The effect of ionic strength, assay constituents, alpha-naphthoflavone (aNF), terfenadine and testosterone on human CYP3A mediated midazolam (MDZ) 1'-hydroxylation (MDZ 1'-OH) and 4-hydroxylation (MDZ 4-OH) in vitro was examined. Increasing concentration of Tris-HCl (Tris) and sodium phosphate (PO4) buffers differentially affected MDZ 1'-OH and MDZ 4-OH formation rates and had a different effect on MDZ metabolism mediated by microsomes containing CYP3A4 versus CYP3A4 and CYP3A5. MDZ metabolism was not affected by PO4 buffer concentration when cumene hydroperoxide (CUOOH) was used as the source of reactive oxygen. Interestingly, the ammonium ion present in the solution of glucose 6-phosphate dehydrogenase was found to inhibit MDZ metabolism. The addition of MgCl2 up to 50 mM and CaCl2 (5-30 mM) had no affect or inhibited MDZ metabolism, respectively. Formation of MDZ 1'-OH by microsomes from adult and fetal liver and expressed CYP3A4 was regioselectively stimulated by aNF (10 microM). In human hepatocytes, aNF stimulated MDZ 1'-OH formation (up to 100%). Terfenadine (20 microM) regioselectively stimulated MDZ 1'-OH formation in Tris (1-200 mM) and PO4 (1-10 mM) buffers by up to 159%. Surprisingly, with expressed CYP3A4, terfenadine (20 microM) inhibited MDZ 1'-OH formation. Terfenadine (20 microM) had little effect on MDZ 1'-OH formation by fetal liver microsomes. Testosterone (10 and 100 microM) regioselectively stimulated (up to 269%) MDZ 4-OH formation by adult liver microsomes and expressed CYP3A4. Testosterone (100 microM) inhibited (> 40%) MDZ 1'-OH and MDZ 4-OH formation by fetal liver microsomes. With adult liver microsomes, aNF and terfenadine had little effect on the Km for MDZ 1'-OH formation. However, the Km for MDZ 4-OH formation was decreased (up to 94%) by 100 microM testosterone. In the presence of CUOOH, no stimulation of MDZ metabolism was observed by aNF, terfenadine or testosterone in adult liver microsomes. These studies indicate that because assay conditions can substantially alter the catalytic activity of CYP3A, caution should be exerted when extrapolating results between in vitro and in vivo, and when results from different laboratories are compared. Further, these results suggest that the stimulation of CYP3A4 may also occur in vivo and, consequently, may have clinical importance.