Plasma sex hormone levels and mortality in disabled older men and women

Aim: To investigate the relationship between circulating sex hormone levels and subsequent mortality in disabled elderly. Methods: This prospective observational study was comprised of 214 elderly subjects aged 70–96 years (117 men and 97 women; mean ± standard deviation age, 83 ± 7 years), receiving services at long‐term care facilities in Nagano, Japan. All‐cause mortality by baseline plasma sex hormone levels was measured. Results: After excluding deaths during the first 6 months, 27 deaths in men and 28 deaths in women occurred during a mean follow up of 32 months and 45 months (up to 52 months), respectively. Mortality rates differed significantly between high and low testosterone tertiles in men, but did not differ significantly between middle and low tertiles. Compared with subjects in the middle and high tertiles, men with testosterone levels in the low tertile (<300 ng/dL) were more likely to die, independent of age, nutritional status, functional status and chronic disease (hazard ratio [HR] = 3.27, 95% confidence interval [CI] = 1.24–12.91). In contrast, the low dehydroepiandrosterone sulfate (DHEA‐S) tertile was associated with higher mortality risk in women (multivariate adjusted HR = 4.42, 95% CI = 1.51–12.90). Exclusion of deaths during the first year and cancer deaths had minimal effects on these results. DHEA‐S level in men and testosterone and estradiol levels in women were not related to mortality. Conclusion: Low testosterone in men and low DHEA‐S in women receiving care at facilities are associated with increased mortality risk, independent of other risk factors and pre‐existing health conditions. Geriatr Gerontol Int 2011; 11: 196–203 .