Journal of the American College of Cardiology
© 2004 by the American College of Cardiology Foundation
Published by Elsevier Inc.
Vol. 43, No. 6, 2004
ISSN 0735-1097/04/$30.00
doi:10.1016/j.jacc.2003.10.036
Acute Coronary Syndromes
Implication of Different Cardiac
Troponin I Levels for Clinical Outcomes
and Prognosis of Acute Chest Pain Patients
Michael C. Kontos, MD,*†§ Rakesh Shah, MD,* Lucie M. Fritz, PHD,‡ F. Philip Anderson, PHD,‡
James L. Tatum, MD,§ Joseph P. Ornato, MD,† Robert L. Jesse, MD, PHD*
Richmond, Virginia
We compared outcomes in patients with non–ST-segment elevation acute coronary syndromes (ACS) according to the degree of cardiac troponin I (cTnI) elevation.
BACKGROUND Controlled trials of high-risk patients have found that troponin elevations identify an even
higher risk subset. It is unclear whether outcomes are similar among a lower risk,
heterogeneous patient group. Also, few studies have reported outcomes other than myocardial
infarction (MI) or death, based on the peak troponin value.
METHODS
Consecutively, admitted patients without ST-segment elevation on the initial electrocardiogram underwent serial marker sampling using creatine kinase (CK), CK-MB fraction, and
cTnI. Patients were grouped according to peak cTnI: negative ⫽ no detectable cTnI; low ⫽
peak greater than the lower limit of detectability but less than the optimal diagnostic value;
intermediate ⫽ peak greater than or equal to the optimal diagnostic value but less than the
manufacturer’s suggested upper reference limit (URL); and high ⫽ peak greater than or equal
to the URL. Thirty-day outcomes included cardiac death, MI based on CK-MB, revascularization, significant disease, and a reversible defect on stress testing. Six-month mortality
was also determined. Negative evaluations for ischemia included nonsignificant disease, no
reversible stress defect, and negative rest perfusion imaging.
RESULTS
Of the 4,123 patients admitted, 893 (22%) had detectable cTnI values. Cardiac events and
positive test results at 30 days and 6-month mortality increased significantly with increasing
cTnI values. Negative evaluations for ischemia were significantly and inversely related to peak
cTnI values. Although adverse events were significantly more common in patients with a low
cTnI value than in those with negative cTnI, negative evaluations for ischemia were frequent.
CONCLUSIONS Increased cTnI values are associated with worse outcomes. Although low cTnI values are
associated with adverse events, they do not have the same implication as higher cTnI values, and
nonischemic evaluations are frequent. (J Am Coll Cardiol 2004;43:958 – 65) © 2004 by the
American College of Cardiology Foundation
OBJECTIVES
Both cardiac troponins T (cTnT) and I (cTnI) have a higher
sensitivity for detecting myocardial necrosis than do traditional cardiac markers (1,2). They also add significant
incremental diagnostic and prognostic value to routine
clinical and electrocardiographic (ECG) variables for identifying patients at risk of cardiac events (3,4). These advantages led a joint committee of the European Society of
Cardiology and American College of Cardiology (ESC/
ACC) to recommend that patients who have detectable
cardiac troponin, as a result of myocardial ischemia, be
diagnosed as having acute myocardial infarction (MI) (5).
From the Departments of *Internal Medicine, Cardiology Division; †Emergency
Medicine; ‡Pathology, Clinical Chemistry Division; and §Radiology, Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia. This study
was presented in part at the Sessions of the American Heart Association Scientific
Meetings, November 2002, Anaheim, California. Dr. Kontos—Speakers Bureaus
and/or served as a consultant for Eli Lily and Co., Centocor, Cor Therapeutics,
Millenium, Genentech, Aventis, and Merck and Co, Inc.; Dr. Tatum—Speakers
Bureau for DuPont; Dr. Ornato— consultant for Genentech and Aventis; Dr.
Jesse—Speakers Bureaus and/or served as a consultant for Eli Lily and Co., Centocor,
Scios, Cor Therapeutics, Millenium, Genentech, Aventis, Genentech, Merck and Co,
Inc., Dade Behring, and DuPont.
Manuscript received May 20, 2003; revised manuscript received September 30,
2003, accepted October 6, 2003.
Patients with troponin elevations appear to benefit preferentially from antiplatelet (6 – 8) and antithrombotic treatment (9), as well as early coronary angiography (2). As a
result, these treatment strategies have been incorporated
into recommendations for the treatment of patients with
non–ST-segment elevation acute coronary syndromes (ACS)
(3,4). However, these recommendations were based primarily
on studies that evaluated outcomes in higher risk patients
enrolled in ACS trials. In addition, the definition of positive
troponin varies substantially among the studies (2,6,7,10 –12).
Therefore, the conclusions drawn are difficult to apply and may
not be valid in lower risk patients, such as those undergoing an
evaluation in the emergency department (ED), particularly
those with lower troponin concentrations. In this study, we
assessed outcomes based on different peak cTnI values in a
large, consecutive group of patients admitted from the ED for
exclusion of myocardial ischemia.
METHODS
This study was performed at a 600-bed, inner-city hospital
with approximately 85,000 ED visits and 1,500 coronary
JACC Vol. 43, No. 6, 2004
March 17, 2004:958–65
Abbreviations and Acronyms
ACS
⫽ acute coronary syndromes
CCU
⫽ coronary care unit
CI
⫽ confidence interval
CK
⫽ creatine kinase
cTnI
⫽ cardiac troponin I
cTnT
⫽ cardiac troponin T
CV
⫽ coefficient of variation
ED
⫽ emergency department
ESC/ACC ⫽ European Society of Cardiology/
American College of Cardiology
LLD
⫽ lower limit of detectability
MI
⫽ myocardial infarction
MPI
⫽ myocardial perfusion imaging
URL
⫽ upper reference limit
care unit (CCU) admissions per year. The chest pain
protocol used at our institution has been described in detail
previously (13). In brief, after the initial evaluation, patients
thought to be at high risk (those with ischemic ECG
changes and those with known coronary disease experiencing typical symptoms) are admitted directly to the CCU.
Patients considered at low to moderate risk of ACS (e.g.,
absence of ischemic ECG changes) undergo further risk
stratification using rest myocardial perfusion imaging (MPI)
(13). Moderate-risk patients are admitted as observation
patients and undergo serial marker sampling to exclude MI.
Low-risk patients undergo rest MPI in the ED and are
discharged and scheduled for outpatient stress testing if the
images are either negative or unchanged from previous
studies. Markers are not routinely measured in these patients. Patients with positive MPI are considered high risk
and admitted for MI exclusion.
A total of 4,567 consecutive patients were admitted to the
CCU from June 1996 through March 2000. For patients with
multiple admissions during each study period, only the first
admission was included in the analysis. Patients with STsegment elevation who met criteria for fibrinolytic therapy (n
⫽ 230) or did not have an 8-h cTnI (124 for assay 1; 101 for
assay 2) were excluded, leaving 4,123 patients who formed the
study cohort. This study was approved by the Office of
Research Subjects Protection for the Conduct of Human
Research.
Markers. All patients underwent serial testing for creatine
kinase (CK) and CK-MB by mass assay at 0, 3, 6, and 8 h
and cTnI at 0 and 8 h after presentation. Patients who had
positive markers or recurrent or continuing symptoms had
sampling continued at 6- to 8-h intervals until a peak value
was reached or a diagnosis was made. The peak cTnI value
during the first 24 h was used for analysis.
Centrifuged plasma was filtered before CK-MB and
cTnI analyses. Two different diagnostic assays for CK-MB
and cTnI were used. From June 1996 to May 1998, the
Opus Magnum Analyzer (Behring Diagnostics, Boston,
Massachusetts) was used. The lower limit of detectability
(LLD) for this assay was 0.5 ng/ml, and the suggested
Kontos et al.
Troponin I and Outcomes
959
diagnostic value for MI (upper reference limit [URL]) was
2.5 ng/ml (14). The 99th percentile for patients without
coronary disease was ⬍0.5 ng/ml. The coefficients of variation (CV) reported by the manufacturer were 12% and 5%
for cTnI values of 3.0 and 19 ng/ml, respectively. The CV
values for lower cTnI values were not available. Outcomes
for these patients have been reported in detail previously
(15). The current study represents expansion of outcomes
assessed, as well as analysis using different cTnI diagnostic
values from a previous study (15). From May 1998 to April
2000, cTnI was analyzed using the Bayer assay (Bayer
Corp., Tarrytown, New York). The LLD for this assay was
0.1 ng/ml, and the suggested diagnostic value for MI was
0.9 mg/ml (10). The 99th percentile determined in 200
patients without evidence of coronary disease was ⬍0.1
ng/ml. The CV for this assay at our institution was ⬍15%
at a cTnI concentration of 0.3 ng/ml. For CK-MB, an URL
of 8.0 ng/ml was used for both assays.
For each of the two assays, we chose an optimal diagnostic value for cTnI (Opus, 1.0 ng/ml; Bayer, 0.3 ng/ml),
which improved specificity without affecting sensitivity,
compared with the reference standard CK-MB MI definition. Sensitivity (96% vs. 97%) and specificity (93% vs. 92%)
at the optimal diagnostic value were similar and not statistically different for the two assays. In addition, they were
equivalent to the recent ESC/ACC definition of MI.
Patients were then separated into four groups according to
peak cTnI values: 1) negative ⫽ no detectable cTnI; 2) low
⫽ peak cTnI values greater than or equal to the LLD and
less than the optimal diagnostic value; 3) intermediate ⫽
peak cTnI values greater than or equal to the optimal
diagnostic value but less than the URL; and 4) high ⫽ peak
cTnI values greater than or equal to the URL.
The results from the two assays were subsequently combined because the areas under the receiver-operating characteristics curve for the two assays (Opus assay, 0.977 ⫾
0.04; Bayer assay, 0.971 ⫾ 0.04) were comparable and not
significantly different, indicating similar diagnostic test performance. In addition, there was no significant difference
between the two groups in the baseline characteristics
associated with adverse cardiac outcomes and test results in
ACS patients (Table 1).
Outcomes. Patient outcomes occurring within 30 days of
admission were assessed. When a patient experienced more
than one cardiac event, only one was considered, based on
the following order: cardiac death, criteria for MI at the
time of admission, late CK-MB MI (⬎24 h but within 30
days), revascularization, significant coronary disease, and a
reversible defect on stress imaging. Negative evaluations for
ischemia were assessed in a similar fashion and included
nonsignificant disease on angiography, a normal or fixed
defect on stress MPI, and negative rest MPI during the ED
evaluation. Cardiac death was defined as death due to MI,
coronary artery disease, or arrhythmia. In addition, sixmonth cardiac mortality was also assessed. Mortality was
determined using chart review, scripted phone interviews,
960
Kontos et al.
Troponin I and Outcomes
JACC Vol. 43, No. 6, 2004
March 17, 2004:958–65
Table 1. Demographics and Risk Factors for the Two Populations
Age (yrs)
Age ⱖ65 yrs
Male
Hypertension
Diabetes
Tobacco use
Total risk factors
Previous revascularization
Previous bypass surgery
Previous angioplasty
Prior MI
CK-MB MI
Ischemic ECG changes
All Patients
(n ⴝ 4,123)
Assay 1
(n ⴝ 1,943)
Assay 2
(n ⴝ 2,180)
58 ⫾ 14 (58)
1,312 (32%)
2,038 (49%)
2,709 (66%)
1,308 (32%)
1,757 (43%)
2.1 ⫾ 1.2 (2)
1,039 (25%)
528 (13%)
725 (18%)
600 (15%)
341 (8.4%)
493 (12%)
58 ⫾ 14 (58)
630 (12%)
946 (49%)
1,271 (65%)
615 (32%)
809 (42%)
2.1 ⫾ 1.2 (2)
464 (24%)
255 (13%)
340 (18%)
286 (15%)
176 (9.1%)
215 (11%)
58 ⫾ 14 (58)
682 (31%)
1,092 (50%)
1,438 (66%)
693 (32%)
948 (44%)
2.2 ⫾ 1.2 (2)
531 (24%)
273 (13%)
385 (18%)
315 (14%)
170 (7.8%)
278 (13%)
There were no significant differences between variables in the two groups. Data are presented as the mean value ⫾ SD (median)
or number (%) of subjects.
CK-MB ⫽ creatine kinase-MB; ECG ⫽ electrocardiogram; MI ⫽ myocardial infarction.
and death registry data from the Virginia Division of Health
Statistics and the Social Security Death Index. The
definition of CK-MB MI was an elevation of CK-MB
ⱖ8.0 ng/ml, with a relative index ⱖ4.0 (CK-MB ⫻
100/total CK), in association with a characteristic marker
rise and fall. Significant coronary disease was defined as
ⱖ60% left main stenosis and ⱖ70% stenosis in a major
coronary artery, its branches, or a bypass graft supplying
myocardium at ischemic risk. Stress testing was performed
using symptom-limited exercise or pharmacologic stress with
single-photon emission computed tomographic MPI and was
considered abnormal if there was a reversible defect. Rest MPI
was considered positive if there was a perfusion defect in
conjunction with abnormal wall motion or thickening (13). An
ischemic ECG was defined as transient ST-segment elevation,
ST-segment depression ⱖ1 mm, or ischemic T-wave inversion
(symmetrical T-wave inverted ⱖ2 mm).
Statistical analysis. Results were compared using the Student t test for continuous variables and chi-square analysis for
dichotomous variables. A value of p ⱕ 0.05 was considered
significant in most cases. When differences in cardiac outcomes
were compared among the different cTnI groups, a value p ⬍
0.017 was required to correct for multiple comparisons. Cardiac mortality was shown using the Kaplan-Meier method,
with comparisons made using the log-rank test. Significance of
the trend toward an increasing incidence of adverse outcomes
and a decreasing incidence of negative ischemic evaluations
was assessed using linear regression analysis. A Cox regression
model was used to model the hazard of six-month cardiac
mortality as a function of cTnI positivity. The Wald test was
used to test for all pairwise difference in hazard rates. The
Bonferroni correction was used for multiple testing. Statistical
analyses were performed using SAS version 8.2 (Cary, North
Carolina).
RESULTS
The characteristics and demographic variables of the 4,123
patients are shown in Table 1. A total of 346 patients (8.4%)
had CK-MB MI. Of the 893 patients (22%) who had
detectable cTnI values, 425 (10%) had cTnI values greater
than the URL (high cTnI), 198 (4.8%) had cTnI values less
than the URL but greater than or equal to the optimal
diagnostic value (intermediate cTnI), and 270 (6.5%) had
cTnI values less than the optimal diagnostic value but
greater than the LLD (low cTnI). Serial testing made a
difference in classifying these patients, as the initial cTnI
often did not reflect the final classification. Among the low
cTnI group, 45% had negative cTnI initially. Similarly, 49%
of patients in the intermediate cTnI group had initial cTnI
values that were negative (29%) or low (20%), and 52% of
those who had high cTnI values had initial values that were
negative (29%), low (13%), or intermediate (9%). Therefore,
if only the initial cTnI were used, 50% of the patients with
detectable cTnI values would have been misclassified.
Thirty-day and six-month follow-up was complete in
92% and 96%, respectively. Patients who did not have
follow-up were significantly less likely to have cTnI elevations and revascularization, were younger, had fewer risk
factors for coronary disease, and were more likely to have an
evaluation negative for ischemia. Additional diagnostic
testing was performed within 30 days of admission in 73%
of patients (n ⫽ 3,020). A total of 1,508 patients (37%)
underwent coronary angiography, with 579 (14%) undergoing revascularization. Stress MPI was performed in 1,116
patients (27%) who did not undergo coronary angiography,
and 396 patients (9.6%) had negative rest MPI during the
initial ED evaluation but did not undergo further diagnostic
testing, other than serial marker sampling. At least one end
point (cardiac death or CK-MB MI, or additional diagnostic testing) was present in 76%, 69%, 67%, and 92% of
patients who had negative, low, intermediate, or high peak
cTnI values, respectively. The distribution of test results
based on peak cTnI values are shown in Table 2.
Six-month cardiac mortality is shown in Figure 1. There
was a highly significant (p ⬍ 0.0001) stepwise increase in
cardiac mortality as peak cTnI increased. After correcting
Kontos et al.
Troponin I and Outcomes
JACC Vol. 43, No. 6, 2004
March 17, 2004:958–65
961
Table 2. Cardiac Outcomes and Test Results Based on Peak Cardiac Troponin I Value
None (n ⫽ 3,229)
Low cTnI (n ⫽ 270)
Intermediate cTnI (n ⫽ 198)
High cTnI (n ⫽ 426)
Death
MI
Revasc.
Sig Dz
(ⴙ) Stress
Non-Sig Dz
(ⴚ) Stress
(ⴚ) MPI
22 (0.7%)
6 (2.2%)
6 (3.0%)
30 (7.0%)
8 (0.2%)
3 (1.1%)
29 (15%)
306 (72%)
306 (9.5%)
42 (16%)
49 (25%)
186 (44%)
541 (17%)
59 (22%)
67 (34%)
264 (62%)
298 (9.2%)
19 (7.0%)
6 (3.0%)
14 (3.3%)
488 (15%)
37 (14%)
20 (10%)
31 (7.3%)
967 (30%)
56 (21%)
22 (11%)
15 (3.5%)
353 (11%)
29 (11%)
9 (4.5%)
5 (1.2%)
A patient could have more than one event or test result. Data are expressed as the number (%) of subjects.
cTnI ⫽ cardiac troponin I; MI ⫽ myocardial infarction; MPI ⫽ myocardial perfusion imaging; Revasc. ⫽ revascularization; Sig Dz ⫽ significant disease; (⫹) ⫽ positive;
(⫺) ⫽ negative.
for multiple comparisons, mortality was significantly higher
in all cTnI groups, including the low cTnI group, compared
with the negative cTnI group (p ⬍ 0.01). In addition,
mortality was significantly different between the low and
high cTnI groups. Differences between the low and intermediate cTnI groups and the intermediate and high cTnI
groups were not statistically significant. Compared with the
negative cTnI group, the hazard ratio for the low cTnI
group was 2.5 (95% confidence interval [CI] 1.4 to 4.4), 3.9
for the intermediate cTnI group (95% CI 2.3 to 6.8), and
6.1 for the high cTnI group (95% CI 4.2 to 8.7).
Thirty-day outcomes based on peak cTnI values are
shown in Figure 2. In each of the outcome categories, the
trend toward an increasing incidence of ischemic outcomes
with increasing peak cTnI was highly significant. Figure 2
also shows the differences between each cTnI group. Only a
small minority of patients who had low or negative cTnI
values met CK-MB criteria for MI. In contrast, the majority
of patients with high cTnI values had CK-MB MI. The
results were essentially unchanged after excluding patients
who had CK-MB MI at the time of admission (Fig. 3).
The incidence of evaluations that were negative for
ischemia showed an inverse relation to peak cTnI values,
decreasing as peak cTnI values increased. An evaluation that
was negative for ischemia (CK-MB MI excluded and either
nonsignificant coronary disease, no reversible stress defect,
or negative rest MPI) was found in 54% of those without
cTnI elevations, 42% of those with low cTnI values (p ⬍
0.001 vs. negative cTnI group), 22% of those with intermediate values (p ⬍ 0.001 vs. low cTnI group), and only 8.7%
of those with high cTnI values.
Because the distinction between negative cTnI and low
cTnI values was of particular interest, the two groups were
further analyzed. When the two groups were compared, the
differences in six-month cardiac mortality was significant (p
⬍ 0.02) (Fig. 1). Events at 30 days (including the combinations of death and MI; death, MI, and revascularization;
and death, MI, and significant disease) also differed significantly (Figs. 2 and 3). When patients in the low cTnI group
who had an ischemic evaluation were compared with those
who did not, the only variables different were the presence
of an ischemic ECG (22% vs. 10%, p ⬍ 0.05) and ECG
evidence of previous MI (20% vs. 9%, p ⬍ 0.05).
When outcome and test results were analyzed individually, a similar gradient appeared, with the incidence of
death, MI, revascularization, and significant disease increasing as cTnI values increased. Also, the incidence of negative
cardiac evaluations decreased as peak cTnI values increased
Figure 1. Kaplan-Meier curves of six-month cardiac mortality for the different cardiac troponin I (cTnI) values. The trend toward increased mortality and
increased cTnI was highly significant (p ⬍ 0.001). In addition, mortality was significantly different (p ⬍ 0.01) between the negative (neg) cTnI and low,
intermediate (inter), and high cTnI groups.
962
Kontos et al.
Troponin I and Outcomes
JACC Vol. 43, No. 6, 2004
March 17, 2004:958–65
Figure 2. Positive outcomes based on peak cardiac troponin I (cTnI) values. The trend toward an increased number of events for increasing peak cTnI values
was highly significant (p ⬍ 0.001). *p ⬍ 0.001. #p ⬍ 0.02. Open bars ⫽ no troponin I; vertically striped bars ⫽ low troponin I; horizontally striped
bars ⫽ intermediate troponin I; diagonally striped bars ⫽ high troponin I. MI ⫽ myocardial infarction; Revasc ⫽ revascularization; Sig Dz ⫽ significant
disease.
(Table 2). The only exception was a positive stress test
result, which was likely due to coronary angiography being
performed as the preferred initial evaluation.
DISCUSSION
In this large study of ED chest pain patients without
ST-segment elevation on the initial ECG, we found a
significant incremental increase in cardiac events and positive test results as the peak cTnI value increased. Low peak
cTnI values predicted cardiac events, but negative evalua-
tions for ischemia were also common. Importantly, these
relationships persisted after excluding CK-MB MI at the
time of admission as an end point.
We separated patients into four groups based on their
peak cTnI value, using the URL, LLD, and optimal
diagnostic value as decision limits. The first two were
chosen because they are the most commonly used values in
research studies and clinical practice (2,10,16,17). The
third—the optimal diagnostic value—was chosen using
receiver operating characteristic curve analysis, which allows
Figure 3. Positive outcomes based on peak cardiac troponin I (cTnI) values after excluding creatine kinase-MB myocardial infarction (MI) at the time of
admission as an end point. The trend toward an increased number of events for increasing peak cTnI values was highly significant (p ⬍ 0.001). *p ⬍ 0.02.
Open bars ⫽ no troponin I; vertically striped bars ⫽ low troponin I; horizontally striped bars ⫽ intermediate troponin I; diagonally striped bars ⫽
high troponin I. Revasc ⫽ revascularization; Sig Dz ⫽ significant disease.
JACC Vol. 43, No. 6, 2004
March 17, 2004:958–65
more appropriate selection of a diagnostic value through
optimization of sensitivity and specificity (18). This value
was similar to the one proposed by the ESC/ACC (10%
CV) (5).
Non-MI end points. The majority of reports analyzing
outcomes based on troponin positivity used only the “hard”
end points of MI and death. However, these outcomes
occur in only a minority of patients; therefore, this approach
provides only a limited assessment of the diagnostic ability
of troponin. It also fails to identify patients in whom
outcomes may be impacted by subsequent treatment, such
as revascularization. Studies that have reported other outcomes, such as stress testing (19,20) or coronary angiography (21), typically included too few patients to analyze
outcomes based on different troponin levels. Only two
studies did include sufficient numbers, allowing stratification by troponin concentration (22,23). Although performed in a population of predominately high-risk ACS
patients, both studies reported results similar to ours:
increasing troponin values were associated with a higher
prevalence of positive stress tests, significant coronary disease, and revascularization procedures (22,23). Our results
extend this finding to lower risk patients. In addition, we
found that negative ischemic evaluations increased with
decreasing cTnI values.
Low cTnI and outcomes. Because cardiac troponin is not
normally found in the blood, small amounts of damage,
which may not meet traditional criteria for MI, can be
detected. Studies performed in high-risk ACS patients did
not find any concentration of detectable troponin that was
not associated with increased risk (1,2,5,24). This observation led to the current recommendation that, in the proper
setting, any detectable troponin should be considered
pathologic and indicative of MI (5). In addition to identifying troponin as the preferred diagnostic marker, the
recommendations also specified that the diagnostic cut-off
value should be ⬎99th percentile. The LLD for the two
assays used in the present study met this criteria, as they
were undetectable in large control groups.
Several studies that used the LLD for cTnI (2,25) and
cTnT (2) observed that the presence of these minor elevations was associated with a significantly higher cardiac event
rate, compared with the absence of detectable troponin. In
contrast, other studies performed in lower risk patient
populations found no such difference (16,26 –28). This
discrepancy can be attributed to two factors: patient risk and
cohort size. Unlike patients in most ACS trials which have
a high prevalence of MI, most ED patients have a low
overall risk. Greater numbers of patients are therefore
required to demonstrate a significant difference in outcomes.
The large number of patients in the current study allowed us
to demonstrate a stepwise increase in events with increasing
cTnI values, confirming that even small cTnI elevations
have prognostic significance.
However, evaluations that were negative for ischemia
were common in patients with low cTnI values. One
Kontos et al.
Troponin I and Outcomes
963
explanation is that previous treatment and preprocedural
delays resulted in thrombus resolution by the time angiography was performed (29). Positive cTnI values in patients
with negative imaging could be related to the relative
insensitivity of MPI, as approximately 3% to 5% of myocardium must be ischemic for detection (30). Another
possible mechanism is the release of small amounts of
troponin resulting from global ischemia and patchy necrosis,
such as during a hypertensive crisis or severe heart failure,
rather than through prolonged ischemia and infarction from
epicardial coronary artery disease. However, the high number of patients with low cTnI elevations who did not have
events suggests that most discrepancies represent analytical
false-positive results due to the assays themselves, rather
than necrosis.
Troponin I values just above the LLD have been a source
of considerable diagnostic confusion. In contrast to ACS
trials, only a minority of ED patients with chest pain have
diagnostic ECG changes. Because atypical presentations are
frequent, the presence of troponin elevations is often the
primary, if not only, criterion used to diagnose these
patients as having MI. Spurious elevations have been
frequent enough in some studies (31–33) so that the term
“troponinosis” (33) has been coined to describe them.
A high prevalence of analytical false-positive results has
important implications for the treatment of chest pain
patients. Troponin-positive patients appear to benefit preferentially from more intensive antiplatelet and antithrombotic therapy (6,7,12,16,34), as well as early coronary
angiography with revascularization, when appropriate
(2,12). Recommendations for the treatment of patients with
non–ST-segment elevation ACS now use the detection of
troponin to guide diagnostic and therapeutic measures (3,4).
Given the high frequency of negative subsequent evaluations in chest pain patients with low troponin elevations,
such intensive treatment may not be warranted in all
patients (16,35). Rather, treatment decisions should be
based on clinical variables indicative of high risk, such as an
ischemic ECG or previous MI.
The high frequency of analytical false-positive troponin
values near the LLD led to the ESC/ACC recommendation
of ⱕ10% CV at the diagnostic value (5). Almost none of the
currently available cTnI assays (36 –38) or the current cTnT
assay (39) are able to achieve this. Thus, fulfilling this
requirement will require a higher diagnostic value, often two
to three times the LLD. Values that fall between the LLD
and the diagnostic value are therefore indeterminate and are
not to be considered diagnostic of MI.
It is unclear what these indeterminate troponin values
should be called. Based on the ESC/ACC definition, they
would be considered negative. However, our data, as well as
those of others (2), indicate that these lower cTnI values
have prognostic value, so disregarding them is not appropriate. Our data suggest that until assays meeting the
criteria specified by the ESC/ACC are available, the use of
two decision limits, as suggested by the National Academy
964
Kontos et al.
Troponin I and Outcomes
of Clinical Biochemists (40), may be necessary. The MI
would be diagnosed if the troponin value exceeded the
diagnostic limit, whereas lower values would be considered
equivocal, with further clinical correlation and evaluation
required.
Study limitations. Although the current study is observational, it has several advantages over many previous studies
in which troponin analyses were reported as substudies of
controlled clinical trials. In most multicenter trials, sampling
was performed at only one time point, reducing accuracy. In
contrast, we performed serial sampling in all patients. The
prespecified inclusion and exclusion criteria used in multicenter trials frequently lead to high-risk, relatively homogeneous populations (1,6 –12) in which women and the
elderly are under-represented (17). Assessing a test’s performance in only a high-risk population results in substantial
bias, thus limiting the ability to generalize the results.
Therefore, studies such as ours are important to complement the information obtained from randomized, clinical
trials. Although this was a single-center study, the patient
characteristics and outcomes, including the proportion of
patients who had CK-MB MI (25,28,41,42) and troponin
elevations without CK-MB MI (25,27,28,42), were similar
to those reported in both single- and multicenter studies,
indicating that these results should be generalizable to other
sites, as long as the definition of troponin positivity is
consistent with the recent consensus recommendations (5).
Not all patients underwent diagnostic testing beyond
serial marker analysis. This reflects actual clinical practice in
which individual patient variables and clinical judgment
ultimately determine what, if any, additional testing is
necessary. However, the trend toward increased events as
cTnI values increased was consistent across the spectrum of
outcome events. Also, physicians were not blinded to the
cTnI values; therefore, further care was based in part on
these results. Although clinical trials blind physicians to
troponin values, myocardial markers, including cTnI, are
routinely assessed in all patients admitted to most U.S.
hospitals (43). The range of diagnostic testing performed in
patients with and without cTnI elevations suggests that the
decisions were based on the entire clinical presentation
rather than the cTnI results alone. We used cardiac rather
than all-cause mortality as an end point. In contrast to
clinical trials, in which the majority of patients have a high
prevalence of coronary disease, many of the patients included in the current study had a nonischemic cause for
their chest pain, which would have increased the overall
mortality in the lower risk patients. Although follow-up was
not complete, it was comparable to previous studies, and
patients without follow-up were at lower risk. Our use of
the Virginia Death Registry, as well as the Social Security
Death Index, which has been demonstrated to have a high
specificity (44), further supports the veracity of our conclusions.
Conclusions. Cardiac troponin I elevations are associated
with an increased adverse event rate, although low cTnI
JACC Vol. 43, No. 6, 2004
March 17, 2004:958–65
values are frequently associated with a high rate of negative
evaluations for ischemia. Therefore, the degree of troponin
elevation, as well as clinical variables such as previous MI
and an ischemic ECG, should be considered when deciding
treatment.
Reprint requests and correspondence: Dr. Michael C. Kontos,
Room 7-074, Heart Station, North Hospital, P.O. Box 980051,
Medical College of Virginia, 12th and Marshall Streets, Richmond, Virginia 23298-0051. E-mail: [email protected].
REFERENCES
1. Lindahl B, Venge P, Wallentin L, the FRISC Study Group. Relation
between troponin T and the risk of subsequent cardiac events in
unstable coronary artery disease. Circulation 1996;93:1651–7.
2. Morrow DA, Cannon CP, Rifai N, et al. Ability of minor elevations
of troponins I and T to predict benefit from an early invasive strategy
in patients with unstable angina and non-ST elevation myocardial
infarction. JAMA 2001;286:2405–12.
3. Bertrand ME, Simoons ML, Fox KA, et al. Management of acute
coronary syndromes: acute coronary syndromes without persistent
ST-segment elevation. Recommendations of the Task Force of the
European Society of Cardiology. Eur Heart J 2000;21:1406 –32.
4. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA guidelines
for the management of patients with unstable angina and non–STsegment elevation myocardial infarction: a report of the American
College of Cardiology/American Heart Association Task Force on
Practice Guidelines (Committee on the Management of Patients with
Unstable Angina). J Am Coll Cardiol 2000;36:970 –1062.
5. Alpert JS, Thygesen K, Bassand JP, et al. Myocardial infarction
redefined—a consensus document of the Joint European Society of
Cardiology/American College of Cardiology Committee for the redefinition of myocardial Infarction. J Am Coll Cardiol 2000;36:959 –
69.
6. Hamm CW, Heeschen C, Goldmann B, et al., the Chimeric 7E3
AntiPlateleT in Unstable angina REfractory to standard treatment
(CAPTURE) Study Investigators. Benefit of abciximab in patients
with refractory unstable angina in relation to serum troponin T levels.
N Engl J Med 1999;340:1623–9.
7. Newby LK, Ohman EM, Christenson RH, et al. Benefit of glycoprotein IIb/IIIa inhibition in patients with acute coronary syndromes
and troponin T positive status: the paragon B troponin T study.
Circulation 2001;103:2981–6.
8. Heeschen C, Hamm C, Goldmann B, Deu A, Langenbrink L, White
HD, the PRISM Study Investigators. Troponin concentrations for
stratification of patients with acute coronary syndromes in relation to
therapeutic efficacy of tirofiban. Lancet 1999;354:1757–62.
9. Morrow DA, Antman EM, Tanasijevic M, et al. Cardiac troponin I
for stratification of early outcomes and the efficacy of enoxaparin in
unstable angina: a TIMI-11B substudy. J Am Coll Cardiol 2001;36:
1812–7.
10. Morrow DA, Rifai N, Tanasijevic MJ, Wybenga DR, de Lemos JA,
Antman EM. Clinical efficacy of three assays for cardiac troponin I for
risk stratification in acute coronary syndromes: a Thrombolysis In
Myocardial Infarction (TIMI) 11B substudy. Clin Chem 2000;46:
453–60.
11. Morrow DA, Antman EM, Tanasijevic M, et al. Cardiac troponin I
for stratification of early outcomes and the efficacy of enoxaparin in
unstable angina: a TIMI-11B substudy. J Am Coll Cardiol 2000;36:
1812–7.
12. Lindahl B, Toss H, Siegbahn A, Venge P, Wallentin L. Markers of
myocardial damage and inflammation in relation to long-term mortality in unstable coronary artery disease. N Engl J Med 2000;343:
1139 –47.
13. Tatum JL, Jesse RL, Kontos MC, et al. Comprehensive strategy for
the evaluation and triage of the chest pain patient. Ann Emerg Med
1997;29:116 –23.
14. Wu AH, Feng YJ, Contois JH, Pervaiz S. Comparison of myoglobin,
creatine kinase-MB, and cardiac troponin I for diagnosis of acute
myocardial infarction. Ann Clin Lab Sci 1996;26:291–300.
Kontos et al.
Troponin I and Outcomes
JACC Vol. 43, No. 6, 2004
March 17, 2004:958–65
15. Kontos MC, Anderson FP, Alimard R, Ornato JP, Tatum JL, Jesse
RL. Ability of troponin I to predict cardiac events in patients admitted
from the emergency department. J Am Coll Cardiol 2000;36:1818 –23.
16. Heeschen C, Hamm CW, Goldmann B, Deu A, Langenbrink L,
White HD. Troponin concentrations for stratification of patients with
acute coronary syndromes in relation to therapeutic efficacy of tirofiban. Lancet 1999;354:1754 –62.
17. Lee PY, Alexander KP, Hammill BG, Pasquali SK, Peterson ED.
Representation of elderly persons and women in published randomized
trials of acute coronary syndromes. JAMA 2001;286:708 –13.
18. Zweig MH, Campbell G. Receiver-operating characteristic (ROC)
plots: a fundamental evaluation tool in clinical medicine. Clin Chem
1993;39:561–77.
19. Newby LK, Kaplan AL, Granger BB, Sedor F, Califf RM, Ohman
EM. Comparison of cardiac troponin T versus creatine kinase-MB for
risk stratification in a chest pain evaluation unit. Am J Cardiol
2000;85:801–5.
20. Fearon WF, Lee FH, Froelicher VF. Does elevated cardiac troponin I
in patients with unstable angina predict ischemia on stress testing?
Am J Cardiol 1999;84:1440 –2.
21. Benamer H, Steg PG, Benessiano J, et al. Elevated cardiac troponin I
predicts a high-risk angiographic anatomy of the culprit lesion in
unstable angina. Am Heart J 1999;137:815–20.
22. Lindahl B, Andren B, Ohlsson J, Venge P, Wallentin L, the FRISK
Study Group. Risk stratification in unstable coronary artery disease:
additive value of troponin T determinations and pre-discharge exercise
tests. Eur Heart J 1997;18:762–70.
23. Safstrom K, Lindahl B, Swahn E, the FRISC Study Group. Risk
stratification in unstable coronary artery disease— exercise test and
troponin T from a gender perspective. J Am Coll Cardiol 2000;35:
1791–800.
24. Antman EM, Tanasijevic MJ, Thompson B, et al. Cardiac-specific
troponin I levels to predict the risk of mortality in patients with acute
coronary syndromes. N Engl J Med 1996;335:1342–9.
25. Polanczyk CA, Lee TH, Cook EF, et al. Cardiac troponin I as a
predictor of major cardiac events in emergency department patients
with acute chest pain. J Am Coll Cardiol 1998;32:8 –14.
26. Fuchs S, Kornowski R, Mehran R, et al. Cardiac troponin I levels and
clinical outcomes in patients with acute coronary syndromes: the
potential role of early percutaneous revascularization. J Am Coll
Cardiol 1999;34:1704 –10.
27. Sayre MR, Kaufmann KH, Chen IW, et al. Measurement of cardiac
troponin T is an effective method for predicting complications among
emergency department patients with chest pain. Ann Emerg Med
1998;31:539 –49.
28. Johnson PA, Goldman L, Sacks DB, et al. Cardiac troponin T as a
marker for myocardial ischemia in patients seen at the emergency
department for acute chest pain. Am Heart J 1999;137:1137–44.
29. Heeschen C, van den Brand MJ, Hamm CW, Simoons ML. Angiographic findings in patients with refractory unstable angina according
to troponin T status. Circulation 1999;100:1509 –14.
30. Verani MS, Jeroudi MO, Mahmarian JJ, et al. Quantification of
myocardial infarction during coronary occlusion and myocardial sal-
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
965
vage after reperfusion using cardiac imaging with technetium-99m
hexakis 2-methoxyisobutyl isonitrile. J Am Coll Cardiol 1988;12:
1573–81.
Khan IA, Tun A, Wattanasauwan N, et al. Elevation of serum cardiac
troponin I in noncardiac and cardiac diseases other than acute coronary
syndromes. Am J Emerg Med 1999;17:225–9.
Wright SW, Sawyer DB, Chyu S, Goldhaber SZ. Elevations of
troponin I levels in patients without evidence of myocardial injury.
JAMA 1997;278:2144.
Ng SM, Krishnaswamy P, Morrisey R, Clopton P, Fitzgerald R,
Maisel AM. Mitigation of the clinical significance of spurious elevations of cardiac troponin I in settings of coronary ischemia using serial
testing of multiple cardiac markers. Am J Cardiol 2001;87:994 –7.
Lindahl B, Venge P, Wallentin L, the FRagmin during InStability in
Coronary artery disease (FRISC) Study Group. Troponin T identifies
patients with unstable coronary artery disease who benefit from
long-term antithrombotic protection. J Am Coll Cardiol 1997;29:
43–8.
Roe MT, Harrington RA, Prosper DM, et al. Clinical and therapeutic
profile of patients presenting with acute coronary syndromes who do
not have significant coronary artery disease. Circulation 2000;102:
1101–6.
Jaffe AS, Ravkilde J, Roberts R, et al. It’s time for a change to a
troponin standard. Circulation 2000;102:1216 –20.
James S, Lindahl B, Venge P, et al. No beneficial effect of 24 to 48
hour abciximab infusion in aspirin and heparin treated acute coronary
syndrome patients with elevated troponin without early revascularization procedures: a GUSTO-IV ACS substudy (abstr). J Am Coll
Cardiol 2001;37 Suppl A:326A.
Panteghini M. Present issues in the determination of troponins and
other markers of cardiac damage. Clin Biochem 2000;33:161–6.
Diderholm E, Andren B, Frostfeld G, et al. The prognostic and
therapeutic implications of increased troponin T levels and ST
depression in unstable coronary artery disease: the FRISC II invasive
troponin T electrocardiogram substudy. Am Heart J 2002;143:760 –7.
Wu AHB, Apple FS, Warshaw MM, Valdes R Jr., Jesse RL, Gibler
WB. National Academy of Clinical Biochemistry Standards of Laboratory Practice recommendations for use of cardiac markers in
coronary artery disease. Clin Chem 1999;45:1104 –21.
Selker HP, Beshansky JR, Griffith JL, et al. Use of the acute cardiac
ischemia time-insensitive predictive instrument (ACI-TIPI) to assist
with triage of patients with chest pain or other symptoms suggestive of
acute cardiac ischemia: a multicenter, controlled clinical trial. Ann
Intern Med 1998;129:845–55.
Zimmerman J, Fromm R, Meyer D, et al. Diagnostic marker cooperative study for the diagnosis of myocardial infarction. Circulation
1999;99:1671–7.
Apple FS, Murakami M, Panteghini M, et al. International survey on
the use of cardiac markers. Clin Chem 2001;47:587–8.
Newman TB, Brown AN. Use of commercial record linkage software
and vital statistics to identify patient deaths. J Am Med Inform Assoc
1997;4:233–7.