The lack of a therapeutic effect or unsupportable side-effects can lead to substitution of one antidepressant by another. The present study investigated potential modifications to the pharmacokinetic profile of milnacipran at steady-state when it is substituted for fluoxetine without any washout period. The open-label, multiple dose, three-period study was carried out in 12 evaluable healthy volunteers. A reference period (period 1) comprising a 3.5-day treatment with milnacipran at 50 mg b.i.d. was followed, after a 5–10-day washout, by 3 weeks of administration of 20 mg fluoxetine once daily (period 2), immediately followed by a further 3.5 days of administration of milnacipran at 50 mg b.i.d. (period 3). Blood samples collected at each period were analysed for milnacipran, N-dealkyl milnacipran, fluoxetine and norfluoxetine. Potential drug–drug interactions were evaluated by comparing milnacipran pharmacokinetic parameters between periods 1 and 3. A steady-state of fluoxetine and its metabolite was effectively reached by the end of the 3-week period. A steady-state of milnacipran was reached on day 2 of both periods 1 and 3. Trough concentrations of milnacipran were 66 and 65 ng/ml before and after the fluoxetine administration period, respectively. Cmax values were 226 and 248 ng/ml. When comparing the kinetic parameters of milnacipran before and after fluoxetine treatment, all the 90% confidence intervals were in the 20% range. No significant difference in the adverse events of milnacipran was observed before or after fluoxetine administration.