Introduction
Acute cholecystitis is a common medical condition, conventionally treated with cholecystectomy,
either open or laparoscopic [1]. Nearly 700,000 cholecystectomies are performed annually in the United States, making
it one of the most common surgeries performed [2], however, many patients with cholecystitis may be poor operative candidates due
to comorbid conditions that increase risk for perioperative morbidity and mortality
[3].
Percutaneous transhepatic gallbladder drainage (PTGBD) has demonstrated efficacy in
temporary decompression of the gallbladder [4]
[5]
[6] PTGBD is limited by presence of severe coagulopathy, anatomically inaccessible gallbladders,
and is associated with adverse events (AEs) including catheter dislodgment, cellulitis,
pneumothorax, bleeding, and infection [7]
[8]. In addition, there is a high recurrence rate of cholecystitis if the catheter is
removed [9].
Endoscopic gallbladder drainage was developed to supplant the need for PTGBD in treatment
of cholecystitis in patients deemed unfit for surgery due to comorbidities or advanced
malignancy unable to undergo cholecystectomy [10]. Endoscopic options include endoscopic retrograde cholangiopancreatography (ERCP)
with transpapillary placement of a plastic stent into the cystic duct and endoscopic
ultrasound-directed gallbladder drainage (EUS-GBD) [11]. ERCP and EUS-GBD have demonstrated similar efficacy, however, EUS-GBD has a higher
technical success rate, particularly when the gallbladder is distended or the cystic
duct is inaccessible due to tortuosity or obstruction [12].
Cirrhosis has historically been considered a relative, if not absolute, contraindication
to laparoscopic cholecystectomy [3]. Gallstone prevalence in patients with cirrhosis ranges from 25 % to 30 %, being
at least twice that of the general population with a higher incidence in decompensated
cirrhosis, irrespective of etiology [13]. Untreated, cholecystitis with or without gallstone disease can lead to repeat admissions
for pain, or may lead to gallbladder perforation and death [14]. Sparse data exist on use of EUS-GBD in management of cholecystitis in patients
with cirrhosis. The current study aimed to describe the outcome of EUS-GBD for cholecystitis
in patients with cirrhosis.
Patients and methods
This retrospective series involved two large tertiary referral centers with two therapeutic
endoscopists (THB and JN) performing EUS-GBD. The institutional review board at each
center approved this study. All consecutive adult patients (age ≥ 18 yr) who underwent
EUS-GBD for cholecystitis between August 1, 2014 and December 31, 2018 were identified
using the electronic medical record. Patients without documented cirrhosis were excluded
from analysis. Clinical and procedural data were collected, including etiology of
cirrhosis, Child-Pugh score at time of procedure, model for end-stage liver disease
sodium (MELD-Na) score at time of procedure, indication for EUS-GBD, endoscopic data
(length and diameter of stent, anastomotic location, procedural findings), procedure-related
AEs, post-procedural symptoms, and clinical success, when available. The primary endpoint
was the rate of technical success, defined as EUS-guided placement of a lumen-apposing
metal stent (LAMS) from the gastrointestinal lumen to the gallbladder. AEs were graded
according to the American Society for Gastrointestinal Endoscopy lexicon [15]. Hepatic decompensation within 30 days following the procedure was defined according
to the European Association for the Study of the Liver guidelines [16]. Sepsis was defined according to the Society of Critical Care Medicine [17]. Clinical success was defined as relief of symptoms of cholecystitis.
Statistical analyses were performed using Stata version 15.1 (StataCorp, Texas, United
States). All continuous variables are expressed as mean ± standard deviation and categorical
variables are expressed as proportions (%). Because of a small number of events, we
did not perform a logistic regression analysis. P < 0.05 was considered statistically significant. Neither sample size nor statistical
power was calculated as this was not a comparative study.
Prior to undertaking the procedure, all patients were informed of the risks, benefits,
and alternatives of proceeding with EUS-GBD and each provided written informed consent.
General anesthesia and fluoroscopy were used in all cases. Patients with acute cholecystitis
were continued on their antibiotic therapy; in the absence of infection, pre-procedural
antibiotics were not routinely administered. The technique of EUS-GBD was performed
as follows: a standard therapeutic channel oblique linear echoendoscope (GF-UCT180,
Olympus America, Center Valley, Pennsylvania, United States) was passed into the duodenum,
stomach (3 cases) or jejunum (1 case) to visualize the gallbladder. Interposing vessels
were identified by Doppler so that they could be avoided. One of two approaches was
used. In the first, a 19G needle (Expect, Boston Scientific, Marlborough, Massachusetts,
United States) preloaded with water-soluble contrast was used to puncture the gastrointestinal
lumen into the gallbladder and entry was confirmed by contrast injection ([Fig. 1]). The needle was flushed with saline and a 0.025-inch, 450-cm hydrophilic-tipped
guidewire (VisiGlide, Olympus) was passed into the gallbladder ([Fig. 2]). The needle was withdrawn and the LAMS (AXIOS-EC, Boston Scientific, Marlborough,
Massachusetts, United States) and electrocautery device was introduced through the
working channel. Current was applied to the cautery tip and advanced into the gallbladder
over the guidewire, with the distal portion deployed into the gallbladder and proximal
into the gastrointestinal lumen ([Fig. 3]). In the latter cases and when the gallbladder was distended, the electrocautery-enhanced
LAMS was passed directly into the gallbladder without a guidewire. In all technically
successful cases performed by one endoscopist (THB) a 7Fr double pigtail plastic stent
was placed within the LAMS over a guidewire.
Fig. 1 EUS-guided needle placement into the gallbladder confirmed with contrast injection
under fluoroscopy.
Fig. 2 Guidewire passed into gallbladder in order to secure position and aid stent placement.
Fig. 3 Endoscopic view of duodenum following cholecystoduodenostomy with lumen apposing
metal stent and plastic pigtail stent placement within.
Results
A total of 15 patients (UNC:12, WVU:3, 9 females, 6 males) met inclusion criteria
and underwent EUS-GBD during the study period. Mean patient age was 61 ± 17.1 years,
mean MELD-Na 15 ± 7, and Child-Pugh classes were A (n = 3), B (n = 10) and C (n = 2)
([Table 1]). Etiology of cirrhosis was HCV (n = 2), alcohol (n = 4), non-alcoholic fatty liver
disease (n = 8), and autoimmune hepatitis (n = 1). Indications for EUS-GBD included
acute cholecystitis (n = 11), chronic cholecystitis (n = 2), and recurrent gallstone
pancreatitis (n = 2). Nine patients had undergone surgical evaluation and been deemed
to have a prohibitively high risk for perioperative morbidity and mortality. Three
patients had previously had a percutaneous drain placed and EUS-GBD was being performed
for gallbladder drain internalization. Three patients were septic at the time of the
procedure.
Table 1
Patient demographics.
|
EUS-GBD patients with cirrhosis (n = 15)
|
|
|
|
Age, mean ± SD
|
61
|
17.1
|
|
Number of females (%)
|
9
|
60
|
|
Etiology of cirrhosis
|
|
|
|
Chronic Hepatitis C virus, n (%)
|
2
|
13
|
|
Alcohol, n (%)
|
4
|
27
|
|
Non-alcoholic fatty liver disease, n (%)
|
8
|
53
|
|
Autoimmune, n (%)
|
1
|
7
|
|
Child-Pugh class
|
|
|
|
A
|
3
|
20
|
|
B
|
10
|
67
|
|
C
|
2
|
13
|
|
Ascites grade at time of procedure
|
|
|
|
0
|
5
|
33
|
|
1
|
7
|
47
|
|
2
|
3
|
20
|
|
3
|
0
|
0
|
|
INR at time of procedure
|
|
|
|
< 1.5
|
8
|
53
|
|
1.5 – 2.5
|
7
|
47
|
|
MELD-Na, mean ± SD
|
15
|
7
|
|
Indication for EUS-GBD
|
|
|
|
Acute cholecystitis, n (%)
|
11
|
73
|
|
Chronic cholecystitis, n (%)
|
2
|
13
|
|
Recurrent gallstone pancreatitis, n (%)
|
2
|
13
|
|
Surgical consultation for cholecystectomy, n (%)
|
9
|
60
|
|
Indwelling percutaneous cholecystostomy tube, n (%)
|
3
|
20
|
EUS-GBD, endoscopic ultrasound-guided gallbladder drainage; SD, standard deviation;
INR, international normalized ratio; MELD, model for end-stage liver disease.
The technical success rate for EUS-GBD was 93.3 % (n = 14) ([Table 2]). Mean procedure time was 64 ± 59 minutes and initial puncture site was duodenum
(n = 11), stomach (n = 3) and jejunum (n = 1) and portion of gallbladder used for
drainage was neck (n = 4) and body (n = 11). Mean gallbladder wall thickness was 7 ± 3.6 mm
and the gallbladder was distended in all cases except one (93.3 %). EUS-GBD was performed
using a LAMS in 14 patients and a non-foreshortening self-expandable metal stent (Viabil,
W.L. Gore, Flagstaff, Arizona, United States) in one. Stent lengths were 10 mm (n = 14)
and 40 mm (n = 1) with diameters of 10 mm (n = 12) or 15 mm (n = 3). In one case,
a second overlapping stent of 40 mm × 10 mm was required to bridge the distance between
gallbladder and gastrointestinal lumen. In nine cases (60 %), a plastic pigtail stent
was placed within the metal stent to decrease risk for stent occlusion; plastic stents
were lengths of 3 cm (n = 1), 4 cm (n = 7) or 7 cm (n = 1) with diameters of 7 French
(n = 8) or 10 French (n = 1). In the two patients with recurrent gallstone pancreatitis,
one patient had a pre-procedural magnetic resonance cholangiopancreatography demonstrating
no stones in the common bile duct and the other patient had an ERCP with removal of
choledocholithiasis prior to EUS-GBD.
Table 2
Procedural and outcomes data.
|
EUS-GBD in patients with cirrhosis (n = 15)
|
|
|
|
Technical success, n (%)
|
14
|
93.3
|
|
Procedure time, mean ± SD, minutes
|
64
|
59
|
|
Initial puncture site
|
|
|
|
Duodenum
|
11
|
|
|
Stomach
|
3
|
|
|
Jejunum
|
1
|
|
|
Portion of gallbladder used
|
|
|
|
Neck
|
4
|
|
|
Body
|
11
|
|
|
Gallbladder wall thickness, mean ± SD, mm
|
7
|
3.6
|
|
Stent diameters
|
|
|
|
10 mm
|
12
|
|
|
15 mm
|
3
|
|
|
Coaxial plastic stent placement, n (%)
|
9
|
60
|
|
Clinical outcome
|
|
|
|
Outpatient cases, n (%)
|
3
|
20
|
|
Length of hospital stay post- procedure, mean ± SD, days
|
2.9
|
3.9
|
|
Adverse events (severity grade)
|
|
|
|
Postprocedural pancreatitis, mild
|
1
|
6.7
|
|
Stent misdeploy requiring surgery, severe
|
1
|
6.7
|
|
Patients with hepatic decompensation following EUS-GBD, n (%)
|
1
|
6.7
|
|
Eventual EUS-GBD stent exchange with plastic stent, n (%)
|
7
|
46.6
|
|
Length of time stent left in place prior to exchange, mean ± SD, days
|
179
|
230
|
|
Patients who went on to orthotopic liver transplant, n (%)
|
1
|
6.7
|
|
Length of follow-up, median (IQR)
|
174
|
58, 674
|
|
Patients with clinical success, n (%)
|
14
|
93.0
|
|
Deaths during follow-up period, n (%)
|
1
|
6.7
|
EUS-GBD, endoscopic ultrasound-guided gallbladder drainage; SD; standard deviation,
IQR, interquartile range.
The EUS-GBD stent was endoscopically removed in seven patients (46.6 %) after a mean
of 179 ± 230 days; cholecystitis did not recur in any of these patients. No patients
later underwent cholecystectomy, as their operative candidacy did not improve. One
patient went on to receive an orthotopic liver transplant; [18]
[19] operative documentation noted increased surgical complexity due to the dense scarring
of the porta hepatis with the duodenum adherent to the gallbladder fossa, making dissection
extremely challenging due to deformed anatomy and scar tissue. The duodenotomy created
by the EUS-GBD was oversewn with gastric decompression via nasogastric tube and total
parenteral nutrition. Eight patients did not have their stents removed and EUS-GBD
served as destination therapy. Fourteen patients (93.3 %) went on to achieve clinical
success, with relief of symptoms of cholecystitis.
Two AEs occurred in this cohort (13.3 %). One patient developed pancreatitis 1 week
after stent placement and was managed supportively; this was considered a mild AE.
In a second patient, the LAMS was maldeployed with the distal end in the gallbladder
and the proximal end between the gastric wall and gallbladder; this was managed surgically
with an ultimate favorable clinical outcome, but was considered a severe AE. One patient
with Child-Pugh class C cirrhosis was admitted 3 weeks after the procedure for hepatic
decompensation including gastrointestinal hemorrhage, hepatic encephalopathy, and
worsening ascites. Mean length of follow-up post-procedure was 373 ± 367.3 days. One
death (6.6 %) occurred due to underlying malignancy in a patient with stage IV pancreatic
ductal adenocarcinoma; this occurred 37 days following the procedure after the patient
was on home hospice care.
Discussion
Patients with cirrhosis and symptomatic gallbladder disease offer unique challenges
and require a tailored approach to therapy. This group of patients is more likely
to develop surgical complications including bleeding due to impaired coagulation,
infection due to compromised immune function, and poor wound healing due to malnutrition
and/or ascites [20]. Cholecystectomy is a particularly high-risk surgery in this patient population,
particular in patients with Child-Pugh Class C cirrhosis, hypoalbuminemia, and portal
hypertension [21] Management of cirrhotic patients with acute or chronic cholecystitis has largely
employed percutaneous drainage for gallbladder decompression. For most patients in
this demographic, the drain is often destination therapy, without plans for cholecystectomy
at a later date as their operative candidacy does not improve over time. Percutaneous
gallbladder drains carry a risk of dislodgement, which may be increased by ascites
or hepatic encephalopathy. Another option for management of cholecystitis in cirrhotic
patients is transpapillary cystic duct placement via ERCP [22]. This procedure may be challenging due to acute angulation in the cystic duct or
obstruction from cholecystolithiasis. In addition, long-term drainage is a concern
due to the narrow diameter of the plastic stent used. A third nonsurgical option for
management of gallbladder disease has arisen in recent years in the form of EUS-GBD,
however, little data exist on its use in patients with cirrhosis.
The current study was a multicenter, retrospective analysis of consecutive patients
with cirrhosis who underwent EUS-GBD for management of non-operative gallbladder disease.
This was a heterogeneous group with a wide age range (33 – 91 years) and diverse underlying
etiologies for liver disease. The MELD-Na score was a mean of 15, which correlates
in most studies with an estimated 3-month mortality of 6.0 % [23]. The mortality rate in our cohort (6.6 %) approximates this estimate with one patient
death during follow up; however, the cause of death was not directly related to cirrhosis,
but rather, metastatic pancreatic cancer.
In our cohort, one patient with cirrhosis due to autoimmune hepatitis underwent EUS-GBD
for management of acute cholecystitis and later went on to receive an orthotopic liver
transplant. On review of pre-procedural imaging, ERCP with transpapillary cystic duct
stenting was not felt to be possible and PTGBD was limited by hepatic encephalopathy
with concern for high risk of drain dislodgement. Though EUS-GBD was felt to be the
best option at the time of the patient’s critical illness from acute cholecystitis,
we feel that in the majority of cases, EUS-GBD should not be used in patients being
considered for liver transplantation. The added surgical complexity from gallbladder
scarring and duodenotomy may prohibit the patient from undergoing the life-saving
operation, an outcome that no endoscopist would wish for their patient if it can be
prevented. The decision to perform EUS-GBD should be made in a multidisciplinary fashion,
with agreement from the transplant surgery team when applicable.
Long-term placement of an expandable metal stent into the gallbladder is not advisable
as there are concerns for delayed bleeding and perforation. In addition, removal after
a prolonged indwelling period may not be possible due to degradation of the stent
covering; this could hypothetically lead to obstruction from tissue hyperplasia. However,
there is currently no agreed upon duration to leave an expandable metal stent in place
following EUS-GBD. Our typical approach to stent exchange is similar to the strategy
employed in walled-off pancreatic necrosis drainage. In most cases of EUS-GBD, we
electively replace the stent after tract maturation (2 – 4 weeks) with plastic pigtail
stents to allow for continuous drainage without concern for recurrent cholecystitis.
This plastic stent can be placed in most circumstances with a forward-viewing upper
endoscope and provides an opportunity to assess the fistulous tract and extract accessible
gallstones if present.
Limitations of the current study include retrospective design and relatively limited
sample size. In addition, few patients with advanced cirrhosis, as determined by Child-Pugh
Class, were included in this study. This is likely due to patient selection factors
including increased risk for post-procedural bleeding, technical complexity of stent
placement with comorbid ascites, and risk for infection due to impaired immune function.
Future studies should be performed in a prospective fashion, ideally in a randomized
controlled trial against either PTGBD or ERCP with transpapillary cystic duct stent
placement.
Conclusion
In summary, the current study shows the feasibility of using EUS-GBD in patients with
cirrhosis and gallbladder disease who are deemed nonsurgical candidates due to perioperative
risk. This approach, using a LAMS, is safe and efficacious in management of gallbladder
disease in this patient population. We believe such an approach will be preferred
over percutaneous therapy, however, further studies are needed to determine optimal
patient selection and procedural technique. As the growing burden of non-alcoholic
fatty liver disease leads to an increasing patient population with comorbid cirrhosis
and gallstone disease, further innovation is needed to effectively care for these
patients.
