True 99th centile of high sensitivity cardiac troponin for hospital patients: prospective, observational cohort study
BMJ 2019; 364 doi: https://doi.org/10.1136/bmj.l729 (Published 13 March 2019) Cite this as: BMJ 2019;364:l729
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
This an extremely useful study, and also potentially extremely valuable from a research ethics standpoint. After wide consultation and careful consideration (Health Research Authority, Confidentiality Advisory Group & British Cardiac Patients Association) patient consent was deemed unnecessary - not only for additional testing of drawn blood, but also for (selective) withholding, from patients and clinicians, of troponin assay results above the recommended upper limit of normal (ULN). Had patient consent been necessary, the bureaucratic load for the study would have increased enormously [1], and arguably may well have killed it. It would therefore be extremely interesting to learn more of the discussions and decisions surrounding the ethics of this study.
We think that such insight could guide other researchers on acceptable strategies to help reduce bureaucracy associated with unnecessary consenting of patients. It might make possible studies that would otherwise be impossible, and would almost certainly reduce costs and clinicians' time whilst still maintaining the quality, integrity and safety of clinical research - upon which evidence based medicine crucially depends.
1. Bureaucracy is strangling clinical research. BMJ 2019;364:l1097
Competing interests: No competing interests
The use or overuse of diagnostic testing – particular laboratory testing – is a question for our time, as ever more novel biomarkers are at our disposal [1,2].
It is tempting to infer certainty of dichotomous classification into normality or abnormality from laboratory testing - in deference to the low analytical error rate of biochemistry. As such, there is often a large overestimation of the positive predictive value of a test [3]. However, health may be considered a relative and not an absolute state. Clinicians operate in shades of grey.
Each test we utilise should be performed on the understanding that it shifts the pre-test probability of having an illness, as it seldom gives absolute certainty one way or the other. Mariathas and colleagues’ paper neatly draws attention to this fact [4].
1. Freedman DB. Towards Better Test Utilization – Strategies to Improve Physician Ordering and Their Impact on Patient Outcomes. EJIFCC 2015; 26(1): 15-30
2. Mandl KD & Manrai AK. Potential Excessive Testing at Scale. Biomarkers, Genomics, and Machine Learning. JAMA 2019; 321(8): 739-740
3. Manrai AK et al. Medicine’s Uncomfortable Relationship With Math. Calculating Positive Predictive Value JAMA 2014;174(6):991-993.
4. Mariathas M et al. True 99th centile of high sensitivity cardiac troponin for hospital patients: prospective, observational cohort study BMJ 2019; 364:l729
Competing interests: No competing interests
Referring to the recent article, we would like to raise a few points about the use of cardiac troponin (cTn) and its interpretation.
The value of cTn is in aiding the clinician to rule out or diagnose acute myocardial infarction (AMI) in patients with the appropriate clinical milieu. In this study, only 9.1% (n = 1829) of the 20,000 patients had AMI or clinical indications for cTn testing; of this group tested, 6.1% (n = 122) had AMI. As mentioned in the article, several factors such as age, sex, glomerular filtration rate, left ventricular function and the presence of major inflammatory conditions can cause falsely elevated cTn values. The cTn 99th percentile upper reference limits (URL) derived from a healthy population would reduce interference from other background conditions which elevate troponin as opposed to hospital patients.
In some quarters, the use of a single troponin to rule out AMI is advocated; however, there is danger that early presenters may be missed. Serial testing is essential to ascertain the presence or absence of myocardial infarction even if an initial reading is normal in a patient with suspicion of AMI. Although cTn above the 99th percentile URL is a diagnostic criteria for AMI, guidelines exist for the use of the percentage increment between initial and second test values; for initial baseline values ≤ the 99th percentile URL, a 50-60% rise in cTn is regarded as significant, while a >20% change is considered significant for the initial cTn >99th percentile URL (1).
The Beckman Coulter Access AccuTnI+3 assay used in this study is a contemporary assay. Beckman has only recently unveiled their high-sensitivity cTnI assay. Results on one assay cannot be compared to another (2).
This study adds to literature that we must be mindful of other causes of raised troponin that can lead to spurious result interpretation if used inappropriately.
References:
(1) Thygesen K, Alpert JS, Jaffe AS, Chaitman BR, Bax JJ, Morrow DA, et al. Fourth universal definition of myocardial infarction (2018). Eur Heart J 2019; 40:237-69.
(2) Christenson RH, Jacobs E, Uettwiller-Geiger D, Estey MP, Lewandrowski K, Koshy TI, et al. Comparison of 13 commercially available cardiac troponin assays in a multicentre north American study. J Appl Lab Med 2017; 1:544-61.
Competing interests: No competing interests
The Authors conducted a large study in patients presenting to a hospital to assess value of high sensitivity cardiac troponin I. We highly appreciate their encouraging efforts on opening a new avenue of information and broadening our knowledge on this not-much explored subject. But when we went through the study, we found something strange in Table 2.
This table has a title, "Distribution of hs-cTnI (ng/L) according to location when the biochemistry test was requested”. So the table locates a patient in a hospital when the blood test for cardiac enzyme - a marker of myocyte injury - was requested. In the table, we observe that 39.02% (48) patients admitted to critical care units had hs- cTnI >ULN , while only 21.56% (58) admitted to cardiac wards had the finding. So it appears that those patients in critical care units have a higher probability of having a rise in “cardiac” enzymes on their blood tests than those admitted to cardiac wards. And patients of critical care units have been found to have the highest level of marker of cardiac injury too in this study. This observation is somewhat contrary to conventional wisdom. Conventionally we learn that cardiac wards house cardiac patients and they are likely to have sustained cardiac injury and hence have high levels of cardiac enzymes due to their release from cardiac cells.
So we pondered over it and imagined something. Although the Authors write under the title ‘Limitations of the study’ that they did not go to diagnosis of their patient population as the blood sample for the study was drawn during presentation to hospital whereas final diagnosis is made later on (and which can be retrieved by discharge summary data). What happens in my hospital is that when all the beds in cardiac wards are full, patients are initially admitted to medical wards or to ICU, and later on they are shifted to cardiac wards - when beds are available there. Therefore critical care units in my hospital may have several unstable cardiac patients as well - some of which may have a rise in their cardiac enzyme level. So we want to know if a similar practice is adopted in this hospital too, where this exploratory study is carried out. We are interested to know if patients in the critical care unit in this study included a few cardiac patients too, which may have resulted in the observation of mixed results. We know that the Authors did not make a protocol to gather information about the final diagnosis of their patients, but were cardiac patients admitted to critical care units according to hospital policy, and included in the study? That may be one explanation for the discovery of a higher cardiac enzyme level in patients admitted to critical care units, as compared to those admitted to cardiac wards.
Competing interests: No competing interests
When Conventional troponin testing was initially introduced, it aided rapid, accurate diagnosis or exclusion of Myocardial infarction. The high predictive value of conventional assays enabled rapid discharge of patients presenting to Emergency departments. But the newer high-sensitivity troponin (hs-T) assays have inevitably traded off specificity for sensitivity in general hospital population.[1] [2]. For instance, the introduction of sensitive assays at Brigham and Women's Hospital seem to have doubled positive results with no change in incidence of acute coronary syndrome. [1]. The costlier newer assays have paradoxically increased overall health costs due to increased hospital admissions and interventions.[2] [3] .
Interpretation of raised levels from a hs-T assay in the context of clinical presentation and use of pre-test probability is good in theory but it is inevitably subjective and hence exposes a clinician to significant medico-legal risk if raised troponin levels are ignored in a patient with high mortality risk.[4].
This cohort study adds weight to the suggestion that the cheaper conventional troponin assays should be used in most settings except in evaluation of chest pain in cardiac units and emergency departments.
Use of a less sensitive test in clinical practice is not without precedent. For instance, ultra-sensitive PSA (prostate specific antigen) tests are not widely used in NHS because they do not provide significant additional clinical value in most prostate cancer patients.[5].
References
1 Bhoi S, Verma P, Vankar S, et al. High sensitivity troponins and conventional troponins at the bedside. Int J Crit Illn Inj Sci 2014;4:253–6. doi:10.4103/2229-5151.141471
2 Kaambwa B, Ratcliffe J, Horsfall M, et al. Cost effectiveness of high-sensitivity troponin compared to conventional troponin among patients presenting with undifferentiated chest pain: A trial based analysis. Int J Cardiol 2017;238:144–50. doi:10.1016/j.ijcard.2017.02.141
3 High-Sensitivity Cardiac Troponin for the Rapid Diagnosis of Acute Coronary Syndrome in the Emergency Department: A Clinical and Cost-Effectiveness Evaluation. CADTH Technol Overv 2013;3.https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3711526/ (accessed 15 Mar 2019).
4 Mariathas M, Allan R, Ramamoorthy S, et al. True 99th centile of high sensitivity cardiac troponin for hospital patients: prospective, observational cohort study. BMJ 2019;364:l729. doi:10.1136/bmj.l729
5 Ellis WJ, Vessella RL, Noteboom JL, et al. Early detection of recurrent prostate cancer with an ultrasensitive chemiluminescent prostate-specific antigen assay. Urology 1997;50:573–9. doi:10.1016/S0090-4295(97)00251-3
Competing interests: No competing interests
Mariathas et al investigated the distribution of high sensitivity cardiac troponin I concentrations in 20000 consecutive patients undergoing a blood test for any clinical reason at a UK hospital [1]. The results showed that one in 20 (5.4%, n=1080) had a high sensitivity cardiac troponin I greater than the recommended upper limit of normal. There was no clinical suspicion of acute myocardial infarction for most of these patients. Advancing age, male sex, decreasing estimated glomerular filtration rate, and inpatient location (vs. outpatient department) were independent predictors of a patient having a high sensitivity cardiac troponin I concentration greater than the recommended upper limit of normal. These results suggest that using a threshold value of upper limit of normal to diagnosis acute myocardial infarction is less specific.
Actually, previous studies and expert consensus already pointed out that elevated troponin in itself does not indicate acute myocardial infarction because it can be caused by varied factors and conditions [2, 3]. The statistical issue of pre-test probability should be considered when interpreting the result of elevated troponin. This means the diagnosis of myocardial infarction using troponin levels should be in the context of the pre-test probability of acute coronary syndrome. Factors associated with a high pre-test probability of acute coronary syndrome including typical symptoms, ischemic ECG changes or wall-motion abnormalities on echocardiography, and the presence of coronary artery disease risk factors or history of coronary artery disease. Even with an elevated troponin, the post-test probability for acute coronary syndrome is still low in a patient with low pre-test probability of acute coronary syndrome. Therefore, it was recommended that the testing of troponin should be performed only if clinically indicated for suspected myocardial infarction [3].
However, there is actually a dilemma in the clinical practice for many doctors in China. Although many Chinese doctors know that it is not specific for myocardial infarction, the testing of troponin is still widely ordered by them, especially in wards such as intensive care units. Due to the unfriendly doctor-patient relationship in China [4], doctors are afraid of missing a diagnosis, which leads to many unnecessary subsequent tests. We hope that this dilemma will be resolved in the near future.
Yu-Ming Xu, Peng-Peng Niu, Bo Song. The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China.
References
1. Mariathas M, Allan R, Ramamoorthy S, et al. True 99th centile of high sensitivity cardiac troponin for hospital patients: prospective, observational cohort study. BMJ. 2019 Mar 13;364:l729. doi: 10.1136/bmj.l729.
2. Eggers KM, Lind L, Venge P, Lindahl B. Factors influencing the 99th percentile of cardiac troponin I evaluated in community-dwelling individuals at 70 and 75 years of age. Clin Chem. 2013 Jul;59(7):1068-73. doi: 10.1373/clinchem.2012.196634.
3. Newby LK, Jesse RL, Babb JD, et al. ACCF 2012 expert consensus document on practical clinical considerations in the interpretation of troponin elevations: a report of the American College of Cardiology Foundation task force on Clinical Expert Consensus Documents. J Am Coll Cardiol. 2012 Dec 11;60(23):2427-63. doi: 10.1016/j.jacc.2012.08.969.
4. Jie L. New generations of Chinese doctors face crisis. Lancet. 2012 May 19;379(9829):1878. doi: 10.1016/S0140-6736(12)60774-0.
Competing interests: No competing interests
Re: True 99th centile of high sensitivity cardiac troponin for hospital patients: prospective, observational cohort study
The study by Mariathas and colleagues illustrates the confusion which clinicians seem to find with the development of troponin assays of increasing analytical sensitivity. First there is the problem with nomenclature. In their study they consider that they have evaluated a high sensitivity assay. Unfortunately, this is not correct as the assay they have chosen, the Beckman Coulter access would be considered to be a contemporary sensitive assay although its analytical characteristics approach those of high sensitivity assays. Similarly, and in common with many other studies reporting evaluation of troponin assays, they do not report the assay characteristics correctly. They report the limit of quantification as the 10% CV at 20 ng/L. The limit of quantification is in fact the 20% CV (at 20 ng/L) and the limit of detection is reported in the package insert as <10 ng/L(1).
The authors have measured troponin in all comers exempting only those where a diagnosis of myocardial infarction is under consideration. Their findings merit some more detailed analysis. In the outpatient setting they found a 99th percentile of 65 ng/L. This is identical to the value we obtained in a randomly selected population from general practice(2). This population was thoroughly investigated and when patients with comorbidities or underlying structural heart disease were excluded, the 99th percentile fell to 40 ng/L. Our study is in agreement with others which show that troponin elevation in the apparently healthy population indicates underlying cardiac disease(3) and is prognostic in the medium to long-term(4;5). A second point should be made here is that they found that 2% of the outpatient population exceeded 40 ng/L. By definition, the 99th percentile is 1% of the population and their finding illustrates very clearly the problem of skewed distributions affecting the value obtained for the 99th percentile. The authors also find that as troponin is measured in a progressively more unwell proportion of the population (patients do not normally attend the Emergency Department for fun) the proportion of troponin is which are elevated increases and when the sickest population is considered (patients on intensive care) the proportion is the greatest. This is not a novel finding, troponin elevation in intensive care is well documented and is prognostic(6;7).
The authors mention the binary use of troponin. The shift to the 99th percentile requires an understanding of the effect of the redefinition of myocardial infarction. The previous definition included a biomarker component of twice the upper reference limit of the assay. This corresponds to a probability of obtaining an abnormal value of 0.0044%. Compare this to 1% of the 99th percentile. Originally, decision limits for cardiac troponin were based on matching the original WHO diagnosis. These values were typically 5-15 times the 99th percentile value. This is comparable with the 99th percentile the authors obtained for their ED patients and inpatients.
The use of high sensitivity assays plus the 99th percentile has supported rapid decision-making strategies which can be used for early rule in and ruling out of acute myocardial injury(8) and the ability to distinguish between acute and chronic injury. This has the potential for rapid patient categorisation and discharge. The authors demonstrate that blanket troponin testing is counter-productive due to the large number of other conditions which will cause cardiac damage (all of which have an adverse prognosis). The implication from the study is that they would suggest that an arbitrary much higher value is used for troponin decision-making. The likelihood is if this were to occur that the gains from using troponin in identifying patients missed by previous diagnostic strategies would be lost and more importantly the current problem of under diagnosis of women with cardiac disease exacerbated. Troponin should only be requested when a diagnosis of acute coronary injury and intervention is considered and it must always be remembered that the diagnosis of myocardial infarction is clinical.
Reference List
1. IFCC Committee on Clinical Applications of Cardiac Bio-Markers (C-CB). Contemporary Cardiac Troponin I and T Assay Analytical Characteristics Designated by Manufacturer. IFCC . 2018.
Ref Type: Electronic Citation
2. Collinson PO, Heung YM, Gaze D, Boa F, Senior R, Christenson R, Apple FS. Influence of population selection on the 99th percentile reference value for cardiac troponin assays. Clin Chem 2012;58:219-25.
3. McKie PM, Heublein DM, Scott CG, Gantzer ML, Mehta RA, Rodeheffer RJ et al. Defining high-sensitivity cardiac troponin concentrations in the community. Clin Chem 2013;59:1099-107.
4. de Lemos JA, Drazner MH, Omland T, Ayers CR, Khera A, Rohatgi A et al. Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA 2010;304:2503-12.
5. Blankenberg S, Salomaa V, Makarova N, Ojeda F, Wild P, Lackner KJ et al. Troponin I and cardiovascular risk prediction in the general population: the BiomarCaRE consortium. Eur Heart J 2016.
6. Ostermann M, Lo J, Toolan M, Tuddenham E, Sanderson B, Lei K et al. A prospective study of the impact of serial troponin measurements on the diagnosis of myocardial infarction and hospital and six-month mortality in patients admitted to ICU with non-cardiac diagnoses. Crit Care 2014;18:R62.
7. Ostermann M, Ayis S, Tuddenham E, Lo J, Lei K, Smith J et al. Cardiac Troponin Release is Associated with Biomarkers of Inflammation and Ventricular Dilatation During Critical Illness. Shock 2017;47:702-8.
8. Roffi M, Patrono C, Collet JP, Mueller C, Valgimigli M, Andreotti F et al. 2015 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation: Task Force for the Management of Acute Coronary Syndromes in Patients Presenting without Persistent ST-Segment Elevation of the European Society of Cardiology (ESC). Eur Heart J 2016;37:267-315.
Competing interests: No competing interests