Modeling creatine-kinase MB concentrations following coronary artery bypass grafting

Abstract

An increase in cardiac biomarkers is a prerequisite to diagnose periprocedural myocardial infarction (PMI) after coronary artery bypass grafting (CABG). Early-phase risk detection may be aided by modeling time-dependent serum creatine kinase-MB (CK-MB) concentrations. This study aimed to model the kinetics of CK-MB, while identifying its influencing factors. Patients who underwent elective CABG and had CK-MB measurements within 72 hours postoperatively were included. The primary outcome was the modeled post-hoc kinetics of CK-MB in patients without potential PMI. These patients were defined as having no potential PMI in case of absence of: ischemic electrocardiographic abnormalities, imaging abnormalities, in-hospital cardiac arrest, mortality, or postoperative unplanned catheterization. A web-based application was created using mixed-effect modeling to provide an interactive and individualized result. 635 patients underwent elective isolated CABG, resulting in 1589 CK-MB measurements. Of these, 609 patients (96%) had no potential PMI, while 26 (4%) had potential PMI. Male sex, aortic cross-clamp time, and cardioplegia type significantly impacted CK-MB concentrations. The diagnostic accuracy of the model had an area under the ROC curve of 82.8% (72.6-90.2%). A threshold of 7 μg/L yielded a sensitivity of 94% and a specificity of 80% (positive predictive value, 17%; negative predictive value, 99%) for excluding potential PMI in our own study population. CK-MB release after CABG depends on the timing of measurement, sex, aortic cross-clamp time, and cardioplegia type. The model at https://www.cardiomarker.com/ckmb can be validated, reproduced, refined, and applied to other biomarkers.