Today's Wall St. Journal reports that Merck's most powerful defense witness may have become a distinct liability. [subscription required]. This piece elaborates on Ted Frank's Dec. 10 entry, below.
Merck's witness, Dr. Alise Reicin, was an author of a key NEJM study comparing Vioxx to a competitor's product. The November 2000 study stated that in a study of Vioxx versus naproxen, an older, generic painkiller, 0.4% of Vioxx users had heart attacks compared with 0.1% of naproxen users. The study's aggregate numbers, which weren't in the article, showed that 17 out of about 4,000 Vioxx users had heart attacks versus four of 4,000 naproxen users. Merck explained the difference by concluding that naproxen helps prevent heart attacks, though no clinical trials have shown that to be the case.
The article didn't mention an additional three heart attacks among Vioxx users, which occurred after the article was submitted for publication. These deaths raised the total heart attacks to 20 of 4,000 patients, or 0.5%. The additional deaths didn't affect Merck's conclusions at all. Nor was there any misstatement -- the facts were exactly as reported as of the date of submission for publication.
Merck said last week that it excluded the three heart attacks because they took place after the cut-off date the scientists had set for measuring benefits or risks. And the attorney for Merck noted that the additional heart attacks wouldn't have changed the conclusions of the journal article. Despite this, plaintiff in the federal Vioxx case on Friday asked for a mistrial on grounds of fraud. The judge has taken the motion under advisement, and will presumably only rule on it if Merck wins a jury verdict (the jury is still out as of this writing).
So, let's see, the failure to submit information irrelevant to a medical journal article has diminished the credibility of an author of the article. Makes sense to me!