Abstract: Objective  To analyze the relationship between the expression of serum Krüppel-like factor 4 (KLF4), thioredoxin-interacting protein (TXNIP) and vascular calcification (VC) in maintenance hemodialysis (MHD) patients.  Methods  Patients who received MHD at Baoding First Hospital from March 2019 to March 2024 were selected as the study subjects. They were divided into a calcification group and a non-calcification group based on the presence or absence of VC. Data on gender, age, Body Mass Index (BMI), dialysis vintage, smoking, alcohol consumption and primary disease were collected. Serum levels of albumin, blood phosphorus, blood calcium, blood creatinine, urea nitrogen and fasting blood glucose were measured using a DXI-800 automatic biochemical analyzer. Hemoglobin was detected by an automated hematology analyzer, intact parathyroid hormone (iPTH) was detected by electrochemiluminescence. The urea removal index (Kt/V) and glomerular filtration rate were calculated. The levels of KLF4 and TXNIP were determined by enzyme-linked immunosorbent assay (ELISA). The Pearson correlation method was used for correlation analysis. Multivariate Logistic regression and risk analysis were performed to evaluate the effects of KLF4 and TXNIP on VC. The predictive value was evaluated using the receiver operating characteristic (ROC) curve, and the clinical application value of the model was evaluated by decision curve analysis (DCA).  Results  The KLF4 and TXNIP levels in the calcification group were significantly higher than those in the non-calcification group (t=7.312, 8.419; both P＜0.001). Pearson analysis showed that KLF4 was positively correlated with age, dialysis vintage, fasting blood glucose, blood phosphorus, and iPTH (r=0.425, 0.328, 0.396, 0.550, 0.386), and negatively correlated with hemoglobin and albumin (r=-0.397, -0.315). TXNIP was positively correlated with age, dialysis vintage, fasting blood glucose, blood phosphorus, and iPTH (r=0.407, 0.501, 0.421, 0.385, 0.314), and negatively correlated with hemoglobin and albumin (r=-0.437, -0.418), all P＜0.001 Multivariate Logistic analysis showed that KLF4 [OR (95% CI) =2.591 (1.434～4.683), P=0.002] and TXNIP [OR (95% CI) =1.926 (1.505～2.466), P＜0.001] were both independent influencing factors for VC (P＜0.05). Risk analysis showed that the risk of VC in patients with high levels of serum KLF4 and TXNIP was 1.483 (1.149～1.915) times and 1.432 (1.125～1.822) times that of patients with low levels, respectively. The ROC curve showed that the AUC of the combined KLF4 and TXNIP for VC prediction was 0.930 (Z combined VS KLF4=2.232, P=0.026) (Z combined VS TXNIP=2.432, P=0.015). The DCA indicated that when the high-risk threshold probability ranges from 0.26 and 0.95, the combined model has high clinical application value in predicting the occurrence of VC. Conclusion  Patients with VC after MHD exhibit high expression of KLF4 and TXNIP. Both factors are related to the occurrence of VC in MHD patients. The combined of KLF4 and TXNIP has high prediction value of VC in MHD patients.

Key words: Maintenance hemodialysis, Vascular calcification, Krüppel-like factor 4, Thioredoxin interacting protein