Journal of the American College of Cardiology

2012 by the American College of Cardiology Foundation

Published by Elsevier Inc.

Vol. 59, No. 17, 2012

ISSN 0735-1097/$36.00

CORRESPONDENCE

Letters to the Editor

Dichotomizing High-Sensitivity
Cardiac Troponin T Results
and Important Analytical
Considerations
We read with great interest the paper by Body et al. (1) that
presented intriguing data on a possible early rule-out protocol for
acute myocardial infarction in patients presenting with undetectable cardiac troponin T (cTnT) measured with the high-sensitivity
assay ([hs-cTnT], undetectable 3 ng/l). Using this approach,
nearly one-quarter of the population could be ruled out for acute
myocardial infarction at presentation with, the investigators contend, possibly no need for serial troponin testing in this group (1).
However, at these lower concentrations, there are other important
factors, in addition to imprecision, that need to be considered.
First, there is no standardization in reporting hs-cTnT results,
with different laboratories using different cutoffs and lower limits
for reporting (e.g., limit of the blank [LoB] 3 ng/l, which Body
et al. used versus limit of detection 5 ng/l) (2). It would be
interesting to see how this rule-out algorithm would change if the
higher limit of detection of 5 ng/l were used as opposed to the LoB
of 3 ng/l.
Second, the presence of hemolysis will also negatively affect
the cTnT concentrations, with larger concentrations of hemoglobin resulting in proportionally lower cTnT concentrations
(3). This is not a trivial problem, as a recent audit in our
hospital over 1 month identified over 10% (217 of 2,085) of
cTnT results were on hemolyzed samples. Thus, for hemolyzed
specimens whose reported hs-cTnT concentrations are 3 ng/l,
a repeat specimen may be required to confirm that the cTnT
concentration is truly below the LoB and not due to this
negative interferent.
Third, this approach of using the LoB to rule out acute
myocardial infarction may not be transferable to other assays that
are perhaps more analytically sensitive. For example, in a cohort of
chest pain patients presenting early to an emergency department,
more than 25% of these patients had hs-cTnT concentrations
below the LoB; however, in contrast, all of these patients had
detectable cardiac troponin I concentrations as measured with an
investigational-use high-sensitivity cardiac troponin I assay (4). It
is clear from the work of Body et al. (1) that changes in how we use
and report high-sensitivity cardiac troponin assays are required to
achieve the most optimum care for patients presenting with chest
pain. However, the true benefit for early rule-out may be less than
reported after including other important outcomes requiring hospital admission (i.e., all of the major adverse cardiac events) and
after considering important analytical issues.
*Peter A. Kavsak, PhD
Andrew Worster, MD

Downloaded From: http://content.onlinejacc.org/ on 01/16/2015

*Juravinski Hospital and Cancer Centre (Core Lab Section)

711 Concession Street
Hamilton, Ontario L8V 1C3
Canada
E-mail: kavsakp@mcmaster.ca
doi:10.1016/j.jacc.2011.10.904
Please note: Dr. Kavsak has received grants/honorarium from Abbott, Beckman
Coulter, Ortho Clinical Diagnostics, Randox, and Roche.
REFERENCES

1. Body R, Carley S, McDowell G, et al. Rapid exclusion of acute

myocardial infarction in patients with undetectable troponin using a
high-sensitivity assay. J Am Coll Cardiol 2011;58:13329.
2. Saenger AK, Beyrau R, Braun S, et al. Multicenter analytical
evaluation of a high-sensitivity troponin T assay. Clin Chim Acta
2011;412:748 54.
3. Li A, Brattsand G. Stability of serum samples and hemolysis interference on the high sensitivity troponin T assay. Clin Chem Lab Med
2011;49:335 6.
4. Kavsak PA, Hill SA, Bhanich Supapol W, Devereaux PJ, Worster A.
Biomarkers for predicting serious cardiac outcomes at 72 hours in
patients presenting early after chest pain onset with symptoms of acute
coronary syndromes. Clin Chem 2012;58:298 302.

The D-Dimer Approach for

Troponin in the Diagnosis of
Myocardial Infarction
Is it Really Useful?
We read with interest the paper by Body et al. (1). In a cohort of
703 individuals with acute chest pain, Body et al. demonstrated
that the use of a high-sensitivity troponin assay coupled with a low
cutoff point (any detectable level) yields 100% sensitivity for
recognizing myocardial infarction, leading to a perfect negative
predictive value. Thus, the investigators concluded, this strategy
could be used to reduce unnecessary hospital admissions. We
intend to further analyze the implications of this approach.
First, in order to reduce the number of patients unnecessarily
admitted to the hospital, a test should have an improved ability to
recognize healthy individuals that can be safely discharged. The ability
to recognize healthy people is defined as specificity. As one lowers
the cutoff point of a test, sensitivity improves at the cost of specificity.
In fact, along with the increase in sensitivity from 85% to 100%, the
investigators reported a decrease in specificity from 82% to 34% as the
lower cutoff was adopted. Because a smaller number of healthy people
will be identified, it is highly questionable if this approach really
reduces unnecessary admissions. Even if different cutoff points were
adopted to rule out and rule in infarction, a gray zone of confusion
would be created, leading to considerable doubt if this strategy would
be useful in clinical practice.

Correspondence

JACC Vol. 59, No. 17, 2012

April 24, 2012:15703

Second, if the lower cutoff point is used only to rule out infarction,
how many patients will be discharged by this new approach? Body et
al. showed that 28% of the cohort had negative troponin, ensuring no
infarction. However, the study did not report how many of those were
really discharged. From those patients, some could have very typical
chest pain, characterizing unstable angina; some could have ischemic
electrocardiogram changes; and others could have other serious causes
of chest pain that prevented discharge. Therefore, a negative troponin
does not necessarily mean discharge and the actual number of patients
in which the troponin result helped the decision is not clear in the
paper.
The universal definition of infarction takes the 99th percentile
of troponin as the cutoff point (2), providing good diagnostic
accuracy (85% sensitivity and 82% specificity) (1). Before trading
this accuracy for a higher sensitivity at the expense of specificity
(so-called D-dimer approach), clinical evidence should demonstrate a real advantage over the traditional way of troponin
interpretation. The definitive level of evidence will be ideally
provided by randomized clinical trials comparing the 2 strategies.
*Luis C. L. Correia, MD, PhD
Marcia Noya-Rabelo, MD, MSc
*Medical School of Bahia
Avenue Princesa Leopoldina, 19/402
40.150-080, Salvador, Bahia
Brazil
E-mail: lccorreia@terra.com.br
doi:10.1016/j.jacc.2011.10.905

Figure 1

Algorithm for the Diagnosis of AMI

AMI acute myocardial infarction; hs-cTnT high-sensitivity cardiac troponin T.

Downloaded From: http://content.onlinejacc.org/ on 01/16/2015

1571

REFERENCES

1. Body R, Carley S, McDowell G, et al. Rapid exclusion of acute

myocardial infarction in patients with undetectable troponin using a
high-sensitivity assay. J Am Coll Cardiol 2011;58:13329.
2. Thygesen K, Alpert JS, White HD, et al., for the Joint ESC/ACCF/
AHA/WHF Task Force. Universal definition of myocardial infarction.
Circulation 2007;116:2634 53.

Reply
We thank Drs. Kavsak and Worster and Drs. Correia and
Noya-Rabelo for their interest in our paper (1). We understand the
concerns of Drs. Correia and Noya about the specificity of our
proposed use of high-sensitivity cardiac troponin T (hs-cTnT) (1).
This approach may have more in common with the use of brain
natriuretic peptide than D-dimer. Some levels of brain natriuretic
peptide are diagnostic of either the presence of heart failure or its
absence (2). Other patients have levels that are not sufficiently high or
low enough to be diagnostic and require additional testing. This is the
case with hs-cTnT. Some patients will have acute myocardial infarction (AMI) ruled out immediately, whereas others still require serial
testing to establish a diagnosis, as per current practice (Fig. 1). This
will maintain overall specificity while avoiding serial testing in a large
group, reducing emergency department crowding and its negative
consequences while facilitating accurate patient evaluation and care.
Not all patients with an initial hs-cTnT 3 ng/l will be eligible
for early discharge. We strongly believe that hs-cTnT is an adjunct
to care and not a substitute for clinical evaluation. For example, 22
(11.3%) of the patients with hs-cTnT 3 ng/l had ischemic