Background Suicidal thoughts and behaviors could be associated to a poor response to antidepressant treatment, but the exclusion of suicidal patients from randomized clinical trials restricts the available knowledge. In this study, we aimed at defining more precisely the response to antidepressants among suicidal patients and the threshold of suicidality that best predicts a poor response. Method We investigated the short‐term response to a new antidepressant treatment of 4,041 depressed outpatients depending on their suicidal status (passive or active suicidal ideation (SI), history of suicide attempts [SAs]), either self‐rated or clinician‐rated. Depression outcomes, measured with the Hospital Anxiety and Depression Scale, and remission rates were compared depending on suicidal status at baseline using logistic regression models. Results Using either a qualitative or a quantitative approach to measure SI, we found that suicidal patients were less likely to improve or attain remission, but not more likely to worsen, than nonsuicidal patients. In the multivariate analyses, SI (odds ratio [OR] = 1.40; 95% confidence interval [CI]: 1.18–1.65) and a history of SA (OR = 1.39; 95% CI: 1.16–1.66) were the best predictors of nonremission, independently of the class of antidepressant treatment. Conclusion Antidepressant treatment seems to be less effective among those patients that need it most. Clinical trials including suicidal patients are needed to investigate specific treatment options.