Development and validation of a new predictive model for in-hospital postoperative major adverse cardiovascular and cerebrovascular events after general anesthesia in nonagenarians undergoing non-cardiac surgery.

Background: Major adverse cardiac and cerebrovascular events (MACCE) following noncardiac surgery are the main cause of perioperative mortality. However, there are few evidence-based prediction models available for predicting the risk of MACCE. We aimed to analyze the risk factors of MACCE in patients aged 90 and older and to construct a prediction model, ultimately leading to the development of a nomogram. Methods: This review study included clinical data from 872 patients aged 90 and older who underwent non-cardiac surgery under general anesthesia between 2015 and 2024. The outcome of interest was in-hospital postoperative MACCE. Logistic regression was employed to identify risk factors and to establish a nomogram for predicting the risk of MACCE. Calibration curves, C-index, and decision curves were used to evaluate the predictive model. An external cohort was used to compare the performance between our model and the widely used revised cardiac risk index (RCRI) score. Results: 112 patients (12.84%) experienced in-hospital MACCE. The final model identified four predictors, including emergency surgery, neutrophil/lymphocyte ratio (NLR) ≥ 11.2, D-dimer ≥ 3.6 mg/L, and postoperative admission to the ICU. The nomogram demonstrated strong discriminative ability with a C statistic of 0.853 and maintained its performance during 10-fold cross-validation with a C statistic of 0.784. Compared to the RCRI score, our predictive model performed better in the validation test (C statistic = 0.853 vs. 0.693). Conclusions: The predictors including NLR, D-dimer, emergency surgery, postoperative 24-hour ICU admission could better predict MACCE than RCRI score in patients greater than 90 years old undergoing non-cardiac surgery undergoing general anesthesia.