| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
|
|
|
||||||
|
||||||||
Electronic Letters to:
|
|
Electronic letters published:
![[Read eLetter]](/icons/misc/sectionDOWN.gif){kind=icon}
Plasma 99th percentile reference limit for cardiac troponin
Physiological circulating cardiac troponin. Is the 99th percentile achievable & clinically relevan
|
|
|||
|
Daylily S. Ooi, Medical Biochemist Department of Laboratory Medicine, Ottawa Hospital, Canada, George A. Wells, Chair and Professor, Department of Epidemiology and Community Medicine, University of Ottawa, Canada
Send letter to journal:
dsooi{at}ottawahospital.on.ca Daylily S. Ooi, et al.
|
The paper by Apple et al (1) on the 99th percentile reference limits for serum cardiac markers provides much needed and useful information as we grapple with the use of these limits in acute coronary syndromes (ACS). There are two points we would like to add. Firstly, one notices that only Roche’s cardiac troponin T (cTnT) assay has the same 99th percentile limit as quoted by the manufacturer (<0.010 ug/L). The value is consistent across gender and ethnic groups, and below the 10% coefficient of variation (CV) cut-off. While the former is obviously advantageous, the latter has implications regarding estimation of imprecision. The authors stated that only the Tosoh assay met the 10% CV target at the 99th percentile limit. With the Roche cTnT value of <0.010 ug/L, precision cannot be determined in the conventional way; rather, one has to determine how often a <0.010 ug/L remains as such on re-analysis. We have found <0.010 ug/L very reproducible (2). On repeat testing of 141 clinical samples with initial cTnT values of < 0.040 ug/L, 108 of 111 samples remained as <0.010 ug/L; of the 30 samples with measurable cTnT initially, only 1 sample (cTnT 0.016 ug/L) gave a value of <0.010 ug/L. This gives an error rate of 2.8% if <0.010 ug/L is used as the cut-off. In contrast, the error rate with a numerical cut-off will be considerably higher as equal numbers will fall on either side of the mean in a Gaussian distribution. To illustrate this, we used data from our laboratory to determine the error rate that would arise using 0.030 ug/L as the cut-off. Applying the theoretical CV of 10% and 90% confidence limits, we calculated the number of results (area of the Gaussian curve) that would fall on the other side of the cut-off for each value. Of 154 results between 0.021 and 0.039 ug/L (+ 3SD), 22 of 85 results initially reported as 0.029 ug/L or lower would on repeat give a value 0.030 ug/L and above, and 25 of 69 with initial values of > 0.030 ug/L would give the opposite error, for a total misclassification rate of 31%. Although the actual percentage will vary depending on the distribution around the cut-off, it will not be substantially different as error arises mainly from the variability around the cut-off. This also explains why improving CV will not reduce the misclassification rate drastically. As all cTnI assays have numerical cut-offs, one could project the error rates to be in this region. We confirmed this high negative predictive value of cTnT <0.010 ug/L for ACS. In an audit of cardiac markers utilization in our hospital (a tertiary care general hospital with an affiliated cardiac centre where total creatine kinase (CK) is performed on all cTnT requests), 34 out of 373 (9%) patients with initial cTnT <0.010 ug/L, and 53 of 239 (22%) with cTnT between 0.010 and 0.099 ug/L, had ACS. When taking temporal CK changes into consideration, the difference was more marked. Where there was no subsequent doubling of CK, only 2% of patients with initial cTnT <0.010 ug/L had ACS, while 15% of patients with measurable cTnT had ACS; where CK had doubled, the percentages were considerably higher at 15% and 43% respectively. This brings us to the second point. The authors discussed the increased diagnosis of ACS resulting from these lower reference limits. This is to be expected, as not only would minor ischemia be detected, other myocardial pathology can result in release of intracellular contents. Minor elevations have been observed in congestive heart failure, cardiomyopathies, end-stage renal failure, sepsis and hypovolemic shock; however, they do not show the temporal changes seen with ACS. We advocate that all laboratories performing cTnT capitalize on the robustness of the <0.010 ug/L result, both analytically and clinically, and to use it instead of a measured value such as the 10% CV cut-off. Additionally, we stress the importance of temporal changes in cardiac markers in the management of patients with suspected ACS when using these low reference cut-offs based on 99th percentile of reference population. References: 1. Apple FS, Quist HE, Doyle PJ, Otto AP, Murakami MM. Plasma 99th percentile reference limits for cardiac troponin and creatine kinase MB mass for use with European Society of Cardiology/American College of Cardiology Consensus recommendations. Clin Chem 2003;49:1331-6. 2. Ooi DS, Black HS. Low-end precision of cardiac troponin T – does it meet recommended guidelines? (abstract). Clin Biochem 2002;35:339. |
|||
|
|
|||
|
David C Gaze, Cardiac Research Scientist St Georges Hospital London, Paul O Collionson
Send letter to journal:
davidtroponin{at}hotmail.com David C Gaze, et al.
|
The recent paper by Apple and colleagues describing the plasma 99th percentile reference limits for cardiac troponin (cTn) and creatine kinase MB mass (1), along with the accompanying editorial (2) raises some interesting questions that require informed discussion. It has been demonstrated that any elevation (above the 99th centile of normal) of cardiac troponin T (cTnT) identifies patients at an increased risk of subsequent mortality (3). A caveat with this study is that it represents a highly selected clinical trial population and not the general chest pain group. The recent redefinition of acute myocardial infarction (4) places emphasis on having an analytical coefficient of variation of <10% at the 99th centile. What is the evidence base for a figure of 10% apart from expert consensus? Why this figure has been selected deserves discussion and justification. As shown in this paper, no assay currently meets these criteria, apart from the AIA 600 II cTnI assay (Tosoh Bioscience). In our experience of this assay (unpublished data), we agree that the CV at this level is below 10% (0.04 ug/L) whilst the 99th percentile of 484 apparently healthy individuals is 0.08 ug/L. Yeo and colleagues (5), who have assayed plasma pools at four levels of cTn over an 8-10 week period on 5 cTnI assays and the Roche Elecsys cTnT assay, have argued for, and determined a 20% coefficient of variation. Quality specifications for cTn assays described by both the International Federation of Clinical Chemistry (IFCC), (6) and National Association of Clinical Biochemistry (NACB), (7), do not have a justification or a standard protocol for determination of 10% CV. There is a need for a clear definition of how this should be achieved experimentally. Pagani and colleagues representing the Committee of Standardization of Markers of Cardiac Damage (C-SMCD) by the International Federation of Clinical Chemistry have evaluated the 99th percentile and 10% CV concentration of 14 commercial troponin assays (8). The authors have calculated the 10% CV/99th percentile ratio. The protocol used was two replicates of eight serum pools per run per day over 20 days using two reagent lots and three calibrations. This protocol reflects routine use of the analyser and is a more realistic assessment of analytical variation. The second question is whether this actually has clinical relevance. Is there truly a reference level for cTnT and cTnI? This has not been our experience with the Roche cTnT assay, the Tosoh AIA 600 II, Beckman AccuTnI and Dade-Behring RxL cTnI assays to date. Clinically reporting levels with imprecision higher than 10% could lead to misclassification of myocardial infarction. Can the clinical chemistry laboratory reliably offer precise cTn measurements at such low levels? Currently, the answer is no. Newer assays are demonstrating better but not satisfactory precision (9). Advances in cTn assays are occurring. It is time that standard protocols and definitions are written with scientific and experimental justification to advance quality specifications for future assays. This is currently being addressed by the IFCC (6). David C. Gaze Paul O. Collinson References: 1. Apple FS, Quest HE, Doyle PJ, Otto AP, Murakami MA. Plasma 99th Percentile Reference Limits for Cardiac Troponin and Creatine Kinase MB Mass for Use with European Society of Cardiology/American College of Cardiology Consensus Recommendations. Clinical Chemistry 2003;49:1331-1336 2. Giannitisis E & Katus HA. 99th Percentile and Analytical Imprecision of Troponin and Creatine Kinase-MB Mass Assays: An objective Platform for Comparison of Assay Performance. Clinical Chemistry 2003;49:1248-1249 3. James S, Armstrong P, Califf R, Simmons M, Venge P, Wallentin L, Lindahl B. Troponin T Levels and Risk of 30-Day Outcomes in Patients with the Acute Coronary Syndrome: Prospective Varification in the GUSTO-IV Trial. Am. J. Med. 2003;115:178-184 4. The Joint European Society of Cardiology/American College of Cardiology Committee. Myocardial infarction redefined - a consensus document of the Joint European Society of Cardiology/American College of Cardiology Committee for the Redefinition of Myocardial Infarction. J Am Coll. Cardiol. 2000;36:959-69. 5. Yeo K, Quinn-Hall K, Bateman S, Fischer G, Wieczorek S, Wu A. Functional sensitivity of cardiac troponin assays and its implications for risk stratification for patients with acute coronary syndromes. In Adams J, Apple F, Jaffe A, Wu A eds. Markers in cardiology: current and future clinical applications. Am. Heart Association Monograph Series. Armonk, NY: American Heart Association 2002:23-30. 6. Panteghini M, Gerhardt W, Apple FS, Dati F, Ravkilde J, Wu AHB. Quality Specifications for Cardiac Troponin Assays. Clin. Chem. Lab. Med. 2001;39:174-178 7. Wu AHB, Apple FS, Gibler WB, Jesse RL, Warshaw NM, Valdes R Jr. National Academy of Clinical Biochemistry Standards of Laboratory Practice: recommendations for use of cardiac markers in coronary artery disease. Clin. Chem. 1999;45:1104-21 8. Pagani F, Yeo J, Apple F, Christenson R, Dati F, Mair J, Ravkilde J, Wu A, Panteghini M. Evaluation of the imprecision at the low-range concentration of he assays for cardiac troponin determination. Clin. Chem. 2003;49(S6):A34-5 9. Uettwiller-Geiger D, Wu AHB, Apple FS, Jevans AW, Venge P, Olson MD et al. Multicenter evaluation of an automated assay for troponin I. Clin. Chem. 2002;48:869-76 |
|||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH |
eLetters: