In the studies reported here, the ability of atomoxetine hydrochloride (Strattera) to inhibit or induce the metabolic capabilities of selected human isoforms of cytochrome P450 was evaluated. Initially, the potential of atomoxetine and its two metabolites, N-desmethylatomoxetine and 4-hydroxyatomoxetine, to inhibit the metabolism of probe substrates for CYP1A2, CYP2C9, CYP2D6, and CYP3A was evaluated in human hepatic microsomes. Although little inhibition of CYP1A2 and CYP2C9 activity was observed, inhibition was predicted for CYP3A (56% predicted inhibition) and CYP2D6 (60% predicted inhibition) at concentrations representative of high therapeutic doses of atomoxetine. The ability of atomoxetine to induce the catalytic activities of CYP1A2 and CYP3A in human hepatocytes was also evaluated; however, atomoxetine did not induce either isoenzyme. Based on the potential of interaction from the in vitro experiments, drug interaction studies in healthy subjects were conducted using probe substrates for CYP2D6 (desipramine) in CYP2D6 extensive metabolizer subjects and CYP3A (midazolam) in CYP2D6 poor metabolizer subjects. Single-dose pharmacokinetic parameters of desipramine (single dose of 50 mg) were not altered when coadministered with atomoxetine (40 or 60 mg b.i.d. for 13 days). Only modest changes (approximately 16%) were observed in the plasma pharmacokinetics of midazolam (single dose of 5 mg) when coadministered with atomoxetine (60 mg b.i.d. for 12 days). Although at high therapeutic doses of atomoxetine inhibition of CYP2D6 and CYP3A was predicted, definitive in vivo studies clearly indicate that atomoxetine administration with substrates of CYP2D6 and CYP3A does not result in clinically significant drug interactions.