Performance of the Expanded Cardiovascular Risk Prediction Score for Rheumatoid Arthritis Is Not Superior to the ACC/AHA Risk Calculator.

Abstract

Cardiovascular (CV) risk estimation calculators for the general population do not perform well in patients with rheumatoid arthritis (RA). An RA-specific risk calculator has been developed, but did not perform better than a risk calculator for the general population when validated in a heterogeneous multinational cohort. In a cohort of patients with new-onset RA from northern Sweden (n = 665), the risk of CV disease was estimated by the Expanded Cardiovascular Risk Prediction Score for Rheumatoid Arthritis (ERS-RA) and the American College of Cardiology/American Heart Association algorithm (ACC/AHA). The ACC/AHA estimation was analyzed, both as crude data and when adjusted according to the recommendations by the European League Against Rheumatism (ACC/AHA × 1.5). ERS-RA was calculated using 2 variants: 1 from patient and physician reports of hypertension (HTN) and hyperlipidemia [ERS-RA (reported)] and 1 from assessments of blood pressure (BP) and blood lipids [ERS-RA (measured)]. The estimations were compared with observed CV events. All variants of risk calculators underestimated the CV risk. Discrimination was good for all risk calculators studied. Performance of all risk calculators was poorer in patients with a high grade of inflammation, whereas ACC/AHA × 1.5 performed best in the high-inflammatory patients. In those patients with an estimated risk of 5-15%, no risk calculator performed well. ERS-RA underestimated the risk of a CV event in our cohort of patients, especially when risk estimations were based on patient or physician reports of HTN and hyperlipidemia instead of assessment of BP and blood lipids. The performance of ERS-RA was no better than that of ACC/AHA × 1.5, and neither performed well in high-inflammatory patients.