Background There are conflicting data on outcomes in patients with atrial fibrillation treated with non-vitamin K oral anticoagulants across varying body mass index (BMI). We investigated cardiovascular and bleeding risk in patients with atrial fibrillation with varying BMI. Methods and Results Observational cohort study from the Swedish oral anticoagulation registry between August 2, 2011, and December 31, 2018. Primary outcomes were mortality, ischemic stroke/systemic embolism, and major bleeding. Baseline BMI (kg/m2) was analyzed continuously and categorized: underweight (<18.5 kg/m2), normal weight (18.5 to <25 kg/m2), preobesity (25 to <30 kg/m2), and obesity class I to III (30 to <35, 35 to <40, and ≥40 kg/m2). Adjusted Cox models and nonlinear relationships of BMI were modeled using restricted cubic splines. Non-vitamin K oral anticoagulant-treated patients with atrial fibrillation were included (n=26 047). At baseline, 602 (2.3%) were underweight, 9101 (34.9%) were normal weight, 9970 (38.3%) were preobese, 4280 (16.4%) were obese class I, 1486 (5.7%) were obese class II, and 608 (2.3%) were obese class III. Underweight and obesity class III were in adjusted continuous analysis associated with increased mortality and major bleeding, with lowest risk observed among preobese patients with BMI 28.2 and 26.2 kg/m2, respectively. In adjusted categorical analysis, underweight was associated with increased mortality (hazard ratio [HR], 1.77 [95% CI, 1.57-1.99]) and nonsignificant higher risk of major bleeding (HR, 1.23 [95% CI, 0.95-1.58]). Similarly, obesity class III was associated with increased mortality (HR, 1.67 [95% CI, 1.40-1.99]) and major bleeding (HR, 1.68 [95% CI, 1.29-2.17]). No significant association was observed between higher BMI and ischemic stroke/systemic embolism. Conclusions In non-vitamin K oral anticoagulant-treated patients with atrial fibrillation, there appears to be a U-shaped relationship between BMI and mortality and major bleeding.