General Practice

Antidepressant drug use in primary care: a record linkage study in Tayside, Scotland

BMJ 1996; 313 doi: (Published 05 October 1996) Cite this as: BMJ 1996;313:860
  1. T M MacDonald, reader in clinical pharmacologya,
  2. A D McMahon, biostatisticiana,
  3. I C Reid, professor of psychiatrya,
  4. G W Fenton, professor of psychiatrya,
  5. D G McDevitt, professor of clinical pharmacologya
  1. a Medicines Monitoring Unit, Department of Clinical Pharmacology, and Department of Psychiatry, University of Dundee, Ninewells Hospital and Medical School, Dundee DD1 9SY
  1. Correspondence to: Dr T M MacDonald, Medicines Monitoring Unit.
  • Accepted 17 June 1996

Depression is a common psychiatric disorder in general practice.1 There is consensus that antidepressants should be given in sufficient dose and for sufficiently long if treatment is to be effective.2 Tricyclic antidepressants carry warnings about their use in patients with cardiac diseases, and they have well known anticholinergic effects that could cause predictable toxicity. Selective serotonin uptake inhibitors are thought to be devoid of such effects. We studied the use and safety of antidepressant drugs in the community, with reference to their labelled warnings.

Patients, methods, and results

We studied patient specific dispensed prescribing3 for antidepressant drugs and admissions to hospital for possible adverse effects over 14 months using a record linkage database that covers the population of Tayside, Scotland (about 400 000 subjects).4 The dose and duration of use of antidepressants was determined, as was the number of subjects who had labelled contraindications to and cautions against taking tricyclic antidepressants.

Overall, 20 195 patients received 85 315 prescriptions for an antidepressant drug (14 988 patients received tricyclic antidepressants, 3571 received selective serotonin reuptake inhibitors, and 1636 received both). In comparison with published guidelines,2 tricyclic antidepressants were generally given in subtherapeutic doses (10 804 (72%) subjects received only subtherapeutic doses, 1125 (7%) therapeutic or both doses, and 3059 (21%) unknown doses). Selective serotonin reuptake inhibitors tended to be given in therapeutic doses (281 (8%) received only subtherapeutic doses, 2828 (79%) therapeutic or both doses, and 642 (13%) unknown doses). In both drug groups most prescribing was of short duration (38% of subjects </= 30 days, 58% </= 60 days, 68% </= 90 days).

Despite labelled warnings, 26 out of 16 624 subjects (0.2%) were dispensed a tricyclic antidepressant and 7 out of 5207 subjects (0.1%) a selective serotonin reuptake inhibitor within 30 days of admission for a myocardial infarction. Fourteen subjects (0.1%) had been admitted with heart block, 179 subjects (1.1%) with an arrhythmia, 86 subjects (0.5%) with acute glaucoma, 159 (1.0%) with thyroid disease, 165 (1.0%) with epilepsy, and 1149 (6.9%) with a cardiovascular disorder; all had had one or more prescriptions for a tricyclic antidepressant. Selective serotonin reuptake inhibitors had been prescribed to two (0.04%) patients admitted for heart block, 82 (1.6%) for arrhythmia, 23 (0.4%) for glaucoma, 49 (0.9%) for thyroid disease, 46 (0.9%) for epilepsy, and 357 (6.9%) for a cardiovascular disorder. Tricyclic antidepressants were dispensed to 6303 subjects (37.9%) and selective serotonin reuptake inhibitors to 1897 subjects (36.5%) who had also been prescribed a cardiovascular drug as listed in the second chapter of the British National Formulary. Antidepressants did not result in increased admission for known adverse effects (table 1), with the exception of a significantly increased risk of admission for acute retention of urine with tricyclic antidepressants, the odds ratio compared with non-exposure being 1.9 (1.2 to 2.9).

Table 1

Numbers of events, exposures, event rates, and adjusted odds ratios with 95% confidence intervals for each end point studied

View this table:


We found that tricyclic antidepressants were commonly used in low doses that were unlikely to be of benefit in treating major depression,2 which confirmed a recent report.5 Antidepressants were commonly given for short and probably ineffective durations. Tricyclic drugs were given to some patients who had labelled warnings, but, despite this, and perhaps because tricyclic antidepressants were prescribed in low doses, we found no increased risks of admission for adverse cardiovascular outcomes with tricyclic drugs. There was an increase in the risk of acute retention of urine, possibly because of the anticholinergic effect of tricyclic drugs.

These results are reassuring in that tricyclic antidepressants seem to be safe in the doses used in clinical practice. Although antidepressants are given for indications other than depression, the results are worrying because tricyclic antidepressants were mostly prescribed in low doses and both types of antidepressant were given only short term. These two factors mean that the treatment of major depression is less likely to be effective. Further research should determine why antidepressant drugs are prescribed in this way.


  • Funding Research grant from SmithKline Beecham Pharmaceuticals. The Medicines Monitoring Unit is supported by the Medicines Control Agency.

  • Conflict of interest The study was funded by SmithKline Beecham Pharmaceuticals.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract