A new composite model including metabolic syndrome, alanine aminotransferase and cytokeratin-18 for the diagnosis of non-alcoholic steatohepatitis in morbidly obese patients

P>Background Non-invasive approaches are useful to differentiate simple steatosis from non-alcoholic steatohepatitis (NASH) in obese and morbidly obese patients. Aim To develop a new scoring system to diagnose definitive NASH. Methods Preoperative clinical and biological data including serum caspase 3-generated cytokeratin-18 fragments (CK18) and surgical liver biopsies were obtained from 464 morbidly obese patients who had undergone bariatric surgery. The cohort was divided into two groups: training group (n = 310) and validation group (n = 154). Definitive NASH was defined according to Kleiner's classification with a Non-alcoholic fatty liver disease Activity Score (NAS) >= 5. Results Alanine aminotransferase (ALT), CK18 fragments and the presence of metabolic syndrome were independent predictors for discriminating patients with NAS >= 5 in the training group. These three parameters were used to carry out a scoring system for the prediction of NAS >= 5. Whereas serum CK18 fragment alone had an area under the receiver operating characteristic (AUROC) curve = 0.74, AUROC curves of the scoring system were 0.88 and 0.83 in the training group and the validation group, respectively. Conclusion A simple and non-invasive composite model (the Nice Model) including metabolic syndrome, ALT and CK18 fragments is able to predict accurately a non-alcoholic fatty liver disease activity score >= 5 in morbidly obese subjects.