Accumulating evidence supports the hypothesis that antidepressant treatment prevents suicide. Most of this evidence consists of studies demonstrating decreases in suicide in parallel with increased use of antidepressants in different populations. Since ecological studies cannot establish causality, this individual-based study, analyzing temporal patterns in the toxicology of all suicides in Sweden 1992-2005, aimed at testing the hypothesis that the substantial decrease in suicide during the period was caused by the increased use of antidepressants.

During the eleven years, 16,937 suicides, and 33,426 controls (natural or accidental deaths) were investigated by toxicological screening. The annual numbers of detections of antidepressants, zolpidem, zopiclone, tramadole, and dextropropoxyphene were analysed in suicides and controls.

The annual number of suicides decreased by 332 cases (18.6 %). The expected number of antidepressant positive suicides, as determined only by their increased use in controls, was 741. If the decrease in suicide was caused by antidepressants, the expected number of suicides with antidepressants in toxicology should be reduced with 332 cases i.e. 409. If the decrease in suicide was not caused by antidepressants the expected number of positive cases should be reduced by 18.6 %, i.e. 603. The observed number was 420, indicating causality. The other drugs were detected as expected if they had not caused the decrease in suicide.

This controlled individual-based naturalistic study indicates that the increased use of antidepressants was the cause of the decrease in suicide previously demonstrated in ecological studies. Definite conclusions cannot be drawn, however.