This huge systematic review (Cipriani et al 2018) does not add anything to the knowledge we already had about depression pills.
Briefly, the effects as estimated on the Hamilton scale are very similar to those reported in a another huge meta-analysis in early 2017 (Jakobsen JC, Katakam KK, Schou A, et al. Selective serotonin reuptake inhibitors versus placebo in patients with major depressive disorder. A systematic review with meta-analysis and Trial Sequential Analysis. BMC Psychiatry 2017;17:58). See p4 bottom right and p150 in the Appendix for these results. The average effect of these drugs, SMD 0.30, is less than what is clinically relevant.
There were two primary outcomes, response rate and drop-out for any reason.
As explained in the Discussion of Jakobsen’s meta-analysis, response rates, defined as “the total number of patients who had a reduction of ≥50% of the total score on a standardised observer-rating scale for depression”, are flawed and should not be trusted.
The other primary outcome was called acceptability in the paper, defined as treatment discontinuation measured by the proportion of patients who withdrew for any reason. The results for this outcome are shown in Figure 3 and on p148 in the Appendix. They are not reliable. We have done such an analysis based on clinical study reports of placebo-controlled trials we have obtained from European drug regulators (submitted for publication). I will not reveal the details before our results have been published, but our results are very different from those shown in the Lancet paper. The use of clinical study reports is crucial. The manufactures had excluded patients from their analyses, which we were able to include. This is generally not possible for the type of systematic review done by Cipriani et al. So, these results are also flawed and should not be trusted.
Peter C. Gøtzsche
Nordic Cochrane Centre