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

BACKGROUND

Gastric varices are found in patients with portal hypertension. Incidence of bleeding from gastric varices is relatively low, but tends to be more severe, and is associated with higher mortality than esophageal variceal bleeding.

AIMS AND OBJECTIVES

To compare the prevalence and types of gastric varices in cirrhosis versus extrahepatic portal venous obstruction (EHPVO) and the results of endoscopic N-butyl-2-cyanoacrylate (NBC, glue) injection.

METHODS

Eighty six patients presenting with bleeding from gastric varices between August 2010 and August 2011 were retrospectively analyzed.

RESULTS

Of 86 patients, 65% (n = 56) were cirrhotics and 35% (n = 30) had EHPVO. Distribution of types of gastric varices showed GOV1 in 14% (n = 8) of cirrhotics vs. 7% (n = 2) of EHPVO, GOV2 in 80% (n = 45) of cirrhotics vs. 53% (n = 16) of EHPVO, IGV1 in 40% (n = 12) of patients with EHPVO vs. 4% (n = 2) cirrhotics. The patients were treated with NBC injections. The mean volume of glue injected was 3.7 ± 2.58 ml over a median of 1 session (range: 1-8). The total volume of glue required was lower in cirrhotics (3.2 ± 2 ml vs. 4.7 ± 3.1 ml, p < 0.05) than in EHPVO patients. Twenty (36%) of cirrhotics required >1 sessions of glue injection as compared to 17 (57%) of EHPVO patients. Over mean follow up of 12 months, rebleeding (9% vs. 10%) and mortality (11% vs. 3%) were similar in patients with cirrhosis and EHPVO.

CONCLUSIONS

In patients with bleeding from gastric varices, GOV2 is more common in cirrhotics and IGV1 in patients with EHPVO. Patients with EHPVO required higher total volume of glue and more glue sessions for gastric varix obturation.