Critical Psychiatry Textbook, Chapter 8: Depression and Mania (Affective Disorders) (Part Five)


Editor’s Note: Over the next several months, Mad in America is publishing a serialized version of Peter Gøtzsche’s book, Critical Psychiatry Textbook. In this blog, he continues the discussion of the pharmaceutical industry’s manipulation of depression pill trials for children and adolescents. Each Monday, a new section of the book is published, and all chapters are archived here.

The large TADS study of fluoxetine funded by NIH was seriously misreported

There has been one independent trial of fluoxetine for adolescents, the US National Institutes of Health’s Treatment of Adolescent Depression Study (TADS), published in 2004.322 This trial was very large and has been highly influential.

Adolescents with depression (n = 439) were randomised to four treatment groups: fluoxetine alone (n = 109), cognitive behavioural therapy (CBT) alone (n = 111), fluoxetine with CBT (n = 107) or a pill placebo (n = 112) during TADS’ acute phase (12 weeks).

The investigators reported that combination treatment with fluoxetine and CBT “offered the most favourable trade-off between benefit and risk for adolescents with major depressive dis-order.” However, the reporting has been widely criticised. There are issues with study design, statistical reporting and interpretation, discrepancies between article abstracts and their content in the over 30 publications by the TADS team, and misreporting of harms.323

Photo of a white child covering ears with handsA 2020 systematic review criticised 19 international clinical practice guidelines for their reliance on TADS findings without considering the failure of the TADS authors to report adequately on drug harms.324

The TADS authors claimed efficacy and safety for fluoxetine, which is the standard mantra for the drug industry whatever the results are, but both claims are wrong. The effect was not clinically relevant, and there were twice as many suicidal events in patients randomised to fluoxetine than in patients randomised to placebo.322,325

To this day, the reporting on harms remains highly deficient.323 Two researchers who wanted to redress this got access to summary data via the National Institutes of Health.323 These data indicated that, of the 30 serious adverse events recorded during the study’s acute phase, 12 were suicide attempts among children taking SSRIs, compared with only two attempts among children not taking SSRIs.

Next, the researchers tried to get access to the case record forms and narratives, which are essential for a rigorous reanalysis, which the MedDRA (Medical Dictionary for Regulatory Activities) coded terms and severity ratings do not allow. The researchers’ previous experience with restoring GlaxoSmithKline (GSK) study 329 of paroxetine in children and adolescents had shown that this additional step is very important in order to correct the errors made by the original investigators, which changed the harms significantly in disfavour of paroxetine.300

However, Duke University, where the trial data were lodged, refused to hand over serious adverse event forms from the trial even though they had signed an agreement about delivering the data.323 Their arguments for refusal were invalid.

The researchers then tried to get the missing data from Eli Lilly, which provided the drug for the trial and had received all the reports of serious adverse events from the investigators, but Lilly refused to release the data and also to have any of the correspondence published.

The researchers also tried to get the data from the FDA but were told it would take at least two years before they came up in the queue.

Other depression pills are also unsafe and the pediatric trials are manipulated

The increase in the risk of suicide and violence is not limited to fluoxetine. It is a class effect. My research group used the clinical study reports of the placebo-controlled trials of SSRIs and SNRIs, and we found that these drugs increase suicidality and aggression 2-3 times among children and adolescents, odds ratios 2.39, (1.31 to 4.33) and 2.79 (1.62 to 4.81), respectively.326

Prior to the development of the SSRIs, there had been 15 randomised trials of tricyclics and related compounds in children and adolescents, all negative.327 There was a clinical consensus that children did not get endogenous depression.279 They might be miserable and unhappy, but this was situational and would respond to psychosocial interventions. Linked to this, there were almost no child psychiatrists with expertise in psychopharmacology.

The SSRIs at that time had an anxiolytic action but were marketed as depression pills in part to skirt around clinical concerns that any new anxiolytic would necessarily produce dependence as the benzodiazepines had.328 A 2017 meta-analysis of published trials in children and adolescents confirmed that SSRIs are essentially anxiolytic drugs, as the effect sizes were significantly larger for anxiety (0.56) and obsessive-compulsive disorder (0.39) than for depression (0.20).329

Our reanalysis of the two pivotal fluoxetine trials279 made it clear that this drug should never have been approved for use in children and adolescents. But once approved, it paved the way for approval of other ineffective and dangerous SSRIs.

After licensing fluoxetine for children, based on studies negative on their primary endpoint, the FDA issued an approvable letter in October 2002 for paroxetine, which came to light because of a court case:330 “We agree [with GSK] that … the results from Studies 329, 377, and 701 failed to demonstrate the efficacy of Paxil in pediatric patients with MDD [major depressive disorder]. Given the fact that negative trials are frequently seen, even for antidepressant drugs that we know are effective, we agree that it would not be useful to describe these negative trials in labeling.”

This is one of the most horrible statements I have ever seen a drug regulator make. “The drug didn’t work, but we know it works.” If so, then why bother doing randomised trials? This is how practitioners of homeopathy or Chinese medicine and other quacksters argue.

In the initial 2001 publication of study 329, which was a trial of paroxetine in depressed minors, GSK claimed paroxetine was safe and effective.331 But an internal document from 1998 reveals that GSK knew the study demonstrated its drug to be ineffective, which GSK considered would be commercially unacceptable to publish.332,333 The document states that the “good bits of the study would be published.”

The study was negative for efficacy on all eight protocol-specified outcomes and positive for harm. But GSK tortured the data until they confessed.332,334 The paper didn’t leave any trace of the torture; in fact, it falsely stated that the new outcomes were declared a priori – a classical Texas sharpshooter fraud.

Based on this information, New York State’s Attorney General lodged a fraud action against GSK in 2004.122 The settlement of this action made it possible to access data on study 329 and restore it in a manner that demonstrated paroxetine’s lack of efficacy and increase in suicidal events in contrast to the original publication,331 which was fraudulent. Seven children on paroxetine versus one on placebo demonstrated suicidal or self-injurious behaviour.300 In the published paper, five cases of suicidal thoughts and behaviour were listed as “emotional lability” and three additional cases of suicidal ideation or self-harm were called “hospitalisation.”122 When the FDA demanded the company to review the data again, there were four additional cases of intentional self-injury, suicidal ideation, or suicide attempt, all on paroxetine.

The first author on the fraud, Martin Keller, double-billed his travel expenses; was offered $25,000 for each vulnerable teenager; received hundreds of thousands of dollars to fund research that wasn’t being conducted; received hundreds of thousands of dollars from drug companies every year that he didn’t disclose; lectured for patients and their relatives on drug company money, which he didn’t reveal; and his honoraria were whitewashed.122

Keller’s misdeeds didn’t harm his career, likely because his department had received $50 million in research funding. A spokesperson from Brown University School of Medicine said that “Dr Keller’s research regarding Paxil complied with Brown’s research standards.” I see.

The journal that published Keller’s paper, Journal of the American Academy of Child and Adolescent Psychiatry, was complicit in the fraud. Although the journal’s editors were shown evidence that the article misrepresented the science, they refused to convey this information to the medical community and to retract the article.332 An explanation for this passivity can likely be found by following the money that goes to the journal’s owner.

GSK illegally pushed paroxetine for use in children, although it wasn’t approved for children, and withheld trial results showing paroxetine was ineffective.335 The ruthless marketing worked. I have described many heart-breaking stories about children and young adults who were not mentally ill in any way, but who killed themselves by hanging or other violent means because of the harms of the depression pills they took.6:219,7:79 These people were prescribed depression pills because of insomnia, break-up with a girlfriend, stress at work or at school, and other every-day problems.

The approval of fluoxetine for depression in children and adolescents and the publication of many articles since, often ghost-written, claiming efficacy for a number of SSRIs swept away the idea of relying on psychotherapy and other forms of support.279

The FDA was also deceived by other drug companies. In 2002, when GSK applied to get paroxetine approved for children, the FDA wrote to the company:330

“You did not provide any analysis of ECG interval data for the controlled studies. The results provided for studies 701 and 704 consisted of a count of the numbers of patients with ECG abnormalities. In study 329, ECG abnormalities were considered adverse events but were not otherwise analyzed. In order to complete our review of this application, we are requesting that you submit the typical kind of analyses conducted for these type of data; i.e., an analysis of mean change from baseline for measured ECG intervals.”

The FDA furthermore criticised a table that did not show any data from the placebo groups and listed paroxetine-treated children whose adverse events had been coded as hostility, emotional lability, or agitation but did not include psychiatric adverse events that were coded under other terms. FDA requested the narrative case summaries for those events that were either serious or resulted in premature discontinuation.

It is unbelievable that such information was not provided in the application. GSK was also asked to provide its “rationale for coding suicide attempts and other forms of self-injurious behavior under the WHOART term ‘emotional lability.’”

Paroxetine seemed to stunt growth, just as we found for fluoxetine. The FDA requested GSK to statistically test their data on height and weight and to conduct juvenile animal studies to evaluate paroxetine’s effects on growth and neurological, behavioural, and reproductive development. However, as soon as a drug is approved, the drug company tend to “forget” everything the drug regulator has requested. This seemed also to be the case this time. I reviewed FDA’s package insert for paroxetine in 2022,336 and there was nothing that suggested that GSK had done the requested animal studies even though they were very important.

The FDA package insert for fluoxetine shows how dangerous these drugs are.33 It describes a meta-analysis of short-term placebo-controlled trials. For every 1000 children or adolescents treated with drug instead of placebo for a median duration of only two months, there were 14 additional cases of suicidality. The number needed to harm one kid was therefore only 71.

In 2004, the FDA issued a black-box warning on depression pills based on a meta-analysis that showed that the rate of suicidal thinking or suicidal behaviour was 4% among young patients on a depression pill and only 2% among those on placebo, which was a statistically significant difference.303,337 However, when the FDA published the doubling of the suicide risk in children in a medical journal, they called it a “modestly increased risk.”338

While the FDA was reviewing the data, the academics at the medical schools who had published positive results of these drugs were worried and issued a report in January 2004 defending the effectiveness of the drugs and disputing evidence that their use increased suicidal behaviour.339 The academic researchers had contacted the companies to get access to the data they had themselves generated, but some drug companies refused to turn over the data. This decision could not be disputed because the medical schools, in agreeing to run the trials, had signed agreements with the drug makers that kept the data confidential.

Academic medical centres in the United States have set up clinical trial offices and openly court the industry, offering the services of their clinical faculties and easy access to patients.340 Instead of fighting the corruption of academic integrity, the academics participate in a race to the ethical bottom, making it less and less likely that any outsiders will ever get to see the data. Science has shaded into marketing and the professors end up as promoters, while some industry scientists are sickened by the process they have become involved in,341 but there is nothing they can do.

The textbook that has only psychiatrists as authors noted that some people experience agitation or anxiety at the beginning of treatment, especially at young ages, with a possible worsening of suicidal thoughts.18:238 It is not especially in young people; it is not limited to the beginning of treatment but can happen at any time; and it is far worse than just thoughts. Some children kill themselves because of the pills’ harms.7

It is cruel that most psychiatric leaders say—even on Danish national TV, which Lars Kessing did342—that depression pills can be given safely to children because there wasn’t a statistically significant increase in suicides in the trials, only in suicidal thoughts and behaviour, as if there is no relation between the two.

The psychiatrists reward the companies for their fraud while they sacrifice the children. We all know that a suicide starts with suicidal thinking followed by preparations and suicide attempts.


To see the list of all references cited, click here.


Mad in America hosts blogs by a diverse group of writers. These posts are designed to serve as a public forum for a discussion—broadly speaking—of psychiatry and its treatments. The opinions expressed are the writers’ own.


Mad in America has made some changes to the commenting process. You no longer need to login or create an account on our site to comment. The only information needed is your name, email and comment text. Comments made with an account prior to this change will remain visible on the site.


  1. Thank You for your hard work and dedication Dr Gøtzsche,
    Depression as an illness was invented, with Prozac coming onto the market. If you register with a doctor in the UK it is required of you to fill in a form, which at one level or another, might suggest you need an ‘anti depressant’.

    Report comment

    • If depression isn’t an illness would you accept it as a valid symptom of other medical conditions?

      It sounds like a small distinction but a doctor who views depression as a symptom of illness will take a very different approach than a doctor who views depression as the illness.

      Report comment

  2. DNA test are now available for pediatric & adults before any psychological drugs are to be administered. These test show CLEARLY what drugs are compatible w your DNA & which ones are TOXIC to you and can cause suicidal tendency.
    ANY WISE PHYSICIAN will now seek this out b4 prescribing ! DONT be a human guinea pig or a victim of medical negligence.
    ASK for a DNA test. Most insurance companies would rather pay for this than years if bad therapy or meds that don’t work.

    Report comment

    • I don’t believe that is completely true. They can apparently test for people who are “slow metabolizers” of different drugs and so perhaps avoid creating homicidal/suicidal ideation by SOME people taking SSRIs, but these tests appear to be far from foolproof, and just as importantly, almost no physicians regularly use such testing in their practice, despite your assertion that “ANY WISE PHYSICIAN” will do so. I’d be happy to be proven wrong about this, but that’s my understanding of the current situation. I have yet to meet a person currently engaged in the MH system who was offered such testing by their doctor, though I’ve heard of a few who demanded it themselves. Do you have any links to recent studies showing anything more promising than what I’ve identified above?

      Report comment

      • Thank you for your comment.
        I believe the tests you refer to are for metabolizing enzymes labeled CYP450 + subfamilies (such as 2D6, 3A4, etc.)
        These metabolizers are important in breaking down drugs for absorption and excretion, but can incur adverse effects due to the person’s inherited balance or imbalance. (E.g. these are often grouped into: Rapid Metabolizers, which “use up” the drugs quickly, “Normal”, and Slow Metabolizers, which retain the drugs by slowing absorption/excretion and often result in toxicity).
        These differences in the amount or lack of enzymes helps to explain why drugs affect people differently – for example why some become more agitated (Rapid Metabolizers) and others don’t “Improve” right away but experience a buildup leading to toxicity (Slow Metabolizers).
        Then to further complicate matters, different drugs interact if they share certain enzymes that act as “substrates”, so ideally all persons taking more than one psychmed should be aware of this, and possibly get tested.
        But as far as I have researched for myself, this type of testing is drug-specific and only recommended if you have a concern with a prescribed drug or combination of drugs. Concerning severe reactions to psych meds I have personally had, I narrowed down the CYP subfamilies by “connecting the dots” of drugs that were most harmful, then finding out the psychopharmacology (listed usually on Wikipedia) to see which CYP enzyme produced the adverse effects and/or interactions.
        In my case the most problematical were CYP2D6 (for many neuroleptics) and 3A4 (which I believe also metabolizes alcohol). I did try to find out how to get tested, but was discouraged from doing so – I probably might pursue the testing if I were to continue with any neuroleptic, but I haven’t taken any for many years. But I do avoid alcohol.

        I haven’t researched this recently, so do check it out yourself, but I hope this helps someone. (I only wish that psychiatrists would have explained this to me before I had so many “adversities”!)

        Report comment

        • As I intimated, I don’t think most psychiatrists understand or care about fast- vs. slow-metabolizers and so forth. I do appreciate your research, and that was specifically what I was recalling and referring to.

          The difficulty with this approach is that it totally begs the question of what we are supposedly “treating” with these drugs. What we’re doing is manipulating brain chemicals and hoping we get lucky and find a drug that a person can live with that helps reduce his/her “symptoms” without saddling him/her with intolerable adverse effects. But of course, we have no idea if brain chemicals are in fact a causal agent in any of these cases, or how to distinguish the ones that are (if they exist) by any kind of objective means. It’s not really very different from observing that some people can take opioids or drink alcohol and be fine, while some are quickly overwhelmed by the drug’s effects and others seem to take a lot to have any effect at all. All fine information to have, but it hardly substitutes for figuring out and addressing WHY this person is distressed enough to be heading down to the bar every night to imbibe, and what we might do to alter that reality!

          Report comment

      • How can we get reliable drug facts into hands of people who are otherwise not coming to this venue or who are overwhelmed by the ocean of facts? Would QR codes be a useful tool?
        Thank you, Dr. Gøtzsche, for doing the “heavy lifting” for the rest of us.

        Report comment

  3. I am fascinated by this topic as it hits home for me. I was one of those teens that was given Paxil when I was only 11 years old. I can’t help to think about how much different my life could be and have been without it. When my behaviour didn’t improve the dosage went up and I took it for seven years. Where is my justice or even an apology?

    Report comment