Patency of stents covered with polytetrafluoroethylene in patients treated by transjugular intrahepatic portosystemic shunts: long-term results of a randomized multicentre study.

An 80% dysfunction rate at 2 years limits the use of transjugular intrahepatic portosystemic shunts (TIPS) in the treatment of complications of portal hypertension. The use of covered stents could improve shunt patency; however, long-term effect and safety remain unknown. Eighty patients randomized to be treated by TIPS either with a covered stent (Group 1) or an uncovered prosthesis (Group 2) were followed-up for 2 years. Doppler US was performed every 3 months. Angiography and portosystemic pressure gradient measurement were performed every 6 months or whenever dysfunction was suspected. Actuarial rates of primary patency in Groups 1 and 2 were 76% and 36% respectively (P=0.001). Clinical relapse occurred in four patients (10%) in Group 1 and 12 (29%) in Group 2 (P<0.05). Actuarial rates of being free of encephalopathy were 67% in Group 1 and 51% in Group 2 (P<0.05). Probability of survival was 58% and 45% at 2 years, respectively, in Groups 1 and 2 (NS). The mean Child-Pugh score improved only in Group 1 (from 8.1+/-1.6 to 7+/-2.2 at 2 years -P<0.05). We also compared the Doppler-US parameters between patent and dysfunctioning shunts. In patent shunts, the mean velocity within the portal vein was significantly higher but the performance of Doppler-US was not accurate enough to predict shunt dysfunction. In conclusion, the improvement in TIPS patency by using covered prostheses is maintained over time with a decreased risk of encephalopathy, while the risk of death was not increased.

Differential effects of ammonia on the benzodiazepine modulatory site on the GABA-A receptor complex of human brain.

Ammonia is a key factor in the pathogenesis of encephalopathies associated with liver failure. A direct effect of ammonia on GABAergic neurotransmission was proposed as a mechanism that may explain its neurotoxic effect on the basis of electrophysiological and biochemical studies performed in animal models of liver failure. In the present study, we investigated using a radiometric assay the effect of ammonia on the binding of GABA-A receptor ligands to membranes from normal human brains. Ammonium tartrate significantly decreased the maximal binding of [3H]flunitrazepam to well-washed frontal cortical membranes (366+/-63 fmol/mg protein in absence of ammonia versus 294.1+/-51 fmol/mg protein in presence of 2 mM ammonia; p<0.05). The efficacy of the effects of ammonia was within the millimolar range (IC50=4.8 mM). This effect was not seen in cerebellum or hippocampus. Ammonia exposure decreased the maximal binding of [3H]flumazenil (284.9+/-24.2 fmol/mg protein in absence of ammonia versus 146.4+/-15.6 fmol/mg protein in presence of 2 mM ammonia; p<0.01). This effect was seen with a greater potency (Imax=32.4%) and a lower IC50 (0.1 mM). Inhibition of [3H]flumazenil binding was significant in all brain regions. The apparent ammonia-induced decrease of [3H]flunitrazepam and [3H]flumazenil binding was due to a decrease in the binding affinities of these ligands for the benzodiazepine site. In contrast, ammonium tartrate exposure did not cause significant changes to the binding of [3H]muscimol in any brain region. These findings demonstrate that ammonia interacts negatively with components of the benzodiazepine-associated site at the GABA-A receptor complex in human brain in contrast to previous reports in the rat, and thus, does not support the notion that ammonia directly activates the GABA-A receptor complex resulting in increased GABAergic neurotransmission in human hepatic encephalopathy. These findings also suggest that positron emission tomography studies in cirrhotic patients using [11C]flumazenil may be underestimating GABA-A receptor sites depending upon the degree of hyperammonemia of the patient.