Background and study aims: Both endoscopic obturation and transjugular intrahepatic portosystemic shunts (TIPSs) have proven to be effective in preventing rebleeding from gastric varices. This study compared the efficacy and complications of these modalities. Patients and methods: Cirrhotic patients with acute bleeding from gastric varices were considered for inclusion. After initial control, eligible patients were randomly allocated to two groups: TIPS (n = 35) and obturation using cyanoacrylate (n = 37). In the cyanoacrylate group, treatment was repeated regularly until gastric varices were obliterated. Patients of both groups received regular follow-up. The end points were gastric variceal rebleeding or death. Results: Stent shunt insertion was successful in all TIPS patients, and mean portal pressure gradient decreased from 21.4 ± 7.5 mm Hg to 7.5 ± 3.5 mm Hg (P < 0.001). Variceal obliteration was achieved in 19 patients in the cyanoacrylate group (51 %) compared with seven TIPS patients (20 %) (P < 0.02). After a median follow up of 33 months, upper gastrointestinal bleeding occurred in 15 TIPS patients (43 %) and 22 cyanoacrylate patients (59 %) (P = 0.12). Rebleeding from gastric varices was encountered in four TIPS patients (11 %) and 14 cyanoacrylate patients (38 %) (P = 0.014; odds ratio 3.6, 95 %CI 1.2 - 11.1). Blood transfusion requirements were lower in the TIPS group than in the cyanoacrylate group (P < 0.01). Survival and frequency of complications were similar in both groups. Conclusions: TIPS proved more effective than glue injection in preventing rebleeding from gastric varices, with similar survival and frequency of complications.

To evaluate the authors' experience with a technique for management of thrombosed Brescia-Cimino arteriovenous fistula. Forty patients with 42 thrombosed arteriovenous fistulas were percutaneously treated. Thrombosis occurred within 24 hours of attempted angioplasty in five fistulas, between 24 and 72 hours in 27, and longer than 72 hours in 10. Thrombosed fistulas were approached in a retrograde fashion followed by direct balloon dilation with 5-8-mm balloon catheters. If retrograde catheterization failed to cross the arterial anastomosis, an antegrade puncture directly into the thrombosed drainage vein close to the anastomosis was performed with ultrasonographic guidance, as an aid to catheterize the arterial inflow. Thrombolytic therapy with infusion of urokinase directly into the thrombus was performed in selected patients with visible thrombus that had compromised blood flow in the partially restored vascular access. Postintervention primary and secondary patency was calculated by using Kaplan-Meier analysis. Patency rates between patients without and with urokinase infusion were examined by using the log-rank test. Anatomic success was achieved in 39 (93%) of 42 fistulas; and clinical patency, in 38 (90%) of 42 fistulas. Postintervention primary and secondary patencies (including initial technical failure) at 6, 12, and 18 months were 81% and 84%, 70% and 80%, and 63% and 80%, respectively. No significance of patency rate between patients without and with urokinase infusion was found (P =.912). Three patients died of unrelated causes at 1, 2, and 5 months after the procedures. No major complications were encountered. High anatomic success and excellent clinical patency can be achieved in the salvage of thrombosed arteriovenous fistulas. Percutaneous restoration of arteriovenous fistulas should be attempted before surgical recreation to optimize outcome in patients undergoing hemodialysis.

Hepatic resection for huge hepatocellular carcinoma (HCC) is challenging. The role of multimodality nonsurgical therapy for HCC larger than 10 cm is unclear. We retrospectively investigated 131 HCC patients with main tumors larger than 10 cm in diameter seen between October 1990 and October 2001. Fifty-six patients (group A) underwent hepatectomy and 75 patients (group B) underwent nonsurgical multidisciplinary therapy including hepatic arterial infusion, transcatheter arterial embolization, and percutaneous acetic acid injection. Patients in group B were older, had lower serum albumin levels, and there were more patients with liver cirrhosis and great vessel invasion. Median survivals of group A and B patients were 17 months and 7 months, respectively (p < 0.001). But the 1-, 3-, 5-year survival rates in group B using 38 patients undergoing 3 or more sessions of nonsurgical treatment were not significantly worse than those for group A using 53 patients with followup (57.1%, 19.0%, 16.3% versus 60.7%, 24.5%, 24.5%, respectively). Group A patients had 37.7% and 71.7% recurrence rates at 6 and 12 months, respectively, after operation, and they had a significantly higher frequency of overall extrahepatic recurrence compared with group B patients (43.4% versus 18.7%, p = 0.005). In group B, only 3 of 35 patients younger than 60 years had tumor shrinkage after nonsurgical treatment modalities in comparison to 17 of 40 patients in the elderly group (p = 0.003). Younger patients had a significantly higher prevalence of hepatitis B surface antigen positivity (85.7% versus 47.5%) and infiltrating tumor growth pattern (74.3% versus 45.0%) compared with older patients. Our study suggests that the advantage of hepatic resection in patients with huge HCC is marginal. An effective adjuvant therapy is needed to improve outcomes after hepatic resection. The experience in using nonsurgical treatment shows that the result is poor in young patients compared with that in elderly patients.