Introduction
Esophageal obstruction or gastric compression caused by loculated malignant ascites
are occasional clinical problems that lack a standard therapeutic approach. Development
of malignant ascites usually portends a poor prognosis, with an average 6-month survival
rate of 11 %, except in cases of ovarian cancer, where 5-year survival can exceed
27 % [1]. If ascites occurs in a diffuse form, repeat paracentesis or tunneled catheter systems
can readily be applied for palliation of symptoms. In cases of loculated ascites,
however, these approaches might not be technically feasible.
Endoscopic ultrasound (EUS)-guided drainage is now a well-established procedure for
pancreatic fluid collections (PFC) using either plastic pigtail stents or fully-covered
self-expanding metal stents (FCSEMS) [2]. PFC usually occur following acute pancreatitis or pancreatic surgery. In the setting
of malignant ascites, however, there are only case reports of successful application
of EUS-guided drainage. Using FCSEMS, symptomatic relief in three cases of loculated
malignant ascites could be achieved [3].
We report two cases where plastic pigtails stents were being used to treat loculated
malignant ascites causing esophagogastric obstruction.
Case reports
Patient 1
A 36-year-old female with stage IV rectal cancer was admitted due to nausea, dysphagia,
anorexia and dyspnea. She had been diagnosed with synchronous metastatic rectal cancer
20 months earlier, and had undergone neoadjuvant chemoradiotherapy, abdominoperineal
resection, partial liver resection and several lines of systemic chemotherapy. Two
weeks prior to admission, she had received a radiologically placed, tunneled peritoneal
drainage catheter for palliation of malignant chylous ascites.
Upon admission, this catheter had to be removed for infection with Staphylococcus aureus. A computed tomography (CT) scan revealed free intraperitoneal fluid next to several
loculated fluid collections, one of them herniating from the epigastrium to the lower
posterior mediastinum, causing compression of the esophagus at the level of the diaphragmatic
hiatus. The fluid collection was inaccessible to transabdominal needle puncture due
to its subdiaphragmatic position in the area of atypically resected liver segments
3 and 4 ([Fig. 1]). A gastroscopy was performed showing total collapse of the distal esophagus caused
by bulging of its left anterior wall. The stenosis could be passed with the endoscope
by applying slight pressure.
Fig. 1 Patient 1, EUS-guided drainage of loculated ascites herniating from resected liver
area to the mediastinum. a Contrast-enhanced CT scan prior to the procedure, sagittal, and b, c coronal views, showing cyst-shaped ascites obstructing the air-filled esophagus (asterisk)
with stomach (st) adjacent to the cyst (cy). Endoscopic views prior to the procedure,
d showing trapped fluid in the esophagus oral to the stenosis and e the stenosis itself in the forward and f in the retroflexed view. g EUS view of puncture with a 10 F cystotome. Fluoroscopic views of the procedure,
showing h linear echoendoscope with guidewire and i deployed double-pigtail plastic stent. Multiple staples from partial liver resection
can also be seen.
Using a therapeutic echoendoscope (Pentax EG-3630U, Tokyo, Japan), an 80 × 80-mm almost
echo-free cyst was identified as the cause of obstruction and punctured from a position
approximately 2 cm oral to the esophagogastric junction, using a 19 G Echotip needle
(Cook Medical, Limerick, Ireland). The 550 mL of fluid removed was slightly turbid
and milky appearing. Following the procedure, the patient felt immediate relief of
her dysphagia, but the symptoms had returned by the next morning. Therefore, the procedure
was repeated twice with the same result each time. After discussing the experimental
character of the procedure with the patient, a standard 0.035-inch guidewire was inserted
into the cavity after re-puncture from the gastric cardia under fluoroscopic guidance.
The cystogastrostomy tract was established using a 10 French cystotome (Cook Medical),
and a 10 French 5-cm double-pigtail plastic stent (Cook Medical) was placed into the
cyst. After the procedure the patient experienced ongoing clinical improvement of
her dysphagia. An episode of fever quickly resolved with antibiotic therapy. No other
adverse events (AEs) occurred, and the pigtail drainage was left in place.
The patient died 4 weeks later from progression of her rectal cancer.
Patient 2
A 56-year-old female with progressive stage IV ovarian cancer was admitted due to
nausea and vomiting to the point where even intake of oral medication was impossible.
She had been diagnosed with localized cancer 7 years earlier, but had subsequently
developed hepatic, lymphatic, and peritoneal spread with ascites, and received several
lines of systemic chemotherapy.
Gastroscopy revealed a near-complete constriction of the gastric cavity at the level
of corpus and fundus by bulging of the inferior gastric wall. This corresponded to
loculated ascitic fluid in the omental bursa seen on abdominal ultrasound and CT scan
([Fig. 2]). Using a therapeutic echoendoscope (Pentax EG-3870UTK, Tokyo, Japan), the 85 × 50-mm
echo-free cyst containing few septae could be identified and punctured as described
above from a position in the gastric corpus, leaving a 10 French 5-cm double-pigtail
plastic stent in place. The procedure resulted in partial improvement of symptoms,
but repeat gastroscopy still showed significant narrowing of the gastric lumen. Upon
repeat EUS, only part of the loculated ascitic fluid collection appeared sufficiently
drained, but another, separate part measuring 61 × 34 mm did not. Repeat drainage
with another double-pigtail stent led to radiologic complete remission of the culprit
ascites pocket and enabled recommencement of oral intake. No AEs were observed.
Fig. 2 Patient 2, EUS-guided drainage of loculated ascites in the omental bursa causing
near-complete constriction of the gastric cavity. a Non-enhanced CT scan prior to the procedure, showing cystic fluid collection (cy)
compressing the stomach (asterisk). b Endoscopic forward view of the stomach prior to the procedure. c Contrast-enhanced CT scan (arterial phase) 2 weeks after completion of drainage,
showing collapsed cyst containing plastic stents. d Endoscopic view immediately after placement of 2nd stent.
Subsequently, another line of chemotherapy was started, leading to oncological partial
remission. At the time of submission of this report, after 6 months of follow-up,
the patient had progressive ascites requiring repeat paracentesis. Yet she was still
in clinical remission of her gastric compression symptoms with the double-pigtail
drainage left in place.
Discussion
Malignant ascites is a frequent clinical problem giving rise to a variety of symptoms
such as abdominal swelling and pain, nausea, anorexia, fatigue, and dyspnea, that
can significantly impair quality of life. Diuretics are of little value in this setting
except for a subgroup of patients with portal hypertension due to liver metastases
[4]. Some form of drainage procedure is often mandated as part of supportive care measures.
Drainage methods include repeat paracentesis, permanent peritoneal ports, tunneled
peritoneal catheters, central venous catheters and peritoneovenous shunts. The latter
have high rates of serious AEs including mortality from pulmonary edema, disseminated
intravascular coagulation and thromboembolism. Frequently used techniques are repeat
paracentesis and tunneled peritoneal catheters.
We described two cases of loculated ascites that lacked apposition to the abdominal
wall due to their deep epigastric position. Hence paracentesis or insertion of a peritoneal
catheter did not appear safely feasible because organs like liver or stomach were
interposed. Yet the patients were severely symptomatic from esophagogastric obstruction
symptoms. For this particular situation of an encysted symptomatic fluid collection,
we thought it appropriate to apply the method of EUS-guided drainage that is well
established in PFC [5]. In both scenarios, immediate contact of encapsulated fluid to the gastric or duodenal
wall allows a drainage procedure that minimizes risk of spillage of intestinal contents
to the free peritoneal cavity.
In the setting of malignant ascites, efficacy and safety of merely diagnostic EUS-guided
paracentesis has been demonstrated [6]. EUS-guided therapeutic drainage of loculated malignant fluid collections has been
reported with technical and clinical success in three cases using a lumen-apposing
metal stent (LAMS) [3]. Up to now, there is no clear evidence to suggest that expandable metal stents are
superior to plastic stents [5]. In an ongoing randomized trial comparing LAMS with multiple plastic stents for
treatment of walled-off pancreatic necrosis, the protocol had to be changed towards
early removal of LAMS due to serious AEs including bleeding [7]. The authors hypothesize that plastic stents likely gravitate towards the gastrointestinal
lumen as the PFC resolves, whereas LAMS remain in place with the resultant friction
against regional vasculature precipitating bleeding. In cancer patients with malignant
ascites, the necessary duration of drainage is unclear at the time of stent insertion,
and avoidance of complications that are related to possibly long treatment duration
is of utmost importance. For these reasons, we believe that plastic pigtail stents
could be more appropriate in the setting of drainage of malignant ascites. In case
of viscous fluids, either LAMS or multiple plastic stents (allowing inter-stent flow)
could be a therapeutic option.
Three cases of successful EUS-guided drainage using double-pigtail plastic stents
to obviate the need for repeat paracentesis have recently been described in this journal
[8]. In these patients malignant ascites was not described as loculated. Their follow-up
was short, however, as all patients died within days to weeks after the procedure.
One concern in non-loculated ascites would be that deliberately creating a connection
of the gastric lumen with the free peritoneal space could promote infectious peritonitis.
In our patients, however, ascites was strictly loculated thus minimizing this risk.
Still, the possibility of intraperitoneal leakage is one of the drawbacks of this
method, in which case we would have removed the stents. Therefore, it should be reserved
for patients receiving palliative care after evaluation of alternatives. However,
once the drainage is in place, the resulting peri-drainage inflammation will reduce
this risk over time.
One unanswered question regards timing of stent removal. A lower incidence of recurrence
of PFC without AEs like stent clogging or infection was reported in a randomized controlled
trial when plastic stents were removed after 12 months compared to removal after 2
months [9]. We opted to leave stents in place as long as they appeared effective and no AEs
were observed. Another concern we had prior to the procedure was whether different
pressure in the thoracic and the abdominal compartment would hamper drainage of a
mediastinal fluid collection through a stent inserted into the subdiaphragmatic compartment
of the same collection. This, however, did not seem to prevent clinical success in
the case of our patient. The chosen approach could also be applied to ascites of non-malignant
origin other than PFCs. However, because there is limited experience it should be
restricted to those with clear demarcation and after consideration of all other therapeutic
options.
Obviously, the reported number of cases is too small to show efficacy and safety of
the method. Further studies regarding EUS-guided drainage of loculated malignant ascites
are needed.
Conclusion
We described for the first-time application of the well-established technique of EUS-guided
drainage using plastic stents to address loculated symptomatic ascites not amenable
to paracentesis. One of our patients is still in remission of her esophagogastric
obstruction symptoms after 6 months of follow-up.
