Introduction
The severity of acute pancreatitis is classified as mild, moderate or severe depending
on the presence or absence of organ failure, and local or systemic complications.
Local complications of acute pancreatitis include acute peri-pancreatic fluid collection
(APFC), acute necrotic collection (ANC), pancreatic pseudocyst (PPC) and walled-off
necrosis (WON) [1].
WON, whether sterile or infected, is characterized by a well-defined capsule around
necrotic pancreatic material, usually occurring 4 weeks after initial insult. It contains
varying amounts of solid debris. Treatment of WON is indicated when it becomes symptomatic,
i. e. abdominal pain, infection, gastric outlet obstruction, or biliary compression
[2].
Management of WON is drainage, which can be done via endoscopic, surgical or percutaneous
methods. However, endoscopic drainage, especially with endoscopic ultrasound (EUS),
is the preferred modality [3]
[4]
[5].
Initially, plastic stents were used for EUS guided drainage of WON; however, the solid
component within WON frequently obstructed them, leading to insufficient drainage
[6]. This led to the advent of fully covered self-expandable metal stents (FCSEMS),
which were used to drain WON with encouraging results as these had a wide caliber;
however, an increased risk of spontaneous stent migration was present because of their
longer length and the absence of lumen-apposing properties [7]
[8].
Short and biflanged dedicated FCSEMSs such as NAGI (Taewoong Medical, Gyeonggi-do,
South Korea), and AXIOS (Boston Scientific Corporation, United States) stents were
subsequently used to drain WON with good results as the number of interventions required
to resolve WON were significantly lower with these stents compared to conventional
long FCSEMS or plastic stents [9]
[10]
[11]. Over time, these short FCSEMS became popular.
Traditionally, after a dedicated FCSEMS insertion, endoscopic intracavitary debridement
(direct passage of the endoscope into the cavity with necrosectomy) is used, and is
associated with a high rate of complications [12]. Whether these complications can be avoided by adopting a better thought out and
more acceptable method of endoscopic necrosectomy, is an issue for research [2]. Intracavitary flushing using a water pump has been described in the literature
with good results [13].
In the present study, we evaluated the safety and efficacy of EUS guided WON drainage
by FCSEMS placement accompanied by intracavitary lavage using a solution containing
hydrogen peroxide and adopting a step-up approach of endoscopic necrosectomy.
Methodology
The current study is a prospective open label study, conducted at a single tertiary
center in North India between January 2014 and January 2016 after approval from the
institute’s ethics committee. A total of 64 patients were finally included in the
study.
The inclusion criteria were all patients > 18 years with symptomatic WON. EUS guided
drainage was performed after obtaining informed consent. Exclusion criteria included
age < 18 years, pregnancy and coagulopathy (international normalized ratio [INR] > 1.5,
platelets < 50 000). Anticoagulant and antiplatelet agents were withheld for 5 days
before the procedure (in six patients).
Primary end points were completion of drainage with improvement in symptoms or major
adverse events. Secondary end points were minor adverse events related to procedures.
Initial procedure details
All EUS-guided drainage procedures (Step 1) were performed in an inpatient setting.
All procedures were performed with the patient in the left lateral position under
propofol sedation administered by an anesthetist. All patients received prophylactic
antibiotic (1 g cefoperazone and 1 g sulbactam, iv), 2 hours before beginning the
procedure.
A 3.7-mm channel, linear therapeutic echo-endoscope (GFUCT180; Olympus, Tokyo, Japan)
was used to access the WON via the stomach or duodenal wall under fluoroscopic guidance.
CO2 was used during the procedure in all of the patients. The WON was assessed for size,
wall thickness, interposing collaterals, and percentage of solid debris. The WON cavity
was punctured with a 19-gauge EUS needle (Expect Slimline by Boston Scientific, Echotip
Ultra by Cook), the fluid sample was aspirated and sent for microbial analysis (Gram
stain, culture, and sensitivity). Then a 0.035-inch stiff guidewire (Jagwire, Straight
tip, Boston Scientific) was inserted through the needle into the cavity, and coiling
into the cavity was confirmed with fluoroscopy. Over the wire, a 6 F cystotome (Endo-flex
GmbH, Dusseldorf, Germany) was used to dilate the puncture site and this was further
dilated with a wire guided 6 mm balloon dilator (Hurricane; Boston Scientific). Then,
the stent delivery catheter was advanced over the guidewire across the WON wall, and
a dedicated FCSEMS (Nagi; Taewoong Medical, Gyeonggi-do, South Korea) was deployed
under endosonographic and fluoroscopic visualization. The Nagi stent used for drainage
was short in length (20 or 30 mm) with 16 mm caliber.
Post-procedure follow-up and antibiotic protocol
All patients received prophylactic antibiotic (1 g cefoperazone and 1 g sulbactam,
iv), 2 hours before beginning the procedure and this was continued at 12-hour intervals
at the same dose for 48 hours. Patients were assessed after 48 hours. If patients
were asymptomatic, they were discharged on oral antibiotics (cefuroxime 500 mg tablet,
12 hourly) for 5 days. For patients who had not responded after 48 hours (rising total
leukocyte count (TLC), increase in temperature or heart rate, development of organ
dysfunction), antibiotic treatment was upgraded according to the fluid culture and
antimicrobial sensitivity reports, and a repeat session of necrosectomy was performed.
Endoscopic necrosectomy (Step 2) was performed by lavage with 1 L of a solution containing
3 % hydrogen peroxide (1:20 dilution). Irrigation was done with the aid of a 9.9 mm
diameter gastroscope (GIFH180; Olympus, Tokyo, Japan), which was advanced into the
cavity through the FCSEMS and irrigation was performed using a water pump and subsequent
lavage carried out. All of the liquid/debris was suctioned from the cyst cavity as
well as from the stomach. Necrotic material obstructing the stent lumen, leading to
hampered drainage, was removed by a snare or Roth net forceps under fluoroscopy. No
other debris was removed mechanically ([Video 1]).
Video 1 Endoscopic ultrasound-guided pancreatic necrosectomy using a step-up approach for
walled-off pancreatic necrosis.
For patients who were symptomatic even after two sessions of lavage (96 hours of FCSEMS
placement), a non-contrast CT scan of the abdomen was performed to examine the adequacy
of the drainage. In the case of persistence of collection, either repeat necrosectomy
or percutaneous drainage (PCD) was performed. In patients in whom there was a separate
collection or if it was not amenable to endoscopy, PCD was inserted (Step 3). In patients
in whom there was endoscopic amenable collection, a repeat session of necrosectomy
was carried out. Repeat endoscopic lavages of the WON were performed every 48 hours
( [Fig. 1]).
Fig. 1 Flow diagram showing the study design. FCSEMS, fully covered self-expanding metal
stent; PCD, percutaneous drainage; WOPN, walled-off pancreatic necrosis.
All adverse events such as bleeding, whether life threatening or minor, and stent
migration were noted. Massive bleeding can be defined as follows: (i) blood loss exceeding
circulating blood volume within a 24-hour period, (ii) blood loss of 50 % of circulating
blood volume within a 3-hour period, (iii) blood loss exceeding 150 mL/min, or (iv)
blood loss that necessitates plasma and platelet transfusion. Torrential bleeding
was defined when there was massive bleeding along with rapidly developing hypotension.
Follow-up visits
Follow-up endoscopy for stent removal was performed 12 weeks after initial drainage
after documentation of resolution of WON on non-contrast computed tomography (CT).
During removal of FCSEMS, a pancreatogram was obtained under fluoroscopy. If the pancreatogram
showed disruption of the pancreatic duct in the head or body region, it was bridged
using a plastic stent. If the pancreatic duct appeared disconnected on the pancreatogram,
no further action was taken. These patients were prospectively followed over a period
of 6 months.
Results
Our study was a prospective open label study, conducted between January 2014 and January
2016. In total, 78 patients with symptomatic walled-off pancreatic necrosis (WOPN)
in the form of abdominal pain, fever, and gastric outlet obstruction leading to failure
to thrive, who were amenable to endoscopic drainage, were allocated to EUS guided
drainage by FCSEMS. In total, 14 patients were excluded from the study (10 patients
did not provide consent, 3 patients had coagulopathy, and 1 patient was pregnant (26
weeks)), so 64 patients (mean age 36 years, range 30 – 56; 52 male and 12 female)
were studied ( [Table1]).
Tab. 1
Demographic profile of the patients and adverse events.
|
Age (range), years
|
36 (30 – 56)
|
|
Sex, M/F
|
52:12
|
|
Etiology – Alcoholic/Gallstone/Hypercalcemia/Idiopathic, n
|
30:26:2:6
|
|
Fever and pain in the abdomen/Recurrent vomiting/Pain in the abdomen, n
|
40:16:8
|
|
Duration of symptoms (range), days
|
36 (30 – 96)
|
|
Initial diameter of WON (range), cm
|
12.5 (9 – 18.5)
|
|
Technical success, n (%)
|
64 (100)
|
|
Clinical success, n (%)
|
58 (90.6)
|
|
Primary end point, n
|
|
|
18
|
|
|
40
|
|
|
5
|
|
Adverse events
|
8
|
|
|
Bleeding, 3
|
|
|
Bleeding, 2; migration, 3
|
|
Follow-up, months
|
6
|
FCSEMS, fully covered self-expanding metal stent; PCD, percutaneous drainage; WON,
walled-off necrosis.
The etiology of the pancreatitis leading to WON was alcohol in 46.9 % (n = 30), gallstones
in 40.6 % (n = 26); hypercalcemia in 3.1 % (n = 2), and idiopathic in 9.4 % (n = 6).
The symptoms that required drainage were fever and abdominal pain in 62.5 % (n = 40),
recurrent vomiting and failure to thrive in 25 % (n = 16), and abdominal pain alone
in 12.5 % (n = 8). The mean duration of symptoms from the initial event was 36 days
(range 30 – 96 days). Mean initial WON diameter was 12.5 cm (range 9.0 – 18.5 cm)
on CT and/or EUS.
Technical success was achieved in 100 % of the patients and clinical success was achieved
in 90.6 % of patients (n = 58). The primary end point was achieved with FCSEMS alone
in 28.1 % (n = 18), FCSEMS with necrosectomy in 62.5 % (n = 40), and FCSEMS with PCD
in 7.8 % (n = 5) of patients. In total, 1.6 % (n = 1) of patients required salvage
surgery. On subgroup analysis, it was found that patients with more than 40 % solid
debris on EUS were likely to need necrosectomy compared to FCSEMS alone (32 (80 %)
versus 2 (5 %)).
Mean necrosectomy sessions required were 3.2 (range 1 – 5). Four patients required
1 session, 26 patients 2 sessions, 6 patients 3 sessions, 3 patients 4 sessions, and
1 patient required 5 sessions.
Growth of bacteria in aspirated fluid was detected in 37.5 % (n = 24). Sixteen patients
showed growth of Escherichia coli and 8 showed Klebsiella pneumoniae ssp. pneumoniae. Antibiotics were upgraded according to the antimicrobial sensitivity test.
Average hospital stay was 8.3 days.
Adverse events
Adverse events were encountered in 8 (12.5 %) patients. Life threatening bleeding
was the major event encountered in 4.7 % of patients (n = 3). One patient developed
torrential bleeding associated with hypotension 2 months after FCSEMS insertion. On
CT angiography of the abdomen, there was a pseudoaneurysm of the splenic artery ( [Fig. 2]) close to the FCSEMS, secondary to impinging and this required salvage surgery,
i. e. exploratory laparotomy and ligation of the splenic artery pseudoaneurysm with
splenectomy, and the patient survived. He required 8 units of packed red blood cells
(PRBC) and 6 units of plasma along with inotropes. Two patients developed massive
bleeding (2 months, 2.5 months) after the last session of necrosectomy secondary to
pseudoaneurysm of the splenic artery close to the FCSEMS, and which was managed successfully
with radiological intervention. One patient (bleeding at 2 months) required 4 units
of PRBC while another patient required a total of 5 units of PRBC with 2 units of
plasma. There was also impingement of the splenic artery in both of these patients
secondary to NAGI stent emplacement.
Fig. 2 Splenic artery pseudoaneurysm with FCSEMS on CT angiography of the abdomen.
Minor adverse events were non-life threatening bleeding in 3.1 % of patients (n = 2)
and stent migration in 4.7 % (n = 3). One patient had stent migration to the rectum
( [Fig. 3]) 2 months after the initial procedure, and it was spontaneously expelled. Another
patient had stent migration to the WON cavity ( [Fig.4 a,b]) after 2 weeks and this was withdrawn and reapplied. One patient developed recurrent
vomiting and failure to thrive after 6 weeks. Non-contrast computed tomography (NCCT)
of the abdomen ( [Fig. 5]) revealed a collapsed cavity, one flange in the cavity and the other flange was
obstructing the lumen of the duodenum leading to gastric outlet obstruction (GOO).
This stent was removed and the patient became asymptomatic. There were no adverse
events related to hydrogen peroxide, i. e. bloating, nausea or bad taste in the mouth.
Fig. 3 FCSEMS migration to rectum on CT of the abdomen.
Fig. 4 a Fluoroscopic image of migrated FCSEMS in WON cavity. b Endoscopic image of migrated FCSEMS in WON cavity.
Fig. 5 FCSEMS causing gastric outlet obstruction on CT of the abdomen.
Follow-up
Follow-up endoscopy for stent removal was performed 12 weeks after initial drainage
after documentation of resolution of WON on NCCT. During removal of the FCSEMS, a
pancreatogram was obtained under fluoroscopy.
In 4.7 % (n = 3) of patients, plastic stenting of the pancreatic duct was carried
out for disruption of the pancreatic duct in the head and body region. In the remaining
95.3 % (n = 61) of patients, the pancreatic duct appeared disconnected on the pancreatogram,
and no other action was taken.
The mean stenting period was 72 days (range 62 – 106 days).
Patients were followed up for 6 months after removal of the FCSEMS both clinically
as well as with trans-abdominal ultrasound. All of the patients were asymptomatic
and there were no recurrences of WON.
Discussion
The present study showed that, in symptomatic WON, endoscopic lavage of a cavity using
a large volume of saline solution with hydrogen peroxide and without any debridement
is a safe and effective method with a high technical and clinical success rate.
Endoscopic approaches to manage WON have progressed from making small transmural tracts
for irrigation to making large tracts that allow the endoscope to move directly into
cavities and perform endoscopic necrosectomy, reducing the need for adjuvant PCD [14].
The technical success rate of FCSEMS in WON reported in the literature ranges from
91 % to 100 % and technical failure encountered was related to the technique or device
used [9]
[10]
[11]
[15]. In our study, FCSEMS were successfully deployed in 100 % of the patients. Successful
resolution of WON using FCSEMS has been reported in 75 % to 90 % of patients [10]
[11]
[16]. In our study, clinical success was achieved in 90.8 %, and we could achieve incremental
success with each step using the step-up approach.
The clinical success after the index drainage procedure was 28 %. The incremental
success added with each subsequent step was 72 % for endoscopic necrosectomy by saline
irrigation with hydrogen peroxide, and 8 % for PCD insertion. One patient required
salvage surgery.
Reported adverse events in pancreatic fluid collection (PFC) drainage using metal
stents range from 5 % to 10 % [16]
[17]. The rates of adverse events with dedicated FCSEMS were 5.4 % as reported by Rinninella
et al. [16] and 9.8 % by Siddiqui et al. [17]. In our study, adverse events were documented in 12.5 % of patients. Three patients,
who developed life threatening bleeding, required special attention. They had documented
splenic artery pseudoaneurysms on CT angiography. Initial CT did not reveal any pseudoaneurysm.
In all three patients, there were significant impingements of the splenic artery by
flanges of the FCSEMS and all the stents were long (30 mm). We assume that the pseudoaneurysms
had occurred either because of the disease process or because of the long stent or
its shape, so we recommend that short length and smooth flanged FCSEMS are used as
they are associated with fewer complications.
The migration rate of metal stents ranged from 0 % to 11 % in various studies. Chandran
et al. [10] and Walter et al. [11] reported migration in 4 of 54 and 3 of 61 PFCs, respectively, during direct endoscopic
necrosectomy (DEN). Migration of the stent can be either external (into the gastrointestinal
lumen) or internal (into the WON cavity). External migration, particularly when associated
with resolution of WON, does not require further intervention. For internal migration
of a stent, it can be salvaged if detected during the procedure. Clinically significant
migration of stents, whether internal or external, can occur occasionally and requires
repositioning them to their originally intended site. In this study, stent migration
occurred in 3 patients (internal in 1, external in 2), all of which could be salvaged.
One stent migration needed special attention. As reported earlier, this rare complication
of partial stent migration led to GOO. In this case, the stent was long (30 mm), and
we note that the short stent is associated with fewer complications.
A large number of patients with WON having significant solid debris may require re-intervention
because of its non-resolution with or without occlusion of the FCSEMS. Endoscopic
re-intervention is generally required to remove this clinically significant adherent
or non-adherent debris. Non-adherent debris may clog the stent lumen after the fluid
within the WON drains out, and may require declogging which can be carried out using
various endoscopic devices, especially the snare. Debris adhering to the wall of the
WON requires a more intense cleaning approach, known as direct endoscopic necrosectomy
(DEN).
There are a lot of published data on the use of DEN and this approach has a high risk
of complications [18]
[19]. In a study of 93 patients who underwent direct transluminal endoscopic removal
of pancreatic necrotic material (mean of six interventions), there was an 80 % rate
of clinical success; however, there was also a 26 % complication rate and a 7.5 %
mortality rate at 30 days [18]. In another study of 104 patients, clinical success was achieved in 91 % of patients
with a mean time to resolution of 4.1 months after initial DEN. Complications occurred
in 14 % of patients and included 5 retrogastric perforations or pneumoperitoneum [19]. In comparison to most published data on FCSEMS for drainage of WON, our method
is minimally invasive, as we only performed large volume saline irrigation, without
any intracavitary debridement. Smith et al. [2] used 700 – 1200 mL of saline for forceful irrigation and lavage using a waterjet.
We used almost the same amount of saline, i. e. 1 L in addition to hydrogen peroxide.
Possibly irrigation with hydrogen peroxide and a large volume of saline solution causes
liquefaction of necrosis and may also control local sepsis. We believe that repetitive
lavage and the presence of a sufficiently large conduit to allow the semiliquid debris
to flow out were the key factors for success. This aspect of drainage makes our procedure
safe as previous studies had significant complications, especially bleeding and local
infection secondary to necrosectomy. Our study shows complete resolution of symptomatic
WOPN in all of the patients with minimal adverse events.
The removal of FCSEMS is straightforward and uneventful in the majority of patients.
The mean stenting period was 72 days (range 62 – 106 days) and all stents could be
safely removed; however, some authors have suggested safe early removal of FCSEMS
by 3 weeks to prevent embedding in tissue [20].
Our study had several strengths. First, we followed a systematic step-up approach
in managing WON with FCSEMS and demonstrated incremental benefit with each step. Second,
by using the less invasive irrigation technique, the aggressive step of DEN and its
associated complications can be avoided, and it is also time consuming.
Our study had some limitations. It was a single center study and comparison with established
methods was not done. Only a randomized, multicentric comparative study can answer
the question whether this method is more efficacious and safer.
We conclude that the step-up approach for the management of symptomatic WON using
endoscopic drainage with FCSEMS followed by repeated endoscopic irrigation of the
cavity using a saline solution containing hydrogen peroxide, is a minimally invasive,
safe, and effective method with a high technical and clinical success rate. We also
conclude that long stents are more frequently associated with complications. Although
our results are promising, future large and randomized studies are needed to validate
our technique.
