Higher Cerebral Small Vessel Disease Burden is Associated with Smaller Hematoma Volume in Mixedâlocation Intracerebral Hemorrhage
Mangmang Xu,Yajun Cheng,Shihong Zhang,Shuting Zhang,Quhong Song,Lukai Zheng,Meng Liu,Ming Liu
DOI: https://doi.org/10.1111/micc.12705
2021-01-01
Microcirculation
Abstract:Objective To study the relationship between cerebral small vessel disease (CSVD) and hematoma volume in mixed-location intracerebral hemorrhage (ICH), and non-mixed ICH (hypertensive arteriopathy/cerebral amyloid angiopathy-related ICH). Methods We consecutively collected patients with primary ICH with MRI. Mixed-location ICH was defined as having ICH or cerebral microbleeds (CMBs) in both lobar and deep regions. CSVD markers including lacunes, white matter hyperintensities (WMH), CMBs, and enlarged perivascular spaces (EPVS) were assessed on brain MRI during hospitalization. Multivariable binary logistic regression (>= 30 ml vs. <30 ml) and linear regression analyses (log-transformed hematoma volume as dependent variable) were implemented to explore the association between CSVD and hematoma volume. Results Of the 167 included patients, 69 (41.3%) had mixed-location ICH, with higher prevalence of lacune, more CMB count, higher WMH score and total CSVD score than those with non-mixed ICH (all p < .001). Higher WMH score was associated with lower risk of hematoma volume >= 30 ml (adjusted OR 0.521, 95% CI 0.299-0.908, p = .021) in patients with mixed-location ICH. Also, multivariable linear regression showed the association of smaller hematoma volume with higher CSVD burden, especially in mixed-location ICH (beta = -0.349, p = .019 for CMB >= 5; beta = -0.183, p < .001 for WMH score; beta = -0.456, p = .002 for EPVS>20 in basal ganglia and/or centrum semiovale; beta = -0.256, p = .002 for CSVD score), while these relationships were not observed in non-mixed ICH. Conclusions Higher CSVD burden is associated with smaller hematoma volume in mixed-location ICH, but not in non-mixed ICH, which is novel and needs further studies with larger sample size to confirm our results and explore the underlying mechanisms.
What problem does this paper attempt to address?