Some of the morbidity associated with chronic hemodialysis is thought to result from retention of large molecular weight solutes that are poorly removed by diffusion in conventional hemodialysis. Hemodiafiltration combines convective and diffusive solute removal in a single therapy. The hypothesis that hemodiafiltration provides better solute removal than high-flux hemodialysis was tested in a prospective, randomized clinical trial. Patients were randomized to either on-line postdilution hemodiafiltration or high-flux hemodialysis. The groups did not differ in body size, treatment time, blood flow rate, or net fluid removal. The filtration volume in hemodiafiltration was 21 +/-1 L. Therapy prescriptions were unchanged for a 12-mo study period. Removal of both small (urea and creatinine) and large (ss(2)-microglobulin and complement factor D) solutes was significantly greater for hemodiafiltration than for high-flux hemodialysis. The increased urea and creatinine removal did not result in lower pretreatment serum concentrations in the hemodiafiltration group. Pretreatment plasma beta(2)-microglobulin concentrations decreased with time (P< 0.001); however, the decrease was similar for both therapies (P = 0.317). Pretreatment plasma complement factor D concentrations also decreased with time (P<0.001), and the decrease was significantly greater with hemodiafiltration than with high-flux hemodialysis (P = 0.010). The conclusion is that on-line hemodiafiltration provides superior solute removal to high-flux hemodialysis over a wide molecular weight range. The improved removal may not result in lower pretreatment plasma concentrations, however, possibly because of limitations in mass transfer rates within the body.

Hemodialysis is associated with the formation of platelet-leukocyte aggregates. Whether this phenomenon is hemodialysis (HD) membrane dependent is unclear. To evaluate this process, we examined respectively platelet activation (anti-CD41, anti-CD62, and antifibrinogen monoclonal antibodies [MoAb] binding), leukocyte activation (CD11b expression), and the appearance of platelet specific antigens on leukocytes as an index of platelet-leukocyte aggregation during HD using 3 different membrane materials, Cuprophan, Hemophan, and polysulfone. Flow cytometric techniques and specific MoAb were used. All parameters were assayed 5 min after initiation of HD to avoid the confounding variable of leukopenia and resultant cell subpopulation analysis. Platelet activation (anti-CD62 and antifibrinogen binding) occurred only with Cuprophan. All 3 membranes induced equivalent increases in CD11b expression on neutrophils and similarly increased the binding of anti-CD41. to neutrophils, reflecting an increment in the formation of platelet neutrophil aggregates. However, only Cuprophan induced an increase in anti-CD62 binding to neutrophils, suggesting that the aggregated platelets linked to neutrophils were activated. Increased anti-CD41 binding by monocytes was similarly observed with all 3 membranes. However, only polysulfone induced an increase in CD11b expression and fibrinogen binding to monocytes. We conclude that while the formation of platelet leukocyte aggregates appears to be a universal phenomenon in HD occurring with a variety of membrane types, subtypes of this phenomenon consisting of activated platelets and fibrinogen binding may be membrane dependent. This phenomenon may serve as a new biocompatibility parameter and may shed light on some of the biologic consequences of hemodialysis.

Dialysis neutropenia is the result of pulmonary sequestration of neutrophils after complement activation by the dialyzer membrane. Increased expression of neutrophil adhesion receptors, such as CD11b/CD18, suggests that neutrophil adhesion to the capillary endothelium is a possible mechanism. An alternative hypothesis is that the complement fragment C5a modulates neutrophil mechanical properties via the cytoskeleton-largely filamentous actin (F-actin)-stiffening them and thereby slowing their passage through the pulmonary capillaries. To investigate this hypothesis, we developed an assay to measure the F-actin content of neutrophils in whole blood using flow cytometry and the stain NBD-phallacidin. We measured neutrophil F-actin content during hemodialysis of patients with polysulfone (N = 6), Hemophan (N = 6), and Cuprophan membranes sterilized with either ethylene oxide (N = 5) or steam (N = 6). Cell counts, neutrophil and monocyte CD11b expression and plasma C5a concentrations were also measured. The results confirm the strong relationship between the degree of neutropenia, increases in CD11b expression and plasma C5a levels reported by previous researchers. Modulation of the F-actin content of neutrophils was also strongly related to C5a levels, indicating that the neutrophil cytoskeleton is active during dialysis. Modeling of cell counts suggests that with Cuprophan a substantial fraction of neutrophils and monocytes are sequestered before they even pass through the dialyzer, suggesting some form of systemic activation of these cells. Evidence for systemic activation was also seen in measurements of F-actin content, but not CD11b expression, a finding that strengthens the case for the involvement of the cytoskeleton in dialysis neutropenia.

Hemodialysis is associated with simultaneous changes in leukocytes and platelets, but it is unclear whether these alterations affect the interactions between these cell types. To evaluate this process, we examined the appearance of platelet specific antigens (CD41) on leukocytes as an index of platelet-leukocyte aggregation during hemodialysis using three different synthetic membranes. Patients with end-stage renal disease (ESRD) on long-term hemodialysis treatment were enrolled. Flow cytometric techniques and platelet specific monoclonal antibodies (MoAb) that recognize the glycoprotein complex on resting and activated platelets (anti-CD41), the activated GPIIb-IIIa complex receptor (anti-LIBS1), and the p selectin GMP140, that is exposed on platelet plasma membrane after activation and platelet degranulation (anti-CD62), were used. Subjects with ESRD had a lower predialysis platelet surface expression of CD41 and LIBS1 compared to normal controls, but unchanged CD62 expression. In parallel, patients with ESRD manifested a uniformly reduced platelet-leukocyte microaggregates predialysis compared to normal controls. When examined across the dialyzer, however, an increase in platelet-neutrophil and platelet-monocyte microaggregates was observed with all three synthetic membranes at both 15 and 30 minutes after initiation of dialysis. This phenomenon could be duplicated in vitro by physiologic concentrations of the platelet specific agonist ADP, but not by the complement factors C3a or C5a. We conclude that platelet-leukocyte aggregates occur during dialysis likely related to a primary platelet activation mechanism. This phenomenon may serve as a new biocompatibility parameter and may shed light on some of the biologic consequences of hemodialysis.

A method for determining local transmembrane fluid movement in a commercial hemodialyzer at low dialysate flow rates by measuring changes along the dialyzer length in the local concentration of a marker macromolecule added to the dialysis solution has been developed. The method was evaluated in vitro at zero net ultrafiltration using dialyzers containing polysulfone (n = 4) and cuprophane (n = 3) membranes. The local concentration of the marker macromolecule along the dialyzer length was higher than the input dialysate concentration only during experiments with dialyzers containing polysulfone membranes. These observations provide direct empirical evidence that fluid movement in the dialysate to blood direction, i.e., backfiltration, occurs during hemodialysis with this highly permeable membrane. Net rates of backfiltration for the dialyzer containing the polysulfone membrane were also calculated from changes in the local concentration of the marker macromolecule and mass balance considerations. The calculated backfiltration rates increased with increasing blood flow rate and trended upward with increasing dialysate flow rate. The described methodology provides a novel approach for the further characterization of fluid and solute transport during hemodialysis with highly permeable membranes.

The expression of MHC class II (HLA-DR) and complement receptor (CR1) surface molecules on CD 14+ monocytes were compared with the production of the monokine interleukin-1 beta (IL-1 beta) in patients with endstage renal disease undergoing maintenance haemodialysis (HD) with hollow fibre dialyzers containing cellulose (CE, n = 8) and polysulfone (PS, n = 7) membranes. Monocyte staining was performed in blood samples obtained at the beginning and after 3 h of HD. Analysis of surface marker expression was done by immunofluorescence staining and flow cytometry analysis. Specific fluorescence intensity for both CR1 and class II (HLA-DR) antigens was increased in patients treated with CE membranes at the beginning of a dialysis treatment when compared to healthy control values. Interestingly, after 3 h on dialysis a further increase was noted for CR1 complement receptor expression whereas the increased HLA-DR expression was no longer detectable. In contrast, specific fluorescence intensity for both antigens was not significantly different from controls, either before or after dialysis, in patients treated with PS. The capacity of peripheral blood mononuclear cells to produce IL-1 beta spontaneously in vitro in the two patient groups was found to parallel results on phenotypic expression of surface molecules. The present study demonstrates that functional signs of monocyte activation, as evidenced by an augmented IL-1 beta production, in some patients on long-term HD correlate with an increased expression of two functionally important monocyte surface marker molecules.

Backfiltration of dialysis solution can occur during hemodialysis with highly permeable membranes. A method has recently been developed for determining backfiltration rates in vitro at low dialysate flow rates by measuring changes in the local dialysate concentration of a marker macromolecule via sampling ports added to the hemodialyzer housing. In the present study, the influence of net ultrafiltration on backfiltration rates was determined for five commercial dialyzers containing membranes with different water permeabilities. In vitro experiments were performed (n = 3) using freshly donated whole blood at blood flow rates of 200 and 340 ml/min and at a dialysate flow rate of 100 ml/min. At zero net ultrafiltration, backfiltration rates increased with increasing membrane water permeability and ranged from 0.9 to 6.9 ml/min. At a net ultrafiltration rate of 10 ml/min, backfiltration was eliminated for dialyzers containing membranes with water permeabilities of < 30 ml/h/mm Hg but remained significant for dialyzers with higher membrane water permeabilities. Therefore, despite a significant net ultrafiltration rate, backfiltration may still occur during hemodialysis with highly permeable membranes.

Peripheral blood mononuclear cells (PBMC) from haemodialysis patients treated with different types of membranes were isolated, incubated in vitro, and extracellular and cell‐associated interleukin 1 (IL‐1) assayed by radioimmunological methods. Extracellular IL‐1 was low and not different from controls, regardless of the dialyser used. In contrast, cell‐associated IL‐1 was increased in patients treated with dialysers containing low‐Hux Cuprophan (CU, n = 5) and polyacrylonitrile sheet membrane (AN69, n = 5), Patients treated with diaiysers containing highflux polysulphone (PS, n= 7), and polymethylmethacrylate (PMMA, n= S), exhibited no increase in cell‐associated IL‐l under these conditions. To elucidate the mechanism of the activation, aqueous extracts of dialysers containing CU, PS, and AN69 were tested for their ability to induce IL‐1 generation in PBMC from healthy donors. Extracts from unrinsed CU‐containing dialysers caused significant IL‐1 synthesis and release, whereas incubation with extracts from dialysers containing PS and AN69 sheet membranes did not. Hence, although both CU and AN69 sheet type dialysers result in activation of blood monocytes the mechanism of action appears to be different. We speculate that functional signs of PBMC activation as evidenced by increased spontaneous IL‐1 production observed in some patients on long‐term haemodialysis may result from extractable dialyser membrane material while in other instances direct cell membrane interactions or endotoxin transfer from the dialysate may be relevant.

An ex vivo model of hemodialysis was used to evaluate the effect of dialysis membranes on phagocytic cell function. Blood was withdrawn continuously from healthy, non-uremic donors, heparinized, and pumped, single pass, through membrane modules under conditions which simulated normal dialysis conditions. The membrane modules contained membranes of cellulose, DEAE-substituted cellulose, or polysulfone. Blood was obtained from the module outlets for determination of complement activation, phagocyte elastase release, zymosan-induced phagocyte chemiluminescence, and monocyte interleukin-1 production. Significantly less complement activation occurred with the polysulfone and DEAE-substituted cellulose membranes than with cellulose membranes. Normal monocyte interleukin-1 production was not stimulated by any of the membranes used. In contrast, the cellulosic, but not the polysulfone, membranes primed the oxidative burst of the phagocytes and caused them to release elastase. DEAE-substituted cellulose had a lesser effect on elastase release than did cellulose and elastase release correlated significantly with the degree of complement activation. However, the correlation between complement activation and priming of phagocyte oxidative burst was weak, suggesting that membranes affect phagocyte oxidative metabolism through more than one mechanism. We conclude that some dialysis membranes stimulate the bacteriacidal functions of normal phagocytic cells, in part through complement-dependent mechanisms.

Monocytes from patients treated by long-term haemodialysis with cellulosic membranes can show functional signs of activation depending on the dialyser module. To elucidate the mechanism of this phenomenon, aqueous extracts of various hollow-fibre dialysers were tested for their ability to induce in-vitro interleukin-1 generation in peripheral blood mononuclear cells from healthy donors. Extracts were prepared by passing 0.5 litres of sterile endotoxin-free water from the dialysate to the blood compartment of the dialyser. Different dialysers containing cellulosic membranes, i.e. cuprophan, and hemophan, were studied. Extracts were dried by lyophilisation and resuspended in cell culture medium before incubation with peripheral blood mononuclear cells for 18 hours at 37 degrees C in a 5% CO2 atmosphere. Interleukin-1 was assayed by biological or radioimmunological methods. Extracts from steam-autoclaved, dry-stored, or gamma-sterilised, wet-stored cuprophan and hemophan modules resulted in interleukin-1 activity that did not differ from negative controls. By contrast, extracts from cuprophan caused significant interleukin-1 production when prepared from ethylene oxide-sterilised, dry-stored dialysers. This monokine-inducing activity could not be neutralised by the addition of polymyxin B and was heat unstable, indicating that the cell-activating stimulus was not endotoxin. Extensive rinsing of the module with water before extract preparation totally mitigated the in-vitro production of interleukin-1. Our results suggest that the extract-induced activation of peripheral blood mononuclear cells found with some dialysers containing membranes of cellulosic origin cannot be exclusively related to the membrane polymer, but depends on a number of other parameters such as sterilisation and storage mode of a given membrane.

The capacity of peripheral blood mononuclear cells (PBMC) to undergo spontaneous and induced interleukin-1 (IL-1) production was examined in 10 healthy controls and in 10 patients undergoing intermittent hemodialysis (HD) using cuprophane membranes. Extracellular and cytoplasmic IL-1 activity was assayed by biologic and radioimmunologic methods. Spontaneous extracellular IL-1 activity was low in the patient and the control groups. However, spontaneous cytoplasmic IL-1 activity was greatly increased in the patients. Both assays gave comparable results. In contrast, the IL-1 response to lipopolysaccharide was significantly decreased in PBMC from the patients compared to controls. From these results we conclude that PBMC from HD patients show functional signs of activation, but an impaired response to exogenous stimuli, such as endotoxin.