ORIGINAL ARTICLES

Selective Shunts for Portal Hypertension:

Current Role of a 21-Year Experience
Hector Orozco, Miguel Angel Mercado, Jorge Granados Garcia,
Jorge Hernandez-Ortiz, Manuel Tielve, Carlos Chan, and Juan Luis Contreras
SEE EDITORIAL ON PAGE 552
The results of treatment of hemorrhagic portal
hypertension with selective shunts over a 21-year
period in a selected patient population are reported. Patients selected for surgical treatment
had good cardiopulmonary and renal function,
and most also had adequate liver function (141
Child-Pugh class A, 59 class B). Among 734
patients treated surgically for bleeding portal
hypertension, 221 had selective shunts (168 distal
splenorenal and 53 splenocaval shunts). Global
operative mortality (in the 21-year period) was
14% and 12% for Child-Pugh A patients. Operative
mortality in Child-Pugh A patients in the last 5
years was only 5%. The rate of rebleeding was 6%,
rate of incapacitating encephalopathy was 5%,

he concept of selectivity for portosystemic

shunting based on the principle of selective
deviation of venous blood from the esophagogastrosplenic area to the systemic circulation, maintaining the mesoportal blood flow intact, was developed in the 1960s by Warren et al.1 One of the more
favorable characteristics of these types of shunts is
that results are almost universally reproducible,
and indeed many centers have shown encouraging
results.2 Other therapies have emerged for the
treatment of portal hypertension. In the last 25
years, the roles of pharmacotherapy,3 sclerotherapy,4 transjugular intrahepatic portosystemic
shunting (TIPS),5 and liver transplantation (LT) 6 in
the treatment of patients with portal hypertension
have been defined. Nevertheless, preference for one
of these therapies is obvious in some places or
groups. Herein, we report our experience with
selective shunt surgery for the treatment of bleeding portal hypertension, its evolution as well as its
current role and interaction with other forms of
therapy.

Materials and Methods

Patients admitted to our hospital with active portal
hypertensive bleeding are treated with supportive mea-

and rate of survival was 65% at 15 years (last 5

years: 88% at 1 year and 85% at 5 years). Good
quality of life was demonstrated in 80% of surviving patients. Shunt patency was 94%. Postoperative portal blood flow changes occurred in 23% of
cases (8% diameter reduction, 14% thrombosis).
Compared with other forms of therapy (pharmacotherapy, sclerotherapy, and transjugular intrahepatic shunting), only liver transplantation offers
similar results for these patients. In countries in
which liver transplantation is not routinely performed, shunting with selective shunts is the
treatment of choice for patients with good liver
function.
Copyright r 1997 by the American Association for
the Study of Liver Diseases

sures and blood transfusions. Transendoscopic sclerotherapy and pharmacotherapy are the initial specific
measures for treatment, and in some cases a Sengstaken
tube is used. Only refractory cases with good liver
function (Child-Pugh class A and some class B) are
considered for emergency surgery. This is an exceptional
situation because in most patients who have no response
to initial measures, liver function deteriorates after
bleeding. Patients liver function is reclassified when
bleeding has been controlled and after recovery. ChildPugh class A patients, and some class B patients, are
considered for surgical treatment. Child-Pugh class C
patients are entered in the long-term sclerotherapy
program. Child-Pugh class C patients who fulfill other
criteria are the ones who are considered for the liver
transplantation program, and few cases are considered
for TIPS.
Low-risk patients are considered for surgical treatment if the criteria listed in Table 1 are fulfilled. These

From the Portal Hypertension Clinic, Instituto Nacional De La

Nutricion, Salvador Zubiran,Mexico City, Mexico.
Address reprint requests to Hector Orozco, M.D., Surgical
Division, Instituto Nacional de la Nutricion, Salvador Zubiran,
14000 Tlalpan, Mexico, D.F.
Copyright r 1997 by the American Association for the Study of
Liver Diseases
1074-3022/97/0305-0001$3.00/0

Liver Transplantation and Surgery, Vol 3, No 5 (September), 1997:pp 475-480

475

476

Orozco et al

Table 1. Patient Selection Criteria for Portal

Hypertensive Surgery
Good cardiopulmonary function
Good renal function
Good liver function
Albumin .3.0 g/dL
Total bilirubin ,2 mg/dL
Prothrombin time ,2 seconds
No ascites
No encephalopathy
Good nutritional status

criteria select only cases in which hemorrhagic portal

hypertension is the main problem.
Good cardiopulmonary function is defined when
results of cardiac evaluation (clinical, electrocardiographic, echocardiographic) are within normal limits.
Patients need an adequate cardiac reserve to tolerate a
major surgical procedure. In some cases, cardiac catheterization (Swan-Ganz catheter) is performed to rule out
pulmonary hypertension (an event sometimes associated
with portal hypertension). Patients with pulmonary
hypertension and a low cardiac output are excluded from
surgical treatment.
Renal function is evaluated by determination of
creatinine concentration and levels. If any abnormality is
detected in the results of these tests or of urine analyses,
a more detailed examination is carried out. This includes
a 24-hour creatinine depuration, as well as kidney and
urinary tract ultrasound. It is important for the patient to
have a healthy left kidney when selective shunting is
planned. Patients who have coraliform lithiasis of the
kidney are excluded.
The surgical procedures performed in our hospital are
those that preserve portal blood flow: selective shunting
(distal splenorenal, splenocaval), devascularization (our
modification of the Sugiura-Futagawa operation), and
low-diameter shunting (currently under long-term evaluation). The procedure is selected according to anatomic
characteristics, guided mainly by splanchnic angiography (Fig. 1). After selective injection of the celiac trunk
and superior mesenteric artery, the venous phase is
evaluated. The quality and diameter of the splenic,
superior mesenteric, and portal veins are carefully evaluated, as well as the characteristics of the left renal vein
and inferior vena cava. If the patient has good veins for
shunting, distal splenorenal shunting is performed. Direct and indirect splenocaval shunting are performed
when the splenic vein is long enough to reach the vena
cava; indirect splenocaval shunting (terminoterminal
splenorenal shunt) is considered in some cases in which
a classic distal splenorenal shunting is difficult.
If the splenic and renal veins are not suitable for

shunting, the patient is considered for our modification

of the Sugiura-Futagawa procedure.7 Veins can be considered unsuitable for shunting for reasons ranging from
total thrombosis of the splenic, superior mesenteric, and
portal veins to patent veins with inadequate distances
between them. Thin vessels are also considered unsuitable for shunting. Low-diameter shunts are considered
for patients with a history of splenectomy or multiplesession sclerotherapy (which preclude esophageal transection because of the inflammatory process and scar
tissue).
A total of 734 patients underwent surgery in the
portal hypertension clinic over a 21-year period (1973 to
1994). Of these, 221 were treated with selective shunting
(168 distal splenorenal and 53 splenocaval shunts).
Clinical histories were reviewed, and special attention
was given to rebleeding, encephalopathy rate, survival,
and quality of life. Angiographic follow-up of the shunt
and portal vein was performed yearly in available patients. Portal vein characteristics (diameter, flow) and
the patency of shunts were evaluated.
Patients general data are shown in Table 2. Surgery
was elective in 210 cases; these were the cases analyzed.
Ten Child-Pugh class C patients were excluded. Rebleeding was defined as the presence of hematemesis and/or
melena with hemoglobin level reduction. Encephalopathy was determined by clinical evaluation (carried out by
the group of hepatologists) and number connection test
when necessary. Patients who received treatment for
encephalopathy were also recorded.
Follow-up was performed on an ambulatory basis.
Patients were evaluated every 3 months for the first year,
then every 3 or 6 months for life. In the first 2
postoperative weeks, angiography was performed in all

Figure 1. Preoperative angiography. Venous

phase of the splenic artery selective injection,
showing the splenic and portal veins. Simultaneously, inferior vena cava catheterization and a
left renal vein angiography have been performed.
Vessels are adequate for shunting.

Selective Shunts for Variceal Bleeding

477

Shunt Patency
Table 2. General Data on Child-Pugh Class A and B
No. of patients
Age (mean)
Child-Pugh class A (n)
Child-Pugh class B (n)
Liver disease (n)
Alcoholic
Posthepatitis
Idiopathic pulmonary hypertension
Portal thrombosis

200
49.6 years
141
59
76
98
24
2

In the early postoperative period, 8 cases of shunt

thrombosis were recorded (Fig. 2). These patients
had rebleeding. Some patients underwent surgery
again (our modification of the Sugiura-Futagawa),
and others were treated with sclerotherapy. In 4
cases, late shunt thrombosis was shown. Thus the
total thrombosis rate was 6%.
Postoperative Portal Vein Angiography
Reduction in the diameter of the portal vein was
seen in 15 patients (8%) who developed the

cases. The status of the shunt (patency) was observed, as

well as the portal vein characteristics. Postoperative
thrombosis of the vein was also recorded. After the first
year, angiography was repeated between the fourth and
sixth postoperative year.
During follow-up, patientslife quality was evaluated.
Good quality of life was defined as ability to perform
normal activities without restriction, with the patient
going to the hospital for appointments and showing no
hospital dependence. Bad quality of life was defined as
inability to perform normal activities, with hospital and
physician dependence.

Results
Elective Cases
Of patients who underwent elective surgery, 141
were classified as Child-Pugh class A, and 59 were
class B.
Operative Mortality
Global mortality was 14%, and mortality among
Child-Pugh class A patients was 12% (18 of 141).
Our operative mortality with the Child A group in
the last 5 years of the study was 5%. In most cases,
operative mortality was attributable to liver failure.
Rebleeding
Rebleeding occurred in 11 (6%) of the 169 surviving patients; 8 of these patients had shunt thrombosis (4%).
Encephalopathy
By standard clinical evaluation, encephalopathy
was shown in 26 cases (14%). Incapacitating encephalopathy was seen in only 8 cases (4.4%).

Figure 2. (A) Distal splenorenal shunt. Venous

phase of the splenic artery selective injection,
showing the flow of the splenic vein completely
directed to the systemic circulation. (B) Nuclear
magnetic resonance showing the abdominal venous systems. Patent shunt is clearly visible, as
well as the mesoportal system. The main advantage over angiography is that no vessel punctures are needed.

478

Orozco et al

so-called pancreatic syphon (Fig. 3). Five of these

patients had alcoholic cirrhosis (Fig. 4).
Postoperative portal vein thrombosis was seen
in 25 cases (14%). Clinical manifestations of portal
vein thrombosis were protean. Some patients had
no symptoms and had only an angiographic finding, and others developed progressive liver failure
with hepatorenal syndrome and death (6 patients).
Acute postoperative thrombosis developed in 14
cases, and the clinical picture included poor postoperative evolution, with ascites and severe liver
failure (downgrading the patient to Child-Pugh
class C). Recovery of liver function was slow in
most of these cases.
Survival
Survival was 56% at 10 years and 48% at 15 years
for the whole group. For Child-Pugh class A
patients, it was 75% in the first year, 68% at 5 years,
and 65% at 15 years (Fig. 5). Quality of life was
good 14 in 80% of the survivors.
Last 5 Years
In the last 5 years, 148 patients underwent surgical
procedures that preserved portal blood flow, including 53 selective shunting, 41 low-diameter shunting, and 54 devascularization procedures (our
modification of the Sugiura-Futagawa operation).
For all cases, the criteria listed in Table 1 were
used. All procedures were performed by the same
two surgeons (H. Orozco and M. A. Mercado).
Operative mortality was 5.7%.
Survival for Child-Pugh class A patients who

Figure 4. Pancreatic syphon. Portal blood flow is

directed through collaterals to the low pressure
splenopancreatic area. Portal blood flow is lost.
We have observed this event mainly in alcoholic
patients, but also in some posthepatitic cases.

underwent shunting was 88% at 1 year and 85% at

5 years.

Discussion

Figure 3. Portal blood flow after distal splenorenal shunt. Diameter of portal vein is maintained in
80% of the cases in the postoperative period.

Surgery for portal hypertension has evolved in

recent decades. During the 1950s and 1960s, most
patients were considered for surgical treatment. No
case selection was done, and only procedures that
deprived the liver of portal blood flow were performed (end-to-side or side-to-side portocaval
shunting); poor results were the rule, particularly
in the long-term follow-up and in liver function.8
Patients died of liver failure rather than bleeding.
Thus, randomized controlled trials showed that no
real benefit was obtained with these procedures. At
that time, sclerotherapy appeared as an alternate
therapy for these patients. Initial reports of good

Selective Shunts for Variceal Bleeding

479

Figure 5. Survival curves according

to Kaplan-Meier. (See text.)

results appeared with this therapy in the 1970s and

in the first years of the 1980s; prospective randomized controlled trials were begun to compare surgical treatment with sclerotherapy.9,10 Results varied
from one trial to another but showed that survival
was similar with both forms of therapy, and better
bleeding control was achieved with the surgical
treatment.11 At the beginning of 1980, pharmacotherapy in the form of b-blockers also appeared.
The role and indications of b-blockers in management of rebleeding were not clear; there are currently signs that they could be useful for prophylactic treatment in patients who have never bled
before.3 Liver transplantation came of age in the
same decade and was also used to treat patients
with bleeding portal hypertension; good results
were obtained in patients with poor liver function.6
In the last years, TIPS appeared as an alternative
form of therapy. Several studies have shown that
long-term patency of the shunt is low and that the
encephalopathy rate (and certainly deteriorated
liver function) is high after the procedure. TIPS are
now used as a bridge for liver transplantation.5
In our own experience, we believe that surgery
(procedures that preserve portal blood flow) is the
treatment of choice for patients with good liver
function who can undergo elective surgery.
The results reported here, which reflect 21 years
of experience with selective shunts, show that
long-term survival and good quality of life can be
achieved in a well-selected patient population. We
have developed a large experience with procedures
that preserve portal blood flow.12,13 For cases in
which a shunt is not feasible, our modification of

the Sugiura-Futagawa operation offers excellent

results.13 In our experience with more than 150
patients, we have had results comparable to those
obtained with selective shunts. If the ideal conditions for a shunt are not found in a certain patient,
this kind of devascularization can be used also
achieving a low rebleeding and encephalopathy
rate, as well as good long-term survival.
We have had a global operative mortality of 6%,
a low rebleeding rate (5%), and a low incapacitating encephalopathy rate (4%) with this approach in
recent years. Approximately 80% of our patients
have a good postoperative quality of life.14 Only
orthotopic liver transplantation (OLT) can offer
better long-term results in this subset of patients.
Indeed, Child-Pugh class A patients who undergo
shunting have a postoperative quality of life comparable to that obtained with OLT.
These findings have been confirmed recently by
Hermann et al.15 In a 4-year period, they compared
selective shunting (distal splenorenal), TIPS, and
OLT. Although the patient population differed
considerably between groups (low-risk for shunting, poor liver function for the remaining two
groups), the investigators found interesting differences. Operative mortality was 0% for the shunting
group, 22% for the TIPS group, and 12% for the
OLT group. The rate of rebleeding was 0% for OLT,
3% for distal splenorenal shunt (DSRS), and 29%
for TIPS. Encephalopathy was lower in transplant
and shunt groups (0% in OLT, and 6% in DSRS, and
31% for TIPS). Survival was higher in the shunt
group.
Salam 16 raises the question of the preferred

480

Orozco et al

treatment method in developing countries. We

believe surgery for portal hypertension has a welldefined role around the world (low-risk, elective
cases). This role is surely expanded in countries in
which OLT is not routine. For countries in which
OLT is not a routine procedure (we believe that
almost 95% of the countries in the world are in this
condition), selective shunt surgery is the best form
of treatment for low-risk elective cases. In addition,
first world countries with good OLT programs have
begun to reconsider selective shunting surgery for
low-risk patients with hemorrhagic portal hypertension because of the limited number of donor
organs.
In most instances, a history of selective shunt
surgery does not preclude liver transplantation.
Several reports have shown that liver transplantation can be performed safely in patients with a
history of a selective shunt surgery.17 In our small
liver transplantation program, we do not consider
Child-Pugh class A patients. Child-Pugh class A
patients with history of variceal bleeding have the
same long-term survival (free of complications
related to immunosuppression) as liver transplantation patients. With regard to other forms of therapy,
we prefer surgery over b-blockers and sclerotherapy for our selected population because of the
high rebleeding rate of these two therapies (almost
50% in our experience). We consider these therapies to be the treatment of choice for patients with
poor liver function who are not candidates for liver
transplantation and have been bleeding. TIPS, in
our experience, has a limited role in low-risk
elective cases. They have a very high dysfunction
rate, are expensive, and promote liver failure and
encephalopathy in a considerable number of cases
when they work. We conclude that selective shunting has a relevant place in low-risk elective cases.
The key issue to obtain good results with this kind
of therapy is patient and procedure selection.

References
1. Warren WD, Zeppa R, Fomon JJ. Selective transsplenic
decompression of esophageal varices by distal splenorenal shunt. Ann Surg 1967;166:437-452.
2. W. Dean Warren Memorial Issue. Am J Surg 1990;160:
1-144.

3. Rodrguez-Perez F, Groszmann RJ. Pharmacologic

treatment of portal hypertension. Gastroenterol Clin
North Am 1992;21:15-40
4. Westaby D, Williams R. Status of sclerotherapy for
variceal bleeding in 1990. Am J Surg 1990;160:32-36.
5. Helton WS, Belshaw A, Althaus S, Park S, Coldwell D,
Johansen K, et al. Critical appraisal of the angiographic
portocaval shunt (TIPS). Am J Surg 1993;165:566-571.
6. Ringe B, Lang H, Tusch G, Pichlmayr R. Role of liver
transplantation in management of esophageal variceal
hemorrhage. World J Surg 1994;18:233-239.
7. Mercado MA, Takahashi T, Rojas G, Prado E, Hernandez J, Tielve M, et al. Ciruga en hipertension portal:
En que paciente y cual operacion? Surgery for portal
hypertension: Which patient and which operation? Rev
Invest Clin 1993;45:329-337.
8. Malt RA. Portosystemic venous shunts. N Engl J Med
1976;295:24-29,80-86.
9. Warren WD, Henderson JM, Millikan WJ, Galambos JT,
Brooks WS, Riepe S, et al. Distal splenorenal shunt
versus endoscopic sclerotherapy for long-term management of variceal bleeding: Preliminary report of a prospective randomized trial. Ann Surg 1986;203:454-459.
10. Rikkers LF, Burnett DA, Volentine GD, Buchi KN, Cormier
RA. Shunt surgery versus endoscopic sclerotherapy for
long term treatment of variceal bleeding: Early results of
a randomized trial. Ann Surg 1987;206:261-269.
11. Henderson JM, Kutner MA, Millikan WJ, Galambos JT,
Riepe SP, Brooks WS, et al. Endoscopic variceal sclerosis compared with distal splenorenal shunt to prevent
recurrent variceal bleeding in cirrhosis: A prospective
randomized trial. Ann Intern Med 1990;112:262-269.
12. Orozco H, Mercado MA, Takahashi T, Garca-Tsao G,
Guevara L, Hernandez-Ortz J, Hernandez-Cendejas A.
Role of the distal splenorenal shunt in management of
variceal bleeding in Latin America. Am J Surg 1990;160:
86-89.
13. Orozco H, Mercado MA, Takahashi T, Hernandez J,
Capellan JG, Garca-Tsao G. Elective treatment of
bleeding varices with the Sugiura operation over 10
years. Am J Surg 1992;163:585-589.
14. Orozco H, Mercado MA, Takahashi T, Rojas G, Hernandez J, Tielve M. Survival and quality of life after portal
blood flow preserving procedures in patients with portal
hypertension and liver cirrhosis. Am J Surg 1994;168:
232-234.
15. Hermann RE, Henderson JM, Vogt DP, Mayes JT,
Geisinger MA, Agnor C. Fifty years of surgery for portal
hypertension at the Cleveland Clinic Foundation: Lessons and prospects. Ann Surg 1995;221:459-468.
16. Salam A. Discussion. Ann Surg 1995;221:466.
17. Mazzaferro V, Todo S, Tzakis AG, Stieber AC, Makowka
L, Starzl TE. Liver transplantation in patients with
previous portosystemic shunt. Am J Surg 1990;160:111116.