Objective: The objective of this study was to examine whether the established positive association between birth weight and childhood psychometric intelligence is seen within singleton sibling pairs from the same family as well as between nonsiblings.
Methods: We examined the association of intrauterine growth (measured as birth weight standardized for gender and gestational age) with psychometric intelligence (measured using the Moray House picture test) at 7 years old in a birth cohort of 9792 individuals who were singleton births occurring in Aberdeen, Scotland, between 1950 and 1956. We further compared this association within siblings with that between nonsiblings in the cohort; this family-based analysis included 1645 sibling pairs (N = 3290 individuals).
Results: There was a positive linear association between birth weight and childhood psychometric intelligence at age 7 in the whole cohort, which remained with adjustment for a range of potential confounding factors. A one standard deviation increase in birth weight for gestational age z score was associated with a greater intelligence score in a regression model adjusting for sex, year of birth, paternal social class, maternal height, age, gravidity, and birth outside of marriage. The mean age difference between the siblings within each family pair was 2.2 years. In the family-based analysis there was no strong association between birth weight for gestation age z score and intelligence within sibling pairs from the same family, but there was a positive association between nonsiblings; the difference in these effects being unlikely to be due to chance. With additional adjustment for social class, maternal height, age, gravidity, and birth outside of marriage, the within-sibling pair effect was unaltered and the nonsibling effect attenuated, although an apparently robust positive association remained. In these adjusted analyses there was still evidence that the within-sibling effect differed from that between nonsiblings. We found no evidence that the main effects or the family-based analyses differed between males and females.
Discussion: Our family-based analyses are consistent with one previous large family-based study that included >2500 sibling pairs and found no within-sibling-pairs association between birth weight and childhood intelligence, but did not make a direct statistical comparison between the within-sibling-pairs association and that between nonsiblings. In a second large study that included only sibling pairs of the same sex, in males there was a within-sibling-pairs association between birth weight and childhood intelligence. However, for females there was no within-sibling-pairs association. The authors commented that this sex difference was puzzling and needed replication. Although we had less power than this earlier study to assess sex differences, the point estimates and statistical tests in our study suggested that there was no sex difference.
Conclusions: The lack of any association within sibling pairs from the same family suggests that the association between birth weight and childhood intelligence in the general population of singletons is largely explained by fixed family factors that are closely matched in siblings of a similar age. These factors include family socioeconomic characteristics, parental education and intelligence, genetic factors and fixed maternal factors, such as her behaviors, size, and metabolic and cardiovascular health that are constant from one pregnancy to the next and could therefore affect her offspring growth and intelligence across all pregnancies.