Professionalism/Martin Keller, SmithKline Beecham, and Study 329

Study 329 is a clinical trial testing the efficacy of Paroxetine (an established adult antidepressant known by the trade names Paxil and Seroxat) for major depression in adolescents. SmithKline Beecham (later GlaxoSmithKline) owned the patent for Paroxetine and hoped Study 329 would yield positive results that would enable them to license the drug for adolescents. Despite disappointing study results, SmithKline Beecham proceeded with their claim that Paroxetine is effective in adolescents.

Study DetailsEdit

The protocol for the study called for a multicenter, double-blind, placebo-controlled experiment aimed at determining the efficacy of Paroxetine for treating depression in adolescents aged 12–18.[1] The 300 participants in the study were to be between ages 12 inclusive and 19 exclusive, had to have been diagnosed with major depression as according to the DSM-III-R for at least 8 weeks, and had to have a score of 12 or greater on the 17-item Hamilton Depression Scale, or HAM-D. These participants were split equally and randomly into one of three categories: Paroxetine, Placebo, or an already established antidepressant Imipramine. Each category would be given their prescription for an 8-week acute phase and followed with a 6-month extension period, where the acute phase was the main point of focus. In this phase, efficacy was measured through tracking the total HAM-D and K-SADS-L score change from beginning to end, specifically checking if the HAM-D score had lowered to a maximum of 8 or by at least 50%. The study protocol was approved August 26, 1993 with a single amendment being approved April 17, 1994.

The study argues that Paroxetine is “generally well tolerated and effective for major depression in adolescents.”[2] The experimental results tell a different story as 11 of the patients in the paroxetine group suffered adverse effects during the trial. One patient suffered from rebound headaches, two had worsening depression, five had increased in suicidal thoughts or actions, two became more aggressive, and one experienced “euphoria/expansive mood”. Seven of these patients were hospitalized, but only the rebound headache case was considered the result of Paroxetine.

Journal ProposalEdit

In April 1998, Sally Laden and John Romankiewicz of Scientific Therapeutics Information approached SmithKline Beecham’s Ivan Gergel to write and edit a journal article for the Paroxetine study.[3] This journal article was intended to be published within the Journal of the American Medical Association with plans to have Martin Keller as the primary author and up to three additional authors as reviewers. Alongside this, Jim McCafferty of SmithKline Beecham would be included as a primary contact for data and coordination. In December of 1999, Sally Laden faxed to Jim McCafferty a letter outlining that the Journal of the American Medical Association had declined the article and that revisions would have to be made and looked over by all authors before being submitted to the American Journal of Psychiatry.[4]


Study IssuesEdit

The study concluded that Paroxetine significantly outperformed the placebo.[2] The actual numerical results, however, only show significance for a mild improvement through the usage of Paroxetine, as the percentage of patients with a significant lowering of their HAM-D score differed by approximately 11%. Alongside this, the study itself downplays the adverse effects that Paroxetine had shown. The discussion brings up that a fairly large number of participants developed “various psychiatric events”, but none of them were linked to the drug. In comparison to the 10 participants who underwent psychiatric events, the discussion lists 3 participants undergoing similar events for the Imipramine group and no description for the placebo group past having 2 serious adverse effects.

There were also notable concerns with the journal article. The journal article was rejected from the Journal of the American Medical Association and presumably the American Journal of Psychiatry. The article ended up being published in the Journal of the American Academy of Child and Adolescent Psychiatry, or JAACAP. The article also was published with 22 authors, mostly MDs and PhDs. Included in this list of authors is James McCafferty, who had only served to give Scientific Therapeutics Information the information needed to write the article. Sally Laden, despite being a major editor and writer for the article, was not listed as an author. Instead, a footnote was included that acknowledges Laden’s work as “Editorial assistance.”

Another likely reason for this study was that it would extend SmithKline Beecham's patent on Paroxetine for another six months.[5]

SmithKline Beecham's PositionEdit

Prior the publication of an article based on Study 329 in the JAACAP, a Paroxetine position paper was circulated within SmithKline Beecham. It recognized that the clinical results from Study 329 were "insufficiently robust to support a label change” because the drug only "showed trends in efficacy" but "no statistical significant difference from placebo."[6] Because the data was unsuitable to ask for a label change, they concluded that they did not need to submit the results of the study to the regulatory authorities. A memo attached to the paper stated that not even a safety statement would be published because Marketing confirmed it would be "unacceptable commercially."[7] The goal of the position paper was to "effectively manage the dissemination of these data in order to minimize any potential negative commercial impact."[6]

GlaxoSmithKline (Glaxo Wellcome merged with SmithKline Beecham in 2000[8]) pushed forward with marketing Paroxetine to adolescents through a common practice in medicine called off-label prescriptions. Off-label prescriptions are when a physician prescribes a medicine for a condition it was not approved to treat, but in the physician's professional opinion, the medicine would help the patient.[9] In 2001, Zachary Hawkins of Paxil Product Management sent a memo to "All Sale Representatives Selling Paxil" lauding the study published in JAACAP. He claimed "Paxil demonstrates REMARKABLE Efficacy and Safety in the treatment of adolescent depression."[10] The memo also claims that "Paxil was significantly more effective than placebo" for the HAM-D, CGI, and K-SADS criteria[10] which is patently false. It repeated the conclusion of the journal article that "Paxil was generally well tolerated in this adolescent population."[10]

BBC ExposureEdit

In 2002, BBC Panorama ran a story on Seroxat (trade name for Paroxetine in the UK) and delved into the details of how GlaxoSmithKline (GSK) spun the results of Study 329. This was the first major public exposure for Study 329 and produced a lot of negative coverage for GSK. Panorama had only planned to make one or two episodes about Seroxat and GSK, but due to massive public outcry and listener submitted testimonials, they made four episodes related to user experience with Seroxat and drug licensing.

After reaching negative results in Study 329, GSK decided to emphasize the purported benefits and downplay the risks. And because the evidence was not in for a label change of the drug, they had to convince doctors that the drug was safe for children. Panorama alleges that the paper published on Study 329 was mostly ghost written but listed many doctors and researchers with good reputations as authors to enhance the legitimacy of the paper's conclusion. The chief author was Martin Keller, a respected Brown University Psychiatry professor. For a chief author, Keller professed to "tend not to look at those [data analytic tables]. I do better with words than I do with symbols."[5] He was also aware of the extent that the paper was ghostwritten because he exchanged communications with the ghostwriter. He compliments the writer for doing "a superb job with this. Thank you very much. It is excellent" and adds "some rather minor changes from me."[5] Later he is told to "please re-type [the study] on your letterhead."[5] Another author, Dr. Neal Ryan at the University of Pittsburgh, asked GSK for advice on how to respond to the reporter's questions about Paroxetine.[5]

Medical EthicsEdit

The UK's Medicines and Healthcare products Regulatory Authority (MHRA) opened an investigation into GSK about whether they had acted criminally in withholding clinical trial data from regulators. In 2008, the MHRA opted to decline prosecution of GSK despite clear evidence that GSK knew about the side effects and dubious efficacy of Paroxetine in adolescents and actively withheld that information.[11] The decision not to prosecute was not due to a lack of evidence against GSK but because there was enough ambiguity and loopholes in the regulations in 2003 to "make a criminal prosecution impossible." One loophole GSK exploited was that the EU legislation did not "require companies to disclose adverse events from trials in groups of patients for whom the medicine was not licensed."[12] Companies are required to report to MHRA any adverse reactions from their product, but this could be construed to only apply during "normal conditions of use."[11] Neither clinical trials nor off-label prescriptions legally count as normal use. However, the UK 1968 Medicines Act (incidentally passed after the thalidomide fiasco) does require companies to report adverse effects from clinical trials, but only for trials taking place in the UK (Study 329 occurred in the US). Furthermore, a failure to report was not a criminal offense.[11]

GSK did not engage in illegal behavior or, at least, in behavior that could be reliably argued to be illegal. However, they were certainly operating in a morally gray area. MHRA assumed that "[the pharmaceutical] industry has a very strong vested interest in not actually stepping over the line."[11] In a letter to GSK following the decision to not prosecute, Kent Woods (CEO of MHRA) asserts that "moral responsibility now needs to be insisted upon by the unambiguous force of the law."[13]


Martin Keller, GlaxoSmithKline, and Ethical BehaviorEdit

In 2003, the MHRA [14] banned the use of Paroxetine and several other antidepressants for use in adolescents. In 2012, the United States Department of Justice fined GlaxoSmith Kline $3 billion to settle criminal and civil liabilities for several drug-related violations including misbranding Paxil.[15] Misrepresenting study data for financial gain is unethical, but GSK's significant financial losses relating to their cavalier treatment of Study 329 should also serve as warning. The false representation of the journal article's authors also highlights the unethical nature of claiming work that is not one's own. This is especially unethical when done to put the weight of one's professional license behind the work.

Further ResearchEdit

Topics for future study could include:

  • Journal rules of retraction
  • Industry influence on regulatory agencies
  • Clinical trial laws and reporting serious adverse side effects
  • Antidepressant use generally in adolescents
  • Other adolescent trials of antidepressants


  1. SmithKline Beecham Pharmaceuticals. Study Drug:BRL 29060/Paroxetine(Paxil). (1994, April 17). Retrieved from
  2. a b Keller, M. B., Ryan, N. D., & Wagner, K. D. (2002). Paroxetine In Adolescent Major Depression. Journal of the American Academy of Child & Adolescent Psychiatry, 41(4), 364. doi:10.1097/00004583-200204000-00004
  3. Laden, S. K., & Romankiewicz, J. A. (1998, April 3). Adolescent Depression STUDY 329 Proposal for a Journal Article. Retrieved from
  4. Laden, S. (1999, December 7). RE: Publication Status - Paroxetine Adolescent Depression Study [Letter to Jim McCafferty]. SmithKline Beecham Pharmaceuticals, Collegeville, PA.
  5. a b c d e BBC Panorama. (2007). Secrets of the drug trials[Episode transcript]. Retrieved from
  6. a b Wilson, J. (1998). Wilson Position Piece Phase III (Rep.). Retrieved May 6, 2019, from
  7. Westaway, J. (1998, October 14). Seroxat/Paxil in Adolescent Depression [E-mail].
  8. GSK today: 2000 – present. (n.d.). Retrieved May 6, 2019, from
  9. Off-Label Drugs: What You Need to Know. (2015, September 01). Retrieved from
  10. a b c Hawkins, Z. (n.d.). [Letter written August 16, 2001 to All Sales Representatives Selling Paxil]. Retrieved May 6, 2019, from
  11. a b c d Mcgoey, L., & Jackson, E. (2009). Seroxat and the Suppression of Clinical Trial Data: Regulatory Failure and the Uses of Legal Ambiguity. Journal of Medical Ethics, 35(2), 107-112. doi:10.1136/jme.2008.025361
  12. A legal framework for drug safety. (2008). Bmj, 336(7644). doi:10.1136/bmj.39518.627060.47
  13. Woods, K. (n.d.). [Letter written March 6, 2008 to Dr. Jean-Pierre Garnier]. Retrieved from
  14. Kondro, W., & Sibbald, B. (2004, March 02). Drug company experts advised staff to withhold data about SSRI use in children. Retrieved from
  15. GlaxoSmithKline to Plead Guilty and Pay $3 Billion to Resolve Fraud Allegations and Failure to Report Safety Data. (2015, May 22). Retrieved from