EDITOR,--I M Anderson and B M Tomenson's paper1 updates our meta-analysis comparing drop out rates in patients taking selective serotonin reuptake inhibitors and tricyclic antidepressants2 but excludes newer tricyclic drugs. Despite this difference, which would tend to favour the selective serotonin reuptake inhibitors, the results are nearly identical. The table compares drop out rates in the three published meta-analyses.1 2 3 Clearly, despite the different inclusion criteria, the results broadly agree: the difference is small and of dubious clinical and economic importance.