Köhler-Forsberg and coauthors conclude from their findings that ‘switching to the other drug resulted in significant improvement in depression scores’, and argue for switching as a ‘viable approach to achieve response among patients with MDD [major depressive disorder]’.^{Reference Köhler-Forsberg, Larsen, Buttenschøn, Rietschel, Hauser and Souery1} We doubt, however, that the method of the trial allows any such conclusion: It is well established that (a) depression can improve over time due to its episodic course,^{Reference Lehmann2} (b) depression has a large placebo response,^{Reference Walsh, Seidman, Sysko and Gould3} and (c) the efficacy of antidepressants grows over time, even after 8 or 12 weeks.^{Reference Henssler, Kurschus, Franklin, Bschor and Baethge4}

To tease apart possible reasons for improvement, it is necessary to compare, in a randomised setting, patients who switch with a group of patients continuing the antidepressant that has been ineffective so far (continuation arm).^{Reference Bschor, Kern, Henssler and Baethge5} In the study by Köhler-Forsberg et al, however, the entire group of patients switched to another antidepressant. It is therefore logically impossible to pin down the mechanism leading to the improvement seen in the study: Is it, indeed, because of the switch, to the prolonged treatment with any sort of antidepressant, or simply to the passing of time?

In a recent systematic review and meta-analysis, we found only four studies comparing switch strategies with continuation.^{Reference Bschor, Kern, Henssler and Baethge5} Not one of the trials resulted in a statistically significant advantage of switching over continuation with the first antidepressant (summary estimate: statistically non-significant standardised mean difference of –0.17 in favour of continuation (95% CI: –0.59 to 0.26)). One of the four studies, however (reference 11 in the paper by Köhler-Forsberg et al), revealed a statistically significant disadvantage for patients in the switch arm. Interestingly, the authors had used similar antidepressants to the study by Köhler-Forsberg et al: the primarily noradrenergic tricyclic antidepressant desipramine and the serotonergic selective serotonin reuptake inhibitor citalopram. In conclusion, neither in earlier attempts nor, we believe, in the current investigation, has switch been shown to be a promising option for patients unresponsive to antidepressant treatment.