Untargeted salivary metabolomics for pediatric chronic kidney disease diagnosis

Chronic kidney disease (CKD) results in severe morbidity and mortality in children; thus early diagnosis is essential to minimize long-term effects. This research aimed to unveil potential saliva biomarkers for improving early CKD diagnosis, as a non-invasive alternative to blood. An untargeted metabolomics method based on LC-QTOF-MS was applied to unstimulated and acid-stimulated saliva samples from children and adolescents suffering from CKD and healthy subjects. Partial Least Squares Method-Discriminant Analysis (PLS-DA) revealed different metabolic profiles in unstimulated and acid-stimulated saliva samples. The features significantly altered were identified based on their MS and MS/MS spectra and the analytical standards. Prolylproline, leucyl-leucine and pyroglutamyl-leucine were found to be increased in unstimulated saliva of CKD patients, while creatine, n-butyrylcarnitine and palmitic acid were altered in stimulated saliva. Hierarchical cluster analysis showed better grouping with biomarkers altered in stimulated saliva samples. According to Receiver Operating Characteristic (ROC) curves generated using PLS-DA discriminant method, these biomarkers (AUC = 0.999) also provided better prediction accuracy than those in unstimulated samples (AUC = 0.874), demonstrating the relevance of the collection method in the metabolic profile of saliva. Hence, as long as the collection method is considered, the salivary metabolome was proven to be a tool of choice for pediatric CKD diagnosis.