Clinical Characteristics and Natural History of Quasi-Moyamoya Disease

ARTICLE IN PRESS

Clinical Characteristics and Natural History of Quasi-Moyamoya Disease Meng Zhao, MD,*,†,‡ Zhiqin Lin, MD,*,† Xiaofeng Deng, MD,*,†,‡ Qian Zhang, MD,*,†,‡ Dong Zhang, MD,*,†,‡ Yan Zhang, MD,*,†,‡ Rong Wang, MD,*,†,‡ Shuo Wang, MD,*,†,‡ Zhongli Jiang, MD,*,† Hao Wang, MD,*,† and Jizong Zhao, MD*,†,‡

Background: Quasi-moyamoya disease (quasi-MMD) is a rare cerebrovascular disease and its clinical features and natural history remain unclear. The aim of the study is to describe the clinical characteristics and the natural histories of this disease, with analysis of the risk factors for future cerebrovascular events. Methods: We identified 64 patients with quasi-MMD from 693 moyamoya vasculopathy patients referred to our hospital between 2011 and 2015. Demographic data, associated disorders, clinical manifestation, angiographic findings, natural history, and risk factors for cerebrovascular events were analyzed. Results: Patients included in the study had a mean age of 31.5 years. A unimodal age distribution was noted. Atherosclerosis was the most frequently associated disorder. Forty-five (70.3%) patients had ischemic events as their initial clinical manifestation and 14 (21.9%) patients presented as hemorrhagic stroke. The majority of patients presented with Suzuki grades 3 and 4 (20.3% and 42.2%). The annual risk of cerebrovascular events was 19.4% per patient-year. Prior hemorrhage (HR 2.77, 95% CI 1.20-6.41) and ischemic stroke (HR 2.77, 95% CI 1.26-6.07) were 2 risk factors for future events. Conclusions: Several clinical characteristic differences were observed in our mainland China cohort compared with the Japanese and European cohorts. The annual risk of cerebrovascular events was relatively high in quasi-MMD patients. Patients with prior hemorrhage and ischemic stroke were inclined to have future cerebrovascular events. Close follow-up is needed for these patients. Key Words: Quasi-moyamoya disease—moyamoya disease—clinical characteristics—natural history. © 2017 National Stroke Association. Published by Elsevier Inc. All rights reserved.

From the *Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Beijing, China; †China National Clinical Research Center for Neurological Diseases, Beijing, China; and ‡Center of Stroke, Beijing Institute for Brain Disorders, Beijing, China. Received October 6, 2016; revision received November 19, 2016; accepted December 24, 2016. Grant support: This study was supported by the “13th Five-Year Plan” on National Science and Technology supporting plan (2015BAI12B04), the National Natural Science Foundation of China (81371292), and the Beijing Municipal Administration of Hospitals’ Mission Plan, Code: SML20150501. Address correspondence to Jizong Zhao, MD, Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, No. 6 Tiantanxili, Dongcheng District, Beijing 100050, China. E-mail: [email protected]. 1052-3057/$ - see front matter © 2017 National Stroke Association. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jstrokecerebrovasdis.2016.12.025

Journal of Stroke and Cerebrovascular Diseases, Vol. ■■, No. ■■ (■■), 2016: pp ■■–■■

1

ARTICLE IN PRESS M. ZHAO ET AL.

2

Introduction Moyamoya vasculopathy is a disorder characterized by progressive narrowing or occlusion of the intracranial internal carotid artery (ICA), with formation of abnormal vascular collateral networks at the base of the brain.1 This vasculopathy was first described by Takeuchi and Shimizu in 1957.2 Moyamoya disease (MMD) corresponds to isolated and primary moyamoya vasculopathy. Quasi-moyamoya disease (quasi-MMD) refers to moyamoya vasculopathy associated with various disease entities. Although terms like moyamoya syndrome, moyamoya phenomenon, and moyamoya-like vasculopathy all have been used to describe such disorder, it is clearly defined as quasi-MMD in the Guidelines for Diagnosis and Treatment of Moyamoya Disease.3 Despite major clinical and experimental advances in MMD, few researches put the spotlight on quasi-MMD.4-7 A Japanese nationwide survey reported that the prevalence and annual incidence of quasi-MMD are approximately 10 times lower than those of MMD.4 Because of its rarity, the clinical features and the natural history of this disease remain unclear. Thus, limitations to predict the prognosis and determine the treatment strategy still existed. Therefore, we conducted a study of the natural history and clinical characteristics of quasi-MMD to help elucidate the clinical features of this rare disease and to help guide management decisions in these patients.

Materials and Methods Patient Selection The study was approved by the Beijing Tiantan Hospital Research Ethics Committee. From 2011 to 2015, a total of 693 patients were identified as moyamoya vasculopathy in the Tiantan Hospital Stroke Center moyamoya database. Of these patients, 64 were diagnosed with quasi-MMD based on the criteria prepared by the Research Committee on Spontaneous Occlusion of the Circle of Willis (Moyamoya Disease) of Japan.3 According to these criteria, patients were diagnosed with quasi-MMD if they had typical angiographic MMD features associated with disorders as follows: atherosclerosis, autoimmune disease, meningitis, neurofibromatosis type I, brain tumor, Down’s syndrome, head injury, irradiation, Turner syndrome, Alagille syndrome, Williams syndrome, Noonan syndrome, Marfan syndrome, nodular sclerosis, Ito nevus, incontinence of pigment, Hirschsprung disease, diabetes mellitus IA, Prader–Willi syndrome, Wilms tumor, primary oxalosis, sicklemia, Fanconi anemia, spherocyte, eosinophilic granuloma, plasminogen abnormality II, leptospirosis, enterovirus infection, protein S deficiency, pyruvate kinase deficiency, fibrous dysplasia, polycystic kidney, retinitis pigmentosa, and oral contraceptives.3 A universally accepted definition of atherosclerosis had not been estab-

lished. Determination of atherosclerosis-associated quasiMMD was performed by 2 authors (Y.Z. and R.W.) independent of each other. The criteria were developed primarily based on cerebrovascular imaging with supporting evidence from the patient’s clinical profile. Any patient with a radiography demonstration similar to MMD and who met 3 of the following 4 qualifications was identified as an atherosclerosis-associated quasi-MMD patient: (1) imaging demonstrated eccentric calcified plaques in the intracranial vasculature or the carotid bifurcation; (2) confirmed atherosclerosis in other site; (3) unilateral stenosis; and (4) hyperlipidemia. Other concurrent disorders were confirmed by clinical test or by medical history.

Chart Review We retrospectively reviewed the clinical records for each patient from presentation. Previous clinic records were also obtained and reviewed if necessary. Baseline information was noted at initial presentation including the following: age, sex, ethnicity, mode of presentation, hypertension, cerebral aneurysm, familial MMD, hyperlipidemia, diabetes mellitus, hyperthyroidism, significant alcohol and/or tobacco use, oral contraceptive use, autoimmune disease, and parental stroke history. Clinical onset symptoms were divided into ischemic symptoms (including transient ischemic attacks [TIAs] and stroke), hemorrhage (including subarachnoid hemorrhage [SAH], intracerebral hemorrhage [ICH], and intraventricular hemorrhage [IVH]), seizure, and headache. Headache was accounted as initial manifestation if it was the only symptom of the patient. Radiological studies including digital subtraction angiography (DSA), magnetic resonance imaging (MRI), computed tomography angiography (CTA), and CT perfusion were evaluated. DSA was routinely performed on all the patients at admission to our hospital and repeated if necessary during their course of disease. Two independent and blinded interventional neurosurgeons (Y.Z. and R.W.) were enrolled to analyze angiography data for stenosis sites, extent of involvement, sources of collateral flow, and other cerebrovascular abnormalities. The angiographic stages of quasi-MMD were estimated according to the Suzuki angiographic stage classification.1 Disagreements were resolved by consensus.

Clinical Follow-up The follow-up period was defined as the time between the initial clinical presentation and the last clinical followup or revascularization, whichever occurred first. If a patient had had a diagnosed moyamoya before the referral to our hospital, we required the information of initial clinical presentation by doing a chart review of the patient’s earlier clinical notes. Thus, the follow-up period was extended. Trained study staff interviewed patients or their caregivers by telephone. We inquired about the following:

ARTICLE IN PRESS CHARACTERISTICS AND NATURAL HISTORY OF QUASI-MMD

(1) new ischemic events; (2) new hemorrhagic events; (3) TIA frequency; and (4) death events. We defined recurrent ischemic events as new symptomatic neurological deterioration lasting at least 24 hours that was not attributable to a nonischemic cause, or new symptomatic neurological deterioration accompanied by neuroimaging evidence of new brain infarction; hemorrhagic events were defined as radiographic presentation of ICH, SAH, or IVH. TIAs were not defined as a recurrent cerebrovascular event. The annual event risk was the ratio of the number of cerebrovascular events (initial presentation excluded) and the total number of patientyears of follow-up.

Statistical Analysis Descriptive summaries were reported as mean ± standard deviation for continuous variables and as frequency or percentage for categorical variables. We performed both univariate and multivariate time-to-event analyses to identify the risk factors associated with future cerebrovascular events. The log-rank test and Cox proportional hazards regression model using future cerebrovascular events (ischemic events and hemorrhagic events) as censoring events were employed. Risk factors were first assessed for association with the cerebrovascular events in univariate analysis using log-rank tests. All those factors with P ≤ .20 were included as candidate independent variables for multivariate analysis using the Cox proportional hazards model. All statistical analyses were carried out using R (R core team, Vienna, Austria) statistical program.

3

Table 1. Clinical features of 64 patients with quasi-MMD Clinical features

Value

Total number of patients Male/female ratio Age (%) Mean <18 ≥18 Ethnicity (%) Han Mongols History of risk factors (%) Familial MMD Hypertension Aneurysm Smoking or alcohol use (%) Diabetes (%) Hyperlipidemia (%) Initial symptom (%) Hemorrhage TIA Infarction Headache Seizure Type of hemorrhage (%) ICH IVH SAH

64 33/31 31.5 ± 13.89 10 (15.6) 54 (84.4) 63 (98.4) 1 (1.6) 2 (3.1) 16 (25.0) 2 (3.1) 8 (12.5) 6 (9.4) 21 (32.8) 14 (21.9) 23 (35.9) 22 (34.4) 3 (4.7) 2 (3.1) 4 (6.3) 8 (12.5) 2 (3.1)

Abbreviations: ICH, intracranial hemorrhage; IVH, intraventricular hemorrhage; MMD, moyamoya disease; SAH, subarachnoid hemorrhage; TIA, transient ischemic attack.

Results General Patient Characteristics Epidemiological information of the 64 quasi-MMD patients is summarized in Table 1. Patients included in the study had a mean age of 31.5 ± 13.9 years (range, 2-62 years). The female/male ratio is .9:1 (male = 33 [51.6%]). Female predominance could not be seen in adult nor in pediatric patients in our cohort. Age distribution of the quasi-MMD patients was found to have a unimodal appearance: the highest peak of detection rate was observed at ages 30-40 years. Both women and men followed the same pattern in age distribution (Fig 1). The associated disorders are summarized in Figure 2. We found atherosclerosis in 32 (50.0%) patients, hyperthyroidism in 12 (18.8%) patients, anemia in 5 (7.8%) patients, tumor/radiation in 5 (7.8%) patients, renal hypertension in 2 (3.1%) patients, diabetes mellitus IA in 3 (4.7%) patients, tetralogy of Fallot in 1 (1.6%) patient, and brain injury in 1 (1.6%) patient.

Clinical Manifestation Initial clinical manifestations are summarized in Table 1. Forty-five (70.3%) patients had ischemic events as their

initial clinical manifestation; 23 (35.9%) patients presented with TIA; and 22 (34.4%) patients presented with infarction. Fourteen (21.9%) patients presented as hemorrhagic stroke. Most of the hemorrhages were IVH (n = 8, [57.1%]), while ICH came second (n = 4, [28.6%]). Three of the 4 patients presented as ICH are atherosclerosisassociated quasi-MMD patients. These 3 patients also had hypertension. The other patient had anemia. These 4 ICH patients presented with sudden contralateral motor weakness at onset. Two hemorrhagic quasi-MMD patients presented as SAH (14.3%). Ventricular drainage (VD) or ventriculoperitoneal (VP) shunt placement was performed in 3 patients whose CT demonstrated severe hydrocephalus before admission to our hospital. Other modes of presentation included headache (n = 3, [4.7%]) and seizure (n = 2, [3.1%]). Adults had a higher rate of hemorrhagic presentation than children (22.6% versus 9.1%).

Angiographic Findings Angiographic findings are detailed in Table 2. Half of the quasi-MMD patients (n = 32, [50%]) are demonstrated as unilateral lesion. The majority of patients presented with Suzuki grades 3 and 4. Steno-occlusive

ARTICLE IN PRESS M. ZHAO ET AL.

4

Figure 1.

Age distribution of the quasi-moyamoya patients. (A) The whole group. (B) Male. (C) Female.

lesion was seen in ICA the most (56 lesion sites in total, 28 on each side), and middle cerebral artery came second (33 lesion sites in total, 23 on the right side and 10 on the left side). For anterior cerebral artery, only 2 on the right side and 1 on the left side reached the qualification

of severe stenosis. The spontaneous collateral vessels from ophthalmic artery or external carotid artery could be seen in 28.1% (n = 18) of the cases. CT perfusion was available in 53 patients. Fifty (94.3%) patients showed various kinds of hypoperfusion and 34 (64.2%) patients demonstrated decreased cerebral blood flow (CBF).

Follow-up

Figure 2.

Associated disorders of quasi-moyamoya patients.

Table 2. Summary of angiography findings

Suzuki stage 1 2 3 4 5 6 Normal Severe stenosis-occlusion ICA ACA MCA No severe stenosis

Right side

(%)

Left side

(%)

4 6 14 19 9 1 11

6.3 9.4 21.9 29.7 14.1 1.6 17.2

3 6 10 20 3 1 21

4.7 9.4 15.6 31.3 4.7 1.6 32.8

28 2 23 11

43.8 3.1 35.9 17.2

28 1 10 25

43.8 1.6 15.6 39.1

Abbreviations: ACA, anterior cerebral artery; ICA, internal carotid artery; MCA, middle cerebral artery.

In our 64 patients, the mean follow-up was 3.8 years (range, .1-16.4 years). During a total of 191.0 patientyears, 8 hemorrhages and 29 ischemic events reoccurred, corresponding to an overall annual hemorrhage risk of 4.2% per patient-year and an annual ischemic event rate of 15.2% per patient-year. In the present study, all of the quasi-MMD patients with a hemorrhagic presentation experienced cerebrovascular events during follow-up, and 8 of the events were hemorrhagic. Most of the recurrent hemorrhages were IVH (n = 7). ICH happened in 1 patient. All of the patients had IVH as initial presentation. Patients had motor weakness (n = 3) and altered consciousness (n = 5) as presentation when hemorrhage reoccurred. In the hemorrhagicpresented quasi-MMD patients, the annual rebleeding risk was 18.6% per patient-year. As for the ischemic-presented patients, the annual event risk is 19.4%. Patients presented with cerebral infarction had an annual event risk of 28.7% per patient-year, whereas patients presented with TIA had an annual event risk of 4.1% per patient-year. Among the 29 patients who had recurrent ischemic events, 21 (72.4%) were diagnosed based on new site of infarction with radiological evidence (CT, 15; MRI, 6) and 8 were diagnosed based on the persistent neurological symptoms lasting longer than 24 hours. Lacunar infarcts (LACIs) were detected in almost half of the patients (n = 10, [47.6%]); partial anterior circulation infarcts came second (n = 7, [33.3%]); and posterior circulation infracts accounted for 4 cases (19.0%). No patient had total anterior circulation infarcts. Twelve ischemic sites were in right hemispheres and 9 in left hemispheres. These newly developed infarction sites were all located at the same hemisphere as the initial infarction. We compared events-free survival of 4 most common associated disorder groups in our study using log-rank test for multiple comparisons (as shown in Fig 3, F). We found no

ARTICLE IN PRESS CHARACTERISTICS AND NATURAL HISTORY OF QUASI-MMD

significance in this multiple comparison (P = .60). Then multiple comparison tests were adjusted for pairwise test using Bonferroni method (K = 6). We detected no significance between each of the paired groups (all P > .05).

Risk Factors for Future Events In the univariate analysis, we identified that prior hemorrhage (HR 2.77, 95% CI 1.20-6.41), prior infarction (HR 2.77, 95% CI 1.26-6.07), and TIAs (HR .11, 95% CI .06.22) were associated with future cerebrovascular events (all P < .05). VD or VP shunt placement and hemiparesis with aphasia (both P < .2) were also entered into multivariate analysis. In Cox proportional hazards regression model, prior hemorrhage (HR 5.67, 95% CI 1.1528.04, P = .03) and prior infarction (HR 5.84, 95% CI 1.2028.35, P = .03) were independently associated with future cerebrovascular events. The Kaplan–Meier curves were shown in Figure 3. A summary of univariate and multivariate analyses of risk factors for hazard events is provided in Table 3.

Discussion Despite major advances of clinical and laboratory resources in MMD, a few research have been reported on clinical characteristics and natural history of quasiMMD. To the best of our knowledge, the natural history of Asian quasi-MMD patients has not been described in the literature before. This large cohort study extends our knowledge of the clinical features, radiography findings, and natural history of patients with quasi-MMD.

Clinical Characteristics Some large moyamoya cohort research had been conducted in recent years, but quasi-MMD was not focused in these studies.8,9 A nationwide survey conducted in Japan reported that the annual incidence ratio of quasi-MMD to MMD is .10, and the prevalence ratio is .07.10 Unlike in Asia, the proportion of quasi-MMD in moyamoya vasculopathy seems to be higher in Western countries.6,11 The prevalence ratio of quasi-MMD to MMD in our database (.10, 64/649) is similar to the Japanese study. Whether quasi-MMD is an indistinct disease is a controversial issue.12 A recent European study concluded that quasi-MMD represents a distinct disease with different ethnic clinical features.7 Some clinical features of quasi-MMD presented in our study differed from those in previous reports. Studies in Japan or Western countries7,13-17 demonstrated a female/ male ratio from 1.57 to 4.25. However, female predominance of quasi-MMD is not noted in our series (female/male, .94). The bipeak age distribution was widely observed in Japanese and European studies, but our cohort did not demonstrate it. The age distribution of our study is similar

5

to the moyamoya vasculopathy cohort reported in Taiwan.5 The fact that half of our patients were associated with atherosclerosis, which usually affected middle-aged people or older, may explain the single peak distribution. Similar to previous reports,4,5,18 atherosclerosis and hyperthyroidism are the top 2 associated disorders (44/64, [68.8%]) in our study. However, a proportion of atherosclerosis-associated quasi-MMD is higher in our cohort (n = 32, [50%]) compared to 29.0%-32.4% in previous studies.4,5 In accordance with the Japanese study,4 the clinical characteristics of this atherosclerosisassociated quasi-MMD were similar to the characteristics of the whole cohort. However, unilateral lesion was seen in 18 (66.7 %) cases of their atherosclerosis patients, whereas in our cohort, nearly all of the atherosclerosis-associated quasi-MMD patients (31/32) presented as unilateral lesion. Besides histopathology diagnosis, high-resolution MRI is helpful in differentiating MMD artery from atherosclerotic stenosis. Imaging presentation of eccentric artery stenosis with a relatively larger outer vessel wall diameter is usually observed in atherosclerosis-associated quasi-MMD.19,20 Because of the lack of histopathological confirmation and the limited high-resolution MRI data, atherosclerosis-associated quasi-MMD may be overdiagnosed in our study. Some observational studies of the relationship between hyperthyroidism and quasi-MMD have been reported. Hyperthyroidism-associated quasi-MMD often presented as unilateral stenosis or occlusion.21,22 But only 1 hyperthyroid patient presented as unilateral quasiMMD in our study. Ischemic events were the major clinical manifestation of these patients. Cerebrovascular hemodynamic changes due to thyrotoxicosis might be the cause.21 In Li’s case series of 21 hyperthyroidism-associated moyamoya patients,23 cerebral infarction was found in 19 patients (90.4%), and IVH was observed in 1 case (4.8%). In our cohort, 4 of the 12 (33.3%) hyperthyroidismassociated patients had infarction presentation, and 2 (16.7%) had initial presentation of hemorrhage. Also, the findings of the current study do not support the previous research that the ischemic symptoms of quasiMMD disease could disappear after control of hyperthyroidism.21,22 Despite regularly taking antithyroid medications, 3 of the 4 ischemia patients experienced ischemic events again. The clinical impact of hyperthyroidism on quasi-MMD is still not thoroughly elucidated. Quasi-MMD exhibits a variety of clinical presentations, with ischemic presentation such as TIA and infarction being the most common in our cohort. This is consistent with the previous large cohort of MMD research in Japan and China.4,8 As for the hemorrhagic manifestation, IVH and ICH are the main types of hemorrhage in quasi-MMD patients, which is similar to definitive MMD. In the nationwide survey in Japan, intracranial hemorrhage is less frequent in quasi-MMD than in MMD (7% versus 40%). In contrast, more hemorrhagic presentations

ARTICLE IN PRESS 6

M. ZHAO ET AL.

Figure 3. Kaplan–Meier curves for future events-free survival. (A) Hemorrhage-presented patients compared to non–hemorrhage-presented patients. (B) Hemorrhage-presented patients compared to non–hemorrhage-presented patients in the first 12 months. (C) Infarction-presented patients compared to TIApresented patients. (D) Suzuki grade 0-2 patients compared to Suzuki grade 3-5 patients. (E) Male-to-female patients. (F) Future events-free survival of different associated disorders in the first 12 months. Abbreviation: TIA, transient ischemic attack.

in quasi-MMD were observed (n = 14, [21.9%]) in our cohort. Half of the quasi-MMD patients (n = 7) with hemorrhagic presentation are atherosclerosis associated. In these 7 patients, 3 had hypertension, 3 had aspirin and/or clopidogrel treatment for secondary prevention of stroke after TIA, and 1 had no risk factor. Hypertensive ICH

might be the cause of bleeding in 3 patients. Our cohort had a higher hemorrhagic proportion than the quasiMMD nationwide survey conducted in Japan (4%).4 But it is largely in line with the hemorrhagic rate in definite MMD both in China and in Japan.8,24 Our study showed that ICH in quasi-MMD at presentation bears more

ARTICLE IN PRESS CHARACTERISTICS AND NATURAL HISTORY OF QUASI-MMD

7

Table 3. Risk factors for future cerebrovascular events Univariate analysis

Multivariate analysis

Clinical features

HR

95% CI

P value

HR

95% CI

P value

Age presentation <18 Female Prior hemorrhage Prior infarction TIA as initial symptom Atherosclerosis Hyperthyroidism Anemia Diabetes (all types) Hypertension Suzuki grade <3 Unilateral lesion Seizure Consciousness disturbance Aphasia Hemiparesis Aphasia and hemiparesis Headache VD/VP shunt

.77 .98 2.77 2.77 .11 1.09 .61 1.67 .71 1.02 .82 1.02 .91 1.06 1.35 1.55 2.12 .99 2.51

.32-1.84 .50-1.89 1.20-6.41 1.26-6.07 .06-.22 .56-2.11 .27-1.37 .46-5.99 .21-2.46 .48-2.20 .39-1.74 .53-1.99 .23-3.57 .43-2.61 .54-3.35 .80-3.02 .72-6.24 .41-2.38 .42-14.92

.58 .95 <.001 <.001 <.001 .80 .30 .33 .64 .95 .62 .94 .89 .89 .47 .21 .06 .98 .11

— — 5.67 5.84 .39 — — — — — — — — — — — 1.86 — 1.29

— — 1.15-28.04 1.20-28.35 .06-2.34 — — — — — — — — — — — .79-4.39 — .35-4.69

— — .03 .03 .30 — — — — — — — — — — — .15 — .70

Abbreviations: CI, confidence interval; HR, hazard ratio; TIA, transient ischemic attack; VD/VP shunt, ventricular drainage or ventriculoperitoneal shunt before admission to our hospital.

resemblance with primary ICH, with basal ganglia being the most common site. Three of the 4 ICH patients had hypertension. Taking the site into consideration, these 3 hemorrhagic events might be due to hypertension. SAH is a rare presentation in quasi-MMD and definitive MMD. In a previous study, SAH accounted for 9.7% of MMD/ quasi-MMD cases.25 Other than that study, only case reports exist. Thus, the rate of SAH is not clear in quasi-MMD. In our institution, SAH was more commonly observed in quasi-MMD patients than in definitive MMD patients (3.1% versus 1.7%). No asymptomatic quasiMMD patient was detected in the present study. In a demographic Japanese study in 2008, patients with asymptomatic MMD were detected more than expected. The brain checkup system was attributable to the identification of asymptomatic patients in Japan.13,26 The relatively low economic status in some area of mainland China restricted the availability of DSA and CTA checkup. We believe that a certain amount of asymptomatic quasiMMD patients was not identified in China. Around half of our quasi-MMD patients have unilateral lesion and most of the severe stenosis-occlusion lesions were seen in the ICA. This is in accordance with recent reports.4 The main associated disease of quasi-MMD is atherosclerosis, and atherosclerosis often affects one side of the intracranial artery. Cerebrovascular lesion of other associated diseases such as neurofibromatosis type 1 disease also has a high ratio of unilateral presentation.27,28 In terms

of cerebral perfusion, hypoperfusion is commonly seen in quasi-MMD patients and CBF was impaired in almost all cases.4 Hypoperfusion was also observed in most of the cases in our study, but we also found a lower rate (63.2%) of decreased CBF overall. As CBF decreases after the reserve capacity has failed in the development of hypoperfusion, a possible explanation for this might be that the patients referred to our hospital were in a relatively early stage of perfusion disorder.

Natural History Rebleeding and future ischemic stroke are the main causes of poor clinical outcome in MMD, and possibly quasi-MMD.29 The aim of the quasi-MMD treatment is to avoid the progression of the vasculopathy and to prevent future cerebrovascular events. Therefore, understanding the natural history of quasi-MMD, especially related risk factors of future hazard events, is of great importance. Although quasi-MMD has certain heterogeneity due to its different underlying causes, we found no significant association between different disorders and events-free survival (Fig 3, F). In MMD, subsequent episodes of bleeding occurred at a higher rate in patients with a hemorrhagic presentation, and subsequent ischemic events occurred at a higher rate in patients with an ischemic presentation.30,31 Our cohort has confirmed these 2 risk factors in quasiMMD. The North American study on quasi-MMD also

ARTICLE IN PRESS M. ZHAO ET AL.

8

reported that higher event risk could be seen among smokers.6 However, smoking history, hypertension, hemiparesis at presentation, and higher radiography demonstration grade are not identified as risk factors for future cerebrovascular events in our study. Our study demonstrates that only prior infarction and prior hemorrhage are independent risk factors for a subsequent cerebrovascular event. Rebleeding has been reported with various incidence rates in moyamoya patients.29,30 In Liu’s cohort,32 25 (41.7%) of 60 conservatively treated hemorrhagic patients experienced intracranial hemorrhages during a follow-up period, which ranged from 12 to 300 months. In Morioka’s study,33 29 incidents of rebleeding were observed in 21 moyamoya patients. In our study, 8 of 14 quasi-MMD patients with initial hemorrhagic presentation experienced rebleeding during follow-up. All of the 8 patients who had rebleeding had IVH as initial presentation. Seven of the recurrent hemorrhage sites were the same as initial presentation and 1 had ICH in basal ganglia. Morioka reported that angiographic dilatation and branch extension of the anterior choroidal and posterior communicating arteries are predictors of hemorrhage, mainly IVH and ICH, in MMD patients.33 Of the 14 patients who had hemorrhagic events as presentation in our cohort, only 3 had angiographic dilatation and branch extension of the anterior choroidal or posterior communicating arteries. Aneurysms are another cause of hemorrhage in moyamoya patients. The association of MMD and aneurysm has been reported in 3%-14% of moyamoya patients.34 We only detected 2 unruptured aneurysms in our cohort. However, small peripheral artery aneurysms that arose from the collateral or moyamoya vessels could rupture and regress spontaneously. Thus, we could not rule out the possibility of aneurysm rupture being the cause of intracranial hemorrhage.35 The annual risk of these hemorrhagic events was highest in the first year and decreased sharply afterward. Such pattern of unbalanced annual risk is also observed in other cerebral vascular diseases like arteriovenous malformation.36 Despite agreement on the role of prior hemorrhage as a risk factor for recurrent hemorrhage, multiple other factors for rebleeding are mentioned in the literature: age >36 years at onset, microbleeds on MRI, female sex, and stroke presentation within 3 years.6,37,38 The role of these risk factors is controversial. In our study, only initial hemorrhagic presentation is a risk factor for rebleeding, other risk factors are insignificant. Ischemia is the most frequent symptom in both quasiMMD and MMD.4-6,9,14,39 The majority of the quasi-MMD patients in our study (n = 45, [70.3%]) presented with ischemic symptoms. The future cerebrovascular event risk for ischemic patients of different studies varies significantly. Hallemeier reported a cumulative 5-year risk of stroke of 82% in adult patients with ischemic presentation15; Chiu,14 Gross,6 Kuroda,24 Cho,30 and Kobayashi29 reported

annual stroke risk ranging from 3.2% to 13.3%. In our study, the annual stroke risk for ischemic patients is 19.4%. Actually, the annual cerebrovascular event risk could be even higher as we did not include recurrent TIAs as ischemic events. Most of the initial and recurrent infarctions in our cohort are LACI, which is likely due to occlusion of small penetrator arteries arising from the large arteries of the circle of Willis or from the basilar artery.40,41 Thus, it is usually located deep within the cerebral hemispheric white matter, basal ganglia, the internal capsule, or the thalamus. The limited number of studies on quasi-MMD did not report the subtype of infarctions. A study of definite MMD infarction based on diffusion-weighted imaging reported age-specific appearance of infarct patterns.42 In contrast to our findings, gyral locations were the main pattern (44.0%) and LACI only took 7.7% of the infarction pattern in their cohort. The explanation of the discrepancy could be the different cause of infarction in definite moyamoya and quasi-MMD. Infarction in definite MMD is largely attributed to the hypoperfusion caused by progressive stenosis of major arteries, whereas thromboembolism, which could induce a pattern of multiple small infarctions, is a major cause of the infarctions in quasi-MMD.43 The high event risk makes the need of intervention for quasi-MMD patients imperative. In the ischemic moyamoya patients, studies validated that revascularization surgery could reduce the risk of future ischemic events significantly.15,44 With hemorrhage being the most common cause of morbidity and mortality in moyamoya patients, effective treatment modalities should be devised in hemorrhagic patients. But for these patients, the value of revascularization surgery is controversial.29,32,45-47 Currently, 2 general modalities of revascularization are used: direct and indirect.27 A recent meta-analysis compared these common surgical modalities; the pooled analysis showed that a direct bypass was better at preventing long-term hemorrhage in MMD than was an indirect bypass (OR .26; 95% CI .09-.79; P = .02).48 Further studies are needed to determine the optimal interventions in quasi-MMD patients with prior ischemic stroke or hemorrhage.

Limitations The study is limited by its retrospective design. This is a single-center cohort from a tertiary referral hospital. Thus, selection bias is unavoidable. Some quasiMMD patients who had relatively mild presentation could be missed. Furthermore, these patients could possibly have a natural history without future stroke events. Thus, the event rates might be inflated. The mean follow-up durations may not be long enough to evaluate the full natural history of the disease, even though the follow-up period in the present study was among the longest to date in literature.

ARTICLE IN PRESS CHARACTERISTICS AND NATURAL HISTORY OF QUASI-MMD

Conclusions In conclusion, clinical features of quasi-MMD in mainland China were revealed in the present study. Several differences with previous studies on quasi-MMD in our cohort should be noted, especially the relatively high event rates during follow-up. Prior hemorrhage and ischemic stroke were 2 risk factors for future events. Further study is needed to reduce rebleeding and ischemic stroke risk in quasi-MMD patients.

References 1. Suzuki J, Kodama N. Moyamoya disease—a review. Stroke 1983;14:104-109. 2. Takeuchi K, Shimizu K. Hypoplasia of the bilateral internal carotid arteries. Brain nerve 1957;9:37-43. 3. Research Committee on the Pathology and Treatment of Spontaneous Occlusion of the Circle of Willis, Health Labour Sciences Research Grant for Research on Measures for Intractable Diseases. Guidelines for diagnosis and treatment of moyamoya disease (spontaneous occlusion of the circle of Willis). Neurol Med Chir (Tokyo) 2012;52:245-266. 4. Hayashi K, Horie N, Izumo T, et al. Nationwide survey on quasi-moyamoya disease in Japan. Acta Neurochir 2014;156:935-940. 5. Wei Y-C, Liu C-H, Chang T-Y, et al. Coexisting diseases of moyamoya vasculopathy. J Stroke Cerebrovasc Dis 2014;23:1-7. 6. Gross BA, Du R. The natural history of moyamoya in a North American adult cohort. J Clin Neurosci 2013;20:4448. 7. Acker G, Goerdes S, Schmiedek P, et al. Characterization of clinical and radiological features of quasi-moyamoya disease among European Caucasians including surgical treatment and outcome. Cerebrovasc Dis 2016;42:464-475. 8. Duan L, Bao X-Y, Yang W-Z, et al. Moyamoya disease in China: its clinical features and outcomes. Stroke 2012;43:56-60. 9. Liu X, Zhang D, Shuo W, et al. Long term outcome after conservative and surgical treatment of haemorrhagic moyamoya disease. Journal of Neurology, Neurosurgery & Psychiatry 2013;84:258-265. 10. Hayashi K, Horie N, Suyama K, et al. An epidemiological survey of moyamoya disease, unilateral moyamoya disease and quasi-moyamoya disease in Japan. Clin Neurol Neurosurg 2013;115:930-933. 11. Khan N, Schuknecht B, Boltshauser E, et al. Moyamoya disease and moyamoya syndrome: experience in Europe; choice of revascularisation procedures. Acta Neurochir 2003;145:1061-71, discussion 1071. 12. Yamamoto S, Koh M, Kashiwazaki D, et al. Is quasimoyamoya disease a uniform disease entity? A threedimensional constructive interference in steady state imaging study. J Stroke Cerebrovasc Dis 2016;25:1509-1516. 13. Baba T, Houkin K, Kuroda S. Novel epidemiological features of moyamoya disease. Journal of Neurology, Neurosurgery & Psychiatry 2008;79:900-904. 14. Chiu D, Shedden P, Bratina P, et al. Clinical features of moyamoya disease in the United States. Stroke 1998;29:1347-1351. 15. Hallemeier CL, Rich KM, Grubb RL, et al. Clinical features and outcome in North American adults with moyamoya phenomenon. Stroke 2006;37:1490-1496.

9

16. Kraemer M, Heienbrok W, Berlit P. Moyamoya disease in Europeans. Stroke 2008;39:3193-3200. 17. Kuriyama S, Kusaka Y, Fujimura M, et al. Prevalence and clinicoepidemiological features of moyamoya disease in Japan. Stroke 2008;39:42-47. 18. Goyal MS, Hallemeier CL, Zipfel GJ, et al. Clinical features and outcome in North American adults with idiopathic basal arterial occlusive disease without moyamoya collaterals. Neurosurgery 2010;67:278-285. 19. Kim J-M, Jung K-H, Sohn C-H, et al. High-resolution MR technique can distinguish moyamoya disease from atherosclerotic occlusion. Neurology 2013;80:775776. 20. Kim YJ, Lee DH, Kwon JY, et al. High resolution MRI difference between moyamoya disease and intracranial atherosclerosis. Eur J Neurol 2013;20:1311-1318. 21. Kim SJ, Heo KG, Shin HY, et al. Association of thyroid autoantibodies with moyamoya-type cerebrovascular disease: a prospective study. Stroke 2010;41:173-176. 22. Li H, Zhang Z-S, Dong Z-N, et al. Increased thyroid function and elevated thyroid autoantibodies in pediatric patients with moyamoya disease: a case-control study. Stroke 2011;42:1138-1139. 23. Li D, Yang W, Xian P, et al. Coexistence of moyamoya and Graves’ diseases: the clinical characteristics and treatment effects of 21 Chinese patients. Clin Neurol Neurosurg 2013;115:1647-1652. 24. Kuroda S, Houkin K. Moyamoya disease: current concepts and future perspectives. Lancet Neurol 2008;7:1056-1066. 25. Wan M, Han C, Xian P, et al. Moyamoya disease presenting with subarachnoid hemorrhage: Clinical features and neuroimaging of a case series. Br J Neurosurg 2016;29:804-810. 26. Ohno Y, Kawamura T, Tamakoshi A, et al. Epidemiology of diseases of unknown etiology, specified as “intractable diseases”. J Epidemiol 1996;6:S87-S94. 27. Scott RM, Smith ER. Moyamoya disease and moyamoya syndrome. N Engl J Med 2009;360:1226-1237. 28. Horn P, Pfister S, Bueltmann E, et al. Moyamoya-like vasculopathy (moyamoya syndrome) in children. Childs Nerv Syst 2004;20:382-391. 29. Kobayashi E, Saeki N, Oishi H, et al. Long-term natural history of hemorrhagic moyamoya disease in 42 patients. J Neurosurg 2000;93:976-980. 30. Cho W-S, Chung YS, Kim JE, et al. The natural clinical course of hemodynamically stable adult moyamoya disease. J Neurosurg 2015;122:82-89. 31. Choi JU, Kim DS, Kim EY, et al. Natural history of moyamoya disease: comparison of activity of daily living in surgery and non surgery groups. Clin Neurol Neurosurg 1997;99(Suppl 2):S11-S18. 32. Liu X-J, Zhang D, Wang S, et al. Clinical features and long-term outcomes of moyamoya disease: a single-center experience with 528 cases in China. J Neurosurg 2015;122:392-399. 33. Morioka M, Hamada J-I, Kawano T, et al. Angiographic dilatation and branch extension of the anterior choroidal and posterior communicating arteries are predictors of hemorrhage in adult moyamoya patients. Stroke 2003;34:90-95. 34. Kawaguchi S, Sakaki T, Morimoto T, et al. Characteristics of intracranial aneurysms associated with moyamoya disease. Acta Neurochir 1996;138:1287-1294. 35. Liu P, Lv X-L, Liu A-H, et al. Intracranial aneurysms associated with moyamoya disease in children: clinical features and long-term surgical outcome. World Neurosurg 2016;94:513-520.

ARTICLE IN PRESS M. ZHAO ET AL.

10 36. Da Costa L, Wallace MC, Brugge Ter KG, et al. The natural history and predictive features of hemorrhage from brain arteriovenous malformations. Stroke 2009;40:100-105. 37. Morioka M, Hamada J-I, Todaka T, et al. High-risk age for rebleeding in patients with hemorrhagic moyamoya disease: long-term follow-up study. Neurosurgery 2003;52:1049-1054, discussion 1054–1055. 38. Sun W, Yuan C, Liu W, et al. Asymptomatic cerebral microbleeds in adult patients with moyamoya disease: a prospective cohort study with 2 years of follow-up. Cerebrovasc Dis 2013;35:469-475. 39. Kleinloog R, Regli L, Rinkel GJE, et al. Regional differences in incidence and patient characteristics of moyamoya disease: a systematic review. Journal of Neurology, Neurosurgery & Psychiatry 2012;83:531-536. 40. Rovira A, Griv E, Rovira A, et al. Distribution territories and causative mechanisms of ischemic stroke. Eur Radiol 2005;15:416-426. 41. Bamford J, Sandercock P, Dennis M, et al. Classification and natural history of clinically identifiable subtypes of cerebral infarction. The Lancet 1991;337:1521-1526. 42. Cho HJ, Jung YH, Kim YD, et al. The different infarct patterns between adulthood-onset and childhood-onset

43.

44.

45.

46.

47.

48.

moyamoya disease. Journal of Neurology, Neurosurgery & Psychiatry 2010;82:38-40. Kastrup A, Schulz JB, Mader I, et al. Diffusion-weighted MRI in patients with symptomatic internal carotid artery disease. J Neurol 2002;249:1168-1174. Starke RM, Komotar RJ, Hickman ZL, et al. Clinical features, surgical treatment, and long-term outcome of adult moyamoya patients. J Neurosurg 2009;111:936-942. Fujii K, Ikezaki K, Irikura K, et al. The efficacy of bypass surgery for the patients with hemorrhagic moyamoya disease. Clin Neurol Neurosurg 1997;99(Suppl 2):S194S195. Houkin K, Kamiyama H, Abe H, et al. Surgical therapy for adult moyamoya disease: can surgical revascularization prevent the recurrence of intracerebral hemorrhage? Stroke 1996;27:1342-1346. Huang Z, Ding X, Men W, et al. Clinical features and outcomes in 154 patients with haemorrhagic moyamoya disease: comparison of conservative treatment and surgical revascularization. Neurol Res 2015;37:886-892. Sun H, Wilson C, Ozpinar A, et al. Perioperative complications and long-term outcomes after bypasses in adults with moyamoya disease: a systematic review and meta-analysis. World Neurosurg 2016;92:179-188.