Arterial aneurysms associated with cerebral arteriovenous malformations: classification, incidence, and risk of hemorrhage

Full access

Object. The goal of this study was to develop a classification system for aneurysms associated with arteriovenous malformations (AVMs) based on their anatomical and pathophysiological relationships and to determine the incidence and bleeding rates for these aneurysms as well as the effects of AVM treatment on their natural history.

Methods. Of 632 patients with AVMs, intranidal aneurysms were found in 35 (5.5%) and flow-related aneurysms in 71 (11.2%). Patients with intranidal aneurysms presented more frequently with hemorrhage (72% compared with 40%, p < 0.001) and had a 9.8% per year risk rate of bleeding during follow-up review. Twelve (17%) of the patients with flow-related aneurysms associated with an AVM presented with hemorrhage from an aneurysm, whereas 15 (21%) bled from their AVM. Seventeen patients underwent angiography after AVM treatment (mean 2.25 years). Of 23 proximal aneurysms, 18 (78.3%) were unchanged, four (17.4%) were smaller, and one (4.3%) had disappeared, whereas four (80%) of five distal aneurysms regressed completely and one was unchanged. Sixteen patients underwent angiography after partial AVM treatment (mean 3.8 years). In cases with less than a 50% reduction in the AVM, no aneurysms regressed, although two enlarged and bled. In cases with greater than a 50% reduction in the AVM, two of three distal aneurysms disappeared and five proximal aneurysms were unchanged.

Conclusions. Arterial aneurysms associated with cerebral AVMs may be classified as intranidal, flow-related, or unrelated to the AVM nidus. Intranidal aneurysms have a high correlation with hemorrhagic clinical presentation and a risk of bleeding during the follow-up period that considerably exceeds that which would be expected in their absence. Patients with flow-related aneurysms in association with an AVM may present with hemorrhage from either lesion. Aneurysms that arise on distal feeding arteries near the nidus have a high probability of regressing with substantial or curative AVM therapy.

Abstract

Object. The goal of this study was to develop a classification system for aneurysms associated with arteriovenous malformations (AVMs) based on their anatomical and pathophysiological relationships and to determine the incidence and bleeding rates for these aneurysms as well as the effects of AVM treatment on their natural history.

Methods. Of 632 patients with AVMs, intranidal aneurysms were found in 35 (5.5%) and flow-related aneurysms in 71 (11.2%). Patients with intranidal aneurysms presented more frequently with hemorrhage (72% compared with 40%, p < 0.001) and had a 9.8% per year risk rate of bleeding during follow-up review. Twelve (17%) of the patients with flow-related aneurysms associated with an AVM presented with hemorrhage from an aneurysm, whereas 15 (21%) bled from their AVM. Seventeen patients underwent angiography after AVM treatment (mean 2.25 years). Of 23 proximal aneurysms, 18 (78.3%) were unchanged, four (17.4%) were smaller, and one (4.3%) had disappeared, whereas four (80%) of five distal aneurysms regressed completely and one was unchanged. Sixteen patients underwent angiography after partial AVM treatment (mean 3.8 years). In cases with less than a 50% reduction in the AVM, no aneurysms regressed, although two enlarged and bled. In cases with greater than a 50% reduction in the AVM, two of three distal aneurysms disappeared and five proximal aneurysms were unchanged.

Conclusions. Arterial aneurysms associated with cerebral AVMs may be classified as intranidal, flow-related, or unrelated to the AVM nidus. Intranidal aneurysms have a high correlation with hemorrhagic clinical presentation and a risk of bleeding during the follow-up period that considerably exceeds that which would be expected in their absence. Patients with flow-related aneurysms in association with an AVM may present with hemorrhage from either lesion. Aneurysms that arise on distal feeding arteries near the nidus have a high probability of regressing with substantial or curative AVM therapy.

The association of arteriovenous malformations (AVMs) with intracranial aneurysms has been well documented in numerous reports.1,2,4–7,13,15,18,26,27 It is generally believed that aneurysms lead to a higher risk of hemorrhage as the initial presentation and also as part of the clinical course and natural history of patients with this combination of lesions.2,4,16,17,21,28,29 Aneurysms in a patient with an AVM may occur on unrelated vessels and be found coincidentally or they may be associated with the AVM on a pathophysiological basis due to an underlying vascular defect or as a result of the dynamic interaction between hemodynamic stimuli, vasoactive substances, and locally generated growth factors leading to structural and functional alterations through the process of vascular remodeling.10,14,19,23,30

There have been several attempts to categorize AVM-related aneurysms,4,5,9,15,17,21,24,26,27 but a widely accepted system of classification based on their anatomical and pathophysiological relationship to the AVM has yet to be developed and validated. Ideally, such a system of classification would have predictive value with respect to the risk of future hemorrhage, as well as to the potential impact of hemodynamic changes resulting from AVM treatment on associated saccular aneurysms. Most publications have dealt with descriptive aspects of these associated lesions, whereas less emphasis has been placed on appropriate surgical and endovascular management priorities.

An imperfect understanding of the relationship between AVMs and associated aneurysms has led to widely varying treatment strategies. Concern that abrupt elimination of an AVM might put aneurysms located along feeding arteries at immediate risk of distention and subsequent rupture has led some to recommend treating the aneurysm before the AVM.2 Alternatively, the reduction of flow through feeding arteries following AVM elimination has prompted others to recommend elimination of the AVM first, with the expectation that resulting hemodynamic alterations may lead to diminution or complete regression of related aneurysms.15 In an effort to clarify some of these issues, we have reviewed our experience with a relatively large, unselected AVM patient population.

Clinical Material and Methods
Patient Population

Six hundred thirty-two patients with angiographically verified cerebral AVMs were evaluated by the University of Toronto Brain Vascular Malformation Study Group between 1987 and 1997. Demographic features; clinical presentation; size, location, and angioarchitectural features of the AVM; presence or absence of intracranial aneurysms; therapeutic interventions; results of treatment; subsequent hemorrhages; and dates and results of follow-up angiography were recorded for each patient. All data were prospectively entered into a computer database beginning in 1989. Data accrued prior to that time were entered into the database retrospectively. Permanent x-ray files containing relevant cross-sectional and angiographic imaging studies were available for each patient. All patients with an angiographically confirmed arterial aneurysm were included in this analysis.

Classification of Aneurysms

Aneurysms were categorized as intranidal, flow-related, or unrelated to the AVM. Aneurysms were designated intranidal if they filled early during angiography, before substantial venous filling had occurred, and if they were localized within the boundaries of the AVM nidus (Fig. 1). Simple arterial ectasias, infundibulae, venous pouches, and variceal dilations were excluded. Saccular aneurysms arising along the course of arteries that eventually supplied the AVM were classified as flow related. Flow-related aneurysms were subclassified as proximal if they were located on the supraclinoid internal carotid artery (ICA), the circle of Willis, the middle cerebral artery (MCA) up to and including the primary bifurcation, the anterior cerebral artery (ACA) up to and including the anterior communicating artery (ACoA), or the vertebrobasilar trunk. All flow-related aneurysms beyond these locations were subclassified as distal (Fig. 2). Aneurysms that occurred on arteries with no supply to the AVM were classified as unrelated (Table 1).

Fig. 1.
Fig. 1.

Left: Right ICA angiogram, arterial phase, demonstrating an intranidal aneurysm (arrow) within an AVM supplied by MCA branches and the anterior choroidal artery. Right: Superselective angiogram in which the intranidal aneurysm (curved arrow) is better delineated, with the tip of a microcatheter (long arrow) positioned within the distal anterior choroidal artery.

Fig. 2.
Fig. 2.

Angiograms. Left: Occipital AVM supplied by the posterior cerebral artery (PCA). A proximal flow-related aneurysm is present at the basilar artery bifurcation. Right: Small posterior temporal AVM (short arrow) with supply from the distal PCA. A distal flow-related aneurysm is present on the feeding artery (long arrow), with the aneurysm fundus superimposed on an anteriorly directed draining vein.

TABLE 1

Categories of arterial aneurysms associated with AVMs

TypeAneurysm Features
intranidalw/in AVM nidus, fills early during angiography
flow-related
 proximalproximal on major artery w/ eventual supply to AVM*
 distaldistal on AVM feeding artery
unrelatedon artery unrelated to AVM supply

Located on ICA, circle of Willis, MCA (M1 segment or bifurcation), or vertebrobasilar trunk (see Classification of Aneurysms).

Results
Patient Demographics

Within the entire AVM patient population, 97 patients (15.3%) were found to have at least one aneurysm. There were 52 males (54%) and 45 females (46%) in this group with an age at the time of clinical presentation ranging from 8 to 76 years (mean 37 years). The period of follow up from the initial clinical presentation was calculated as beginning at the time of the first documented hemorrhagic event or diagnostic angiogram. The mean duration of follow up from the time of clinical presentation for all 97 patients was 7.4 years (range 0–37 years, median 4.5 years). The length of the prospective follow-up period, which commenced at the first clinical evaluation in our institution, averaged 1.8 years (range 0–10 years).

Angiographic Results

Intranidal aneurysms were identified in 35 patients (5.5% of the entire AVM patient population), flow-related aneurysms in 71 (11.2%), and unrelated aneurysms in five patients (0.8%). Fifty-seven patients had only flow-related aneurysms; 22 had only intranidal aneurysms; 13 had both flow-related and intranidal aneurysms; four had only unrelated aneurysms; and one patient had both flow-related and unrelated aneurysms.

Superselective angiography with microcatheterization of feeding artery pedicles was performed in 257 patients (41%). Of the 97 patients with aneurysms, 57 (59%) underwent superselective angiography. Specific angioarchitectural features, including those indicating intranidal aneurysms, were more clearly delineated with superselective angiography. However, a comparison of the incidence of aneurysms identified on superselective and global angiograms is not possible because many patients underwent superselective catheterization during an embolization procedure which was recommended, in part, because of an aneurysm identified on the initial diagnostic angiogram. Patients undergoing superselective angiography were a select subpopulation with AVM angioarchitectural features that were believed to require endovascular therapy, thus creating a bias for any comparison with the entire AVM patient population.

Clinical Presentation

Hemorrhage was the most frequent clinical presentation in the entire AVM patient population (243 of 632 patients [38%]). The clinical presentation of the 97 patients with aneurysms is detailed in Table 2. A hemorrhagic presentation was significantly more frequent in the subgroup with intranidal aneurysms (72%) than the entire AVM population without aneurysms (36%) or the subgroup with only flow-related or unrelated aneurysms (40%) (chi-square test, p < 0.001). The presentation in five patients with unrelated aneurysms was hemorrhage from the AVM in three, subarachnoid hemorrhage (SAH) from aneurysm rupture in one, and third nerve paresis due to compression from a large posterior communicating artery aneurysm in one.

TABLE 2

Clinical presentation of 97 patients with aneurysms and AVMs

No. of Patients
PresentationW/ Intranidal Aneurysm (35 patients)W/O Intranidal Aneurysm (62 patients)
bleeding from AVM22 (63%) 13 (21%) 
bleeding from aneurysm3 (9%) 10 (16%) 
bleeding, unknown source 2 (3%) 
seizure7 (20%) 20 (32%) 
other3 (9%) 17 (27%) 

Flow-Related Aneurysms

There were 123 flow-related aneurysms in 71 patients. Eighty-four aneurysms were of the proximal type and 39 were of the distal type. The location and size of flow-related aneurysms are presented in Table 3. Twenty-nine patients (40.8%) presented with hemorrhage. The source of the hemorrhage was determined by the location and distribution of blood on cross-sectional imaging studies correlated with angiography and by findings at surgery. Twelve patients (41.4%) bled from an aneurysm and 15 (51.7%) from the AVM; in two patients (6.9%) with distal flow-related aneurysms the source of the hemorrhage could not be conclusively determined as arising from the aneurysm or the AVM nidus. Of the 12 flow-related aneurysms that presented with hemorrhage, five were distal (12.8% of 39 distal aneurysms) and seven were proximal (8.3% of 84 proximal aneurysms).

TABLE 3

Location of 123 flow-related aneurysms in 71 patients

LocationNo. of Aneurysms
proximal aneurysms84*
 ICA26
 MCA19
 ACoA15
 basilar bifurcation14
 P1–P2 junction5
 superior cerebellar artery3
 vertebrobasilar junction1
 anterior inferior cerebellar artery1
distal aneurysms39
 PCA13
 ACA/pericallosal artery9
 superior cerebellar artery8
 MCA7
 posterior inferior cerebellar artery2

Eighty small (<12 mm) and four large (12–25 mm) aneurysms.

Thirty-eight small and one large aneurysm.

No significant correlation was found between the size of the AVMs and the type, number, or size of the aneurysms with which they were associated. There was a trend for flow-related aneurysms to occur with greater frequency in larger AVMs. Aneurysms were identified in 31 (7.6%) of 408 small AVMs, 36 (18.4%) of 196 medium AVMs, and four (14.3%) of 28 large AVMs. The frequency of hemorrhagic presentation and the source of the hemorrhage based on AVM size are presented in Table 4.

TABLE 4

Hemorrhagic presentation of AVMs associated with flow-related aneurysms in 71 patients according to size of AVM nidus*

PresentationSmall (31 patients)Medium (36 patients)Large (4 patients)
bleeding from AVM582
bleeding from aneurysm1020
bleeding from unknown source200
total17102

Small = <3 cm; medium = 3–6 cm; large = >6 cm.

Effect of AVM Treatment on Flow-Related Aneurysms

Seventeen patients with 28 flow-related aneurysms underwent delayed follow-up angiography after AVM cure (mean 2.25 years, range 6 months–8 years). Aneurysms that were treated directly at the time of surgical or endovascular AVM treatment or located on segments of the parent vessel that underwent retrograde thrombosis were excluded from analysis. Of 23 proximal aneurysms, 18 (78.3%) were unchanged (Fig. 3), four (17.4%) were smaller, and one (4.3%) regressed completely. However, four (80%) of five distal aneurysms disappeared (Fig. 4), whereas only one (20%) was unchanged. There were no instances of SAH from a flow-related aneurysm after AVM cure in the follow-up period.

Fig. 3.
Fig. 3.

Angiograms showing a proximal flow-related aneurysm that remains unchanged despite AVM cure. Left: Left ICA angiogram showing extensive crossflow through the ACoA and an ACA aneurysm (arrow) associated with a large right-hemisphere AVM. Right: One year after staged embolization with complete AVM elimination, there is substantial diminution in the caliber of the left ICA and ACA, and there is no longer crossflow through the ACoA. The ACA aneurysm is unchanged.

Fig. 4.
Fig. 4.

Angiograms showing a distal flow-related aneurysm that regressed completely following AVM cure. Left: Right ICA angiogram demonstrating an enlarged frontal branch supplying an AVM (curved arrow) and a distal flow-related aneurysm at the bifurcation of the callosomarginal artery (straight arrow). Right: Following embolization, repeated angiography performed 6 months later reveals that the AVM has been eliminated and the aneurysm is no longer present (arrow).

Sixteen patients with 30 flow-related aneurysms underwent follow-up angiography after subtotal AVM therapy (mean 3.8 years, range 6 months–15 years). In 11 patients the size of the AVM nidus was reduced by less than 50%. There were 20 proximal and two distal aneurysms, none of which decreased on follow-up angiography. Two of the patients experienced SAH from proximal flow-related aneurysms during the follow-up period (Fig. 5). In five patients the AVM nidus was reduced by more than 50%. There were five proximal and three distal aneurysms. None of the proximal aneurysms regressed, whereas two (67%) of the distal aneurysms regressed completely (Fig. 6). There were no episodes of aneurysmal SAH during the follow-up period in patients whose AVM nidus was reduced by more than 50%.

Fig. 5.
Fig. 5.

Angiograms demonstrating a flow-related aneurysm that enlarged and bled after partial AVM therapy. Left: Vertebral artery angiogram showing a dorsal midbrain AVM with supply from numerous perforating branches arising from the upper basilar artery and the PCA. A small aneurysm is present at the origin of the left superior cerebellar artery (arrowhead). Right: Seven years after undergoing stereotactic radiosurgery with minimal change in the AVM size, the patient presented with SAH. Angiography demonstrated that the aneurysm had enlarged (arrow). The aneurysm was confirmed as the source of hemorrhage at the time of surgical clipping.

Fig. 6.
Fig. 6.

Angiograms displaying a distal flow-related aneurysm that regressed completely following subtotal AVM treatment. Left: Lateral vertebral artery angiogram showing a large occipital AVM with variceal dilation of a draining vein (arrowheads) superimposed over the perimesencephalic segment of the PCA. Center: Following the first stage of embolization there is substantial reduction of the AVM. Repeated angiography, early arterial phase, demonstrating a previously obscured aneurysm arising from the distal PCA (arrow). Right: Final angiographic result after the second stage of embolization with marked decrease in the size of the AVM. The aneurysm has disappeared (arrow).

Natural History of Intranidal and Flow-Related Aneurysms Without Treatment

Thirteen patients with intranidal aneurysms did not receive treatment for their AVM at the time of the initial diagnosis. The initial presentation was hemorrhage from the AVM in nine, seizures in three, and hydrocephalus in one case. Only one of the intranidal aneurysms had a “pseudoaneurysm” appearance.9 The average duration of the clinical follow-up period until the last clinic assessment or definitive AVM treatment for all patients was 11 years (range 1.5–24 years). During 143 patient-years of follow up there were 14 hemorrhages, resulting in an annual risk of bleeding of 9.8%. Nine patients had one hemorrhage during the follow-up period (including three of the four patients with nonhemorrhagic presentation) and one patient had five separate hemorrhages.

All patients with flow-related aneurysms who experienced aneurysmal bleeding underwent definitive surgical or endovascular aneurysm obliteration. There were only four patients with associated flow-related aneurysms who did not receive treatment for their aneurysms or AVMs. All four presented with seizures. The average follow-up period from the time of initial diagnosis was 14.25 years (range 7–19 years). During 57 patient-years of follow up, three patients experienced hemorrhage from their AVM, for an average risk of bleeding of 5.3% per year. None of these patients experienced aneurysm bleeding.

Discussion
Etiology of AVM-Associated Aneurysms

Although the anatomical and clinical features of aneurysms associated with AVMs have been the focus of several studies, little is understood about their pathogenesis. It is generally accepted that hemodynamic factors and a hyperdynamic circulatory state account, at least in part, for their occurrence.8,10,12,13,15,18 This theory is supported by the observation that the incidence and distribution of aneurysms on feeding arteries supplying AVMs are greatly in excess of what would be expected in the absence of the AVM20 and that some aneurysms regress after AVM removal.15,26

The vasculature is a dynamic structure, capable of sensing biochemical and hemodynamic alterations and of changing itself through the local production of mediators that influence both structure and function.10,14 A “high-flow angiopathy” consisting of intimal thickening and destruction, elastic degeneration, irregular thickening and thinning of vessel walls, and alteration of the muscular layer has been described in a rabbit model of a high-flow arteriovenous shunt.23 However, given that flow-related and intranidal aneurysms occur in association with AVMs of all sizes and flow rates, yet only in a minority of all AVMs, their development must be the result of complex interactions of host-specific and hemodynamic factors.14,19,30

The flow-related saccular aneurysms that we and others have found to be associated with cerebral AVMs cannot be distinguished on the basis of their angiographic or histological features from those aneurysms that occur in the absence of any other vascular malformation.1,2,4,20,22 The question may be raised as to whether intranidal aneurysms or distal flow-related aneurysms located near the AVM nidus identified on angiography following a hemorrhage represent true aneurysms that were present prior to bleeding or “pseudoaneurysms” developing from the rupture of a thin-walled vessel.9,21 This issue is of some importance, because these aneurysms are believed to increase the risk of hemorrhage from AVMs significantly, and their identification may lead to a greater sense of urgency with respect to AVM treatment.17 However, if a substantial proportion of intranidal aneurysms are really pseudoaneurysms that arise as a result of hemorrhage, then it is inappropriate to extrapolate a high risk of bleeding to intranidal aneurysms found on AVMs that have not bled.

Pseudoaneurysms may be suspected if they appear on angiography or magnetic resonance imaging as vascular cavities of irregular shape within or at the margins of a hematoma. These aneurysms can only be confirmed as pseudoaneurysms on the basis of histological examination or if a comparison with previous vascular examinations confirms the aneurysm as a new angioarchitectural feature. This acquired nature secondary to a hemorrhagic event is pathognomonic. Garcia-Monaco, et al.,9 reported finding pseudoaneurysms on angiography in 15 cases (8%) in a population of 189 patients with AVMs. None of the pseudoaneurysms was confirmed histologically. Five were treated with embolization, one was treated surgically, and eight of the remaining nine pseudoaneurysms resolved on follow-up angiography. Marks and colleagues17 reported on 15 patients with intranidal aneurysms, all of whom had a history of bleeding. Three patients underwent surgical excision, and in two cases the aneurysms could be located in the pathological specimens obtained. Histological evaluation demonstrated these aneurysms as thin-walled vascular structures rather than pseudoaneurysms secondary to previous hemorrhage.

In our study population there were five patients with intranidal aneurysms or distal flow-related aneurysms located near the AVM nidus who presented with hemorrhage and were found to have a hematoma surrounding the aneurysm on cross-sectional imaging. None of these patients underwent angiography prior to their hemorrhage and none of the aneurysms was available for histological study, but their angiographic and cross-sectional imaging features were compatible with a diagnosis of pseudoaneurysm. However, the issue of diagnosis of pseudoaneurysm on imaging studies alone is thrown into question by the fact that two other patients in our series had large, irregular intranidal aneurysms demonstrated on angiography in the absence of a history of bleeding and without signal abnormalities on magnetic resonance imaging that would suggest remote subclinical hemorrhage.

Incidence of Aneurysms in Association With AVMs

Data on the prevalence of unruptured saccular aneurysms vary considerably according to study design, study population, and aneurysm characteristics. It was concluded in a recent systematic review of autopsy and angiographic studies on this issue that saccular aneurysms are found in approximately 2% of the adult population. The vast majority of these aneurysms are small and have an annual risk of rupture of less than 1%.25

The incidence of aneurysms associated with AVMs reported in the literature has varied widely, with most series reporting between 10% and 20%, up to as high as 58%.3,11,15,18,22,26,27,30 Lack of a uniform system of classification and nomenclature has prevented direct comparison between series. Anderson and Blackwood1 reported autopsy findings in a series of nine patients with AVMs. In five cases (55.6%) there were associated aneurysms, all of which were localized to the circle of Willis or the MCA. No intranidal aneurysms were described. The total number of patients with AVMs treated at their institution during the same time period was not mentioned; therefore, the incidence of associated aneurysms cannot be extrapolated to the general AVM patient population.

Turjman and associates27 reported a series of 100 consecutive patients who underwent preembolization superselective angiography. In this highly selected population, aneurysms were identified in 58%, a figure approximating the results of the autopsy study by Anderson and Blackwood.1 However, 70% of the aneurysms were intranidal or arose from perforating vessels. Caution must be exercised in the interpretation of studies in which extremely high incidences of intranidal aneurysms are reported, because their angiographic diagnosis can be rather subjective and viewer dependent, in comparison with the diagnosis of saccular aneurysms arising from proximal vessels in relation to the circle of Willis or along feeding artery pedicles.

We found intranidal aneurysms in 5.5% of our entire patient population. This figure is substantially lower than that reported by Turjman and associates,27 reflecting a lack of selection bias for our patient population as a whole, and perhaps a more conservative interpretation of angiographic findings. The significantly increased incidence of hemorrhagic presentation in our patients with intranidal aneurysms compared with those without them suggests that we have identified a subpopulation with unique angioarchitectural and clinical features.

In this study, saccular aneurysms on arteries unrelated to the AVM were found in five patients (0.8%). One presented with SAH and one was symptomatic, causing a third nerve palsy. We have no reason to believe that the prevalence or risk of rupture of incidental, unrelated aneurysms in patients with AVMs is different from their counterparts without AVMs.25 Flow-related aneurysms were found in 11.2% of all patients with AVMs, confirming that the incidence of saccular aneurysms is much higher in the presence of an AVM than in the general population. The fact that almost 90% of patients with AVMs did not have a flow-related aneurysm suggests that augmentation of flow alone is insufficient to cause their development.

Risk of Hemorrhage

In this study, 29 (40.8%) of 71 patients with flow-related aneurysms presented with hemorrhage. The incidence of hemorrhage as the initial presentation was not significantly different from those patients without aneurysms, but it is interesting to note that the source of the hemorrhages was almost equally divided between an aneurysm and the AVM. This distribution is similar to the experience in the Cooperative Study of Intracranial Aneurysms and Subarachnoid Hemorrhage22 and that reported by Graf, et al.,11 whereas Batjer, et. al., found that 78% of hemorrhagic events in a similar patient population were due to the aneurysm. In their series, all of the patients who suffered aneurysmal hemorrhage bled from atypical distal aneurysms. Our experience differs in that the incidence of hemorrhage from proximal aneurysms (8.3%) and distal aneurysms (12.8%) was similar.

Because hemorrhage is the most serious risk for the patient with an AVM, it is important to identify characteristics that may be used to predict which patients are at the highest risk of bleeding. These patients should be expected to derive the greatest benefit from therapeutic intervention. In addition to the presence of intranidal aneurysms, several angioarchitectural features of AVMs have been found to correlate with a clinical presentation of hemorrhage, including deep venous drainage, venous aneurysms or outflow compromise, feeding by perforating vessels, and deep or periventricular location.3,12,16,24,28 Patients with small AVMs present more frequently with hemorrhage, likely reflecting a lower incidence of seizures or progressive neurological deficits in comparison with large AVMs, rather than an intrinsic predisposition to bleed.3,11,22

With respect to predicting future hemorrhage, this study is limited by the fact that many patients who presented with hemorrhage or angiographic features believed to be associated with a high likelihood of bleeding, particularly intranidal aneurysms, were treated without undergoing a period of follow-up observation. Given the morbidity associated with AVM hemorrhage, in these instances it was not possible to withhold therapy, such as surgery or embolization, that permanently altered the angioarchitecture of these lesions and, presumably, their natural history risks. However, the 9.8%-per-year risk of AVM hemorrhage in the small number of untreated patients with intranidal aneurysms is considerably higher than the generally reported figures of 2 to 4% observed in other series.3,11,22

The risk of future hemorrhage in patients with unruptured AVMs and unruptured aneurysms was studied by Brown and associates.4 Among 91 patients with unruptured AVMs seen at the Mayo Clinic, 16 (17.5%) had unruptured intracranial saccular aneurysms. During follow-up review, intracranial hemorrhage occurred in six (38%) of the 16 patients with aneurysms and AVMs. The source of the hemorrhage was the AVM in one patient, the aneurysm in one, and in four it could not be determined. The risk of intracranial hemorrhage among the 16 patients with coexisting aneurysm and AVM was 7% at 1 year, compared with 3% for those with AVM alone. At 5 years, the risk persisted at 7% per year, whereas it decreased to 1.7% per year in those with an AVM unassociated with aneurysm.

We only had four patients with unruptured flow-related aneurysms and unruptured AVMs who were followed prospectively without treatment. In this small group the bleeding rate was 5.3% per year, with all hemorrhages arising from the AVM. In those patients with unruptured, untreated flow-related aneurysms and AVMs that were either untreated or only partially treated, there were 20 patients with 118 patient years of follow up. There were two aneurysmal hemorrhages, resulting in an annual rate of aneurysm rupture of 1.7%. Therefore, in patients with flow-related aneurysms and AVMs, if left untreated, the risk of bleeding seems to be greater for the AVM than for the aneurysm, although hemorrhage from either source is possible. The total combined rate of bleeding from aneurysms and AVMs during follow up in this small group of patients is approximately 7%, which is consistent with the findings of Brown, et al.4

Flow-Related Aneurysms: Effect of AVM Treatment

The need for treatment of flow-related aneurysms that are discovered incidentally in association with AVMs is a matter of some controversy. Specific intervention in all cases may be unnecessary, because there are numerous reports of spontaneous regression following treatment of the AVM.5,13,15,26 However, the impact of partial or complete AVM obliteration on the natural history of these associated aneurysms remains unclear.

This study is the first in which the angiographic and clinical effects of AVM treatment on flow-related aneurysms was evaluated systematically in a prospective fashion. The subclassification of flow-related aneurysms into distal and proximal types appears to be useful in predicting the response of the aneurysm to AVM therapy. Distal aneurysms had a high probability of regression with substantial or curative AVM treatment, with 80% completely regressing following AVM obliteration and 67% disappearing after reduction of the AVM nidus by more than 50%. Proximal aneurysms regressed infrequently, with only 17% decreasing in size and 4% disappearing after AVM cure. None of the proximal aneurysms regressed with partial AVM treatment. Two patients experienced SAH due to aneurysm rupture during the follow-up period as a result of aneurysm enlargement after partial AVM treatment with less than 50% reduction of the AVM nidus. One of these patients had been treated with stereotactic radiosurgery and one with a combination of stereotactic radiosurgery and embolization.

These results are interesting in light of recent experimental observations reported by Gao, et al.,8 who used a computer simulation model to estimate the magnitude of pressure changes along the vascular tree during stepwise occlusion of an AVM. Prior to AVM obliteration, there is a progressive decrease in arterial pressure along feeding artery pedicles. As the AVM is occluded and flow decreases, the mean arterial pressure increases in a nonlinear fashion, with 70 to 90% occlusion being required to increase the pressure halfway to the eventual maximum value. As the AVM is gradually occluded, the blood velocity and shear stress decrease while the arterial pressure increases. Furthermore, the pressure and flow changes near the circle of Willis are relatively minor, although they are quite profound in distal vessels near the AVM nidus.

The fact that the arteriovenous shunting associated with AVMs induces relative hypotension in the feeding arteries along which flow-related aneurysms develop suggests that rapid or turbulent blood flow is a more important hemodynamic factor in aneurysm formation than intravascular pressure. Although AVM obliteration may acutely increase the transmural pressure across an aneurysm's dome and place it at increased risk for rupture,2,7 this has not occurred in our experience. Our finding that only distal aneurysms have a high probability of regression, but only if more than 50% of the AVM nidus is eliminated, corroborates the predictions of the computer model that substantial hemodynamic effects are limited to the distal vessels near the AVM nidus.8

Conclusions

Arterial aneurysms associated with cerebral AVMs may be classified as intranidal, flow related, or unrelated to the AVM nidus. Our experience confirms that intranidal aneurysms have a high correlation with hemorrhagic clinical presentation and a risk of bleeding during follow up that considerably exceeds that which would be expected in their absence. Patients with flow-related aneurysms in association with an AVM may present with hemorrhage from either lesion. The risk of bleeding during the follow-up period, regardless of clinical presentation, seems to be higher for the AVM than for associated aneurysms. Those aneurysms arising on distal feeding arteries near the nidus have a high probability of regressing with substantial or curative AVM therapy.

Acknowledgments

The authors thank Drs. M. C. Wallace and M. Tymianski for their contribution to the neurosurgical care of many of these patients and for their support of the Brain Vascular Malformation Study Group.

References

  • 1.

    Anderson RMDBlackwood W: The association of arteriovenous angioma and saccular aneurysm of the arteries of the brain. J Pathol Bacteriol 77:1011101959J Pathol Bacteriol 77:

  • 2.

    Batjer HSuss RASamson D: Intracranial arteriovenous malformations associated with aneurysms. Neurosurgery 18:29351986Neurosurgery 18:

  • 3.

    Brown RD JrWiebers DOForbes Get al: The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 68:3523571988J Neurosurg 68:

  • 4.

    Brown RD JrWiebers DOForbes SG: Unruptured intracranial aneurysms and arteriovenous malformations: frequency of intracranial hemorrhage and relationship of lesions. J Neurosurg 73:8598631990J Neurosurg 73:

  • 5.

    Cunha e Sa MJStein BMSolomon RAet al: The treatment of associated intracranial aneurysms and arteriovenous malformations. J Neurosurg 77:8538591992J Neurosurg 77:

  • 6.

    Deruty RMottolese CSoustiel JGet al: Association of cerebral AVM and cerebral aneurysm. Diagnosis and management. Acta Neurochir 107:1331391990Acta Neurochir 107:

  • 7.

    Fuwa IMatsukado YKaku Met al: Enlargement of a cerebral aneurysm associated with ruptured AVM. Acta Neurochir 80:65681986Acta Neurochir 80:

  • 8.

    Gao EYoung WLPile-Spellman Jet al: Cerebral AVM feeding artery aneurysms: a theoretical model of intravascular pressure changes after treatment. Neurosurgery 41:134513581997Neurosurgery 41:

  • 9.

    Garcia-Monaco RRodesch GAlvarez Het al: Pseudoaneurysms within ruptured intracranial arteriovenous malformations: diagnosis and early endovascular management. AJNR 14:3153211993AJNR 14:

  • 10.

    Gibbons GHDzau VJ: The emerging concept of vascular remodeling. N Engl J Med 330:143114381994N Engl J Med 330:

  • 11.

    Graf CJPerret GETorner JC: Bleeding from cerebral arteriovenous malformations as part of their natural history. J Neurosurg 58:3313371983J Neurosurg 58:

  • 12.

    Kader AYoung WLPile-Spellman Jet al: The influence of hemodynamic and anatomic factors on hemorrhage from cerebral arteriovenous malformations. Neurosurgery 34:8018081994Neurosurgery 34:

  • 13.

    Kondziolka DNixon BJLasjaunias Pet al: Cerebral arteriovenous malformations with associated arterial aneurysms: hemodynamic and therapeutic considerations. Can J Neurol Sci 15:1301341988Can J Neurol Sci 15:

  • 14.

    Lasjaunias P: A revised concept of the congenital nature of cerebral arteriovenous malformations. Intervent Neuroradiol 3:2752811997Lasjaunias P: A revised concept of the congenital nature of cerebral arteriovenous malformations. Intervent Neuroradiol 3:

  • 15.

    Lasjaunias PPiske RTerbrugge Ket al: Cerebral arteriovenous malformations (C. AVM) and associated arterial aneurysms (AA). Analysis of 101 C. AVM cases with 37 AA in 23 patients. Acta Neurochir 91:29361988Acta Neurochir 91:

  • 16.

    Marks MPLane BSteinberg GKet al: Hemorrhage in intracerebral arteriovenous malformations: angiographic determinants. Radiology 176:8078131990Radiology 176:

  • 17.

    Marks MPLane BSteinberg GKet al: Intranidal aneurysms in cerebral arteriovenous malformations: evaluation and endovascular treatment. Radiology 182:3553601992Radiology 182:

  • 18.

    Miyasaka KWolpert SMPrager RJ: The association of cerebral aneurysms, infundibula, and intracranial arteriovenous malformations. Stroke 13:1962031982Stroke 13:

  • 19.

    Mullan SMojtahedi SJohnson DLet al: Embryological basis of some aspects of cerebral vascular fistulas and malformations. J Neurosurg 85:181996J Neurosurg 85:

  • 20.

    Okamoto SHanda HHashimoto N: Location of intracranial aneurysms associated with cerebral AVM: statistical analysis. Surg Neurol 22:3353401984Surg Neurol 22:

  • 21.

    Perata HJTomsick TATew JM Jr: Feeding artery pedicle aneurysms: association with parenchymal hemorrhage and arteriovenous malformations of the brain. J Neurosurg 80:6316341994J Neurosurg 80:

  • 22.

    Perret GNishioka H: Report on the Cooperative Study of Intracranial Aneurysms and Subarachnoid Hemorrhage. Section VI. Arteriovenous malformations. An analysis of 545 cases of cranio-cerebral arteriovenous malformations and fistulae reported to the cooperative study. J Neurosurg 25:4674901966J Neurosurg 25:

  • 23.

    Pile-Spellman JMBaker FKLiszczak TMet al: High-flow angiopathy: cerebral blood vessel changes in experimental chronic arteriovenous fistula. AJNR 7:8118151986AJNR 7:

  • 24.

    Pritz MB: Ruptured supratentorial arteriovenous malformations associated with venous aneurysms. Acta Neurochir 128:1501621994Pritz MB: Ruptured supratentorial arteriovenous malformations associated with venous aneurysms. Acta Neurochir 128:

  • 25.

    Rinkel GJEDjibuti MAlgra Aet al: Prevalence and risk of rupture of intracranial aneurysms. A systematic review. Stroke 29:2512561998Stroke 29:

  • 26.

    Suzuki JOnuma T: Intracranial aneurysms associated with arteriovenous malformations. J Neurosurg 50:7427461979J Neurosurg 50:

  • 27.

    Turjman FMassoud TFViñuela Fet al: Aneurysms related to cerebral arteriovenous malformations: superselective angiographic assessment in 58 patients. AJNR 15:160116051994AJNR 15:

  • 28.

    Turjman FMassoud TFViñuela Fet al: Correlation of the angioarchitectural features of cerebral arteriovenous malformations with clinical presentation of hemorrhage. Neurosurgery 37:8568621995Neurosurgery 37:

  • 29.

    Willinsky RLasjaunias PTerbrugge Ket al: Brain arteriovenous malformations: analysis of the angio-architecture in relationship to hemorrhage (based on 152 patients explored and/or treated at the hospital de Bicêtre between 1981 and 1986). J Neuroradiol 15:2252371988J Neuroradiol 15:

  • 30.

    Yaşargil MG: New York: Thieme1987182190Yaşargil MG:

Dr. Redekop was supported by a Detweiler Traveling Fellowship from the Royal College of Physicians and Surgeons of Canada.

Article Information

Address reprint requests to: Gary Redekop, M.D., F.R.C.S.(C), Division of Neurosurgery, The University of British Columbia, 323-C, 700 West 10th Avenue, Vancouver, British Columbia V52 4E5, Canada.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Left: Right ICA angiogram, arterial phase, demonstrating an intranidal aneurysm (arrow) within an AVM supplied by MCA branches and the anterior choroidal artery. Right: Superselective angiogram in which the intranidal aneurysm (curved arrow) is better delineated, with the tip of a microcatheter (long arrow) positioned within the distal anterior choroidal artery.

  • View in gallery

    Angiograms. Left: Occipital AVM supplied by the posterior cerebral artery (PCA). A proximal flow-related aneurysm is present at the basilar artery bifurcation. Right: Small posterior temporal AVM (short arrow) with supply from the distal PCA. A distal flow-related aneurysm is present on the feeding artery (long arrow), with the aneurysm fundus superimposed on an anteriorly directed draining vein.

  • View in gallery

    Angiograms showing a proximal flow-related aneurysm that remains unchanged despite AVM cure. Left: Left ICA angiogram showing extensive crossflow through the ACoA and an ACA aneurysm (arrow) associated with a large right-hemisphere AVM. Right: One year after staged embolization with complete AVM elimination, there is substantial diminution in the caliber of the left ICA and ACA, and there is no longer crossflow through the ACoA. The ACA aneurysm is unchanged.

  • View in gallery

    Angiograms showing a distal flow-related aneurysm that regressed completely following AVM cure. Left: Right ICA angiogram demonstrating an enlarged frontal branch supplying an AVM (curved arrow) and a distal flow-related aneurysm at the bifurcation of the callosomarginal artery (straight arrow). Right: Following embolization, repeated angiography performed 6 months later reveals that the AVM has been eliminated and the aneurysm is no longer present (arrow).

  • View in gallery

    Angiograms demonstrating a flow-related aneurysm that enlarged and bled after partial AVM therapy. Left: Vertebral artery angiogram showing a dorsal midbrain AVM with supply from numerous perforating branches arising from the upper basilar artery and the PCA. A small aneurysm is present at the origin of the left superior cerebellar artery (arrowhead). Right: Seven years after undergoing stereotactic radiosurgery with minimal change in the AVM size, the patient presented with SAH. Angiography demonstrated that the aneurysm had enlarged (arrow). The aneurysm was confirmed as the source of hemorrhage at the time of surgical clipping.

  • View in gallery

    Angiograms displaying a distal flow-related aneurysm that regressed completely following subtotal AVM treatment. Left: Lateral vertebral artery angiogram showing a large occipital AVM with variceal dilation of a draining vein (arrowheads) superimposed over the perimesencephalic segment of the PCA. Center: Following the first stage of embolization there is substantial reduction of the AVM. Repeated angiography, early arterial phase, demonstrating a previously obscured aneurysm arising from the distal PCA (arrow). Right: Final angiographic result after the second stage of embolization with marked decrease in the size of the AVM. The aneurysm has disappeared (arrow).

References

1.

Anderson RMDBlackwood W: The association of arteriovenous angioma and saccular aneurysm of the arteries of the brain. J Pathol Bacteriol 77:1011101959J Pathol Bacteriol 77:

2.

Batjer HSuss RASamson D: Intracranial arteriovenous malformations associated with aneurysms. Neurosurgery 18:29351986Neurosurgery 18:

3.

Brown RD JrWiebers DOForbes Get al: The natural history of unruptured intracranial arteriovenous malformations. J Neurosurg 68:3523571988J Neurosurg 68:

4.

Brown RD JrWiebers DOForbes SG: Unruptured intracranial aneurysms and arteriovenous malformations: frequency of intracranial hemorrhage and relationship of lesions. J Neurosurg 73:8598631990J Neurosurg 73:

5.

Cunha e Sa MJStein BMSolomon RAet al: The treatment of associated intracranial aneurysms and arteriovenous malformations. J Neurosurg 77:8538591992J Neurosurg 77:

6.

Deruty RMottolese CSoustiel JGet al: Association of cerebral AVM and cerebral aneurysm. Diagnosis and management. Acta Neurochir 107:1331391990Acta Neurochir 107:

7.

Fuwa IMatsukado YKaku Met al: Enlargement of a cerebral aneurysm associated with ruptured AVM. Acta Neurochir 80:65681986Acta Neurochir 80:

8.

Gao EYoung WLPile-Spellman Jet al: Cerebral AVM feeding artery aneurysms: a theoretical model of intravascular pressure changes after treatment. Neurosurgery 41:134513581997Neurosurgery 41:

9.

Garcia-Monaco RRodesch GAlvarez Het al: Pseudoaneurysms within ruptured intracranial arteriovenous malformations: diagnosis and early endovascular management. AJNR 14:3153211993AJNR 14:

10.

Gibbons GHDzau VJ: The emerging concept of vascular remodeling. N Engl J Med 330:143114381994N Engl J Med 330:

11.

Graf CJPerret GETorner JC: Bleeding from cerebral arteriovenous malformations as part of their natural history. J Neurosurg 58:3313371983J Neurosurg 58:

12.

Kader AYoung WLPile-Spellman Jet al: The influence of hemodynamic and anatomic factors on hemorrhage from cerebral arteriovenous malformations. Neurosurgery 34:8018081994Neurosurgery 34:

13.

Kondziolka DNixon BJLasjaunias Pet al: Cerebral arteriovenous malformations with associated arterial aneurysms: hemodynamic and therapeutic considerations. Can J Neurol Sci 15:1301341988Can J Neurol Sci 15:

14.

Lasjaunias P: A revised concept of the congenital nature of cerebral arteriovenous malformations. Intervent Neuroradiol 3:2752811997Lasjaunias P: A revised concept of the congenital nature of cerebral arteriovenous malformations. Intervent Neuroradiol 3:

15.

Lasjaunias PPiske RTerbrugge Ket al: Cerebral arteriovenous malformations (C. AVM) and associated arterial aneurysms (AA). Analysis of 101 C. AVM cases with 37 AA in 23 patients. Acta Neurochir 91:29361988Acta Neurochir 91:

16.

Marks MPLane BSteinberg GKet al: Hemorrhage in intracerebral arteriovenous malformations: angiographic determinants. Radiology 176:8078131990Radiology 176:

17.

Marks MPLane BSteinberg GKet al: Intranidal aneurysms in cerebral arteriovenous malformations: evaluation and endovascular treatment. Radiology 182:3553601992Radiology 182:

18.

Miyasaka KWolpert SMPrager RJ: The association of cerebral aneurysms, infundibula, and intracranial arteriovenous malformations. Stroke 13:1962031982Stroke 13:

19.

Mullan SMojtahedi SJohnson DLet al: Embryological basis of some aspects of cerebral vascular fistulas and malformations. J Neurosurg 85:181996J Neurosurg 85:

20.

Okamoto SHanda HHashimoto N: Location of intracranial aneurysms associated with cerebral AVM: statistical analysis. Surg Neurol 22:3353401984Surg Neurol 22:

21.

Perata HJTomsick TATew JM Jr: Feeding artery pedicle aneurysms: association with parenchymal hemorrhage and arteriovenous malformations of the brain. J Neurosurg 80:6316341994J Neurosurg 80:

22.

Perret GNishioka H: Report on the Cooperative Study of Intracranial Aneurysms and Subarachnoid Hemorrhage. Section VI. Arteriovenous malformations. An analysis of 545 cases of cranio-cerebral arteriovenous malformations and fistulae reported to the cooperative study. J Neurosurg 25:4674901966J Neurosurg 25:

23.

Pile-Spellman JMBaker FKLiszczak TMet al: High-flow angiopathy: cerebral blood vessel changes in experimental chronic arteriovenous fistula. AJNR 7:8118151986AJNR 7:

24.

Pritz MB: Ruptured supratentorial arteriovenous malformations associated with venous aneurysms. Acta Neurochir 128:1501621994Pritz MB: Ruptured supratentorial arteriovenous malformations associated with venous aneurysms. Acta Neurochir 128:

25.

Rinkel GJEDjibuti MAlgra Aet al: Prevalence and risk of rupture of intracranial aneurysms. A systematic review. Stroke 29:2512561998Stroke 29:

26.

Suzuki JOnuma T: Intracranial aneurysms associated with arteriovenous malformations. J Neurosurg 50:7427461979J Neurosurg 50:

27.

Turjman FMassoud TFViñuela Fet al: Aneurysms related to cerebral arteriovenous malformations: superselective angiographic assessment in 58 patients. AJNR 15:160116051994AJNR 15:

28.

Turjman FMassoud TFViñuela Fet al: Correlation of the angioarchitectural features of cerebral arteriovenous malformations with clinical presentation of hemorrhage. Neurosurgery 37:8568621995Neurosurgery 37:

29.

Willinsky RLasjaunias PTerbrugge Ket al: Brain arteriovenous malformations: analysis of the angio-architecture in relationship to hemorrhage (based on 152 patients explored and/or treated at the hospital de Bicêtre between 1981 and 1986). J Neuroradiol 15:2252371988J Neuroradiol 15:

30.

Yaşargil MG: New York: Thieme1987182190Yaşargil MG:

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 11 11 11
Full Text Views 58 58 32
PDF Downloads 83 83 55
EPUB Downloads 0 0 0

PubMed

Google Scholar