Familial resemblance of body mass index and familial risk of high and low body mass index. A study of young men in Sweden

OBJECTIVE: Male familial correlations in body mass index (BMI) were studied using a large nationwide dataset of Swedish military conscripts examined at age 18-19y. DESIGN: Record linkage was performed between Sweden's Military Service Conscription Register and the Multiple-Generation Register, enabling the identification of 92 869 families containing at least two brothers. Data on BMI at conscription was available for 196 743 sons and for 19 972 fathers. Pairs of relatives were used to estimate Pearson correlation coefficients for BMI, and to examine whether there are BMI intervals that show particularly strong familial resemblance. RESULTS: All biological family relations showed highly significant correlations for BMI: 0.28 (95% CI 0.27-0.29) for father-son pairs; 0.36 (0.35-0.37) for full-brothers, 0.21 (0.18-0.24) for maternal half-brothers, and 0.11 (0.08-0.14) for paternal half-brothers. Also, a significant correlation, of 0.06 (0.01-0.11), was found for non-biological quasi father-son relations. Full-brothers were more similar with respect to BMI at age 18 than father-son pairs, and maternal half-brothers were more similar than paternal half-brothers. The familial risk of having BMI values above or below various cut-offs was found to be equally strong for low BMI as for high BMI values. CONCLUSION: The almost twice as strong BMI correlation between maternal half-brothers as between paternal half-brothers illustrates the importance of factors of non-additive genetic origin, to the familial aggregation of BMI. The significant BMI association found between biologically unrelated individuals from the same family emphasizes that assortative mating (and regional clustering) should be taken into account when the heritability of BMI is estimated.