The study's findings were dramatic: 20 patients on Vioxx had heart attacks. Only four had heart attacks on the comparison drug.

But that's not the study that prompted Merck & Co. to yank Vioxx off the market two weeks ago. That study, the VIGOR study, was published in The New England Journal of Medicine in 2000, and was used to bolster Vioxx's reputation for causing fewer gastrointestinal side effects than other non-steroidal anti-inflammatory drugs.

The heart attacks were mentioned briefly, and the marketing blitz started.

"More money was spent on marketing Vioxx than on Pepsi and Budweiser," said Dr. John Abramson, a physician who wrote "Overdosed America: How the Pharmaceutical Companies Distort Medical Knowledge, Mislead Doctors and Compromise Your Health" (HarperCollins, $24.95). "The marketing trumped the facts."

Fast forward to Sept. 30 of this year, when Merck announced it was pulling Vioxx off the market, after a new study found patients had double the number of cardiovascular events after 18 months on the drug.

The new study wasn't carried out to check on the long-term cardiovascular safety of Vioxx; Merck was trying to prove Vioxx's usefulness for another condition, preventing recurrence of colorectal polyps. But the researchers ran headlong into the cardiovascular complications that they had explained away in 2000, when Vioxx was compared to naproxen, an older drug of the same class as ibuprofen.

In his book, Abramson points out that the VIGOR research plan acknowledged that patients taking naproxen might have fewer cardiovascular complications and blood clots than those on Vioxx, a selective COX-2 inhibitor.

The COX-2 group of drugs are thought to be easier on the stomach. But the ibuprofen/naproxen group also inhibits a second enzyme, known as COX-1, which makes platelets less "sticky." That action may decrease the chance of a bad blood clot - of particular concern for rheumatoid arthritis patients, who are at risk for heart disease anyway.

But the difference in heart attacks between naproxen and Vioxx was fivefold. Yet no alarms were sounded.

The authors of the VIGOR study put forth several explanations. First, they argued, the difference was an indication of naproxen's cardio-protective effect rather than of any risks associated with Vioxx. But no clinical trials had ever shown that naproxen curtailed heart attacks.

The authors made another argument too: They said the patients who suffered heart attacks were at high risk for heart disease to begin with.

The clinical trial guidelines had barred participants with heart disease. So the authors went back after the trial ended and re-classified patients, saying they had found some who actually had pre-existing disease factors.

And while the VIGOR paper disclosed the heart attacks, it left out other complications such as stroke, unstable angina and venous blood clots, saying those would be analyzed in a separate paper.

Dr. Sidney Wolfe, from the Public Citizen Health Research Group consumer organization, was convinced Vioxx was the cause of the heart attacks, and he placed the drug on his Do-Not-Use list of prescription drugs (see www.worstpills.org).

Analysts at the FDA were also troubled, according to briefing documents for a meeting on Feb. 8, 2001, that are still available on the FDA's Web site.

More than 4,000 patients had been in the group taking Vioxx and a similar number were taking naproxen. The FDA had data showing that 61 Vioxx patients had dropped out of the VIGOR trial because of adverse cardiovascular events, compared to 21 on naproxen; that 10 Vioxx patients dropped out because of cerebrovascular accidents, compared to three on naproxen; that 12 Vioxx patients dropped out because of heart attacks, compared to three on naproxen. Overall, 20 on Vioxx had heart attacks - compared to four on naproxen.

"This analysis could lead one to conclude that naproxen, with a 51 percent risk reduction compared to rofecoxib [Vioxx], would be the preferred drug," notes a Feb. 1, 2001, memo from Dr. Shari L. Targum, a medical officer at the FDA, still posted on the Web. Another FDA memo questioned the authors' retrospective re-classification of patients' risk, saying the "clinical judgment at the time of enrollment appears to be more relevant."

Merck officials said in a statement last week that the VIGOR study "was not designed to evaluate differences in cardiovascular events," even though data were collected about adverse cardiovascular events.

FDA officials said the agency had added a warning about cardiovascular risks on Vioxx's label in 2002, and had required monitoring of cardiovascular events in the new study.

One could argue the drug was worth keeping on the market because it caused less GI bleeding. But FDA analysts had their doubts on this too.

Three patients on Vioxx died of GI complications during the VIGOR study; only one on naproxen died. That may have been by chance. But from May 1999 to October 2000, according to the 2001 memos, the FDA received reports that 37 people died while on Vioxx - of GI complications.

© 2004 All Rights Reserved