IMAGING-DETECTED FEATURES OF HAND OSTEOARTHRITIS ASSOCIATE WITH SYMPTOMS AND RADIOGRAPHIC CHANGE OVER TIME: A SYSTEMATIC REVIEW AND META-ANALYSIS OF OBSERVATIONAL STUDIES

Background: Osteoarthritis (OA) commonly affects joints in the hand. The natural history of hand OA is not well understood, and the local determinants of symptoms and structural changes over time remain unclear. Objectives: To investigate, in both cross-sectional and prospective studies, the association between imaging (ultrasound [US] and magnetic resonance imaging [MRI]) features and symptoms of hand OA, and to examine in prospective studies whether imaging-detected features at baseline predict subsequent clinical and radiographic outcomes. Methods: A systematic literature search was conducted in five databases including Medline, Web of Science, EMBASE, CINAHL and AMED in April 2018. The search was designed to capture published observational studies on the use of US and MRI in hand OA with no language restrictions. Odds ratios (OR), risk ratios (RR), and 95% confidence interval (CI) between [1] imaging features and hand OA symptoms at baseline, and [2] baseline-imaging features and follow-up outcomes were extracted and pooled using random effects model. Outcomes were defined as either incidence or progression of pre-existing features. Risk of bias assessment was performed using the Newcastle-Ottawa Scales. Heterogeneity and publication bias were assessed. Results: The search identified 2818 citations, which reduced to 2216 after duplicate removal. Screening of titles and abstracts found 140 articles which met the inclusion criteria. After full text screening, 25 were included for analysis, including 452 participants (87% women) for US and 298 participants (86% women) for MRI with mean ages 60.3 and 62.5, respectively. Imaging-detected structural OA features were preferentially found in distal interphalangeal joints (DIPJs) followed by carpometacarpal (CMCJ) and proximal interphalangeal (PIPJ) joints. Metacarpophalangeal joints were least affected. However, the distribution pattern was different for inflammatory features for which the CMCJ was the most affected, and with no clear difference between DIPJs and PIPJs (Figure 1). Of 10 US and 5 MRI studies examining association at baseline, joint tenderness was associated with US osteophytes (pooled ORs 2.30, 95% CI 1.90-2.79), grey-scale synovitis (3.00, 2.33-3.84), synovial effusion (2.92, 2.29-3.72), and power Doppler (PD) (2.30, 1.68-3.15). Similar relationships were observed with MRI features (Figure 2). Six studies did not find any association between imaging features and self-reported outcomes. However, association was observed with US- and MRI-detected synovitis in one study each, and MRI-detected structural features in two. Statistical pooling was not possible for these outcomes due to heterogeneous data. Of the 9 US and 5 MRI studies for prediction, a dose-dependent relationship was observed between baseline PD and radiographic change at follow-up (Figure 3). Similar results were observed for MRI features and Kellgren-Lawrence change. The pooled ORs (95% CI) was 2.66 (1.88, 3.78) for bone marrow lesions, and 2.18 (1.53, 3.10) and 4.7 (3.08, 7.18) for grades 1 and 2 synovitis, respectively. Data to predict change in clinical outcomes however, were lacking. Conclusion: Imaging-detected inflammatory features and osteophytes associate with joint tenderness. In addition, imaging-detected inflammatory changes at baseline predict future development and progression of structural OA changes, indicating that inflammation may precede radiographically-detectable structural changes. Disclosure of Interests: Abasiama Obotiba: None declared, Subhashisa Swain: None declared, Jaspreet Kaur: None declared, Khalid Yaseen: None declared, Michael Doherty Grant/research support from: AstraZeneca funded the Nottingham Sons of Gout study, Consultant of: Advisory borads on gout for Grunenthal and Mallinckrodt, Weiya Zhang Consultant of: Grunenthal for advice on gout management, Speakers bureau: Bioiberica as an invited speaker for EULAR 2016 satellite symposium, Abhishek Abhishek Grant/research support from: AstraZeneca and OxfordImmunotech, Speakers bureau: Menarini pharmaceuticals