LAUSR

OBJECTIVE To describe cross-sectional and longitudinal associations between magnetic resonance imaging (MRI)-detected osteophytes (OPs) and knee structural abnormalities and knee pain in older adults. METHOD A prospective population-based cohort study of 895 participants aged 50-80 years (mean age 62 years, 50% female) was performed. T1-or T2-weighted fat suppressed MRI was used to assess knee OPs, cartilage volume, cartilage defects and bone marrow lesions (BMLs) at baseline and after 2.6 years. Radiographically-detected OPs were scored according to the Osteoarthritis Research Society International (OARSI) atlas. Knee pain was assessed using a self-administered questionnaire at baseline, 2.6 and 5 years later. RESULTS 85% of participants had MRI-detected OPs at baseline, while 10% of participants had radiographically-detected OPs. Cross-sectionally, higher gardes of MRI-detected OPs in all compartments were significantly, independently and site-specifically associated with higher prevalences of cartilage defects and BMLs, lower cartilage volume and higher prevalence of knee pain. Longitudinally, higher gardes of baseline MRI-detected OPs site-specifically predicted greater risks of any increase in cartilage defects or BMLs, and loss of cartilage volume in medial and lateral tibiofemoral (LTF) and total compartments over 2.6 years in multivariable analyses. These significant associations were similar in those without radiographically-detected OPs. MTF and total OP scores were significantly associated with change in total knee pain over 2.6 and 5 years but these became non-significant after adjustment for cartilage defects and BMLs. CONCLUSION MRI-detected knee OPs are common and appear to be clinically relevant to knee structural changes in older adults.