Bronchial Artery Embolization for HemoptysisDue to Benign Diseases: Immediate andLong-Term ResultsAkira Kato,1 Sho Kudo,1 Koichi Matsumoto,1 Tetsuhiro Fukahori,1 Toshihisa Shimizu,1

Akira Uchino,1 Shinichiro Hayashi2

1Department of Radiology, Saga Medical School, Nabeshima-5-1-1, Saga, 849, Japan2Department of Internal Medicine, Saga Medical School, Nabeshima-5-1-1, Saga, 849, Japan

AbstractPurpose: To clarify the immediate effect and long-termresults of bronchial artery embolization (BAE) for hemop-tysis due to benign diseases and the factors influencing theoutcomes.Methods:One hundred and one patients (aged 34–89 years)received bronchial artery embolization with polyvinyl alco-hol particles and gelatin sponge for massive or continuingmoderate hemoptysis caused by benign pulmonary diseasesand resistant to medical treatment.Results:After BAE, bleeding stopped in 94 patients (94%).The immediate effect was unfavorable in cases where feedervessels were overlooked or the embolization of the intercos-tal arteries was insufficient. Long-term cumulative hemop-tysis nonrecurrence rates after the initial embolization were77.7% for 1 year and 62.5% for 5 years. In bronchitis (n 59) and active tuberculosis (n 5 4) groups, an excellent(100%) 5-year cumulative nonrecurrence rate was obtained.The rate was lower in groups with pneumonia/abscess/pyo-thorax (n 5 8) or with pulmonary aspergillosis (n 5 9)(53.3%, 1-year cumulative nonrecurrence). There werehigher incidences of early recurrence among patients withmassive hemorrhage or more marked vascularity and sys-temic artery–pulmonary artery shunt in angiography: how-ever, these trends were not statistically significantConclusions:BAE can yield long-term benefit in patientswith hemoptysis due to benign diseases. Technical problemsin the procedure had an impact on the short-term effect. Thedegree of hemorrhage or the severity of angiographical find-ings were not significant factors affecting the outcome. Themost significant factor affecting long-term results was

whether the inflammation caused by the underlying diseasewas medically well controlled.

Key words: Bronchial artery—Therapeutic embolization—Benign pulmonary diseases

Bronchial artery embolization (BAE) has been established asan effective treatment for massive hemoptysis or hemoptysisthat is resistant to medical treatment [1–3]. There are severalreports concerning the long-term effects of BAE includingthe effect on neoplastic lesions [3–6]. However, there havebeen no reports regarding the factors that influence thelong-term results of BAE performed in patients with benignpulmonary diseases who have the potential for long-termsurvival.

This study was undertaken 1) to clarify the immediateeffect and long-term results of BAE for massive or chronicmoderate hemoptysis that were resistant to medical treat-ment, both of which are attributable to benign pulmonarydiseases; and 2) to clarify the factors that influence theseresults.

Materials and MethodsThe present study was conducted in 101 patients (73 men and 28women) aged 34-89 years (mean 60.4 years) who received BAE orbronchial and other systemic artery embolization for massive orcontinuing moderate hemoptysis resistant to medical treatment andcaused by benign pulmonary diseases. The study spanned a periodof 13 years and 3 months (from May 1984 to August 1997). Ofthese patients, 100 were involved in the investigation of the imme-diate effect and long-term results of BAE. We excluded one patientwho developed a minor complication during angiography prior toinjection of the embolic material and subsequently left the hospitalagainst medical advice.Correspondence to:A. Kato, M.D.

CardioVascularand InterventionalRadiology

© Springer-Verlag New York, Inc. 2000 Cardiovasc Intervent Radiol (2000) 23:351–357DOI: 10.1007/s002700010062

Table 1 describes the underlying diseases of these patients.Bronchiectasis (n 5 26) and inactive tuberculosis (n 5 20) werethe two most common diseases. All the patients with pulmonaryaspergillosis (n 5 9) had chronic pulmonary tuberculosis as acomplication. We classified one group as having idiopathic hemop-tysis (n 5 19) in which plain chest roentogenography (n 5 19),bronchoscopy (n 5 19), chest computed tomography (CT) (n 59), or clinical findings failed to reveal the underlying diseases.

The amount of hemoptysis at the time of initial BAE was lessthan 100 ml/day in 55 patients, between 100 and 500 ml/day in 39patients, and more than 500 ml/day (considered massive hemopty-sis) in 7 patients.

One hundred twenty-two embolization procedures were per-formed on 100 patients: one time in 86 patients, two times in 10patients, three times in 3 patients, and seven times in 1 patient. Thearteries embolized in the procedure were bronchial artery alone in81 patients, bronchial and intercostal arteries in 17 patients, acombination of bronchial, intercostal, and branches of the subcla-vian artery in 1 patient, and intercostal arteries and branches of thesubclavian artery in 1 patient. The embolic material used in themajority of patients (n 5 87) was a combination of polyvinylalcohol (PVA) particles 200–1000mm in diameter and strippedgelatin sponge (GS) 1–2 mm square. In the early phase of thisstudy, PVA alone was used in nine patients and a coil in onepatient. Three patients received GS cubes alone. Embolization wasperformed when angiography revealed hypervascularity or bronchi-al-pulmonary arterial shunts on the hemorrhaging side. The non-hemorrhaging side was also embolized when the angiographicfindings were severe. We started using 3.0 Fr microcatheters in1992 and have performed superselective embolization in 13 pa-tients.

At our institution, bronchoscopy was performed initially in allpatients with hemoptysis to detect and treat the bleeding sites. Forthose patients with relatively large amounts of hemoptysis orhemoptysis resistant to medical treatment, angiography and embo-lization were contemplated based on a discussion between respira-tory physicians and radiologists. Emergency BAE was performed in9 out of 101 patients because they had relatively massive hemop-tysis that was resistant to the endoscopic treatment. The remainingpatients with continuing moderate hemoptysis resistant to medicaltreatment received scheduled BAE.

Investigation of the underlying diseases, amounts of hemor-rhage, and the follow-up of patients’ conditions was conducted byreviewing the patients’ medical records and interviewing the pa-tients by telephone. The observation of 13 patients who receivedpneumonectomy after the embolization was terminated at the time

of surgery. The shortest observation period was 10 days for apatient who died due to recurrent hemoptysis after the emboliza-tion. The longest observation was 158 months, and the meanobservation period was 45.8 months. Of the patients whose recordsshowed no recurrent hemoptysis, those with observation periods ofup to 60 months answered a telephone questionnaire regarding thepresence/absence, severity, and timing of any recurrent hemoptysis.

Our study examined the following points: 1) overall immediateeffect of embolization and the long-term result; 2) immediate effectand long-term result according to underlying disease; 3) relationbetween the amount of hemorrhage and the long-term result; 4)relation between long-term result and angiographic findings (thedegree of vascularity and extent of bronchial or systemic-pulmo-nary arterial shunts); and 5) incidence and gravity of complications.

The immediate effect was assessed by the nonrecurrence rate ofhemoptysis for the first month after BAE. The long-term resultswere assessed by the 1-, 3-, and 5-year cumulative hemoptysisnonrecurrence rates. The relations between long-term results andvarious factors were analyzed by the generalized Wilcoxon test. Inthis study, recurrence of hemoptysis was defined as expectorationof clot or fresh blood after BAE.

Angiographic findings were analyzed by three radiologists. Thedegree of vascularity was classified macroscopically into threelevels: mild (brush-like neovascularity alone), moderate (brush-likeneovascularity with mildly dilated and tortuous vessels), andmarked (cluster of markedly dilated and tortuous vessels) (Fig. 1).The extent of bronchial or systemic-pulmonary arterial shunts wasclassified into three levels: none, mild (visible only in the peripheralregions of pulmonary arteries), and marked (reverse flow near thepulmonary hilum) (Fig. 2).

ResultsImmediate Effect and Long-Term Results

Immediate effect of embolization.After embolization, bleed-ing stopped in 94 of 100 patients (Table 2). Of the sixpatients who had recurrent hemoptysis within the first monthafter embolization, one patient with pulmonary aspergillosisdied of recurrent hemoptysis 10 days after the procedure. Inthis patient the intercostal arteries, which were feeding thelesion, were not embolized because the hemoptysis was mildand the catheter placement was not secure (Fig. 3). However,the shunt to the pulmonary artery was still present at com-pletion of the embolization of the bronchial artery. Onepatient with pulmonary abscess had recurrent hemoptysis 10days after the BAE. This patient was observed conserva-tively, but died due to massive hemoptysis 3 months later.This was also likely due to insufficient embolization of theintercostal arteries. Three patients (two with pulmonary ab-scess and one with inactive tuberculosis) underwent opera-tions after recurrent hemoptysis as their conditions werejudged to be difficult to control by embolization. The otherpatient with inactive tuberculosis received reembolizationand had no further recurrence of hemoptysis until the patientdied of other disease (13 months after embolization). Of thethree patients who had angiography immediately after therecurrence of hemoptysis, we found an overlooked feeder

Table 1. Underlying diseases

No. of patients

Bronchiectasis 26Inactive tuberculosis 20Pulmonary aspergillosis 9Bronchitis 9Pneumonia/abscess/pyothorax 8Active tuberculosis 4Pneumoconiosis 4Atypical mycobacteriosis 1Wegener granulomatosis 1Idiopathic hemoptysis 19Total 101

352 A. Kato et al.: BAE for Hemoptysis Due to Benign Diseases

vessel in one and recanalization of the embolized feedervessels in two.

Long-term results of embolization.Table 2 and Figure 4show cumulative nonrecurrence rates after initial emboliza-tion and after repeat embolizations. The cumulative nonre-currence rates after initial embolizations were 77.7% at 1

Fig. 1. Degrees of vascularity detected by angiography.Vascularity was classified macroscopically by three radiolo-gists as mild: brushlike neovascularity alone (A), moderate:

brushlike neovascularity with mildly dilated and tortuous ves-sels (B), or marked: cluster of markedly dilated and tortuousvessel (C).

Fig. 2. Extent of bronchial or systemic-pulmonary arterialshunts. Extent was classified as none, mild (visible only in theperipheral regions of the pulmonary arteries) (A), or marked

(pulmonary arteries visible close to the pulmonary hilum) (B,C).

Table 2. Cumulative nonrecurrent rates of hemoptysis after initial BAE andrepeat BAE

Cumulative nonrecurrent rates (%)

1 month 1 year 3 years 5 years

Initial BAE only 94.0 77.7 69.6 62.5Repeat BAE 95.0 84.3 77.2 71.6

A. Kato et al.: BAE for Hemoptysis Due to Benign Diseases 353

year, 69.6% at 3 years, and 62.5% at 5 years. The rates afterrepeat procedures were 84.3% at 1 year, 77.2% at 3 years,and 71.6% at 5 years.

Immediate Effect and Long-Term Results inEach Underlying Disease

The immediate effect and long-term results for each under-lying disease are shown in Table 3. In the bronchitis andactive tuberculosis groups, an excellent 5-year cumulativenonrecurrence rate (100%) was obtained. In the pulmonaryaspergillosis and pneumonia/abscess/pyothorax groups,1-year cumulative nonrecurrence rate was 53.3% for eachgroup. There was a statistically significant difference be-tween these groups and the other disease groups (p , 0.05,generalized Wilcoxon test). The remaining disease groupsshowed intermediate rates with no statistically significantdifferences among any of the groups.

Relationship Between Amount of HemorrhageBefore the Procedure and Long-Term Results

The massive hemoptysis group (500 ml/day) showed thehighest incidence of early recurrence, but this difference wasnot statistically significant (Fig. 5).

Relationship Between Angiographic Findingsand Long-Term Results

Relationship between degree of vascularity and recurrence-free interval.Vascularity was mild in 18 patients, moderatein 33 patients, and marked in 49 patients. There was a higherincidence of early recurrence in patients with more markedvascularity; however, this trend was not statistically signif-icant (Fig. 6).

Fig. 3. A patient withbronchopulmonary aspergillosis whodied 10 days after the BAE ofrecurrent massive hemoptysis. A Therewas a single right bronchial artery withmassive shunts to the pulmonaryartery. B Shunts to the pulmonaryartery were found from the 4th and 5thintercostal arteries. Catheterization ofthe intercostal arteries and distalembolization by bypassing theradicular branches (arrows) wereplanned. However, we were unable toperform the procedure because of thetight winding of the vessel.

Table 3. Cumulative nonrecurrent rates of hemoptysis according to disease

Cumulative nonrecurrent rates (%)

1 month 1 year 3 years 5 years

Bronchiectasis (n 5 26) 100 70.2 65.5 55Inactive tuberculosis (n 5 20) 95 80 69.3 69.3Pulmonary aspergillosis (n 5 9) 88 53.3 53.3 53.3Bronchitis (n 5 9) 100 100 100 100Pneumonia/abscess/pyothorax (n 5 8) 63.3 53.3 53.3 53.3Active tuberculosis (n 5 4) 100 100 100 100Pneumoconiosis (n 5 4) 100 75 50 50Idiopathic hemoptysis (n 5 19) 100 93.8 63.5 63.5

Atypical mycobacteriosis (n 5 1): recurrence 57 months after BAEWegener granulomatosis (n 5 1): died of renal failure 10 days after BAE without recurrence

354 A. Kato et al.: BAE for Hemoptysis Due to Benign Diseases

Relationship between extent of bronchial or systemic-pulmo-nary arterial shunts and recurrence-free interval.Shuntswere not observed in 29 patients, mild in 42 patients, andmarked in 29 patients. There was a higher incidence of earlyrecurrence in patients with more marked shunts; however,this trend was not statistically significant (Fig. 7).

Incidence and Gravity of Complications

Bronchial arterial damage caused by the catheter tip or by theinjected embolic material was found in 5 (5%) of 101 pa-tients (2 intimal dissection, 2 rupture of bronchial artery, and1 arterial penetration by the guidewire). Focal aortic dissec-tion was found in one patient (1%). Of these patients, tworeceived coil embolization to repair the damaged vessels.The remaining four had no further treatment, and no prob-lems were found while the patients were being observed. No

serious complications such as spinal infarction were encoun-tered.

DiscussionPrevious studies reported nonrecurrence rates of 77%–91%for the immediate effect of BAE for hemoptysis [1–8]. Amore favorable rate (94%) was obtained in our study. Itappeared that the immediate effect was unfavorable in caseswhere feeder vessels associated with the hemoptysis wereoverlooked or where the embolization of the intercostalartery was insufficient. Another cause of recurrent hemop-tysis within 1 month was early recanalization. We experi-enced two cases of recurrent hemoptysis, possibly due tounexpected early recanalization of the vessel within 1 month,which had been considered adequately embolized. We used

Fig. 4. Cumulative nonrecurrence rates after initial embo-lization and after repeat embolization.

Fig. 5. Cumulative nonrecurrence rates of massive (.500ml/day), moderate (100–500 ml/day), and mild (,100 ml/day)hemoptysis. The massive hemoptysis group showed thehighest incidence of early recurrence, but this difference wasnot statistically significant.

Fig. 6. Cumulative nonrecurrence rates according to thedegree of vascularity. The marked hyperplasia group showedthe highest incidence of early recurrence, but the differencewas not statistically significant.

Fig. 7. Cumulative nonrecurrence rates according to theextent of bronchial or systemic-pulmonary arterial shunts.The marked shunt group showed the highest incidence ofearly recurrence, but the difference was not statistically sig-nificant.

A. Kato et al.: BAE for Hemoptysis Due to Benign Diseases 355

primarily a combination of PVA and GS as the embolicmaterial since GS prevents aggregation of PVA particles inthe catheter. However, in our method the amount of GS mayexceed that of PVA. This may explain the early recanaliza-tion observed in this study.

PVA is a nonabsorbable substance commonly used toembolize visceral or central nervous system lesions [9–12].In histopathologic studies, PVA induces a rapid, strong for-eign body reaction in the wall of the embolized vessel andperivascular tissue, however, this change is seen only in theacute phase [13–15]. In addition, no complications associ-ated with PVA embolotherapy were reported in previousstudies of BAE [16, 17]. Thus, PVA is considered a safeembolic material in embolization of the bronchial artery andalso the intercostal artery. Furthermore, the favorable imme-diate result in our study suggests that its use may prevent therecurrence of hemoptysis due to early recanalization of bron-chial artery.

As for the long-term effect on hemoptysis, the cumulativenonrecurrence rates in patients who underwent BAE onlyonce were 77.7% at 1 year, 69.6% at 3 years, and 62.5% at5 years. These results were similar to those reported previ-ously [3–8]. However, by performing additional BAE forthose patients who had recurrent hemoptysis, these ratesincreased to 84.3% at 1 year, 77.2% at 3 years, and 71.6% at5 years. This result suggests that outcome can be improvedby repeating BAE even in patients with recurrent hemopty-sis. This improvement is most likely to occur when there-embolization allows compensation for procedural insuffi-ciencies, such as overlooking feeder vessels, in the initialBAE.

Our results revealed a trend toward early recurrent he-moptysis in the massive hemoptysis group and the groupwith angiographically marked vascularity with markedshunts although it was not statistically significant. The re-sults in the pneumonia/abscess/pyothorax and pulmonaryaspergillosis groups were unfavorable, similar to those inprevious reports [1, 4, 6, 18]. Our study showed that factorsother than underlying disease have no statistically significantimpact on the long-term results. The clinical significanceobserved in the groups with pneumonia/abscess/pyothoraxor with pulmonary aspergillosis appeared to be the difficultyin controlling the pulmonary inflammatory lesions by med-ical treatment. The size of the embolic material used in BAEis large enough to preserve the capillary beds. Therefore,even with nonabsorbable embolic material like PVA, thetreatment is palliative over an extended period because of thepossible development of collateral vessels and recanalizationof embolized vessels. Thus, the management of inflamma-tion attributable to the underlying diseases is critically im-portant.

No serious complications occurred in our patients, whichleads us to conclude that BAE is a relatively safe procedure.However, one patient died 10 days after BAE due to hem-orrhage, presumably from the intercostal artery that was notembolized. The proper strategy for intercostal artery embo-

lization has been discussed extensively but remains contro-versial. It has been claimed that spinal infarction can beavoided with the use of large GS strips (33 10 mm) evenwhen the angiographic image of the intercostal artery depictsspinal arteries [3]. Yet, one report insists that even by usinglarge GS strips, intercostal artery embolization caused spinalinfarction [19]. Other authors reported that intercostal arteryembolization caused spinal infarction in patients who hadprevious BAE or surgery, though no spinal arteries had beendepicted [20]. This complication could be avoided by per-forming embolization from the peripheral portion of theartery distal to the origin of the connecting branches to thespine using a microcatheter [19]. In the patient who died ofrecurrent hemoptysis, we did withdraw the intercostal arteryembolization because the hemoptysis was mild and the mi-crocatheter could not be securely placed beyond the origin ofthe radicular arteries. The best approach in such a case is stillunknown, though we should have performed the proximalintercostal artery embolization with a larger embolic mate-rial, given the fact that the patient had no visible spinal arteryon the angiogram.

In conclusion, BAE is apparently a safe and effectivemethod of treatment for hemoptysis caused by benign dis-ease. In BAE for massive or continuing moderate hemopty-sis due to benign pulmonary diseases, the technical difficultyof the procedure had the largest impact on the short-termeffects. The amount of hemorrhage or the severity of angio-graphic findings were not significant factors affecting thelong-term outcome. For prevention of recurrent hemoptysisit was important after performing BAE to medically controlthe inflammation induced by the underlying disease.

References1. Remy J, Arnaud A, Fardon H, Giraud R, Voisin C (1977) Treatment of

hemoptysis by embolization of bronchial arteries. Radiology 122:33–372. Ramakantan R, Bandekar VG, Gandhi MS, Aulakh BG, Deshmukh HL

(1996) Massive hemoptysis due to pulmonary tuberculosis: Controlwith bronchial artery embolization. Radiology 200:691–694

3. Uflacker R, Kaemmerer A, Nerves C, Picon PD (1983) Management ofmassive hemoptysis by bronchial artery embolization. Radiology 146:627–634

4. Uflacker R, Kaemmerer A, Picon PD, Rizzon CFC, Neves CMC,Oliveira ESB, Oliveira MEM, Azevedo SNB, Ossanai R (1985) Bron-chial artery embolization in the management of hemoptysis: Technicalaspect and long-term results. Radiology 157:637–644

5. Rabkin JE, Astafjev VI, Gothman LN, Grigorjev YG (1987) Transcath-eter embolization in the management of pulmonary hemorrhage. Radi-ology 163:361–365

6. Hayakawa K, Tanaka F, Torizuka T (1992) Bronchial artery emboli-zation for hemoptysis: Immediate and long-term results. CardiovascIntervent Radiol 15:154–158

7. Matsumoto S, Kishikawa T, Kudo S, Matsuio Y, Yamada H, Katoh O(1991) Bronchial and non-bronchial systemic artery embolization forhemoptysis due to benign lung diseases: Immediate effect and long-term results. Nippon Acta Radiol 51:1027–1036

8. Tamura S, Kodama T, Otsuka N, Kihara Y, Nisikawa K, Yuki Y,Samejima M, Uwada O, Watanabe K, Minoda S (1993) Embolotherapyfor persistent hemoptysis: The significance of pleural thickening. Car-diovasc Intervent Radiol 16:85–88

9. Chuang VP, Soo CS, Wallace S (1980) Ivalon embolization in abdom-inal neoplasms. AJR 136:729–733

10. Kusano S, Murata K, Ohuchi H, Motohashi O, Atari H (1987) Low-

356 A. Kato et al.: BAE for Hemoptysis Due to Benign Diseases

dose particulate polyvinyl alcohol embolization in massive small intes-tinal hemorrhage: Experimental and clinical results. Invest Radiol 22:388–392

11. Berenstein A, Kricheff II (1981) Microembolization techniques ofvascular occlusion: Radiologic, pathologic and clinical correlation.AJNR 2:261–267

12. Scialfa G, Scoatti G (1985) Superselective injection of polyvinyl alco-hol microemboli for the treatment of cerebral arteriovenous malforma-tions. AJNR 6:957–960

13. Germano IM, Davis RL, Wilson CB, Hieshima GB (1992) Histopatho-logical follow-up study of 66 cerebral arteriovenous malformationsafter therapeutic embolization with polyvinyl alcohol. J Neurosurg76:607–614

14. Tomashefski JF, Cohen AM, Doershuk CF (1988) Long-term his-topathological follow-up of bronchial arteries after therapeutic embo-lization with polyvinyl alcohol (Ivalon) in patients with cystic fibrosis.Hum Pathol 19:555–561

15. Castan˜eda-Zuniga WR, Sanchez R, Amplatz K (1978) Experimental

observations on short- and long-term effects of arterial occlusion withIvalon. Radiology 126:783–785

16. Cremaschi P, Naschimbene C, Vitulo P, Catanese C, Rota L, BarazzoniGC, Cornalba GP (1993) Therapeutic embolization of bronchial artery:A successful treatment in 209 cases of relapse hemoptysis. Angiology44:295–299

17. Cohen AM, Doershuk CF, Stern RC (1990) Bronchial artery emboli-zation to control hemoptysis in cystic fibrosis. Radiology 175:401–405

18. Katoh O, Kishikawa T, Yamada H, Matsumoto S, Kudo S (1990)Recurrent bleeding after arterial embolization in patients with hemop-tysis. Chest 97:541–546

19. Tanaka N, Yamakado K, Murashima S, Takeda K, Mastumura K,Nakagawa T, Takano K, Ono M, Hattori T (1997) Superselectivebronchial artery embolization for hemoptysis with a coaxial microcath-eter system. J Vasc Interv Radiol 8:65–70

20. Vujic I, Pyle R, Parker E, Mithoefer J (1980) Control of massivehemoptysis by embolization of intercostal arteries. Radiology 137:617–620

A. Kato et al.: BAE for Hemoptysis Due to Benign Diseases 357