Association of Key Magnetic Resonance Imaging Markers of Cerebral Small Vessel Disease With Hematoma Volume and Expansion in Patients With Lobar and Deep Intracerebral Hemorrhage | Cerebrovascular Disease | JAMA Neurology | JAMA Network
[Skip to Navigation]
Figure 1.  Details of Patient Selection
Details of Patient Selection

CT indicates computed tomographic; ICH, intracerebral hemorrhage; and MRI, magnetic resonance imaging.

Figure 2.  Hematoma Expansion in a Patient With Lobar Intracerebral Hemorrhage With Disseminated Cortical Superficial Siderosis
Hematoma Expansion in a Patient With Lobar Intracerebral Hemorrhage With Disseminated Cortical Superficial Siderosis

A and B, Noncontrast computed tomographic axial sections are shown 6 hours and 18 hours after onset of left-sided weakness and altered consciousness in a 78-year-old patient with a history of cognitive dysfunction demonstrating a large frontoparietal acute intracerebral hemorrhage (asterisk) and significant hematoma expansion (arrows). C, Magnetic resonance imaging T-weighted gradient-recalled echo axial sequence performed 10 hours after onset shows disseminated cortical superficial siderosis (arrowheads) and the absence of accompanying microbleeds.

Table 1.  Baseline Characteristics of the Study Population
Baseline Characteristics of the Study Population
Table 2.  Ordinal Logistic Regression Analysis for Quintiles of Final Intracerebral Hemorrhage (ICH) Volumea
Ordinal Logistic Regression Analysis for Quintiles of Final Intracerebral Hemorrhage (ICH) Volumea
Table 3.  Nominal Logistic Regression of Variables Associated With Hematoma Expansiona
Nominal Logistic Regression of Variables Associated With Hematoma Expansiona
1.
Qureshi  AI, Tuhrim  S, Broderick  JP, Batjer  HH, Hondo  H, Hanley  DF.  Spontaneous intracerebral hemorrhage.  N Engl J Med. 2001;344(19):1450-1460.PubMedGoogle ScholarCrossref
2.
Fazekas  F, Kleinert  R, Offenbacher  H,  et al.  Pathologic correlates of incidental MRI white matter signal hyperintensities.  Neurology. 1993;43(9):1683-1689.PubMedGoogle ScholarCrossref
3.
Greenberg  SM, Vernooij  MW, Cordonnier  C,  et al; Microbleed Study Group.  Cerebral microbleeds: a guide to detection and interpretation.  Lancet Neurol. 2009;8(2):165-174.PubMedGoogle ScholarCrossref
4.
Charidimou  A, Linn  J, Vernooij  MW,  et al.  Cortical superficial siderosis: detection and clinical significance in cerebral amyloid angiopathy and related conditions.  Brain. 2015;138(pt 8):2126-2139.PubMedGoogle ScholarCrossref
5.
Wardlaw  JM, Smith  EE, Biessels  GJ,  et al; Standards for Reporting Vascular Changes on Neuroimaging (STRIVE v1).  Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration.  Lancet Neurol. 2013;12(8):822-838.PubMedGoogle ScholarCrossref
6.
Lou  M, Al-Hazzani  A, Goddeau  RP  Jr, Novak  V, Selim  M.  Relationship between white-matter hyperintensities and hematoma volume and growth in patients with intracerebral hemorrhage.  Stroke. 2010;41(1):34-40.PubMedGoogle ScholarCrossref
7.
Selekler  K, Erzen  C.  Leukoaraiosis and intracerebral hematoma.  Stroke. 1989;20(8):1016-1020.PubMedGoogle ScholarCrossref
8.
Inzitari  D, Giordano  GP, Ancona  AL, Pracucci  G, Mascalchi  M, Amaducci  L.  Leukoaraiosis, intracerebral hemorrhage, and arterial hypertension.  Stroke. 1990;21(10):1419-1423.PubMedGoogle ScholarCrossref
9.
Fisher  CM.  Pathological observations in hypertensive cerebral hemorrhage.  J Neuropathol Exp Neurol. 1971;30(3):536-550.PubMedGoogle ScholarCrossref
10.
Fisher  CM.  Hypertensive cerebral hemorrhage: demonstration of the source of bleeding.  J Neuropathol Exp Neurol. 2003;62(1):104-107.PubMedGoogle ScholarCrossref
11.
Davis  SM, Broderick  J, Hennerici  M,  et al; Recombinant Activated Factor VII Intracerebral Hemorrhage Trial Investigators.  Hematoma growth is a determinant of mortality and poor outcome after intracerebral hemorrhage.  Neurology. 2006;66(8):1175-1181.PubMedGoogle ScholarCrossref
12.
Dowlatshahi  D, Demchuk  AM, Flaherty  ML, Ali  M, Lyden  PL, Smith  EE; VISTA Collaboration.  Defining hematoma expansion in intracerebral hemorrhage: relationship with patient outcomes.  Neurology. 2011;76(14):1238-1244.PubMedGoogle ScholarCrossref
13.
Mayer  SA, Brun  NC, Begtrup  K,  et al; FAST Trial Investigators.  Efficacy and safety of recombinant activated factor VII for acute intracerebral hemorrhage.  N Engl J Med. 2008;358(20):2127-2137.PubMedGoogle ScholarCrossref
14.
Anderson  CS, Heeley  E, Huang  Y,  et al; INTERACT2 Investigators.  Rapid blood-pressure lowering in patients with acute intracerebral hemorrhage.  N Engl J Med. 2013;368(25):2355-2365.PubMedGoogle ScholarCrossref
15.
Delcourt  C, Huang  Y, Wang  J,  et al; INTERACT2 Investigators.  The second (main) phase of an open, randomised, multicentre study to investigate the effectiveness of an intensive blood pressure reduction in acute cerebral haemorrhage trial (INTERACT2).  Int J Stroke. 2010;5(2):110-116.PubMedGoogle ScholarCrossref
16.
Qureshi  AI, Palesch  YY, Martin  R,  et al.  Interpretation and Implementation of Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT II).  J Vasc Interv Neurol. 2014;7(2):34-40.PubMedGoogle Scholar
17.
Beslow  LA, Ichord  RN, Gindville  MC,  et al.  Frequency of hematoma expansion after spontaneous intracerebral hemorrhage in children.  JAMA Neurol. 2014;71(2):165-171.PubMedGoogle ScholarCrossref
18.
Mayer  SA, Davis  SM, Skolnick  BE,  et al; FAST Trial Investigators.  Can a subset of intracerebral hemorrhage patients benefit from hemostatic therapy with recombinant activated factor VII?  Stroke. 2009;40(3):833-840.PubMedGoogle ScholarCrossref
19.
Charidimou  A, Martinez-Ramirez  S, Reijmer  Y,  et al.  Total magnetic resonance imaging burden of small vessel disease in cerebral amyloid angiopathy: an imaging-pathologic study of concept validation.  JAMA Neurol. 2016;73(8):994-1001. PubMedGoogle ScholarCrossref
20.
Brouwers  HB, Raffeld  MR, van Nieuwenhuizen  KM,  et al.  CT angiography spot sign in intracerebral hemorrhage predicts active bleeding during surgery.  Neurology. 2014;83(10):883-889.PubMedGoogle ScholarCrossref
21.
Delgado Almandoz  JE, Yoo  AJ, Stone  MJ,  et al.  Systematic characterization of the computed tomography angiography spot sign in primary intracerebral hemorrhage identifies patients at highest risk for hematoma expansion: the spot sign score.  Stroke. 2009;40(9):2994-3000.PubMedGoogle ScholarCrossref
22.
Gurol  ME, Irizarry  MC, Smith  EE,  et al.  Plasma β-amyloid and white matter lesions in AD, MCI, and cerebral amyloid angiopathy.  Neurology. 2006;66(1):23-29.PubMedGoogle ScholarCrossref
23.
Rorden  C, Karnath  HO, Bonilha  L.  Improving lesion-symptom mapping.  J Cogn Neurosci. 2007;19(7):1081-1088.PubMedGoogle ScholarCrossref
24.
von Elm  E, Altman  DG, Egger  M, Pocock  SJ, Gøtzsche  PC, Vandenbroucke  JP; STROBE Initiative.  The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.  Lancet. 2007;370(9596):1453-1457.PubMedGoogle ScholarCrossref
25.
Gurol  ME.  Molecular neuroimaging in vascular cognitive impairment.  Stroke. 2016;47(4):1146-1152. PubMedGoogle ScholarCrossref
26.
Chen  YW, Gurol  ME, Rosand  J,  et al.  Progression of white matter lesions and hemorrhages in cerebral amyloid angiopathy.  Neurology. 2006;67(1):83-87.PubMedGoogle ScholarCrossref
27.
Charidimou  A, Peeters  AP, Jäger  R,  et al.  Cortical superficial siderosis and intracerebral hemorrhage risk in cerebral amyloid angiopathy.  Neurology. 2013;81(19):1666-1673.PubMedGoogle ScholarCrossref
28.
Linn  J, Wollenweber  FA, Lummel  N,  et al.  Superficial siderosis is a warning sign for future intracranial hemorrhage.  J Neurol. 2013;260(1):176-181.PubMedGoogle ScholarCrossref
29.
Shoamanesh  A, Martinez-Ramirez  S, Oliveira-Filho  J,  et al.  Interrelationship of superficial siderosis and microbleeds in cerebral amyloid angiopathy.  Neurology. 2014;83(20):1838-1843.PubMedGoogle ScholarCrossref
30.
Greenberg  SM, Eng  JA, Ning  M, Smith  EE, Rosand  J.  Hemorrhage burden predicts recurrent intracerebral hemorrhage after lobar hemorrhage.  Stroke. 2004;35(6):1415-1420.PubMedGoogle ScholarCrossref
31.
van Etten  ES, Auriel  E, Haley  KE,  et al.  Incidence of symptomatic hemorrhage in patients with lobar microbleeds.  Stroke. 2014;45(8):2280-2285.PubMedGoogle ScholarCrossref
32.
Charidimou  A, Martinez-Ramirez  S, Shoamanesh  A,  et al.  Cerebral amyloid angiopathy with and without hemorrhage: evidence for different disease phenotypes.  Neurology. 2015;84(12):1206-1212.PubMedGoogle ScholarCrossref
33.
Falcone  GJ, Biffi  A, Brouwers  HB,  et al.  Predictors of hematoma volume in deep and lobar supratentorial intracerebral hemorrhage.  JAMA Neurol. 2013;70(8):988-994.PubMedGoogle ScholarCrossref
34.
Greenberg  SM, Nandigam  RN, Delgado  P,  et al.  Microbleeds versus macrobleeds: evidence for distinct entities.  Stroke. 2009;40(7):2382-2386.PubMedGoogle ScholarCrossref
35.
Charidimou  A, Pantoni  L, Love  S.  The concept of sporadic cerebral small vessel disease: a road map on key definitions and current concepts.  Int J Stroke. 2016;11(1):6-18.PubMedGoogle ScholarCrossref
36.
Brouwers  HB, Chang  Y, Falcone  GJ,  et al.  Predicting hematoma expansion after primary intracerebral hemorrhage.  JAMA Neurol. 2014;71(2):158-164.PubMedGoogle ScholarCrossref
Original Investigation
December 2016

Association of Key Magnetic Resonance Imaging Markers of Cerebral Small Vessel Disease With Hematoma Volume and Expansion in Patients With Lobar and Deep Intracerebral Hemorrhage

Author Affiliations
  • 1Hemorrhagic Stroke Research Program, Department of Neurology, Massachusetts General Hospital Stroke Research Center, Harvard Medical School, Boston
  • 2Department of Neurology, Leiden University Medical Center, Leiden, the Netherlands
  • 3Brain Center Rudolf Magnus, Department of Neurosurgery, University Medical Center Utrecht, Heidelberglaan, Utrecht, the Netherlands
  • 4Division of Neurocritical Care and Emergency Neurology, Massachusetts General Hospital, Harvard Medical School, Boston
JAMA Neurol. 2016;73(12):1440-1447. doi:10.1001/jamaneurol.2016.2619
Key Points

Question  What is the association of magnetic resonance imaging markers of small vessel disease with hematoma volume and expansion in patients with intracerebral hemorrhage (ICH)?

Findings  In this analysis of data collected from consecutive patients admitted with primary lobar or deep ICH to a single tertiary care medical center, the presence of cortical superficial siderosis was an independent variable associated with larger lobar ICH volume in patients with lobar ICH. The absence of cerebral microbleeds was associated with larger lobar ICH volume in both patients with lobar or deep ICH and with hematoma expansion in patients with lobar ICH.

Meaning  This study sheds light on the mechanisms of ICH formation and growth and provides an analytical framework for future studies aimed at limiting hematoma expansion.

Abstract

Importance  Hematoma expansion is an important determinant of outcome in spontaneous intracerebral hemorrhage (ICH) due to small vessel disease (SVD), but the association between the severity of the underlying SVD and the extent of bleeding at the acute phase is unknown to date.

Objective  To investigate the association between key magnetic resonance imaging (MRI) markers of SVD (as per the Standards for Reporting Vascular Changes on Neuroimaging [STRIVE] guidelines) and hematoma volume and expansion in patients with lobar or deep ICH.

Design, Setting, and Participants  Analysis of data collected from 418 consecutive patients admitted with primary lobar or deep ICH to a single tertiary care medical center between January 1, 2000, and October 1, 2012. Data were analyzed on March 4, 2016. Participants were consecutive patients with computed tomographic images allowing ICH volume calculation and MRI allowing imaging markers of SVD assessment.

Main Outcomes and Measures  The ICH volumes at baseline and within 48 hours after symptom onset were measured in 418 patients with spontaneous ICH without anticoagulant therapy, and hematoma expansion was calculated. Cerebral microbleeds, cortical superficial siderosis, and white matter hyperintensity volume were assessed on MRI. The associations between these SVD markers and ICH volume, as well as hematoma expansion, were investigated using multivariable models.

Results  This study analyzed 254 patients with lobar ICH (mean [SD] age, 75 [11] years and 140 [55.1%] female) and 164 patients with deep ICH (mean [SD] age 67 [14] years and 71 [43.3%] female). The presence of cortical superficial siderosis was an independent variable associated with larger ICH volume in the lobar ICH group (odds ratio per quintile increase in final ICH volume, 1.49; 95% CI, 1.14-1.94; P = .004). In multivariable models, the absence of cerebral microbleeds was associated with larger ICH volume for both the lobar and deep ICH groups (odds ratios per quintile increase in final ICH volume, 1.41; 95% CI, 1.11-1.81; P = .006 and 1.43; 95% CI, 1.04-1.99; P = .03; respectively) and with hematoma expansion in the lobar ICH group (odds ratio, 1.70; 95% CI, 1.07-2.92; P = .04). The white matter hyperintensity volumes were not associated with either hematoma volume or expansion.

Conclusions and Relevance  In patients admitted with primary lobar or deep ICH to a single tertiary care medical center, the presence of cortical superficial siderosis was an independent variable associated with larger lobar ICH volume, and the absence of cerebral microbleeds was associated with larger lobar and deep ICHs. The absence of cerebral microbleeds was independently associated with more frequent hematoma expansion in patients with lobar ICH. We provide an analytical framework for future studies aimed at limiting hematoma expansion.

Introduction

Hypertensive arteriopathy and cerebral amyloid angiopathy (CAA) are common cerebral small vessel diseases (SVDs) in older individuals and are the 2 most common causes of deep and lobar spontaneous intracerebral hemorrhage (ICH), respectively.1 Imaging biomarkers, including white matter hyperintensities (WMHs),2 cerebral microbleeds (CMBs),3 and cortical superficial siderosis (cSS),4 can detect the presence of and quantify the severity of the underlying SVD-related damage.5

However, data are limited regarding the specific association between magnetic resonance imaging (MRI) markers and the extent of acute-phase bleeding (namely, ICH volume and hematoma expansion).6-8 Pathological evidence suggests that hematoma growth occurs in an “avalanche” fashion, with initial bleeding causing secondary mechanical shearing of peripheral vessels, leading to larger hematoma volume and greater expansion.9,10 The severity of microvascular changes is associated with vessel fragility and may influence the cascade of secondary ruptures. Neuroimaging markers of SVD severity are good candidates to provide further insight regarding the pathogenesis of ICH formation and ongoing hematoma expansion once it occurs.

Significant hemorrhage expansion occurs in more than one-third of patients with acute-phase ICH,11 is associated with a poor prognosis,12 and has drawn attention as a potential therapeutic target.13-16 Refining our ability to anticipate significant hematoma expansion would have major implications for patient selection in trials17 and treatment stratification in future clinical practice.18 Investigating the associations between MRI markers of SVD severity and the extent of acute-phase bleeding is critical to provide an analytical framework for future studies aimed at limiting hematoma expansion. Therefore, the objective of the present study was to investigate whether the presence and severity of SVD, assessed by validated MRI markers, correlate with ICH volume and hematoma growth in lobar and deep primary parenchymal brain bleeds (as per the Standards for Reporting Vascular Changes on Neuroimaging [STRIVE] guidelines5).

Methods
Study Design

This study was a retrospective analysis of prospectively collected data from consecutive patients admitted with spontaneous lobar or deep ICH to a single tertiary care medical center (Massachusetts General Hospital, Boston). It was approved by the Partners Human Research Committee at the Massachusetts General Hospital. Informed written or verbal consent was obtained by patients and family members or waived by the institutional review board at Massachusetts General Hospital.

Patient Selection

A total of 696 consecutive patients admitted with primary lobar or deep ICH to a single tertiary care medical center and included in an ongoing prospective cohort study19 were screened for eligibility between January 1, 2000, and October 1, 2012. Data were analyzed on March 4, 2016. As shown in Figure 1, patients meeting the following criteria were included in the ICH volume analysis: (1) at least 1 noncontrast computed tomographic (CT) scan within 48 hours after symptom onset showing spontaneous ICH, (2) clinical MRI obtained within 90 days from stroke onset, and (3) MRI with adequate sequences of sufficient quality for cerebral microbleeds, cortical superficial siderosis, and WMH volume assessments. For the hematoma expansion analysis, patients were included if they had a CT scan performed in the first 24 hours after onset and follow-up CT scan within 48 hours after symptom onset. Patients were excluded if they (1) had any suspected cause of secondary hemorrhage, (2) were receiving anticoagulant therapy at symptom onset (including warfarin sodium), (3) had the index hemorrhage located in the brainstem or cerebellum or had multiple hemorrhages, or (4) had a primary isolated intraventricular hemorrhage.

Data Collection

Patient recruitment and data collection have been previously described.19,20 A known time of onset was recorded only if it was witnessed or confirmed by the patient within a 15-minute margin of error. Otherwise, time since onset of symptoms was recorded as unknown.

All patients were treated according to a standard institutional protocol during the recruitment period. A current version is available online (https://www2.massgeneral.org/stopstroke/treatmentProtocols.aspx).

Imaging Acquisition and Analysis
Computed Tomography

Neuroradiologists or neurologists (G.B., E.S.v.E., A.C., and M.E.G.) unaware of clinical and outcome data ascertained hemorrhage locations based on the following criteria: hemorrhages originating at the cortical and subcortical junction were considered lobar, and bleeds exclusively involving the thalamus or basal ganglia were considered deep. The ICH volumes on baseline and follow-up CT were calculated by trained staff (A.V. and M.J.J.) unaware of clinical data with a semiautomated planimetric method using available software as previously described (Analyze 10.0; Mayo Clinic).21 The presence of intraventricular blood at baseline was also recorded.

Final ICH volume was defined as the largest ICH volume measured within 48 hours after symptom onset, and the corresponding time since onset was recorded (hereafter time to final CT). Time since onset was trichotomized as within 6 hours, more than 6 hours, or unknown. In the subset of patients who underwent baseline computed tomographic angiography, spot signs were read as previously described in detail.20 Significant hematoma expansion was defined as an increase in volume between baseline and follow-up CT exceeding 6 mL or 33% of the baseline volume.12

Magnetic Resonance Imaging

The MRI images were obtained using a 1.5-T system (Signa; GE Medical Systems). Imaging included the following; (1) at least whole-brain T2-weighted, T2*-weighted gradient-recalled echo (repetition time/echo time of 750/50 milliseconds, 5-mm section thickness, and 1-mm intersection gap), or susceptibility-weighted (repetition time/echo time of 27/20 milliseconds and 1.5-mm section thickness) imaging; (2) fluid-attenuated inversion recovery (repetition time/echo time of 10 000/140 milliseconds, inversion time of 2200 milliseconds, and 5-mm section thickness); and (3) T1-weighted sequences (repetition time/echo time of 400/2.5 milliseconds and 5-mm section thickness).

The CMBs were defined as punctate, hypointense foci (<5 mm in diameter) on gradient-recalled echo, distinct from vascular flow voids and leptomeningeal hemosiderosis.3 The WMH volume calculation was performed using a semiautomated planimetric method as previously described.22,23 Cortical superficial siderosis was defined as curvilinear signal loss on gradient-recalled echo following the gyral cortical surface, and its presence was classified as focal (restricted to ≤3 sulci) or disseminated (≥4 sulci).4 Peri-ICH siderosis, defined as cSS in contact with ICH or separated by less than 3 cSS free sulci from the hemorrhage, was not considered for cSS evaluation. All MRI analyses were performed and recorded by trained investigators (G.B., E.S.v.E., A.C., E.A., and K.E.H.) unaware of clinical and outcome data.

Statistical Analysis

For the statistical analysis, patients were grouped into 2 categories, namely, patients with lobar ICH or deep ICH. Clinical characteristics were compared between the 2 groups using univariable analyses.

Final ICH volume was categorized into quintiles, and each potential factor was analyzed using univariable ordinal regression analysis. Variables with significant associations (predetermined P < .10) on univariable analyses were tested using a fully adjusted ordinal regression model (correcting for sex, age, presence of spot sign, and time to final CT). Prespecified additional analyses included linear regression (univariable and multivariable) for log-transformed final ICH volume (to approximate normal distribution).

For variables associated with hematoma expansion, univariable analysis was performed using appropriate tests. All associations significant at P < .10 were entered in a nominal logistic model adjusting for sex, age, presence of spot sign, baseline ICH volume, and time to baseline CT based on previous studies. The underlying assumptions for the valid use and interpretation of each of these statistical tests were assessed and met.

For variables associated with hematoma expansion, univariable analysis was performed using appropriate tests. All associations significant at P < .10 were entered in a nominal logistic model adjusting for sex, age, presence of spot sign, baseline ICH volume, and time to baseline CT based on previous studies.

Statistical analyses were performed with a software program (JMP Pro 12; SAS Institute Inc). P < .05 was considered statistically significant. All tests of significance were 2-tailed. This article was prepared in accordance with Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.24

Results
Study Population

A total of 418 patients without anticoagulant therapy (254 with lobar ICH and 164 with deep ICH) were included in the study (Figure 1). The included patients (vs excluded patients) were more likely to be female (50.5% [211 of 418] vs 44.2% [123 of 278], P = .05), have hypertension (73.2% [306 of 418] vs 54.0% [150 of 278], P < .001), be less likely to have deep ICH (39.2% [164 of 418] vs 48.2% [134 of 278], P = .002), have larger baseline ICH volume (mean [SD], 29 [2] vs 23 [1] mL; P = .001), have a lower international normalized ratio (median, 1.0; interquartile range [IQR], 1.0-1.1 vs 1.2; IQR, 1.0-2.4; P = .001), and be seen later since onset of symptoms (median, 7; IQR, 3-20 vs 5; IQR, 3-11 hours).

Among the included patients, those with deep ICH had an overall higher burden of vascular risk factors except for younger age (P < .05 for all). Clinical characteristics and radiologic outcomes among subgroups are summarized in Table 1.

Variables Associated With Final ICH Volume

Patients with lobar ICH had larger final ICH volume compared with patients with deep ICH. The median values were 25 mL (IQR, 11-44 mL) vs 8 mL (IQR, 4-19 mL) (P < .001).

Univariable analyses in each subgroup showed that patients with larger final ICH volume were more likely to be younger and (in the deep ICH group only) to be male. Also associated with larger final ICH volume were presence of spot sign, longer time to final CT, intraventricular hemorrhage at baseline, absence of CMBs, and (in the lobar ICH group only) presence of cSS (eTable 1 in the Supplement).

All significant univariable associations (younger age, presence of spot sign, time to final CT >6 hours, intraventricular hemorrhage at baseline, and hemorrhagic markers of SVD [absence of CMBs and presence of cSS]) remained significant on multivariable testing in both groups except for shorter time to final CT in the deep ICH group. These results are summarized in Table 2.

The presence of cSS and the absence of CMBs were independently associated with larger final ICH volume in the lobar ICH group (odds ratios per quintile increase in final ICH volume, 1.49; 95% CI, 1.14-1.94; P = .004 and 1.41; 1.11-1.81; P = .006, respectively). The absence of CMBs was also independently associated with larger ICH volume in the deep ICH group (odds ratio per quintile increase in final ICH volume, 1.43; 95% CI, 1.04-1.99; P = .03).

Although only marginally associated with larger final ICH volume in the univariable analysis (β estimate for log WMH, −0. 19; 95% CI, −0.42 to 0.02; P = .09 in the lobar ICH group and −0.20; 95% CI, −0.50 to 0.09; P = .17 in the deep ICH group), larger WMH volume was forced in the multivariable model. However, it did not show a significant association with hematoma volume.

The prespecified analysis using a multivariable linear model for log-transformed final ICH volume estimates showed similar results. These findings are summarized in eTable 2 in the Supplement.

Variables Associated With Hematoma Expansion

Of the 418 included patients, 321 (193 in the lobar ICH group and 128 in the deep ICH group) had both baseline and follow-up CT within 48 hours after symptom onset to allow hematoma expansion assessment. Details are shown in Figure 1.

Patients who did not undergo follow-up CT within 48 hours after symptom onset (and were therefore excluded from this analysis) had lower baseline ICH volume (mean [SD], 16 [24] vs 26 [29] mL; P < .05) and longer time to baseline CT (22.7% [22 of 97] underwent imaging within 6 hours vs 43.6% [140 of 321] for patients with follow-up, P < .001) but did not differ otherwise from the included patients. Patients in the lobar ICH group did not experience significantly more hematoma expansion than patients in the deep ICH group (20.2% [39 of 193] vs 13.3% [17 of 128], P = .13).

Univariable analyses of the factors associated with hematoma expansion showed that patients with expansion were more likely to have shorter time to baseline CT (within 6 hours in 61.5% [24 of 39] vs 26.6% [41 of 154] of patients without expansion, P < .001) and were less likely to have CMBs (absence of CMBs in 66.7% [26 of 39] vs 48.1% [74 of 154], P = .048) in the lobar ICH group. These results are summarized in Figure 2 and eTable 3 in the Supplement.

In a multivariable model, the presence of spot sign remained a significant factor associated with hematoma expansion in both groups. In patients with lobar ICH, younger age (odds ratio for each additional year, 0.95; 95% CI, 0.91-0.99; P = .03) and the absence of CMBs (odds ratio, 1.70; 95% CI, 1.07-2.92; P = .04) also remained independently associated with hematoma expansion (Table 3). Diagnostic performances of CMBs and cSS for hemorrhage expansion are summarized in eTable 4 in the Supplement.

Discussion

In this study, we showed that key hemorrhagic imaging markers of SVD are associated with the extent of bleeding and hematoma expansion in patients with acute-phase ICH. The presence of cSS was independently associated with larger lobar ICH volume. The absence of accompanying CMBs was also independently associated with larger ICH volume in both lobar and deep hemorrhages and with hematoma expansion in the lobar ICH group. Finally, we found no evidence of a statistically significant association between WMH volume and the extent of bleeding. These MRI markers (CMBs, cSS, and WMHs) have been shown to correlate with both clinical and molecular imaging markers of SVD severity in previous studies.22,25,26 Therefore, the findings are of key importance because they add insight to our understanding of ICH development and growth, linking the underlying SVD severity with acute-phase features of brain hemorrhage.

Cortical superficial siderosis emerged in our study as a significant factor associated with larger lobar ICH volume. It is an established marker of CAA and is thought to result from blood leakage of involved cortical and leptomeningeal vessels.27 Cortical superficial siderosis is an increasingly recognized marker of focally active disease4 and is known to be associated with hemorrhagic risk in patients with CAA.28 Available evidence suggests that the presence of cSS indicates vessel fragility,29 in line with the fact that cSS is one of the strongest markers of hemorrhage recurrence in published series of patients with lobar ICH.27,28 Our results suggest that the presence of cSS identifies patients who once a vessel ruptures will experience the most bleeding, potentially from secondary ruptures or fragile vessels. It may be that the presence of cSS can mark those patients at highest risk of hematoma expansion in the hyperacute phase.

Another striking finding of this study is that the absence of accompanying CMBs was consistently correlated with larger ICH volume in both groups and with hematoma expansion in the lobar ICH group. Previous data from investigations that did not include cSS suggest that the presence of more CMBs is associated with higher risk of incident and recurrent ICH.30 More recent investigations identified lobar microbleeds as a strong marker of the presence of CAA but did not necessarily find a positive correlation between microbleed counts and bleeding risk.31 Finally, other recent data showed an inverse association between CMBs and the presence of cSS, suggesting different vasculopathic processes for these pathological findings.29,32 Another group has similarly found that patients with ICH and CMBs are less likely to show a spot sign on CT angiography, an established marker of ongoing bleeding.33 Our data showed that higher CMBs are associated with a lower extent of acute-phase ICH, in line with the recent reports referenced above. Previous work has highlighted the differences between microhemorrhages and macrohemorrhages.9,34 In patients with CAA, those with high vs low CMBs were more likely to have thicker vessel walls. As a result, patients with multiple CMBs may effectively experience a rupture when it occurs, leading to asymptomatic CMBs or smaller macrobleeds.34 Therefore, a high frequency of CMBs may reflect vessels with protection from ongoing bleeding through secondary ruptures in the acute phase once the initial ICH occurs. Both the presence of cSS and the absence of CMBs in those patients with larger hemorrhage volumes, being putative markers of more important vessel fragility, may indirectly support the avalanche model of ICH development and growth, in which after an initial rupture of a single culprit vessel ongoing bleeding is maintained by secondary ruptures from surrounding vessels.9,10

Mural changes of comparable magnitude are also present in sporadic nonamyloid microvasculopathy, in which inadequate wall remodeling may also protect from secondary ruptures.35 In a key pathological study, Fisher9 described the sites of secondary rupture surrounding a hemorrhage as being more likely to occur in vessels without arteriolosclerosis. In turn, mechanisms similar to those reported in CAA-related hemorrhages may account for the greater extent of bleeding in deep ICH without CMBs. Further studies that include experimental models are needed to test those hypotheses.

We found no association between WMH volume and either hematoma expansion or ICH volume. The absence of correlation between WMH volume and ICH volume contradicts a previous study6 that showed a strong and independent correlation between those variables. This finding held true in our sample as a whole but was confounded by the fact that lobar bleeds, typically larger than their deep counterparts,33 occur in older patients, who have more white matter damage due to aging, with the causative underlying SVD being more frequently CAA.6-8,26 Specific studies addressing this assumption are needed to clarify the association between WMH-related brain alteration and acute-phase extent of bleeding in patients with ICH.

Taken together, our findings suggest that there may be value in the use of acute-phase MRI in patients with ICH because an understanding of the underlying SVD can potentially identify those destined for more bleeding and worse outcomes. Future prospective studies may evaluate whether this information can be used to guide specific treatment strategies.

There are limitations to the present analysis, with the first being the number of exclusions, in part because of the unavailability of imaging (acquired for clinical diagnosis workup and decision making), leading to potential selection (volume analysis) and attrition (hematoma expansion analysis) biases. It is understandable that clinical teams selected which patients should undergo MRI to confirm the type of underlying SVD or rule out an underlying secondary cause of ICH more frequently for those with lobar bleeds. This rationale is supported by the fact that the excluded patients were more likely to have deep ICH and have hypertension. A similar tendency for lobar ICH cohorts to receive MRI more often is observed in most published large case series of ICH.

Another shortcoming inherent to all studies investigating hematoma expansion using imaging acquired at the discretion of a clinical team is that only those individuals with follow-up CT were included in the hematoma expansion analysis. It has been demonstrated that early withdrawal of care or death disproportionately results in the absence of follow-up imaging.36 It may be that those individuals also experienced more hematoma expansion, but this information could not be included in the present analysis, narrowing the external validity of our study.

However, the comparison of patients who did not benefit from follow-up CT in our sample showed conversely that those with less severe bleeds seen later after symptom onset were less likely to fulfill our inclusion criteria for the hematoma expansion analysis. This finding indicates that a consequent portion of the excluded patients was not included because of late presentation or clinical stability rather than more severe bleeds, with both factors closely reflecting clinical practice and being hard to delineate in retrospective analyses. In turn, this result limits the interpretation of our analyses to patients seen early in the course of their disease. In addition, we were unable to adjust the analyses for the type of blood-sensitive sequence, and a small proportion of patients received susceptibility-weighted imaging, which is known to be more sensitive than gradient-recalled echo for blood breakdown products.

Finally, the present study was performed at a tertiary care medical center. Therefore, our study cohort may not reflect the wider range of patients with ICH who would be seen in a community-based sample.

Conclusions

In this study, the presence of cSS was associated with larger lobar ICH volume, and the absence of CMBs was associated with larger lobar and deep ICHs. In addition, the absence of CMBs was independently associated with more frequent hematoma expansion in patients with lobar ICH. This study adds insight to our understanding of the mechanisms of ICH formation and growth and provides an analytical framework for future studies aimed at limiting hematoma expansion. Our findings support a broader role for acute-phase MRI in patients with ICH.

Back to top
Article Information

Corresponding Author: Joshua N. Goldstein, MD, PhD, Division of Neurocritical Care and Emergency Neurology, Massachusetts General Hospital, Harvard Medical School, Zero Emerson Place, Ste 3B, Boston, MA 02114 (jgoldstein@partners.org).

Accepted for Publication: June 1, 2016.

Published Online: October 10, 2016. doi:10.1001/jamaneurol.2016.2619

Author Contributions: Drs Boulouis and van Etten contributed equally to this work. Drs Boulouis and van Etten had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Boulouis, van Etten, Morotti, Rosand, Goldstein, Gurol.

Acquisition, analysis, or interpretation of data: Boulouis, van Etten, Charidimou, Auriel, Morotti, Pasi, Haley, Brouwers, Ayres, Vashkevich, Jessel, Schwab, Gurol.

Drafting of the manuscript: Boulouis, van Etten, Goldstein, Gurol.

Critical revision of the manuscript for important intellectual content: van Etten, Charidimou, Auriel, Morotti, Pasi, Haley, Brouwers, Ayres, Vashkevich, Jessel, Schwab, Viswanathan, Greenberg, Rosand, Goldstein, Gurol.

Obtained funding: Schwab, Viswanathan, Greenberg, Rosand, Goldstein, Gurol.

Conflict of Interest Disclosures: Dr Goldstein reported receiving consulting and research funds from CSL Behring and Boehringer Ingelheim. No other disclosures were reported.

Funding/Support: This work was supported by grants K23-NS083711 (Dr Gurol), R01-AG026484 (Dr Greenberg), and R01-NS073344 (Dr Rosand) from the National Institutes of Health. Dr Boulouis was supported by a J. William Fulbright Scholarship and a Monahan Foundation Biomedical Research Grant.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, or interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Qureshi  AI, Tuhrim  S, Broderick  JP, Batjer  HH, Hondo  H, Hanley  DF.  Spontaneous intracerebral hemorrhage.  N Engl J Med. 2001;344(19):1450-1460.PubMedGoogle ScholarCrossref
2.
Fazekas  F, Kleinert  R, Offenbacher  H,  et al.  Pathologic correlates of incidental MRI white matter signal hyperintensities.  Neurology. 1993;43(9):1683-1689.PubMedGoogle ScholarCrossref
3.
Greenberg  SM, Vernooij  MW, Cordonnier  C,  et al; Microbleed Study Group.  Cerebral microbleeds: a guide to detection and interpretation.  Lancet Neurol. 2009;8(2):165-174.PubMedGoogle ScholarCrossref
4.
Charidimou  A, Linn  J, Vernooij  MW,  et al.  Cortical superficial siderosis: detection and clinical significance in cerebral amyloid angiopathy and related conditions.  Brain. 2015;138(pt 8):2126-2139.PubMedGoogle ScholarCrossref
5.
Wardlaw  JM, Smith  EE, Biessels  GJ,  et al; Standards for Reporting Vascular Changes on Neuroimaging (STRIVE v1).  Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration.  Lancet Neurol. 2013;12(8):822-838.PubMedGoogle ScholarCrossref
6.
Lou  M, Al-Hazzani  A, Goddeau  RP  Jr, Novak  V, Selim  M.  Relationship between white-matter hyperintensities and hematoma volume and growth in patients with intracerebral hemorrhage.  Stroke. 2010;41(1):34-40.PubMedGoogle ScholarCrossref
7.
Selekler  K, Erzen  C.  Leukoaraiosis and intracerebral hematoma.  Stroke. 1989;20(8):1016-1020.PubMedGoogle ScholarCrossref
8.
Inzitari  D, Giordano  GP, Ancona  AL, Pracucci  G, Mascalchi  M, Amaducci  L.  Leukoaraiosis, intracerebral hemorrhage, and arterial hypertension.  Stroke. 1990;21(10):1419-1423.PubMedGoogle ScholarCrossref
9.
Fisher  CM.  Pathological observations in hypertensive cerebral hemorrhage.  J Neuropathol Exp Neurol. 1971;30(3):536-550.PubMedGoogle ScholarCrossref
10.
Fisher  CM.  Hypertensive cerebral hemorrhage: demonstration of the source of bleeding.  J Neuropathol Exp Neurol. 2003;62(1):104-107.PubMedGoogle ScholarCrossref
11.
Davis  SM, Broderick  J, Hennerici  M,  et al; Recombinant Activated Factor VII Intracerebral Hemorrhage Trial Investigators.  Hematoma growth is a determinant of mortality and poor outcome after intracerebral hemorrhage.  Neurology. 2006;66(8):1175-1181.PubMedGoogle ScholarCrossref
12.
Dowlatshahi  D, Demchuk  AM, Flaherty  ML, Ali  M, Lyden  PL, Smith  EE; VISTA Collaboration.  Defining hematoma expansion in intracerebral hemorrhage: relationship with patient outcomes.  Neurology. 2011;76(14):1238-1244.PubMedGoogle ScholarCrossref
13.
Mayer  SA, Brun  NC, Begtrup  K,  et al; FAST Trial Investigators.  Efficacy and safety of recombinant activated factor VII for acute intracerebral hemorrhage.  N Engl J Med. 2008;358(20):2127-2137.PubMedGoogle ScholarCrossref
14.
Anderson  CS, Heeley  E, Huang  Y,  et al; INTERACT2 Investigators.  Rapid blood-pressure lowering in patients with acute intracerebral hemorrhage.  N Engl J Med. 2013;368(25):2355-2365.PubMedGoogle ScholarCrossref
15.
Delcourt  C, Huang  Y, Wang  J,  et al; INTERACT2 Investigators.  The second (main) phase of an open, randomised, multicentre study to investigate the effectiveness of an intensive blood pressure reduction in acute cerebral haemorrhage trial (INTERACT2).  Int J Stroke. 2010;5(2):110-116.PubMedGoogle ScholarCrossref
16.
Qureshi  AI, Palesch  YY, Martin  R,  et al.  Interpretation and Implementation of Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial (INTERACT II).  J Vasc Interv Neurol. 2014;7(2):34-40.PubMedGoogle Scholar
17.
Beslow  LA, Ichord  RN, Gindville  MC,  et al.  Frequency of hematoma expansion after spontaneous intracerebral hemorrhage in children.  JAMA Neurol. 2014;71(2):165-171.PubMedGoogle ScholarCrossref
18.
Mayer  SA, Davis  SM, Skolnick  BE,  et al; FAST Trial Investigators.  Can a subset of intracerebral hemorrhage patients benefit from hemostatic therapy with recombinant activated factor VII?  Stroke. 2009;40(3):833-840.PubMedGoogle ScholarCrossref
19.
Charidimou  A, Martinez-Ramirez  S, Reijmer  Y,  et al.  Total magnetic resonance imaging burden of small vessel disease in cerebral amyloid angiopathy: an imaging-pathologic study of concept validation.  JAMA Neurol. 2016;73(8):994-1001. PubMedGoogle ScholarCrossref
20.
Brouwers  HB, Raffeld  MR, van Nieuwenhuizen  KM,  et al.  CT angiography spot sign in intracerebral hemorrhage predicts active bleeding during surgery.  Neurology. 2014;83(10):883-889.PubMedGoogle ScholarCrossref
21.
Delgado Almandoz  JE, Yoo  AJ, Stone  MJ,  et al.  Systematic characterization of the computed tomography angiography spot sign in primary intracerebral hemorrhage identifies patients at highest risk for hematoma expansion: the spot sign score.  Stroke. 2009;40(9):2994-3000.PubMedGoogle ScholarCrossref
22.
Gurol  ME, Irizarry  MC, Smith  EE,  et al.  Plasma β-amyloid and white matter lesions in AD, MCI, and cerebral amyloid angiopathy.  Neurology. 2006;66(1):23-29.PubMedGoogle ScholarCrossref
23.
Rorden  C, Karnath  HO, Bonilha  L.  Improving lesion-symptom mapping.  J Cogn Neurosci. 2007;19(7):1081-1088.PubMedGoogle ScholarCrossref
24.
von Elm  E, Altman  DG, Egger  M, Pocock  SJ, Gøtzsche  PC, Vandenbroucke  JP; STROBE Initiative.  The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.  Lancet. 2007;370(9596):1453-1457.PubMedGoogle ScholarCrossref
25.
Gurol  ME.  Molecular neuroimaging in vascular cognitive impairment.  Stroke. 2016;47(4):1146-1152. PubMedGoogle ScholarCrossref
26.
Chen  YW, Gurol  ME, Rosand  J,  et al.  Progression of white matter lesions and hemorrhages in cerebral amyloid angiopathy.  Neurology. 2006;67(1):83-87.PubMedGoogle ScholarCrossref
27.
Charidimou  A, Peeters  AP, Jäger  R,  et al.  Cortical superficial siderosis and intracerebral hemorrhage risk in cerebral amyloid angiopathy.  Neurology. 2013;81(19):1666-1673.PubMedGoogle ScholarCrossref
28.
Linn  J, Wollenweber  FA, Lummel  N,  et al.  Superficial siderosis is a warning sign for future intracranial hemorrhage.  J Neurol. 2013;260(1):176-181.PubMedGoogle ScholarCrossref
29.
Shoamanesh  A, Martinez-Ramirez  S, Oliveira-Filho  J,  et al.  Interrelationship of superficial siderosis and microbleeds in cerebral amyloid angiopathy.  Neurology. 2014;83(20):1838-1843.PubMedGoogle ScholarCrossref
30.
Greenberg  SM, Eng  JA, Ning  M, Smith  EE, Rosand  J.  Hemorrhage burden predicts recurrent intracerebral hemorrhage after lobar hemorrhage.  Stroke. 2004;35(6):1415-1420.PubMedGoogle ScholarCrossref
31.
van Etten  ES, Auriel  E, Haley  KE,  et al.  Incidence of symptomatic hemorrhage in patients with lobar microbleeds.  Stroke. 2014;45(8):2280-2285.PubMedGoogle ScholarCrossref
32.
Charidimou  A, Martinez-Ramirez  S, Shoamanesh  A,  et al.  Cerebral amyloid angiopathy with and without hemorrhage: evidence for different disease phenotypes.  Neurology. 2015;84(12):1206-1212.PubMedGoogle ScholarCrossref
33.
Falcone  GJ, Biffi  A, Brouwers  HB,  et al.  Predictors of hematoma volume in deep and lobar supratentorial intracerebral hemorrhage.  JAMA Neurol. 2013;70(8):988-994.PubMedGoogle ScholarCrossref
34.
Greenberg  SM, Nandigam  RN, Delgado  P,  et al.  Microbleeds versus macrobleeds: evidence for distinct entities.  Stroke. 2009;40(7):2382-2386.PubMedGoogle ScholarCrossref
35.
Charidimou  A, Pantoni  L, Love  S.  The concept of sporadic cerebral small vessel disease: a road map on key definitions and current concepts.  Int J Stroke. 2016;11(1):6-18.PubMedGoogle ScholarCrossref
36.
Brouwers  HB, Chang  Y, Falcone  GJ,  et al.  Predicting hematoma expansion after primary intracerebral hemorrhage.  JAMA Neurol. 2014;71(2):158-164.PubMedGoogle ScholarCrossref
×