The influencing factors and establishment of early warning model for unplanned withdrawal of CRRT in elderly patients with acute heart failure

He Shuya; Ding Lingling

doi:10.3969/j.issn.1002-1949.2025.07.010

Abstract：Objective To explore the influencing factors of unplanned withdrawal of continuous renal replacement therapy(CRRT)in elderly patients with acute heart failure(AHF)and to construct an early warning model.Methods A total of 186 elderly patients with AHF who underwent CRRT and were admitted to the CCU of Cardiology Department of Xiangyang Central Hospital from April 2022 to June 2024 were prospectively included as the research subjects.According to the unplanned withdrawal of CRRT,the patients were respectively included in the occurrence group(n=42)and the non-occurrence group(n=144).A baseline data questionnaire for the patients was designed to compare the differences in baseline data between the two groups.Multivariate analysis was conducted to clarify the influencing factors of unplanned withdrawal of CRRT and to construct an early warning model.Then,the model was prospectively verified by using 31 elderly patients with AHF admitted to the hospital from August 2024 to March 2025.Results There were 42 out of 186 elderly AHF patients who experienced unplanned withdrawal of CRRT,with an incidence rate of 22.58％.Through univariate and multivariate Logistic regression analysis,the results showed that poor catheter outlet function,high hourly CRRT treatment volume,and high hematocrit(PCV)at the time of admission to CCU were risk factors for unplanned withdrawal of CRRT in elderly AHF patients(OR＞1,P＜0.05),while high activated partial thromboplastin time(APTT)at the time of admission to CCU was a protective factor(OR＜1,P＜0.05).A warning model for unplanned withdrawal of CRRT in elderly AHF patients was constructed based on influencing factors with good discrimination(C-index=0.936)and high accuracy.Decision curve showed that the net benefit rate was always greater than 0 within the threshold range of 0.04-0.99,indicating good clinical practicality.The prospective validation results showed that the sensitivity corresponding to this model was 80％and the specificity was 100％.Conclusions Poor catheter outlet function,high hourly CRRT treatment volume,high PCV upon admission to CCU,and low APTT are the influencing factors for unplanned withdrawal of CRRT in elderly AHF patients.The warning model established based on this has good accuracy and discrimination,and has good ability to distinguish and predict.