London, UK - Dr Steven Haffner (University of Texas Health Science Center at San Antonio), a peer reviewer for the New England Journal of Medicine of the original rosiglitazone (Avandia, GlaxoSmithKline [GSK]) meta-analysis by Dr Steven Nissen (Cleveland Clinic, OH) et al, is reported to have leaked the paper to GSK weeks ahead of publication [1].
Nissen's meta-analysis, suggesting an increased risk of MI with rosiglitazone, was published on the New England Journal of Medicine website on May 21, 2007. But according to a statement released today by Sen Charles Grassley (R-IA), who launched an inquiry into rosiglitazone last year, Haffner confirmed to congressional investigators that he had faxed a draft of the study to GSK weeks before it was published [2].
The revelation was first published by Nature in a news article in its January 31 issue, which reports that 17 days before it was published, Haffner faxed his copy of the paper to Alexander Cobitz, a GSK employee whom Haffner knew from working on an earlier clinical trial of the drug. "Why I sent it is a mystery," Haffner told Nature. "I don't really understand it. I wasn't feeling well. It was bad judgment."
In the Nature article, Haffner says he has given many talks for the company but declined to say how much he had earned from them. "I've got a considerable amount of money. I didn't do it to raise my income or anything like that," he told the journal. In a letter to GSK [3], Grassley states: "According to documents filed at the FDA, GSK has paid Dr Haffner around $75 000 in consulting fees and speaking honoraria since 1999."
Nissen's meta-analysis, suggesting an increased risk of MI with rosiglitazone, caused a major furor, with huge media attention, resulting in a large fall in GSK's share price and new warnings issued by the FDA and other regulatory authorities about the drug. The receipt of the paper before publication would have given GSK valuable time to put together a response to the data and contributed to the company's decision to unblind the ongoing RECORD clinical trial of the drug.
GSK: "We acted responsibly"
GSK spokesperson Nancy Pekarek confirmed to heartwire that the company had received a copy of Nissen's paper from Haffner on May 3, 2007. She said Haffner, who consulted for GSK on Avandia, had concerns and questions regarding the methodology of the analysis and sent the article to GSK for advice from company statisticians. She added that GSK did not solicit the paper and did not respond to Haffner's request for statistical advice and that no money changed hands in relation to the paper. "We believe GSK acted appropriately and responsibly in responding to the situation," she said. Pekarek added that GSK had already conducted its own meta-analysis, which had raised similar concerns about rosiglitazone, and was in discussions with the FDA as to whether to unblind the ongoing RECORD clinical trial to find out more information about cardiovascular risk with the drug and that knowledge of the Nissen meta-analysis "was an additional factor that led us to unblind that trial." The RECORD trial was unblinded in mid-May, before Nissen's meta-analysis was published, but the data were said to be "insufficient" to determine whether rosiglitazone increased the risk of MI.
Congress investigates
In his statement, Grassley notes that contrary to some press reports last year, Grassley himself did not receive a copy of the Nissen study before publication, but "you can imagine my surprise when I learned that one of GlaxoSmithKline's own consultants leaked a copy of the study to GlaxoSmithKline.
"Not only did Dr Haffner breach his agreement with the New England Journal of Medicine to properly peer-review the Nissen study, but he violated practically every tenet of independence and integrity held sacred by the major medical journals," Grassley said. He added that his letter to GSK asks what action it took after it received the leaked manuscript.
"It is only good-quality science that separates modern pharmaceuticals from old-fashioned snake oil. Over the past few years, my investigations have found that the Food and Drug Administration has a very cozy relationship with drug companies. I have also discovered that drug companies spend big bucks to influence which drugs doctors prescribe. Finally, I have shown that some drug companies intimidate scientists who speak up about bad drugs. Now it appears that even peer-reviewed science is not completely without its own problems."
Nissen "hurt"
Nissen commented to heartwire that Haffner's leak of the paper was "very troubling. . . . The scientific review process is sacrosanct. It is an honor and a responsibility to hold a manuscript in confidence. To provide it to a third party for commercial reasons is a fundamental breach of integrity." Nissen said he was particularly hurt by Haffner's actions given Haffner's "attack" on Nissen in the media after the manuscript was published. At that time, Haffner criticized Nissen and the New England Journal of Medicine on theheart.org about the way in which they published the article. "How he could attack me in such a way for conducting this analysis when he had broken his confidentiality agreement with the New England Journal of Medicine and sent the article to GSK I find particularly shocking," Nissen told heartwire.
The New England Journal of Medicine released a statement saying: "We consider the peer-review process to be confidential. Any breach of ethics by a reviewer would be taken very seriously by the editors but would be handled as a private matter."
Haffner was unavailable for comment.
 |
 |
 |
 |
|
 |
 |
 |
 |
 |
|
|
January 30, 2008 03:56 (EST)
|
|
 |
 |
 |
 |
 |
|
No Mystery
"Why I sent it is a mystery," Haffner told Nature. No mystery here - Money Makes the Mare GO ! |
|
 |
|
|
|
January 30, 2008 07:01 (EST)
|
|
|
|
|
|
|
|
It's all politics
unfortunate scenario all around. i don't think there has ever been any love lost between the 2 steves. it would be interesting to see how much dr. n has been paid by the makers of pioglit since 1999. they paid him 25K just to publish their own metanalysis which he had published in another 'peer' reviewed jrnl. also would be interesting how much he is paid in clinical research trials, like prasugrel where he said it has a place even though it proved to reduce MI, it actually DID kill people NNK 300 - yet he is paid quite well in clinical trials for the makers of that drug and he supports this approval. it's all fishy stuff |
|
|
|
|
|
January 30, 2008 07:38 (EST)
|
|
|
|
|
|
|
|
re:
"I don't really understand it. I wasn't feeling well. It was bad judgment"
Not feeling well? Would expect an excuse like that from Roger Clemens, not an academic researcher. |
|
|
|
|
|
January 31, 2008 08:36 (EST)
|
|
|
|
|
|
|
|
Forest for the trees!
How charming that we might excuse or allow for this sort of academic immorality by suggesting that the author of the reviewed manuscript also has financial ties to Big Pharma. I recognize that I may sound naïve here, but I consider Dr. Haffner’s action deplorable. I’m certain he is a decent man, and I know he's a brilliant epidemiologist and scientist, but I cannot see how he could ever be invited to review another manuscript for a top-tier academic journal. |
|
|
|
|
|
January 31, 2008 03:46 (EST)
|
|
|
|
|
|
|
|
who hasn't made terrible mistakes?
His motivations should be understood. He should apologize and we should all move on. Who here hasn't made a terrible mistake and paid for it? I don't think he should pay for it with the rest of his career. He did not commit fraud, fabrication of data, or deceit; cardinal academic sins. Forget it and move on. |
|
|
|
|
|
January 31, 2008 05:03 (EST)
|
|
|
|
|
|
|
|
Cardinal Sins
While we are talking about Cardinal sins, why don't we include breach of fiduciary duty, ethics, and confindentiality and scientific intergrity ? I'd hope that breaking these "duties" would be considered a academic Cardinal Sin.
I believe in forgiveness, but physicians, esp. leaders in the field are expected to be held to a higher standard. Let's first admit how egregious an offense this was, in the realm of science and , this is in terms of minimizing publishing bias and maintaining the quality of literature, then we can ....move on.
|
|
|
|
|
|
January 31, 2008 06:03 (EST)
|
|
|
|
|
|
|
|
holier than thou
Yes he breeched the integrity of the peer review process, and it is doubtful that he will ever be asked to review at N Engl J Med again. But let's not confuse fraud, fabrication of data, lying/deceit with what occurred in this case. I doubt he ever signed his name to a confidentiality agreement with the journal; few of us have (unless one is an editor). What I'm most concerned is that this was sent to GSK, but I believe they knew anyway because S.Nissen had contacted them for data for his meta-analysis before hand! Thus what are the ramifications of his action - did anyone die as a result? Did the paper fail to get published? No to each of these. Yes it was egregious, but no it was not a cardinal sin (you have to wonder about his judgement though...).
I think we have all had days where we regretted something that we did. |
|
|
|
|
|
January 31, 2008 07:55 (EST)
|
|
|
|
|
|
|
|
mea culpa
Your points are sound, as usual; my criticism needed to be more tempered. Regrettably one has come to expect a certain amount of contrivance for financial gain within industry, but it remains a bitter pill when an academic thought-leader engages in similar behavior. Surely somethings shouldn't be for sale, for instance the NEJM manuscript review process. |
|
|
|
|
|
January 31, 2008 08:14 (EST)
|
|
|
|
|
|
|
|
for sale
I don't think anything was sold or purchased.
No one is more distraught or disgusted with this than Dr. H himself, i'm sure.
I don't know if he was thinking that perhaps gsk would have insight or information that would help with his review.
Dr. N has been battering rosi for nearly 12 months prior to this article and perhaps Dr. H thought it was a biased presentation.
I'm sure in retrospect he wishes he had simply asked gsk for all of their analysis and he had a statistician go over it instead of faxing Dr. N's analysis to gsk.
Poor judgement in a review process, no one hurt, no action taken - his reputation is damaged for a degree of integrity. |
|
|
|
|
|
January 31, 2008 08:41 (EST)
|
|
|
|
|
|
|
|
interesting
'Regrettably one has come to expect a certain amount of contrivance for financial gain, but it remains a bitter pill when an academic thought-leader engages in similar behavior.'
I can't think of a single opinion leader who doesn't have at least some sort of tie to industry. Unfortunately, when you check out the financial disclosures at the end of each guideline update, one can see that we are all easily bought and sold like cattle in the marketplace. Now that I think about it, doesn't that mean we should take these 'unbiased' treatment guidelines with a huge chunk of salt?
I think S.Nissen gives all of his honoraria to charity though... |
|
|
|
|
|
February 1, 2008 09:13 (EST)
|
|
|
|
|
|
|
|
confused now
Isn't this is a bit beyond than the usual tie. 'His motivations should be understood.' If not financial, then what? |
|
|
|
|
|
February 1, 2008 12:51 (EST)
|
|
|
|
|
|
|
|
The thought leaders are never impartial
I agree with AM here- I am sure Dr H is a very nice man but even if one blames the subconsious , his loyalty has been bought- why else did he do it-This is a real problem in academia and all the thought leaders have a conflict- even if the money is donated away- Dr N' s entire career has been built on industry sponsored research and he obviously will be conflicted one way ( towards those who support his research) or another (his obvious dislike for Merck/Schering and GSK)
I find that the much maligned private practise guys are usually more balanced when it comes to industry
|
|
|
|
|
|
February 1, 2008 02:58 (EST)
|
|
|
|
|
|
|
|
completely agree
I meet with the industry guys all the time, but so far have yet to take (or ask) for any money for research or otherwise. Unfortunately almost all good randomized trials are supported by industry and that is just the price we pay. On top of that, panellists for all the treatment guidelines, which favour drugs or devices, often seem to have ties to industry. My advice: keep your head down, stay away from drug talks, and read the literature yourself. Having some decent epi training goes along way in terms of critiquing the evidence, much of it is poor. |
|
|
|
|
|
February 1, 2008 04:11 (EST)
|
|
|
|
|
|
|
|
Keep head down and dr. n charity
Yes Dan he gives his $$ to charity - his own chartibable foundation - no conflict there :o) haha
I do think that fair balanced fda approved educational lectures have value. i always learn something and if you do read and have balance i would hope that most clinicians can make unbiased decisions. there are bad eggs in all industries.
pharma has certainly been beneficial in discovery, mgd care is a different story - the jury is still out whether mgd care has improved care/outcomes. mc |
|
|
|
|
|
February 1, 2008 10:56 (EST)
|
|
|
|
|
|
|
|
Voilation of Ethics
It is difficult to understand some of the soul searching going on here. It is pretty obvious that Steven Haffner seriously violated ethics. The peer review process is confidential and he violated it because of his allegiance to GSK from whom he had obviously a cozy relationship and had gotten lot of money. Then he comes out bashing Dr. Nissen which was likely in great part motivated by the above and then his possibly his feelings for Dr. Nissen. So for him to say "Why I sent it is a mystery," "I don't really understand it. I wasn't feeling well” is laughable and does not fool even a kindergartner. I don’t recall when I last heard such a laughable excuse. While is true that most researchers cannot avoid Drug company influences, it does not justify these kind of abuses, nor give any justification to do and publish corrupted research that all too common these days. There is something also called as morals, scruples, ethics and integrity. If these are so easily discarded when presented with the big buck and the opportunity, why is we are thought to have a higher ethical code than a commoner when we abandon our code in gay abandon and compromise the very soul of our profession from within ?
This episode with Haffner reflects far more than just small infarction, but reflects something far grater. It reflects the depths of the corruption of research integrity by the Pharmaceutical industry We all know that writers who write manuscripts are far more often than not biased towards their publications and the drug companies for whom they have done research for and got paid from. The peer-review is literally the last wall between that research and the common physician who very often does not have the time or the competence to analyze all the data in detail for every study that comes across his table, and often trusts the peer-reviews to give him the balanced view of the publication. If the Pharma can make even a the top peer-reviewers easily compromise their ethics by their cozy relationship with him, there is literally no check left for compromised and corrupted data from making its way to the common physicians and consequently medicines of questionable benefits and even dangerous consequences from making into patient’s mouth |
|
|
|
|
|
February 2, 2008 09:29 (EST)
|
|
|
|
|
|
|
|
peer review is a failure
In what other domain in life are we judged by people who know us but who do not reveal themselves to us? An open, transparent, and fully accountable peer review system would have prevented this problem. Had Dr Nissen, his colleague Dr Wolski, the editors at N Engl J Med, and the (likely) numerous peer reviewers on this manuscript all known each other, behaviour such as this would likely not have occurred.
Peer review is a dismal failure. Authors routinely recommend their friends as referees, suggestions for reviewers which are taken up by harried editors. A study in JAMA of the peer review process showed that manuscripts with suggested peer reviewers often received more favourable review.
The British Medical Journal is one of the few mainstream journals (other than the open access journals such as BMC) that allows open transparent peer review. The authors know who they are getting. The reviewers know who they are reviewing. I can't think of another domain in life where the judges are blinded to the defendants - the court system? the olympics? PhD dissertations? elections to congress?
It's the nonsensical aspects of the peer review system that made this happen. Also, the fact that science should be somehow kept under 'wraps' until published, and not shared or disseminated to others, is ridiculous, and does not take into account the fact of abstract presentations at congresses and routine inter-investigation dissemination by email, telephone, and in person. |
|
|
|
|
|
February 7, 2008 01:04 (EST)
|
|
|
|
|
|
|
|
Personal issues???
i cannot help to consider the possibility of personal dislike between the two protagonists |
|
|
|
|
|
February 10, 2008 05:29 (EST)
|
|
|
|
|
|
|
|
ACCORD
Considering the ACCORD findings of worse outcomes with intensive glycemic control, has anyone considered that the issue with rosiglitazone is the increased morbidity of excellent glycemic control?
Perhaps the problem with rosiglitazone is that it works too well. I would suggest that Nissen re-evaluate the findings and correct for glycemic control. Perhaps rosiglitazone is off the hook.
|
|
|
|
|
|
February 10, 2008 05:47 (EST)
|
|
|
|
|
|
|
|
weird rationale
Actually I don't understand that rationale. In PROACTIVE, pioglitazone reduced the need for insulin therapy by 50% and yet ended up preventing cardiovascular events. Whereas rosiglitazone increases them. Clearly there are inter-molecular differences at play across species in the TZD drug class. |
|
|
|
|
|
February 10, 2008 09:40 (EST)
|
|
|
|
|
|
|
|
Rosi vs Actos study, n=2
Just random observation that I have had a couple of patients in whom I stopped roziglitazone after the news broke. Unfortuantely, I was unable to control their blood sugar on Actos as well as it was controlled on rosi. |
|
|
|
|
|
February 11, 2008 02:49 (EST)
|
|
|
|
|
|
|
|
rosi-pio rehash
the accord study could not lay blame on any one therapy nor was rosi the culprit. higher mort was seen in all treatments in the intensive group. they were also being 'flogged' with aggressive therapy. we will have to wait for the publication and draw more conclusions then but there are not enough events i think even then for a conclusion to be made.
again rosi has not been implicated in death - adopt, dream and record all nullified thad individually and together.
we can speculate that su's are not the best for post mi pts with DM (the evidence supports this), that acarbose isn't good for people with chronic gi probs, the metformin isn't good with stage 4 ckd, perhaps stage 3 ckd higher doses or if pt gets contrast study, that tzd's aren't good in uncorrected anemia, current chf, etc...
I think the key is a pt over 60 with dm and hx cad perhaps shouldn't be pummeled with intense glucose control unless all comorbid features are addressed and i would be very careful with high dose su tx.
PROactive with pio showed positive secondary endpoint reduction of death, mi, stroke RRR 16% with ARR over 2% over 3 yr study in also a high risk group. Fortunately rosi wasn't the culprit in accord. |
|
|
|
|
|
February 11, 2008 10:03 (EST)
|
|
|
|
|
|
|
|
insulin to blame?
The data from ACCORD will raise eyebrows, more questions, but very few answers unless poor interpretation of subset of data are overemphasized.
Hopefully, the lay press will not blow this out of proportion, although I doubt it.
The reps for TZDs are already saying that they were "cleared of blame". Is there anyone not spinning these data?
Remember to be cautious of any data blaming or exonerating any DM drugs from this study since it was not powered nor designed to look at this information. Remember that Capricorns and Libra both benefit from ASA/Streptokinase in AMI despite what the subgroup analysis from ISIS-2 suggested.
Daniel |
|
|
|
|
|
February 11, 2008 12:01 (EST)
|
|
|
|
|
|
|
|
or, focus intensity on what works:
-BP
-RAS
-microalbuminuria regression
-LDL
-HDL
-smoking
-diet
-activity
-weight
These are the things that lead to the deaths of type 2's. The decidedly mixed picture on glucose-lowering and events in Type II DM is not without reason (or precedent, going back all the way to the UDP in the 70s). |
|
|
|
|
|
February 12, 2008 09:27 (EST)
|
|
|
|
|
|
|
|
ACCORD
or we could focus on how amazing it was that this study anticipated 5% mortality/year and we saw less than 1.5% mortality/year in each group of high risk people with DM for over 10 years with significant athero.
|
|
|
|
|
|
February 12, 2008 12:12 (EST)
|
|
|
|
|
|
|
|
that is why NNT and ARR taken from clinical trials is misleading
A point I've been trying to make all along...patients in clinical trials do not resemble the mixed higher risk bag we see in the clinic or on the ward or in the unit. That is why when people quote these astronomically small absolute risk reductions or astronomically high NNTs, I always say, wait a minute, isn't it the RRR you should be quoting, and apply this to the patient (higher risk) in front of you? |
|
|
|
|
|
|
Blinklist
delicious
Digg
Facebook
Furl
Google
LinkedIn
ma.gnolia
Mixx
Reddit
Stumbleupon
Twitter
Y! Bookmarks
Yahoo Buzz
