The undrained abdominal abscess is associated with a high degree of mortality.¹ Traditionally, abdominal and pelvic abscesses have been drained by percutaneous techniques or surgically. Of these, percutaneous drainage is usually recommended as the first-line treatment as it is minimally invasive and requires a shorter duration of hospitalization compared to surgery.² Theoretically, endoscopic ultrasound (EUS)-guided abscess drainage has the same advantages as percutaneous drainage and also eliminates the need for an external drain. EUS-guided drainage has recently become the standard treatment for peripancreatic fluid collection;³ however, few studies have evaluated its efficacy for drainage of non-peripancreatic fluid collections. Here, we review the current status of EUS-guided abscess drainage at three sites: the liver, pelvis, and mediastinum.

Liver abscesses occur most commonly in elderly patients and are associated with malignant biliary obstruction, biliary disease, hepatic trauma, bacteremia, and amebiasis. Most are drained percutaneously, although 14 case reports and three retrospective studies have been published regarding EUS-guided drainage (Table 1).^4–20 In these previous reports, the most common indication for EUS-guided liver abscess drainage was failed medical therapy and inability to drain the abscess percutaneously. Seewald et al⁴ first described EUS-guided liver abscess drainage in 2005, reporting that the entire left lobe and most of the central segments of the liver were well visualized and accessible using this technique. Compared with the percutaneous transhepatic approach, EUS-guided drainage has several advantages: (1) direct access to the cavity in the left lobe of the liver via the trans-gastric route, (2) clear visualization of interposed vessels using color Doppler ultrasonography, and (3) avoidance of transcutaneous infection. EUS-guided drainage of liver abscess appears safe in selected patients in whom a trans-gastric approach can be used to minimize the distance between the abscess and the ultrasound transducer. In most reports, the abscess was located in the left lobe or the caudate lobe, but it is not possible to visualize and access all liver segments by EUS. The right lobe should be visualized from the duodenum, as the long and rigid tips of EUS preclude maneuverability. However, right lobe drainage can be challenging using EUS. Yamamoto et al¹⁶ reported the usefulness of EUS-guided liver drainage for an abscess in the right lobe in which percutaneous drainage was difficult due to Chilaiditi syndrome. Although there are fewer reports of drainage in the right lobe than the left lobe, these have been considered successful.^14–17

Table 1 . Summary of Literature on Endoscopic Ultrasound-Guided Liver Abscess Drainage.

Author (year)	Abscesses (n)	Location	Approach	Drainage method	Technical success (%)	Clinical success (%)	Adverse events
Seewald et al (2005)4	1	Left lobe	TG	7 Fr NDT	100	100	None
Ang et al (2009)5	1	Left lobe	TG	10 Fr, 8 Fr PS	100	100	None
Noh et al (2010)6	3	Caudate lobe (2) Gastrohepatic space (1)	TG (2) TD (1)	7 Fr PS 7 Fr NDT	100	100	None
Itoi et al (2011)7	2	Caudate lobe Left lobe	TD TG	7 Fr PS 5 Fr NDT	100	100	None
Keohane et al (2011)8	2	Caudate lobe	TG	7 Fr, 10 Fr PS	100	100	None
Medrado et al (2013)9	1	Left lobe	TG	10 mm PCSEMS	100	100	Intra-abscess stent migration
Alcaide et al (2013)10	1	Left lobe	TG	10 mm LAMS, PS	100	100	None
Kawakami et al (2014)11	1	Left lobe	TG	16 mm dedicated wide FCSEMS	100	100	None
Koizumi et al (2015)12	1	Left lobe	TG	5 Fr NDT	100	100	None
Kodama et al (2015)13	1	Left lobe	TG	7 Fr PS, 6 Fr NDT → 10 mm FCSEMS	100	100	None
Tonozuka et al (2015)14	7	Left lobe (6) Right lobe (1)	TG (6) TD (1)	16 mm dedicated wide FCSEMS or 8–10 mm FCSEMS, 5 Fr or 6 Fr NDT	100	71.4	None
Ogura et al (2016)15	8	Left lobe (6) Right lobe (2)	TG (6) TD (2)	10 mm FCSEMS, 7 Fr PS	100	100	None
Yamamoto et al (2017)16	1	Right lobe	TD	5 Fr NDT	100	100	None
Carbajo et al (2019)17	9	Left lobe (4) Right lobe (5)	TG TD	FCSEMS LAMS	77.8	77.8	Details unknown
Venkatesh et al (2020)18	1	Caudate lobe	TG	7 Fr PS (2)	100	100	None
Chandra and Chandra (2021)19	3	Caudate lobe (1) Left lobe (2)	TG	8 Fr PS 8–10 Fr NDT	100	100	None
Molinario et al (2021)20	1	Left lobe	TG	20 mm LAMS, 8.5 Fr PS	100	100	None

The procedure for EUS-guided liver abscess drainage is the same as that for peripancreatic fluid collection. The liver abscess is visualized using a convex echoendoscope. The left lobe and the caudate lobe are visualized from the stomach, and the right lobe is visualized mainly from the duodenum. After the liver abscess is identified, it is punctured with a needle. The first reported case by Seewald et al⁴ used a 22G needle, but a 19G needle has been used in most other cases. A guidewire is coiled within the liver abscess and tract dilation is performed. Mechanical dilators and balloon dilators are mainly used for dilation. Electric cautery has been used in placement of a lumen-apposing metal stent (LAMS) or dedicated wide fully covered self-expanding metal stent (FCSEMS).^10,11,14,17 After tract dilation, stent placement is performed. Table 1 lists the stents used in previous reports. Until 2011, there were reports only of a plastic stent and nasocystic tube use. In 2013, Medrado et al⁹ reported using a partially covered self-expanding metal stent (PCSEMS). The combination of a pigtail plastic stent and FCSEMS became the drainage system used by most endoscopists. Recent studies have reported insertion of an FCSEMS. The review by Chin and Asokkumar²¹ stated that FCSEMSs have several advantages compared to plastic stents: (1) more rapid drainage due to the wider diameter of the metal stent, (2) minimized risk of pus leakage, (3) enhanced ability to remove solid necrotic debris, and (4) fewer repeated procedures. Although these theoretical advantages over plastic stents remain unvalidated, the review included a variety of stents and high levels of technical and clinical success were found in all cases.

There are no reports of difficulty with the drainage technique; however, large abscesses and the presence of solid components may require additional endoscopic drainage. Tonozuka et al¹⁴ reported a clinical success rate of 71.4% at the first session. In two patients with large abscesses and solid components, resolution was achieved after the second session. Direct endoscopic necrosectomy was required in one case, and the final clinical success rate was 100%. Few adverse events and no fatalities have been associated with EUS-guided liver abscess drainage. Medrado et al⁹ reported a case of intra-abscess stent migration in which trans-gastric drainage was performed using a 60 mm × 10 mm PCSEMS, with no adverse events. At 2 weeks after the procedure, EUS revealed a reduction in the size of the liver abscess and intra-abscess stent migration. A 10-Fr double pigtail stent was inserted inside the SEMS to maintain the drainage channel, and there was full clinical resolution by 8 weeks. Carbajo et al¹⁷ reported three cases of gastrointestinal bleeding that were managed conservatively and one perforation that was closed during the procedure. However, as this report included other upper abdominal abscesses, it was unclear whether adverse events had occurred in the patients with liver abscesses. Tonozuka et al¹⁴ reported spontaneous migration of an FCSEMS into the digestive tract without any adverse events. Stent migration is a limitation of FCSEMSs, although the most recent LAMSs have an anti-migratory system to prevent its occurrence. The main problem is internal stent migration, and there is no clinical problem with stent dislocation after fistula formation and clinical resolution of the abscess. Stent removal, when necessary, was performed without difficulty after resolution of the abscess.

Our case is shown in Fig. 1–3. A patient with pancreatic cancer complicated by infection had a large hepatic cyst in the medial segment of the left lobe. EUS-guided drainage was performed.

Figure 1. Contrast-enhanced computed tomography reveals a large hepatic cyst with early wall enhancement and a fluid-fluid level within the cyst.

Figure 2. Endoscopic ultrasound (EUS)-guided liver abscess drainage. (A) A pyogenic liver abscess is punctured with a 19G needle under EUS visualization from the stomach. (B) A 0.025-inch guidewire is coiled within the abscess. (C) After dilation with a tapered dilator, a 10 mm × 8 cm fully covered self-expanding metal stent (FCSEMS) and a 7-Fr double pigtail stent is inserted into the FCSEMS to prevent kinking.

Figure 3. Two weeks later, the fully covered self-expanding metal stent (FCSEMS) was removed, and the Double Pit stent was retained to maintain the tract patency. Computed tomography obtained six weeks later revealed complete shrinkage of the cavity (arrowhead).

Pelvic abscess is a relatively common complication of obstetric and colorectal surgery. And can also occur in Crohn’s disease and diverticulitis. The management of pelvic abscess can be technically challenging due to anatomical considerations. Traditionally, pelvic abscess is treated by surgery or percutaneous drainage. Numerous other therapeutic modalities have been described, including transrectal, transvaginal, and transgluteal drainage.^22–24 Table 2 summarizes the literature on EUS-guided pelvic drainage,^25–36 for which reports are much more numerous compared with other areas such as the liver and mediastinum.

Table 2 . Summary of Literature on Endoscopic Ultrasound-Guided Pelvic Drainage.

Author (year)	Abscesses (n)	Approach	Drainage method	Technical success (%)	Clinical success (%)	Adverse events
Giovannini et al (2003)25	9	TR	8.5 Fr/10 Fr PS	100	88.8	Abdominal pain (1) Fever (2)
Attwell et al (2003)26	1	TS	Two 5 Fr PS	100	100	Fever
Varadarajulu and Drelichman (2007)27	4	TR/TS	10 Fr drain	100	75	None
Trevino et al (2008)28	4	TR	7 Fr PS and 10 Fr drain	100	100	None
Varadarajulu and Drelichman (2009)29	25	TR/TS	7 Fr PS in 25, plus 10 Fr drain in 10	100	96	None
Ramesh et al (2013)30	38	TR (27) TC (11)	PS ± drain	TR/TC 100/100	TR/TC 96.3/70	None
Hadithi and Bruno (2014)31	8	TR (6) TS (2)	One or two 7 Fr PS	100	100	None
Ratone et al (2016)32	7	TR	Two 7 Fr PS	100	100	None
Javed and Ho (2016)33	1	TR	LAMS	100	100	None
Poincloux et al (2017)34	37	TR (34) TS (4)	Aspiration only in 4, one or more PS in 29, LAMS in 4	100	91.9	Perforation (1) Stent migration (1) Rectal discomfort (1)
Mudireddy et al (2018)35	8	TR	LAMS	100	87.5	Details unknown
Donatelli et al (2021)36	7	TR	LAMS	85.7	85.7	None

Giovannini et al²⁵ first described EUS-guided drainage and aspiration of pelvic abscesses in 2003. They reported successful transrectal stent insertion into the fluid collection in nine patients, among whom complete drainage was achieved without relapse in eight and surgical drainage was performed in one who had incomplete drainage. In three patients, only aspiration was possible; of these, two developed recurrence of the abscess and required surgical treatment. Attwell et al²⁶ described EUS-guided drainage for diverticular abscess as a bridge to surgery. Varadarajulu and Drelichman²⁷ reported that transrectal stents can clog easily, particularly with fecal matter or pus, and when left long term can cause perirectal pain, migrate spontaneously, or cause retrograde infection. In 2007, they reported the usefulness of EUS-guided pelvic drainage using a drainage catheter. A combined technique using an EUS-guided transrectal drainage catheter and a transmural pigtail stent placement was adopted in 2008.²⁸ A short-term (36–48 hours) drainage catheter provided access for continued abscess irrigation, and a medium-term (2 weeks) stent maintained a patent transmural tract to ensure abscess resolution. Abscesses were resolved in all four patients. The efficacy of this combined approach was prospectively validated in a cohort of 25 patients.²⁹ The procedures were technically successful in all patients, the treatment success rate was 96%, and there were no adverse events. The combined technique minimizes the possibility of accidental dislocation of the drainage catheter and enables continuous irrigation, and has the additional advantages of minimizing patient discomfort and facilitating quick discharge of patients from hospital after removal of the drainage catheter. The 2013 study by Ramesh et al³⁰ found no difference between transcolonic and transrectal drainage in terms of rates of technical success, clinical success, or adverse events. When evaluated by etiology, treatment success was significantly lower for diverticular abscess compared with others (25% vs 97%; P = 0.002). In two case series of EUS-guided pelvic abscess drainage with one or more 7-Fr plastic pigtail stents,^31,32 technical success and clinical success were 100% in each, without adverse events. All reports so far have used drainage catheters with or without double pigtail plastic stents. In 2016, Javed and Ho³³ reported using a LAMS for successful transluminal drainage of perirectal abscess. Interval imaging one month later showed resolution of the abscess and the stent was removed successfully. Poincloux et al³⁴ reported long-term outcomes of EUS-guided pelvic drainage in 2017. Of 37 consecutive patients who were treated for perirectal or perisigmoid abscess via EUS-guided aspiration or drainage using a plastic stent or LAMS, the technical and clinical success rates were 100% and 91.9%, respectively. One patient required surgery and one required best supportive care due to persistent abscess. Early adverse events were perforation requiring surgery (n = 1), stent migration (n = 1), and rectal discomfort (n = 1). At a median follow-up of 64 months, two patients had experienced abscess recurrence. In recent years, two studies have reported transrectal LAMS for EUS-guided pelvic drainage in multiple patients.^35,36

In this way, EUS-guided pelvic drainage has been performed by various and nonstandardized methods, such as needle aspiration, aspiration with or without tract dilation, and placement of plastic and/or LAMS. Double pigtail plastic stents were used in all cases. It is recommended that double pigtail plastic stents be as short as possible, particularly if drainage is done in the lower rectum, because of the risk of discomfort if the stent touches the anal canal.^32,34 The risk of leakage must also be considered when using plastic stents. Recent reports suggest that LAMS placement is advantageous in terms of preventing leaks and discomfort.^33–37 Of four patients treated with LAMS in the study by Poincloux et al,³⁴ one developed recurrence and one with diverticular abscess had a perforation. EUS-guided pelvic drainage offers a viable alternative that can minimize the need for surgical intervention. It is noteworthy that higher complication rates have been reported for abscesses with diverticular etiology.³⁰ There are no reports of tubular FCSEMS for EUS-guided pelvic drainage. When performing EUS-guided pelvic drainage, it is necessary to consider the etiology when selecting an appropriate drainage method.

Our case is shown in Fig. 4–6. A patient with terminal pancreatic cancer developed a pelvic abscess that showed no clinical improvement with antibiotic therapy. EUS-guided transrectal drainage was performed. The stents were retained and after infection control, the patient was able to spend remaining time at home.

Figure 4. Contrast-enhanced computed tomography reveals an abscess anterior to the rectum (arrowhead).

Figure 5. Endoscopic ultrasound (EUS)-guided pelvic drainage of the same abscess. (A) The abscess is punctured with a 19G needle under forward-viewing EUS visualization from the rectum (arrowhead). (B) A 0.025-inch guidewire is coiled within the abscess. (C) After dilation with an electric cautery, a 10 mm × 8 cm fully covered self-expanding metal stent (FCSEMS) and a 7-Fr double pigtail stent are inserted inside the FCSEMS to prevent kinking.

Figure 6. Follow-up computed tomography scan of the same patient was obtained the next day and showing shrinkage of the abscess (arrowhead).

Mediastinal abscesses with uncontrolled infection can be fatal. Surgery has been used as the first-line treatment, but there have been many reports over the years of the diagnostic usefulness of EUS-guided fine needle aspiration via the transesophageal approach for mediastinal lesions.^38–40 However, there are still few reports regarding the usefulness of EUS-guided mediastinal drainage: only seven case reports and one prospective study (Table 3).^41–47 Fritscher-Ravens et al⁴⁸ first reported the successful aspiration of a mediastinal abscess using EUS in 2000, and Kahaleh et al⁴¹ first reported EUS-guided mediastinal drainage using a 7-Fr single pigtail stent in 2004. In 2005, Wehrmann et al⁴² reported endoscopic debridement of para-esophageal and mediastinal abscesses in a prospective case series of 20 patients. Simple drainage with an 8.5-Fr double pigtail stent was sufficient in 4 cases, puncture was impossible in 1 case, and debridement was performed after drainage in the remaining 15. In this report, both EUS-guided and endoscopic approaches were used for drainage, but no further breakdown of other details was provided. The clinical success rate was 100%, but one patient died from a massive pulmonary embolism one day after successful debridement. No other adverse events were observed. In all other case reports, treatment was by EUS-guided drainage only. Use of a pigtail plastic stent has been described in four reports,^37,41,43,45 and use of a nasobiliary drainage tube (NDT) in two.^44,47 Shibuya et al⁴⁷ reported several advantages of external drainage with EUS-guided mediastinal drainage. First, irrigation with saline followed by drainage enables complete cleansing of the abscess cavity, and drainage failure can be prevented by changing the position of the tube and the patient’s body. Second, culture samples can be taken repeatedly. Third, pharyngeal discomfort is less likely to occur with external drainage than with internal drainage. Fourth, it is easy to perform external drainage without substantial technical skill even in narrow working spaces, such as very close to the upper esophageal sphincter. In addition, an oral elemental diet can be ingested with monitoring via the NDT. However, external drainage using NDT is limited by the diameter of the drainage route. The limitation in diameter can lead to treatment failure through disruption of necrotic tissue. Consiglieri et al⁴⁶ reported the usefulness of 10-mm LAMS. Balloons and tapered dilators are the main devices used for dilation, and there are no reports of electric cautery for EUS-guided mediastinal drainage. Two case reports, including one from our institution, have described a transgastric approach to the mediastinum.^37,45 High technical and clinical success rates were reported in all cases, although Piraka et al⁴³ reported pneumothorax and mediastinitis in one patient who underwent drainage alone. In addition, pneumoderma, mediastinal emphysema, and bilateral pneumothorax developed in one of our patients who underwent transesophageal EUS-guided rendezvous.⁴⁹ Although this is a different case to abscess drainage, when performing EUS-guided mediastinal drainage it is important to confirm that the abscess is firmly encapsulated and separated from the pleural cavity. There are few reports of EUS-guided mediastinal drainage, and it is necessary to consider the advantages and disadvantages when selecting a drainage method.

Table 3 . Summary of Literature on Endoscopic Ultrasound-Guided Mediastinal Drainage.

Author (year)	Abscesses (n)	Approach	Drainage method	Technical success (%)	Clinical success (%)	Adverse events
Kahaleh et al (2004)41	1	TE	7 Fr PS (single pigtail)	100	100	None
Wehrmann et al (2005)42	20	TE	8.5 Fr PS alone in 4, plus debridement in 15	95	100	Pulmonary embolism (1)
Piraka et al (2009)43	2	TE	Two 10 Fr PS (double pigtail)	100	100	Pneumothorax (1) Mediastinitis (1)
Mahady et al (2011)44	1	TE	7 Fr NDT	100	100	None
Saxena et al (2014)45	1	TG	Two 7 Fr PS (double pigtail)	100	100	None
Consiglieri et al (2015)46	1	TE	10 mm LAMS	100	100	None
Shibuya et al (2019)47	1	TE	6 Fr NDT	100	100	None
Aritake et al (2021)37	1	TG	Three 7 Fr PS (double pigtail)	100	100	None

Our case is shown in Fig. 7–9. A patient with esophageal cancer developed a mediastinal abscess following perforation of the esophagus. There was no clinical improvement with antibiotic therapy, and EUS-guided transesophageal mediastinal drainage was considered as a bridge to surgery. Surgery was performed after infection control.

Figure 7. Endoscopic ultrasound (EUS)-guided drainage of the same mediastinal abscess. (A) EUS depicts the abscess deep within an esophageal tumor (arrowhead). (B) The abscess is punctured with a 19G needle, and a 0.025-inch guidewire is coiled inside the abscess. (C) After dilation with a tapered dilator, a 6-Fr nasobiliary drainage tube is inserted.

Figure 8. Coronal computed tomography reveals air within a mediastinal abscess (arrowhead).

Figure 9. Follow-up coronal computed tomography of the same patient obtained 8 days after mediastinal drainage shows the position of the nasobiliary drainage tube and resolution of the abscess (arrowhead).

We have reviewed the current status of EUS-guided abscess drainage. Although there are no comparative studies with percutaneous drainage or surgery, EUS-guided abscess drainage has high clinical and technical success rates without serious adverse events. It is necessary to consider the advantages and disadvantages when selecting a drainage method.

None.

Dr. Mizuno reports the following grants, none of which are connected with the submitted work: grants from Yakult Honsha, AstraZeneca, Novartis, MSD, ASLAN Pharmaceuticals, Incyte, Ono Pharmaceuticals, Seagen, Taiho Pharmaceutical and Eisai; personal fees from Yakult Honsha, AstraZeneca, Novartis, FUJIFILM Toyama Chemical and MSD.