Objective: The authors conducted a systematic review and meta-analysis to determine whether the risk of psychosis is higher in past or future episodes in patients with major depression with psychotic features than in patients with nonpsychotic depression.

Method: PubMed, Embase, and PsycINFO were searched, and studies were selected that 1) identified patients with unipolar major depression, 2) made diagnoses of psychosis based on the presence of delusions or hallucinations, 3) characterized past or subsequent episodes as psychotic or nonpsychotic, and 4) were published in English. Two meta-analyses were then conducted using data from patients having index depressive episodes with or without psychosis at study entry to determine the risk of any prior or subsequent psychotic episode and the risk of psychosis in all episodes.

Results: Twelve studies met the inclusion criteria, and altogether they included 546 psychotic and 1,583 nonpsychotic patients with unipolar depression. In seven of the studies, the risk ratio for a prior or subsequent psychotic episode in patients whose index depressive episode was psychotic compared with those whose index episode was nonpsychotic was 9.98 (95% CI=4.75, 20.94). In eight studies, the risk ratio for psychosis among all episodes of depression in the subgroups with psychotic and nonpsychotic index episodes was 7.24 (95% CI=5.03, 10.43). Differences in risk of psychosis between these subgroups remained robust when potential sources of heterogeneity were explored.

Conclusions: The findings support the hypothesis that psychotic depression runs true to form, and they support the distinction between psychotic and nonpsychotic depression. Because patients with psychotic depression are at high risk for psychosis in future episodes, determination of effective preventive treatments is imperative.