Relationships between plasma concentrations of trazodone and m-chlorophenylpiperazine (m-CPP) and the clinical effects were studied in 26 patients (12 males and 14 females) with major depression during three weeks' treatment of 150 mg/d trazodone using an open-study design. Depressive symptoms were evaluated by Montgomery Asberg Depression Rating Scale (MADRS), and subjective side effects were assessed by UKU side effects rating scale (UKU) before treatment and at weekly intervals. Plasma concentrations of trazodone and m-CPP were measured by HPLC. There were significant linear relationships between the steady-state plasma concentration (Css) of trazodone and both the final MADRS score (rs = -0.529, P < 0.01) and the percent improvement at 3 weeks (r = 0.442, P < 0.05). Moreover, the proportion of responders (a final MADRS score of 10 or less) was significantly higher in the group with a trazodone concentration greater than 714 ng/mL (6/8 vs 3/18, P = 0.008). No significant correlation was found between UKU score and the Css for either compound nor between the UKU score and the ratio of m-CPP/trazodone. The current study suggests that a therapeutic response is dependent on the plasma concentration of trazodone but not m-CPP and that a plasma trazodone concentration of about 700 ng/mL may be a threshold for a good therapeutic response.