Ultrafiltration rate as a dose surrogate in pre-dilution hemofiltration

Transcription

1 The International Journal of Artificial Organs / Vol. 30 / no. 2, 2007 / pp Artificial Kidney and Dialysis Ultrafiltration rate as a dose surrogate in pre-dilution hemofiltration Z. HUANG 1, J.J. LETTERI 2, W.R. CLARK 2,3, W. ZHANG 4, D. GAO 5,6, C. RONCO 7 1 Department of Mechanical Engineering, Widener University, Chester, PA - USA 2 Gambro Renal Products, Lakewood, CO - USA 3 Nephrology Division, Indiana University School of Medicine, Indianapolis, IN - USA 4 Renal Division, Renji Hospital, Shanghai Second Medical University, Shanghai - China 5 Department of Mechanical Engineering, University of Washington, Seattle, WA - USA 6 Department of Mechanical Engineering, University of Kentucky, Lexington, KY - USA 7 Nephrology Department, St. Bortolo Hospital, Vicenza - Italy ABSTRACT: For critically ill patients treated with continuous hemofiltration (HF), doses recently shown to improve survival can usually be achieved only in the pre-dilution mode. However, use of the pre-dilution mode results in reduced treatment efficiency, relative to post-dilution at the same ultrafiltration rate (Q f ) and blood flow rate (Q b ). The objective of this study is to determine the effect of Q f on removal parameters for solutes over a wide molecular weight spectrum in pre-dilution HF. Experiments were performed in an isovolemic, plasma-based pre-dilution system with Q b =200 ml/min. Removal parameters were measured for a 1.2 m 2 polysulfone hemofilter (HF1200, Minntech) at Q f values of 20, 40, and 60 ml/min, corresponding to 17, 34 and 51 ml/h/kg for a 70 kg patient (N=3 hemofilters for each Q f ). Clearance of urea and creatinine (small solute surrogates) was derived from plasma and ultrafiltrate concentrations at 30, 60, 120, 180, and 240 min while clearance of vancomycin and inulin (middle molecule surrogates) was estimated from changes in plasma concentrations over time. In addition, the sieving coefficient (SC) of vancomycin and inulin was measured at the same time points and at baseline (T=0 min). Our findings indicate pre-dilution had a predictable effect on clearance for each solute, as clearance increased linearly with Q f. Sieving coefficient values were not significantly influenced by either Q f or time and the equivalence of SC values in the middle molecule range suggest attenuation of secondary membrane effects. These data indicate filter performance can largely be preserved despite high Q f values by use of predilution. Moreover, Q f appears to be a reasonable dose surrogate in pre-dilution HF. (Int J Artif Organs 2007; 30: ) KEY WORDS: Hemofiltration, Solute, Ultrafiltration, Dose, Dilution INTRODUCTION Hemofiltration (HF) was originally applied in the context of end-stage renal disease (ESRD) (1) and early clinical evidence indicated its superiority over hemodialysis (HD) with respect to both cardiovascular stability (2) and clearance of relatively large molecular weight uremic toxins (3). However, the broadest use of HF is currently for critically ill patients with acute renal failure (ARF) (4, 5). In ARF, HF is usually delivered on a continuous basis and is referred to as continuous venovenous HF (CVVH), which is one modality in the spectrum of continuous renal replacement therapy (CRRT) (6, 7). The continuous therapies are preferred by many clinicians over standard intermittent HD in patients with significant hemodynamic instability (8). Relative to standard HD, the major benefits of CRRT relate to its overall greater volume removal and dose delivery capabilities (8-13), despite the relatively / $15.00/0 Wichtig Editore, 2007

2 Huang et al slow rate at which volume and solute removal occur. CRRT is now the most popular dialysis treatment for critically ill ARF patients in Europe and Asia (14). At least until recently, the ultrafiltration rate (Q f ) in CVVH has typically been in the 1-2 l/h range (5, 10, 12). However, in response to recent outcome data published by Ronco and colleagues (15), prescription of significantly higher Q f values is occurring. These investigators reported a direct relationship between daily ultrafiltrate volume and survival in critically ill patients treated with CVVH. Daily ultrafiltrate volumes of 55 L or more (on average) were associated with a 30-day mortality of approximately 50% while more standard ultrafiltrate volumes (mean 31 l/d) were associated with a 30-day mortality of approximately 65%. Therapy in which an Q f of greater than 50 l/d is prescribed has recently been termed high volume hemofiltration (16). In addition to the Ronco study, several other lines of clinical evidence suggest the application of convective CRRT influences patient outcome favorably, especially when applied early (17-19). The primary solute removal mechanism in HF is convection (20). In post-dilution HF, the mode employed in the Ronco study, the relationship between solute clearance and Q f is quite straightforward (21). In this situation, solute clearance is determined primarily by and related directly to the solute s sieving coefficient (SC) and the Q f (22). (Sieving coefficient is defined as the ratio of the solute concentration in the filtrate to the simultaneous plasma concentration.) Consequently, the concept from the Ronco study that ultrafiltrate volume is a surrogate for treatment dose (i.e., solute clearance) is reasonable. However, post-dilution HF is limited inherently by the attainable blood flow rate (Q b ) and the patient s hematocrit (Hct). More specifically, the ratio of Q f to the plasma flow rate delivered to the filter, termed the filtration fraction (FF), is the limiting factor (23): Previous studies suggest FF values > 20-30% in post dilution may be undesirable due to hemoconcentrationrelated effects on filter performance (24). A fundamental difference between post-dilution and pre-dilution CVVH is the fact the latter is not filtration fraction-constrained. Therefore, the pre-dilution mode avoids post-dilution s hemoconcentration-related effects on hemofilter performance and is being increasingly used [1] for CVVH therapy. However, the above mass transfer benefits must be weighed against the predictable dilution-induced reduction in plasma solute concentrations, one of the driving forces for convective solute removal (25, 26). The extent to which this reduction occurs is determined mainly by the ratio of replacement fluid rate to Q b (27, 28). For reasons described above, the relationship between clearance and Q f may not be as predictable in predilution, relative to the case of post-dilution. Consequently, the claim that Q f is a dose surrogate in pre-dilution HF needs to be demonstrated. To this end, the purpose of the present study is to investigate the effect of Q f on solute removal parameters in an experimental pre-dilution CVVH system. These parameters are measured for solutes of varying molecular weight (MW). MATERIALS AND METHODS Hemofilter and solutes The hemofilter used for all experiments was HF1200 (Minntech Renal Systems, Minneapolis, MN), having a 1.25 m 2 surface area membrane composed of polysulfone and a water permeability (K uf ) of 37 ml/h/mmhg. A wide MW spectrum of solutes was investigated (29). These solutes were urea (60 Dalton [Da]), creatinine (112 Da), vancomycin (1,448 Da), and inulin (5,200 Da). The first two solutes are small solute surrogates while the latter two are middle molecule surrogates. Six liters of bovine plasma were used as the blood compartment test solution for each experiment. The duration of each experiment was 240 minutes. Three experiments (N=3) were performed for each Q f value; a new hemofilter was used for each experiment. Experimental setup The closed-loop experimental system is shown in Figure 1 (29). In this system, a conventional hemodialysis machine (550, Baxter Healthcare Co,. McGaw Park IL) was used for blood (plasma) recirculation and ultrafiltration control. Consistent with clinical CVVH, a Q b of 200 ml/min was used. Ultrafiltration rate was set at 20 ml/min, 40 ml/min or 60 ml/min, equating to 28.8, 57.6, or 125

3 Therapy dose in pre-dilution hemofiltration equations appear as Equations [2] and [3] below, respectively (21): [2] [3] Fig. 1 - Experimental pre-dilution hemofiltration system l/d, respectively, for continuous operation. The plasma-containing reservoir was situated on a heated stirrer to maintain the experimental temperature at 37 o C ± 1 o C. The entire reservoir and heated stirrer were placed on a balance (EA35EDE-1, Sartorius Corp. NY) to allow for continuous weighing of the reservoir. Because a net Q f of zero (i.e., zero-balance) was employed in all experiments, the absolute Q f was equal to the reinfusion (substitution fluid) rate. Zero-balance was ensured by maintaining the weight of the reservoir constant. The desired Q f was achieved by manipulations in transmembrane pressure through the use of manually adjusted clamps on the arterial and venous tubing. Fresh reinfusion solution (acetate-based dialysate) was prepared just before each experiment and its flow rate maintained by a precision pump (Fluid Metering, INC, Syosset, NY). Data collection and analysis For small solute (urea and creatinine) clearance determinations, simultaneous blood (arterial and venous lines) and ultrafiltrate samples were obtained at five different time points during each experiment: 30, 60, 120, 180, and 240 min. At each time point, the clearance determination represents the mean of the blood-side and ultrafiltrate-side values (see below). For each experiment, the reported small solute clearances are the mean of instantaneous clearances measured at the above time points. The blood-side and ultrafiltrate-side clearance In these equations, Cl is clearance (ml/min); Q bi is blood (plasma) flow rate at inlet (ml/min); C bi is solute concentration at blood (plasma) inlet (mg/dl); C bo is solute concentration at blood (plasma) outlet (mg/dl); and C f is solute concentration in the ultrafiltrate (mg/dl). For the middle molecular surrogates (vancomycin and inulin), the clearance calculation is based on the reduction of concentration in the reservoir over the experimental time. The equation is as follows: In this case, Q f refers to net Q f. Since the experiments were zero-balanced, the equation reduces to: where b is slope of the regression line of experimental time vs natural logarithm of solute concentration, t the elapsed time (minute), and V 0 the volume of reservoir at time t 0 (ml) (29). The SC of vancomycin and inulin was also measured at 0, 30, 60, 120, 180, and 240 minutes and calculated by the following equation: Urea nitrogen, creatinine, and vancomycin concentrations were measured in duplicate by the Cobas Mira-S autoanalyzer (Roche Diagnostic System, Inc. Somerville, NJ). Inulin concentration was measured spectrophotometrically (DU 640, Beckman Instruments, Inc., Schaumburg, IL). Student s t-test was used for statistical comparisons of the effect of Q f on clearance and SC while analysis of variance (ANOVA) was employed to assess the effect of Q f and time on clearance and SC. Results are expressed as mean ± SD (N=3). Differences were considered statistically significant at P < 0.05 level. [4] [5] [6] 126

4 Huang et al Clearance (ml/min) Clearance (ml/min) Time (minutes) Ultrafiltration rate (ml/min) Fig. 2 - Urea clearance (ml/min) as a function of time (ultrafiltration rate = 20 ml/min). Fig. 3 - Solute clearance (ml/min) as a function of ultrafiltration rate (ml/min). RESULTS During each experiment, clearance determinations for small solutes were made at a series of time points. In Figure 2, urea clearance data from a representative experiment are shown as a function of time. No significant decrement in urea clearance is observed over the 240 min experimental period. Therefore, subsequently reported clearance data for both urea and creatinine represent the mean value for all five time points during a given experiment. The relationship between solute clearance and Q f for urea, creatinine, vancomycin and inulin appears in Figure 3. Overall, these data are consistent with a convective therapy for two reasons. First, for each solute, the clearance-q f relationship is linear, confirming a direct relationship between these two parameters. Second, for a given Q f over the solute MW range investigated, clearance is not strongly dependent on molecular weight, at least in comparison to hemodialysis. Specifically, very little difference in clearance is observed between the two small solutes and between the two middle molecule surrogates as a function of Q f. On the other hand, reflecting its diffusive basis, HD is associated with much larger differences in clearance over the same MW range (30). That the linearity of the relationship between clearance and Q f is preserved over the entire Q f range suggests the use of a relatively high Q f does not impair hemofilter performance in pre-dilution HF. As Q f increases, polarization effects (31) and solutemembrane interactions (32) may result in impaired filter importance. This may manifest as a reduction in hydraulic permeability or more importantly in clearance, especially of larger molecular weight substances. The use of pre-dilution appears to attenuate polarization and secondary membrane effects and is the most likely explanation for the observed findings. The sieving properties of the polysulfone hemofilter at different ultrafiltration rates were also assessed. The SC values for middle MW solutes as a function of experimental time appear in Figure 4. No significant SC changes were observed over the 240 min treatment time for either of these solutes. Moreover, except for minor differences observed at certain time points with inulin, ultrafiltration rate also has no significant impact on the sieving coefficient values for these solutes. These results differ from those observed previously in post-dilution hemofiltration (22) and provide further evidence membrane performance is relatively preserved with the use of pre-dilution. A relatively predictable effect of Q f on solute clearance was observed in the experimental predilution CVVH system over a wide MW spectrum of test 127

5 Therapy dose in pre-dilution hemofiltration Sieving coefficient Qs = 20 ml/min Qs = 40 ml/min Qs = 60 ml/min Sieving coefficient Qs = 20 ml/min Qs = 40 ml/min Qs = 60 ml/min Time (minutes) Time (minutes) Fig. 4 - Sieving coefficient as a function of time for vancomycin (a) and inulin (b). solutes. From a theoretical perspective, the efficiency of pre-dilution CVVH can be evaluated by introducing a dilution factor D f (33): where Q s is reinfusion (substitution) flow rate (ml/min). Under isovolemic conditions (Q s = Q f ), this parameter represents the ratio of the pre-dilution clearance to postdilution clearance at a given Q f value. For the Q f values of 20, 40, and 60 ml/min used in the present study, the dilution factor D f is estimated to be 91%, 83%, and 77%, respectively. Based on the assumption of a SC of unity for urea and creatinine, equation [8] can be used to estimate solute clearance: For this approach, estimated urea and creatinine clearance values are 18.2, 33.2, and, 46.2 ml/min at Q uf values of 20, 40, and 60 ml/min, respectively. These estimates compare favorably with the experimentally derived values (Tab. I). The close agreement between experimental and predicted results suggests an orderly relationship between Q f and solute clearance (dose) in pre-dilution CVVH, confirming the former is a reasonable surrogate for the latter in this therapy mode. [7] [8] Two caveats apply to the above analysis. First, it should be emphasized that the predictable relationship between Q f and dose exists at a given blood flow rate. An increase in blood flow rate attenuates the substitution fluid s dilution effect while a decrease in blood flow rate has the opposite influence (26, 27). Second, the current study was performed with plasma rather than whole blood as the blood compartment fluid. For solutes such as urea and creatinine, the additional distribution volume in the intracellular space and the rate of solute transfer across red blood cell membranes are further considerations in such an analysis. Finally, a useful application of our data is to develop a model which predicts pre-dilution clearance of a specific non-adsorbing solute as a TABLE I - SOLUTE CLEARANCE (ml/min) AT DIFFERENT ULTRAFILTRATION RATES Solutes Ultrafiltration rate (ml/min) Urea 18.3± ±1.6* 42.3±4.6* Creatinine 17.6± ±2.1* 42.5±2.9* Vancomycin 15.3± ±2.2* 37.4±0.9* inulin 16.7± ±1.4* 37.9±2.4* *: P< 0.01 vs. 20 (ml/min). 128

6 Huang et al Clearance (ml/min) function of SC and Q f. To achieve this, equation [8] was used for various combinations of Q b and Q s over a clinically relevant Q f range of 0 to 100 ml/min. For the Q f range evaluated, the relationship between clearance and ultrafiltration rate is essentially linear (Fig. 5). However, the slope of each line is determined by the specific combination of Q b and SC. Specifically, at a fixed Q b, solute SC and slope are directly correlated. Therefore, at a given Q f, a higher SC is associated with a higher solute clearance. Likewise, for a given solute SC, Q f and slope are also directly correlated, manifesting as higher solute clearance with increasing Q f. DISCUSSION Ultrafiltration rate (ml/min) Fig. 5 - Predicted relationship between clearance and ultrafiltration rate for solutes of varying sieving coefficient. Ronco et al s recent study establishing an association between post-dilution CVVH dose and outcome has led to attempts to extrapolate their data to other forms of CRRT. Although both post-dilution and pre-dilution CVVH employ convection as the solute removal mechanism, the operating parameters and solute removal capabilities of the hemofilter may differ significantly for these two therapies. Both the water and solute permeability of an ultrafiltration membrane are influenced by the phenomena of concentration polarization (31) and secondary membrane formation (34, 35). The exposure of an artificial surface to plasma results in the non-specific, instantaneous adsorption of a layer of proteins, the composition of which generally reflects that of the plasma itself. Therefore, plasma proteins such as albumin, fibrinogen, and immunoglobulins form the bulk of this secondary membrane. Moreover, the plasma total protein concentration also influences this phenomenon. This layer of proteins, by serving as an additional resistance to mass transfer, effectively reduces both the water and solute permeability of an extracorporeal membrane. Evidence of this is found in comparisons of solute sieving coefficients determined before and after exposure of a membrane to plasma or other protein-containing solution (36-38). Although concentration polarization primarily pertains to plasma proteins, it is distinct from secondary membrane formation. Concentration polarization specifically relates to ultrafiltration-based processes and applies to the kinetic behavior of an individual solute. Accumulation of a solute that is predominantly or completely rejected by a membrane used for ultrafiltration of plasma occurs at the blood compartment membrane surface. This surface accumulation causes the solute concentration just adjacent to the membrane surface (i.e., the submembranous concentration) to be higher than the bulk (plasma) concentration. Conditions which promote the polarization process are high Q f (high rate of convective transport), low blood flow rate (low shear rate or membrane sweeping effect), and the use of post-dilution (rather than pre-dilution) replacement fluids (increased local solute concentrations) (21). For a given solute and rate of replacement fluid administration, post-dilution HF provides higher solute clearance than does pre-dilution HF, provided filter operation is unimpaired (20). As discussed below, the relative inefficiency of the latter mode is related to the dilution-related reduction in solute concentrations, which decreases the driving force for convective mass transfer. Despite its superior efficiency with respect to replacement fluid utilization, in post-dilution HF filtration fraction is a significant constraint, as discussed above. At high filtration fraction values, concentration polarization and secondary membrane effects may become prominent 129

7 Therapy dose in pre-dilution hemofiltration with the resultant possibility of compromised hemofilter performance. From a mass transfer perspective, the use of predilution has several potential advantages over postdilution. First, both hematocrit and blood total protein concentration are reduced significantly prior to the entry of blood into the hemofilter. This effective reduction in the red cell and protein content of the blood attenuates the secondary membrane and concentration polarization phenomena described above, resulting in improved mass transfer (39). Predilution also favorably impacts mass transfer due to augmented flow in the blood compartment, because pre-filter mixing of blood and replacement fluid occurs. This achieves a relatively high membrane shear rate, which also reduces solute-membrane interactions. Finally, pre-dilution may also enhance mass transfer for some compounds by creating concentration gradients that induce solute movement out of red blood cells (39). Because the pre-dilution mode is not filtration fraction-limited, high solute clearances can be achieved by prescription of a relatively high Q f, which can overcome the dilutionrelated loss of efficiency. The potentially detrimental impact of secondary membrane and concentration polarization effects is directly proportional to solute MW. Our experimental data suggest these phenomena are largely attenuated by the use of pre-dilution. One indication of this is the preservation of sieving coefficients as a function of time for both middle molecule surrogates (vancomycin and inulin). Another indication is that, in a manner similar to that observed for the small solutes, middle molecule clearances are a linear function of Q f. It should be emphasized that neither of the middle molecule surrogates nor the polysulfone membrane used in this study are known to be particularly hydrophobic. Therefore, adsorption effects most likely are not an important consideration. Our study indicates hemofilter performance can be preserved under high Q f conditions in pre-dilution CVVH. Moreover, because an orderly relationship exists between Q f and solute clearance, our data suggest Q f is a reasonable dose surrogate in predilution CVVH, as has been suggested for the postdilution mode. Finally, because Q f is not constrained by filtration fraction, the pre-dilution mode appears to be a rational approach for delivering high-dose CVVH. As has been emphasized, both pre-dilution and post-dilution CVVH have distinct advantages and disadvantages. In order to exploit the advantages and minimize the disadvantages of each mode, simultaneous pre-dilution and post-dilution ( mixed dilution ) HF and hemodiafiltration have been applied for ESRD patients (40, 41). In a similar manner, some newer-generation CRRT machines are capable of providing mixed-dilution therapy (42) and this may be the most attractive option for delivering high-dose CVVH in the future. Irrespective of the infusion mode, to achieve high dose delivery, blood flow rates higher than those traditionally used in CRRT will need to be applied. CONCLUSION Address for correspondence: Zhongping Huang, PhD Department of Mechanical Engineering Widener University Chester, PA USA zhuang@mail.widener.edu 130

8 Huang et al REFERENCES 1. Henderson LW, Besarab A, Michaels AS, Bluemle LW. Blood purification by ultrafiltration and fluid replacement (diafiltration). Trans Am Soc Artif Intern Organs 1967; 13: Baldamus CA, Ernst W, Frei U, Koch KM. Sympathetic and hemodynamic response to volume removal during different forms of renal replacement therapy. Nephron 1982; 31: Colton CK, Henderson LW, Ford C, Lysaght MJ. Kinetics of hemodiafiltration. I. In vitro transport characteristics of a hollow-fiber blood ultrafilter. J Lab Clin Med 1975; 85: Ronco C, Bellomo R: Continuous renal replacement therapy: Evaluation in technology and current nomenclature. Kidney Int 1998; 53 (suppl 66): S Macias WL, Mueller BA, Scarim SK, Robinson M, DW Rudy. Continuous veno-venous hemofiltration: An alternative to continuous arteriovenous hemofiltration and hemo-diafiltration in acute renal failure. Am J Kidney Dis 1991; 18: Clark, WR, Ronco C. Continuous renal replacement techniques. Contrib Nephrol 2004; 144: Clark WR, Ronco C. Renal replacement therapy in acute renal failure: Solute removal mechanisms and dose quantification. Kidney Int 1998; 53 (suppl 66): S Bellomo R, Ronco C. Continuous versus intermittent renal replacement therapy in the intensive care unit. Kidney Int 1998; 53 (suppl 66): S Clark WR, Mueller BA, Kraus MA, Macias WL. Dialysis prescription and kinetics in acute renal failure. Adv Ren Replace Ther 1997; 4: Clark WR, Mueller BA, Alaka KJ, Macias WL. A comparison of metabolic control by continuous and intermittent therapies in acute renal failure. J Am Soc Nephrol 1994; 4: Clark WR, Mueller BA, Kraus MA, Macias WL. Solute control by extracorporeal therapies in acute renal failure. Am J Kidney Dis 1996; 28 (suppl 3): S Clark WR, Mueller BA, Kraus MA, Macias WL. Extracorporeal therapy requirements for patients with acute renal failure. J Am Soc Nephrol 1997; 8: Liao Z, Zhang W, Poh CK, Huang Z, Hardy PA, Kraus MA, Clark WR, Gao D. Kinetic comparison of different acute dialysis therapies. Artif Organs 2003; 27: Mehta RL, Letteri JM. Current status of renal replacement therapy for acute renal failure. A survey of US nephrologists. The National Kidney Foundation Council on Dialysis. Am J Nephrol 1999; 19: Ronco C, Bellomo R, Homel P, Brendolan A, Dan M, Piccinni P, La Greca G. Effects of different doses in continuous veno-venous haemofiltration on outcomes of acute renal failure. A prospective randomised trial. Lancet 2000; 356: Tetta C, Bellomo R, Kellum J, et al. High volume hemofiltration in critically ill patients: Why, when, and how? Contrib Nephrol 2004; 144: Honore PM, Jamez J, Wauthier M, et al. Prospective evaluation of short-term, high-volume isovolemic hemofiltration on the hemodynamic course and outcome in patients with intractable circulatory failure resulting from septic shock. Crit Care Med 2000; 28: Piccinni P, Dan M, Barbacini S, Carraro R, Lieta E, Marafon S, Zamparetti N, Brendolan A, D Intini V, Tetta C, Bellomo R, Ronco C. Early isovolaemic haemofiltration in oliguric patients with septic shock. Intensive Care Med 2006; 32: Page B, Viellard-Baron A, Chergui K, et al. Early venovenous haemofiltration for sepsis-related multiple organ failure. Crit Care 2005; 9: R Clark WR, Ronco C. CRRT efficiency and efficacy in relation to solute size. Kidney Int 1999; 56 (suppl 72): S3-S Henderson LW. Pre vs post dilution hemofiltration. Clin Nephrol 1979; 11: Henderson LW. Biophysics of ultrafiltration and hemofiltration. In: Jucobs C, ed. Replacement of renal function by dialysis (4th ed). Dortdrecht: Kluwer Academic Publishers, 1996; Huang Z, Henderson LW, Gao D, Clark WR. Hemofiltration and hemodiafiltration for end-stage renal disease. In: Chronic kidney disease: Dialysis and transplantation. Amsterdam, The Netherlands: Elsevier Saunders; 2004 p Bellomo R, Ronco R: Continuous renal replacement therapy in the intensive care unit. Intensive Care Med 1999; 25: Troyanov S, Cardinal J, Geadah D, et al. Solute clearances during continuous venvenous haemofiltration at various ultrafiltration flow rates using Multiflow-100 and HF1000 filters. Nephrol Dial Transplant 2003; 18: Clark WR, Turk JE, Kraus MA, Gao D. Dose determinants in continuous renal replacement therapy. Artif Organs 2003; 27: Ahuja A, Rodby R, Huang Z, Gao D, Zhang W, Clark WR. Effect of pre-dilution replacement fluid administration on solute clearance in high-volume hemofiltration (Abstract). J Am Soc Nephrol 2003; 14: 734A. 28. Ledebo I. Principles and practice of hemofiltration and hemodiafiltration. Artif Organs 1998; 22: Scott MK, Mueller BA, Clark WR. Dialyzer-dependent 131

9 Therapy dose in pre-dilution hemofiltration changes in solute and water permeability with bleach reprocessing. Am J Kidney Dis 1999; 33: Henderson LW, Colton CK, Ford CA, Lysaght MJ. Kinetics of hemodiafiltration. II. Clinical characterization of a new blood cleansing modality. J Lab Clin Med 1975; 85: Kim S. Characteristics of protein removal in hemodiafiltration. Contrib Nephrol 1994; 108: Clark WR, Hamburger RJ, Lysaght MJ. Effect of membrane composition and structure on performance and biocompatibility in hemodialysis. Kidney Int 1999; 56: Zhang W, Huang Z, Ahuja A, Rodby R, Clark WR, Gao D. Blood flow rate effects in high-dose pre-dilution CVVH (Abstract). J Am Soc Nephrol 2003; 14: 742A. 34. Rockel A, Hertel J, Fiegel P, Abdelhamid S, Panitz N, Walb D. Permeability and secondary membrane formation of a high flux polysufone hemofilter. Kidney Int 1986; 30: Clark WR, Macias WL, Molitoris BA, Wang NHL. Plasma protein adsorption to highly permeable hemodialysis membranes. Kidney Int 1995; 48: Ofsthun NJ, Zydney AL. Importance of convection in artificial kidney treatment. Contrib Nephrol 1994; 108: Langsdorf LJ, Zydney AL. Effect of blood contact on the transport properties of hemodialysis membranes: A twolayer model. Blood Purif 1994; 12: Morti SM, Zydney AL. Protein-membrane interactions during hemodialysis: Effects on solute transport. ASAIO J 1998; 44: Ofsthun NJ, Colton CK, Lysaght MJ. Determinants of fluid and solute removal rates during hemofiltration. In: Henderson LW, Quellhorst EA, Baldamus CA, Lysaght MJ, eds. Hemofiltration. Berlin: Springer-Verlag, 1986; Pedrini LA, De Cristafaro V. On-line mixed hemodiafiltration with a feedback for ultrafiltration control: Effect on middle-molecule removal. Kidney Int 2003; 64: Canaud B, Levesque R, Krieter D, et al. On-line hemodiafiltration as routine treatment of end-stage renal failure: Why pre- or mixed dilution mode is necessary in on-line hemodiafiltration today? Blood Purif 2004; 22 (suppl 2): S Salvatori, G, Ricci Z, Bonello M, et al. First clinical trial for a new CRRT machine: The Prismaflex. Int J Artif Organs. 2004; 27: