BackgroundCurrently a number of risk factors (RF) are considered to be responsible for radiological progression of knee osteoarthritis (OA), nevertheless key predictors of OA progression have not yet been established.ObjectivesTo identify RF predicting radiological progression of knee joint osteoarthritis (OA) in a 5 year multicenter prospective study.MethodsThis study of RF predicting knee OA progression was the first with multicenter prospective design ever conducted in Russia. The study included 344 female patients 40–75 y.o with primary stage I-III knee OA (ACR criteria) from 6 centres. Radiological stage was identified by Kellgren J.- Lawrence J. grading scale. The follow up (FUP) was 5 years. Individual patients’ files described 90 parameters. Instrumental diagnostic methods included plain radiography of knee joints, dual energy X-ray absorptiometry (DEXA) of lumbar spine L1–4, femoral neck and subchondral tibia, ultrasound (US) and MRI examination of knee joints. OA progression was verified based on evolution of radiological stage. At baseline 24 pts (7%) had stage I OA, 227 (66%) – stage II, and 93 (27%) – stage III. Discriminant analysis was applied to verify most reliable RF predicting radiological progression.ResultsRadiological progression was documented in 45% participants during 5 year FUP. The groups with and without progression were comparable in terms of age and disease duration (»<0,05). Pts who progressed suffered more intensive knee pain – 68(52–72) vs 41(30–63) mm, » <0,01, had higher body weight – 82(77–93) vs 72(65–81) kg, » <0,01, had higher rates of knee synovitis (US) 44% vs 26%, »=0,03, (RR=1,67, 95% CI 1,07–2,59) and mid-tibia bone marrow oedema – 60% vs 28%, » <0,01 (RR=2,12, 95% CI 1,34–3,35). The discriminant analysis showed that knee pain, excessive body weight, synovitis and mid-tibia bone marrow oedema (MRI) can be considered as predictors of OA radiological progression. A model capable of predicting OA course in an individual patient with high 88% accuracy, 87.7% sensitivity and 70% specificity has been developed based on identified RF and their coefficients. Area under the ROC-curve 0, 921 (95% CI 0,875–0,966).ConclusionsKnee pain, excessive body weight, synovitis and bone marrow oedema should be considered as key RF predicting knee OA radiological progression.Disclosure of InterestNone declared