Amino Acid Profiles Distinguish Glycemic Phenotypes in Childhood and Predict 10-Year Risk of Dysglycemia in Young Adults: The BCAMS Study

Amino acid (AA) dysregulation has been associated with metabolic diseases. However, its potential as a childhood determinant of adult diabetes risk remains understudied, particularly among children with normoglycemia. Leveraging a 10-year follow-up cohort, we investigated whether AA profiles distinguished childhood glycemic phenotypes and predicted future diabetes risk. Among 947 children from the Beijing Children and Adolescents Metabolic Syndrome (BCAMS) study, participants were categorized as having normoglycemia with insulin sensitivity (NG-IS), normoglycemia with insulin resistance (NG-IR), or impaired fasting glucose (IFG), and plasma AA profiles were assessed by targeted metabolomics. At follow-up, 297 participants with baseline normoglycemia underwent a 2-h oral glucose tolerance test to assess incident dysglycemia, and five adipokines were measured. After adjusting for age, sex, and relevant covariates, distinct AA signatures were identified cross-sectionally across glycemic groups, with a core cluster (Ile, Val, Tyr, Phe, Ala, Pro, and branched-chain AA score) independently associated with IR and IFG. Prospectively, children in the NG-IR group had a 4.3-fold higher risk of incident dysglycemia than those in the NG-IS group. Six AAs (Ile, Leu, Tyr, Phe, Lys, and Hyl) significantly predicted 10-year dysglycemia risk, with 22.6–42.1% of their associations estimated through follow-up IR, with the largest proportion observed for Ile. In separate Ile–IR models, estimated proportions through follow-up leptin and high-molecular-weight adiponectin were 72% and 38%, respectively. Furthermore, dietary Ile restriction in obese mice reduced body weight by 25% and leptin by 50% and improved IS by 22%. These findings identify childhood plasma AA dysregulation as an early biomarker for diabetes risk, and Ile emerges as a potentially modifiable target. Article Highlights Prospective evidence linking childhood amino acid (AA) dysregulation with future diabetes risk remains limited. We examined whether AA profiles distinguished childhood glycemic phenotypes and predicted dysglycemia risk in the Beijing Children and Adolescents Metabolic Syndrome (BCAMS) cohort. AA alterations were already detectable in children with normoglycemia and insulin resistance; six AAs were associated with 10-year dysglycemia risk, and their addition to clinical variables improved prediction. Ile showed the largest estimated indirect association with incident dysglycemia through follow-up insulin resistance; dietary Ile restriction in obese mice produced concordant metabolic changes, supporting AA profiling for early risk identification and Ile as a candidate modifiable target.