Authors' reply

It is correct that some patients with psychotic depression in the studies by Bruijn et al (1996) and van den Broek et al (2004) were given haloperidol as adjunctive treatment. However, as mentioned in our article, in our intention-to-treat analysis we counted these patients as having dropped out. So, additional treatment with haloperidol did not affect the validity of our findings regarding patients receiving antidepressant monotherapy.

We agree, as mentioned in our article, that the quality of the constituent studies and the small sample sizes does influence the outcome of our meta-analysis regarding the comparison of antidepressant monotherapy v. the combination of an antidepressant and an antipsychotic. But to say that these data favour the combination is statistically not true and surely is no sound basis for contemporary clinical practice to use the combination. Moreover, the data indicate that there is no evidence for the clinical belief that an antidepressant alone is ineffective. Thus, we maintain our conclusion that both antidepressant monotherapy and the combination of an antidepressant and an antipsychotic are appropriate options for patients with psychotic depression.