Tuesday, February 5, 2019

Should Troponin be a Vital Sign? Perhaps, but only if Interpreted Using Pre-test Probability.

An 80-something non-english speaking patient complained of vertigo.  He was asked multiple times about chest pain or dyspnea, but repeatedly denied any such symptoms.

His vital signs were all normal.  His exam was normal.

He had a triage ECG at time 0 (ECG-1):
Computer read, with Physician overread:
Sinus rhythm.  Possible right ventricular conduction delay.

What do you think?

When I saw this without any other information, I said it was very suspicious for a high lateral MI.  In aVL, there is a tiny QRS with 0.5 mm of STE, and there is reciprocal ST depression in inferior leads.  If you see this, you must very closely question the patient about any chest symptoms, and even if no relevant symptoms, to be certain to look for old ECG, and to do serial ECGs and troponins.

There actually was an old ECG from about 4 years prior (ECG-2), which I'm not sure anyone saw:
Here the limb leads have much higher QRS voltage, without a correspondingly larger ST-T voltage.
This means that the ST-T is out of proportion to the QRS on ECG-1

The clinicians did not see the finding on ECG-1.

The clinicians later stated they had had no suspicion of ACS, but that the faculty wanted a troponin anyway.  So troponin was ordered, and it returned some time within the next few hours (uncertain exactly when they saw it):

Troponin I = 27 ng/mL (900 x the upper limit of normal)

A 2nd ECG was recorded at 3 hours (ECG-3):
Computer, and physician overread: 
Minimal ST depression
What do you think?

The STE in aVL and STD in III and aVF is more pronounced, and now there is STD in V2 and V3.  This makes it diagnostic of a posterolateral OMI (especially with the troponin!).

A 3rd ECG was recorded at 3.5 hours (ECG-4):
Computer, and physician overread: 
Minimal ST depression
This is about the same.
This was in the chart:
"Troponin found to be elevated to 27. Patient denied chest pain on initial review of symptoms. Was now endorsing chest pain which began 30 minutes ago. Upon further questioning, he states that he has had intermittent chest pain since yesterday. Pain worsens when lying flat and improves with sitting up. Given 324 mg aspirin. Repeat ECG shows modest ST elevation in I and aVL and depression in inferior leads."

The cath lab was activated.

LM 30% ostial.
LAD 80% mid
LCx occluded mid (acute infarct lesion)
RCA 80% mid. distal stent patent.
PCI mid LCx

So this is an OMI (Occlusion Myocardial Infarction), but not a STEMI

Decreased left ventricular systolic performance, mild/moderate.
The estimated left ventricular ejection fraction is in the range of
Regional wall motion abnormality (WMA):
Regional WMA: Inferobasal (this is another name for posterior), akinetic, small.
Regional WMA: Lateral, large, hypokinetic.

The patient went into cardiogenic shock and ultimately died of this MI.


There are many, cardiologists especially, who think that Emergency Physicians order too many troponins.  Many come back "positive" and trigger many often unnecessary downstream tests.  In our UTROPIA study (NCT02060760), from which there are many publications, 85% of "positive" troponins (at least one value above the 99th %-ile upper reference limit (URL), were either type II MI or non-MI myocardial injury (acute or chronic myocardial injury).

Sandoval Y et al. Type 1 and 2 Myocardial Infarction and Myocardial Injury: Clinical Transition to High-Sensitivity Cardiac Troponin I.  American Journal of Medicine 130(12): 1431-1439, Dec. 2017.

Important: this was not due to high sensitivity troponin; the number of troponins greater than the 99% URL was lower with high sensitivity troponin I than with contemporary troponin I.

We published another study (see below) showing that, as expected, patient selection greatly influences the positive predictive value of troponin.

My opinion on this:

It is not that we necessarily order too many troponins.  However, we use the 99th percentile cutoff as a cutoff for diagnosis of type I MI, and too often assign a diagnosis of type I MI simply because there is a troponin above that URL.  It is this interpretation of the test which is often inappropriate, not necessarily the ordering of the test.  Only in the right clinical context is a slightly elevated troponin highly suspicious for type 1 MI.

However, no matter the clinical context, a troponin that is 30x, 100x, or 500x the URL is indeed most likely to be type I MI.

This figure comes from the Diamond T study (all type 1 MI were NSTEMI, not STEMI):
Notice that the 6 hour value (far right) is very low for type 2 MI.
Sandoval Y.  Nelson SE.  Smith SW.  Schulz KM.  Murakami M.  Pearce LA.  Apple FS.  Cardiac Troponin Changes to Distinguish Type 1 and Type 2 Myocardial Infarction and 180-Day Mortality Risk.  Acute Cardiovascular Care 2014;3(4):317-325.

The following study, of which I am also a co-author, could be interpreted to conclude that one should not order troponins unless there is a very high pretest probability of type 1 MI.  

I would interpret the study in the following way:

When pretest probability is low, then the troponin must be very high in order to suspect type 1 MI.  

When pretest probability is very high, the troponin need not be so high.

I have the same approach with ECGs:

If the patient has a low pretest probability and the ST elevation or T-wave is not clearly diagnostic, then don't call it OMI without further testing.  If the presentation for MI is classic, then the highly suggestive ST-T has a much higher post-test probability.

Learning Points:

1. It was wise to order a troponin on this patient who does not speak English and is very old and at high risk of MI, even though his chief complaint is not suggestive of MI.  Had his troponin come back only a bit elevated, then it would not be diagnostic of type 1 MI.  But it returned so high that type I MI is by far the most likely diagnosis.

2. Of course, recognition of the ECG findings would be ideal, but this is very difficult.


Shah A.  Sandoval Y.  Noaman A.  Sexter A.  Vaswani A.  Smith SW.  Gibbins M.  Griffiths M.  Chapman AR.  Strachan FE.  Anand A.  Denvir MA.  Adamson PE.  D'Souza MS.  Gray AJ.  McAllister DA.  Newby DE.  Apple FS.  Mills NL.  Patient selection for high-sensitivity cardiac troponin testing and the diagnosis of myocardial infarction.  BMJ 2017;359:j4788 doi: 10.1136/bmj.j4788 (Published 2017 November 07)


Objective To evaluate how selection of patients for high sensitivity cardiac troponin testing affects the diagnosis of myocardial infarction across different healthcare settings.
Design Prospective study of three independent consecutive patient populations presenting to emergency departments.
Setting Secondary and tertiary care hospitals in the United Kingdom and United States.
Participants High sensitivity cardiac troponin I concentrations were measured in 8500 consecutive patients presenting to emergency departments: unselected patients in the UK (n=1054) and two selected populations of patients in whom troponin testing was requested by the attending clinician in the UK (n=5815) and the US (n=1631). The final diagnosis of type 1 or type 2 myocardial infarction or myocardial injury was independently adjudicated.
Main outcome measures Positive predictive value of an elevated cardiac troponin concentration for a diagnosis of type 1 myocardial infarction.
Results Cardiac troponin concentrations were elevated in 13.7% (144/1054) of unselected patients, with a prevalence of 1.6% (17/1054) for type 1 myocardial infarction and a positive predictive value of 11.8% (95% confidence interval 7.0% to 18.2%). In selected patients, in whom troponin testing was guided by the attending clinician, the prevalence and positive predictive value were 14.5% (843/5815) and 59.7% (57.0% to 62.2%) in the UK and 4.2% (68/1631) and 16.4% (13.0% to 20.3%) in the US. Across both selected patient populations, the positive predictive value was highest in patients with chest pain, with ischaemia on the electrocardiogram, and with a history of ischaemic heart disease.
Conclusions When high sensitivity cardiac troponin testing is performed widely or without previous clinical assessment, elevated troponin concentrations are common and predominantly reflect myocardial injury rather than myocardial infarction. These observations highlight how selection of patients for cardiac troponin testing varies across healthcare settings and markedly influences the positive predictive value for a diagnosis of myocardial infarction.

Comment by KEN GRAUER, MD (2/5/2019):
Excellent case by Dr. Smith — both for emphasizing important clinical concepts regarding optimal use of Troponin in the ED — and — for a number of insightful subtleties on interpretation of serial tracings, with comparison to a prior tracing on the patient. I focus my comments on these ECG subtleties.
  • For clarity — I have combined in Figure-1 the first 2 ECGs done in the ED, as well as the prior tracing done on this patient.
Figure-1: The first 2 ECGs done in this case, together with a prior tracing done 4 years earlier (See text).
Clinical presentation of this case was especially challenging = an elderly, non-English speaking patient who initially denied any hint of chest pain. The prior ECG ( = TOP tracing in Figure-1) was not available at the time the patient was first seen — so initial assessment was based solely on the 1st ECG = MIDDLE tracing in Figure-1. My thoughts regarding this initial ECG done in the ED were the following:
  • As per Dr. Smith — there is a tiny QRS complex in lead aVL, with what appears to be 0.5 mm of coved ST elevation. There is an equally tiny QRS complex in lead III, with subtle-but-real scooping depression that looks to be the “mirror-image” of the subtle ST elevation in lead aVL. Similar suggestion of reciprocal change is seen in lead aVF — but not in lead II. While I thought these changes were far from definitive — I agree with the excellent point emphasized by Dr. Smith = this clearly could represent OMI (high-lateral MI?), and the case needs immediate attention.
  • Additional Thoughts about this 1st ECG: There is an RSr’ in lead V1 — with a QRS complex of normal duration. The question arises as to whether this represents incomplete RBBB? Technically, it does not — because there is no terminal s wave in lateral leads I and V6. I generally acknowledge this finding by simply writing, “RSr’ in lead V1”. Of note — transition is early (a prominent R wave is already forming by V2, and the R wave becomes predominant by lead V3). Of note — T waves are upright in all 6 chest leads. Often with complete or incomplete RBBB — there is ST-T depression in lead V1, and sometimes also in V2 and V3. Given our concern about possible subtle high-lateral OMI — this raises the question whether the upright T waves in leads V1 and V2 of this 1st ECG might be abnormal and reflect ischemia. I wasn’t certain of the answer to this from assessment of this single tracing.
  • There are tiny initial q waves in multiple leads — including leads I, II, III, aVL, aVF, and V4,5,6. I suspected these small and narrow q waves seen in multiple leads were not significant.
A Prior ECG (done 4 years earlierwas found TOP tracing in Figure-1. It is interesting to compare this prior ECG with the initial ECG done in the ED.
  • As per Dr. Smith — QRS amplitude in the limb leads was greater on the prior tracing. In addition — there is slight axis shift (ie, the QRS complex is almost entirely negative in lead III of the prior tracing — whereas it appears to be isoelectric in the 1st ECG). This is relevant — because the T wave inversion that we see isolated to lead III in the prior tracing is not necessarily abnormal (ie, T wave inversion may be normal when isolated to either lead III or aVF if the QRS complex in the same lead is predominantly negative).
  • We do not see any hint of ST coving or elevation in lead aVL in the prior ECG. That said, given the axis shift that has occurred, with different QRS morphology now seen in aVL, I wasn’t sure about the significance (if any) of this change.
  • A monophasic R wave, equal in amplitude to the thin S wave in lead V1 is seen in the prior tracing. There was no hint of the RSr’ seen in the 1st ECG. Transition was again early in the prior tracing (R=S by V2, with a predominant R wave by V3). Of note — the T wave was also positive in all 6 chest leads on this prior tracing — so this T wave positivity in all 6 chest leads is not a new finding.
  • BOTTOM LINE — I didn’t think the prior tracing clarified the situation. We were still left with potential concern regarding the changes noted above for the initial ECG.
My primary purpose in commenting on this case arises from the 2nd ECG = BOTTOM tracing in Figure-1. When assessed by comparing this 2nd ECG to the 1st ECG — definitive diagnosis of OMI can now be made!
  • In the 3 hours that passed since the 1st ECG was done — there is now clearly more coved ST elevation in lead aVL — and — more mirror-image ST depression (reciprocal changein leads III and aVF. In addition, the 3rd inferior lead ( = lead II) now shows suggestion of reciprocal ST depression.
  • In the anterior chest leads, since the 1st ECG was done — there is now J-point ST depression in leads V2 and V3 that was not previously present.
  • In the lateral chest leads — there now appears to be slight-but-real ST elevation, with beginning T wave inversion in leads V5 and V6. Although admittedly difficult to assess because of baseline artifact in leads V5 and V6 in both tracings — I do think the new ST elevation in these leads is real.
  • BOTTOM LINE  Even without the markedly increased Troponin in this case — the sequential ECG changes described above that developed between the time the 1st and 2nd ECGs were done (a time span of 3 hours) are diagnostic of acute postero-lateral OMI.
SUGGESTION  The ECG findings and sequential ECG changes described for the 3 tracings shown in Figure-1 are subtle! That said, they are perfectly consistent with the clinical history and cath findings. It is worth another look if you did not confidently diagnose acute OMI after seeing the 2nd ECG, and comparing it to the 1st ECG done 3 hours earlier.

P.S. As per Arron Pearce, instead of a sinus rhythm in the 1st ECG (MIDDLE tracing in Figure-1) — there is probably a Low Atrial Rhythm, because a clear upright P wave is lacking in lead II — the P wave appears to be negative in leads III and aVF — and, the P wave is upright in leads I and aVL. Note how sinus rhythm returns in the 2nd ECG done 3 hours later (the P wave now is upright in lead II, as well as in the other inferior leads)
  • Lack of a long lead II rhythm strip, baseline artifact and very low amplitude atrial activity all contribute to making rhythm assessment of the 1st ECG quite challenging. That said — a low atrial rhythm clearly appears to be present.
  • Clinically — the clinical significance of a transient, non-accelerated low atrial rhythm is minimal.
  • Our THANKS again to Dr. Smith for presenting this case.


  1. In such situations set the ECG device to 20 mm/mV (or even more, if available) and record again.

  2. What could be the reason of high R wave in v1 in previous ecg...

  3. Great case with excellent comments and an especially good example of an OMI without STEMI (or rather, without STE that meets STEMI criteria).

    I would be very suspicious that the first ECG had a LA/LL lead wire switch. This is the first thing that goes through my mind when I notice that Lead II does not have the largest upright P wave. While the P waves normalized with the 2nd ECG that was performed, there were also subtle changes in Lead III (the QRS in III is reversed in such an exchange and this is reflected in the first two tracings in the ED). And there are also probably subtle differences in aVL and aVF, which would have been switched in an LA/LL exchange. The sinus rate after "conversion" to sinus rhythm is also slower than the rate of the "low atrial" rhythm that would have had to usurp the sinus rhythm; in other words, a FASTER "lower atrial" rhythm was usurped by a SLOWER sinus rhythm.


Recommended Resources