Circulating Androgen Concentrations and Risk of Incident Heart Failure in Older Men: The Cardiovascular Health Study

Abstract

Background Circulating androgen concentrations in men decline with age and have been linked to diabetes and atherosclerotic cardiovascular disease (ASCVD). A similar relationship has been reported for low total testosterone and incident heart failure (HF) but remains unstudied for free testosterone or the more potent androgen dihydrotestosterone (DHT). We hypothesized that total/free testosterone are inversely related, sex hormone-binding globulin is positively related, and total/free DHT bear a U-shaped relationship with incident HF. Methods and Results In a sample of men from the CHS (Cardiovascular Health Study) without atherosclerotic cardiovascular disease or HF, serum testosterone and DHT concentrations were measured by liquid chromatography-tandem mass spectrometry, and sex hormone-binding globulin by immunoassay. Free testosterone or DHT was calculated from total testosterone or total DHT, sex hormone-binding globulin, and albumin. We used Cox regression to estimate relative risks of HF after adjustment for potential confounders. In 1061 men (aged 76±5 years) followed for a median of 9.6 years, there were 368 HF events. After adjustment, lower calculated free testosterone was significantly associated with higher risk of HF (hazard ratio [HR], 1.14 [95% CI, 1.01-1.28]). Risk estimates for total testosterone (HR, 1.12 [95% CI, 0.99-1.26]), total DHT (HR, 1.10 [95% CI, 0.97-1.24]), calculated free dihydrotestosterone (HR, 1.09 [95% CI, 0.97-1.23]), and sex hormone-binding globulin (HR, 1.07 [95% CI, 0.95-1.21]) were directionally similar but not statistically significant. Conclusions Calculated free testosterone was inversely associated with incident HF, suggesting a contribution of testosterone deficiency to HF incidence among older men. Additional research is necessary to determine whether testosterone replacement therapy might be an effective strategy to lower HF risk in older men.

Keywords: aging; heart failure; men; sex hormones; testosterone.

Figure 1. Graphs of predicted hazard ratios (HRs) for heart failure by hormone measures, in SD increments.

HRs are predicted from cubic spline models adjusting for age and race with knots at percentiles determined using Akaike's information criterion, selected from among models with 0 to 9 evenly spaced knots and 3 to 7 knots at Harrell's specified locations. LRT P value compares cubic spline model to a linear model. The density plot displayed just above the x axis indicates the distribution of the hormone along the spline. AIC indicates Akaike's information criterion; DHT, dihydrotestosterone; LRT, likelihood ratio test; SHBG, sex hormone–binding globulin; and T, testosterone.

Figure 2. Associations of circulating hormone measures with incident heart failure. Model 1 adjusts for age and race.

Model 2 adjusts for age, race, body mass index, smoking, alcohol use, physical activity, systolic blood pressure, antihypertensive medication, total and high‐density lipoprotein cholesterol, and estimated glomerular filtration rate. cfDHT indicates calculated free dihydrotestosterone; cfT, calculated free testosterone; DHT, dihydrotestosterone; HR, hazard ratio; SHBG, sex hormone–binding globulin; and T, testosterone.