P128 Predicting Persistent Inflammatory Arthritis in patients with Palindromic Rheumatism: Development of Scores for Risk Stratification

Abstract

Abstract Background/Aims Palindromic rheumatism (PR) is a recognised precursor of rheumatoid arthritis (RA). Up to 50% of PR patients will progress to persistent inflammatory arthritis (PIA) but risk factors for progression remain unclear. Analysis of this large UK prospective PR cohort has been used to develop the first risk stratification tool to predict progression to PIA. Methods The Leeds PR cohort was recruited from regional primary and secondary care referrals. PR was defined as ‘a confirmed history/physical examination consistent with episodes of joint pain and swelling that returned to normal in the absence of an alternative diagnosis’. Clinical, serological and immunogenetic parameters were assessed at baseline and three monthly for the first year, then six monthly. Progression to PIA was defined as clinical synovitis persisting for >3 weeks as confirmed by a rheumatologist. Prediction models were developed by performing a Cox regression analysis, then applying LASSO to select the most important risk variables (using R). Results 161 PR patients were followed between July 2008 and January 2023. 33% (53/161) developed PIA and 91% of progressors met criteria for RA. 90% of patients were DMARD naive at baseline. Female gender, anti-cyclic citrullinated peptide (anti-CCP) positivity, rheumatoid factor (RF) positivity, age>40, smoking (ever) and a typical interval between attacks of < 1 month were all associated with progression (Table 1). Cox regression analysis showed female gender (HR 2.2CI1.1-4.7, p = 0.03) and anti-CCP>3x upper limit of normal (HR 5.9, CI 2.0-17.4, p = 0.002) were strongly associated with progression. 'Ever’ and 'Two-year’ prediction models were created to calculate risk of developing PIA (Table 1). Patients could be stratified into high, moderate and low risk groups, with progression rates of 53% (34/64), 25% (19/77) and 0% respectively in the ‘Ever’ model and 37% (27/ 73), 16% (11/69) and 0% respectively in the ‘Two-year’ model. Conclusion In patients with PR, female gender, positive anti-CCP, positive RF, age >40, smoking and a typical interval between attacks of < 1 month were predictive of progression to PIA. Combining these factors allows feasible risk stratification in clinical practice. High and moderate risk groups should be monitored in secondary care and considered for interventional trials. Disclosure R. Chowdhury: None. F. Shuweihdi: None. N. Sidhu: None. L. Duquenne: None. L. Garcia-Montoya: None. J. Nam: None. A. Di Matteo: None. K. Harnden: None. D. Sahin-Eroglu: None. P. Emery: Consultancies; BMS, AbbVie, MSD, Pfizer, Novartis, and Roche. Honoraria; Abbvie, Gilead, Lilly, Novartis. Grants/research support; AbbVie, BMS, Lilly, Samsung. K. Mankia: Honoraria; Abbvie, Lilly, UCB, Galapagos. Grants/research support; Gilead, Lilly.