Ketoconazole is extensively used as an index inhibitor of cytochrome P450-3A (CYP3A) activity in vitro and in vivo, but the mechanism of ketoconazole inhibition of CYP3A still is not clearly established. Inhibition of metabolite formation by ketoconazole (seven concentrations from 0.01 to 1.0 µm) was studied in human liver microsomes (n = 4) at six to seven substrate concentrations for triazolam, midazolam, and testosterone, and at two substrate concentrations for nifedipine. Analysis of multiple data points per liver sample based on a mixed competitive-noncompetitive model yielded mean inhibition constant K(i) values in the range of 0.011 to 0.045 µm. Ketoconazole IC50 increased at higher substrate concentrations, thereby excluding pure noncompetitive inhibition. For triazolam, testosterone, and midazolam α-hydroxylation, mean values of α (indicating the 'mix' of competitive and noncompetitive inhibition) ranged from 2.1 to 6.3. However, inhibition of midazolam 4-hydroxylation was consistent with a competitive process. Determination of K(i) and α based on the relation between 50% inhibitory concentration values and substrate concentration yielded similar values. Pre-incubation of ketoconazole with microsomes before addition of substrate did not enhance inhibition, whereas inhibition by troleandomycin was significantly enhanced by pre-incubation. Ketoconazole inhibition of triazolam α- and 4-hydroxylation, midazolam α-hydroxylation, testosterone 6β-hydroxylation, and nifedipine oxidation appeared to be a mixed competitive-noncompetitive process, with the noncompetitive component being dominant but not exclusive. Quantitative estimates of K(i) were in the low nanomolar range for all four substrates. © 2011 The Authors. JPP © 2011 Royal Pharmaceutical Society.