CHADS, CHADS-VASc and RCHADS scores predict mortality in patients with coronary artery disease.

Few studies to date address the predictive ability of CHADS-VASc and RCHADS in CAD patients. Our aim is to investigate the prognostic performance of CHADS, CHADS-VASc and RCHADS scores in patients with coronary artery disease (CAD). Angiographically obstructive CAD patients were enrolled. The prognostic performance of the three risk scores was evaluated using Cox hazards models. In addition, we compared their predictive values by calculating C statistics, net reclassification improvement (NRI) and integrated discrimination improvement (IDI). The endpoints are death from any cause and cardiovascular death. Of 3295 subjects with CAD, the mean CHADS, CHADS-VASc and RCHADS scores are 1.2 ± 1.0, 2.4 ± 1.4, and 1.6 ± 1.4, respectively. The CHADS-guided risk classification is markedly distinct from CHADS--VASc- and RCHADS-guided ones. Over a median follow-up of 24 months, a total of 290 (rate 4.00/100 person-year) deaths occurred, and 163 (rate 2.2/100 person-year) were attributed to cardiovascular deaths. Event rates increase by CHADS, CHADS-VASc and RCHADS (P for trend <0.001). The multivariate analyses show 60, 111 and 82% higher risk of mortality per unit increase of CHADS, CHADS-VASc and RCHADS scores, respectively. Comparing with CHADS score (c-statistic = 0.61), CHADS-VASc (c-statistic 0.65, NRI 0.52 and IDI 0.06, P for all <0.05) and RCHADS (c-statistic 0.66, NRI 0.43 and IDI 0.09, P for all <0.05) scores provide better discrimination and reclassification for mortality. Also, CHADS-VASc and RCHADS have comparable predictive ability of mortality to the GRACE score. The CHADS, CHADS-VASc and RCHADS scores are simple yet robust prognostic tools in CAD patients.