Back in 2001, an infamous study was published in the Journal of the American Academy of Child and Adolescent Psychiatry that declared Glaxo's Paxil antidepressant - called Seroxat in the UK - was "generally well tolerated and effective for major depression in adolescents." Known as study 329, the findings were used to widely promote the drug, which became a huge seller.
Of course, the study was later held in disrepute after it was learned the results didn't tell the whole story. In fact, 329 was one of three studies cited by former New York Attorney General Eliot Spitzer, who filed a suit charging Glaxo with "repeated and persistent fraud,” alleging the drugmaker had promoted positive findings, but hadn’t publicized unfavorable data (back story).
As it turns out, study 329, which already had a sordid history that included ghostwriting charges (here's some background), were worse than imagined. A new study in the International Journal of Risk & Safety of Medicine discloses that, after sifting through some 10,000 documents that surfaced during Paxil litigation, highly selective reporting was used to skew the results favorably.
The latest study prompted a complaint concerning possible scientific misconduct to be leveled against study 329's lead author, Martin Keller, a psychiatry professor at Brown University, by David Egilman, a clinical associate professor in community health at Brown. We are awaiting a reply from Keller and Brown provost David Kertzer, who was sent the complaint by Egilman. UPDATE: A Brown spokesman declines to comment.
Back to the new study. A key finding about 329: "There was no significant difference between the paroxetine and placebo groups on any of the eight pre-specified outcome measures. However, by the time the data were analysed, many other new measures had been added to the list of secondary outcomes."
For instance, along the way, "four of the eight negative outcome measures specified in the protocol were replaced with positive ones." In fact, the study finds that in 329 "...disclosures of overdose and mania were edited out, and serious adverse events on (Paxil) were attributed to other causes." And the original authors of 329 didn't conduct any analysis to determine whether side effects occurred more commonly on Paxil compared with a placebo.
By their calculations, though, the researchers write that serious adverse events in 329 were significantly more likely to occur in patients taking Paxil (12 percent) vs. placebo (2 percent). Adverse events requiring hospitalization were 6.5 percent on Paxil vs. none on placebo. Severe nervous system side effects - 18 percent on Paxil vs. 4.6 percent on placebo.
"Our detailed case study of proprietary documents from GSK regarding this study adds to the evidence that flaws in industry-funded research can be severe and difficult to detect," the researchers write. "The documents reveal that the published conclusions of study 329 and information provided by Glaxo to health professionals understated adverse effect rates and emphasised post-hoc measures that were not consistent with the unpublished, protocol-defined primary and secondary outcomes."
[We should note that the latest study points out that the lead author was contacted by plaintiffs' attorneys to conduct a review of Paxil data as part of their litigation against Glaxo. This is noted in the study, by the way].
49 Comments
Ed, I think your link is wrong. There is no link to the study, only to a 2004 Science article about Spitzer's allegations.
Hi Nathan,
Thanks for pointing that out. Hopefully, it's fixed now. I appreciate the heads up.
ed
To All,
GSK Documents can be found here: http://www.paxilharmschildren.com
Hmmmm,... victims had to hire an Attorney because GSK,.. with blatant disregard for human life,... withheld clinical trial data, knowing full well, that it would maime and kill tens of thousands of children,.. all for the sake of profits.
Thanks for pointing out that article. It was quite sad reading it. Frankly, it's embarrassing to be affiliated (by profession) with scientists with such lack of morals. I can understand (just slightly) how the scientists could change the primary endpoints as they are looking for efficacy. As a scientist, I recognize that there is sometimes a very thin line between looking for evidence of your hypothesis and "massaging" the evidence to support your hypothesis. I can forgive them for this mistake.
The unforgivable and unexplainable mistake they made was the twisting of adverse events data. This was just plain immoral. Paxil clearly had significantly more side effects than placebo -- yet in the final draft of the paper they wrote that "only headache was considered by the treating investigator to be related to paroxitine treatment." I sure hope that the lead author of this study is no longer employed in this industry. We don't need his type here.
There was one encouraging thing I learned from this paper: the peer-review system seems to work reasonably well. The reviewers repeatedly questioned the author about lack of differentiating between primary and secondary endpoints as well as his interchangeable use of "response" and "remission". At first the author seems to just use confusing language to cover up the truth. In the end, (due to the reviews comments) the author had to outright LIE in order to get the paper accepted. From what I can see, the reviewers did their job well. You can’t expect reviewers to catch lies – only inconsistencies.
However, I don't understand why the other coauthors didn't object to the paper. I'm guessing that they never read the final version. Sometimes I'll read a couple early drafts of a paper that someone is authoring for me, but I don't take the time to re-read the final draft just prior to submission. You can bet I won't make that mistake again...
The process of grinding out multiple post hoc secondary analyses that are then cherry-picked to suit the marketing message is called HARKing -- Hypothesizing After the Results are Known. Study 329 has a secure place in the HARKing hall of fame, er, infamy.
I agree with Nathan's comment that manipulating the risk profile was even more unethical than massaging the benefit profile. Sadly, that is by now not unusual. Another recent case study is here: http://hcrenewal.blogspot.com/2008/01/antipsychotic-drugs-for-depression.html
Wouldn't you know, Keller and a bunch of his cronies show up in this case, too.
There might be indications as to how much how many knew amongst the drafts, peer reviews, letters, memos, emails, minutes, depositions, medical queries, dear health care professional letters and information for sales representatives relating to the study at
http://www.healthyskepticism.org/documents/PaxilStudy329.php
"The unforgivable and unexplainable mistake they made was the twisting of adverse events data. This was just plain immoral."
Thank you for this Nathan. Now you can begin to understand some of the anger that parents have about the paxil situation. This alteration of scientific data cost my son 4 years of his life. This is immoral, and the children are paying the price.
I've gotten past the anger, but watching my child have to live with the long term effects of short term paxil use is heartbreaking. No human being, child or adult, should have to deal with the ramifications of a drug that was marketed as "safe and non addictive", when the situation was far from the truth.
James P McCafferty/DEV/PHRD/SB_PLC (is PHRD policies and human resources development in this case?) seems to know something as it looks very much as if he is advising how to word things so that a "high" percentage (11/93, 12%) of serious AEs in the paroxetine group sounded more acceptable.
http://www.healthyskepticism.org/documents/documents/19990721McCaffertytoLaden2.pdf
Take a look also at Jon Jureidini's analyis of the citations of the 329 report on the Healthy Skepticism page, and you can see how much influence one "inaccurately" reported study can have.
The peer review and publishing process is worthy of a discussion in itself. I know that it has taken a very long time to find a journal willing to take the legal risk involved in publishing this latest paper, and you can see what sort of response the authors got from their initial criticisms here as well:
http://www.healthyskepticism.org/documents/PaxilStudy329.php
I'm not sure on this one - but is it an attempt by a number of the company's more influential people (see from and to addresses) to subtly influence physicians to prescribe paxil off label to children?
http://www.healthyskepticism.org/documents/documents/010425BattintoLaden.pdf
More people implicated given they all 'worked hard' on Study 329?
http://www.healthyskepticism.org/documents/documents/040514KellertoCarp.pdf
David J Carpenter (gsk) Ryan, Wagner, Emslie, Stober, Perara, Lipschitz (in addition to the ones in previous documents) ?
Are they the same Dr Karen Wagner and Dr Graham Emslie as appear to be associated with TMAP and TeenScreen?
You can't trust anyone these days. It sometimes seems that the more letters they have after their name, the more 'important' they believe they are, the less trustworthy they are. Or at least it seems the case when they find it cosily profitable to become closely tied with drugmakers.
NJ Residents are footing the Medicaid Bill for Children in the NJ Foster Care System who are being prescribed Paxil,.. the youngest being 3 years of age.
One should ask NJ Attorney General Ann Milgram how she sleeps at night, knowing that NJ's most vulnerable,.. Foster Care Kids will be maimed or killed, because Dr.s are haphazardly prescribing this deadly drug. And the NJ Taxpayer is being ripped off.
Children are profitable. They will soon be the next generation with perceptions that drugs are the way of life. They'll be moved, as they already are, onto equally risky drugs such as zyprexa when antidepressants like Paxil, Effexor, Prozac etc, are unfashionable.
I'd be asking: Why are children being given any drugs where adverse effects are protected under court order and why does the system allow protection of drug 'secrets' when it is not in the public interest to do so. Children have died already on antipsychotics and antidepressants.
Forgot to mention: Children have also died on stimulants prescribed for ADHD. The whole problem needs addressing. The 'children prescribed psychoactive drugs' problem.
pg,
You will find this interesting. NJ Senator Shirley Turner addressed these issues yesterday at the Budget and appropiations hearing,.. Click on listen,.. April 28 at (1:57) and also on April 23 (:20)
http://www.njleg.state.nj.us/media/archive_audio2.ASP?KEY=SBAB&Session=2008
Thanks Lisa for the links - I've got one running but I'm not sure its the right one yet, given that at my end there's two dated 28th and no time on either :) Its very late here, so will probably come back to them tomorrow.
pg,.. You want 10 am on 28 and 2pm on 23
pg,.. Scroll the top line,.. it should show you the time
Lisa thanks - I see it now. I'll do it tomorrow because its very late at night here. I think I was probably listening to the right one just now because it was talking about mental health and the 'state', but I'm too tired to concentrate and take it in. It's also an 'unknown' topic to me as its not how things work at this end, so I need to be wide awake to take the content on board :-) Nite.
Senator Turners testimoney starts just before the 2 hour mark on the April 28th 10am session.
The one positive that can come out of this new information is the recognition of the adverse effects as they truly are. Adults experience similar problems which are dismissed by medicine. It's time that the adverse reaction reports and those who have taken these drugs are listened to seriously. Being victimized again by having problem with the drug denied only makes the problem worse.
This sound's familiar, getting a DeJeVue...think I heard it from you guy's,...Lisa & Laurie
GSK should have been prosecuted for human rights abuses over this..
"Severe nervous system side effects - 18 percent on Paxil vs. 4.6 percent on placebo"
I sometimes have to refrain my anger(and my use of language) when discussing Seroxat(Paxil) and GSK..
The horror which this drug has unleashed on unsuspecting people who were already in a vulnerable state is a human rights abuse on a magnitude of a mental health genocide ...
It disgusts me ..
If someone was to suggest marketing a drug which increases suicidal thoughts to depressed people the consensus would surely be .. That's an insane notion?...
Well, that is exactly what GSK did with Paxil.. They sold a drug which they knew would harm people and make their condition worse. What is even more sinister is that for a long time they blamed the patients original condition for the side effects.. And to add insult to injury the drug is the hardest of all anti-depressants to quit... To me that is nothing short of a criminal human rights abuse...
It's great to see these figures here at pharmalot, and around the web. Eventually docs will be forced not to prescribe any new prescriptions of paxil. A horrific part of the paxil experience for my family was the fact that the adverse affects were not acknowledged, and my child received all kinds of mental health labels. GSK is as corrupt as it gets, and I love that this information is being circulated.
It's refreshing to read someone in pharma profession acknowledge the wrong doing Nathan. I realize GSK actions do not represent all in the industry.
"I realize GSK actions do not represent all in the industry."
I think this is a key point. Those of us who have had family suffer with the use of this drug are portrayed as antipharma. I'm not. But as Nathan recognized in this study, our anger is well placed in this one situation. Hopefully this will enlighten those who prescribe Paxil and prevent the horror that our children went through from happening to others. When I read the complete study the misrepresentation of the side effects of "hostility, worsening depression emotional lability" turned into "headache" it makes me ill. These exact side effects are exactly what we had to fight to get recognized and warned against, yet they were documented the whole time. The "severe neurological symptoms" I have witnessed in my son. Pseudoparkinsonism, violent twitching, loss of autonomic control of blood pressure, altered gag reflex...all symptoms that he has had for the last 3 years. The physical and mental toll this has taken on him has been life altering. It's only in the last few months that he is starting to see resolution of these symptoms. Yes, I'm venting....but those who call us extremists, crazy and "scientologists" need to understand what we have lived through and now we have the scientific paper that we can carry with us, since the first hand reports have been dismissed as "anecdotal". The authors of this study have my eternal gratitude.
Ok, so here's a question that I'm sure you guys have thought about: Who is responsible for this failure?
Is it the lead investigator? Is it GSK? Is it the coauthors on the study? Is it the marketing people who promoted the study?
It seems to me that the vast majority of the responsibility rests on the lead author. But the other parties do bare some blame as well. What is he guilty of? (in a courtroom – not in your mind) Certainly of scientific fraud. Anything else?
I certainly don’t agree with Truthman that GSK should be prosecuted for "crimes against humanity". It seems to me that the lead author pulled the wool over everyone's eyes -- even those at his own company. A company probably can be held financially responsible for the injuries -- but I doubt if they can be held legally responsible.
'A company probably can be held financially responsible for the injuries — but I doubt if they can be held legally responsible.'
Nathan, the fact that GSK has had this information for years now, and is still promoting this medication makes them legally responsible, imo.
Reading what you have, do you think paxil should still be prescribed to children?
A recent scenerio: A brilliant child, one that was on the honour role at school was stressed out with exams, sports etc. Dad sends her to the doctor for a physical to ensure all was well, she looked like hell. The doc tells this child that paxil will relieve the anxiety and get her through exams. The child is 16 so the parents are not told about the medication. The child takes the paxil, and dad finds her hanging from the basement rafters. Who's responsible? Does it matter? The fact is, yet another child dies even after this information has been exposed.
What I don't 'get', is why with all this information is GSK still promoting paxil, and why doesn't anyone give a damn and stop it...
"Is it the lead investigator? Is it GSK? Is it the coauthors on the study? Is it the marketing people who promoted the study?"
Good question. I don't believe it's a simple answer. There's a long line of responsibility. Authors and investigators are hired by the company paying for the clinical trial, so in this case I think the responsibility applies to all involved. If you look at the position piece on this trial, the "Target" statement pretty much lays out GSK's attitude on the priority with Paxil and children.
http://www.ahrp.org/risks/SSRI0204/GSKpaxil/pg5.html
If you read the last page of this position paper, you can see where the focus is on how to publish the "positive" data, not on publishing the full study.
I think if I had seen GSK come public about the risks they had found and emphatically warn doctors to NOT prescribe this to children I might feel different about their involvement. But instead we see pharma sponsored groups fighting to keep this drug on the market, GSK reps visiting politicians when there is any hint that the public may get the full story and a total distain for those who advocate for the truth on Paxil. My son's Paxil journey started with samples(alot of samples) from a child and adolescent psychiatrist. Samples don't get to an office like that by accident. So, yes, marketing bears some of the blame too.
"A company probably can be held financially responsible for the injuries — but I doubt if they can be held legally responsible"
GSK fights it's financial responsibility with preemption and the knowledge that few attorneys have the resources to go up against them. This is what angers me the most. They had the information on the risks yet deny those same effects when it happens to a real person in real world use. How is that an honest practice?
Quick comment with the proviso that I haven't read all the relevant documents, etc.. So just answering in general terms.
The company as a company can be held accountable (in criminal of civil proceedings) if suppressed/skewed study can be shown to have been part of an organized plan to misrepresent risks/benefits of Paxil. Again, the Neurontin case comes to mind.
The finer points come down to who are the intended recipients of the misrepresentation - FDA, docs, public, Medicare, etc.. - and to what end.
Nathan : I certainly don’t agree with Truthman that GSK should be prosecuted for “crimes against humanity
Well if not a "crime againts humanity" Nathan, then what ?
"Corporate manslaughter?"? "Grievous bodily harm?" "Scientific fraud resulting in death of consumers?" "Death by misadventures from greed?" "Manslaughter for profit"?
Take your pick.. They all stink.. And they are all human rights abuses..
Nathan" It seems to me that the lead author pulled the wool over everyone’s eyes — even those at his own company. A company probably can be held financially responsible for the injuries — but I doubt if they can be held legally responsible"
Thats entirely speculative Nathan..
What is a pharmaceutical company? It's an organiisation made up of people. And those people (particularly at the top) make the life or death decisions. To suppress or not to suppress negative data? Is that the question? Or is the question.. who from the company makes the final decisions? And would that( or those) individuals not be the ones who hold the ultimate responsibilty and absoulte accountability?
Let me put it this way: Have you ever heard of a corporation being charged with murder or manslaughter? I certainly haven't. I think those kind of crimes are reserved for individuals. Even if the corporation is charged and convicted, so what? Who will you lock up? All the employees? The managers? Which managers? The current ones or the ones in place at the time of the coverup? What about the people who SHOULD have known about the coverup but didn't bother doing thier homework? What about the coauthors who didn't bother reading the final manuscript? Were any of these people aware of the implications of the decisions they made? Hindsight is always 20/20.
I agree something terribly wrong happened. I'm just not sure that the solution is very clear.
Here's a question for Jack2 or someone else with more clinical knowledge than me: Could this happen in today's environment? It appears that the reviewers did not have access to the raw data of the clinical trial. Aren't there laws today that require clinical data to be deposited upon completion? Therefore, wouldn't the reviewers be able to cross check the paper with the raw data to make sure this kind of thing doesn't happen again?
Hi Nathan - A partial answer. I still haven't read all the relevant Paxil stuff. But the below from the Depatment of Justice's summary of the Neurontin prosecutions that Ed has linked a number of times.
"“The Department of Justice is committed to rooting out and prosecuting health care fraud,” said Associate Attorney General Robert McCallum. “It is of paramount importance that the Department use every legal tool at its disposal to assure the health and safety of the consumers of America’s health care system, and to pursue companies and individuals that steal from the taxpayers and inflict suffering on patients and families."
Note the reference to "companies and individuals that steal from the taxpayers and inflict suffering on patients and families."
The latter, particularly in the context of the rest of the report, demonstrates that companies as companies - and not only specific individuals - can be held liable for "inflicting suffering."
As far as what you do, the usual is monetary fines and, sometimes, jail terms for managers who had central responsbility. Clearly, and for several reasons, it would not be in the public interest to lock up Warner Lambert!
Keep in mind that the DOJ has assumed that accountability would also be enforced by injured plaintiff's or their survivors' civil suits.
As many times discussed, with preemption, that disappears.
Hi Nathan,
1. I agree that the co-authors definetly should have had access to the raw data, and (if they bothered) could have double-checked it. Some areas for me though I wouldn't be able to double-check. While I've got basic college stats, I can't double-check my stats groups work. They understand the stats concepts and stats programs much better than I do. However, while I don't know my ANOVAs from my ANCOVAs, I can count up an event (eg. suicides). Would everyone on a paper bother to double-check it in practice? I doubt it.
2. As for reviewers it depends on the journal. For example Annals of Internal Medicince required my company to submit the raw data recently for their stats analysis. I think the higher the journal's impact factor the more likely the raw data will be required. I don't know if this requirement only applies to industry funded studies, but I think it must apply to all clinical trials. I also don't know if the raw data is reviewed only by stats people who work for the journal or also the external reviewers.
3. Previously, all top medical journals would only publish studies posted on clinicaltrials.gov. This rule meant the VAST majority of trials went on the website since (1) people only run trials they think can work and (2) people want those trials in the most widely read journals possible. Now, by FDA rules all trials (beyond basic PK studies) must be posted to clinicaltrials.gov - which doesn't really change practices (since it was done anyway). All this prevents a company from sweeping under the rug a negative trial.
Now (as of 2008) all data must be published for a given trial after a certain amount of time - if you can't get a journal to publish it you must at least post results to clinicaltrials.gov (it's about 1 year but it depends on the seriousness of the disease studied). The requirements for what you actually need to disclose are not clearly defined. I'm not sure everyone will post enough details so other people can perform rigorous meta-analyses.
Can it happen today?
Well, you need to put the primary endpoint on clinicaltrials.gov, so you won't be able to change it and not have people find out about it.
Could people hide important safety data? If the journal makes you submit all data you couldn't do it without out-and-out lying by forging docs. That certainly makes it harder. It also pushes people from something they might be able to ethically rationalize in their own head (omission) to something they can't ethically rationalize in their own head (falsification).
I have no idea, but I would say most scientific misconduct consists of selective interpertation or omission. Very little of it is lying or fabrication. So I think it would be harder today.
Nathan Let me put it this way: Have you ever heard of a corporation being charged with murder or manslaughter? I certainly haven’t. I think those kind of crimes are reserved for individuals.
Have you ever heard of the phrase "there's a first time for everything"?
Corporations are made up of people Nathan.. And these people commit the crimes which result in death sometimes.. They are the corporate face , and they should be prosecuted..
Jack2 - It is my understanding that posting on clinicaltrials.gov is a voluntary effort outside of federally funded agencies (eg. NIH, etc.), and is not mandatory for pharma companies, although some pharma companies have made it their stated policy to post all their trials. (Drugs studied for efficacy under an IND that are for the treatment of Serious or Life-Threatening Diseases or conditions are also required to be posted. Post-marketing studies though would not fall in this category)
I believe that (at least some) top journals do make posting on clinicaltrials.gov a requirement for publication.
Access to 'Raw data' would still not help you with coding issues.
Jack2,
I actually disagree w/Janet.. Access to Raw Data,.. "WOULD HELP".. with coding issues!!!!!!!!!!!!!!!!!!!!!!
Janet, "Care to challenge me"?,,,,
Maybe a colonoscopy is needed here,.. to get rid of the bullshit!
[...] For the background, let me take the easy way out, and quote Ed Silverman from PharmaLot: [...]
Lisa - I think we are using different terminology. When I referred to 'raw data', it was after coding was already done and the data was in a database.
Jack2 - thanks for your comments. It sounds like the system has improved a lot from the days of the "329 Study" and hopefully something like that won't happen again.
I went to th lead author's website yesterday -- he's still a professor at Brown and sits on countless advisory boards. It's ashame that such blatent abuse of his position didn't cost him his career.
Posting something on clinicaltrials.gov before the trial and some results after the trial will become FDA required during 2008. For serious/life threatening diseases it phases in earlier (I think this spring, so maybe already), and for other diseases it happens later this year. My understanding is: 1. this only affects new trials 2. the 1 yr deadline for posting results of the trial runs from last patient out. 3. this will actually change practice minimally, since almost everyone already did it, because they wanted to publish in a top journal. My company already put its trials on the website. The journals really set the standard because they already insisted the sponser (industry or otherwise) post the trial at inception on the website. 4. when posting results, the guidelines, the last time I saw them, were vague. I seriously doubt you will see the whole database on the website. I don't think you will see as much as you would if the trial was published as a 5-10 page summary in a journal, let alone a 3,000 page CSR. I think it will just be a few paragraphs - but I confess I don't really know. 5. Some journals will view posting on clinicaltrials.gov as publication of the data, which will disqualify you from submitting to that journal. Sometimes a manufacturer doesn't publish a study because they overshoot about which journal will accept the study, and the review/revision/rejection process can take over a year. Maybe they submit to JAMA when they only had a small shot of getting into JAMA - but months go by as they get reviewed, revised and rejected. Then they miss the deadline. This is different from the Enhance trial delays, and would affect the myriad of important, but lower profile trials. 6. Small, early trials are exempt (example: a phase I PK study), since they could require companies to release secret information (example: the molecular structure).
Are the "days of the 329 study" over yet? I don't think so.
Peer reviewers may be experts in their field, but if they are not quite up to the job of interpreting complicated statistics, what is to stop sponsors publishing unjustifiably positive spin they have massaged out of the numbers, even if those numbers are available? How many clinicians actually check for results on clinicaltrials.gov or on companies websites? How many can understand the stats themselves?
Once the genie is out of the bottle, it's very hard to get it back in. According to Jureidini, in 2008 "study 329" is still being reported unequivocally in some papers as positive.
It's hard to know how to deal with papers like '329. Someone relatively new in the field who is looking for some background references would stumble across the '329 study and assume it is positive. How would they know otherwise? Maybe in cases like this, there should be some sort of perminent "addendum" to the paper to warn people of the misleading information that it contains. This is a problem with the entire scientific literature system - not just this paper.
Nathan, the fact that Paxil didn't recieved FDA approval for Paxil for children should throw up the big red flag.
The current prescribing of paxil for children is being done on a "word of mouth" by psychiatrists, which always begs the question, why are ancedotal positive effects used to prescribe, yet negative anecdotal effects dismissed?
In 329 we have documentation of those negative effects(even with very short use time), yet we battle every day for clear warnings for parents to make informed decisions. GSK fought hard with the FDA to water down the black box warning and succeeded in getting "causal role" removed...well, here is a study that confirms that causal role. How is a parent supposed to know the truth if the label is so misleading?
I'd still like to know who McCafferty and Oakes reported to, within SKB/GSK.
Matt