a Department of Medical Epidemiology and Biostatistics
b Unit of Clinical Epidemiology, Department of Medicine, Karolinska Institutet, Stockholm, Sweden
c Department of Womens and Childrens Health, Uppsala University, Uppsala, Sweden
| ABSTRACT |
|---|
|
|
|---|
METHODS. This was a population-based cohort study of 357768 Swedish males born as singletons without congenital malformations between 1973 and 1981. The main outcome measure was intellectual performance at military conscription.
RESULTS. Compared with men born with appropriate birth weight for gestational age, men born light for gestational age suffered an increased risk of low intellectual performance after adjustment for maternal and socioeconomic factors. The increase in risk of low intellectual performance related to a decrease in birth weight for gestational age was similar between families and within families. Men born short or with a small head circumference for gestational age were also at increased risk of low intellectual performance, both when adjusting for maternal and socioeconomic factors and within families.
CONCLUSIONS. We found that all of the studied dimensions of restricted fetal growth are independently associated with increased risks of low intellectual performance and that these associations are only partly mediated by socioeconomic or familial factors.
Key Words: birth weight head circumference intellectual performance socioeconomic factors familial factors
Abbreviations: SDSSD score ORodds ratio CIconfidence interval
PRENATAL NUTRITION IS an important determinant of maturation and functional development of the brain. Intrauterine malnutrition may deprive the brain of nutrients necessary for cellular structure and growth.1 In addition, malnutrition may, during critical periods of brain development, amplify functional errors in brain development.2 Numerous studies have found that infants born small for gestational age are at increased risk of impaired cognitive function and low intellectual performance.36 We have reported previously that birth weight, birth length, and head circumference for gestational age are independently associated with intellectual performance in early adulthood.7
However, intellectual performance is attributed to a combination of genetic and environmental factors. It is estimated that
50% of the individual differences in intelligence in adults are the result of genetic effects.810 Reviews of studies on the association between socioeconomic status and child development have shown that children from low socioeconomic class score lower in terms of school achievement and intelligence tests compared with children from families with a higher socioeconomic class.11,12 Major pathways by which low socioeconomic class hampers cognitive development in childhood may be impaired health status at birth and decreased access to social, educational, and material resources that reduce the negative effects of perinatal complications.13,14
Genetic and environmental factors are also of substantial importance for fetal growth. Maternal and fetal genetic effects are estimated to account for 40% to 70% of the variation in birth weight,15,16 and parental socioeconomic status is one of the most universally reported factors influencing birth weight.17,18 Thus, if there are genetic and social similarities in mechanisms of fetal growth restriction and intellectual performance, such factors may confound the association between fetal growth and subsequent intellectual performance.
In the present population-based study, we used prospectively collected data on birth characteristics, social factors, and intellectual performance on >350000 young Swedish males conscripted for military service. Because we also have information about full siblings, we investigated whether the association between size at birth and intellectual performance in young adulthood is influenced by familial factors, including shared environmental and common genetic factors.
| METHODS |
|---|
|
|
|---|
The Medical Birth Register has data on >99% of all births in Sweden since 1973. We used the birth register to collect data on mothers age and parity and infant birth weight, birth length, head circumference, and gestational age. All of the births and deaths are validated each year against the Register of Total Population and Population Changes, using the mothers and the infants unique National Registration Numbers. The Medical Birth Register was recently validated, and the quality of the variables included in the present investigation is considered high (results from the validation study are available online at www.sos.SE/FULLTEXT/112/2002-112-4/2002-112-4.pdf [in Swedish]).
The Swedish Conscript Register includes information about Swedish males conscripted for military service. The conscription examination is mandatory and enforced by law, but males with known severe handicaps, congenital malformations, or chronic diseases are generally not conscripted (
23% in each birth cohort). In the studied cohort, almost all (99%) conscripted men undertook their conscription examination between 17 and 19 years of age. The registry was used to collect data on weight, height, and intellectual performance at conscription.
The Population and Housing Census of 1990 was conducted by Statistics Sweden, and the following data from the census was used: households highest socioeconomic and education categories and households family structure. The Swedish Multi-Generation Register includes information about first-degree relatives for residents born in Sweden in 1932 and later. We used the Multi-Generation Register to identify the full brothers of the males in the defined study population.
Studied Cohort
The Swedish Birth Register included information on 458371 live-born males born between 1973 and 1981. To achieve higher homogeneity within the studied cohort, we excluded males born to mothers of non-Nordic nationality, males born with congenital malformations, males born before 28 or after 43 completed weeks gestation, and multiple births (n = 53634). Another 3473 males died before 18 years of age and therefore were excluded. Of the remaining 401264 males, 377527 (94%) were conscripted between the years 1991 and 2000. In all, data on intellectual performance were available for 357768 (95%) of the conscripted males.
Measures
Birth weight, birth length, and head circumference for gestational age were standardized according to the Swedish birth weight, birth length, and head circumference standards19 and were expressed in SD score (SDS) for gestational age. We define >2 SDS below the mean birth weight for gestational age as light for gestational age, between 2 and +2 SDS as appropriate weight for gestational age, and >2 SDS as heavy for gestational age. Definitions of birth length and head circumference followed the same procedure. Gestational age was estimated from the date of the last menstrual period and stratified into very preterm (
31 completed weeks gestation), moderately preterm (3236 weeks gestation), term (3741 weeks gestation), and postterm (
42 weeks gestation) births. Birth weight, birth length, and head circumference for gestational age
5 SDs from the mean were considered implausible and, hence, considered missing. The gestational-age distribution was scrutinized according to the method described by Alexander et al,20 and erroneous values of gestational age were considered missing.
At military conscription, Swedish men undergo a careful health examination. Intellectual performance is measured in 4 dimensions (logical/inductive, verbal, spatial, and theoretical/technical) and conducted through a time-limited test. The test consists of 160 questions, 40 for each dimension, and has been computerized since 1994, which precludes observer bias. The results on intellectual performance are presented as standard 9 (stanine) scores. Low intellectual performance was defined as a score of
2. Individuals scoring
2 in intellectual performance can be expected to have difficulties in coping with basic education programs. Weight is measured in kilograms (in light indoor clothes), and height is measured in centimeters (without shoes). Body mass index (BMI) was calculated as the ratio between adult weight and squared adult height (kg/m2). Underweight was defined as a BMI of <18.5, normal weight as a BMI between 18.5 and 24.9, overweight as a BMI between 25 and 29.9, and obesity as a BMI of
30. Growth in height was calculated by subtracting birth length for gestational age (SDS) from height at conscription (SDS).
The variables coming from the Population and Housing Census were classified according to recommendations set fourth by Statistics Sweden. Within each household, we used the highest socioeconomic category, which was classified in the following manner: unskilled blue-collar workers; skilled blue-collar workers; low-level white-collar workers; intermediate-level white-collar workers; high-level white-collar workers; and self-employed. Households highest education was similarly classified into 9-year compulsory school, upper secondary school 2 years, upper secondary school 3 years, higher education <3 years, and higher education
3 years. The family structure of the household was categorized as living with both biological parents, only living with biological mother, only living with biological father, and living with neither biological parent.
Statistical Methods
The cohort was analyzed with multiple logistic regression to calculate risks of low intellectual performance by using SAS (SAS Institute, Inc, Cary, NC) Proc Genmod. To adjust for potential correlation between siblings, generalized estimating equations were fitted for this data. Odds ratios (ORs) were calculated from the regression coefficients by exponentiation and are presented with 95% confidence intervals (CIs).
To control for familial factors, we estimated the risk of low intellectual performance within and between families. Familial effects on the risk of low intellectual performance include common genetic and shared environmental factors. Full brothers share, on average, 50% of their genes and generally also share environment during childhood and adolescence. These analyses were restricted to males that had at least 1 full brother in the defined study population (in all, 106513 individuals, 48767 families with 2 brothers, 2846 families with 3 brothers, 109 families with 4 brothers, and 1 family with 5 brothers, among whom 96189 [90%] were conscripted and had measurements on intellectual performance).
In the analyses of familial effects on the association between exposure (anthropometric measurements at birth) and outcome (low intellectual performance), the exposure is decomposed into between-family and within-family components. The between-family component is measured by the family mean of the measurement under study (
i; i = family number), and the within-family component is measured by the individual deviation from the family mean (Xij
i; j = individual in family i). The between-family component estimates the expected excess in risk of low intellectual performance for a 1 SDS decrease in their family mean. The within-family component estimates the expected excess in risk of low intellectual performance for a 1 SDS difference from the family mean. To establish whether an association between exposure and outcome is attributed to familial effects, a comparison between the within and between effect is needed. If familial factors are of importance, they would tend to make family members more alike. Hence, we would expect the within-family component to be less than the between-family component. If the within-family component remains significant, this indicates that the association between exposure and outcome depends on other factors than familial ones. Generalized estimating equations were also fitted for these data to adjust for correlation between siblings.
| RESULTS |
|---|
|
|
|---|
|
|
Next, we investigated whether the associations between birth characteristics (birth weight, birth length, and head circumference for gestational age), growth in height, and intellectual performance were confounded by familial (early environmental and/or common genetic) factors. Birth characteristics and growth in height were associated with risk of low intellectual performance, both in analyses between and within families (Table 3). However, the associations were slightly weaker in the analyses within families compared with analyses between families. For example, when we analyzed head circumference, the between family effect, represented by reducing the familys mean head circumference for gestational age with 1 SDS, led to a 10% increase in risk of low family mean intellectual performance. The within-family effect showed that a boy born with a head circumference for gestational age 1 SDS below the family mean suffered a slightly lower (5%) increase in risk of low intellectual performance. Similar differences in between-family and within-family effects were found with regard to birth weight and birth length for gestational age.
|
| DISCUSSION |
|---|
|
|
|---|
A recent review suggested that demographic and social factors are more important than fetal growth in determining later cognitive ability.21 However, although cognitive function is influenced by social factors, studies have found that birth weight is independently associated with cognitive function also after adjusting for social factors and maternal characteristics, including mothers age and parity.22,23 In the present study, the association between birth weight and intellectual performance was only partly explained by maternal demographic and socioeconomic factors. In addition, we also found that the association between birth weight and cognitive function was not explained by other anthropometric measures at birth or later in life.
We found that risks of low intellectual performance at 18 years related to measures of fetal growth were, if anything, slightly weaker within than between families. This indicates that some of the association between birth weight for gestational age and intellectual performance may be attributable to familial (common genetic and shared environmental) factors. Results from previous sibling and twin studies are conflicting. In sibling studies, Matte et al24 and, most recently, Lawlor et al25 reported that the association between birth weight and intellectual performance was not confounded by familial factors. In contrast, Record et al26 found an association between birth weight and verbal reasoning scores in the overall cohort but not within siblings. One twin study suggests that the association between birth weight and intellectual performance in adolescence is primarily mediated by common environmental factors.27 Meanwhile, results from another twin study suggest that the association between birth weight and intelligence is mediated by genetic factors, at least in childhood.28
Previous studies have found a positive relationship between head circumference and intellectual performance.7,29 In the present study, a small head circumference for gestational age increased the risk of low intellectual performance, but the effect was slightly lower within families. Thus, some of the variability in intellectual performance in adolescence related to head circumference at birth could be attributable to familial (common genetic and shared environmental) factors. Cognitive function and intellectual performance is developed through a mixture of genetic and environmental effects.9 Reviews of genetic studies have found that genetic differences account for at least half of the variability observed in intelligence8,30 and that heritability in interindividual differences in brain volume and structure is high.31,32 In adults, the association between brain volume and intelligence is completely mediated by common underlying genetic factors.33,34 These and our findings support the hypothesis that the association between head circumference at birth and intellectual performance may be partly mediated by familial factors, that is, shared environmental and common genetic factors.
We found that low birth weight for gestational age is associated with risk of low intellectual performance also after adjusting for head circumference for gestational age. In malnourished fetuses, shunting of fetal blood to the central nervous system will result in preferential perfusion of the fetal brain, often referred to as brain sparing.35 In pregnancies complicated by fetal growth restriction, this may lead to a proportionally smaller reduction in head circumference than body weight.1,36,37 Martyn et al38 suggested that, in the absence of an association between fetal growth and cognitive function, adaptations (including brain sparing) made by fetal growth-restricted infants seem to be largely successful in maintaining their brain development. However, results from another study suggest that cognitive development may be affected despite this brain-sparing effect.39 Growth-restricted infants may also have a reduction in brain volume despite apparently normal head circumference.40 Thus, low birth weight for gestational age, irrespective of head circumference at birth, may negatively affect brain organization and subsequent intellectual performance.
This nationwide population-based study permitted powerful comparisons of subgroups within the studied population. Intellectual performance tests were conducted on most of the conscripts. Recall bias was precluded, because data were collected prospectively. The studied cohort was homogeneous, which reduces the possibilities of unmeasured confounding factors. However, we were restricted to the data available in the registries and were unable to account for potential confounders, such as smoking and alcohol habits of both the studied males and their parents and obstetric and perinatal complications. The study was restricted to males in young adulthood, and the interpretation of the results should be restricted to males.
| CONCLUSIONS |
|---|
|
|
|---|
| ACKNOWLEDGMENTS |
|---|
| FOOTNOTES |
|---|
Address correspondence to Niklas Bergvall, MSc, Department of Medical Epidemiology and Biostatistics, Karolinska Institute, Box 281, SE-171 77 Stockholm, Sweden. E-mail: niklas.bergvall{at}meb.ki.se
The authors have indicated they have no financial relationships relevant to this article to disclose.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
M. Desai and M. D. Begg A Comparison of Regression Approaches for Analyzing Clustered Data Am J Public Health, August 1, 2008; 98(8): 1425 - 1429. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. Yang, J. Lynch, E. S. Susser, and D. A. Lawlor Birth Weight and Cognitive Ability in Childhood Among Siblings and Nonsiblings Pediatrics, August 1, 2008; 122(2): e350 - e358. [Abstract] [Full Text] [PDF] |
||||
![]() |
M.-T. Tsou, M.-W. Tsou, M.-P. Wu, and J.-T. Liu Academic achievement of twins and singletons in early adulthood: Taiwanese cohort study BMJ, July 21, 2008; 337(jul21_1): a438 - a438. [Abstract] [Full Text] [PDF] |
||||
![]() |
K. Heinonen, K. Raikkonen, A.-K. Pesonen, E. Kajantie, S. Andersson, J. G. Eriksson, A. Niemela, T. Vartia, J. Peltola, and A. Lano Prenatal and Postnatal Growth and Cognitive Abilities at 56 Months of Age: A Longitudinal Study of Infants Born at Term Pediatrics, May 1, 2008; 121(5): e1325 - e1333. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. L. Nielsen, C. Dethlefsen, H. T. Sorensen, J. F. Pedersen, and L. Molsted-Pedersen Cognitive Function and Army Rejection Rate in Young Adult Male Offspring of Women With Diabetes: A Danish population-based cohort study Diabetes Care, November 1, 2007; 30(11): 2827 - 2831. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. H. Whitaker, J. F. Feldman, J. M. Lorenz, S. Shen, F. McNicholas, M. Nieto, D. McCulloch, J. A. Pinto-Martin, and N. Paneth Motor and cognitive outcomes in nondisabled low-birth-weight adolescents: early determinants. Arch Pediatr Adolesc Med, October 1, 2006; 160(10): 1040 - 1046. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||