The pharmacokinetics and safety of duloxetine were evaluated in a single-blind, placebo-controlled, escalating multiple-dose study in 12 healthy male subjects. In the treatment group (n = 8), duloxetine was administered orally at a starting dose of 20 mg twice daily (bid) and escalated at weekly intervals to 30 mg bid, then to 40 mg bid. The observed plasma concentration-time data at all three dose levels were adequately described by a one-compartment model with a first-order absorption rate constant. The mean oral clearance, apparent volume of distribution, and half-life values were 114 L/h (range: 44 to 218 L/h), 1943 L (range: 803 to 3531 L), and 12.5 h (range: 9.2 to 19.1 h), respectively. Somnolence, nausea, and dry mouth were observed following the initial dose, but they resolved with continuing drug administration. Duloxetine was not associated with clinically significant changes in blood pressure (BP) or heart rate (HR) measured in the standing position. However, in recumbent position, small increases in systolic (≤ 9 mmHg) and diastolic (≤ 5 mmHg) BP and small decreases in HR (≤ 6 beats/min) were observed. Abrupt discontinuation of duloxetine was associated with a small increase in mean HR (≤ 12beats/min). In 3 subjects, abrupt discontinuation was also associated with transient sleep disturbance. No clinically important changes in electrocardiograms, cardiac intervals, clinical laboratory tests, and neurological functions were observed. These results indicate that duloxetine exhibits linear pharmacokinetics with respect to dose and duration of treatment and that a multiple oral dose regimen starting at 20 mg bid and gradually escalating up to 40 mg bid was generally well tolerated.