The development of new antidepressant drugs has raised many questions concerning their benefits, risks and costs relative to the older established compounds, the tricyclic antidepressants. Nevertheless, it must be borne in mind that all of the newer compounds [atypical tricyclics, selective serotonin reuptake inhibitors (SSRIs) and reversible inhibitors of monoamine oxidase] are no more effective than the tricyclic antidepressants (TCAs). The reasons for this limited efficacy are unclear, as are predictors of response. The upshot is the problem of managing the refractory depressive and many stratagems have been proposed. Adverse effects vary greatly from class to class and to some extent within class. The TCAs affect many systems reflecting their multiple binding capabilities; the SSRIs have side effects stemming from their actions on central 5-HT mechanisms. By and large, SSRIs are somewhat better tolerated than TCAs, but whether this translates into better compliance is not known. Rare putative side effects, such as suicidal ideation, have received undue attention but further studies are justified. The pharmacokinetics of antidepressants are typically complex and routine monitoring of plasma drug concentrations is not helpful. The cost of even the most expensive antidepressants is but,a fraction of the direct costs of unsuccessful treatment, let alone the indirect costs.