Thursday, May 24, 2007

Expanded Avandia Story Now Echoes Familiar Themes

The story about the possible cardiovascular adverse effects of rosiglitazone (Avandia, made by GlaxoSmithKline [GSK]) just continues to get more interesting, now that more reporters are digging around about it. In particular, today Stephanie Saul and Gardiner Harris wrote in the New York Times that the US Food and Drug Administration (FDA) had been warned in 2000, seven years ago, about possible cardiovascular side effects of Avandia.

A leading diabetes doctor sent the Food and Drug Administration a letter seven years ago that warned of the heart risks of the drug Avandia.

The documents, found in a reporter’s search of the F.D.A.’s database, indicate that the agency had been warned of safety concerns with the Type 2 diabetes treatment Avandia, and that the drug’s maker, GlaxoSmithKline, was seeking to minimize Avandia’s risks....

The letter in 2000 to the F.D.A. was written by Dr. John B. Buse, chief of endocrinology at the University of North Carolina in Chapel Hill, who is about to become the president of the American Diabetes Association. His letter from seven years ago sounded an alarm about Avandia, citing 'a worrisome trend in cardiovascular deaths and severe adverse events' among patients using the drug.

In his letter to the agency, dated March 15, 2000, Dr. Buse was highly critical of the drug maker’s marketing of Avandia, accusing the company of 'pervasive and systemic' efforts to play down the drug’s risks and overstate its benefits.
Dr. Buse said that he does not generally prescribe Avandia to patients. He has been an outspoken critic of the drug in medical education meetings, some of them sponsored by Takeda and Eli Lilly, which jointly market a competing drug, Actos.

He was also an investigator in a study comparing Avandia with Actos, sponsored by Eli Lilly, that showed Actos had better effects on cholesterol than Avandia.

Dr. Buse said yesterday that he wrote the letter in 2000 in response to an F.D.A. petition filed by Dr. Sidney Wolfe, a consumer activist, who had asked the agency to place warning labels on Rezulin, Avandia and Actos.

Dr. Wolfe’s Health Research Group, a part of Public Citizen, has long warned patients not to use any of those drugs. At the time, the F.D.A. was considering removing Rezulin from the market, and Dr. Buse objected. Rezulin was made by Parke-Davis, a division of the Warner Lambert Company.

'The way I felt about it, after several years of clinical availability of Rezulin, we kind of understood the problems with it,' he said, 'and we didn’t understand the problems of Actos and Avandia.'

Dr. Buse analyzed data submitted to the F.D.A. in support of Actos and Avandia and came to the conclusion that there was a 'hint, a whisper' of cardiac-related deaths with Avandia as well as evidence of negative effects on cholesterol.

Referring to Avandia by its generic name, rosiglitazone, and to Rezulin as troglitazone, Dr. Buse wrote in the letter, 'I do not believe that rosiglitazone will be proven safer than troglitazone in clinical use under current labeling of the two products.' He added: 'In fact, rosiglitazone may be associated with less beneficial cardiac effects or even adverse cardiac outcomes.'

In the 2000 letter, Dr. Buse asked the agency to call for head-to-head studies of all the drugs.

Yesterday, he said, 'I would say that in the last several years, there has not been a study that’s made me feel better about this.'

The Times reporters also uncovered evidence that GSK had been warned about over-zealous marketing of Avandia that played down its possible side effects.

The F.D.A. was conducting its own investigation of Avandia marketing and found that company representatives were denying the existence of changes on the drug’s label that the F.D.A. had already ordered, which were meant to flag Avandia’s risks to the heart and liver.

The agency sent the drug maker a warning letter in July 2001, citing misleading statements made by company representatives at a recent meeting of the American Academy of Clinical Endocrinologists, where the agency had sent undercover investigators. The F.D.A’s letter criticized company posters displayed at the meeting, saying they did not carry adequate warnings.

It was the third time that the agency had chastised the company about its promotion of Avandia. Because of the repeated warnings, the agency demanded that the company send out a letter to doctors specifically warning them of the risks. The company subsequently sent out that 'Dear Doctor' letter on Sept. 6, 2001.

One wonders in retrospect whether the unfortunate timing of that letter, just before the attacks on the World Trade Center in New York, diverted attention from it.

In any event, this story has acquired even more elements that ought to be familiar to readers of Health Care Renewal. First the story seemed to be about how the designs of clinical studies of rosiglitazone make them less likely to find clear evidence of cardiovascular risk, how failure of the results of some of these studies to be published lessened attention to then, and how the published reports of the bigger studies were written to divert attention from cardiovascular risk.

Now there are new, but familiar elements:

  • Early warnings from at least one whistle-blower were apparently ignored.
  • The pharmaceutical promoted the drug enthusiastically, apparently again in a way designed to distract attention from particular adverse effects.
  • Just to muddy the water, the whistle-blower had financial relationships to manufacturers of competing drugs, possibly leading to questions about whether his whistle-blowing may have been influenced by them.

So let me reiterate some implications.

Failing to Disclose Clinical Research Results Violates the Trust of the Research Subjects

People who perform clinical research, that is studies of real, live people, especially experimental studies, that is, randomized controlled clinical trials have an obligation to their study subjects to make the results of these studies as widely available as possible. Otherwise, they have violated the trust of their study subjects, who participated under the assumption that their efforts were going to advance science and health care.

It makes no ethical sense to think that because a drug (or other) company pays for a clinical study, i.e., a study of real people, it owns the data from the study, and has no obligation to see that the results of the study are made public. (The exception would be if the subjects were told plainly the study was done for marketing or other commercial purposes, that is, to benefit the company, not science or health care. One suspects that if patients were told this, they would demand such high fees for participating that the conduct of such studies would be prohibitively expensive.)

It is very worrisome that at least one high FDA official was quoted as seemingly disturbed by the wider distribution of the results of clinical trials. The New York Times reported yesterday,

Some experts also believe that releasing the results of hundreds of studies involving drugs or medical devices might create confusion and anxiety for patients who are typically not well prepared to understand the studies or to put them in context.

'I would be very concerned about wholesale posting of thousands of clinical trials leading to mass confusion,' said Dr. Steven Galson, the director for the Center for Drug Evaluation and Research at the F.D.A.

Recently, a report issued by the Institute of Medicine, a part of the National Academy of Sciences, recommended that the F.D.A. release all summaries of study data it had collected in the process of approving new drugs as well as all post-marketing studies of those products.

The F.D.A. rejected the first recommendation as overly burdensome and Dr. Galson, the director of the F.D.A.’s drug evaluation and research, said that the agency already released much of this information. 'It is not that we are philosophically opposed to it, but the work would be enormous, he said.

There is an argument for making sure that clinical trial results are disclosed in a clear, non confusing way. That may take some work and cost some money. But failing to disclose the results because of worries about "confusion," again, violates the trust of the study subjects.

Patients and Physicians Need to be Very Skeptical About Drug and Device Marketing

Once again, it appears that a drug company was pushing one its products beyond what the clinical research evidence supported. In this day and age, patients and physicians should be extremely skeptical of any and all marketing claims by drug, device, and biotechnology companies (and all other health care organizations, for that matter). Instead, we ought to be looking at the biomedical science and clinical research evidence.

Pervasive Conflicts of Interest in Health Care Make Figuring Out What is Best for Patients Even Harder

This story is muddied by the apparent conflicts of interest of the major players, even the "good guys." Dr Buse, the apparent whistle-blower about the hazards of Avandia, had financial ties that may have made him more critical of that drug. Thus, his warning message may have not been so easy to interpret. Even Dr Steve Nissen, who dug out and meta-analyzed the data about Avandia, has been criticized for his conflicts of interest, per this report by Michelle Fay Cortez for Bloomberg,

While drugmakers pay for much of the research that Nissen conducts, the nonprofit Cleveland Clinic provides his salary -- one bit of data that the doctor won't disclose. To distance himself from industry, Nissen set up a separate foundation to receive consulting income and speaker fees.

The arrangement hasn't shielded Nissen from criticism that he sometimes goes too far to promote medicines that he studied for drugmakers.

Sales of Crestor [rosuvastatin, made by AstraZeneca] swelled after a study from Nissen showed the cholesterol medication reversed signs of heart disease in some patients. He praised the drug from the podium of a medical meeting in March 2006, saying the study showed 'unequivocal' evidence that disease retreated with high-dose Crestor therapy.

He took over as president of the meeting sponsor, the American College of Cardiology, the same evening. Crestor sales have climbed at least 50 percent in every quarter since then, rising 62 percent to $628 million in the first three months of 2007.

His handling of the Crestor study casts doubt on Nissen's status as a public-health defender, says Sidney Wolfe, the director of Public Citizen's Health Research Group in Washington. Nissen wouldn't have been so bullish if he hadn't received the research grant, Wolfe says.

'This paper, and the way it was played, is an example of what happens with financial conflicts of interest,' Wolfe says. 'If he had not been an investigator for the study, I don't think he would have put quite the same spin on the results.'

So the Avandia story turns into another argument that physicians, other health care professionals, and leaders of health care organizations should rid themselves of conflicts of interest that might influence their professional, academic, or research work on behalf of patients or public health.

Oh, and finally,

Failing to Disclose Research Results Is Bad for Patients

According to the evidence-based medicine paradigm, patients and physicians should together make decisions about health care interventions based on a critical review of the best available evidence from clinical research, combined with the physicians' knowledge of the clinical context and biology, with consideration of the patient's values. Hiding relevant evidence (because a sponsoring company feels it owns the evidence and the evidence is not favorable to that company's products) distorts this decision making process, creates pseudo-evidence, and will lead to patients not getting the best tests and treatments for them. That is bad for patients.

ADDENDUM: See also comments on GoozNews, Retired Doc's Thoughts, PharmaGossip, and again PharmaGossip, PharmaLot, and again Pharmalot.


Martin said...

I think you are absolutely right when you write that refusing to publish clinical trial data violates the trust between research subjects and the sponsor. It seems to me that IRBs should take this into account when evaluating the risk-benefit profile of clinical research protocols.

Anonymous said...

Martin: Hear, hear.

More comment at Derek Lowe's In The Pipeline:

Anonymous said...

Steve Nissen did all of his analysis using data that GSK posts on their website for anyone to read. Makes all your heated talk of hiding evidence seem kind of silly.

Anonymous said...

"...Makes all your heated talk of hiding evidence seem kind of silly"

Not if you look at their marketing:

Aubrey Blumsohn said...

Actually I think he only had access to GSK's filtered tabulated summary data. Data = raw data. That remains hidden.

David J. Phillips said...

To quote yourself: "Once again, it appears that a drug company was pushing one its products beyond what the clinical research evidence supported."

Contrary to what Nissen purports, the PROactive Trial DID NOT prove that Actos has a more benign CVD profile than AVANDIA:

On reading the NEJM meta-view, troubled with Nissen’s comment that the observed risks of rosiglitazone might not be a "class effect" of thiazolidinediones:

“Pioglitazone is a related agent also widely used to treat type 2 diabetes mellitus. However, unlike rosiglitazone, pioglitazone has been studied in a prospective, randomized trial of cardiovascular outcomes, called Prospective Pioglitazone Clinical Trial in Macrovascular Events (PROACTIVE). The primary end point, a broad composite that included coronary and peripheral vascular events, showed a trend toward benefit from pioglitazone (hazard ratio, 0.90; P=0.095). A secondary end point consisting of myocardial infarction, stroke, and death from any cause showed a significant effect favoring pioglitazone (hazard ratio, 0.84; P=0.027).”

In plain English, “a trend toward benefit” is not material. Specifically, the primary end point occurred in 21% of patients on pioglitazone, compared with 25.5% of the patients on placebo. Pioglitazone's total risk reduction of 10% was not considered statistically significant. The study was powered to show a 20% reduction in the composite of primary end-point events.

Albeit the secondary end point—death, nonfatal heart attack, and stroke—was statistically significant, Nissen failed to mention heart failure occurred more frequently in the pioglitazone group (10.8% of patients) than in the placebo group (7.5% of patients), with 5.7% versus 4.1% being hospitalized.

David J. Phillips, Publisher


Anonymous said...

I agree about Actos