Construction of a predictive model for the risk of low cardiac output syndrome after heart valve replacement based on the decision tree method

WANG Ziwei; CHEN Si

doi:10.12102/j.issn.2095-8668.2026.05.008

Abstract：Objective:To construct a predictive model for the risk of low cardiac output syndrome(LCOS)after heart valve replacement(HVR)based on the decision tree method.Methods:A total of 260 patients who underwent HVR surgery in our hospital from June 1,2023 to June 30,2025 were selected as the study subjects and divided into a control group(n=213)and an LCOS group(n=47)according to the postoperative LCOS occurrence.Logistic regression analysis was used to analyze the influencing factors of LCOS after HVR surgery,a decision risk prediction model was established,and area under the receiver operating characteristic(ROC)curves were plotted to evaluate its predictive efficacy.Results:A total of 47 of the 260 patients who underwent HVR surgery developed LCOS postoperatively,with an incidence rate of 18.08%.Logistic regression analysis showed that age,body mass index(BMI),electrolyte imbalance,intraoperative blood loss≥20%,cardiopulmonary bypass(CPB)time,and aortic cross-clamp(ACC)time were risk factors for LCOS after HVR surgery(P<0.05).A decision tree model constructed based on these risk factors consisted of 3 layers,15 nodes,and 8 terminal nodes,with intraoperative blood loss as the first layer,indicating that intraoperative blood loss was the most important influencing factor.ROC curve results showed that the AUC of the decision tree model(0.901)was higher than that of the Logistic regression model(0.852),and the difference was statistically significant(Z=2.019,P=0.043).Conclusions:The occurrence of LCOS after HVR surgery is influenced by multiple factors.The risk prediction model for LCOS after HVR surgery constructed based on influencing factors has good predictive efficacy which can provide a basis for clinicians to identify high-risk patients early.