Extended Data Fig. 3: Graphical description of the study design.

Graphical description of the study design, visualizing temporal anchors of exposure (new-onset postoperative atrial fibrillation), outcome/follow-up (ischemic stroke/transient ischemic attack), covariates (baseline demographics, procedure related factors, comorbidities and preexisting medication) and primary effect modifier (oral anticoagulation). The cohort entry date (CED), serving as the primary anchor date for patients entering the study analysis, was the date of the patients' index surgery (Day 0). We excluded patients with preexisting atrial fibrillation, age below 18 years, ASA physical status greater than 4 or missing data for exposure, outcome or covariates. Covariates were assessed during a window extending one year prior to the CED if not otherwise specified [-365,-1]. The time window for the primary exposure - postoperative atrial fibrillation (POAF) - was between day 0 and the end of day post-CED. Oral anticoagulation prescriptions were included between postoperative day 1 and the end of day 365. The outcome ischemic stroke was assessed between day 31 and 365 after surgery. Abbreviations: y, years; ASA, American Society of Anesthesiologists; POAF, new-onset postoperative atrial fibrillation.