Ab0874 Combined Use of Score2 and Qrisk3 Increases Identification of Ankylosing Spondylitis Patients at High Cardiovascular Risk: Results From the Carma Project Cohort After 7.5 Years of Follow-Up

Background: Ankylosing Spondylitis (AS) patients are at increased risk for cardiovascular (CV) disease. The Systematic Coronary Risk Evaluation (SCORE) has been used to estimate CV risk over the years. The EULAR group recommended the use of the modified EULAR SCORE 2015/2016 (mSCORE) for patients with rheumatic diseases. The SCORE was recently updated to a new prediction model (SCORE2). Another index, the QRESEARCH risk estimator version 3 (QRISK3), calculates the risk of presenting a fatal and non-fatal CV event at 10 years. There is great concern that indices developed for the general population may underestimate CV risk in patients with rheumatic diseases. Of note, in patients with rheumatoid arthritis (RA) a study showed that the QRISK3 was superior to the mSCORE in identifying patients at high CV risk¹. Another work disclosed that the combined use of the QRISK3 and the mSCORE made it possible to identify patients with RA and high risk of CV events². Objectives: To establish the predictive value of the QRISK3 in identifying patients with AS at high risk of CV events and mortality. To determine whether to combine QRISK3 with another CV risk algorithm; SCORE, mSCORE, or SCORE2 may increase identification of patients with AS and high-risk CV disease. Methods: The information of 684 patients with AS from the Spanish prospective project CARdiovascular in ReuMAtology (CARMA) without history of CV events, who were followed up in rheumatology outpatient clinics of tertiary centers for 7.5 years, was review. Risk chart algorithms were retrospectively tested using baseline data. Survival analysis was performed in two ways: Firstly, we considered as failure the compound of CV event or death due to any cause. Secondly, we considered as failure any CV event, with death due to any non-CV cause as competitive risk. The discriminative ability of any CV scale to identify failures, and the association of CV scale with CV event was studied with Cox regression. The results of the survival analysis were presented as hazard ratios (HR) with their 95% confidence intervals and two-tailed p values. We built seven models: and combinations of the QRISK3 with a scale from the "SCORE family". To compare all models obtained in each survival analysis, we estimated the Information Criteria: Akaike (AIC) and Bayesian (BIC). Results: 33 patients with AS experienced CV events. All four scales were associated with CV events (p<0.001), with HR [1.05 (1.03, 1.07); 1.09 (1.06, 1.12); 1.06 (1.04, 1.08); and 1.15 (1.10, 1.19)] for QRISK3, SCORE, mSCORE and SCORE2, respectively. QRISK3 correctly discriminated between people with lower and higher risk although the percentage of events accumulated in 7.5 years was lower than the expected according to the risk established by QRISK3. SCORE correctly discriminated between people with lower and higher risk. After 7.5 years of follow-up, the cumulative hazard about equals the expected percentage of events. mSCORE showed the same discrimination ability than SCORE, although the percentage of events predicted was higher than the percentage that occurred. SCORE2 had strong discrimination ability according to the CV risk. For people with SCORE2=5%, 10% or 15%, the percentage of events occurred by year 7.5 of follow-up was about the same that was predicted. Combining QRISK3 with any other scale improved the model. This was especially true for the combination of QRISK3 and SCORE2 which achieved the lowest AIC (406.70) and BIC (415.66), so this combination would be the best predictive model. Conclusion: In AS patients from the CARMA project the four algorithms tested accurately discriminated those patients with higher CV risk and those with lower CV risk. A model that includes QRISK3 and SCORE2 combined the best discrimination ability of QRISK3 with the best calibration of SCORE2. REFERENCES: [1] Corrales A, et al. Carotid plaques as predictors of cardiovascular events in patients with Rheumatoid Arthritis. Results from a 5-year-prospective follow-up study. Seminars in Arthritis and Rheumatism. 2020;50(6):1333-8. [2] Corrales A, et al. Combined use of QRISK3 and SCORE as predictors of carotid plaques in patients with rheumatoid arthritis. Rheumatology (Oxford). 2021;60(6): 280-7. Acknowledgements: To the Spanish collaborative CARMA group for the study of the cardiovascular disease in inflammatory arthritis. Disclosure of Interests: None declared.