Low testosterone concentrations and prediction of future heart failure in men and in women: evidence from the large FINRISK97 study

Abstract

Aims: The increased incidence of heart failure in men suggests that endogenous sex hormones might play a role in the development of heart failure, but epidemiological data remain sparse. Here, we evaluated the predictive value of low testosterone levels on future heart failure in the large population-based FINRISK97 study. Methods and results: Baseline serum testosterone concentrations were measured in 7855 subjects (3865 men and 3990 women) of the FINRISK97 study. During a median follow-up (FU) of 13.8 years, a total of 564 heart failure events were recorded. The age-adjusted baseline testosterone levels did not differ significantly between subjects developing incident heart failure during FU and those without incident events during FU (men: 16.6 vs. 17.1 nmol/L, P = 0.75; women: 1.15 vs. 1.17 nmol/L, P = 0.32). Relevant statistically significant correlations of testosterone levels were found with high-density lipoprotein cholesterol levels (R = 0.22; P < 0.001), body mass index (R = −0.23; P < 0.001), and waist-to-hip ratio (R = −0.21; P < 0.001) in men, while statistically significant correlations in women were negligible in effect size. In sex-stratified Cox regression analyses, taking age into account, a quite strong association between low testosterone and incident heart failure was found in men [hazard ratio (HR) 1.51 (95% confidence interval, CI: 1.09–2.10); P = 0.020 for lowest vs. highest quarter], but not in women [HR 0.70 (95% CI: 0.49–0.98); P = 0.086 for lowest vs. highest quarter]. Nevertheless, this association turned non-significant after full adjustment including body mass index and waist-to-hip ratio, and testosterone levels were no longer predictive for incident heart failure—neither in men [HR 0.99 (95% CI: 0.70–1.42); P = 0.77 for lowest vs. highest quarter] nor in women [HR 0.92 (95% CI: 0.64–1.33); P = 0.99 for lowest vs. highest quarter]. Accordingly, Kaplan–Meier analyses did not reveal significant association of testosterone levels with heart failure. Conclusions: Low levels of testosterone do not independently predict future heart failure.