Pfizer Says Internal Studies Show No Celebrex Risks

By GARDINER HARRIS and ALEX BERENSON

Published: February 5, 2005

Pfizer has told federal drug officials that the company's analysis of 41 of its own patient trials of the arthritis and pain drug Celebrex finds no evidence that the drug increases the risk of heart problems.

But in an interview, a Pfizer official acknowledged that the analysis did not include outside studies that indicated the drug raised the risk of heart problems.

The company's written analysis is part of a welter of documents that the Food and Drug Administration released yesterday for public review ahead of hearings beginning Feb. 16. At the hearings, the agency will assess the safety records of Celebrex and Bextra, made by Pfizer; Vioxx from Merck; and other drugs in the class known as COX-2 inhibitors.

Various studies have raised questions about the cardiac safety of those drugs, and Merck took Vioxx off the market last fall in response to evidence of the drug's dangers.

The Pfizer analysis covered tests involving 44,000 patients but did not include results from studies conducted by independent researchers or those at other drug makers.

Dr. Joseph Feczko, Pfizer's president for worldwide development, acknowledged in an interview yesterday that a few studies had found increased heart risks for patients taking Celebrex and Bextra.

A 1999 trial in Alzheimer's disease patients that Pfizer sponsored but never published found an increased risk of heart problems in those taking Celebrex. And the National Institutes of Health ended a study in December after finding that high doses of Celebrex more than tripled the incidence of heart attacks and strokes in patients.

"They are what they are," Dr. Feczko said of those two trials. But "the overwhelming preponderance of evidence" indicates the drugs are safe, he said.

Celebrex had $3.3 billion in sales last year, making it one of the biggest-selling drugs in the world. Bextra had $1.3 billion in sales. Sales of both drugs have declined sharply since December.

Over all, Pfizer has performed much less research on Bextra than on Celebrex, Dr. Feczko said. Most of the company's studies of Bextra have been short term, with many lasting only two weeks. As a result, Pfizer has less data to support its contention that Bextra is safe, he said.

Offering an unusual glimpse into the deliberations of the F.D.A., the documents show that in 2000 agency officials largely agreed that Celebrex appeared safe, but that in 2001 they found reasons to worry about the safety of Bextra.

The documents also show that agency officials disagreed forcefully with Merck officials in 2001 and 2002 as the sides wrestled over label language that would balance the benefits and risks of Vioxx.

During that period, some meetings were canceled as "substantial distance" separated the positions of Merck and F.D.A. officials, the documents show. The sides eventually agreed to revise Vioxx's label to indicate that the drug had been shown to increase the risk of heart problems.

The documents also show that the F.D.A. has concluded that Arcoxia, Merck's experimental drug that it had hoped would be approved for sale last year by the F.D.A., harms the heart and is not particularly effective at protecting against ulcers. Merck announced last year that the F.D.A. needed more information before approving the drug.

An F.D.A. summary of all of the documents stated that many questions about the safety of COX-2 drugs remain. "Based upon all of the information currently available, many in the public, the scientific community, and F.D.A. have raised questions about whether there should be continued marketing of COX-2" pain pills, the summary said.

The documents detail serious concerns that F.D.A. reviewers had about Bextra and a related drug called paracoxib before they approved Bextra for sale in 2002. Those concerns are among the reasons that the agency never approved paracoxib for sale and refused to approve Bextra as a treatment for acute pain.

In November 2001, Dr. Kent Johnson, an F.D.A. medical officer who was reviewing the Bextra application, noted that one study of Bextra and paracoxib found that the drugs were linked to high rates of heart attacks and strokes in patients who had had bypass surgery. In addition, an earlier study found that high doses of Bextra raised blood pressure, possibly increasing heart risk.

"There may be a component of increased thrombogenicity associated with this agent," Dr. Johnson wrote. Thrombogenicity means a tendency to produce blood clots, which can lead to heart attacks or strokes.

Still, Dr. Johnson recommended that low doses of Bextra be approved for the treatment of arthritis, since arthritis studies had not shown increased risks from Bextra.

In its report to the F.D.A, Pfizer draws distinctions between Celebrex and Bextra, arguing that Celebrex probably increases blood pressure less than drugs like Advil, while Bextra probably has very similar effects.

Dr. Feczko of Pfizer explained that the company felt it was not as important to study Bextra extensively because the company believed that the drug was similar to Celebrex.