Keywords
cyanoacrylate glue - gastrointestinal bleeding - hemostasis - duodenal varices - portal
hypertension
Introduction
Ectopic varices are large portosystemic venous collaterals situated in locations other
than the gastroesophageal region and account for up to 1 to 5% of all variceal bleeding.[1]
[2] Although their exact prevalence is not known, duodenal varices (DVs) are estimated
to account for 17 to 40% of all ectopic varices with a higher prevalence noted in
patients with extrahepatic portal hypertension (EPH) than those with intrahepatic
portal hypertension (IPH).[3]
[4]
[5]
[6] Patients with DVs may present with hematemesis or massive lower gastrointestinal
(GI) bleeding and should be considered in all patients with portal hypertension who
present with GI bleeding particularly when both upper and lower GI endoscopies fail
to identify the source of bleeding. Furthermore, management of DVs involves a multidisciplinary
approach and includes medical, endoscopic, interventional radiological, and surgical
modalities depending on patient's condition, the acuteness of bleeding, site of varices,
and available clinical expertise. Endoscopic intervention is the initial therapeutic
modality for the management of acute bleeding DVs as it is less invasive compared
with interventional radiological and surgical procedures; however, given their rarity
and sporadic nature, literature is limited to case reports, small case series, and
reviews, with no well-established treatment guidelines regarding their management.
Hence, we undertook this study to report our experience in the management of DVs at
our center.
Materials and Methods
We performed a retrospective analysis of all patients with portal hypertension who
underwent endoscopy at St. John's Medical College Hospital, Bengaluru, between January
2010 and December 2021. Inclusion criteria included patients who were diagnosed to
have bleeding primarily from DVs as well as those in whom DVs were diagnosed routinely
on endoscopy performed as part of evaluation of portal hypertension. Medical records
were reviewed from our hospital database with regard to demographic data of patients,
clinical features, endoscopic and imaging findings of DVs, and endoscopic treatment
of bleeding DVs. All patients presenting with suspected portal hypertension-related
bleeding were admitted, and after initial resuscitation underwent esophagogastroduodenoscopy.
Resuscitation included stabilization with intravenous fluids and/or administration
of packed red cell transfusion, intravenous infusion of terlipressin (in the initial
years 0.5–1 mg 6th hourly, followed by continuous infusion 2 mg/24 hours in the later
years) and ceftriaxone injection as per guidelines.[7] After endoscopic confirmation, patients with DV bleeding (n = 21) received endoscopic injection of n-butyl-2-cyanoacrylate 1 mL undiluted injection
followed by 2 mL normal saline flush, while patients with nonbleeding DVs (n = 20) were initiated on beta-blocker prophylaxis and followed at regular intervals.
An informed consent was obtained from all patients before undertaking endoscopy. Patients
were closely followed up for rebleeding by a combination of history and laboratory
check (for drop in hemoglobin). Given the retrospective nature of the study, Institutional
Ethics Review Board approval was deemed not mandatory. Although the prophylactic role
of beta-blockers in patients with DVs remains debatable, all our patients with DVs
were initiated on nonselective beta-blocker (propranolol) and followed up at regular
intervals.
Results
Overall, 5,892 patients underwent endoscopy for portal hypertension during the study
period and we identified 41 patients with DVs, 29 of whom were males and the mean
age (± SD [standard deviation]) of the study population was 38.1 (± 13.7) (range:
5–75 years). DVs accounted for 0.42% (21/4889) of all portal hypertensive-related
bleedings. Extrahepatic portal vein obstruction (EHPVO) was the most common etiology
of DVs noted in 26 patients (63%), followed by cirrhosis in 14 patients (34%), and
noncirrhotic portal fibrosis (NCPF) in 1 patient (3%). Of the total 41 DVs, 21 were
in the duodenal bulb, while 17 were in the second part of duodenum and three in the
third part of duodenum ([Table 1]).
Table 1
Clinical and endoscopic characteristic of study population
|
Total number of patients
|
n = 41
|
|
Male:female
|
29:12
|
|
Mean age (±standard deviation) (in years)
|
38.146 ± 13.739
|
|
Etiology of duodenal varices
|
|
Extrahepatic portal vein obstruction[a]
|
26
|
|
Cirrhosis
|
14
|
|
a. Alcohol
|
7
|
|
b. Chronic hepatitis B
|
4
|
|
c. Nonalcoholic steatohepatitis
|
3
|
|
Noncirrhotic portal fibrosis
|
1
|
|
History of previous surgery/abdominal procedure
|
n = 3
|
|
Frey's procedure for chronic pain secondary to chronic calcific pancreatitis
|
1
|
|
Splenectomy and devascularization surgery
|
1
|
|
Kasai's procedure for extrahepatic biliary atresia
|
1
|
|
Location of duodenal varices
|
|
Duodenal bulb
|
21
|
|
Second part of duodenum (D2)
|
17
|
|
Third part of duodenum (D3)
|
3
|
|
Presence of coexistent esophagogastric varices
|
|
Esophageal varices
|
13
|
|
Gastric varices
|
2
|
|
Both esophageal and gastric varices
|
14
|
|
Endoscopic treatment for duodenal varices (cyanoacrylate glue)
|
n = 21
|
|
Control of index bleeding after endotherapy
|
21 (100%)
|
|
Patients with rebleeding on follow-up
|
4/21 (19%)
|
|
Success of endotherapy in repeat bleedings
|
¾ (75%)
|
a Including three patients with chronic calcific pancreatitis.
Overall, 21 patients presented primarily with upper gastrointestinal (UGI) bleeding
from DVs, while in 20 other patients DVs were detected on endoscopy performed for
evaluation of portal hypertension. Endoscopic cyanoacrylate glue injection successfully
achieved hemostasis of the index bleeding in all 21 patients with DV bleeding ([Fig. 1]). Four of these patients, however, had recurrent DV bleeding between 2 weeks and
12 months of follow-up necessitating repeat sessions of cyanoacrylate glue injections.
Following repeat endotherapeutic sessions, three patients remained free of further
bleedings, while one patient (female, aged 25 years with underlying EHPVO) continued
to present with recurrent massive bleeds, episodes, not amenable to endotherapy and
ultimately underwent duodenal resection and surgical shunt procedure. A mean of 1.25
endoscopic sessions was required for near eradication of the DVs (range: 1–4).
Fig. 1 (A) i. Endoscopy showing active bleeding from the duodenal varices. ii. Repeat endoscopy
after 2 weeks of glue injection in the same patient. (B) Endoscopy showing duodenal varices in second part of duodenum with stigmata of recent
bleeding. (C) Large conglomerate of duodenal varices in third part of duodenum.
Discussion
DVs, although rare, can manifest as severe, life-threatening massive UGI bleeding
in patients with portal hypertension, and they usually coexist with esophagogastric
varices.[8]
[9]
[10]
[11]
[12] In our series, we identified a total of 41 patients with DVs of whom 21 presented
primarily with DV bleeding. In the remainder (n = 20), DVs were diagnosed during routine endoscopic evaluation of portal hypertension.
Our series in a fairly large number of patients demonstrates that endoscopic cyanoacrylate
glue injection of DVs achieves complete and durable hemostasis in an overwhelmingly
large number of patients including a small proportion of patients presenting with
recurrent DV bleedings. Only one patient with recurrent episodes of massive DV bleeding
was subjected to surgery after four sessions of endoscopic glue injection ([Fig. 2]). Of note, a total of three patients in our study had history of previous abdominal
surgeries ([Table 1]).
Fig. 2 Flowchart depicting algorithm of our study.
In a study that enrolled a total of 169 patients with ectopic variceal bleeding, 17%
of the bleeding was attributed to DVs.[3] This study also found that DVs were common in patients with EPH. Another study involving
173 patients reported that the mean age of patients with ectopic varices was 62.3
years and that 32.9% of patients with ectopic varices had DVs.[4] DVs are more prevalent in patients with portal hypertension secondary to portal
or splenic vein thrombosis than in patients with IPH.[13]
[14] The mean (±SD) age of our study population was 38.14 (±13.73) years that is quite
younger than that noted in the above studies. This finding is because the most common
underlying etiology of DVs in our study was EHPVO that was noted in 26 patients (63%),
followed by cirrhosis in 14 patients (34%), and NCPF in one patient (3%). Given the
regional variation in the causation of portal hypertension, there is considerable
heterogeneity in the etiology and prevalence of DVs. In the west, most bleeding from
ectopic varices is usually associated with IPH as the prevalence of EPH is low.[13]
[15] EHPVO, on the other hand, is relatively common in the Indian subcontinent that is
reflected in our study and these patients tend to be younger than those with portal
hypertension due to other causes.[16] Another study that enrolled 25 patients treated with endoscopic glue injection for
duodenal variceal bleeding showed that the most common cause of duodenal varix was
EHPVO in 19 (76%) followed by chronic liver disease in 5 (20.0%) patients.[17] This study also showed that five (20.0%) patients had rebleeding after initial endotherapy,
and no death was observed during the follow-up that is in concurrence with our own
study. The high rate of rebleeding (19.04%) in our study could be related to the severity
of underlying liver disease as patients with advanced cirrhosis are more likely to
present with rebleeds.[18]
Western data have shown that ectopic varices are more frequently noted in the duodenal
bulb, while a Japanese study reported that the most common site of occurrence of DV
was the descending duodenum.[4]
[19] In our own study, DVs were most common in the duodenal bulb and were seen in 21
patients, followed by second part of duodenum in 17 patients, while 3 patients had
varices in the third part of duodenum. Therefore, it is advocated that endoscopists
attempt to reach as far as possible into the duodenum as most of the DVs are within
the reach of a standard endoscope.[13] Approximately 25% of patients with DVs also have associated esophagogastric varices.
While 31.7% of our patients with DVs had associated esophageal varices, 4.88% patients
had associated gastric varices, and 34.14% had combined esophagogastric varices.
Varices of the small bowel and colon are commonly seen in patients with IPH who have
previously undergone abdominal surgery. We had three such patients with previous history
of abdominal surgeries ([Table 1]). Although the exact cause for this is not known, it is postulated that the postoperative
inflammation impedes hepatic blood flow and in turn causes portal hypertension.[20]
The role of endoscopic therapy in nonbleeding DV is not clear; however, endoscopic
injection of glue has been shown to be safe and effective therapeutic modality for
those with bleeding DV but complications such as rebleeding or thromboembolic complications
have been reported with its use.[10]
[21] A randomized controlled trial supports the use of endoscopic cyanoacrylate glue
in achieving rapid hemostasis in patients with isolated gastric varices including
DVs.[22] Apart from four patients presenting with rebleeding on follow-up, none of our patients
had documented thromboembolic events following cyanoacrylate injection. Available
literature supports the use of secondary interventional procedures like transjugular
intrahepatic portosystemic shunting (TIPS), balloon-occluded retrograde transvenous
obliteration band ligation (BRTO) with surgical procedures such as duodenectomy, and
gastroduodenectomy if all other treatment measures have failed in patient with recurrent
DV bleeding. Limited data from case reports and case series has shown that (TIPS)
and BRTO effectively achieved hemostasis in patients with bleeding DVs; however, apart
from being more invasive, these modalities are often associated with a high risk of
precipitating hepatic encephalopathy, notably with TIPS.[19]
[23]
[24]
[25]
[26] In addition, both modalities remain limited to hospitals with technical expertise
in radiological intervention. Previous studies have demonstrated that surgical procedures
are associated with a high rate of rebleeding and mortality for bleeding ectopic varices,
particularly in patients with cirrhosis and portal hypertension.[2]
[25] Medication alone has limited efficacy in cases of duodenal variceal bleeding.[23]
Endoscopic banding has been recommended if the diameter of the varix is less than
the diameter of the endoscope, and some have reported successful application of bands
for ligation of DV, while others have questioned the usage of band ligation based
on animal studies, which suggest a higher likelihood of perforation.[3]
[27]
[28]
[29] A case of fatal rebleeding after band ligation of DV has also been reported.[30] We achieved fair amount of success with glue injection and hence had no opportunity
to fall back on banding. Endoscopic glue injection is fraught with complications such
as pulmonary and systemic arterial embolization, portal and mesenteric vein thrombosis,
sepsis, and variceal ulceration due to extrusion of glue.[31] Apart from transient abdominal discomfort noted in five patients post-endoscopic
glue injection, no major adverse events or mortality was reported in our study.
Our study has limitations, one of which is the retrospective nature of the study.
As the treatment of DVs was successfully achieved with endoscopic cyanoacrylate glue
injection, we have no experience with other modalities of therapy to compare with.
Our study strength is the fairly large number of patients with DVs who were successfully
managed with glue injection. This modality can easily be applied in resource-limited
settings.
Conclusion
Bleeding DVs may be the first manifestation of underlying portal hypertension, and
careful endoscopic screening of the duodenum should be performed in all patients presenting
with UGI bleeding to look for DVs. Our experience suggests that early diagnosis with
prompt endoscopic intervention with cyanoacrylate glue injection of the DVs is a safe
and effective treatment modality when performed by experienced endoscopists in the
acute setting.
