The authors sought to determine the prognosis of depression in old age by reviewing original literature on this topic from three computer databases. All reports used at least 20 patients over age 60, followed up for at least 1 year, with affective status as outcome, and a quantitative meta-analysis was performed to combine results. Sixteen hospital-based and five community-based studies were selected, involving 1,487 and 249 depressed subjects, respectively. Most studies had serious methodological limitations. For combined results, 60% of patients in the hospital-based studies were well or had relapses-with-recovery, and 14%-22% were continuously ill; 19%-34% of community subjects were well; 27% were continuously ill, and most of the remainder had died. Physical illness, cognitive impairment, and severe depressive symptoms were frequently but inconsistently related to poor prognosis. Future studies must attend to composition of study populations and outcome assessment and control of extraneous prognostic factors.