Measurement of liver fat by magnetic resonance imaging: relationships with body fat distribution, insulin sensitivity and plasma lipids in healthy men

Abstract

We compared the use of magnetic resonance imaging (MRI) as a test for liver fat content (LFAT) with proton magnetic resonance spectroscopy (MRS) and investigated its relationship with body fat distribution, insulin sensitivity, plasma lipids and lipoproteins. LFAT was quantified by MRI and MRS in 17 free-living, healthy men with a wide range of body mass indexes. Fasting adiponectin was measured by immunoassay and insulin resistance by homeostasis assessment (HOMA) score. Intraperitoneal, retroperitoneal, anterior subcutaneous and posterior subcutaneous abdominal adipose tissue masses (ATMs) were determined by MRI. Measurements of LFAT by MRI and MRS were highly correlated (r = 0.851, p < 0.001). In univariate regression analysis, LFAT by MRI was also significantly correlated with plasma triglycerides (TGs), insulin, HOMA score, carbohydrate intake and the masses of all abdominal adipose tissue compartments (p < 0.05). LFAT was inversely correlated with plasma adiponectin (r = -0.505, p < 0.05). In multivariate linear regression analysis including plasma adiponectin and age, intraperitoneal ATM was an independent predictor of LFAT (beta-coefficient = 0.587, p = 0.024). Moreover, intraperitoneal ATM was also an independent predictor of HOMA score after adjusting for LFAT, plasma adiponectin and age (beta-coefficient = 0.810, p = 0.010). Conversely, LFAT was a significant predictor of plasma TG concentration after adjusting for adiponectin, intraperitoneal ATM, HOMA and age (beta-coefficient = 0.751, p = 0.007). Similar findings applied with LFAT measured by MRS. These data suggest that MRI is as good as MRS to quantify liver fat content. Our data also suggest that liver fat content could link intraabdominal fat with insulin resistance and dyslipidaemia.