AJR:181, July 2003

71

Symptomatic Intrahepatic Portosystemic Venous Shunt:

Angiographic Findings and Transcatheter Embolization with an Alternative Approach

OBJECTIVE.

Intrahepatic portosystemic venous shunt is relatively rare and not well recog-nized. Awareness of intrahepatic communications is important because they can cause encepha-lopathy, and most of these shunts can be completely cured by transcatheter embolization. In thisstudy, we describe the angiographic findings and transcatheter embolization techniques usingseveral approaches for the treatment of intrahepatic portosystemic venous shunt.

MATERIALS AND METHODS.

Between 1989 and 2001, we treated 10 patients withsymptomatic intrahepatic portosystemic venous shunt by performing transcatheter embolizationwith Gianturco coils, fibered platinum coils,

detachable balloons, and detachable microcoils us-ing one of three approaches to access the portal venous system: transileocolic obliteration (

n

= 2),percutaneous transhepatic obliteration (

n

= 4), or retrograde transcaval obliteration (

n

= 4).

RESULTS.

In all patients, complete obliteration or nearly complete obliteration was con-firmed angiographically, and symptoms related to portal–systemic encephalopathy improvedafter treatment. Complications were observed in three patients: adhesive ileus in a patienttreated by transileocolic obliteration and thrombosis of intrahepatic portal branches in two pa-tients treated by percutaneous transhepatic obliteration.

CONCLUSION.

On angiography, two types of intrahepatic portosystemic venous shuntwere seen: intrahepatic portal venous–hepatic venous communication and intrahepatic portalvenous–perihepatic venous communication. Transcatheter embolization is effective for treat-ment of intrahepatic portosystemic venous shunt. Retrograde transcaval obliteration is theleast invasive technique and is recommended as the first choice for treatment of portosystemicvenous shunt except in patients with multiple shunts.

ntrahepatic portosystemic venousshunt is a rare condition defined ascommunication between the intrahe-

patic portal vein and systemic veins, includingthe hepatic vein and perihepatic vein, via ananomalous intrahepatic venous channel. After theintroduction of CT, MR imaging, and sonogra-phy, intrahepatic portosystemic venous shunt hasbeen encountered more frequently [1–3]. Clinicalmanifestations of intrahepatic portosystemicvenous shunt depend on the shunt flow; a high-flow shunt might cause hepatic encephalopathyand hypoglycemia.

Conservative therapy (restriction of protein,ingestion of lactulose, oral administration ofnonabsorbable antibiotics) [4], surgery (portalvein ligation or hepatic lobectomy), and trans-catheter embolization have been used for thetreatment of intrahepatic portosystemic venousshunt. Transcatheter embolization is a well-es-tablished, useful, and less invasive treatment

for several vascular diseases. In this report, wedescribe methods of embolization with threeapproaches in respective types of intrahepaticportosystemic venous shunt, and we refer tothe pathogenesis of the condition based on an-giographic findings.

Materials and Methods

Patients

Between 1989 and 2001, 10 patients (two menand eight women; age range, 33–77 years; meanage, 58.3 years) with symptomatic intrahepatic por-tosystemic venous shunt were treated by transcathe-ter embolization at multiple related facilities. Thepatients had experienced various symptoms, includ-ing disturbance of consciousness (

n

= 6), tremors (

n

=4), disorientation (

n

= 2), and somnolence (

n

= l),that were thought to be caused by portal–systemicencephalopathy. The diagnoses of intrahepatic por-tosystemic venous shunt were based on CT andsonography performed before embolization in all

Shuichi Tanoue

1

Hiro Kiyosue

1

Eiji Komatsu

2

Yuzo Hori

3

Tohru Maeda

2

Hiromu Mori

1

Received June 7, 2002; accepted after revision December 6, 2002.

1

Department of Radiology, Oita Medical University, 1-1, Idaigaoka, Hasama-machi, Oita-gun, Oita, 879-5593, Japan. Address correspondence to H. Mori.

2

Department of Radiology, Oita Prefectural Hospital, 476, Bunyo, Oita-shi, Oita, 870-8511, Japan.

3

Department of Radiology, Nagatomi Neurosurgical Hospital, Omichi-Machi, Oita-shi, Oita, 870-0822, Japan.

AJR

2003;181:71–78

0361–803X/03/1811–71

© American Roentgen Ray Society

I

Dow

nloa

ded

from

ww

w.a

jron

line.

org

by 1

86.1

51.6

3.12

on

07/0

9/14

fro

m I

P ad

dres

s 18

6.15

1.63

.12.

Cop

yrig

ht A

RR

S. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved

72

AJR:181, July 2003

Tanoue et al.

patients, and conservative therapies were tried ini-tially. In two patients, intrahepatic portosystemicvenous shunt was associated with liver cirrhosis.

Laboratory test results and transcatheter portalvenous pressure measurements were as follows: se-rum ammonium ratio, 1.50–5.00 (mean, 2.65); Fis-cher’s ratio, 0.94–1.65 (mean, 1.31); and portalvenous pressure, 5.0–18.0 cm H

2

O (mean, 11.0 cmH

2

O). Serum ammonium ratios are defined as ratiosof serum ammonium levels relative to normal valuesbecause the units and normal values differed amongthe institutions where patients were treated.

Transcatheter Embolization

Transcatheter embolization was performed viaone of the following three access routes: transileo-colic obliteration, percutaneous transhepatic oblit-eration, and retrograde transcaval obliteration.These access techniques are described and are il-lustrated in Figure 1.

Transileocolic obliteration

.—After exposure ofthe distal ileum under a small abdominal incision,a catheter was advanced into the portal venoussystem via the ileocolic vein.

Percutaneous transhepatic obliteration.—

Afterpercutaneous puncture of the intrahepatic portalbranch under sonographic guidance, a catheterwas advanced into the portal venous system.

Retrograde transcaval obliteration

.—Two cathe-ters were advanced in a retrograde manner into theportal venous system via bilateral transfemoral

venous access. One flexible and straight catheter(Tracker 38, BSJ, Tokyo, Japan) was advanced intothe main portal vein through the shunt to obtain aportogram and measure portal venous pressure be-fore and during embolization. Another catheter wasused for placement of embolic materials.

Retrograde transcaval obliteration and percuta-neous transhepatic obliteration were performedwith the patient under local anesthesia, and trans-ileocolic obliteration were performed with the pa-tient under epidural anesthesia.

These approaches were selected as follows: Ifthe patient had one large shunt or a few shunts, ret-rograde transcaval obliteration was selected as theaccess route for the initial treatment. If we failedwith this approach, another approach was at-tempted. If the patient had multiple shunts in aunilateral lobe, percutaneous transhepatic oblitera-tion was selected. Transileocolic obliteration wasselected only if multiple shunts were present in thebilateral lobe or if treatments with other ap-proaches had failed.

Embolization was performed by transileocolicobliteration in two patients, percutaneous transhe-patic obliteration in four patients, and retrogradetranscaval obliteration in four patients. Several em-bolic materials including Gianturco coils (WilliamCook Europe, Bjaeverskov, Denmark), detachableballoons (BSJ), and microcoils were used. Four pa-tients were treated using Gianturco coils only, onewas treated with Gianturco coils and detachable bal-loons, and five were treated with Gianturco coils and

microcoils (fibered platinum coils, Detach Coil Sys-tem [William Cook Europe], or both). Three to 30coils that ranged from 5 to 15 mm in diameter wereused in each patient. When a 5-French cathetercould not be advanced into the shunt vessels becauseof their extreme tortuosity, we used microcathetersand microcoils. As a follow-up examination within3–12 months after the procedure, CT, sonography, orboth were performed. The clinical follow-up periodranged from 24 to 156 months.

Evaluation

All data including clinical data, radiologic find-ings, and clinical outcomes were collected retro-spectively. The angiographic findings evaluated bythree radiologists were the type of drainage vein,multiplicity, and associated intrahepatic venous ab-normalities. The intrahepatic portosystemic venousshunt was classified by the type of drainage vein andmultiplicity. Possible approaches for several types ofshunt, their technical success rates, and complica-tions were investigated. We evaluated the effects ofthese procedures on clinical symptoms, laboratorytest results, and portal venous pressures. Clinicalsymptoms, except for consciousness level, wereevaluated without any scale by degree of patient’scomplaint. Consciousness level was evaluated usingthe Glasgow Coma Scale. Laboratory tests includedserum ammonium ratios and Fischer’s ratios. Thechanges in these data were analyzed using Wil-coxon’s signed rank test.

A B C

Fig. 1.—Drawings illustrate three approaches to access intrahepatic portosystemic venous shunts.A, For transileocolic obliteration, catheter (open arrow) is advanced into portal venous system via ileocolic vein (solid arrow) through small abdominal incision.B, For percutaneous transhepatic obliteration, catheter (arrow) is advanced into portal venous system after percutaneous puncture of intrahepatic portal branch.C, For retrograde transcaval obliteration, two catheters are retrogradely advanced into portal venous system through shunt vessel (arrowhead) via bilateral transfemoralvenous access. One catheter (open arrow), which is advanced into main portal vein through shunt, is straight catheter used for portography and to measure portal venouspressure during procedure. Other catheter (solid arrow) is used to place embolic materials.

Dow

nloa

ded

from

ww

w.a

jron

line.

org

by 1

86.1

51.6

3.12

on

07/0

9/14

fro

m I

P ad

dres

s 18

6.15

1.63

.12.

Cop

yrig

ht A

RR

S. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved

Angiography and Embolization of Portosystemic Venous Shunts

AJR:181, July 2003

73

Results

Transvenous portography was performed inall patients, and eight of 10 patients underwenttransarterial portography before embolization.Transarterial or

transvenous portography showedmultiple shunts in five patients and single shuntsin the other five patients.

Intrahepatic portosystemic venous shuntswere divided into two types according to thedrainage vein. One was a communication be-tween the intrahepatic portal vein and the hepaticvein, whereas the other was a communicationbetween the intrahepatic portal vein and the in-

ferior vena cava via the perihepatic veins (adre-nal vein

or inferior phrenic veins). The formertype of shunt was identified in eight patients in-cluding two patients with multiple shunts.None of the cases of intrahepatic portalvenous–hepatic venous shunt were associatedwith liver cirrhosis. Portograms or hepaticvenograms showed the eight cases were asso-ciated with intrahepatic vein anomalies in-cluding five portal vein aneurysms (Fig. 2),one portal vein anastomosis (Fig. 3), and twohepatic vein anastomoses (Fig. 4). The lattertype of shunt, the portal venous–perihepatic

venous shunt, was observed in two patientswith liver cirrhosis.

Three of the five patients with multiple in-trahepatic portosystemic venous shunts under-went percutaneous transhepatic obliteration. Inthe remaining two patients, retrograde trans-caval obliteration was tried but failed initiallybecause of technical difficulty in approachingthe shunts. These two patients were subse-quently treated by transileocolic obliteration(Fig. 5). One of the five patients with a singleshunt underwent percutaneous transhepaticobliteration. Because the shunt was not com-pletely obliterated after the initial procedure,reembolization was subsequently performedand the shunt was completely occluded. Theremaining four patients were treated by retro-grade transcaval obliteration (Fig. 6). Reembo-lization was required in one of these patientsbecause of residual shunt flow. In nine of the10 patients, intrahepatic portosystemic venousshunts were shown to be completely obliter-ated on angiograms obtained after the oblitera-tion procedures. Nearly complete obliterationwas achieved in the remaining patient who hadmultiple shunts.

Procedure-related complications were ob-served in three patients: adhesive ileus wasseen in a patient treated by transileocolicobliteration a few days after the procedure,occlusion of the left portal vein due to coilmigration was found in a patient treated bypercutaneous transhepatic obliteration, andthrombosis of the left portal venous branchrelated to the puncture procedure was ob-

Fig. 2.—Transcaval retrograde por-togram shows intrahepatic porto-systemic venous shunt withaneurysmal dilatation (arrow) in 48-year-old man. Arrowhead indicatesa catheter advanced into main por-tal vein via shunt.

Fig. 3.—Transcaval retrograde hepatic venogram shows intrahepatic portosystemic venousshunt between left hepatic vein and left portal vein in 53-year-old woman. Note portal venousanastomosis of medial branches (arrowheads).

Fig. 4.—Transcaval retrograde hepatic venogram shows hepaticvenous anastomosis between right hepatic vein and accessory hepaticvein (arrow) in 64-year-old woman.

Dow

nloa

ded

from

ww

w.a

jron

line.

org

by 1

86.1

51.6

3.12

on

07/0

9/14

fro

m I

P ad

dres

s 18

6.15

1.63

.12.

Cop

yrig

ht A

RR

S. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved

74

AJR:181, July 2003

Tanoue et al.

served in a patient treated by percutaneoustranshepatic obliteration. The latter two com-plications were seen on follow-up CT afterthe procedure and did not cause any clinicalsymptoms. No other complications, includ-ing peritoneal complications, were encoun-tered after the treatments.

Symptoms related to portal–systemic en-cephalopathy completely disappeared in

eight patients and improved in two patientsafter treatment. Serum ammonium ratios sig-nificantly decreased and Fischer’s ratios in-creased after embolization (Figs. 7A and7B). The portal venous pressure tended to in-crease to levels above those recorded beforetreatment (Fig. 7C). In the two patients withliver cirrhosis, the pressures after treatmentwere above the reference value.

Discussion

Portal–systemic encephalopathy usuallyoccurs in patients with portal hypertensionand is mainly induced by liver cirrhosis, al-though it can sometimes occur in patientswho do not have liver cirrhosis. Recent im-aging studies have identified intrahepaticportosystemic venous shunt in the lattergroup of patients [1–22].

A B

Fig. 5.—Multiple intrahepatic portosystemic venous shunts in 62-year-old woman who did not have cirrhosis and who presented with memory disturbance and trembling.Blood examination revealed hyperammonemia and low Fischer’s ratio. Patient was treated by transileocolic obliteration.A, Transileocolic portogram revealed multiple intrahepatic portosystemic venous shunts in left lobe (arrowheads). Gianturco coils (William Cook Europe, Bjaeverskov, Den-mark) and fibered microcoils were placed into shunt vessels.B, Transileocolic portogram obtained after embolization shows complete obliteration of intrahepatic portosystemic venous shunts.

A B

Fig. 6.—Single intrahepatic portosystemic venous shunt in 72-year-old woman who did not have cirrhosis and who presented in coma. Blood examination revealed hyper-ammonemia and low Fischer’s ratio. Patient was treated by retrograde transcaval obliteration.A, Retrograde transcaval portography was performed with catheter advanced into portal vein via shunt vessel (arrowhead). Portogram shows intrahepatic portosystemicvenous shunt between right portal vein and accessory hepatic vein (open arrow) with portal vein aneurysm (solid arrow). Gianturco coil (William Cook Europe, Bjaever-skov, Denmark), detachable microcoils, and fibered platinum microcoils were positioned in shunt just before aneurysmal dilatation.B, Portogram obtained after procedure shows complete obliteration of intrahepatic portosystemic venous shunt.

Dow

nloa

ded

from

ww

w.a

jron

line.

org

by 1

86.1

51.6

3.12

on

07/0

9/14

fro

m I

P ad

dres

s 18

6.15

1.63

.12.

Cop

yrig

ht A

RR

S. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved

Angiography and Embolization of Portosystemic Venous Shunts

AJR:181, July 2003

75

Although surgical occlusion had been usedfor treatment of patients with this condition inthe past [20, 23], transcatheter embolizationand its usefulness have been reported in recentyears [24–28]. In our patients, clinical symp-toms and laboratory data improved immedi-ately after occlusion treatment. Although theportal venous pressures recorded after treat-ment tended to be higher than the levels beforetreatment, this relative increase in pressure wasnot associated with overt clinical problems.

Transcatheter embolization was performedusing one of three routes to access the intrahe-patic portosystemic venous shunts: transileo-colic obliteration, percutaneous transhepaticobliteration, or retrograde transcaval oblitera-tion. Of the patients in our study group, twowith multiple shunts underwent transileocolicobliteration. This procedure offers the easiestcontrol of catheters through its anterograde ac-cess route. On the other hand, transileocolicobliteration is the most invasive among thethree techniques used by our team because itrequires an abdominal incision with the patientunder general anesthesia or epidural tubing.Furthermore, this procedure carries the risk ofadhesion because it requires an abdominal in-cision. One of our patients experienced adhe-sive ileus after the procedure. Therefore,transileocolic obliteration should be limited topatients with multiple shunts located in bothlobes of the liver.

Percutaneous transhepatic obliteration isuseful for patients with contralateral distribu-tion of shunts because it offers good cathetercontrol of the portal vein contralateral to thepunctured side. In our patients, three with mul-tiple shunts in identical segments and one witha single shunt underwent percutaneous trans-hepatic obliteration. The procedure resulted insuccessful and complete obliteration as con-firmed angiographically. Although percutane-ous transhepatic obliteration is a relativelyinvasive technique, Ohta et al. [29] reportedthat 16.5% of their patients treated using per-cutaneous transhepatic catheterization devel-oped procedure-related complications.

Retrograde transcaval obliteration is theleast invasive technique, but it has some appli-cable limitations with regard to the type, num-ber, and location of intrahepatic portalsystemic venous shunts. The presence of alarge shunt close to the inferior vena cava al-lows easy catheterization of the main portalvein and retrograde portography to confirm theshunt and embolization. Because the inferiorvena cava and the proximal portion of the he-patic vein have relatively large diameters, sup-porting catheters during placement of embolicmaterial is difficult. Therefore, the cathetertype and shape and the embolic materialshould be selected carefully. We recommendthe use of a preshaped catheter adjusted to thehepatic vein, and detachable coils or a combi-

nation of microcatheters and microcoils. In ourgroup, four patients with a single shunt andone patient with two shunts in the adjacent he-patic vein were successfully treated by retro-grade transcaval obliteration. We believe thatthis technique should be applied to patientswith a small number of shunts.

The cause of intrahepatic portosystemicvenous shunt is not completely understood; how-ever, two major theories have been proposed. Thefirst is the congenital origin theory, which sug-gests the persistence of the communication be-tween the portal and hepatic venous systems thatoccurs during embryonal development. The sec-ond is the acquired theory, which suggests theshunt results from portal hypertension, trauma, orrupture of a portal vein aneurysm. Intrahepaticportosystemic venous shunt in patients who donot have liver cirrhosis or a history of trauma arethought to be congenital in origin. Embryologi-cally, the intra- and extrahepatic portal venoussystems develop by the selective persistence ofvitelline and umbilical systems between thefourth week and third month of fetal life (Fig. 8).In the congenital origin theory, intrahepatic por-tosystemic venous shunt is thought to representpersistent communication between cranial andcaudal hepatic sinusoids formed by vitellineveins and umbilical vein.

Macroscopically evident intrahepatic porto-systemic venous shunts have been classifiedaccording to their morphology [l, 4–6]. In the

20

10

BeforeEmbolization

Po

rtal

Ven

ou

s P

ress

ure

(cm

H2O

)

AfterEmbolization

3

2

1

BeforeEmbolization

Fis

cher

’s R

atio

AfterEmbolization

4

5

3

2

1

BeforeEmbolization

Ser

um

Am

mo

niu

m R

atio

AfterEmbolization

A B C

Fig. 7.—Graphs show changes in laboratory data and portal venous pressures after treatment in study group. A, Serum ammonium levels decreased significantly (p < 0.01) after treatment. Data are expressed as ratios relative to normal values because units and normal values dif-fered among institutions. B, Fischer’s ratios increased significantly (p = 0.028) after treatment. This value was not measured in four patients.C, Portal venous pressure increased significantly (p = 0.018) after treatment. Dotted lines represent range of normal values. In two patients, portal venous pressure washigher than normal values both before and after treatment; both patients had associated liver cirrhosis. Pressure levels were within normal range in other patients. Portalvenous pressure was not measured in three patients.

Dow

nloa

ded

from

ww

w.a

jron

line.

org

by 1

86.1

51.6

3.12

on

07/0

9/14

fro

m I

P ad

dres

s 18

6.15

1.63

.12.

Cop

yrig

ht A

RR

S. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved

76

AJR:181, July 2003

Tanoue et al.

our study, we divided the intrahepatic porto-systemic venous shunts into two types on thebasis of pathogenic mechanisms. One is ashunt that consists of an intrahepatic portalvenous–hepatic venous pathway, whereas theother is a shunt that consists of an intrahepaticportal venous–perihepatic venous pathwaythat includes the inferior phrenic veins, adrenalvein, and paraumbilical veins.

Intrahepatic Portal Venous–Hepatic Venous Pathway

This type of intrahepatic portosystemicvenous shunt is depicted as tubular or aneurys-

mal communication (single or multiple) be-tween intrahepatic portal veins and hepaticveins. To our knowledge, 42 cases have been re-ported in the English-language literature [l, 2,4–21, 24, 30–39]. These cases include 31 sim-ple types and 11 multiple types. Portal vein an-eurysms were reported in 29 (69%) of these 42cases. Most of these cases (76%) were not asso-ciated with liver cirrhosis. Eight of our patientshad this type of shunt.

Six (75%) of these pa-tients had neither cirrhosis nor any other hepaticdisease. Associated anomalies of the hepaticvessels, which included portal vein aneurysm,

hepatic venous anastomosis, and portal veinanastomosis, were observed (Figs. 2–4). Chag-non et al. [2] indicated that portal vein aneu-rysms including intrahepatic portosystemicvenous shunts are congenital. The latter twoanomalies have not, to our knowledge, been re-ported previously. These anomalies are thoughtto be formed in the hepatic sinusoid during fetaldevelopment. Because of the low rate of coex-isting liver cirrhosis and high rate of coexistinganomalies, intrahepatic portal venous–hepaticvenous shunts are likely to be of the congenitalorigin type.

A B

C D

Fig. 8.—Schematic drawings of normaldevelopment of intrahepatic portal andhepatic venous systems. A, Drawing shows embryo at 5 weeks’gestation. Vitelline venous plexus issurrounded by liver cords to form he-patic sinusoids. Bilateral umbilicalveins (UV) form sinusoids. CV = cardinalvein, HS = hepatic sinusoid, D = duode-num, VV = vitelline vein.B, Drawing shows embryo at 8 weeks’gestation. Sinusoids start to develop,forming portal and hepatic venous sys-tems. Right umbilical vein and cranialportion of left umbilical vein are re-gressed. Dorsal communication be-tween caudal vitelline veins persists aspart of main portal vein. Note tubularstructure between left umbilical veinand inferior vena cava, which is calledductus venosus. DV = ductus venosus,LVV = left vitelline vein. C, Drawing shows fetus at 12 weeks’gestation. Note advanced differentialgrowth of portal and hepatic venoussystems. Presence of residual commu-nication between hepatic venous sys-tem and portal venous system at thisstage corresponds with intrahepaticportosystemic venous shunt afterbirth. DV = ductus venosus, SMV = su-perior mesenteric vein, IVC = inferiorvena cava. D, Drawing shows fetus with normallydeveloped portohepatic venous systembefore birth. HV = hepatic vein, DV =ductus venosus, PV = portal vein.

Dow

nloa

ded

from

ww

w.a

jron

line.

org

by 1

86.1

51.6

3.12

on

07/0

9/14

fro

m I

P ad

dres

s 18

6.15

1.63

.12.

Cop

yrig

ht A

RR

S. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved

Angiography and Embolization of Portosystemic Venous Shunts

AJR:181, July 2003

77

Intrahepatic Portal Venous–Perihepatic Venous Pathway

Radiologically, this type of intrahepatic por-tosystemic venous shunt shows some commu-nications between the intrahepatic portal veinand perihepatic veins, and it drains into the in-ferior vena cava. Intrahepatic portosystemicvenous shunts through persistent paraumbili-cal veins are occasionally encountered in pa-tients with cirrhosis. Paraumbilical veins (alsocalled veins of Sappey) are known as potentialcommunications between the intrahepatic por-tal vein and veins of the abdominal wall [40].Other perihepatic veins, such as the inferiorphrenic veins and capsular vein, might consti-tute the communicating venous system andform portosystemic venous shunt in this type.

Intrahepatic portosystemic venous shunt be-tween the right portal vein and inferior venacava via venous structures around the rightlobe are less common. To our knowledge, 22cases of this type have been reported in the En-glish-language literature [1, 3, 21, 22, 31, 32,41–48], and these shunts were described as be-ing located in the bare area and posteroinferioraspect of the right lobe. A high incidence ofliver cirrhosis (71.4%) is reported in patientswith this type of shunt. Our two patients withthis type of shunt had liver cirrhosis. Becauseof the high rate of coexisting liver cirrhosis,the acquired theory could explain this type ofshunt, and the perihepatic veins are thoughtto develop in association with portal hyper-tension as intra- and extrahepatic collateralpathways from existing venous structures in-cluding paraumbilical veins, inferior phrenicveins, and adrenal vein.

Intrahepatic portosystemic venous shunts aredivided into two main types: intrahepatic portalvenous–hepatic venous communication and in-trahepatic portal venous–perihepatic venouscommunication. On the basis of the angio-graphic findings and clinical manifestations, theformer type is considered to be of congenital ori-gin, whereas the latter is thought to be an ac-quired condition associated with portalhypertension. Selection of the most suitable ac-cess route based on the morphology of shunt isan important aspect of treatment of intrahepaticportosystemic venous shunt by transcatheter em-bolization. In selected cases, retrograde tran-scaval obliteration is a useful, safe, and lessinvasive technique than the other options.

References

1. Mori H, Hayashi K, Fukuda T, Matsunaga N, Fu-tagawa W, Nagasaki M. Intrahepatic portosys-

temic venous shunt: occurrence in patients withand without liver cirrhosis.

AJR

1987;149:711–714

2. Chagnon SF, Vallee CA, Barge J, Chevalier LJ,Le Gal J, Blery MV. Aneurysmal portahepaticvenous fistula: report of two cases.

Radiology

1986;159:693–695 3. Itai Y, Kurosaki Y, Saida Y, Niitsu M, Kuramoto

K. CT and MRI in detection of intrahepatic porto-systemic shunts in patients with liver cirrhosis.

JComput Assist Tomogr

1994;18:768–773 4. Watanabe A. Portal-systemic encephalopathy in

non-cirrhotic patients: classification of clinicaltypes, diagnosis and treatment.

J GastroenterolHepatol

2000;15:969–979 5. Park JH, Cha SH, Han JK, Han MC. Intrahepatic

portosystemic venous shunt.

AJR

1990;155:527–528 6. Chevallier P, Oddo F, Souci J, Diaine B, Padovani

B. Macroscopic intrahepatic portosystemic venousshunt: review of the literature and reclassification.

JRadiol

2001;81:597–604 7. Araki T, Ohtomo K, Kachi K, et al. Magnetic res-

onance imaging of macroscopic intrahepatic por-tal-hepatic venous shunts.

Gastrointest Radiol

1991;16:221–224 8. Kawakatsu M, Vilgrain V, Belghiti J, Flejou JF,

Nahum H. Association of multiple liver cell ade-nomas with spontaneous intrahepatic portohe-patic shunt.

Abdom Imaging

1994;19:438–440 9. Gheorghiu D, Leibowits O, Bloom RA. Case re-

port: asymptomatic aneurysmal intrahepaticporto-hepatic venous shunt—diagnosis by ultra-sound.

Clin Radiol

1994:49:64–65 10. Horiguchi Y, Kitano T, Imai H, Ohsuki M,

Yamauchi M, Itoh M. Intrahepatic portal-sys-temic shunt: its etiology and diagnosis.

Gastroen-terol Jpn

1987;22:496–502 11. Tanaka S, Kitamura T, Fujita M, et al. Intrahepatic

venous and portal venous aneurysms examined bycolor Doppler flow imaging.

J Clin Ultrasound

1992;20:89–98 12. Jabra AA, Taylor GA. Ultrasound diagnosis of

congenital intrahepatic portosystemic venousshunt.

Pediatr Radiol

1991;21:529–530 13. Tarazov PG. Spontaneous aneurysmal intrahe-

patic portosystemic venous shunt.

Cardiovasc

In-tervent

Radiol

1994;17:44–45 14. Wittich G, Jantsch H, Tscholakoff D. Congenital

portosystemic shunt diagnosed by combined real-time and Doppler sonography.

J Ultrasound Med

1985;4:315–318 15. Okuda K, Kanda Y, Fukuyama Y, Musha H, Sum-

ikoshi T. Hepatic vascular anomalies in nonpara-sitic cysts of the liver.

Acta Hepatogastroenterol(Stuttg)

1976;23:110–113 16. Kozuka S, Sassa R, Kakumu S. An enormous in-

trahepatic shunt between portal vein and hepaticone.

Angiology

1975;26:365–371 17. Charnsangavej C, Soo CS, Bernardino ME,

Chuang VP. Portal-hepatic venous malformation:ultrasound, computed tomographic, and angio-graphic findings.

Cardiovasc Intervent Radiol

1983;6:109–111 18. Takayasu K, Moriyama N, Shima Y, et al. Sponta-

neous portal-hepatic venous shunt via an intrahe-patic portal vein aneurysm.

Gastroenterology

1984;86:945–948

19. Hondo T, Teragawa H, Munemori M, et al. Portal-hepatic venous shunt through a portal aneurysmcomplicated by hepatic encephalopathy and pulmo-nary hypertension.

Intern Med

1997;36:790–793 20. Miyauchi A, Okada S, Hashimoto T, et al. Surgi-

cal treatment of an enormous aneurysmal porta-hepatic venous fistula: report of a case.

SurgToday

1995;25:855–858 21. Santamaria G, Pruna X, Serres X, Inaraja L, Zuasna-

bar A, Castellote A. Congenital intrahepatic porto-systemic venous shunt: sonographic and magneticresonance imaging.

Eur Radiol

1996;6:76–78 22. Florio F, Nardella M, Balzano S, Giacobbe A, Perri

F. Congenital intrahepatic portosystemic shunt.

Cardiovasc Intervent Radiol

1998;21:421–424 23. Jeng LB, Chen ME. Intrahepatic portohepatic

venous shunt: a case report of successful surgicalresection.

Arch Surg

1993;128:349–352 24. Ohmoto K, Furui S, Saito M, Kokubo T, Itai Y, Ito

M. Enormous intrahepatic communication be-tween the portal vein and hepatic vein.

Clin Ra-diol

1986;37:513–514 25. Okada Y, Endo T, Kusano S, Yoshida M. Multiple

intrahepatic portosystemic venous shunts: treat-ment with steel-coil embolization.

AJR

1991;157:971–973

26. Maeda T, Mori H, Aikawa H, Komatsu E. Thera-peutic embolization of intrahepatic portosystemicshunts by retrograde transcaval catheterization.

Cardiovasc Intervent Radiol

1993;16:245–247 27. Yoshimitsu K, Andoh H, Kudo S, Matsuo Y, Shi-

moda Y. Multiple intrahepatic portosystemicvenous shunts: treatment by portal vein embo-lization.

Cardiovasc Intervent Radiol

1993;16:49–51

28. Kim IO, Cheon JE, Kim WS, et al. Congenital in-trahepatic venous shunt: treatment with coil em-bolisation.

Pediatr Radiol

2000;30:336–338 29. Ohta M, Hashizume M, Kawanaka H, et al. Com-

plications of percutaneous transhepatic catheter-ization of the portal venous system in patientswith portal hypertension.

J Gastroenterol Hepatol

1996;11:630–634 30. Naganuma H, Ishida H, Niizawa M, Masamune O.

Macroscopic intrahepatic portal-hepatic venousshunt not associated with chronic hepatic disease.Abdom Imaging 1994;19:153–156

31. Kudo M, Tomita S, Tochio H, Minowa K, TodoA. Intrahepatic portosystemic venous shunt: diag-nosis by color Doppler imaging.

Am J Gastroen-terol

1993;88:723–729 32. Mori H, Hayashi K, Amamoto Y. Membranous

obstruction of the inferior vena cava associatedwith intrahepatic portosystemic shunt.

Cardio-vasc Intervent Radiol

1986;19:209–213 33. Hanatate F, Matsuoka H, Mizuno Y, Kitamura T.

Porto-hepatic venous shunt via portal vein aneu-rysm with splenomegaly.

J Gastroenterol

1995;30:786–789

34. Kumano H, Kinoshita H, Hirohashi K. Aneurysm ofintrahepatic portal vein shown by percutaneous tran-shepatic portography. (letter)

AJR

1994;163:1000 35. Yamaguchi T, Kubota Y, Seki T, et al. Acquired

intrahepatic portal vein aneurysm.

Dig Dis Sci

1992;37:1769–1771 36. Stroz P, Asch M, Salem S, Sniderman K, Gallinger S.

Transhepatic embolization of a portohepatic venous

Dow

nloa

ded

from

ww

w.a

jron

line.

org

by 1

86.1

51.6

3.12

on

07/0

9/14

fro

m I

P ad

dres

s 18

6.15

1.63

.12.

Cop

yrig

ht A

RR

S. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved

78

AJR:181, July 2003

Tanoue et al.

fistula.

J Vasc Interv Radiol

1996;7:705–707 37. Sakai T, Katoh T, Asada K, Yagi K, Yoshida R. Por-

tal hepatic venous shunt via an intrahepatic portalvein aneurysm.

Intern Med

1992;31:899–903 38. Ito Y, Tarao K, Tamai S, et al. Portal vein aneurysm

in the liver associated with multiple vascular mal-formations.

J Gastroenterol

1994;29:776–781 39. Bezzi M, Mitchell DG, Needleman L, Goldberg

BB. Iatrogenic aneurysmal portal-hepatic venousfistula: diagnosis by color Doppler imaging.

J Ul-trasound Med

1998;7:457–461 40. Martin BF, Tudor RG. The umbilical and paraum-

bilical vein of man.

J Anat 1980;130:305–322 41. Park JH, Cha SH, Han JK, Han MC. Intrahepatic

portosystemic venous shunt. AJR 1990;155:527–528 42. Lalonde L, VanBeers B, Trigaux JP, Delos M, Prin-

got J. Focal nodular hyperplasia in association withspontaneous intrahepatic portosystemic venousshunt. Gastrointest Radiol 1992;17:154–156

43. Kadoya M, Takashima T, Matsui O, Kitagawa K,Chohtoh S. Unusual portosystemic collateral pen-etrating the liver parenchyma. Gastrointest Ra-diol 1985;10:115–118

44. Kudo M. Intrahepatic portosystemic venous shuntin liver cirrhosis: is it congenital or acquired? AJR1993;160:421–422

45. Ohnishi K, Hatano H, Nakayama T, Kohno K,Okuda K. An unusual portal-systemic shunt, most

likely through a patent ductus venosus: a case re-port. Gastroenterology 1983;85:962–965

46. Nagafuchi Y, Watanabe H, Kitahara K, et al. In-trahepatic portosystemic venous shunt: report oftwo cases assessed by helical computed tomogra-phy. Scand J Gastroenterol 1996;31:1132–1135

47. Ardic S, Deniz E, Arbatli M, Ardic FA, Aysun A.A rare intrahepatic portacaval tubular shunt in apatient with spastic paraparesis. Australas Radiol1997;41:181–184

48. Kanematsu M, Hoshi H, Imaeda T, Yokoyama R.Three-dimensional CT demonstration of intrahe-patic portosystemic venous shunt draining into theinferior vena cava. Br J Radiol 1997;70:418–420

The ARRS foundation, The Roentgen Fund, sponsors worthy educational, scientific, and research initiatives that contribute to excellence in radiology and medicine. A gift to The Roentgen Fund means progress in radiology. For more information call 800-438-2777.

Dow

nloa

ded

from

ww

w.a

jron

line.

org

by 1

86.1

51.6

3.12

on

07/0

9/14

fro

m I

P ad

dres

s 18

6.15

1.63

.12.

Cop

yrig

ht A

RR

S. F

or p

erso

nal u

se o

nly;

all

righ

ts r

eser

ved