S204 Osteoarthritis and Cartilage Vol. 16 Supplement 4 Conclusions: We showed that CRP levels are associated with OA- related knee pain in women. This might indicate that there is an in- flammatory component present in knee OA that contributes to knee pain experienced in OA. In addition, we showed that BMI modifies the effect of CRP levels on KOA. This may indicate a difference in pathogenesis of OA in overweight women compared to women with a healthy weight. For example, mechanical load is a major contributor in overweight women, whilst in women with a healthy weight systemic inflammation might play a more prominent role. Another explanation could be that we do not see the effect in overweight women because of confounding by co-morbidity such as cardiovascular disease or diabetes, although we tried to adjust for this in our analysis. 470 COMPARING PLACEBO RESPONSE IN NORTH AMERICAN AND EUROPEAN PATIENTS WITH KNEE OA S.J. Bartlett, C.O. Bingham III. Johns Hopkins, Baltimore, MD, USA Purpose: In OA trials, rates of placebo response are often 50%+. Despite this, little is known about sociodemographic or clinical predictors. We hypothesized that rates of response and predictors would be similar between patients in North America (NA) vs. Europe (EUR). Methods: Data were drawn from 622 males and females randomized to the placebo arm of a large multinational clinical trial designed to evaluate the clinical effects of a bisphosphonate on knee OA. Patients were 40-80 years of age who reported knee pain due to OA on most days during 1 of the prior 3 months, morning stiffness lasting <30 min or knee crepitus according to the ACR criteria for knee OA. No minimum level of pain was required. Patients also had 1+ osteophytes, joint space width (JSW) of 2-4 mm in the medial tibiofemoral compartment, and a medial compartment < lateral. WOMAC pain and function scales were administered at baseline and 6 months. OARSI 2004 Responder Criteria were used to classify participants at 6 months. Results: 552 (83%) patients completed the 6 month evaluations. Overall, patients were mostly female (63%) and white (79.2%) with a mean (±SD) age of 62.0±8.8 years, BMI of 29.8±4.6, and JSW of 3.0±0.6 mm. EUR patients were significantly (p < 0.01) older, more likely to be female, lighter and use less analgesic medication. EUR patients also had significantly (p < 0.01) higher WOMAC pain, function and patient global assessment scores at baseline. More participants in the EUR as compared to the NA sites met the criteria for treatment response (23.5% vs. 21.0%, respectively), although the difference was not statistically significant (p = 0.488). A breakdown of patients meeting criteria for high response and improvement by cohort is shown below. At baseline, placebo responders had significantly lower scores for WOMAC pain (31.6±1.5 vs. 41.6±1.1; p < 0.001), function (37.8±1.7 vs. 44.9±1.2; p < 0.001) and patient global assessment (50.2±2.2 vs. 55.4±1.1; p = 0.032). No statistically significant differences were found between responders and non-responders for age, gender, body mass index, baseline JSW, OARS grade, osteophytes or use of pain medica- tion. In multivariate logistic regression models, even with adjustment for baseline sociodemographic and OA-related characteristics, baseline pain was the only consistent predictor of placebo response in both joint and separate models. Conclusions: Using OARSI Response Criteria, more than 1 in 5 partic- ipants who were in the placebo arm met stringent criteria for treatment response. Responders were more likely to report moderate levels of pain and impairments in physical function than non-responders at baseline. Overall, there do not appear to be sociodemographic or OA-related characteristics reliably associated with placebo response. Pain: Pathophysiology 471 QUANTITATIVE SENSORY TESTING IN OSTEOARTHRITIS OF THE KNEE P.A. Dieppe 1 , S. Ayis 2 , S. Clarke 2 , D. Simmons 2 , S. Williams 2 , M. Fallon 3 , L. Colvin 3 . 1 University of Oxford, Oxford, UNITED KINGDOM, 2 University of Bristol, Bristol, UNITED KINGDOM, 3 University of Edinburgh, Edinburgh, UNITED KINGDOM Purpose: To assess skin sensitisation over the medial aspect of the knee in people with osteoarthritis (OA) of the knee, using quantitative sensory testing (QST) and relate this to their pain experiences Methods: 28 patients with established medial compartment tibiofemoral joint OA were recruited to the study. They were interviewed about pain, and they completed WOMAC, HADs and MGill pain questionnaires. QST was then used to assess any area of altered sensation in the skin, test for mechanical sensitivity and pain thresholds, and assess skin thermal sensitivity. Results: The group included 20 women and 8 men with a mean age of 62.8 years, knee pain of an averge of 5 years duration, and radiographic evidence of medial compartment knee OA. Their mean WOMAC and HADs scores were 47.8 (range 16-73) and 14.2 (6-31) respectively. 26/28 had an area of altered sensation over the medial aspect of the index knee. Sensitivity to touch was lower in that knee, but pain threshold higher than in the contralateral knee. 19/28 had thermal allodynia, severe in 4. Those with cold allodynia had higher WOMAC and HADs scores than the whole group, and were more likely to use McGill pain descriptors suggesting of neuropathic pain. Conclusions: Sensory processing is altered in most patients with knee OA, and in some cases neuropathic pain may make an important contri- bution to the pain experience. This has implications for therapy. 472 SYNOVITIS INCREASES THE RISK OF KNEE PAIN IRRESPECTIVE OF STATUS OF X-RAY OA: THE MOST STUDY K. Baker 1 , A. Grainger 2 , J. Niu 1 , M. Clancy 1 , A. Guermazi 1 , M. Crema 1 , L. Hughes 3 , J. Buckwalter 4 , M. Nevitt 5 , D. Felson 1 . 1 BUMC, Boston, MA, USA, 2 U Leeds, Leeds, UNITED KINGDOM, 3 UAB, Birmingham, AL, USA, 4 U Iowa, Iowa City, IA, USA, 5 UCSF, San Francisco, CA, USA Purpose: Pain from knee OA is a key symptom in the decision to seek medical care and an important antecedent to disability. Synovitis has been hypothesized as one pathological feature that causes pain. Synovial thickening on contrast enhanced (CE) MRI has been shown to correlate well with synovitis on histology. On non CE MRIs synovial thickening cannot be assessed and on these images synovitis has been weakly and inconsistently associated with pain. We used CE MRIs, to assess synovial thickening in relation to knee pain severity among subjects in the Multicenter Osteoarthritis Study (MOST). Methods: MOST is a NIH-funded longitudinal observational study of risk factors for knee structural changes and occurrence of pain in persons age 50-79 with or at high risk of knee OA. An unselected subset of partici- pants who volunteered obtained 1.5 CE MRI of one knee at the 30 month clinic visit. Synovitis was scored 0-3 in 4 compartments (suprapatellar pouch, medial and lateral parapatellar recesses and infrapatellar fat pad) and 0-1 in 2 compartments (medial and lateral posterior condylar) (Rheumatol 44:1569 2005). We categorized synovitis in the whole knee as severe (>1 compartment scored 3), moderate (>2 compartments scored 2 but no 3s), mild (>4 compartments scored 1 or 1 scored 2), normal/questionable (<4 compartments scored as 1). Inter-reader kappa was 0.9 (p < 0.001). Subjects were asked about their knee pain severity using WOMAC. Each of the 5 items in the WOMAC pain scale is scored from 0 (no pain) to 4 (extreme pain). We used the maximum item score to define the categories of knee pain severity: no-pain (all 5 items scored 0 prior to a clinic visit), mild-pain (maximum score of any item 3). We examined the association between synovitis and pain severity using a logistic regression model adjusting for age, sex, BMI, MRI bone marrow edema, and effusions. We also examined if the effect of synovitis on pain severity was modified by radiographic OA status. Table 1 All subjects Synovitis Worst pain on any WOMAC pain question Normal/ questionable (n = 151) Mild (n = 233) Moderate/severe (n = 69) None (n = 158) 69 (46%) 79 (34%) 10 (15%) Mild (n = 149) 45 (29%) 79 (34%) 25 (36%) Moderate/severe/extreme (n = 146) 37 (25%) 75 (32%) 34 (49%) Adj OR for mild pain vs no pain (95% CI)* 1.0 1.4 (0.9, 2.4) 3.1 (1.2, 8.0) p for trend 0.02 Adj OR for moderate/ severe/extreme pain vs no pain (95% CI)* 1.0 1.9 (1.0, 3.3) 4.8 (1.8, 12.3) p for trend 0.001 Subjects KL < 2, no PF OA n = 132 n = 149 n = 16 None 62 (47%) 61 (41%) 5 (31%) Mild/moderate/severe/extreme 70 (53%) 88 (59%) 11 (69%) Adj OR for pain (95% CI)* 1.0 1.4 (0.8, 2.3) 2.2 (0.6, 8.1) p for trend 0.15 *Adjusted for age, sex, BMI and MRI bone marrow edema and effusions