A research director for Pfizer was positively buoyant after reading that an important medical conference had just featured a study claiming that the new arthritis drug Celebrex was safer on the stomach than more established drugs.
“They swallowed our story, hook, line and sinker,” he wrote in an email to a colleague.
The truth was that Celebrex was no better at protecting the stomach from serious complications than other drugs. It appeared that way only because Pfizer and its partner, Pharmacia, presented the results from the first six months of a yearlong study rather than the whole thing.
The companies had a lot riding on the outcome of the study, given that Celebrex’s effect on the stomach was its principal selling point. Earlier studies had shown it was no better at relieving pain than common drugs — like ibuprofen — already on the market.
The research chief’s email, sent in 2000, is among thousands of pages of internal documents and depositions unsealed recently by a federal judge in a long-running securities fraud case against Pfizer. While the companies’ handling of the research was revealed a dozen years ago, the documents provide a vivid picture of the calculation made by Pfizer at the time and its efforts ever since to overcome doubts about the drug.
The documents suggest that officials made a strategic decision during the early trial to be less than forthcoming about the drug’s safety. They show that executives considered attacking the trial’s design before they even knew the results and disregarded the advice of an employee and an outside consultant who had argued the companies should disclose the fact that they were using incomplete data.
In one email, an associate medical director at Pharmacia (which was later bought by Pfizer) disparaged the way the study was being presented as “data massage,” for “no other reason than it happens to look better.”
In another, a medical director at Pfizer described it as “cherry-picking the data” even as officials were publicly boasting of the study’s success. Dr. M. Michael Wolfe, a gastroenterologist who had cautiously praised the study in a medical journal at the outset, said after reviewing the new documents: “I always try to give investigators the benefit of the doubt, but these communications make it quite challenging for me.”
The importance of Celebrex to Pfizer is indisputable. It is one of the company’s best-selling drugs, racking up more than $2.5 billion in sales, and was prescribed to 2.4 million patients in the United States last year alone.
The drug is the last of the so-called COX-2 inhibitor pain drugs, after Vioxx and Bextra were withdrawn in 2004 and 2005 because of safety concerns.
Some of the Celebrex’s detractors contend that its risks are still not fully understood, and argue that Pfizer is dragging its feet on a study — now nearly six years old — evaluating the drug’s heart risks. The study is scheduled to end in May 2014, the same month that Celebrex loses its patent protection and sales of the drug are expected to plunge.
Then and now, Pfizer has defended its decision to release partial results from the 2000 study and denies any intent to deceive. Company officials have said the drug has demonstrated its worth and safety. The proof, they say, is that 33 million Americans have taken it. “The bottom line is Celebrex is a very important option for many of these patients,” said Dr. Steve Romano, head of the medicines development group in Pfizer’s primary care unit.
The decision by Pfizer and Pharmacia to withhold crucial data became widely known in 2001, after the Food and Drug Administration released the study’s full results. The revelations, along with similar reports of withheld data by other drug companies, led to calls for reforms in the way data from clinical trials is published, including in The Journal of the American Medical Association, which ran an article featuring the partial results from the study.
The withheld data also led to a lawsuit, filed in 2003, by several pension funds that charged that by handling the results the way they did, Pfizer and Pharmacia had misled investors and were responsible for a drop in Pharmacia’s stock value when the full results were revealed.
Lawyers for Pfizer and for the pension funds declined to comment. In a statement, company officials said they were confident they would prevail when all the evidence was heard. “The few documents handpicked by lawyers suing Pfizer and being reported by The New York Times are not a fair representation of this body of evidence,” the company said.
The documents show that in February 2000, Pharmacia employees came up with a game plan on how they might present the findings once they were available.
“Worse case: We have to attack the trial design if we do not see the results we want,” a memo read. It went on: “If other endpoints do not deliver, we will also need to strategize on how we provide the data.”
Another document, a slide, proposed explaining poor results through “statistical glitches.”
Pfizer officials said the memo appears to reflect discussions by some Pharmacia employees about both the Celebrex study and a similar study of Vioxx. The slides, the company said, appeared to have been prepared before the results were known and discussed several situations.
While officials were boasting of the study’s success, employees behind the scenes were questioning its value. In September 2000, Dr. Emilio Arbe, a Pharmacia associate medical director, expressed his reservations.
After describing the decision to use the limited results as “data massage,” Dr. Arbe wrote, “I wouldn’t feel too comfortable presenting a fudged version of the facts.”
In May 2001, Dr. Mona Wahba, who worked on Celebrex, sent an email to colleagues describing as “cherry-picking” a new analysis that also used six months of the results. Pfizer officials said Dr. Wahba’s email was sent after the full study became known. In a deposition, Dr. Wahba said she did not recall what she meant.
Dr. Samuel Zwillich, who wrote the “hook, line and sinker” email, testified in another deposition that his comment probably had to do with his concerns around a lesser claim that Celebrex led to less blood loss than other drugs. Through a Pfizer spokesman, he declined to comment.
Pfizer has argued that presenting the limited data was legitimate because so many people taking a comparison drug, diclofenac, dropped out, biasing the later results.
The controversy over the safety and effectiveness of Celebrex continues today. Celebrex and Vioxx, which was made by Merck, brought in billions in sales almost as soon as they were introduced in the 1990s.
But the excitement skidded to a halt in 2004, when Merck withdrew Vioxx after studies linked it to an increased risk for heart attacks. Some studies indicated that Celebrex, too, carried elevated risks. In part to address those concerns, Pfizer announced in 2005 that it was starting a trial that would compare the heart risk for Celebrex with ibuprofen, the drug in Advil and Motrin, and naproxen, which is sold as Aleve. The trial is not expected to be finished until 2014 when the Celebrex patent expires.
Dr. Steven Nissen, the Cleveland Clinic cardiologist who is overseeing the trial, said Pfizer has spent hundreds of millions of dollars and enrolled 18,000 patients. Recruiting has been difficult, he said, in part because European Union countries have barred patients with heart risks.
Dr. Nissen dismissed claims from critics that Pfizer has been delaying the trial out of fears about its outcome. “The last thing in the world I want to do is to be sitting here twiddling my thumbs with a public health concern,” he said.
Others were not so sure.
“One could draw conclusions,” Dr. Alastair J. J. Wood, who was chairman of the F.D.A. advisory panel that examined COX-2 inhibitors, said recently. He is a partner at Symphony Capital, which invests in drug development. “It clearly would have been nice to have had this information long ago.”
Despite its success, Celebrex’s place among pain drugs is not settled. According to Pfizer, 93 percent of insured patients have access to Celebrex.
Some doctors said Celebrex has advantages because it can be taken once a day and studies have shown that it causes less stomach discomfort than other drugs, although some have argued those types of findings are not always reliable.
“You’re dealing with softer, subjective endpoints,” said Garret FitzGerald, chairman of pharmacology at the University of Pennsylvania.
There is still no clinical proof that Celebrex is better at preventing serious gastrointestinal injuries.
Dr. David Borenstein, who conducted some trials of Celebrex for Pharmacia, said the drug relieved pain in some patients when other drugs failed. “It’s easy to talk about theoreticals when you’re not hurting,” he said.