BMI in the Trois‐Rivières study: Child–adult and child–parent relationships

This study evaluated intraindividual child–adult and interindividual child–parent relationships of body mass index (BMI) using data from the Trois‐Rivières semilongitudinal study of growth and development. Intraindividual correlations between age 12 and 35 years were substantial (r² = 36% of variance in women, 30% of variance in men). Interindividual child–parent correlations for mothers and fathers age 36.6 ± 0.4 and 39.5 ± 0.4 years, respectively, were very low to low for daughters age 12 years (r = 0.09, NS and 0.34, P < 0.001 vs. father and mother, respectively) but all very low for sons age 12 years (r = 0.07, NS and 0.16, NS vs. father and mother, respectively). A multiple regression analysis predicted adult BMI from the individual's BMI at 10, 11, 12 years plus the maternal and paternal BMIs as calculated from self‐reported heights and weights. The BMI at age 12 years was a better predictor of adult BMI than the parental BMI in both men and women (P < 0.001) and multivariate analysis revealed that this index at age 12 years was the sole significant predictor of adult BMI for both men and women. The results from our study do not support the hypothesis that parental BMI is a stronger predictor of adult BMI than childhood BMI. However, useful information for the prediction and prevention of adult overweight can be obtained from the BMI at age 12 years. Our results suggest that environmental influences may be the major factor in the present obesity epidemic. Am. J. Hum. Biol. 15:187–191, 2003. © 2003 Wiley‐Liss, Inc.