Validation of the non-alcoholic fatty liver disease activity score

Background The non-alcoholic fatty liver disease (NAFLD) activity score (NAS) is a scoring system designed by the Nonalcoholic Steatohepatitis (NASH) Clinical Research Network (CRN) to encompass the spectrum of NAFLD and evaluate histological changes. However, the NAS and the correlation between the NAS and a diagnosis of NASH have not been validated outside the NASH CRN. Aim To validate the NAS outside the NASH CRN. Methods This study retrospectively examined liver biopsies from adults with NAFLD or steatohepatitis obtained from January 2003 to May 2010. Biopsy specimens were evaluated twice in a blinded manner by a single hepatopathologist, once to determine a diagnosis (steatohepatitis or steatosis/not-steatohepatitis), and a second time to determine the NAS. Results A total of 386 liver biopsies were evaluated. Mean age of patients at time of biopsy was 49.9 +/- 10.2 years. NASH was found in 51% of the patients. For NAS >= 5 as a diagnosis of steatohepatitis and NAS <5 for not-steatohepatitis, the sensitivity was 57%, specificity: 95%, negative predictive value (NPV): 68% and positive predictive value (PPV): 93%. Lowering the NAS to >= 4 as a diagnosis of steatohepatitis increased the sensitivity to 85% with a decrease in specificity to 81%; NPV: 84%, PPV: 82% and Cohen's kappa 0.658. Conclusions The NAFLD activity score is a valid scoring system encompassing the spectrum of NAFLD with an excellent level of agreement between the histological diagnosis and the NAFLD activity score. A NAFLD activity score >= 4 has optimal sensitivity and specificity for predicting steatohepatitis, and is the recommended value for admission into an interventional trial for NASH.