Performance of the new PREVENT score for indicating subclinical atherosclerosis in rheumatoid arthritis.

Abstract

Patients with rheumatoid arthritis (RA) are at increased risk of atherosclerotic cardiovascular disease (ASCVD), yet traditional 10-year CVD risk calculators - such as the 2013 Pooled Cohort Equations (PCE) - underperform in RA. Our objective was to determine the performance of the PCE vs the updated PREVENT algorithm in indicating presence of subclinical coronary and carotid atherosclerosis in RA. Patients with RA without known CVD underwent chest CT to quantify a coronary artery calcium (CAC) score and carotid artery ultrasound to determine presence of carotid artery plaque. The PCE score and three 10-year PREVENT scores - Total CVD, ASCVD and Total CVD with urine albumin-to-creatinine ratio (uACR) - were calculated; scores were compared in their performance to indicate CAC and presence of carotid plaque. The PCE score was significantly different from each of the PREVENT scores (P < 0.001). The PCE and the PREVENT scores were significantly associated with all levels of CAC and carotid plaque. In predicting CAC >0, the PREVENT ASCVD score area under the receiver operating characteristic curve (AUROC) was 0.723, lower than the PCE AUROC of 0.775 (P = 0.034). For carotid plaque, there was no difference in AUROC between the PCE and the PREVENT ASCVD, PREVENT Total CVD score or PREVENT Total CVD + uACR score. The newly updated PREVENT scores are not better than the prior PCE for indicating subclinical coronary and carotid atherosclerosis in RA. Because the PCE score already underperforms in RA patients, it is unlikely that the PREVENT algorithm will adequately predict CVD event risk in this population.