INTRODUCTION: Bleeding from gastric varices (GV) is associated with a high mortality rate. Injection of cyanoacrylate (CYA) using standard gastroscope has demonstrated to achieve higher hemostasis and lower rebleeding rates compared to band ligation or sclerotherapy. Nevertheless CYA treatment is known to be associated with significant adverse events. Pulmonary embolism due to CYA injection is a serious and sometimes fatal complication of this therapy. These patients usually have respiratory symptom, however this complication can be present in asymptomatic patients, being demonstrated only by a pathological CT scan. On the other hand, it has been described that the risk of glue embolism dependent on the volume of CYA injected, being significantly greater with high volumes. Other complications related to CYA injection are hemorrhage from injection site ulcers, fever, peritonitis, needle impaction, and even death. Also the injection material can cause serious damage to the endoscope.

Currently, endoscopic injection of CYA can be performed by direct visualization using a standard gastroscope or guided by Endoscopic Ultrasound (EUS) with injection of CYA alone or in combination with coils. The injection of coils in conjunction with CYA may reduce or eliminate the risk of glue embolization as coils can function as a scaffold to retain CYA within the varix and may decrease the amount of glue injection needed to achieve obliteration. It has been previously demonstrated that treatment under EUS guidance may have some benefits. It allows a precise targeting of the varix lumen or afferent feeding veins, being the vessel obstructed with less amount of CYA than used for the "blind" injection by standard endoscopy, reducing the risk of glue embolism. EUS can confirm varix obliteration by Doppler effect and also the visualization of GV is not impaired by blood or food in the stomach, thus it can be used in the setting of active hemorrhage.

On the other hand, despite the EUS-injection of coils and CYA have shown very promising results, it is extremely important to have experience performing this technique, as the injection in a wrong vessel could be catastrophic. It has been develop in our unit a modification of Binmoeller´s original technique by performing the injection of coils and CYA by EUS and fluoroscopic guidance. The performance of an angiography guided by EUS and under fluoroscopy evaluation, represents a modification to the original technique that add some benefits: 1) The injection of water-soluble contrast allows identifying the variceal flow and the afferent feeding vessel (left gastric vein) and avoiding splenic vessels embolization. 2) Also, if there is a shunt it can be detected and the injection of CYA can be avoided. 3) In case of active bleeding the angiography allows visualizing which varix is bleeding, and target the correct vessel.

The aim of this study is to show the results and potential benefits by adding an angiography to the original EUS-guided injection of coils in combination with CYA technique, to determinate and confirm the feeder vessel in GV treatment.

MATERIALS AND METHODS

Study design: It will be an interventional, descriptive and prospective study, performed in the Ecuadorian Institute of Digestive Disease (IECED), Omni Hospital Academic Tertiary Center Ecuador. Patients have been included from July 2015 and will continue to be included up to October 2017. The study protocol and consent form has been approved by the institutional review board, and will be conducted according to the declaration of Helsinki. Written informed consent will be obtained from all subjects.

Population selection, inclusion and exclusion criteria: Patients above 18 years old with gastric varices (GV) on the initial standard diagnostic upper endoscopy will be enrolled. GV will be classified according to Sarin and Kumar classification. Only gastro-esophageal varices type II (GOV II) (fundal varices communicating with esophageal varices) and isolated gastric varices type I (IGV I) (fundal varices within a few centimeters of the gastric cardia) will be included. Gastro-esophageal varices type I (GOV I) will be excluded as they can be successfully treated by endoscopic band ligation. Patients with active bleeding and history of previous bleeding due to GV (secondary prophylaxis) will be included as well as patients with high-risk GV suitable for primary prophylaxis according to Baveno VI consensus. The exclusion criteria will be pregnancy, concurrent hepatorenal syndrome and/or multi-organ failure, platelet count less than 50,000/ml or International Normalized Rate (INR) >2, esophageal stricture, splenic or portal vein thrombosis and allergy to iodine.

Endoscopic Procedure and Technique: All procedures will be performed in a hospital-based interventional endoscopy setting, where radiology is available and by one endoscopist (C.R.M), under general anesthesia, with tracheal intubation, with the patient in supine position and using antibiotic prophylaxis. EUS examination will be performed using a 3.8 mm working channel linear-array therapeutic echoendoscopes (EG 3870UTK; Pentax, Hamburg, Germany), attached to an ultrasound console (Avius Hitachi, Tokyo, Japan). The echoendoscope will be positioned in the distal esophagus at the level of the cardia (anterograde trans-esophageal, transcrural approach) to visualize the gastric fundus and intramural varices.

Detection of the feeder vessel by EUS: Once positioned, water will be instilled to fill the gastric fundus, in order to improve acoustic coupling and visualization of GV. EUS color Doppler imaging will be used to allow direct visualization of the varices flow. Following the gastric vessel, and by using the fine-flow doppler color of the ultrasound console, the vessel or vessels will be followed from the cardia (distal part of the esophagus) to the proximal part of the esophagus (2 to 3 cm) above the cardia, to detect the feeding vessel. The feeding vessel is characterized by the convergence of all vessels. Once the feeding vessel has been found, a 19 gauge EUS-fine needle aspiration (FNA) (Expect® flexible; Boston Scientific®, Marlborough, USA) will be used to access the vessel, the stylet will be withdrawn and a 20 ml syringe with negative pressure will be used to evaluate blood return, confirming intravascular location. In order to prevent blood clotting of the needle tip, 5 ml of saline solution will be instilled. Moreover, the flush with saline solution allows to confirm the flow of the gastric vessels under B-mode ultrasound visualization.

Using angiography to confirm the target vessel (feeder vessel) and flow traject: In the final diagnostic step, an angiography will be performed using 5 to 10 mL of water-soluble contrast (Ultravist, Bayer, Ecuador) under fluoroscopic evaluation, in order to ensure intravascular location and study varix flow direction (afferent or efferent). Angiography technical success will be measured, determined as the possibility to define the flow trajectory vessel under fluoroscopy.

EUS-guided deployment of Coils and Cyanoacrylate injection: Finally, patients will be treated by EUS-deployment of coils followed by CYA injection. The intravascular embolization coils that will be used (10-16 mm coiled diameter, 12-20 cm straight lengths, 0.035 inches diameter, Nester Embolization Coil; Cook Medical) will be delivered into the vessel through the FNA needle, using the stylet as a pusher. Special attention will be paid not place the needle tip on the counter wall due to the risk of perforation, bleed, coil extrusion and to allow enough space for the coil to curl.

The CYA that will be used is the 2-Octyl-CYA (Dermabond; Ethicon, Piscataway, NJ). It will be injected always after the coils are deployed and then 1 mL of saline solution will be used to flush the glue completely through the needle. The diameter (10-16 mm), number of coils and the volume of the 2-Octyl-CYA injected will be calculated according to the diameter of the vessel measured on EUS and by the angiography of the vessel. After 30 seconds, once CYA has been solidified and the risk of bleeding on the puncturing site decrease, the needle will be withdrawn. Obliteration of the vessel will be confirmed 3 minutes later using Doppler imaging. Complete obliteration of the treated varix will be defined as absence of Doppler flow on EUS.

A direct visualization of the gastric varices will be performed after the EUS procedure to determine the effect of the obliteration of the feeder vessel by using a standard videoendoscope. After the procedure, patients will be observed for 2 hours in the recovery room before being discharged of the unit. In cases of primary prophylaxis the treatment will be ambulatory.

Finally, a 1, 3, 9 and 12 months follow-up will be performed in order to confirm the results of the EUS-therapeutic by EUS and by upper endoscopy.

Technical success will be considered as target of the vessels and deployment of the proposed EUS-therapeutic. Treatment failure and early rebleeding will be considered when bleeding occurred the first five days and between 5 and 90 days respectively.

Statistical analysis: Demographic characteristics, previous medication and radiological images will be described. Quantitative variables will be described in mean or median, and standard deviation or range, as corresponding. Qualitative variables will be described in percentages. A Kolmogorov-Smirnov test with a p value <0.05 will be considered to be statistically significant. The IECED institutional biostatistician will review statistical methodology of the study. Statistical calculations will be performed in SPSS software suite v.22.

Inclusion Criteria:

• Patients above 18 years old with gastric varices (GV) on the initial standard diagnostic upper endoscopy will be enrolled. GV will be classified
• Only gastro-esophageal varices type II (GOV II) (fundal varices communicating with esophageal varices) and isolated gastric varices type I (IGV I) (fundal varices within a few centimeters of the gastric cardia) will be included (according to Sarin and Kumar classification).
• Patients with active bleeding and history of previous bleeding due to GV (secondary prophylaxis)
• Patients with high-risk GV suitable for primary prophylaxis according to Baveno VI consensus.

Exclusion Criteria:

• Gastro-esophageal varices type I (GOV I) will be excluded as they can be successfully treated by endoscopic band ligation.
• Pregnancy,
• Hepatorenal syndrome
• Multi-organ failure,
• platelet count less than 50,000/ml
• International Normalized Rate (INR) >2,
• esophageal stricture,
• Splenic vein thrombosis
• Portal vein thrombosis
• Allergy to iodine.