Diagnostic value of high sensitive troponin T in chest pain patients with no persistent ST-elevations

Abstract

Objectives. The aim of this study was to compare the early diagnostic value of the Roche high-sensitive troponin T (Hs-TnT) and that of conventional troponins. Design. A total of 233 consecutive chest pain patients without ST-elevations were included. Hs-TnT was compared with two conventional assays (Roche troponin T [fourth generation] and Beckman Coulter Accu-TnI) on admission and at two hours. Results. When acute Myocardial Infarction (MI) was defined by conventional troponins and prespecified decision limits (Hs-TnT ≥ 14 ng/l, conventional TnT ≥ 0.04 μg/l, and Accu-TnI ≥ 0.06 μg/l) were used, Hs-TnT had a higher sensitivity but a lower specificity than conventional troponins both on admission and after two hours. When the biomarkers were compared in a ROC analysis there were no significant differences with regard to AUC. When acute MI was defined by Hs-TnT, the diagnostic performance of Hs-TnT remained very high (on admission: sensitivity 96%, specificity 85%, at two hours: sensitivity 99%, specificity 83%) whereas that of conventional troponins became lower, mainly because of lower sensitivity. Conclusion. In conclusion, when acute MI is defined by a high sensitive troponin assay, the use of Hs-TnT improves the early diagnostic accuracy compared with conventional troponins. By measuring Hs-TnT it seems possible to exclude acute MI already within the first few hours from admission.

Key words::

• Troponin
• chest pain
• myocardial infarction
• diagnosis

Early assessment of patients with chest pain and no persistent ST-elevation is still a difficult task for the clinician and a major health issue for the community. Although new cardiac markers, such as troponins, have improved the early detection of myocardial necrosis, levels of these biomarkers do not rise until several hours after a myocardial infarction (MI) and normal levels do not exclude the presence of an acute coronary syndrome (ACS) (1–4). Thus, exclusion of ACS is often a time- and resource-consuming process. Therefore, new and faster strategies for early rule-in or rule-out of ACS are needed.

Recently, a new generation of fully automated troponin assays with higher sensitivity and better precision in the lower range has been developed (5–8). In a recent landmark study evaluating several new sensitive troponin assays, the Roche high-sensitive troponin T (Hs-TnT) had a better early diagnostic value than the conventional troponin T (TnT) assay when MI was defined by conventional troponin assays used in the clinical routine (9). However, in the future the diagnosis of acute MI will be based on high-sensitive troponin assays identifying smaller MIs than previous conventional troponin assays. To what extent Hs-TnT is better than conventional troponins to detect MI:s defined by the new sensitive troponin assays is still uncertain. Moreover, a more sensitive assay will have a lower specificity due to an increase detection rate of non-ischemic myocardial damage (e.g. myocarditis, pulmonary embolism and heart failure). So far, there have been no comparison made between Hs-TnT and conventional troponin assays concerning early detection of acute myocardial damage regardless of cause.

Therefore, the purpose of the present study was to compare the early diagnostic value of Hs-TnT and that of conventional troponins with regard to: (1) MI defined by conventional troponins, (2) myocardial damage of any cause defined by conventional troponins, (3) MI defined by Hs-TnT, and (4) myocardial damage of any cause defined by Hs-TnT.

Material and methods

Study population and clinical assessment

A total of 233 patients, admitted to the coronary care unit at the Karolinska University Hospital, Huddinge, between August 2006 and January 2008 were eligible for enrolment in this prospective observational study. The inclusion criterion was chest pain or other symptoms suggestive of ACS with the last onset of symptoms within 12 hours from admission. The exclusion criterion was presence of persistent ST-segment elevations. The patients were consecutively included except for temporary interruptions of the study due to high work load at the coronary care unit. All patients received oral and written information about the study and written informed consent was obtained before entering the study. The study was conducted according to the principles of the Declaration of Helsinki and was approved by the local ethics committee.

All patients underwent clinical assessment including clinical history, physical examination, standard 12-lead ECG, ECG-monitoring and serial measurement of conventional cardiac markers up to 9–12 hours after admission. The cardiac marker used in the clinical routine were either cardiac TnT (4^th generation) from Roche Diagnostics measured with a Modular system or cardiac troponin I using the Stratus CS analyzer from Dade-Behring. All other examinations and the treatments were left to the discretion of the individual cardiologist. Clinical data were prospectively collected and entered into the local RIKS-HIA/SWEDEHEART database.

Adjudication of the final diagnosis

The final discharge diagnosis was set by the individual cardiologist and then adjudicated by two endpoint evaluators (DM and TJ) who had access to all data presented in patient records except the results from the tested troponin assays. If there was a disagreement between the evaluators a consensus was reached. An acute MI was defined according to current guidelines (4). Thus, an acute MI was diagnosed if there was a rise and/or fall of cardiac troponin with at least one value above the 99^th percentile of a normal population at a level of imprecision of less than 10% (cardiac TnT [4^th generation] ≥0.04 μg/l and cardiac troponin I [Stratus CS] ≥0.10 μg/l), together with symptoms or ECG-signs of ischemia. Also type of acute MI was defined according to current guidelines. Thus, type 1 was defined as a spontaneous myocardial infarction due to a primary coronary event, and type 2 as a non-thrombotic myocardial infarction secondary to ischemia due to either increased oxygen demand or decreased supply, e.g. tachyarrhythmia, severe infection or anemia in patients with known coronary artery disease. Patients who were considered to have unstable angina not fulfilling the criteria for acute MI were divided into those with (dynamic ECG-changes, positive stress test or a coronary angiogram showing at least one significant stenosis) or without objective signs of ischemia. Patients without ACS were classified to have (1) other cardiac, (2) other non-cardiac or unknown cause to their symptoms. Patients were then classified into those with and without acute myocardial damage regardless of whether the criteria for acute MI were fulfilled or not. Acute myocardial damage was diagnosed if there was a rise and/or fall of cardiac troponin with at least one value above the 99^th percentile of a normal population, at a level of imprecision of less than 10%. To be able to assess the early diagnostic value of the tested assays to detect MI and myocardial damage based on results of Hs-TnT during the first 12 hours from admission, a reclassification were then made by the two endpoint evaluators using the same criteria as above.

Blood samples and tested troponin assays

Blood samples were obtained on admission, two and 12 hours later. After centrifugation, serum was stored frozen at 270°C until analyzed. Serum TnT was determined by the fourth generation assay (Roche Diagnostics). The limit of detection is 0.01 μg/l with a 10% CV at 0.035 μg/l. The upper reference level, defined as the 99^th percentile of healthy controls, is below the lower limit of the analytical range. Hs-troponin T was determined by a high-sensitivity Troponin T assay (Roche Diagnostics) having a limit of detection of 2 ng/l, a 99^th percentile of healthy controls of 14 ng/l and a CV of less than 10% below the 99^th percentile. For both troponin T methods a Modular Analytics E170 (Roche Diagnostics) was used. For conventional troponin I measurement the Beckman Coulter Access AccuTnI assay was used on a UniCel DxI 800 system. The assay has a limit of detection of 0.01 μg/l, a 99^th percentile of healthy controls of 0.04 μg/l and a CV of less than 10% at 0.06 μg/l.

Statistical analysis

Continuous baseline characteristics are summarized by median with interquartile range (IQR) and categorical variables as number and percentages. Categorical variables were analyzed using the χ² test and continuous variables using the Mann-Whitney U test. To compare the diagnostic value of Hs-TnT and conventional troponin assays, we determined the cumulative sensitivity, specificity, positive and negative predictive value on admission and two hours after admission, using prespecified decision limits (the 99^th percentile of a normal population and the lowest level with CV less than 10%). To compare the diagnostic value of tested troponin assays regardless of chosen decision limits receiver operating characteristic (ROC) curves were generated and the area under the curves calculated. All data analysis was performed using the Statistical Package for Social Sciences (SPSS 17) software (SPSS Inc., Chicago, Illinois, USA).

Results

The baseline characteristics and findings on admission are listed in Table I. Of 233 patients, 114 (49%) were classified as having acute MI according to conventional troponins. Eighty-five were classified as having a type 1 MI and 29 as having a type 2 MI. Of all acute MI patients, 84 (74%) underwent coronary angiography during hospitalization, 68 had at least one significant stenosis and 59 were subsequently revascularized. Unstable angina without MI was present in 28 (12%) patients, of which 18 had objective signs of ischemia. Other cardiac causes to the symptoms (such as acute heart failure or myocarditis) were present in six (3%) patients (of which one had acute elevated troponins and one with chronic elevated troponins due to chronic kidney disease) and other non cardiac or unknown causes in 85 (36%) patients. According to the conventional troponins used in clinical practice, four patients were judged to have acute myocardial damage without a MI (one with myocarditis, one with pulmonary embolism, one who had undergone recent pacemaker implantation, and one with unknown reason to the rise in troponin levels) and four patients were judged to have persistent elevated troponins due to chronic heart failure or chronic kidney disease.

Table I. Baseline characteristics (n = 233).

Variables	All n = 233 n (%)	MI n = 114 n (%)	No MI n = 119 n (%)	p-value
Age (median, 25th–75th perc.)	65 (55–76)	72 (60–79)	61 (51–70)	<0.001
Men	155 (67)	81 (71)	74 (62)	0.152
Hypertension	117 (50)	63 (55)	54 (45)	0.131
Diabetes Mellitus	53 (23)	32 (28)	21 (18)	0.058
Current smoker	40 (17)	17 (15)	23 (19)	0.372
Previous cardiovascular disease
Myocardial infarction	70 (30)	31 (27)	39 (33)	0.353
Heart Failure	20 (9)	10 (9)	10 (8)	0.920
Revascularization	49 (21)	17 (15)	32 (27)	0.025
Stroke	14 (6)	9 (8)	5 (4)	0.236
Time from onset of symptoms (median, 25th–75th perc.)	5.3 (3.3–7.5)	5.7 (3.6–8.0)	5.0 (3.0–7.0)	0.018
ECG on admission
Rhythm
Sinus rhythm	198 (85)	90 (79)	10 (91)	0.012
Atrial fibrillation	24 (10)	19 (17)	85 (4)	0.002
ST-depression ≥0.05 mV	99 (43)	71 (62)	28 (24)	<0.001
Left bundle branch block	11 (5)	4 (4)	7 (6)	0.393

MI, myocardial infarction; perc., percentile.

The cumulative sensitivity, specificity, positive and negative predictive value of the tested biomarkers to detect MI and acute myocardial damage according to conventional troponins are listed in Table II. When the prespecified decision limits (Hs-TnT ≥ 14 ng/l, TnT ≥ 0.04 μg/l, and Accu-TnI ≥ 0.06 μg/l) for the tested assays were used, Hs-TnT had a higher sensitivity but a lower specificity than TnT and Accu-TnI both on admission and after two hours. When lower decision limits (the 99^th percentile in a healthy population) were used for TnT and Accu-TnI, all three tested methods had similar sensitivity and specificity. When the biomarkers were compared in a ROC analysis there were no significant differences with regard to AUC (Table III).

Table II. Diagnostic value of the tested assays to detect myocardial infarction and acute myocardial damage according to conventional troponins.

Time	Method	Cut-off	Sensitivity	Specificity	PPV	NPV
Myocardial infarction according to conventional troponins
0 h	Hs-TnT	≥14 ng/l	97	74	78	97
	TnT	≥0.01 μg/l	95	82	83	94
		≥0.04 μg/l	79	94	93	82
	TnI	≥0.04 μg/l	95	80	82	94
		≥0.06 μg/l	90	92	92	91
2 h	Hs-TnT	≥14 ng/l	99	71	77	99
	TnT	≥0.01 μg/l	99	79	82	99
		≥0.04 μg/l	88	92	92	88
	TnI	≥0.04 μg/l	99	71	77	99
		≥0.06 μg/l	96	89	89	96
Acute myocardial damage according to conventional troponins
0 h	Hs-TnT	≥14 ng/l	97	76	81	97
	TnT	≥0.01 μg/l	95	85	87	94
		≥0.04 μg/l	79	96	96	82
	TnI	≥0.04 μg/l	95	83	85	94
		≥0.06 μg/l	90	95	94	90
2 h	Hs-TnT	≥14 ng/l	99	73	80	99
	TnT	≥0.01 μg/l	99	82	85	99
		≥0.04 μg/l	87	95	94	87
	TnI	≥0.04 μg/l	99	74	80	99
		≥0.06 μg/l	96	91	92	95

Table III. Area under Receiver Operating Curve (ROC) for the tested troponin assays to detect myocardial infarction and acute myocardial damage according to conventional troponins.

	Hs-TnT		Conventional TnT		Accu-TnI
	AUC	(95% CI)	AUC	(95% CI)	AUC	(95% CI)
Myocardial infarction according to conventional troponins
0 h	0.95	(0.92–0.98)	0.93	(0.89–0.97)	0.94	(0.90–0.97)
2 h	0.96	(0.94–0.99)	0.96	(0.93–0.98)	0.97	(0.94–0.99)
Acute myocardial damage according to conventional troponins
0 h	0.97	(0.94–0.99)	0.95	(0.92–0.98)	0.96	(0.92–0.99)
2 h	0.98	(0.96–0.99)	0.97	(0.95–0.99)	0.98	(0.96–1.00)

Hs-TnT, High sensitive troponin T; AUC, Area under curve.

Of 28 patients who were classified to have unstable angina but without MI according to conventional troponins used in the clinical routine, seven (25%) patients had acutely elevated Hs-TnT at two hours. The corresponding numbers for TnT ≥ 0.01 μg/l, TnT ≥ 0.04 μg/l, Accu-TnI ≥ 0.04 μg/l and Accu-TnI ≥ 0.06 μg/l were three (10%), zero (0%), five (18%) and one (4%), respectively. In patients with unstable angina and Hs-TnT <14 ng/l at both time points, none had an increase of more than 2 ng/l during the first two hours from admission.

When patients were according to levels of Hs-TnT reclassified into those with or without acute MI and with or without acute myocardial damage, 131 (56%) patients were classified as having a MI and 135 (58%) patients as having an acute myocardial damage. There was one patient who was classified as having an acute myocardial damage according to the conventional troponins used in the clinical routine but not according to Hs-TnT. There were 18 patients who were classified as having acute myocardial damage according to Hs-TnT but not according to conventional troponins used in the clinical routine. These patient were classified as unstable angina or angina (n = 10), other cardiac causes (n = 2) or other or unknown cause (n = 6) when conventional troponins were used. The diagnostic value of the tested biomarkers to detect MI and acute myocardial damage according to Hs-TnT are listed in Table IV. Regardless of chosen cut-off value, Hs-TnT performed better than TnT and Accu-TnI both on admission and after two hours, with a higher sensitivity and a comparable specificity. High sensitive troponin T had the largest AUC both on admission and after two hours (Table V).

Table IV. Diagnostic value of the tested assays to detect myocardial infarction and acute myocardial damage according to Hs-TnT.

Time	Method	Cut-off	Sensitivity	Specificity	PPV	NPV
Myocardial infarction according to Hs-TnT
0 h	Hs-TnT	≥14 ng/l	98	82	90	95
	TnT	≥0.01 μg/l	90	84	91	83
		≥0.04 μg/l	70	91	93	64
	TnI	≥0.04 μg/l	88	82	90	81
		≥0.06 μg/l	80	92	94	73
2 h	Hs-TnT	≥14 ng/l	100	79	89	100
	TnT	≥0.01 μg/l	95	83	91	90
		≥0.04 μg/l	77	89	93	70
	TnI	≥0.04 μg/l	93	76	87	87
		≥0.06 μg/l	86	91	94	79
Acute myocardial damage according to Hs-TnT
0 h	Hs-TnT	≥14 ng/l	97	85	92	94
	TnT	≥0.01 μg/l	89	88	93	82
		≥0.04 μg/l	70	94	96	63
	TnI	≥0.04 μg/l	88	85	91	79
		≥0.06 μg/l	79	96	97	72
2 h	Hs-TnT	≥14 ng/l	100	84	92	100
	TnT	≥0.01 μg/l	95	88	93	90
		≥0.04 μg/l	78	94	96	70
	TnI	≥0.04 μg/l	93	81	90	87
		≥0.06 μg/l	87	96	97	79

Table V. Area under Receiver Operating Curve (ROC) for the tested troponin assays to detect myocardial infarction and acute myocardial damage according to Hs-TnT.

	Hs-TnT		Conventional TnT		Accu-TnI
	AUC	(95% CI)	AUC	(95% CI)	AUC	(95% CI)
Myocardial infarction according to Hs-TnT
0 h	0.93	(0.89–0.97)	0.89	(0.84–0.94)	0.89	(0.84–0.94)
2 h	0.94	(0.90–0.98)	0.91	(0.87–0.96)	0.92	(0.87–0.96)
Acute myocardial damage according to Hs-TnT
0 h	0.95	(0.92–0.98)	0.91	(0.87–0.95)	0.91	(0.87–0.96)
2 h	0.97	(0.94–0.99)	0.94	(0.91–0.98)	0.94	(0.91–0.98)

Hs-TnT: High sensitive troponin T, AUC: Area under curve.

Discussion

About 20% of medical emergencies are patients with symptoms suggestive of ACS (10). Rapid rule-in and rule-out of ACS is therefore essential. This process has so far been hampered by a delayed increase in circulating levels of cardiac troponins and present guidelines recommend serial measurement of troponins for 6–12 hours before acute MI can be safely excluded (1–4). Recently, new troponin assays with lower limit of detection and improved precision have been introduced. Recent studies suggest an improved early diagnostic accuracy with these new assays (9,11).

The advantage of high-sensitive troponin T

In the present study, the advantage of using Hs-TnT compared with conventional troponins was limited when acute MI was defined by levels of conventional troponins. High sensitive troponin T had a higher sensitivity but a lower specificity when prespecified decision limits were tested. However, by lowering the decision limit for TnT and Accu-TnI below the level of which CV 10% is reached, the early diagnostic value of the three tested methods were similar. When the assays were compared regardless of chosen decision limits in a ROC analysis there were no significant differences between the methods. This is in contrast to a recently published study by Reichlin et al., in which Hs-TnT had a much better early diagnostic performance than conventional troponins (9). A probable cause to this discrepancy is a longer time delay from onset of symptoms to blood sampling in the present study. In the study by Reichlin et al. the improved diagnostic performance was seen at presentation. After 1–2 hours from presentation there was no significant difference regarding AUC when Hs-TnT and TnT were compared. The median time delay was 5.3 (3.3–7.5) hours in our study. In the work by Reichlin et al. a time delay of less than three hours was present in 31% of the patients (compared to 19% in our study), but other data regarding time delay are sparse.

In the future, MI will be defined by the new more sensitive troponin assays. So far, there has been no study comparing the early diagnostic value of Hs-TnT and that of conventional troponins when MI is defined by sensitive troponins. In the present study, the number of patients classified to have acute MI increased by 15% when Hs-TnT was used. The early diagnostic performance of HS-TnT remained very high whereas that of TnT and Accu-TnI became lower, mainly because of lower sensitivity.

Exclusion of acute myocardial infarction

Although present guidelines have remained rather conservative, recommending repeated blood sampling and measurements 6–12 hours after admission, there are studies suggesting that acute MI can be excluded more rapidly even with conventional troponin assays. Eggers et al. (12) included 195 consecutive chest pain patients of which 22% were proven to have an acute MI. A decision limit of 0.07 μg/l (99^th percentile in a healthy population) for troponin I determined on a Stratus CS analyzer resulted in 100% sensitivity already three hours after admission. In another study by Collinson et al. (13), including 786 chest pain patients at low risk of MI, a decision limit of 0.02 μg/l for TnT two hours after admission and at least six hours from onset of symptoms could safely exclude acute MI. In a recent study by Keller et al. (11) serial measurement with a high sensitive troponin I method identified all patients with acute MI three hours after admission. In the present study, acute MI defined by conventional troponins could be excluded by all three tested assays two hours after admission. However, only Hs-TnT measurements could at two hours safely exclude acute MI defined by Hs-TnT.

Exclusion of unstable angina

A majority (75%) of those who were classified as having unstable angina without MI did not show any signal of myocardial damage. Thus, these patients still need to be identified by their history, ECG-changes and positive stress tests. However, patients with no dynamic ECG-changes and low levels of troponin have in general an excellent short-term prognosis and do not benefit from intensified antithrombotic treatment or early revascularization (14–17). These patients may therefore be discharged early and further evaluated as outpatients.

Exclusion of acute myocardial damage

Studies comparing diagnostic tests regardless of chosen decision limits often use ROC-analysis and calculation of AUC and thereby assume that sensitivity and specificity are equally important. However, the lower specificity of a more sensitive troponin assay is usually caused by the detection of non-ischemic myocardial damage, which may be equally important to detect. Moreover, in clinical practice it is sometimes difficult to discriminate between ischemic and non-ischemic causes of acute myocardial damage. Therefore, in the current study we also compared the ability of the assays to diagnose acute myocardial damage regardless of cause. This did result in larger AUC for all tested troponins. The reasons for the rather moderate effect are probably that the AUC was already very large and that the most obvious cases of non-ischemic symptoms were not included in the study, resulting in a low rate of non-ischemic troponin elevations.

Strengths and limitations

The strength of this study is the consecutive inclusion of unselected high-risk patients, resulting in a high prevalence of acute MI. Highly sensitive tests are required to reach a high negative predictive value in such population. The present study has also limitations. The final discharge diagnosis was based on two different conventional troponin assays that were used in the clinical routine. However, this has also been the fact in other studies (9). The sample size was rather small and the median time delay from onset of symptoms until blood sampling rather long. Therefore, firm conclusions regarding the diagnostic value of the tested biomarkers and the exact time point at which acute MI can be safely excluded cannot be made, especially in patients with a short period of symptoms before admission. Moreover, in this observational study the clinical effect of an improved early diagnostic value of Hs-TnT cannot be quantified. A randomized study is needed for that purpose.

Conclusions

When acute MI is defined by a high sensitive troponin assay, the use of Hs-TnT improves the diagnostic accuracy compared with conventional troponins. By measuring Hs-TnT it was possible in the current fairly small study to exclude acute MI already at two hours from admission. The results from this study and previous ones (9,11–13) suggest that the time needed to exclude MI with serial measurements of troponins is shorter than suggested by current guidelines. Since rise of troponin levels only indicates myocardial damage and not the cause of myocardial damage and patients may have ACS without troponin elevations, a careful history, physical examination and an ECG should still be obtained together with troponin measurements in the early assessment of these patients.

NOTICE OF CORRECTION

[ePub ahead of print] 24 March 2011, DOI: 10.3109/14017431.2011.565792. The Early Online version of this article published ahead of print on 24 March 2011 contained an error in the abstract and on page 2 where the unit for TnT and TnI were shown as g/L instead of as μg/L. This has now been corrected.

Acknowledgments

We are indebted to the staff at the CCU at Karolinska University Hospital, Huddinge and Margareta Berglund (RN) for all assistance. Financial support was provided through a regional agreement on medical training and clinical research (ALF) between Stockholm County Council and Karolinska Institute. This study was also supported by grants from the Swedish Heart and Lung Foundation, the National Board of Health and Welfare and Roche Diagnostics Scandinavia AB who also provided Hs-TnT and TnT test kits. The authors report no conflicts of interest. The authors alone are responsible for the content and writing of the paper.