Intended for healthcare professionals


Suicide and antidepressants

BMJ 1995; 310 doi: (Published 28 January 1995) Cite this as: BMJ 1995;310:205
  1. J Guy Edwards
  1. Honorary clinical senior lecturer University Department of Psychiatry, Royal South Hants Hospital, Southampton SO9 4PE

    Controversies on prevention, provocation, and self poisoning continue

    Antidepressants are mostly prescribed for depression and its complications. Foremost among these is suicide, but, despite these drugs having been available for over 35 years, no unequivocal evidence that they prevent suicide exists. Ironically, some antidepressants have been alleged to provoke self destructive behaviour, and patients sometimes kill themselves with the drugs prescribed to treat their depression.

    Antidepressant drugs are more effective than placebo in treating acute episodes of affective disorders, and they may help prevent relapses and recurrences of depression. Although they also decrease the score on the “suicide item” on the Hamilton rating scale for depression, this is not a valid measure of suicidal intent or a predictor of future self destruction. Suicide is rare, and small differences in suicide rates during treatment with different antidepressants could be due to chance or unsatisfactory matching between groups. Thus, saying that antidepressants prevent suicide could be more a statement of faith than of fact.

    Depression has long been considered to be an uncommon paradoxical adverse reaction to antidepressants.1 Maprotiline has been thought to provoke suicide in some patients because a controlled trial found that more suicides occurred in patients receiving the drug than in those receiving placebo (although the difference was not significant).2 Significantly more suicidal attempts or gestures occurred in the group given maprotiline, but populations who show such behaviour are different from those who commit suicide. Fluoxetine has similarly been suspected of provoking suicidal behaviour because violent self destructive thoughts were reported in six patients treated with the drug.3 Reviews suggested that the suicidal ideation was more likely to have been due to the disorders being treated,4 while a meta-analysis found that suicide was not significantly more common during treatment with fluoxetine than with tricyclic antidepressants or placebo.5 More recently, prescription event monitoring has not found a higher incidence of suicide during treatment with fluoxetine than with fluvoxamine, another selective serotonin reuptake inhibitor.6 Thus no totally convincing evidence exists that antidepressants either provoke or prevent suicide.

    What is more certain, however, is that people kill themselves by taking overdoses of antidepressants and that tricyclic drugs introduced into practice before 1970 are more lethal than those introduced since. A study of deaths from self poisoning between 1976 and 1984 showed that the older drugs have a higher fatal toxicity index (the number of deaths due to overdose per million prescriptions).7 The research, updated to 1987, showed that fluvoxamine also has a low fatal toxicity index,8 while the most recent update (to 1992) showed that fluoxetine, paroxetine, and sertraline similarly have low lethality in overdose (p 221).9 Despite the limitations of the methodology, especially uncertainty over the causes of death, the quantities of drugs and other substances taken, and the medical condition of the patients, the results provide strong circumstantial evidence that death due to overdoses of antidepressants is more likely to occur in those prescribed older tricyclic drugs than those prescribed newer antidepressants. This is consistent with the known cardiotoxic effects of the older drugs and recent work showing that overdoses of dothiepin are more likely than those of other antidepressants to cause seizures and cardiac arrhythmias, which are regarded as intermediate outcome measures.10

    In view of these observations and the knowledge that patients survive large overdoses of selective serotonin reuptake inhibitors, some commentators have suggested that prescribing older tricyclic drugs for depressed patients while safer drugs are available is unethical and irresponsible. To do so, they say, would justify any claim for negligence that ensued. It is necessary, however, to see the risks in a clinical and epidemiological perspective.

    Firstly, although the lifetime prevalence of suicide in patients with major depression is high, a much smaller proportion take their lives during the relatively short period when they are prescribed antidepressants as first line treatment than at other times. More important than the choice of drug in most non-suicidal patients is a thorough assessment of suicidal risk. Although there are no accurate predictors of suicide in the long term, patients known to be at high risk in the short term should be targeted for special care. The ethical and medicolegal questions that should be asked therefore are not only which antidepressant was prescribed but how competently the risk of suicide was assessed; what support, supervision, and treatment other than an antidepressant were given; and why that particular antidepressant was prescribed for the patient at that time. Merely handing a patient a prescription for an antidepressant (however safe in overdose it might be) without a proper assessment and care plan could be regarded as negligent.

    Secondly, only about 4% of all suicides are due to overdoses of single antidepressants.11 It is not known what proportion of these overdoses is taken during treatment—that is, when choice is more relevant—and statistics on deaths due to overdoses of antidepressants include those due to accidental overdoses and overdoses of other people's supplies, often taken by people without psychiatric disorders.

    Thirdly, it is not known whether different suicide rates in patients given different antidepressants are influenced by doctors' perception of risk; if this is the case it will confound attempts to determine whether the prescribing of one drug or group of drugs is more or less likely to result in suicide. Isaacson et al recently showed that groups of patients treated with different antidepressants may not be comparable with respect to diagnosis and severity of illness.12 Amitriptyline is chosen more commonly for patients with severe depression and depression associated with severe insomnia, while lofepramine, mianserin, moclobemide, and possibly amitriptyline seem to be chosen more commonly for patients prone to suicide.

    Finally, patients treated with tricyclic antidepressants may not be at greater overall risk of suicide (by any method) than those treated with safer drugs, as people who are determined to kill themselves will choose a means of doing so. Consistent with this are data from the Office of Population Censuses and Surveys showing that, although deaths due to self poisoning in England and Wales have decreased, those resulting from more violent methods have increased.11 Even more convincing is the report in this week's journal by Jick et al (p 215), which shows, in an open cohort study with a nested casecontrol analysis carried out in general practices in Britain, that the risk of suicide during treatment with 10 different antidepressants (new and old) is similar.13 In the debate on selective serotonin reuptake inhibitors versus tricyclic drugs the safety in overdose of selective serotonin reuptake inhibitors has been offered as the most compelling reason for their use as first line treatment; the work of Jick et al militates against this argument.


    1. 1.
    2. 2.
    3. 3.
    4. 4.
    5. 5.
    6. 6.
    7. 7.
    8. 8.
    9. 9.
    10. 10.
    11. 11.
    12. 12.
    13. 13.
    View Abstract