Development and Validation of a Model to Predict Hypothermia in Patients Undergoing Continuous Renal Replacement Therapy: A Retrospective Study

Background: Hypothermia is a significant complication of Continuous Renal Replacement Therapy (CRRT) and is associated with severe adverse outcomes. This study aimed to develop a hypothermia risk prediction model integrating environmental parameters, treatment variables, and biomarkers. Methods: A total of 2791 patients undergoing CRRT at the Xinqiao Hospital, Army Medical University between January 2022 and February 2025were enrolled. Patients were randomly divided into a training set (n=1955) and a test set (n=836) in a ratio of 7:3. LASSO regression followed by multivariate logistic regression analysis were used to select risk factors for hypothermia during CRRT. A nomogram based on independent risk factors was constructed and validated by the receiver operating characteristic curve (AUC), calibration curve and decision curve analysis (DCA). Results: In the training set, a total of 163 (8.3%) patients develop hypothermia. LASSO regression multivariate logistic regression analysisrevealed that baseline body temperature (OR=9.81), lactate (OR=1.18), treatment duration (OR=1.02), and C-reactive Protein (OR=1.00) were independent risk factors for hypothermia. The model nomogram based on these factors achieved an AUC of 0.858 for predicting hypothermia in the test set. The calibration curve and DCA showed the reliability of the nomogram. Conclusion: The clinical nomogram developed in this study is the first to incorporate multi-dimensional indicators, providing an accurate predictive tool for personalized thermal management in patients undergoing CRRT. This instrument is significant for reducing the incidence of hypothermia and optimizing the allocation of healthcare resources.