Association between ischemic stroke and coronary artery disease severity in patients with atrial fibrillation: big data evidence supporting a regional paradox

Introduction Since atrial fibrillation (AF) is strongly associated with ischemic stroke, the CHA2DS2-VASc score is the commonly used scale for stroke risk stratification. However, the conventional angiographic threshold of ≥50% stenosis may not reliably represent meaningful coronary artery disease (CAD), as intermediate stenosis is often discordant with hemodynamic significance. Given that the CHA2DS2-VASc score was derived from a Danish cohort, the definition of CAD used for the vascular component should account for regional variation, particularly in East Asia; in the context of the East Asian Paradox. Purpose This study aimed to validate whether this predictive association is optimal in East Asian population. Methods We identified 27,964 patients who underwent cardiac computed tomography or invasive coronary angiography between 2006 and 2025. From this cohort, a total of 2,898 patients were included who had AF without history of myocardial infarction, angina and previous percutaneous coronary intervention. The primary endpoint was the ischemic stroke incidence within 3 years after the baseline, classified by coronary artery stenosis (CAS) severity: None (0%), mild (1–49%), moderate (50–69%), and severe (≥70%). To reduce confounding, inverse probability weighting was applied, and the risks were estimated using Cox hazards models. Results The 3-year ischemic stroke incidence demonstrated a significant increasing trend across stenosis grades: None 5.6%, mild 4.3%, moderate 7.9%, severe 9.7% (trend p < 0.001). The stroke incidence curves among non-severe groups did not differ significantly (minimum log-rank p = 0.292), whereas the severe group curve diverged significantly (moderate vs. severe: log-rank p < 0.001). Adjusted hazard ratios (HRs) for the primary endpoint were: mild 0.90 (95% CI 0.68-1.18; p = 0.431); moderate 1.16 (95% CI 0.89-1.50; p = 0.275); severe 1.48 (95% CI 1.12-1.95; p = 0.005). Modifying the CAS threshold from ≥50% to ≥70% within the CHA2DS2-VASc framework improved specificity by +4.93 percentage points (from 0.137 to 0.187, a 35.9% relative increase; 95% CI +4.29 to +5.57 pp; p < 0.001), while the sensitivity slightly declined by -2.03 percentage points (from 0.923 to 0.902; 95% CI -3.51 to -0.55 pp; p = 0.013), for the oral anticoagulation recommendation. Conclusion The non-severe CAS (<70%) did not significantly increase the ischemic stroke risk in East Asian patients with AF. Aligning with the East Asian Paradox, where higher bleeding risks emphasizes greater diagnostic specificity, raising the CAS threshold to ≥70% may refine anticoagulation strategies. Further investigation into the net clinical benefit of this modification, explicitly considering bleeding risk, is warranted.CAD_figure