1216-P: Plasma Amino Acid Profile as a Predictive Biomarker for Visceral Adiposity in Japanese Americans

Background/Aim: Greater visceral fat area (VFA) is associated with multiple metabolic conditions. We sought to determine whether plasma amino acid concentrations (AA) predict VFA. Methods: In plasma samples from Japanese American participants in a study of diabetes risk factors, we used mass spectrometry to measure fasting levels of 20 AA in 351 individuals who were free from diabetes (mean age 56 y, 51% male). We measured height, weight, waist circumference (WC), and VFA and subcutaneous fat area (SFA). Using AA that were significantly associated with VFA in univariate analyses, we fitted a multivariable linear regression model to create a VFA prediction index. We calculated area under receiver-operating characteristic curves (AUROC) for ability of AA to discriminate between low vs. high VFA dichotomized at median values (men - 99.3 cm2, women - 65.6 cm2) and compared our AA estimates to WC AUROC. Results: All AA significantly discriminated between high and low VFA as all AUROC significantly exceeded the null value of 0.5 (Table 1). However, WC AUROC was significantly greater than all AA. Our new AA-based VFA prediction index performed better at distinguishing between high and low VFA than a previously reported index (Yamakado M et al., Clin Obesity, 2012, AUROC 0.780 vs. 0.676, p-value