In men, testosterone concentration declines gradually with ageing and has been linked to many adverse consequences of ageing. This prospective study included 794 Caucasian men aged 50-91 (median 73.6) and examined the association of endogenous testosterone levels at baseline with mortality in older men followed for an average of 11.8 years. Men with total testosterone levels below the 25th percentile (241 ng/dl) had a 40% higher risk of death compared to men with higher testosterone concentrations. This relationship was independent of age, obesity, and lifestyle choices. Total testosterone concentrations were also found to be inversely associated with weight, body mass index, waist-to-hip ratio, systolic and diastolic blood pressure, fasting plasma glucose, serum insulin, triglycerides, C-reactive protein and leptin levels, and positively related to HDL cholesterol and adiponectin levels (all p<0.01). Moreover, mortality associated with low levels of testosterone was found to be independent of the metabolic syndrome, diabetes, and prevalent cardiovascular disease. This is the first prospective population-based study to demonstrate that low serum testosterone increases the risk of death in older men. However, in this population, testosterone levels above the median did not confer any additional survival benefit, putting a damper on the notion of widespread testosterone therapy for ageing in men.