High Sensitivity Troponin T and Acute Myocardial Infarction: One and Done?

High Sensitivity Troponin T and Acute Myocardial Infarction: One and Done?



There has been a lot of publicity about evaluation of chest pain patients in the emergency department (ED) with high sensitivity troponin testing. In the past with older troponin assays, clinicians would evaluate patients, get an ECG, and an initial set of cardiac biomarkers. The subsequent set of biomarkers would be performed at 6-8 hours later before determination of disposition. In the past few years, several studies have been published evaluating point of care troponins,  sensitive troponins, and high sensitivity troponins which have changed our practice and evaluation of these patients.  An early version of a study was recently released in the Journal of the American College of Cardiology (JACC) stating that for ED chest pain patients, we may be able to discharge patients from the ED with an initial normal ECG and single high sensitivity troponin T (hs-cTnT). So is it true… one and done?

Google Hangout on Air with Dr. Holzmann

Citation: High Sensitivity Troponin T study1

What they did:

  • Retrospective study of 14,636 patients age ≥ 25 years presenting to an ED with chest pain, with at least one hs-cTnT and ECG over two year period in Stockholm, Sweden
  • Compared patients with hs-cTnT of < 5 ng/L, 5 to 14 ng/L, and > 14 ng/L

Primary and Secondary Outcomes:

  • Primary Outcomes: Fatal or non-fatal type 1 MI within 30 days of ED visit
  • Secondary Outcomes: MI within 180 days and 365 days after the ED visit
  • Also calculated hazard ratio for all-cause mortality in patients with first hs-cTnT of < 5 ng/L for admitted versus directly discharged from ED patients


  • Myocardial InfarctionMean age of 55 years old
  • 61% (8,883 patients) of patients had initial hs-cTnT < 5 ng/L
    • 39 had diagnosis of MI and 2 died at 30 days
      • 15  of these 39 had no ECG changes
  • 21% of patients had initial hs-cTnT between 5 – 14 ng/L
  • 18% of patients had initial hs-cTnT > 14 ng/L
  • Hospitalization rate of 21% in patients with initial hs-cTnT < 5 ng/L
    • 90% of these patients had a second hs-cTnT < 5 ng/L
    • 10% of these patients had a second hs-cTnT of between 5 – 14 ng/L or > 14 ng/L
    • 14% of patients were discharged home with diagnosis of MI
  • If initial hs-cTnT < 5 ng/L AND ECG with no signs of ischemia:
    • NPV for MI and death at 30 days 99.8% and 100%, respectively
    • Absolute risk for MI 0.17%
    • No significant difference in the risk of death within 365 days between patients discharged directly form the ED vs admitted to hospital (HR 0.73 with CI 0.48 – 1.12)


  • Retrospective study
  • No external validation study

Author’s Conclusion: Patients with chief complaint of chest pain presenting to the ED with an initial hs-cTnT of < 5 ng/L and no signs of ischemia on ECG have minimal risk of MI and/or death at 3o days and can be safely discharged from the ED.

My Thoughts: Sensitivity vs Specificity

15 of 8,883 (0.2%) patients had a diagnosis of MI with initial hs-cTnT of < 5 ng/dL. But on the flip side only 676 of 2,579 (26%) patients with a hs-cTnT > 14 ng/dL had a diagnosis of MI at 30 days. Another way of stating this is the significance of a positive test is significantly reduced with higher sensitivity troponins (i.e. for what we gain in sensitivity we are giving up in specificity).

From a practical standpoint, would you then admit all your patients with indeterminant hs-cTnT’s (i.e. ≥ 5 ng/dL), knowing that the majority of these patients don’t have an AMI? The Associated Press recently published:

Dr. Allan Jaffe, a cardiologist at the Mayo Clinic, said the problem is not what the test rules out, but what it might falsely rule in. It’s so sensitive that it can pick up troponin from heart failure and other problems and cause unnecessary tests for that.

I look forward to seeing additional followup studies looking at hs-cTnT utility and generalizability. 

For more thoughts, also check out:

Bandstein N, Ljung R, Johansson M, Holzmann M. Undetectable high-sensitivity cardiac troponin T level in the emergency department and risk of myocardial infarction. J Am Coll Cardiol. 2014;63(23):2569-2578. [PubMed]

Expert Peer Review

Salim does a great job reviewing this study and points out a number of weaknesses here. Whether we like it or not, hscTnT is coming to an ED near you (if it hasn’t already) and we need to understand the test and how to use it. This test is extremely sensitive. No one argues this point. Unfortunately, as with all tests, a high sensitivity comes with the cost of a poor specificity. A lot of patients will have intermediate or high levels that don’t have disease. These patients will be subjected to further testing. The further testing may simply be a second troponin but could also be CTCA or cardiac catheterization. These tests come with numerous risks and minimal benefits.

Goldstein et al demonstrated that patients who got CTCA instead of traditional care had a 4-fold risk of invasive interventions without an improvement in important outcomes [1]. deFilippi et al. showed the same results when comparing stress test to cardiac catheterization in 2001. The group that went right to cath had an 11% revascularization rate as compared to 4% in the stress group with no change in patient centered outcomes [2]. Additionally, revascularization for patients without STEMI has been shown to be without benefit by both the Cochrane group in 2010 and by Stergiopoulos et al in 2012 [3,4].

What does this study add to the discussion? One huge take home is that ECG findings should not be ignored. In the small group (n = 39) of patients with a hs-cTnT < 5 ng/L who had MIs, 24 of them had ECG changes. We must be vigilant about reviewing the ECG regardless of the type of troponin assay we are using. If you have a negative hs-cTnT and negative ECG, patients can go home but can we get there without this test? I argue that we can for many patients. Skeptics of our clinical abilities often quote the Pope article in NEJM in 2000 that we have a 2% “miss rate” for MI. In that study, 19 patients with MIs were sent home out of 11,000 ED patients with chest pain. That’s actually a miss rate of 0.18% (the same miss rate quoted in Bandstein study) [5]. Yes, we missed 2% of the 900 MI patients but that doesn’t take into account all the other patients with chest pain. I think the 2% number is misleading and leads us to marginalize our skills.

The bottom line here is that the hs-cTnT isn’t an evil test but it will become the bane of our clinical existence if we treat like a panacea to chest pain admissions. This test, as all others, must be applied to the correct group of patients. We must prognosticate before we start diagnostic workups. Worster and colleagues eloquently make this plea in their editorial in Annals in 2012 [6]. We as a specialty need to make sure that the hs-cTnT doesn’t become the d-dimer of this decade.


  1. Goldstein JA et al. A Randomized Controlled Trial of Multi-Slice Coronary Computed Tomography for Evaluation of Acute Chest Pain. JACC 2007; 49(8): 863-871.
  2. deFilippi CR et al. Randomized Comparison of a Strategy of Predischarge Coronary Angiography Versus Exercise Testing in Low-Risk Patients in a Chest Pain Unit: In-Hospital and Long-Term Outcomes. JACC 2001; 37(8): 2042-9.
  3. Hoenig MR, Aroney CN, Scott IA. Early invasive versus conservative strategies for unstable angina and non-ST elevation myocardial infarction in the stent era (Review). Cochrane Database Syst Rev 2010; 17(3)
  4. Stergiopoulos K, Brown DL. Initial Coronary Stent Implantation With Medical Therapy vs Medical Therapy Alone for Stable Coronary Artery Disease. Arch Int Med 2012; 172(4): 312-9.
  5. Pope JH et al. Missed Diagnoses of Acute Cardiac Ischemia in the Emergency Department. NEJM 2000; 342: 1163-70.
  6. Worster A, Kavask PA, Brown M. Risk Stratification in the Era of High-Sensitivity Troponin Assays. Ann Emerg Med 2012; 59(2): 126-7.
Anand Swaminathan, MD MPH
Assistant Residency Director and Assistant Professor of Emergency Medicine, Bellevue/NYU; Faculty Editor of EM Lyceum

Salim Rezaie, MD

Salim Rezaie, MD

ALiEM Associate Editor
Clinical Assistant Professor of EM and IM
University of Texas Health Science Center at San Antonio
Founder, Editor, Author of R.E.B.E.L. EM and REBEL Reviews
  • Matt Anderson

    Great review! We need more and more of these types of resources!!!

    First, a negative troponin = low-risk (0.17) which is well within my risk threshold for discharge for cardiac pathology as the underlying etiology of my patient’s presentation. Doesn’t this negate the need for ED chest pain observation units, ED chest pain pathways for stress testing or need for immediate admission for dynamic stress testing (ie these patients are low-risk, can go home and see their primary medical doctor to discuss the need for stress testing — which unless done with nuc med or echo are horrible screening tests anyways) ??? This alone would save our health care system millions. IMHO, evidence like this study ’empowers’ ED physicians to make the right choice for the patient and the system.

    Second, so 0.17% had an MI with 2 deaths. I can’t find specifically what these patients die from in the article or supplements? Looking at the supplements, the negative troponins that I believe are ‘important’ are those with subsequent STEMI. Troponins generally taken within 1h were negative (1 patient presented 18d later) and these patients ‘presented’ themselves with EKG changes consistent with a STEMI — so I’m no sure the negative result matters? The other ‘misses’ were NSTEMI, some of which received PCI — some that didn’t; thought Newmen did a review on repeat PCI (ie I don’t see anything in the supplements showing whether the PCI was the patient’s first or second or third) — wasn’t there no mortality benefit in repeat cases (ie medical optimization as good as medical optimization + PCI)?

    Third, 14% of patients with an ‘elevated’ level were diagnosed with an MI. 86% had something else — well this is a low-specificity for ACS — these patients often have a reason for their elevated troponin and need to be admitted to the hospital. Sepsis, PE, CHF, etc etc etc can cause elevations as well — which need to be treated as well?

    Fourth, I also think of troponin as a marker of ‘prognostician’. The absolute risk of death at 30d=2.56, 180d=8.38 and 1yr=13.3 which gives some prognostic value to the patient’s overall health, expected course and likely need for improved medical optimization with close follow-up and serial evaluations (ie from a primary care standpoint).

    Just a few of my thoughts. Thanks for the great review.

    • Salim R. Rezaie

      Hello Matt,
      Thank you for your comments. Very thoughtful and thorough.

      1. I would argue that in low risk patients (i.e. non-concerning history, ECG w/o ischemic changes, and negative TnI or T at 0 and 2 hours of presentation) can be discharged home without observation. This has NPV of 99% according to the ADAPT Trial. So I don’t keep these patients for observation stays in my clinical practice…which is a completely different conversation.

      2. I agree with you….subsequent STEMI should be the gold standard we are held to, but unfortunately it isn’t at least in the US.

      3. Specificity of hsTn testing is a huge issue. For example of all the positive troponins almost 30% had actual MI. This is a huge problem. Most of these patients will be admitted. I agree that most of these patients had something else, but unfortunately most of them get admitted to cardiology until MI is ruled out, but this takes up beds, time, and money.

      4. I like that it prognosticates, but from an EM standpoint, If we can get them into a clinic or cardiology follow up before 30days does it matter what happens at 1 year? Maybe it does, but I am not sure it changes my practices of admission or discharge.

      Thank you again for reading and your comments.


      • Matt Anderson


        1. Do you even need the 2h troponin now with this new literature? Also, what about moderate to high-risk patients — I don’t see that there was any exclusion criteria for ‘high-risk’ patients?

        2. Maybe STEMI isn’t the ‘gold standard’ but other than medical optimization there isn’t great evidence (ie risk-benefit analysis) for NSTEMI tx with anti-coagulation and likely medical optimization is all they need anyways. We need a PESI-like risk-stratification for NSTEMIs too b/c I would argue that medical optimization is all these patients probably need.

        3. I would say again — there’s a reason in most cases for an elevated troponin and its prognosticate value for ‘bad outcomes’ in most cases would warrant admission anyways (ie specific exceptions – marathon runner with chest pain and negative EKG, etc etc). I talk to our cardiologist/hospitalist ‘all of the time’ about elevated troponins and I triage them to most likely etiology ‘cardiac or medical’ so they receive the appropriate evaluation for their presentation. I think the point here is… a troponin does not equal ACS but isn’t elevated for a reason and given the associated absolute risk found in this study — needs to be further evaluated. Moreover, I haven’t found my patients etiology to be overly difficult to differentiate between a type I or type II NSTEMI — but with that said we have no idea if a large number of these patients were misclassified as a type II versus a type I NSTEMI in this study either — could be affected by local practice, particular cardiologists (ie some call everything a MI — others always call a MI something else). This could affect the 30% number.

        4. I think it depends on the patients risk: low, moderate or high. With each receiving the appropriate follow-up. I think it is important to look at from an inpatient standpoint as well, exs) CHF with an elevated troponin ! 3 the risk of mortality during that admission, PE with elevated troponin increased risk of mortality, etc.

  • noobmedstudent

    hi, i was just wondering if this study adjusted the NPV and absolute risk values for age, sex and other comorbidities? could this potentially be a potential confounder to the study?

    • Salim R. Rezaie


      You bring up an excellent question….remember: A Cox proportional hazard model is a popular model used in survival analysis that can be used to assess the importance of various covariates in the survival times of individuals or objects through the hazard function. In addition, the quantitative impact of these variables on important lifetime variables of interest (such as median survival) can be described. In other words it assigns the probability that a particular individual has an event by dividing the risk (actually, the hazard) for that person by the sum of all the hazards for all of the people who are at risk.

      So for this study, we want to ensure there is NO bias between the admitted group vs the not admitted group to ensure that we are NOT biasing our results based on one group having more DM, being male, being older or history of MI, etc….Hope this helps.