Incremental prognostic value of renal function for stroke prediction in atrial fibrillation

Background: Renal function has been associated with an increased stroke risk in patients with atrial fibrillation (AF). However, whether renal function incrementally adds to risk prediction in both anticoagulated and non-anticoagulated patients with AF is unclear. Methods: We used data from the Outcomes Registry for Better Informed Treatment of AF (ORBIT-AF)—a national, prospective, outpatient AF registry in patients aged >18 years (2010−2011). The association between baseline renal function and risk of stroke/systemic embolism (SSE) was evaluated in proportional hazards models adjusting for stroke risk score components. We compared discrimination of 2-year outcomes using C-indices and evaluated calibration by comparing event rates in ORBIT-AF to published rates from an external clinical trial population (ROCKET AF) and an observational cohort (ATRIA). Results: Among 9743 patients included in the analysis, the median age was 75 years (interquartile range [IQR] 67–82), 89.5% were white, 43% were female, and 76% were taking oral anticoagulation (OAC). Over a median follow-up of 2.3 years, 214 SSE events occurred (1.00 per 100 patient-years). Continuous creatinine clearance (CrCl) was not associated with SSE risk after adjusting for other clinical factors (components of CHADS₂ or CHA₂DS₂-VASc). Discrimination for predicting stroke (C-index; 95% CI) was similar for R₂CHADS₂ (0.65; 0.61–0.69), CHADS₂ (0.65; 0.61–0.69), and CHA₂DS₂-VASc (0.66; 0.62–0.70). Conclusions and relevance: In a community patient population with AF, renal dysfunction was not independently associated with embolic risk beyond other established risk factors in either OAC-treated or untreated patients. Additional study is needed to identify clinical factors that incrementally add to stroke risk prediction.