Abstract: Objective  To investigate the risk factors for the development of dialysis disequilibrium syndrome (DDS) in patients on maintenance hemodialysis (MHD) and to construct a predictive model. Methods  A total of 321 patients with MHD in our hospital from August 2019 to August 2022 were selected and patients were randomly divided into the training group (n=225) and the internal verification group (n=96) according to the proportion of 7:3. The two groups were divided into the DDS subgroup and the non-DDS subgroup according to the occurrence of DDS. The incidence of imbalance syndrome, demographic characteristics and biochemical indicators of the two groups were analyzed. The logistic regression model and random forest model were constructed based on the data of the training group. Then Parallel internal and external validation was performed in the groups.  Results Univariate analysis showed that in the training and internal validation groups, age (t=32.154, 24.618, both P＜0.001), number of dialysis sessions per week (t=10.632, 8.211, both P＜0.001), epilepsy (χ2=4.647, 7.248, P=0.031, 0.007), hemoglobin (t= 21.366, 15.476, all P＜0.001), cognitive impairment (χ2=4.644, 5.403, P=0.031, 0.020), urea nitrogen (t=21.284, 13.058, all P＜0.001), and albumin (t=13.094, 9.018, all P＜0.001) between the DDS subgroups and the non-DDS subgroup were significant differences (P＜0.05). Logistic regression analysis showed that the number of dialysis sessions per week (OR=6.360, 95% CI: 1.968 to 20.554, P＜0.001), cognitive impairment (OR=8.404, 95% CI: 2.446 to 28.877, P＜0.001), hemoglobin (OR=4.889, 95% CI: 1.436 to 16.645, P＜0.001), albumin (OR=0.596, 95% CI: 0.447 to 0.794, P＜0.001), and urea nitrogen (OR=4.429, 95% CI: 1.879 to 10.441, P＜0.001) were factors influencing the occurrence of DDS in patients (P＜0.05). The top 5 influencing factors for the occurrence of DDS were obtained in the following order: urea nitrogen, cognitive impairment, hemoglobin, albumin, and number of dialysis sessions per week. Based on the above factors, logistic regression models and random forest models for the occurrence of DDS in patients were constructed, and internal validation showed that there was no significant difference between the two models in predicting the AUC of DDS in patients, and external validation showed that there was no significant difference between the two models and the actual results.  Conclusion  The influence of DDS in MHD patients is due to urea nitrogen, cognitive impairment, hemoglobin, albumin, and weekly dialysis times. The prediction model built based on the above factors is reliable and provides a certain reference for clinical treatment identification of DDS.

Key words: Maintenance hemodialysis, Imbalance syndrome, High risk factors