Eur Radiol (2010) 20: 2663–2670 DOI 10.1007/s00330-010-1851-2 INTERVENTIONAL Jonathan L. Hart Zaid Aldin Philip Braude Claire L. Shovlin James Jackson Embolization of pulmonary arteriovenous malformations using the Amplatzer vascular plug: successful treatment of 69 consecutive patients Received: 3 March 2010 Revised: 25 April 2010 Accepted: 27 May 2010 Published online: 24 June 2010 # European Society of Radiology 2010 J. L. Hart : Z. Aldin : J. Jackson Department of Imaging, Imperial College Healthcare NHS Trust, London, England Abstract Objective The technique of embolization of pulmonary arteriovenous malformations (PAVMs) with the Amplatzer vascular plug (AVP) has been reported, but no large series has evaluated the effectiveness of this relatively new embolic device. The purpose of this study is to assess the role of AVPs in the treatment of PAVMs. Materials and methods Sixty-nine consecutive patients underwent embolization of pulmonary arteriovenous malformations between September 2006 and December 2008. Clinical, procedural, and physiological data were reviewed retrospectively. Results Of 161 PAVMs, 120 (75%) were successfully embolized with Amplatzer vascular plugs alone. Complete and rapid P. Braude : C. L. Shovlin Department of Respiratory Medicine, Imperial College Healthcare NHS Trust, London, England C. L. Shovlin National Heart and Lung Institute of Cardiovascular Sciences, Imperial College London, London, England J. Jackson ()) Hammersmith Hospital, Du Cane Road, London, W12 0HS, UK e-mail: james.jackson@imperial.nhs.uk Tel.: +44-208-3833485 Fax: +44-208-3833121 Introduction Pulmonary arteriovenous malformations (PAVMs) are high-flow, low-pressure communications between the pulmonary arterial and venous systems allowing a rightto-left shunt and subsequent hypoxemia [1]. The morphology of these communications is variable, ranging from complex vascular structures supplying and draining an aneurysmal sac, to much smaller caliber telangiectatic vessels. The majority of PAVMs occur in the context of hereditary hemorrhagic telangiectasia (HHT), in which there is a 50% incidence [2–4]. Sporadic PAVMs are less common, accounting for fewer than 10% of cases. The direct arteriovenous communication bypasses the important “filter capacity” of the pulmonary capillaries, and this predisposes to cerebrovascular complications including stroke, transient ischemic attack, and cerebral abscess [5]. Transcatheter embolization is established as the preferred occlusion of feeding vessels was easily achieved at the site of arteriovenous communication without complication. Particularly small or tortuous feeding arteries supplying 27 complex and 14 simple PAVMs were occluded with coils. There have been no documented instances of recanalization on follow-up. Conclusion Amplatzer vascular plugs allow the rapid and safe distal occlusion of the majority of PAVMs. Keywords Pulmonary . Arteriovenous malformation . Embolization . Technique . Vascular plug treatment for PAVMs in order to reduce the risk of paradoxical embolization. The technique had previously been recommended only for PAVMs that have one or more feeding vessels of 3 mm or larger [6–9], but concern was expressed about the 3 mm rule [10]. Since stroke risk has now been shown to be independent of the size of PAVMs [5], embolization is now recommended for all PAVMs amenable to the procedure. Hypoxemia due to right-to-left shunting is often well tolerated due to the associated low pulmonary vascular resistance, but a subjective improvement in exercise tolerance is often reported following embolization, even in those individuals without symptoms before treatment [1]. Embolization with MR-compatible steel or platinum coils is the mainstay of treatment in most centers and requires a meticulous technique; unsurprisingly, the experience of the operator has been associated with improved outcomes [11]. In the majority of cases, 2664 embolization involves occlusion of the feeding vessel with one or more coils deployed immediately proximal to the aneurysmal sac in order to avoid the occlusion of normal pulmonary artery branches. Occasionally the specific anatomy of a lesion may necessitate an alternative strategy such as packing of the aneurysm sac when the neck is particularly short or wide or the use of occlusion balloons to control flow during coil deployment in especially large, high-flow feeding vessels. The Amplatzer vascular plug (AGA Medical, Plymouth, MN, USA) is a relatively new occlusive device made of a self-expanding cylindrical nitinol mesh that is especially suited for embolization of large high-flow vessels. It has been used in the arterial and venous systemic circulation for a number of indications including iliac artery occlusion prior to endovascular repair of aortoiliac aneurysms or pelvic surgery, occlusion of transjugular intrahepatic portosystemic shunts, and testicular vein embolization [12, 13]. Several recent reports have described its use for treatment of PAVMs but, to date, no large series addressing its effectiveness at treating these lesions has been published [14–18]. We report our experience with the Amplatzer vascular plug, which has been used for PAVM embolization at our institution since September 2006. Materials and methods Study group The study group comprised 69 consecutive patients who underwent embolization of pulmonary arteriovenous malformations between September 2006 and December 2008. femoral venous approach to document the anatomy of treatable lesions. In patients in whom the MDCT had demonstrated unilateral PAVMs, that side only was studied at the time of angiography. The Grollman pigtail catheter was then exchanged for a 90 cm long, 6 Fr, straight sheath (Cordis, Ascot, Berkshire, UK) through which a 100 cm, 5 Fr Headhunter catheter (Cook Europe, Bjaeverskov, Denmark) was introduced. Selective catheterization of the feeding vessel(s) to each treatable PAVM was performed with this catheter sheath combination. Once a suitable position had been achieved as distally as possible within the feeding vessel beyond any branches to normal lung, the 5 Fr catheter was removed, and embolization was performed through the sheath with an AVP. The size of the AVP selected for embolization was approximately 1.5 to 2 times the caliber of the feeding vessel, which was itself estimated by comparing it by eye with the diameter of the sheath tip on pre-embolization arteriograms. Satisfactory positioning of the AVP was confirmed in most cases by repeat arteriography via the sheath before its detachment from the introducing wire. If suboptimally positioned, the AVP was resheathed and redeployed in a more appropriate site. Additional feeding vessels to a PAVM were treated in the same manner except when their small size or tortuosity precluded the introduction of the sheath to a suitable distal position, in which case embolization was performed through the 5 Fr catheter with MR compatible coils. The duration of the embolization intervention was determined by the number and complexity of the malformations requiring embolization and patient tolerance to the procedure. In general, procedure duration was between 90 and 120 min. All patients were discharged on the day following the procedure. Patients were readmitted for subsequent embolizations as required until complete occlusion of all significant lesions was achieved. Pre-procedure assessment Physiological tests All patients underwent a full clinical evaluation prior to treatment. Baseline chest radiographs were obtained and standard blood tests included full blood count, coagulation screen, and liver function tests (to exclude hepatopulmonary syndrome). If recent cross-sectional imaging from the referring hospital was not available, a non-contrast multidetector CT (MDCT) of the thorax [19] was obtained to confirm the diagnosis of PAVM and identify the approximate number and size of lesions prior to intervention. Embolization procedure All embolizations were performed by a single interventional radiologist. Prophylactic intravenous antibiotics were administered in all cases (1 g vancomycin) 1 h before each procedure. The pulmonary artery pressure was measured in all individuals at the time of angiography. Selective right and/or left pulmonary digital subtraction arteriograms were then obtained with a 7 French Grollman angled pigtail catheter (Cook Europe, Bjaeverskov, Denmark) via a Arterial oxygen saturation was measured with the patient breathing room air using a pulse oximeter (Biox 3740; Ohmeda, Hatfield, Hertfordshire, UK) and an ear probe. Measurements were taken every 60 s for 10 min in the erect and supine positions, with the result expressed as the mean of the last four readings. The nonparametric Wilcoxon’s matched pairs signed rank test was used to compare the pre- and post-embolization arterial oxygen saturation. To compensate for multiple comparison tests, only P values of less than 0.01 were considered statistically significant. Where more than one procedure was performed, the post-embolization value corresponds with the recording taken after the most recent procedure. Follow-up Patients were invited for review in a dedicated PAVM clinic 3–6 months following the procedure and annually thereafter. Given the pattern of tertiary referral for patients 2665 with PAVMs from across the UK, in the absence of any symptoms or concerns, a small number opted for followup at the referring hospital or declined follow up. Data are presented for all individuals who attended for a follow-up assessment. Results No major complications occurred. Minor pleuritic-type chest pain was experienced by four (6%) patients. A single patient developed asymptomatic atrial fibrillation, which resolved spontaneously. In one case a minor, selflimiting hemoptysis occurred following the procedure. In two cases, the anatomy of the lesion required occlusion of the arterial supply to a small segment of immediately adjacent normal lung, but both patients remained asymptomatic following the procedure. Clinical features Physiological tests The study group comprised 28 male and 41 female patients, with a mean age of 44.6 years (range 16–78). The presenting clinical features of these patients are summarized in Table 1. A definitive diagnosis or a possible diagnosis of HHT was made in 52 (75%) patients and 10 (15%) patients, respectively, according to Curacao criteria [20]. Seven (10%) individuals had isolated PAVMs. Ten patients had undergone previous embolization of PAVMs (as part of a previously reported series [21]): seven had recurrent lesions that had been previously embolized with MR-compatible coils, and in three patients pre-existing PAVMs noted at the time of their original procedures, but not treated because of their small size, had grown to a size amenable to embolization. A single patient had undergone a lobectomy 19 years previously to remove a symptomatic PAVM. Embolization Eighty-three procedures were performed in 69 patients. A total of 161 PAVMs (mean 2.3 lesions per patient; range 1–12) comprising 115 simple lesions and 46 complex lesions with feeding arteries ranging between 2 and 13 mm in diameter were embolized. A total of 120 lesions were treated with Amplatzer vascular plugs alone (75%). Smaller additional feeding arteries to 27 complex PAVMs required coil occlusion. Fourteen PAVMs with small tortuous feeding arteries were treated with coils alone. The mean pulmonary artery pressure prior to embolization was 14.8 mmHg (range 7–26 mmHg). Pulmonary arterial hypertension, defined as a mean pulmonary artery pressure in excess of 25 mmHg, was identified in three patients. Table 1 Clinical features at presentation Symptoms Respiratory Dyspnea Hemoptysis Embolic Cerebral abscess Transient ischemic attack Cerebrovascular accident Peripheral abscess Migraine Asymptomatic Number Percentage 28 9 41 13 2 5 5 2 9 15 3 7 7 3 13 22 Statistically significant improvement in supine and erect systemic arterial oxygen saturations was demonstrated following completion of embolization treatments (Table 2). Results are comparable with those from our previous series in which MR-compatible coils were used [6, 7, 21]. Follow-up Clinical follow-up data were available in 51 (74%) patients at a mean of 9.6 months (range 1–25 months). Twenty-five out of 26 patients with impaired exercise tolerance or dyspnea prior to the procedure reported an improvement in symptoms following treatment (two additional patients who had not described impaired exercise tolerance prior to the procedure nevertheless described an increased exercise capacity following treatment). No further events were documented in patients with previous neurological complications attributable to PAVMs (a single patient reported a neurological event related to a known cerebral AVM). Eleven patients required additional embolization sessions to achieve occlusion of all PAVMs of treatable size. Repeat angiography prior to these subsequent treatments demonstrated that vessels previously treated by AVPs remained occluded. In a single patient, repeat pulmonary angiography for investigation of hemoptysis demonstrated filling of the treated PAVM via multiple intrapulmonary collaterals that were successfully embolized with MR-compatible coils; again, the vessel treated with an AVP remained occluded. Five patients underwent follow-up CT at a mean of 12.4 months following embolization (range 2–34 months). In all of these, there was complete occlusion of the feeding vessels treated with AVPs with decompression of the PAVM venous sac (Figs. 1 and 2). Discussion The technique of percutaneous catheter embolization of pulmonary arteriovenous malformations involves delineation of the lesion by pulmonary angiography, followed by super-selective catheterization and occlusion of the feeding artery. It is established as the preferred treatment to reduce 2666 Table 2 Systemic arterial oxygen saturation pre- and post-embolization treatments the risk of embolic complications [6–9]. A substantial number of reports have demonstrated good immediate outcomes and long-term effectiveness, with recurrence rates ranging from 0 to 22% [10, 22–28]. Recurrences are also amenable to retreatment by transcatheter embolotherapy with durable results. Despite good technical outcomes in the vast majority of cases, residual lesions remain in many patients due to the presence of feeding vessels that are too small to be occluded with currently available techniques. Although the risk of infective embolic complications is reduced, the lifelong requirement for antibiotic prophylaxis prior to dental and surgical treatment persists. The Amplatzer vascular plug is a relatively new occlusive device made of a self-expanding cylindrical nitinol mesh that can be deployed rapidly and can be repositioned before final release. It is particularly suitable for embolization of large high-flow vessels such as those found in pulmonary arteriovenous malformations. AVP1, AVP2, AVP3, and AVP4 devices are available; they vary in their configuration and range of available diameters. The AVP4 has only recently been introduced into clinical practice and differs from the other devices in that it can be introduced through a conventional diagnostic catheter with a 0.038 inch lumen. The AVP1 device was used for all but one PAVM in this series. Fig. 1 A 57-year-old woman with hereditary hemorrhagic telangiectasia (HHT), several large bilateral PAVMs, and history of CVA. a Left lower lobe pulmonary angiogram demonstrates large basal PAVM of simple type with a feeding vessel measuring approximately 6 mm in diameter at its point of communication with the dilated venous sac. b Control film showing 10 mm diameter AVP after exiting the delivery sheath but still attached to its wire. c Selective arteriogram shows good position of AVP at the neck of the PAVM. There is still flow through the malformation immediately after AVP insertion. d Selective arteriogram a few minutes later demonstrates complete vessel occlusion with preservation of normal proximal pulmonary artery branches. e Axial CT image through the lung bases performed before PAVM embolization demonstrates the large venous sac in the left lower lobe. A further PAVM is present on the right side, which was also embolized with an AVP. f Axial CT image through the lung bases at the same level as e performed 2 years and 10 months after PAVM embolization shows a portion of the AVP used for vessel occlusion on the right side. The venous sacs in both lower lobes have decompressed completely. Other images (not shown) demonstrated only small linear markings on both sides at the sites of the venous sacs and confirmed complete occlusion of other embolized PAVMs Measurement Pre-embolization Median Interquartile range Post-embolization Median Interquartile range Pa a Systemic arterial oxygen saturation (%) Supine (n=62) Erect (n=62) 94 89, 96 94 92, 96 96 95, 97 <0.001 96 95, 97 <0.001 Wilcoxon’s matched pairs signed rank test 2667 Fig. 2 A 48-year-old man with hereditary hemorrhagic telangiectasia (HHT), relatively small bilateral PAVMs, and history of CVA. a Axial CT image through the right lung demonstrates a small PAVM adjacent to the oblique fissure, which had a feeding vessel measuring 4 mm in diameter. b Control film and c angiogram after releasing a 6 mm diameter AVP demonstrate its optimal positioning at the neck of the PAVM. d Arteriography a few minutes later demonstrates complete vessel occlusion. e CT image 10 months later demonstrates AVP and complete decompression of the venous sac The device has platinum and iridium marker bands on either end and a stainless steel microscrew on the proximal band attached to a 135-cm-long delivery wire. The diameter of the AVP1 ranges from 4 to 16 mm in 2 mm increments; all sizes of AVP1 can be deployed via a 6 French sheath with a 2 mm lumen although the introduction of the 14 mm and 16 mm plugs may be difficult, and a 7 French sheath with a luminal diameter of 2.3 mm is recommended when these devices are used. There are a number of important advantages of AVPs over coils for the treatment of PAVMs; they are described below. “anchor” technique of coil embolization, in which the first loop of a coil is purposely deployed in a normal pulmonary artery branch proximal to the venous sac, is utilized by some practitioners to overcome the problem of coil migration. The very nature of this technique, however, means that the occlusion is performed proximal to, and often at some distance from, the PAVM neck. In our experience with the AVP1 device, this problem is overcome in most PAVMs; the vascular sheath through which the plug is to be deployed can usually be placed at the neck of the PAVM and can, in some instances, be introduced into the venous sac itself. The plug can then be introduced with the sheath in this position and deployed during sheath withdrawal (Figs. 1, 2, and 3). Ability to achieve a very distal and safe embolization in the majority of PAVMs It is recognized by all practitioners who regularly perform PAVM embolization that it is desirable to achieve a very distal occlusion of the feeding vessel to a PAVM, preferably at the neck of the venous sac. This not only reduces the risk of occluding branches to adjacent normal lung but may also reduce the likelihood of persistent perfusion of the venous sac by bronchial collaterals and of pulmonary artery recanalization. Such a distal occlusion is often very difficult to achieve with metallic coils, particularly when the feeding vessel is large, because of the risk of coil migration through the sac into the systemic circulation with potentially disastrous complications. The Complete vessel occlusion with a single device The embolization of a feeding vessel to a PAVM, particularly when of a large diameter, will usually require the use of several metallic coils in order to achieve complete occlusion. In our experience complete pulmonary artery occlusion, even of feeding vessels measuring 12 mm in diameter, is achievable with a single AVP1 (Figs. 1, 2, and 3). This has a number of advantages: 1. It speeds up the procedure such that a larger number of PAVMs can be embolized in a single session. 2668 Fig. 3 A 37-year-old woman with hereditary hemorrhagic telangiectasia (HHT), single moderately sized right lower lobe PAVM, known cerebral AVM, and history of transient ischemic attack. a Coned image of portion of CXR demonstrates well-defined nodule in the right lower zone consistent with a PAVM. b Right pulmonary artery angiogram shows PAVM with 5 mm feeding vessel, an aneurysmal venous sac, and early venous return. c Angiogram after positioning an 8 mm diameter AVP demonstrates its optimal positioning at the neck of the PAVM. d Arteriography a few minutes later demonstrates complete vessel occlusion with preservation of normal pulmonary artery branches. e Coned image of portion of CXR 8 months after embolization demonstrates AVP and disappearance of venous sac 2. It allows occlusion of a shorter length of vessel. When coils are used to occlude large pulmonary feeding arteries, a coil nest measuring several centimeters in length is often required to achieve complete occlusion. 3. There is a theoretical reduction in the likelihood of thrombus embolizing through the PAVM into the systemic circulation during vessel occlusion. During the deployment of several coils in the feeding vessel to a PAVM, the catheter and coil manipulation adjacent to previously positioned coils on which thrombus is forming risks displacement of some of this thrombus. Such manipulations are not required when using the AVP1. A recent paper demonstrated a mean vessel occlusion time of just over 3 min for AVPs used to treat 12 simple PAVMs, which the authors suggested could minimize the opportunity for systemic embolization of microthrombi from the device surface [29]. Oversizing of the device is recommended, and this means that the choice of too small a plug is very unlikely. Furthermore, this oversized plug does not displace the sheath tip during deployment. Use of the AVP to treat PAVMs has been reported by several authors, but to our knowledge, ours is the first report of a large series of patients successfully treated with the device. No technical difficulties were encountered in deploying the device, and adequate target vessel occlusion was achieved in all cases where it was used. Fourteen PAVMs (8.7%) with particularly small or tortuous vessels were, however, unsuitable for treatment with the device, and these were occluded with coils. No major complications arose, and the few that did occur were minor with no long-term sequelae. Clinical outcomes in terms of symptom improvement and increase in systemic arterial oxygen saturations were similar to our previous experiences using MR-compatible coils [6, 7, 21]. The major limitation of this study is the small number of patients who have undergone follow-up imaging of treated PAVMs. Further follow-up will be required to fully establish the effectiveness of the AVP in the longer term, but on the basis of the angiographic follow-up available in 12 of our patient group (at a mean interval of 7 months), and CT follow-up in 5 patients (at a mean interval of 12.4 months) we have had no concerns regarding its Reduced requirement for accurate sizing of the AVP to achieve safe vessel occlusion One of the main difficulties with coil occlusion of PAVMs is the choice of the correct coil size. Too small a coil risks migration through the venous sac into the systemic circulation, whilst too large a coil will not form the tight “nest” required for vessel occlusion and will often displace the catheter tip from its distal position. This is not a problem with the AVP. 2669 effectiveness. Recanalization of PAVMs after embolization has been reported, however, by Fidelman et al. [30]. They described “spontaneous reperfusion of two PAVMs within seven weeks of initially successful embolization with AVPs” and suggested that “deposition of coils proximal to the AVP may decrease the chance of PAVM reperfusion.” As we have already documented, the additional use of coils proximal to an AVP has never been required in our series, and we have seen no instances of recanalization in those patients in whom follow-up imaging has been performed. It is possible that instances of recanalization have been missed as CT is not a part of our routine follow-up protocol, but none of our patients who have been followed up in clinic have shown any decline in their oxygen saturation or suffered paradoxical emboli. It is noteworthy that the recanalized PAVM demonstrated in the figures in the report by Fidelman et al. had been treated by positioning an AVP within its feeding artery approximately 5 cm proximal to the venous sac, and this is the most likely cause of the observed reperfusion. The authors of this paper comment upon the need to occlude PAVMs as close as possible to the arteriovenous communication to prevent recurrence as it is well recognized that recanalization of PAVMs treated with coils is more likely the more proximally they are placed in the feeding artery. That this malpositioned AVP was the principal cause of the reperfusion is supported by the fact that recanalization has not been documented in any of the PAVMs treated in our own series in which the AVPs were positioned at the PAVM neck. Conclusion The majority of PAVMs can be treated effectively with Amplatzer vascular plugs, and distal feeding artery occlusion can be achieved with these devices in all but the smallest and most tortuous feeding arteries. Optimal positioning and release of these devices is often safer and faster than coils. As with all forms of embolotherapy, the interventional radiologist is best served by having more than one treatment option, including both occlusion devices and coils. Meticulous technique, modified according to the vascular architecture of the individual PAVM, is required to achieve the best results. Acknowledgments Dr. Jackson has received lecturing fees from AGA Medical during the previous 2months. References 1. Shovlin CL, Jackson JE (2010) Pulmonary arteriovenous malformations and other pulmonary vascular abnormalities. In: Mason RJ (ed) Murray and Nadel’s textbook of respiratory medicine, 5th edn. Elsevier Saunders, Philadelphia 2. Cottin V, Plauchu H, Bayle JY, Barthelet M, Revel D, Cordier JF (2004) Pulmonary arteriovenous malformations in patients with hereditary hemorrhagic telangiectasia. Am J Respir Crit Care Med 169:994– 1000 3. Faughnan ME, Granton JT, Young LH (2009) The pulmonary vascular complications of hereditary haemorrhagic telangiectasia. Eur Respir J 33:1186–1194 4. Sabbà C, Pasculli G, Lenato GM, Suppressa P, Lastella P, Memeo M, Dicuonzo F, Guant G (2007) Hereditary hemorrhagic telangiectasia: clinical features in ENG and ALK1 mutation carriers. J Thromb Haemost 5:1149– 1157 5. Shovlin CL, Jackson JE, Bamford KB, Jenkins IH, Benjamin AR, Ramadan H, Kulinskaya E (2008) Primary determinants of ischaemic stroke/brain abscess risks are independent of severity of pulmonary arteriovenous malformations in hereditary haemorrhagic telangiectasia. Thorax 63:259–662 6. Jackson JE, Whyte MK, Allison DJ, Hughes JM (1990) Coil embolization of pulmonary arteriovenous malformations. Cor Vasa 32:191–196 7. Dutton JAE, Jackson JE, Hughes JMB, Whyte MK, Peters AM, Ussov W, Allison DJ (1995) Pulmonary arteriovenous malformations: results of treatment with embolization in 53 patients. Am J Roentgenol 165:1119– 1125 8. Haitjema TJ, Overtoom TT, Westermann CJ, Lammers JW (1995) Embolisation of pulmonary arteriovenous malformations: results and follow up in 32 patients. Thorax 50:719–723 9. White RI Jr, Pollak JS, Wirth JA (1996) Pulmonary arteriovenous malformations: diagnosis and transcatheter embolotherapy. J Vasc Interv Radiol 7:787–804 10. Pollak JS, Saluja S, Thabet A, Henderson KJ, White RI (2006) Clinical and anatomical outcomes after embolotherapy of pulmonary arteriovenous malformations. J Vasc Interv Radiol 17:35–45 11. Milic A, Chan RP, Cohen JH, Faughnan ME (2005) Reperfusion of pulmonary arteriovenous malformations after embolotherapy. J Vasc Interv Radiol 16:1675–1683 12. Cil B, Peynircioğlu B, Canyiğit M, Geyik S, Ciftçi T (2008) Peripheral vascular applications of the Amplatzer vascular plug. Diagn Interv Radiol 4:35–39 13. Rimon U, Heldenberg E, Golan G, Shinfeld A, Garniek A (2008) Amplatzer vascular plug: expanding the applications spectrum. Cardiovasc Interv Radiol 31(Suppl 2):S84–S87 14. Conti V, Fiorucci F, Serpilli M, Patrizi A, Paone G, Neri P, Giannunzio G, Pirozzo MG, Fiorucci C, Lucantoni G (2008) Osler-Rendu-Weber syndrome: congenital arteriovenous intrapulmonary fistula treated using a percutaneous Amplatzer plug. Eur Rev Med Pharmacol Sci 12:213–216 2670 15. Beck A, Dagan T, Matitiau A, Bruckheimer E (2006) Transcatheter closure of pulmonary arteriovenous malformations with Amplatzer devices. Catheter Cardiovasc Interv 67:932–937 16. Tabori NE, Love BA (2008) Transcatheter occlusion of pulmonary arteriovenous malformations using the Amplatzer vascular plug II. Catheter Cardiovasc Interv 71:940–943 17. Ferro C, Rossi UG, Bovio G, Seitun S, Rossi GA (2007) Percutaneous transcatheter embolization of a large pulmonary arteriovenous fistula with an Amplatzer vascular plug. Cardiovasc Intervent Radiol 30:328–331 18. Anderson PE, Kieldsen AD (2007) Occlusion of pulmonary arteriovenous malformations by use of vascular plug. Acta Radiol 48:496–499 19. Remy J, Remy-Jardin M, Wattinne L, Deffontaines C (1992) Pulmonary arteriovenous malformations: evaluation with CT of the chest before and after treatment. Radiology 182:809–816 20. Shovlin CL, Guttmacher AE, Buscarini E, Faughnan M, Hyland R, Westermann CJJ, Kjeldsen A, Plauchu H (2000) Diagnostic criteria for hereditary haemorrhagic telangiectasia (OslerWeber-Rendu syndrome). Am J Med Genet 91:65–66 21. Shovlin CL, Tighe HC, Davies RJ, Gibbs JSR, Jackson JE (2008) Embolisation of pulmonary arteriovenous malformations: no consistent effect on pulmonary artery pressure. Eur Respir J 32:162–169 22. White RI, Lynch-Nyhan A, Terry P, Buescher PC, Farmlett EJ, Charnas L, Shuman K, Kim W, Kinnison M, Mitchell SE (1988) Pulmonary arteriovenous malformations: techniques and long-term outcome of embolotherapy. Radiology 169:663–669 23. Remy-Jardin M, Dumont P, Brillet PY, Dupuis P, Duhamel A, Remy J (2006) Pulmonary arteriovenous malformations treated with embolotherapy: helical CT evaluation of long-term effectiveness after 2-21-year follow-up. Radiology 239:576–585 24. Lee DW, White RI Jr, Egglin TK, Pollak JS, Fayad PB, Wirth JA, Rosenblatt MM, Dickey KW, Burdge CM (1997) Embolotherapy of large pulmonary arteriovenous malformations: long-term results. Ann Thorac Surg 64:930–939 25. Mager JJ, Overtoom TT, Blauw H, Lammers JW, Westermann CJ (2004) Embolotherapy of pulmonary arteriovenous malformations: long-term results in 112 patients. J Vasc Interv Radiol 15:451–456 26. Saluja S, Sitko I, Lee DW, Pollak J, White RI (1997) Embolization of pulmonary arteriovenous malformations: long term results. Ann Thorac Surg 64:930–940 27. Brillet PY, Dumont P, Bouaziz N, Duhamel A, Laurent F, Remy J, Remy-Jardin M (2007) Pulmonary arteriovenous malformation treated with embolotherapy: systemic collateral supply at multidetector CT angiography after 2-20-year follow-up. Radiology 242:267–276 28. Lacombe P, Lagrange C, Beauchet A, El Hajjam M, Chinet T, Pelage JP (2009) Diffuse pulmonary arteriovenous malformations in hereditary hemorrhagic telangiectasia: long-term results of embolization according to the extent of lung involvement. Chest 135:1031–1037 29. Aal AKA, Hamed MF, Biosca RF, Saddekni S, Raghuram K (2009) Occlusion time for Amplatzer vascular plug in the management of pulmonary arteriovenous malformations. Am J Roentgenol 192:793–799 30. Fidelman N, Gordon RL, Bloom AI, LaBerge JM, Kerlan RK Jr (2008) Reperfusion of pulmonary arteriovenous malformations after successful embolotherapy with vascular plugs. J Vasc Interv Radiol 19:1246–1250