NPR logo

Medical Journal Publishes Correction on Vioxx Study

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript
Medical Journal Publishes Correction on Vioxx Study

Research News

Medical Journal Publishes Correction on Vioxx Study

Medical Journal Publishes Correction on Vioxx Study

  • Download
  • <iframe src="" width="100%" height="290" frameborder="0" scrolling="no" title="NPR embedded audio player">
  • Transcript

The New England Journal of Medicine has published a correction to a Vioxx study that has figured prominently in Merck's legal defense of the drug. The medical journal says that a reinterpretation of the study shows Vioxx increases cardiovascular risk soon after patients start taking the painkiller. Merck says the study continues to show Vioxx was safe for the first 18 months of use.


It's MORNING EDITION, from NPR News. I'm Steve Inskeep.


And I'm Susan Stamberg, in for Renee Montagne.

Vioxx in the news again. Seems an unusual public disagreement broke out yesterday between The New England Journal of Medicine and the makers of Vioxx, both issuing statements about a study published last year on the drug. The medical journal corrected a key finding of the study, while Merck stands behind the original.

Vioxx was withdrawn from the market in 2004, after this study showed it doubled the risk of heart problems. NPR's Snigdha Prakash reports the correction brings attention to how the prestigious journal handled the study in the first place.


The dispute between the journal and Merck was triggered by an arcane statistical error in the original study. The New England Journal says the study, called a prove, no longer shows that Vioxx was safe for the first 18 months that patients took the drug. Merck says that conclusion is still true.

Statistician Janet Wittes, who helped analyze a similar study about the risks of the painkiller Celebrex, agrees with The New England Journal that the conclusion is wrong. But, she says it's been wrong all along, and the reasons she reels off have nothing to do with the statistical error that the journal and Merck are arguing over.

Dr. JANET WITTES (President, Statistics Collaborative, Inc.): Number one, the fact that you don't see harm during an initial period of taking a drug doesn't mean that nothing has happened.

PRAKASH: Wittes uses the effect of radiation as an example. Just because the effects may not be evident for years, she says, doesn't mean that a brief exposure wasn't harmful. The same could be true of Vioxx, she says.

Dr. WITTES: Number two, more people in the Vioxx group than in the placebo group dropped off of the study.

PRAKASH: Making it likely that heart attacks and strokes in the Vioxx group would be undercounted.

Dr. WITTES: Oh, and number three, that wasn't even the right picture.

PRAKASH: Merck used a graph to show that patients on Vioxx and those on placebo had the same low risk of heart attack and stroke for the first 18 months of the study. And, that among Vioxx patients, the risks started to increase at 18 months. But Wittes says that type of graph cannot show how the risk changes from one month to the next.

And she isn't alone in her opinions. Before the paper was published, four of the five scientists who reviewed it for The New England Journal raised so many questions about the claim that Vioxx was safe for the first 18 months that the editor in charge of the paper asked the authors of the Vioxx paper to delete the claim. But that didn't happen.

The New England Journal's editor-in-chief, Jeffrey Drazen, says there was a negotiation between the authors and the Journal's editors. Drazen saw another flaw in the paper that he wanted fixed even more.

Dr. JEFFREY DRAZEN (Editor-in-Chief, The New England Journal of Medicine): And we pressured the authors to put that data in the paper, and in so doing, we said we were willing to allow them to continue to put in the information about the 18 month separation as long as they clearly labeled it as a post-hoc analysis.

PRAKASH: Or a speculative analysis. Former editor-in-chief of the New England Journal, Marsha Angell, says editors negotiate with authors all the time. But, Angell says...

Dr. MARSHA ANGELL (Former Editor-in-Chief, The New England Journal of Medicine): In this case, I would not have yielded an inch, if I were the editors. I would have watched them like a hawk.

PRAKASH: After all, Angell says, The New England Journal had already been burned once by Merck.

In November, 2000, it published a paper on the largest Vioxx study ever done. That paper gave a full account of Vioxx's benefits, but said little about its cardiovascular risks.

Dr. ANGELL: And that should have made the editors especially skeptical of anything that was said by a company-sponsored trial of the same drug, particularly in the face of the drug having been withdrawn and of all the lawsuits. I mean, in a sense, you can say, burn me once, shame on you, burn me twice, shame on me.

PRAKASH: Jeffrey Drazen says he disagrees.

Dr. DRAZEN: We felt that it was appropriate to allow the authors to speculate. And I respectfully disagree with Dr. Angell. I think that we were on top of this and we're on top of this again with this correction.

PRAKASH: Yesterday's correction won't be the last word on this issue. Data from the study is being analyzed by independent statisticians, and an updated analysis of the study's results recently submitted to the Food and Drug Administration will likely also be published in the journal. It's already known that in that analysis there is no 18-month period of safety for Vioxx patients.

Snigdha Prakash, NPR News, Washington.

STAMBERG: Read about health risks linked to Vioxx at

Copyright © 2006 NPR. All rights reserved. Visit our website terms of use and permissions pages at for further information.

NPR transcripts are created on a rush deadline by Verb8tm, Inc., an NPR contractor, and produced using a proprietary transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative record of NPR’s programming is the audio record.