Obesity is a significant risk factor for cardiometabolic diseases, but individuals within the same BMI category exhibit heterogeneity in their cardiometabolic profiles. East Asians, for instance, develop type 2 diabetes (T2D) at lower BMIs and younger ages compared to Caucasians, suggesting that factors beyond BMI contribute to cardiometabolic risk. One such factor is the deposition of visceral and ectopic organ fat, particularly in the liver and pancreas, which is implicated in insulin resistance and dyslipidemia. Current assessment of ectopic fat relies on expensive and/or invasive methods, highlighting the need for circulating biomarkers for early detection. Previous metabolomic studies have identified potential markers for non-alcoholic fatty liver disease (NAFLD), but biomarkers for early asymptomatic liver fat deposition and pancreatic fat remain largely unknown. This study aimed to identify plasma metabolite markers that predict fat deposition in the pancreas, liver, and the visceral-to-subcutaneous adipose tissue (VAT/SAT) ratio, using MRI/MRS, and compare their predictive performance with clinical measurements in a cohort of Caucasian and Chinese women from the TOFI_Asia study. The VAT/SAT ratio was included as it's an indicator of body fat distribution, with VAT considered more detrimental than subcutaneous adipose tissue (SAT).

The existing literature strongly links visceral and ectopic fat accumulation to increased risk of cardiometabolic diseases such as type 2 diabetes and cardiovascular disease, independent of BMI. However, the lack of readily available, non-invasive methods for measuring ectopic fat deposition necessitates the search for reliable circulating biomarkers. While some studies have explored metabolomic profiles in relation to non-alcoholic fatty liver disease (NAFLD), identifying specific markers for early, asymptomatic ectopic fat in both liver and pancreas remains an unmet need. The study acknowledges the scarcity of research on circulating biomarkers for pancreatic fat deposition, with previous studies yielding inconclusive results. The importance of identifying these biomarkers has been emphasized by recent position statements from leading organizations. This current study aims to address this gap.

This cross-sectional study utilized data from the TOFI_Asia study, focusing on 68 female participants (34 Chinese, 34 Caucasian) aged 20-70 years with BMIs between 20-45 kg/m². Participants were excluded if they experienced significant weight changes in the preceding three months, had undergone bariatric surgery, were on glucose-related medications, or had a history of relevant diseases. Fasting venous blood samples were collected for metabolomic analysis, and body composition was assessed using dual-energy X-ray absorptiometry (DXA) and MRI/MRS. Anthropometric and clinical measurements (height, weight, waist and hip circumferences, blood pressure, fasting plasma glucose, HbA1c, liver function tests, lipid profile, and glucoregulatory peptides) were also recorded. MRI/MRS was used to quantify fat content in the abdomen (VAT and SAT), pancreas, and liver. Untargeted metabolomics using LC-MS was performed on plasma samples. Data preprocessing included normalization, feature filtering, and metabolite annotation. Multivariate partial least squares (PLS) regression with unbiased variable selection (MUVR) was used to identify metabolites predicting pancreatic fat, liver fat, and VAT/SAT ratio. The performance of PLS models using selected metabolites was compared with those using clinical markers and a combined set of markers. Univariate linear regression was used to assess the association of individual metabolites with fat depots, adjusting for covariates such as ethnicity, total adiposity, VAT/SAT ratio, and age.

Multivariate PLS analysis identified 56, 64, and 31 metabolites associated with pancreatic fat, liver fat, and VAT/SAT ratio, respectively. Variable selection significantly improved model fit and predictive power. For pancreatic fat, sulfolithocholic acid remained significantly associated after adjusting for various covariates, though no metabolite remained significantly associated after adjusting for age. Liver fat was associated with several metabolites, including dihydrosphingomyelin (dhSM (d36:0)), phosphatidylethanolamines, diacylglycerols (DGs), and triacylglycerols (TGs), with many of these associations remaining significant after adjusting for covariates. Several DGs and TGs were significantly associated with VAT/SAT ratio after adjusting for covariates. Notably, the metabolite markers generally provided better prediction of ectopic fat levels than clinical markers alone, while combining both types of markers did not significantly improve the model beyond using only the metabolite markers.

This study successfully identified several plasma metabolites as potential markers for visceral and ectopic fat deposition, particularly in the liver and VAT/SAT ratio. The superior predictive performance of the metabolite markers compared to clinical markers highlights their potential value in early detection of ectopic fat accumulation. The finding that only sulfolithocholic acid remained consistently associated with pancreatic fat after multiple adjustments suggests a specific metabolic pathway related to this fat depot. The identification of multiple lipid species (DGs and TGs) associated with liver fat and VAT/SAT suggests a complex interplay of lipid metabolism in these fat depots. This study contributes significantly to the understanding of the metabolic pathways associated with ectopic fat accumulation and provides potential candidates for non-invasive screening and early intervention strategies. Future studies should investigate the identified metabolites in larger, longitudinal studies to validate their predictive ability and potential causal roles.

This study demonstrates the utility of untargeted metabolomics in identifying novel plasma metabolite markers for visceral and ectopic fat. Several metabolites showed strong associations with liver fat and VAT/SAT ratio, while sulfolithocholic acid was linked to pancreatic fat. These findings suggest potential for developing non-invasive screening tools for early detection of ectopic fat accumulation. Further research is needed to validate these findings in larger, more diverse populations and to elucidate the underlying mechanisms.

The cross-sectional nature of the study limits causal inferences. The relatively small sample size and the focus on women may affect the generalizability of the findings. Further investigation is needed to replicate these findings in larger, more diverse cohorts and to explore the potential for these markers in different populations and genders.