Charidimou Andreas, Boulouis Gregoire, Pasi Marco, Auriel Eitan, van Etten Ellis S, Haley Kellen, Ayres Alison, Schwab Kristin M, Martinez-Ramirez Sergi, Goldstein Joshua N, Rosand Jonathan, Viswanathan Anand, Greenberg Steven M, Gurol M Edip
From the Hemorrhagic Stroke Research Program (A.C., G.B., M.P., E.A., E.S.v.E., K.H., A.A., K.M.S., S.M.-R., A.V., S.M.G., M.E.G.), Department of Neurology, Massachusetts General Hospital Stroke Research Center, Harvard Medical School, Boston; Université Paris-Descartes (G.B.), INSERM UMR 894, Department of Neuroradiology, Centre Hospitalier Sainte-Anne, Paris, France; and Division of Neurocritical Care and Emergency Neurology (J.N.G., J.R.), Massachusetts General Hospital, Harvard Medical School, Boston.
Neurology. 2017 Mar 21;88(12):1157-1164. doi: 10.1212/WNL.0000000000003746. Epub 2017 Feb 22.
To assess MRI-visible enlarged perivascular spaces (EPVS) burden and different topographical patterns (in the centrum semiovale [CSO] and basal ganglia [BG]) in 2 common microangiopathies: cerebral amyloid angiopathy (CAA) and hypertensive arteriopathy (HA).
Consecutive patients with spontaneous intracerebral hemorrhage (ICH) from a prospective MRI cohort were included. Small vessel disease MRI markers, including cerebral microbleeds (CMBs), cortical superficial siderosis (cSS), and white matter hyperintensities (WMH), were rated. CSO-EPVS/BG-EPVS were assessed on a validated 4-point visual rating scale (0 = no EPVS, 1 = <10, 2 = 11-20, 3 = 21-40, and 4 = >40 EPVS). We tested associations of predefined high-degree (score >2) CSO-EPVS and BG-EPVS with other MRI markers in multivariable logistic regression. We subsequently evaluated associations with CSO-EPVS predominance (i.e., CSO-EPVS > BG-EPVS) and BG-EPVS predominance pattern (i.e., BG-EPVS > CSO-EPVS) in adjusted multinomial logistic regression (reference group, BG-EPVS = CSO-EPVS).
We included 315 patients with CAA-ICH and 137 with HA-ICH. High-degree CSO-EPVS prevalence was greater in CAA-related ICH vs HA-related ICH (43.8% vs 17.5%, < 0.001). In multivariable logistic regression, high-degree CSO-EPVS was associated with lobar CMB (odds ratio [OR] 1.33, 95% confidence interval [CI] 1.10-1.61, = 0.003) and cSS (OR 2.08, 95% CI 1.30-3.32, = 0.002). Deep CMBs (OR 2.85, 95% CI 1.75-4.64, < 0.0001) and higher WMH volume (OR 1.02, 95% CI 1.01-1.04, = 0.010) were predictors of high-degree BG-EPVS. A CSO-EPVS-predominant pattern was more common in CAA-ICH than in HA-ICH (75.9% vs 39.4%, respectively, < 0.0001). CSO-PVS predominance was associated with lobar CMB burden and cSS, while BG-EPVS predominance was associated with HA-ICH and WMH volumes.
Different patterns of MRI-visible EPVS provide insights into the dominant underlying microangiopathy type in patients with spontaneous ICH.
评估两种常见微血管病，即脑淀粉样血管病（CAA）和高血压性动脉病（HA）中，MRI可见的血管周围间隙增宽（EPVS）负担及不同的拓扑学模式（半卵圆中心[CSO]和基底节区[BG]）。
纳入来自前瞻性MRI队列的自发性脑出血（ICH）连续患者。对包括脑微出血（CMB）、皮质表面铁沉积（cSS）和白质高信号（WMH）在内的小血管病MRI标志物进行评分。采用经过验证的4分视觉评分量表评估CSO-EPVS/BG-EPVS（0 = 无EPVS，1 = <10，2 = 11 - 20，3 = 21 - 40，4 = >40个EPVS）。在多变量逻辑回归中，我们测试了预先定义的高度（评分>2）CSO-EPVS和BG-EPVS与其他MRI标志物的关联。随后，在调整后的多项逻辑回归中（参照组，BG-EPVS = CSO-EPVS），我们评估了与CSO-EPVS优势（即CSO-EPVS > BG-EPVS）和BG-EPVS优势模式（即BG-EPVS > CSO-EPVS）的关联。
我们纳入了315例CAA-ICH患者和137例HA-ICH患者。与HA相关的ICH相比，CAA相关的ICH中高度CSO-EPVS患病率更高（分别为43.8%和17.5%，<0.001）。在多变量逻辑回归中，高度CSO-EPVS与脑叶CMB相关（比值比[OR] 1.33，95%置信区间[CI] 1.10 - 1.61，= 0.003）和cSS相关（OR 2.08，95% CI 1.30 - 3.32，= 0.002）。深部CMB（OR 2.85，95% CI 1.75 - 4.64，<0.0001）和更高的WMH体积（OR 1.02，95% CI 1.01 - 1.04，= 0.010）是高度BG-EPVS的预测因素。CSO-EPVS优势模式在CAA-ICH中比在HA-ICH中更常见（分别为75.9%和39.4%，<0.0001）。CSO-PVS优势与脑叶CMB负担和cSS相关，而BG-EPVS优势与HA-ICH和WMH体积相关。
MRI可见的EPVS的不同模式为自发性ICH患者潜在的主要微血管病类型提供了见解。
