Procalcitonin (PCT) is used as a diagnostic marker in critically ill patients to monitor the course of the disease and to assess the efficacy of therapeutic measures especially during sepsis, severe sepsis or septic shock [1,2]. Numerous studies indicate that the follow-up of PCT plasma concentrations can be used to assess the efficacy of therapeutic measures as well as the outcome in patients with severe bacterial infections, sepsis or multiple organ dysfunction. After elimination of a focus of infection, rapidly declining concentrations were indicative of a good prognosis of the disease [3-6]. Particularly after successful surgical treatment of a focus of infection, e.g. during acute peritonitis, PCT plasma concentrations rapidly declined after elimination of the focus  indicating a good prognosis with high sensitivity and specifity . Also in patients with sepsis and multiple organ dysfunction syndrome, a decline of plasma levels of PCT was correlated with survival of the patients in various studies, whereas increasing or continuously elevated concentrations were found in patients with poor prognosis. Other inflammatory parameters, such as C-reactive protein (CRP), granulocyte-elastase or various cytokines, were not correlated with the prognosis as PCT .
Organ dysfunction is a frequent complication in critically ill patients, particularly acute renal failure often occurs early in septic patients and those at risk of multiple organ dysfunction. Today, continuous venovenous haemodiafiltration (CVVHDF) is the therapy of choice for acute renal failure in these patients. Because PCT is a protein with a low molecular weight (molecular mass approximately 13 kDa) [1,2,7], dialysis of the protein through the filter membrane may occur, as known from other proteins of a similar molecular weight. Thus, plasma levels and elimination kinetics of PCT might be altered during haemodiafiltration and therapeutic and diagnostic decisions would be compromized, when CVVHDF alters plasma concentrations. Because there are insufficient data to show whether PCT will be eliminated in significant amounts during CVVHDF, we have analysed both the elimination of PCT into the filtrate during CVVHDF as well as the course of PCT plasma levels in patients undergoing CVVHDF. To determine whether filtration itself induces a PCT response, plasma concentrations were followed-up during CVVHDF also in a control group of patients, who initially had low PCT plasma levels.
Patients and methods
Patients, in whom elevated PCT plasma concentrations were measured (> 3 ng mL−1) and in whom CVVHDF was required due to acute oliguric renal failure (diuresis < 400 mL day−1), were included in the study. The underlying disease was either sepsis, severe sepsis or septic shock according to ACCP/SCCM criteriae , due to different aetiologies (peritonitis, pneumonia, multiple trauma and subsequent multiple organ dysfunction syndrome). Nineteen critically ill patients (of whom 14 were male) were included prospectively into this study during the period from May 1, 1998 to December 31, 1999. The median age was 68.8 years (53.3-74.7 years, 25/75 percentiles). Median plasma PCT concentration before CVVHDF was 20 ng mL−1, ranging to 80 ng mL−1. Nineteen periods of CVVHDF were analysed, seven of them up to a maximum duration of 24 h. PCT was measured in the prefilter plasma and the filtrate at 5 min, 15 min and 1, 2, 4, 6, 12, 24 h after setup of CVVHDF.
To analyse the possible increase of PCT by CVVHDF, 21 patients with low PCT plasma concentrations prior to CVVHDF (< 2 ng mL−1) were prospectively followed-up, and PCT plasma concentrations were measured before and 12-18 h after the onset of CVVHDF. In these patients, PCT concentrations were measured daily by the routine clinical laboratory and not as a separate batch.
The study was conducted in accordance with the principles of the Helsinki Declaration, the aims of the study were examined by the local ethical committee and informed consent was obtained from the patients or their relatives.
Venous access was obtained by a 12 F Sheldon catheter inserted either into the femoral or internal jugular vein. CVVHDF was performed using a hollow-fibre Renaflo®II PSHF 1200 hemofilter (Baxter, Germany) and a pump-assisted circuit (BM11 and BM 14, Baxter, Germany). A lactate-buffered fluid (SH-04, SH-14, Baxter, Germany), supplemented with potassium chloride when required, was used for dialysis. The filtrate fluid was pumped countercurrent to blood through the dialysate compartment of the filter at a rate of 33.33 mL min−1. The blood flow was set to 80 mL min−1 and net filtration rate was set to −100 mL h−1. Heparin was infused through a prehaemofilter port with an initial dosage of 500 U h−1 and the infusion rate was adjusted by monitoring of the PTT and activated clotting time (ACT). The surface area of the Baxter PSHF 1200 hemofilter was 1.25 m2, the priming volume of the system 83 mL .
Blood samples were collected prefilter in a 5-mL polyethylene monovette and centrifuged immediately after collection (10 min, 800 g, 4°C). The prefilter blood samples represent the patients' systemic venous blood concentrations. The filtrate was collected directly from the outlet of the haemofilter (polyethylene syringe, Sarsted, Germany). Both plasma and filtrate, respectively, were frozen immediately (−20°C) until measurement.
Blood samples were taken immediately after onset of CVVHDF (t=0) and at the following times: 5 min, 15 min, 1h, 2h, 4h, 6h, 12h and 24h. A minimum duration of dialysis of at least 24h was scheduled; however, clotting of the filters limited the duration of dialysis in some patients. PCT was measured with the Lumitest® PCT (B.R.A.H.M.S.-Diagnostica GmbH, Hennigsdorf, Germany) as described previously . When PCT was measured in the filtrate, samples and the respective standard solutions (20 μL each) were diluted with 20 μL of zero-serum, containing approximately 5% of human albumin, because filtrate has a low protein content. The stability of PCT in the filtrate was confirmed by storage of PCT containing samples of the filtrate for 24 h and comparison with immediately frozen samples. The stability was > 90% even after 24 h (data not shown). As demonstrated previously, also freezing and thawing does not significantly affect plasma concentrations . The blood flow and filtrate flow were kept constant during the whole measurement at 80 mL min−1 and 33 mL min−1 respectively.
The following formulae were used:
Filtrate clearance (mL min−1), Cfil Qfil/Cin
Coefficient of filtration to plasma, Cfil/Cin
Mass transfer filtrate (ng min−1), Vfil × Cfil
Plasma volume (mL), 0.0375 body weight 
Volume of the extracellular space (mL), ECS, 0.225 body weight 
Mfil/Mplasma × 24; Mfil/MECS × 24, amount of PCT, eliminated into the filtrate compared with the estimated amount of plasma PCT and extracellular space PCT, respectively, given as a percentage and extrapolated to 24 h.
Cin, PCT concentration prefilter (ng mL−1); Cfil, PCT concentration of filtrate (ng mL−1); Qfil, filtrate flow (33.33 mL min−1); Vfil, volume of filtrate per min (mL min−1); Mfil, mass of PCT eliminated in the filtrate within 1 h; Mplasma, MECS, estimated mass of PCT in the plasma and ECS, respectively.
Data are expressed as median and 25 and 75 percentile of the original data or as percentage of the initial values (time=5 min) where appropriate. The Wilcoxon test for dependent samples was used to analyse the course of prefilter venous PCT plasma concentrations during haemofiltration. Statistical significance was assumed as P < 0.05.
During CVVHDF, using a polysulphone membrane, PCT in significant amounts passed the filter membrane and was measurable in the filtrate fluid. Two hours after set-up of the CVVHDF device, approximately 4% of the PCT concentrations of the plasma were recovered in the filtrate fluid. The filtrate clearance calculated to approximately 3-4 mL min−1 after 12-24 h of CVVHDF. However, during the initial first hour of dialysis, only moderate amounts of PCT were measurable in the filtrate fluid. The results of the measurements are given for each time point in Table 1. Concentrations, measured immediately after the onset of CVVHDF (t=0) are not shown, because the sample of the filtrate are diluted by the priming solution at this time, and plasma levels of the patients were not different from those measured only 5 min later. Because the bypass circuit clotted in some patients after several hours of filtration, nine patients were followed-up for 12 h and seven patients for an observation period of 24 h. Thus, the 24-h values do not represent the entire sample of patients and should be interpreted accordingly.
When the PCT plasma concentrations of the patients were analysed during and after CVVHDF, no significant decline of prefilter plasma concentrations was observed (Tables 1 and 2). Also, PCT was not induced by CVVHDF in patients with low PCT concentrations. In 21 patients of the control group, in whom PCT was < 2 ng mL−1 prior to CVVHDF, PCT did not increase significantly 12-18 h after the onset of filtration (Table 2).
To compare the amount of PCT, eliminated by CVVHDF with the estimated amount of PCT within the plasma, we have also calculated the plasma volume and the PCT content of the plasma in each patient according the above given formula. Thus, during a 24-h period of dialysis, approximately 10% of the estimated amount of plasma PCT can be removed from the plasma by CVVHDF. However, plasma levels of PCT were not significantly reduced during the course of CVVHDF. We thus have also calculated the estimated ECS, because the distribution volume of PCT may exceed that of the plasma volume. Assuming the distribution volume of PCT to equal that of the ECS, approximately 1% of the ECS-PCT is eliminated during a 24-h period of CVVHDF, which consistently explains the weak influence of CVVHDF on PCT plasma levels.
During CVVHDF, procalcitonin is partially removed from the plasma into the filtrate fluid. An initial filtrate clearance of 0.4 rising to approximately 4 mL min−1 after 24h of CVVHDF and a coefficient of dialysis rising from 0.01 to 0.1 indicate that finally up to 10% of plasma PCT approximately was removed from the plasma into the filtrate at the current setting of the CVVHDF device. The coefficient of dialysis, which describes the ratio of PCT between the plasma and the filtrate, is only approximately half of the sieving coefficient, measured for PCT during CVVHF . Thus, only approximately half the amount of PCT crosses the membrane, when compared with convective flow during CVVHF.
When compared with other proteins of low molecular weight, the sieving coefficient of PCT is within a comparable range as described for cytokines and other proteins. The Baxter PSHF 1200 hemofilter is a glycerin free polysulphone membrane, characterized by sieving coefficients ranging from 1.0 for urea and creatinine, 0.07 for myoglobin (MW 17 kDa), 0.42 for IL 1β (MW 17 kDa) to 0.891 for cytochrome C (MW 12 kDa). Larger substances, like albumin (MW 65 kDa) or for example TNFα RII (MW 75 kDa) are filtrated only in very low amounts (sieving coefficients 0.005 and 0.01 respectively) [9, 14, 15]. For cytokines, sieving coefficient in the range from 0.02 to 0.6, as measured by Heering et al., van Brommel et al., and Hoffmann et al. were reported, e.g. 0.02-0.22 for TNFα, 0.04 for IL-6 (MW 23-30 kDa) and 0.12-0.67 for IL-8 (MW 10 kDa).
During the initial period of CVVHDF, only low concentrations of PCT were found in the filtrate. Because a mass balance, and thus a possible adsorption to the membrane, is not measurable during dialysis due to effects of haemoconcentration and haemodilution according to the oncotic pressure of the plasma of each patient, the remaining PCT during this period cannot be completely followed-up. However, as analysed previously during CVVHF, using the same filter membrane, adsorption to the membrane is the most likely explanation for the low filtrate clearance in this initial period also for CVVHDF . During CVVHF, a negative mass balance error was calculated, indicating thoroughly the adsorbance of PCT to the PSHF 1200 membrane. We therefore conclude, that adsorption to the membrane occurs also during CVVHDF and accounts for the low PCT concentration of the filtrate during the first hour of dialysis.
PCT plasma levels of the patients were not altered during CVVHDF, which is consistent with previous findings reported for CVVHF. However, a steady rate of PCT production cannot be assumed in patients with sepsis during the entire observation period. However, stimulation of PCT production by the filtration device, was excluded when PCT was measured in patients with low PCT concentrations prior and 12-18 h after the onset of CVVHDF or CVVHF. Also, when the course of PCT was evaluated in the group of patients with high PCT concentrations, there was no consistent tendency towards increasing concentrations of PCT during, prior and the time after CVVHDF (data not shown in detail): approximately one-third of the patients presented with equal PCT concentrations the day prior or after CVVHDF, whereas in some patients PCT concentrations increased, but also decreased in some patients in the same observation period. Thus, a bias through a natural increase of PCT due to the underlying disease can be excluded in the majority of patients analysed in this study. The data render it more likely that the distribution volume of PCT in the body exceeds that of the plasma volume, which would explain the above results. According to the data published by Nylen and his colleagues , PCT may be synthesized in numerous cell types, organs and tissues. PCT is secreted from these cells rapidly upon the initial stimulus into the extracellular space. This may result in a continuous liberation of extracellular deposed PCT into the circulation. Thus, elimination of PCT from the plasma may cause a gradient of concentrations from the ECS to the plasma only, affecting PCT plasma concentrations to a minor degree, because the pool of PCT within the body largely exceeds the amount of PCT within the plasma. The maintenance of the plasma concentrations thus requires a production rate similar to the elimination rate. To confirm this hypothesis, further studies need to be performed. Obviously, the removal of much higher amounts of PCT would be required to obtain a measurable effect on PCT plasma concentrations in the patient group averages.
Furthermore, unchanged plasma concentrations during CVVHF have likewise been reported for other proteins with a similar sieving coefficient during sepsis. For example, plasma concentrations of IL-1β and IL-1 Ra (MW 17 kDa each), though being filtrated (sieving coefficient 0.42 and 0.41 respectively), were not significantly influenced by CVVHF, even when higher sieving coefficients indicate a more easy penetration of the membrane for these proteins [14-16]. However, these proteins have a half-life time far less than PCT.
In summary, measurable amounts of procalcitonin were eliminated also during haemodiafiltration, however, CVVHDF did not influence PCT plasma concentrations or kinetics in a significant way. Obviously a greater pool of PCT than that residing in the plasma volume is present in the body, although a varying production rate during the observation period of CVVHDF cannot be excluded. Given our data, we conclude that PCT can be used as a diagnostic parameter also in patients with renal failure undergoing artificial renal replacement therapy.
List of abbreviations: C concentration; CI clearance; CRP C-reactive protein; CVVHDF continuous veno-venous haemodiafiltration; ECS extracellular space; kDa kilodalton; Hct haematocrit; MBE mass balance error; MODS multiple organ dysfunction syndrome; MW molecular mass; PCT procalcitonin; Q flow; V volume
The authors thank Mrs. I. Witte for skilful technical assistance.
1 Meisner M. PCT-Procalcitonin. Ein Neuer, Innovativer Infektionsparameter Biochemische und klinische Aspekte.
Stuttgart: Georg Thieme-Verlag, 2000.
2 Meisner M. PCT, Procalcitonin - a New, Innovative Infection Parameter.
Berlin: B.R.A.H.M.S.-Diagnostica GmbH, 1996.
3 Gramm HJ, Dollinger P, Beier W. Procalcitonin - ein neuer Marker der inflammatorischen Wirtsantwort. Longitudinalstudien bei Patienten mit Sepsis und Peritonitis. Chir Gastroenterol
(Suppl. 2): 51-54.
4 Reith HB, Lehmkuhl P, Beier W, Högby B. Procalcitonin - ein prognostischer Infektionsparameter bei der Peritonitis. Chir Gastroenterol
(Suppl. 2): 47-50.
5 Reith HB, Mittelkötter U, Debus ES, Lang J, Thiede A. Procalcitonin (PCT) immunreactivity in critical ill patients on a surgical ICU. In: The Immune Consequences of Trauma, Shock and Sepsis.
Bologna: Monduzzi Editore, 1997 1: 673-677.
6 Meisner M, Tschaikowsky K, Palmaers T, Schmidt J. Comparison of procalcitonin (PCT) and C-reactive protein (CRP) plasma concentrations at different SOFA scores during the course of sepsis and MODS. Crit Care
7 Oczenski W, Fitzgerald RD, Schwarz S. Procalcitonin: a new parameter for the diagnosis of bacterial infection in the peri-operative period. Eur J Anaesthesiol
8 Anonymous. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference: Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Crit Care Med
9 Baxter. Baxter Renaflo II Hemofilter, Product Infomation.
Unterschleissheim, Germany: Baxter GmbH, D-85716, 1999.
10 Meisner M. Procalcitonin: Erfahrungen mit einer neuen Messgroesse fuer bakterielle Infektionen und systemische Inflammation. J Lab Med
11 Meisner M, Tschaikowsky K, Schnabel S, Schmidt J, Katalinic A, Schüttler J. Procalcitonin - Influence of temperature, storage, anticoagulation and arterial or venous asservation of blood samples on procalcitonin concentrations. Eur J Clin Chem Clin Biochem
12 Niemer M, Nemes C, Lundsgaard-Hansen P, Blauhut B. (eds) Datenbuch Intensivmedizin.
Stuttgart: Gustav Fischer-Verlag, 1992.
13 Meisner M, Hüttemann E, Lohs T, Kasakov L, Reinhart K. Plasma concentrations and clearance of procalcitonin during continuous veno-venous hemofiltration in septic patients. Shock
2000 (in press).
14 Heering P, Morgera S, Schmitz FJ et al.
Cytokine removal and cardiovascular hemodynamics in septic patients with continuous venovenous hemofiltration. Intensive Care Med
15 van Brommel EFH, Hesse CJ, Jutte NHPM, Zietse R, Bruining HA, Weimar W. Impact of continous hemofiltration on cytokines and cytokine inhibitors in oliguric patients suffering from systemic inflammatory response syndrome. Renal Failure
16 Hoffmann JN, Hartl WH, Deppisch R, Faist E, Jochum M, Inthorn D. Effect of hemofiltration on hemodynamics and systemic concentrations of anaphylatoxins and cytokines in human sepsis. Intensive Care Med
17 Nylen E, Muller B, Snider R et al.
Pathophysiological significance of calcitonin precursors in sepsis and systemic inflammation (Abstract). Shock