Superficial Cerebellar Microbleeds and Cerebral Amyloid Angiopathy

Abstract

Background and Purpose— The differentiation between cerebral amyloid angiopathy (CAA) and hypertensive small vessel disease in primary intracerebral hemorrhage is mainly based on hemorrhagic neuroimaging markers in the supratentorial regions, and the cause for cerebellar microbleeds remains unknown. Our aim was to investigate whether superficial cerebellar microbleeds are more likely to be related to CAA rather than hypertensive small vessel disease. Methods— Two hundred seventy-five consecutive patients with intracerebral hemorrhage were retrospectively reviewed from a prospectively maintained hospital-based stroke registry. Eighty-five (33.1%) patients had cerebellar microbleeds and were categorized into superficial (gray matter, vermis), deep (white matter, deep nucleus, cerebellar peduncle), or mixed type based on the location of cerebellar hemorrhagic lesions. Amyloid imaging was obtained using 11C-Pittsburgh Compound B–positron emission tomography in a subgroup of patients. The associations between cerebellar microbleed locations and the type of small vessel disease (CAA versus hypertensive small vessel disease) based on distribution of supratentorial hemorrhagic lesions as well as other magnetic resonance imaging and positron emission tomography markers were analyzed. Results— The presence of cerebellar microbleed was independently associated with supratentorial microbleed and lacunar infarcts (both P <0.01). Strictly superficial cerebellar microbleeds were significantly related to CAA–intracerebral hemorrhage, cortical superficial siderosis and high-grade enlarged perivascular space in centrum semiovale (all P <0.05); deep or mixed cerebellar microbleeds were related to hypertension and deep microbleed (all P <0.05). In multivariable models, superficial cerebellar microbleeds were independently associated with CAA–intracerebral hemorrhage ( P =0.03). Of 33 patients assessed by amyloid positron emission tomography, cerebral and cerebellar amyloid load (standardized uptake value ratio) was higher in patients with superficial cerebellar microbleeds compared with deep/mixed cerebellar microbleeds (cerebrum standardized uptake value ratio [reference: cerebellum] 1.33±0.24 versus 1.05±0.09, P <0.001; cerebellum standardized uptake value ratio [reference: pons] 0.58±0.08 versus 0.51±0.09, P =0.03). Conclusions— Patients with strictly superficial cerebellar microbleeds are associated with a clinicoradiological diagnosis of CAA as well as increased cerebral and cerebellar amyloid deposition on Pittsburgh Compound B–positron emission tomography, suggesting underlying CAA pathology.