Pfizer Says 1999 Trials Revealed Risks With Celebrex

By ALEX BERENSON and GARDINER HARRIS

Published: February 1, 2005

Celebrex, the popular arthritis and pain medicine from Pfizer, sustained another blow yesterday when the company acknowledged that a 1999 clinical trial found that elderly patients taking the drug were far more likely to suffer heart problems than patients taking a placebo.

The company sought to play down the trial's significance yesterday, saying that it was flawed. But the disclosure of the study contradicts earlier public statements by Pfizer, the world's largest drug maker, about the safety of Celebrex. Concerns about the risks of Celebrex have risen since last fall, when Merck withdrew a similar drug, Vioxx, after a trial linked it to heart attacks.

Responding to the Vioxx withdrawal, Pfizer said in October that no completed study had ever shown any increased heart risks related to Celebrex. Then, in December, when a colon cancer prevention study by the company indicated that Celebrex was associated with heart attacks and strokes, Pfizer said the results were unexpected.

Now the company has acknowledged that the 1999 study, which was intended to examine whether Celebrex could treat Alzheimer's disease, found that the number of Celebrex patients suffering heart attacks was almost four times that of those taking a placebo. Pfizer's own analysis found the difference statistically significant.

But the study was never published and not submitted to the Food and Drug Administration until June 2001, four months after the F.D.A. conducted a major review of the safety of Vioxx and Celebrex. Two doctors who participated in that review said they had not known about the 1999 study until yesterday. One of them, Dr. Kenneth Brandt, a professor of medicine at Indiana University School of Medicine, said that if the safety panel had known about the study, it might have recommended that both Vioxx and Celebrex be taken with greater caution. That panel decided to recommend that Vioxx, but not Celebrex, carry a warning about its cardiovascular risks. That difference is one of the main reasons Celebrex had greater sales than Vioxx.

The 1999 study is further evidence that Celebrex is dangerous, said Sidney M. Wolfe, a director at Public Citizen, a consumer-advocacy group that has asked the F.D.A. to ban the medicine.

Dr. Wolfe publicized the 1999 study yesterday, after finding it last week on a new Web site where Pfizer and other drug companies have begun to post some clinical trial results. Dr. Wolfe said the results had not been on the site a few weeks earlier.

"It's a clear signal that I would have loved to have known about four years ago," he said.

But Dr. Gail Cawkwell, Pfizer's medical team leader for Celebrex, said yesterday that the study's importance should not be overstated. Many other trials have shown that Celebrex is safe, and that the medicine is an important treatment for arthritis patients, she said.

Celebrex is one of the world's most widely prescribed medicines, with sales of $3.3 billion last year. About 25 million people have taken the drug since Pfizer introduced it in 1999, and analysts estimate that Celebrex contributes at least $2 billion to Pfizer's pretax annual profits. But its sales have slowed since the Vioxx withdrawal last fall raised safety questions about the entire class of drugs, which are known as COX-2 inhibitors and include another Pfizer medication, Bextra.

After the 1999 study was disclosed yesterday, Pfizer's shares fell 19 cents, to $24.16, their lowest level since 1997. The shares have fallen 21 percent since Merck withdrew Vioxx.

Pfizer, whose chief executive is Henry A. McKinnell Jr., has strongly defended Celebrex's safety, calling the drug an important alternative for arthritis patients. The company maintained that stance yesterday, saying that the 1999 trial was flawed and too small to be meaningful.

Dr. Cawkwell said yesterday that the 1999 study that showed Celebrex was ineffective in treating Alzheimer's disease had been presented at a conference in Sweden in 2000. But she said she did not know whether the study's safety data had been presented.

At the time, Celebrex was manufactured by Pharmacia, not Pfizer, although Pfizer marketed the drug. Pfizer bought Pharmacia for $60 billion in late 2002, largely to gain complete control of Celebrex and Bextra, another arthritis drug.

A spokesman for Fred Hassan, Pharmacia's former chief executive, declined to comment. Philip Needleman, one of the discoverers of Celebrex, did not return a phone message.

Dr. Cawkwell said she did not know why Pharmacia did not send the results to the F.D.A. in time for the broader review. After Pfizer bought Pharmacia, it assured patients and doctors that it had no evidence that Celebrex was dangerous.

On Oct. 1, after Merck's decision to withdraw Vioxx, Pfizer said in a press release, "The evidence distinguishing the cardiovascular safety of Celebrex has accumulated over years in multiple completed studies, none of which has shown any increased cardiovascular risk for Celebrex."

In fact, the 1999 Alzheimer's study found that 22 out of 285 patients taking Celebrex suffered heart attacks, strokes and other heart problems. Only 3 of 140 patients taking a placebo had similar problems. Even accounting for the difference in the sizes of the two groups, Celebrex users were almost four times as likely to suffer heart problems.

In the study, patients took 400 milligrams of Celebrex daily, a dose commonly used by arthritis patients. "A statistically significant difference favoring placebo in adverse events was observed," the study's authors explained.

Dr. Cawkwell said the difference might have resulted from the fact that the Celebrex patients in this trial happened to be sicker than patients given the placebo. That difference was unintentional and due solely to random chance, she said. "The patients had very different cardiovascular histories," she said.