Various definitions of myocardial infarction type 5 after coronary artery bypass grafting (CABG) have been proposed (myocardial infarction [MI-5], also known as peri-procedural MI), using different biomarkers and different and arbitrary cut-off values. This meta-analysis aims to determine the expected release of high-sensitivity cardiac troponin T (hs-cTnT) after CABG in general and after uncomplicated surgery and off-pump CABG in particular. A systematic search was applied to 3 databases. Studies on CABG as a single intervention and reporting on postoperative hs-cTnT concentrations on at least 2 different time points were included. All data on hs-cTnT concentrations were extracted, and mean concentrations at various points in time were stratified. Eventually, 15 studies were included, encompassing 2,646 patients. Preoperative hs-cTnT was 17 ng/L (95% confidence interval [CI] 13 to 20 ng/L). Hs-cTnT peaked at 6 to 8 hours postoperatively (628 ng/L, 95% CI 400 to 856 ng/L; 45x upper reference limit [URL]) and was still increased after 48 hours. In addition, peak hs-cTnT concentration was 614 ng/L (95% CI 282 to 947 ng/L) in patients with a definite uncomplicated postoperative course (i.e., without MI). For patients after off-pump CABG compared to on-pump CABG, the mean peak hs-cTnT concentration was 186 ng/L (95% CI 172 to 200 ng/L, 13 × URL) versus 629 ng/L (95% CI 529 to 726 ng/L, 45 × URL), respectively . In conclusion, postoperative hs-cTnT concentrations surpass most of the currently defined cut-off values for MI-5, even in perceived uncomplicated surgery, suggesting thorough reassessment. Hs-cTnT release differences following on-pump CABG versus off-pump CABG were observed, implying the need for different cut-off values for different surgical strategies.
Peri-procedural myocardial infarction (MI-5) after coronary artery bypass grafting (CABG), is a particular entity within the spectrum of myocardial infarction etiologies, and its separate definition should be applied in the first 48 hours after coronary surgery. To define peri-procedural MI, the universal definition of MI (UDMI) was proposed, with the incorporation of cardiac troponin (cTn) T (cTnT) or I (cTnI). In the latest UDMI-4, a cTn rise of >10-times the upper reference limit (URL) is required for the diagnosis of MI-5, in conjunction with imaging or electrocardiographic evidence. However, the scientific basis for this cut-off value is weak, and the authors acknowledge that this value was arbitrarily chosen. , In response, another definition was proposed by the Society of Cardiovascular Angiography and Interventions (SCAI) in 2013. For the SCAI definition, if cTn is used, an isolated rise of >70 × URL confirms the diagnosis of MI-5 without the need for other electrocardiographic or imaging evidence, whereas a rise of >35 × URL warrants accompanying imaging or electrocardiographic abnormalities. In response to these controversies and variations in MI-5 definitions, and to evaluate the legitimacy of the arbitrarily chosen diagnostic cTn cut-off values using the current high-sensitivity assays (hs-cTnT), the current systematic review and meta-analysis aim to determine the expected release of hs-cTnT after CABG in relation to definitions of MI in general, and patients with uncomplicated postoperative courses and different uses of cardiopulmonary bypass in particular.
A review protocol was drafted and registered in the PROSPERO registry (CRD42020175979; approval date: April 28, 2020). The current report was written following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement.
All studies reporting on patients undergoing isolated CABG were eligible for inclusion. Studies with patients exclusively treated in the acute phase (i.e., during or directly after an acute coronary syndrome) were excluded to ensure homogeneous and normal baseline hs-cTnT concentrations.
Studies enrolling patients treated with on-pump CABG or off-pump CABG (OPCAB) and reporting on postoperative hs-cTnT concentrations were included. Studies were excluded when they described patients that were treated by CABG with concomitant cardiac surgical procedures (i.e., valve, rhythm, and/or aortic surgery).
The main outcome was hs-cTnT concentration after CABG reported on at least 2 different points in time. Secondary outcomes were MI as a complication of CABG, the definition of myocardial infarction employed (UDMI-1,-3, or -4 or SCAI), and short-term mortality. All points in time of hs-cTnT measurement, about the timing of surgery, were extracted and uniformly reported.
Randomized studies and prospective and retrospective observational studies were deemed eligible.
Potentially eligible studies were searched through electronic scientific databases. A comprehensive search query was applied to the MEDLINE and EMBASE databases and the Cochrane library using terms such as “coronary artery bypass grafting,” “high-sensitivity cardiac troponin,” and alternative spelling through Ovid ( Supplementary Material 1 ). The last search was performed on March 23, 2021. Two experienced reviewers performed the search (JD, ED). The process of study selection was performed in a blinded standardized manner.
Study characteristics such as study type and enrollment period were extracted from all studies. Baseline patient characteristics, including age and gender, were extracted as well, in addition to the type of surgery performed (on-pump CABG or OPCAB). The predefined T 0 point in time was registered for all studies. All postoperative hs-cTnT concentrations, at every possible point in time, were extracted from the included study. Statistical conversion was performed if hs-cTnT was not reported as mean and SD, applying Wan’s method. Because studies reported hs-cTnT concentrations on nonuniform points in time, stratification for these points in time was performed. Based on the available data in the included studies, points in time were stratified in preoperative, 0 to 4 hours, 6 to 8 hours, 10 to 12 hours, 16 to 24 hours, 36 to 48 hours, and >48 hours postoperatively.
For this research question, the risk of bias assessment in terms of evaluation of a potential bias in favor of 1 treatment over another, does not seem fully applicable, as the aim was to evaluate postoperative hs-cTnT concentrations. Although it may be unlikely that the study design influences these values, we performed a risk of bias assessment using The Jadad Scale for Reporting Randomized Clinical Trials for randomized controlled trials, to determine study quality. To homogeneously assess this quality in the different observational studies, it was imperative to use a tool that was able to assess single- and 2-armed studies in the same manner. As reported in another meta-analysis, the Newcastle-Ottawa-Scale for assessment of cohort studies as recommended by the Cochrane Collaboration can be adapted for assessment of single-arm studies. The adaptation concerned removing 1 of 4 questions in the selection process (i.e., Question 2 in Selection). Two experienced investigators performed the quality assessment (SH, JD), whereby potential disagreements were resolved by consulting the senior author (AM). Studies with a Jadad scale score of 3 or less were considered low-quality, and those with a Newcastle-Ottawa-Scale score of 5 or less.
All statistical analyses were performed by 2 researchers with extensive expertise in statistics, specifically applied to meta-analyses. Due to the inherently different nature of single- and multiple-armed studies, we chose to report all hs-cTnT concentrations per study arm of the multiple-armed studies. Multiple-armed studies were allocated a number between brackets (i.e., [2]) after the study name. All available hs-cTnT concentrations were collected for the different points in time. Then, every study was allocated a certain weight per stratified point in time using inverse variance weighting. Data on hs-cTnT were pooled per point in time using random-effects modeling and reported as ng/L with 95% confidence intervals (CI). The I 2 -test assessed statistical heterogeneity. Post-hoc subgroup analyses were performed for studies reporting separately on hs-cTnT concentrations after on-pump CABG and OPCAB. In addition, studies that exclusively reported on an uncomplicated postoperative course were submitted to post-hoc analyses as well. Finally, peak hs-cTnT values of studies reporting on mortality as a clinical outcome were denoted to identify a potential relation using meta-regression (random-effects modeling). All analyses were performed using open-source software packages (OpenMetaAnalyst, available from http://www.cebm.brown.edu/openmeta/ ).
The study selection process is displayed in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flowchart in Supplementary Material 2 . The search identified 3,529 publications using the 3 databases. Duplicates were excluded (n = 662). Based on the title and abstract screening, 2,515 studies were excluded. Reasons for full-text exclusion were presented in the flowchart. After scrutinizing all reports, 15 studies were included for qualitative and quantitative analysis.
Study characteristics are listed in Table 1 . All studies were published between 2012 and 2020. Seven studies comprised prospective observational cohorts, , , , 6 studies were randomized controlled trials , , , , , and 2 studies were retrospective analyses. , Of the 9 retrospective and prospective observational studies, 8 studies were single-armed , , , , and 1 report had 3 study arms. The 15 studies encompassed data of 2,646 patients, of which 22% were female (n = 592). The mean age of the total patient group was 65 years (SD 11).
| First Author | Year | Country | Study Type | Study Arms | Patients | Age (Mean, SD) | Female, n (%) | Surgery Type |
|---|---|---|---|---|---|---|---|---|
| Candilio | 2014 | United Kingdom | retrospective | 3 | 54 | 66 (9) | 11 (20%) | on-pump CABG |
| Freiermuth | 2016 | Switzerland | RCT | 2 | 30 | 65 (8) | 4 (13%) | on-pump CABG |
| Ge | 2018 | China | retrospective | 1 | 398 | 64 (8) | 81 (20%) | OPCAB |
| Jakobsen | 2012 | Norway | RCT | 1 | 60 | 64 (7) | 5 (8%) | on-pump CABG |
| Koppen | 2019 | Norway | prospective | 1 | 626 | 66 (13) | 130 (21%) | on-pump CABG |
| Laurikka | 2014 | Finland | prospective | 1 | 32 | 69 (8) | 9 (28%) | on-pump CABG |
| Lucchinetti | 2012 | Canada | RCT | 2 | 55 | 61 (9) | 5 (9%) | on-pump CABG |
| Machado | 2019 | Portugal | prospective | 1 | 600 | 61 (10) | 177 (30%) | on-pump CABG and OPCAB |
| Markman | 2017 | Australia | prospective | 1 | 15 | 63 (12) | 4 (27%) | on-pump CABG and OPCAB |
| Mehdiani | 2017 | Germany | prospective | 1 | 213 | 68 (10) | 55 (26%) | on-pump CABG and OPCAB |
| Narducci | 2014 | Italy | prospective | 1 | 38 | 69 (11) | 11 (29%) | on-pump CABG |
| Nederlof | 2017 | Netherlands | RCT | 2 | 29 | 68 (9) | 0 | on-pump CABG |
| Petäjä | 2016 | Finland | prospective | 1 | 428 | 68 (10) | 86 (20%) | CABG type not reported |
| Pruthi | 2020 | India | RCT | 2 | 30 | 57 (8) | 3 (10%) | OPCAB |
| Tosun | 2013 | turkey | RCT | 2 | 38 | 62 (10) | 11 (29%) | on-pump CABG |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
