(The New York Times News Service) - The prescription pain relievers Vioxx and Celebrex surged onto the scene two years ago with an unprecedented marketing effort by their makers and an avalanche of sales. Patients demanded the drugs. And doctors took out their prescription pads, often not even trying to dissuade people who would have done just as well with a cheap alternative, like acetaminophen or aspirin.

Tens of millions of prescriptions have been written, and the drugs remain wildly popular. But now there are troubling questions about whether Vioxx may have an unexpected side effect - a very slight increase in the risk of heart attack.

The risk is hypothesized, not proved; it was suggested in a single Vioxx study and not in studies of Celebrex. And even one of the researchers who is most worried, Dr. Eric Topol, a heart disease expert at the Cleveland Clinic, said, "We are not talking about big risks here." He added that at most, the drugs might increase the risk of a heart attack by 0.3 percent to 0.5 percent.

In interviews, leading arthritis specialists who have no financial ties to Merck, the maker of Vioxx, or Pharmacia, which makes Celebrex, say that they are not concerned and that they continue to prescribe the drugs for patients who may have heart disease.

But the Food and Drug Administration recently sent Merck a warning letter, telling the company that its marketing campaigns had to discuss the possibility that Vioxx might increase the risk of heart attacks. And some heart specialists say they are now telling patients that they may want to consider taking other drugs for pain.

At issue is the subtle question of what counts as evidence. The new drugs, known as cox-2 inhibitors, were developed because scientists had discovered a way to make a drug that could alleviate pain and inflammation like aspirin, ibuprofen and naproxen, but without their tendency to cause stomach pain, bleeding or ulcers in a small percentage of people.

Huge clinical trials involving thousands of arthritis patients - the principal intended users of the new drugs - showed that the new drugs performed as expected. They were as effective in relieving inflammation as aspirin and the other nonsteroidal anti-inflammatory drugs that are available over the counter. And the incidence of stomach ulcers and their complications plummeted.

The heart attack hypothesis was unexpected, said Dr. Laura Demopoulos, a cardiologist who directs Merck's cardiovascular clinical research. It arose in a company study of 8,000 rheumatoid arthritis patients who were randomly assigned to take Vioxx or naproxen. The study, known as Vigor, was designed to look for serious gastrointestinal side effects.

The study found that patients taking Vioxx had less than half the incidence of these disorders - 2.08 percent of them had the complications, compared with 4.49 percent of the patients taking naproxen.

But the study also found, unexpectedly, that 101 Vioxx patients, or 2.5 percent, had serious cardiovascular problems during the study, like strokes, blood clots in a vein or heart attacks. In contrast, just 46 naproxen patients, or 1.1 percent, had similar cardiovascular problems. The heart attack incidence in the Vioxx patients was 0.5 percent - 20 patients - compared with 0.1 percent - four patients - who took naproxen.

The problem was, What did it mean? Ordinarily, if statisticians are looking for convincing evidence, they insist that a study be designed from the start to find it. In this case, the study was designed to look for gastrointestinal effects.

If the cardiovascular effects are real, and not spurious, it is not clear why. Did naproxen protect patients from heart attacks? Or did Vioxx make them more vulnerable?

Scientists scoured published papers, scrutinized data from thousands of patients in other Vioxx trials, and did laboratory studies and studies with monkeys. The company hypothesized, Demopoulos said, that all their data were consistent with the idea that naproxen, like aspirin, protects patients from heart attacks, rather than that Vioxx causes them.

"The totality of the data support that perspective," Demopoulos said. But none of the findings to date are enough to prove that the issue is fully resolved, and that lack of proof is why the the FDA demanded that Merck explain both sides of the hypothesis.

It is not known whether naproxen has the same protective effect as aspirin, and it remains a possibility, the agency said, that Vioxx itself caused the heart attacks by making blood more likely to clot. Greg Reaves, a spokesman for Merck, said the company took the warning letter "extremely seriously."

In the meantime, Dr. Steve Geis, the group vice president for clinical research of Pharmacia, said his company's drug, Celebrex, was getting tarred with the same brush. Pharmacia's studies never showed any increase in heart attacks or strokes in patients taking Celebrex, he said.

"We say, you must look at the data with the individual drugs and not make a sweeping statement about all cox-2 inhibitors," Geis said. He added that the company constantly heard from doctors and patients who were worried that Celebrex caused heart disease. "They write us, they call us," Geis said. And, he said, every time he addresses a medical meeting, he can count on being asked about it. In question-and-answer sessions after his talks, he said, someone always says, "Tell us about the cardiovascular effects."

Usually, Geis said, he has already discussed the data in his talk, but he takes it out again. "We systematically go through our data," he said, and he explains again that the Celebrex studies found no such effect.

Copyright 2001 The New York Times News Service. All rights reserved.