Background: Atrial Fibrillation (AF) is a major cause of embolic stroke. CHADS2 and CHA2DS2VASc scores (Congestive heart failure, Hypertension, Age, Diabetes, Stroke, Peripheral vascular disease) are used as risk stratification instruments to select anticoagulation therapy. Both scores have been well validated in the general population, however applicability to end-stage renal disease patients is unclear. Previous reports suggest that the CHADS2 score seems to have applicability to this population but the applicability of the CHA2DS2VASc score is unknown. Methods: All incident adult dialysis cases from the United States Renal Data System (USRDS) 2005-2008, were queried for a diagnosis of cerebrovascular accident (CVA), AF or flutter and individual parameters from the CHA2DS2VASc and CHADS2 scores using the International Classification of Diseases-9 (ICD-9) diagnosis codes and data from the Centers for Medicare and Medicaid Services form 2728. The CHA2DS2VASc and CHADS2 score were calculated for each patient before an incident of stroke or up to the last date of service. Logistic regression (LR) was used to estimate the odds ratio of the CHADS2 or CHA2DS2VASc score on stroke, accounting for number of patient-years and controlling for anticoagulation therapy. Results: There were 443,890 incident dialysis patients between 2005 and 2008, 35,147 had a diagnosis of AF or flutter of whom 83% were white, 14% black and 3% other. 42% were female and the average age at their incident service was 75.4 years (SD=9.0). 12% were identified as receiving anticoagulation. 96% were on hemodialysis and 4% on peritoneal dialysis. After adjustment for anticoagulation therapy, the odds ratio for stroke increased by 1.22 for every one-unit increase in CHA2DS2VASc and 1.26 for every one-unit increase in CHADS2. Conclusion: The CHA2DS2VASc and CHADS2 scoring systems predict the risk for stroke in patients with AF or flutter undergoing dialysis. Dialysis patients with AF or flutter have a significant risk for stroke; these two scores have applicability in this population and may be utilized to guide clinicians in the possible need for anticoagulation to decrease the risk of CVA.