November 30, 2004

Vioxx E-Mail Exchange Shows Questions

WASHINGTON -- Government questions about "basic data integrity" derailed publication of a report raising safety concerns about arthritis drug Vioxx on the eve of a contentious congressional hearing, according to an e-mail exchange.

Without a clear agency go-ahead, Food and Drug Administration safety reviewer David Graham withdrew the paper from consideration for publication in The Lancet medical journal. The paper said high doses of Vioxx tripled the risk of heart problems compared with a rival painkiller.

The whistleblower told Congress on Nov. 18 that drug regulators fumbled on safety with Vioxx. Graham said the dysfunctional agency left America virtually defenseless against the next Vioxx debacle and he named five other problematic drugs on the market.

The Lancet's editor, Richard Horton, had already pointed to "lethal weaknesses" in FDA's regulatory oversight regarding Vioxx in a scathing editorial published in early November.

Graham's paper was a peer-reviewed version of research he presented this summer in France. It would have noted that heart attacks and sudden cardiac deaths were higher among Kaiser Permanente patients taking high or low-dose Vioxx, compared with Celebrex.

The FDA said Monday it didn't try to thwart publication of the article, but at the same time noted that Vioxx already had been withdrawn from the market. Merck & Co. cited an increased risk of heart attack and stroke when it pulled the drug Sept. 30.

According to the e-mails, FDA superiors peppered Horton with last-minute concerns about the article's scientific accuracy. USA Today published a story on the exchange Monday.

"You will not be surprised if I say that I was a little taken aback to get your call," Horton replied by e-mail from England. "It is very unusual, indeed, for a member of the employing institution of an author to contact us in the middle of the review and publication process of a manuscript. Such an intervention could easily be perceived as an improper injunction in our peer review process."

FDA reviewers "identified scientific issues with the paper," Dr. Steven Galson, acting director of the agency's Center for Drug Evaluation and Research, wrote Horton on Nov. 12. "Some of the questions involve basic data integrity. (Graham) has declined to address these issues, and we think it's important that your readers know this."

Graham's study examined the records of 1.4 million Kaiser Permanente patients taking typical doses of Vioxx as well as daily doses greater than 25 milligrams. High doses of Vioxx tripled the risk of heart attack and sudden cardiac death. Even typical Vioxx doses, however, were more dangerous for the heart than Celebrex, the research team found.

In an e-mail sent a few hours later on Nov. 12, Galson pointed to discrepancies between the number of heart attacks and sudden cardiac deaths in an earlier paper vs. the Lancet manuscript. The earlier paper attributed 5 cases of heart attack and sudden cardiac death to high doses of Vioxx. The Lancet draft paper raised that to 10 cases.

"As these changes could affect the interpretation of the study, it would seem to be important to determine, before publication, whether these changes were made as part of a pre-specified process, who was involved in making the decisions, and when these new cases were added," Galson wrote.

In addition, the Lancet draft gave slightly different findings for such anti-inflammatories as naproxen and ibuprofen.

Galson's e-mail to Horton said the agency "would be happy" to submit its concerns as a formal letter to be published in the Lancet. "However, I hope these issues can be resolved without making that necessary."

Kaiser's David Campen, one of the paper's co-authors, said after submitting an abstract to the American College of Rheumatology that the team enlisted more researchers to examine their analysis. The team further refined what constituted a high dose of Vioxx.

In 30 to 40 cases, Kaiser's electronic database was unclear about dose. The research team reviewed those borderline cases one by one without knowing health outcomes to determine whether those people were taking high or typical doses of Vioxx.

Horton defended Graham's integrity and pointed to detailed responses Graham e-mailed in reply to earlier FDA questions. By Nov. 14, the Lancet editor said he was satisfied that FDA questions had been addressed.

However, with no FDA reply in hand by the noon Nov. 16 deadline, Graham withdrew the paper.


On the Net: