OBJECTIVES: The aim of this study is to assess the health system and patient costs associated with antidepressant (AD) use considering the presence and persistence of depression and anxiety. METHODS: The data was retained from a population-based health survey on 2004 community dwelling older adults aged ≥ 65 years participating in the ESA (Étude sur la Santé des Aînés) study. Depression and anxiety were assessed using DSM-IV criteria and measured at 2 time points one year apart. Medical and non-medical costs were considered. Medication and health service use and costs were identified from provincial administrative databases.  The excess costs associated with AD use as a function of mental health status was analysed using generalized linear models with a gamma distribution (log link), controlling for a number of factors. RESULTS: The prevalence of antidepressant use reached 15.5%: SSRIs followed by TCAs were the most common.  Significantly higher health care costs (Δ: $2840, Wald χ2=60.00, df=1, p<0.0001) were associated with AD use.  Among antidepressant users, the results did not show any differences in costs when accounting for dosage, the number of episodes of use and the presence of antidepressant switches. Among persistent cases of depression and anxiety the use of AD was associated with lower adjusted total costs reaching CDN $2724 and CDN $2114, respectively. The use of AD can result in cost savings reaching $154.6 million and $118.4 million per 1 000 000 population, for persistent cases of depression and anxiety. CONCLUSIONS:  This study showed important cost savings associated with AD use in persistent cases of depression and anxiety. Future studies should focus on further exploring potential cost savings associated with different classes of AD in the treatment of different clinical profiles of depression and anxiety and this in the older adult population.