DIFFUSIVE AND CONVECTIVE TRANSFER OF CYTOKINE-INDUCING BACTERIAL PRODUCTS ACROSS HEMODIALYSIS MEMBRANES

The widespread use of bicarbonate dialysate, and high-flux and high-efficiency dialyzers have raised concerns regarding the transmembrane passage of bacterial products from the dialysate into the blood compartment. To study the mechanisms as well as magnitude of the transmembrane transfer of bacterial products from the dialysate, we developed a computerized in vitro dialysis model which provides continuous pressure recording from the arterial, venous, dialysate inflow and outflow ports. By virtue of a computer controlled on-line infusion pump, this model permits control of ultrafiltration/backfiltration. Heparinized (10 U/ml) whole blood (150 ml) was circulated through the blood compartment for 120 minutes at 100 ml/min. Bicarbonate dialysate contaminated with Pseudomonas maltophilia filtrate was circulated through the dialysate compartment at 100 ml/min. A two-point pressure of +10 mm of Hg (ultrafiltration) was maintained for the first 60 minutes and -10 mm of Hg (backfiltration) for the next 60 minutes. Whole blood samples (10 ml) were drawn from the blood at 0, 60 and 120 minutes. Peripheral blood mononuclear cells (PBMC) harvested from these samples were incubated for 24 hours in tissue culture medium. In addition, 0.5 ml samples of dialysate were collected at 0, 60 and 120 minutes and incubated with PBMC from the same donor for 24 hours. After 24 hour incubation, total cell-associated IL-1Ra and IL-1 beta were measured by specific radioimmunoassay. Paired experiments were performed with eight high-flux synthetic membranes (polyamide) and eight low-flux cellulose membranes (hemophan). Cytokine production is expressed as pg/2.5 million PBMC. During the mandatory ultrafiltration phase of the experiment (first hour), the production of IL-1Ra by PBMC from the blood compartment rose from 515 +/- 118 to 785 +/- 209 with polyamide membranes, and from 1175 +/- 365 to 3865 +/- 1847 with hemophan membranes, suggesting diffusive transport of bacterial products across the membrane. In contrast, at the end of the backfiltration phase (second hour), there was no further rise in the production of IL-1Ra by PBMC from the blood compartment (702 +/- 123 with polyamide, and 2284 +/- 886 with hemophan). The production of IL-1Ra by PBMC harvested from the blood compartment of polyamide membranes at 0, 60 and 120 minutes was lower than that with hemophan membranes (P = 0.01). In contrast to IL-1Ra, the production of IL-1 beta by PBMC harvested from the blood compartment with both membranes was uniformly low. The results of this study demonstrate the diffusive transfer of bacterial products across dialysis membranes. Therefore, any condition that increases diffusive transport such as dialyzers with large surface areas and high blood and dialysate flow rates could potentially increase the reverse transfer of bacterial products from the dialysate. This risk is not greater for synthetic high-flux membranes such as polyamide. Further, IL-1 beta is a less sensitive indicator than IL-1Ra of the transmembrane passage of cytokine-inducing substances from the dialysate to blood compartment.