Coronary artery disease (CAD) remains one of the leading causes of death and disability worldwide. Coronary artery bypass graft (CABG) surgery is one of the most important therapeutic strategies for patients with multivessel CAD. Despite recent advancements in surgical techniques, CABG still causes significant mortality and morbidity, especially given the ageing population and increasing prevalence of risk factors such as diabetes, obesity, chronic kidney disease and hypertension. Among the many factors associated with worse clinical outcomes after CABG, the occurrence of procedural MI (PMI; type 5 MI) is a critical determinant of clinical outcomes. The most sensitive and specific cardiac biomarkers for detecting PMI are serum elevations of high-sensitivity cardiac troponin (hscTn), which is released from cardiomyocytes following injury. It comprises two isoforms: high-sensitivity cardiac troponin T (hscTnT) and high-sensitivity cardiac troponin I (hscTnI). Both are highly specific to the myocardium, enabling their use as biomarkers for assessing PMI following cardiac surgery.
Current definitions of PMI following cardiac surgery in the Fourth Universal Definition of MI (UDMI) or by the Society for Cardiovascular Angiography and Interventions (SCAI) or Academic Research Consortium (ARC)-2 have been based on diagnostic criteria rather than prognostic ones, with cut-off cardiac troponin (cTn) thresholds arbitrarily chosen because of a lack of available evidence when these definitions were proposed. 1–3 Despite this, a number of clinical studies and meta-analyses have reported that, providing the procedural elevation in cTn is accompanied by new evidence of ischaemia, these definitions for PMI are prognostically important, impacting on short- and long-term clinical outcomes following cardiac surgery.4
Current definitions of PMI require the presence of new evidence of ischaemia, but this can be challenging to ascertain following cardiac surgery, given the difficulties in assessing chest pain and ECGs, along with diagnosing ECG and imaging evidence of new ischaemia. As such, a number of clinical studies and meta-analyses have investigated the prognostic impact of PMI following cardiac surgery as defined by isolated elevations in cardiac biomarkers such as creatine kinase MB (CK-MB) and cTn. These have been associated with worse short- and long-term clinical outcomes, but the specific cut-off thresholds for defining prognostic PMI are not clear.
The Fourth UDMI criteria for defining procedural myocardial injury require a cTn value >99th percentile upper reference limit (URL) during the first 48 hours following cardiac surgery, occurring from a normal baseline cTn value. However, based on this definition, the vast majority of patients undergoing cardiac surgery will experience procedural myocardial injury.5 To address this, higher cut-off thresholds for isolated elevations in standard cTn and CK-MB in the 48-hour post-procedure period have been proposed to indicate the presence of prognostically significant procedural myocardial injury following cardiac surgery.2,3,6 Several studies have evaluated the prognostic impact of isolated hscTn elevations on short- and long-term clinical outcomes and have proposed much higher cut-off thresholds for defining PMI that have prognostic significance and may be used to guide patient management and be used as clinical endpoints in future cardiac surgery trials. However, the cut-off thresholds of hscTn elevation do vary depending on the type of cardiac surgery performed, the timing of the blood sampling and whether hscTnT or hscTnI are used. Hence, it is challenging to propose a single cut-off threshold for defining prognostic elevations in hscTn post-cardiac surgery.7 Therefore, there is a need for a more universally accepted and widely used diagnostic and prognostic criteria for PMI in CABG.
The current systematic review has two aims. The first is to summarise the current available evidence using hscTn to evaluate the presence of PMI, by estimating the accuracy of hscTn according to sensitivity and specificity. The second is to evaluate the diagnostic and prognostic cut-off thresholds for defining PMI after CABG.
Methods
Literature Search
A comprehensive search of electronic databases was conducted using PubMed, Embase, Web of Science and Cochrane Library until December 2024. The literature search was confined to English publications with no region restriction. The detailed literature search strategy is provided in the Supplementary Material.
Study Selection and Inclusion
After removing duplicates, two researchers independently screened the titles and abstracts of the remaining records to select articles for full-text assessment. Any discrepancies were resolved through a group meeting with a third investigator. To ensure completeness, the reference lists of the included studies were meticulously reviewed and a citation search was also conducted on all eligible studies.
The inclusion criteria were as follows: participants were adults with PMI and high-sensitivity troponin tests after CABG or CABG concomitant with valve surgery; the index was the high-sensitivity troponin test; the target condition was PMI; and the study types were randomised controlled trials or observational studies (cross-sectional or longitudinal). The following were excluded: articles written in languages other than English; reviews, comments, protocols, editorials, letters, case reports, or animal trials; studies without hscTn to define PMI; and studies with sample sizes <20.
Data Extraction
Studies that met our inclusion criteria and those from which complete information about diagnostic and prognostic value could be derived were selected for inclusion. Data extraction was initially conducted by two independent reviewers via a standardised Excel (Microsoft) table. Basic characteristics including the first author, year of publication and country where the study was performed were extracted.
Results
Twenty-seven studies met the inclusion criteria; 20 studies used hscTnT and seven used hscTnI. The flow chart of the literature search and included studies is shown in Figure 1.
High-sensitivity Cardiac Troponin T in Coronary Artery Bypass Graft Surgery
The characteristics and major findings for hscTnT in PMI diagnosis after CABG are listed in Supplementary Table 1.4,8–26 Among the studies, there were 10 retrospective studies, six randomised clinical trials and four prospective observational studies. There were multiple types of CABG, including on-pump CABG (the majority), off-pump CABG and CABG concomitant with valve surgery. All studies used the Elecsys hscTnT assay (Roche), except one German study that used the ENZYMUN assay (Roche).
The diagnostic values of hscTnT in predicting short-term outcomes in CABG are listed in Table 1. The primary outcomes include operative mortality, 30-day mortality, in-hospital major adverse cardiovascular events (MACE), 30-day adverse events, postoperative in-hospital complications and Type 5 MI. The cut-off range, area under the curve (AUC) and 95% CI, reported sensitivity, specificity and Youden J value are listed. The seven studies listed in Table 1 report eight cut-off ranges. Based on the reported data, 120 × URL had the highest sensitivity (100%) in predicting type 5 MI, while 100 × URL had the highest specificity (99%). Regarding Youden J (sensitivity + specificity-1), 120 × URL had the highest diagnostic value.
The prognostic value of hscTnT for short-term mortality prediction in CABG is shown in Table 2. Three studies were included; two defined the primary outcome as 1-year mortality and one as 180-day mortality. Based on the current reported data, 35 × URL had the highest sensitivity for predicting 1-year all-cause mortality (84%) and 170 × URL had the highest specificity. Regarding Youden J, 70 × URL had the highest Youden J index.
The prognostic value of hscTnT in predicting short-term and mid-term outcomes is shown in Supplementary Tables 2 and 3 for different reported cut-off ranges. For short-term mortality and 1-year all-cause mortality, 170 × URL and 70 x URL had the highest OR, respectively.
High-sensitivity Cardiac Troponin I in Coronary Artery Bypass Graft Surgery
Seven studies reported the diagnostic and prognostic value of hscTnI after CABG surgery; most were published after 2022. The basic characteristics are listed in Supplementary Table 4.27–33 Among the seven studies, six were retrospective and one was prospective. Most studies used the ARCHITECT STAT (Abbott) hscTnI immunoassay; one study used the hscTnI assay by Beckman Coulter.
Regarding short-term outcome prediction (Table 3), four studies reported sensitivity and specificity. Among them, 80 × URL had the highest sensitivity, while 307 × URL had the highest specificity. Regarding the Youden J index, 307 × URL had the highest diagnostic value. Up to the end date of the literature search, no data have been available concerning the application of hscTnI in predicting long-term outcomes.
Regarding short-term outcome prediction assessment (Supplementary Table 3), three studies were available for data extraction and 500 × URL had the highest OR for outcome prediction. Regarding long-term outcome assessment, one study demonstrated that 36 × URL was predictive of the risk of 1-year mortality (Supplementary Table 4).
Discussion
Summary of Evidence
The findings of this review underscore the need for a more universal definition for PMI diagnosis using hscTn. As is well known, the profiles of hscTn release are surgery type-, patient- and comorbidity-specific. These observations align with previous literature, highlighting the need for a universally accepted definition of PMI.4–7
Significance of High-sensitivity Cardiac Troponin in Diagnosing PMI after Coronary Artery Bypass Graft Surgery
PMI is defined as an isolated elevation in cardiac biomarkers above the URL in the 48-hour postoperative period. Because of the nature of cardiac surgery, this level of cardiac biomarker elevation occurs in virtually all patients undergoing CABG surgery. As in cardiac surgery, the nature of the CABG surgery involves multiple manipulations, all of which may lead to increases in biomarkers.
Recent high-sensitivity assays have demonstrated that cardiac troponin levels are increased in many diseases affecting the myocardium and offer crucial prognostic data. Many studies have shown that the hscTnT assay has better diagnostic and prognostic accuracy than the traditional troponin T assay in a variety of patient populations with cardiovascular diseases.34 The pattern of change in biomarkers can help in distinguishing graft-related and non-graft-related failure after cardiac surgery.7
Defining the Threshold for PMI Diagnosis
Different definitions and criteria will lead to different analysis outcomes. As such, there is an urgent need for a universal definition that can apply to different clinical scenarios. The diagnostic criteria for different types of cardiac surgeries vary. In addition, the studies vary in terms of the surgery type, concomitant procedures, timing used to define the end points and timing of the hscTn test, as well as the assay of hscTn, the underlying complexity of the disease and whether patients had experienced previous MI within 30 days.
Current definitions recommend cardiac troponin values >10, 35 and 70 × URL after CABG procedures (with or without additional new signs of ischaemia) for the diagnosis of PMI.35 Recent studies provide evidence that thresholds for cardiac biomarkers defining PMI used in most consensus statements are too low to predict mortality after cardiac surgery.4,26,27,30,36 This is consistent with our observations.
The large international VISION cardiac surgery study evaluated the prognostic impact of hscTnI elevations within the postoperative 24-hour period on 30-day mortality in 13,862 cardiac surgery patients recruited from 24 hospitals in 12 countries.27 The authors found that the >10 × URL threshold was exceeded in 98% of patients and that the ≥35 × URL and ≥70 × URL thresholds were exceeded in 89% and 75% of patients, respectively, confirming that the cut-off thresholds in cTn elevation used for defining PMI in the Fourth UDMI and SCAI are too low. Interestingly, the 24-hour hscTnI threshold associated with an increased risk of 30-day mortality was >210 × URL after isolated CABG or aortic valve replacement; this decreased to 59 × URL at days 2 and 3 post-surgery. For other cardiac surgeries, the 24-hour hscTnI threshold was 499 × URL; this decreased to 96 × URL at days 2 and 3 post-surgery. In a subsequent large, single-centre cohort study of 8,292 patients undergoing cardiac surgery, the hscTnT cut-off peak value within 72 hours of surgery associated with 30-day mortality varied from 41 × URL for aortic valve replacement, to 134 × URL for other cardiac surgery and to 170 × URL for isolated CABG surgery.26 Furthermore, higher troponin levels were observed at all postoperative timepoints in on-pump CABG than in off-pump CABG surgeries. Interestingly, elevations in hscTnT above this prognostic cut-off were associated with an increase in 5-year mortality for patients receiving CABG and other cardiac surgery, but not for those receiving aortic valve replacement. A recent study investigated the impact of postoperative hscTnI elevations within 48 hours and their implications for postoperative clinical decision making in 4,684 consecutive patients undergoing elective isolated CABG at a single centre.30 The authors found that 161 patients (3.5%) underwent invasive coronary angiography after surgery, of whom 86 (53.4%) underwent repeat revascularisation. An optimal cut-off value for peak hscTnI of >500 × URL was significantly associated with repeat revascularisation within 48 hours after surgery, 30-day MACE and all-cause mortality after a median follow-up of 3.1 years. Interestingly, the addition of serial hscTnI measurements showed no additional benefit in patients with electrocardiographic or echocardiographic abnormalities or haemodynamic instability. Early postoperative hscTnI elevations (before 12 hours) also had a low yield for clinical decision-making; only later elevations (at 12–16 hours postoperatively) using a threshold of 307 × URL were significantly associated with repeat revascularisation with an area under the curve of 0.92.
Following cardiac surgery, PMI due to graft-related and non-graft-related failures have different mechanisms and patterns of biomarker release.31 This suggests that surgery-, sex-, time- and comorbidity-specific cut-off values are necessary for clinical significance, highlighting the necessity for distinct cardiac biomarker cut-off values tailored to each strategy, based on clinically significant outcomes rather than arbitrary thresholds.
Study Limitations
The study has several limitations. The study is a systematic review without meta-analysis. More evidence is needed to perform a high-quality meta-analysis. Other types of cardiac surgeries (valvular heart surgery, aortic surgery and heart transplantation) were not included in the systematic review. Given the lack of available literature and the differences in cardiac surgical populations (background risk of mortality), surgery types, assays, timing of assay and so on, it is challenging to define cut-off prognostic thresholds of hscTn. Furthermore, the optimal threshold level of hscTn elevation that provides optimal sensitivity and specificity regarding PMI after cardiac surgery has not been defined.
Future Directions
The utmost priority is to establish a universally accepted diagnostic and prognostic threshold for PMI. Current studies use heterogeneous cut-off values, often derived from small cohorts without accounting for variables such as age, renal function or surgical technique. Future multicentre studies should employ large, diverse populations to validate thresholds adjusted for baseline troponin levels and comorbidities.
In the era of multimodality diagnostic tools and rapid development of artificial intelligence, a combined model of biomarker, CT angiography, MRI and echocardiographic data would be helpful in addressing this problem. Combining hscTn with advanced imaging (e.g. cardiac MRI for late gadolinium enhancement) or novel biomarkers may differentiate ischaemic necrosis from reversible injury and from graft-related and non-graft-related PMI.37 Additionally, intraoperative factors, such as haemodynamic changes (e.g. low blood pressure, low cardiac output and low oxygenation, etc.), transit-time flowmetry linking intraoperative graft quality with postoperative troponin kinetics profiles, may help address this crucial clinical problem.38,39 Based on the current literature, elevated postoperative hscTn levels are associated with short-term adverse outcomes; however, the long-term morbidity and follow-up results need to be further validated. Validation of long-term results should be incorporated in large cohorts with longer follow-up periods.
Procedural Myocardial Injury and Infarction as a Clinical Trial Endpoint
The choice of definition of PMI has been shown to influence the outcome of several recent clinical trials including ISCHEMIA, SYNTAXES and EXCEL.40–42 This underscores the need to have a universal definition of PMI following either percutaneous coronary intervention or cardiac surgery that can be used as clinical endpoints for trial comparing outcomes between PCI and cardiac surgery. As such, further research is needed to better define PMI when used as a primary composite endpoint in clinical trials; however, which definition should be used needs further study. Although ARC-2- and Society for Cardiovascular Angiography and Interventions (SCAI) SCmodel-defined PMI are more specific and thus more prognostic for cardiac death, the Fourth UDMI definition is more sensitive for detecting events and remain independently associated with cardiac mortality.36 More data from adequately powered cohorts are needed to compare prognosis of PMI among PCI and cardiac surgery patients across definitions to allow a fair comparison of event rates in a clinical trial.
Conclusion
HscTn has been proven to be of diagnostic and prognostic value for PMI after CABG. Overall, 120 × URL and 70 × URL of hscTnT are the most effective cut-off values for diagnosing PMI for short-term and long-term MACE after CABG, respectively. The optimal threshold value for hscTnI is 307 × URL for diagnosing short-term MACE. To conclude, despite accumulating evidence, there are still no universally accepted definitions and diagnostic criteria for PMI after cardiac surgery. Such universally accepted and accurate definitions supported by robust evidence are urgently needed, although this will be challenging given the heterogeneity of cardiac surgery.
Clinical Perspective
- The occurrence of procedural MI (PMI) is a critical determinant of clinical outcomes following cardiac surgery.
- Although elevations in blood levels of high-sensitivity cardiac troponin following cardiac surgery have been associated with worse clinical outcomes, the optimal diagnostic and prognostic cut-off values for defining prognostic PMI remain uncertain.
- However, defining the prognostic cut-off thresholds for high-sensitivity cardiac troponin elevation following cardiac surgery remains challenging given the varying conditions such as the type of surgery, the underlying patient risk, the troponin assay used and the timing of the assay.