OBJECTIVES Physician decisions regarding initial anti-depressant dosing levels may impact treatment outcomes. This study examined initial dosing levels and their associations with adherence among currently available selective serotonin and serotonin-norepinephrine reuptake inhibitors (SSRIs and SNRIs). METHODS A total of 115,284 patients initiating treatment on an SSRI or SNRI during 2005 were selected for analysis from a large managed-care claims database. Patients were assigned to a drug cohort on the basis of their most recent prescription for one of the studied medications. The study included only patients with one or more ICD-9 coded diagnoses for depression or anxiety within +/- 3 months of initiation of treatment with their index drug, an initial 30 days supply of the index medication, and continuous eligibility of at least 6 months prior and 12 months following treatment initiation. Subtherapeutic dosing levels of each studied medication were determined on the basis of label information. Outcome measures included early discontinuation (no second prescription following initiation) and days on therapy with the index medication. All reported differences were significant at the <0.01 level. RESULTS A total of 13.6% of all study patients initiated antidepressant treatment at a subtherapeutic dose. The share of patients initiating treatment at subtherapeutic levels varied substantially across medications: paroxetine, most frequently (30.1%) and escitalopram, least frequently (2.7%). Across all studied medications, subtherapeutic initial dosing was associated with higher levels of early discontinuation (29.9%) and fewer average days on therapy (164.1) than initial dosing at therapeutic doses or higher (24.3% and 176.8 days, respectively). These associations were directionally consistent across all studied medications. CONCLUSIONS Initial dosing of antidepressants at subtherapeutic levels was associated with more frequent early discontinuation and fewer average days on treatment. Prescribers should be cognizant of the potential impact of initial dosing decisions on subsequent adherence to therapy.