Percutaneous Endovascular Occlusion of Symptomatic Coronary Arteriovenous Fistulas with Cyanoacrylate

Coronary artery (coronary-cameral) fistula is an uncommon condition where an abnormal communication exists between a coronary artery and a cardiac chamber, pulmonary artery or systemic vein. Usually the termination is to the right side of the heart, occasionally to the left, and either the right or the left coronary artery may be involved.1 Transcatheter occlusion of coronary artery fistula has become a safe and effective treatment method, avoiding surgical repair and its associated complications.2 Transcathteter occlusion has been performed with a number of devices or materials including coils,2–7 detachable balloons,7–9 Rashkind double umbrellas,10,11 Amplatzer duct occluders,10,12 polyvinyl alcohol particles13,14 and more recently covered stents (or stent-grafts).15–17 Only one other previous case report describes the use of yet another embolic agent, cyanoacrylate (CA) (Histoacryl, Melsungen, Germany), for transcatheter closure of a coronary artery fistula.18 We describe our first four such cases treated with cyanoacrylate and discuss its properties and uses as an embolic agent. Materials and Methods The four patients with coronary arterial fistulas that were treated in our series included one child and three adults. Their presenting features and site of the fistulas are shown in Table 1. Informed consent was obtained from the three adults and the parents of the child prior to each procedure. No institutional review board approval was required for the performance of any of the procedures as the use of cyanoacrylate was considered an established method of embolization in other vascular territories. All cases were given a weight-appropriate intravenous bolus dose of heparin at the onset of the procedure. After placement of an initial guiding catheter in the proximal aspect of the relevant coronary artery (5 Fr in the child; 6–7 Fr in adults), a microcatheter was placed co-axially through this and used to superselectively catheterize the branch involved in the fistula. In all cases we were able to advance the microcatheter up to and even through the fistula itself. A Turbo Tracker microcatheter (Target Therapeutics, Boston Scientific Corporation, Fremont, Calif.) was used in the first three cases, and two Renegade microcatheters (Target Therapeutics) were used in the fourth. These were manipulated over a Transend EX soft-tip 0,014 inch microwire (Target Therapeutics) in all cases. After placement of the microcatheter in an appropriate position for CA delivery, the CA mixture was prepared in a separate set of stainless steel bowls, uncontaminated by saline or other fluids. In each case a mixture of CA and Lipiodol (Guerbet, Paris, France) and Tungsten powder (Balt, Montmorency, France) was prepared and drawn up into a 1cc luer-lock syringe. The concentrations of the CA mixtures used are summarized in Table 2. A 5% Dextrose solution was used for flushing the microcatheter prior to injections of the CA mixture. Contrast uncontaminated by saline was placed in a third bowl to be used for check arteriograms prior to the glue injection. The 5% dextrose solution and clean contrast medium were drawn up into color-coded 5 ml luer lock syringes. Over a new clean drape placed on the table, the 5% dextrose solution and the CA containing syringe were then sequentially attached to the hub of the microcatheter. The CA mixture was then injected under real-time fluoroscopic guidance. Once a suitable glue cast had formed in the target vessel, slight negative pressure was applied by drawing back upon the plunger of the 1 cc syringe, after which the second operator upon verbalcommand of the first rapidly but smoothly removed the microcatheter. Check arteriography was then performed through the guiding catheter. All microcatheters through which the CA had been injected were discarded after use. Further adjunctive procedures or embolic agents used are listed in Table 2. In Case 1, we were initially unsure about whether or not the fistula was completely occluded after the CA injection hence the placement of two standard platinum microcoils (Target Therapeutics Boston Scientific Corp, Fremont, Calif.) in the same vessel as a “back-up” embolization. In Case 2, a 33% CA mixture was used. Some initial escape of this mixture through the fistula was noted at the onset of the injection, with approximately 0.1 cc of the mixture escaping distally through the fistula and into the pulmonary artery and lungs before increasing the rate of the CA injection allowed “catch-up” of the glue mixture with good adherence to the vessel wall and eventual occlusion of the fistula. In Case 3, we elected to initially create a distal nest of coils within the markedly ectatic distal aspect of the fistula in order to trap any escaping CA so as to avoid the previous complication of pulmonary embolism. A higher CA concentration was also used in this case. No visible distal escape of the CA mixture was noted during the injection. In Case 4 the highest concentration of CA was used (66%) without the creation of an additional coil nest. Two injections of CA were required into the afferent vessel before it was fully occluded. After the final check arteriograms the femoral sheaths were later removed when the activated clotting time had decayed to around 160 seconds. Experienced interventional cardiologists performed the initial coronary catheterizations, whereas the CA mixtures were prepared and injected by an experienced neurointerventionalist in all cases. Results The complications and eventual clinical outcome of all cases are also listed in Table 2. In Case 1, secondary coil placement was probably unnecessary as both feeding arteries (one each from the right and left coronary arteries) were in fact no longer seen to be supplying the fistula after the glue injection indicating that the glue mixture had penetrated the entire length of the fistula sealing the openings of each contributing vessel as originally planned. Transient ST segment elevation changes, lasting for less than a minute, were noted on the electrocardiogram (ECG) during test arteriography in the right coronary artery following the CA and coil embolizations. Cardiac enzyme levels were not measured as these changes did not recur and the patient remained asymptomatic after the procedure. We postulate that transient myocardial ischaemia may have resulted from the contrast medium perfusing areas of myocardium adjacent to the fistula that had previously been rendered relatively ischaemic due to a local “steal” phenomenon by the fistula. Another theoretical mechanism for these changes could be local intimal injury by the CA inducing transient platelet-mediated vasospasm. However, we have never found any evidence of this phenomenon following CA injections in central nervous system or other vascular shunts. Had the ST-segment elevation persisted then one could have postulated that inadvertent CA embolization due to reflux of CA into an occult proximal branch of the coronary artery or to embolization of a CA fragment adherent to the microcatheter as the latter was withdrawn. Despite the presence of a co-existent small ASD, this patient’s symptoms and effort tolerance improved considerably after closure of the fistula. In Case 2, we underestimated the flow through the fistula that resulted in a small volume of CA lost into the pulmonary vessels producing a small symptomatic posterobasal infarction in the left lung that responded well to treatment with antibiotics and non-steroidal anti-inflammatory drugs over the next few days. After this case we realized that a higher concentration of CA and/or a coil nest trap or flow-reduction mechanism should be considered in future cases, as was successfully used in Case 3. Case 4 involved a HIV-positive patient with hypertrophic cardiomyopathy (HOCM) in addition to a coronary artery fistula. There was no evidence of ischaemia on resting or stress ECG testing. After the embolization procedure he complained of mild pleuritic chest pain. Although no glue escape had been noticed during the CA injection, an unenhanced computed tomographic (CT) scan was nevertheless performed to exclude the possibility of pulmonary embolism, which proved to be negative for this finding. As we have no other explanation for the cause of this post-procedural pain we must assume it to be related to pulmonary embolism, possibly due to embolization of a small clot fragment as no glue was visualized in the lungs on the CT scan. Although there was a slight elevation of the cardiac enzymes following the procedure, there were no accompanying ischaemic ECG changes. The chest pain responded well to non-steroidal anti-inflammatory drugs. The patient was discharged 2 days later, only to be readmitted shortly thereafter with central chest pain. At this time, the ECG showed changes suggestive of acute pericarditis but not of acute ischaemia. Echocardiography revealed that the fistula was occluded. It was decided to repeat a coronary angiogram in view of an elevated troponin T level (0.8 ng/ml). This confirmed widely patent epicardial vessels and occlusion of the fistula. The patient developed acute renal failure on the day after the second angiogram. This responded well to 36 hours of peritoneal dialysis. The likely explanation for the development of the renal failure was the presence of an HIV- nephropathy aggravated by a combination of dehydration due to fasting, contrast injection and the use of non-steroidal anti-inflammatory agents. The elevation of the Trop T level could thus have been related to the renal dysfunction rather than being of cardiac origin. The patient made a complete recovery from the renal failure and remains asymptomatic to date. Discussion Cyanoacrylate, in medically pure form, has been used for a number of years as a liquid embolic agent principally in the field of neurointervention for the endovascular treatment of brain and spinal cord arteriovenous malformation and dural arteriovenous fistulas.19–21 CA has also been used in the management of craniofacial arteriovenous malformations,22 varicoceles,23 esophageal and gastric varices24,25 and a number of other non-neurological indications.26 The original CA compound, Isobutyl cyanoacrylate or IBCA, is no longer available having being replaced in clinical use by N-butyl-2-cyanoacrylate (NBCA).27 NBCA is available under the name Histoacryl (B. Braun, Melsungen, Germany). Cyanoacrylate polymerizes immediately upon contact with an ionic medium such as saline or blood. It is non-opaque radiographically and must be rendered so by mixing the CA with iophendylate (Pantopaque) or ethiodized oil (Lipiodol). In addition to opacifying the CA mixture, the use of iophendylate or Lipiodol to dilute the CA mixture also prolongs the speed of the polymerization.20,21,26 Glacial acetic acid has also been used to prolong the polymerization time.28 The concentration of the CA mixture i.e. the relative ratio of glue to diluting agent, is judged mainly according to the rate of flow through a lesion i.e. the contrast transit time from artery to vein through the lesion, as well as from the calibre of the target vessel and the vascular architecture of the target, i.e. fistula or plexiform nidus. A plexiform nidus traps the CA to a greater degree and more readily than a fistula and therefore generally requires a lower concentration mixture to achieve greater nidal penetration and filling. Arteriovenous fistulas are high flow single or multiple large calibre lesions. Occlusion of these with CA requires rapid polymerization times and therefore a relatively higher CA concentration to avoid inadvertent passage of the CA through to the venous circulation before the fistula is closed. The glue concentration (glue to lipiodol ratio) used must be matched to the velocity of flow through and calibre and degree of tortuosity of the target vessel as well as the distance between the intended site of obstruction and the catheter tip. Unfortunately this is a relatively inexact science and requires a large degree of personal judgement and experience. This also explains the variable concentrations of glue used in our cases. CA is always delivered through a microcatheter placed in a co-axial manner through a larger guiding catheter. The microcatheter system must first be flushed with a non-ionic fluid, such as a 5% dextrose solution, so as to avoid premature polymerization within the catheter. In the case of a fistula, the tip of the microcatheter should be curved or placed at a bend in the target vessel so as to immediately trap the CA as it emerges from the microcatheter tip. The injection is continued until the target vessel is fully occluded, which may occur in less than a second. Slight negative pressure is applied by traction on the plunger of the syringe, and then the microcatheter is rapidly removed to avoid it becoming glued in-situ. Once removed the microcatheter is then discarded as it becomes blocked by the CA and cannot be reused. As the CA sets it creates a solid cast within the vessel or nidal lumen resulting in immediate occlusion. Vessel wall changes following CA injection include an initial inflammatory response followed by a later chronic granulomatous reaction with fibrosis within 1 month.21,27 Heat release occurs as part of the polymerization process but its role in producing vessel wall damage is unknown. There may also be a direct toxic reaction by the CA itself on the vessel wall. There is no evidence to date that NBCA is carcinogenic in humans. Being a liquid, the CA can be extremely difficult to control in a high flow fistula. Two potential complications associated with CA usage include the inadvertent occlusion of non-target vessels leading to ischaemia and necrosis of normal tissues, and passage of CA through the arteriovenous shunt leading to downstream vascular occlusion or pulmonary or systemic embolization.34,35 Flow-reduction techniques can be used to afford better control of the CA injection. In neurointerventional procedures this can be partly achieved by producing a transient fall of the mean blood pressure to around two-thirds of the baseline level.20 However, such a transient drop in blood pressure may not always be well tolerated in cardiac patients. Inflation of a balloon in the vessel proximal or distal to the micro-catheter tip may be a reasonable alternative. Complete flow arrest is inadvisable because some forward flow is required to carry the CA into, and slightly beyond, the fistula site. Metallic coils can also be deployed in a proximal or distal position relative to the site of the fistulous communication so as to also reduce the flow. Coil placement distal to the injection site will also act as a glue trap possibly preventing inadvertent distal embolization of CA beyond the fistula. This would result in fewer coils being required than would be required for occlusion of a fistula with coils alone. In our experience, the detachable coils, such as the Guglielmi Detachable Coil (Target Therapeutics, Boston Scientific Corporation, Fremont, Calif.), are best suited for this purpose, being safer to deploy in a high flow situation. We would not advocate the use of CA as the primary embolic material of choice where the outlet of the fistula is to a left-sided heart chamber, nor would we advocate its use in larger calibre vessels (6–8 mm) without the use of concomitant flow reduction or a coil “filter.” CA should not be used in very large calibre vessels (> 8 mm). The use of CA as an embolic agent is both difficult and dangerous. Given the availability and relative ease of use of microcoil systems and other devices the question arises as to why we have elected to use CA in the setting of coronary artery fistulas. The retention of coils or other mechanical devices in a high-flow situation can be difficult and may not always occlude the fistula satisfactorily. One advantage of CA is that it can be injected right through the fistula from the arterial to the venous side occluding the confluences of multiple feeding arteries within the fistula as was seen in Case 1. The use of a detachable coil system such as the Guglielmi Detachable coil system is safer but far more expensive. At a cost of $500.00–600.00 per coil, CA is a much cheaper agent that allows immediate and permanent occlusion. A single ampoule of CA costs around $15.00, and the cost of 2 or 3 ampoules of CA, an ampoule of Lipiodol and, if required to render the mixture more opaque, tungsten powder, will still add up to no more than $150.00 for the CA mixture. Despite the advantage of the cheaper cost of CA compared with other embolic agents, two of our cases developed complications necessitating further investigation and treatment, and additional coils were placed in Cases 1 and 3 adding further to the final overall costs of treatment in our small series. Despite our previous experience of several hundred CA injections in the treatment of well over 100 adult and paediatric cranial and spinal AV shunts we found it difficult to adapt this technique to the closure of coronary artery fistulas. This was in part due to cardiac movement making fluoroscopic observation during the CA injections very difficult, coupled with the injection of CA mixtures in a high-flow situation without the benefit of flow reduction. In conclusion, cyanoacrylate is a useful but difficult embolic agent to use. It is cheap, permanent and can penetrate into vascular channels into which other embolic agents may be unable to reach and occlude adequately. We strongly advocate that the use of CA as an embolic agent for procedures such as coronary artery fistula occlusion should be restricted to practitioners experienced in its usage, or be utilized under the strict supervision of highly experienced operators, and then only in situations that would benefit from the use of CA rather than other embolic materials or devices.

1. Amplatz K, Moller JH. Coronary artery fistula. In: Amplatz K and Moller JH (Eds.) Radiology of Congenital Heart Disease. St Louis: Mosby Year Book;1993:407–414. 2. McMahon CJ, Nihill MR, Kovalchin JP, Mullins CE, Grifka RG. Coronary artery fistula: Management and intermediate-term outcome after transcatheter coil embolization. Tex Heart Inst J 2001;28:21–25. 3. Eicken A, Sebening W, Genz T, et al. Coil embolization therapy in congenital coronary arterial fistulas. Paediatr Cardiol 2003;24:149–153. 4. Buheitel G, Ludwig J, Hofbeck M. Transcatheter occlusion of a large coronary artery fistula with new detachable platinum microcoils. Cardiol Young 2001;11:571–573. 5. Qureshi SA, Reidy JF, Bin Alwi M, et al. Use of interlocking detachable coils in embolization of coronary arteriovenous fistulas. Am J Cardiol 1996;78:110–113. 6. Issenberg HJ. Transcatheter coil closure of a congenital coronary arterial fistula. Am Heart J 1990;120:1441–1443. 7. Reidy JF, Anjos RT, Qureshi SA, et al. Transcathter embolization in the treatment of coronary artery fistulas. J Am Coll Cardiol 1991;18:187–192. 8. Skimming JW, Gessner IH, Victoria BE, Mickle JP. Percutaneous transcatheter occlusion of coronary artery fistula using detachable balloons. Pediatr Cardiol 1995;16:38–41. 9. Hartnell GG and Jordan SC. Balloon embolisation of a coronary arterial fistula. Int J Cardiol 1990;29:381–383. 10. Okubo M, Nykanen D, Benson LN. Outcomes of transcatheter embolization in the treatment of coronary artery fistulas. Cathet Cardiovasc Interv 2001;52:10–517. 11. Perry SB, Rome J, Keane JF, et al. Transcatheter closure of coronary artery fistulas. J Am Coll Cardiol 1992;20:205–209. 12. Hakim F, Madoni A, Goussous Y, et al. Transcatheter occlusion of a large coronary artery fistula using the new Amplatzer duct occluder. Cathet Cardiovasc Diagn 1998;45:155–157. 13. Bennet JM and Maree E. Successful embolization of a coronary arterial fistula. Int J Cardiol 1989;23:405–406. 14. Strunk BL, Hieshimia GB, Shafton EP. Treatment of congenital coronary arteriovenous malformations with microparticulate embolization. Cathet Cardiovasc Diagn 1991;22:133–136. 15. Atmaca Y, Altin T, Özdöl C, et al. Coronary-pulmonary artery fistula associated with right heart failure: Successful closure of fistula with a graft stent. Angiology 2002;53:613–616. 16. Balanescu S, Sangiorgi G, Medda M, et al. Successful concomitant treatment of a coronary-to-pulmonary artery fistula and a left anterior descending artery stenosis using a single covered stent-graft: A case report and literature review. J Interv Cardiol 2002;15:209–213. 17. Mullasani AS, Umesan CV, Kumar KJ. Transcatheter closure of coronary artery to pulmonary artery fistula using covered stents. Heart 2002;87:60. 18. Vajifdor BU, Gupta A, Lokhandwala YY, Kerkar PG, Limaye U, Ramakantan R. Transcatheter glue occlusion of bilateral coronary-pulmonary arterial fistulae. Indian Heart J 1999;51:209–211. 19. Zanetti PH and Sherman FE. Experimental evaluation of a tissue adhesive as an agent for the treatment of aneurysms and arteriovenous anomalies. J Neurosurg 1972;36:72–79. 20. Aletich VA and Debrun GM. Intracranial arteriovenous malformations: The approach and technique of cyanoacrylate embolizaion. In: Connors JJ and Wojak JC (Eds). Interventional Neuroradiology: Strategies and Practical Techniques. Philadelphia: WB Saunders Co. 1999;240–258. 21. Kerber CW and Wong W. Liquid acrylic adhesive agents in interventional neuroradiology. Neurosurg Clin North Am 2000;11:85–99. 22. Han MH, Seong SO, Kim HD, Chong KH, Yeon KM, Han MC. Craniofacial arteriovenous malformation: Pre-operative embolization with direct puncture and injection of n-butyl cyanoacrylate. Radiology 1999;211:661–666. 23. Kunnen MD and Comhaire F. Nonsurgical cure of varicocele by transcatheter embolization of the internal spermatic vein(s) with a tissue adhesive (Histoacryl Transparent). In: Castañeda-Zúñiga WR and Tadavarthy SM (Eds). Interventional Radiology Baltmore: Williams and Wilkins;1992:73–101. 24. D’Imperio N, Piemontese A, Baroncini D, Billi P, Borioni D, Dal Monte PP, Borello P. Evaluation of undiluted N-butyl-2-cyanoacrylate in the endoscopic treatment of upper gastrointestinal tract varices. Endoscopy 1996; 28: 239-243 25. Dhiman RK, Chalwa Y, Taneja S, et al. Endoscopic sclerotherapy of gastric variceal bleeding with N-butyl-2-cyanoacrylate. J Clin Gastroenterol 2002;35:222-227. 26. Pollak JS and White RI. The use of cyanoacrylate adhesives in perpheral embolization. J Vasc Intervent Radiol 2001;12:907–913. 27. Brothers MF, Kaufmann JCE, Fox AJ, Deveikis JP. N-Butyl-2-cyanoacrylate - Substitute for IBCA in interventional neuroradiology: Histopathological and polymerization time studies. Am J Neuroradiol 1989;10:777–786. 28. Spiegel SM, Viñuela F, Goldwasser MJ, et al. Adjusting the polymerization time of isobutyl-2-cyanoacrylate. Am J Neuroradiol 1996;7:109–112. 29. Pelz DM, Lownie SP, Fox AK, Hutton LC. Symptomatic pulmonary complications from liquid acrylate embolization of brain arteriovenous malformations. Am J Neuroradiol 1995;16:19–26. 30. Hwang SS, Kim HH, Park SH, et al. N-butyl-2-cyanoacrylate pulmonary embolism after endoscopic injection sclerotherapy for gastric variceal bleeding. J Comput Assist Tomogr 2001;25: 16–22.