Celecoxib (Cele) v placebo or NSAIDs for rheumatoid arthritis (RA) or osteoarthritis (OA)*

Outcomes at 12–52 weeksPatient groupComparisonsWeighted event ratesRBI (95% CI)NNT (CI)
Met ACR-20 responder criteriaRACele v placebo38% v 26%49% (25 to 78)9 (6 to 14)
Cele v NSAIDs29% v 28%†4% (−20 to 36)†Not significant
Improvement in numberof painful/tender jointsCele v placebo48% v 34%39% (21 to 61)8 (6 to 13)
Cele v NSAIDs44% v 40%†9% (−10 to 32)†Not significant
Withdrawal because of any adverse effectAllCele v placebo8.8% v 6.2%†48% (15 to 92)†39 (24 to 117)†
Withdrawal because of GI adverse effectCele v placebo3.3% v 2.0%68% (7 to 165)80 (44 to 450)
Withdrawal because of any adverse effectCele v NSAIDs8.5% v 9.7%14% (−4 to 28)Not significant
Withdrawal because of GI adverse effectCele v NSAIDs3.6% v 6.3%†46% (29 to 58)37 (26 to 67)†
Mean difference (CI)
*ACR-20 = American College of Rheumatology index; GI = gastrointestinal; NSAIDs = non-steroidal anti-inflammatory drugs; WOMAC = Western Ontario and McMaster universities osteoarthritis index. Other abbreviations defined in glossary; weighted event rates, NNT, NNH, and CI calculated from data in article using fixed effects.
†Significant heterogeneity was present; a random effects model was used.
‡Statistically significant difference favouring celecoxib.
WOMAC pain score (0–20)OACele v placebo−1.35 (−1.74 to −0.97)‡
Cele v NSAIDs−0.01 (−0.73 to 0.72) †
WOMAC joint stiffnessScore (0–8)Cele v placebo−0.65 (−0.80 to −0.49)‡
Cele v NSAIDs0.06 (−0.11 to 0.22)