Epidemiology of meniscus position: associations with knee symptoms and osteoarthritis

Abstract: Meniscal extrusion (ME) in the knee joint is defined as when the peripheral border of the meniscus is substantially located outside the joint margin. Prior studies have reported that ME is associated with meniscal tears, meniscal degeneration, and the presence of knee osteoarthritis (OA). Medial ME of the body of 3 mm or more, as seen on knee Magnetic Resonance Imaging (MRI) has a wide acceptance to be regarded as “pathologic”. However, it is still unclear if ME is associated with knee pain. My aims with this thesis were to: i) study the normal ME with its change over time and its relationship with meniscus tear/damage in knees without OA; ii) to scrutinize the widely accepted 3 mm cut-off for “pathological” medial ME; iii) determine an optimal cut-off for ME associated with radiographic knee OA, bone marrow lesions (BMLs) and cartilage damage; and iv) investigate the association between medial ME and pain in knees without radiographic OA.In paper I we used 118 subjects from the Osteoarthritis Initiative (OAI) “non-exposed” reference cohort (aged 45-79 years, free of knee pain, radiographic knee OA and risk factors for knee OA) and in papers 2-4 1004 subjects from the community based Framingham Osteoarthritis Study (aged 50–90 years, selection not made on the basis of knee problems). MRI´s were read for ME, cartilage coverage, BMLs and cartilage damage. Knee x-rays were read according to the Kellgren and Lawrence scale. I estimated changes in ME over 4 years using repeat knee MRI. I evaluated the 3 mm cut-off and estimated a new cut-off for “pathologic” extrusion. The odds ratio (OR) for pain as outcome, with ME as exposure was estimated and adjusted for age, sex, body mass index, meniscal tear, BMLs, cartilage damage and previous knee injury as confounders.In the OAI-cohort we found only minor increase in medial ME over 4 years. In the Framingham cohort ME was on average 2.7 mm medially and 1.8 mm laterally. Cartilage coverage was about 30% of ipsilateral cartilage surface. Meniscal damage was associated with more ME medially. A new estimated 4 mm cut-off maximizes the sum of sensitivity and specificity for the OA features radiographic OA, BML and cartilage damage. The OR for knee pain was 1.15 (95% confidence interval 0.97, 1.37) per 1 mm more ME, indicating only a weak association with pain.In conclusion my thesis delivers an array of novel normative data for meniscus body position on MRI, as well as its associations with meniscal damage and knee OA. Medial ME is strongly associated with knee OA. The cut-off value 4 mm may be a more optimal cut-off to use than 3 mm to be considered ”pathologic”. Further, medial ME per se seems to have a weak association with pain in knees free of radiographic OA..