- Email: [email protected]
Hepatic infarction following selective hepatic artery embolization with microcoils for iatrogenic biliary hemorrhage
Hepatology Research 30 (2004) 42–50
Hepatic infarction following selective hepatic artery embolization with microcoils for iatrogenic biliary hemorrhage Manabu Hashimoto∗ , Yoko Akabane, Jyouichi Heianna, Etuko Tate, Koichi Ishiyama, Toshiaki Nishii, Jiro Watarai Department of Radiology, Akita University School of Medicine, Hondo 1-1-1, Akita City, Akita 010-8543, Japan Received 5 February 2004; received in revised form 12 April 2004; accepted 7 May 2004 Available online 2 July 2004
Abstract Purpose: The aim of this study was to explore the relation of collateral filling to ischemic or infarcted liver following selective embolization of hepatic artery with microcoils in patients with iatrogenic hemobilia. Methods: We performed retrospective analysis of clinical outcomes and post-embolization angiograms in eight patients (mean age of 66 years) studied over the last 7 years. Hemobilia occurred after percutaneous biliary drainage (n = 5) and percutaneous hepatic biopsy (n = 3). Causes of bleeding were pseudoaneurysm (n = 6), arterial laceration (n = 1), and direct hepatic artery-to-biliary duct fistula (n = 1). We placed microcoils in the subsegmental (n = 4) or segmental branch (n = 2), or both branches (n = 2), distal and proximal to the bleeding point. Results: We obtained complete hemostasis in all patients (100%). Four patients had no hepatic infarction after embolization. Normal filling of the distal part of the embolized branch through collaterals was seen on post-embolization films. Four patients with no collateral filling experienced liver infarction in the area corresponding to embolized branch. One patient with severe portal stenosis died of hepatic failure. Conclusion: Hepatic infarction is related to lack of immediate collateral flow. © 2004 Elsevier B.V. All rights reserved. Keywords: Liver; Hemobilia; Embolization; Complication
1. Introduction
2. Materials and methods
Transcatheter arterial embolization (TAE) has been established as an alternative to surgical intervention for obtaining hemostasis for gastroduodenal hemorrhage or hemobilia that can occur after various surgical and interventional treatments [1–7]. Although TAE is considered to be relatively safe, several complications may occur, such as liver abscess, liver ischemia, or liver failure [4,5,7–9]. To enhance the safety of TAE, it is important to understand complications of the procedure. In this study, we analyze the incidence and distribution of ischemic or infarcted liver following selective hepatic artery embolization with microcoils in patients with iatrogenic hemobilia. Then, we investigate the relation of collateral filling to ischemic or infarcted liver.
We performed retrospective analysis of clinical outcomes and post-embolization angiographic findings in eight consecutive patients undergoing selective hepatic embolization for iatrogenic hemobilia from 1996 to 2002 (Table 1). The patients comprised two women and six men with a mean age of 66 years. Seven patients had malignancies: hilar cholangiocarcinoma (n = 5), intrahepatic cholangiocarcinoma (n = 1), metastatic liver tumor (n = 1). The remaining patient had chronic hepatitis. None of the patients had a history of hepatic viral infection (hepatitis B or C) or liver cirrhosis. Hemobilia occurred after percutaneous transhepatic biliary drainage (PTBD) in five patients and percutaneous liver biopsy in three. The clinical manifestations were bleeding from a PTBD tube in two patients and hematemesis in six. Emergency gastrointestinal endoscopy established a diagnosis of hemobilia in the six patients with hematemesis. All patients were referred to our department within 2 weeks of PTBD or biopsy because hemostasis could not be obtained through conservative treatment.
∗
Corresponding author. Fax: +81 188 36 2623. E-mail address: [email protected] (M. Hashimoto).
1386-6346/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.hepres.2004.05.002
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
43
Table 1 Patient background No.
Age (year)
Sex
Disease
Underlying interventional procedure
Bleeding site
Angiographic findings
Portal flow abnomality
1 2
67 65
M M
Hilar cholangiocarcinoma Hilar cholangiocarcinoma
Drainage Drainage
Laceration Pseudoaneurysm
– –
3
63
F
Hilar cholangiocarcinoma
Drainage
Pseudoaneurysm
–
4 5
67 61
M M
Hilar cholangiocarcinoma Intrahepatic cholangiocarcinoma
Drainage Biopsy
A3 Right posterior segmental branch Right anterior segmental branch A8 A8
Pseudoaneurysm Pseudoaneurysm
6 7 8
72 69 68
F M M
Hilar cholangiocarcinoma Metastatic liver tumor Chronic hepatitis
Drainage Biopsy Biopsy
A3 A8 A6
Pseudoaneurysm Pseudoaneurysm A-B fistula
Anterior branch occlusion AP shunts (R) anterior branch occlusion Right PV stenosis – –
A3: left lateral anterior branch; A6: right posterior inferior branch; A8: right anterior superior branch; R: right lobe of the liver; Anterior branch: anterior branch of right portal vein; PV: portal vein; AP: hepatic artery-to-portal vein; A-B: hepatic artery-to-bile duct.
Pre-embolization angiographic findings included hepatic artery branch pseudoaneurysm in six patients, arterial transgression by the drainage catheter in one, and direct hepatic artery-to-biliary fistula in one. Three patients had portal abnormalities on angiograms. One patient with hilar cholangiocarcinoma had severe stenosis in the right portal vein. One with intrahepatic cholangiocarcinoma had occlusion of the anterior branch of the right portal vein. Hepatic artery-to-portal vein (AP) shunts throughout the anterior segment of the right lobe were also seen in this patient. The remaining one patient had occlusion of the right anterior portal vein. These portal abnormalities were caused by hepatic hilar tumors. We performed selective hepatic arterial embolization as follows. Diagnostic hepatic angiography was performed via
the femoral artery and included selective celiac, superior mesenteric, and common hepatic arteriography. After obtaining a diagnosis of pseudoaneurysm, arterial laceration, or direct hepatic artery-to-biliary fistula, a microcatheter (Microferret, William Cook Europe, Denmark) was inserted into a 4.2-F diagnostic angiography catheter (Hanako, Tokyo, Japan), which remained in the common, proper, or right or left hepatic artery. We advanced a fine guidewire (Transend EX, Boston Scientific Target, CA) through the microcatheter and into the corresponding arterial branch and distal to the pseudoaneurysm, laceration, or fistula. The microcatheter was then threaded over the fine guidewire. We placed adequately sized platinum microcoils (Vortex, Boston Scientific Target) in the corresponding branch distal and proximal to the bleeding point. Embolization was performed un-
Table 2 Patient outcomes No. Embolized artery Opacification of distal part through intrahepatic collaterals
Liver infarction
1 2
Extrahepatic (+) +
– –
37/26 43/25
34/24 42/23
None None
Uneventful Uneventful
+
–
38/34
35/29
None
Uneventful
110/105 61/56
98/96 203/159
None None
Uneventful Died of malignancy 22 months after embolization Died of hepatic failure 14 days after embolization Uneventful Uneventful
3
4 5
6
7 8
Area of infarction on CT
AST/ALT values (U/L) Pre
Re-bleeding Outcomes
Post
A3 Right posterior segmental branch Right anterior segmental branch A8 A8 + right anterior segmental branch A3 + left hepatic artery
+ A8 (−), A5 (+)
– S8
Patchy distribution
A2 (−), A3 (−), A4 (+)
Lateral segment
Throughout lateral segment
49/21
1251/663
None
A8 A6
– –
S8 S6
Wedge-shaped Ill-defined
10/11 93/75
598/692 143/109
None None
A2: left lateral posterior branch; A3: left lateral anterior branch; A4: left medial branch; A5: right anterior inferior branch; A6: right posterior inferior branch; A8: right anterior superior branch; S6: right posterior inferior segment; S8: right anterior superior segment. Lateral segment: left lateral segment. Pre: pre-embolization; Post: post-embolization.
44
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
Fig. 1. Images obtained in a 67-year-old man (no. 1). (a) Left hepatic arteriogram shows laceration of the left lateral anterior branch of the hepatic artery (arrowhead) secondary to percutaneous transhepatic catheter placement in the anterior branch of the left lateral bile duct (B3). Another catheter is inserted from the dorsal branch of the left lateral bile duct. Massive bleeding was seen from the drainage route after removal of the catheter in the B3. (b) Angiogram after TAE shows normal filling of the distal part of the embolized branch through gastric artery (arrowheads). Liver infarction did not occur after TAE in this patient.
til the flow ceased completely; angiograms through the microcatheter showed elimination of the focal arterial lesions from the circulation and stagnant contrast medium in the proximal part of the corresponding vessels. The placement sites of the coils varied from patient to patient according to the vascular anatomy and the sites of the bleeding. After embolization, we performed immediate hepatic angiography through the diagnostic angiography catheter, which had remained in the common, proper, or right or left hepatic artery as a control. Images were obtained by digital subtraction angiography.
Clinical examination, complete blood count, and liver function tests were performed in all patients within 1 week of TAE. The patients with suspected hepatic complications underwent emergent CT (plain or plain and contrast-enhanced CT). We also performed plain and contrast-enhanced CT in 7 patients 1 month after embolization for follow-up. The diagnosis of liver ischemia or infarction was based on a temporary increased in transaminase levels and development of low-attenuation areas corresponding to the embolized arterial branches on emergent CT. We evaluated the presence of low-attenuation areas in pre-contrast or
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
pre-contrast and post-contrast during equilibrium phase CT images.
3. Results Selective embolization with microcoils was performed successfully in all eight patients at the level of the subseg-
45
mental (n = 4) or segmental branches (n = 2), or both (n = 2) (Table 2). Clinical course after embolization was uneventful in four of the patients. Blood examinations and CT images (obtained 1 month after embolization) showed no remarkable changes related to embolization. The post-embolization films showed normal filling of the peripheral branches of the embolized artery through collaterals (Fig. 1).
Fig. 2. Images obtained in a 61-year-old man (no. 5) show development of liver infarction after TAE. The patient had undergone liver biopsy for liver tumor around the hepatic hilum. (a) Common hepatic arteriogram shows a pseudoaneurysm (arrow) of the right anterior superior branch of the right hepatic artery (A8) (arrowhead). (b) Selective arteriogram of the anterior branch of the right hepatic artery shows the pseudoaneurysm and opacification of the right anterior superior (P8, small arrowheads) and inferior (P5, large arrowhead) branches of the portal vein. Anterior inferior branch of the right hepatic artery (A5) is also seen (arrows). After advancement of a microcatheter distally, we placed microcoils around the pseudoaneurysm in this branch. (c) Common hepatic arteriogram obtained after embolization shows absence of opacification of the aneurysm and a branch of the A8 (arrowhead in (a)). Other branches of the A8 are seen (small arrow) through intrahepatic collateral filling but peripheral parts are poorly visualized. Distal parts of the A5 (arrows) are adequately filled through intrahepatic collaterals. Enhanced CT scans obtained 4 days after TAE show liver infarction in the two portions of the right anterior superior segment corresponding to the areas of absence of filling of the branch of A8 (d) and the poorly visualized branch (e), respectively.
46
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
Fig. 2. (Continued ).
The other four patients had liver infarction in the areas corresponding to the embolized arterial branches. They experienced no abdominal pain or fever after the embolization procedure, but we noted a temporary increased in transaminase levels (maximum value of AST: 203, 1251, 598 and 143 U/L; ALT: 159, 663, 692 and 109 U/L, respectively, in these four patients) indicating a transient liver ischemia. Emergent CT showed development of low-attenuation areas corresponding to the embolized arterial branches. On CT, the infarction was shown to occur entirely in the involved segment in one patient. One patient had a wedge-shaped infarction. Hepatic infarction appeared as a patchy distribution in one patient (Fig. 2) and as an ill-defined area of low attenuation with liver surface predominance in another (Fig. 3). The values of the measured transaminases gradually recov-
ered to normal within 2 weeks. On control angiography, opacification of the distal portion of the embolized branches was not obtained through collaterals in the areas of infarction (Figs. 2 and 3). Infarctions did not occur in the areas in which peripheral branches were filled through intrahepatic collaterals in two of four patients. Clinical follow-up of from 2 weeks to 60 months (mean, 21 months) was obtained on all eight patients. No recurrence of bleeding was observed; however, one patient (no. 6) with hepatic infarction died of hepatic failure 14 days after TAE. This patient had severe stenosis of the right portal vein. Follow-up CT of the other three patients with hepatic infarction revealed gradual resolution of the infarcted area. One patient (no. 4) died of associated malignancy after 22 months.
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
47
Fig. 2. (Continued ).
4. Discussion Various focal arterial lesions leading to hepatic arterial-biliary fistula, such as pseudoaneurysm, arterial laceration, and direct hepatic arterial-biliary fistula, can occur after surgical and interventional procedures of the liver. The purpose of embolization therapy for these focal lesions is to eliminate focal arterial lesions from the circulation and to obtain thrombosis of this segmental pedicle. Most previously published papers have described embolization for pseudoaneurysms [1,2,4,7,10–13]. Microcoils should be used initially for TAE and should be placed both proximally and distally to the bleeding point [5,7,14] because TAE done with particle materials (such as Gelfoam) often increase intravascular pressure when the particles are released from the catheter and induce rupture of the pseudoaneurysm [5]. With the advent of and subsequent advancements in microcatheter systems, it has become relatively easy to place microcoils proximally and distally to the bleeding site, even in small-caliber branches [2,15] such as those in our patients. The selective embolization method with microcatheter and microcoils has a high success rate for achieving hemostasis [2,5]; complete hemostasis can be obtained in most patients when the microcatheter is advanced for enough into the target vessel that the microcoils can be placed distally and proximally to the bleeding point. Furthermore, this method can avoid the needless occlusion of a normal branch. The success rate was 100% in our patient series. Previous reports have stated that some patients had complications associated with embolization of the hepatic arteries after hemostasis, with complication rates ranging from 11% [7] to 60% [4,5]. None of these reports described precisely how hepatic infarction occurred after TAE. Although there were a limited number of patients in our study, some
observations can be made. We found a strong association between the occurrence of infarction and the findings on the post-embolization film. When there was no opacification of the distal part of the embolized branch through collaterals (most of them were intrahepatic collaterals) on control angiography, hepatic infarction occurred in the area supplied by the embolized artery. Infarction did not occur entirely in the involved segments in all patients. Although collaterals were not seen on the post-embolization film, we think the development of collaterals following embolization can modify the extent of infarction. In our patients, various transaminase levels were elevated, and there was rough correlation between the extent of the infarcted areas and these levels. When the distal branch filled normally through collaterals, infarction did not occur. It is not clear why arterial blood flow in the peripheral part of the embolized artery can not be maintained through collaterals in some patients; it may depend on vascular anatomy. We also speculate that we may have deposited the distal microcoils more distally in the corresponding branch. More distally placed microcoils might prevent blood flow through the collaterals. If this is so, we should a attempt to place microcoils as close as possible to the bleeding point to preserve the lumen of the peripheral portion of the embolized artery and to ensure that blood flow can be maintained through (intrahepatic) collaterals. In the patient with AP shunts, hepatic infarction also occurred in the area where intrahepatic collaterals developed after embolization (Fig. 2e). Both arterial and portal perfusion through the AP shunt was decreased after embolization in this area compared with that on pre-embolization films. Hepatic infarction tends to occur in this condition [16–19]. However, infarction did not occur in the area where the distal part of the embolized artery was well seen through filling
48
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
by intrahepatic collaterals. The portal branches in this area were also opacified on post-embolization films. We think the arterial flow through intrahepatic collaterals in this segment was enough to perfuse both the portal and arterial systems. The patient with severe right portal vein stenosis died of hepatic failure after TAE. There was potential liver dysfunction caused by decreased portal flow in the right lobe of this liver. Infarction occurred entirely in the normal left lateral segment. We speculate that the remaining right lobe could not compensate for the lack of function of the left lateral seg-
ment and could not tolerate the acute damage caused by infarction. According to previous reports [16,17], catastrophic necrosis and death have occurred after hepatic embolization in the treatment of hereditary hemorrhagic telangiectasis in patients with AP and PV (portal vein-to-hepatic vein) shunts. Those authors felt that decrease in effective portal flow into the liver parenchyma is the most important predisposing factor for hepatic failure after hepatic embolization. In other reports [4,5,18,19], hepatic embolization resulted in fatal outcomes, especially in patients who underwent pancreaticoduodenectomy. One report stated that verification of
Fig. 3. Images obtained in a 68-year-old man (no. 8) show development of small liver infarctions on the liver surface after TAE. The patient had undergone percutaneous liver biopsy for chronic hepatitis. (a) Celiac arteriogram shows a direct hepatic artery-to-biliary duct fistula (arrow) in the right posterior inferior segment. (b) Image obtained after diagnostic angiography shows opacification of the biliary system. (c) Common hepatic arteriogram obtained after embolization shows absence of filling of the distal part of the embolized branch and the fistula. (d) Plain CT obtained 10 days after TAE shows an ill-defined low density area (arrow) on the liver surface.
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
49
Fig. 3. (Continued ).
patent portal venous flow is important before embolization in these patients [5]. One patient with diminished portal flow in the embolized area did not have infarction. Peripheral flow in the embolized artery was seen through intrahepatic collaterals in this patient. Interventional radiologists should pay special attention when doing hepatic embolization in patients with decreased portal flow. There are some limitations to our study. First, it is a retrospective study. Second, we performed selective embolization with microcoils for hemobilia in a limited number of patients. It is not yet clear from post-embolization films whether hepatic infarction occurs in all patients without collaterals. We evaluated the presence of collaterals on angiograms obtained immediately after embolization. Collaterals that adequately perfuse liver parenchyma of the
embolized area may develop gradually after embolization. Third, associated portal vein injury such as only a small arterioportal shunt from previous interventional procedures might be presented. It could not be detected on the angiogram. Associated tiny portal vein injury might be a reason for necrosis of liver post-embolization. However, we think absence of collaterals on post-embolization films can be a predisposing factor for hepatic infarction after selective embolization with microcoils. Forth, we can not predict subsequent infarction before embolization. We do not think it necessary to change the treatment algorithm; however, we believe that microcoils should be placed as close as possible to the bleeding point to preserve the lumen of the peripheral portion of the embolized artery and to avoid unnecessary occlusion of a normal branch.
50
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
In conclusion, our results indicate that liver infarction following selective hepatic arterial embolization with microcoils may occur in patients in whom no opacification of the distal part of the embolized branch can be visualized through collaterals on post-embolization films.
References [1] Baker KS, Tisnado J, Cho SR, et al. Splanchnic artery aneurysms and pseudoaneurysms: transcatheter embolization. Radiology 1987;163:135–9. [2] Okazaki M, Higashihara H, Koganemaru F, et al. Emergent embolization for control of massive hemorrhage from a splanchnic artery with a new coaxial catheter system. Acta Radiol 1992;33:57–62. [3] Meyerovits MF, Levin DC, Boxt LM. Superselective catheterization of small-caliber arteries with new highly visibility guide wire. AJR 1985;144:785–7. [4] Okuno A, Miyazaki M, Ito H, et al. Nonsurgical management of ruptured pseudoaneurysm in patients with hepatobiliary pancreatic diseases. Am J Gastroenterol 2001;96:1068–71. [5] Reber PU, Baer HU, Patel AG, et al. Superselective microcoil embolization: treatment of choice in high-risk patients with extrahepatic pseudoaneurysms of the hepatic arteries. J Am Coll Surg 1998;186:325–30. [6] Delgadillo X, Berney T, de Perrot M, et al. Successful treatment of a pseudoaneurysm of the cystic artery with microcoil embolization. JVIR 1999;10:789–92. [7] Nicholson T, Travis S, Ettles D, et al. Hepatic artery angiography and embolization for hemobilia following laparoscopic cholecystectomy. CVIR 1999;22:20–4. [8] Kim W, Clark TWI, Baum RA, et al. Risk factors for liver abscess formation after hepatic chemoembolization. JVIR 2001;12:965–8.
[9] Cohen SE, Safadi R, Verstromdig A, et al. Liver-spleen infarcts following transcatheter chemoembolization: a case report and review of the literature on adverse effect. Dig Dis Sci 1997;42:938– 43. [10] Mitchell SE, Shuman LS, Kaufman SL, et al. Biliary catheter drainage complicated by hemobilia: treatment by balloon embolotherapy. Radiology 1985;157:645–52. [11] Ahmed A, Samuels SL, Keeffe EB, et al. Delayed fatal hemorrhage from pseudoaneurysm of the hepatic artery after percutaneous liver biopsy. Am J Gastroenterol 2001;96:233–7. [12] Hagiwara A, Yukioka T, Shimazaki S, et al. Delayed hemorrhage following transcatheter arterial embolization for blunt hepatic injury. CVIR 1993;16:380–3. [13] Triller J, Krebs T, Ackerman D. Superselective embolization of traumatic renal pseudoaneurysm with a tracker-18 catheter and microcoils. Eur Radiol 1993;3:261–3. [14] Inoguchi H, Mii S, Sakata H, et al. Intrahepatic pseudoaneurysm after surgical hemostasis for a delayed hemorrhage due to blunt liver injury: report of a case. Surg Today 2001;31:367–70. [15] Okazaki M, Higashihara H, Koganemaru F, et al. A coaxial catheter and steerable guidewire used to embolize branches of the splanchnic arteries. AJR 1990;155:405–8. [16] Garcia-Tsao G, Korzenik JR, Young L, et al. Liver disease in patients with hereditary hemorrhagic telangiectasia. N Eng J Med 2000;343:931–6. [17] Miller FJ, Whiting JH, Korzenik JR, et al. Caution with use of hepatic embolization in the treatment of hereditary hemorrhagic telangiectasia. Radiology 1999;213:928. [18] Chung JW, Park JH, Han JK, et al. Hepatic tumors: predisposing factors for complication of transcatheter oily chemoembolization. Radiology 1996;198:33–40. [19] Sato N, Yamaguchi K, Shimizu S, et al. Coil embolization of bleeding visceral pseudoaneurysms following pancreatectomy: the importance of early angiography. Arch Surg 1998;133:1099–102.
Hepatic infarction following selective hepatic artery embolization with microcoils for iatrogenic biliary hemorrhage Manabu Hashimoto∗ , Yoko Akabane, Jyouichi Heianna, Etuko Tate, Koichi Ishiyama, Toshiaki Nishii, Jiro Watarai Department of Radiology, Akita University School of Medicine, Hondo 1-1-1, Akita City, Akita 010-8543, Japan Received 5 February 2004; received in revised form 12 April 2004; accepted 7 May 2004 Available online 2 July 2004
Abstract Purpose: The aim of this study was to explore the relation of collateral filling to ischemic or infarcted liver following selective embolization of hepatic artery with microcoils in patients with iatrogenic hemobilia. Methods: We performed retrospective analysis of clinical outcomes and post-embolization angiograms in eight patients (mean age of 66 years) studied over the last 7 years. Hemobilia occurred after percutaneous biliary drainage (n = 5) and percutaneous hepatic biopsy (n = 3). Causes of bleeding were pseudoaneurysm (n = 6), arterial laceration (n = 1), and direct hepatic artery-to-biliary duct fistula (n = 1). We placed microcoils in the subsegmental (n = 4) or segmental branch (n = 2), or both branches (n = 2), distal and proximal to the bleeding point. Results: We obtained complete hemostasis in all patients (100%). Four patients had no hepatic infarction after embolization. Normal filling of the distal part of the embolized branch through collaterals was seen on post-embolization films. Four patients with no collateral filling experienced liver infarction in the area corresponding to embolized branch. One patient with severe portal stenosis died of hepatic failure. Conclusion: Hepatic infarction is related to lack of immediate collateral flow. © 2004 Elsevier B.V. All rights reserved. Keywords: Liver; Hemobilia; Embolization; Complication
1. Introduction
2. Materials and methods
Transcatheter arterial embolization (TAE) has been established as an alternative to surgical intervention for obtaining hemostasis for gastroduodenal hemorrhage or hemobilia that can occur after various surgical and interventional treatments [1–7]. Although TAE is considered to be relatively safe, several complications may occur, such as liver abscess, liver ischemia, or liver failure [4,5,7–9]. To enhance the safety of TAE, it is important to understand complications of the procedure. In this study, we analyze the incidence and distribution of ischemic or infarcted liver following selective hepatic artery embolization with microcoils in patients with iatrogenic hemobilia. Then, we investigate the relation of collateral filling to ischemic or infarcted liver.
We performed retrospective analysis of clinical outcomes and post-embolization angiographic findings in eight consecutive patients undergoing selective hepatic embolization for iatrogenic hemobilia from 1996 to 2002 (Table 1). The patients comprised two women and six men with a mean age of 66 years. Seven patients had malignancies: hilar cholangiocarcinoma (n = 5), intrahepatic cholangiocarcinoma (n = 1), metastatic liver tumor (n = 1). The remaining patient had chronic hepatitis. None of the patients had a history of hepatic viral infection (hepatitis B or C) or liver cirrhosis. Hemobilia occurred after percutaneous transhepatic biliary drainage (PTBD) in five patients and percutaneous liver biopsy in three. The clinical manifestations were bleeding from a PTBD tube in two patients and hematemesis in six. Emergency gastrointestinal endoscopy established a diagnosis of hemobilia in the six patients with hematemesis. All patients were referred to our department within 2 weeks of PTBD or biopsy because hemostasis could not be obtained through conservative treatment.
∗
Corresponding author. Fax: +81 188 36 2623. E-mail address: [email protected] (M. Hashimoto).
1386-6346/$ – see front matter © 2004 Elsevier B.V. All rights reserved. doi:10.1016/j.hepres.2004.05.002
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
43
Table 1 Patient background No.
Age (year)
Sex
Disease
Underlying interventional procedure
Bleeding site
Angiographic findings
Portal flow abnomality
1 2
67 65
M M
Hilar cholangiocarcinoma Hilar cholangiocarcinoma
Drainage Drainage
Laceration Pseudoaneurysm
– –
3
63
F
Hilar cholangiocarcinoma
Drainage
Pseudoaneurysm
–
4 5
67 61
M M
Hilar cholangiocarcinoma Intrahepatic cholangiocarcinoma
Drainage Biopsy
A3 Right posterior segmental branch Right anterior segmental branch A8 A8
Pseudoaneurysm Pseudoaneurysm
6 7 8
72 69 68
F M M
Hilar cholangiocarcinoma Metastatic liver tumor Chronic hepatitis
Drainage Biopsy Biopsy
A3 A8 A6
Pseudoaneurysm Pseudoaneurysm A-B fistula
Anterior branch occlusion AP shunts (R) anterior branch occlusion Right PV stenosis – –
A3: left lateral anterior branch; A6: right posterior inferior branch; A8: right anterior superior branch; R: right lobe of the liver; Anterior branch: anterior branch of right portal vein; PV: portal vein; AP: hepatic artery-to-portal vein; A-B: hepatic artery-to-bile duct.
Pre-embolization angiographic findings included hepatic artery branch pseudoaneurysm in six patients, arterial transgression by the drainage catheter in one, and direct hepatic artery-to-biliary fistula in one. Three patients had portal abnormalities on angiograms. One patient with hilar cholangiocarcinoma had severe stenosis in the right portal vein. One with intrahepatic cholangiocarcinoma had occlusion of the anterior branch of the right portal vein. Hepatic artery-to-portal vein (AP) shunts throughout the anterior segment of the right lobe were also seen in this patient. The remaining one patient had occlusion of the right anterior portal vein. These portal abnormalities were caused by hepatic hilar tumors. We performed selective hepatic arterial embolization as follows. Diagnostic hepatic angiography was performed via
the femoral artery and included selective celiac, superior mesenteric, and common hepatic arteriography. After obtaining a diagnosis of pseudoaneurysm, arterial laceration, or direct hepatic artery-to-biliary fistula, a microcatheter (Microferret, William Cook Europe, Denmark) was inserted into a 4.2-F diagnostic angiography catheter (Hanako, Tokyo, Japan), which remained in the common, proper, or right or left hepatic artery. We advanced a fine guidewire (Transend EX, Boston Scientific Target, CA) through the microcatheter and into the corresponding arterial branch and distal to the pseudoaneurysm, laceration, or fistula. The microcatheter was then threaded over the fine guidewire. We placed adequately sized platinum microcoils (Vortex, Boston Scientific Target) in the corresponding branch distal and proximal to the bleeding point. Embolization was performed un-
Table 2 Patient outcomes No. Embolized artery Opacification of distal part through intrahepatic collaterals
Liver infarction
1 2
Extrahepatic (+) +
– –
37/26 43/25
34/24 42/23
None None
Uneventful Uneventful
+
–
38/34
35/29
None
Uneventful
110/105 61/56
98/96 203/159
None None
Uneventful Died of malignancy 22 months after embolization Died of hepatic failure 14 days after embolization Uneventful Uneventful
3
4 5
6
7 8
Area of infarction on CT
AST/ALT values (U/L) Pre
Re-bleeding Outcomes
Post
A3 Right posterior segmental branch Right anterior segmental branch A8 A8 + right anterior segmental branch A3 + left hepatic artery
+ A8 (−), A5 (+)
– S8
Patchy distribution
A2 (−), A3 (−), A4 (+)
Lateral segment
Throughout lateral segment
49/21
1251/663
None
A8 A6
– –
S8 S6
Wedge-shaped Ill-defined
10/11 93/75
598/692 143/109
None None
A2: left lateral posterior branch; A3: left lateral anterior branch; A4: left medial branch; A5: right anterior inferior branch; A6: right posterior inferior branch; A8: right anterior superior branch; S6: right posterior inferior segment; S8: right anterior superior segment. Lateral segment: left lateral segment. Pre: pre-embolization; Post: post-embolization.
44
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
Fig. 1. Images obtained in a 67-year-old man (no. 1). (a) Left hepatic arteriogram shows laceration of the left lateral anterior branch of the hepatic artery (arrowhead) secondary to percutaneous transhepatic catheter placement in the anterior branch of the left lateral bile duct (B3). Another catheter is inserted from the dorsal branch of the left lateral bile duct. Massive bleeding was seen from the drainage route after removal of the catheter in the B3. (b) Angiogram after TAE shows normal filling of the distal part of the embolized branch through gastric artery (arrowheads). Liver infarction did not occur after TAE in this patient.
til the flow ceased completely; angiograms through the microcatheter showed elimination of the focal arterial lesions from the circulation and stagnant contrast medium in the proximal part of the corresponding vessels. The placement sites of the coils varied from patient to patient according to the vascular anatomy and the sites of the bleeding. After embolization, we performed immediate hepatic angiography through the diagnostic angiography catheter, which had remained in the common, proper, or right or left hepatic artery as a control. Images were obtained by digital subtraction angiography.
Clinical examination, complete blood count, and liver function tests were performed in all patients within 1 week of TAE. The patients with suspected hepatic complications underwent emergent CT (plain or plain and contrast-enhanced CT). We also performed plain and contrast-enhanced CT in 7 patients 1 month after embolization for follow-up. The diagnosis of liver ischemia or infarction was based on a temporary increased in transaminase levels and development of low-attenuation areas corresponding to the embolized arterial branches on emergent CT. We evaluated the presence of low-attenuation areas in pre-contrast or
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
pre-contrast and post-contrast during equilibrium phase CT images.
3. Results Selective embolization with microcoils was performed successfully in all eight patients at the level of the subseg-
45
mental (n = 4) or segmental branches (n = 2), or both (n = 2) (Table 2). Clinical course after embolization was uneventful in four of the patients. Blood examinations and CT images (obtained 1 month after embolization) showed no remarkable changes related to embolization. The post-embolization films showed normal filling of the peripheral branches of the embolized artery through collaterals (Fig. 1).
Fig. 2. Images obtained in a 61-year-old man (no. 5) show development of liver infarction after TAE. The patient had undergone liver biopsy for liver tumor around the hepatic hilum. (a) Common hepatic arteriogram shows a pseudoaneurysm (arrow) of the right anterior superior branch of the right hepatic artery (A8) (arrowhead). (b) Selective arteriogram of the anterior branch of the right hepatic artery shows the pseudoaneurysm and opacification of the right anterior superior (P8, small arrowheads) and inferior (P5, large arrowhead) branches of the portal vein. Anterior inferior branch of the right hepatic artery (A5) is also seen (arrows). After advancement of a microcatheter distally, we placed microcoils around the pseudoaneurysm in this branch. (c) Common hepatic arteriogram obtained after embolization shows absence of opacification of the aneurysm and a branch of the A8 (arrowhead in (a)). Other branches of the A8 are seen (small arrow) through intrahepatic collateral filling but peripheral parts are poorly visualized. Distal parts of the A5 (arrows) are adequately filled through intrahepatic collaterals. Enhanced CT scans obtained 4 days after TAE show liver infarction in the two portions of the right anterior superior segment corresponding to the areas of absence of filling of the branch of A8 (d) and the poorly visualized branch (e), respectively.
46
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
Fig. 2. (Continued ).
The other four patients had liver infarction in the areas corresponding to the embolized arterial branches. They experienced no abdominal pain or fever after the embolization procedure, but we noted a temporary increased in transaminase levels (maximum value of AST: 203, 1251, 598 and 143 U/L; ALT: 159, 663, 692 and 109 U/L, respectively, in these four patients) indicating a transient liver ischemia. Emergent CT showed development of low-attenuation areas corresponding to the embolized arterial branches. On CT, the infarction was shown to occur entirely in the involved segment in one patient. One patient had a wedge-shaped infarction. Hepatic infarction appeared as a patchy distribution in one patient (Fig. 2) and as an ill-defined area of low attenuation with liver surface predominance in another (Fig. 3). The values of the measured transaminases gradually recov-
ered to normal within 2 weeks. On control angiography, opacification of the distal portion of the embolized branches was not obtained through collaterals in the areas of infarction (Figs. 2 and 3). Infarctions did not occur in the areas in which peripheral branches were filled through intrahepatic collaterals in two of four patients. Clinical follow-up of from 2 weeks to 60 months (mean, 21 months) was obtained on all eight patients. No recurrence of bleeding was observed; however, one patient (no. 6) with hepatic infarction died of hepatic failure 14 days after TAE. This patient had severe stenosis of the right portal vein. Follow-up CT of the other three patients with hepatic infarction revealed gradual resolution of the infarcted area. One patient (no. 4) died of associated malignancy after 22 months.
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
47
Fig. 2. (Continued ).
4. Discussion Various focal arterial lesions leading to hepatic arterial-biliary fistula, such as pseudoaneurysm, arterial laceration, and direct hepatic arterial-biliary fistula, can occur after surgical and interventional procedures of the liver. The purpose of embolization therapy for these focal lesions is to eliminate focal arterial lesions from the circulation and to obtain thrombosis of this segmental pedicle. Most previously published papers have described embolization for pseudoaneurysms [1,2,4,7,10–13]. Microcoils should be used initially for TAE and should be placed both proximally and distally to the bleeding point [5,7,14] because TAE done with particle materials (such as Gelfoam) often increase intravascular pressure when the particles are released from the catheter and induce rupture of the pseudoaneurysm [5]. With the advent of and subsequent advancements in microcatheter systems, it has become relatively easy to place microcoils proximally and distally to the bleeding site, even in small-caliber branches [2,15] such as those in our patients. The selective embolization method with microcatheter and microcoils has a high success rate for achieving hemostasis [2,5]; complete hemostasis can be obtained in most patients when the microcatheter is advanced for enough into the target vessel that the microcoils can be placed distally and proximally to the bleeding point. Furthermore, this method can avoid the needless occlusion of a normal branch. The success rate was 100% in our patient series. Previous reports have stated that some patients had complications associated with embolization of the hepatic arteries after hemostasis, with complication rates ranging from 11% [7] to 60% [4,5]. None of these reports described precisely how hepatic infarction occurred after TAE. Although there were a limited number of patients in our study, some
observations can be made. We found a strong association between the occurrence of infarction and the findings on the post-embolization film. When there was no opacification of the distal part of the embolized branch through collaterals (most of them were intrahepatic collaterals) on control angiography, hepatic infarction occurred in the area supplied by the embolized artery. Infarction did not occur entirely in the involved segments in all patients. Although collaterals were not seen on the post-embolization film, we think the development of collaterals following embolization can modify the extent of infarction. In our patients, various transaminase levels were elevated, and there was rough correlation between the extent of the infarcted areas and these levels. When the distal branch filled normally through collaterals, infarction did not occur. It is not clear why arterial blood flow in the peripheral part of the embolized artery can not be maintained through collaterals in some patients; it may depend on vascular anatomy. We also speculate that we may have deposited the distal microcoils more distally in the corresponding branch. More distally placed microcoils might prevent blood flow through the collaterals. If this is so, we should a attempt to place microcoils as close as possible to the bleeding point to preserve the lumen of the peripheral portion of the embolized artery and to ensure that blood flow can be maintained through (intrahepatic) collaterals. In the patient with AP shunts, hepatic infarction also occurred in the area where intrahepatic collaterals developed after embolization (Fig. 2e). Both arterial and portal perfusion through the AP shunt was decreased after embolization in this area compared with that on pre-embolization films. Hepatic infarction tends to occur in this condition [16–19]. However, infarction did not occur in the area where the distal part of the embolized artery was well seen through filling
48
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
by intrahepatic collaterals. The portal branches in this area were also opacified on post-embolization films. We think the arterial flow through intrahepatic collaterals in this segment was enough to perfuse both the portal and arterial systems. The patient with severe right portal vein stenosis died of hepatic failure after TAE. There was potential liver dysfunction caused by decreased portal flow in the right lobe of this liver. Infarction occurred entirely in the normal left lateral segment. We speculate that the remaining right lobe could not compensate for the lack of function of the left lateral seg-
ment and could not tolerate the acute damage caused by infarction. According to previous reports [16,17], catastrophic necrosis and death have occurred after hepatic embolization in the treatment of hereditary hemorrhagic telangiectasis in patients with AP and PV (portal vein-to-hepatic vein) shunts. Those authors felt that decrease in effective portal flow into the liver parenchyma is the most important predisposing factor for hepatic failure after hepatic embolization. In other reports [4,5,18,19], hepatic embolization resulted in fatal outcomes, especially in patients who underwent pancreaticoduodenectomy. One report stated that verification of
Fig. 3. Images obtained in a 68-year-old man (no. 8) show development of small liver infarctions on the liver surface after TAE. The patient had undergone percutaneous liver biopsy for chronic hepatitis. (a) Celiac arteriogram shows a direct hepatic artery-to-biliary duct fistula (arrow) in the right posterior inferior segment. (b) Image obtained after diagnostic angiography shows opacification of the biliary system. (c) Common hepatic arteriogram obtained after embolization shows absence of filling of the distal part of the embolized branch and the fistula. (d) Plain CT obtained 10 days after TAE shows an ill-defined low density area (arrow) on the liver surface.
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
49
Fig. 3. (Continued ).
patent portal venous flow is important before embolization in these patients [5]. One patient with diminished portal flow in the embolized area did not have infarction. Peripheral flow in the embolized artery was seen through intrahepatic collaterals in this patient. Interventional radiologists should pay special attention when doing hepatic embolization in patients with decreased portal flow. There are some limitations to our study. First, it is a retrospective study. Second, we performed selective embolization with microcoils for hemobilia in a limited number of patients. It is not yet clear from post-embolization films whether hepatic infarction occurs in all patients without collaterals. We evaluated the presence of collaterals on angiograms obtained immediately after embolization. Collaterals that adequately perfuse liver parenchyma of the
embolized area may develop gradually after embolization. Third, associated portal vein injury such as only a small arterioportal shunt from previous interventional procedures might be presented. It could not be detected on the angiogram. Associated tiny portal vein injury might be a reason for necrosis of liver post-embolization. However, we think absence of collaterals on post-embolization films can be a predisposing factor for hepatic infarction after selective embolization with microcoils. Forth, we can not predict subsequent infarction before embolization. We do not think it necessary to change the treatment algorithm; however, we believe that microcoils should be placed as close as possible to the bleeding point to preserve the lumen of the peripheral portion of the embolized artery and to avoid unnecessary occlusion of a normal branch.
50
M. Hashimoto et al. / Hepatology Research 30 (2004) 42–50
In conclusion, our results indicate that liver infarction following selective hepatic arterial embolization with microcoils may occur in patients in whom no opacification of the distal part of the embolized branch can be visualized through collaterals on post-embolization films.
References [1] Baker KS, Tisnado J, Cho SR, et al. Splanchnic artery aneurysms and pseudoaneurysms: transcatheter embolization. Radiology 1987;163:135–9. [2] Okazaki M, Higashihara H, Koganemaru F, et al. Emergent embolization for control of massive hemorrhage from a splanchnic artery with a new coaxial catheter system. Acta Radiol 1992;33:57–62. [3] Meyerovits MF, Levin DC, Boxt LM. Superselective catheterization of small-caliber arteries with new highly visibility guide wire. AJR 1985;144:785–7. [4] Okuno A, Miyazaki M, Ito H, et al. Nonsurgical management of ruptured pseudoaneurysm in patients with hepatobiliary pancreatic diseases. Am J Gastroenterol 2001;96:1068–71. [5] Reber PU, Baer HU, Patel AG, et al. Superselective microcoil embolization: treatment of choice in high-risk patients with extrahepatic pseudoaneurysms of the hepatic arteries. J Am Coll Surg 1998;186:325–30. [6] Delgadillo X, Berney T, de Perrot M, et al. Successful treatment of a pseudoaneurysm of the cystic artery with microcoil embolization. JVIR 1999;10:789–92. [7] Nicholson T, Travis S, Ettles D, et al. Hepatic artery angiography and embolization for hemobilia following laparoscopic cholecystectomy. CVIR 1999;22:20–4. [8] Kim W, Clark TWI, Baum RA, et al. Risk factors for liver abscess formation after hepatic chemoembolization. JVIR 2001;12:965–8.
[9] Cohen SE, Safadi R, Verstromdig A, et al. Liver-spleen infarcts following transcatheter chemoembolization: a case report and review of the literature on adverse effect. Dig Dis Sci 1997;42:938– 43. [10] Mitchell SE, Shuman LS, Kaufman SL, et al. Biliary catheter drainage complicated by hemobilia: treatment by balloon embolotherapy. Radiology 1985;157:645–52. [11] Ahmed A, Samuels SL, Keeffe EB, et al. Delayed fatal hemorrhage from pseudoaneurysm of the hepatic artery after percutaneous liver biopsy. Am J Gastroenterol 2001;96:233–7. [12] Hagiwara A, Yukioka T, Shimazaki S, et al. Delayed hemorrhage following transcatheter arterial embolization for blunt hepatic injury. CVIR 1993;16:380–3. [13] Triller J, Krebs T, Ackerman D. Superselective embolization of traumatic renal pseudoaneurysm with a tracker-18 catheter and microcoils. Eur Radiol 1993;3:261–3. [14] Inoguchi H, Mii S, Sakata H, et al. Intrahepatic pseudoaneurysm after surgical hemostasis for a delayed hemorrhage due to blunt liver injury: report of a case. Surg Today 2001;31:367–70. [15] Okazaki M, Higashihara H, Koganemaru F, et al. A coaxial catheter and steerable guidewire used to embolize branches of the splanchnic arteries. AJR 1990;155:405–8. [16] Garcia-Tsao G, Korzenik JR, Young L, et al. Liver disease in patients with hereditary hemorrhagic telangiectasia. N Eng J Med 2000;343:931–6. [17] Miller FJ, Whiting JH, Korzenik JR, et al. Caution with use of hepatic embolization in the treatment of hereditary hemorrhagic telangiectasia. Radiology 1999;213:928. [18] Chung JW, Park JH, Han JK, et al. Hepatic tumors: predisposing factors for complication of transcatheter oily chemoembolization. Radiology 1996;198:33–40. [19] Sato N, Yamaguchi K, Shimizu S, et al. Coil embolization of bleeding visceral pseudoaneurysms following pancreatectomy: the importance of early angiography. Arch Surg 1998;133:1099–102.