CORRESPONDENCE TOPIC
Commentary: The ARUBA Trial A randomized trial of unruptured brain arteriovenous malformations (ARUBA) was recently published, and the authors have concluded that medical management alone is superior to medical management with interventional therapy for the prevention of death or stroke in patients with unruptured brain arteriovenous malformations (AVMs) followed up for 33 months. This is the first randomized trial to compare conservative therapy with intervention for unruptured AVMs. The authors of the ARUBA trial aimed to address a valid question, and their efforts are to be commended. However, the study design and execution have serious flaws that undermine the validity of their conclusions. Mohr et al1 recently reported a randomized trial of unruptured brain arteriovenous malformations (ARUBA). This prospective, multicenter, parallel-design, randomized, controlled trial was intended to compare intervention with medical management for unruptured brain arteriovenous malformations (bAVMs). Eligible patients were 18 years of age or older with no previous hemorrhage or intervention and with bAVMs that were deemed suitable for complete obliteration. Of the 1740 patients screened, 726 were deemed eligible. Of those 726 eligible patients, 323 refused to enroll, and 177 patients had their management decision made by the treating team outside of the randomization process. Only 7 (3%) patients dropped out of the trial. The primary outcome measure was death or symptomatic stroke, and the secondary outcome was clinical impairment, defined as a score of 2 or higher on the modified Rankin Scale. The authors initially planned to enroll 800 patients for an 87.5% power to detect a 40% reduction in the risk of symptomatic stroke or death. Due to the low number of patients enrolled in the study, the goal sample size was reduced to 400 patients for an 80% power to detect a 46% reduction in risk. However, based on recommendations of an independent data safety monitoring committee, the study ultimately stopped enrolling patients after randomizing 226 patients between April 2007 and April 2013. Of those 226 patients, 114 were randomized to intervention and 109 to medical management. At the time of analysis, 53 of the treatment patients had not completed therapy and 20 had not initiated therapy. Seven patients randomized to medical management ultimately crossed over to the intervention arm, although no hemorrhages were identified in these patients. Interventional therapy included any neurosurgical, endovascular, or radiotherapy procedure (single or multiple) deemed appropriate by local ARUBA investigators to achieve complete bAVM obliteration. In the intervention group, 5 were treated with neurosurgical procedures, 30 with embolization, 31 with radiotherapy, and 28 with multimodality therapy. The average followup was 33 months and was balanced between the intervention and medical management arms. The primary endpoint was achieved in 35 (30.7%) of the 114 patients randomized to treatment and in
E96 | VOLUME 75 | NUMBER 1 | JULY 2014
Correspondence
11 (10.1%) of the 109 patients in the medical management arm. Secondary outcome was achieved in 24 (46.2%) patients receiving intervention and in 8 (15.1%) of the medically managed patients at 30 months. At 36 months, the secondary outcome was achieved by 17 (38.6%) of the treatment patients and 6 (14%) of the patients in the medical group. The authors conclude that medical management is superior to intervention in preventing death or stroke in patients with unruptured bAVMs followed for 33 months. The authors of the ARUBA trial aimed to address a valid question and their efforts are to be commended. However, the study design and execution have serious flaws that undermine the validity of their conclusions. To start, the authors fail to explain why such a large number of eligible patients, who did not refuse enrollment, were not randomized. Of the 726 eligible patients, 323 refused and 226 were enrolled. The other 177 patients were managed outside of the randomization process. There is no discussion regarding the characteristics of these patients and their lesions. The authors do acknowledge that potentially useful outcome data could be obtained from these patients. However, they briefly explain that participating centers did not respond to offers to create such a registry. Allowing treatment of eligible and consenting patients outside of the protocol is a critical flaw in design, introducing a layer of selection bias that need not complicate this study. Additionally, the study design lacks any standardization of the treatment arm. A recent large meta-analysis of treated bAVMs reported obliteration rates of 96% for surgical resection, 38% for stereotactic radiosurgery, and 13% for embolization.2 This analysis included ruptured and unruptured AVMs as well as a larger proportion of higher Spetzler-Martin grade AVMs. The neurosurgical literature supports these findings, and current recommendations, based on pooled-analysis, are surgical resection as the primary therapy for Spetzler-Martin grade I and II bAVMs.3 Without further detail regarding the treatment arm, it remains difficult to resolve why only 5 patients received surgical resection alone when 76 patients in the treatment arm had grade I or II bAVMs. The authors also provide no discussion regarding the treatment details of the radiotherapy and embolization groups. What type of radiotherapy was used and with what doses? Were serial or single embolizations performed and why? Serial embolizations certainly increase the risk of adverse events compared with single therapeutic options. Additionally, embolization is generally not considered a curative procedure, as evidenced by the previously mentioned obliteration rate of 13%, and is mainly used in combination therapies. It is concerning that 30 of the treatment patients were managed with embolization alone. The authors attempt to justify the very high rate of hemorrhage or stroke in the interventional therapy group (30.7%) by citing
www.neurosurgery-online.com
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
CORRESPONDENCE
recently published complication rates in a large systematic metaanalysis. The authors quote complication rates of 29% for surgery, 25% for embolization, and 13% for radiosurgery.2 The authors do mention that the meta-analysis includes lesions that have hemorrhaged in addition to those that have not. Unfortunately, they do not explore the population that they are making a comparison with any further. First, the quoted complication rates include all complications. The actual meta-analysis separates out severe complications, which include death and permanent neurological deficit. Remembering that the ARUBA primary outcome was death or symptomatic stroke and the secondary outcome was clinical impairment with a modified Rankin Scale score of 2 or higher, it makes more sense to compare the ARUBA outcomes with the severe complication rates in the meta-analysis. The rates for severe complications were 7.4% (range, 0%-40%) for surgery, 6.6% (range, 0%-28%) for embolization, and 5.1% (range, 0%-21%) for radiosurgery. All of these rates are significantly below what is presented in the ARUBA trial. Second, 84% of the patients, 11,521 of 13,698 included in the meta-analysis, presented with a ruptured AVM. Clearly patients presenting with cerebral hemorrhage are a very different population from that of the ARUBA trial. Last, in the metaanalysis referenced by the ARUBA authors, 61% of the patients had Spetzler-Martin grades equal to or less than 3. At time of analysis, 93% of the ARUBA trial patients randomized to interventional therapy had Spetzler-Martin grades equal to or less than 3. The Spetzler-Martin grading system has been well validated in the literature as a predictor of surgical morbidity and mortality. Given that the ARUBA trial included a much greater proportion of low-grade lesions, outcomes are expected to be better compared with the meta-analysis presented by the authors. On closer examination, the authors attempting to support their findings with previously published literature is a thinly veiled misrepresentation of their study population. There is reference to evaluation of lesion eradication by cerebral angiography in the Methods section. However, the authors do not include data regarding eradication in the article. The temporal relationship between treatment and achievement of primary and secondary outcomes is important when considering radiotherapy or serial embolization. Radiotherapy has a significantly delayed therapeutic effect. Serial embolizations will not provide therapeutic effect until complete eradication is achieved. Of the 114 patients randomized to treatment, 53 had not completed therapy and 20 had not initiated therapy at the time of analysis. Published literature has shown that there is no significant protection from the risk of hemorrhage provided by partial treatment of an AVM.4-7 Any events in the treatment arm that occurred before eradication should be noted by the authors. Given the lack of information and heterogeneity regarding therapies, it is impossible to understand what the treatment arm
NEUROSURGERY
represents. That said, it is unclear to what the authors are concluding medical management is superior. It seems irresponsible to conclude superiority of medical management for bAVMs when the treatment arm provides therapies that are seemingly well below the current standard of care for eradication. The length of follow-up is the most obvious pitfall of this analysis. The 33-month follow-up for a disease with a lifelong risk of stroke and hemorrhage is clearly inadequate to make conclusions. The authors' proposed follow-up of 5 years is also comparatively short. Treatment of bAVMs compresses the risk associated with these lesions. It may require more than the planned follow-up period to demonstrate the benefit of eradicating these lesions. The ARUBA trial is the first randomized, controlled trial attempting to evaluate intervention vs medical management for unruptured bAVMs. Unfortunately, errors in study design and execution as well as a relative lack of information regarding the treatment arm and the enrollment process invalidate the authors' conclusions. Indeed, the fact that in the medically managed group there was a 10% incidence of death or stroke over a time interval of less than 3 years emphasizes the need to provide a lowrisk cure for these patients. Disclosure The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.
Jonathan Russin Robert Spetzler Phoenix, Arizona 1. Mohr JP, Parides MK, Stapf C, et al. Medical management with or without interventional therapy for unruptured brain arteriovenous malformations (ARUBA): a multicentre, non-blinded, randomised trial. Lancet. 2013;383 (9917):614-621. 2. van Beijnum J, van der Worp HB, Buis DR, et al. Treatment of brain arteriovenous malformations: a systematic review and meta-analysis. JAMA. 2011;306(18):2011-2019. 3. Spetzler RF, Ponce FA. A 3-tier classification of cerebral arteriovenous malformations. Clinical article. J Neurosurg. 2011;114(3):842-849. 4. Lunsford LD, Kondziolka D, Flickinger JC, et al. Stereotactic radiosurgery for arteriovenous malformations of the brain. J Neurosurg. 1991;75(4):512-524. 5. Karlsson B, Lax I, Söderman M. Risk for hemorrhage during the 2-year latency period following gamma knife radiosurgery for arteriovenous malformations. Int J Radiat Oncol Biol Phys. 2001;49(4):1045-1051. 6. Miyamoto S, Hashimoto N, Nagata I, et al. Posttreatment sequelae of palliatively treated cerebral arteriovenous malformations. Neurosurgery. 2000;46(3):589-594; discussion 594-585. 7. Fournier D, TerBrugge KG, Willinsky R, Lasjaunias P, Montanera W. Endovascular treatment of intracerebral arteriovenous malformations: experience in 49 cases. J Neurosurg. 1991;75(2):228-233. 10.1227/NEU.0000000000000357
VOLUME 75 | NUMBER 1 | JULY 2014 | E97
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
CORRESPONDENCE
ERRATUM Response to Journal Club: Magnetic Resonance Imaging-Guided Focused Laser Interstitial Thermal Therapy for Intracranial Lesions: Single-Institution Series. Neurosurgery. 2014;74(5):565. The first sentence should have read: “We thank Dr Rohatgi and colleagues for their excellent Journal Club submission and appreciate their favorable assessment of our experience in laser interstitial thermal therapy for intracranial lesions.” DOI: 10.1227/01.neu.0000450985.53210.72 Copyright © by the Congress of Neurological Surgeons
E98 | VOLUME 75 | NUMBER 1 | JULY 2014
www.neurosurgery-online.com
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
Commentary: The ARUBA Trial A randomized trial of unruptured brain arteriovenous malformations (ARUBA) was recently published, and the authors have concluded that medical management alone is superior to medical management with interventional therapy for the prevention of death or stroke in patients with unruptured brain arteriovenous malformations (AVMs) followed up for 33 months. This is the first randomized trial to compare conservative therapy with intervention for unruptured AVMs. The authors of the ARUBA trial aimed to address a valid question, and their efforts are to be commended. However, the study design and execution have serious flaws that undermine the validity of their conclusions. Mohr et al1 recently reported a randomized trial of unruptured brain arteriovenous malformations (ARUBA). This prospective, multicenter, parallel-design, randomized, controlled trial was intended to compare intervention with medical management for unruptured brain arteriovenous malformations (bAVMs). Eligible patients were 18 years of age or older with no previous hemorrhage or intervention and with bAVMs that were deemed suitable for complete obliteration. Of the 1740 patients screened, 726 were deemed eligible. Of those 726 eligible patients, 323 refused to enroll, and 177 patients had their management decision made by the treating team outside of the randomization process. Only 7 (3%) patients dropped out of the trial. The primary outcome measure was death or symptomatic stroke, and the secondary outcome was clinical impairment, defined as a score of 2 or higher on the modified Rankin Scale. The authors initially planned to enroll 800 patients for an 87.5% power to detect a 40% reduction in the risk of symptomatic stroke or death. Due to the low number of patients enrolled in the study, the goal sample size was reduced to 400 patients for an 80% power to detect a 46% reduction in risk. However, based on recommendations of an independent data safety monitoring committee, the study ultimately stopped enrolling patients after randomizing 226 patients between April 2007 and April 2013. Of those 226 patients, 114 were randomized to intervention and 109 to medical management. At the time of analysis, 53 of the treatment patients had not completed therapy and 20 had not initiated therapy. Seven patients randomized to medical management ultimately crossed over to the intervention arm, although no hemorrhages were identified in these patients. Interventional therapy included any neurosurgical, endovascular, or radiotherapy procedure (single or multiple) deemed appropriate by local ARUBA investigators to achieve complete bAVM obliteration. In the intervention group, 5 were treated with neurosurgical procedures, 30 with embolization, 31 with radiotherapy, and 28 with multimodality therapy. The average followup was 33 months and was balanced between the intervention and medical management arms. The primary endpoint was achieved in 35 (30.7%) of the 114 patients randomized to treatment and in
E96 | VOLUME 75 | NUMBER 1 | JULY 2014
Correspondence
11 (10.1%) of the 109 patients in the medical management arm. Secondary outcome was achieved in 24 (46.2%) patients receiving intervention and in 8 (15.1%) of the medically managed patients at 30 months. At 36 months, the secondary outcome was achieved by 17 (38.6%) of the treatment patients and 6 (14%) of the patients in the medical group. The authors conclude that medical management is superior to intervention in preventing death or stroke in patients with unruptured bAVMs followed for 33 months. The authors of the ARUBA trial aimed to address a valid question and their efforts are to be commended. However, the study design and execution have serious flaws that undermine the validity of their conclusions. To start, the authors fail to explain why such a large number of eligible patients, who did not refuse enrollment, were not randomized. Of the 726 eligible patients, 323 refused and 226 were enrolled. The other 177 patients were managed outside of the randomization process. There is no discussion regarding the characteristics of these patients and their lesions. The authors do acknowledge that potentially useful outcome data could be obtained from these patients. However, they briefly explain that participating centers did not respond to offers to create such a registry. Allowing treatment of eligible and consenting patients outside of the protocol is a critical flaw in design, introducing a layer of selection bias that need not complicate this study. Additionally, the study design lacks any standardization of the treatment arm. A recent large meta-analysis of treated bAVMs reported obliteration rates of 96% for surgical resection, 38% for stereotactic radiosurgery, and 13% for embolization.2 This analysis included ruptured and unruptured AVMs as well as a larger proportion of higher Spetzler-Martin grade AVMs. The neurosurgical literature supports these findings, and current recommendations, based on pooled-analysis, are surgical resection as the primary therapy for Spetzler-Martin grade I and II bAVMs.3 Without further detail regarding the treatment arm, it remains difficult to resolve why only 5 patients received surgical resection alone when 76 patients in the treatment arm had grade I or II bAVMs. The authors also provide no discussion regarding the treatment details of the radiotherapy and embolization groups. What type of radiotherapy was used and with what doses? Were serial or single embolizations performed and why? Serial embolizations certainly increase the risk of adverse events compared with single therapeutic options. Additionally, embolization is generally not considered a curative procedure, as evidenced by the previously mentioned obliteration rate of 13%, and is mainly used in combination therapies. It is concerning that 30 of the treatment patients were managed with embolization alone. The authors attempt to justify the very high rate of hemorrhage or stroke in the interventional therapy group (30.7%) by citing
www.neurosurgery-online.com
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
CORRESPONDENCE
recently published complication rates in a large systematic metaanalysis. The authors quote complication rates of 29% for surgery, 25% for embolization, and 13% for radiosurgery.2 The authors do mention that the meta-analysis includes lesions that have hemorrhaged in addition to those that have not. Unfortunately, they do not explore the population that they are making a comparison with any further. First, the quoted complication rates include all complications. The actual meta-analysis separates out severe complications, which include death and permanent neurological deficit. Remembering that the ARUBA primary outcome was death or symptomatic stroke and the secondary outcome was clinical impairment with a modified Rankin Scale score of 2 or higher, it makes more sense to compare the ARUBA outcomes with the severe complication rates in the meta-analysis. The rates for severe complications were 7.4% (range, 0%-40%) for surgery, 6.6% (range, 0%-28%) for embolization, and 5.1% (range, 0%-21%) for radiosurgery. All of these rates are significantly below what is presented in the ARUBA trial. Second, 84% of the patients, 11,521 of 13,698 included in the meta-analysis, presented with a ruptured AVM. Clearly patients presenting with cerebral hemorrhage are a very different population from that of the ARUBA trial. Last, in the metaanalysis referenced by the ARUBA authors, 61% of the patients had Spetzler-Martin grades equal to or less than 3. At time of analysis, 93% of the ARUBA trial patients randomized to interventional therapy had Spetzler-Martin grades equal to or less than 3. The Spetzler-Martin grading system has been well validated in the literature as a predictor of surgical morbidity and mortality. Given that the ARUBA trial included a much greater proportion of low-grade lesions, outcomes are expected to be better compared with the meta-analysis presented by the authors. On closer examination, the authors attempting to support their findings with previously published literature is a thinly veiled misrepresentation of their study population. There is reference to evaluation of lesion eradication by cerebral angiography in the Methods section. However, the authors do not include data regarding eradication in the article. The temporal relationship between treatment and achievement of primary and secondary outcomes is important when considering radiotherapy or serial embolization. Radiotherapy has a significantly delayed therapeutic effect. Serial embolizations will not provide therapeutic effect until complete eradication is achieved. Of the 114 patients randomized to treatment, 53 had not completed therapy and 20 had not initiated therapy at the time of analysis. Published literature has shown that there is no significant protection from the risk of hemorrhage provided by partial treatment of an AVM.4-7 Any events in the treatment arm that occurred before eradication should be noted by the authors. Given the lack of information and heterogeneity regarding therapies, it is impossible to understand what the treatment arm
NEUROSURGERY
represents. That said, it is unclear to what the authors are concluding medical management is superior. It seems irresponsible to conclude superiority of medical management for bAVMs when the treatment arm provides therapies that are seemingly well below the current standard of care for eradication. The length of follow-up is the most obvious pitfall of this analysis. The 33-month follow-up for a disease with a lifelong risk of stroke and hemorrhage is clearly inadequate to make conclusions. The authors' proposed follow-up of 5 years is also comparatively short. Treatment of bAVMs compresses the risk associated with these lesions. It may require more than the planned follow-up period to demonstrate the benefit of eradicating these lesions. The ARUBA trial is the first randomized, controlled trial attempting to evaluate intervention vs medical management for unruptured bAVMs. Unfortunately, errors in study design and execution as well as a relative lack of information regarding the treatment arm and the enrollment process invalidate the authors' conclusions. Indeed, the fact that in the medically managed group there was a 10% incidence of death or stroke over a time interval of less than 3 years emphasizes the need to provide a lowrisk cure for these patients. Disclosure The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.
Jonathan Russin Robert Spetzler Phoenix, Arizona 1. Mohr JP, Parides MK, Stapf C, et al. Medical management with or without interventional therapy for unruptured brain arteriovenous malformations (ARUBA): a multicentre, non-blinded, randomised trial. Lancet. 2013;383 (9917):614-621. 2. van Beijnum J, van der Worp HB, Buis DR, et al. Treatment of brain arteriovenous malformations: a systematic review and meta-analysis. JAMA. 2011;306(18):2011-2019. 3. Spetzler RF, Ponce FA. A 3-tier classification of cerebral arteriovenous malformations. Clinical article. J Neurosurg. 2011;114(3):842-849. 4. Lunsford LD, Kondziolka D, Flickinger JC, et al. Stereotactic radiosurgery for arteriovenous malformations of the brain. J Neurosurg. 1991;75(4):512-524. 5. Karlsson B, Lax I, Söderman M. Risk for hemorrhage during the 2-year latency period following gamma knife radiosurgery for arteriovenous malformations. Int J Radiat Oncol Biol Phys. 2001;49(4):1045-1051. 6. Miyamoto S, Hashimoto N, Nagata I, et al. Posttreatment sequelae of palliatively treated cerebral arteriovenous malformations. Neurosurgery. 2000;46(3):589-594; discussion 594-585. 7. Fournier D, TerBrugge KG, Willinsky R, Lasjaunias P, Montanera W. Endovascular treatment of intracerebral arteriovenous malformations: experience in 49 cases. J Neurosurg. 1991;75(2):228-233. 10.1227/NEU.0000000000000357
VOLUME 75 | NUMBER 1 | JULY 2014 | E97
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
CORRESPONDENCE
ERRATUM Response to Journal Club: Magnetic Resonance Imaging-Guided Focused Laser Interstitial Thermal Therapy for Intracranial Lesions: Single-Institution Series. Neurosurgery. 2014;74(5):565. The first sentence should have read: “We thank Dr Rohatgi and colleagues for their excellent Journal Club submission and appreciate their favorable assessment of our experience in laser interstitial thermal therapy for intracranial lesions.” DOI: 10.1227/01.neu.0000450985.53210.72 Copyright © by the Congress of Neurological Surgeons
E98 | VOLUME 75 | NUMBER 1 | JULY 2014
www.neurosurgery-online.com
Copyright © Congress of Neurological Surgeons. Unauthorized reproduction of this article is prohibited.
