Abstract:Research highlights • Profound perturbation of metabolome preceding T2D onset was revealed in a prospective Chinese population-based cohort with a follow-up period of almost 16 years. • Insulin resistance rather than β-cell dysfunction was strongly associated with metabolic shifts before T2D. • The impact of metabolites on T2D risk could be mediated by insulin resistance. • Discovery of previously unknown independent associations between metabolites and incident T2D. • The inclusion of metabolites selected by machine learning significantly improved the predictive power of traditional clinical prediction models. Introduction Type 2 diabetes (T2D) is a heterogeneous metabolic disease with large variations in the relative contributions of insulin resistance and β-cell dysfunction across different glucose tolerance subgroups and ethnicities. A more precise yet feasible approach to categorize risk preceding T2D onset is urgently needed. This study aimed to identify potential metabolic biomarkers that could contribute to the development of T2D and investigate whether their impact on T2D is mediated through insulin resistance and β-cell dysfunction. Methods A non-targeted metabolomic analysis was performed in plasma samples of 196 incident T2D cases and 196 age- and sex-matched non-T2D controls recruited from a long-term prospective Chinese community-based cohort with a follow-up period of ∼16 years. Results Metabolic profiles revealed profound perturbation of metabolomes before T2D onset. Overall metabolic shifts were strongly associated with insulin resistance rather than β-cell dysfunction. In addition, 188 out of the 578 annotated metabolites were associated with insulin resistance. Bi-directional mediation analysis revealed putative causal relationships among the metabolites, insulin resistance and T2D risk. We built a machine-learning based prediction model, integrating the conventional clinical risk factors (age, BMI, TyG index and 2hG) and 10 metabolites (acetyl-tryptophan, kynurenine, γ-glutamyl-phenylalanine, DG(18:2/22:6), DG(38:7), LPI(18:2), LPC(P-16:0), LPC(P-18:1), LPC(P-20:0) and LPE(P-20:0)) (AUROC = 0.894, 5.6% improvement comparing to the conventional clinical risk model), that successfully predicts the development of T2D. Conclusions Our findings support the notion that the metabolic changes resulting from insulin resistance, rather than β-cell dysfunction, are the primary drivers of T2D in Chinese adults. Metabolomes as a valuable phenotype hold potential clinical utility in the prediction of T2D. Graphical abstract Download : Download high-res image (106KB) Download : Download full-size image

A novel model for predicting prolonged stay of patients with type-2 diabetes mellitus: a 13-year (2010–2022) multicenter retrospective case–control study

A Novel Risk Score for Type 2 Diabetes Containing Sleep Duration: A 7-Year Prospective Cohort Study among Chinese Participants.

A SuperLearner approach for predicting diabetic kidney disease upon the initial diagnosis of T2DM in hospital

Length of Stay Prediction for Clinical Treatment Process Using Temporal Similarity

Understanding variations and influencing factors on length of stay for T2DM patients based on a multilevel model

The Utility of the Adjusted-Optimize-hf Risk Model for Predicting In-Hospital Length of Stay in the Chinese Population.

Analyzing the factors that are involved in length of inpatient stay at the hospital for diabetes patients

Predicting three-month fasting blood glucose and glycated hemoglobin changes in patients with type 2 diabetes mellitus based on multiple machine learning algorithms

Development and assessment of novel machine learning models to predict medication non-adherence risks in type 2 diabetics

Predicting the Development of Type 2 Diabetes in a Large Australian Cohort Using Machine-Learning Techniques: Longitudinal Survey Study

Machine Learning Model for Predicting Risk Factors of Prolonged Length of Hospital Stay in Patients with Aortic Dissection: a Retrospective Clinical Study

Development and validation of a prediction model estimating the 10-year risk for type 2 diabetes in China

Predicting the risk of subclinical atherosclerosis based on interpretable machine models in a Chinese T2DM population

Predictive models of medication non-adherence risks of patients with T2D based on multiple machine learning algorithms

Clinical study of different prediction models in predicting diabetic nephropathy in patients with type 2 diabetes mellitus

Clinical nomogram prediction model to assess the risk of prolonged ICU length of stay in patients with diabetic ketoacidosis: a retrospective analysis based on the MIMIC-IV database

1233-P: Prediction of Type 2 Diabetes Occurrence Using Machine Learning Model

Comparing the accuracy of four machine learning models in predicting type 2 diabetes onset within the Chinese population: a retrospective study

Ten metabolites-based algorithm predicts the future development of type 2 diabetes in Chinese

Predicting the Risk of Incident Type 2 Diabetes Mellitus in Chinese Elderly Using Machine Learning Techniques

A patient network-based machine learning model for disease prediction: The case of type 2 diabetes mellitus