The first published studies of Arcoxia, the drug that Merck & Co. hopes will take the place of its withdrawn painkiller Vioxx, are getting mixed reviews from doctors, some of whom say the results do not make a case for the medication's approval.

The critics cite not just Arcoxia's side effects but also that Merck tested it against diclofenac, an older painkiller known to raise heart risks. A fairer comparison would have been to a medicine that does not do that, such as naproxen, sold as Aleve, they say.

"The development program for Arcoxia is fatally flawed," said Dr. Steven Nissen, a Cleveland Clinic cardiologist who formerly headed the Food and Drug Administration's cardiac drug advisory panel. "My advice to the FDA is that they should not approve this drug."

Dr. David Graham, an FDA drug safety expert who has criticized his agency's handling of Vioxx, agreed.

"It's my own suspicion that this study was intentionally designed to minimize the possibility of their having a repeat of what happened with VIGOR," the study that revealed Vioxx's heart risks, he said.

The new research "raises red flags," Graham said, and gave his personal opinion that the drug should not be approved. "There's every reason to think it suffers from the same problems" as Vioxx, he said.

Merck stock rose 47 cents to $43.59 in morning trading on the New York Stock Exchange.

Vioxx was a blockbuster arthritis drug when Merck pulled it from the market in September 2004 over safety concerns. Like Arcoxia, it is a cox-2 inhibitor, a class of drugs originally developed to be gentler on the stomach and hopefully easier to take long-term than traditional pain relievers.

Arcoxia already is widely used in other countries, and Merck has been seeking approval to sell it in the United States. In August, the company said preliminary data from unpublished studies showed that Arcoxia did not increase heart risks more than diclofenac, sold as Cataflam and Voltaren by Switzerland-based Novartis AG.

On Monday, pooled results of three studies of Arcoxia were to be published in the British journal The Lancet and presented at an American Heart Association conference in Chicago.

They involved 34,701 arthritis patients, average age 63, in the United States and 37 other countries. One-third had heart disease or multiple risk factors for it.

After an average of 18 months on either drug, clot-related heart problems were statistically similar — 320 among the 17,412 who started the study on Arcoxia, and 323 among the 17,289 on diclofenac, said lead author Dr. Christopher Cannon of Brigham and Women's Hospital in Boston.

About one percent of each group suffered a heart attack, stroke or death within a year of starting on their drug. That rose to about 2.5 percent three years later.

Arcoxia users were 30 percent less likely to have ulcers, stomach bleeding or other gastrointestinal problems (176 cases versus 246 on diclofenac), but the rate of serious problems was similar.

"Most of the GI bleeding in this study was probably microscopic, meaning it wasn't clinically important," the FDA's Graham said.

Slightly more than half of each group stopped taking whichever drug they had been assigned. Side effects were the reason for 20 percent of the discontinuations. More people quit Arcoxia because of high blood pressure; more stopped diclofenac because of stomach or liver problems. Congestive heart failure was rare but more common among those receiving a higher dose of Arcoxia than among those on a lower dose or diclofenac.

Cannon consults for Merck and other drug makers. Nissen currently heads a large study funded by Merck's chief rival in this field, Pfizer Inc., though he said he gets no salary from it and that all research revenue goes to the Cleveland Clinic Foundation.

The Arcoxia study "isn't the answer to everything" about painkillers, but it provides one piece of the puzzle, Cannon said.

"A picture is emerging here," said co-author Garret FitzGerald, a University of Pennsylvania scientist who believes that all cox-2 drugs bring additional heart and stroke risks because they work through a common mechanism.

"If you have people that are at risk of heart disease ... you stay away from these drugs at the cox-2 end of the spectrum and tilt towards the other end, where naproxen lies," he said.

Conversely, if someone's risk of ulcer or stomach bleeding is high, a cox-2 might make more sense, he said. Pfizer's Celebrex is the only one still on the market in the U.S.