A Clinical Comparison of Definite Moyamoya Disease Between South Korea and Japan
Background and Purpose The goal of the present study was to clarify whether South Korean patients with moyamoya disease have clinical features similar to those of Japanese patients.
Methods From 26 South Korean neurosurgical institutes, 296 definite cases were collected and analyzed statistically. These cases were then compared with 731 Japanese definite cases registered to the Research Committee on Moyamoya Disease of the Ministry of Health and Welfare, Japan.
Results The Korean age distribution patterns showed two peaks that were similar to those seen in Japanese patients. The incidence of adult moyamoya disease in South Korea, however, was 20% higher than that in Japanese patients. The family occurrence rate was 1.8% in Koreans. The incidence of cerebral infarction and bleeding in Koreans was statistically higher, whereas transient ischemic attack and seizure were less than those in Japanese subjects. The incidence of infarction in children and that of hemorrhage in both children and adults were also statistically higher in Koreans. The incidence of hemorrhage was higher in females than in males. Both the age at onset and sex affected the disease type. Although most Japanese patients underwent direct bypass surgery and/or combined indirect bypass procedures, single encephaloduroarteriosynangiosis was performed on 87.6% of all surgical cases in Koreans. Despite the higher incidence of hemorrhagic type in South Korea, the outcomes of the patients were similar to those of the Japanese patients.
Conclusions This study suggests that the clinical background of moyamoya disease in South Korea is essentially similar to that in Japan.
Spontaneous occlusion of the circle of Willis (moyamoya disease) is a cerebrovascular disorder first described in the 1960s in Japan and remains a clinical entity with unknown etiology.1 Since the Research Committee on Spontaneous Occlusion of the Circle of Willis (Moyamoya Disease) of the Ministry of Health and Welfare, Japan (RCMJ) first proposed the angiographic criteria for diagnosis in 1979, RCMJ has monitored the occurrence of this disease, extensively studied the clinicopathological and epidemiological features, and established treatment modalities. Moyamoya disease is characterized by angiographic findings of bilateral steno-occlusive changes in the terminal portion of the internal carotid artery and a coexisting abnormal vascular network (moyamoya vessels) in the base of the brain.1 The onset age distribution pattern has two peaks, with a higher peak at 5 years of age and a lower peak from 30 to 49 years of age. The former is termed the juvenile type and the latter the adult type. The pathophysiology of the juvenile type is ischemia due to steno-occlusive changes of the circle of Willis, while that of the adult type is usually bleeding caused by a breakdown of collaterals formed at a younger age.
According to the diagnostic criteria in 1988,2 definite cases should fulfill the aforementioned angiographic features on both sides of the brain. Unilateral involvement is also classified as probable cases because some of them later develop into bilateral involvement.2 3 4 The Table⇓ shows the diagnostic criteria recently modified for the diagnosis by MRI.5 6 Any disorders that can possibly affect the cerebrovasculature, such as head trauma, radiation, and von Recklinghausen’s disease, should be excluded from the diagnosis. Most of the reported epidemiological studies, however, have included not only definite cases but also probable cases as well as the systemic disorders associated with “angiographic moyamoya” because the number of patients is limited.7 8
Remarkable regional differences exist in the frequency of the reported patients with moyamoya disease in the world.7 This disease is mostly found in Asians or Mongoloids, although more than 6000 cases have been reported all over the world.7 In South Korea, a survey of this disease was first conducted in 1988.8 The clinical features of Koreans, including both probable cases and cases with angiographic findings similar to those for moyamoya disease, were, however, found to be different from those of the Japanese. It is therefore important to strictly apply the criteria for the diagnosis and examine whether there are any racial or regional differences in moyamoya disease. This report illustrates the current clinical features of definite moyamoya disease in South Korea and compares them with those for Japanese patients.
Subjects and Methods
In 1995, a collaborative study was organized with Korean neurosurgeons as a main subject of RCMJ. The diagnostic criteria2 (Table⇑) and registration cards were distributed to 29 institutions with neurosurgical residency programs in South Korea. The registered cases included patients diagnosed as having moyamoya disease by the end of June 1995. A total of 504 cases were thus registered. Because 83 cases lacked sufficient angiographic findings, they were returned for confirmation. Finally, a total of 451 cases with full angiographic findings were collected. Definite cases have moyamoya vessels on both sides. Because moyamoya vessels are hardly seen on the angiographic stages 1, 2, and 6 of Suzuki and Takaku,9 definite cases were selected as follows: (1) in adults an angiogram showing either stage 3, 4, or 5 bilaterally, and (2) in children (<16 years) either stage 3, 4, or 5 on one side and stage 1, 2, 3, 4, 5, or 6 on the other side. Finally, 296 cases were classified as definite. There were 103 probable cases and 52 nonmoyamoya cases that did not meet the angiographic diagnostic criteria. Among the 296 definite cases, 4 cases with systemic disease (2 fibromuscular dysplasia, 1 thyrotoxicosis, 1 cerebral palsy) were excluded from the analysis based on the clinical criteria.
To compare the Korean data with the Japanese data, 731 definite Japanese cases were selected from the registry database and then were analyzed with the same angiographic criteria. Any significance between the two groups was analyzed by a χ2 analysis; a stepwise logistic regression analysis was used for the multivariate analysis.
Age at Onset
The age distribution patterns at onset are shown in Fig 1⇓. Both countries showed two peak patterns. The patients under 10 years of age showed the first highest peak and constituted approximately 36% and 55% of all Korean and Japanese cases, respectively. The childhood onset rate was significantly less (P<.0001) in Koreans (47%) than in the Japanese (67%). Although the second peak was seen in adolescence, Korean females did not show any apparent peak at approximately 30 years of age.
Sex and Familial Occurrence
A significant female predominance was seen in both countries. The Korean female/male ratio was 1.30 and significantly less (P<.0379) than that (1.75) in Japan. The family occurrence rates were 1.8% and 6% in South Korea and Japan, respectively. No significant difference was thus observed between the two countries.
Fig 2⇓ compares the disease types in both countries. In South Korea the incidence of cerebral infarction and bleeding was significantly higher (P<.0001, P<.0001, respectively) and the incidence of transient ischemic attack and seizure was less (P<.0001, P=.025, respectively) than that in Japan. Children showed a significantly higher incidence (P<.005) of transient ischemic attack and seizure than adults in both countries (Fig 3⇓). In South Korea infarction is also higher in children than adults (P=.0024). Approximately 80% of the children had disease of the ischemic type in both countries. In a multivariate analysis of age at onset, sex, and angiographic stage, the age at onset (childhood onet) again showed the highest correlation with the presentation of ischemic type (odds ratio=0.0902 with confidential intervals from 0.0503 to 0.162; P<.0001).
In contrast, the incidence of hemorrhagic onset in South Korea was 42.4%, which was approximately 20% higher than that in Japan (19.1%). Furthermore, 69% of all Korean adult patients presented with the hemorrhagic type, while hemorrhage was seen in 51% of all Japanese adult patients. Female patients constituted approximately 64% and 71% of the hemorrhagic cases in South Korea and Japan, respectively. A significant correlation was seen between females and hemorrhagic type in both countries (P<.05). In a multivariate analysis of the age at onset, sex, and angiographic stage, the age at onset (adult onset) again showed the highest correlation for hemorrhagic type (odds ratio=26.7 with confidential intervals from 12.7 to 56.1; P<.0001), while the second highest correlation was for the patient to be a female (odds ratio=2.34 with confidential intervals from 1.23 to 4.45; P=.008). The factors involved in the disease type were thus age at onset and sex. The angiographic stage showed no correlation with disease type.
The rebleeding rate in Korean patients (8%) was slightly less than that in Japanese patients (11%), while no significant difference was observed between them.
In South Korea the incidence of stage 5 was significantly higher in adults than in children (P=.0141), while no significant difference was observed between males and females. On the other hand, the incidence of stage 3 was less (0.0245) and that of stage 5 was higher (P=.0298) in Japanese females than in males, while there was no significant difference in the angiographic stages between children and adults. The overall incidences were less in stage 3 (P=.0014) and higher in stage 4 (P=.0002) in South Korea than in Japan.
Surgery was performed in approximately 56% (164 cases) and 76% (555 cases) of all patients in South Korea and Japan, respectively. The rest of the patients were conservatively followed up or medically treated with anticoagulants, antifibrinolytics, or anticonvulsants. In Japan, bypass surgery constituted 94% (521 cases) of all surgical procedures, and the percentage was higher than in South Korea (77.4%, 127 cases). In South Korea single indirect bypass surgery—encephaloduroarteriosynangiosis (EDAS) or encephalomyosynangiosis (EMS)—was the main procedure (95.3%). In Japan, however, single indirect bypass surgery constituted only 40.1% of all bypass procedures. Direct bypass surgery (26.7%), combined bypass surgery (direct plus indirect [24.8%]), or a combination of indirect procedures (8.34%) such as EDAS, EMS, and EDAMS (encephaloduroarteriomyosynangiosis) constituted nearly 60%.
For hemorrhagic cases (n=124) in South Korea, he-matoma removal and/or ventricular drainage was performed in 48 cases (38.7%), and 13 of them later received bypass surgery. In Japan 63.6% of 140 hemorrhagic cases received surgical treatment, and bypass surgery was performed in 46.1% of all surgical cases.
The outcomes were determined at the point of final examination on outpatients’ records. With a mean follow-up period of 4.7 years, more than 80% of all cases showed a good outcome. The main factors affecting the outcome were analyzed among age at onset, sex, disease type, and surgery. Bypass surgery showed a significantly higher probability of an excellent outcome (Karnofsky’s Performance Scale: 100% to 90%) than the nonsurgical group (P=.0001). In contrast, patients with ventricular drainage and hematoma removal showed a significantly lower probability for an excellent outcome than the nonsurgical group (P=.0010). Furthermore, 70% of all patients showed either a disappearance or improvement of symptoms. Adult patients showed a lower Karnofsky’s Performance Score, including death. There was no significant difference between males and females regarding outcome. A further relationship between disease type and outcome showed that hemorrhagic type tended to show a poor outcome compared with nonhemorrhagic type. However, 75% of hemorrhagic cases also showed as good an outcome as that seen in nonhemorrhagic cases.
The mean follow-up period in Japan was 6.6 years. Adult patients again showed a significantly lower Karnofsky’s Performance Score, including death, than children (P<.002). In South Korea, the outcome of females was slightly worse than that in Japanese females, probably because of the higher incidence of hemorrhage in Korean females.
A part of the Japanese population seems to have an ancestry similar to that of Koreans, and thus there are a large number of patients with moyamoya disease in both countries. The characteristics of Korean moyamoya disease as reported in 1988,8 however, were somewhat different from those seen in the Japanese because (1) two peaks in the onset age distribution pattern were not observed and (2) both adult and hemorrhagic cases were apparently predominant. The treatment and outcome of the patient were also unclear at that time. In this study we selected only definite cases and excluded patients with both systemic disease that might cause angiographic findings similar to moyamoya disease and unilateral involvement (probable). Among the 504 registered Korean cases, approximately 20% of them were reported as definite even though they were actually probable or nonmoyamoya cases. These findings thus suggest that the diagnostic criteria of moyamoya disease are still poorly understood, not only in South Korea but also in other nations. Even in Japan definite and probable cases are often analyzed together. However, among 180 Japanese cases classified as probable, only 12 cases (6.7%) progressed to the definite type.10 Houkin et al3 also suggested that adult probable cases are distinct from definite moyamoya disease cases based on a clinical and biochemical analysis of basic fibroblast growth factor. It is thus necessary to analyze definite cases separately from probable cases because there might be different genetic and clinical backgrounds between them.
Korean patients with definite moyamoya disease showed a relatively higher incidence of hemorrhagic and adult onset, although the overall clinical background was similar to that of Japanese cases. However, there might be a bias of data collection because the Korean data were collected only from neurosurgical institutes and not from pediatric or internal medicine institutes. Furthermore, Korean patients also showed a higher angiographic stage, such as stages 4 and 5, especially in adults. In addition to genetic and epidemiological differences, a delay in medical attention and the diagnosis of moyamoya disease can also affect these findings. However, this study suggests that the clinical background of Korean moyamoya disease is quite similar to that seen in Japan. The pathophysiology,11 treatment selection, and outcome in Korean hemorrhagic-type moyamoya disease might therefore provide us with further information on how to manage hemorrhagic cases.
Further cooperative studies based on a random strata analysis are necessary to better understand the similarities and differences in this disease among nations. We recently observed that some alleles in human leukocyte antigen class II genes are positively associated with the occurrence of moyamoya disease.12 Worldwide surveys and randomized studies are therefore desirable to clarify the racial significance and genetic background of this disease.
This study was supported by a grant from the Ministry of Health and Welfare, Japan, for the Research Committee on Spontaneous Occlusion of the Circle of Willis (Moyamoya Disease). The authors would like to express heartfelt thanks to the following South Korean neurosurgeons who kindly cooperated with the registration of their cases: Jong Oung Doh (National Medical Center), Youn Kim (Chungnam National University), Sang Chul Kim (Taegu Fatima Hospital), Moon Soo Shin (Jeonju Presbyterian Medical Center), Byung Yeon Choi (Yeungnam University), Jae Gon Moon (Kosin University), Man Bin Yim (Keimyung University Dongsan Medical Center), Sung Don Kang (Wonkwang University), Bark Jang Byun (Soonchunhyang University), Dae Jo Kim (Maryknoll Hospital) Seon Gil Choi (Hangang Sacred Heart Hospital), Byung Gyu Lee (Ewha Womans University), Joong Uhn Choi (Yonsei University), Hyoung-Ihl Kim (Chonbuk National University), Jun Hyuk Song (Korea University), Moon Chul Kim (Taegu Catholic University), Young Jin Lim (Kyung Hee University), Jong Woo Han (Gyeong Sang National University), Shin Jung (Chonnam University Hospital), Jong Sik Suk (Chung-Ang University), Byung Duk Kwun (Ulsan University Asan Medical Center), Soo Chun Kim (Inje University Pusan Paik Hospital), Hyung Dong Kim (Dong A University), Chang Wha Choi (Pusan National University), and Hwa Seong Park (Wallace Memorial Baptist Hospital). We also thank Chiharu Noguchi and Takako Chiga for their secretarial assistance.
Reprint requests to Kiyonobu Ikezaki, MD, PhD, Department of Neurosurgery, Neurological Institute, Kyushu University Faculty of Medicine, 812–82 Fukuoka, Japan.
- Received July 28, 1997.
- Revision received September 5, 1997.
- Accepted September 5, 1997.
- Copyright © 1997 by American Heart Association
Kitamura K, Fukui M, Oka K, Matsushima T, Kurokawa T, Hasuo K. Moyamoya disease. In: Toole JF, eds. Handbook of Clinical Neurology 11. Amsterdam, Netherlands: Elsevier Science Publishing Co; 1989:293–306.
The Research Committee of Spontaneous Occlusion of the Circle of Willis (Moyamoya Disease) of the Ministry of Health and Welfare, Japan: Guidelines for diagnosis. Clin Neurol Neurosurg. 1997;99(suppl 2):S238–S240.
Choi KS. Moyamoya disease in Korea: a cooperative study. In: Suzuki J, ed. Advances in Surgery for Cerebral Stroke. Tokyo, Japan: Springer-Verlag; 1988:107–109.
Ikezaki K, Inamura T, Kawano, T, Fukui M. Clinical features of probable moyamoya disease in Japan. Clin Neurol Neurosurg. 1997;99(suppl 2):S173–S177.
Inoue T, Ikezaki K, Sasazuki T, Matsushima T, Fukui M. Analysis of class II genes of human leukocyte antigen in the patients with moyamoya disease. Clin Neurol Neurosurg. 1997;99(suppl 2):S234–S237.