Early response to COX-2 inhibitors as a predictor of overall response in osteoarthritis: pooled results from two identical trials comparing etoricoxib, celecoxib and placebo

OBJECTIVE: We evaluated whether early response to NSAIDs predicted later response, and when this was established. METHODS: We evaluated pooled data from two identical 26-week, double-blind, randomized trials comparing once-daily etoricoxib 30 mg (n = 475), celecoxib 200 mg (n = 488) and placebo (n = 244) in patients with knee or hip OA. The present analysis was limited to the 12-week placebo-controlled period. Patient-level OMERACT-OARSI response was determined at 2, 4, 8 and 12 weeks. The proportion of patients who maintained response status between these times was determined from binomial distribution using the exact method. RESULTS: After 12 weeks of treatment, there were significantly more responders in the etoricoxib (59.8%) and celecoxib (57%) groups compared with placebo (34%; P < 0.001 for etoricoxib or celecoxib vs placebo). About 77.2% of the patients receiving etoricoxib, 75.4% celecoxib and 58% placebo (P = 0.001 vs etoricoxib; P = 0.003 vs celecoxib) who were responders at 2 weeks were also responders at 12 weeks. When comparing response agreement (responder or non-responder) at 2 weeks and 12 weeks, 74.3% of the patients receiving etoricoxib, 73.2% celecoxib and 71.3% placebo had the same response status (kappa-coefficient 0.459, 0.449 and 0.357, respectively). There were small incremental increases in agreement between Weeks 4 and 8 and 12 weeks. Logistic regression showed that agreement was not affected by index joint (P = 0.965). CONCLUSIONS: The overwhelming majority of the patients who responded to treatment by 2 weeks remained responders at 12 weeks, with response status largely established within 2 weeks of treatment initiation. Early identification of NSAID response or non-response may allow clinicians to better and more rapidly adjust symptomatic OA management.