Clinical Question

What is the best treatment for depression with psychotic features?

Evidence-Based Answer

An antidepressant such as imipramine (Tofranil) or sertraline (Zoloft), possibly with the addition of an antipsychotic, is the preferred initial pharmacologic treatment for psychotic depression.

Practice Pointers

About 15 percent of patients with major depression exhibit psychotic features such as hallucinations or delusions. Electroconvulsive therapy often is used to treat psychotic depression and has been shown to be effective in comparisons with sham electroconvulsive therapy. Typically, however, it is not considered first-line therapy, and it has not been compared with pharmacologic treatment in a randomized controlled trial (RCT). Most patients initially are treated pharmacologically, and this review by Wijkstra and colleagues compared an antidepressant alone, an antipsychotic alone, and the combination of both.

Only 10 RCTs with a total of 548 patients were identified after a thorough search of the literature. Although only double-blinded studies with appropriate randomization and intention-to-treat analysis were included, most studies were small (mean of 55 patients), and differing dosages and treatment protocols made it inappropriate to combine results from different studies. The conclusions that could be drawn from this literature, therefore, are limited.

The authors found that an antidepressant alone was effective (four RCTs; relative risk [RR] 2.06; 95% confidence interval [CI], 1.41 to 3.00), but an antipsychotic alone was not. Four studies compared different anti-depressants and found that imipramine was more effective than mirtazapine (Remeron) and fluvoxamine (Luvox), that sertraline was more effective than paroxetine (Paxil), and that there was no difference between fluvoxamine and venlafaxine (Effexor). Based on these limited data, the preferred initial agents are imipramine and sertraline. There was a nonsignificant trend toward greater effectiveness of the combination of antidepressant and antipsychotic over an antidepressant alone (two studies, RR 1.44; 95% CI, 0.86 to 2.41). An important practice point for these patients that was not addressed by this meta-analysis is that they should be strongly considered for hospitalization.¹