To the Editor The meta-analysis by Kalfas et al attempts to examine a safety issue in a sample of trials where strong effects are unlikely to show up. Previous research found that antidepressant withdrawal symptoms are predominantly mild and transient when treatment duration is less than 3 months (but occasionally, severe symptoms still occur), whereas the risk of severe withdrawal effects is notably increased after 6 months of treatment. For example, Rosenbaum et al showed that after a minimum treatment duration of 4 months (mean, 11 months), abrupt placebo-interruption of sertraline and paroxetine treatment for 5 to 8 days caused, on average, 5.7 (Hedge g = 0.9) and 7.8 (Hedge g = 1.1) incident withdrawal symptoms, compared with 0.2 for fluoxetine, indicating severe withdrawal effects in many patients. Surprisingly, the trial by Rosenbaum et al was not included in the Kalfas et al main analysis of continuous withdrawal scores. Instead, the authors relied on 11 short-term trials: 6 of 8-week duration, 4 of 12-week duration, and 1 of 26-week duration.