Association of Serum Testosterone at 12 Years with a Subsequent Increase in Bone Mineral Apparent Density at 18 Years: A Longitudinal Study of Boys in Puberty.

Background: Cross-sectional studies have associated serum testosterone with bone mineral density (BMD). However, there is a shortage of prospective longitudinal studies in this domain, leaving it unclear whether changes in testosterone level precede changes in BMD. Objectives: To examine the association between serum testosterone concentration at the age of 12 years and a subsequent increase in BMD by the age of 18 years. Methods: Eighty-eight boys with a mean age of 12.1 +/- 0.7 (time point 1 [T1]) and 18.0 +/- 0.7 (T2) were investigated. For both time points, serum testosterone was measured from venous blood samples. Total body (TB) and lumbar spine (LS) BMD and bone mineral apparent density (BMAD) were measured. As different brands of DEXA machines were used at T1 and T2, we calculated SD scores (SDS) from samples at T1 and T2 and their change (Delta). As covariates, bone age at T1 and physical activity (PA) by accelerometer at T1 and T2 were measured. Results: Serum testosterone at T1 was positively correlated with TB BMD at T2 (r = 0.28; p < 0.01), Delta TB BMAD SDS (r = 0.47; p < 0.0001) and Delta LS BMAD SDS (r = 0.23; p < 0.05). When additionally controlling for bone age and total PA at T1, the correlation between testosterone at T1 and Delta TB BMAD SDS remained significant (r = 0.32; p < 0.05). Conclusions: Serum testosterone concentration at the age of 12 years is associated with a subsequent increase in TB BMAD by the age of 18 years. This supports the inference that testosterone levels in early puberty may influence subsequent bone mineral accrual.