Response to nonsteroidal anti-inflammatory drugs in African American subjects with osteoarthritis of the knee

Objectives: Celecoxib is an effective treatment for osteoarthritis (OA); however, its efficacy and safety profile in African-Americans has not been specifically studied. Differences in therapeutic response between racial/ethnic populations are reported in other diseases but have not been thoroughly investigated in OA.  This study was designed to compare analgesic efficacy, tolerability, and safety of celecoxib, naproxen, and placebo in an African American population with OA of the knee.

Methods: African American subjects aged ≥ 45 years with OA of the knee in a flare state were randomized in a double-blind, parallel-group trial to receive celecoxib 200 mg qd, naproxen 500 mg bid, or placebo for 6 weeks. The trial had a non-inferiority design.  Primary end point was change from baseline in the Patient's Assessment of Arthritis Pain.  Secondary efficacy outcomes included other measures of pain, functionality, and quality of life assessments, evaluation of safety and upper gastrointestinal (UGI) tolerability.

Results: 322 subjects (80% female, mean age 58 years [range 45-83], mean duration of OA >5 years) were randomized, 69 discontinued prematurely.  For the primary end point, celecoxib was shown to be as effective as naproxen in reducing OA pain. Similar efficacy between celecoxib and naproxen was observed in secondary outcomes (Patient's Global, Physicians' Global, WOMAC, APS). Improvements in primary and secondary outcome measures were numerically greater in the active treatments compared with the placebo groups; however, few of these improvements reached statistical significance. Celecoxib demonstrated favorable UGI tolerability compared to naproxen; fewer subjects experienced moderate/severe nausea, abdominal pain or dyspepsia.

Conclusion: Celecoxib was as effective as naproxen in relieving pain associated with OA of the knee in African American patients in a trial designed to evaluate this population.  Few significant differences were observed between the active treatments and placebo, possibly due to a strong placebo effect or differences in flares.