A 4-year follow-up of 75 patients was conducted to investigate outcome after recovery from an episode of mania. Predictors of an unfavorable outcome included poor occupational status prior to index episode, history of previous episodes, history of alcoholism, psychotic features and symptoms of depression during the index manic episode, male gender, and interepisode affective symptoms at 6 months' follow-up. The mortality risk during the follow-up period was 4%. The identification of specific risk factors depended on the definition of outcome and the length of follow-up.