However, the findings do not mean all osteoarthritis patients should be given celecoxib, Dr Theodore Pincus of Vanderbilt University School of Medicine in Nashville, the study's lead author, told. He said there is "a group of patients who do better with acetaminophen and we should recognizs that."

Acetaminophen is recommended for osteoarthritis patients because of concerns about the gastrointestinal side effects of non-steroidal anti-inflammatory drugs (NSAIDs) and because of studies suggesting that the drug is as effective as ibuprofen and naproxen.

However, the researchers point out that recent reports indicate acetaminophen may be less effective for the disease than previously thought, and COX-2 inhibitors, such as celecoxib, have fewer gastrointestinal side effects than other NSAIDs.

In two studies, patients were randomly assigned to receive six weeks of 200 mg celecoxib daily, 1,000 mg acetaminophen four times daily or placebo. Patients were then switched to a second treatment. The first study involved 524 patients and the second, 556 patients.

Among patients in all three groups, the rate of adverse events reported ranged from 23 to 29 percent. No significant difference was seen among the three groups for gastrointestinal side effects.

On average, patients on celecoxib showed a 22 per cent improvement in osteoarthritis scores. Scores improved by 13 per cent in patients on acetaminophen and by 8 percent in patients on placebo. Patients' pain scores improved 28 percent with celecoxib, 18 percent with acetaminophen, and 10 percent with placebo.

More patients preferred celecoxib than acetaminophen, and more preferred acetaminophen than placebo. In the first study, 53 percent preferred celecoxib versus 24 percent for acetaminophen. In the second study, 50 percent preferred celecoxib and 32 percent preferred acetaminophen.

Patient preference for acetaminophen versus placebo was 37 to 28 percent in the first study and 48 to 24 percent in the second.

Source: Annals of the Rheumatic Diseases, July 2004.