It is important to predict CYP3A enzyme induction in the drug-discovery process to avoid adverse effects in clinical. In the present study, we constructed a method to correct the variability of in vitro CYP3A induction assays and thereby a method for the prediction of CYP3A induction in the clinical setting. Induction assays were performed in vitro using HepaRG cells and seven typical inducers. An index value was determined for enzyme induction, termed the relative factor (RF), from the ratio of the concentration of the inducers to the reference standards. Using RF as an index, variation among the assays was reduced. A good relationship was obtained between the ratio of the free plasma concentration at steady-state (C(ss,u)) to RF (expressed as C(ss,u)/RF) and the in vivo induction response. Using rifampicin as a reference standard, compounds with a C(ss,u)/RF value greater than 7.31 nmol l(-1) may induce CYP3A in vivo in humans.