A Cochrane review [withdrawn from publication] included 9 studies with a total of 211 subjects. Examined drugs included mifepristone [RU-486], ketoconazole, metyrapone and DHEA. Three trials were in patients with psychotic major depression (pMDD), five trials in non-psychotic major depression and one trial in bipolar disorder. Of the eight trials recording outcome as 50% reduction in HAM-D scores, there was no significant difference between treatment or control (RR 0.89, 95% CI 0.76 to 1.05), although the mean change in HAM-D scores indicated a significant difference in favour of treatment (WMD -4.54, 95% CI -6.78 to -2.29). Examination of data from the studies in non-psychotic depressive episodes showed a clearer antidepressant effect in favour of treatment (HAM-D 50% reduction: RR 0.72, 95% CI 0.56 to 0.91). Examination of psychotic subtype separately showed no antidepressant or overall BPRS effect (HAM-D 50% reduction: RR 0.98, 95% CI 0.79 to 1.22), but an effect in favour of treatment in BPRS PSS scores, and continuous (WMD) outcomes.
Comment: The quality of evidence is downgraded by inconsistency (heterogeneity in interventions and outcomes), by imprecise results (few patients and wide confidence intervals) and by study quality (inadequate or unclear allocation concealment).