A previously unpublished study by Pfizer Inc. of its arthritis drug Celebrex found more than four years ago that users had a potentially significant increase in risk of heart attacks and strokes, casting doubt on the company's insistence that it had no hint of the problem until recently.
The study of 425 Alzheimer's patients was completed in 2000 but was made accessible to the public only recently on a drug industry Web site. When posted last month, the report of the clinical trial did not initially indicate there was evidence of an increased cardiovascular risk from the drug. But an updated account of the study posted last week acknowledged that there was a statistically significant difference in the number of heart-related problems between users of Celebrex and of the placebo.
The Celebrex study was not designed to look into heart and stroke problems.
Audio: The Post's Marc Kaufman discusses a 2000 medical study that found an increased risk of heart attack linked to the use of Celebrex.
Women Still Treated Less for Heart Disease (The Washington Post, Feb 2, 2005)
For RLS Patients, a Challenging Quest for Relief (The Washington Post, Feb 1, 2005)
Anti-Inflammatory Eating (The Washington Post, Jan 26, 2005)
New Study Criticizes Painkiller Marketing (The Washington Post, Jan 25, 2005)
But First, A Word From Your Doctor (The Washington Post, Jan 25, 2005)
More Heart News
The results, highlighted yesterday by the consumer group Public Citizen, raise the possibility that Pfizer was more aware of the cardiovascular risks of its drug than it has acknowledged. After a federally sponsored trial of Celebrex was stopped in December because of concerns about an increase in heart attacks and strokes, company chairman and chief executive Henry A. McKinnell said that the drug would stay on the market and that "these clinical trial results are new."
Two months earlier, Merck & Co. withdrew its arthritis painkiller Vioxx from the market after finding a significant risk of heart and stroke problems in patients who used the drug more than 18 months. Both Celebrex and Vioxx are COX-2 inhibitors, a class of drugs hailed in the late 1990s as a major advance for arthritis sufferers. But because of the recent worries about increased heart and stroke risk, the Food and Drug Administration has scheduled an unusual three-day expert advisory meeting for later this month to discuss the fate of the entire class.
After the Vioxx withdrawal, Pfizer continued to advertise Celebrex heavily to the public until the FDA pressured it to stop.
Pfizer dismissed the results of the Alzheimer's study yesterday, saying that it was too small to be meaningful and that it was unreliable because those in the Celebrex arm of the trial were found to have had considerably more cardiovascular risk factors to begin with than those in the placebo group. In a statement, the company also said that it had shared the study results with the FDA in June 2001 and that an independent panel of safety experts had monitored the 52-week study for any significant health problems.
But Sidney M. Wolfe of Public Citizen, which has petitioned the FDA to take all COX-2 drugs off the market, called the study a "smoking gun." He said the company and the FDA should have made the results broadly known to the public long ago.
"This was the first randomized and placebo-controlled study to show these cardiovascular problems as statistically significant, and Pfizer had a responsibility to inform the public about it," Wolfe said. "If Pfizer sent this information to the FDA like they say and the agency sat on it, then the FDA has been incredibly irresponsible as well."
The Alzheimer's study was initiated by Pfizer to see whether the drug could help slow the onset of the brain disease. It was started before Celebrex was approved as an arthritis drug and was not designed to look into the then-theoretical question of whether COX-2 inhibitors might cause heart and stroke problems.
Patients who met the criteria for the study were placed randomly into either the drug or the placebo group. The Pfizer statement said that despite the randomizing, the Celebrex group ended up with a disproportionately large number of people with preexisting risks for heart attacks and strokes, making the results unreliable.
Wolfe, however, said his analysis showed convincingly that the group taking Celebrex was 3.6 times as likely to experience heart problems or a stroke as those taking the placebo. In reaching that conclusion, Wolfe combined all the study's heart- and stroke-related adverse events into one category, rather than leaving them in distinct sub-categories such as cardiac failure, angina, cardiac arrhythmia and stroke -- as the study did. That grouping, Wolfe said, made the adverse events statistically significant -- showing 22 events in the Celebrex group and three in the placebo group -- while the company's analysis did not.
Wolfe sent his assessment for review to Bruce Psaty, co-director of the University of Washington cardiovascular research center and an expert on cardiac drug safety. Psaty, who testified at last fall's high-profile Senate hearing on the Vioxx withdrawal and FDA drug safety procedures, said in an interview that Pfizer appeared to have presented its cardiovascular data in a way to minimize the evidence of risks. In addition, he said that the company should have published its results much earlier.
"This study was completed in 2000 but we're just learning about it in 2005, and that's just unacceptable," Psaty said. "How can physicians be expected to responsibly prescribe drugs to their patients if we don't know the true risks and benefits?"
Although the Celebrex Alzheimer's study was finished in late 2000, it was not available for an FDA expert panel that met in early 2001 to assess the cardiovascular risks of COX-2 inhibitors. In its statement, Pfizer said the study results were sent to the FDA in the summer of 2001.
The company also said that the results of the study were presented in a scientific meeting in 2000. But it was unclear whether safety issues were discussed or only the question of whether the drug was effective in treating Alzheimer's -- the study said it was not. The company did not respond to requests for additional information.