Citation

Neutel CI, Patten SB. Ann. Epidemiol. 1997; 7(8): 568-574.

Affiliation

Health Canada, Ottawa, Ontario, Canada.

Copyright

(Copyright © 1997, American College of Epidemiology, Publisher Elsevier Publishing)

DOI

unavailable

PMID

9408553

Abstract

PURPOSE: Recent publications imply the existence of associations between psychotropic drugs use and risk of suicide. Some studies have measured the tissue level of these drugs in suicide deaths, while others compared toxicity indices, defined as number of suicide deaths per million prescriptions for individual antidepressants. Few of these studies used unexposed controls. The objective of this cohort study was to evaluate suicide attempts in subjects recently exposed to benzodiazepines and/or antidepressants, as compared to unexposed controls. METHODS: A population of 225,796 persons with prescriptions for benzodiazepines were selected from the Saskatchewan Health Data Bases. Controls consisted of 97,862 individuals who did not receive benzodiazepines. RESULTS: Stratifying the populations into antidepressant users and non-antidepressant users indicated that nonantidepressant users had statistically significant associations between suicide attempts and benzodiazepine use (odds ratio (OR) = 6.2), antipsychotic use, (OR = 2.6), and a history of past treatment for drug/alcohol abuse (OR = 13.4). Antidepressant users showed a statistically significant relation only with past treatment for drug/alcohol use (OR = 5.8). It is argued that the large OR for antidepressant use is due to confounding by indication. If so, the concept of toxicity index is misleading and should not be used. CONCLUSIONS: The association between benzodiazepine use and attempted suicide is especially high for nonantidepressant users, for the young, and for males. Whether this relationship is causal or not, physicians should be aware of the high potential for suicide attempts when prescribing benzodiazepines for patients in these high risk groups.

Language: en