Background Whether meniscal extrusion and bone marrow lesions (BMLs) independently predict the risk of knee osteoarthritis (OA) remains to be clarified. Yet, while prospective cohort studies have shown that both meniscal extrusion and BMLs increase the risk of progressive knee OA, these studies have focused on the medial rather than the lateral knee joint, and examined people with established knee OA. It remains unclear whether these structural abnormalities predate incident knee OA, or whether meniscal extrusion or BMLs independently predict progressive disease.
Objectives The aim of this 72 month longitudinal study was to examine whether compartmental baseline meniscal extrusion and BMLs were independently associated with incident and progressive knee OA, as measured by i) cartilage volume loss, ii) incident or progressive radiographic OA (ROA), and iii) total knee replacement (TKR).
Methods Data was extracted from the Osteoarthritis Initiative (OAI) cohort. Participants were grouped according to the absence (Kellgren-Lawrence (KL) grade ≤1, n=2120) or presence (KL ≥2, n=2249) of ROA. Baseline meniscal extrusion, BMLs and cartilage volume were assessed on magnetic resonance imaging (MRI). Cartilage volume of the tibial plateau was assessed at baseline and 72 months, while ROA was assessed at baseline and 48 months. TKR was assessed at 72 months.
Results In those with ROA, the presence of a baseline meniscal extrusion (independent of BMLs) was associated with accelerated cartilage loss (medial tibia: -2.1%/annum vs. -1.5%; lateral tibia: -2.6%/annum vs. -1.6%; both p<0.001), progressive ROA and TKR (OR range 1.4 to 1.8; 95% CI range 1.1 to 2.9, p≤0.04). The presence of a baseline BML (independent of meniscal extrusion) was associated with accelerated cartilage loss (medial tibia: -2.1%/annum vs -1.6%; lateral tibia: -1.9%/annum vs -1.6%; p≤0.02), progressive ROA and TKR (OR range 1.5 to 2.4; 95% CI range 1.1 to 3.4, p≤0.02). In those with no ROA, a baseline medial meniscal extrusion was associated with accelerated cartilage loss (medial tibia: -2.1%/annum vs. -1.2%, p<0.001) and a baseline medial BML with incident ROA (OR 1.7, 95% CI 1.1 to 2.9, p=0.03).
Conclusions Baseline meniscal extrusion and BMLs predict incident and progressive knee OA and represent important and independent structural targets for patient selection in disease-modifying OA drug (DMOAD) trials.
Disclosure of Interest J. Martel-Pelletier Shareholder of: ArthroLab Inc., A. Teichtahl: None declared, F. Cicuttini: None declared, F. Abram Employee of: ArthroLab Inc., Y. Wang: None declared, J.-P. Pelletier Shareholder of: ArthroLab Inc.