Materials and Methods
Retrospective analysis of prospectively maintained database of patients who underwent
EUS-guided liver abscess drainage from 2019 to May 2024 was done. Demographics, clinical
data, procedural data, and adverse events were collected. Abscesses were analyzed
for their location, number, size, puncture distance, indication of drainage, route
of access, and the endoprosthesis which was used for drainage.
Results were reported as mean or median (range) for quantitative variable and percentage
for categorical variable.
All patients were on intravenous (IV) Metronidazole and IV broad-spectrum antibiotics.
None of these enrolled patients had any previous attempt of PCD or EUS-guided drainage.
These patients were taken up for EUS-guided aspiration or transmural drainage under
conscious sedation.
Consent for transmural drainage was taken in all patients. Aspiration was done in
subcapsular abscess with predominantly liquefied contents. Transmural drainage was
undertaken in ruptured abscesses or subcapsular abscesses with non-liquefied contents.
Detailed segmental localization of abscess was done by ultrasonography (USG) or computed
tomography scan. EUS evaluation was done to decide the shortest (transgastric/transduodenal/transesophageal)
and safest route to access the abscess cavity. After securing a vessel-free approach,
the abscess was punctured with 19G needle (Echotip Ultra: Cook-Endoscopy; [Fig. 1]). When the plan was aspiration, active suction of the abscess was done to aspirate
all the liquefied contents. In cases where aspiration yield was insignificant, transmural
drainage was undertaken. Steps of transmural drainage: a 0.035-inch guide-wire (Jagwire:
Boston Scientific) was coiled in the abscess cavity. Serial dilatation was performed
using a 6F Cystotome (Endoflex Germany) using electrocautery (ERBE VIO-300D), followed
by a 7F and 10F Soehendra dilator (Cook-Endoscopy). Active aspiration was done with
a 10F Soehendra dilator with flushing requirement standby if catheter got blocked.
A 10F nasocystic drain (NCD) was then placed for transmural drainage under EUS and
fluoroscopic guidance, ensuring all side holes of NCD were within the abscess cavity
([Fig. 2] and [Video 1]). In our first patient, a combination of 8F double pig tail (DPT) stent and 8F NCD
was used, thereafter we shifted to 10F NCD.
Fig. 1 Abscess punctured with 19G needle under EUS visualization. EUS, endoscopic ultrasound.
Fig. 2 Fluoroscopic image of 10F NCD deployed transesophageal following tract dilatation.
Video 1 6.9 by 6.1 cm large heterogeneous abscess, in segment 6. A 19G EUS needle was introduced
into abscess cavity transduodenal. This was followed by coiling of 0.035-inch guidewire
in the abscess cavity. Serial dilatation was performed using a 6F Cystotome followed
by a 7F and 10F Soehendra dilator. Active aspiration was done with a 10F Soehendra
dilator. A 10F NCD was then deployed for transmural drainage.
NCD was put on intermittent suction until there was spontaneous drainage aided by
gravity. USG was done 24 to 48 hours after the procedure. NCD was internalized into
the stomach with a Ryles tube when there was complete resolution of clinical symptoms
with disappearance or at least >50% reduction of abscess size on repeat imaging (nasal
end of 18F Ryles tube is cut at 1.5 cm, inverted, and reinserted; this funnel-shaped
end thus created acts as an inserter to push/internalize the cut nasal end of NCD
into the stomach). USG was done weekly thereafter for 4 weeks.
Technical success was defined as the ability to access and drain the abscess. Clinical
success was defined as complete resolution of clinical symptoms with at least 50%
reduction in abscess size on repeat imaging.
Results
This study was conducted in an area endemic for amoebic liver abscess. In all, 312
patients underwent liver abscess drainage during the study time period. In total,
266 (85.25%) patients underwent PCD under ultrasound guidance and in 46 (14.75%) EUS-guided
drainage was done either in isolation or with PCD.
A total of 46 patients (44 males and 2 females) who underwent EUS-guided liver abscess
drainage were included in the study. Transmural drainage was done in 31 (67.4%; [Table 1]) and aspiration in 15 (32.6%; [Table 2]).
Table 1
Characteristics of patients who underwent transmural drainage
|
S/NO
|
Age
|
Location
|
Route
|
Abscess
|
Indication
|
Size (cm)
|
Puncture
distance (cm)
|
Endoprosthesis
|
|
1
|
30/M
|
Caudate
|
TG
|
S(AB)
|
RA
|
8.6 × 7.8
|
<2
|
8F DPT & 8FNCD
|
|
2
|
55/M
|
Segment 3
|
TG
|
MA(AB)
|
RA
|
7.3 × 5.7
|
<2
|
10F NCD
|
|
3
|
22/M
|
Segment 3
|
TG
|
S(AB)
|
SC
|
9.3 × 7.1
|
<2
|
10F NCD
|
|
4
|
65/M
|
Caudate
|
TG
|
S(PY)
|
RA
|
5.8 × 5.6
|
<2
|
Two 10F NCD
|
|
5
|
40/M
|
Segment 2
|
TE
|
S(AB)
|
RA
|
6.5 × 5.3
|
<2
|
10F NCD
|
|
6
|
40/M
|
Segment 3
|
TG
|
S(PY)
|
SC
|
7.1 × 6.9
|
<2
|
10F NCD
|
|
7
|
35/M
|
Caudate
|
TG
|
S(AB)
|
SC
|
10.1 × 8.1
|
<2
|
10F NCD
|
|
8
|
18/M
|
Segment 4
|
TD
|
MA(AB)
|
SC
|
6.3 × 6.2
|
2–4
|
10F NCD
|
|
9
|
23/M
|
Segment 5
|
TD
|
S(AB)
|
SC
|
6.8 × 6.1
|
<2
|
10F NCD
|
|
10
|
60/F
|
Caudate
|
TE
|
S(AB)
|
SC
|
7.9 × 7.5
|
<2
|
10F NCD
|
|
11
|
30/M
|
Segment 3
|
TG
|
S(AB)
|
SC
|
6.5 × 5.7
|
<2
|
10F NCD
|
|
12
|
40/M
|
Segment 8
|
TE, TS
|
S(AB)
|
SC
|
5.0 × 4.6
|
2–4
|
10F NCD
|
|
13
|
31/M
|
Segment 8
|
TE, TS
|
MA(AB)
|
SC
|
6.4 × 6.1
|
2–4
|
10F NCD
|
|
14
|
32/M
|
Segment 4
|
TG
|
S(AB)
|
SC
|
7.5 × 6.3
|
<2
|
10F NCD
|
|
15
|
46/M
|
Segment 3
|
TG
|
S(AB)
|
RA
|
7.0 × 6.5
|
<2
|
10F NCD
|
|
16
|
55/F
|
Segment 8
|
TE, TS
|
S(AB)
|
RA
|
7.2 × 6.4
|
2–4
|
10F NCD
|
|
17
|
42/M
|
Segment 3
|
TG
|
S(AB)
|
SC
|
6.7 × 5.8
|
<2
|
10F NCD
|
|
18
|
40/M
|
Segment 8
|
TD, TS
|
S(AB)
|
SC
|
7.4 × 6.9
|
>4
|
10F NCD
|
|
19
|
50/M
|
Segment 6
|
TD
|
MA(AB)
|
SC
|
5.9 × 4.1
|
<2
|
10F NCD
|
|
20
|
40/M
|
Segment 6
|
TD
|
S(AB)
|
RA
|
6.9 × 6.1
|
<2
|
10F NCD
|
|
21
|
38/M
|
Segment 3
|
TG
|
S(AB)
|
S C
|
6.6 × 6.6
|
<2
|
10F NCD
|
|
22
|
24/M
|
Caudate
|
TG
|
MA(AB)
|
RA
|
6.6 × 5.1
|
<2
|
10F NCD
|
|
23
|
56/M
|
Segment 8
|
TD, TS
|
MA(AB)
|
RA
|
5.4 × 5.1
|
>4
|
10F NCD
|
|
24
|
40/M
|
Segment 3
|
TG
|
MA (PY)
|
SC
|
8.8 × 6.9
|
<2
|
10F NCD
|
|
25
|
45/M
|
Caudate
|
TG
|
S(AB)
|
RA
|
5.8 × 5.5
|
<2
|
10F NCD
|
|
26
|
22/M
|
Segment 3
|
TG
|
MA(AB)
|
SC
|
8.5 × 5.5
|
<2
|
10F NCD
|
|
27
|
52/M
|
Caudate
|
TG
|
S(AB)
|
SC
|
6.6 × 5.2
|
<2
|
10F NCD
|
|
28
|
56/M
|
Segment 6
|
TD
|
MA(AB)
|
SC
|
7.6 × 6.6
|
<2
|
10F NCD
|
|
29
|
38/M
|
Segment 5
|
TD
|
S(AB)
|
S C
|
6.5 × 6.4
|
<2
|
10F NCD
|
|
30
|
71/M
|
Segment 2
|
TE
|
MA(AB)
|
SC
|
6.7 × 5.4
|
<2
|
10F NCD
|
|
31
|
51/M
|
Segment 3
|
TG
|
MA(AB)
|
SC
|
7.2 × 5.4
|
<2
|
10F NCD
|
Abbreviations: AB, amoebic; MA, multiple abscess; NCD, nasocystic drain; PY, pyogenic;
RA, ruptured abscess; SA, single abscess; SC, subcapsular abscess; TD, transduodenal;
TE, transesophageal; TG, transgastric; TS, trans-segmental.
Table 2
Characteristics of patients who underwent aspiration
|
S/NO
|
Age
|
Location
|
Route
|
Abscess
|
Indication
|
Size (cm)
|
Puncture distance
|
|
1
|
27/M
|
Caudate
|
TG
|
MA
|
SC
|
6.1 × 5.8
|
<2
|
|
2
|
64/M
|
Caudate
|
TE
|
MA
|
SC
|
5.2 × 5.1
|
<2
|
|
3
|
26/M
|
Segment 4
|
TD
|
MA
|
SC
|
3.9 × 3.6
|
<2
|
|
4
|
46/M
|
Segment 5
|
TD
|
MA
|
SC
|
5.6 × 5.2
|
<2
|
|
5
|
55/M
|
Segment 4
|
TD
|
S
|
SC
|
4.9 × 4.5
|
<2
|
|
6
|
54/M
|
Segment 4
|
TD
|
S
|
SC
|
5.3 × 4.7
|
<2
|
|
7
|
46/M
|
Segment 6
|
TD
|
MA
|
SC
|
5.6 × 5.2
|
<2
|
|
8
|
56/M
|
Caudate
|
TG
|
MA
|
SC
|
4.8 × 4.6
|
<2
|
|
9
|
48/M
|
Segment 2
|
TG
|
MA
|
SC
|
5.3 × 4.9
|
<2
|
|
10
|
55/M
|
Segment 3
|
TG
|
MA
|
SC
|
5.0 × 4.6
|
<2
|
|
11
|
47/M
|
Segment 7
|
TD, TS
|
MA
|
SC
|
5.9 × 5.4
|
>4
|
|
12
|
31/M
|
Segment 2
|
TG
|
S
|
SC
|
7.3 × 4.8
|
<2
|
|
13
|
22/M
|
Segment 8
|
TD TS
|
MA
|
SC
|
6.6 × 5.4
|
>4
|
|
14
|
35/M
|
Segment 4
|
TD
|
MA
|
SC
|
5.6 × 6.2
|
<2
|
|
15
|
65/M
|
Segment 4
|
TD
|
MA
|
SC
|
6.2 × 5.4
|
<2
|
Abbreviations: MA, multiple abscess; NCD, nasocystic drain; SA, single abscess; SC,
subcapsular abscess; TD, transduodenal; TE, transesophageal; TG, transgastric; TS,
trans-segmental.
Indication of drainage was radiologically inaccessible abscesses in all. The mean
age of included patients was 41.51 ± 1.89 years. Etiology of abscesses was amoebic
in 43 (93.5%), pyogenic in 3 (6.5%). The mean size of abscess collection was 6.55 ± 0.33 cm.
The size was 5.55 ± 0.76 cm in the aspiration group and 7.03 ± 0.49 cm in patients
who underwent transmural drainage. The puncture distance was <2 cm in 38, 2–4 cm and
>4 cm in 4 each.
Location of abscess was caudate lobe in 10 (21.73%), segment 2 in 4 (8.7%), segment
3 in 11 (23.996), segment 4 in 7 (15.2%), segment 5 in 3 (6.5%), segment 6 in 4 (8.7%),
segment 7 in 1 (2.17%), and segment 8 in 6 (13%). Access was transesophageal in 7
(15.2%), transduodenal in 17 (36.9%), and transgastric in 22 (47.8%).
Seven patients underwent trans-segmental drainage. Segment 8 was accessed through
segment 5 or caudate lobe in three patients each, and segment 7 through segment 6
in one. 10F NCD was used as endoprosthesis in all patients except one. In our first
patient, we placed the 8F DPT stent, which got blocked, thereafter we placed an 8F
NCD alongside the stent. One patient with liver cirrhosis with pyogenic abscess required
placement of two 10F NCDs.
In 16.6% (3/18) patients where initial plan was aspiration; however, the yield was
inadequate. These patients underwent transmural drainage and were included in the
transmural group. In total, 13.3% (2/15) patients in the aspiration group required
a repeat session.
Multiple abscesses were seen in 23 (50%). These patients underwent a combination of
PCD for radiologically accessible and EUS-guided drainage for radiologically inaccessible
abscess. Ten patients presented with ruptured abscess.
Technical and clinical success of EUS-guided drainage was 100%. NCD was internalized
into stomach after a mean duration of 2.2 days. All patients improved clinically with
resolution of abscess cavity on follow-up USG. The mean duration of hospital stay
from the endoscopic procedure to discharge was 7.6 days.
Liver abscess resolved on imaging in all patients with a mean period of 23.6 days.
None had any recurrence on the median follow-up period of 560 days (range: 140–1,100).
Only complication observed was self-resolving biliary communication following transmural
drainage, seen in 8 (25.8%). In one patient, there was no drainage on aspiration following
NCD placement as the catheter had got kinked. Repeat transmural drainage was done
in the same setting. NCD was removed at 4 to 6 weeks in all our patients.
Discussion
There are only 18 case reports/series and 6 small retrospective studies in the literature
on EUS-guided drainage of liver abscess ([Table 3]). The present study, largest reported till date, further validates the role of EUS-guided
drainage in radiologically inaccessible liver abscesses.
Table 3
Summary of literature on endoscopic ultrasound-guided liver abscess drainage
|
Study
|
Year
|
Location of abscess
|
Approach
|
Endoprosthesis for drainage
|
Complications
|
|
Seewald et al (2005), Germany[3]
|
2005
|
Lateral segment of left lobe
|
Proximal TG
|
7 F NCD
|
None
|
|
Ang et al (2009), Singapore[4]
|
2009
|
Left subhepatic space collection
|
TG
|
8F and 10F × 7 cm
|
None
|
|
Noh et al (2010), Korea[5]
|
2010
|
Gastrohepatic space
|
TG
|
7F DPS
|
None
|
|
|
Caudate lobe of liver
|
TG
|
7F DPS
|
None
|
|
|
Caudate lobe of liver with portacaval extension
|
TD
|
2, 7F DPS with NCD
|
None
|
|
Itoi et al (2011), Japan[6]
|
2011
|
Between pancreas and caudate lobe of liver
|
TD
|
7F SS and 5F NCD
|
None
|
|
|
Caudate lobe
|
TG
|
7F DPS and 5F NCD
|
None
|
|
Keohane et al (2011)[7]
|
2011
|
Caudate lobe (2)
|
TG
|
7F DPS
|
None
|
|
|
|
TG
|
10F DPS
|
None
|
|
Ivanina et al (2012)[8]
|
2012
|
Caudate lobe
|
TG
|
NCD
|
NCD traversing the esophagus, new para-esophageal collection
|
|
Medrado et al (2013)[9]
|
2013
|
Left lobe
|
TG
|
PCSEMS, 60 × 10 MM
|
Stent migration in the abscess 2 weeks, 10Fr DPS inserted with in SEMS
|
|
Alcaide et al (2013)[10]
|
2013
|
Left lobe
|
TG
|
LAMS (Axios 10 × 10 mm)
|
None
|
|
Kawakami et al (2014)[11]
|
2014
|
Left lobe
|
TG
|
BFMS (Nagi 16 × 30 mm)
|
None
|
|
Koizumi et al (2015)[12]
|
2014
|
Left lobe
|
TG
|
5F NCD
|
None
|
|
Kodama et al (2015)[13]
|
2015
|
Left lobe
|
TG
|
6Fr NCD inadvertent removal later replaced by FCSEMS 10 mm × 120 mm
|
None
|
|
Study
|
Year
|
Location of abscess
|
Approach
|
Endoprosthesis for drainage
|
Complications
|
|
Tonozuka et al (2015)[14]
|
2015
|
Left lobe (6)
|
TG (6)
|
FCSEMS (16 mm × 2 cm for lesion near the wall and 10 mm 6–8 cm for lesions away from
wall
|
None
|
|
|
Right lobe (1)
|
TD (1)
|
FCSEMS
|
None
|
|
Ogura et al (2016)[15]
|
2015
|
Left lobe (6)
|
TG (6)
|
FCSEMS (10 mm × 6, 8, 12 cm), 7Fr DPS inserted in the stent
|
None
|
|
|
Right lobe (2)
|
TD (2)
|
FCSEMS
|
None
|
|
Yamamoto et al (2017)[16]
|
2017
|
Right lobe
|
TD
|
5F NCD, internalized after 6 days
|
None
|
|
Carbajo et al (2019)[17]
|
2019
|
Left lobe (3)
|
TG (3)
|
FCSEMS (60 × 10 and 40 × 10 mm)
|
1 bleed and 1 perforation managed conservatively
|
|
|
Right lobe (6)
|
TD (6)
|
LAMS (2) (10 × 10 mm and 15 mm)
|
None
|
|
Rana et al (2020)[18]
|
2020
|
Left lobe
|
TG (10)
|
Two 7F DPS
|
1 repeat procedure and exchange of stent
|
|
|
Caudate lobe
|
TE (4)
|
|
None
|
|
Chandra and Chandra (2021), India[19]
|
|
Caudate
|
TG (1)
|
8F DPS and 8F NCD active aspiration with biliary dilator before stenting
|
None
|
|
|
Left lobe and segment IV
|
TG (2)
|
10F NCD (2)
|
None
|
|
Molinario et al (2021),
Italy[20]
|
|
Left lobe
|
TG (1)
|
LAMS (Axios 10 × 20 mm), 8.5 Fr × 3 cm DPS
Placed inside the LAMS
|
None
|
|
Zanwar and Agarwal (2022), India[21]
|
2022
|
Caudate lobe
|
TG (1)
|
7F DPS with 7F NCD
|
None
|
|
Shah et al (2023),
India[22]
|
2023
|
Caudate lobe
|
TG (8)
|
DPS (6) (aspiration) 1
LAMS (1)
|
None
|
|
Shahid et al (2023)[23]
|
2023
|
Left lobe
|
TG (3O)
|
LAMS (2) FCMS (8)
PCSEMS (20)
|
None
|
Abbreviations: BFMS, biflanged metal stent; DPS, double pigtail plastic stent; FCSMES,
fully covered self-expending metal stent; LAMS, lumen apposing stent; NCD, nasocystic
drain; PCSEMS, partial covered self-expending metal stent; SS, straight stent; TD,
transduodenal; TE, transesophageal; TS, trans-segmental.
There is no consensus on EUS-guided liver abscess drainage. Summarizing available
literature,[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23] EUS allows easy access to left lobe and central right lobe liver segments. Commonest
indication of drainage is radiologically inaccessible left lobe abscess. Transgastric
is the preferred route for left lobe abscesses and transduodenal for right lobe abscesses.
The choice of endoprosthesis is individualized on endoscopists' experience?
Questions needing addressal are: (1) which endoprosthesis is to be preferred in relation
to location, route of drainage, abscess contents, or in ruptured abscesses? (2) How
safe is transesophageal and transduodenal drainage? (3) Is trans-segmental drainage
of distantly located right lobe liver segments feasible?
The present study has shown 10F NCD as an endoprosthesis has excellent technical and
clinical success, irrespective of the route (transesophageal, transgastric, or transduodenal),
location (distant or in close vicinity to the lumen), thick abscess contents, and
in ruptured abscesses.
NCD made drainage more predictable, by allowing active aspiration and/or flushing,
if it gets blocked. We could diagnose biliary communication, by monitoring the drain
output. In ruptured abscesses with adjacent perihepatic collection, active aspiration
could decompress the collection obviating the need for additional PCD or EUS-guided
drainage. NCD placement is suitable for transduodenal and transesophageal drainage,
being less invasive as it requires concordant tract dilatation. NCD is easy to deploy
in the duodenum where there is limited maneuverability. In distantly located abscess,
NCD is a safer option. When abscess contents are thick, one can place multiple 10F
NCDs and achieve the same result, as we get by upgrading percutaneous catheter (PCC)
to a wider diameter. The only drawback of NCD is that like PCD it is an external drainage;
however, we could internalize the NCD into stomach in 2 days in majority of our patients.
Placement of a7F NCD alone was successful in draining large abscess (11by 7cm), as
reported by Seewald et al. Koizumiet al, yamamoto et al reported success with 5F NCD.
We used 10F NCD as it is equivalent to 10F PCC used for liver abscess drainage. To
cover for long length of NCD compared with PCC, we added active aspiration of thick
contents until there was spontaneous drainage aided by gravity.
Successful drainage with DPT stents 7F to 10F, single or multiple, as endoprosthesis
has been reported.[4]
[5]
[7]
[18]
[22] These small-caliber stents work if the abscess cavity is within 1 to 2 cm from the
lumen and the contents of abscesses are predominantly liquefied. In our first patient,
we deployed 8F stent that got blocked, due to thick contents and the patient's condition
worsened. Unpredictability of drainage is a drawback with plastic stents, which may
lead to clinical deterioration in patients. Plastic stent deployment requires liberal
tract dilatation, thus, it is not a suitable option for transduodenal and transesophageal
drainage.
Placing NCD alongside the stent offers a wicking effect and provides the option of
performing active lavage of abscess cavity. This has worked in our first case. Successful
transmural drainage with a stent and NCD placement has been reported by Itoi et al[6] in two, and Noh et al[5] and Zanwar et al[21] in one case each. The present study has shown 10F NCD alone is sufficient, thus,
there is no need to add stents and make the procedure more invasive.
All recent publications have used partially or fully covered self-expandable metal
stent (SEMS) or lumen-apposing metal stent (LAMS) for transmural drainage.[9]
[10]
[11]
[13]
[14]
[15]
[17]
[20]
[22]
[23] These wide-bore stents cause rapid initial decompression and predictable sustained
drainage, with an option of direct endoscopic necrosectomy.[14] Fully covered self-expandable metal stents (FCSEMSs) can be deployed in distantly
located abscess, whereas LAMS in abscesses located in close vicinity to gastrointestinal
(GI) lumen. SEMS or LAMS placement requires technical expertise and all reported cases
are from high-volume centers performing EUS-guided interventions. Even in expert hands,
there are reports of severe complications like stent migration, perforation, and bleeding.[2]
[17]
[23] In addition, the large rent following LAMS or SEMS removal requires endoscopic closure
if there is no fistula formation between GI lumen and abscess cavity. The present
study has shown that the technique of 10 F NCD deployment is simple, safe, and least
invasive. Thus, with comparable clinical efficacy in the reported literature, 10F
NCD compared with SEMS or LAMS allows wider clinical application of EUS-guided liver
abscess drainage. This is especially relevant in third-world countries with limited
resources. Large multicenter studies comparing EUS-guided drainage with a FCSEMS or
LAMS and a 10F NCD are needed.
There are 10 reported cases of EUS-guided transduodenal drainage of right lobe liver
abscess.[14]
[15]
[16]
[17] FCSEMS or LAMS was used as endoprosthesis in nine and a 5F NCD was used in one.
Plastic stents are not used, as their placement requires liberal tract dilatation,
which increases the chances of peritoneal spillage of luminal and abscess contents.
We drained 14 (30.4%) right lobe liver abscesses with 10F NCD. Segments 5 and 6 could
easily be approached from the duodenum. Segment 8 could be accessed, trans-segmentally
through segment 5 or caudate lobe and segment 7 accessed through segment 6.
Trans-segmental drainage allows EUS-guided drainage of segment 7 and segment 8, which
were thought to be inaccessible to EUS-guided drainage. We found abscesses reported
in segment 7 or segment 8 on cross-sectional imaging to involve adjacent EUS assessable
segment 6, segment 5, or caudate lobe.
Furthermore, under EUS guidance, one can avoid intervening vessels by changing the
path traversed by the needle by torquing the scope, this allows trans-segmental access
of distantly located abscesses ([Video 2]). We accessed segment 8 abscess transcaudate, transesophageal alongside of inferior
vena cava.
Video 2 Large heterogenous abscess deep to the IVC in segment 8. Window available for access
by torquing the scope. 19G EUS needle introduced transcaudate, transesophageal alongside
of IVC.
We preferred NCD as endoprosthesis for trans-segmental drainage as it requires concordant
tract dilatation, thus least invasive.
In our study, six patients underwent transesophageal drainage. Rana et al[18] had reported transesophageal drainage in three patients. The present study further
supports safety of transesophageal route for EUS-guided drainage. NCD is a safer option
for transesophageal drainage as its deployment is least invasive with no risk of stent
migration.
Twenty-three patients included in the present study had multiple abscesses. These
patients underwent percutaneous aspiration and/or drainage in combination with EUS-guided
aspiration and/or drainage for the radiologically inaccessible abscesses. EUS-guided
drainage complemented PCD in patients with multiple abscesses.
The standard technique reported for EUS-guided drainage is the same as for pseudocyst
drainage.[2] However, in contrast to pseudocyst, the wall of abscess cavity is not adherent to
the luminal wall.
Thus, we advocate concordant tract dilatation using a Soehendra dilator, as this minimized
chances of intra-peritoneal spillage or mediastinal contamination. In transcaudate,
transesophageal segment 8 drainage, cystotome was avoided and a 6F Soehendra dilator
was used for tract dilatation. We have improvised by adding the step of active aspiration
of abscess contents with a 10F Soehendra dilator, this minimized peritoneal spillage
and hastened the clinical recovery.[19]
In total, 8 (25.8%) patients had biliary communication following transmural drainage
which resolved conservatively in all. A higher incidence compared with PCD could be
explained by intermittent suction which was done until there was spontaneous drainage
aided by gravity. NCD was internalized in these patients after biliary drainage dried
up. Catheter kinking was likely due to a manufacturing error, not a true complication
but a periprocedural event, corrected by repeat transmural drainage in the same setting.
Novelty of the present study lies in using 10F NCD for the first time[19] as endoprosthesis for EUS-guided transmural drainage in the largest patient cohort
reported till date with excellent safety and clinical efficacy. We could achieve the
same results as with 10F PCD as the standard of care for liver abscesses requiring
drainage. Novelty of our present study is reiterated by the wider applicability of
10F NCD as an endoprosthesis irrespective of route (transgastric, transesophageal,
or transduodenal), abscess location (close vicinity or distant), thick abscess contents,
and in ruptured abscesses. The present study also includes the first report in world
literature of EUS-guided trans-segmental, transesophageal, or transduodenal drainage
of segment 8 and transduodenal drainage of segment 7 abscesses.
A major limitation of our study is its retrospective and noncomparative design. Majority
of patients included had amoebic etiology, only 3 (7.5%) were pyogenic. However, we
do expect the same result in pyogenic and other etiologies of liver abscess.
