How Sanofi-Aventis Won Approval For A Heart Drug
41 CommentsBy Ed Silverman // May 30th, 2010 // 11:22 am
Richard Page thinks Multaq is an excellent new drug for treating atrial fibrillation, and the chairman of the department of medicine at the University of Wisconsin School of Medicine and Public Health, last year authored a large study in The New England Journal of Medicine that led to FDA approval. But his involvement allowed Sanofi-Aventis to dictate the terms, according to The Milwaukee Journal-Sentinel.
Although he vouched for the accuracy and completeness, he never saw the raw data. The drugmaker, which paid for the study, collected that info and performed the analysis without an external audit. “These companies, if they were falsifying data, wouldn’t be kept in business if that were found out,” Page tells the paper. “I was satisfied and remain satisfied that the study was conducted in an appropriate way…the FDA examined the original data and was satisfied with the conduct of the trial. Any suggestion that the sponsor had any influence on the conduct of the trial interpretation is inaccurate…We knew how it was analyzed. We just didn’t do the analysis ourselves, and we didn’t go out and hire other statisticians to do that. This was a trial sponsored by Sanofi-Aventis and they paid for the statisticians. There is a sense of trust that they won’t falsify data.”
As it turns out, Page and the other co-authors had financial ties to Sanofi-Aventis at the time of the study. Two authors worked for the drugmaker and owned stock, while Page and four others were consultants or speakers (see the disclosures listed at the end of the study here). A spokeswoman for Sanofi-Aventis, which was recently tagged for clandestine Multaq recommendations on some Internet sites (see here), says the drugmaker collected and managed the data but that a steering committee and a data monitoring committee had access to analyzed data, but not the raw data. Neither Sanofi-Aventis nor Page has been accused of unethical conduct, the paper notes. But as the paper notes, Multaq was linked to problems before its approval.
A separate 2003 trial, which was published in 2008, was stopped because more Multaq patients were dying than those on a placebo. Nonetheless, the NEJM study described Multaq as having significantly reduced cardiovascular deaths, and FDA advisory panel members criticized that claim a month later (here’s the transcript). “This is a drug that killed a lot of people in this first trial, and so there is a huge amount of concern,” said panel member Sidney Wolfe of Public Citizen tells the paper, adding that there is a lot of off-label use.
The NEJM does not have a policy forbidding such studies and NEJM editor Greg Curfman notes that clinical trial data belongs to drugmakers. “We are living in the real world here. They own it,” he tells the paper. “It’s their data.” He defended the decision to publish the study and called it an exploratory finding, although that word never appears in the article. Before such a claim can be made, another clinical trial must be done, and Curfman maintained that doctors understand what the finding means.
The NEJM study concluded Multaq reduced cardiovascular deaths, but the benefit was not significant under the original study design, according to Sanjay Kaul, an FDA panel member who is also director of the vascular physiology and thrombosis research lab at Cedars-Sinai Heart Institute. Calculating the cardiovascular death rate was a secondary measure. The primary endpoint combined hospitalization for a cardiovascular reason or death, which yielded a 24 percent benefit with Multaq. The FDA approved the drug on that basis.
But as Kaul tells the paper, the study was supposed to stop at 4,300 patients, although at that point, there was no significant benefit in reducing cardiovascular deaths. However. the study was extended beyond the planned cutoff date and 328 more patients were enrolled - and there were now five deaths among those in the placebo group and one in the Multaq group. Sanofi and the authors decided that was enough to report a significant cardiovascular death benefit.
patrons99
Great post, Ed. The pattern continues. It’s very clear that pharma shills (M.D.s with financial ties to Sanofi-Aventis) are involved. This is not the first time that a prominent medical journal has published “research” of this type. Dare we trust its conclusions? Are there serious questions here as to COI and bias. Was ghostwriting involved?
I hope that some postmarketing outcome studies were required for approval. Why would a cardiologist prescribe it without knowing the cardiac death benefit? Now that it’s “FDA-approved”, it will now require a voluntary recall Sanofi-Aventis to ever get it off the market. How likely is that?
Brad
I believe the Vioxx paper was ghostwritten by Merck. What about the Sanofi-Aventis Multaq paper?
It just seems wrong for the journal to contact the drug company if they have questions about the study. Perhaps the way to do it is to have the drug company listed as sole author.
patrons99
I’m sure this NEJM “research” publication, and others, will be widely disseminated by Sanofi-Aventis sales reps to “learned intermediaries”. Off-label promotions likely began the minute FDA approval was granted. It won’t take long to recoup their 1 billion or so in development costs. Let’s hope there’s not too much underreporting of cardiac deaths and that someone at FDA is monitoring them, faithfully.
Paul
The FDA did the analysis. They are not paid consultants.
ex-FDAer
Actually FDA drug review is 100% paid for by pharma. Or as Janet Woodcock calls them, Clients.
keiner
@ex-FDAer
So the FDA is actually for free for tax payers? LOL!
patrons99
Generally speaking, in a study or studies wherein the investigators, with financial ties to the corporate sponsor, deems an SAE, e.g. death, to be “not drug-related”, isn’t that cause for some serious doubt, because of the COI?
patrons99
@ ex-FDAer - How did Pharma become “clients” of FDA? I thought Congress originally created the FDA and enacted the FD&CA to protect and serve the public. What ever happened? I think I know the answer to this one, but I’d certainly like to hear your view point.
patrons99
@ Brad - “Perhaps the way to do it is to have the drug company listed as sole author.”
You might be on to something here. Who would do the peer-review? Wouldn’t off-label drug promotion take on an entirely different meaning? It wouldn’t be peer-to-peer anymore. Wouldn’t medical practice guidelines take on an entirely new meaning?
John
Patrons, I don’t think much weight is attached to the investigator’s assessment of causuality, it is more about the raw numbers of AEs in the treated and control groups.
Sidney Wolfe’s quote makes a great quote for news articles, but was not accepted at face value by committee chair Robert Herrinson, who forced Wolfe to backpedal.
DR. HERRINSON: Dr. Wolfe, we saw an Analysis this morning, based on simulation, that there may be upwards of a 20 percent type I error in the ANDROMEDA findings. So to say that we know it kills people — a little harsh?
DR. WOLFE: Well, I think that unless one believes that it is by chance — I think people have, at least around the table — the company may be stronger, for obvious and understandable reasons, in thinking that it’s a chance finding — I think that is enough of a likelihood that it is probable….
ex-FDAer
There’s no such thing as a free lunch.
Historically (go to the library of congress an look at the records fo the closed door discussions) the FD&CA was passed to protect the beef trusts and reignite confidence in the safety of meat products after the 1905 publication of The Jungle, and thereby bolster meat sales and profits. It was never about protecting the public.
The term clients came with the coming of PDUFA and was introduced by Woodcock. We can only do what we’re paid to do. They can take 90% of the reviewers and put them on project helping us develop new biased statistical methods etc.. Then that’s what happens. If they say take sign off from the reviewers and place it in the hands of a few managers who we will send you it’s done. No PDUFA controls all everything the drug division does.
Not enough time. Place 10 new foreign reviewers on a wine and dine program with Pharma so they become buddy buddy. Take good reviewrs and overwhelm them until they crack and leave or can be forced out. Place the new reviews on multiple reviews with drugs with 150,000 pages of documents and even more data files. Don’t train them in how to access the data files,give them a 2 month deadlines and no help. Step back and laugh. All to meet those bought and paid for PDUFA goals.
Resistence if futile. If you do you will never work anyway ever again. Of course this does’t even count the threats against your small children.
John
Whoa, threats against your children? Let’s hear more about that.
ex-FDAer
I would like to but I’m afraid if I do the resulting retaliation would be so much worse.
There’s a fine line at places between what I can say and what I can’t say and even if I say things that I’m allowed to FDA could still sue me and mess me up even more than they already have.
There is absolutely no way we can protect the public health. Big pharma wants a big drug approved, that’s that. We may delay things a little while but that just results in retaliation and elimination.
As Andy Von E says 3 strikes and you’re out.
Congress and the whistleblower protections are a complete joke. Congress will only give you a platitude of “thanks for reaching out to us” (don’t call us we’ll call you). Congress cares more when 2 dozen dogs are killed than when children, newborn infants, and the elderly are killed or maimed.
I would love to be able to provide the hard evidence but our systems says I can’t. They’re commercial secrets.
pharmavet
I’m glad that Pharma is able to afford the best expert outside statistical consultants to come to proper conclusions on data. About one year ago there was a suggestion made that Vytorin might be carcinogenic. Sir Richard Peto, one of the world’s preeminent statisticians made a compelling statistical argument that any tumors associated with Vytorin were no more likely to occur than by random chance alone. FDA accepted Dr. Peto’s analysis and concluded that there was no increased tumor risk.
There are expert biostatisticians at both FDA and industry. Unfortunately both entities sometimes have to bring in outside experts to refute claims of investigators who are given the data set and use inappropriate statistical methods to reach erroneous conclusions that could needlessly cause concern among patients.
M. Black
Sounds like a big game of hot-potato to me…
MB
ex-FDAer
These statistical methods are typically not able to confirm or deny a relationship due to the small number of cases. You also have to look at other factors such as biological plausibility. However even that is problematic. The best we can freqeuntly do is say there’s a possiblity and lets continue to watch and take action when we’re more certain. Unfortunately as we saw with Vioxx this approach is subject to abuse. As they all are.
Ideally I would like to see open HONEST public hearings discussing the strengths and weaknesses of each approach in a particular situation and not have it packed with too many people with any particular agenda or point of view. Then decide on what additional work may be needed (basic science, statistical analysis, observational data, public health implications, etc..)
Personally, I think I’m wishing for pie in the sky. This is money and politics and it will never happen.
ex-FDAer
By the way I’ve seen FDA statisticians withhold information that would have resulted in drug turndownd. Simply because they were afraid of possible retaliation and loss of their jobs.
pharmavet
Ex-FDAer, you would probably therefore not be surprised by the statement of one FDA senior manager, who once said “we’re not in the business of reviewing drugs; we’re in the business of approving drugs”.
ex-FDAer
No I’m not. I’ve heard variations of that inside FDA, i.e. “Our job is to approve drugs!”
patrons99
ex-FDAer -
Without naming names, can you tell us who made threats against your children? FDA or Pharma? Hobbs Act extortion and retaliation are both federal crimes, unless I’m sadly mistaken. Did it just happen once? You used the pleural “threats”. This is matter for DOJ Criminal Division.
ex-FDAer
FDA.
It is a matter for FDA management, HHS-OIG, and OSC. All of whom have documented track records of ignoring complaints or worse yet being involved in retaliation against anyone who does bring a complaint. As for the issue of plural threats lets just say that things went downhill from there.
It is not a matter for DOJ criminal division unless the above 3 working together forward a complaint to them. Even then DOJ has to want to actually do something. Otherwise nothing happens.
patrons99
ex-FDAer - “It is not a matter for DOJ criminal division unless the above 3 working together forward a complaint to them. Even then DOJ has to want to actually do something. Otherwise nothing happens.”
I could not agree more. We might wish to throw FBI into the mix, too.
ex-FDAer
I’ve spoken to my colleagues at FDA. Every single one says nothing is changing. I’m not surprised. It stinks from the head to the tail and it would take a wholesale house cleaning.
Numerous people were purged during the last administration, i.e. those reviewers who tried to do their jobs ethically.
The people who remain saw what happened not only to me but saw reviewers who were foreign nationals have their green cards revoked because they wouldn’t go along. They saw people physically removed by burly security guards. They saw retailiation including firings, attempted firings, and criminial prosecution. of anyone who went to congress. They were shown the following slide in a presentation http://verydemotivational.com/2009/12/21/where-did-that-penguin-get-cymbals/ and then told polar bears eat penguins.
Unless and until very senior managers at FDA who are responsible for things, e.g Janet Woodcock, John Jenkins, Sandy Kweder, Bob Temple, and a number of division directors as well as their goons are held accountable. Things are just going to get worse and what we’re seeing now is just window dressing.
Attempting to coopt these people is useless. They’ve been through it with prior administrations, 1/2 step for every 2 steps forward for pharma. You need something that will go 8 steps in the other direction in a single leap.
patrons99
ex-FDAer -
How do we get 8 steps forward in a single leap? Legislative reform? I’m troubled by the number of VERY pro-pharma laws on the books. For the Supreme Court to reach them all could take decades.
Even with better lawmaking, there still must be some political will to enforce the law. I completely agree with your view that the senior FDA management “as well as their goons” must be held accountable.
Perhaps, the court of public opinion should play a greater role.
Justice in MI
This is a smaller aspect of the larger discussion, but as I read the Sentinal story FDA did _not_ approve a cardio benefit claim that S-A wanted. So, to this point, the core story here sounds like its S-A once again manipulating NEJM rather than (primarily) FDA. Kind of like the CLASS saga.
patrons99
@ John 7:56 AM -
I’m not so sure that Dr Wolfe really needed to “backpedal” much.
This drug may prove to be Godsend. I don’t know. Personally, I will reserve judgment on that, for now. Even after FDA-approval, the question remains, do people with the more severe unstable recently changed heart status actually have an increased rate of death with it?
Dr Kaul and Dr Karnowsky seemed to share some of the same concerns that Dr Wolfe voiced.
At page 88 of 334, Dr Kaul stated, “Well, I had this question about did dronedarone improve symptoms in ATHENA? I mean, signs and symptoms of heart failure went in the wrong direction.”
At page 110 of 334, Dr Karnowsky stated, “With respect to heart failure deaths — I thought this was interesting because of the potential for this drug to be a negative inotrope. The vast majority of the people who died in ANDROMEDA died from heart failure deaths.”
At page 111 of 334, Dr Karnowsky stated, “But also of interest to me were the number of patients — and they’re highlighted and underlined — they’re highlighted in red — of evidence of worsening heart failure. These are symptoms that often occur with heart failure, peripheral edema, fatigue, asthenia and dyspnea. All went with about 25 to 30 events favoring placebo.”
At page 114 of 334, Dr Karnowsky stated, “So my conclusions are the results of ADONIS, EURIDIS and ATHENA suggest a benefit in delay of recurrence of afib. ANDROMEDA study suggests subjects with heart failure have an adverse outcome. ATHENA study suggests no overall adverse mortality outcome in patients without severe heart failure and a decrease in afib hospitalization, but I don’t know where the cross-over point is and where the neutral point is for saying this patient should not be treated with dronedarone because of their heart failure.”
At page 117 of 334, “DR. KARKOWSKY: Well, I would argue that it wasn’t just mortality that went in the wrong direction in ANDROMEDA. It was heart failure hospitalizations. It was all hospitalizations for cardiovascular reason.”
At page 295 of 334, Dr Kaul stated, “…the quality of the data regarding cardiovascular death are suspect…”
At page 300-301 of 334, Dr Wolfe stated,
“In the real world, there is no question — there is no question that you’re going to have people on the in-between group that are sicker than they were in ATHENA — and maybe merging into some of the characteristics that caused the deaths in ANDROMEDA.”
“…the people with the more severe unstable recently changed heart status actually had an increased rate of death with it.”
patrons99
@ JiM -
“So, to this point, the core story here sounds like its S-A once again manipulating NEJM rather than (primarily) FDA.”
“the NEJM study described Multaq as having significantly reduced cardiovascular deaths”
Are you suggesting that perhaps there might be a “ghost” exerting it’s will on NEJM? Perish the thought. Was NEJM actually manipulated by S-A? Hmmm. Such a thing could never happen!
“Before such a claim can be made, another clinical trial must be done, and Curfman maintained that doctors understand what the finding means.”
Well, some doctors probably do “understand what that means”. Sounds a bit like pharma-sponsored “bilge”, doesn’t it? Can the NEJM publication really be trusted?
ex-FDAer
Don’t think companies don’t manipulate the FDA too. It’s just that whatever they can’t get past FDA or have FDA senior management squash they go out and put misleading versions of the data in journals.
Companies try to write their study reports to make the same claims but FDA reviewers have the actual clinical trial designs and the raw data and they can call them on these falso claims.
Journals don’t get that stuff and simple look at what the companies wrote.
We often see published trials of studies we review that companies are making claims about in the study that we totally rejected. Happens all the time.
However sometimes when you find certain stuff FDA management suppresses it and will get rid of the reviewer so that the drug can get approved before generics kill the entire market and also so it won’t raise issues for the other drugs in the class.
patrons99
ex-FDAer -
It sure sounds like there isn’t much that the FDA won’t do to keep their Pharma clients happy. Kinda leaves the public in a tough spot. I mean, who’s really looking out for the public safety.
In another thread you said something pretty amazing,
“As for dead people. On average a drug that is taken off the market is only taken off after 5 years and a minimum of 10,000 patients per year are killed. Until it’ reaches these levels no one is likely to notice the dead bodies and will attribute things to something else, e.g. it was just his time.
So a lack of bodies may simply mean no one is looking.”
http://www.pharmalot.com/2010/05/generic-drugmakers-lose-preemption-argument/#comment-505008
I, too, have reason to believe that “a lack of bodies may simply mean no one is looking”.
Justice in MI
Right. Agree with all above, and Curfman essentially says so. This a couple of paragraphs I’ve included before from an article quoting Peter Honig, then a Merck VP (perhaps he still is):
Dr. Honig said Merck’s promise to publish the outline of its studies well before they are completed was an important way to ensure the integrity of scientific publishing. He said that companies sometimes changed the parameters of their studies after they had been completed to make the results look better.
“When I was at the F.D.A.,” Dr. Honig said, “I would read peer-reviewed articles in journals and realize that the endpoints weren’t the original ones from the study because I had reviewed the original protocols myself.”
pharmavet
Ex FDAer. Don’t discount the power of the published paper. I’m not sure how far you go back, but one of my first assignments in clinical research in the early 1980’s was to assemble what the FDA called a “paper NDA” for new indication. This was before FDA routinely required clinical trials for Supplemental NDA’s for new claims. These basically consisted of a literature review of published trials supporting the new claim and were an inexpensive way of expanding your PI. Even today, companies routinely cite published data when submitting Citizens Petitions supporting or refuting a new claim, or to counter claims made by a competitor.
ex-FDAer
I don’t totally discount published papers.
As an aside I’ve also learned to listen to patients and not discount anything they tell me even though it’s not what my current thinking or education tells me.
But having reviewed many clinical studies at the FDA I came to the same conclusion as Peter Honig quoted above.
When I know the clinical study design, the primary objectives, the raw data, and the appropriate statistical analysis and know that the company does too either because they did do it appropriately or because we had to redo an appropriate analysis and then told them. Then when I find that instead their published studies are misleading and likely intentionally so. Under these scenarios I find that in the world that we currently live in you simply don’t know which published papers to trust and which ones not to.
patrons99
ex-FDAer - “Under these scenarios I find that in the world that we currently live in you simply don’t know which published papers to trust and which ones not to.”
I’ve been saying that for a long time, too. You might as well pile up all of the medical literature over the last 20 years and burn it.
I feel sorry for the new doctors just out of training. Without sources of unbiased CME, they will only too readily become brainwashed. I know it sounds a trifle heretical, but medical practice guidelines cannot be trusted.
Unless recently trained doctors really work at it, they will find it VERY difficult to find CME that is trustworthy. I won’t name names of the most egregious CME sites. They know who they are.
patrons99
@ ex-FDAer - “I’ve spoken to my colleagues at FDA. Every single one says nothing is changing. I’m not surprised. It stinks from the head to the tail and it would take a wholesale house cleaning.”
If you will permit me a brief digression, catalyzed by your comment. Have we become complicit by our silence?
It seems as though over the last year or so of blogging, I’m starting to see a number things re: Pharma and FDA, just keep repeating themselves, with no real change for the better…in fact, things are arguably getting worse.
I recently found a quote from Ian Crane which I’d like to share, in an article titled “Big Pharma & Social Conditioning” on December 17, 2007.
http://foodcode.blogspot.com/2007/12/big-pharma-social-conditioning.html
“It’s time to choose whether you will be complicit by your silence…or whether you will take responsibility and be part of the change you want to see!”
ex-FDAer
I was not silent. Others were deaf.
Justice in MI
What a great line, ex-FDAer. And how many contexts it applies to!
How many times have we heard some version of: “If the public really knew X, they’d riot in the streets.”
But somehow they never do. (almost) Keeping people busy and just “tryin to survive” (to go back to an earlier discussion on students and Adderall) is as powerful a social control mechanism as it gets.
pharmavet
Ex-FDAer, I think that we have to be fair to the research and publication processes. Anyone who has spent time in a lab or doing clinical research knows that research itself tends to be somewhat messy, and the way you see data presented in a nice orderly fashion in a publication is not necessarily the order in which the experiments happened in real life. A publication by definition is creating a coherent narrative out of a series of experiments performed to test an hypothesis, and that’s all. My doctoraal dissertation ran 237 pages, and not every single experiment or blind alley that I ran into in five years as a grad student is included in the thesis. I published five papers based on my thesis research; each one of them reflected a different part of the overall research. Therefore, some selectivity was needed in deciding which data was most relevant to which paper. This is not misleading.
Publications have limited space. If FDA, for example requires a table of treatment-emergent adverse events > 1% incidence, that tends to be a long table. The same table might be submitted to a peer-reviewed journal, and the editors will insist on shortening the table to incidence > 5%. This is not misleading; it is simply a response to the reviewer’s demands. The only other alternative would be not to have the paper published at all, which I don’t think anyone wants to see.
ex-FDAer
Pharmavet,
While I agree with everything you say. That is not the situation I am talking about.
I am talking about studies for additional indications. There may be 1 or 2 phase III studies that were performed where the drug simply didn’t work. There were clear primary objectives and the study didn’t show any difference and we rejected the claim for the indication. Instead the companies go ahead and publish them showing secondary endpoints and claiming efficacy. Or they change the statistical methods so they can get a significant result but they violate the underlying assumption in doing so.
Other claims might be our drug has an onset of 1.5 hours but the first point they measured was 3 hours so they took the average of 0 and 3 hours. 3 hours is the earliest that they have any evidence for. Maybe there is an earlier onset but they could have just as easily said onset is within 30 seconds.
Toxicity is a whole ballgame in and of itself and I was not talking about toxicity or AEs. I was simply talking about clearly false information on efficacy being published that the FDA has already reviewed and rejected and warned the company not to promote.
pharmavet
Thanks, ex-FDAer. If a companiy publishes a efficacy paper based on a secondary objective and simply drops the paper off with a doctor as a “leave behind” I don’t see that there is that much that can be done with it. If the the secondary indication is stated a priori in the “Methods” section of the paper and the pre-stated statistical methods were used to reach the conclusion, and if the writer uses the appropriate qualifiers in the Discussion section of the paper, such as the data “suggest” rather than “support” proof of efficacy, than this is not misleading. I would even support making the same statement (suugest rather than support)about a primary endpoint where the p-value fell between 0.05 and 0.10, although that p-value may not meet with regulatory approval. I still don’t know if we’re onn the same page on this matter, but these are just a few thoughts.
BTW, you and I agree that any promotional material with efficacy claims or indications based on secondary endpoints should at least merit an NOV from DDMAC.
vince
As Pharmavet said “……Sir Richard Peto, one of the world’s preeminent statisticians made a compelling statistical argument that any tumors associated with Vytorin were no more likely to occur than by random chance alone. FDA accepted Dr. Peto’s analysis and concluded that there was no increased tumor risk.”…….but when Jupiter showed a decrease in cancer deaths well did the industry call in the hired gun ‘don’t think so” the study they published showed a reduction in mortality rate ‘a rarity in statin trails driven by the decrease in cancer deaths.’ So this is the new model add patients o lengthen trails or shorten them .Wait for the moment when you can claim a significant benefit immediately stop the study send checks all around.
LF Velez
RE: Patrons99’s query about ghostwriting…
The standards for authorship clearly say that for someone to qualify for “author” status, they need to have been involved in data analysis/collection, have contributed substantially to the manuscript AND had final approval of the manuscript. All three items need to be satisfied [which is why I, as a writer, am never allowed to be listed: I may have analyzed data and contributed substantially, but I don't have final approval over publication]. Dr. Page didn’t fulfill all of the criteria either, so, yes, there’s ghostwriting and ghostauthorship going on.