Extreme Prematurity: Risk and Resiliency

Abstract

Individuals born extremely preterm (before 28 weeks of gestation) comprise only about 0.7% of births in the United States and an even lower proportion in other high resource countries. However, these individuals account for a disproportionate number of children with cerebral palsy, intellectual deficit, autism spectrum disorder, attention deficit hyperactivity disorder, and epilepsy. This review describes two large multiple center cohorts comprised of individuals born extremely preterm: the EPICURE cohort, recruited 1995 in the United Kingdom and the Republic of Ireland, and the Extremely Low Gestational Age Newborn (ELGAN), recruited 2002–2004 in five states in the United States. The primary focus of these studies has been neurodevelopmental disorders, but also of interest are growth, respiratory illness, and parent- and self-reported global health and well-being. Both of these studies indicate that among individuals born extremely preterm the risks of most neurodevelopmental disorders are increased. Early life factors that contribute to this risk include perinatal brain damage, some of which can be identified using neonatal head ultrasound, bronchopulmonary dysplasia, and neonatal systemic inflammation. Prenatal factors, particularly the family’s socioeconomic position, also appear to contribute to risk. For most adverse outcomes, the risk is higher in males. Young adults born extremely preterm who have neurodevelopmental impairment, as compared to those without such impairment, rate their quality of life lower. However, young adults born extremely preterm who do not have neurodevelopmental impairments rate their quality of life as being similar to that of young adults born at term. Finally, we summarize the current state of interventions designed to improve the life course of extremely premature infants, with particular focus on efforts to prevent premature birth and on post-natal efforts to prevent adverse neurodevelopmental outcomes

Introduction: Public Health Significance of Extreme Prematurity

Extremely preterm birth, defined as birth before 28 weeks of gestation, occurs infrequently in the United States (about 0.7% of births) but contributes disproportionately to the prevalence of chronic health and developmental disorders among children. Of the approximately 28,000 extremely premature neonates born in the U.S. each year, about 75% survive into childhood.(1) In the perinatal period, extreme prematurity is associated with an increased risk of injury to nearly every organ system, but most prominently, to brain(2), eyes(3), lungs(4), heart(5), kidneys(6), and the gastrointestinal system(7). Consequently, chronic disorders of neurodevelopment, cardiopulmonary function, and renal function affect a larger proportion of children born extremely preterm as compared to those born at term. Whether this is true also for gastrointestinal disorders is not known.

Most consequential among prematurity-associated chronic disorders are the neurodevelopmental impairments, such as cerebral palsy and intellectual disability. Unfortunately, interventions to prevent neurodevelopmental impairments associated with extreme prematurity have had only a modest impact on the frequencies of these disorders.(8–10) This is true for efforts to prevent premature birth,(11) perinatal pharmacologic interventions to prevent the neonatal antecedents of brain damage,(12–15) and interventions to enrich social and learning environments.(16) While impressive progress has been made in reducing mortality risk over the past several decades,(17) surviving extremely premature neonates still face a very high risk of neurodevelopmental impairments.(18–21) In a large United States cohort of individuals born extremely prematurely, 11% had cerebral palsy, 25% had moderate to severe cognitive impairment,(21) 7% had autism spectrum disorders and 7.5% had epilepsy.(21) Thus individuals born extremely prematurely have about a 30-fold higher risk of cerebral palsy,(21, 22) a 9-fold higher risk of severe cognitive impairment,(20) a 7-fold higher risk of epilepsy,(23) and a 4-fold higher risk of autism spectrum disorders.(24)

Despite substantial research efforts, the causes of neurodevelopmental impairments and other chronic disorders among individuals born extremely prematurely remain unclear. Few studies have followed extremely premature children from birth into adolescence and few have collected the extensive biological and developmental data necessary to support rigorous investigations on early life etiologies, such as prenatal environmental exposures. This lack of understanding about etiology presents a critical barrier to prevention of neurodevelopmental impairments among individuals born extremely premature. It is likely that knowledge gained about the biological basis underlying chronic disorders in children born extremely prematurely is pertinent to term-born children, even though their prevalence of these disorders is lower.

In this review we will first summarize key findings from longitudinal studies of cohorts of individuals born extremely preterm, with a focus on the EPICURE study of neonates born in the United Kingdom and the ELGAN Study of neonates born in five states in the United States. We then will describe the limited number of evidence-based interventions that improve child outcomes after extremely preterm birth. We will conclude by describing several research themes that emerge from research focused on improving outcomes of extremely preterm neonates, including the use of biomarkers of developmental processes such as inflammation and epigenetic regulation of gene expression; the consideration of positive child health as an outcome; and the consideration of how the prenatal and postnatal environment, viewed broadly, could influence outcomes of individuals born extremely preterm.

Research Approaches to the Problem of Extreme Prematurity

Observational Studies

Observational studies, particularly those with longitudinal follow up, are a rich source of information about the prevalence of, and risk and resiliency factors for, chronic health and developmental disorders among individuals born extremely preterm. In the following sections we will summarize findings from two longitudinal studies of extremely preterm neonates – the EPICURE study and the ELGAN Study. Table 1, while not a comprehensive list of all assessments used in these studies, summarizes the methods used to evaluate cognition, behavior, and growth, as well as several other domains of function in which individuals born preterm, as compared to those born at term, have an increased risk for adverse outcomes. Table 2 compares the EPICURE and ELGAN cohorts with regard to sample characteristics and frequencies of some of the primary outcomes.

Table 1.

Follow up evaluations in the EPICURE and ELGAN cohort studies.

Domain	Assessment used	Selected citations
EPICURE - assessments at 30 months
Cognitive	Bayley Scales of Infant Development-Second Edition(58)	(26)
Motor	Standardized neurological examination, based on Amiel-Tison(207)	(26)
Vision and hearing	Clinician’s classification of functional ability	(26)
Behavior	Child Behavior Checklist	(41)
anthropometrics	Measurement of weight, length, head circumference	(26)
Respiratory system	Assessment by pediatrician	(42)
EPICURE - assessments at 6 years
Cognitive	Kaufman Assessment Battery for Children	(29)
Motor	3 items from the Movement Assessment Battery for Children: time to post 12 coins; heel-toe walking (15 steps), and 1-leg balance (> 20 seconds	(29)
Executive function	A Developmental NEuroPSYchological Assessment (NEPSY)	(32)
Academic achievement	Survey of child’s school teacher	(32)
Behavior and psychiatric outcomes	parent-report Strengths and Difficulties Questionnaire	(39, 41)
anthropometrics	Measurement of weight, length, head circumference	(34)
Respiratory system	Assessment by pediatrician; history based on core questions from the International Study of Asthma and Allergies in Childhood; peak expiratory flow measured with Wright peak flow meter	(42)
Blood pressure	Measurement of sitting blood pressure	(34)
EPICURE - assessments at 11 years
Cognitive	Kaufman Assessment Battery for Children	(29)
Behavior and psychiatric outcomes	parent-report Strengths and Difficulties Questionnaire, Social Communication Questionnaire (SCQ), structured psychiatric interveiw	(40, 41)
Anthropometrics	Measurement of weight and height	(38)
Respiratory system	history based questions from the International Study of Asthma and Allergies in Childhood; spirometry	(43)
Quality of life	Health Utilities Index Mark 3 multiattribute utility	(27)
EPICURE - assessments at 16 years
Behavior	parent-report Strengths and Difficulties Questionnaire	(41)
EPICURE - assessments at 19 years
Cognitive	Wechsler Abbreviated Scale of Intelligence, Second Edition	(208)
Behavior and psychiatric outcomes	Achenbach Adult Self-Report Scale (participant report), Clinical Interview Schedule-Revised (participant interview), and parent-report Strengths and Difficulties Questionnaire, Broader Autism Phenotype Questionnaire (BAPQ), the Empathy Quotient questionnaire, and the Frankfurt Test and Training of Facial Affect Recognition.	(41, 209)
Anthropometrics	Measurement of weight and height	(38)
Metabolic	Fasting blood glucose, blood lipids	(38)
Blood pressure	Measurement of sitting blood pressure	(38)
Quality of life	Health Utilities Index Mark 3 multiattribute utility	(27)
ELGAN - assessments at 12 months corrected age
Motor	Standardized neuromotor examination	(56)
ELGAN - assessments at 24 months corrected age
Cognitive	Bayley Scales of Infant Development - Second Edition	(60, 61)
Motor	Standardized neuromotor examination	(56, 75)
Behavior	Child Behavior Checklist, Modified Checklist for Autism in Toddlers	(76–78)
Anthropometrics	Measurement of weight, length, and head circumference	(210, 211)
ELGAN - assessments at 10 years of age
Cognitive	Differential Ability Scales, A Developmental NEuroPSYchological Assessment (NEPSY)	(63)
Motor	Gross Motor Function Classification System, Manual Ability Classification System	(193)
Academic achievement	Wechsler Individual Achievement Test-III	(20)
Behavior and psychiatric outcomes	Child Symptom Inventory, Social Communication Questionnaire, Social Responsiveness Scale, Autism Diagnostic Interview-Revised, Autism Diagnostic Observation Scales	(24, 80)
Anthropometrics	Measurement of weight, height, and head circumference	(85)
Epilepsy evaluation	Validated structured interview of parent or guardian	(23)
Quality of life	PedsQL (JW Varni)	(68)
ELGAN - assessments at 15 years of age
Cognitive	Wechsler Abbreviated Scale of Intelligence-II full-scale IQ (FSIQ) and the NIH Cognitive Toolbox Battery (NTCB) verbal/crystallized and fluid cognition measures
Academic achievement	Wechsler Individual Achievement Test-III
Behavior and psychiatric outcomes	Child Behavior Checklist, Mini International Neuropsychiatric Interview for Children and Adolescents, Behavior Rating Inventory of Executive Function
Anthropometrics	Measurement of weight, height, head circumference, waist circumference
Blood pressure	Measurement of sitting blood pressure
Epilepsy evaluation	Validated structured interview of parent or guardian
Brain structure	Magnetic resonance imaging

Table 2.

Characteristics and outcomes of EPICURE and ELGAN cohorts

	EPICURE	ELGAN
Characteristic
Location, years of recruitment	United Kingdom and Republic of Ireland, 1995	Massachusetts, Connecticut, Michigan, Illinois, North Carolina, 2002–2004
Number enrolled	811	1506
Gestational age	20–25 weeks	23–27 weeks
Outcomes
Number survived to NICU discharge (%)	314 (39%)	1200 (80%)
Age at Bayley and CP exam	30 months	24 months adjusted age
Cerebral palsy	12%	11.4%
Bayley MDI < 70	41%	25%
Age at ADHD and IQ assessment	11 years	10 years
IQ < 70	21%	17% verbal IQ; 15% non-verbal IQ
ADHD	11.5%	11.5% inattentive ADHD; 2.8% hyperactive ADHD; 5.5% combined ADHD
asthma	25%	38%
Factors associated with increased risk *
of cognitive impairment	Fewer years of maternal education, lack of antenatal glucocorticoid treatment of mother, premature rupture of membranes, male sex, bronchopulmonary dysplasia	Fewer years of maternal education, maternal obesity, minority race, male sex, small for gestational age, bronchopulmonary dysplasia, bacteremia, ultrasound-identified cerebral white matter injury
of motor impairment	Ultrasound-identified brain injury, longer exposure to postnatal steroids	Male sex, ultrasound-identified cerebral white matter injury

Abbreviations: ELGAN – Extremely Low Gestational Age; CP – cerebral palsy; MDI – Mental Development Index; ADHD – attention deficit hyperactivity disorder; IQ – intelligence quotient

^*bold text indicates risk factors found in both EPICURE and ELGAN

Observational Studies: The EPICURE Study

The EPICURE Study [website: https://www.ucl.ac.uk/womens-health/research/neonatology/epicure] is one of the most informative collaborative research ventures in the history of neonatology. The first phase of the study occurred in the 10 months of 1995 from March 1 to Dec 31, and remarkably, involved all 276 maternity units in the United Kingdom and the Republic of Ireland. Information was recorded for all deliveries between 20 and 25 weeks 6 days of gestation. For babies who were provided neonatal care, a full record of the hospitalization was collected. Of the 4004 births meeting enrollment criteria, 811 infants were admitted for neonatal care and 314 were discharged home (39% survival rate). Only two babies born at 22 weeks survived (9% of babies admitted for neonatal care). Survival among those admitted at later gestational ages was 26 babies (20%) at 23 weeks, 100 babies (34%) at 24 weeks, and 186 babies (52%) at 25 weeks.(25)

In addition to providing information about survival and neonatal morbidities among this group of extremely preterm individuals, EPICURE has completed follow up assessments of study participants at around 2.5, 6, 11, and 19 years of age. In addition, the Strength and Difficulties questionnaire was completed by parents when their child was around 16 years old. (See Table 1.)

One of the most important strengths of the EPICURE Study has been the inclusion of a control group of study participants who were born at term gestation. One of the most important challenges in EPICURE and other longitudinal studies in children is cohort attrition. Ninety-two percent (n=283) of surviving EPICURE participants participated in follow up evaluations at 2.5 years of age;(26) however fewer than half (n=129) of those who were evaluated at 2.5 years were also evaluated at 19 years.(27)

Cognitive Outcomes

At 30 months of adjusted age, the cohort was assessed with the Bayley Scales of Infant and Toddler Developmental-Second Edition (BSID-II) and a standardized neuromotor examination. The mean score on the BSID-II Mental Development Index (MDI), a measure of early cognitive functioning, was 84, slightly more than one standard deviation below the mean score in the normative reference sample, in which the mean score is 100 and the standard deviation is 15. Delays in early cognitive function, as indicated from BSID-II Mental Development Index scores, were associated with male sex, fewer years of maternal education, lack of antenatal steroid treatment of the mother, premature rupture of the membranes, and bronchopulmonary dysplasia (BPD).(28)

In a second wave of follow-up, at six years of age,(29) the rate of follow-up was 78% (n=241), lower than the rate at 30 months (92%). Fully 21% of the cohort (n=50) had cognitive impairment, defined as a score on the Kaufman Assessment Battery for Children more than two standard deviations below the mean in the standardization sample. Moreover, when compared to the distribution of intelligence quotient (IQ) scores among a group of classmates delivered at full term (the control group), 41% of the EPICURE cohort had cognitive impairment, defined as a score more than 2 standard deviations below the mean for the control group

In the most recent wave of follow up of the EPICURE cohort, at 19 years of age, young adults born extremely preterm scored lower on all neuropsychological tests, and 60% had an impairment in at least one neuropsychological domain. Most prevalent were deficits in general cognitive functioning and visuomotor abilities. Perhaps most concerning was the finding of an increase in the proportion of the EPICURE cohort with intellectual deficit.(30) This increase could not be attributed to bias introduced by sample attrition, since study participants who were evaluated at 11 years but not at 19 years were from lower socioeconomic backgrounds and had lower IQ and/or cognitive scores.

Motor Outcomes

At 30 months of age, 10% of the EPICURE sample had severe neuromotor impairment, 2% were blind or perceived only light, and 3% had uncorrectable hearing impairment or required hearing aids. Overall, 49% (95% confidence limits: 43–55%) had an impairment of neuromotor, sensory, or communication function, and 23% (95% confidence limits: 18–28%) had severe impairment.(26) Adverse motor outcomes were associated with ultrasound-identified brain injury and longer duration of postnatal corticosteroid treatment, an intervention used frequently for infants at high risk for developing BPD.(4, 31)

At six years of age, the rate of disabling cerebral palsy (12%) was similar to the rate of severe neuromotor impairment of 10% found at 30 months. Among children who had been identified as having severe disability at 30 months, 86% had moderate-to-severe impairment at 6 years, but milder forms of impairment at 30 months were not predictive of outcome at 6 years. Even among children without cerebral palsy, the EPICURE cohort, as compared with classmates born full-term, performed worse on tests of motor, visuospatial, and sensorimotor function, even when adjusting for overall cognitive scores. The functional impact of lower scores on tests of cognitive, visuospatial, motor, and attention-executive function lies in part in the associations of these scores with academic performance.(32)

Growth

At 30 months of age, children born extremely preterm birth had lower weight, height, head circumference, body mass index (BMI), and mid upper arm circumference as compared to population norms. Parent-reported feeding problems after discharge from neonatal intensive care (NICU), and longer time on supplemental oxygen (in nearly all cases given to treat hypoxemia due to BPD) were associated with worse growth through 30 months adjusted age.(33)

As was the case at 30 months, at six years, growth outcomes among children born extremely preterm were lower than normative data.(34) Extremely preterm children were 1.2 standard deviations (SD) lighter and 0.97 SD shorter, and had BMI that was 0.95 SD lower and head circumference that was 1.3 SD lower. The EPICURE cohort had exhibited “catch up” growth(35) between 30 months and 6 years, with an increase of 0.37 SD in weight, 0.42 SD in height, and 0.13 SD in head circumference. Related to growth, parent-reported eating problems were more prevalent among the extremely preterm children in EPICURE as compared to controls, and children with eating problems were shorter and lighter and had lower BMI at six years of age. (36) Continued tracking of growth in EPICURE participants indicated that at 19 years of age, extremely preterm individuals were shorter and lighter relative to controls, but had similar proportions with BMI > 25 kg/m² (23.4%) and > 30 kg/m² (6.3%). Pubertal development did not differ for the extremely preterm and control groups.(37) At 19 years, 8.7% of the extremely preterm group had metabolic syndrome. Compared to those without metabolic syndrome, these individuals had lower birth weights and greater increases in weight z-scores during the first 30 months of life. BMI and systolic blood pressure at 19 years were positively correlated to change in weight z-score between 2.5 and 6 6 years of age.(38)

Behavior and psychiatric outcomes

Behavioral outcomes at 6 years were based on questionnaires completed by parents and teachers. Boys more often had behavioral problems than girls, and both boys and girls born extremely preterm were more likely to have behavioral problems as compared to controls born at term.(39) The most frequently reported behavioral problems among EPICURE participants were hyperactivity (31%), and attention problems (33%), which were present, respectively, in 9% and 7% of classmate controls not born preterm.

At the third wave of follow up of the EPICURE cohort, at age 11 years, extremely preterm children were more than three times as likely to have a psychiatric disorder than controls, with nearly one fourth of the sample affected. The most prevalent psychiatric disorder was attention deficit/hyperactivity disorder (ADHD; 11.5%). Extremely preterm children, relative to controls, had 10-fold the odds of ADHD inattentive subtype but were not more likely to have ADHD combined subtype. Extremely preterm children also had an increased likelihood of anxiety disorders and autism spectrum disorders; the latter was found in 8% of extremely preterm children. Risk factors for psychiatric disorders after extremely preterm birth included both cognitive impairment and early traumatic experiences encountered by the child and parent are potential contributors to risk.(40)

The longitudinal nature of the EPICURE study, with evaluations at 6, 11, 16, and 19 years of age, has allowed investigators to evaluate trajectories of behavioral, attention, social, and emotional problems from childhood to young adulthood. These outcomes have been assessed in both the EPICURE cohort and full-term controls using the parent-reported Strengths and Difficulties questionnaire. Important findings from the evaluation of trajectories of behavioral outcomes between 6 and 19 years : 1) the extremely preterm group had higher scores for total difficulties; 2) these behavioral difficulties had a larger impact on home life, friendships, and activities among the group born extremely preterm; 3) hyperactivity/inattention and peer problems contributed the most to difference between extremely preterm and controls groups; and 4) a clinically significant behavioral screen at 30 months and cognitive impairment were predictive of more behavioral difficulties between age 6 and 19 years.(41)

Respiratory and Cardiovascular Outcomes

Besides growth, the other health outcomes assessed at six years were respiratory disorders, blood pressure, and a measure of arterial stiffness. Children born extremely preterm were twice as likely to have been diagnosed with asthma. The subset of those born extremely preterm with a history of BPD were two to three times more likely than those without a history of BPD to have wheezed in the past year, have wheeze or cough at night, or have used steroids or bronchodilators. However, BPD was not associated with having been diagnosed with asthma. In children born extremely preterm, exposure to tobacco smoke was associated with an increased likelihood of wheeze.(42) A more comprehensive assessment of respiratory outcomes at 11 years of age revealed that 56% of children born extremely preterm had abnormal baseline spirometry and 25% had been diagnosed with asthma, a rate twice that of controls born at term. Even in the subset with no respiratory symptoms over the previous 12 months, 48% had abnormal baseline spirometry, supporting the need for continued monitoring of respiratory outcomes throughout childhood and, probably, beyond. (43) A measure of arterial stiffness, but not blood pressure, was increased among children born extremely preterm children as compared to controls, and arterial stiffness increased with decreasing forced expiratory volume in 1 second.(44)

Quality of Life

Based on the foregoing information about outcomes after extremely preterm birth, it is not surprising that when the EPICURE cohort was assessed for quality of life at 19 years of age, their scores were lower than those reported by controls born at term. In the subset of EP individuals without neurodevelopmental impairment, the median quality of life score was 0.91, as compared to 0.97 among controls. (A difference of 0.03 to 0.05 is considered clinically significant.) Adults born extremely preterm who had developmental impairments, as compared to those without such impairments, rated their quality of life lower. Parents rated their children’s quality of life lower than did the children themselves. Most concerning was the finding that the quality of life for this group of young adults born extremely preterm decreased between ages 11 and 19. The EPICURE investigators speculated that this decline in quality of life might be due to difficulty transitioning to adulthood.(27)

Observational Studies: The ELGAN Study

The ELGAN Study was designed by Alan Leviton, assisted by Nigel Paneth, Karl Kuban, and Michael O’Shea. The over-arching hypothesis of the ELGAN Study is that perinatal inflammation contributes to brain and related disorders among individuals born extremely preterm.(45–52) To address this hypothesis, emphasis was placed on collection of data about potential perinatal initiators of inflammation and biological specimens (placenta, maternal blood, and neonatal blood) that could be evaluated for biomarkers of inflammation, such as white blood cell infiltration of the umbilical cord and increased levels of blood proteins indicative of an inflammatory response, such as C-reactive protein, interleukins, and tumor necrosis factor-α.(1) Related to potential initiators of inflammation, data were collected about indicators of mother’s socioeconomic position (e.g., maternal education and health insurance); chronic illnesses in the mother (e.g., obesity, diabetes, asthma); and acute maternal infections. Biomarkers of inflammation included histological evidence of placental inflammation and levels of inflammation-related proteins (e.g., C-reactive protein) in neonatal dried blood spots collected on postnatal days 1, 7, 14, 21, and 28.(53, 54)

In the first phase of ELGAN (Table 1), 1506 neonates born before 28 weeks of gestation at 14 hospitals in the United States were enrolled at birth and followed through 24 months adjusted age. In the first phase of the ELGAN Study (ELGAN-1), special efforts were directed towards collecting highly reliable data about brain-related outcomes, including ultrasound-identified brain injury(55) and neurodevelopmental outcomes at 24 months of adjusted age,(56, 57) when a valid diagnosis of cerebral palsy can almost always be made. The assessment at 24 months also included the BSID-II; a screening test for autism, the Modified Checklist for Autism in Childhood (MCHAT), and an assessment of problem behaviors based on parent report, the Child Behavior Checklist (CBCL). Of the 1200 infants who were alive at 24 months (79.7%), 1056 (88%) were evaluated at around 24 months of corrected age.

In the second phase of ELGAN, funding was provided in order to evaluate the relationship of neonatal systemic inflammation to cognitive outcomes; thus the only ELGAN study participants who were recruited for this phase were those from whom data were available about neonatal systemic inflammation (n=966). At ten years of age, 889 (92% of those recruited) returned for evaluation of cognitive function, behavioral problems, autism spectrum disorder, epilepsy, and anthropometric outcomes, as well as a pilot study of about 20% of study participants, using brain magnetic resonance imaging (MRI).

In the third phase of ELGAN, named ELGAN-ECHO because this phase was made possible through a collaboration with the Environmental Influences on Child Health Outcomes (ECHO) Study, when participants were 15 years old, the focus was broader to include assessment of the child’s physical and psychosocial environment, cognition, behavior and psychiatric outcomes, epilepsy, brain structure assessed with MRI, sleep health, and positive health. The last of these refers to a construct that is relatively novel in studies of child health outcomes, and we will expand on this construct later in this paper.

Cognitive Outcomes

At 24 months corrected age, 84% of survivors were assessed with the BSID-II, a measure of early cognitive function that is somewhat predictive of a child’s IQ.(58) The mean BSID-II MDI was 83, almost the same as the mean MDI in the EPICURE cohort; 25% had MDI < 70 (2 standard deviations below the normative mean) and 15% had a MDI < 55 (3 standard deviations below the normative mean) (unpublished data). As hypothesized, neonatal systemic inflammation, particularly if present on multiple days in the first two postnatal weeks, was associated with BSID-II MDI < 55, which is 3 standard deviations below the mean in the reference sample(59) (Table 2). Antenatal factors associated, in multivariate analyses, with an MDI < 55 were male sex, maternal obesity, mother having less than 12 years of formal education,(60) and minority race, while postnatal factors associated with MDI < 55 included ultrasound-identified cerebral white matter injury,(61) and necrotizing enterocolitis requiring surgical intervention and accompanied by bacteremia.(62)

At ten years of age cognition was assessed with an IQ test, the Differential Ability Scales-II (DAS-II), and measures of executive function. Verbal and non-verbal IQ, along with two subtests from the.DAS-II and 5 subtests from A Developmental NEuroPSYchological Assessment (NEPSY) were used to identify four groups of ELGAN participants with similar cognitive function profiles.(63) Overall, 17% of ELGAN participants were classified as having a “moderately impaired” cognition profile, with IQ and executive function test scores between 1.5 and 2.5 standard deviations below the normative mean on all 9 measures used to derive the cognitive profiles, and 8% were classified as having a “severely impaired” profile, with scores 3 to 4 standard deviations below normative values. Antenatal factors associated with worse cognitive outcome at ten years included maternal pre-pregnancy obesity,(64) mother’s having no more than high school education,(65) male sex,(21) and fetal growth restriction. Postnatal factors associated with worse cognitive outcome were small-for-gestational age,(66) neonatal systemic inflammation,(67) bronchopulmonary dysplasia,(68) bacteremia,(69) and ultrasound-identified cerebral white matter injury.(70) The presence of Lactobacillus in the placenta(71) and elevated levels of neurotrophic proteins(72) were associated with better cognitive outcome. In the subset of study participants who underwent brain MRI, higher grey matter, white matter, and cerebellum and brainstem volumes were significantly correlated with higher IQ.(73)

Motor Outcomes

In the first phase of ELGAN, the functional outcome that was of primary interest was cerebral palsy; thus extensive efforts directed towards collecting highly reliable data about neurological abnormalities and the diagnosis of cerebral palsy at 24 months of adjusted age.(56, 57) At 24 months adjusted age, the overall prevalence of cerebral palsy in the 1056 children who were examined (88% of survivors), 11.4% (120) were classified as having cerebral palsy; of these children, 31% had diparesis, 17% had hemiparesis, and 52% had quadriparesis. This overall frequency of cerebral palsy is very similar to what was found in the EPICURE cohort at 30 months and 6 years of age. Neonatal systemic inflammation, particularly if present on multiple days in the first two weeks, was associated with an increased risk of cerebral palsy and also an increased risk of ultrasound indicators of neonatal brain injury,(74) which are predictive of cerebral palsy.(75) Male sex and severe bronchopulmonary dysplasia, defined as the requirement for mechanical ventilation at 36 weeks postmenstrual age, were associated with an increased risk of cerebral palsy. (Table 2)

Behavior and Psychiatric Outcomes

In the first wave of follow up of ELGAN participants, behavior was assessed with the CBCL, a broad screen for problem behaviors, and the MCHAT, a screen for autism spectrum disorder. Based on parent report on the CBCL, 10.7% of ELGAN participants had an attention problem, defined as a score above the 93^rd percentile on the empirically based attention problem scale, and 8.1% for the attention Diagnostic and Statistical Manual of Mental Health Disorders-oriented scale.(76) Antenatal factors associated with an increased risk of attention problems at 24 months were no maternal formal education beyond high school, secondhand tobacco smoke exposure during pregnancy, and culture of two or more microorganisms from the placenta.(76) In common with cognitive and motor impairments, attention problems at 24 months was associated with neonatal systemic inflammation.(77) Based on the MCHAT, 22% (230/1040) ELGAN participants screened positive for increased autism risk. However, the frequency of a positive screen varied greatly across subgroups with or without a neurodevelopmental impairment. For example, 76% of individuals with quadriparetic cerebral palsy screened positive on the MCHAT, as compared to 18% of individuals without cerebral palsy.(78) Further, for subgroups with BSID-II < 55, 55–69, ≥70, the frequency of positive MCHAT was 62%, 35%, and 11%. These findings suggest that among individuals born extremely preterm, neurodevelopmental impairments could led to false positive M-CHAT screening.

In the second phase of ELGAN, behavioral and psychiatric outcomes were assessed with a rigorous evaluation for autism, which included initial screening with the Social Communication Questionnaire (SCQ), followed by the Autism Diagnostic Interview-Revised (ADI-R) for those with scores ≥11, and the Autism Diagnostic Observation Schedule-2 for those who met ADI-R criteria for autism. Based on this approach, 7% of the ELGAN cohort was diagnosed with autism spectrum disorder (ASD), a frequency about 4 times higher than found in school-age children in the United States.(24) Risk for ASD varied from 15% among those born at 23–24 weeks to 3.4% among those born at 27 weeks. Males and children with a history of being small-for-gestational age were at increased risk for ASD,(24) as were children whose neonatal ultrasound showed persistent ventricular enlargement, an indicator of cerebral white matter injury.(70) As was true for other neurodevelopmental impairments, increased risk of ASD was associated with neonatal systemic inflammation.(79)

In addition to the autism evaluation, the second phase of ELGAN included a screen for psychiatric symptoms, the Child Symptom Inventory Fourth Edition (CSI-4). ELGAN participants were more likely to screen positive for several psychiatric disorders as compared with normative samples, with 15% screening positive for 1 disorder, 7% for 2, 3% for 3, and 4% for 4 or more disorders. A positive screen for inattentive ADHD was found in 11.5% of the ELGAN cohort, as compared to 3.6% of a normative sample; hyperactive ADHD was found in 2.8% versus 2% in the normative sample; and combined ADHD was found in 5.5% as compared to 1.6% in the normative sample. After ADHD, the next most frequent CSI-4 positive screen was for general anxiety, which was found in 5.3% of the ELGAN cohort, as compared to 1.6% of the normative sample.(80) Parent-reported ADHD at 10 years was associated with maternal obesity, mother’s age < 21, maternal smoking during pregnancy, male sex, persistent ventricular enlargement (an ultrasound indicator of cerebral white matter injury), and severe retinopathy of prematurity.(81)

Epilepsy

At 10 years ELGAN investigators used a validated approach to diagnose epilepsy that consisted of a structured screening interview by a research assistant, which, if positive, was followed by a structured interview by a pediatric neurologist. Using this approach, 7% of the ELGAN cohort was diagnosed with epilepsy,(23) about 4 times higher than in the general population of school-aged children in the United States. Risk factors for epilepsy in the ELGAN cohort include intraventricular hemorrhage,(70) maternal eligibility for Medicaid, an indicator of lower household income, maternal fever during pregnancy, ultrasound-identified cerebral white matter injury, and severe bronchopulmonary dysplasia, defined as the requirement for mechanical ventilation at 36 weeks postmenstrual age.(82) Epilepsy was the only neurodevelopmental impairment associated with tobacco smoke exposure during pregnancy.(83)

Respiratory and Cardiovascular Outcomes

In addition to neurodevelopmental impairments, analyses of risk factors have been completed for asthma and overweight/obese at ten years of age, chronic health conditions present in 38%(84) and 24%(85), respectively, of the ELGAN participants evaluated at ten years of age. An unexpected finding was that bronchopulmonary dysplasia was not associated with asthma at ten years of age; whereas mother’s eligibility for Medicaid, an indicator of household income, was associated with asthma.(84) Similarly, unmarried mother, tobacco smoke exposure during pregnancy, and maternal obesity, all of which are more prevalent among mothers with indicators of socioeconomic disadvantage (Medicaid eligibility, eligibility for food stamps, fewer years of maternal education, and single marital status), were associated with overweight and/or obesity at ten years of age. Thus, social disadvantage appears to be a risk factor for at least two highly prevalent chronic health conditions (asthma and obesity) among individuals born extremely preterm.

Quality of Life and Positive Child Health

Parents completed the Pediatric Quality of Life Inventory at 10-year study visits.(86) Children with a history of BPD, a history of fetal growth restriction, or who screened positive for at least two psychiatric disorders at 10 years old, had lower quality of life compared to other children born extremely preterm.(66, 68, 80) Quality of life was also strongly associated with the Positive Child Health Index. (87) The Positive Child Health Index quantifies the number of developmental or medical disorders for each child. Therefore, children within fewer developmental or medical disorders had higher parent reported quality of life.

The Positive Child Health Index was developed to investigate the broad concept of positive child health. Health is defined by the Institute of Medicine as children’s ability and potential to: ”a) develop and realize their potential; b) satisfy their needs; and c) develop the capacities that allow them to interact successfully with their biological, physical, and social environments”. (88) A lower Positive Child Health Index at 10 years old, indicating more developmental or medical disorders, was associated with indicators of lower maternal socioeconomic status, maternal obesity, and maternal asthma. (89) Ongoing work with the ELGAN cohort include measurement of the Positive Child Health Index at older ages as well as parent or child reported well-being measures. Well-being measures, a complementary strategy to measure positive child health, have been developed through the Patient-Reported Outcomes Measurement Information System (PROMIS) to quantify parent or child reported perception of general health, meaning and purpose, or life satisfaction.(90, 91)

Summary of Themes from Observational Studies of Extremely Preterm Infants

In this section we will draw from both the EPICURE and ELGAN studies, as well as other epidemiologic studies, to synthesize information on 5 themes that are particularly relevant to the risk of chronic health disorders among individuals born extremely preterm: socioeconomic disadvantage, maternal health disorders, male sex, neonatal morbidities, perinatal inflammation, and 2 innovative research approaches that hold promise for advancing our understanding of risk and resiliency among individuals born extremely preterm: biomarkers of developmental processes and multi-dimensional outcomes.

Socioeconomic Disadvantage

As is the case among children born at term, among those born extremely preterm socioeconomic disadvantage is consistently associated with worse neurodevelopmental(60, 65, 92–94) and health outcomes(84, 85, 95). Likely mediators between socioeconomic factors and adverse child health outcome include less optimal nutrition,(96) increased maternal stress,(97) environmental pollution,(98–100) structural racism,(97, 101, 102) less optimal learning environments,(94, 103, 104) and less access to health care.(105, 106) In the ELGAN cohort, social disadvantage has been associated with neonatal systemic inflammation(107) and with epigenetic changes in the placenta,(108) suggesting two possible mediating pathways for the relationship between socioeconomic disadvantage and adverse outcomes in the offspring.(109)

Maternal Health

Maternal health disorders, such asthma, obesity, and diabetes, have been associated with a variety of adverse outcomes among individuals born preterm.(64, 89) Some of these conditions are associated with systemic inflammation,(110) raising possibility that inflammation mediates the relationship between maternal health and outcomes in the offspring. (111) Maternal health disorders occur more often among mothers with socioeconomic disadvantage which, as discussed above, is a risk factor for worse outcomes among children born extremely preterm.

Sex Differences in Outcomes

In both the EPICURE(28) and ELGAN Studies(21) adverse outcomes occurred more frequently in boys. An understanding of the biological mechanisms underlying this association could point towards biological targets for interventions to improve outcome after extremely preterm birth. One possibility is that the heightened inflammatory response to inflammatory initiators in male neonates might contribute to this group’s higher risk of neurodevelopmental impairments.(112) It is also possible that sexual dimorphism in the placenta contributes to sex differences in child outcomes.(113)

Neonatal Morbidities

The neonatal morbidity that has most consistently been associated with adverse neurodevelopmental outcomes is ultrasound-identified cerebral white matter injury.(70, 114–122) Less consistently, these adverse outcomes have been associated with severe bronchopulmonary dysplasia,(123, 124) necrotizing enterocolitis requiring surgery,(62, 125) sepsis,(69, 126) and severe retinopathy of prematurity.(127–129) These neonatal morbidities co-occur so frequently that it is challenging to identify effects of individual factors.(130) Most of these neonatal morbidities occur more often in male infants, so neonatal morbidities could mediate some of the sex differences in outcomes, discussed above.(131) Each of these morbidities has been associated with neonatal systemic inflammation,(132–137) suggesting a potential mechanism by which these neonatal morbidities, which often resolve during infancy, could be related to long-term disruption of brain development and neurological disorders later in life.(138–140)

Biomarkers of Perinatal Inflammation

In an epidemiological study published more than 45 years ago, findings suggested that perinatal infection outside of the brain was strongly associated with cerebral white matter injury.(141) In preclinical models (e.g., newborn kittens,(142) neonatal rats,(143) fetal sheep,(144) and preterm non-human primates(145)), injection of endotoxin, which initiates a systemic inflammatory response, can lead to neuroinflammation and disruption of cerebral white matter.(140, 142, 144, 146–148) These findings provide a plausible mechanism for the association between chorioamnionitis during pregnancy and cerebral palsy in the offspring.(149, 150) Among individuals born extremely preterm, elevated levels of biomarkers of inflammation, such as C-reactive protein and specific interleukins, have been associated with an increased risk of cerebral palsy,(151–153) cognitive impairment,(67) microcephaly,(154) lower brain volume on MRI,(73) ADHD,(155) and autism spectrum disorder.(79)

Neonatal systemic inflammation could play a central role in the etiology of adverse neurodevelopmental outcomes among individuals born extremely preterm. (See Figure 2.) As reviewed above, maternal health conditions, such as pre-pregnancy obesity, have been associated with both neonatal systemic inflammation and neurodevelopmental impairment in the offspring. Similarly, neonatal complications, such as bronchopulmonary dysplasia, are associated with both neonatal inflammation and adverse neurodevelopmental outcomes, suggesting the possibility that neonatal systemic inflammation mediates at least some of the putative effect of bronchopulmonary dysplasia on neurodevelopment. Also, inflammation could mediate, at least in part, sex differences observed between the outcomes of females and males born extremely preterm. Finally, specific environmental factors, such as air pollution,(99, 156) arsenic,(157, 158) and maternal psychosocial stress,(109, 159) could initiate systemic and/or neuroinflammation and provide a mechanistic link between environmental factors and adverse neurodevelopmental outcomes.

Figure 2.

Conceptual model for relationships among prenatal initiators of preterm birth, neonatal complications of extreme prematurity, neonatal systemic inflammation, and quality of life for individuals born preterm. Examples of environmental exposures that contribute to preterm birth include air pollution and racism. Examples of specific maternal conditions that contribute to preterm birth include severe preeclampsia and placenta abrutption.

Based on findings from preclinical models, prenatal intrauterine exposure to inflammation can alter the offspring’s response to postnatal initiators of inflammation. A prior exposure to inflammation that results in dampening of the response to a subsequent exposure is referred to as a pre-conditioning exposure;(148, 160) a prior exposure that results in a heightened response to a subsequent exposure is referred to as a sensitizing exposure.(146, 161) The presumed modification by an earlier occurring factor of a subsequent factor’s effect has been referred to as the “two-hit” model of pathogenesis. Two studies from the ELGAN cohort provide some preliminary support for this model. In one, the association between small-for-gestational age (an indicator of fetal growth restriction) and cognitive impairment was especially predominant among neonates with a postnatal systemic inflammatory response, as though the prenatal exposure to fetal growth restriction sensitized the individual to the later exposure to inflammation.(162) In the other, infants whose placenta had evidence of inflammation and whose blood contained elevated levels of inflammatory biomarkers were at higher risk for neurodevelopment impairment than infants with only placental inflammation or only neonatal systemic inflammation.(163)

In addition to pre-conditioning and sensitization, the strength of association between neonatal systemic inflammation and neurodevelopmental outcomes might be influenced by the levels of “endogenous protectors”,(51) such as anti-inflammatory cytokines,(164) neurotrophins,(155, 165, 166) growth factors,(167) and angiogenic factors.(168) Preliminary support for a protective role of this group of proteins in extremely preterm neonates comes from several analyses of data from the ELGAN cohort, in which higher levels of one or more putative endogenous protectors seem to decrease the strength of association between elevations of inflammation-related proteins and neurodevelopmental impairment.(79, 155, 169)

Biomarkers of Developmental Processes

In the past several decades observational research has made use of biomarkers to provide more precise measures of exposures and biological processes that underlie adverse outcomes among individuals born extremely preterm. Biomarker studies have been revolutionized by the development of methods using specimens of blood archived as dried blood spots to measure proteins (e.g., Hemoglobin A1c, cystatin C, and C-reactive protein), ,(170, 171) viruses,(172) environmental pollutants,(173) fetal nutrition,(174, 175) and genetic factors such as deoxyribonucleic acid (DNA) CpG methylation,(176) messenger ribonucleic acid (mRNA),(177) and DNA mitochondrial copy number(178). In the ELGAN Study, initial biomarker studies focused on protein biomarkers of inflammation; findings from these studies have been summarized above.(59, 67, 72–74, 77, 79, 107, 132, 152, 154, 162, 169, 179–181) More recent studies have utilized archived placenta specimens of ELGAN participants to assay DNA CpG methylation, a marker of epigenetic control of gene expression. One objective of measuring epigenetic variation in placenta is to evaluate the potential for environmental exposures during pregnancy to influence fetal development and health of the offspring. For example, microorganisms in placenta were associated with differential methylation of over 1000 genes, including proteins that are involved in immune/inflammatory responses,(182) which, as described above, could influence the risk of neurodevelopmental impairment among individuals born extremely preterm.(183) Indicators of social disadvantage, such as fewer years of maternal education, have been associated with differential CpG methylation in placenta of ELGAN participants, suggesting a potential role for the placenta in transducing the effects of socioeconomic environment on the fetus.(108) Finally, the placenta CpG methylation profile has been associated with both cognitive(71) and behavioral outcomes in the ELGAN cohort.(184) Current biomarker studies in the ELGAN cohort are utilizing data on environmental metals, assayed in archived umbilical cords, and CpG DNA methylation, mRNA, and proteins, measured in both placenta and neonatal blood.

Interventions to Improve the Life Course of Individuals Born Preterm

Prevention of Preterm Birth

Primary prevention of the adverse outcomes that occur frequently among individuals born extremely preterm involves prevention of extremely preterm delivery. Unfortunately, few interventions have been demonstrated, through rigorous clinical trials, to prevent extremely preterm birth. A 2018 overview of Cochrane Systematic Reviews of interventions during pregnancy to prevent preterm birth concluded that systematic reviews reported “clear evidence of benefit to prevent specific populations of pregnant women from giving birth early, including midwife-led continuity models of care versus other models of care for all women; screening for lower genital tract infections for pregnant women less than 37 weeks’ gestation and without signs of labor, bleeding or infection; and zinc supplementation for pregnant women without systemic illness. Cervical cerclage (a stitch around the cervix) showed clear benefit for women with singleton pregnancy and high risk of preterm birth only.”(185) In women who present early in pregnancy with cervical dilatation, there is limited evidence, from small observational studies, that emergency cervical cerclage may prolong pregnancy, but no data are available about long term outcomes in the mother and newborn.(186)

One of the most effective approaches to prevention of preterm birth is weekly intramuscular injections of 17-hyroxyprogesterone caproate (17-OHPC) beginning between 16–21 weeks gestation. In a meta-analysis of 6 randomized trials, 17-OHPC treatment of mothers with a history of prior preterm delivery reduced the risk of very preterm (before 32 weeks of gestation) delivery by 32%, and reduced the risk of neonatal death by 59%.(187) Among women at increased risk for preeclampsia, such as women with chronic hypertension, obesity, African-American ancestry, age over 40 years, family history of preeclampsia, or prior history of preeclampsia, low dose aspirin (81 milligrams/day), initiated after 12 weeks of gestation, lowers the odds of preterm birth by 20%.(188) In a meta-analysis, treatment of mothers with gestational diabetes, which largely focuses on management of blood glucose levels, was associated with a 25% reduction in the odds of preterm birth, although this effect was not quite statistically significant (odds ratio: 0.75; 95% confidence limits: 0.56, 1.01).(189) Promising, but underdeveloped, approaches to prevention of preterm birth include modifications to the maternal diet during pregnancy(190) and decreasing exposures to environmental stressors, such as air pollution and psychosocial stress.(191, 192)

Prevention of Neurodevelopmental Impairment among Individuals Born Extremely Preterm

Despite the introduction of some effective preventive prenatal interventions, extremely preterm birth occurs in about 0.7% of United States births, and approximately 75% of extremely preterm neonates survive each year, representing some 21,000 individuals. Since about 6700 (32%) will develop a neurodevelopmental impairment(193), prevention of neurodevelopmental impairments among individuals born extremely preterm should be a public health priority. What follows below is a brief summary of the most important prenatal and postnatal interventions that improve outcome of pregnancies complicated by preterm delivery, including both pharmacological and non-pharmacological therapies as well as hospital-based and home-based interventions.

When an expectant mother develops a pregnancy complication that is likely to lead to preterm delivery, such as preterm labor, preterm premature rupture of membranes, severe preeclampsia, suspected placenta abruption, or severe intrauterine growth restriction), prompt antenatal treatment with magnesium sulfate is an evidence-based intervention to decrease the risk of cerebral palsy in the offspring,(194) and antenatal treatment with betamethasone or dexamethasone reduces the risk of developmental delay.(195)

Postnatal approaches to prevention of neurodevelopmental impairment have frequently used the strategy of preventing neonatal morbidities, such as intraventricular hemorrhage and bronchopulmonary dysplasia, which have been associated with neurodevelopmental impairments later in childhood(127, 128) and are presumed to be causal risk factors for these impairments. Among neonates born extremely low birth weight (a group which overlaps highly with neonates born extremely preterm), indomethacin prevents severe intraventricular hemorrhage, but does not have clear benefit on long-term neurodevelopmental outcome.(10) In specific subgroups of extremely preterm infants, caffeine,(15) vitamin A,(196) and postnatal glucocorticoids are evidence-based practices to reduce the risk of bronchopulmonary dysplasia. However, there is little evidence that these interventions reduce the risk of neurodevelopmental impairment.(196–198) When assessed in a large multi-center trial, high dose erythropoietin, a neuroprotective strategy that showed great promise in preclinical models, was not associated with improved neurodevelopmental outcomes.(199) Similarly, conclusive evidence is lacking that nutritional interventions, such as docosahexaenoic acid, decrease the risk of neurodevelopmental impairment.(200, 201)

A variety of non-pharmacological interventions in the NICU have been designed to support the development of preterm infants, including infant massage, individualized modifications of the social and physical environment during neonatal intensive care, and supportive interventions to enhance parents’ well-being and parenting skills. It is not clear which interventions might decrease the risk of neurodevelopmental impairment because few large multi-center randomized trials of these interventions have been conducted in which neurodevelopmental outcomes were evaluated as the primary outcome. However, one of these interventions deserves comment: Newborn Individualized Developmental Care and Assessment Program (NIDCAP) for organized developmentally appropriate, family-centered care of fragile premature newborns.(202) This intervention is based largely on pioneering research by Dr. Heidelise Als focused on using each individual preterm infant’s behavioral cues to modify the environment and care practices to which that infant is exposed in the neonatal intensive care unit, with the goal of “supporting and enhancing each infant’s strengths and self-regulation capacities.(203) NIDCAP is a multi-faceted approach with involvement of health professionals and parents. Despite evidence suggestive of beneficial effects of NIDCAP, a meta-analysis involving 627 individuals concluded that NIDCAP did not improve neurodevelopmental outcome.(204)

A wide variety of post-hospital programs have been evaluated in randomized trials.(205) Most have involved less than 200 study participants; an exception is the multi-center Infant Health and Development Program (IHDP), which randomized 985 infants at increased risk for developmental impairment because of preterm birth or low birth weight. Three hundred and seventy-seven were randomized to an intensive early intervention program and 608 infants were randomized to standard care.(8, 16) Children randomized to the intervention group were visited in their home by professionals with expertise in child development; visits were weekly for the first year and biweekly for years 2 and 3. Home visitors provided parents with games to use with their child and provided guidance on the child’s cognitive, language and social development and management of problems identified by the parent. In addition, children in the intervention group attended child development centers 5 days per week in the second and third year of life. Due to its relatively large sample size, the IHDP contributed a majority of individuals to a systematic review of post-hospital early interventions to improve developmental outcome, and the findings from that review are consistent with those from the IHDP. Specifically, early intervention programs for preterm infants have a positive influence on cognitive and motor outcomes during infancy, with cognitive benefits persisting into preschool age. However, these benefits were largely attenuated at school age, emphasizing the need for interventions with sustained effects on the developmental outcomes of individuals born preterm.(8)

Summary: Interventions to Improve the Life Course of Individuals Born Preterm

Over the past half-century, multi-center randomized trials have identified several prenatal interventions to improve outcome for individuals born extremely preterm. In women with a history of prior preterm birth, 17- OHPC reduces the risk of recurrent preterm birth. In mothers for whom preterm delivery is expected, betamethasone and magnesium sulfate improve neurodevelopmental outcome. Among infants born extremely preterm, neurodevelopmental outcome interventions that appear to hold the greatest promise are those that support neurodevelopment by enhancing parents’ abilities to foster developmental progress in their child.

Summary: Extreme prematurity: risk and resiliency

Longitudinal cohort studies performed over the past three decades indicate that extreme prematurity is an important cause of neurodevelopmental impairments, psychiatric disorders, asthma, cardiovascular, and renal disease among children. These observational studies have identified both risk factors as well as resiliency factors that influence the likelihood of chronic disorders that occur more frequently among individuals born extremely preterm, as compared to those born at term. Randomized trials have confirmed the effectiveness of prenatal pharmacological interventions for decreasing the risk of preterm delivery and pre- and post-natal interventions for decreasing the risk of adverse outcomes when preterm delivery occurs. Despite the widespread availability of these preventive measures, at least in high-income countries, rates of neurodevelopmental impairments remain unacceptably high.

In 2006, a decade after the initiation of the EPICURE Study, the EPICURE investigators launched EPICURE2 and recruited another population-based sample of extremely preterm newborns. To this point in time, comparing outcomes in EPICURE2 to EPICURE has indicated that while survival had improved, cognitive and educational outcomes had not improved, and the rate of neurodevelopmental impairment had not decreased (206). Thus, continued investigation is needed to characterize outcomes, identify risk and resiliency factors, and test the effectiveness of preventive measures for chronic developmental and health disorders among individuals born extremely preterm, towards the goal of eliminating extreme prematurity as a contributor to adverse outcomes in children.

Figure 1.

Frequencies for neurodevelopmental outcomes: comparison of the ELGAN cohort with reference data from the United States

Joseph RM et al. Pediatrics. 2016; Kuban KC et al. J Pediatr. 2008;153(4):466–72; Douglass LM et al. Pediatr Neurol. 2017;73:13–9; Joseph RM et al. Autism Res. 2017;10(2):224–32; Dvir Y et al. J Dev Behav Pediatr. 2019;40(9):725–34; Zablotsky B et al. Pediatrics. 2019;144(4).

Table 3.

Risk factors for neurodevelopmental impairment in the ELGAN cohort.

Outcome	Outcome assessment	Risk factor	Findings†	reference
Cerebral palsy	Standardized neurological examination at two year adjusted age(56)	Ultrasound-defined white matter injury¶	Quadriplegia: ORwhite matter: 28 (95% CI: 12, 64); ORwhite matter + IVH: 19 (9.2, 39) Diplegia: ORwhite matter: 5.6 (1.5, 21); ORwhite matter + IVH: 4.5 (1.5, 13) Hemiparesis: ORwhite matter: 5.6 (0.6, 50); ORwhite matter + IVH: 15 (4.9, 46)	(114)
		Bronchopulmonary dysplasia	Quadriplegia: ORBPD: 1.6 (0.8,3.2); ORBPD + vent: 5.7 (2.5, 13) Diplegia: ORBPD: 2.1 (0.8,5.0); ORBPD + vent: 4.2 (1.3, 14) Hemiparesis: ORBPD: 2.7 (0.7, 11); ORBPD + vent: 1.2 (0.1, 13)	(124)
		Male sex	GMFCS 1–2: 1.6 (1.2, 2.3) GMFCS 3–5: 2.1 (1.1, 4.2)	(21)
		Necrotizing enterocolitis requiring surgery	Quadriplegia: 1.1 (0.2, 5.2) Diplegia: 1.3 (0.1, 11)	(62)
		Bacteremia after first postnatal week	Quadriplegia: 1.4 (0.7, 2.7) Diplegia: 0.9 (0.4, 2.4)	(62)
		Intermittent/Sustained systemic inflammation of TNF-αφ	Quadriplegia: 1.6 (0.8, 3.1) Diplegia: 3.5 (1.4, 8.6) Hemiparesis: 2.5 (0.8, 7.5)	(152)
		Intermittent/Sustained systemic inflammation of IL-8φ	Quadriplegia: 2.1 (1.03, 4.2) Diplegia: 3.1 (1.2, 8.1) Hemiparesis: 3.1 (1.0, 9.9)	(152)
Cognitive Impairment	Differential Abilities Scale and A Developmental NEuroPSYchological Assessment (NEPSY-II) at 10 years of age	Ultrasound-defined white matter injury¶	ORwhite matter: 5 (2.1, 12.0) ORwhite matter + IVH: 4.5 (2.5, 8.1)	(70)
		Bronchopulmonary dysplasia	BPD associated with increased odds of test z-scores < 1 on tests of intelligence, executive function, and academic achievement; stronger associations for BPD plus ventilator, as compared to BPD without ventilator	(123)
		Male sex	Moderate impairment: 1.8 (1.1, 2.7) Severe impairment: 2.1 (1.2, 3.8)	(21)
		Bacteremia after first postnatal week	bacteremia associated with increased odds of test z-scores < 1 on tests of intelligence	(69)
		Intermittent/Sustained systemic inflammation of TNF-αφ	Moderate impairment: 2.8 (1.6, 4.6) Severe impairment: 2.2 (1.1, 4.3)	(67)
		Intermittent/Sustained systemic inflammation of IL-8φ	Moderate impairment: 2.2 (1.3, 3.9) Severe impairment: 2.8 (1.4, 5.5)	(67)
Autism Spectrum Disorder	Sequential assessments with Social Communication Questionnaire, Autism Diagnostic Interview-Revised, and Autism Diagnostic Observation Schedule	Ultrasound-defined white matter injury¶	ORwhite matter: 0.7 (0.1, 5.9) ORwhite matter + IVH: 0.6 (0.2, 1.8)	(70)
		Male sex	2.0 (1.1, 3.6)	(21)
		Birth weight z-score < −2	9.9 (3.3, 30) [OR for autism with normal IQ]	(212)
		Intermittent/Sustained systemic inflammation of TNF-αφ	1.1 (0.5, 2.7)	(79)
		Intermittent/Sustained systemic inflammation of IL-8φ	0.9 (0.3, 2.4)	(79)
		Intermittent/Sustained systemic inflammation of serum amyloid association proteinφ	2.5 (1.2, 5.3)	(79)
Epilepsy		Ultrasound-defined white matter injury¶	ORwhite matter: 7.6 (2.9, 20.1) ORwhite matter + IVH: 4.9 (2.3, 20.4)	(70)
		Bronchopulmonary dysplasia, requiring ventilator support at 36 weeks postmenstrual age	4.5 (2.2, 9.2)	(82)

^†when available, adjusted odds ratios are presented

^¶white matter injury defined as ventricular persistent enlargement or echolucent lesions in the cerebrum

^φTop quartile level on at least 2 of 3 measurements in first 2 postnatal weeks; Assays of proteins in dried blood spots(53)

Abbreviations: NESPY-II - A Developmental NEuroPSYchological Assessment; OR – odds ratio; CI – confidence interval

Call outs.

1. Individuals born extremely prematurely have about a 30-fold higher risk of cerebral palsy, a 9-fold higher risk of severe cognitive impairment, a 7-fold higher risk of epilepsy, and a 4-fold higher risk of autism spectrum disorders.

2. Adults born extremely preterm who had developmental impairments, as compared to those without such impairments, rated their quality of life lower.

3. Likely mediators between socioeconomic factors and adverse child health outcome include less optimal nutrition, increased maternal stress, environmental pollution, structural racism, less optimal learning environments, and less access to health care.

4. Indicators of social disadvantage, such as fewer years of maternal education, have been associated with differential CpG methylation in placenta of ELGAN participants, suggesting a potential role for the placenta in transducing the effects of socioeconomic environment on the fetus.

Abbreviations:

17-OHPC

17-hyroxyprogesterone caproate

ADHD

attention deficit hyperactivity disorder

ADI-R

Autism Diagnostic Interview – Revised

ADOS-2

Autism Diagnostic Observation Schedule-Second Edition

BPD

bronchopulmonary dysplasia

BMI

body mass index

BSID-II

Bayley Scales of Infant and Toddler Developmental-Second Edition

CBCL

Child Behavior Checklist

CSI-4

Child Symptom Inventory- Fourth Edition

DAS-II

Differential Abilities Scale-II

ECHO

Environmental Influences on Child Health Outcomes

ELGAN

Extremely Low Gestational Age Newborn

IHDP

Infant Health and Development Project

IQ

intelligence quotient

MCHAT

Modified Checklist for Autism in Toddlers

MDI

Mental Development Index

MRI

magnetic resonance imaging

NEPSY

NEuroPSYchological Assessment

NICU

neonatal intensive care unit

NIDCAP

Newborn Individualized Developmental Care and Assessment Program

PCHI

positive child health index

SCQ

Social Communication Questionnaire