Intended for healthcare professionals


Depression, antidepressants, and accidents

BMJ 1995; 311 doi: (Published 07 October 1995) Cite this as: BMJ 1995;311:887
  1. J Guy Edwards, Honorary clinical senior lecturer
  1. University Department of Psychiatry, Royal South Hants Hospital, Southampton SO14 0YG

    Pharmacological concerns need epidemiological elucidation

    Patients with depressive disorders could be more liable to accidents than others for several reasons. These include impaired attention and concentration (often due to preoccupation with morbid or worrying thoughts); anxiety, which usually accompanies depression; irritability; agitation or retardation; fatigue after sleepless nights; and weakness due to insufficient food intake. In addition, depressed patients may take risks with little regard to the consequences, make suicidal gestures or attempts that go wrong, and take alcohol or other substances that impair functioning to relieve distress. In keeping with these considerations are the reductions in performance in tests of cognitive and psychomotor function in depressed patients.1

    If depression causes or contributes to accidents in these ways then therisk of accidents might be expected to fall as the disorder improves in response to treatment with antidepressant drugs (or other treatment). On the other hand, the unwanted effects of antidepressants may adversely affect cognitive and psychomotor function (especially in those with pre-existing cerebral impairment), thereby increasing the risk of accidents. This is particularly the case in patients who do not respond to treatment and thus are impaired by both the depression and the drug. Drug induced effects may be clinically obvious, with, for example, somnolence due to oversedation; mild effects may be evident only on formal psychometric testing. In tests of higher central nervous system functions the older tricyclic antidepressants, such as amitriptyline and dothiepin, and some newer drugs, such as mianserin and trazodone, cause more impairment than the newer tricyclic, lofepramine, and the selective serotonin reuptake inhibitors.2

    Driving performance can be tested in car simulators fitted with computer controlled steering wheels and brakes, with images of changing traffic conditions projected on to screens wrapped around the vehicle. Driving in real life is assessed at test sites and, with dual controls, on the open road. Although such tests have been largely abandoned in Britain because of safety and medicolegal concerns, not one accident occurred in Dutch studies involving more than 1000 volunteers and patients, who collectively drove more than one million kilometres (J F O'Hanlon, personal correspondence, 1995). When a specially instrumented car was used on a road tricyclic antidepressants were shown to impair tracking ability (that is, control of lateral position),3 although the impairment largely resolves after a week's treatment owing to adaptation.4 6 In these studies reversible inhibitors of monoamine oxidase A and selective serotonin reuptake inhibitors had little or no detrimental effect on driving.4 5 6 Dothiepin given in a therapeutic dose at night for three weeks also had no appreciableadverse effects on driving.7

    As the older tricyclic drugs have been shown in drug trials to cause more sedation than newer antidepressants, such as selective serotonin reuptake inhibitors and reversible inhibitors of monoamine oxidase A, and more impairment in tests of cognition and psychomotor performance, it has been assumed that they are responsible for many of the accidents that occur on the road, at work, and in the home. However, the predictive validity of laboratory tests (especially those involving the administration of single doses of a drug to volunteers) as reflections of what happens in the real world has been questioned,1 8 9 and no epidemiological data exist to support the view that substituting newer antidepressants for older tricyclic drugs would lead to fewer accidents.

    Many skilled tasks can be performed without undue effort and with spare processing capacity available, and it has been suggested that tasks entailing the processing of information are measures of potential performance (integrity of the CNS) rather than of actual performance.10 Furthermore, patients may draw on their previous experience and compensate for small drug induced decreases in functioning. Although adverse effects of tricyclic antidepressants on critical flicker fusion (an index of arousal and ability to process information) could persist for weeks,11 adaptation to other psychomotor effects including decreases in driving performance have been shown to occur during continued treatment.4 5 6 7


    Studies of elderly drivers have shown that treatment with cyclic antidepressants is associated with an increased risk of vehicle crashes in which injuries are sustained; in addition, a dose-effect relation exists.12 13 This suggests a contributory role for these drugs, although inability to control for some potentially confounding variables, especially the health status of the drivers, does not allow definite conclusions to be reached. A study of loss of consciousnesswhile driving found risk factors other than antidepressants, such as working long hours, alcohol consumption, and organic brain diseases.14 Drugs may, however, contribute to traffic crashes through more subtle mechanisms.

    Accidents may also result from falls caused by postural hypotension (mediated by the anti-(alpha) adrenoceptor effect of antidepressants) and, rarely, drug induced cardiac arrhythmias and convulsions. Elderly people, particularly those with disorders affecting stability and mobility (for example, neurological diseases, defective vision, vertigo, and arthropathies) are especially vulnerable. Patients treated concurrently with other drugs that decrease blood pressure, such as diuretics, are also at higher risk. Falls may result in Colles' fractures and fractures of the neck of femur, although femoral fractures mostly occur spontaneously in elderly people with low bone mass (osteopenia) related to age, and the role of drugs that sedate and impair postural control in the aetiology of hip fractures has been questioned.15

    Consider newer antidepressants for patients who:

    • are elderly, especially if cognitively impaired

    • have severe disorders affecting stability or mobility

    • experience persistent sedation or transient hypotension when taking tricyclic antidepressants

    • require combinations of drugs with central nervous system or hypotensive effects

    • misuse or are suspected of misusing alcohol or illegal drugs

    In contrast to older tricyclic antidepressants, newer antidepressants are relatively free of both sedative and anti-(alpha) adrenoceptor effects and may therefore be less liable to cause or contribute to falls and fractures. Supporting this is the low incidence of such drug related complications found in prescription event monitoring studies of fluvoxamine16 and other selective serotonin reuptake inhibitors (J G Edwards et al, unpublished findings). Comparable prescription event monitoring data on tricyclic antidepressants are not available.

    Insufficient epidemiological data are available to indicate the extent to which antidepressants (particularly new drugs) cause or contribute to accidents. This is especially true in the case of traffic accidents because of the many interacting variables and the many methodological and practical obstacles to research in this field. Nevertheless, given the known pharmacological effects of the older tricyclic drugs, patients should be warned of the potential risks while driving17 and while working in dangerous situations, especially when treatment is started, the dose is increased, and the drugs are taken with other substances that affect cognition and psychomotor performance. Until more data are available it is best to err on the side of safety and prescribe newer antidepressants, such as selective serotonin reuptake inhibitors and reversible inhibitors of monoamine oxidase A for those at high risk.

    But pharmacological concerns need more epidemiological support before these drugs are recommended as first line treatment in all patients.18


    View Abstract