Are Selective Serotonin Reuptake Inhibitors a Risk Factor for Adolescent Suicide?

Abstract

Can J Psychiatry. 2009;54(2):69-71. In 1959, at a meeting in Cambridge celebrating Imipramine - a potent serotonin reuptake inhibitor - a series of speakers distinguished between the recognized risk of suicide stemming from electroconvulsive therapy rolling back psychomotor retardation before depressive suicidality was alleviated and imipramine-induced agitation and suicidality.1 Thereafter, the received clinical wisdom in Europe was that antidepressants could trigger suicidality. When SSRIs came on stream, controlled trials had become an important method of evaluating their clinical effects. Had these trials been meta-analyzed for suicidal acts, it is now clear that as of 1 988 they would have shown a doubling of the risk of suicidal acts on SSRIs, compared with placebo.2 Against this background, a series of articles between 1 990 and 1992 describing an early onset, dose-related emergence of suicidality on fluoxetine that cleared on discontinuation and reemerged on reexposure was not surprising.3 Using all standard causality algorithms, these studies demonstrated a convincing causal link between treatment and adverse effect. The original article described 6 nonfatal cases but omitted a death by suicide in a person aged 14 years with obsessivecompulsive disorder.4 King et al5 described identical effects in children to those in adults. While claiming that there was no significant difference in suicidal act rates between active treatment and placebo, all meta-analyses of clinical trial data from 1991 to 2005 in fact reported an excess of suicidal acts on active treatment.6 The excess would have been greater but for the fact that these analyses included, under the heading of placebo, acts that happened during the pre- and (or) post-randomization phase of the trials analyzed.6 A 2003 analysis of placebo-controlled trials of antidepressants in anxiety concluded - on the basis of 11 suicides in 12 914 patients on active treatment, compared with 0 suicides in 3875 patients on placebo - that anxiety was a risk factor for suicide, suggesting a profound bias against recognizing the risks stemming from treatment.7 In 2004, the suicide risk of antidepressants came to the fore as linked to possible inefficacy of these agents in children, providing evidence for widespread failure to report trials and ghostwriting of those published. FDA analyses of the data confirmed a doubling of the risk of suicidal behaviours on active treatment, compared with placebo. Among those concerned that these data may deter patients from seeking treatment, it has been claimed that: there were in fact no deaths by suicide in these trials, increased risk may stem from patients on active agents verbalizing their ideation or acts, and, since the launch of SSRIs, national suicide rates have largely fallen. A series of articles have shown suicidal act rates falling in depressed patients after the institution of treatment, and there have been suggestions of increased pediatric suicide rates in the United States and Holland since warnings were placed on antidepressants, although these uncontrolled studies may demonstrate little more than the fact that patients with any condition, from influenza to depression, seek or receive treatment when at their worst, and the natural history of these disorders leads to an improvement in the post-consultation period, even if no effective treatments are available. The argument from national suicide rates omits data from the 1960s and 1970s when national suicide rates climbed, even though antidepressant prescriptions among those at greatest suicide risk increased maximally. The 1990s' data fail to distinguish between the 80% of antidepressants administered in longer-term treatment to patients at little risk from their treatment, and the at-risk group of first-exposure patients who take possibly no more than 20% of antidepressant prescriptions in any 1 year. Once such distinctions are made, it is possible to reconcile an increased exposure to a treatment-induced risk with falling national suicide rates. …