Osteoarthritis and Cartilage Vol. 15, Supplement C C141 Results: The study included 65 women and 35 men, with 77% being White, 18% African American, 2% Asian, and 3% Other. Mean age of the subjects was 63.5 years (± 8.7). 54% of the subjects reported daily vitamin D supplementation, with 31% taking <400 IU vitamin D and 23% taking ≥400 IU. Mean base- line serum (OH)D level was 31.6 (± 12.6) ng/ml, with 47% of subjects found to be vitamin D deficient. Many subjects were vitamin D deficient in spite of supplementation (42% who took ≤400 IU vitamin D, and 30% of those taking >400 IU vitamin D). However, after adjustment for age, gender, race, BMI, edu- cation, and month of blood draw, vitamin D supplementation was associated with decreased risk of vitamin D deficiency (subjects taking 1-400 IU vitamin D vs. no-taking, OR = 0.51, 95% CI, 0.16-1.63; subjects taking >400 IU vitamin D vs. no-taking, OR = 0.26, 95% CI, 0.07-1.04), p trend = 0.047. In these models, lower vitamin D levels were inversely associated with education (college graduate or above vs. under, OR = 0.28, 95% CI, 0.09- 0.84), and positively associated with African-American race (OR = 4.10, 95% CI, 0.94-18.0). Conclusions: Many individuals with knee OA are vitamin D deficient in spite of daily vitamin D supplementation. Current recommendation with respect to vitamin D supplementation may require reconsideration for individuals with OA. 249 LUMIRACOXIB IMPROVES PAIN AND FUNCTIONAL STATUS IN PATIENTS WITH PRIMARY HIP OSTEOARTHRITIS: A 13-WEEK, MULTICENTER, RANDOMIZED, PLACEBO-CONTROLLED, ACTIVE COMPARATOR TRIAL T.J. Schnitzer 1 , I.D. Dattani 2 , B. Seriolo 3 , H. Schneider 4 , S. Yu 5 , L. Tseng 5 , R. Rebuli 6 1 Northwestern Center for Clinical Research, Chicago, IL; 2 Prairie Clinical Research Group, Saskatoon, SK, Canada; 3 Ricercatore Confermato, Clinica Reumatologica, Università di Genova, Genova, Italy; 4 Facharzt für Orthopädie/Sportmedizin, Bad Nauheim, Germany; 5 Novartis Pharmaceuticals Corporation, East Hanover, NJ; 6 Novartis Pharma AG, Basel, Switzerland Purpose: Hip osteoarthritis (OA) is a debilitating musculoskeletal disorder, characterized by chronic pain, affecting health related quality of life. Lumiracoxib is a selective COX-2 inhibitor that provides an alternative to traditional NSAIDs for the treatment of OA. The present study was designed to demonstrate the efficacy and safety of lumiracoxib (100 mg od) compared with placebo in patients with primary hip OA. Celecoxib (200 mg od) was used as the positive control. Methods: The primary efficacy variables included patient’s global assessment of disease activity (0-100 mm VAS), WOMAC 3.1 LK pain sub-scale score and WOMAC difficulty performing daily activities (DPDA) sub-scale score (at Week 13). Prespec- ified secondary efficacy variables included response to treat- ment according to Osteoarthritis Research Society International (OARSI) criteria at Weeks 4, 8 and 13; and minimal clinically im- portant improvement (MCII) for hip OA pain by visit. The efficacy in the subgroup of patients ≥65 years was also prespecified to be analyzed for these primary outcomes. Results: A total of 1262 primary hip OA patients aged ≥40 years (490 were ≥65 years of age) were randomized to receive lumiracoxib 100 mg od (n=427), celecoxib 200 mg od (n=419), or placebo (n=416) over a period of 13 weeks. Lumiracoxib was significantly superior to placebo in all primary efficacy variables and provided similar pain relief compared with celecoxib at 13 weeks (in both overall population and ≥65 years sub group, Table 1). The treatment response in <65 age group was also similar to the overall population and statistically significant compared to placebo. Table 1. Efficacy of lumiracoxib versus placebo for primary outcomes (estimated difference after 13 weeks) Lumiracoxib vs Placebo LSM difference (95% CI) Overall Population ≥65 years Patient’s global assessment of disease activity (VAS, mm) -8.59 (-11.84, -5.33)* -9.60 (-14.94, -4.25)* WOMAC pain (0-20) scores -1.12 (-1.63, -0.60)* -1.63 (-2.48, -0.78)* WOMAC DPDA (0-68) scores -3.58 (-5.24, -1.91)* -4.16 (-6.90, -1.42)** *p<0.001, **p=0.003. Similar results for all primary endpoints were observed for cele- coxib vs placebo in the overall population and patients ≥65 years. A greater proportion of patients in the lumiracoxib group (63.0%, 67.0%, 65.8%) had a treatment response according to OARSI criteria at Weeks 4, 8 and 13 compared with placebo (45.4%, 50.0%, 53.6%) and the response was similar to that in the cele- coxib group (63.7%, 67.5%, 67.8%). Lumiracoxib and celecoxib were statistically superior to placebo in the achievement of MCII for target hip OA pain after 4 weeks (p≤0.001; p=0.004), 8 weeks (both p≤0.001), and 13 weeks (both p≤0.001) of treatment. The incidence of adverse events in the lumiracoxib, celecoxib and placebo groups was 56.4%, 53% and 47.6%, respectively. The overall percentage of serious adverse events (SAEs) was 1.9%, 1.0% and 2.2% in the respective groups. There were no fatal SAEs in the lumiracoxib and placebo groups. Two deaths oc- curred in the celecoxib group that were not suspected to be study drug related. Two patients (0.5%) had AST and/or ALT elevations >3 x ULN in the placebo group versus no elevations in the active treatment groups. Conclusions: Lumiracoxib 100 mg od and celecoxib 200 mg od were well tolerated and provided similar pain relief and improve- ment in functional status in patients with primary hip OA. Efficacy was maintained in patients ≥65 years. 250 IDENTIFYING FACTORS TO PREDICT PLACEBO RESPONDERS IN KNEE OSTEOARTHRITIS CLINICAL TRIALS S.J. Bartlett, M.S. Solinger, C.O. Bingham III Johns Hopkins University, Baltimore, MD Purpose: The placebo effect is inherent to all treatments. Phar- macotherapy trials assess whether specific effects of an agent are significantly greater than non-specific influences (placebo) on treatment outcomes. In OA trials, rates of placebo response often approach or exceed 50%. Despite this, very little is known about sociodemographic or clinical predictors of placebo response. Methods: Data were drawn from placebo arm of North Amer- ican patients in a large clinical trial designed to evaluate the clinical effects of risedronate on knee OA (KOSTAR). OARSI pain criteria (20% and 50% decrease in WOMAC pain) were used to classify participants at 6 months as placebo respon- ders. We examined whether placebo response was associated with baseline demographic, clinical, and radiographic parameters including: age, gender, race, BMI, work status, WOMAC (pain, stiffness, function), patient global assessment, OA at other sites, NSAID, analgesic and glucosamine use, joint space width (JSW) and osteophyte grade. Results: Of 311 OA patients who started the trial, 269 (86%) completed the 6 month evaluations. Subjects were mostly female (60%) and white (84%) with a mean (± SD) age of 60.4 ± 8.9, BMI of 30.2 ± 5.0, OARSI osteophyte grade of 1.6 ± 0.6 and baseline pain (WOMAC) of 35.1 ± 21.5. Nearly half (49%) reported a 20% decrease in pain while 29% reported a 50% decrease, with no between-group differences by gender or