The effectiveness of hemodialysis therapy for the treatment of chronic renal failure is determined in part by the mass transfer characteristics of the dialysis device (i.e., the hemodialyzer/artificial kidney). 1–4 Uremia has been characterized by the retention of numerous types of solutes. 5 The most widely studied group of uremic markers are the low molecular weight (MW) nitrogenous waste products, such as urea and creatinine. These molecules have high diffusivity because of their low MW and lack of protein binding. 6,7 Diffusive mass transport through the hemodialyzer membranes depends upon diffusion and partition coefficients, number and size of membrane pore, membrane thickness, and difference in solute concentration across the membrane. An increase in solute concentration gradient across the membrane will increase the rate of solute transport. To increase the concentration gradient, one needs to reduce the solute concentration in the dialysate flow. This may be achieved by increasing the dialysate flow rate. However, studies have shown that the solute concentration varies at a cross-section of the dialysate flow. 8 The MRI studies of the present authors, 9 as well as numeric simulation, showed that the flow in dialysate compartments is not uniform (Figure 1). Although spacing filaments or curved hollow fibers (Moiré structure) were used in the dialyzers to prevent the dialysate flow maldistribution, 10–12 wide diameter channels and some gaps in the packing structure may still be present and result in nonuniform dialysate flow distribution, called the “channeling phenomena.”13 However, effect of the channeling upon the solute clearance at different locations in a dialyzer has not been directly and quantitatively examined. In this report, a novel experiment approach is presented to test the hypothesis that hollow fibers at different regions in a given hemodialyzer may contribute differently to the solute clearance by measuring the local clearance of urea (MW 60) and creatinine (MW 113) using two high flux dialyzers.
Materials and Methods
Dialyzers and Solutes
The authors tested dialyzer A (cellulose triacetate hollow fibers; surface area 1.9 m2; ultrafiltration coefficient Kuf = 36 ml/hr/mm Hg) and dialyzer B (polyethersulfone hollow fibers; surface area 1.6 m2; Kuf = 73 ml/hr/mm Hg).
The solutes used for this study were urea (MW 60 Dalton) and creatinine (MW 113 Dalton). The solute concentration in “blood” (water solution with urea and creatinine) was 46 mg/dl urea and 15 mg/dl creatinine. Hemodialysis experiments and clearance measurements were performed with the experimental settings showed in Figure 2. Single pass of the “blood” and dialysate (water) flows were used in experiments.
Preparation of Dialyzers with Segmented Annular Rings
Before the experiment, the hemodialyzer’s manifold was opened, and different annular rings were drawn on surface of the potting material (at the blood inlet) in the dialyzer header. First, adhesive tape was used to cover the entire flat surface of the potting material. Concentric annular rings with equal surface areas were drawn on the surface. All the tape, except for the selected ring, was then removed, leaving one annular ring area still covered by tape (ring tape). An epoxy resin was used to seal or block the hollow fiber openings in all areas that were uncovered by the tape. After the epoxy dried, the remaining ring tape was removed (fibers in this ring were unblocked and open to blood flow). The process above was repeated on new hemodialyzers to make different concentric annular rings. Details of the concentric annular ring arrangement in each hemodialyzer and the corresponding micropicture are shown in the following figures:Figure 3 shows the three concentric annular rings for dialyzer A with sealed and blocked regions, Figure 4 shows four rings for the dialyzer B, and Figure 5 shows rings with fibers openings unblocked or blocked by epoxy resin.
Three A dialyzers and four B dialyzers were used for the experiments. It was assumed that the fibers were uniformly distributed in a dialyzer. Therefore, concentric rings with equal surface areas had equal numbers of hollow fibers. For dialyzer A, we divided the header surface into three equal area concentric rings with the radii shown in Figure 3 (area of each concentric ring is approximately 320 mm2). For dialyzer B, because its plastic housing diameter was larger than dialyzer A, the authors divided the header surface into four equal area concentric rings, with radii shown in Figure 4 (area of each concentric ring is approximately 443 mm2).
The experiments were performed using two single pass circuits (Figure 2). The “blood” from a reservoir on a balance was pumped through the unblocked hollow fibers in the dialyzer and returned to a different reservoir on the same balance. Any decrease in blood mass was attributed to ultrafiltration flux into the dialysate flow. The ultrafiltration rate was set by adjusting the pressure drop along the fiber bundle. Deionized water was used as dialysate in experiments. The primary component of reservoir was 4 L deionized water.
To simulate a clinical condition, the blood side flow rate was set to 120 ml/min passing through the unblocked fibers. The equivalent blood flow rate (if all fibers in the three rings are unblocked) for dialyzer A is approximately 3 × 120 = 360 ml/min; for dialyzer B (if all fibers in four rings are unblocked), it is approximately 4 × 120 = 480 ml/min (this is to simulate the typical blood flow rate 300 ml/min ∼ 500 ml/min for clinical use). The deionized water based dialysate was set to 500 ml/min, 800 ml/min, and 1,000 ml/min, respectively, to evaluate the effects of dialysate flow upon the solute clearance.
A balance was used to monitor the ultrafiltration rate of 5 ml/min by adjusting a clamp on the tubing to achieve desired transmembrane pressure (Figure 2).
Data Collection and Analysis
Solution samples at blood compartment inlet and outlet were collected at different times. The urea nitrogen and creatinine concentration in the samples were determined using Cobas-Mira (Roche Diagnostics, Somerville, NJ) calorimetric assays.
The clearance (C) was calculated by using the following equation. EQUATION
where Qbi is the blood inlet flow rate (ml/min), Cbi is the solute concentration at the blood flow inlet (mg/dl), Cbo is the solute concentration at the blood flow outlet (mg/dl), and QUFR is the ultrafiltration rate (ml/min).
A Student Newman Keuls post hoc test was performed to determine significances of and differences between dialysate flow rates and between ring positions, as well as the interaction between the dialysate flow rate and the ring position. The analysis was performed using SPSS for Windows version 11.
As indicated in Materials and Methods, the “blood” flow rate was fixed in this study. The dialysate flow rates were 500, 800, and 1,000 ml/min, respectively. Two types of hemodialyzers were used. Urea and creatinine clearance attributed by different dialyzer rings was determined at each of three blood-dialysate flow combinations. The solute concentration data at blood flow outlet and the clearance data are shown in Tables 1 and 2 for dialyzer A and Tables 3 and 4 for dialyzer B. For a given dialyzer, effects of ring/location (Position) and dialysate flow rate (Flow) upon the solute clearance were studied by analyzing the data in these tables. Position was given a numeric value of 1–3 for dialyzer A and 1–4 for dialyzer B.
The analyses for dialyzer A show the following:
- With urea as the solute, both the flow and the position have significant effects upon the solute clearance, and there is an interaction between the position and flow. There is little difference between rings 1 and 2, but ring 3 exhibits higher values. The effect of flow is significant at all three ring positions, showing as much as a 10% increase for average clearances between the highest and lowest flows tested. These effects are diagrammed in Figure 6. The error bars show the 95% confidence limits for each position at each of the three flow rates, and it is apparent that the means of rings 1 and 2 are close at all three dialysate flow rates.
- With creatinine as the test solute, a similar pattern is observed. Both effects of the position and flow are significant, and there is a significant interaction between the two. However, with creatinine, the small difference between rings 1 and 2 is distinguishable, perhaps because of the smaller errors in the replicate measurements. As with the urea, there is as much as a 10% difference between the lowest and highest flow rates. The results with the 95% confidence limits are diagrammed in Figure 7.
The analyses for dialyzer B show the following:
- With urea as the test solute, there is a significant difference in clearance caused by both position and flow rate. There is no significant interaction between the two. The inner ring produces significantly lower clearances than the outer three, but the differences between the outer three are not statistically significant. The data are summarized in Figure 8.
- With creatinine as the test solute, both position and flow rate have a significant effect upon clearance, and there is an interaction between these two variables. For the pooled flow rates, there is nearly a 15% increase in clearance as one goes from the core annulus to the outer annulus, compared with only a 5% increase in clearance as flow is increased from 500 ml/min to 1,000 ml/min. The effect of flow rate is more important to the core annulus than to the outer rings, as shown in Figure 9.
The lower clearance in core region of the dialyzers may be caused by the channeling of the dialysate flow or by the reported facts 8–13 that dialysate flow rate is lower at the core region of hollow fiber bundle because of the fiber resistance to the dialysate flow in the radial direction (i.e., dialysate flow must pass the layer of fibers before reaching the core region). As a result, the solute clearance at core region was lower. A faster dialysate flow was generated by increasing hydraulic pressure at dialysate flow inlet. The high pressure can not only generate high flow rate but can also push more dialysate flow into the core region. This may explain why the clearance increases with a dialysate flow rate increase. The different effects of dialysate flow rate upon the solute clearance in dialyzer A and dialyzer B may be attributed by the different membrane material and hollow fibers bundle arrangement in these two dialyzers.
The understanding that hollow fibers and dialysate flows at different regions within the same hemodialyzer contribute differently to the overall clearance is important to future attempts to design more efficient dialyzers. The method described in this study appeared to be very useful in evaluating effects of the position and the flow upon the local clearance. The following conclusions were made from this study:
- At different annular regions inside the dialyzers, the local clearance can be significantly different, with the lowest values measured at the core annulus of both dialyzers. The implication is that dialysate does not perfuse the fibers uniformly, perhaps because the fiber bundles are not distributed uniformly throughout the housing, or because there are associations between fibers that cause channeling.
- Increasing the dialysate flow rate can increase dialyzer clearances, but this increase is not experienced uniformly over the cross-section of the dialyzer. Again, this may reflect differences of fiber packing densities across the cross-section of the device.
- The effect of increasing dialysate flow rate on the solute clearance was greater for dialyzer B than for dialyzer A.
This work was supported, in part, by Baxter Health Care Corporation (McGaw Park, IL), as well as by the One-Hundred Talent Program and the ChangJiang Scholar Program of China. The authors thank Dr. Michael Brier of University of Louisville for his help in statistical analysis.
1. Klein E, Holland F, Lebeouf A, Donnaud A, Smith JK: Transport and mechanical properties of hemodialysis hollow fibers. J Membrane Science 1: 371–396, 1976.
2. Sternby J, Jonsson S, Ledebo I: Hemodiafiltration: technical aspects. Contrib Nephrol 96: 86–98, 1992.
3. Leypoldt JK: Dialyzers and dialysis membranes research directions. ASAIO J 42: 133–135, 1996.
4. Jaffrin MY: Convective mass transfer in hemodialysis. Artif Organs 19: 1162–1171, 1995.
5. Langsdorf L, Zydney A: Diffusive and convective solute transport through hemodialysis membranes: a hydrodynamic analysis. J Biomed Mater Res 28: 573–582, 1994.
6. Vanholder R, Smet R: Pathophysiologic effects of uremic retention solutes. J Am Soc Nephrol 10: 1815–1823, 1999.
7. Clark WR, Shinaberger JH: Effect of dialysate-side mass transfer resistance on small solute removal in hemodialysis. Blood Purif 18: 260–263, 2000.
8. Frank A, Lipscomb GG, Dennis M: Visualization of concentration fields in hemodialyzers by computed tomography. J Membrane Science 175: 239–251, 2000.
9. Poh CK: Characterization of fluid transport in artificial kidneys. Master thesis. University of Kentucky, 2000.
10. Ronco C, Scabardi M, Brendolan A, Crepaldi C, Lagreca G: Impact of spacing filaments external to hollow fibers on dialysate flow distribution and dialyzer performance. Int J Artif Organs 5: 261–266, 1997.
11. Vander Velde C, Leonard EF: Theoretical assessment of the effect of flow maldistributions on the mass efficiency of artificial organs. Med Boil Eng Comput 23: 224–229, 1985.
12. Ronco C, Ballestri M: New developments in hemodialyzers. Blood Purif 18: 267–275, 2000.
13. Ronco C, Brendolan A, Crepaldi C, Rodighiero M, Scabardi M: Blood and dialysate flow distributions in hollow-fiber hemodialyzer analyzed by computerized helical scanning technique, J Am Soc Nephrol
13(Suppl): 53–61, 2002.