Prozac use for childhood depression skewed by single flawed medical trial

A reanalysis shows that a single small 2006 Iranian clinical trial of 40 children comparing fluoxetine (Prozac) with nortriptyline produced an implausibly large effect size and has disproportionately influenced paediatric antidepressant evidence. The trial’s data imply unusually uniform changes in every participant’s depression score, lack a description of ethics approval, and have not been defended by the original authors despite queries. When that trial’s data are excluded from major meta-analyses, the large apparent benefit of fluoxetine disappears; the positive findings reported in 2016 and 2020 meta-analyses depend on including this outlier, whereas a 2021 Cochrane review that omitted it found only a small and clinically unimportant advantage.

The episode illustrates how a single “zombie” trial can corrupt guideline evidence and why new procedures are now used to flag untrustworthy studies. Publishers and journal editors are being asked to investigate the 2006 paper, and systematic-review groups have protocols to remove suspect trials. Authors of the reanalysis urge guideline committees to re-evaluate recommendations for fluoxetine in children and adolescents, while stopping short of telling clinicians to halt prescriptions immediately; the broader conclusion is that medical guidance should be revisited when foundational trials are demonstrably flawed.