Abstract
CONTEXT: Testosterone supplementation is increasingly widespread and has well-established beneficial effects on sexual function and metabolic health. However, there remains uncertainty regarding associated cardiovascular risks. OBJECTIVE: Human genetics studies demonstrated that Mendelian randomization approaches recapitulate the beneficial effects of testosterone therapy; here we apply this to cardiovascular disease. DESIGN: We performed a Mendelian randomization study to assess the causal effect of higher circulating testosterone on coronary artery disease (CAD). We also tested the phenotypic association between measured circulating testosterone and CAD in the cohort of men aged 40 to 69. PATIENTS OR OTHER PARTICIPANTS: Testosterone genetic instrument data were derived from 425 097 European ancestry adults from the UK Biobank study and CAD from single nucleotide polymorphism-level summary statistics from 1 165 690 individuals in CARDIoGRAMplusC4D. MAIN OUTCOME MEASURE(S): CAD as defined in CARDIoGRAMplusC4D was the main outcome. In longitudinal analyses, CAD was defined according to medical records and self-report. RESULTS: We found that higher genetically predicted circulating testosterone conferred a higher risk of CAD in men [odds ratio (OR): 1.17, 95% confidence interval (CI) 1.07-1.27, P = 3.32 × 10-4]. There was no evidence of an effect in women (OR: 1.01, 95% CI 0.94-1.10, P = .73). The genetic association in men appeared to be mediated by higher blood pressure. In longitudinal observational analyses, a directionally opposite association was observed in men, likely arising due to confounding by type 2 diabetes and body mass index. CONCLUSION: These data suggest that increased testosterone may increase the risk of cardiovascular disease and that this safety concern should be a focus in future clinical trials for testosterone supplementation.