Gastric varices (GV) are present in approximately 20% of patients with portal hypertension. GVs bleed less frequently than esophageal varices (EV), but bleeding is more severe and mortality is higher. Risk factors for gastric variceal bleeding (GVB) include variceal location (particularly the fundus), size, overlying red signs, and advanced liver disease.
balloon-occluded retrograde transvenous obliteration
endoscopic band ligation
gastric variceal bleeding
isolated gastric varices
transjugular intrahepatic portosystemic shunt
Classification of GV
The commonly used classification system introduced by Sarin comprises four types of GV based on gastric location and relationship with EV (Fig. 1). Gastroesophageal varices (GOV) either are an extension of EV for 2 to 5 cm along the lesser curve of the stomach (GOV1) or extend along the greater curve into the fundus (GOV2). Isolated gastric varices (IGV) are located either in the fundus (IGV1) or in other parts of the stomach (IGV2). GOV1 account for 74% of all GV, but the incidence of bleeding is highest with fundal varices (IGV1 and GOV2). Bleeding IGV1 require the exclusion of isolated splenic vein thrombosis because therapy consists of splenectomy.
Management of Acute Bleeding and Prevention of Rebleeding
Similarly to patients with EV, patients with suspected GVB should be managed in an intensive care unit with careful blood volume resuscitation (target hemoglobin ∼8 g/dL), administration of prophylactic antibiotics (intravenous ceftriaxone in high-risk patients), correction of significant clotting abnormalities, and endotracheal intubation in patients at risk for aspiration. Despite the lack of efficacy data for GVB, vasoactive agents (e.g., octreotide) are administered because of their favorable safety profile and potential benefit. Variceal tamponade with a large capacity balloon (e.g., a Linton-Nachlas tube) is used as bridge therapy in patients with massive or uncontrollable bleeding.
The optimal approach to GVB remains undefined because of a lack of robust evidence-based data from large randomized trials and the inclusion of different types of GV among studies. Several endoscopic, interventional radiologic, and surgical treatment options are available. The selection of a particular treatment modality is influenced in part by patient characteristics, type of GV, available expertise, technical/anatomical considerations, and therapy-specific contraindications.
Band Ligation and Sclerotherapy
The management of GOV1 is similar to that of EV, and endoscopic band ligation (EBL) is preferred to sclerotherapy because of a lower risk of complications. However, EBL and sclerotherapy may be ineffective therapies for fundal varices (GOV2 and IGV1) because of high rebleeding rates and the potential for massive bleeding from large treatment-induced ulcers. EBL can be used to control active fundal variceal bleeding if it is the only available endoscopic option, but referral to a center for performance of more definitive therapy (e.g., cyanoacrylate injection) is recommended for prevention of rebleeding (secondary prophylaxis).
Cyanoacrylate (glue) injection is the recommended first-line therapy for GVB,[3, 4] with initial control of bleeding in 90% to 100% of patients and rebleeding rates <15% in recent series. Small controlled trials also favor cyanoacrylate over sclerotherapy and EBL with regard to initial hemostasis, rebleeding, and/or complication rates. Cyanoacrylate is not approved by the US Food and Drug Administration for intravariceal injection, but it is used off-label with increasing frequency in the United States.
Cyanoacrylate injection requires careful patient selection and attention to technique for an optimal outcome. Dynamic computed tomography is recommended to determine the presence of spontaneous large splenorenal or gastrorenal shunts (GRS). These shunts are associated with the risk for systemic glue embolism. A standardized injection protocol enhances procedural safety and efficacy, although technical variations exist depending on the type of cyanoacrylate used and local expertise (Table 1). Complete variceal obturation, including tributaries, is attempted at the initial session, and repeat endoscopy is performed in ∼4 weeks to confirm GV obliteration or treat residual varices (Fig. 2). Following variceal obliteration, surveillance endoscopy is typically performed at 3- to 6-month intervals to monitor for variceal recurrence. A glue cast may remain visible for weeks to months after injection. Beta-blocker therapy alone is inferior to cyanoacrylate injection for secondary prophylaxis of GVB, and its use in conjunction with cyanoacrylate provides no additional benefit with regard to prevention of rebleeding and mortality.
|N-Butyl-2 Cyanoacrylate||Octyl-2 Cyanoacrylate|
|Mixed with Lipiodol||Yes > No||No|
|Injected volume per puncture site||1-2 mL||2-4 mL|
|Maximum volume injected per treatment session||≤5 mL||≤10 mL|
The incidence of cyanoacrylate-related complications is low, which include bleeding from early glue cast extrusion, sepsis, and distant glue embolism. The presence of intracardiac or intrapulmonary shunts (as assessed by contrast echocardiogram) and a large GRS heighten the risk for systemic glue embolization. The embolic risk may be minimized with the introduction of novel techniques, such as EUS-guided coil/glue injection and the combined approach of endoscopic cyanoacrylate injection with angiographic balloon occlusion of the GRS (Fig. 3), although additional appraisal of these modalities is warranted.[15, 16]
The clinical experience regarding injection of thrombin or fibrin sealants for GVB is limited, with reported rates of 70% to 100% for initial hemostasis and 0% to 50% for rebleeding. Embolic complications have not been described, but leakage of thrombin into the systemic circulation via a GRS could potentially induce disseminated intravascular coagulation or pulmonary embolism. Controlled studies on the efficacy and safety of these agents for GVB are needed before they can be recommended for use. The use of detachable stainless steel or nylon snares for ligation of GV has been described, although ligation-induced ulcer leading to life-threatening hemorrhage remains a concern.
Transjugular Intrahepatic Portosystemic Shunt (TIPS)
TIPS is effective at controlling GVB in over 90% of cases, with reported rebleeding rates of 15% to 40% within 1 year. In one randomized trial, TIPS was more effective than cyanoacrylate injection for prevention of recurrent bleeding (11% versus 38%), although survival was similar and incidence of encephalopathy was higher in the TIPS group (26% versus 3%). GV may bleed at significantly lower portal pressures than EV, and TIPS may be less effective in this subgroup of patients in whom the hepatic venous pressure gradient is low and a large GRS is present. Because cyanoacrylate may be more cost-effective than TIPS,[20, 21] the latter is generally used as rescue therapy after failed endoscopic intervention for acute GVB. Following the acute episode, either cyanoacrylate or TIPS can be recommended for prevention of recurrent bleeding from fundal varices.
Balloon-Occluded Retrograde Transvenous Obliteration (BRTO)
BRTO is possible only in the presence of a GRS, which is present in up to 85% of patients with GV. The technique involves instillation of sclerosant or foam into the GV via a balloon-occluding catheter placed through the GRS. BRTO is used primarily in Japan for prevention of initial bleed and secondary prophylaxis of GVB, whereas its use in the United States and Europe is sporadic. Uncontrolled data show excellent initial control of bleeding (>90%), low rebleeding rates (0%-9%), and high variceal eradication rates (75%-100%). Potential adverse effects include fever, ascites, pleural effusions, and the development of EV in up to two-thirds of patients. Partial splenic embolization preceding BRTO resulted in a significantly reduced incidence of EV compared with BRTO alone (9% versus 45%) by reducing blood inflow into the portal vein. Where available, BRTO can be considered in patients with GVB who have failed endoscopic therapy and are poor candidates for TIPS.
Surgical intervention is considered a last resort when endoscopic and interventional radiologic therapies fail to control GVB in patients with good liver function (Child-Pugh class A).
Prevention of Initial Bleeding
Beta-blockers may be considered for prevention of first GVB (primary prophylaxis) even though their value is unproven in this setting. A recent, randomized, three-arm trial favored cyanoacrylate injection over no therapy (for prevention of bleeding and survival) and propranolol (for prevention of bleeding only), although these findings require validation due to statistical uncertainty accounting for multiple comparisons in this study. Although BRTO is used in Asia for primary prophylaxis of GV at risk for bleeding, it is not standard practice in most Western countries, and randomized trials are needed before this approach can be recommended.
The management of bleeding GOV1 is the same as bleeding esophageal varices. Endoscopic cyanoacrylate injection is considered first-line therapy for management of acute fundal variceal bleeding (GOV2 and IGV1). If cyanoacrylate is unsuccessful or unavailable, TIPS is recommended. BRTO is a potential alternative in patients who are poor candidates for TIPS. Following the acute episode, either cyanoacrylate or TIPS is recommended for prevention of recurrent bleeding from fundal varices (Fig. 4). No formal recommendations can be made with regard to specific therapies for prevention of first bleeding from GV in the absence of robust evidence-based data.