Why hyper-troponin-aemia does not always equal acute myocardial infarction

Myocardial infarction ECG

Can you remember the time when it took several days to rule out a myocardial infarction? Do you remember running serial CK, LDH and ALT? Did you run this in patients whose history was a carbon copy of the textbook, while anyone with anything other than central crushing chest pain radiating to the left arm in association with diaphoresis, vomiting and 5 cardiac risk factors was reassured providing that their ECG didn’t show diagnostic ST elevation? You know, the days when CK was considered an ‘early marker’ of myocardial necrosis because levels rose within the first 24 hours? If not, I guess you’re younger than me. I can remember those days, even if I was a mere medical student at the time.

Cardiac troponin has changed the world since then. Over the past 13 years since troponin was incorporated into international guidelines, we’ve gotten pretty used to being able to ‘rule out’ acute myocardial infarction (AMI) just 12 hours after the patient’s peak symptoms. Cardiac troponin assays have become so cardiac specific that we’ve not had to worry about running ‘screening’ tests for AMI in the ED. As we increasingly recognise that patients with AMI more often than not present with ‘atypical’ symptoms and with increasingly litigious culture, we’ve become used to running troponin tests for all and sundry, safe in the knowledge that patients without AMI will have undetectable troponins thanks to the wonderfully high specificity of the assay.

Of course, in 2009 when the first high sensitivity troponin assay was reported, we welcomed the introduction of this novel technology. Now, we could potentially bring forward the point at which we can confidently ‘rule out’ AMI to as early as 3 hours after presentation. For the first time, we can detect troponin levels in apparently healthy individuals. In fact, a high sensitivity assay can detect circulating levels in over half of apparently healthy individuals by definition. As an emergency physician with an understanding that lowering a diagnostic cut-off will always act to increase sensitivity and lower specificity, I was particularly interested in looking at cut-offs below the 99th percentile. Maybe accepting a lower cut-off would enable AMI to be ‘ruled out’ using a single test at the time of presentation – and the early evidence does look fairly promising.

What are the benefits of high sensitivity troponin?

The advantages of these new assays is firstly in their improved precision and analytical sensitivity. Improved precision means that, if we tested the same sample a number of times, the results would be closer together. This is similar to the concept of inter-observer reliability, which I’m sure we’re all familiar with. Improved analytical sensitivity means that they can detect (and quantify) smaller concentrations of troponin than was previously possible, thus meaning that we can find troponin in the blood of people who are (or seem to be) entirely healthy. This means that the assays are unequivocally better from the laboratory’s perspective. But what does it mean for the clinician?

The levels that we ‘see’ in the lab report are higher than they used to be, i.e. a standard troponin T result may have been 0.01ng/ml (10ng/L) but the high sensitivity assay will read at approximately 0.035ng/ml (35ng/L). This actually means that we diagnose more AMIs. More patients with acute coronary syndromes will have positive troponins – i.e. we diagnose more NSTEMI and less unstable angina (which is a troponin negative state). This is great because unstable angina is a difficult diagnosis to make and it requires further investigation even after troponin testing. Now, that problem is reduced – more patients are identified as being ‘at risk’ using the troponin alone. Once we’ve identified them, we can treat them.

Chronic troponin elevations

Of course, the levels of patients without AMI are also higher. We always knew that conditions like heart failure and renal impairment were associated with chronic troponin elevations. With high sensitivity assays, even more of these patients will have troponin elevations. What’s more, we’ll see patients with advanced age and cardiovascular risk factors who have troponin levels above the 99th percentile even when they’re apparently healthy. Thus, we are likely to see many patients in our practice who have a troponin level that is above the 99th percentile but that is not actually abnormal for them. How do we know? Well, until we reach the stage where routine troponin testing takes place in primary care, we will only know by serial testing. A chronic troponin elevation will not change (more than would reasonably be expected given the precision of the assay), whereas the troponin level will rise and/or fall in AMI. How much is a rise and/or fall? For troponin T, probably ≥10ng/L is the best answer we have – but that’s a topic for later.
What do you do with a chronic troponin elevation? Not a lot. You recognise that this is an adverse prognostic marker and you consider what you can do to optimise primary/secondary prevention. You also consider the true diagnosis. Is this unstable angina? (Patients with chronic troponin elevations can get this just as much, if not more, than those with entirely ‘normal’ troponin levels). Is it another important diagnosis? But you don’t treat the troponin – and you don’t need an angiogram, unless there is another clinical reason to do so.

Type 2 myocardial infarction

The other common reason for non-AMI troponin elevation in the Emergency Department is type 2 myocardial infarction. These patients have had AMI. They have a rise and/or fall of troponin on serial testing. Biochemically, we can’t differentiate them from patients with type 1 AMI caused by plaque rupture, erosion or fissure. However, the aetiology is entirely different. Type 2 AMI is caused by an imbalance of supply and demand in the coronary circulation. There are many causes. The key to differentiating type 1 from type 2 AMI is the clinical context. With high sensitivity troponin assays, we will of course diagnose more type 2 AMIs. This means that we really do have to think carefully when we see a positive troponin result. Just as the troponin result can’t tell us whether the patient has unstable angina, it also can’t tell us whether an AMI is type 1 or type 2. Only a clinician can do that!
Type 2 myocardial infarction is caused by an underlying condition, e.g. sepsis, gastrointestinal haemorrhage, a dysthyrhmia, etc. It can occur in patients with or without coronary disease, although it is of course going to be more common in those who have pre-existing coronary disease because it will take less to decompensate the supply-demand balance in those patients, who have stenotic coronary heart disease. Particularly if the AMI has been precipitated by a relatively mild insult, these patients may still warrant cardiac investigation and management of underlying coronary heart disease. However, the benefit of antiplatelet and antithrombotic treatment remains unproven in this context. What is for sure is that the underlying condition should be treated.
So there you have it. A brief rundown of why hypertroponinaemia does not always equal AMI. Even when it does, we must be careful to differentiate between type 1 and type 2 AMI, as the management will be very different in each context. Not all patients with a high troponin need to see a cardiologist. In fact, if we refer every patient with hypertroponinaemia to the cardiologists, even those with genuine disease will suffer as the condition will be greeted with an inevitable cynicism if the prevalence of significant disease among those referred is low. In a nutshell, here’s what to do with a raised troponin:

(a) Repeat the troponin level and assess for a rise and/or fall.

(b) If there is no rise and/or fall, this is not AMI. Think about the patient: is this someone you expect to have a chronic troponin elevation or is this totally unexpected?

(c) If the troponin is elevated and there is a rise and/or fall on serial testing, think about the clinical context. Is there an underlying condition causing a supply-demand imbalance? If so, this is most likely a type 2 AMI. Treat the underlying condition first, then consider whether you need to investigate for underlying coronary artery disease.

(d) If the history is compatible with an acute coronary syndrome, there is no supply-demand imbalance and the troponin is elevated with a rise and/or fall, then this is most likely to be a type 1 AMI. These patients benefit from all the proven treatments for acute coronary syndrome.




  1. Edd Carlton

    Great post Rick, am constantly battling in Poole because a “negative” trop is deemed <30ng/l for our assay as suggested by the manufacturer for our hsTrop assay. Have introduced a policy that no patient should be discharged with a single trop greater than the 99th centile value. Not even the cardiologists seem to know what to do with patients with values between 14-30!

    1. richardbody (Post author)

      Hi Edd, Thanks for the comment. Great point. There does seem to be some resistance to adopting the 99th percentile cut-off (i.e. 14ng/L for hs-cTnT) as per the Universal Definition of Myocardial Infarction. I guess some people believe there will be so many false positives that you can’t possibly glean any useful information at this level. Unfortunately this is throwing the baby out with the bathwater because (used well) you can get loads of extremely useful information at this level. Nick Mills showed very nicely that adopting the 99th percentile cut-off leads to better outcomes for patients.

      I don’t think it’s the manufacturer (Roche) that suggests adopting a cut-off above the 99th percentile – I’m sure it’s a local thing. The precision of the assay is fine at the 99th percentile. By adopting 30ng/L they’re pretty much losing all the advantages of a ‘high sensitivity’ troponin assay – and with this way of using it, they may as well be running the previous generation (non high sensitivity) assay.

      See you soon, Edd!

  2. T

    How many hours post chest pain, does a negative troponin mean no MI? I always thought 12, now im hearing 6.Can you clarify? Does it vary between trop I And trop t?

    1. richardbody (Post author)

      Hi T,

      Thanks for the comment. First, to answer the question about troponin I or troponin T, it’s a bit more complicated than that. Not all troponin ‘I’s or ‘T’s are the same. There are lots of troponin I assays and their performance really varies. They vary from high sensitivity (Abbott ARCHITECT hs – not marketed yet) to sensitive (Siemens Ultra, which is almost a high sensitivity assay) to not particularly sensitive at all (no examples given for risk of offending people!!) There are even ‘super sensitive’ assays around now (e.g. Singulex Erenna) but these are currently for research use only. The Singulex Erenna uses single molecule counting technology! The more sensitive an assay is from an analytical point of view, the greater the chance that it will have a higher diagnostic sensitivity at an earlier time point. With troponin T, there are only two assays around: the Roche high sensitivity troponin T assay and the Radiometer assay, which is a point of care test, albeit a pretty good one. In general, most point of care tests will be less sensitive than laboratory-based assays. However, you have to appraise the evidence for each assay individually in order to know how it really works – because they are all very different.

      Now, on to your question about 6 hours versus 12 hours. This is a big question – and I’ll hopefully write a whole blog post on just this topic soon. First, we need to know what our reference point is. This is absolutely vital.

      Do we start the clock:

      (a) When the patient arrives in the hospital, or
      (b) When the patient says their peak symptoms occurred?

      In the UK, we start the clock at (b). Elsewhere, the clock starts at (a). Some people suggest that (b) is unfeasible because patients can’t be relied on. I’d counter that because I very rarely find a patient who can’t tell me when the symptoms occurred. If they were woken from sleep by pain, we just ask what time they were woken. Sometimes we might not get a precise time (to the minute!) but patients usually know whether the symptoms peaked an hour ago or 24 hours ago! Anyhow, either approach is totally acceptable according to international guidelines – but we should still recognise the differences.

      If the clock starts on arrival, a test 6 hours later will be an average of 9 or 10 hours after peak symptoms – as the median time to presentation is usually 3 to 4 hours. If the patient’s peak symptoms were >12h ago, the 6-hour test will be taken >18 hours after peak symptoms.

      Why is this important? Well, because the reference standard in many of the large recent studies is a troponin test 6 hours after arrival. That was the case in the APACE study, for example, which formed the basis for the 2009 NEJM paper by Keller et al. In that study, a sensitive troponin I test 3 hours after arrival (Siemens Ultra) was 100% sensitive for AMI (compared to a test 6 hours after arrival using a non-sensitive assay). However, when stratified by time from symptom onset, the sensitivity was only 95.5% at 6 to 12 hours from symptom onset, i.e. you miss 4.5% of AMIs.

      How do you reconcile those two figures? The only way I can think of that you can get those two findings in the same study is if the 6-hour reference standard was missing some patients.

      In the UK, our reference standard is a troponin test at least 12 hours from peak symptoms. The burden of proof is on the side of the change. There is no direct evidence (as far as I’m aware) from large and rigorous studies to demonstrate that a test 6 hours after symptom onset is just as good as a test 12 hours after symptom onset.

      Therefore, my advice is to follow the NICE guidance and keep the reference standard as a test 10-12 hours from symptom onset. We might be able to shorten that – but we should get some robust direct evidence first.

      I should warn you that my take is a little controversial. There are others in this field who take alternative views. Please do me a favour though. Be cheeky and ask them for the direct evidence. Then share it with me if they have some!

      Thanks again for the comment. Apologies for a long reply,


  3. Prof Montage (@ProfMontage)

    Thanks Rick, I agree entirely!
    See my video at http://ow.ly/ka4Ht

    1. Simon Carley

      Like the videos Prof!


    2. richardbody (Post author)

      Love the videos! Thanks for sharing!

  4. Pingback: High sensitivity troponin

  5. Pingback: SMACC Gold: Cullen v Body

  6. Pingback: acute myocardial infarction lab tests | My Blog

  7. Pingback: acute myocardial infarction nejm | My Blog

  8. Pingback: acute myocardial infarction laboratory test | My Blog

  9. Pingback: acute myocardial infarction nos | My Blog

  10. Pingback: acute myocardial infarction uk | My Blog

  11. Pingback: Cardiology Case 04: An unusual ECG - St.Emlyn's

Thanks so much for following. Viva la #FOAMed

Translate »