Prevalence and predictors of non-alcoholic steatohepatitis in patients with morbid obesity.

Abstract

BACKGROUND\nNon-alcoholic fatty liver disease (NAFLD) is highly prevalent in morbid obesity (MO). A considerable proportion of patients with MO have non-alcoholic steatohepatitis (NASH). Liver biopsy (LB) is the only procedure that reliably differentiates NASH from other stages of NAFLD, but its invasive nature prevents it from being generalisable. Hence, non-invasive assessment is critical in this group of patients.\n\n\nOBJECTIVES\nTo report NAFLD/NASH prevalence in a cohort of patients with MO and to identify predictors of NASH.\n\n\nMETHODS\nFifty-two consecutive patients subjected to bariatric surgery in a University hospital in Spain underwent LB. Anthropometric, clinical and biochemical variables were registered. According of the results of the LB, individuals were classified by whether they had NASH or not. Multiple logistic regression analysis was performed to identify independent factors associated with NASH.\n\n\nRESULTS\nNAFLD was reported in 94.2% of the patients, simple steatosis was present in 51.92% and NASH in 42.31%. Meanwhile, 17.3% of patients exhibited significant fibrosis (≥F2). HIGHT score for NASH risk was established using five independent predictors: systemic Hypertension, Insulin resistance, Gamma-glutamyl transferase, High density lipoprotein cholesterol and alanine Transaminase. This score ranges from 0 to 7 and was used to predict NASH in our cohort (area under the receiver operator characteristic curve 0.846). A score of 4 or greater implied high risk (sensitivity 77.3%, specificity 73.3%, positive predictive value 68%, negative predictive value 81.5%, accuracy 75%).\n\n\nCONCLUSIONS\nNAFLD is practically a constant in MO with a considerable proportion of patients presenting NASH. The combination of five independent predictors in a scoring system may help the clinician optimise the selection of patients with MO for LB.