Pfizer acknowledges that patients taking Celebrex in the trial involving 425 patients--performed to assess the drug as a treatment for Alzheimer's disease--suffered heart attacks at nearly four times the rate of patients taking placebos.

The company counters, however, that information on the study has been public since 2000. Pfizer also claims that it is not possible to draw clinically meaningful conclusions from the study alone, given variations in patients' medical history. As recently as last October, Pfizer claimed that no completed study of Celebrex had pointed to increased heart risks.

The confrontation comes two weeks before an FDA conference on COX-2 inhibitors, a class of drugs that has been at the center of calls for more public information from clinical trials on drug safety.

Last month, Public Citizen petitioned FDA to ban COX-2 inhibitors on the basis of clinical data that had come to light indicating health risks for several drugs in the category. Sydney M. Wolfe, director of Public Citizen's health research group, says information recently posted on an industry-sponsored website about Pfizer's Celebrex trial supports his group's contention that the drugs are unsafe and that drug companies have suppressed results.

"Public Citizen's cherry-picking of data is alarmist and misconceived," says Joseph M. Feczko, president of worldwide development at Pfizer. Feczko adds that an independent panel monitored the study and that an abstract was presented at a scientific congress in Stockholm in 2000. Wolfe tells C&EN that the conference abstract was a "whitewash" that did not include the negative results that Pfizer posted on the website.

Observers say the confrontation illustrates the potential for confusion as the public gains access to difficult-to-interpret data.

According to Michael A. Valentino, chief consultant for pharmacy benefits management at Veterans Health Administration (VHA), interpreting clinical trial data is a problem even for medical researchers. "For every study that shows harm, you can find two that don't. And vice versa," Valentino says. "You really need to dig a lot deeper."

VHA looks to FDA for guidance, he says. "But we don't stop there. We do our own evaluation and critiques of published studies and, more importantly, we look at experience in our own patient population."

Valentino says VHA is still prescribing and dispensing Celebrex. Meanwhile, managed care giant Kaiser Permanente says it will stop dispensing Pfizer's other COX-2 drug, Bextra (valdecoxib), pending further evaluation of the drug's safety.