When editors of the Journal of the American Medical Association sent medical expert M. Michael Wolfe an unpublished study on the blockbuster arthritis drug Celebrex last summer, he was impressed by what he read.
Tested for six months in a company-sponsored study involving more than 8,000 patients, the drug was associated with lower rates of stomach and intestinal ulcers and their complications than two older arthritis medicines diclofenac and ibuprofen.
JAMA’s editors wanted to rush the findings into print, and Wolfe and a colleague provided a cautiously favorable editorial to accompany it. But in February, when Wolfe was shown the complete data from the same study as a member of the Food and Drug Administration’s arthritis advisory committee, he said he saw a different picture.
“We were flabbergasted,” he said.
The study already completed at the time he wrote the editorial had lasted a year, not six months as he had thought, Wolfe learned. Almost all of the ulcer complications that occurred during the second half of the study were in Celebrex users. When all of the data were considered, most of Celebrex’s apparent safety advantage disappeared.
“I am furious, I wrote the editorial. I looked like a fool,” said Wolfe, a Boston University gastroenterologist. “But all I had available to me was the data presented in the article.”
JAMA’s editor, Catherine D. DeAngelis, said the journal’s editors were not informed about the missing data. “I am disheartened to hear that they had those data at the time that they submitted [the manuscript] to us,” she said. “We are functioning on a level of trust that was, perhaps, broken.”
The study’s 16 authors included faculty members of eight medical schools. All authors were either employees of Pharmacia, Celebrex’s manufacturer, or paid consultants of the company. For company-sponsored studies, JAMA now requires a statement, signed by an author who is not employed by the company, taking “responsibility for the integrity of the data and the accuracy of the data analyses,” DeAngelis added.
Steven Geis, a vice president for clinical research of Pharmacia and one of the authors, said that only the first six months of data were presented because, after that, more patients withdrew from the comparison groups than from the Celebrex group, biasing later findings. He said a three-member executive committee, composed of authors who were not Pharmacia employees, approved the decision.
“The intention really was not to be deceptive in any way,” he said. “People thought that six months was the appropriate analysis.”
With inclusion of the later data, “the actual difference between Celebrex and [the other drugs] are not as wide as they were at six months,” he acknowledged. “But I think in the end, it does show that Celebrex has a superior safety profile.”
After reviewing the full study, the FDA’s arthritis advisory committee concluded that Celebrex offers no proven safety advantage over the two older drugs in reducing the risk of ulcer complications, said FDA spokesman Susan Cruzan. The company has requested a change in the drug’s labeling to state that it is indeed safer, but the FDA has asked for additional information before making a decision.
Meanwhile, the JAMA article and editorial have likely contributed to Celebrex’s huge sales. “When the JAMA article comes out and confirms the hype, that probably has more impact than our labeling does,” said Robert J. Temple, director of medical policy at the FDA’s Center for Drug Evaluation and Research.
James Wright, a professor of clinical pharmacology at the University of British Columbia, said he complained to JAMA after noticing differences between the published report and the data presented to the FDA. He praised the Public Citizen’s Health Research Group, a consumer organization, for filing a lawsuit that led to the agency’s putting all drug studies presented to its advisory committees on its public Web site.
“Otherwise, we still wouldn’t know this,” Wright said. “We would still be in the dark.”