score < 7. The antepartum and intrapartum characteristics to predict the combined adverse neonatal outcome are comparable to the prognostic factors of perinatal asphyxia. (see appendix 3). The AUC for the antepartum model was 0.74 (0.73 -0.74 95%CI) and for the integrated antepartum and intrapartum model the AUC was 0.77 (0.76 – 0.78 95%CI).

The predicted probabilities for adverse neonatal outcome ranged from 0.49 to 84%. Again we marked the mean observed proportion for an adverse neonatal outcome versus the mean observed proportion of an intervention for FTP for both nulliparous and multiparous women. The lower left quadrant is showing laboring women with a predicted probability of adverse neonatal outcome of 0.40 – 4% and a predicted probability of failure to progress below 5%. One could imagine that if, for a certain nulliparous women the probability of FTP is very low (more than half the mean observed proportion for nulliparous women) the clinician is allowed to proceed spontaneous delivery (with strict fetal surveillance) as long the baby does not appear to be in imminent distress and a cesarean delivery can be prevented.


8


149

Figure 5. Predicted antepartum probability of adverse neonatal outcome versus obstetric intervention due

to failure to progress.





Adverse neonatal outcome: Perinatal asphyxia and / or intervention for suspected fetal distress

Figure 6. Predicted antepartum and intrapartum probability of adverse neonatal outcome versus obstetric

intervention due to failure to progress.





Adverse neonatal outcome: Perinatal asphyxia and / or intervention for suspected fetal distress


Chapter 8

150

disCussion

Principle findingsIn this study we developed and internally validated four prognostic models, two for predicting obstetric intervention due to failure to progress and two for perinatal asphyxia. The models are based on a combination of ante partum characteristics and combined antepartum and intrapartum characteristics and showed moderate to good calibration and discrimination. Based on the prediction models, we developed a two-dimensional graphical tool, in which for each laboring woman the predictions for failure to progress and perinatal asphyxia were expressed as coordinates on a two-dimensional plane. Overall, the chance of an intervention due to failure to progress increased with an increased risk of perinatal asphyxia. This tool could assist clinicians in the assessment of the risks of adverse pregnancy and neonatal outcomes and guide subsequent interventions during labor and delivery.

Strengths and limitationsThe existing prediction models for failure to progress and perinatal asphyxia (or adverse neonatal outcome) were developed on a cohort of Dutch women with high-risk pregnancies. 10, 11 However, a study on the limitations in the clinical prediction of intrapartum perinatal asphyxia showed that a significant proportion of intrapartum fetal asphyxia occurred in pregnancies with no risk factors or pregnancies deemed to be low risk. 13 Also, a 2013 study in the US found that 10% of all first-time mothers with otherwise normal pregnancies had a cesarean section for failure to progress. 27 These findings suggest that there is a need for predictive models for both perinatal asphyxia and failure to progress that can be applied not only to high-risk pregnancies but to low-risk pregnancies as well. One of the strengths of this study is that our prediction models take this into account as we used a 10 year national dataset from Perinatal Registry of the Netherlands, which included 96%l low- and high-risk singleton pregnancies between 37+0 to 41+6 completed weeks of gestation so we expect the tool to be generalizable in both primary and secondary care settings, i.e. high and low-risk pregnancies. Another strength of this study is that the validity of the new prediction models is supported by the fact that we used such a large sample size with a significant number of events per variable. The general rule of thumb suggests 5, 10 or even 20 events per variable to achieve reliable predictor-outcome associations. 28

There are also some limitations to this study. In the PRN perinatal asphyxia is defined as an infant having an Apgar score of less than 7 at 5 minutes of life. However, some studies have shown that low 5-minute Apgar scores alone are not conclusive markers of an intrapartum hypoxic event. 29 Low Apgar scores may be evidence of prior asphyxia but may also be due to a host of other causes, ranging from the effects of maternal


8


151

medication, to congenital abnormalities of the fetus. 30 However, using a cohort of only term singletons without congenital anomalies we do not believe this influences our analysis, as in this subgroup of newborns a low 5-minute Apgar score is caused generally by intra-uterine asphyxia. 31-33 Another limitation of this study is that potential predictors like the Bishop score, body mass index and maternal smoking were not taken into account because there was no information on these characteristics in the national registry database. 14

Comparing the outcomes to the common literatureThe existing published model for perinatal asphyxia had 6 antepartum and intrapartum characteristics.10. In this study, the new model for perinatal asphyxia, using both antepartum and intrapartum variables, was developed using 21 characteristics (including all 6 from the previous model). The performance measurements of our models are comparable to the existing model, underlining the complexity of the predicting of perinatal asphyxia. The existing model for failure to progress used a total of 13 characteristics 11 while our new model for failure to progress, using 22 characteristics (including 11 from the previous model). While it could be argued that the new models are oversaturated with too many variables some of which occur very infrequently such as previous intra-uterine fetal death (IUFD) and previous preterm birth, we chose to use the backward selection on a model with all candidate predictors to avoid over fitting and selection bias, and thus kept all the variables that remained after the backward selection. Also, while their incidence might be negligible, studies have demonstrated a significantly increased risk of perinatal asphyxia in woman with a history of an IUFD and/or preterm pregnancy. 10, 11, 27 Furthermore, the previous prognostic models assessed the risk of these outcomes in only high-risk women and were developed separately but not combined. But as both outcomes can occur at the same time we feel that there is a benefit to integrating both dimensions into a single two-dimensional prognostic graph as we did in our study.

Clinical implicationsThe ultimate goal for every pregnant woman is the birth of is a healthy child in the less harmful way for herself. The characteristics used in the development of the prediction models are readily available in clinical practice and are generally registered in national registries. Thus, integrating the two-dimensional tool into the Electronic Medical Record (EMR) would make it inexpensive and easy to apply the tool before or shortly after the onset of labor. Employing such a model, in conjunction with other methods of corroborating the predictions of the fetal and maternal risk assessments made by the tool, could subsequently guide clinicians in the choice of interventions during labor and delivery.


Chapter 8

152

Besides this timely prognostication, it could also aid in organizational aspects of the delivery ward (e.g. the presence of personnel who can perform an obstetrical intervention and availability of operating theatre). We found that the addition of intrapartum variables did marginally influence the prediction of non-progressive labor, while the impact of the intrapartum predictors on the prediction of asphyxia was negligible. These results suggesting that clinicians are able to predict the laboring ‘chances’ before the onset of labor, and allow adequate counseling of the pregnant women in the more peaceful antepartum period. Using the two-dimensional decision tool, it is easily visualized what the chances of a laboring women are compared to other nulliparous or multiparous women. In case a multiparous woman with probabilities for both an intervention for FTP and risk of adverse neonatal outcome far above the mean for multiparous women, one could question if these women should undergo trial of labor at all. On the other hand, for nulliparous women it is important to prevent an unnecessary caesarean delivery, as these interventions has long-term consequences for her subsequent pregnancy.

In conclusion, adverse pregnancy outcomes such as failure to progress (mother) and perinatal asphyxia (child) can be attributed to events during labor and delivery but may also have their origins in the antepartum period. As both outcomes can occur simultaneously it could be of great benefit to clinicians to have a two-dimensional decision graph that provides information on both the risk for an intervention for FTP as well as perinatal asphyxia in order to accurate counseling of pregnant women and guiding labor management. After external validation and proof of generalizability, this model could eventually be used in high- and low-risk pregnancies in obstetric care settings outside the Netherlands.


8


153

referenCes

1. Shipp TD, Zelop CM, Repke JT, Cohen A, Caughey AB, Lieberman E. Labor after previous cesarean: influence of prior indication and parity. Obstetrics and gynecology 2000;95:913-6.

2. Beard RW, Rivers RP. Fetal asphyxia in labour. Lancet 1979;2:1117-9.

3. Cheng YW, Hopkins LM, Laros RK, Jr., Caughey AB. Duration of the second stage of labor in multiparous women: maternal and neonatal outcomes. American journal of obstetrics and gynecology 2007;196:585 e1-6.

4. Allen VM, Baskett TF, O’connell CM, Mckeen D, Allen AC. Maternal and perinatal outcomes with increasing duration of the second stage of labor. Obstetrics and gynecology 2009;113:1248-58.

5. Dalili H, Nili F, Sheikh M, Hardani AK, Shariat M, Nayeri F. Comparison of the four proposed Apgar scoring systems in the assessment of birth asphyxia and adverse early neurologic outcomes. PloS one 2015;10:e0122116.

6. Aslam S, Molloy EJ. Biomarkers of multiorgan injury in neonatal encephalopathy. Biomarkers in medicine 2015;9:267-75.




10. Westerhuis ME, Schuit E, Kwee A, et al. Prediction of neonatal metabolic acidosis in women with a singleton term pregnancy in cephalic presentation. American journal of perinatology 2012;29:167-74.

11. Schuit E, Kwee A, Westerhuis ME, et al. A clinical prediction model to assess the risk of operative delivery. BJOG : an international journal of obstetrics and gynaecology 2012;119:915-23.

12. Low JA, Pickersgill H, Killen H, Derrick EJ. The prediction and prevention of intrapartum fetal asphyxia in term pregnancies. American journal of obstetrics and gynecology 2001;184:724-30.


Chapter 8

154

13. Low JA, Simpson LL, Tonni G, Chamberlain S. Limitations in the clinical prediction of intrapartum fetal asphyxia. American journal of obstetrics and gynecology 1995;172:801-4.

14. Gopalani S, Bennett K, Critchlow C. Factors predictive of failed operative vaginal delivery. American journal of obstetrics and gynecology 2004;191:896-902.

15. Peregrine E, O’brien P, Omar R, Jauniaux E. Clinical and ultrasound parameters to predict the risk of cesarean delivery after induction of labor. Obstetrics and gynecology 2006;107:227-33.

16. Heinonen S, Saarikoski S. Reproductive risk factors of fetal asphyxia at delivery: a population based analysis. Journal of clinical epidemiology 2001;54:407-10.

17. Barkovich AJ, Truwit CL. Brain damage from perinatal asphyxia: correlation of MR findings with gestational age. AJNR American journal of neuroradiology 1990;11:1087-96.


19. Park KH. Transvaginal ultrasonographic cervical measurement in predicting failed labor induction and cesarean delivery for failure to progress in nulliparous women. Journal of Korean medical science 2007;22:722-7.


21. Biesheuvel CJ, Vergouwe Y, Steyerberg EW, Grobbee DE, Moons KG. Polytomous logistic regression analysis could be applied more often in diagnostic research. Journal of clinical epidemiology 2008;61:125-34.

22. Harrell FE, Jr., Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Statistics in medicine 1996;15:361-87.

23. Sun GW, Shook TL, Kay GL. Inappropriate use of bivariable analysis to screen risk factors for use in multivariable analysis. Journal of clinical epidemiology 1996;49:907-16.

24. Atkinson A. A note on the generilzed information on criterion for choice of a model. Biometrika 2010;67:413-8.

25. Miller ME, Hui SL, Tierney WM. Validation techniques for logistic regression models. Statistics in medicine 1991;10:1213-26.

26. Collins SG, Et Al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement. The Britisch Medical Journal 2014;350.


8


155

27. Boyle A, Reddy UM, Landy HJ, Huang CC, Driggers RW, Laughon SK. Primary cesarean delivery in the United States. Obstetrics and gynecology 2013;122:33-40.

28. Courvoisier DS, Combescure C, Agoritsas T, Gayet-Ageron A, Perneger TV. Performance of logistic regression modeling: beyond the number of events per variable, the role of data structure. Journal of clinical epidemiology 2011;64:993-1000.


30. Freeman JM, Nelson KB. Intrapartum asphyxia and cerebral palsy. Pediatrics 1988;82:240-9.





Chapter 8

156

Appen

dix

1.

Uni

varia

ble

asso

ciat

ions

bet

wee

n id

entifi

ed p

redi

ctor

s of

inte

rven

tion

for f

ailu

re to

pro

gres

s or

sus

pect

ed fe

tal d

istre

ss.

FTP v

s Sp

onta

neo

us

vagin

al

FD v

s. S

ponta

neo

us

vagin

al

Chara

cter

istic

Reg

ress

ion c

oef

fici

ent

Odds

ratio

Reg

ress

ion c

oef

fici

ent

Odds

ratio (

95

%CI)

Mate

rnal a

ge

(yea

rs)

- 0.0

0372

6710

0.90

(0.8

8 - 0

.91)

- 0.0

0381

7541

1.07

(1.0

5 - 1

.09)

Ges

tational a

ge

(wee

ks

days)

- 0.0

1592

674

1.02

(1.0

2 - 1

.02)

- 0.0

1307

512

1.01

(1.0

1 - 1

.01)

Non-C

auca

sian e

thnic

ity

-0.2

8545

2895

40.

75 (0

.74

- 0.7

6)-0

.000

7225

262

1.00

(0.9

8 - 1

.02)

Low

soci

o-e

conom

ic s

tatu

s-0

.102

3665

50.

90 (0

.98

- 0.9

1)0.

0845

5502

1.09

(1.0

7 - 1

.10)

Nulli

paro

us

1.73

0638

5.65

(5.5

8 - 5

.71)

1.48

8397

4.43

(4.3

7 - 4

.49)

Pre

vious

cesa

rean s

ection

1.57

5978

4.83

(4.7

5 - 4

.92)

1.49

1172

4.44

(4.3

4 - 4

.45)

Hyper

tensi

ve d

isea

se0.

6898

975

1.99

(1.9

6 - 2

.03)

1.00

1274

22.

72 (2

.67

- 2.7

7)

Dia

bet

es m

ellit

us

0.52

4771

41.

69 (1

.62

- 1.7

7)0.

5183

034

1.68

(1.5

9 - 1

.78)

Art

ifici

al r

epro

duct

ion t

echniq

ue

0.65

4853

11.

92 (1

.89

- 1.9

6)0.

8951

874

2.45

(2.4

0 - 2

.49)

SGA

P5

0.78

4521

00.

46 (0

.44

- 0.4

7)0.

9346

314

2.55

(2.4

9 - 2

.60)

LGA

P5

0.90

3147

72.

47 (2

.43

- 2.5

1)-0

.250

9473

0.78

(0.7

5 - 0

.80)

Ute

rus

myom

ato

sis

0.56

6600

31.

76 (1

.53

- 2.0

3)0.

5665

425

1.76

(1.4

8 - 2

.10)

Coagula

tion d

isord

er0.

1173

412

1.12

(1.0

0 - 1

.27)

0.27

0919

81.

31 (1

.13

- 1.5

2)

Endocr

inolo

gic

al d

isea

se- 0

.020

0103

410.

98 (0

.86

- 1.1

2)0.

0065

7263

91.

01 (0

.85

- 1.2

0)

Neo

nata

l fem

ale

gen

der

0.22

4696

41.

25 (1

.24

- 1.2

6)0.

3274

976

1.39

(1.3

7 - 1

.41)

Vagin

al b

leed

ing

0.42

5118

31.

53 (1

.41

- 1.6

6)0.

6397

930

1.90

(1.7

2 - 2

.09)

Pre

vious

IUFD

-0.8

1047

914

0.44

(0.3

8 - 0

.52)

-0.0

8401

521

0.92

(0.8

0 - 1

.06)

Pre

vious

IUG

R-1

.420

5954

0.24

(0.1

8 - 0

.32)

-0.1

5453

210.

86 (0

.70

- 1.0

5)


8


157

Pre

vious

pre

term

bir

th-0

.928

7657

0.40

(0.3

5 - 0

.45)

-0.2

1291

490.

81 (0

.71

- 0.9

1)

Labor

that

start

s in

sec

ondary

care

0.60

7283

31.

84 (1

.82

- 1.8

5)1.

0379

514

2.82

(2.7

9 - 2

.86)

Induct

ion o

f la

bor

0.37

8095

21.

46 (1

.44

- 1.4

8)0.

8712

997

2.39

(2.3

5 - 2

.43)

Ruptu

red m

embra

nes

> 2

4h

0.78

3065

92.

19 (2

.15

- 2.2

2)0.

4673

715

1.60

(1.5

6 - 1

.63)

Ox

yto

cin a

ugm

enta

tion

1.61

5337

5.03

(4.9

8 - 5

.08)

1.26

0662

3.53

(3.4

8 - 3

.58)

Non-E

pid

ura

l anes

thes

ia

(morp

hin

e)0.

9025

485

1.56

(1.5

4 - 1

.58)

0.65

8929

51.

90 (1

.93

- 1.9

7)

Epid

ura

l anes

thes

ia0.

5126

652

1.67

(1.6

4 - 1

.70)

0.44

5751

12.

47 (2

.42

- 2.5

2)

Mec

oniu

m s

tain

ed a

mnio

tic

fluid

0.36

6394

11.

44 (1

.42

- 1.4

7)0.

8398

541

2.31

(2.2

8 - 2

.36)

Chapter 8

158

Appendix 2. Univariable associations between identified predictors of perinatal asphyxia.

Characteristic Regression coefficient

Odds Ratio (95%CI)

Maternal age (years) -0.0048 0.97 (0.94 - 0.99)

Gestational age (weeks days) 0.0038 1.06 (1.03 - 1.08)

Non-Caucasian ethnicity 0.2968 1.35 (1.29 - 1.41)

Low socio-economic status 0.2235 1.25 (1.20 - 1.30)

Nulliparous 0.6608 1.94 (1.87 - 2.01)

Previous caesarean section 0.5165 1.68 (1.56 - 1.80)

Pre-existing hypertension 0.2176 1.24 (0.99 - 1.57)

Pre-existing diabetes 1.0999 3.04 (2.33 - 3.87)

Hypertensive disease 0.5872 1.80 (1.71 - 1.89)

Diabetes mellitus 0.6856 1.98 (1.74 - 2.26)

Subfertility -0.2832 0.75 (0.47 - 1.21)

Artificial reproduction technique 0.4538 1.57 (1.49 - 1.66)

SGA < P5 0.8804 2.41 (2.27 - 2.56)

LGA > P95 0.4470 1.56 (1.46 - 1.67)

Uterus myomatosis 0.3442 1.41 (0.86 - 2.31)

Coagulation disorder 0.4194 1.52 (1.06 - 2.18)

Endocrinological disease 0.0378 1.04 (0.65 - 1.65)

Neonatal female gender 0.2549 0.77 (0.75 - 0.80)

Vaginal bleeding 0.5492 1.73 (1.34 - 2.23)

Previous IUFD 0.1516 1.16 (0.82 - 1.65)

Previous IUGR -0.3589 0.70 (0.36 - 1.35)

Previous very preterm birth 0.1027 1.11 (0.73 - 1.69)

Previous preterm birth -0.1417 0.87 (0.62 - 1.22)

Labor that starts in secondary care 0.4972 1.64 (1.59 - 1.70)

Induction of labor 0.4964 1.64 (1.57 - 1.72)

Ruptured membranes > 24h 0.3348 1.40 (1.32 - 1.48)

Oxytocin augmentation 0.006405 1.01 (1.01 - 1.02)

Non-Epidural anesthesia (morphine) 0.6371 1.89 (1.81 - 1.98)

Epidural anesthesia 0.8121 2.25 (2.13 - 2.38)

Meconium stained amniotic fluid 0.6303 1.88 (1.79 - 1.97)

8


159

Appen

dix

3 M

ultiv

aria

ble

asso

ciat

ions

bet

wee

n id

entifi

ed p

redi

ctor

s an

d ad

vers

e ne

onat

al o

utco

me.

AN

TEPA

RTU

MA

NTE

+ IN

TRA

PARTU

M

Chara

cter

istic

Reg

ress

ion c

oef

fici

ent

Odds

ratio (

95

%C)

Reg

ress

ion c

oef

fici

ent

Odds

ratio

(95

%C)

Inte

rcep

t-8

.803

2-8

.227

3

Mate

rnal a

ge

(yea

rs)

0.03

261.

26 (1

.24

- 1.2

7)0.

0274

1.21

(1.2

0 - 1

.23)

Ges

tational a

ge

(wee

ks

days)

0.01

341.

22 (1

.21

- 1.2

3)0.

0111

1.18

(1.1

7 - 1

.19)

Non-

Cauca

sian e

thnic

ity

0.19

071.

21 (1

.19

- 1.2

3)0.

1168

1.12

(1.1

0 - 1

.14)

Low

soci

o-e

conom

ic s

tatu

s0.

0906

1.09

(1.0

8 - 1

.11)

0.06

361.

06 (1

.05

- 1.0

8)

Nulli

paro

us

1.41

864.

13 (4

.07

- 4.1

9)-1

.373

93.

94 (3

.88

- 4.0

0)

Pre

vious

cesa

rean s

ection

1.55

204.

72 (4

.61

- 4.8

3)1.

2337

3.44

(3.3

6 - 3

.52)

Hyper

tensi

ve d

isea

se0.

6374

1.89

(1.8

6 - 1

.93)

0.15

141.

16 (1

.14

- 1.1

9)

Dia

bet

es m

ellit

us

0.49

771.

64 (1

.56

- 1.7

4)0.

3622

1.09

(1.0

3 - 1

.16)

Art

ifici

al r

epro

duct

ion t

echnolo

gy

0.63

911.

89 (1

.86

- 1.9

3)0.

3426

1.41

(1.3

8 - 1

.43)

SGA

 P

95

-0.3

399

0.71

(0.6

9 - 0

.73)

-0.4

221

0.66

(0.6

4 - 0

.68)

Coagula

tion d

isord

er0.

2735

1.31

(1.1

4 - 1

.52)

-0.0

116

0.99

(0.8

6 - 1

.14)

Neo

nata

l fem

ale

gen

der

0.31

110.

73 (0

.72

- 0.7

4)0.

3175

0.73

(0.7

2 - 0

.74)

Vagin

al b

leed

ing

0.52

091.

68 (1

.53

- 1.8

5)0.

2052

1.23

(1.1

1 - 1

.35)

Pre

vious

IUFD

0.68

081.

98 (1

.73

- 2.2

6)0.

3707

1.45

(1.2

7 - 1

.66)

Pre

vious

pre

term

bir

th0.

3338

1.40

(1.2

3 - 1

.58)

0.29

191.

34 (1

.18

- 1.5

2)

Seco

ndary

care

at

onse

t of

labor

XX

0.70

592.

02 (1

.98

- 2.0

5)

Induct

ion o

f la

bor

XX

0.32

881.

39 (1

.36

- 1.4

1)

Ox

yto

cin a

ugm

enta

tion

XX

0.43

511.

53 (1

.51

- 1.5

5)

Epid

ura

l anes

thes

iaX

X0.

0408

1.06

(1.0

4 - 1

.07)

Mec

oniu

m s

tain

ed a

mnio

tic

fluid

XX

0.78

812.

20 (2

.16

- 2.2

4)


Chapter 9 Summary and general discussion

Chapter I

162

summary

The work presented in this thesis focuses on perinatal asphyxia related morbidity and mortality among term singleton pregnancies. In this thesis we aimed:1. to study trends and risk factors for perinatal asphyxia at term, 2. to explore the practice variation in obstetric interventions among general hospitals

in the Netherlands,3. to develop a two-dimensional prognostic tool for the simultaneous prediction of risk

of perinatal asphyxia and risk of obstetric interventions for non-progressive labor.

In chapter 2, we described temporal trends in perinatal asphyxia at term in the Netherlands. The prevalence of perinatal asphyxia (Apgar score < 7) was 0.85% (1,285/151,680) and severe asphyxia 0.16% (262/151,680) (Apgar score < 4). Between 1999 and 2010 the risk of perinatal asphyxia decreased significantly with approximately 6%, this translates to a decrease of 80 neonates who suffer from asphyxia per year. And, although we observed stable rates in the total obstetric interventions (around 20%), the decline in perinatal asphyxia was accompanied by a significant increasing number of interventions for suspected fetal distress. The proportion of interventions for fetal distress increased from 5.9% to 7.7%. This trend towards increasing obstetric interventions for suspected fetal distress was most pronounced for Cesarean deliveries, a possible effect of improved fetal monitoring. One would expect that a rising trend in obstetric interventions for suspected fetal distress, together with a decrease in perinatal asphyxia at term, would contribute to a decrease in perinatal mortality for this cause. However, the contribution of perinatal asphyxia (87%) to the overall perinatal mortality did not change during the study period. The continuous and relative high risk of perinatal mortality within the first week after birth (0.98 per 1000 births) requires attention and underlines the need for improved risk assessment of pregnant women with high(er) risk for adverse pregnancy outcomes.

In chapter 3, we described the practice variation in elective and emergency caesarean rates among nulliparous women delivering a fetus in cephalic position at term in general hospitals in the Netherlands. Of the 458 712 infants born, 1.8% were delivered by elective caesarean delivery. Elective caesarean rates varied between 0.72% and 5.2% among the hospitals, whereas the emergency caesarean rates varied between 8.2% and 28%. Higher rates of elective caesarean sections could result in better neonatal outcomes, although hospitals performing few elective caesarean sections have no significant increased risk of adverse neonatal outcomes. Neonates born after intended vaginal delivery in hospitals with restraint in emergency caesareans have an increased risk for neonatal mortality and morbidity. This was mainly seen among infants who were delivered by caesarean (aOR 1.69, 95%CI 1.18 – 2.41 and

Summary and general discussion

9

163

aOR 1.63, 95%CI 1.31 – 2.12 respectively), suggesting there may be deficiencies in the timing or selection of cases during the delivery process.

In chapter 4, we examined the recurrence risk of perinatal asphyxia in the subsequent pregnancy. In current practice, obstetric history is used as an important determinant to assess the risks in subsequent pregnancies and influences clinical management. Risk for an Apgar score < 7 was 0.99% among first term singleton pregnancies and overall halved in the subsequent pregnancies (0.50%). It was shown that women with birth asphyxia of the first born have twice the risk of renewed asphyxia at the next birth compared to women without birth asphyxia of their first child. This increased risk was independent of the mode of delivery of the first born. Despite the low prevalence of asphyxia, an obstetric history of perinatal asphyxia and the interaction with other risk factors, like intra-uterine growth retardation, should be incorporated in the risk estimation for the subsequent pregnancy.

In chapter 5, we presented the risk of asphyxia related adverse neonatal outcomes after medically assisted reproduction (MAR). In this observational study cohort of over 45,000 children conceived after MAR, we found that MAR singletons had an increased risk of a 5 minute Apgar score < 4 compared to spontaneous conceived (SC) singletons (aOR 1.29, 95%CI 1.14 – 1.46). Because subfertility patients differ from fertile patients on numerous factor contributing to adverse neonatal and pregnancy outcomes, we used propensity score matching analysis to minimize the selection bias and maternal confounding factors. After propensity score matching the risk of adverse neonatal outcomes was similar between MAR infants and SC infants, suggesting that characteristics of subfertile women are associated with the increased prevalence of health problems in term MAR newborns. Until it is possible to identify women at risk for adverse neonatal outcomes, obstetric caregivers classify MAR pregnancies as potentially high risk. This contributes to high rates of obstetric interventions and their potential negative consequences for both mother and child.

Perinatal asphyxia plays a major role in perinatal mortality at term and in chapter 6 we presented the most important risk factors for this adverse neonatal outcome. We found that nulliparity, non-western ethnicity, smoking, being small for gestational age SGA and medical conditions indicating high-risk pregnancy are the most important risk factors for asphyxia related mortality at term. Population Attributable Risk (PAR), based on the Relative Risk and Prevalence of a certain risk indicator, were calculated to investigate the potential improvement that could be reached when eliminating that certain risk factor. The highest PAR was found for SGA below 10th percentile (17%), nulliparity (20%) and smoking (16%). Also, prelabor classification of high-risk pregnancy based on medical features had a high PAR (39%). As in the Netherlands


Chapter I

164

women with low risk of pregnancy complications receive care in a separate primary care system, with less strict fetal surveillance, optimal antepartum risk assessment is of crucial importance to improve perinatal outcomes. In women who started labor in a primary care setting, the PAR of SGA below the 10th percentile was 24% as compared to 13% for women who started labor in secondary care. The contribution of SGA, particularly in women who started labor in primary care, show the potential improvement of adequate antenatal detection of SGA.

In chapter 7 we assessed the association of the five individual components of the Apgar score with neonatal mortality and morbidity among term singletons. Logistic regression and penalized variable selection showed that only heart rate score of zero (aOR 18, 95%CI 3.7 – 86) and respiratory effort with a score of zero (aOR 6.0, 95%CI 1.3 – 28) were independently associated with poor neonatal outcome in our study sample. And although the Apgar score is probably the most used prognostic test in obstetrics, we showed that a reduced model based on both these variables had comparable predictive performance in terms of discrimination as the current model.

In chapter 8, we described the development of a two-dimensional prediction tool, in which for each laboring woman the predictions for failure to progress (FTP) and perinatal asphyxia were assessed simultaneously. The predicted probabilities for an intervention due to failure to progress varied between 0.30% to nearly 100%. The intrapartum model consisted of 22 variables and showed good calibration and an AUC of 0.78 (95%CI 0.77 – 0.79). Multivariable regression analysis showed that, nulliparity, previous caesarean delivery, artificial reproduction techniques and hypertensive diseases are most strongly associated with interventions for FTP. The intrapartum model for the prediction of perinatal asphyxia consisted of 21 variables and showed a limited discrimination with an AUC of 0.67 (95%CI 0.66 – 0.68). Again previous caesarean delivery, non-Caucasian ethnicity and hypertensive diseases were identified as important predictors, together with fetuses being small for gestational age and. The predicted probabilities for perinatal asphyxia varied between 0.10% and 13%. Combining the predicted probabilities into a two-dimensional prediction model showed that the probability of an intervention due to FTP increased with an increased risk of perinatal asphyxia. However, in some cases the risk of FTP was high while the risk of perinatal asphyxia was low and vice versa. In current clinical practice, individual risk assessment and antenatal counseling of women at risk for adverse outcomes for both the mother and child become increasingly important. We believe that, after external validation, integrating the two-dimensional prediction graph in obstetric care, in conjunction with other methods of corroborating the predictions of adverse outcomes, could guide clinicians in the choice of interventions during labor and delivery.



9

165

General disCussion

Clinical implications and future researchIt is already known that certain maternal, fetal and pregnancy characteristics are associated with an increased risk of perinatal asphyxia and related pregnancy complications at term. We introduced some new insights in the interrelationship among these factors and provided a basis for more individualized risk assessment and counseling of pregnant women with respect to their risk of an adverse (asphyxia related) pregnancy outcome at term. On the other hand, we must conclude that perinatal asphyxia at term is still a difficult event to predict. We were not able to develop a robust prognostic model that can be immediately applied in clinical practice. This is partly due the absence of some relevant variables in the dataset of the Netherlands Perinatal Registry (PRN), for example the Bishop score at start of labor, maternal smoking habits and Body Mass Index (BMI), but moreover dedicated to the complexity of predicting rare events as perinatal asphyxia.

Risk factors for perinatal asphyxiaMaternalThe general health, physical condition and lifestyle factors of pregnant women influence her risk for complications during pregnancy and delivery. Women with a pre-existing medical condition, for example hypertension, are at increased risk of pregnancy complications and perinatal asphyxia due to these complications. 1 With the advancing age in which women get pregnant nowadays, the likelihood of pre-existing comorbidities and the need for medical assisted reproduction are increasing. 2 We have shown, using propensity score matching techniques, that the increased risk of impaired neonatal outcomes among term singletons conceived by MAR are at least partly explained by the maternal and parental characteristics of subfertile couples. The same problem of confounding factors applies to maternal ethnicity. Non–western ethnicity was previously described as a risk factor for adverse neonatal outcomes at term. 3 We confirmed this association for asphyxia related complications. However, we cannot distinguish whether the observed differences can be explained by genetic differences or by (un)known other confounding factors.Behavioral factors like smoking during pregnancy are strongly associated with adverse outcomes.4 Our results show the great potential of health care programs designed to reduce the number of pregnant women smoking, as this contributes to the decreased perinatal mortality at term. Smoking not only directly increases the risk of perinatal mortality, it also works as a mediator due to increased risk for intrauterine growth retardation (IUGR) among smoking women. 5


Chapter I

166

FetalThe increased risk of asphyxia related mortality of fetuses being small for gestational age (SGA), and the interplay between fetal growth and many maternal (hypertension), pregnancy (MAR), and lifestyle factors (smoking) underline the importance of early detection of IUGR. 6 However, consensual scientific evidence on the accuracy of detecting IUGR by ultrasound screening is lacking. At present, a nationwide study on prenatal detection of IUGR by routine third trimester ultrasound screening is being performed (The IRIS – study). 7 The outcome of this study might influence obstetricians’ behavior in the future.

Obstetric historyAfter a previous vaginal delivery at term, in general a pregnant woman is considered at low risk for pregnancy complications in the next birth. This also counts for women who delivered a baby suffering from perinatal asphyxia. According to our findings, in which women with birth asphyxia of the first born had twice the risk of renewed asphyxia at the next birth, this low risk classification is remarkable. So despite the low prevalence of asphyxia at term, an obstetric history of birth asphyxia and the interaction with other risk factors should be incorporated in the risk estimation for subsequent pregnancies.

Obstetric interventions and pregnancy outcomesHistorically, in the Netherlands there is a conservative attitude towards caesarean delivery. Our caesarean delivery rates around 10% for term singletons in cephalic presentation are among the lowest in Western countries. 8 Although many Western countries focus on lowering caesarean delivery rates, too few caesarean deliveries may effect neonatal outcomes adversely. On the other hand, a caesarean delivery does not guarantee improved neonatal outcomes. The rising trend in obstetric interventions for fetal distress in the Netherlands, together with a decrease in perinatal asphyxia did not result in change in perinatal mortality. This underlines the difficulties in timing and selection of cases during the delivery process. This delicate balance surrounding caesarean deliveries together with the presented practice variation in caesarean delivery rates should be debated and emphasizes the need for more general management in obstetric healthcare.

Individual risk assessment for adverse pregnancy outcomesTo improve poor pregnancy outcomes, risk estimation and individual prediction of adverse health outcomes is needed to allow targeted interventions. Before one can apply individualized risk estimation for a certain health outcome such as perinatal asphyxia and related mortality and morbidity, first the association between potential risk factors and the proposed outcome needs to be understand completely, thus allowing more targeted intervention.



9

167

The development and internal validation of our prognostic model for perinatal asphyxia and combining the risks for perinatal asphyxia and an intervention due to failure to progress (FTP) in a two-dimensional prediction graph is an important next step towards individualized risk assessment of women during pregnancy and delivery. The characteristics used in the development of the prediction models are readily available in clinical practice and are generally recorded in national registries. To assess the two-dimensional prediction graph’s performance one should aim to integrate the tool into the Electronic Medical Record (EMR) and validate this tool in a large obstetric population.

The limited model’s performance to predict perinatal asphyxia impedes at present its clinical usefulness and emphasizes the difficulty of predicting asphyxia at term. The main challenge is not only to collect data of a large number of pregnant women and their offspring’s, but also to include more potentially relevant variables and standardization of the definition of these variables. Ideally, an integrated tool into the medical record of pregnant women should be modifiable at all time during pregnancy and labor by the continuous addition of maternal, fetal, pregnancy and labor characteristics and measurements. Such a continuously adjustable decision tool could guide clinicians with subsequent interventions during pregnancy, labor and delivery.


Chapter I

168

referenCes


2. Sauer MV. Reproduction at an advanced maternal age and maternal health. Fertility and sterility 2015;103:1136-43.

3. Ravelli AC, Schaaf JM, Eskes M, Abu-Hanna A, De Miranda E, Mol BW. Ethnic disparities in perinatal mortality at 40 and 41 weeks of gestation. Journal of perinatal medicine 2013;41:381-8.

4. Lanting CI, Buitendijk SE, Crone MR, Segaar D, Bennebroek Gravenhorst J, Van Wouwe JP. Clustering of socioeconomic, behavioural, and neonatal risk factors for infant health in pregnant smokers. PloS one 2009;4:e8363.

5. Cooke RW. Smoking, intra-uterine growth retardation and sudden infant death syndrome. International journal of epidemiology 1998;27:238-41.

6. Henningsen AK, Pinborg A, Lidegaard O, Vestergaard C, Forman JL, Andersen AN. Perinatal outcome of singleton siblings born after assisted reproductive technology and spontaneous conception: Danish national sibling-cohort study. Fertility and sterility 2011;95:959-63.

7. http://www.zonmw.nl/nl/projecten/project-detail/the-iugr-risk-selection-study-iris-study/samenvatting/ (vol 2015).




9

169


Appendices Samenvatting in het Nederlands

List of co-authors and their contribution to the manuscript


PhD Portfolio

Curriculum vitae

Appendices

172

NEDERLANDSE SAMENVATTING

AchtergrondPerinatale asfyxie, of wel zuurstoftekort van het ongeboren kind, is een niet veel voorkomende, maar wel een ernstige complicatie die kan leiden tot sterfte of mogelijk lange termijn gevolgen voor de motorische en mentale ontwikkeling van de pasgeborene. Perinatale asfyxie kan zowel voorafgaand als tijdens de bevalling optreden.

In Nederland, met ongeveer 175 000 bevallingen per jaar en een prevalentie van perinatale asfyxie van 1.3%, betekend dit dat er per jaar 2200 kinderen worden geboren met een vorm van zuurstofgebrek. De diagnose perinatale asfyxie wordt gesteld aan de hand van een lage Apgar score en/of pH waarden van het navelstrengbloed. De Apgar score werd in 1952 ontwikkeld door Virginia Apgar voor een snelle en gestructureerde observatie van de conditie van de pasgeborene. De Apgar score bestaat uit 5 onderdelen waarvoor de pasgeborene 0, 1 of 2 punten krijgt, de maximale score bedraagt dus 10 punten en een totaal score van ≥ 7 wordt gezien als normaal. De Apgar score wordt 1, 5 en 10 minuten na de geboorte beoordeeld. Hieronder is een schematische weergave van de Apgar score weergegeven.

O punten 1 punt 2 puntenAdemhaling Geen Onregelmatig /

zwak huilenGoed door huilen

Pols en hartslag Geen < 100 slagen/minuut > 100 slagen/minuutSpierspanning Slap Beetje beweging,

matige spierspanningActieve bewegingen, sterke spierspanning

Aspect / kleur van de huid

Gehele lichaam bleek of blauw

Alleen blauwe extremiteiten

Gehele lichaam roze

Reactie op prikkels Geen Reactie, maar niet huilen Actief wegduwen, goed huilen

Het European PERISTAT project, een project dat de geboortezorg in Europa monitort en alle zwangerschapsuitkomsten tussen de Europese landen met elkaar vergelijkt, liet zien dat ondanks dalende trends de perinatale sterfte (sterfte rondom de geboorte) in Nederland tot de hoogste van Europa behoort. Uit literatuur onderzoek is bekend dat 85% van deze sterfte rondom de geboorte is toe te wijzen aan een van de zogenaamde BIG 4 determinanten: 1. vroeggeboorte; een bevalling die plaatsvindt voor de 37e week, 2. intra-uteriene groeivertraging; baby’s die te klein zijn, 3. aangeboren afwijkingen van het kind en 4. zuurstofgebrek rondom de bevalling. Verder weten we dat bij a terme geboren kinderen, dit zijn kinderen die worden geboren bij een zwangerschapsduur tussen de 37 weken + 0 dagen en 41 weken + 6 dagen, het aandeel van perinatale asfyxie aan de totale sterfte >75% bedraagt.

A

173

Samenvatting in het Nederlands

Tijdens de zwangerschap en bevalling wordt de klinische conditie van de foetus goed in de gaten gehouden, dit kan onder meer door het luisteren naar de foetale harttonen of een continue registratie van de foetale hartslag met een hartfilmpje (CTG). Dit wordt ook wel ‘foetale bewaking’ genoemd. De afgelopen decennia is de foetale bewaking enorm toegenomen in Westerse landen en als gevolg hiervan is er ook een toename van het aantal interventies, keizersneden en vaginale kunstverlossingen, tijdens de bevalling. Een interventie tijdens de baring, met name een keizersnede, is niet zonder risico voor de moeder en heeft vergaande consequenties voor eventuele volgende zwangerschappen.

Reeds voor de bevalling maakt de verloskundige en/of gynaecoloog al een risico inschatting voor de eventuele risico’s voor zowel moeder en kind om zo een veilige zwangerschap en bevalling te kunnen waarborgen. Deze risico inschatting gebeurt aan de hand van kenmerken die betrekking hebben op de moeder (vb. leeftijd en aanwezigheid van ziekten), het ongeboren kind (groei) en de zwangerschap (eerdere keizersnede). Een adequate risico inschatting zorgt ervoor dat vrouwen met een verhoogd risico op zwangerschapscomplicaties tijdig kunnen worden herkend en voor hen optimale zorg krijgen tijdens de zwangerschap en bevalling.

Dit proefschrift wil inzicht verschaffen in de risico factoren en de ontwikkeling van prognostische modellen voor perinatale asfyxie en daarmee samenhangende complicaties bij a terme eenling zwangerschappen.

In dit proefschrift worden de volgende onderwerpen onderzocht:1. Identificeren van risico factoren voor perinatale asfyxie gerelateerde morbiditeit en

mortaliteit bij a terme eenling zwangerschappen.2. Praktijkvariatie in obstetrische interventies in Nederlandse ziekenhuizen en hun

samenhang met neonatale en maternale uitkomsten.3. De ontwikkeling van een prognostisch model om perinatale asfyxie en gerelateerde

complicaties beter te kunnen voorspellen.

Hieronder worden de belangrijkste bevindingen samengevat:

In hoofdstuk 2 beschrijven we de trends in perinatale asfyxie bij a terme zwangerschappen in Nederland tussen 1999 en 2010. De prevalentie van milde asfyxie (Apgar score < 7) was 0.85% (1,285/151,680) en van ernstige asfyxie 0.16% (262/151,680). In deze tijdsperiode is het risico op asfyxie significant gedaald met ongeveer 6%, wat overeenkomt met een daling van 80 kinderen per jaar.

Appendices

174

Tegelijk met deze daling in asfyxie was er een significante stijging waarneembaar in het aantal interventies voor foetale nood, deze stegen van 5.9% in 1999 tot 7.7% in 2010. Echter ondanks de stijging in het aantal interventies en de daling in perinatale asfyxie was er geen daling waarneembaar in perinatale sterfte. Het aandeel van asfyxie gerelateerde sterfte aan de totale sterfte a term bleef met 87% constant in de studie periode.

In hoofdstuk 3 beschrijven we de praktijkvariatie in electieve (geplande) en ongeplande sectio caesarea (keizersneden) bij vrouwen die zwanger zijn van hun eerste kind. Van de 458, 712 kinderen die geboren zijn in de studie periode werd 1.8% geboren middels electieve sectio. Echter het percentage electieve sectio’s varieerde tussen 0.72% en 5.2% in de algemene Nederlandse ziekenhuizen. Voor ongeplande sectio’s was de variatie nog groter, namelijk tussen de 8.2% en 28%. Hogere percentages electieve sectio’s lijken te resulteren in betere neonatale uitkomsten, echter ziekenhuizen die minder electieve sectio’s deden dan gemiddeld hadden geen significante stijging van slechte neonatale uitkomsten. Kinderen die worden geboren na een ongeplande sectio caesarea in ziekenhuizen waar ze terughoudend zijn met het uitvoeren van sectio’s hebben een verhoogd risico op neonatale mortaliteit (sterfte) en morbiditeit.

In hoofdstuk 4 wordt het herhaal risico op perinatale asfyxie beschreven. Voor vrouwen die bevallen van hun eerste kind is het risico op een Apgar score < 7 0.99% in de a terme periode. Voor vrouwen die al eerder een kind hebben gebaard is dit risico gehalveerd tot 0.50%. Echter vrouwen die bij de eerste zwangerschap een kind hebben gebaard met een Apgar score < 7, hebben bij de volgende bevalling wederom een verhoogd risico op een kind met perinatale asfyxie. Dit risico is gelijk aan vrouwen die bevallen van hun eerste kind en treedt zowel op na een vaginale partus als keizersnede van het eerste kind. Dit herhaalrisico op asfyxie zou moeten worden meegenomen bij de risico inschatting van de volgende zwangerschap.

In hoofdstuk 5 presenteren we de uitkomsten van zwangerschappen ontstaan middels geassisteerde reproductie technieken (ART) als intra-uteriene inseminatie (IUI) en in vitro fertilisatie (IVF). In een observationeel cohort van meer dan 45 000 ART zwangerschappen vonden we dat eenlingen geboren na een ART zwangerschap een verhoogd risico hadden op perinatale asfyxie (Apgar < 4) in vergelijking tot eenlingen geboren na een natuurlijk ontstane zwangerschap (aOR 1.29, 95%CI 1.14 – 1.46). We weten dat subfertile vrouwen (vrouwen die ART ondergaan) verschillen van vrouwen die natuurlijk zwanger worden, ze zijn bijvoorbeeld vaak ouder en hebben vaker bijkomende ziekten als hoge bloeddruk. Daarom hebben we propensity score technieken gebruikt om vrouwen met een ART zwangerschap te matchen aan vrouwen


A

175


met een natuurlijke zwangerschap, zodat we een homogeen studiecohort kregen. In dit gematchte cohort bleek dat het risico op perinatale asfyxie voor ART vrouwen gelijk was aan dat van vrouwen die zwanger waren middels natuurlijke conceptie. Dit suggereert dat de karakteristieken van deze subfertile vrouwen die ART ondergaan bijdragen aan de slechtere uitkomsten in deze zwangerschappen.

Nullipariteit, niet-westerse etniciteit, roken tijdens de zwangerschap, groeivertraging van de foetus en op basis van medische condities (vb. hoge bloeddruk) geclassificeerde hoog risico zwangerschappen zijn allen belangrijke risico factoren voor asfyxie gerelateerde sterfte a term. Deze worden besproken in hoofdstuk 6. Populatie Attributieve Risico’s (PAR) werden berekend om te zien wat de potentiele gezondheidswinst zou kunnen zijn indien een bepaalde risico factor zou kunnen worden geëlimineerd. De hoogste PAR werd, naast nullipariteit (20%) en herkende hoog risico zwangerschappen (39%), gevonden voor groeivertraging (17%) en roken (16%). Voor vrouwen die hun bevalling starten in de eerste lijn (geclassificeerde laag risico zwangerschappen) is de PAR voor groeivertraging toegenomen tot 24% ten opzichte van vrouwen die hun bevalling starten in de tweede lijn (geclassificeerde hoog risico zwangerschappen). De bijdrage van groeivertraging aan a terme sterfte in de gehele populatie, en specifiek bij vrouwen die hun bevalling starten in de eerste lijn, toont aan dat adequate antenatale detectie van groeivertraging zou kunnen leiden tot een afname van a terme sterfte.

In hoofdstuk 7 beschrijven we de associatie tussen de verschillende componenten van de Apgar score en neonatale mortaliteit en morbiditeit bij a terme eenling zwangerschappen. Logistische regressie analyse laat zien dat alleen nul punten voor hartslag (a OR 18, 95%CI 3.7 – 86) en nul punten voor ademhaling (aOR 6.0, 95%CI 1.3 – 28) onafhankelijk geassocieerd zijn met een slechte neonatale uitkomst. Ondanks dat de Apgar score de meest gebruikte prognostische test is in de verloskunde, laat onze studie zien dat een beperkt model met alleen deze twee variabelen vergelijkbare prognostische kenmerken heeft.

In hoofdstuk 8 beschrijven we de ontwikkeling van een tweedimensionale beslissingstool waarmee we voor vrouwen het risico op een interventie vanwege het onvoldoende vorderen van de ontsluiting of uitdrijving uitzetten tegen het risico op perinatale asfyxie tijdens de bevalling. Voor vrouwen met een a terme eenling zwangerschap varieerde het risico op een interventie voor niet vorderende van de baring tussen de 0.30 en bijna 100%. Het intrapartum model, gebaseerd op 22 variabelen, toonde een goede kalibratie (voorspelde risico’s kwamen overeen met de geobserveerde risico’s) en had een AUC (Area Under the Curve; dit beschrijft het onderscheidend vermogen) van 0.78 (95%CI 0.77 – 0.79). Multivariate regressie analyse selecteerde nullipariteit, eerdere sectio caesarea, ART en hoge bloeddruk als


Appendices

176

de belangrijkste risico factoren voor een interventie voor niet vorderen van de baring. Het intrapartum model voor perinatale asfyxie bestond uit 21 variabelen en had een beperkter onderscheidend vermogen met een AUC van 0.67 (95%CI 0.66 – 0.68). De risico’s op perinatale asfyxie varieerde van 0.10% tot 13%. Een eerdere sectio, niet-kaukasische etniciteit en hoge bloeddruk waren samen met groeivertraging van de foetus de belangrijkste voorspellers voor perinatale asfyxie.

De voorspelde risico’s voor perinatale asfyxie en een interventie vanwege een niet-vorderende baring werden gecombineerd in een tweedimensionale beslissingstool. Hieruit bleek dat over het algemeen het risico op een interventie vanwege niet vorderen toenam met een verhoogd risico op perinatale asfyxie. Echter, het komt ook voor dat het risico op een interventie hoog was en het risico op perinatale asfyxie juist laag. Gezien het feit dat de individuele risico inschatting van de risico’s voor moeder en kind steeds belangrijker worden, denken wij dat de tweedimensionale beslissingstool gynaecologen en verloskundigen kan ondersteunen bij hun antenatale counseling van zwangere vrouwen en keuzes voor interventies tijdens de bevalling.


A

177



Appendices

178

lisT of Co-auThors and Their ConTribuTion To The manusCripT

Trends in birth asphyxia, obstetric interventions and perinatal mortality among

term singletons: a nationwide cohort study.


S. Ensing: study concept and design, acquisition of data, data analysis, interpretation of data,

drafting the manuscript, final approval of the manuscript.

A. Abu-Hanna: data analysis, critically reviewing the manuscript, final approval of the manuscript.

J.M. Schaaf: study concept and design, data analysis, final approval of the manuscript.

B.W.J. Mol: interpretation of data, critically reviewing the manuscript, final approval of the

manuscript.

A.C.J. Ravelli: study concept and design, supervision of data acquisition and analysis, interpretation

of data, critically reviewing the manuscript, final approval of the manuscript.

Are higher rates of obstetric interventions associated with a better neonatal

outcome? A population wide cohort study.



J. Kaandorp study concept and design, interpretation of data, critically reviewing the manuscript,

final approval of the manuscript.

J. B. Derks: critically reviewing the manuscript, final approval of the manuscript.

A. Abu-Hanna: critically reviewing the manuscript, final approval of the manuscript.


manuscript.




A

179

List of co-authors

Recurrence risk of low Apgar score among term singletons: a population-based

cohort study.




J.M. Schaaf: study concept and design, data analysis, final approval of the manuscript.



manuscript.



Risk of poor neonatal outcome at term after medically assisted reproduction: a

propensity score-matched study.




A. Abu-Hanna: supervision of data and analysis, critically reviewing the manuscript, final approval of

the manuscript.

T.J.Roseboom: interpretation of data, critically reviewing the manuscript, final approval of the

manuscript.

S. Repping: critically reviewing the manuscript, final approval of the manuscript.

F. van der Veen: critically reviewing the manuscript, final approval of the manuscript.


manuscript.

A.C.J. Ravelli: study concept and design, supervision of data acquisition, interpretation of data,

critically reviewing the manuscript, final approval of the manuscript.


Appendices

180

Maternal and neonatal risk factors for asphyxia related perinatal mortality at term.



F. Groenendaal: interpretation of data, critically reviewing the manuscript, final approval of the

manuscript.

M. Eskes: critically reviewing the manuscript, final approval of the manuscript.



manuscript.



Predictive capacity of the five individual components of the 5 minute Apgar score on

neonatal mortality and morbidity in term infants.



M.C. Engelen: acquisition of data, critically reviewing the manuscript, final approval of the manuscript.

P. Tamminga: acquisition of data, critically reviewing the manuscript, final approval of the manuscript.

A. Abu-Hanna: analysis of data, interpretation of data, critically reviewing the manuscript, final

approval of the manuscript.

B.W.J. Mol: critically reviewing the manuscript, final approval of the manuscript.




A

181

List of co-authors

The development of a clinical prediction model to support clinicians in the

simultaneous assessment of the risks of failure to progress and fetal asphyxia in

term pregnancies.



U. Ogba: study concept and design, acquisition of data, data analysis, drafting the manuscript,


E. Schuit: study concept and design, supervising data analysis, critically reviewing the manuscript,


A. Abu-Hanna: supervising analysis of data, interpretation of data, critically reviewing the manuscript,



manuscript.

A.C.J. Ravelli: study concept and design, supervision of data acquisition, interpretation of data,

critically reviewing the manuscript, final approval of the manuscript.


Appendices

182

lisT of publiCaTions

Effects of IVF and maternal characteristics on perinatal outcomes: a population-

based study using siblings.

Seggers J, Pontesilli M, Ravelli, ACJ, Painter RC, Hadders - Algra M, Heineman MJ, Repping S, Mol

BWJ, Roseboom TJ, Ensing S.

Accepted for publication in Fertil Steril.

Subsequent pregnancy outcome after preterm breech delivery, a population based

cohort study.

Bergenhenegouwen L, Ensing S, Ravelli ACJ, Schaaf JM, Kok M, Mol BWJ

Accepted for publication in J Matern Fetal Neonatal Med

Preterm Breech Presentation: A Comparison of planned Intended Vaginal and

Intended planned Cesearean Delivery.

Bergenhenegouwen L, Vlemmix F, Ensing S, Schaaf J., vd Post JAM, Abu-Hanna A, Ravelli AC, Mol

BWJ, Kok M

Accepted for publication in Obstet & Gynecol

Risk of poor neonatal outcome at term after medically assisted reproduction: a

propensity score-matched study.

Ensing S, Abu-Hanna A, Roseboom TJ, Repping S, van der Veen F, Mol BW, Ravelli AC.


Trends in birth asphyxia, obstetric interventions and perinatal mortality among

term singletons: a nationwide cohort study.

Ensing S, Abu-Hanna A, Schaaf JM, Mol BW, Ravelli AC.


Impact of computer-based pregnancy-induced hypertension and diabetes decision

AIDS on empowering pregnant women.

Aslani A, Tara F, Ghalighi L, Pournik O, Ensing S, Abu-Hanna A, Eslami S.

Healthc Inform Res. 2014 Oct;20(4):266-71

Term breech deliveries in the Netherlands: did the increased cesarean rate affect

neonatal outcome? A population-based cohort study.

Vlemmix F, Bergenhenegouwen L, Schaaf JM, Ensing S, Rosman AN, Ravelli AC, Van Der Post JA,

Verhoeven A, Visser GH, Mol BW, Kok M.



A

183


Recurrence risk of low Apgar score among term singletons: a population-based

cohort study.

Ensing S, Schaaf JM, Abu-Hanna A, Mol BW, Ravelli AC.


Influence of the 20-week anomaly scan on prenatal diagnosis and management of

fetal facial clefts.

Ensing S, Kleinrouweler CE, Maas SM, Bilardo CM, Van der Horst CM, Pajkrt E.

Ultrasound Obstet Gynecol. 2014 Aug;44(2):154-9


Appendices

184

phd porTfolio

Name PhD student: S. Ensing

PhD period: July 2012 – September 2015

PhD supervisors: Prof. dr. B.W.J. Mol, Prof. dr. A. Abu-Hanna

and dr. A.C.J. Ravelli

PhD training Year Workload (ECTS)

General Courses

Academic Medical Center, Amsterdam

The AMC World of Science 2012 0.7

Pubmed Biomedical Sciences 2012 0.2

Reference Manager 2012 0.1

Evidence Based Searching 2012 0.1

Oral Presentation in English 2012 0.8

Practical Biostatistics 2012 1.1

Specific Courses


Systematic Reviews 2013 0.3

Advanced topics in Biostatistics 2013 2.1

Clinical Epidemiology 2013 0.6

Advanced topics in Clinical Epidemiology 2013 1.1

Computing in R 2013 0.4

Genetic Epidemiology 2014 1.1

Erasmus University Medical Center, Erasmus Summer/Winter

School, Netherlands Institute for Health Sciences, Rotterdam

Conceptual Foundation of Epidemiologic Study Design 2012 0.7

Social Epidemiology 2012 0.7

Pharmaco-epidemiology and Drug-safety 2013 1.9

Advanced Topics in Decision Making in Medicine 2013 1.9

Topics in Meta-analysis 2013 0.7

Clinical Decision Analysis 2013 0.7

14th conference on Artificial Intelligence in Medicine, Murcia

Evaluating Prognostic Models in Medicine 2013 0.5


A

185

phd porTfolio

Name PhD student: S. Ensing

PhD period: July 2012 – September 2015

PhD supervisors: Prof. dr. B.W.J. Mol, Prof. dr. A. Abu-Hanna

and dr. A.C.J. Ravelli


General Courses


The AMC World of Science 2012 0.7

Pubmed Biomedical Sciences 2012 0.2

Reference Manager 2012 0.1

Evidence Based Searching 2012 0.1

Oral Presentation in English 2012 0.8

Practical Biostatistics 2012 1.1

Specific Courses


Systematic Reviews 2013 0.3

Advanced topics in Biostatistics 2013 2.1

Clinical Epidemiology 2013 0.6

Advanced topics in Clinical Epidemiology 2013 1.1

Computing in R 2013 0.4

Genetic Epidemiology 2014 1.1

Erasmus University Medical Center, Erasmus Summer/Winter

School, Netherlands Institute for Health Sciences, Rotterdam

Conceptual Foundation of Epidemiologic Study Design 2012 0.7

Social Epidemiology 2012 0.7

Pharmaco-epidemiology and Drug-safety 2013 1.9

Advanced Topics in Decision Making in Medicine 2013 1.9

Topics in Meta-analysis 2013 0.7

Clinical Decision Analysis 2013 0.7

14th conference on Artificial Intelligence in Medicine, Murcia

Evaluating Prognostic Models in Medicine 2013 0.5

PhD Portfolio


Academic Medical Center, Amsterdam and Erasmus University

Medical Center, Erasmus Summer/Winter School,

Netherlands Institute for Health Sciences, Rotterdam

Clinical Epidemiologist – Registered Epidemiologist A 2012-2015

Seminars

Weekly department lunch meetings 2012-2014 3.5

Weekly department seminars 2012-2014 3.5

Weekly department journal club

Department of Obstetrics and Gynecology 2012-2014 3.5

Department of Medical Informatics 2012-2014 3.5

Presentations

Oral presentation Pijlerdag Pijler Foetomaternale Geneeskunde 2013 0.5

Recurrence risk of birth asphyxia.

Poster presentation SMFM San Francisco 2013 0.5

Trends in birth asphyxia, obstetric interventions

and perinatal mortality among term singletons.

Poster presentation SMFM San Francisco 2013 0.5

Recurrence risk of low Apgar score among term

singletons: a population-based cohort study.

Poster presentation WEON Utrecht 2013 0.5

Risk of birth asphyxia at term after Assisted Reproduction

Technology (ART), A Propensity Score Analysis

Poster presentation SMFM New Orleans 2014 0.5

Risk of poor neonatal outcome at term after

medically assisted reproduction: a propensity score-matched study.

Poster presentation SMFM New Orleans 2014 0.5

Prediction of perinatal Asphyxia in term singletons

Poster presentation SMFM San Diego 2015 0.5

Risk factors for asphyxia related perinatal mortality at term.

Poster presentation SMFM San Diego 2015 0.5

Predictive capacity of the five individual components

of the 5 minute Apgar score on neonatal mortality and morbidity.


Appendices

186


(Inter)national conferences

SMFM’s 33rd Annual Pregnancy Meeting, San Francisco 2013 0.25

SMFM’s 34rd Annual Pregnancy Meeting, New Orleans 2014 0.25

Jaarlijkse conferentie Werkgroep Epidemiologisch 2013 0.25

Onderzoek Nederland (WEON)

Teaching

Tutoring skills education to second year medical students 2013 0.1

Student coaching/mentoring scientific research project 2014 1.0

Student coaching/mentoring scientific research project 2015 1.0

Course “Verloskundige zorg en gegevensvastlegging” 2012 – 2014 0.25

Course “Beslisondersteuning in de zorg: vroeggeboorte 2012 – 2014 0.25

Course Matchingsday Medical informatics 2014 0.25


A

187

PhD Portfolio


Appendices

188

CurriCulum ViTae

Sabine Ensing was born on the 6th of April in 1986 in Nijmegen, the Netherlands. She grew up in Breda with two younger brothers with her parents. After graduating from secondary school (cum laude) at the Onze Lieve Vrouwelyceum in Breda in 2004 she continued with medical school at the University of Amsterdam. She was an active member of the L.A.N.X. student association. For her scientific internship she did a research project at the department of prenatal diagnosis in the Academic Medical Center in Amsterdam on the influence of the 20-week anomaly scan on prenatal diagnosis and management of fetal facial clefts. After she completed an obstetrical internship in Suriname, she graduated from medical school (cum laude) in June 2011. After graduation she started working as a resident on the department of Obstetrics and Gynaecology of the Zaans Medisch Centrum in Zaandam. In the summer of 2012 prof. dr. B.W.J. Mol, prof. dr. A. Abu-Hanna and dr. A.C.J. Ravelli gave her the opportunity to start a PhD project on the risk factors of perinatal asphyxia among term pregnancies. The results of this PhD project resulted in this thesis. During her PhD project she completed a structured curriculum of education and training as epidemiologist. In October 2014 she started her residency in Obstetrics and Gynaecology in the Spaarne Gasthuis in Hoofddorp under the supervision of dr. A. Vollebregt.


A

189

Curriculum vitae


Appendices

190

danKWoord

Een proefschrift schrijf je niet alleen, er zijn dan ook veel mensen die op wat voor manier dan ook een bijdrage hebben geleverd aan de totstandkoming van dit ‘boekje’. Een aantal van hen wil ik specifiek benoemen.

Prof. dr. A. Abu-Hanna, beste Ameen. Ik liep al 8 jaar in het AMC rond, maar ik was nog nooit op de medische informatiekunde geweest, jouw domein. Het was even wennen, maar al snel ging er een hele statistische / “R” wereld voor mij open. Veel dank voor het geduld waarmee je me de vaardigheden hebt uitgelegd en je vaak reddende mail met de R-code als ik er even niet uitkwam. Ik heb ontzettend veel van je geleerd en je altijd kritische blik op het schrijfwerk tilde mijn stukken naar een hoger plan. Dank voor de prettige samenwerking. Ik wens je veel succes voor de toekomst en alle goeds voor jou en je gezin!

Prof. dr. B.W.J. Mol, beste Ben Willem. Vol bewondering kijk ik hoe je alle (promovendi) ballen in de lucht houdt en steeds weer twee stappen op ons vooruit bent. Ons wekelijks overleg ging lang niet altijd over mijn proefschrift, des te meer over de (verloskundige) zorg in bredere zin. Dank voor de kansen die je me hebt gegeven om betrokken te zijn bij verschillende onderzoeken, dit verruimde mijn blik. Al moest ik je soms afremmen om mijn eigen proefschrift niet in het gedring te laten komen, uiteindelijk kwamen tot een goed evenwicht. Het is inspirerend om met je samen te werken en ik hoop dat dan ook nog een tijd voort te zetten. Het is goed te horen dat het jou en je gezin goed vergaat in Australië.

Dr. A.C.J. Ravelli, beste Anita. Mijn collega’s waren vaak een beetje jaloers op een co-promotor als jij en dat is zeker terecht! Je hebt me ingewijd in de epidemiologie en gezorgd dat ik me snel thuis voelde op de KIK. Uren hebben we samen aan de artikelen gezeten, bijschaven, andere analyses runnen. Na een overleg met jou met als hoofdvraag: “Sabine wat wil je nu eigenlijk zeggen met dit artikel?”, zag ik het weer helemaal zitten. Je deur staat altijd open en het was ook erg gezellig om samen te kletsen over het kopen van een huis, vakanties etc. Veel dank voor je steun, frisse blik op mijn onderzoek, motiverende mails en enthousiasme als co-promotor! Ik wens je het allerbeste en kom nog wel eens binnenvallen.

De leden van de promotiecommissie: Prof. dr. G.J. Bonsel, Prof. dr. A.H.L.C. van Kaam, dr. A. Kwee, Prof. dr. T.J. Roseboom, Prof. dr. E.W. Steyerberg en Prof. dr. F van der Veen dank ik voor het kritisch beoordelen van dit proefschrift en voor de mogelijkheid om dit proefschrift ten overstaan van hen te verdedigen.


A

191

Dankwoord

De Stichting Perinatale Registratie Nederland wil ik bedanken voor het gebruik maken van de dataset waarop het merendeel van de artikelen in dit proefschrift zijn gebaseerd. En daarnaast wil ik natuurlijk ook de zorgverleners bedanken voor het rapporteren van de gegevens.

Veel dank gaat ook uit naar de mede-auteurs van de artikelen, dank voor jullie kritische woorden en bruikbare suggesties. Marie – Claire, uren heb je oude NICU statussen doorgespit op zoek naar de 5 componenten van de APGAR score, dank daarvoor. Uloma, thanks for your enthusiasm and hard work. Without your tireless efforts Chapter 8 was never finished on time.

Double daters van de KIK. Leonie, Jos, Tom, Marjan, Wouter, Erik, Ellen, Nick en Airin, wat was het gezellig in de rechter gang. Op woensdagen hadden we de juiste break van de week en draaide de koffiemachine van de appie overuren, onze computers iets minder. Ik hoop nog vaak bij een pubquiz of squash avondje te kunnen aansluiten. Marjan, je hebt me wegwijs gemaakt op de KIK en daardoor voelde ik me als nieuwkomer meteen thuis op J1B. Wat hebben we veel gelachen, de fruitsapjes en liters thee waren niet aan te slepen en zodra de zon scheen een drankje bij de oranje bouwkeet. Het waren 2 top jaren. Owja onze kerstboom hangt er nog steeds “gezellig” bij!

Lieve collega’s van de gyn, een paar keer per week klom ik de trap op naar H4 en daar was het altijd feest, met drop en stroopwafels in overvloed! Wat hebben we gelachen op de skiweekenden, radlers bij Seppies en dan over de grashelling naar beneden. Of met zijn allen naar the Swedisch Housemaffia in San Francisco, dansen tussen Amerikaanse tienermeisjes gekleed als Victoria Secret’s Angels. Janneke, Larisa en Brenda, onze trip naar NYC was legendarisch, bij the Starbucks heet ik Shelby! Hannah, hoe wij 3 kg zwitsersekaas hebben opgegeten zonder buikpijn is mij nog steeds een raadsel?! Kai en Laris de reis naar Cuba was prachtig! Ik hoop dat er nog vele vrimibo’s mogen volgen met zijn allen.

Lieve vrienden en familie, dank voor jullie steun en interesse in mijn onderzoek de afgelopen jaren. Manon, Loes, Simone, Ank en Marloes, na 10 jaar nog altijd even soepel en puur! Vol verwondering en met trots kijk ik naar hoe jullie in het leven staan, stelletje powervrouwen. Mijn jaargenoten van het eerste uur, vriendinnen in goede en slechte tijden, ik kijk vol geluk terug op alle mooie momenten samen en kijk uit naar de vele avonden samen en weekendjes weg die nog komen gaan. San, Roos en Ank, we zijn vier drukke baasjes, maar eenmaal bij elkaar een niet te stoppen wervelwind, in elk geval niet voordat de wijn op is en de lichten gedoofd! Jullie zijn prachtig en zijn er altijd, ook als ik er even niet was. Let the sun shine!


Appendices

192

Lieve collega’s uit het Spaarne, bedankt dat jullie me meteen thuis hebben laten voelen op de afdeling, VK, poli etc. Er zijn al veel leuke, grappige en vooral ook leerzame momenten geweest en ik kijk uit naar mijn resterende tijd als AIOS bij jullie. Binnenkort weer bowlen?!

Oud-collega’s uit het ZMC, groen als gras kwam ik als piepjonge ANIOS bij jullie terecht. Jullie hebben me bij de hand genomen en me wegwijs gemaakt binnen de gynaecologie. Bijzonder dank aan Neriman, als moederoverste waak je over al je assistenten en geef je ons het juiste zetje in de goede richting.

Mijn paranimfen, Loes en Rianne. Lieve Ryan, mijn Bredase rots in de branding! Dit jaar kennen wij elkaar 25 jaar en ook al zien we elkaar minder sinds we allebei in een andere stad gingen studeren, nog steeds begrijp je me beter dan wie dan ook. Ik ben trots dat jij vandaag naast mij staat. Loes, vriend, in Paramaribo wist ik het al zeker, als ik ooit ga promoveren dan hoor jij daar bij. Je ongekende enthousiasme voor alles en iedereen is aanstekelijk, wat ben je toch een mooi mens.

Mijn kleine grote broertjes, Daan en Paul. Ik mag in mijn handjes knijpen met broers zoals jullie, die me met Jaki en Roos ook nog eens twee fantastische schoonzusjes hebben opgeleverd. Altijd zijn jullie daar en altijd geïnteresseerd in wat ik aan het doen ben, ook als ik weer eens afhaak door dienst of deadlines. We hebben alle drie een totaal andere weg bewandeld, maar wat ben ik trots op jullie. Daan, blijf jezelf en De Parel trouw dan komt het goed. Paul, kleine krullenbol, hoe gezellig dat jij nu ook in Amsterdam woont. Ik beloof, na vandaag heb ik meer tijd en kom ik vaker even aanwaaien.

Lieve papa en mama, dankzij jullie ben ik geworden wie ik nu ben en jullie hebben daarmee de basis gelegd voor dit boekje. Altijd kan ik op jullie terug vallen, Breda is nog steeds goed thuiskomen. Pap, ook al wordt de rest van de familie soms moe van ons eeuwige gepraat over de gezondheidszorg, ik koester deze momenten en onze soms felle discussies. Wat jij bereikt hebt, maakt mij trots! Lieve mama, zoals jij zijn er weinig, altijd geïnteresseerd en een luisterend oor. Maar je kwam ook met de juiste de afleiding als ik even verzonk in mijn werk. Dank je wel. Hoe jij in het leven staat, met je zorgzaamheid en betrokkenheid naar iedereen, is voor mij een groot voorbeeld.


A

193

Dankwoord

UvA-DARE (Digital Academic Repository) · Sabine Ensing op 18 december 2015 om 10.00 uur...

Documents

Transcript of UvA-DARE (Digital Academic Repository) · Sabine Ensing op 18 december 2015 om 10.00 uur...