Suicides by Violent Means in Individuals Taking SSRIs and Other Antidepressants A Postmortem Study in Sweden, 1992–2004 Seena Fazel, MRPsych, MD,*y Martin Grann, PhD, y Johan Ahlner, MD, PhD, z and Guy Goodwin, FRCPsych, DPhil* Abstract: A number of reports have linked consumption of selective serotonin reuptake inhibitors (SSRIs) with suicide by violent methods. We aimed to determine whether suicides with postmortem evidence of SSRI consumption are more likely to have used violent methods compared with suicides with no detectable antidepressants. Blood samples from all suicides in Sweden during 1992–2004 were examined. Suicides were classified into those who died by violence and nonviolent (self-poisoning) methods using information from police records and autopsy. In addition, we investigated proportions of violent suicide in individuals who died with detectable levels of tricyclic and other antidepressants. The sample consisted of 14,691 suicides. Of the 1958 suicides with detectable levels of SSRIs, 1247 were by violent means (63.7%) compared with 7835 of 11,045 suicides (70.9%) in antidepressant-free group (x 2 1 = 7.6; P < 0.01). When subdivided by gender and age-bands, we found specific groups with signifi- cantly lower proportions of violent suicides compared with the antidepressants-free group, including men aged 15–24 years. (J Clin Psychopharmacol 2007;27:503–506) O n their introduction from the 1980s, selective serotonin reuptake inhibitors (SSRIs) were seen as a notable therapeutic advance. They are effective in treating depres- sion and anxiety but cause fewer side effects than the older tricyclic antidepressants (TCAs). However, SSRIs can produce an initial increase in anxiety or agitation, and it has been suggested since the early 1990s that this might mediate acts of violence. The evidence for this was largely anecdotal and included high-profile legal cases that sought to cast individual SSRIs as drugs that cause murder or violent suicide. 1 Nevertheless, the overall balance between benefit and harm has come increasingly to be questioned, notably in relation to the use of SSRIs in children. 2 In 2004, the US Food and Drug Administration and the UK Medicines and Healthcare Products Regulatory Agency reviewed the clinical trial data and concluded that SSRIs may marginally increase the risk of self-harm. 3,4 This has also often been understood to imply a similarly increased risk of suicide, which is hazardous because the relationship between suicide and self-harm is weak, 5 and any such increase is only likely to be detectable in large populations because suicide is a relatively rare event. Although the evidence regarding SSRI use and completed suicide from trial data is limited because of the low base rate of suicide, exclusion of suicidal individuals at baseline, and short follow-up times, meta- analyses have found no significant differences in the placebo and treatment arms. 6–8 Routinely collected population-level information provides the necessary scale for suicide research, but analysis requires appropriate comparison of either time trends or meaningful subgroups. 9 One category of routinely collected data—from postmortem studies—provides an important approach to confirm the nature of the exposure away from prescription data to forensic proof of antidepres- sant use. Nonexperimental studies in the general population have explored associations between SSRI prescribing and suicidal thoughts and acts 4,10 but are unable to account for differences between the prescriptions rates used in such studies, which is known, and actual antidepressant use, which is not. If the SSRIs are agitating, as some have argued, then one would expect an increase in the levels of suicide by violent means compared with individuals who are not taking SSRIs. A number of reports have linked SSRIs with violent suicides. 1,11,12 In addition, other changes such as hostility have been noted. 13 We aimed to examine the possible asso- ciation with violent suicides using postmortem data. Suicides in individuals without antidepressants in their blood (a substantial majority of population cases 14 ), can be compared with suicides where there is evidence of recent SSRI or TCA use. If SSRIs increase the risk of suicide by violent means, this should be expressed as a higher proportion of such deaths in those with detectable blood levels of SSRIs compared with individuals without toxicological evidence of recent antidepressant use. By contrast, suicides in those with evidence of TCA exposure should show the opposite Brief Report 503 Journal of Clinical Psychopharmacology  Volume 27, Number 5, October 2007 *Department of Psychiatry, University of Oxford, Oxford, UK; yCentre for Violence Prevention, Karolinska Institute, Stockholm and zDepartment of Forensic Toxicology, Linkoping University Hospital, Linkoping, Sweden. Received December 21, 2006; accepted after revision June 11, 2007. Address correspondence and reprint requests to Seena Fazel, MRPsych, MD, University Department of Psychiatry, Warneford Hospital, Oxford OX3 7JX, United Kingdom. E-mail: seena.fazel@psych.ox.ac.uk. Copyright * 2007 by Lippincott Williams & Wilkins ISSN: 0271-0749/07/2705-0503 DOI: 10.1097/jcp.0b013e31814ce3ef Copyright @ Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.