The loss of efficacy of fluoxetine in pediatric depression: explanations, lack of acknowledgment, and implications for other treatments, 2025, Plöderl

Chandelier

Chandelier

Senior Member (Voting Rights)
The loss of efficacy of fluoxetine in pediatric depression: explanations, lack of acknowledgment, and implications for other treatments

Plöderl, Martin; Lyus, Richard; Horowitz, Mark A.; Moncrieff, Joanna

Abstract​

Objectives​

Fluoxetine is among the most used antidepressants for children and adolescents and frequently recommended as first-line pharmacological treatment for pediatric depression. However, in contrast to earlier studies and reviews, a Cochrane network meta-analysis from 2021 concluded that the estimated efficacy of fluoxetine was no longer clinically meaningful. We aimed to explain the discrepant findings between the recent Cochrane review and earlier reviews, and to explore if this was acknowledged in guidelines and treatment recommendations appearing since then.

Study Design and Setting​

Meta-analytical aggregation of trial results over time, exploring potential biases, and a nonsystematic search for recent treatment guidelines/recommendations from major medical organizations.

Results​

The estimated efficacy of fluoxetine in clinical trials declined over time into the range of clinical equivalence with placebo when more recent studies were included in analyses and when considering common thresholds of clinical significance. This remains unacknowledged in treatment guidelines and related publications, including some that continue to recommend fluoxetine as first-line pharmacological treatment. Finally, we find that the loss of efficacy over time is likely explained by biases such as the novelty bias or by variations of expectancy effects.

Conclusion​

The seeming lack of clinically meaningful efficacy of fluoxetine for the treatment of pediatric depression needs to be considered by those who develop treatment recommendations as well as by patients and clinicians. The biases we observed are not only relevant in the evaluation of fluoxetine and other antidepressants for pediatric depression, but also for any new treatment.

Web | DOI | Journal of Clinical Epidemiology
 
www.theguardian.com

Prozac ‘no better than placebo’ for treating children with depression, experts say

Exclusive: Clinical guidelines should change to avoid exposing young people to potentially harmful side-effects, researchers say
www.theguardian.com www.theguardian.com

AI summary:
  • A new meta-analysis of 12 trials (1997–2024) finds that fluoxetine (Prozac) offers no clinically meaningful benefit over placebo for treating depression in children and adolescents.
  • Researchers say potential side-effects—such as weight gain, sleep and concentration problems, and increased suicidal ideation—outweigh any minimal benefit.
  • The study highlights “novelty bias” in early, more positive trials, while later studies failed to confirm any real advantage.
  • Experts argue that clinical guidelines in the UK, US, and Canada should stop recommending Prozac for under-18s, as current advice is not aligned with evidence.
  • NICE and some psychiatrists urge caution in interpretation, noting that treatment decisions weigh multiple factors and that antidepressants may still be considered under specialist supervision.
 
Utsikt said:
That seems a bit misguided. Why would specialist supervision improve the effect of the drug?
Click to expand...
It seems to me as if there’s no substance, "just words" as @rvallee would say.

From the end of the article:
Prof Allan Young, chair of the Royal College of Psychiatrists’ Academic Faculty, said that the study should be interpreted with “caution”. “Clinical guidelines weigh many factors beyond average effect size, including safety, feasibility, and patient preferences. It is important that prescribed medication demonstrate consistent evidence and safety data,” he said.
Click to expand...
 
The paper said:
Finally, we find that the loss of efficacy over time is likely explained by biases such as the novelty bias or by variations of expectancy effects.
Click to expand...
That is not the likely explanation, which is, obviously, that the studies that showed efficacy were flawed and the evidence was always wrong. The bias in medical research is simply too overwhelming, a failed trial is considered a failure, and the effect is 100x larger with non-pharmaceutical trials, where it's actually more popular than ever to openly press on the scale to achieve the desired outcome.

Actually, exact same thing as scientific evidence that the Moon is made of cheese. It never was made of cheese. If such a study existed in the literature, it wouldn't reflect anything other than the fact that that study was wrong.

For a system that is supposed to be built on checks and balances, academia is severely lacking in any such thing. It's an idea, but it hasn't been realized for real. Instead we have a terrible mix of the old system, egos before evidence, alongside a failed system that makes it appear as if the old opinions built on egos are worth a damn. It's literally the worst of both worlds, with none of the best of either.
The biases we observed are not only relevant in the evaluation of fluoxetine and other antidepressants for pediatric depression, but also for any new treatment.
Click to expand...
This really captures it perfectly. It makes it as if this problem isn't widespread and affects the entirety of EBM. Instead it's treated in total isolation and only as far as things that could happen in the future. This is exactly like bad politicians trying to get over a scandal saying they are "focused on the future".

All of this is current and applies to all of EBM. You look at how evidence-based medicine has evolved and how it's just an endless stream of low quality low effort 'pilot' studies. There is zero appetite in the profession to fix any of those issues, because it would require admitting to mistakes, and that's not allowed and no one can make them. The whole thing is massively dysfunctional.
 
You must log in or register to reply here.
Back
Top Bottom