Objective. To investigate the relation between school difficulties and being born small for gestational age (SGA) at full term in adolescents and young adults.
Methods. A total of 236 full-term singletons who were born SGA (birth weight and/or length below the third percentile) from 1971 through 1978 and 281 full-term singletons who were born appropriate for gestational age (AGA; between the 25th and 75th percentiles) from the maternity registry of Haguenau, France. Participants were evaluated at a mean age of 20.6 (±2.1) years. The outcomes measured were late entry into secondary school (normal age: 11 years) and failure to take or pass the baccalaureate examination at the end of secondary school (normal age: 18 years).
Results. Late entry into secondary school was more frequent for the SGA than the AGA children (odds ratio: 2.3) after adjustment for maternal age and educational level, parental socioeconomic status, family size, and gender. A significantly higher proportion of term SGA adolescents failed to take or pass the baccalaureate examination than AGA adolescents (odds ratio: 1.6). SGA participants with a smaller head circumference entered secondary school late more often than SGA participants with a larger head circumference, but the association was not significant after adjustment.
Conclusion. Being born SGA at term is associated with poorer school performance at 12 and 18 years. Fetal adaptation to conditions that retard growth during gestation may not be successful in maintaining brain development.
Intrauterine growth retardation (IUGR) is associated with increased perinatal mortality and morbidity1 and postnatal growth failure2; it also contributes to such late-onset disorders as cardiovascular disease, insulin resistance, and non–insulin-dependent diabetes.2,,3 The metabolic adaptations made by a fetus that is undernourished in utero may program these later diseases.4 IUGR may be related to identifiable genetic or toxic factors or to reduced nutrient or oxygen delivery to the fetus, mostly related to placental dysfunction.5Experimental data show that undernutrition at vulnerable periods during brain development may have long-term effects on brain architecture and differentiation that can affect learning and memory.6,,7
Studies of neurobehavioral and cognitive outcomes in several groups of young school children born small for gestational age (SGA) have yielded conflicting results.8 Some found no difference with participants who were appropriate for gestational age (AGA) for cognitive9,,10 or school performance indicators,11 whereas others reported minor neurologic dysfunction12 or cognitive deficits.13 There are limited studies from the recent era on follow-up of adolescents or adults born SGA at term; these observed lower educational achievement14 and intelligence scores15 than among control participants. In a recent well-conducted study with a high number of SGA participants, Strauss16 observed lower school performance at 16 years and lower income or fewer professional or managerial jobs at 26 years in SGA than in AGA patients. Nonetheless, other studies of cognitive development among term SGA children often involved small sample size and lack of comparison group and did not always control adequately for social factors that may influence test results. This study aimed to investigate whether term SGA adolescents were at risk of impaired school performance during childhood and adolescence.
Participants were identified from a population-based birth registry in and around the city of Haguenau, France. This registry recorded information on all pregnancies, deliveries, and perinatal events from 1971 to 1985.17 Gestational age was determined from the date of the last menstrual period and by physical examination during pregnancy and was confirmed by ultrasound measurements when available. Local standard growth curves by gestational age and gender were derived from all live births registered (nearly 20 000 births). The growth standards of the population of Haguenau are different from those of the population of France in general because of persons of northern origin who are taller: comparing with reference curves of Leroy and Lefort18 established in the Paris region in 1971 and commonly used in France, the 50th percentile was higher by approximately 100 g for Haguenau's infants born at ≥37 weeks. SGA was defined as birth weight or length below the third percentile of these curves. All full-term (≥37 weeks' gestation) singleton patients who were born SGA from 1971 to 1978 (N = 452) were selected.2 The control participants were 451 full-term singletons who were (AGA; birth weight between the 25th and 75th percentiles). The first AGA infant listed in the registry after each SGA birth was selected as a control.
Sixteen of the 903 patients were excluded because of aberrant growth measurements at birth, and 10 were excluded because of adoption; 41 died, 32 (7%) in the SGA group and 9 (2%) in AGA group. Of the 836 survivors who could be included, therefore, 517 (62%) participated in the study (236 SGA and 281 AGA). Participants were contacted by letter at the address of the parents' home or a more recent address known by the hospital. All participants (16–24 years old) underwent detailed clinical investigations and completed a questionnaire, as previously described.2
The questionnaire included information on schooling, from which we selected 2 outcome measures. The first was age at starting secondary school (normally 11). Participants who were older (≥12 years) had repeated at least 1 grade in primary school (from 6 to 11 years). For this analysis of academic results, we excluded 10 adolescents who had psychomotor handicaps or were institutionalised (all were SGA) and 10 who had severe chronic illnesses or malformations that might have interfered with school performance. Among them, 8 born SGA patients were excluded for malformation of inner ear and vertebral malformation (1), dwarfism (1), congenital heart defects (3), adrenocortical hyperfunction (1), fibrocystic disease (1), and Gilbert disease (1), and 2 born AGA were excluded for malformation of urinary system (1) and medullar aplasia (1). Age at entry into secondary school was known for the 218 who were born SGA and 279 who were born AGA.
The second outcome measurement was the final examination of secondary school, the baccalaureate. This diploma is not obtained by adolescents who failed it and those who took a short vocational program. The normal age for the baccalaureate is approximately 18 years. The mean (standard deviation: SD) age of the population at the follow-up was 20.6 ± 2.1 years (range: 16–24 years). We analyzed baccalaureate success only among participants who were at least 19 years old.
To evaluate the severity of growth retardation, birth weight, length, and head circumference were calculated as SD scores corrected for gestational age and gender. We evaluated intrauterine nutritional status with the ponderal index (PI) [birth weight in g/(birth length in cm)3] × 100, corrected for gestational age, according to Miller and Hassanein.19
Participation was lower among the SGA (57%) than the AGA group (64%;P = .04). SGA participants and nonparticipants did not differ for birth weight or length, head circumference, or PI. Within the SGA group, participation was lower among those whose mothers were unmarried or smoked during pregnancy and those whose fathers had a low educational level or socioeconomic status. Among the AGA group, the participation was lower among males and among children whose mothers were born abroad.
The study protocol was reviewed and approved by the faculty ethics committee, and all participants and parents gave signed written consent.
SGA and AGA participants were compared for both school indicators, means were compared with the 2-tailed t test, and differences in proportions were compared with χ2 tests. The associations between SGA status and school indicators were estimated as adjusted odds ratios and their 95% confidence intervals (CIs), with unconditional logistic regression used to control for confounders. Confounders selected were maternal age at birth (<25, ≥25 years), maternal educational level (missing, primary or less, higher), gender, number of children in the family (<3, ≥3), parents' socioeconomic status (highest of mother's or father's occupation in 1994–1995: low, high, no occupation). We also examined the outcome variables among SGA patients by symmetry of growth retardation (asymmetric PI <third percentile; symmetric PI ≥third percentile) and by the severity of growth retardation for birth weight, length, and head circumference (≤ or >2 SD). SAS (SAS Institute, Cary, NC) and BMDP (BMDP Statistical Software, Inc, Los Angeles, CA) were used for statistical analysis.
Table 1 shows the clinical characteristics at birth of the 2 study groups. The groups did not differ significantly for gender or gestational age. Among the SGA group, 101 (47%) had birth weight ≤2 SD, 100 (47%) had birth length ≤2 SD, 32 (19%) had birth head circumference ≤−2 SD, and 46 (21%) had a birth PI <third percentile.
The main social characteristics of the AGA and SGA groups are shown inTable 2. Parents of those born SGA more frequently were not married and were of low socioeconomic status. Mothers of those born SGA more often were smokers. Among those born AGA, a late entry into secondary school was related to younger maternal age, larger family size, missing or low maternal educational level, and parental low socioeconomic status or unemployment (Table 2). We therefore adjusted the relation between school indicators and the SGA status for these parameters. As Table 3shows, the proportion of late entrants into secondary school was higher among SGA than the AGA group, and more SGA participants failed to take or pass the baccalaureate examination.
Among the SGA group, the severity of IUGR described by a birth weight or length <2 SD was not related to school achievements. Late entry into secondary school and no baccalaureate were more frequent among SGA participants whose head circumference was <2 SD, but this difference was not significant after adjustment for confounders (Table 4). In the SGA group, significantly fewer participants with an asymmetric body proportion at birth (PI <third percentile) failed to take or pass the baccalaureate compared with patients with symmetric proportions (PI ≥third percentile). Participants with IP <third percentile had a significantly lower birth weight, a higher height, and a lower head circumference than participants with IP ≥third percentile. The difference was no longer significant after adjustment on head circumference (odds ratio: 0.4; 95% CI: 0.15–1.08; P = .06).
This study shows that SGA participants have an increased risk of impaired school performance in late childhood and adolescence than AGA participants. Our findings are not due to low socioeconomic status, which has been found to be related to SGA birth20 and school performance as we have adjusted for it. Nevertheless, the contribution of parental intelligence, measured by IQ, was not investigated,21 and we could not assess the quality of the home environment.22
Studies on younger children showed conflicting results regarding psychomotor development.8 At adolescence, 2 studies were performed with few participants.14,,23 Paz et al14 found, after adjustment for parental socioeconomic status, that 34 SGA males at age 17 had lower levels of academic achievement (more of them either did not complete 12 years of schooling or finished their studies at a vocational school) and that 71 SGA females had lower intelligence scores than those who were AGA. Westwood et al23 found a trend toward lower intelligence test scores for 33 adolescent singleton SGA (versus AGA) patients.
Results from the 1970 British Birth Cohort in which school performance was assessed at 16 and 26 years were published recently.16At 16 years, those born SGA (birth weight <fifth percentile) demonstrated small but significant deficits in academic achievement, and teachers were less likely to rate those who were born SGA in the top 15th percentile of the class compared with those who were born AGA. At 26 years, the 489 adults who were born SGA did not demonstrate any differences in years of education, but they were less likely to have professional or managerial jobs and reported lower income than the 6981 AGA participants. This study had methodological strength with a high number of participants and adjustment for social confounders. Our results confirm these recent findings in a different country of an equivalent level of care, in a study that also has a sufficient number of SGA participants and is adjusted adequately.
Studies of younger children have not always detected significant differences in intelligence tests or school indicators.8Most developmental studies of SGA young children have used IQ tests, which may lack the specificity or sensitivity to identify subtle effects on such school-related skills as spatial working memory and recognition memory. Subtle developmental disabilities might not be obvious in early childhood and can only be reliably tested for later in life. For young adults or adolescents, direct measures of economic success such as educational attainment, occupation, and income also may be useful to assess the ability to cope with problems encountered in life.16
We must emphasize that our study defined SGA by stringent criteria (<third percentile, taking gestational age and gender into account): the results of previous studies among children or adolescents may have been confounded by definitions of IUGR as a birth weight below the 15th or 10th percentile or as ≤2500 g. Only the smallest infants (<fifth percentile) showed increased mortality and morbidity.1
Our follow-up rate was 62%, high for a study at 20 years from a population-based registry and higher than in the British Birth Cohort16 at 26 years (53%). The bias as a result of nonresponders in studies that deal with child development generally decreases the likelihood of finding an association, because children with poorer development or/and parents with low socioeconomic status tend to participate less in follow-up studies.24 This pattern was verified in our study. Accordingly, the association found in our study most probably is an underestimation.
A small head circumference in SGA children has been found to be associated with poorer neurodevelopmental outcome during childhood.9,,25,26 Strauss16 found that both head size at 5 years and being SGA at birth were predictors of occupation and income at adult age. In our study, although we observed this association, the difference was not significant after adjustment for confounders; this might, however, be attributable to a lack of power of the study. The effect on childhood neurodevelopment of body proportionality at birth also has been debated,8,,1327–29and the findings were inconsistent. Our study found SGA patients with a normal PI at birth (symmetric SGA) to have higher rates of low school achievement as young adults than the asymmetric SGA patients but not significant after adjustment for head circumference.
Normal brain development requires adequate oxygenation, nutrition, and hormonal balance as well as avoidance of toxic substances. A sufficient number of cells, appropriate migration, and circuit formation are critical as are plasticity and compensation pathways.6,,7 The endocrine system, most important for fetal growth, involves the insulin-like growth factors (IGFs), which may regulate nutrient partitioning between the maternal, placental, and fetal compartments.30 Although growth hormone-induced IGF-I production is not required for normal brain growth or cognitive development,31 a critical role for IGF-I in brain development has been suggested by the report of a child with mental retardation who carried a deletion of the IGF-I gene.32Whether prolonged IGF-I deficiency induced by fetal undernutrition in utero33 might alone cause deleterious effects on normal brain and cognitive development requires further study. Thyroid status also has been postulated to play a critical role in the development of the central nervous system, and circulating free thyroid hormone concentrations have been shown to be lower in fetuses with IUGR.34 Additional studies also are necessary to determine whether this results in hypothyroidism at a tissue level, especially in the central nervous system. The evaluation of these factors and whether and how they affect the developing nervous system thus remain unclear. IUGR also is a known risk factor for hypoglycemia during the neonatal period, and a retrospective evaluation recently demonstrated that recurrent episodes of hypoglycemia during the neonatal period are associated with persistent neurodevelopmental deficits in preterm SGA infants at least until 5 years of age.35 Measures to prevent hypoglycemia should be applied during this critical period to minimize the risk of subsequent neurodevelopmental deficits.
Reduced fetal oxygen or nutrient availability may have late consequences, not only for final height and metabolic disorders but also for learning during childhood, late adolescence, and, probably, adulthood. Intervention strategies should be studied to ameliorate school results in these children. Data from other sources that enable careful controlling for confounding factors will help assess the risk of impaired development as a function of the different causes of IUGR.
This work was supported by Pharmacia Upjohn Laboratories.
We thank D. Collin, A. Pinet, E. Mairot, and W. C. Walter for assistance with the field work, and M. Kaminski for discussion about the analysis.
- Received July 10, 2000.
- Accepted November 9, 2000.
Reprint requests to (B.L.) Inserm U149, 16 avenue Paul Vaillant Couturier, 94807 Villejuif, France.
- IUGR =
- intrauterine growth retardation •
- SGA =
- small for gestational age •
- AGA =
- appropriate for gestational age •
- SD =
- standard deviation •
- PI =
- ponderal index •
- CI =
- confidence interval •
- IGF =
- insulin-like growth factor
- Leger J,
- Levy-Marchal C,
- Bloch J,
- et al.
- Hales CN,
- Barker DJP,
- Clark PMS,
- et al.
- Barker DJP
- Papiernik E,
- Bouyer J,
- Dreyfus J,
- et al.
- Miller HC,
- Hassanein K
- ↵Broman SH, Nichols PL, Kennedy WA. Prenatal and Early Developmental Correlates. Hillsdale, NJ: Lawrence Erlbaum Associates; 1975
- ↵Caldwell BM, Bradley RH. Home Observation for Measurement of the Environment. Little Rock: University of Arkansas; 1979
- Westwood M,
- Kramer MS,
- Munz D,
- Lovett JM,
- Watters GV
- Harvey D,
- Prince J,
- Parkinson C,
- Campbell S
- Villar J,
- Smeriglio V,
- Brown CH,
- Klein RE
- Kranzler JH,
- Rosenbloom AL,
- Martinez V,
- Guevara-Aguirre J
- Leger J,
- Noel M,
- Limal JM,
- Czernichow P
- Kilby MD,
- Verhaeg J,
- Gittoes N,
- Somerset DA,
- Clark PMS,
- Franklyn JA
- ↵Besson Duvanel C, Fawer CL, Cotting J, Hohlfeld P, Mattieu JM. Long-term effects of neonatal hypoglycemia on brain growth and psychomotor development in small for gestational age preterm infants.J Pediatr. 1999;134:492–498
- Copyright © 2001 American Academy of Pediatrics