Background Chronic treatment with dual cyclo-oxygenase (Cox-1/Cox-2) inhibitors is known to increase the risk of serious gastrointestinal (GI) adverse events (AE). Rofecoxib, a selective inhibitor of Cox-2 (coxib) exhibits an efficacy profile comparable to dual Cox-1/Cox-2 inhibitors. In a meta-analysis of several OA trials and in a large prospective study of RA patients, rofecoxib resulted in a markedly lower incidence of serious GI outcomes (i.e. perforation, ulcer and GI bleeding) versus comparator NSAIDs.

Objectives To assess the GI tolerability of rofecoxib compared to naproxen in the treatment of patients with OA in the primary care setting.

Methods 5597 patients with OA were treated for 3-months in a randomised, controlled trial conducted in the U. S. and Sweden. Patients with clinically defined OA of the knee, hip, hand or spine previously treated with an NSAID, coxibs, or acetaminophen were randomised to receive either naproxen (500 BID) or rofecoxib (25 mg QD). Patients using low dose aspirin for cardiovascular prophylaxis were allowed to enter the trial. The primary endpoint was GI tolerability as defined by the incidence of discontinuations due to any GI AEs. The secondary endpoint was usage of concomitant GI medication to treat GI symptoms initiated during the study. Efficacy and overall safety and tolerability were assessed at office visits at baseline, 6 and 12 weeks and by telephone at weeks 3 and 9.

Results The majority of patients were female (71%). The mean age was 63 + 11 years. The racial composition was White (87%), African American (9%), and other racial origin (4%). 12% used low dose aspirin during the trial. The primary study joints were: knee (50%), spine (24%), hand (16%) and hip (10%). In 90% of patients, joints other than their primary joint were involved by OA. 92% had OA symptoms for more than one year. For prior OA therapy, 62% of patients used single therapy with NSAIDs, 7% used acetaminophen alone, and 30% used both. The most common cardiovascular and GI secondary diagnoses at study entry were hypertension (45%) and heartburn (13%). Treatment groups were similar in baseline characteristics. The primary hypothesis, that there would be lower incidence of discontinuations due to GI AEs in patients treated with rofecoxib, was confirmed (rofecoxib: 5.9%; naproxen: 8.1%; p = 0.005). The secondary hypothesis, that rofecoxib patients would require less GI concomitant medication than naproxen patients, also was confirmed (rofecoxib: 9.1%; naproxen: 11.2%, p = 0.014). The incidences of lower extremity oedema (rofecoxib: 3.5%, naproxen: 3.8%) and hypertension (rofecoxib: 2.9%, naproxen: 2.5%) were similar. Concomitant use of low dose aspirin, did not significantly affect relative rates of discontinuation due to AEs, serious AEs or drug related AEs. No significant difference in OA efficacy was noted between treatments for Patient Global Assessment of Disease Status [Wk 6: 11.6 mm and 10.8 mm; Wk 12: 10.4 mm and 9.6 mm; for rofecoxib and naproxen, respectively), the AUSCAN OA Hand Index or incidence of discontinuation due to lack of efficacy.

Conclusion In patients with OA of the knee, hip, hand or spine, which included patients concomitantly treated with low dose aspirin, rofecoxib 25 mg QD demonstrated superior GI tolerability and resulted in significantly less use of concomitant GI medications while demonstrating similar efficacy compared to naproxen 500 mg BID.