HEPATOLOGY, October 1998, p. 1154-1158, Vol. 28, No.
4
Clinical Challenge
Management of Ectopic Varices
Ian D. Norton1, James C.
Andrews2, and Patrick S. Kamath1
From the 1 Division of Gastroenterology and
Hepatology and the 2 Department of Diagnostic Radiology,
Mayo Clinic and Foundation, Rochester, MN
INTRODUCTION
The term "ectopic varices" is sometimes reserved for
abnormally dilated veins associated with gastrointestinal mucosa
and, therefore, with the potential for gastrointestinal hemorrhage.
However, the term has also been used loosely to describe
portosystemic collateral veins in the abdominal wall and
retroperitoneum. The distinction between "ectopic varices" and
collaterals that are commonly found on the abdominal wall and
retroperitoneum of patients with portal hypertension is one of
semantics. Thus, ectopic varices may be best defined as large
portosystemic venous collaterals occurring anywhere in the abdomen
except in the cardioesophageal region.
Ectopic varices are an unusual cause of gastrointestinal
hemorrhage, but account for up to 5% of all variceal
bleeding.1 The clinician caring for patients with
gastrointestinal bleeding must be aware of this entity, because
diagnosis and management of ectopic varices differ from that of
esophagogastric varices. Furthermore, the prognosis from bleeding
ectopic varices may be poor, with one study quoting 40% mortality
at initial bleed from duodenal varices.2 The literature
on this subject consists mainly of small series and case reports
with no randomized trials of therapeutic modalities. However, a
review of the literature does provide sufficient information from
which rational management decisions can be made.
PREVALENCE
Ectopic varices account for between 1% and 5% of all variceal
bleeding. 1,3 Ectopic varices are a relatively common
finding at endoscopy in patients with portal hypertension. The
prevalence seems to be related to the cause of the portal
hypertension and the technique used to show the varices. In
patients with intrahepatic portal hypertension, duodenal varices
are seen in 40% of patients undergoing angiography,3
whereas anorectal varices have been reported in between 10% and 40%
of cirrhotic patients undergoing colonoscopy. 4,5 It is
important to differentiate anal varices from hemorrhoids: Anal
varices collapse with digital pressure, whereas hemorrhoids do
not.4 In patients with portal hypertension caused by
obstruction of the portal or splenic veins, duodenal varices are
more prevalent than in patients with intrahepatic portal
hypertension. The prevalence is higher if angiography is used to
show varices. In fact, most patients with portal or splenic vein
thrombosis are likely to have duodenal varices shown on
angiography. 6,7 The majority of patients with duodenal
varices visualized on endoscopy have extrahepatic portal
hypertension. In contrast to duodenal varices, it appears that most
cases of varices in other portions of the small intestine and
colonic varices are seen in patients with intrahepatic portal
hypertension who have previously undergone abdominal
surgery.6 In the west, because the prevalence of
extrahepatic portal hypertension is low, most bleeding from ectopic
varices is usually associated with intrahepatic portal
hypertension. 6,8 Stomal varices are a particularly
common cause of ectopic varices and occur in patients with
intrahepatic portal hypertension caused by primary sclerosing
cholangitis.6
Although ectopic varices can occur at several sites, bleeding
ectopic varices are most commonly found in the duodenum and at
sites of previous bowel surgery including stomas. In a review of
169 cases of bleeding ectopic varices, 17% occurred in the
duodenum, 17% in the jejunum or ileum, 14% in the colon, 8% in the
rectum, and 9% in the peritoneum. In the review, 26% bled from
peristomal varices and a few from infrequent sites such as the
ovary and vagina.
ETIOPATHOGENESIS
Ectopic varices are natural portosystemic shunts, resulting
from portal hypertension, and may occur at any site in the gut.
Collaterals occur where the portal venous system is in apposition
to systemic veins and are typically in the gastroesophageal
junction and in the anorectum but can also occur around the
umbilicus, ovaries, and bare area of the liver. Resistance within
these collateral vessels is initially high; in contrast, resistance
in the portal vasculature is low. Thus, blood flows preferentially
through the portal venous circulation, and collateral vessels
remain collapsed. However, with the development of portal
hypertension, intrahepatic vascular resistance increases, allowing
the portosystemic collaterals to open. In extrahepatic portal
venous obstruction when the defined portosystemic collaterals are
obstructed, less well-defined collaterals, such as in the duodenum
and colon, open up. Surgery that results in apposition of bowel to
abdominal wall or other structures drained by the systemic venous
circulation also results in collaterals being formed in unusual
sites. The best recognized of these ectopic varices is stomal
varices formed at ileostomy sites in patients with primary
sclerosing cholangitis who have undergone colectomy with creation
of an ileostomy for inflammatory bowel disease, usually ulcerative
colitis.9
Ectopic varices account for between 1% and 5% of all variceal
bleeding. 1,3 Ectopic varices are a relatively common
finding at endoscopy in patients with portal hypertension. The
prevalence seems to be related to the cause of the portal
hypertension and the technique used to show the varices. In
patients with intrahepatic portal hypertension, duodenal varices
are seen in 40% of patients undergoing angiography,3
whereas anorectal varices have been reported in between 10% and 40%
of cirrhotic patients undergoing colonoscopy. 4,5 It is
important to differentiate anal varices from hemorrhoids: Anal
varices collapse with digital pressure, whereas hemorrhoids do
not.4 In patients with portal hypertension caused by
obstruction of the portal or splenic veins, duodenal varices are
more prevalent than in patients with intrahepatic portal
hypertension. The prevalence is higher if angiography is used to
show varices. In fact, most patients with portal or splenic vein
thrombosis are likely to have duodenal varices shown on
angiography. 6,7 The majority of patients with duodenal
varices visualized on endoscopy have extrahepatic portal
hypertension. In contrast to duodenal varices, it appears that most
cases of varices in other portions of the small intestine and
colonic varices are seen in patients with intrahepatic portal
hypertension who have previously undergone abdominal
surgery.6 In the west, because the prevalence of
extrahepatic portal hypertension is low, most bleeding from ectopic
varices is usually associated with intrahepatic portal
hypertension. 6,8 Stomal varices are a particularly
common cause of ectopic varices and occur in patients with
intrahepatic portal hypertension caused by primary sclerosing
cholangitis.6
On rare occasions varices may be present in the colon in the
absence of portal hypertension caused by anomalies of portal venous
outflow, such as congenital anomalies of normal portosystemic
anastomoses,10 abnormal vessel structure,11
or arteriovenous fistulae.12 Familial occurrence of
colonic varices has also been described.11
Wall tension is thought to be the major determinant of
hemorrhage from varices.13 Tension in the varix wall is
proportional to transmural pressure across the vessel wall and the
radius of the vessel. Therefore, the major determinants of rupture
of ectopic varices, as in esophageal varices, are likely to be
vessel size and portal pressure.
MANAGEMENT OF BLEEDING ECTOPIC VARICES
Because of the infrequency with which bleeding ectopic
varices present, there have been no randomized trials on the
management of this condition and it is unlikely that there ever
will be such trials. Therefore, management depends on local
expertise and the cause of portal hypertension. The management of
varices in the abdominal wall and retroperitoneum (also called
intra-abdominal varices), of stomal varices, and of varices around
the biliary tree is different from that of other ectopic varices
and is discussed separately.
Most patients with ectopic variceal hemorrhage present with
sudden, profuse melena or hematochezia. Duodenal varices may also
be associated with hematemesis. Bleeding from ectopic varices is
suspected in all patients with portal hypertension and
gastrointestinal bleeding in whom upper and lower gastrointestinal
endoscopy have not shown a bleeding source.
Initial management involves clinical assessment and
hemodynamic stabilization. The use of octreotide to reduce
splanchnic blood flow and variceal pressure may be beneficial in
patients bleeding from esophagogastric varices.14 The
role of this drug in the control of bleeding from ectopic varices
has not been addressed. It is our routine to commence an octreotide
infusion (50 µg/h) in all patients presenting with suspected
portal hypertension-related bleeding. It is unclear whether this
intervention helps, and is only justified in light of the safety
profile of octreotide. Similarly, terlipressin, which has been
beneficial in the management of bleeding esophageal
varices15 may be tried, but this drug is not currently
available in the United States.
Because brisk upper gastrointestinal bleeding can result in
hematochezia, all subjects with suspected variceal hemorrhage
undergo emergency upper endoscopy as a first line investigation. If
the upper endoscopy fails to show a source of upper
gastrointestinal hemorrhage, colonoscopy is performed after a rapid
colonic purge, typically 3 L of polyethylene glycol solution
delivered via a nasogastric tube. In the article by Lebrec and
Benhamou,6 22% of patients with ectopic varices bled
from the colon or rectum. Colonic (Fig. 1) and rectal varices are
identified as serpiginous vessels projecting into the lumen. If the
colon appears normal and bleeding continues, angiography may be
useful in identifying varices or in localizing a nonvariceal source
of hemorrhage. Even in the face of massive bleeding, conventional
transfemoral angiography may not show the site of variceal
bleeding.
View
Larger Version
|
Fig. 1. Colonic varices as seen at colonoscopy. The bleeding
site is shown (arrow). |
Most ectopic varices are within reach of standard
endoscopy.6 Several studies have reported successful
therapy of duodenal varices with sclerosant injection.16
Our experience with sclerosant has been less than satisfactory,
especially with colonic varices. This may be because sclerosant may
be diluted in large varices beyond a concentration adequate to
obliterate the varix. Successful control of bleeding after the
injection of cyanoacrylate or thrombin/sclerosant combination has
also been reported.17 There are no data regarding band
ligation of ectopic varices. It is our opinion that if the entire
varix cannot be banded there is the risk of causing a wide defect
in the varix after sloughing of the band, rendering the banding
technique unsafe for large ectopic varices. We recommend banding
only if the diameter of the varix is less than the diameter of the
endoscope.
Extensive literature exists regarding the efficacy of
embolization therapy using radiological techniques in the
short-term therapy of bleeding esophagogastric varices. A number of
techniques have been used to embolize the varices, including steel
coils, thrombin, gelfoam, collagen, and autologous blood
clot18 either alone or in combination with a variety of
sclerosants. In the acute setting, these techniques control
bleeding in up to 94% of cases. 19,20 Unfortunately,
these techniques do not decompress the portal system, resulting in
high rebleeding rates over 1 year.18 The experience with
percutaneous embolotherapy for ectopic varices is limited to case
reports.21-23 The transhepatic approach is used to
catheterize the portal vein; the catheter is passed into the
mesenteric veins, and angiography is directly performed. The
transhepatic approach is preferred, especially in the patient with
active bleeding, because access to the portal venous system is much
quicker than when the transjugular approach is used. The ectopic
varices and their drainage into the systemic venous system are
identified (Fig. 2).
Unlike arterial embolization for gastrointestinal hemorrhage, the
goal in this setting is not to occlude the bleeding site itself,
but to occlude the feeding vein (that is, the vein on the portal
venous side) to the ectopic varices. This allows the varices to
drain into the low pressure systemic veins, while protecting them
from the high pressure portal system (Fig. 2). Steel coils are the
preferred agent, because they provide a permanent, focal occlusion.
They are available in a variety of sizes, which allows occlusion of
large veins without difficulty. The transhepatic approach to
variceal embolization may be difficult to apply in patients with
portal vein occlusion. The transjugular approach may be used if a
transjugular intrahepatic portosystemic shunt (TIPS) is planned
immediately after embolization of the varix. We restrict this
approach only to patients who are on a waiting list for liver
transplantation.
View
Larger Version

View
Larger Version

View
Larger Version
|
Fig. 2. (A) Angiographic view of the colonic varices
(arrowheads) shown in Fig. 1. The transhepatic catheter
is in the feeding mesenteric vein branch. (B) A late image from the
same series shows the large draining systemic vein. (C) Angiography
shows stasis in the feeding vessels (arrow) and disappearance of
the varices.
|
If the patient continues to bleed in spite of embolization of
the ectopic varices, options include either creating a TIPS or
proceeding with surgery. At surgery, the varix is ligated and, in
some circumstances, a portosystemic shunt is created. The choice of
TIPS or surgery depends on the underlying liver function. Surgery
is preferred in patients of Child-Pugh class A and in patients with
an extrahepatic portal vein thrombosis as the cause of portal
hypertension.
Several recent publications have emphasized the role of TIPS
in the management of bleeding ectopic varices caused by
intrahepatic portal hypertension.24-29 In the largest
series to date, nine patients with intestinal varices received
TIPS.26 Five of these patients were Child-Pugh class C
at the time that the procedure was performed. Six were actively
bleeding at the time of the procedure. Five of these patients
stopped bleeding with the TIPS and one required embolization of the
portal feeding vessel to stop bleeding. There was no procedural
mortality and no recurrence of bleeding in the follow-up period
(median, 7 months). However, two patients died within 5 days of the
procedure from multisystem failure, and three more died within 6
months. This illustrates the generally bleak prognosis in patients
of poor liver function who bleed from portal hypertension.
Furthermore, up to 50% of TIPS will stenose by 6 months. It appears
that in patients with intrahepatic portal hypertension, TIPS offers
a highly effective means of controlling hemorrhage in the short
term; if patients have good liver function and long-term survival
is expected, a surgical portosystemic shunt may be preferred. It
must be emphasized that only a nonselective portosystemic shunt,
such as portocaval, mesocaval, or central splenorenal shunt, will
adequately decompress ectopic varices. In case of extensive
thrombosis of the portal venous system, nonconventional
portosystemic shunts using large collateral vessels may be
performed.30
There are no data available regarding the use of
adrenergic -blockers or
nitrates in the long-term management of patients with ectopic
varices. A meta-analysis of randomized, controlled trials comparing
a nonselective -blockade
versus placebo showed the efficacy of -blocker therapy for secondary prophylaxis of bleeding
from esophagogastric varices.31 Therefore, it seems
logical that these modalities could be tried in patients with
ectopic varices. However, in one study, -blocker therapy did not appear to be effective in
the management of peristomal varices.9 We recommend the
use of -blocking agents
in the prevention of rebleeding or when patients cannot undergo a
surgical shunt. This is usually when portal venous thrombosis is so
extensive that a suitable vein (8 mm or more in diameter) cannot be
found for anastomosis with a systemic vein.
Intra-abdominal Varices. Veins in the abdominal wall
may rupture externally. More often, veins around the falciform
ligament of the liver rupture into the peritoneal cavity. Fewer
than 20 such cases have been reported in the literature. External
rupture is easily recognized and treated with local compression and
surgical ligation. Rupture of intra-abdominal varices is
characterized by a rapid increase in ascites, particularly in the
presence of abdominal pain, and is associated with a decrease in
the hematocrit. The diagnosis is confirmed by a heavily
blood-stained ascitic fluid tap. The diagnosis is further suggested
by computed tomography scan demonstration of retroperitoneal
varices and intra-abdominal hemorrhage.
Management of this situation is made more difficult by the
relative contraindication to transhepatic obliteration of the
varices caused by the accompanying ascites. The transjugular,
intrahepatic approach to the portal system is another possibility
for angiographic embolization therapy. In appropriate subjects,
TIPS placement at the time of variceal obliteration may be
considered, but in an acutely ill patient it should probably be
deferred unless the initial embolization therapy fails to arrest
hemorrhage. Emergency surgery may be performed in these patients
with the aim of ligating the bleeding varix, but this carries a
high mortality.32 Because such patients are usually
quite ill at the time of surgery, no attempt is made to create a
portosystemic shunt.
Stomal Varices. The most common scenario for stomal
varices is in patients with ileostomies after proctocolectomy for
inflammatory bowel disease with associated primary sclerosing
cholangitis.9 Stomal varices are recognized by the
purplish hue around the stoma. The patient is usually aware of the
bleed early in its clinical course and may be taught how to control
the hemorrhage with pressure. Probably because the patient can
control hemorrhage easily, the mortality from a stomal variceal
bleed is relatively low (3%-4%).27 Thus, a relatively
conservative approach to stomal varices is recommended.
Simple local measures, such as pressure dressings,
epinephrine-soaked gauze, gel foam, and suture ligation, have all
been used with success for the initial bleeding
episode.27 Injection sclerotherapy has also been used,
but this may result in mucosal ulceration, peristomal skin
necrosis, and stricturing of the stomal orifice.33
Mucocutaneous disconnection of the stoma with ligation of the
portosystemic channels has been performed under local anesthesia
and is a less invasive therapy than refashioning the entire stoma,
but long-term results with both procedures are usually suboptimal.
We do not recommend either of these procedures as effective
long-term treatment of stomal varices. In patients with
uncontrolled or recurrent bleeding we advocate transhepatic
obliteration of the bleeding stomal varices. In some patients this
is insufficient to prevent further bleeding and a portosystemic
procedure should be contemplated. Because many patients with stomal
varices are candidates for liver transplantation, a TIPS procedure
is preferred to a surgical portosystemic shunt.
Biliary Varices. In extrahepatic portal vein
thrombosis, varices may occur around the gallbladder in 30% of
patients.34 They are usually asymptomatic and require no
therapy. Varices can also occur in portal venous obstruction around
the common bile duct resulting in extrahepatic biliary obstruction,
cholangitis, and hemobilia. The appearances of a biliary tree on
endoscopic retrograde cholangiopancreatography may mimic that of
primary sclerosing cholangitis.35 Optimal therapy
requires both a surgical portosystemic shunt as well as a biliary
drainage procedure, which may be either surgical, endoscopic, or
radiological.
SUMMARY
The management of ectopic varices is difficult and must be
performed by a multidisciplinary team approach of hepatologists,
endoscopists, interventional radiologists, and surgeons. Although
there are no large series on this subject, based on our experience
and the literature of case reports, small series, and reviews, we
present this systematic approach for any patient presenting with
bleeding from possible ectopic varices (see Fig. 3).
View
Larger Version
|
Fig. 3. An algorithmic approach to the patient with bleeding
ectopic varices. |
Footnotes
Abbreviations: TIPS, transjugular intrahepatic portosystemic
shunt.
Address reprint requests to Patrick S. Kamath, MD, Division
of Gastroenterology and Hepatology, Mayo Clinic, 200 First Street
SW, Rochester, MN 55905. Fax: (507) 284-0538; e-mail: kamath.patrick@mayo.edu.
Received August 3, 1998; accepted August 10, 1998.
References
Copyright © 1998 by the American Association for
the Study of Liver Diseases.
|