Metabolic signatures and risk of type 2 diabetes in a Chinese population: an untargeted metabolomics study using both LC-MS and GC-MS

Abstract

Aims/hypothesis: Metabolomics has provided new insight into diabetes risk assessment. In this study we characterised the human serum metabolic profiles of participants in the Singapore Chinese Health Study cohort to identify metabolic signatures associated with an increased risk of type 2 diabetes. Methods: In this nested case–control study, baseline serum metabolite profiles were measured using LC-MS and GC-MS during a 6-year follow-up of 197 individuals with type 2 diabetes but without a history of cardiovascular disease or cancer before diabetes diagnosis, and 197 healthy controls matched by age, sex and date of blood collection. Results: A total of 51 differential metabolites were identified between cases and controls. Of these, 35 were significantly associated with diabetes risk in the multivariate analysis after false discovery rate adjustment, such as increased branched-chain amino acids (leucine, isoleucine and valine), non-esterified fatty acids (palmitic acid, stearic acid, oleic acid and linoleic acid) and lysophosphatidylinositol (LPI) species (16:1, 18:1, 18:2, 20:3, 20:4 and 22:6). A combination of six metabolites including proline, glycerol, aminomalonic acid, LPI (16:1), 3-carboxy-4-methyl-5-propyl-2-furanpropionic acid and urea showed the potential to predict type 2 diabetes in at-risk individuals with high baseline HbA1c levels (≥6.5% [47.5 mmol/mol]) with an AUC of 0.935. Combined lysophosphatidylglycerol (LPG) (12:0) and LPI (16:1) also showed the potential to predict type 2 diabetes in individuals with normal baseline HbA1c levels (<6.5% [47.5 mmol/mol]; AUC = 0.781). Conclusions/interpretation: Our findings show that branched-chain amino acids and NEFA are potent predictors of diabetes development in Chinese adults. Our results also indicate the potential of lysophospholipids for predicting diabetes.