Cytokine dialysis: an ex vivo study.

To test the hypothesis that dialysis using a new large pore membrane would achieve effective cytokine removal, blood from six volunteers was incubated with endotoxin (1 mg) and then circulated through a closed circuit with a polyamide membrane (nominal cut-off: 100 kDa). Hemodialysis was conducted at 1 or 9 L/hr of dialysate flow at the start of circulation and after 2 and 4 hours. The peak dialysate/plasma concentration ratios were 0.92 for interleukin (IL)-1beta, 0.67 for IL-6, 0.94 for IL-8, 0.33 for tumor necrosis factor (TNF)-a, and 0.11 for albumin. The dialysate/plasma ratios for all cytokines and albumin were decreased with increased dialysate flow from 1 to 9 L/hr (p < 0.05). Clearances for IL-1beta, IL-6, and IL-8, however, were significantly improved with increased dialysate flow (p < 0.01). There was no increase in TNF-a clearance (not significant) and a decrease in albumin clearance (p < 0.01). The peak clearance at 9 L/hr was 33 ml/min for IL-1beta, 19 for IL-6, 51 for IL-8, 11 for TNF-alpha, and 1.2 for albumin. No adsorption of cytokines was observed. We conclude that cytokine dialysis is achievable through a membrane with a high cut-off point with negligible albumin loss. These findings support the technical feasibility of this new approach to blood purification in sepsis.