Frequencies of MRI-detected structural pathology in knees without radiographic OA and worsening over three years: How relevant is contralateral radiographic osteoarthritis?

PURPOSE: To test whether radiographically normal knees with contralateral radiographic knee osteoarthritis (ROA) (i.e. ‘early OA model’) exhibit MRI-defined structural tissue pathology to a greater extent and show higher rates of progression compared to knees with bilateral radiographically normal knees without risk factors (‘healthy reference’). METHODS: We included 154 knees from the Osteoarthritis Initiative without ROA (Kellgren-Lawrence = 0), but with definite ROA (Kellgren-Lawrence ≥2) in the contralateral knee, and 78 participants from the OAI healthy reference cohort (without any signs of radiographic OA, knee pain or risk factors in either knee). Effusion-synovitis, Hoffa-synovitis, bone marrow lesions (BMLs), cartilage lesions, meniscus morphology and - extrusion and osteophytes were assessed at year 1 (Y1) and year 4 (Y4). Frequencies of features for both groups at Y1 and rates of worsening from Y1 to Y4 were compared using Fisher's exact test. RESULTS: 69% (early OA model) vs. 46% (healthy reference) had baseline Hoffa-synovitis, 26% vs. 19% effusion-synovitis, 27% vs. 13% femorotibial (FT) BMLs, 77% vs. 50% FT cartilage lesions, 36% vs. 9% meniscal damage, 51% vs. 24% meniscus extrusion, and 92% vs. 74% FT osteophytes. Apart from effusion-synovitis, all differences were statistically significant. For structural worsening, statistically significant differences were observed for FT cartilage (p = 0.03) and FT osteophytes (p = 0.01). CONCLUSION: MRI structural abnormalities are substantially more frequent and more progressive in radiographically normal knees with contralateral osteoarthritis than in 'healthy reference' controls. Compared with published data, they also are more frequent compared to radiographically normal knees “at risk”, without contralateral knee OA.