Efficacy and Safety of the Medium Cut-Off Elisio HX Dialyzer

Postdilution online hemodiafiltration (HDF) is currently the dialysis treatment achieving the most favorable results in terms of uremic toxin removal, above predilution HDF and high-flux hemodialysis (HD) [1‒3]. Scientific evidence of the superiority of HDF versus HD in overall and cardiovascular survival has been demonstrated [4‒7], and consequently, it can currently be considered in terms of efficacy as the standard conventional HD treatment [8], although today it is still not the most used.

Expanded HD (HDx) with medium cut-off (MCO) membranes, designed to improve the permeability of dialysis membranes, has been incorporated into clinical practice, and although these dialyzers can only be used in the HD modality, these dialyzers could provide an alternative to HDF, since they achieve the same removal performance as postdilution HDF [9‒14]. This is related to the tailored cut-off of the pores combined with an internal architecture that allows MCO membranes to optimize the internal convection and, thus, increased removal capacities for middle molecules and large middle molecules than standard high-flux HD treatments. While maintaining a high solute clearance of less than 10 kDa, such membranes were developed to improve the clearance of medium to high molecular weight (MW) solutes (in the range of 10–50 kDa) [10].

The pharmaceutical industry has developed and refined its dialyzers to achieve a higher level of purification, adapting to new treatment modalities, which has considerably increased the supply of dialyzers. Thus, Nipro developed an MCO dialyzer named Elisio HX. In 2021, obtaining the CE mark, this dialyzer became commercially available in several European countries. As in the previous Elisio H generation, the membrane is made of polyethersulfone and polypropylene housing. The only difference is the increased pore size.

This study aimed to assess the safety and efficacy of this new MCO dialyzer and compare it with contrasted HD, HDx, and postdilution HDF treatments. Removal of a wide range of MW molecules was assessed, and safety was evaluated with blood and dialysate albumin loss.

All prevalent adult patients on a maintenance HD program in our unit were considered for inclusion. Patients were excluded if they were in unstable clinical condition, with a scheduled living donor transplantation, or did not accept to participate. The study finally included 18 patients, 13 men and 5 women, with a mean age of 69.8 ± 12 years (range 51–89) on a regular HD program. Among them, there were 6 diabetics, 11 had high blood pressure, their mean Charlson index was 7.39 ± 2.43, the mean hemoglobin was 11.5 ± 0.99 mg/dL, the ferritin was 345 ± 162 ng/mL, the albumin was 3.9 ± 0.29 g/dL, the prealbumin was 0.273 ± 0.078 g/L, the D₃-vitamin was 24.02 ± 14.3 ng/mL, and the PTH was 268 ± 144 pg/mL. Patients were dialyzed with a 5008 Fresenius monitor, suitable for online HDF. Vascular access was an autologous arteriovenous fistula in 14 patients, tunneled catheter in 3 patients, and prosthetic arteriovenous fistula in the other. Underlying renal diseases were chronic glomerulonephritis (4 patients), nephroangiosclerosis (3 patients), tubulointerstitial nephropathy (2 patients), systemic disease (2 patients), diabetic nephropathy (1 patient), polycystic kidney disease (1 patient), urological cause (1 patient), and undiagnosed nephropathy (4 patients). The anticoagulation used was low MW heparin (tinzaparin) in 13 patients and heparin sodium in 3, and the remaining 2 patients were dialyzed without heparin. Net fluid removal was set individually, depending on the patient’s clinical needs. All patients were anuric with a urine volume inferior to 50 mL/day. Each patient acted as their own control receiving five different sessions with their usual parameters (dialysis duration 285 ± 19 min, blood flow rate [Qb] 444 ± 16 mL/min, and dialysate flow rate [Qd] 400 mL/min) at the same day of the week independently if it was in the long or short interdialytic period:

1. High-flux FX80 Cordiax^TM, helixone, Fresenius Medical Care, in HD.

2. High-flux Elisio H19^TM, polyethersulfone, Nipro, in HD.

3. MCO Elisio HX19^TM, polyethersulfone, Nipro, in HD.

4. MCO Theranova 400^TM, polyarylethersulfone, Baxter, in HD.

5. High-flux FX80 Cordiax^TM, helixone, Fresenius Medical, in postdilution HDF.

Dialyzers were automatically primed with the 5008 monitors. Online HDF was performed with postdilution infusion and automatic infusion flow. The order of the different treatment sessions was randomly assigned. The dialyzer characteristics are summarized in Table 1. Blood and dialysis fluid samples for analyses were taken from each patient in the same dialysis session of the week.

The dialysis parameters collected in each session were as follows: real duration, dialyzer, Qb, Qd, recirculation index measured by the temperature module, arterial pressure, venous pressure, three-point measurement (including blood outlet, dialysate inlet, and dialysate outlet) for transmembrane pressure (TMP) calculation, initial and final hematocrit automatically measured by the blood volume monitor biosensor, initial and final body weight, the volume of blood processed, and replacement volume. Laboratory measurements included concentrations of urea (60 Da), creatinine (113 Da), ß₂-microglobulin (11,800 Da), myoglobin (17,200 Da), prolactin (23,000 Da), α₁-microglobulin (33,000 Da), α₁-acid glycoprotein (41,000 Da), and albumin (66,000 Da) in serum at the beginning and at the end of each session to calculate the percentage reduction ratio (RR) of these solutes. Free immunoglobulin light chains (FLCs) were also measured, kappa FLC (kFLC) with a MW of 22,500 Da, and lambda FLC (λFLC) with a MW of 45,000 Da.

The final concentration of ß₂-microglobulin, myoglobin, prolactin, kFLC, α₁-microglobulin, α₁-acid glycoprotein, λFLC, and albumin was corrected for the degree of hemoconcentration and the distribution volume (approximate extracellular volume) according to Bergström and Wehle [15]. Urea and creatinine were measured by molecular absorption spectrometry, albumin and β₂-microglobulin were measured by immunoturbidimetry, and myoglobin and prolactin were measured by indirect enzyme immunoassay; all of them were performed in an Atellica Solution analyzer (Siemens Healthineers, Tarrytown, NY, USA). Finally, α₁-acid glycoprotein, α₁-microglobulin, kFLC, λFLC were measured by immunonephelometry using the BNII analyzer (Siemens Healthineers).

A proportional part of the dialysis fluid was collected throughout the treatment to quantify albumin loss employing a reverse perfusion pump. A global removal score (GRS) [9] was also calculated with the following formula: ([UreaRR + β₂-mRR + myoglobinRR + prolactinRR + α₁-microglobulinRR + α₁-acid glycoproteinRR − albuminRR]/6).

The results are expressed as the arithmetic mean ± standard deviation. Quantitative parameters were analyzed with the Student t test for paired data. Parametric data were analyzed with ANOVA for repeated measurements, followed by Bonferroni’s post hoc test. p < 0.05 was considered statistically significant. Analyses were performed using SPSS software version 23 (SPSS, Chicago, IL, USA).

We observed proper tolerance to filters, while no adverse reactions in the connection and disconnection or during the HD or HDF sessions in the studied population. Replacement fluid in online postdilution HDF was 30.9 ± 4 L (range 18–34 L) with helixone HDF treatment.

There were no differences in dialysis parameters: Qb, total blood processed, vascular access recirculation, real session duration, initial weight, final weight, weight gain, initial and final hematocrit measured by the dialysis monitor, arterial pressure, and venous pressure (Table 2). As expected, the TMP and the replacement volume were significantly higher in HDF sessions than in the other sessions in HD.

Similar urea, creatinine, and phosphate RRs were observed (Table 3). The only statistically significant difference was observed in urea RRs between Elisio HX19 in HDx and FX80 Cordiax in HD treatment.

The RRs of β₂-microglobulin, myoglobin, and prolactin with HDF and HDx treatments were significantly higher than those obtained with the two dialyzers in HD treatments (Table 3). In addition, FX80 in HDF obtained statistically superior results to both dialyzers in HDx (Table 3).

In the high MW range, the α₁-microglobulin RRs with HDF and HDx were significantly higher than those obtained in both HD treatments (Table 3). The α₁-acid glycoprotein RRs with helixone in HDF were significantly higher than those obtained in HD and Elisio HX treatments (Table 3).

The values of kFLC RRs were above 60% in all treatments. The kFLC RR was significantly higher with HDF than with HD and HDx (Table 3); the kFLC RR was higher with both HDx dialyzers than with the two HD ones.

The values of λFLC RRs varied between 20% and 60%. The RRs of λFLC were also significantly higher with HDF than with HD and HDx; λFLC RR was higher with both HDx dialyzers than with the two HD ones (Table 3).

The mean albumin RR was less than 11% in all study situations, and there were no significant differences (Table 3). The mean amount of dialysate albumin loss was less than 3 g in all situations, and significant differences were observed between different study situations (Fig. 1). Albumin losses with both HD dialyzer were less than 1 g, and HDx and HDF treatments were between 1.5 and 2.5 g (Table 3).

The highest GRS values were obtained with the FX80 Cordiax in HDF treatment and were statistically superior to HD and HDx treatments. The GRS values with HDx treatments were statistically higher than those obtained with HD (Fig. 2). There were no differences between Elisio HX and Theranova.

The present study is, to our knowledge, the first to analyze in vivo data on the performance of the new MCO dialyzer, Elisio HX, in HDx treatments. We compared this dialyzer with commonly used dialyzers suitable for both HD and HDF (helixone) and specific dialyzers for HDx. This study shows that the solute permeability of this dialyzer is higher than that of other dialyzers in HD, similar to that of the MCO Theranova dialyzer and slightly lower than that of the helixone dialyzer in HDF postdilution. The Elisio HX dialyzer showed excellent safety, with a lower RR of albumin in blood and dialysate albumin loss of less than 2.5 g per dialysis session.

The helixone membrane was chosen for this study as it is representative of high-flux dialyzers that allow HD and HDF and it is one of the most widely used in our dialysis unit. We used the latest generation helixone (FX Cordiax series) fabricated by nanotechnology, which has proven to be superior to previous generations (FX series) [9].

This study confirms the superiority of HDF and HDx over high-flux HD, showing the HDx as the alternative that came closest to postdilution HDF and clearly superior to HD, as previously published in the literature [16‒21], reinforcing the importance of both the choice of the dialyzer and the treatment modality to obtain optimal performance. The optimization of the backfiltration of the Elisio HX dialyzer is not linked to the classic reduction of the internal diameter of the fibers as happens in the Theranova one (180 μm), but through keeping the inner diameter of the fibers at 200 μm. Moreover, backfiltration is particularly compensated by lengthening the fibers from 236 mm of the Theranova to 281 mm of the Elisio. We also evaluated the removal of free light chains. In comparison with the study by Kirsch et al. [16], using the same dialyzers (helixone and MCO membranes), the RRs of kFLC were somewhat higher, especially in HD (36% vs. 65%), and were more similar in HDx (66% vs. 77%) and in HDF (72% vs. 84%). The results were similar with λFLC in HD (13% vs. 25%), HDx (42.5% vs. 45%), and post-HDF (38% vs. 57%). Both light chains, 22 and 45 kDa, could be considered as a good marker not only for the management of multiple myeloma, but also as a good differential marker of depurative efficacy, especially differential in the MW range of 40–45 kDa. The differences observed in the RR between α₁-acid glycoprotein and λFLC, with similar MWs, are logically due to more components than the molecular size itself. In this case, there is an evident influence of the free fraction of the λFLC, unlike α₁-acid glycoprotein, being able to partially bind to proteins, in addition to the different electric charges and isoelectric points (pI) between them.

The safety of MCO dialyzers is ensured by restricting pore sizes to limit albumin losses below 5 g per session [22, 23]. In this regard, most published studies report that MCO membranes lead to a higher albumin loss than HD and show inconsistent results compared to HDF [1, 2, 9, 13, 16, 24]. In this study, Elisio HX dialyzer in HDx treatment has maintained a totally safe behavior, with a global albumin RR of around 10% and a mean of dialysate albumin loss of 1.6 g per session, higher than HD treatments as previously described, but slightly lower although without reaching statistical significance compared to Theranova (mean albumin loss 2.0 g) and HDF (mean albumin loss 2.3 g). Although albumin loss could be considered clinically tolerable in all study treatments, it is necessary to mention that MCO membranes should only be used in HD mode to avoid increased albumin losses with a progressive reduction in serum levels [25].

A limitation of the study is that there was no long-term clinical follow-up with this new dialyzer. Although all the dialyzers evaluated are synthetic, biocompatibility aspects such as complement activation, proinflammatory cytokines, platelet activation, or endothelial damage have not been analyzed in this study. The performance of dialysis therapy is usually evaluated by solute removal, and the quality of dialysis therapies should be evaluated by clear outcome studies such as randomized controlled trials [26]. The mass of solute collected in the dialysate corresponds only to its elimination by diffusion and convection, obviating the adsorption one, which cannot be quantified routinely; however, to assess the overall loss of albumin (adsorption, diffusion, and convection), the albumin RR in the blood was reported.

In conclusion, the results of this study show how the new MCO dialyzer, Elisio HX, offers excellent behavior and tolerance, with good efficiency and complete safety. This dialyzer represents an upgrade when compared to their predecessor, Elisio H, in HD treatment and is very close to the removal capacity of HDF treatment. There were no significant differences in RRs between both MCO dialyzers compared in this study. This improvement has been achieved thanks to a suitable albumin loss. However, as a MCO dialyzer, its use should be limited in the HD modality, avoiding convective treatments. Because of the large number of currently available dialyzers, studies such as the present are needed to allow a personalized dialysis prescription.

We would like to express our gratitude to all participating patients, as well as to all the staff of the Dialysis Section of Hospital Clínic of Barcelona, for their collaboration in this study and enthusiasm.

This study was approved by the Clinical Research Ethics Committee of Hospital Clínic of Barcelona, approval number HCB/2018/1099. All study subjects gave their written informed consent. Data collection has followed the Regulation (EU) 2016/679 (General Data Protection Regulation), its subordinate national and regional laws, and the Declaration of Helsinki principles.

The authors declare no financial support for the project. Francisco Maduell has received consultancy fees and lecture fees from Baxter, Fresenius Medical Care, Medtronic, Nipro, Toray, and Vifor. The other authors have no conflict of interest to declare.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Francisco Maduell conceived the study. José Jesús Broseta, Diana Rodríguez-Espinosa, Jimena Del Risco Zeballos, Miquel Gomez, Lida M. Rodas, Marta Arias-Guillén, Manel Vera, Néstor Fontseré, Maria del Carmen Salgado, and Nayra Rico acquired the data. Francisco Maduell and José Jesús Broseta analyzed the data, made the figures, and drafted the paper. All authors have revised the drafts and approved the final one.

The data that support the findings of this study are available from the corresponding author, Francisco Maduell, upon reasonable request.