Non-alcoholic steatohepatitis (NASH) patients are at increased risk for progression to cirrhosis. The aim of this study was to assess all-cause and liver-specific mortality in a cohort of non-alcoholic fatty liver disease (NAFLD) patients.

Biopsy-proven NAFLD patients with and without NASH from two historic databases were included. Clinico-demographic information from the time of biopsy was available. Mortality data were obtained from National Death Index-Plus and used for estimating overall and cause-specific mortality. The non-parametric Kaplan–Meier method with log-rank test and multivariate analyses with Cox proportional hazard model were used to compare cohorts.

Two hundred eighty-nine NAFLD patients were included (50.3 ± 14.5 years old, 39.4 % male, 78.6 % Caucasian, 46.0 % obese, 26.0 % diabetic, 5.9 % with family history of liver diseases). Of these, 59.2 % had NASH whereas 40.8 % had non-NASH NAFLD. NASH patients were predominantly female, had higher aspartate aminotranserase, alanine aminotransferase and fasting serum glucose. During follow-up (median 150 months, maximum 342 months), patients with NASH had higher probability of mortality from liver-related causes than non-NASH NAFLD patients (p value = 0.0026). In the entire NAFLD cohort, older age [aHR = 1.07 (95 % CI = 1.05–1.10)] and presence of type II diabetes [aHR = 2.09 (1.39–3.14)] were independent predictors of overall mortality. However, in addition to age [aHR = 1.06 (1.02–1.10)] having histologic NASH [aHR = 9.16 (2.10–9.88)] was found to be an independent predictor of liver-related mortality. Additionally, presence of type II diabetes was associated with liver-related mortality [aHR = 2.19 (1.00–4.81)].

This long-term follow-up of NAFLD patients confirms that NASH patients have higher risk of liver-related mortality than non-NASH. Additionally, patients with NAFLD and type II diabetes are at highest risk for overall and liver-related mortality.