Background Loneliness has a significant influence on both physical and mental health. Few studies have investigated the possible associations of loneliness with mortality risk, impact on men and women and whether this impact concerns the situation of being alone (social isolation), experiencing loneliness (feeling lonely) or both. The current study investigated whether social isolation and feelings of loneliness in older men and women were associated with increased mortality risk, controlling for depression and other potentially confounding factors. Method In our prospective cohort study of 4004 older persons aged 65-84 years with a 10-year follow-up of mortality data a Cox proportional hazard regression analysis was used to test whether social isolation factors and feelings of loneliness predicted an increased risk of mortality, controlling for psychiatric disorders and medical conditions, cognitive functioning, functional status and sociodemographic factors. Results At 10 years follow-up, significantly more men than women with feelings of loneliness at baseline had died. After adjustment for explanatory variables including social isolation, the mortality hazard ratio for feelings of loneliness was 1.30 [95% confidence interval (CI) 1.04-1.63] in men and 1.04 (95% CI 0.90-1.24) in women. No higher risk of mortality was found for social isolation. Conclusions Feelings of loneliness rather than social isolation factors were found to be a major risk factor for increasing mortality in older men. Developing a better understanding of the nature of this association may help us to improve quality of life and longevity, especially in older men. © 2011 Cambridge University Press.