Cytochrome P450 (CYP) inhibition and induction are the key mechanisms in drug-drug interactions. Aside from clinical studies, primary human hepatocytes may represent the most appropriate experimental system for the evaluation of CYP induction in humans. A consensus of an international panel on the present status and future research directions in the application of primary human hepatocytes in the evaluation of CYP-induction is presented here. The following observations are concluded to be generally true: (1) Human hepatocytes isolated from both biopsy samples and transplantable livers are suitable for induction studies. (2) Hormonally-defined media can be used for the evaluation of CYP induction. (3) Isozyme-selective induction of CYP1A and 3A by known inducers are observed. (4) Reproducibility of induction could be improved by using hepatocytes plated as confluent cultures. (5) Induction could be observed for hepatocytes treated at 1-3 days after culturing. (6) Treatment duration of 2 days in general leads to near maximal induction. (7) In general, there is a good qualitative correlation between human hepatocyte results in vitro and clinical observations in vivo. (8) When the same inducers were evaluated in independent laboratories, similar data were generally observed. We conclude that primary human hepatocytes represent an appropriate model for mechanistic evaluation of CYP induction and as a screening tool for CYP induction potential of xenobiotics. A set of data acceptance criteria are proposed: (1) Positive response should be observed with concurrent positive control chemicals; (2) reproducible observation should be observed with multiple human donors; (3) for negative response, the doses used should not be cytotoxic; and (4) replicate treatment and/or multiple dose treatment should be performed to allow statistical analysis. Future studies should include the further development of on: (1) The inducibility of CYP isozymes other than CYP1A and 3A, and phase II enzymes; (2) further development of culturing condition to allow optimal gene expression; (3) evaluation of the involvement of nonparenchymal cells on CYP induction of parenchymal cells; (4) the and validation of quantitative approaches to extrapolate in vitro data to in vivo data; (5) evaluation of possible individual variations and potential genetic polymorphism in inducibility; (6) further definition of species differences in CYP induction; (7) development of a 'normal' human hepatocyte cell line for CYP induction studies; (8) improvement of cryopreservation procedure of human hepatocytes; (9) definition of the molecular mechanisms of CYP induction; and (10) evaluation of the induction of phase II metabolic pathways.