OBJECTIVE: Children born at <32 weeks' gestation are at increased risk of intellectual impairment. Few studies have focused on the majority of preterm children born at 32 to 36 weeks' gestation. We aimed to investigate the association between the full range of gestational ages at birth and the risk of not completing basic school.
METHODS: This longitudinal, register-based study included all live-born infants in Denmark from 1988 to 1989. Data were obtained from national registers. School achievements were evaluated by using the examination marks. The association between gestational age and not completing basic school was estimated, taking into account the effect of their parents' educational level, being small for gestational age, plurality, and cerebral palsy.
RESULTS: The cohort constituted 120 585 infants, of whom 118 281 (98.6%) were alive in 2007. Of these infants, 5.01% (n = 5.928) were born before 37 weeks' gestation. Of the subjects born before 37 weeks' gestation, 11.5% (95% confidence interval: 10.7–12.4) did not complete basic school compared with 7.5% (95% confidence interval: 7.3–7.6) of those born at term. The percentage of subjects who did not complete basic school increased with decreasing gestational age. The increase was steeper at <31 weeks (4.2% per week) than at 31 to 36 weeks' gestation (0.5% per week).
CONCLUSIONS: The risk of not completing basic school increased with decreasing gestational age. The risk was moderate at ≥31 weeks' gestation and increased steeply at <31 weeks' gestation. The increase at <31 weeks' gestation was only partly explained by cerebral palsy.
WHAT'S KNOWN ON THIS SUBJECT:
Children born very preterm are at increased risk of having difficulties in school. The majority of preterm children are born moderately preterm. Few studies have found that moderate preterm birth is associated with having difficulties in school.
WHAT THIS STUDY ADDS:
Moderate preterm birth was associated with school difficulties persisting through basic school. The risk of not completing basic school increased with decreasing gestational age. The absolute increase was minimal above 32 weeks' gestation but more pronounced below 31 weeks' gestation.
It is widely known that children born very preterm, before 32 weeks' gestation, are at increased risk of having difficulty in school. Studies from all over the world have revealed an increased risk of cognitive difficulties, such as learning disabilities, grade failure, and placement in special classes, appearing once the child reaches school age.1,–,4 The majority (∼85%) of preterm children, however, are born moderately preterm (ie, between 32 and 37 weeks' gestation). Recently, the proportion of moderately preterm children has increased from 4.40% in 1995 to 5.36% in 2004 in Denmark,5 and a similar rise has been observed in the United States, in which 10.6% of all children are born moderately preterm.6,7 A vast majority of people born after 32 weeks' gestation experience an uncomplicated neonatal course and are subsequently judged as functioning normally and usually are discharged without follow-up. The few studies that have investigated the outcome of people born after 32 weeks' gestation, however, have found that moderate preterm birth is associated with school difficulties. Thus, a British study8 revealed that up to one-third of children born between 32 and 35 weeks' gestation have school difficulties at 7 years of age despite an uncomplicated neonatal course in most cases. An increased risk of school difficulties in children born between 32 and 36 weeks also was found in studies from North America and Denmark9,10 in which children from kindergarten to grade 5 and at 10 years of age, respectively, were evaluated. The British study lacked a control group, which makes the effect of preterm birth difficult to interpret, and the studies from North America and Denmark8,–,10 did not investigate whether school difficulties found in the first years of school were temporary or permanent. Other studies11,–,17 have focused on school outcome in relation to birth weight, and it is well established that a low birth weight (<2500 g) is associated with having school difficulties. A low birth weight, however, can be the result of both a low gestational age and being small for gestational age (SGA), and the separate significance of these covariates is not clear.18,–,20
In our register-based study, we aimed to describe the association between gestational age and the risk of not completing basic compulsory 9-year school requirements by using data on all live-born infants from 1988 to 1989 in Denmark.
This study was a register-based follow-up study of all live-born children born from 1988 to 1989 in Denmark. All residents in Denmark receive a fixed personal identification number (central population register number) shortly after birth from the Danish civil registration system. Using this number, Statistics Denmark holds the demographic and vital statistics of all people in various registers, which makes it is possible to link information from these registers for research purposes.
The study was approved by the Danish Data Protection Agency. Danish legislation allows this type of study without individual consent and ethical approval.
Demographic measures were extracted in December 2006, except for the data on mortality and education, which were obtained in October 2008. All live births in 1988 and 1989 were identified in the fertility database. Data on gestational age, birth weight, gender, and plurality were obtained. Plurality was coded dichotomously as singleton or multiple birth. People without information on gestational age and/or birth weight (n = 653) were excluded. In the fertility database, gestational age was given as full weeks of gestation (range: 22–45). We calculated a birth-weight z score by using the fetal growth standards published by Marsal et al21 and excluded people with extreme values (ie, ±6 SDs from the mean birth weight [n = 21]) (Fig 1). To examine the graduated influence of growth retardation, the z score also was used to identify infants with a birth weight between 1 and 2 SDs below the mean for gestational age (SGA − 1 SD), between 2 and 3 SDs (SGA − 2 SDs), and below 3 SDs (SGA − 3 SDs). Survival rates at 28 and 365 days and at follow-up were calculated from data from the cause-of-death register.
From the educational register, we extracted information on the highest achieved level of education for both parents by 2007 and categorized subjects according to the parent with the highest level into 3 groups according to the International Standard Classification of Education (ISCED). The ISCED is based on the achieved qualifications rather than the length of the education: (1) those with basic education (compulsory basic education corresponding to 9 years' attendance, ISCED 0–2A); (2) skilled professionals (equivalent to secondary education, ISCED 3C–4C); and (3) those with tertiary education (different levels of professional education, ISCED 5A–6). Diagnoses of cerebral palsy were obtained from the National Patient Register, which holds information on all discharge diagnoses from secondary health care. Diagnoses made since 1993 were coded according to the International Classification of Diseases, 10th Edition (ICD-10) classification and, before that, according to the ICD-8 classification by using ICD-8 codes 34.30 to 34.50 and ICD-10 codes G80 to G82. Information on emigration was extracted from the central population register to identify people who had emigrated before or during the years when basic school was expected to be completed.
In Denmark, children usually start school at 6 to 11 years of age, and the compulsory basic school requirement is of 9 years' duration. Children have to attend either a public-sector school or obtain education at an equal level in a private school system. The final examination after the ninth grade is identical for all pupils in mainstream schools. To qualify for university, it is necessary to complete an additional 3-year education comparable to high school or A-level education. Admission to high school depends on a general assessment and on the results of the compulsory school-leaving examination. Some private schools, which represent 1% to 2% of the ninth-grade pupils, do not offer an examination after ninth grade, and some children who attend public-sector schools do not take the examination, primarily because of severe learning disabilities or school difficulties for other reasons. The majority of those who do not obtain the ninth-grade examination fail to do so because of learning disabilities.
From the Ministry of Education we received information of individual grades from the ninth-grade examinations after the compulsory 9 years of school. Grades were given as the average result of a general examination of academic achievement, which is scored on a scale of 0 to 13; the average performance is 8, and a score below 6 is considered not satisfactory (ie, failed). The examination grades are supplemented by teacher ratings (the same 0- to 13-point scale). The examination covers major domains of academic achievement, including Danish, foreign languages, mathematics, hard science, and social science. It is a legal requirement for all schools, except special schools for pupils with severe learning disabilities, to report the examination grades and teacher evaluations to the Danish Ministry of Education. We used the average of the examination grade and the teacher evaluations completed through ninth grade. If the grades or teacher's evaluation were not available, the person was considered to have “not completed basic school.” Students will not be given a grade if they do not attend the examination or are not at school adequately on a daily basis. The normal age of finishing ninth grade is 15 to 16 years. We used a cohort of subjects aged 18 to 19 years, taking a delay into account.
The risk of not completing basic school was analyzed by means of logistic regression models, including the following explanatory variables: gestational age in weeks, gender, multiple births, being SGA by −1, −2, and −3 SDs, parents' educational level, and diagnosis of cerebral palsy, both 1 at a time in univariate models and simultaneously in multiple models. The odds ratio (OR) approximates the relative risk.22
The average mark (average of examination grade and teacher ratings) when completing basic school was analyzed in a multiple linear regression model that included cerebral palsy status, gender, multiple birth, SGA by −1, −2, and −3 SDs, and parents' educational level as explanatory variables. Average marks followed a normal distribution and were expressed as z scores. Statistic analyses were performed by using SAS 9.1 (SAS Institute, Cary, NC).
Of 120 585 live-born children registered between 1988 and 1989 in the birth register, a total of 118 281 (98.6%) were alive in 2007. Of these children, 5.01% (n = 5.928) were born before 37 weeks' gestation. The neonatal mortality rate among children born earlier than 32 weeks' gestation was 24.5%, and for those born between 32 and 36 weeks it was 1.78% compared with 0.16% at 37 to 42 weeks' gestation (Table 1). In the whole population, 20.8% (n = 24 551), 4.2% (n = 4991), and 0.6% (n = 743) were SGA by −1, −2, and −3 SDs, respectively. At follow-up in 2007, 26.9%, 3.06%, and 0.72% of children born with a gestational age less than 32 weeks, 32 to 36 weeks, and at term, respectively, had died. Likewise, 72 (7.1%), 49 (1.00%), and 314 (0.25%) had a diagnosis of cerebral palsy in the 3 gestational-age groups, respectively (Table 2).
Of the children who did not complete basic school, 2108 had emigrated at the age when they would have completed basic school (1.75% preterm and 1.82% term). Of the children born before 37 weeks' gestation, 11.5% (95% confidence interval [CI]: 10.7–12.4) did not complete basic school compared with 7.5% (95% CI: 7.3–7.6) of those born at term.
Figure 2 shows the percentage of surviving children who completed basic school and ORs of preterm children who did not complete basic school. The percentage of children who did not complete basic school increased stepwise with decreasing gestational age, from 6.8% (95% CI: 6.5–7.1) at 41 weeks' gestation to 28.0% (95% CI: 19.9–36.1) at less than 28 weeks' gestation. This relation (Fig 2A) was much steeper below 31 weeks' gestation and gives a regression line of −4.2% per week compared with the regression line above 31 weeks' gestation of −0.5% per week. The crude OR of not completing basic school was 1.32 (95% CI: 1.15–1.52) to 3.26 (95% CI: 2.26–4.71) between 36 and 30 weeks' gestation, including a nonsignificant lower OR at a gestational age of 31 to 32 weeks, but it rose to 4.82 (95% CI: 3.22–7.21) at a gestational age of less than 28 weeks.
Parents' educational level was used as measure of socioeconomic status in the logistic regression analyses (Table 3). These data were missing in 0.44% of cases (n = 512); the missing data were equally distributed across all gestational-age groups. The parents' educational level was associated with not completing basic school, low versus high educational level (OR: 5.06 [95% CI: 4.76–5.38]), and medium versus high educational level (OR: 1.87 [95% CI: 1.77–1.97]). A progressively stronger association was seen for SGA of −1 SD (OR: 1.45 [95% CI: 1.38–1.53]), −2 SDs (OR: 1.91 [95% CI: 1.75–2.08]), and −3 SDs (OR: 2.80 [95% CI: 2.32–3.38]), and an even stronger association for cerebral palsy (OR: 14.07 [95% CI: 11.63–17.02]) was found. Three multiple logistic regression analyses were performed: model 1 described status before birth; model 2 added variables known right after birth; and model 3 further added the diagnosis of cerebral palsy. In models 1 and 2, the parents' educational level (OR: 5.02 [95% CI: 4.72–5.34]) and being SGA, multiple births, and gender reduced the OR in the gestational-age groups above 27 weeks, whereas a diagnosis of cerebral palsy (model 3) had the greatest impact in the gestational-age groups below 36 weeks' gestation. A weak interaction between SGA and gestational age was eliminated by adjusting for the effect of cerebral palsy.
An inverse linear correlation was found between gestational age and age at completing basic school. The age increased 0.016 years (95% CI: 0.0177–0.0145) by each gestational-age week, and there was no significant break at 31 weeks. The difference between individuals born earlier than 28 weeks' gestation and at term was 4.24 months. Only small differences were found according to mean marks by gestational ages. Children with a gestational age of 41 weeks were found to have the highest mean z score (0.027 [0.013–0.0.41]), and this score decreased by 0.01 (0.006–0.012) by each gestational week, which results in a difference of 0.2 SDs in mean z score across gestational age (Fig 3). Children born at earlier than 28 weeks' gestation had a mean z score of 0.062 (95% CI: −0.127 to 0.251), and, although not statistically significant, it differed positively from the tendency of other mean grades. Testing separately the examination grades and teachers' rating, we found the same effect across gestational-age groups.
In this national register–based study that included 118 281 children, of whom 5.01% (n = 5928) were born preterm, we found that the percentage of preterm children who did not complete basic school increased week by week according to decreasing gestational age, from 6.8% at a gestational age of 41 weeks to 28.0% at a gestational age of less than 28 weeks. There was a significant association between moderate preterm birth and not completing basic school (crude OR: 1.32–3.26 at a gestational age of 36–30 weeks), although the association was partly explained by the parents' educational level and being SGA (adjusted OR: 1.14–3.03). Thus, our data corroborate that moderate preterm birth in itself is associated with a risk of not completing basic school. The absolute risk increase, however, was only 1% to 3%. For children born earlier than 31 weeks, however, the risk increased by −4.2% per gestational week. Cerebral palsy was an important risk factor, although after adjusting for cerebral palsy the OR was still as high as 3.13 at less than 28 weeks' gestation. Similar to previous studies,20,23 we also found that being SGA is a risk factor for school difficulties. We expected that this effect would be more pronounced in the lowest gestational-age groups; however, that was not the case. It could be attributed to the low survival rate of the smallest SGA-born individuals in our cohort, which left us with less power to explore this issue.
We made use of routine data from the Danish national registers, which are updated annually at the individual level and have a high level of completeness. Use of these registers is a major advantage, because the bias related to follow-up is minimized. The proportion of pupils who did not complete basic school was as high as 7.7% in our entire cohort. In Denmark, ∼1.5% of pupils attend free schools that have special permission not to evaluate after 9 years of school attendance, and these young adults manage equally well compared with students from public-sector schools in later academic education (Ministry of Education, unpublished data, 2003). Thus, accounting for the pupils attending free schools, the 7.7% of the children who do not complete basic school after 9 years is probably overestimated by ∼20%. However, we did not try to adjust our results accordingly, because we lacked the exact information concerning free schools. The Economic Council of the Labor Movement in Denmark found in 2002 that 8.4% of a birth cohort did not complete basic school. In addition, it found that of this group ∼15% proceeded in basic school at the 10th-grade level, whereas 30% started vocational training for youngsters and 55% dropped out of the educational system (J. Mehlbye, MPA, verbal communication, 2009). Thus, the outcome measure of not completing basic school is important and indicates that a significant proportion of these pupils have experienced learning difficulties, although the kinds of difficulties have not been described in detail.
The internal validity of this study is strengthened by the consistent expected outcome of known confounders. By testing the association between preterm birth and not completing basic school by including social confounders in the analyses, we found no reason to suspect that our results are confounded by a different proportion of preterm children attending free schools, although we have no direct data to illuminate this relationship.
Our incidence of cerebral palsy is overestimated. The Danish Cerebral Palsy Registry, which covers the east of Denmark, revealed a significantly lower incidence of cerebral palsy in 1987–1990, especially in the term group.24 For the purpose of this study, the cerebral-palsy variable is likely to be a relevant measure of brain dysfunction that appears before the age of 5 years.
Our results confirm findings from previous studies that moderate preterm birth is associated with school difficulties. We have no data on special educational support, because these data are not centrally registered, but our data suggest that the increased risk of school difficulties in preterm children found during basic school in other studies10,14,17 seems to persist rather than being temporary.
It is not clear whether the increased risk of school difficulties in preterm children is caused by a general cognitive deficit or by more specific cognitive dysfunctions. In a British study, a high correlation was seen between examination results and IQ scores in the general population.25 We have no data to illuminate this relation in Denmark. The most notable finding in our study, however, is that the relation between gestational age and not completing basic school corresponds closely in shape with a meta-analysis of the Bavarian and EPIcure studies of IQs in children born preterm. The curve that describes the association between IQ and gestational age showed a significant break similar to ours at 31 weeks' gestation, when the IQ declined steeply.26 In absolute terms, the absolute risk difference for not completing basic school, attributable to preterm birth, also is in the same order of magnitude as the risk of developing a subnormal IQ (less than −2 SDs) (ie, rising to 20% earlier than 28 weeks' gestation).27 The relatively weak relationship between gestational age and average marks in our study could be attributed to a fairly uncompetitive approach to teaching and learning in basic schools in Denmark.
Completing compulsory basic school requirements in Denmark is a milestone that is necessary to progress in the educational system and thereby obtain professional qualifications; therefore, it has importance as a clinically significant outcome. Our findings correspond with 2 recent register-based studies from Scandinavia that investigated the social integration of premature-born people as adults. These studies revealed a similar association between preterm birth and educational level as adults. The negative effect increased with decreasing gestational age, and, as in our study, the effect was moderate but statistically significant for people who were born moderately preterm.28,29
Moderate preterm birth is associated with school difficulties that persist through basic school. The risk of not completing basic school increased with decreasing gestational age. The absolute increase was minimal at more than 32 weeks' gestation, but at less than 31 weeks' gestation it has practical importance. It is only explained partly by cerebral palsy.
This work was financially supported by the Sara and Ludvig Elsass Foundation, the Dagmar Marshall Foundation, the Hartmanns Foundation, the A. P. Møller and Chastine Mc-Kinney Møllers Foundation, and the Aase og Ejnar Danielsens Foundation.
- Accepted August 13, 2010.
- Address correspondence to Rene Mathiasen, MD, PhD, Department of Neonatology, Copenhagen University Hospital, Rigshospitalet, Blegdamsvej 9, 2100 Copenhagen Ø, Denmark. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- SGA =
- small for gestational age •
- ISCED =
- International Standard Classification of Education •
- ICD =
- International Classification of Diseases •
- OR =
- odds ratio •
- CI =
- confidence interval
- Litt J,
- Taylor HG,
- Klein N,
- Hack M
- Saigal S,
- Hoult LA,
- Streiner DL,
- Stoskopf BL,
- Rosenbaum PL
- Langhoff-Roos J,
- Kesmodel U,
- Jacobsson B,
- Rasmussen S,
- Vogel I
- Hamilton BE,
- Miniño AM,
- Martin JA,
- Kochanek KD,
- Strobino DM,
- Guyer B
- Huddy CLJ,
- Johnson A,
- Hope PL
- Kirkegaard I,
- Obel C,
- Hedegaard M,
- Henriksen TB
- Shenkin SD,
- Starr JM,
- Pattie A,
- Rush MA,
- Whalley LJ,
- Deary IJ
- Elgen I,
- Sommerfelt K,
- Markestad T
- Pinto-Martin J,
- Whitaker A,
- Feldman J,
- et al
- Hutton JL,
- Pharoah POD,
- Cooke RWI,
- Stevenson RC
- Nomura Y,
- Halperin JM,
- Newcorn JH
- Woodward M
- Lindstrom K,
- Winbladh B,
- Haglund B,
- Hjern A
- Copyright © 2010 by the American Academy of Pediatrics