S368 POSTERS 1.9-4.2), p < 0.001 of weight gain, significant reduction in ALT con- centrations was noted, median 49 U/L (IQR 34-60), p < 0.001. Similarly, GGT improved mean reductions, men: 26 U/L (95% CI 11-40), p < 0.001; women: 31 U/L (17-45), p < 0.001. 44 patients remained on pioglitazone after 4 years. Of these, 23 were those with abnormal LFTs at baseline in whom significant improvement in ALT mean reduction, 32 U/L (95% CI 16-47), p = 0.008, and HbA1c mean reduction, 1.2% (95% CI 0.7-1.8), p < 0.001 were observed despite 8.3 kg (95% CI 2.6-11.5), p = 0.004, of weight gain over 4 years. However, improvement in GGT was only sustained in men {mean changes, men: -44 U/L [95% CI (-77)–(-11)], p = 0.04; women: -29 U/L [95% CI (-71)– (+13)], p = 0.33}. There were significant correlations between changes in HbA1c and ALT (r = 0.23, p = 0.005), and between changes in HbA1c and GGT (r = 0.35, p < 0.001). Conclusion: Pioglitazone treatment resulted in sustained reductions in ALT and, at least in men, GGT concentrations despite significant weight gain over 4 years. These changes were associated with improvement in HbA1c, thus supporting the role played by insulin resistance in NAFLD. 986 CHRONIC ETHANOL CONSUMPTION MODULATE THE EXPRESSION OF PEROXISOME PROLIFERATOR- ACTIVATED RECEPTOR (PPAR) IN C57BL/6J ANIMAL MODEL J.E. Yeon, Jeong H. Kim, Y.K. Jung, M.K. Joo, Ji H. Kim, J.J. Park, J.S. Kim, Y.T. Bak, K.S. Byun. Department of Internal Medicine, Korea University College of Medicine Guro Hospital, Seoul, South Korea E-mail: 93haan@hanmail.net Background and Aims: Alcohol induced liver injury is caused by increased intracellular NADH concentration that lead to the accumulation of lipid in the hepatocyte. Nonalcoholic fatty liver disease (NAFLD), that have similar pathologic findings, is caused by insulin resistance. Peroxisomal proliferator-activated receptor (PPAR) regulate many factors related lipid and glucose metabolism. PPAR agonist have been tried for treatment of NAFLD. But In chronic alcoholic liver disease, role of PPAR agonist is not established. The aim of this study was to evaluate the expression of PPAR in liver, muscle and adipose tissue in chronic ethanol fed animal model. Methods: C57Bl/6J mice were divided control group (Gr.C: control group, n = 9) and alcohol group (Gr.E: ethanol group, n = 19). Mices were fed Lieber-deCarli control and ethanol diet and ethanol accounted for 18% of the calorie content. After 8 weeks, the liver, fat and muscle tissues were obtained. PPAR-a, d/b, g, soluble carrier family 2, member 4 (Glut4), carnitine palmitoylacyltransferase (CPT) I, long chain acylcoA dehydrogenase (LCAD), acylcoA oxidase (AOX), Fatty acid transport protein (FATP), uncoupling protein 3 (UCP3) mRNA were quantified by real time PCR. Results: In Gr.E, the PPAR-a protein expression in the liver was reduced to 65% of Gr.C. The PPAR-g mRNA and protein expressions in the liver or the muscle of Gr.E didn’t show significant change compared with Gr.C. The PPAR-d/b mRNA and protein expressions in muscle was significantly reduced to 30% of Gr.C. In Gr.E, mRNA expression was reduced compared with Gr.C as follows; the GLUT4 20% in muscle, the CPT 50% in fat. the UCP3, AOX, FATP and LCAD in Gr.E was significantly elevated compared with Gr.C. Conclusions: Chronic alcohol consumption in animal model reduced PPAR-d expression and change mRNA expression of various proteins that related to free fatty acid transport and oxidation, glucose metabolism, electron transport of mitochondria. So, further studies are warranted about the role of PPAR agonist treatment in chronic alcoholic liver disease. 987 A COMPARISON BETWEEN THE HEPATORENAL ULTRASOUND INDEX AND STEATOTEST FOR THE NON-INVASIVE QUANTIFICATION OF LIVER STEATOSIS S. Zelber-Sagi 1 , M. Webb 1 , V. Ratziu 2 , T. Poynard 2 , Z. Halpern 1 , R. Oren 1 . 1 Dept. Gastroenterology, Tel Aviv Sourasky Medical Center, Tel-Aviv, Israel; 2 Universit´ e Pierre et Marie Curie, Hˆ opital Piti´ e Salpˆ etri` ere, Paris France E-mail: zelbersagi@bezeqint.net Background: Quantification of liver steatosis is clinically relevant in diseases such as NAFLD but cannot be done by regular ultrasound, which only provides a qualitative assessment with mediocre sensitivity and significant observer variability. The aim was to assess the hepatorenal ultrasound index (HRI) for ultra- sonographic quantification of liver steatosis in comparison with SteatoTest (Biopredicitve, France) (ST), a biochemical surrogate marker of liver steatosis (Poynard, Comp Hepatol, 2005;4:10). Methods: A cross-sectional study of a sub-sample of the Israeli National Health Survey. Exclusion criteria were any known etiology for liver dis- ease. Participants underwent an abdominal ultrasound that was performed by the same operator. The HRI, which has been previously validated against liver biopsy (AASLD, Boston 2006), was calculated based on the ratio between the echogenicity of the liver and that of the right kidney cortex. ST classes were S0 (equivalent to no histological steatosis), S1 (1-5%), S2 (6-33%) and S4 (34–100%). Results: Three hundred and forty volunteers (53% male, 31% with a bright liver pattern) were included. ST predicted the presence of steatosis in 33% (S1 and above) and of advanced steatosis in 21% (S2 and above). There was a significant correlation between the HRI and ST (r = 0.51, P < 0.001). HRI provided a reliable quantification of steatosis as predicted by ST. The mean HRI was: 1.1 for S0, 1.3 for S1, 1.5 for S2, 1.8 for S3-S4 (P  0.03). The diagnostic accuracy of HRI for the diagnosis of steatosis above 5%, as predicted by ST, was then calculated. The area under the ROC curve was 0.83 (0.77-0.9, 95% CI). The optimal HRI cut-off for the prediction of steatosis above 5% was 1.2, with a sensitivity of 86% and specificity of 71%. Almost all (98%) patients with steatosis on ultrasound (bright liver pattern) and 17% of patients with no bright liver pattern had HRI values above this HRI threshold. Conclusion: HRI is a sensitive and quantitative non-invasive method for steatosis quantification. Remarkably, HRI can diagnose small amounts of liver fat that would be missed by conventional ultrasound. Non-invasive quantitative diagnostic methods for liver steatosis should be validated in future studies and therapeutic trials.