Time-Dependent Changes in Extracellular-to-Intracellular Water Ratios and Risk of All-Cause Mortality in Peritoneal Dialysis Patients

Rationale & Objective: Bioelectrical impedance spectroscopy (BIS) is a valuable tool for assessing body composition in patients on peritoneal dialysis (PD), as body composition can significantly change during long-term PD treatment. We aimed to clarify the relationship between time-dependent changes in the extracellular-to-intracellular water ratio (EIR) and all-cause mortality in patients undergoing long-term PD. Study Design: A single-center, retrospective cohort study. Setting & Participants: This study enrolled adult PD patients who underwent body composition measurements from our bioelectrical impedance spectroscopy-aided fluid management program from May 1, 2012, to April 30, 2023. Patients were followed up until death, withdrawal from PD, or the end of the study. Exposure: EIR and its covariates. Outcomes: All-cause mortality. Analytical Approach: Time-dependent Cox proportional hazards, restricted cubic splines, and propensity score matching analyzed risk factors and mortality. Results: Among the 468 PD patients, the median follow-up period was 2.0 (interquartile range [IQR] 0.6-4.3) years. Three distinct temporal EIR trajectories are found after grouping baseline EIR (bEIR) into quartiles (Pinteraction < 0.001). Additionally, the optimal cutoff value of bEIR for predicting survival was 0.87. A high bEIR (≥0.87) was associated with a greater risk of mortality than a low bEIR (<0.87, crude hazard ratio, 6.095; P < 0.001). The multivariable time-dependent Cox model showed that the risk of mortality increased by nearly 50% for every 0.1 increase in the EIR (P < 0.001), without significant nonlinearity (P = 0.07). After propensity score matching, the EIR remained an important predictor of the risk of mortality. Limitations: The study had a retrospective design. Conclusions: The EIR, regarded as either a time-independent or time-dependent variable, is a crucial predictor of mortality in PD patients under regular monitoring of fluid status. Our results also suggest that an EIR of ≥0.87 would be a useful target for predicting the long-term mortality of PD patients in clinical practice. Plain-language Summary: Fluid overload-induced cardiovascular disease has long been one of the leading causes of mortality in dialysis patients. In this study, we used time-dependent Cox proportional hazards regression analysis and found that the extracellular-to-intracellular water ratio (EIR), as measured by bioelectrical impedance spectroscopy, can predict the mortality risk in peritoneal dialysis patients. Over the course of dialysis, the bEIR quartile groups exhibited 3 distinct trajectories. Except for the fourth bEIR quartile (Q4) group, which showed an initial decrease in EIR due to intervention for fluid overload, the other two groups demonstrated an upward trend, with the first bEIR quartile (Q1) group showing the greatest increase.