Nature: one 2006 trial of 40 children may have skewed Prozac paediatric depression meta-analyses

On 2 October 2026, Nature news reported a secondary analysis arguing that a single 2006 Iranian clinical trial of 40 children may have wrongly swayed meta-analyses in favour of fluoxetine (Prozac) for depression in children and adolescents. The new work, published in Cochrane Evidence Synthesis and Methods, is led by researchers including consultant child psychiatrist Richard Lyus, psychiatrist Florian Naudet and clinical psychotherapist Martin Plöderl.

What the re-analysis found

Many guidelines cite a 2016 Lancet network meta-analysis and a 2020 Lancet Psychiatry update that found Prozac better than placebo for paediatric depression. A 2021 Cochrane review instead found only a small, unimportant advantage. The new team reports that disagreement largely turns on whether the 2006 trial (Prozac versus nortriptyline) is included: its effect size was extraordinarily large compared with other paediatric Prozac trials, and re-running the 2016 and 2020 analyses without that trial removed the clear benefit. The 2021 Cochrane review had already omitted it because of how it selected comparison studies.

Caveats — and what authors are not saying

Authors describe the 2006 trial as a possible “zombie” study with data so flawed as to be untrustworthy (orderly identical score shifts in both arms; no ethics-approval description). Trial investigators did not respond to the team or to Nature. Co-author of the earlier favourable meta-analyses Andrea Cipriani told Nature he agrees the 2006 data are problematic and that teams had noted very low confidence and risk of bias. Lyus says guideline committees should review recommendations and weigh whether use can still be justified — the authors stop short of calling for clinicians to stop prescribing Prozac for paediatric depression. This article summarises a secondary analysis of published trials/meta-analyses with inherent limitations; it is not personal medical advice. Patients and caregivers should talk with a qualified clinician before changing any medicine.