CONTEXT: Late-preterm infants (LPIs) born at 34 to 36 weeks' gestation are increasingly regarded as being at risk for adverse developmental outcomes. To date, the early childhood development of LPIs has not been systematically considered.
OBJECTIVE: To undertake a broad examination of literature relating to early childhood development at the ages of 1 to 7 years of LPIs born at 34 to 36 weeks' gestation.
METHODS: We conducted a systematic review of early childhood outcomes in LPIs by using 9 electronic databases (January 1980 to March 2010). Bibliographies were reviewed. After examination of abstracts, ineligible studies were excluded. A specifically designed data-extraction form was used. The methodologic quality of included studies was assessed by using well-documented quality-appraisal guidelines.
RESULTS: Of 4581 studies, 10 (3 prospective and 7 retrospective cohort) were included. Studies were heterogeneous, and poorer outcomes were reported among LPIs in relation to neurodevelopmental disabilities, educational ability, early-intervention requirements, medical disabilities, and physical growth in comparison to term-born children. No identified study used healthy nonadmitted LPIs as a comparison group for admitted LPIs.
CONCLUSIONS: Evidence suggests that LPIs are at increased risk of adverse developmental outcomes and academic difficulties up to 7 years of age in comparison to term infants. An infant control group matched for gestational age has not been used; thus, for LPIs, the effect of neonatal admission on longer-term outcomes has not been fully explored. Systematic measurement of early childhood outcomes is lacking, and focused long-term follow-up studies are needed to investigate early childhood development after late-preterm birth.
- late-preterm infants
- neonatal admission
- early childhood
- developmental outcomes
- language development
Infants born preterm and admitted to neonatal intensive care (NIC) have unique and significant developmental concerns throughout early childhood.1,–,3 Much of the research to date has focused on those infants born extremely premature (23–28 weeks' gestation) or at a very low (<1500 g) or extremely low (<1000 g) birth weight. However, even those infants admitted for NIC for only short periods of time or for less invasive care may still be at risk of negative neurodevelopmental outcomes.4,5 Over recent years, there has been an increasing focus on the early outcomes of the late-preterm subgroup of premature infants. Late-preterm infants (LPIs) are defined as those born between 34 and 36 weeks' gestation6,7 and account for up to 75% of all preterm births8; there was a reported 25% increase in late preterm births from 1990 to 2006.9 Although some of these infants are not admitted for NIC but, rather, nursed with their mothers or in the special care setting only, the proportion of all infants admitted for NIC attributable to this late-preterm group is significant. Data relating to admission rates are not widely available, but sources have suggested that LPIs account for 20% to 25% of all NIC admissions.10,11
LPIs have unique and particular concerns in the neonatal period, including an increased risk of mortality in comparison to term infants.12,–,15 Furthermore, compared with term infants, LPIs are at significant risk for increased morbidity including hypothermia, hypoglycemia, hyperbilirubinemia, respiratory distress, poor feeding, and nutritional compromise in the early neonatal period.16,–,19 Disturbance of infant brain development during the third stage of pregnancy has also been reported.20,21 Thus, it is plausible to propose that long-term morbidity may indeed be a reality and that general developmental immaturity may persist in LPIs.
Currently, clinical practice in the early care of LPIs is varied, and the long-term effect of neonatal care on this population of infants remains largely unknown. There is a dearth of research relating to the early childhood development of LPIs,22,23 and further research has been called for in this infant group.6
To date, the number of studies on and the quality of information relating to the various facets of development of LPIs throughout early childhood has not been systematically considered. In light of this fact, we undertook a systematic review of the literature to explore current understanding of this significant group of NIC graduates.
The objective of this review was to examine studies of early childhood cognitive, motor, speech, and language development, health, and growth at the ages of 1 to 7 years of LPIs born at 34 to 36 weeks' gestation. Particular attention was given to the gestational age of the comparison groups used (whether healthy term-born infants or healthy, nonadmitted LPIs).
A comprehensive literature search was undertaken to identify literature from January 1980 through March 2010. The search strategy involved searching electronic databases, inspecting bibliographies of retrieved articles, and hand-searching the published literature.
We searched the Medline (from 1980), CINAHL (Cumulative Index to Nursing and Allied Health Literature) (from 1982), Embase (from 1980), PsycINFO (from 1987), and Maternity and Infant Care (from 1980) databases. Searches were also undertaken of LILACS (from 1982) and the Science Citation Index (from 1980) via the ISI Web of Knowledge. Alongside this, CDSR (Cochrane Database of Systematic Reviews [from 1992]), CENTRAL (Cochrane Controlled Trial Register [from 1992]), and National Health Service Centre for Reviews and Dissemination (including DARE [Database of Abstracts of Reviews of Effects]) were searched via Wiley. Ongoing and current research was identified through the UK Clinical Research Network and the National Research Register Archive.
The search was undertaken by using the following search terms: late preterm; near term; 34 to 36 weeks; moderately preterm; preterm; premature; neonatal intensive care; child development; long-term outcome; neurodevelopment; early childhood; cognitive; motor; speech; language; health; and growth and development. Appendix 1 highlights a single electronic search strategy.
Selection of Eligible Studies
A range of study methodologies were reviewed, including randomized controlled trials (including follow-up of randomized controlled trials), prospective and retrospective cohort studies (including longitudinal studies), case-control studies, and case-series studies. A number of articles were excluded after an initial review of titles and abstracts. After this process, a study-selection panel (Ms McGowan and Drs Alderdice, Holmes, and Johnston) agreed on the inclusion of articles. Studies were deemed ineligible for inclusion if any of the following applied: data-collection dates were not reported; data were collected before 1980; the number of study participants was too small (<30 participants); participants did not meet specified criteria for gestation (34–36 weeks only) and age (1–7 years); infants were recorded by birth weight only; or the research methodology was not adequately described. Only studies written in or translated into English were included. Full-text analysis was conducted for included studies. Authors of the early literature pertaining to preterm infant outcomes used birth weight as a measure of prematurity, which poses a well-recognized source of bias (potential inclusion of term intrauterine growth-retarded infants as “premature” and the potential for large-for-gestational-age infants to not be defined as preterm). In light of this potential, studies that solely used birth weight as the defining criterion were not included in this review.
Data were extracted by using a specifically designed data-extraction form that included the authors and year of publication; study design; number, gestation, and admission status of study and comparison-group infants; exclusion criteria; age at assessment; and primary objective. Available summary results were then tabulated. The original search methodology was designed to examine 5 specific domains of childhood development: cognitive development; motor development; speech and language development; health; and physical growth. There are larger bodies of research relating to specific developmental domains for other premature infant groups; however, this review of developmental outcomes of LPIs has identified broad outcomes assessed using widely varying scales and measures. Therefore, a descriptive methodology was chosen and a narrative synthesis was undertaken.
Included studies investigated developmental outcomes of infants after late-preterm birth. There is limited consensus on how to appraise the quality of such “outcome” studies.24 However, Hayden et al25 provided a useful quality-assessment framework for appraising evidence relating to prognosis or health outcomes. Their framework appraises 6 areas of potential bias: study participation; study attrition; prognostic factor measurement; confounding measurement and account; outcome measurement; and analysis. Alongside this framework, the International Society for Pharmacoeconomics and Outcomes Research (ISPOR) retrospective database checklist26 was used to assess the quality of data sources in those studies that used retrospective databases. Appendix 2 shows the adapted quality-assessment framework used. Published abstracts were not assessed for quality.
Ten studies considered the early childhood development of LPIs (see Table 1). The study-selection process is outlined in Fig 1. The combined initial database searches retrieved 4581 potentially relevant studies, of which 4192 were excluded on the basis of the title and abstract (including 46, which had not been translated into English). After a more detailed review, 222 did not meet all of the inclusion criteria and were excluded. In total, 167 articles were considered, and another 150 of them were excluded for the following main reasons: the studies included all preterm infants with no specific subgroup analysis of the defined late-preterm gestational age group, or the studies considered only short-term (up to 1-year) outcomes or considered developmental outcomes beyond the scope of the review (eg, behavior and attention). In total, 17 studies were considered by a panel of 4 investigators, and 10 were deemed eligible for detailed discussion in this review. Our stringent exclusion criteria excluded a number of studies on the basis of the “late-preterm” gestational age definition used (ie, not within the 34–36 weeks' gestational age group). Excluded studies that presented data relating to the wider group of infants born between 32 and 36 weeks' gestation have been detailed for reference (see Table 2).27,–,35 For example, Darlow et al27 considered 2-year outcomes in an entire preterm cohort, including a subgroup of infants who were born at 33 to 36 weeks' gestation; although excluded on the basis of gestational age, this study provided useful comparative data for the wider group.
Description of Included Studies
Included studies are summarized in Table 1. Of 10 studies that detailed infants born late preterm, 4 studies focused solely on the late-preterm group,36,–,39 and 6 studies included a subgroup of infants born at 34 to 36 weeks' gestation within the infant population considered.40,–,45 Most studies were conducted within developed nations: the United States (6),36,–,39,42,45 France (1),40 England (1),44 Norway (1),41 and Brazil (1).43 All of the included studies had been undertaken within the past 10 years. Seven of the studies had a retrospective cohort design,36,37,39,41,42,44,45 and 3 were prospective observational studies.38,40,43 Reporting of neonatal comorbidities was limited in all studies; however, the neonatal admission status of infants studied were as follows: 3 studies considered LPIs admitted for NIC37,39,40; 4 studies were of birth cohorts that included all LPIs41,–,44; 2 included LPIs defined as “healthy”36 or “without neonatal compromise that would qualify them for developmental follow-up”45; and admission status of the infants in 1 study were not reported.38
The results detailed in the following paragraphs are based on a narrative synthesis of studies identified in the existing literature relating to developmental outcomes of LPIs within 5 key areas: neurodevelopmental disabilities; educational ability; early-intervention requirement; medical disabilities; and physical growth. A summary of these results is shown in Table 3.
Early Childhood Outcomes
Neurodevelopmental disabilities have been defined as “a group of heterogeneous conditions that share a disturbance in the acquisition of basic developmental skills in a chronologically appropriate manner.”46 This broad definition may include motor impairment such as cerebral palsy (CP), global developmental delay, intellectual disability, or developmental language impairments. Six studies considered neurodevelopmental disabilities in LPIs38,–,42 (see Table 3). Three of these studies (2 prospective observational studies and 1 retrospective study) used standardized assessment tools.38,39,40 Marret et al40 examined infants born between 30 and 34 weeks' gestation and performed subgroup analysis of infants born at 34 weeks' gestation. Using the Kaufman-ABC assessment, they identified moderate cognitive impairment (Mental Processing Composite [MPC] scores of 70–84) in 18.6% of the infants born at 34 weeks' gestation and severe impairment (MPC scores of <70) in 5.3% of the infants. CP (according to the European CP Network definition) was noted in ∼1% of those born at 34 weeks' gestation, which is 10-fold that in the general population. Gyamfi38 assessed infants born late preterm at a mean age of 48 months (range: 32–64 months) by using the Ages & Stages Questionnaire to assess communication, gross motor, fine motor, problem-solving, and personal-social skills. This secondary analysis was undertaken with infants born after a randomized, placebo-controlled trial of 17 α-hydroxyprogesterone caproate, which was used to prevent preterm delivery. Contrary to other studies, the authors found similar Ages & Stages Questionnaire scores in both LPIs and term infants despite increased composite neonatal morbidity scores in LPIs. Baron et al39 conducted a retrospective cohort study considering neuropsychological outcomes of LPIs when compared with term infants at 3 years of age using the Differential Ability Scales (2nd ed) and a range of neuropsychological tests; relative deficits were observed in comparison with term infants in 2 aspects: visuospatial ability and verbal fluency. Deficits in attention/working memory, language, nonverbal reasoning, and manual coordination/dexterity were not found between the groups.
Three further retrospective cohort studies used linked data sets to analyze neurodevelopmental disabilities, 2 according to International Classification of Disease (ICD) codes41,42 and 1 according to database definitions36 (see Table 3). Moster et al41 reported data from a large cohort of premature infants in Norway, including data from a late-preterm subgroup, by using compulsory national registries to document medical disabilities and developmental outcomes. Significant increased risk ratios for CP (2.7 [95% confidence interval (CI): 2.2–3.3]) and developmental delay (1.6 [95% CI: 1.4–1.8]) were reported when compared with healthy term-born infants. Across this cohort of preterm infants, a continuous relationship between decreasing gestational age at birth and increased adverse outcomes without an obvious threshold was observed. Petrini et al42 reported the rate of developmental delay among LPIs as 12.2 per 1000 children and an adjusted hazard ratio of 1.25 (95% CI: 1.01–1.54) compared with those of the healthy term-born reference group. The adjusted hazard ratio for CP in this study was 3.39 (95% CI: 2.54–4.52) compared with that of term infants. Morse et al36 reported “developmental delay/disability” based on the requirement for early-intervention services up to 36 months; enrollment in such a program stipulates that an infant must attain a score of 1.5 SDs below the mean on a standardized developmental assessment. This requirement was 36% higher for healthy LPIs (stay of <72 hours) than for healthy term-born infants. An increased relative risk of disability in prekindergarten at 3 and 4 years was reported (adjusted relative risk: 1.13 and 1.10, respectively).
A significant proportion of current literature relates to educational ability, including infant groups between the ages of 3 and 7 years. Within all of the age groups, LPIs showed poorer academic performance and greater difficulty with school-related activities. Two key areas were considered: academic ability and special education requirement.
Three studies considered academic ability by using direct infant assessments and teacher ratings.36,44,45 An early prospective cohort study by Huddy et al44 encompassed infants born between 32 and 35 weeks' gestation and included 2 subgroups of infants born between 34 and 35 weeks' gestation. No comparison group was included. School problems in children were defined in 1 aspect of this study as those who had a score of >3 in a skill area using a 5-point scale of difficulty in completing tasks, as assessed by teachers. Children born between 34 and 35 weeks' gestation showed poor performance (with a score of >3) in writing/composition (34% and 33%, respectively), fine motor skills (29% and 33%), mathematics (29% and 31%), speaking/listening (18%), reading (21% and 22%), and physical education (8% and 9%). Morse et al36 evaluated 4 school outcomes between healthy late-preterm and term infants: not ready to start school; retention in kindergarten; suspension in kindergarten; and special educational status (see Table 3). They reported statistically significant differences between LPIs and term infants in 3 of the outcomes considered, and results of additional analysis indicated that infants born at 34 weeks' gestation were more likely to be “not ready for school” than infants born at 35 or 36 weeks' gestation. Chyi et al45 compared learning difficulties between groups of infants born at 32 to 33 and 34 to 36 weeks' gestation and term infants from kindergarten through to fifth grade (only data for children up to 7 years old are reported here) by using the Early Childhood Longitudinal Study-Kindergarten Cohort data set. An adapted assessment including Peabody, Woodcock, Kaufman, and primary tests of cognition, reading, and math ability revealed that LPIs had lower reading and math scores than term infants in kindergarten and first grade (P < .05). The risk of LPIs obtaining below-average academic rating scores remained elevated at first grade for reading and math in comparison to that of term infants.
The second reported measure of school performance is that of a special education requirement. Chyi et al45 considered the presence of an Individualized Education Program or special education services as an indicator of learning difficulties. They concluded that, although the results were nonsignificant, more LPIs were enrolled in the Individualized Education Program than term infants both in kindergarten (8.04% vs 6.18%) and first grade (10.54% vs 7.48%); special education services enrollment was also greater in kindergarten and first grade. Huddy et al44 also noted that support from a “nonteaching” assistant at school was required by 24% of children who were born at 34 to 35 weeks' gestation; however, no comparison group was available. Furthermore, Morse et al36 reported that LPIs were at increased risk of being assigned “exceptional student status” in comparison to term infants.
In a retrospective cohort study, Kalia et al37 considered the requirement for early intervention (therapeutic services) by using a unique comparison group of very preterm infants (<32 weeks' gestation). Overall, the percentage uptake of early-intervention services was lower for LPIs when compared with those in the very preterm group. However, after controlling for neonatal comorbidities, including 5-minute Apgar scores, receipt of caffeine for apnea of prematurity, bronchopulmonary dysplasia, respiratory distress syndrome, and length of stay, in the very preterm infant group, there was no significant difference in enrollment in early-intervention services between the late-preterm and very preterm infants.
Medical disabilities were not reported extensively within the late-preterm outcomes literature. However, the authors of 3 studies did report sensory impairment: visual impairment/blindness; hearing impairment/deafness; and seizure disorders.40,–,42 Moster et al41 recorded medical disability as the occurrence of a single ICD code (see Table 3) denoting any of the above-listed conditions and reported a rate of 0.3% in the 34- to 36-weeks' gestation group (adjusted relative risk: 1.5 [95% CI: 1.2–1.8]; P < .001) compared with term infants. This analysis excluded ∼2.3% of late-preterm children who died before their fifth birthday. Visual (0.8%) and hearing (1.5%) impairment were also reported by Marret et al40 and did not vary significantly from infants born at 30 to 33 weeks. Seizure disorders were reported by Petrini et al42 with an adjusted hazard ratio of 1.27 (95% CI: 0.69–2.32) at 34 to 36 weeks compared with term infants. The general health status of LPIs beyond 1 year was not identified in any of the studies.
Physical growth was considered a primary outcome in 1 population-based cohort of all births during 1 year from a middle-income country: the 2004 Pelotas Cohort (southern Brazil).43 The authors reported rates of underweight, stunting, and wasting in a late-preterm subgroup of all preterm infants in comparison with term infants. LPIs were at increased risk of underweight and stunting at 12 and 24 months of age (adjusted OR: 3.36 [95% CI: 1.56]7.23] and 2.30 [95% CI: 1.40–3.77], respectively). Wasting was significantly different between the groups in the first year of life; only a small number of the infants were reported to have wasting at both 12 and 24 months. Gyamfi38 also considered height and weight in LPIs in comparison with term infants at 48 months (range: 32–64 months) as a secondary outcome, and no significant difference between the 2 groups was reported.
Quality Assessment of Studies
Our quality assessment of studies using the ISPOR Retrospective Database Checklist26 and guidelines for assessing quality in prognostic studies25 is summarized in Table 4 and identified the following methodologic concerns. First, regarding study design and sample size, 5 retrospective cohort studies used existing data sets,36,37,41,42,45 and only 1 of these studies fully outlined the quality assessment of the original data sources.42 In addition, outcomes were defined on the basis of existing data rather than being purposefully selected as measures of infant development. Although loss to follow-up and missing data are inherent concerns in longitudinal cohort studies, 8 of 9 of the included studies did account to some degree for missing data or loss to follow-up.36,37,40,–,45 Prospective and observational studies also tended to be limited by small sample sizes, and 2 studies lacked a comparison group.40,44 Second, the use of nonstandardized outcome measurements such as neurologic examination, caregiver report, and early-intervention uptake increases the risk of measurement bias and makes comparability between studies difficult. Indeed, although standardized ICD codes were used for defined outcomes in 2 studies,41,42 the authors highlighted the subjective nature of diagnosis and the potential for missing data. Finally, although adjustment for previously reported potentially confounding perinatal, maternal, and socioeconomic factors was noted across the included studies, neonatal and childhood comorbidities were not considered at length.
LPIs constitute an epidemiologically significant group of preterm infants and NIC graduates, yet the extent of adverse developmental outcomes in their early years remains largely underresearched. In this comprehensive review of the literature, 10 studies relating to early childhood development up to the age of 7 years were identified. Only 4 of these studies focused solely on LPIs, which highlights a paucity of focused research on this particular group of infants. However, the authors of all but 1 of the included studies reported similar trends of adverse early childhood developmental outcomes in the late-preterm group as a whole.
LPIs were at increased risk of neurodevelopmental disabilities up to 7 years, poorer performance on standardized testing, and increased diagnoses of developmental delay in comparison to term infants. Significant development of the infant brain takes place during the last 4 to 6 weeks of pregnancy; there is a fourfold increase in cortical volume during the third trimester20 and an accrual of 35% of brain weight during the last 6 weeks of gestation.21 Preterm delivery as an interruption of these processes may contribute to adverse neurodevelopmental outcomes experienced by LPIs. In addition, complex medical problems in the early neonatal period may further compound the negative effect of early birth and associated neonatal admission. It is interesting that 1 included study, which reported a requirement for early intervention, noted a similar uptake of services between LPIs and very preterm infants (<32 weeks) after adjustment for neonatal comorbidities,37 which further highlights the influence of morbidity experienced in the early neonatal period on later childhood development. Despite this observation, only 3 studies specifically reported the NIC admission status of infants. It is of particular note that the neonatal admission status of infants and comorbidities experienced within the late-preterm group and their relationship to ensuing developmental outcomes have not been well addressed. At the time of this review, no identified study had used healthy nonadmitted LPIs as a comparison group for complicated, admitted LPIs.
A consistent observation throughout the review is that LPIs have more favorable outcomes than very preterm infants but less favorable outcomes than term infants. There seems to be a continuous relationship between decreasing gestational age and increasing risk of adverse outcomes such as neurodevelopmental disabilities and academic performance. This “scale” of prematurity identifies an important aspect of LPI development with an emphasis not on severe disability but potentially more and multiple subtle developmental concerns. Winders-Davis47 highlighted the complexity of identifying these milder disabilities and the adverse effect they have on global development and noted that if undetected by school age, these milder disabilities may have a negative cumulative effect on development. This effect was apparent in those studies relating to school performance and academic ability, in which LPIs performed less well than their term-born peers and required more special education or academic support.36,45 Developmental follow-up of infants born at late-preterm gestations during preschool years may help identify and alleviate subtle difficulties and potential learning problems encountered at school age.48 As the number of LPIs delivered each year continues to rise,9 the requirement for early intervention and early educational input becomes increasingly significant both in clinical follow-up and educational policy and planning.
The overarching aim of this review was to gain an understanding of early childhood development in the LPI population. Although LPIs were previously considered similar to term infants, emerging evidence suggests that significant adverse developmental outcomes do exist among LPIs, which further indicates that longer-term outcomes of prematurity remain a concern even for those infants born at the more optimistic late-preterm stages of pregnancy.
In recent years, LPIs have increasingly been regarded as “at-risk” rather than “low-risk” infants. They are born developmentally immature and with increased neonatal health concerns compared with term infants. The impact of early neonatal care on longer-term outcomes has not yet been well considered; comorbidities, neonatal admission, and surrounding factors have not been fully explored. Systematic measurement of early childhood outcomes, such as those already considered for extremely preterm infant groups, is lacking in the late-preterm population. There is a real need for focused long-term follow-up studies to investigate early childhood development after late-preterm birth.
Ms McGowan receives a PhD studentship from the Department for Employment and Learning (Northern Ireland).
- Accepted January 14, 2011.
- Address correspondence to Fiona A. Alderdice, PhD, School of Nursing and Midwifery Medical Biology Centre, 97 Lisburn Road, Belfast, BT9 7BL, United Kingdom. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- NIC =
- neonatal intensive care •
- LPI =
- late-preterm infant •
- ISPOR =
- International Society for Pharmacoeconomics and Outcomes Research •
- CP =
- cerebral palsy •
- ICD =
- International Classification of Diseases •
- CI =
- confidence interval
- Caravale B,
- Tozzi C,
- Albino G,
- Vicari S
- Raju TNK,
- Higgins RD,
- Stark AR,
- Leveno KJ
- McCall E,
- Craig S
- Young PC,
- Glasgow TS,
- Li X,
- Guest-Warnick G,
- Stoddard G
- Wang ML,
- Dorer DJ,
- Fleming MP,
- Catlin EA
- Shapiro-Mendoza CK,
- Tomashek KM,
- Kotelchuck M,
- et al
- Altman DG
- Kerstjens JM,
- Butcher PR,
- Molen van der JL,
- et al
- Morse SB,
- Zheng H,
- Tang Y,
- Roth J
- Kalia JL,
- Visintainer P,
- Brumberg HL,
- Pici M,
- Kase J
- Gyamfi C
- Huddy CL,
- Johnson A,
- Hope PL
- Shevell MI
- McCain M,
- Mustard JF,
- Shanker S
- Copyright © 2011 by the American Academy of Pediatrics