Objective: To evaluate impact of age, gender, race and BMI on NSAID efficacy in OA. Methods: Pfizer Clinical Trials Registry was searched for studies meeting the following criteria: randomized parallel-group design; Ն1 treatment group receiving celecoxib Ն200 mg daily; Ն1 placebo, NSAID, or rofecoxib comparator group; planned duration of Ն6 weeks; study completed and report finalized by October 31, 2004; 1 primary OA efficacy end point and pain visual analog scale (VAS) of 100 mm collected. In analyzing efficacy, we used the minimal clinically important difference (MCID), which proposes that patients are unable to detect a small difference in efficacy. For MCID a change of Ն10 mm on the VAS from baseline to last observation was used to define responders. Percent responders and differences in responder rates were calculated for placebo and NSAIDs combined for age, gender, BMI, and race. For each treatment, responder rates were compared statistically for categorical data (gender and race) using a test of incidence rates within each treatment category, and CMH for overall association. For continuous variables (BMI and age), inference was calculated on the slope of the regression of VAS change on predictor variables. No adjustments in the nominal alpha level were made for the multiple analyses. Results: Fifteen studies were included: 21,798 OA patients received celecoxib (61%), diclofenac (19%), naproxen (10%), ibuprofen (1%), rofecoxib (2%), or placebo (8%). The majority of patients were female (73%) and Caucasian (78%). Median (range) for age and BMI were: 63 years (18 -96) and 29 (11-149), respectively. NSAID response was influenced significantly by age (P ϭ 0.0099), BMI (P ϭ 0.0002), gender and race (P Յ 0.0001); placebo response was influenced by age (P ϭ 0.0002) and race (P Յ 0.0001). Differential response between NSAIDs and placebo was significantly influenced by all variables (P Յ 0.0001). Conclusions: Age, gender, BMI, and race are important predictors of NSAID response, while only age and race influence placebo response. Predictors of the highest benefits of pain reduction with NSAIDs versus placebo are: older age, female sex, being overweight, and Caucasian race.