Exploratory Metabolomic Characterization of HbA1c-Defined Glycemic Groups in the Qatar Biobank Cohort

Glycemic status is commonly defined using glycated hemoglobin (HbA1c), classifying individuals as normoglycemic, prediabetic, or diabetic. However, these categories may not fully capture the underlying metabolic complexity associated with early dysglycemia and disease progression. This study aimed to investigate metabolomic alterations across HbA1c-defined glycemic groups using a bioinformatics-guided approach. Targeted metabolomic profiling of EDTA plasma samples from 1752 Qatar Biobank participants was performed using ultra-performance liquid chromatography–tandem mass spectrometry (UPLC-MS/MS). Participants were classified as normoglycemic (NG), prediabetic (PD), or diabetic (DM) according to established HbA1c criteria. Metabolite data were integrated with clinical variables and analyzed using multivariate modeling, ROC analysis, and pathway enrichment to identify discriminatory metabolic features across glycemic states. Distinct metabolomic profiles were observed across the glycemic spectrum. The NG–PD comparison showed modest discrimination (AUC 0.64–0.69), with alterations primarily involving steroid-conjugated, lipid-related, and amino acid-associated metabolites, indicating early metabolic alterations beyond glucose metabolism. In contrast, the PD–DM comparison demonstrated stronger discrimination (AUC ~0.90), driven predominantly by carbohydrate-related metabolites, including meta_glucose, mannose, mannonate, fructose, and reduced 1,5-anhydroglucitol (1,5-AG), reflecting progressive glucose-centric metabolic dysregulation. Metabolomic profiling revealed distinct patterns of metabolic alterations across HbA1c-defined glycemic groups, with modest differences between normoglycemia and prediabetes involving steroid-, lipid-, and amino acid-related metabolites, and more pronounced carbohydrate-related differences in diabetes. These findings provide exploratory insights into the metabolic heterogeneity associated with early dysglycemia and warrant further investigation in independent and longitudinal cohorts.