RESEARCH DESIGN AND METHODS - Analyses were conducted on the cohort of the Massachusetts Male Aging Study, a population-based random sample of men aged 40-70. Of the 1,709 men enrolled in 1987-1989 (T-1), 1,156 were followed up 7-10 years later (T-2). Testosterone and SHBG levels at T-1 were used to predict new cases of diabetes between T-1 and T-2.

RESULTS - After controlling for potential confounders, diabetes at follow-up was predicted jointly and independently by lower baseline levels of free testosterone and SHBG. The odds ratio for future diabetes was 1.58 for a decrease of ISD in free testosterone (4 ng/dl) and 1.89 for a 1SD decrease in SHBG (16 nmol/l), both significant at P < 0.02.

CONCLUSIONS - Our prospective findings are consistent with previous, mainly cross-sectional reports, suggesting that low levels of testosterone and SHBG play some role in the development of insulin resistance and subsequent type 2 diabetes.