Filtration flow rate at low convection volume in continuous hemodiafiltration using cytokine adsorbing hemofilter does not affect cytokine clearance in an experimental model

Abstract

Background

In recent years, continuous kidney replacement therapy (CKRT) with a cytokine adsorbing hemofilter (CAH) has been used in clinical practice to treat acute kidney injury associated with hypercytokinemia. Two types of CAH are available, including polymethyl methacrylate (PMMA) and polyethylenimine-coated polyacrylonitrile (AN69ST), each having distinct adsorption mechanisms. PMMA adsorbs substances with hydrophobic bases through hydrophobic interactions, resulting in occlusion of the membrane pores. AN69ST adsorbs positively charged substances through electrostatic bonds because its bulk layer is negatively charged. In both CAH, the adsorption efficiency of cytokines with large molecular weights is likely affected by filtration rather than by diffusion transfer. These adsorbing-type membranes have limitations in terms of filtration flow rate because of their low water permeability. The relationship between the adsorption effect and the filtration flow rate in CAH-CKRT has not been fully investigated.

Purpose

The effect of low-convection volume settings below 600 mL/h on cytokine adsorption characteristics was experimentally investigated in CKRT.

Materials and methods

Test solutions, including albumin (4.5%), creatinine (10 mg/dL), interleukin (IL)-6 (1000 pg/mL), and IL-8 (1000 pg/mL), were prepared. The test solution was circulated through the experimental circuit of continuous hemodiafiltration (CHDF) with a total convection volume ranging from 0 to 600 mL/h. The test solution was circulated through the experimental circuit of hemodiafiltration (CHDF) with various dialysate flow rates (QD; 0, 200, 400, and 600 ml/h) and filtration flow rates (QF; 0, 200, 400, and 600 ml/h). Samples immediately before and after CAH from the sampling port of the circuit were collected seven times at 1-min intervals. Clearances of creatinine, IL-6, and IL-8 were calculated for each setting of the proportions of QD and QF during CKRT using PMMA and AN69ST.

Results

Creatinine clearance increased with increasing QD and QF, regardless of the membrane type. There were no significant differences in the adsorption clearance of IL-6 and IL-8 among the different settings, regardless of the membrane type.

Conclusion

The results of this study indicate that the cytokine clearance at a low convection volume in CAH-CKRT was not affected by the filtration flow rate.

Background

In intensive care medicine, hypercytokinemia can lead to organ failure in critically ill patients [1]. Acute kidney injury (AKI) is an organ failure caused by hypercytokinemia. Patients with AKI seldom undergo intermittent blood purification therapy. Continuous kidney replacement therapy (CKRT) is generally performed in patients with AKI and hemodynamic derangement. Cytokine elimination is anticipated in addition to kidney replacement therapy when CKRT is performed for AKI associated with hypercytokinemia [2].

Because CKRT is generally performed for 24 h, it is executed at lower blood convection volumes, including filtration flow rate (QF) and dialysate flow rate (QD), than intermittent hemodiafiltration. Two methods are employed in CKRT for cytokine removal: increasing filtration volume and using a cytokine-adsorbing hemofilter (CAH). Ronco et al. reported that increased filtration volume improves AKI prognosis by removing inflammatory mediators [3].

In recent years, CKRT using CAHs has been used in general practice to improve hemodynamics in patients with AKI by adsorbing and removing inflammatory cytokines [4,5,6,7,8]. The filters used in CAH-CKRT include polymethyl methacrylate (PMMA) and polyethylenimine-coated polyacrylonitrile (AN69ST). These two filters exhibit distinct adsorption mechanisms. PMMA adsorbs substances from the pores of membranes via hydrophobic interactions [9], whereas AN69ST adsorbs substances via electrostatic bonds, because its bulk layer is negatively charged [10].

Therefore, adequate CKRT settings for cytokine elimination have not been fully investigated. In this study, the effects of QD and QF settings on cytokine adsorption characteristics were experimentally investigated for CKRT with low blood convection volumes.

Materials and methods

An experimental model of continuous hemodiafiltration was prepared using a test solution containing cytokines. In this experimental model, the effects of different QD and QF settings on cytokine clearance were evaluated.

Preparation of test solutions

The test solution for the experimental circulation was prepared as follows: bovine albumin (Nacalai Tesque, Kyoto, Japan) was added to 2.2 L of phosphate-buffered saline (PBS) solution (GIBCO^®, Life Technologies, MA), and the final concentration was adjusted to 4.5%. A total of 2200 mg of creatinine (Nacalai Tesque, Kyoto, Japan), 22 µg of IL-8 (pI 8.5; 8 kDa) (Sigma-Aldrich, St. Louis), and 22 µg of IL-6 (pI 5.3; 21 kDa) (Sigma-Aldrich) were added to the PBS solution adjusted to 4.5% albumin concentration. The concentration of each solute was adjusted to 10, 1000, and 1000 pg/mL for creatinine, IL-6, and IL-8, respectively. The pH of the test solution was adjusted to 7.4. The test solution was heated to 37 °C and stirred continuously using a magnetic stirrer.

Experimental methods

The experimental circuit is illustrated in Fig. 1. The test solution was drawn from the inlet circuit and discharged from the circuit outlet. The hemofilters used for the experimental circuits were PMMA (CH-1.0N®, Toray Medical Co., Ltd., Tokyo, Japan) and AN69ST (SepXiris100®, Baxter Limited, Tokyo, Japan). The flow rate of the test solution (QT) was adjusted to 150 mL/min. The convection volumes of the settings followed the QD and QF. The proportions of QD and QF were (1) 0 mL/h, 0 mL/h (0/0); (2) 200 mL/h, 0 mL/h (200/0); (3) 400 mL/h, 0 mL/h (400/0); (4) 600 mL/h, 0 mL/h (600/0); (5) 0 mL/h, 200 mL/h (0/200); (6) 0 mL/h, 400 mL/h (0/400); and (7) 0 mL/h, 600 mL/h (0/600). Cytokines were circulated through the circuit for 10 min for stabilization before sampling. Samples of the test solution were collected seven times: before and after the hemofilter, and every minute thereafter the start of CKRT. The samples were stored at −80 °C until further analysis. The albumin concentration in each sample was determined using the biuret method with a dedicated reagent and creatinine concentration was determined using an enzymatic method. The concentrations of IL-6 and IL-8 in each sample were measured using a commercially available ELISA kit, following the manufacturer’s instructions.

Fig. 1

Schematic view of experimental circuits. Stars are indicated as sampling points

The clearance of each solute was determined using the following equation, based on previous studies [11]:

CKRT clearance (ml/min) = (Cti − Cto)/Cti × (QB—QF) + QF.

(Cti and Cto represent the concentrations of substances before and after hemofiltration, respectively).

Since the cytokine clearance of CAH is considered to be dependent on three factors: diffusion, filtration, and adsorption [12], cytokine clearance was defined as adsorption clearance regardless of the setting conditions.

Statistical analysis

Continuous variables are expressed as medians (interquartile ranges). Statistical analyses were performed using R (version 4.20), and unpaired t-tests were used to compare two groups. The Friedman test was used for more than three groups and the Tukey test was used as a post hoc test when there was a statistically significant difference. Statistical significance was set at P < 0.05.

Result

The clearance of each solute was compared for each membrane type and convection volume setting condition. The clearance of creatinine, IL-8, and IL-6 were shown for each condition in this order; since IL-8 and IL-6 were not detected on the filtrate side, we assumed that, if they decreased before or after the membrane, they were adsorbed on the membrane itself and defined the clearance as adsorption clearance.

Creatinine clearance (Fig. 2)

Fig. 2

Creatinine clearance of CHDF using PMMA and AN69ST. Box plot, center lines indicate the medians; box limits indicate the 25th and 75th percentiles; whiskers extend to minimum and maximum values. White boxes indicate no settings, light-grey boxes indicate QD settings, and dark-grey boxes indicate QF settings. QF, filtration flow rate; QD, dialysate flow rate; CHDF, continuous hemodiafiltration; PMMA, polymethyl methacrylate; AN69ST, polyethyleneimine-coated polyacrylonitrile. *P < 0.01. †P < 0.05 versus 0/0(QD/QF). #: P < 0.01 versus 0/0(QD/QF)

In the 0/0 setting, creatinine clearance of PMMA was 1.0 mL/min (0.4–1.3), while it was significantly higher in all convection volume settings than in 0/0 settings. Creatinine clearance increased with higher QD and QF values. Similar to PMMA, AN69ST exhibited a higher clearance depending on the convection volume. However, in the AN69ST clearance at the same convection volume, QF was significantly higher at 11.7 mL/min (11.3–12.1) in the 0/600 setting than at 9.2 mL/min (7.2–10.0) in the 600/0 setting (P = 0.011).

Adsorption clearance of IL-8 (Fig. 3)

Fig. 3

IL-8 adsorption clearance of CHDF using PMMA and AN69ST. Box plot, center lines indicate the medians; box limits indicate the 25th and 75th percentiles; whiskers extend to minimum and maximum values. White boxes indicate no settings, light-grey boxes indicate QD settings, and dark-grey boxes indicate QF settings. IL-8, interleukin-8; QF, filtration flow rate; QD, dialysate flow rate; CHDF, continuous hemodiafiltration; PMMA, polymethyl methacrylate; AN69ST, polyethyleneimine-coated polyacrylonitrile

No significant differences in adsorption clearance were identified for PMMA in any convection volume setting (P = 0.717). The mean adsorption clearance of IL-8 in all settings in PMMA was 17.1 mL/min (11.9–22.2). No significant differences in adsorption clearance were identified for AN69ST in any convection volume setting (P = 0.569). The mean adsorption clearance of IL-8 in all settings in AN69ST was 46.2 mL/min (40.4–49.9).

Adsorption clearance of IL-6 (Fig. 4)

Fig. 4

IL-6 adsorption clearance of CHDF using PMMA and AN69ST. Box plot, center lines show the medians; box limits indicate the 25th and 75th percentiles; whiskers extend to minimum and maximum values. White boxes indicate no settings, light-grey boxes indicate QD settings, and dark-grey boxes indicate QF settings. IL-6, interleukin-6; QF, filtration flow rate; QD, dialysate flow rate; CHDF, continuous hemodiafiltration; PMMA, polymethyl methacrylate; AN69ST, polyethyleneimine-coated polyacrylonitrile

PMMA showed no significant difference in adsorption clearance at any convection volume setting. The mean adsorption clearance of IL-6 in all settings in PMMA was 32.4 mL/min (29.7–26.2), which was higher than that of IL-8 (P < 0.001). No significant differences in adsorption clearance according to the convection volume were observed for AN69ST. The mean adsorption clearance of IL-6 in all settings in AN69ST was 42.5 mL/min (35.4–49.8), which was lower than that of IL-8 (P < 0.05).

Discussion

Among the hemofilters used in CKRT, PMMA and AN69ST adsorb various proinflammatory cytokines such as IL-6 and IL-8 [10]. Unlike PS membranes, these membranes have lower water permeability. Therefore, continuous venovenous hemofiltration with CAHs is prone to fouling [13] and membrane coagulation due to platelet activation [14]. It is difficult to perform CKRT using CAHs in high filtration rate settings. In previous reports, the QF during CAH-CKRT with PMMA ranged from 300 to 500 mL/h, which was substantially lower than the convection volume recommended by the Kidney Disease Improving Global Outcomes [6]. Hence, there was not an adequate filtration rate at low convection volume in CAH-CKRT. In this study, the effects of QD and QF on the removal characteristics of small and large molecular weight substances were investigated in vitro using an experimental model of CAH-CKRT with test solutions. In conventional CKRT without CAHs, it was assumed that a higher filtration rate would increase the removal rate of the mediators. However, our results showed that the concentrations of IL-6 and IL-8 did not differ significantly between QD and QF. Unlike previous reports using filtration membranes [15], no relationship was found between the convection volume and the adsorption properties of high-molecular-weight mediators in this experiment. Therefore, it is not necessary to use a filtration modality to perform CAH-CKRT. In clinical settings, when CAH-CKRT is conducted, large molecular weight substances may be adequately removed, even under conditions that primarily remove small-molecular-weight substances.

Because the adsorption mechanism of PMMA involves occlusion of the pores of PMMA membranes, it is assumed that the adsorption capacity is promoted by the pressure caused by the filtration rate [13]. No significant differences were observed in the adsorption clearance of IL-6 and IL-8 after CAH-CKRT using PMMA in any setting. At low convection volumes, the internal filtration pressure derived from the flow of the test solution may have affected adsorption more than that of QF [16]. This may explain why our results differed from those of previous reports on CKRT with filtration membranes. The adsorption clearance of IL-6 (21 kDa) exceeded that of IL-8 (8 kDa) in PMMA under all conditions. The radius of IL-6 is 1.9 nm, whereas the membrane pore size of PMMA is 10 nm [17]. Therefore, the pore size of PMMA is suitable for IL-6 adsorption. These mechanisms may result in differences in IL-6 and IL-8 adsorption clearance rates.

The clearance of creatinine, a low molecular weight substance, increases with an increase in convection volume. However, in this study, there was no significant difference in creatinine clearance between QD and QF with the same convection volume. CHD, with a QD of 600 mL/h, is considered the most efficient modality for achieving both cytokine adsorption and creatinine removal in CAH-CKRT using PMMA. In addition, the adsorption efficiency of PMMA can be further enhanced by increasing the effectiveness of internal filtration, such as by elongating the shape of the housing.

No significant differences were observed in the adsorption clearance of IL-6 and IL-8 in any setting when CKRT was performed using AN69ST. As the bulk layer on the inner surface of AN69ST is negatively charged, positively charged substances can be adsorbed via electrostatic coupling. Proteins, including cytokines, generally contain both dissociative and polar functional groups. Their charge states change depending on the pH of the solvent or blood. The isoelectric point of IL-8 is 8.5, so it was considered to be positively charged in the test solution with the pH adjusted to 7.4 [18]. Therefore, IL-8 can be easily adsorbed onto the bulk layer of the inner surface of AN69ST. We found that the adsorption clearance of IL-8 was higher than the theoretical clearance by filtration, which is consistent with the aforementioned mechanism. The isoelectric point of IL-6 is 5.3, and it is negatively charged in the test solution adjusted to pH 7.4, suggesting that the adsorption capacity by electrostatic coupling is relatively small. The results of this study showed that adsorption clearance of IL-6 occurred even with CKRT using AN69ST. One reason for this is the formation of charged residues on the surface of the amino acids constituting IL-6. Electrostatic binding by the cations of these charged residues is assumed to have occurred [19]. Because the removal of large molecular weight substances, such as IL-6, from the filtrate rarely occurs, the clearance performance is considered to be due to adsorption.

In contrast, creatinine clearance in AN69ST was the highest in the 0/600 setting. The reason for the higher clearance in the 0/600 setting than that in the 600/0 setting at the same convection volume may be related to the membrane structure of AN69ST, which is a bulk layer containing water and not pores through which the material can pass directly, as in general filtration membranes. The bulk layer is a membrane structure that allows small molecules to pass through. This bulk layer is involved in the movement of small molecules, resulting in a high creatinine clearance value in the 0/600 setting, where the QF is the highest. However, creatinine clearance in the 0/600 setting is a very minor difference compared with that in the 600/0 setting, and is unlikely to affect treatment in clinical practice. Therefore, when using AN69ST for CAH-CKRT with the goal of removing IL-6 and IL-8, CHD with a QD of 600 mL/h may be appropriate for reducing the risk of membrane coagulation.

Sepsis involves cytokines and other mediators [15]. In CKRT for septic AKI, the adsorptive removal of large-molecular-weight substances may be sufficient when CAHs are used under conditions targeting small-molecular-weight clearance. Therefore, further clinical investigations are required.

Limitations

Because this was an in vitro comparative study using test solutions, the experiments were conducted under conditions in which no substances physically inhibited filtration or adsorption. In clinical practice, cells and proteins present in the blood reduce the efficiency of filtration and adsorption [16]. The possibility that the results of this study deviate from those of other studies when used in clinical settings cannot be ruled out. Clinical studies under similar conditions are necessary to confirm the results of this study. Moreover, because this study was conducted at a low convection volume, the results may not be applicable when the purification rate is increased.

Furthermore, this experiment was conducted over a short period of time. This does not indicate that the adsorption clearance reported here can always be obtained, because the actual treatment may last for a longer period of time, and the adsorption performance of the filter may deteriorate.

Conclusions

In this study, the effects of QD and QF on the removal of small- and large-molecular-weight substances during CAH-CKRT were investigated in vitro. At convection volumes below 600 mL/h, there was no relationship between the convection volume and the adsorption clearance of large molecular weight substances. CKRT settings based on QF may not be necessary for the removal of substances with high molecular weights, such as IL-6 and IL-8. Further in vivo studies are required to determine appropriate settings for CAH-CKRT.