Abstract Body (Do not enter title and authors here): Background: Elevated 1-hour glucose levels during a 75g oral glucose tolerance test (OGTT) are strong predictors of progression to type 2 diabetes (T2D) and chronic metabolic diseases. Cardiometabolic abnormalities and underlying pathophysiology related to elevated 1-hour glucose may be precisely captured by combined untargeted fasting blood metabolites.
Hypothesis: We investigated whether a fasting blood metabolomic risk score (MRS) for predicting 1-hour glucose could be a target for improving impaired glucose tolerance, β-cell function, and long-term glycemic trajectories, as well as for reducing the risks of T2D and coronary heart disease (CHD), through an integrated metabolomics analysis of data from dietary intervention trials and prospective cohort studies. We also analyzed genome-wide circulating microRNAs to determine whether microRNAs in plasma derived from key metabolic organs may serve as novel biomarkers regulating MRS changes.
Methods: Untargeted blood metabolomics and a frequently sampled 75g OGTT were performed in the participants from the OmniCarb trial (n=162). In an independent weight-loss dietary intervention trial (POUNDS Lost), temporal changes in MRS and plasma miRNAs measured by genome-wide sequencing in response to weight-loss diet interventions were calculated. Associations of MRS at baseline and its 10-year changes with the long-term risk of incident T2D and CHD were prospectively investigated in participants from the Nurses’ Health Study (NHS).
Results: We developed a fasting blood MRS for predicting 1-hour glucose levels (Pearson r = 0.8) using a machine-learning elastic net regression model. MRS levels were significantly associated with half-day postprandial glucose excursions and insulin secretion after 5-week feeding interventions varying in carbohydrate amount and glycemic index. In the POUNDS Lost, diet-induced changes in MRSs were related to 2-year trajectories of glucose metabolism; circulating microRNAs regulating cardiometabolic abnormalities were pivotal factors influencing these changes. In the NHS, women in the top 20% of MRS had a multivariate-adjusted relative risk of 3.80 (95% CI: 2.22, 6.51) for T2D and 1.48 (1.04, 2.12) for CHD compared to those in the lowest 20%. Additionally, 10-year increases in MRS were linearly associated with higher T2D risk.
Conclusions: Our findings indicate that fasting blood MRS predicting elevated 1-hour glucose reflects disease pathophysiology and is a target for preventing T2D and CHD.