Numerous groups have described the rat as an in vivo model for the assessment and prediction of drug-drug interactions (DDIs) in humans involving the inhibition of cytochrome P450 3A forms. Even for a well-established substrate-inhibitor pair like midazolam-ketoconazole, however, the magnitude of the DDI in rats (e.g. 1.5- to 5-fold) does not relate to what is observed clinically (e.g. 5- to 16-fold). Because nonlinear substrate pharmacokinetics (PK) may result in a weaker interaction, it was hypothesized that the lower magnitude of interaction observed in rats was due to the saturation of metabolic pathway(s) of midazolam at the doses used (10-20 mg/kg). Therefore, the inhibitory effects of ketoconazole were reevaluated at lower oral (1 and 5 mg/kg) and intravenous (IV) (1 mg/kg) doses of midazolam. In support of the hypothesis, oral exposure at 5 mg/kg dose of midazolam was 18-fold higher compared to that at 1 mg/kg. Furthermore, when the interaction was investigated at the lower midazolam dose (1 mg/kg), ketoconazole increased the IV and oral exposure of midazolam by 7-fold and 11-fold, respectively. A weaker DDI (1.5- to 1.8-fold) was observed at the higher oral midazolam dose. Collectively, these results suggest that the lower reported interaction in rats is likely due to saturation of midazolam clearance at the doses used. Therefore, when the rat is used as a DDI model to screen and differentiate compounds, or predict CYP3A inhibition in humans, it is important to use low doses of midazolam and ensure linear PK.