Regarding the direct costs and the social value of depression, the decision of an antidepressant treatment prescription must be optimized as much as possible. The development of a personalized medicine in psychiatry may reduce treatment failure, intolerance or resistance, and hence burden and costs of affective disorders.

There is hope that biomarkers will be found to guide treatment selection. It might be of decisive interest to be able to assess an individual's metabolism activity. We propose here to explore the relationship between the activity of drug-metabolizing enzymes (DME) and transporters- assessed by a phenotypic approach and the efficacy of antidepressants. We will focus on venlafaxine (V) that provides a reasonable second-step choice for patients with depression and is used extensively in psychiatric practice, and the metabolism of which involves several cytochromes (CYP) P450 enzymes and the transporter P-gp.

Thus, the primary objective of this study is to study the correlation between the concentration of V and its metabolite ODesmethylV (V+ODV) and drug metabolism variability assessed by a phenotypic approach, in patients with major depressive disorder and MADRS ≥ 20 despite 4 weeks of V at 150mg or less

Inclusion Criteria:

• Patient (Hospitalized or outpatient) with major depressive disorder and MADRS ≥ 20 at visit of selection
• Patients non responders to V after 4 weeks of V at 150mg or less
• Decision of the psychiatrist to increase the dose of V at visit of selection
• Understanding of French language and able to give a written inform consent.
• Informed consent signed to participate to the study
• Individuals covered by social security regimen

Exclusion Criteria:

• Patients treated by more than one antidepressant
• Patients currently treated with one of the drug substrate of the cocktail
• Sensitivity or contra-indication to any of the substrate drugs used
• Current pregnancy, desire to get pregnant, or breastfeeding
• Bipolar disorder and schizophrenia