High-sensitivity troponin measurements: challenges and opportunities for the laboratory and the clinician

There has been a paradigm shift in the role of the laboratory in both acute and chronic management of cardiovascular disease. Cardiac biomarkers are now central to the diagnosis and management of acute cardiac disease. Measurement of the cardiac troponins, cardiac troponin T (cTnT) and cardiac troponin I (cTnI) became incorporated within the redefinition ¹ and then definition of acute myocardial infarction ² (AMI) and was recognized by the European Society of Cardiology (ESC) as central to the management pathway of suspected acute non-ST elevation myocardial infarction (NSTEMI). ³

The shift to the use of cTnT and cTnI measurement was based on the early demonstration that in patients with a final diagnosis excluding AMI and a presumptive diagnosis of unstable angina, approximately one-third had detectable troponin values, and this detectable troponin was associated with an adverse prognosis.^4,5 Measurement of cardiac troponin was demonstrably superior, both diagnostically and prognostically, but was also considerably (typically 10–50 times) more expensive than measurement of ‘cardiac enzymes’. In order to justify the cost, troponin measurement was ‘sold’ to the cardiology and emergency medicine physicians as the ‘golden ticket’. It was totally cardiac specific, and allowed a diagnosis of AMI to be reliably made and needed only be measured once 10–12 h from onset of chest pain/admission. Troponin measurement was embraced with enthusiasm, and some would say clinical judgement was abandoned. At this time, troponin assays were relatively insensitive, troponin could not be detected in the normal healthy individual, and decision limits were optimized for equivalence with diagnosis based on the original WHO criteria for AMI. Hence, the claims for troponin could, at that point, be justified and the troponin era began. There were arguments about relative differences between assays, although the main argument was about troponin elevation in renal failure. It was an inconvenient truth that both cTnT and cTnI were elevated in renal failure, even with the assays then in use. ⁶ It was also overlooked that cardiac death, especially cardiac sudden death, is the commonest cause of death in patients with renal failure. ⁷

The original assays have been progressively developed, redeveloped and reformulated to produce the current generation of sensitive and now high-sensitivity (hs) assays. An hs assay is not measuring a new form of troponin. It has been pragmatically defined as one where the 10% CV is below the 99th percentile of a reference population, and which has the ability to measure troponin above the limit of detection of the assay in at least 50% of that population. ⁸ The impetus for improved assay performance has in part come from the manufacturer’s desire to improve assay performance, but has been driven by the redefinition of MI itself. The original redefinition document specified the 99th percentile as decision level and recommended that this should be measured with a coefficient of variation (CV) ≤10%. ¹ It was only the universal definition that explicitly specified a rise or fall of troponin (a delta change). ² Progressive improvement in assay sensitivity has resulted in an increasing range of clinical conditions where troponin elevation is detected in patients who do not have AMI. This led to the observation by Bob Jesse, an Emergency Medicine physician that ‘when troponin was lousy assay it was a great test, but now that it’s a great assay it’s a lousy test’. ⁹

However, it must be remembered that it has been consistently shown that troponin elevation, whatever the cause, is prognostic. In acute situations with a defined pathology, such as pulmonary embolism, patients with an elevated troponin have a significantly worse prognosis. ¹⁰ Similarly, in the acutely ill in intensive care, troponin elevation is similarly associated with a worse outcome.^11,12 This is also true for chronic disease; in patients with heart failure^13,14 or renal impairment,^15,16 troponin elevation is associated with a worse long-term prognosis. In population studies, troponin elevation above the 99th percentile is associated with subclinical disease^17,18 and with a higher long-term risk.^19–21 There is a perception that the shift to an hs assay results in many more patients with troponin elevation which lacks clinical significance. It has been shown that the lowering of the diagnostic threshold identifies patients at significant risk who will benefit from treatment^22,23 and that the number of patients diagnosed as having myocardial infarction (MI) does not dramatically increase. ²⁴ High-sensitivity troponin measurement improves the detection of AMI in women. ²⁵ In addition, it is frequently overlooked that the majority of people who present with chest pain have a troponin within the reference interval. A troponin within the 99th percentile indicates a good prognosis; a troponin above the 99th percentile indicates that the patient has underlying disease. This may not necessarily be AMI but still needs appropriate assessment.

Strategies for the use of sensitive and hs troponin assays for the differential diagnosis of patients with chest pain

These can be considered into three categories: strategies based on the 99th percentile, strategies utilizing improved assay imprecision and strategies based on novel decision thresholds.

Conventional 99th percentile strategies

This is the concept that most clinicians and laboratories are familiar with and is incorporated in the redefinition and universal definition of MI. The original NICE recommendation of a single measurement 10–12 h from admission/onset of chest pain reflected financial and assay performance imperatives. Subsequently, guidelines recommended measurement on admission and 6–9 h. The initial publications using sensitive assays showed excellent diagnostic accuracy on admission (area under the receiver operator characteristic curve of 0.96) with performance apparently independent of sampling time.^27,28 There are some caveats here, however. Diagnosis was not compared with an hs assay as reference standard or even a conventional assay at the 99th percentile achievable at that time and the sample set included patients presenting with ST segment elevation MI (where biochemistry is not part of early management). The use of an admission and 3-h sample was initially recommended by the ESC ²⁹ and subsequently, following an evidence and cost-effectiveness review, by NICE. ³⁰ Again there are caveats. In a well-executed study with hs troponin as part of the reference diagnosis, it was shown in a large prospective observational series that 100% sensitivity for a final diagnosis of MI was only achieved at 8 h from symptom onset (typically 6 h from admission). ³¹ The NICE review is also not concerned solely with diagnosed accuracy but also takes into account cost-effectiveness. So while a 3-h diagnostic strategy should be applied, it should be used as part of appropriate clinical assessment. Further testing or admission may be appropriate in some patients irrespective of a ‘normal’ 3-h troponin value.

Accelerated 99th percentile strategies

These have been actively investigated by the Australasian Emergency Medicine community. They have shown that the combination of formal selection of a low-risk group by structured clinical assessment or use of a risk scoring tool can be combined with troponin measurement on admission and at 2 h for rapid exclusion of MI.^32–36 Here, the emphasis is not only on a diagnosis of MI, but also identification of a population at low risk of future cardiac events that can be safely discharged and will not re-present to the hospital within 30 days. The strategy therefore does not guarantee 100% detection of all AMI patients but does identify those who re-attend. When this approach is used but with an hs troponin assay, rather than a conventional sensitive assay, the cohort could be extended into a higher risk population with equal diagnostic efficiency. ³⁷ They have further extended their studies by using 2-h delta troponin and shown that the combination of a 2-h delta plus the 99th percentile yields optimal diagnostic performance in this population. ³⁸

Strategies utilizing improved assay imprecision

These are based on the ability to confirm or exclude acute myocardial injury combining the use of a delta troponin with a decision threshold, usually near the 99th percentile. Patients are split into three categories based on troponin measurement on admission and 1 h later combined with the calculation of a 1-h delta value. Those with a troponin less than the 99th percentile or decision limit and no delta change are considered ruled out for AMI. Those with a value above the limit plus a delta change rule in for AMI and those who do not exceed the limit but have a significant delta change or exceed the limit without a delta change are in an intermediate category requiring further investigation. The initial study used troponin T with a decision limit of 12 ng/L ³⁹ and a delta of 3 ng/L and claimed a sensitivity and negative predictive value of 100%. A subsequent multicentre study has confirmed these findings although with a slightly lower sensitivity, ⁴⁰ and there has been a similar study of hs cTnI. ⁴¹ The findings have potential for widespread application and are also included in current ESC ⁴² guidelines. Although the focus is on AMI, in reality, the algorithm rules in and rules out acute myocardial injury rather than AMI and further validation is required.

Novel decision thresholds for cardiac troponin

The strategies described above are conventional and based on the universal definition and aimed at diagnosis. Novel decision thresholds are based on the ability of a troponin level to predict risk of a cardiac event during short-term follow-up (typically 30 days), hence are prognostic and not diagnostic strategies. The objective has been to define a low value of troponin, usually at or about the detection limit of the assay, which allows rule out of clinically significant myocardial injury on first presentation to the hospital, hence measurement of a single sample on first admission. This approach has been shown to be effective for both hs cTnT^43,44, with a sensitivity of 99.8% ⁴³ and 99.9% for cTnI, ⁴⁵ but it must be remembered that such a strategy is predicated on an acceptable rate of misdiagnosis and does not achieve 100% sensitivity. In addition, it may be less effective on patients presenting very early (less than 2 h from onset of infarction) and must always be combined with risk stratification and clinical judgement.

hs assays for chronic disease management

The ability of hs troponin measurement to provide prognostic information in both the apparently healthy and those with chronic disease is well documented, and, as discussed above, underpins the use of hs troponin in decision making in acute chest pain. The question then arises can hs troponin measurement be used to monitor treatment effectiveness. Current risk assessment tools for primary prevention of cardiovascular disease have a relatively poor performance, and it has been shown that addition of hs cTn measurement improves risk stratification.^46,47 Analysis of a subset of samples from a randomized controlled trial of pravastatin showed that troponin values at baseline (measured by an hs cTnI method) predicted risk of a cardiac event during follow-up (Mills, presented ESC annual meeting 2014).⁴⁸ In addition, in those randomized to treatment, where cTnI fell, risk was reduced compared to those where cTnI remained elevated. In patients with renal disease, a different dialysis regime reduced troponin elevation. ⁴⁹ These are observational studies and intervention trials are required, as observation or risk prediction may not translate into an effective intervention, ⁵⁰ but the findings are intriguing.

In conclusion, the use of hs troponin measurements offers significant possibilities to improve patient flow in patients presenting with chest pain or suspected AMI. An example of a diagnostic algorithm based on current evidence is shown in Figure 1.

Figure 1.

Diagnostic algorithm for rapid management of chest pain using high-sensitivity troponin measurements.