Design analyses of a new selective filtration system for removal of uremic toxins.

A new system for removal of uremic toxins is described, The system is based on selective removal of small (urea, creatinine) and large uremic toxins through a multimembrane filter/dialyzer unit. Theoretical calculations show that the system is capable of removing the desired range of molecules from uremic plasma, while retaining a large portion of non-toxic elements in the blood. The SElective DUal Filtration ARtificial Kidney can be optimized and assembled into a single compact unit, which is simple to use and requires minimal monitoring equipment. Preliminary clinical trials indicate that SEDUFARK is a feasible system for treatment of uremic patients.

Transport of uremic toxins through conventional hemodialysis membranes.

Human mononuclear phagocytes cultured in vitro were tested after preincubation with uremic plasma dialyzed in vitro and the effects of pre and post hemodialysis plasma were compared with the effect of dialyzates equilibrated with uremic plasma in vivo. Factors in uremic blood plasma inhibiting phagocytosis of Candida albicans could not be removed by means of conventional hemodialysis with cuprophan (CPN) or polyacrylonitrile (PAN) membranes. Hemofiltrates of uremic blood plasma produced by PAN membranes caused a significant inhibition of phagocytes in our test system while CPN hemofiltrates did not.

Removal of uremic toxins and regeneration of hemofiltrate by a selective dual hemofiltration artificial kidney (SEDUFARK) system.

A new system for the treatment of patients with end-stage renal failure has been developed. The SElective DUal Filtration ARtificial Kidney (SEDUFARK) is based on a dual filtration technique combined with low flow dialysis. Primary hemofiltrate containing large molecular weight substances is regenerated through a second filter followed by dialysis before the fluid containing essential substances of low molecular weight is returned to the patient. The ability of SEDUFARK to purify blood from uremic patients has been evaluated using a human macrophage assay system. Uremic plasma post SEDUFARK treatment was significantly less toxic than matched pre treatment plasma and the SEDUFARK regeneration system was found superior to conventional hemofiltration treatment.

Kinetics of intravenous saline infusion and selective IgG removal in rabbits.

Selective removal of approximately 60% of the plasma immunoglobulin G (IgG) mass in conscious rabbits was studied and compared with similar procedures combined with intravenous saline infusions equal to 7.5% body weight. Plasma concentrations of 125I-IgG and endogenous IgG were employed in model analysis to examine if saline infusions could be used to shift IgG from extra- to intravascular pool, thereby making more protein available for removal by extracorporeal plasma treatment. After IgG removal, the fractional metabolic clearance and the extra- to intravascular transfer coefficient were 40-50% lower than before IgG removal, and model simulations indicated that this may be caused by IgG binding. Saline infusion resulted in 40% more IgG mass in plasma 24 h after treatment compared with procedures without saline. Model analysis indicated that the increased IgG mass in plasma after saline procedures could be explained by a three- to fourfold increase in lymphatic clearance. Crystalloid infusions may be a method to increase the efficacy of repeated plasma exchange treatment.

Plasma concentrations and transperitoneal transport of native insulin and C-peptide in patients on continuous ambulatory peritoneal dialysis.

The insulin and C-peptide response to glucose (50 g), given intraperitoneally or enterally, and the elimination rate of these compounds has been studied in five nondiabetic patients on continuous ambulatory peritoneal dialysis (CAPD). The fasting C-peptide concentrations were three to ten times the normal values, whereas the fasting plasma insulin concentrations were within normal limits. After intraperitoneal glucose administration, a more marked hyperglycemia (P less than 0.05) and a more long lasting hyperinsulinemia (P less than 0.05) were found than after the enteral glucose load. The relative change in plasma C-peptide was slower and less pronounced in both experiments. Estimated total body clearance (Kt) for insulin was higher than for C-peptide (P less than 0.01), but dialysis clearance (Kd) for C-peptide was higher than for insulin in both experiments (P less than 0.01). The markedly elevated fasting C-peptide concentrations in plasma can be explained only partly by the absence of normal kidney function and suggests a continuously increased production of C-peptide during CAPD treatment. This was not reflected by the fasting plasma insulin concentrations. C-peptide measurements in plasma and dialysate during CAPD could be helpful in evaluating the beta-cell function in patients in need of exogenous insulin.

Toxicity of middle molecules: clinical evaluation using a selective filtration artificial kidney.

The effect of four uremic patients, treated 4-6 months with a new artificial kidney system aimed to remove molecules between 10,000-40,000 daltons returning substances with mol wt 200-10,000 back to the patients, was compared with the effect of conventional hemodialysis. The patients treated with this system obtained a more stable hemoglobin concentration without blood transfusions. They got an increased mean nerve conduction velocity and their plasma increased in quality as culture medium on human mononuclear phagocytes grown in vitro.

Beta 2-microglobulin kinetics during haemofiltration.

To study the kinetics of beta 2-microglobulin during haemofiltration, seven patients with end-stage renal failure were treated with the AN 69 (acrylonitrile), Duo-Flux (cellulose acetate) and F 60 (polysulphone) haemofilter. Low beta 2-microglobulin sieving coefficients and a highly negative filter mass balance error were observed during the initial phase of treatment with AN 69 but not with Duo-Flux or F 60, indicating a high degree of beta 2-microglobulin adsorption by AN 69. Total removal of beta 2-microglobulin was calculated by addition of the total amount adsorbed by the membrane and the total amount recovered in the collected ultrafiltrate. With AN 69 and F 60, total removal of beta 2-microglobulin amounted to 393 +/- 135 (SD) and 316 +/- 35 mg per treatment, while total removal with Duo-Flux was 242 +/- 79 mg per treatment. Thus, highly permeable membranes such as AN 69 or F 60 used in a haemofiltration mode may nearly balance the presumed generation of beta 2-microglobulin in uraemic patients. During treatment, an increase of the calculated beta 2-microglobulin distribution volume occurred with all three membranes, probably representing extra-to-intracellular water shifts. The water shifts occurring during haemofiltration reduce the value of precision of beta 2-microglobulin kinetics and limit the value of the plasma level decrease as an index of beta 2-microglobulin removal.

Generation and removal of anaphylatoxins during hemofiltration with five different membranes.

The generation of anaphylatoxins (C3a and C5a) during hemofiltration with polysulfone (F60), acrylonitrile (AN69 HF), polyamid (FH77), polyacrylonitrile (PAN200) and cellulose-acetate (Duo-Flux) membranes were investigated. The 'net' production of C3a in the blood compartment during the treatment was 3,028 micrograms (Duo-Flux), 1,845 micrograms (FH77), 270 micrograms (F60), -2,388 micrograms (AN69 HF) and -5,189 micrograms (PAN200), the latter two negative values reflecting a high adsorption of C3a to the membrane material. Corresponding values for C5a were 185 micrograms (AN69 HF), 108 micrograms (F60), 83 micrograms (PAN200), 6 micrograms (FH77) and -133 micrograms (Duo-Flux). The sieving coefficient remained stable for C5a throughout the treatment, while it fell significantly for C3a with the F60 membrane (0.44-0.25) and increased significantly for the FH77 membrane (0.14-0.33). The cellulose-acetate membrane device produced the most pronounced drop in white blood cell counts and the acrylonitrile membrane the least. The Duo-Flux device was significantly less biocompatible than the other membranes investigated, while little differences were found between AN69 HF, F60, FH77 and PAN200. When hemodialysis is compared with hemofiltration employing the same membranes, the latter appears to be a more biocompatible treatment modality partly due to convective removal of activated complement from the blood compartment into the filtrate.

Definitions of differences and changes in peritoneal membrane water transport properties.

A survey is given comparing measurements of transperitoneal water transport in different clinical situations with analyses based on the so-called "pore theory." This model links the measured changes to physical alterations of the peritoneal membrane. The calculations include "equivalent pore radius," effective "membrane area" and diffusive length, the transport resistance of the unstirred dialysate layer, and the residual intraperitoneal volume after dialysate drainage. The clinical appearances include individual differences in transperitoneal transport characteristics, changes in transperitoneal transport over time on continuous ambulatory peritoneal dialysis (CAPD) and during peritonitis, the pharmacological effect on the transport properties, and the effect of peritoneal catheter dislocation on ultrafiltration capacity. The main conclusions are as follow: During CAPD treatment the measurement of intraperitoneal solute equilibration and "mass-transfer-area coefficients" for urea and creatinine is less sensitive than the measurement of ultrafiltration volume in revealing peritoneal membrane changes. Differences and changes found have mostly a combined physical explanation, but one is more or less dominant. Changes in peritoneal membrane area seem to be the most dominant cause of changes in transperitoneal transport during time on CAPD and when sodium nitroprusside was added to the peritoneal dialysate. Changes during peritonitis can be explained by changes in pore radius and depth. Individual differences can be explained by differences in "membrane" area and in resistance of the unstirred dialysate fluid. High residual dialysate volume can give rise to clinical problems and should be considered when placing the catheter in the peritoneal cavity.

Removal, generation and adsorption of beta-2-microglobulin during hemofiltration with five different membranes.

Removal of beta 2-microglobulin (B2MG) from uremic plasma during hemofiltration in 5 patients using polysulfone (F60), acrylonitrile (AN69), polyamid (FH77), polyacrylonitrile (PAN200) and cellulose-acetate (Duo-Flux) membranes was investigated. The reduction in plasma concentration of B2MG (corrected for hemoconcentration during treatment) was greatest for the F60 hemofilter (61%, sieving coefficient 0.55 at end of treatment) and the AN69 (48%, sieving coefficient 0.45 at end of treatment, while the PAN200 membrane showed no reduction in plasma concentration of B2MG during treatment. Mass balance for B2MG showed that there was a continued net adsorption of B2MG in the AN69 membrane (about 120 mg) throughout the treatment, while insignificant amount of B2MG was absorbed by the F60 membrane. The results with the PAN200 showed a 'negative adsorption', indicating release of B2MG from cells into the plasma as blood passed through this filter. Comparison of reduction in plasma concentrations, sieving coefficients, plasma clearances and filter mass balances showed that the question of B2MG removal during hemofiltration is a complex issue where adsorption to the membrane, 'shedding' from blood cells, membrane pore size and possibly also 'redistribution' of B2MG in the body must be considered.

Analysis of immunoglobulin G kinetics in the non-steady state.

The effect of specific intravascular IgG depletion on IgG catabolism, generation, and intrabody transfer has been studied in rabbits. In contrast to previous studies, the radiolabeled IgG kinetics were analyzed in the non-steady state. A two-pool model was used to determine IgG distribution, catabolism, generation, and intrabody mass transfer after intravenous injection of 125I-IgG. Circulating IgG was then specifically removed by plasma perfusion through a Protein-A Sepharose column in an extracorporeal circuit. Based on the two-pool analysis, IgG catabolic clearance fell after IgG removal (1.0 ml/hr vs. 0.7 ml/hr), and mean generation rate was unchanged. Plasma levels rose 20 hours after IgG removal as a result equally of contributions from intrabody transfer and of generation. Model parameters from plasma 125I decay analysis overestimated plasma 125I levels in the first 24 hours after removal, although predicted endogenous levels corresponded well with experimental results over a 7-day period. Rapid intravenous infusion of a 7% body weight volume of saline solution during IgG removal resulted in 50% greater plasma levels of 125I-IgG 24 hours after removal. This indicated that an increased lymphatic flow had occurred, resulting in increased IgG transfer from the extravascular to the intravascular space. The two-pool model adequately describes circulating IgG levels after specific IgG removal. Catabolic clearance was found to be a function of IgG level, whereas generation does not appear to be similarly dependent. Both the two-pool model and saline infusion procedure may be applied directly to the planning and optimization of plasma exchange therapy regimens in human autoimmune disease.

Transport of small, middle, and large molecular weight substances in a dual filtration artificial kidney.

Theoretical calculations based on in vitro transport characteristics of membranes presently used in long-term clinical evaluation of a SElective DUal Filtration ARtificial Kidney (SEDUFARK) show that the system is comparable with hemofiltration in removal of urea and creatinine, and for substances greater than 15,000 daltons. Removal of middle molecular weight substances with SEDUFARK treatment is smaller or in the same region as conventional hemodialysis using cuprophane dialyzers. Analysis of G-15 chromatograms from the different in vivo SEDUFARK compartments confirmed these results, hence indicating that uremic serum contain substances larger than the middle molecular weight range which act toxic on a biological test system.

Clearance and synthesis rates of beta 2-microglobulin in patients undergoing hemodialysis and in normal subjects.

Retention of beta 2-microglobulin in patients undergoing hemodialysis is associated with a beta 2-microglobulin-derived amyloidosis. Removal of beta 2-microglobulin by renal replacement therapy has been proposed for the prevention of this amyloidosis. Currently, however, data on the beta 2-microglobulin synthesis rate in patients undergoing hemodialysis are scarce, and consequently it remains speculative how much removal would be necessary to counterbalance synthesis. The plasma kinetics of iodine 131-labeled beta 2-microglobulin were therefore examined in 11 patients with anuria who were undergoing long-term hemodialysis. Five healthy persons served as controls. Kinetic modeling of the plasma curves showed that the data fitted a two-pool model (r2 greater than 0.96) consisting of a rapid 2 to 4 hour distribution phase followed by a less steep curve, described by the plasma (metabolic) clearance (Clp). Synthetic rates were calculated from Clp and the beta 2-microglobulin steady state plasma concentration (plus beta 2-microglobulin removal during hemodialysis in the case of high flux hemodialysis). The results showed a significantly higher Clp in normal controls as compared with patients undergoing hemodialysis (65.5 +/- 12.8 ml/min (mean +/- SD) versus 3.4 +/- 0.7 ml/min). In contrast, the beta 2-microglobulin synthesis rate in the patient group (3.10 +/- 0.79 mg/kg/day) was not significantly different from that of normal controls (2.40 +/- 0.67 mg/kg/day), which was due to markedly elevated beta 2-microglobulin plasma concentrations in the patients (37.6 +/- 14.1 mg/L vs 1.92 +/- 0.27 mg/L). These findings suggest that the presence of end-stage renal disease does not have a significant impact on the beta 2-microglobulin generation rate. The degree of accumulation of beta 2-microglobulin in patients undergoing hemodialysis seems to depend on the loss of renal excretory function.

Biocompatibility aspects of cellophane, cellulose acetate, polyacrylonitrile, polysulfone and polycarbonate hemodialyzers.

The biocompatibility of cuprammonium rayon (Cu), cellulose acetate, polysulfone (Ps) and polyacrylonitrile hollow-fiber dialyzers and a polycarbonate-polyether flat plate dialyzer has been investigated. The Cu dialyzer resulted in more complement activation and a greater degree of leukopenia than the others, while the Ps hollow-fiber dialyzer appeared to be the most biocompatible of the membrane equipment in this study. These results were confirmed by in vitro evaluations and microscopic examinations of the different dialyzers.

Removal of uraemic toxins by haemofiltration with different membranes. The benefit of regenerating haemofiltrate using a newly developed system.

Based on earlier results which indicate that certain uraemic toxins (verified by an in vitro assay system) are larger than 10,000 daltons - a new system for treatment of endstage renal failure, SElective DUal Filtration ARtificial Kidney (SEDUFARK) has been developed. This system consists of a multimembrane filter/dialyser unit making removal of small molecules (mol wt less than 200) and substances in the mol wt range 10,000 - 40,000 possible without exchange of body fluids. Evaluation of pre and post treatment uraemic plasma with the bio assay showed that SEDUFARK was superior to conventional haemofiltration (CHF) with polyacrylonitrile (PAN) or Gambro Lundia Major High Flux 1.36 m2 (CPN) membranes as filters.

Glucose, insulin and C-peptide kinetics during continuous ambulatory peritoneal dialysis.

The insulin and C-peptide kinetics due to glucose (50 g), given intraperitoneally or enterally has been compared in five non-diabetic patients on continuous ambulatory peritoneal dialysis (CAPD). The fasting C-peptide concentrations were three to ten times the normal values whereas the fasting plasma insulin concentrations were within normal limits. After intraperitoneal glucose administration a more marked hyperglycaemia (p less than 0.05) and a more long lasting hyperinsulinaemia (p less than 0.05) was found than after the enteral glucose load. The relative change in plasma C-peptide was slower and less pronounced in both experiments. C-peptide concentration in plasma did not differ significantly between the two experiments. Estimated total body clearance (Kt) for insulin was higher than for C-peptide (p less than 0.01), but dialysis clearance (Kd) for C-peptide was higher than for insulin in both experiments (p less than 0.01).

Long-term changes in transperitoneal water transport during continuous ambulatory peritoneal dialysis.

9 patients were observed prospectively during 14-40 months 003 continuous ambulatory peritoneal dialysis (CAPD) treatment. From start of CAPD, each patient recorded dwell time, drained ultrafiltration volume (delta V), initial glucose concentration in dialysate, dialy fluid intake, body weight and blood pressure on a special form. These data, together with monthly values for albumin, urea, creatinin, phosphate, glucose and beta 2-microglobulin in plasma and in instilled dialysate, were later fed into a specially designed computer program to compare changes in the monthly mean (+/- SEM) values. During 5 episodes of peritonitis, daily changes in delta V were also computed. A long-term increase in delta V was found in 4 and a decrease in 5 patients. In all 9 patients delta V changed intermittently. All changes were most pronounced for long dwell times as compared to shorter dwell exchanges. The decrease in delta V started within the first 12 months of treatment. In the daily routine were aware of decreased ultrafiltration capacity in 3 patients only. Intermittent monthly changes in delta V could partly be correlated to changes in daily fluid intake. No correlations were found between long-term changes in delta V and fluid intake. All except 1 patient gained progressively in body weight, but without correlations to fluid balance, blood pressure and plasma albumin concentration. At the start of the observation period, most patients loosing delta V during this study appeared to have a more permeable membrane with a higher absorption rate of glucose and higher equilibration ratios for creatinine and beta 2-microglobulin in 5-hours drained dialysate as compared with the other patients. However, this was not statistically different between the two groups of patients. During the observation period, most patients with decreased delta V also increased transperitoneal solute transport, while the solute transport decreased in patients with increasing delta V, but these changes were only significant for some patients. During peritonitis, delta V decreased significantly 1 day before any other signs of peritonitis. All changes in delta V were most pronounced for long dwell times as compared with short dwell times. It is suggested that changes in ultrafiltration can be related to altered permeability of the peritoneal membrane, which appear earlier and more frequent than suggested by others, and any loss of delta V can be explained by a more permeable ('open') peritoneal membrane. It is also possible that different diseases act differently on the permeability of the peritoneal membrane.

Selection of filters and evaluation in vitro of the selective dual filtration artificial kidney, SEDUFARK.

Transport properties of different artificial kidney membranes have been studied employing substances with molecular weights ranging from 60 daltons (urea) to 62,000 daltons (dextran). The results from these studies were used to select filters for the SElective DUal Filtration ARtificial Kidney, SEDUFARK. Selection criteria for filters to be used in long-term clinical trials with SEDUFARK were based on biological tests showing toxicity of large molecular weight substances in uremic blood plasma and clinical advantages of returning middle molecular weight substances to the patient. The filters selected were used for evaluation in vitro of the complete SEDUFARK system. The results confirmed a theoretical analysis of this system, demonstrating that selective removal of substances within given molecular weight ranges from uremic blood plasma is feasible with a combination of commercial filters. The efficiency of the system, however, is limited due to the "cut-off" characteristics and hydraulic permeabilities of membrane types currently available.

High-flux synthetic versus cellulosic membranes for beta 2-microglobulin removal during hemodialysis, hemodiafiltration and hemofiltration.

Efficient removal of beta 2 microglobulin (beta 2-M) in end-stage renal failure patients is a continuing preoccupation, as the incidence and severity of dialysis-associated amyloidosis are increasing. To evaluate comparative beta 2-M removal we studied six stable end-stage renal failure patients during high-flux 3-h haemodialysis, haemodia-filtration, and haemofiltration, using acrylonitrile, cellulose triacetate, polyamide and polysulphone capillary devices. The reduction of plasma beta 2-M, total removal in ultrafiltrate/dialysate, and beta 2-M sieving coefficients were measured by RIA. The results suggest that convection plays the major role in beta 2-M removal when high-flux synthetic membranes are used in combination with high blood flow rates. In contrast, using the cellulose triacetate membrane under investigation, beta 2-M removal is diminished when ultrafiltration rates are increased. Accordingly, in any future prospective study on the role of beta 2-M retention in the amyloidogenesis, it is recommended that high-flux synthetic membranes be employed rather than the type of high-flux cellulosic membranes used in this study. The modality with which these synthetic membranes are used is probably less important, as long as maximum convective transport rates are obtained. Under present conditions, this will imply haemofiltration or haemodiafiltration rather than haemodialysis.