The biochemical variations induced in human primary hepatocyte cultures by reference activators of xenoreceptor CAR (NR1I3) and PXR (NR1I2), i.e., rifampicin, phenobarbital, and 6-(4-chlorophenyl)imidazo[2,1-b] [1,3]thiazole-5-carbaldehyde O-3,4-dichlorobenzyl) oxime (CITCO), were investigated using a global metabonomics approach. Cultured human hepatocytes were treated with the three drugs before analysis of intracellular and extracellular media by ultra performance liquid chromatography/time-of-flight-mass spectrometry (UPLC/TOF-MS) technique, in order to list endogenous compounds potentially related to a PXR or CAR induction mechanism and to identify drug metabolites related to each treatment. The emphasis was put on the quality of the analytical data (dilution/filtration strategy before data processing) and on the appropriate pattern recognition techniques. In cellular media, the most significant variations seen in the data are not related to the treatments but to the source of hepatocytes, illustrating the importance of the genetic and/or environmental background in human liver experiments. However when applying classical multivariate statistical approaches (principal component analysis (PCA) and orthogonal partial least squares (O-PLS)), the statistical weight due to drug metabolites, present only in the treated groups, hinders the interpretation because of their predominance compared to most of the changes seen in endogenous metabolites. A new statistical approach, called shared and unique structure (SUS) plot, enabling the comparison of different treatments having the same control has been applied, allowing separation of clearly exogenous variables (drug metabolites) from endogenous biomarkers. Endogenous variables (either up- or down-regulated) have been attributed specifically to the impact of rifampicin (PXR ligand), CITCO (CAR ligand), and phenobarbital (CAR and PXR activator) on the biological regulation pathways of the hepatocytes. This global approach coupled to a statistical pretreatment of the data, enabling the separate capture of both drug related and drug induced biomarkers, represents a powerful technique for future mechanistic studies using cellular tools.