Merck Pulled Vioxx From the Market in 2004 After Linking It to Tens of Thousands of Heart Attacks. The FDA Didn't Formally Revoke Its Approval Until 2022.

On May 20, 1999, the FDA approved Merck's application to sell rofecoxib under the brand name Vioxx, an arthritis and pain pill from the new COX-2 inhibitor class. Merck's own 1999 annual report put it on the cover: Vioxx was "the biggest, fastest, and best launch ever." Within five years the company would pull it from every pharmacy shelf in the world, an FDA safety officer would tell Congress it had likely caused tens of thousands of American deaths, and Merck would eventually pay out billions to settle the lawsuits — while its actual FDA marketing approval sat on the books, technically valid, for another eighteen years.

None of the cardiovascular risk that ended Vioxx was undiscovered science by the time the drug came off the market. It showed up in a company-funded clinical trial in March 2000, in a form clear enough that Merck's own head of research emailed colleagues the same day to say the signal was "clearly there." What followed was four and a half years of a drug staying on the market, a sales force trained to change the subject, and a regulator whose own scientists say it wasn't built to stop any of it.

The Biggest, Fastest, Best Launch Ever

Vioxx was the second COX-2 inhibitor to reach the U.S. market, three months behind Pfizer's Celebrex. Merck made up the gap with money. The company fielded roughly 3,000 to 4,500 sales representatives to market the drug to physicians and spent $67 million on advertising in the first four months of 2000 alone — more than any pharmaceutical company had spent promoting a single drug up to that point. The bet paid off: Vioxx was prescribed to more than 20 million Americans, generated over $2.5 billion in annual sales at its peak, and was eventually taken by an estimated 80 million patients worldwide before it disappeared from shelves.

• FDA approval: May 20, 1999, for osteoarthritis, acute pain, and menstrual pain
• First-year marketing spend: $67 million in advertising in four months, a record at the time
• Peak annual sales: more than $2.5 billion

The Study That Found the Problem It Was Looking For

The VIGOR trial — Vioxx Gastrointestinal Outcomes Research — was designed to prove Vioxx caused fewer stomach ulcers and GI bleeds than naproxen, an older painkiller. It did. It also found something the trial wasn't built to explain away: patients on Vioxx suffered heart attacks at roughly five times the rate of patients on naproxen, 79 serious cardiovascular events in the Vioxx group versus far fewer on naproxen. The study's authors, writing in the New England Journal of Medicine in November 2000, attributed the gap to naproxen having a protective effect on the heart rather than to Vioxx causing harm.

Merck's head of research, Edward Scolnick, did not need the published paper's framing to see the problem. In an internal email sent within hours of the VIGOR results in March 2000, he wrote that the "CV events were clearly there," and separately told a colleague in April 2000 that he was in "minor agony" over fears Vioxx was causing heart attacks and strokes. In 2005, the NEJM's editors issued a formal "Expression of Concern," stating that at least two of the VIGOR paper's authors had known — more than four months before publication — about three additional heart attacks in the Vioxx group that had been excluded from the submitted data using an earlier cutoff date for cardiovascular events than for the gastrointestinal events the trial was designed to measure.

A Sales Training Game Called "Dodge Ball"

Merck did not update Vioxx's label to reflect the VIGOR cardiovascular data until April 2002, roughly a year and a half after the results were known internally. In the meantime, the company trained its sales force on how to handle doctors who had read the same study. An internal 16-page training document, later made public and nicknamed "Dodge Ball Vioxx," walked reps through physician "obstacles" — questions like "I am concerned about the cardiovascular effects of Vioxx" — and scripted the correct response as literally: "DODGE!" One approved line for reps facing a doctor asking about the heart attack data: "Doctor…because the study is not on the label, I cannot discuss the details with you."

Reps were also issued a "Cardiovascular Card" describing Vioxx's effect on the heart in favorable terms rather than presenting the VIGOR signal, according to internal documents released by the House Committee on Government Reform. On November 8, 2001, Scolnick wrote to colleagues in another internal email: "I will NOT sign off on any cardiac warning" for the label.

• Vioxx cardiovascular warning added to the label: April 2002, roughly 18 months after Merck's own research chief called the VIGOR heart attack signal "clearly there"
• Sales force size trained on the "Dodge Ball" script and Cardiovascular Card: approximately 3,000 representatives

Eighteen Months Into the APPROVe Trial

The drug's end came from a study that wasn't even about heart disease. Merck was running the APPROVe trial to see whether Vioxx could prevent colorectal polyps from recurring. On September 27, 2004, the company told the FDA it was halting APPROVe early: patients taking Vioxx for longer than 18 months were showing a statistically significant increase in confirmed cardiovascular events — heart attack and stroke — compared to placebo. Three days later, on September 30, 2004, Merck announced a voluntary worldwide withdrawal of Vioxx from the market.

By the time of the withdrawal, an estimated 80 million people had taken the drug worldwide since its 1999 launch.

Two to Four Jumbo Jets a Week

On November 18, 2004, David Graham, associate director for science in the FDA's Office of Drug Safety, testified before the Senate Finance Committee that the agency he worked for was "incapable of protecting America against another Vioxx." He described the FDA's Center for Drug Evaluation and Research as "broken," and pointed to what he called an inherent structural conflict: the same office that approves a drug is also the office responsible for later taking regulatory action against it.

Graham told the committee that a conservative estimate put the toll at nearly 28,000 excess cases of heart attack or sudden cardiac death attributable to Vioxx, and that a broader analysis put the range at 88,000 to 139,000 excess heart attacks and strokes, of which 30 to 40 percent were likely fatal. He translated the estimate into an image for the senators: at 150 to 200 people per flight, a toll in that range was "the rough equivalent of 500 to 900 aircraft dropping from the sky," or two to four jumbo jetliners crashing every week for five years. Graham later said FDA superiors had pressured him ahead of the hearing not to testify and dismissed his findings as "junk science"; in January 2005, his peer-reviewed analysis, estimating 88,000 to 140,000 excess cases of serious coronary heart disease in the U.S. over Vioxx's five years on the market, was published in The Lancet.

27,000 Lawsuits, One Guilty Plea

Merck was hit with tens of thousands of individual lawsuits from patients and families alleging heart attacks and strokes caused by Vioxx. On November 9, 2007, the company announced a $4.85 billion settlement resolving roughly 27,000 claims — split into a $4 billion fund for heart attack claims and an $850 million fund for ischemic stroke claims — without admitting fault. Kenneth Frazier, then Merck's general counsel, told reporters on the announcement call that "without this settlement, the litigation might very well stretch on for years."

The lawsuits weren't the end of it. In November 2011, Merck's subsidiary Merck Sharp & Dohme pleaded guilty to a single misdemeanor criminal charge — introducing a misbranded drug into interstate commerce — tied to its promotion of Vioxx for rheumatoid arthritis between 1999 and 2002, before the FDA had approved it for that use. The company paid a $321.6 million criminal fine and a further $628.4 million to settle related civil claims, including allegations it made false statements about Vioxx's cardiovascular safety — roughly $950 million combined, on top of the $4.85 billion already paid to injured patients.

What the FDA Changed

Congress used the 2007 renewal of FDA user-fee legislation, the Food and Drug Administration Amendments Act, to give the agency new post-market authority it had lacked throughout the Vioxx years: the power to require additional safety studies after approval, order label changes unilaterally, and impose formal Risk Evaluation and Mitigation Strategies on high-risk drugs. In May 2008, the FDA launched the Sentinel Initiative, an electronic surveillance system built to scan real-world insurance claims and health records for safety signals across more than 100 million patient lives — the kind of active, systematic post-market monitoring that, had it existed in 2000, might have caught what Edward Scolnick's inbox already knew.

Merck's own request to formally withdraw its Vioxx new drug applications did not reach the FDA until October 7, 2021, and the agency did not act on it until September 13, 2022. For almost eighteen years after Merck pulled Vioxx from every pharmacy shelf on earth, citing an elevated risk of heart attack and stroke, the drug remained, on paper, an FDA-approved medicine.

Filed from public reporting:

• FDA — "FDAAA Implementation: Highlights One Year After Enactment"
• Federal Register — "Merck Sharp & Dohme Corp.; Withdrawal of Approval of New Drug Applications for VIOXX (Rofecoxib) Tablets and Suspension"
• New England Journal of Medicine — "Expression of Concern: Bombardier et al., 'Comparison of Upper Gastrointestinal Toxicity of Rofecoxib and Naproxen in Patients with Rheumatoid Arthritis'"
• New England Journal of Medicine — "Failing the Public Health — Rofecoxib, Merck, and the FDA"
• NPR — "Timeline: The Rise and Fall of Vioxx"
• Senate Finance Committee — Testimony of David J. Graham, MD, MPH (November 18, 2004)
• Nature — coverage of Graham et al., The Lancet, "Risk of acute myocardial infarction and sudden cardiac death in patients treated with cyclo-oxygenase 2 selective and non-selective non-steroidal anti-inflammatory drugs"
• NPR — "Documents Suggest Merck Tried to Censor Vioxx Critics"
• CBS News — "Merck Knew Vioxx Dangers In 2000"
• NPR — "Merck CEO Gilmartin Steps Down"
• CNN Money — "Merck agrees to pay $4.85B to settle Vioxx lawsuits"
• U.S. Department of Justice — "U.S. Pharmaceutical Company Merck Sharp & Dohme Sentenced in Connection with Unlawful Promotion of Vioxx"