On the contrary to EV, gastric varices (GV) are present in a smaller proportion of patients with cirrhosis (20%). It is known that a hemorrhage from cardiofundal varices is less frequent but often more severe and not easily controlled, providing a higher risk of recurrent bleeding and mortality (when compared to EV)
[41][53]. While standard EGD still represents the gold standard in detecting EV, EUS has better sensitivity in diagnosing GV
[54], with a detection rate two times higher
[55]. According to some authors, EUS can evaluate ectopic duodenal varices
[56], easily distinguish thickened gastric folds from small GV
[57], and help in the diagnosis of portal gastropathy, showing diffuse thickening of the gastric wall with dilated paragastric veins (differential diagnosis to “watermelon stomach”)
[33]. EUS also has an important role in the characterization of GV, visualization of treatment in progress, and confirmation of obliteration using Doppler
[58][59]. Furthermore, EUS can easily measure the size of GV, which directly correlates with their flow volume
[60]. Nowadays, the standard endoscopic management of fundal GV in acute bleeding or selective therapy is endoscopic cyanoacrylate (CYA) injection, which can be complicated with fever, chest pain, post-injection ulcers, re-bleeding (15 to 30%), embolic events (the incidence increases with the amount of CYA injected), or death. One of the advantages of EUS includes the identification of GV in the setting of acute bleeding when blood and clots in the gastric lumen disable an adequate endoscopic view
[61]. Since the risk factors for re-bleeding include varix size, presence of para-gastric veins
[62], and deficiency of complete obliteration of the GV or of the perforating vascular channels, which are unavailable for detection or eradication during standard EGD, a possible therapeutic role of EUS is arising. A retrospective study on 101 patients treated with glue injection after an episode of GV hemorrhage showed significantly lower re-bleeding rates in those patients in whom EUS was aggressively used during follow-up with the intention of achieving a complete obliteration of variceal veins
[63]. EUS-guided hemostasis of GV (with different available methods: injection of CYA, coils, coils with CYA injection, thrombin, or coils with an absorbable gelatin sponge (AGS)) allows assessment of the variceal blood flow, selective targeting of the varices with very exact treatment into the lumen of varix or into its feeding vessel (lowering the required dose of adhesive agent), and monitoring of the obliteration results (confirmation of cessation of variceal blood flow using Doppler and the presence of echogenic GV)
[64]. Several studies evaluated the role of CYA injection alone, either in primary prophylaxis
[65][66] or acute GV bleeding
[67][68][69], with an overall GV obliteration rate of 100% for the first group and 77–100% for the second. The re-bleeding rate was 0% and 5%, respectively, and severe complications were detected in the second group and included pulmonary embolism and splenic infarct in 5% of cases. As it was mentioned earlier, EUS-guided CYA injection has a risk of distal embolization (embolization to the pulmonary arteries and systemic embolism), and a multidisciplinary assessment is required to evaluate the potential presence of a septal defect prior to EUS-guided CYA injection. Furthermore, EUS-guided coil injection (either with or without CYA injection that can be delivered using standard 22- or 19-gauge needles used for FNA) may be a future method of choice to reduce the risk of embolization due to providing primary hemostasis
[70][71]. In the literature, limited data showing the role of coil injection alone in treating GV are available. Five studies evaluated coil injection in primary prophylaxis
[67][72], both primary prophylaxis and acute GV bleeding
[73][74] and in secondary prophylaxis
[75]. The results showed a GV obliteration rate of 70–100%, with no re-bleeding complications, but one event of major bleeding during the procedure was detected. Several groups of authors encourage the use of glue injection and coils in combination (primary prophylaxis, acute GV bleeding, and secondary prophylaxis), believing in their synergistic activity of hemostasis and reducing the risk of re-bleeding and distal embolization
[64][67][70][75][76][77][78][79]. The overall GV obliteration rate was 40–100%, with up to a 20% re-bleeding rate. A retrospective trial that compared EUS-guided CYA injection to EUS-guided coil placement showed similar rates of varix obliteration (complete obliteration was more likely to be achieved in the coil group after a single endoscopic session) and re-bleeding rates. It was shown that patients treated with CYA injection had significantly higher adverse events, but the number of sessions needed was fewer in patients receiving coil embolization
[67]. Some of the adverse events associated with coil placement (with or without CYA injection) include abdominal pain, fever, minor and major bleeding, coil migration, and extrusion of coils into the gastric lumen
[67][73][75]. Severe complications also included pulmonary embolism in up to 25% of cases
[64][77]. In addition to synthetic tissue adhesives, such as CYA, some of the biologic tissue adhesives that have been studied for GV obliteration include thrombin (converts fibrinogen to fibrin and promotes clot production) and AGS that is prepared from purified porcine gelatin and can absorb up to 45 times its weight in whole blood. The studies showed that EUS-guided thrombin injection had no procedure-related complications, making it safe to use in this indication
[80]. EUS-guided coil placement followed by AGS injection is a well-tolerated procedure, providing positive results in small case series
[81][82]. EUS-guided CYA injection with/without coiling has also been used for duodenal varices
[83]. Despite the abovementioned results, the specific role of EUS-guided coil/CYA injection in primary prophylaxis of EGV is not clear yet. Based on the available data, the treatment strategy should imply aggressive retreatment of any residual GV seen on follow-up EUS, with the intention of achieving their complete obliteration. It is advocated that EUS-guided coil and CYA injection have the best efficacy in the treatment of GV. Due to the previously mentioned advantages of coil placements, EUS-guided coil insertion has been given a preference over CYA injection. In one single-center study, a retrospective cohort of patients with active/recent bleeding or high-risk GV treated with direct endoscopic injection was compared with a prospective cohort of similar patients treated with EUS-guided fine needle injection (EUS-FNI). It was concluded that EUS-FNI is the preferred treatment strategy, which is substantiated by results showing decreased rates of bleeding in the EUS-guided CYA injection group of patients with active or recently bleeding GV
[20][69]. According to retrospective analysis that compared patients who underwent EUS-guided coil injection with patients who underwent a standard EGD injection of CYA for secondary prophylaxis of GV, the EUS group had a significantly lower rate of re-bleeding
[76]. In conclusion, EUS does not have an established role in clinical practice to investigate PH yet. The only distinct indication for EUS-guided treatment is the failure of standard EGD in GV bleeding control
[4][61]. In the future, EUS might provide an alternative approach to transjugular intrahepatic portosystemic shunts in cases of refractory ascites and refractory variceal bleed, and more studies are needed before its eventual implementation in humans.