Combining Imaging and Genetics to Predict Recurrence of Anticoagulation-Associated Intracerebral Hemorrhage

Supplemental Digital Content is available in the text. Background and Purpose: For survivors of oral anticoagulation therapy (OAT)–associated intracerebral hemorrhage (OAT-ICH) who are at high risk for thromboembolism, the benefits of OAT resumption must be weighed against increased risk of recurrent hemorrhagic stroke. The ε2/ε4 alleles of the apolipoprotein E (APOE) gene, MRI-defined cortical superficial siderosis, and cerebral microbleeds are the most potent risk factors for recurrent ICH. We sought to determine whether combining MRI markers and APOE genotype could have clinical impact by identifying ICH survivors in whom the risks of OAT resumption are highest. Methods: Joint analysis of data from 2 longitudinal cohort studies of OAT-ICH survivors: (1) MGH-ICH study (Massachusetts General Hospital ICH) and (2) longitudinal component of the ERICH study (Ethnic/Racial Variations of Intracerebral Hemorrhage). We evaluated whether MRI markers and APOE genotype predict ICH recurrence. We then developed and validated a combined APOE-MRI classification scheme to predict ICH recurrence, using Classification and Regression Tree analysis. Results: Cortical superficial siderosis, cerebral microbleed, and APOE ε2/ε4 variants were independently associated with ICH recurrence after OAT-ICH (all P<0.05). Combining APOE genotype and MRI data resulted in improved prediction of ICH recurrence (Harrell C: 0.79 versus 0.55 for clinical data alone, P=0.033). In the MGH (training) data set, CSS, cerebral microbleed, and APOE ε2/ε4 stratified likelihood of ICH recurrence into high-, medium-, and low-risk categories. In the ERICH (validation) data set, yearly ICH recurrence rates for high-, medium-, and low-risk individuals were 6.6%, 2.5%, and 0.9%, respectively, with overall area under the curve of 0.91 for prediction of recurrent ICH. Conclusions: Combining MRI and APOE genotype stratifies likelihood of ICH recurrence into high, medium, and low risk. If confirmed in prospective studies, this combined APOE-MRI classification scheme may prove useful for selecting individuals for OAT resumption after ICH.