Intended for healthcare professionals


Chest pain units

BMJ 2002; 325 doi: (Published 20 July 2002) Cite this as: BMJ 2002;325:116

Evidence of their usefulness is limited but encouraging

  1. Mike Clancy, consultant in emergency medicine
  1. Emergency Medicine, Southampton General Hospital, Southampton SO16 6YD

    Patients presenting to hospital with chest pain represent a substantial burden to the NHS. About 500 000 patients attend emergency departments in the United Kingdom each year with chest pain, 1 2 and 20–30% of all medical admissions are for acute chest pain.3 Currently most clinicians working in emergency departments rely on the history, clinical examination, and electrocardiogram (ECG) to decide which patients have acute coronary syndromes and need admission and which to send home. Given the unreliability of these tools alone to either rule in or rule out the diagnosis of acute myocardial infarction4 and unstable angina, it is not surprising that 2–4% of patients with acute myocardial infarction have been sent home from American emergency departments with a high case fatality rate and medicolegal costs. 5 6 The position in the United Kingdom is uncertain, but a recent study identified that 6% of patients who were discharged from an emergency department had prognostically significant myocardial damage.7 Equally, of those patients admitted for further investigation of their chest pain, fewer than half will have acute coronary syndromes.8 The possible inappropriate admission of the majority of these patients has considerable cost consequences for the NHS.

    Chest pain units, first developed in the United States, attempt to improve diagnostic accuracy, shorten length of hospital stay, and save money.9 Such units take patients who have already had a full history, examination, and ECG and do not have an acute coronary syndrome and have been assessed to be at low or moderate risk. Usually these units are located within emergency departments and are protocol driven. Typically, patients are closely monitored for 6–12 hours, subjected to a battery of biochemical tests, serial ECGs, and often ST segment monitoring and an exercise ECG. If all these tests are negative the patient is sent home, but if tests are positive or equivocal the patient is admitted for further investigation and treatment. However, the units vary in their selection criteria, length of observation, use of cardiac tests, and discharge criteria.

    Have chest pain units led to any improvement in diagnostic accuracy and clinical outcome? Most of the evidence comes from North America and shows that chest pain units are safe, with no significant difference in event rate, reattendance, or readmission rate compared with conventional care.10 However, these studies were not sufficiently powered to show improvement in rare events (mortality of patients inadvertently sent home), and in many studies all those who served as controls were admitted. Diagnostic certainty was increased and length of stay reduced. The economic analyses are predominantly North American and suggest modest savings. Often the economic perspective taken is departmental rather than from a healthcare system or societal perspective. Whether such savings would be made in the United Kingdom with its very different practices (much less interventional radiology, and higher discharge rates from emergency departments) is uncertain: there is little evidence to guide us.

    In the United Kingdom a small but increasing number of emergency departments are running chest pain units, and the limited evidence available of their diagnostic performance is encouraging.11 These are very different from the new and more common chest pain clinics, which deal with patients judged by their primary care practitioner not to require emergency care. Patients are seen by cardiologists and may undergo provocative cardiac testing, but the clinics do not usually provide observation and biochemical testing.

    The government's focus on coronary heart disease, with its recommendation that patients or their doctors call for an ambulance in the event of symptoms that suggest acute myocardial infarction,12 make it almost certain that the large numbers of patients with chest pain currently seen in emergency departments will increase. If the patient group is to be dealt with safely and efficiently then practice will need to change. The difficulty is that the problem is already here and getting worse, and the ideal evidence from a UK multicentre randomised controlled trial is absent. We have to decide urgently whether the systematic approach to the diagnosis of chest pain in those patients who present as emergencies, such as the approach offered by chest pain units, is likely to be better than existing care. If the answer is yes then the investment in such units is needed. A method of dealing with the growing wave of patients with chest pain in emergency departments is needed. Otherwise, the wards will be swamped with patients who do not need to be there, and as the pressure to avoid admission inexorably rises, so will rates of inappropriate discharge from emergency departments.


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