An increased cytokine production, correlated with long term complications of uremic disease, has been described during hemodialysis. To identify possible differences in the cytokine release of differently sterilized membranes, we enrolled six uremic patients on chronic hemodialysis. The patients underwent dialysis with ETO-sterilized low-flux polysulphone membranes (F6, Fresenius AG) for at least three months (A1), they were then switched to steam-sterilized polysulphone membranes (F6-HPS Fresenius AG) and further evaluations after one (B1) and two months (B2) were carried out. A final evaluation (A2) was made one month after switching back to F6 dialyzers. At each time period, samples were drawn to measure IL-1beta released by cultured mononuclear cells (MN). Moreover, dialysate samples were collected to test endotoxin levels. C3a and C5a levels were assessed at 0, 5, 15 and 60 min from starting hemodialysis. Anti-ETO IgE levels were also assayed at A1, B1 and A2. The LAL test revealed a good quality dialysate. The mean pre-dialysis IL-1beta levels were 215 pg/million cells at A1; falling to 49 at B1, and 54 at B2 (p<0.01); there was then a sharp rebound at A2: 284, p<0.01. Post-dialysis levels followed the same pattern. No correlation between the dialysate endotoxin level and cytokine release was found. Complement activation did not change and in all the phases of the study no anti-ETO IgE was detected in any of the subjects. Our data suggest that the steam sterilized polysulphone membrane induces a lower cytokine release than the ETO sterilized membrane, although the mechanism by which it does so remains to be clarified.