A Two-Filter Adaptation to Achieve Enhanced Hemodialysis Performance

Abstract

Hemodialysis (HD) technology, pivotal in managing end-stage kidney disease, has witnessed significant advancements. Yet, the high cost of novel equipment often restricts its usage in resource-limited settings. This study introduces a two-filter adaptation to conventional HD machines, aimed at enhancing toxin removal while maintaining cost-effectiveness. Using a benchtop experimental setup, the performance of the adapted system was compared with that of standard HD. The results demonstrated that the two-filter system improved urea clearance rates by 54% compared with standard HD, without increasing albumin loss or causing additional hemolysis. In a pilot study of four HD patients, the modified setup achieved a higher single-pool Kt/V (1.82) and urea-reduction ratio (80%). These findings underscore the potential of this adaptation to enhance HD machine efficiency without additional patient risks, thereby offering a feasible solution for improving access to advanced renal therapies in under-resourced areas. Further clinical trials with larger populations are warranted to validate these benefits and evaluate middle-molecule clearance for comparison with hemodiafiltration (HDF).

1. Introduction

Since Kolff’s pioneering work in the 1940s with rotating cellophane tubing, hemodialysis (HD) technology has continually evolved to address the significant morbidity and mortality associated with end-stage kidney disease (ESKD) patients [1,2]. Recent advancements in dialyzer technology have enhanced the distribution of dialysate flow within the dialysate compartment. These improvements were shown to enhance urea clearance without increasing dialysate flow rates [3]. In the context of limited healthcare budgets, it is not always feasible to purchase advanced dialyzers or invest in new equipment; therefore, cost-effective methods to enhance existing machines are highly desirable, especially in areas with constrained resources.

In pursuit of more equitable access to advanced renal replacement therapies in resource-limited settings, we introduce a novel method for reconfiguring conventional HD machines to improve elimination of uremic solutes. Double-filter HD circuits have been explored previously, including the convective-controlled double high-flux hemodiafiltration (HDF) setup by Pisitkun et al. which utilizes back-filtration through serial dialyzers within the blood line [4]. Here, we describe a simpler two-filter, post-dilution substitution configuration on a standard HD machine. The configuration includes an additional ‘secondary filter’ using standard connectors and tubing. The dialysate inflow to the ‘secondary filter’ is split into two streams: one for dialysis and ultrafiltration, and another for fluid substitution. The modified setup was evaluated with urea clearance, albumin loss, and hemolysis levels using benchtop models with human blood. A pilot clinical study in HD patients was performed to evaluate dialysis adequacy using single-pool Kt/V and urea-reduction ratio (URR).

2. Materials and Methods

2.1. Bench Model

Bench models of HD and the two-filter adaptation setups are shown in Figure 1. The hemodialyzers, both the ‘primary filter’ and the ‘secondary filter’, used in this study were the Elisio™ 17H (Nipro, Bridgewater, NJ, USA). The tubing used was Tygon S3TM E-3603 (Saint-Gobain, Akron, OH, USA). Human whole blood purchased from commercial supplier BioChemed for research purposes was used for the bench model testing. The blood and dialysate were circulated using two Hei-FLOW Core 600 peristaltic pumps (Heidolph, Schwabach, Germany). A Manostat 72-310-000 (Barrington, IL, USA) served as the ultrafiltration pump. All pumps operated on a peristaltic mechanism. Four flow meters were placed in accordance with the SURDIAL^TM 55 HD machine (Nipro, Bridgewater, NJ, USA). The flow was adjusted manually using these three pumps based on the desirable flow parameters (

Table 1. Target parameters for conventional HD and the two-filter configuration.

Parameter	Value
Dialysate flow rate (QD)	600 mL/min
Blood flow rate (QB)	400 mL/min
Substitution rate (QS)	44 mL/min
Ultrafiltration rate (Quf)	8.3 mL/min

) prior to dialysis and a small adjustment was made during the dialysis to match the parameters closely.

Citrated human whole blood was acquired from BioChemed Services and spiked with urea to reach a concentration of 30 mM prior to each dialysis experiment, during which a citrate-based, bicarbonate-buffer dialysate was used to dialyze 500 mL urea-spiked human whole blood for 12 min. Blood and dialysate samples (~1 mL) were collected in intervals at 0, 1.5, 3, 6, 9 min. Additional dialysate samples (~1.5 mL) were collected at the beginning and the end of the experiment (0 and 12 min). Plasma was obtained from the collected whole blood samples after centrifugation at 2000× g for 10 min and analyzed for urea concentration via a urease/Berthelot reagent-based urea assay. Human serum albumin (HSA) in the dialysate samples was analyzed using an enzyme-linked immunosorbent assay (Abcam, Eugene, OR, USA). The percentage hemolysis of blood samples was determined via spectrophotometry. Three independent sessions were performed for the bench test.

Considering the blood is in a closed-loop system in a much smaller volume (500 mL) and the blood is flowing at a rate of 400 mL/min, the clinical urea clearance formula which involves blood urea nitrogen (BUN) levels before and after dialysis is not suitable.

$$K=Q_B\times {\displaystyle \frac{{BUN}_{arterial}-{BUN}_{venous}}{{BUN}_{arterial}}}$$

(1)

To adjust for the volume of blood in a closed-loop system affecting urea clearance, Chu derived a formula that models physiological concentrations and clearance [5].

$${\displaystyle \frac{d(V_B\times C_B)}{dt}}=G-K\times C_B$$

(2)

In this formula, $V_B$ represents total blood volume, $C_B$ denotes the urea concentration in the blood, $t$ is the time, $G$ represents urea generation rate, and $K$ represents the clearance, either by the kidneys or through a hemodialyzer. In the HD unit, the volume remains consistent over time and urea is not produced within the HD system. Solving the differential equation yields:

$$K=-{\displaystyle \frac{V_B}{t}}ln{\displaystyle \frac{C_{B_t}}{C_{B_0}}}$$

(3)

where $C_{B_0}$ represents the initial urea concentration in blood and $C_{B_t}$ represents urea concentration at $t$. To reduce the influence of unstable initial mixing of the bench model, $K$ can be estimated over a predefined early window $t$ after stabilization.

2.2. Clinical Evaluation

A pilot clinical evaluation was conducted to test the efficacy of the novel setup. Four patients undergoing HD at the Institute of Specialized Renal Care (Instituto de Atención Renal Especializada IARE) in Quito, Ecuador, participated in this study. Participants included two men and two women with a mean age of 49.5 years. The duration of their maintenance dialysis therapy prior to this study averaged 10.25 years, ranging from 2 to 20 years.

The patients first underwent conventional HD therapy in January 2013, followed by conventional HDF from January to March 2013. From April to July 2013, patients received the two-filter configuration treatment. Two patients were treated using a short daily regimen consisting of 2 h sessions, 5 days per week. The other two patients received a regimen consisting of 4 h sessions, 3 times per week.

Hemodialysis was performed using a Nipro hemodialysis system with a single blood pump and volumetric ultrafiltration control (Figure 2). A citrate-based, sodium bicarbonate-buffer dialysate at a sodium concentration of 137 mEq/L was used. The secondary dialyzer partitions the dialysate to the dialysate inlet of the primary dialyzer and downstream of the primary dialyzer. The flow was manually adjusted, and the replacement volume was determined using a visual indicator in the venous chamber. A clear liquid ring forms at the top of the chamber, where its thickness directly correlates with the replacement rate. A ring of 1 cm represents a 100 mL/min replacement rate with a maximum rate of 200 mL/min.

3. Results

3.1. Bench Model Performance

In the bench test, as shown in Figure 3, the blood urea level in the two-filter configuration demonstrated a steeper reduction in the first 2 min compared to the conventional HD setup. Urea clearance (K) numbers calculated using Equation (3) were 165.41 ± 22.38 mL/min for the HD configuration and 254.90 ± 12.17 mL/min for the two-filter configuration (

Table 2. Summary of performance of conventional HD and two-filter configuration.

Performance	Conventional HD	Two-Filter Configuration	p-Value *
Urea Clearance	165.41 ± 22.38 mL/min	254.90 ± 12.17 mL/min	0.003
Albumin Loss	0.23 ± 0.025 µg/mL	0.06 ± 0.036 µg/mL	0.014
Percentage Change in Hemolysis	12.26 ± 16.7%	8.87 ± 9.15%	0.78

* Paired, two-tailed t-test.

). The URRs were determined to be 64.25% and 83.67% for the HD and two-filter configurations, respectively. The urea clearance of the two-filter configuration is notably higher, comparable to the performance of HDF, which typically leads to a urea clearance 10–15% higher than that achieved by the diffusive mode alone [6].

One of the major concerns about the removal of uremic toxins by convection is the accompanied albumin loss [7,8]. Therefore, albumin loss during our dialysis experiments was monitored by measuring the HSA concentrations in the dialysate before and after dialysis. As shown in Table 2, HSA concentrations in the dialysate were around 0.23 ± 0.025 µg/mL after dialysis using the HD configuration and around 0.06 ± 0.036 µg/mL in the two-filter configuration. These values seem to be comparable and also negligible in comparison to the approximate 35 g/L in the blood pre-dialysis. As there exists increased pressure difference due to ultrafiltration with the two-filter adaptation, percentage hemolysis was also assessed. A slight increase in percentage hemolysis was observed in both configurations, as shown in Figure 4 and listed in Table 2. The noticeable increase in percentage hemolysis is likely attributed to the use of non-sterile dialysate, plastic particles from tubing, as well as the positive pressure (>70 psi) generated by the peristaltic pumps [9,10,11]. Nonetheless, there is no significant difference between the HD and two-filter configurations with respect to percentage hemolysis, suggesting that the modified configuration did not increase the possibility of hemolysis.

3.2. Clinical Performance

In the clinical study, dialysis adequacy was assessed using Kt/V and URR (

Table 3. Comparison of hemodialysis efficiency between conventional and two-filter system.

	Conventional HD	Conventional Hemodiafiltration (HDF)	Two-Filter Configuration
Kt/V	1.60	1.70	1.82
URR (%)	70	75	80

). During the conventional HD treatment phase, patients had a mean Kt/V of 1.6 and a URR of 70%. After transitioning to conventional HDF therapy, these values increased to 1.7 and 75%, respectively. Following implementation of the two-filter configuration setup, further improvement was observed, with a Kt/V of 1.82 and a URR of 80%. All three modalities exceeded the National Kidney Foundation’s recommended target of 1.4 per session for thrice-weekly treatment [12]. The URRs for all modalities needed to exceed the 65% limit to be considered adequate [13,14].

4. Discussion

This study evaluated the adequacy of a novel dialysis configuration through benchtop simulation and clinical testing. The setup is structurally similar to post-dilution HDF but implemented on a standard hemodialysis machine without the need for an additional substitution fluid pump.

During standard HD operation, a single dialyzer (primary filter) is connected to a HD machine, as shown in Figure 1B. In this study, we utilized a benchtop replication setup based on the operation of HD machines, in particular SURDIAL 55, to evaluate the performance and safety of a two-filter adaptation, as shown in Figure 1A, which involves attaching two dialyzers (i.e., primary and secondary filters) to a SURDIAL 55 externally using standard tubing and connectors. While the benchtop setup necessitates manual modulation of the peristaltic pump to achieve the target ultrafiltration rate, the SURDIAL 55 has built-in volumetric control by ceramic pumps with user-defined ultrafiltration parameters. Setting appropriate flow rates on the HD machine and partitioning dialysate flow using the ‘secondary filter’ allows for an enhanced rate of ultrafiltration than in a typical HD configuration. By partitioning a portion of the dialysate inflow for substitution, the actual dialysate outflow is lower than the dialysate outflow expected by the machine. As such, the dialysate is pumped out at a flow rate higher than anticipated, resulting in a higher actual ultrafiltration rate. While it is possible to increase the ultrafiltration rate manually in conventional HD, this may result in dehydration in the patient if increased excessively [15]. With substitution fluid infused into the blood, this would not be an issue.

The Elisio 17H dialyzer has an in vitro KoA of 1614 mL/min, resulting in a urea clearance of approximately 340 mL/min at a Q_b of 400 mL/min and a Q_d of 600 mL/min [16]. The discrepancy between the bench model results and the dialyzer performance parameter is primarily due to the differences in experimental setup as the manufacturer’s values are calculated based on single-pass measurements between input and output, using a standardized aqueous solution according to the European Norm (EN 1283); whereas our study used a closed-loop system with 500 mL of whole blood continuously cycled through the circuit [17]. It is challenging to compare clearance values directly. The main purpose here is to compare the relative performance between the two setups under identical conditions.

Under the exact same operative conditions, the increased urea clearance of the two-filter setup compared to the single-filter setup (conventional HD) suggests that the two-filter adaptation provides improved clearance of urea and likely other small solutes [18]. As a standard assessment of dialysis efficiency, the results confirmed that the two-filter configuration is highly efficient [13,14]. The bench models focus on comparing the urea clearance efficiency between the two setups, rather than aligning K_urea values with existing literature. This is due to the calculation method for K_urea, which differs from traditional approaches. Our method considers the limitations in blood volume available for testing and the sampling approach in a closed-loop circulation system. In this system, the entire blood volume of 500 mL is repeatedly cycled through for dialysis, which takes around 1.25 min per cycle at a blood flow rate of 400 mL/min, making accurate post-dialysis urea concentration measurements challenging.

Albumin loss and percentage hemolysis were evaluated as preliminary safety indicators. Albumin loss and percentage hemolysis did not differ between configurations in short benchtop runs. Because the benchtop test used stored whole blood, and the obstruction in blood flow may lead to hemolysis as red blood cells experience high pressure, the results should be interpreted as relative comparisons rather than absolute safety metrics. In clinical settings, an elevated blood flow rate is not reported to result in hemolysis, provided that the needle size is sufficiently large and the needles are positioned correctly [11,15]. In this study, the percentage hemolysis close to 10% pre-dialysis (Figure 4) is likely attributable to the condition of the purchased whole blood, which had been in storage and shipment for several days prior to the experiment and was therefore more susceptible to hemolysis [19].

While the study did not assess the removal of endotoxins and bacterial products by the secondary filter, it may offer additional benefits by adsorbing and filtering endotoxins and bacterial contaminants from the dialysate, potentially reducing the risk of contamination and dialyzer fouling. Further investigations are needed to evaluate these effects [20,21].

A clinical study was conducted to assess the feasibility of the two-filter configuration in practice. The two-filter configuration achieved the highest mean single-pool Kt/V and URR among HD and conventional HDF. The increased clearance observed is consistent with the bench results and supports the hypothesis that a convenient two-filter configuration on existing HD machines can enhance hemodialysis adequacy.

5. Conclusions

This work shows that a simple two-filter, post-dilution substitution configuration operated on an unmodified HD platform can yield improvement in small-solute removal at the proof-of-concept stage. Evidence comes from a benchtop model for within-model comparisons and a pilot clinical study of four maintenance HD patients reporting higher Kt/V and URR values. Given the small clinical sample and short exposure, these findings should not be interpreted as having efficacy or safety equivalence with standard HDF.

Clinical applicability requires future work on safety, middle-molecule clearance and biocompatibility over routine session durations with direct benching against post-dilution HDF. The main contribution of the present work is the engineering feasibility that convection can be generated on HD-only hardware without additional pumps or software. If subsequent patient studies confirm efficacy and safety, this approach could provide a practical alternative when HDF is not routinely available.