- Email: [email protected]
Embolization of Peripheral Vascular Malformations with Ethylene Vinyl Alcohol Copolymer (Onyx)
Embolization of Peripheral Vascular Malformations with Ethylene Vinyl Alcohol Copolymer (Onyx) Furuzan Numan, MD, Alp Ömerog˘lu, MD, Batuhan Kara, MD, Murat Cantas¸demir, MD, I˙brahim Adaletli, MD, and Fatih Kantarcı, MD
PURPOSE: To demonstrate the feasibility and preliminary efficacy of endovascular embolization of peripheral congenital vascular malformations (VMs) with use of a nonadhesive liquid embolic agent, Onyx. MATERIALS AND METHODS: Nine patients with a mean age of 20.8 years had local low-flow (n ⴝ 4), local high-flow (n ⴝ 3), or diffuse high-flow (n ⴝ 2) VMs located in the upper or lower extremities. In all patients, endovascular embolization was performed via the superselective catheterization of arterial feeders of VMs with use of microcatheters in a coaxial technique. A total of 15 embolization procedures were performed with Onyx, which was composed of 6%, 8%, or 20% ethylene vinyl alcohol copolymer dissolved in dimethyl sulfoxide. RESULTS: In two of four patients with local low-flow VMs, the lesions were embolized completely. In the other two patients with local low-flow VMs, embolizations were incomplete. The remaining five high-flow lesions of local (n ⴝ 3) or diffuse (n ⴝ 2) types were also embolized incompletely. In all patients with local low-flow VMs and in one patient with a local high-flow VM, clinical signs and symptoms were resolved significantly. Other patients did show clinical benefit from embolization to varying degrees. CONCLUSION: In our experience in a limited number of cases, Onyx promises and provides important advantages over conventional embolic agents in the endovascular transcatheter embolization of congenital peripheral VMs. However, as with other embolic agents, it is far from perfect. J Vasc Interv Radiol 2004; 15:939 –946 Abbreviations:
DMSO ⫽ dimethyl sulfoxide, NBCA ⫽ n-butyl cyanoacrylate, PVA ⫽ polyvinyl alcohol, VM ⫽ vascular malformation
CONGENITAL peripheral vascular malformations (VMs) are uncommon lesions that can present a myriad of clinical presentations. Surgery is usually unfeasible or inefficient in the treatment of peripheral VMs (1,2). En-
From the Department of Radiology (F.N., B.K., M.C., F.K.), 3400, Cerrahpasa Medical Faculty, Istanbul University, Istanbul; Department of Radiology, Izzet Baysal Medical Faculty (A.Ö.), Abant Izzet Baysal University, Bolu; and Department of Radiology (I˙.A.), Gaziantep University Medical Faculty, Gaziantep, Turkey. Received December 23, 2003; revision requested February 18, 2004; final revision received April 1; accepted April 12. Address correspondence to F.N.; E-mail: [email protected] None of the authors have identified a conflict of interest. © SIR, 2004 DOI: 10.1097/01.RVI.0000130862.23109.52
dovascular transcatheter embolization has been suggested as the primary therapeutic option or as a presurgical intervention to reduce bleeding and maximize successful resection. The agents available for endovascular transcatheter embolization include polyvinyl alcohol (PVA) particles, absolute alcohol, acrylic polymers, coils, and n-butyl cyanoacrylate (NBCA). The Onyx Liquid System (Embolyx; Micro Therapeutics, Irvine, CA) is a new nonadhesive liquid embolic agent that is currently being evaluated in the treatment of cerebral arteriovenous VMs (3–5). Peripheral use of Onyx has been published in two case reports (6,7), one in traumatic pseudoaneurysms and the other in treatment of pelvic arteriovenous VMs. The aim of this study is to report the feasibility
and preliminary efficacy of endovascular embolization of peripheral congenital VMs with use of Onyx.
MATERIALS AND METHODS Patient Population From October 2001 to August 2003, nine patients (four men, five women) with a mean age of 20.8 years (range, 11–33 years) underwent endovascular transcatheter arterial embolization with use of Onyx in the treatment of peripheral VMs. Table 1 summarizes patients’ age, sex, VM type and location, signs and symptoms at presentation, and previous therapeutic procedures. Of nine patients, two had diffuse high-flow VMs located at the shoulder
939
940
•
Embolization of Peripheral Vascular Malformations with Onyx
September 2004
JVIR
Table 1 Summary of Cases Patient No.
Age (y)/ Sex
Lesion Type
1
18/M
2
11/M
3
12/F
4
33/M
5
30/F
6
16/F
7
27/F
8
15/M
Low flow (local) Low flow (local) Low flow (local) Low flow (local) High flow (local) High flow (local) High flow (local) High flow (diffuse)
9
26/F
High flow (diffuse)
Lesion Location Right gluteus
Signs and Symptoms
Previous Treatment –
Left calf
Pain, swelling, varicosity, erythema, discoloration Pain in activity, discoloration, erythema
Right thigh
Pain, swelling, discoloration
–
Right gluteus
Pain, swelling, varicosity, erythema, discoloration Pain, swelling, varicosity, erythema, discoloration Pain, swelling, erythema, discoloration, scaling, ulceration, disabling of extremity Pain, swelling, varicosity, discoloration, thrill, congestive heart failure Pain, swelling, varicosity, discoloration, erythema, scaling, disabling of extremity, thrill, congestive heart failure Pain, swelling, varicosity, discoloration, erythema, scaling, disabling of extremity, thrill, congestive heart failure
Left forearm Right calf Left arm Left shoulder and upper extremity Left shoulder and upper extremity
and upper extremity. Previous unsuccessful surgical management was performed in one of these patients and several endovascular transcatheter embolization procedures were performed previously with use of PVA particles in the other one. Seven of nine patients had local VMs, which were of low-flow type in four and high-flow type in three. Two patients with local high-flow VMs had undergone previous treatment. One of them had unsuccessful surgery and the other had previous endovascular transcatheter embolization with use of NBCA. One patient with a local lowflow VM had undergone surgery and then embolization with NBCA and PVA particles. Signs and symptoms at presentation included pain, swelling, multiple varicosities, thrill, trophic changes of the skin, functional impairment of the involved extremities, high cardiac output, and congestive cardiac failure (Table 1). Embolization Technique All patients underwent diagnostic angiography via the femoral artery route under local anesthesia just before the embolization procedures. Catheterization was performed via a transfemoral approach with use of
standard coaxial techniques. A dimethyl sulfoxide (DMSO)– compatible microcatheter/microwire combination (Rebar microcatheter and Silver Speed micro– guide wire; Micro Therapeutics) was used to perform superselective catheterization of the feeding arteries. Onyx embolization required flushing of the microcatheter with saline solution first and then with 0.4 mL DMSO to fill the microcatheter’s “dead space.” After that, the Onyx was aspirated into a 1-mL syringe, and 0.25 mL of this amount was injected slowly for 40 seconds to fill the microcatheter and replace the DMSO in the dead space. Finally, Onyx was injected at a volume and rate low enough to prevent reflux of the agent around the microcatheter under fluoroscopic guidance. The injection was maintained and repeated until the distal end of the feeding arteries and the abnormal malformed vascular communications were completely occluded. Because the estimation of injection volume of Onyx is very difficult, especially in high-flow lesions, as a result of the complex structure of the abnormal vascular communications, the volume and rate of Onyx to be injected was determined only by fluoroscopic monitoring during injection. We did not perform an occlusion injection test of contrast material to determine the
–
Surgery and PVA/NBCA embolization Surgery NBCA embolization – PVA embolization Surgery
volume of Onyx before the injection. The reliable criteria used to determine the volume and rate of injection were based on the penetration of the Onyx to the abnormal vascular communications and avoidance of reflux. Because Onyx is not adhesive, multiple injections could be performed. In multiple injections, we used the same catheter to catheterize another feeding artery, provided that the microcatheter had been withdrawn first and then flushed with DMSO and physiologic saline solution again before each catheterization process. Onyx is supplied in ready-to-use vials. Each vial contains ethylene vinyl alcohol copolymer, DMSO, and tantalum powder for radiopacity. Ethylene vinyl alcohol copolymer is formed from 48 mol/L ethylene and 52 mol/L vinyl alcohols. The polymer is dissolved in DMSO and prepared in different concentrations. Onyx 6.0% contains 6.0% copolymer and 94% DMSO, Onyx 8.0% contains 8.0% copolymer and 92% DMSO, and Onyx 20% contains 20% copolymer and 80% DMSO. The lower the concentration of the copolymer, the less viscous the agent and the more distal penetration can be achieved. In most patients, we used Onyx 6% and Onyx 8% to achieve more distal penetration. However, for a fistulous component of a diffuse
Volume 15
Number 9
high-flow VM in one patient (patient 9), we required fast polymerization and therefore used Onyx 20%. Systemic anticoagulation was achieved during the procedures with intraarterial heparin in 5,000-IU bolus injections. For an effective and comfortable intervention, the pain during embolizations was primarily eliminated by analgesics and/or sedatives. General anesthesia should be reserved for children and patients who would not be able to tolerate the pain even with administration of analgesics and/or sedatives. Study Endpoints and Definitions Diffuse VMs were defined by involvement of the entire extremity, whereas local VMs were defined by involvement of a restricted area of an extremity. VMs were also subdivided into high-flow and low-flow groups according to the speed of flow through the lesion and the rate of shunting and/or abnormal communication between the arterial and venous components. In high-flow lesions, arteries and veins were visualized through the arterial phase of the arteriogram. If the venous component of the lesion appeared after the arterial phase, the lesion was defined as a low-flow VM. Technical success was defined as complete if the postembolization arteriograms revealed no flow in the VM. In cases of incomplete technical success, flow through the abnormal vascular communications of VMs reduced but the filling of the VM still remained. We further defined incomplete success qualitatively as having moderately or significantly reduced flow through a VM. Clinical success was defined as total or partial improvement of clinical signs and symptoms such as pain, swelling, erythema, discoloration, scaling, ulceration, disabling of extremity, thrill, and congestive heart failure. An unsuccessful clinical result was defined by persistence of clinical signs or additional symptoms related to ischemic complications secondary to nontarget embolization. Major complications were defined as nontarget embolization with acute ischemia and pulmonary embolism. Minor complications included post-
Numan et al
embolization syndrome, gluing of the catheter to the vessel wall, and vasospasm related to the angiotoxicity of DMSO. Our institutional guidelines do not require institutional review board approval in these circumstances, and Declaration of Helsinki principles were followed.
RESULTS Number of embolization procedures, amount of injected Onyx, complications, angiographic results of embolization procedures, clinical outcome, and length of clinical follow-up are given in Table 2. A total of 15 embolization procedures were performed in nine patients. In five cases, embolization procedures were performed once; in three cases, it was performed two times; and in one case, it was performed four times. Multiple injections in single embolization procedures were performed in most cases, and several arterial feeders were embolized during these 15 embolization procedures. In most embolization procedures, the same microcatheter was used for repeat catheterization and injection of the embolic agent. In 15 embolization procedures, a total of 69 mL of Onyx was used. The mean volume of injected Onyx in each procedure was 4.6 mL. Postembolization angiograms revealed total occlusion in two patients with local lowflow VMs. In the remaining two patients with local low-flow VMs, embolizations were incomplete but achieved significantly reduced flow through the malformation. In all patients with high-flow VMs (diffuse, n ⫽ 2; local, n ⫽ 3), embolizations were also incomplete. In three of these cases, we achieved moderately reduced flow through the VM. In two patients with local high-flow VMs, a postembolization angiogram showed significantly reduced flow through the lesion. One of these cases was clinically unsuccessful as a result of nontarget embolization and the other was clinically successful (Fig 1). In all patients with local low-flow VMs (n ⫽ 4), clinical signs and symptoms were significantly resolved as shown in Table 2. In one of two patients with completely occluded local low-flow VMs (Fig 2), only mild discoloration persisted on follow-up and
•
941
other symptoms disappeared (patient 1). All symptoms were resolved in the other patient (patient 2). In one of two patients (patient 4) with incompletely occluded local low-flow VMs, only mild discoloration and mild swelling were observed during the follow-up period, whereas the other patient (patient 3) was free of symptoms. In a patient with an incompletely occluded local high-flow VM (patient 5), embolization relieved the pain but other symptoms persisted. Patient 6, who had an incompletely occluded local high-flow VM, reported that symptoms had worsened following the procedure, likely as a result of nontarget embolization, despite significantly reduced flow through the lesion. In patient 7, who had an incompletely occluded local high-flow VM, all symptoms regressed. In two patients with incompletely occluded diffuse high-flow VMs (patients 8 and 9), we observed improvement of congestive heart failure, and only mild discoloration and varicosity remained. There were no complications related to the procedure in six patients (four with local low-flow VMs, one with diffuse high-flow VM, and one with local high-flow VM). In the remaining three patients, who had highflow VMs (local, n ⫽ 2; diffuse, n ⫽ 1), reflux of Onyx caused nontarget embolization. In patient 5, reflux to the ulnar artery resulted in occlusion of the artery without ischemia. In patient 6, reflux to the anterior and posterior tibial arteries caused peripheral ischemia and worsening of skin condition as a result of occlusion of these arteries. In patient 9, who had diffuse involvement of the left upper extremity, we observed transient ischemia and worsening of skin condition as a result of minimal arterial reflux of the embolic agent to the brachial artery. Patients 5 and 9 underwent conservative management. Patient 6 was referred to undergo plastic surgery and treated with total excision of the lesion and reconstruction with free latissimus dorsi muscle transfer. The patients with partially successful procedures were followed up and evaluated with physical examination during follow-up visits. No clinical progression was detected in any of these patients. No instances of gluing of the catheter to the vessel wall occurred. There was no vasospasm related to angiotox-
942
•
Embolization of Peripheral Vascular Malformations with Onyx
September 2004
JVIR
Table 2 Results of Onyx Embolization Onyx No. of Onyx Volume Embolizations Concentration (mL)
Patient No.
Lesion Type
1
Low flow (local)
1
6%
7.5
–
Complete
2
Low flow (local) Low flow (local)
1
6%
3
–
Complete
1
6%
3
–
4
Low flow (local)
2
6%
6
5
High flow (local)
2
6%
9
6
High flow (local)
1
6%
6
7
High flow (local)
1
6% 8%
3 1.5
8
High flow (diffuse)
2
6%
9
High flow (diffuse)
4
6% 8% 20%
3
12
9 7.5 1.5
Complications
Embolization Results
Clinical Results Only mild discoloration remained Complete success
Incomplete Complete success (significantly reduced flow) Only mild – Incomplete discoloration and (significantly swelling remained reduced flow) Only pain resolved Reflux resulting Incomplete (moderately in ulnar reduced artery flow) occlusion without ischemia or nerve palsy Reflux resulting Incomplete Unsuccessful in ATA and (significantly PTA reduced occlusion, flow) peripheral ischemia and worsening of skin change Complete success – Incomplete (significantly reduced flow) Improvement of – Incomplete congestive heart (moderately failure; mild reduced discoloration and flow) varicosity remained Improvement of Incomplete Reflux to congestive heart (moderately brachial failure; mild reduced artery discoloration and flow) resulting in varicosity remained transient peripheral ischemia and worsening of skin changes
Follow-up (mo) 18 18 25
15
17
18
3
21
28
ATA ⫽ anterior tibial artery; PTA ⫽ posterior tibial artery.
icity of DMSO. In the follow-up period, we did not observe any pulmonary embolism in these cases.
DISCUSSION There is a wide variety of classifications for VMs based on clinical, histopathologic, and hemodynamic prop-
erties. The classification introduced by Mulliken and Glowacki (8) was based on the clinical and histologic characteristics of VMs. They categorized these lesions as hemangiomas, and VMs consisted of arterial, capillary, venous, or lymphatic channels, or any combination of these components. However, this classification system is
not useful for determination of therapeutic options. To select the accurate treatment method, Jackson et al (9) classified VMs according to the vascular dynamics of the lesion, ie, low flow and high flow. These terms indicate the speed of flow through the lesion and the rate of shunting between the arterial and venous components. We
Volume 15
Number 9
Numan et al
•
943
Figure 1. (a) Selective left subclavian artery injection demonstrates a local high-flow VM in the left arm of patient 7. (b) Brachial artery injection depicts an arterial feeder aneurysm. (c) Postembolization arteriography shows Onyx cast at the neck of the aneurysm and total exclusion of the aneurysm from the circulation. (d) Postembolization arteriography reveals significantly reduced flow through the VM.
Figure 2. (a) Right internal iliac arteriogram shows a local right posteromedial gluteal low-flow VM in patient 1. (b) A postembolization selective right internal iliac arteriogram demonstrates total occlusion of the lesion. Note the Onyx casts and vasospasm caused by catheter manipulation.
used this categorization based on hemodynamic properties of VMs. The management of congenital VMs can present a difficult therapeutic
challenge. Surgical treatment is notoriously difficult and carries a high likelihood of recurrence (1). Endovascular transcatheter embolization of periph-
eral VMs has become more widely accepted as a first therapeutic option for many VMs (2,10). Endovascular management of pe-
944
•
Embolization of Peripheral Vascular Malformations with Onyx
ripheral VMs begins with finding the appropriate embolic agent to match the morphology and hemodynamic status of the lesion. The ideal embolic agent would be nonclumping, highly radiopaque, nontoxic, nonallergenic, and inexpensive. It would allow easily controlled injection through a microcatheter and would provide permanent occlusion. Most important, it should have the ability to penetrate and occlude the abnormal foci of vascular communications in VMs. Because embolization coils and detachable balloons provide proximal occlusion and cannot penetrate the foci, their role in the endovascular management of complex VMs is limited. PVA particles may be used in some cases, but determining the appropriate size of particles to use is difficult, leading to the risk of pulmonary embolism. Recanalization or development of arterial collaterals are other possible problems with PVA embolization (11). Unlike particles, which may be arrested at a precapillary level and cannot always penetrate the abnormal foci of vascular communications in VMs because of their size, liquid embolic agents can pass to the capillary level and penetrate deep into foci. This feature of liquid embolic agents makes them more feasible in the endovascular treatment of VMs. Several liquid embolic agents are available for the endovascular transcatheter embolization of peripheral VMs. The main liquid embolic agents used in endovascular procedures are cyanoacrylates. NBCA is the preferred agent among them and is often referred to as glue (12). The substance polymerizes immediately on contact with blood or other ionic fluids. Polymerization of NBCA results in an exothermic reaction that destroys the vessel wall. Because of rapid polymerization of NBCA when in contact with blood, precise and safe occlusion can be difficult to achieve. In addition, after the injection of NBCA, the microcatheter should be removed quickly so that the catheter tip does not adhere to the vessel wall (12,13). Other disadvantages of NBCA are that the microcatheter cannot be reused during the same procedure and transembolization angiograms cannot be performed. Another liquid embolic agent, absolute alcohol (ethanol), is also widely
used to treat certain types of VMs. It is usually preferred as a sclerosing agent for the treatment of predominantly venous lesions by direct percutaneous injections. However, it can be damaging if it reaches the capillary bed of skin, and it may cause significant softtissue swelling, which may subsequently lead to compartment syndrome. Ethanol has a direct toxic effect on the endothelium that activates the coagulation system and causes occlusion of the vessel lumen by means of thrombosis. As a result of its progressive and permanent occlusion effect, it has also been used in the endovascular management of VMs. However, if large amounts of absolute alcohol enter the systemic circulation, toxic effects including central nervous system depression, hemolysis, or cardiac arrest may occur (14). The physical and biologic properties of embolic agents affect their level of occlusion, which constitutes the heart of the matter in endovascular treatment of peripheral VMs. Traditionally, several embolic agents have been used in the endovascular treatment of peripheral VMs, but they are not always able to penetrate the abnormal foci of vascular communications. Onyx is a new biocompatible liquid embolic agent consisting of ethylene vinyl alcohol copolymer dissolved in various concentrations of DMSO. Micronized tantalum powder, added for radiopacity, is used in higher concentrations in the latest generations of Onyx. When this mixture contacts aqueous media such as blood, DMSO rapidly diffuses from the mixture, causing in situ precipitation and solidification of the polymer, with formation of an elastic soft, spongy embolus without adhesion to the vascular wall (3,4). This polymerization process is time-dependent and mainly affected by the amount of ethylene within the copolymer. As one decreases the percentage of ethylene, the polymer becomes softer and spongier and exhibits a different solidification rate and behavior in blood (6). In addition, concentration of the copolymer dissolved in DMSO determines the viscosity of the embolic agent, which is especially important for choosing the proper concentration of copolymer during a procedure. The lower the concentration of the copolymer, the less viscous the agent is and the more distal pene-
September 2004
JVIR
tration can be achieved in low-flow lesions. In high-flow VMs and for the fistulous components of VMs, higher concentrations of copolymer should be preferred for fast polymerization and to avoid venous reflux. Because Onyx is mechanically occlusive but not adherent to vascular tissues, theoretically, there is no risk of catheter gluing to the arterial wall. Another technical advantage of Onyx is the possibility to perform transembolization angiography, which enables assessment of the requirement for additional embolic agent. The biologic and technical advantages of Onyx over other liquid embolic agents, and the reported results in the endovascular treatment of cerebral arteriovenous VMs with Onyx, encouraged us to use it in embolization of peripheral VMs. As we detected in our cases, one of the advantages of this embolic agent was its ability to conform to the shape of the tortuous arteries supplying the VM. This property provided good penetration to the abnormal foci of vascular communications, which is one of the major issues of endovascular treatment. Onyx does not precipitate in the presence of its solvent, DMSO. Preloading of the microcatheters with DMSO before Onyx injection provides longer injection time and more controllable embolization. Another advantage of Onyx over tissue adhesives is preservation of the option for subsequent surgical resection of the VM. Because Onyx is nonadhesive and maintains the vessel wall integrity, surgical removal of the VM after Onyx embolization seems to be easier than with NBCA (3,15). The nontarget embolization caused by reflux to the parent artery was the most frequent problem we encountered in the endovascular treatment of peripheral VMs with Onyx. This complication occurred in patients with high-flow VMs (patients 5, 6 and 9; Table 2). Regarding our experience in these cases, we think the reasons for the reflux were the complex unpredictable angioarchitecture of highflow lesions, short arterial feeders close to the parent artery, and poor radiopacity of the former generation of Onyx. These features made it difficult to decide the proper injection rate and amount of Onyx to be used. A test injection with contrast media before
Volume 15
Number 9
delivery of Onyx is not always helpful in the estimation of the threshold for reflux as a result of viscosity differences between Onyx and contrast media and hemodynamic changes during the occlusion of the feeder. During the final stage of the injection, the pressure in the feeding artery possibly increases to a critical threshold. After that point, Onyx may easily reflux to the parent artery. In addition to the parent artery reflux, there is always the theoretical risk of venous reflux, especially in high-flow VMs. If Onyx passes through the high-flow fistulous component of the VM before polymerization, it may result in venous reflux. We observed that the refluxed Onyx in the venous site tends to move but stay along the vessel wall. This may be because of its viscosity and laminar flow. In patient 7, who had a local VM in the left upper extremity, we observed that Onyx passed through the high-flow lesion into one of the main drainage veins. However, we did not observe pulmonary embolism because Onyx was polymerized along the proximal subclavian vein before it reached the pulmonary bed. During the follow-up period, the patient reported no side effects and there were no signs of venous thrombosis in the affected extremity. Considering this observation, the risk of pulmonary embolism seems to be much lower with Onyx than with PVA particles and NBCA. Controlled and slow injection of embolic agent and/or reduction of flow through the VM with external compression to the involved extremity are usually helpful to stagnate the high flow and prevent venous reflux. Onyx with higher concentrations of copolymer (8% or 20%) should also be used for embolization of high-flow fistulous components. In patients 7, 8, and 9, we used external compression (a tourniquet was administered to the involved extremity), which caused stagnation of flow. In addition to this maneuver, the use of 8% and 20% Onyx allowed us to embolize the fistulous components because of its faster polymerization. In patient 7, we used external compression and 8% Onyx to occlude the neck of arterial feeder aneurysms of the VM (Fig 1b,c). In case 9, 20% Onyx allowed us to occlude the fistulous components of the lesion.
Numan et al
Arterial and/or venous reflux did not occur in any patients with lowflow VMs. We did not experience any nontarget embolization in this type of VM. Compared with low-flow VMs, the risk of arterial reflux and nontarget embolization is considerably higher in high-flow VMs. Because VMs are dynamic entities, their complex morphology and hemodynamic status may change in time, rendering them untreatable by endovascular means. In some VMs, whether high- or low-flow type, there are multiple tiny short arterial (parasitic) feeders. Also, especially during endovascular treatment of high-flow VMs, after occlusion of the major feeders, only parasitic feeders were left behind, which makes selective catheterization impossible. In patient 8, another session was planned after two sessions of embolization, but control angiograms revealed that the lesion was still filled, mainly by parasitic feeders. Therefore, we decided not to perform repeat embolization and simply perform follow-up with the patient. The pain during injection is a disadvantage of the use of Onyx, as we observed in all cases. The pain is probably caused by DMSO, as all of the patients reported pain during the injection of DMSO. To perform more effective and comfortable embolization, it is necessary to use moderate sedation or, if needed, general anesthesia. Because of the hemodynamic properties and ever-changing morphology of VMs, total occlusion of them, especially the high-flow types, is very difficult to achieve. Therefore, the interventional radiologist should focus on palliation rather than eradication of these lesions. Peripheral VMs may have severe outcomes like inability to use the involved extremity, cardiac failure resulting from high cardiac output, risk of posttraumatic hemorrhage, and cosmetic/psychological problems in young patients. The primary goal of endovascular treatment should be to minimize these severe outcomes of VMs and to maximize the patient’s quality of life. In our study, although VMs in seven patients were embolized incompletely, six showed signs of clinical improvement. In this aspect, embolization with Onyx was very helpful in two of our patients with diffuse high-flow VMs. They had
•
945
experienced pain, swelling, disabling of the involved extremity, skin changes, and cardiac failure, and although embolization could not occlude the lesions completely, flow through the lesions was moderately reduced and cardiac function improved in both patients. In one of them (patient 9), swelling of the involved extremity regressed significantly and a better cosmetic result was achieved. In five patients with incompletely occluded local VMs, embolization was clinically successful to varying degrees. In one patient (patient 6; Table 2), embolization was clinically unsuccessful. The major disadvantage of Onyx is its high cost, which increases with higher concentrations of copolymer. It restricts the availability of Onyx compared with other embolic agents like PVA or NBCA, which are less expensive. In conclusion, long polymerization and long injection times, which enable good nidal penetration and more controlled administration, are important advantages of Onyx over other liquid embolic agents. However, like other embolic agents, Onyx is far from perfect. In our experience, Onyx may not have a revolutionary contribution in the treatment of diffuse high-flow VMs, but it can provide improvement in clinical status. It can also be used effectively in local low-flow VMs. More studies with larger patient populations are needed to determine the exact role of this new liquid embolic agent in endovascular transcatheter embolization of peripheral VMs. References 1. Raso AM, Rispoli P, Trogolo M, Sisto G, Castagno PL. Venous and arteriovenous vascular malformations: diagnostic and therapeutic considerations regarding 239 patients observed in the 1978 –1991 period. J Cardiovasc Surg 1993; 34:63– 65. 2. White RI Jr, Pollak J, Persing J, Henderson KJ, Thomson JG, Burdge CM. Long-term outcome of embolotherapy and surgery for high-flow extremity arteriovenous malformations. J Vasc Interv Radiol 2000; 11:1285–1295. 3. Jahan R, Murayama Y, Gobin YP, Duckwiler GR, Vinters HV, Vinuela F. Embolization of arteriovenous malformations with Onyx: clinicopathological experience in 23 patients. Neurosurgery 2001; 48:984 –995. 4. Murayama Y, Vinuela F, Ulhoa A, et al.
946
•
Embolization of Peripheral Vascular Malformations with Onyx
Nonadhesive liquid embolic agent for cerebral arteriovenous malformations: preliminary histopathological studies in swine rete mirabile. Neurosurgery 1998; 43:1164 –1175. 5. Warakaulle DR, Aviv RI, Niemann D, Molyneux AJ, Byrne JV, Teddy P. Embolisation of spinal dural arteriovenous fistulae with Onyx. Neuroradiology 2003; 45:110 –112. 6. Cantasdemir M, Kantarci F, Mihmanli I, Numan F. Embolization of profunda femoris artery branch pseudoaneurysms with ethylene vinyl alcohol copolymer (Onyx). J Vasc Interv Radiol 2002; 13:725–728. 7. Castaneda F, Goodwin SC, Swischuk JL, et al. Treatment of pelvic arteriovenous malformations with ethylene vinyl alcohol copolymer (Onyx). J Vasc Interv Radiol 2002; 13:513–516.
8. Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: A classification based on endothelial characteristics. Plast Reconstr Surg 1982; 69:412– 422. 9. Jackson IT, Carreno R, Potparic Z, Hussain K. Hemangiomas, vascular malformations, and lymphovenous malformations: classification and methods of treatment. Plast Reconstr Surg 1993; 91:1216 –1230. 10. Widlus DM, Murray RR, White RI, et al. Congenital arteriovenous malformations: tailored embolotherapy. Radiology 1988; 169:511–516. 11. Schönholz CJ, Mendaro E, Sergio S, Hurvitz D. Embolization in peripheral territory. In: Heuser RR, Henry M. Textbook of peripheral vascular interventions, first edition. London: Martin Dunitz, 2004;447– 456.
September 2004
JVIR
12. Pollak JS, White RI Jr. The use of cyanoacrylate adhesives in peripheral embolization. J Vasc Interv Radiol 2001; 12:907–913. 13. Barr JD, Hoffman EJ, Davis BR, Edgar KA, Jacobs CR. Microcatheter adhesion of cyanoacrylates: comparison of normal butyl cyanoacrylate to 2-hexyl cyanoacrylate. J Vasc Interv Radiol 1999; 10:165–168. 14. Suh JS, Shin KH, Na JB, Won JY, Hahn SB. Venous malformations: sclerotherapy with a mixture of ethanol and lipiodol. Cardiovasc Intervent Radiol 1997; 20:268 –273. 15. Taki W, Yonekawa Y, Iwata H, Uno A, Yamashita K, Amemiya H. A new liquid material for embolization of arteriovenous malformations. AJNR Am J Neuroradiol 1990; 11:163–168.
PURPOSE: To demonstrate the feasibility and preliminary efficacy of endovascular embolization of peripheral congenital vascular malformations (VMs) with use of a nonadhesive liquid embolic agent, Onyx. MATERIALS AND METHODS: Nine patients with a mean age of 20.8 years had local low-flow (n ⴝ 4), local high-flow (n ⴝ 3), or diffuse high-flow (n ⴝ 2) VMs located in the upper or lower extremities. In all patients, endovascular embolization was performed via the superselective catheterization of arterial feeders of VMs with use of microcatheters in a coaxial technique. A total of 15 embolization procedures were performed with Onyx, which was composed of 6%, 8%, or 20% ethylene vinyl alcohol copolymer dissolved in dimethyl sulfoxide. RESULTS: In two of four patients with local low-flow VMs, the lesions were embolized completely. In the other two patients with local low-flow VMs, embolizations were incomplete. The remaining five high-flow lesions of local (n ⴝ 3) or diffuse (n ⴝ 2) types were also embolized incompletely. In all patients with local low-flow VMs and in one patient with a local high-flow VM, clinical signs and symptoms were resolved significantly. Other patients did show clinical benefit from embolization to varying degrees. CONCLUSION: In our experience in a limited number of cases, Onyx promises and provides important advantages over conventional embolic agents in the endovascular transcatheter embolization of congenital peripheral VMs. However, as with other embolic agents, it is far from perfect. J Vasc Interv Radiol 2004; 15:939 –946 Abbreviations:
DMSO ⫽ dimethyl sulfoxide, NBCA ⫽ n-butyl cyanoacrylate, PVA ⫽ polyvinyl alcohol, VM ⫽ vascular malformation
CONGENITAL peripheral vascular malformations (VMs) are uncommon lesions that can present a myriad of clinical presentations. Surgery is usually unfeasible or inefficient in the treatment of peripheral VMs (1,2). En-
From the Department of Radiology (F.N., B.K., M.C., F.K.), 3400, Cerrahpasa Medical Faculty, Istanbul University, Istanbul; Department of Radiology, Izzet Baysal Medical Faculty (A.Ö.), Abant Izzet Baysal University, Bolu; and Department of Radiology (I˙.A.), Gaziantep University Medical Faculty, Gaziantep, Turkey. Received December 23, 2003; revision requested February 18, 2004; final revision received April 1; accepted April 12. Address correspondence to F.N.; E-mail: [email protected] None of the authors have identified a conflict of interest. © SIR, 2004 DOI: 10.1097/01.RVI.0000130862.23109.52
dovascular transcatheter embolization has been suggested as the primary therapeutic option or as a presurgical intervention to reduce bleeding and maximize successful resection. The agents available for endovascular transcatheter embolization include polyvinyl alcohol (PVA) particles, absolute alcohol, acrylic polymers, coils, and n-butyl cyanoacrylate (NBCA). The Onyx Liquid System (Embolyx; Micro Therapeutics, Irvine, CA) is a new nonadhesive liquid embolic agent that is currently being evaluated in the treatment of cerebral arteriovenous VMs (3–5). Peripheral use of Onyx has been published in two case reports (6,7), one in traumatic pseudoaneurysms and the other in treatment of pelvic arteriovenous VMs. The aim of this study is to report the feasibility
and preliminary efficacy of endovascular embolization of peripheral congenital VMs with use of Onyx.
MATERIALS AND METHODS Patient Population From October 2001 to August 2003, nine patients (four men, five women) with a mean age of 20.8 years (range, 11–33 years) underwent endovascular transcatheter arterial embolization with use of Onyx in the treatment of peripheral VMs. Table 1 summarizes patients’ age, sex, VM type and location, signs and symptoms at presentation, and previous therapeutic procedures. Of nine patients, two had diffuse high-flow VMs located at the shoulder
939
940
•
Embolization of Peripheral Vascular Malformations with Onyx
September 2004
JVIR
Table 1 Summary of Cases Patient No.
Age (y)/ Sex
Lesion Type
1
18/M
2
11/M
3
12/F
4
33/M
5
30/F
6
16/F
7
27/F
8
15/M
Low flow (local) Low flow (local) Low flow (local) Low flow (local) High flow (local) High flow (local) High flow (local) High flow (diffuse)
9
26/F
High flow (diffuse)
Lesion Location Right gluteus
Signs and Symptoms
Previous Treatment –
Left calf
Pain, swelling, varicosity, erythema, discoloration Pain in activity, discoloration, erythema
Right thigh
Pain, swelling, discoloration
–
Right gluteus
Pain, swelling, varicosity, erythema, discoloration Pain, swelling, varicosity, erythema, discoloration Pain, swelling, erythema, discoloration, scaling, ulceration, disabling of extremity Pain, swelling, varicosity, discoloration, thrill, congestive heart failure Pain, swelling, varicosity, discoloration, erythema, scaling, disabling of extremity, thrill, congestive heart failure Pain, swelling, varicosity, discoloration, erythema, scaling, disabling of extremity, thrill, congestive heart failure
Left forearm Right calf Left arm Left shoulder and upper extremity Left shoulder and upper extremity
and upper extremity. Previous unsuccessful surgical management was performed in one of these patients and several endovascular transcatheter embolization procedures were performed previously with use of PVA particles in the other one. Seven of nine patients had local VMs, which were of low-flow type in four and high-flow type in three. Two patients with local high-flow VMs had undergone previous treatment. One of them had unsuccessful surgery and the other had previous endovascular transcatheter embolization with use of NBCA. One patient with a local lowflow VM had undergone surgery and then embolization with NBCA and PVA particles. Signs and symptoms at presentation included pain, swelling, multiple varicosities, thrill, trophic changes of the skin, functional impairment of the involved extremities, high cardiac output, and congestive cardiac failure (Table 1). Embolization Technique All patients underwent diagnostic angiography via the femoral artery route under local anesthesia just before the embolization procedures. Catheterization was performed via a transfemoral approach with use of
standard coaxial techniques. A dimethyl sulfoxide (DMSO)– compatible microcatheter/microwire combination (Rebar microcatheter and Silver Speed micro– guide wire; Micro Therapeutics) was used to perform superselective catheterization of the feeding arteries. Onyx embolization required flushing of the microcatheter with saline solution first and then with 0.4 mL DMSO to fill the microcatheter’s “dead space.” After that, the Onyx was aspirated into a 1-mL syringe, and 0.25 mL of this amount was injected slowly for 40 seconds to fill the microcatheter and replace the DMSO in the dead space. Finally, Onyx was injected at a volume and rate low enough to prevent reflux of the agent around the microcatheter under fluoroscopic guidance. The injection was maintained and repeated until the distal end of the feeding arteries and the abnormal malformed vascular communications were completely occluded. Because the estimation of injection volume of Onyx is very difficult, especially in high-flow lesions, as a result of the complex structure of the abnormal vascular communications, the volume and rate of Onyx to be injected was determined only by fluoroscopic monitoring during injection. We did not perform an occlusion injection test of contrast material to determine the
–
Surgery and PVA/NBCA embolization Surgery NBCA embolization – PVA embolization Surgery
volume of Onyx before the injection. The reliable criteria used to determine the volume and rate of injection were based on the penetration of the Onyx to the abnormal vascular communications and avoidance of reflux. Because Onyx is not adhesive, multiple injections could be performed. In multiple injections, we used the same catheter to catheterize another feeding artery, provided that the microcatheter had been withdrawn first and then flushed with DMSO and physiologic saline solution again before each catheterization process. Onyx is supplied in ready-to-use vials. Each vial contains ethylene vinyl alcohol copolymer, DMSO, and tantalum powder for radiopacity. Ethylene vinyl alcohol copolymer is formed from 48 mol/L ethylene and 52 mol/L vinyl alcohols. The polymer is dissolved in DMSO and prepared in different concentrations. Onyx 6.0% contains 6.0% copolymer and 94% DMSO, Onyx 8.0% contains 8.0% copolymer and 92% DMSO, and Onyx 20% contains 20% copolymer and 80% DMSO. The lower the concentration of the copolymer, the less viscous the agent and the more distal penetration can be achieved. In most patients, we used Onyx 6% and Onyx 8% to achieve more distal penetration. However, for a fistulous component of a diffuse
Volume 15
Number 9
high-flow VM in one patient (patient 9), we required fast polymerization and therefore used Onyx 20%. Systemic anticoagulation was achieved during the procedures with intraarterial heparin in 5,000-IU bolus injections. For an effective and comfortable intervention, the pain during embolizations was primarily eliminated by analgesics and/or sedatives. General anesthesia should be reserved for children and patients who would not be able to tolerate the pain even with administration of analgesics and/or sedatives. Study Endpoints and Definitions Diffuse VMs were defined by involvement of the entire extremity, whereas local VMs were defined by involvement of a restricted area of an extremity. VMs were also subdivided into high-flow and low-flow groups according to the speed of flow through the lesion and the rate of shunting and/or abnormal communication between the arterial and venous components. In high-flow lesions, arteries and veins were visualized through the arterial phase of the arteriogram. If the venous component of the lesion appeared after the arterial phase, the lesion was defined as a low-flow VM. Technical success was defined as complete if the postembolization arteriograms revealed no flow in the VM. In cases of incomplete technical success, flow through the abnormal vascular communications of VMs reduced but the filling of the VM still remained. We further defined incomplete success qualitatively as having moderately or significantly reduced flow through a VM. Clinical success was defined as total or partial improvement of clinical signs and symptoms such as pain, swelling, erythema, discoloration, scaling, ulceration, disabling of extremity, thrill, and congestive heart failure. An unsuccessful clinical result was defined by persistence of clinical signs or additional symptoms related to ischemic complications secondary to nontarget embolization. Major complications were defined as nontarget embolization with acute ischemia and pulmonary embolism. Minor complications included post-
Numan et al
embolization syndrome, gluing of the catheter to the vessel wall, and vasospasm related to the angiotoxicity of DMSO. Our institutional guidelines do not require institutional review board approval in these circumstances, and Declaration of Helsinki principles were followed.
RESULTS Number of embolization procedures, amount of injected Onyx, complications, angiographic results of embolization procedures, clinical outcome, and length of clinical follow-up are given in Table 2. A total of 15 embolization procedures were performed in nine patients. In five cases, embolization procedures were performed once; in three cases, it was performed two times; and in one case, it was performed four times. Multiple injections in single embolization procedures were performed in most cases, and several arterial feeders were embolized during these 15 embolization procedures. In most embolization procedures, the same microcatheter was used for repeat catheterization and injection of the embolic agent. In 15 embolization procedures, a total of 69 mL of Onyx was used. The mean volume of injected Onyx in each procedure was 4.6 mL. Postembolization angiograms revealed total occlusion in two patients with local lowflow VMs. In the remaining two patients with local low-flow VMs, embolizations were incomplete but achieved significantly reduced flow through the malformation. In all patients with high-flow VMs (diffuse, n ⫽ 2; local, n ⫽ 3), embolizations were also incomplete. In three of these cases, we achieved moderately reduced flow through the VM. In two patients with local high-flow VMs, a postembolization angiogram showed significantly reduced flow through the lesion. One of these cases was clinically unsuccessful as a result of nontarget embolization and the other was clinically successful (Fig 1). In all patients with local low-flow VMs (n ⫽ 4), clinical signs and symptoms were significantly resolved as shown in Table 2. In one of two patients with completely occluded local low-flow VMs (Fig 2), only mild discoloration persisted on follow-up and
•
941
other symptoms disappeared (patient 1). All symptoms were resolved in the other patient (patient 2). In one of two patients (patient 4) with incompletely occluded local low-flow VMs, only mild discoloration and mild swelling were observed during the follow-up period, whereas the other patient (patient 3) was free of symptoms. In a patient with an incompletely occluded local high-flow VM (patient 5), embolization relieved the pain but other symptoms persisted. Patient 6, who had an incompletely occluded local high-flow VM, reported that symptoms had worsened following the procedure, likely as a result of nontarget embolization, despite significantly reduced flow through the lesion. In patient 7, who had an incompletely occluded local high-flow VM, all symptoms regressed. In two patients with incompletely occluded diffuse high-flow VMs (patients 8 and 9), we observed improvement of congestive heart failure, and only mild discoloration and varicosity remained. There were no complications related to the procedure in six patients (four with local low-flow VMs, one with diffuse high-flow VM, and one with local high-flow VM). In the remaining three patients, who had highflow VMs (local, n ⫽ 2; diffuse, n ⫽ 1), reflux of Onyx caused nontarget embolization. In patient 5, reflux to the ulnar artery resulted in occlusion of the artery without ischemia. In patient 6, reflux to the anterior and posterior tibial arteries caused peripheral ischemia and worsening of skin condition as a result of occlusion of these arteries. In patient 9, who had diffuse involvement of the left upper extremity, we observed transient ischemia and worsening of skin condition as a result of minimal arterial reflux of the embolic agent to the brachial artery. Patients 5 and 9 underwent conservative management. Patient 6 was referred to undergo plastic surgery and treated with total excision of the lesion and reconstruction with free latissimus dorsi muscle transfer. The patients with partially successful procedures were followed up and evaluated with physical examination during follow-up visits. No clinical progression was detected in any of these patients. No instances of gluing of the catheter to the vessel wall occurred. There was no vasospasm related to angiotox-
942
•
Embolization of Peripheral Vascular Malformations with Onyx
September 2004
JVIR
Table 2 Results of Onyx Embolization Onyx No. of Onyx Volume Embolizations Concentration (mL)
Patient No.
Lesion Type
1
Low flow (local)
1
6%
7.5
–
Complete
2
Low flow (local) Low flow (local)
1
6%
3
–
Complete
1
6%
3
–
4
Low flow (local)
2
6%
6
5
High flow (local)
2
6%
9
6
High flow (local)
1
6%
6
7
High flow (local)
1
6% 8%
3 1.5
8
High flow (diffuse)
2
6%
9
High flow (diffuse)
4
6% 8% 20%
3
12
9 7.5 1.5
Complications
Embolization Results
Clinical Results Only mild discoloration remained Complete success
Incomplete Complete success (significantly reduced flow) Only mild – Incomplete discoloration and (significantly swelling remained reduced flow) Only pain resolved Reflux resulting Incomplete (moderately in ulnar reduced artery flow) occlusion without ischemia or nerve palsy Reflux resulting Incomplete Unsuccessful in ATA and (significantly PTA reduced occlusion, flow) peripheral ischemia and worsening of skin change Complete success – Incomplete (significantly reduced flow) Improvement of – Incomplete congestive heart (moderately failure; mild reduced discoloration and flow) varicosity remained Improvement of Incomplete Reflux to congestive heart (moderately brachial failure; mild reduced artery discoloration and flow) resulting in varicosity remained transient peripheral ischemia and worsening of skin changes
Follow-up (mo) 18 18 25
15
17
18
3
21
28
ATA ⫽ anterior tibial artery; PTA ⫽ posterior tibial artery.
icity of DMSO. In the follow-up period, we did not observe any pulmonary embolism in these cases.
DISCUSSION There is a wide variety of classifications for VMs based on clinical, histopathologic, and hemodynamic prop-
erties. The classification introduced by Mulliken and Glowacki (8) was based on the clinical and histologic characteristics of VMs. They categorized these lesions as hemangiomas, and VMs consisted of arterial, capillary, venous, or lymphatic channels, or any combination of these components. However, this classification system is
not useful for determination of therapeutic options. To select the accurate treatment method, Jackson et al (9) classified VMs according to the vascular dynamics of the lesion, ie, low flow and high flow. These terms indicate the speed of flow through the lesion and the rate of shunting between the arterial and venous components. We
Volume 15
Number 9
Numan et al
•
943
Figure 1. (a) Selective left subclavian artery injection demonstrates a local high-flow VM in the left arm of patient 7. (b) Brachial artery injection depicts an arterial feeder aneurysm. (c) Postembolization arteriography shows Onyx cast at the neck of the aneurysm and total exclusion of the aneurysm from the circulation. (d) Postembolization arteriography reveals significantly reduced flow through the VM.
Figure 2. (a) Right internal iliac arteriogram shows a local right posteromedial gluteal low-flow VM in patient 1. (b) A postembolization selective right internal iliac arteriogram demonstrates total occlusion of the lesion. Note the Onyx casts and vasospasm caused by catheter manipulation.
used this categorization based on hemodynamic properties of VMs. The management of congenital VMs can present a difficult therapeutic
challenge. Surgical treatment is notoriously difficult and carries a high likelihood of recurrence (1). Endovascular transcatheter embolization of periph-
eral VMs has become more widely accepted as a first therapeutic option for many VMs (2,10). Endovascular management of pe-
944
•
Embolization of Peripheral Vascular Malformations with Onyx
ripheral VMs begins with finding the appropriate embolic agent to match the morphology and hemodynamic status of the lesion. The ideal embolic agent would be nonclumping, highly radiopaque, nontoxic, nonallergenic, and inexpensive. It would allow easily controlled injection through a microcatheter and would provide permanent occlusion. Most important, it should have the ability to penetrate and occlude the abnormal foci of vascular communications in VMs. Because embolization coils and detachable balloons provide proximal occlusion and cannot penetrate the foci, their role in the endovascular management of complex VMs is limited. PVA particles may be used in some cases, but determining the appropriate size of particles to use is difficult, leading to the risk of pulmonary embolism. Recanalization or development of arterial collaterals are other possible problems with PVA embolization (11). Unlike particles, which may be arrested at a precapillary level and cannot always penetrate the abnormal foci of vascular communications in VMs because of their size, liquid embolic agents can pass to the capillary level and penetrate deep into foci. This feature of liquid embolic agents makes them more feasible in the endovascular treatment of VMs. Several liquid embolic agents are available for the endovascular transcatheter embolization of peripheral VMs. The main liquid embolic agents used in endovascular procedures are cyanoacrylates. NBCA is the preferred agent among them and is often referred to as glue (12). The substance polymerizes immediately on contact with blood or other ionic fluids. Polymerization of NBCA results in an exothermic reaction that destroys the vessel wall. Because of rapid polymerization of NBCA when in contact with blood, precise and safe occlusion can be difficult to achieve. In addition, after the injection of NBCA, the microcatheter should be removed quickly so that the catheter tip does not adhere to the vessel wall (12,13). Other disadvantages of NBCA are that the microcatheter cannot be reused during the same procedure and transembolization angiograms cannot be performed. Another liquid embolic agent, absolute alcohol (ethanol), is also widely
used to treat certain types of VMs. It is usually preferred as a sclerosing agent for the treatment of predominantly venous lesions by direct percutaneous injections. However, it can be damaging if it reaches the capillary bed of skin, and it may cause significant softtissue swelling, which may subsequently lead to compartment syndrome. Ethanol has a direct toxic effect on the endothelium that activates the coagulation system and causes occlusion of the vessel lumen by means of thrombosis. As a result of its progressive and permanent occlusion effect, it has also been used in the endovascular management of VMs. However, if large amounts of absolute alcohol enter the systemic circulation, toxic effects including central nervous system depression, hemolysis, or cardiac arrest may occur (14). The physical and biologic properties of embolic agents affect their level of occlusion, which constitutes the heart of the matter in endovascular treatment of peripheral VMs. Traditionally, several embolic agents have been used in the endovascular treatment of peripheral VMs, but they are not always able to penetrate the abnormal foci of vascular communications. Onyx is a new biocompatible liquid embolic agent consisting of ethylene vinyl alcohol copolymer dissolved in various concentrations of DMSO. Micronized tantalum powder, added for radiopacity, is used in higher concentrations in the latest generations of Onyx. When this mixture contacts aqueous media such as blood, DMSO rapidly diffuses from the mixture, causing in situ precipitation and solidification of the polymer, with formation of an elastic soft, spongy embolus without adhesion to the vascular wall (3,4). This polymerization process is time-dependent and mainly affected by the amount of ethylene within the copolymer. As one decreases the percentage of ethylene, the polymer becomes softer and spongier and exhibits a different solidification rate and behavior in blood (6). In addition, concentration of the copolymer dissolved in DMSO determines the viscosity of the embolic agent, which is especially important for choosing the proper concentration of copolymer during a procedure. The lower the concentration of the copolymer, the less viscous the agent is and the more distal pene-
September 2004
JVIR
tration can be achieved in low-flow lesions. In high-flow VMs and for the fistulous components of VMs, higher concentrations of copolymer should be preferred for fast polymerization and to avoid venous reflux. Because Onyx is mechanically occlusive but not adherent to vascular tissues, theoretically, there is no risk of catheter gluing to the arterial wall. Another technical advantage of Onyx is the possibility to perform transembolization angiography, which enables assessment of the requirement for additional embolic agent. The biologic and technical advantages of Onyx over other liquid embolic agents, and the reported results in the endovascular treatment of cerebral arteriovenous VMs with Onyx, encouraged us to use it in embolization of peripheral VMs. As we detected in our cases, one of the advantages of this embolic agent was its ability to conform to the shape of the tortuous arteries supplying the VM. This property provided good penetration to the abnormal foci of vascular communications, which is one of the major issues of endovascular treatment. Onyx does not precipitate in the presence of its solvent, DMSO. Preloading of the microcatheters with DMSO before Onyx injection provides longer injection time and more controllable embolization. Another advantage of Onyx over tissue adhesives is preservation of the option for subsequent surgical resection of the VM. Because Onyx is nonadhesive and maintains the vessel wall integrity, surgical removal of the VM after Onyx embolization seems to be easier than with NBCA (3,15). The nontarget embolization caused by reflux to the parent artery was the most frequent problem we encountered in the endovascular treatment of peripheral VMs with Onyx. This complication occurred in patients with high-flow VMs (patients 5, 6 and 9; Table 2). Regarding our experience in these cases, we think the reasons for the reflux were the complex unpredictable angioarchitecture of highflow lesions, short arterial feeders close to the parent artery, and poor radiopacity of the former generation of Onyx. These features made it difficult to decide the proper injection rate and amount of Onyx to be used. A test injection with contrast media before
Volume 15
Number 9
delivery of Onyx is not always helpful in the estimation of the threshold for reflux as a result of viscosity differences between Onyx and contrast media and hemodynamic changes during the occlusion of the feeder. During the final stage of the injection, the pressure in the feeding artery possibly increases to a critical threshold. After that point, Onyx may easily reflux to the parent artery. In addition to the parent artery reflux, there is always the theoretical risk of venous reflux, especially in high-flow VMs. If Onyx passes through the high-flow fistulous component of the VM before polymerization, it may result in venous reflux. We observed that the refluxed Onyx in the venous site tends to move but stay along the vessel wall. This may be because of its viscosity and laminar flow. In patient 7, who had a local VM in the left upper extremity, we observed that Onyx passed through the high-flow lesion into one of the main drainage veins. However, we did not observe pulmonary embolism because Onyx was polymerized along the proximal subclavian vein before it reached the pulmonary bed. During the follow-up period, the patient reported no side effects and there were no signs of venous thrombosis in the affected extremity. Considering this observation, the risk of pulmonary embolism seems to be much lower with Onyx than with PVA particles and NBCA. Controlled and slow injection of embolic agent and/or reduction of flow through the VM with external compression to the involved extremity are usually helpful to stagnate the high flow and prevent venous reflux. Onyx with higher concentrations of copolymer (8% or 20%) should also be used for embolization of high-flow fistulous components. In patients 7, 8, and 9, we used external compression (a tourniquet was administered to the involved extremity), which caused stagnation of flow. In addition to this maneuver, the use of 8% and 20% Onyx allowed us to embolize the fistulous components because of its faster polymerization. In patient 7, we used external compression and 8% Onyx to occlude the neck of arterial feeder aneurysms of the VM (Fig 1b,c). In case 9, 20% Onyx allowed us to occlude the fistulous components of the lesion.
Numan et al
Arterial and/or venous reflux did not occur in any patients with lowflow VMs. We did not experience any nontarget embolization in this type of VM. Compared with low-flow VMs, the risk of arterial reflux and nontarget embolization is considerably higher in high-flow VMs. Because VMs are dynamic entities, their complex morphology and hemodynamic status may change in time, rendering them untreatable by endovascular means. In some VMs, whether high- or low-flow type, there are multiple tiny short arterial (parasitic) feeders. Also, especially during endovascular treatment of high-flow VMs, after occlusion of the major feeders, only parasitic feeders were left behind, which makes selective catheterization impossible. In patient 8, another session was planned after two sessions of embolization, but control angiograms revealed that the lesion was still filled, mainly by parasitic feeders. Therefore, we decided not to perform repeat embolization and simply perform follow-up with the patient. The pain during injection is a disadvantage of the use of Onyx, as we observed in all cases. The pain is probably caused by DMSO, as all of the patients reported pain during the injection of DMSO. To perform more effective and comfortable embolization, it is necessary to use moderate sedation or, if needed, general anesthesia. Because of the hemodynamic properties and ever-changing morphology of VMs, total occlusion of them, especially the high-flow types, is very difficult to achieve. Therefore, the interventional radiologist should focus on palliation rather than eradication of these lesions. Peripheral VMs may have severe outcomes like inability to use the involved extremity, cardiac failure resulting from high cardiac output, risk of posttraumatic hemorrhage, and cosmetic/psychological problems in young patients. The primary goal of endovascular treatment should be to minimize these severe outcomes of VMs and to maximize the patient’s quality of life. In our study, although VMs in seven patients were embolized incompletely, six showed signs of clinical improvement. In this aspect, embolization with Onyx was very helpful in two of our patients with diffuse high-flow VMs. They had
•
945
experienced pain, swelling, disabling of the involved extremity, skin changes, and cardiac failure, and although embolization could not occlude the lesions completely, flow through the lesions was moderately reduced and cardiac function improved in both patients. In one of them (patient 9), swelling of the involved extremity regressed significantly and a better cosmetic result was achieved. In five patients with incompletely occluded local VMs, embolization was clinically successful to varying degrees. In one patient (patient 6; Table 2), embolization was clinically unsuccessful. The major disadvantage of Onyx is its high cost, which increases with higher concentrations of copolymer. It restricts the availability of Onyx compared with other embolic agents like PVA or NBCA, which are less expensive. In conclusion, long polymerization and long injection times, which enable good nidal penetration and more controlled administration, are important advantages of Onyx over other liquid embolic agents. However, like other embolic agents, Onyx is far from perfect. In our experience, Onyx may not have a revolutionary contribution in the treatment of diffuse high-flow VMs, but it can provide improvement in clinical status. It can also be used effectively in local low-flow VMs. More studies with larger patient populations are needed to determine the exact role of this new liquid embolic agent in endovascular transcatheter embolization of peripheral VMs. References 1. Raso AM, Rispoli P, Trogolo M, Sisto G, Castagno PL. Venous and arteriovenous vascular malformations: diagnostic and therapeutic considerations regarding 239 patients observed in the 1978 –1991 period. J Cardiovasc Surg 1993; 34:63– 65. 2. White RI Jr, Pollak J, Persing J, Henderson KJ, Thomson JG, Burdge CM. Long-term outcome of embolotherapy and surgery for high-flow extremity arteriovenous malformations. J Vasc Interv Radiol 2000; 11:1285–1295. 3. Jahan R, Murayama Y, Gobin YP, Duckwiler GR, Vinters HV, Vinuela F. Embolization of arteriovenous malformations with Onyx: clinicopathological experience in 23 patients. Neurosurgery 2001; 48:984 –995. 4. Murayama Y, Vinuela F, Ulhoa A, et al.
946
•
Embolization of Peripheral Vascular Malformations with Onyx
Nonadhesive liquid embolic agent for cerebral arteriovenous malformations: preliminary histopathological studies in swine rete mirabile. Neurosurgery 1998; 43:1164 –1175. 5. Warakaulle DR, Aviv RI, Niemann D, Molyneux AJ, Byrne JV, Teddy P. Embolisation of spinal dural arteriovenous fistulae with Onyx. Neuroradiology 2003; 45:110 –112. 6. Cantasdemir M, Kantarci F, Mihmanli I, Numan F. Embolization of profunda femoris artery branch pseudoaneurysms with ethylene vinyl alcohol copolymer (Onyx). J Vasc Interv Radiol 2002; 13:725–728. 7. Castaneda F, Goodwin SC, Swischuk JL, et al. Treatment of pelvic arteriovenous malformations with ethylene vinyl alcohol copolymer (Onyx). J Vasc Interv Radiol 2002; 13:513–516.
8. Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: A classification based on endothelial characteristics. Plast Reconstr Surg 1982; 69:412– 422. 9. Jackson IT, Carreno R, Potparic Z, Hussain K. Hemangiomas, vascular malformations, and lymphovenous malformations: classification and methods of treatment. Plast Reconstr Surg 1993; 91:1216 –1230. 10. Widlus DM, Murray RR, White RI, et al. Congenital arteriovenous malformations: tailored embolotherapy. Radiology 1988; 169:511–516. 11. Schönholz CJ, Mendaro E, Sergio S, Hurvitz D. Embolization in peripheral territory. In: Heuser RR, Henry M. Textbook of peripheral vascular interventions, first edition. London: Martin Dunitz, 2004;447– 456.
September 2004
JVIR
12. Pollak JS, White RI Jr. The use of cyanoacrylate adhesives in peripheral embolization. J Vasc Interv Radiol 2001; 12:907–913. 13. Barr JD, Hoffman EJ, Davis BR, Edgar KA, Jacobs CR. Microcatheter adhesion of cyanoacrylates: comparison of normal butyl cyanoacrylate to 2-hexyl cyanoacrylate. J Vasc Interv Radiol 1999; 10:165–168. 14. Suh JS, Shin KH, Na JB, Won JY, Hahn SB. Venous malformations: sclerotherapy with a mixture of ethanol and lipiodol. Cardiovasc Intervent Radiol 1997; 20:268 –273. 15. Taki W, Yonekawa Y, Iwata H, Uno A, Yamashita K, Amemiya H. A new liquid material for embolization of arteriovenous malformations. AJNR Am J Neuroradiol 1990; 11:163–168.