J Neurol Neurosurg Psychiatry 84:258-265 doi:10.1136/jnnp-2012-302236
  • Cerebrovascular disease
  • Research paper

Long term outcome after conservative and surgical treatment of haemorrhagic moyamoya disease

  1. Jizong Zhao
  1. Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, Beijing, China
  1. Correspondence to Dr J Zhao, Department of Neurosurgery, Beijing Tiantan Hospital, Capital Medical University, 6 Tiantanxili, Chongwen District, Beijing 100050, China; liuxingju006{at}
  1. Contributors XL designed the study, wrote the manuscript, researched the data and contributed to the discussion. DZ designed the study, contributed to the discussion and edited the manuscript. WS contributed to the discussion and edited the manuscript. YZ contributed to the discussion and edited the manuscript. RW contributed to the discussion and edited the manuscript. JZ designed the study, researched the data, reviewed/edited the manuscript and contributed to the discussion.

  • Received 9 January 2012
  • Revised 9 May 2012
  • Accepted 7 August 2012
  • Published Online First 25 September 2012


Objective To investigate the long term outcomes after conservative and surgical treatment for haemorrhagic moyamoya disease.

Methods 97 consecutive patients with haemorrhagic moyamoya disease from 1997 to 2009 were enrolled in this study (mean age 31±10 years; range 5–56 years). We reviewed the clinical charts and radiographs of patients at the first bleeding episode. Follow-up was obtained prospectively by questionnaires and radiographic examinations. Outcomes were compared based on initial treatment (conservative vs surgical).

Results After a median follow-up of 7.1 years, 21 of the 97 (21.7%) patients developed a second episode of bleeding, and six patients (6.2%) died of intracranial rebleeding. The median interval from initial episode to subsequent rebleeding was 9.1 years (0.1–23.2 years). 17 of 43 (37.1%) conservatively treated patients and four of 54 (7.4%) surgically treated patients experienced a rebleeding event (OR 8.1; 95% CI 2.4 to 26.8; p<0.001). There was a difference in the Kaplan–Meier curve of rebleeding between the two groups (Breslow test p=0.047; log rank test p=0.05). The rebleeding ratio in patients who underwent direct bypass was lower than that in patients treated with indirect bypass alone (0% vs 28.5%, 95% CI 1.0 to 1.9; p=0.002). No significant correlation was found between rebleeding and the patient's age, sex, location of haemorrhage, hypertension status or presence of cerebral aneurysm (p>0.05).

Conclusions There is a high risk of rebleeding after the first haemorrhagic episode in Chinese patients with haemorrhagic moyamoya disease. Revascularisation surgery can improve regional blood flow and have greater efficacy at preventing rebleeding than conservative treatment.


Moyamoya disease is a chronic cerebrovascular disease characterised by bilateral steno-occlusive changes at the terminal intracranial internal carotid arteries and their major branches. There are two main phenotypes of moyamoya disease in Asian populations: ischaemic type moyamoya disease, which is common in children; and haemorrhagic type moyamoya disease, which is mostly seen inadults.1 ,2 At present, the exact aetiology and natural history of the disease remain unknown. Recent studies have suggested that surgical revascularisation is effective in the treatment of ischaemic moyamoya disease, especially in the paediatric population.3–5 However, for haemorrhagic moyamoya disease there has been a lot of controversy over ideal treatment protocols. There is little information on comparing long term outcomes of surgical and conservative therapy of haemorrhagic moyamoya disease. Kawaguchi et al and Fujii et al have tried to investigate the effect of surgical treatment on recurrent haemorrhages and their studies have suggested a benefit of surgical revascularisation to prevent recurring haemorrhages.6 ,7 However, due to the limitation of small sample size and short observation period, methods to prevent intracranial rebleeding in patients with moyamoya disease have yet not been devised.8–10

Therefore, preventing intracranial rebleeding in patients with haemorrhagic moyamoya disease has been an evolving topic of study. In this study, we aim to evaluate the 10 year effectiveness and medium to long term outcomes of surgery versus conservative management for patients with haemorrhagic moyamoya disease.

Clinical material and methods

Patient selection

This study was approved by the Beijing Tiantan Hospital Research Ethics Committee. From 1997 to 2009, a total of 246 patients with moyamoya disease were identified. A retrospective review of clinical charts and image data was performed. Ninety-seven (39.8%) patients were identified with intracranial haemorrhage as the initial symptom. Inclusion criteria included: (1) cerebral digital subtraction contrast angiography (DSA) that revealed severe stenosis or occlusion of the distal internal carotid or proximal middle and anterior cerebral arteries with prominent lenticulostriate ‘moyamoya collaterals’; and (2) patients had experienced at least one intracranial haemorrhage which was verified by CT scanning, MRI or lumbar puncture. Patients whose initial onset was marked by ischaemia but subsequently suffered from intracranial haemorrhage were excluded from this review.


Fifty-four (55.6%) of the 97 haemorrhagic moyamoya patients underwent surgical revascularisation procedures based on preoperative cerebral DSA and cerebral perfusion studies: 29 were treated with superficial temporal artery to middle cerebral artery (STA–MCA) anastomosis, 11 with STA–MCA anastomosis combined with encephalo-duro-arterio- synangiosis (EDAS), four with EDAS, eight with multiple burr holes and two with carotid artery adventitial denudations (table 1). Microscope integrated intraoperative indocyanine green angiography was used to demonstrate bypass patency. Patients who underwent surgical treatment but not surgical revascularisation were divided within the conservative treatment group, including two cases of ventricular drainages, six haematoma evacuations and three ventriculo-peritoneal shunting. There were no significant differences in general conditions at admission between the conservative and surgical groups (table 2).

Table 1

Patient characteristics

Table 2

Comparison between the conservative and surgical treatment at admission

Clinic follow-up

CT perfusion or single photon emission CT was used to test the variability of cerebral blood flow (CBF) of patients in the surgical revascularisation group at 3 and 6 months postoperatively. DSA was obtained at 3 months after STA–MCA anastomosis. Follow-up was calculated from the time of the first cerebral haemorrhage. The primary follow-up events included death and recurrent cerebral haemorrhage. Evaluation of outcome status was according to the modified Rankin Scale (mRS) and Barthel Index of Activities of Daily Living score (ADL): working full time (ADL=100, mRS=0 or 1), living independently (60≤ ADL <100, mRS=1 or 2), living with help (20≤ ADL <60, mRS=3), disability (ADL <20, mRS ≥4). mRS scores ≤2 were defined as a good outcome, and scores of 3 or more were defined as a poor outcome.

Statistical analysis

Statistical analysis was performed using SPSS V.16.0. Categorical variables were analysed using a Pearson χ2 test or Fisher exact test. Continuous variables were compared using independent Student t tests. Risk of subsequent rebleeding was determined using the Kaplan–Meier technique, with comparisons made using the log rank or Breslow testing. Risk factors for rebleeding were analysed by multivariable Cox regression analysis. The following risk factors were entered as covariates: age, gender, hypertension, smoking and alcohol consumption, complication with cerebral aneurysm, family history and type of haemorrhage. Statistical significance was set at p<0.05 for 95% CI.


Clinic characteristics

Of the 246 patients, there were 112 men and 134 women. Average age was 25±13 years. The ratio of men to women in the 97 haemorrhagic patients (1.0:2.1) was higher than in the whole sample (χ2 analysis; OR 1.9; 95% CI 1.1 to 3.1; p=0.009). Of these 97 patients, there were 33 men and 64 women aged 5–56 years (mean 30±11 years). Six patients (6.2%) were children (age <12 years). History of smoking, alcohol use, hypertension, cerebral aneurysm and familial moyamoya disease were noted in 27 patients. The most common initial symptom was sudden headache, loss of consciousness or hemiparesis. Intraventricular haemorrhage (IVH) was the most common type of bleeding on CT scans (table 1). According to Suzuki's11 classification of moyamoya disease demonstrated on angiography, stages III and IV were the most common types, accounting for 59% of patients.

Episode of rebleeding

Twenty-one of 97 patients (21.7%) suffered cerebral rebleeding during a median follow-up period of 7.1 years (range 2.0–25.0); of these, six patients had two rebleeding episodes and one patient had three rebleeding episodes (there were six men and 15 women with an average age of 27±10 years (5–42)). The shortest interval from initial episode to subsequent rebleeding was 1 month and the longest was 23 years, with a median interval of 9.1 years. The highest number of patients developed rebleeding after the first haemorrhage after 6–10 years (eight cases; 42.8%), followed by >15 years (five cases; 23.8%), 2–5 years (three cases; 14.3%) and within 1 year (three cases; 14.3%). Recurrent rebleeding between 11 and 15 years was not frequent, occurring in only one case (figure 1). Six of the 21 patients who suffered rebleeding died during hospitalisation (mortality rate 28.6%). Four patients presented with neurological deficits. IVH and intracerebral haemorrhage (ICH) were the two most common types of rebleeding, accounting for 13 cases (61.9%), followed by ICH (28.6%). Details of rebleeding are shown in table 3.

Figure 1

Time interval between the initial episode and rebleeding.

Table 3

Summary of 21 patients with recurrent intracranial haemorrhage

Although revascularisation procedures were used, four of 54 patients (7.4%) suffered cerebral rebleeding (at 6 months, 3 years, 8 years and 10 years). Compared with the operated hemisphere, one patient experienced rebleeding in the opposite hemisphere (table 3, surgical revascularisation group, patient No 2), two patients experienced rebleeding in the same hemisphere (patient Nos 1 and 3) and one patient experienced two SAHs (patient No 4). Of these, two patients died and two patients had moderate to severe disability (however, two patients who developed extensive cerebral infarction after STC–MCA were not included in the re-haemorrhage group in terms of morbidity and mortality because there was not enough time for any re-haemorrhaging event). Seventeen of 43 (37.1%) patients who were treated conservatively experienced rebleeding, while four (9.3%) patients died of rebleeding. The rate of rebleeding in patients treated with revascularisation surgery was lower than for those without surgery (OR 8.1; 95% CI 2.4 to 26.8; p<0.001) (table 4) while no significant differences were found in mortality between the two groups (OR 1.7; 95% CI 0.3 to 8.4; p=0.73). Patients who underwent STA–MCA or STA–MCA combined with EDAS had lower rates of rebleeding than those treated with indirect revascularisation alone (OR 1.4; 95% CI 1.0 to 1.9; p=0.002).

Table 4

Comparison of outcomes between the conservative and surgical treatment

Kaplan–Meier analysis for risk of rebleeding is shown in figure 2, demonstrating that the probability of rebleeding is higher with conservative treatment. The Breslow test revealed that there was a significant difference between the two rebleeding curves (χ2=3.93, p=0.047). However, this difference was barely significant in the log rank test (χ2=3.85, p=0.05). There was no significant difference in the rate of rebleeding between patients with a history of risk factors and those without a history. No significant correlations were observed between rebleeding and patient age, gender, hypertension, aneurysm or type of haemorrhage.

Figure 2

Kaplan–Meier curve for rebleeding free survival in patients treated by the two methods.

Follow-up image findings

CBF findings

Seventy-two patients received at least one CT perfusion or single photon emission CT 6 months after discharge, including 46 surgically treated patients and 26 conservatively treated patients. In the revascularisation group, 36 (78.3%) patients showed an increase in cerebral perfusion (figure 3, figure 4), six showed no change and four showed a decrease. In the conservative group, 17 (65.3%) patients had no change in perfusion, six had decreased perfusion and three patients had slightly increased cerebral perfusion. The improvement in cerebral perfusion in the revascularisation group was significantly better than in the conservative group (OR 17.5; 95% CI 4.5 to 68.0, p<0.001).

Figure 3

(A) Digital subtraction angiography (DSA) of a 40-year-old man with moyamoya disease. The left side shows stenosis of the middle cerebral (MCA) artery and development of moyamoya vessels. Preoperative CT perfusion (CTP) shows hypoperfusion in the bilateral frontal area, and increased mean transit time (MTT) and time to peak (TTP). (B) DSA 1 month after left superficial temporal artery (STA)–MCA anastomosis shows that the MCA was supplied by the STA, and CTP shows increased cerebral blood flow; MTT and TTP were normal. (C) DSA 2.5 years after operation shows reduction of moyamoya vessels and normal CTP.

Figure 4

A 30-year-old woman. (A) Digital subtraction angiography (DSA) of the left side of the internal carotid artery and external carotid artery, and moyamoya vessels are faintly shown. CT perfusion (CTP) shows a significant decrease in cerebral blood flow and increases in mean transit time and time to peak. (B) DSA 3.2 years after left superficial temporal artery–middle cerebral artery (STA–MCA) bypass shows reduction of moyamoya vessels and numerous collateral arteries from the STA covering the left MCA area. This patient showed normal development on CTP.

DSA and MR angiography findings

Due to the limitations of the medical environment, DSA studies were only performed in a few patients postoperatively, but MR angiography or CT angiography was performed in many patients several years after discharge. In 14 patients, DSA studies were performed 1–6 years after STA–MCA bypass. Of 17 sides in these cases, reduction in moyamoya vessels was observed in 13 sides (figure 3, figure 4); well developed revascularisation was observed in two sides treated by EDAS.

Functional outcomes

A total of seven deaths occurred: six patients died of rebleeding and one patient died of another cause. In the 54 patients treated with revascularisation surgery, there were no intracranial haemorrhagic complications after surgery. However, two patients developed extensive cerebral infarction after STC–MCA; one died 10 days after decompressive craniotomy and one survived with hemiparalysis. None of the other patients experienced postoperative events except for one patient with postoperative seizure and three cases of subdural effusion. During the follow-up period, according to ADL, 32 of 54 (59.2%) patients were employed in a full time occupation, 16 (29.6%) were living at home self-sufficiently, two (3.7%) had disability and three (5.5%) had died. In the conservative group, 13 (30.2%) were employed in a full time occupation, 20 (46.0%) were living at home self-sufficiently (six (13.9%) reported intermittent headache), six (13.9%) had disability and four (9.3%) patients had died of rebleeding (table 4).


In recent decades, more and more patients with moyamoya disease have been found in the Chinese population.12 ,13 In our cohort, patients with haemorrhagic type moyamoya disease were less frequent than those with ischaemic type, accounting for 39%, higher than reports from North America and Europe.14 ,15 Clinical characteristics of haemorrhagic moyamoya patients were similar to those in Japan. It primarily affected adult females, with a female to male ratio of 2.1:1.0. The proportion of females with haemorrhagic disease was much higher than those with ischaemic disease. In addition, we found that most haemorrhagic patients had no ischaemic symptoms prior to the initial bleeding episodes, although chronic cerebral hypoperfusion had already existed, according to CBF imaging. Most patients will visit a doctor only if they have sudden onset of severe headache, vomiting, coma or hemiparalysis, resulting from cerebral haemorrhage. Therefore, finding effective methods of early diagnosis in these patients is essential, but to our knowledge there are currently no reports on this.

The exact cause of bleeding or rebleeding in moyamoya disease remains unknown. Compared with normal intracranial blood vessels, histopathological examination showed that the walls of the collateral vessels had worsened elasticity, became thinner and more brittle, and were more likely to develop intracranial microaneurysms.16 ,17 Rupture of the collateral vessels at the base of the skull was considered a primary reason for this. Ikezaki et al 18 reported that patients with hypertension had a higher risk of rebleeding (27.8%) and mortality (60.0%). The presence of asymptomatic microbleeds on MRI was also mentioned as a predictor of rebleeding in previous reports,9 while other reported risks of rebleeding include age and aneurysm complications.19 ,20 In our study, IVH or ICH with IVH was the most common type, either at the first bleeding or on rebleeding. This was usually the site of the collateral vessels. Although six (6.1%) patients had the complication of cerebral aneurysm, rebleeding occurred in only one. No significant correlation was found between bleeding or rebleeding and age, sex, hypertension, aneurysm or type of haemorrhage.

Haemorrhagic moyamoya disease was less frequent than the ischaemic type but it caused higher morbidity and mortality. Rebleeding was the main cause of death. Kobayashi et al 6 investigated the long term natural history of 42 haemorrhagic moyamoya patients and showed that 33.3% patients experienced a second episode of bleeding during a mean follow-up period of 80.6 months. More importantly, after rebleeding, the mortality rate rose from 6.8% to 28.6%, and the rate of good recovery decreased from 45.5% to 21.4%. Morioka et al 17 found that up to 61.1% of patients experienced rebleeding during a mean follow-up period of 12.7±7.1 years. However, this was not the case in North American patients presenting with haemorrhage; they had a relatively low risk of re-haemorrhage.14 As shown in this study, 17 (37.1%) conservatively treated patients had recurrent bleeding, over half experienced rebleeding 10 years after the initial onset (the longest was 23 years) and four (23.5%) patients died of rebleeding (at 8 years, 9 years, 10 years and 15 years) (table 3). Consequently, prevention of cerebral rebleeding, especially long term prevention, is the key for treatment for haemorrhagic moyamoya disease in Asian patients.

Surgical revascularisation can improve cerebral perfusion and reduce the number of moyamoya vessels, which is considered to reduce the occurrence of rebleeding. However, previous studies have not confirmed this.10 ,18 ,21 ,22 There have been a few studies comparing surgical and medical treatment in haemorrhagic moyamoya disease.6–8 ,16 Yoshida et al 10 conducted an extensive study of 21 patients with haemorrhagic moyamoya disease. The results revealed that rebleeding occurred in one of eight (12.5%) patients who had bypass surgery and in five of 13 (38.5%) patients who did not, over a mean period of 14.2 years. However, there were no statistically significant differences in rebleeding rates between the two groups because of the small sample size. In this study, four patients (7.4%) who underwent revascularisation surgery experienced rebleeding; the rebleeding rate in revascularisation surgery was significantly lower than that in the conservative group (37.1%). Forty-eight of 54 (89%) patients in the surgery group had a full time occupation or were living self-sufficiently, which was higher than the 76% in the conservative group. Mortality from rebleeding in the surgery group (50%) tended to be higher than that in the non-surgery group (23.5%), but this was not statistically significant. In addition, statistical analysis using the Kaplan–Meier rebleeding curve (figure 2), showed that revascularisation surgery had a stronger effect on preventing short term rebleeding.

There is still no standard with respect to the ideal revascularisation surgery for haemorrhagic moyamoya disease. Kawaguchi et al 6 reported that compared with EDAS, STA–MCA bypass had greater efficacy in the prevention of rebleeding in patients with haemorrhagic moyamoya disease. However, according to another study, no significant difference was observed in the rate of rebleeding among various types of surgical revascularisation.21 In our study, there were no cases of rebleeding in 40 patients who underwent STA–MCA alone or combined STA–MCA and EDAS. Among the remaining 14 cases treated with indirect revascularisation surgery, rebleeding occurred in four patients. At this time, we cannot conclude that STA–MCA was more efficacious than other methods of revascularisation because of the limited number of indirect bypass procedures.

Although great advances have been made in our understanding of surgical treatments for moyamoya disease in recent years, a significant portion of patients have developed serious postoperative complications, including postoperative ischaemic injury, seizures, haemorrhage and subdural effusions.4 ,23–26 Mesiwala et al 4 reported that patients who underwent direct revascularisation were more likely to develop postoperative complications. In our study, intraoperative indocyanine green was used to evaluate bypass patency for all patients treated with STA–MCA. However, there were still two patients who experienced complicated postoperative cerebral infarction, and one of them died. We were surprised by the abnormally high rate of serious cerebral infarction in the STA–MCA group. Possible causes may include longer time the MCA was clamped, intraoperative or perioperative blood pressure fluctuations, unexpected haemodynamic changes in the cerebral artery after revascularisation and anaesthesia. Data have shown that surgical revascularisation accelerates progressive stenotic changes of the ICA in moyamoya disease.27 Therefore, intraoperative and postoperative haemodynamic monitoring was necessary for patients treated with revascularisation.

Several limitations of the study must be noted. First, these results cannot be generalised to the impact of revascularisation on subsequent haemorrhage in the majority of moyamoya cases who did not present with haemorrhage as their first symptom prior to revascularisation. Second, scientific comparisons between various revascularisation interventions cannot be made due to the retrospective and non-randomised nature of this study. Selection biases might have assigned patients with different features and known or unknown risks to the treatment versus the conservative groups. Third, long term follow-up angiographic data were not obtained in most of these patients, although 1 year follow-up imaging data were available. These biases may have an impact on the validity of conclusions. Therefore, we are unable to determine whether one revascularisation technique is superior to another. Large randomised controlled trials comparing various revascularisation interventions are necessary.

Despite these inherent limitations, our findings offer useful information about the treatment of haemorrhagic moyamoya disease and its natural history. The study reveals that there is a high risk of rebleeding in haemorrhagic moyamoya disease in China. Patients are primarily adult women, and rebleeding tends to occur at any time after the original bleeding. Revascularisation interventions showed greater efficacy than conservative treatment in reducing the risk of rebleeding. This study also found that the rate of rebleeding in STA–MCA bypass seemed to be lower than that in indirect bypass.


  • Funding This study was supported by ‘11th Five-Year Plan’ National Science and Technology supporting plan (2006BAI01A13). Beijing Municipal S&T Commission (D101107049310001).

  • Competing interests None.

  • Ethics approval This study was approved by the Beijing Tiantan Hospital Research Ethics Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.


Visit the full archive of podcasts for JNNP here >>

Free sample
This recent issue is free to all users to allow everyone the opportunity to see the full scope and typical content of JNNP.
View free sample issue >>

Don't forget to sign up for content alerts so you keep up to date with all the articles as they are published.