Background Previous studies have shown that synovitis is more common in erosive vs. non-erosive hand OA, but it is unknown whether a higher prevalence is explained by more structural damage. If erosive hand OA represents a separate more inflammatory phenotype, inflammation should possibly play a more important role in disease progression in erosive vs. non-erosive hand OA.
Objectives First, we explored whether synovitis (irrespective of radiographic damage) and radiographic progression were more common in erosive vs. non-erosive hand OA patients. Secondly, we explored whether synovitis was equally associated with radiographic progression in erosive and non-erosive hand OA patients.
Methods In total 26 and 39 patients from the Oslo hand OA cohort (59 women, mean age 68) were classified as non-erosive and erosive based on absence/presence of ≥1 radiographic erosions. The 2nd-5th interphalangeal joints were screened for synovitis at baseline using contrast-enhanced 1.0T MRI of dominant hand and ultrasound-detected grey-scale (GS) synovitis and power-Doppler (PD) activity of both hands (0-3 scales). Bilateral hand radiographs were obtained at baseline and 5-year follow-up. First, we explored whether erosive hand OA status was associated with synovitis irrespective of radiographic OA using Generalized Estimating Equations. Secondly, we explored whether erosive hand OA status was associated with radiographic progression (increasing osteophytes, joint space narrowing or erosions) independent of baseline synovitis. In the same model, we added an interaction term between erosive hand OA status and synovitis, and we repeated the analyses for erosive and non-erosive hand OA separately. Analyses were adjusted for age and sex.
Results Synovitis was most common in erosive hand OA. The median (IQR) sum score of MRI synovitis was 5 (2.5-7) and 7 (4-10) in non-erosive and erosive hand OA patients, respectively (0-24 scale). The difference was even larger for GS synovitis with median (IQR) sum scores of 1 (1, 4) and 6 (3, 10), respectively (0-48 scale). PD activity was infrequently present in erosive hand OA only. Radiographic progression was also significantly more common in erosive hand OA patients. Twice as many joints showed progression in erosive vs. non-erosive hand OA patients (31.5% vs. 15.8%).
Radiographic severity, but not erosive hand OA status, was associated with synovitis when both variables were included in same model (Table). Both erosive hand OA status and synovitis at baseline predicted radiographic progression independent of each other (Table). A significant interaction was found between GS synovitis grade 2-3 and erosive hand OA only (OR 12.55, 95% CI 2.39-65.86). In stratified analyses, a clear dose-response association was found between synovitis and progression in erosive hand OA, whereas the pattern was less clear for non-erosive hand OA (data not shown).
Conclusions Higher prevalence of synovitis in erosive hand OA is related to more structural damage in these patients. Disease progression is higher in erosive hand OA irrespective of more synovitis. Synovitis seems to be of importance for disease progression in both erosive and non-erosive hand OA, although the associations seem to be somewhat stronger for erosive hand OA.
Disclosure of Interest None declared