The study was approved by the local Ethics Committee (Comitato Etico Area Vasta Emilia Centro).

All methods were performed in accordance with the Declaration of Helsinki as well as with our institutional ethics principles, guidelines and regulations.

Given the observational and retrospective nature of the study, patients’ consent was inferred in line with the general consent expressed upon hospital admission and, as requested by the ethics committee, specifically updated at the first medical re-contact with the patient, if any.

We retrospectively enrolled all consecutive patients ≥ 18 years old who underwent CABG, alone or in association with other cardiac surgery/surgery on the ascending aorta, at S. Orsola-Malpighi Hospital from September 1, 2018, to September 1, 2019.

We analyzed each patient’s clinical record and collected demographic data as well as recent and past medical history, including the presence of cardiovascular risk factors and cardiovascular events such as previous MI, previous percutaneous and/or surgical coronary revascularization, previous cardiac surgery, and previous episodes of heart failure. Moreover, we gathered data about patients’ comorbidities, including the presence of peripheral artery disease, symptomatic chronic lung disease, chronic renal disease, and history of cerebrovascular disease (previous stroke and/or transient ischemic attack). Most of such variables were defined according to the European System for Cardiac Operative Risk Evaluation (EuroScore) II risk model, which we also calculated for each patient [10]. In particular: acute coronary syndromes were defined according to the related current guidelines [11]; previous MI was categorized as recent if occurring within the past 90 days before the date of current surgery; dyspnea and angina were graded according respectively to the New York Heart Association and to the Canadian Cardiovascular Society classification systems [12, 13]; previous heart failure was defined as a previous hospitalization for heart failure and/or signs/symptoms of heart failure and/or left ventricular ejection fraction ≤ 35% [14]; peripheral artery disease was defined as extra-cardiac arteriopathy manifested as any among lower limb claudication, carotid occlusion or > 50% stenosis, amputation due to arterial disease, previous or planned intervention on the abdominal aorta, limb arteries or carotids; symptomatic chronic lung disease was defined as long term use of bronchodilators or steroids for lung disease; poor mobility was defined as severe impairment of mobility secondary to musculoskeletal or neurological dysfunction with need of walking aid; the estimated glomerular filtration rate (eGFR) was calculated via the Cockcroft-Gault formula [15]; endocarditis was considered active when patients were still on antibiotic treatment for endocarditis at the time of surgery.

We also collected data on the surgical intervention. Surgical coronary revascularization was performed as indicated by current guidelines [2]. Based on the degree of urgency, surgery was classified as follows: salvage surgery in the case of need for cardiopulmonary resuscitation en-route to the operating theatre or before induction of anesthesia, emergency surgery in the case of need for operation before the beginning of the next working day after decision to intervene has been taken, urgent surgery in the case of all other patients whose admission and surgery were not scheduled, and elective surgery in the case of routine admission for operation. Critical preoperative state was defined as the pre-operative presence of any among ventricular tachycardia, ventricular fibrillation, aborted sudden death, cardiopulmonary resuscitation, ventilation (before entry into the anesthetic room), hemodynamic support by means of inotropes, intra-aortic balloon pump (IABP) or ventricular assist device, or acute renal failure (new-onset anuria or oliguria).

Furthermore, pre-operative and post-operative laboratory findings were reviewed, including hs-cTn I, along with pre-operative and post-operative electrocardiographic and echocardiographic findings, as well as post-operative angiographic findings when available.

Finally, post-operative events, including death, stroke (ischemic or hemorrhagic), reintervention, need of circulatory support (IABP and/or extracorporeal membrane oxygenation, ECMO), renal replacement therapy, final diagnosis of PMI, and total hospitalization length were recorded.

In line with the Fourth Universal Definition of Myocardial Infarction [5] CABG-related PMI was defined as an elevation of hs-cTn I values > 10 times the 99th percentile Upper Reference Limit (URL) in patients with normal baseline hs-cTn I values; for patients with known elevated but stable (≤ 20% variation) or falling baseline hs-cTn I, post-procedural hs-cTn I must be > 10 times the 99th percentile URL and rise by > 20% with respect to the pre-procedural value. An additional criterion was also required among the development of new pathological Q waves on electrocardiogram (ECG), the imaging evidence of new loss of viable myocardium or of new regional wall motion abnormality in a pattern consistent with an ischemic etiology, or the finding of new graft or native coronary artery occlusion at angiography.

Cardiac troponin values had been repeatedly measured at fixed time intervals after surgery by using the Access hs-cTn I assay by Beckman Coulter. The 99th percentile URL established in a population of healthy adults were, respectively, 11.6 ng/L for females and 19.8 ng/L for males [16].

The prespecified clinical outcomes of the study were represented by in-hospital death or PMI, in-hospital death, PMI, and LVEF decrease ≥ 10% after CABG surgery. In particular, we investigated the power of hs-cTn I in predicting the above-mentioned clinical outcomes.

Continuous variables are presented as means ± standard deviations (SD) or medians and interquartile ranges (IR), as indicated depending on the normality of the distribution, while categorical variables are presented as numbers and percentages. The non parametric Wilcoxon rank-sum test was applied in order to compare continuous variables while the Chi-squared test was applied in order to compare categorical variables.

Univariable logistic regression analysis was used to evaluate the unadjusted association between LVEF decrease ≥ 10% after CABG surgery as this was the only outcome which resulted significantly associated with hs-cTn I values in our study population. The following variables were examined at univariable logistic analysis as significantly associated with LVEF decrease ≥ 10%: hs-cTn I in the first 24 h following CABG, previous cardiac surgery, number of surgical procedures, cardiopulmonary bypass time, aortic cross clamp time, new regional wall motion abnormalities at post-operative echocardiogram, and PMI.

Receiver-operating characteristic (ROC) curves of the post-operative hs-cTn I values identified as predictor of LVEF decrease ≥ 10% were performed to evaluate the diagnostic performance in terms of discrimination power [area under the receiver operator characteristic curve (AUC)]. The Youden index was then used to identify the optimal cut-off value.

P-values less than 0.05 were deemed statistically significant.  

All analyses were performed with IBM SPSS Statistics package for Windows, version 25.0 (BM Co., Armonk, NY, USA).