Depressive illness is a common, often unrecognised and untreated condition with substantial associated costs, particularly indirect costs (e.g. lost productivity and absenteeism). The improved tolerability profile of fluoxetine and associated lower discontinuation rates, the relative safety of the drug in overdosage and its similar efficacy compared with tricyclic antidepressants have provided the main rationale for using this agent in depressed patients. Pharmacoeconomic analyses of fluoxetine have mainly sought to determine whether its higher acquisition cost in comparison with tricyclic antidepressants can be offset by reductions in other costs and whether the use of this agent as first-line therapy can be justified. Studies have also attempted to determine whether the selective serotonin reuptake inhibitors (SSRIs) can be distinguished from one another on pharmacoeconomic grounds; overall efficacy and tolerability of these agents appear to be similar, although tolerability data are conflicting. Most analyses have been of a retrospective database or clinical decision analytic model design; two prospective trials (one conducted in a naturalistic setting) have been conducted. These studies have mainly considered direct treatment costs only from the perspective of the healthcare payer. Available evidence suggests that overall total direct healthcare costs for patients who start antidepressant therapy with fluoxetine are similar to, or lower than, those for patients who start therapy with tricyclic agents or other SSRIs. Offsetting of the higher acquisition cost of fluoxetine compared with that of tricyclic agents may be accounted for by lower in- and outpatient costs with fluoxetine, a possible lower risk of absenteeism from work and lower mean total medical costs associated with acute overdosage. Between-treatment differences in drug use patterns may also, in part, explain the observed differences in total healthcare costs between fluoxetine and other antidepressants. In particular, patients beginning therapy with fluoxetine are more likely to receive treatment regimens that meet minimum recommended guidelines for dosage and duration and are less likely to require treatment switching/augmentation than those recei-ving tricyclic antidepressants or other SSRIs as initial therapy. In addition, fewer fluoxetine than tricyclic antidepressant recipients discontinue therapy early, and fewer fluoxetine recipients require upward dosage titration or concomitant anxiolytic/hypnotic medications than patients receiving other SSRIs. In conclusion, fluoxetine is a well established antidepressant which possesses tolerability and safety advantages over the tricyclic agents. The available cost analyses show that these benefits can be obtained without additional overall cost to the healthcare provider. Cost advantages observed to date for fluoxetine over other SSRIs require confirmation.