Serial Transcranial Doppler Flow Velocity and Cerebral Blood Flow Measurements for Evaluation of Cerebral Vasospasm after Subarachnoid Hemorrhage

Transcription

1 Serial Transcranial Doppler Flow Velocity and Cerebral Blood Flow Measurements for Evaluation of Cerebral Vasospasm after Subarachnoid Hemorrhage Makoto MIZUNO, Shigeyoshi NAKAJIMA, Takeshi SAMPEI, Hiromi NISHIMURA, Hiromu HADEISHI, Akifumi SUZUKI, Nobuyuki YASUI, and Edgar NATHAL-VERA Department of Surgical Neurology, Research Institute for Brain and Blood Vessels-Akita, Akita Abstract Serial transcranial Doppler (TCD) and cerebral blood flow (CBF) examinations were performed in 73 patients with subarachnoid hemorrhage (SAH) due to ruptured intracranial aneurysm to evaluate cerebral vasospasm. Twenty-six (35.6%) of the 73 patients developed ischemic neurological symptoms associated with cerebral vasospasm, which were reversible in all except four patients (5.5%) who demonstrated low-density areas associated with vasospasm on computed tomographic scans. In general, the flow velocities in the middle cerebral arteries began to increase soon after onset of SAH, reaching the maximum between days 8 and 10, subsequently decreasing gradually. There was no signifi cant difference in the highest value and the time course of flow velocities between symptomatic vasospasm and asymptomatic vasospasm patients. Patients with symptomatic vasospasm demon strated two typical time courses of flow velocities: rapid increases in flow velocities that preceded the clinical manifestations of vasospasm (16 patients, 61.5%), and no rapid increases in flow velocities despite the presence of ischemic symptoms (10 patients, 38.5%). In the latter, angiograms demon strated vasospasm in segments distal to those evaluated by TCD examination. These results showed that the degree of cerebral vasospasm cannot be assessed only by the absolute flow velocities. CBF was measured two to 10 (mean 4.7) times within 3 weeks of SAH using the <SUP>133</SUP>Xe intravenous injec tion method. The CBF value remained stable even during the period of major risk of vasospasm. However, the CBF was significantly lower in patients with symptomatic vasospasm on days 8, 9, 10, 13, 14, and 15, when compared with patients without symptomatic vasospasm. Patients who devel oped ischemic neurological symptoms due to vasospasm without increased flow velocity also showed a low CBF value. The combination of serial TCD and CBF measurements, using the former for ear ly detection of vasospasm and the latter for evaluation of cerebral ischemia, is at present the best method for diagnosis of vasospasm. Key words: transcranial Doppler, flow velocity, cerebral blood flow, cerebral vasospasm, subarachnoid hemorrhage, ruptured intracranial aneurysm Introduction Cerebral vasospasm occurring after subarachnoid hemorrhage (SAH) is difficult to predict and evaluate, partly because of the lack of detection methods suitable for use in the acute phase. Angiography is the most reliable method to detect vasospasm, but the method is invasive and time-con suming, so cannot be used for serial monitoring of Received July 17, 1992; Accepted July 5, 1993 vasospasm. The transcranial Doppler (TCD) technique allows noninvasive monitoring of the blood flow velocity in the main trunk of cerebral arteries. '"3"5 The method is generally used for evaluating acute hemodynamic response such as intraoperative monitoring of cerebral blood flow (CBF),''9) assessment of col lateral flow, 14,20 or vessel reactivity to C02 2,4,11,11) and chronic hemodynamic disturbances such as in tracranial stenosis.13,18) TCD examination can detect vasospasm after SAH,') and as increased flow veloc

2 ity may precede the clinical manifestations of ischemia in SAH patients,") it is useful in the evalua tion of cerebral vasospasm.8",",",",22) However, the incidence of false negative results by the method is not negligible.",") This study analyzed a series of 73 patients with SAH due to ruptured aneurysm who underwent TCD flow velocity and CBF measurements to iden tify factors causing false negative indications of vasospasm. Clinical Materials and Methods This study included 73 patients who underwent TCD flow velocity and regional CBF (rcbf) measure ments after SAH due to ruptured intracranial aneurysm. There were 41 females and 32 males, aged from 29 to 81 years (mean 53.3 yrs). In all cases the ruptured aneurysm was verified by angiographic ex amination. The locations of the ruptured aneurysm in the 73 patients were: 28 (38.4%) arising from the middle cerebral artery (MCA), 24 (32.9%) from the anterior communicating artery, 17 (23.3%) from the internal carotid artery, three (4.1 %) from the basilar artery, and one (1.4%) from the anterior cerebral artery (ACA). The neurological grades (Hunt & Kosnik) of the patients on admission were: 46 (63.0%) grade II, 14 (19.2%) grade III, 12 (16.4%) grade IV, and one (1.4%) grade V. Surgery was per formed in the acute stage in all patients except four who showed severe angiographic vasospasm. The progressive neurological symptoms corresponding to the period of vasospasm, such as hemiparesis or aphasia, were defined as the symptomatic ischemic neurological symptoms (symptomatic vasospasms). I. TCD measurements Mean blood flow velocities in the bilateral horizon tal portions of the MCA and the ACA were measured daily using a TCD EME TC2-64 (Eden Medizinische Elektronik, Uberlingen, Germany) from day 1 to day 20 after the last bleeding epi sode. The transtemporal approach was utilized, from a sampling depth of mm with intervals of 5 mm, using the same anatomical point in each patient for all measurements. In patients with mid line aneurysms (arising from the ACA or basilar ar tery), the side demonstrating the higher values was selected for analysis, and for cases with aneurysms arising in the internal carotid artery or MCA, the highest values on the operated side were used. II. CBF measurements Serial measurements of CBF by the 133Xe in travenous injection method (10 mci) were used to calculate the cerebral hemispheric average value of the initial slope index, using a rcbf Analyzer BI 1400 (Valmet, Tampere, Finland) with 14 detectors in one hemisphere. Two to 10 measurements of CBF (mean 4.7 times) were performed SAH. within 3 weeks of III. Cerebral angiography Cerebral angiography was performed in all pa tients on admission, and repeated during hospitaliza tion between days 7 and 35 (mean 17.4) to ensure oc clusion of the aneurysm and evaluate arterial spasm as the angiographic constriction in 36 patients with repeat angiography between days 7 and 17. The angiographic constriction at the same segment evaluated by the TCD method was calculated as follows: constriction = vessel diameter on _ vessel diameter on angiogram on admission postoperative angiogram vessel diameter on angiogram on admission The magnification rates in both studies (angio gram on admission and postoperative angiogram) were the same. IV. Management of cerebral vasospasm Postoperatively, the cerebrospinal fluid drainage (spinal or ventricular) positioned during surgery was used to maintain normal intracranial pressure and prevent hydrocephalus. Patients received intrave nous infusion of hydroxyethyl starch (Hespander ; Kyorin, Tokyo) ( ml/day) to prevent dehy dration beginning the day after operation. Patients demonstrating an increased flow velocity also re ceived dobutamine at a maintenance dosage of 8-25,ug/kg/min (mean 12.4). Patients expected to de velop vasospasm due to severe SAH also received dobutamine even if not demonstrating an increased flow velocity. All patients were therefore managed by hyperdynamic therapy under normal blood volume and normal blood pressure conditions.') V. Statistical analysis Statistical analysis used the Wilcoxon test, with significance at a p value of 0.05 or less. Illustrative Case Reports Case 1: This 56-year-old male was admitted to our in stitution 40 minutes after onset of SAH. On admis sion he was alert with only severe headache and minimal neck stiffness (Hunt & Kosnik grade II).

3 Fig. 3 Case 1. Time course of flow velocities measured by TCD monitoring at the right distal M, (upper: 45 mm depth), right prox imal M, (middle: 55 mm depth), and right A, (lower: 75 mm depth) portions. See text for the velocity values. Fig. 1 Case 1, a 56-year-old male with ruptured right Fig. 2 MCA aneurysm. upper: CT scans on admis sion, showing some blood in the right Sylvian and interhemispheric fissures, and the supra sellar and ambient cisterns. lower: CT scans on day 14, showing no evidence of low-density areas associated with cerebral vasospasm. Case 1. left: Right internal carotid angiogram on day 0, showing a MCA aneurysm (arrow). right: Right internal carotid angiogram on day 16, showing occlusion of the aneurysm and lumen narrowing (arrowheads) in the right A, portion (67.5% constriction). Computed tomography (CT) revealed some blood in the right Sylvian and interhemispheric fissures, and the suprasellar and ambient cisterns (Fig. 1 upper). Right carotid angiography showed a MCA aneurysm (Fig. 2 left). He underwent aneurysm surgery 4 hours after onset through a right frontotemporal craniotomy. TCD monitoring was started on day 1. The flow velocities on day 1 were 34 cm/sec either in the right proximal M1 or in the right distal M1, and 60 cm/sec in the right A,. The flow velocities increased rapidly on day 4 (132 cm/sec in the right proximal M1 portion) with delayed ischemic neuro logical symptoms manifesting from day 6. On day 10, the flow velocity in the right distal M1 portion began to increase rapidly to 156 cm/sec, and on day 14 that in the right A, portion increased to 138 cm/ sec. Subsequently, the flow velocities gradually de creased with disappearance of the ischemic neuro logical deficits (Fig. 3). CT on day 14 showed no evidence of low-density areas associated with cere bral vasospasm (Fig. 1 lower). Postoperative right carotid angiography on day 16 showed severe vaso spasm in the right A, portion (67.5% constriction) (Fig. 2 right). The CBF in the right cerebral hemi sphere was 45.5, 37.4, and 34.0 ml/ 100 g/min on days 1, 8, and 14, respectively. Case 2: This 64-year-old male was admitted to our in stitution 2 hours after onset of SAH. On admission he was drowsy with severe headache and neck stiffness but without other neurological symptoms (Hunt & Kosnik grade III). CT showed severe SAH with blood in all cisterns and multiple lacunar in farcts in the bilateral basal ganglia (Fig. 4 upper). Right carotid angiography 3 hours after onset show ed a right MCA aneurysm and severe arteriosclerosis (Fig. 5 left). He underwent surgery 7 hours after

4 Fig. 6 Case 2. Time course of flow velocities measured by TCD monitoring at the right distal M, (upper: 50 mm depth), right prox imal M, (middle: 60 mm depth), and right A, (lower: 70 mm depth) portions. See text for the velocity values. Fig. 4 Case 2, a 64-year-old male with ruptured right MCA aneurysm. upper: CT scans on admis sion, showing severe SAH with blood in all cisterns and additional multiple lacunar in farcts in the bilateral basal ganglia. lower: CT scans on day 11, showing no evidence of low-density areas due to cerebral vasospasm. sec in the right distal M1, and 40 cm/sec in the right A,. No further increases were subsequently detected (Fig. 6). However, he showed ischemic neurological symptoms (lethargy, disorientation, poor verbal response, restless state) from day 8. The CBF in the right cerebral hemisphere was 43.1, 35.1, 33.3, 33.7, 51.3, 48.6, and 48.9 ml/ 100 g/min on days 3, 4, 8, 9, 10, 11, and 15, respectively. CBF decreased on days 4, 8, and 9. CT on day 11 showed no evidence of low-density areas due to cerebral vasospasm (Fig. 4 lower). Postoperative right carotid angiography on day 21 showed arterial narrowing in the M2 and M3 portions of the MCA (Fig. 5 right), which were thought to be residual vasospasm. Results Fig. 5 Case 2. left: Right internal carotid angiogram on day 0, showing a MCA aneurysm (arrow). right: Right internal carotid angiogram on day 21, showing occlusion of the aneurysm and lumen narrowing (arrowheads) in the right M2 and M3 portions of the MCA. onset. The aneurysm was clipped through a right frontotemporal craniotomy. TCD monitoring was started on day 1. The flow velocities on day 1 were 50 cm/sec in the right proximal M, portion, 56 cm/ I. Changes in flow velocities Figure 7 shows the changes in the daily mean flow velocity in the horizontal portion of the MCA in the 73 patients after SAH. In general, the flow velocity began to increase soon after onset of SAH, reaching the maximum between days 8 and 10, and subse quently decreasing gradually (Fig. 7 upper). All pa tients with the aneurysm not located in the midline demonstrated flow velocities on the operated side significantly higher than those on the opposite side on days 6-8 and (Fig. 7 lower). Twenty-six (35.6%) of the 73 patients developed ischemic neurological symptoms associated with cerebral vasospasm. These symptoms were reversible in all except four cases (5.5%). In our series, irreversi

5 ble symptomatic vasospasm was associated with the appearance of low-density areas on CT scans. The highest flow velocities measured during the 20 day period and angiographic constriction in the patients with symptomatic and asymptomatic vasospasm are shown in Table 1. There was only a significant difference between the flow velocities for the group with low-density areas on CT scan and the group with asymptomatic vasospasm. No significant dif ference was found in the angiographic constriction of vessels between the three groups, although the constriction in patients with symptomatic vaso spasm (including the group with low-density areas) tended to be more severe than in patients without symptomatic vasospasm. The changes in flow veloc ity in patients with symptomatic and asymptomatic vasospasm also showed no difference except on days 13 and 18. There were two typical time courses for flow velocities in patients with symptomatic vasospasm. One group showed rapid increases in flow velocities that preceded the clinical manifestation of vasospasm (16 patients, 61.5%) (illustrative Case 1), while the other group developed ischemic symptoms without an increase in flow velocity (10 patients, 38.5%) (illustrative Case 2). In the first group, vasospasm could be diagnosed reliably before the ap pearance of clinical manifestations. Three patients in this group developed cerebral infarction due to vasospasm (Fig. 8 upper). In the second group, angiography demonstrated vasospasm in segments distal to those investigated by TCD flow analysis, so the vasospasm could not be detected in the early stage. One patient in this group developed cerebral in farction (Fig. 8 lower). II. CBF measurements Figure 9 shows the daily mean CBF value in the 73 patients with SAH in the same hemisphere used for TCD monitoring. The CBF tended to remain stable even during the period of major risk of Fig. 7 upper: Time course of the daily mean flow velocities in the horizontal portion of the MCA of 73 patients after SAH. The flow veloc ity values increased soon after onset of SAH, with maximum values recorded between days 8 and 10. lower: Time course of the daily mean flow velocities in the horizontal portion of the MCA on the operated ( ) and non-operated (o) sides of 45 patients with ruptured internal carotid artery or MCA aneurysms. Flow veloc ities measured on the operated side were significantly higher than those on the opposite side (*p < 0.05, **p < 0.01). vasospasm, but the CBF in patients who developed symptomatic vasospasm was significantly lower on days 8, 9, 10, 13, 14, and 15. Serial CBF measure ment was more useful than serial TCD examination Table 1 Highest mean flow velocities obtained during the 20 day period and angiographic constriction in patients with symptomatic and asymptomatic vasospasm (VS)

6 Fig. 9 Time courses of the daily mean CBF in patients with symptomatic ( ) and asymptomatic ((D) vasospasm. The CBF in patients with symp tomatic vasospasm was significantly lower on days 8, 9, 10, 13, 14, and 15 (*p < 0.05, **p < 0.01). Fig. 8 Two typical time courses of flow velocity in pa for the evaluation of cerebral ischemia caused by vasospasm. Moreover, all patients who showed no increase in flow velocity but developed ischemic neurological symptoms associated with vasospasm demonstrated tients with symptomatic vasospasm. upper: Three patients who developed cerebral infarc tion due to vasospasm after demonstrating rapid increases in flow velocities. lower: One patient who developed cerebral infarction due to vasospasm without a rapid increase in flow velocities. a low CBF value. Discussion Our series of patients showed that flow velocity began to increase soon after onset of SAH, and reached a maximum between days 8 and 10, gradual ly decreasing thereafter. Angiographic studies during the period of vasospasm showed arterial narrowing in the same segments demonstrating increased flow velocities. We also found that patients with lateral aneurysms demonstrated significantly higher flow velocities on the operated side than on the non operated side. This result disagrees with the theory that removal of subarachnoid clots reduces the severi ty of vasospasm, and is attributed to the rapid action of vasoconstrictive substances in vessels without in tact arachnoid sheets') and to the vasoreaction associ ated with operative manipulation. One problem in interpreting the results of TCD in vestigation has been the diagnostic utility of absolute flow velocity values. Maximum mean flow velocities in the range of cm/sec are reportedly not critical and never lead to brain infarction, while flow velocities over 200 cm/sec represent a dangerous situation with a high possibility of cerebral ischemia.1,211 In our study, we found no significant difference between the absolute flow velocities in pa tients with and without symptomatic vasospasm. Some patients had flow velocities above 200 cm/sec without ischemic symptoms, while others had values around 120 cm/sec with neurological manifesta tions. This demonstrates that flow velocities are in fluenced not only by the vessel diameter, but also by other factors such as age, PaCO2i blood pressure, flow volume, intracranial pressure, and drugs in use. Moreover, treatment for vasospasm such as hyper volemic-hypertensive therapy') or hyperdynamic therapy') will further influence the flow velocity, so the significance of isolated values is reduced. At present, not enough studies have evaluated the in fluence of these factors on flow velocity. Clearly, although flow velocity is influenced more by the af fected vessel diameter than by other factors, the ex tent of cerebral ischemia cannot be assessed only by

7 the absolute flow velocity. The best predictor for symptomatic vasospasm before the appearance of ischemic symptoms was a rapid increase in flow velocity within a few days. Se rial TCD examinations could therefore be very use ful for predicting vasospasm, in agreement with Seiler et al.,`) who found that a rapid increase in flow velocity is an indicator of severe ischemia. How ever, many patients developed symptomatic vaso spasm despite low or moderate flow velocities. The ischemic symptoms were thought to arise from vaso spasm in more distal segments judging by the angio graphic findings, and may be due to removal of subarachnoid clots extending only to the major cis terns, so some blood remained in the distal portions of the arteries. Klingelhofer et al. 10) pointed out that with a pronounced increase in intracranial pres sure, evaluation of the severity and time course of vasospasm based solely upon mean flow velocity can lead to false negative results. In our patients, the intracranial pressure may have increased despite the presence of cerebrospinal fluid drainage result ing in vasospasm which could not be detected by serial TCD examination. In some of our patients, migration of vasospasm from the proximal to distal portions of the M, and A, was detected. A rapid increase in flow velocity is therefore indicative of vasospasm, but low velocities do not exclude the development of vasospasm. CBF measurement is a useful technique for detec tion of cerebral ischemia, and has been suggested for sequential monitoring of vasospasm, as the values of CBF are closely related to the degree of vaso spasm.s." 23) We always perform serial CBF measure ments by the 133Xe intravenous method in patients with suspected cerebral vasospasm after SAH. Although the method does not provide three-dimen sional analysis, it provides useful information on the cerebral circulation and is highly reproducible. Pa tients injected with a 10 mci 133Xe bolus receive a radiation dose of Gy in the whole body, 0.39 Gy in the lung, and 0.29 Gy in the brain. These doses are thought to cause no hazard to the patients, even if repeated every day for 10 days. Repeated quantitative measurement of CBF allows evaluation of the changes in CBF due to cerebral ischemia caused by cerebral vasospasm, as well as evaluation of the effect of hyperdynamic treatment. In our series, the daily mean value of CBF after SAH remained stable even during the period when vasospasm is expected to occur. This contrasts with the rapid changes in flow velocities in the same period, and may be due to the effects of the hyper dynamic therapy. All patients with symptomatic vasospasm, even with normal flow velocities, demonstrated low CBF values during the phase of vasospasm that correlated with ischemia, so CBF measurement was more useful than TCD examina tion. Nevertheless, it should be recognized that the results of CBF studies are affected by many factors and the number of studies is limited for each patient, so variations in CBF between each examination re main uncertain. The combination of serial TCD and CBF studies, using the former for the detection of vasospasm and the latter for evaluation of cerebral ischemia, is at present the best method for the diagnosis of cerebral vasospasm, and is of great benefit in the general management of patients with SAH. References 1) Aaslid R, Huber P, Nornes H: Evaluation of cerebrovascular spasm with transcranial Doppler ultrasound. J Neurosurg 60: 37-41, ) Aaslid R, Lindegaard KF, Sorteberg W, Nornes H: Cerebral autoregulation dynamics in humans. Stroke 20: 45-52, ) Aaslid R, Markwalder TM, Nornes H: Noninvasive transcranial Doppler ultrasound recording of flow ve locity in basal arteries. J Neurosurg 57: , ) Bishop CCR, Powell S, Rutt D, Browse NL: Transcranial Doppler measurement of middle cerebral artery blood flow velocity: A validation study. Stroke 17: , ) Furguson GG, Farrar JK, Meguro K, Peerless SJ, Drake CG, Barnett HJM: Serial measurement of CBF as a guide to surgery in patients with ruptured in tracranial aneurysm. J Cereb Blood Flow Metab 1: , ) Hadeishi H, Mizuno M, Suzuki A, Yasui N: Hyper dynamic therapy for cerebral vasospasm. Neurol Med Chir (Tokyo) 30: , ) Halsey JH, Mcdowell HA, Gelmon S, Morawetz RB: Blood velocity in the middle cerebral artery and regional cerebral blood flow during carotid en darterectomy. Stroke 20: 53-58, ) Harders AG, Gilsbach JM: Time course of blood ve locity changes related to vasospasm in the circle of Willis measured by transcranial Doppler ultrasound. J Neurosurg 66: , ) Kassell NF, Peerless SJ, Durward QJ, Back DW, Drake CG, Adams HP: Treatment of ischemic deficits from vasospasm with intravascular volume ex pansion and induced hypertension. Neurosurgery 11: , ) Klingelhofer J, Sander D, Holzgraefe M, Bischoff C, Conrad B: Cerebral vasospasm evaluated by transcranial Doppler ultrasonography at different in tracranial pressures. J Neurosurg 75: , 1991

8 11) Knuckey NW, Fox RA, Surveyor I, Stokes BAR: Ear ly cerebral blood flow and computerized tomography in predicting ischemia after cerebral aneurysm rup ture. J Neurosurg 62: , ) Kyoi I, Hashimoto H, Tokunaga H, Morimoto T, Hiramatsu K, Tsunoda S, Tada T, Utsumi S: Time course of blood velocity change and clinical symp toms related to cerebral vasospasm and prognosis after aneurysmal surgery. No Shinkei Geka 17: 21 30, 1989 (in Japanese) - 13) Lindegaard KF, Bakke SJ, Aaslid R, Nornes H: Dop pler diagnosis of intracranial artery occlusive disorders. J Neurol Neurosurg Psychiatry 49: , ) Lindegaard K-F, Bakke SJ, Grolimund P, Aaslid R, Huber P, Nornes H: Assessment of intracranial hemodynamics in carotid artery disease by transcranial Doppler ultrasound. J Neurosurg 63: , ) Lindegaard KF, Lundar T, Wiberg J, Sioberg D, Aaslid R, Nornes H: Variations in middle cerebral artery blood flow investigated with noninvasive transcranial blood velocity measurements. Stroke 18: , ) Markwalder TM, Grolimund P, Seiler RW, Roth F, Aaslid R: Dependency of blood velocity in the middle cerebral artery on endtidal carbon dioxide partial pressure: A transcranial ultrasound Doppler study. J Cereb Blood Flow Metab 4: , ) Mizuno M, Asakura K, Hadeishi H, Nishimura H, Sampei T, Suzuki A, Yasui N: Combination of serial transcranial Doppler examinations and cerebral blood flow studies for the management of cerebral vasospasm after subarachnoid hemorrhage. No Shinkei Geka 18: , 1990 (in Japanese) 18) Moritake K, Yonekawa Y, Nagasawa S, Kaneko T, Handa H: The prospect of usefulness of transcranial Doppler sonography in cerebrovascular surgery. No Shinkei Geka 15: , 1987 (in Japanese) 19) Nelson RJ, Perry S, Harries TK, Pickard JD: Transcranial Doppler ultrasound studies of cerebral autoregulation and subarachnoid hemorrhage in the rabbit. J Neurosurg 73: , ) Schneider PA, Ringelstein EB, Rossman ME, Dilley RB, Sobel DF, Otis SM, Bernstein EF: Importance of cerebral collateral pathways during carotid en darterectomy. Stroke 19: , ) Seiler RW, Grolimund P, Aaslid R, Huber P, Nornes H: Cerebral vasospasm evaluated by transcranial ultrasound correlated with clinical grade and CT visualized subarachnoid hemorrhage. J Neurosurg 64: , ) Sekhar LN, Wechsler LR, Yonas H, Luyckx K, Obrist W: Value of transcranial Doppler examination in the diagnosis of cerebral vasospasm after subarachnoid hemorrhage. Neurosurgery 22: , ) Yamagata S, Kikuchi H, Ihara I, Nagata I, Morooka Y, Naruo Y, Koizumi T, Hashimoto K, Minamikawa J, Miyamoto S, Mitsuno K, Matsumoto M, Yamazoe N, Akiyama Y: Cerebral blood flow as a prognostic indication in subarachnoid hemorrhage. Neurol Med Chir (Tokyo) 28: , 1988 (in Japanese) Address reprint requests to: M. Mizuno, M.D., Depart ment of Surgical Neurology, Research Institute for Brain and Blood Vessels-Akita, 6-10 Senshu-kubota machi, Akita 010, Japan.