telling the truth as a liability…

Posted on Wednesday 26 December 2012

I wasn’t satisfied with the timeline on that last post [the lesson of Study 329: an unfinished symphony…]. It didn’t capture the story in context. Here’s how it actually went. The Study 329 article came out in 2001, but the actual dates of the study itself were April 20th, 1994 to May 7th, 1997 [begun shortly after Paxil was launched in 1993]:
Efficacy of Paroxetine in the Treatment of Adolescent Major Depression: A Randomized, Controlled Trial
Journal of the American Academy of Child and Adolescent Psychiatry, 2001, 40(7):762–772.
Conclusions: Paroxetine is generally well tolerated and effective for major depression in adolescents.

By the time it was published, GSK had funded two further completed Clinical Trials of Paxil in Children and Adolescents:

  • Study 329 Dates: 20 April 1994 to 07 May 1997
  • Study 377 Dates: 26 April 1995 to 15 May 1998
  • Study 701 Dates: 20 March 2000 to 24 January 2001

Neither Study 377 nor Study 701 were published:

  • SKB [GSK] didn’t submit Study 329 to the FDA for a pediatric approval, but Studies 329 and 377 were submitted in 2002 as part of a package for a pediatric extension of exclusivity. The FDA said of Study 329:
  • Drs. Juriedini and Tonkin wrote their letter to the JAACAP criticizing the paper, it was published in 2003 [the lesson of Study 329: naked Emperors, fractious Queens… ] along with the author’s response. They said:
    We believe that the Keller et al. study shows evidence of distorted and unbalanced reporting that seems to have evaded the scrutiny of your editorial process… Thus a study that did not show significant improvement on either of two primary outcome measures is reported as demonstrating efficacy. Given that the research was paid for by Glaxo-Smith-Klein, the makers of paroxetine, it is tempting to explain the mode of reporting as an attempt to show the drug in the most favorable light…
  • Added to that, Attorney General Elliot Spitzer had successfully prosecuted GSK in 2004 in New York, focusing on Study 329. As part of that settlement, GSK was required to publish the results of their Clinical Trials on-line. They did publish summaries [but did not put up any raw data - a movement… ].
  • But most damaging of all, in 2004 after a meta-analysis and hearings, the FDA issued its black box warning on SSRIs, particularly for youth.
So as things stood in 2004, GSK wasn’t looking so hot. They had their Study 329 in print being handed out by sales reps, yet it was being increasingly questioned. It had been declared a failed study by the FDA. The meta-analysis of the SSRIs in Adolescents was suggestive enough for the FDA to add a black box warning to the package insert. GSK was sitting on two other unpublished Clinical Trials that showed neither efficacy nor safety for Paxil – certainly different from the reported results in the Keller et al paper. It looked like they were intentionally withholding negative studies [which they were]. So their wish to publish a review article including all three of their Clinical Trials was a CYA move both with efficacy and safety. That’s what had Dr. Keller so stirred up. We don’t have the first draft of that paper, but apparently Keller et al didn’t look so good, because the other two [Studies 377 and 701] didn’t come close to confirming what they had reported. Thus his indefensible rant:
In the final version of the review as it was published, GSK did fold all three studies together, reporting an increased incidence of suicidal thinking, and didn’t single out Study 329:
Evaluation of suicidal thoughts and behaviors in children and adolescents taking paroxetine.
by Apter A, Lipschitz A, Fong R, Carpenter DJ, Krulewicz S, Davies JT, Wilkinson C, Perera P, and Metz A.
Journal of Child and Adolescent Psychopharmacology. 2006 16[1-2]:77-90.
Conclusions: Adolescents treated with paroxetine showed an increased risk of suicide-related events. Suicidality rating scales did not show this risk difference. The presence of uncontrolled suicide risk factors, the relatively low incidence of these events, and their predominance in adolescents with MDD make it difficult to identify a single cause for suicidality in these pediatric patients.

And GSK followed the review by quickly publishing Studies 377 and 701, though they were 5 and 8 years old by the publication time:

An International, Multicenter, Placebo-Controlled Trial of Paroxetine in Adolescents with Major Depressive Disorder
by Ray Berard, Regan Fong, David J. Carpenter, Christine Thomason, and Christel Wilkinson
Journal of Child and Adolescent Psychopharmacology. 2006 16[1-2]:59–75.
Conclusions: No statistically significant differences were observed for paroxetine compared with placebo on the two prospectively defined primary efficacy variables. Paroxetine at 20–40 mg/day administered over a period of up to 12 weeks was generally well tolerated.
Paroxetine Treatment in Children and Adolescents With Major Depressive Disorder: A Randomized, Multicenter, Double-Blind, Placebo-Controlled Trial
Journal of the American Academy of Child and Adolescent Psychiatry. 2006 45[6]:709-719.

Conclusions: Paroxetine was not shown to be more efficacious than placebo for treating pediatric major depressive disorder.

Notice that the latter had three authors in common with Study 329, was published in the same Journal [JAACAP], and was accepted by the same Editor, Dr. Myna Dulcan. In that last paper, they summarized:
"Paroxetine has been studied in three randomized, double-blind, placebo-controlled trials in depressed children and/or adolescents. The results of the first two trials, which enrolled adolescents only, produced equivocal results overall (Berard et al., 2006; Keller et al., 2001). The first, a study comparing paroxetine, imipramine, and placebo in adolescents with major depressive disorder (MDD), failed to show a difference between paroxetine and placebo on the primary outcome (Hamilton Depression Rating Scale score <8 or 50% decrease on it). However, several secondary outcome measures did show differences between paroxetine and placebo, but not between imipramine and placebo (Keller et al., 2001). Thus, there was some evidence of positive benefit for adolescents taking paroxetine in the first RCT. The second study, which was an international study of both inpatients and outpatients, did not show a difference between paroxetine and placebo on any outcome measures. The third study, described here, was the only one to include patients younger than 12 years."
Notice, "The results of the first two trials, which enrolled adolescents only, produced equivocal results overall… The first, a study comparing paroxetine, imipramine, and placebo in adolescents with major depressive disorder (MDD), failed to show a difference between paroxetine and placebo on the primary outcome…"

By any criteria, the conclusion to the 2001 Keller et al article is wrong ["Paroxetine is generally well tolerated and effective for major depression in adolescents"]. That’s not the reason that some of us have been so persistent in pushing for retraction. The reason is that it was wrong at the time the article was published, and they knew it was wrong, but published it anyway. We want to make a statement about the integrity of the scientific literature. Now, eleven years later, the process of getting that article into print and using it for false advertising is known in precise detail. The offending pharmaceutical company has settled the case and paid a $3 B fine. So the question on the table now is "Why doesn’t the Journal of the American Academy of Child and Adolescent Psychiatry retract the article?" I don’t know the answer, but I expect it has to do with things lawyers say – like this:

Journal Article Didn’t Mislead, Drug Company Asserts
Chronicle of Higher Education
October 1, 2012
[full text available on-line]
To the Editor:
As counsel for GlaxoSmithKline LLC in its recently concluded settlement with the United States government, I write to correct some significant factual inaccuracies in "Academic Researchers Escape Scrutiny in Glaxo Fraud Settlement" [The Chronicle, August 6]. The piece focuses on a peer-reviewed journal article published in 2001 in the Journal of the American Academy of Child and Adolescent Psychiatry reporting the results of a clinical trial sponsored by Glaxo­SmithKline [Study 329] of Paxil® [paroxetine hydrochloride] for the treatment of major depressive disorder in adolescents [the article was "Efficacy of Paroxetine in the Treatment of Adolescent Major Depression: A Randomized, Controlled Trial," by M.B. Keller et al., in the July 2001 issue]. Unfortunately, your piece incorrectly states that, as part of its government settlement, GlaxoSmithKline admitted that the journal article "was part of the fraud" and "constituted scientific fraud." In fact, GlaxoSmithKline made no such admission and vigorously disputes that the journal article was false, misleading, or fraudulent.
As part of its settlement with the government, GlaxoSmithKline pleaded guilty to a misdemeanor violation of federal law related to the marketing of Paxil for use by patients under 18, an unapproved, "off label" use for Paxil. While the law permits physicians to prescribe medications for such off-label uses, manufacturers are not permitted to market them for such purposes. The charging document filed by the United States, known as an information, contains many allegations that are either inaccurate or incomplete, that tell only part of the story, and that draw unwarranted conclusions from disputed facts. Throughout the investigation, the settlement process, and the guilty-plea proceeding, GlaxoSmithKline made clear to the government and to the court that it did not agree with all of the factual allegations made in the information. In particular, GlaxoSmithKline has consistently denied and disputed the allegations that the journal article was false or misleading. Nothing in the settlement or the guilty plea has changed Glaxo­SmithKline’s position on the journal article. Most important, Glaxo­SmithKline did not admit and has not admitted that the journal article was part of any fraud or that it was false or misleading.
Your piece also incorrectly states that the journal article "was actually written by Glaxo-hired authors to overstate the benefits and understate the risks of a highly profitable Glaxo drug." While GlaxoSmithKline did hire a medical writer to provide editorial assistance to the clinical investigators, a fact that the journal article itself acknowledged, control over the contents of the manuscript remained at all times with the clinical-investigator authors who provided substantial comments on and input into the manuscript. GlaxoSmithKline believed then and continues to believe that the journal article reflects the honestly held views of the clinical-investigator authors.
Thomas H. Lee II

Editor’s note: The federal "criminal information" document in the Paxil case repeatedly describes the journal article as "false and misleading." It also notes that GlaxoSmithKline distributed the article to its marketing department and its sales representatives, who used it to promote the use of the drug for children and adolescents. In addition, the criminal information says that a contractor hired by Glaxo­SmithKline wrote the first draft of the article and incorporated changes made in the article by researchers and another GlaxoSmithKline employee. The plea agreement between GlaxoSmithKline and the U.S. Department of Justice states that GlaxoSmithKline "expressly and unequivocally admits that it committed the crimes charged in the information, and is in fact guilty of those offenses."
I have no idea why the GSK lawyer wrote this absurd letter. I’m sure it has something to do with some liability he envisions. I expect there are some equally paranoid lawyer[s] advising the JAACAP. Whatever their reason for seeing telling the truth as a liability doesn’t matter. Such considerations have absolutely no place in a scientific dialog…
    December 26, 2012 | 3:30 PM

    Study 329, JAACAP, and Trust

    Why should any clinician and/or researcher reading a JAACAP abstract today in 2012/2013 care about this over 10 year old paper and the upcoming 10 year anniversary of the last published exchange on it in the JAACAP?


    The heart of the matter is not the primary endpoints. Everyone concedes that paroxetine did not separate from placebo on those.
    What we know now only because of the documents released as a result of state and federal prosecution, and which otherwise would have remain known in full only to GSK (and presumably the primary authors), is that the secondary analyses break down as follows:
    “There was no significant efficacy difference between paroxetine and placebo on the two primary outcomes or six secondary outcomes in the original protocol. At least 19 additional outcomes were tested. Study 329 was positive on 4 of 27 known outcomes (15%).”

    What we know now only because of the release of the original annonymized patient level data as a result of state and federal prosecution, and which otherwise would have remain known in full only to GSH (and presumably the primary authors), is this additional point:
    Only 1 of those 4 positive (out of 27) remains positive with a Bonferroni correction needed for the number of outcomes reported in the paper (8 outcomes) let alone the inclusion of the at least 19 additional negative outcomes not reported in the original abstract or Ryan et al’s 2003 response.

    The primary readers of the abstracts and correspondence in JAACAP are obviously clinicians and researchers. Thus the heart of the matter is not a legalistic one.

    Review the abstract here:

    Review Dr. Ryan’s first draft of the final thing that JAACAP has seen fit to publish on this matter, the Ryan et al response in May 2003:

    Could even an astute reader of the abstract, along with the last thing that JAACAP has published on this matter (Ryan et al’s response in May 2003), reasonably be expected to be aware of the additional secondary endpoint analyses that were not reported? And, in addition could they have reasonably been expected to conclude that the authors were choosing to “fairly interpret” the above findings as a “strong signal of efficacy” that in anything but a legalistic sense could be felt to support the conclusion stated in the abstract that “Paroxetine is generally well tolerated and effective for major depression in adolescents.”?
    The findings of a statistically significant positive separation from placebo on one post-hoc secondary endpoint (had there been statistical correction for multiple analyses), or 4 positive post-hoc secondary endpoints. In the context of 2 negative primary endpoints, all 6 pre-specified secondary endpoints, and at least 15 additional secondary endpoints analyzed. Of which the reader is aware only as a result of the release of documents during the last decade of litigation? Documents and data that originally GSK, and the primary authors, would have no reason to believe would ever become fully publically available. That are now available as a result for review by any impartial third parties who would wish to check the accuracy of the above summary.
    There may not necessarily be anything technically inaccurate in the abstract and letter above. It appears the editorial judgment of the journal can assure the reader that the abstract, and later author responses, could be relied on to be technically accurate. If that represents the standard of the journal then so be it.

    The journal has given no explicit indication that the handling of the abstract and later responses in any way falls below its standards. No indication that anything about the abstract and later author response falls short enough of their standard a study being “fairly interpreted” as to provide a basis for further comment.

    As a result of the release of documents and data surrounding litigation related to Study 329 and the resultant paper there is more known about the handling of this paper than about almost any other study in the literature.

    This is from the editorial response published in 2011 in The Spine Journal resulting from concerns about the integrity of papers that had been published many years before in that journal. They have not retracted those papers, however they did not appear comfortable with giving the impression that they had no serious issue with what had occurred. They wrote in part:
    “Clearly, the entire concept of peer-reviewed literature, systematic topic reviews, and evidence-based clinical decision-making rests on the assumption that the published literature being reviewed has sufficient integrity to make the exercise worthwhile.”

    “The core of our professional faith, as Spengler points out, is to first do not harm. It harms patients to have biased and corrupted research published. It harms patients to have unaccountable special interests permeate medical research. It harms patients when poor publication practices become business as usual.”

    And this from Dr. Goldacre’s recent 2012 book “Bad Pharma” when referencing the $3 billion dollar fine in the GSK settlement (where the 2001 AACAP paper was exhibit #1 of the DOJ complaint):
    “But the most dangerous tactic of all is the industry’s enduring claim that these problems are all in the past. This is deeply harmful because it repeats the insult of all the fake fixes we have seen throughout this book: and it repeats the insult of all the fake fixes we have seen throughout this book: and it is this recurring pattern of flat denial that allows the problems to persist.”

    This is what JAACAP has seen fit to publish regarding the 2001 JAACAP paper since Dr. Ryan et al’s May 2003 response:


    December 26, 2012 | 3:31 PM

    a classic illustration of the consequences of “plea bargaining” by prosecutors

    as the lawyer claims GSK is “guilty” (by a plea bargaining agreement) of a misdemeanor


    “In the United States, the federal government generally considers a crime punishable with incarceration for one year or less to be a misdemeanor.[1] All other crimes are considered felonies[citation needed]. Many states also employ this distinction[citation needed].

    A misdemeanor is considered a crime of low seriousness, and a felony one of high seriousness[citation needed]. A principal of the rationale for the degree of punishment meted out is that the punishment should fit the crime.[2][3][4] One standard for measurement is the degree to which a crime affects others or society. Measurements of the degree of seriousness of a crime have been developed.[5]”

    “…. examples of misdemeanors may include: petty theft, prostitution, public intoxication, simple assault, disorderly conduct, trespass, vandalism, reckless driving, possession of marijuana and in some jurisdictions first-time possession of certain other drugs, and other similar crimes.”

    December 26, 2012 | 3:48 PM

    “There was no significant efficacy difference between paroxetine and placebo on the two primary outcomes or six secondary outcomes in the original protocol. At least 19 additional outcomes were tested. Study 329 was positive on 4 of 27 known outcomes (15%).”
    Came from:

    December 26, 2012 | 3:53 PM
    December 26, 2012 | 4:24 PM

    Sorry, perhaps because you are retired, but, the CPT issue that will be devastating colleagues after Jan 1 2013 is a bigger issue than rehashing this Paxil matter.

    December 28, 2012 | 12:53 PM

    Someone should videotape an interview with Martin Keller about this.

    annie UK
    December 28, 2012 | 1:36 PM

    Dear Mr. Hassman, MD
    You have said ‘rehashing this Paxil matter.’
    Make no mistake, people have been murdered by Seroxat in the UK, and so rehashing this Paxil matter , is of significant importance to us, in the UK.
    Do not offend, Mickey, that he is retired. He is the most intelligent, most aware, most scrupulous, most dedicated, person, that I like to read.
    Do not offend Seroxat in the UK, and do not offend, Mickey……

    January 1, 2013 | 6:13 PM

    Joel, rehashing something that has never been put to bed shouldn’t be a flaw.

Sorry, the comment form is closed at this time.