Objective: To determine the prospective association between endogenous sex hormones and the development of type 2 diabetes in older men and women.
Research design and methods: A standardized medical history was obtained, an oral glucose tolerance test was performed, and plasma samples for sex hormones and covariates were collected from ambulatory, community-dwelling men and women at baseline from 1984 to 1987. Approximately 8 years later (1992-1996), another medical history was obtained, an oral glucose tolerance test was performed, fasting and 2-h insulin levels were measured, and the homeostasis model assessment for insulin resistance (HOMA-IR) was evaluated. This report is based on the 294 men and 233 women, aged 55-89 years, who completed both visits and who did not have diabetes as determined by history or glucose tolerance test at baseline, as well as women who were postmenopausal and not taking replacement estrogen.
Results: In age-adjusted correlation analyses, total testosterone was inversely and significantly related to subsequent levels of fasting and postchallenge glucose and insulin in men, whereas bioavailable testosterone and bioavailable estradiol were positively and significantly related to fasting and postchallenge glucose and insulin in women (all P <0.05). There was similar significant association with insulin resistance (HOMA-IR) in unadjusted and multiply adjusted analyses (P <0.05). There were 26 men and 17 women with new (incident) diabetes. The odds for new diabetes were 2.7 (95% CI 1.1-6.6) for men in the lowest quartile of total testosterone and 2.9 (1.1-8.4) for women in the highest quartile of bioavailable testosterone.
Conclusions: Low testosterone levels in men and high testosterone levels in women predict insulin resistance and incident type 2 diabetes in older adults.