Abstract
BACKGROUND: Emerging data from longitudinal studies suggest that low sex steroid concentrations in men are associated with increased cardiovascular risk and mortality. The impact of longitudinal trajectory patterns from serial sex steroid and gonadotrophin measurements on the observed associations is unknown to date. METHODS: We prospectively evaluated 254 elderly men (mean age, 75·5 years) of the Framingham Heart Study with up to four serial measurements of serum total testosterone (TT), dehydroepiandrosterone sulphate (DHEAS), follicle-stimulating hormone (FSH), luteinizing hormone (LH) and total estradiol (EST); and constructed age- and multivariable-adjusted Cox proportional hazard regression models relating baseline hormone concentrations and their mean, slope and variation over time (modelled as continuous and categorized into quartiles) to the incidence of clinical cardiovascular disease (CVD) and all-cause mortality at 5- and 10-year follow-up. RESULTS: We observed no association between baseline concentrations of sex steroids, gonadotrophins and their trajectories with incident clinical CVD over 5- and 10-year follow-up. Although higher baseline TT concentrations were associated with lower mortality risk at 5 years (hazard ratio per quartile increment, 0·74; 95% confidence interval, 0·56-0·98), correction for multiple statistical testing (P < 0·005) rendered this association statistically nonsignificant. Repeat analyses at the 10-year follow-up time point also demonstrated no significant association between sex steroids, gonadotrophins or their trajectories and mortality. CONCLUSION: Investigating longitudinal trajectory patterns of serial sex steroid and gonadotrophin measurements, the present study found no consistent associations with incident clinical CVD and all-cause mortality risk in elderly men from the community.