OBJECTIVES: Antidepressants can cause undesirable drug-drug interactions when taken concomitantly with certain medications. Elderly patients may be particularly vulnerable to antidepressant interactions due of slower metabolism and utilization of multiple pharmacotherapies. The objective of this study was to determine rates of potential drug-drug interactions involving antidepressants in a geriatric population and their impact on prescription adherence. METHODS: Data were from the MarketScan® Medicare Database, a claims database from retirees with employer-sponsored Medicare supplemental insurance. Subjects were age >65 years, new antidepressant users, and had a depression diagnosis between 7/1/2001-12/31/2006. Potential drug-drug interactions involving at least one antidepressant and another drug with overlapping days supplied were identified over the one year following antidepressant initiation using MicroMedex DRUG-REAX® software. Multinomial logistic regression and bivariate statistics were used to evaluate the association between potential interactions and whether patients discontinued, refilled or switched their first antidepressant prescription. RESULTS: Among the 39,512 patients who met the inclusion criteria, 25.4% had potential contraindicated or major interactions, 36.1% had moderate interactions, and 38.5% had minor or no interactions. Compared to the moderate/minor/no interactions groups, the contraindicated/major group had a greater prevalence of medical comorbidities and higher comorbidity indices (p<0.001). Amitriptyline hydrochloride was involved in 19.1% of the potential contraindicated/major interactions. Tramadol hydrochloride and oxycodone, opioid analgesics, were the most common medications with contraindicated/major interactions. Presence of contraindicated/major and moderate interactions was associated with an increase probability of switching to a different antidepressant of 23 and 11 percentage points, respectively (p<.001) and decreased probability of discontinuing of 3.5 and 2.5 percentage points (p<0.001), after controlling for age, gender, pre-period mental disorders and medical comorbidities, antidepressant treatment type, and use of other psychotropic medications. CONCLUSIONS: Elderly antidepressant users frequently use medications with the potential for interactions with their antidepressant medication. There is a need for antidepressants with improved interaction profiles.