TABLE 5.

Metabolomics biomarkers for drug toxicity.

Study	Sample/biofluids	Cohorts (size n)	Metabolome fraction	Analytical apparatus	Statistic methods	Metabolite biomarkers
Zhao et al. (2017)	Rat liver and serum	Rats (zero, low, high PZA dose) x (male, female), 10 each group	Total metobolome	1H NMR spectroscopy	PCA, OSC-PLS-DA, STT, non-parametric MWT, FDR	Increased LDL/VLDL, lactate, decreased BCAAs, glucose, taurine (high dose vs. zero dose)
Cao et al. (2018)	urine	TB pre-dose (25), TB-DILI (11, post-dose), non-DILI (49, post-dose)	total metobolome	UPLC-MS/MS	PCA, OPLS-DA, ANOVA test	pyroglutamate, isocitrate, citrate, xanthine; urate, cis-4-octenedioic acid, cis-aconitate, hypoxanthine
Deng et al. (2022)	Liver and serum bile acids, liver tissues	4 groups of mice (n = 8 each), treated with INH and/or RIF for 21 days	metabolome	LC-MS/MS	PLS-DA, one-way analysis of variance, Pearson’s test	CA, DCA, LCA, TDCA, TUDCA were potential biomarkers for early detection of RIF-induced liver injury

ANOVA = analysis of variance, BCAA = branched chains amino acids, CA = cholic acid, DCA = deoxycholic acid, FDR = hochberg and benjamini false discovery rate, INH = isoniazid, LCA = lithocholic acid, MWT = Mann–Whitney test, OPLS-DA = orthogonal PLS-DA, OSC-PLS-DA = orthogonal signal correction PLS-DA, PCA = principal component analysis, PLS-DA = partial least-squares discriminant analysis, PZA = pyrazinamide, (V) LDL = (very) low-density lipoprotein, STT = Student’s t-test, TDCA = taurodeoxycholic acid, TUDCA = tauroursodeoxycholic acid.