How Do We Treat Arteriovenous Malformations (Tips and Tricks)?
Young Soo Do, MD, Kwang-Bo Park, MD, and Sung Ki Cho, MD
Ethanol embolization is sufficient to eliminate or improve symptoms of arteriovenous malformations (AVMs) in a high percentage of patients, but with substantial risk of minor and major complications. Inadvertent embolization must be avoided by superselective catheterization or direct puncture of the nidus. According to the angiographic morphology of the nidus, AVMs of the trunk and extremities can be classified into four types: type I (arterio-venous fistulae), type II (arteriolo-venous fistulae), type IIIa (arteriolo-venulous fistulae without dilation of the fistula), and type IIIb (arteriolo-venulous fistulae with dilation of the fistula). The described angiographic classification provides considerable information concerning the characteristics of AVMs in the body and extremities, the optimal therapeutic approach, and the likely therapeutic outcome. Tech Vasc Interventional Rad 10:291-298 © 2007 Elsevier Inc. All rights reserved. KEYWORDS: arteriovenous malformation, ethanol, embolization, angiography
Congenital arteriovenous malformations (AVMs) are a major challenge in medical practice, which are often associated with serious symptoms such as heart failure, neuropathy, pain, and bleeding. Embolotherapy has been a primary mode of treatment for the management of AVMs with improvements in catheter systems and selective techniques, which can be achieved through the transarterial, transvenous, or direct puncture approaches.
Despite the importance of angiography, most reports have traditionally described the angiographic features of AVMs in terms of feeding arteries, the complex network of arteriovenous shunts (referred to as the nidus), and draining veins. However, AVMs are so diverse and diffuse in nature that such descriptions are too broad to accurately represent the individual
characteristics of an AVM. Furthermore, such descriptions have limitations in terms of aiding therapeutic decision making by interventional radiologists and in terms of the exchange of angiographic data between therapeutic teams.
Recently, to help the therapeutic decision making for performing embolotherapy, we have proposed four prototypes of peripheral AVMs based on the angiographic morphology of the nidus. In this article, we described therapeutic results and approaches to the ethanol embolization of AVMs according to a newly proposed angiographic classification of AVMs of the body and extremities.
Techniques and Methods Patient Selection Criteria All patients referred to our hospital had serious problems. These included severe pain, a pulsating growing mass, ulceration at the lesion site, bleeding at the ulcerated skin, vaginal bleeding, dysmenorrhea, bone overgrowth at the lesion site, symptoms or signs of congestive heart failure consisting of dyspnea on exertion and cardiomegaly, ischemic symptoms at the extremities of the distal part of AVMs as a result of a steal effect through a high-flow shunt, or their combinations. The patients usually had more than one symptom and sign. Except severe infection associated with AVMs, the AVMs of extremity and trunk were treated by ethanol embolization.
Diagnosis and Follow-up of AVMs The diagnoses of AVMs were made based on the clinical and imaging examinations. The lesion was first assessed by a physical examination. Multidetector computed tomography (MDCT) or magnetic resonance (MR) imaging besides the selective angiography should be performed to obtain the detailed anatomical and hemodynamic information of AVMs. MDCT or MR imaging are also good imaging modalities for the evaluation of residual lesions or recurrence after treatment.
The Angiographic Classification of AVMs It is well known that the nidus of AVMs is fundamentally a conglomeration of arteriovenous fistulae. AVMs of the body and extremities can be classified into four types according to their angiographic morphologies (Fig. 1), as follows: type I (arterio-venous fistulae), when at most three separate arteries are shunted to a single draining vein; type II (arteriolo-venous fistulae), when multiple arterioles are shunted into a single draining vein; type III (arteriolo-venulous fistulae), when multiple shunts are located between the arterioles and venules. In this type, if the fistula unit of the nidus is observed as a blush or fine striation on angiography, it is subdivided into either; type IIIa (arteriolo-venulous fistulae with nondilated fistula), when the fistula unit of the nidus is observed as a complex vascular network; or type IIIb (arteriolo-venulous fistulae with dilated fistula). This classification can be regarded as a modification of the previous classification proposed by Houdart and coworkers, because they classified intracranial AVMs into three types based on the morphology of the nidus—arterio-venous, arteriolo-venulous, and arteriolo-venulous fistulae.
Type I (arterio-venous fistulae) AVMs: No more than three separate arteries shunt to the initial part of a single venous component.
Type II (arteriolo-venous fistulae): Multiple arterioles shunt to the initial part of a single venous component, in which the arterial components show a plexiform appearance on angiography.
Type IIIa (arteriolo-venulous fistulae with nondilated fistula): Fine multiple shunts are present between arterioles and venules and appear as a blush or fine striation on angiography.
Type IIIb (arteriolo-venulous fistulae with dilated fistula): Multiple shunts are present between arterioles and venules and appear as a complex vascular network on angiography.
Procedures All of the ethanol embolization was performed under general anesthesia to control the pain, and Swan-Ganz and arterialline monitoring were performed in most patients to monitor the pulmonary artery pressure and arterial blood pressure. The objective of the staged ethanol embolization was to embolize not vascular feeders but all or part of the nidus until the complete resolution or improvement in the clinical symptoms and signs was achieved. In every session of embolotherapy, baseline selective and superselective angiographic studies were performed to determine the exact flow characteristics of the AVMs. After considering not only the angiographic features of the AVMs but also other factors such as important normal arterial or venous branches arising in very close proximity to a malformation and extreme arterial tortuosity or previous surgery precluding a successful transarterial catheterization, the approach of vascular access to attack the nidus itself was chosen. Transarterial, transvenous catheterization using a coaxial catheter, and/or percutaneous direct puncture was needed to reach the nidus being embolized. When focusing only on the morphology of the nidus, the type I and II AVMs can be embolized through any of the three approaches: transarterial, transvenous, or direct puncture approaches. The type IIIa AVMs can be treated only through a transarterial approach because they are too fine to be punctured directly. The type IIIb AVMs can be treated through either the transarterial or direct puncture approaches. In the type III AVMs, embolization through the transvenous approach is contraindicated because embolic materials introduced by a transvenous approach will not reach the shunts but block the venous drainage. Therefore, this will result in hypertension in the shunts and aggravation of AVMs if the arterial contribution is not disrupted. By the definition of this classification, the type II AVMs have multiple tortuous feeding arterioles and a large dilated venous component. Therefore, it is important to attack the venous component of the nidus through a direct puncture or transvenous approaches rather than through the transarterial approach. Also, a direct puncture and transvenous approach can avoid the risk of an inadvertent occlusion of the adjacent normal vessels by the transarterial approach. In the mixed type AVMs, the transarterial and direct puncture approaches of embolotherapy, which can treat both components of the AVMs simultaneously, are preferred.
To determine the volume of ethanol used during embolizations and the rate of injection, test injections of the contrast medium were performed under fluoroscopic monitoring. The amount of ethanol used was based on the amount of contrast medium required to fill the AVM nidus without opacifying normal vessels. When maximum amounts of ethanol were used, the total dose did not exceed 1 mL/kg of body weight. In some cases, to achieve vascular stasis, a proximal inflow occlusion was performed during the ethanol injection using either an intravascular occlusion balloon catheter for the trunk and pelvis area or external pneumatic blood pressure cuffs for extremities. Of the embolic materials, absolute ethanol (99%) was used in most cases and 50-70% ethanol with a nonionic contrast medium (Xenetix 300; Guerbet, Cedex, France) was used in some AVMs of the hands or feet. Additional coil embolizations (Nester coil; Cook, Bloomington, IN) of the dominant outflow vein were performed in Type II AVMs by direct puncture or transvenous catheterization to reduce the amount of ethanol and to stabilize the thrombosis in the dominant outflow vein as Yee and Yakes previously reported (Fig. 2). After waiting 5 to 10 minutes after the ethanol injection, an arteriogram was performed to determine whether AVMs had been embolized completely. Complete embolization of at least one compartment of the AVMs required the meticulous repetition of the technique described earlier. When an elevated pulmonary artery pressure (more than 25 mm Hg) was sustained at the end of the session or a large amount of absolute ethanol was injected, the patient was kept at the intensive care unit for close pulmonary artery pressure monitoring and continuous administration of nitroglycerine (0.3-3.0 µg/kg/min). All patients were closely followed up to identify any immediate or delayed complications.
As a rule, additional embolization was recommended if the symptoms and signs remained or the AVMs remained at follow-up imaging studies. The recommended time interval between embolotherapy is more than 2 months.
A diagrammatic presentation of approaches to ethanol embolotherapy according to the angiographic type.
(A) The main target of type II AVMs is the venous component of the nidus. Therefore, the mainstay therapeutic approaches are transvenous (TV) and direct puncture (DP). Before ethanol embolization, coil embolization of the venous component of the nidus through a transvenous or direct puncture approach is often required to reduce the amount of ethanol and to stabilize the thrombosis within the large venous component.
(B) Type IIIa AVMs. Only the transarterial (TA) approach is available. Because the fistula is too fine, direct puncture of the AVMs is not possible.
(C)Type IIIb AVMs. These AVMs can be treated properly via transarterial (TA) and direct puncture (DP) approaches. However, if there are no obstacles in terms of access and safe embolization, the transarterial approach is preferred because of familiarity with arterial angiographic findings, the ease of detecting a normal artery arising near an AVM, and the direct puncture hazard whereby ethanol can leak into adjacent soft tissues. The treatment of type III AVMs via a transvenous approach is contraindicated. The approaches used for mixed types are combinations of those used to treat the individual types, but an approach that simultaneously treats all types present is preferred.
Evaluation of Angiographic Studies and Clinical Data
If an AVM that had been confined to one anatomic lesion at the initial workup showed two or more different types by angiography, the AVM was considered to be a mixed type of the two or more different types present. The therapeutic outcomes were assessed according to symptoms and signs and degree of devascularization at angiography. Cure was defined as complete resolution of clinical symptoms and signs, with 100% devascularization of AVMs at angiography. Partial remission was defined as complete resolution or an improvement in clinical symptoms and signs, with 50-99% AVM devascularization at angiography. No remission was defined as an improvement or no change in clinical symptoms and signs, with<50% devascularization at angiography. Aggravation was defined as a worsening of clinical symptoms and signs, regardless of the degree of AVM devascularization at angiography. Cure and partial remission were considered to be effective (successful) therapeutic outcomes of ethanol embolization.
Results of assessing the angiographic types of the 66 AVMs based on the classifications are given in Table 1. The most common form of AVMs was type IIIb (45%, 30/66 AVMs). There were no type I AVMs. Twenty-one (32%) AVMs were categorized into the mixed form. Fifty-one (77%) AVMs had a type IIIb component, 22 (33%) AVMs had a type II component, and 14 (21%) AVMs had a type IIIa component.
The therapeutic outcomes according to the types of AVMs are also summarized in Table 1. Ethanol embolization was effective in 49 (cure 21, partial remission 28) (74%) of 66 patients (Figs. 3 and 4). Of the 19 patients waiting for further treatment, 9 patients were in the partial remission group and 10 patients were in the no remission group.
With statistics, the therapeutic outcome for type II AVMs was better than for any other types (P < 0.05). No statistical difference in the therapeutic outcomes was found between the other types of AVMs (P > 0.05).
Complications related to ethanol embolotherapy according to the angiographic types of 66 AVMs are detailed in Table 2. Thirty-two (48%) of 66 patients had complications and there was no procedure-related mortality. Fifty-three (21%) complications occurred in 255 staged embolotherapy procedures. Four patients had major and minor complications together during the procedures. There were 43 minor complications in 28 patients (43/255 sessions, 17%; 28/66 patients, 42%). There were 10 major complications in 8 patients (10/255 sessions, 4%; 8/66 patients, 12%).
Figure 3 A 21-year-old male with type II AVM in the left calf. (A-C) Pretreatment posteroanterior angiography (arterial, late arterial, venous phase) shows plexiform arteriolar component (arrows) and a large venous component (arrowheads). (D) On superselective angiography, the components of a type II AVM can be easily identified.
Figure 4 A 42-year-old male with type IIIb AVM in the left calf. (A and B) Posteroanterior angiography (arterial and venous phase) shows multiple arteriolar and venulous components of the nidus. (C and D) Posteroanterior angiography (arterial and venous phase) after six embolotherapy sessions by the transarterial and direct puncture approaches shows a obliteration of 50% of the AVM.