This 12-week, multicenter, randomized, double-blind, placebo-controlled, dose-ranging trial evaluated tramadol ER (extended-release tramadol) in the management of osteoarthritis pain. Adults with knee and/or hip osteoarthritis and baseline pain intensity of ≥40 on a 100-mm visual analog scale (0 = no pain, 100 = extreme pain) received once-daily tramadol ER 100 mg (n = 201), 200 mg (n = 199), or 300 mg (n = 199), celecoxib 200 mg (n = 202; to test model sensitivity), or placebo (n = 200). Coprimary efficacy variables were Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain subscale, WOMAC physical function subscale, and patient global assessment of disease activity. Tramadol ER 300 mg significantly improved patient global assessment scores compared with placebo (P ≤ 0.05), but not the other 2 coprimary efficacy variables. Tramadol ER 200 and 100 mg were not significantly different from placebo for the coprimary efficacy variables. Daily diary arthritis pain intensity scores improved significantly for tramadol ER 300 and 200 mg compared with placebo. WOMAC joint stiffness subscale, physician's global assessment, arthritis pain intensity in index and nonindex joints, and overall sleep quality scores improved significantly for tramadol ER 300 mg compared with placebo. Significant differences in efficacy between celecoxib and placebo validated the model sensitivity. Adverse events occurred more frequently with tramadol ER than placebo in the gastrointestinal (nausea, constipation, diarrhea) and central nervous (dizziness, headache) systems. In this study, tramadol ER 300 mg was effective in the management of moderate to severe painful osteoarthritis of the hip or knee. A large, increasing placebo response during the study may have contributed to the lack of statistical separation between tramadol ER 200 or 100 mg and placebo.