Objective: To compare the incidence of selected spontaneously reported adverse events (AEs) in patients with osteoarthritis (OA) treated with rofecoxib (VIOXX, 12.5 mg qd) or Arthrotec (diclofenac 50 mg/misoprostol 200 mcg bid). Methods: Double-blind, parallel-group, 6-week study of patients aged ≥ 40 years with a clinical diagnosis of OA treated with rofecoxib or Arthrotec. Primary endpoint: self-reported diarrhea; secondary endpoints: abdominal pain, discontinuations due to AEs, GI AEs and NSAID-type GI AEs (ie., acid reflux, dyspepsia, epigastric discomfort, heartburn, nausea, vomiting). Results: Among 483 patients (80.3% females, mean age 62.1), the rofecoxib group vs the Arthrotec group respectively reported diarrhea 6.2% vs 16.2% (p<0.001); drug-related diarrhea 3.7% vs 16.2% (p<0.001); one or more clinical AEs 52.9% vs 73.0% (p<0.001); GI AEs 28.9% vs 48.5% (p<0.001); NSAID-type GI AEs 18.6% vs 29.9% (p = 0.004); discontinuations due to abdominal pain 0.4% vs 3.7% (p<0.05); and discontinuations due to any AE 4.1% vs 9.1% (p = 0.029). No significant differences were observed in efficacy. Conclusion: Rofecoxib 12.5 mg qd has improved GI tolerability and similar efficacy compared to Arthrotec (diclofenac 50mg/misoprostol 200 mcg bid).