Obesity and type 2 diabetes mellitus (T2DM) frequently coexist and share overlapping metabolic disturbances. Identifying integrated biochemical and clinical markers may support improved disease classification. A prospective study published in Diabetes, Metabolic Syndrome and Obesity: Targets and Therapy developed and evaluated a multivariable diagnostic model incorporating serum metabolites and clinical parameters.

The analysis included 139 individuals: 81 with obesity, 25 with T2DM, and 33 healthy controls. Serum amino acids and fatty acids were quantified using liquid chromatography-tandem mass spectrometry (LC-MS/MS), alongside anthropometric and laboratory assessments. Data were divided into training and validation sets in a 7:3 ratio. Among 54 variables that showed significant group differences (p < 0.05), 44 were associated with diagnosis in univariable logistic regression. Feature reduction using least absolute shrinkage and selection operator (LASSO), recursive feature elimination (RFE), and random forest (RF) converged on seven variables: alanine (Ala), histidine (His), glutamine (Gln), interleukin-10 (IL-10), age, fasting blood glucose (FBG), and aryl hydrocarbon receptor (AHR).

A support vector machine (SVM) classifier built on these features achieved an area under the curve (AUC) of 0.998 in the training cohort and 0.958 in validation. Individuals with obesity or T2DM had elevated Ala and reduced Gln, His, and IL-10 levels (p < 0.01). Glutamine and histidine concentrations correlated positively with IL-10 (r = 0.46 and 0.48; p < 0.001). The model establishes a 7-marker panel for disease classification and indicates that specific amino acids and IL-10 may participate in shared metabolic pathways relevant to obesity and T2DM.