SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

Aim  To investigate whether infants born late preterm have poorer cognitive outcomes than term-born infants.

Method  A cohort study based on the Avon Longitudinal Study of Parents and Children. Cognitive measures were assessed between the ages of 8 and 11 years. Exposure groups were defined as moderate/late preterm (32–36 weeks’ gestation) or term (37–42wk). Regression models were used to investigate the association between gestational age and IQ.

Results  Seven hundred and forty-one infants (5.4% of total eligible population; 422 males, 319 females; mean (SD) birthweight 2495g [489]) were born between 32 and 36 weeks’ gestation. The analysis was based on 6957 infants with IQ data at age 11 (50% of eligible infants). In the adjusted model, children born moderately and late preterm had similar IQ scores to peers born at term (mean difference [95% confidence interval] −0.18 [−1.88 to 1.52]). However, the preterm infants had a higher risk of having special educational needs at school (odds ratio 1.56 [1.18–2.07]).

Interpretation  Despite an increased risk of special educational needs, there is little evidence of a reduction in IQ, memory, or attention measures at school age in children born between 32 and 36 weeks’ gestation. Although interpretation is limited by the amount of missing data, further work is needed to identify why these infants have increased educational needs.


Abbreviation
ALSPAC

Avon Longitudinal Study of Parents and Children

What this paper adds

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information
  •  Moderate and late preterm infants have increased risk of educational needs.
  •  Moderate and late preterm infants have similar measures of memory, attention, and IQ compared with term peers.
  •  In this cohort, the association between moderate and late preterm birth and educational needs does not appear to be mediated through poorer IQ measures.

The long-term effects of preterm birth are well documented in infants born before 32 weeks’ corrected gestation.1 Extreme preterm birth is associated with increased risks of cerebral palsy,2 impairments in executive functioning,3 as well as reduction in more global measures of cognition,1 resulting in poorer educational achievement.4 However, the long-term outcome of infants born only moderately preterm is less clear. Previous studies have suggested increased rates of educational difficulties in infants born before 37 weeks’ gestation5–7 and there is some evidence that measures of cognition may also be worse8. At present these infants receive little specific educational input or formal neurodevelopmental follow-up, unlike infants born at earlier gestations, and represent a much larger population than those born at earlier gestations.5 Recently the definitions of moderate (32–33 weeks’ gestation) and late (34–36wk) preterm birth have been used to identify this group of infants.

The aim of the present study was to investigate whether infants born at moderate and late preterm gestations (32–36wk) have poorer cognitive, memory, attention, or school outcomes in childhood than those born at term.

Method

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

This study is based on the Avon Longitudinal Study of Parents and Children (ALSPAC), an ongoing longitudinal study containing data on over 14 000 infants.9 The cohort includes children born in the Bristol area of the UK, from April 1991 to December 1992. Further information about the study can be found on the ALSPAC website (http://www.alspac.bristol.ac.uk).

Data on gestational age were routinely recorded in the clinical notes. If the gestation was recorded as less than 37 weeks (based on last menstrual period, ultrasound, or paediatric assessment), then gestational age was confirmed by a single paediatrician after reviewing the clinical records.

Cognitive functions were assessed at annual research clinics at 8 to 11 years of age. Parents and children were invited to attend by letter. If there was no response a postal and then a telephone/personal contact was attempted, and a final letter was sent if there was still no response after 3 months. The primary outcome was IQ, assessed when study participants were aged 8 years using a shortened version of the Wechsler Intelligence Scale for Children, 3rd edition (WISC-III).10

Short-term memory was tested using the non-word repetition test11 at the same clinic, whereas working memory was assessed at the 11-year clinic using the digit span-score.12 Attention skills were also assessed at the 8-year clinic using the Sky Search Test, and Dual Attention tasks taken from the Test of Everyday Attention for Children.13 Finally the tests of reading skills were derived from the Neale Analysis of Reading Ability14 at the 9-year clinic. The Neale Analysis of Reading Ability is a test of oral reading based on a series of short narratives. Details of these tests are available on request.

At the age of 8 years, the child’s teacher was sent a standardized questionnaire which asked, ‘Has this child ever been recognized as having special educational needs?’

The following social and perinatal factors were recorded for the infants. Social factors included maternal age, socio-economic group15 and education, car ownership, housing, crowding index (number of household members per room), and ethnicity. Perinatal factors were sex, parity, birthweight, length and head circumference, mode of delivery, and maternal hypertension and pyrexia.

Statistical analysis

The dataset contained information on 13 843 infants born alive at between 32 weeks and 42 completed weeks of gestation. Exposure groups were defined as preterm (32–36wk) or term (37–42wk).

Linear regression models were used to investigate the association between gestational group and the cognitive measures. In the models, gestational group was used as the categories defined above (i.e. a binary variable with preterm=1). All cognitive measures were standardized within the dataset to a mean of 100 and standard deviation of 15, with higher scores indicating better performance. Dual attention scores from the Test of Everyday Attention for Children required log transformation to give a normally distributed measure. The models used random effects to adjust for possible clustering of data within multiple births. Adjustment for possible confounders was performed by adding the covariates to the models, in blocks of common variables (e.g. socio-economic factors).

As not all children completed all parts of the testing, each analysis contained a different number of participants, with 6382 children having complete IQ data, 653 having partial data, and 6808 having no IQ data. Thus, we report data from 6957 children for summary IQ score, and 6132 with data on special educational needs. A total of 3459 infants had no data on any outcome and were therefore not included in any analysis.

Only 8878 children had complete data on all confounder variables. To minimize any potential selection bias in the multivariable models a multiple imputation data technique (chained equations) was used to impute the missing covariate data, allowing us to report on the same number of children for crude and adjusted analyses.16 For binary outcomes, logistic regression models were derived in the same way as the linear regression models, and the population attributable risk fraction was calculated from the final model.17 Owing to previous work suggesting modification of an association with socio-economic status,18 maternal educational and socio-economic groups were tested to see if they modified the relationship between moderate and late preterm birth and IQ and special educational needs.

Several sensitivity analyses were performed including (1) using only those infants with complete data, rather than the imputed dataset; (2) restricting the analysis to presumed well infants (in contrast to all infants): those born by spontaneous vaginal delivery, who did not require resuscitation after birth, and who were well grown (within 1SD of the mean birthweight for their gestation); (3) splitting the preterm group into moderate and late preterm gestations; and (4) using IQ as a binary outcome (<80 or 75). In one final analysis the association between gestation and special educational need was repeated, including the cognitive measures as covariates.

All analyses were conducted with Stata 10 (StataCorp, College Station, TX, USA). All data are presented as odds ratio (OR) (95% confidence interval [CI]), mean (SD), mean difference (95% CI), median (interquartile range), or number (%). Univariable comparisons were made with the χ2, Student’s t-test, or Mann–Whitney U test as appropriate. Ethical approval for the study was obtained from the ALSPAC Law and Ethics Committee and the local research ethics committees.

Results

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

Sample

Table I shows the characteristics of the eligible infants (n=13 843). Seven hundred and forty-one (5.4%) infants were classified as late or moderately preterm (422 males, 319 females; mean [SD] birthweight 2495g [489]). The mean gestation (SD) in the preterm infants was 35.0 completed weeks (1.2) whereas in the reference group (51.4% males; mean [SD] birthweight 3456g [485]) was 39.7 (1.3). A breakdown of the gestational groups is shown in Appendix SI (supplementary material published online only). Data were complete only for some covariates, and consequently the denominator for different measures varies.

Table I.   Characteristics of study population, n (%)
MeasureNumber with dataTerm infants (n=13 102)Moderate/late preterm infants, (n=741)p
  1. SDs are given for means of normally distributed continuous variables, and percentages for proportions. aCSE, Certificate in Secondary Education (commonly taken at 16y of age); vocational, City & Guilds (intermediate level), technical, shorthand or typing, or other qualification; O level, Ordinary level (commonly taken at 16y of age); A level, Advanced level (commonly taken at 18y of age), state enrolled nurse, state registered nurse, City & Guilds (final or full level), or teaching qualification; degree, university degree.

Pre-pregnancy factors
 Maternal age13 84328y 0mo27y 8mo0.088
 Maternal socio-economic group11 435   
  Professional 659 (6.1)27 (4.6) 
  Managerial 3362 (31.0)177 (30.2) 
  Skilled non-manual 4152 (38.3)236 (40.3) 
  Skilled manual 1267 (11.7)79 (13.5) 
  Semi-skilled 1145 (10.6)53 (9.0) 
  Unskilled 264 (2.4)14 (2.4)0.365
 Mother’s highest educational qualificationa12 321   
  CSE 2316 (19.8)163 (26.0) 
  Vocational 1148 (9.8)62 (9.9) 
  O level 4071 (34.8)206 (32.9) 
  A level 2635 (22.5)133 (21.2) 
  Degree 1524 (13.0)63 (10.1)0.002
 Housing12 903   
  Mortgaged or owned 8988 (73.6)488 (70.7) 
  Rented from municipality 1721 (14.1)119 (17.3) 
  Private rented 1504 (12.3)83 (12.0)0.069
 Car ownership12 91110929 (89.4)596 (86.8)0.029
 Crowding index (people per room)12 708   
  <0.5 4998 (41.5)288 (43.1) 
  0.5–0.75 6223 (51.7)332 (49.7) 
  0.75–1 576 (4.8)33 (4.9) 
  1+ 243 (2.0)15 (2.3)0.785
 Non-white ethnicity13 538692 (5.4)64 (8.9)<0.001
Antenatal and intrapartum factors
 Primiparous12 8175342 (44.0)323 (47.0)0.102
 Maternal hypertension12 577350 (3.0)31 (4.4)0.026
 Maternal pyrexia12 57765 (0.6)5 (0.7)0.562
 Multiple birth13 843202 (1.5)137 (18.5)<0.001
 Delivery12 575   
  Spontaneous cephalic 9033 (76.1)421 (60.2) 
  Emergency Caesarean section 700 (5.9)144 (20.6) 
  Elective Caesarean section 495 (4.2)36 (5.2) 
  Instrumental 1485 (12.5)67 (9.6) 
  Breech 163 (1.4)31 (4.4%)<0.001
Infants and postpartum factors
 Male13 8416736 (51.4)422 (57.0%)0.003
 Birthweight (g), mean (SD)13 7933456 (485)2495 (489)<0.001
 Birth length (cm), mean (SD)11 29250.9 (2.4)47.2 (2.5)<0.001
 Head circumference (cm), mean (SD)11 94534.9 (1.4)32.7 (1.8)<0.001
 Apgar at 1min, mean (SD)12 5558.4 (1.4)7.9 (1.8)<0.001
 Apgar at 5min, mean (SD)12 5419.5 (0.7)9.3 (1.0)<0.001
 Received resuscitation12 577945 (7.8)136 (19.5%)<0.001
 Died before 8y of age13 84352 (0.4)11 (1.5%)<0.001

Moderate or late preterm infants were more likely to have mothers with lower levels of educational achievement (p=0.002), be male (p=0.003), had lower birthweights, lengths, and head circumferences, with lower Apgar scores at 1 and 5 minutes, and were more likely to have received resuscitation at birth than term infants (all other comparisons, p<0.001). Their profile of delivery differed (p<0.001), in particular the preterm infants were more likely to be born by emergency Caesarean section (20.6% vs 5.9%) and were more likely to die before 8 years of age (1.5% vs 0.4%, p<0.001).

Missing data

In common with all longitudinal studies, some of the original cohort did not attend follow-up. A total of 6886 children (49.7%) did not have a recorded summary IQ score at 8 years; these infants were more likely to have needed resuscitation (9.2% vs 8.1%, p=0.033), were smaller (e.g. birthweight; 3379g vs 3340g, p<0.001), and they differed on most socio-economic measures (e.g. were born to younger mothers; 26y 10mo vs 29y 2mo, p<0.001) than those included in the main analysis. However, there was no evidence that children with missing outcome data were more likely to be preterm (5.6% vs 5.1%, p=0.165). Differences between children with and without summary IQ scores are reported in Appendix SII (supplementary material published online only). Children born preterm were also no more likely than their term peers to have missing data on special educational needs (422 [57.0%] vs 7289 [55.6%], p=0.482).

Cognitive and educational outcomes

In the univariable analysis, preterm infants had slightly lower verbal (p=0.081), performance (p=0.087), and summary IQ scores (p=0.059) than term infants (Table II).

Table II.   Mean (SD) IQ scores split by gestational age
MeasureNumber with data (% of eligible)Term infantsModerate/late preterm infantsp
IQ measures
 Verbal IQ6897 (50)107 (17)106 (17)0.081
 Performance IQ6890 (50)100 (17)98 (17)0.087
 Summary IQ6957 (50)104 (16)103 (16)0.059
Tests of memory
 Non-word repetition6970 (50)100 (15)98 (14)0.003
 Span score6601 (48)100 (15)99 (14)0.123
Tests of attention
 Sky Search Attention6797 (49)100 (15)98 (15)0.025
 Dual Attention score5254 (38)100 (15)99 (15)0.099
Tests of reading
 Accuracy6523 (47)100 (15)97 (16)0.001
 Read per minute6507 (47)100 (15)98 (16)0.015
 Comprehension6523 (47)100 (15)98 (15)0.003

In the unadjusted model there was only weak evidence that infants born preterm had lower IQ scores than those born at term (mean IQ difference -1.38 [−3.20 to 0.44], p=0.137) and this association attenuated further after correction for socio-economic factors and in the fully adjusted model (−0.18 [−1.88 to 1.52], p=0.835) (Table III). Repeating the analysis using the subtests produced similar results.

Table III.   Difference in mean (95% CI) Wechsler Intelligence Scale for Children scores for moderate/late preterm compared with term infants
MeasureUnadjusted mean differencepAdjusted for social factorsapFully adjusted mean differencea,bp
  1. Data are mean difference (95% confidence interval [CI]) in scores between late preterm and term infants derived from linear regression models. aAdjusted for ethnicity, housing, crowding, maternal education, socio-economic group, car ownership, and age. bFurther adjusted for sex, parity, weight, length and head circumference at birth, mode of delivery, and maternal hypertension and pyrexia.

Verbal subtest
 Information (n=7017)−0.43 (−0.77 to −0.08)0.015−0.37 (−0.69 to −0.05)0.024−0.29 (−0.63 to 0.04)0.082
 Arithmetic (n=7002)−0.47 (−0.92 to −0.02)0.042−0.33 (−0.77 to 0.11)0.140−0.22 (−0.67 to 0.24)0.347
 Vocabulary (n=6984)−0.33 (−0.81 to 0.15)0.179−0.17 (−0.62 to 0.27)0.446−0.04 (−0.50 to 0.42)0.869
 Comprehension (n=6943)0.04 (−0.37 to 0.44)0.8610.07 (−0.32 to 0.47)0.7160.13 (−0.28 to 0.54)0.543
 Similarities (n=7017)0.00 (−0.44 to 0.44)0.9940.09 (−0.33 to 0.50)0.6770.18 (−0.25 to 0.61)0.416
Performance subtest
 Picture completion (n=6993)−0.49 (−0.44 to 0.35)0.8080.02 (−0.37 to 0.41)0.9230.26 (−0.14 to 0.67)0.206
 Coding (n=7010)−0.37 (−0.70 to −0.04)0.0270.31 (−0.64 to 0.01)0.056−0.22 (−0.55 to 0.10)0.175
 Picture arrangement (n=6919)−0.25 (−0.77 to 0.26)0.338−0.19 (−0.71 to 0.32)0.4620.00 (−0.53 to 0.53)0.998
 Block design (n=6970)−0.19 (−0.61 to 0.22)0.360−0.10 (−0.50 to 0.31)0.618−0.03 (−0.45 to 0.38)0.884
 Object assembly (n=6608)−0.25 (−0.68 to 0.17)0.238−0.19 (−0.61 to 0.22)0.363−0.12 (−0.55 to 0.31)0.575
Summary scores
 Verbal IQ (n=6987)−1.43 (−3.27 to 0.42)0.129−0.83 (−2.5 to 0.85)0.333−0.33 (−2.06 to 1.40)0.708
 Performance IQ (n=6890)−1.33 (−3.20 to 0.54)0.163−0.95 (−2.75 to 0.84)0.297−0.12 (−1.96 to 1.73)0.900
 Summary IQ (n=6957)−1.38 (−3.20 to 0.44)0.137−0.88 (−2.54 to 0.78)0.297−0.18 (−1.88 to 1.52)0.835

Preterm infants also had worse scores for short-term memory (p=0.003), but similar scores for working memory (p=0.123). They performed worse on all three tests of reading, and had lower scores on the Sky Search test (p=0.025) with less evidence for a lower score in the Dual Attention task (p=0.099).

In the unadjusted regression model, preterm infants had lower scores in one of the memory domains (non-word repetition; −2.19 [−3.83 to −0.56], p=0.009), one of the attention measures (Sky Search; −1.85 [−3.52 to −0.17], p=0.030), and all three of the reading measures (Table IV). All measures were attenuated after adjustment for confounders, leaving weak evidence that there was a lower score in non-word repetition (−1.55 [−3.18 to 0.09], p=0.064) and stronger evidence for a lower reading accuracy score (−1.97 [−3.61 to −0.32], p=0.019).

Table IV.   Linear regression difference in memory, attention, and reading scores for moderate/late preterm compared with term infants
  1. Data are mean difference (95% confidence interval [CI]) in scores between late preterm and term infants. aAdjusted for ethnicity, housing, crowding, maternal education, socio-economic group, car ownership, and age. bAdjusted for sex, parity, weight, length and head circumference at birth, mode of delivery, and maternal hypertension and pyrexia.

MeasureUnadjusted mean differencepAdjusted for social factors)apFully adjusted mean difference)a,bp
Tests of memory
 Non-word repetition (n=6970)−2.19 (−3.83 to −0.56)0.009−2.00 (−3.58 to −0.41)0.014−1.55 (−3.18 to 0.09)0.064
 Span score (n=6601)−1.33 (−3.03 to 0.37)0.125−1.04 (−2.71 to 0.62)0.220−1.04 (−2.77 to 0.68)0.236
Tests of attention
 Sky Search Attention (n=6797)−1.85 (−3.52 to −0.17)0.030−1.84 (−3.53 to −0.15)0.032−1.29 (−3.00 to 0.42)0.138
 Dual Attention score (n=5254)−1.52 (−3.37 to 0.33)0.106−1.32 (−3.17 to 0.53)0.161−1.12 (−3.03 to 0.79)0.250
Tests of reading
 Accuracy (n=6523)−2.96 (−4.67 to −1.25)0.001−2.41 (−4.01 to −0.80)0.003−1.97 (−3.61 to −0.32)0.019
 Read per minute (n=6507)−2.03 (−3.74 to −0.33)0.019−1.64 (−3.26 to −0.03)0.046−1.18 (−2.84 to 0.47)0.162
 Comprehension (n=6523)−2.32 (−4.03 to −0.61)0.008−1.71 (−3.27 to −0.15)0.032−0.98 (−2.58 to 0.61)0.227

Preterm infants were more likely to have special educational needs identified (110 [34.5%] vs 1355 [23%], p<0.001). There was strong evidence in all models that infants born preterm remained more likely to have special educational needs than term infants (unadjusted: OR 1.76 [1.38–2.30], p<0.001; fully adjusted: OR 1.56 [1.18–2.07], p=0.002). The population attributable risk fraction was 1.7% (0.6–2.8%).

There was little evidence to suggest that maternal socio-economic status (pinteraction=0.397) or maternal education (pinteraction=0.407) modified any relationship between gestational age and IQ or special educational needs (pinteraction= 0.199 and pinteraction=0.468 respectively). Restricting the analysis to those infants with complete data (n=5228, mean IQ difference −1.47 [−3.46 to 0.51], p=0.147) or to those infants born by spontaneous vaginal delivery, who did not require resuscitation after birth and who were well grown (n=3494, mean IQ difference 0.64 [−2.01 to 3.30], p=0.635), produced similar results to the main analysis. Splitting the preterm group into moderate (n=6655, mean IQ difference 1.52 [−0.60 to 3.64], p=0.147) and late (n=6905, mean IQ difference −0.74 [−2.57 to 1.09], p=0.430) preterm infants produced similar, if less precise, results. There was little evidence that infants born preterm were more likely to have an IQ score below 70 (OR 1.24 [0.60–2.57, p=0.570] or 85 (OR 0.86 [0.60–1.24], p=0.432) than term peers. Repeating the analysis adjusting for the cognitive measures did not appreciably attenuate the association between preterm status and having special educational needs (OR 1.45 [1.05–2.02], p=0.025).

Discussion

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

Using a prospective cohort, we have found that infants born at moderate and late preterm (32–36 weeks gestational age) needed extra educational support at primary school than their term peers. Although we have found little evidence of an association with IQ, memory, or attention to explain this association, we did see a small effect in one domain of reading (although results remained in the ‘normal’ range) and evidence for an increasing risk of a poor score in two other areas of reading. Although the mothers of the preterm infants had lower educational qualifications than those of the term group, maternal socio-economic position or education did not appear to interact with the association between gestation and IQ.

An association between moderate and late preterm birth and reduced educational achievement has been reported in several other papers.5–7,18 However, it is important to consider the causal routes investigated and the assumptions made in the modelling. Any association between preterm birth and long-term cognitive deficits could well be due to the causes of preterm birth rather than any consequence of it. Although we were able to control for several common prenatal pathologies (e.g. multiple births, poor in utero growth), some degree of residual or uncontrolled confounding is possible. In particular specific parental pathology (e.g. mental health problems) may well be correlated with both gestation at birth and cognitive outcomes. Furthermore, we have made assumptions around the confounding nature of certain measures (e.g. growth centiles), although the any causal pathway between gestation and cognition is likely to be complex.

The weak association found with IQ was largely attenuated by adjusting for socio-economic measures that were likely to represent patterning of health behaviours. Further weak associations were seen with reading, memory, and attention measures, which may explain some of the increased educational needs, but the effects seen were small (all less than a quarter of a standard deviation difference). However, the cognitive measures available here did not seem to appreciably attenuate the association between gestation and special educational needs. The reason for the discrepancy between special educational needs without obvious deficits in cognition in this cohort is unclear. It is possible that other background factors may be related to the need for extra educational support in the moderate or late preterm group (but not mediated through cognition), including the mother’s own educational level and qualifications, ethnic background, and the higher number of multiple births. Another explanation is that multiple small cognitive deficits, difficult to measure on single tests, exist in these infants. Individually these deficits may be difficult to quantify, but together with socio-economic factors they have a cumulative effect resulting in school failure in later life. Alternatively other, unmeasured, subtle deficits in information processing and learning may underlie the educational problems.

Although the calculated population impact may appear low (1.7%), given the many other factors likely to impact on an individual’s need for special education (including the factors corrected for in the analysis like socio-economic patterns and intra-uterine growth restriction), and the relative rarity of the exposure, it is perhaps surprising that in nearly 2% of children who require special education, the need may be due to moderate or late preterm birth.

The proportion of infants born between 32 and 36 weeks’ gestation (5.4%) appears consistent with other recent work (6.6%7). However, although ALSPAC is a relatively large study, the study power was reduced by the poor uptake of IQ testing at age 8. Despite this limitation the point estimates derived were fairly precise, and it would seem unlikely that important differences in IQ existed that we were unable to detect. A potential limitation is the use of a short version of the WISC-III, and the use of multiple testers, but a validation undertaken showed acceptable correlations with the full test. A further limitation was that the assessment of special education needs was undertaken by questionnaire (completed by the child’s teacher), although the results are consistent with other recent work.5–7,18

Selection bias from missing data is an important limitation of any interpretation. Most outcome measures were only available on 50% of the eligible cohort although there was no evidence that infants born preterm were less likely to attend the IQ testing at 8 years or have a data on their special educational needs status. It is felt unlikely that this would produce important selection bias for this group of infants, but a multiple imputation technique was used to try to reduce any impact of missing confounders, and restricting the analysis to infants with complete data produced similar results to the main analysis. Although 75% of infants were included in at least one of the analyses reported here, the large amount of missing outcome data must be acknowledged in any interpretation of this work.

However, although there is little evidence to support differential rates of follow-up between the preterm and term groups, the demographics of infants in the preterm and term groups differed, and it is difficult to assess the impact that this complex patterning will have had on the results. Consequently the interpretation, and generalization of these results, should be undertaken with caution.

The weakness of the evidence for an association between late and moderate preterm birth and cognition was unexpected, as previous studies had found a robust association with poor educational achievement. One study that was able to report measures of IQ at 7 years of age18 found lower IQ scores in late and moderate preterm (33–37 weeks’ gestation) compared with term infants (>37wk) (88.9 vs 93.5), although the effect appeared restricted to infants of lower socio-economic status. A further study8 has also suggested worse cognitive skills at a conscription examination in adults who were born preterm. However, in both these studies the infants were born over 30 years ago and therefore may not represent the outcomes of late preterm infants born more recently, particularly in the era of antenatal steroids and surfactant therapies.

Morse et al. 6, reported on over 7000 late preterm infants born in Florida in the mid-1990s, and showed a strong association between kindergarten education performance and birth between 34 and 36 weeks, whereas Lindström et al. 5, reported increased risks of disability requiring economic assistance, worse educational achievement, and lower net salary in individuals who had been born between 33 and 36 weeks’ gestation. As discussed above, the weak associations with cognitive measures in our study are unlikely to be explained simply by chance. However, even the most evident association (with reading accuracy) was small and it is likely that educational measures are more socially influenced than cognitive measures.19 An important predictor of school performance may be age at school entry, with preterm infants more likely to attend school a year earlier than if they had been born on their due date (e.g. approximately 15% of 32-week infants would be expected attend school a year ‘early’). Indeed previous work with extremely preterm infants has suggested that a significant proportion of school failure may be due to this.20

Conclusions

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

Although mortality was higher, there was little evidence of a reduction in IQ in children born moderately and late preterm compared with their peers born at term. The preterm infants had slightly lower scores in several measures of reading, memory, and attention, although the effect sizes seen were small and only one (reading accuracy) reached conventional levels of statistical significance. Further work is needed to identify why infants born between 32 and 36 weeks’ gestation appear to have increased educational needs.

Acknowledgements

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information

We are grateful to all the families who took part in this study, the midwives for their help in recruiting them, and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists, and nurses. The UK Medical Research Council, the Wellcome Trust, and the University of Bristol provide core support for ALSPAC.

References

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information
  • 1
    Wood NS, Marlow N, Costeloe K, Gibson AT, Wilkinson AR. Neurologic and developmental disability after extremely preterm birth EPICure Study Group. N Engl J Med 2000; 343: 37884.
  • 2
    Marlow N, Wolke D, Bracewell MA, Samara M. Neurologic and developmental disability at six years of age after extremely preterm birth. N Engl J Med 2005; 352: 919.
  • 3
    Marlow N, Hennessy EM, Bracewell MA, Wolke D, Group EPS. Motor and executive function at 6 years of age after extremely preterm birth. Pediatrics 2007; 120: 793804.
  • 4
    Wolke D, Samara M, Bracewell M, Marlow N, Group EPS. Specific language difficulties and school achievement in children born at 25 weeks of gestation or less. J Pediatr 2008; 152: 25662.
  • 5
    Lindstrom K, Winbladh B, Haglund B, Hjern A. Preterm infants as young adults: a Swedish National Cohort study. Pediatrics 2007; 120: 707.
  • 6
    Morse SB, Zheng H, Tang Y, Roth J. Early school-age outcomes of late preterm infants. Pediatrics 2009; 123: e6229.
  • 7
    Chyi LJ, Lee HC, Hintz SR, Gould JB, Sutcliffe TL. School outcomes of late preterm infants: special needs and challenges for infants born at 32 to 36 weeks gestation. J Pediatr 2008; 153: 2531.
  • 8
    Ekeus C, Lindström K, Lindblad F, Rasmussen F, Hjern A. Preterm birth, social disadvantage, and cognitive competence in Swedish 18- to 19-year-old men. Pediatrics 2010; 125: e3773.
  • 9
    Golding J, Pembrey M, Jones R. ALSPAC-the Avon Longitudinal Study of Parents and Children I. Study methodology. Paediatr Perinat Epidemiol 2001; 15: 7487.
  • 10
    Wechsler D. Manual for the Wechsler Intelligence Scale for Children. 3rd edn. San Antonio, TX: The Psychological Corporation, 1991.
  • 11
    Gathercole SE, Baddeley AD. The Children’s Test of Non-word Repetition. Sidcup, UK: The Psychological Corporation, 1996.
  • 12
    Case R, Kurland D, Goldberg J. Operational efficiency and the growth of short-term-memory span. J Exp Child Psychol 1982; 33: 386404.
  • 13
    Manly T, Anderson V, Nimmo-Smith I, Turner A, Watson P, Robertson IH. The differential assessment of children’s attention: the Test of Everyday Attention for Children (TEA-Ch), normative sample and ADHD performance. J Child Psychol Psychiatry 2001; 42: 106581.
  • 14
    Neale MD. Neale Analysis of Reading Ability. 2nd edn. UK; NFER-Nelson Windsor, 1997.
  • 15
    HMSO. Standard Occupational Classification. London, UK: HMSO, 1991.
  • 16
    Royston P. Multiple imputation of missing values: update. Stata J 2005; 5: 52736.
  • 17
    Greenland S. Modeling and variable selection in epidemiologic analysis. Am J Public Health 1989; 79: 3409.
  • 18
    Nomura Y, Halperin JM, Newcorn JH, et al. The risk for impaired learning-related abilities in childhood and educational attainment among adults born near-term. J Pediatr Psychol 2009; 34: 40618.
  • 19
    Odd DE, Rasmussen F, Gunnell D, Lewis G, Whitelaw A. A cohort study of low Apgar scores and cognitive outcomes. Arch Dis Child Fetal Neonatal Ed 2008; 93: F11520.
  • 20
    Johnson S, Hennessy E, Smith R, Trikic R, Wolke D, Marlow N. Academic attainment and special educational needs in extremely preterm children at 11 years of age: the EPICure study. Arch Dis Child Fetal Neonatal Ed 2009; 94: F2839.

Supporting Information

  1. Top of page
  2. Abstract
  3. What this paper adds
  4. Method
  5. Results
  6. Discussion
  7. Conclusions
  8. Acknowledgements
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
DMCN_4315_sm_AppendixS1-S2.pdf38KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.