Background: Intravariceal injection of the tissue adhesive Histoacryl (nbutyl-2-cyanoacrylate) will induce sclerosis of gastric varices, however its role in the treatment of variceal bleeding is not clear. The aim of this study was to document the outcome of patients treated with Histoacryl injection therapy (HIT) at The Wellesley Central Hospital. Methods: The patient records of all patients treated with HIT at the Wellesley Central Hospital were retrospectively reviewed and pertinent data recorded. Results: Between 1986 and 1999, 112 patients were treated with HIT on 149 occasions. The average age was 61 (range 27-80) with 64 men and 50 women. The variceal etiology was alcohol in 51 (46%), hepatitis B in 7 (6%), venous thrombosis in 6 (5%), hepatitis C in 5 (4%), other in 4 (4%) and unknown or unrecorded in 39 (35%). The varices were esophageal in 30 (27%), gastric in 31 (28%), gastric and esophageal in 44 (39%), duodenal in 2 (2%), gastric and duodenal in 3 (2%) and esophageal and duodenal in 2 (2%). Gastric varices were the site of HIT in 76 patients. The indication for endoscopy at index HIT was active bleeding in 72 (64%), recent bleeding in 21 (19%), routine follow up in 18 (16%) and occult bleeding in 1 (1%). Active bleeding was seen at endoscopy on 31 occasions in 25 patients of which 13 were gastric and 12 esophageal. Bleeding was controlled at endoscopy in 28/31 and was not recorded in 3. Bleeding was induced at endoscopy on 8 occasions (banding 3, HIT 3, overtube 1, spontaneous 1), all of which were controlled with initiation of or continued HIT. Stigmata of recent bleeding were present at endoscopy 16 times in 16 patients. Histoacryl was injected as 50% Histoacryl in Lipiodol (Ethiodized oil) (mean 3ml per treatment, range 1-10ml). Adjunct endoscopic treatment was performed on 27 occasions including banding on 25/27 occasions (24 esophageal) and epinephrine and metal clips on one occasion each. One patient had splenic artery embolisation and another had splenectomy and gastrectomy. Upper gastrointestinal tract bleeding within 30 days of the index HIT for gastric varices occurred on 6 occasions in 5 patients (7%). Complications occurred in 2 patients with an esophageal ulcer in 1 and cerebral embolism in 1. Inhospital mortality occurred in 14 patients (12%), due to bleeding in 6/14. Conclusions: HIT is a safe and effective method for the control of bleeding in gastric varices. It may also be an effective method for rescue from variceal bleeding induced by other endoscopic techniques.

(77%). Coded archival sera were tested for HP positivity using a commercial ELISA kit (Wampole Labs, Cranbury, NJ) according to the manufacturer's instructions. Each sample was run in duplicate without knowledge of clinicopathological information. Serology results are reported as positive, negative, or equivocal based on controls provided in each kit. Equivocals, representing 2% of the total, were counted as positive.