Therapeutic drug monitoring studies of selective serotonin reuptake inhibitor (SSRI) antidepressants thus far failed to identify a clear concentration-response relationship in major depression. Majority of the previous studies defined clinical response as 50% or greater reduction from baseline in depression rating scale scores. Because many patients who meet these criteria still present symptoms associated with functional impairment, there is a need to consider "remission" as an alternative end point in concentration-response analyses of SSRIs. The present 12-week prospective study investigated the relationship between fluvoxamine (an SSRI) plasma concentration and remission in outpatients with depression. We used a flexible dose titration study designed to mimic clinical practice within the therapeutic dose range of fluvoxamine (25-200 mg/d). Receiver operating characteristics (ROC) curve was computed to determine the optimal fluvoxamine plasma concentration for remission using 269 concentration data obtained from 80 patients. Analysis of the ROC curve from the entire study sample did not reveal a fluvoxamine concentration significantly predicting remission. By contrast, ROC analysis specifically in patients with moderate to severe depression (N = 51; baseline 17-item Hamilton Rating Scale for Depression score ≥ 20) found a fluvoxamine concentration of 61.4 ng/mL as a significant predictor of remission. In conclusion, therapeutic drug monitoring may be useful for rational titration and individualization of fluvoxamine dose and predicting remission in patients with moderate to severe depression, who may presumably display lesser placebo component in pharmacodynamic response.