OBJECTIVE: We investigated whether the oral administration of a low dose (75 micro g) of midazolam, a CYP3A probe, can be used to measure the in vivo CYP3A activity. METHODS: Plasma concentrations of midazolam, 1'OH-midazolam and 4'OH-midazolam were measured after the oral administration of 7.5 mg and 75 micro g midazolam in 13 healthy subjects without medication, in four subjects pretreated for 2 days with ketoconazole (200 mg b.i.d.), a CYP3A inhibitor, and in four subjects pretreated for 4 days with rifampicin (450 mg q.d.), a CYP3A inducer. RESULTS: After oral administration of 75 micro g midazolam, the 30-min total (unconjugated + conjugated) 1'OH-midazolam/midazolam ratios measured in the groups without co-medication, with ketoconazole and with rifampicin were (mean+/-SD): 6.23+/-2.61, 0.79+/-0.39 and 56.1+/-12.4, respectively. No side effects were reported by the subjects taking this low dose of midazolam. Good correlations were observed between the 30-min total 1'OH-midazolam/midazolam ratio and midazolam clearance in the group without co-medication (r(2)=0.64, P<0.001) and in the three groups taken together (r(2)=0.91, P<0.0001). Good correlations were also observed between midazolam plasma levels and midazolam clearance, measured between 1.5 h and 4 h. CONCLUSION: A low oral dose of midazolam can be used to phenotype CYP3A, either by the determination of total 1'OH-midazolam/midazolam ratios at 30 min or by the determination of midazolam plasma levels between 1.5 h and 4 h after its administration.