Papers

Relative mortality from overdose of antidepressants

BMJ 1995; 310 doi: https://doi.org/10.1136/bmj.310.6974.221 (Published 28 January 1995) Cite this as: BMJ 1995;310:221
  1. John A Henry, consultant physiciana,
  2. Carol A Alexander, research assistanta,
  3. Ersin K Sener, research fellowa
  1. a National Poisons Unit, Guy's Hospital, London SE1 9RT
  1. Correspondence to: Dr Henry.
  • Accepted 16 December 1994

Abstract

Objective: To compare the fatal toxicities of antidepressant drugs in 1987–92.

Design: Retrospective epidemiological review of prescription data of the Department of Health, Scottish Office Home and Health Department, and Welsh Health Common Services Authority (excluding data from most private general practices and most hospitals), and mortality data from the Office of Population Censuses and Surveys and General Register Office in Scotland.

Setting: General practice, England, Scotland, and Wales.

Main outcome measures: Deaths per million prescriptions and deaths per defined daily dose.

Results: 81.6% (1310/1606) of deaths from antidepressant overdose were due to two drugs, amitriptyline and dothiepin. The overall average of deaths per million prescriptions was 30.1. The overall rate for tricyclic drugs was 34.14 (95% confidence interval 32.47 to 38.86; P<0.001), monoamine oxidase inhibitors 13.48 (6.93 to 22.19; P<0.001), atypical drugs 6.19 (4.04 to 8.80; P<0.001), and selective serotonin reuptake inhibitors 2.02 (0.64 to 4.17; P<0.001). The numbers of deaths per million prescriptions of amoxapine, dothiepin, and amitriptyline were significantly higher than expected, while nine drugs had a significantly lower number of deaths per million prescriptions than expected. Analysis of deaths per defined daily dose showed a similar pattern.

Conclusions: Safety in overdose should be considered in risk-benefit and cost-benefit considerations of antidepressants. A switch in prescribing, from drugs with a high number of deaths per million prescriptions to drugs with a low number, could reduce the numbers of deaths from overdose. Although this form of suicide prevention can be implemented easily and immediately, its introduction needs to be considered against the higher costs of some of the newer drugs.

Key messages

  • Key messages

  • Antidepressant drugs such as the selective serotonin reuptake inhibitors and lofepramine have low toxicity in overdose

  • Deaths from overdose can be prevented by switching prescribing from older, tricyclic drugs; such a move may contribute to government targets for reducing suicide

  • Choice of a first line antidepressant should be based on several factors, including adverse effect profile, compliance, overdose safety, and cost, and the antidepressant should be appropriate to the patient and his or her clinical condition

  • Patients with evidence of suicidal ideation should be given special consideration—admission to hospital may be indicated

Introduction

While little demonstrable difference exists between antidepressants in terms of efficacy,1 toxicity in overdose varies widely.2 We compared the fatal toxicities of antidepressants currently available in Britain individually and by group during 1987–92, during which time the selective serotonin reuptake inhibitors were introduced.

Methods

Antidepressants were assigned to four classes: monoamine oxidase inhibitors, tricyclic antidepressants, selective serotonin reuptake inhibitors, and the so called “atypical”3 antidepressants. We obtained numbers of deaths in England, Wales, and Scotland due to acute poisoning by a single antidepressant.4 5 The statistics and research division of the Department of Health supplied data on the number of antidepressant prescriptions for general medical practices within the NHS for England, Wales, and Scotland for 1987–9 and for England for 1990–2; for 1990–2 the Scottish data were provided by the Scottish Office Home and Health Department and the Welsh data by the Welsh Health Common Services Authority. Private general practice and most hospitals were excluded; figures for 1991 and 1992 include dispensing practices. About three quarters of all drug prescriptions are written by general practitioners.6 7 Most patients with psychiatric disorders, however, are treated by general practitioners rather than by psychiatrists8 9 10 (90% to 98% of depressed patients9 10 11). The exclusion of hospital prescriptions from our analysis should not, therefore, appreciably affect the outcome of this study.

We calculated the number of deaths per million prescriptions during the six years for all the drugs taken together; for each of the four groups of antidepressants; and for each drug individually. The x2 test was applied to the groups of antidepressants. The expected numbers of deaths are given for the individual drugs, with Fisher's exact test (one tailed) applied to the data. Confidence limits are calculated as x+/-1.96 SD (x).

Using the prescribed data, we considered each preparation of each drug analysed, multiplied the strength of the preparation by the quantity prescribed, and divided this by the defined daily dose values (obtained from the World Health Organisation).12 13 We calculated the number of defined daily doses per prescription for each drug and the number of deaths per million defined daily doses prescribed.

Results

The mean annual number of deaths due to overdose with a single antidepressant over the six years was 268 (range 238 to 288). The tricyclic drugs were implicated in most deaths (table I), with two drugs—amitriptyline and dothiepin—accounting for 81.6% of all deaths. Tables II and III show the figures for mortality and data for prescriptions for deaths per million prescriptions for the four groups of drugs for the six years. The tricyclic antidepressants as a group had a significantly higher number of deaths per million prescriptions than expected compared with all the antidepressants taken together (P<0.001). The monoamine oxidase inhibitors as a group had a lower than expected number of deaths per million prescriptions (P<0.001). The groups of atypical antidepressants and selective serotonin reuptake inhibitors each had the lowest number of deaths per million prescriptions (P<0.001). Tables IV and V list each drug in each group with its number of deaths per million prescriptions, which is used to rank them within groups, and shows that three of the tricyclic agents (dothiepin, amitriptyline, and amoxapine) had a significantly higher number of deaths per million prescriptions than expected. A further three drugs from this group (lofepramine, clomipramine and trimipramine) had a significantly lower number of deaths per million prescriptions than expected when compared with all antidepressants. One monoamine oxidase inhibitor (phenelzine) had a significantly lower number of deaths per million prescriptions. Two of the atypical drugs (mianserin and trazodone) had a significantly lower number of deaths per million prescriptions. Three of the selective serotonin reuptake inhibitors (fluoxetine, fluvoxamine, and paroxetine) had a lower number of deaths per million prescriptions. No deaths were recorded for five drugs, all of which had low prescription figures (table III). Calculation of data with defined daily doses showed a pattern that was broadly similar to the data derived from deaths per million prescriptions.

TABLE I

Mean yearly numbers of deaths from tricyclic and other antidepressants, 1987-92

View this table:
TABLE II

Fatal poisonings and deaths per million prescriptions for deaths from single antidepressant, by groups of drug. Values in parentheses are 95% confidence intervals

View this table:
TABLE III

Fatal poisonings and deaths per million defined daily doses for deaths from single antidepressants, by groups of drug. Values in parentheses are 95% confidence intervals

View this table:
TABLE IV

Fatal poisonings and deaths per million prescriptions for deaths from single antidepressants, by individual antidepressant. Values in parentheses are 95% confidence intervals

View this table:
TABLE V

Fatal poisonings and deaths per million defined daily doses for deaths from single antidepressants, by individual antidepressant. Values in parentheses are 95% confidence intervals

View this table:

Discussion

The atypical antidepressants form a heterogeneous group of drugs, while the other antidepressants fall into three pharmacologically homogeneous groups. The selective serotonin reuptake inhibitors share a common mechanism of action, despite their remarkably differing chemical structures (fluvoxamine is a monocyclic agent, fluoxetine a bicyclic agent, sertraline a naphthylamine derivative, and paroxetine a phenyl piperidine derivative). They also had the lowest toxicity in overdose of the groups of drugs studied here. This accords with clinical experience.14 15 16 17 The numbers of deaths per million prescriptions of antidepressants have been shown to be inversely related to their serotonin reuptake inhibition activity,18 but this relation may be coincidental with their structural properties.

Shortcomings of a study of this nature include systematic error in prescription or mortality data, or both; confounding by prescriber biases or patient biases—such as, an inability to distinguish between the use of “first line” and “second line” drugs—or both biases; and the fact that no allowance was made for prescribing for indications other than those of concern to the study. These factors, however, are unlikely to affect the conclusions of our study as most biases run in favour of the tricyclic drugs, which are more widely used as first line drugs.19 It is also possible that selective serotonin reuptake inhibitors are being prescribed to patients at greater risk of overdose.20 These drugs may also have a different role from tricyclic drugs in patients with depression resistant to treatment—for example, paroxetine3—or with suicidal thoughts—for example, fluoxetine21 and fluvoxamine.22 Our data thus provide a useful guide to the relative toxicities of drugs and an indication of the needs of prescribing policy and correlate with the results of median lethal dose in animals.23 Our data also agree with alternative indices based on deaths per million standard quantity units or deaths per thousand kilograms of drug prescribed.24

The use of data on defined daily dose is gaining popularity internationally, mainly because defined daily doses provide a standardised technical measure of drug use that is not influenced by strength of dosage form. The dose is an assumed average daily dose for the main indication of a drug as determined by the Nordic Council on Medicines,13 and this causes problems in its use in a study like ours. Problems include variation in the number of prescriptions that depressed patients have each year; discontinuation or non-compliance, especially for tricyclic drugs,25 which increases the apparent market share of other antidepressants; and prescription of drugs at doses below or above the defined daily dose. These factors might be expected to bias toxicity data in favour of the tricyclic drugs,26 but our analysis shows a strong correlation in the ranking of the two indices.

The differing toxicities of antidepressants in overdose should be considered against the wider issue of costs. The selective serotonin reuptake inhibitors are relatively expensive and are generally prescribed at effective daily doses, which can hardly be reduced to save on costs. Ironically, it is the cheaper, tricyclic antidepressants that are more likely to be prescribed in subtherapeutic doses10 27 to avoid adverse effects.10 27 28 The widely perceived cost advantage of tricyclic drugs is based almost entirely on market price alone29; many other factors influence the cost to the NHS, and virtually all models of clinical practice give the selective serotonin reuptake inhibitors a cost advantage.29 30 31 One cost effectiveness study estimated that (paunds sterling)19000–173000 per life year was gained from deaths from overdose prevented by a switch in routine first line use from tricyclic drugs to selective serotonin reuptake inhibitors.32 This finding rested, however, on the assumption that the only differences between the different groups of antidepressant drugs lay in the market prices of the drugs and the numbers of deaths from overdose. Many other factors would need to be considered for a valid comparison. Overdose itself has cost implications, regardless of whether death results. The tricyclic drugs are more likely to lead to medical complications and admissions to intensive therapy units.33

The British Association for Psychopharmacology concluded that the newer antidepressants have real advantages over the older ones in terms of safety and tolerability when given in an effective dose.34 Low toxicity in overdose is an important consideration in the drug treatment of depressed patients. Although it can be argued that a failed suicide attempt may only delay the ultimate outcome, it is widely accepted that a suicide attempt is a poor predictor of further attempts35; about 10% of those who fail to kill themselves with acute overdose go on to successful suicide.36 Furthermore, a drug overdose might bring the patient under closer medical supervision and lead to a better outcome to the depressive episode. While the present study shows that fatal overdose is an important problem, untreated depression may be even more important.26

As a result of the “Defeat Depression” campaign by the Royal Colleges of Psychiatrists and General Practitioners the diagnosis of depression may increase over the next few years. Wider use of drugs with a high number of deaths per million prescriptions could lead to an increase in the number of successful suicides from antidepressant overdose. Conversely, wider use of drugs with a low number of deaths per million prescriptions would make a contribution towards the government's targets targets for the year 2000 of reducing the number of suicides.

EKS is supported by the Turkish Education Institute and the Turkish ministry of health.

References

  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.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
View Abstract