In their recent paper based on German old-age pensioners, Kühntopf and Tivig [1] show that early retirement is associated with considerably higher mortality in men. This is in line with previous reports from British, Danish, US, German and Greek populations showing an increased mortality risk related to retirement, especially in the case of early retirement [2–6]. As pointed out by Kühntopf and Tivig, interpretation of these results is complicated, since a “Healthy worker selection effect” may be operating. To reduce this bias, they used information on credited periods of disease in the public insurance system [1]. Other strategies include adjustment for baseline medical problems [2, 6], using a time lag during follow up [5] or exclusion of subjects retiring for health reasons [3, 4]. It is however questionable, whether these strategies have been sufficient to eliminate the effect of health on retirement.