Renal Dysfunction as a Predictor of Stroke and Systemic Embolism in Patients with Nonvalvular Atrial Fibrillation: Validation of the R2CHADS2 Index in the ROCKET AF and ATRIA Study Cohorts

Jonathan P. Piccini; Susanna R. Stevens; YuChiao Chang; Daniel E. Singer; Yuliya Lokhnygina; Alan S. Go; Manesh R. Patel; Kenneth W. Mahaffey; Jonathan L. Halperin; Günter Breithardt; Graeme J. Hankey; Werner Hacke; Richard C. Becker; Christopher C. Nessel; Keith A. A. Fox; Robert M. Califf

CIRCULATIONAHA.112.107128 Published online before print December 3, 2012

Abstract

Background—We sought to define the factors associated with the occurrence of stroke and systemic embolism in a large, international atrial fibrillation (AF) trial.

Methods and Results—In ROCKET AF, 14,264 patients with nonvalvular AF and creatinine clearance (CrCl) ≥30 mL/min were randomized to rivaroxaban or dose-adjusted warfarin. Cox proportional hazards modeling was used to identify factors at randomization independently associated with the occurrence of stroke or non-central nervous system (CNS) embolism based on intention-to-treat analysis. A risk score was developed in ROCKET AF and validated in ATRIA, an independent AF patient cohort. Over a median follow-up of 1.94 years, 575 (4.0%) patients experienced primary endpoint events. Reduced CrCl was a strong, independent predictor of stroke and systemic embolism, second only to prior stroke or transient ischemic attack (TIA). Additional factors associated with stroke and systemic embolism included elevated diastolic blood pressure and heart rate, and vascular disease of the heart and limbs (C-index 0.635). A model including CrCl (R2CHADS2) improved net reclassification index (NRI) by 6.2% when compared with CHA2DS2VASc (C-statistic=0.578) and 8.2% when compared with CHADS2 (C-statistic=0.575). The inclusion of estimated glomerular filtration rate <60 and prior stroke or TIA in a model with no other covariates led to a C-statistic of 0.590. Validation of R2CHADS2 in an external, separate population improved NRI by 17.4% (95% CI 12.1-22.5%) relative to CHADS2.

Conclusions—In patients with nonvalvular AF at moderate to high risk of stroke, impaired renal function is a potent predictor of stroke and systemic embolism. Stroke risk stratification in patients with AF should include renal function.