Merck & Co. withheld information about heart problems suffered by patients taking the company’s painkiller Vioxx from the manuscript of a journal article published in 2000, the New England Journal of Medicine said yesterday.
The journal said pertinent information about the problems had been deleted from the manuscript, and the article’s authors, who included two Merck employees, failed to disclose information about three heart attacks, though they knew about the incidents at least four and half months before publication.
Had the data been reported, the study, which was sponsored by Merck, would have shown that the risk of heart attack for people taking Vioxx was five times higher than for those using naproxen, another painkiller, the Journal reported in an on-line ”expression of concern.”
Vioxx became a blockbuster drug that Merck did not pull from the market until four years later, after another study showed long-term use increased the risk of heart attacks and strokes. The drug is now the subject of thousands of lawsuits, and the problems with the Journal article were revealed during a deposition.
The Journal has asked the authors for a correction. Besides the two Merck employees, there were 11 other authors, all of whom were academic scientists who had financial ties to the company.
The lead author, Dr. Claire Bombardier, said in an e-mail that she believed the study had ”appropriately disclosed the data” and that the authors are preparing a formal response to the Journal editorial.
Four of the authors, including Dr. Alise Reicin, Merck’s vice president of clinical research, did not reply to telephone calls late yesterday seeking comment.
In a statement, the company said it ”promptly and appropriately disclosed the results” of the trial, and that the additional heart attacks ”did not materially change” the article’s conclusions.
Yesterday, jurors began deliberations in the third of thousands of lawsuits alleging harm from Vioxx. Testimony included details of the heart risks uncovered by the clinical trial that was the subject of the Journal report and whether Merck was forthcoming about what and when it knew about safety risks posed by Vioxx.
Dr. Gregory D. Curfman, the Journal’s executive editor, said it learned of the new information about two weeks ago.
”We believed that this new information cast in doubt not only some of the data in the article, but also some of the conclusions in the article,” Curfman said. ”As editors, it’s our responsibility to stand behind the work that we publish. And we felt we could no longer do that.”
The results of the clinical trial, published by the Journal on the Nov. 23, 2000, sought to show that Vioxx eased pain without the stomach bleeding associated with older painkillers. In addition to other outcomes, scientists running the trial tracked such serious side effects as heart attacks, stroke and sudden cardiac death.
Curfman was subpoenaed and on Nov. 21 he was interviewed by plaintiff attorneys pursuing a Vioxx lawsuit. He was shown a July 5, 2000 memorandum that made clear at least two of the study authors knew before the article was published that some heart attack data was missing.
On Oct. 5, 2004 days after Vioxx was removed from the market, Journal editors had examined a computer diskette that had been submitted along with the paper manuscript. It showed that details about heart attacks and other cardiovascular problems were deleted two days before the publication received the manuscript on May 18, 2000. The diskette also indicated that the computer used to delete that information was owned by Merck.
Curfman said the more complete numbers raise questions about Merck’s contention.
Dr. John Abramson, author of ”Overdosed America,” called the New England Journal retraction ”enormous.”
Abramson said the original study had cemented Vioxx’s reputation as a safer alternative to some other painkillers.
”When articles are published in our most respected medical journals, they become medical knowledge and doctors act on that information believing they’re providing the best medical care to their patients,” he said.
Dr. Jerome Kassirer, editor-in-chief of the New England Journal of Medicine from 1991 to 1999, said that in his tenure, ”any time there was a hint of any kind of abnormality, we tried to correct it.”
While medical journals review submitted studies, sending them to a panel of specialists for vetting, journal editors fundamentally rely on the veracity of researchers, Kassirer said. And that relationship, he said, is not likely to change even with the disclosure about the Vioxx study.
Journals could request every piece of data collected in research books, Kassirer said, but that is neither practical nor foolproof.
”I suppose they could ask for the books, but when you are publishing four or five original articles every week, there’s no way you can look at the books. And besides that, the books could be cooked,” Kassirer said. ”The fact is, you can’t be a data detective, and you have to rely on the researchers to give you honest data. There really is no other choice.”
Merck said it reported complete data including the information withheld from the Journal article to the FDA in 2000 and also to federal advisers during a February 2001 committee meeting. The FDA approved Vioxx in May 1999.
Dr. Alastair Wood, who chaired a federal advisory meeting in February 2005 that resulted in a vote that would permit Vioxx back on the market, said the new data could have put more pressure on the FDA to request additional safety studies.
It is unclear whether the disclosure will affect the 7,000 Vioxx lawsuits that have been filed nationwide or the case currently before jurors in Texas. Changes to a computer file, by themselves, do not prove intent, said Victor Schwartz.
”Did someone at Merck deliberately change the numbers? That’s the issue,” said Schwartz, general counsel for The American Tort Reform Association. ”If you were trying a case, you would have to know the ‘why’ of it. And one should not assume that they were simply taken out to make the numbers look better.”