Introduction
Acute pancreatitis is one of the most common medical conditions encountered by medical
professionals worldwide. While most cases resolve without sequelae, between 10 % and
20 % of patients will develop more serious adverse events such as pancreatic necrosis,
with an increased rate of morbidity and mortality [1]. Furthermore, approximately 30 % of patients with necrotizing pancreatitis will
develop a secondary infection in the weeks following, usually 3 to 4 weeks after the
onset of necrosis [2]. If left untreated, infected pancreatic necrosis has a markedly elevated mortality
rate; however, aggressive supportive care and intervention significantly improves
outcomes [3]
[4]. Ideally, therapeutic intervention is delayed to allow maturation of the collection
to form walled-off pancreatic necrosis (WOPN).
The optimal interventional modality for the treatment of WOPN remains controversial.
Historically, surgical necrosectomy was performed; however with adverse event rates
of 40 % to 70 % and mortality rates as high as 56 % reported, this procedure possess
its own toxicity and is typically approached with trepidation [5]
[6]. The PANTER study demonstrated that a step-up approach, starting with CT-guided
percutaneous drain placement was superior to up-front open surgical necrosectomy,
thus demonstrating the value of a minimally invasive approach for this condition [7]. However, many of these patients eventually still required surgical necrosectomy.
More recently, direct transgastric endoscopic necrosectomy involving endoscopic ultrasound
(EUS)-guided, translumenal drainage and debridement has been shown to be an effective
treatment for WOPN with an improved safety profile [8]
[9]
[10]. However, this approach can be complex and is limited to necrosis abutting the stomach
or duodenum.
Direct percutaneous endoscopic necrosectomy was first described in 2000 as a novel
minimally invasive intervention for debridement and washout of WOPN with two series
more recently reported in China and India ( [Table 1]) as well as several case reports [11]
[12]
[13]
[14]
[15]
[16]. This novel interventional endoscopy approach utilizes a percutaneous access tract
previously placed by interventional radiology methods to directly access the abscess
or necrosis for debridement and washout using flexible endoscopy. This allows the
patient to avoid major surgery and typically requires only moderate sedation. Furthermore,
direct percutaneous endoscopic necrosectomy could possibly be used for various types
of intra-abdominal fluid collections, regardless of anatomic location, provided that
it can be first accessed by interventional radiology techniques. A recent clinical
series from India illustrated the efficacy and safety of direct percutaneous endoscopic
necrosectomy in the treatment of infected WOPN [12]. In this study, we intend to confirm these results in a US referral center and further
evaluate the clinical value of direct percutaneous endoscopic necrosectomy in the
treatment of other types of intra-abdominal fluid collections and necrosis.
Table 1
Published case series involving direct percutaneous endoscopic necrosectomy.
|
Study
|
Participants
|
Time to intervention (mean)
|
Number of necrosectomies (mean)
|
Average hospital stay
(median)
|
Adverse events
|
Resolution
|
|
Dhingra et al
|
15
|
39.2 days
|
5
|
54 days
|
Fistula, bleeding
|
93 %
|
|
Mui et al
|
13
|
n/a
|
3
|
84 days
|
Fistula
|
67 %
|
Patients and methods
In this retrospective cohort study, 12 consecutive patients undergoing direct percutaneous endoscopic necrosectomy over
the period of 2007 to 2014 were identified. All patients had previously undergone
CT-guided percutaneous drain placement without resolution of their symptoms due to
the presence of solid necrosis and/or loculations. Time to intervention was defined
as the number of days between the onset of symptoms and the first percutaneous endoscopic
necrosectomy. The primary endpoint was complete removal of all percutaneous drains
without recurrence of clinical symptoms. Time to resolution was defined as the number
of days between the first endoscopic intervention and the date in which all drains
were removed.
Endoscopic Technique
Prior to endoscopy, all patients had their percutaneous drains upsized to 24- to 28-F
diameter to accommodate the endoscope through the body wall access point ( [Fig.1 a]). Once accessed, a fluid sample was collected and sent for amylase level and cytology.
The patients were then placed under either moderate sedation with fentanyl and midazolam
or general anesthesia if clinically warranted. The percutaneous drain(s) was then
removed and standard 8.8-mm upper endoscope (GIF-Q180; Olympus Inc., Center Valley,
PA) was introduced through the established tract into the necrotic cavity ( [Fig.1 b] and [Fig.1 c]). The cavity was visualized ([Fig. 1 d]), lavaged with normal saline, and necrotic debris then removed using blunt removal
and washout. A standard polypectomy snare was typically used through the scope to
mobilize and remove solid debris ([Fig. 1 e]). Large necrotic pieces were sequentially removed over the course of the procedure,
and once debridement was satisfactory, the endoscope removed and the percutaneous
drain replaced over a guidewire. If needed, repeat percutaneous debridement would
occur within a few days until all necrotic material was removed ([Fig. 1 f]). Drain output was monitored and drains were downsized and then removed once output
fell below 30 mL per day and cross-sectional imaging confirmed resolution ( [Fig.1 g]). Patients were subsequently followed in clinic over the course of 1 year to monitor
for recurrence of signs or symptoms and need for further intervention.
Fig. 1 a Axial CT scan with WOPN (red arrow) with percutaneous drain in place. b Endoscopic image of sinus tract. c Fluoroscopy of percutaneous endoscope accessing WOPN via sinus tract d Initial visualization of necrotic cavity prior to debridement. e Removal of necrotic material with polypectomy snare. f Endoscopic image of necrotic cavity after debridement. g Follow-up axial CT scan 9 months after direct percutaneous endoscopic necrosectomy
demonstrating complete resolution of WOPN.
Results
A total of 12 patients underwent direct percutaneous endoscopic necrosectomy over
the study time period ([Table 2]). The majority of patients (75 %) were female with an average age of 51. Ten patients
(83 %) had been diagnosed with WOPN, one (8 %) with omental necrosis after resection
of a gastrointestinal stromal tumor, and one (8 %) with bilateral, necrotic, loculated
hepatic abscesses. Six patients (50 %) presented with manifestations of marked disease
severity including sepsis and multiorgan failure. The median time from onset of symptoms
until the first necrosectomy was 85 days (range 21 – 248) ( [Table 3]). The mean number of necrosectomies performed was 2.3.
Table 2
Patient demographics and procedure details.
|
Sex, age
|
Etiology of necrosis
|
Sepsis/
multiorgan failure
|
Size of necrosis (cm)
|
Number of percutaneous drains
|
Time to intervention (days)
|
Number of PEN
|
Time to resolution
(days)
|
Adverse events
|
1-year sustained resolution
|
|
M, 63
|
Pancreatic
|
No
|
10.5 × 3.7
|
1
|
99
|
2
|
57
|
Fistula
|
Yes
|
|
M, 46
|
Hepatic
|
Yes
|
11.9 × 9.7 12.7 × 6.4
|
3
|
21
|
4
|
74
|
None
|
Yes
|
|
F, 51
|
Pancreatic
|
Yes
|
23.8 × 15.5
|
1
|
61
|
2
|
171
|
None
|
Yes
|
|
F, 34
|
Pancreatic
|
Yes
|
21.2 × 14.9 13.5 × 7.6 12.3 × 5.5
|
3
|
68
|
5
|
123
|
None
|
Yes
|
|
F, 65
|
Pancreatic
|
Yes
|
14.7 × 3.8
|
2
|
87
|
4
|
n/a
|
None
|
n/a
|
|
F, 73
|
Omental
|
No
|
9.8 × 3.5
|
1
|
86
|
1
|
31
|
None
|
Yes
|
|
F, 45
|
Pancreatic
|
No
|
14.1 × 12.9
|
1
|
79
|
1
|
10
|
None
|
Yes
|
|
F, 26
|
Pancreatic
|
No
|
6.4 × 2.5
|
1
|
248
|
1
|
19
|
None
|
Yes
|
|
F, 53
|
Pancreatic
|
No
|
11.6 × 5.2
|
1
|
124
|
1
|
57
|
None
|
Yes
|
|
M, 41
|
Pancreatic
|
Yes
|
7.0 × 3.2
|
3
|
53
|
1
|
59
|
None
|
Yes
|
|
F, 61
|
Pancreatic
|
No
|
12.4 × 2.5
|
1
|
159
|
1
|
210
|
None
|
Yes
|
|
F, 54
|
Pancreatic
|
Yes
|
10.6 × 9.6
|
2
|
84
|
5
|
40
|
None
|
n/a
|
PEN, percutaneous endoscopic necrosectomy
Table 3
Group analysis of procedure details and outcomes.
|
Time to intervention (days, median)
|
85
|
|
Number of percutaneous necrosectomies performed (mean)
|
2.3
|
|
Time to resolution (days, median)
|
57
|
Complete removal of percutaneous drains was accomplished in 11 patients (92 %). The
one patient in whom the drains were not removed died 3 months after her last pancreatic
endocrine neoplasm (PEN) from metastatic colon cancer. The median time from the initial
PEN to complete removal of drains was 57 days. No adverse events were observed in
11 patients (92 %). One patient (8 %) experienced a persistent sinus tract fistula,
which has been treated conservatively. There were no procedure-related mortalities.
Ten patients (83 %) completed 1 year of outpatient follow up, none of whom required
further intervention. One patient has had drains removed for 7 months (at the time
of publication) and has not yet completed 1-year follow up. One patient died of metastatic
colon cancer 3 months after her last necrosectomy with drains in place. No patients
required surgery or repeat percutaneous access after necrosectomy.
Discussion
Infected WOPN is a life-threatening adverse event of acute pancreatitis. Previous
studies have shown that a more minimally invasive, step-up approach with percutaneous
drainage is superior to up-front surgery; however, these patients often fail drainage
and ultimately require surgical necrosectomy which carries high morbidity and mortality
rates [3]
[7]. More recently, several endoscopic modalities have been developed to improve or
replace the step-up approach and avoid surgical necrosectomy [3]
[8]
[9]. Direct percutaneous endoscopic necrosectomy has shown initial promise in this arena,
however, the supporting data are limited [11]
[12].
In this study, we have demonstrated the effectiveness and safety of direct percutaneous
endoscopic necrosectomy in the treatment of infected WOPN as well as other intra-abdominal
fluid collections. When compared with the recent Indian study by Dhingra et al, the
current series demonstrates a lower mean number of necrosectomy procedures required
per patient (2.3 vs. 5). We also had a longer median time from onset of symptoms until
the initial percutaneous necrosectomy (85 days vs. 39 days). This longer delay may
have allowed further maturation of the walled-off fluid collection, thereby allowing
for a more complete and aggressive debridement per session, which facilitated fewer
total sessions per patient. We had successful removal of all percutaneous drains in
all but one patient with a median time to resolution of 57 days. Furthermore, there
were no mortalities and only one minor adverse event. In comparison with the Chinese
series by Mui et al, the current series had a significantly higher clinical success
rate (92 % vs. 66 %). We attribute this difference to their use of a small 5-mm choledochoscope,
thereby limiting the extent and efficacy of debridement.
Several other minimally invasive interventions have been studied in the treatment
of WOPN, particularly direct transgastric endoscopic necrosectomy. Indeed, evidence
demonstrates that direct transgastric endoscopic necrosectomy is effective with superior
mortality and morbidity rates when compared with surgical approaches in the management
of WOPN in appropriate patients [3]
[8]
[9]
[10]. In the GEPARD study, 93 patients with infected WOPN underwent direct transgastric
endoscopic necrosectomy with an 80 % clinical success rate [9]. However, they also experienced a 26 % adverse event rate and 7.5 % 30-day mortality
rate. In another multicenter study, 104 patients with WOPN underwent direct transgastric
endoscopic necrosectomy with a success rate of 91 % and an adverse event rate of 14 %
with one periprocedural death [10]. Thus, although a transgastric approach offers excellent clinical success rates
and improved safety profiles when compared with surgical approaches, the overall morbidity
and mortality rates of 25 and 7 are still significant [3]. Here we have demonstrated a comparable success rate, but with an improved risk
profile when compared with reported rates of surgical necrosectomy and even direct
transgastric endoscopic necrosectomy.
We believe direct percutaneous endoscopic necrosectomy offers several advantages over
the transgastric approach in certain situations. First, percutaneous access along
a predefined tract avoids many of the inherent complexities of a translumenal approach,
which may explain the superior safety profile reported here. Second, direct percutaneous
endoscopic necrosectomy can be performed using conscious sedation in an endoscopy
suite rather than the general anesthesia often required for prolonged per-oral endoscopies.
Third, the percutaneous approach is not limited to collections with abutment to the
stomach or duodenum but rather, can be utilized for any intra-abdominal fluid collection
accessible to interventional radiology techniques, such as the successful bilateral
hepatic abscess debridement and washout and omental necrosis debridement included
in this series.
There are several limitations to our study and this technique worth noting. First,
a well-known potential adverse event of percutaneous drainage of pancreatic fluid
collections is pancreaticocutaneous fistula formation, as was seen in one patient
in this study. Our patient experienced only mild discomfort, but these fistulas have
the potential to become infected and cause long-term problems. Second, one of the
advantages of direct transgastric endoscopic necrosectomy is complete internalization
of all hardware and close proximity of the target to the drainage site without traversing
intervening bowel or abdominal vessels. In that respect, each patient is unique and,
in our experience, many patients will be best served with direct transgastric endoscopic
necrosectomy or rarely a surgical approach, depending on their anatomic and clinical
features. Third, this was a retrospective study with no control group for comparison
of outcomes. Finally, this complex procedure should only be performed in a tertiary
care center with expert interventional endoscopists and appropriate surgical availability.
Despite these limitations, this is the largest US report of outcomes from direct percutaneous
endoscopic necrosectomy and supports the efficacy and safety of this approach.
In conclusion, direct percutaneous endoscopic necrosectomy is a safe and effective
intervention for intra-abdominal fluid collections and necrosis in appropriately selected
patients. Our study demonstrates a high clinical success rate with minimal adverse
events. This modality offers several new advantages including use of conscious sedation,
improved accessibility, and an excellent safety profile.
