Background: Diabetic ketoacidosis (DKA) is a serious complication of type 1 diabetes (T1D), arising from relative insulin deficiency and leading to hyperglycemia, ketonemia, and metabolic acidosis. Early detection and treatment are essential to prevent severe outcomes. This pediatric case-control study utilized plasma metabolomics to explore metabolic alterations associated with DKA and to identify predictive metabolite patterns.
Methods: We examined 34 T1D participants, including 17 patients admitted with severe DKA and 17 age- and sex-matched individuals in insulin-controlled states. A total of 215 plasma metabolites were analyzed using proton nuclear magnetic resonance and direct-injection liquid chromatography/mass spectrometry. Multivariate statistical methods, machine learning techniques, and bioinformatics were employed for data analysis.
Results: After adjusting for multiple comparisons, 65 metabolites were found to differ significantly between the groups (28 increased and 37 decreased). Metabolomics profiling demonstrated 100% accuracy in differentiating severe DKA from insulin-controlled states. Random forest analysis indicated that classification accuracy was primarily influenced by changes in ketone bodies, acylcarnitines, and phosphatidylcholines. Additionally, groups of metabolites (ranging in number from 8 to 18) correlated with key clinical and biochemical variables, including pH, bicarbonate, glucose, HbA1c, and Glasgow Coma Scale scores.
Conclusions: These findings underscore significant metabolic disturbances in severe DKA and their associations with critical clinical indicators. Future investigations should explore if metabolic alterations in severe DKA can identify patients at increased risk of complications and/or guide future therapeutic interventions.
Keywords: Diabetic ketoacidosis; Metabolites; Metabolomics; Panels; Pathways; Pediatric.
© 2024. The Author(s).