Skip to main content
NCBI home page
Therapeutic Advances in Gastroenterology logoLink to Therapeutic Advances in Gastroenterology
. 2016 May 1;9(4):635–647. doi: 10.1177/1756283X16645050

Endoscopic ultrasound-assisted gastrointestinal hemostasis: an evolving technique

Deepanshu Jain 1, Nirav Thosani 2, Shashideep Singhal 3,✉
PMCID: PMC4913341  PMID: 27366229

Abstract

Gastrointestinal bleeding can range from grossly visible blood in stool or vomitus to microscopic bleed. However, any kind of bleeding can lead to potential life-threatening consequences. A small proportion of patients with gastrointestinal bleeding remain refractory to initial endoscopic hemostasis. While some are successfully managed by repeat endoscopic intervention, a few fail to respond or are not amenable to endoscopic hemostasis. As of now, the next level of intervention is passed on to either surgeons or interventional radiologists. There is new evidence suggesting the increased utility of endoscopic ultrasound (EUS) in diagnosis and treatment of culprit vascular lesions across the gut. In addition, EUS-assisted technique has also been used in the primary prevention of bleeding from gastroesophageal varices. In this review article, we have summarized case series and reports describing the use of EUS-assisted hemostasis. Indications, techniques, complications and success rates reported are discussed. While most of the authors describe their experience with primary and secondary treatment of gastric varices, treatment of other gastrointestinal lesions with EUS assisted hemostatic techniques is also discussed.

Keywords: coil, endoscopic ultrasound, gastrointestinal bleeding, glue, hemostasis, injection, technique

Introduction

Endoscopic ultrasound (EUS), first introduced by Dr Eugene DiMagno in the 1980s, increased endoscopists’ ability to better visualize the internal pathology via high-frequency ultrasound transducer [DiMagno et al. 1980]. Over time, EUS has evolved from a diagnostic to a therapeutic or prophylactic modality. Gastrointestinal (GI) bleeding refractory to initial endoscopic hemostasis accounts for up to 15–20% of total cases [Vazquez-Sequeiros and Olcina, 2010]. These lesions may be amenable to more advanced interventions such as angiography, coil embolization, glue application, or transjugular intrahepatic portosystemic shunt (TIPS) [Levy and Chak, 2009]. More recently, endoscopists have reported their experience of using these treatment modalities with EUS guidance. With EUS the identification of the bleeding vessel is easier and helps in documenting the success of endoscopic therapy [Levy and Chak, 2009; Wong, 2004; Wong et al. 2007; Kuramochi et al. 2007]. There is evidence to suggest that Doppler ultrasound is a better predictor of risk of rebleeding than endoscopic stigmata [Levy and Chak, 2009]; specifically, the absence of a Doppler ultrasound signal after therapy has been associated with a low risk of rebleeding [Kohler and Riemann, 1993]. The use of EUS allows endoscopists to visualize the bleeding vessel in real time (Figure 1), deliver the therapy precisely (Figure 2) and confirm the cessation of blood flow (Figure 3). In this review article, we have summarized case series and reports describing the use of EUS-assisted hemostasis. Indications, techniques, limitations and complications reported are discussed.

Figure 1.

Figure 1.

Open in a new tab

Tumor with Doppler showing vessel.

Figure 2.

Figure 2.

Open in a new tab

Glue injection using a 25 gauge fine aspiration needle.

Figure 3.

Figure 3.

Open in a new tab

Obliteration of vessel within tumor after glue injection.

Materials and methods

An extensive English language literature search was done through November 2015 using PubMed, Medline and Google to identify the peer-reviewed original and review articles using key words endoscopic ultrasound and hemostasis. Human-only articles were selected. The references of pertinent studies were manually searched to identify additional relevant studies. The indication, procedural details, technical and clinical success rate, complications and limitations for each study were reviewed.

Results

Only 13 original articles were considered appropriate to be included in the review article. Out of them, five were case reports from India [Sharma and Somasundaram, 2010; Rana et al. 2011], the UK [Krystallis et al. 2012] and the US [Kinzel et al. 2014; Ribeiro et al. 2001]. There were three case series from France [Gonzalez et al. 2012], Spain [Romero-Castro et al. 2007] and the US [Lahoti et al. 2000]. Two retrospective studies (analyzing prospectively maintained databases) from the US [Binmoeller et al. 2011] and Spain [Romero-Castro et al. 2013]; two retrospective [Law et al. 2014; Bhat et al. 2015] from the US and one prospective study [Seewald et al. 2008] from China were also included in the review. All studies have been summarized in Table 1.

Table 1.

Descriptive summary of each study.

Study no. Author/ year/ location Number of patients Diagnosis Indication Intervention Active bleeding at time of EUS-guided intervention Hemostasis achieved Obliteration of blood flow Rebleeding Procedure-related complication (other than rebleeding)
1 Sharma et al. [2010] India 1 Varix (single) Hemorrhagic shock and lack of localization of bleeding vessel EUS-guided histoacryl glue injection Y Y Decrease in flow reported, incomplete obliteration N N
2 Rana et al. [2011] India 1 Duodenal varix Primary treatment EUS guided cyanoacrylate glue injection N N/A Y N N
3 Krystallis et al. [2011] UK 1 Varices Hemorrhagic shock, lack of localisation of bleeding vessel, Unsuitable candidate for transjugular intra-hepatic porto-systemic shunt (TIPS) EUS guided thrombin injection (seven sessions of active trt over a span of 18 months) Y Y DNA N N
4 Kinzel et al. [2014] USA 1 Varix (single) Failure of ethanolamine oleate injection therapy EUS guided coil and 2-octyl-cyanoacrylate injection Y Y Y N N
5 Ribeiro et al. [2001] US 1 Dieulafoy lesion (single) Failure of rubber-band ligation 20 MHz catheter US probe (via therapeutic endoscope) guided banding ( on first visit), Echoendoscope guided thermal contact therapy (7 F Bicap probe) and absolute alcohol injection (on second visit) Y (on first visit), N (on second visit) Y (on first visit), N/A (second visit) DNA (first visit), Y (second visit) N N
6 Gonzalez et al. [2011] France 8 1. Gastric varix- 3/8
2. Gastroduodenal artery aneurys/fundal aneurysmal arterial malformation: 3/8
3. Dieulafoy lesion: 2/8		 1. Refractory bleeding: Failed endoscopic hemostasis: 6/8
2. Primary treatment		 EUS guided-
1. Cyanoacrylate + lipiodol mixture injection: 6/8
2. Polidocanol injection: 2/8		 Y (7/8)
N (1/8)		 Y: 7/7 Y: 7/8
N: 1/8 (rebleed at 3 months requiring second EUS session)		 Y: 1/8 rebleed at 3 months, required second EUS session, beyond that no recurrence N
7 Romero-Castro et al. [2006] Spain 5 Varices Pilot study EUS-guided N-butyl-2-cyanoacrylate-lipiodal mixture injection N N/A 100% after 1.6 sessions (mean) N N
8 Lahoti et al. 1999, U.S.A 5 Varices Pilot Study EUS guided sclerotherapy with Sodium morrhuate N N/A 100% after 2.2 sessions (mean) N Y- 1/5
1. Esophageal stricture- responsive to dilation treatment
9 Binmoeller et al. 2011, U.S.A 30 Varices (1.3/ patient) Varix > 1cm in size on EUS, active or recent bleeding attributed to varices, poor candidate or failure of TIPSS EUS guided coil and 2-octyl-cyanoacrylate (CYA) glue treatment Y (2/30) Y 23/24 after one session, one patient required second session 4/24 (6 lost to follow up) (all were nonvariceal) N
10 Romero-castro et al. 2013, Spain 11 Varices TIPSS failure/ poor candidate/ patient refusal EUS guided Coiling N N/A 90.9% after 1.3 sessions (mean) 1/11 (secondary to esophageal varices) N
19 Varices TIPSS failure/ poor candidate/ patient refusal EUS guided N-butyl-2-cyanoacrylate and lipiodal mixture injection Y (10/19) Y 100% after 1.5 sessions (mean) N Y- 11/19
1. Chest pain- 1/19
2. Fever- 1/19
3. Pulmonaary embolism- 9/19
11 Law et al. 2014, U.S.A 17 1. Esophageal cancer- 1/17
2. Gastric Dieulafoy lesion-1/17
3. Gastrointestinal stromal tumor- 4/17
4. Marginal ulcer post RYGB- 1/17
5.Duodenal ulcer- 1/17
6. Duodenal Dieulafoy lesion- 1/17
7. Brunner gland hematoma- 1/17
8. Duodenal metastasis- 2/17
9. Pancreatic pseudoaneurysm- 1/17
10. Rectal arteriovenous malformation- 1/17
11. Rectal aberrant vessel- 2/17
12. Rectal invasion by prostate cancer- 1/17		 Failure of primary treatment/ unsuitable candidates for additional therapies-
1. Failed EGD (Range-1 to 4, Median-2.5)- 16/17
2.Failed endoscopic interventions- 15/17
3. Failed radiologic embolization- 4/17
4. Prior surgical intervention- 3/17		 EUS guided-
1. N-butyl-2-cyanoacrylate and lipiodal mixture injection (5/17)
2. Coil embolisation (4/17)
3. Hyaluronate (2/17)
4. Ethanol (3/17)
5. Band ligation (1/17)
6. Combination of band ligation & ethanol (1/17)
7. Combination of epinephrine, snare ligature and polypectomy (1/17)		 14/17 Y- 13/14
N- 1/14		 1. Complete cessation-
a.) Single session- 10/17
b.) Two sessions- 1/17
2. Decreased blood flow- 5/19
3. DNA- 1/17		 Y- 3/17
a. 1/17- rebleed after 38 months (gastric Dieulafoy lesion)
b. 1/17- daily bleed despite EUS guided therapy (rectal invasion by prostate cancer)
c. 1/17- rebleed multiple times (rectal aberrant artery)		 N
12 Bhat et al. 2015, U.S.A 152 Gastroesophageal varices (GEV)-
1. GEV type 2- 9/152
2. Isolated gastric varices (IGV)- 143/152		 1. Primary treatment-
a) Active bleeding- 7/152 (5%)
b) History of bleeding or stigmata of recent hemorrhage- 105/152 (69%)
2. Prophylactic treatment (never bled)- 40/152 (26%)		 EUS guided coil and CYA glue treatment Y (7/152) Y- 7/7 EUS confirmed obliteration of varix- 93/100
(Number of patients with EUS follow up- 100/152)		 Y- 20/125 (16%)-
1. Secondary to gastric varix- 10/20 (50%) (4/10 secondary to extrusion of coil and glue complex)
(Number of patients with at least 1 month F/U- 125)		 Y- 5/125 (with at least 30 day follow up)
1. Abdominal pain- 4/125 (3%)
2. Symptomatic pulmonary embolization (1 week post procedure)- 1/125 (1%)
13 Lee et al. 2000, China 47 (On demand group- OD) Gastroesophageal varices (GEV)-
1. GEV type 1- 16/47
2. GEV type 2- 20/47
3. Isolated gastric varices (IGV)- 11/47		 Primary treatment for cirrhotic patients presenting with bleeding from GV on endoscopy (July 1993 to June 1996) Endoscopic injection therapy with cyanoacrylate + lipiodol (free hand injection) Y (23/47)-
1. Active spurting- 19/47
2. Oozing- 4/47
N (24/47)
1. Red dot- 11/47
2. Fibrin clot- 13/47		 Y- 45/47 (95.7%)
N- 2/47 (required ballon tamponade and vasoactive agents)		 N/A Y- 33/47 (Total)
1. Secondary to GV- 27/47
a) Early GV bleed (<= 48 hr)- 6/47 (12.8%)
b) Late GV bleed (> 48 hr)- 21/47 (44.7%)		 Y- 9/47
1. Post injection ulcer- 8/47
2. Pulmonary emboli (unconfirmed)- 1/47
3. Aspiration- 0/47
4. Septicemia- 0/47
54 (Repeated Injection group- RI) Gastroesophageal varices (GEV)-
1. GEV type 1- 20/54
2. GEV type 2- 18/54
3. Isolated gastric varices (IGV)- 16/54		 Primary treatment for cirrhotic patients presenting with bleeding from GV on endoscopy (July 1996 to October 1998) EUS guided cyanoacrylate + lipiodol mixture injection Y (26/54)-
1. Active spurting- 20/54
2. Oozing- 6/54
N (28/54)
1. Red dot- 15/54
2. Fibrin clot- 13/54		 Y- 52/54 (96.3%)
N- 2/54 (required ballon tamponade and vasoactive agents)		 Y- 43/54 (79.6%) after 2.2 + 1.7 sessions (mean)
1. GEV 1- 14/20
2. GEV 2- 14/18
3. IGV- 15/16		 Y- 19/54 (Total)
1. Secondary to GV- 14/54
a) Early GV bleed (<= 48 hr)- 4/54 (7.4%)
b) Late GV bleed (> 48 hr)- 10/54 (18.5%)		 Y- 22/54
1. Post injection ulcer- 19/54 (6/19 were bleeding requiring topical spray therapy)
2. Pulmonary emboli (unconfirmed)- 1/54
3. Aspiration- 1/54
4. Septicemia- 1/54
Open in a new tab

Indications

Refractory bleeding is described in patients who have undergone first-line endoscopic intervention by upper endoscopy or colonoscopy but later present with overt or occult bleeding. Although the primary treatment modality for esophageal varices has been the banding or the use of beta-blockers, management of gastric, duodenal or rectal varices remains debatable. There is expanding evidence for EUS-assisted technique in achieving hemostasis for patients with refractory bleeding [Sharma et al Sharma and Somasundaram, 2010; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Law et al. 2014], primary prevention [Rana et al. 2011; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Lee et al. 2000] or secondary treatment of variceal bleeding [Gonzalez et al. 2012; Binmoeller et al. 2011; Romero-Castro et al. 2013; Bhat et al. 2015; Lee et al. 2000]. Two studies reported the use of EUS-assisted hemostasis in the setting of hemorrhagic shock secondary to nonlocalization of the bleeding vessel by routine endoscopic techniques [Sharma and Somasundaram, 2010; Krystallis et al. 2012]. Two authors presented their data on EUS-assisted hemostasis among patients with gastric varices, who either have large varices (>1cm) or have failed or were poor candidates for TIPS [Binmoeller et al. 2011; Romero-Castro et al. 2013]. A study by Law and colleagues reported the utility of EUS in achieving hemostasis in the setting of failed endoscopic therapy, radiological intervention and surgery [Law et al. 2014].

Site and etiology of gastrointestinal bleed

Both upper and lower GI bleeds are amenable to EUS-assisted hemostasis. Romero-Castro and colleagues reported two case series with 5 and 30 patients respectively, who were managed with EUS for underlying gastric varices [Romero-Castro et al. 2007; 2013]. Similarly, a case series of 30 patients undergoing EUS-assisted procedure for gastric varices (with a mean of 1.3 varices per patient) was reported by Binmoeller and colleagues [Binmoeller et al. 2011]. Two more authors reported their experience in using the EUS-assisted hemostasis technique as the prophylactic or secondary treatment for the gastroesophageal varices (GEV) among 152 and 54 patients respectively [Bhat et al. 2015; Lee et al. 2000]. Another study described the use of EUS in managing patients with esophageal varices in a case series of five patients [Lahoti et al. 2000].

In a case report, authors described a case of gastric Dieulafoy lesion with recurrent bleeding where EUS was helpful in both diagnosis and management of the lesion [Ribeiro et al. 2001]. Authors of another study reported the role of EUS in managing gastric lesions like varices, Dieulafoy lesions and arterial malformations [Gonzalez et al. 2012]. Sharma and Somasundaram described a patient with a bleeding rectal varix, controlled with the aid of EUS [Sharma and Somasundaram, 2010]. Three authors reported their experience with EUS-assisted treatment of duodenal varices [Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014]. Law and colleagues reported their case-by-case experience in achieving EUS-assisted hemostasis in a variety of lesions, ranging from those of oncogenic etiology like primary esophageal cancer, metastatic duodenal disease, gastric tumors like GI stromal tumors (GIST), prostate cancer eroding into rectum; vascular abnormalities like gastric and duodenal Dieulafoy lesions, aberrant rectal vessels and arteriovenous malformations; to those of miscellaneous etiology like marginal ulcer post Roux-en-Y gastric bypass (RYGB) procedure, pancreatic pseudoaneurysm and Brunner gland hematoma [Law et al. 2014].

Source of gastrointestinal bleed

There are multiple case series and case reports describing the efficacy of EUS-assisted hemostasis for venous bleeds [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015; Lee et al. 2000], with minimal data on safety and efficacy of EUS-assisted hemostasis for arterial bleeds [Ribeiro et al. 2001; Law et al. 2014].

Technique

Echoendoscope

Radial- [Rana et al. 2011; Krystallis et al. 2012] or linear- array [Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015] therapeutic echoendoscopes with color Doppler have been used by most endoscopists for localizing and targeting the feeding vessels.

Puncture needle

Both 19 gauge (G) [Kinzel et al. 2014; Gonzalez et al. 2012; Binmoeller et al. 2011; Romero-Castro et al. 2013; Bhat et al. 2015] and 22 G [Kinzel et al. 2014; Romero-Castro et al. 2007, 2013; Law et al. 2014; Bhat et al. 2015] fine needle aspiration (FNA) needles have been used most frequently for EUS-assisted coiling whereas a 22 G FNA needle has been preferred for EUS-assisted injection therapy (sclerosant, glue and desiccants).

Type of intervention

EUS increases endoscopists’ ability to better characterize the pathologic vascular lesion. It helps to localize the feeding blood vessel, which can be then targeted specifically. Authors have reported use of EUS-assisted injection therapy [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Law et al. 2014; Lee et al. 2000], coiling [Romero-Castro et al. 2013; Law et al. 2014], banding [Ribeiro et al. 2001; Law et al. 2014], thermal-contact therapy [Ribeiro et al. 2001] and combination therapies like glue injection and coiling [Kinzel et al. 2014; Binmoeller et al. 2011; Bhat et al. 2015], ethanol injection and band ligation [Law et al. 2014], or epinephrine injection along with snare ligation and polypectomy [Law et al. 2014]. A lot of diversity has been seen in the agent of use for the EUS-assisted injection therapy, ranging from glue therapy with histoacryl glue for rectal varix [Sharma and Somasundaram, 2010], N-butyl-2-cyanoacrylate for esophageal cancer [Law et al. 2014], gastroesophageal varix [Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lee et al. 2000], gastric GIST [Law et al. 2014], gastroduodenal artery malformation [Gonzalez et al. 2012], Dieulafoy lesion [Gonzalez et al. 2012], duodenal varix [Rana et al. 2011], duodenal ulcer [Law et al. 2014], and duodenal metastatic disease [Law et al. 2014]. 2-octyl-cyanoacrylate glue has been used successfully for duodenal and gastric varices [Kinzel et al. 2014; Binmoeller et al. 2011], while desiccants like alcohol have been used for duodenal metastatic disease [Law et al. 2014], pancreatic pseudoaneurysm [Law et al. 2014], rectum-invading prostate cancer [Law et al. 2014], and gastric Dieulafoy lesion [Ribeiro et al. 2001]. Sclerosants like sodium morrhuate have been used for esophageal varices [Lahoti et al. 2000], hyaluronate for gastric GIST [Law et al. 2014] and thrombin injections for duodenal varices [Krystallis et al. 2012]. There is a paucity of trials comparing the efficacy across different modes of EUS-assisted hemostasis. Romero-Castro and colleagues compared the EUS-assisted glue injection therapy against coiling as the method of achieving hemostasis and obliteration of blood flow for gastric varices. The study revealed that EUS-assisted coiling subgroup required fewer endoscopies and tended to have fewer adverse events than the glue-therapy subgroup [Romero-Castro et al. 2013].

The quantity of coil or injection agent required alone or in combination to achieve hemostasis is not uniform, as shown in Table 2. In most studies, the decision was made by the endoscopist, during procedure, based on the size of target lesion and Doppler findings of persistent blood flow after initial intervention [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015; Lee et al. 2000]. For large blood vessels, use of two coils has been described [Binmoeller et al. 2011; Romero-Castro et al. 2013; Law et al. 2014; Bhat et al. 2015]. In few cases, the endoscopists have used combination treatment with injection therapy and coils [Kinzel et al. 2014; Binmoeller et al. 2011; Bhat et al. 2015]. The number of required coils varied depending on whether the endoscopist had planned to use it alone [Romero-Castro et al. 2013; Law et al. 2014] or in combination with injection therapy [Kinzel et al. 2014; Binmoeller et al. 2011; Bhat et al. 2015] to achieve hemostasis, and the same principle applied for the quantity of injection agent [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015; Lee et al. 2000]. Romero-Castro and colleagues reported a wide range for the number of coils (2–13) required per patient to achieve hemostasis in the absence of concomitant injection therapy [Romero-Castro et al. 2013]. The size of coil was determined, based on the size of the target vessel on Doppler ultrasound. Romero-Castro and colleagues reported using a 20% larger-sized coil than the size of the target vessel [Romero-Castro et al. 2013], similar to that of Law and colleagues who reported using coils, 1.25 times the size of target-vessel diameter. Also, the size of coil has been reported to vary from as small as 6 mm [Law et al. 2014] to as large as 20 mm [Binmoeller et al. 2011; Romero-Castro et al. 2013]. The quantity of injection therapy also varies with the type of agent [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015; Lee et al. 2000] and has been reported to be as low as 0.2 ml of 99% alcohol [Law et al. 2014] or 0.5 ml of N-butyl-2-cyanoacrylate glue [Bhat et al. 2015] per session, to as high as 4 ml of sodium morrhuate [Lahoti et al. 2000], 7.5 ml of alcohol [Law et al. 2014] or 10 ml of epinephrine [Law et al. 2014]. For the concern of possible pulmonary embolism linked to glue therapy, Romero-Castro and colleagues restricted the amount of injection therapy to 1 ml per session [Romero-Castro et al. 2013].

Table 2.

Procedure specific details across each study.

S.No Author/ Year/ Location Type of Endoscope Size of Needle Injection material Number of patients Quantity/Concentration/Number of sessions
1 Sharma et al. 2010, India Color doppler rectal EUS Data Not Available (DNA) Histoacryl glue 1 1 ml, single session
2 Rana et al. 2011, India Radial echoendoscope DNA Cyanoacrylate (CYA) injection 1 1.5 ml of glue, single session
3 krystallis et al. 2011, U.K Radial color doppler EUS and double channel therapeutic endoscope DNA Human Thrombin 1 46 ml (250 U/ml) in 7 sessions over 18 months
4 kinzel et al. 2014, U.S.A Linear array echoendoscope 19 G FNA needle for coiling, 22 G FNA needle for sclerotherapy 2-octyl-cyanoacrylate and coiling 1 2 ml, single session, one coil deployed
5 Ribeiro et al. 2001, U.S.A Pulsed and color doppler with electronic scanning echoendoscope and linear array US endoscope DNA Absolute alcohol 1 2.5 ml, single session
6 Gonzalez et al. 2011, France Linear endoscope with doppler enhanced EUS 19 G FNA needle CYA + lipiodol mixture 6 2 ml, single session
Polidocanol injection 2 1/2 patient- 2% (4 ml), single session
1/2 patient- 2% (4 ml), two sessions
7 Romero-castro et al. 2006, Spain Linear array echoendoscope with pulsed and color doppler 22 G needle N-butyl-2-cyanoacrylate-lipiodal (1:1) mixture 5 1.6 ml (mean) (1-2 ml, range) over 1.6 (mean) sessions per patient [2/5 patients- single session, 3/5 patients- two sessions]
8 Lahoti et al. 1999, U.S.A Linear array scanning echoendoscope 2.5 mm catheter varijet injector needle Sodium morrhuate 5 2-4 ml per injection site, 2.2 sessions per patient (2-3, range)
9 Binmoeller et al. 2011, U.S.A Forward view curved linear array echoendoscope or therapeutic channel curvilinear array echoendoscope 19 G FNA needle 2-octyl-cyanoacrylate and coiling 30 1.4 ml (mean) per varix, 1.3 (mean) number of varix/patient, 23/24 patients- single session, 1/24 patient- 2 sessions
Every patient received coiling at each session, number of coils per session (not reported, endoscopist determined the need of second coil based on persitent doppler flow and size of remaining varix)
10 Romero-castro et al. 2013, Spain Linear array therapeutic echoendoscope 19 G FNA needle Coil 11 5.8 + 1.2 (mean) coils per patient (2-13, range), 9/11 patients- single session, 1/11 patient- two sessions, 1/11 patient- three session
Linear array therapeutic echoendoscope 22 G FNA needle N-butyl-2-cyanoacrylate-lipiodal (1:1) mixture 19 1 ml per session, 1.5 + 0.1 ml (mean) per patient (1- 3 ml, range), 10/19 patients- single session, 8/19 patients- two sessions, 1/19 patients- three session
11 Law et al. 2014, U.S.A Curvilinear therapeutic echoendoscope 22 G FNA needle N-butyl-2-cyanoacrylate-lipiodal (1:1) mixture 5 2-7 ml (total injected amount), single session
Coil 4 Native coil diameter- 1.25 times the target vessel diameter (6- 10 mm), 1-2 coils/patient, single session
Hyaluronate injection 2 1/2 patient- 1 ml, single session
1/2 patient- 3-4 ml/session, two sessions
Ethanol injection 3 0.2-7.5 ml (total injected amount), single session
Band ligation 1 6 bands, single session
Band ligation + Ethanol injection 1 4 ml of 99% alcohol injection + 3 bands, single session
Epinephrine injection + snare ligation + polypectomy 1 10 ml of 1:1000 epinephrine injection, single session
12 Bhat et al. 2015, U.S.A Forward viewing curved linear array echoendoscope or therpeutic channel oblique viewing CLA echoendoscope 19 G or 22 G Echotip CYA glue + Coiling 152 Mean number of coils (range)- 1.4 (1- 4)
Mean volume of CYA (range)- 2 ml (0.5- 6)
13 Lee et al. 2000, China Endoscope (no EUS) N/A CYA + lipiodol mixture 47 (On demand group- OD) Each injection was constituted from 0.5 ml of CYA + 0.7 ml of lipiodol
Median numbner of injections (range)- 2 (1-6)
Average number of sessions- 1.3 + 0.5
Echoendoscope and forward vieweing gastroscope DNA 54 (Repeated injection group- RI) Each injection was constituted from 0.5 ml of CYA + 0.7 ml of lipiodol.
Median numbner of injections (range)- 3 (1-8)
Average number of sessions- 2.2 + 1.7
Open in a new tab

The optimal amount of therapeutic agents for EUS-assisted hemostasis remains debatable and is driven by multiple endoscopic parameters.

Outcome

Hemostasis and obliteration of blood flow

Out of total 306 cases [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015; Lee et al. 2000], 71 cases were reported to have evidence of active bleeding at the time of EUS-assisted index therapeutic procedure. In 93% of the cases (66/71), hemostasis (stoppage of active bleeding) was achieved post procedure. Romero-Castro and colleagues reported obliteration of blood flow in five out of five gastric varices cases (100%) post EUS-assisted glue therapy after a mean of 1.6 sessions [Romero-Castro et al. 2007]. Similarly, complete cessation of blood flow in five out of five esophageal varices cases (100%) was achieved after a mean of 2.2 sessions of EUS-assisted sclerotherapy [Lahoti et al. 2000]. In another study, complete obliteration of blood flow in 23/24 gastric varices cases was reported after single session of combination treatment with EUS-assisted coiling and glue therapy, whereas 1/24 cases required two sessions of treatment [Binmoeller et al. 2011]. Romero-Castro and colleagues also reported complete cessation of blood flow in 19/19 gastric varices cases post EUS-assisted glue therapy after a mean of 1.5 sessions of treatment, in contrast to 10/11 gastric varices cases post EUS-assisted coiling after a mean of 1.3 sessions of treatment [Romero-Castro et al. 2013]. Out of five case reports, three authors reported complete cessation of blood flow post procedure [Rana et al. 2011; Kinzel et al. 2014; Ribeiro et al. 2001], one achieved only a decrease in blood flow [Krystallis et al. 2012] and one reported no documentation with this respect [Sharma and Somasundaram, 2010]. Law and colleagues reported complete cessation of blood flow in 10/17 patients (58.8%) after a single session of treatment, in 1/17 patients (5.9%) after two sessions of treatment and decrease in blood flow in 5/17 patients (29.4%) after a single session of treatment. In another study, complete cessation of blood flow was achieved after a single treatment session in all except one (7/8) that required a second session [Gonzalez et al. 2012]. Complete obliteration of GEV was achieved in 93/100 (93%) [Bhat et al. 2015] and 43/47 (91.5%) [Lee et al. 2000] patients with EUS-assisted intervention.

Follow up

Most authors report a variable follow up of the patients, ranging from a minimum of 3 weeks to a maximum of 120 months [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015; Lee et al. 2000]. After achieving initial hemostasis on index endoscopy, the endoscopist should re-evaluate the patients at shorter follow-up intervals with repeat EUS, until there is documented cessation of blood flow in the underlying feeding vascular supply and thereafter the follow-up interval can be increased and done based on any clinical or laboratory signs of bleeding.

Rebleeding

Out of 306 total patients [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015; Lee et al. 2000], 48 patients had a repeat episode of GI bleed post procedure. On further evaluation, only 28 patients (9.2%) had a rebleed from either the recurrence of primary etiology or lack of complete obliteration of blood flow from the therapeutic EUS-assisted treatment performed on the index procedure. In one study [Gonzalez et al. 2012], 1/8 patients (12.5%) experienced a rebleed, whereas in another study, the authors reported rebleeding in 3/17 patients (17.6%) [Law et al. 2014]. In both of these scenarios, the bleeding was from the same primary etiology and the authors were unable to achieve complete cessation of blood flow on index EUS-assisted intervention [Gonzalez et al. 2012; Law et al. 2014]. Bhat and colleagues reported rebleeding from the same primary etiology in 10/125 patients (8%). Only 3 out of 93 patients (3.2%) had rebleeding despite complete obliteration of gastric fundal varices (GFV) on index procedure secondary to recurrence of new GFV in the setting of persistent portal hypertension [Bhat et al. 2015]. Only 4/125 (3.2%) patients had rebleeding secondary to extrusion of the coil and glue from the primary lesion [Bhat et al. 2015]. Lee and colleagues reported rebleeding in 14/54 patients (25.9%) post EUS-assisted intervention for GFV. Only four patients had early rebleeding within 48 hours of the primary intervention and the rest occurred beyond 48 hours [Lee et al. 2000]. Authors reported that complete obliteration of GFV was achieved in 43/54 patients, out of which 10 (23.3%) had documented recurrence of GFV and seven (16.3%) presented with bleeding after a mean follow up of 14.8 + 10 months [Lee et al. 2000].

Cessation of blood flow in the feeding vessel for the target lesion seems to correlate with decreased risk of rebleeding from the primary lesion. Recurrence of primary etiology secondary to persistent underlying pathology is a risk for rebleeding.

Postprocedural complications (excluding gastrointestinal bleed)

Out of 306 cases [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007, 2013; Lahoti et al. 2000; Binmoeller et al. 2011; Law et al. 2014; Bhat et al. 2015; Lee et al. 2000], 39 patients (12.8%) were reported to have procedure-related complications. Of these, 11 patients (3.5%) were symptomatic, whereas the remaining 28 patients (9.2%) remained asymptomatic. In the case series of five patients who had EUS-assisted sclerotherapy for the esophageal varix, one patient developed esophageal stricture that was successfully managed with dilatation treatment [Lahoti et al. 2000]. Romero-Castro and colleagues reported that 11/30 patients had adverse events related to the EUS-assisted cyanoacrylate treatment [Romero-Castro et al. 2013]. Among these 11 patients, two were symptomatic [secondary to fever and chest pain (in the absence of pulmonary embolism)], whereas the remaining nine patients had asymptomatic pulmonary embolism (secondary to glue injection) diagnosed on follow-up CT chest [Romero-Castro et al. 2013]. In the study by Bhat and colleagues, out of 125 patients with at least a month of follow up, authors reported one case of symptomatic pulmonary embolism (treated with anticoagulation) and four patients with self-limited abdominal pain [Bhat et al. 2015]. Lee and colleagues reported postprocedural complication in 22/54 patients (40.7%). As many as 19 patients had ulceration at the injection site but all were asymptomatic. Only 6 out of 19 patients had bleeding from the ulcer and required topical adrenaline spray. Two patients required antibiotic therapy to treat aspiration pneumonia and septicemia, respectively [Lee et al. 2000]. One patient had postprocedural cough suspected to be secondary to pulmonary embolism but not confirmed [Lee et al. 2000]. There were no other adverse events reported in the rest of the studies [Sharma and Somasundaram, 2010; Rana et al. 2011; Krystallis et al. 2012; Kinzel et al. 2014; Ribeiro et al. 2001; Gonzalez et al. 2012; Romero-Castro et al. 2007; Binmoeller et al. 2011; Law et al. 2014].

Glue embolism is a rare but serious side effect of glue-injection therapy for hemostasis. In the past decade, many studies have reported that limiting the amount of glue agent (N-butyl 2-cyanoacrylate) to less than 1 ml per injection and diluting the glue agent with lipiodal solution decreases the incidence of embolic events [Seewald et al. 2008]. Slower flow rate secondary to a smaller glue volume injected in one session is a potential explanation for lower rate of embolism seen in such patients [Fujii-Lau et al. 2015]. Use of EUS has been postulated to help in minimizing the risk of embolic event by identifying the target-feeding blood vessel, thereby decreasing the amount of injected glue needed to achieve hemostasis. Use of concomitant therapy with EUS-assisted coiling helps reduce the incidence of embolism by providing a scaffold to retain the subsequently injected glue in addition to reducing the required volume of glue to achieve hemostasis [Binmoeller et al. 2011]. Despite following the recommended preventive measures, authors reported the incidence of glue embolism to be as high as 47.4% (9/19) [Romero-Castro et al. 2013]. Interestingly, all were asymptomatic requiring no treatment. Similarly, one case of symptomatic pulmonary embolism (requiring anticoagulation treatment) was reported, out of 125 GFV patients treated with EUS-assisted cyanoacrylate glue and coiling [Bhat et al. 2015]. This highlights the importance of closely following up the patients, postintervention. Also, no study has yet reported the long-term effect of glue embolism of pulmonary vasculature, that, if proven, will further highlight the importance of diagnosing it, even in the asymptomatic patients.

Another alternative to prevent glue embolism, is to replace it with alternative therapeutic options. EUS-assisted coiling is a potential replacement option in the near future. EUS-assisted coiling is comparable with EUS-assisted glue therapy for achieving hemostasis with the added advantage of absence of any procedure-related complications including embolism [Romero-Castro et al. 2013]. Technically speaking, endoscopists have reported that EUS-assisted coiling is more challenging than EUS-assisted injection therapy, although there was no significant difference in the respective procedure time [Romero-Castro et al. 2013]. EUS-assisted therapy for hemostasis is an evolving technique and specifically, experience with EUS-assisted coiling in the absence of concomitant injection therapy for hemostasis is still in its baby steps.

Summary and future directions

EUS-assisted technique to prevent or treat GI bleeding is a safe, minimally invasive and efficacious alternative approach to surgery or interventional radiology in selected patients who fail the standard endoscopic approach. With present techniques and accessories, management of gastric varices with the EUS-assisted approach seems to be a reasonable first-line approach. The technique has a potential role, not only in assisting in diagnosis by more accurate visualization of the bleeding vascular anomaly but also helps in targeted-injection therapy or placement of coil in the feeding vessels and finally in documenting the obliteration of flow in the bleeding vessel with help of the Doppler technique, resulting in decreased risk of rebleeding. Preliminary reports appear promising, but large multicentric prospective studies are needed in the future to adopt it as a standard alternative therapy for refractory GI bleeding and as primary treatment for gastric varices. With further experience and development of more sophisticated accessories, the arena of EUS-assisted hemostasis is likely to expand across different kinds of GI vascular lesions.

Footnotes

Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Conflict of interest statement: None of the authors has any conflicts of interest or financial relationship with the company that produces or distributes the device described in the review article.

Contributor Information

Deepanshu Jain, Department of Internal Medicine, Albert Einstein medical center, Philadelphia, PA, USA.

Nirav Thosani, Division of Gastroenterology, Hepatology and Nutrition, University of Texas Health Science Center at Houston, Houston, TX, USA.

Shashideep Singhal, Division of Gastroenterology, Hepatology and Nutrition, University of Texas Health Science Center at Houston, 6431 Fannin, MSB 4.234, Houston, TX 77030, USA.

References

  1. Bhat Y., Weilert F., Fredrick R., Kane S., Shah J., Hamerski C., et al. (2015) EUS-guided treatment of gastric fundal varices with combined injection of coils and cyanoacrylate glue: a large U.S. experience over 6 years (with video). Gastrointest Endosc [Epub ahead of print]. [DOI] [PubMed] [Google Scholar]
  2. Binmoeller K., Weilert F., Shah J., Kim J. (2011) EUS-guided transesophageal treatment of gastric fundal varices with combined coiling and cyanoacrylate glue injection (with videos). Gastrointest Endosc 74: 1019–1025. [DOI] [PubMed] [Google Scholar]
  3. DiMagno E., Buxton J., Regan P., Hattery R., Wilson D., Suarez J., et al. (1980) Ultrasonic endoscope. Lancet 8169: 629–631. [DOI] [PubMed] [Google Scholar]
  4. Fujii-Lau L., Law R., Wong Kee Song L., Gostout C., Kamath S., Levy M. (2015) Endoscopic ultrasound (EUS)-guided coil injection therapy of esophagogastricand ectopic varices. Surg Endosc July 3. [Epub ahead of print]. [DOI] [PubMed] [Google Scholar]
  5. Gonzalez J, Giacino C, Pioche M, Vanbiervliet G, Brardjanian S, Ah-Soune P, et al. (2012) Endoscopic ultrasound-guided vascular therapy: is it safe and effective? Endoscopy 44: 539–542. [DOI] [PubMed] [Google Scholar]
  6. Kinzel J., Pichetshote N., Dredar S., Aslanian H., Nagar A. (2014) Bleeding from a duodenal varix - a unique case of variceal hemostasis achieved using EUS-guided placement of an embolization coil and cyanoacrylate. J Clin Gastroenterol 48: 362–364. [DOI] [PubMed] [Google Scholar]
  7. Kohler B., Riemann J. (1993) Endoscopic injection therapy of Forrest II and III gastroduodenal ulcers guided by endoscopic Doppler ultrasound. Endoscopy 25: 219–223. [DOI] [PubMed] [Google Scholar]
  8. Krystallis C., McAvoy N., Wilson J., Hayes P., Plevris J. (2012) EUS-assisted thrombin injection for ectopic bleeding varices—a case report and review of the literature. Q J Med 105: 355–358. [DOI] [PubMed] [Google Scholar]
  9. Kuramochi A., Imazu H., Kakutani H., Uchiyama Y., Hino S., Urashima M. (2007) Color Doppler endoscopic ultrasonography in identifying groups at a high-risk of recurrence of esophageal varices after endoscopic treatment. J Gastroenterol 42: 219–224. [DOI] [PubMed] [Google Scholar]
  10. Lahoti S., Catalano M., Alcocer E., Hogan W., Geenen J. (2000) Obliteration of esophageal varices using EUS-guided sclerotherapy with color Doppler. Gastrointest Endosc 51: 331–333. [DOI] [PubMed] [Google Scholar]
  11. Law R., Fujii-Lau L., Wong-Kee-Song L., Gostout C., Kamath P., Abu-Dayyeh B, et al. (2014) Efficacy of endoscopic ultrasound guided hemostatic interventions for resistant nonvariceal bleeding. Clin Gastroenterol Hepatol 13: 808–12.e1 [Epub 19 September 2014]. [DOI] [PubMed] [Google Scholar]
  12. Lee Y., Chan F., Ng E., Leung V., Law K., Yung M., et al. (2000) EUS-guided injection of cyanoacrylate for bleeding gastric varices. Gastrointest Endosc 52: 168–174. [DOI] [PubMed] [Google Scholar]
  13. Levy M., Chak A. (2009) EUS 2008 Working Group document: evaluation of EUS-guided vascular therapy. Gastrointest Endosc 69: S37–S42. [DOI] [PubMed] [Google Scholar]
  14. Rana S., Bhasin D., Rao C., Singh K. (2011) Endoscopic ultrasound-guided treatment of bleeding duodenal varix. Ind J Gastroenterol 30: 280–281. [DOI] [PubMed] [Google Scholar]
  15. Ribeiro A., Vazquez-Sequeiros E., Wiersema M. (2001) Doppler EUS-guided treatment of gastric Dieulafoy’s lesion. Gastrointest Endosc 53: 807–809. [DOI] [PubMed] [Google Scholar]
  16. Romero-Castro R., Ellrichmann M., Ortiz-Moyano C., Subtil-Inigo J., Junquera-Florez F., Gornals J., et al. (2013) EUS-guided coil versus cyanoacrylate therapy for the treatment of gastric varices: a multicenter study (with videos). Gastrointest Endosc 78: 711–721. [DOI] [PubMed] [Google Scholar]
  17. Romero-Castro R., Pellicer-Bautista F., Jimenez-Saenz M., Marcos-Sanchez F., Caunedo-Alvarez A., Ortiz-Moyano C., et al. (2007) EUS-guided injection of cyanoacrylate in perforating feeding veins in gastric varices: results in 5 cases. Gastrointest Endosc 66: 402–407. [DOI] [PubMed] [Google Scholar]
  18. Seewald S., Ang T., Imazu H., Naga M., Omar S., Groth S., et al. (2008) A standardized injection technique and regimen ensures success and safety of N-butyl-2-cyanoacrylate injection for the treatment of gastric fundal varices (with videos). Gastrointest Endosc 68: 447–454. [DOI] [PubMed] [Google Scholar]
  19. Sharma M., Somasundaram A. (2010) Massive lower GI bleed from an endoscopically inevident rectal varices: diagnosis and management by EUS (with videos). Gastrointest Endosc 72: 1106–1108. [DOI] [PubMed] [Google Scholar]
  20. Vazquez-Sequeiros E., Olcina J. (2010) Endoscopic ultrasound guided vascular access and therapy: a promising indication. World J Gastrointest Endosc 2: 198–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Wong R. (2004) Endoscopic Doppler US probe for acute peptic ulcer hemorrhage. Gastrointest Endosc 60: 804–812. [DOI] [PubMed] [Google Scholar]
  22. Wong R, Farooq F., Chak A. (2007) Endoscopic Doppler US probe for the diagnosis of gastric varices (with videos). Gastrointest Endosc 65: 491–496. [DOI] [PubMed] [Google Scholar]

Articles from Therapeutic Advances in Gastroenterology are provided here courtesy of SAGE Publications

RESOURCES