Codeine is O-demethylated by cytochrome P450 2D6 (CYP2D6) to form the more potent drug morphine, accounting for much of codeine's analgesic and dependence-producing properties. Because morphine production can be decreased by inhibition of CYP2D6, the authors hypothesized that CYP2D6 inhibition could be used to treat codeine dependence. A randomized, double-blind, placebo-controlled trial was conducted. All patients received brief behavioral therapy. Two weeks of baseline monitoring were followed by 8 weeks of daily treatment with fluoxetine or quinidine (two potent CYP2D6 inhibitors) or placebo. Thirty patients were assessed (all white, age 40 + 12 years using 127 + 79 mg/day of codeine [mean + SD]), and 17 entered treatment. Eight patients remained in the study by treatment week 8. Quinidine > fluoxetine > placebo inhibited CYP2D6 as reflected in the change of the O-demethylation of dextromethorphan, a specific CYP2D6 probe. At treatment week 8, placebo, quinidine, and fluoxetine reduced mean daily codeine intake by 57%, 56%, and 51% of baseline intake respectively; there was no difference among treatment groups. In this small sample, CYP2D6 inhibitors did not appear to have a useful role in the treatment of codeine dependence.