Peak Troponin Level and Revascularisation Rate in Type 2 MI

Type 2 MI (T2MI) is a well-defined entity, but it lacks evidence to support clinical decision-making surrounding management, particularly the role of invasive angiography revascularisation. Troponin T level has been correlated with mortality rate in T2MI and is anecdotally used as a parameter to identify significant coronary lesions that may benefit from revascularisation with either percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG).

Therefore, we sought to explore the association between peak troponin level in the first 24 hours of presentation and the rates of revascularisation in patients presenting with type 1 MI (T1MI) and non-T1MI, with a view that further insights into this relationship may guide selection of patients who may benefit from referral for coronary angiography with a view to revascularisation.

The Fourth Universal Definition of MI provides a mechanistic approach to the diagnostic classification of myocardial injury, which is defined as a troponin level above the 99th percentile.¹ A rise and/or fall in troponin concentration indicates a dynamic pathophysiological mechanism, while elevated troponin concentration without this dynamic change implies a more chronic pathological process, that is, chronic myocardial injury. Acute events encompass T1MI, T2MI and acute myocardial injury. Acute myocardial injury differs from acute MI by the absence of cardiac ischaemia, as evidenced by a lack of clinical symptoms consistent with an acute coronary syndrome, ECG changes indicating coronary occlusion, or cardiac imaging changes suggesting coronary territory infarction or ischaemia. T1MI results from plaque rupture or erosion causing acute ischaemia; this is in contrast to T2MI and acute myocardial injury, which result from either mismatched myocardial oxygen supply and demand, or direct myocardial inflammation. While types of myocardial injury have been clearly defined, only the management of T1MI has been standardised.^2,3

Based on an understanding of its acute pathophysiology, the management strategies for T1MI are embedded in international guidelines with an emphasis on providing coronary angiography with a view to revascularisation to attain the strongly evidenced reductions in recurrent ischaemic events.^3,4 Notably, troponin T concentration in T1MI has an established predictive and prognostic value with regard to the likelihood of receiving PCI or CABG, and mortality, respectively, justifying its utility in the decision-making for an early invasive approach.^5,6 There are no guidelines for the approach to non-T1MI despite its known association with higher mortality rates.^7–9

Clinical insights that might inform the differentiation of patients who may or may not benefit from coronary revascularisation in the many presenting with non-T1MI remain challenging to ascertain, and in the absence of strong evidence, clinical decision-making is often guided by available signals including peak troponin level. Extrapolating the relationship between troponin level and the likelihood of requiring revascularisation in T1MI to those presenting with T2MI seems intuitive, however, this approach remains understudied.

Method

Study Design

This study is a subanalysis of a statewide study based in South Australia, Australia, exploring the impact of high-sensitivity cardiac troponin T (hs-cTnT) reporting on clinical outcomes up to 12 months. Data were captured from February 2020 to February 2021, 6 months before and 6 months after the hs-cTnT reporting change. Also note that troponin T is used in a statewide system in South Australia. This subanalysis explores the relationship between peak cardiac TnT level in the first 24 hours of presentation (‘peak cTnT’) and the rate of revascularisation.

Study Population

All patients aged 18 years or older with any degree of myocardial injury, defined as a cTnT level above the upper limit of normal (>14 ng/l) in the first 24 hours of emergency department (ED) presentation, were included in the study. All cTnT concentrations were extracted from the electronic records for each patient. Patients were excluded from analysis if they were under the age of 18, had undergone coronary angiography in the 6 months prior to their presentation, or had a prior ED cTnT assessment in the 6 months prior to presentation.

Study Protocol

Within the primary study, two validated machine learning-based algorithms trained to classify clinical presentations according to the Fourth Universal Definition of MI using the same assay and patients sampled from a prior ED population, were used to identify each episode as T1MI, T2MI or acute myocardial injury or chronic myocardial injury.^1,10 Discrimination of the type of myocardial injury was informed by troponin dynamics, systemic biochemical derangement, ECG information, and past history. Details of the algorithms’ training and validation have been published elsewhere, and the diagnostic performance for the discrimination of myocardial injury type, using area under the curve (AUC), was 0.96 for XGBoost and 0.95 for Deep Learning.¹⁰ Given their similar incidence and clinical outcome profile, patients with T2MI and those with acute myocardial injury were combined into one group in this analysis, denoted as ‘T2MI/acute myocardial injury’. The specific systemic aetiologies leading to T2MI/acute myocardial injury were not recorded, given that these are subjectively defined and lack verifiable diagnosis. Additionally, a comparison of systemic parameters of patients presenting with T1MI and T2MI/acute myocardial injury (Table 1) shows significant physiological differences between the groups, irrespective of underlying aetiology. This significant objective difference in physiological derangement further undermines the subjectively classified aetiologies of T2MI/acute myocardial injury as a measurable parameter.

Data Collection and Outcome Measures

This analysis focused on the relationship between peak cTnT concentration and the subsequent likelihood of revascularisation. Revascularisation methods encompassed PCI and/or CABG.

Data on all-cause mortality, new or recurrent MI (new cTnT elevation >50% after the first 48 hours from index presentation, confirmed by ICD10-AM clinical codes I21–I25), over the subsequent 12 months were also collected.

Statistical Analysis

Baseline characteristics, provision of coronary angiography and revascularisation were stratified by type of myocardial injury. The chi-squared test was used for analysis of dichotomous outcomes between groups, while the Kruskal–Wallis test was used for continuous outcomes. The association between the type of myocardial injury and death or recurrent MI was adjusted for age, sex, renal function and peak cTnT using a Cox regression model with the reference population defined as patients who presented with T1MI. Given that coronary revascularisation can occur only in patients who have undergone coronary angiography, an analysis confined to this subgroup was undertaken. Rates of coronary angiography and revascularisation were stratified by numeric intervals of peak cTnT in the combined cohort and for each type of myocardial injury. The intervals of peak cTnT were defined as: 100–499; 500–999; 1,000–4,999; and =5,000. A Cochrane–Armitage test of trend was used to investigate the association between peak cTnT and the rate of revascularisation in patients who underwent coronary angiography. Locally weighted scatterplot smoothing (LOWESS) was used to present the relationship between peak cTnT and revascularisation in patients with T1MI and T2MI/acute myocardial injury who underwent coronary angiography. All statistical analysis was conducted using Stata 18.0, and p<0.05 was considered statistically significant.

Results

Patient Population

A total of 14,425 patients matching the study inclusion criteria were included: 1,011 (7%) were categorised as having T1MI, 1,819 (13%) were categorised as having T2MI/acute myocardial injury, and 11,595 (80%) were categorised as having chronic myocardial injury. The majority of patients were older than 60 years of age. There was a female predominance in the cohort, with 725 women (71.7%) in the T1MI group, 894 (49.1%) in the T2MI/acute myocardial injury group and 6,438 (55.5%) in the chronic myocardial injury group.

Table 2 lists the baseline characteristics, including traditional cardiovascular risk factors, for the study population. Patients presenting with non-T1MI were older and had more comorbidities than those presenting with T1MI: those presenting with T1MI had a median age of 66 years, compared with those in the T2MI/acute myocardial injury and chronic myocardial injury groups who had median ages of 78 years and 79 years, respectively; and those with non-T1MI were more likely to have hypertension, heart failure and diabetes.

Mortality and Event Recurrence

Overall, the 12-month mortality rate was significantly higher in those who presented with T2MI/acute myocardial injury (28.75%; HR 1.87; 95% CI [1.43–2.43]; p<0.001) and chronic myocardial injury (17.02%; HR 3.51; 95% CI [2.68–4.61]; p<0.001), than in those who presented with T1MI (6.63%; Table 3). Increasing peak cTnT concentration was associated with a higher mortality rate in non-T1MI (HR 1.01 per 100 ng/l cTnT; 95% CI [1.01–1.02]; p<0.001). Rates of subsequent T1MI overall were lower in those presenting with non-T1MI, but this was not associated with peak cTnT concentration (HR 1.00 per 100 ng/l cTnT; 95% CI [1.00–1.01]; p=0.249).

Rates of Coronary Angiography and Revascularisation

A substantially greater proportion of patients underwent coronary angiography in those presenting with T1MI (80.8%) than in those presenting with T2MI/acute myocardial injury (6.4%) and chronic myocardial injury (2.0%). There was a lower mortality rate for all patients who underwent coronary angiography versus conservative management, with 43 (5.3%) versus 24 (12.4%) for T1M1; 16 (13.8%) versus 507 (29.8%) for T2MI/acute myocardial injury; and 11 (4.7%) versus 1,963 (17.3%) for chronic myocardial injury. Notably in those who presented with T2MI/acute myocardial injury, those who underwent angiography had a lower 12-month mortality than those who did not undergo angiography (HR 0.43; 95% CI [0.26–0.71]; p=0.001) despite a higher rate of subsequent T1MI at 12 months (HR 5.08; 95% CI [2.59–9.96]; p<0.001). A similar pattern was observed in those presenting with chronic myocardial injury (Table 4). No significant difference in subsequent MI was noted in patients presenting T1MI who did or did not undergo coronary angiography. Of those who underwent invasive coronary angiography, 54.4%, 43.8% and 35% of patients presenting with T1MI, T2MI/acute myocardial injury and chronic myocardial injury, respectively, were subsequently revascularised. Increasing peak cTnT was modestly associated with the likelihood of revascularisation in patients presenting with T1MI (OR 1.022 per 100 ng/l cTnT; 95% CI [1.008–1.037]; p=0.002; Figure 1). This association was not seen in those presenting with T2MI/acute myocardial injury across the same range of peak troponin levels who had coronary angiography (OR 1.031 per 100 ng/l; 95% CI [0.948–1.121]; p=0.472; Figure 2).

Discussion

On analysis of this dataset there are three key observations: peak cTnT correlates with increased mortality in non-T1MI; peak cTnT did not correlate with rate of revascularisation in T2MI/acute myocardial injury; and rate of angiography was not associated with lower event recurrence in non-T1MI, particularly in those presenting with T2MI/acute myocardial injury. In keeping with the existing data, we have shown that peak cTnT correlates with increased mortality in T2MI.⁸ Mortality in the setting of T2MI/acute myocardial injury is largely driven by the underlying process and influenced by comorbidities that often diminish physiological reserve; and troponin level is known to correlate with degree of severity of myocardial injury.⁶ Thus, it is unsurprising that higher peak cTnT correlates with increased mortality in T2MI. While troponin levels in T1MI have been shown to have predictive value with regards to likelihood of revascularisation, we did not observe this association in patients presenting with T2MI/acute myocardial injury.⁵

Stratification of peak cTnT against rates of coronary revascularisation in those who underwent coronary angiogram was not significant in those presenting with T2MI/acute myocardial injury. This is likely to be reflective of the understood mechanisms of T2MI and acute injury: given that the degree of myocardial injury is the result of the magnitude of systemic illness interacting with the amount of myocardium under stress, cTnT elevation in the setting of T2MI would not be expected to correlate well with revascularisation requirement. Conversely, cTnT elevation in the setting of obstructive coronary lesions causing T1MI is expected to correlate with revascularisation requirement, given that more proximal lesions are expected to cause greater insult and are anatomically more amenable to revascularisation, hence the subtle positive trend of revascularisation likelihood with greater peak cTnT concentration that we observed in patients presenting with T1MI. Despite the lack of association between rates of revascularisation in T2MI/acute myocardial injury and peak cTnT, the presence of a small proportion of these patients requiring revascularisation at all is notable. Although we have demonstrated that peak cTnT cannot be used to guide revascularisation in T2MI/acute myocardial injury, further work to determine the parameters to guide selection and the optimal method of evaluation would be of interest.

Morbidity of T2MI is known to be driven by intercurrent illness and comorbidities, neither of which are remedied by revascularisation.¹¹ Therefore, given that revascularisation is yet to be proven to reduce the morbidity of T2MI, its role in the management of T2MI is in question. As an observational study, this work was not equipped to quantify a mortality benefit associated with angiography in T2MI/acute myocardial injury. However, it was noted that those presenting with T2MI/acute myocardial injury who underwent coronary angiography had a lower 12-month mortality rate, but a higher rate of subsequent MI than those who did not undergo coronary angiography. The lower mortality rate in those who underwent coronary angiography is potentially influenced by selection bias: patients presenting with T2MI/acute myocardial injury who underwent angiography are likely to have a higher degree of physiological fitness. Patients referred for invasive investigation are expected to be able to survive the procedure and subsequent management and thus a degree of fitness is required, which would also confer a baseline survival advantage over those deemed not sufficiently fit to be referred. The higher rate of MI recurrence despite coronary angiography is likely to have been influenced by the high incidence of non-T1MI, and perhaps also reflective of the mechanisms of MI as described above. However, it is particularly notable that the population of patients with T2MI/acute myocardial injury who underwent angiography had a higher rate of subsequent T1MI than those who did not undergo angiography.

It is possible that the occurrence of T2MI/acute myocardial injury heralds future events and reinforces the well-established consideration of T2MI/acute myocardial injury as a cardiovascular risk factor. We observe that there exists a clinically selected, albeit small, population of patients with T2MI/acute myocardial injury who are at risk of future T1MI, which raises the question of how exactly these patients were selected. Unfortunately, it is unknown as to which parameters the clinicians used to refer these patients for coronary angiography. Given that we have shown that peak cTnT did not correlate with initial revascularisation requirement, we are prompted to consider whether other avenues of investigation could serve to better delineate this population at risk. CT coronary angiography (CTCA) is a common method of coronary assessment with evolving purpose. Coronary plaque volume and attenuation on CTCA have been shown to have predictive value with regards to future cardiac events.¹² CTCA may be a reasonable initial avenue of coronary investigation in patients presenting with T2MI to help stratify future risk and identify patients in whom secondary preventative measures are warranted.

A number of limitations may have impacted our findings. It is unknown whether cTnT level was a considered factor in the selection of patients for angiography: this information was not obtained from clinicians, although we do observe this pattern anecdotally. It is also notable that data were collected during a time of reporting change of troponin concentration. Although these factors may have impacted the observed trends, no relationship with peak cTnT was seen in those with T2MI/acute myocardial injury who were referred for revascularisation. The machine learning-based algorithm used to phenotype each episode, although validated, was not backed by clinical adjudication, the typical means of diagnosis.¹⁰ Additionally, peak cTnT is in fact a parameter used by the algorithm for categorisation. Thus, it is possible that the potential for miscategorisation and/or pre-influenced phenotyping affected the observed associations, especially in the comparisons between each type of myocardial injury.

Factors that influenced the decision to revascularise, such as lesion characteristics, were not measured; additionally, the COVID pandemic may have negatively impacted the referrals for revascularisation in patients with non-T1MI: invasive coronary assessment of troponin elevations attributed to myocarditis, or other aetiologies of non-T1MI, may have been restricted in a pandemic context. However, this does not negate the identification of those presenting with T2MI/acute myocardial injury who did proceed to revascularisation, an observation that prompts future interest; the notably small proportion of patients who were revascularised may have hidden associations that might be otherwise more noticeable with a larger population.

Conclusion

T2MI and acute myocardial injury are associated with increased cardiovascular risk, however, peak cTnT concentration in patients with T2MI did not predict revascularisation. Furthermore, we observed that coronary angiography was not associated with a reduction in subsequent MI in patients presenting with T2MI and acute myocardial injury. As such, we propose that non-invasive coronary evaluation with CTCA could play a role in further risk stratification of patients presenting with T2MI. The role of revascularisation in T2MI is also of future interest.