Background Underlying mechanisms leading to pain in hand osteoarthritis (OA) remain unclear. Ultrasonography studies showed that synovial inflammation may contribute to the presence of pain. Magnetic resonance imaging (MRI) can additionally visualize bone marrow lesions (BML), a feature which has shown to be associated with pain, especially in knee OA.
Objectives The objective of this study was to examine the association between MRI features and pain.
Methods Cross-sectional data were used of the ongoing HOSTAS (Hand OSTeoArthritis in Secondary care) study, which includes consecutive patients diagnosed by their treating rheumatologist with primary hand OA. Patients who received a contrast enhanced MRI of the DIP 2-5 and PIP 2-5 joints of the right hand were included for this analysis. During physical examination the number of joints painful upon palpation (0-30) was assessed. Self-reported pain in the right hand was evaluated by the visual analogue scale (VAS) (0-100) and Michigan Hand outcomes Questionnaire (MHQ) (0-100).
MR images were scored following the Oslo hand OA score: synovitis (0-3), BMLs and flexor tenosynovitis (FTS). Additionally, extensor tenosynovitis (ETS) was scored. Due to low numbers, the grades of BMLs and tenosynovitis were merged and resulted in: BMLs (0-2), FTS (0-1), ETS (0-1). The reliability of the scoring was good.
Odds Ratios (OR) with 95% confidence intervals (CI) were calculated using generalised estimating equations (GEE) to associate MRI abnormalities with pain upon palpation in DIP and PIP joints, while adjusting for age, sex, BMI and patient effect. Additionally, ORs with 95% CI were calculated using logistic regression for reporting self-reported pain.
Results 848 joints in 106 patients (83% women, mean age 60.1 yrs, median BMI 26.9 kg/m2 ) were scanned; in 93 patients physical examination was performed within 3 weeks of the MRI. Synovitis was present in 94% of the patients and BMLs in 56%, while FTS (16%) and ETS (31%) were scored less often. The median (range) number of affected joints per patient was 3 (0-8), 1 (0-6), 0 (0-4) and 0 (0-8) for synovitis, BMLs, FTS and ETS, respectively. Synovitis was especially seen in DIP 2-3 and PIP 2-5, BMLs in DIP 2-3 and PIP 2, FTS in PIP 3 and ETS in PIP5. The median (range) number of painful joints upon palpation was 1 (0-8), the median (range) VAS pain 36 (0-83) and MHQ pain 45 (0-95).
GEE were performed in the 93 patients with physical examination. Synovitis (grade 3 vs 0: OR 5.36 (95%CI 2.79-10.30)) and BMLs (2 vs 0: 6.33 (2.92-13.72)) were associated with pain in the joint upon palpation, as FTS and ETS were not. Additional analyses including synovitis and BMLs, showed independent associations for synovitis (3 vs 0: 3.54 (1.91-6.57)) and BMLs (2 vs 0: 3.50 (1.59-7.70)). No association was seen between the number of joints with MRI abnormalities and VAS pain or the MHQ pain subscale.
Conclusions In hand OA patients, synovitis and BMLs on MRI were associated with pain upon palpation, where tenosynovitis was not. Our results suggest a role for inflammation and subchondral bone processes in the pathogenesis of hand OA.
Disclosure of Interest None declared