Development and validation of a predictive model for postoperative hepatic dysfunction in Stanford type A aortic dissection

Abstract

Abstract To investigate the risk factors for postoperative hepatic dysfunction (HD) in patients undergoing acute Stanford type A aortic dissection (ATAAD) surgery and to develop an individualized prediction model. We retrospectively analyzed cardiac surgery patients with ATAAD treated at our hospital from January 2020 to March 2024, dividing them into 7:3 training and validation cohorts and grouping them into HD and non-HD categories based on postoperative liver function. Least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression were used to identify independent predictive factors for postoperative HD, which formed the basis of a nomogram prediction model. We assessed model accuracy, calibration and clinical utility using C-statistics, calibration plots and decision curve analysis (DCA) curves. Internal validation with 1000 Bootstrap resamples was performed to reduce overfitting bias. LASSO and multivariate logistic regression identified key risk factors for HD in ATAAD patients, including chronic kidney disease, preoperative creatinine, international normalized ratio (INR), red blood cell (RBC) transfusion volume, peak intraoperative lactate, aortic cross-clamping time greater than 99 min, and reoperation. Based on these factors, a nomogram prediction model was successfully developed. The Hosmer–Leme show test yielded a p value of 0.952, indicating a good model fit. The area under the curve (AUC) values in the training and validation cohorts were 0.856 (95% CI 0.777–0.936) and 0.958 (95% CI 0.915–1) respectively, indicating good discriminatory power. The calibration curve shows that the bias corrected line is close to the ideal line. The DCA curve indicates that the use of the nomogram provides greater net clinical benefit. The AUC values before and after Bootstrap validation were 0.860 (95% CI 0.795–0.924) and 0.858 (95% CI 0.795–0.924), respectively, reflecting stable model performance and minimal risk of overfitting. The internally validated prognostic nomogram demonstrates excellent discriminative power, calibration, and clinical utility for predicting the risk of HD in patients who have undergone ATAAD surgery. This allows for an individualized evaluation and the optimization of clinical outcomes.

Keywords

• Stanford type A aortic dissection
• Hepatic dysfunction
• Predictive model