See the corresponding editorial in this issue, pp J Neurosurg 118: , 2013 AANS, 2013

Transcription

1 See the corresponding editorial in this issue, pp J Neurosurg 118: , 2013 AANS, 2013 The failure of preoperative ethylene-vinyl alcohol copolymer embolization to improve outcomes in arteriovenous malformation management: case series Clinical article Michael Kerin Morgan, M.D., 1 Andrew Stewart Davidson, M.S., 1 Stavros Koustais, M.B.B.S., 1 Mary Simons, M.App.Sc., 2 and Elizabeth Anne Ritson, M.B.B.S. 3 1 Department of Neurosurgery, Australian School of Advanced Medicine; 2 Macquarie University Library; and 3 Macquarie University Hospital, Macquarie University, Sydney, New South Wales, Australia Object. Ethylene-vinyl alcohol copolymer embolization is increasingly used preoperatively in the resection of brain arteriovenous malformations (AVMs). However, the case for embolization improving the outcome of resection has not been evaluated. In this paper the authors set out to compare outcomes after surgery for brain AVMs in 2 consecutive periods of practice. In the first period, selective embolization was used without the use of ethylene-vinyl alcohol copolymer. In the second period, selective embolization with ethylene-vinyl alcohol copolymer was performed. Methods. A consecutive case series (prospectively collected data) was retrospectively analyzed. Adverse outcomes were considered to be an outcome modified Rankin Scale score greater than 2 due to embolization or surgery. Results. A total of 538 surgical cases were included. The percentages of adverse outcomes were as follows: 0.34% for Spetzler-Martin AVMs less than Grade III (1 of 297 cases); 5.23% (95% CI 2.64% 9.78%) for Grade III AVMs (9 of 172 cases); and 17% (95% CI 10% 28%) for AVMs greater than Grade III (12 of 69 cases). There was no improvement in outcomes from the first period to the second period. The adverse outcome for Grade III brain AVMs in the first period was 5.2% (7 of 135 cases) and in the second period (after ethylene-vinyl alcohol copolymer was introduced) it was 5.4% (2 of 37 cases). For AVMs greater than Grade III, the adverse outcome was 12% (6 of 49 cases) in the first period and 30% (6 of 20 cases) in the second period. Conclusions. Outcomes for brain AVM surgery were not improved by ethylene-vinyl alcohol copolymer embolization. Preoperative embolization of high-grade AVMs with an ethylene-vinyl alcohol copolymer did not prevent those hemorrhagic complications which embolization is hypothesized to prevent based on theoretical speculations but not demonstrated in practice. ( Key Words arteriovenous malformation brain vascular disorders ethylene-vinyl alcohol copolymer Onyx embolization surgery Introduced by Luessenhop and Spence14 for the management of brain AVMs more than 50 years ago, embolization has evolved dramatically. As stated in Youmans authoritative textbook, During the past two decades, the role of presurgical embolization of AVMs has become well accepted and firmly established. 29 In their 1995 series review (32 series with a total of 1246 cases) of embolization for brain AVM, Frizzel and Fisher 7 Abbreviations used in this paper: AVM = arteriovenous malformation; EVACE = ethylene-vinyl alcohol copolymer embolization; GOS = Glasgow Outcome Scale; mrs = modified Rankin Scale. demonstrated a cure rate of 5% (in cases without surgery) and a morbidity of 8%. These series were reported prior to the introduction of EVACE. With innovation in embolization techniques, it would be anticipated that this efficacy and/or morbidity rate would have improved. Although embolization is commonly performed preoperatively to assist in the excision of brain AVMs, the effectiveness of this embolization in assisting the subsequent surgery is hard to measure. Viñuela and colleagues 32 attempted to correlate surgeons opinions as to perceived benefit from embolization. They concluded that embolization needed to achieve a volume reduction of the brain 969

2 M. K. Morgan et al. AVM of at least 50% before assistance to the surgeon might be possible and a reduction of at least 75% to be of definite benefit. However, a 75% reduction in volume was not often achieved at the time of that study. In their series of 49 cases, Pasqualin and colleagues 23 reported that there was benefit from embolization in terms of operation time and blood loss. However, this finding has not been reproduced in a subsequent larger series. With the introduction of EVACE, the prospect of achieving a greater degree of nidal penetration and obliteration was promising from case series reports. 9 13,15,20,22,24,27,31,33,34 However, the superiority of EVACE over other embolizing agents to assist surgery has not been clearly established. 13 Furthermore, the benefit of embolization used preoperatively has not been demonstrated. We have reported on the changing role of preoperative embolization. 19 Our analysis led to our discontinuing the use of embolization at the time of this report for Spetzler-Martin Grade I and II brain AVMs. This was because the sum of risks for both procedures was considered greater than the risk associated with surgery alone. However, innovations in embolization and neurosurgical techniques since then call for a reassessment of the role of embolization. One innovation was the introduction of EVACE. We aimed to review the literature of EVACE and also to report on our experience with preoperative embolization for the management of brain AVM for 2 consecutive periods of practice: Period 1, in which preoperative embolization was used prior to the introduction of EVACE; and Period 2, in which embolization was used after the introduction of EVACE but Spetzler-Martin Grade I or II brain AVMs were not embolized. Methods Literature Review The Cochrane Database of Systematic Reviews (Cochrane Reviews), Database of Abstracts of Reviews of Effects (Other Reviews), and Cochrane Central Register of Controlled Trials (Clinical Trials) were searched. The following international trial registries were searched: Clinical Trials Search ( org/), Current Controlled Trials ( com/), and ClinicalTrials.gov ( Two online databases were searched using the following strategies: Ovid MEDLINE ( ) was searched using the following MeSH (Medical Subject Headings) terms and keywords: (ethylene-vinyl alcohol copolymer or ethylene vinyl alcohol copolymer or onyx) AND (exp intracranial arteriovenous malformations/ or avm) AND (surg* or neurosurg* or microsurgery or pre-operative or preoperative). The results were limited to case series of more than 10 patients and clinical trials. Scopus was searched using the following Title, Abstract, Keyword terms: ( ethylene-vinyl alcohol copolymer OR ethylene vinyl alcohol copolymer OR onyx) AND ( brain arteriovenous malformation OR cerebral arteriovenous malformation OR intracranial arteriovenous malformation OR avm) AND (surg* OR neurosurg* OR microsurgery OR pre-operative OR preoperative OR resection). Results were limited to articles, reviews, conference papers, and English-language literature. All abstracts found were read to identify relevant articles. The resulting papers were combined and the final list was checked against hand searches of 3 journals: Journal of Neurosurgery, Neurosurgery, and American Journal of Neuroradiology. Relevant articles were identified and full texts were obtained. Series were included if more than 40 cases were treated with EVACE and if permanent severe disability (including mrs score > 2, GOS score < 5, and Barthel Index < 50 or severe ) could be deduced. Patient Population This study was approved by the Macquarie University Human Ethics Committee and was performed in accordance with institutional ethics committee guidelines. Since 1989, the senior author (M.K.M.) has prospectively collected data on all consecutively enrolled AVMs in a specifically designed AVM database including demographic, clinical, radiological, and treatment-related information. The database specifically included all patients who underwent consultation for but were not considered for surgery. The decision to offer surgery was based on treatment recommendations previously published by the senior author. 5,16 Data were collected in 703 consecutively enrolled patients. A total of 538 patients undergoing resection were eligible for inclusion in the study. Outcomes assessment was performed using the mrs administered preoperatively, at 6 weeks, and at 12 months of follow-up. 6,25,30 Patients with a 12-month mrs score greater than 2 were considered to have attained a poor outcome. To differentiate between the neurological effects of brain AVM at presentation (hemorrhage and focal neurological deficit) and the effects of treatment, adverse outcomes were attributed to one of the following factors at the 6-week clinical assessment: natural history, surgery, embolization, or focused irradiation. For the purposes of this article, adverse outcomes due to surgery also included those due to planned preoperative embolization. The 2 consecutive periods studied were a period when a policy of selective embolization was used (prior to EVACE; Period 1, ) and a second period when a policy of selective embolization with EVACE was used but no patient with a Spetzler-Martin Grade I or II brain AVM underwent embolization (Period 2, 2005 June 2008). The decision to select patients for consideration for embolization was made if the anticipated volume reduction through embolization was thought likely to decrease the risks of subsequent surgery. Embolization was not offered if this was not thought likely to be successful. The decision as to when embolization was thought to be complete was made when the proceduralists believed that further embolization was unlikely to bring additional benefit, when risks of the procedure were deemed to increase, or when a complication of embolization occurred. All embolization procedures were performed under general anesthesia with careful blood pressure control. Normally the largest and straightest of the feeding arteries was chosen for EVACE injection. Superselective angiog- 970

3 Preoperative embolization of brain AVMs raphy was performed to delineate the portion of the nidus that was planned to be filled and to determine whether the branch selected also filled branches to critical brain. In this latter instance, an alternate pedicle was selected. When the EVACE agent was being injected, the distribution of the material was closely monitored so that injection was stopped when reflux was recognized. The injection was recommenced after an interval and continued if forward penetration occurred but again was discontinued when reflux was noted. This continued until the goal had been reached. Utmost care was taken to avoid compromising venous outflow. Following embolization, all patients were managed in intensive care with blood pressure control following the protocol previously reported by the senior author for perioperative surgical management. 18 Statistical Analysis The modified Wald method was used to calculate the 95% confidence interval for a proportion. 1,21 Comparison between groups was made using the chi-square test for linear trends or the Fisher exact test to derive a 2-sided p value using the method to sum small P. Statistical significance was demonstrated for p < Results Literature Review of EVACE Outcomes The Cochrane Database of Systematic Reviews found one relevant systematic review, and the Cochrane Central Register of Controlled Trials found another. 13,26 Sixty-six results were obtained from the MEDLINE search strategy, and 106 results were obtained from the Scopus search strategy. From the weekly topic alert for intracranial arteriovenous malformations in 4 databases, an additional relevant reference was identified. Sufficient information regarding permanent neurological morbidity and mortality could be found in 14 publications (Fig. 1). 9 13,15,20,22,24,27,31,33,34 One of these publications was excluded, as it was assumed to be a subset of cases from another study that was included. The analysis included case series and clinical trials in which the use of EVACE was used contemporaneously with other agents. There was no publication that directly provided evidence that preoperative use of EVACE was of benefit to surgery. However, there were some publications that estimated the volume or size reduction of brain AVM achieved by EVACE. Morbidity was not reported in a consistent way. The mrs score or other standard rankings were inconsistently reported and could not be used in our compilation analysis. For the purpose of our analysis, morbidity was considered to be significant when it could be assessed from the report that it was due to the embolization procedure and resulted in a permanent neurological deficit. In those series in which the morbidity was reported as nondisabling and disabling, only the disabling cases were included in our analysis ,15 In publications in which the mrs was used, patients with an mrs score greater than 2 due to the embolization procedure were included. 9,21,22,24,27,33,34 In one series that used the Barthel Index, an index less than 50 was included. 13 In a series utilizing the GOS, a score of less than 5 was included. 31 From 13 case series, including 12 observational case series and 1 clinical trial, the morbidity and mortality (including mortality; adverse outcomes with mrs score > 2, Barthel Index < 50, GOS score < 5, or a rating of severe ; and urgent surgery required due to embolization) from EVACE was totaled to be 6.6% (95% CI 5.3% 8.2%; 80 of 1211 brain AVMs in 1210 reports). Spetzler-Martin Grade I and II brain AVMs constituted 41% of 982 cases in publications from these series when grade was specified. The goal of EVACE could not be confidently deduced from these papers. Therefore, the intention-to-treat to cure with embolization alone versus the intention-to-treat to assist subsequent management by surgery or radiosurgery is generally unknown. The number of cases cured by embolization alone (taking into account reported late recanalization) was 27% (327 of 1211 brain AVMs) ranging from 0% to 50.6%. Few papers pro- Fig. 1. Case series of more than 40 patients reporting disabling morbidity who underwent embolization for a brain AVM. Outcomes included severe permanent disability (mrs score > 2, Barthel Index < 50, GOS score < 5, or a rating of severe ), death, and emergency surgery. Spetzler-Martin Grades I/II, III, and IV/V were seen in 41.3%, 32.2%, and 26.5%, respectively, of all cases. Twenty-seven percent of AVMs were completely obliterated. The horizontal lines represent the 95% CI, with the small vertical line in the center indicating the point estimate. The dotted line represents the mean from all series (6.6%). n = number of patients. 971

4 M. K. Morgan et al. vided enough detail regarding outcomes for the re spective Spetzler-Martin grades. However, one report provided a breakdown demonstrating that cure was achieved in 98% (154 of 157 cases) of Grade I and II brain AVMs. 27 Clinical Case Series The distribution of cases, outcome data, and comparisons between the 2 consecutive periods of practice are reported in Table 1. These periods were as follows: Period 1 (445 cases), when a policy of selective embolization was used (prior to EVACE); and Period 2 (93 cases), when a policy of selective embolization with EVACE was used, but no patient with a Spetzler-Martin Grade I or II brain AVM underwent embolization. The overall percentage of cases with an mrs score greater than 2 was 0.34% (1 of 297 cases) for Spetzler- Martin grade less than III, 5.23% (9 of 172 cases) for Spetzler-Martin Grade III, and 17% (12 of 69 cases) for Spetzler-Martin grade greater than III. These differences were significant (p < ). Those undergoing embolization had significantly more adverse events (16% [14 of 89 cases]) than those not undergoing embolization (1.78% [8 of 449 cases]). Comparing the 2 periods, Period 1 had fewer adverse outcomes (3.15% [14 of 445 cases]) than Period 2 (8.6% [8 of 93 cases]). This significant difference (p = 0.04) was lost when adjusted for Spetzler-Martin grade. For Spetzler-Martin Grades I and II, very few patients were treated with preoperative embolization, and all of these embolizations were performed in Period 1. Outcomes leading to an mrs score greater than 2 due to surgery for Spetzler-Martin Grade I and II brain AVMs were no different between Periods 1 and 2 (0.4% [1 of 261 cases] and 0% [0 of 36 cases], respectively). For Spetzler-Martin Grade III AVMs, 17% of cases (23 of 135) were preoperatively embolized during the first period and 24% (9 of 37) during the second period. There was no statistically significant difference in the percentage embolized between these 2 periods. Outcomes leading to an mrs score greater than 2 due to surgery for Spetzler-Martin Grade III brain AVMs were no different between Periods 1 and 2 (5.2% [7 of 135 cases] and 5.4% [2 of 37 cases], respectively). For Spetzler-Martin Grade IV and V brain AVMs, 43% (21 of 49 cases) were preoperatively embolized during the first period, and 80% (16 of 20 cases) during the second period. Adverse outcomes leading to an mrs score greater than 2 due to surgery for Spetzler-Martin Grade IV and V brain AVMs were more common in the second period utilizing EVACE. However, this difference failed to reach significance. Adverse outcomes for each of these 2 periods for Spetzler-Martin Grades IV and V were 12% (6 of 49 cases) and 30% (6 of 20 cases), respectively. These differences were not statistically significant. Patients who underwent embolization and sustained an adverse outcome leading to an mrs score greater than 2 due to embolization or surgery are summarized in Table 2. Of the 14 cases, 10 were due to intraoperative or postoperative hemorrhage in the bed of the brain AVM and 4 were due to a nonhemorrhagic source of loss of critical brain. All 7 deaths were associated with hemorrhagic complications. The graphic representation of comparisons in the contribution to adverse outcome due to intraoperative hemorrhage or hemorrhage within the brain AVM bed following surgery for Spetzler-Martin Grade III or higher (grouped by period and whether embolized) is provided in Fig. 2. Discussion Evidence for a Benefit From Preoperative Embolization From the Literature Review There is no Class 1 or 2 evidence for a benefit from Fig. 2. Left: Adverse outcomes from surgery during the 2 consecutive periods as stratified by Spetzler-Martin (SM) grade clusters. The vertical lines indicate the 95% CIs. The horizontal lines within the 95% CIs represent the point estimates, which are also given as percentages. Right: Adverse outcomes due to intraoperative hemorrhage or hemorrhage in the brain AVM bed following surgery for Spetzler-Martin AVMs Grade III or higher are provided for those who underwent embolization and those who did not for each of the 2 consecutive periods. The vertical lines represent the upper limits of the 95% CIs. The tops of the boxes represent the point estimates, which are also given as percentages. Period 1 represents the pre-evace period (15% [65 of 445] cases underwent embolization). Period 2 represents the EVACE period (27% of 93 cases underwent embolization). 972

5 Preoperative embolization of brain AVMs TABLE 1: Distribution of cases, pooled outcome data, and comparisons between the 2 periods of brain AVM surgery* Variable No. of Cases Cause of Complication Resulting in Permanent Deficit SM Grade <III, mrs Score >2 Due to Treatment Value SM Grade III, mrs Score >2 Due to Treatment SM Grade >III, mrs Score >2 Due to Treatment Comparison all cases no embol 449 embol 89 total 538 detected during op or angiography 0.34% (0 2.08%), 1/ % ( %), 9/172 17% (10 28%), 12/69 SM grade p < Period 1 no embol 380 intraop hemorrhage or hemorrhage in AVM bed after op 0%, 0/ %, 3/112 0%, 0/28 nonhemorrhagic source 0.4%, 1/ %, 2/112 0%, 0/28 embol 65 intraop hemorrhage or hemorrhage in AVM 0%, 0/21 4.3%, 1/23 19%, 4/21 bed after op nonhemorrhagic source 0%, 0/21 4.3%, 1/23 9.5%, 2/21 subtotal 445 all complications 0.38% (0 2.4%), 1/ % (2.4 11%), 7/135 12% (5.4 25%), 6/49 Period 2 no embol 68 intraop hemorrhage or hemorrhage in AVM bed after op 0%, 0/36 3.6%, 1/28 0%, 0/4 nonhemorrhagic source 0%, 0/36 3.6%, 1/28 0%, 0/4 embol 25 intraop hemorrhage or hemorrhage in AVM NA 0%, 0/9 25%, 4/16 bed after op nonhemorrhagic source NA 0%, 0/9 12.5%, 2/16 subtotal 93 all complications 0% (0 11%), 0/36 5.4% (0.6 19%), 2/37 30% (14 42%), 6/20 comparison of Periods 1 & 2 NS NS NS * embol = embolization; NA = not applicable; NS = not significant; SM = Spetzler-Martin. Values in these columns are presented as follows: point estimate (95% CI), number of cases with mrs grade greater than 2/total number of cases for grade and period. The point estimates (95% CIs), number of cases with mrs score > 2 for no embolization, and embolization during the entire study period are as follows: 1.78% (0.84% 3.54%), 8 of 449 and 16% (9.5% 25%), 14 of 89, respectively; p <

6 M. K. Morgan et al. TABLE 2: Details of complications in 14 surgical cases with mrs scores greater than 2 treated by preoperative embolization Age (yrs), Sex AVM Size & Location SM Grade Period Presentation Outcome mrs Score 48, M 5 cm, lt temporal III 1 intractable seizures on multiple anticonvulsants w/ memory loss & homonymous hemianopia nonhemorrhagic deficit from embol 3 20, M 2.5 cm, hypothalamic III 1 repeat hemorrhage w/ memory loss & homonymous hemianopia delayed hemorrhage from shunt pass 6 20, M 6 cm, rt parietal IV 1 single hemorrhage w/ complete recovery delayed postop hemorrhage 6 32, M 9 cm, rt occipital IV 1 intractable seizures on multiple anticonvulsants & homonymous hemianopia intraop hemorrhage from meningeal source 6 52, F 6 cm, lt cerebellar V 1 single hemorrhage & ataxia delayed postop hemorrhage 6 23, F 7 cm, lt occipital V 1 multiple hemorrhage & homonymous hemianopia intraop hemorrhage from meningeal source 6 40, M 5.5 cm, lt occipital IV 1 single hemorrhage & quadrantanopia intraop hemorrhage 3 31, F 12 cm, lt posterior hemisphere V 1 progressive nonhemorrhagic hemiparesis & homonymous vision loss deficit from resection of critical brain 3 32, F 7 cm, parietal V 2 multiple hemorrhages & progressive neurocognitive decline (confined to bed) intraoperative hemorrhage 6 59, M 6 cm, rt frontal V 2 single hemorrhage & hemiparesis intraoperative hemorrhage 3 45, F 9 cm, rt hemisphere temporal & V 2 intractable seizures on multiple anticonvulsants & progressive hemiparesis deficit from resection of critical brain 3 parietal lobes 62, F 5 cm, lt basal ganglia & dien- V 2 single hemorrhage & hemiparesis embol brain hemorrhage 6 cephalon 20, M 5 cm, rt paraventricular IV 2 progressive nonhemorrhagic hemiparesis embol brain hemorrhage 6 31, M 5 cm, rt premotor frontal lobe IV 2 progressive nonhemorrhagic memory loss deficit from resection of critical brain 3 974

7 Preoperative embolization of brain AVMs embolization. However, evaluating the benefit of preoperative embolization presents many challenges. First, there is the negotiated extent to which embolization is planned between the interventional neuroradiologist and the surgeon performing the subsequent surgery. The individual variability in this negotiation will be large and cannot be easily reflected in pooled data gathered from many institutions. This negotiation must take into account the acceptable risk transfer from surgeon to interventional neuroradiologist. Second, there is the measure as to what constitutes benefit to surgery. One would hope to see fewer total management complications related to intraoperative or postoperative hemorrhage in the total management of patients undergoing preoperative embolization. Such data are not easily found in the literature. Making surgery easier, by transforming a brain AVM to a tumor (by embolization), may not make the overall management safer. Surrogate measurements of benefits (such as surgeon opinion, total blood loss, and time of surgical procedure) are all challenged by either subjectivity or the lack of matched controlled comparisons. There are no retrospective case series that have identified a benefit of preoperative embolization for all patients undergoing embolization. This would not be of concern except that the morbidity and mortality from the embolization case series are not inconsequential, and for those patients not deriving a benefit from preoperative embolization, these risks may be unacceptable. Prior to the introduction of EVACE, the morbidity associated with embolization of brain AVMs varied considerably between case series reports. In a 1995 review of 32 case series (1246 cases) that should be regarded as the historical starting point of modern neuroradiological intervention for brain AVMs (or perhaps the end point of a previous era), permanent morbidity occurred in a mean of 8% of cases (range 0% 24%), mortality in a mean of 1% of cases (range 0% 6%), and cure in a mean of 5% of cases (range 0% 67%). 7 This compares with the 13 large series reported utilizing EVACE, and suggests that disabling permanent morbidity, mortality, and urgent surgery occur in 6.6% of cases. 9 13,15,20,22,24,27,31,33,34 Cure was achieved by EVACE alone in 27% of patients. For the 13 EVACE series reviewed, Spetzler-Martin Grade I and II AVMs constituted 41% of embolization cases. For the 73% of cases not cured by embolization, it is important to understand the proportion that proceeded to curative intervention as the risks of embolization with EVACE are clearly significant and can only be justified by a high likelihood of subsequent cure. The distribution of the morbidity, mortality risk, and cure rate among the EVACE series for Spetzler-Martin grade is only known for one of these reports. 27 In this report, the morbidity (mrs score > 2), mortality, and cure rate associated with treating AVMs with a Spetzler-Martin Grade less than III were 0%, 0.6%, and 98%, respectively. This contrasts with the morbidity (mrs score > 2), mortality, and cure rate associated with treating Spetzler-Martin grade greater than II of 7.8%, 2.1%, and 11.5%, respectively. If such a low cure rate for Spetzler-Martin Grade greater than II brain AVMs is generalized to all of these series, the risk reduction to subsequent management strategies (whether conservative, radiosurgery, or microsurgery) would need to be great to justify the 10% combined major morbidity and mortality of EVACE. A reduction in operative time or blood loss is insufficient justification for the use of EVACE unless a reduction in morbidity and mortality of the subsequent operative procedure of greater than 10% can be achieved. There has been only one randomized controlled trial comparing embolization with N-butyl cyanoacrylate with that of EVACE. 13 In that trial the authors concluded that there was no difference. However, study size constrains this interpretation. We conclude that no clear-cut benefit has yet been demonstrated from the literature for the current level of use of preoperative embolization of brain AVM with EVACE. Evidence for a Benefit From Preoperative Embolization From Our Case Series Our results suggest that embolization will be unlikely to assist in the management of any Spetzler-Martin Grade I or II brain AVM when used as a presurgical measure (Fig. 2). This is because we are very confident that these AVMs can be effectively removed surgically with a risk of producing a complication resulting in an mrs score greater than 2 in less than 2% (0.3% adverse outcome [95% CI 0% 1.9%]) of cases without preoperative embolization. We have previously reported on our results for low-grade brain AVMs and have suggested that surgery is not likely to be surpassed by either embolization on its own or by radiosurgery in terms of cure and morbidity (mrs score > 1 in 0.7% with an upper 95% confidence of 3%). 5,17 Our present study confirms that these results can be achieved without the assistance of preoperative embolization. When EVACE is reported to be successful as a curative management option for Spetzler-Martin Grade I and II AVMs, it has been claimed that cure rate of 98%, with the low morbidity (2.5%) and mortality (0.6%) rates, is at least as good as the surgical results. 17,27 This gives an estimated total morbidity and mortality of 3.2% (95% CI 1.2% 7.4%). This is not as good as our microsurgical results. In addition, EVACE was not offered as first-line treatment when 1) the AVM was so small that the pedicle could hardly be recognized (that is, the catheterization of the feeding pedicle was anticipated to be risky); 2) the AVM was in the brainstem or in a deep location where it was fed by the anterior choroidal artery, so that the reflux might result in neurological deficit; or 3) the feeding vessels of the AVM were originating as en-passage vessels, and there was not a sufficient segment for Onyx reflux. 27 This differs from the experience with surgery where surgery can be offered with low risk to all patients in whom the natural history is deemed significantly less favorable. 5 In our series, for Grade III brain AVMs there was an increase in the percentage of patients treated with preoperative embolization (from 17% [23 of 135 cases in the period prior to EVACE] to 24% [9 of 37 cases treated with EVACE]). This was not seen to reduce the complication rate (Table 1 and Fig. 2). In our series the number of cases with an adverse outcome (mrs score > 2) increased from 3.0% to 5.4% with the introduction of EVACE into the choice of preoperative embolization agents. Our current strategy of no embolization has too few cases (8 cases) 975

8 M. K. Morgan et al. to compare with earlier periods but is currently tracking favorably with previous experience. Our results also confirm that, although only a small number of Spetzler-Martin Grade IV and V brain AVMs were surgically treated, there remains a high risk of adverse outcome even with preoperative embolization (Fig. 2). Twenty percent of patients with Spetzler-Martin Grade IV and V AVMs in our series had an adverse outcome, leading to an mrs score greater than 2. Of these adverse outcomes, embolization with EVACE accounted for 2 deaths. In addition, there is no evidence from our experience that embolization reduced the risk of hemorrhage during, or as a consequence of, surgery. Despite a significant increase in the number of patients who underwent embolization in Period 2, there was an increase in morbidity (although this did not reach statistical significance). Furthermore, these adverse outcomes were mostly due to hemorrhage, the specific complication that embolization was performed to prevent (Table 2). This experience considerably influenced our strategy and selection of patients when it became apparent that embolization may have emboldened us into treating brain AVMs that would not be otherwise surgically treated without preoperative embolization. Good evidence-based management decisions are challenging with regard to the complex and highly variable care of the patient with a brain AVM. The addition of preoperative embolization in the management of brain AVMs is considered well accepted and firmly established. 29 However, the benefits of embolization need to clearly outweigh the additional risks, complexity, and expense of a management pathway that incorporates embolization. The efficacy of preoperative embolization must be, to a large extent, a subjective judgment by the management team. To know what occurred, and then to guess what would have occurred with an alternative treatment, requires conjecture that can be significantly influenced by experience and bias. Comparing outcomes of patients receiving preoperative embolization with those who do not is clearly inappropriate even when corrected for Spetzler-Martin grade because of the complexity of the variables that inform the treatment decision. Construction of a randomized trial would be of extraordinary difficulty and likely inappropriate due to the large number of variables that need to be considered for the trial, 4 including the large number of variables that influence a decision as to how a patient with a brain AVM is best managed, the rarity of a disease with an annual incidence of presentation less than 1.5/100,000 population, 2,3,8,28 and the variability in proceduralists skills, approaches, and techniques. We suggest that insight into the potential beneficial role for preoperative embolization can be gained by analyzing and reflecting upon individual institutional practice. While this approach suffers from many problems of interpretation and generalization of results, it is a way of providing a practical lens to view practices that may have profound effects on patients. Such experiences should be shared to ensure that established practices are always questioned and that the basis for treatment regimen decisions remain reasonable. The difficulty of making a judgment that embolization may benefit the surgeon is evident from the publication of Viñuela and colleagues, 32 who first addressed the question of quantifying the assistance to surgery by preoperative embolization. The authors found that only 31% of patients selected for embolization prior to excision of brain AVMs achieved a 75% obliteration rate (at a time prior to the introduction of EVACE), a rate (in the opinion of Viñuela et al.) of obliteration below which was of doubtful benefit to the subsequent surgery. With greater penetration of the nidus with EVACE, the expectations for a greater proportion of cases effectively assisted by embolization today could be reasonably postulated. However, in our opinion, such a postulation has not translated into evidence of improved clinical outcomes from either the literature or our results. Conclusions Ethylene-vinyl alcohol copolymer embolization is a cause of serious complications in a small proportion of treated brain AVMs. In addition, there are significant cost implications. However, there is a lack of evidence of a benefit from preoperative EVACE in reducing serious morbidity and mortality. Furthermore, the use of embolization may give the surgeon inappropriate reassurance of protection from complications, such as intraoperative hemorrhage, and influence decisions to recommend surgically treating brain AVMs over conservative management had embolization not been available. With careful selection of brain AVM cases and using surgical strategies and techniques specifically for brain AVM, embolization can be avoided. The temptation to approach brain AVM surgery by converting a brain AVM to a tumor through embolization, and then utilizing the surgical techniques well understood by all neurosurgeons in the resection of brain tumors, is to be resisted when a surgical technique that can remove a brain AVM without embolization is available. Disclosure The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper. Author contributions to the study and manuscript preparation include the following. Conception and design: Morgan. Acquisition of data: Morgan, Koustais, Simons. Analysis and interpretation of data: Morgan, Koustais, Simons. Drafting the article: Morgan, Ritson. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Morgan. Statistical analysis: Morgan, Davidson. References 1. Agresti A, Coull BA: Approximate is better than exact for interval estimation of binomial proportions. Am Stat 52: , Al-Shahi R, Bhattacharya JJ, Currie DG, Papanastassiou V, Ritchie V, Roberts RC, et al: Prospective, population-based detection of intracranial vascular malformations in adults: the Scottish Intracranial Vascular Malformation Study (SIVMS). Stroke 34: , ApSimon HT, Reef H, Phadke RV, Popovic EA: A populationbased study of brain arteriovenous malformation: long-term treatment outcomes. Stroke 33: , Atkinson RP, Awad IA, Batjer HH, Dowd CF, Furlan A, Giannotta SL, et al: Reporting terminology for brain arteriovenous 976

9 Preoperative embolization of brain AVMs malformation clinical and radiographic features for use in clinical trials. Stroke 32: , Davidson AS, Morgan MK: How safe is arteriovenous malformation surgery? A prospective, observational study of surgery as first-line treatment for brain arteriovenous malformations. Neurosurgery 66: , Farrell B, Godwin J, Richards S, Warlow C: The United Kingdom transient ischaemic attack (UK-TIA) aspirin trial: final results. J Neurol Neurosurg Psychiatry 54: , Frizzel RT, Fisher WS III: Cure, morbidity, and mortality associated with embolization of brain arteriovenous malformations: a review of 1246 patients in 32 series over a 35-year period. Neurosurgery 37: , Gabriel RA, Kim H, Sidney S, McCulloch CE, Singh V, Johnston SC, et al: Ten-year detection rate of brain arteriovenous malformations in a large, multiethnic, defined population. Stroke 41:21 26, Gao K, Yang XJ, Mu SQ, Li YX, Zhang YP, Lü M, et al: Embolization of brain arteriovenous malformations with ethylene vinyl alcohol copolymer: technical aspects. Chin Med J (Engl) 122: , Hamada J, Kai Y, Morioka M, Kazekawa K, Ishimaru Y, Iwata H, et al: A mixture of ethylene vinyl alcohol copolymer and ethanol yielding a nonadhesive liquid embolic agent to treat cerebral arteriovenous malformations: initial clinical experience. J Neurosurg 97: , Hauck EF, Welch BG, White JA, Purdy PD, Pride LG, Samson D: Preoperative embolization of cerebral arteriovenous malformations with onyx. AJNR Am J Neuroradiol 30: , Katsaridis V, Papagiannaki C, Aimar E: Curative embolization of cerebral arteriovenous malformations (AVMs) with Onyx in 101 patients. Neuroradiology 50: , Loh Y, Duckwiler GR: A prospective, multicenter, randomized trial of the Onyx liquid embolic system and N-butyl cyanoacrylate embolization of cerebral arteriovenous malformations. Clinical article. J Neurosurg 113: , Luessenhop AJ, Spence WT: Artificial embolization of cerebral arteries. Report of use in a case of arteriovenous malformation. JAMA 172: , Maimon S, Strauss I, Frolov V, Margalit N, Ram Z: Brain arteriovenous malformation treatment using a combination of Onyx and a new detachable tip microcatheter, SONIC: shortterm results. AJNR Am J Neuroradiol 31: , Morgan MK: Classification and decision making in treatment and perioperative management, including surgical and radiosurgical decision making, in Winn HR (ed): Youmans Neurological Surgery, ed 5. Philadelphia: Saunders, 2004, Vol 2, pp Morgan MK, Rochford AM, Tsahtsarlis A, Little N, Faulder KC: Surgical risks associated with the management of Grade I and II brain arteriovenous malformations. Neurosurgery 61 (1 Suppl): , Morgan MK, Winder M, Little NS, Finfer S, Ritson E: Delayed hemorrhage following resection of an arteriovenous malformation in the brain. J Neurosurg 99: , Morgan MK, Zurin AAR, Harrington T, Little N: Changing role for preoperative embolisation in the management of arteriovenous malformations of the brain. J Clin Neurosci 7: , Mounayer C, Hammami N, Piotin M, Spelle L, Benndorf G, Kessler I, et al: Nidal embolization of brain arteriovenous malformations using Onyx in 94 patients. AJNR Am J Neuroradiol 28: , Newcombe RG: Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 17: , Panagiotopoulos V, Gizewski E, Asgari S, Regel J, Forsting M, Wanke I: Embolization of intracranial arteriovenous malformations with ethylene-vinyl alcohol copolymer (Onyx). AJNR Am J Neuroradiol 30:99 106, Pasqualin A, Scienza R, Cioffi F, Barone G, Benati A, Beltramello A, et al: Treatment of cerebral arteriovenous malformations with a combination of preoperative embolization and surgery. Neurosurgery 29: , Pierot L, Januel AC, Herbreteau D, Barreau X, Drouineau J, Berge J, et al: Endovascular treatment of brain arteriovenous malformations using onyx: results of a prospective, multicenter study. J Neuroradiol 36: , Rankin J: Cerebral vascular accidents in patients over the age of 60. II. Prognosis. Scott Med J 2: , Ross J, Al-Shahi Salman R: Interventions for treating brain arteriovenous malformations in adults. Cochrane Database Syst Rev 2010:CD003436, Saatci I, Geyik S, Yavuz K, Cekirge HS: Endovascular treatment of brain arteriovenous malformations with prolonged intranidal Onyx injection technique: long-term results in 350 consecutive patients with completed endovascular treatment course. Clinical article. J Neurosurg 115:78 88, Stapf C, Mast H, Sciacca RR, Berenstein A, Nelson PK, Gobin YP, et al: The New York Islands AVM Study: design, study progress, and initial results. Stroke 34:e29 e33, Starke RM, Lavine SD, Meyers PM, Connolly ES Jr: Adjuvant endovascular management of brain arteriovenous malformations, in Winn HR (ed): Youmans Neurological Surgery, ed 6. Philadelphia: Elsevier, 2011, Vol 4, pp UK-TIA Study Group: United Kingdom transient ischaemic attack (UK-TIA) aspirin trial: interim results. Br Med J (Clin Res Ed) 296: , van Rooij WJ, Sluzewski M, Beute GN: Brain AVM embolization with Onyx. AJNR Am J Neuroradiol 28: , Viñuela F, Dion JE, Duckwiler G, Martin NA, Lylyk P, Fox A, et al: Combined endovascular embolization and surgery in the management of cerebral arteriovenous malformations: experience with 101 cases. J Neurosurg 75: , Weber W, Kis B, Siekmann R, Kuehne D: Endovascular treatment of intracranial arteriovenous malformations with onyx: technical aspects. AJNR Am J Neuroradiol 28: , Xu F, Ni W, Liao Y, Gu Y, Xu B, Leng B, et al: Onyx embolization for the treatment of brain arteriovenous malformations. Acta Neurochir (Wien) 153: , 2011 Manuscript submitted November 16, Accepted November 27, Please include this information when citing this paper: published online January 25, 2013; DOI: / JNS Address correspondence to: Michael K. Morgan, M.D., Health and Medical Development, Macquarie University, 2 Technology Place, Sydney, New South Wales 2109, Australia. michael. morgan@mq.edu.au. 977