Importance of liver fat indices for diagnosis of hepatic steatosis

Abstract

Nonalcoholic fatty liver disease (NAFLD) strongly associates with obesity and type 2 diabetes and may lead to cirrhosis and hepatocellular carcinoma. Fatty liver indices based on laboratory and anthropometric parameters are increasingly used to predict steatosis in clinical studies, which requires solid validation. To this end, we examined the accuracy of three recently published indices – NAFLD liver fat score, fatty liver index and hepatic steatosis index – by comparison with exact quantification of liver fat content (HCL) by 1H-magnetic resonance spectroscopy (1H-MRS) in 92 non-diabetic humans (58 ± 12 years, BMI 26 ± 4 kg/m2). A 4-point 75-g oral glucose tolerance test was performed to assess insulin sensitivity and β-cell function. Employing 1H-MRS, 17 out of 92 subjects had fatty liver. HCL ranged from 0.03 to 39.01% with a median of 2.5 [0.6;4.2] and correlated with insulin sensitivity (composite whole body insulin sensitivity index, ISIcomp; r =-0.46, p < 0.001) and β-cell function (Disposition Index, DI; r = 0.36, p < 0.001). The tested indices yielded areas under the receiver operating curve (AROCs) from 0.70 to 0.79 and correlated with HCL (r = 0.40, 0.50 and 0.46, all p < 0.001), ISIcomp (r =-0.70, -0.62 and -0.53, all p < 0.001) and DI (r = 0.56, 0.47 and 0.48, all p < 0.001). Even after adjustments for age, sex and HCL, the indices remained independently associated with insulin sensitivity (β=-0.33, -0.19 and -0.04, all p < 0.001) and β cell function (β= 0.23, 0.13 and 0.03, all p < 0.001).