Repaglinide is presently recommended by the U.S. Food and Drug Administration as a clinical CYP2C8 probe, yet current in vitro and clinical data are inconsistent concerning the role of this enzyme in repaglinide elimination. The aim of the current study was to perform a comprehensive investigation of repaglinide metabolic pathways and assess their contribution to the overall clearance. Formation of four repaglinide metabolites was characterized using in vitro systems with differential complexity. Full kinetic profiles for the formation of M1, M2, M4, and repaglinide glucuronide were obtained in pooled cryopreserved human hepatocytes, human liver microsomes, human S9 fractions, and recombinant cytochrome P450 enzymes. Distinct differences in clearance ratios were observed between CYP3A4 and CYP2C8 for M1 and M4 formation, resulting in a 60-fold M1/M4 ratio in recombinant (r) CYP3A4, in contrast to 0.05 in rCYP2C8. Unbound K(m) values were within 2-fold for each metabolite across all in vitro systems investigated. A major system difference was seen in clearances for the formation of M2, which is suggested to be a main metabolite of repaglinide in vivo. An approximately 7-fold higher unbound intrinsic clearance was observed in hepatocytes and S9 fractions in comparison to microsomes; the involvement of aldehyde dehydrogenase in M2 formation was shown for the first time. This systematic analysis revealed a comparable in vitro contribution from CYP2C8 and CYP3A4 to the metabolism of repaglinide (<50%), whereas the contribution of glucuronidation ranged from 2 to 20%, depending on the in vitro system used. The repaglinide M4 metabolic pathway is proposed as a specific CYP2C8 probe for the assessment of drug-drug interactions.