Thirty-four studies (n=16,221) with 67 treatment arms or conditions (42 with and 25 without lithium treatment) were included. Three of the studies were double-blind randomised controlled trials. The study design of the other 31 studies was not indicated.
Overall. The estimated rate for completed suicides was 0.174 (95% CI: 0.138, 0.215) with lithium treatment versus 0.942 (95% CI: 0.743, 1.180) without; the RR was 5.43 (95% CI: 3.57, 8.25); the reduction was 81.5%. The estimated rates for suicide attempts were 0.312 (95% CI: 0.204, 0.451) and 4.646 (95% CI: 4.104, 5.239) with and without lithium treatment, respectively; the RR was 14.9 (95% CI: 8.41, 26.4); the reduction was 93.3%. The estimated rates for all suicidal acts (completed suicides and suicide attempts) were 0.210 (95% CI: 0.172, 0.253) and 3.100 (95% CI: 2.803, 3.422) with and without lithium treatment, respectively; the RR was 14.8 (95% CI: 8.54, 25.6); the reduction was 93.2%.
Bipolar disorder. The estimated rate for all suicidal acts was 0.295 (95% CI: 0.216, 0.394) with lithium treatment versus 6.102 (95% CI: 5.320, 6.966) without; the RR was 20.7 (95% CI: 11.9, 35.8); the reduction was 95.2%.
Major affective disorder. The estimated rate for all suicidal acts was 0.193 (95% CI: 0.151, 0.244) with lithium treatment versus 2.119 (95% CI: 1.829, 2.441) without; the RR was 11.0 (95% CI: 5.37, 22.4); the reduction was 90.9%.
Further results of comparisons of suicide rates found with and without lithium to general population rates were also presented, as were secular trends in suicide rates and effects of lithium treatment.