Context: Low testosterone is associated with all-cause mortality, but the relationship with cause-specific mortality is uncertain.
Objective: Our objective was to explore associations between testosterone and its related hormones and cause-specific mortality.
Design: This was a population-based cohort study.
Setting and participants: Demographic and clinical predictors of mortality, and testosterone, SHBG, and LH were measured from 2001-2004 in 3637 community-dwelling men aged 70-88 yr (mean, 77 yr).
Main outcome measure: Cause of death was obtained via electronic record linkage until December 31, 2008.
Results: During a mean follow-up period of 5.1 yr, there were 605 deaths. Of these, 207 [34.2%; 95% confidence interval (CI) = 30.4-38.1%] were due to cardiovascular disease (CVD), 231 to cancer (38.2%; 95% CI = 34.3-42.1%), 130 to respiratory diseases (21.5%; 95% CI = 18.2-24.8%), and 76 to other causes (12.6%; 95% CI = 9.9-15.2%). There were 39 deaths attributable to both cancer and respiratory diseases. Lower free testosterone (hazard ratio = 1.62; 95% CI = 1.20-2.19, for 100 vs. 280 pmol/liter), and higher SHBG and LH levels were associated with all-cause mortality. In cause-specific analyses, lower free testosterone (sub-hazard ratio = 1.71; 95% CI = 1.12-2.62, for 100 vs. 280 pmol/liter) and higher LH predicted CVD mortality, while higher SHBG predicted non-CVD mortality. Higher total testosterone and free testosterone levels (sub-hazard ratio = 1.96; 95% CI = 1.14-3.36, for 400 vs. 280 pmol/liter) were associated with mortality from lung cancer.
Conclusions: Low testosterone predicts mortality from CVD but is not associated with death from other causes. Prevention of androgen deficiency might improve cardiovascular outcomes but is unlikely to affect longevity otherwise.