Heart test questions

Behind the Headlines

Friday November 14 2008

Newspaper reports that a common heart test is 'worthless' in assessing the risk of heart attack are an exaggeration of the research results they are based on.

The Daily Mail says, “An NHS test to assess a patient's risk of heart attack is worthless” and “Simply talking to patients about their symptoms would be be more effective” than using an electrocardiogram, or ECG. (An ECG produces a graph of heart activity from electrical signals in the chest.)

The newspaper has overstated the implications of this well-conducted study commissioned by the NHS. Researchers found that using an ECG did not offer much additional benefit when used alongside a thorough assessment by a doctor, but only in the diagnosis of suspected angina in non-emergency situations.

ECGs remain an invaluable tool in emergency settings for investigating signs of heart ischaemia (lack of oxygen) or heart attack, in addition to a number of other rhythm and conduction abnormalities.

Where did the story come from?

This research was conducted by Doctors Neha Sekhri, Gene Feder and colleagues from Newham University Hospital, University College London, the University of Bristol and Barts and London Queen Mary’s School of Medicine and Dentistry. The study was funded by the National Health Service’s research and development programme, service delivery and organisation. The study was published in the peer-reviewed British Medical Journal.


What kind of scientific study was this?

This was a diagnostic cohort study where researchers were investigating the usefulness of the ECG when added to clinical questioning of patients who attend chest pain clinics with suspected angina.


Most chest pain clinics in the UK use ECGs when diagnosing people with suspected angina. Commonly a ‘resting ECG’ (i.e. not during activity) is routinely performed when people present symptoms of angina.

Another form of the test may be given, the ‘exercise ECG’, which records the activity of the heart during exercise (usually while the patient is on a treadmill or exercise bike). The exercise ECG is routinely provided in 59% of chest pain clinics in the UK, according to the researchers.

ECGs are used to investigate whether people are at high risk of coronary heart disease or other heart events. The researchers here questioned whether the resting and exercise ECGs add anything useful to an assessment of this risk for people attending chest pain clinics with new-onset chest pain and no history of heart disease.

People who had been referred by their GP to one of six chest pain clinics in the UK between January 1996 and December 2002 were potentially eligible for this study. From these, patients were included in the study if they met the following criteria:

  • They had chest pain and were without a diagnosis of coronary artery disease.
  • They did not have missing data.
  • They were not of black or ‘unspecified’ ethnicity.

This left a total of 4,848 eligible for this study.

Patients were interviewed by a clinician and details were recorded in a database. Information included age, ethnicity, sex, symptom duration, smoking, history of high blood pressure, diabetes, pulse, systolic blood pressure and descriptions of chest pain. From this clinical data, doctors made a diagnosis of 'angina', 'non-cardiac chest pain' or 'other'.

Resting ECG results were then taken for each patient and recorded in the database as either normal or abnormal. 4,848 of the total cohort had an exercise ECG too, recorded as positive, negative or equivocal for ischaemia (i.e. uncertain). For a proportion of the total patients (1,422 of them) more detailed information was available from an exercise ECG. Patients were followed up until December 2003 and mortality records and records of hospital admissions were collected until then.

Researchers then analysed what factors predicted outcome in these patients, i.e. lifestyle factors and results on ECGs.

Through complex statistical methods, they constructed models that would tell them how well the three types of investigations predicted patient outcome, i.e. basic clinical assessment (discussion with doctor) alone; basic clinical assessment plus resting ECG; and basic clinical assessment plus both resting and exercise ECG (summary results and then detailed results).

By doing this they were investigating whether ECG provides an additional diagnostic benefit on top of the clinical discussion with doctors at presentation.


What were the results of the study?

On average, people were followed up for 2.5 years. The smaller groups who also had an exercise ECG were followed up for a slightly shorter time (about 2.2 years). Resting ECGs didn’t add any value to a basic clinical assessment in terms of being able to predict negative outcomes. Exercise ECGs added only a little.


While there was evidence that on their own, abnormal readings on ECGs were able to predict poor outcomes, they didn’t add much to prediction from the basic assessment. 47% of all events during follow-up occurred in people with a negative exercise ECG result, i.e. showing a ‘normal’ result. There were only small differences in the ability of the test to predict outcome when the researchers considered the risk-level of patients at baseline.


What interpretations did the researchers draw from these results?

Researchers conclude that ECGs add little incremental value to an assessment of risk of negative cardiac outcome through clinical assessment (i.e. recording history and symptom assessment by a clinician).


They say that these results emphasise the importance of the clinical assessment and the need for more effective means of assessing future risk for these patients. Given that a large number of events occurred in people with ‘normal’ exercise ECGs the researchers say that this emphasises the ‘limitations of using ECGs for risk assessment’.


What does the NHS Knowledge Service make of this study?

This large cohort study has shown that, on their own, abnormal ECG results can predict negative cardiovascular outcomes and death in patients with suspected angina. However, it also showed that these tests do not appear to add value towards prognosis when given alongside a thorough clinical assessment.


There are several points to keep in mind when interpreting the results:

  • It is important to highlight the fact that abnormal results on the ECGs on their own were able to predict adverse outcomes, so the study findings do not suggest that these tests do not work, only that, for people with suspected angina, they add little to a thorough clinical assessment.
  • Predicting long-term risk of cardiovascular outcome is not a straightforward task. The accompanying editorial to this article suggests the results from this study highlight the importance of ensuring that the clinical assessment is thorough in the first place.
  • This study was carried out in a select group of patients: those referred by their GPs to chest pain clinics with suspected angina. The results don’t therefore apply to the use of ECG in other settings such as emergency settings or in patients with other presentations.

Patients with suspected angina do not always demonstrate changes on either resting or exercise ECG and therefore thorough clinical assessment on an individual basis is vital when making, or excluding, a diagnosis of angina.

However, the ECG remains an essential investigation for demonstrating signs of heart ischaemia (lack of oxygen, as may be seen during an angina episode) or infarction (heart attack), in addition to a number of other heart rhythm and conduction abnormalities, and should therefore be used in all presentations that are suggestive of chest pain to alert to need for monitoring or urgent treatment.

Analysis by Bazian

Edited by NHS Choices


How helpful is this page?

Average rating

Based on 0 ratings

All ratings

0  ratings
0  ratings
0  ratings
0  ratings
0  ratings

Add your rating

Is your blood pressure too high?

At least 25% of adults have high blood pressure. Get tips on how to keep healthy

Salt: the facts

Most people in the UK eat too much salt. But cutting back is possible, if you know how