ERS-RA cardiovascular risk prediction varies by event definition in BRASS validation

Cardiovascular disease is more frequent in rheumatoid arthritis (RA), but predicting individual risk remains challenging. Investigators externally validated the expanded risk score for RA (ERS-RA) in the single-center BRASS registry, integrating registry questionnaires with electronic health records (EHR) and Medicare/Medicaid claims. The observational cohort included 1,332 participants aged 20–80 years without prior cardiovascular disease, followed for a mean of 6.8 years. Mean age was 54 years, and 84% were female.

A less restrictive definition identified cardiovascular events in any of the three data sources; a more restrictive definition included events in EHR and claims. Events occurred in 164 participants (12.3%) under the less restrictive definition and 69 (5.2%) under the more restrictive definition. The mean ERS-RA predicted risk was 7.1%, with 424 participants (32%) scoring at least 7.5%. C-statistics were 0.66 (95% CI 0.62–0.71) and 0.80 (95% CI 0.76–0.85), respectively. Calibration was reported as good, with better performance under the more restrictive definition.

These findings support ERS-RA's potential role in cardiovascular risk assessment while highlighting sensitivity to event ascertainment. The single-center, predominantly female cohort limits generalizability. Only the abstract was available, preventing detailed assessment of event definitions and calibration methods; the study does not establish whether using the score improves clinical outcomes.