Untargeted metabolomics for plasma biomarker discovery for early chronic kidney disease diagnosis in pediatric patients using LC-QTOF-MS

Abstract

Pediatric chronic kidney disease (CKD) is a clinical syndrome characterized by renal hypofunction occurring due to gradual and irreversible kidney damage that can further progress over time. New biomarkers may help early diagnosis of pediatric patients suffering from CKD and improve the outcome. Untargeted metabolomics based on LC-QTOF-MS has been used to find new biomarkers for the early diagnosis of CKD in plasma from pediatric patients. In order to avoid any bias in the determination of statistically significant entities as a consequence of the data analysis method followed, two different chemometric approaches have been used, Mass Profiler Professional (MPP) software and Matlab R2015a software. Metabolic fingerprints of control and CKD pediatric patients were compared and five metabolites which showed a significant change common to both data analysis procedures were identified. Sphingosine-1-phosphate, n-butyrylcarnitine, cis-4-decenoylcarnitine and an unidentified feature with 126.0930 m/z were found to be increased in plasma from pediatric patients with CKD, whereas bilirubin was significantly decreased. A partial least squares discriminant analysis model built with these 5 entities classified correctly 96% of the samples. In addition, when considering only early CKD patients against controls, a performance of 97% was obtained. Thus, these promising metabolites could be suitable biomarkers for the early diagnosis of pediatric CKD in a clinical setting.