Commandment 1: collaborate with the laboratory and the emergency department

Analytical issues and how emergency department (ED) physicians think are often an anathema to clinicians. Collaboration is essential. Issues to be shared with the laboratory include when and how to evaluate potential false-positive and false-negative values, what cut-off values to use, and how to decide when a changing pattern is present. In regard to the ED, it is important to recognise that ED physicians never want to miss a patient at risk or to do so rarely. This means their practice, if they are unsure, is to admit patients. Since all patients with elevated cTn are at risk, many such patients are admitted. Those with a dynamic pattern require admission; however, many without a dynamic pattern can be evaluated as outpatients, but only if there is agreement concerning a facile pathway for that activity.

These issues will become more important with high sensitivity assays which are more prone to analytical confounds and will detect more subtle elevations not due to coronary artery disease, making follow-up even more important.

Commandment 2: understand some analytical considerations

One needs to know the 99th centile value for the assay in use locally. There is variability in the sensitivity of assays and our journals are not careful about these distinctions. Laboratory analysts are well versed in this area. In addition, the laboratory can be helpful when possible analytical issues arise.

The 99th centile value has been shown in almost every study to maximise diagnostic sensitivity and specificity for finding patients at risk. This is because present day assays do not come close to normal values so all elevations are important. The values for all assays are listed in a table from the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) website (table 1). As assay sensitivity increases, one will unmask larger numbers of elevations due to disease processes other than ischaemia. Thus, cTn elevations are not and should not be considered specific for acute myocardial infarction (AMI). The large number of new elevations observed open new insights into the causes of cardiac injury. Most of these elevations are associated with an adverse prognosis. Elevations with a changing pattern are usually caused by acute disease and thus patients are at risk over a shorter period than those associated with chronic disease which tend not to change notably over a few hours. Acute elevations are best evaluated in hospital. Many of those that are chronic can be evaluated in the outpatient setting.2–4

Analytical false-positive elevations should be suspected when values are elevated and do not fit the clinical situation and/or are elevated and do not change over time. The most common problem is fibrin interference. If suspected, the sample should be re-centrifuged and rerun. Interfering substances, mostly antibodies that can cause false-positives and false-negatives, also exist. Asking the laboratory to treat the sample with additional blocking antibodies (all assays contain some) is the easiest way to unmask this problem. In addition, the failure of samples to dilute linearly can be a hint to the presence of an interfering substance.5

These problems will be more important with high sensitivity cTn (hscTn) assays because when measuring tiny amounts of protein, small changes may be important. Antibodies to cardiac troponin I reduce values but usually do not cause false-negatives. However, with high sensitivity assays, that possibility exists. There also are circulating antibodies to cardiac troponin T (cTnT).6 Haemolysis of even modest extent reduces cTnT and increases cardiac troponin I (cTnI) with some assays7 (figure 1). Thus, with high sensitivity assays, samples will need to be obtained more carefully. Finally, macrokinases, a combination of immunoglobulins and troponin, can occur.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Pronounced effects of haemolysis on cardiac troponin assays, including the high sensitivity cardiac troponin T assay. Circles, squares, and triangles identify samples with varying degrees of cardiac troponin values in the absence of haemolysis. cTNI, cardiac troponin I; cTNT, cardiac troponin T. Reproduced with permission from Bais.7

Commandment 3: make the diagnosis of AMI based on cTn and the clinical scenario

The diagnosis of AMI rests on a changing pattern of cTn values in the proper clinical situation (table 2). The clinical scenario may be classic in some, but many patients—especially women, diabetics, postoperative patients, and the elderly—present atypically. Many patients have supply–demand imbalance and elevations in cTn for that reason rather than plaque rupture, a so called type 2 AMI.2 The diagnosis of AMI should be reserved for when there is a component due to a coronary abnormality, be it plaque disruption, fixed coronary artery disease or endothelial dysfunction. The problem is that it is often hard to tell even with angiography. Often a clear ‘culprit lesion’ is not present. A synthesis of the clinical data may be the best that can be done. Designating patients with heart failure or tachycardia/hypotension and normal hearts and the like as AMIs is not recommended. Those patients should be indicated as having cardiac injury, but not AMI.

EDs are under pressure to evaluate patients expeditiously. Therefore biomarkers are obtained on a large number of patients at the start of the evaluation so they will be available as needed. This leads to values being obtained in large numbers of patients at low risk, especially in the USA where the frequency of acute coronary syndrome (ACS—unstable angina and non-ST elevation AMI) is below 10%. This changes interpretation of cTn values. It important to make a pretest assessment of the probability of ischaemic heart disease before one obtains the cTn values, and then integrate the values into the evaluation8 (figure 2). If the pretest probability for ischaemia is high and cTn is elevated, the diagnosis of AMI is made. If the clinical situation is not suggestive of ischaemia, alternative diagnoses need to be considered. Some other aetiologies for cTn elevations such as sepsis, pulmonary embolism, acute respiratory failure and others can cause a rising pattern.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Baysian approach to use of high sensitivity cardiac troponin T (hsTNT) for the diagnosis of acute coronary syndrome. cTNT, cardiac troponin T. Reproduced with permission from Diamond and Kaul.8

The tension with hscTn is that the more sensitive the assay, the lower the specificity of any given minor elevation for ischaemic injury. That does not mean that more sensitive assays do not detect more ACS patients; it simply reduces the proportion. The ability of these assays to detect small amounts of cardiac damage will detect more patients with ACS and do it more rapidly. However, because of increased sensitivity, one will also detect many elevations related to other disease processes, some acute and some chronic. Myocarditis now appears to be the most common mimicker of AMI9 and the most common cause of an elevated cTn in patients with ACS and normal coronary arteries (table 3). Values can be increasing or can be stable. A subset of patients, particularly women, can have ACS or AMI without a clear culprit lesion.

With present day assays, most patients (80%) rule in for AMI within 2–3 h of onset of the symptoms.10 This trend will improve further with higher sensitivity assays which may allow a rule in diagnosis in large numbers at presentation. This benefit requires the use of the 99th centile value and appropriate assays, and not some of the insensitive assays still on the market. With this approach, there is no need for the use of other so called ‘rapidly increasing markers’.11

Implementing use of a changing pattern is challenging. If patients present late, one may reach peak values and values may not be increasing. Using a changing pattern is valuable mostly when cTn values are low. There are very few causes of pronounced elevations. The four that are known are an occasional analytical problem, an occasional patient with renal failure, myocarditis, and AMI. The level that is substantially elevated is assay dependent and requires local determination and, if exceeded, should lead to immediate admission. Lower levels are less frequently due to ischaemic heart disease. For example, cTnT is elevated in 0.7% of subjects, mostly because of underlying ischaemic heart disease, heart failure and/or renal failure. The maximum value published is 0.29 ng/ml.12 Therefore, values above that imply something more acute is present. Utilising a delta is helpful with these lower values. With most acute ischaemic presentations, the delta is substantial in size but the minimum delta is unclear. If there is a change of 50% or 60% with most assays (not all), an acute event is likely.13 If the delta is more modest, one has to rely on clinical judgement. One way to provide guidance eventually will be to define biological and analytical variation together and argue that a significant delta should exceed that value. Another approach that is necessary with contemporary assays that cannot measure normal subjects and thus cannot determine biological variation is to use analytical variation. When troponin values are elevated, assay variability is such that one can rely on a 20% change as significant.3 4 However, for low values, a larger change is necessary. Minor differences should not necessarily be considered significant. Importantly, the presence of coronary artery disease does not mean that a given cTn elevation is indicative of unstable disease, especially with hscTn.

With high sensitivity troponin assays, this issue will be more challenging. Sensitive assays will be sensitive to clinical changes and even minor analytic problems. For example, >50% of patients admitted for pacemaker placement have elevated hscTnT values, and after implantation >80% had elevations. Thus, 30% had a rising pattern. This is one reason why some have argued for delaying implementation of these assays until better data are developed.14

Commandment 4: rule out myocardial infarction differently than ruling it in

If one rules in the vast majority of patients early, one should be able to send the rest home rapidly depending on the miss rate one is willing to tolerate. EDs would like an event rate of <1% in. Given this, a different evaluation to define those cTn values below which one is very unlikely to have an elevation subsequently or an event is necessary. Present guidelines suggest patients are monitored for 6 h.2 3

Higher sensitivity assays are apt to help. Eventually, it may be that up to 50% of patients can be ruled out on the initial sample; however, the exact metrics for this are not clear at present. What is key is the duration from the onset of symptoms suggestive of acute ischaemia to the evaluation. If by 6 h after the onset of symptoms hscTn is not elevated, one is unlikely to have an AMI and/or be at risk for short term adverse outcomes. However, it is not always easy to determine the onset of symptoms. If one is unsure, one should use the time of presentation as the zero time. Because of this need, other rule out markers such as copeptin deserve study. However, it is likely that with high sensitivity assays, the use of such markers will be short lived.

Often after exclusion of AMI with troponin, clinicians feel obligated to do stress tests to evaluate the possibility of coronary disease. Studies validating stress testing were done before hscTn assays. Even with present assays, and particularly with ultrasensitive assays, it is not clear whether stress testing provides incremental value. Novel high sensitivity assays have been reported to manifest a change in values if stress tests are positive.

Commandment 5: use common sense to interpret elevations of cTn in patients who are critically ill

Critically ill patients with cardiac involvement are at higher risk than others. Large numbers of investigations have documented elevations in 30–50% of patients. The mechanism depends on the clinical circumstances. Patients who have acute respiratory failure have a 30-fold increase in mortality when cTn is elevated.15 In the absence of an elevated cTn the prognosis in these patients, despite other respiratory markers of severity, is excellent (figure 3). This increased risk is not specifically linked to coronary artery disease, although smokers who make up much of this group might have a high incidence of coronary heart disease. In patients who are septic, most elevations are thought to be due toxic cytokines such as tumour necrosis factor (TNF) and heat shock proteins.16 They presage an adverse short and long term prognosis. In patients with gastrointestinal bleeding on the other hand, elevations appear related to hypotension and/or tachycardia and are only prognostic long term.17 One needs to think individually about patients and each specific acute disease state. Some of these patients could have AMI, but how to screen for that group is unclear. Even if AMI is present, it is thought that the majority of these patients probably have coronary artery disease and/or endothelial dysfunction that is fixed, and then have severe supply–demand abnormalities as one might expect with hypotension, tachycardia and/or the use of catecholamines. This group may not need anticoagulation and/or an early invasive approach. It is not clear whether potentially cardiotoxic but also potentially life saving agents should be discontinued, but moderating their use may be prudent. Thus, sensible suggestions about management may frequently be all that can be done. Once these patients are discharged, their prognosis is still substantially worse. At that time, a search for cardiovascular risk factors such as hypertension, hyperlipidaemia, valvular heart disease and the like would be advised.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Short and long term survival in patients with acute respiratory failure by cardiac troponin T (cTNT) value. Reproduced with permission from Vasile et al.15

These same considerations are better documented in the postoperative circumstance where even minor elevations of cTn are prognostic both short and long term. There is better documentation in this setting that the majority of elevations are attributable to the supply–demand imbalance which should be aggressively sought and corrected. These patients also deserve outpatient scrutiny subsequently in the hope of improving long term prognosis.18

This is an area that is likely to explode. We are just beginning to interrogate situations where concomitant cardiac injury may be more apt to occur such as in elderly patients undergoing surgery. In addition, with high sensitivity assays, larger numbers of elevations will be found. It may be that some of the elevations may be so minor that differences between the ‘normals’ and the ‘abnormals’ are slight and it will be difficult to show prognostic and/or therapeutic benefits from intervention. This is not the case at present.

Commandment 6: do not be intimidated by elevations in patients with renal failure19

Elevations in patients with end stage renal disease are common, but lesser degrees of renal dysfunction do not cause elevations. There are more elevations of cTnT than cTnI. The best explanation for this is that there are different cleavage products due to renal failure which inhibit cTnI assays but not the cTnT assay. cTn is not cleared renally. Elevations identify high risk patients. In an analysis of over 700 patients awaiting renal transplant, the prognostic significance of troponin overwhelmed all other predictors (table 4). Although there is a relationship between elevations of cTnT and coronary artery disease, mortality could not be linked to coronary heart disease. The presence of left ventricular hypertrophy and elevated cTn is a major risk factor for mortality.

Elevations associated with renal failure increase slowly over time as the duration of dialysis increases. However, in the short term, they are stable and usually modest. Acute events can be diagnosed when values have a rising and/or falling pattern. It is prudent to obtain baseline levels in patients with end stage renal disease for risk stratification, so that one can observe changes more objectively when patients present acutely.

Commandment 7: take the baseline value of cTn into account with percutaneous coronary intervention

It has been difficult to understand why modest elevations after percutaneous coronary intervention (PCI) are associated with an adverse prognosis. It is now clear that individuals at risk for mortality and/or recurrent AMI are not those who develop de novo elevations but those with elevations at baseline. This group includes patients with stable coronary artery disease. It appears that elevations mark an adverse anatomic substrate. Consequently, these patients do worse. Initially, most studies were done utilising insensitive assays or high cut-off values. Thus, large numbers of patients had elevated baseline cTn values that were not accounted for. Once this effect was unmasked, it became clear that the patients who developed substantial cTn elevations post-PCI were those with baseline elevations. Elevations from a normal baseline are less frequent and, with rare exception, usually modest. Commonly, minor elevations pre-PCI are accentuated post-PCI, and if high cut-off values are used, it is only post-PCI when they are identified.

If one utilises the pre-PCI elevation values when one does prognostic evaluation, the prognostic importance of post-PCI cTn and/or creatine kinase MB (CK-MB) values is totally obviated.19 However, the baseline value is substantially prognostic. It is for this reason that recent guidelines indicate that, in the absence of a normal baseline cTn, one cannot diagnose post-PCI injury because it is impossible to discern whether the elevation is related to the initial injury or the PCI.1 The one exception is if there is an elevated baseline cTn that stays constant or goes down over time. In that circumstance the criteria for reinfarction of a 20% increase is suggested. The most interesting dataset validating this approach was recently published (figure 4). Baseline elevations were highly prognostic. Elevations arising from a normal baseline had borderline statistical significance, largely because of the inclusion of patients who had non-cardiac events post-PCI.20

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Predicted effects of cardiac troponin T values pre- and post-percutaneous coronary intervention (PCI). Reproduced with permission from Prasad et al.20

Consequently, post-PCI injury occurs, it can be diagnosed by elevations in cTn if the baseline is normal, but elevations have little if any prognostic significance for cardiac events. If the baseline is elevated, one should not attempt to make the diagnosis of post-PCI injury.

High sensitivity assays are likely to strengthen these findings.

Commandment 8: it takes multiple parameters to make the diagnosis of AMI following bypass surgery2

Cardiac surgery damages myocardium and elevations in troponin are ubiquitous. The extent of the injury depends on the details of the cardioplegia, the anaesthesia, the complexity of the procedures, the duration of the crossclamp time, etc. Thus, substantial variability based on these approaches occurs. Given this heterogeneity of cTn values after surgery, the idea of finding one cut-off value that would suffice for the diagnosis of AMI of all cardiac surgeries was not possible. Therefore, the guidelines suggest a fivefold increase to suspect the diagnosis and, if present, ancillary criteria to confirm it. This low cut-off value permits the diagnosis even with less invasive off pump approaches, but will invariably be met with more extensive surgeries. One could characterise biomarker release associated with each type of procedure, anaesthesia and cardioprotective regimen and to develop metrics for AMI when the degree of damage is outside of what should be expected.

There is a reasonable correlation between the magnitude of the cTn elevations and MRI determined findings. Most MRI findings are apical, subendocardial and thus reflect the details of the cardiac procedure and not coronary problems. Very high values are associated with transmural defects and are more associated with graft or native vessel occlusion, but the overlap is large. High sensitivity assays will not change these issues since we are already dealing with high levels.

Commandment 9: do not forget drug toxicities as an aetiology for cTn elevations

We now can detect important cardiovascular drug toxicities. Cardinale et al documented increases in cTn in response to Adriamycin and elevations presaged subsequent reductions in ventricular function. They then showed that angiotensin receptors blockers reduced increases in cTn and obviated reductions in left ventricular function.21 Similar data are now available for trastuzumab.22 High sensitivity assays will facilitate these findings with other agents. In addition, toxicities such as carbon monoxide poisoning23 cause elevations in cTn which have prognostic importance. cTn elevations may provide a hint to that diagnosis.

Pharmaceutical companies are already using novel high sensitivity research assays to evaluate new potentially cardiotoxic compounds. Such studies will likely unmask additional elevations with potentially important pathophysiological implications.

Commandment 10: be cautious with cTn elevations post-exercise

Elevations in cTn occur transiently after extreme exercise. These findings led to arguments over whether the release was due to reversible or irreversible injury. There still are no answers. However, if cTn elevations are modest and disappear within 24 h, they are likely to be benign short term. The longer term prognosis is unclear. One argument that is clinically relevant and may be influenced by high sensitivity assays is that usually elevations persist for only 24 or 36 h. The lack of prolonged elevations has been used to suggest that the structural pool of cTn is not degraded and thus irreversible injury is not occurring. Data with high sensitivity assays seem to suggest that some individuals do have persistent elevations which could change thinking in this area.24