Analysis of risk factors and development of a predictive model for IABP application in post-cardiac valve replacement patients

TL;DR

This study identifies risk factors and creates a predictive model to determine if patients will need an intra-aortic balloon pump after heart valve replacement surgery.

Contribution

The study introduces a predictive model using five preoperative variables to assess the need for IABP after heart valve replacement.

Findings

Five independent risk factors for IABP use were identified: age, stroke volume, cardiac output, cardiac index, and left ventricular end-systolic diameter.

The predictive model demonstrated excellent discrimination with an AUC of 0.946 in the training set and 0.933 in the validation set.

The model showed good calibration and clinical utility according to the Hosmer-Lemeshow test and decision curve analysis.

Abstract

To identify risk factors for intra-aortic balloon pump (IABP) requirement following heart valve replacement surgery (HVRS) and to develop a predictive model.

This retrospective cohort study analyzed 161 HVRS patients (October 2023 to January 2025) from the First Affiliated Hospital of Xinjiang Medical University. Patients were stratified into IABP (n = 58) and non-IABP (n = 103) groups. Independent risk factors were identified through univariate analysis, LASSO regression, and multivariate logistic regression. The cohort was randomly split into training and validation sets (7:3 ratio) for model development and internal validation. Model performance was assessed using receiver operating characteristic (ROC) curves, Hosmer-Lemeshow calibration, and decision curve analysis (DCA).

Significant differences were observed between groups across multiple parameters (all P < 0.05), including demographics, inflammatory markers, cardiac biomarkers, and echocardiographic indices. Multivariate analysis identified five independent risk factors for postoperative IABP use: age (OR = 1.138, 95% CI: 1.067–1.226), stroke volume (SV) (OR = 1.155, 95% CI: 1.060–1.296), cardiac output (CO) (OR = 5.700, 95% CI: 2.700–12.040), cardiac index (CI) (OR = 4.982, 95% CI: 2.879–10.119), and left ventricular end-systolic diameter (LVESD) (OR = 1.463, 95% CI: 1.157–1.849). The prediction model showed excellent discrimination in both the training set (AUC = 0.946, 95% CI: 0.910–0.982) and the validation set (AUC = 0.933, 95% CI: 0.876–0.990). Good calibration was indicated by Hosmer-Lemeshow test (P > 0.05 for both sets), and decision curve analysis confirmed the model's clinical utility.

A model incorporating five routinely available preoperative variables effectively stratifies the risk of requiring IABP after HVRS, demonstrating strong discriminatory performance and potential clinical applicability for preoperative risk assessment.

Figures