Résumé : The management of treatment-resistant depression is a much debated issue. In particular, the evidence supporting the commonly suggested sequential use of antidepressants from 2 different pharmacological classes is weak. This retrospective study was undertaken to investigate whether there is a better response in nonresponders switched to a different class of antidepressants (across-class) compared with nonresponders switched to an antidepressant from the same class (within-class).