Background: We hypothesized that levels of circulating sex steroids i) change with stages of chronic kidney disease (CKD) and ii) are associated with clinical outcomes.

Methods: We conducted a prospective observational study. 221 patients (143 males, 78 females) with CKD comprising 49 patients with CKD III-IV, 102 dialysis patients, and 70 kidney transplant recipients (KTR) were followed at a tertiary referral centre to death or kidney transplant. Baseline serum sex steroid levels were repeatedly measured by liquid chromatography/ tandem mass spectrometry.

Results: In males, but not in females, both testosterone (p=0.003) as well as estradiol (p<0.02) levels were lowest in dialysis patients and highest in KTR. Over a median follow up of 8.5 years (interquartile range 3.8-9.2), 52 men (36%) died, and 24 (17%) received a kidney transplant. In multivariate analyses using Cox proportional hazard models up to 9.6 years, testosterone predicted mortality independent of baseline age, body mass index, and renal disease status (p=0.02), and circulating levels of brain natriuretic peptides or cardiac troponin T (p<0.05). An increase in testosterone by 1 nmol/L was associated with a 9.8% (95% confidence interval 3.1-16.3%) decrease in mortality. By contrast, sex steroid levels were not associated with mortality in females.

Conclusions:             Low testosterone levels predict mortality in males, independent of established and novel predictors of mortality in CKD.