Outcomes associated with small changes in normal-range cardiac markers
Original Contribution
Outcomes associated with small changes in normal-range cardiac markers
Nolan McMullin MD a, Christopher J. Lindsell PhD b, Lei Leia, John Mafi a, Preeti Jois-Bilowich MD c, Venkataraman Anantharaman MD d,
Charles V. Pollack Jr. MA, MD e, Judd E. Hollander MD e, W. Brian Gibler MD b,
James W. Hoekestra MD f, Deborah Diercks MD g, W. Frank Peacock MD a,?
for the EMCREG i*trACS Investigators
aThe Cleveland Clinic, Cleveland, OH, USA
bThe University of Cincinatti, Cincinatti, OH, USA cUniversity of Florida, Gainesvile, FL, USA dSingapore General Hospital, Singapore eUniversity of Pennsylvania, Philadelphia, PA, USA fWake Forest University, Winston-Salem, NC, USA gUniversity of Sacramento, Davis, Davis, CA, USA
Received 10 June 2009; revised 22 July 2009; accepted 18 August 2009
Abstract
Introduction: Troponin concentrations rising above an institutional cutpoint are used to define acute Myocardial necrosis, yet it is uncertain what outcomes are associated with fluctuations in troponin that do not exceed this level. We evaluate the association between troponin fluctuations below an institutional Upper limit of normal and acute coronary syndrome (ACS).
Materials and methods: This was a post hoc analysis of the Internet tracking registry of ACS (i*trACS), which describes patients presenting to emergency departments (EDs) with suspected ACS across the spectrum of risk. Patients were included in this registry if they were at least 18 years old and had suspected ACS at the time of their ED visit. Inclusions in this analysis required that patients had at least 1 cardiac marker (creatine kinase-MB [CK-MB], troponin T, or troponin I) drawn twice within 6 hours of presentation, with both measures being below the institution’s upper limit of normal. A marker change was defined as either an increase or decrease that exceeded 15% of the institutional upper limit of normal. Acute coronary syndrome was defined as a positive stress test, documented myocardial infarction, coronary revascularization, or death within 30 days of their ED admission.
Results: Of 17 713 patient visits, 2162 met inclusion and exclusion criteria. There were 1872 patient visits with 2 troponin results and 1312 with 2 CK-MB results. Patient visits with increasing troponin had increased odds of ACS compared with those with stable troponin levels (odds ratio, 3.6; 95%
* Corresponding author. Desk E-19, Emergency Department, The Cleveland Clinic, Cleveland, OH, 44195, USA. Tel.: +1 216 445 4546; fax: +1 216 445 4552.
E-mail address: [email protected] (W.F. Peacock).
0735-6757/$ - see front matter (C) 2011 doi:10.1016/j.ajem.2009.08.016
confidence interval, 1.4-9.2). Changing CK-MB and decreasing troponin were not associated with increased odds of ACS.
Conclusions: Small increases in troponin concentration below the upper limit of normal are associated with increased odds of ACS.
(C) 2011
Introduction
The importance of cardiac markers in the diagnosis of acute myocardial infarction (MI) has been summarized by the Joint European Society of Cardiology and the American College of Cardiology consensus document, and marker elevations exceeding the institutional definition of normal are indicative of myocardial necrosis [1]. However, troponin suffers from an initial sensitivity deficit at emergency department (ED) presentation. In 1 study using an enzyme- linked immunosorbent troponin T assay (Boehringer Mann- heim) with an upper limit of normal defined as 0.2 ug/L, the sensitivity was 9.5% for predicting adverse outcomes in low- risk chest pain patients [2]. Thus, serial measurements are commonly used to improve the accuracy for detecting cardiac injury. Monitoring serial cardiac markers provides high sensitivity and specificity for acute MI [3-5] and is commonly performed in chest pain centers and ED Observation Units. This use of cardiac markers has been incorporated into a recent Clinical Policy statement by the American College of Emergency Physicians [6].
By definition, cardiac marker concentrations below an institutional cutpoint may be interpreted as “normal.” In EDs and chest pain centers, patients are frequently ruled out for MI and discharged home, provided their cardiac biomarkers are consistently below this cutpoint. However, marker changes that do not exceed the institutional definition of normal may indicate that some level of myocardial necrosis has occurred. Clinical outcomes in patients with such changes are not well described. If these patients who have detectable, but still “normal,” cardiac biomarkers are at increased risk of adverse outcomes, then further research is necessary to help define how to optimally manage these patients. Our purpose was to determine if small changes in troponin and/or creatine kinase- MB (CK-MB), below the upper limit of normal, are associated with acute or 30-day adverse events.
Materials and methods
i*trACS study design and setting
This was a post hoc analysis of the Internet tracking registry of acute coronary syndromes (i*trACS); a multicen- ter registry of 17 713 patients with possible acute coronary syndrome (ACS) presenting to 1 of 8 EDs in the United States and 1 in Singapore. i*trACS contains a Representative sample of patients with suspected ACS across the entire
spectrum of risk. Enrolling institutions and specific methods of data collection are detailed elsewhere [7]. Briefly, i*trACS is a convenient sample of patients prospectively enrolled between June 1999 and August 2001. Patients were included if they were older than 18 years and were suspected of having ACS, primarily indicated by ordering of a 12-lead electro- cardiogram (ECG) or cardiac biomarkers. Patients were excluded if they were transferred from another institution or if the ECG was obtained for non-ACS evaluations.
Data collection
Each enrolling site used standardized case report forms that included patient demographics, medical history, medica- tions, presenting symptoms, initial cardiac biomarkers, and the treating physician’s 12-lead ECG interpretation. Emer- gency department and hospital course were obtained by either chart review or daily follow-up of patients admitted to the hospital. Thirty-day outcomes were obtained by telephone interview, chart review, and social security death registry. Follow-up was completed in 95.6% of enrolled patients.
Inclusion and exclusion
For this analysis, patients were included if they had a cardiac marker (troponin I, troponin T, or CK-MB) measured twice within 6 hours of presentation, with both being below the institutional upper limit of normal. Two independent groups were defined, one evaluating changing troponin (either troponin I or troponin T) and the other for changing CK-MB. Mimicking clinical emergency medicine practice, no age, race, or sex corrections were applied to CK-MB values. The upper limit of normal was defined using the individual hospital recommended cutoff for MI [8]. For the analysis, each patient was defined by the marker cutpoint from his or her institution of enrollment.
Predictor variables
The primary predictor for this analysis was the change in cardiac marker concentrations. A change in concentration equivalent to 15% of the institutional upper limit of normal was selected as the smallest variation that would exceed all participating institution’s laboratory platform coefficient of variation [8]. Patients were stratified by the magnitude of changes from baseline: decreasing cardiac markers were
|
|
1 |
3.50 |
0.53 |
||
|
Site |
2 |
5.80 |
0.87 |
0.25 |
0.04 |
|
Site |
3 |
5.80 |
0.87 |
0.25 |
0.04 |
|
Site |
4 |
5.80 |
0.87 |
0.25 |
0.04 |
|
Site |
5 |
5.00 |
0.75 |
2.00 |
0.30 |
|
Site |
6a a |
7.00 |
1.05 |
0.20 |
0.03 |
|
Site |
6b |
5.00 |
0.75 |
1.50 |
0.23 |
|
Site |
7 |
5.00 |
0.75 |
0.20 |
0.03 |
|
Site |
8 |
8.00 |
1.20 |
2.00 |
0.30 |
|
Site |
9 |
5.00 |
0.75 |
2.00 |
0.30 |
|
a Site 6 used 2 different assays during the study period. |
|||||
defined as an absolute decline exceeding 15% of the institutional upper limit of normal, stable markers had no absolute change greater than 15% of the institutional upper limit of normal, and increasing markers were defined as an absolute increase exceeding 15% of the institutional upper limit of normal. The cutoffs and change thresholds for the different sites are shown in Table 1. For one site, the institutional upper limit of normal for troponin was a factor of 10 lower than other sites; troponin levels at this site were not included in analysis.
Table 1 Site-specific marker cutpoints, with calculation of 15% change relative to the institutional upper limit of normal, that defined moving marker results
CK-MB Troponin
Institutional Minimum Institutional Minimum upper limit change upper limit change
of normal not to be of normal not to be considered considered
stable stable
The Acute coronary ischemia-Time Insensitive Predictive Instrument (ACI-TIPI) was computed as a baseline measure of disease severity [9]. The ACI-TIPI can be interpreted as the probability of ischemia and is computed as a function of age, sex, chest pain properties, and ECG findings.
Outcome variable
Patients were considered to have an ACS if, during their hospital visit or within 30 days of the ED visit, there was positive testing, documented MI, revascularization, or death [8]. Positive testing was defined as greater than 70% stenosis in any vessel at cardiac catheterization, positive myocardial perfusion imaging, or a positive noninvasive provocative test
|
for ischemia. Revascularization was defined as a documen- |
Normal |
7.0 (0.2) |
665.0 (0.4) |
5.0 (0.2) |
|
tation of percutaneous coronary intervention with or without |
ACI-TIPI a (n = 1577) |
35.9 (19.2) |
27.6 (16.3) |
33.0 (16.6) |
|
stent placement, or a diagnostic related group (DRG) code |
ACS |
11 (29.7) |
275 (14.2) |
8 (32.0) |
|
indicating revascularization (percutaneous coronary inter- vention, coronary artery stent placement, or Coronary artery bypass grafting). Myocardial infarction was defined based on documented evidence or a DRG code of acute MI. Death included all-cause mortality. |
Data are given as means and SDs for continuous variables, and frequencies and percents for categorical variables. CAD, coronary artery disease; CHF, congestive heart failure. a The ACI-TIPI could not be computed for all subjects due to missing data. |
|||
Analysis
Data are described using means and SDs for continuous variables and frequencies and percents for categorical variables. The 95% confidence intervals (CIs) for differences between groups are given. Logistic regression was used to model the odds of ACS, and because a single patient could have multiple visits, each visit was considered to be an independent event. Data were analyzed using SPSS v 16.0 (SPSS Inc, Chicago, Ill).
Results
Of 17 713 patient visits included in i*trACS, 2623 had either 2 troponins or 2 CK-MB assays performed within 6 hours of presentation. Of these, 85 had a positive troponin, 186 had a positive CK-MB, and 164 had both a positive troponin and a positive CK-MB; these subjects have been excluded from analysis. There were 26 additional subjects with only 1 troponin at site 2; these were also excluded. Overall, 2162 patient visits were included, 1872 patient visits had 2 troponin assays (Table 2), and 1312 patient visits had 2 CK-MB assays (Table 3).
Table 2 Demographics, risk factors, and outcomes for patients with rising, falling, or stable troponin
Change in Troponin (n = 2021) Decreasing Stable Increasing
|
Age (y) |
60.4 (15.7) |
55.1 (14.6) |
56.8 (15.3) |
|
Female |
22 (59.5) |
1016 (52.3) |
15 (60.0) |
|
Male |
15 (40.5) |
925 (47.7) |
10 (40.0) |
|
White |
20 (54.1) |
822 (42.3) |
12 (48.0) |
|
African American |
7 (18.9) |
711 (36.6) |
9 (36.0) |
|
Other |
10 (27.0) |
408 (21.0) |
4 (16.0) |
|
Diabetes |
8 (25.0) |
404 (29.3) |
10 (45.5) |
|
Hypertension |
26 (81.3) |
1031 (74.8) |
17 (77.3) |
|
Hyperlipidemia |
9 (28.1) |
483 (35.1) |
6 (27.3) |
|
Angina |
11 (34.4) |
217 (15.7) |
7 (31.8) |
|
CAD |
13 (40.6) |
478 (34.7) |
8 (36.4) |
|
CHF |
2 (6.3) |
156 (11.3) |
3 (13.6) |
|
Current smoker |
5 (27.8) |
643 (46.8) |
7 (41.2) |
|
Recent smoker ECG diagnostic categor |
2 (11.1) y |
135 (9.8) |
4 (23.5) |
|
Acute MI |
0 (0.0) |
13 (0.7) |
0 (0.0) |
|
Acute ischemia |
1 (2.8) |
95 (5.2) |
2 (8.7) |
|
Early repolarization |
1 (2.8) |
47 (2.6) |
0 (0.0) |
|
Nondiagnostic |
27 (75.0) |
992 (54.7) |
16 (69.6) |
|
|
54.0 (14.6) |
57.9 |
(16.8) |
|
Female 33 (36.3) |
601 (50.6) |
18 |
(54.5) |
|
Male 58 (63.7) |
586 (49.4) |
15 |
(45.5) |
|
White 42 (46.2) |
540 (45.5) |
20 |
(60.6) |
|
African American 43 (47.3) |
515 (43.4) |
11 |
(33.3) |
|
Other 6 (6.6) |
133 (11.2) |
2 |
(6.1) |
|
Diabetes 23 (33.8) |
235 (28.6) |
13 |
(54.2) |
|
Hypertension 52 (76.5) |
609 (74.0) |
19 |
(79.2) |
|
Hyperlipidemia 26 (38.2) |
312 (37.9) |
8 |
(33.3) |
|
Angina 16 (23.5) |
160 (19.4) |
6 |
(25.0) |
|
CAD 28 (41.2) |
261 (31.7) |
10 |
(41.7) |
|
CHF 8 (11.8) |
75 (9.1) |
3 |
(12.5) |
|
Current smoker 29 (44.6) |
430 (48.0) |
6 |
(25.0) |
|
Recent smoker 8 (12.3) |
87 (9.7) |
5 |
(20.8) |
|
Acute MI 0 (0.0) |
6 (0.5) |
0 |
(0.0) |
|
Acute ischemia 4 (4.7) |
53 (4.8) |
2 |
(6.7) |
|
Early repolarization 3 (3.5) |
28 (2.5) |
0 |
(0.0) |
|
Nondiagnostic 50 (58.8) |
605 (54.9) |
20 |
(66.7) |
|
Normal 28 (32.9) |
410 (37.2) |
8 |
(26.7) |
|
ACI-TIPI a (n = 1125) 28.0 (17.0) |
27.2 (16.1 |
28.2 |
(20.1 |
|
ACS 14 (15.4) |
174 (14.6) |
8 |
(24.2) |
|
Data are given as means and SDs for continuous. CAD, coronary artery disease; CHF, congestive heart failure. a The ACI-TIPI could not be computed for all subjects due to missing data variables and frequencies and percents for categorical variables. |
|||
Table 4 presents the raw outcomes associated with changing troponin and CK-MB. Table 5 shows the odds for ACS in patients with changing markers compared with those with stable markers. Unadjusted models and models adjusted for probability of ischemia using the ACI-TIPI are shown; adjustment using the ACI-TIPI suggests that any finding for a changing marker is an effect over and above known predictors of ACS. Patients with both increasing and decreasing troponin had greater odds of ACS than patients
Table 3 Demographics, risk factors, and outcomes for patients with rising, falling, or stable CK-MB
Change in CK-MB (n=1311) Decreasing Stable Increasing
with stable troponin, regardless of adjustment. Patients with decreasing CK-MB had lower odds of ACS than those with stable CK-MB. Those with increasing CK-MB had similar odds of ACS to those with a stable CK-MB.
Although Composite outcomes were associated with increasing troponin, for patient visits with either increasing or decreasing troponin, there was not a statistically increased probability of ischemia as defined by ACI-TIPI. Compared with patients with stable troponin, the absolute difference in ACI-TIPI was 5.1% (95% CI, -4.1% to 14.4%) for those
with decreasing troponin, and 4.4% (95% CI -6.1% to
14.9%) for those with increasing troponin. There was no difference in the probability of ischemia between those with increasing troponin and those with decreasing troponin (mean difference in ACI-TIPI, -0.8%; 95% CI, -14.7% to 13.2%). Similar results were obtained for CK-MB: for those with decreasing CK-MB compared with stable CK-MB, the mean difference was 0.7% (95% CI, -3.9% to 5.3%). For those with stable CK-MB compared with those with increasing CK-MB, the mean difference was 1.0% (95% CI, -6.7% to 8.6%). For those with decreasing CK-MB compared with those with increasing CK-MB, the mean difference was 0.25% (95% CI, -8.5% to 9.0%).
Discussion
Few studies consider the association between ACS and cardiac marker fluctuations below the institutional upper limit of normal, and our findings suggest diagnostic value when these changes are noted. Our data demonstrate that small changes in troponin not exceeding the institutional upper limit of normal are associated with increased risk of ACS. Although it makes sense that a rising troponin, even below the normal limits, may be a warning of occult cardiac ischemia, a decreasing troponin did not portend ACS despite the fact that theoretically it could be the result of a resolving MI, which could also be associated with ACS [10].
It is important to note that the magnitude of change in this study was small, only 15% of the institutional presentation
value. Because the odds of adverse events are more than tripled when marker troponin was increasing, this suggests that even the smallest increases in troponin are clinically significant. Others have reported that many clinicians do not consider patients with very low to marginal Troponin elevations to be at risk of adverse events, an assumption our data refute [11-13].
Considering detectable troponin to be the product of myocyte death [1], it might be argued that any measurable troponin is likely to be associated with ACS. However in our analysis, those with increasing troponin concentrations were at increased risk for adverse events as compared with the stable or declining troponin cohort. This may suggest that in patients with dynamic troponin changes, there is a more acute pathologic process not present in patients with detectable but stable troponin levels.
Changing troponin and changing CK-MB provided different information. We found no differences between those with a stable CK-MB and those with either an increasing or decreasing CK-MB. The significance of this finding is unclear but may reflect the lower specificity of CK-MB, as compared with troponin. Alternatively, this may be the result of the shorter biologic half-life of CK-MB or the fact that at very low levels, many CK-MB assays have inadequate precision.
From a clinical perspective, the value of small troponin changes must be considered in the context of other tests. Although a normal ECG was found most often in the stable troponin population (occurring in nearly 40%; see Table 2), this should not be reassuring; acute ischemia was rarely detected by ECG, and the ECG was normal or nondiagnostic in approximately 90% of all patients. This suggests that small changes in troponin identify a cohort of patients at risk for ACS not otherwise detected by the ECG.
Several limitations are apparent that must be considered in the interpretation of our findings. Our analysis relied on registry data, the shortcomings of which have been previously described [7]. Most relevant to this study, the
data are subject to workup bias and inclusion bias. Also, although we have identified a variable that could improve risk stratification in the ED patient at risk for ACS, we are unable to comment on the effect interventions would have in the cohort of patients with small troponin changes below the institutional upper limit of normal. Whether their rate of adverse outcomes would be decreased by more aggressive ACS therapy is unknown.
It might be argued that small changes in cardiac markers can be accounted for by variations in the test performance and therefore are not clinically significant. However, we chose 15% as a cutoff point for defining fluctuations in troponin because the recommended coefficient of variation for troponin assays, as defined by the International Federation of Clinical Chemistry and Laboratory Medicine, is 10% at the 99th percentile [14]. This should mitigate the introduction of troponin changes into our data set due to testing variability as opposed to variability due to changes in an individual patient’s troponin levels. However, although we defined marker changes as being a difference greater than 15% the upper limit of normal between 2 measurements no more than 6 hours apart, smaller changes may confer clinical significance but were undetectable in our analysis.
Finally, we did not control for the effect of pathology known to be associated with increased troponin concentra- tions (eg, renal failure). It might be that it is precisely these patients that comprise the stable marker groups, albeit with preclinical disease. If this is the case, our findings are further strengthened by the ability to differentiate between those with underlying pathology not associated with a current ischemic event and those with an acute event that might benefit from intervention.
Conclusions
Any increase in troponin concentration within 6 hours of ED evaluation, irrespective of the institutional upper limit of normal, is associated with an increased risk of ACS. Although these data should not change current clinical practice, it will hopefully prompt further research into how patients, with troponin changes below the upper limit of normal, should be managed.
References
- The Joint European Society of Cardiology/American College of Cardiology Committee. 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.
- Peacock WF, Emerman CL, McErlean ES, Deluca SA, van Lente F, Rao JS, et al. Prediction of short and long term outcomes by troponin-T in low risk patients evaluated for acute coronary syndromes. Ann Emerg Med 2000;35:213-20.
- Fesmire FM, Hughes AD, Fody EP, Jackson AP, Fesmire CE, Gilbert MA, et al. The Erlanger chest pain evaluation protocol: a one-year experience with serial 12-lead ECG monitoring, two-hour delta serum marker measurements, and selective nuclear Stress testing to identify and exclude acute coronary syndromes. Ann Emerg Med 2002;40: 584-94.
- Fesmire FM. A rapid rule out protocol for identification and exclusion of acute myocardial infarction: continuous 12-lead ECG monitoring in conjunction with the 2-hour ?CK-MB. Am J Emerg Med 2000;18: 698-702.
- Hollander JE. The future of cardiac biomarkers: new concepts and emerging technologies for emergency physicians. EMCREG 2005;4:2-3.
- Fesmire FM, Decker WW, Diercks DB, Ghaemmaghami C, Nazarian D, Brady W, et al. Clinical policy: critical issues in the evaluation and management of adult patients with Non-ST-segment elevation acute coronary syndromes. Ann Emerg Med 2006;48:270-301.
- Lindsell CJ, Anantharaman V, Diercks D, Han JH, Hoekstra JW, Hollander JE, et al. The Internet Tracking Registry of Acute Coronary Syndromes (i*trACS): a multicenter Registry of patients with suspicion of ACS reported using the Standardized reporting Guidelines for Emergency Department Chest Pain Studies. Ann Emerg Med 2006;48:666-77.
- Storrow AB, Lindsell CJ, Han JH, et al. Discordant cardiac biomarkers: frequency and outcomes in emergency department patients with chest pain. Ann Emerg Med 2006:660-5.
- Storrow AB, Lindsell CJ, Han JH, Slovis CM, Miller KF, Gibler WB, et al. Performance of a population-based cardiac risk stratifi- cation tool in Asian patients with chest pain. Acad Emerg Med 2005;12:423-30.
- Newby LK, Storrow AB, Gibler WB, Garvey JL, Tucker JF, et al. Bedside multimarker testing for risk stratification in chest pain units: the chest pain evaluation by creatine kinase-MB, myoglobin, and troponin I study. Circulation 2001;103:1832-7.
- Van Lente F, McErlean ES, DeLuca SA, Peacock WF, Rao JS, Nissen SE. Ability of troponins to predict adverse outcomes in patients with renal insufficiency and Suspected acute coronary syndromes: a case- matched study. J Am Coll Cardiol 1999;33:471-8.
- Antman EM. Decision making with cardiac Troponin tests. N Engl J Med 2002;346:2079-82.
- Henrikson CA, Howell EE, Bush DE, Miles JS, Meininger GR, Friedlander T, et al. Prognostic usefulness of marginal troponin T elevation. Am J Cardiol 2004;93(3):275-9.
- Panteghini M, Gerhardt W, Apple FS, Dati F, Ravkilde J, Wu AH. Quality specifications for cardiac troponin assays. Clin Chem Lab Med 2001;39:175-9.