Endoscopic Ultrasound-Guided Drainage of Difficult-to-Access Liver Abscesses: A Large Single-Center Experience

Abstract

Objective

Endoscopic ultrasound (EUS)-guided drainage can be a minimally invasive alternative approach in draining radiologically inaccessible liver abscesses. However, experience with EUS-guided drainage is limited with no guidelines available.

The objective of this study was to share our experience in EUS-guided drainage of radiologically inaccessible liver abscesses.

Materials and Methods

Retrospective analysis of prospectively maintained database of patients who underwent EUS-guided liver abscess drainage from 2017 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 used for drainage.

Statistical Analysis

Results were reported as mean or median (range) for quantitative variable and percentage for categorical variable.

Results

A total of 46 patients (44 males and 2 females) underwent EUS-guided liver abscess drainage. Transmural drainage was done in 31 (67.4%) and aspiration in 15 (32.6%) patients. The mean size of abscess collection was 6.55 ± 0.33 cm. 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 (24.4%), segment 2 in 4 (8.7%), segment 3 in 11 (23.9%), 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 patient. 10F nasocystic drain (NCD) was used as endoprosthesis in all patients who underwent transmural drainage except one. Technical and clinical success of EUS-guided drainage was 100%.

Conclusion

EUS-guided drainage with its excellent safety profile and clinical success should always be considered for draining radiologically inaccessible liver abscesses before contemplating surgical drainage. 10F NCD works well as endoprosthesis with excellent results irrespective of approach, location, thick abscess contents, and in ruptured abscesses. EUS-guided trans-segmental drainage is technically feasible and clinically effective.

Keywords

endoscopic ultrasound - endoprosthesis - liver abscess drainage - percutaneous drainage - radiologically inaccessible abscess

Introduction

Liver abscesses requiring drainage are conventionally managed by interventional radiology-guided percutaneous drainage (PCD) with concomitant antibiotic therapy.[1] Radiologically difficult-to-access abscesses are managed with laparoscopic or open surgery, which carries a high rate of morbidity and mortality. Endoscopic ultrasound (EUS) allows easy access to the left lobe and central right lobe liver segments. EUS-guided drainage can be a minimally invasive alternative approach in such patients.[2] The available preliminary data suggest EUS-guided drainage to be safe, effective, and technically feasible. However, experience with EUS-guided drainage is limited with no guidelines till date. The objective of this study was to share our experience in EUS-guided drainage of radiologically inaccessible liver abscesses.

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.

Conclusion

EUS-guided drainage with its excellent safety profile and clinical success should always be considered for draining radiologically inaccessible liver abscesses before contemplating surgical drainage. 10F NCD works well as endoprosthesis with excellent results irrespective of approach, location, thick abscess contents, and in ruptured abscesses. EUS-guided trans-segmental drainage is technically feasible and clinically effective.

Conflict of Interest

None declared.

Acknowledgments

None.

Earlier Presentation

• Oral abstract in ISGCON VARANASI 2024.

• Asian EUS cup in Asian EUS Congress June 2023.

• Video abstract (Abstract ID 1353–891) in APDW 2021.

• Abstract ENDO 2022 Kyoto Japan published in DEN 2022.

• Case series Endoscopy Int open 2021;9(1):E 35–40.

Study was conducted as per the guidelines in the Declaration of Helsinki.

Ethics Statement

The study was performed in accordance with the Code of Ethics of the World Medical Association.

Approval of Research Protocol by an Institutional Review Board

N/A.

Inform Consent

Yes.

Registry and Registration No. of the Study

N/A.

Animal Studies

N/A.

• References

• 1 Sharma S, Ahuja V. Liver abscess:complications and treatment. Clin Liver Dis (Hoboken) 2021; 18 (03) 122-126
• 2 Chin YK, Asokkumar R. Endoscopic ultrasound-guided drainage of difficult-to-access liver abscesses. SAGE Open Med 2020; 8: 2050312120921273
• 3 Seewald S, Imazu H, Omar S. et al. EUS-guided drainage of hepatic abscess. Gastrointest Endosc 2005; 61 (03) 495-498
• 4 Ang TL, Seewald S, Teo EK, Fock KM, Soehendra N. EUS-guided drainage of ruptured liver abscess. Endoscopy 2009; 41 (2, Suppl 2): E21-E22
• 5 Noh SH, Park DH, Kim YR. et al. EUS-guided drainage of hepatic abscesses not accessible to percutaneous drainage (with videos). Gastrointest Endosc 2010; 71 (07) 1314-1319
• 6 Itoi T, Ang TL, Seewald S. et al. Endoscopic ultrasonography-guided drainage for tuberculous liver abscess drainage. Dig Endosc 2011; 23 (Suppl. 01) 158-161
• 7 Keohane J, Dimaio CJ, Schattner MA, Gerdes H. EUS-guided transgastric drainage of caudate lobe liver abscesses. J Interv Gastroenterol 2011; 1 (03) 139-141
• 8 Ivanina E, Mayer I, Li J. et al. EUS- guided drainage of hepatic abscess. Gastrointest Endosc 2012; 75 (04) AB114
• 9 Medrado BF, Carneiro FO, Vilaça TG. et al. Endoscopic ultrasound-guided drainage of giant liver abscess associated with transgastric migration of a self-expandable metallic stent. Endoscopy 2013; 45 (Suppl. 02) E331-E332
• 10 Alcaide N, Vargas-Garcia AL, de la Serna-Higuera C, Sancho Del Val L, Ruiz-Zorrilla R, Perez-Miranda M. EUS-guided drainage of liver abscess by using a lumen-apposing metal stent (with video). Gastrointest Endosc 2013; 78 (06) 941-942
• 11 Kawakami H, Kawakubo K, Kuwatani M. et al. Endoscopic ultrasonography-guided liver abscess drainage using a dedicated, wide, fully covered self-expendable metallic stent with flared-end. Endoscopy 2014; 46 (01) 982-983
• 12 Koizumi K, Masuda S, Uojima H. et al. Endoscopic ultrasound-guided drainage of an amoebic liver abscess extending into the hepatic subcapsular space. Clin J Gastroenterol 2015; 8 (04) 232-235
• 13 Kodama R, Saegusa H, Ushimaru H, Ikeno T, Makino M, Kawaguchi K. Endoscopic ultrasonography-guided drainage of infected intracystic papillary adenocarcinoma of the liver. Clin J Gastroenterol 2015; 8 (05) 335-339
• 14 Tonozuka R, Itoi T, Tsuchiya T. et al. EUS-guided drainage of hepatic abscess and infected biloma using short and long metal stents (with videos). Gastrointest Endosc 2015; 81 (06) 1463-1469
• 15 Ogura T, Masuda D, Saori O. et al. Clinical outcomes of endoscopic ultrasound-guided liver abscess drainage using self-expandable covered metallic stent (with video). Dig Dis Sci 2016; 61 (01) 303-308
• 16 Yamamoto K, Itoi T, Tsuchiya T, Tanaka R, Nagakawa Y. EUS–guided drainage of hepatic abscess in the right lobe of a liver of a patient with chilaiditi syndrome. VideoGIE 2017; 2 (11) 299-300
• 17 Carbajo AY, Brunie Vegas FJ, García-Alonso FJ. et al. Retrospective cohort study comparing endoscopic ultrasound-guided and percutaneous drainage of upper abdominal abscesses. Dig Endosc 2019; 31 (04) 431-438
• 18 Rana SS, Ahmed S, Sharma R, Gupta R. Safety and efficacy of EUS-guided drainage of liver abscess: a single-center experience. Endosc Ultrasound 2020; 9 (05) 350-351
• 19 Chandra S, Chandra U. Endoscopic ultrasound-guided transgastric drainage of radiologically inaccessible left lobe liver abscess involving segment 4, caudate lobe, and left lateral segments using a modified technique. Endosc Int Open 2021; 9 (01) E35-E40
• 20 Molinario F, Rimbaş M, Pirozzi GA. et al. Endoscopic ultrasound-guided drainage of a fungal liver abscess using a lumen-apposing metal stent: case report and literature review. Rom J Intern Med 2021; 59 (01) 93-98
• 21 Zanwar S, Agarwal M. Endosonography-guided caudate lobe liver abscess drainage: a case and review of methods. J Dig Endosc 2022; 13: 119-124
• 22 Shah J, Jena A, Singh AK, Jearth V, Dutta U. Endoscopic ultrasound (EUS)-guided drainage of caudate lobe abscess: a single center experience. Surg Laparosc Endosc Percutan Tech 2023; 33 (06) 682-687
• 23 Shahid H, Tyberg A, Sarkar A. et al. EUS-guided versus percutaneous liver abscess drainage: a multicenter collaborative study. Endosc Ultrasound 2023; 12 (06) 472-476

Publication History

Article published online:
19 March 2025