Faecal metabolomics reveals amino acid, nucleotide and sphingolipid metabolic remodelling in paediatric urolithiasis

Abstract Paediatric urolithiasis often recurs, yet routine clinical measures capture only part of its metabolic heterogeneity. We therefore profiled faecal metabolic alterations and evaluated candidate multi-metabolite markers in children with paediatric urolithiasis. Faecal samples from 30 children with calcium-containing mixed upper urinary tract stones and 30 healthy controls underwent LC-MS-based untargeted metabolomics. Differential features were identified for exploratory analysis. PCA, OPLS-DA, KEGG over-representation, pathway topology, GSEA and network analyses were used to characterise metabolic patterns. A fixed six-metabolite L2-regularised logistic regression model was evaluated using nested leave-one-out cross-validation. The comparison retained 35,349 features and 2,932 annotated compounds, including 7,143 differential features and 623 differential compounds. OPLS-DA separated the paediatric urolithiasis and healthy-control groups. Nominal KEGG analyses implicated beta-alanine, nucleotide and glycine/serine/threonine metabolism. GSEA yielded nominal positive enrichment for caffeine metabolism, ABC transporters, sphingolipid metabolism and signalling, cAMP signalling and galactose metabolism. A 351-node database network connected amino-acid, nucleotide, beta-alanine and sphingolipid-related modules. The fixed six-metabolite model yielded a cross-validated AUC of 0.946. Faecal metabolomics revealed coordinated, but heterogeneous, candidate metabolic differences in paediatric urolithiasis. The strongest exploratory signals involved amino-acid derivatives, nucleotides, beta-alanine and sphingolipid-related metabolism.