Suicide in primary affective disorders revisited: A systematic review by treatment era

Background: We reviewed suicide rates in affective disorder and their variation with electroconvulsive therapy (ECT) and antidepressant availability. Method. Suicide rates were calculated from 75 follow-up studies, identified by systematic literature searches and analyzed for differences over time eras characterized by the availability of specific treatments. Data Sources and Study Selection: MEDLINE, EMBASE, BIOSIS Previews, and Psychological Abstracts literature searches were conducted for the years 1966 to 1995, References from review articles identified from these sources from 1985 onward and textbook references were also included. Publications prior to 1966 were obtained from article references identified for the period 1966 to 1976 and reviews. Inclusion criteria were (1) articles written in English, French, or German; (2) sample size > 30; (3) age at recruitment between 18 and 64 years for each subject; (4) sample had to contain subjects hospitalized at time of recruitment; and (5) naturalistic follow-up of at least 6 months. Results: Suicide rates decreased with longer follow-up periods. For follow-up periods over 20 years, the mean rate was 3.76/1000 person-years (95% confidence interval [CI] = 2.35 to 5.17). Suicides accounted for 12.3% (95% CI = 8.52 to 16.04) of all deaths in samples in which 40% or more of patients had died. For studies with minimal overlap between eras, the mean suicide rate differed significantly between eras (pretreatment, before 1940: 6.3/1000; ECT treatment, 1940 to 1959: 5.7/1000; antidepressant treatment, 1960 onward: 3.3/1000; F = 31.4, df = 2,42; p < .001). Conclusion: The risk of suicide in follow-up studies of affective disorder has decreased compared to that reported in previous reviews. The availability of ECT and antidepressants may have contributed to this decrease, but prescription of these treatments cannot be assumed for all patients.