Renal Disease and Stroke Risk in Atrial Fibrillation

Abstract & Commentary

By John P. DiMarco, MD, PhD, Professor of Medicine, Division of Cardiology, University of Virginia, Charlottesville.

Source: Piccini JP, et al. Renal dysfunction as a predictor of stroke and systemic embolism in patients with nonvalvular atrial fibrillation validation of the R2 CHADS2 index in the ROCKET AF (Rivaroxaban Once-daily, oral, direct factor Xa inhibition Compared with vitamin K antagonism for the prevention of stroke and Embolism Trail in Atrial Fibrillation) and ATRIA (AnTicogulation and Risk factors In Atrial fibrillation) study cohorts. Circulation 2013;127:224-232.

This study examines the influence of renal dysfunction on stroke risks in two atrial fibrillation (AF) study groups. The ROCKET AF trial was a trial comparing fixed-dose rivaroxaban, a direct factor Xa inhibitor, with adjusted-dose warfarin for prevention of thromboembolic events. Clinical variables were collected at study enrollment. A total of 14,264 patients were randomized. The median patient age at entry was 74 years with a median CHADS2 score of 3.0. During median follow-up of almost 2 years, 4% of the population (575 patients) experienced a thromboembolic endpoint. Data on creatinine clearance were then combined with the CHADS2 and CHADS2 VASC scores and found to be a significant predictor of stroke or systemic embolism. The strength of association of creatinine clearance with stroke was secondary only to that of prior stroke or TIA. When creatinine clearance was examined as a continuous variable, the hazard ratio for a thromboembolic event increased 12% per 10 mL per minute decrease in creatinine clearance. A R2 CHADS2 risk score was then calculated with an additional 2 points ascribed for a creatinine clearance < 60 mL per minute. In the ROCKET AF cohort, the R2 CHADS2 model was associated with a C statistic of 0.587 compared with 0.575 for the CHADS2 score alone and 0.578 for the CHA2DS2 VASC score. A model that included only creatinine clearance < 60 mL per minute and prior stroke or TIA had a C statistic of 0.590. These models were then applied to data from the ATRIA study cohort. The ATRIA cohort is an epidemiologic study of 13,559 adult patients with nonvalvular AF who were followed from 1996-2003 at Kaiser Permanente of northern California. In the ATRIA study cohort, the R2 CHADS2 risk score had a similar C statistic (discriminant power) to the CHADS2 score but net stroke risk reclassification improved 17.4%. The authors conclude that impaired renal function is a potent predictor of stroke and should be included in risk stratification schemes for patients with AF.


Other studies have shown that moderate-to-severe renal dysfunction increases stroke risk in patients with AF. Patients with severe renal disease were excluded from ROCKET AF, so the observations here apply mainly to those with mild-to-moderate disease. As always, physicians prescribing anticoagulants must also consider the risk of bleeding and renal disease as a risk factor for bleeding. Therefore, the net gain by adding and “R” to CHADS2 may not be very great. We must also remember that the new oral anticoagulants are cleared by the kidney and that leaves warfarin as the only option for patients with advanced renal disease.