Salivary metabolite signatures of children with and without dental caries lesions

A metabolomic approach was used to analyze endogenous metabolites and to correlate with a specific biological state. The analysis of salivary metabolites is a growing area of investigation with potential for basic and clinical applications. Analyses of children's saliva in different dentitions and with or without caries could potentially reveal a specific profile related to oral disease risk. Nuclear Magnetic Resonance (NMR) is well suited for mixture analysis followed by Principal Component Analysis combined with Linear Regression (PCA-LR) statistics and was used to identify differences in the salivary metabolites. The classificatory analysis was performed using PCA-LR based on 1,000 cross-validation bootstrap runs from both classifiers in order to increase the data information from a small sample size. The PCA-LR presented a statistically good classificatory performance for children with and without caries with an accuracy of 90.11 % (P < 0.001), 89.61 % sensitivity (P < 0.001), and 90.82 % specificity (P < 0.001). Children with caries lesions presented higher levels of several metabolites, including lactate, fatty acid, acetate and n-butyrate. Saliva from subjects with different dentition stages was also analyzed. Although the salivary samples were poorly classified, permanent dentition presented increased levels of acetate, saccharides and propionate. The NMR data and PCA-LR were able to classify saliva from children with or without caries, with performance indexes comparable to the partial least-squares regression discriminant analysis (PLS-DA) results also performed. Our data also showed similar salivary metabolite profiles for healthy subjects despite the differences in their oral hygiene habits, socioeconomic status and food intake.