Hemostasis of Acute Gastric Variceal Bleeding

Fig. 13.1
Large fundal varices in cluster of grapes formation

The Sarin classification is commonly used to categorize gastric varices on the basis of their location in the stomach and relationship with esophageal varices [1] (Fig. 13.2). Gastroesophageal varices (GOV) are esophageal varices in continuity with gastric varices that extend along the gastric lesser curvature (GOV1) or toward the fundus (GOV2). Isolated gastric varices (IGV) occur in the absence of esophageal varices and are located either in the fundus (IGV1) or elsewhere in the stomach, such as the body, antrum, or pylorus (IGV2). The presence of IGV1 at endoscopy should raise suspicion and evaluation for splenic vein thrombosis (e.g., from pancreatitis) for which the treatment of choice is splenectomy.


Fig. 13.2
Sarin classification of gastric varices . Gastroesophageal varices (GOV) are esophagogastric varices that extend along the lesser curvature of the stomach (GOV1) or toward the fundus (GOV2). Isolated gastric varices (IGV) occur in the absence of esophageal varices and are located in the fundus (IGV1) or elsewhere in the stomach, such as the body, antrum, or pylorus (IGV2)

Gastric varices are typically secondary to portal hypertension, which is the result of increased portal blood flow and intrahepatic vascular resistance. Blood that normally flows through the portal vein into the liver is now impeded, resulting in diverted flow to alternate venous pathways. Increased pressure into the gastric veins forms fundal varices (IGV1 and GOV2), while increased flow into the gastroepiploic veins forms IGV2 [5]. Gastric varices are typically associated with spontaneous portosystemic shunts, namely, splenorenal or gastrorenal shunts, which drain through the left renal vein. A higher percentage of gastric varices are associated with gastrorenal shunts compared to esophageal varices [6]. These shunts may explain, in part, the lower portal pressures that can be encountered with bleeding gastric varices. The estimated total blood flow through these shunts can be as high as 1.7 l/min, which allows for decompression of the portal system and lowering of the transhepatic pressure gradient [7].

GOV1 are considered an extension of esophageal varices and the management of bleeding GOV1 is, therefore, similar to that of esophageal varices (Video 13.1). Limited data are available regarding the management of IGV2, but its treatment generally mirrors that of GOV2 and IGV1. Herein, the management of GVH from fundal varices (IGV1 and GOV2) will be the focus of this chapter, with emphasis on endoscopic therapy.

Initial Management

The medical management of acute variceal bleeding is detailed in a separate chapter. In brief, initial management of GVH is similar to that of esophageal variceal hemorrhage and consists of hemodynamic stabilization with blood transfusion, as appropriate, prophylactic antibiotics, and administration of a vasoactive drug, such as octreotide or terlipressin. Patients with acute GVH should be managed in an intensive care unit initially, and endotracheal intubation for airway protection is recommended in the setting of active bleeding or other factors that place the patient at risk for aspiration (e.g., encephalopathy). The transfusion strategy should be restrictive in nature, with target hemoglobin of 7–8 g/dl, since transfusions above this level can elevate portal pressure and increase bleeding [8, 9]. Due to the lower venous pressure gradient needed for GVH compared to esophageal variceal hemorrhage, a higher dose of a vasoconstrictor may be required to decrease portal pressure and reduce portal and collateral blood flow [10]. Although there is evidence that octreotide is beneficial for esophageal variceal bleeding, there are no dedicated studies that have examined the role of vasoactive agents in the setting of GVH [11].

Massive bleeding from gastric varices may require initial balloon tamponade as temporary therapy. For this purpose, a variety of balloon devices are available (e.g., Sengstaken-Blakemore, Linton-Nachlas, and Minnesota tubes), although the Linton-Nachlas tube is preferred for GVH due to its larger gastric balloon capacity (600 ml) for more effective tamponade of fundal varices. Although balloon tamponade may provide immediate hemostasis, sustained hemostasis is unlikely, with high rebleeding rates following balloon deflation. The maintenance of balloon tamponade for longer than 24 h may result in ischemic necrosis and perforation [12], and it should only be used as a bridging measure to more definitive therapy.

Endoscopic Therapy

Endoscopic Band Ligation

Although the use of endoscopic band ligation (EBL) for treatment of esophageal varices is well supported in the literature, its application for fundal variceal hemorrhage is limited. Initial case series revealed EBL to be safe and effective for the control of acute GVH [13, 14], but subsequent randomized controlled trials showed rebleeding rates as high as 60–70 % [15]. One prospective study that compared EBL to cyanoacrylate injection of gastric varices revealed comparable initial hemostasis (100 and 89 %, respectively), but a significantly higher rebleeding rate in the EBL group (72 % vs. 32 %) [16]. EBL is considered an ineffective treatment for sustained hemostasis of fundal variceal hemorrhage due to post-banding ulcers and recurrent hemorrhage after band sloughing (Fig. 13.3), as well as failure to completely ligate the deep aspect of the varix and its feeder vessel(s). Detachable snares or loops have also been used to ligate gastric varices. This technique was associated with low rebleeding rates (0–10 %) in two small studies [17, 18], although the technical challenge in loop placement and risk of torrential bleeding at the site of post-ligation ulcer have restricted the use of these devices in practice. Although EBL or loop ligation is not recommended as primary therapy for fundal variceal bleeding, these techniques can be considered in certain circumstances to temporarily arrest active bleeding from a fundal varix in order to buy time toward preparation for more definitive therapy (e.g., cyanoacrylate injection).


Fig. 13.3
Massive bleeding from post-banding ulcers overlying fundal varices


Endoscopic injection of sclerosants results in endothelial damage and thrombosis, leading to vascular obliteration [19]. Available sclerosing agents include absolute alcohol, fatty acid derivatives (e.g., ethanolamine oleate and sodium morrhuate), and synthetic chemicals (e.g., sodium tetradecyl sulfate and polidocanol). A sclerotherapy needle is passed through the working channel of the endoscope and the sclerosant is injected either into (intravariceal) or immediately adjacent (paravariceal) to the varix. Intravariceal injection results in direct occlusion of the vessel, while paravariceal injection occludes the vessel by submucosal fibrosis of tissue around the varix, leading to tamponade.

Although sclerotherapy in esophageal variceal hemorrhage is effective, with control of active bleeding in about 90 % of patients [2023], its use in GVH is less impressive. The rates of initial hemostasis range from 44 to 92 %, with high rebleeding rates of 30–90 % and poor rates of eventual variceal obliteration [2429]. Adverse events related to sclerotherapy include post-sclerotherapy ulceration with delayed bleeding and bacteremia/sepsis. Sclerotherapy is not recommended as first-line therapy for GVH unless no other options are available.

Cyanoacrylate Injection

Supportive Evidence

The use of cyanoacrylate (glue) injection for the treatment of gastric varices was first described by Soehendra et al. in 1986 [30] and currently constitutes first-line treatment for GVH, where available [2]. There is evidence supporting the use of cyanoacrylate injection, with several large case series (encompassing 121–613 patients per study) reporting >90 % control of bleeding and <15 % rebleeding rates [31]. The most feared complication of glue injection is embolization [32]. However, the rate of clinically relevant glue embolization was only 0.7 % (1 pulmonary, 1 cerebral, and 3 splenic) in the largest case series of cyanoacrylate injection for gastric varices that enrolled over 750 patients [33]. The overall complication-related mortality was 0.53 % and included three deaths from sepsis and one death from rebleeding after early-onset glue cast extrusion. Early-onset (<3 months) rebleeding was 4.4 %.

Sarin et al. showed cyanoacrylate injection to be more effective than alcohol sclerotherapy in achieving gastric variceal obliteration in a small randomized controlled trial of 37 patients [24]. In another prospective nonrandomized trial, cyanoacrylate injection was better than sclerotherapy with regard to acute hemostasis (93 % vs. 67 %), although the rebleeding rates were not different (25 % vs. 30 %) [26].

Two randomized controlled trials have compared EBL with cyanoacrylate injection. A trial of 60 patients showed improved outcomes favoring glue injection over EBL in terms of initial hemostasis (87 % vs. 45 %), rebleeding (31 % vs. 54 %), and mortality (29 % vs. 48 %) [15]. A larger randomized study of 97 patients demonstrated similar rates of initial hemostasis (93 % for both), but a higher rebleeding rate in the EBL group (3-year cumulative rate of 72 % vs. 27 %) [34]. In a retrospective study comparing EBL (n = 18) to cyanoacrylate injection (n = 19), the rates of initial hemostasis were similar (89 % vs. 100 %), but the rebleeding rate was significantly less in the cyanoacrylate group (32 % vs. 72 %) [16].

Cyanoacrylate Monomers

Cyanoacrylate monomers differ primarily in the length of their alkyl groups. The two monomers that are currently used for GVH are enbucrilate (N-butyl 2-cyanoacrylate) and ocrylate (2-octyl cyanoacrylate). Enbucrilate polymerizes at a faster rate than ocrylate and is usually mixed with the oily contrast agent, Lipiodol (Guerbet LLC, Bloomington, IN, USA), to slow its polymerization rate and minimize glue solidification within the injection catheter before it reaches the varix. The use of Lipiodol also allows for fluoroscopic visualization of the mixture during injection, if desired. The enbucrilate-to-Lipiodol ratio ranges from 1:1 to 1:6 in published studies [35, 36]. A commercial formulation of enbucrilate adds methacryloxysulfolane to slow the polymerization rate.

Technique for Cyanoacrylate Injection

The technical steps for cyanoacrylate injection are outlined in Table 13.1. After administering prophylactic antibiotics, the endoscopic procedure is performed preferably with a therapeutic channel gastroscope coupled to a water irrigation pump. Silicone oil should be used to coat the tip of the endoscope, as well as to flush the instrument channel to minimize the risk of glue adherence that can lead to endoscope damage. The injection needle catheter should be primed with either sterile water for enbucrilate injection or saline for ocrylate injection. Saline should not be used for enbucrilate injection because it accelerates its polymerization rate, which may lead to premature clogging of the catheter.

Table 13.1
Technical steps for cyanoacrylate injection

1.Coat the endoscope tip with silicone oil and flush oil through the instrument channel to minimize the risk of glue adherence

2.Prime the injection needle catheter with either sterile water for enbucrilate injection or saline for ocrylate injection

3.Confirm that the initial injection with water or saline is free flowing into the varix and not forming a submucosal bleb

4.Inject the cyanoacrylate into the varix in aliquots of 0.5–1 ml. If used undiluted, enbucrilate must be rapidly injected over a few seconds to avoid premature glue solidification. Due to its longer polymerization rate, ocrylate must be used undiluted and slowly injected over 30–45 s

5.After the glue has been injected, flush out the dead space of the catheter with sterile water or saline

6.Retract the needle from the varix while continuously flushing to keep the needle patent for possible repeat glue injection

7.If there is no bleeding at the puncture site, palpate the varix with a blunt tip catheter or closed forceps. If the varix is still soft, additional glue injections are performed

Once the target varix is punctured, the initial injection of saline or sterile water, depending on the type of cyanoacrylate used, is carefully observed to ensure the solution flows freely into the varix and does not form a submucosal bleb. The varix is then injected with aliquots of 0.5–1 ml of the glue. Injection of more than 1 ml of glue per aliquot may increase the risk of embolization. The injection time will vary depending on the choice of cyanoacrylate and degree of dilution with Lipiodol. If the injection is too slow, the glue may solidify in the needle. If used undiluted, enbucrilate must be rapidly injected over seconds to minimize premature glue solidification. Because of its longer polymerization time, ocrylate must be used undiluted and is slowly injected over 30–45 s (Video 13.2). After the glue has been injected, the catheter’s dead space is flushed with sterile water or saline, and the needle is withdrawn from the varix while continuously flushing solution to keep the needle patent for possible repeat glue injection. If there is no bleeding at the punctured site, the varix is palpated with a blunt tip instrument to confirm adequate obturation, as evidenced by a hardened varix. If the varix is still “soft,” then additional glue injections are performed (Fig. 13.4). Follow-up endoscopy several weeks later may show retained glue cast at the puncture site(s) (Video 13.3).


Fig. 13.4
(a) Large isolated gastric varix in the fundus (IGV1). (b) Endoscopic injection of cyanoacrylate into the varix. (c) Follow-up endoscopy at 8 weeks shows partial glue extrusion at the puncture site. (d) Obliterated gastric varix on surveillance endoscopy at 6 months

Cyanoacrylate Injection for Primary Prophylaxis

The role of primary prophylaxis (prevention of a first bleed) is established in the management algorithm of esophageal varices [2], but no clear guidelines exist with regard to primary prophylaxis of gastric varices. One randomized trial evaluated the role of endoscopic glue injection for primary prophylaxis of gastric varices. A total of 89 patients with fundal varices were randomized to cyanoacrylate injection, beta-blocker therapy, or no treatment. After a mean follow-up of 26 months, the probability of bleeding was 13 % in the cyanoacrylate group, 28 % in the beta-blocker group, and 45 % in the no-treatment group. Survival was higher in the cyanoacrylate group compared to the no-treatment group (90 % vs. 72 %). Predictors of bleeding were variceal size, model for end-stage liver disease (MELD) score, and presence of portal hypertensive gastropathy. Further studies are needed before endorsing endoscopic glue injection for primary prophylaxis of gastric varices, although some centers, including ours, have offered this option on a case-by-case basis, particularly in patients with gastric varices at high risk for bleeding (e.g., large fundal varices with prominent red wale markings or hematocystic spots).

Cyanoacrylate Injection for Secondary Prophylaxis

Despite excellent initial hemostasis following endoscopic therapy, rebleeding occurs at rates of 10–30 % [10]. A randomized trial of 67 patients with prior bleeding from fundal varices compared cyanoacrylate injection with beta-blocker treatment [37]. Rebleeding (15 % vs. 55 %) and mortality (3 % vs. 25 %) were significantly lower in the cyanoacrylate group. If secondary prophylaxis (prevention of recurrent bleeding) is pursued, repeat sessions are often needed to ensure complete eradication. However, a standardized protocol for repeat therapy has not been established. Some centers advocate for retreatment, as appropriate, at intervals of 2–12 weeks to ensure complete gastric variceal obliteration [35, 38].

While the role of combined endoscopic and medical therapies to prevent rebleeding is better established for esophageal varices [39], such combination treatment has not been well studied for gastric varices. In one randomized trial of 95 patients who bled from fundal varices, repeat cyanoacrylate injection every 3–4 weeks until eradication was compared to repeat cyanoacrylate injection plus a nonselective beta blocker [40]. After a mean follow-up of 20.3 months, there was no difference in rebleeding or survival rates between the two groups.

Thrombin Injection

Thrombin assists in hemostasis by converting fibrinogen to a fibrin clot, as well as enhancing local platelet aggregation. Human thrombin is pooled from human plasma donors and is typically injected in aliquots of 1 ml per injection site, with an average dose of 1500–2000 units [31]. After the initial report of its use in 1947 by Daly [41], numerous small uncontrolled observational studies have shown thrombin to be an effective initial hemostatic agent for the treatment of gastric varices, with successful hemostasis in 70–100 % of patients and relatively low rebleeding rates [4247]. There are no controlled trials of thrombin injection for gastric varices to date, although one trial that compared ethanolamine injection, with or without thrombin, showed lower rates of bleeding from the injection site in the thrombin group [48]. The cost of thrombin is substantially higher than cyanoacrylate, and further studies are needed before thrombin can be recommended as a primary treatment option for GVH.

EUS-Guided Therapy

EUS-Guided Cyanoacrylate Injection

Most centers perform glue injection under endoscopic guidance, which may result in injection adjacent to rather than within the varix. Data from sclerotherapy suggest that up to 60 % of injections are actually paravariceal in nature [49]. EUS-guided glue injection is attractive as it enables sonographic visualization for precise glue delivery into the variceal lumen. Furthermore, the technique allows for visualization of deeper varices as well as feeder vessels [50], which can be targeted separately. EUS can improve detection and visualization of gastric varices [51], especially in the setting of active bleeding which may obscure the endoscopic field of view. The monitoring of gastric varices with EUS and repeat injections until complete obliteration have been shown to decrease rebleeding rates [52], and color Doppler can be used to confirm complete variceal obliteration with absence of blood flow.

EUS-Guided Coil Injection

A large variety of embolization coils are available for transcatheter vascular use. Many of these fit through EUS fine-needle aspiration (FNA) needles and can be utilized for EUS-guided angiotherapy. The coils used at our institution are made of Inconel, a nickel-based superalloy. The coils contain radially extending, synthetic fibers that help induce clot formation and hemostasis. The coils are MRI conditional and can be used in a static magnetic field of 3 T or less. A variety of coil sizes and lengths are available. A 0.035-in. coil will fit through a 19-gauge FNA needle (Fig. 13.5). Straightened coil lengths range from 2 cm to 15 cm, with coiled diameters of 2 mm to 20 mm and approximate number of loops ranging from 1.9 to 5.6. Smaller 0.018-in. coils are also available and will fit through a 22-gauge FNA needle. Coil selection depends on the size of the varix, but typically a coiled diameter of 10–20 mm is optimal.


Fig. 13.5
A 0.035-in. embolization coil is advanced through a 19-gauge FNA needle and resumes a coiled configuration as it exits the needle tip

EUS-guided coil injection for acute variceal bleeding was initially reported in 2008 [53]. Embolization was accomplished using microcoils through a 22-gauge FNA needle for variceal obliteration of ectopic varices surrounding a choledochojejunal anastomosis. A retrospective trial of 30 patients comparing EUS-guided coil injection to EUS-guided cyanoacrylate injection revealed similar obliteration rates, but fewer endoscopy sessions were required in the coil group (82 % vs. 53 % obliteration in a single treatment session) [54]. Of note, the intended therapy was for coil injection when feasible, but only 11 of 30 patients underwent such therapy due to technical difficulties hindering the use of coils. The rate of adverse events was significantly higher in the cyanoacrylate group (58 % vs. 9 %), although 9 of the 11 adverse events in the cyanoacrylate group were asymptomatic pulmonary glue embolisms found on routine post-procedure CT. This indicates that glue embolization is occurring more commonly than appreciated, but that it rarely causes symptoms.

EUS-guided angiotherapy requires additional training and expertise in interventional EUS and performance of the technique appears currently limited to a very small number of tertiary centers. EUS-guided angiotherapy faces several challenges compared to standard endoscopic techniques, including a smaller echoendoscope channel size with limited suction capability and the required ultrasound processor, which makes bedside endoscopy in the intensive care unit difficult. Identification of the feeder vessel can also be challenging and time-consuming. Accidental injection of cyanoacrylate into an efferent vessel would not provide variceal obliteration and could increase the risk of embolization. Lastly, EUS-guided therapy is more suitable for localized gastric varices from a single feeder vessel, whereas a diffuse variceal network may not be amenable to EUS-guided therapy [55].

EUS-Guided Combined Coil and Cyanoacrylate Injection

Our center has developed an EUS-guided approach consisting of coil placement followed immediately by glue injection into the same varix. We theorized that the coil provides a scaffold for glue retention at the site of intravariceal injection. We believe the combination of coil and glue may enhance the rates of hemostasis and variceal obliteration while decreasing the risk of glue embolization.

The procedural protocol at our center for combined EUS-guided coil and glue injection is as follows (Table 13.2 and Fig. 13.6):

Table 13.2
Technique for EUS-guided coil and glue therapy

1.Standard upper endoscopy for classification of varices

2.Perform EUS with curvilinear array echoendoscope with intraluminal water filling

3.Puncture target varix with 19- or 22-gauge FNA needle primed with saline. Verify intravariceal needle position with blood aspiration. Deliver coil into varix with needle stylet as a pusher

4.Re-aspirate blood to ensure needle position is still within the varix. Inject 1 ml of 2-octyl cyanoacrylate over 30–45 s, followed by saline to flush glue through the needle’s dead space

Only gold members can continue reading. Log In or Register to continue

May 30, 2017 | Posted by in GASTROENTEROLOGY | Comments Off on Hemostasis of Acute Gastric Variceal Bleeding
Premium Wordpress Themes by UFO Themes