Journal Corrects Vioxx Risk Study, Citing Bad Math

By Maggie Fox, Health and Science Correspondent

WASHINGTON (Reuters) – A critical study showing the heart risks of Merck and Co.’s now-withdrawn arthritis drug Vioxx contained math errors that wrongly showed the risks changed over time, the New England Journal of Medicine reported on Monday.

The journal and the authors, including a team at Merck, corrected the May 2005 study to show the risks do not, as originally shown, greatly increase after 18 months.

In fact, it is not possible to tell when the risk of heart attack or stroke shoot up, Journal editor Dr. Jeffrey Drazen said.

Merck withdrew Vioxx from the market in September of 2004 after a three-year study showed it doubled the risk of heart attack and strokes in patients taking it for at least 18 months.

Last month, Merck said it was correcting one of the statistical methods used to analyze data from that key study, but said the change did not alter the results of that clinical trial.

But Drazen said the mistake was significant enough to change some of the conclusions of the trial — including the important 18-month cutoff.

"They found this mistake. It turned out they did the wrong math. The proportional hazards calculation in fact was not significant," Drazen said in a telephone interview.

"You couldn’t make this argument that the hazards weren’t proportionate over time. It meant a lot of their conclusions they had originally drawn no longer could be accepted."

The correction to the data from the study, called APPROVe, involves statistical and graphical methods used to assess whether the relative risk of Vioxx compared to placebo was constant or if it changed over time.

The original conclusions suggested that the biggest risk did not become apparent until 18 months after a patient first took the arthritis drug.

"Because of the mathematical errors, the editors allowed some statements in the original file which we no longer feel are supported by the weakened evidence," Drazen said.

Merck is facing more than 11,500 product liability lawsuits from people claiming to have been harmed by Vioxx, which earned $2.5 billion a year before it was pulled.

Vioxx, known generically as rofecoxib, is in a class of drugs known as COX-2 inhibitors. These drugs were ironically designed to be safer alternatives to aspirin, ibuprofen and other non-steroidal anti-inflammatory drugs or NSAIDs, which can cause sometimes fatal stomach bleeding.

Drazen defended the Journal’s handling of the study, which has been criticized.

"The papers always said that although it took 18 months for you to appreciate a difference in risk, it didn’t mean you had to be exposed to the drugs for 18 months," Drazen said.

In other words, a patient could take Vioxx for three months, or six months, and not have the bad side-effects until 18 months later.