点击显示 收起
the Department of Neurosurgery, University of Illinois at Chicago.
Abstract
Background and Purpose— Symptomatic vertebrobasilar disease (VBD) carries a high risk of recurrent stroke. We sought to determine whether a management algorithm consisting of quantitative hemodynamic assessment could stratify stroke risk and guide the need for intervention.
Methods— All patients with symptomatic VBD at our institution are evaluated by a standard protocol including quantitative magnetic resonance angiography (QMRA). Patients are stratified on the basis of the presence or absence of distal flow compromise. Those with low distal flow are offered intervention (surgical or endovascular); all patients receive standard medical therapy. We reviewed the clinical outcome of patients managed with this protocol from 1998 to 2003.
Results— Follow-up was available for 47 of 50 patients over a mean interval of 28 months. Stroke and combined stroke/transient ischemic attack free survival at 2 years was calculated using the Kaplan–Meier curve. Patients with normal distal flow (n=31) had an event-free survival of 100% and 96%, respectively. Comparatively, patients with low distal flow (n=16) experienced a 71% and 53% event-free survival, demonstrating a significantly higher risk of recurrent ischemia (P=0.003). Patients with low flow who subsequently underwent treatment (n=12) had an 82% event-free survival. Cox proportional hazards analysis demonstrated that flow status affected event-free survival regardless of covariates.
Conclusions— Patients with symptomatic VBD demonstrating low distal flow on QMRA appear to have a high risk of stroke; conversely, those with normal flow seem to have a benign course and may be optimally managed with medical therapy alone.
Key Words: blood flow stenosis stroke transient ischemic attack vertebrobasilar insufficiency
Introduction
Atherosclerotic disease of the vertebrobasilar system is an important etiology of posterior circulation stroke.1 Compared with the anterior circulation, large or small vessel occlusive disease is a more prevalent source of stroke than thromboembolism.2 Symptomatic vertebrobasilar disease (VBD), particularly if it affects intracranial vessels, carries a high stroke risk, averaging 10% to 15% per year despite medical therapy.3,4 Advances in technology, especially in the arena of endovascular therapy, have created new options for treatment of VBD. However, the selection criteria for appropriate candidates for either surgical or endovascular revascularization remain uncertain.5
We sought to determine whether a management algorithm for symptomatic VBD consisting of quantitative hemodynamic assessment could identify patients at high risk for stroke and guide the need for intervention. Existing modalities for evaluating hemodynamic insufficiency have been unsatisfactory for detection of posterior circulation compromise.6 As an alternative modality, phase-contrast quantitative magnetic resonance angiography (QMRA) is a noninvasive tool for measuring blood flow in intracranial vessels. At our institution, we incorporated QMRA analysis into our diagnostic work-up of all patients presenting with vertebrobasilar insufficiency (VBI). Patients with symptomatic VBD are considered for intervention if evidence of flow compromise is present and treated medically otherwise. We reviewed the subsequent risk of stroke in the presence or absence of distal flow compromise because this may define a population of patients most likely to benefit from revascularization.
Methods
Patient Selection and Evaluation
Clinical data for all patients referred with a diagnosis of VBI between 1998 and 2003 were collected and reviewed. At the time of referral, each patient underwent a thorough neurological evaluation and imaging with computed tomography or MRI to evaluate for mass lesions or infarct. If there was suspicion for a vestibular etiology for symptoms, vestibular and audiometric testing was performed. If warranted, medical and cardiac work-up including echocardiogram, 24-hour electrocardiographic (Holter) monitoring, and laboratory tests for hypercoagulability and platelet function were performed to assess for other potential causes of stroke, such as cardioembolism. Magnetic resonance angiography (MRA) or 4-vessel cerebral angiography was performed to evaluate the anatomy of the cerebral vasculature. The presenting and follow-up clinical data of each patient were reviewed at the time of evaluation by the multidisciplinary neurovascular group of our institution, which convened weekly and consists of representatives from neurology, neuroradiology, interventional neuroradiology, neuropsychiatry, and neurosurgery. Clinical and imaging interpretations were made prospectively by the multidisciplinary group, but the data were collected and analyzed retrospectively. The study was approved by the institutional review board, and data collection and recording was according to institutional guidelines.
Patients were considered to have VBI if they demonstrated characteristic symptoms of dizziness or vertigo, transitory bilateral motor or sensory deficits, ataxia or clumsiness, dysarthria, dysphagia, or transitory diplopia or other visual symptoms. Isolated dizziness was not felt to be adequate to diagnose VBI. Angiography was recommended in all patients with MRA findings consistent with >50% stenosis. The majority of patients (84%) underwent angiography. If angiography could not be performed for medical reasons such as renal failure or was declined by the patient, a high-resolution 3-T MRA was used to make the determination of stenosis. For the purposes of this study, if angiography or MRA reports did not specify the percentage of stenosis, descriptive terms such as "severe" or "high-grade" stenosis were used as acceptable evidence of >50% stenosis. After evaluation, those patients with >50% vertebrobasilar stenosis or occlusion on angiography or MRA and symptoms of VBI were characterized as having symptomatic VBD.
Transient ischemic attack (TIA) was defined as a transient focal neurological deficit attributed to ischemia lasting <24 hours. A stroke was defined as a focal neurological deficit lasting >24 hours attributable to cerebral infarction. The occurrence of these events was determined by the multidisciplinary neurovascular group. Follow-up evaluation was routinely recommended at 6-month intervals but occurred more or less frequently dependent on the individual case. Follow-up was performed primarily through clinic visits but also by phone contact if the patient lived remotely or failed to appear for a clinic appointment. A review of the patient’s condition and clinical data were routinely performed by our multidisciplinary group at the time of follow-up.
Blood Flow Measurements
Patients with symptomatic VBD all underwent quantitative flow measurements of the extracranial and intracranial arteries using QMRA. The technique of blood flow quantification by QMRA has been described previously.7 The technique is now implemented and enhanced in commercially available software called the NOVA (Noninvasive Optimal Vessel Analysis) system (VasSol, Inc.).
Management Algorithm
The flow measurements of the posterior circulation were interpreted on the basis of the algorithm shown in Figure 1 to represent low or normal flow distal to the site of VBD. Flows in specific vessels were felt to be reduced if they were >20% lower than the lower limit of baseline normative ranges available for those vessels (supplemental Table I, available online at http://www.strokeaha.org). This threshold was based on existing literature on cerebral blood flow and perfusion assessment that indicates that blood flow reduction in the range of 20% to 25% in distal flow is correlated with ischemia.9–11 The vessels used to designate flow status were those that reflect the distal territory of the vertebrobasilar tree, namely the basilar artery (BA) itself, and the posterior cerebral arteries (PCAs). If the PCA was found to be anatomically fetal, it no longer represented a distal outflow of the vertebrobasilar system and was therefore not considered in designation of flow status. On the basis of our criteria for flow reduction, the BA was designated as reduced if flow was measured to be <120 cc/min, and the PCAs if flows were <40 cc/min. For a designation of "low flow," the BA and PCAs had to be reduced. In borderline cases, for example, with 1 PCA below threshold and the other maintained, additional criteria were used, as outlined in Figure 1.
Patients were recommended medical therapy alone or intervention (surgery or endovascular therapy) coupled with medical therapy based on the designation of their flow status. Patients with normal distal flow as outlined by the algorithm were felt to be experiencing primarily either embolic phenomenon or, more commonly, small vessel disease. All these patients were managed medically with anticoagulants or antiplatelet agents. Patients with reduced distal flow were felt to be experiencing a regional flow compromise, which would be amenable to flow augmentation by angioplasty/stenting or surgical bypass procedures. These patients were offered intervention in addition to medical therapy. Patients were followed clinically, and episodes of TIA or new/recurrent stroke were noted.
Statistical Analysis
Ischemic events in the territory of the posterior circulation were used as the primary end point. Stroke and combined stroke/TIA were evaluated as separate events. Survival was defined as the interval between presentation and stroke or stroke and TIA. Stroke and stroke/TIA-free survival was estimated with the Kaplan–Meier curve at 24 months. The event-free survival is provided with 95% CI as an indicator of the precision of the estimate. Clinical variables between groups were compared using unpaired 2-tailed Student’s t test or the chi square test where appropriate. Univariate analysis to compare survival curves between groups was performed with the log-rank test. The effect of age, gender, lesion location (intracranial versus extracranial; basilar versus other), stroke as presenting symptom, degree of stenosis (occlusion versus >80% versus 50% to 80%), and type of medical therapy after diagnosis (anticoagulant versus antiplatelet versus both) was examined. Cox proportional hazards analysis was used to assess the effect of covariates showing significance of P<0.10 on univariate analysis. All analyses were performed with STATA (Intercooled version 6; Stata Corporation).
Results
Patient Demographics
the time period January 1998 and December 2003, 101 patients were referred for evaluation for symptoms suggestive of VBI. The most prominent symptom prompting referral was dizziness. Patients presenting with acute dissections and presumed embolic strokes were excluded from analysis, but patients presenting with chronic dissection and persistent vertebrobasilar stenosis were included. After the standard work-up, 61 patients were considered to have symptomatic VBD as the source of their symptoms; 50 of these patients had adequate QMRA studies and were managed according to our flow-based management algorithm, 3 of whom were lost to follow-up, leaving 47 patients available for analysis. Their demographic features and clinical characteristics are outlined in the Table.
Demographic and Clinical Data for Patients With Symptomatic VBD
Management Algorithm
Of the 50 patients, 33 (66%) showed normal distal flow and were managed medically. Seventeen patients demonstrated low flow and were offered intervention, 12 subsequently underwent revascularization, and 5 deferred treatment. Follow-up was available in 47 of the 50 patients, as outlined below, over an average period of 27.9 months (range 4 to 65 months; median 23 months).
Normal Distal Flow (n=31)
Thirty-one patients in the normal distal flow category were managed with medical therapy, with either anticoagulants, antiplatelet agents, or both. Treatment regimens often changed during the period of follow-up, with a tendency to switch to antiplatelet agents alone over time. During the follow-up period, none of these patients experienced new or recurrent strokes. Two patients experienced TIAs 6 and 38 months after initial presentation. Over a total of 80.2 person years of follow-up, this resulted in a 0% stroke and 2.5% stroke/TIA rate per person year.
Low Distal Flow (n=16)
Sixteen patients with significant flow reduction in the vertebrobasilar system were candidates for treatment. Twelve patients underwent intervention and were censored from analysis of stroke rates at the time of intervention; their subsequent course is described in the section below. In 4 cases, intervention was refused by the patient, or other medical comorbidities made intervention high risk. All patients were receiving anticoagulants or antiplatelets. During the 10.6 person years of follow-up, 2 patients experienced strokes, and an additional patient experienced recurrent TIA, resulting in an 18.9% stroke and 28.3% stroke/TIA rate per person year.
Low Distal Flow With Intervention (n=12)
In the 12 patients who underwent intervention, endovascular therapy was offered if technically feasible, and angioplasty/stenting was performed in 3 cases. The remaining 9 cases were treated with a variety of surgical revascularization procedures, including extracranial–intracranial bypass and carotid–vertebral transposition. Two patients experienced strokes after therapy: 1 perioperatively after a failed superficial temporal artery-PCA bypass graft, and another 6 months after angioplasty. Over 18.4 person years of follow-up, this resulted in a 10.9% yearly ischemic event rate.
Risk of Stroke and TIA
Stroke and combined stroke/TIA-free survival was examined in the 3 groups of patients using the Kaplan–Meier method (Figure 2). Patients with normal distal flow (n=31) had a stroke and stroke/TIA-free survival of 100% and 96% (95% CI, 77% to 99%), respectively, at 24 months. Comparatively, patients with low distal flow (n=16) had a 71% (CI, 23% to 92%) and 53% (CI, 12% to 83%) event-free survival, demonstrating a significantly higher risk of recurrent ischemia (P=0.003). Patients with low flow after treatment (n=12) had an 82.5% (CI, 46% to 95%) stroke and stroke/TIA-free survival.
For patients with intracranial VBD (n=34), stroke-free survival was 93% at 24 months overall. The risk was focused in the low flow group, which demonstrated 66% stroke-free survival compared with 100% stroke-free in the normal flow group.
On univariate analysis of the medically treated patients, only flow status (low versus normal distal flow; P=0.003) and presence of basilar disease (P=0.03) significantly affected combined stroke/TIA-free survival. Multivariate analysis with Cox proportional hazards analysis demonstrated that flow status affected event-free survival regardless of lesion location (P=0.03).
Discussion
Patients with symptomatic VBD, especially those with intracranial disease, are at high risk of stroke. Posterior circulation ischemic events are more often the result of either local perforator ischemia or regional hypoperfusion secondary to large vessel disease than thromboembolic artery-to-artery embolic events, which are a prominent feature of carotid stenosis.2 However, the anatomy of collateral vessels in the posterior circulation is such that stenosis in 1 vertebral may be compensated by the other or by flow through the posterior communicating arteries. Subsequently, severe stenosis may not cause significant flow reduction distally. Imaging methods used for assessment of hemodynamic compromise in the anterior circulation12–14 have been less effective in identifying cerebrovascular compromise in the posterior circulation.6 As a result, decision making in the management of VBD has been based primarily on the degree of stenosis, with those >50% to 70% designated as "hemodynamically significant."15 We propose that direct measurement of blood flow in the vertebrobasilar tree can provide information regarding the collateral patterns and hemodynamic effects of VBD. The premise of our management algorithm was that patients without distal flow compromise, although they may still remain at risk for stroke via local perforator ischemia/small vessel disease or emboli, are unlikely to benefit from interventions to augment flow.
Using QMRA to assess posterior circulation hemodynamics, we identified a population with symptomatic VBD and normal distal flow that demonstrated a 100% stroke-free survival at 2 years with medical management alone. Existing retrospective studies have estimated the risk associated with intracranial stenosis (50% to 99%) to be particularly high, at 10% to 15% per year.3,4 Mouffarij et al reported an 80% stroke-free survival at 2 years in a retrospective study of patients with distal vertebral artery or BA stenosis of 50%.16 The recent study by Qureshi et al of 102 patients with symptomatic intracranial VBD (>50% stenosis) demonstrated that 14% of patients experienced recurrent stroke, with a stroke-free survival of 72% at 2 years.3 In comparison, our group of medically managed patients with intracranial VBD (n=34) had a better overall 2-year stroke-free survival of 92%, which likely reflects a preselection of patients felt to be at high risk (because of low flow), for intervention. The stroke risk was concentrated in the low flow patients, whereas the normal flow patients had a 100% stroke-free survival at 2 years. Therefore, even in a population typically defined as high risk, assessment of distal flow can distinguish a subset with a more benign course. Although proximal vertebral disease is generally felt to carry a lower risk of recurrent ischemia,17 our analysis suggests that it is the presence or absence of distal large vessel flow compromise rather than the location of the disease that is most pertinent in predicting stroke risk. However, the small number of stroke end points in this study may lessen our power to detect the significance of other potential predictors of stroke risk.
Reduced basilar flow rates measured by magnetic resonance have been described previously in patients with VBI.18 Our results demonstrate that patients can be stratified on this basis, and those without distal flow compromise are at significantly lower stroke risk, even when compared with published stroke rates in similar patients.3,4,16 Defining appropriate candidates for intervention is especially important given that the efficacy and risks of revascularization strategies have not been fully characterized. Endovascular therapy with angioplasty and stenting, particularly for intracranial disease, has become increasingly feasible over the last decade but has been tempered by concerns for distal embolization of plaque material or fatal vessel rupture.19,20 Technical success rates are reported to range from 83% to 100% for both extracranial21,22 and intracranial disease,23–26 with 88% to 93% of patients remaining symptom-free within a year of the intervention.22,26 However, despite recent series suggesting reduced complication rates,23–25 stroke and death rates up to 35% are still reported.27,28 The efficacy of angioplasty in long-term stroke reduction and long-term patency is not well established. Surgical revascularization using bypass carries higher risks and lower patency rates than that seen from anterior circulation bypass, with an approximate 20% risk of major morbidity/mortality and success rates of 75% to 90%.29,30
Because of the retrospective nature of this analysis, there is a potential for selection, recall, and ascertainment bias. For example, a patient with a fatal ischemic event may be more likely to lack follow-up. However, only 3 patients were lost to follow-up (2 normal flow and 1 low flow), and their outcomes do not qualitatively change the results. The lack of angiographic confirmation of stenosis in all cases is another potential source of selection bias if MRA underestimates stenosis and angiography is performed less frequently in patients designated as normal flow. However, the majority of patients did undergo confirmatory angiography, and high-quality 3-T MRA was used otherwise. Furthermore, repeat analysis excluding patients without angiography yields equivalent results. Recall bias is another potential concern because intervals between follow-ups often span periods of 6 to 12 months. Although patients may not accurately recall TIAs, it is unlikely that a stroke would be under-reported. Ascertainment of end point events was not performed in a blinded fashion, which may also lead to bias, but determination of ischemic events was made in a standard fashion by a multidisciplinary group. Other identifiable and unidentified factors may also alter the stroke risk and are difficult to address in this study. For example, the type of antithrombotic/antiplatelet therapy was frequently altered by the referring physician during the course of a patient’s follow-up. The heterogeneity of treatments and changes over time may affect stroke risk in ways that are not accounted for in our analysis. Consequently, the results need to be interpreted cautiously. A prospective standardized application of our management algorithm would be the next step in verifying these results.
Conclusion
Despite the potential shortcomings of this study, a management algorithm using blood flow measurements appears to stratify patients with symptomatic VBD into 2 groups: (1) no distal flow compromise, likely embolic or local perforator hemodynamic compromise, unlikely to benefit from flow augmentation, and carrying a low risk of recurrent stroke on medical therapy; and (2) distal flow compromise, regional hemodynamic compromise with high risk of stroke, potentially benefiting from flow augmentation by endovascular or surgical means.
Blood Flow Measurements (available online only at http://www.strokeaha.org)
An axial 2D or 3D time-of-flight MRA is performed first. The acquired images are then transmitted to a PC workstation where the NOVA software (VasSol, Inc.) is used to create a rotating 3D surface rendering of the vasculature, including the circle of Willis, using a marching cube algorithm7 (Figure I, top). From the scan line (Figure I, bottom) calculated by a line-fitting algorithm,7 a double-oblique scan is performed using gated 2D phase-contrast magnetic resonance imaged perpendicular to the vessel axis. A flow report including the mean volumetric flow rate (cc/min) on each vessel of interest is created using the NOVA software. All the images were acquired with the use of 1.5-T or 3-T GE MRI scanner (General Electric). Flow measurements performed on 1.5-T or 3-T scanners are equivalent.8
Footnotes
F.T.C. has financial interest in VasSol, Inc. M.Z. is employed by VasSol, Inc.
References
Caplan LR. Vertebrobasilar disease. Adv Neurol. 2003; 92: 131–140.
Caplan LR. Treatment of patients with vertebrobasilar occlusive disease. Compr Ther. 1986; 12: 23–28.
Qureshi AI, Ziai WC, Yahia AM, Mohammad Y, Sen S, Agarwal P, Zaidat OO, Suarez JI, Wityk RJ. Stroke-free survival and its determinants in patients with symptomatic vertebrobasilar stenosis: a multicenter study. Neurosurgery. 2003; 52: 1033–1039.
Anonymous. Prognosis of patients with symptomatic vertebral or basilar artery stenosis. The Warfarin-Aspirin Symptomatic Intracranial Disease (WASID) study group. Stroke. 1998; 29: 1389–1392.
Benesch CG, Chimowitz MI. Best treatment for intracranial arterial stenosis 50 years of uncertainty. The WASID investigators. . Neurology. 2000; 55: 465–466.
Haase J, Magnussen IB, Ogilvy CS, Ojemann RG, Meyer FB, Quest DO, Lawton MT, McDougall CG, Spetzler RF, Samson D, Takahashi A, Yoshimoto T, Ausman JI, Charbel FT, Debrun G, Aletich V, Dujovny M. Evaluating patients with vertebrobasilar transient ischemic attacks. Surg Neurol. 1999; 52: 386–392.
Zhao M, Charbel FT, Alperin N, Loth F, Clark ME. Improved phase-contrast flow quantification by three-dimensional vessel localization. Magn Reson Imaging. 2000; 18: 697–706.
Zhao M, Curcio A, Clark ME, Zhou X, Charbel FT. In vitro validation of MR volumetric flow measurement. Proceedings of the 2004 International Workshop on Flow and Motion. 2004; 148–149.
Jennett WB, Harper AM, Gillespie FC. Measurement of regional cerebral blood-flow during carotid ligation. Lancet. 1966; 2: 1162–1163.
Jawad K, Miller D, Wyper DJ, Rowan JO. Measurement of CBF and carotid artery pressure compared with cerebral angiography in assessing collateral blood supply after carotid ligation. J Neurosurg. 1977; 46: 185–196.
Charbel FT, Zhao M, Amin-Hanjani S, Hoffman W, Du X, Clark ME. A patient-specific computer model to predict outcomes of the balloon occlusion test. J Neurosurg. 2004; 101: 977–988.
Ogasawara K, Ogawa A, Yoshimoto T. Cerebrovascular reactivity to acetazolamide and outcome in patients with symptomatic internal carotid or middle cerebral artery occlusion: a xenon-133 single-photon emission computed tomography study. Stroke. 2002; 33: 1857–1862.
Kleiser B, Widder B. Course of carotid artery occlusions with impaired cerebrovascular reactivity. Stroke. 1992; 23: 171–174.
Grubb RL Jr, Derdeyn CP, Fritsch SM, Carpenter DA, Yundt KD, Videen TO, Spitznagel EL, Powers WJ. Importance of hemodynamic factors in the prognosis of symptomatic carotid occlusion. J Am Med Assoc. 1998; 280: 1055–1060.
Marks MP, Marcellus M, Norbash AM, Steinberg GK, Tong D, Albers GW. Outcome of angioplasty for atherosclerotic intracranial stenosis. Stroke. 1999; 30: 1065–1069.
Moufarrij NA, Little JR, Furlan AJ, Leatherman JR, Williams GW. Basilar and distal vertebral artery stenosis: long-term follow-up. Stroke. 1986; 17: 938–942.
Moufarrij NA, Little JR, Furlan AJ, Williams G, Marzewski DJ. Vertebral artery stenosis: long-term follow-up. Stroke. 1984; 15: 260–263.
Kato T, Indo T, Yoshida E, Iwasaki Y, Sone M, Sobue G. Contrast-enhanced 2d cine phase MR angiography for measurement of basilar artery blood flow in posterior circulation ischemia. AJNR. 2002; 23: 1346–1351.
Takis C, Kwan ES, Pessin MS, Jacobs DH, Caplan LR. Intracranial angioplasty: experience and complications. AJNR. 1997; 18: 1661–1668.
Volk EE, Prayson RA, Perl J II. Autopsy findings of fatal complication of posterior cerebral circulation angioplasty. Arch Pathol Lab Med. 1997; 121: 738–740.
Chastain HD,2nd, Campbell MS, Iyer S, Roubin GS, Vitek J, Mathur A, Al-Mubarak NA, Terry JB, Yates V, Kretzer K, Alred D, Gomez CR. Extracranial vertebral artery stent placement: in-hospital and follow-up results. J Neurosurg. 1999; 91: 547–552.
Higashida RT, Tsai FY, Halbach VV, Dowd CF, Smith T, Fraser K, Hieshima GB. Transluminal angioplasty for atherosclerotic disease of the vertebral and basilar arteries. J Neurosurg. 1993; 78: 192–198.
Gomez CR, Misra VK, Liu MW, Wadlington VR, Terry JB, Tulyapronchote R, Campbell MS. Elective stenting of symptomatic basilar artery stenosis. Stroke. 2000; 31: 95–99.
Mori T, Kazita K, Chokyu K, Mima T, Mori K. Short-term arteriographic and clinical outcome after cerebral angioplasty and stenting for intracranial vertebrobasilar and carotid atherosclerotic occlusive disease. AJNR. 2000; 21: 249–254.
Nahser HC, Henkes H, Weber W, Berg-Dammer E, Yousry TA, Kuhne D. Intracranial vertebrobasilar stenosis: angioplasty and follow-up. AJNR. 2000; 21: 1293–1301.
Rasmussen PA, Perl J II, Barr JD, Markarian GZ, Katzan I, Sila C, Krieger D, Furlan AJ, Masaryk TJ. Stent-assisted angioplasty of intracranial vertebrobasilar atherosclerosis: an initial experience. J Neurosurg. 2000; 92: 771–778.
Gress DR, Smith WS, Dowd CF, Van Halbach V, Finley RJ, Higashida RT. Angioplasty for intracranial symptomatic vertebrobasilar ischemia. Neurosurgery. 2002; 51: 23–27.
Levy EI, Horowitz MB, Koebbe CJ, Jungreis CC, Pride GL, Dutton K, Purdy PD. Transluminal stent-assisted angioplasty of the intracranial vertebrobasilar system for medically refractory, posterior circulation ischemia: early results. Neurosurgery. 2001; 48: 1215–1221.
Ausman JI, Diaz FG, Vacca DF, Sadasivan B. Superficial temporal and occipital artery bypass pedicles to superior, anterior inferior, and posterior inferior cerebellar arteries for vertebrobasilar insufficiency. J Neurosurg. 1990; 72: 554–558.
Hopkins LN, Budny JL. Complications of intracranial bypass for vertebrobasilar insufficiency. J Neurosurg. 1989; 70: 207–211.