Long-Term Outcome and Safety of Histoacryl Glue Sclerotherapy for Bleeding Gastric Varices

Abstract

Prior to the introduction of N-butyl-2-cyanoacrylate (Histoacryl) glue, the prognosis of gastric variceal haemorrhage was poor. Long term efficacy and safety of Histoacryl has been established in several Asian studies but there is no comparable data in Western populations. We thus report our experience in using glue to manage gastric variceal haemorrhage.A retrospective review of records of patients seen with bleeding gastric varices treated with Histoacryl from 1994-2005 was undertaken. Histoacryl was injected as a 1:1 mixture with lipiodol using a 21-gauge sclerotherapy needle. Twenty-one patients were analysed. Median follow up was 30 months (11 days-88 months). Data collected included patient demographics, Child-Pugh (CP) score, cause of portal hypertension, variceal location and transfusion requirements. Primary haemostasis was defined as control of bleeding at the time of endoscopy. Early in the study, repeat sclerotherapy was only undertaken for re-bleeding; later, repeat treatments were performed to obliterate varices. The median age was 54 years (range 32-79). Seventeen patients had cirrhosis, 13 secondary to alcohol and HCV; 4 patients had extrahepatic portal hypertension. CP score at presentation was A in 3, B in 8 and C in 6. Three patients had portal vein and 2 others splenic vein thrombosis. All received octreotide at presentation. Seventeen patients were actively bleeding and 4 had evidence of recent bleeding. Median blood transfusion requirement was 9 units (range 0-37 units). Location of varices was fundal in 17 patients and proximal lesser curve in 4 with 17 having associated oesophageal varices. Primary haemostasis was achieved in 18 (86%). Of the others, 1 achieved haemostasis with re-treatment, 1 proceeded to surgery, and the final patient died. Early re-bleeding occurred in 5 (25%), with successful re-treatment in 4 (80%); the final patient required a TIPS. Median injected glue volume per session was 5 ml (2-9 ml). Complications included 3 patients with radiologic but not clinical evidence of distant glue embolisation (pulmonary vessels 2, iliac and adrenal vein 1). There was no treatment-related mortality. Survival at 30-days was 90%, and 71% at 1 year; most deaths liver related. During long term follow up, the only re-bleed was at 13 months in a patient with splenic vein thrombosis who had a curative splenectomy; there were no long term glue-related sequelae. Our results confirm that, in a Western population, Histoacryl sclerotherapy is an effective and safe treatment for bleeding gastric varices with results similar to Asian studies. It should be considered as first line treatment for gastric variceal bleeding.