Predictive value of electrocardiographic, echocardiographic and hematological parameters for predicting new‑onset atrial fibrillation in patients with ST‑segment elevation myocardial infarction after percutaneous coronary intervention

Aim. To assessment predictive potential of electrocardiographic, echocardiographic and hematological parameters for predicting new‑onset atrial fibrillation (AF) in patients with ST‑segment elevation myocardial infarction after percutaneous coronary intervention (PCI), as well as to develop novel prognostic models based on machine learning. Material and methods. This single‑center prospective study included 733 patients with ST‑segment elevation myocardial infarction. Two following groups were identified: the first — 57 (7,8%) patients with new‑onset postoperative AF after PCI, and the second — 676 (92,2%) patients without cardiac arrhythmia. To predict AF, univariate and multivariate logistic regression, decision trees, CatBoost gradient boosting were used. Results. Comparative analysis of electrocardiography, echocardiography, hematological and clinical data demonstrated that following parameters has the highest predictive potential: neutrophil‑to‑eosinophil ratio (NER) >48,7 (odds ratio (OR) 7,1), Killip class >2 acute heart failure (OR 4,44), erythrocyte sedimentation rate (ESR) >36 mm/h (OR 4) and systemic inflammatory response index (SIRI) >5 (OR 3,8). The best prognostic model of new‑onset AF after PCI (AUC=0,806) included 9 following categorical predictors: NER >48,7 conventional units, SIRI >5 conventional units, erythrocyte sedimentation rate >36 mm/h, PQ >200 ms, 600 ms< RR >1200 ms, pulmonary artery systolic pressure >33,5 mm Hg, age >66 years, TIMI <3 and Killip class >2 acute heart failure. Conclusion. Improving the accuracy of predicting new‑onset AF after PCI can be achieved by expanding the range of potential predictors and using modern explainable artificial intelligence technologies.