The elimination of procalcitonin and the course of plasma concentrations during continuous veno-venous haemodiafiltration were measured in patients with sepsis or multiple organ dysfunction syndrome, because these patients are a main target group for the measurement of procalcitonin and often require renal replacement therapy. Procalcitonin was measured in the prefilter plasma and the filtrate at 5 min, 15 min and 1, 2, 4, 6, 12, 24 h after set-up of continuous veno-venous haemodiafiltration. In a prospective study, 19 patients with plasma levels of procalcitonin > 3 ng mL-1 and acute oliguric renal failure treated with continuous veno-venous haemodiafiltration using a polysulphone membrane, were evaluated for the study of clearance. Twenty-one control patients (procalcitonin < 2 ng mL-1) were studied to determine whether filtration itself induced a procalcitonin response. No interventions were required. In patients with low procalcitonin concentrations (procalcitonin < 2 ng mL-1) continuous veno-venous haemodiafiltration did not cause a rise in procalcitonin. In patients with increased procalcitonin plasma concentrations (> 3 ng mL-1), the protein was removed through the polysulphone membrane, with a final clearance of 4 mL min-1 after the initial adsorption period (clearance 0.4-0.9 mL min-1 during the first hour of continuous veno-venous haemodiafiltration). Thus, on the average, approximately 10% of plasma concentrations were measurable in the filtrate ultimately. However, procalcitonin plasma levels were not significantly altered during continuous veno-venous haemodiafiltration (86% of the initial concentration after 24 h). Although procalcitonin is removed from the plasma during continuous veno-venous haemodiafiltration in measurable amounts plasma procalcitonin concentrations did not change significantly during haemodiafiltration. Procalcitonin thus can also be used as a diagnostic parameter in patients undergoing continuous veno-venous haemodiafiltration.