The Troubled Life of Study 329: Consequences of Failure to Retract
If someone were to ask the surviving authors of Study 329 the question: “Knowing what you know now, if you had to do it over, would you agree to participate in that study again?”, many would probably say no. The real title of Study 329 was “Efficacy of Paroxetine in the Treatment of Adolescent Major Depression, A Randomized Controlled Trial”. It was published in the Journal of the American Academy of Child and Adolescent Psychiatry (JAACAP) in July, 2001. Although it is probably fairly typical of medical research studies, it has been plagued by more than its share of troubles. Next week, fourteen years and two months after it was published, it is about to take yet another hit, when the Restored version is published.
There were many obstacles to doing the work to write Restoring Study 329: A randomized, controlled trial of the efficacy and harms of paroxetine and imipramine in the treatment of adolescent major depression. The biggest was getting access to the data. It took years to get the data, months to analyze it, and another year to get a version of the restored study that BMJ felt comfortable with. The “Restoration” of Study 329 may not have occurred if Study 329 had been retracted. If JAACAP had heeded the earliest calls for retraction, much of the unfortunate history of Study 329 could have been avoided.
In hindsight, the authors may wish that they had not been so adamant that retraction was unthinkable. Of course, there are probably many flawed studies that remain un-retracted. Many of the pitfalls into which Study 329 fell resulted from bad timing, coincidence, an astute and determined journalist, a Senator who found conflicts of interest in medicine unacceptable, and the unwavering commitment of a small group of medical experts who refused to give up and go away. How could those authors have foreseen?
Study 329’s problems started to surface right after it was published. Several doctors wrote letters to the JAACAP Editor with probing questions, mostly centred on the psychiatric side effects of paroxetine, and the measures used to claim its efficacy in treating adolescents. The authors responded and the questioners did not pursue their concerns further. Except one. Child Psychiatrist Jon Jureidini, M.D. from the University of Adelaide remained convinced that there were serious methodological problems hiding real harms, and he never stopped pushing to expose the truth.
After Study 329 was accepted but before it was published, a Wyoming jury awarded $6.4 million to the relatives of retired oilman Don Schell (Tobin v.SmithKline Beecham). Forty-eight hours after Mr. Schell had been prescribed paroxetine (Paxil), he put bullets through the heads of his wife, Rita, his daughter, Deborah, and his granddaughter, Alyssa. Then he shot and killed himself. The jury decided that Paxil was responsible for the tragedy.
Scottish journalist Shelley Jofre, having learned about the Schell case, did some research on the drug and arranged to do a program on paroxetine, known as Seroxat in the U.K., for BBC’s Panorama. The segment, Secrets of Seroxat, aired in October 2002. It revealed that the drug can cause suicidal and violent thoughts and behaviours, and that many people experience serious withdrawal problems. The show received such a huge public response that a second episode was developed, Emails from the Edge.
Following Secrets of Seroxat, the U.K. public was highly sensitized to the potential problems associated with paroxetine. In December of the year that Emails from the Edge was broadcast, the Medicines and Healthcare Products Regulatory Agency (MHRA) issued an advisory warning to physicians against prescribing SSRI antidepressants, including Seroxat, to people under age 18.
In June of 2004, New York State Attorney General Elliot Spitzer filed a lawsuit accusing British drug giant GlaxoSmithKline PLC of “repeated and persistent fraud” for concealing known problems with efficacy and safety of Paxil (paroxetine) for children and adolescents. Evidence was largely based on documents that had come to light in the Tobin case. The suit was settled for $2.5 million plus a promise by GSK that they would make their drug trial data accessible. Later that year, the FDA required that a “black box” warning label be added for all SSRI antidepressants, including Paxil.
In 2008, US Senator Chuck Grassley investigated serious violations of conflict of interest policies among high-profile academic psychiatrists, including Martin Keller, lead author of Study 329, and three of his coauthors. Senator Grassley’s investigation brought attention to the fact the nominal authors in many research studies are not the real authors. This was the case in Study 329. GSK hired a “ghostwriter”, Sally Laden of STI, to write the study article under GSK direction.
Over the next few years, many lawsuits involving suicidality and violence caused by paroxetine were settled. In addition, Paxil and other SSRI drugs became the target of lawsuits for causing birth defects. The FDA recognized suicidality and violence as side effects of SSRI antidepressants in all age groups and expanded the required “black box” warning.
In 2012, the U.S. Department of Justice brought an action in U.S. District Court to recover damages and civil penalties from GSK under the False Claims Act, and damages and other monetary relief under common law and equity for causing the submission of false or fraudulent claims to federal health care programs. Three drugs were implicated, including Paxil. This action was settled for $3 billion, the largest settlement of its kind.
All in all, the past decade was not an unqualified success for Paxil and its manufacturer. Still, the conclusion of Study 329, that “Paroxetine is generally well tolerated and effective for major depression in adolescents”, has stayed officially intact, since the study was never retracted, and has been widely cited.
In 2013, Peter Doshi and a number of other researchers published a paper in the BMJ entitled “Restoring invisible and abandoned trials: a call for people to publish the findings”. This is part of a larger campaign to make all data from drug trials available for public scrutiny.
In August, 2015, BMJ accepted for publication a new analysis of the data from the original Study 329. Restoring Study 329: A randomized, controlled trial of the efficacy and harms of paroxetine and imipramine in the treatment of adolescent major depression will be published online sometime the week of Sept 14, and in the hard copy journal later that week. The new team of authors, including those most active in lobbying for retraction, has taken the same data and reached startlingly different conclusions than the original. Their conclusions should lead us to reassess our assumptions about drug safety regulation.
Stay tuned to this blog as the details continue to unfold. See also 1boringoldman for a thorough ongoing analysis of this remarkable saga…
Today on the Study 329 site we have added a new section under “Background” which provides a summary of the attempts to get the original Study 329 retracted, together with links to the underlying documents.