Lymphocyte subsets in dialyser eluates: a new parameter of bioincompatibility?

During haemodialysis (HD), several adverse reactions in peripheral blood can occur, which have been attributed to the bioincompatibility of the dialyser membrane. Utilizing a dialyser elution technique, we have demonstrated that polymorphonuclear cells (PMN) manifested non-membrane dependent signs of activation during HD with cellulose triacetate (CTA), cuprammonium (CU) and polysulphone (PS) membranes. In the present study, we employed this elution technique to investigate the influence of HD with these membranes on lymphocytes.

Eight patients were studied during HD with CTA, CU, and PS dialysers in a randomized crossover design. Dialyser elution was performed after 3 h of HD. Besides total leukocyte count and differentiation, lymphocyte subpopulations and activation status in peripheral blood and dialyser eluates were analysed by flow cytometry.

Only with CU was a significant leukocyte decrease observed in peripheral blood at 30 min (P<0.001). Neither the total number of lymphocytes nor the proportion of T(CD3+) and B(CD19+) cells had markedly changed after HD with either membrane. Meanwhile, all membranes induced a relative decline in natural killer cells -NK(CD3-/CD16+/56+)- at the end of dialysis, although this was only significant for CTA (P=0.04). As for the T-lymphocyte subsets, the proportion of CD4+ cells had markedly increased after three hours of HD with all three dialysers, CTA and PS being significant (P<0.05). Dialyser eluates contained 33.8-82.2 x 10(6) cells, CTA yielding the highest cell counts. The majority (81-91%) of the eluted cells consisted of PMN dialyser eluates versus peripheral blood: P<0.05), whereas only few lymphocytes were found (4-13%, absolute 2.6 x 10(6)). Lymphocyte subpopulations in dialyser eluates were comparable to peripheral blood at t 180 in case of CTA and CU. In contrast PS eluates contained significantly fewer T-cells (37%), but more B-cells (22%) and NK-cells (30%) in comparison with peripheral blood at 180 min (peripheral blood: 79, 6 and 16% respectively; P<0.05). The expression of activation markers on T-cells (HLA-DR, CD25) in dialyser eluates was comparable with peripheral blood. Conclusions. The absolute number of lymphocytes in dialyser eluates of CTA, CU, and PS dialysers was low (mean 2.6 x 10(6)) in comparison with peripheral blood (mean 1.4 x 10(9)/l). Whereas non-selective adhesion occurred in CU and CTA dialysers, a selective adhesion pattern of lymphocyte subpopulations was observed in case of PS, suggesting a difference in bioincompatibility. Apparent T-cell activation was not noted, either in peripheral blood or in dialyser eluates. Because PMN in the dialyser eluates of three different membranes showed similar activation patterns in a previous study, we hypothesize that eluted lymphocyte, rather than PMN, represent a preferable parameter of bioincompatibility.