Polygenic Risk Factor Profiling and Risk of Intracerebral Hemorrhage In Patients with Atrial Fibrillation on Apixaban
Santiago Clocchiatti-Tuozzo1, Cyprien Rivier2, Shufan Huo2, Emily Gilmore5, Ashkan Shoamanesh6, Hooman Kamel7, Santosh Murthy8, Lucila Ohno-Machado3, Kevin Sheth9, Thomas Gill4, Guido Falcone10
1Neurology, Yale University, Department of Neurology, 2Yale University, 3Department of Biomedical Informatics and Data Science, 4Internal Medicine, Yale University, 5Yale University School of Medicine, 6McMaster University, 7Neurology, Weill Cornell Medical College, 8Weill Cornell Medicine, 9Yale University Division of Neuro and Critical Care, 10Yale School of Medicine
Objective:
To evaluate whether a combined polygenic risk factor profile (PRFP) increases the risk of intracranial hemorrhage (ICH) in atrial fibrillation (AF) patients on apixaban.
Background:
Apixaban is the most widely used anticoagulant in AF patients, and ICH is its most severe adverse effect. Hypertension, diabetes, dyslipidemia, and chronic kidney disease are well-known risk factors for ICH. However, the relationship between their combined PRFP and ICH risk in AF patients on apixaban remains unclear.
Design/Methods:
We conducted a prospective genetic association study within All of Us, including participants >50y with AF, taking apixaban, and without prior history of ischemic stroke or ICH. PRFP was calculated using five polygenic risk scores (for systolic blood pressure, type 2 diabetes, LDL and HDL, and glomerular filtration rate), as well as APOE epsilon 4 and 2 genotypes, ascertained using variants rs429358 and rs7412. PRFP was categorized into favorable (<20%), neutral (20%-80%), and adverse (>80%) risk. The outcome was incident ICH (intraparenchymal, subdural, or subarachnoid hemorrhage) after apixaban initiation.
Results:
In 2,088 participants (mean age 71y, 45% female) with 2.9 years of median follow-up, 26 cases of ICH occurred (cumulative incidence: 1.5% [95%CI: 1.00–2.20]). Participants with an adverse PRFP had more than three times the risk of ICH compared to those with a favorable PP (HR: 3.38, 95%CI: 1.09–10.50, p-trend=0.005). Incorporating polygenic information improved the predictive accuracy of clinical scores for ICH, with c-statistics rising from 0.68 to 0.75 (p=0.01).
Conclusions:
Among AF patients on apixaban, adverse PRFPs for the most important risk factors significantly increase the risk of ICH compared to favorable PRFPs. Additionally, incorporating PRFP data enhances the predictive power of clinical prediction scores for ICH. Future studies should explore whether PRFP can improve clinical decision-making in AF patients.
Disclaimer: Abstracts were not reviewed by Neurology® and do not reflect the views of Neurology® editors or staff.