Development and validation of a risk prediction model for hypoproteinemia after adult cardiac valve surgery: implications for clinical care.

Objective: To construct and validate a risk prediction model for hypoproteinemia in adults following cardiac valve surgery with cardiopulmonary bypass (CPB), providing medical staff with an effective tool for early identification and intervention.

Methods: This retrospective cohort study analyzed clinical data from 259 patients who underwent CPB-assisted heart valve surgery at the Department of Cardiovascular Surgery, First Affiliated Hospital of China University of Science and Technology, between January and December 2023. Patients were divided into two groups based on whether their serum albumin levels fell below 35 g/L within 48 hours postoperatively: the hypoproteinemia group ( = 144) and the non-hypoproteinemia group ( = 115). Least absolute shrinkage and selection operator (LASSO) regression was used to identify candidate predictors, followed by multivariate logistic regression to determine independent risk factors.

Results: Among the 259 patients, 144 developed hypoproteinemia, yielding an incidence rate of 55.60%. LASSO regression identified nine variables associated with hypoproteinemia, and multivariate logistic regression confirmed eight independent predictors. Hypertension, chest infection, frailty, and preoperative heart failure were identified as independent risk factors (OR > 1,  < 0.05), while higher body mass index (BMI), red blood cell (RBC) count at admission, platelet count at admission, and albumin level at admission were protective factors (OR < 1,  < 0.05). The predictive model constructed using the nine LASSO-selected variables demonstrated good discrimination, with an area under the ROC curve (AUC) of 0.823 (95% CI [0.774-0.873]). The Hosmer-Lemeshow test showed no significant difference between predicted and observed outcomes ( = 0.737), indicating good model calibration.

Conclusion: The incidence of postoperative hypoproteinemia in this cohort was 55.60%. The developed nomogram model, based on key clinical predictors, demonstrated strong calibration and discrimination, offering a practical tool for identifying patients at high risk of hypoproteinemia following valve surgery.