BACKGROUND: Gastric variceal hemorrhage is one of the primary manifestations of left-sided portal hypertension (LSPH). The hemorrhage is fatal and requires safe and effective interventions.
OBJECTIVE: To evaluate the clinical safety and efficacy of modified endoscopic ultrasound (EUS)-guided selective N-butyl-2-cyanoacrylate (NBC) injections for gastric variceal hemorrhage in LSPH.
METHODS: A retrospective observational study of patients with LSPH-induced gastric variceal hemorrhage was conducted. Preoperative EUS evaluations were performed. Enrolled patients were divided into modified and conventional groups according to the NBC injection technique. The final selection of NBC injection technique depended on the patients\' preferences and clinical status. The technical and clinical success rates, operation time, NBC doses, perioperative complications, postoperative hospital stay, and recurrent bleeding rates were analyzed, respectively.
RESULTS: A total of 27 patients were enrolled. No statistically significant differences were observed between the two groups regarding baseline characteristics. In comparison to patients in the conventional group, patients in the modified group demonstrated significantly reduced NBC doses (2.0 ± 0.6 mL vs 3.1 ± 1.0 mL; P = 0.004) and increased endoscopic operation time (71.9 ± 11.9 min vs 22.5 ± 6.7 min; P < 0.001). Meanwhile, the two groups had no significant difference in the technical and clinical success rates, perioperative complications, postoperative hospital stay, and recurrent bleeding rates.
CONCLUSIONS: Modified EUS-guided selective NBC injections demonstrated safety and efficacy for LSPH-induced gastric variceal hemorrhage, with advantages of reduced injection dose and no radiation risk. Drawbacks were time consumption and technical challenge.

BACKGROUND: Left-sided portal hypertension (LSPH), also known as sinistral portal hypertension or regional portal hypertension, refers to extrahepatic portal hypertension caused by splenic vein obstruction or stenosis. N-butyl-2-cyanoacrylate (NBC) has been widely used in the endoscopic hemostasis of portal hypertension, but adverse events including renal or pulmonary thromboembolism, mucosal necrosis and gastrointestinal (GI) bleeding may occur after treatment. Herein, we report successfully managing gastric variceal (GV) hemorrhage secondary to LSPH using modified endoscopic ultrasound (EUS)-guided selective NBC injections.
METHODS: A 35-year-old man was referred to our hospital due to an upper GI hemorrhage. Gastroscopy revealed GV hemorrhage and computed tomography venography (CTV) confirmed LSPH. The patient requested endoscopic procedures and rejected surgical therapies including splenectomy. EUS-guided selective NBC injections were performed and confluences of gastric varices were selected as the injection sites to reduce the injection dose. The \"sandwich\" method using undiluted NBC and hypertonic glucose was applied. No complications occurred. The patient was followed up regularly after discharge. Three months later, the follow-up gastroscopy revealed firm gastric submucosa with no sign of NBC expulsion and the follow-up CTV showed improvements in LSPH. No recurrent GI hemorrhage was reported during this follow-up period.
CONCLUSIONS: EUS-guided selective NBC injection may represent an effective and economical treatment for GV hemorrhage in patients with LSPH.

OBJECTIVE: To focus on procedure-related complications, evaluate their incidence, analyze the reasons and discuss the solutions.
METHODS: Overall, 628 endoscopic gastric variceal obturation (EGVO) procedures (case-times) with NBC were performed in 519 patients in the Department of Endoscopy of the Third Affiliated Hospital of Sun Yat-Sen University from January 2011 to December 2016. The clinical data of patients and procedure-related complications of EGVO were retrospectively analyzed.
RESULTS: In the 628 EGVO procedures, sticking of the needle to the varix occurred in 9 cases (1.43%), including 1 case that used lipiodol-diluted NBC and 8 cases that used undiluted NBC (P = 0.000). The needle was successfully withdrawn in 8 cases. Large spurt bleeding occurred in one case, and hemostasis was achieved by two other injections of undiluted glue. The injection catheter became blocked in 17 cases (2.71%) just during the injection, and 4 cases were complicated with the needle sticking to the varix. Large glue adhesion to the endoscope resulted in difficulty withdrawing the endoscope in 1 case. Bleeding from multiple sites was observed in the esophagus and gastric cardia after the endoscope was withdrawn. Hemostasis was achieved by 1% aethoxysklerol injection and intravenous somatostatin. The ligation device stuck to the varices in two cases during the subsequent endoscopic variceal ligation. In one case, the ligation device was successfully separated from the esophageal varix after all bands were released. In another case, a laceration of the vein and massive bleeding were observed. The bleeding ceased after 1% aethoxysklerol injection.
CONCLUSIONS: Although EGVO with tissue glue is usually safe and effective, a series of complications can occur during the procedure that may puzzle endoscopists. There is no standard operating procedure for addressing these complications. The cases described in the current study can provide some reference for others.