Association between baseline meniscal extrusion on MRI and long-term incident knee osteoarthritis in two different cohorts

Abstract

Purpose: Previously, we identified a significant association between baseline meniscal extrusion and the short-term incidence of knee osteoarthritis (KOA). To validate these findings, we now evaluated the long-term incidence of KOA in knees with baseline meniscal extrusion, using two different cohorts. Methods: We used data from the PROOF study, a preventive RCT which evaluated a high-risk population of 407 middle-aged overweight women (BMI ≥ 27 kg/m2), and a subcohort of the Rotterdam Study (RS), a prospective population-based cohort study. Meniscal extrusion was defined as grade 2 or 3 according to the MOAKS criteria (≥ 3 mm). The primary outcome measure was incident radiographic or clinical KOA, assessed at 6.6 years (PROOF) and 5.1 years (RS). Radiographic incidence was defined as Kellgren and Lawrence (KL) ≥ grade 2 in knees with KL < 2 at baseline. Incident clinical KOA was defined fulfilling the ACR criteria at follow-up, without clinical KOA at baseline. With generalized estimating equations, we determined the association of knees with and without baseline meniscal extrusion and incident KOA, adjusted for the baseline differences. Furthermore, we computed the population attributable risk percentage (PAR%) of meniscal extrusion for both outcomes in both cohorts. Results: In PROOF, 437 knees were available for analysis of which 23% had baseline meniscal extrusion. There were statistically significant baseline differences between knees with and without meniscal extrusion regarding age, history of knee injury, varus alignment, strength quadriceps muscle, Heberden’s nodes, osteophytes, bone marrow lesions (BML) and meniscal pathologies. Incident radiographic KOA after 6.6 years was borderline significantly higher in knees with meniscal extrusion compared to those without (31% vs. 10%, adjusted OR 1.95, 95% CI 0.95, 3.98, PAR 11%). Incident clinical KOA was significantly higher (25% vs. 14%, adjusted OR 2.33, 95% CI 1.19, 4.59, PAR 18%). In the RS cohort, 890 knees were available for analysis of which 46% had baseline meniscal extrusion. There were significant baseline differences regarding BML, osteophytes and meniscal pathologies. Incidence of radiographic KOA after 5.1 years was significantly higher in knees with extrusion (8% vs 2%, adjusted OR 2.54, 95% CI 1.12, 5.76, PAR 30%). Incidence of clinical KOA was borderline significantly higher (6% vs. 2%, adjusted OR 2.05, 95% CI 0.9, 4.65, PAR 31%). Conclusions: Baseline meniscal extrusion is largely independently associated with the long-term incidence of both radiographic and clinical KOA, providing validation for our previous results after short-term follow-up. A high number of incident cases were attributable to meniscal extrusion.