New Definition and Natural History of Patients With Diffuse Pulmonary Arteriovenous Malformations*

Transcription

1 Original Research PULMONARY VASCULAR DISEASE New Definition and Natural History of Patients With Diffuse Pulmonary Arteriovenous Malformations* Twenty-Seven Year Experience Paola Pierucci, MD; Joshua Murphy, MD; Katharine J. Henderson, MS; Deborah A. Chyun, MSN, PhD, RN; and Robert I. White, Jr, MD Background: Patients with diffuse pulmonary arteriovenous malformations (PAVM), a small but important subset of the PAVM population, have significant morbidity and mortality rates. Methods: Thirty-six patients (21 female and 15 male) with diffuse PAVM from a cohort of 821 consecutive patients with PAVM were evaluated. Diffuse PAVM were categorized angiographically: involvement of one or more segmental pulmonary arteries in one or both lungs. Hereditary hemorrhagic telangiectasia (HHT) status, gender, presence or absence of large (> 3-mm diameter artery) focal PAVM, oxygen saturations, complications including hemoptysis, years of follow-up, and survival were tabulated. Results: HHT was present in 29 of 36 patients (81%), and diffuse PAVM were more commonly bilateral (26 of 36 patients, 72%) than unilateral (10 of 36 patients, 28%) [p 0.02]. Female gender was associated with bilateral diffuse PAVM (19 of 26 patients, 73%) [p 0.01]. Focal PAVM were present in both groups but more commonly in patients with bilateral involvement (16 of 26 patients, 62%) [p 0.02]. Initial oxygen saturations (pulse oximetry, standing) of patients with unilateral and bilateral diffuse PAVM were 87 7% and 79 8% (mean SD), respectively (p 0.02). The last or current values for patients with unilateral and bilateral involvement are 95 3% and 85 7%, respectively (p < ). Nine deaths occurred, and all were in patients with bilateral involvement. Deaths were due to hemoptysis of bronchial artery origin (n 2), hemorrhage from duodenal ulcer (n 1), spontaneous liver necrosis (n 3), brain hemorrhage (n 1), brain abscess (n 1), and operative death during attempted lung transplant (n 1). Conclusions: Patients with diffuse PAVM are a high-risk group, and yearly follow-up is recommended. (CHEST 2008; 133: ) Key words: arteriovenous malformations; diffuse; hereditary hemorrhagic telangiectasia; pulmonary angiography; transcatheter embolization Abbreviations: HHT hereditary hemorrhagic telangiectasia; PAVM pulmonary arteriovenous malformations; TCE transcatheter embolotherapy; TIA transient ischemic attack; URI upper respiratory infection Pulmonary arteriovenous malformations (PAVM) are most commonly connections between a segmental pulmonary artery and vein. 1,2 They are associated with hereditary hemorrhagic telangiectasia (HHT) and are a cause of brain abscess, stroke, and lung hemorrhage in up to 50% of untreated patients. 3 6 Rarely, PAVM are diffuse. Faughnan et al 7 defined diffuse PAVM as every segment of one or more lobes diffusely involved by small PAVM. Patients with this pattern of involvement have a greater incidence of complications including brain abscess and other infections compared to patients with focal PAVM Unlike the vast majority of patients with focal PAVM who are successfully treated by transcatheter embolotherapy (TCE), management of patients with diffuse PAVM is more difficult, and the approach to treatment is less well defined. In a cohort of 16 patients reported in 2000, there was minimal rise in oxygen saturation after TCE even though all seg- CHEST / 133 / 3/ MARCH,

2 mental arteries in both lungs were proximally occluded. Those with unilateral involvement had a better response. 7 Since the 2000 report, we have seen an additional 21 patients with diffuse PAVM and followed up 15 of the 16 originally reported patients. The purpose of this report is to further describe the natural history of patients with diffuse PAVM and to present a new definition and approach to treatment. Materials and Methods Demographics and Clinical Presentation From a cohort of 821 consecutive patients with PAVM seen by the senior author between May 1978 and December 2006, 36 patients had diffuse PAVM and were included in this report. With informed consent and approval of the internal review board, charts were reviewed. HHT status, gender, age at first presentation, the presence of large focal PAVM, degree of diffuse involvement, oxygen saturations, years of follow-up, morbidity, and mortality were noted. The patients were classified into two categories: those with diffuse involvement of one lung (unilateral) or both lungs (bilateral). Years of follow-up were calculated from the date of the first visit to the last visit or date of death. HHT was diagnosed according to current criteria published by Shovlin et al. 11 All patients underwent complete diagnostic pulmonary angiography with measurement of pulmonary artery pressures. Definition of Diffuse PAVM Previously, a patient was classified as having diffuse PAVM when all segmental branches of one lobe were involved. 7 Since there was no attempt to distinguish the degree of involvement of each segment in each lobe in the first article, 7 we believed that a more precise classification was needed. For inclusion in this article, diffuse involvement was defined as minimally one segment diffusely involved. If only 1 of 18 lung segments was diffusely involved, this qualified for entry into this series (Fig 1, 2). If there were five or more segments involved in the right lung or four or more segments involved in the left lung, the extent was considered 50%. Our new definition is quite restrictive and should not be confused with patients having multiple focal PAVM, involving one or more segmental arteries, who after TCE have nearnormal-appearing pulmonary circulation (Fig 3). Patients with *From the Department of Respiratory Medicine (Dr. Pierucci), School of Medicine, Policlinic of Bari, Bari, Italy; Department of Pediatrics, Section of Cardiology (Dr. Murphy), and Department of Diagnostic Radiology, Section of Interventional Radiology (Ms. Henderson and Dr. White), Yale University School of Medicine, New Haven, CT; and Yale University School of Nursing (Dr. Chyun), New Haven, CT. Manuscript received August 10, 2007; revision accepted November 12, Reproduction of this article is prohibited without written permission from the American College of Chest Physicians ( org/misc/reprints.shtml). Correspondence to: Katharine J. Henderson, MS, Yale University School of Medicine, Department of Diagnostic Radiology, 333 Cedar St, Room 5039 LMP, New Haven, CT 06520; katharine.henderson@yale.edu DOI: /chest Figure 1. Left anterior oblique selective right pulmonary angiogram early and late arterial images in patient 2. Top, A: diffuse involvement of the posterior and medial segment arteries is easily seen while the anterior and lateral segments are uninvolved. Bottom, B: in late arterial phase, early venous drainage into the left atrium is seen from the diffusely involved posterior medial segments. multifocal PAVM have a higher rise in oxygen saturation after TCE. These patients tend to resemble patients with simple focal PAVM in that the rise in oxygen saturation after TCE is higher and more durable than in patients with diffuse involvement. Embolotherapy Our technique was changed from central to peripheral occlusion empirically since the first article 7 was published. We reasoned that since hemoptysis, due to hypertrophied bronchial 654 Original Research

3 Figure 2. Shallow left anterior oblique views of super selective right pulmonary angiograms in late arterial phases in patient 28. In both the anterior segment (top, A) and the lateral segment (bottom, B), diffuse involvement from each subsegmental branch is noted. At the end of each subsegmental branch is tiny almost telangiectasia-type abnormalities. The pulmonary veins from each of these segments are already opacified. arteries, in our severely symptomatic patients is noted with increasing frequency, we should modify our technique. Previously, we had occluded all segmental pulmonary arteries to one lobe by proximal occlusion. This undoubtedly contributed to bronchial artery enlargement and collateral filing of distal pulmonary arteries. With our new technique, our goal was to block peripherally only the most severely affected segments, one lung at a time. The most severely involved segmental arteries were occluded from peripheral to central by dense packing of pushable Figure 3. A young patient with multifocal PAVM in the right lung. Earlier in her life, surgical resection of her right superior segment had been performed. The anterior (top left, A) and lateral (top right, B) right pulmonary angiogram demonstrated small PAVM distally in right lower lobe and right middle lobe. Center left, C, and center right, D: super selective pulmonary angiogram in anterior segment of right lower lobe demonstrates multiple focal PAVM with early venous drainage (PV pulmonary vein). Bottom left, E, and bottom right, F: Super selective right pulmonary angiogram with catheter beyond right middle lobe and anterior segment lower lobe demonstrates normal pulmonary artery anatomy following distal occlusions of each small focal PAVM. This distinction between multifocal and diffuse PAVM is often difficult and is not certain until a post-tce angiogram is obtained. No early veins are seen, and angiogram is almost normal in the postocclusion angiogram. fibered coils in the new patients treated since our previous publication (Fig 4).7 This new approach was designed to exclude the diffusely involved segmental arteries from the pulmonary circulation, thus redistributing the pulmonary blood flow to less involved portions of lung. Complications and Hemoptysis Complications including coil migration, inadvertent air embolus with angina, or transient ischemic attack (TIA), pleurisy, and CHEST / 133 / 3 / MARCH,

4 Statistical Analysis All patient data were entered in an ACCESS (Microsoft Corporation; Redmond, WA) database and imported in statistical analysis software (SAS, Version 9.1; SAS Institute; Cary, NC) for analysis. Means, SDs, and percentages were used to describe the subjects. Comparisons between subjects with unilateral vs bilateral disease were conducted with t tests (continuous data) and 2 (categorical data) analysis. Results Thirty-six of 821 consecutive patients (4.4%) had diffuse PAVM involving one or more segmental pulmonary arteries. Twenty-one patients were female and 15 were male, with a mean age of years ( SD) at the time of presentation. HHT was present in 29 of 36 patients (81%). Mean follow-up was years (range, 0.12 to 26 years). PAVM were morphologically of the type seen in HHT and quite different from the type seen in hepatopulmonary syndrome or abdominal heterotaxy. 12 Complications are reported in Tables 1, 2. Nine of the 36 patients (25%) died during the follow-up period, and they all had bilateral involvement. None of the patients with unilateral involvement died, and this difference is statistically significant (p 0.03). Unilateral Diffuse PAVM Patients Among the 10 patients with unilateral involvement, 2 were female and 8 were male, and 7 of the 10 patients had HHT (Tables 1, 3). Two of the 10 patients were seen prior to Mean age at the first visit was years (range, 4 to 37 years), and mean follow-up time was years (range, 0.12 to 16 years). Eight of the 10 patients were seen after At the time of their first visit, 3 of the 10 patients had complications that included brain abscess (n 2) and hemoptysis (n 1). Since 1998, three patients have had significant complications and have recovered. All patients are living and attending school full-time or working. Figure 4. Postocclusion angiogram from the patient in Figure 1 (patient 28) after peripheral occlusion of most involved segments in right lower lobe. The new approach of occluding from distal to proximal using tightly packed coils is demonstrated in the posterior, anterior, and lateral segments of right lower lobe. thrombophlebitis occurring during embolotherapy or up to 30 days after TCE were considered early events. Complications that occurred 30 days after TCE were considered late events and included hemoptysis and neurologic events. Hemoptysis was defined as 100 ml of blood per day. Bilateral Diffuse PAVM Among the 26 patients with bilateral diffuse involvement, 19 were female and 7 were male, and 22 patients had HHT (Tables 2, 4). Thirteen of the 26 patients were seen prior to Mean age at the first visit was years (range, 0.5 to 71 years), and the mean follow-up time was years (range, 0.4 to 26 years). At the time of the first visit, 7 of 26 patients (27%) had one brain abscess and 2 patients had two brain abscesses. Nine of the 26 patients (35%) had a stroke/tia, and 1 patient had two strokes. 656 Original Research

5 Table 1 Patients With Unilateral Diffuse PAVM* Patient No. HHT Sex Age, yr Stroke/TIA BA HEMO Stroke/TIA BA HEMO Y/F 1 N M 33 N Y N N N N 16 2 Y F 31 N N N Y N N 10 3 Y M 22 N N N N Y N 4 4 Y M 37 N N N N N N 6 5 Y F 10 N N N N N Y 6 6 N M 23 N N N Y N N 1 7 N M 25 N N N N N N 1 8 Y M 7 N N N N N Y 6 9 Y M 4 N N N N N N 4 10 Y M 7 N N N N N N 1 *N no; Y yes; F female; M male; BA brain abscess; HEMO hemoptysis; Y/F years of continuous follow-up. Patients 1 and 2 were reported in a previous paper. 7 Patients 3 to 10 were seen after Patient 10 has not been treated. At presentation. Four of the 26 patients had hemoptysis. The first patient (patient 14) had not been treated. 7 The second patient (patient 12) had massive hemoptysis, and she was treated successfully by bronchial artery embolization. 7,13 Hemoptysis developed again in 2003, and she died in ,13 The third patient (patient 20) had many episodes of hemoptysis occurring spontaneously or with hard coughing during an upper respiratory infection (URI). He underwent bronchoscopy twice, and diffuse telangiectases of his tracheobronchial tree were demonstrated. He spontaneously improved after 1996 and has had no further Table 2 Patients With Bilateral Diffuse PAVM* Patient No. HHT Sex Age, yr Stroke/TIA BA HEMO Stroke/TIA BA HEMO Y/F 11 N F 29 N N N N N N Y F 24 N Y Y N N Y M 15 Y Y N N Y N Y F 14 N Y Y N N N 5 15 Y F 39 Y N N Y N N Y M 6 Y N N N N N 5 17 Y F 42 Y Y N N N N 6 18 Y F 20 Y N N Y Y Y F 17 Y N N Y N N Y M 20 Y N N N N Y Y F 50 Y Y N Y N N 9 22 N M 23 N N N N N N N F 0.8 N N N N N N Y F 12 N N N N N N 3 25 N F 0.5 N N N N N N Y F 43 N N N N N N 5 27 Y F 71 N N N Y N N 3 28 Y F 6 N N N N N N 4 29 Y F 4 N N N N N N 8 30 Y F 49 N N N N 2 N 9 31 Y F 28 N N N N N N 7 32 Y F 5 N N N N N N 5 33 Y M 9 N N N N N N 1 34 Y M 15 N N N N N N Y M 9 N N N N N N 4 36 Y F 31 N N N Y 1 1 spinal abscess N 10 *See Table 1 for expansion of abbreviations. Patients 11 to 23 were reported in a previous article. 7 Patients 24 to 36 were seen after At presentation. Patients who died during follow-up. CHEST / 133 / 3/ MARCH,

6 Table 3 Patients With Unilateral Diffuse PAVM: Distribution and Oximetry* Right Left Focal PAVM Oxygen Saturation, % Patient No. HHT Sex 50% 50% 50% 50% Unilateral Bilateral Initial Current 1 N M Y N N N Y N Y F Y N N N N N Y M N N N Y N N Y M N N N Y N N Y F Y N N N N N N M N N N Y N N N M N Y N N N N N/A N/A 8 Y M N Y N N Y N Y M Y N N N N N Y M N Y N N N N *Initial and current oxygen saturations are values in room air, usually standing when first seen by us; current values are latest values up to December Patient 10 was treated in another institution in May 2007 and finished after peripheral-to-central occlusion of three segments with an oxygen saturation of 98%. Patients 11 to 23 were reported in a previous article. 7 Patients 3 to 10 were seen after N/A not available; see Table 1 for expansion of abbreviations. 50% five or more segments of the right lung or four or more segments of the left lung diffusely involved; 50% fewer than five segments of the right lung and fewer than four segments of the left lung diffusely involved. hemoptysis. One other patient (patient 18) had massive spontaneous hemoptysis and was evaluated at another institution in She underwent partial right lower lobe lobectomy with resolution of hemoptysis and no recurrence. Since the first article, one patient had a fatal brain abscess. A second patient had a brain abscess with recovery, and a third patient had a significant paraspinal abscess with recovery. In addition, five patients had stroke/tia, all with recovery. Of the 17 patients living, 14 patients are attending school or working full-time. Three patients are considered disabled. Mortality During the 27-year follow-up period, no patient with unilateral diffuse PAVM died. Of the 26 patients with bilateral diffuse PAVM, there have been nine deaths, six of which occurred during the interval follow-up. These six deaths were due to brain abscess (patient 13), cerebral hemorrhage (patient 25), hemoptysis (patient 12), and liver complications (patients 17, 19, and 27). Two of the three patients died from subacute liver hemorrhage and biliary necrosis, a well-described syndrome, occurring in patients with HHT and liver malformations. 14 A third death occurred in an elderly woman who had progressive ammonia intoxication, most probably due to enlarged portal-to-hepatic vein communications. Mean age at the first visit of patients who died (25 16 years) was not significantly different than that of those who are still living (24 23 years) [p 0.22]; oxygen saturations at first visit were 78 7% vs 82 9%, respectively (p 0.18). While there was a trend for lower oxygen saturation in those who ultimately died, it was not significantly different from those who are living. Embolotherapy No patients had pulmonary hypertension, although some of the patients with liver disease had mild elevations of pulmonary artery pressure commensurate with elevations of left atrial pressure. Six of 36 patients underwent angiography and pressures without TCE. Three had temporary occlusion of the lower lobe pulmonary arteries as described previously without a rise 10 mm of Pao 2, and were not thought to be candidates for embolization. 7 The other three patients were relatively asymptomatic and did not require TCE. Two of 10 patients (20%) with unilateral diffuse PAVM and 17 of 26 patients (65%) with bilateral diffuse PAVM also had focal PAVM (p 0.02) [Tables 3, 4]. Those with arteries 3 mm were occluded as close to the aneurysmal sac as possible Peripheral-to-central occlusion of diffusely involved segments was performed in severely symptomatic patients using standard coil techniques. 16 This technique was utilized for all new patients requiring embolization and any of the first 15 reported who were severely symptomatic in follow-up. 7 Initial values of oxygen saturation for patients with unilateral and bilateral diffuse PAVM were 87 7% and 79 8%, respectively (p 0.02). The last or most current value obtained for patients with unilateral and bilateral involvement are 95 3% and 85 7%, respectively (p ) [Tables 3, 4]. 658 Original Research

7 Table 4 Patients With Bilateral Diffuse PAVM: Distribution and Oximetry* Right Left Focal PAVM Oxygen Saturation, % Patient No. HHT Sex 50% 50% 50% 50% Unilateral Bilateral Initial Current 11 N F Y N Y N N Y Y F Y N Y N N Y Y M Y N Y N N Y Y F Y N N Y N N 67 N/A 15 Y F Y N Y N N Y Y M Y N Y N N Y Y F Y N Y N N Y Y F Y N Y N N Y Y F Y N Y N Y N Y M Y N Y N N Y Y F Y N Y N N Y N M Y N Y N N N N F Y N Y N Y N 75 N/A 24 Y F Y N Y N Y N N F Y N Y N N N Y F Y N Y N N Y Y F Y N Y N N Y Y F Y N N Y N N Y F Y N Y N N N Y F Y N Y N N Y Y F N Y Y N N N Y F Y N Y N N N Y M N Y N Y N N Y M Y N N Y N N Y M Y N Y N N Y Y F Y N Y N Y N *Initial and current oxygen saturations are values in room air, usually standing, when first seen by us; current values are latest values up to December Patients 11 to 23 were reported in a previous article. 7 Patients 24 to 36 were seen after See Tables 1 and 3 for abbreviations and definitions. Complications Thirteen of 30 patients had early complications after embolization (Tables 1, 2). These included self-limited pleurisy in 11 of 12 patients and transient angina in 1 of 12 patients, most probably due to inadvertent air introduced at the time of embolization. These complications were transient with full recovery. All patients are heparinized during embolotherapy because of a thrombophlebitis in one patient reported previously. 7 Hemoptysis Six of 36 patients had hemoptysis sometime during their course: 1 patient with unilateral involvement and 1 patient with bilateral involvement before treatment, 7 and 4 patients with late involvement after TCE (Tables 1, 2). Hemoptysis in Patients With Unilateral Diffuse PAVM One patient had viral URI, leading to small amounts of hemoptysis that progressed to more severe hemoptysis (patient 5). She was successfully treated by bronchial artery embolization and has not recurred. She is aggressively treated with antitussives with codeine and phenergan to prevent hard coughing during URI. The second patient with unilateral diffuse PAVM treated initially by segmental TCE of three segments of the right lower lobe has hemoptysis that occurs either spontaneously or with URI (patient 8). Hemoptysis in Patients With Bilateral Diffuse PAVM Four of the patients with bilateral diffuse PAVM had hemoptysis. Patient 12 and 14 are discussed in the Mortality section. 7 Patient 20, with significant hemoptysis due to bronchial telangiectases, spontaneously improved; and patient 18 underwent a partial right lower lobectomy. No telangiectasia was seen on bronchoscopy, and she has not had a recurrence in 8 years. Discussion Patients with diffuse PAVM represent a small and severely affected subset of patients with PAVM, CHEST / 133 / 3/ MARCH,

8 most of whom have HHT. Very little has been published on these patients since our first attempt to characterize them. 7 This project was initiated by our observations that mortality was increasing; and hemoptysis, a relatively infrequent complication in patients with focal PAVM, was frequently seen in patients with diffuse PAVM, irrespective of TCE. These observations are consistent with the conclusion of Faughnan et al 7 that these patients are much more susceptible to complications of their disorder and require frequent re-evaluation. In addition, we began to question our approach of occluding diffusely involved segmental arteries proximally. Complications were less in the unilateral group, and their response to TCE was better than those with bilateral involvement. An interesting finding was that large focal PAVM were more common in the bilateral group. We still believe it is very important at the initial evaluation of patients with diffuse PAVM to treat all focal PAVM by TCE as has been described previously In our original article, 7 diffuse PAVM were defined as all segmental branches of one lobe affected. Many patients with diffusely involved segmental arteries will have normal segmental arteries in the same lobe. We believe it is more accurate to simply report segments diffusely involved. This distinction may be important because we have changed our approach to therapy. 7 Patients with only two or three segments diffusely involved have a sustained rise in oxygen saturation after peripheral-to-central occlusion. It is worth emphasizing that among our entire PAVM population, there are many with multifocal PAVM (Fig 3). It can be difficult to be certain which type is present. Patients with multifocal PAVM respond well to TCE and after occluding all PAVM with 2- to 3-mm arteries, and the resultant pulmonary angiogram will look almost normal. These patients were not included in our 2000 article nor in our current report. Patients with bilateral involvement have greater mortality rates than patients with unilateral involvement. Three of the six deaths in this expanded series were potentially preventable. One patient with a second brain abscess 21 years after his first brain abscess did not receive prompt medical attention. Any patient with diffuse or focal PAVM should seek medical evaluation for a headache lasting 24 h. The second potentially preventable death was due to massive hemoptysis and delay in seeking medical attention. Finally, an infant with diffuse bilateral involvement died from cerebral hemorrhage, presumably due to an undiagnosed cerebral arteriovenous malformation. 14 The other three deaths were due to liver complications, which, while rare in patients with HHT without PAVM, were more common in our patients, perhaps because of shunting of poorly oxygenated blood through multiple liver shunts. Up to 80% of patients with HHT have liver shunts In the absence of hypoxemia, these shunts cause symptoms due to high output failure in a small minority of patients. 22 In our patients, two had a syndrome of liver hemorrhage, pain, and elevated alkaline phosphatase, which has been referred to as liver disintegration. 23 The third patient had progressive ammonia intoxication and did not have acute symptoms. Neither age nor oxygen saturation were predictable of which patients were likely to succumb. At our initial evaluation, all focal PAVM with arteries 3 mm are occluded, and this remains our primary indication for embolotherapy. Based on our follow-up of patients since 2000, we are now occluding segmental pulmonary arteries with diffuse involvement from very distal to proximal with pushable fibered coils Our new approach to treatment of diffusely involved segmental arteries is performed only in very symptomatic patients. Symptoms that may be improved include fatigue and dyspnea in the absence of anemia. Many of our patients are relatively asymptomatic, despite very low oxygen saturation on room air, reflecting in our opinion the absence of airway issues or pulmonary hypertension. In these patients, we prefer to delay treatment until they become symptomatic. Hemoptysis has emerged in this expanded series as an important symptom that requires management. In most of our patients, hemoptysis is associated with URI. Rarely does it become massive ( 100 ml/d) if coughing is treated vigorously with antitussives with codeine. As part of a workup of a patient with hemoptysis, bronchoscopy is performed to identify telangiectasia of the tracheal bronchial mucosa. A recent report 24 identified approximately one third of patients with treated focal PAVM as having enlarged bronchial arteries. It is not clear whether enlarged bronchial arteries developed as a response to occlusion of focal PAVM or were part of the HHT syndrome. Anecdotally, we have noted this as well, and when a patient with diffuse hemoptysis does not respond to antitussives but progresses to massive hemoptysis, bronchial embolotherapy has been successful in controlling this symptom. When considering bronchial embolotherapy, large-particle embolization should be done. 13 Because of the small number of patients reported with diffuse PAVM, it is still difficult to make firm recommendations about management. However, we strongly recommend yearly follow-up visits. 660 Original Research

9 Conclusions Diffuse PAVM are associated with significant morbidity and mortality. In this article, we define diffuse pulmonary malformation as involvement of a single segmental rather than a whole lobe. Hemoptysis, often occurring with URI, should be treated with antitussives. If hemoptysis progresses, bronchial embolotherapy with large particles should be performed. We recommend yearly follow-up of this group of patients because they are at high risk for complications and they require reassessment on an ongoing basis. References 1 White RI Jr, Mitchell SE, Barth KH, et al. Angioarchitecture of pulmonary arteriovenous malformations: an important consideration before embolotherapy. AJR Am J Roentgenol 1983; 140: White RI Jr, Pollak JS, Wirth JA. Pulmonary arteriovenous malformations: diagnosis and transcatheter embolotherapy. J Vasc Interv Radiol 1996; 7: Guttmacher AE, Marchuk DA, White RI Jr. Hereditary hemorrhagic telangiectasia. N Engl J Med 1995; 333: Shovlin CL, Letarte M. Hereditary haemorrhagic telangiectasia and pulmonary arteriovenous malformations: issues in clinical management and review of pathogenic mechanisms. Thorax 1999; 54: Haitjema T, Disch F, Overtoom TT, et al. Screening family members of patients with hereditary hemorrhagic telangiectasia. Am J Med 1995; 99: Gossage JR, Kanj G. Pulmonary arteriovenous malformations: a state of the art review. Am J Respir Crit Care Med 1998; 158: Faughnan ME, Lui YW, Wirth JA, et al. Diffuse pulmonary arteriovenous malformations: characteristics and prognosis. Chest 2000; 117: Sands A, Dalzell E, Craig B, et al. Multiple intrapulmonary arteriovenous fistulas in childhood. Pediatr Cardiol 2000; 21: Hamada H, Terai M, Okajima Y, et al. Angiographical and computed tomographic findings in diffuse pulmonary arteriovenous fistulas. Int J Cardiol 1997; 59: Han S, Lim DJ, Chung YG, et al. The multiple brain abscesses associated with congenital pulmonary arteriovenous malformations: a case report. J Korean Med Sci 2002; 17: Shovlin CL, Guttmacher AE, Buscarini E, et al. Diagnostic criteria for hereditary hemorrhagic telangiectasia (Rendu- Osler-Weber syndrome). Am J Med Genet 2000; 91: Krowka MJ. Hepatopulmonary syndrome and portopulmonary hypertension: implications for liver transplantation. Clin Chest Med 2005; 26: Vinaya KN, White RI Jr, Sloan JM. Reassessing bronchial artery embolotherapy with newer pherical embolic materials [letter]. J Vasc Interv Radiol 2004; 15: Morgan T, McDonald J, Anderson C, et al. Intracranial hemorrhage in infants and children with hereditary hemorrhagic telangiectasia (Osler-Weber-Rendu syndrome). Pediatrics 2002; 109:E12 15 Pollak JS, Saluja S, Thabet A, et al. Clinical and anatomic outcomes after embolotherapy of pulmonary arteriovenous malformations. J Vasc Interv Radiol 2006; 17:35 44; 16 White RI Jr, Pollak JS. Controlled delivery of pushable fibered coils for large vessel embolotherapy. In: Springer, ed. Vascular embolotherapy: a comprehensive approach. Berlin, Germany: Jafar Golzarian, 2006; Pelage JP, Lacombe P, White RI Jr. Pulmonary arteriovenous malformations. In: Springer, ed. Vascular embolotherapy a comprehensive approach. Berlin, Germany: Jafar Golzarian, 2006; White RI Jr, Pollak JS. Pulmonary arteriovenous malformations. In: Peripheral vascular interventions. Philadelphia, PA: Lippincott Williams & Wilkins, 2008 (in press) 19 Garcia-Tsao G, Korzenik JR, Young L, et al. Liver disease in patients with hereditary hemorrhagic telangiectasia. N Engl J Med 2000; 343: Ianora AA, Memeo M, Sabba C, et al. Hereditary hemorrhagic telangiectasia: multi-detector row helical CT assessment of hepatic involvement. Radiology 2004; 230: Buscarini E, Plauchu H, Garcia Tsao G, et al. Liver involvement in hereditary hemorrhagic telangiectasia: consensus recommendations. Liver Int 2006; 26: Garcia-Tsao G. Liver involvement in hereditary hemorrhagic telangiectasia (HHT). J Hepatol 2007; 46: Blewitt RW, Brown CM, Wyatt JI. The pathology of acute hepatic disintegration in hereditary haemorrhagic telangiectasia. Histopathology 2003; 42: Brillet PY, Dumont P, Bouaziz N, et al. Pulmonary arteriovenous malformation treated with embolotherapy: systemic collateral supply at multidetector CT angiography after year follow-up. Radiology 2007; 242: CHEST / 133 / 3/ MARCH,