Etiology of Primary Cerebellar Intracerebral Hemorrhage Based on Topographic Localization

Abstract

BACKGROUND: Cerebellar intracerebral hemorrhage (cICH) is often attributed to hypertension or cerebral amyloid angiopathy (CAA). However, deciphering the exact etiology can be challenging. A recent study reported a topographical etiologic relationship with superficial cICH secondary to CAA. We aimed to reexamine this relationship between topography and etiology in a separate cohort of patients and using the most recent Boston criteria version 2.0. METHODS: We performed a retrospective analysis of consecutive patients with primary cICH admitted to a tertiary academic center between 2000 and 2022. cICH location on brain computed tomography/magnetic resonance imaging scan(s) was divided into strictly superficial (cortex, surrounding white matter, vermis) versus deep (cerebellar nuclei, deep white matter, peduncular region) or mixed (both regions). Magnetic resonance imaging was rated for markers of cerebral small vessel disease. We assigned possible/probable versus absent CAA using Boston criteria 2.0. RESULTS: We included 197 patients; 106 (53.8%) were females, median age was 74 (63–82) years. Fifty-six (28%) patients had superficial cICH and 141 (72%) deep/mixed cICH. Magnetic resonance imaging was available for 112 (57%) patients (30 [26.8%] with superficial and 82 [73.2%] with deep/mixed cICH). Patients with superficial cICH were more likely to have possible/probable CAA (48.3% versus 8.6%; odds ratio [OR], 11.43 [95% CI, 3.26–40.05]; P <0.001), strictly lobar cerebral microbleeds (51.7% versus 6.2%; OR, 14.18 [95% CI, 3.98–50.50]; P <0.001), and cortical superficial siderosis (13.8% versus 1.2%; OR, 7.70 [95% CI, 0.73–80.49]; P =0.08). Patients with deep/mixed cICH were more likely to have deep/mixed cerebral microbleeds (59.2% versus 3.4%; OR, 41.39 [95% CI, 5.01–341.68]; P =0.001), lacunes (54.9% versus 17.2%; OR, 6.14 [95% CI, 1.89–19.91]; P =0.002), severe basal ganglia enlarged perivascular spaces (36.6% versus 7.1%; OR, 7.63 [95% CI, 1.58–36.73]; P =0.01), hypertension (84.4% versus 62.5%; OR, 3.43 [95% CI, 1.61 to −7.30]; P =0.001), and higher admission systolic blood pressure (172 [146–200] versus 146 [124–158] mm Hg, P <0.001). CONCLUSIONS: Our results suggest that superficial cICH is strongly associated with CAA whereas deep/mixed cICH is strongly associated with hypertensive arteriopathy.