Risk Factors for Completed Suicide in Bipolar Disorder

Background: Because Chinese bipolar patients in Taiwan, unlike Western patients, exhibit low comorbidity of substance abuse disorders, this retrospective and controlled study of completed suicide in bipolar patients explored the risk period and other risk factors for such an outcome.

Method: All acute inpatients with bipolar I disorder (DSM-IV) were followed from date of admission after January 1, 1985, until December 31, 1996, in regard to their death. The patients were followed by record linkage to the Death Certification System in Taiwan, which was issued throughout 1996. Nineteen female and 24 male patients died as a result of suicide within this period. Forty-one of 43 of the total number of patients were matched with 1 living bipolar individual (as a control subject) for age, sex, and date of admission. Demographic data, family history, and clinical characteristics were collected from the patients’ medical records and were formally confirmed at every admission.

Results: The lifetime prevalence of alcohol/drug use disorders was 14.6% in suicide completers. Thirty suicide completers (69.8%) revealed duration of illness of at least 7 years at the time of death. The latency period from the presumed time of onset to completing suicide averaged 12.2 years. The mean age at the first suicide attempt was 31.1 years among 43 completers and 10 living controls who had ever attempted suicide. Conditional logistic regression revealed a strong association of suicide (p<.001) with the following factors: onset with mood-congruent psychotic feature (adjusted odds ratio [OR]=0.18, 95% confidence interval [CI]=0.04 to 0.74), positive first-degree family history of completed suicide (adjusted OR=15.08, 95% CI=1.39 to 163.50), and making a suicide attempt at least once in 7 years of illness (adjusted OR=4.96, 95% CI=1.03 to 23.83). There appeared to be no significant difference in fasting levels of serum cholesterol or blood sugar between the suicide completers and the living controls.

Conclusion: The first 7 to 12 years subsequent to onset of affective illness and age less than 35 years may be the high-risk periods for suicide in bipolar disorder. Those bipolar disorder patients who have a first-degree family history of suicide and who have more suicide attempts (at least once in 7 years of illness) are likely to commit suicide. Symptomatology (e.g., mood congruence of psychotic features) at the time of presumed disease onset may potentially differentiate subgroups of bipolar patients with various levels of suicide risk.