Randomized controlled pilot study of 2 weeks' treatment with high cutoff membrane for hemodialysis patients with...

Chronic inflammation in hemodialysis (HD) patients is associated with cardiovascular complications and mortality. Circulating immune active proteins in the molecular range 15-45 kD that cannot be efficiently cleared by high-flux (HF) dialysis may be causally involved. We intended to test the feasibility of using a high cutoff (HCO) dialyzer in chronic HD patients and its influence on inflammation and monocyte activation. The Gambro HCO1100 dialyzer was compared to a conventional HF membrane in a randomized double-blind crossover trial in 19 chronic HD patients selected for the presence of elevated serum C-reactive protein levels. Patients were treated for six consecutive dialysis sessions (2 weeks) with each membrane. Safety analysis recorded adverse events and albumin losses through the protein-leaking membranes. Efficacy analysis observed reductions in the number of proinflammatory (CD14+CD16+) monocyte subpopulations in circulating blood. Treatment with the HCO membrane was well tolerated, although the number of adverse events was slightly higher. Despite significant serum albumin loss (from 34.1 ± 2.7 to 29.6 ± 3.0 g/L; P < 0.01), there was no need to supplement albumin, and rising activity of cholinesterase during HCO treatment indicated compensation by enhanced hepatic synthesis. The HCO membrane cleared high amounts of proinflammatory cytokines, but did not reduce predialysis inflammatory monocytes and markers. Although the time of HD session was extended, the study was hampered by a lower Kt/V in the HCO compared to the HF period. Treatment of chronic HD patients with this HCO dialyzer for 2 weeks is tolerable in terms of albumin loss and able to clear proinflammatory cytokines; however, this was not sufficient to decrease monocyte activation. Therefore, a more selective, less albumin-leaking membrane is desirable to allow prolonged high-efficient dialysis with more effective cytokine clearance.