Angiographic Dilatation and Branch Extension of the Anterior Choroidal and Posterior Communicating Arteries Are Predictors of Hemorrhage in Adult Moyamoya Patients
Background and Purpose— The cause of intracranial bleeding in moyamoya disease patients is still unknown. To identify factors that contribute to bleeding, we assessed the angiographic findings of moyamoya disease patients.
Methods— We examined angiograms obtained from 107 moyamoya patients; 70 manifested ischemic and 37 had hemorrhagic lesions. Patients with intracerebral aneurysms or both hemorrhagic and ischemic lesions in the same cerebral hemisphere were not included. Patients were divided into those <20 years of age (n=47) and those ≥20 years of age (n=60). The right and left hemispheres in each patient were individually classified as hemorrhagic, ischemic, or asymptomatic. Each hemisphere was assessed for dilatation and branch extension of the anterior choroidal artery (AChA) and posterior communicating artery (P-CoM) and for the degree of proliferation of basal moyamoya vessels. These data were then statistically analyzed for correlation with intracranial bleeding events.
Results— The degree of proliferation of basal moyamoya vessels was not statistically correlated with hemorrhagic events. On the other hand, there was a correlation between hemorrhage and dilatation and abnormal branching of the AChA. In 27 of 37 hemorrhagic hemispheres (73.0%), this artery was dilated, and its abnormal branches served as collateral supply vessels to other regions. This phenomenon was observed in 4 of 5 hemorrhagic hemispheres from young patients; it was noted in fewer than one third of ischemic and asymptomatic hemispheres from this age group. Similarly, 71.9% of hemorrhagic hemispheres from adult patients manifested AChA dilatation and branching, and the difference between hemorrhagic hemispheres and those that were ischemic or asymptomatic was statistically significant (P<0.01). Although the incidence of dilatation and abnormal branching of the P-CoM was relatively low in hemorrhagic hemispheres from adult patients (18.8%), it was significantly higher than in the ischemic and asymptomatic hemispheres from this age group. Using dilatation and abnormal branching of the AChA and/or P-CoM as assessment criteria, we obtained high specificity (86.4%) and sensitivity (84.4%) for hemorrhagic events in adult moyamoya patients.
Conclusions— In adult moyamoya patients, dilatation and abnormal branching of the AChA and/or P-CoM are strong predictors of hemorrhagic events.
Moyamoya disease is an unusual form of chronic, occlusive cerebrovascular disease characterized by bilateral stenosis or occlusion at the terminal portion of the internal carotid artery (ICA) and an abnormal vascular network (the so-called moyamoya vessels) at the base of the brain.1–4 Many patients with moyamoya disease experience cerebral ischemic or hemorrhagic events; pediatric patients suffer primarily ischemic attacks, whereas hemorrhage is prevalent among adults.1,3 Although revascularization surgery can prevent ischemic attacks,5–8 the cause of intracranial bleeding is unclear, and strategies to prevent its occurrence remain to be developed.
Because the site of intracranial bleeding in many moyamoya patients is the basal ganglia, thalamus, and near lateral ventricle wall,9–11 basal moyamoya vessels have been suspected as its origin.1,12–15 However, in some patients with hemorrhage, few basal moyamoya vessels are found, and their location is away from the bleeding site (Figure 1). We noted that among hemorrhagic moyamoya patients, including some exhibiting this phenomenon, the anterior choroidal artery (AChA) was frequently dilated and exhibited abnormal extension of many branches (Figure 1B, 1D, and 1E) and that the bleeding site was within the AChA territory. Therefore, we examined the arteries that normally supply the central or basal brain region in an effort to identify characteristic features of hemorrhagic moyamoya disease.
Subjects and Methods
A total of 170 patients with moyamoya disease underwent treatment at Kumamoto University Hospital and its affiliated hospitals beginning in 1969. Of these, 100 patients experienced ischemic events, including transient ischemic attacks or cerebral infarction; 47 presented with hemorrhage, including subarachnoid and intraventricular hemorrhage or intracerebral hematoma; and 23 had no or other symptoms. We focused only on patients with hemorrhagic or ischemic symptoms. Patients with intracerebral aneurysms and those with both hemorrhagic and ischemic lesions in the same hemisphere were excluded from this study. Finally, we could obtain 107 intra-arterial angiograms of 107 patients (ischemic, n=70; hemorrhagic, n=37) at the time of onset.
We focused on vessels that supply the basal regions of the brain such as the basal ganglia, thalamus, and near lateral ventricle where intracranial bleeding frequently occurs in moyamoya patients.1,4,9–11 Our focus was on the AChA, posterior communicating artery (P-CoM), and basal moyamoya vessels, except for the ethmoidal and vault moyamoya vessels because the territories they feed are not frequent bleeding sites.
The right and left cerebral hemispheres (n=214) in each patient were separately identified as ischemic (including those with infarction and transient ischemic attacks, n=102), hemorrhagic (n=37), or asymptomatic (n=75). The AChA in each hemisphere was recorded as grade 0 (normal), 2 (dilated with distal branching), or 3 (dilated with abnormal branches serving as collateral supply vessels to other regions; Figure 2A through 2D). Basal moyamoya vessels were graded as absent (grade 0), few in number (grade 1), and densely proliferated (grade 2; Figure 2A, 2E, and 2F). Findings on the P-CoM were recorded as negative; normal, dilated, or positive; or dilated with abnormal branch extensions (Figure 2G). We also noted the site of occlusion or stenosis (O/S) of the ICA as following; site 1, ICA top or A1/M1; site 2, just distal to AChA; site 3, between the AChA and P-CoM; or site 4, proximal to P-CoM (Figure 3).
The 107 patients were divided into a young (<20 years of age; n=47) and an adult (≥20 years of age; n=60) group because childhood moyamoya patients tend to manifest ischemia and adults predominantly experience hemorrhage. We examined the correlation between angiographic findings and symptoms using the Mann-Whitney U and χ2 tests. Differences of P<0.01 were considered statistically significant.
We examined 107 angiograms (214 hemispheres) from 107 patients (ischemia, n=70; hemorrhage, n=37); the characteristics of these patients are shown in Table 1. The mean±SD age of ischemic patients was 21.3±18.0 years (range, 1.6 to 67.8 years); it was 40.0±14.6 years (range, 5.3 to 66.8 years) for patients with hemorrhage. There were 102 ischemic, 37 hemorrhagic, and 75 asymptomatic hemispheres. The 37 hemorrhages were made up of 1 subarachnoid hemorrhage, 9 intraventricular hemorrhages, 14 intracerebral hematomas, and 13 lesions that were of both intracerebral hematoma and intraventricular hemorrhage origin. Among 47 young patients (94 hemispheres), only 5 patients (5 hemispheres) manifested hemorrhage; of 60 adult patients (120 hemispheres), 32 (32 hemispheres) had lesions of hemorrhagic origin.
Angiographic Findings on Moyamoya Vessels and the AChA
Compared with ischemic and asymptomatic hemispheres, those with hemorrhage tended to manifest a higher degree of moyamoya vessel proliferation, regardless of the age of the patient (Table 2). However, the difference was not statistically significant.
Overall, in 33 of 37 hemorrhagic hemispheres (89.2%), the AChA was dilated and manifested branching (grades 1 and 2). More than 70% of the hemorrhagic hemispheres in both age groups had grade 2 AChA; the difference was statistically significant compared with ischemic and asymptomatic hemispheres (P<0.01).
Angiographic Findings on the P-COM
As shown in Table 2, approximately one fifth of the hemorrhagic hemispheres in both age groups showed P-CoM dilatation and branching. The difference between hemorrhagic and nonhemorrhagic hemispheres was statistically significant at P<0.01 (χ2 test) in both age groups.
Site of O/S of the ICA
Figure 3 demonstrates the distribution of the O/S site in the ICA. In 80.6% of adult hemorrhagic hemispheres, the lesion was just distal to the AChA (site 2); this site was involved in 45.9% and 20.8% of ischemic and asymptomatic hemispheres, respectively. The incidence of site 4 and 3 O/S was highest in ischemic hemispheres, and site 4 O/S was not found in hemorrhagic hemispheres.
Predictive Indicator(s) of a Hemorrhagic Event
To determine the most important factor(s) for predicting a hemorrhagic event in adult moyamoya patients, we assessed different parameters for their positive or negative predictive value, specificity, and sensitivity for hemorrhagic events (Table 3). Although the findings of grade 1 or 2 AChA showed the highest sensitivity (90.6%), the specificity was not as high (70.5%). Finally, our findings suggest that the combination of a dilated AChA in which the branches provide collateral flow (grade 2) or dilated P-CoM with abnormal branch extensions represents an indicator with good sensitivity (84.4%) and specificity (86.4%) for predicting a hemorrhagic event.
Our study shows a strong correlation between AChA dilatation and branching and hemorrhagic events in patients with moyamoya disease. Changes in the P-CoM were also correlated. In moyamoya patients without aneurysms, the main bleeding site has been thought to be the site of rupture of moyamoya vessels.1,12–15 However, in some patients with hemorrhage, these vessels are few or are located away from the bleeding site (Figure 1). Bleeding was attributable to an intraventricular hematoma in 35% to 50% of patients reported by others9–11; this was true in 22 (59.5%) of our 37 hemorrhagic moyamoya cases. However, because moyamoya vessels usually do not exist near the ventricle wall, it is difficult to accept that their rupture results in intraventricular hematoma. Suzuki and Kodama3 suggested that there is a tendency for arteries near the ventricle wall to rupture.
Under normal conditions, the AChA supplies areas such as the medial portion of the globus pallidus, anterior perforating substance, internal capsule, choroid plexus, tail of the caudate nucleus, and thalamus. The normal P-CoM gives off branches to the optic chiasm, oculomotor nerve, tuber cinereum, cerebral crura, ventral thalamus, and caudal portion of the caudate nucleus.16,17 Thus, the normal branches of the AChA and P-CoM provide blood to areas near the lateral ventricle, basal ganglia, and thalamus. In moyamoya patients, these normal branches sometimes extend to other regions, and their function as collateral vessels results in blood flow increases. Rupture of dilated branches of the AChA and/or P-CoM may produce intraventricular or thalamic hematomas, hematomas near the lateral ventricle, and hematomas of the basal ganglia that are frequently found in hemorrhagic moyamoya patients. It is considered that the postmortem data may support (or disprove) our hypothesis. Thus, we examined 3 autopsy cases with hemorrhagic moyamoya disease and tried to identify the bleeding vessels. However, we could not identify the branches of AChA and P-CoM. Although it is difficult to pinpoint the bleeding vessels, in our series, hemorrhagic events coincided with dilatation and branch extension of the AChA and/or P-CoM (Table 3), suggesting that these vessels may have been the origin of bleeding.
According to the morphometric analysis of Yamashita et al,15 most dilated arteries show fibrosis and marked attenuation of the media with occasional segmentation of the elastic lamina. As a result of hemodynamic stress or aging, dilated arteries with attenuated walls may acquire a predisposition for focal protrusion (microaneurysm formation) of the arterial wall. Subsequent rupture is thought to be a mechanism resulting in hemorrhage in patients with moyamoya disease. Such microscopic arterial changes may occur in the branches of the AChA and P-CoM.
The normal AChA is ≈0.5 mm in diameter and 3.0 cm in length. On angiograms, the last visible portion of the AChA enters the plexus of the temporal horn, visualized as the angiographic “brush” of the plexus. Furthermore, the fine vessels of the normal P-CoM are scarcely visible angiographically.16,17 Thus, the angiographically evident dilatation and branching of the AChA and P-CoM recorded in our study may signal increased blood flow and hemodynamic stress in these vessels. In ≈30% of nonhemorrhagic hemispheres from our young moyamoya patients, the dilated AChA had branches that served as collateral vessels; in adult ischemic and asymptomatic hemispheres, this phenomenon was more rare (Table 2). This suggests that in young patients the AChA may function as a major collateral route and that it may play the same role for moyamoya vessels.
In nonhemorrhagic hemispheres, blood flow in the AChA decreases or disappears as the O/S site gets closer to the area proximal to the P-CoM (Figure 3). Furthermore, if the stenotic lesion is located just proximal to the P-CoM, neither the dilated AChA nor the basal moyamoya vessels are visible on angiograms. We posit this as the reason that no basal moyamoya vessels were discernible in about one fourth of the ischemic and asymptomatic hemispheres examined (Table 2). We also suspect that O/S just distal to the AChA induces the greatest change in this vessel and results in severe hemodynamic stress. In adults with hemorrhagic moyamoya, the site of O/S may have stopped progressing at a younger age, and the presence of prolonged hemodynamic stress and/or some factors related to aging18 may eventually have led to bleeding.
Intracranial bleeding is a major prognostic factor in moyamoya patients,9,11 and currently we have no reliable indicator(s) for predicting a hemorrhagic event. In our series, angiograms were obtained within a few days of the initial hemorrhagic event, and the findings reflect the status before any rebleeding episodes. Of 32 adult patients who suffered hemorrhage, 17 who did not undergo bypass surgery experienced rebleeding during the follow-up period. Of these, 16 (94.1%) manifested a dilated AChA whose branches served as collateral vessels and/or a dilated P-CoM on angiograms obtained before the rebleeding episode. From these findings and the data shown in Table 3, we suggest that dilatation and branch extension of the AChA and P-CoM are sensitive and specific predictive indicators of rebleeding.
Direct bypass surgery is thought to prevent recurrent intracranial bleeding in moyamoya patients.19–21 A direct bypass may decrease the blood flow in the AChA and P-CoM, and the findings reported here may help to identify moyamoya patients at greatest risk for hemorrhage and thus make it possible to apply preventive surgical therapy before the occurrence of a potentially lethal repeated hemorrhage. However, because this study was retrospective, a large prospective study is needed to confirm our conclusions.
This work was supported by grants from the Research Committee on Spontaneous Occlusion of the Circle of Willis of the Ministry of Health and Welfare of Japan (1999). We wish to thank the members of Kumamoto Stroke Data Bank for the clinical data presentations.
- Received March 6, 2002.
- Revision received August 8, 2002.
- Accepted August 13, 2002.
Masuda J, Ogata J, Yamaguchi T. Moyamoya disease. In: Barnett HJM, Mohr JP, Stein BM, Yatsu FM, eds. Stroke: Pathophysiology, Diagnosis, and Management. New York, NY: Churchill Livingston; 1998: 815–832.
Suzuki J, Kodama N. Moyamoya disease: a review. Stroke. 1983; 14: 104–109.
Karasawa J, Touho H, Ohnishi H, et al. Long-term follow-up study after extracranial-intracranial bypass surgery for anterior circulation ischemia in childhood moyamoya disease. J Neurosurg. 1982; 77: 84–89.
Kinugasa K, Mandai S, Kamata I, et al. Surgical treatment of moyamoya disease: operative technique for encephalo-duro-arterio-myo-synangiosis, its follow-up, clinical results and angiograms. Neurosurg. 1993; 32: 527–531.
Yoshida Y, Yoshimoto T, Shirane R, et al. Clinical course, surgical management, and long-term outcome of moyamoya patients with rebleeding after an episode of intracranial hemorrhage: an extensive follow-up study. Stroke. 1999; 30: 2272–2276.
Iwama T, Hashimoto N, Murai BN, et al. Intracranial rebleeding in moyamoya disease. J Clin Neurosci. 1997; 1997: 4: 169–172.
Oka K, Yamashita M, Sadoshima S, et al. Cerebral haemorrhage in moyamoya disease at autopsy. Virchows Arch (Pathol Anat). 1981; 392: 247–261.
Yamashita M, Oka K, Tanaka K. Histopathology of the brain vascular network in moyamoya disease. Stroke. 1983; 14: 50–58.
Krayenbuhl H, Yasargil MG. Radiologic anatomy and topography of cerebral vessels. In: Huber P, ed. Cerebral Angiography. New York, NY: Georg Thieme Verlag; 1982: 74–80.
Taveras JM, Wood EH. Cerebral Angiography. In: Robbins LL, ed. Diagnostic Neuroradiology. Baltimore, Md: Williams & Wilkins Co; 1976: 581–832.
Mauro AJ, Johnson ES, Chikos PM, et al. Lipophyalinosis and miliary microaneurysms causing cerebral hemorrhage in a patient with moyamoya: a clinicopathological study. Stroke. 1980; 11: 405–412.
Houkin K, Kamiyama H, Abe H, et al. Surgical therapy for adult moyamoya disease: can surgical revascularization prevent the recurrence of intracranial hemorrhage? Stroke. 1996; 27: 1342–1346.
Kawaguchi S, Okuno S, Sakaki T. Effect of direct bypass on the prevention of future stroke in patients with the hemorrhagic variety of moyamoya disease. Stroke. 2000; 93: 397–401.
Okada Y, Shima T, Nishida M, et al. Effectiveness of superficial temporal artery–middle cerebral artery anastomosis in adult moyamoya disease: cerebral hemodynamics and clinical course in ischemic and hemorrhagic varieties. Stroke. 1998; 29: 625–630.