POS0380 THE ASSOCIATION OF BONE MARROW LESIONS WITH COLLATERAL LIGAMENT LESIONS IN INTERPHALANGEAL JOINTS; THE HOSTAS COHORT

<h3>Background:</h3> Inflammation or activity at the insertion site of the collateral ligament ("enthesitis") of the finger joints has been hypothesized to influence lesions of the collateral ligaments. However, it is unknown to what extent lesions co-exist, nor is their association over time. More knowledge would aid in understanding the development of hand osteoarthritis (OA) and potentially provide therapeutic targets. <h3>Objectives:</h3> To investigate 1) to what extent bone activity (in the form of bone marrow lesions (BMLs)) adjacent to collateral ligaments and collateral ligament lesions of the PIP and DIP (proximal and distal interphalangeal) finger joints are present over time on magnetic resonance imaging (MRI) in hand OA patients and 2) the longitudinal association between these BMLs with collateral ligament lesions. <h3>Methods:</h3> We used data from the Hand OSTeoArthritis in Secondary care cohort, consisting of patients with primary hand OA. MRI measurements at cohort entry and two- and four-years follow-up were used. Each radial and ulnar site of the proximal and distal interphalangeal joint of digits 2-5 of the right hand (n=16 per patient) was scored 1) for BMLs according to the HOAMRIS scoring method (definition: "A lesion within the trabecular bone with signal characteristics consistent with increased water content (i.e., high signal intensity on STIR images or T2w FS images) and with ill-defined margins"); and 2) for collateral ligament lesion using the OSLO scoring method ("a non-visible or non-continuous collateral ligament on T1w fs images"). BMLs in the ulnar or radial one-third of the joint were regarded "adjacent to the collateral ligament". Intra reader-reliability exceeded 0.95 for both BMLs and collateral ligament lesions. Logistic regression analyses and odds ratios (ORs) with 95% confidence intervals (95%CI) were estimated using generalized estimating equations to quantify the association between BMLs and collateral ligament lesions over time, adjusted for patient effect. <h3>Results:</h3> We included 263 patients (mean age 60.5 years, 84% women, mean BMI 27.5), of whom 213 (81%) had data on year four. At baseline, BML adjacent to a collateral ligament was present for 63 patients (18%), for whom 113/4,206 joint sides (3%) were affected. This proportion was similar at year four (89/3,396 (3%)). Presence of these BMLs fluctuated over time on joint side level (<b>Table 1</b>). Any collateral ligament lesion was present for 129 patients at baseline (49%), for whom 500/4,206 joint sides (12%) were affected. Collateral ligament lesions remained present once there (<b>Table 1</b>), and increased in prevalence over time, with 17% of joint sides affected at year four. A BML adjacent to a collateral ligament lesion was present for 64 joint sides (1%) at baseline (OR adjusted for patient effect 7.3 (95%CI 4.6;11.7) and 64 (2%) at year four (12.7 (95%CI 8.1;20.5) (<b>Table 2</b>)). In joint sides with no collateral ligament lesions at baseline (n=3,706), those with BML at baseline had an adjusted odds ratio of 3.6 (95% CI 1.5;8.9) for having a collateral ligament lesion at year two versus those without a BML at baseline (adjusted for patient effect) (<b>Table 2</b>). For baseline BML versus collateral ligament lesion at year four, an association remained (4.8 (95%CI 2.2;10.4)). In joint sides with no BML at baseline (n=4,142), those with collateral ligament lesion at baseline had an odds ratio of 9.3 (95%CI 3.9;22.2) for having a BML at year two versus those without a collateral ligament lesion at baseline (adjusted for patient effect) (<b>Table 2</b>). For baseline versus year four instead of two, this association increased in strength: 11.1 (95%CI 5.9;21.1). <h3>Conclusion:</h3> BMLs were rare and their presence fluctuated over time on joint side level, while collateral ligament lesions were more common, and remained present during follow-up. A consistent association between these lesions was found during follow-up, and collateral ligament lesions were more associated with future BMLs than vice-versa. Future research should investigate the exact biological processes underlying these associations. <h3>REFERENCES:</h3> <b>NIL.</b> <h3>Acknowledgements:</h3> <b>NIL.</b> <h3>Disclosure of Interests:</h3> Sietse Terpstra grant from Dutch arthritis fund for doing research, all paid to the institution, Lotte A. van de Stadt: None declared, Monique Reijnierse ASAS, ISS Seed grant, Rolf Groenwold: None declared, Frits Rosendaal: None declared, Margreet Kloppenburg grant from Dutch Arthritis Fund, all paid to the institution.