||Consistent with previous studies, our group provided a preliminary report stating that mortality of patients with affective disorders is higher than expected unless lithium prophylaxis is controlled for several years. The objective of this study was to analyze whether specific causes of mortality are diminished during prolonged lithium treatment. Causes of death registered from January 1, 1980, to December, 2002 among 1411 consecutive outpatients admitted to the lithium clinic, Department of Neurosciences, University of Cagliari, from January 1, 1980, to June, 2000 were coded according to the International Classification of Diseases, 9th revision. Survival analysis was performed by means of Cox proportional hazard modeling, adjusted for age and gender. Prolonged attendance at the lithium clinic favored survival to death from external causes. Patients who attended for less than 2 years had an increased risk of dying from external causes compared with patients who attended for 2 years or more. Risk was highest (hazard ratio, 15.8; 95% confidence interval, 2.1–119.0) at 3 years, based on 16 events and only 1, respectively. Among external causes, suicide accounted for the majority of cases (80%). Among patients under controlled treatment for at least 2 years, only 1 committed suicide during regular attendance, whereas 9 committed suicide an average of 28 months (interquartile range, 9.5-45.4) after abandoning the lithium clinic.