Statistics from Altmetric.com
What is already known on this topic
Delivery before 39 weeks gestation increases the risk of cognitive problems. School performance in late preterm and early term children remains under-researched.
What this study adds
Late preterm and, to a lesser extent, early term delivery is associated with an increased risk of poorer school performance at 7 years.
Late preterm (34–36 weeks gestation) and early term (37–38 weeks gestation) birth is common and increasing in incidence worldwide.1 For example, these groups comprise 5% and 18% of all live births in England, respectively.2 Despite the large proportion of late preterm and early term births, there is a paucity of prospective research into their long-term outcomes.
The associations between extreme preterm birth (<28 weeks gestation) and poorer cognitive outcomes and lower levels of academic attainment are well established.3 ,4 Recent studies suggest that children born moderately/late preterm and early term have an increased risk of poorer health, cognitive development and school performance.5–8 A national English study of children in their first year of school, demonstrated that late preterm and early term children attain lower performance compared to their full-term peers.9 Given there are many important determinants of school performance, this research was undertaken to track the academic performance trajectory of children born at different gestational ages through key developmental stages, to determine whether the academic performance gap identified at 5 years persists at 7 years of age in this cohort. School performance by 7 years is an important outcome to assess because it is strongly associated with future qualifications, socioeconomic status and health.10–12
Gestational age in weeks was grouped into categories defined by the American Congress of Obstetricians and Gynecologists (full-term 39–41 weeks gestation, early term 37–38 weeks gestation, late preterm 34–36 weeks gestation, moderately preterm 32–33 weeks gestation, very preterm <32 weeks gestation).13 Measuring the spectrum of gestational ages enables investigation of a dose-response association between gestation and school performance.
Millennium Cohort Study
The Millennium Cohort Study (MCS) is a UK nationally representative longitudinal study of 18 818 children born in 2000–2001.14 Participants were randomly selected from the Child Benefit Records, a database registering >98% of children in the UK and contacted at 9 months of age for the first survey.15 There was a deliberate oversampling of areas with higher proportions of ethnic minorities and social disadvantage to enable sufficient representation of these populations. Surveys were performed at 9 months, 3, 5 and 7 years of age, collecting a wide range of demographic and health data, and linking consented education and health information from other sources.14 ,15
Gestational age was derived from the mother's report of the expected due date in weeks taken at the 9-month survey, which has been shown to have high agreement with routine hospital records except for >42 weeks gestation.16 The post-term group was excluded from the analysis due to potential misclassification and differences in causal pathways leading to post-term compared to preterm birth.
School performance using Key Stage 1
In England, all state-funded and some private schools progress through the national curriculum comprising of four ‘key stages’ which begin in year 1 (age 5–6 years) and are completed by the end of year 11.17 Key Stage 1 (KS1) covers coursework completed between ages 5–7 years in five key domains: reading, writing, speaking and listening, mathematics and science. The KS1 statutory assessments comprise teacher evaluations of the student's academic achievement throughout the school year in each domain according to uniform criteria, aided by standardised KS1 tests in most subject areas.18 At KS1, children generally perform between level 1 (below expected level) to level 3 (considerably above the expected level), with adequate performance categorised as achieving level 2 or above.18 For consenting MCS parents, the KS1 results were obtained from the Department of Education's National Pupil Database.
Exclusions and missing data
This study included MCS families who responded at 9 months and 7 years of age with known gestational age, who were born and attending school in England (figure 1). Of the 18 818 children recruited at age 9 months, 13 543(72%) responded to the survey at 7 years; 62% of these children were living in England. Children were also excluded if the mother was not the main respondent, or gestational age was unknown, implausible for birth weight or below 23 weeks or above 42 weeks. 94% of eligible MCS participants had parental consent for record linkage of KS1 results and 77% were successfully linked. This resulted in a total study population of 6031 children.
The most common sources of missing data were from study attrition (5275 (28%) at 7 years which were accounted for using non-response weights19) and missing KS1 results (1811 (23% unweighted) of eligible participants). Table 1 shows selected demographic factors of children with missing KS1 results. The percentage of missing KS1 results were similar among all gestational age groups (21–27%). Missing KS1 results were more common in children with non-white ethnicity, single-parent households, and mothers who had never worked.
KS1 results were converted into binary outcomes with adequate performance defined as level 2 or 3, and below the expected level as level 1 or below. The proportion of children performing below expected levels for each outcome was compared in each gestational age group against the full-term reference group. Children who attained level 2 or above in reading, writing and mathematics were categorised as achieving adequate general school performance,17 and the primary outcome was defined as not having reached this level. The secondary outcome examined the proportion of children performing below the expected level within individual subjects. The tertiary outcome was the proportion of children who were categorised as ‘considerably above average’ (level 3) generally and in specific subjects.
As study outcomes were common, risk ratios were estimated (rather than ORs) using modified Poisson regression to adjust for potential confounders.20 The child's sex and age within the school year (oldest—born September to December; middle—born January to April; youngest—born May to August) were adjusted for in all models. For children born between May and August 2001, who were also born before full-term, we also assessed whether they would have been placed in the following school year for corrected age. While the overall number of children who fell into this group was small (53 out of 6031), it affected significant numbers of children born very preterm/moderately preterm (25 out of 136; 18%) or born late preterm (28 out of 360; 8%).
Other variables likely to affect school performance were adjusted for if they were independently associated with the outcome (p<0.05): maternal age at delivery, maternal education, maternal socioeconomic status, marital status, multiple births, whether the child was firstborn, and smoking during pregnancy (all collected at 9 months). Six children had missing information on some confounding variables and were excluded from the final adjusted results.
All analyses were performed in Stata and allowed for non-response and the disproportionately stratified and clustered sampling using the ‘survey commands’.21
The MCS was approved by the Multicentre Research Ethics Committee. Our analysis of KS1 data was performed via the UK Secure Data Service (University of Essex, Colchester, UK) which required strict adherence to confidentiality guidelines. Due to statistical disclosure control guidelines, small sample sizes (n≤10) could not be published, so these have been removed from the data and replaced with an asterisk (*).
Among the 6031 children included in our study, 21% were born early term and 6% born late preterm, similar to national statistics.2 table 2 shows descriptive characteristics of the children according to gestational age. As expected, increasing prematurity was associated with multiple birth, low birth weight, neonatal intensive care, and shorter breastfeeding duration. Very preterm children were more likely to have single parents, lower maternal education and employment levels.
The percentage of MCS children who achieved an adequate level in each KS1 assessment was similar to the national data for England (see online supplementary table), and 18% of full-term children performed below the expected KS1 level. This proportion increased with decreasing gestational age from 21% in early term, 25% in late preterm, 31% in moderately preterm and 43% in very preterm children (table 3). After multivariable adjustment, there was a statistically significant increased risk of poor performance in those born very preterm (adjusted RR1.78, 95% CI 1.24 to 2.54) moderately preterm (adjusted RR 1.71, 95% CI 1.15 to 2.54) and late preterm (adjusted RR 1.36, 95% CI 1.09 to 1.68), but not in those born early term (adjusted RR 1.07, 95% CI 0.94 to 1.23).
The gradient of effect between gestational age and KS1 performance was demonstrable particularly in reading and writing (table 3). Early term children had a statistically significantly increased risk of poorer performance compared to full-term children in all subjects except writing with effect sizes (RR) ranging between 1.22–1.38. Late preterm children performed worse than full-term children in reading (adjusted RR 1.55, 95% CI 1.20 to 2.00) and writing (adjusted RR 1.35, 95% CI 1.07 to 1.71) only.
Gestational age was not strongly associated with achieving ‘considerably above average’ except that very preterm children were less likely to perform ‘considerably above average’ in general KS1 compared to full-term children (adjusted RR 0.92, 95% CI 0.88 to 0.96) (table 4).
This study demonstrates that at age 7 years, children born at lower gestations are at increased risk of poorer academic performance than their full-term peers. Children born very preterm were most likely to achieve below the expected level in general KS1, but moderately preterm and late preterm children had a 71% and 36% increased risk respectively. Early term children performed similarly to full-term children in general KS1 measures, however, they performed consistently poorer in all five subject domains except writing. The association between gestational age and school performance appears to follow a sigmoid gradient. Compared to full-term children, very preterm children have the highest risk of poor school performance, which lessens with moderately and late preterm children, and a small effect in early term children. Very preterm children are less likely to perform considerably above average, but the other gestational age groups are not strongly associated with this outcome. Although the adverse effects of late preterm and early term birth are small in comparison with factors such as gender, parental education and school attendance,9 ,11 ,22 they may augment these other risk factors for poor school performance. For example, 24% of boys and 15% of girls performed below the expected level in KS1. The additional effect of late preterm delivery (adjusted RR 1.36, 95% CI 1.09 to 1.68) would increase these risks to 33% and 20%, respectively.
The strengths of this study are that it examines a large, nationally representative cohort which has sufficient power to analyse the full range of gestational ages. The KS1 assessment is statutory, validated and based on subjective (teacher observation) and objective (test) components which together provide a measure of the child's performance throughout the year. This study also provides subject-specific data, supporting previous studies on the association between gestational age and school performance,5 ,23 and thus enabling identification of specific skill-sets which may be influenced by gestational age.
The main limitation of this analysis was missing data, with 28% study attrition and 23% missing KS1 results. Study attrition was addressed using non-response weights. The children with missing KS1 results were more likely to be very preterm, have higher rates of poorer health, special educational needs and lower socioeconomic status. Therefore, the true effect of gestational age on academic performance may be underestimated due to under-representation of these children.
Literature on the association between late preterm and early term birth and school performance is mixed, with some studies reporting small adverse effects,7 ,23 ,24 others reporting language difficulties only25 ,26 and some studies finding no effect.27 ,28 Our findings are consistent with most large cohort studies, demonstrating that decreasing gestational age is associated with a small increased risk in most general and subject-specific measures.9 ,23 ,24 ,29
This study follows a preceding analysis of school performance in the MCS cohort at age 5 years.9 Quigley's study demonstrated that early term and late preterm children had increased difficulties in ‘communication, language and literacy’ and ‘mathematical development’ compared to full-term children.9 These groups continue to have poorer performance in reading, writing (late preterm only) and mathematics (early term, but not late preterm) at age 7 years. Providing that there are a multitude of factors which determine school performance, it is striking that being born just a few weeks early continues to be associated with academic performance. These findings are supported by another UK study of KS1 performance, which found moderate to late preterm children (32–36 weeks gestation) were less likely to be successful in achieving level 2 or above in reading, writing and mathematics.5
The mechanisms underlying this association of decreasing gestational age and poorer school performance are likely to be multifactorial. Considerable brain growth occurs during the final trimester, with the brain at 34 and 36 weeks gestation weighing only 65% and 80% of the full-term brain, respectively.30 Foetal brain development is complex and occurs in highly specific orders and timeframes. Even small disruptions caused by a shortened gestation may have long-term ramifications.
Physiological immaturity of the late preterm or early term infant may also be another explanatory mechanism. Delivery prior to full-term increases the stress-mediated response, and predisposes to temperature and blood sugar instability which may affect brain development.31 ,32
Another potential explanation is the role of behaviour and attention in school performance. Research suggests that late preterm and early term children have higher rates of attention-deficit hyperactivity disorder33 and problematic behaviours in teacher and parent rating scales.7 ,34
Our results have identified that late preterm and early term birth are associated with an increased risk of poorer school performance. These risks may be inherent to preterm birth, or reflective of the maternal and foetal conditions which predispose to birth before full term. Nonetheless, these risks should be considered in decision making of deliveries prior to full-term. Children born after 33 weeks gestation do not receive routine neurodevelopmental follow-up, but increased vigilance of the potential risks associated with late preterm/early term birth and academic underperformance may enable earlier identification and management of potential school difficulties particularly if other risk factors are present. Although individual risks are small, late preterm and early term children comprise approximately one quarter of all live births, and thus public health implications may be substantial through potentially increased special educational needs, lower educational attainment and lower future income.8 ,35
Future research is required to discern performance trajectories of infants born late preterm and early term, identify high-risk subgroups, and develop effective monitoring and intervention strategies to improve school performance in children born prior to full term.
We would like to thank the families who participated in the Millennium Cohort Study. We would also like to thank the UK Data Archive for the Millennium Cohort Study datasets and the Secure Data Service for allowing us access to the school data. Although all efforts are made to ensure the quality of the materials, neither the original data creators, depositors or copyright holders, the funders of the Data Collections, nor the UK Data Archive bear any responsibility for the accuracy or comprehensiveness of these materials.
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.