Abstract 16629: Comparison of Clinical and Genetic Risk Model for Ischemic Stroke After Atrial Fibrillation Catheter Ablation

Abstract

Introduction: CHA 2 DS 2 -VASc score (CVS) is used to estimate risk of stroke in atrial fibrillation (AF) patients. But, CVS does not include genetic risk. We explored clinical and genetic risks for ischemic stroke (IS) after AF catheter ablation (AFCA). Methods: We used genome-wide association study (GWAS) data in European (34217 IS, 406111 controls) and Asian (22664 IS, 152022 controls) to develop polygenic risk score (PRS) for IS in 2897 independent Korean AF patients (Median 60 years, male 74.0%, paroxysmal AF 66.3%, anticoagulation rate 60% in 1367 patients with CVS ≥ 2 who underwent AFCA. Patients with IS within 2 weeks after AFCA were excluded. In the independent AF cohort during median 50 [IQR 28-94] months follow-up, 48 experienced IS after AFCA. We developed two different CVS-combined PRS models in the independent cohort using European and Asian GWAS data. Using p-value <5e-8, rs635634 and rs671 were used in CVS-combined European PRS (CVS-EPRS) and CVS-combined Asian PRS (CVS-APRS) model, respectively in the independent cohort. We compared C index of risk models to investigate prediction power for IS. Results: CVS was higher in patients with IS than in their counterparts (3.0 IQR [1.0, 4.0] vs. 1.0 [1.0, 2.0], p<0.001). Overall C index was 0.716 in CVS model, 0.711 in CVS-EPRS model, and 0.725 CVS-APRS model. For 1-year IS after AFCA, CVS-APRS model improved C index as compared to CVS model (0.834 vs. 0.813, p=0.037). For 3-year IS after AFCA, C index was higher in CVS-APRS model rather than CVS-EPRS model (0.746 vs. 0.721, p=0.030). In patients with CVS < 2, both CVS-APRS (0.849, p=0.013) and CVS-EPRS (0.766, p=0.027) models showed better prediction for 1-year IS compared to CVS model (0.523). Conclusions: In AF patients with AFCA, genetic polymorphism with clinical risk factors contributed to predict 1-year IS. Genetic contribution was more obvious in AF patients with CVS <2. However, ethnicity might be an important factor in adapting genetic contributions to predicting IS.