Greater Severity of Steatosis Is Associated with a Higher Risk of Incident Diabetes: A Retrospective Longitudinal Study.

BACKGRUOUND: Fatty liver is associated with increased risk of developing type 2 diabetes. We aimed to evaluate whether the severity of hepatic steatosis is associated with incident diabetes. METHODS: We conducted a longitudinal analysis using data from 1,798 participants who underwent a comprehensive health checkup and abdominal computed tomography (CT). We assessed the association between baseline liver attenuation value on non-contrast CT images and risk of incident diabetes. All the participants were categorized into three groups based on the baseline liver attenuation value on non-contrast CT images: without hepatic steatosis (>57 Hounsfield unit [HU]), mild hepatic steatosis (41-57 HU), and moderate to severe hepatic steatosis (≤40 HU). RESULTS: During a median follow-up period of 5 years, 6.0% of the study participants progressed to diabetes. The incidence of diabetes was 17.3% in the moderate to severe hepatic steatosis group, 9.0% in the mild steatosis group, and 2.9% in those without hepatic steatosis. In a multivariate adjustment model, as compared with participants without hepatic steatosis, those with moderate to severe steatosis had a hazard ratio (HR) of 3.24 (95% confidence interval [CI], 1.64 to 4.2) for the development of diabetes, and those in the mild steatosis group had a HR of 2.33 (95% CI, 1.42 to 3.80). One standard deviation decrease in mean CT attenuation values of the liver was associated with a 40% increase in the development of diabetes (multivariate adjusted HR, 1.40; 95% CI, 1.2 to 1.63). CONCLUSION: We found a positive association between severity of hepatic steatosis and risk of incident diabetes. Greater severity of steatosis was associated with a higher risk of incident diabetes.