- Email: [email protected]
Comparison of troponin T versus creatine kinase-MB in suspected acute coronary syndromes
Comparison of Troponin T Versus Creatine Kinase-MB in Suspected Acute Coronary Syndromes Ellen S. McErlean, RN, MSN, CCRN, Sue A. Deluca, RN, BSN, Frederick van Lente, Franklin Peacock IV, MD, J. Sunil Rao, PhD, Craig A. Balog, BS, and Steven E. Nissen, MD
PhD,
Limitations of creatine kinase-MB (CK-MB) have led to alternative biochemical markers, including troponin T (TnT), to detect myocardial necrosis. Limited data are available regarding the predictive value of this new marker in patients with chest pain of uncertain etiology. Therefore, we prospectively compared CK-MB and TnT in a broad population with suspected acute coronary syndromes, including those admitted to a short-stay chest pain unit. CK-MB, quantitative TnT levels, and a rapid bedside assay were performed at 0, 4, 8, and 16 hours. Adverse events, including infarction, recurrent ischemia, coronary surgery, need for catheterization and/or intervention, stroke, congestive heart failure, or death, were identified by chart review and by follow-up phone call at 6 months. Of 707 patients, 104 were excluded for creatinine >2 mg/dl or incomplete data, leaving a total cohort of 603 patients. Coronary Care Unit admissions were 18%, intermediate care admissions were 14%, telemetry admissions is 21%, and admissions to 24-hour short-stay area were 47%. TnT (at 0.1 ng/ml) and CK-MB were positive in a similar proportion of patients (20.4% and 19.7%, respectively); however, the patients identified by TnT and CK-MB were
not identical. In-hospital adverse events occurred in 37.1% with no differences in positive predictive value for the markers (p ⴝ NS). If CK-MB and TnT were negative, the early adverse event rate was 27%. No cardiac marker was positive by 16 hours in 54.9% of patients with an adverse event. Six-month follow-up was obtained in 576 of the 603 patients (95.5%). One hundred fifty-five late adverse events occurred in 134 patients (23.3%) at an average of 3.3 ⴞ 2.5 months after discharge. If both markers were negative, the late event rate was 20.2% and did not increase in patients with positive CK-MB or TnT >0.2 ng/ml. However, the late event rate was substantially higher (52.9%) in those with intermediate TnT levels of 0.1 to 0.2 ng/ml (p ⴝ 0.002). Thus, TnT is a suitable alternative to CK-MB in patients with suspected acute coronary syndromes. The rapid bedside assay is comparable to quantitative TnT and may enable early diagnosis and triage. A negative cardiac marker value (TnT or CK-MB) does not necessarily confer a low risk of complication in patients presenting with acute chest pain to an emergency department. 䊚2000 by Excerpta Medica, Inc. (Am J Cardiol 2000;85:421– 426)
or 20 years, CK-MB isoenzymes have represented the “gold standard” for identification of myocarF dial necrosis. However, CK-MB is abnormal only in
However, most studies examined patients with a high pretest probability of acute myocardial infarction, did not use treating physicians who were blinded to the data, and followed the cohort for a short duration, typically 30 days. Accordingly, we designed a prospective study to compare CK-MB, quantitative TnT, and a qualitative TnT bedside assay in patients with acute chest pain.
patients with necrosis and requires sampling over time to confirm diagnosis. Therefore, CK-MB is of limited value for rapid triage decisions in the emergency department. CK-MB levels poorly predict long-term outcome. For example, patients with non–Q-wave acute myocardial infarction usually have lower levels of CK-MB, but a greater likelihood of recurrent ischemia. These limitations have led to development of alternative markers, including troponin T (TnT).1– 8 Some studies indicate that elevated TnT levels in unstable angina portend a higher morbidity rate.9 –12 From the Department of Cardiology, the Department of Laboratory Medicine, and the Department of Emergency Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio. This study was supported by Roche Diagnostics, Boehringer Mannheim Corporation, Indianapolis, Indiana. Manuscript received May 17, 1999; revised manuscript received and accepted September 20, 1999. Address for reprints: Ellen S. McErlean, RN, MSN, CCRN, The Cleveland Clinic Cardiovascular Coordinating Center, The Cleveland Clinic Foundation/F25, 9500 Euclid Avenue, Cleveland, Ohio 44195. E-mail: [email protected]. ©2000 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 85 February 15, 2000
METHODS
Patient population: Patients were eligible for the study if they presented to the emergency department with chest pain of suspected cardiac origin within 24 hours of onset. Patients were admitted to the coronary care unit, intermediate (stepdown) unit, telemetry, or an emergency department 24-hour short-stay unit. Exclusion criteria included cardiopulmonary resuscitation within 7 days, angioplasty or thrombolytic therapy within 3 weeks, vasopressors, serum creatinine ⬎2.0 mg/dl or long-term dialysis, or a major surgical procedure within 7 days. Study protocol: The protocol was approved by the Institutional Review Board at the Cleveland Clinic Foundation. After receiving informed consent from 0002-9149/00/$–see front matter PII S0002-9149(99)00766-3
421
duration of previously stable angina; and (3) ST depression or T-wave inIntermediate version with spontaneous pain. QElectrocardiogram wave acute myocardial infarction (n ⫽ 241) was defined as new Q waves ⬎0.04 56 (49.1%) ms or ST-segment elevation lasting 46 (48.4%) ⬎24 hours (followed by T-wave in46 (49.5%) version) with typical chest pain or a characteristic time-related increase 52 (48.2%) in CK-MB. Non–Q-wave myocardial infarction required a time-related increase in CK-MB levels and 1 of the following: (1) typical chest pain, or (2) ST-segment depression ⬎0.1 mV or an increase of ⬎0.1 mV in limb leads or ⬎0.2 mV in precordial leads; T-wave inversion; or both. A positive catheterization required ⱖ1 lesion with ⬎70% stenosis. Final discharge diagnosis was assigned by treating physicians using customary criteria. Statistical analysis: Demographics are presented as mean ⫾ SD and analyzed using a 2-sided p value (significance set at 0.05) or analysis of variance. Sensitivity and specificity were evaluated by setting up appropriate r by c contingency tables. Associations between risk factors and outcome predictions were performed using chi-square analyses. Adjustments were performed using Fisher’s exact test (to deal with small expected cell frequencies).
TABLE I Relation Between Initial Electrocardiogram and Cardiac Marker Positivity
Cardiac Marker TnT ⬎0.1 ng/ml TnT ⬎0.2 ng/ml Bedside positive TnT ⬎0.2 ng/ml CK-MB ⬎8.8 ng/ml
Positive Electrocardiogram (n ⫽ 144)
Negative Electrocardiogram (n ⫽ 143)
45 (39.5%) 39 (41.1%) 38 (40.9%)
13 (11.4%) 10 (10.5%) 9 (9.7%)
44 (40.7%)
12 (11.1%)
the patients, TnT and CK-MB samples were obtained at presentation and at 4, 8, and 16 hours. Physicians and nursing staff were blinded to all TnT results and the 4-hour CK-MB values. The baseline, 8-, and 16hour CK-MB results were reported to the clinical service. A 12-lead electrocardiogram was recorded at presentation, on admission to the hospital, and daily thereafter. Additional electrocardiograms were obtained for episodes of suspected ischemia. The electrocardiograms were interpreted in a core laboratory by a reader who was blinded to the results. Early and late follow-up: Adverse events were evaluated at 2 time points: early (in-hospital) and late (6 months after discharge). A retrospective chart review was performed by an investigator who was blinded to the data to determine adverse events, including death, catheterization and/or intervention, coronary surgery, stroke, recurrent ischemia, reinfarction, and congestive heart failure. At 6 months, nurses conducted phone interviews to determine the incidence and type of late adverse events. Analytic methods: For each sampling, 10 ml of whole blood was transported to the laboratory within 30 minutes. Within 45 minutes, 2 technologists independently interpreted qualitative TnT using an assay system with a positive threshold of 0.2 ng/ml (Cardiac T Rapid Assay, Roche Diagnostics, Boehringer Mannheim Corporation, Indianapolis, Indiana). Total CK with CK-MB mass were determined using plasma, which was frozen for later TnT quantitation. TnT level was determined using the Enzymune Cardiac T Troponin T Assay (Roche Diagnostics, Boehringer Mannheim Corporation, Indianapolis, Indiana). A total CK-MB of ⬎8.8 ng/ml was considered positive to detect patients with a normal total CK, but elevated MB fraction. Electrocardiographic analysis: Electrocardiograms were classified as positive for any of the following: new or unknown left bundle branch block; ST elevation or depression ⬎1 mm in 2 contiguous leads; or R greater than S in leads V1 or V2. A negative electrocardiogram required absence of the following: Q waves, negative T waves, ST deviation ⫾ 1 mm, new or unknown left bundle branch block, and new hemiblocks. All other abnormalities, such as T-wave inversion or nonspecific ST-T wave deviations were categorized as intermediate probability of ischemia. Patient classification: The definition of unstable angina required 1 of the following: (1) classic angina at rest; (2) abrupt increase in the frequency, severity, or 422 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
RESULTS
Patient demographics: A total of 707 patients were recruited in the emergency departments of the Cleveland Clinic or Kaiser Parma Hospital between December 1995 and January 1997. Fifty-two patients were excluded for a creatinine ⬎2.0 mg/dl and another 52 for incomplete data. The final cohort (n ⫽ 603) had a mean age of 59 years (range 23 to 98); 58% were men; 36% were African-American, 53% had history of coronary disease; 23% prior bypass surgery; and 19% previous angioplasty. The median time from chest pain to presentation was 4.7 hours. Forty-seven percent were managed in an emergency department chest pain unit, 18% in the coronary care unit, 14% in intermediate care, and 21% in a telemetry unit. Final diagnoses included Q-wave acute myocardial infarction in 7.5%, non–Q-wave myocardial infarction in 7.2%, unstable angina in 22.3%, suspected coronary syndrome in 18.8%. and non-cardiac chest pain in 44%. Electrocardiographic findings: Of the 603 patients who were evaluated, electrocardiograms synchronized with blood draws and technically adequate laboratory data were available in 532 (88%). Table I illustrates initial electrocardiographic tracing and subsequent marker positivity. All comparisons between positive, negative, and intermediate electrocardiographic groups were significant with a chi-square p value of ⬍0.001. However, no marker was positive in ⬎40% of the patients with an abnormal electrocardiogram at presentation. Conversely, about 10% of patients with an elevation of ⱖ1 markers had a normal electrocardiogram. The largest category of patients with an FEBRUARY 15, 2000
TABLE II Incidence of In-Hospital Adverse Events Event Type Any adverse event Positive catheterization Coronary intervention Recurrent ischemia Coronary bypass Congestive heart failure Reinfarction Stroke Death
TABLE III Positive Predictive Value of Cardiac Markers
No. of % % Entire Cohort Patients Complications (n ⫽ 603) 224 103 56 54 41 17
NA 46.0% 25.0% 24.0% 18.0% 8.0%
37.1 17.1 9.3 9.0 6.8 2.8
6 2 1
2.7% 0.9% 0.4%
1.0 0.3 0.2
Cardiac Marker
No. of Patients With Adverse Events (%)
No. of Patients With Positive Results
CK-MB ⬎8.8 ng/ml TnT ⬎0.1 ng/ml TnT ⬎0.2 ng/ml TnT bedside positive ⬎0.2 ng/ml TnT and CK-MB negative
119 123 103 101
87 94 83 81
455
(73.1) (76.4) (80.6) (80.2)
123 (27.0)
TABLE IV Sensitivity of Cardiac Markers
elevated marker, slightly ⬍50%, had a nondiagnostic electrocardiogram. Temporal relation of marker positivity: The time course for CK-MB and TnT positivity were similar. Of 119 patients who were eventually positive for CK-MB, 59.5% were positive at the 0-hour sample, 28.7% at 4 hours, another 8.8% became positive at 8 hours, and 2.9% were first positive at the 16-hour sample. Of 123 patients with elevated TnT values (⬎0.1 ng/ml), 58.3% were positive at the 0 hour, an additional 30.6% at 4 hours, 6.9% at 8 hours, and 4.2% at 16 hours. Statistically, none of the differences in time of positivity were significant when comparing TnT with CK-MB. Cardiac marker results: Bedside TnT was positive in the smallest number of patients, 101 (16.7%), whereas quantitative TnT was ⬎0.1 ng/ml in the largest number, 123 (20.4%). CK-MB was abnormal in 119 patients (19.7%). Many, but not all, patients positive for TnT were also positive for CK-MB and vice versa. CK-MB was positive in 25 patients in whom no TnT value was ⬎0.1 ng/ml. TnT was ⬎0.2 ng/ml in 7 patients with negative CK-MB values. TnT was ⬎0.1 ng/ml in 29 patients with negative CK-MB. Bedside TnT correlated closely with the quantitative assay (0.2 ng/ml threshold). The overall concordance between the 2 methods of measurements was 99.4%. In-hospital adverse events: A total of 224 in-hospital adverse events occurred in the 603 patients (37.1% of the cohort). The number and incidence of the complications are listed in Table II. Because individual complications are not mutually exclusive, the totals are ⬎100%. The most common adverse event was catheterization that revealed significant coronary disease, which occurred in 103 patients (46% of complications, 17.1% of the cohort). The positive predictive value of the markers for in-hospital events is shown in Table III. Differences between CK-MB and TnT were not statistically significant. The odds ratio for an adverse event in patients with an abnormal marker was approximately 9:1, differing only slightly for the various markers or thresholds. All markers showed statistically significant odds ratios with 95% confidence limits ⬎2:1 for all markers. Sensitivity of cardiac markers: Although the specificity of cardiac markers was high, the sensitivity was moderate. As illustrated in Table IV, 224 of the 603
Cardiac Marker
No. of Events (n ⫽ 224)
CK-MB ⬎8.8 ng/ml TnT ⬎0.1 ng/ml TnT ⬎0.2 ng/ml TnT bedside positive ⬎0.2 ng/ml TnT and CK-MB negative
% Events
87 94 83 81
38.8 42.0 37.1 36.2
123
54.9
TABLE V Type of Adverse Events in Cardiac Marker Negative Group Adverse Event Recurrent ischemia Cath/intervention Cath only Coronary angioplasty Congestive heart failure Coronary bypass Reinfarction Stroke Death
No. of Patients (n ⫽ 123)
% Cardiac Marker Negative Group
37 102 84 18 10
30.1 82.9 68.3 14.6 8.1
19 4 1 0
15.4 3.3 0.8 0.0
Cath ⫽ catheterization.
patients ultimately experienced an adverse in-hospital event, but no cardiac marker was positive within the first 16 hours in 123 patients (54.9%) of these patients. The types of adverse events in this group are listed in Table V. A total of 84 patients underwent catheterization and 78 had complete angiographic data and no intervention. The incidence of significant coronary disease among these 78 patients was 60.3% (n ⫽ 47). Late adverse events: Six-month follow-up was obtained in 576 of the 603 patients (95.5%). Late adverse events (n ⫽ 155) occurred in 134 patients (23.3%) at an average interval of 3.3 ⫾ 2.5 months after discharge. The types of late events are shown in Table VI. Of 18 deaths, 10 had a peak TnT level ⬎0.1 ng/ml, and 6 had a positive CK-MB (p ⫽ NS). Compared with the patients who had negative results for both markers, late adverse events were not more prevalent in patients with a positive CK-MB (30.4%) or TnT level ⬎0.2 ng/ml (27.1%) (p ⫽ NS). However, the late event rate was higher (52.9%) in those with intermediate troponin levels ranging from 0.1 to 0.2 ng/ml (p ⫽ 0.001).
CORONARY ARTERY DISEASE/TROPONIN T VS CK–MB IN CORONARY SYNDROMES
423
TABLE VI Distribution of Late Adverse Events Event Type Positive catheterization and/or coronary angioplasty Recurrent ischemia Coronary bypass Congestive heart failure Reinfarction Stroke Death
No. of Events (%) 25 (16.0) 75 13 14 5 5 18
(48.0) (8.0) (9.0) (3.0) (3.0) (12.0)
Discordant marker results: Some patients in the study had an elevated level for one, but not both markers. This discordant group was analyzed using 2 thresholds for TnT positivity: ⬎0.1 ng/ml and ⬎0.2 ng/ml. For 25 patients who were CK-MB positive, but TnT negative at the 0.1 ng/ml threshold, the in-hospital event rate was 28% (p ⫽ NS). Conversely, the early event rate was 48.3% for 29 patients with TnT ⬎0.1 ng/ml, who were CK-MB negative (p ⫽ 0.014). Eleven patients had TnT ⬎0.2 ng/ml, but negative CK-MB, with an event rate of 45% (p ⫽ NS). Twelve patients had a positive bedside TnT, but negative CK-MB, with an adverse event rate of 50% (p ⫽ NS). Late adverse events in the cohort with discordant markers were analyzed in a similar fashion. In patients with only positive CK-MB, the late event rate was 30% (p ⫽ NS). In those with TnT ⬎0.2 ng/ml, but negative CK-MB, the late event rate was 27.3% (p ⫽ NS). Finally, the late event was higher (42.9%, p ⫽ 0.012) in the group with negative CK-MB and TnT ⬎0.1 ng/ml, when compared with patients who had both negative markers.
DISCUSSION For 2 decades, CK isoenzymes have played a pivotal role in the diagnosis of acute coronary syndromes. CK-MB values are universally employed to select patients for admission and to triage patients toward more aggressive in-hospital management. Recently, the development of newer cardiac markers, such as TnT, has enabled alternative approaches to clinical decision-making in acute coronary syndromes. Despite the growing popularity of the newer markers, limited data are available regarding the sensitivity, specificity, and prognostic value of these new assays in a broad patient population. Accordingly, we correlated clinical outcomes with measured TnT and CK-MB values in a diverse patient population presenting to the emergency department with chest pain of uncertain etiology. The present study differs in several respects from other recent studies of TnT. Most investigators employed cardiac marker values to confirm the clinical impression already formed by patient history and/or electrocardiogram.8,9,11,12 We sought to examine TnT in patients presenting to the emergency department with an uncertain diagnosis, including those with a low to intermediate risk level. Our study included a substantial number (47%) of patients managed in a 424 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
24-hour short-stay unit. In this population, cardiac marker levels are commonly employed to make triage decisions. Accordingly, this population represents a more rigorous test of the clinical utility of any new marker. Several salient findings emerged from our study of TnT and CK-MB in this diverse population. Using a threshold of 0.1 ng/ml, TnT and CK-MB were positive in a similar proportion of patients, approximately 20% of the entire cohort. However, the patients identified by TnT and CK-MB were not identical—25 patients (4.1%) were only CK-MB positive, and 29 patients (4.8%) were only TnT positive. Thus, approximately 1 of every 5 positive patients was detected by only 1 of the 2 cardiac markers examined in this study. These data demonstrate that CK-MB and TnT at a threshold of 0.1 ng/ml can be used interchangeably, but that maximum sensitivity is achieved only if both markers are measured. Setting the threshold for TnT positivity at 0.2 ng/ml reduced sensitivity, although it did not improve positive predictive value. Thus, patients with TnT ⬎0.2 ng/ml were no more likely to experience an adverse in-hospital or late event than those with a TnT ⬎0.1 ng/ml. Performance of the bedside point-of-care assay was indistinguishable from quantitative TnT at a threshold of 0.2 ng/ml. TnT and CK-MB were comparable in identifying patients at high risk for in-hospital adverse events. However, no cardiac marker identified most of the patients who ultimately experienced an in-hospital adverse event. Thus, the rate of marker positivity ranged from 36% to 42% for patients who suffered an early adverse event; 55% of the group who experienced an in-hospital event were negative for all cardiac markers. Accordingly, these data do not support the application of any cardiac marker as a means to identify patients who can be discharged from the emergency department with minimal risk of adverse outcomes. After discharge, patients experienced a moderate incidence of adverse events (23.3%) and mortality (3.1%) within the first 6 months. The incidence of late events was similar in patients with positive CK-MB or TnT at a threshold of 0.2 ng/ml (30.4% and 27.1%, respectively). For either marker, the rate of late events was not statistically different from those with negative enzymes. Thus, at these thresholds, neither cardiac marker can be employed to identify patients at increased risk for late complications. However, there was a strikingly high incidence of late adverse events (55.6%) in patients with intermediate TnT levels (0.1 to 0.2 ng/ml). Compared with a TnT threshold of 0.2 ng/ml or a positive CK-MB, an intermediate TnT value conferred a statistically significant increase in risk of late events (p ⫽ 0.001). It seems likely that intermediate levels of TnT (0.1 to 0.2 ng/ml) can identify patients with incomplete myocardial necrosis, thereby identifying a subgroup with additional myocardium at risk. Accordingly, patients with TnT of 0.1 to 0.2 ng/ml should be considered candidates for more aggressive in-hospital and/or postdischarge management. FEBRUARY 15, 2000
The temporal relation between marker positivity and time from symptom onset is a critical question in determining the role of any cardiac marker in early triage decisions. In the present study, the median time from chest pain onset to presentation at the emergency department was 4.7 hours. TnT and CK-MB were initially positive in the first sample for approximately 60% of the patients, with an additional 30% becoming positive at the 4-hour sample. The 8-hour blood sample only captured an additional 7% to 8%, and a very small number (3% to 4%) became positive initially only at 16 hours. With respect to the time course of positivity, TnT did not offer any advantage over CKMB, a finding that confirms previous observations.13 Thus, the addition of TnT to a battery of cardiac markers should not be used to justify a shorter length of stay for patients at risk. Both quantitative TnT and a rapid bedside assay were evaluated in this study. The bedside assay was essentially comparable in performance to the quantitative assay at a threshold of 0.2 ng/ml. Although there was a similarity, there are important advantages of a bedside, point-of-care testing system. A binary positive or negative value is available within 20 minutes, which is a turnaround time unlikely to be matched by a typical busy hospital laboratory. However, there were also important disadvantages of the version of the bedside test examined in our study, particularly the threshold of 0.2 ng/ml, rather than 0.1 ng/ml. Commercial products with a lower threshold for the bedside qualitative assay are now available. Correlation with electrocardiogram: In our study, an initial negative or nondiagnostic electrocardiogram was present in more than half of the patients in whom ⱖ1 cardiac marker ultimately became positive. Conversely, no marker was positive in ⬎1⁄3 of the patients who had an electrocardiogram that was consistent with an acute coronary syndrome. The dissociation between the initial electrocardiogram and subsequent CK-MB levels in acute coronary syndromes is well described in the literature.1,2,14,15 These findings emphasize the limitations of a single electrocardiogram “snapshot” in predicting those patients who will ultimately receive a diagnosis of an acute coronary syndrome. Other studies: The present study contrasts with recent investigations suggesting that a negative TnT confers a sufficiently low risk to enable discharge from the emergency department.16 None of the markers successfully predicted ⬎42% of subsequent inhospital adverse events. This finding is consistent with previous observations that the long-term outcome of patients who “ruled out” for myocardial infarction was relatively poor.17–19 In contrast, Hamm et al16 found a very low adverse event rate in patients with negative TnT (1.1% at 30 days) and in patients without STsegment elevation who presented with chest pain within 12 hours. However, practitioners in the Hamm study were not blinded to the troponin results and adverse events were limited to death and nonfatal myocardial infarction within 30 days. We expanded the definition of adverse events to include recurrent
ischemia, significant coronary disease by catheterization, and need for revascularization because these events are associated with morbidity and require increased health care resources. We followed patients for a longer duration (6 months) with adverse events occurring an average of 3.3 ⫾ 2.5 months after discharge. Several limitations of our study require discussion. Although physicians responsible for the care of patients were blinded to TnT values, they had access to the standard CK-MB results at 0, 8, and 16 hours, which were routinely used for decision-making. Thus, CK-MB played a role in management of these patients, whereas TnT results were not available for this purpose. Physicians often used CK-MB results to provide a rationale for catheterization, an affect not possible for TnT values. Therefore, the results may be skewed toward better performance for CK-MB over TnT as a predictor of outcome. These findings have significant implications for use of TnT in clinical practice. Serial CK-MB and/or TnT levels should be obtained for a minimum of 12 hours after presentation. If positive, either marker portends a threefold increase in risk for an in-hospital adverse event. CK-MB and TnT are complimentary in identifying patients at risk for early complications—about 20% of patients are positive for only 1 marker. TnT at the 0.1 ng/ml threshold offers additional information over the previously used 0.2 ng/ml threshold and is more closely comparable to CK-MB. Patients with TnT values between 0.1 to 0.2 ng/ml are at particularly high risk and warrant aggressive early management and careful follow-up. The bedside TnT test provides improved turnaround times, but will require a lower threshold (0.1 ng/ml) for optimal sensitivity. Importantly, a negative TnT value does not confer a low risk of an adverse outcome. Accordingly, rapid “rule out” within the emergency department is not feasible using currently available methods. Finally, this study illustrates the importance of prospective evaluation of newer cardiac markers before accepting these techniques into clinical practice algorithms. 1. Gibler WB, Lewis LM, Erb RE, Makens PK, Kaplan BC, Vaughn RH, Biagini
AV, Blanton JD, Campbell WB. Early detection of acute myocardial infarction in patients presenting with chest pain and non diagnostic ECGs: serial CK-MB sampling in the emergency department. Ann Emerg Med 1990;19:1359 –1366. 2. Lee TH, Weisberg MC, Cook EF, Daley K, Brand DA, Goldman L. Evaluation of creatine kinase and creatine kinase MB for dagnosing myocardial infarction: clinical impact in the emergency room. Arch Intern Med 1987;147:115–121. 3. Ravikilde J, Horder M, Gerhardt W, Ljungdahl L, Petterson T, Tryding N, Moller BH, Hamfelt A, Graven T, Asburg A, Helin M, Penttila I, Thygesen K. Diagnostic performance and prognostic value of serum troponin T in suspected acute myocardial infarction. Scand J Clin Lab Invest 1993;53:677– 685. 4. Katus HA, Remppis A, Scheffold T, Diederich KW, Kuebler W. Intracellular compartmentation of cardiac troponin-T and its release kinetics in patients with reperfused and nonreperfused myocardial infarction. Am J Cardiol 1991;67: 1360 –1367. 5. Katus HA, Remppis A, Neumann FJ, Diedrich KW, Kuebler W. Diagnostic efficiency of troponin-T measurements in acute myocardial infarction. Circulation 1991;83:902–912. 6. Bakker AJ, Koelemay MW, Gorgels JP, van Vlies B, Smits R, Tijssen JG, Haagen FD. Troponin-T and myoglobin at admission: value of early diagnosis of acute myocardial infarction. Eur Heart J 1994;15:45–53. 7. Wu AHB, Valdes R Jr, Apple FS, Gornet T, Stone MA, Mayfield-Stokes S, Ingersoll-Stroubos AM, Wiler B. Cardiac troponin-T immunoassay for diagnosis of acute myocardial infarction. Clin Chem 1994;40:900 –907.
CORONARY ARTERY DISEASE/TROPONIN T VS CK–MB IN CORONARY SYNDROMES
425
8. Antman EM, Grudzien C, Sacks DB. Evaluation of a rapid bedside assay for detection of serum cardiac troponin T. JAMA 1995;273:1279 –1282. 9. Hamm CW, Ravkilde J, Gerhardt W, Jorgensen P, Peheim E, Ljungdahl L, Godmann B, Katus HA. The prognostic value of serum troponin T in unstable angina. N Engl J Med 1992;327:146 –150. 10. Seino Y, Tomita Y, Takano T, Hayakawa H. Early identification of cardiac events with serum troponin T in patients with unstable angina. Lancet 1993;342: 1236 –7. 11. Lindahl B, Venge P, Wallentin L. Relation between troponin T and the risk of subsequent cardiac events in unstable coronary artery disease. Circulation 1996;93:1651–1657. 12. Magnus Ohman E, Armstrong PW, Christenson RH, Granger CB, Katus HA, Hamm CW, O’Hanesian MA, Wagner GS, Kleiman NS, Harrell FE Jr, et al. Cardiac troponin t levels for risk stratification in acute myocardial ischemia. N Engl J Med 1996;335:1333–1341. 13. Mair J, Morandell D, Genser N, Lechleiter P, Dienstl F, Puschendorf B. Equivalent early sensitivities of myoglobin, creatine kinase MB mass, creatine kinase isoform ratios, and cardiac troponins I and T for acute myocardial infarction. Clin Chem 1995;41:1266 –1272. 14. Rouan GW, Lee TH, Cook EF, Brand DA, Weisberg MC, Goldman L.
426 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
Clinical characteristics and outcome of acute myocardial infarction patients with initially normal or nonspecific electrocardiograms (report from the Multicenter Chest Pain Study). Am J Cardiol 1989;64:1087–1092. 15. Rude RE, Poole WK, Muller JE, Turi Z, Rutherford J, Parker C, Roberts R, Rabbe DS, Gold HK, Stone PH. Electrocardiographic and clinical criteria for recognition of acute myocardial infarction based on analysis of 3,697 patients. Am J Cardiol 1983;52:936 –942. 16. Hamm CW, Goldmann BU, Heeschen C, Kreymann G, Berger J, Meinertz T. Emergency room triage of patients with acute chest pain by means of rapid testing for cardiac troponin T or troponin I. N Engl J Med 1997;337:1648 –1653. 17. Karlson BW, Herlitz J, Richter A, Sjolin M, Hjalmarson A. Prognosis in patients with ST-T changes but no rise in serum activity as compared with non-Q-wave infarction. Cardiol 1991;79:271–279. 18. Schroeder JS, Lamb KH, Hu M. Do patients in whom myocardial infarction has been ruled out have a better prognosis after hospitalization than those surviving infarction? N Engl J Med 1980;303:1–5. 19. Bakker AJ, Koelemay MJ, Gorgels JP, van VliesB, Smits R, Tijssen JG, Haagen FD. Failure of new biochemical markers to exclude acute myocardial infarction at admission. Lancet 1993;342:1220 –1222.
FEBRUARY 15, 2000
PhD,
Limitations of creatine kinase-MB (CK-MB) have led to alternative biochemical markers, including troponin T (TnT), to detect myocardial necrosis. Limited data are available regarding the predictive value of this new marker in patients with chest pain of uncertain etiology. Therefore, we prospectively compared CK-MB and TnT in a broad population with suspected acute coronary syndromes, including those admitted to a short-stay chest pain unit. CK-MB, quantitative TnT levels, and a rapid bedside assay were performed at 0, 4, 8, and 16 hours. Adverse events, including infarction, recurrent ischemia, coronary surgery, need for catheterization and/or intervention, stroke, congestive heart failure, or death, were identified by chart review and by follow-up phone call at 6 months. Of 707 patients, 104 were excluded for creatinine >2 mg/dl or incomplete data, leaving a total cohort of 603 patients. Coronary Care Unit admissions were 18%, intermediate care admissions were 14%, telemetry admissions is 21%, and admissions to 24-hour short-stay area were 47%. TnT (at 0.1 ng/ml) and CK-MB were positive in a similar proportion of patients (20.4% and 19.7%, respectively); however, the patients identified by TnT and CK-MB were
not identical. In-hospital adverse events occurred in 37.1% with no differences in positive predictive value for the markers (p ⴝ NS). If CK-MB and TnT were negative, the early adverse event rate was 27%. No cardiac marker was positive by 16 hours in 54.9% of patients with an adverse event. Six-month follow-up was obtained in 576 of the 603 patients (95.5%). One hundred fifty-five late adverse events occurred in 134 patients (23.3%) at an average of 3.3 ⴞ 2.5 months after discharge. If both markers were negative, the late event rate was 20.2% and did not increase in patients with positive CK-MB or TnT >0.2 ng/ml. However, the late event rate was substantially higher (52.9%) in those with intermediate TnT levels of 0.1 to 0.2 ng/ml (p ⴝ 0.002). Thus, TnT is a suitable alternative to CK-MB in patients with suspected acute coronary syndromes. The rapid bedside assay is comparable to quantitative TnT and may enable early diagnosis and triage. A negative cardiac marker value (TnT or CK-MB) does not necessarily confer a low risk of complication in patients presenting with acute chest pain to an emergency department. 䊚2000 by Excerpta Medica, Inc. (Am J Cardiol 2000;85:421– 426)
or 20 years, CK-MB isoenzymes have represented the “gold standard” for identification of myocarF dial necrosis. However, CK-MB is abnormal only in
However, most studies examined patients with a high pretest probability of acute myocardial infarction, did not use treating physicians who were blinded to the data, and followed the cohort for a short duration, typically 30 days. Accordingly, we designed a prospective study to compare CK-MB, quantitative TnT, and a qualitative TnT bedside assay in patients with acute chest pain.
patients with necrosis and requires sampling over time to confirm diagnosis. Therefore, CK-MB is of limited value for rapid triage decisions in the emergency department. CK-MB levels poorly predict long-term outcome. For example, patients with non–Q-wave acute myocardial infarction usually have lower levels of CK-MB, but a greater likelihood of recurrent ischemia. These limitations have led to development of alternative markers, including troponin T (TnT).1– 8 Some studies indicate that elevated TnT levels in unstable angina portend a higher morbidity rate.9 –12 From the Department of Cardiology, the Department of Laboratory Medicine, and the Department of Emergency Medicine, The Cleveland Clinic Foundation, Cleveland, Ohio. This study was supported by Roche Diagnostics, Boehringer Mannheim Corporation, Indianapolis, Indiana. Manuscript received May 17, 1999; revised manuscript received and accepted September 20, 1999. Address for reprints: Ellen S. McErlean, RN, MSN, CCRN, The Cleveland Clinic Cardiovascular Coordinating Center, The Cleveland Clinic Foundation/F25, 9500 Euclid Avenue, Cleveland, Ohio 44195. E-mail: [email protected]. ©2000 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 85 February 15, 2000
METHODS
Patient population: Patients were eligible for the study if they presented to the emergency department with chest pain of suspected cardiac origin within 24 hours of onset. Patients were admitted to the coronary care unit, intermediate (stepdown) unit, telemetry, or an emergency department 24-hour short-stay unit. Exclusion criteria included cardiopulmonary resuscitation within 7 days, angioplasty or thrombolytic therapy within 3 weeks, vasopressors, serum creatinine ⬎2.0 mg/dl or long-term dialysis, or a major surgical procedure within 7 days. Study protocol: The protocol was approved by the Institutional Review Board at the Cleveland Clinic Foundation. After receiving informed consent from 0002-9149/00/$–see front matter PII S0002-9149(99)00766-3
421
duration of previously stable angina; and (3) ST depression or T-wave inIntermediate version with spontaneous pain. QElectrocardiogram wave acute myocardial infarction (n ⫽ 241) was defined as new Q waves ⬎0.04 56 (49.1%) ms or ST-segment elevation lasting 46 (48.4%) ⬎24 hours (followed by T-wave in46 (49.5%) version) with typical chest pain or a characteristic time-related increase 52 (48.2%) in CK-MB. Non–Q-wave myocardial infarction required a time-related increase in CK-MB levels and 1 of the following: (1) typical chest pain, or (2) ST-segment depression ⬎0.1 mV or an increase of ⬎0.1 mV in limb leads or ⬎0.2 mV in precordial leads; T-wave inversion; or both. A positive catheterization required ⱖ1 lesion with ⬎70% stenosis. Final discharge diagnosis was assigned by treating physicians using customary criteria. Statistical analysis: Demographics are presented as mean ⫾ SD and analyzed using a 2-sided p value (significance set at 0.05) or analysis of variance. Sensitivity and specificity were evaluated by setting up appropriate r by c contingency tables. Associations between risk factors and outcome predictions were performed using chi-square analyses. Adjustments were performed using Fisher’s exact test (to deal with small expected cell frequencies).
TABLE I Relation Between Initial Electrocardiogram and Cardiac Marker Positivity
Cardiac Marker TnT ⬎0.1 ng/ml TnT ⬎0.2 ng/ml Bedside positive TnT ⬎0.2 ng/ml CK-MB ⬎8.8 ng/ml
Positive Electrocardiogram (n ⫽ 144)
Negative Electrocardiogram (n ⫽ 143)
45 (39.5%) 39 (41.1%) 38 (40.9%)
13 (11.4%) 10 (10.5%) 9 (9.7%)
44 (40.7%)
12 (11.1%)
the patients, TnT and CK-MB samples were obtained at presentation and at 4, 8, and 16 hours. Physicians and nursing staff were blinded to all TnT results and the 4-hour CK-MB values. The baseline, 8-, and 16hour CK-MB results were reported to the clinical service. A 12-lead electrocardiogram was recorded at presentation, on admission to the hospital, and daily thereafter. Additional electrocardiograms were obtained for episodes of suspected ischemia. The electrocardiograms were interpreted in a core laboratory by a reader who was blinded to the results. Early and late follow-up: Adverse events were evaluated at 2 time points: early (in-hospital) and late (6 months after discharge). A retrospective chart review was performed by an investigator who was blinded to the data to determine adverse events, including death, catheterization and/or intervention, coronary surgery, stroke, recurrent ischemia, reinfarction, and congestive heart failure. At 6 months, nurses conducted phone interviews to determine the incidence and type of late adverse events. Analytic methods: For each sampling, 10 ml of whole blood was transported to the laboratory within 30 minutes. Within 45 minutes, 2 technologists independently interpreted qualitative TnT using an assay system with a positive threshold of 0.2 ng/ml (Cardiac T Rapid Assay, Roche Diagnostics, Boehringer Mannheim Corporation, Indianapolis, Indiana). Total CK with CK-MB mass were determined using plasma, which was frozen for later TnT quantitation. TnT level was determined using the Enzymune Cardiac T Troponin T Assay (Roche Diagnostics, Boehringer Mannheim Corporation, Indianapolis, Indiana). A total CK-MB of ⬎8.8 ng/ml was considered positive to detect patients with a normal total CK, but elevated MB fraction. Electrocardiographic analysis: Electrocardiograms were classified as positive for any of the following: new or unknown left bundle branch block; ST elevation or depression ⬎1 mm in 2 contiguous leads; or R greater than S in leads V1 or V2. A negative electrocardiogram required absence of the following: Q waves, negative T waves, ST deviation ⫾ 1 mm, new or unknown left bundle branch block, and new hemiblocks. All other abnormalities, such as T-wave inversion or nonspecific ST-T wave deviations were categorized as intermediate probability of ischemia. Patient classification: The definition of unstable angina required 1 of the following: (1) classic angina at rest; (2) abrupt increase in the frequency, severity, or 422 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
RESULTS
Patient demographics: A total of 707 patients were recruited in the emergency departments of the Cleveland Clinic or Kaiser Parma Hospital between December 1995 and January 1997. Fifty-two patients were excluded for a creatinine ⬎2.0 mg/dl and another 52 for incomplete data. The final cohort (n ⫽ 603) had a mean age of 59 years (range 23 to 98); 58% were men; 36% were African-American, 53% had history of coronary disease; 23% prior bypass surgery; and 19% previous angioplasty. The median time from chest pain to presentation was 4.7 hours. Forty-seven percent were managed in an emergency department chest pain unit, 18% in the coronary care unit, 14% in intermediate care, and 21% in a telemetry unit. Final diagnoses included Q-wave acute myocardial infarction in 7.5%, non–Q-wave myocardial infarction in 7.2%, unstable angina in 22.3%, suspected coronary syndrome in 18.8%. and non-cardiac chest pain in 44%. Electrocardiographic findings: Of the 603 patients who were evaluated, electrocardiograms synchronized with blood draws and technically adequate laboratory data were available in 532 (88%). Table I illustrates initial electrocardiographic tracing and subsequent marker positivity. All comparisons between positive, negative, and intermediate electrocardiographic groups were significant with a chi-square p value of ⬍0.001. However, no marker was positive in ⬎40% of the patients with an abnormal electrocardiogram at presentation. Conversely, about 10% of patients with an elevation of ⱖ1 markers had a normal electrocardiogram. The largest category of patients with an FEBRUARY 15, 2000
TABLE II Incidence of In-Hospital Adverse Events Event Type Any adverse event Positive catheterization Coronary intervention Recurrent ischemia Coronary bypass Congestive heart failure Reinfarction Stroke Death
TABLE III Positive Predictive Value of Cardiac Markers
No. of % % Entire Cohort Patients Complications (n ⫽ 603) 224 103 56 54 41 17
NA 46.0% 25.0% 24.0% 18.0% 8.0%
37.1 17.1 9.3 9.0 6.8 2.8
6 2 1
2.7% 0.9% 0.4%
1.0 0.3 0.2
Cardiac Marker
No. of Patients With Adverse Events (%)
No. of Patients With Positive Results
CK-MB ⬎8.8 ng/ml TnT ⬎0.1 ng/ml TnT ⬎0.2 ng/ml TnT bedside positive ⬎0.2 ng/ml TnT and CK-MB negative
119 123 103 101
87 94 83 81
455
(73.1) (76.4) (80.6) (80.2)
123 (27.0)
TABLE IV Sensitivity of Cardiac Markers
elevated marker, slightly ⬍50%, had a nondiagnostic electrocardiogram. Temporal relation of marker positivity: The time course for CK-MB and TnT positivity were similar. Of 119 patients who were eventually positive for CK-MB, 59.5% were positive at the 0-hour sample, 28.7% at 4 hours, another 8.8% became positive at 8 hours, and 2.9% were first positive at the 16-hour sample. Of 123 patients with elevated TnT values (⬎0.1 ng/ml), 58.3% were positive at the 0 hour, an additional 30.6% at 4 hours, 6.9% at 8 hours, and 4.2% at 16 hours. Statistically, none of the differences in time of positivity were significant when comparing TnT with CK-MB. Cardiac marker results: Bedside TnT was positive in the smallest number of patients, 101 (16.7%), whereas quantitative TnT was ⬎0.1 ng/ml in the largest number, 123 (20.4%). CK-MB was abnormal in 119 patients (19.7%). Many, but not all, patients positive for TnT were also positive for CK-MB and vice versa. CK-MB was positive in 25 patients in whom no TnT value was ⬎0.1 ng/ml. TnT was ⬎0.2 ng/ml in 7 patients with negative CK-MB values. TnT was ⬎0.1 ng/ml in 29 patients with negative CK-MB. Bedside TnT correlated closely with the quantitative assay (0.2 ng/ml threshold). The overall concordance between the 2 methods of measurements was 99.4%. In-hospital adverse events: A total of 224 in-hospital adverse events occurred in the 603 patients (37.1% of the cohort). The number and incidence of the complications are listed in Table II. Because individual complications are not mutually exclusive, the totals are ⬎100%. The most common adverse event was catheterization that revealed significant coronary disease, which occurred in 103 patients (46% of complications, 17.1% of the cohort). The positive predictive value of the markers for in-hospital events is shown in Table III. Differences between CK-MB and TnT were not statistically significant. The odds ratio for an adverse event in patients with an abnormal marker was approximately 9:1, differing only slightly for the various markers or thresholds. All markers showed statistically significant odds ratios with 95% confidence limits ⬎2:1 for all markers. Sensitivity of cardiac markers: Although the specificity of cardiac markers was high, the sensitivity was moderate. As illustrated in Table IV, 224 of the 603
Cardiac Marker
No. of Events (n ⫽ 224)
CK-MB ⬎8.8 ng/ml TnT ⬎0.1 ng/ml TnT ⬎0.2 ng/ml TnT bedside positive ⬎0.2 ng/ml TnT and CK-MB negative
% Events
87 94 83 81
38.8 42.0 37.1 36.2
123
54.9
TABLE V Type of Adverse Events in Cardiac Marker Negative Group Adverse Event Recurrent ischemia Cath/intervention Cath only Coronary angioplasty Congestive heart failure Coronary bypass Reinfarction Stroke Death
No. of Patients (n ⫽ 123)
% Cardiac Marker Negative Group
37 102 84 18 10
30.1 82.9 68.3 14.6 8.1
19 4 1 0
15.4 3.3 0.8 0.0
Cath ⫽ catheterization.
patients ultimately experienced an adverse in-hospital event, but no cardiac marker was positive within the first 16 hours in 123 patients (54.9%) of these patients. The types of adverse events in this group are listed in Table V. A total of 84 patients underwent catheterization and 78 had complete angiographic data and no intervention. The incidence of significant coronary disease among these 78 patients was 60.3% (n ⫽ 47). Late adverse events: Six-month follow-up was obtained in 576 of the 603 patients (95.5%). Late adverse events (n ⫽ 155) occurred in 134 patients (23.3%) at an average interval of 3.3 ⫾ 2.5 months after discharge. The types of late events are shown in Table VI. Of 18 deaths, 10 had a peak TnT level ⬎0.1 ng/ml, and 6 had a positive CK-MB (p ⫽ NS). Compared with the patients who had negative results for both markers, late adverse events were not more prevalent in patients with a positive CK-MB (30.4%) or TnT level ⬎0.2 ng/ml (27.1%) (p ⫽ NS). However, the late event rate was higher (52.9%) in those with intermediate troponin levels ranging from 0.1 to 0.2 ng/ml (p ⫽ 0.001).
CORONARY ARTERY DISEASE/TROPONIN T VS CK–MB IN CORONARY SYNDROMES
423
TABLE VI Distribution of Late Adverse Events Event Type Positive catheterization and/or coronary angioplasty Recurrent ischemia Coronary bypass Congestive heart failure Reinfarction Stroke Death
No. of Events (%) 25 (16.0) 75 13 14 5 5 18
(48.0) (8.0) (9.0) (3.0) (3.0) (12.0)
Discordant marker results: Some patients in the study had an elevated level for one, but not both markers. This discordant group was analyzed using 2 thresholds for TnT positivity: ⬎0.1 ng/ml and ⬎0.2 ng/ml. For 25 patients who were CK-MB positive, but TnT negative at the 0.1 ng/ml threshold, the in-hospital event rate was 28% (p ⫽ NS). Conversely, the early event rate was 48.3% for 29 patients with TnT ⬎0.1 ng/ml, who were CK-MB negative (p ⫽ 0.014). Eleven patients had TnT ⬎0.2 ng/ml, but negative CK-MB, with an event rate of 45% (p ⫽ NS). Twelve patients had a positive bedside TnT, but negative CK-MB, with an adverse event rate of 50% (p ⫽ NS). Late adverse events in the cohort with discordant markers were analyzed in a similar fashion. In patients with only positive CK-MB, the late event rate was 30% (p ⫽ NS). In those with TnT ⬎0.2 ng/ml, but negative CK-MB, the late event rate was 27.3% (p ⫽ NS). Finally, the late event was higher (42.9%, p ⫽ 0.012) in the group with negative CK-MB and TnT ⬎0.1 ng/ml, when compared with patients who had both negative markers.
DISCUSSION For 2 decades, CK isoenzymes have played a pivotal role in the diagnosis of acute coronary syndromes. CK-MB values are universally employed to select patients for admission and to triage patients toward more aggressive in-hospital management. Recently, the development of newer cardiac markers, such as TnT, has enabled alternative approaches to clinical decision-making in acute coronary syndromes. Despite the growing popularity of the newer markers, limited data are available regarding the sensitivity, specificity, and prognostic value of these new assays in a broad patient population. Accordingly, we correlated clinical outcomes with measured TnT and CK-MB values in a diverse patient population presenting to the emergency department with chest pain of uncertain etiology. The present study differs in several respects from other recent studies of TnT. Most investigators employed cardiac marker values to confirm the clinical impression already formed by patient history and/or electrocardiogram.8,9,11,12 We sought to examine TnT in patients presenting to the emergency department with an uncertain diagnosis, including those with a low to intermediate risk level. Our study included a substantial number (47%) of patients managed in a 424 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
24-hour short-stay unit. In this population, cardiac marker levels are commonly employed to make triage decisions. Accordingly, this population represents a more rigorous test of the clinical utility of any new marker. Several salient findings emerged from our study of TnT and CK-MB in this diverse population. Using a threshold of 0.1 ng/ml, TnT and CK-MB were positive in a similar proportion of patients, approximately 20% of the entire cohort. However, the patients identified by TnT and CK-MB were not identical—25 patients (4.1%) were only CK-MB positive, and 29 patients (4.8%) were only TnT positive. Thus, approximately 1 of every 5 positive patients was detected by only 1 of the 2 cardiac markers examined in this study. These data demonstrate that CK-MB and TnT at a threshold of 0.1 ng/ml can be used interchangeably, but that maximum sensitivity is achieved only if both markers are measured. Setting the threshold for TnT positivity at 0.2 ng/ml reduced sensitivity, although it did not improve positive predictive value. Thus, patients with TnT ⬎0.2 ng/ml were no more likely to experience an adverse in-hospital or late event than those with a TnT ⬎0.1 ng/ml. Performance of the bedside point-of-care assay was indistinguishable from quantitative TnT at a threshold of 0.2 ng/ml. TnT and CK-MB were comparable in identifying patients at high risk for in-hospital adverse events. However, no cardiac marker identified most of the patients who ultimately experienced an in-hospital adverse event. Thus, the rate of marker positivity ranged from 36% to 42% for patients who suffered an early adverse event; 55% of the group who experienced an in-hospital event were negative for all cardiac markers. Accordingly, these data do not support the application of any cardiac marker as a means to identify patients who can be discharged from the emergency department with minimal risk of adverse outcomes. After discharge, patients experienced a moderate incidence of adverse events (23.3%) and mortality (3.1%) within the first 6 months. The incidence of late events was similar in patients with positive CK-MB or TnT at a threshold of 0.2 ng/ml (30.4% and 27.1%, respectively). For either marker, the rate of late events was not statistically different from those with negative enzymes. Thus, at these thresholds, neither cardiac marker can be employed to identify patients at increased risk for late complications. However, there was a strikingly high incidence of late adverse events (55.6%) in patients with intermediate TnT levels (0.1 to 0.2 ng/ml). Compared with a TnT threshold of 0.2 ng/ml or a positive CK-MB, an intermediate TnT value conferred a statistically significant increase in risk of late events (p ⫽ 0.001). It seems likely that intermediate levels of TnT (0.1 to 0.2 ng/ml) can identify patients with incomplete myocardial necrosis, thereby identifying a subgroup with additional myocardium at risk. Accordingly, patients with TnT of 0.1 to 0.2 ng/ml should be considered candidates for more aggressive in-hospital and/or postdischarge management. FEBRUARY 15, 2000
The temporal relation between marker positivity and time from symptom onset is a critical question in determining the role of any cardiac marker in early triage decisions. In the present study, the median time from chest pain onset to presentation at the emergency department was 4.7 hours. TnT and CK-MB were initially positive in the first sample for approximately 60% of the patients, with an additional 30% becoming positive at the 4-hour sample. The 8-hour blood sample only captured an additional 7% to 8%, and a very small number (3% to 4%) became positive initially only at 16 hours. With respect to the time course of positivity, TnT did not offer any advantage over CKMB, a finding that confirms previous observations.13 Thus, the addition of TnT to a battery of cardiac markers should not be used to justify a shorter length of stay for patients at risk. Both quantitative TnT and a rapid bedside assay were evaluated in this study. The bedside assay was essentially comparable in performance to the quantitative assay at a threshold of 0.2 ng/ml. Although there was a similarity, there are important advantages of a bedside, point-of-care testing system. A binary positive or negative value is available within 20 minutes, which is a turnaround time unlikely to be matched by a typical busy hospital laboratory. However, there were also important disadvantages of the version of the bedside test examined in our study, particularly the threshold of 0.2 ng/ml, rather than 0.1 ng/ml. Commercial products with a lower threshold for the bedside qualitative assay are now available. Correlation with electrocardiogram: In our study, an initial negative or nondiagnostic electrocardiogram was present in more than half of the patients in whom ⱖ1 cardiac marker ultimately became positive. Conversely, no marker was positive in ⬎1⁄3 of the patients who had an electrocardiogram that was consistent with an acute coronary syndrome. The dissociation between the initial electrocardiogram and subsequent CK-MB levels in acute coronary syndromes is well described in the literature.1,2,14,15 These findings emphasize the limitations of a single electrocardiogram “snapshot” in predicting those patients who will ultimately receive a diagnosis of an acute coronary syndrome. Other studies: The present study contrasts with recent investigations suggesting that a negative TnT confers a sufficiently low risk to enable discharge from the emergency department.16 None of the markers successfully predicted ⬎42% of subsequent inhospital adverse events. This finding is consistent with previous observations that the long-term outcome of patients who “ruled out” for myocardial infarction was relatively poor.17–19 In contrast, Hamm et al16 found a very low adverse event rate in patients with negative TnT (1.1% at 30 days) and in patients without STsegment elevation who presented with chest pain within 12 hours. However, practitioners in the Hamm study were not blinded to the troponin results and adverse events were limited to death and nonfatal myocardial infarction within 30 days. We expanded the definition of adverse events to include recurrent
ischemia, significant coronary disease by catheterization, and need for revascularization because these events are associated with morbidity and require increased health care resources. We followed patients for a longer duration (6 months) with adverse events occurring an average of 3.3 ⫾ 2.5 months after discharge. Several limitations of our study require discussion. Although physicians responsible for the care of patients were blinded to TnT values, they had access to the standard CK-MB results at 0, 8, and 16 hours, which were routinely used for decision-making. Thus, CK-MB played a role in management of these patients, whereas TnT results were not available for this purpose. Physicians often used CK-MB results to provide a rationale for catheterization, an affect not possible for TnT values. Therefore, the results may be skewed toward better performance for CK-MB over TnT as a predictor of outcome. These findings have significant implications for use of TnT in clinical practice. Serial CK-MB and/or TnT levels should be obtained for a minimum of 12 hours after presentation. If positive, either marker portends a threefold increase in risk for an in-hospital adverse event. CK-MB and TnT are complimentary in identifying patients at risk for early complications—about 20% of patients are positive for only 1 marker. TnT at the 0.1 ng/ml threshold offers additional information over the previously used 0.2 ng/ml threshold and is more closely comparable to CK-MB. Patients with TnT values between 0.1 to 0.2 ng/ml are at particularly high risk and warrant aggressive early management and careful follow-up. The bedside TnT test provides improved turnaround times, but will require a lower threshold (0.1 ng/ml) for optimal sensitivity. Importantly, a negative TnT value does not confer a low risk of an adverse outcome. Accordingly, rapid “rule out” within the emergency department is not feasible using currently available methods. Finally, this study illustrates the importance of prospective evaluation of newer cardiac markers before accepting these techniques into clinical practice algorithms. 1. Gibler WB, Lewis LM, Erb RE, Makens PK, Kaplan BC, Vaughn RH, Biagini
AV, Blanton JD, Campbell WB. Early detection of acute myocardial infarction in patients presenting with chest pain and non diagnostic ECGs: serial CK-MB sampling in the emergency department. Ann Emerg Med 1990;19:1359 –1366. 2. Lee TH, Weisberg MC, Cook EF, Daley K, Brand DA, Goldman L. Evaluation of creatine kinase and creatine kinase MB for dagnosing myocardial infarction: clinical impact in the emergency room. Arch Intern Med 1987;147:115–121. 3. Ravikilde J, Horder M, Gerhardt W, Ljungdahl L, Petterson T, Tryding N, Moller BH, Hamfelt A, Graven T, Asburg A, Helin M, Penttila I, Thygesen K. Diagnostic performance and prognostic value of serum troponin T in suspected acute myocardial infarction. Scand J Clin Lab Invest 1993;53:677– 685. 4. Katus HA, Remppis A, Scheffold T, Diederich KW, Kuebler W. Intracellular compartmentation of cardiac troponin-T and its release kinetics in patients with reperfused and nonreperfused myocardial infarction. Am J Cardiol 1991;67: 1360 –1367. 5. Katus HA, Remppis A, Neumann FJ, Diedrich KW, Kuebler W. Diagnostic efficiency of troponin-T measurements in acute myocardial infarction. Circulation 1991;83:902–912. 6. Bakker AJ, Koelemay MW, Gorgels JP, van Vlies B, Smits R, Tijssen JG, Haagen FD. Troponin-T and myoglobin at admission: value of early diagnosis of acute myocardial infarction. Eur Heart J 1994;15:45–53. 7. Wu AHB, Valdes R Jr, Apple FS, Gornet T, Stone MA, Mayfield-Stokes S, Ingersoll-Stroubos AM, Wiler B. Cardiac troponin-T immunoassay for diagnosis of acute myocardial infarction. Clin Chem 1994;40:900 –907.
CORONARY ARTERY DISEASE/TROPONIN T VS CK–MB IN CORONARY SYNDROMES
425
8. Antman EM, Grudzien C, Sacks DB. Evaluation of a rapid bedside assay for detection of serum cardiac troponin T. JAMA 1995;273:1279 –1282. 9. Hamm CW, Ravkilde J, Gerhardt W, Jorgensen P, Peheim E, Ljungdahl L, Godmann B, Katus HA. The prognostic value of serum troponin T in unstable angina. N Engl J Med 1992;327:146 –150. 10. Seino Y, Tomita Y, Takano T, Hayakawa H. Early identification of cardiac events with serum troponin T in patients with unstable angina. Lancet 1993;342: 1236 –7. 11. Lindahl B, Venge P, Wallentin L. Relation between troponin T and the risk of subsequent cardiac events in unstable coronary artery disease. Circulation 1996;93:1651–1657. 12. Magnus Ohman E, Armstrong PW, Christenson RH, Granger CB, Katus HA, Hamm CW, O’Hanesian MA, Wagner GS, Kleiman NS, Harrell FE Jr, et al. Cardiac troponin t levels for risk stratification in acute myocardial ischemia. N Engl J Med 1996;335:1333–1341. 13. Mair J, Morandell D, Genser N, Lechleiter P, Dienstl F, Puschendorf B. Equivalent early sensitivities of myoglobin, creatine kinase MB mass, creatine kinase isoform ratios, and cardiac troponins I and T for acute myocardial infarction. Clin Chem 1995;41:1266 –1272. 14. Rouan GW, Lee TH, Cook EF, Brand DA, Weisberg MC, Goldman L.
426 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
Clinical characteristics and outcome of acute myocardial infarction patients with initially normal or nonspecific electrocardiograms (report from the Multicenter Chest Pain Study). Am J Cardiol 1989;64:1087–1092. 15. Rude RE, Poole WK, Muller JE, Turi Z, Rutherford J, Parker C, Roberts R, Rabbe DS, Gold HK, Stone PH. Electrocardiographic and clinical criteria for recognition of acute myocardial infarction based on analysis of 3,697 patients. Am J Cardiol 1983;52:936 –942. 16. Hamm CW, Goldmann BU, Heeschen C, Kreymann G, Berger J, Meinertz T. Emergency room triage of patients with acute chest pain by means of rapid testing for cardiac troponin T or troponin I. N Engl J Med 1997;337:1648 –1653. 17. Karlson BW, Herlitz J, Richter A, Sjolin M, Hjalmarson A. Prognosis in patients with ST-T changes but no rise in serum activity as compared with non-Q-wave infarction. Cardiol 1991;79:271–279. 18. Schroeder JS, Lamb KH, Hu M. Do patients in whom myocardial infarction has been ruled out have a better prognosis after hospitalization than those surviving infarction? N Engl J Med 1980;303:1–5. 19. Bakker AJ, Koelemay MJ, Gorgels JP, van VliesB, Smits R, Tijssen JG, Haagen FD. Failure of new biochemical markers to exclude acute myocardial infarction at admission. Lancet 1993;342:1220 –1222.
FEBRUARY 15, 2000