| 초록 |
Objectives: With growing interest in incremental hemodialysis (HD), residual kidney function (RKF) is a key determinant of dialysis prescription. Conventional RKF assessment requires interdialytic urine collection for more than 24 hours to calculate urea/creatinine clearance, but prolonged collections are often impractical. Cystatin C and beta-2 microglobulin (β2M), relatively larger middle molecules, are expected to show negligible clearance with low-flux membranes; therefore, when low-flux membranes are used, peri-dialytic concentrations of these markers may better reflect RKF and help estimate clearance-based RKF. Methods: Eighteen incident HD patients were enrolled, and all dialysis sessions were performed using low-flux membranes. After initiation of incremental HD, between days 8 and 10, blood samples were obtained immediately post-dialysis and before the subsequent session for blood urea nitrogen, creatinine, cystatin C, and β2M. Urine volume and collection time were recorded from immediately after the day-8 session to immediately before the day-10 session, and urea/creatinine clearance was calculated as the reference RKF (Figure 1). Results: Linear regression models predicting clearance-based RKF were developed using (1) pre-dialysis cystatin C (CysCpre) plus pre-dialysis β2M (β2Mpre), (2) averaged cystatin C (CysCavr) plus averaged β2M (β2Mavr) (mean of post- and pre-dialysis values), and (3) CysCavr alone. Performance was assessed by distributional similarity (mean±SD), Spearman’s rho between measured and predicted RKF, root mean square error (RMSE), and P40 accuracy (predictions within ±40% of measured RKF). The CysCavr-only model showed the best overall performance (rho=0.701, RMSE=2.18, P40=44.4%) (Table 1). Conclusion: In incident HD patients treated with low-flux membranes, peri-dialytic cystatin C can estimate clearance-based RKF and may provide a practical blood-based alternative to prolonged urine collection. Larger prospective datasets are needed to refine and validate these models. |