Innovative techniques for and results of

portal vein reconstruction in living-

related liver transplantation
Ibrahim K. Marwan, MD, Ahmed Tarek I. Fawzy, MD, Hiroto Egawa, MD, Yukihiro Inomata, MD,
Shinji Uemoto, MD, Katsuhiro Asonuma, MD, Tetsuya Kiuchi, MD, Michihiro Hayashi, MD, Shiro
Fujita, MD, Yasuhiro Ogura, MD, and Koichi Tanaka, MD, Kyoto, Japan

Background. Portal vein reconstruction is a crucial factor affecting the outcome of a successful living-
related liver transplantation. We describe here our experience with portal vein reconstruction in 314
cases of living-related liver transplantation with use of novel surgical modalities to enable the transplant
surgeons to deal with any size mismatch between the donor’s and recipient’s portal veins.
Methods. Portal vein reconstruction was classified into 2 major groups, anastomosis without and with
a vein graft. When there was no stenosis of the recipient portal vein and the diameter was the same, the
portal trunk was used for anastomosis. When the diameter mismatch was minimal, branch patch anas-
tomosis was feasible. When the recipient portal vein was significantly stenotic and the portal vein of the
graft was long enough, we removed the stenotic trunk and constructed an anastomosis between the graft
portal vein and the confluence of the recipient portal vein. When the graft portal vein was short, a vein
graft was interposed. The vein patch technique was preferable when the diameter of the graft vein was
not large enough for the interposition technique.
Results. Anastomosis without vein graft included trunk anastomosis (n = 156), branch patch anasto-
mosis (n = 39), and confluence anastomosis (n = 22). Anastomosis with vein graft used the interposi-
tion technique (n = 77) and vein patch technique (n = 27). The origin of the grafts was mostly from the
maternal left ovarian vein (70%) or the paternal inferior mesenteric vein (27%). Complications related
to portal vein reconstruction occurred in 16 (5%) patients: portal vein thrombosis in 8, stenosis in 7,
and fatal rupture in 1 patient. The incidence of complications was similar for all techniques except for
confluence anastomosis.
Conclusion. Our innovative techniques should be helpful for overcoming diameter or length mismatches
in portal vein reconstruction in pediatric liver transplantation. (Surgery 1999;125:265-70.)

From the Department of Transplantation Immunology, Faculty of Medicine, Kyoto University, Kyoto, Japan,
and the Liver Institute and Department of Surgery, Faculty of Medicine, Al Menoufyia University, Cairo Egypt

PORTAL VEIN RECONSTRUCTION IS a crucial factor for
success in liver transplantation. In pediatric liver
transplantation occlusion and stenosis of the recip-
ient portal vein, often observed in biliary atresia,
and size mismatch between the donor and the
recipient portal veins in the case of reduced-size
grafts present specific issues regarding surgical
techniques for portal vein reconstruction. In living-
related liver transplantation (LRLT) especially
availability of vascular grafts from living donors is
limited.1 Techniques that use an interposition vein
graft consisting of donor ovarian vein, inferior
mesenteric vein, saphenous vein, or iliac vein grafts
have been reported.1-3 However, the interposition
graft is not sufficient in all cases, so it becomes
necessary to overcome the diameter mismatch
between the donor and the recipient portal veins
when the diameter of the available graft is small.
In this study we present our experience with portal
vein reconstruction over the last 7 years describing
different surgical modalities and focusing on our
novel portal vein patch graft technique.

PATIENTS AND METHODS

Accepted for publication Sept 27, 1998.
Reprint requests: Hiroto Egawa, MD, Department of
Transplantation Immunology, Transplant Surgery, Kyoto
University Hospital, 54 Kawahara-cho, Shogoin, Sakyo-ku,
Kyoto 606, Japan.
Copyright © 1999 by Mosby, Inc.
0039-6060/99/$8.00 + 0 11/56/95114
Between June 1990 and August 1997, 314
patients with end-stage liver disease underwent
321 LRLT at Kyoto University Hospital.
There were 193 male and 121 female recipients.
Their ages ranged from 1 month to 61 years (6.8 ±
9.9 years, mean ± SD) and their weight from 3.1 to

SURGERY 265
266 Marwan et al Surgery
March 1999

Table I. Recipient profiles

Sex
Male 193 (61.5%)
Female 121 (38.5%)
Age (mean) 1 mo-61 y (6.8 ± 9.9 y)
Age distribution (y)
<1 85 (27%)
1-2 58 (18.5%)
2-10 94 (30%)
10-18 50 (16%)
>18 27 (9.5%)
Weight (kg) (mean) 3.1-77.8 (18.8 ± 16)
Primary diagnosis
Biliary atresia 239 (76%)
Metabolic diseases 26 (8.3%)
Fulminant hepatitis 12 (3.8%)
Liver cirrhosis 12 (3.8%)
Alagille’s syndrome 10 (3.2%)
Intrahepatic cholestasis 4 (1.3%)
Primary biliary cirrhosis 6 (2%)
Hepatic malignancy 3 (1%)
Budd-Chiari syndrome 2 (0.6%)
Fig 1. Portal vein reconstruction is classified into 2 major
Retransplantation 7 (2.2%)
groups: group 1, anastomosis without vein graft including
trunk anastomosis (n = 156), branch patch anastomosis (n
= 39), or confluence anastomosis (n = 22); group 2, anas-
tomosis with vein graft including interposition technique vein reconstruction into 2 major groups, namely,
(n = 77) and patch graft technique (n = 27). group 1, anastomosis without a vein graft, and group
2, anastomosis with a vein graft, as shown in Fig 1.
Group 1 included trunk anastomosis, branch
77.8 kg (18.8 ± 16 kg). Age distribution and diag- patch anastomosis at the bifurcation of the right
nostic indications are shown in Table I. Biliary atre- and left branches of the recipient portal vein, and
sia was the most frequent disease (76%). confluence anastomosis at the confluence of the
There were 175 male donors and 146 female superior mesenteric vein and the splenic vein of
donors (165 fathers, 141 mothers, 7 brothers, 3 sis- the recipient.
ters, 3 husbands, 1 wife, and 1 daughter). Their Group 2 included the interposition graft tech-
ages ranged from 21 to 59 years (34 ± 7.6 years) nique and the patch technique, described briefly
and their weight from 39 to 95 kg (58.8 ± 9.9 kg). here. In the interposition graft technique a vein graft
The graft weight ranged from 160 to 630 g (283 is interposed between the graft’s portal vein and the
± 75.8 g) and graft-recipient body-weight ratios recipient superior mesenteric vein (SMV)–splenic
(Graft weight [g]/Recipient body weight [g] × 100 vein (SPV) confluence. The vein grafts used were
[%]) from 0.48% to 10.1% (2.4% ± 1.46%). The usually the left ovarian vein from a female donor and
grafts were lateral segments (segments 2 and 3) in the inferior mesenteric vein from a male donor. The
215 cases (67%), the left lobe (segments 2, 3, and vein graft was anastomosed to the recipient’s SMV-
4 including the middle hepatic vein) in 77 cases SPV confluence with 7/0 Maxon (Davis & Geck) run-
(24%), extended left lateral segments (segments 2, ning suture without growth factor before implanta-
3, and part of segment 4 not including the middle tion of the liver graft. The patch technique is our
hepatic vein) in 28 cases (8.7%), and the right lobe original innovation designed to overcome diameter
(segments 5, 6, 7, and 8) in 1 case (0.3%). mismatch between an available living vascular graft
All operations were performed with informed and a graft portal vein, as shown in Fig 2.
consent of the recipients or their parents and of Indication of techniques. Modalities for portal
the donors and approved by the Ethics Committee vein reconstruction were chosen according to the
of Kyoto University. The donor and recipient oper- diameter, wall status, and length of the recipient’s
ations are reported elsewhere.3,4 portal vein and the diameter and length of the
Surgical techniques for portal vein reconstruc- graft portal vein. The most important thing in por-
tion. We divided the surgical techniques for portal tal anastomosis is to obtain enough diameter and
Surgery Marwan et al 267
Volume 125, Number 3

Fig 2. Recipient’s portal vein should be carefully dissected all the way down to SMV-SPV confluence, after
which small Statineski clamp is applied (2a). Portal vein (PV) is incised anteriorly along its axis until point of
intersection of axes of portal vein, SMV, and SPV at confluence is reached. Incision is then extended along
axes of SMV and SPV for a short distance followed by trimming of edges (2a) to make transverse length of gap
created at confluence C-D match short limb of rectangular vein graft C´-D´. This graft is created by bisection
of vein graft harvested from donor ovarian vein or inferior mesenteric vein or from recipient’s own portal vein
or inferior mesenteric vein (2b). Vein graft is mounted over opening in recipient portal vein and fixed with 4
stay sutures at corners (2c). Transverse limb CC´-DD´ is sutured with 7/0 Maxon in interrupted fashion and 2
long sides of the patch graft (AA´-CC´and BB´-DD´) are sutured with 7/0 Maxon sutures in continuous run-
ning fashion (2d). Anastomosis between donor’s portal vein and recipient’s portal vein is completed with
patch graft in end-to-end fashion with 7/0 Maxon running sutures.

length to allow for adequate portal flow and to
obtain size matching between the graft portal vein
and the recipient portal vein. Fig 3 shows our deci-
sion tree for portal vein reconstruction. At dissec-
tion of the recipient portal vein during hepatecto-
my we assessed the recipient portal vein and the
amount of portal flow by unclamping the portal
vein. When there was no stenosis leading to
reduced portal flow, we used the native trunk for
anastomosis. When there was diameter mismatch,
branch patch anastomosis was more viable than
trunk anastomosis. When the trunk was significant-
ly stenotic, approximately smaller than 4 mm in
diameter, the native trunk was removed to obtain
adequate portal flow. The portal vein was usually
enough large at the level of the confluence. The
next issue was the distance between the confluence
of a recipient and the graft portal vein. When the
graft portal vein could reach the confluence, we
made an anastomosis between the graft portal vein
and the confluence. When the graft portal vein
could not reach easily the confluence, we needed
the vein graft. Because the origins of vein grafts
were limited in living donors, the diameter of the
vein graft was not always large enough. The inter-
position technique was used when the diameter
and the length of the vein graft were optimal,
whereas the patch technique was preferable when
the diameter of the vein graft was too small for the
interposition technique.
RESULTS
The recipient portal vein diameter ranged from
1.8 to 16 mm (6.5 ± 2.7 mm) and that of the donor
portal vein ranged from 4.4 to 11.5 mm (7.7 ± 1.5
mm). There were 71 recipients (22%) with a portal
vein diameter of less than 4 mm. Distribution of por-
tal vein reconstruction is shown in Table II. The patch
technique was used for 27 patients (8.9%). The ori-
gins of the vein grafts were mentioned in Table III.
Complications related to portal vein reconstruc-
tion occurred in 16 of 321 LRLTs (5%). Portal vein
268 Marwan et al Surgery
March 1999

Fig 3. Decision tree to choose techniques for portal vein (PV) reconstruction.

Table II. Modality of portal vein reconstruction and complications

Thrombus Stenosis Rupture Total in technique
Group 1: without vein graft
Trunk (n = 156) 4 (2.6%) 4 (2.6%) 0 8 (5.1%)
Branch patch (n = 39) 2 (5.6%) 0 0 2 (5.1%)
Confluence (n = 22) 0 0 0 0 (0%)
Total in group 1 (n = 217) 6 (2.8%) 4 (1.8%) 0 10 (4.6%)
Group 2: with vein graft
Interposition (n = 77) 2 (2.6%) 2 (2.6%) 1 (1.3%) 5 (6.5%)
Patch (n = 27) 0 1 (3.7%) 0 1 (3.7%)
Total in group 2 (n = 104) 2 (2.0%) 3 (2.9%) 1 (1.0%) 6 (6%)
Total overall (n = 321) 8 (2.5%) 7 (2.2%) 1 (0.3%) 16 (5%)

Table III. Techniques and origin of vein grafts bosis had the episodes within 3 days after trans-
Technique Origin No. plantation. Their graft-recipient weight ratios were
Interposition Left ovarian vein 58 (75%) 3.05, 4.09, and 4.36. Two others with hypoxemia
(n = 77) resulting from intrapulmonary shunt had fatal por-
Inferior mesenteric vein 18 (23%) tal thrombosis 1 month after transplantation. All 5
External iliac vein 2 (2%) underwent immediate thrombectomy. One of
Patch (n = 27) Left ovarian vein 14 (51.9%) them survived. Another 3 had portal thrombosis
Inferior mesenteric vein 10 (37%) later than 2 years after transplantation. Two of
Left gastric vein 2 (7.4%) them had been complicated with refractory cholan-
Own portal vein 1 (3.7%) gitis, presumably from ABO-incompatible trans-
plantation. One patient, whose thrombosis had
been diagnosed early, received intraportal infusion
thrombosis was seen in 8, stenosis in 7, and rupture with heparin through a silicone tube left after per-
of the anastomotic line in 1, which was fatal, as cutaneous transhepatic portography and followed
shown in Table IV. There were no portal vein com- by administration of oral anticoagulants. The
plications in patients with confluence anastomosis, remaining 2 patients have been followed up with-
and the incidence of complications was similar out treatment for portal thrombus because there
among the remaining 4 techniques, ranging from was no portal vein both radiologically and histolog-
3.7% to 6.5%. Three of the 8 patients with throm- ically at the diagnosis of the portal vein thrombosis
Surgery
Volume 125, Number 3
Table IV. Management and prognosis of portal vein complications
Complication Management
Thrombosis Thrombectomy + anticoagulants
Stenosis PTBD
Rupture Exploration
PTBD, Percutaneous transhepatic balloon dilation.
*Mortality rate for patients without portal vein complications was 16%.
and the symptoms of portal hypertension were
acceptable. Five of 7 patients with stenosis had
annular stricture at the anastomotic line that
responded to percutaneous transhepatic balloon
dilatation, whereas the remaining 2 were cases
resulting from extraluminal fibrin plugs requiring
excision of the fibrin ring. One patient had fatal
anastomotic rupture of portal aneurysm because of
severe peritoneal sepsis. There were no complica-
tions such as pancreatitis relating to the dissection
of the confluence.
DISCUSSION
In our LRLT series biliary atresia was the most
frequent indication. Patients with this disorder are
likely to have repeated attacks of cholangitis or
laparotomies leading to stenosis of the portal vein.
Only 1 patient had occlusion through the portal
vein to the superior mesenteric vein where the
portal flow could not be reconstructed, but all
other patients had a patent vessel at the conflu-
ence of the superior mesenteric vein and the
splenic vein. The key concept for portal recon-
struction is to obtain enough size and length to
allow for adequate portal flow as described above.
In almost all cases with a stenotic portal vein, the
diameter of the portal vein at the confluence is
wide enough to allow for adequate flow. We per-
formed a direct anastomosis between the portal
vein of the hepatic graft and the portal vein at the
confluence, but this anastomosis requires a suffi-
ciently long portal vein of the hepatic graft. In
LRLT the portal vein of the hepatic graft is the left
portal branch, which is often short. To solve this
problem, we placed an interposed vein graft
between the confluence and the graft portal vein
or widened the diameter of the stenotic portal vein
by means of the innovative patch graft technique.
We usually obtain the left ovarian vein from a
maternal donor or the inferior mesenteric vein
from a paternal donor because these can be safely
obtained from the same surgical field where the
hepatic graft was harvested. When the diameter of
the vein graft is adequate, the graft can then be
Marwan et al 269
Prognosis
Alive Dead Mortality (%)*
4 4 50
7 — 0
— 1 100
used as an interposed vein graft, but a vein graft
with a small diameter can be used only when treat-
ed with the patch technique.
The incidence of portal vein thrombosis in the
pediatric age group has been reported to be between
0% and 33% and is generally much greater than that
observed in adult recipients (1% to 2%).5-9 In our
study it was 5%. Several factors influence the devel-
opment of portal vein complications in children. A
reduced graft from a large donor has a large portal
vein, and pediatric patients with biliary atresia have a
small, sclerotic portal vein. Moreover, compression
by the hepatic graft, especially in cases of small babies
receiving large grafts, and the low pressure of the
portal vein related to collaterals could increase the
risk of thrombosis.9 Results obtained with the conflu-
ence technique were the best and those with the
patch technique second best, although there was no
statistical significance. Results of the other tech-
niques were similar.
The mortality for portal vein thrombosis was
high, whereas interventional treatment was found
to be effective for stenosis. Greater effort is need-
ed to prevent portal thrombosis by considering eti-
ologic factors such as surgical techniques and
medical background. Two patients with fatal
thrombosis had hypoxemia as a result of intrapul-
monary shunt accompanied by high hematocrit.
After this experience we managed to prevent
thrombosis in such patients by prophylactic anti-
coagulant therapy. Furthermore, daily Doppler
ultrasonography during the early postoperative
period is mandatory for early detection of vascular
complications before they develop into fatal portal
vein thrombosis. We lost a chance to treat throm-
bosis for patients with cholangitis related to ABO-
incompatible transplantation because of late diag-
nosis. Hepatic blood flow should be followed by
Doppler ultrasonography in the long term for
such patients even when the liver function and
general condition are stable.
In summary, our innovative techniques proved
to be helpful for overcoming diameter or length
mismatch in portal vein reconstruction in LRLT.
270 Marwan et al Surgery
March 1999

REFERENCES Vascular complications after orthotopic liver transplanta-

1. Tokunaga Y, Tanaka K, Yamaoka Y, Ozawa K. Portal vein
graft in living related hepatic transplantation. J Am Coll
Surg 1994;178:297-9.
2. Broelsch CE, Whitington PF, Emond JC, et al. Liver trans-
plantation in children from living related donors. Ann Surg
1991;214:428-39.
3. Saad S, Tanaka K, Inomata Y, et al. Portal vein reconstruc-
tion in pediatric liver transplantation from living donors.
Ann Surg 1998;227:275-81.
4. Yamaoka Y, Ozawa K, Tanaka A, et al. New device for harvest-
ing the hepatic graft from a living donor. Transplantation
1991;52:157-60.
5. Langnas AN, Marujo W, Stratta RJ, Wood RP, Shaw B Jr.
tion. Am J Surg 1991;161:76-83.
6. Funaki B, Rosenblum JD, Leef JA, et al. Portal vein stenosis
in children with segmental liver transplantation: treatment
with percutaneous transhepatic venoplasty. AJR Am J
Roentgenol 1995;165:161-5.
7. Kalayoglu M, D’Alessandro AM, Knechtle SJ, Eckhoff DE,
Pirsch JD, Judd R, et al. Long term results of liver trans-
plantation for biliary atresia. Surgery 1993;114:711-8.
8. Tanaka K, Uemoto Y, Tokunaga Y, et al. Living related liver
transplantation in children. Am J Surg 1994;168:41-8.
9. Millis JM, Seaman DS, Piper JB, et al. Portal vein thrombo-
sis and stenosis in pediatric liver transplantation.
Transplantation 1996;62:748-5.

Acknowledgment
We would like to thank the reviewers listed below who contributed their time recently to review
manuscripts for Surgery. These individuals, as well as members of the Editorial Board, commit their
time and careful consideration to ensure that articles published in Surgery reflect the highest stan-
dards of scholarship and relevance.
Andrew L. Warshaw
Michael G. Sarr
Editors in Chief
Altman, R. Peter Delmonico, Francis L.
Columbia-Presbyterian Medical Center Massachusetts General Hospital
Bannon, Michael P. DeMeester, Thomas R.
Mayo Clinic University of Southern California School of
Blend, Michael Medicine
University of Illinois Hospital Fan, S. T.
Breen, Lisa Queen Mary Hospital, University of Hong Kong
Brigham and Women’s Hospital Fink, Mitchell P.
Bromberg, Jonathan S. Beth Israel Deaconess Medical Center
University of Michigan Fong, Yuman
Brooks, David C. Memorial Sloan-Kettering Cancer Center
Brigham and Women’s Hospital Foster, Roger S., Jr
Cady, Blake Crawford Long Hospital of Emory University
Women and Infants Hospital of Rhode Island Gloviczki, Peter
Cambria, Robert A. Mayo Clinic
Medical College of Wisconsin Gluckman, Jack L.
Carter, Edwin C. University of Cincinnati Medical Center
Massachusetts General Hospital Gorelick, Fred S.
Cho, Sang I. Veterans Administration Medical Center
Boston Medical Center Grendall, James H.
Choti, Michael A. New York Hospital, Cornell Medical Center
Johns Hopkins Hospital Hallet, John W.
Clark, Orlo H. Mayo Clinic
University of California, San Francisco/Mount Hannon, Edward L.
Zion Medical Center State University of New York Albany School of
Dawes, Lillian G. Public Health
Northwestern University Medical School –continued on page 317