Background: The unplanned interruption during continuous renal replacement therapy (CRRT) can not only decrease the therapeutic outcomes of critically ill patients but also increase the healthcare costs and nurses’ workload. This study aimed to identify the independent factors of unplanned interruption during CRRT to construct a corresponding predictive model to guide clinical practice.
Methods: The clinical data of critically ill patients with CRRT admitted to our hospital (n = 331) were retrospectively analyzed. They were divided into a planned interruption group (n = 238) and an unplanned interruption group (n = 93). Their clinical data such as gender, the blood flow rate, and the filtration fraction were compared between the groups by single-factor analysis. Then, statistically significant variables were analyzed by multivariate analysis to construct a predictive model of unplanned interruption during CRRT, and a receiver operating characteristic (ROC) curve was plotted to validate the model’s predictive value.
Results: The single-factor analysis revealed that gender, chronic diseases, the blood purification catheter function, platelet before CRRT initiation, the blood flow rate, pre-dilution or dialysis, the filtration fraction, and the frequency of blood pump stops in both groups were statistically significant. These eight variables were included in the multivariate analysis showing that chronic diseases (OR: 3.063, 95% CI: 1.200-7.819), the blood purification catheter function (OR: 4.429, 95% CI: 1.270-15.451), the blood flow rate (OR: 0.928, 95% CI: 0.900-0.957), and the frequency of blood pump stops (OR: 1.339, 95% CI: 1.231-1.457) were the independent factors for the unplanned interruption of CRRT. These four factors were used to construct a predictive model and its ROC curve revealed that the area under the curve (AUROC) was 0.952 (95% CI: 0.930-0.973) with a cut-off point of “P = 0.0030” indicating a sensitivity of 0.882 and specificity of 0.899.
Conclusion: Chronic diseases, the blood purification catheter function, the blood flow rate, and the frequency of blood pump stops are independent factors of the unplanned interruption during CRRT. A predictive model constructed using these factors has a certain predictive value for the occurrence of unplanned interruption during CRRT.