Magnetic resonance-detected subchondral bone marrow and cartilage defect characteristics associated with pain and X-ray-defined knee osteoarthritis

Abstract

OBJECTIVE To assess whether the presence of subchondral bone marrow abnormalities (bone marrow edema (BME)) and cartilage defects, determined by magnetic resonance imaging (MRI), would explain the difference between painful osteoarthritis of the knee (OAK) compared with painless OAK or pain without OAK. METHOD Four groups of women (30 per group), aged 35-55 years, were recruited from the southeast Michigan Osteoarthritis cohort (group 1: painful OAK; group 2: painless OAK; group 3: knee pain without OAK; and group 4: no OAK or knee pain). OAK was defined by a Kellgren-Lawrence score of 2 or greater, while pain was based on self-report. BME and cartilage defects were identified from MRI. RESULTS BME lesions were identified in 56% of all knees. BME lesions were four times (95% CI=1.7, 8.7) more likely to occur in the painless OAK group as compared with the group with pain, but no OAK. BME lesions >1cm were more frequent (OR=5.0; 95% CI=1.4, 10.5) in the painful OAK group than all other groups. While the frequency of BME lesions was similar in the painless OAK and painful OAK groups, there were more lesions, >1cm, in the painful OAK group. About 75% of all knees had evidence of some cartilage defect, of which 35% were full-thickness defects. Full-thickness cartilage defects occurred frequently in painful OAK. One-third of knees with full-thickness defects and 47% of knees with cartilage defects involving bone had BME >1cm. Women with radiographic OA, full-thickness articular cartilage defects, and adjacent subchondral cortical bone defects were significantly more likely to have painful OAK than other groups (OR=3.2; 95% CI=1.3, 7.6). CONCLUSION The finding on MRI of subchondral BME cannot satisfactorily explain the presence or absence of knee pain. However, women with BME and full-thickness articular cartilage defects accompanied by adjacent subchondral cortical bone defects were significantly more likely to have painful OAK than painless OAK.