Martenyi et al (Reference Martenyi, Brown and Zhang2002) suggest that fluoxetine is effective and well-tolerated in the prevention of relapse of post-traumatic stress disorder (PTSD) for up to 6 months. I think that this statement needs careful consideration.

First, the authors start by randomising patients into a placebo group and a fluoxetine group; the latter is later subdivided into a fluoxetine/placebo group and a fluoxetine/fluoxetine group. We see the outcome results of both the groups initially treated with fluoxetine, but those of the placebo/placebo group are not included in the paper.

Second, the authors dismiss the issue of discontinuation-emergent adverse effects, referring to a study by Rosenbaum et al (Reference Rosenbaum, Fava and Hoog1998). That study, also sponsored by Eli Lilly, concluded that fluoxetine had fewer adverse events than other selective serotonin reuptake inhibitors. However, fluoxetine was used up to a maximum dose of 60 mg/day with a mean dose close to 25 mg/day, whereas in the Martenyi et al study, the maximum dose was 80 mg/day and the mean close to 50 mg/day - double that in the Rosenbaum et al study. This is more significant as the results are not analysed on an intention-to-treat basis. Martenyi et al state that there were no significant differences when comparing drop-outs due to adverse events, but if we compare the total number of patients discontinuing the study, the percentages are almost double for those switched to placebo compared with those continued on fluoxetine (33.4% v. 17.3%).

Third, the authors mention that the reason behind the failure to show significant differences in the improvement of symptoms between the two treatment groups is the result of inconsistent patient self-rating. Could it not simply be that there are no differences?

The study addresses an important area, but the interpretation of the results should have been more rigorous.