Perioperative myocardial injury is a frequent complication observed in patients undergoing cardiovascular surgery. For individuals undergoing surgical revascularisation, the monitoring of myocardial injury plays a crucial role in perioperative management.1,2 Moreover, perioperative myocardial injury is associated with reduced long-term survival.
The identification of myocardial injury and possible salvage intervention may improve the outcomes of coronary revascularisation. However, the detection of such injuries is affected by the methods chosen, the definition or cut-off criteria applied and the impact of cardiopulmonary bypass. Indeed, the clinical significance of perioperative myocardial injury can vary significantly based on the criteria used.
ECG, cardiovascular MRI and blood biomarkers are commonly used to identify myocardial injury. Of the cardiac-specific biomarkers, troponin I (TnI) is the most widely used for monitoring myocardial injury.3 Previous data have indicated that TnI is superior to ECG and creatinine kinase MB in predicting clinically significant myocardial injury after coronary artery bypass grafting (CABG).4 Several cohort studies have demonstrated that perioperative elevations in cardiac biomarkers are associated with increased early postoperative complications and poorer long-term outcomes.5,6 However, the potential confounding effect of cardiopulmonary bypass on myocardial injury complicates the interpretation of these findings. Furthermore, the precise distribution of TnI concentrations and the underlying mechanisms contributing to their elevation remain incompletely understood. Therefore, the aim of the present study was to investigate clinical variables associated with elevations in TnI, as well as the implications of TnI concentrations for both early and long-term patient outcomes.
Methods
This was a retrospective single-centre study conducted in a large tertiary heart institute. The study was approved by the Institutional Review Board of Anzhen Hospital (No. 2024047X).
We conducted a retrospective review of patients who underwent off-pump (OP) CABG at Anzhen Hospital between 2011 and 2022. Data on TnI and high-sensitivity (hs) TnI concentrations were collected. Individuals diagnosed with acute MI when admitted or those undergoing emergency CABG were excluded from the study. Patients with preoperative TnI >5 μg/l or preoperative hsTnI >5,000 pg/ml were also excluded. Baseline diagnoses, including hyperlipidaemia, were extracted from the patient history. TnI concentrations were measured using the AQT90 FLEX TnI test kit (Radiometer Medical; upper reference limit 0.023 μg/l), whereas hsTnI concentrations were measured using a UniCel DxI 800 Access Immunoassay System (Beckman Coulter; upper reference limit 11.6 pg/ml). TnI values were transformed into hsTnI equivalents by multiplying by 1,000.
Graft flow was measured using the VeriQ and MiraQ Cardiac systems (Medistim) during surgery, with the results archived in patients’ medical records. Patient follow-up was conducted via telephone. For those patients lost to follow-up, the last visit to Anzhen Hospital served as the censoring date.
Continuous data are expressed as either the median with interquartile range (IQR), and were compared using the Mann–Whitney U-test, or as the mean ± SD, with the significance of differences analysed using independent t-tests. Categorical data are presented as numbers and percentages and were compared using the Chi-squared test. For propensity score matching, patients who died in hospital were excluded. Unbalanced variables in the comparison of baseline characteristics were moved to the propensity score matching analysis.
Statistical analyses and data plotting were conducted using IBM SPSS Statistics version 25 and GraphPad Prism version 8.0.
Results
In all, 19,196 patients were included in the present study. The median age of patients was 63 years (IQR 57–68 years) and 14,423 (75.1%) were male. Of the patients analysed, 18,487 underwent postoperative TnI tests and 10,743 underwent hsTnI tests. The median peak TnI was 0.31 μg/l (IQR 0.14–0.87 μg/l), while the median peak high-sensitivity TnI was 781 pg/ml (IQR 386–1,792 pg/ml). TnI concentrations were log transformed, with the distribution of concentrations shown in Figure 1. Based on TnI concentrations in the present study and previously published data, patients were categorised into a low TnI group (TnI <10 μg/l or hsTnI <10,000 pg/ml) and a high TnI group (TnI ≥10 μg/l or hsTnI ≥10,000 pg/ml).7 The baseline characteristics of patients in both groups, as well as for the entire study cohort, are presented in Table 1. Conditional backward regression analysis revealed that increased age (OR 1.016; 95% CI [1.010–1.023]; p<0.001), hypertension (OR 1.415; 95% CI [1.251–1.601]; p<0.001), a history of cerebrovascular disease (OR 1.262; 95% CI [1.061–1.501]; p=0.009), diabetes (OR 0.785; 95% CI [0.696–0.885]; p<0.001) and hyperlipidaemia (OR 0.863; 95% CI [0.770–0.968]; p=0.012) were independently associated with postoperative TnI concentrations.
The high TnI group had a higher incidence of intra-aortic balloon pump (IABP) insertions (17.8% versus 2.9%; p<0.001), a higher incidence of extracorporeal membrane oxygenation (ECMO) support (3.6% versus 0.1%; p<0.001) and a greater number of in-hospital deaths (2.7% versus 0.2%; p<0.001) than the low TnI group. Figure 2 shows the predictive ability of peak TnI concentrations for IABP insertion (Figure 2A; area under the curve [AUC]=0.708), ECMO support (Figure 2B; AUC=0.877), in-hospital death (Figure 2C; AUC=0.775) and a composite outcome (IABP insertion, ECMO support, and in-hospital death) (Figure 2D; AUC=0.712). After surgery, 48 patients underwent postoperative coronary angiography; of these patients, 21 subsequently underwent percutaneous coronary intervention. In addition, 37 patients underwent a second CABG after re-exploration. Of the 58 patients who underwent repeat revascularisation, 36 were in the high TnI group. The probability of re-intervention on coronary arteries was significantly higher among patients with high TnI (2.81% versus 0.12%; p<0.001).
To elucidate the operative factors associated with elevated TnI, a propensity score matching analysis was conducted. After matching, 1,210 patients in the high TnI group and 1,210 patients in the low TnI group entered the next round of analysis. As indicated in Supplementary Table 1, baseline characteristics were well-balanced between the two groups. Graft flow data were available for 1,168 patients in the high TnI group and for 1,170 patients in the low TnI group. The number of grafts to the left anterior descending artery (LAD), the left circumflex coronary artery (LCX) and the right coronary artery (RCA) were compared between the two groups. The number of grafts to the LAD was comparable between the high and low TnI groups (1.47 ± 0.54 versus 1.46 ± 0.52, respectively; p=0.879; Figure 3A), but the high TnI group had a lower number of grafts to the LCX (0.71 ± 0.58 versus 0.81 ± 0.57, p<0.001; Figure 3B) and RCA (0.89 ± 0.53 versus 0.95 ± 0.53, p=0.011; Figure 3C). Consequently, the total number of grafts was lower in the high TnI group than in the low TnI group (3.07 ± 0.91 versus 3.22 ± 0.89; p<0.001; Figure 3D).
Similarly, the high TnI group had lower flow in the LCX (33 [IQR 21–55] versus 41 [IQR 25–67] ml/min; p<0.001; Figure 4B) and RCA (30 [IQR 18–50] versus 35 [IQR 22–55]; p<0.001; Figure 4C) than the low TnI group, but flow in the LAD (56 [IQR 33–86] versus 58 [IQR 35–92] ml/min, respectively; p=0.070; Figure 4A) was comparable between the two groups. Total flow to the coronary vessels was lower in the high TnI group than in the low TnI group (110 [IQR 73–163] versus 128 [IQR 87–184] ml/min, respectively; p<0.001; Figure 4D).
The long-term outcomes of the matched cohort were examined, encompassing the vital status of 2,364 patients (high TnI 1,173; low TnI 1,191). The median follow-up time was 1.74 years (IQR 0.28–3.68 years). As shown in Figure 5, individuals with a high peak TnI level postoperatively had significantly reduced long-term survival (HR 2.59; 95% CI [1.76–3.82]; p<0.001).
Discussion
Blood TnI concentrations are commonly used for monitoring myocardial injury. Previous studies have investigated the impact of elevated TnI concentrations on perioperative outcomes. However, there is a scarcity of data concerning causative factors and long-term consequences. In this study, we observed an association between postoperative TnI levels and both baseline characteristics and the grafts transplanted in a sizable OPCABG cohort. An increased postoperative TnI peak may be attributed to incomplete revascularisation of the LCX and RCA. Elevated TnI levels were further correlated with an increased frequency of IABP insertion, greater ECMO support and a higher incidence of in-hospital mortality. In addition, long-term survival after CABG was reduced compared with survival in individuals with lower TnI levels.
Various cut-off values for TnI have been established to identify periprocedural MI. The number of affected patients can differ significantly when using different criteria, leading to variations in the clinical significance of periprocedural myocardial injury.8–10 In this study, we opted for an arbitrary cut-off value to categorise patients into a high TnI group and a low TnI group. The cut-off value was decided upon based on the distribution of TnI concentrations in our cohort and insights derived from prior studies.7,11,12 Thielmann et al. reported a TnI cut-off of 10.5 ng/ml as the most effective in distinguishing between the release of TnI ‘in general’ and inherent TnI release after CABG.7
Periprocedural myocardial injury following coronary revascularisation is not an uncommon occurrence. Many factors influence TnI release, including mechanical manipulation, ischaemia–reperfusion injury, native coronary artery injury, graft failure, perioperative tachyarrhythmia, increased left ventricular end-diastolic pressure, cardiopulmonary bypass and cardioplegic arrest.13,14 In our cohort, all patients underwent isolated OPCABG. Because OPCABG avoids the impact of cardiac arrest on myocardial injury, the increase in TnI is primarily attributed to other causes. However, an early randomised study by Pegg et al. found that on-pump beating heart CABG was associated with more irreversible myocardial injury and higher troponin levels.15 This finding contrasts with a recent study by Matsuhashi et al., who reached the opposite conclusion.16 More observational data are needed to address this apparent discrepancy. Of note, in the present study, hypertension and age were associated with increased TnI after CABG, whereas diabetes and hyperlipidaemia were linked to lower TnI concentrations. This differs from a prior small study on myocardial injury following percutaneous coronary intervention.17 A possible explanation could be that the diabetes patients in the present study had more severe microvascular disease and were more acclimated to ischaemic conditions.
Numerous studies have consistently demonstrated an association between elevated TnI levels and increased perioperative complications and mortality.18,19 This correlation holds for patients undergoing both cardiac and non-cardiac surgery, with elevated preoperative TnI being linked to heightened perioperative risk. In the present study, we specifically excluded patients with elevated TnI before surgery and selected a high cut-off value for postoperative TnI. This approach was implemented to minimise the impact of non-ischaemic injuries. As anticipated, a high postoperative TnI concentration was found to be associated with an increased likelihood of requiring mechanical support and experiencing in-hospital mortality. The 17.8% IABP insertion rate at suggests that the majority of patients with elevated TnI levels maintain haemodynamic stability. This elevation in TnI is more likely attributable to native vessel occlusion rather than graft failure affecting larger vessels. This was further supported by the low early reintervention rate after the initial CABG in the present study.
Elevated TnI levels were also found to be associated with an increased long-term risk of all-cause mortality, suggesting that the myocardial injury incurred was permanent. In alignment with these findings, a recent study on beating-heart CABG by Kim et al. similarly indicated that a high postoperative TnI peak was linked to poorer long-term prognosis.20 Interestingly, these results contrast with another recent study on troponin T by Polzl et al., who found an association between troponin T and perioperative mortality but not 5-year mortality.21 Nevertheless, more studies have consistently demonstrated an increased risk of long-term mortality in patients with elevated perioperative TnI. An early meta-analysis suggested that the early elevation of troponin was associated with an increased risk of intermediate- and long-term mortality.19 However, this association was influenced by factors such as timing, cut-off values and assay methods.22 Moreover, Ranasinghe et al. revealed that serial TnI data collected 72 h postoperatively provided the best prediction for mid-term mortality.23 Litwinowicz et al. reported that perioperative MI was associated with increased early and long-term mortality.24 Despite these data on long-term impacts, the underlying mechanism has been unclear. Poor native coronary artery and unmet myocardial perfusion may be important causes.
An important finding of the present study is the association between elevated TnI levels and the extent of revascularisation. The grafts and flow to the LAD were comparable between the high and low TnI groups. However, significant differences were observed between the two groups with regard to grafts and flow to the LCX and RCA. These arteries present challenges with thinner branches and more difficult anastomosis. A lack of ideal target vessels is more likely to occur, and the exposure of these vessels and branches is limited, particularly in off-pump procedures. Numerous studies have indicated that OPCABG is associated with a higher incidence of incomplete revascularisation compared with on-pump CABG.25–28 Moreover, incomplete revascularisation has been linked to increased long-term mortality.29,30 Our study provides more detailed insights, revealing that incomplete revascularisation is more likely to occur in the LCX and RCA. Interestingly, this was in alignment with the study by Benedetto et al., who reported that off-pump surgery was associated with a lower rate of revascularisation of the LCX and RCA and resulted in increased 3-year mortality.31 In situations where achieving complete revascularisation in OPCABG proves challenging, an alternative strategy, including additional percutaneous coronary intervention, may reduce early complications and enhance long-term outcomes.
In the present study, we selected the peak postoperative TnI level as a representative indicator of myocardial injury. This decision was guided by prior evidence suggesting that peak troponin levels provide a reliable surrogate for the extent of myocardial damage. A small clinical study demonstrated that troponin T typically reaches its maximum concentration approximately 24 hours after CABG in patients with an otherwise uncomplicated perioperative course.32 In addition, Wang et al. reported that elevations in TnI within the first 48 hours after surgery were significantly associated with worse long-term outcomes, including increased mortality and adverse cardiovascular events.6 These observations underscore the importance of not only the magnitude but also the timing of the increase in troponin when interpreting its prognostic implications. Given the variability in the time course of troponin release depends on surgical factors, patient comorbidities and myocardial perfusion status, the temporal pattern of biomarker elevation may offer important insights into the underlying pathophysiology. For example, an early and rapid rise in TnI may reflect acute perioperative ischaemia or technical complications, whereas a delayed or sustained elevation may suggest ongoing myocardial stress or suboptimal revascularisation. Therefore, incorporating both the peak value and its timing could enhance the clinical utility of TnI as a biomarker in the risk stratification and postoperative management of patients undergoing CABG. Future studies should explore the prognostic value of serial TnI measurements and the potential role of integrating temporal biomarkers into predictive models for long-term outcomes.
The present study has certain limitations that should be acknowledged. First, the retrospective nature of the study, which was conducted in a large tertiary centre, introduces the possibility of selection bias; therefore, the study conclusions may not be readily generalisable to other centres. Second, TnI testing was performed at different times and using different methods. Only peak values were analysed, and the heterogeneity in data collection may have introduced unknown impacts on the results. Not all patients in the study underwent both TnI and hsTnI testing. Consequently, the conversion by multiplying TnI values by 1,000 to approximate hsTnI may have introduced methodological heterogeneity and increased variability within the dataset. In addition, the long-term follow-up only included survival data. The absence of analysis on functional status and adverse events weakens the conclusions drawn regarding late outcomes.
Conclusion
An increase in circulating TnI following OPCABG was found to be associated with increased early and late mortality. The higher TnI levels may be associated with incomplete revascularisation, particularly affecting the LCX and RCA.
Clinical Perspective
- Periprocedural myocardial injury may be associated more with incomplete revascularisation than early graft failure following off-pump coronary artery bypass grafting.
- A comprehensive strategy including hybrid approaches to achieve complete revascularisation may improve the outcomes of off-pump coronary artery bypass grafting.