Kidney assistance therapy, a term that more accurately reflects the limited therapeutic capabilities of dialytic modalities conventionally referred to as kidney replacement therapy, sustains life by removing retained uremic toxins and excess fluid, while operating within a dynamic biologic milieu shaped by comorbidities, intercurrent illness, diet, medications, and the evolving course of residual kidney function [1]. Neither uremia nor its dialytic treatment can therefore be comprehensively characterized by a single clearance measure. At the molecular level, retained solutes differ in size, protein binding, tissue distribution, generation, and biological activity. At the patient level, nutritional reserve, inflammatory burden, cardiovascular physiology, and tolerance of fluid removal may vary across individuals and over time, as we have recently emphasized [2]. Finally, at the technical level, advances in dialysis membranes, convective therapies, and transport modeling are expanding what kidney assistance therapies can accomplish. Science and bedside practice need to connect these dimensions and translate dialysis treatment capabilities into benefits that matter to patients.
This Special Issue of Toxins, “Contemporary and Emerging Modalities of Kidney Assistance Therapy for Patients with Kidney Dysfunction Requiring Dialysis”, brings together three original investigations and six reviews spanning uremic toxin biology, hemodialysis, hemodiafiltration, and peritoneal dialysis. These contributions move beyond asking whether greater removal is better to examining which targets are biologically relevant, what governs their removal, and how therapy might be adapted to individual patients.
Cozzolino et al. [3] trace the evolution of uremic toxin classification from the 2003 European Uremic Toxin Work Group categories of small water-soluble compounds, middle molecules, and protein-bound solutes toward a multidimensional framework incorporating biological effects, associations with clinical outcomes, and the capabilities of different removal technologies. The authors caution that while molecular size and protein binding help predict dialyzability, they do not, by themselves, establish pathogenicity. The review also considers the potential of multi-omics, advanced adsorption and membrane technologies, and computational approaches to connect candidate toxins with therapeutic targets. Such classification can guide investigation and technology selection, but a retained solute must still be shown to contribute meaningfully to disease, and its modification to improve outcomes, before it can be treated as an actionable target.
Two original investigations illustrate why biological context matters. In a prospective cohort of 448 patients receiving hemodialysis, Bui et al. [4] found that a higher baseline predialysis leptin-to-adiponectin ratio was associated with lower all-cause mortality (adjusted hazard ratio for the highest versus lowest tertile, 0.18; 95% confidence interval, 0.07–0.46). The association persisted after adjustment for demographic, clinical, and laboratory covariates, including serum albumin and interleukin-6; nonetheless, this observation does not establish causality or that changing the ratio would improve survival. These observations suggest that, rather than identifying either adipokine as inherently beneficial or harmful, the association may reflect differences in nutritional status, body composition, and inflammation that also relate to survival. Espi et al. [5] combined observations in patients receiving hemodialysis with cell-culture experiments and murine CKD models induced by an adenine-enriched diet or five-sixths nephrectomy; an additional adenine-fed group was immunized during active inflammation. Plasma measurements of total and free indoxyl sulfate, p-cresyl sulfate, p-cresyl glucuronide, indole-3-acetic acid, and hippuric acid, as well as 3-carboxy-4-methyl-5-propyl-2-furanpropionic acid and uric acid, showed no significant associations with the assessed vaccine responses, and indoxyl sulfate did not directly impair T–B cell cooperation in their experimental system. C-reactive protein was the only variable independently associated with vaccine-response status in multivariable analysis; each 1 mg/L higher C-reactive protein concentration was associated with 14% lower odds of being classified as a vaccine responder (adjusted odds ratio, 0.86; 95% confidence interval, 0.76–0.96). In the adenine-induced CKD model, immunization during the inflammatory phase was accompanied by higher interleukin-6 levels, impaired antibody responses, and smaller draining lymph nodes. Together, these studies caution that an association between a circulating marker and an outcome should not be taken as evidence that removing or modifying that marker will improve the outcome.
Hemodiafiltration also warrants a more nuanced perspective on treatment performance. Lang et al. [6] synthesize the rationale, clinical evidence, safety, and implementation requirements for high-volume hemodiafiltration. Adding substantial convection to diffusion improves the removal of many middle molecules relative to conventional hemodialysis; for β2-microglobulin, the review cites approximately 10–20% higher reduction ratios with hemodiafiltration than with high-flux hemodialysis. Protein-bound uremic toxins, such as indoxyl sulfate and p-cresyl sulfate, illustrate, however, an important limitation of enhanced convective clearance. Only the unbound fraction can cross the dialyzer membrane; as this fraction is removed, additional toxin dissociates from albumin, but strong protein binding limits the amount available for clearance. Furthermore, toxins distributed outside the vascular compartment may redistribute into plasma during and after dialysis, contributing to post-dialysis rebound. Consequently, although high-volume hemodiafiltration can enhance the removal of some protein-bound toxins compared with high-flux hemodialysis, the incremental benefit is often modest or variable, and true clinical significance remains to be determined. The review also emphasizes that safe delivery of high-volume hemodiafiltration requires vascular access capable of sustaining the blood-flow rates needed to achieve the prescribed convection volume; HDF-capable equipment; validated production of sterile, nonpyrogenic substitution fluid; trained staff; and ongoing microbiological surveillance and equipment maintenance.
Two complementary reviews by Daugirdas discuss the distinction between removal and biological exposure with hemodiafiltration. The first review applies phosphate kinetic modeling and estimates approximately 15–20% greater dialytic phosphate clearance with high-volume hemodiafiltration in a modeled four-hour session with blood flow of 350 mL/min, dialysate flow of 500 mL/min, and substitution-fluid infusion of 100 mL/min (24 L per session) [7]. The Author also calculates how modest differences in dietary phosphate intake and absorption, phosphate-binder dose, or residual kidney clearance can obscure this advantage when predialysis serum phosphate is the endpoint [7]. The bottom line is that treatment comparisons must account for solute generation or entry, distribution, and nondialytic elimination, rather than equating a blood concentration with delivered dialytic clearance. The second review asks whether differences in outcomes between hemodiafiltration and hemodialysis may also reflect mechanisms other than enhanced removal of larger molecules, including small-solute clearance, intradialytic hemodynamics, ultrapure dialysis fluid, and reduced exposure to infused saline [8]. These mechanisms are all biologically plausible and potentially complementary, but their respective contributions to clinical outcomes are not established.
Zhou et al. [9] examine intermittent infusion hemodiafiltration, a distinct strategy that alternates brief backfiltration of ultrapure dialysate with subsequent ultrafiltration. A commonly described regimen backfilters approximately 200 mL of ultrapure dialysate across the dialyzer membrane into the blood compartment of the extracorporeal circuit every 30 min (about 150 mL/min during each infusion), for a total of about 1.4 L during a four-hour session. This cyclic pattern is intended to transiently support intravascular volume and may help sustain tissue perfusion while preserving the net fluid-removal goal. The review brings together physiological studies and relatively small clinical comparisons reporting fewer hypotension-related interventions in selected patients, as well as observations on middle-molecule removal and albumin loss. However, intermittent infusion hemodiafiltration should not be equated with online hemodiafiltration at either higher or lower convection volumes: the former uses brief backfiltration pulses, whereas the latter uses sustained convective removal with continuous substitution. Comparative long-term clinical effects remain to be established. The contribution broadens the discussion from how much fluid is removed to how fluid removal and replacement are distributed within a treatment.
Individualization should also be considered at the interface between treatment technology and patient phenotype. Murea et al. [10] propose tailoring dialyzer membrane selection with consideration of membrane permeability, pore selectivity, adsorption, and biocompatibility alongside residual kidney function, inflammatory and nutritional status, protein-bound toxin burden, and hemodynamic vulnerability. Building on principles already incorporated into clinical practice in countries such as Japan, the framework balances solute removal, preservation of essential proteins, and membrane availability, with reassessment as the patient’s phenotype evolves. For example, preserved residual kidney function may temper the need to intensify middle-molecule removal, whereas a patient with substantial inflammatory or amyloidogenic burden may prompt consideration of a membrane with different clearance characteristics. Medium-cut-off and more permeable membranes can expand middle-molecule removal, but albumin leakage may be consequential in patients with limited nutritional reserve; further on, filters with significant adsorptive properties offer a different potential route for solutes poorly addressed by diffusion and convection alone. This phenotype-based framework offers a conceptual framework for individualizing dialyzer selection, keeping in mind that prospective studies are needed to determine whether systematically applying this approach improves patient outcomes compared with conventional membrane selection.
A route to personalization is also explored in peritoneal dialysis. Swapnasrita et al. [11] compared six computational models using experimental data from pigs and clinical data from patients to predict solute concentrations and fluid transport occurring during dialysis. With inputs consisting of measurements of blood and dialysate solute concentrations, instilled and drained fluid volumes, and dialysate composition, the six models estimated solute transport across the peritoneal membrane and predicted changes in dialysate solute concentrations over a four-hour dwell. The three-pore model additionally estimated intraperitoneal fluid volume, ultrafiltration, and lymphatic fluid absorption. Among the six models, the three-pore and Waniewski models generally provided the closest agreement between predicted and measured solute concentrations. The three-pore model also accurately reproduced measured ultrafiltration but required substantially more computational time, whereas the faster Waniewski model sometimes generated physiologically implausible estimates of sodium transport. This work provides a promising foundation for future studies that would develop and refine patient-specific models of peritoneal transport that could eventually guide individualized dialysis prescriptions by predicting solute clearance and ultrafiltration under different treatment conditions.
Collectively, these nine contributions, alongside the broader nephrology literature, help shape an emerging research agenda for kidney assistance therapy. Candidate uremic toxins should be evaluated as potential mediators rather than assumed to be actionable agents because they accumulate; studies of clearance should distinguish intradialytic removal from sustained solute exposure; and innovations in membranes, convection, volume management, and transport modeling should be tested against clinically important benefits and harms. Prospective studies should determine whether phenotype-informed treatment choices improve symptoms, hemodynamic tolerance, preservation of nutritional status and residual kidney function, quality of life, and longer-term outcomes. Powerful advances will be defined not only by how much more dialysis can remove, but by how precisely its biological targets, treatment mechanisms, and trade-offs are aligned with each patient’s changing needs.
We thank the authors, reviewers, and editorial team for their contributions to this Special Issue and to the continuing development of kidney assistance therapy.
Conflicts of Interest
Dr. Fülöp is an employee of the United States Veterans Health Administration. However, the views and opinions expressed herewith do not reflect the official views or opinions of and are not endorsed by the United States Veteran Health Administration or the United States Government. The remaining authors declare no conflict of interest.
References
- Murea, M.; Deira, J.; Kalantar-Zadeh, K.; Casino, F.G.; Basile, C. The spectrum of kidney dysfunction requiring chronic dialysis therapy: Implications for clinical practice and future clinical trials. Semin. Dial. 2022, 35, 107–116. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Zsom, L.; Zsom, M.; Abdul Salim, S.; Fülöp, T. Subjective global assessment of nutrition, dialysis quality, and the theory of the scientific method in Nephrology practice. Artif. Organs 2020, 44, 1021–1030. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Cozzolino, M.; Magagnoli, L.; Ciceri, P. From Physicochemical Classification to Multidimensional Insights: A Comprehensive Review of Uremic Toxin Research. Toxins 2025, 17, 295. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Bui, T.-A.V.; You, A.S.; Kalantar, S.S.; Yoon, J.; Narasaki, Y.; Sy, J.; Hanna, R.; Daza, A.; Guerrero, Y.; Dang, A.; et al. Leptin and Adiponectin as Uremic Adipokines: Associations with Survival in a Prospective Hemodialysis Cohort. Toxins 2025, 17, 525. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Espi, M.; Charmetant, X.; Fusil, F.; Mathieu, C.; Legras, M.; Pelletier, C.; Glorieux, G.; Soulage, C.; Koppe, L.; Thaunat, O. Chronic Kidney Disease-Associated Defect in Humoral Immune Response Is Driven by Inflammation. Toxins 2026, 18, 104. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Lang, T.; Zawada, A.M.; Wang, X.; Raimann, J.G.; Stauss-Grabo, M.; Grobe, N. Hemodiafiltration in Dialysis Practice: Global Adoption, Toxin Clearance, Inflammation Control, and Clinical Safety. Toxins 2026, 18, 329. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Daugirdas, J.T. Quantifying the Contribution of Nondialytic Factors Affecting Predialysis Serum Phosphate Level When Comparing Hemodiafiltration with Hemodialysis. Toxins 2026, 18, 179. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Daugirdas, J.T. Factors Beyond High-Molecular-Weight Toxin Removal That Might Affect Survival in Comparisons of High-Volume Hemodiafiltration with Hemodialysis. Toxins 2026, 18, 365. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Zhou, X.; Sun, J.; Miao, L. Intermittent Infusion Hemodiafiltration: A Narrative Review of an Emerging Dialysis Modality. Toxins 2025, 17, 442. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Murea, M.; Awad, A.S.; Niyyar, V.D.; Fülöp, T.; Yamashita, A.C.; Tomo, T.; Abe, M. Membrane-to-Patient Optimization: Individualized Dialyzer Selection for Extracorporeal Dialysis. Toxins 2026, 18, 156. [Google Scholar] [CrossRef] [Scilit] [PubMed]
- Swapnasrita, S.; de Vries, J.C.; Stachowska-Piętka, J.; Öberg, C.M.; Gerritsen, K.G.F.; Carlier, A. Comparing Computational Peritoneal Dialysis Models in Pigs and Patients. Toxins 2025, 17, 329. [Google Scholar] [CrossRef] [Scilit] [PubMed]
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2026 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license.