Liver failure, whether fulminant or acute-on-chronic, can cause high mortality.1,2 Although liver transplantation is the preferred therapy for liver failure, a shortage of liver donors is prompting the development of extracorporeal artificial liver support systems.3 A loss of the detoxifying function of the liver causes the accumulation of toxins, and especially protein-bound toxins, in the human body and, consequently, multiple organ failure.4–6 Various approaches have been proposed to remove these toxins to bridge patients to liver transplantation, such as hemodialysis, hemofiltration, hemoperfusion, plasma perfusion, and plasma exchange.6,7 However, these techniques have limited application because of their respective disadvantages. For example, hemodialysis and hemofiltration are effective only for water-soluble toxins6,8; hemoperfusion and plasma perfusion may cause a loss of valuable substances from patients; hemoperfusion, plasma perfusion, and plasma exchange can result in serious immune problems in patients.9 In the 1990s, the development of albumin dialysis technology provided a new hope in the treatment of liver failure.10 Currently, this technology is also extended to the treatment of cholestatic patients with intractable pruritus11 and the preoperative management of jaundiced patients with hilar cholangiocarcinoma.12
Albumin dialysis, or bound solute dialysis,13–15 is a detoxifying method in which albumin solution is used as dialysate in a hemodialysis system (Figure 1): the blood containing toxins passes through inside of hollow fibers, whereas the albumin dialysate flows outside hollow fibers counter-currently. In the process, free toxins in blood are transferred to dialysate because of the transmembrane diffusion and convection, bind to albumin in dialysate, and are removed eventually. In the literature,16–18 the support system often runs in open-loop albumin dialysis mode (OLM; the dialysate containing albumin is discarded after passing through the dialyzer, as shown in Figure 1B). Open-loop albumin dialysis mode requires a large volume of albumin dialysate and causes a waste of expensive albumin. There are some other important studies on the support systems where the albumin dialysate operates in closed-loop mode,19,20 but in these systems, the removal of toxins are based on the adsorbent adsorption mechanism (albumin is only a carrier) not the albumin binding mechanism (albumin is not only a carrier but also a remover). Therefore, there is still interest in closed-loop albumin dialysis mode (CLM; the albumin dialysate flows through a closed system, as shown in Figure 1A) for removing protein-bound toxins only by albumin.5,16,21 Much meaningful theoretical work on the mass transfer model of albumin dialysis and the effect of ultrafiltration on albumin dialysis in OLM has been performed by Patzer et al.13–15 However, the detailed comparison between CLM and OLM has not been theoretically reported. Closed-loop albumin dialysis mode is a possible way to help to decrease the use of dialysate (or increase the device portability) and to realize the regeneration and reuse of albumin (or reduce the treatment cost). However, certain issues need to be clarified beforehand: the first issue is whether CLM is better than OLM in terms of the detoxification efficiency and the usage efficiency of albumin; the second is how we can reach a high efficiency of CLM in practice.
Therefore, in this work, CLM in albumin dialysis was investigated to clarify the above issues. Under a fixed amount of albumin in dialysate, CLM was experimentally compared with OLM in terms of the detoxification efficiency and the usage efficiency of albumin (the usage efficiency of albumin is the ratio of the amount of bilirubin removed from blood to the amount of albumin used in dialysate). In addition, the factors affecting the efficiency of CLM were also discussed with the help of a theoretical model that considers the detailed local ultrafiltration across the fiber membrane. The work in this study is of significance to the design of a portable artificial liver system.
Materials and Methods
Bovine serum albumin (Fraction V, molecular weight = 66,000 Da, ≥96%) and bilirubin (molecular weight = 584.66) were obtained from Sangon Biotech (Shanghai, China). Dimethyl sulfoxide (DMSO, ≥99.0%), sodium carbonate (Na2CO3, anhydrous, ≥99.8%), sodium chloride (NaCl, ≥99.5%), potassium dihydrogen phosphate (KH2PO4, ≥99.5%), disodium hydrogen phosphate dodecahydrate (NaH2PO4•12H2O, ≥99.0%), and potassium chloride (KCl, ≥99.5%) were from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). Hydrochloride acid (HCl, 1.0 N) was purchased from the Chemistry Store at the University of Science and Technology of China (Anhui, China). The two peristaltic pump systems used (BT100-2J) were obtained from LongerPump Co., Ltd. (Hebei, China). The hollow fiber dialyzer was a Gambro 6LR dialyzer (Gambro, Lakewood, CO).
Phosphate buffered saline (PBS) was prepared beforehand. The PBS was then used to prepare a solution containing 20 g/L albumin and 14.7, 17.7, or 21.4 mg/dl bilirubin in dim light as the blood. In the preparation,22 DMSO (4 ml) was added to the weight boat to dissolve the bilirubin powder, 0.25 M Na2CO3 (6 ml) solution was used to help dissolution, 1.0 N HCl solution was added drop by drop to adjust the pH value of the solution to approximately 7.2. The PBS was also used to prepare a solution containing 1, 2, or 20 g/L albumin as the dialysate.
Closed-Loop and Open-Loop Albumin Dialysis Mode Experiments
In CLM experiments (Figure 1A), the operating conditions were as follows: in blood side, the blood volume was 500 ml, the blood flow rate was set to 180 ml/min, the concentration of albumin in blood was 20 g/L, and the concentration of bilirubin in blood was 14.7, 17.7, or 21.4 mg/dl. In dialysate side, the dialysate volume was 500 ml, the dialysate flow rate was set to 180 ml/min, the concentration of albumin in dialysate was 20 g/L, and the net ultrafiltration rate was kept 0 ml/min.
In OLM experiments (Figure 1B), the operating conditions were as follows: in blood side, the blood volume was 500 ml, the blood flow rate was set to 180 ml/min, the albumin concentration in blood was 20 g/L, and the bilirubin concentration in blood was 14.7, 17.7, or 21.4 mg/dl. In dialysate side, the dialysate volume was 5,000 ml, the dialysate flow rate was set to 30 ml/min, the albumin concentration in dialysate was 2 g/L, and the net ultrafiltration rate was kept 0 ml/min.
Before experiments, the dialyzer was primed with the PBS for approximately 60 minutes. In experiments, the priming solution was removed at the beginning (the lumen priming solution was ~115 ml and the shell priming solution was ~110 ml), two magnetic stirrers were used to continuously mix blood and dialysate, and samples (200 μL) were collected from the blood reservoir at 0, 15, 30, 45, 60, 80, 100, 120, 140, 160, and 180 minutes. The experiments were conducted at room temperature (25°C) in dim light. In all experiments, the albumin dialysate was used only once.
In this test, all samples were analyzed using clinical biochemistry methods by a chemistry analyzer (Beckman Coulter AU480, Tokyo, Japan). In a test for albumin, the bromocresol green method was used, and in a test for bilirubin, the diazonium salt method was used.22 Each sample was tested three times and an averaged value was used.
To compare the transport of toxins in CLM and OLM, a theoretical mass transfer model was used, as shown in the Appendix. In this work, two key parameters, the total transmembrane mass transfer coefficient kA for bilirubin and the equilibrium binding constant of bilirubin to albumin, were estimated based on the experimental data, and they were approximately 8.0 × 102 ml/min and 0.8 × 107 L/mol, respectively. The two fitted parameters are reasonable according to the literature.15,23,24 The main forms of the bilirubin purchased were experimentally determined: 94% unconjugated bilirubin and 6% conjugated bilirubin. In simulation, the solution viscosity was assumed the same as the one of water at room temperature. The value of pressure at the inlet of the blood side was assumed 2,000 Pa, which does not affect the results. Other parameters used are listed in Table 1.
In this work, all boundary conditions are listed in Table 2. First, Equations 2–5 were solved by Runge–Kutta. A regula-falsi method14 was used to adjust Qd,in at z = L. Then, Equations 7–10 were solved by the same methods to converge Cs,tl,d,in at z = L (Cs,tl,d,in is zero in OLM, whereas variable in CLM). Finally, Equations 11 and 12 were used to obtain the toxin and albumin concentrations in reservoirs at each time step. A mesh refinement analysis was performed to obtain a proper space step so that the error caused by the grids was less than 0.5%.
Results and Discussion
The Comparison of the Detoxification Efficiency Between Closed-Loop and Open-Loop Albumin Dialysis Mode
The detoxification efficiency is a key parameter for assessing the performance of an artificial liver support system. In this work, under the operating condition of a fixed amount of albumin in dialysate (10 g), we performed a comparison of the detoxification efficiency between CLM and OLM (Figure 2, the parameter λ is the ratio of the bilirubin molar number to the albumin molar number in the blood). In CLM, the concentration of albumin in dialysate is usually high, and the dialysate, with a fixed small volume, can be always operated at a high flow rate. In OLM, the concentration of albumin in dialysate is usually low, and the dialysate, with a fixed large volume, only can be operated at a low flow rate for a long dialysis time. The results show that when CLM was used and the dialysis time was 3 hours, the concentrations of bilirubin in blood decreased from 14.7, 17.7, and 21.4 mg/dl to 7.7, 9.0, and 10.8 mg/dl, respectively. However, in OLM, the concentrations of bilirubin decreased only to 11.3, 12.7, and 13.9 mg/dl, respectively. The comparison indicates that according to the detoxification efficiency in the 3 hour dialysis, CLM is better than OLM.
In practice, the usage efficiency of albumin is another important indicator for assessing an artificial liver support system because of the high cost of albumin. In Figure 2, A–C, the usage efficiencies of albumin in CLM were 3.5, 4.3, and 5.3 mg/g, respectively, whereas in OLM, they were 1.7, 2.5, and 3.7 mg/g, respectively. Here, the unit mg/g means the amount of bilirubin removed from blood over the amount of albumin used in dialysate. The comparison indicates that in the 3 hour dialysis, the usage efficiency of albumin in CLM is higher. In addition, the results also imply that the advantage of CLM is less significant when the concentration of bilirubin in blood is higher (i.e., when λ is larger). The reason is that when λ is large, due to a fixed small volume in CLM, the increase of the bilirubin concentration in dialysate results in the rapid decrease of the transmembrane bilirubin concentration difference. In this work, some results were obtained according to the theoretical model appended, the validity of which was confirmed with the experimental data (circles and dots in Figures 2 and 7A). In Figure 2C, there is a difference between theoretical and experimental results, especially at the beginning in CLM. This deviation might be caused by the flow field change induced by the random movement of fibers. It might also result from the model precision because the model neglected the effects of some minor factors on toxin transport, such as the inlet and outlet structure, the dialysate flow misdistribution, the nonuniform membrane property, and the boundary layer.
The Effects of Albumin Concentrations in Closed-Loop and Open-Loop Albumin Dialysis Mode
In albumin dialysis, albumin is a carrier for albumin-bound bilirubin, and thus, the concentration of albumin in dialysate has substantial effects on the bilirubin clearance. In OLM, the concentration of albumin in dialysate is usually low, and the volume of dialysate is large; as a result, the regeneration of albumin is not convenient. Closed-loop albumin dialysis mode can compensate for this shortcoming, but the ensuing question is whether it is better to use a small volume of highly concentrated albumin solution as dialysate if the amount of albumin is kept constant (under this situation, the concentration of albumin in dialysate increases as the volume of dialysate decreases).
Figures 3 and 4 show the comparison of the effect of the albumin concentration in dialysate between OLM and CLM when the amount of albumin used in dialysate and the flow rate of dialysate were fixed. The results indicate that in CLM, a higher concentration of albumin in dialysate can result in a lower concentration of bilirubin in blood (Figures 3A and 4A). That is, toxins can be removed more rapidly when the concentration of albumin in dialysate is higher. If the albumin concentration in dialysate is higher, more unbound bilirubin molecules will be bound to albumin molecules in dialysate, generating a higher driving force for the transport of unbound bilirubin molecules from the blood side to the dialysate side. Therefore, in CLM, for a given amount of albumin in dialysate, the highly concentrated albumin dialysate results in the fast removal of toxins. Here, the normalized bilirubin concentration in Figures 3–8 is the ratio of the real-time bilirubin concentration to the initial bilirubin concentration in blood.
For a given amount of albumin in dialysate, in OLM, a higher concentration of albumin in dialysate can result in a faster decrease of the concentration of bilirubin in blood at the beginning (Figures 3B and 4B). However, due to the dialysate volume decrease caused by the increase in the albumin concentration, the dialysis time is shortened, and thus, the toxin clearance is lower. In addition, the comparison between Figures 3 and 4 shows that the effect of the albumin concentration in dialysate on the detoxification efficiency is more significant when the bilirubin concentration is higher (when λ is larger). The results presented here imply that it is promising to use CLM as a substitute for OLM.
The Effects of Dialysate Flow Rates in Closed-Loop and Open-Loop Albumin Dialysis Mode
In the process of hemodialysis, to increase the clearance of the uremic toxins, one method is to increase the dialysate flow rate. Here, one still can take the same solution as above to remove the albumin-bound toxins quickly. However, it is important to note that the effect of the above method sometimes is limited. For example, in CLM, the increase in the detoxification efficiency when the dialysate flow rate increases from 30 to 90 ml/min is significant whereas not when the dialysate flow rate increases from 90 to 180 ml/min (Figures 5A and 6A). In addition, our results also show that the above method becomes more efficient when the bilirubin concentration is higher (Figure 5Avs. Figure 6A).
For a fixed amount of albumin in dialysate, in OLM, the concentration of albumin in dialysate determinates the volume of dialysate. Under this situation, if the dialysis time is determinate, the flow rate of dialysate is also determinate. Then, if the flow rate of dialysate increases, the dialysis time will decrease, and thus, the clearance of toxins will drop (Figures 5B and 6B). For OLM, to increase the flow rate of dialysate and not reduce the fixed dialysis time, one has to decrease the concentration of albumin in dialysate accordingly so that the volume of dialysate is sufficient for the dialysis process. This to some extent can increase the detoxification efficiency, especially when the concentration of bilirubin in blood is high (Figure 7Avs. Figure 7B).
The Effects of the Volume of Blood and the Amount of Albumin in Dialysate in Closed-Loop and Open-Loop Albumin Dialysis Mode
In the above analysis, the conclusions drawn are based on a small volume of blood. Then, what are the results of the comparison between CLM and OLM when the volume of blood is large? Our results show that when the volume of blood is large, under a fixed small amount of albumin in dialysate, both CLM and OLM cannot offer adequate binding sites for the unbound bilirubin from blood. However, due to the fresh albumin dialysate always through the dialyzer, OLM can provide a larger continuous driving force for bilirubin. Therefore, the detoxification efficiency of OLM may be more than that of CLM, but is still very low (Figure 8A). To increase the detoxification efficiency of OLM, one has to increase the quantity of albumin in dialysate. However, as the quantity of albumin in dialysate increases, the detoxification efficiency of CLM will be quickly more than that of OLM (Figure 8B). Therefore, compared with OLM, CLM is more promising.
Albumin dialysis has been clinically proven effective and widely used in extracorporeal artificial liver support systems. However, the need for a large quantity of expensive albumin motivates researchers to find not only a more efficient but also a more economical operating mode for artificial liver systems.
In this work, the albumin dialysate was operated in two modes: CLM and OLM. The detoxification efficiencies and the usage efficiencies of albumin in dialysate in both modes were compared experimentally and theoretically. The results show that in terms of the detoxification efficiency and the usage efficiency of albumin in the 3 hour dialysis for removing albumin-bound bilirubin in blood, CLM is better than OLM, especially when the concentration of bilirubin in blood is not so high.
In practice, under a given quantity of albumin, one may further increase the performance of CLM by means of increasing the flow rate of the albumin dialysate or using the highly concentrated albumin dialysate. However, the effect of the above methods sometimes is limited, especially when the concentration of toxins in blood is low. In addition, under a fixed small amount of albumin in dialysate, the performance of OLM may approach or even surpass that of CLM in the 3 hour dialysis when the bilirubin concentration in blood or the volume of blood is very high. However, as the amount of albumin in dialysate increases, the detoxification efficiency of CLM will be quickly more than that of OLM. For the sake of the cost of albumin and the convenience of albumin regeneration, CLM is also highly recommended. As a concentrated albumin solution is used, the volume of the dialysate is reduced, which will enable significant improvements in the design of portable extracorporeal liver supporting systems. It should be noted that this study is fully based on the results obtained in the artificial environment that is different from the clinical albumin dialysis. In clinical practice, one needs to take into consideration the patient’s actual situation (e.g., the volume of blood, the generation of toxins, and the type of toxins) and then choose the proper treatment plan (e.g., the volume of dialysate, the amount of albumin used in dialysate, and the flow rate of dialysate) so that toxins can be removed quickly.
1. Ellis A, Wendon J. Circulatory, respiratory, cerebral, and renal derangements in acute liver failure: Pathophysiology and management. Semin Liver Dis. 1996;16:379–388
2. Vienken J, Christmann H. How can liver toxins be removed? Filtration and adsorption with the Prometheus system. Ther Apher Dial. 2006;10:125–131
3. McCashland TM, Shaw BW Jr, Tape E. The American experience with transplantation for acute liver failure. Semin Liver Dis. 1996;16:427–433
4. Lee UE, Friedman SL. Mechanisms of hepatic fibrogenesis. Best Pract Res Clin Gastroenterol. 2011;25:195–206
5. Sen S, Mookerjee RP, Davies NA, Williams R, Jalan R. Review article: The molecular adsorbents recirculating system (MARS) in liver failure. Aliment Pharmacol Ther. 2002;16(Suppl 5):32–38
6. Stange J. Extracorporeal liver support. Organogenesis. 2011;7:64–73
7. Dunlop EH, Weston MJ, Gazzard BG, Langley PG, Mellon PJ, Williams R. Artificial liver support based on haemoperfusion of adsorbents. Biomed Eng. 1975;10:213–218
8. Bouré T, Vanholder R. Biochemical and clinical evidence for uremic toxicity. Artif Organs. 2004;28:248–253
9. Rozga J, Umehara Y, Trofimenko A, Sadahiro T, Demetriou AA. A novel plasma filtration therapy for hepatic failure: Preclinical studies. Ther Apher Dial. 2006;10:138–144
10. Kreymann B, Seige M, Schweigart U, Kopp KF, Classen M. Albumin dialysis
: Effective removal of copper in a patient with fulminant Wilson disease and successful bridging to liver transplantation: A new possibility for the elimination of protein-bound toxins. J Hepatol. 1999;31:1080–1085
11. Leckie P, Tritto G, Mookerjee R, Davies N, Jones D, Jalan R. ‘Out-patient’ albumin dialysis
for cholestatic patients with intractable pruritus. Aliment Pharmacol Ther. 2012;35:696–704
12. Regimbeau JM, Fuks D, Chapuis-Roux E, et al. Addition of Molecular Adsorbent Recirculating System (MARS(®)) Albumin Dialysis
for the Preoperative Management of Jaundiced Patients with Hilar Cholangiocarcinoma. Case Rep Gastroenterol. 2013;7:396–403
13. Patzer JF 2nd, Bane SE. Bound solute dialysis. ASAIO J. 2003;49:271–281
14. Patzer JF 2nd, Safta SA, Miller RH. Slow continuous ultrafiltration with bound solute dialysis. ASAIO J. 2006;52:47–58
15. Patzer J. Principles of bound solute dialysis. Ther Apher Dial. 2006;10:118–124
16. Sauer IM, Goetz M, Steffen I, et al. In vitro
comparison of the molecular adsorbent recirculation system (MARS) and single-pass albumin dialysis
(SPAD). Hepatology. 2004;39:1408–1414
17. Collins KL, Roberts EA, Adeli K, Bohn D, Harvey EA. Single pass albumin dialysis
(SPAD) in fulminant Wilsonian liver failure: A case report. Pediatr Nephrol. 2008;23:1013–1016
18. Boonsrirat U, Tiranathanagul K, Srisawat N, et al. Effective bilirubin
reduction by single-pass albumin dialysis
in liver failure. Artif Organs. 2009;33:648–653
19. Magosso E, Ursino M, Colì L, Baraldi O, Bolondi L, Stefoni S. A modeling study of bilirubin
kinetics during Molecular Adsorbent Recirculating System sessions. Artif Organs. 2006;30:285–300
20. Patzer I, John F. Thermodynamic considerations in solid adsorption of bound solutes for patient support in liver failure. Artif Organs. 2008;32:499–508
21. Krisper P, Stauber RE. Technology insight: Artificial extracorporeal liver support–how does Prometheus compare with MARS? Nat Clin Pract Nephrol. 2007;3:267–276
22. Ding W, Zou L, Sun S, Li W, Gao D. A new method to increase the adsorption of protein-bound toxins in artificial liver support systems. Artif Organs. 2014;38:954–962
23. Meyer TW, Walther JL, Pagtalunan ME, et al. The clearance of protein-bound solutes by hemofiltration and hemodiafiltration. Kidney Int. 2005;68:867–877
24. Weisiger RA, Ostrow JD, Koehler RK, et al. Affinity of human serum albumin for bilirubin
varies with albumin concentration and buffer composition: Results of a novel ultrafiltration method. J Biol Chem. 2001;276:29953–29960
25. Rohacova J, Sastre G, Marin ML, Miranda MA. Dansyl labeling to modulate the relative affinity of bile acids for the binding sites of human serum albumin. J Phys Chem B. 2011;115:10518–10524
26. Wanwimolruk S, Birkett DJ, Brooks PM. Structural requirements for drug binding to site II on human serum albumin. Mol Pharmacol. 1983;24:458–463
27. Ding W, Zhou X, Heimfeld S, Reems JA, Gao D. A steady-state mass transfer
model of removing CPAs from cryopreserved blood with hollow fiber modules. J Biomech Eng. 2010;132:011002
28. Legallais C, Catapano G, Harten Bv, Baurmeister U:. A theoretical model to predict the in vitro performance of hemodiafilters. J Membr Sci. 2000;168:3–15
29. Villarroel F, Klein E, Holland F. Solute flux in hemodialysis and hemofiltration membranes. Trans Am Soc Artif Intern Organs. 1977;23:225–233
30. Benavente J, Jonsson G. Electrokinetic characterization of composite membranes: Estimation of different electrical contributions in pressure induced potential measured across reverse osmosis membranes. J Membr Sci. 2000;172:189–197
Appendix: Theoretical Modeling
In albumin dialysis, albumin is the key to removing protein-bound toxins. Based on the literature,25,26 albumin has more than one site for binding toxins. However, to simplify the issue studied in this study, we assumed that albumin has only one binding site for bilirubin.13,14 The association between albumin and bilirubin can be described as follows:
where a · s denotes the albumin–bilirubin compound, s denotes the free or unbound bilirubin, a denotes the free albumin, KB denotes the equilibrium binding constant (
), and C denotes the concentration. For a given association constant, the above binding kinetic can be used to obtain the relationship of the concentration of free bilirubin and the total concentration of bilirubin.
To make the theoretical model as simple as possible for practical use, we assumed that the effects of the inlet and outlet structure, the dialysate flow misdistribution, the nonuniform membrane property, and the boundary layer on toxin transport are negligible. Transmembrane diffusion and convection are two driving forces for transporting free toxins during albumin dialysis. For a traditional hemodialyzer, diffusive mass transfer is dominant, whereas for currently used hemodiafilters, convective mass transfer is dominant. Transmembrane convection, i.e., local ultrafiltration, is strongly affected by the membrane structure, the blood flow rate, and the dialysate flow rate, and this convection varies along the axial direction. In the pioneering work presented by Patzer et al.,13–15 local ultrafiltration was assumed to be even and equal to the net ultrafiltration divided by the fiber length. Here, we comprehensively considered the effect of detailed local ultrafiltration. Regarding the calculation of local ultrafiltration, the following equations can be used27,28:
where the subscripts b and d denote the blood and dialysate sides, respectively; ri and ro are the inner and outer radii of hollow fibers, respectively. RM is the inner radius of the dialyzer; LP is the hydraulic permeability of the membrane; P is the pressure; Q is the flow rate; μ is the viscosity; η is the fiber number; L is the effective fiber length; and Jv is the local ultrafiltration rate.
In this work, we assumed that only free bilirubin molecules cross membrane, blood, and dialysate are well mixed along the dialyzer,13–15 and for a high-flux dialyzer, the osmotic/oncotic pressure difference is negligible compared with the hydraulic pressure difference. In practice, if the osmotic pressure difference (Δπ) is considered, the right side of Equation 6 should be changed to
Considering the effect of local ultrafiltration, based on mass conservation, the transport of the unbound bilirubin can be described as follows14,29:
Based on the binding kinetic, albumin and bilirubin concentrations on both sides of hollow fibers can be described as follows:
where Cs is the concentration of free bilirubin; Cs,tl is the total concentration of bilirubin; Ca,tl is the total concentration of albumin; K is the mass transfer coefficient; A is the total membrane area; f is a function of the Pe number14,29:
; and σ is the reflection coefficient.
In CLM, the total concentrations of bilirubin and albumin in the blood and dialysate reservoirs can be written as
where the subscripts in and out denote the reservoir inlet and outlet, respectively; the subscripts br and dr denote the blood and dialysate reservoirs, respectively; and V is the volume.