Background: Perioperative cardiac arrest (CA) is a devastating surgical complication, yet its epidemiology and risk factors across diverse surgical populations are not well-characterized. Methods: We conducted a population-based cohort study of 315,642 adult participants who underwent surgery within the prospective UK Biobank. The primary endpoint was incident perioperative CA. Survival probabilities were estimated using Kaplan-Meier analysis, with comparisons made between patients with and without CA. To identify robust predictors of CA-related mortality, we applied the Boruta algorithm, a machine learning-based feature selection method. Results: Over a median follow-up of 13.1 years, 2,311 (0.73%) patients experienced perioperative CA. CA was associated with a dramatically higher mortality rate compared to non-CA surgical controls (62.8% vs. 5.0%; P<0.001). While patients aged ?70 years had the highest overall risk of CA, the peak CA-attributable mortality occurred unexpectedly in middle-aged patients (45-49 years). A machine learning approach identified 12 robust predictors for CA-related death, which included not only established factors (age, prior myocardial infarction, diabetes) but also key markers of systemic inflammation (e.g., pan-immune-inflammation value, neutrophil-lymphocyte ratio) and hematologic dysfunction (e.g., anemia, elevated creatinine). Conclusions: CA carries a profoundly high risk of death across all surgical settings. The discovery that middle-aged patients experience peak attributable mortality, combined with the strong predictive power of inflammatory and hematologic biomarkers, challenges current risk-assessment paradigms. These findings underscore the need to incorporate these measurable biomarkers into preoperative screening to better identify and optimize high-risk individuals beyond traditional clinical factors.
Zheng et al. (Tue,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: