♦ Epidemiology (Table 15.1)

Although first described in Japan,1 ^, 4 ^, 5 subsequent studies have reported on the epidemiology of moyamoya in the United States, Korea, Taiwan, and Europe.6 ^– 10 Across the Japanese studies, the female-to-male ratio has ranged from 1.8 to 2.2:1 and the incidence from 0.35 to 0.94 per 100,000.4 ^, 5 ^, 11 The reported incidence has increased in recent years, paralleling improvements in imaging technology. Although studies of other ethnicities also show a female predilection,8 ^– 10 the incidence of the disease is notably lower: 0.048 per 100,000 in a Taiwanese study,7 0.052 and 0.086 per 100,000 in American studies,8 ^, 9 and one-tenth the incidence in Japan in a European study.10 Interestingly, the American study of Uchino et al9 demonstrated ethnicity-specific incidence rate ratios of 4.6 for Asian-Americans, 2.2 for African Americans, and 0.5 for Hispanics as compared with Caucasians. The incidence of moyamoya among Asian-Americans—0.28 per 100,000—approaches that reported among Japanese studies. Taken together with a 10% prevalence of familial cases among Japanese studies,5 ^, 11 a genetic component to the idiopathic form of moyamoya is implicated. Supporting this premise, abnormalities on chromosomes 3, 6, and 17 have been identified in association with moyamoya disease.

Epidemiologic studies have demonstrated a bimodal age distribution with one peak involving children under 10 years of age and another involving adults approximately in their fourth decade.4 ^, 6 ^– 11 This dichotomous age distribution parallels the differing clinical presentations between age groups.

♦ Natural History and Clinical Presentation

The majority of pediatric cases of moyamoya disease present with ischemic symptoms secondary to progressive occlusion of vessels of the circle of Willis ( Table 15.2 ). This is closely paralleled in pediatric series of moyamoya syndrome in association with sickle cell disease and Down syndrome with 67 to 100% of patients presenting with transient ischemic attack (TIAs) or strokes.12 Symptoms are commonly precipitated by events that reduce blood flow to the brain in the setting of a tenuous vascular supply such as dehydration or hyperventilation. Reductions in PaCO₂—as in hyperventilation (particularly pertinent in crying children)—induce constriction of already maximally dilated cerebral vessels. Dehydration can reduce blood pressure and promote hypercoagulable states, both of which can adversely affect blood flow through stenotic parent vessels and small collaterals.

In pediatric patients, recurrent TIAs or those that alternate sides can suggest the diagnosis of moyamoya. Up to 25% of previously healthy moyamoya patients become dependent on full-time care or die within 3 years.13 In the study of Olds et al,14 nearly 90% of untreated patients continued to have ischemic symptoms during an average follow-up period of 3.5 years. This relentless course is also seen in patients with moyamoya syndrome and sickle cell disease, with 58% of children having new strokes despite optimal medical therapy.12

Other clinical presentations include seizures (up to 23% of pediatric cases)3 ^, 6 and choreiform movements in up to 11% of cases.3 ^, 15 These movements may result from ischemia in the basal ganglia or mass effect on these areas from dilated collaterals. Dilation of meningeal and leptomeningeal collaterals may explain headache as a presenting symptom in 3 to 6% of cases.3 ^, 15 ^, 16 Rupture of these collaterals resulting in hemorrhage occurs relatively infrequently in the pediatric population (2.5 to 9%)3 ^, 6 ^, 17 and may portend a worse prognosis.

In contrast, hemorrhage is typically one of the most common presentations among the adult moyamoya population, occurring in up to two thirds of cases18 ^– 20 ( Table 15.3 ). Most hemorrhages are intraventricular (IVH) or intraparenchymal (IPH). Of those with IPH, approximately three fourths are in the basal ganglia or thalamus. Mortality rates after hemorrhage have been reported to be as high as 18%. Approximately one third of patients rebleed, with a reported annual rate of 7%.21 After a second hemorrhage, mortality is as high as 67%. Rehemorrhage can occur in a delayed fashion, with reports of bleeding at 20 years following the initial ictus.22 In the study of Kobayashi et al,21 five patients rebled at least 10 years after their first bleed, with a mean time of 6.5 years between hemorrhages. The authors also noted that the second bleed often occurred in a different location, suggesting diffuse vulnerability of moyamoya vessels. Saeki et al23 noted that females were at higher risk factor of rebleeding.

The female sex was also found to be a significant risk factor for radiographic progression. In the general adult moyamoya natural history study of Kuroda et al,24 63 adults were followed prospectively for a mean of 6.1 years, and progression was seen in 17.4% of hemispheres (23.8% of patients) over a mean interval of 5 years. The odds ratio for disease progression in males was 0.2. More impressive are the clinical consequences of untreated moyamoya. An American study reported a 65% risk of recurrent stroke within 5 years of initial symptomatic presentation; this risk increased to 82% among patients with bilateral moyamoya and ischemic symptomatic presentation.25 Even for the limited number of patients without any evident clinical symptoms—that is, only radiographic disease—there remains a 27% risk of symptomatic stroke over 5 years. A recent Japanese study of asymptomatic radiographic moyamoya reported a 3.2% annual risk of ischemic or hemorrhagic events.26 In this population, radiographic studies portended the progressive nature of this disease; 20% of hemispheres showed evidence of completed strokes and 40% exhibited disturbed cerebral hemodynamics.

♦ Diagnosis

Radiographic investigation of children and adults presenting with ischemic or hemorrhagic symptoms often begins with computed tomography (CT). Among those with ischemic symptoms, small hypodense areas are often seen in the basal ganglia, deep white matter, periventricular regions, or watershed zones. Hemorrhagic presentation is often seen in the ventricular system, basal ganglia, thalamus, or medial temporal lobes.

Magnetic resonance imaging (MRI)/magnetic resonance angiography (MRA) typically ensues, at times demonstrating acute infarcts (best seen with diffusion-weighted imaging [DWI]). The “ivy sign”—sulcal linear hyperintensity—seen on T2 fluid-attenuated inversion recovery (FLAIR) is representative of slow flow. Diminished flow voids in the internal carotid artery (ICA), middle cerebral artery (MCA), and anterior cerebral artery (ACA) territories and prominent collateral flow voids in the basal ganglia and thalamus are characteristic of moyamoya.

Angiograms have been the imaging modality of choice to assess the severity of disease. In their classic 1969 study, Suzuki and Takaku2 stratify the angiographic appearance of moyamoya disease into six stages ( Table 15.4 ). Appropriate visualization of basal collaterals, small vessel occlusions, and associated vascular pathology often requires conventional digital subtraction angiography (DSA).

Identification of preexisting spontaneous collaterals from the external carotid circulation is critical for surgical planning, and external carotid artery (ECA) injections should be administered and internal carotid and vertebral artery imaging should be performed. The risk of angiography in moyamoya patients is not significantly different from the risk in patients with other cerebrovascular disease, with overall complication rates of less than 1%.

Other diagnostic evaluations that may be useful in patients with moyamoya include electroencephalography (EEG) and cerebral blood flow studies. Specific alterations of EEG recordings are usually observed only in pediatric patients and include posterior or centrotemporal slowing, a hyperventilationinduced diffuse pattern of monophasic slow waves (called “build-up”), and a characteristic “rebuild-up” phenomenon. Rebuild-up looks identical to the build-up slow waves seen in non-moyamoya patients, but differs in the timing of its presentation. Build-up occurs during hyperventilation, whereas rebuild-up occurs after its completion, indicating a diminished cerebral perfusion reserve.

Cerebral blood flow studies, including xenon-CT, positron emission tomography (PET), and single photon emission computed tomography (SPECT) with acetazolamide challenge, may be employed in the initial diagnostic evaluation and to assess subsequent response to surgery. Xenon-CT has been used to measure cerebral blood flow both prior to and following treatment; its shortcomings include a long acquisition time, limited availability, and susceptibility to motion artifact. PET imaging typically demonstrates reduced regional cerebral blood flow, an elevation of regional oxygen extraction fraction, and an elevation of regional cerebral blood volume. SPECT imaging in moyamoya helps to determine regional cerebral blood flow, particularly in vessels that are not seen angiographically. These studies may help to quantify blood flow, serve as a baseline prior to the institution of treatment, and occasionally aid in treatment decisions.