Preservation of Residual Kidney Function in the Management of Patients on Incident Hemodialysis

The kidneys regulate extracellular volume with limited and temporary deviations in response to volume depletion or excess. A decline in kidney function decreases the body's ability to control deviations of fluid status. Historically, residual kidney function (RKF) was thought to decline rapidly in most patients starting chronic hemodialysis, becoming negligible within the first few months of treatment.¹ More recent studies, however, indicated that RKF may be better preserved with biocompatible dialysis membranes, ultrapure dialysis water, and hemodiafiltration. The rate of RKF decline is affected by the etiology of kidney failure, comorbidities, and other patient-specific factors. More frequent hemodialysis, higher ultrafiltration rate, and intradialytic hypotension are also associated with RKF decline.¹ Loss of RKF significantly increases the risk of fluid overload² and is strongly associated with mortality in patients on hemodialysis.³

The importance of RKF is widely acknowledged in peritoneal dialysis. Over the past decade, the significance of RKF in hemodialysis has gained increased attention.⁴ Patients on dialysis face two challenges related to fluid status: on the one hand, the risk of fluid overload due to impaired kidney function, and on the other hand, the danger of overzealous fluid removal by ultrafiltration. Both conditions have been associated with adverse outcomes.⁵ In this situation, preservation of RKF can have a mitigating effect. Of note, urine volume increases during the interdialytic period, likely driven by extracellular volume expansion.⁶ This observation indicates that avoidance of fluid depletion may aid RKF preservation. Systematic quantitation of RKF may indicate a decline and thus instigate countermeasures.

The Bioimpedance Spectroscopy To maintain Renal Output (BISTRO) study is a pragmatic, multicenter, randomized controlled trial that compared the rate of GFR decline between two fluid management strategies in patients on hemodialysis.⁷ All participants, who were undergoing incident hemodialysis, had >500 ml/d of urine output or >3 ml/min per 1.73 m² of GFR, estimated as the average of creatinine and urea clearances.⁷ One group was treated according to the fluid management proforma aiming to avoid excessive volume depletion at the discretion of the treating providers. In the other group, target weights were adjusted with additional assessments by bioimpedance spectroscopy. The authors reported no between-group differences in ultrafiltration volume, target weight or posthemodialysis weight, and GFR decline during the 2-year trial period. Notably, the overall GFR decline in the BISTRO study was markedly slower than that in previous reports.¹

In this issue of Kidney360, Belcher et al. report a retrospective analysis of the BISTRO trial,⁸ where the study team lifted the randomized group allocation and analyzed the trial data in a pooled fashion. The systematic data collection of the study allowed for more robust analyses compared with other real-world data studies on the basis of routinely collected clinical assessments; for example, the BISTRO study collected 24-hour urine samples every 3 months throughout the 2-year study period. The study showed an association of higher GFR at dialysis initiation with better survival. The authors also found a nonsignificant increased mortality risk with a greater GFR decline. Then, the authors used joint modeling and reported an association of greater RKF (i.e., higher GFR or larger urine volume) at any time point with better survival.

This retrospective study by Belcher et al. benefitted from the prospective design of the main study with a granular collection of GFR data and a follow-up period up to 50 months. Furthermore, the authors used three different statistical models to analyze the association of RKF with survival to corroborate the robustness of the findings. The first approach was the Cox proportional hazards model using the exposure variable of interest only at baseline. However, this model did not account for longitudinal changes in GFR. The second approach was a time-varying Cox model to address this limitation. However, the time-varying Cox model used the last-observation-carried-forward approach for missing exposure measurements and did not account for informative dropout. These are important limitations given the association between RKF decline and mortality. The third approach was joint modeling, a statistical method that allows for the simultaneous analysis of longitudinal data and time-to-event data. Joint modeling typically involves two submodels, one for longitudinal data (e.g., the mixed-effects model for repeated GFR measurements) and one for time-to-event data (e.g., the Cox model for all-cause death). These submodels are linked by shared random effects or latent variables, which capture the underlying individual-specific characteristics influencing both processes. Joint modeling reduces bias in parameter estimates by accounting for the association between longitudinal and time-to-event data. Additional advantages of joint modeling include improved accounting for measurement errors in longitudinal datasets and better handling of missing data, including informative dropout. The approach provides a comprehensive and more accurate analysis of complex longitudinal and survival data and is a powerful tool for understanding the interplay between repeated measurements and time-to-event outcomes. In the study by Belcher et al., RKF measurements were regularly obtained only during the 2-year trial period. Therefore, the rate of RKF decline after 2 years was assumed to remain the same as the rate observed during the first 2 years. In addition, the observed association between higher baseline GFR and better survival is well known to be subject to lead time bias, which could have persisted even with joint modeling.

The exact reasons for the survival benefit of maintained RKF are not entirely understood, and several potential causal pathways seem reasonable.⁴ Importantly, RKF contributes to the clearance of middle-sized molecules and protein-bound solutes. Preserved RKF allows for more flexible fluid and dietary management, alleviating dietary restrictions and possibly improved nutritional status. Preserved RKF also improves fluid and BP control and thus may reduce interdialytic weight gain and the risk of intradialytic hypotension. Patients with preserved RKF report better quality of life and often require lower doses of erythropoiesis-stimulating agents for anemia management. Preserved RKF is also associated with better control of hyperphosphatemia and lower pill burden from phosphorus binders.

Supported by previous findings, the results by Belcher and the BISTRO team⁸ underscore the need to develop and implement interventions to preserve RKF. However, it is not entirely clear how to operationalize RKF-preserving strategies. In the BISTRO trial, the addition of bioimpedance to the fluid volume management protocol did not influence target weight, RKF preservation, or survival. In the main study, the aim of bioimpedance was to avoid excessive fluid depletion. It seems that the BISTRO investigators did an excellent job in avoiding overzealous ultrafiltration, irrespective of the use of bioimpedance. However, we hesitate to generalize the findings by a dedicated specifically trained study team to the larger population of patients on hemodialysis for several reasons. First, interdialytic weight gain does not report fluid overload,⁹ so a purely weight-based approach to fluid management can be misleading. Second, while clinical assessment of fluid status is a fundamentally important skill, its sensitivity to determine fluid overload or depletion is poor¹⁰ and often fails to detect subtle, but clinically important, volume disturbances. Even minor, clinically hard-to-detect derangements of fluid status are associated with increased mortality.⁵ Hence, we suggest the implementation of protocols including an objective assessment of fluid status, such as bioimpedance. In addition to its capability to assess fluid status, bioimpedance also provides body composition information. Given the known effect of RKF on nutrition and metabolism, the association between dynamic changes of body composition and RKF are a clinically relevant, yet understudied, research topic.

In summary, the results of this retrospective analysis of the BISTRO trial indicate a protective effect of preserved GFR on survival. To make informed decisions on fluid management for the individual patient, tools allowing objective assessments will be important. Studies on how to implement such tools as bioimpedance spectroscopy most effectively for the purpose of RKF preservation are imperative.

Supplementary Material

kidney360-6-012-s001.pdf ^{(1.3MB, pdf)}

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/KN9/A786.

Funding

None.

Author Contributions

Conceptualization: Yoshitsugu Obi, Jochen G. Raimann.

Supervision: Yoshitsugu Obi.

Writing – original draft: Yoshitsugu Obi, Jochen G. Raimann.

Writing – review & editing: Yoshitsugu Obi, Jochen G. Raimann.