The role of frailty in the association between testosterone levels and mortality risk in older men.

Higher testosterone levels were associated with lower risk of all-cause mortality in older men.

• Hazard ratio 0.892 (95% CI 0.830–0.959; P = 0.002) per unit increase in testosterone
• Analysis based on 1478 men aged 60 years or more from NHANES 2011–2016
• Cox proportional hazards models were used to evaluate the association

Higher testosterone levels were associated with lower risk of cardiovascular disease mortality in older men.

• Hazard ratio 0.820 (95% CI 0.716–0.939; P = 0.004) per unit increase in testosterone
• Analysis based on the same NHANES cohort of 1478 men aged 60 years or more
• Cox proportional hazards models were used

Frailty index was strongly associated with both all-cause and cardiovascular disease mortality.

• HR 1.648 per 0.1 increase in frailty index for all-cause mortality
• HR 1.723 per 0.1 increase in frailty index for cardiovascular disease mortality
• Frailty was measured using a frailty index derived from NHANES data

Frailty mediated 35.9% of the total association between testosterone and all-cause mortality, consistent with near-complete mediation.

• Mediation analysis conducted on the log-hazard scale using the product-of-coefficients approach and Sobel tests
• The proportion mediated was calculated as the ratio of the indirect to total effect
• After adjustment for frailty, the residual direct effect of testosterone on all-cause mortality was attenuated and no longer statistically significant
• The pattern was described as 'consistent with near-complete mediation'

Frailty mediated 21.7% of the association between testosterone and cardiovascular disease mortality, consistent with partial mediation.

• The indirect pathway through frailty accounted for 21.7% of the total association with cardiovascular disease mortality
• After adjustment for frailty, the direct effect was reduced but its effect size and confidence interval indicated remaining direct pathways
• This pattern supported 'a partial rather than complete mediation' for cardiovascular disease mortality

The study sample consisted of 1478 older men drawn from the National Health and Nutrition Examination Survey 2011–2016.

• All participants were aged 60 years or more
• Data came from three NHANES cycles: 2011–2012, 2013–2014, and 2015–2016
• Mediation analysis used the product-of-coefficients approach and Sobel tests to quantify indirect effects via the frailty index