$84 POSTERS Conclusions: Noradrenaline seems to be as effective and safe as terli- pressin in patients with HRS. If confirmed, these results could support the use of noradrenaline, a cheap and easily available drug, in the management of these patients. I• HEPATIC ENCEPHALOPATHY AFTER TRANSJUGULAR INTRAHEPATIC PORTOSYSTEMIC SHUNT (TIPS): PTFE-COVERED STENT-GRAFTS VERSUS BARE STENTS S. Angeloni 1, G. Nicolini 1, E Nicolao 1, L. Ridola 1, EM. Salvatori 2, E Fanelli 2, A.E Attili 1, M. Merli 1 , O. Riggio 1. 1H Gastroenterology, 2Dept. of Radiology, University "La Sapienza" of Rome, Italy Background and Aims: PTFE-covered stent-grafts appear to have the potential to improve TIPS patency. However, to a greater efficacy in the control of variceal rebleeding and ascites/hydrothorax, a long-term patency of the shunt might also increase the risk of hepatic encephalopathy. The aim of this study was to prospectively assess the incidence of hepatic encephalopathy (HE) with the new PTFE-covered stent-grafts in comparison with a group of controls treated with bare stents. Patients and Methods: Eighty-seven cirrhotics were treated with bare stents and 61 with PTFE-covered stent-grafts; all patients were followed by the same medical team according to a prospective protocol for a diagnostic work-up and a surveillance strategy. Results: The two groups were comparable for age, sex, etiology, and severity of cirrhosis. At least one episode of HE occurred in 27 of the 61 patients in the study group and in 55 of the 87 patients in the control group. The probability of remaining free of HE was similar in both groups: 52.9% (95%CI: 39.4 66.4) at one year and 49.8% at two years (95%CI: 35.7 63.8%) inthe stent-graft group versus 41.1% (95%CI: 30.4 51.8) at one year and 38% (95% CI: 32.5 48.7%) at two years in the control group; Log rank test, p 0.30. However, the total number of HE episodes was higher in the covered stent group (3.7 vs 6.3 episodes/100 month patient). Moreover, six of the 61 patients treated with covered stents and none in the bare stent group suffered of recurrent HE which did not respond to the medical therapy and required shunt reduction (Chi-square 7.53; p 0.002). At multivariate analysis, the use of covered stent-grafts and high level of serum creatinine were shown to be independent factors for the occurrence of persistent intractable HE. The cumulative probability of remaining free of shunt dysfunction in PTFE-covered stent group was significantly higher than in control group: 83.7% (95%CI: 73.3 94.1%) versus 48.2% (95%CI: 36.8 59.6%) at two years, respectively; Log rank test, p 0.0003. Cumulative survival was similar in both groups. Conclusions: PTFE-covered stent-grafts are more likely to remain patent over time, but severe recurrent HE requiring the shunt reduction is a new emerging complication of these covered stents. I-~ ASSESSMENT OF PROGNOSIS IN CRITICALLY ILL ADULT PATIENTS WITH CHRONIC LIVER DISEASE REQUIRING INTENSIVE CARE ADMISSION, A SINGLE CENTRE EXPERIENCE OF 363 PATIENTS M. Austin 1, W. Bernal 1, G. Auzingel 2, E. Size12, J. Wendon 1. 1Institute of Liver Studies, 2Liver Intensive Therapy Unit, King k College Hospital, Denmark Hill, London, UK Background: Intensive care unit (ICU) admission is limited, expensive and often refused for complications of advanced chronic liver disease (CLD). The identification of patients who benefit from ICU support is important and determined by organ failure (OF) scoring systems. In a cohort of patients admitted to a Liver ICU we examined early predictors of ICU survival and the accuracy of OF scoring systems. Methods: Consecutive patients with CLD admitted to ICU between 1/1/1999 and 1/3/2005 were studied. Demographic features, indication for admission and severity of CLD and OF were examined. Discriminative power for prediction of ICU survival with MELD, Child@ugh (CP), APACHE II and SOFA scores generated were examined using receiver operating characteristic (AUROC) curves. Statistical testing with non- parametric, data is n (%) or median (inter-quartile range) including univariate and multiple logistic regression analysis. Results: 363 patients (39% female) were studied; median age was 50yrs (40 58), CP score 12 (10 13), MELD 25 (17.7 32.1), SOFA 12 (10 15) and APACHE II 23 (17 29). 110 (30%) admitted with upper GI bleeding (UGIB) and 225 (62%) for other indications. Overall survival was 138/363 (38%), median length of ICU stay of 4 days (2 10) in survivors and 9 (3 17) in non-survivors. Surviving patients had lower CP, SOFA, MELD and APACHE II scores (all p < 0.00001). 58/110 (53%) admitted with UGIB survived (p <0.0003), had lower OF scores (all p <0.0001) and fewer CP grade C disease (p 0.002). Multiple logistic regression highlighted cardiovascular (cvs) (OR 4.4, p <0.0000001), renal (OR 3.4, p<0.0000001) and neurological (cns) (OR 3.6, p<0.0000001) organ failure as key indicators of prognosis (3 key OF 98% mortality). Age > 50 (OR 2 p 0.012) or CP > 11 (OR 2.5 p <0.001) with 1 key OF predicted 70% mortality. On ROC analysis AUROC was 0.807 for SOFA, 0.780 for APACHE II, 0.726 for MELD and 0.721 for CP scores with overlap of 95% confidence intervals between all scores. Conclusions: Patients with advanced CLD requiring ICU care had high mortality. Discrimination between survivors and non-survivors was great- est using the SOFA score, but the presence of key organ failures on admission provided a simple tool for predicting outcome in this group of patients. I-~ SAFE USE OF PEG-IFN AND RIBAVIRIN IN HCV- CIRRHOTIC PATIENTS UNDERGOING PARTIAL SPLENIC EMBOLIZATION R. Bfircena 1 , A. Moreno 2, J.R. Foruny1, C. Quereda 2, J. Blazquez 3, L.A. Gil-Grande 1, J. Moreno 1, M.J. Perez-Elias 2, A. Antela 2, J. Sanchez 3, M. Rivero 1, M. Rodriguez-Gandia 1, S. Moreno 2. 1Servicio de Gastroenterologia, 2Servicio de Enfermedades Infecciosas, S Servicio de Radiologia Vascular Intervencionista, Hospital Ramon y Cajal, Madrid, Spain Background and Alms: Partial splenic embolization (PSE) is a non- surgical alternative for HCV-cirrhotic patients with hypersplenism and pancytopenia, especially thrombocytopenia, that precludes the use of pegylated-interferon (peg-IFN)-based therapies. We aimed to assess the safety and efficacy of peg-IFN/RBV in cirrhotic patients undergoing PSE specifically aimed to improve blood parameters enough to allow peg- IFN/RBV. Methods: Between May 2002 and October 2005, 18 HCV-cirrhotic pa- tients have undergone PSE at a tertiary center in Madrid, Spain, prior to peg-IFN/RBV. We report the outcomes of 16 patients (89%) that have already started therapy, with complete follow-up data in 9 (56%). Results: The mean MELD score and age before PSE were 11+3.56 (6 17) and 43+9 years; 12 subjects were male (75%), and 8 were HIV-coinfected (50%). The most frequent HCV genotype was 1 (n 12, 75%). Five patients (31%) underwent PSE and started HCV-therapy following liver transplantation. Before PSE, the mean hemoglobin, platelet and neutrophil values were 12.5+1.5 g/dl, 47,150+14,304 cells/ml (below 50,000 cels/ml in 9 cases, 56%) and 1351+888 cells/ml. PSE not only improved the mean platelet (p 0.0001), neutrophil (p 0.005) and hemoglobin (p 0.002) levels, but also the mean prothrombin activity (p 0.001), INR (p 0.024) and MELD score (p 0.01). A mean of 15+9 weeks after PSE all patients started weight-adjusted RBV (mean dose 14.40+3.17 mg/kg/day) and full- dose peg-IFN-~-2b (n 8) or peg-IFN-~-2a (n 8). Baseline mean HCV- RNA was 5.7+0.91 logl0 IU/ml. Peg-IFN dose has not been reduced in any subject, even in 3 patients (19%) with early withdrawal after 13, 32 and 70 days, respectively. Five patients (31%) received erythropoietin (with RBV dose adjustments in only one) due to hemoglobin below 10 g/dl.