Re-analysis of controversial Paxil study shows drug “ineffective and unsafe” for teens

The antidepressant Paxil isn’t safe or effective for teens after all, says a re-analysis of a 2001 study published today in The BMJ.

The original 2001 paper in Journal of the American Academy of Child and Adolescent Psychiatry — study 329, as it’s known — helped greenlight use of the drug (generically known as paroxetine) in young people. But it’s faced accusations of ghostwriting, undisclosed conflicts of interest, and issues with data analysis since publication.

According to a BMJ feature, also published today:

The disparity between what the manufacturer and study authors claim the trial found and what other parties say the data show was an important element in the US Department of Justice’s criminal charges against [drugmaker GlaxoSmithKline]. In 2012, GSK was fined a record $3bn (£2bn; €2.7bn), in part for fraudulently promoting paroxetine.

And over the years, many researchers have called for the retraction of the 2001 paper, a stance bolstered by the new paper.

The new analysis in BMJ includes documents that were previously confidential, such as “about 77,000 pages of de-identified individual case report forms” posted by GSK “after negotiation.” It concludes:

On the basis of access to the original data from Study 329, we report a reanalysis that concludes that paroxetine was ineffective and unsafe in this study.

The new paper found a discrepancy between instances of harm listed in the appendix of the original paper, and those reported on case report forms:

We discovered adverse events recorded onto case report forms but not transcribed into the patient level listings of adverse events in appendix D of the [clinical study report].

The original paper also downplayed the harm that they did report, according to the BMJ authors. The first author of the original paper, and the accompanying clinical study report:

…mostly reported suicide related events as “emotional lability.”

The original paper, “Efficacy of paroxetine in the treatment of adolescent major depression: a randomized, controlled trial,” was cited 334 times, according to Thomson Scientific’s Web of Knowledge. It found:

Paroxetine is generally well tolerated and effective for major depression in adolescents.

In a two-page statement [PDF] sent to us by Martin Keller, first author on the original paper and a Professor Emeritus of Psychiatry and Human Behavior at Brown University, nine of the 22 authors voice “strong disagreement” with the reanalysis. They take issue with critiques of sponsor involvement, the analysis of side effects, and the charge that it was ghostwritten (the “report was authored mainly by the academic investigators with industry collaboration,” they write.) Ultimately, they conclude

…to describe our trial as “misreported” is pejorative and wrong, both from consideration of best research practices at the time, and in terms of a retrospective from the standpoint of current best practices.

The BMJ feature authored by the journal’s associate editor, Peter Doshi, highlights the swift disconnect in the early ’00s between the skepticism of the trial, and the glowing version of Paxil presented to the public:

In 2002, a US Food and Drug Administration officer who formally reviewed the trial reported that “on balance, this trial should be considered as a failed trial, in that neither active treatment group showed superiority over placebo by a statistically significant margin.” Yet this same year, according to the New York State Attorney General’s office, which sued GSK, over two million prescriptions were written for children and adolescents in the United States, all off-label, after a marketing campaign that characterized Study 329 as demonstrating “REMARKABLE Efficacy and Safety.”

On study329.org, the authors of the new paper detail attempts to have the original paper retracted. Those have been unsuccessful, The BMJ feature explains:

It is often said that science self corrects. But for those who have been calling for a retraction of the Keller paper for many years, the system has failed. None of the paper’s 22 mostly academic university authors, nor the journal’s editors, nor the academic and professional institutions they belong to, have intervened to correct the record. The paper remains without so much as an erratum, and none of its authors—many of whom are educators and prominent members of their respective professional societies—have been disciplined.

The current editor of JAACAP, Andres Martin, has stood by the paper, notably in an assembly of the American Academy of Child and Adolescent Psychiatry, which publishes the journal, last October. The BMJ feature reports:

According to the minutes, members heard how Martin had investigated the matter thoroughly by consultation with the authors, the Committee on Publication Ethics (COPE), clinical experts, “a whole range of attorneys, and more.” Martin’s assessment, completed in July 2010, concluded that no further action was necessary.

However, there’s no public information about that investigation:

No specific findings from Martin’s investigation are recorded in the minutes, and Martin did not respond to multiple requests for comment from The BMJ.

The BMJ even interviewed our own Ivan Oransky (under embargo):

Ivan Oransky, cofounder of the Retraction Watch blog, says that transparency is vital. “GSK agreed to pay a $3bn fine and you’re [Martin] saying you had completely different results? Great. Show me.”

Oransky described Martin’s silence as part of the “typical scientific playbook.” “It has certainly been our experience that journals and researchers and institutions can be incredibly stubborn about failing to retract a paper, about ignoring calls, or not responding favourably to calls to retract.”

We contacted Martin for this piece ourselves,  and will update if we hear back.

In response to a call for retraction addressed to Brown University, The BMJ feature reports that Edward Wing, former dean of medicine and biological sciences, wrote in 2011:

The University takes seriously any questions about the soundness of faculty-conducted research. While we cannot comment on individual personnel cases, we do take appropriate actions whenever we receive substantive concerns about the conduct of research. The University gathers relevant information, conducts internal reviews if appropriate, and makes any adjustments or corrections that may be warranted.

That does not shock the leadership here at Retraction Watch, as The BMJ reports: 

Retraction Watch’s Oransky says he is not surprised by Brown’s silence. “Universities have become more corporate than many corporations.” Oransky argues that institutions like Brown should confront the problem. Instead of silence, the message from Brown should be: “There is fraud in science; it’s pretty rare, but here is what we’re doing to get rid of it because we’re upset about it.”

“The controversy will not end if they simply stop talking about it. If anything, it will get worse.”

We contacted the Brown communications office, and will update this post if we hear back.

The paper is the first publication of a BMJ initiative called “Restoring invisible and abandoned trials” (RIAT), designed to open up conversations with old data. But its success will depend on who else is willing to talk — BMJ co-author Joanna Le Noury told us that, in this and future RIAT cases, reanalysis:

…could be relatively quick and easy do if both the pharmaceutical companies cooperated and academic journals were less nervous.

An editorial in BMJ entitled “Liberating the data from clinical trials,” also published today, notes that reanalysis is rare, and often revealing:

In a recent review, Ebrahim and colleagues identified just 37 published reanalyses of clinical trials. Only five were conducted by investigators not associated with the original report. A third of the reanalyses led to interpretations that were different from those of the original articles.

As for the status of the original Study 329, BMJ co-author John Nardo told us:

Of course it should be retracted, but at this point, “the ball’s in their court.”

Update 9/18/15 11:02 a.m. eastern: The authors of the BMJ study have posted a response to the two-page letter from Keller at al that was forwarded to us.

Like Retraction Watch? Consider making a tax-deductible contribution to support our growth. You can also follow us on Twitter, like us on Facebook, add us to your RSS reader, and sign up on our homepage for an email every time there’s a new post. Click here to review our Comments Policy.