Childhood BMI is inversely associated with pubertal timing in normal-weight but not overweight boys

Abstract

ABSTRACT Background An inverse association between childhood body mass index (BMI; in kg/m2) and pubertal timing is well established for girls. Among boys, studies are scarce and the results inconclusive. Objective We aimed to determine the association between childhood BMI and age at peak height velocity (PHV) in boys. Design We collected height and weight measurements between 6.5 and 22 y of age for boys born 1945–1961 (original cohort; n = 31,971; mean ± SD childhood BMI: 15.74 ± 1.41; age at PHV: 14.06 ± 1.11 y) and 1981–1996 (replication cohort; n = 1465; childhood BMI: 16.47 ± 2.06; age at PHV: 13.71 ± 1.08 y) attending schools in Gothenburg, Sweden, and examined at mandatory military conscription. Age at PHV was obtained from curve-fitting of measured heights with the use of a modified Infancy-Childhood-Puberty model. Results In the original cohort, childhood BMI was inversely associated with age at PHV (P < 0.001) and a significant quadratic term for childhood BMI (P < 0.001) indicated the nonlinearity of this association. Via piecewise linear regression, we identified a threshold for the association at a childhood BMI of 18.42. A significant inverse association was observed below (β: −0.17 y/BMI unit; 95% CI: −0.18, −0.16 y/BMI unit) but not above (β: 0.02 y/BMI unit; 95% CI: −0.03, 0.06 y/BMI unit) this childhood BMI threshold. For every unit increase in childhood BMI, age at PHV was ∼2 mo earlier up to the childhood BMI threshold. Similar results were observed in the replication cohort, demonstrating a significant inverse association below (β: −0.16; 95% CI: −0.21, −0.11) but not above (β: −0.03; 95% CI: −0.11, 0.05) the childhood BMI threshold. The identified threshold was close to the cutoffs for overweight at 8 y of age, and childhood BMI was inversely associated with age at PHV below but not above the overweight cutoffs. Conclusions The present findings establish an inverse association between childhood BMI and pubertal timing in normal-weight but not overweight boys.