P808 A WEB-BASED PROGRAM FOR IMPLEMENTING LIFESTYLE CHANGES TOWARDS HEALTHY DIET AND HABITUAL PHYSICAL ACTIVITY IN NAFLD

activity. It has been used as a noninvasive measure of hepatic steatosis, but has not been widely validated and not examined at all in a multiethnic population. We evaluated the FLI using populationbased data on non-Hispanic whites, non-Hispanic blacks, and Mexican Americans from the third U.S. National Health and Nutrition Examination Survey, 1988–1994. We also explored whether a better fitting model for the U.S. population could be derived. Methods: We studied 5,676 adults negative for viral hepatitis markers who were examined in the morning after an overnight fast and had abdominal ultrasound data on hepatic steatosis. Results: The prevalence of moderate-severe hepatic steatosis was 19%. For the FLI, a receiver operating characteristic (ROC) curve had an area under the curve (AUC) (95% confidence interval (CI)) of 0.780 (0.751–0.801) (Figure), compared to an AUC of 0.84 (0.81–0.87) in the Italian data. A new index generated from logistic regression analysis that included age, race-ethnicity, waist circumference, diabetes status, fasting insulin, and alanine aminotransferase (ALT) activity predicted hepatic steatosis with an AUC (95%CI) of 0.810 (0.784–0.829) (Figure). Comparing the curves, at 90% specificity the new curve had better sensitivity (0.55 vs. 0.49) and at 90% sensitivity it had better specificity (0.39 vs. 0.27). Excluding heavy drinkers had little effect. Conclusions: In the multiethnic U.S. population, an index incorporating age, race-ethnicity, waist circumference, diabetes, insulin, and ALT predicted hepatic steatosis more accurately than the FLI.