7  Management of Unruptured Brain Arteriovenous Malformations

Thomas J. Sorenson and Giuseppe Lanzino

Abstract

The management of unruptured arteriovenous malformations (AVMs) is controversial. Whereas the average risk of rupture is approximately 2% per year, hemorrhage from ruptured AVM carries a significant morbidity and mortality. However, treatment of brain AVMs is not without the risk of severe complications. In the past few years, the publication of the ARUBA study has further increased the controversy surrounding the best treatment strategy for unruptured AVMs. The ARUBA trial notwithstanding, we feel that invasive treatment is indicated for young patients if the risk of treatment is acceptably low. In this chapter, we review current knowledge on the natural history of unruptured AVMs, the results of the ARUBA study, and our current approach to a patient with an unbled AVM.

Keywords: arteriovenous malformations, natural history, ARUBA study, hemorrhage, stereotactic radiosurgery, microsurgery, embolization

7.1  Introduction

Brain arteriovenous malformations (AVMs) are relatively uncommon vascular lesions that have traditionally been detected after hemorrhage, with notable long-term morbidity and case fatality. However, with widespread utilization of neuroimaging, AVMs are increasingly diagnosed incidentally or in the presence of minor neurological symptoms without hemorrhage.¹ This has created a paradigm shift and, today, most patients evaluated with an AVM do not have clinical history of hemorrhage. In such cases, the treating physician is faced with the dilemma of submitting the patients to the risk of invasive treatment or submitting the patients to a small, definite risk of hemorrhage but leaving the AVM untreated, subjected to natural history. In recent years, the controversy about the best therapeutic modality in patients with unruptured AVMs has been heightened by the results of A Randomized Trial of Unruptured Brain Arteriovenous Malformations (ARUBA) study. In this chapter, we will review current knowledge on the natural history of untreated, unruptured AVMs, the results of the ARUBA study, the results of further investigations, and life after ARUBA.

7.2  Natural History of Unruptured AVMs

Sound data about the natural history of untreated, unruptured AVMs are limited and, for the majority, based on retrospective, clinical series. The main challenge in obtaining high-quality data to base clinical decision upon is the low incidence of the end point (rupture), the need for long-term (several years) follow-up, and the fact that unruptured AVMs are a very heterogeneous group in relation to their location, mode of presentation, size, and angioarchitecture. The issue is further complicated by the observation that an important subset of patients without a clinical history of hemorrhage demonstrates either radiological or histological findings consistent with prior clinically silent hemorrhage.²

For a long time, the best data available on the natural history of AVMs were based on a large, single-center Finnish cohort, initially analyzed by Ondra et al³ and more recently updated by Hernesniemi et al.⁴ Of 631 AVM patients admitted at Helsinki University Central Hospital between 1942 and 2005, 238 were medically managed and considered for analysis. Preadmission AVM rupture occurred in 139 patients. Average follow-up time was 13.5 years (range, 1 month to 53.1 years). A total of 77 patients experienced AVM hemorrhage during 3,222 person-years, which represents a 2.4% annual rupture rate.

The main strength of this cohort is related to the long-term follow-up and the fact that the Finnish population was very stable during the period under study, making possible to account for every patient, even many years after the initial diagnosis. The limitations of this cohort are that patients with prior history of hemorrhage are not excluded when calculating the hemorrhage rate, which could be slightly misleading. We were also not able to recalculate the hemorrhage rate as the natural history exclusively of patients who did present with hemorrhage was not defined. Additionally, limitations are related to the fact that the cohort under study spanned many years, during which there have been dramatic changes in diagnostic tools and therapeutic approach to these lesions. These limitations notwithstanding, data from this cohort suggested that the yearly risk of hemorrhage was 2.4% per patient per year. More definitive and contemporary data on the natural history of patients with unruptured AVMs come from the ARUBA study, which showed that for all-comers with unruptured AVMs, the annual rate of bleeding was similar, at 2% per year.

7.3  Outcome after First Hemorrhage

Another important issue in assessing the risk of treatment versus the natural history is to understand the clinical outcome after first hemorrhage of an AVM. After an AVM-associated hemorrhage (AVM-ICH), 72% of patients with AVM have a good functional outcome (mRS < 2).⁵ The outcome from hemorrhage in a patient with an AVM is much more favorable than patients who suffer a spontaneous ICH (not secondary to AVM rupture). Patients with AVM-ICH are younger and have lower blood pressure and better Glasgow Coma Scores than patients with spontaneous hemorrhage (sICH).⁶ Case fatality at 1 month was 11% for those with AVM-ICH.

However, recent population studies have shown that overall prognosis after AVM rupture, though better than sICH, is not quite as benign as single-center studies had suggested. In the Scottish study, the case fatality rate after AVM-ICH was 11% at 1 month and 13% at 2 years. The percentage of patients who were dead or dependent (mRS > 3) at 1 year was 40%. Case fatality rate is age dependent and was 9.2% at 1 month in patients younger than 60 years and 21% in those older than 60 years.⁷

To select those patients at higher risk of hemorrhage from an originally unruptured AVM, several publications have analyzed factors predictive of higher risk of hemorrhage. The main limitations of these studies are that these are usually retrospective and the analysis of the risk factors (especially those related to angioarchitecture) is retrospective in patients who have already presented with hemorrhage, which does not always allow for an accurate understanding of prehemorrhage angioarchitecture. Considering these important methodological limitations, there are still some factors that are likely associated with a higher risk of bleeding. Most of these factors are related to the angioarchitecture and include (1) presence of venous varices and evidence of outflow obstruction of the drainage pathways; (2) presence of intranidal aneurysms, which indicate areas of hemodynamic stress within the nidus; and (3) presence of feeding pedicle aneurysms.

Though controversy exists about the predictive ability of size, location, and patient age to increased risk of rupture, it is our feeling that mode and symptoms of presentation are most important in trying to understand if a patient might bleed. In other words, there are some situations where the onset of symptoms might indicate that hemodynamic factors might have acutely changed in an AVM, thus causing the symptoms. Therefore, we think that it is not only important to analyze patient symptoms and reason of AVM diagnoses, but also important to make careful analysis of risk factors associated with AVM in question. It is also our belief that for unruptured AVMs, the risk of rupture is highest soon after diagnosis, and then starts decreasing if no hemorrhage occurs during a period of a few years. This observation is corroborated by the Finnish cohort in which it was observed that the risk of bleeding was much higher during the first 5 years after diagnosis than thereafter.

As mentioned above, anecdotal observations and recent data have suggested that an important subset of patients without clinical history of hemorrhage has radiological or histological findings consistent with a silent bleed. Whether these patients are at increased risk of bleeding than patients without such history is a matter of debate. Like other asymptomatic, often prevalent cerebrovascular disorders, such as intracranial aneurysms, cavernous malformations, and carotid stenosis, not all unruptured AVMs are similar, and it will be important in the future to try to identify those patients at higher risk of bleeding who might be more likely to benefit from treatment and for whom the risk of invasive treatment might be justified.

7.4  ARUBA (A Randomized Trial of Unruptured Brain Arteriovenous Malformations) Results and Critiques

ARUBA was the first randomized trial that attempted to determine the best approach for management of unruptured brain AVMs, specifically whether the medical management or interventional therapy posed a greater risk to the patient. Given the risks associated with invasive treatment, the ARUBA investigators aimed to answer the questions of whether invasive treatment of unruptured AVMs is indeed better than conservative management in patients with unruptured AVMs.

The primary end point measured in the trial was time to death or symptomatic stroke. Symptomatic stroke was defined as onset of neurological deficit or headache with corresponding neuroimaging of new hemorrhage or infarction. The secondary end point measured was clinical impairment at 5 years (mRS score ≥ 2).

The authors screened 1,514 patients across 39 active clinical sites in 9 countries and found 1,014 patients ineligible, 323 refusing participation, and 42 unable to participate. In total, 226 patients were randomized into two cohorts: interventional therapy (n = 116) or conservative management (n = 110). Based on modern clinical practice, three interventional therapies were utilized alone or in a multimodal approach: neurosurgery alone (n = 5), embolization alone (n = 30), stereotactic radiosurgery alone (n= 31), embolization and neurosurgery (n=12), embolization and radiosurgery (n = 15), or embolization, neurosurgery, and radiosurgery (n = 1). Following randomization, median time to first intervention was 76 days. There were no statistically significant differences between the two cohort characteristics at randomization or treatment. Seven patients crossed over from conservative management to interventional therapy, while three patients assigned to interventional therapy suffered an outcome event between randomization and intervention and crossed over to conservative management. Seven patients discontinued their participation in the trial at follow-up (average: 33 months). At randomization, all patients had Spetzler-Martin (SM) grade scores < 4 and most (62%) were ≤ 2 and all patients had mRS score ≤ 1. Patient enrollment was halted by the NINDS/NIH at the recommendation of the data and monitoring board.

During follow-up, a primary event occurred in 10 (8%) patients in the conservative management cohort versus 36 (36.7%) patients in the interventional therapy cohort (► Table 7.1). This significant difference between cohorts held true for the secondary event (clinical impairment) as well, with 8 (15.1%) patients from conservative management cohort versus 24 (46.2%) patients from interventional therapy cohort having an mRS score ≥ 2 at 30 weeks’ follow-up.⁸

Related posts:

Surgical Approaches and Nuances for Lobar Arteriovenous Malformations Vein of Galen Aneurysmal Malformation Classification of Brain Arteriovenous Malformations and Fistulas Natural History of Arteriovenous Fistulas, Clinical Presentation, and Indications for Treatment Surgery of Basal Ganglia, Thalamic, and Brainstem Arteriovenous Malformations Surgical Treatment of Cerebellar Arteriovenous Malformations

Stay updated, free articles. Join our Telegram channel

Join