In patients with chronic kidney disease (CKD), signs and symptoms typically develop gradually [1-3]. As kidney function declines progressively, the condition may eventually lead to kidney failure. At this stage, the kidneys can no longer effectively eliminate metabolic waste, resulting in the accumulation of toxins and disturbances in electrolyte balance [4]. The definitive treatment for kidney failure is kidney replacement therapy (KRT) through a kidney transplant. Still, this option is not always feasible or successful, necessitating other KRT modalities, namely hemodialysis (HD) or peritoneal dialysis [5-9]. HD can be divided into conventional HD, hemodiafiltration (HDF), and expanded HD [10].

Conventional HD uses solute diffusion and limited ultrafiltration, using a high-flux membrane inserted in a hollow-fiber dialyzer. Conventional HD is linked to “residual risk” resulting from the accumulation of inadequately removed medium-sized compounds. Meanwhile, HDF uses convection and diffusion to eliminate small- and medium-sized molecules. A significant distinction between HDF and conventional HD is the total ultrafiltered volume per session (dialytic convective dose). HDF can achieve a significantly higher convective dose (10- to 15-fold), leading to a 2-fold increase in middle-molecular-weight solute clearance [11,12].

HDF was initially introduced in the United States by Lee Henderson in 1973 [13]. On-line generation of substitution fluid for HDF was then implemented in the mid-1980s [14]. HDF has since undergone substantial development. See et al (2019) [15] found that HDF was associated with a significantly lower risk of all-cause mortality (adjusted hazard ratio [HR] of 0.79 and 0.88 for patients in Australia and New Zealand, respectively). An association was also found between HDF and reduced cardiovascular mortality, with an adjusted HR of 0.78 in Australian patients. Using the French Renal Epidemiology and Information Network Registry, Mercadal et al (2016) [16] found that the HRs for all-cause and cardiovascular mortality in patients treated exclusively with HDF were 0.77 and 0.66, respectively.

HDF can be conducted with dialysis membranes comprising different raw materials, including cellulose, polysulfone (PS), polyamide, polymethylmethacrylate (PMMA), and polycarbonate [17]. Clinicians must consider several factors to determine which dialyzer membrane is most suitable for dialysis: performance, biocompatibility, cytotoxicity, sterilization, and endotoxin retention [18]. This systematic review and meta-analysis will focus on the efficacy of PMMA dialyzers compared to that of PS in postdilution HDF in patients with kidney failure requiring long-term HD. A PS membrane has an asymmetric structure where the pore size enlarges from inside to outside and a thin skin layer with very small pore sizes to control solute separations. Conversely, a PMMA membrane has a uniform pore size structure, and the entire membrane controls solute separations. Therefore, the separation characteristics of PMMA are broader than those of PS membranes. Furthermore, PS membranes remove uremic toxins from the blood via diffusion and filtration, while PMMA membranes also use adsorption in addition to diffusion and filtration [19-21]. However, PMMA membranes have lower permeability than PS membranes, increasing the risk of fouling and platelet-induced membrane coagulation [22].

A dialyzer’s efficacy is assessed based on the clearance of solutes, including urea, creatinine, phosphate, uric acid, albumin, α1-microglobulin, ß2-microglobulin, myoglobin, prolactin, and κ and λ free light chains [23,24]. Our review aims to answer the following question: in patients with kidney failure requiring long-term HD, does the PMMA dialyzer membrane demonstrate a higher efficacy than the PS dialyzer membrane? The PICO (Population, Intervention, Comparison, and Outcomes) framework will be used during the article search; “P (population)” refers to patients with kidney failure requiring long-term HD, “I (intervention)” refers to HDF using PMMA dialyzer membrane, “C (comparison)” refers to HDF using the PS dialyzer membrane, and “O (outcome)” refers to the efficacy of solute clearance.

The search and screening for the systematic literature review is expected to conclude in May 2026. Data extraction, quality appraisal, and subsequent data synthesis will begin in June 2026. The review should be completed by October 2026, and the study results will be published in 2027.

PS is an amorphous polymer with excellent thermal and chemical stability [35]. The PS membrane is asymmetric, implying that for a membrane thickness of 30 microns, a 1-micron layer controls the separation process, while the remainder has structural functions. Conversely, the PMMA membrane is symmetric, with the whole thickness involved in the separation process [21]. The PMMA membrane adds adsorption to the existing HD diffusion and convection processes. While diffusion and convection remove small and medium-sized molecules, adsorption enables the removal of medium- and high-molecular-weight molecules [22].

The removal of toxins and waste products from the patient’s blood defines the adequacy of HD. Adequate dialysis can improve patient survival and quality of life [36-39]. Kt/V calculation is the most widely used method to assess dialysis adequacy, where K is the effective dialyzer solute clearance, t is the duration of the dialysis session, and V is the solute distribution volume [40]. Currently, small solute clearance is regarded as the best measure of HD adequacy. Kt/V urea, also known as fractional urea clearance, is a widely used tool to assess dialysis efficacy in daily clinical practice [41].

The use of urea as the sole representation of dialysis adequacy tends to oversimplify this concept. In this context, urea does not represent the kinetic behavior of other solutes, including inorganic solutes such as phosphate, middle molecules such as β2-microglobulin, and protein-bound uremic toxins such as p-cresyl sulfate, whose detrimental effects are linked with inflammation, oxidative stress, and cardiovascular disease [42-47]. Hence, the concept of adequacy is evolving to include nonurea parameters [48].

HD adequacy is partly determined by dialyzer membrane clearance [49]. Moreover, membrane properties, including pore size, wall thickness, and surface area, determine solute clearance. The uremic toxins HD removes are divided into four categories: (1) small water-soluble compounds (eg, urea) with a size of <500 Da, (2) a middle-molecular-weight of 500-15,000 Da, (3) large molecules with a size of >15,000 Da, and (4) protein-bound molecules (eg, p-cresol of 108 Da). However, when describing the in vitro performance of dialyzers, many studies only considered the results obtained with small solutes and middle-molecular-weight molecules [50].

Middle-molecular-weight molecules include several cytokines, adipokines, growth factors, and other signaling proteins significantly elevated in dialysis patients compared to that in people with normal kidney function. The serum levels of interleukin (IL)-1β, IL-6, and IL-18, and κ and λ free light chains are also elevated in patients with advanced CKD [51]. These molecules play key roles in the inflammation-cardiovascular disease pathway and are linked with chronic inflammation, atherosclerosis, structural cardiac disease, and protein-energy wasting [52,53].

While conventional HD can effectively remove small uremic molecules (<0.5 kDa), including urea, it poorly removes middle-molecular-weight molecules in the 0.5-60 kDa range. Thus, HDF was invented to improve the clearance of middle-molecular-weight molecules [54]. The clinical application of on-line generation of replacement fluid HDF has been limited in the United States and has predominantly advanced in Europe. Increasing numbers of patients in Japan and other Asian countries have recently opted for HDF treatment [55,56].

The HDF removal efficiency is mainly determined by the replacement fluid volume, while the removal characteristic is affected by the dilution method. Postdilution HDF, when combined with a protein-nonleakage hemodiafilter, is superior in removing small to middle-molecular-weight molecules. Conversely, predilution HDF with a protein-leakage–type hemodiafilter is preferred for removing middle- to high-molecular weight molecules [57]. Predilution HDF efficiently removes middle-molecular-weight molecules of up to α1-microglobulin (26-30 kDa). However, the biological activity of α1-microglobulin is unclear, and the uremic toxins in its vicinity are also unknown. Meanwhile, postdilution HDF increases the removal efficiency for molecular weights up to those of β2-microglobulin (11-12 kDa), which has been established as a uremic toxin [57].

Views differ on the impact of dialyzer characteristics on patient outcomes and the characteristics that the clinician should prioritize. The impact also relies on the dialyzer’s application, where convective therapies such as HDF outperform diffusive therapies such as HD [58]. Although no ideal dialyzer can efficiently remove all types of uremic toxins while retaining vitamins and other essential elements, expanding the range of uremic toxin clearance should benefit patients. Nevertheless, further research is required to demonstrate the superiority of these new dialyzers [59].

Our study will include stable patients with kidney failure undergoing thrice weekly HD sessions. We recognize that the duration of each dialysis session (in minutes) might differ between studies. Therefore, we will only include studies with similar dialysis prescriptions and the same modality, postdilution HDF, to limit bias. To our knowledge, this study will be the first to conduct a systematic review and meta-analysis of the available evidence regarding comparing clearance between PMMA and PS dialyzer membranes for postdilution HDF in patients with kidney failure.

Our study may have several limitations. Even if the studies used the same membrane material (PS or PMMA), if they used dialyzers from different manufacturers or the same manufacturer but different versions (type, design, or specification), bias could result from different clearance capacities. Second, clearance is subject to methodological heterogeneity, as the time point for each blood sampling collection might differ between studies. We will not specify time points as inclusion criteria due to the risk of eliminating potentially relevant studies.

Third, our study will not assess some aspects potentially affecting the fluid removal rate. We will not evaluate the patients’ primary kidney disease or other health problems potentially affecting the blood protein level (albumin). Moreover, we will not assess the quality of blood vessels or HD lumen catheters used for dialysis, which potentially affect dialysis efficiency through resistance and subsequently influence the fluid removal rate.

In summary, our systematic review and meta-analysis may help compare the solute clearance of PMMA and PS dialyzer membranes for postdilution HDF in people with kidney failure. The results may help clinicians determine the most suitable membrane for HDF for their routine HD practice. The data may also provide a starting point for further research on the impacts of various dialyzer membranes on the HDF solute removal rate.