Progression of White Matter Injury After Intracerebral Hemorrhage: A Magnetic Resonance Imaging Study

Original Article

Progression of White Matter Injury After Intracerebral Hemorrhage: A Magnetic Resonance Imaging Study Deren Wang1,2, Casey Norton2, Johanna Helenius2, Xiaomeng Xu2,3, Ming Liu1, Magdy Selim2, Vasileios-Arsenios Lioutas2

-

BACKGROUND: White matter injury (WMI) has been observed after experimental intracerebral hemorrhage (ICH). The supporting clinical data have been sparse. We assessed the presence, extent, and progression of WMI in patients with ICH.

(OR, 1.073; 95% CI, 1.019e1.130; P [ 0.007) were predictors of WMIP after adjustment. Multivariate analyses showed an independent association between WMIP and unfavorable 3-month outcomes (OR, 5.196; 95% CI, 1.059e25.483; P [ 0.042).

METHODS: We performed a retrospective review of data from 65 consecutive patients with spontaneous supratentorial ICH who had undergone baseline brain magnetic resonance imaging (MRI) within 7 days of ICH onset and repeat MRI afterward. We used the Fazekas scale (FZKS) to grade the severity of WMI. The clinical and imaging characteristics of the patients with and without WMI progression (WMIP) were compared using uni- and multivariate logistic regression analyses.

-

-

RESULTS: We observed WMIP in 23 patients (35.4%). WMIP was noted in both hemispheres but more commonly ipsilateral to the ICH (33% vs. 21%). The mean total FZKS score had increased from 3 (interquartile range [IQR], 1e4) at baseline to 4 (IQR, 2e5) on repeat MRI (P < 0.0001). Patients with lobar ICH had a greater median FZKS score than those with deep ICH (median, 3; IQR, 2e4; vs. 1.5, IQR, 1e3.25; P [ 0.027). The baseline parenchymal ICH volume (odds ratio [OR], 1.067; 95% confidence interval [CI], 1.018e 1.119; P [ 0.007) and ventricular volume on baseline MRI

-

Key words Intracerebral hemorrhage - Magnetic resonance imaging - Prognosis - White matter hyperintensities -

Abbreviations and Acronyms CI: Confidence interval CMB: Cerebral microbleeding FLAIR: Fluid-attenuated inversion recovery FZKS: Fazekas scale ICH: Intracerebral hemorrhage IQR: Interquartile range IVH: Intraventricular hemorrhage MRI: Magnetic resonance imaging mRS: Modified Rankin scale OR: Odds ratio PV: Periventricular SC: Subcortical

WORLD NEUROSURGERY -: e1-e11, - 2019

CONCLUSIONS: WMI will progress over time in patients with ICH, and WMIP has been associated with worse outcomes. This novel finding could represent a potential therapeutic target. Future prospective larger studies are needed to confirm our findings.

INTRODUCTION

A

fter intracerebral hemorrhage (ICH), blood can dissect along the fiber tracts into the white matter, destroying axonal integrity and causing demyelination and dysfunction of brain structures distant from the hemorrhage.1 Preclinical animal studies have demonstrated the occurrence of white matter injury (WMI), including axonal damage and demyelination, at the edge of the hematoma and its temporal and spatial progression after experimental ICH.2 Little is known about the occurrence and progression of WMI and its effect on the outcome in patients with ICH. A few clinical studies have shown that the

WMH: White matter hyperintensity WMI: White matter injury WMIP: White matter injury progression From the 1Department of Neurology, Center of Cerebrovascular Diseases, West China Hospital, Sichuan University, Chengdu, China; 2Stroke Division, Department of Neurology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts, USA; and 3Department of Neurology, Jinling Hospital, Medical School of Nanjing University, Nanjing, China To whom correspondence should be addressed: Magdy Selim, M.D., Ph.D. [E-mail: [email protected]] Deren Wang and Casey Norton contributed equally to the present study. Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2019.02.089 Journal homepage: www.journals.elsevier.com/world-neurosurgery Available online: www.sciencedirect.com 1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.

www.journals.elsevier.com/world-neurosurgery

e1

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

presence of radiological markers of WMI, in particular, white matter hyperintensities (WMHs), is independently associated with larger hematoma volumes and worse outcomes after spontaneous ICH.3-6 This finding suggests that WMI could be an important prognostic factor and that its prevention or progression might improve the outcomes for patients with ICH. Whether WMI can occur or progress as a consequence of ICH and whether WMI progression (WMIP) over time could affect the outcomes for patients with ICH is unclear. Therefore, we undertook the present study to assess the presence, extent, and progression of WMI, as assessed by WMHs, after ICH in the clinical setting and to determine its effects on long-term functional outcomes. We also aimed to determine whether WMI severity and WIMP would vary by ICH location (lobar vs. deep). METHODS The institutional review board at Beth Israel Deaconess Medical Center approved the present study. We adhered to the STROBE (strengthening the reporting of observational studies in epidemiology) guidelines (available at: www.strobe-statement.org).7 Patient Selection We retrospectively reviewed our prospectively collected ICH database for consecutive patients with spontaneous ICH who had been admitted from January 2008 to May 2017. We identified patients with spontaneous supratentorial ICH who had had a baseline brain magnetic resonance imaging (MRI) scan performed within 7 days of ICH onset and a repeat brain MRI scan
14 days afterward. Patients with a secondary cause of ICH (i.e., underlying aneurysm, vascular malformation, brain neoplasm or metastasis, head trauma, dural sinus thrombosis, hemorrhagic transformation of ischemic infarction, infection), those with bilateral hemorrhage, infratentorial ICH, primary intraventricular hemorrhage (IVH), and those who had undergone surgical evacuation were excluded from the present study. We also excluded patients whose repeat brain MRI had been performed because of a new ischemic stroke or recurrent ICH. Data Collection We retrieved the baseline clinical and demographic information, including age, sex, ethnicity, preadmission modified Rankin scale (mRS), Glasgow coma scale, and systolic and diastolic blood pressure from the initial evaluation. The pre-existing risk factors recorded included a history of hypertension, diabetes mellitus, hyperlipidemia, atrial fibrillation, coronary artery disease, ICH, and ischemic stroke or transient ischemic attack, current smoking, and alcohol consumption. We also recorded the medications in use at admission, such as antiplatelet agents, anticoagulants, statins, antihypertensive agents, and antidiabetic drugs. The laboratory results on admission extracted included serum glucose, estimated glomerular filtration rate (MDRD [modification of diet in renal disease] calculation), hemoglobin A1c, calculated lowdensity lipoprotein cholesterol, partial thromboplastin time, international normalized ratio, and platelet count. In addition, the date of ICH symptom onset, when known, or the last time the patient was known to be symptom free and the date of the baseline and repeat brain MRI scans were recorded. The mRS

e2

www.SCIENCEDIRECT.com

score was determined at the clinic follow-up visits by experienced and certified investigators. An unfavorable outcome was defined as an mRS score of 3e6.8 Image Review and Analysis The initial computed tomography scan taken at presentation was reviewed to determine the ICH location, the presence of IVH, and the hematoma volumes. We measured the ICH and IVH volumes using the National Institutes of Health image-processing software, as previously described.3,9 In brief, the regions of interest were manually drawn by tracing the perimeters of the hematoma in each slice, throughout the hemorrhagic lesion. The traced regions of interest in every slice were then summed after adjusting for slice thickness to yield the hematoma volume. We previously demonstrated the high intra- and inter-rater reliability for these volumetric measurements (intraclass correlation coefficient, >0.98).3,9 All brain MRI studies were performed in accordance with a standardized institutional protocol using a 1.5T magnet and including T1-weighted, T2-weighted, and fluid-attenuated inversion recovery (FLAIR) and T2* gradient-echo sequences. Baseline axial T2-weighted images were used to measure the volume of the ventricular system (including the lateral, third, and fourth ventricles) using National Institutes of Health image-processing software and the same procedures. WMI was defined as supratentorial WMHs on T2-weighted and FLAIR sequences and isointense or hypointense (although not as hypointense as the cerebrospinal fluid) on T1-weighted sequences. In addition, we assessed for presence of cerebral microbleeds (CMB) and lacunar infarcts. All imaging markers were defined in accordance with the STRIVE (standards for reporting vascular changes on neuroimaging) criteria.10 The severity of WMI was graded on axial FLAIR sequences using the Fazekas scale (FZKS) by visual assessment of both periventricular (PV) (0, absent; 1, caps or pencil lining; 2, smooth halo; 3, irregular periventricular hyperintensities extending into the deep white matter) and subcortical (SC) areas (0, absent; 1, punctuate foci; 2, beginning confluence of foci; 3, large confluent areas; Supplemental Figure 1).11,12 The total FZKS score was calculated by adding the PV and SC scores and was separately assessed for each hemisphere (ipsilateral and contralateral to the ICH). The total FZKS score for the entire brain was calculated by the greatest scores from the PV and SC areas from any side of the hemisphere. Deep gray matter and brainstem hyperintensities, perivascular spaces, and lacunar infarcts of presumed vascular origin were excluded. An experienced operator who was unaware of the clinical data, patient identity, and sequential ordering of the baseline and repeat MRI scans assessed the severity of WMI. The baseline and follow-up images were assessed in a random manner. The images were reviewed independently by 2 raters on 20 randomly selected scans. The intraclass and interclass correlation coefficients for the total FZKS score were 0.97 (95% confidence interval [CI], 0.92e0.99) and 0.95 (95% CI, 0.88e0.98), respectively. Therefore, only 1 operator analyzed the remaining MRI scans. WMIP was defined as an increase of
1 point from the baseline to the repeat total FZKS score. Statistical Analysis Categorical variables are reported as proportions. Continuous variables are reported as the mean  standard deviation or median and interquartile range (IQR). The normality of the data was

WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.089

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

examined using the Shapiro-Wilk test. The total FZKS score was not normally distributed (P < 0.001), and the Wilcoxon signedranks test was used to compare the FZKS scores between the baseline and repeat MRI scans. The results for categorical variables were compared between groups with and without WMIP using the c2 or Fisher exact tests. The results for continuous variables were compared using 1-way analysis of variance or the Mann-Whitney U test. Binary logistic regression was used to evaluate the association between the risk factors and WMIP. Factors with P  0.10 on univariate analysis were analyzed by binary logistic regression. This P value threshold was less stringent than the traditional cutoff of P < 0.05 to reduce the possibility of omitting factors that might have been related to dependent variables. The interval from the baseline to repeat MRI scans was forced into logistic model to reduce its effect on WMIP, regardless of the P value found on univariate analysis between groups. Binary logistic regression was also used to evaluate the association between WMIP and unfavorable outcomes. A 2-tailed value of P < 0.05 was defined as the threshold of statistical significance. All statistical analyses were performed using SPSS, version 16 (IBM, Armonk, New York, USA). RESULTS Our final cohort included 65 patients with spontaneous supratentorial ICH who had both baseline and follow-up MRI scans available. The steps for cohort creation, reasons for exclusion, and numbers excluded are summarized in Figure 1. The characteristics of the patients (included, n ¼ 65; excluded, n ¼ 246) considered for our study are shown in Table 1. The patients included in the present analysis were younger (P ¼ 0.007), and had a lower preadmission mRS score (P < 0.001), a greater proportion of lobar ICH (P < 0.001), and a lower prevalence of hypertension (P ¼ 0.010), atrial fibrillation (P ¼ 0.038), history of previous ICH (P ¼ 0.021), and oral anticoagulant use (P ¼ 0.032). In addition, fewer of the included patients had a reduced estimated glomerular filtration rate <60 mL/min/1.73 m2 on admission (P < 0.001). Finally, the included patients had a lower proportion of unfavorable outcome at discharge (P ¼ 0.002). The age of the included patients ranged from 24 to 88 years (mean, 65.6  14.9). O the 65 included patients, 51 had lobar ICH and 14 had deep ICH. IVH was present in 15 patients (23.1%) and was found in 19.6% of the patients with lobar ICH compared with 35.7% of those with deep ICH (P ¼ 0.363). The mean interval from ICH onset to the baseline MRI scan was 2.2  1.5 days, and the interval from the baseline MRI to the repeated MRI scan was 121.6  94.8 days. The severity of WMI, as assessed by the FZKS score, was significantly greater in those with lobar compared with deep ICH on both the baseline and the repeat MRI scans. This was noted in the ipsilateral and contralateral hemispheres and in the entire brain (Table 2). WMIP was noted in both the ipsilateral and the contralateral hemispheres to the ICH, and the median total FZKS score of the entire brain had increased from a median of 3 (IQR, 1e4) at baseline to 4 (IQR, 2e5) on the repeat MRI scan (P < 0.001; Table 3). When stratified by ICH location (deep vs. lobar), we observed a similar pattern in the patients with lobar

WORLD NEUROSURGERY -: e1-e11, - 2019

Figure 1. Flowchart of patient selection. ICH, intracerebral hemorrhage; IVH, intraventricular hemorrhage; MRI, magnetic resonance imaging.

ICH, but to a lesser extent in patients with deep ICH (Table 3). Although the median total FZKS score for the entire brain increased from 1.5 to 2.5 on the repeat MRI scan (P ¼ 0.03) in the deep ICH group, the changes in the FZKS score in the ipsilateral (P ¼ 0.06) and contralateral (P ¼ 0.12) hemispheres were less pronounced. A total of 23 patients (35.4%) showed evidence of WMIP on the repeat MRI scan. WMIP was observed ipsilateral to the ICH in 21 patients (32.3%) and contralateral to the ICH in 14 patients (21.5%; P < 0.0001; Figure 2). In 12 patients (18.5%), the WMIP was bilateral. WMIP was observed as early as 44 days after the baseline MRI. None of the patients exhibited regression of WMI between the scans, and the proportion of patients experiencing WMIP was not different between those with lobar (17 of 51; 33.3%) and deep (6 of 14; 42.9%) ICH (P ¼ 0.73). The observed increase in the FZKS score was not paralleled by a similar increase in the markers of small vessel disease (i.e., CMB and lacunar infarcts). The baseline MRI demonstrated the presence of CMB in 24 patients (37%) and lacunar infarcts in 7 (11%). However, only 4 patients (6%) developed new CMB: 1 (4%) in the WMIP group and 3 (8%) in the non-WIMP group (P ¼ 0.6). Three patients (5%) had developed new lacunar infarcts on the repeat MRI scan: 2 (8.7%) in the WMIP group and 1 (2.4%) in the non-WIMP group (P ¼ 0.27). The characteristics of the patients with versus without WMIP are listed in Table 4. The demographic, clinical, and radiological characteristics of the patients with and without WMIP for the whole ICH cohort were largely comparable. The interval from the baseline MRI to the repeat MRI did not differ between the groups (P ¼ 0.683), and the FZKS score on the baseline MRI scan did not differ between the patients with and without WMIP (P ¼ 0.51). Only the baseline parenchymal ICH volume (P ¼ 0.007), ventricular volume

www.journals.elsevier.com/world-neurosurgery

e3

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

Table 1. Patient Characteristics Stratified by Study Inclusion Included (n [ 65)

Excluded (n [ 246)

P Value

OR (95% CI)

65.6  14.9

70.8  13.6

0.007

NA

Female sex (n; %)

32 (49.2)

110 (44.7)

0.516

1.199 (0.694e2.073)

White race (n; %)

52 (80.0)

175 (71.1)

0.152

1.623 (0.833e3.163)

Preadmission mRS score (median; IQR)

0 (0)

0 (1)

< 0.001

NA

Admission GCS score (median; IQR)

14 (4)

15 (4)

0.595

Characteristic Age (years)

BP on admission

NA

SBP (mm Hg)

160.0  32.1

163.5  30.2

0.413

DBP (mm Hg)

86.4  19.7

85.4  19.1

0.694

Hypertension

39 (60.0)

187 (76.0)

0.010

0.473 (0.266e0.842)

Diabetes mellitus

11 (16.9)

58 (23.6)

0.251

0.660 (0.324e1.346)

Hyperlipidemia

26 (40.0)

86 (35.0)

0.451

1.240 (0.708e2.174)

6 (9.2)

50 (20.3)

0.038

0.399 (0.163e0.976)

11 (16.9)

41 (16.7)

0.961

1.019 (0.491e2.113)

Pre-existing risk factors (n; %)

Atrial fibrillation Coronary artery disease ICH history

1 (1.5)

26 (10.6)

0.021

0.132 (0.018e0.993)

IS or TIA history

7 (10.8)

38 (15.4)

0.340

0.661 (0.280e1.557)

Current smoking

14 (21.5)

35 (14.2)

0.150

1.655 (0.829e3.303)

Alcohol consumption

18 (27.7)

45 (18.3)

0.094

1.711 (0.909-3.219)

Antihypertensive

34 (52.3)

149 (60.6)

0.229

0.714 (0.412-1.237)

Antiglycemic

10 (15.4)

39 (15.9)

0.926

0.965 (0.453-2.055)

Antiplatelet

23 (35.4)

109 (44.3)

0.195

0.688 (0.390-1.214)

Preadmission medications (n; %)

Oral anticoagulant

7 (10.8)

56 (22.8)

0.032

0.409 (0.177e0.948)

Statin

19 (29.2)

99 (40.2)

0.104

0.613 (0.339e1.109)

129.3  43.2

138.6  67.1

0.288

NA

4 (6.2)

70 (28.6)

< 0.001

0.164 (0.057e0.468)

Laboratory data on admission Glucose (mg/dL) eGFR <60 mL/min/1.73 m (n; %) 2

5.9  1.0

6.1  1.2

0.364

NA

LDL-C (mg/dL)

99.0  36.3

92.5  35.1

0.308

NA

PTT (seconds)

28.3  5.0

29.2  5.6

0.256

NA

9 (13.8)

51 (20.9)

0.201

0.608 (0.282e1.312)

247.3  90.2

225.2  81.1

0.058

NA

< 0.001

3.358 (1.767e6.382)

HbA1c (%)

INR >1.3 (n; %) Platelet count (K/mL) Neuroimaging findings (n; %) Hematoma location Lobar

51 (78.5)

128 (52.0)

Nonlobar

14 (21.5)

118 (48.0)

IVH (n; %)

15 (23.1)

83 (33.7)

0.100

0.589 (0.312e1.111)

mRS score at discharge of 3e6

29 (44.6)

162 (65.9)

0.002

0.418 (0.240e0.728)

OR, odds ratio; CI, confidence interval; NA, not applicable; mRS, modified Rankin scale; IQR interquartile range; GCS, Glasgow coma scale; BP, blood pressure; SBP, systolic blood pressure; DBP, diastolic blood pressure; ICH, intracerebral hemorrhage; IS, ischemic stroke; TIA, transient ischemic attack; eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c; LDL-C calculated low-density lipoprotein cholesterol; PTT, partial thromboplastin time; INR, international normalized ratio; IVH, intraventricular hemorrhage.

e4

www.SCIENCEDIRECT.com

WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.089

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

Table 2. Total Fazekas Scores Stratified by Intracerebral Hemorrhage Location and Imaging Study Fazekas Scale Score MRI Study

Lobar ICH

Deep ICH

P Value

Entire brain, baseline

3 (2e4)

1.5 (1e3.25)

0.027

Entire brain, repeat

4 (2e5)

2.5 (1e3.25)

0.049

Ipsilateral, baseline

3 (2e4)

1 (1e3.25)

0.025

Ipsilateral, repeat

4 (2e5)

2 (1e3.25)

0.038

Contralateral, baseline

3 (2e4)

1.5 (1e3.25)

0.039

Contralateral, repeat

4 (2e5)

2 (1e3.25)

0.036

Data presented as median (interquartile range). MRI, magnetic resonance imaging; ICH intracerebral hemorrhage.

on the initial MRI scan (P ¼ 0.031), and the presence of IVH (P ¼ 0.004) emerged as predictors of WMIP on univariate analysis. After adjustment by binary logistic regression model, only the baseline parenchymal ICH volume (odds ratio [OR], 1.067; 95% CI, 1.018e 1.119; P ¼ 0.007) and ventricular volume on the initial MRI scan (OR, 1.073; 95% CI, 1.019e1.130; P ¼ 0.007) showed an independent association with WMIP (Table 5). Overall, 29 patients (44.6%) had an unfavorable outcome (mRS score >2) at discharge, and 13 patients (20.0%) had an mRS score >2 at 3 months. The proportion of patients with an unfavorable outcome at discharge was 52.2% for patients with WMIP and 40.5% for those without WMIP (OR, 1.604; 95% CI, 0.58e4.47; P ¼ 0.518). At 3 months, however, the proportion of patients with

an mRS score >2 was 34.8% in the WMIP group compared with 11.9% in the non-WMIP group (OR, 3.947; 95% CI, 1.11e14.03; P ¼ 0.049) on univariate analysis (Table 4). This association between WMIP and unfavorable 3-month outcomes remained significant (OR, 5.196; 95% CI, 1.059e25.483; P ¼ 0.042) after adjustment for confounding factors with P  0.10 (Glasgow coma scale score on admission, presence of IVH, baseline parenchymal ICH volume, and ventricular volume on the initial MRI scan) and forcing the baseline FZKS score into the model because the baseline lesion burden can be a predictor of subsequent progression. We found no association between the baseline entire brain FZKS score and 3-month outcomes (median score, 4; IQR, 2e4.5; vs. median score, 3; IQR, 1e4; P ¼ 0.33).

Table 3. Total Fazekas Scale Scores at Baseline Compared With Those at Repeat Magnetic Resonance Imaging for Overall Cohort Stratified by Intracerebral Hemorrhage Location Total Fazekas Scale Score Variable

Baseline MRI Study

Follow-Up MRI Study

P Value

Ipsilateral hemisphere

2 (1e4)

3 (2e5)

< 0.0001

Contralateral hemisphere

3 (1e4)

3 (2e4)

< 0.0001

Entire brain

3 (1e4)

4 (2e5)

< 0.0001

All ICH cohort (n ¼ 65)

Lobar ICH cohort (n ¼ 51) Ipsilateral hemisphere

3 (2e4)

4 (2e5)

< 0.0001

Contralateral hemisphere

3 (2e4)

4 (2e5)

0.001

Entire brain

3 (2e4)

4 (2e5)

< 0.0001

1 (1e3.25)

2 (1e3.25)

0.06

Deep ICH cohort (n ¼ 14) Ipsilateral hemisphere Contralateral hemisphere

1.5 (1e3.25)

2 (1e3.25)

0.12

Entire brain

1.5 (1e3.25)

2.5 (1e3.25)

0.03

Data presented as median (interquartile range). MRI, magnetic resonance imaging; ICH, intracerebral hemorrhage.

WORLD NEUROSURGERY -: e1-e11, - 2019

www.journals.elsevier.com/world-neurosurgery

e5

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

Figure 2. Example of white matter injury progression. (A) Baseline (acute) magnetic resonance imaging scan showing a minimal amount of periventricular white matter hyperintensity on the left temporal horn;

DISCUSSION We found radiological evidence for WMIP in approximately one third of patients after ICH. This was noted in both lobar and deep ICHs and more frequently in the ipsilateral hemisphere. Furthermore, the ICH volume at presentation and the ventricular volume on the initial MRI scan were independently associated with WMIP. Also, the presence of WMIP was associated with unfavorable outcomes (defined as an mRS score, 3e6) at 3 months. These findings suggested that WMIP might represent a potential therapeutic target after ICH to improve patient outcomes. Several previous population-based and hospital-based cohort studies have examined WMIP in normal elderly patients, patients with ischemic stroke, and patients with Alzheimer’s disease.13-15 However, similar clinical studies of ICH have been lacking until very recently.16 The temporal and spatial progression of WMI has been described after experimental ICH2 and, recently, in the clinical setting.16 Our results demonstrated similar findings in our ICH cohort. In our study, WMIP, as assessed by the FZKS score, had developed in 35.4% of the patients, similar to the rate observed in the Rotterdam scan study (39%).17 Unlike the Rotterdam study, which was a prospective, population-based cohort study that had randomly selected participants aged 71  7 years and assessed WMIP during a 3-year interval, our study was limited to patients with ICH who were younger (65.6  14.9 years) and had a much shorter imaging follow-up period. This implies that the rate of WMIP might be accelerated after ICH. We found no difference in the interval between the repeat MRI scans, the prevalence of hypertension, or the use of antihypertensive agents between patients with and without WMIP, suggesting that our observations were unlikely to have been related to the merely elapse of time or imbalances in hypertension. We also found no parallel progression in other markers of small vessel

e6

www.SCIENCEDIRECT.com

Fazekas scale score, 1 (white arrow). (B) Repeat magnetic resonance imaging scan, 85 days later, showing more confluent periventricular white matter hyperintensities on the left temporal horn (black arrow).

disease, in particular, CMB and lacunar infarcts in our patients, suggesting that the increase in the FZKS score was unlikely to have been related to ICH-mediated progression of small vessel disease. Although hypertension has been the most consistently identified risk factor for WMI,18-20 the evidence has been inconsistent regarding its effect on WMIP. High blood pressure emerged as an independent predictor of WMI in the Rotterdam study17 and several others,21,22 but not in our study. These disparate findings could be attributed to the differences in the interval between MRI scans in the 2 studies or an artifact of our small sample size. However, our findings are not inconsistent with those of the multinational Leukoaraiosis and Disability study, in which hypertension was also not a risk factor for WMIP.23 Several studies have previously examined the association between the extent of WMI and outcomes for patients with ICH.4-6 They found that the severity of WMI on the baseline scan was associated with worse functional outcomes or mortality after spontaneous ICH.4-6 We found no difference in the baseline severity of WMI between patients with versus without WMIP, and the severity of WMI on the baseline MRI scan was not associated with unfavorable 90-day outcomes in our study. Therefore, our finding of an association between WMIP and functional outcomes after ICH might be novel in this regard. We found that the severity of WMI was greater with lobar than with deep ICH. Previous studies have reported that the patterns of WMI might be different between hypertensive arteriopathy and cerebral amyloid angiopathy.24 We found that a larger parenchymal ICH volume on the initial scan was independently associated with WMIP. The attributed pathophysiological mechanisms predisposing to WMI are speculative at present. One explanation is that the increased intracranial pressure caused by a direct mass effect of a large parenchymal hematoma

WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.089

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

Table 4. Patient Characteristics Stratified by White Matter Injury Progression WMIP (n [ 23)

No WMIP (n [ 42)

P Value

OR (95% CI)

68.3  15.4

64.1  14.7

0.287

NA

Female sex (n; %)

9 (39.1)

23 (54.8)

0.228

0.53 (0.19e1.49)

White race (n; %)

18 (78.3)

34 (81.0)

1.000

0.85 (0.24e2.97)

Preadmission mRS score (median; IQR)

0 (0)

0 (0)

0.247

NA

Admission GCS score (median; IQR)

13 (4)

15 (4)

0.092

NA

SBP (mm Hg)

163.3  28.6

158.1  34.1

0.373

NA

DBP (mm Hg)

86.3  14.6

86.5  22.2

0.959

NA

Hypertension

15 (65.2)

24 (57.1)

0.525

1.41 (0.49e4.03)

Diabetes mellitus

4 (17.4)

7 (16.7)

1.000

1.05 (0.27e4.06)

Hyperlipidemia

10 (43.5)

16 (38.1)

0.672

1.25 (0.45e3.51)

Atrial fibrillation

4 (17.4)

2 (4.8)

0.174

4.21 (0.71e25.04)

Coronary artery disease

2 (8.7)

9 (21.4)

0.335

0.35 (0.07e1.78)

Characteristic Age (years)

BP on admission

Pre-existing risk factors (n; %)

ICH history

1 (4.3)

0 (0)

0.354

NA

IS or TIA history

3 (13.0)

4 (9.5)

0.691

1.43 (0.29e7.00)

Current smoking

4 (17.4)

10 (23.8)

0.775

0.67 (0.19e2.45)

Alcohol consumption

6 (26.1)

12 (28.6)

0.831

0.88 (0.28e2.78)

Antihypertensive

11 (47.8)

23 (54.8)

0.592

0.76 (0.27e2.10)

Antiglycemic

3 (13.0)

7 (16.7)

0.978

0.75 (0.17e3.23)

Antiplatelet

10 (43.5)

13 (31.0)

0.313

1.72 (0.60e4.92)

Oral anticoagulant

3 (13.0)

4 (9.5)

0.691

1.43 (0.29e7.00)

Statin

7 (30.4)

12 (28.6)

0.875

1.09 (0.36e3.33)

131.1  40.6

128.3  45.0

0.802

NA

2 (8.7)

2 (4.8)

0.610

1.90 (0.25e14.50)

Preadmission medication (n; %)

Laboratory data on admission Glucose (mg/dL) eGFR <60 mL/min/1.73 m (n; %) 2

5.9  1.0

5.9  1.0

0.970

NA

LDL-C (mg/dL)

95.1  14.7

100.7  42.4

0.674

NA

PTT (seconds)

28.1  4.1

28.4  5.4

0.813

NA

3 (13.0)

6 (14.3)

1.000

0.90 (0.20e3.99)

262.7  94.9

238.8  87.5

0.310

NA

2.6  1.6

2.0  1.4

0.137

NA

128.2  93.7

118.0  96.3

0.683

NA

0.730

0.67 (0.20e2.23)

17 (73.9)

34 (81.0)

HbA1c (%)

INR >1.3 (n; %) Platelet count (K/mL) Neuroimaging findings Interval from ICH onset to baseline MRI (days) Interval from baseline to repeat MRI (days) Hematoma location (n; %) Lobar

WMIP, white matter injury progression; OR, odds ratio; CI, confidence interval; NA, not applicable; mRS, modified Rankin scale; IQR, interquartile range; GCS, Glasgow coma scale; BP, blood pressure; SBP, systolic blood pressures; DBP, diastolic blood pressures; ICH, intracerebral hemorrhage; IS, ischemic stroke; TIA, transient ischemic attack; eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c; LDL-C, calculated low-density lipoprotein cholesterol; PTT, partial thromboplastin time; INR, international normalized ratio; MRI, magnetic resonance imaging; IVH, intraventricular hemorrhage; CMB, cerebral microbleeding. Continues

WORLD NEUROSURGERY -: e1-e11, - 2019

www.journals.elsevier.com/world-neurosurgery

e7

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

Table 4. Continued Characteristic Nonlobar

WMIP (n [ 23)

No WMIP (n [ 42)

P Value

OR (95% CI)

6 (26.1)

8 (19.0)

10 (43.5)

5 (11.9)

0.004

5.69 (1.64e19.78)

23.7  21.3

12.5  10.6

0.007

NA

3.0  6.7

1.5  5.6

0.354

NA

31.4  18.9

22.8  12.6

0.031

NA

2 (2e4)

3 (1e4.25)

0.64

NA

Baseline contralateral

3 (1.25e4)

3 (1e4.25)

0.49

NA

Baseline entire brain

3 (2e4)

3 (1e4.25)

0.51

NA

10 (42)

14 (34)

0.6

1.37 (0.49e3.89)

2 (8)

5 (12.5)

0.7

0.63 (0.11e3.57)

mRS score at discharge, 3e6 (n; %)

12 (52.2)

17 (40.5)

0.518

1.60 (0.58e4.47)

mRS score at 3 months, 3e6 (n; %)

8 (34.8)

5 (11.9)

0.049

3.95 (1.11e14.03)

IVH (n; %) 3

Parenchymal ICH volume (cm ) IVH volume (cm3) 3

Ventricular volume on initial MRI (cm ) Fazekas scale score (median; IQR) Baseline ipsilateral

CMB present on baseline MRI (n; %) Lacunar infarcts present on baseline MRI (n; %) Unfavorable outcome

WMIP, white matter injury progression; OR, odds ratio; CI, confidence interval; NA, not applicable; mRS, modified Rankin scale; IQR, interquartile range; GCS, Glasgow coma scale; BP, blood pressure; SBP, systolic blood pressures; DBP, diastolic blood pressures; ICH, intracerebral hemorrhage; IS, ischemic stroke; TIA, transient ischemic attack; eGFR, estimated glomerular filtration rate; HbA1c, hemoglobin A1c; LDL-C, calculated low-density lipoprotein cholesterol; PTT, partial thromboplastin time; INR, international normalized ratio; MRI, magnetic resonance imaging; IVH, intraventricular hemorrhage; CMB, cerebral microbleeding.

might reduce the cerebral perfusion pressure and cause secondary ischemic injury to the white matter.1,25,26 Another explanation is that lysis of larger hematomas might release more toxic hemoglobin degradation products, resulting in the greater generation of free radicals or inflammatory cytokines causing WMI.1,25,26 We also found that a larger ventricular volume on the initial MRI scan was associated with WMIP. Evidence has linked hydrocephalus to the occurrence of WMI and its progression,27 presumably as a result of reduced blood flow in the periventricular white matter28,29 and the mechanical pressure of hydrocephalus.30 However, we also found that although the mean volume of IVH did not differ between the patients with and without WMIP, we found a trend for an association between the presence of IVH and WMIP. Collectively, these findings have led us to ponder whether the association between ventricular

enlargement and WMIP in our study was merely an epiphenomenon of a larger parenchymal ICH volume. Our study had some limitations attributable to its retrospective nature, nonstandardized selection for MRI assessments, and small sample size. A relatively small proportion of our patients with ICH had undergone a follow-up MRI scan after discharge, which raised the possibility of selection bias and the nongeneralizability of our results. This was largely due to a practice bias at our institution, because MRI tends to be more commonly performed for patients with lobar ICH to assess for the presence of microbleeding or to exclude an underlying secondary cause and not for those with deep ICH, which are often thought to be hypertensive in nature. Also, not all patients who had undergone a baseline MRI scan had undergone a repeat MRI scan. Our policy is to repeat MRI 8e12 weeks after the initial MRI scan to rule out an underlying lesion that might have

Table 5. Binary Logistic Regression Analysis of Risk Factors for White Matter Injury Progression Characteristic

Adjusted P Value

Adjusted OR

95% CI

Admission GCS score

0.603

0.944

0.759e1.174

Interval from baseline to repeat MRI

0.133

1.006

0.998e1.014

IVH

0.304

2.193

0.491e9.790

Parenchymal ICH volume

0.007

1.067

1.018e1.119

Ventricular volume on initial MRI

0.007

1.073

1.019e1.130

Baseline FZKS score

0.138

0.677

0.405e1.133

OR, odds ratio; CI, confidence interval; GCS, Glasgow coma scale; MRI, magnetic resonance imaging; IVH, intraventricular hemorrhage; ICH, intracerebral hemorrhage; FZKS, Fazekas scale.

e8

www.SCIENCEDIRECT.com

WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.089

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

been missed on the initial MRI study. Therefore, patients for whom the initial MRI study was informative or who had died or had been lost to follow-up did not undergo a follow-up MRI scan. We also used the FZKS, a semiquantitative visual rating scale, to assess for the presence of WMI, instead of a fully quantitative automated volumetric method. This was largely driven by the limited availability of DICOM (digital imaging and communications in medicine) images in a number of our patients, which prohibited us from automated volumetric assessments of white matter. However, previous studies have reported a significant correlation between the visual rating using the FZKS and quantitative volumetric assessments of WMI.31 Although the FZKS is well-validated, it has a ceiling effect and uses broad categories for severity, which could have led to an underestimation of WMI.13,19,32 Also, the 90-day period for the assessment of WMIP and functional outcomes could have been short. A longer follow-up period would have been more optimal. However, we were able to detect significant changes in WMI and an association with functional outcomes, despite the relatively short follow-up period. Another point that merits discussion and further research is that diverse pathologic processes, not solely demyelination or inflammation, might underlie the appearance of

REFERENCES 1. NINDS ICH Workshop Participants. Priorities for clinical research in intracerebral hemorrhage: report from a National Institute of Neurological Disorders and Stroke workshop. Stroke. 2005;36: e23-e41. 2. Wasserman JK, Schlichter LC. White matter injury in young and aged rats after intracerebral hemorrhage. Exp Neurol. 2008;214:266-275. 3. Lou M, Al-Hazzani A, Goddeau RP, Novak V, Selim M. Relationship between white-matter hyperintensities and hematoma volume and growth in patients with intracerebral hemorrhage. Stroke. 2010;41:34-40. 4. Caprio FZ, Maas MB, Rosenberg NF, et al. Leukoaraiosis on magnetic resonance imaging correlates with worse outcomes after spontaneous intracerebral hemorrhage. Stroke. 2013;44:642-646. 5. Sykora M, Herweh C, Steiner T. The association between leukoaraiosis and poor outcome in intracerebral hemorrhage is not mediated by hematoma growth. J Stroke Cerebrovasc Dis. 2017;26: 1328-1333. 6. Tveiten A, Ljøstad U, Mygland Å, Naess H. Leukoaraiosis is associated with short- and long-term mortality in patients with intracerebral hemorrhage. J Stroke Cerebrovasc Dis. 2013;22:919-925. 7. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The strengthening the reporting of observational studies in epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370: 1453-1457.

WMHs on MRI studies. Therefore, we could not entirely rule out that some of the observed hyperintensities were related to an underlying subclinical small vessel disease and that this could have influenced the functional outcomes, which has been extensively described as predicting outcomes. The preceding limitations can be only addressed in a prospective, systematic, well-designed, study. The use of advanced structural MRI techniques such as diffusion tensor imaging could further characterize and quantify the white matter changes at a microstructural level. Regardless, our study has provided preliminary and intriguing hypothesis-generating findings to lead the way for future advancements in this understudied area of ICH. CONCLUSIONS We have provided evidence suggesting that WMI progresses over time in patients with ICH and that WMIP is associated with worse outcomes, in accordance with recently reported studies.16 Our findings suggest that WMI might represent a potential therapeutic target for ICH. Future prospective larger studies are needed to confirm our findings and to carefully probe the mechanisms of WMI after ICH.

8. de Haan R, Limburg M, Bossuyt P, van der Meulen J, Aaronson N. The clinical meaning of Rankin “handicap” grades after stroke. Stroke. 1995;26:2027-2030.

16. Chen X, Chen X, Chen Y, Xu M, Yu T, Li J. The impact of intracerebral hemorrhage on the progression of white matter hyperintensity. Front Hum Neurosci. 2018;12:471.

9. Yao X, Xu Y, Siwila-Sackman E, Wu B, Selim M. The HEP score: a nomogram-derived hematoma expansion prediction scale. Neurocrit Care. 2015;23: 179-187.

17. van Dijk EJ, Prins ND, Vrooman HA, Hofman A, Koudstaal PJ, Breteler MMB. Progression of cerebral small vessel disease in relation to risk factors and cognitive consequences: Rotterdam Scan study. Stroke. 2008;39:2712-2719.

10. Wardlaw JM, Smith EE, Biessels GJ, et al. Neuroimaging standards for research into small vessel disease and its contribution to ageing and neurodegeneration. Lancet Neurol. 2013;12:822-838. 11. Fazekas F, Chawluk JB, Alavi A, Hurtig HI, Zimmerman RA. MR signal abnormalities at 1.5 T in Alzheimer’s dementia and normal aging. AJR Am J Roentgenol. 1987;149:351-356. 12. Helenius J, Henninger N. Leukoaraiosis burden significantly modulates the association between infarct volume and National Institutes of Health stroke scale in ischemic stroke. Stroke. 2015;46: 1857-1863. 13. Prins ND, Scheltens P. White matter hyperintensities, cognitive impairment and dementia: an update. Nat Rev Neurol. 2015;11:157-165. 14. Ramirez J, McNeely AA, Berezuk C, Gao F, Black SE. Dynamic progression of white matter hyperintensities in Alzheimer’s disease and normal aging: results from the Sunnybrook Dementia Study. Front Aging Neurosci. 2016;8:62. 15. Debette S, Markus HS. The clinical importance of white matter hyperintensities on brain magnetic resonance imaging: systematic review and metaanalysis. BMJ. 2010;341:c3666.

WORLD NEUROSURGERY -: e1-e11, - 2019

18. Allan CL, Zsoldos E, Filippini N, et al. Lifetime hypertension as a predictor of brain structure in older adults: cohort study with a 28-year followup. Br J Psychiatry J Ment Sci. 2015;206:308-315. 19. Smith EE, Saposnik G, Biessels GJ, et al. Prevention of stroke in patients with silent cerebrovascular disease: a scientific statement for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2017;48: e44-e71. 20. Swan GE, DeCarli C, Miller BL, et al. Association of midlife blood pressure to late-life cognitive decline and brain morphology. Neurology. 1998;51: 986-993. 21. Schmidt R, Fazekas F, Kapeller P, Schmidt H, Hartung HP. MRI white matter hyperintensities: three-year follow-up of the Austrian Stroke Prevention Study. Neurology. 1999;53:132-139. 22. Markus HS, Hunt B, Palmer K, Enzinger C, Schmidt H, Schmidt R. Markers of endothelial and hemostatic activation and progression of cerebral white matter hyperintensities: longitudinal results of the Austrian Stroke Prevention Study. Stroke. 2005;36:1410-1414. 23. Gouw AA, van der Flier WM, Fazekas F, et al. Progression of white matter hyperintensities and

www.journals.elsevier.com/world-neurosurgery

e9

ORIGINAL ARTICLE DEREN WANG ET AL.

incidence of new lacunes over a 3-year period: the Leukoaraiosis and Disability study. Stroke. 2008;39: 1414-1420. 24. Charidimou A, Boulouis G, Haley K, et al. White matter hyperintensity patterns in cerebral amyloid angiopathy and hypertensive arteriopathy. Neurology. 2016;86:505-511. 25. Yeo SS, Choi BY, Chang CH, et al. Periventricular white matter injury by primary intraventricular hemorrhage: a diffusion tensor imaging study. Eur Neurol. 2011;66:235-241. 26. Zuo S, Pan P, Li Q, Chen Y, Feng H. White matter injury and recovery after hypertensive intracerebral hemorrhage. BioMed Res Int. 2017;2017: 6138424. 27. Krauss JK, Regel JP, Vach W, et al. White matter lesions in patients with idiopathic normal pressure hydrocephalus and in an age-matched control group: a comparative study. Neurosurgery. 1997; 40:491-495 [discussion: 495-496].

e10

www.SCIENCEDIRECT.com

WMI PROGRESSION AFTER ICH

28. Momjian S, Owler BK, Czosnyka Z, Czosnyka M, Pena A, Pickard JD. Pattern of white matter regional cerebral blood flow and autoregulation in normal pressure hydrocephalus. Brain J Neurol. 2004;127(Pt 5):965-972. 29. Ziegelitz D, Starck G, Kristiansen D, et al. Cerebral perfusion measured by dynamic susceptibility contrast MRI is reduced in patients with idiopathic normal pressure hydrocephalus. J Magn Reson Imaging. 2014;39:1533-1542. 30. Jang SH, Choi BY, Chang CH, et al. The effects of hydrocephalus on the periventricular white matter in intracerebral hemorrhage: a diffuser tensor imaging study. Int J Neurosci. 2013;123:420-424.

lesions on MRI: visual rating and volumetrics. Neurology. 2004;62:1533-1539.

Conflict of interest statement: Deren Wang was supported by the National Natural Science Foundation of China (grants 81870923 and 81620108009) and the National Key R&D Program of China (grants 2017YFC1308401 and 2016YFC1300500-505). Ming Liu was supported by the National Natural Science Foundation of China (grant 81620108009) and the National Key R&D Program of China (grant 2016YFC1300500-505). Magdy Selim was supported by the National Institute of Neurological Disorders and Stroke (grant U01NS 074425) and American Heart Association (grant 15CSA24540001). Received 29 October 2018; accepted 8 February 2019

31. Kapeller P, Barber R, Vermeulen RJ, et al. Visual rating of age-related white matter changes on magnetic resonance imaging: scale comparison, interrater agreement, and correlations with quantitative measurements. Stroke. 2003;34:441-445.

Citation: World Neurosurg. (2019). https://doi.org/10.1016/j.wneu.2019.02.089

32. Prins ND, van Straaten ECW, van Dijk EJ, et al. Measuring progression of cerebral white matter

1878-8750/$ - see front matter ª 2019 Elsevier Inc. All rights reserved.

Journal homepage: www.journals.elsevier.com/worldneurosurgery Available online: www.sciencedirect.com

WORLD NEUROSURGERY, https://doi.org/10.1016/j.wneu.2019.02.089

ORIGINAL ARTICLE DEREN WANG ET AL.

WMI PROGRESSION AFTER ICH

SUPPLEMENTARY DATA

Supplemental Figure 1. Axial magnetic resonance images illustrating the Fazekas scale for white matter hyperintensities (WMHs). (A) Fazekas score 0. (B) Fazekas score 1, pencil-thin periventricular WMHs (white arrow) and faint punctate deep WMHs (white

WORLD NEUROSURGERY -: e1-e11, - 2019

arrow). (C) Fazekas score 2, periventricular WMHs, more visible on the left side; greater than the simple, pencil-thin pattern but not confluent. (D) Fazekas score 3, confluent periventricular and deep WMHs.

www.journals.elsevier.com/world-neurosurgery

e11