Résumé : OBJECTIVES. Evidence in favour of switching between selective serotonin reuptake inhibitor (SSRI) and tricyclic (TCA) antidepressants in treatment resistant depression has been tested in a few studies only, consequently a prospective study was undertaken to evaluate the impact of switching strategies. METHODS. One hundred eighty-nine patients who failed to respond to a previous antidepressant were randomised to four arms: firstly they received citalopram or desipramine for a 4-week period; secondly, those who failed to respond were treated for a further 4-week period with the same antidepressant (citalopram-citalopram and desipramine-desipramine arms) or switched to the alternate one (citalopram-desipramine and desipramine-citalopram arms). RESULTS. There was no difference in the first 4-week phase between patients receiving citalopram versus desipramine in Hamilton Rating Scale for Depression (HRSD), Montgomery-Asberg Depression Rating Scale (MADRS), and Clinical Global Impression (CGI) scores. In the second 4-week phase remitter rates were higher among non-switched patients (P = 0.04). Moreover, considering HRSD and MADRS, switched patients reported significantly higher scores (P ≤ 0.02 for both scales at each time-point). CONCLUSIONS. This study supports the thesis that switching from an SSRI to a TCA (and vice versa) in non-responders to a 4-week trial of an SSRI/TCA is not associated with improved response. The result goes in the opposite direction to that predicted by current guidelines.