The potential for cytochrome P450 (CYP) 2D6 substrates to interact with desvenlafaxine (administered as desvenlafaxine succinate) and paroxetine was evaluated. In an open-label, crossover study, 20 healthy volunteers (aged 21-50) were randomized to 2 series of 9 days each of desvenlafaxine (100 mg/d) or paroxetine (20 mg/d), separated by a 5-day washout. The CYP2D6 substrate desipramine (50 mg) was administered alone on day 1 and coadministered on day 6 of dosing with either desvenlafaxine or paroxetine. CYP2D6 genotype was determined at baseline. Based on least squares geometric mean ratios between reference (desipramine alone) and test treatments, desvenlafaxine produced minor increases in desipramine area under the plasma concentration versus time curve (AUC; 36%) and peak plasma concentration (C(max); 30%) (vs paroxetine: 419%, 90%, respectively; both P < .001). Desvenlafaxine produced little change in 2-hydroxydesipramine AUC (16% increase) and C(max) (0%) versus paroxetine (18% and 82% decreases, respectively; P = .008, P < .001, respectively), indicating that desvenlafaxine, especially at the recommended therapeutic dose of 50 mg/d for major depressive disorder in the United States, has little potential to interact with CYP2D6 substrates.