Abstract
BackgroundExcess visceral obesity and ectopic organ fat is associated with increased risk of cardiometabolic disease. However, circulating markers for early detection of ectopic fat, particularly pancreas and liver, are lacking.MethodsLipid storage in pancreas, liver, abdominal subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) from 68 healthy or pre-diabetic Caucasian and Chinese women enroled in the TOFI_Asia study was assessed by magnetic resonance imaging/spectroscopy (MRI/S). Plasma metabolites were measured with untargeted liquid chromatography–mass spectroscopy (LC–MS). Multivariate partial least squares (PLS) regression identified metabolites predictive of VAT/SAT and ectopic fat; univariate linear regression adjusting for potential covariates identified individual metabolites associated with VAT/SAT and ectopic fat; linear regression adjusted for ethnicity identified clinical and anthropometric correlates for each fat depot.ResultsPLS identified 56, 64 and 31 metabolites which jointly predicted pancreatic fat (R2Y = 0.81, Q2 = 0.69), liver fat (RY2 = 0.8, Q2 = 0.66) and VAT/SAT ((R2Y = 0.7, Q2 = 0.62)) respectively. Among the PLS-identified metabolites, none of them remained significantly associated with pancreatic fat after adjusting for all covariates. Dihydrosphingomyelin (dhSM(d36:0)), 3 phosphatidylethanolamines, 5 diacylglycerols (DG) and 40 triacylglycerols (TG) were associated with liver fat independent of covariates. Three DGs and 12 TGs were associated with VAT/SAT independent of covariates. Notably, comparison with clinical correlates showed better predictivity of ectopic fat by these PLS-identified plasma metabolite markers.ConclusionsUntargeted metabolomics identified candidate markers of visceral and ectopic fat that improved fat level prediction over clinical markers. Several plasma metabolites were associated with level of liver fat and VAT/SAT ratio independent of age, total and visceral adiposity, whereas pancreatic fat deposition was only associated with increased sulfolithocholic acid independent of adiposity-related parameters, but not age.
Highlights
ObjectivesThe goals of this study were firstly to identify candidate metabolite markers that predict pancreas, liver and visceral adipose tissue (VAT)/subcutaneous adipose tissue (SAT) ratio, and elucidate the associated metabolic changes; and secondly to compare the predictive performance of these metabolite markers with a range of clinical measurements associated with each fat depot identified from the present cohort
Obesity has been long recognised as a risk factor for cardiometabolic disease and subsequent complications [1], individuals within each body mass index (BMI)
After metabolite selection by methods with Unbiased Variable selection in R (MUVR) and removal of redundant features, 56 (91% identified), 64 (95% identified), and 31 (100% identified) variables were associated with pancreatic fat, liver fat, and visceral adipose tissue (VAT)/subcutaneous adipose tissue (SAT) respectively
Summary
The goals of this study were firstly to identify candidate metabolite markers that predict pancreas, liver and VAT/SAT ratio, and elucidate the associated metabolic changes; and secondly to compare the predictive performance of these metabolite markers with a range of clinical measurements associated with each fat depot identified from the present cohort
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call