Objective: To assess the impact of prior medication use and comorbidities on treatment initiation with duloxetine vs. venlafaxine XR for patients with major depression disorder (MDD) using retrospective claims data. Methods: Using the PharMetrics Database, we studied commercially insured individuals aged 18-64 who initiated treatment with duloxetine or venlafaxine XR between July 2005 and July 2006, and had =1 prior diagnosis with MDD and continuous enrollment during 12 months prior to initiation date. Initiation was defined as the first use of a medication preceded by three months without a prescription of the same medication. Chi-square and Logistic regression analysis of patients' demographics, past-year medication use and comorbidities assessed predictors of initiations with duloxetine vs. venlafaxine XR. Results: A total of 12,662 patients (73.8% female) initiated treatment with duloxetine, and 14,801 (72.1% female) with venlafaxine XR. Compared to venlafaxine XR patients, significantly more duloxetine patients received =3 unique antidepressants (39.5% vs. 25.2%), =3 unique pain medications (25.8% vs. 15.1%), SSRIs (55.5% vs. 41.2%), TCAs (12.5% vs. 7.5%), analgesics (63.6% vs. 51.6%), anticonvulsants (31.3% vs. 19.2%), or hypnotics (31.5% vs. 22.0%), and had =8 unique co-morbid medical conditions (38.8% vs. 29.5%) and diagnoses with pain (76.4% vs. 67.1%) (all p-values <0.001). Regression results revealed that the significant factors for duloxetine initiation vs. venlafaxine XR were prior use of =3 unique antidepressants (OR=1.34), =3 unique pain medications (OR=1.24), SSRIs (OR=1.51), TCAs (OR=1.19), analgesics (OR=1.12), anticonvulsants (OR=1.45), hypnotics (OR=1.25), and prior medical comorbidities of pain (OR=1.11) (all p-values <0.001). Conclusion: The results suggest that duloxetine patients with MDD are more likely to have more medical conditions and complex prior medication treatments than venlafaxine XR patients.