Factors associated with suicide attempts in 648 patients with bipolar disorder in the Stanley Foundation Bipolar Network

Abstract

Background: Clinical factors related to suicide and suicide attempts have been studied much more extensively in unipolar depression compared with bipolar disorder. We investigated demographic and course-of-illness variables to better understand the incidence and potential clinical correlates of serious suicide attempts in 648 outpatients with bipolar disorder.

Method: Patients with bipolar I or II disorder (DSM-IV criteria) diagnosed with structured interviews were evaluated using self-rated and clinician-rated questionnaires to assess incidence and correlates of serious suicide attempts prior to study entry. Clinician prospective ratings of illness severity were compared for patients with and without a history of suicide attempt.

Results: The 34% of patients with a history of suicide attempts, compared with those without such a history, had a greater positive family history of drug abuse and suicide (or attempts); a greater personal history of early traumatic stressors and more stressors both at illness onset and for the most recent episode; more hospitalizations for depression; a course of increasing severity of mania; more Axis I, II, and III comorbidities; and more time ill on prospective follow-up. In a hierarchical logistic regression, a history of sexual abuse, lack of confidant prior to illness onset, more prior hospitalizations for depression, suicidal thoughts when depressed, and cluster B personality disorder remained significantly associated with a serious suicide attempt.

Conclusion: Our retrospective findings, supplemented by prospective follow-up, indicate that a history of suicide attempts is associated with a more difficult course of bipolar disorder and the occurrence of more psychosocial stressors at many different time domains. Greater attention to recognizing those at highest risk for suicide attempts and therapeutic efforts aimed at some of the correlates identified here could have an impact on bipolar illness-related morbidity and mortality.