The aim of the current study was to evaluate the association between changes in non-alcoholic fatty liver disease (NAFLD) over time and risk of incident diabetes mellitus (DM). In total,… Click to show full abstract
The aim of the current study was to evaluate the association between changes in non-alcoholic fatty liver disease (NAFLD) over time and risk of incident diabetes mellitus (DM). In total, 3047 subjects without underlying DM were followed up for 14 years from the Anseong-Ansan cohort. NAFLD status was determined biennially using the hepatic steatosis index (HSI), and subjects were clustered into seven groups according to changes in HSI, body mass index (BMI), and homeostatic model assessment of insulin resistance (HOMA-IR): none, persistent, transient, transient resolved, resolved, incident, and recurrent NAFLD (Groups 1–7, respectively). Predictive abilities were compared between the dynamics of HSI and single time points. Regarding the changes in HSI, the risk of incident DM was highest in Group 2 (hazard ratio [HR] 2.710; P < 0.001), followed by Groups 7 (HR 2.062; P < 0.001) and 3 (HR 1.559; P = 0.027). The predictive ability for DM was powerful in order of HOMA-IR, HSI and BMI. The dynamics of NAFLD were less predictive of incident DM than single time-point NAFLD. In conclusion, NAFLD is more useful than BMI in predicting incident DM. However, NAFLD status at single time points can better predict incident DM than dynamic changes in HSI.
               
Click one of the above tabs to view related content.