Share this article on:

On-Line Dialysate Infusion to Estimate Absolute Blood Volume in Dialysis Patients

Schneditz, Daniel*; Schilcher, Gernot; Ribitsch, Werner; Krisper, Peter; Haditsch, Bernd*; Kron, Joachim

doi: 10.1097/MAT.0000000000000086

It was the aim to measure the distribution volume and the elimination of ultra-pure dialysate in stable hemodialysis patients during on-line hemodiafiltration (HDF). Dialysate was automatically infused as a volume indicator using standard on-line HDF equipment. Indicator concentration was noninvasively measured in the arterial blood-line (using the blood volume monitor, Fresenius Medical Care, Bad Homburg vor der Höhe, Germany), and its time course was analyzed to obtain the elimination rate and the distribution volume V t at the time of dilution. Blood volume at treatment start (V 0) was calculated accounting for the degree of intradialytic hemoconcentration. Five patients (two females) were studied during 15 treatments. Two to six measurements using indicator volumes ranging from 60 to 210 ml were done in each treatment. V 0 was 4.59 ± 1.15 L and larger than the volume of 4.08 ± 0.48 L estimated from anthropometric relationships. The mean half-life of infused volume was 17.2 ± 29.7 min. Given predialysis volume expansion V 0 was consistent with blood volume determined from anthropometric measurements. Information on blood volume could substantially improve volume management in hemodialysis patients and fluid therapy in intensive care patients undergoing extracorporeal blood treatment. The system has the potential for complete automation using proper control inputs for BVM and HDF modules of the dialysis machine.

From the *Institute of Physiology, Medical University of Graz, Graz, Austria; Department of Internal Medicine, Clinical Division of Nephrology, Medical University of Graz, Graz, Austria; and KfH-Kuratorium für Dialyse und Nierentransplantation, KfH Kidney Center Berlin-Köpenick, Berlin, Germany.

Submitted for consideration December 2013; accepted for publication in revised form March 2014.

Part of this study was supported by Fresenius Medical Care, Bad Homburg vor der Höhe, Germany.

Disclosure: The authors have no conflicts of interest to report.

Correspondence: Daniel Schneditz, PhD, Institute of Physiology, Center for Physiologic Medicine, Medical University of Graz, Harrachgasse 21/5, 8010 Graz, Austria. Email:

Volume management plays an important role in chronic and acute renal replacement therapy. In hemodialysis, the removal of fluid by ultrafiltration within short periods of time is one of the most important factors triggering intradialytic hypotension, whereas excess blood volume plays a critical role in the pathogenesis of hypertension.1–3 In acute treatments, excessive fluid administered to maintain hemodynamic stability leaks into tissues aggravating diffusive transport of metabolites.4–7 Therefore, the magnitude of intravascular volume plays an important role in both settings.

Sensors have been introduced in chronic therapy to follow the ultrafiltration-induced changes in blood volume by continuous measurement of hemoconcentration with the aim to adjust treatment settings and to improve treatment prescriptions.8–11 The success of this approach is in debate.12,13 The volume effect of different solutions to treat hemodynamic instability during dialysis has also been studied.14,15 However, in spite of so-called relative blood volume measurements, the magnitude of intravascular volume remains unknown. Hence, there have been attempts to measure blood volume during extracorporeal blood treatment using noninvasive extracorporeal sensors and technology.16–19 All approaches require a defined perturbation by endo- or exogenous indicators. The controlled removal or delivery of fluid volume is most appealing as it is inherent to the process of renal replacement therapy, and, therefore, it is also easily quantified using current dialysis technology. This is especially the case with on-line hemodiafiltration (HDF) where ultra-pure dialysate can directly be infused into the patient. Because the indicator volume is exactly known, the resulting change in hemoconcentration measured by available on-line techniques may be used to identify the magnitude of the distribution volume and the escape rate of the intravascular volume indicator.

It was the aim of this study to measure the distribution and elimination of a volume indicator during on-line HDF and to quantify the blood volume and the indicator half-life of ultra-pure dialysate in a group of stable hemodialysis patients.

Back to Top | Article Outline

Material and Methods

The distribution volume of a volume indicator such as saline added to whole blood is given as

where V ind is the indicator volume (in L), V is the distribution volume before the addition of indicator volume (in L), and where C ind is the volume concentration of the indicator (in vol/vol) after indicator mixing and equilibration. The equations for V and C ind are derived in the Appendix.

Back to Top | Article Outline

Indicator Measurement

Indicator concentration was determined by serial measurements of blood water concentration at a sampling period of 5 s together with the exact time and volume of indicator injected using the blood volume monitor (BVM) and dedicated data acquisition software (Fresenius Medical Care [FMC], Bad Homburg vor der Höhe, Germany) (Figure 1).20 The BVM was also used to determine the relative blood volume changes (RBV, in vol/vol) throughout the whole treatment.21 The measuring site was located in the arterial limb of the extracorporeal circulation.

Figure 1

Figure 1

Analysis of dilution curves was done off-line using a Microsoft Excel macro for Macintosh PC. Data were sampled from the original record for a 10 min window around the time t of the dilution measurement. The period preceding the dilution was used to quantify any trend resulting from simultaneous ultrafiltration and vascular refilling and to estimate baseline water content (W) (Figure 1, right panel). Serial data were corrected for this trend, and the time course of indicator concentration C ind was then calculated as a function of water content in the mixture (W m) according to Equation 10 (Appendix; Figure 2, left panel). Indicator water content W ind was assumed as 99.1% (≈0.9% saline). The exponential decline determined from a time window between approximately 1 min after indicator maximum and 10 min after infusion start was used to back-extrapolate the concentrations to the time of indicator maximum (Figure 2). The distribution volume before the addition of indicator was then calculated from Equation 1.

Figure 2

Figure 2

Distribution volumes (V t) measured at different time points (t) during dialysis were corrected for the relative blood volume change measured at these time points (RBV t, in vol/vol) and back-calculated to the beginning of the treatment (V 0) where RBV 0 = 1 to account for effects of ongoing ultrafiltration and vascular refilling:

Anthropometric blood volume (V a) was estimated from body height (H, in m) and body mass at dry weight (M, in kg) using the formula developed by Nadler et al.22:

with specific constants for male (a = 0.3669, b = 0.03219, c = 0.6041) and female (a = 0.3561, b = 0.03308, c = 0.1833) subjects.

Specific volumes (V s, in ml/kg) were obtained by normalizing anthropometric blood volumes and measured distribution volumes to body mass at dry weight.

Back to Top | Article Outline


Patients gave written informed consent to participate in the study as approved by the Ethics Committee of the Medical University of Graz, Austria.

Ultra-pure bicarbonate dialysis was delivered by a 4008H dialysis machine equipped with an on-line HDF module and a BVM utilizing FMC blood-lines with the incorporated BVM measuring cell and F70S (FMC) dialyzers. Dialysate flow was either 500 or 800 ml/min and delivered at a temperature of 36°C. Blood flows and dialysate [Na+] were set as prescribed. All studies were done in the on-line HDF mode using either pre- or postdilution configurations. Infusion volumes in HDF patients were set as prescribed. In patients usually treated by hemodialysis, the volume of the substitution fluid was 5 L.

Indicator dilution was applied by means of the “bolus” function of the HDF module. This function delivers ultra-pure dialysate in multiples of 30 ml up to a total amount of 210 ml at a constant infusion rate of approximately 150 ml/min during the operation of on-line HDF either into the arterial (predilution) or into the venous blood-line (postdilution) (Figure 3). The volume is delivered with an accuracy of better than ±1.5%.

Figure 3

Figure 3

Dilutions were done between automatic transmembrane pressure (TMP) tests (recognized by blinking of TMP light-emitting diodes on the dialysis monitor) usually occurring every 12.5 min in the machine used in this study and producing a biphasic blood volume transient (Figure 1, left panel) that interferes with the volume measurement.

The ultrafiltration volume was set as prescribed to reach the clinical patient dry weight and augmented by the total amount of anticipated infusion volume. This volume was removed by a constant ultrafiltration rate maintained throughout the duration of the treatment.

Back to Top | Article Outline

Statistical Analysis

Data are presented as mean ± standard deviation unless otherwise specified. Linear regression was done by the method of least squares. Continuous data were compared by nonparametric tests (StatView 4.5, Abacus Concepts Inc., Berkeley, CA). A probability p < 0.05 was assumed as significant to reject the null hypothesis.

Back to Top | Article Outline


Five patients (two females) were studied during 15 treatments so that each patient was at least studied on two different occasions. Basic patient and treatment data are summarized in Table 1. Two to six dilution measurements using indicator volumes V ind ranging from 60 to 210 ml were done in each treatment so that 61 measurements were available for final analysis (Table 2). The dilutions were delivered without technical difficulties, were not noticed by patients, and were well tolerated.

Table 1

Table 1

Table 2

Table 2

An example of on-line data is shown in Figure 1, whereas Figure 2 shows indicator concentrations extracted from basic concentration curves. The curves are characteristic for slow dilution into the venous blood-line with typical delays among injection (t = 0), appearance (t app), and maximum (t max) of indicator concentration measured at the extracorporeal sampling site.

Two patients had a central venous access where mixed venous blood entered the extracorporeal circulation and passed the arterial measuring site. In this case, the initial phase of the dilution curve was systematically different from that obtained with a peripheral arteriovenous access where arterial concentrations were sampled (Figure 4).

The decline in dialysate volume after the injection was identified from exponential curve fits. The rate constant (k) was significantly different from zero (p < 0.0001, One-Sample Sign test) and negative in 58 studies (Table 2). The half-life t 1/2 of indicator volume identified from rate constants was 17.2 ± 29.7 min. Indicator concentrations estimated at the time of indicator maximum (t), corresponding distribution volumes (V t), and distribution volumes corrected for the relative change in blood volume between measuring time points are given in Table 2. Volumes estimated at t and at the beginning of dialysis were larger than anthropometric blood volumes (p < 0.0001, Friedman test). A comparison of distribution volumes and specific blood volumes obtained in individual patients and within single treatments is shown in Figure 5.

Back to Top | Article Outline


This study shows that dialysate infused as a bolus during on-line HDF produces a distinct dilution and volume expansion with an appreciable half-life and is feasible to identify cardiovascular parameters such as blood volume and volume residence time.

Saline has been used to measure cardiac output and central blood volume in experimental applications and during hemodialysis before.23–25 Dialysate is comparable to saline in its physical characteristics with the important advantage that it is available at the proper temperature and osmotic concentration and that it can be delivered under ultra-pure conditions with on-line HDF. One important difference to classic indicator dilution is the duration of the infusion. Unlike other techniques, the infusion is not instantaneous, but indicator volume is infused over a short albeit definite period of time. As a consequence, the infusion overlaps with the transit of indicator recirculating through the body and returning to the measuring site. Thus, the first transit of indicator passing through the heart and lungs and measured in arterial blood cannot be separated from subsequent transits, and the slow injection cannot be used to identify cardiac output using the classic Hamilton approach.26 However, the concentration of indicator after the injection can be used to identify the distribution volume and the elimination from the vascular space. The decrease in volume indicator observed in almost all dilutions indicates a slow loss of infused dialysate from the intravascular compartment. A plausible explanation for this loss is the reduction in plasma colloid osmotic pressure induced by the dilution thereby changing the microvascular filtration equilibrium so that outward filtration outweighs inward microvascular refilling.27 From the rate constant, the half-life of dialysate was estimated in the range of 10 to 15 min, which is somewhat longer than what has been reported for crystalloid solutions.28

The distribution volumes determined within the same treatment or within the same patient (on the occasion of subsequent treatments) were quite reproducible. The smallest volume was obtained in a patient with bilateral leg amputation (patient FR, V 0 = 3.6 ± 0.6 L, V a = 3.5 L, Figure 5). In the other patients, the volumes were consistently larger than those expected from anthropometrically derived estimates. This difference was not seen in in vitro studies were the experimental distribution volume was exactly known and where the bias of absolute volume estimations was only 1.3 ± 2.1%.29 The blood volume excess observed in dialysis patients is plausible as significant deviations from anthropometric estimates can be expected for the beginning of dialysis, before excess body fluid is removed from the patient by ultrafiltration. Mitra et al.30 for example observed that blood volumes derived from indocynanine-green (ICG) dilution were 20% larger than those estimated from four different anthropometric methods. Such discrepancies where not observed in a companion ICG study where blood volumes measured late during dialysis were not different from anthropometric estimates.31 Another explanation for increased blood volumes is the expansion of the circulation by the extracorporeal blood circuit by up to 300 ml. Part of the priming volume contained in the extracorporeal circulation is usually infused when the patient is connected to the extracorporeal circulation. Because of these effects in the magnitude of 0.5 L, we think that the measured distribution volume is a good estimate of total circulating blood volume. A comparison of initial distribution volumes to established reference volumes was not possible at that point. This is a limitation of this study.

Could the measured indicator concentrations have been too low and thus distribution volumes be overestimated? Access recirculation produces an early peak, is easily recognized, and will increase rather than decrease the arterial concentration of an indicator injected into the venous bloodline.32 Is indicator lost from the circulation during the first cardiopulmonary transit? Others have injected 50 ml of saline to measure cardiac output and have not reported such loss during pulmonary transit.25 One explanation for a possible underestimation of saline or dialysate is the preferred sequestration of cell-free volume indicator in the compliant venules and vessels of the microcirculation so that the increase in volume concentration measured in large vessels is blunted.33 Such an effect can be expected with the addition of volume described here as well as with the removal of volume by ultrafiltration described previously.34

Indicator concentration in arterial blood reaches a pronounced peak and declines more rapidly during the first few minutes after infusion, whereas the later phase appears to be adequately described by a mono-exponential decline with a reduced rate constant (Figure 4, left panel). The arterial overshoot can be explained by the superposition of cardiopulmonary indicator transits with mixed venous concentrations and is akin to the effects of so-called cardiopulmonary recirculation in hemodialysis.35 Consequently, such an overshoot was not observed when dilutions were done with central-venous accesses measuring mixed venous blood (Figure 4, right panel).

Figure 4

Figure 4

Figure 5

Figure 5

The approach described here is not limited to the ultrasonic measurement of blood water but can also be applied to on-line hemoglobin and hematocrit measurements available with almost all on-line HDF machines. It is, however, important that the extracorporeal sampling site is located upstream of the infusion site to record arterial or mixed venous patient concentrations, depending on the type of vascular access. Automatic volume infusions provided by the machine are very helpful. This bypasses the requirement of manual infusion and avoids all risks associated with such manipulation. The test should also be possible with different forms of hemofiltration used in acute treatments. The technique has the potential for complete automation with appropriate changes in system software only. Changes in hardware are not required.

Infusion of saline appears to defeat the purpose of volume removal during hemodialysis and other forms of renal replacement therapy. However, in complicated treatments, on-line infusion of dialysate will help to improve hemodynamic stability because of intravascular volume expansion and at the same time provide information whether stability is related to a reduced intravascular volume or to impaired vascular tone. In acute therapy, fluid requirements are assessed using various measures of fluid responsiveness.36 Interestingly, little information is available on the efficiency of infusions to expand the intravascular space in acute treatments.28,37 This test may therefore help to assess to which degree infusion volume remains within the circulation and to establish a measure for intravascular volume efficiency. Infusion volume leaking into the tissue has lost its hemodynamic effect while impairing tissue oxygenation because it aggravates peripheral edema.

In conclusion, on-line dilution of dialysate is easily done and produces a reproducible and distinct dilution curve to estimate the distribution volume of isotonic dialysate as well as the escape rate of fluid into peripheral fluid compartments. Given the factors that tend to increase the blood volume of dialysis patients, the initial distribution volume is consistent with blood volume determined from anthropometric measurements. Such information could be of interest for volume management in hemodialysis patients and more importantly for the management of fluid therapy in intensive care patients undergoing extracorporeal blood treatment. We call upon manufacturers to consider implementing this measuring technique into machines for treatment of chronic and acute renal failure.

Back to Top | Article Outline


The total mass (M, in kg) of a fluid is given by its density (ρ, in kg/L) and volume (V, in L) as

Therefore, the mass of a solution after mixing (index m) is given as the sum of the mass before mixing and the mass of added indicator (index ind) as

The mass of water (M W, in kg) in a fluid is given by the mass of the fluid (ρV) and the water content (W, in kg/kg) of that fluid as

Accordingly, mass balance for water before and after addition of a volume indicator is given as

Insertion of Equation 5 into Equation 7 and eliminating V m ρ m yields

or in more general notation

Here, C ind refers to the volume concentration of the indicator (in vol/vol). With water as indicator, C ind is given as

The water content of the indicator is known; the water content before and after dilution can be measured with available technology.21,38 Fluid density at a given water content W and temperature T (in °C) can be determined from the following relationship39:

The ratio ρ/ρ ind is largely independent of the range of hematocrit values encountered in hemodialysis. For the purpose of estimating V 0 in this setting, the density ratio was assumed as constant (ρ/ρ ind=1.05).

In case of isotonic dilution, the indicator concentration can also be derived from the relative blood volume (RBV, in vol/vol) measured before and after (index m) addition of volume indicator as

Relative blood volume changes can be measured by a variety of techniques.40

Back to Top | Article Outline


The authors thank A. Wüpper, Fresenius Medical Care, Bad Homburg vor der Höhe, Germany, for financial support and J. Gross for excellent technical help with the studies.

Back to Top | Article Outline


1. Kim KE, Neff M, Cohen B, et al. Blood volume changes and hypotension during hemodialysis. Trans Am Soc Artif Intern Organs. 1970;16:508–514
2. Lewicki MC, Kerr PG, Polkinghorne KR. Blood pressure and blood volume: Acute and chronic considerations in hemodialysis. Semin Dial. 2013;26:62–72
3. Daugirdas JT. Pathophysiology of dialysis hypotension: An update. Am J Kidney Dis. 2001;38:S11–S17
4. Vincent JL, Gottin L. Type of fluid in severe sepsis and septic shock. Minerva Anestesiol. 2011;77:1190–1196
5. Grände P-O, Persson JHahn GR. Microvascular fluid exchange. Clinical Fluid Therapy in the Perioperative Setting. 2011 Cambridge, UK Cambridge University Press:120–126
6. Cordemans C, De Laet I, Van Regenmortel N, et al. Aiming for a negative fluid balance in patients with acute lung injury and increased intra-abdominal pressure: A pilot study looking at the effects of PAL-treatment. Ann Intensive Care. 2012;2(suppl 1):S15
7. Malbrain ML, Van Regenmortel N. Fluid overload is not only of cosmetic concern: Exploring a new hypothesis. ICU Management. 2012;12:30–33
8. Mann H, Stiller S, Schallenberg U, Thömmes A. Optimizing dialysis by variation of ultrafiltration rate and sodium concentration controlled by continuous measurement of circulating blood volume. Contrib Nephrol. 1989;74:182–190
9. Steuer RR, Leypoldt JK, Cheung AK, Harris DH, Conis JM. Hematocrit as an indicator of blood volume and a predictor of intradialytic morbid events. ASAIO J. 1994;40:M691–M696
10. Paolini F, Bosetto A. Biofeedback systems architecture. Adv Ren Replace Ther. 1999;6:255–264
11. Barth C, Boer W, Garzoni D, et al. Characteristics of hypotension-prone haemodialysis patients: Is there a critical relative blood volume? Nephrol Dial Transplant. 2003;18:1353–1360
12. Davenport A. Can advances in hemodialysis machine technology prevent intradialytic hypotension? Semin Dial. 2009;22:231–236
13. Dasselaar JJ, van der Sande FM, Franssen CF. Critical evaluation of blood volume measurements during hemodialysis. Blood Purif. 2012;33:177–182
14. Nette RW, Krepel HP, van den Meiracker AH, Weimar W, Zietse R. Specific effect of the infusion of glucose on blood volume during haemodialysis. Nephrol Dial Transplant. 2002;17:1275–1280
15. van der Sande FM, Luik AJ, Kooman JP, Verstappen V, Leunissen KML. Effect of intravenous fluids on blood pressure course during hemodialysis in hypotensive-prone patients. J Am Soc Nephrol. 2000;11:550–555
16. Schallenberg U, Stiller S, Mann H. A new continuous haemoglobinometric measurement of blood volume during haemodialysis. Life Support Syst. 1987;5:293–305
17. Leypoldt JK, Cheung AK, Steuer RR, Harris DH, Conis JM. Determination of circulating blood volume by continuously monitoring hematocrit during hemodialysis. J Am Soc Nephrol. 1995;6:214–219
18. Schneditz D, Roob JM, Vaclavik M, Holzer H, Kenner T. Noninvasive measurement of blood volume in hemodialysis patients. J Am Soc Nephrol. 1996;7:1241–1244
19. McMahon MP, Campbell SB, Shannon GF, Wilkinson JS, Fleming SJ. A non-invasive continuous method of measuring blood volume during haemodialysis using optical techniques. Med Eng Phys. 1996;18:105–109
20. Schneditz D, Kaufman AM, Levin NW. On-line measurement of blood water concentration in the extracorporeal circulation of hemodialysis patients. Proc. IEEE EMBS. 1991;13:1651–1652
21. Schneditz D, Pogglitsch H, Horina J, Binswanger U. A blood protein monitor for the continuous measurement of blood volume changes during hemodialysis. Kidney Int. 1990;38:342–346
22. Nadler SB, Hidalgo JH, Bloch T. Prediction of blood volume in normal human adults. Surgery. 1962;51:224–232
23. Moser M, Kenner T. Blood flow and blood volume determinations in aorta and in coronary circulation by density dilution. Basic Res Cardiol. 1988;83:577–589
24. Schneditz D, Probst W, Kubista H, Binswanger U. Kontinuierliche Blutvolumenmessung im extrakorporellen Kreislauf mit Ultraschall. Nieren- und Hochdruckkrankheiten. 1991;20:649–652
25. Krivitski NM, Depner TA. Cardiac output and central blood volume during hemodialysis: Methodology. Adv Ren Replace Ther. 1999;6:225–232
26. Lassen NA, Henriksen O, Sejrsen PShepherd JT, Abboud FM. Indicator methods for measurement of organ and tissue blood flow Handbook of Physiology. Section 2: The Cardiovascular System. 1983 Bethesda, MD American Physiological Society:21–63 Volume 3
27. Schneditz D, Roob JM, Oswald M, Pogglitsch H, Moser M, Kenner T. Nature and rate of vascular refilling during hemodialysis and ultrafiltration. Kidney Int. 1992;42:1425–1433
28. Hahn GRHahn GR. Crystalloid fluids. Clinical Fluid Therapy in the Perioperative Setting. 2011 Cambridge, UK Cambridge University Press:1–10
29. Schneditz D, Christmann H, Krämer M. Indicator dilution by controlled infusion of ultra-pure dialysate. ASAIO J. 2002;48:174
30. Mitra S, Chamney PW, Greenwood RN, Farrington K. Serial determinations of absolute plasma volume with indocyanine green during hemodialysis. J Am Soc Nephrol. 2003;14:2345–2351
31. Schneditz D, Haditsch B, Jantscher A, Ribitsch W, Krisper P. Absolute blood volume and hepato-splanchnic blood flow measured by indocyanine green kinetics during hemodialysis. ASAIO J. 2014 doi: 10.1097/MAT.0000000000000075.
32. Schneditz D, Polaschegg HD, Levin NW, et al. Cardiopulmonary recirculation in dialysis. An underrecognized phenomenon. ASAIO J. 1992;38:M194–M196
33. LaForte AJ, Lee LP, Rich GF, Skalak TC, Lee JS. Blood volume redistribution from a passive elastic permeable microcirculation due to hypovolemia. Am J Physiol. 1994;266(6 Pt 2):H2268–H2278
34. Dasselaar JJ, Lub-de Hooge MN, Pruim J, et al. Relative blood volume changes underestimate total blood volume changes during hemodialysis. Clin J Am Soc Nephrol. 2007;2:669–674
35. Schneditz D, Kaufman AM, Polaschegg HD, Levin NW, Daugirdas JT. Cardiopulmonary recirculation during hemodialysis. Kidney Int. 1992;42:1450–1456
36. Cannesson MHahn GR. Non-invasive guidance of fluid therapy. Clinical Fluid Therapy in the Perioperative Setting. 2011 Cambridge, UK Cambridge University Press:103–111
37. Svensén C, Hahn RG. Volume kinetics of Ringer solution, dextran 70, and hypertonic saline in male volunteers. Anesthesiology. 1997;87:204–212
38. Schneditz D, Kenner T, Heimel H, Stabinger H. A sound speed sensor for the measurement of total protein concentration in disposable, blood perfused tubes. J Acoust Soc Am. 1989;86:2073–2080
39. Schneditz D, Heimel H, Stabinger H. Sound speed, density and total protein concentration of blood. J Clin Chem Clin Biochem. 1989;27:803–806
40. Dasselaar JJ, Huisman RM, DE Jong PE, Franssen CF. Relative blood volume measurements during hemodialysis: Comparisons between three noninvasive devices. Hemodial Int. 2007;11:448–455

ultra-pure dialysate; indicator dilution; volume kinetics; blood volume; half-life

Copyright © 2014 by the American Society for Artificial Internal Organs