s / Osteoarthritis and Cartilage 22 (2014) S57–S489 S398 matrix remodeling process induced by an inflammatory microenvironment suggested by the presence of synovial hypervascularization. Figure 1. Prolidase protein expression: A) Hipertrophy synovial of patient with OA B) Normal synovium 710 MANUAL LABOR OCCUPATIONS INCREASE THE SHORT-TERM INCIDENCE OF KNEE OSTEOARTHRITIS IN JAPANESE, BUT NOT CAUCASIAN SUBJECTS K.M. Leyland y, M.T. Sanchez-Santos y, S. Muraki z, A. Judge y, T.D. Spector x, D.J. Hart x, J. Newton y, C. Cooper ky, N. Yoshimura z, N.K. Arden yk. yUniv. of Oxford, Oxford, United Kingdom; zUniv. of Tokyo, Tokyo, Japan; x St. Thomas’ Hosp., London, United Kingdom; kUniv. of Southampton, Southampton, United Kingdom Purpose: To determine the interacting role of occupation and obesity on the risk of incident knee osteoarthritis (OA) among Caucasian and Japanese women. Methods: We used data from women in two prospective communitybased cohorts, the UK based Chingford study and the Japanese Research on Osteoarthritis/osteoporosis Against Disability (ROAD) study. Incident radiographic knee osteoarthritis (RKOA) was defined as having Kellgren and Lawrence (K/L) grade 0 or 1 in both knees at baseline and a K/L grade 2 or greater at follow up. Follow-up x-rays were taken after a median of 4 years after baseline for Chingford and 3 years in ROAD. The cut-off for a normal BMI was 25 kg/m2 for the Caucasian subjects and 23 kg/m2 for the Japanese subjects. Occupation was divided into manual versus non-manual (i.e. sedentary/moderate activity) for both populations based on self-assessment. Multivariable logistic regression models assessed of risk of obesity, occupation and their interaction on incident RKOA, controlling for age, hand grip strength, walking distance, smoking and alcohol intake. Results: 12.4% of Caucasian and 14.5% of Japanese women developed incident RKOA in at least one knee. Being overweight/obese was associated with an increased risk of incident RKOA in both populations (Caucasian OR 1.7 [95% CI 1.0, 2.8]; Japanese OR 2.5 [95% CI 1.3, 4.9]), however having a manual labour occupation was only significantly associated with incident RKOA in the Japanese subjects (OR 2.8 [95% CI 1.3, 5.8]). No significant interaction between BMI category and occupation was found for either population (Chingford p1⁄40.070; ROAD p1⁄40.291), however the Japanese population showed a slight increase in risk for manual labour occupations in the overweight/obese category Table 1 Profile of Participants by presence or absence of back pain