As kidney function declines and progresses to end-stage kidney disease (ESKD), kidneys lose their ability to remove waste products effectively. Consequently, patients with ESKD require kidney replacement therapy such as dialysis or kidney transplantation. Hemodialysis (HD) involves the movement of blood and dialysates across the HD membrane to ensure efficient dialysis. Water and solutes pass through semipermeable membranes via distinct separation mechanisms such as diffusion and ultrafiltration. These mechanisms rely on the pressure gradient created by the patient’s blood to allow water and solutes to move toward the dialysate side. HD is critical for removing waste products and regulating the fluid balance, particularly for clearing uremic toxins. For decades, continuous efforts have been made to optimize uremic toxin removal using HD membranes. Standard high-flux dialysis membranes are not highly effective at removing middle-molecular-weight uremic toxins, particularly those larger than 15 kDa. Although online hemodiafiltration (HDF) utilizing high-flux membranes combines convection and diffusion to broaden the range of uremic toxins, it also presents challenges, such as the need for high vascular access flow, ultrapure water, and large convective volumes. Recent advancements in dialysis membranes aim to enhance convective clearance while limiting albumin loss to improve the elimination of middle-molecular-weight toxins [1]. Consistent with these advancements, a medium cut-off (MCO) membrane, a next-generation dialysis membrane, was developed to enable expanded HD by effectively removing the middle molecules. This review aims to provide a comprehensive overview of MCO membranes and their clinical implications, discussing their advantages over traditional dialysis membranes and their potential to improve patient outcomes in ESKD.

HD membrane development is a continuous process. HD membranes are classified into cellulosic and synthetic types, based on their composition [2]. Recently, synthetic membranes have become predominant, and a classification method based on the ultrafiltration coefficient has been proposed. HD membranes were initially classified as having low or high permeability based on differences in water membrane permeability. Subsequent research further refined this classification into low- and high-flux membranes, based on their ability to remove fluids and molecules [3]. Low- and high-flux membranes are categorized according to their ultrafiltration coefficients (Kuf), which represent the pore sizes of the membranes. High-flux dialyzers are defined as having a Kuf > 20 mL/h/mmHg, whereas low-flux dialyzers are defined as having a Kuf < 10 mL/h/mmHg. Furthermore, acknowledging the critical role of albumin in human health, researchers have revised the classification to incorporate factors such as water permeability, beta-2 microglobulin clearance, and albumin-related parameters [4].

As kidney function declines, uremic toxins accumulate and are traditionally classified as small water-soluble molecules, protein-bound solutes, or middle molecules [5]. When not used in HDF mode, high-flux dialysis membranes can remove uremic toxins with a molecular weight of approximately 15 kDa [6]. Accordingly, uremic toxins > 15 kDa are classified as large middle-molecules, with further subdivisions based on the molecular weight and clearance capacity of the dialysis membranes. Recently, experts have suggested refining the classification of middle molecular uremic toxins, emphasizing the need to consider their molecular structure, removal methods, and correlation with clinical symptoms, which calls for an update of the current framework [7]. Uremic toxicity has harmful effects on various organs and metabolic processes. Previous studies have demonstrated that uremic toxins with molecular weights greater than 15 kDa, such as cytokines, adipokines, and hormones, contribute to chronic inflammation, atherosclerosis, structural heart disorders, and secondary immunodeficiency [6]. Therefore, it is important to clarify the relationship between the accumulation of specific toxins, their diverse physicochemical characteristics, and their clinical symptoms. A more detailed classification of uremic toxins has been proposed based on the removal capabilities of the currently available dialysis membranes [7]. Molecules are categorized as: small (< 0.5 kDa), small-middle (0.5–15 kDa), medium-middle (> 15–25 kDa), large-middle (> 25–58 kDa), and large (> 58–170 kDa). High-flux dialyzers used in standard HD can clear molecules up to 15 kDa, with this threshold increasing to 25 kDa when used in HDF. A new class of membranes, the MCO membrane, allows the removal of molecules of up to 56 kDa. Compared to high-flux membranes, MCO membranes demonstrate superior clearance efficiency for larger molecules within the 25–56 kDa range.

The MCO membrane, a new-generation dialysis membrane, was developed to enable expanded HD by effectively removing middle molecules up to 56 kDa. The characteristics and classification of the HD membranes are summarized in Table 1. In comparison, HCO membranes developed earlier to eliminate free light chains (kappa and lambda) in patients with myeloma kidney are associated with significant albumin loss owing to their large pore size, which limits their use for short-term therapeutic purposes [8].

An MCO membrane was developed to enable the clearance of large middle-molecules while minimizing albumin loss. The MCO membrane was designed with tailored pore sizes and optimized to accommodate different molecular sizes while achieving an effective pore size (Stokes–Einstein radius) [9]. The Stokes–Einstein radius of the molecular weight cutoff was aligned with the effective pore radius of the membrane. MCO membranes feature an improved pore size distribution and adopt a tighter configuration to enable effective removal of large middle-molecules in clinical practice while minimizing albumin loss [6,10]. Moreover, the MCO membrane increases internal filtration by reducing the inner diameter and wall thickness compared to traditional high-flux membranes, thereby enhancing its convective flow capacity (Fig. 1) [9]. This results in greater permeability owing to significant internal filtration in the proximal section, driven by an increased end-to-end pressure drop [11]. The improved convective transport along the fibers facilitates the removal of larger molecules with low diffusion coefficients. Adequate backfiltration in the distal part of the MCO membranes eliminates the need for exogenous substitution fluids [12]. These unique sieving and filtration properties enable MCO membranes to provide expanded clearance of uremic toxins in the 15–45 kDa range, offering a performance comparable to that of high-volume HDF and superior to that of standard high-flux HD membranes. However, some earlier studies have reported that the removal efficiency of middle molecules, such as myoglobin and kappa- and lambda-free light chains, was superior in case of HDF compared with MCO membranes [13,14]. Other studies have demonstrated superior reduction ratios for MCO membranes [15]. Notably, the convective volumes used in these HDF studies varied considerably—30.4 ± 4.1 L per session in the study by Maduell et al. [13] and 23.5 ± 3.8 L in the study by Reque et al. [15]—suggesting that higher convective volumes may enhance solute clearance. Consequently, HDF requires specific conditions to achieve optimal performance, including high convective volumes, high vascular access flow rates, and consistent use of ultrapure dialysate. These technical requirements may limit their broad application in clinical settings. In contrast, MCO membranes, which can be used with a conventional dialysis infrastructure, may offer a more practical and accessible alternative for improving middle-molecule clearance in routine clinical practice. However, evidence on the long-term clinical impact of MCO membranes, including their effects on cardiovascular events and both cardiovascular and all-cause mortality, remains limited, highlighting the need for further research.

There have been concerns that using an MCO membrane for HD could result in higher albumin loss than with high-flux HD and HDF [16]. In a 12-month study of 638 patients undergoing extended HD with an MCO membrane, serum albumin levels showed minimal reduction, with a nadir of 3.9 g/dL [17]. Variations remained within 5% of the baseline, and mean levels remained within the normal range (3.5–5.5 g/dL). Previous studies have shown that serum albumin loss is comparable between MCO membrane use, high-flux HD, and HDF [18-20]. Over three years of treatment with MCO membranes, an evaluation of clinical safety revealed no significant differences in serum albumin levels over time compared to the HD group (mean monthly change difference = -0.0003, p = 0.855) [21]. Additionally, albumin loss with the use of an MCO membrane was lower than that observed with peritoneal dialysis [22]. Thus, although the use of MCO membranes may lead to a slight decrease in serum albumin levels, the reduction was not significant, and the levels generally remained within the normal range, indicating that they can be safely used.

HD is a life-saving therapy; however, it can have environmental impacts as it consumes considerable amounts of water and energy, and generates significant waste. Additionally, the amount of water used for HD varies depending on the treatment and modality applied. In this context, MCO membranes offer eco-friendly properties. HDF consumes substantially more water and dialysis fluid than conventional or expanded HD [55]. This results in increased power consumption and waste production. Moreover, HDF requires the use of a second sterilizing ultrafilter, which further increases the filter usage and contributes to additional waste production. MCO membranes offer benefits comparable to HDF but require less water because of their filtration–backfiltration capabilities [56]. Additionally, the use of an MCO membrane can reduce both power consumption and waste production compared to HDF. Furthermore, MCO membranes are smaller and lighter than high-flux membranes [56], which helps reduce the weight of medical waste. Beyond the clinical effects of MCO membranes, these environmental impacts align with the goals of green dialysis [57,58].

Most studies using MCO membranes for HD have reported no additional adverse events or serious complications associated with the procedure. Although MCO dialyzers show promising results in removing middle-molecule uremic toxins, there are limited long-term data on their impact on patient outcomes such as survival rates and cardiovascular health. In a Korean three-year cohort study comparing MCO membranes with high-flux membranes, there were no significant differences in clinical efficacy and safety outcomes during the treatment period [21]. The levels of inflammatory cytokines remained stable, and there were no differences in the rates of death, cardiovascular events, infections, or hospitalization between the groups. Additional clinical trials should consider the diverse characteristics of patients undergoing dialysis in order to determine the long-term safety and potential benefits of MCO membranes.

MCO membranes are not effective in removing gut-derived protein-bound uremic toxins (PBUTs) and large-molecule uremic toxins [20,59]. PBUTs such as indoxyl sulfate and p-cresyl sulfate are associated with cardiovascular disease in patients with CKD [60]. Previous studies have shown that post-dialysis plasma levels of indoxyl sulfate and p-cresyl sulfate are not significantly different among high-flux HD, HDF, and MCO-HD [20].

MCO dialyzers tend to be more expensive than conventional high-flux membranes, hindering their widespread adoption, particularly in resource-limited settings. However, the reduction in hospitalization rates and shorter hospital stays associated with MCO membrane use could lower overall healthcare costs, making it important to consider these potential savings in the evaluation [48,49]. For instance, one study reported that the use of MCO membranes was significantly associated with fewer hospitalization days and reduced doses of medications, including ESA, iron supplements, antihypertensive drugs, and insulin [61]. This study estimated a 23% reduction in the annual hospitalization-related costs in the MCO group. Another study suggested that the use of an MCO membrane could lead to a 45% reduction in hospitalization rates and approximately USD 6,098 in annual healthcare cost savings per patient [49]. Future research, particularly large-scale randomized controlled trials, are required to confirm and strengthen the findings of this study. Further cost-effectiveness analyses are warranted to comprehensively evaluate the economic implications of adopting MCO membranes in various health care settings.

MCO membranes represent a significant advancement in HD, offering improved removal of large middle-molecules of uremic toxins compared to traditional HD membranes while maintaining minimal loss of albumin. MCO membranes have been shown to improve patient QOL, reduce ESA dose and resistance, lower hospitalization rates, and decrease overall healthcare costs. However, important clinical questions remain unresolved, including the long-term safety of MCO membranes, their cost-effectiveness across diverse healthcare settings, and their impact on adverse outcomes, such as cardiovascular events and all-cause mortality. It is also important to identify the patient population that may benefit the most from MCO membrane therapy, as its effects could vary depending on individual patient characteristics. Future research should address these gaps through large-scale, multicenter randomized controlled trials and real-world data analyses. As nephrology continues to evolve toward patient-centered and precision-based care, the role of MCO membranes should be further explored as part of a broader strategy to improve dialysis quality and long-term patient outcomes.