CLB-08931; No. of pages: 7; 4C: Clinical Biochemistry xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem
The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay P.O. Collinson a,⁎, D. Gaze a, S. Goodacre b a b
St George's Hospital and Medical School, London, United Kingdom University of Sheffield, United Kingdom
a r t i c l e
i n f o
Article history: Received 12 November 2014 Received in revised form 16 December 2014 Accepted 17 December 2014 Available online xxxx Keywords: Troponin High sensitivity Myocardial infarction
a b s t r a c t Objectives: The aim of this study is to determine the imprecision profile, 99th percentile and diagnostic efficiency of a new high sensitivity cardiac troponin I (cTnI) assay. Methods: Total imprecision was assessed by following CLSI protocol EP15-A.14. Serum pools prepared from sera of known high cardiac troponin concentrations were adjusted by dilution with serum considered to be troponin free. Determination of the 99th-percentile reference value examined a fully characterized population that had undergone non-invasive cardiac imaging. Diagnostic accuracy utilised samples from the point of care arm of the RATPAC trial (Randomised Assessment of Treatment using Panel Assay of Cardiac markers), set in the emergency departments of six hospitals. Blood samples were taken on admission and 90 min from admission. Diagnosis was based on the universal definition of myocardial infarction utilising laboratory measurements of cardiac troponin performed at the participating sites together with measurements performed in a core laboratory and compared by construction of receiver operator characteristic curves. Results: Total imprecision was 4%–12.1% with 10% CV of 7 ng/L. cTnI was measureable in 99.5% of the samples. Troponin values were influenced by gender but not by age. The 99th percentile was 14.8 ng/L (18.1 males, 8.6 females). Progressive filtering of the population reduced the 99th percentile. For the diagnosis of MI on admission the area under the curve was 0.92, statistically indistinguishable from four other assays studied (0.90–0.94). Conclusion: The analytical performance of the new assay meets the criteria for a high sensitivity troponin assay. © 2015 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Introduction High sensitivity troponin assays allow earlier diagnosis of acute myocardial infarction (AMI) than conventional troponin assays (1–3) and replace other cytoplasmic biomarkers for early diagnosis (4,5). Guidelines recommend implementation of the 99th percentile as the upper reference limit for an abnormal result, and studies have shown that implementation of values at or near this threshold improves patient outcomes (6). Criteria which should be met for a troponin assay to be designated high sensitivity have been proposed (7–9). In this study, the Abbott high sensitivity cardiac troponin I (cTnI) assay underwent analytical and clinical validation to see if it met the proposed criteria for a high sensitivity assay. Evaluation examined assay imprecision, reference interval values and clinical diagnostic performance
⁎ Corresponding author at: Department of Clinical Blood Sciences, St George's Hospital and Medical School, Jenner Wing, Cranmer Terrace, London SW17 0RE, United Kingdom. E-mail address: [email protected] (P.O. Collinson).
compared with two other contemporary sensitive cTnI assays and one high sensitivity cardiac troponin T (cTnT) assay. Methods Assay imprecision Assay imprecision was assessed by following CLSI protocol EP15A.14 using human serum sample pools. Serum pools were prepared from residual samples by selection of sera of known high cardiac troponin concentrations. A base serum pool was created and individual sample pools were prepared by dilution with serum with undetectable troponin by measurement using a contemporary sensitive cTnI assay (cTnI ultra, Siemens Diagnostics, Camberley UK). The objective was to cover the range of cTnI values if possible from close to the limit of detection (1.1 ng/L), across the expected 99th percentile, the WHO diagnostic cut off for MI diagnosis and to a high value representative of unequivocal AMI diagnosis. A total of 13 pools were analysed (20 individual samples in total for each pool, a total of 260 samples) with 4 samples
http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017 0009-9120/© 2015 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
2
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
from each pool measured daily over 5 consecutive days). All samples were stored frozen at − 20 °C prior to analysis, then thawed, mixed and centrifuged prior to analysis. Reference interval Ethical permission for the study was obtained from the local research ethics committee in accordance with the declaration of Helsinki. The background normal population group comprised subjects N 45 years old randomly selected from seven representative local community practices (10). Subjects who agreed to participate were further characterised with clinical details collected by questionnaires plus blood pressure measurement, spirometry, electrocardiography (ECG) and echocardiography. They were venesected for fasting serum glucose, and creatinine. An aliquot of serum stored frozen at − 70 °C and which had not been previously thawed was used for the reference interval study. Values assessed on spirometry included forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC), and their ratio (FEV1%). Spirometry was defined as abnormal if FEV1% b 60% (obstructive defect), if FEV1% b70% with FEV1 b 80% of the predicted value (obstructive defect), and if FEV1% N 70% with both FEV1 and FVC b 80% of the predicted (restrictive defect). Two-dimensional echocardiography was performed with a SONOS 4500 machine (Philips, Eindhoven, The Netherlands) using second harmonic imaging. Left ventricular ejection fraction (LVEF) was calculated quantitatively using Simpson's apical biplane method taking the average of three readings (11). Borderline or worse left ventricular systolic dysfunction (LVSD) was defined as LVEF b50%. LV mass was calculated using the Devereux-modified American Society of Echocardiography equation, with left ventricular hypertrophy (LVH) defined as LV mass index N134 g/m2 for men and N110 g/m2 for women (12). Valvular regurgitation was assessed qualitatively on a five-point scale (nil or trivial, mild, mild-to-moderate, moderate, and severe). Valvular stenosis was assessed by peak pressure gradient and estimated valve area, and again ascribed the same five-point scale. Significant valve disease was taken as mild-to moderate or worse. Diastolic parameters assessed included isovolumic relaxation time (IVRT), mitral inflow peak E wave velocity (E), peak A wave velocity (A), E/A ratio, and E wave deceleration time (E decel). Diastolic heart failure (DHF) was defined according to the European Study Group on Diastolic Heart Failure guidelines. The following three populations were studied: unselected apparently normal healthy individuals; subjects with no previous history of vascular disease, diabetes mellitus or taking any cardioactive drugs; subjects with no history of vascular disease, diabetes mellitus, hypertension, or heavy alcohol intake and receiving no cardiac medication; whose blood pressure was ≤140/90 mm Hg as the mean of two readings; whose fasting blood glucose was b6 mmol/l, whose eGFR was N 60 mL/min/1.73 m2; and who had no significant valvular heart disease, LVH, DHF, or regional wall motion abnormalities on echocardiography and had an LVEF N50%.
Exclusion criteria for enrolment were; ECG changes for myocardial infarction or high-risk acute coronary syndrome (N1 mm ST deviation or N 3 mm inverted T waves), known coronary heart disease presenting with prolonged (N1 h) or recurrent episodes of cardiac-type pain, proven or suspected serious non-coronary pathology (e.g. pulmonary embolus), co-morbidity or social problems that require hospital admission, an obvious non-cardiac cause (e.g. pneumothorax or muscular pain), more than 12 h since their most significant episode of pain, previous participants in the study, those unable to understand the trial information and those unwilling to consent. Research nurses then used emergency department and hospital inpatient notes to record management decisions at initial attendance and admission, and to identify subsequent attendances or admissions and the major adverse cardiac events (MACE) of readmission with MI or unstable angina, need for revascularisation and death up to three months. All those eligible for enrolment were then randomised to either the POCT or CLT arm. Patients randomised to the CLT arm were managed according to the usual hospital protocol of serial sampling and biomarker measurement in the hospital laboratory (CLT). Patients randomised to POCT were scheduled to have a blood sample drawn on admission and at 90 min from admission for POCT measurement. An additional sample for subsequent biomarker measurement was drawn at the same time as the POCT sample and the serum was separated and frozen prior to −20 ° C prior to transfer to long-term storage at −70 ° C in the central laboratory. The admission and 90 minute samples were subsequently analysed for cTnI and cTnT in the trial central laboratory for this study. In addition, patients randomised to POCT had further samples taken and analysed by the local laboratory according to the local clinical protocol and clinician's discretion unless a decision was made to discharge them on the basis of POCT measurements. Patients enrolled in the POCT arm where managed according to POCT measurements, clinical assessment and additional laboratory measurements in accordance with the trial protocol. Decision to discharge or admit on the basis of POCT measurements (supplemented by laboratory measurements if requested) was at the clinician's discretion. Only samples from the POCT arm (0 and 90 minute samples) were used in the clinical evaluation. Final diagnostic classification was performed by two independent clinicians with access to all the relevant information, utilising the 99th percentile value for the cardiac troponin value from POCT measurement, from the local laboratory (where available) and from cTnI measurements performed in the central laboratory. All patients had POCT measurement with a cTnI method which meets current analytical goals. Four of the local laboratories used a troponin method which meets the current analytical goals for the 99th percentile, one used a cTnI method which just fails to reach these goals and one used the current generation cTnT method. Central laboratory measurements were performed using a cTnI method that meets current analytical goals, the Siemens ultra cTnI assay. Analytical methods
Clinical performance Samples used were taken from patients enrolled in the Randomised Assessment of Panel Assay of Cardiac Markers (RATPAC) Trial. Ethical permission for this study was obtained from the Leeds East Research Ethics Committee (07/Q1206/22) and the study was performed in accordance with the declaration of Helsinki. Full details of the RATPAC trial had been published (13,14). The RATPAC trial was a prospective randomised controlled trial comparing point of care testing (POCT) with conventional management based on central laboratory testing (CLT). The population studied was the low risk chest pain population in whom biomarker measurement is integral to clinical decision-making. Samples from this study have now been analysed to examine the diagnostic performance of the new generation high sensitivity troponin assays. The study enrolled patients 18 years or older presenting with acute chest pain to the emergency department (ED) of 6 participating hospitals were screened for eligibility.
Glucose (coupled hexokinase method, analytical range 0.3–38.8 mmol/L) and creatinine (Jaffe alkaline picrate calibrated to a reference isotope dilution mass spectrometry method, analytical range 27–2210 μmol/L) were measured using a Beckman LX20 by the manufacturer's recommended procedures. Both assays had coefficients of variation (CV) of less than 5% across the analytical range. Estimated glomerular filtration rate (eGFR) was then calculated by the corrected diet modification of renal disease equation (15). POCT measurements were performed using the Stratus CS (Siemens Healthcare Diagnostics). The analytical characteristics of the assays were as follows: cTnI detection limit 30 ng/L, analytical range 30 to 50,000 ng/L, and inter assay CV 4.0–8.2% (67 to 344 ng/L). The 99th centile of the assay is 70 ng/L. At the individual sites cardiac troponin was measured as follows; Siemens cTnI ultra (3 sites) 99th percentile limit 40 ng/L, Abbott cTnI (1 site) 99th percentile limit 50 ng/L, Beckman
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
3
AccuTnI (1 site) 99th percentile limit 60 ng/L (locally derived value) and Roche cTnT (1 site) 99th percentile limit 10 ng/L. Central laboratory assays were as follows. For cTnI, the Architect hsTnI (Abbott Diagnostics), range 1.1–50,000 ng/L 10% CV 4.7 ng/L, Siemens cTnI Ultra (ADVIA Centaur, Siemens Healthcare Diagnostics), range 6–50,000 ng/L. 10% CV is 30 ng/L with a 99th centile of 40 ng/L. Beckman-Coulter AccuTnI enhanced (Access 2, Beckman-Coulter), range 1–100,000 ng/L, 10% CV 30 ng/L, and 99th centile 40 ng/L. For cTnT, the Roche diagnostics high sensitivity cardiac troponin T (hs-cTnT) assay (Elecsys 2010, Roche diagnostics), detection limit 3 ng/, upper limit of 10,000 ng/L, 10% CV is 13 ng/L with a 99th centile of 14 ng/L. All assays were performed by operators blinded to the clinical and diagnostic information. The Abbott hsTnI measurements were performed on an instrument in routine clinical use for a change of immunoassays. Statistical methods All statistical analyses were performed using Analyse-it for Microsoft Excel (version 2.30), www.analyse-it.com, by nonparametric methods. Absolute values for total imprecision, repeatability (within run) and between run imprecision were calculated. The imprecision profile was estimated from imprecision curves using a power curve fitting equation and estimating the 10% and 20% points from the fitted curve and by interpolation from the calculated values. The 99th percentile was calculated as the absolute single upper 99th percentile value of the one tailed distribution. Association of age and gender with cTnI was assessed by non-parametric comparison (Mann–Whitney). Diagnostic accuracy was compared by calculation of receiver operator characteristic (ROC) curves using final diagnosis and MACE as the dichotomous variables and comparison of area under the curve (AUC) by the Delong Delong Clarke–Pearson method. Sensitivity and specificity were compared by construction of 2 × 2 tables and Fishers exact probability test. Results Imprecision Total imprecision was estimated as 7.3 ng/L at the 10% point (95% confidence interval for CV 6.8 to 20.8%) with a 20% CV at 1.55 ng/L (Table 1 & Fig. 1). The 10% CV for repeatability (within run imprecision) was 1.75 ng/L and between run imprecision b2% at values greater than 5 ng/L and 3% at the lowest troponin value directly measured (4.4 ng/L). 99th percentile Full data was available on 599/1392 individuals (296 male, 303 female) who attended for screening. The median age was 58 years
Table 1 Imprecision data. Total imprecision
Within run imprecision
cTnI (ng/L)
CV (%)
Confidence interval
CV (%)
Confidence interval
4.4 5.0 6.8 6.9 8.9 10.8 11.6 43.6 54.3 74.9 367.1 693.7 1261.6
12.1 12.0 12.6 11.2 6.9 9.2 9.1 5.3 3.9 6.9 6.9 4.4 4.0
(8.0–24.7) (8.2–21.9) (8.6–22.9) (7.2–24.6) (4.6–14.1) (5.7–22.6) (5.9–20.1 (3.6–9.6) (2.7–6.8) (4.6–14.1) (4.4–15.1) (3.2–6.9) (3.0–5.8)
7.0 8.4 8.5 5.2 4.2 3.9 4.7 3.6 2.8 4.2 3.5 3.7 4.0
(5.2–10.8) (6.2–13.1) (6.3–13.2) (3.8–8.1) (3.1–6.5) (2.9–6.1) (3.4–7.2) (2.6–5.5) (2.1–4.4) (3.1–6.5) (2.6–5.5) (2.7–5.8) (2.9–6.2)
Fig. 1. Imprecision curve.
(range 45–89, lower quartile 51, upper quartile 67 years). There was no significant difference in age or sex distribution across all of the subgroups studied. The distribution in cardiac troponin between the randomly selected population, those selected by questionnaire and those selected by questionnaire, examination, ECG and imaging is summarised in Fig. 2. Progressive screening reduced the number of outliers. There were statistically significant differences in population according to gender in the three groups with higher values seen in males. In the unscreened population, cTnI correlated with age but this correlation disappeared in the other two groups. Troponin was detectable in 99.5% of the population. Reference intervals are summarised in Table 2. There were insufficient numbers to produce a meaningful male and female distribution for the highly screened population. The highly screened male, but not female, population distribution corresponded to a normal distribution (data shown in Supplementary Fig. 1). Excluding one outlier yielded a 99th percentile of 8.5 ng/L in the female population but the number of samples was small, 115. Diagnostic test performance Data was available for 342 patients, 202 male 140 female median age 53.8 years (range 23.7–90.6 lower quartile 44.2, upper quartile 64.7). Median time from onset of chest pain to first sample was 53.8 min (range 55–7630, lower quartile 168.7, upper quartile 370 min). The AUC for the admission sample was 0.91 (CI 0.83–0.98), for the 90 minute sample 0.98 (CI 0.95–1.00) and for the peak value 0.90 (CI 0.81–0.99) and each (admission, 90 minute and peak) was statistically indistinguishable from the other troponin assays. Data for the admission sample is shown in Fig. 3. Full details of the area under the ROC curves and AUCs are given in the online data appendix (Supplementary Figs. 2 and 3 and Supplementary Tables 1–3). The Abbott hsTnI assay predicted MACE with the AUC on admission of 0.84 (CI 0.73–0.96) and for the peak value of 0.84 (CI 0.70–0.97) with the same diagnostic efficiency as the other assays (Fig. 4 for the admission sample, Supplementary Fig. 4, and Supplementary Tables 4 and 5). Because of the male–female difference, the data was divided and the analysis repeated for males and females separately. There was no significant difference in test performance for females although the Abbott hsTnI assay was statistically better than the Siemens assay on the admission sample in males for both diagnosis and prognosis (data not shown). There was a trend to an influence of gender on diagnostic efficiency when using the 99th percentile but due to small numbers of positive cases this failed to reach statistical significance. When a diagnostic discriminant of 21 ng/L was used overall diagnostic sensitivity was 61.3% (CI 42.2–78.2) on admission with a specificity of 98.5% (CI 96.2–99.6).
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
4
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
Fig. 2. Distribution of cTnI values in the three groups studied. 99th percentiles were as follows: All data 21.0 ng/L (n = 599), Questionnaire exclusion 12.3 ng/L (n = 366), and highly selected population 8.7 ng/L (n = 208).
For males at a diagnostic discriminant of 21 ng/L sensitivity was 77.8% (CI 52.4–93.6). Using a sex-specific cut-off of 28.6 ng/L for males sensitivity was reduced to 66.7% (CI 41.0–86.7) p = 0.71 for comparison of sensitivity at the two different cut-offs, with specificity remaining at 98.1% (CI 94.6–99.6). In female sensitivity at 21 ng/L was 38.5% (CI 13.9–68.4)
and increased to 69.2% (CI 38.6–90.9) p = 0.24 for comparison of sensitivity at the two different cut-offs with a small decrease in specificity from 99.0% (CI 94.8–100) to 96.2% (CI 90.4–98.9) p = 0.37 for comparison of specificity at the two different cut-offs when a sex specific cut off of 9.9 ng/L is used.
Table 2 Reference intervals for three population groups studied. Population
99th percentile ng/L all data [confidence interval] (n)
99th percentile ng/L—male (n)
99th percentile ng/L—female (n)
p (male vs. female)
All Plus questionnaire screening Plus normal imaging, ECG, spirometery and biochemistry
21.0 [11–44] (599) 12.3 [6.2–21] (366) 8.7 (208)
28.5 (296) 18.3 (171)
9.9 (303) 9.5 (195)
b0.0001 0.0002 0.0009
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
5
Fig. 3. Receiver operating characteristic curves for samples taken on admission for a final diagnosis of acute myocardial infarction. cTnI A 1 = Abbott hs cTnI admission sample, cTnI CS 1 = Stratus CS cTnI admission minute sample, cTnI S 1 = Siemens ultra cTnI admission sample, cTnI B 1 = Beckman cTnI admission sample, and cTnT 1 = Roche hs cTnT admission sample.
Discussion Surveys of routine use of cardiac biomarkers have shown that the manufacturer's datasheets and published literature are the commonest source of performance data and decision limits (16,17). It is therefore
important to provide independent assessment of claims by manufacturer. Performance claims within datasheets are not always reflected by experience in routine clinical use. This is because most laboratory analysers in routine clinical use are performing multiple sequential measurements on different sample types and of different analytes. They are therefore
Fig. 4. Receiver operating characteristic curves for samples taken on admission to predict major adverse cardiac events. cTnI A 1 = Abbott hs cTnI admission sample, cTnI CS 1 = Stratus CS cTnI admission minute sample, cTnI S 1 = Siemens ultra cTnI admission sample, cTnI B 1 = Beckman cTnI admission sample, and cTnT 1 = Roche hs cTnT admission sample.
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
6
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
functioning under less than ideal conditions. This study performed measurements on an analyser in routine clinical use. The 10% CV imprecision claim for the Abbott hs cTnI assay is 4.7 ng/L. We found a value of 7.3 ng/L at the 10% point (95% confidence interval for CV 6.8 to 20.8%). This represents the difference between the CV to be expected in routine clinical use and the CV obtained under more optimal conditions. The value obtained is similar to but higher than the 5.6 ng/L found in a multicentre evaluation of this analyser (18). The CV obtained is below the 99th percentile for the assay and the reproducibility of the assay was excellent with the repeatability between results less than 2% with a 10% CV of 1.75 ng/L. Clinically this will translate into good clinical performance. The 10% CV is one third of the combined 99th percentile and the repeatability means that sequential measurements for detection of a delta troponin value can be undertaken over short time intervals with a high degree of confidence. Recent recommendations for high sensitivity assays suggest that should very short time intervals between sequential sampling can be used, certainly on admission, 90 min, 2 h and 3 h post admission (19–22) and possibly admission and 1 h from admission (23). In order to use such short time intervals safely a high degree of repeatability between sequential measurements is required. It would seem that the Abbott hs cTnI assay meets these requirements. The results of the reference range study are consistent with previous studies on high sensitivity troponin assays and show that progressive exclusion of probable underlying structural heart disease causes a decreased troponin upper reference limit to be obtained, more marked in men than in women (10,24). A distinct difference seen between men and women with a higher value in men, supports the use of gender specific cut-offs (25). The diagnostic efficiency of the assay was similar to that of two contemporary sensitive assays. However, the number of patients with a final diagnosis of AMI in the sample population was relatively modest at 10.7% (33/309) although it did correspond to a low risk chest pain group. It has been suggested that the Abbott hs cTnI assay shows greater sensitivity in the early phase of AMI (26). Care must be taken in interpreting apparent analytical diagnostic performance differences as they may not always translate into significant differences in prognostic ability (27). There are limitations to this study. The samples for the reference population had been in long term storage (7 years). Preliminary data on freeze thaw stability and studies from other cTnI methods did not suggest that significant sample deterioration would have occurred but this cannot be categorically excluded, but the data are consistent with other studies. The reduction in sample numbers on extensive screening is a consequence of the study design but the data are consistent with previously published work of this type. The relatively small numbers of patients with a final diagnosis of AMI, although representative of an Emergency Department population, the group of interest, means that the study is not powered enough to detect small differences. This would require a large multicentre study. Conclusion The Abbott high sensitivity troponin I assay meets the performance characteristics of a high sensitivity assay as recently recommended (7,8). The diagnostic performance in patients with chest pain is comparable with other contemporary and high sensitivity assays. Acknowledgements The authors thank Abbott diagnostics for providing free reagents and supporting the cost of analyses of the samples by an independent laboratory as an unrestricted grant. Abbott diagnostics had no influence over the study design, analytical process, data analysis or preparation, submission or findings of the manuscript. The authors have no conflict of interest and have received no honoraria in respect of the study or any related activities.
Professor Collinson acts as a guarantor in respect of data integrity and study transparency and as study sponsor.
Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.clinbiochem.2014.12.017.
References [1] Melanson SE, Morrow DA, Jarolim P. Earlier detection of myocardial injury in a preliminary evaluation using a new troponin I assay with improved sensitivity. Am J Clin Pathol 2007;128:282–6. [2] Reichlin T, Hochholzer W, Bassetti S, Steuer S, Stelzig C, Hartwiger S, et al. Early diagnosis of myocardial infarction with sensitive cardiac troponin assays. N Engl J Med 2009;361: 858–67. [3] Keller T, Zeller T, Peetz D, Tzikas S, Roth A, Czyz E, et al. Sensitive troponin I assay in early diagnosis of acute myocardial infarction. N Engl J Med 2009;361:868–77. [4] Eggers KM, Oldgren J, Nordenskjold A, Lindahl B. Diagnostic value of serial measurement of cardiac markers in patients with chest pain: limited value of adding myoglobin to troponin I for exclusion of myocardial infarction. Am Heart J 2004;148:574–81. [5] Collinson P, Goodacre S, Gaze D, Gray A. Very early diagnosis of chest pain by point-ofcare testing: comparison of the diagnostic efficiency of a panel of cardiac biomarkers compared with troponin measurement alone in the RATPAC trial. Heart 2012;98:312–8. [6] Mills NL, Churchhouse AM, Lee KK, Anand A, Gamble D, Shah AS, et al. Implementation of a sensitive troponin I assay and risk of recurrent myocardial infarction and death in patients with suspected acute coronary syndrome. JAMA 2011;305:1210–6. [7] Apple FS. A new season for cardiac troponin assays: it's time to keep a scorecard. Clin Chem 2009;55:1303–6. [8] Apple FS, Collinson PO. Analytical characteristics of high-sensitivity cardiac troponin assays. Clin Chem 2012;58:54–61. [9] Apple FS, Jaffe AS, Collinson P, Mockel M, Ordonez-Llanos J, Lindahl B, et al. IFCC educational materials on selected analytical and clinical applications of high sensitivity cardiac troponin assays. Clin Biochem 2014. [10] Collinson PO, Heung YM, Gaze D, Boa F, Senior R, Christenson R, et al. Influence of population selection on the 99th percentile reference value for cardiac troponin assays. Clin Chem 2012;58:219–25. [11] Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, Feigenbaum H, et al. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr 1989; 2:358–67. [12] Devereux RB, Alonso DR, Lutas EM, Gottlieb GJ, Campo E, Sachs I, et al. Echocardiographic assessment of left ventricular hypertrophy: comparison to necropsy findings. Am J Cardiol 1986;57:450–8. [13] Goodacre SW, Bradburn M, Cross E, Collinson P, Gray A, Hall AS. The Randomised Assessment of Treatment using Panel Assay of Cardiac Markers (RATPAC) trial: a randomised controlled trial of point-of-care cardiac markers in the emergency department. Heart 2011;97:190–6. [14] Goodacre S, Bradburn M, Fitzgerald P, Cross E, Collinson P, Gray A, et al. The RATPAC (Randomised Assessment of Treatment using Panel Assay of Cardiac markers) trial: a randomised controlled trial of point-of-care cardiac markers in the emergency department. Health Technol Assess 2011;15:1–108. [15] Levey AS, Coresh J, Greene T, Marsh J, Stevens LA, Kusek JW, et al. Expressing the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem 2007;53:766–72. [16] Collinson P, Pulkki K, Suvisaari J, Ravkilde J, Stavljenic-Rukavina A, HammererLercher A, et al. How well do laboratories follow guidelines on cardiac markers? The cardiac marker guideline uptake in Europe study. Clin Chem 2008;54:448–9. [17] Collinson PO, Dieijen-Visser MP, Pulkki K, Hammerer-Lercher A, Suvisaari J, Ravkilde J, et al. Evidence-based laboratory medicine: how well do laboratories follow recommendations and guidelines? The Cardiac Marker Guideline Uptake in Europe (CARMAGUE) Study. Clin Chem 2012;58:305–6. [18] Krintus M, Kozinski M, Boudry P, Capell NE, Koller U, Lackner K, et al. European multicenter analytical evaluation of the Abbott ARCHITECT STAT high sensitive troponin I immunoassay. Clin Chem Lab Med 2014;52:1657–65. [19] Hamm CW, Bassand JP, Agewall S, Bax J, Boersma E, Bueno H, et al. ESC guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: the task force for the management of acute coronary syndromes (ACS) in patients presenting without persistent ST-segment elevation of the European Society of Cardiology (ESC). Eur Heart J 2011;32:2999–3054. [20] Collinson P, Gaze D, Goodacre S. Comparison of contemporary troponin assays with the novel biomarkers, heart fatty acid binding protein and copeptin, for the early confirmation or exclusion of myocardial infarction in patients presenting to the emergency department with chest pain. Heart 2014;100:140–5. [21] Cullen L, Mueller C, Parsonage WA, Wildi K, Greenslade JH, Twerenbold R, et al. Validation of high-sensitivity troponin I in a 2-hour diagnostic strategy to assess 30-day outcomes in emergency department patients with possible acute coronary syndrome. J Am Coll Cardiol 2013;62:1242–9. [22] Than M, Aldous S, Lord SJ, Goodacre S, Frampton CM, Troughton R, et al. A 2-hour diagnostic protocol for possible cardiac chest pain in the emergency department: a randomized clinical trial. JAMA Int Med 2014;174:51–8.
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx [23] Reichlin T, Schindler C, Drexler B, Twerenbold R, Reiter M, Zellweger C, et al. One-hour rule-out and rule-in of acute myocardial infarction using high-sensitivity cardiac troponin T. Arch Intern Med 2012;172:1211–8. [24] Koerbin G, Abhayaratna WP, Potter JM, Apple FS, Jaffe AS, Ravalico TH, et al. Effect of population selection on 99th percentile values for a high sensitivity cardiac troponin I and T assays. Clin Biochem 2013;46:1636–43. [25] Eggers KM, Johnston N, James S, Lindahl B, Venge P. Cardiac troponin I levels in patients with non-ST-elevation acute coronary syndrome—the importance of gender. Am Heart J 2014;168:317–24.
7
[26] Rubini GM, Twerenbold R, Reichlin T, Wildi K, Haaf P, Schaefer M, et al. Direct comparison of high-sensitivity-cardiac troponin I vs. T for the early diagnosis of acute myocardial infarction. Eur Heart J 2014;35:2303–11. [27] Venge P, Lindahl B. Cardiac troponin assay classification by both clinical and analytical performance characteristics: a study on outcome prediction. Clin Chem 2013;59:976–81.
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
Contents lists available at ScienceDirect
Clinical Biochemistry journal homepage: www.elsevier.com/locate/clinbiochem
The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay P.O. Collinson a,⁎, D. Gaze a, S. Goodacre b a b
St George's Hospital and Medical School, London, United Kingdom University of Sheffield, United Kingdom
a r t i c l e
i n f o
Article history: Received 12 November 2014 Received in revised form 16 December 2014 Accepted 17 December 2014 Available online xxxx Keywords: Troponin High sensitivity Myocardial infarction
a b s t r a c t Objectives: The aim of this study is to determine the imprecision profile, 99th percentile and diagnostic efficiency of a new high sensitivity cardiac troponin I (cTnI) assay. Methods: Total imprecision was assessed by following CLSI protocol EP15-A.14. Serum pools prepared from sera of known high cardiac troponin concentrations were adjusted by dilution with serum considered to be troponin free. Determination of the 99th-percentile reference value examined a fully characterized population that had undergone non-invasive cardiac imaging. Diagnostic accuracy utilised samples from the point of care arm of the RATPAC trial (Randomised Assessment of Treatment using Panel Assay of Cardiac markers), set in the emergency departments of six hospitals. Blood samples were taken on admission and 90 min from admission. Diagnosis was based on the universal definition of myocardial infarction utilising laboratory measurements of cardiac troponin performed at the participating sites together with measurements performed in a core laboratory and compared by construction of receiver operator characteristic curves. Results: Total imprecision was 4%–12.1% with 10% CV of 7 ng/L. cTnI was measureable in 99.5% of the samples. Troponin values were influenced by gender but not by age. The 99th percentile was 14.8 ng/L (18.1 males, 8.6 females). Progressive filtering of the population reduced the 99th percentile. For the diagnosis of MI on admission the area under the curve was 0.92, statistically indistinguishable from four other assays studied (0.90–0.94). Conclusion: The analytical performance of the new assay meets the criteria for a high sensitivity troponin assay. © 2015 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Introduction High sensitivity troponin assays allow earlier diagnosis of acute myocardial infarction (AMI) than conventional troponin assays (1–3) and replace other cytoplasmic biomarkers for early diagnosis (4,5). Guidelines recommend implementation of the 99th percentile as the upper reference limit for an abnormal result, and studies have shown that implementation of values at or near this threshold improves patient outcomes (6). Criteria which should be met for a troponin assay to be designated high sensitivity have been proposed (7–9). In this study, the Abbott high sensitivity cardiac troponin I (cTnI) assay underwent analytical and clinical validation to see if it met the proposed criteria for a high sensitivity assay. Evaluation examined assay imprecision, reference interval values and clinical diagnostic performance
⁎ Corresponding author at: Department of Clinical Blood Sciences, St George's Hospital and Medical School, Jenner Wing, Cranmer Terrace, London SW17 0RE, United Kingdom. E-mail address: [email protected] (P.O. Collinson).
compared with two other contemporary sensitive cTnI assays and one high sensitivity cardiac troponin T (cTnT) assay. Methods Assay imprecision Assay imprecision was assessed by following CLSI protocol EP15A.14 using human serum sample pools. Serum pools were prepared from residual samples by selection of sera of known high cardiac troponin concentrations. A base serum pool was created and individual sample pools were prepared by dilution with serum with undetectable troponin by measurement using a contemporary sensitive cTnI assay (cTnI ultra, Siemens Diagnostics, Camberley UK). The objective was to cover the range of cTnI values if possible from close to the limit of detection (1.1 ng/L), across the expected 99th percentile, the WHO diagnostic cut off for MI diagnosis and to a high value representative of unequivocal AMI diagnosis. A total of 13 pools were analysed (20 individual samples in total for each pool, a total of 260 samples) with 4 samples
http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017 0009-9120/© 2015 The Canadian Society of Clinical Chemists. Published by Elsevier Inc. All rights reserved.
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
2
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
from each pool measured daily over 5 consecutive days). All samples were stored frozen at − 20 °C prior to analysis, then thawed, mixed and centrifuged prior to analysis. Reference interval Ethical permission for the study was obtained from the local research ethics committee in accordance with the declaration of Helsinki. The background normal population group comprised subjects N 45 years old randomly selected from seven representative local community practices (10). Subjects who agreed to participate were further characterised with clinical details collected by questionnaires plus blood pressure measurement, spirometry, electrocardiography (ECG) and echocardiography. They were venesected for fasting serum glucose, and creatinine. An aliquot of serum stored frozen at − 70 °C and which had not been previously thawed was used for the reference interval study. Values assessed on spirometry included forced expiratory volume in 1 s (FEV1), forced vital capacity (FVC), and their ratio (FEV1%). Spirometry was defined as abnormal if FEV1% b 60% (obstructive defect), if FEV1% b70% with FEV1 b 80% of the predicted value (obstructive defect), and if FEV1% N 70% with both FEV1 and FVC b 80% of the predicted (restrictive defect). Two-dimensional echocardiography was performed with a SONOS 4500 machine (Philips, Eindhoven, The Netherlands) using second harmonic imaging. Left ventricular ejection fraction (LVEF) was calculated quantitatively using Simpson's apical biplane method taking the average of three readings (11). Borderline or worse left ventricular systolic dysfunction (LVSD) was defined as LVEF b50%. LV mass was calculated using the Devereux-modified American Society of Echocardiography equation, with left ventricular hypertrophy (LVH) defined as LV mass index N134 g/m2 for men and N110 g/m2 for women (12). Valvular regurgitation was assessed qualitatively on a five-point scale (nil or trivial, mild, mild-to-moderate, moderate, and severe). Valvular stenosis was assessed by peak pressure gradient and estimated valve area, and again ascribed the same five-point scale. Significant valve disease was taken as mild-to moderate or worse. Diastolic parameters assessed included isovolumic relaxation time (IVRT), mitral inflow peak E wave velocity (E), peak A wave velocity (A), E/A ratio, and E wave deceleration time (E decel). Diastolic heart failure (DHF) was defined according to the European Study Group on Diastolic Heart Failure guidelines. The following three populations were studied: unselected apparently normal healthy individuals; subjects with no previous history of vascular disease, diabetes mellitus or taking any cardioactive drugs; subjects with no history of vascular disease, diabetes mellitus, hypertension, or heavy alcohol intake and receiving no cardiac medication; whose blood pressure was ≤140/90 mm Hg as the mean of two readings; whose fasting blood glucose was b6 mmol/l, whose eGFR was N 60 mL/min/1.73 m2; and who had no significant valvular heart disease, LVH, DHF, or regional wall motion abnormalities on echocardiography and had an LVEF N50%.
Exclusion criteria for enrolment were; ECG changes for myocardial infarction or high-risk acute coronary syndrome (N1 mm ST deviation or N 3 mm inverted T waves), known coronary heart disease presenting with prolonged (N1 h) or recurrent episodes of cardiac-type pain, proven or suspected serious non-coronary pathology (e.g. pulmonary embolus), co-morbidity or social problems that require hospital admission, an obvious non-cardiac cause (e.g. pneumothorax or muscular pain), more than 12 h since their most significant episode of pain, previous participants in the study, those unable to understand the trial information and those unwilling to consent. Research nurses then used emergency department and hospital inpatient notes to record management decisions at initial attendance and admission, and to identify subsequent attendances or admissions and the major adverse cardiac events (MACE) of readmission with MI or unstable angina, need for revascularisation and death up to three months. All those eligible for enrolment were then randomised to either the POCT or CLT arm. Patients randomised to the CLT arm were managed according to the usual hospital protocol of serial sampling and biomarker measurement in the hospital laboratory (CLT). Patients randomised to POCT were scheduled to have a blood sample drawn on admission and at 90 min from admission for POCT measurement. An additional sample for subsequent biomarker measurement was drawn at the same time as the POCT sample and the serum was separated and frozen prior to −20 ° C prior to transfer to long-term storage at −70 ° C in the central laboratory. The admission and 90 minute samples were subsequently analysed for cTnI and cTnT in the trial central laboratory for this study. In addition, patients randomised to POCT had further samples taken and analysed by the local laboratory according to the local clinical protocol and clinician's discretion unless a decision was made to discharge them on the basis of POCT measurements. Patients enrolled in the POCT arm where managed according to POCT measurements, clinical assessment and additional laboratory measurements in accordance with the trial protocol. Decision to discharge or admit on the basis of POCT measurements (supplemented by laboratory measurements if requested) was at the clinician's discretion. Only samples from the POCT arm (0 and 90 minute samples) were used in the clinical evaluation. Final diagnostic classification was performed by two independent clinicians with access to all the relevant information, utilising the 99th percentile value for the cardiac troponin value from POCT measurement, from the local laboratory (where available) and from cTnI measurements performed in the central laboratory. All patients had POCT measurement with a cTnI method which meets current analytical goals. Four of the local laboratories used a troponin method which meets the current analytical goals for the 99th percentile, one used a cTnI method which just fails to reach these goals and one used the current generation cTnT method. Central laboratory measurements were performed using a cTnI method that meets current analytical goals, the Siemens ultra cTnI assay. Analytical methods
Clinical performance Samples used were taken from patients enrolled in the Randomised Assessment of Panel Assay of Cardiac Markers (RATPAC) Trial. Ethical permission for this study was obtained from the Leeds East Research Ethics Committee (07/Q1206/22) and the study was performed in accordance with the declaration of Helsinki. Full details of the RATPAC trial had been published (13,14). The RATPAC trial was a prospective randomised controlled trial comparing point of care testing (POCT) with conventional management based on central laboratory testing (CLT). The population studied was the low risk chest pain population in whom biomarker measurement is integral to clinical decision-making. Samples from this study have now been analysed to examine the diagnostic performance of the new generation high sensitivity troponin assays. The study enrolled patients 18 years or older presenting with acute chest pain to the emergency department (ED) of 6 participating hospitals were screened for eligibility.
Glucose (coupled hexokinase method, analytical range 0.3–38.8 mmol/L) and creatinine (Jaffe alkaline picrate calibrated to a reference isotope dilution mass spectrometry method, analytical range 27–2210 μmol/L) were measured using a Beckman LX20 by the manufacturer's recommended procedures. Both assays had coefficients of variation (CV) of less than 5% across the analytical range. Estimated glomerular filtration rate (eGFR) was then calculated by the corrected diet modification of renal disease equation (15). POCT measurements were performed using the Stratus CS (Siemens Healthcare Diagnostics). The analytical characteristics of the assays were as follows: cTnI detection limit 30 ng/L, analytical range 30 to 50,000 ng/L, and inter assay CV 4.0–8.2% (67 to 344 ng/L). The 99th centile of the assay is 70 ng/L. At the individual sites cardiac troponin was measured as follows; Siemens cTnI ultra (3 sites) 99th percentile limit 40 ng/L, Abbott cTnI (1 site) 99th percentile limit 50 ng/L, Beckman
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
3
AccuTnI (1 site) 99th percentile limit 60 ng/L (locally derived value) and Roche cTnT (1 site) 99th percentile limit 10 ng/L. Central laboratory assays were as follows. For cTnI, the Architect hsTnI (Abbott Diagnostics), range 1.1–50,000 ng/L 10% CV 4.7 ng/L, Siemens cTnI Ultra (ADVIA Centaur, Siemens Healthcare Diagnostics), range 6–50,000 ng/L. 10% CV is 30 ng/L with a 99th centile of 40 ng/L. Beckman-Coulter AccuTnI enhanced (Access 2, Beckman-Coulter), range 1–100,000 ng/L, 10% CV 30 ng/L, and 99th centile 40 ng/L. For cTnT, the Roche diagnostics high sensitivity cardiac troponin T (hs-cTnT) assay (Elecsys 2010, Roche diagnostics), detection limit 3 ng/, upper limit of 10,000 ng/L, 10% CV is 13 ng/L with a 99th centile of 14 ng/L. All assays were performed by operators blinded to the clinical and diagnostic information. The Abbott hsTnI measurements were performed on an instrument in routine clinical use for a change of immunoassays. Statistical methods All statistical analyses were performed using Analyse-it for Microsoft Excel (version 2.30), www.analyse-it.com, by nonparametric methods. Absolute values for total imprecision, repeatability (within run) and between run imprecision were calculated. The imprecision profile was estimated from imprecision curves using a power curve fitting equation and estimating the 10% and 20% points from the fitted curve and by interpolation from the calculated values. The 99th percentile was calculated as the absolute single upper 99th percentile value of the one tailed distribution. Association of age and gender with cTnI was assessed by non-parametric comparison (Mann–Whitney). Diagnostic accuracy was compared by calculation of receiver operator characteristic (ROC) curves using final diagnosis and MACE as the dichotomous variables and comparison of area under the curve (AUC) by the Delong Delong Clarke–Pearson method. Sensitivity and specificity were compared by construction of 2 × 2 tables and Fishers exact probability test. Results Imprecision Total imprecision was estimated as 7.3 ng/L at the 10% point (95% confidence interval for CV 6.8 to 20.8%) with a 20% CV at 1.55 ng/L (Table 1 & Fig. 1). The 10% CV for repeatability (within run imprecision) was 1.75 ng/L and between run imprecision b2% at values greater than 5 ng/L and 3% at the lowest troponin value directly measured (4.4 ng/L). 99th percentile Full data was available on 599/1392 individuals (296 male, 303 female) who attended for screening. The median age was 58 years
Table 1 Imprecision data. Total imprecision
Within run imprecision
cTnI (ng/L)
CV (%)
Confidence interval
CV (%)
Confidence interval
4.4 5.0 6.8 6.9 8.9 10.8 11.6 43.6 54.3 74.9 367.1 693.7 1261.6
12.1 12.0 12.6 11.2 6.9 9.2 9.1 5.3 3.9 6.9 6.9 4.4 4.0
(8.0–24.7) (8.2–21.9) (8.6–22.9) (7.2–24.6) (4.6–14.1) (5.7–22.6) (5.9–20.1 (3.6–9.6) (2.7–6.8) (4.6–14.1) (4.4–15.1) (3.2–6.9) (3.0–5.8)
7.0 8.4 8.5 5.2 4.2 3.9 4.7 3.6 2.8 4.2 3.5 3.7 4.0
(5.2–10.8) (6.2–13.1) (6.3–13.2) (3.8–8.1) (3.1–6.5) (2.9–6.1) (3.4–7.2) (2.6–5.5) (2.1–4.4) (3.1–6.5) (2.6–5.5) (2.7–5.8) (2.9–6.2)
Fig. 1. Imprecision curve.
(range 45–89, lower quartile 51, upper quartile 67 years). There was no significant difference in age or sex distribution across all of the subgroups studied. The distribution in cardiac troponin between the randomly selected population, those selected by questionnaire and those selected by questionnaire, examination, ECG and imaging is summarised in Fig. 2. Progressive screening reduced the number of outliers. There were statistically significant differences in population according to gender in the three groups with higher values seen in males. In the unscreened population, cTnI correlated with age but this correlation disappeared in the other two groups. Troponin was detectable in 99.5% of the population. Reference intervals are summarised in Table 2. There were insufficient numbers to produce a meaningful male and female distribution for the highly screened population. The highly screened male, but not female, population distribution corresponded to a normal distribution (data shown in Supplementary Fig. 1). Excluding one outlier yielded a 99th percentile of 8.5 ng/L in the female population but the number of samples was small, 115. Diagnostic test performance Data was available for 342 patients, 202 male 140 female median age 53.8 years (range 23.7–90.6 lower quartile 44.2, upper quartile 64.7). Median time from onset of chest pain to first sample was 53.8 min (range 55–7630, lower quartile 168.7, upper quartile 370 min). The AUC for the admission sample was 0.91 (CI 0.83–0.98), for the 90 minute sample 0.98 (CI 0.95–1.00) and for the peak value 0.90 (CI 0.81–0.99) and each (admission, 90 minute and peak) was statistically indistinguishable from the other troponin assays. Data for the admission sample is shown in Fig. 3. Full details of the area under the ROC curves and AUCs are given in the online data appendix (Supplementary Figs. 2 and 3 and Supplementary Tables 1–3). The Abbott hsTnI assay predicted MACE with the AUC on admission of 0.84 (CI 0.73–0.96) and for the peak value of 0.84 (CI 0.70–0.97) with the same diagnostic efficiency as the other assays (Fig. 4 for the admission sample, Supplementary Fig. 4, and Supplementary Tables 4 and 5). Because of the male–female difference, the data was divided and the analysis repeated for males and females separately. There was no significant difference in test performance for females although the Abbott hsTnI assay was statistically better than the Siemens assay on the admission sample in males for both diagnosis and prognosis (data not shown). There was a trend to an influence of gender on diagnostic efficiency when using the 99th percentile but due to small numbers of positive cases this failed to reach statistical significance. When a diagnostic discriminant of 21 ng/L was used overall diagnostic sensitivity was 61.3% (CI 42.2–78.2) on admission with a specificity of 98.5% (CI 96.2–99.6).
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
4
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
Fig. 2. Distribution of cTnI values in the three groups studied. 99th percentiles were as follows: All data 21.0 ng/L (n = 599), Questionnaire exclusion 12.3 ng/L (n = 366), and highly selected population 8.7 ng/L (n = 208).
For males at a diagnostic discriminant of 21 ng/L sensitivity was 77.8% (CI 52.4–93.6). Using a sex-specific cut-off of 28.6 ng/L for males sensitivity was reduced to 66.7% (CI 41.0–86.7) p = 0.71 for comparison of sensitivity at the two different cut-offs, with specificity remaining at 98.1% (CI 94.6–99.6). In female sensitivity at 21 ng/L was 38.5% (CI 13.9–68.4)
and increased to 69.2% (CI 38.6–90.9) p = 0.24 for comparison of sensitivity at the two different cut-offs with a small decrease in specificity from 99.0% (CI 94.8–100) to 96.2% (CI 90.4–98.9) p = 0.37 for comparison of specificity at the two different cut-offs when a sex specific cut off of 9.9 ng/L is used.
Table 2 Reference intervals for three population groups studied. Population
99th percentile ng/L all data [confidence interval] (n)
99th percentile ng/L—male (n)
99th percentile ng/L—female (n)
p (male vs. female)
All Plus questionnaire screening Plus normal imaging, ECG, spirometery and biochemistry
21.0 [11–44] (599) 12.3 [6.2–21] (366) 8.7 (208)
28.5 (296) 18.3 (171)
9.9 (303) 9.5 (195)
b0.0001 0.0002 0.0009
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
5
Fig. 3. Receiver operating characteristic curves for samples taken on admission for a final diagnosis of acute myocardial infarction. cTnI A 1 = Abbott hs cTnI admission sample, cTnI CS 1 = Stratus CS cTnI admission minute sample, cTnI S 1 = Siemens ultra cTnI admission sample, cTnI B 1 = Beckman cTnI admission sample, and cTnT 1 = Roche hs cTnT admission sample.
Discussion Surveys of routine use of cardiac biomarkers have shown that the manufacturer's datasheets and published literature are the commonest source of performance data and decision limits (16,17). It is therefore
important to provide independent assessment of claims by manufacturer. Performance claims within datasheets are not always reflected by experience in routine clinical use. This is because most laboratory analysers in routine clinical use are performing multiple sequential measurements on different sample types and of different analytes. They are therefore
Fig. 4. Receiver operating characteristic curves for samples taken on admission to predict major adverse cardiac events. cTnI A 1 = Abbott hs cTnI admission sample, cTnI CS 1 = Stratus CS cTnI admission minute sample, cTnI S 1 = Siemens ultra cTnI admission sample, cTnI B 1 = Beckman cTnI admission sample, and cTnT 1 = Roche hs cTnT admission sample.
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
6
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx
functioning under less than ideal conditions. This study performed measurements on an analyser in routine clinical use. The 10% CV imprecision claim for the Abbott hs cTnI assay is 4.7 ng/L. We found a value of 7.3 ng/L at the 10% point (95% confidence interval for CV 6.8 to 20.8%). This represents the difference between the CV to be expected in routine clinical use and the CV obtained under more optimal conditions. The value obtained is similar to but higher than the 5.6 ng/L found in a multicentre evaluation of this analyser (18). The CV obtained is below the 99th percentile for the assay and the reproducibility of the assay was excellent with the repeatability between results less than 2% with a 10% CV of 1.75 ng/L. Clinically this will translate into good clinical performance. The 10% CV is one third of the combined 99th percentile and the repeatability means that sequential measurements for detection of a delta troponin value can be undertaken over short time intervals with a high degree of confidence. Recent recommendations for high sensitivity assays suggest that should very short time intervals between sequential sampling can be used, certainly on admission, 90 min, 2 h and 3 h post admission (19–22) and possibly admission and 1 h from admission (23). In order to use such short time intervals safely a high degree of repeatability between sequential measurements is required. It would seem that the Abbott hs cTnI assay meets these requirements. The results of the reference range study are consistent with previous studies on high sensitivity troponin assays and show that progressive exclusion of probable underlying structural heart disease causes a decreased troponin upper reference limit to be obtained, more marked in men than in women (10,24). A distinct difference seen between men and women with a higher value in men, supports the use of gender specific cut-offs (25). The diagnostic efficiency of the assay was similar to that of two contemporary sensitive assays. However, the number of patients with a final diagnosis of AMI in the sample population was relatively modest at 10.7% (33/309) although it did correspond to a low risk chest pain group. It has been suggested that the Abbott hs cTnI assay shows greater sensitivity in the early phase of AMI (26). Care must be taken in interpreting apparent analytical diagnostic performance differences as they may not always translate into significant differences in prognostic ability (27). There are limitations to this study. The samples for the reference population had been in long term storage (7 years). Preliminary data on freeze thaw stability and studies from other cTnI methods did not suggest that significant sample deterioration would have occurred but this cannot be categorically excluded, but the data are consistent with other studies. The reduction in sample numbers on extensive screening is a consequence of the study design but the data are consistent with previously published work of this type. The relatively small numbers of patients with a final diagnosis of AMI, although representative of an Emergency Department population, the group of interest, means that the study is not powered enough to detect small differences. This would require a large multicentre study. Conclusion The Abbott high sensitivity troponin I assay meets the performance characteristics of a high sensitivity assay as recently recommended (7,8). The diagnostic performance in patients with chest pain is comparable with other contemporary and high sensitivity assays. Acknowledgements The authors thank Abbott diagnostics for providing free reagents and supporting the cost of analyses of the samples by an independent laboratory as an unrestricted grant. Abbott diagnostics had no influence over the study design, analytical process, data analysis or preparation, submission or findings of the manuscript. The authors have no conflict of interest and have received no honoraria in respect of the study or any related activities.
Professor Collinson acts as a guarantor in respect of data integrity and study transparency and as study sponsor.
Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.clinbiochem.2014.12.017.
References [1] Melanson SE, Morrow DA, Jarolim P. Earlier detection of myocardial injury in a preliminary evaluation using a new troponin I assay with improved sensitivity. Am J Clin Pathol 2007;128:282–6. [2] Reichlin T, Hochholzer W, Bassetti S, Steuer S, Stelzig C, Hartwiger S, et al. Early diagnosis of myocardial infarction with sensitive cardiac troponin assays. N Engl J Med 2009;361: 858–67. [3] Keller T, Zeller T, Peetz D, Tzikas S, Roth A, Czyz E, et al. Sensitive troponin I assay in early diagnosis of acute myocardial infarction. N Engl J Med 2009;361:868–77. [4] Eggers KM, Oldgren J, Nordenskjold A, Lindahl B. Diagnostic value of serial measurement of cardiac markers in patients with chest pain: limited value of adding myoglobin to troponin I for exclusion of myocardial infarction. Am Heart J 2004;148:574–81. [5] Collinson P, Goodacre S, Gaze D, Gray A. Very early diagnosis of chest pain by point-ofcare testing: comparison of the diagnostic efficiency of a panel of cardiac biomarkers compared with troponin measurement alone in the RATPAC trial. Heart 2012;98:312–8. [6] Mills NL, Churchhouse AM, Lee KK, Anand A, Gamble D, Shah AS, et al. Implementation of a sensitive troponin I assay and risk of recurrent myocardial infarction and death in patients with suspected acute coronary syndrome. JAMA 2011;305:1210–6. [7] Apple FS. A new season for cardiac troponin assays: it's time to keep a scorecard. Clin Chem 2009;55:1303–6. [8] Apple FS, Collinson PO. Analytical characteristics of high-sensitivity cardiac troponin assays. Clin Chem 2012;58:54–61. [9] Apple FS, Jaffe AS, Collinson P, Mockel M, Ordonez-Llanos J, Lindahl B, et al. IFCC educational materials on selected analytical and clinical applications of high sensitivity cardiac troponin assays. Clin Biochem 2014. [10] Collinson PO, Heung YM, Gaze D, Boa F, Senior R, Christenson R, et al. Influence of population selection on the 99th percentile reference value for cardiac troponin assays. Clin Chem 2012;58:219–25. [11] Schiller NB, Shah PM, Crawford M, DeMaria A, Devereux R, Feigenbaum H, et al. Recommendations for quantitation of the left ventricle by two-dimensional echocardiography. American Society of Echocardiography Committee on Standards, Subcommittee on Quantitation of Two-Dimensional Echocardiograms. J Am Soc Echocardiogr 1989; 2:358–67. [12] Devereux RB, Alonso DR, Lutas EM, Gottlieb GJ, Campo E, Sachs I, et al. Echocardiographic assessment of left ventricular hypertrophy: comparison to necropsy findings. Am J Cardiol 1986;57:450–8. [13] Goodacre SW, Bradburn M, Cross E, Collinson P, Gray A, Hall AS. The Randomised Assessment of Treatment using Panel Assay of Cardiac Markers (RATPAC) trial: a randomised controlled trial of point-of-care cardiac markers in the emergency department. Heart 2011;97:190–6. [14] Goodacre S, Bradburn M, Fitzgerald P, Cross E, Collinson P, Gray A, et al. The RATPAC (Randomised Assessment of Treatment using Panel Assay of Cardiac markers) trial: a randomised controlled trial of point-of-care cardiac markers in the emergency department. Health Technol Assess 2011;15:1–108. [15] Levey AS, Coresh J, Greene T, Marsh J, Stevens LA, Kusek JW, et al. Expressing the Modification of Diet in Renal Disease Study equation for estimating glomerular filtration rate with standardized serum creatinine values. Clin Chem 2007;53:766–72. [16] Collinson P, Pulkki K, Suvisaari J, Ravkilde J, Stavljenic-Rukavina A, HammererLercher A, et al. How well do laboratories follow guidelines on cardiac markers? The cardiac marker guideline uptake in Europe study. Clin Chem 2008;54:448–9. [17] Collinson PO, Dieijen-Visser MP, Pulkki K, Hammerer-Lercher A, Suvisaari J, Ravkilde J, et al. Evidence-based laboratory medicine: how well do laboratories follow recommendations and guidelines? The Cardiac Marker Guideline Uptake in Europe (CARMAGUE) Study. Clin Chem 2012;58:305–6. [18] Krintus M, Kozinski M, Boudry P, Capell NE, Koller U, Lackner K, et al. European multicenter analytical evaluation of the Abbott ARCHITECT STAT high sensitive troponin I immunoassay. Clin Chem Lab Med 2014;52:1657–65. [19] Hamm CW, Bassand JP, Agewall S, Bax J, Boersma E, Bueno H, et al. ESC guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: the task force for the management of acute coronary syndromes (ACS) in patients presenting without persistent ST-segment elevation of the European Society of Cardiology (ESC). Eur Heart J 2011;32:2999–3054. [20] Collinson P, Gaze D, Goodacre S. Comparison of contemporary troponin assays with the novel biomarkers, heart fatty acid binding protein and copeptin, for the early confirmation or exclusion of myocardial infarction in patients presenting to the emergency department with chest pain. Heart 2014;100:140–5. [21] Cullen L, Mueller C, Parsonage WA, Wildi K, Greenslade JH, Twerenbold R, et al. Validation of high-sensitivity troponin I in a 2-hour diagnostic strategy to assess 30-day outcomes in emergency department patients with possible acute coronary syndrome. J Am Coll Cardiol 2013;62:1242–9. [22] Than M, Aldous S, Lord SJ, Goodacre S, Frampton CM, Troughton R, et al. A 2-hour diagnostic protocol for possible cardiac chest pain in the emergency department: a randomized clinical trial. JAMA Int Med 2014;174:51–8.
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
P.O. Collinson et al. / Clinical Biochemistry xxx (2015) xxx–xxx [23] Reichlin T, Schindler C, Drexler B, Twerenbold R, Reiter M, Zellweger C, et al. One-hour rule-out and rule-in of acute myocardial infarction using high-sensitivity cardiac troponin T. Arch Intern Med 2012;172:1211–8. [24] Koerbin G, Abhayaratna WP, Potter JM, Apple FS, Jaffe AS, Ravalico TH, et al. Effect of population selection on 99th percentile values for a high sensitivity cardiac troponin I and T assays. Clin Biochem 2013;46:1636–43. [25] Eggers KM, Johnston N, James S, Lindahl B, Venge P. Cardiac troponin I levels in patients with non-ST-elevation acute coronary syndrome—the importance of gender. Am Heart J 2014;168:317–24.
7
[26] Rubini GM, Twerenbold R, Reichlin T, Wildi K, Haaf P, Schaefer M, et al. Direct comparison of high-sensitivity-cardiac troponin I vs. T for the early diagnosis of acute myocardial infarction. Eur Heart J 2014;35:2303–11. [27] Venge P, Lindahl B. Cardiac troponin assay classification by both clinical and analytical performance characteristics: a study on outcome prediction. Clin Chem 2013;59:976–81.
Please cite this article as: Collinson PO, et al, The clinical and diagnostic performance characteristics of the high sensitivity Abbott cardiac troponin I assay, Clin Biochem (2015), http://dx.doi.org/10.1016/j.clinbiochem.2014.12.017
