Which patients with chest pain need a troponin test?

There are two easy ways to get troponin testing wrong. One is to send it automatically because somebody has chest pain. The other is to decide against it because the discomfort sounds like indigestion, the patient has no conventional risk factors and, frankly, they look far too well.

The actual evidence is rather unkind to many of those reassuring instincts.

In this podcast, Professor Rick Body examines how much we can trust the pain description, risk factors and clinical judgement when deciding who needs a troponin.

Listen and record your CPD with MedPod Learn

How much does the pain description help?

Rick’s original prospective study of symptoms and signs in suspected acute coronary syndrome (ACS) recruited 796 ED patients with suspected cardiac chest pain. Acute myocardial infarction was diagnosed in 148 of them, giving a pre-test probability of 18.6%, rounded to 19% in the podcast.¹ It’s important to look carefully at who was studied. These were not 796 unselected people who happened to mention chest discomfort at triage; clinicians had already decided that myocardial infarction was plausible enough to investigate.

Within that group, some of the descriptions we tend to regard as “cardiac” were surprisingly unhelpful. Pain described as heavy or crushing shifted the probability of MI only from around 19% to 22%. Even more awkwardly, describing the discomfort as indigestion-like produced essentially the same probability.¹ Rick has previously gone through these figures on St Emlyn’s in his discussion of who we should investigate for acute coronary syndrome, and they are a useful antidote to the idea that the patient’s choice of adjective can safely determine whether we send a troponin.

Some findings were considerably more useful in the other direction. Vomiting was associated with an MI probability of about 41% in this cohort, while patients observed to be sweating in the ED had a probability of about 59%. Pain radiating to the right arm or both arms also increased the likelihood.¹

Sweating and vomiting should make us more concerned. Their absence does not offer the same reassurance. “Not sweaty, not vomiting and the pain isn’t crushing” is not a rule-out strategy.

That fits with a point we have made before at St Emlyn’s when considering which patients really need an ECG on arrival. History, ECG and biomarkers all contribute information, but none should be treated as an isolated binary answer to the question of whether the presentation represents ACS.

Does “typical” chest pain help?

We often defend ourselves against the weakness of individual symptoms by saying that nobody practises medicine one symptom at a time. We recognise patterns. Perhaps heavy pain on its own is weak, but surely an experienced clinician can listen to the whole story and decide whether it sounds typical for ACS?

Unfortunately, “typical” and “atypical” do not rescue us as much as we might hope.

In a prospective UK study of 912 patients with chest pain and a non-diagnostic ECG, clinicians classified the presentation as typical or atypical for ACS. The ability of that judgement to discriminate patients with and without acute myocardial infarction was barely better than chance, with an area under the ROC curve of 0.54. More experienced clinicians were somewhat better at using typicality to rule disease in, but overall discrimination remained poor.² We discussed that paper previously in our St Emlyn’s journal club on typical chest pain symptoms.

The broader Asia-Pacific work from Greenslade and colleagues reached a similar uncomfortable conclusion. Symptoms differed between populations, but most individual symptoms had limited diagnostic value for ACS. In other words, the problem is not solved by simply assembling several weak descriptors and calling the resulting pattern “typical”.³

The history still helps us understand what happened, identify high-risk features and look for alternative diagnoses. What it cannot justify is using “atypical” as a reason to stop thinking about ACS.

What about cardiovascular risk factors?

Hypertension, smoking, diabetes and hyperlipidaemia influence the longer-term risk of coronary artery disease. But once a patient presents with symptoms that could be MI, can the absence of those risk factors tell us that today’s episode is not an infarction?

Rick and colleagues studied 796 patients with suspected cardiac chest pain. Among those with none of the five conventional risk factors recorded, 12.2% had an acute MI. Among people with four or five, the figure was 21.3%. The overall discrimination of risk-factor count was extremely poor, with an area under the ROC curve of 0.49. The original Resuscitation paper therefore concluded that traditional chronic coronary risk factors were not useful for confirming or excluding AMI in this setting.⁴

That finding has since been revisited in the high-sensitivity troponin era. In a 14-hospital English cohort of 1,273 patients with suspected cardiac chest pain, 14.1% had an adjudicated MI. The probability was still 9.7% in patients with no recorded conventional risk factors and 23.5% in those with four or five.⁵

Risk factors still belong in the assessment, including in validated decision aids. Their absence alone is a poor reason to withhold investigation when MI remains plausible.

Surely clinical judgement can sort this out?

Clinical gestalt fares rather better, although not quite well enough to do the job unaided.

In the multicentre validation study discussed by Rick, clinicians recorded whether they thought ACS was definitely not, probably not, possible, probable or definite. Their judgement did track with disease. The overall C-statistic was 0.75, so clinicians were extracting useful information from the presentation.⁶

Among the 60 patients clinicians judged to be “definitely not” ACS, three met the study’s ACS endpoint: MI or a major adverse cardiac event within 30 days. That is a negative predictive value of 95% in the supposedly safest group.⁶

At the opposite end, calling a presentation “definitely” ACS produced a positive predictive value of 71.2%. The multicentre gestalt validation study concluded that clinical judgement alone was not sufficiently safe either to rule ACS in or to rule it out.⁶

That feels closer to everyday practice than saying clinical judgement is useless. It isn’t. Gestalt changes probability and tells us who concerns us more. It simply does not provide the certainty we sometimes attribute to it.

So who should actually have a troponin?

There is a temptation at this point to respond to all this uncertainty by measuring troponin in everybody. Rick does not argue for that, and neither does current guidance.

NICE specifically advises against using high-sensitivity troponin in people in whom ACS is not suspected. When ACS is suspected, troponin should be used within an appropriate validated diagnostic pathway, taking account of the clinical presentation, ECG, timing of symptoms and the possibility of chronically elevated troponin.⁷ ESC guidance similarly recommends high-sensitivity troponin and rapid rule-in/rule-out algorithms for patients with suspected NSTE-ACS.⁸

A sensible approach is therefore:

1. Assess whether ACS is plausible. Use the history, examination and ECG. Look for a coherent alternative explanation; a single reassuring descriptor does not provide one.
2. Do not rule out MI on reassuring labels alone. “Atypical”, mild or indigestion-like discomfort, or the absence of conventional risk factors, is not a sufficient reason to withhold investigation.
3. Avoid indiscriminate testing. If another diagnosis convincingly explains the presentation and ACS is no longer suspected, do not add troponin as a screening test for ACS.
4. Use your validated local pathway. If MI remains a plausible explanation for the symptoms, investigate using your local high-sensitivity troponin pathway. The strategy depends on the assay, symptom timing and protocol.

There is one final complication. A positive troponin is not itself a diagnosis of myocardial infarction. The newly published Fifth Universal Definition of Myocardial Infarction retains the distinction between myocardial injury and infarction: MI requires evidence of acute myocardial injury together with evidence that the injury is ischaemic.⁹ Rick has recently written about the changes in detail in his St Emlyn’s guide to the Fifth Universal Definition and you can hear more about this on Episode 297 of the podcast.

The patient whose discomfort is mild, transient or described as indigestion may genuinely have something benign. The job is to establish why we believe that, rather than allowing the description itself to become the diagnosis. If MI remains a reasonable explanation after that assessment, the evidence supports moving on to a validated troponin strategy rather than trying to talk ourselves out of the test.

And if you want to hear Rick dismantle each of our favourite reassuring features one by one, the podcast is probably more entertaining than I have made it sound.

About Rick Body

Professor Rick Body is Professor of Emergency Medicine at the University of Manchester and an Honorary Consultant in Emergency Medicine. His research focuses particularly on the early diagnosis of acute coronary syndromes and high-sensitivity cardiac troponin; he is also Chair of the NICE Interventional Procedures Advisory Committee and, since January 2026, Editor-in-Chief of the Emergency Medicine Journal. He is a Senior Editor at St Emlyn’s.

References

1. Body R, Carley S, Wibberley C, et al. The value of symptoms and signs in the emergent diagnosis of acute coronary syndromes. Resuscitation. 2010;81:281-286.
2. Carlton EW, Than M, Cullen L, et al. ‘Chest pain typicality’ in suspected acute coronary syndromes and the impact of clinical experience. American Journal of Medicine. 2015;128:1109-1116.e2.
3. Greenslade JH, Cullen L, Parsonage WA, et al. Examining the signs and symptoms experienced by individuals with suspected acute coronary syndrome in the Asia-Pacific region. Annals of Emergency Medicine. 2012;60:777-785.e3.
4. Body R, McDowell G, Carley S, Mackway-Jones K. Do risk factors for chronic coronary heart disease help diagnose acute myocardial infarction in the Emergency Department? Resuscitation. 2008;79:41-45.
5. Bali T, Boda S, Reynard C, Body R. The association between cardiac risk factors and the probability of acute myocardial infarction in the emergency department. European Journal of Emergency Medicine. 2021;28:233-237.
6. Oliver G, Reynard C, Morris N, Body R. Can Emergency Physician Gestalt “Rule In” or “Rule Out” Acute Coronary Syndrome? Academic Emergency Medicine. 2020;27:24-30.
7. NICE CG95: Recent-onset chest pain of suspected cardiac origin.
8. 2023 ESC Guidelines for the management of acute coronary syndromes.
9. Mills NL, Newby LK, Zaman S, et al. Fifth Universal Definition of Myocardial Infarction (2026). JACC. 2026;88:1314-1368.

Learning from podcasts?

If podcasts form part of your CPD, MedPod Learn helps you turn listening into documented learning. You can use it across all your podcasts, not just St Emlyn’s, to record your listening, make notes, answer questions, reflect on what you have learned and create a record of your CPD.