Summary
Cardiac troponin (cTn) assays have quickly gained in analytical sensitivity to become what are termed ‘high-sensitivity cardiac troponin’ (hs-cTn) assays, bringing a flurry of dense yet incomplete literature data. The net result is that cTn assays are not yet standardized and there are still no consensus-built data on how to use and interpret cTn assay results. To address these issues, the authors take cues and clues from multiple disciplines to bring responses to frequently asked questions. In brief, the effective use of hs-cTn hinges on knowing: specific assay characteristics, particularly precision at the 99th percentile of a reference population; factors of variation at the 99th percentile value; and the high-individuality of hs-cTn assays, for which the notion of individual kinetics is more informative than straight reference to ‘normal’ values. The significance of patterns of change between two assay measurements has not yet been documented for every hs-cTn assay. Clinicians need to work hand-in-hand with medical biologists to better understand how to use hs-cTn assays in routine practice.
Résumé
L’évolution rapide des méthodes de dosage des troponines cardiaques (cTn) vers une meilleure sensibilité analytique (cTn de haute sensibilité, ou cTn HS) s’accompagne de nombreuses données de la littérature mais encore incomplètes. En l’absence de standardisation des cTn et de données consensuelles sur l’utilisation et l’interprétation des résultats, les auteurs de cette revue proposent, à partir d’une revue de la littérature, et de façon multidisciplinaire, des éléments de réponses aux questions fréquemment posées. En conclusion, le bon usage des cTn HS repose sur la connaissance : des caractéristiques propres de la méthode utilisée, en particulier de la précision obtenue au 99 e percentile d’une population de référence ; des facteurs de variation de la valeur du 99 e percentile ; de la forte individualité des dosages de cTn HS, pour lesquels la notion de cinétique individuelle est plus informative que la simple référence à des valeurs usuelles. La significativité des variations entre deux dosages, n’est pas encore documentée pour toutes les méthodes HS. La collaboration entre cliniciens et biologistes est nécessaire à une meilleure utilisation des troponines au quotidien.
Background
International guidelines on myocardial infarction (MI) diagnosis recommend running a cardiac troponin (cTn) assay in suspected MI patients unless they present ST-segment elevation (suspected lone-event non-ST-segment elevation MI [NSTEMI]). The need to observe an increase in troponin over the 99th percentile of a reference population together with the significant assay-to-assay variation make it necessary to use what are dubbed sensitive or hypersensitive cTn assays. cTn assays are rapidly gaining in analytical sensitivity. Published data on this latest generation of more sensitive assays are dense, but are still incomplete. Furthermore, cTn assays are not yet standardized and there are still no consensus-built data on how to use and interpret high-sensitivity cTn (hs-cTn) assay results.
Given this context, three French academic societies – the Société française de médecine d’urgence (SFMU) for emergency medicine, the Société française de cardiologie (SFC) for cardiology and the Société française de biologie clinique (SFBC) for clinical biology – have joined forces to co-propose an integrated French-language document that, through a review of the literature, tackles the issue of how to use troponin assays properly. The document adopts a ‘Question and Answer’ format to connect with grass-roots practitioners, and is written to provide clinicians and biologists with the most routine-relevant conclusions possible, including a series of boxes headed ‘In practice/takeaways’, which recap the key messages.
Terms and definitions
What does assay sensitivity mean?
An assay that qualifies as sensitive or hypersensitive (qualifiers arbitrarily grouped under the term ‘high-sensitivity’ in this paper) is an assay that demonstrates greater analytical sensitivity and precision than the conventional method it is built on. The word ‘sensitive’ refers to the assay, not to the biomarker itself.
From an analytical standpoint, analytical sensitivity is the smallest measurable analyte concentration above the limit of detection. Here, sensitivity is determined by the slope of the calibration curve. Higher sensitivity increases the possibility of getting low variations between two assays, as their respective signals will be significantly different ( Fig. 1 ). In other words, a method’s sensitivity is also its ability to precisely and reliably differentiate between two different concentrations.
- •
The word ‘sensitive’ refers to the assay, not to the biomarker itself.
- •
Analytical sensitivity is the smallest difference in concentration measurable by the assay.
What does analytical precision mean?
The analytical precision of an assay is an evaluation of the degree of dispersion in serial test results on a single sample; it is expressed as the analytical coefficient of variation (CV) of the assay, where CV = mean/standard deviation × 100, given as a percentage (%). Analytical precision splits into a repeatability strand (intraseries precision) and a reproducibility strand (interseries precision).
A measurement method’s lower-range limits are defined by: the Limit of Blank (LoB; the concentration below which 95% of results are expected to be found when replicates [ n > 60] of a sample containing no analyte [i.e. a biomarker-free sample] are measured); the Limit of Detection (LoD; the lowest detectable analyte concentration likely to be reliably distinguished from the LoB; it is determined based on the LoB and the standard deviation of replicates of a sample containing a low, but non-zero, biomarker concentration); the Limit of Quantitation (LoQ; the smallest value obtained at a predefined CV). In the specific subfield of cTn, and for conventional assay methods, 10% CV is the analytical limit adopted for MI diagnosis .
Precision, LoD and LoQ vary from cTn assay to cTn assay, so they need to be verified by the laboratories and, if appropriate, communicated to the clinicians, to optimize the results interpretation process. Within the laboratory competency accreditation framework (standard ISO 15189), the analytical characteristics of hs-cTn assays are to be verified as stipulated in document SH-GTA 04 (see ‘What precautions does the laboratory need to take when transitioning to a hs-cTn assay?’).
- •
The analytical precision of an assay is given by the CV at a given analyte concentration.
- •
Analytical characteristics (sensitivity, precision, LoD, LoQ) are assay specific.
What is an hs-cTn assay and are all hs-cTn assays essentially similar?
An hs-cTn assay possesses better sensitivity and better analytical precision than the ‘conventional’ assay it is built on. High-sensitivity assays offer 4-fold to 10-fold greater analytical sensitivity than conventional methods.
Apple and Collinson short-listed two basic criteria for defining whether a cTn assay is ‘highly sensitive’: precision of a reference population; and proportion of measurable concentrations in healthy individuals above the assay’s LoD . For an assay to qualify as’ high-sensitivity’, it has to demonstrate ≤ 10% total imprecision at the 99th percentile value and be able to quantitate at least 50% of healthy individuals .
cTn assays – whether high-sensitivity or conventional – are still not standardized at this time. Standardization efforts are hampered by a combination of factors, chiefly the heterogeneity of the circulating cTn forms that the assays can recognize, post-translational modifications to cTn isoforms and immunoassay response modifications tied to interferences and autoantibodies.
Compounding the issue, results do not directly correlate from technique to technique. In theory, the assays are equimolar, which means they should identically recognize all circulating forms. In practice, however, the distribution of circulating cTn forms in a given patient at a given time can vary, thus producing different responses in different assays. Consequently, the results given by different assays are not directly transposable from test to test.
The net result is that each assay method has its own characteristics and its own cut-off thresholds ( Table 1 ) . Consequently, it is vital that a patient’s cTn concentrations are monitored with the same assay. The high-sensitivity assays add to a long list of conventional cTn assays. In most cases, the high-sensitivity method has replaced the conventional assay (e.g. fourth-generation cTnT [cardiac troponin T] upgraded to hs-cTnT at Roche Diagnostics, ARCHITECT cTnI [cardiac troponin I] upgraded to ARCHITECT hs-cTnI at Abbott, etc.). The analytical modifications integrated in high-sensitivity assays are patent protected and are rarely published.
| Manufacturer | Analyser | LoB | LoD | LoQ a | 99th percentile values | 99th percentile CV |
|---|---|---|---|---|---|---|
| Conventional assays | ||||||
| Abbott | AxSYM | 0.020 μg/L | NA | 0.160 μg/L | 0.040 μg/L | 14% |
| ARCHITECT | < 0.010 μg/L | NA | 0.032 μg/L | 0.028 μg/L | 14% | |
| i-STAT | 0.020 μg/L | NA | 0.100 μg/L | 0.080 μg/L | 17% | |
| Alere | Triage Cardio3 b | 0.002 μg/L | 0.010 μg/L | 0.040 μg/L | 0.020 μg/L | 17% |
| Beckman | Access Accu | 0.010 μg/L | NA | 0.060 μg/L | 0.040 μg/L | 14% |
| bioMérieux | VIDAS Ultra | < 0.010 μg/L | < 0.010 μg/L | 0.110 μg/L | 0.010 μg/L | 28% |
| Radiometer | AQT90 TnT b | NA | 0.008 μg/L | 0.026 μg/L | 0.017 μg/L | 15% |
| AQT90 TnI b | NA | 0.010 μg/L | 0.039 μg/L | 0.023 μg/L | 17% | |
| Response Biomedical | RAMP | 0.030 μg/L | 0.030 μg/L | 0.210 μg/L | 0.100 μg/L | 20% |
| Siemens | Dimension RxL | 0.040 μg/L | NA | 0.140 μg/L | 0.070 μg/L | 15–22% |
| IMMULITE | 0.100 μg/L | NA | 0.220 μg/L | 0.190 μg/L | 11% | |
| Tosoh | AIA II 2G | 0.060 μg/L | NA | NA | 0.060 μg/L | 9% |
| High-sensitivity assays | ||||||
| Abbott | ARCHITECT c | 1.3 ng/L | 1.9 ng/L | 4.7 ng/L | 26.2 ng/L | 4.0% |
| Beckman Coulter | Access c | 2 ng/L | 3 ng/L | 8.6 ng/L | 8.6 ng/L | 10% |
| Mitsubishi | PATHFAST b | 2 ng/L | 8 ng/L | 14 ng/L | 29 ng/L | 5.0% |
| Ortho Clinical Diagnostics | VITROS ECi c | 7 ng/L | 12 ng/L | 34 ng/L | 34 ng/L | 10% |
| Roche Diagnostics | Elecsys/Modular E/cobas e | 3 ng/L | 5 ng/L | 13 ng/L | 14 ng/L | 9% |
| Siemens Healthcare | ADVIA Centaur | 6 ng/L | NA | 30 ng/L | 40 ng/L | 9% |
| Vista | 15 ng/L | NA | 45 ng/L | 45 ng/L | 10% | |
| Vista HS c | 0.5 ng/L | 3 ng/L | 9 ng/L | 5% | ||
| Stratus CS b | 30 ng/L | NA | 60 ng/L | 70 ng/L | 10% | |
| Dimension EXL | 10 ng/L | 17 ng/L | 50 ng/L | 56 ng/L | 10% | |
| Tosoh | AIA 3G | 8 ng/L | 20 ng/L | 35 ng/L | 40 ng/L | NA |
a The smallest concentration of troponin obtained with reliable precision, i.e. with a 10% CV (see ‘What does analytical precision mean?’).
b Point-of-care biochemistry analyser.
The gain in analytical sensitivity from conventional to high-sensitivity assays has translated into a change in unit, where high-sensitivity assay results are expressed in ng/L instead of μg/L – a change that simplifies the reporting of results statements.
- •
An assay qualifies as’ high-sensitivity if it demonstrates ≤ 10% total imprecision at the 99th percentile of a reference population and is able to quantitate at least 50% of healthy individuals.
- •
cTn assays – whether high-sensitivity or conventional – are not standardized.
- •
Not all cTn assays share the same level of precision at low cTn concentrations.
- •
Each cTn assay has its own characteristics and cut-off thresholds, which makes it vital to always use the same assay to monitor the same patient.
- •
The results of hs-cTn assays are expressed in ng/L.
What is an MI?
Defining an MI has always been tricky. The problem is that you have to identify a clinical-anatomical process based on a combination of indirect criteria, as ‘hard’ anatomopathological findings are rarely used in human subjects. For the academic societies, this means that defining an MI involves determining relevant ‘proxy’ criteria and their upper-bound/lower-bound thresholds.
Why does the universal definition of an MI combine diagnostic criteria with a series of infarction types?
Since the early 1960s, there have been regular efforts to establish these criteria, to realign them as the technologies emerge and improve, and to establish some kind of consensus. Up until WHO-MONICA, which served as the benchmark until the late 1990s, an MI was defined based on an electrocardiogram (ECG) and pathological Q-waves. The validation of troponins as a reliable specific biomarker of myocardial injury led first the biochemistry societies then the cardiology societies to refocus the definition onto biomarkers, with troponins leading the way. The European Society of Cardiology (ESC)/American College of Cardiology redefinition of an MI in 2000 marked this turning point.
Initial fuzziness in the redefinition, together with swift and steady progress in the assay techniques, prompted a retweaked ‘redefinition’ of MI a few years later. In 2007, the biochemistry and cardiology societies joined forces to propose what was dubbed a ‘universal definition’ of an MI, where ‘universal’ means that the definition pools all clinical settings qualified as an MI; this has gained very broad consensus. The universal definition gave an even more central role to cTn. The third version – a second update – of the universal definition was published in 2012 .
This fairly radical definition of an MI has had two major consequences: a strong jump in the incidence of clinical settings qualified as infarction ; and, perhaps of even more concern, mounting confusion over where to draw the line between the concept of an MI and other clinical settings associated with myocardial injury. This second concern only emerged with the improvement in assay procedures. Back in 2000, when MI was ‘redefined’, the performance quality of cTn assays put them on a par with creatine kinase assays, and only enabled them to pick up the high concentrations typical of relatively large MIs and certain forms of myocarditis.
To get a sharper picture of MI in light of its redefinition, it was necessary to look beyond the first strand of baseline biochemical, clinical, ECG and imaging characteristics used to define an MI. The definition thus integrated a second ‘clinical classification’ strand, split into five types (numbered 1 to 5).
The definition is only universal if it is taken as a whole – i.e. the criteria plus the clinical types. This makes it crude and, in fact, ultimately wrong to simply say that an individual has had an ‘infarction’ – you need to say that they have had a ‘type- n infarction’, as it is the type that effectively indicates the immediate severity for triage and treatment strategy.
- •
Definitions of MI are based on combinations of consensus-based indirect criteria.
- •
Increase in cTn concentration was first put forward as a pivotal criterion in 2000 and has since taken centre stage.
- •
The improvement in cTn assays achieved by reducing cTn specificity for infarction compelled the definition to introduce the concept of subcategories or ‘types’ of infarction on top of the consensus-built clinical criteria for defining myocardial necrosis.
Which criteria were finally selected for identifying an MI?
The universal definition of an MI works up from the general definition of myocardial injury with cell necrosis: ‘cardiac biomarker values (preferentially troponin) with at least one value above the 99th percentile of the upper reference limit (URL)’.
For this myocardial necrosis to qualify as an infarction, it has to be associated with at least two other conditions. First, a compulsory condition of haemodynamic kinetic evidence that the myocardial injury occurred in a clinical setting of acute myocardial ischaemia: ‘detection of rise and/or fall’. Second, there must be at least one further criterion designed to help confirm that the clinical setting is consistent with recent myocardial ischaemia: clinical symptoms of ischaemia; development of pathological Q-waves in the ECG; new or presumed-new significant ST-segment changes or new left bundle branch block (LBBB); imaging evidence of new loss of viable myocardium or new regional wall motion abnormality; identification of an intracoronary thrombus by angiography (or autopsy).
Note that neither the scale of the rise (or fall) nor its upper-bound/lower-bound cut-offs are detailed.
- •
The definition of infarction is centred on the concept of myocardial necrosis/injury, which is itself defined by a cTn assay value above the 99th percentile of a reference population.
- •
To strengthen the probability that this damage is ischaemic, the definition specifies that myocardial injury has to been combined with conditions based on cTn kinetics (rise and/or fall) plus either clinical, ECG or imaging criteria.
How are the different infarction types defined?
This first strand of the definition is already a source of confusion, as it refers back to a broad and diverse range of clinical infarction settings. To address the issue, the definition goes on to propose five types of MI: two – types 1 and 2 – are clinical; three – types 3, 4 and 5 – are wholly arbitrary.
Clinical definitions: types 1 and 2
Type 1 is defined as spontaneous (sometimes dubbed ‘wild’) MI related to ischaemia caused by a primary coronary event, such as atheromatous plaque erosion and/or rupture, fissuring or dissection, along with fresh intracoronary thrombus.
Type 2 is defined as instances of myocardial injury with cell necrosis where a condition other than plaque rupture contributes to an imbalance between myocardial oxygen supply and/or demand, e.g. anaemia, respiratory failures, arrhythmias, hypotension or hypertension with or without left ventricular hypertrophy. By extension, coronary artery spasm and coronary embolism not tied to plaque rupture are classified under type 2.
It is abundantly clear that the definitions of type 1 and type 2 have nothing to do with ST-segment elevation or non-ST-segment elevation. ST-segment elevation MI (STEMI) is practically always type 1. NSTEMI will split into either type 1 or type 2 depending on clinical features.
Arbitrary definitions: types 3 to 5
These definitions cover settings where the clinical picture points strongly to infarction, but the classical criteria are missing (type 3), and settings where infarction is clearly identified, but applying the classical criteria strictly is irrelevant (types 4a and 5).
Type 3 is defined as sudden unexpected cardiac death with symptoms suggestive of myocardial ischaemia, accompanied by presumably new ST-segment changes or new LBBB, occurring before biomarkers were obtained or before cardiac biomarker values would be increased.
Type 4a is defined as MI related to percutaneous coronary intervention, arbitrarily defined by an elevation of cTn values over 5-fold the 99th percentile URL, and symptoms suggestive of myocardial ischaemia and/or new ischaemic ECG changes or new LBBB and/or angiographical or imaging-demonstrated criteria.
Type 4b is defined as MI associated with stent thrombosis, as documented by coronary angiography or at autopsy in the setting of myocardial ischaemia. The threshold used is the 99th percentile URL, associated with a rise and/or fall of cardiac biomarker values.
Type 5 is defined as MI associated with coronary artery bypass graft, arbitrarily defined by an elevation of cTn values over 10-fold the 99th percentile URL in a patient with normal baseline values, plus symptoms suggestive of myocardial ischaemia and/or new ischaemic ECG changes or new LBBB and/or angiographical or imaging-demonstrated criteria.
The text accompanying the third version of this universal definition spends more time justifying the selected criteria than clarifying them .
- •
The first strand of the definition is already a source of confusion, as it refers back to a broad and diverse range of clinical infarction settings.
- •
The second strand of the definition specifies the criteria for five types of MI, with widely different mechanisms, prognoses and treatment options.
- •
It is the type of infarction that indicates its immediate severity for triage and treatment strategy; MI type must always be stated.
- •
The two types that are most relevant to clinical practice are type 1 (caused by plaque rupture) and type 2 (caused by an imbalance between myocardial oxygen supply and demand); only type 1 MIs are covered under strategies and treatments recommended in the guidelines.
hs-cTn and MI
How are hs-cTn cut-offs determined and what does the 99th percentile of a normal reference population mean in practice?
The cut-off is the value at the 99th percentile of a reference population. This reference population should ideally be representative of the general population.
The 99th percentile value of a cardiac biomarker is the cut-off value below, which 99% of values obtained in the normal reference population fall.
The cut-off is tough to establish, as there is no established consensus on the characteristics of a ‘normal’ population. The demographics of the subjects enrolled in the reference population should be known, and investigations should be completed to confirm the absence of heart disease. Tougher selection criteria tend to result in lower 99th percentile values. The sample has to be large enough, and – ideally – the reference population should be characterized in terms of cardiological health .
For the so-called conventional cTn assays, if imprecision at the 99th percentile is too high (CV > 10%), the lowest concentration with a CV of 10% should be used as the cut-off . For hs-cTn procedures, the cut-off is the 99th percentile as, by definition, the CV at this level is ≤ 10% .
In practice, the 99th percentile URL with its 95% confidence interval is determined from a homogeneous population, following approved guideline procedure, using a non-parametric test . As the proportion of subjects with detectable troponin levels increases, the 99th percentile calculated gains in precision. It takes at least 300 subjects with detectable troponin levels to calculate a 99th percentile to a probability of 95% . European academic society taskforces have singled out this point as one of the critical issues surrounding hs-cTn assays . Age, sex and renal function can all influence the 99th percentile .
It is recommended that the assay reference values of the medical laboratory population are verified (and readapted if appropriate). International medical laboratory accreditation standard ISO 15189 stipulates that the verification of biological analysis reference intervals should be based on at least 100 healthy subjects (which, in this context, means free of chest pain and MI) .
- •
The cut-off value for hs-cTn assays is the 99th percentile.
- •
The 99th percentile value of a cardiac biomarker is the cut-off value below which 99% of values obtained in the population under study fall.
- •
The 99th percentile value can vary for a given assay and a given target reference population according to a series of factors that need to be identified.
- •
Standard ISO 15189 stipulates that the verification of biological analysis reference intervals is to be based on at least 100 healthy subjects.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
