Ten years' follow-up of 472 patients following transjugular intrahepatic portosystemic stent-shunt insertion at a single centre.

BACKGROUND: Transjugular intrahepatic portosystemic stent-shunt (TIPSS) is increasingly used for the management of portal hypertension. We report on 10 years' experience at a single centre. METHODS: Data held in a dedicated database was retrieved on 497 patients referred for TIPSS. The efficacy of TIPSS and its complications were assessed. RESULTS: Most patients were male (59.4%) with alcoholic liver disease (63.6%), and bleeding varices (86.8%). Technical success was achieved in 474 (95.4%) patients. A total of 13.4% of patients bled at portal pressure gradients < or = 12 mmHg, principally from gastric and ectopic varices. Procedure-related mortality was 1.2%. The mean follow-up period of surviving patients was 33.3 +/- 1.9 months. Primary shunt patency rates were 45.4% and 26.0% at 1 and 2 years, respectively, while the overall secondary assisted patency rate was 72.2%. Variceal rebleeding rate was 13.7%, with all episodes occurring within 2 years of TIPSS insertion, and almost all due to shunt dysfunction. The overall mortality rate was 60.4%, mainly resulting from end-stage liver failure (42.5%). Patients who bled from gastric varices had lower mortality than those from oesophageal varices (53.9% versus 61.5%, P < 0.01). The overall rate of hepatic encephalopathy was 29.9% (de novo encephalopathy was 11.5%), with pre-TIPSS encephalopathy being an independent predicting variable. Refractory ascites responded to TIPSS in 72% of cases, although the incidence of encephalopathy was high in this group (36.0%). CONCLUSIONS: TIPSS is effective in the management of variceal bleeding, and has a low complication rate. With surveillance, good patency can be achieved. Careful selection of patients is needed to reduce the encephalopathy rate.

Induced hyperammonemia alters neuropsychology, brain MR spectroscopy and magnetization transfer in cirrhosis.

Hyperammonemia is a universal finding after gastrointestinal hemorrhage in cirrhosis. We administered an oral amino acid solution mimicking the hemoglobin molecule to examine neuropsychological changes, brain glutamine levels, and brain magnetization transfer ratio (MTR). Forty-eight metabolically stable patients with cirrhosis and no evidence of "overt" hepatic encephalopathy (HE) were randomized to receive 75 g of amino acid solution or placebo; measurements were performed before and 4 hours after administration. Neuropsychological tests included the Trails B Test, Digit Symbol Substitution Test, memory subtest of the Randt battery, and reaction time. Plasma was collected for ammonia and amino acid measurements, and brain metabolism was studied using proton magnetic resonance (MR) spectroscopy in the first 16 randomized patients. In 7 other patients, MTR was measured. A significant increase in ammonia levels was observed in the amino acid group (amino acid group, 76 +/- 7.3 to 121 +/- 6.4 micromol/L; placebo, 83 +/- 3.3 to 78 +/- 2.9 micromol/L; P <.001). Neuropsychological function improved significantly in the placebo group, but no significant change in neuropsychological function was observed in the amino acid group. Brain glutamate/glutamine (Glx)/creatine (Cr) ratio increased significantly in the amino acid group. MTR decreased significantly from 30 +/-2.9 to 23 +/- 4 (P <.01) after administration of the amino acid solution. In conclusion, an improvement in neuropsychological test results followed placebo, which was not observed in patients administered the amino acid solution. Induced hyperammonemia resulted in an increase in brain Glx/Cr ratio and a decrease in MTR, which may indicate an increase in brain water as the operative mechanism.