Hyaluronic acid (HA) and tissue inhibitor of metalloproteinase 1 (TIMP-1) are reliable markers of liver fibrosis and are closely linked to the proinflammatory status. In this pilot cohort study, we attempted to identify a clinical score that would predict the severity of nonalcoholic fatty liver disease (NAFLD) based on clinical variables and serum markers of fibrosis and inflammation. The cohort included 46 patients with histologically confirmed NAFLD (76.1% male; mean age, 43+/-13 years; mean body mass index [BMI], 27.8+/-3.5). Serum transforming growth factor beta (TGF-beta), HA, TIMP, and matrix metalloproteinase (MMP) levels were measured with commercial enzyme-linked immunoassay (ELISA) kits. Demographic features and clinical and laboratory findings were subjected to univariate and multivariate binary logistic regression analysis to construct the mathematical model. Receiver operating characteristic curve (ROC) analysis was used to identify a threshold value for diagnosis of NASH and to assess its sensitivity and specificity. Serum levels of HA and TIMP-1 were statistically different in patients with nonalcoholic steatohepatitis (NASH) (P<0.05). Logistic regression analysis of several clinical variables indicated patient age as the only independent predictor of NASH (odds ratio [OR], 1.129, 95% confidence interval [CI], 1.019-1.251, P=0.020). The mathematical model constructed on the basis of these results included age, TIMP-1, and HA levels. A value of 148.27 or more identified patients with NASH with 85.7% sensitivity, 87.1% specificity, and negative and positive predictive values of 96.4% and 60%, respectively. This model seems to represent a reliable noninvasive tool for excluding the presence of NASH. If validated in larger prospective cohort studies, it might be useful for determining when a liver biopsy is actually warranted in patients with NAFLD.