The purpose of this study was to investigate the association between vitamin D analogs and peak bone mineral density (BMD, g/cm(2)) in young men. The cohort consisted of 78 healthy young males with a mean age of 22.6 years at baseline. BMD of the total body, hip, and spine and lean body mass were measured at baseline and at follow-up 2 years later. Blood samples were assayed for 25-hydroxyvitamin D(2) (25OHD(2)), 25-hydroxyvitamin D(3) (25OHD(3)), and 25-hydroxyvitamin D (25OHD) at baseline using high-performance liquid chromatography. Levels of 25OHD(3) significantly correlated to BMD at all sites and to lean body mass (r = 0.23-0.35, P < 0.05). In contrast, levels of 25OHD(2) significantly negatively correlated with BMD of the total body (r = -0.28, P = 0.01) and spine (r = -0.27, P = 0.02). BMD was then adjusted for the influence of age, body weight, body height, and physical activity (hours/week). Level of 25OHD(3) was then found to be an independent predictor of BMD of the total body (beta = 0.24, P = 0.03) and spine (beta = 0.25, P = 0.03), while level of 25OHD(2) was an independent negative predictor at the same sites (beta = -0.23 for both, P = 0.03). There was a negative association between levels of 25OHD(3) and 25OHD(2) (r = -0.31, P = 0.006). In summary, our novel results suggest an inverse relationship between 25OHD(3) and 25OHD(2) and an opposite relationship of these vitamin D analogs to BMD in young men.