The Concept of Perioperative Cardiac Risk: Literature Review

Highlights This review introduces for the first time an integrative conceptual model of high perioperative cardiac risk, combining three key components: surgical risk, patient vulnerability, and dynamic monitoring, enabling a systematic approach to patient stratification prior to noncardiac surgery. It demonstrates a critical gap between the high predictive accuracy of modern machine learning models (AUROC > 0.90) and the lack of their clinical validation in real-world practice, particularly in the Russian population, highlighting the need for a national risk calculator. For the first time in Russian literature, the role of MINS (myocardial injury after noncardiac surgery) as a major driver of postoperative mortality (contributing 3.9% to overall mortality) is analyzed in detail, justifying the shift from reactive to preventive perioperative cardiac risk management based on routine monitoring of high-sensitivity troponins and natriuretic peptides. Abstract This review focuses on contemporary approaches to the assessment and management of perioperative cardiac risk in patients undergoing non-cardiac surgery. The topic's relevance stems from the high frequency of major adverse cardiovascular events (MACE), a leading cause of perioperative mortality. The work analyzes the paradigm shift from the question “is surgery feasible?” to the strategy “how to improve outcomes?” which includes pre-operative optimization (prehabilitation), active monitoring, and a multidisciplinary approach. Based on a literature search, the epidemiology and key risk factors are reviewed, with detailed classifications and definitions of complications (MACE, myocardial injury after noncardiac surgery (MINS), intraoperative critical incidents). Particular attention is paid to the role of biomarkers (NT-proBNP/BNP for prediction, high-sensitivity troponin for diagnosing MINS) and hemodynamic control in risk stratification and early detection of myocardial injury. A comparative analysis of prognostic tools is conducted: from traditional clinical scales (Lee index, NSQIP) to modern machine learning models demonstrating high accuracy. Problems with validation, clinical interpretability, and the integration of these tools into routine practice are noted. Contemporary recommendations for risk minimization, based on staged personalized assessment, patient optimization, and active postoperative monitoring, are summarized. The conclusion emphasizes the need to develop integrative, transparent, and clinically applicable algorithms that combine data from scales, biomarkers, and dynamic monitoring to transition from reactive to preventive management of perioperative cardiac risk.