Pooled analysis of GI tolerability of 21 randomized controlled trials of celecoxib and nonselective NSAIDs

Abstract

Objective: To compare the gastrointestinal (GI) tolerability of celecoxib and nonselective nonsteroidal anti-inflammatory drugs (NSAIDs) at approved doses in patients with common musculoskeletal conditions.

Research design and methods: This was a retrospective, pooled analysis of studies selected from the Pfizer Corporate Clinical Trials Registry. Study selection criteria were: (1) Data available as of October 31, 2004; (2) Randomized, parallel-group study design and planned treatment duration of >or= 2 weeks; (3) At least one nonselective NSAID (naproxen, ibuprofen, or diclofenac) as a comparator; (4) At least one arm with 200 mg or 400 mg celecoxib per day; (5) Patients with osteoarthritis (OA), adult rheumatoid arthritis (RA), or ankylosing spondylitis (AS). Data were pooled by treatment and by subject from the safety analysis population of each included study. Joint primary end points were the combined incidence of tolerability-related GI adverse events (AEs) (abdominal pain, dyspepsia, nausea, diarrhea, and flatulence), and time to study discontinuation due to any GI AE.

Results: In all, 21 studies met the selection criteria. Across the safety analysis populations of the included studies, 7797 patients received celecoxib total daily dose 200 mg/day, 6653 received celecoxib total daily dose 400 mg/day, 2953 received naproxen, 499 received ibuprofen, and 5643 received diclofenac. Tolerability-related GI AEs were reported by significantly fewer celecoxib-treated patients (16.0%) than by those treated with naproxen (24.3%), ibuprofen (24.2%), or diclofenac (19.9%) (p < 0.0001 vs. each comparator). Time to study discontinuation due to any GI AE was significantly longer for celecoxib than for naproxen (p < 0.0001), ibuprofen (p = 0.002), or diclofenac (p = 0.048). In the RA subpopulation (n = 2857), there was no significant difference between the celecoxib and naproxen or ibuprofen groups in incidence of tolerability-related GI AEs and GI AEs.

Limitations: The limitations are inherent to the retrospective analysis design.

Conclusions: In this pooled analysis of celecoxib at approved doses in OA, RA, and AS, fewer celecoxib-treated patients in the overall population had tolerability-related GI AEs than patients treated with naproxen, ibuprofen, or diclofenac. In addition, celecoxib-treated patients had a significantly longer time to study discontinuation due to GI AEs.