Many antidepressants have been introduced as racemic drugs, the enantiomers of which may differ in some of their pharmacodynamic and pharmacokinetic properties. This review argues that each enantiomer of a chiral antidepressant should be evaluated according to its individual characteristics rather than by extrapolation from the racemate, or by assumptions based on the stereoselective characteristics of other enantiomeric drugs. For example, in some cases the enantiomers' pharmacodynamic and therapeutic properties can be complementary, which suggests that the racemate should be used clinically. In other cases where enantiomers show qualitatively similar but quantitatively different properties to the racemate, using a single enantiomer might be more appropriate. In yet further cases, a distomer may induce the metabolism of the eutomer, enantiomers may be metabolised by different enzymes, there may be a different profile of drug-drug interactions, and therapeutic drug monitoring may be simpler. Therefore, this review exemplifies the principle that each enantiomer of a chiral antidepressant should be evaluated according to its individual pharmacological, pharmacokinetic and pharmacogenetic characteristics. These factors are discussed in relation to five chiral antidepressants: trimipramine, mianserin, mirtazapine, fluoxetine and citalopram. It is hoped that an appreciation of the stereoselective differences between enantiomers will facilitate improvements in the benefit:risk ratio of drugs used in the management of depression. Copyright 2001 John Wiley & Sons, Ltd.