Retinopathy of Prematurity and Multiple

Postnatal Infections in Preterm Neonates: Delays

in White Matter Development with Poorer

Neurodevelopmental Outcomes.

by

Torin James Alexander Glass

A thesis submitted in conformity with the requirements

for the Degree of Master of Science

Institute of Medical Science

University of Toronto

© Copyright by Torin James Alexander Glass 2018

ii

Retinopathy of Prematurity and Multiple Postnatal Infections in

Preterm Neonates: Delays in White Matter Development with Poorer

Neurodevelopmental Outcomes.

Torin James Alexander Glass

Master of Science

Institute of Medical Science

University of Toronto

2018

Abstract

Preterm birth is a common cause of neurodevelopmental disorders in childhood.

Little is known about the outcome of infants with severe retinopathy of prematurity (ROP)

and multiple infections in the postnatal period. This thesis describes their associations

with abnormal brain development using multi-modal MR imaging and standardized

outcome assessments. Those infants with severe ROP had delayed brain maturation of

the posterior white matter and optic radiations, with poorer 18 month cognitive and motor

outcomes. Compared to fewer episodes of infection, three or more postnatal infections

was associated with delayed maturation of the posterior limb of the internal capsule

(PLIC), corpus callosum and the optic radiations, with poorer 36 month motor outcomes.

Our findings support that severe ROP and multiple postnatal infections in very preterm

newborns are associated with decreased brain maturation and poorer

iii

neurodevelopmental outcomes, and that advancements in these disorders have the

potential to improve outcomes.

iv

Acknowledgements

Firstly, I thank Dr. Steven Miller for his mentoring and superb insight into the needs

of my learning as a trainee and developing researcher. I have felt continually supported,

challenged and encouraged throughout completing this work. During my time at SickKids

Hospital and the University of Toronto I have learnt so much and I am grateful for the time

and energy that was invested in the development of my research and clinical skills.

Secondly, I thank Dr. Vann Chau for his unending support and friendship, and for

his commitment to my success as a clinician and researcher. It was in part due to his

encouragement, kindness and laughter in the work environment that I came to the

University of Toronto in the first place. Thank you for the words of wisdom and the

thousands of coffees that have fueled me through the last couple of years.

I also sincerely thank Dr. John Sled and Dr. Margot Taylor for their guidance and

assistance as members of my advisory committee. Their combined analytical assessment

improved greatly the quality of the work within this thesis and have made me a better

researcher.

To the members of the Miller and Tam labs and the Neonatal Neurology fellows,

nurses and NPs, thank you for making my time in Toronto some of the best of my

professional life. Your unending encouragement of my work provided me with the

inspiration to complete it and has educated me in the importance of surrounding oneself

with truly good and inspiring people. I would especially like to thank Emily Tam, Diane

Wilson, Claire Watt, Julianne Schneider, Lara Leiser, Mariam Ayed, Amr Al-Shahed,

Elana Pinchefsky, Dalit Cayam-Rand, Isabel Benavente-Fernandez and Asma Al-

v

Mazroei and for their support through the years and for the laughter we have shared in

the process of my fellowship. In addition, this project would not have been possible

without the contributions of Jessie Guo, Justin Foong, Emma Duerden, Janet Rigney,

Ruth Grunau, Anne Synnes and Ken Poskitt.

Special thanks go to my family, whose continued support and encouragement has

always pushed me to be the best that I can be and in large part are the reason why I am

where I am today. My biggest thanks to my wife, Michelle, and son, Ronan, who have

joined me on this adventure to Toronto and whose strength and humility provided the

backbone upon which I was able to complete this work.

vi

Contributions

The author was responsible for the writing and preparation of this original thesis. All of

the work presented, including the planning, analysis and writing of the original research,

was performed by the author, with the guidance and expertise of the individuals listed

below. The following contributions to the work in this thesis are formally and inclusively

acknowledged:

Dr. Vann Chau: Provided clinical assessments and data described in Chapters 2 – 3 as

well as producing the DTI and MRSI database used in Chapters 2 and 3.

Dr. Emma Duerden and Mr. Justin Foong: Assisted in the development of the TBSS

analysis method and provided the figures in Chapters 2 and 3.

Dr. Jessie Guo: Provided the brain segmentation volumes used in Chapter 3.

Dr. Anne Synnes and Dr. Ruth Grunau: Contributed to the overall planning and

assessments for the follow-up and outcomes in Chapters 2 and 3.

Dr. Jane Gardiner: Provided the retinopathy of prematurity scoring in Chapters 2 and 3.

Dr. Ken Poskitt: Provided the MRI analysis and injury scoring in Chapters 2 and 3.

Dr. Jillian Vinall: Contributed DTI analysis to the study in Chapter 2.

vii

Table of Contents

Contents Page

Acknowledgments…………………………………………………………………………. iv

Contributions………………………………………………………………………………. vi

Table of Contents…………………………………………………………………………. vii

List of Abbreviations………………………………………………………………………. xi

List of Figures …………………………………………………………………………….. xiv

List of Tables ……………………………………………………………………………… xv

1 Literature Review……………………………………………………………………….. 1

1.1 Introduction………………………………………………………………………… 2

1.2 Preterm Birth………………………………………………………………………. 3

1.2.1 Overview…………………………………………………………………….. 3

1.3 Brain Injury in Preterm Infants…………………………………………………… 6

1.3.1 White Matter Injury…………………………………………………………. 6

1.3.2 Intraventricular Hemorrhage………………………………………………. 10

1.3.3 Cerebellar Hemorrhage……………………………………………………. 11

1.4 Retinopathy of Prematurity………………………………………………………. 12

1.4.1 Background…………………………………………………………………. 12

1.4.2 Pathophysiology……………………………………………………………. 15

1.4.3 Outcomes…………………………………………………………………… 17

1.5 Infection in Preterm Infants………………………………………………………. 18

1.5.1 Background…………………………………………………………………. 18

1.5.2 Classification………………………………………………………………... 18

1.5.3 Pre-Natal Infections………………………………………………………… 19

1.5.4 Postnatal Infections…………………………………………………………. 20

viii

1.5.5 Neonatal Inflammation……………………………………………………… 22

1.5.6 Neonatal Hypoxia-ischemic Injury…………………………………………. 24

1.6 Magnetic Resonance Imaging……………………………………………………. 26

1.6.1 Imaging of Preterm Newborn………………………………………………. 26

1.6.2 Overview of MRI basics…………………………………………………….. 27

1.6.3 MRI Sequences……………………………………………………………… 28

1.6.4 Diffusion Weighted Imaging and Diffusion Tensor Imaging ……………. 29

1.6.5 Magnetic Resonance Spectroscopic Imaging……………………………. 35

1.7 Neurodevelopmental Outcomes………………………………………………….. 39

1.7.1 Background………………………………………………………………….. 39

1.7.2 Bayley Scales of Infant and Toddler Development……………………… 39

1.7.3 Peabody Developmental Motor Scales…………………………………… 40

1.7.4 Cerebral Palsy……………………………………………………………….. 41

1.8 Hypothesis, Major Goal and Specific Aims……………………………………… 43

1.8.1 Hypothesis……………………………………………………………………. 43

1.8.2 Major Goal……………………………………………………………………. 43

1.8.3 Specific Aims…………………………………………………………………. 44

2 Severe Retinopathy of Prematurity predicts Delayed White Matter Maturation

and Poorer Neurodevelopment at 18 months CA………………………………….. 45

2.1 Introduction…………………………………………………………………………. 46

2.2 Material and Methods……………………………………………………………… 47

2.2.1 Participants…………………………………………………………………… 47

2.2.2 Clinical Characteristics……………………………………………………… 47

2.2.3 MRI Studies………………………………………………………………….. 48

2.2.4 Diffusion Tensor Imaging…………………………………………………… 48

2.2.5 Developmental Follow-up…………………………………………………… 50

ix

2.2.6 Data Analysis…………………………………………………………………. 50

2.3 Results………………………………………………………………………………. 51

2.3.1 Clinical Characteristics……………………………………………………… 51

2.3.2 Retinopathy of Prematurity…………………………………………………. 52

2.3.3 Brain Injury…………………………………………………………………… 54

2.3.4 White Matter Maturation…………………………………………………….. 54

2.3.5 Developmental Outcomes………………………………………………….. 57

2.4 Discussion………………………………………………………………………….. 58

2.4.1 Limitations……………………………………………………………………. 60

2.5 Conclusions………………………………………………………………………… 61

3 Multiple Postnatal Infections in Preterm Newborns are associated with delayed Maturation of

Motor Pathways at Term-equivalent age and Poorer Motor Outcomes at

3 years…………………………................................................................................ 62

3.1 Introduction…………………………………………………………………………. 63

3.2 Material and Methods……………………………………………………………… 64

3.2.1 Study Population…………………………………………………………….. 64

3.2.2 Clinical Characteristics……………………………………………………… 65

3.2.3 MR Brain Imaging……………………………………………………………. 65

3.2.4 Magnetic Resonance Spectroscopic Imaging……………………………. 66

3.2.5 Diffusion Tensor Imaging…………………………………………………… 66

3.2.6 Developmental Follow-up…………………………………………………… 68

3.2.7 Data Analysis………………………………………………………………… 68

3.3 Results……………………………………………………………………………… 69

3.3.1 Clinical Characteristics……………………………………………………… 69

3.3.2 Infection Characteristics…………………………………………………….. 70

3.3.3 MR Imaging………………………………………………………………….. 73

x

3.3.4 MRSI and DTI Imaging……………………………………………………… 74

3.3.5 Developmental Outcomes………………………………………………….. 76

3.4 Discussion………………………………………………………………………….. 79

3.4.1 Limitations……………………………………………………………………. 84

3.5 Conclusions………………………………………………………………………… 84

4 Summary of Main Findings and Future Directions…………………………………… 86

4.1 Conclusions………………………………………………………………………… 87

4.2 Future Directions…………………………………………………………………… 88

4.2.1 Retinopathy of Prematurity and Neurodevelopment…………………….. 88

4.2.2 Multiple Postnatal Infections and Neurodevelopment…………………… 93

References ………………………………………………………………………………… 104

xi

List of Abbreviations

ω3-PUFA – ω3 Polyunsaturated fatty acids

3D – Three-dimensional

ADHD – Attention-deficit hyperactivity disorder

BPD – Bronchopulmonary dysplasia

BSID-III – Bayley Scales of Infant and Toddler Development-III

CA – Corrected age

CHO – Glycerophosphocholine + phosphocholine

CI – Confidence interval

CNS – Central nervous system

CONS – Coagulase-negative staphylococcus aureus

CR – Creatine + phosphocreatine

CSF – Cerebrospinal fluid

DCD – Developmental coordination disorder

DHA - Docosahexaenoic acid

DNA – Deoxyribonucleic acid

DTI – Diffusion tensor imaging

DWI – Diffusion weighted imaging

EPO – Erythropoietin

FA – Fractional anisotropy

GA – Gestational age

GABA – Gamma-aminobutyric acid

GBS – Group B streptococcus

IGF-1 – Insulin-like growth factor 1

IL – Interleukin

INS – Myo-inositol

IQ – Intelligence quotient

IQR – Interquartile range

IVH – Intraventricular hemorrhage

xii

LAC – Lactate

LMP – Last menstrual period

MD – Mean diffusivity

MEG - Magnetoencephalogram

MHz – Megahertz

MR – Magnetic resonance

MRI – Magnetic resonance imaging

MRSI – Magnetic resonance spectroscopic imaging

NAA – N-acetylaspartate + n-acetylaspartylglutamate

NEC – Necrotising enterocolitis

NICU – Neonatal intensive care unit

OCT – Optical coherence topography

OL – Oligodendrocyte

OR – Odds ratio

PDA – Patent ductus arteriosus

PDMS-2 – Peabody developmental motor scales 2

PLIC – Posterior limb of the internal capsule

PMA – Post-menstrual age

PPM – Parts per million

PRR – Protein recognition receptor

PVL – Periventricular leukomalacia

RF – Radiofrequency

RNFL – Retinal nerve fibre layer

ROI – Region of interest

ROP – Retinopathy of prematurity

ROS – Reactive oxygen species

RR – Risk ratio

SNAPPE-II – Score for neonatal acute physiology with perinatal extension-II

TBSS – Tract-based spatial statistics

TE – Echo time

xiii

TNF-α – Tumor necrosis factor alpha

TR – Repetition time

VEGF – Vascular endothelial growth factor

WMI – White matter injury

xiv

List of Figures Page

1.1 Number of Neonatal Morbidities and Probability of Poor Outcomes at 5 years… 6

1.2 WMI Severity in the Preterm Infant………………………………………………..... 8

1.3 Pathophysiology of Diffuse and Axonal WMI…………………………..………….. 8

1.4 Retinopathy of Prematurity scoring scale…………………………………………… 13

1.5 Retinopathy of Prematurity stages of severity……………………………………… 14

1.6 Pathogenic Mechanisms in Retinopathy of Prematurity…………………………... 16

1.7 Representation of the 3D diffusion of an ellipsoid…………………………………. 30

1.8 Diffusion Tensor Imaging colour map of a Preterm Infant Brain…………………. 32

1.9 TBSS model of Age-specific Templates……..……………………………………... 34

1.10 MRSI example from a Preterm Infant……………………………………………… 38

2.1 Flow chart of study enrollment………………………………………………………. 52

2.2 Mean FA map and results of Severe ROP………..……………….………………. 54

2.3 TBSS model of Severe ROP results…………..………………….………………… 56

3.1 Flow chart of study enrollment………………………………………………………. 70

3.2 MRSI graph results of NAA/CHO ratio in white matter and basal ganglia……… 74

3.3 TBSS results of Multiple Infections compared to no Infection group…………….. 76

3.4 BSID-III Outcomes at 36 months across groups…………………………………… 78

xv

List of Tables Page

2.1 Demographics and Imaging results in Infants with and without Severe ROP….. 53

2.2 Follow-up Assessments at 18 months CA…………………………………………. 57

2.3 Multivariate Linear Regression Analysis of Outcomes at 18 months CA……….. 58

3.1 Demographics and Imaging in Infants with Postnatal Infections……………….... 72

3.2 Infection Location and Organism…………..………………………………………… 73

3.3 BSID-III Outcomes at 36 months CA by Infection groups……..…………………. 77

3.4 Outcome Odds for Poorer Outcomes at 36 months CA……..……………………. 79

- 1 -

Chapter 1

Literature Review

- 2 -

1.1 Introduction

Retinopathy of prematurity (ROP) and postnatal infection are important factors in

children born preterm, which are often associated with poor neurodevelopmental

outcomes. The pathophysiologies of ROP and infection share several characteristics with

inflammation and oxygenation alterations likely leading to impairments of the developing

brain. Infections, in particular postnatal bacteremia and fungal infections, are an

independent risk factor in the development of severe ROP in extremely preterm infants

(Manzoni et al, 2006; Tolsma et al, 2011). This association is possibly explained by the

combined effects of systemic inflammation and hypoxia-ischemia in the susceptible

preterm infant (Chen et al, 2011; Dammann, 2010), though the complete effects of each

remains uncertain. Furthermore, infection has been shown to have effects on insulin-like

growth factor 1 (IGF-1) and vascular endothelial growth factor (VEGF), growth factors

known to have critical roles in the growth of normal and abnormal vascularization of the

developing retina (Biswas et al, 2006; Heemskerk et al, 1999).

Structural MRI brain imaging of preterm infants has the potential to predict long-

term outcomes, with abnormalities in motor pathway tracts shown to be associated with

a higher likelihood of developing neurodevelopmental impairments (de Vries et al, 2011;

Guo et al, 2017). Advanced MRI techniques, such as diffusion tensor imaging (DTI) to

measure fractional anisotropy, have described delayed development of motor pathways

in children with cerebral palsy (Thomas et al, 2005; Yoshida et al, 2010), further

strengthening the role of MRI in predicting outcomes.

- 3 -

While much is known about postnatal infection and severe ROP, and their

associated developmental outcomes, descriptions of the alterations in the brain of infants

with these conditions are few. The research presented in this thesis examines the

association of severe ROP and of multiple postnatal infections with brain development of

preterm infants using multi-modal MRI methods, and describes the neurodevelopmental

outcomes of children with these disorders.

1.2 Preterm Birth

1.2.1 Overview

Every year approximately 15 million infants world-wide are born preterm (before

37 completed weeks of gestation), with complications stemming from prematurity the

number one cause of death in infants under 5 years of age (Blencowe et al, 2012).

Surviving preterm infants have higher rates of cognitive dysfunction, motor impairments,

attention-deficit hyperactivity disorder (ADHD) and autism than infants born at term

(Bhutta et al, 2002; Johnson et al, 2010; Wood et al, 2000). This leads to an increased

requirement for medical services resulting in a greater cost to the medical system than

term-born infants (Petrou, 2005).

Infants born prematurely are roughly grouped into three categories based upon the

completed weeks of gestation since the last menstrual period (LMP), approximated with

antenatal ultrasound measurements, and referred to as their “gestational age” (GA).

Infants born at <28 weeks GA are termed “extremely preterm”, 28 to <32 weeks GA as

“very preterm”, and 32 to <37 weeks as “moderate to late preterm”. The GA at birth carries

- 4 -

a wealth of information about the potential risks of poor outcomes resulting from greater

immaturity of the lungs, brain and other developmental systems. More medical

interventions and an increased risk of their potential complications are consequences of

a lower GA at birth (Leviton et al, 2005) with GA being more informative than birthweight

for these risks.

Preterm birth often occurs as a result of several maternal (e.g. pre-eclampsia,

diabetes), obstetric (e.g. short cervix, placenta previa, chorioamnionitis) and/or fetal

(multiple gestations, genetic disorders) causes. Poorer outcomes in an infant are often

linked to specific causes of preterm labour, with several perinatal factors associated with

delivery, such as asphyxia, inflammation, fetal growth restriction and major birth defects

all occurring more commonly in those infants with poorer outcomes (Delorme et al, 2016;

Nelson and Blair, 2015).

There is significant epidemiological data describing the marked worldwide

variations in outcomes of infants born preterm (Hossain et al, 2015; Shah et al, 2016).

Current data supports the resuscitation of infants ≥23 weeks GA, considered by many

experts in the field to be the limit of viability for the human infant. A “gray zone” still exists

<23 weeks and <500 grams where infant survival to neonatal intensive care unit (NICU)

discharge ranges between 0% and 50%, with wide variation between centres (Rysavy et

al, 2015; Seri and Evans, 2008; Stoll et al, 2010). Neurodevelopmental outcomes of

preterm-born infants have been shown to be significantly impaired compared to their

peers born at term-age, with the greatest percentage of impairments seen within the

extremely preterm group (Marlow et al, 2005; Stoll et al, 2010). However, a significant

proportion of children born near the limits of viability at 23-24 weeks will perform within

- 5 -

the expected school norms at early school age, providing meaningful reassurance that

resuscitation of these infants should be continued (Garfield et al, 2017; James et al,

2017).

In helping to understand the clinical factors that have the greatest impact upon the

long-term development of the infant born preterm, large epidemiologic studies have been

reported. One particular study of 5 year outcomes reported increased mortality and

morbidity independently with each of brain injury, bronchopulmonary dysplasia (BPD) and

retinopathy of prematurity (ROP) (Schmidt et al, 2015). They were also able to show that

in those infants with a combination of morbidities the risk for poor outcomes increased

significantly from 11% in those with none of the above morbidities to 62% in those with

all three (Figure 1.1) (Schmidt et al, 2015). What causes such a large percentage of

infants with these conditions to have poor outcomes remains an important area of

research in the care of the preterm infant.

- 6 -

Figure 1.1 Number of Neonatal Morbidities and Probability of Poor Outcomes at 5

years. Probability of poor 5-year outcome in infants; 95% CI shown by error bars. Linear

line indicates predictions based on the fitted morbidity count model. Horizontal dotted

line indicates the overall probability of poor 5 year outcome. Republished from Schmidt

et al (2015) Journal of Pediatrics with permission from Elsevier.

1.3 Brain injury in Preterm Infants

1.3.1 White Matter Injury

Severe white matter injury (WMI), known today as cystic periventricular

leukomalacia (PVL), was first described by Banker and Laroche in 1962 based upon

pathological specimens of infants who were found to have lesions consisting of necrosis

and macrophage activity within the periventricular white matter (Banker and Larroche,

1962). The identified cause of PVL in the majority of these cases was severe anoxia, with

- 7 -

WMI now seen as the major form of brain injury recognized in survivors of preterm birth,

with the greatest risk to infants <32 weeks PMA (Volpe, 2009).

PVL is often caused by cerebral ischemia because of hemodynamic instability with

hypoxia-ischemia or inflammation as a result of intrauterine or postnatal infections.

During the development of the cerebral vascular bed in the preterm infant, vascular

watershed regions develop. These areas are vulnerable to reductions in perfusion as they

are furthest from the arterial supply and their blood and oxygen delivery supply is at the

border of two vascular territories. In the preterm infant brain, the white matter surrounding

the periventricular region is in a major watershed zone. Thus, PVL has historically been

considered a form of watershed ischemic brain injury of the preterm infant.

During the initial phase of WMI the major cells thought to degenerate are the pre-

oligodendrocyte (OL) lineage cells. Axonal injury can also occur and is a prominent

feature of cystic WMI where necrosis is present, such as seen in classical cystic PVL.

Though in a recent series, cystic PVL comprised only about 5% of the total burden of WMI

(Haynes et al, 2008; Kinney and Back, 1998). Disturbances in the normal myelination

pathway are initiated by the selective vulnerability of the late pre-OL in the most common

form of WMI at 23 – 32 weeks GA (Back and Miller, 2014; Back et al, 2001).

- 8 -

Figure 1.2 Severity of WMI. MRI examples of (A) mild (B) moderate and (C) severe white

matter injury on T1 1mm sagittal images of preterm infants. Scoring scale adapted from

Miller et al (2005) Journal of Pediatrics. Arrows highlight the areas of T1 hyperintensity

which are marked as areas of WMI.

Figure 1.3 WMI pathophysiology. Proposed pathogenic mechanisms with distinct

differences in periventricular leukomalacia [PVL] (upper pathway) and diffuse WMI (lower

pathway). More severe events of hypoxia-ischemia results in cystic PVL, with milder

events resulting in selective pre-oligodendrocyte (OL) death and myelination failure.

- 9 -

Reprinted from Back and Miller (2014), Annals of Neurology with permission from

publisher John Wiley and Sons.

The predominant form of WMI in the preterm infant is diffuse WMI, which is more

commonly seen in those infants with necrotizing enterocolitis (NEC), infection, and

hypoxia-ischemia, and which results in poorer outcomes than expected for a neonates’

GA (Glass et al, 2008; Shah et al, 2008). Through hypoxia-ischemia and inflammation,

glutamate-mediated injury occurs to the pre-OL cell leading to selective cell death and

loss of maturation-dependent cellular processes (Khwaja and Volpe, 2008). It remains

unknown the extent to which the individual and combined effects of hypoxia-ischemia or

inflammation result in WMI, with white matter cellular injury described in animal models

of both (Hagberg et al, 2002). There remains significant debate in the literature

considering the effects of both inflammation and hypoxia-ischemia in WMI, with some

experts suggesting it is predominantly a result of inflammation (Gilles et al, 2017) and

others suggesting it’s predominantly from hypoxia-ischemia (Hagen et al, 2014). What

seems most likely is that the developing cells are susceptible to the combined effects of

both inflammation and hypoxia-ischemia, resulting in cellular injury (Back, 2006; Back and

Rivkees, 2004; Khwaja and Volpe, 2008; Penn et al, 2016). In addition, inflammation can

be induced by hypoxia-ischemia, resulting in a greater cellular impairment, while hypoxia-

ischemia can occur from inflammation, such as from hypotension during an infection. The

best description of these effects refers to the “upstream” mechanisms of hypoxia-

ischemia and inflammation in activating brain microglia, followed by the “downstream”

mechanisms of excitotoxicity with generation of free radicals resulting in the subsequent

- 10 -

death of the vulnerable pre-OL (Khwaja and Volpe, 2008). Regardless of cause, as we

gain a greater understanding of the pathophysiology of WMI we will be better prepared to

evaluate new strategies in its prevention and treatment.

The outcomes of children with classical PVL range from CP to milder motor

impairments (Fazzi et al, 1994; Hamrick et al, 2004; Miller et al, 2000; Spittle et al, 2011).

The outcomes of diffuse WMI are less robust in the literature with motor and cognitive

impairments described (Woodward et al, 2006; Woodward et al, 2012). The outcomes of

infants following diffuse WMI are best predicted by the location of the injury, with presence

of frontal lobe WMI the strongest predictor of poor motor, language and cognitive

outcomes (Guo et al, 2017). With the development of advanced MR techniques in the

preterm infant we are gaining further information about white matter maturation (Hoon et

al, 2009; Miller et al, 2002), connectivity (Smyser et al, 2013) and regional abnormalities

(Pierson et al, 2007) seen in white matter injury of the preterm infant.

1.3.2 Intraventricular Hemorrhage

Intraventricular hemorrhage (IVH) is a condition in which there is hemorrhage in

the ventricular system of the brain, or the sub ependymal zone immediately adjacent to

it. IVH is especially common in preterm infants with a highest period of injury at <30 weeks

due to the fragility of the germinal matrix at this time. The germinal matrix is a region of

blood vessels immediately adjacent to the lateral ventricles which arises during fetal

development and usually disappears before 35 weeks GA. When born prematurely,

infants are at risk of rupturing the germinal matrix, particularly those infants requiring

resuscitation with fluctuations in blood pressure (Ballabh, 2010). IVH is categorized into

grades of severity with grade I considered mild, grade II moderate and grades III and IV

- 11 -

severe. IVH grades I and II have a low possibility of long-term damage as the blood does

not cause excessive pressure or occlude the normal ventricular system. However, in

grades III and IV there is frequently blood occluding the ventricular system resulting in a

backflow of fluid, potentially irreversibly injuring the brain. Post-hemorrhagic ventricular

dilatation often develops slowly in causing injury to the brain parenchyma, and while the

optimal management of this condition remains uncertain, treatment with a CSF reservoir

and/or ventricular-peritoneal shunt remains standard therapy. Improvements in antenatal

care of the pregnancy at-risk for preterm birth has resulted in reduced rates of IVH for

those infants exposed to antenatal steroids across all gestational ages (Wei et al, 2016),

suggesting preventative therapies can effectively reduce brain injury in this population.

The extent of the injury and the resultant neurodevelopmental impairments remains a

function of the severity of the hemorrhage and the location of any parenchymal injury

(McCrea and Ment, 2008). When accompanied by parenchymal injury there is a greater

risk of cerebral palsy, low mental and motor scores and visual and hearing impairments

(Vohr et al, 2003); though without associated parenchymal injury the risk of adverse

outcomes is low (Han et al, 2002; Linsell et al, 2016; O'Shea et al, 1998).

1.3.3 Cerebellar Hemorrhage

The cerebellum is a compact and particularly important anatomical region involved

in neurodevelopment of the preterm infant, encompassing more than 3.5 neurons for

every neuron in the cerebrum (Herculano-Houzel, 2010). Its role in motor, language and

cognitive development is evident. Injury to the cerebellum, such as a hemorrhage, is

shown to result in significant adverse outcomes with greater rates of autism, motor

impairments and an abnormal neurological examination (Tam et al, 2011; Tam et al,

- 12 -

2009; Ure et al, 2016; Wang et al, 2014). The size of cerebellar hemorrhage is an

important predictor of outcomes, with small hemorrhages (<4mm on MRI) not seen to be

associated with poorer outcomes (Steggerda et al, 2013). In addition, the location of the

hemorrhage is also of importance, with injury to the vermis thought to confer a greater

risk of impairments than injury to a hemisphere (Hashimoto et al, 1995). Delays in the

cerebellar development with poorer neurodevelopmental outcomes are also seen

following cerebellar hemorrhage (Lee et al, 2016; Messerschmidt et al, 2008). With the

exploration and expansion of research into the cerebellum the medical community has

learned that the cerebellum in the preterm infant is “rapidly developing, vulnerable and

clinically important” (Volpe, 2009).

1.4 Retinopathy of Prematurity

1.4.1 Background

Retinopathy of prematurity (ROP) is a vascular proliferation disorder of the

developing retina, and is the leading cause of visual impairment and blindness in infants

born preterm (Rivera et al, 2011). First described in the late 1940s, ROP appeared

suddenly in those infants who were the first to receive supplemental oxygen in closed

incubators. Initially called “retrolental fibroplasia,” meaning “proliferation of fibrous tissue

behind the lens,” ROP had severe consequences with retinal detachment leading to

blindness in the most severe cases (Hellstrom et al, 2013). As the shift away from

unrestrained oxygen therapy has been made in NICUs throughout the developed world,

- 13 -

there remains a delicate balance of oxygenation between lung maturity and survival with

poorer visual outcomes (Hellstrom et al, 2013).

ROP severity is scored clinically by a pediatric ophthalmologist via a dilated eye

examination using an indirect ophthalmoscope with examinations starting in those infants

born <31 weeks GA. The first examination is normally planned for around 4 weeks

following delivery. Two components are utilized in scoring, - the zone and the grade. The

location of the abnormal vessel growth, referred to as the “zone”, is determined by

proximity to the central vision and optic nerve. Zone 1 includes abnormal vessel growth

in the central retina, whereas zone 3 is when abnormal vessel growth occurs in the

peripheral retina (Figure 1.4).

Figure 1.4 Retinopathy of Prematurity (ROP) scoring scale. Showing zone borders

and clock hours used to describe the location and extent of ROP. Figure originally

published in Pediatrics (2006), reproduced with permission from the American Academy

of Pediatrics.

The second component, the “stage,” ranges from 1 to 5, with stages 1 and 2

considered mild, and stages 3 to 5 considered severe, on which treatment is usually

- 14 -

provided according to the Early Treatment for ROP trials (Early treatment of Retinopathy

of prematurity cooperative group, 2003) (Figure 1.5).

Figure 1.5 Stages of Retinopathy of Prematurity. Stage 1 is characterized by a thin

demarcation line between non-vascularized and vascularized retina, stage 2 by a ridge,

stage 3 by extraretinal fibrovascular proliferation, stage 4 by part retinal detachment, and

stage 5 by total retinal detachment. Figure originally published by Hellstrom et al (2013)

in The Lancet; reproduced with permission from Elsevier.

The current mainstay of treatment is with laser photocoagulation therapy to stop

the growth of the abnormal vessels and to prevent retinal detachment. Other approved

therapies include the intravitreal injection of anti-VEGF agent Bevacizumab (Trade name

Avastin) (Gunther and Altaweel, 2009) and cryotherapy in which reginal retinal destruction

was performed through freezing of the area (Pearce et al, 1998). Other therapies of

interest include scleral buckling, which is aimed to treat retinal detachments, propranolol,

to reduce the progression of ROP, and recombinant humanized IGF-1 to prevent

- 15 -

excessive vessel formation through reducing VEGF (Hellstrom et al, 2003; Hellstrom et

al, 2013). The neurodevelopmental outcomes following treatment for severe ROP are still

not understood.

1.4.2 Pathophysiology

During embryogenesis, the retina, ciliary body, iris and optic nerves arise from the

diencephalon, which goes on to form the caudal forebrain; thus the retina develops as

part of the central nervous system (CNS). The retina includes layers of cells with three

neural cells including the photoreceptor cells, bipolar cells and ganglion cells, and a fourth

layer of pigmented epithelial cells. These layers are complete in their development by

around 16 weeks PMA, though are not vascularized. Normal retinal vascularization

continues until term age through the balance of growth factors of insulin-like growth factor

(IGF-1), vascular endothelial growth factor (VEGF), erythropoietin (EPO) and ω3 poly-

unsaturated fatty acids (ω3 PUFA), among others. Following preterm birth, normal

vascularization is inhibited and normal vascularization may continue if the optimal

environment for retinal development is achieved. However, should there be a relative

imbalance of growth factors with increasing metabolism, hyperoxia or hypoxia, the

potential for abnormal retinal neovascularization occurs (Hellstrom et al, 2013).

Of particular importance to reducing the development of ROP is the prevention of

hyperoxia with reduced oxygen supplementation in the care of the preterm infant.

Because of improved NICU care and practices, ROP is relatively uncommon in those

infants born >30 weeks PMA, though a significant proportion of infants <28 weeks will

have at least some retinal vasoproliferation. Furthermore, infants who are small for

gestational age, and those with hyperglycemia, hyperinsulinemia, and/or postnatal

- 16 -

infections, are at greater risk for ROP and are more likely to require treatment to preserve

vision (Hellstrom et al, 2013).

Figure 1.6: Retinopathy of Prematurity Pathophysiology. (A) In-utero: normal vascular

growth with low oxygen tension. (B) Phase 1: retinal vascularization is inhibited by

hyperoxia and loss of growth-factors from the placenta. Blood-vessel growth stops and

with retinal maturation hypoxia results. (C) Phase 2: hypoxic retina stimulates expression

of oxygen-regulated factors such as erythropoietin (EPO) and vascular endothelial growth

factor (VEGF) which in turn stimulate retinal neovascularization with Insulin-like growth

factor 1 (IGF-1). (D) Resolution: retinopathy may be prevented or treated with prevention

of phase 1 or inhibition of phase 2 with laser therapy or an antibody. ω3 PUFA= ω3

polyunsaturated fatty acids. Figure originally published by Hellstrom et al (2013) in The

Lancet; reproduced with permission from Elsevier.

Low IGF-1 has a strong association with later ROP due to poor retinal vascular

growth. This is likely due to its impacts upon VEGF, which is an important factor in normal

vascularization (Hellstrom et al, 2013) (Figure 1.6). IGF-1 is also an important growth

- 17 -

factor for the fetus in-utero and remains an important regulator of glucose metabolism in

postnatal development (Cheng et al, 2000). Low IGF-1 concentrations in the early

postnatal period in babies born preterm correlates strongly with later ROP and other

comorbidities, including a slower growth rate of brain volume, a marker of cerebral

development (Hansen-Pupp et al, 2011; Hellstrom et al, 2003). Low IGF-1 concentrations

are nutrition-dependent markers and can be reduced following starvation, infection and

stress (Demendi et al, 2012). In addition, hyperglycemia within neonates, absence of ω3

PUFAs in diet and low levels of EPO are identified as other potential risks for increased

ROP (Hellstrom et al, 2013).

1.4.3 Outcomes of Infants with ROP

In assessing the neurodevelopmental outcomes of children with ROP, Schmidt et al

reported an odds ratio of 4 for death or disability at 5 years, greater than that of brain

injury or BPD (Schmidt et al, 2015) with a 3 to 4 times greater risk for non-visual

disabilities than those without ROP (Msall et al, 2000; Schmidt et al, 2014). While vision

abnormalities are common in preterm infants (O'Connor et al, 2002), visual outcomes

have improved in the CRYO-ROP (Cryotherapy for Retinopathy of Prematurity Group,

1996), early treatment for ROP (Early Treatment For Retinopathy Of Prematurity

Cooperative, 2003) and Avastin trials (Martinez-Castellanos et al, 2013; Mintz-Hittner et

al, 2011). Common visual disturbances seen in long-term follow-up of children with ROP

include strabismus, amblyopia and refraction abnormalities (O'Connor et al, 2002). The

impacts of these visual disturbances are thought to be minor, though when severe, early

visual-motor coordination and vestibular system dysregulation can effect developmental

mechanisms (Goyen et al, 2006; Prechtl et al, 2001). The association between severe

- 18 -

ROP and brain maturation and long-term neurodevelopmental outcomes remains an area

in which little is known.

1.5 Infection in Preterm infants

1.5.1 Background

In many large epidemiological studies of preterm infants worldwide, infection has

been determined to be a strong predictor of poorer outcomes compared to non-infected

infants (Kiechl-Kohlendorfer et al, 2009; Mitha et al, 2013; Rand et al, 2016; Schlapbach

et al, 2011; Stoll et al, 2004; Van der Ree et al, 2011). Rates of infection in preterm infants

range between 20 – 65% within industrialized countries, with even greater rates reported

in developing countries, where mortality rates are also significantly higher (Adams-

Chapman and Stoll, 2006; Orsi et al, 2009; Stoll et al, 2004; Zaidi et al, 2005). It is for

these reasons that infection remains an important determinant of childhood outcomes and

prevention strategies remain a key initiative of the World Health Organization.

1.5.2 Classification

Traditionally, neonatal infections are divided into two categories based upon the

timing of the infection in relation to birth; [1] early-onset infection, classified as presenting

within the first 72 hours of life, and [2] late-onset infection, presenting past 72 hours of

life. This division has been utilized to assist in the differentiation of the early-onset

infections, which are thought to be a result of maternal risk factors or transmission and

acquisition of the infection via the birth canal, in comparison to late-onset infections which

- 19 -

are more often related to acquisition in the hospital environment. It is because of these

factors that they are also referred to as “pre-natal” and “postnatal” infections.

1.5.3 Pre-Natal Infection

Infections occurring in the first 72 hours of life appear to have a different

pathophysiology than that of postnatal infections, evident by transmission via the maternal

bloodstream or via direct exposure to the organism ascending to within the uterus or from

delivery through the vaginal canal. Chorioamnionitis, the most common pre-natal

infection, involves infection of the fetal membranes, placenta or amniotic fluid, and is

commonly diagnosed in the third trimester with associated maternal fever, leukocytosis,

uterine tenderness and foul-smelling amniotic fluid with fetal tachycardia. Current

treatment of chorioamnionitis consists of appropriate intravenous antibiotic treatment

early in the maternal illness and may include postnatal antibiotics following delivery. Pre-

natal infection is a known cause for stillbirth as well as a common cause of preterm

delivery (Goldenberg et al, 2008; McClure et al, 2010). Infants delivered following a pre-

natal infection have a higher mortality rate and are commonly infected with organisms

such as Escherichia coli and group B Streptococcus (GBS), which are less commonly

seen in postnatal infections (Jiang et al, 2004; Stoll et al, 2005).

The literature with regards to whether pre-natal infection confers a greater risk of

brain injury and poorer outcomes remains unclear with some studies reporting increased

rates of cerebral palsy and cystic PVL (Dammann et al, 2002; Dammann and Leviton,

1998; Grether and Nelson, 1997; Mitha et al, 2013), while other studies report no

significant difference in WMI, brain maturation or poorer outcomes (Chau et al, 2009;

Grether et al, 2003). While the neurodevelopmental outcomes resulting from pre-natal

- 20 -

infection alone remain unclear, there is research supporting that the combined influence

of pre-natal and postnatal infections together may confer a stronger risk for cerebral injury

and cerebral palsy (Mitha et al, 2013; Yanni et al, 2017).

1.5.4 Postnatal Infection

Infections occurring after 72 hours of life are more common than pre-natal

infections and have a far greater association with poorer outcomes, with a higher

percentage of infected infants developing motor and/or cognitive impairments in follow-

up compared to uninfected infants (Jiang et al, 2004; Stoll et al, 2004). Studies reporting

long-term outcomes at school age continue to support the association of infection with

greater frequency of cognitive and motor dysfunction, as well as ADHD and other mental

health impairments (Mitha et al, 2013; Rand et al, 2016).

Published research of clinical cohorts exploring postnatal infection have

investigated the different types of infections separately, with the hypothesis that sepsis

and meningitis would be more likely to result in poorer outcomes. However, among all

infants with postnatal infections there is little difference between those with “clinical

infection alone” and those with sepsis and/or meningitis (Stoll et al, 2004). The same has

been shown for necrotizing enterocolitis (NEC), thought to convey an increased

inflammatory response, resulting in a greater risk of WMI and motor impairments (Shah

et al, 2008; Stoll et al, 2004). In multiple cohorts of preterm infants, postnatal infection

was associated with an increased risk of WMI, as well as worsened WMI seen on

subsequent MRIs, highlighting it as an important risk factor for brain injury in the preterm

newborn (Chau et al, 2009; Glass et al, 2008). In addition to increases in WMI, infants

with postnatal infection have delayed brain development using measures of white matter

- 21 -

maturation with increased average diffusivity, decreased white matter fractional

anisotropy and lower MRSI metrics of brain maturation (Chau et al, 2012). The link

between multiple postnatal infections and development of the preterm brain has not been

thoroughly investigated and remains an area of interest in the field.

Common organisms involved in postnatal infection include gram-negative and

gram-positive bacteria, as well as fungal species, with coagulase-negative

staphylococcus (CONS) species the most common (Orsi et al, 2009). Other organisms

such as Klebsiella pneumoniae, Escherichia coli, Enterococcus species, Enterobacter

species and Candida species are less commonly seen (Orsi et al, 2009). Among the

organisms implicated in postnatal infections, it has been shown that infants are at risk of

poorer outcomes as a result of any infectious organism (Stoll et al, 2004), suggesting that

the impacts on the brain are not specific to one organism, but rather related to the

inflammatory response that occurs during the infection and the impacts on the

hemodynamic circuitry and immune system of the body.

Implementing tighter infection control measures within NICUs, such as increased

hand hygiene by health care workers, reduces infection rates, particularly of hospital-

acquired infections (Schelonka et al, 2006; Won et al, 2004). The implementation of

improved infection control measures can result in reductions in the length of hospital stay

and of other co-morbidities, and is even associated with improved outcomes at 2 years in

one study (Davis et al, 2016). However, it is not fully understood whether these

procedures, and other more costly measures, need to be put in place for an entire NICU

unit, or whether the most vulnerable infants would benefit better through an individualized

approach. Additionally, there is no testing of these personalized or “precision” infection

- 22 -

control measures specifically targeted at those infants at greatest risk of adverse

outcomes.

1.5.5 Neonatal Inflammation

When an infectious pathogen makes contact with a mature host’s immune system,

the innate immune system identifies and eliminates the invading pathogen through the

activation of pattern recognition receptors (PRRs). Within the preterm infant, however,

the innate immune system is not well developed, with a deficiency of PRRs, which results

in an unbalanced and potentially harmful inflammatory response (Kan et al, 2016; Van

der Poll et al, 2017). This heightened vulnerability of the preterm infant brain to infection

was first illustrated and described by Gilles in 1976 when a bacterial lipopolysaccharide,

injected systemically into newborn kittens, caused a leukoencephalopathy, though it

resulted in no effect in the mature cat (Gilles et al, 1976). The activation of toll-like

receptors within the human epithelium of the bloodstream, known to be activated by

bacterial lipopolysaccharides, is believed to be one mechanism involved in the activation

of the local host response on the blood brain barrier (Strunk et al, 2014). This results in

the production of various inflammatory mediators, with cytokines, prostaglandins and

reactive oxygen species (ROS) released into the CNS causing direct cytotoxicity on the

brain; and may occur even without direct bacterial invasion into the CNS, such as in

meningitis (Strunk et al, 2014).

Increased levels of cytokines tumor necrosis factor (TNF-α) and interleukins (IL-

1β, IL-6, IL-8, IL-10, IL-12, IL-17 and IL-18) have been investigated in the fetal and

neonatal responses to infection, with some shown to have a pro-inflammatory role within

the developing immune system (Dammann and Leviton, 1997; Leviton et al, 2005; Van

- 23 -

der Poll et al, 2017). In those infants with higher levels of these pro-inflammatory markers,

cerebral injury is more commonly seen, with greater extent of white matter injury and PVL

(Duggan et al, 2001; Viscardi et al, 2004). In animal models of fulminant infection, blocking

some of these cytokines has been shown to result in increased survival and has been

suggested as a potential therapeutic target against the hyper-immunity in sepsis (Flierl et

al, 2008).

The impact of multiple inflammatory events has been investigated in what is referred

to as the “two-hit” model, in which an initial inflammation sensitizes the vulnerable immune

system to a subsequent inflammatory event. The second inflammation event is thought

to have a threshold stimulus lower than is required for the initial event and results in the

release of excitotoxic inflammatory markers, leading to cerebral injury and potential

epigenetic alterations (Fleiss and Gressens, 2012; Fleiss et al, 2015; Yanni et al, 2017).

Both pre-natal and postnatal stimuli have been shown to sensitize the immune system

with increased levels of pro-inflammatory cytokines (Yanni et al, 2017). Pre-natal

infections, and in particular chorioamnionitis, are often associated with a fetal

inflammatory response in which pro-inflammatory markers are increased and thought to

act as an initial trigger (Wang et al, 2007), with some studies reporting a greater incidence

of WMI and cerebral palsy as a result (Leviton et al, 2010; Yanni et al, 2017). Postnatal

events, such as prolonged mechanical ventilation, NEC and postnatal infection, often

trigger similar inflammatory responses to those seen in pre-natal infections, resulting in a

significant release of pro-inflammatory markers (Aden et al, 2010; Hagberg et al, 2015;

Volpe, 2008). Furthermore, it has been shown that exposure of the developing brain to

multiple inflammatory events leads to excitotoxicity, with greater mitochondrial

- 24 -

impairment and weakened vascular integrity potentially causing direct brain injury or

interfering with normal CNS development (Boisse et al, 2004; Fleiss et al, 2015; Wang et

al, 2009; Yanni et al, 2017), and potentially altering the epigenome (Fleiss and Gressens,

2012). Multiple inflammation events have the potential to result in an altered immune

response leading to a chronic inflammatory condition.

1.5.6 Neonatal Hypoxic-Ischemia Injury

The preterm infant brain is particularly susceptible to the effects of hypoxia-

ischemia as reflected in the development of white matter injury and PVL. The

susceptibility of the OL progenitor cell is a maturation-dependent phenomenon, which is

represented by a greater resistance of the mature OL cells to the effects of oxidative

stress than the immature pre-OL cell (Back et al, 1998). What results is a critical period

of vulnerability for the preterm infant during which irreversible injury can occur to the OL

cell, resulting in myelination dysfunction. The over-expression of the immune system

during this period may contribute to the cerebral injury as a consequence of the pressure-

passive and poor auto-regulatory system of the preterm infant (du Plessis, 2009; Khwaja

and Volpe, 2008). We know that cardiovascular collapse commonly occurs during an

acute infection (Healy et al, 2004), which when during a vulnerable period, may potentiate

the hypoxic-ischemic injury (Khwaja and Volpe, 2008). Animal models have shown the

combined effects of hypoxia-ischemia insults and infection can worsen neuronal injury,

with particular periods in which there is greater susceptibility to injury (Eklind et al, 2005;

Larouche et al, 2005).

There are various mechanisms which make the preterm infant brain vulnerable to

the effects of hypoxia (Gopagondanahalli et al, 2016). Firstly, the preterm brain receives

- 25 -

a relatively reduced global and regional supply of blood compared to the term infant brain,

which may result in a limited margin for reduced supply (Altman et al, 1988; Pichler et al,

2014). Additionally, arterial extension into the developing brain is incomplete resulting in

poorly vascularized end-zone watershed regions, particularly of the periventricular white

matter regions (Altman et al, 1988; Inage et al, 2000). Include these factors with the poor

cardiac output commonly seen in the preterm infants (Kluckow and Evans, 2000), and the

poor correlation between blood pressure monitors and cerebral blood flow (Weindling and

Kissack, 2001), and it becomes clear that poor cerebral blood flow is likely

underestimated in the clinical care of the preterm infant.

The pressure passive blood pressure system of the preterm infant brain is

vulnerable to alterations in blood pressure, in which the greatest fluctuations are seen in

the smallest infants (Soul et al, 2007). The regulation of blood pressure in the preterm

infant has a poor reserve for fluctuations and is dysfunctional in response to hypotension

(Munro et al, 2004). Treatment with inotropes and fluids, and other appropriate therapies

for low blood pressure, can improve cerebral blood flow, reducing cerebral ischemia

(Munro et al, 2004). Poor cerebral autoregulation is associated with a higher frequency

of IVH and WMI, thus preserving a stable mean arterial blood pressure is paramount in

reducing injury to the developing brain (O'Leary et al, 2009; Vesoulis and Mathur, 2017).

Further potentiating the blood pressure fluctuations in the brain are the effects of

hyperoxygenation and hypocarbia, resulting in vasoconstriction of the blood vessel wall,

with greater amounts of brain injury (Fujimoto et al, 1994; Khwaja and Volpe, 2008).

Hypocarbia is also a strong factor in the development of severe ROP, with changes in the

- 26 -

oxygen targets and more awareness of hyperoxygenation and hypocarbia helping to

reduce the severity and incidence of ROP (Sears et al, 2009).

1.6 Magnetic Resonance Imaging

1.6.1 Imaging the Preterm newborn

Brain imaging in the very preterm newborn has developed within the last two

decades to be the gold standard assessment for brain injury, and assists in identification

of those infants at high-risk for adverse outcomes in infancy and childhood (Anderson et

al, 2017; Woodward et al, 2006). Magnetic resonance imaging (MRI) of the preterm

newborn, in many instances without the need for sedation, is a safe and reliable technique

with a greater sensitivity for white matter abnormalities than cranial ultrasound (Inder et

al, 2003; Miller et al, 2003; Plaisier et al, 2012). Following the use of routine term-

equivalent MRI in some centres, 25-33% of infants were found to have brain

abnormalities, many of which were not detected with conventional cranial ultrasound

(Kidokoro et al, 2014; Neubauer et al, 2017). The timing of the MRI is an important

consideration in detecting injury in the preterm infant as WMI is more easily seen in the

preterm period, and may be undetectable on subsequent imaging at term-equivalent age

(Kersbergen et al, 2014; Martinez-Biarge et al, 2016). As a result of this, it is

recommended that serial imaging of the preterm brain be done as this improves the

sensitivity for WMI, providing a more definite prediction of childhood outcomes (Martinez-

Biarge et al, 2016; Sarkar et al, 2015). However, MRI does have its limitations with the

current clinical scanners unable to visualize microscopic WMI, the least severe form of

- 27 -

WMI, which remains a neuropathological diagnosis, and suggests that MRI is liable to

underestimating the full extent of WMI (Back et al, 2012; Volpe, 2017). With the addition

of advanced MR techniques, these milder injuries and delays in development may be

more readily assessed allowing a further understanding of the long-term effects of

prenatal and postnatal injuries (de Vries et al, 2011; Guo et al, 2017; Kwon et al, 2014;

Rutherford et al, 2005; Thomas et al, 2005).

1.6.2 Overview of MRI Basics

MRI utilizes the nuclear magnetic resonance properties particularly of hydrogen

protons, which when combined with radiofrequency (RF) waves, results in energy release

which can be detected, interpreted and transformed into images. The images are formed

as a result of the physical properties of the molecules and their response to the magnetic

field. When first placed within a strong magnetic field, such as that provided by the main

B0 field, the hydrogen protons within the body polarize. The RF pulse is then used to kick

spins off their axes, which results in excitation when applied at the correct resonance

frequency, flipping the natural spins of the hydrogen atoms into the anti-parallel state.

This is known as the “Larmor Frequency” and it depends on the field strength of the

magnet (42.6 MHz/Tesla). As the spins return, or relax, from the high energy anti-parallel

state back to their equilibrium they release energy in the form of an oscillation magnetic

field at the Larmor frequency.

For visualization, the vector representation is used such that when excited by the

RF pulse, the net storage of energy is represented by the tipping of the vector onto its

side, away from the z-axis and into the x-y planes, the amount referred to as the ‘Flip

angle’. Following this, the vector begins to precess around the z-axis slowly returning to

- 28 -

equilibrium through the relaxation phase. The relaxation of the protons can be divided

into two processes: T1 and T2 relaxation, described below.

1.6.3 MRI sequences

T1 relaxation, also referred to as spin-lattice relaxation, refers to the exponential

recovery of the protons to the z-axis. T2 relaxation, also known as spin-spin relaxation,

refers to the loss of phase coherence among excited protons that leads to exponential

decay of the component of the magnetization in the transverse plane. Variations in the

relaxation rates of T1 and T2 signal occurs as a result of the diverse concentrations of

molecules in the different tissues within the brain. Fat, which has tighter bonds to

surrounding structures than water, releases energy quickly; while water results in a slower

relaxation, resulting in water having a higher T1 and higher T2 signal than fat.

The signal received by the receiver coils during the relaxation phase, is acquired

in a temporary image space, also referred to as ‘k-space’, in which the digitized MR

signals are stored during data acquisition. The images are then produced using the

mathematical conversion known as the “Fourier transform”. The k-space is expressed as

a summation of overlapping sinusoids that are used to produce an anatomical image.

The repetition time (TR) and echo time (TE) play key roles in the contrast of a MR

image due to the variations in the recovery and relaxation times of different tissues. The

TR refers to the time, in milliseconds, between successive RF excitation pulses, with the

TE reflecting the time, also in milliseconds, from the application of the RF excitation pulse

and the peak signal induced by the coil. T1-weighted sequences maximize T1 contrast

- 29 -

by utilizing a short TR and short TE, increasing tissue contract. T2-weighted sequences

use a longer TE and long TR, minimizing the T1 contrast.

Two-dimensional multi-slice imaging is produced by sequentially exciting and

collecting data from one slice region at a time. The thickness of the tissue excited in a

single image slice of the MR image is determined by changes in the RF pulse and/or the

gradient strength. These images are then stacked together to produce a combined 3D

image of the brain. Three-dimensional (3D) volumetric imaging is produced by exciting a

thick slab of tissue, which is then spatially encoded by traversing the 3D k-space,

reconstructed in 3D and resliced for image review.

1.6.4 Diffusion Weighted Imaging and Diffusion Tensor Imaging

Diffusion weighted imaging (DWI) is a MR technique that measures free motion of

water in tissue, based upon measuring the random Brownian motion of water molecules

within a voxel of tissue. Unlike the free diffusion, or movement, of water inside a container,

diffusion of water inside a voxel of the brain is hindered by cellular membrane boundaries

in both the intracellular and extracellular compartments. Using these properties, DWI can

be utilized in assessing micro-structural architecture and is sensitive to cell changes as a

result of ischemia and those seen in highly cellular tumours. Images are produced by

adding extra gradients to standard MR imaging sequences, such as a T2-weighted echo

planar sequence, which sensitizes the sequences to molecular diffusion. The diffusion

gradients are equal in magnitude and centered on a 180-degree refocusing

radiofrequency pulse. DWI images are produced as static hydrogen protons are

unaffected by the diffusion gradient and will retain their signal, whereas water molecules

- 30 -

moving in the axis of the gradient will accumulate phase by the first gradient, but not the

second and hence will lose their signal.

Diffusion tensor imaging (DTI) is another quantitative MR method that uses the

same properties of diffusion of water molecules in conventional DWI imaging, though

does so using a Gaussian model of diffusion. The Gaussian distribution is defined by a

3x3 symmetric, positive definite matrix, with 3 orthogonal (mutually perpendicular)

eigenvectors and 3 positive eigenvalues in each voxel. The direction of fastest diffusion,

the principle eigenvector (λ1), is referred to as axial diffusivity, with the other eigenvectors

(λ2, and λ3) termed radial diffusivity. These eigenvectors, in combination with the 3

eigenvalues define an ellipsoid reflecting the diffusion probability within each voxel

(Figure 1.7).

Figure 1.7 Representation of the 3D diffusion ellipsoid. Presented as an ellipsoid

with three unit eigenvectors (ε1, ε2, and ε3), with corresponding lengths (λ1, λ2, and λ3),

the eigenvalues.

DTI characterizes the 3D spatial distribution of water diffusion in each voxel of the

MR image, providing an indirect measure of microstructure (Beaulieu, 2002; Miller et al,

λ1 ε1

λ2 ε2

λ3 ε3 y

z

x

- 31 -

2002; Mukherjee et al, 2002). Within each voxel it is possible to infer the axial diffusivity

(λ1), the radial diffusivity (λ2 and λ3), and mean diffusivity (MD), a measure of the average

size of diffusion in each voxel, where Dxx, Dyy and Dzz are the diagonal terms of the

diffusion tensor:

MD = λ1 + λ2 + λ3 = Dxx + Dyy + Dzz = Trace

3 3 3

Molecules diffuse differently within tissues depending on the type, integrity,

architecture and the presence of barriers generating a quantitative anisotropy. Anisotropy

refers to the diffusion of water which may be unrestricted in all directions, such as within

CSF, termed “isotropic”, or restricted to a single plane, such as in the white matter

pathways of the mature brain, characterized as “anisotropic” (Soares et al, 2013).

Fractional anisotropy (FA), the normalized measure of the fraction of the tensor’s

magnitude due to anisotropic diffusion, corresponds to the degree of directionality and

ranges from 0 (isotropic diffusion) to 1 (anisotropic diffusion), averaged between bilateral

brain regions in mean FA analyses (Soares et al, 2013). FA increases with white matter

maturation, and is thought to reflect the maturation of the oligodendrocyte lineage and

early events of myelination within the preterm brain (Drobyshevsky et al, 2005; Miller et

al, 2002). Using this technology, DTI can be displayed by condensing the tensor

information into one number, referred to as a “scalar”. The tensor can also be represented

using glyphs, a small 3D representation of the major eigenvector or whole tensor, or

presented as 4 numbers in producing colour-coded FA map. The colour FA map is

- 32 -

produced with each colour representing the primary diffusion direction within each voxel

of the image (Figure 1.8) (Soares et al, 2013).

Figure 1.8 DTI colour map. DTI map of a preterm infant brain at 32 weeks PMA in the

axial plane at the level of the superior white matter (left) and basal ganglia (right). Colour

coding is reflected by red colour representing the left-to-right orientation within the image,

green the posterior-to-anterior and blue the inferior-to-superior axis of diffusion.

Another way of viewing the DTI image includes estimations of the course of the

white matter tracts through the brain, termed tractography. The most common approach,

streamline tractography, involves following the path of successive principle eigenvectors

from a given voxel origin, and is a form of “deterministic” tractography. “Probabilistic”

tractography, another widely utilised method, relies upon connection probabilities

between voxels and is more reliable at reconstructing crossing fibers than deterministic

- 33 -

tractography (O'Donnell and Westin, 2011). Despite significant promise with tractography,

both methods have their limitations and challenges with interpretation, in part as a result

of the complexity of the crossing fibres within the brain, but also as a result of the large

numbers of potential false positive or negative tracts that can result from the analysis.

Tract-based spatial statistics (TBSS; http://fsl.fmrib.ox.uk/fsl/fslwiki/TBSS), a semi-

automated and 3D assessment of FA, provides a less user-dependent measure of FA

than complementary ROI-based analyses. TBSS utilizes “voxel-based morphometry,” in

which each subject’s FA image is registered to a standard space, following which

voxelwise statistics are applied to find areas that correlate to the covariate of interest.

Through TBSS, FA data can be projected onto a group mean FA tract skeleton which can

be used to display the voxelwise cross-subject statistics as an assessment of white matter

integrity differences (Smith et al, 2006). The normalization of each image to a standard

space is a crucial step to the TBSS model, but can lead to misalignment and issues with

the normalization of the data. This is especially true for the preterm infant brain which is

undergoing rapid changes in size and shape within the normal fetal third trimester. The

FA images are aligned using voxelwise nonlinear registration, which is driven by the FA

images. Secondly, the mean of the aligned FA image is utilised to produce a skeletonised

mean FA image. Following that, each subjects’ aligned FA image is projected on the

skeleton filling the skeleton with FA values, on which cross-subject voxelwise statistics

can be performed. In producing voxelwise statistical models with a correction for multiple

comparisons, a Gaussian random field theory thresholding approach is often applied for

analysis (Smith et al, 2006). If the spatial width of this Gaussian filter is not chosen

appropriately then the signal-to-noise ratio will be reduced resulting in poorer power of

- 34 -

the model and more imperfections. Cluster-size thresholding, a permutation-based

approach to thresholding, can help to reduce this without the use of a Gaussian filter,

while still allowing the appropriate statistical test. By using a cluster size determined by

500 permutations of the cluster size using Randomise v.2.9 within the functional MRI of

the Brain Software Library (http://sel.fmrib.ox.ac.uk/fsl/fslwiki/Randomise) the familywise

error rate is lessened while maintaining the ability to search the entire FA skeleton for

regions of significant difference (Smith et al, 2006). In showing cross-subject voxelwise

significance, a threshold of P <0.05, equivalent to the 95th percentile of the distribution, is

used for the clusters, and corrected for multiple comparisons across space. Using these

techniques, a preterm infant model for assessing TBSS with the white matter skeleton

separated into four categories based upon GA at scan was developed by Duerden et al

(Duerden et al, 2015) (Figure 1.9).

Figure 1.9 Tract-based spatial statistic (TBSS) model. Age-specific models for preterm

infants are shown here. Green regions reflect the white and grey matter skeleton template

for the respective age group overlaid upon an axial T1-image.

- 35 -

1.6.5 Magnetic Resonance Spectroscopic Imaging

Proton magnetic resonance spectroscopic imaging (MRSI) uses the signals

produced by the different metabolite molecules within the voxel to produce spectra of

individual peaks. These peaks are based upon the concentrations of metabolites,

represented by parts per million (ppm) on the spectroscopic map. The principle behind

MRSI is that the distribution of electrons around hydrogen protons in the different

metabolites results in slight alterations to the magnetic field. Because these metabolites

have molecular concentrations 10,000 times lower than water, the signal produced is

much lower, thus larger voxels are required and the water signal has to be suppressed

(Posse et al, 2012). The most commonly used MRSI method involves using conventional

phase-encoded imaging in which the phase encoding gradient amplitude is incremented

once per TR. Phase encoding modulates the signal phase and amplitude of the MR signal

before detection of the signal frequency, which ensures that subsequent phase and

amplitude modification due to the chemical shift is independent of spatial encoding (Posse

et al, 2012). Suppression of the water signal is then achieved by using either a chemical

shift selective or inversion recovery technique, which when the Fourier transform is

applied, separates the signal into individual frequencies. The greater the magnetic field

strength, the greater the signal-to-noise ratio, which allows smaller voxels to be used in

producing the metabolite spectra. Another technique increasingly used is echo-planar

MRSI, which has strong localization performance and large volume coverage with

improved spatial and temporal resolution (Posse et al, 2012). This process utilizes a

zigzag trajectory in k-t-space creating a series of gradient echoes that are modulated by

- 36 -

the chemical shift evolution and relaxation. Through reformatting, the data format is

equivalent to conventional phase-encoded MRSI.

As with conventional MR, the TE is an important factor in the information obtained.

With a short TE of 30 msec, metabolites with both short and long T2 relaxation times are

observed (Brandao and Domiques, 2003). A longer TE of 270 msec results in only

metabolites with a long T2 to be seen. An intermediate length TE of 144 msec is often

used as it allows the imaging of lactate as an inverted doublet at 1.3 ppm. MRSI

metabolites of importance in the brain include N-acetyl-aspartate + N-

acetylaspartylglutamate (NAA) at 2.0 ppm, lactate (LAC) at 1.3 ppm,

glycerophosphocholine + phosphocholine (CHO) at 3.2 ppm, creatine + phosphocreatine

(CR) at 3.0 ppm and myo-Inositol (INS) at 3.5 ppm (Card et al, 2013) (Figure 1.10). A

long TE spectra at 270 msec produces a spectra of primarily NAA, CR and CHO, while

the other metabolites are best seen with a short TE with a high signal to noise ratio. Other

metabolites of interest include glutamate/glutamine and gamma-aminobutyric acid

(GABA) [all at 2.2 – 2.4 ppm], which have been shown to correlate with white matter injury

(Wisnowski et al, 2013). NAA is a neuronal marker, present in high concentrations in the

brain, and synthesized in neurons (Moffett et al, 2007). CHO is a measure of increased

cellular turnover or membrane breakdown, and is often elevated in tumours and

inflammatory disorders, while CR provides a measure of brain metabolism and energy

stores. INS participates in phospholipid metabolism and plays a role in cellular message

transduction and a proposed marker for gliosis. Unfortunately, many notable metabolites

are not represented in the MRSI spectra, with neurotransmitters acetylcholine, dopamine

- 37 -

and serotonin all absent, either as a result of low concentrations or molecules that don’t

respond to conventional MRSI techniques.

The metabolites of interest are commonly presented as ratios, which helps to

control for changes in brain water content, the T1 and T2 relaxation times of water and

possible CSF fluid contamination of the voxel in use (Card et al, 2013). Using absolute

metabolite concentrations is becoming more common, though its use in vivo is difficult

and requires intensive post-processing requirements. CHO is commonly used as a

denominator in metabolite ratios due to the fact that its concentration reduces over time

as the rapid brain growth of the neonate slows in infancy. During the preterm period

dramatic increases in CR/CHO and NAA/CHO ratios are seen with decreases in INS/CR

and INS/CHO ratios (Card et al, 2013). LAC/CR and LAC/CHO ratios have been shown

to correlate with the severity of asphyxia at birth and likely reflects lactate as a marker of

anaerobic metabolism seen in tissue infarction. White matter injury has previously been

shown to result in lower NAA/CHO and NAA/CR ratios (Card et al, 2013; Chau et al,

2009).

- 38 -

Figure 1.10 Magnetic Resonance Spectroscopic Imaging (MRSI) example. Example of a

preterm infant scanned at term with a 6mm voxel located in the left basal ganglia with a

long echo time (TE) of 144 ms on a 3T magnet. The peaks of Cr2 (phosphocreatine), Cho

(choline), Cr (creatine), and N-acetylaspartate (NAA) are shown at their respective parts

per million position on the spectroscopy (ppm) on the x-axis with the MR signal peak

amplitude on the y-axis.

- 39 -

1.7 Neurodevelopmental Outcomes

1.7.1 Background

Standardized developmental assessments are important in the early detection of

developmental delays both for the eligibility requirements of early intervention programs,

but also in the evaluation of therapies in the attempt to improve childhood developmental

outcomes (Johnson and Marlow, 2006). While no assessment battery is perfect and

several standardized assessments for infants are available, the Bayley Scales of Infant

and Toddler Development and the Peabody Developmental Motor scales, along with their

revisions, are validated and widely reported assessment methods (Anderson et al, 2010).

1.7.2 The Bayley Scales of Infant and Toddler Development

The Bayley Scales of Infant and Toddler Development, 3rd edition (BSID-III) are

meant to examine all facets of a young child’s development with norm-referenced,

standardised testing that can be done for infants 1 to 42 months of age (Bayley, 2005).

Initially developed in 1969 as the first edition, changes were made in the second edition

utilized from 1993 – 2006, with the 3rd edition used since 2006. Originally standardized to

1,700 infants, toddlers and pre-schoolers in the United States of America, the scoring

scale reflects an assessment of healthy individuals and does not include at-risk or

disabled populations (Bayley, 1993; Bayley, 2005). Qualified therapists, through the

administration of a battery of developmentally appropriate tests of child interaction for

infants, toddlers and pre-school age children, calculate an individual infant’s cognitive,

language and motor composite scores. Social-emotional and adaptive behaviour

assessments are also included and conducted via parent questionnaires. The scoring

- 40 -

scales for each composite have a standardized and normalized mean of 100 with

standard deviation of 15, with scores ≤85 considered below average and <70 considered

severely abnormal (Bayley, 2005). When assessing the scores across timelines and

previous literature it is important to note that there is considered to be a large difference

between the scales on the BSID-III and the previously used BSID, 2nd edition, particularly

at the lower score values (Moore et al, 2016).

Cognitive scoring is done by assessing the way the child thinks, reacts and learns

about the world around them through their interaction with familiar and unfamiliar objects.

The language scale includes two components, with the receptive communication (RC)

part assessing the ability for the child to recognize sounds and understand words and

directions. Expressive communication (EC) assesses how well the child communicates

with sounds, gestures or words. The motor scale also has two parts with fine motor (FM)

assessing how well the child uses his or her fingers to manipulate objects, while the gross

motor (GM) part looks at how well the child moves his or her body within their

environment.

1.7.3 Peabody Developmental Motor Scales

The Peabody Developmental Motor Scales, 2nd edition (PDMS-2), was first

published in 2000, and is a developmental assessment focused on the motor skills of

children from birth through to age 5 years (Folio and Fewell, 2000). The PDMS-2 was

validated on 2,003 children residing in 46 U.S. States and one Canadian province, and

matched to a normative sample of children <5 years with regard to geographic region,

sex, race, rural or urban residence, ethnicity, family income, parent education and

disability (Folio and Fewell, 2000). The testing contains six subtests which are divided

- 41 -

into the following categories: reflexes, stationary, locomotion, object manipulation,

grasping and visual-motor integration. All of the sub-tests allow the formation of a total

motor quotient, considered the best estimate of overall motor abilities. The composite

scores for each of gross motor and fine motor quotients are a combination of a number

of the subsets. The PDMS-2 composite scores have high correlation with the BSID-III

composite motor score, greatest in those infants >18 months of age (Connolly et al, 2012).

1.7.4 Cerebral Palsy

Cerebral palsy (CP) is the most common movement disorder in children, with an

estimated incidence of 2.1 per 1,000 live infants (Oskoui et al, 2013). Some of the first

descriptions of CP were made by Hippocrates in the 5th century BCE, with its first

description in modern medicine in the 19th century by William John Little, and first termed

“cerebral palsy” by William Osler (Panteliadis et al, 2013). CP is defined as “a group of

permanent disorders of the development of movement and posture, causing activity

limitation, that are attributed to non-progressive disturbances that occurred in the

developing fetal or infant brain” (Rosenbaum et al, 2006). Often accompanied with the

disorders of motor function are disturbances of sensation, perception, cognition,

communication and behaviour, epilepsy and secondary musculoskeletal problems. The

diagnosis of CP is made following an assessment by a pediatrician or neurologist familiar

with the development of children. While many factors contribute to the risk of CP, preterm

birth has a strong prevalence among children with CP. With many improvements in NICU

care over the past several decades, the incidence has decreased significantly in those

infants born >1000 grams (Sellier et al, 2016), though has remained stable in those infants

born <28 weeks or <1000 grams, with rates as high as 112 (95% CI, 70-180) per 1000

- 42 -

live births (Conde-Agudelo and Romero, 2009; Oskoui et al, 2013; Robertson et al, 2007;

van Haastert et al, 2011). As a measure of severe motor impairment, several NICU

practices have been shown to reduce the likelihood of CP in the preterm infant with

antenatal betamethasone (French et al, 2004), magnesium sulfate (Nelson and Grether,

1995; Rouse et al, 2008), postnatal caffeine therapy (Schmidt et al, 2007), increased

caesarian section rate and antenatal antibiotics (van Haastert et al, 2011) all showing a

reduced incidence. Despite these improvements in NICU care, CP remains a common

lifelong condition that results in significant impairments. Postnatal infection and severe

ROP are strong predictors of CP with double the risk in those infants with infection (Stoll

et al, 2004), and a three-fold increase of motor impairments seen in those infants with

severe ROP (Schmidt et al, 2014). Efforts to reduce the incidence of CP and its impact

upon the developing child need to be continued.

- 43 -

1.8 Hypothesis, Major Goal and Specific Aims

1.8.1 Hypothesis

I hypothesize that retinopathy of prematurity (ROP) and multiple postnatal

infections will be associated with delays in the brain maturation evident on multi-modal

MR imaging with associated poorer neurodevelopmental outcomes.

1.8.2 Major Goal

My goal is to identify whether retinopathy of prematurity and/or multiple infections

in the preterm infant are associated with poorer outcomes and/or delayed white matter

maturation. Through this investigation I hope to assist in altering the care of the preterm

infant to optimize developmental outcomes and provide insight into which infants are at

greatest risk of poor outcomes and would benefit most from early intervention practices.

These findings will expand the understanding of these factors in preterm infant

development following complications of preterm birth and contribute to the understanding

of these variables in the brain development of the preterm infant, while potentially

supporting further research in the field.

- 44 -

1.8.3 Specific Aims

1. To characterize the brain maturation and neurodevelopmental outcomes of

extremely preterm infants with severe ROP.

2. To describe the association of multiple postnatal infections with developmental

outcomes and brain maturation of very preterm infants.

- 45 -

Chapter 2

Severe Retinopathy of Prematurity predicts

delayed white matter maturation and poorer

neurodevelopment at 18 months CA

This chapter is modified from work published in the Archives of Disease in Childhood -

Fetal and Neonatal Edition: Glass TJA et al. Severe retinopathy of prematurity predicts

delayed white matter maturation and poorer neurodevelopment Archives of Disease in

Childhood - Fetal and Neonatal Edition Published Online First: 23 May

2017. doi:10.1136/archdischild-2016-312533

The work is published here with copyright permission from the BMJ Publishing Group

Ltd.

A link to the published paper can be found at:

http://fn.bmj.com/content/early/2017/05/22/archdischild-2016-312533

- 46 -

2.1 Introduction

Retinopathy of prematurity (ROP) is a proliferative disorder of retinal

vascularisation that is most severe in extremely preterm neonates born at less than 28

weeks gestational age (GA). Primary therapeutic strategies for ROP include prevention

of abnormal vessel development with tight control of supplemental oxygen therapy and,

when severe, treatment with either retinal laser photocoagulation therapy or, more

recently, intravitreal anti-vascular endothelial growth factor (VEGF) (Gunther and

Altaweel, 2009; Lee and Dammann, 2012). Several large cohort studies of preterm

neonates have shown that severe ROP is associated with lower cognitive and motor

scores in early childhood (Bassler et al, 2009; Harrell and Brandon, 2007; Hellstrom et al,

2013; Schmidt et al, 2003). Despite favourable visual outcomes, severe ROP continues

to strongly predict non-visual disabilities independent of brain injury (Schmidt et al, 2014).

Given this link between severe ROP and non-visual disabilities, we hypothesized

that severe ROP would be associated with white matter maturational delay on advanced

MRI. White matter maturation is well recognized as an important predictor of

neurodevelopmental outcomes in infants born preterm (Chau et al, 2013).

The objectives of this prospective cohort study of extremely preterm neonates

were to determine the association of severe ROP with: (1) early brain development as

measured by MR diffusion-tensor imaging (DTI) and tract-based spatial statistics (TBSS);

and (2) motor and cognitive outcomes at 18 months corrected age (CA). We tested the

hypothesis that severe ROP would be associated with abnormalities in early brain

- 47 -

microstructural development and with worse cognitive and motor development at 18

months CA follow-up.

2.2 Material and Methods

2.2.1 Participants

This study was approved by the University of British Columbia/Children’s and

Women’s Health Centre of British Columbia Research Ethics Board. As part of a larger

prospective study of neonates born 24 to 32 weeks GA, informed consent was obtained

from parents/guardians prior to recruitment of the neonates from April 2006 to September

2013 at British Columbia Women’s Hospital (BCWH), the major provincial tertiary-level

neonatal referral center. Neonates were excluded if they had (1) clinical evidence of a

congenital malformation or syndrome, (2) congenital infection, or (3) ultrasound evidence

of a large parenchymal hemorrhagic infarction (>2 cm) (Papile et al, 1978). This cohort

has been described previously addressing separate hypotheses (Adams et al, 2010;

Brummelte et al, 2012; Chau et al, 2009; Chau et al, 2013; Duerden et al, 2015). Only

extremely preterm neonates born ≤ 28 weeks GA were included in this sub-study as they

are the subset at greatest risk of severe ROP.

2.2.2 Clinical Characteristics

Neonates were screened by a pediatric ophthalmologist at BCWH as per the

International Classification of ROP and the maximal ROP severity in sequential

assessments was included in the analysis (International Committee for the Classification

of Retinopathy of Prematurity, 2005). Severe ROP was defined as ROP requiring

- 48 -

treatment as per the Early Treatment for Retinopathy of Prematurity (ETROP) study

(Early treatment of Retinopathy of prematurity cooperative group, 2003). Intravitreal anti-

VEGF treatment was not used at our institution at the time of the study. Clinical

characteristics were collected systematically by chart review, as previously described;

(Chau et al, 2012; Chau et al, 2009): with bronchopulmonary dysplasia (BPD) defined as

oxygen therapy beyond 36 weeks PMA, hypotension defined as any treatment for low

blood pressure, culture positive infection as any positive blood, urine, cerebrospinal fluid

or respiratory culture, and necrotising enterocolitis (NEC) defined as stages 2 and 3 of

Bell’s criteria (Bell et al, 1978).

2.2.3 MRI Studies

Neonates were scanned first when clinically stable in the preterm period and again

at term-equivalent age. MRI studies were completed without pharmacological sedation

on a Siemens 1.5 Tesla Avanto scanner with 3D coronal volumetric T1-weighted and axial

fast-spin echo T2-weighted images. Neonates were scanned using an MR-compatible

isolette (Lammers Medical Technology, Luebeck, Germany) and specialized neonatal

head coil (Advanced Imaging Research, Cleveland, OH). An experienced

neuroradiologist, blinded to the participant’s medical history, reviewed the images and

recorded the WMI, IVH, ventriculomegaly and cerebellar hemorrhage severity according

to scales previously described (Chau et al, 2009; Miller et al, 2005).

2.2.4 Diffusion Tensor Imaging

DTI reflects the water movement of an ellipsoid space with axial diffusion (λ1), the

preferred movement of water along the white matter tracts, and radial diffusion (λ2 and

- 49 -

λ3) reflecting the orthogonal planes. DTI characterizes the 3-dimensional (3D) spatial

distribution of water diffusion in each voxel of the MR image, providing an indirect

measure of microstructural integrity (Beaulieu, 2002; Miller et al, 2002; Mukherjee et al,

2002). Mean FA, the average directionality of diffusion increases with white matter

maturation, reflecting the maturation of the oligodendrocyte lineage and early events of

myelination (Drobyshevsky et al, 2005; Miller et al, 2002). Diffusion tensor imaging

parameters of FA, λ1, λ2 and λ3 were measured in seven manually placed white matter

regions of interest (ROI) and acquired using a multi-repetition single-shot echo planar

sequence (Chau et al, 2009). ROIs were excluded if significant motion artifact was

present. The ROI areas were acquired in the anatomical regions of the calcarine, anterior,

central and posterior white matter regions, optic radiations, splenium of the corpus

callosum and posterior limb of the internal capsule (PLIC). These regions have been

shown previously to correlate with motor, language and cognitive outcomes (Chau et al,

2009; Chau et al, 2013).

TBSS, a semi-automated and 3D DTI assessment of mean FA, provides a less

user-dependent measure of FA than complementary ROI-based analyses. TBSS was

performed using functional MRI of the Brain software library (FSL) and, after correction

for eddy currents, each DTI volume was registered to a non-DTI volume for each subject

(Smith et al, 2006). The TBSS FA data were projected onto a mean FA tract skeleton,

which was used to apply the voxelwise cross-subject statistics (Smith et al, 2006). We

used five age-based templates and were able to compare neonate groups with a

standardised analysis and a calculated voxel significance threshold of p <0.05 adjusted

for PMA at scan (Duerden et al, 2015). White matter abnormalities detected with TBSS

- 50 -

at term-equivalent age have been shown to predict neurodevelopmental outcomes in

preterm neonates, with increased FA at 2 years CA being associated with improved

outcomes (Counsell et al, 2002; Duerden et al, 2015).

2.2.5 Developmental Follow-up

At 18 months CA, neurodevelopment was assessed with the Bayley Scales of

Infant and Toddler Development-III (BSID-III) cognitive, language and motor composite

scores with a mean of 100 and standard deviation of 15 (Bayley, 2005). Assessors were

qualified therapists blinded to the imaging findings of the participants. Severe cerebral

palsy (CP) was defined as any diagnosis by an experienced clinician prior to or at the 18

month assessment of CP. Hearing impairment was diagnosed when audiograms showed

hearing threshold >70 dB. Visual acuity was assessed sequentially by the treating

ophthalmologist using varied assessment techniques depending upon age at follow-up

and collected with a retrospective chart review. Visual impairment was defined as best

visual acuity <20/200. Socioeconomic status was assessed as number of years of

maternal education.

2.2.6 Data Analysis

Statistical analysis was performed using Stata 14.1 (StataCorp, 2015). Clinical

characteristics were compared using Fisher’s exact and the Kruskall–Wallis tests for

categorical and continuous data, respectively, with a statistical significance of p <0.05.

The association of severe ROP and other clinical variables with WMI was tested with

univariate logistic regression. The mean values of FA, averaged bilaterally, were

compared between neonates with and without severe ROP, in a generalized least

- 51 -

squares regression model for repeated measures, adjusting for PMA at MRI scan and

multiple ROIs with a p<0.05. We examined the relationship of ROP with FA modified by

PMA at MRI scan, considering p<0.1 as significant due to the interaction term.

2.3 Results

2.3.1 Clinical Characteristics

Of the 234 neonates born 24-32 weeks GA, 126 were born at 24-28 weeks GA.

Ninety-eight (79%) extremely preterm neonates were assessed for ROP at BCWH and

were included in the analysis; infants not assessed were older at birth but had no

difference in other demographics. Neonates were born at a median GA of 26.0 weeks

(interquartile range [IQR], 25.0−26.9 weeks). Early MRI scans were completed in the 98

neonates assessed for ROP at median 32.4 weeks (IQR, 30.3-35.0 weeks) and term-

equivalent MRI was performed in 87 (89%) at median PMA 39.8 weeks (IQR, 38.3-41.6

weeks) (Figure 2.1).

- 52 -

Figure 2.1 Study flow chart. Flow chart of study enrollment, ROP treatment, MRI

scans and follow-up. Ages presented are median, with corrected age used for Bayley-III

Assessment.

2.3.2 Retinopathy of Prematurity

Of the 98 neonates in this cohort, 67 (68%) had any stage of ROP (stage 1 = 7,

stage 2 = 47, stage 3 = 13, stage 4 = 0, stage 5 = 0), and 19 (19%) neonates required

treatment for severe ROP as per criteria defined in the ETROP study at a median PMA

of 37.9 weeks (IQR, 36.1 – 39.3 weeks). Neonatal parameters associated with severe

ROP included GA at birth, birth weight, birth length, head circumference at birth, BPD,

Birth GA 24 – 28 weeks GA 26.0 weeks

N = 126

MRI #1 GA 32.4 weeks

N=98

Excluded as ROP assessment not done

N = 28

Laser Photocoagulation Treatment

GA 37.9 weeks N=19

MRI #2 GA 39.8 weeks

N=87

Bayley-III Assessment 18.6 months N=83 (85%)

Died N = 4 Withdrew = 7

Lost to follow-up = 4

- 53 -

NEC and hypotension (Table 2.1). Culture positive infection, histologic chorioamnionitis

and multiple gestation were not significantly different in those with and without severe

ROP.

Table 2.1 Demographic, clinical characteristics and imaging findings of neonates 24-28

weeks GA with and without severe ROP treated with retinal laser therapy. Number (%) or

median (Interquartile range)

No Severe ROP

n=79

Severe ROP

n=19

P-value

Male sex 38 (48%) 14 (74%) 0.07

Multiple gestation 5 (6%) 4 (21%) 0.07

Age at birth (weeks) 26.1 (25.3-27.1) 25.4 (24.7-25.9) <0.01

Weight at birth (grams) 845 (745-991) 700 (630-795) <0.01

Length at birth (cm) 34 (32-36) 32 (31-34) 0.02

Head Circumference at birth (cm)

24 (22.5-25) 22.5 (22-23) <0.01

Conventional ventilation (days) 16 (5-33) 51 (37-59) <0.01

Histologic chorioamnionitis 39 (51%) 9 (53%) >0.99

Hypotension 37/79 (47%) 15/19 (79%) 0.02

Bronchopulmonary dysplasia (BPD)

22/79 (28%) 11/19 (58%) 0.02

Necrotizing enterocolitis (NEC) 3/79 (4%) 4/19 (21%) 0.03

Culture positive infection 46/79 (58%) 15/19 (79%) 0.12

Intraventricular hemorrhage (IVH)

43/78 (55%) 10/19 (53%) >0.99

White matter injury (WMI) 24/78 (31%) 8/19 (42%) 0.42

Cerebellar hemorrhage 16/78 (21%) 2/19 (11%) 0.51

Severe WMI and/or IVH 12/78 (15%) 3/19 (16%) >0.99

- 54 -

2.3.3 Brain Injury

Ninety-eight early and 87 term-equivalent scans were scored. Findings included:

WMI in 30 (35%) neonates, severe WMI in 9 (11%), IVH in 47 (55%) neonates, severe

IVH in 5 (6%), ventriculomegaly in 27 (32%), and cerebellar haemorrhage in 16 (19%).

Severe ROP was not associated with an increased risk of WMI, IVH, ventriculomegaly or

cerebellar haemorrhage even when comparing the most severely affected forms of WMI,

IVH, and ventriculomegaly separately (all p>0.05) (Table 2.1).

2.3.4 White Matter Maturation

Mean FA in neonates with severe ROP was significantly lower for the posterior

white matter (effect size, -0.02; 95% confidence interval (CI), -0.04 to -0.004; p=0.02)

which was more pronounced over time on interaction analysis (p=0.08) (Figure 2.2). A

trend was seen in the FA of the optic radiations (effect size, -0.02; 95% CI, -0.39 to 0.001;

p=0.07) (Figure 2.2).

- 55 -

Figure 2.2 Mean Fractional Anisotropy (FA) values. Mean FA in those with and without

severe retinopathy of prematurity (ROP) by post menstrual age at scan in (A) posterior

white matter [p=0.02], (B) optic radiations [p=0.07], (C) splenium of the corpus callosum

[p=0.50], and (D) calcarine white matter [p=0.78]. (E) Diffusion tensor imaging region of

interest model map at the level of the high centrum semiovale and (F) basal ganglia:

(AWM) anterior white matter, (CWM) central white matter, (PWM) posterior white matter,

(PLIC) posterior limb of the internal capsule, (SCC) splenium of the corpus callosum, (OR)

optic radiations, (CWM) calcarine white matter.

C D

E F

- 56 -

Using voxelwise regression analysis in TBSS, white matter regions with

significantly lower FA in severe ROP included the optic radiations, PLIC and external

capsule on scans at 34-37 weeks (n=6/30) and 42+ weeks GA (n=3/24) (Figure 2.3).

Figure 2.3 Tract based spatial statistics (TBSS) model. TBSS using semi-automated

preterm and term neonate templates. White matter regions where neonates with severe

retinopathy of prematurity (ROP) differ from those with no severe ROP [p<0.05], corrected

for age at scan, shown in blue overlaid on the FA white matter skeleton in yellow. The

number of subjects in each sample with ROP is reflected by the numerator, and total

subjects in the sample, the denominator. R = right, L = left, P = posterior, A=anterior.

FA on ROI analysis in the anterior white matter (effect size, 0.001; p=0.93), central

white matter (effect size, -0.002; p=0.84), splenium of the corpus callosum (effect size, -

Axial Sagittal

L

L L

L R

R R

R A

A

P

P

- 57 -

0.010; p=0.50), PLIC (effect size, -0.01; p=0.19) and calcarine white matter (effect size, -

0.002; p=0.78) did not differ significantly in neonates with and without ROP. In the

posterior white matter and optic radiations, there was no difference in the relationship with

FA in neonates with severe ROP in the radial axes (λ2 and λ3) and the axial diffusion axis

(λ1).

2.3.5 Developmental Outcomes

A total of 83 (85%) infants were assessed at 18 months CA follow-up (Table 2);

there were no significant differences in the 7 who withdrew from the study after at least

one MRI scan or the 8 who died or were lost to follow-up.

Table 2.2 18 month corrected age (CA) follow-up assessments of neonates 24-28 weeks

GA with and without severe ROP treated with retinal laser therapy. Number (%) or median

(Interquartile range)

No Severe ROP

n=67

Severe ROP

n=16

P-value

BSID-III Cognitive score 105 (100-115) 95 (88-105) 0.02

BSID-III Motor score 94 (85-110) 85 (75-93) 0.01

BSID-III Language score 96 (83-112) 89 (83-103) 0.32

Severe cerebral palsy 6/67 (9%) 1/16 (6%) >0.99

Visual impairment 1/67 (1%) 0/16 (0%) >0.99

Hearing impairment 2/67 (3%) 0/16 (0%) >0.99

Infants with severe ROP had lower BSID-III motor and cognitive scores, with no

difference in language scoring relative to infants without severe ROP. There was no

difference in years of maternal education between groups (p=0.91). In a multivariate

- 58 -

model, the relationship between severe ROP and motor and cognitive scores remained

significant when adjusting for GA at birth and severe WMI and/or IVH (Table 2.3).

Table 2.3 Multivariate linear regression analysis. 18-month corrected age Bayley

Composite scoring adjusted for GA at birth and severe white matter injury and/or

intraventricular hemorrhage.

Effect size (95% CI) P-value

BSID-III Cognitive Composite -9.22 (-17.69 - -0.75) 0.03

BSID-III Motor Composite -10.77 (-19.87 - -1.66) 0.02

BSID-III Language Composite -3.45 (-13.80 – 6.89) 0.51

One infant in the “no severe ROP” group was unable to complete the BSID-III

assessment due to severe CP and cortical visual impairment; when assigned a value of

2 standard deviations below the mean, there was no meaningful difference in the findings.

There was no difference in the rate of cerebral palsy, hearing or visual impairment in

infants with and without severe ROP (Table 2.2). Visual acuity assessments were

available in 12 (63%) newborns with severe ROP at a median age of 3.9 years (IQR, 3.1-

5.0 years). Median visual acuity was 20/30 OD and 20/40 OS in infants with ROP; no

infants met guidelines for severe visual impairment.

2.4 Discussion

In a prospective longitudinal cohort of extremely premature neonates, we found a

significant maturational delay of the posterior white matter regions with delay on TBSS

- 59 -

analysis in the optic radiations, PLIC and external capsule in those neonates with severe

ROP, which to our knowledge has not been previously reported. These novel findings

occurred independently of severe white matter injury or intraventricular hemorrhage and

likely indicate similar vital mechanisms in the development of severe ROP with poor

maturation of the white matter. It is well known that circulating levels of insulin-growth

factor 1 (IGF-1) are important modifiers in the circulatory levels of VEGF and the

development of severe ROP (Harrell and Brandon, 2007). Within the brain, IGF-1 has

been shown to have a significant role in the augmentation and utilisation of glucose across

all neural cell lineages (Cheng et al, 2000; Dercole and Ye, 2008). Furthermore, mean

IGF-1 concentrations between birth and 35 weeks PMA in preterm neonates correlate

with total brain, white matter, grey matter, and cerebellar brain volumes, with poorer

cognitive development seen in those infants with the slowest rate of increase (Hansen-

Pupp et al, 2013; Hansen-Pupp et al, 2011). These factors support severe ROP as a

marker of adverse brain development, independent of visual outcomes, and are

highlighted by our 18 month CA follow-up assessments showing lower cognitive and

motor developmental scores in those with severe ROP, consistent with larger cohorts

(Bassler et al, 2009; Schmidt et al, 2003; Schmidt et al, 2015). Our study emphasizes

how adverse motor and cognitive developmental outcomes are associated with

maturational delays on DTI and TBSS independent of the traditional markers of brain

injury in the preterm neonate.

Maturational delays in the optic radiations are consistent with previous studies

using DTI and TBSS in preterm neonates that showed a relationship between severe

ROP and delayed white matter microstructural development in the optic radiations at

- 60 -

term-equivalent age, independent of white matter injury (Bassi et al, 2008; Thompson et

al, 2014). Further research at 7 years of age showed delayed optic radiation

microstructure in children with a history of severe ROP compared to those with milder

ROP (Thompson et al, 2014). We have shown that these delays are present in the late

preterm period and involve non-visual pathways, highlighting an early link of severe ROP

with poor white matter maturation in preterm neonates.

When considering the importance of a retinal disorder on the white matter

development of the preterm neonate, retinal nerve fibre layer (RNFL) thickness using

optical coherence tomography (OCT) in preterm neonates can be examined. In a study

comparing RNFL thickness with 18-24 CA developmental scores, thinner RNFL thickness

across the papillomacular bundle correlated with lower cognitive and motor scores

(Rothman et al, 2015). These findings provide an understanding of the associations

between the retina and the underlying white matter pathways, and offer a “window into

the brain” of the connections that the retinal ganglion and photoreceptor cells share with

the white matter pathways. The maturational delays in the motor, visual and visual-

association pathways of the developing brain detected in neonates with severe ROP

suggests a mechanism by which these factors act on neurodevelopment, potentially due

to an imbalance of growth factors.

2.4.1 Limitations

Although we were able to compare a large number of extremely preterm neonates

with serial MRI scans, the timing of the follow-up MRI scans may have affected the

imaging results due to the development of ROP over time as a progressively acquired

disorder. We were unable to control for the timing of the ROP retinal laser therapy and its

- 61 -

possible confounding of the imaging. Visual acuity follow-up was not done in conjunction

with the 18-month CA neurodevelopmental assessments and was not standardised

across the cohort.

2.5 Conclusions

Severe ROP requiring laser photocoagulation therapy is associated with delayed

maturation in the motor, visual and visual-association pathways. Infants with severe ROP

had lower motor and cognitive functioning at 18 months CA, independent of severe brain

injury and GA at birth. More research is needed to determine which potential mechanisms

in severe ROP prevent optimal neurodevelopment, and the impacts on brain maturation

of early and effective treatment of severe ROP.

- 62 -

Chapter 3

Multiple Postnatal Infections in Preterm

Newborns is associated with delayed Maturation

of Motor Pathways at Term-equivalent age and

Poorer Motor Outcomes at 3 years

- 63 -

3.1 Introduction

Preterm newborns born less than 32 weeks gestational age (GA) are at a

substantial risk of postnatal infection with 20 – 65% of newborns suffering from at least a

single infection during this period of significant brain development (Adams-Chapman and

Stoll, 2006; Orsi et al, 2009; Stoll et al, 2004). In a large epidemiologic study, postnatal

infection was found to double the risk of motor impairment and cerebral palsy, while also

greatly increasing the risk of cognitive impairment (Stoll et al, 2004). Similarly, a study of

very preterm infants followed to 9 years of age showed that infants who had postnatal

infection were more likely to have poorer motor development, cognitive delays, school

delays, attention-deficit hyperactivity disorder and other mental health disorders (Rand et

al, 2016). While white matter injury (WMI) is a known complication of postnatal infection,

the majority of infants with postnatal infection do not have punctate WMI on clinical

neuroimaging (Chau et al, 2012; Chau et al, 2009; Glass et al, 2008; Stoll et al, 2004).

Experimental models of WMI with hypoxia-ischemia show inflammation to have additive

effects on the injury present (Eklind et al, 2001). Microscopic WMI, present on

neuropathological studies, has been shown to be more widespread in the brain of infants

with macroscopic WMI, suggesting a more diffuse brain injury may be present, but below

the resolution of current clinical MRI techniques (Buser et al, 2012). Chau et al. reported

that very preterm newborns exposed to infection had reduced measures of white matter

development on MRI, even when adjusting for WMI, which was most prominent in brain

regions important for motor and cognitive development (Chau et al, 2012). This

impairment in the development of the white matter may reflect injury to the pre-OL lineage

cells following hypoxic-ischemic and/or inflammatory events in which the premature

- 64 -

infants’ immune system, and brain, may be primed for further events, increasing their

vulnerability to injury (Hagberg and Mallard, 2005; Wang et al, 2012; Yanni et al, 2017).

Our aim was to investigate the association of multiple postnatal infections with the

white matter development and outcomes of very preterm newborns with the hypothesis

that a greater number of infections would be associated with delayed white matter

development and poorer motor outcomes at 36 months corrected age (CA) compared to

non-infected neonates.

3.2 Material and Methods

3.2.1 Study Population

The study was approved by the University of British Columbia/Children’s and

Women’s Health Centre of British Columbia Research Ethics Board and informed consent

was obtained from parents/guardians prior to recruitment. Preterm neonates 24-32 weeks

gestational age (GA) were recruited into a prospective longitudinal cohort study at British

Columbia Women’s Hospital from April 2006 to September 2013. Infants were excluded

if they had clinical evidence of a congenital malformation or syndrome, congenital

infection, or ultrasound evidence of a large parenchymal hemorrhagic infarction (>2cm),

as these conditions are strongly predictive of neurodevelopmental impairments or early

mortality. This cohort has been described previously to address different hypotheses

(Adams et al, 2010; Brummelte et al, 2012; Chau et al, 2012; Chau et al, 2013; Duerden

et al, 2015; Glass et al, 2017).

- 65 -

3.2.2 Clinical Characteristics

Infection characteristics were collected by systematic chart review. Cultures that

had multiple growths with organisms consistent with contamination were excluded.

Culture positive infection was defined as any positive blood, urine, tracheal aspirate

and/or cerebrospinal fluid culture treated with ≥5 days of antibiotics with “clinical-only

infection” defined as any instance in which there was a clinical concern for infection with

negative cultures in which the antibiotic treatment duration was ≥5 days. A positive culture

with the same organism in each of two separate locations or cultures during a continued

antibiotic course was considered a single infection. Tracheal aspirates required a positive

culture and ≥4 white blood cells per field to be considered a positive culture. An “early

infection” was defined as any infection occurring at <72 hours of postnatal age, and

“postnatal infection” as any infection ≥72 hours after birth, with infections included up to

40 weeks post menstrual age (PMA). Other clinical characteristics were collected via

chart review: histologic chorioamnionitis as confirmed by clinical pathology assessment,

hypotension as any treatment for low blood pressure, patent ductus arteriosus (PDA) as

any requiring pharmacological or surgical treatment, BPD as oxygen therapy beyond 36

weeks PMA, and NEC as stages 2 and 3 of Bell’s criteria (Bell et al, 1978).

3.2.3 MR Brain Imaging

MRI scans were performed early in life when neonates were clinically stable, and

again at term-equivalent age, all without pharmacological sedation. At both time points

MRIs were carried out on a Siemens 1.5 Tesla Avanto scanner using an MR-compatible

isolette (Lammers Medical Technology, Luebeck, Germany) and specialized neonatal

- 66 -

head coil (Advanced Imaging Research, Cleveland, OH). 3D coronal volumetric T1-

weighted and axial fast-spin echo T2-weighted images were performed. An experienced

neuroradiologist (K.J.P.), blinded to the participant’s medical history, reviewed the images

and recorded the severity of WMI, IVH, ventriculomegaly and cerebellar hemorrhage

according to previously described scales (Chau et al, 2009; Miller et al, 2005). The most

severe injury score seen on the preterm or term scan was used in the structural MRI

analysis as the greatest severity of injury and most likely to adversely impact

developmental outcomes and to be associated with greater amounts of microscopic WMI,

not seen on MRI. WMI volumes were calculated on the T1-weighted images with voxels

of abnormal T1 shortening identified as WMI, reviewed by two neonatal neurologists (V.C.

and S.P.M.), then manually segmented with simultaneous coronal, sagittal and axial

views of the brain using Display software (http://www.bic.mni.mcgill.ca/software/Display)

as previously reported (Guo et al, 2017).

3.2.4 Magnetic Resonance Spectroscopic Imaging

MR spectroscopic imaging (MRSI) was used as a measure of neuronal maturation

using quantitative metabolite ratios with bilaterally placed 4mm ROI voxels in six

anatomical regions: the anterior, central and posterior white matter, the caudate, lentiform

nucleus (globus pallidus and putamen), and thalamus. To reflect the overall metabolism

of the regions, we analyzed an average of the three white matter regions, and the three

basal ganglia regions, the caudate, lentiform nucleus, and thalamus regions.

3.2.5 Diffusion Tensor Imaging

- 67 -

DTI is a measure of water diffusion in an ellipsoid space within each 3D voxel of

the MR image. Mean FA, the average directionality of diffusion, increases with white

matter maturation reflecting the maturation of the oligodendrocyte lineage and early

myelination (Drobyshevsky et al, 2005; Miller et al, 2002). DTI parameters of FA, λ1, λ2

and λ3 were acquired using a multi-repetition single-shot echo planar sequence, and

excluded if significant motion artifact was present (Chau et al, 2009). ROI analyses were

manually placed in seven white matter anatomical regions (anterior, central and posterior

white matter regions, optic radiations, splenium of the corpus callosum, genu of the

corpus callosum and PLIC) and three deep grey matter regions (caudate, lentiform

nucleus and thalamus).

TBSS was performed using functional MRI of the brain software library (FSL;

https://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FSL) as previously described (Duerden et al, 2015;

Glass et al, 2017). A diffusion tensor model was fit to the data at each voxel to calculate

the voxelwise FA with the spatial location of alterations in diffusion measures done using

the TBSS pipeline (Smith et al, 2006). The TBSS data were projected onto a mean FA

tract skeleton using age appropriate templates (preterm scans: 27-29 weeks, 30-33

weeks; 34-36 weeks; term scans: 37-41 weeks, ≥42 weeks), which were then used to

apply voxelwise regression cross-subject analysis. Cluster-size thresholding was applied

to the data in which the size of the cluster was determined by 500 permutations by using

Randomise v.2.9 within FSL. A threshold of P<0.05 (95% percentile of the distribution)

was set for the clusters and corrected for multiple comparisons across space in

determining the association of FA and ≥3 infections across subjects within each of the

age-based templates and projected upon the white matter skeleton.

- 68 -

3.2.6 Developmental Follow-up

At 36 months CA, neurodevelopment was assessed with the Bayley Scales of

Infant and Toddler Development-III (BSID-III) cognitive, language and motor composite

scores and Peabody Developmental Motor scales 2nd edition (PDMS-2) all with a mean

of 100 and standard deviation of 15 (Bayley, 2005; Folio and Fewell, 2000). In addition to

the BSID-III motor composite, the PDMS-2 total, gross and fine motor quotient values

were used in the analyses as a more robust assessment of motor impairment at 36

months CA (Folio and Fewell, 2000). Assessments were carried out by qualified

therapists blinded to the imaging findings of the participants. Socioeconomic status was

classified by the self-reported number of years of maternal education. Cerebral palsy was

defined as a confirmed diagnosis made by an experienced pediatrician at the 36 month

assessment.

3.2.7 Data Analysis

Statistical analysis was performed using Stata V.14.2 (StataCorp, 2015). Clinical

and imaging characteristics were compared using Fisher’s exact test for categorical and

the Kruskall–Wallis test for continuous data with a statistical significance threshold of

P<0.05. The association between postnatal infections and other clinical variables with

neurodevelopmental outcomes was tested with univariate logistic regression.

The mean values, averaged bilaterally, of FA and NAA/Choline were compared

between neonates with and without postnatal infection, in a generalized least squares

- 69 -

regression model for repeated measures, adjusting for PMA at MRI scan and multiple

ROIs with a P<0.05. A log-transformed outcome variable for the NAA/choline was used

to determine the percentage differences of the MRS measures (Chau et al, 2012). An

interaction term was examined describing the imaging relationships of postnatal infection

modified by PMA at MRI scan, and considered P<0.1 as significant. The 36 month

outcomes were compared using univariate analysis for multiple infections. An unadjusted

risk ratio and adjusted odds ratio were calculated in assessing the 36 month outcomes

association with the groups of infections.

3.3 Results

3.3.1 Clinical Characteristics

Of the 234 neonates born 24-32 weeks gestation recruited in the cohort, 219 (94%)

completed at least one MRI with a median birth GA of 27.9 weeks (interquartile range

[IQR], 26.0 – 29.7 weeks). Early MRI scans were completed at a median 32.1 weeks (IQR

30.4 – 34 weeks) with term-equivalent scans in 184 (84%) at median 40.2 weeks (IQR

38.7 – 42.0 weeks) (Figure 3.1).

- 70 -

Figure 3.1 Flow chart. Flow chart of study enrollment, MRI scans and follow-up.

Ages presented are median, with corrected age used for Bayley-III and PDMS-2

Assessments.

3.3.2 Infection Characteristics

Of the 219 infants, 110 (50%) had one or more instances of postnatal infection and

109 (50%) had no infections in the postnatal period. Due to the small number of the

“clinical-only infection” group we grouped them with “culture-positive infection” for the

analysis and thus classified each infection event as a “postnatal infection”. There were 54

(25%) infants with one postnatal infection, 29 (13%) had two infections, 19 (9%) had three

infections, 7 (3%) had four infections and 1 (0.5%) had five infections. Infants were

Birth GA 24 – 32 weeks GA 27.9 weeks

N = 234

MRI #1 GA 32.1 weeks

N=219

Excluded as MRI not done

N = 15

MRI #2 GA 40.2 weeks

N=184

Bayley-III and PDMS-2 Assessments

35 months N=175 (82%)

Died N = 4 Withdrew/second scan

not done N= 31

Died N=1 Lost to follow-up N= 8

- 71 -

grouped according to the number of infections, with three or more infections included

together in the analysis due to the small numbers in the four and five infections groups.

One and two infections were presented together in the results section as there were few

differences between the groups on clinical characteristics, neuroimaging and motor

outcomes (Table 3.1).

There were 46 (21%) infants with early infection of whom 21 had no other infection,

and 2 who had positive cultures (1 sepsis, 1 lower respiratory culture), which were

included in the septicemia and lower respiratory analyses respectively. In total there were

202 distinct episodes of postnatal infection among the total of 110 infants. There was a

greater rate of septicemia and lower respiratory infections in the three or more infections

group as well as a greater rate of infection with coagulase-negative Staphylococcus

(CONS) and Enterobacter species organisms compared to the one or two infection group

(Table 3.2). Neonatal characteristics more common among those with higher numbers of

infections included lower GA at birth, lower birth weight, lower birth length, smaller head

circumference at birth, hypotension, PDA, BPD and NEC stage ≥2 (Table 3.1).

- 72 -

Table 3.1 Demographics table. Demographics, clinical characteristics and MRI findings

of newborns 24-32 weeks GA classified by the number of postnatal infections. Number

[%] or median (IQR). Early infection = any infection <72 hours of life; NEC = necrotizing

enterocolitis; IVH = intra-ventricular hemorrhage; WMI = white matter injury

No Infection N=109

One Infection N=54

Two Infections N=29

Three or more Infections

N=27

P-value

Male sex 54 [50] 30 [56] 10 [35] 19 [70] 0.12

Gestational age at birth (weeks)

29.4 (27.7-31.1)

27.3 (25.9-28.6)

25.6 (25.0-27.1)

25.7 (24.9-26.6)

<0.01

Birth weight (grams) 1190 (1020-1376)

909 (789-1190) 835 (663-950) 755 (630-896) <0.01

Birth length (cm) 38 (35-40) 35 (34-39) 33 (31-35) 33 (31-35) <0.01

Birth head circumference (cm)

27 (26-28) 25 (23-27) 24 (22-25) 24 (22-25) <0.01

Twin birth 41 [38] 19 [35] 7 [24] 9 [33] 0.69

Antenatal MgSO4 24 [22] 13 [24] 5 [17] 5 [19] 0.97

Antenatal steroids 11 [10] 7 [13] 2 [7] 4 [15] 0.75

Histologic chorioamnionitis

33 [31] 22 [42] 14 [48] 8 [31] 0.18

Early infection 21 [19] 11 [20] 7 [24] 7 [26] 0.69

Hypotension 21 [19] 23 [43] 14 [48] 23 [85] <0.01

Patent Ductus Arteriosus

35 [32] 27 [50] 21 [72] 25 [93] <0.01

Bronchopulmonary dysplasia

5 [5] 8 [15] 9 [31] 17 [63] <0.01

NEC stage ≥2 1 [1] 1 [2] 2 [7] 4 [15] <0.01

IVH grades 2-4 34 [32] 12 [34] 13 [54] 9 [35] 0.67

WMI 38 [35] 14 [26] 8 [28] 8 [31] 0.49

WMI volume (mm) 35.8 (16.3 – 272.4)

41 (11 – 275) 53 (5 – 892) 17.9 (7.6 – 83.6)

0.60

Cerebellar hemorrhage 5 [5] 6 [17] 7 [29] 7 [27] <0.01

- 73 -

Table 3.2 Infection characteristics. Infection location and organism characteristics divided

by the number of infection groups. Number [%]

3.3.3 MR Imaging

WMI was present on the first MRI in 31% of infants (33 mild, 23 moderate, 12

severe) with 6 new instances of WMI on the second scans (4 mild, 1 moderate, 1 severe).

At least one infection occurred before the first MRI in 99 infants (90%). There was no

increase in WMI seen, either mild or moderate/severe, with increased numbers of

infection, with no increase in WMI when a separate analysis was done for the different

types of infection and organisms (all P>0.05). Similarly, IVH severity and

ventriculomegaly were not associated with infection (all P>0.05) (Table 3.1). Cerebellar

hemorrhage was more common among those infants with postnatal infection (Table 3.1).

One Infection

N=54

Two Infections

N=29

Three or more

Infections N=27

P-value

Clinical-only infection 7 [13] 5 [17] 8 [30] 0.21

Septicemia 16 [30] 17 [59] 17 [63] <0.01

Urinary tract infection 24 [44] 17 [59] 15 [56] 0.41

Lower respiratory infection 6 [11] 10 [34] 16 [59] <0.01

Meningitis 1 [2] 1 [3] 0 [0] >0.99

Coagulase-negative staphylococcus (CONS)

23 [43] 17 [59] 21 [78] 0.01

Enterococcus species 8 [15] 7 [24] 7 [26] 0.41

Escherichia coli (E.coli) 5 [9] 2 [7] 6 [22] 0.17

Klebsiella species 4 [7] 5 [17] 6 [22] 0.13

Enterobacter species 1 [2] 5 [17] 9 [33] <0.01

Candida species 1 [2] 2 [7] 4 [15] 0.08

Other pathogens 7 [13] 11 [38] 9 [33] 0.02

- 74 -

3.3.4 MRSI and DTI Imaging

The NAA/Cho ratio was lower over time in those infants with ≥3 infections in the

white matter (coefficient, -1.4%/week; 95% CI, -2.3% to -0.5%; P<0.01) (Figure 3.2) and

basal ganglia (coefficient, -0.7%/week; 95% CI, -2.5% to -0.1%; P=0.03) (Figure 3.2),

after adjustment for WMI.

- 75 -

Figure 3.2 Graphs of N-acetylaspartate/Choline (NAA/Choline) ratio. Graphs depict

associations of multiple infections with the neurulation of the basal ganglia (top) and white

matter (bottom) included in both preterm and term scans over time. Dots reflect the scatter

plot of the 3 regions included in each analysis and each subject, with the lines reflecting

the linear fit line to the data.

Those with ≥3 infections had lower mean FA over time in the PLIC (coefficient, -

0.005; 95% CI, -0.002 to -0.008, P<0.01). There were no other regions with significant

differences in the FA over time on the ROI analyses (all P≥0.10). The TBSS analysis of

the preterm MRIs 30-34 weeks PMA revealed only delayed FA located within the posterior

corpus callosum. In contrast, on the term scans at 37-42 weeks PMA lower FA was more

widespread and involved the complete corpus callosum, the optic radiations and PLIC

(Figure 3.3).

- 76 -

Figure 3.3 TBSS model for multiple infections. Shown in the pre-term (30-34 weeks GA)

and term (37-42 weeks GA) model for infants with ≥3 infections compared to no infection

group on a white matter skeleton map. The number of infants with ≥3 infections are

indicated by the ‘N=’ with the denominator reflecting the total number of infants. R = right,

L = left.

3.3.5 Developmental Outcomes

Thirty-six month corrected age outcome scores were available in 175 (82% of

survivors) infants (median 35 months, IQR 34 – 37 months). There were no differences

in the rates of WMI, IVH grade ≥3 or neonatal clinical factors (all P>0.05) in the follow-up

and missed follow-up groups. Three infants were unable to complete the testing due to

severe impairment thus were assigned scores of 49 (3.3 standard deviations from the

R R L L

- 77 -

mean). On univariate analysis a greater number of infections in the neonatal period was

significantly associated with poorer BSID-III motor composite score and PDMS-2 total,

gross and fine motor scores, but not language and cognitive scores (Figure 3.4)(Table

3.3).

Table 3.3 BSID-III and PDMS-2 outcomes at 36 months CA. Divided by number of

infection groups. Median (IQR) or Number [%]

No Infection N=93

One Infection N=42

Two Infections N=24

Three or more Infections

N=22

P-value

BSID-III Motor composite score

103 (97 – 115) 103 (94 – 110) 100 (91 – 107) 93 (76 – 107) 0.02

BSID-III Language composite score

109 (103 – 118)

112 (103 -115) 106 (83 – 115) 100 (94 – 112) 0.18

BSID-III Cognitive composite score

100 (95 – 110) 105 (95 – 105) 100 (95 – 105) 100 (90 – 110) 0.57

PDMS-2 Total motor 96 (90 – 102) 94 (88 – 97) 93 (88 – 97) 87 (74 – 94) <0.01

PDMS-2 Gross motor

96 (89 – 102) 94 (87 – 98) 91 (85 – 96) 84 (72 – 94) <0.01

PDMS-2 Fine motor 100 (94 – 103) 94 (91 – 97) 97 (88 – 103) 93 (85 – 97) <0.01

Cerebral Palsy 5 [5] 3 [6] 3 [12] 1 [4] 0.70

- 78 -

Figure 3.4 BSID-III outcomes at 36 months CA. Divided by infection groups with motor,

language and cognitive composite scores reflected by the shades indicated in the

legend. The middle line and box reflect the median and IQR respectively, with the

1.5 IQR reflected by the whiskers and open circles as any outliers. ** p<0.05 (all

others p >0.05)

Unadjusted risk ratio (RR) and adjusted odds ratio (OR) assessments of BSID-III

scores ≤85 and PDMS-2 ≤80, considered clinically significant impairments, are shown

(Table 3.4). These illustrate that a greater number of infections was significantly

associated with lower Bayley-III motor composite scores and poorer performance on the

PDMS-2 total and gross motor testing when adjusted for potential confounding factors,

with less impairment in fine motor scores (Table 3.4). There was no significant increase

- 79 -

in the incidence of cerebral palsy at 36 months CA with multiple infections (P>0.05) (Table

3.3)

Table 3.4 RR and ORs. Unadjusted RR and adjusted OR values for poor BSID-III and

PDMS-2 scores in infants with three or more postnatal infections. Odds (95% CI)

3.4 Discussion

Using multimodal MR imaging methods and follow-up assessments in a cohort of

very preterm newborns our results showed that three or more postnatal infections are

associated with delays in brain maturation, particularly within subcortical nuclei and white

matter implicated in motor function, and with poorer motor outcomes at 36 months CA.

Findings were associated with alterations in brain maturation over time, with differences

most prevalent on the term-equivalent MRI scans reflected by involvement of the PLIC,

corpus callosum and optic radiations. This is in agreement with previous research

suggesting that poorer motor outcomes and cerebral palsy are more likely in infants with

reduced FA in the PLIC, corpus callosum and white matter in cohorts of preterm infants,

implicating these important regions in motor development and highlighting the utility of

Unadjusted RR

P-value

Adjusted OR* P-value

BSID-III Motor composite <85 2.53 (1.10 – 5.85) 0.02 1.99 (1.2 – 3.2) <0.01

BSID-III Language composite <85 5.86 (1.34 – 25.6) <0.01 1.68 (0.94 – 3.00) 0.08

BSID-III Cognitive composite <85 0.55 (0.14 – 2.11) 0.37 0.55 (0.16 – 1.83) 0.33

PDMS-2 Total Motor <80 3.06 (1.27 – 7.4) <0.01 1.93 (1.21 – 3.09) <0.01

PDMS-2 Gross Motor <80 2.07 (1.06 – 4.03) 0.03 1.84 (1.18 – 2.85) <0.01

PDMS-2 Fine Motor <80 1.69 (0.58 – 4.97) 0.33 1.59 (0.88 – 2.90) 0.13

* Adjusted for Gestational age, maternal education, cerebellar hemorrhage and WMI volume

- 80 -

MRI in predicting motor outcomes (Duerden et al, 2015; Huppi et al, 2001; Thomas et al,

2005).

Very preterm newborns are at-risk of multiple postnatal infections due to many

factors. The innate and adaptive immune systems that respond to infectious antigens are

significantly impaired in the preterm newborn, leading the infant to bemore prone to

infections (Cuenca et al, 2013; Lavoie et al, 2010). Preterm infants also have reduced

trans-placental maternal antibody transfer and immature protein recognition receptors

important for bacterial antigen presentation and antimicrobial immune recognition (Kan et

al, 2016; Wynn et al, 2008). The consequence of this is an inappropriate inflammatory

response to infections with activation of toll-like receptors in the innate immune system,

that when associated with hypoxia-ischemia, can result in neural injury (Lehnardt et al,

2003). This may also may lead to priming of the immune system for subsequent events

manifested by greater concentrations of inflammatory markers and free radicals in the

cerebrospinal fluid (CSF), as seen in infants with WMI (Ellison et al, 2005; Inder et al,

2002). Neuropathology analysis of infants with severe WMI has showed that tumour

necrosis factor α (TNF-α) and interleukin-1β (IL-1β) levels are elevated in those infants

with infection, with greater levels than what is seen with hypoxia-ischemia alone, further

supporting their combined effects (Kadhim et al, 2001).

In a large epidemiologic study of over 6000 infants born at extremely low birth

weight (less than 1kg), all types of infection, not just sepsis or meningitis, increased the

risk of severe cognitive and motor impairments (Stoll et al, 2004). We also found no

major difference in the outcomes between the different types or organisms of infection.

We did not observe an increased risk for severe impairments with increasing numbers

- 81 -

of infections, with a low rate of cerebral palsy overall. Rand et al. showed that long-term

neurodevelopmental outcomes at 9 years were more likely to be delayed in infants with

postnatal infection, but with no difference in severe motor outcomes (Rand et al, 2016).

Furthermore, in a small subset of infants, Rand et al. also reported that two or more

infections increased the risk for motor impairments two-fold (RR 5.7 vs 2.6) in addition

to increasing the risk of cognitive impairment (RR 2.1 vs 1.3) (Rand et al, 2016). This is

supported by the literature on the additive effects of subsequent events, as described by

Khwaja and Volpe, in which the preterm infant brain is sensitized following an initial

event of infection or hypoxia-ischemia and has at a greater likelihood to be injured

following subsequent events of inflammation and/or hypoxia-ischemia by a stimulus with

a lower threshold than was necessary for the initial response (Eklind et al, 2001;

Hagberg et al, 2012; Khwaja and Volpe, 2008; Leviton et al, 2013; Yanni et al, 2017).

The brain of the preterm newborn also has an immature blood brain barrier which

makes it especially vulnerable to hypoxia-ischemia injury, particularly of the pre-OL, the

prominent cell injured in WMI of the preterm infant (Back et al, 2001; Back et al, 2002;

Gilles et al, 1976; Wang et al, 2012). In hypoxia-ischemia models of preterm brain injury

a significant increase in pro-inflammatory markers is seen, suggesting that even in the

absence of infection, a significant immune system response occurs (Albertsson et al,

2014; Fleiss et al, 2015). Of the infants in the three or more infection group, 78% had a

CONS infection, which is not known to cause direct cerebral inflammation, but

frequently occurs in conjunction with hypotension (Mallard and Wang, 2012). In our

group of infants with three or more infections 85% had at least a single event of

hypotension, opposed to 45% and 19% in those with one or two and no infections

- 82 -

respectively. Previous work in rats using a lipopolysaccharide injection with hypoxia-

ischemia injury model showed a greater duration of hypoxia was required to induce

injury in older rats than younger rats, indicating the importance of the developmental

window on the systemic response and the vulnerability of the preterm brain to injury

(Eklind et al, 2005). How hypoxia-ischemia and systemic inflammation interact in the

genesis of neonatal brain injury and dysmaturation requires further study (Back and

Miller, 2014; Fleiss and Gressens, 2012).

Another potential contributor to brain dysmaturation includes injury resulting from

hyperoxia with hypocarbia. BPD is a known complication of chronic ventilation in the

preterm baby, with hyperoxia a well-established causative factor in the development of

BPD (Buczynski et al, 2013). Of those infants with three or more infections in our cohort,

65% required long-term oxygen therapy compared to 5% for the no infection group.

Hyperoxia and hypocarbia are potent vasoconstrictors important in the pressure-passive

autoregulatory system of the preterm brain, and are independent risk factors in the

development of WMI (Shankaran et al, 2006). The impacts with which these factors also

contribute to the poorer outcomes seen in infants with multiple infection also needs

further study.

Multiple postnatal infections have been shown to be associated with progressive

WMI on subsequent MRI scans (Chau et al, 2009; Glass et al, 2008). Within our cohort

we found no difference in the rate or volume of MRI-detected WMI, ventriculomegaly or

IVH and the numbers or types of infections. There was an increased rate of cerebellar

hemorrhage among those infants with three or more infections, which may reflect their

lower gestational age at birth, making them more vulnerable to cerebellar hemorrhage.

- 83 -

This has potential implications on the outcomes analysis due to the role that cerebellar

hemorrhage has in association with poor neurological outcomes (Tam et al, 2011).

Adams et al, in a smaller subset of the current cohort, showed that postnatal infection

was associated with slower increase in corticospinal FA over time than non-infected

newborns (Adams et al, 2010). These findings were expanded upon within our study to

show the maturational delays in those infants with three or more infections included the

corpus callosum and white matter regions including the PLIC on both tradition DTI and

TBSS analyses with supportive findings on MRSI analyses (Chau et al, 2012).

Follow-up at 36 months CA revealed that those infants with three or more

infections had poorer motor development than those infants with fewer infections. In

comparison to other large cohorts of preterm newborns with infection, we did not see an

increased risk of cognitive impairment in those infants with higher numbers of infections

(Rand et al, 2016; Stoll et al, 2004). Some of the differences seen from previous cohorts

may be due to improvements in clinical practices and NICU care in recent decades. In

addition, the majority of infants in this cohort are from high-resource settings which are

known to impact cognitive and language outcomes greater than motor development

(Cusson, 2003; Gross et al, 2001; Howard et al, 2011). The differences from previous

literature, with a predisposition for motor tracts and impairments seen in our study, may

be a result of the changes seen in the distribution of WMI, from PVL in older studies, to

more diffuse WMI in contemporary studies, with diffuse WMI resulting in less

widespread axonal changes (Buser et al, 2012). As the motor tracts are some of the first

areas to myelinate, they are also the most susceptible to injury in the preterm brain (Sie

et al, 1997; Welker and Patton, 2012).

- 84 -

Limitations

As described previously in Chau et al (Chau et al, 2012), many of the infants who

had multiple infections would have had later MRI scans due to taking longer to become

clinically stable, which would result in differences in postnatal age at the MRI between

the infection groups. This would, however, favor the delayed images appearing more

mature in the multiple infection groups underestimating the extent of the injury and time

interaction. It should also be considered that the “clinical-only” infections could be the

result of a virus, which would also have implications on the white matter development,

immune system development and pro-inflammatory markers. However, we did not find

neuroimaging abnormalities consistent with known descriptions of viral infections, nor

were there any positive viral cultures. Overall, our findings support the need for further

monitoring of infants with postnatal infections and continued assessment of how to

improve the care of these infants.

Conclusions

Three or more postnatal infections are associated with delays in brain

maturation, particularly in areas of motor function, with poorer motor outcomes at 36

months CA. These results highlight the vulnerability of the preterm brain to multiple

postnatal infections and supports the potential for combined detrimental effects of

inflammation and hypoxia-ischemia within the NICU. Furthermore, it suggests the need

for a personalized approach to infection control in those infants with one or two

infections. Preventing further infections in this vulnerable group of preterm newborns

may have the potential to improve outcomes. More research is needed on the function

and role of the developing immune system and the impact of hypoxia-ischemia and

- 85 -

infection on pro-inflammatory markers, as well as investigating potential therapies to

correct this inflammatory imbalance within the preterm newborn.

- 86 -

Chapter 4

Summary of Main Findings

and

Future Directions

- 87 -

4.1 Conclusions

Through the use of multi-modal MR and outcome assessments I have shown that

severe ROP is associated with delays in brain maturation with poorer outcomes at 18

months. This work supports the need for research that investigates the potential for

recovery in the most severe ROP cases through balancing of growth factors and

preservation of vision, and points to the significance of severe ROP as a marker of

outcomes. Moreover, with similar methods of assessment, we were able to show

impairments in brain maturation in widespread brain regions for those infants with greater

numbers of postnatal infections resulting in poorer motor outcomes at 3 years, thereby

stressing the strong association of infection with poorer brain maturation. This work

supports several basic science and clinical studies on impairment of brain maturation of

the preterm infant and provides clinicians and families with the knowledge to make

informed decisions about likely outcomes and areas in which outcomes can be improved.

Interventional follow-up programs remain an important monitor for those children with

developmental impairments, though their implementation varies by country, location, and

therapist with wide differences in care and follow-up practices. While much is known

about the potential risk factors and diseases that increase the likelihood of poorer

outcomes in preterm infants, many parents and clinicians are left with more questions

than answers at a critical period in development. The work presented in this thesis

supports further research in postnatal infection and severe ROP and highlights each as

a factor which may adversely influence long term outcomes.

- 88 -

4.2 Future Directions

In considering the future potential for this research, several areas of this study warrant

expanded analysis and research, and this work may enhance the development of other

research disciplines.

4.2.1 Retinopathy of Prematurity and Neurodevelopment

In the care of the infant with ROP, the optimal treatment window for this disorder

remains unknown. Current therapies involve serial observations and monitoring, with

treatments provided after the development of “threshold ROP” when it is believed

treatment is needed in order to preserve vision. It is unknown at present whether even

earlier intervention, with treatment on lower stages of ROP than what is currently

provided, improves visual outcomes, neurodevelopmental outcomes and brain

development. Furthermore, the optimal therapy for the treatment of severe ROP remains

unknown. Several methods of treatment are currently in use, or in development, with laser

photocoagulation therapy still considered one of the gold standard therapies, though is

starting to fall out of favour. Intravitreous bevacizumab (Avastin), an anti-VEGF

monoclonal antibody, has been used more recently for severe ROP. Treatment failure

(Patel et al, 2012) and neurodevelopmental outcomes (Morin et al, 2016) have been

shown in one study to be worse among those treated with intravitreous bevacizumab

compared to laser therapy, which raises concerns of the impact of bevacizumab on the

brain. Another therapy of interest includes oral propranolol which was shown in a clinical

trial to reduce the progression of ROP to higher stages that require treatment, however

significant safety issues with bradycardia and hypotension may limit its use (Filippi et al,

- 89 -

2013). Several other monoclonal antibodies are also currently under investigation as

potential ROP treatments, including ranibizumab (Lucentis), aflibercept (Eylea) and

pegaptanib (Macugen). Which method of treatment will be the preferred therapy, or

whether each will be used in a more personalized approach to the stage and location of

the eye disease present, will be determined as each method is validated and compared

through clinical trials.

Visual outcomes of preterm children with ROP are improving, though preterm

children continue to have poorer visual outcomes than their peers without ROP (Al-Otaibi

et al, 2012; Cryotherapy for Retinopathy of Prematurity Group, 2002; Pearce et al, 1998).

While it has been shown, that through quality improvement measures, severe ROP can

be reduced (Lee et al, 2014), severe ROP continues to be a commonly observed disorder

among preterm infants < 28 weeks GA. Prevention of ROP altogether remains an elusive

goal given the complex mechanisms involved in its development, with infection, chronic

oxygen requirements and other complications of preterm birth keys to its pathophysiology.

In each of these conditions, the implications of various growth factors in modulating the

effects on preterm brain development are largely unknown. While there remains ongoing

research into the impact of IGF-1 and VEGF as potential treatments for ROP, there is

also a field of research in using these treatments to improve brain development (Hansen-

Pupp et al, 2011). The use of IGF-1 levels as a biomarker of delayed brain maturation in

the preterm infant has been reported, with investigations underway currently using

recombinant humanized IGF-1 to see if it assists in accelerating postnatal brain

development (Hansen-Pupp et al, 2013), and whether it improves outcomes. There are

also several ongoing clinical trials using IGF-1 therapy in an attempt to improve

- 90 -

neurodevelopment and outcomes in Autism (Riikonen, 2016), Rett’s syndrome (Khwaja

et al, 2014), and Duchenne’s muscular dystrophy (Malik et al, 2012), with overall positive

effects presented thus far. The effects of IGF-1 treatment in preterm infants remains

unknown at present, but with ongoing research and further investigation it may become

a staple therapy for the preterm infant in treating ROP and supporting brain growth.

Animal research of IGF-1 supports its use in neuroprotection with reduced cytokine and

tumour necrosis factor production (Sukhanov et al, 2007) and a reduced infarct volume

in animal models of stroke (DeGeyter et al, 2016). Other populations that appear worth

further investigation into whether IGF-1 can provide neuroprotection or improve brain

growth and neurodevelopmental outcomes, include infants with neonatal infarcts or

strokes, intra-uterine growth restriction (IUGR), and congenital heart disease (CHD).

While my analysis supports severe ROP as resulting in poorer outcomes, we need

further longitudinal assessment to confirm its implications into childhood and beyond.

Assessments at 18 months CA have been shown to have poor correlation with childhood

outcomes and may overestimate cognitive impairments (Hack et al, 2005), hence follow-

up at 3 years and 4.5 years will be important in determining whether these delays persist

into early childhood. Furthermore, there are significant differences in the language

abilities of infants at 18 months (Lung et al, 2009), many of which are considered within

normal range, making the interpretation of language outcomes at this age difficult,

potentially over or underestimating language impairments. It continues to be unknown

which is the best assessment method for predicting poor childhood outcomes, and when

is the optimal timing for assessing infants to accurately predict normal development in

childhood and beyond. Furthermore, significant alterations in the developmental

- 91 -

trajectories of an infant can occur given the wide variation in childhood experiences and

educational supports available to infants, making more accurate predictions of outcome

difficult. With more accurate socioeconomic analysis techniques and calculating the

effectiveness of individual early intervention approaches, we may better understand the

post-NICU factors that are the most important in promoting childhood development.

Further characterization of the impairments of infants with severe ROP is needed,

with developmental visual follow-up assessments. While visual outcomes are much

improved compared to previous decades, infants with severe ROP continue to have mild

visual disturbances compared to their unaffected peers (O'Connor et al, 2002). Visual-

motor integration and visual perception scores have been shown to be highly correlated

with fine motor scores in preterm infants (Goyen et al, 2008). The extent to which these

visual-motor impairments are present in severe ROP, and the impact of these

impairments on infants’ brain development, remains unknown. Similarly, visual

impairments are also common in infants with brain injury (Pike et al, 2008), and it is

unclear to what extent that visual dysfunction impacts motor integration and performance.

Developmental coordination disorder (DCD) is a clinical childhood disorder with

severe implications for childhood cognitive and motor performance, and is commonly

associated with behavioural problems. DCD is hard to detect early in infants and remains

a poorly recognized disorder by clinicians. Despite this, DCD is a common disorder in

very preterm children at school age and is seen frequently in those infants with ROP

(Zwicker et al, 2013). Visual disturbances with poor ocular alignment, binocular vision and

refractive errors have all been shown to occur in a high prevalence in children with DCD

(Creavin et al, 2014). This further suggests that mild impairments to the visual system

- 92 -

have the potential to adversely affect the motor and cognitive control of the developing

brain. To better understand the pathophysiology for these impairments, more research is

needed into the visual outcome monitoring of very preterm children in order to properly

assess the impacts of these visual outcomes on development, and to assess their true

prevalence.

Follow-up MRIs in childhood or adolescence of infants born very preterm are also

needed to monitor whether slower maturation on imaging methods DTI and TBSS

continues over subsequent analysis. This is an important step in providing information on

long-term outcomes from severe ROP, and will allow correlation over time of preterm and

term-equivalent studies. These studies will also be of importance in determining which

factors associated with childhood result in the greatest improvements in brain

development into childhood, as we continue to search for additional interventions to

improve outcomes and maximize each child’s potential.

In addition to the techniques utilized in this thesis, investigating severe ROP with

other measures of brain development such as MRSI and probabilistic tractography may

provide more information about the nature of the maturational deficits present.

Tractography and TBSS methods have been shown to correlate with visual outcome

scores in preterm infants, with delays in the optic radiations (Bassi et al, 2008), though

the link with severe ROP was not explored in this group. Emerging techniques, such as

diffusion kertosis imaging, magnetization transfer ratio, myelin water fraction and

quantitative susceptibility mapping, have all been shown to reflect features of white matter

microstructure (Groeschel et al, 2016), and are techniques of interest in determining the

method best predictive of outcomes in preterm infants. Resting state functional

- 93 -

connectivity MRI at term-equivalent age has previously shown that infants with WMI have

aberrant connectivity, with the degree of variation determined by the severity of injury

(Smyser et al, 2013). The use of functional MRI in those infants with severe ROP could

be explored in this population to determine whether the maturational delays correlate with

not only alterations in the resting state networks, but also in the visual-motor integration

pathways, if utilized in older individuals. As MR methods continue to advance, it is hopeful

that we will be able to visualize better the cortex and its development. Current techniques

with volumetric analysis have shown global cerebral cortex volumes to be reduced in

preterm infants at term-equivalent age (Inder et al, 2005) and in WMI (Inder et al, 1999),

with cortical volumes reduced in cortical lobes in late childhood (Kesler et al, 2004),

though these techniques have been unable to assess regional cortical volumes in

sensitive areas in early infancy. Furthermore, alterations to the cortical volumes from

severe ROP, particularly of visual regions in the primary visual cortex and surrounding

regions, is another way in which neuroimaging can help us to understand the

pathophysiology of these maturational delays, and assist in identifying ways in which we

can improve outcomes.

4.2.2 Multiple postnatal infections and neurodevelopment

In the care of the infant with postnatal infection, infection control measures are

largely provided universally and across units in attempts to reduce the incidence of

infections. Despite this, a large number of preterm infants develop nosocomial infections

during their NICU course. Improvements in NICU infection control have been shown to

reduce infection rates and improve outcomes (Davis et al, 2016), though the specific

features of a sepsis quality improvement project that result in lowered infections is

- 94 -

unknown. Increased hand hygiene results in significant decreases in nosocomial

infections, shown in several studies (Lam et al, 2004; Won et al, 2004). A large-scale

structured quality improvement project that included hand hygiene with stringent catheter-

insertion policies and early advancement of enteral feeds was shown to reduce infections

from 17% to 15% (Wirtschafter et al, 2011). While this is a significant improvement, the

high rate of infants with infection is still concerning. In identifying which factors are the

most important to reducing infection rates, and which can improve outcomes, the

relatively high cost of implementing these procedures can be reduced. Of further interest

is whether these procedures can be implemented in lower income countries where

nosocomial infections contribute to a high rate of morbidity and mortality (Zaidi et al,

2005).

Providing personalized infection control measures to those infants most

susceptible to the adverse effects of infection, is an intriguing concept of care that the

results of this thesis supports as a potential intervention to improve outcomes. Further

investigation is needed into whether, for those infants with one or two infections,

outcomes are improved by more individualized infection control measures, such as

reduced contact, more stringent guidelines for hand washing and line insertion, or other

measures. A personalized care package, such as that implemented to reduce the

incidence of intraventricular hemorrhage, has the potential to improve outcomes (Schmid

et al, 2013). How these measures can be applied, while also allowing for the other

features also important for the growth of the infant, are considerations that would need to

be taken into account.

- 95 -

What also remains an important consideration in reducing the impact of infection

is determining whether there exists a specific period of vulnerability in which postnatal

infections are more likely to impact brain development. With the immature immune system

of the preterm newborn, particularly before 28 weeks, it would seem that this is a period

of particular importance. In addition, considering what we currently understand about the

pre-OL cell and its maximal period of susceptibility, between 23 to 32 weeks GA (Back et

al, 2001), it would seem reasonable to consider the period of maximal impact from

infection during this period as well. Despite this, it is not known whether the timing of the

infections is an important factor, and whether each infection has an equally cumulative

impact on the white matter development. This highlights the question whether all

infections are of equal importance in brain development and long term outcomes, and

stresses what factors need more attention when improving the care provided.

As has been discussed previously in this thesis, inflammation and hypoxia-

ischemia appear to have independent, yet contributory roles on the influence of infection

on the brain. Some of this is seen with the impact of the early physiological alterations

calculated via the Score for Neonatal Acute Physiology-II (SNAP-II), a measure of illness

severity in the first 12 hours of NICU admission which calculates a morbidity and mortality

risk score, and includes early measures of mean blood pressure, temperature,

oxygenation, serum pH among other clinical factors. A higher SNAP-II has been shown

to correlate with greater mortality (Harsha and Archana, 2015), delays in the corticospinal

tracts (Zwicker et al, 2013), and poorer cognitive, behavioural, social, educational, and

neurological outcomes at 10 years (Logan et al, 2017). Recent research in the fields of

inflammation and hypoxia-ischemia injury has improved our understanding of the

- 96 -

contribution of those factors on the brain, and despite some divergence of these fields in

clinical research, there is considerable research to support their joint impact upon the

developing brain (Albertsson et al, 2014; Girard et al, 2008; Khwaja and Volpe, 2008). In

developing and assessing neuroprotective treatments and mechanisms in the preterm

brain, a greater understanding is needed of the numerous factors involved in the

pathophysiologies with which infection, inflammation and hypoxia-ischemia impact the

brain. Investigating the effects of pro-inflammatory biomarkers on preterm brain

development and white matter maturation is another area of interest, in addition to

exploring potential therapies to mitigate these effects. Furthermore, consideration should

be given to whether inflammation and hypoxia-ischemia contribute to other clinical

disorders in the preterm newborn, such as ROP, NEC and BPD.

Another growing area of research includes exploring the effects of many of the

drugs that are used in the treatment of the sick neonate. The impact of antibiotics,

commonly prescribed to infants with infections, is of increasing concern with reports of

poorer outcomes among those exposed, particularly early in life (Greenwood et al, 2014;

Kuppala et al, 2011). Postnatal exposures to hydrocortisone or dexamethasone have

been shown to be related to impaired cerebellar growth, though antenatal betamethasone

did not (Tam et al, 2011). Midazolam, a commonly prescribed drug used for sedation and

analgesia in the NICU, has been shown to result in decreased hippocampus growth, as

well as poorer outcomes (Duerden et al, 2016). Morphine, another analgesic agent, has

been shown to be associated with poorer cerebellar growth (Zwicker et al, 2016),

cognitive and motor outcomes (Grunau et al, 2009) as well as a lower IQ (de Graaf et al,

2011). As a result of these findings, many NICUs are adapting to encourage non-

- 97 -

pharmacological interventions for analgesia with non-nutritive sucking, swaddling, music,

kangaroo care, swaddling and facilitated tucking all used in attempting to reduce the effect

of analgesia on the infant. While unequivocal evidence to support their use is not yet

available, there are encouraging reports of their analgesic abilities (Cignacco et al, 2007).

Sucrose is an agent that has been used extensively with non-nutritive sucking during

painful procedures and has shown favourable results in reducing the effects of pain

(Stevens et al, 2008). As we learn more about the effects of various medications

commonly used in the NICU, we will gain more insight into the areas with which we can

further improve the care provided.

One such area of research that has resulted in significant awareness and changes

in practice has been in the field of neonatal pain. With growth of awareness in the ability

for the preterm neonate to experience pain, we have learned much about how the preterm

infant responds to pain, and how these painful experiences can adversely impact brain

development (Brummelte et al, 2012; Ranger et al, 2013) as well as long term cognitive

IQ scores (Vinall et al, 2014). Increased pain scores in the preterm period results in a

slower increase in the FA of the corticospinal tract, even after adjusting for postnatal risk

factors and gestational age (Zwicker et al, 2013). Neonatal pain has also been shown to

impact the visual-perceptual abilities, using magnetoencephalography (MEG), in school

age children who had greater skin breaking procedures in the NICU period (Doesburg et

al, 2013). Similarly, neonatal pain results in alterations of the clinical responses to pain

(Grunau et al, 2001) and to the hypothalamic-pituitary-adrenal axis with lower cortisol

responses to stress (Grunau et al, 2005). The link between pain and abnormal brain

growth is seen in long term follow-up with decreased cortical thickness of the frontal and

- 98 -

parietal lobes at 8 years of age in those children who had greater number of NICU painful

procedures as very preterm infants (Ranger et al, 2013). While it is clear that pain is an

important factor in the development of the preterm infant, little is known about whether

there are other factors associated with infection that potentiate these impacts.

Considering the significant number of procedures that are part of the management and

monitoring of those infants with infection, it is also unclear how much of the impacts of

multiple infections can be attributed to the increased number of painful procedures.

The microbiome of the preterm infant, which represents all the bacteria living in or

on the host, is a field with which there is growing interest in how to maximize outcomes

through investigating, what has been termed, the “microbiota-gut-brain axis”

(DiBartolomeo and Claud, 2016). Research into the microbiome of meconium in preterm

infants has shown that abnormal gut flora, likely introduced through the swallowing of

amniotic fluid, correlates with a higher likelihood of preterm birth and may contribute to

the early inflammatory response (Ardissone et al, 2014). Investigation into the healthy

microbiota, in a mouse model of behavioural abnormalities and autism, showed

improvements in the behaviours exhibited in those mice treated with probiotics (Hsiao et

al, 2013). This has coincided with an expansion into the investigation of the role of

probiotics in the preterm population, with studies reporting improvements with reduced

death and incidence of NEC (Lin et al, 2008), though without an impact upon sepsis

(Mihatsch et al, 2012) or early neurodevelopmental outcomes (Chou et al, 2010). As we

learn more about the healthy microbiota of the preterm newborn, we will be able to further

explore the impact of infection on this important system, and how treatments, such as

with probiotics, may help improve outcomes in this vulnerable population.

- 99 -

Nutritional requirements and treatments is another field in which significant

research exists, yet the metabolic requirements and optimal nutritional support of the

infant with infection are not known. Increased brain growth in the NICU of the very preterm

infant, catch-up growth in particular, and increased brain volumes are associated with

improved outcomes at discharge and follow-up (Cheong et al, 2016; Franz et al, 2009).

Conversely, brain atrophy, seen on repeat imaging, is associated with poorer cognitive,

motor and behavioural scores at 3 year follow-up (Horsch et al, 2007). Early brain growth

has a direct relationship with early nutrition with improved growth seen with higher protein

intake (Cormack and Bloomfield, 2013), higher amino acid intake (Poindexter et al, 2006),

higher energy and lipid intake (Beauport et al, 2017), as well as better growth with earlier

enteral feeding (Dinerstein et al, 2006). In addition, the composition of the lipids is an

important factor in development with increased levels of docosahexaenoic acid (DHA)

and reduced levels of linoleic acid levels associated with decreased odds of IVH,

increased brain microstructural development and improved developmental scores on

follow-up (Tam et al, 2016). Suboptimal nutrition in the preterm period has also been

shown to be predictive of childhood intelligent quotient (IQ) scores and a higher frequency

of cerebral palsy in late childhood (Lucas et al, 1998). Similarly, increased intake of breast

milk in the first 4 weeks of life in very preterm infants showed significant improvements in

deep grey matter volumes at term (Belfort et al, 2016). These changes did not persist on

follow-up imaging at 7 years, but subjects with greater breast milk intake showed better

performance on IQ scores, mathematics, working memory, and motor functioning (Belfort

et al, 2016). Using advanced MRI techniques, higher energy and fat intake has been

shown to correlate with improved brain growth, particularly of the basal ganglia and

- 100 -

cerebellum, as well as a positive correlation with FA in the PLIC and higher cognitive and

motor scores at 2 years CA (Coviello et al, 2017). Moreover, early establishment of full

enteral breast feeding in very preterm infants has been shown to be a protective factor in

the development of postnatal sepsis, suggesting a potentially protective effect from the

early breast milk (Ronnestad et al, 2005). Reduced rates of infection have also been

observed to be greater in breast milk fed infants than formula fed infants, further

supporting the potential immune modifying benefits of breast milk (Hylander et al, 1998).

While it is clear that early and adequate nutrition, preferably with breast milk, are important

factors in the growth of the preterm brain, we do not have good evidence on the effects

of infection upon the nutritional intake of the preterm infant. Furthermore, in the clinical

treatment of the sick infant, feeds are often held or stopped in the acute management of

infection, and switched to parenteral nutrition. While this treatment may be necessary in

the evaluation and workup for NEC, the reduced enteral feeds provided during this time

may adversely affect the brain growth during a time of particular vulnerability, and is a

potential contributing source of maturational delays and poor development in those

infants with multiple events of infection. More research is needed to further understand

this relationship.

As noted in chapter 3, a significant number of infants in the multiple infection group

had cerebellar hemorrhage. We know that cerebellar hemorrhage is an important factor

in the development of infants, and is associated with several long term outcome

developmental disorders, including motor and cognitive disabilities (Limperopoulos et al,

2007), autism (Wang et al, 2014) and cerebral palsy (Johnsen et al, 2005). What is also

well known is that there are close connections between the cerebrum and cerebellum

- 101 -

with reduced volumes of the cerebellum seen in those infants with unilateral cerebral

injury (Limperopoulos et al, 2005), with impacts seen in the pons of the brainstem in those

infants with large cerebellar hemorrhages (Parodi et al, 2016). Yet, it is not well known

what is the impact of cerebellar hemorrhage on the growth and development of the brain

in the infant with infection, who may have a greater vulnerability to injury. Conversely, it

remains undetermined how cerebellar hemorrhage impacts the outcomes of the child with

infection and whether it factors into the poorer outcomes described.

The study of the epigenetics is a field of significant interest in predicting long term

outcomes in preterm infants. Epigenetics is a field that involves the study of the impacts

of environmental factors upon genetic expression. Through the processes of DNA

methylation and acetylation, genetic signaling can be reduced or enhanced, potentially

resulting in alterations in disease expression. Epigenetic factors have been shown to be

altered in other populations of children with neurodevelopmental disabilities, such as

Rett’s and Fragile X syndromes (Collins et al, 2004; Gu et al, 1996). Various other at-risk

groups of preterm infants have been shown to have methylation alterations with poorer

attention, self-regulation and quality of movements (Lester et al, 2015), traits which are

commonly observed in children born preterm. Bacterial organisms themselves have been

shown to produce “epimutations” in the epigenome with alterations that can potentially

result in disruptions to host cell immune function and pathogen identification (Bierne et al,

2012). Furthermore, in a mouse model of prenatal infection, stable alterations in the DNA

methylation within the cells of the prefrontal cortex and nucleus accumbens have been

described (Richetto et al, 2017). Those alterations were seen in several regions important

for neural function, WNT signaling, and GABA differentiation (Richetto et al, 2017). ROS

- 102 -

have been shown to impact DNA methylation through the formation of oxidized DNA

lesions which are structurally similar to the methylated DNA cytosines (Lewandowska and

Bartoszek, 2011). In addition, ROS can impact the histone-modifying enzymes involved

in acetylation and methylation with direct effects upon the epigenetic fluctuations in the

cell (Simpson et al, 2012). The impact of multiple infections on the epigenome of the

preterm infant, and what epigenome alterations correspond with outcomes, is a field that

is largely unexplored.

In considering alterations to epigenetic factors, the “epigenetic clock” is a feature

of DNA methylation that has been developed as a biomarker of aging (Horvath, 2013). A

preterm infant epigenetic clock using DNA methylation has been shown to be a strong

predictor of GA in two separate birth cohorts, comparable to antenatal ultrasound

methods estimates (Bohlin et al, 2016; Knight et al, 2016). In utilizing this epigenetic clock,

higher maternal socioeconomic status and higher birthweight percentiles were associated

with an accelerated biological age in preterm infants (Knight et al, 2016). Age acceleration

using the epigenetic clock in adults has been shown to be associated with obesity

(Horvath et al, 2014), HIV infection (Horvath and Levine, 2015), Down’s syndrome

(Horvath et al, 2015), Alzheimer’s disease (Levine et al, 2015) and Parkinson’s disease

(Horvath and Ritz, 2015), and is said to predict cancer, cardiovascular and all-cause

mortality (Christiansen et al, 2016; Marioni et al, 2015; Perna et al, 2016). The preterm

epigenetic clock has been used to show that age acceleration is associated with

advanced maternal age, pre-eclampsia, previous fetal demise, lower 1-minute APGAR

score, antenatal betamethasone and female sex; and that age deceleration is associated

with insulin-treated gestational diabetes mellitus in a previous pregnancy (Girchenko et

- 103 -

al, 2017). The impact of many post-natal events upon the epigenetic age of the preterm

infant remains unknown, yet this area has the potential to provide important information

on alterations to genetic expression, and may aid in the understanding of the

pathophysiology of the many neurodevelopmental impairments of the preterm infant.

There is much to be explored and investigated in the preterm infant in improving

outcomes and contributing to the literature of pathophysiology of diseases that impact

infants and children, while also assisting in the expansion of our understanding of brain

development and diseases.

- 104 -

References

Adams-Chapman, I. and B. J. Stoll (2006). "Neonatal infection and long-term neurodevelopmental outcome in the preterm infant." Curr Opin Infect Dis 19(3): 290-297.

Adams, E., V. Chau, K. J. Poskitt, R. E. Grunau, A. Synnes and S. P. Miller (2010). "Tractography-based quantitation of corticospinal tract development in premature newborns." J Pediatr 156(6): 882-888.

Aden, U., G. Favrais, F. Plaisant, M. Winerdal, U. Felderhoff-Mueser, J. Lampa, V. Lelievre and P. Gressens (2010). "Systemic inflammation sensitizes the neonatal brain to excitotoxicity through a pro-/anti-inflammatory imbalance: key role of TNFalpha pathway and protection by etanercept." Brain Behav Immun 24(5): 747-758.

Al-Otaibi, A. G., S. S. Aldrees and A. A. Mousa (2012). "Long term visual outcomes in laser treated threshold retinopathy of prematurity in Central Saudi Arabia." Saudi J Ophthalmol 26(3): 299-303.

Albertsson, A., D. Bi, L. Duan, X. Zhang, J. Leavenworth, L. Qiao, C. Zhu, S. Cardell, H. Cantor, H. Hagberg, C. Mallard and X. Wang (2014). "The immune response after hypoxia-ischemia in a mouse model of preterm brain injury." Journal of Neuroinflammation 11(153).

Altman, D. I., W. J. Powers, J. M. Perlman, P. Herscovitch, S. L. Volpe and J. J. Volpe (1988). "Cerebral blood flow requirement for brain viability in newborn infants is lower than in adults." Ann Neurol 24(2): 218-226.

Anderson, P., C. D. Luca, E. Hutchinson, G. Roberts and L. Doyle (2010). "Underestimation of developmental delay by the new Bayley-III scale." JAMA Pediatrics 164(4): 352-356.

Anderson, P., K. Treyvaud, J. Neil, J. Cheong, R. Hunt, D. Thompson, K. Lee, L. Doyle and T. Inder (2017). "Associations of Newborn Brain Magnetic Resonance Imaging with Long-Term Neurodevelopmental Impairments in Very Preterm Children." J Pediatr 187(58-65).

Ardissone, A. N., D. M. de la Cruz, A. G. Davis-Richardson, K. T. Rechcigl, N. Li, J. C. Drew, R. Murgas-Torrazza, R. Sharma, M. L. Hudak, E. W. Triplett and J. Neu (2014). "Meconium microbiome analysis identifies bacteria correlated with premature birth." PLoS One 9(3): e90784.

Back, S. and S. Miller (2014). "Brain Injury in Premature Neonates: A Primary Cerebral Dysmaturation Disorder?" Ann Neurol 75: 469-486.

Back, S. A. (2006). "Perinatal white matter injury: the changing spectrum of pathology and emerging insights into pathogenetic mechanisms." Ment Retard Dev Disabil Res Rev 12(2): 129-140.

- 105 -

Back, S. A., X. Gan, Y. Li, P. A. Rosenberg and J. J. Volpe (1998). "Maturation-dependent vulnerability of oligodendrocytes to oxidative stress-induced death caused by glutathione depletion." J Neurosci 18(16): 6241-6253.

Back, S. A., N. L. Luo, N. S. Borenstein, J. M. Levine, J. J. Volpe and H. C. Kinney (2001). "Late oligodendrocyte progenitors coincide with the developmental window of vulnerability for human perinatal white matter injury." J Neurosci 21(4): 1302-1312.

Back, S. A., N. L. Luo, N. S. Borenstein, J. J. Volpe and H. C. Kinney (2002). "Arrested oligodendrocyte lineage progression during human cerebral white matter development: dissociation between the timing of progenitor differentiation and myelinogenesis." J Neuropathol Exp Neurol 61(2): 197-211.

Back, S. A. and S. P. Miller (2014). "Brain injury in premature neonates: A primary cerebral dysmaturation disorder?" Ann Neurol 75(4): 469-486.

Back, S. A., A. Riddle, J. Dean and A. R. Hohimer (2012). "The instrumented fetal sheep as a model of cerebral white matter injury in the premature infant." Neurotherapeutics 9(2): 359-370.

Back, S. A. and S. A. Rivkees (2004). "Emerging concepts in periventricular white matter injury." Semin Perinatol 28(6): 405-414.

Ballabh, P. (2010). "Intraventricular hemorrhage in premature infants: mechanism of disease." Pediatr Res 67(1): 1-8.

Banker, B. Q. and J. C. Larroche (1962). "Periventricular leukomalacia of infancy. A form of neonatal anoxic encephalopathy." Arch Neurol 7: 386-410.

Bassi, L., D. Ricci, A. Volzone, J. M. Allsop, L. Srinivasan, A. Pai, C. Ribes, L. A. Ramenghi, E. Mercuri, F. Mosca, D. Edwards, F. M. Cowan, M. A. Rutherford and S. J. Counsell (2008). "Probabilistic diffusion tractography of the optic radiations and visual function in preterm infants at term equivalent age." Brain 131: 573-582.

Bassler, D., B. J. Stoll, B. Schmidt, E. V. Asztalos, R. S. Roberts, C. M. Robertson and R. S. Sauve (2009). "Using a count of neonatal morbidities to predict poor outcome in extremely low birth weight infants: added role of neonatal infection." Pediatrics 123(1): 313-318.

Bayley, N. (1993). The Bayley Scales of Infant Development II. New York, New York Psychological Corporation.

Bayley, N. (2005). The Bayley Scales of Infant Development III. New York, New York, Psychological Corporation.

- 106 -

Beaulieu, C. (2002). "The basis of anisotropic water diffusion in the nervous system - a technical review." NMR Biomed 15(7-8): 435-455.

Beauport, L., J. Schneider, M. Faouzi, P. Hagmann, P. S. Huppi, J. F. Tolsa, A. C. Truttmann and C. J. Fischer Fumeaux (2017). "Impact of Early Nutritional Intake on Preterm Brain: A Magnetic Resonance Imaging Study." J Pediatr 181: 29-36 e21.

Belfort, M. B., P. J. Anderson, V. A. Nowak, K. J. Lee, C. Molesworth, D. K. Thompson, L. W. Doyle and T. E. Inder (2016). "Breast Milk Feeding, Brain Development, and Neurocognitive Outcomes: A 7-Year Longitudinal Study in Infants Born at Less Than 30 Weeks' Gestation." J Pediatr 177: 133-139 e131.

Bell, M., J. Ternberg, R. Feigin, J. Keating, R. Marshall, L. Barton and T. Brotherton (1978). "Therapeutic decisions based upon clinical staging." Annals of Surgery 187(1): 1-7.

Bell, M. J., J. L. Ternberg, R. D. Feigin, J. P. Keating, R. Marshall, L. Barton and T. Brotherton (1978). "Neonatal necrotizing enterocolitis. Therapeutic decisions based upon clinical staging." Ann Surg 187(1): 1-7.

Bhutta, A., M. Cleves and P. Casey (2002). "Cognitive and behavioral outcomes of school-aged children who were born preterm: a meta-analysis " Jama 288(6): 728-737.

Bierne, H., M. Hamon and P. Cossart (2012). "Epigenetics and bacterial infections." Cold Spring Harb Perspect Med 2(12): a010272.

Biswas, P. S., K. Banerjee, P. R. Kinchington and B. T. Rouse (2006). "Involvement of IL-6 in the paracrine production of VEGF in ocular HSV-1 infection." Exp Eye Res 82(1): 46-54.

Blencowe, H., S. Cousens, M. Oestergaard, D. chou, A. Moller, R. Narwal, A. Adler, C. Garcia, S. Rohde, L. Say and J. Lawn (2012). "National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: a systematic analysis and implications." Lancet 379(9832): 2162-2172.

Bohlin, J., S. E. Haberg, P. Magnus, S. E. Reese, H. K. Gjessing, M. C. Magnus, C. L. Parr, C. M. Page, S. J. London and W. Nystad (2016). "Prediction of gestational age based on genome-wide differentially methylated regions." Genome Biol 17(1): 207.

Boisse, L., A. Mouihate, S. Ellis and Q. J. Pittman (2004). "Long-term alterations in neuroimmune responses after neonatal exposure to lipopolysaccharide." J Neurosci 24(21): 4928-4934.

Brandao, L. and R. Domiques (2003). MR Spectroscopy of the Brain, Lippencott Williams and Wilkins.

- 107 -

Brummelte, S., R. E. Grunau, V. Chau, K. J. Poskitt, R. Brant, J. Vinall, A. Gover, A. R. Synnes and S. P. Miller (2012). "Procedural pain and brain development in premature newborns." Ann Neurol 71: 385-396.

Buczynski, B. W., E. T. Maduekwe and M. A. O'Reilly (2013). "The role of hyperoxia in the pathogenesis of experimental BPD." Semin Perinatol 37(2): 69-78.

Buser, J. R., J. Maire, A. Riddle, X. Gong, T. Nguyen, K. Nelson, N. L. Luo, J. Ren, J. Struve, L. S. Sherman, S. P. Miller, V. Chau, G. Hendson, P. Ballabh, M. R. Grafe and S. A. Back (2012). "Arrested preoligodendrocyte maturation contributes to myelination failure in premature infants." Ann Neurol 71(1): 93-109.

Card, D., R. Nossin-Manor, A. Moore, C. Raybaud, J. Sled and M. Taylor (2013). "Brain metabolite concentrations are associated with illness severity scores and white matter abnormalities in very preterm infants." Pediatr Res 74(1): 75-81.

Chau, V., R. Brant, K. J. Poskitt, E. W. Y. Tam, A. Synnes and S. P. Miller (2012). "Postnatal infection is associated with widespread abnormalities of brain development in premature newborns." Pediatr Res 71(3): 274-279.

Chau, V., K. J. Poskitt, D. E. McFadden, T. Bowen-Roberts, A. Synnes, R. Brant, M. A. Sargent, W. Soulikias and S. P. Miller (2009). "Effect of chorioamnionitis on brain development and injury in premature newborns." Ann Neurol 66(2): 155-164.

Chau, V., A. Synnes, R. E. Grunau, K. J. Poskitt, R. Brant and S. P. Miller (2013). "Abnormal brain maturation in preterm neonates associated with adverse developmental outcomes." Neurology 81: 2082-2089.

Chen, M., A. Citil, F. McCabe, K. M. Leicht, J. Fiascone, C. E. Dammann and O. Dammann (2011). "Infection, oxygen, and immaturity: interacting risk factors for retinopathy of prematurity." Neonatology 99(2): 125-132.

Cheng, C., R. Reinhardt, W. Lee, G. Joncas, S. Patel and C. Bondy (2000). "Insulin-like growth factor 1 regulates developing brain glucose metabolism." PNAS 97(18): 10236-10241.

Cheong, J. L., D. K. Thompson, A. J. Spittle, C. R. Potter, J. M. Walsh, A. C. Burnett, K. J. Lee, J. Chen, R. Beare, L. G. Matthews, R. W. Hunt, P. J. Anderson and L. W. Doyle (2016). "Brain Volumes at Term-Equivalent Age Are Associated with 2-Year Neurodevelopment in Moderate and Late Preterm Children." J Pediatr 174: 91-97 e91.

Chou, I. C., H. T. Kuo, J. S. Chang, S. F. Wu, H. Y. Chiu, B. H. Su and H. C. Lin (2010). "Lack of effects of oral probiotics on growth and neurodevelopmental outcomes in preterm very low birth weight infants." J Pediatr 156(3): 393-396.

- 108 -

Christiansen, L., A. Lenart, Q. Tan, J. W. Vaupel, A. Aviv, M. McGue and K. Christensen (2016). "DNA methylation age is associated with mortality in a longitudinal Danish twin study." Aging Cell 15(1): 149-154.

Cignacco, E., J. P. Hamers, L. Stoffel, R. A. van Lingen, P. Gessler, J. McDougall and M. Nelle (2007). "The efficacy of non-pharmacological interventions in the management of procedural pain in preterm and term neonates. A systematic literature review." Eur J Pain 11(2): 139-152.

Collins, A., J. Levenson, A. Vilaythong and e. al. (2004). "Mild overexpression of MeCP2 causes a progressive neurological disorder in mice. ." Human Molecular Genetics(13): 2679-2689.

Conde-Agudelo, A. and R. Romero (2009). "Antenatal magnesium sulfate for the prevention of cerebral palsy in preterm infants less than 34 weeks' gestation: a systematic review and metaanalysis." American Journal of Obstetrics & Gynecology 200(6): 595-609.

Connolly, B., N. Oberg and R. Gatlin (2012). "Concurrent Validity of the Bayley-III and the Peabody Developmental Motor Scale-2." Pediatr Phys Ther 24: 345-352.

Cormack, B. E. and F. H. Bloomfield (2013). "Increased protein intake decreases postnatal growth faltering in ELBW babies." Arch Dis Child Fetal Neonatal Ed 98(5): F399-404.

Counsell, S., E. Maalouf, A. Fletcher, P. Duggan, M. Battin, H. Lewis, A. Herlihy, D. Edwards, G. Bydder and M. Rutherford (2002). "MR Imaging assessment of myelination in the very preterm brain." Am J Neuroradiol 23: 872-881.

Coviello, C., K. Keunen, K. J. Kersbergen, F. Groenendaal, A. Leemans, B. Peels, I. Isgum, M. A. Viergever, L. S. de Vries, G. Buonocore, V. P. Carnielli and M. Benders (2017). "Effects of early nutrition and growth on brain volumes, white matter microstructure and neurodevelopmental outcome in preterm newborns." Pediatr Res.

Creavin, A. L., R. Lingam, K. Northstone and C. Williams (2014). "Ophthalmic abnormalities in children with developmental coordination disorder." Dev Med Child Neurol 56(2): 164-170.

Cryotherapy for Retinopathy of Prematurity Group (1996). "Multicenter Trial of Cryotherapy for Retinopathy of Prematurity." Archives of Ophthalmology 114(4): 417.

Cryotherapy for Retinopathy of Prematurity Group (2002). "Multicenter Trial of Cryotherapy for Retinopathy of Prematurity." Archives of Ophthalmology 120(5): 595.

- 109 -

Cuenca, A., J. Wynn, L. Moldawer and O. Levy (2013). "Role of innate immunity in neonatal infection." Amer J Perinatol 30: 105-112.

Cusson, R. (2003). "Factors influencing language development in preterm infants." Journal of Obstetric, Gynecologic and Neonatal Nursing 32(3): 402-409.

Dammann, O. (2010). "Inflammation and retinopathy of prematurity." Acta Paediatr 99(7): 975-977.

Dammann, O., K. C. Kuban and A. Leviton (2002). "Perinatal infection, fetal inflammatory response, white matter damage, and cognitive limitations in children born preterm." Ment Retard Dev Disabil Res Rev 8(1): 46-50.

Dammann, O. and A. Leviton (1997). "Maternal intrauterine infection, cytokines, and brain damage in the preterm newborn." Pediatric Research 42(1): 1-8.

Dammann, O. and A. Leviton (1998). "Infection remote from the brain, neonatal white matter damage, and cerebral palsy in the preterm infant." Semin Pediatr Neurol 5(3): 190-201.

Davis, J., D. Odd, S. Jary and K. Luyt (2016). "The impact of a sepsis quality improvement project on neurodisability rates in very low birthweight infants." Arch Dis Child Fetal Neonatal Ed.

de Graaf, J., R. A. van Lingen, S. H. Simons, K. J. Anand, H. J. Duivenvoorden, N. Weisglas-Kuperus, D. W. Roofthooft, L. J. Groot Jebbink, R. R. Veenstra, D. Tibboel and M. van Dijk (2011). "Long-term effects of routine morphine infusion in mechanically ventilated neonates on children's functioning: five-year follow-up of a randomized controlled trial." Pain 152(6): 1391-1397.

de Vries, L. S., I. C. van Haastert, M. J. Benders and F. Groenendaal (2011). "Myth: cerebral palsy cannot be predicted by neonatal brain imaging." Semin Fetal Neonatal Med 16(5): 279-287.

DeGeyter, D., A. DeSmedt, W. Stoop, J. DeKeyser and R. Kooijman (2016). "Central IGF-1 receptors in the brain are instrumental to neuroprotection by systemically infected IGF-1 in a rat model for ischemic stroke." CNS Neuroscience and Therapeutics 22: 611-616.

Delorme, P., F. Goffinet, P. Ancel, F.-H. L, B. Langer, C. Lebeaux, L. M. Marchand, J. Zeitlin, A. Ego, C. Arnaud, C. Vayssiere, E. Lorthe, X. Durrmeyer, L. Sentilhes, D. Subtil, T. Debillon, N. Winer, M. Kaminski, C. D'Ercole, M. Dreyfus, B. Carbonne and G. Kayem (2016). "Cause of preterm birth as a prognostic factor for mortality." Obstet Gynecol 127(40-48).

Demendi, C., B. Borzsonyi, Z. Nagy, J. R. Jr, A. Pajor and J. Joo (2012). "Gene expression patterns of insulin-like growth factor 1, 2 (IGF-1, IGF-2) and insulin-like growth factor binding protein 3 (IGFBP-3) in human

- 110 -

placenta from preterm deliveries: influence of additional factors." Eur J Obstet Gynecol Reprod Biol. 160(1): 40-44.

Dercole, A. and P. Ye (2008). "Expanding the mind: insulin-like growth factor 1 and brain development." Endocrinology 149(12): 5958-5962.

DiBartolomeo, M. E. and E. C. Claud (2016). "The Developing Microbiome of the Preterm Infant." Clin Ther 38(4): 733-739.

Dinerstein, A., R. M. Nieto, C. L. Solana, G. P. Perez, L. E. Otheguy and A. M. Larguia (2006). "Early and aggressive nutritional strategy (parenteral and enteral) decreases postnatal growth failure in very low birth weight infants." J Perinatol 26(7): 436-442.

Doesburg, S. M., C. M. Chau, T. P. Cheung, A. Moiseev, U. Ribary, A. T. Herdman, S. P. Miller, I. L. Cepeda, A. Synnes and R. E. Grunau (2013). "Neonatal pain-related stress, functional cortical activity and visual-perceptual abilities in school-age children born at extremely low gestational age." Pain 154(10): 1946-1952.

Drobyshevsky, A., S. K. Song, G. Gamkrelidze, A. M. Wyrwicz, M. Derrick, F. Meng, L. Li, X. Ji, B. Trommer, D. J. Beardsley, N. L. Luo, S. A. Back and S. Tan (2005). "Developmental changes in diffusion anisotropy coincide with immature oligodendrocyte progression and maturation of compound action potential." J Neurosci 25(25): 5988-5997.

du Plessis, A. J. (2009). "The role of systemic hemodynamic disturbances in prematurity-related brain injury." J Child Neurol 24(9): 1127-1140.

Duerden, E., J. Foong, V. Chau, H. Branson, K. Poskitt, R. Grunau, A. Synnes, J. Zwicker and S. Miller (2015). "Tract-based spatial statistics in preterm-born neonates predicts cognitive and motor outcomes at 18 months." AJNR Am J Neuroradiol 36: 1565-1571.

Duerden, E. G., T. Guo, L. Dodbiba, M. M. Chakravarty, V. Chau, K. J. Poskitt, A. Synnes, R. E. Grunau and S. P. Miller (2016). "Midazolam dose correlates with abnormal hippocampal growth and neurodevelopmental outcome in preterm infants." Ann Neurol 79(4): 548-559.

Duggan, P. J., E. F. Maalouf, T. L. Watts, M. H. Sullivan, S. J. Counsell, J. Allsop, L. Al-Nakib, M. A. Rutherford, M. Battin, I. Roberts and A. D. Edwards (2001). "Intrauterine T-cell activation and increased proinflammatory cytokine concentrations in preterm infants with cerebral lesions." Lancet 358(9294): 1699-1700.

Early Treatment For Retinopathy Of Prematurity Cooperative, G. (2003). "Revised indications for the treatment of retinopathy of prematurity: results of the early treatment for retinopathy of prematurity randomized trial." Arch Ophthalmol 121(12): 1684-1694.

- 111 -

Early treatment of Retinopathy of prematurity cooperative group (2003). Revised indications for the treatment of retinopathy of prematurity. Arch Ophthalmol. E. t. o. r. o. p. c. group. 121: 1684-1696.

Eklind, S., C. Mallard, P. Arvidsson and H. Hagberg (2005). "Lipopolysaccharide induces both a primary and a secondary phase of sensitization in the developing rat brain." Pediatr Res 58(1): 112-116.

Eklind, S., C. Mallard, A.-L. Leverin, E. Gilland, K. Blomgren, I. Mattsby-Baltzer and H. Hagberg (2001). "Bacterial endotoxin sensitizes the immature brain to hypoxic-ischaemic injury." Eur J Neurosci 13(6): 1101-1106.

Ellison, V. J., T. J. Mocatta, C. C. Winterbourn, B. A. Darlow, J. J. Volpe and T. E. Inder (2005). "The relationship of CSF and plasma cytokine levels to cerebral white matter injury in the premature newborn." Pediatr Res 57(2): 282-286.

Fazzi, E., S. Orcesi, L. Caffi, A. Ometto, G. Rondini, C. Telesca and G. Lanzi (1994). "Neurodevelopmental outcome at 5-7 years in preterm infants with periventricular leukomalacia." Neuropediatrics 25(3): 134-139.

Filippi, L., G. Cavallaro, P. Bagnoli, M. Dal Monte, P. Fiorini, G. Donzelli, F. Tinelli, G. Araimo, G. Cristofori, G. la Marca, M. L. Della Bona, A. La Torre, P. Fortunato, S. Furlanetto, S. Osnaghi and F. Mosca (2013). "Oral propranolol for retinopathy of prematurity: risks, safety concerns, and perspectives." J Pediatr 163(6): 1570-1577 e1576.

Fleiss, B. and P. Gressens (2012). "Tertiary mechanisms of brain damage: a new hope for treatment of cerebral palsy?" Lancet Neurol 11(6): 556-566.

Fleiss, B., C. Tann, V. Degos, S. Sigaut, J. V. Steenwinckel, A. Schang, A. Kichev, N. Robertson, C. Mallard, H. Hagberg and P. Gressens (2015). "Inflammation-induced sensitization of the brain in term infants." Developmental Medicine & Child Neurology 57(Suppl. 3): 17-28.

Flierl, M., D. Rittirsch, H. Gao, L. Hoesel, B. Nadeau, D. Day, F. Zetoune, J. Sarma, M. Huber-Lang, J. Ferrara and P. Ward (2008). "Adverse functions of IL-17A in experimental sepsis." FASEB J. 22(7): 2198-2205.

Folio, M. R. and R. R. Fewell (2000). Peabody Developmental Motor Scales (PDMS-2). Austin, PRO-ED.

Franz, A. R., F. Pohlandt, H. Bode, W. A. Mihatsch, S. Sander, M. Kron and J. Steinmacher (2009). "Intrauterine, early neonatal, and postdischarge growth and neurodevelopmental outcome at 5.4 years in extremely preterm infants after intensive neonatal nutritional support." Pediatrics 123(1): e101-109.

- 112 -

French, N. P., R. Hagan, S. F. Evans, A. Mullan and J. P. Newnham (2004). "Repeated antenatal corticosteroids: effects on cerebral palsy and childhood behavior." Am J Obstet Gynecol 190(3): 588-595.

Fujimoto, S., H. Togari, N. Yamaguchi, F. Mizutani, S. Suzuki and H. Sobajima (1994). "Hypocarbia and cystic periventricular leukomalacia in premature infants." Arch Dis Child 71(2): F107-110.

Garfield, C., K. Karbownik, K. Murthy, G. Falciglia, J. Guryan, D. Figlio and J. Roth (2017). "Educational performance of children born prematurely." JAMA Pediatrics 171(8): 764-770.

Gilles, F., P. Gressens, O. Dammann and A. Leviton (2017). "Hypoxia-ischemia is not an antecedent of most preterm brain damage: the illusion of validity." Dev Med Child Neurol.

Gilles, F. H., A. Leviton and C. S. Kerr (1976). "Endotoxin leucoencephalopathy in the telecephalon of the newborn kitten." Journal of the Neurological Sciences 27: 183-191.

Girard, S., H. Kadhim, A. Larouche, M. Roy, F. Gobeil and G. Sebire (2008). "Pro-inflammatory disequilibrium of the IL-1 beta/IL-1ra ratio in an experimental model of perinatal brain damages induced by lipopolysaccharide and hypoxia-ischemia." Cytokine 43(1): 54-62.

Girchenko, P., J. Lahti, D. Czamara, A. K. Knight, M. J. Jones, A. Suarez, E. Hamalainen, E. Kajantie, H. Laivuori, P. M. Villa, R. M. Reynolds, M. S. Kobor, A. K. Smith, E. B. Binder and K. Raikkonen (2017). "Associations between maternal risk factors of adverse pregnancy and birth outcomes and the offspring epigenetic clock of gestational age at birth." Clin Epigenetics 9: 49.

Glass, H. C., S. L. Bonifacio, V. Chau, D. Glidden, K. Poskitt, A. J. Barkovich, D. M. Ferriero and S. P. Miller (2008). "Recurrent postnatal infections are associated with progressive white matter injury in premature infants." Pediatrics 122(2): 299-305.

Glass, T. J. A., V. Chau, J. Gardiner, J. Foong, J. Vinall, J. G. Zwicker, R. E. Grunau, A. Synnes, K. J. Poskitt and S. P. Miller (2017). "Severe retinopathy of prematurity predicts delayed white matter maturation and poorer neurodevelopment." Arch Dis Child Fetal Neonatal Ed 102(6): F532-F537.

Goldenberg, R., J. Culhane, J. Iams and R. Romero (2008). "Epidemiology and causes of preterm birth." The Lancet 371(9606): 75-84.

Gopagondanahalli, K. R., J. Li, M. C. Fahey, R. W. Hunt, G. Jenkin, S. L. Miller and A. Malhotra (2016). "Preterm Hypoxic-Ischemic Encephalopathy." Front Pediatr 4: 114.

Goyen, T.-A., K. Lui and R. Woods (2008). "Visual-motor, visual-perceptual, and fine motor outcomes in very-low-birthweight children at 5 years." Developmental Medicine & Child Neurology 40(2): 76-81.

- 113 -

Goyen, T. A., D. A. Todd, M. Veddovi, A. L. Wright, M. Flaherty and J. Kennedy (2006). "Eye-hand co-ordination skills in very preterm infants <29 weeks gestation at 3 years: Effects of preterm birth and retinopathy of prematurity." Early Hum Dev 82(11): 739-745.

Greenwood, C., A. L. Morrow, A. J. Lagomarcino, M. Altaye, D. H. Taft, Z. Yu, D. S. Newburg, D. V. Ward and K. R. Schibler (2014). "Early empiric antibiotic use in preterm infants is associated with lower bacterial diversity and higher relative abundance of Enterobacter." J Pediatr 165(1): 23-29.

Grether, J. and K. Nelson (1997). "Maternal infection and cerebral palsy in infants of normal birth weight." JAMA 278: 207-211.

Grether, J. K., K. B. Nelson, E. Walsh, R. E. Willoughby and R. W. Redline (2003). "Intrauterine exposure to infection and risk of cerebral palsy in very preterm infants." Arch Pediatr Adolesc Med 157(1): 26-32.

Groeschel, S., G. E. Hagberg, T. Schultz, D. Z. Balla, U. Klose, T. K. Hauser, T. Nagele, O. Bieri, T. Prasloski, A. L. MacKay, I. Krageloh-Mann and K. Scheffler (2016). "Assessing White Matter Microstructure in Brain Regions with Different Myelin Architecture Using MRI." PLoS One 11(11): e0167274.

Gross, S., B. Mettelman, T. Dye and T. Slagle (2001). "Impact of family structure and stability on academic outcome in preterm children at 10 years of age." Pediatrics 138: 169-175.

Grunau, R. E., L. Holsti, D. W. Haley, T. Oberlander, J. Weinberg, A. Solimano, M. F. Whitfield, C. Fitzgerald and W. Yu (2005). "Neonatal procedural pain exposure predicts lower cortisol and behavioral reactivity in preterm infants in the NICU." Pain 113(3): 293-300.

Grunau, R. E., T. F. Oberlander, M. F. Whitfield, C. Fitzgerald and S. K. Lee (2001). "Demographic and Therapeutic Determinants of Pain Reactivity in Very Low Birth Weight Neonates at 32 Weeks' Postconceptional Age." Pediatrics 107(1): 105-112.

Grunau, R. E., M. F. Whitfield, J. Petrie-Thomas, A. R. Synnes, I. L. Cepeda, A. Keidar, M. Rogers, M. Mackay, P. Hubber-Richard and D. Johannesen (2009). "Neonatal pain, parenting stress and interaction, in relation to cognitive and motor development at 8 and 18 months in preterm infants." Pain 143(1-2): 138-146.

Gu, Y., Y. Shen, R. A. Gibbs and D. L. Nelson (1996). "Identification of FMR2, a novel gene associated with the FRAXE CCG repeat and CpG island." Nat Genet 13(1): 109-113.

Gunther, J. and M. Altaweel (2009). "Bevacizumab (avastin) for the treatment of ocular disease." Surv ophthalmol 54: 372-400.

- 114 -

Guo, T., E. Duerden, E. Adams, V. Chau, H. Branson, M. Chakravarty, K. Poskitt, A. Synnes, R. Grunau and S. Miller (2017). "Quantitative assessment of white matter injury in preterm neonates." Neurology 88: 1-9.

Hack, M., H. G. Taylor, D. Drotar, M. Schluchter, L. Cartar, D. Wilson-Costello, N. Klein, H. Friedman, N. Mercuri-Minich and M. Morrow (2005). "Poor predictive validity of the Bayley Scales of Infant Development for cognitive function of extremely low birth weight children at school age." Pediatrics 116(2): 333-341.

Hagberg, H., P. Gressens and C. Mallard (2012). "Inflammation during fetal and neonatal life: implications for neurologic and neuropsychiatric disease in children and adults." Ann Neurol 71(4): 444-457.

Hagberg, H. and C. Mallard (2005). "Effect of inflammation on central nervous system development and vulnerability." Curr Opin Neurol 18(2): 117-123.

Hagberg, H., C. Mallard, D. Ferriero, S. Vannucci, S. Levison, Z. Vexler and P. Gressens (2015). "The role of inflammation in perinatal brain injury." Nat. Rev. Neurol. 11: 192-208.

Hagberg, H., D. Peebles and C. Mallard (2002). "Models of white matter injury: comparison of infectious, hypoxic-ischemic, and excitotoxic insults." Ment Retard Dev Disabil Res Rev 8(1): 30-38.

Hagen, M. W., A. Riddle, E. McClendon, X. Gong, D. Shaver, T. Srivastava, J. M. Dean, J. Z. Bai, T. M. Fowke, A. J. Gunn, D. F. Jones, L. S. Sherman, M. R. Grafe, A. R. Hohimer and S. A. Back (2014). "Role of recurrent hypoxia-ischemia in preterm white matter injury severity." PLoS One 9(11): e112800.

Hamrick, S. E., S. P. Miller, C. Leonard, D. V. Glidden, R. Goldstein, V. Ramaswamy, R. Piecuch and D. M. Ferriero (2004). "Trends in severe brain injury and neurodevelopmental outcome in premature newborn infants: the role of cystic periventricular leukomalacia." J Pediatr 145(5): 593-599.

Han, T., M. Bang, J. Lim, B. Yoon and I. Kim (2002). "Risk factors of cerebral palsy in preterm infants." Am J Phys Med Rehabil 81: 297-303.

Hansen-Pupp, I., H. Holger, C. Lofqvist, L. Hellstrom-Westas, V. Fellman, P. Huppi, A. Hellstrom and D. Ley (2013). "Circulatory IGF1 and brain volumes with neurodevelopment in preterm infants." Pediatr Res 74(5): 564-569.

Hansen-Pupp, I., H. Hovel, A. Hellstrom, L. Hellstrom-Westas, C. Lofqvist, E. Larsson, F. Lazeyras, V. Fellman, P. Huppi and D. Ley (2011). "Postnatal decrease in circulating insulin-like growth factor-1 and low brain volumes in very preterm infants." J Clin Endocrinol Metab(96): 1129-1135.

- 115 -

Harrell, S. N. and D. H. Brandon (2007). "Retinopathy of Prematurity: The disease process, classifications, screening, treatment, and outcomes." Neonatal netw 26(6): 371-378.

Harsha, S. S. and B. R. Archana (2015). "SNAPPE-II (Score for Neonatal Acute Physiology with Perinatal Extension-II) in Predicting Mortality and Morbidity in NICU." J Clin Diagn Res 9(10): SC10-12.

Hashimoto, T., M. Tayama, K. Murakawa, T. Yoshiomoto, M. Miyazaki, M. Harada and Y. Kuroda (1995). "Development of the brainstem and cerebellum in autistic patients." Journal of Autism & Developmental Disorders 25(1): 1-18.

Haynes, R. L., S. S. Billiards, N. S. Borenstein, J. J. Volpe and H. C. Kinney (2008). "Diffuse axonal injury in periventricular leukomalacia as determined by apoptotic marker fractin." Pediatr Res 63(6): 656-661.

Healy, C. M., D. L. Palazzi, M. S. Edwards, J. R. Campbell and C. J. Baker (2004). "Features of invasive staphylococcal disease in neonates." Pediatrics 114(4): 953-961.

Heemskerk, V. H., M. A. R. C. Daemen and W. A. Buurman (1999). "Insulin-like growth factor-1 (IGF-1) and growth hormone (GH) in immunity and inflammation." Cytokine & Growth Factor Reviews 10(1): 5-14.

Hellstrom, A., E. Engstrom, A. Hard, K. Albertsson-Wikland, B. Carlsson, A. Niklasson, C. Lofqvist, E. Svensson, S. Holm, U. Ewald, G. Holmstrom and L. Smith (2003). "Postnatal serum insulin-like growth factor I deficiency is associated with retinopathy of prematurity and other complications of premature birth." Pediatrics 112(5): 1016-1020.

Hellstrom, A., L. H. Smith and O. Dammann (2013). "Retinopathy of Prematurity." Lancet 382: 1445-1457.

Herculano-Houzel, S. (2010). "Coordinated scaling of cortical and cerebellar numbers of neurons." Front Neuroanat 4: 12.

Hoon, A. H., Jr., E. E. Stashinko, L. M. Nagae, D. D. Lin, J. Keller, A. Bastian, M. L. Campbell, E. Levey, S. Mori and M. V. Johnston (2009). "Sensory and motor deficits in children with cerebral palsy born preterm correlate with diffusion tensor imaging abnormalities in thalamocortical pathways." Dev Med Child Neurol 51(9): 697-704.

Horsch, S., C. Muentjes, A. Franz and C. Roll (2007). "Ultrasound diagnosis of brain atrophy is related to neurodevelopmental outcome in preterm infants." Acta Paediatrica 94(12): 1815-1821.

Horvath, S. (2013). "DNA methylation age of human tissues and cell types." Genome Biol 14(10): R115.

Horvath, S., W. Erhart, M. Brosch, O. Ammerpohl, W. von Schonfels, M. Ahrens, N. Heits, J. T. Bell, P. C. Tsai, T. D. Spector, P. Deloukas, R. Siebert, B. Sipos, T. Becker, C. Rocken, C. Schafmayer and J. Hampe

- 116 -

(2014). "Obesity accelerates epigenetic aging of human liver." Proc Natl Acad Sci U S A 111(43): 15538-15543.

Horvath, S., P. Garagnani, M. G. Bacalini, C. Pirazzini, S. Salvioli, D. Gentilini, A. M. Di Blasio, C. Giuliani, S. Tung, H. V. Vinters and C. Franceschi (2015). "Accelerated epigenetic aging in Down syndrome." Aging Cell 14(3): 491-495.

Horvath, S. and A. J. Levine (2015). "HIV-1 Infection Accelerates Age According to the Epigenetic Clock." J Infect Dis 212(10): 1563-1573.

Horvath, S. and B. R. Ritz (2015). "Increased epigenetic age and granulocyte counts in the blood of Parkinson's disease patients." Aging (Albany NY) 7(12): 1130-1142.

Hossain, S., P. S. Shah, X. Y. Ye, B. A. Darlow, S. K. Lee, K. Lui, N. Canadian Neonatal, Australian and N. New Zealand Neonatal (2015). "Outcome comparison of very preterm infants cared for in the neonatal intensive care units in Australia and New Zealand and in Canada." J Paediatr Child Health 51(9): 881-888.

Howard, K., G. Roberts, J. Lim, K. J. Lee, N. Barre, K. Treyvaud, J. Cheong, R. W. Hunt, T. E. Inder, L. W. Doyle and P. J. Anderson (2011). "Biological and environmental factors as predictors of language skills in very preterm children at 5 years of age." J Dev Behav Pediatr 32(3): 239-249.

Hsiao, E. Y., S. W. McBride, S. Hsien, G. Sharon, E. R. Hyde, T. McCue, J. A. Codelli, J. Chow, S. E. Reisman, J. F. Petrosino, P. H. Patterson and S. K. Mazmanian (2013). "Microbiota modulate behavioral and physiological abnormalities associated with neurodevelopmental disorders." Cell 155(7): 1451-1463.

Huppi, P. S., B. Murphy, S. E. Maier, G. P. Zientara, T. E. Inder, P. D. Barnes, R. Kikinis, F. A. Jolesz and J. J. Volpe (2001). "Microstructural brain development after perinatal cerebral white matter injury assessed by diffusion tensor magnetic resonance imaging." Pediatrics 107(3): 455-460.

Hylander, M. A., D. M. Strobino and R. Dhanireddy (1998). "Human Milk Feedings and Infection Among Very Low Birth Weight Infants." Pediatrics 102(3): e38-e38.

Inage, Y. W., M. Itoh and S. Takashima (2000). "Correlation between cerebrovascular maturity and periventricular leukomalacia." Pediatric Neurology 22(3): 204-208.

Inder, T., T. Mocatta, B. Darlow, C. Spencer, J. J. Volpe and C. Winterbourn (2002). "Elevated free radical products in the cerebrospinal fluid of VLBW infants with cerebral white matter injury." Pediatr Res 52(2): 213-218.

- 117 -

Inder, T. E., P. S. Huppi, S. Warfield, R. Kikinis, G. P. Zientara, P. D. Barnes, F. Jolesz and J. J. Volpe (1999). "Periventricular white matter injury in the premature infant is followed by reduced cerebral cortical gray matter volume at term." Ann Neurol 46(5): 755-760.

Inder, T. E., S. K. Warfield, H. Wang, P. S. Huppi and J. J. Volpe (2005). "Abnormal cerebral structure is present at term in premature infants." Pediatrics 115(2): 286-294.

Inder, T. E., S. J. Wells, N. B. Mogridge, C. Spencer and J. J. Volpe (2003). "Defining the nature of the cerebral abnormalities in the premature infant: a qualitative magnetic resonance imaging study." J Pediatr 143(2): 171-179.

International Committee for the Classification of Retinopathy of Prematurity (2005). The International Classification of Retinopathy of Prematurity Revisited. Arch Ophthalmol. I. C. f. t. C. o. R. o. Prematurity. 123: 991-999.

James, J., D. Munson, S. B. DeMauro, J. C. Langer, A. R. Dworetz, G. Natarajan, M. Bidegain, C. A. Fortney, R. Seabrook, B. R. Vohr, J. E. Tyson, E. F. Bell, B. B. Poindexter, S. Shankaran, R. D. Higgins, A. Das, B. J. Stoll, H. Kirpalani, H. Eunice Kennedy Shriver National Institute of Child and N. Human Development Neonatal Research (2017). "Outcomes of Preterm Infants following Discussions about Withdrawal or Withholding of Life Support." J Pediatr.

Jiang, J., N. Chiu, F. Huang, H. Kao, C. Hsu, H. Hung, J. Chang and C. Peng (2004). "Neonatal sepsis in the neonatal intensive care unit: characteristics of early versus late onset." Journal of Microbiology, immunology and infection 37(5): 301-306.

Johnsen, S. D., J. B. Bodensteiner and T. E. Lotze (2005). "Frequency and nature of cerebellar injury in the extremely premature survivor with cerebral palsy." J Child Neurol 20(1): 60-64.

Johnson, S., C. Hollis, P. Kochhar, E. Hennessy, D. Wolke and N. Marlow (2010). "Autism Spectrum Disorders in Extremely Preterm Children." Journal of Pediatrics 156(4): 525-531.e522.

Johnson, S. and N. Marlow (2006). "Developmental screen or developmental testing?" Early Human Development 82(3): 173-183.

Kadhim, H., B. Tabarki, G. Verellen, C. De Prez, A. M. Rona and G. Sebire (2001). "Inflammatory cytokines in the pathogenesis of periventricular leukomalacia." Neurology 56(10): 1278-1284.

Kan, B., H. R. Razzaghian and P. M. Lavoie (2016). "An immunological prespective on neonatal sepsis." Trends Mol Med 22(4): 290-302.

- 118 -

Kersbergen, K. J., M. J. Benders, F. Groenendaal, C. Koopman-Esseboom, R. A. Nievelstein, I. C. van Haastert and L. S. de Vries (2014). "Different patterns of punctate white matter lesions in serially scanned preterm infants." PLoS One 9(10): e108904.

Kesler, S. R., L. R. Ment, B. Vohr, S. K. Pajot, K. C. Schneider, K. H. Katz, T. B. Ebbitt, C. C. Duncan, R. W. Makuch and A. L. Reiss (2004). "Volumetric analysis of regional cerebral development in preterm children." Pediatr Neurol 31(5): 318-325.

Khwaja, O. and J. J. Volpe (2008). "Pathogenesis of cerebral white matter injury of prematurity." Arch Dis Child Fetal Neonatal Ed 93(2): F153-161.

Khwaja, O. S., E. Ho, K. V. Barnes, H. M. O'Leary, L. M. Pereira, Y. Finkelstein, C. A. Nelson, 3rd, V. Vogel-Farley, G. DeGregorio, I. A. Holm, U. Khatwa, K. Kapur, M. E. Alexander, D. M. Finnegan, N. G. Cantwell, A. C. Walco, L. Rappaport, M. Gregas, R. N. Fichorova, M. W. Shannon, M. Sur and W. E. Kaufmann (2014). "Safety, pharmacokinetics, and preliminary assessment of efficacy of mecasermin (recombinant human IGF-1) for the treatment of Rett syndrome." Proc Natl Acad Sci U S A 111(12): 4596-4601.

Kidokoro, H., P. J. Anderson, L. W. Doyle, L. J. Woodward, J. J. Neil and T. E. Inder (2014). "Brain injury and altered brain growth in preterm infants: Predictors and prognosis." Pediatrics 134(2): e444-e453.

Kiechl-Kohlendorfer, E., E. Ralser, U. P. Peglow, G. Reiter and R. Trawuger (2009). "Adverse neurodevelopmental outcome in preterm infants: risk factor profiles for different gestational ages." Acta Paediatrica 98: 792-796.

Kinney, H. C. and S. A. Back (1998). "Human oligodendroglial development: relationship to periventricular leukomalacia." Seminars in Pediatric Neurology 5(3): 180-189.

Kluckow, M. and N. Evans (2000). "Low superior vena cava flow and intraventricular haemorrhage in preterm infants." Archives of Disease in Childhood - Fetal and Neonatal Edition 82(3): 188F-194.

Knight, A. K., J. M. Craig, C. Theda, M. Baekvad-Hansen, J. Bybjerg-Grauholm, C. S. Hansen, M. V. Hollegaard, D. M. Hougaard, P. B. Mortensen, S. M. Weinsheimer, T. M. Werge, P. A. Brennan, J. F. Cubells, D. J. Newport, Z. N. Stowe, J. L. Cheong, P. Dalach, L. W. Doyle, Y. J. Loke, A. A. Baccarelli, A. C. Just, R. O. Wright, M. M. Tellez-Rojo, K. Svensson, L. Trevisi, E. M. Kennedy, E. B. Binder, S. Iurato, D. Czamara, K. Raikkonen, J. M. Lahti, A. K. Pesonen, E. Kajantie, P. M. Villa, H. Laivuori, E. Hamalainen, H. J. Park, L. B. Bailey, S. E. Parets, V. Kilaru, R. Menon, S. Horvath, N. R. Bush, K. Z. LeWinn, F. A. Tylavsky, K. N. Conneely and A. K. Smith (2016). "An epigenetic clock for gestational age at birth based on blood methylation data." Genome Biol 17(1): 206.

Kuppala, V. S., J. Meinzen-Derr, A. L. Morrow and K. R. Schibler (2011). "Prolonged initial empirical antibiotic treatment is associated with adverse outcomes in premature infants." J Pediatr 159(5): 720-725.

- 119 -

Kwon, S. H., L. Vasung, L. R. Ment and P. S. Huppi (2014). "The role of neuroimaging in predicting neurodevelopmental outcomes of preterm neonates." Clin Perinatol 41(1): 257-283.

Lam, B. C., J. Lee and Y. L. Lau (2004). "Hand hygiene practices in a neonatal intensive care unit: a multimodal intervention and impact on nosocomial infection." Pediatrics 114(5): e565-571.

Larouche, A., M. Roy, H. Kadhim, A. M. Tsanaclis, D. Fortin and G. Sebire (2005). "Neuronal injuries induced by perinatal hypoxic-ischemic insults are potentiated by prenatal exposure to lipopolysaccharide: animal model for perinatally acquired encephalopathy." Dev Neurosci 27(2-4): 134-142.

Lavoie, P., Q. Huang, E. Jolette, M. Whalen, A. Nuyt, F. Audibert, D. Speert, T. Lacaze-Masmonteil, H. Soudeyns and T. Kollmann (2010). "Profound lack of interleukin (IL)–12/IL-23p40 in neonates born early in gestation is associated with an increased risk of sepsis." J Infect Dis 202(11): 1754-1763.

Lee, J. and O. Dammann (2012). "Perinatal infection, inflammation, and retinopathy of prematurity." Semin Fetal Neonatal Med 17(1): 26-29.

Lee, S. K., P. S. Shah, N. Singhal, K. Aziz, A. Synnes, D. McMillan, M. M. Seshia and E. S. G. Canadian (2014). "Association of a quality improvement program with neonatal outcomes in extremely preterm infants: a prospective cohort study." CMAJ 186(13): E485-494.

Lee, W., H. Al-Dossary, C. Raybaud, J. Young, B. Mordan, H. Whyte, J. Sled, M. Taylor and M. Shroff (2016). "Longitudinal cerebellar growth following very preterm birth." J Magn Reson Imaging 43(1462-1473).

Lehnardt, S., L. Massillon, P. Follett, F. E. Jensen, R. Ratan, P. A. Rosenberg, J. J. Volpe and T. Vartanian (2003). "Activation of innate immunity in the CNS triggers neurodegeneration through a Toll-like receptor 4-dependent pathway." Proc Natl Acad Sci U S A 100(14): 8514-8519.

Lester, B. M., C. J. Marsit, J. Giarraputo, K. Hawes, L. L. LaGasse and J. F. Padbury (2015). "Neurobehavior related to epigenetic differences in preterm infants." Epigenomics 7(7): 1123-1136.

Levine, M. E., A. T. Lu, D. A. Bennett and S. Horvath (2015). "Epigenetic age of the pre-frontal cortex is associated with neuritic plaques, amyloid load, and Alzheimer's disease related cognitive functioning." Aging (Albany NY) 7(12): 1198-1211.

Leviton, A., E. Allred, K. Kuban, J. Hecht, A. Onderdonk, T. O'Shea and N. Paneth (2010). "Microbiologic and histologic characteristics of the extremely preterm infant's placenta predict white matter damage and later cerebral palsy. The ELGAN Study." Pediatr Res 67(1): 95-101.

- 120 -

Leviton, A., E. Blair, O. Dammann and E. Allred (2005). "The wealth of information conveyed by gestational age." J Pediatr 146: 123-127.

Leviton, A., O. Dammann and S. K. Durum (2005). "The adaptive immune response in neonatal cerebral white matter damage." Ann Neurol 58(6): 821-828.

Leviton, A., R. Fichorova, T. O'Shea, K. Kuban, N. Paneth, O. Dammann and E. Allred (2013). "Two-hit model of brain damage in the very preterm newborn: small for gestational age and postnatal systemic inflammation." Pediatr Res 73(3): 362-370.

Lewandowska, J. and A. Bartoszek (2011). "DNA methylation in cancer development, diagnosis and therapy--multiple opportunities for genotoxic agents to act as methylome disruptors or remediators." Mutagenesis 26(4): 475-487.

Limperopoulos, C., H. Bassan, K. Gauvreau, R. L. Robertson, Jr., N. R. Sullivan, C. B. Benson, L. Avery, J. Stewart, J. S. Soul, S. A. Ringer, J. J. Volpe and A. J. duPlessis (2007). "Does cerebellar injury in premature infants contribute to the high prevalence of long-term cognitive, learning, and behavioral disability in survivors?" Pediatrics 120(3): 584-593.

Limperopoulos, C., J. S. Soul, H. Haidar, P. S. Huppi, H. Bassan, S. K. Warfield, R. L. Robertson, M. Moore, P. Akins, J. J. Volpe and A. J. du Plessis (2005). "Impaired trophic interactions between the cerebellum and the cerebrum among preterm infants." Pediatrics 116(4): 844-850.

Lin, H. C., C. H. Hsu, H. L. Chen, M. Y. Chung, J. F. Hsu, R. I. Lien, L. Y. Tsao, C. H. Chen and B. H. Su (2008). "Oral probiotics prevent necrotizing enterocolitis in very low birth weight preterm infants: a multicenter, randomized, controlled trial." Pediatrics 122(4): 693-700.

Linsell, L., R. Malouf, J. Morris, J. J. Kurinczuk and N. Marlow (2016). "Prognostic factors for cerebral palsy and motor impairment in children born very preterm or very low birthweight: a systematic review." Dev Med Child Neurol 58(6): 554-569.

Logan, J. W., O. Dammann, E. N. Allred, C. Dammann, K. Beam, R. M. Joseph, T. M. O'Shea, A. Leviton and K. C. K. Kuban (2017). "Early postnatal illness severity scores predict neurodevelopmental impairments at 10 years of age in children born extremely preterm." J Perinatol 37(5): 606-614.

Lucas, A., R. Morley and T. J. Cole (1998). "Randomised trial of early diet in preterm babies and later intelligence quotient." BMJ 317(7171): 1481-1487.

Lung, F. W., B. C. Shu, T. L. Chiang, P. F. Chen and L. L. Lin (2009). "Predictive validity of Bayley scale in language development of children at 6-36 months." Pediatr Int 51(5): 666-669.

- 121 -

Malik, V., L. R. Rodino-Klapac and J. R. Mendell (2012). "Emerging drugs for Duchenne muscular dystrophy." Expert Opin Emerg Drugs 17(2): 261-277.

Mallard, C. and X. Wang (2012). "Infection-induced vulnerability of perinatal brain injury." Neurol Res Int 2012: 102153.

Manzoni, P., A. Maestri, M. Leonessa, M. Mostert, D. Farina and G. Gomirato (2006). "Fungal and bacterial sepsis and threshold ROP in preterm very low birth weight neonates." J Perinatol 26(1): 23-30.

Marioni, R. E., S. Shah, A. F. McRae, B. H. Chen, E. Colicino, S. E. Harris, J. Gibson, A. K. Henders, P. Redmond, S. R. Cox, A. Pattie, J. Corley, L. Murphy, N. G. Martin, G. W. Montgomery, A. P. Feinberg, M. D. Fallin, M. L. Multhaup, A. E. Jaffe, R. Joehanes, J. Schwartz, A. C. Just, K. L. Lunetta, J. M. Murabito, J. M. Starr, S. Horvath, A. A. Baccarelli, D. Levy, P. M. Visscher, N. R. Wray and I. J. Deary (2015). "DNA methylation age of blood predicts all-cause mortality in later life." Genome Biol 16: 25.

Marlow, N., D. Wolke, M. Bracewell, M. Samara and E. S. Group (2005). "Neurologic and developmental disability at six years of age after extremely preterm birth." New England Journal of Medicine 352(1): 9-19.

Martinez-Biarge, M., F. Groenendaal, K. J. Kersbergen, M. J. Benders, F. Foti, F. M. Cowan and L. S. de Vries (2016). "MRI Based Preterm White Matter Injury Classification: The Importance of Sequential Imaging in Determining Severity of Injury." PLoS One 11(6): e0156245.

Martinez-Castellanos, M. A., S. Schwartz, M. L. Hernandez-Rojas, V. A. Kon-Jara, G. Garcia-Aguirre, J. L. Guerrero-Naranjo, R. V. Chan and H. Quiroz-Mercado (2013). "Long-term effect of antiangiogenic therapy for retinopathy of prematurity up to 5 years of follow-up." Retina 33(2): 329-338.

McClure, E., D. Dudley, U. Reddy and R. Goldenberg (2010). "Infectious causes of stillbirth: a clinical perspective." Clin Obstet Gynecol 53(3): 635-645.

McCrea, H. J. and L. R. Ment (2008). "The diagnosis, management, and postnatal prevention of intraventricular hemorrhage in the preterm neonate." Clin Perinatol 35(4): 777-792, vii.

Messerschmidt, A., R. Fuiko, D. Prayer, P. Brugger, E. Boltshauser, G. Zoder, W. Sterniste, M. Weber and R. Birnbacher (2008). "Disrupted cerebellar development in preterm infants is associated with impaired neurodevelopmental outcome." European Journal of Pediatrics 167(10): 1141-1147.

Mihatsch, W. A., C. P. Braegger, T. Decsi, S. Kolacek, H. Lanzinger, B. Mayer, L. A. Moreno, F. Pohlandt, J. Puntis, R. Shamir, U. Stadtmuller, H. Szajewska, D. Turck and J. B. van Goudoever (2012). "Critical systematic review of the level of evidence for routine use of probiotics for reduction of mortality and prevention of necrotizing enterocolitis and sepsis in preterm infants." Clin Nutr 31(1): 6-15.

- 122 -

Miller, S. P., C. C. Cozzio, R. B. Goldstein, D. M. Ferriero, J. C. Partridge, D. B. Vigneron and A. J. Barkovich (2003). "Comparing the diagnosis of white matter injury in premature newborns with serial MR imaging and transfontanel ultrasonography findings." AJNR Am J Neuroradiol 24(8): 1661-1669.

Miller, S. P., D. M. Ferriero, C. Leonard, R. Piecuch, D. V. Glidden, J. C. Partridge, M. Perez, P. Mukherjee, D. B. Vigneron and A. J. Barkovich (2005). "Early brain injury in premature newborns detected with magnetic resonance imaging is associated with adverse early neurodevelopmental outcome." J Pediatr 147(5): 609-616.

Miller, S. P., M. I. Shevell, Y. Patenaude and A. M. O’Gorman (2000). "Neuromotor spectrum of periventricular leukomalacia in children born at term." Pediatric Neurology 23(2): 155-159.

Miller, S. P., D. B. Vigneron, R. G. Henry, M. A. Bohland, C. Ceppi-Cozzio, C. Hoffman, N. Newton, J. C. Partridge, D. M. Ferriero and A. J. Barkovich (2002). "Serial quantitative diffusion tensor MRI of the premature brain: Development in newborns with and without injury." J Magn Reson Imaging 16(6): 621-632.

Mintz-Hittner, H. A., K. A. Kennedy, A. Z. Chuang and B.-R. C. Group (2011). "Efficacy of intravitreal bevacizumab for stage 3+ retinopathy of prematurity." N Engl J Med 364(7): 603-615.

Mitha, A., L. Foix-L'Helias, C. Arnaud, S. Marret, R. Vieux, Y. Aujard, G. Thiriez, B. Larroque, G. Cambonie, A. Burguet, P. Boileau, J. C. Roze, M. Kaminsku, p. Truffert and P. Ancel (2013). "Neonatal Infection and 5-year Neurodevelopmental Outcome of Very Preterm Infants." Pediatrics 132: e372-e380.

Moffett, J. R., B. Ross, P. Arun, C. N. Madhavarao and A. M. Namboodiri (2007). "N-Acetylaspartate in the CNS: from neurodiagnostics to neurobiology." Prog Neurobiol 81(2): 89-131.

Moore, T., S. Johnson, S. Haider, E. Hennessy and N. Marlow (2016). "Relationship between Test Scores Using the Second and Third Editions of the Bayley Scales in Extremely Preterm Children." Journal of Pediatrics 160(4): 553-558.

Morin, J., T. M. Luu, R. Superstein, L. H. Ospina, F. Lefebvre, M. N. Simard, V. Shah, P. S. Shah, E. N. Kelly, N. Canadian Neonatal and I. the Canadian Neonatal Follow-Up Network (2016). "Neurodevelopmental Outcomes Following Bevacizumab Injections for Retinopathy of Prematurity." Pediatrics 137(4).

Msall, M. E., D. L. Phelps, K. M. DiGaudio, V. Dobson, B. Tung, R. E. McClead, G. E. Quinn, J. D. Reynolds, R. J. Hardy and E. A. Palmer (2000). "Severity of Neonatal Retinopathy of Prematurity Is Predictive of Neurodevelopmental Functional Outcome at Age 5.5 Years." Pediatrics 106(5): 998-1005.

Mukherjee, P., J. H. Miller, J. S. Shimony, J. V. Philip, D. Nehra, A. Z. Snyder, T. E. Conturo, J. J. Neil and R. C. McKinstry (2002). "Diffusion-tensor MR imaging of gray and white matter development during normal human brain maturation." AJNR Am J Neuroradiol 23(9): 1445-1456.

- 123 -

Munro, M. J., A. M. Walker and C. P. Barfield (2004). "Hypotensive extremely low birth weight infants have reduced cerebral blood flow." Pediatrics 114(6): 1591-1596.

Nelson, K. and E. Blair (2015). "Prenatal factors in singletons with cerebral palsy at or near term." New England Journal of Medicine 373: 946-953.

Nelson, K. B. and J. K. Grether (1995). "Can magnesium sulfate reduce the risk of cerebral palsy in very low birthweight infants?" Pediatrics 95(2): 263-269.

Neubauer, V., T. Djurdjevic, E. Griesmaier, M. Biermayr, E. Gizewski and U. Kiechl-Kohlendorfer (2017). "Routine Magnetic Resonance Imaging at Term-Equivalent Age Detects Brain Injury in 25% of a Contemporary Cohort of Very Preterm Infants." PLoS One 12(1): e0169442.

O'Connor, A. R., T. Stephenson, A. Johnson, M. J. Tobin, M. J. Moseley, S. Ratib, Y. Ng and A. R. Fielder (2002). "Long-term ophthalmic outcome of low birth weight children with and without retinopathy of prematurity." Pediatrics 109(1): 12-18.

O'Connor, A. R., T. J. Stephenson, A. Johnson, M. J. Tobin, S. Ratib and A. R. Fielder (2002). "Strabismus in children of birth weight less than 1701 g." Arch Ophthalmol 120(6): 767-773.

O'Donnell, L. J. and C. F. Westin (2011). "An introduction to diffusion tensor image analysis." Neurosurg Clin N Am 22(2): 185-196, viii.

O'Leary, H., M. C. Gregas, C. Limperopoulos, I. Zaretskaya, H. Bassan, J. S. Soul, D. N. Di Salvo and A. J. du Plessis (2009). "Elevated cerebral pressure passivity is associated with prematurity-related intracranial hemorrhage." Pediatrics 124(1): 302-309.

O'Shea, T. M., K. L. Klinepeter, P. J. Meis and R. G. Dillard (1998). "Intrauterine infection and the risk of cerebral palsy in very low-birthweight infants." Paediatr Perinat Epidemiol 12(1): 72-83.

Orsi, G., G. d'Ettorre, A. Panero, F. Chiarini, V. Vullo and M. Venditti (2009). "Hospital-acquired infection surveillance in a neonatal intensive care unit." Am J Infect Control 37: 201-203.

Oskoui, M., F. Coutinho, J. Dykeman, N. Jette and T. Pringsheim (2013). "An update on the prevalence of cerebral palsy: a systematic review and meta-analysis." Dev Med Child Neurol 55(6): 509-519.

Panteliadis, C., P. Panteliadis and F. Vassilyadi (2013). "Hallmarks in the history of cerebral palsy: from antiquity to mid-20th century." Brain & Development 35(4): 285-292.

- 124 -

Papile, L. A., J. Burstein, R. Burstein and H. Koffler (1978). "Incidence and evolution of subependymal and intraventricular hemorrhage: a study of infants with birth weights less than 1,500 gm." J Pediatr 92(4): 529-534.

Parodi, A., L. A. Ramenghi, M. Malova, D. Tortora, M. Severino, G. Morana and A. Rossi (2016). "Crossed Pontine Hemiatrophy Associated with Unilateral Cerebellar Hemorrhage in Premature Infants." Neuropediatrics 47(6): 404-407.

Patel, R. D., M. P. Blair, M. J. Shapiro and S. J. Lichtenstein (2012). "Significant treatment failure with intravitreous bevacizumab for retinopathy of prematurity." Arch Ophthalmol 130(6): 801-802.

Pearce, I. A., F. C. Pennie, L. M. Gannon, A. M. Weindling and D. I. Clark (1998). "Three year visual outcome for treated stage 3 retinopathy of prematurity: cryotherapy versus laser." British Journal of Ophthalmology 82(11): 1254-1259.

Penn, A. A., P. Gressens, B. Fleiss, S. A. Back and V. Gallo (2016). "Controversies in preterm brain injury." Neurobiol Dis 92(Pt A): 90-101.

Perna, L., Y. Zhang, U. Mons, B. Holleczek, K. U. Saum and H. Brenner (2016). "Epigenetic age acceleration predicts cancer, cardiovascular, and all-cause mortality in a German case cohort." Clin Epigenetics 8: 64.

Petrou, S. (2005). "The economic consequences of preterm birth during the first 10 years of life." BJOG: An International Journal of Obstetrics & Gynaecology 112(s1): 10-15.

Pichler, G., P. Y. Cheung, C. Binder, M. O'Reilly, B. Schwaberger, K. Aziz, B. Urlesberger and G. M. Schmolzer (2014). "Time course study of blood pressure in term and preterm infants immediately after birth." PLoS One 9(12): e114504.

Pierson, C. R., R. D. Folkerth, S. S. Billiards, F. L. Trachtenberg, M. E. Drinkwater, J. J. Volpe and H. C. Kinney (2007). "Gray matter injury associated with periventricular leukomalacia in the premature infant." Acta Neuropathol 114(6): 619-631.

Pike, M. G., G. Holmstrom, L. S. Vries, J. M. Pennock, K. J. Drew, P. M. Sonksen and L. M. S. Dubowitz (2008). "Patterns Of Visual Impairment Associated With Lesions Of The Preterm Infant Bran." Developmental Medicine & Child Neurology 36(10): 849-862.

Plaisier, A., M. Roets, C. v. d. Starre, M. Feijen-Roon, P. Govaert, M. Lequin, A. Heemskerk and H. Dudink (2012). "Safety of routine early MRI in preterm infants." Pediatric Radiology 42(10): 1205-1211.

- 125 -

Poindexter, B. B., J. C. Langer, A. M. Dusick, R. A. Ehrenkranz, H. National Institute of Child and N. Human Development Neonatal Research (2006). "Early provision of parenteral amino acids in extremely low birth weight infants: relation to growth and neurodevelopmental outcome." J Pediatr 148(3): 300-305.

Posse, S., R. Otazo, S. Dager and J. Alger (2012). "MR Spectroscopic imaging: principles and recent advances." J Magn Reson Imaging 37: 1301-1325.

Prechtl, H. F. R., G. Cioni, C. Einspieler, A. F. Bos and F. Ferrari (2001). "Role of vision on early motor development: lessons from the blind." Developmental Medicine and Child Neurology 43(03): 198.

Rand, K., N. Austin, T. Inder, S. Bora and L. Woodward (2016). "Neonatal infection and later neurodevelopmental risk in the very preterm infant." J Pediatr 170: 97-104.

Ranger, M., C. M. Chau, A. Garg, T. S. Woodward, M. F. Beg, B. Bjornson, K. Poskitt, K. Fitzpatrick, A. R. Synnes, S. P. Miller and R. E. Grunau (2013). "Neonatal pain-related stress predicts cortical thickness at age 7 years in children born very preterm." PLoS One 8(10): e76702.

Richetto, J., R. Massart, U. Weber-Stadlbauer, M. Szyf, M. A. Riva and U. Meyer (2017). "Genome-wide DNA Methylation Changes in a Mouse Model of Infection-Mediated Neurodevelopmental Disorders." Biol Psychiatry 81(3): 265-276.

Riikonen, R. (2016). "Treatment of autistic spectrum disorder with insulin-like growth factors." Eur J Paediatr Neurol 20(6): 816-823.

Rivera, J., P. Sepieha, J. Joyal, F. Duhamel, Z. Shao, N. Sitaras, E. Picard, E. Zhou, P. Lachapelle and S. Chemtob (2011). "Understanding Retinopathy of Prematurity: Update on Pathogenesis." Neonatology 100: 343-353.

Robertson, C., M. Watt and Y. Yasui (2007). "Changes in the prevalence of cerebral palsy for children born very prematurely within a population-based program over 30 years." JAMA 297(24): 2733-2740.

Ronnestad, A., T. G. Abrahamsen, S. Medbo, H. Reigstad, K. Lossius, P. I. Kaaresen, T. Egeland, I. E. Engelund, L. M. Irgens and T. Markestad (2005). "Late-onset septicemia in a Norwegian national cohort of extremely premature infants receiving very early full human milk feeding." Pediatrics 115(3): e269-276.

Rosenbaum, P., N. Paneth, A. Leviton, M. Goldstein and M. Bax (2006). "A report: the definition and classification of cerebral palsy April 2006." Developmental Medicine & Child Neurology 49(s109): 8-14.

Rothman, A., M. Sevilla, S. Mangalesh, K. Gustafson, L. Edwards, M. Cotten, J. Shimony and e. al (2015). "Thinner retinal nerve fibre layer in very preterm versus term infants and relationship to brain anatomy and neurodevelopment." Am J Ophthalmol 160(6): 1296-1308.

- 126 -

Rouse, D. J., D. G. Hirtz, E. Thom, M. W. Varner, C. Y. Spong, B. M. Mercer, J. D. Iams, R. J. Wapner, Y. Sorokin, J. M. Alexander, M. Harper, J. M. Thorp, Jr., S. M. Ramin, F. D. Malone, M. Carpenter, M. Miodovnik, A. Moawad, M. J. O'Sullivan, A. M. Peaceman, G. D. Hankins, O. Langer, S. N. Caritis and J. M. Roberts (2008). "A randomized, controlled trial of magnesium sulfate for the prevention of cerebral palsy." N Engl J Med 359(9): 895-905.

Rutherford, M., P. Ward and C. Malamatentiou (2005). "Advanced MR techniques in the term-born neonate with perinatal brain injury." Seminars In Fetal & Neonatal Medicine 10: 445-460.

Rysavy, M., L. Lei, E. Bell, A. Das, S. Hintz, B. Stoll, B. Vohr, W. Carlo, S. Shankaran, M. Walsh, J. Tyson, C. Cotten, P. Smith, J. Murray, T. Colaizy, J. Brumbaugh and R. Higgins (2015). "Between-hospital variation in treatment and outcomes in extremely preterm infants." New England Journal of Medicine 372: 1801-1811.

Sarkar, S., S. Shankaran, A. R. Laptook, B. G. Sood, B. Do, B. J. Stoll, K. P. Van Meurs, E. F. Bell, A. Das, J. Barks, H. Generic Database Subcommittee of the Eunice Kennedy Shriver National Institute of Child and N. Human Development Neonatal Research (2015). "Screening Cranial Imaging at Multiple Time Points Improves Cystic Periventricular Leukomalacia Detection." Am J Perinatol 32(10): 973-979.

Schelonka, R., S. Scruggs, K. Dimmitt and W. Carlo (2006). "Sustained reductions in neonatal nosocomial infection rates following a comprehensive infection control intervention." Journal of Perinatology 26(3): 176-179.

Schlapbach, L., M. Aebischer, M. Adams, G. Natalucci, J. Bonhoeffer, P. Latzin, M. Nelle, H. Bucher and B. Latal (2011). "Impact of Sepsis on Neurodevelopmental Outcome in a Swiss National Cohort of Extremely Premature Infants." Pediatrics 128: e348-e357.

Schmid, M. B., F. Reister, B. Mayer, R. J. Hopfner, H. Fuchs and H. D. Hummler (2013). "Prospective risk factor monitoring reduces intracranial hemorrhage rates in preterm infants." Dtsch Arztebl Int 110(29-30): 489-496.

Schmidt, B., E. V. Asztalos, R. S. Roberts, C. M. Robertson, R. S. Sauve and M. F. Whitfield (2003). "Impact of bronchopulmonary dysplasia, brain injury, and severe retinopathy on the outcome of extremely low-birth-weight infants at 18 months: results from the trial of indomethacin prophylaxis in preterms." JAMA 289(9): 1124-1129.

Schmidt, B., P. Davis, E. Asztalos, A. Solimano and R. Roberts (2014). "Association Between Severe Retinopathy of Prematurity and Nonvisual Disabilities at Age 5 Years." JAMA 311(5): 523-524.

- 127 -

Schmidt, B., R. Roberts, P. Davis, L. Doyle, E. Asztalos, G. Opie, A. Bairam, A. Solimano, S. Arnon and R. Sauve (2015). "Prediction of late death or disability at age 5 year using count of 3 neonatal morbidities in very low birth weight infants." J Pediatr 167: 982-986.

Schmidt, B., R. S. Roberts, P. Davis, L. W. Doyle, K. J. Barrington, A. Ohlsson, A. Solimano, W. Tin and G. Caffeine for Apnea of Prematurity Trial (2007). "Long-term effects of caffeine therapy for apnea of prematurity." N Engl J Med 357(19): 1893-1902.

Sears, J. E., J. Pietz, C. Sonnie, D. Dolcini and G. Hoppe (2009). "A change in oxygen supplementation can decrease the incidence of retinopathy of prematurity." Ophthalmology 116(3): 513-518.

Sellier, E., M. J. Platt, G. L. Andersen, I. Krageloh-Mann, J. De La Cruz, C. Cans and N. Surveillance of Cerebral Palsy (2016). "Decreasing prevalence in cerebral palsy: a multi-site European population-based study, 1980 to 2003." Dev Med Child Neurol 58(1): 85-92.

Seri, I. and J. Evans (2008). "Limits of viability: definition of the gray zone." J Perinatol. 28: S4-S8.

Shah, D. K., L. W. Doyle, P. J. Anderson, M. Bear, A. J. Daley, R. W. Hunt and T. E. Inder (2008). "Adverse neurodevelopment in preterm infants with postnatal sepsis or necrotizing enterocolitis is mediated by white matter abnormalities on magnetic resonance imaging at term." J Pediatr 153(2): 170-175.

Shah, P. S., K. Lui, G. Sjors, L. Mirea, B. Reichman, M. Adams, N. Modi, B. A. Darlow, S. Kusuda, L. San Feliciano, J. Yang, S. Hakansson, R. Mori, D. Bassler, J. Figueras-Aloy, S. K. Lee and N. International Network for Evaluating Outcomes of (2016). "Neonatal Outcomes of Very Low Birth Weight and Very Preterm Neonates: An International Comparison." J Pediatr 177: 144-152 e146.

Shankaran, S., J. C. Langer, S. N. Kazzi, A. R. Laptook, M. Walsh, H. National Institute of Child and N. Human Development Neonatal Research (2006). "Cumulative index of exposure to hypocarbia and hyperoxia as risk factors for periventricular leukomalacia in low birth weight infants." Pediatrics 118(4): 1654-1659.

Sie, L. T., M. S. van der Knaap, G. van Wezel-Meijler and J. Valk (1997). "MRI assessment of myelination of motor and sensory pathways in the brain of preterm and term-born infants." Neuropediatrics 28(2): 97-105.

Simpson, N. E., V. P. Tryndyak, M. Pogribna, F. A. Beland and I. P. Pogribny (2012). "Modifying metabolically sensitive histone marks by inhibiting glutamine metabolism affects gene expression and alters cancer cell phenotype." Epigenetics 7(12): 1413-1420.

Smith, S. M., M. Jenkinson, H. Johansen-Berg, D. Rueckert, T. E. Nichols, C. E. Mackay, K. E. Watkins, O. Ciccarelli, M. Z. Cader, P. M. Matthews and a. T. E. J. Behrens. (2006). "Tract-based spatial statistics: Voxelwise analysis of multi-subject diffusion data." NeuroImage 31: 1487-1505.

- 128 -

Smyser, C. D., A. Z. Snyder, J. S. Shimony, T. M. Blazey, T. E. Inder and J. J. Neil (2013). "Effects of white matter injury on resting state fMRI measures in prematurely born infants." PLoS One 8(7): e68098.

Soares, J., P. Marques, V. Alves and N. Sousa (2013). "A hitchhiker's guide to diffusion tensor imaging." frontiers in neuroscience 7(31): 1-14.

Soul, J. S., P. E. Hammer, M. Tsuji, J. P. Saul, H. Bassan, C. Limperopoulos, D. N. Disalvo, M. Moore, P. Akins, S. Ringer, J. J. Volpe, F. Trachtenberg and A. J. du Plessis (2007). "Fluctuating pressure-passivity is common in the cerebral circulation of sick premature infants." Pediatr Res 61(4): 467-473.

Spittle, A. J., J. Cheong, L. W. Doyle, G. Roberts, K. J. Lee, J. Lim, R. W. Hunt, T. E. Inder and P. J. Anderson (2011). "Neonatal white matter abnormality predicts childhood motor impairment in very preterm children." Dev Med Child Neurol 53(11): 1000-1006.

StataCorp (2015). "Stata Statistical Software." Release 14 College Station, TX: StataCorp LP.

Steggerda, S., T. DeBruine, A. V. d. Berg-Huysmans, M. Rijken, L. Leiser, F. Walther and G. v. Wezel-Meijler (2013). "Small Cerebellar Hemorrhage in Preterm Infants: Perinatal and Postnatal Factors and Outcome." Cerebellum 12: 794-801.

Stevens, B., A. Taddio, A. Ohlsson and T. Einarson (2008). "The efficacy of sucrose for relieving procedural pain in neonates-a systematic review and meta-analysis." Acta Paediatrica 86(8): 837-842.

Stoll, B., N. Hansen, E. Bell, S. Shankaran, A. Laptook, M. Walsh, E. Hale, N. Newman, K. Schibler, W. Carlo, K. Kennedy, B. Poindexter, N. Finer, R. Ehrenkranz, S. Duara, P. Sanchez, T. O'Shea, R. Goldberg, K. V. Meurs, R. Faix, D. Phelps, I. Frantz, K. Watterberg, S. Saha, A. Das and R. Higgins (2010). "Neonatal outcomes of extremely preterm infants from the NICHD neonatal research network." Pediatrics 126(443-456).

Stoll, B., N. Hansen, R. Higgins, A. Fanaroff, S. Duara, R. Goldberg, A. Laptook, M. Walsh, M. Oh and E. Hale (2005). "Very Low Birth Weight Preterm Infants With Early Onset Neonatal Sepsis." Pediatr Infect Dis J 24: 635-639.

Stoll, B. J., N. I. Hansen, I. Adams-Chapman, A. A. Fanaroff, S. R. Hintz, B. Vohr and R. D. Higgins (2004). "Neurodevelopmental and growth impairment among extremely low-birth-weight infants with neonatal infection." JAMA 292(19): 2357-2365.

Strunk, T., T. Inder, X. Wang, D. Burgner, C. Mallard and O. Levy (2014). "Infection-induced inflammation and cerebral injury in preterm infants." Lancet Infect Dis 14: 751-762.

- 129 -

Sukhanov, S., Y. Higashi, S. Y. Shai, C. Vaughn, J. Mohler, Y. Li, Y. H. Song, J. Titterington and P. Delafontaine (2007). "IGF-1 reduces inflammatory responses, suppresses oxidative stress, and decreases atherosclerosis progression in ApoE-deficient mice." Arterioscler Thromb Vasc Biol 27(12): 2684-2690.

Tam, E., G. Rosenbluth, E. Rogers, D. Ferriero, D. Glidden, R. Goldstein, H. Glass, R. Piecuch and A. Barkovich (2011). "Cerebellar hemorrhage on MRI in preterm newborns associated with abnormal neurological outcome." J Pediatr 158(2): 245-250.

Tam, E. W., V. Chau, A. J. Barkovich, D. M. Ferriero, S. P. Miller, E. E. Rogers, R. E. Grunau, A. R. Synnes, D. Xu, J. Foong, R. Brant and S. M. Innis (2016). "Early postnatal docosahexaenoic acid levels and improved preterm brain development." Pediatr Res 79(5): 723-730.

Tam, E. W., V. Chau, D. M. Ferriero, A. J. Barkovich, K. J. Poskitt, C. Studholme, E. D. Fok, R. E. Grunau, D. V. Glidden and S. P. Miller (2011). "Preterm cerebellar growth impairment after postnatal exposure to glucocorticoids." Sci Transl Med 3(105): 105ra105.

Tam, E. W., D. M. Ferriero, D. Xu, J. I. Berman, D. B. Vigneron, A. J. Barkovich and S. P. Miller (2009). "Cerebellar development in the preterm neonate: effect of supratentorial brain injury." Pediatr Res.

Thomas, B., M. Eyssen, R. Peeters, G. Molenaers, P. Van Hecke, P. De Cock and S. Sunaert (2005). "Quantitative diffusion tensor imaging in cerebral palsy due to periventricular white matter injury." Brain 128(Pt 11): 2562-2577.

Thompson, D. K., D. Thai, C. E. Kelly, A. Leemans, J.-D. Tournier, M. J. Kean, K. J. Lee, T. E. Inder, L. W. Doyle, P. J. Andersone and R. W. Hunt (2014). "Alterations in the optic radiations of very preterm children—Perinatal predictors and relationships with visual outcomes." NeuroImage Clin 4: 145-153.

Tolsma, K. W., E. N. Allred, M. L. Chen, J. Duker, A. Leviton and O. Dammann (2011). "Neonatal bacteremia and retinopathy of prematurity: the ELGAN study." Arch Ophthalmol 129(12): 1555-1563.

Ure, A. M., K. Treyvaud, D. K. Thompson, L. Pascoe, G. Roberts, K. J. Lee, M. L. Seal, E. Northam, J. L. Cheong, R. W. Hunt, T. Inder, L. W. Doyle and P. J. Anderson (2016). "Neonatal brain abnormalities associated with autism spectrum disorder in children born very preterm." Autism Res 9(5): 543-552.

Van der Poll, T., F. van de Veerdonk, B. Scicluna and M. Netea (2017). "The immunopathology of sepsis and potential therapeutic targets." Immunology 17: 407-420.

Van der Ree, M., J. Tanis, K. V. Braeckel, A. Bos and E. Roze (2011). "Functional impairments at school age of preterm born children with late-onset sepsis." Early Human Development 87: 821-826.

- 130 -

van Haastert, I. C., F. Groenendaal, C. S. Uiterwaal, J. U. Termote, M. van der Heide-Jalving, M. J. Eijsermans, J. W. Gorter, P. J. Helders, M. J. Jongmans and L. S. de Vries (2011). "Decreasing incidence and severity of cerebral palsy in prematurely born children." J Pediatr 159(1): 86-91 e81.

Vesoulis, Z. A. and A. M. Mathur (2017). "Cerebral Autoregulation, Brain Injury, and the Transitioning Premature Infant." Front Pediatr 5: 64.

Vinall, J., S. P. Miller, B. H. Bjornson, K. P. Fitzpatrick, K. J. Poskitt, R. Brant, A. R. Synnes, I. L. Cepeda and R. E. Grunau (2014). "Invasive procedures in preterm children: brain and cognitive development at school age." Pediatrics 133(3): 412-421.

Viscardi, R., C. Muhumuza, A. Rodriguez, K. Fairchild, C. Sun, G. Gross, A. Campbell, P. Wilson, L. Hester and J. Hasday (2004). "Inflammatory Markers in Intrauterine and Fetal Blood and Cerebrospinal Fluid Compartments Are Associated with Adverse Pulmonary and Neurologic Outcomes in Preterm Infants." Pediatr Res 55: 1009-1017.

Vohr, B. R., W. C. Allan, M. Westerveld, K. C. Schneider, K. H. Katz, R. W. Makuch and L. R. Ment (2003). "School-age outcomes of very low birth weight infants in the indomethacin intraventricular hemorrhage prevention trial." Pediatrics 111(4 Pt 1): e340-346.

Volpe, J. (2009). "Cerebellum of the premature infant: rapidly developing, vulnerable, clinically important." Journal of Child Neurology 24(9): 1085-1104.

Volpe, J. J. (2008). "Postnatal sepsis, necrotizing entercolitis, and the critical role of systemic inflammation in white matter injury in premature infants." J Pediatr 153(2): 160-163.

Volpe, J. J. (2009). "Brain injury in premature infants: a complex amalgam of destructive and developmental disturbances." Lancet Neurol 8(1): 110-124.

Volpe, J. J. (2017). "Confusions in Nomenclature: "Periventricular Leukomalacia" and "White Matter Injury"-Identical, Distinct, or Overlapping?" Pediatr Neurol 73: 3-6.

Wang, L., Y. Tu, C. Huang and C. Ho (2012). "JNK signaling is the shared pathway linking neuroinflammation, blood-brain barrier disruption and oligodendroglialapoptosis in the white matter injury of the immature brain." J Neuroinflammation 9(175).

Wang, S., A. Kloth and A. Badura (2014). "The cerebellum, sensitive periods, and autism." Neuron 83: 518-532.

- 131 -

Wang, X., H. Hagberg, C. Nie, C. Zhu, T. Ikeda and C. Mallard (2007). "Dual role of intrauterine immune challenge on neonatal and adult brain vulnerability to hypoxia-ischemia." J Neuropathol Exp Neurol 66(6): 552-561.

Wang, X., L. Stridh, W. Li, J. Dean, A. Elmgren, L. Gan, K. Eriksson, H. Hagberg and C. Mallard (2009). "Lipopolysaccharide sensitizes neonatal hypoxic-ischemic brain injury in a MyD88-dependent manner." J Immunol 183(11): 7471-7477.

Wei, J. C., R. Catalano, J. Profit, J. B. Gould and H. C. Lee (2016). "Impact of antenatal steroids on intraventricular hemorrhage in very-low-birth weight infants." J Perinatol 36(5): 352-356.

Weindling, A. and C. Kissack (2001). "Blood Pressure and Tissue Oxygenation in the Newborn Baby at Risk of Brain Damage." Biol Neonate 79(3-4): 241-245.

Welker, K. M. and A. Patton (2012). "Assessment of normal myelination with magnetic resonance imaging." Semin Neurol 32(1): 15-28.

Wirtschafter, D. D., R. J. Powers, J. S. Pettit, H. C. Lee, W. J. Boscardin, M. Ahmad Subeh and J. B. Gould (2011). "Nosocomial infection reduction in VLBW infants with a statewide quality-improvement model." Pediatrics 127(3): 419-426.

Wisnowski, J., S. Bluml, L. Paquette, E. Zelinski, M. N. Jr, M. Painter, H. Damasio, F. Gilles and A. Panigrahy (2013). "Altered glutamatergic metabolism associated with punctate white matter lesions in preterm infants." PLoS One 8(2): e56880.

Won, S. P., H. C. Chou, W. S. Hsieh, C. Y. Chen, S. M. Huang, K. I. Tsou and P. N. Tsao (2004). "Handwashing program for the prevention of nosocomial infections in a neonatal intensive care unit." Infect Control Hosp Epidemiol 25(9): 742-746.

Wood, N. S., N. Marlow, K. Costeloe, A. T. Gibson and A. R. Wilkinson (2000). "Neurologic and developmental disability after extremely preterm birth. EPICure Study Group." N Engl J Med 343(6): 378-384.

Woodward, L. J., P. J. Anderson, N. C. Austin, K. Howard and T. E. Inder (2006). "Neonatal MRI to predict neurodevelopmental outcomes in preterm infants." N Engl J Med 355(7): 685-694.

Woodward, L. J., C. A. Clark, S. Bora and T. E. Inder (2012). "Neonatal white matter abnormalities an important predictor of neurocognitive outcome for very preterm children." PLoS One 7(12): e51879.

- 132 -

Wynn, J., P. Scumpia, R. Winfield, M. Delano, K. Kelly-Scumpia, T. Barker, R. Ungaro, O. Levy and L. Moldawer (2008). "Defective innate immunity predisposes murine neonates to poor sepsis outcome but is reversed by TLR agonists." Blood 112: 1750-1758.

Yanni, D., S. J. Korzeniewski, E. Allred, R. Fichorova, T. O'Shea, K. Kuban, O. Dammann and A. Leviton (2017). "Both antenatal and postnatal inflammation contribute information about the risk of brain damage in extremely preterm newborns." Pediatr Res 00: 1-6.

Yoshida, S., K. Hayakawa, A. Yamamoto, S. Okano, T. Kanda, Y. Yamori, N. Yoshida and H. Hirota (2010). "Quantitative diffusion tensor tractography of the motor and sensory tract in children with cerebral palsy." Dev Med Child Neurol 52(10): 935-940.

Zaidi, A., W. Huskins, D. Thaver, Z. Bhutta, Z. Abbas and D. Goldmann (2005). "Hospital-acquired neonatal infections in developing countries." The Lancet 365(9465): 1175-1188.

Zwicker, J. G., R. E. Grunau, E. Adams, V. Chau, R. Brant, K. J. Poskitt, A. Synnes and S. P. Miller (2013). "Score for neonatal acute physiology-II and neonatal pain predict corticospinal tract development in premature newborns." Pediatr Neurol 48(2): 123-129 e121.

Zwicker, J. G., S. P. Miller, R. E. Grunau, V. Chau, R. Brant, C. Studholme, M. Liu, A. Synnes, K. J. Poskitt, M. L. Stiver and E. W. Tam (2016). "Smaller cerebellar growth and poorer neurodevelopmental outcomes in very preterm infants exposed to neonatal morphine." J Pediatr 172: 81-87 e82.

Zwicker, J. G., S. W. Yoon, M. Mackay, J. Petrie-Thomas, M. Rogers and A. R. Synnes (2013). "Perinatal and neonatal predictors of developmental coordination disorder in very low birthweight children." Arch Dis Child 98(2): 118-122.