As treatment-resistant depression (TRD) is linked to suicidal behaviors and suicidal risk is a predictor of TRD, depression with current suicidal ideation (SI) may lead to TRD. Early identification of TRD risk factors in patients with depression and current SI is crucial. The aims of our study were: i) to identify risk factors for depression non-remission and TRD in patients with depression and current SI; ii) to assess if SI at baseline mediated the relationship between depression severity at baseline and depression remission at week 6. We analyzed data from two large, prospective, naturalistic French cohorts of adult outpatients with depression (DSM-IV criteria) followed for 6 weeks after starting or changing antidepressants (LUEUR and GENESE). Sociodemographic and clinical characteristics, along with early symptom improvement, were compared between patients with and without current SI using logistic regression models (univariate and multivariate). Patients with antidepressant change or initiation were analyzed separately. Those without depression remission at week 6 after an antidepressant change were considered TRD cases. In patients with antidepressant change, the major predictor of non-remission was poorer early improvement (at week 2) of anxiety. For patients with treatment initiation, SI at baseline mediated the relation between depression severity at baseline and depression remission. Depression severity at baseline alone did not explain depression remission. Clinicians should systematically target with specific pharmacological and non-pharmacological treatments anxiety and SI and assess their changes in the short term to increase the chance of depression remission in depressed patients with current SI.