Objective To explore the risk factors of unplanned withdrawal in critically ill patients undergoing continuous renal replacement therapy (CRRT) in the intensive care unit (ICU), and construct a nomogram model.

Methods The data of 98 critically ill patients were selected. LASSO analysis was used to screen the predictive factors of unplanned withdrawal during CRRT in critically ill patients in the ICU. Logistic regression analysis was used to screen the risk factors of unplanned weaning during CRRT. R software was used to construct a nomogram model of unplanned withdrawal during CRRT in critically ill patients in the ICU, and the nomogram model was internal validated.

Results A total of 276 CRRT treatments were performed on 98 patients, among which 129 were unplanned withdrawal (46.74%). The results of logistic regression analysis showed that calcium ions in the pipeline, dilution method, anticoagulation method and hematocrit were the risk factors of unplanned withdrawal from CRRT, while the blood flow velocity and prothrombin time were the protective factors of unplanned withdrawal from CRRT (P ＜ 0.05). The C-index of the nomogram model for unplanned withdrawal of CRRT was 0.935 (95%CI: 0.908–0.962). The predicted values of the correction curve are basically in agreement with the actual values. The area under the ROC curve of the model was 0.921 (95%CI: 0.892–0.950). The decision curve showed a threshold probability of 1% to 100%, and the nomogram predicted a relatively high net benefit value in unplanned withdrawal from CRRT of critically ill ICU patients.

Conclusions Pipeline ion calcium, dilution method, anticoagulation method, hematocrit, etc. are the risk factors for unplanned weaning in critically ill ICU patients during CRRT, while the blood flow velocity, prothrombin time, etc. are the protective factors for unplanned withdrawal. The chromatogram model constructed based on this has a relatively high predictive value.