THU0094 IN ACPA POSITIVE AT-RISK INDIVIDUALS WITHOUT CLINICAL ARTHRITIS, IS ULTRASOUND SUFFICIENTLY ACCURATE TO PREDICT PROGRESSION TO INFLAMMATORY ARTHRITIS?

Abstract

Background:In a cohort of Anti-Cyclic Citrullinated Peptide Positive (ACPA+) At-Risk of developing inflammatory arthritis (IA) individuals without clinical synovitis, we previously demonstrated the predictive value of Power Doppler for progression depending on the number of joints involved (1). Here we update these results in a larger population combining all ultrasound (US) features and incorporating them in a multivariable analysis with clinical, genetic, immunological and serological markers.Objectives:To investigate the ability of US to predict progression to clinical arthritis using multivariable Cox analysis with and without including other variables.Methods:In a single centre prospective cohort, 488 at risk ACPA+ individuals with new musculoskeletal symptoms underwent an US scan of 30 small joints and 18 tendons at first visit (metacarpophalangeal, interphalangeal and metatarsophalangeal joints and flexor tendons, and extensor carpi ulnaris). The predictive value of US abnormalities (Power Doppler grade ≥ 1 (PD), Grey Scale grade ≥ 2 (GS), or erosion (E) presence) for progression to IA was analysed using Cox regression analysis and adjusted for tenosynovitis (TSV) presence, age, sex, ≥3 ULN CCP2 antibody titre and/or rheumatoid factor titre (RF), early morning stiffness duration, shared epitope (HLA-DRB1*01, *04 and/or *10), number of small joints tender, elevated CRP and intermittent symptoms. A complete dataset considering all variables was available for 324 patients.Results:Consecutive at-risk ACPA+ individuals (n=488, mean age 50.47 years old, 72.9% women) were followed up for at least 24 weeks or up to progression, for a median of 96 weeks (range 0.43-590). 130 of them (26.7%) developed IA after a median of 51.5 weeks (range 0.43 – 486).Multivariable analysis focusing on intra-articular ultrasound features showed that individuals with 1-3 joints with a PD signal or 1-2 with E were twice as likely to develop IA (Table 1), those with ≥ 4 joints with a PD signal were more than six times more likely (Figure 1). All variables data was available in 324 participants, showing a significant predictive value of TSV presence in ≥1 joint (HR= 1.973, p= 0.024, CI= 1.095-3.553), smoking exposure (HR= 2.597, p= 0.003, CI= 1.369-4.929), shared epitope positivity (HR= 1.979, p=0.044, CI= 1.019-3.843), ≥5 small joints tender (HR= 2.111, p= 0.030, CI= 1.073-5.463), and a high titre CCP and/or RF (HR= 4.334, p= 0.003, CI=1.651-11.374).Table 1.Multivariate Cox analysis of joint ultrasound features: non-adjusted for confoundersJoints InvolvedFeatureHRCIP-valuePPVNPVNSmall joints (wrists excluded)PD ≥1:1-3 joints≥4 joints2.1096.5241.364-3.2623.465-12.289<0.0010.001<0.00144%76.5%78.3%488GS ≥2:1-3 joints≥4 joints0.38120.3%30.8%71.9%E presence in ≥1 joint1.8651.206-2.8840.00547.6%76%Conclusion:In at-risk ACPA+ individuals, ultrasound alone - especially Power Doppler- is a powerful predictive factor of progression to IA, therefore of high clinical value for rheumatologists. The multivariable integrated risk prediction values indicates the role of other factors to be considered in future risk model development.