Le trouble de personnalité borderline (TPB) se définit notamment par l’instabilité des relations interpersonnelles et de l’image de soi, ainsi que l’impulsivité des sujets qui en souffrent. Certains auteurs soutiennent que l’hospitalisation de ces patients ne réduit pas leurs comportements suicidaires et pourrait leur être délétère (régression ou renforcement des comportements auto-agressifs). Pour déterminer si l’orientation hospitalière des patients suicidants à haut risque de TPB au décours des urgences influence les récidives des comportements suicidaires à 6 mois, nous avons utilisé les données de l’étude prospective multicentrique FRENCH CRISIS. Le risque de présence d’un TPB a été évalué par un autoquestionnaire de dépistage, le PDQ-4+. Nous avons mesuré la récurrence des comportements suicidaires à 6 mois dans le groupe des patients hospitalisés et celui des non-hospitalisés. Trois cent vingt sujets ont été inclus et répartis entre un groupe ayant un risque élevé d’avoir au moins un TPB ( n = 197), un groupe ayant un risque élevé d’avoir au moins un TP non TPB ( n = 84) et un groupe ayant un faible risque d’avoir un TP ( n = 39). L’hospitalisation au décours des urgences n’est pas associée à une différence de récidive des comportements suicidaires à 6 mois entre les groupes. Borderline personality disorder (BPD) is characterized by a pervasive pattern of instability and impulsivity. There is a high prevalence of BPD patients among those admitted to the emergency department for suicide attempts. However, little empirical research exists to assist clinicians in deciding whether to hospitalize a suicidal patient. Some authors have argued that hospitalization does not prevent suicide and could actually harm these patients, thereby leading to psychosocial regression. Parasuicidal behaviors could be reinforced by the attention given during hospitalization. Our purpose was to determine whether the hospitalization of suicidal patients who have a high risk of BPD after discharge from the emergency department is associated with a recurrence of suicidal behavior at 6 months. We designed a prospective study, acquiring patients from three emergency hospitals. The participants were suicidal subjects admitted for voluntary drug intoxication and were 18 years of age or older. The participants completed the Personality Disorder Questionnaire (PDQ-4+) to assess BPD symptomatology. Information on the recurrence of suicidal behavior at 6 months was obtained by interview of patients and the review of the charts from the 3 hospitals involved in the study. Other assessments included the BDI-13 (severity of depression), the Hopelessness Scale (hopelessness), the TAS-20 (alexythymia), the AUDIT (alcohol disorder) and the MINI (axis I disorders). A total of 606 subjects admitted for a suicide attempt participated in this study. A total of 320 (52.8 %) of the subjects completed the PDQ-4+. The sample was divided into three groups: participants at high risk of having at least one BPD ( n = 197), a group at high risk of having at least one non-BPD PD ( n = 84) and a group with low risk of having a PD ( n = 39). Hospitalization following an emergency was not associated with a recurrence of suicide attempts at 6 months among patients at high risk of BPD. A logistical regression analysis showed pre-hospitalization antidepressant prescription to be associated with recidivism (OR = 2.1, P = .037). Our exploratory study suggests that hospitalization may not increase suicide attempts among patients with BPD when the health organization does not include a specific device such as DBT.