Résumé : Background: It has been proposed that a history of suicide attempts could be a correlate of severe depressive disorder and that suicide attempters (SA) could represent a particular subtype of subjects suffering from major depressive disorder. We investigated clinical and demographic characteristics associated with SA and tested the hypothesis that a history of suicide attempts predicts poor response to antidepressants. Methods: One-hundred-and-forty-one SA and 670 non-SA subjects with major depressive disorder (MDD) were treated for twelve weeks with escitalopram or nortriptyline in GENDEP, a part-randomized multi-center clinical and pharmacogenetic study. Baseline characteristics were compared using linear and logistic regression. Linear mixed models were used to analyse continuous outcomes during the twelve weeks of follow-up. Results: At baseline, SA subjects suffered from more severe depression (mean Montgomery-Asberg Depression Rating Scale: 30.29 (7.61) vs 28.43 (6.54), p = 0.0002), reported higher level of suicidal ideation (1.21 (0.82) vs 0.73 (0.48), p < 0.0001), had a younger age of onset and experienced more depressive episodes, had higher harm avoidance scores and poorer socio-demographic environment than non-SA individuals. However, during the twelve weeks of treatment and after adjustment for baseline severity of depression there was no difference in treatment response between SA and non-SA. Limitations: Due to its retrospective design, it is possible that more severely depressed subjects might report more suicide attempts than less depressed individuals. Conclusions: While SA differed from non-SA in several clinical and demographic characteristics, the antidepressants were similarly effective in SA as in comparably severely depressed subjects without a history of suicide attempts. © 2009 Elsevier B.V. All rights reserved.