Gastric Varices in Cirrhosis vs. Extrahepatic Portal Venous Obstruction and Response to Endoscopic N-butyl-2-cyanoacrylate Injection

BackgroundGastric varices (GV) are found in patients with portal hypertension. Incidence of bleeding from GV is relatively low, but is severe, and associated with higher mortality.Aims and objectivesTo compare the types of GV in cirrhosis vs. extrahepatic portal venous obstruction (EHPVO) and the results of endoscopic N-butyl-2-cyanoacrylate (NBC, glue) injection.MethodsFour hundred and fifty-four patients undergone glue injection of GV for primary prophylaxis or control of bleeding for first episode of bleeding from GV between August 2010 and August 2015 were analyzed.ResultsOf 454 patients, 64% (n = 292) were cirrhotics and 36% (n = 162) had EHPVO. Types of GV were GOV1 in 16.4% (n = 48) of cirrhotics vs. 7.4% (n = 12) of EHPVO, GOV2 in 76.7% (n = 224) of cirrhotics vs. 53.1% (n = 86) of EHPVO, Isolated gastric Varices (IGV1) in 39.5% (n = 64) of patients with EHPVO vs. 6.8% (n = 20) cirrhotics. The patients were treated with NBC injections. The mean volume of glue injected was 2.89 ± 1.59 ml over a median of 1 session (range: 1-7). The total volume of glue required was lower in cirrhotics (2.44 ± 1.17 ml vs. 3.69 ± 1.91 ml, P < 0.05) than in EHPVO patients. One hundred and seventeen (40.1%) of cirrhotics required >1 sessions of glue injection as compared to 102 (63%) of EHPVO patients. Over mean follow up of 14.7 ± 6.46 months, rebleeding (10% vs. 13%) was similar in patients with cirrhosis and EHPVO and mortality (15.4% vs. 2.5%) was higher in cirrhotics than EHPVO.ConclusionIn patients with bleeding from GV, GOV2 are more common in cirrhotics and IGV1 in patients with EHPVO. Patients with EHPVO required higher total volume of glue and more glue sessions for GV obturation.