Q&A

Which is most effective for osteoarthritis of the knee: rofecoxib, celecoxib, or acetaminophen?

Author and Disclosure Information

Geba GP, Weaver AL, Polis AB, Dixon ME, Schnitzer TJ, et al. Efficacy of rofecoxib, celecoxib, and acetaminophen in osteoarthritis of the knee. A randomized trial. JAMA 2002; 287:64-71.


 

ABSTRACT

BACKGROUND: Traditional nonsteroidal antiinflammatory drugs (NSAIDs) and the newer cyclooxygenase-2 enzyme (COX-2) selective inhibitors are recommended as second-line agents in patients with osteoarthritis (OA) who fail to respond to acetaminophen. This study compared the effectiveness of rofecoxib (Vioxx), celecoxib (Celebrex), and acetaminophen (Tylenol) in patients with OA of the knee.

POPULATION STUDIED: This study included 382 patients from 29 US clinical centers with symptomatic OA of the knee for 6 months or longer. All patients had been treated with NSAIDs or acetaminophen for at least 30 days before enrollment, were 40 years of age or older, and retained moderate functional mobility of the knee (American College of Rheumatology functional class I, II, or III). Baseline criteria for OA severity were determined using the Western Ontario McMaster University Osteoarthritis Index (WOMAC) and Investigator Global Assessment of Disease Status scoring. Patients were excluded if they had concurrent medical or arthritic disease or abnormal laboratory results that would have confounded the effectiveness evaluation or increased the risk of complications.

STUDY DESIGN AND VALIDITY: This research was a randomized double-blind controlled study. Allocation to treatment group (using computer-generated assignment) was concealed from enrolling investigators. After a 3-day to 7-day washout period, patients were randomized to receive 12.5 mg rofecoxib once daily, 25 mg rofecoxib once daily, 200 mg celecoxib once daily, or 1000 mg acetaminophen 4 times daily for 6 weeks. Exact matching placebos were used to maintain double-blind conditions. Response was evaluated using intent-to-treat analyses. Early effectiveness, using the WOMAC Index and Patient’s Global Assessment of Response to Therapy (PGART) questionnaires, was defined as occurring within the first 6 days. Later clinical effectiveness was evaluated during office visits using the WOMAC and PGART at weeks 2, 4, and 6.

OUTCOMES MEASURED: The primary outcomes measured were pain on walking, night pain, pain at rest, and morning stiffness (WOMAC Index) and global responses to therapy (PGART).

RESULTS: Seventy-nine percent of patients completed the 6-week follow-up. More patients treated with acetaminophen than patients treated with either rofecoxib or celecoxib discontinued early because of lack of effectiveness (17% vs 8% to 9%; composite number needed to treat for 1 withdrawal because of lack of efficacy = 8). As compared with celecoxib or acetaminophen, WOMAC response over 6 weeks showed that 25 mg rofecoxib once daily provided significantly greater responses in reduction of rest and night pain, composite pain scale, and stiffness scale. Physical function scale results were significantly better with 25 mg rofecoxib once daily than with acetaminophen but were no different from those with celecoxib. PGART response at 6 weeks also showed the best response with 25 mg rofecoxib once daily. Early response results were similar to later response results in showing that the best response was achieved with 25 mg rofecoxib once daily.

RECOMMENDATIONS FOR CLINICAL PRACTICE

In this study, 25 mg rofecoxib once daily was more effective than either celecoxib or acetaminophen in relieving persistent pain and stiffness from knee OA. However, only 1 of 6 patients taking acetaminophen, which is inexpensive and safe, discontinued treatment for lack of efficacy. Therefore, using acetaminophen as first-line therapy is reasonable. Less expensive traditional NSAIDs (eg, ibuprofen or naproxen) have been shown to have similar effectiveness as compared with either rofecoxib or celecoxib in OA. For patients at low risk for serious NSAID-associated gastrointestinal complications, traditional NSAIDs should be the next agents of choice. For patients at high risk, COX-2 selective inhibitors are reasonable second-line agents, since they pose a lower risk of NSAID-associated gastrointestinal complications with long-term use.

Next Article: