eTable. Linear Relationships Between Gestational Age at Birth and 2-Year Neurodevelopment and Social-Emotional Development Scores in Moderate and Late Preterm Children
Customize your JAMA Network experience by selecting one or more topics from the list below.
Cheong JL, Doyle LW, Burnett AC, et al. Association Between Moderate and Late Preterm Birth and Neurodevelopment and Social-Emotional Development at Age 2 Years. JAMA Pediatr. 2017;171(4):e164805. doi:10.1001/jamapediatrics.2016.4805
What is the influence of moderate and late preterm birth on neurodevelopment and social-emotional development at age 2 years?
This longitudinal cohort study found that, compared with term-born children, children born moderate and late preterm are delayed in cognitive, language, and motor development. In addition, they have poorer social competence at 2 years’ corrected age.
Developmental surveillance is important given the risk of developmental delay in moderate and late preterm children.
Moderate and late preterm (MLPT) births comprise most preterm infants. Therefore, long-term developmental concerns in this population potentially have a large public health influence. While there are increasing reports of developmental problems in MLPT children, detail is lacking on the precise domains that are affected.
To compare neurodevelopment and social-emotional development between MLPT infants and term-born control infants at age 2 years.
Design, Setting, and Participants
This investigation was a prospective longitudinal cohort study at a single tertiary hospital. Participants were MLPT infants (32-36 weeks’ completed gestation) and healthy full-term controls (≥37 weeks’ gestation) recruited at birth. During a 3-year period between December 7, 2009, and November 7, 2012, MLPT infants were recruited at birth from the neonatal unit and postnatal wards of the Royal Women’s Hospital, Melbourne, Australia. The term control recruitment extended to March 26, 2014. The dates of the data developmental assessments were February 23, 2012, to April 8, 2016.
Moderate and late preterm birth.
Main Outcomes and Measures
Cerebral palsy, blindness, and deafness assessed by a pediatrician; cognitive, language, and motor development assessed using the Bayley Scales of Infant Development–Third Edition (developmental delay was defined as less than −1 SD relative to the mean in controls in any domain of the scales); and social-emotional and behavioral problems assessed by a parent questionnaire (Infant Toddler Social Emotional Assessment). Outcomes were compared between birth groups using linear and logistic regression, adjusted for social risk.
In total, 198 MLPT infants (98.5% of 201 recruited) and 183 term-born controls (91.0% of 201 recruited) were assessed at 2 years’ corrected age. Compared with controls, MLPT children had worse cognitive, language, and motor development at age 2 years, with adjusted composite score mean differences of −5.3 (95% CI, −8.2 to −2.4) for cognitive development, −11.4 (95% CI, −15.3 to −7.5) for language development, and −7.3 (95% CI, −10.6 to −3.9) for motor development. The odds of developmental delay were higher in the MLPT group compared with controls, with adjusted odds ratios of 1.8 (95% CI, 1.1-3.0) for cognitive delay, 3.1 (95% CI, 1.8-5.2) for language delay, and 2.4 (95% CI, 1.3-4.5) for motor delay. Overall social-emotional competence was worse in MLPT children compared with controls (t statistic mean difference, −3.6 (95% CI, −5.8 to −1.4), but other behavioral domains were similar. The odds of being at risk for social-emotional competence were 3.9 (95% CI, 1.4-10.9) for MLPT children compared with controls.
Conclusions and Relevance
Moderate and late preterm children exhibited developmental delay compared with their term-born peers, most marked in the language domain. This knowledge of developmental needs in MLPT infants will assist in targeting surveillance and intervention.
Prematurity is a recognized risk factor for developmental problems in childhood and adolescence.1 For many years, the focus of research has been on developmental outcomes of infants born very preterm (<32 weeks’ gestation). However, in recent times, data are emerging that children born moderate and late preterm (MLPT) between 32 and 36 weeks’ completed gestation are also at greater risk of developmental problems compared with their term-born peers.2-5 Because MLPT infants comprise the largest group of preterm infants (approximately 80%),6,7 small increases in adverse outcomes in this population have the potential to be a considerable public health burden. Recent reports on early childhood outcomes of MLPT infants highlight deficits in cognitive and motor domains, as well as social functioning.5,8-10 However, limitations of the studies include reliance on questionnaire assessments (rather than direct) and low follow-up rates, which potentially underestimate the prevalence of developmental problems. Furthermore, a better understanding of the specific developmental deficits in MLPT children is needed to inform public health policy, as is targeted surveillance and intervention in developmental domains at greatest risk. This study aimed to fill this gap in the literature by comparing neurosensory outcomes and cognitive, language, motor, and social-emotional development between MLPT children and term-born control children at 2 years’ corrected age. Given that there may be differences in developmental outcomes within the MLPT group, which spans from 32 to 36 weeks’ completed gestation, we also aimed to explore whether there was an association between gestational age at birth and developmental outcomes within this group.
The Late Preterm MRI Study11 is a longitudinal cohort study of brain development and outcomes in MLPT children compared with term-born children. During a 3-year period between December 7, 2009, and November 7, 2012, MLPT infants were recruited at birth from the neonatal unit and postnatal wards of the Royal Women’s Hospital, Melbourne, Australia, the largest of the 3 perinatal centers in the state of Victoria where tertiary neonatal care is provided. The dates of the data neurodevelopmental assessments were February 23, 2012, to April 8, 2016.
It is hospital policy for all infants born less than 36 weeks’ gestation to be admitted to the neonatal unit. Infants with congenital abnormalities or genetic syndromes known to affect development were excluded from the study. Healthy term-born infants (≥37 weeks’ gestation and birth weight ≥2500 g) were recruited from the postnatal wards at the Royal Women’s Hospital. Term infants were excluded from the study if they were unwell at birth, received resuscitation, were admitted to the neonatal nursery, or were identified as having conditions affecting growth or development.
Perinatal, neonatal, and maternal details were recorded by medical record review at the time of recruitment. Social risk was assessed using sociodemographic factors known to be associated with outcomes in preterm infants, including family structure, education of the primary caregiver, employment status of the primary income earner, occupation of the primary income earner, language spoken at home, and maternal age at birth of the child. Each variable was scored on a 3-point scale, where zero represented lowest risk and 2 represented highest risk, and summed to give a total score (range, 0-12). Social risk was then dichotomized to higher (total social risk score of ≥2) or lower (total social risk score of <2) risk.12 All participants were followed up at age 2 years, corrected for prematurity, at The Royal Children’s Hospital Melbourne.
This study was approved by the Human Research Ethics Committee of the Royal Women’s Hospital. Written informed consent was obtained from parents of all participants.
Cerebral palsy was diagnosed in children with abnormal tone and function, after the exclusion of other causes of motor dysfunction, by a pediatrician. The topography and severity of cerebral palsy were classified according to the Gross Motor Function Classification Scale.13 Blindness and deafness were also assessed by a pediatrician. Blindness was defined as having visual acuity of less than 6/60 in the better eye, and deafness was defined as a hearing impairment requiring amplification or a cochlear implant or worse.
Cognitive, language, and motor development were assessed using the Bayley Scales of Infant Development–Third Edition (Bayley-III).14 The Bayley-III is age standardized, is widely used in research and clinical settings, and has published norms that have a mean (SD) of 100 (15). Given the concerns about underestimation of developmental delay in the Australian population,15,16delay in any developmental domain was defined as less than −1 SD relative to the mean for the control children in the present study. We also defined moderate to severe developmental delay as less than −2 SDs relative to the mean for the control children. All assessors (one of us, C.R.P., and other nonauthors) were experienced in both the Bayley-III and neurological assessments and were unaware of group allocation and perinatal history.
Social-emotional and behavioral problems were assessed using the Infant Toddler Social Emotional Assessment (ITSEA).17 The ITSEA is a parent-report questionnaire developed for children aged 12 to 36 months. Aspects of child behavior are rated on 135 items using a 3-point Likert-type rating scale (0 is not true or rarely, 1 is somewhat true or sometimes, and 2 is very true or often [“no opportunity” is available for some questions]). Age- and sex-specific t statistics (mean [SD], 50 ; range, 25-80) are calculated for the following 4 domains: externalizing behavior problems, internalizing behavior problems, dysregulation, and social-emotional competence. The mean scores at or above the 90th percentile for externalizing behavioral problems, internalizing behavioral problems, and dysregulation at or below the 10th percentile for social-emotional competence were defined as being at risk of behavioral or social-emotional problems. The ITSEA has good test-retest reliability and good criterion validity.18
The expected rate of developmental delay for the controls based on previous literature was 16%.19 With a sample size of 201 participants in each group, the study was powered to detect an increase in the prevalence of developmental delay to 27.5% or higher among the MLPT children (equivalent to an odds ratio [OR] of 2) with 80.0% power based on a 2-sided test with type I error of 5.0%, which would represent a clinically important difference.
Participant characteristics were compared between those who were assessed at age 2 years and those who were not using means (SDs) for continuous variables and proportions for categorical variables. Two-year development was compared between groups using linear regression, with a separate model for each developmental outcome. The analysis was first performed unadjusted and then adjusted for social risk, which is known to affect development. Rates of cerebral palsy, blindness, deafness, developmental delay, and social-emotional problems were compared between groups using logistic regression, both unadjusted and adjusted for social risk, which is an important factor influencing development. All regression analyses were fitted using generalized estimating equations and are reported with robust (sandwich) estimates of standard errors to account for clustering because of multiple births within the same family. Linear regression was also used to determine the association between gestational age at birth and developmental outcomes at age 2 years in the MLPT group. Data were analyzed using statistical software (Stata, version 14.0; StataCorp LP).
Two hundred one MLPT infants and an equal number of term-born controls were recruited. Participant characteristics are summarized in Table 1. Compared with controls, the MLPT infants were more likely to be small for gestational age at birth. Compared with mothers of control infants, mothers of MLPT infants had higher rates of pregnancy-related complications, such as preeclampsia and antepartum hemorrhage, and higher rates of assisted conception, multiple births, antenatal corticosteroid use, and cesarean delivery. Moderate and late preterm infants had higher rates of neonatal morbidity compared with controls, although the overall rates were low. Social risk was comparable between the groups.
The follow-up rates were high for both groups at 98.5% (198 of 201) for MLPT children (1 was living abroad and 2 were unable to be contacted) and 91.0% (183 of 201) for controls (1 withdrew from the study, 3 were living abroad and 14 were unable to be contacted). Compared with those who were assessed at age 2 years, MLPT children who were not assessed had younger mothers (26.7 vs 34.1 years) and higher social risk (100% [3 of 3] vs 32.8% [63 of 192]). Compared with those who were assessed at age 2 years, controls who were not assessed had younger gestational age (39.2 vs 39.8 weeks), younger mothers (30.6 vs 33.5 years), and higher birth weight z score (0.80 vs 0.16). Other characteristics were similar.
Two-year outcomes are summarized in Table 2. There were 2 children in the MLPT group with cerebral palsy, one with hemiplegia (Gross Motor Function Classification System level 1) and another with quadriplegia (Gross Motor Function Classification System level 3) (P = .17 for adjusted by logistic regression). No children had blindness or deafness in either group. There was evidence that MLPT children performed more poorly than controls in all domains of development, most marked in language, with a mean difference of approximately 0.7 SD. This finding appeared to be driven equally by performance in the receptive and expressive language subdomains. There was also evidence that the MLPT group had poorer social-emotional competence compared with controls, with a mean difference of 0.4 SD, but were similar to controls in the other emotional and behavioral domains. Adjustment for social risk did not alter the conclusions and had little effect on the mean differences. Adjustment for imbalances in perinatal characteristics, including birth weight z score and multiple birth, did not alter any conclusions.
There was evidence that rates of a developmental delay were higher in the MLPT group than in controls in all domains, with weaker evidence for moderate or severe cognitive delay (Table 3). Compared with controls, MLPT children were more likely to be at risk in the social-emotional competence domain but not other behavioral domains. Adjustment for social risk did not alter any conclusions.
Within the MLPT group, there was little evidence of an association between gestational age at birth and neurodevelopment or social-emotional development at age 2 years. These results are summarized in the eTable in the Supplement.
This study adds to the growing body of evidence that MLPT birth is associated with an increased risk of developmental problems compared with term birth. Using direct, objective, standardized assessments, MLPT children at 2 years’ corrected age performed more poorly in cognitive, language, and motor domains compared with term-born controls. The disparity was greatest in the language domain, where MLPT children had 3 times higher odds of language impairment than their term-born peers, with both receptive and expressive language equally affected. It is also of great concern that MLPT children appear to be at much greater risk of motor impairment, with the odds of MLPT children scoring less than −2 SDs being 9 times higher than their term-born peers (after adjustment for social risk). We also found evidence of poorer social-emotional competence in the MLPT group. The competence domain of the ITSEA reflects aspects of children’s attention, compliance, and motivation, as well as social relationships (eg, play, empathy, and prosocial behavior). It is possible that these early signs of challenge are precursors for some of the school-age behavioral and learning problems described in MLPT children.20,21
In our MLPT sample, cognitive, language, and motor development were delayed compared with term-born controls in the order of 0.5 to 0.7 SD (−6.3 to −11.8 points), which are clinically important differences. Although our findings concur with other studies that MLPT children demonstrate poorer development as early as age 2 to 3 years, the disparity between MLPT children and term-born controls was larger in our study. Voigt et al9 reported a mean difference of −5.4 on the Bayley-II Mental Development Index between MLPT children and term-born controls using a sample from a single center. The population-based Early Childhood Longitudinal Survey–Birth Cohort study reported a mean t statistic difference of −1.4 (95% CI, −2.7 to −0.2) on the Bayley mental score using a short form of the Bayley-II modified specifically for that study.22 They did not report differences in motor development. Only late preterm children (34-36 weeks’ gestation) were included in the Early Childhood Longitudinal Survey–Birth Cohort study. Our findings of greater disparity between groups may be partially explained by our sample, who were recruited from a tertiary hospital and included moderate preterm infants. It is possible that our cohort included a greater proportion of sicker MLPT infants who were admitted to the neonatal nursery than in the general population, especially in the more mature gestational age range.
Rates of developmental delay in all domains were higher in the MLPT children compared with controls. In this study, we calculated rates of delay based on means (SDs) of the contemporaneous control group, which are higher than the normative mean (SD) of 100 (15) in the Bayley-III. Nonetheless, the values of our present full-term control cohort are comparable to data from contemporary term-born controls assessed using the Bayley-III.19,23 Previous studies have similarly reported higher rates of developmental delay in MLPT children than in controls in this age group. Using a parent questionnaire, a large regional cohort from the United Kingdom had an adjusted relative risk of 2.09 (95% CI, 1.19-3.64) for cognitive impairment in MLPT children compared with term-born children at 2 years’ corrected age.5 If anything, this finding may be an underestimation of the risk of delay because the follow-up rates were only approximately 60%. The Early Childhood Longitudinal Survey–Birth Cohort had increased odds of developmental delay at 24 months in late preterm children compared with term-born children for both cognitive (mild delay OR, 1.43; 95% CI, 1.22-1.67) and motor (mild delay OR, 1.58; 95% CI, 1.37-1.83) domains.8 Our findings of higher ORs may be accounted for by differences in samples, assessment tools, and definitions of delay.
Social and emotional behavioral problems in early childhood24 and at school age25,26 have been described in the very preterm population. Early identification of these problems is important because it is increasingly recognized that they are associated with later psychopathological conditions.27,28 Although data on MLPT children are scarce, there are reports of increased externalizing and internalizing problems in MLPT children at preschool and early school age compared with term-born controls.29,30 Recently, delays in social competence (but not other behavioral domains) were identified in MLPT children compared with term-born controls.10 Our results concur with this finding, highlighting the potential importance of social competence as a basis for longer-term developmental and behavioral problems described in school-age MLPT children. It is also possible that the parent-reported concerns about MLPT children’s social and behavioral competence are a reflection of cognitive and language developmental delays in everyday functioning.
The underlying brain basis for developmental problems in MLPT children is poorly understood. The MLPT period involves considerable growth and maturation of the brain. Increases in brain volume, whole-brain weight, and gyral and sulcal development are substantial in this period of late gestation.31 Therefore, there is potential for aberration of these processes after MLPT birth. Our group has recently reported differences in brain volumes, maturation, and microstructure at term-equivalent age in MLPT infants compared with term-born controls.11,32 Larger volumes of total brain tissue, white matter, and cerebellum were associated with better cognitive, language, and motor scores at 2 years’ corrected age in MLPT children.33 Moreover, altered neural activity as measured using functional magnetic resonance imaging has been reported in the primary motor and sensory regions in MLPT children compared with term-born controls.34 Collectively, there is accumulating evidence that alterations in MLPT brain growth and development may underlie some of the deficits described in these children.
In addition to neuroimaging findings, clinical assessment tools in the neonatal period may also provide a valuable indication of later functioning in MLPT infants. For instance, developmental delay in MLPT children may manifest at an earlier age as abnormal neurobehavior. In the same cohort, our group has recently shown that suboptimal neurological and neurobehavioral assessments, including on the Hammersmith Neonatal Neurological Examination and the NICU Network Neurobehavioral Scale, at term-equivalent age were associated with developmental delay at 2 years of age in MLPT infants.35 Further research into neurobehavioral deficits associated with MLPT birth may provide markers for specific developmental deficits in MLPT children.
The strengths of our study are that we recruited a large, prospective cohort and achieved high follow-up rates at age 2 years. We evaluated a range of developmental outcomes using direct criterion standard assessments in addition to parent-report questionnaires. However, we were limited by the inclusion of participants from a single tertiary center, which may have included sicker MLPT infants who were admitted to the neonatal nursery, especially those who were in the more mature gestational age range within the MLPT group. This inclusion potentially limits the generalizability of our findings to MLPT children in the wider population. However, our results are generalizable to the population of MLPT children who were more unwell and needed admission to a neonatal nursery after birth.
In summary, we have presented evidence that MLPT birth is associated with greater morbidities in neurodevelopment and social-emotional competence than term-born birth. This population is a group of infants who traditionally receive little neurodevelopmental surveillance. While there is good evidence that early intervention is effective in improving cognitive outcomes for preterm children up to school age, further research is needed on effective intervention to improve language and motor outcomes specifically in the MLPT population.36 To provide developmental follow-up and intervention to this group, it will be vital to identify risk factors to target those at highest risk of developmental problems given the large numbers of MLPT births. Further research directions into potentially modifiable factors, markers of poor outcome (eg, neuroimaging and neurobehavioral assessments in the newborn period), and the spectrum of deficits at school age and older have the potential to greatly improve the long-term care for this large group of children.
Corresponding Author: Jeanie L. Cheong, MD, Neonatal Services, Royal Women’s Hospital, Level 7, Parkville, Victoria, Australia 3052 (email@example.com).
Accepted for Publication: November 30, 2016.
Published Online: February 6, 2017. doi:10.1001/jamapediatrics.2016.4805
Author Contributions: Dr Cheong had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Cheong, Doyle, Treyvaud, Anderson, Spittle.
Acquisition, analysis, or interpretation of data: Cheong, Doyle, Burnett, Lee, Walsh, Potter, Treyvaud, Thompson, Olsen, Anderson.
Drafting of the manuscript: Cheong, Burnett, Treyvaud.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Cheong, Doyle, Lee.
Obtained funding: Cheong, Thompson, Anderson, Spittle.
Administrative, technical, or material support: Cheong, Walsh, Potter, Treyvaud, Thompson, Olsen.
Study supervision: Walsh, Treyvaud, Anderson, Spittle.
Conflict of Interest Disclosures: None reported.
Funding/Support: This work was supported by the following grants from the Australian National Health and Medical Research Council: project grants 1028822 and 1034516, Centre of Research Excellence grant 1060733, Early Career Fellowship grant 1053787 (Dr Cheong), Senior Research Fellowship grant 1081288 (Dr Anderson), Career Development Fellowship grant 1085754 (Dr Thompson), grant 1053609 (Dr Lee), and grant 1108714 (Dr Spittle). This study was also supported by The Royal Children’s Hospital Foundation and by the Victorian Government’s Operational Infrastructure Support Program.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: We thank the research coordinators, families, and children for their participation in this study.