Current Controversies and the State of the Art in Endovascular Treatment of Vascular Malformations

Vascular anomalies constitute some of the most difficult diagnostic and therapeutic enigmas that can be encountered in the practice of medicine. The clinical presentations are extremely protean and can range from an asymptomatic birthmark to fulminant, life-threatening congestive heart failure. Attributing any of these extremely varied symptoms that a patient may present with to a vascular malformation can be challenging to the most experienced clinician. Compounding this problem is the extreme rarity of these vascular lesions. If a clinician sees one patient every few years, it is extremely difficult to gain a learning curve to diagnose and optimally treat them. Typically, these patients bounce from clinician to clinician only to experience disappointing outcomes, complications, and recurrence or worsening of their presenting symptoms. Vascular malformations are truly an “orphan” disease in the medical world.

Vascular anomalies were first treated by surgeons. The early rationale of proximal arterial ligation of arteriovenous malformations (AVMs) proved totally futile as the phenomenon of neovascular recruitment reconstituted arterial inflow to the AVM nidus. Microfistulous connections became macrofistulous feeders. Complete extirpation of an AVM nidus proved very difficult and extremely hazardous, necessitating suboptimal partial resections. Partial resections could cause an initial good clinical response, but with time, the patient's presenting symptoms recurred or worsened at follow-up (1-3). Because of the significant blood loss that frequently accompanied surgery, the skills of Interventional Radiologists were eventually employed to embolize these vascular lesions preoperatively. This allowed more complete resections; however, complete extirpation of an AVM was still extremely difficult and rarely possible. As catheter delivery systems and embolic agents improved, embolotherapy has since emerged as a primary mode of therapy in the management of vascular anomalies. Anatomically, vascular malformations are often in surgically difficult or inaccessible areas which have led to increased reliance on the sophisticated endosurgical skills of the Interventional Radiologist and Interventional Neuroradiologist in the management of these problematic patients.

As so accurately stated by the former Editor of the American Journal of Vascular Surgery, D. Emerick Szilagyi MD, “We intuitively thought that the only answer of a surgeon to the problem of disfiguring often noisome, and occasionally disabling blemishes and masses, prone to cause bleeding, pain, or other unpleasantness, was to attack them with vigor and with the determination of eradicating them. The results of this attempt at radical treatment were disappointing.” Indeed, of the 82 patients in this patient series only 18 were operated upon. And in these 18 patients 8 were the same or worse post-operatively. 64 patients were considered inoperable (2).

Because the clinical and angiographic manifestations can be extremely varied, hemangiomas and vascular malformations are always difficult to classify. Moreover, numerous descriptive terms have been applied to impressive clinical examples in the hopes of distinguishing them as distinct syndromes. This has resulted in significant confusion in the categorization and treatment of these complex vascular lesions. Some of the confusing terms include congenital arteriovenous aneurysm, cirsoid aneurysm, serpentine aneurysm, capillary telangiectasia, angioma telangiectaticum, angioma arteriale racemosum, angioma simplex, angioma serpingiosum, nevus angiectoides, hemangioma simplex, lymphangioma, hemangiolymphangioma, naevus flammeus, verrucous hemangioma, capillary hemangioma, cavernous hemangioma, and venous angioma.

Based on the landmark research of Mulliken, et al. (4-8), a rational classification of hemangioma and vascular malformations has evolved that should be incorporated into modern clinical practice. This classification system, based on endothelial cell characteristics, has removed much of the confusion in terminology that is present in the literature today. Once all clinicians understand and utilize this important classification system, ambiguity and confusion will be removed and all clinicians can speak a common language. The International Society for the Study of Vascular Anomalies (ISSVA) has adopted this classification system just for these reasons.

In the series of Ul Haq et al (9), 20 patients underwent 21 bleomycin foam embolization procedures, and six procedures involved sclerosant agents in additional to bleomycin. The authors report a per-patient complication rate of 40% (30% minor, 10% major); on a per-procedure basis, the complication rate was 29% (22% minor, 7% major). The authors state that all patients had a decrease in symptoms after treatment.

An interesting comment by the authors is that “the use of bleomycin should be reserved for locations where post procedure swelling would be dangerous.” Yet in their own small 20-patient series, one patient required intubation for 4 days as a result of tongue swelling to maintain the airway. Another patient with a tongue lesion required prolonged intubation and hospitalization for 15 days after treatment with bleomycin and ethanol. The authors state, “swelling was seen only in the area treated with alcohol, not in the area treated with bleomycin foam.” This is a difficult statement to articulate this conclusion. Having treated numerous venous, lymphatic, and mixed venous/ lymphatic lesions in the tongue over the years, I have had only two patients require overnight intubation secondary to tongue swelling. I have never had a patient require prolonged intubation (ie, 4 days or more, as in two of 20 patients in this series) as these authors report. Therefore, swelling is definitely a real possibility with bleomycin use, as swelling is a potential issue with any sclerosant agent.

In regard to the statement that one patient had tongue swelling only in the ethanol-treated part, but not the bleomycin-treated part, venous malformations (VMs) of the tongue are interconnected saccular spaces of varying sizes, and liquid embolic agents flow through endothelial cell–lined spaces and are never limited to specific areas. Injection of one area does not mean that the sclerosant agent stays put: it does flow into other contiguous areas, as does the blood within these interconnecting “tubes.” Therefore, the authors’ statement is challenging to be accurate and definitive as stated.

Other authors have reported the use of bleomycin in the treatment of lymphatic malformations (LMs) and VMs. Hassan et al (10) reported a 71% effectiveness rate and a 14% complication rate, with ulcerations in five of 75 patients. Sainsbury et al (11) reported an 82.7% effectiveness rate with bleomycin and complications that included ulcers, blistering, infections, swelling, and recurrence. In a study of pingyangmycin (bleomycin A5), Bai et al (12) reported a 43% effectiveness rate with interventional treatment alone and a 74% effectiveness rate with interventional treatment and surgery for microcystic LMs (12). Yang et al (13) reported bleomycin injections to be effective in 81% of microcystic LMs and 63% of microcystic LMs, with complications including hematoma, ulcers, fever, and soft-tissue atrophy.

Regarding articles comparing bleomycin and ethanol, Ul Haq et al (9) did not include the 2013 article by Zhang et al (14), a 138-patient series, yet referenced the 2011 article by Spence et al (15) that reported only 17 patients. In their 138-patient series, Zhang et al (14) reported that 95% of patients (71 of 75) were cured or had markedly effective or effective treatment with absolute ethanol. In the bleomycin group, 65% of treatments (41 of 63) were deemed effective (there were no cures or markedly effective treatments in the bleomycin group). The difference in these results was statistically significant, whereas the article by Spence et al (15) could not achieve any statistical significance with only 17 patients. In the large comparison study of Zhang et al (14), absolute ethanol proved markedly superior. The ethanol group had 14 cases of skin necrosis in 75 patients and the bleomycin group had five cases of skin necrosis among 63 patients. Therefore, the use of bleomycin is not immune to swelling or necrosis.

Among articles that report the use of absolute ethanol as the primary sclerosing agent for low-flow VMs, several should be mentioned. Lee et al (16) reported 87 patients, 399 ethanol procedures, and a 95% effectiveness rate of significant or complete malformation ablation, with a 12.4% complication rate. Johnson et al (17) reported the treatment of tongue VMs with a 100% success rate and no major complications. One patient who presented with breathing difficulties secondary to a tongue VM remained intubated for 5 days and was discharged uneventfully. Su et al (18) reported the treatment of head and neck VMs with absolute ethanol, with 56 of 60 patients cured (four with minimal residual VM), and no skin necrosis or nerve damage issues.

Orlando et al (19) reported a 94% rate of complete remission or improvement of symptoms in 39 patients, with 32 having no complications. One had a skin ulcer that was managed conservatively, and three had transient paresthesias. In a study of 87 patients who underwent 305 ethanol injection sessions, Lee et al (20) reported 84.5% of patients with excellent or good results, one transient facial nerve palsy, and no episodes of tissue necrosis.

In my own vascular malformation center, over a 6-year period (January 2002 to December 2007), my team performed a prospective gathering of data regarding ethanol embolizations in VMs of all types in all anatomic locations. There were 1,367 patients under-going 6,798 procedures in which a total of 133,253 mL of absolute ethanol was injected. Transient complications were seen in 5% of procedures (minor blisters, transient nerve injuries, superficial deep vein thrombosis, infections, and disseminated intravascular coagulopathy). Major complications were seen in 0.4% (deep vein thrombosis, n = 4; pulmonary embolism, n = 3; amputation, n = 2; pneumothorax, n =1; gastrointestinal bleeding, n = 1; skin graft, n = 3; permanent nerve injury, n = 1; and cardiopulmonary collapse, n = 3).

Regarding cardiopulmonary collapse with the use of ethanol, this issue has been solved by several investigators (12-15). In a 2010 prospective study, Shin et al (21) reported that, according to their treatment protocol with Swan–Ganz catheter monitoring, if the injections of ethanol do not exceed 0.14 mL/kg every 10 minutes, cardiopulmonary collapse is avoided. This series was of high-flow arteriovenous malformations (AVMs) that can require greater ethanol dose injections. Bisdorff et al (22), in a retrospective review of 71 patients undergoing 162 ethanol treatments of VMs, demonstrated that an ethanol dose/weight ratio greater than 0.24 mL/kg was predictive of systemic toxic effects. Systemic complications were not related to repetitive sclerotherapy sessions. The authors determined that a maximal dose of 0.2 mL/kg per injection could be considered acceptable in VM treatment indications. Cardiopulmonary arrest has also been reported with the use of polidocanol (23). In my institution, we studied this issue as well. With the use of Swan-Ganz catheters, we prospectively monitored 200 consecutive patients (all > 15 y of age) who underwent ethanol embolization of their vascular malformations. It was determined that, when ethanol doses of 0.1 mL/kg or less are injected every 10 minutes, increases in pulmonary artery pressures were minimal if not negligible. Therefore, adherence to this protocol has obviated Swan-Ganz catheter monitoring, and no cardiopulmonary collapse has occurred since. Observational and published prospective data with Swan-Ganz catheter monitoring revealed ethanol dose injections of 0.10–0.14 mL/kg to be without issue. In the retrospective series (24), a dose of 0.20 mL/kg proved safe. One caveat, however, requires elucidation. These dose ranges apply only to patients whose pulmonary artery pressures are in the normal range. In those patients with pulmonary artery hypertension of whatever etiology, these rules do not apply. Patients with pulmonary hypertension, unlike patients with normal pulmonary pressures, have no pulmonary pressure reserves. With the pressures already abnormally elevated, any further evaluations could mitigate cardiopulmonary collapse in an already compromised patient. Therefore, in this patient group, if ethanol embolotherapy is contemplated, Swan-Ganz catheter monitoring is mandatory. After the level of pulmonary hypertension has been determined, ethanol embolotherapy may even be omitted to prevent further elevations that may not be tolerated. If ethanol treatment is determined to be possible, it should be immediately terminated if, during monitoring, there is even a minimal increase in pulmonary artery pressures. Nitroglycerin or adenosine should be available to be administered through the Swan-Ganz catheter into the pulmonary circulation. Inhaled nitric oxide through the endotracheal tube, if available at an institution, should be present if needed by the anesthesiologist. Adenosine is usually preferred to nitroglycerin because of its minimal effect of lowering systemic arterial blood pressures, which is very undesirable during cardiopulmonary collapse situations.

Despite the focus on treatment of only low-flow VMs in the study of Ul Haq et al (9), it is important to also consider and review a 2014 Journal of Vascular and Interventional Radiology article (25) that describes a single-center experience in treating 46 patients with VMs and AVMs over a 13-year period and its retrospective analysis. Vogelzang et al (25) reported on their experience of serially treating 31 patients with VMs and 15 patients with AVMs with absolute ethanol. They reported per-patient major complication rates of 13.3% in AVM treatment (3.9% per procedure) and 9.7% in VM treatment (5.4% per procedure) and per-patient minor complications rates of 6.7% in AVMs (2% per procedure) and 16.1% in VMs (9.8% per procedure). According to Society of Interventional Radiology reporting criteria, the overall complication rate was 24% (11 of 46 patients). Major complications in five patients (11%) included one permanent ulnar neuropathy (in a patient with a forearm VM), one-digit amputation (hand AVM), one leg compartment syndrome requiring fasciotomy (leg AVM), one skin ulceration requiring wound debridement escharectomy (VM), and one transient severe gluteal pain with sciatic neuropathy resolving after 4 weeks (pelvic VM). They further report that 24 patients (52.2%) were totally cured, 12 patients improved (26.1%), 10 patients no change or worse (21.7%). The authors conclude that ethanol produces good outcomes with control of the relief of symptoms in a majority of patients in this series (25).

“Of 141 patients seen from 2005-2011, 22 patients had surgery on their AVMs and 73% had good responses and 20% with minimal improvement; no mean follow-up stated. Only 3 AVM patients had endovascular treatment. In low-flow malformations sodium tetradecyl sulfate (STS) used in all but one of 51 patients. 82% had good response and 18% recurrence rate, again no mean follow-up stated. In total of 73 patients treated, 70% good response, 12% minimal improvement, and 15% had recurrence or no change. (26)” Do et al published a series of 40 AVM patients treated by ethanol alone and ethanol with coils which 68% had significant ablation or complete cure, and 18% had no remission, and 1 patient (2.5%) was worse, with a major complication rate of 3% (5/175 procedures) (27).

A very interesting 2018 publication from Shanghai 9th People’s Hospital authors Jin et al., performed 50 ethanol embolization procedures in 12 patients suffering from chronic skin ulcers refractory to all treatments in the head and neck, hands, and feet. All patient healed their skin ulcers 2-4 weeks post-ethanol embolotherapy and remained healed at a mean clinical follow-up of 15 months. Two patients had focal skin necrosis healing spontaneously within 4 weeks, and 9 patients had minor skin blisters that required no treatment and were self-limited (28).

Fan et al from the 9th People’s Hospital in Shanghai, China reported 8 consecutive patients with symptomatic mandibular AVMs treated with ethanol and coils. 11 embolization procedures were performed on the 8 patients and the amounts of ethanol used ranged from 28 to 50 mL in each session by direct puncture access into the mandible AVM lesion. Fan et al report 6 of 8 patients were arteriographically cured and 2 had significant partial remission awaiting further procedures. 3 of 8 patients had minor complications (29).

Wu et al report 29 patients with ear AVMs that had some form of treatment. Proximal artery ligation in 9 patients caused progression and severe worsening necessitating amputation of 8 patients’ ears and 1 patient underwent embolization. Fifteen patients underwent embolization alone and in 6 patients the disease worsened requiring ear amputation and only 2 patients improved in the remaining 9 patients, with 4 persisting and 3 worsening. Of the 20 patients who had ear amputation 16 (80%) were no better but stable, 3 (15%) improved, and 1 had concurrently an unresectable cervico-facial AVM (30).

In the article by Wu et al (30), ethanol was not used as an embolic agent. In contrast, the article by Jin et al (31) utilized 15 ethanol embolotherapy procedures as the sole means to treat 8 patients with ear AVMs. Six patients (75%) were cured at a mean follow-up of 8 months. The other 2 patients had 80% devascularization and were much improved. Only minor complications requiring no treatments, only observance, were noted. No patients required ear amputation (31).

“The consensus statement in the 2013 International union of Angiology Consensus Document, the authors state among various embolo-sclerotherapy agents, ethanol sclerotherapy procedures are the best long-term outcomes with minimal recurrences.” (32).

In summary, the literature states unequivocally the superiority of absolute ethanol 98% or 96% in the treatment of low-flow malformations. The vast majority of publications state greater than 90% efficacy. Bleomycin is a second-tier embolic agent, as are other sclerosant agents, in low-flow vascular malformation treatment, with fairly consistent published results in the 60%–80% efficacy range. The literature is very clear that, despite rationales for the use of second-tier agents to prevent swelling and ulcerations, these complications do still occur and are never completely avoided.

Experience with the use of ethanol, recognition of where the “nidus” is (rather than a collateral vessel), and knowledge of the vascular supply to nerves is tantamount to lowering complication rates and achieving successful outcomes. With the use of sclerosant embolic agents, especially ethanol, an old adage does apply: “a poor workman blames his tools.” The fact that these authors (9) have published ethanol complication rates in lower ranges than other published series is emblematic of the value of high patient volumes and collaboration in one of the most challenging areas of vascular medicine. The rarity of vascular malformations, compounded by the extreme difficulty in treating these patients, points to the enormity of the problem. Querying many interventional radiologists, interventional neuroradiologists, and fellows training in these subspecialty disciplines, I hear again and again that they have had little or no experience with the use of ethanol as an embolic agent during their training. Therefore, many endovascular specialists are extremely reluctant to use ethanol when they move on to their practices. Often, they use the embolic agents with which they are familiar rather than those with which they are unfamiliar. This is only natural. However, communication between individuals and centers can mutually advance the cumulative experience and knowledge of the endovascular specialists in each. The use of ethanol as an embolic agent in other clinical entities can be a key to gain more experience in its safe use as well. Its use has been reported successful as a primary endovascular treatment for hemorrhaging angiomyolipoma (33), preoperative embolization for renal cell carcinoma as first introduced by Ellman et al (34) and later amplified by Bakal et al (35), or as a palliative treatment for inoperable cases (36). Ethanol can also be very effective as a primary treatment modality for varicocele and its recurrences (37). When collaboration between centers and individuals occurs and experience is shared, knowledge advances, success rates increase, and complication rates decrease, all to benefit our only object: our patients.