AUTHOR(S): Chioffi, Franco, M.D.; Pasqualin, Alberto, M.D.; Beltramello, Alberto, M.D.; Da Pian, Renato, M.D. Department of Neurosurgery (FC, AP, RDP) and Neuroradiology Service (AB), Verona City Hospital, Verona, Italy Neurosurgery 31; 877-885, 1992 ABSTRACT: A SERIES OF 83 patients with cerebral arteriovenous malformations is presented. All patients were evaluated with transcranial Doppler sonography. Thirty-two patients were treated with staged embolization and surgery: 19 of these patients were operated on within 4 weeks of the last embolization, and the remaining 13 patients underwent surgery 6 months or more from embolization. Transcranial Doppler sonography performed 1 day after embolization showed a significant (>60%) reduction of main feeder flow velocity in 72% of patients after the first embolization and in 45% of patients after the second embolization. In no case did such a significant reduction occur after the third embolization. A flow redistribution in the basal vessels (defined as an increase in flow velocity of at least 30% of the initial value) occurred only in patients after the first embolization (64%). On delayed post-embolization studies, complete recovery of flow velocity in the embolized vessel occurred in 46% of patients, and sonographic recruitment of new feeders occurred in the remaining 54%. When main feeder flow velocity (mean) was higher than 120 cm/s after embolization and before surgery, the incidence of postoperative hyperemic complications (cerebral edema and/or intracerebral hematoma) was significantly higher than in patients with a mean flow velocity under 120 cm/s. It is concluded that transcranial Doppler sonography is a valuable method for a noninvasive hemodynamic assessment of shunt flow in arteriovenous malformations, and it permits a physiological monitoring of hemodynamic changes after embolization and allows more precise indications regarding further stages of embolization and timing of surgery after embolization. KEY WORDS: Angiography; Arteriovenous malformation volume; Cerebral arteriovenous malformation; Embolization; Hyperemic complications; Microsurgery; Transcranial Doppler sonography Transcranial Doppler sonography (TCD) provides a useful method for the detection of blood flow velocity in the basal cerebral arteries (1). The relation
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/31/5/877/2549006 by East Carolina University user on 25 March 2018
between peak systolic and end-diastolic velocities reflects the peripheral stream resistance of brain arterioles (1). In patients with cerebral arteriovenous malformations (AVMs) the peripheral stream resistance is very low and blood flow is diverted from the brain to the malformation, causing a rapid shunting of blood. This "sucking" effect causes an altered distribution of blood flow velocity in the basal vessels that is easily detectable on TCD studies (1,16). The amount of the shunt flow through the AVM is very likely related to the occurrence of intra- and postoperative hyperemic complications, defined by some authors as the "normal perfusion pressure breakthrough" (or after-load effect) (5,29,31,32,34,35,41). Recently, the introduction of staged preoperative embolization has permitted a new approach to surgical excision of AVMs, allowing a more "gentle" and sequential reduction of the shunt flow and possibly decreasing the incidence of postoperative hyperemic complications (8,10,18,28,30,39,42). This study was undertaken to evaluate basal hemodynamics in patients with cerebral AVMs and hemodynamic changes occurring after embolization through sequential TCD measurements of flow velocity in AVM feeders. PATIENTS AND METHODS From October 1986 to December 1989, 83 patients with cerebral AVMs were evaluated by TCD sonography. TCD studies were carried out with a 2MHz pulsed frequency transcranial Doppler device with a built-in 64-point Fast Fourier Transformation spectrum analyzer (EME, Ueberlingen, Germany). Systolic, mean, and diastolic flow velocities were measured in centimeters/second and displayed on the screen. The pulsatility index (PI) was defined according to Gosling and King (14) as: PI = (Vs -Vd )/Vm . The common carotid artery was insonated on the neck; the middle cerebral artery (MCA) M1 segment, the anterior cerebral artery (ACA) A1 segment, and the posterior cerebral artery (PCA) P1 segment from the temporal window; the vertebral artery from the mastoid region; and the basilar artery from the foramen magnum window. Patients were divided into five subgroups according to AVM volume: 0 to 10 cm3, 11 to 20 cm3, 21 to 30 cm3, 31 to 50 cm3, and >50 cm3. The calculation of AVM volume was made in the mid- to late arterial phase of the angiogram by multiplying the product of the three main diameters (horizontal and vertical diameters in the anteroposterior projection, and the longitudinal diameter in the lateral projection) by 0.52, as proposed by our group in a recent paper (27). The therapeutic procedures adopted in these patients are shown in Table 1. The group of 32 patients having preoperative embolization were studied in detail. Twenty-one (65%) were under 30 years of age, 10 (31%) were from 30 to 50 years, and 1 was over 50 years. There were 21 men and 11 women. Fifteen of the AVMs were in the right hemisphere (47%), 10 in the left hemisphere (10%), 3 in the corpus callosum (9%), 3 in the cerebellum (9%), and 1 in the juxtapeduncular area (3%). Five (16%) were from 0 to 10 cm3; 9
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Neurosurgery 1992-98 November 1992, Volume 31, Number 5 877 Hemodynamic Effects of Preoperative Embolization in Cerebral Arteriovenous Malformations: Evaluation with Transcranial Doppler Sonography Clinical Study
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/31/5/877/2549006 by East Carolina University user on 25 March 2018
postoperative deficits have also been considered. Transient postoperative deficits were defined as disturbances affecting the motor, speech, visual, and/or mental functions, with complete resolution within 6 months of surgery (only in a few cases within 12 months of surgery). Permanent postoperative deficits were divided into: 1) old deficits, when present before surgery and improved or unchanged after surgery; and 2) new deficits, when appearing after surgery or present before surgery but clearly worsened after surgery. New permanent deficits were considered minor when causing a moderate disability and major when causing a severe disability according to the Glasgow Outcome Scale (37) ; no patient in our series survived in a vegetative state, and no patient died. The period of follow-up ranged from 6 months to 2 years. The data were evaluated using one-way analysis of variance to test whether the means of the five samples (categories of volume: 0-10, 11-20, 21-30, 31-50, and >50 cm3) were different. If so, we used Scheffe's method to determine which samples were different from the others. RESULTS General data In the total series of 83 patients evaluated with TCD, the values of flow velocity detected by TCD in AVM feeders were related to AVM volumes, as shown in Table 2. Although systolic, diastolic, and mean flow velocities increased proportionally to AVM volume, the PI (an indirect parameter of peripheral stream resistance, as defined above) decreased proportionally to AVM volume. In general, the mean differences of the flow velocity values and PI among the samples of AVM volume were highly significant, except for sample 4 (30-50 cm3) vs. sample 5 (>50 cm3). The values referred to these groups of volume were essentially the same; every other comparison between the means of any pair of samples was statistically significant. Independent of the great individual variability, arterial feeders with mean flow velocity >120 cm/s and PI 20 cm3 and in all cerebral AVMs with volumes >30 cm3. In this hemodynamic situation, two patterns were observed on TCD: 1) reversal of the flow direction in the ipsilateral precommunicating segment of the anterior cerebral artery (A1) and posterior cerebral artery (P1); or 2) very low values of flow velocities associated with turbulence in the same arterial segments. Radiographic features of basal vessels, as vessel caliber and ratio of contrast filling between feeders and non-feeders, were related with these TCD findings. In Figure 1, an example of altered blood distribution between the MCA and ACA is shown: the mean flow velocity in the M1 tract is high (118 cm/s), and in the PI low (0.46); comparing the two A1 tracts, the mean flow velocity is significantly lower in the ipsilateral A1 tract than in the contralateral, according to the asymmetrical contrast filling on angiography.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
(28%) from 11 to 20 cm3; 6 (19%) from 21 to 30 cm3; 9 (28%) from 31 to 50 cm3; and 3 (9%) over 50 cm3. Clinical history revealed epilepsy only in 16 patients (50%), hemorrhage only in 9 patients (28%), and epilepsy and hemorrhage in 3 patients (9%). Progressive neurological deficits were present before surgery in two patients (6%), and other clinical features were present in two patients (6%). All patients were investigated by TCD the day before and the day after embolization. These data were collected and analyzed. Selective embolization with surgical sutures (polyene threads) by means of a microcatheter system was performed in all of the patients (a total of 46 procedures), and flow-directed embolization with Silastic sponge was associated with selective embolization in 2 patients (a total of 4 procedures). One procedure was performed in 21 patients (66%) (all selective), two procedures in 4 patients (12%) (1 combined, 3 selective), and more than three procedures in 7 patients (21%) (1 combined, 6 selective). All procedures were done by the transfemoral route. With the exception of a few transient disturbances (headache, vertigo, paresthesia), no significant neurological complication was observed after embolization. Preoperative examination was negative in 28 patients; mild neurological deficits were present in 4 patients (hemianopia in 3, mild motor deficit in 1). Surgery was performed within 10 days after the last embolization in 3 patients, from 11 to 20 days in 6 patients, from 21 to 30 days in 6 patients, from 31 to 60 days in 4 patients, and after 6 months in 13 patients. All patients underwent radical removal of the AVM through a microsurgical approach. Deep controlled hypotension with sodium nitroprusside was induced during surgery in one patient. Mild hypotension with sedation was maintained for the first 24 hours after surgery in most cases; prolonged barbiturate anesthesia in the postoperative course was maintained in a few cases. In the evaluation of intraoperative complications, blood loss of >2000 ml was considered severe; by the term paraventricular bleeding, we have considered long-lasting hemorrhages from small, fragile vessels in the paraependymal area (continuous bleeding for at least 30 min, in spite of accurate bipolar coagulation, with vessels frequently exploding during coagulation and vessel tips easily retracting into the white matter at each coagulation). In the evaluation of postoperative complications, only hyperemic complications were considered. By the term hyperemic complications, as suggested by Batjer (5), we have considered: 1) hematoma in the area of the AVM or in the surrounding tissue not caused by rupture of a residual (angiographically proven) malformation; and 2) edema surrounding the surgical resection, not linked with proximal vascular occlusion, and causing significant shift of the midline structures. No correlation was made between the occurrence of hyperemic complications and the presence of hypoperfused territories on pretreatment angiography. In the evaluation of morbidity, transient
Delayed post-embolization effects Of the group of 32 patients having preoperative embolization, delayed TCD controls were performed in only 13 patients in whom surgery was delayed for various reasons. In all 13 patients, flow velocity of basal arteries months after embolization increased as compared with the immediate post-embolization values. A return to the previous value of mean flow velocity in the embolized vessel occurred in 6 patients (46%) (so called "TCD recanalization"). These data were always confirmed by angiography (Fig. 4). In the other 7 patients, pathological flow velocities were detected for the first time in arteries that were never embolized and were normal or slightly altered on previous TCD examinations. In many of these patients, delayed angiography showed the recruitment of new feeders or the development of a new collateral circulation made by pial or deep
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/31/5/877/2549006 by East Carolina University user on 25 March 2018
medullary vessels (Fig. 5). Relation to postoperative results When considering the group of 32 patients having preoperative embolization followed by surgery, in those with preoperative mean flow velocities 120 cm/s (Table 4). The incidence of postoperative hyperemic complications such as cerebral edema and/or intracerebral hematoma also decreased in patients with mean velocities brought under 120 cm/s. The statistical significance was calculated separately for patients with postoperative hematomas (with or without edema) and for patients with cerebral edema (with or without hematoma). The incidence of transient postoperative deficits was higher in patients with preoperative mean flow velocity >120 cm/s. There was no difference in the incidence of new minor deficits after surgery (Table 5) between AVMs with preoperative mean flow velocities 60% of initial value); and 2) a scant velocity reduction (
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/31/5/877/2549006 by East Carolina University user on 25 March 2018
between peak systolic and end-diastolic velocities reflects the peripheral stream resistance of brain arterioles (1). In patients with cerebral arteriovenous malformations (AVMs) the peripheral stream resistance is very low and blood flow is diverted from the brain to the malformation, causing a rapid shunting of blood. This "sucking" effect causes an altered distribution of blood flow velocity in the basal vessels that is easily detectable on TCD studies (1,16). The amount of the shunt flow through the AVM is very likely related to the occurrence of intra- and postoperative hyperemic complications, defined by some authors as the "normal perfusion pressure breakthrough" (or after-load effect) (5,29,31,32,34,35,41). Recently, the introduction of staged preoperative embolization has permitted a new approach to surgical excision of AVMs, allowing a more "gentle" and sequential reduction of the shunt flow and possibly decreasing the incidence of postoperative hyperemic complications (8,10,18,28,30,39,42). This study was undertaken to evaluate basal hemodynamics in patients with cerebral AVMs and hemodynamic changes occurring after embolization through sequential TCD measurements of flow velocity in AVM feeders. PATIENTS AND METHODS From October 1986 to December 1989, 83 patients with cerebral AVMs were evaluated by TCD sonography. TCD studies were carried out with a 2MHz pulsed frequency transcranial Doppler device with a built-in 64-point Fast Fourier Transformation spectrum analyzer (EME, Ueberlingen, Germany). Systolic, mean, and diastolic flow velocities were measured in centimeters/second and displayed on the screen. The pulsatility index (PI) was defined according to Gosling and King (14) as: PI = (Vs -Vd )/Vm . The common carotid artery was insonated on the neck; the middle cerebral artery (MCA) M1 segment, the anterior cerebral artery (ACA) A1 segment, and the posterior cerebral artery (PCA) P1 segment from the temporal window; the vertebral artery from the mastoid region; and the basilar artery from the foramen magnum window. Patients were divided into five subgroups according to AVM volume: 0 to 10 cm3, 11 to 20 cm3, 21 to 30 cm3, 31 to 50 cm3, and >50 cm3. The calculation of AVM volume was made in the mid- to late arterial phase of the angiogram by multiplying the product of the three main diameters (horizontal and vertical diameters in the anteroposterior projection, and the longitudinal diameter in the lateral projection) by 0.52, as proposed by our group in a recent paper (27). The therapeutic procedures adopted in these patients are shown in Table 1. The group of 32 patients having preoperative embolization were studied in detail. Twenty-one (65%) were under 30 years of age, 10 (31%) were from 30 to 50 years, and 1 was over 50 years. There were 21 men and 11 women. Fifteen of the AVMs were in the right hemisphere (47%), 10 in the left hemisphere (10%), 3 in the corpus callosum (9%), 3 in the cerebellum (9%), and 1 in the juxtapeduncular area (3%). Five (16%) were from 0 to 10 cm3; 9
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
Neurosurgery 1992-98 November 1992, Volume 31, Number 5 877 Hemodynamic Effects of Preoperative Embolization in Cerebral Arteriovenous Malformations: Evaluation with Transcranial Doppler Sonography Clinical Study
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/31/5/877/2549006 by East Carolina University user on 25 March 2018
postoperative deficits have also been considered. Transient postoperative deficits were defined as disturbances affecting the motor, speech, visual, and/or mental functions, with complete resolution within 6 months of surgery (only in a few cases within 12 months of surgery). Permanent postoperative deficits were divided into: 1) old deficits, when present before surgery and improved or unchanged after surgery; and 2) new deficits, when appearing after surgery or present before surgery but clearly worsened after surgery. New permanent deficits were considered minor when causing a moderate disability and major when causing a severe disability according to the Glasgow Outcome Scale (37) ; no patient in our series survived in a vegetative state, and no patient died. The period of follow-up ranged from 6 months to 2 years. The data were evaluated using one-way analysis of variance to test whether the means of the five samples (categories of volume: 0-10, 11-20, 21-30, 31-50, and >50 cm3) were different. If so, we used Scheffe's method to determine which samples were different from the others. RESULTS General data In the total series of 83 patients evaluated with TCD, the values of flow velocity detected by TCD in AVM feeders were related to AVM volumes, as shown in Table 2. Although systolic, diastolic, and mean flow velocities increased proportionally to AVM volume, the PI (an indirect parameter of peripheral stream resistance, as defined above) decreased proportionally to AVM volume. In general, the mean differences of the flow velocity values and PI among the samples of AVM volume were highly significant, except for sample 4 (30-50 cm3) vs. sample 5 (>50 cm3). The values referred to these groups of volume were essentially the same; every other comparison between the means of any pair of samples was statistically significant. Independent of the great individual variability, arterial feeders with mean flow velocity >120 cm/s and PI 20 cm3 and in all cerebral AVMs with volumes >30 cm3. In this hemodynamic situation, two patterns were observed on TCD: 1) reversal of the flow direction in the ipsilateral precommunicating segment of the anterior cerebral artery (A1) and posterior cerebral artery (P1); or 2) very low values of flow velocities associated with turbulence in the same arterial segments. Radiographic features of basal vessels, as vessel caliber and ratio of contrast filling between feeders and non-feeders, were related with these TCD findings. In Figure 1, an example of altered blood distribution between the MCA and ACA is shown: the mean flow velocity in the M1 tract is high (118 cm/s), and in the PI low (0.46); comparing the two A1 tracts, the mean flow velocity is significantly lower in the ipsilateral A1 tract than in the contralateral, according to the asymmetrical contrast filling on angiography.
Redistribution of this article permitted only in accordance with the publisher’s copyright provisions.
(28%) from 11 to 20 cm3; 6 (19%) from 21 to 30 cm3; 9 (28%) from 31 to 50 cm3; and 3 (9%) over 50 cm3. Clinical history revealed epilepsy only in 16 patients (50%), hemorrhage only in 9 patients (28%), and epilepsy and hemorrhage in 3 patients (9%). Progressive neurological deficits were present before surgery in two patients (6%), and other clinical features were present in two patients (6%). All patients were investigated by TCD the day before and the day after embolization. These data were collected and analyzed. Selective embolization with surgical sutures (polyene threads) by means of a microcatheter system was performed in all of the patients (a total of 46 procedures), and flow-directed embolization with Silastic sponge was associated with selective embolization in 2 patients (a total of 4 procedures). One procedure was performed in 21 patients (66%) (all selective), two procedures in 4 patients (12%) (1 combined, 3 selective), and more than three procedures in 7 patients (21%) (1 combined, 6 selective). All procedures were done by the transfemoral route. With the exception of a few transient disturbances (headache, vertigo, paresthesia), no significant neurological complication was observed after embolization. Preoperative examination was negative in 28 patients; mild neurological deficits were present in 4 patients (hemianopia in 3, mild motor deficit in 1). Surgery was performed within 10 days after the last embolization in 3 patients, from 11 to 20 days in 6 patients, from 21 to 30 days in 6 patients, from 31 to 60 days in 4 patients, and after 6 months in 13 patients. All patients underwent radical removal of the AVM through a microsurgical approach. Deep controlled hypotension with sodium nitroprusside was induced during surgery in one patient. Mild hypotension with sedation was maintained for the first 24 hours after surgery in most cases; prolonged barbiturate anesthesia in the postoperative course was maintained in a few cases. In the evaluation of intraoperative complications, blood loss of >2000 ml was considered severe; by the term paraventricular bleeding, we have considered long-lasting hemorrhages from small, fragile vessels in the paraependymal area (continuous bleeding for at least 30 min, in spite of accurate bipolar coagulation, with vessels frequently exploding during coagulation and vessel tips easily retracting into the white matter at each coagulation). In the evaluation of postoperative complications, only hyperemic complications were considered. By the term hyperemic complications, as suggested by Batjer (5), we have considered: 1) hematoma in the area of the AVM or in the surrounding tissue not caused by rupture of a residual (angiographically proven) malformation; and 2) edema surrounding the surgical resection, not linked with proximal vascular occlusion, and causing significant shift of the midline structures. No correlation was made between the occurrence of hyperemic complications and the presence of hypoperfused territories on pretreatment angiography. In the evaluation of morbidity, transient
Delayed post-embolization effects Of the group of 32 patients having preoperative embolization, delayed TCD controls were performed in only 13 patients in whom surgery was delayed for various reasons. In all 13 patients, flow velocity of basal arteries months after embolization increased as compared with the immediate post-embolization values. A return to the previous value of mean flow velocity in the embolized vessel occurred in 6 patients (46%) (so called "TCD recanalization"). These data were always confirmed by angiography (Fig. 4). In the other 7 patients, pathological flow velocities were detected for the first time in arteries that were never embolized and were normal or slightly altered on previous TCD examinations. In many of these patients, delayed angiography showed the recruitment of new feeders or the development of a new collateral circulation made by pial or deep
Downloaded from https://academic.oup.com/neurosurgery/article-abstract/31/5/877/2549006 by East Carolina University user on 25 March 2018
medullary vessels (Fig. 5). Relation to postoperative results When considering the group of 32 patients having preoperative embolization followed by surgery, in those with preoperative mean flow velocities 120 cm/s (Table 4). The incidence of postoperative hyperemic complications such as cerebral edema and/or intracerebral hematoma also decreased in patients with mean velocities brought under 120 cm/s. The statistical significance was calculated separately for patients with postoperative hematomas (with or without edema) and for patients with cerebral edema (with or without hematoma). The incidence of transient postoperative deficits was higher in patients with preoperative mean flow velocity >120 cm/s. There was no difference in the incidence of new minor deficits after surgery (Table 5) between AVMs with preoperative mean flow velocities 60% of initial value); and 2) a scant velocity reduction (
