Troponin T assays - a meeting of minds

Jeanette Marchant discusses high sensitivity cardiac troponin assays and discovers that the best way to find a successful working protocol is to get users together and find a consensus

High sensitivity cardiac troponin (cTn hs) assays identify and measure troponin at previously undetectable levels, enabling earlier diagnosis of acute coronary syndrome (ACS). The new generation of assays raises questions about what information should be reported to provide meaningful data for their use in the clinical setting. While the performance of the assays is well documented, there is a perceived need for users to establish their own working protocol to facilitate obtaining the maximum clinical benefit from the assays.   

With a view to arriving at a general consensus on an initial protocol for the practical use of the Roche Elecsys Troponin T high sensitive (cTnT hs) assay, Roche is pioneering a series of meetings for biochemists and clinicians to share their views. Following an initial meeting in Manchester in 20101, a further group gathered in Cardiff to address key issues relating to the clinical use of the assay.

Chaired by Avril Wayte, consultant clinical biochemist at Betsi Cadwaladr University Health Board, the meeting heard from Dr Evangelos Giannitsis, professor of cardiology at University Hospital, Heidelberg, whose chest pain clinic routinely uses cTnT hs to test patients presenting with symptoms of ACS. Additional input was provided by Dr Martin Myers, consultant clinical biochemist at Royal Preston Hospital, and Dr Richard Body, clinical lecturer in cardiovascular medicine at Manchester Royal Infirmary.     

The universal revised definition of myocardial infarction (MI) requires a rise and/or fall of cardiac troponin to be demonstrated in patients with symptoms of ACS with at least one value above the 99th percentile of the upper reference limit (URL) measured with a coefficient of variation (CV) of <10%2. The new generation of TnT and TnI tests are the first troponin assays to meet this level of sensitivity and precision.

The limits of detection have reached very small numbers, as shown in Table 1. It was agreed that in order to simplify reporting results, units should be reported in nanograms per litre (ng/L) rather than pg/mL, ng/mL or µg/L. Changing the units could also provide an opportunity to educate users about the new assay.    

There was general consensus that the 99th percentile value (14ng/L) should be used as the cut-off for the diagnosis of MI. Since the Elecsys TnT hs provides a LoQ (limit of quantification) value of 13ng/L, users of the assay have confidence in using the 14ng/L value for the diagnostic cut-off. (LoQ is the lowest cTnT concentration that can be reproducibly measured with a between-run coefficient of variation of ≤10%.)

Reporting down to lower levels than the 99th percentile value was considered to be of value as evidence is emerging that decisions can be made for patients who have lower cTnT hs levels at presentation in the emergency department. The debate focused on what level should be used as the lowest value for reporting: the limit of blank (LoB) or the limit of detection (LoD).

LoB is the highest apparent analyte concentration expected to be found when replicates of a blank sample containing no analyte are tested. For the Elecsys TnT hs this is 3ng/L. LoD is the lowest analyte concentration likely to be reliably distinguished from the LoB and at which detection is feasible. LoD is determined by utilising both the measured LoB and test replicates of a sample known to contain a low concentration of analyte. The LoD for the Elecsys TnT hs is 5ng/L.

Heidelberg University reports all numerical values of 3ng/L and above to enable the use of delta changes even when a result is below the 99th percentile. Giannitsis stressed the importance of the interpretation of results by experienced clinicians, rather than by junior doctors, in the emergency department: “The biomarker needs the clinical context,” he added.     

While the imprecision at <14ng/L will be greater than 10%, it was agreed that there appears to be a good argument for reporting down to 3ng/L when used in conjunction with an appropriate delta change.

“No one would diagnose a non-STEMI [non-ST segment elevation myocardial infarction] with a result less than 14 [ng/L], irrespective of the deltas, so when you interpret troponin in levels below 14 you can’t use them as individual results, you would only use them in the context of a change,” said Myers.

In view of the potential clinical value in providing information on low levels of cTnT hs, the group agreed to report numerical values ≥3ng/L, with 14ng/L as the diagnostic cut-off for MI.

Gwyon Jenkins, chemistry laboratory manager at Prince Phillip Hospital, Llanelli, raised the issue of quality control material. The group agreed the availability of appropriate quality control material is paramount, and every effort should be made in the preparation of material with a target value close to the 99th percentile.

At least two measurements of cTnT hs are required to diagnose MI according to universal guidelines that require measuring a rise and/or fall in troponin levels. NICE guidelines also recommend two samples. It was acknowledged that financial constraints could deter some NHS hospitals from paying for two tests, although without two measurements the test could not be used to its full potential, and the ability to differentiate between stable coronary heart disease and acute events would be impaired.

Extremely high concentrations of cTnT hs in one sample might be sufficient to indicate an MI, although to date there are no published data that provide a correlation between cTnT hs level and infarct size. It was not possible to arrive at definitive levels on which clinical decisions could be made on the basis of one sample although it was noted that if cTnT hs levels were 5000ng/L in the clinical context of MI, there would be little point in taking a further measurement.

For patients presenting with clinical signs of myocardial ischaemia, the initial sample should be taken at presentation, and the timing of serial measurements should relate to the time of presentation. Current NICE guidelines refer to ‘times post symptoms’, which is regarded as ambiguous as it is not always possible to determine accurately the timing of the onset of symptoms.

The interval between the first measurement of cTnT hs and subsequent samples is a moot point. Various algorithms were discussed by the group and considered in the context of the NHS setting. The universal definition cites 6-9 hours after presentation, while NICE guidelines recommend a 10-12 hour interval after onset of chest pain, although these are due to be updated.  

At Heidelberg University Hospital, protocol is for cTnT hs to be measured at presentation, at three hours post presentation and again at six hours post presentation. A six-hour ‘rule out’ window is used in patients who remain asymptomatic during their stay in the chest pain unit. Although data presented by Giannitsis showed that all patients with an MI were identified within three hours, with 100% sensitivity and 100% negative predictive value, the German protocol was deemed inappropriate in the UK setting where, if more than one sample is taken, two measurements were more likely to be the norm, rather than three or more3.  

The New Zealand protocol developed by Professor Harvey White of Auckland City Hospital was also discussed4. White’s latest version recommends retesting cTnT hs levels six hours post presentation in patients with baseline levels of <14 ng/L to rule out MI. Retesting at three hours post presentation is recommended in patients with higher cTnT hs values, although it was felt that the shorter timeframe would not affect patient management in the UK setting and six hours might be more appropriate.   

Based on available evidence, the group considered that a second sample should be taken six to nine hours post-presentation. However, it accepted that a repeat sample should not be required for clinical symptoms with diagnostic ECG (STEMI) with raised cTnT hs at presentation. Similarly, a result <14 ng/L at presentation should not require a repeat at six to nine hours if a non-cardiac diagnosis is made in the interim.

MI can be ruled out within six hours provided both cTnT hs levels are <14 ng/L (the universal definition of MI states that there should be a rise/fall with at least one result above the 99th percentile), unless clinical indications e.g. further pain or ECG changes, suggest that further investigation is required.

Currently there are no consensus recommendations on the use of delta change values between serial samples. Delta change values of 20% were recommended in the universal definitions recommended by NICE, but Giannitsis suggested that this is too low to be of value in clinical practice because the biovariability of cTnT hs may exceed this level. His proposed protocol which is published in European Cardiology3 is based on a delta change value of 100%, which provides higher specificity but reduced sensitivity. The higher the delta, the higher the loss of sensitivity. “I’m not sure if 100% is the optimal delta but I think it is the best delta for someone who starts with the test,” he said. With increasing experience, and as further data become available, it may be possible to decrease the delta in future.

Using the 100% delta change, patients with possible MI who have ≥ 100% increase in baseline levels of cTnT hs can be positively identified as having acute MI and treated accordingly, according to Giannitsis.

The algorithm proposed by White uses ≥50% delta change to diagnose MI in patients with baseline levels of cTnT hs of ≥14ng/L-<53ng/L. There was concern expressed by the attendees in using such a clear-cut definition of MI when reporting results. Retesting of troponin is recommended in order to detect an acute rise and or fall, suggesting an acute rather than a chronic change. The magnitude of such rise and/or fall has not been defined and several delta change values ranging from 20% to over 100% have been proposed. So far, there is no data to suggest that higher deltas are associated with worse outcomes. As higher delta change values are likely to increase clinical specificity, avoiding unnecessary coronary angiography – at a moderate loss of sensitivity – some centres might prefer to start with a higher delta change value and to decrease this when emerging evidence indicates a more suitable value. According to Giannitsis, Troponin should be retested not later than six hours after the initial blood draw. Protocols with a complementary testing at three hours might be useful for rapid rule-out protocols, but are still under investigation.

In a field where there are so many variables, Myers suggested that the key parameters were the acuteness and significance of the results. While clinical biochemists can provide guidance to cardiologists, they cannot always give definitive answers, he said, and some patients, especially those with increases between 20% and 100% should not be diagnosed on the cTnT hs alone.       

“There is a danger of going too far down the lines of trying to make a diagnosis based on one clinical test,” noted Dr Geraint Jenkins, consultant cardiologist at Abetawe Bro Morgannwg University. Cardiologists want information from the test results that will enable them to rule out a cardiac event: that the patient’s Tn level is negative and in the clinical context it is safe to send the patient home.

In the rapidly evolving field in which new information is emerging on the new generation cTn hs assays, the group concluded it was prudent at this stage to adopt general principles based on the evidence available to date. The general consensus was that if the change is >100% after the second sample, an acute cardiac event is likely and a further sample should be discouraged.

A delta change of <20%, when both results are less than 14 ng/L, rules out an acute cardiac event. In addition, a delta change of <20% with one result of >14 ng/L means that an acute cardiac event is unlikely and a repeat test is not recommended unless clinical suspicion is high.

Delta changes of 20-100% may require further investigation in order to differentiate between acute and chronic events. It is recommended that NICE reviews the data on delta changes, and that users of cTnT hs review their advice on delta changes as more information becomes available.

Biological variation of cTnT hs levels also needs to be taken into account when interpreting serial test results. Concentrations of cTn can for the first time be measured in healthy individuals with the high sensitive assays, enabling the determination of reference change values (RCV). Published data indicate a short-term (hours) RCV of between 50% and 85%.     

Giannitsis highlighted the role of the new generation assay in improving risk stratification. In patients with chronic heart failure, increases in cTnT hs concentrations – albeit still below the 99th percentile value of 14 ng/L – may indicate adverse outcomes. In most other patients, including ACS and non-ACS, absence of a cTnT hs concentration above the 99th percentile value is prognostically favourable. By contrast, outcomes in patients with or without ACS closely correlate to the presence of a cTnT hs value above the 99th percentile value in a concentration-dependent manner: “Every troponin elevation is meaningful,” he said.

Data show the risk of mortality is highest in non-ACS patients with elevated cTnT hs. Patients with elevated levels of cTnT hs should not be regarded as false positive simply because they have no signs of coronary heart disease, cautioned Giannitsis. “You should look for possible differential diagnosis in order to allow timely and specific treatment of these patients, because this group includes end stage renal disease, heart failure, pulmonary embolism, myocarditis and so forth.”  In patients with non-ACS, any cTnT hs elevation will not exceed 100%, according to Giannitsis.

Participants at the meeting arrived at an initial working protocol for the Elecsys TnT-hs assay in clinical practice based on current available information. It was acknowledged that as new evidence emerges, the protocol may need revising. At the same time, Myers emphasised that implementing the new generation assay in clinical practice would require educating and working with cardiologists.

Table 1 Source IFCC website, Analytical characteristics of commercial Tn assays as stated by manufacturer (Oct 2008)