Abstract Body: Introduction: Intracranial hemorrhage (ICH) is the most severe adverse effect of anticoagulation in atrial fibrillation (AF) patients. Hypertension, diabetes, hyperlipidemia, and chronic kidney disease are well-known cardiovascular risk factors for ICH. However, the relationship between the polygenic profiles (PP) of these risk factors and ICH risk in AF patients on anticoagulation remains unclear. We hypothesize that adverse PP increases the risk of ICH in AF patients on anticoagulation.

Methods: We conducted a prospective genetic association study within All of Us. Participants over 50 with a history of AF treated with apixaban (the most widely used anticoagulant in this population) and no history of ischemic stroke or ICH were included. We calculated the polygenic profile (PP) by combining five standardized polygenic risk scores for systolic blood pressure, type 2 diabetes, low- and high-density lipoproteins, and glomerular filtration rate, along with APOE epsilon 4 and 2 genotypes. These were based on data from recent genome-wide association studies, with APOE genotypes determined by variants rs429358 and rs7412. We categorized PP into three risk groups using a standard approach: favorable (<20%), neutral (20%-80%), and adverse (>80%). The outcome was incident ICH (new intraparenchymal, subdural, or subarachnoid hemorrhage) after apixaban initiation.

Results: A total of 2,088 participants were included in the study (mean age 71 years, 953 [45%] female and 1,743[83%] of European ancestry). After a median follow-up of 2.9 years, 26 participants sustained an ICH (cumulative incidence:1.5%[95%CI:1.00–2.20], Figure 1). Multivariable Cox proportional hazards models showed that when compared to patients with a favorable PP, those with an adverse PP had a more than three-fold increase in the risk of ICH (HR:3.38,95%CI:1.09–10.50,p-trend=0.005). Polygenic information improved the discrimination of risk prediction scores for ICH (c-statistics of 0.68 and 0.75 for models without and with genomic information, respectively (p=0.01, Figure 2).

Conclusions: Our results show that among AF patients on apixaban, an adverse PP for key cardiovascular risk factors significantly increases the risk of ICH compared to those with a favorable PP. Additionally, incorporating PP data enhances the predictive power of clinical prediction scores for ICH. These findings support further research into whether polygenic profiling can improve clinical decision-making in AF patients.