Table 10.

Summary of recent and representative studies aiming to distinguish individuals with ASD from TD individuals using multivariate analysis of potential excretory (urinary and fecal) metabolite biomarkers. Reported sample sizes are the numbers used for classification and do not necessarily reflect the study’s total sample size.

Reference	Study Participants	Experimental Methods	Key Features	Multivariate Technique	Key Results
Nadal-Desbarats et al. (2014)237	30 children with ASD and 28 TD children	Measured urinary metabolite profiles combined from two nuclear magnetic spectroscopy techniques	Minimum number of metabolites combined from both techniques	Partial least squares discriminant analysis	Achieved a prediction accuracy of 83% with 0.92 AUROC
Diémé et al. (2015)238	30 children with ASD and 32 TD children	Evaluated urine metabolite levels using nuclear magnetic spectroscopy and mass spectrometry techniques	46 metabolites combined across techniques	Partial least squares discriminant analysis	Predicted a 16-sample validation set with 0.91 AUROC, 100% sensitivity, and 75% specificity
Gevi et al. (2016)239	30 children with ASD and 30 TD children	Quantified urinary metabolite concentrations through liquid chromatography and mass spectrometry	25 urinary metabolites	Partial least squares discriminant analysis	Classified individuals with 0.89 AUROC
Kang et al. (2018)240	21 children with ASD and 23 TD children	Assess metabolite profiles and microbial compositions in participants’ fecal samples	Five fecal metabolites	Discriminant analysis	With leave-one-out cross-validation, obtained 78% sensitivity and 81% specificity