Background: Data generated with the body composition monitor (BCM, Fresenius) show based on bioimpedance technology, that chronic fluid overload in hemodialysis patients is associated with poor survival. However, removing excess fluid by lowering dry weight can be accompanied by intradialytic and postdialytic complications. Here, we aim at testing the hypothesis that, in comparison to conventional hemodialysis, blood-volume monitored regulation of ultrafiltration and dialysate conductivity (UCR) and/or regulation of ultrafiltration and temperature (UTR) will decrease complications when ultrafiltration volumes are systematically increased in fluid overloaded hemodialysis patients.

Methods/Design: BCM-measurements yield results on fluid overload (in liters), relative to extracellular water (ECW). In this prospective, multicenter, triple-arm, parallel group, cross-over, randomized, controlled clinical trial, we use BCM-measurements, routinely introduced in our 3 maintenance hemodialysis centers shortly prior to the start of the study, to recruit 60 hemodialysis patients with fluid overload (defined as ≥15% ECW). Patients are randomized 1:1:1 into UCR, UTR and conventional hemodialysis groups. BCM-determined, 'final' dry weight is set as -7% ECW postdialysis, and reached by reducing the previous dry weight, in steps of 0.1 kg per 10 kg body weight, during 12 hemodialysis sessions (one study phase). In case of intradialytic complications, dry weight reduction is decreased, according to a pre-specified algorithm. A comparison of intra- and postdialytic complications among study groups constitutes the primary endpoint. In addition, we will assess relative weight reduction, changes in residual renal function, quality of life measures, and predialysis levels of various laboratory parameters including C-reactive protein, troponin T, and N-terminal pro-B-type natriuretic peptide, before and after the first study phase (secondary outcome parameters).

Discussion: Patients are not requested to revert to their initial degree of fluid overload after each study phase, Therefore, the cross-over design of the present study merely serves the purpose of secondary end-point evaluation, for example to determine patient choice of treatment modality. Previous studies on blood volume monitoring have yielded inconsistent results. Since we include only patients with BCM-determined fluid overload, we expect a benefit for all study participants, due to strict fluid management which decreases the mortality risk of hemodialysis patients.

Hypothesis In comparison to conventional hemodialysis, regulation of ultrafiltration and dialysate conductivity (UCR) and/or regulation of ultrafiltration and temperature (UTR) will decrease intra- and postdialytic complications when ultrafiltration volumes are systematically increased in fluid overloaded hemodialysis patients.

Objectives and Outcome Measures The primary objective is to demonstrate superiority of ultrafiltration and dialysate conductivity regulation (UCR) and/or ultrafiltration and temperature regulation (UTR) over conventional hemodialysis, in preventing intra- and postdialytic complications, when fluid overloaded hemodialysis patients receive systematic fluid reduction, to reach a final dry weight of 7% ECW postdialysis. The primary outcome measure is the total number of hemodialysis sessions per patient that were accompanied, intra- or postdialytically, by at least one symptom most likely related to fluid withdrawal (as specified in Table 1), divided by the number of hemodialysis sessions at risk (as by study protocol: 12 sessions per patient in study phase 1). Both groups, UCR and UTR, will be compared against the conventional hemodialysis group, and afterwards against one another, using the two-sided Student's t-test.

The secondary objectives are:

1. To demonstrate superiority of ultrafiltration and dialysate conductivity regulation (UCR) and/or ultrafiltration and temperature regulation (UTR) over conventional hemodialysis, in preventing specific intradialytic complications:

   1. intradialytic cramping
   2. clinically asymptomatic, intradialytic hypotension (>40 mmHg drop in systolic blood pressure within 30 minutes)
   3. clinically symptomatic, intradialytic hypotension (>40 mmHg drop in systolic blood pressure within 30 minutes)
   4. clinically symptomatic, intradialytic hypotension (even if it is not possible to identify a sudden drop in blood pressure, e.g. patients may slowly move towards low blood pressure values [f. ex. below 100 mmHg systolic blood pressure], and report symptoms)
   5. unspecified intradialytic symptoms or events, which are most likely related to fluid withdrawal
   6. patient-reported postdialysis complication, most likely related to fluid withdrawal

   when fluid overloaded hemodialysis patients receive systematic fluid reduction, to reach a final dry weight of -7% ECW postdialysis. The secondary outcome measure for objective [a] is the total number of hemodialysis sessions per patient which were accompanied by the respective symptom most likely related to fluid withdrawal (1 to 6), divided by the number of hemodialysis sessions at risk (as by study protocol: 12 sessions per patient in study phase 1). Both groups, UCR and UTR, will be compared against the conventional hemodialysis group, and afterwards against one another, using the two-sided Student's t-test.

2. To demonstrate superiority of UCR and/or UTR over conventional hemodialysis in allowing patients to reach a lower body weight, relative to his/her postdialysis weight at the beginning of study phase 1 (time zero). The secondary outcome measure for objective [b] is the difference in postdialysis body weight from time zero to the end of study phase 1, divided by the postdialysis body weight at time zero.

   Example: If a patient has a postdialysis body weight of 65 kg at time zero (the last hemodialysis session before the first dry weight reduction), and reaches a postdialysis body weight of 62 kg at the end of the first study phase, his difference in body weight will be 65 kg - 62 kg =3 kg. Relative to this patient's postdialysis body weight at time zero, weight reduction from beginning to end of study phase 1 will be 3/65*100=4.61%.

   The relative weight reduction in both groups, UCR and UTR, will be compared against the conventional hemodialysis group, and afterwards against one another, using the two-sided Student's t-test.

3. To assess if stricter volume control in fluid overloaded patients (using UCR, UTR or conventional hemodialysis) diminishes residual renal function. The secondary outcome measure for objective [c] is the difference in 24h urine volume from time zero to the end of study phase 1, divided by the urine volume at time zero. The relative urine reduction in both groups, UCR and UTR, will be compared against the conventional hemodialysis group, and afterwards against one another, using the two-sided Student's t-test.

4. To assess the amount of sodium transferred to the patient or withdrawn from the patient during conventional hemodialysis, UCR and UTR.

   The amount of sodium removed during hemodialysis is a function of the ultrafiltration volume, the effective diffusion gradient for sodium, and diffusive sodium clearance. The effective diffusion gradient depends on the plasma water - dialysate sodium (DNa) difference, and the Gibbs-Donnan coefficient, the latter being a function of the plasma protein concentration but also being influenced by other ions in the dialysate. The assumption that sodium removal by ultrafiltration is equal to the plasma water sodium concentration, multiplied by the ultrafiltration volume, is a simplification. Electrolyte balances may also be influenced by the membrane charge. Therefore, the amount of sodium transferred will be measured on the dialysate side. Sodium transfer will be calculated from the difference between the mean sodium concentration in partially collected, used dialysate and the fresh DNa concentration, multiplied by the total amount of dialysate used. Results will be compared to a two pool sodium model and parameters of the model will be adjusted using the measured transfer data.

   The above described method is not applicable to the UCR group because the sodium concentration of fresh dialysate is not kept constant during treatment. In order to establish the model, we will measure at least 10 patients during 3 hemodialysis sessions, and due to the formerly stated, we can only use the dialysate from patients in the conventional hemodialysis group and in the UTR group.

   However, after having established this model, sodium transfer for all patients - conventional, UTR, and UCR - can be calculated using this very model.

   The amount of sodium transferred to or withdrawn from the patient will be calculated for each patient, according to the model established in the way stated here above. The secondary outcome measure for objective [d], e.g. the sodium transferred in both groups, UCR and UTR, will be compared against the conventional hemodialysis group, and afterwards against one another, using the two-sided Student's t-test.

5. To assess if stricter volume control by conventional hemodialysis, UCR and/or UTR influences predialysis serum concentrations of (1) C-reactive protein, (2) D-dimer, (3) fibrinogen, (4) troponin T, (5) N-terminal pro-B-type natriuretic peptide. These proteins are used as read-outs for inflammation (1, 2, 3), coagulation (2, 3), and cardiac function (4, 5) and are routinely determined at all three participating centers. The secondary outcome measure for objective [e], e.g. the concentrations of the indicated laboratory parameters in both groups, UCR and UTR, at time zero and at the end of study phase 1, will be compared against the conventional hemodialysis group, and afterwards against one another, using the two-sided Student's t-test. The individual change in these parameters, from time zero to the end of study phase 1, will be compared likewise.

6. To assess if stricter volume control in fluid overloaded patients (using UCR, UTR or conventional hemodialysis) affects the concentration of various proteins that might serve as novel biomarkers (including high density lipoprotein [HDL]-associated surfactant protein B, HDL-associated serum amyloid A, HDL-associated apoC-II, plasma tryptophan, plasma choline, plasma trimethylanine-N-oxide; plasma endotoxin).

   We have recently discovered that various proteins, including those mentioned here above, may render HDL dysfunctional, in hemodialysis patients. Especially with surfactant, we suspect an association with fluid overload. Moreover, plasma tryptophan, plasma choline, and plasma trimethylanine-N-oxide are markers of end stage renal disease. Since one or several of these proteins or molecules might serve as novel biomarkers, we would like to use the context of the present trial to establish diagnostic assays and analyze potential changes in the serum concentration of these proteins, before and after fluid removal.

   Endotoxin, found in the cell membranes of gram-negative bacteria, is a biomarker representing the gut flora. Endotoxin has been shown to be associated with inflammation, nutritional status and even mortality in hemodialysis patients. An incremental rise in endotoxin levels has been shown along with the progression of chronic kidney disease, and especially initiation of hemodialysis. It has been suggested that hemodialysis patients might have high endotoxin levels in the blood due to repeated bacterial translocation from the gut during hemodialysis, secondary to intradialytic changes in blood pressure and/or tissue perfusion. An association with chronic fluid overload has not yet been established, but might be suspected, as a consequence of higher ultrafiltration rates and thus decreased intradialytic stability. Here we will measure endotoxin levels with an amebocyte limulus assay in all study patients, at time zero and at the end of study phase 1.

   The secondary outcome measure for objective [f], for example HDL-associated serum amyloid A in both groups, UCR and UTR, at time zero and at the end of study phase 1, will be compared against the conventional hemodialysis group, and afterwards against one another, using the two-sided Student's t-test. The individual change in these parameters, from time zero to the end of study phase 1, will be compared likewise. The results of this secondary endpoint analysis will be published separately from the clinical results of the present study.

7. To assess if stricter volume control in fluid overloaded patients (using UCR, UTR or conventional hemodialysis) affects quality of life measures. The secondary outcome measures for objective [g] are the mental component summary (MCS) and physical component summary (PCS) derived from the Kidney Disease Quality of Life Short Form (KDQoL-SFTM). The scale of both summary scores is 0-100 (higher indicating better quality of life). MCS and PCS in both groups, UCR and UTR, at time zero and at the end of study phase 1, will be compared against the conventional hemodialysis group, and afterwards against one another, using the two-sided Student's t-test. The individual change in these parameters, from time zero to the end of study phase 1, will be compared likewise.

8. To demonstrate superiority of UCR and/or UTR over conventional hemodialysis in reducing dialysis complications when previously fluid overloaded patients are entering phase 2 and phase 3 of the presented study. In study phases 2 and 3, patients either have to reduce their dry weight further, or else have to maintain their newly reached dry weight. The secondary outcome measures for objective [h] are potentially all of the items described here above, and will be compared against the conventional hemodialysis group, and afterwards against one another, in the same fashion as described here above.

9. To assess which hemodialysis treatment modality will be chosen by the patients at the end of the study. The choice of treatment will be compared as follows: (1) number of patients choosing UCR-treatment against number of patients choosing conventional treatment; (2) number of patients choosing UTR-treatment against number of patients choosing conventional treatment; (3) number of patients choosing UCR-treatment against number of patients choosing UTR-treatment. The statistical test for analyses (1), (2) and (3) will be the two-sided chi square test, for each analysis.

Superiority definition For all objectives listed, superiority will be assumed if a statistically significant difference between one group versus another group can be determined. With regards to the primary endpoint, according to our sample size calculation (see below), this study is suited to detect a minimal difference of 10% between groups.

Inclusion Criteria:

• HD >3 months
• Fluid overload ≥15% of extracellular volume (ECV) predialysis, as determined by BCM-measurement, after one of the 2 the midweek interdialytic intervals
• Informed consent of the patient

Exclusion Criteria:

• none, except those specified by the Declaration of Helsinki