Background: A proportion of patients with embolic stroke of undetermined source have silent atrial fibrillation (AF) or develop AF after the initial evaluation. Better understanding of the risk for development of AF is critical to implement optimal monitoring strategies with the goal of preventing recurrent stroke attributable to underlying AF. The RE-SPECT ESUS trial (Randomized, Double-Blind Evaluation in Secondary Stroke Prevention Comparing the Efficacy and Safety of the Oral Thrombin Inhibitor Dabigatran Etexilate Versus Acetylsalicylic Acid in Patients With Embolic Stroke of Undetermined Source) provides an opportunity to assess predictors for developing AF and associated recurrent stroke. Methods: RE-SPECT ESUS was a randomized, controlled trial (564 sites, 42 countries) assessing dabigatran versus aspirin for the prevention of recurrent stroke in patients with embolic stroke of undetermined source. Of 5390 patients enrolled and followed for a median of 19 months, 403 (7.5%) were found to develop AF reported as an adverse event or using cardiac monitoring per standard clinical care. Univariable and multivariable regression analyses were performed to define predictors of AF. Results: In the multivariable model, older age (odds ratio for 10-year increase, 1.99 [95% CI, 1.78–2.23]; P <0.001), hypertension (odds ratio, 1.36 [95% CI, 1.03–1.79]; P =0.0304), diabetes (odds ratio, 0.74 [95% CI, 0.56–0.96]; P =0.022), and body mass index (odds ratio for 5-U increase, 1.29 [95% CI, 1.16–1.43]; P <0.001) were independent predictors of AF during the study. In a sensitivity analysis restricted to 1117 patients with baseline NT-proBNP (N-terminal prohormone of brain natriuretic peptide) measurements, only older age and higher NT-proBNP were significant independent predictors of AF. Performance of several published predictive models was assessed, including HAVOC (AF risk score based on hypertension, age ≥75 years, valvular heart disease, peripheral vascular disease, obesity, congestive heart failure, and coronary artery disease) and CHA ₂ DS ₂ -VASc (stroke risk score based on congestive heart failure, hypertension, age ≥75 years [doubled], diabetes, previous stroke, transient ischemic attack or thromboembolism [doubled], vascular disease, age 65 to 74 years, and sex category [female]) scores, and higher scores were associated with higher rates of developing AF. Conclusions: Besides age, the most important variable, several other factors, including hypertension, higher body mass index, and lack of diabetes, are independent predictors of AF after embolic stroke of undetermined source. When baseline NT-proBNP was available, only older age and elevation of this biomarker were predictive of subsequent AF. Understanding who is at higher risk of developing AF will assist in identifying patients who may benefit from more intense, long-term cardiac monitoring. Registration: URL: https://www.clinicaltrials.gov ; Unique identifier: NCT02239120.