Low testosterone and sex hormone-binding globulin levels and high estradiol levels are independent predictors of type 2 diabetes in men

Objective: To study the impact of endogenous sex hormone levels in community-dwelling men on later risk for type 2 diabetes. Design: Population-based prospective cohort study. Methods: For the analyses, 1454 men who participated in the fourth Tromso study (1994-1995) were used. Cases of diabetes were retrieved and validated until 31.12.05 following a detailed protocol. The prospective association between sex hormones and diabetes was examined using Cox proportional hazard regression analysis, allowing for multivariate adjustments. Results: There was a significantly lowered multi-adjusted risk for later diabetes with higher normal total testosterone levels, both linearly per S. D. increase (hazard ratio (HR) 0.71, confidence interval (CI) 0.54-0.92) and in the higher quartiles of total testosterone than in the lowest quartiles (HR 0.53, CI 0.33-0.84). A reduced multi-adjusted risk for incident diabetes was also found for men with higher sex hormone-binding globulin (SHBG) levels, both linearly per S. D. increase (HR 0.55, CI 0.39-0.79) and when comparing the third (HR 0.38, CI 0.18-0.81) and the fourth quartile (HR 0.37, CI 0.17-0.82) to the lowest quartile. The associations with total testosterone and SHBG were no longer significant after inclusion of waist circumference to the multivariate models. Estradiol (E(2)) was positively associated with incident diabetes after multivariate adjustments including waist circumference when comparing the second (HR 0.49, CI 0.26-0.93) and the third (HR 0.51, CI 0.27-0.96) quartile to the highest quartile. Conclusion: Men with higher E2 levels had an increased risk of later diabetes independent of obesity, while men with lower total testosterone and SHBG had an increased risk of diabetes that appeared to be dependent on obesity.