Seizures are common after intracerebral hemorrhage (ICH) and are associated with worse outcomes. Risk factors include hematoma size, younger age, and lobar location. It is unknown if primary-ICH etiologies differ in seizure risk. We evaluated the association of primary-ICH etiologies: cerebral amyloid angiopathy (CAA) and hypertension with seizures after ICH.

Baseline demographics, ICH characteristics/etiology, seizures, and medical/surgical treatment data were prospectively collected between 2009-2015 for ICH patients admitted to Columbia University Medical Center. Primary-ICH patients on therapeutic anticoagulants were excluded. Modified Boston Criteria and SMASH-U criteria identified CAA-ICH and hypertension-ICH respectively. Probable/definite CAA-ICH and possible CAA-ICH were evaluated separately.  Logistic regression assessed the association of ICH etiologies (probable/definite CAA-ICH, possible CAA-ICH, hypertension-ICH) and seizures. Separate regression models assessed the association of ICH location (lobar vs non-lobar) with seizures.

Of 305 ICH patients, 55 (18%) were probable/definite CAA-ICH, 46 (15%) possible CAA-ICH and 204 (67%) hypertension-ICH. Probable/definite CAA-ICH patients were significantly older, predominantly female, and had larger hematoma sizes compared to hypertension-ICH. The overall incidence of seizures after ICH was low (5%). After adjusting for ICH score, seizure history and gender, we identified more seizures in probable/definite CAA-ICH compared to hypertension-ICH (OR:4.39; 95%CI: 1.34–14.39; p=0.02). Possible CAA-ICH was not associated with seizures compared to hypertension-ICH (OR:0.62; 95%CI: 0.07-5.42; p=0.67). Though non-significant, probable/definite CAA-ICH was associated with more seizures compared to possible CAA-ICH (OR:5.83; 95%CI: 0.67-50.87; p=0.11). When assessing ICH location (rather than etiology) and seizures, lobar-ICH had more seizures, though non-significant, compared to non-lobar-ICH (OR:2.43; 95%CI: 0.79-7.46; p=0.12).

We identified differences in seizures between ICH etiologies, specifically more seizures in probable/definite CAA-ICH. It is plausible that this may be due to inherent lobar location of probable/definite CAA-ICH, however we did not identify more seizures in possible CAA-ICH. Further investigation is warranted to evaluate for potential electrographic differences between ICH etiologies.