Antidepressants and Suicide

Abstract

At the outset let me declare that in my 35 years of psychiatric practice I have accepted pharmaceutical company hospitality and also support for various research projects, and I have written not only on the use of antidepressants, but also on nonpharmacological methods in the treatment of depression. I believe that I do not have an ideological bias for or against one form of treatment or another, or indeed their combinations; rather, the bottom line is whether or not a person afflicted with depression improves and recovers. Since the introduction of the newer antidepressants in the late 1980s there have been sporadic reports of suicidal ideation and behavior in those who have been placed on Serotonin Specific Re-uptake Inhibitor (SSRI) antidepressants (Teicher et al., 1990; Healy & Whitaker, 2003), which has led to questioning their use. On the other hand, large population studies such as those undertaken by Isacsson (2002) in Sweden, Olfson et al. (2003) in the United States, and Hall et al. (2003) in Australia have been reassuring in noting an association between increased prescribing of antidepressants and a reduction of suicide rates. Nevertheless, there has still been a body of opinion expressing reservations about the use of such antidepressants (Healy & Whitaker, 2003), and, not unexpectedly, there has been ongoing community and professional debate about this important public health issue. In 2004 a remarkable study was published in the Journal of the American Medical Association by Jick et al. This study is important because it addresses a number of concerns that have been raised. Jick et al. had access to the United Kingdom General Practice Research Database, which has enrolled more than three million people since its inception in 1987, and which contains over 35 million patient-years of data available for research. This enabled them to examine the frequency of emergent suicidal behavior following the use of two distinct pharmacological classes of antidepressants, the older tricyclic antidepressants and the newer SSRIs, in large numbers of patients. In fact, the numbers were such that one could never anticipate enlisting so many participants in a formal research project. They identified 159,810 persons aged between 10 and 69 years who had had at least one prescription of amitriptyline, dothiepin (tricyclic antidepressants), fluoxetine, or paroxetine (SSRIs) in the period 1993 to 1999. Subjects were excluded if they had had a recorded prescription for another antidepressant, or if they had diagnoses other than depression, in order to produce a more homogeneous study population. There were 36,165 who had taken amitriptyline, 46,587 who had taken dothiepin, 49,671 who had taken fluoxetine, and 35,465 who had taken paroxetine. The mean number of study antidepressantprescriptions for these persons were similar for each drug. Jick et al. analyzed the data for both nonfatal suicidal behavior as well as suicide. Nonfatal suicidal behavior was a first time record of either suicidal ideation or attempted suicide. Up to four control subjects without suicidal behavior were identified, and matched by age, sex, and duration of contact in the register, for each suicidal subject. For the suicide analysis similar selection criteria were utilized, and up to 10 controls were selected for each suicide. Jick et al. found that there were 555 persons with a first time episode of nonfatal suicidal behavior, including suicidal ideation, and there were 2062 control subjects. There was no difference between the four study drugs in terms of the emergence of suicidal ideation or behavior, and there was no evidence of any effect modification by age or sex. Compared with those who started taking an antidepressant more than 90 days before their first suicidal behavior, those who started taking their antidepressant within 1 to 9 days before developing suicidal behav-