The implications of the adult congenital heart disease anatomic and physiological classification (AP-ACHD) for risk assessment have not been adequately studied. A retrospective cohort study was conducted using data from an ongoing national, multicentre registry of patients with ACHD and atrial arrhythmias (AA) receiving apixaban (PROTECT-AR study, NCT03854149). At enrollment, patients were stratified according to Anatomic class (AnatC, range I to III) and physiological stage (PhyS, range B to D). A follow-up was conducted between May 2019 and September 2021. The primary outcome was a composite of death from any cause, any major thromboembolic event, major or clinically relevant non-major bleeding, or hospitalization. Cox proportional-hazards regression modeling was used to evaluate the risks for the outcome among AP-ACHD classes. Over a median 20-month follow-up period, 47 of 157 (29.9%) ACHD patients with AA experienced the composite outcome. Adjusted hazard ratios (aHR) with 95% confidence intervals (CI) for the outcome in PhyS C and PhyS D were 1.79 (95% CI 0.69 to 4.67) and 8.15 (95% CI 1.52 to 43.59), respectively, as compared with PhyS B. The corresponding aHRs in AnatC II and AnatC III were 1.12 (95% CI 0.37 to 3.41) and 1.06 (95% CI 0.24 to 4.63), respectively, as compared with AnatC I. In conclusion, the PhyS component of the AP-ACHD classification was an independent predictor of net adverse clinical events among ACHD patients with AA.