Parallel Session 16: PATHOGENESIS OF STEATOHEPATITIS S53 Results: The study population included 3,752 and 2,074 patients with chronic and cleared HCV-infection, respectively. The estimated 5-year probabilities of survival were 0.84 (95% CI: 0.83-0.86) in the chronic HCV group and 0.88 (95% CI: 0.87-0.90) in the cleared HCV group. Dur- ing 26,573 person-years of observation (PYR), 822 persons died (mortality rate = 30.9/1,000 PYR, 95% CI: 28.9-33.1). Chronic HCV -infection was a predictor of increased mortality (adjusted MRR 1.37, 95% CI 1.10-1.71) and liver-related mortality (adjusted MRR 2.23, 95% CI 1.34-3.72) after excluding the initial two years of observation representing lead in mortality. Conclusions: Chronic compared to cleared HCV-infection was associated with increased overall and liver-related mortality. Overall mortality in HCV infected patients Time from HCV-RNA test HCV infection Deaths Person years of observation MR (95% CI) Crude MRR (95% CI) Adjusted MRR (95% CI) Overall Cleared 246 9662 25.5 (22.5-28.8) 1 N.A. Chronic 576 16910 34.1 (31.4-37.0) 1.34 (1.15-1.55) N.A. 0-2 years Cleared 135 3756 35.9 (30.4-42.6) 1 1 Chronic 287 6740 42.6 (37.9-47.8) 1.18 (0.97-1.45) 1.05 (0.85-1.29) 2-10 years Cleared 111 5907 18.8 (15.6-22.6) 1 1 Chronic 289 10171 28.4 (25.3-31.9) 1.51 (1.21-1.88) 1.37 (1.10-1.71) Liver related mortality in HCV infected patients Time from HCV-RNA test HCV infection Deaths Person years of observation MR (95% CI) Crude MRR (95% CI) Adjusted MRR (95% CI) Overall Cleared 54 9662 5.6 (4.3-7.3) 1 N.A. Chronic 162 16910 9.6 (8.2-11.2) 1.71 (1.26-2.33) N.A. 0-2 years Cleared 35 3756 9.3 (6.7-13.0) 1 1 Chronic 84 6740 12.5 (10.1-15.4) 1.34 (0.90-1.98) 1.11 (0.74-1.68) 2-10 years Cleared 19 5907 3.2 (2.1-5.0) 1 1 Chronic 78 10171 7.7 (6.1-9.6) 2.38 (1.44-3.94) 2.23 (1.34-3.72) 130 CONSTRUCTION OF A RISK INDEX FOR HEPATOCELLULAR CARCINOMA IN CHRONIC HEPATITIS C PATIENTS R. Tateishi 1 , Haruhiko Yoshida 1 , S. Shiina 1 , Y. Matsuyama 2 , Hideo Yoshida 1 , R. Masuzaki 1 , T. Sato 1 , E. Goto 1 , J. Imamura 1 , Y. Kondo 1 , T. Goto 1 , F. Kanai 3 , N. Kato 1 , M. Omata 1 . 1 Department of Gastroenterology, The University of Tokyo Graduate School of Medicine, 2 Department of Biostatistics, The University of Tokyo Graduate School of Medicine, Tokyo, 3 Department of Medicine and Clinical Oncology, Chiba University Graduate School of Medicine, Chiba, Japan E-mail: tateishi-tky@umin.ac.jp Background and Aim: In the assessment of risk for hepatocellular carcinoma (HCC) development in chronic hepatitis C (CH-C) patients, various factors should be considered such as grade of liver fibrosis and inflammation, sex, and age. We intended to construct a risk index for HCC in CH-C patients in this study. Methods: We enrolled 1459 patients with CH-C who visited our depart- ment between 1994 and 2004 excluding those who had HCC or a history of it, coinfection with hepatitis B virus, or decompensated diseases. We applied cross-validation methods splitting the patients into two groups: training set and testing set. First we constructed 6 models in the training set, from the simplest to the most complicated, including some or all of the following factors: age, sex body mass index, history of diabetes, ALT, ALP, GGT, total bilirubin, platelet count and alpha-fetoprotein (AFP) by logistic regression using HCC development as the only event. Then we validated the models in the test set by receiver operating characteristics (ROC) curve. Results: During the follow-up period until the end of year 2007, HCC was diagnosed in 397 patients. Among 6 models evaluated, a model including age, sex, ALT, platelet count and AFP was selected that provided area under ROC curve of 0.857 (95% confidence interval, 0.828-0.882) not significantly inferior to more complicated models. Scores are assigned to each of the 5 factors according to the estimated regression coefficient in the model (Table). The risk of HCC development in 10 years in patients with 0-9, 10-15, 16-20 and 21-27 points was estimated to be 0.3%, 20.2%, 46.5% and 72.1%, respectively according to the index. Conclusion: We constructed a simple comprehensive index for the assess- ment of HCC development of CH-C patients useful for daily practice. Points 0 1 2 3 4 5 6 8 10 Age (yo) 30 31-40 41-50 51-60 61-70 >70 Gender Female Male PLT (×10 3 /mL) >200 151–200 101–150 100 ALT (× ULN) 1 1.1-2 2.1-3 AFP (ng/mL) 20 20.1–200 >200 Parallel Session 16: PATHOGENESIS OF STEATOHEPATITIS 131 A STUDY ON A RABBIT MODEL OF STEATOHEPATITIS WITH ADVANCED FIBROSIS N. Kawada, T. Ogawa, R. Shiga. Hepatology, Osaka City University Graduate School of Medicine, Osaka, Japan E-mail: kawadanori@med.osaka-cu.ac.jp Background and Aims: To produce an animal model pathologically similar to human steatohepatitis with advanced fibrosis and to study the effect of ezetimibe, an inhibitor of Niemann-Pick C1-Like 1 (NPC1L1) protein, on the development of inflammation and fibrosis of this model. Methods: Japanese white rabbits were fed 0.75% cholesterol diet including 12% corn oil for 9 months. After the sacrifice, the rabbit livers and serums were obtained and used for chemical and histological evaluations. The remaining liver tissues were utilized for evaluating mRNA or protein expressions of IL-1, IL-10, TNF-alpha, TGF-beta, collagen 1A1, TIMPs, and MMPs by real-time PCR. Oxidative stress was evaluated by the level of lipid peroxide and 4-hydroxynonenal. Poteomics and cDNA microarrays were performed for the comprehensive analyses of molecules involved in this model. In another experimental setting, we tested the effect of ezetimibe (1.8 mg/body/day). Results: The livers looked whitish and the surface exhibited irregularity. Steatosis of hepatocytes was evident. Stainings showed that inflamma- tory cells including RAM-11-positive macrophages accumulated in the sinusoids and the formation of fibrotic septum composed of collagens and SMA-positive activated stellate cells crossed over central and portal veins, indicating the reconstruction of parenchyma. Triglyceride, lipid peroxide, and total cholesterol in sera of model rabbits were increased 1.9-, 38-, and 1,277-fold, respectively, compared to the control. Real- time PCR revealed the increase of mRNA levels of IL-1 (2.6-fold), IL-10 (9-fold), TNF-alpha (6-fold), TGF-beta (4-fold), collagen 1A1 (35-fold),