BACKGROUND: The inflammatory biomarker interleukin-6 (IL-6) is associated with mortality in atrial fibrillation (AF). OBJECTIVE: To investigate if repeated IL-6 measurements improve the prognostication for stroke or systemic embolism, major bleeding, and mortality in anticoagulated patients with AF. METHODS: IL-6 levels by ELISA were measured at study entry and at 2 months in 4830 patients in the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) trial with 1.8 years median follow-up. In the Randomized Evaluation of Long-Term Anticoagulation Therapy (RE-LY) trial, IL-6 was measured at study entry, 3, 6, and 12 months in 2559 patients with 2.0 years median follow-up. Associations between a second IL-6 measurement and outcomes, adjusted for baseline IL-6, clinical variables, and other cardiovascular biomarkers, were analyzed by Cox regression. RESULTS: Median IL-6 levels were 2.0 ng/L (interquartile range [IQR] 1.30-3.20) and 2.10 ng/L (IQR 1.40-3.40) at the two time-points in ARISTOTLE, and, in RE-LY, 2.5 ng/L (IQR 1.6-4.3), 2.5 ng/L (IQR 1.6-4.2), 2.4 ng/L (IQR 1.6, 3.9), and 2.4 ng/L (IQR 1.5, 3.9), respectively. IL-6 was associated with mortality; hazard ratios per 50% higher IL-6 at 2 or 3 months, respectively, were 1.32 (95% confidence interval, 1.23-1.41; P < .0001) in ARISTOTLE, and 1.11 (1.01-1.22, P = .0290) in RE-LY; with improved C index from 0.74 to 0.76 in ARISTOTLE, but not in the smaller RE-LY cohort. There were no consistent associations with second IL-6 and stroke or systemic embolism, or major bleeding. CONCLUSIONS: Persistent systemic inflammatory activity, assessed by repeated IL-6 measurements, is associated with mortality independent of established clinical risk factors and other strong cardiovascular biomarkers in anticoagulated patients with AF.