Intended for healthcare professionals


Randomised controlled trial of chest pain units is needed

BMJ 2000; 321 doi: (Published 07 October 2000) Cite this as: BMJ 2000;321:896
  1. Steve Goodacre (steveg{at}, research fellow,
  2. Francis Morris, consultant in accident and emergency medicine,
  3. Simon Capewell, professor of clinical epidemiology
  1. Department of Accident and Emergency Medicine, Northern General Hospital, Sheffield S5 7AU
  2. Department of Public Health, University of Liverpool, Liverpool L69 3GB

    EDITOR—Collinson et al audited prognostically important myocardial damage in patients discharged from the emergency department—a large and important healthcare problem.1 The commentary by Kennedy touches on several key issues as well as the obvious limitations of such a study. As Kennedy says, further follow up data on the seven patients with raised concentrations of troponin would be useful.

    Kennedy also touched on the issue of definitions. The seven missed Mayday patients had myocardial damage but would not satisfy the definitions of acute myocardial infarction held by the World Health Organization or the monitoring trends and determinants in cardiovascular disease (MONICA) study. Collinson et al are, however, correct to point out that such patients have a less favourable prognosis. Should these classic definitions be reconsidered? The 6% of discharged patients with troponin T concentrations above 0.1 ng/ml have a substantial short term mortality and morbidity,2 yet this can be improved by treatment.3 Discharge of such patients should be avoided, even if they fail to meet classic definitions of acute myocardial infarction.

    Chest pain observation units may prevent these inappropriate discharges. One such unit has been operational in the accident and emergency department of the Northern General Hospital in Sheffield since March 1999.4 Similar follow up has been used to monitor those discharged (electrocardiography and troponin T concentration 72 hours after attendance). So far 761 patients have been assessed on the unit, of whom 86% were successfully discharged; 580 (88%) of those discharged attended follow up. Only one patient had a concentration of troponin T>0.1 ng/ml (0.17%). Using the recommended audit tool of Collinson et al in this way suggests that the Sheffield chest pain observation unit performs well. However, longer term follow up data need to be collected, and, ultimately, only a randomised controlled trial can tell us whether the chest pain observation unit is superior to routine care.

    A recent systematic review has examined the efficacy of chest pain observation units and their apparent cost effectiveness,5 but most data come from the United States. We believe that there is an urgent need for a randomised controlled trial in the very different circumstances of the British NHS.


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

    Log in

    Log in through your institution


    * For online subscription