NDT Advance Access published online on August 12, 2008
Nephrology Dialysis Transplantation, doi:10.1093/ndt/gfn457
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A comparison of methods for determining urea distribution volume for routine use in on-line monitoring of haemodialysis adequacy
1 Department of Renal Medicine, St James's University Hospital, Leeds, UK 2 Fresenius Medical Care, Research & Development, Bad Homburg, Germany
Correspondence and offprint requests to: Elizabeth J. Lindley, Department of Renal Medicine, St James's University Hospital, Beckett Street, Leeds LS9 7TF, UK. Tel: +44-113-206-4199; Fax: +44-113-206-6064; E-mail: Elizabeth.Lindley{at}leedsth.nhs.uk
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionAppendix A: Watson formulae...Appendix B: Calculation of...References |
|---|
Background. The availability of haemodialysis machines equipped with on-line clearance monitoring (OCM) allows frequent assessment of dialysis efficiency and adequacy without the need for blood samples. Accurate estimation of the urea distribution volume ‘V’ is required for Kt/V calculated from OCM to be consistent with conventional blood sample-based methods.
Methods. Ten stable HD patients were monitored monthly for 6 months. Time-averaged OCM clearance (KOCM) and pre- and post-dialysis blood samples were collected at each monitored session. The second generation Daugirdas formula was used to calculate the single-pool variable volume Kt/V, (Kt/V)D. Values of V to allow comparison between OCM and blood-based Kt/V were determined from Watson's formula (VWatson), bioimpedance spectroscopy (VBIS), classical urea kinetic modelling (VUKM_C) and a simple computation of V (VUKM_S) from the blood-based Kt/V and KOCMt.
Results. Comparison of KOCMt/V with (Kt/V)D shows that using VWatson leads to significant systematic underestimation of dialysis dose. KOCMt/VBIS agrees with (Kt/V)D to within ± 10%. KOCMt/VUKM_S is, by definition, identical to (Kt/V)D when initially calculated. However, if a historical value of V is used, agreement between KOCMt/V and (Kt/V)D over 6 months varies by 5% for VBIS and 10% for VUKM_S.
Conclusions. When investigating the effect of different treatment strategies on dialysis efficiency, any estimate of V can be used provided it is constant, as K is the relevant parameter. When frequent supervision of actual dialysis dose is required, the greatest consistency between KOCMt/V and the reference, Kt/VD, over time is achieved with VBIS.
Keywords: haemodialysis adequacy; on-line clearance monitoring; urea distribution volume; urea kinetic modelling
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix A: Watson formulae...Appendix B: Calculation of...References |
|---|
Regular measurement of dialysis dose is an essential element in assuring the quality of renal replacement therapy. While blood sampling and measurement of blood urea nitrogen (BUN) is the mainstay of dose assessment, the method is rarely performed more frequently than once a month for both financial and practical reasons. Dialysis machines offering on-line monitoring of dialysis efficiency are now widely available. On-line clearance monitors measure the difference in conductivity between the dialysate entering and leaving the dialyser with two different dialysate inlet electrolyte concentrations. These measurements can be used to calculate the ionic dialysance, which is equal to the effective urea clearance, provided dialysate flow, blood flow and blood electrolyte composition are constant during the measuring interval [1]. On-line clearance monitoring (OCM) allows dialysis dose to be monitored at every treatment with virtually no additional overheads. While it is unlikely that these non-invasive measurements of Kt/V will replace routine blood sampling, OCM affords staff the opportunity to monitor unstable patients more effectively, identify problems quickly and assess the effect of remedial actions.
The use of OCM is based on the finding that ionic dialysance and urea clearance are equivalent. This has been demonstrated in a number of studies that have investigated the influence of recirculation and blood water [2] and shown that KOCM is comparable with blood water urea clearance, taking into account recirculation [3]. The OCM thus provides the effective urea clearance. In order to calculate Kt/V, the patient's urea distribution volume is needed as an input.
The purpose of this study was to assess which of the available methods for determination of V lead to a Kt/V that is consistent with a reference Kt/V based on pre- and post-dialysis blood urea concentrations. As KOCM does not take into account rebound (two-pool) effects, it is appropriate to choose a single-pool Kt/V as the blood-based reference method. For this study, the well-established second generation Daugirdas ‘single-pool variable volume’ spVV Kt/V, (Kt/V)D, was used [4].
Since it is impractical to measure V at every treatment, we evaluated the agreement between the OCM Kt/V and the reference method when a single fixed value of V is used over a period of months.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix A: Watson formulae...Appendix B: Calculation of...References |
|---|
Stable haemodialysis patients on three times weekly treatment at St James's University Hospital, Leeds, were recruited for the 6-month study. The study protocol was approved by the Leeds East Local Ethics Committee, and written informed consent was obtained from each patient.
Changes in target weight were made as necessary, based on blood pressure, intradialytic symptoms and blood volume monitoring as per normal practice. After the initial baseline measurement, each patient was monitored on a monthly basis resulting in seven study sessions per patient. Residual renal function, if any, was determined from an interdialytic (48-h) urine collection.
A 4008H machine (Fresenius Medical Care, Bad Homburg, Germany) equipped with OCM was used for all study sessions. Patients were dialyzed according to their individual prescription with a Fresenius F10HPS or HdF100S dialyzer and a dialysate flow rate of 500 or 800 ml/min. The effective dialysis time and total blood volume processed were recorded and used to calculate the time-averaged blood flow.
The effective urea clearance measured by conductivity monitoring, KOCM, was recorded automatically every 25 min throughout each dialysis session. The OCM module was programmed with the anthropometric V calculated using the Watson formula (see Appendix A), which is the default for this device. The Watson V has been shown to correlate more closely with the modelled urea distribution volume than the volume determined using the Hume–Weyers or Chertow equation in the analysis of the HEMO study [5] and in our own previous study [6].
Bioimpedance measurements were performed before each dialysis using the Hydra bioimpedance analyzer (Model 4200, Xitron Technologies, San Diego, CA, USA). The measurement was carried out using the tetrapolar wrist-to-ankle (non-fistula side) method after the patient had rested for 10 min in a supine position. This method has been validated against gold standard dilution techniques in healthy controls and dialysis patients [7]. Total body water (TBW) at post-dialysis weight (VBIS) was obtained by subtracting the intradialytic weight loss from the measured pre-dialysis TBW.
Blood samples were obtained pre-dialysis (C0) and immediately post-dialysis (Ct). The post-dialysis sample was taken from the arterial needle tubing after slowly reducing the blood pump speed to 50 ml/min and waiting 5 seconds before stopping the pump (to eliminate any recirculated blood from the sample).
The blood-based spVV Kt/V, (Kt/V)D, was calculated using the Daugirdas formula:
|
|
The ‘kinetic’ urea distribution volume was obtained from C0 and Ct using two methods. The ‘classical’ VUKM_C was determined from the two-point classical single-pool variable volume model [8] as described in Appendix B, while the ‘simple’ VUKM_S was calculated by dividing KOCMt by the blood-based (Kt/V)D.
Non-invasive Kt/V was calculated by multiplying KOCM with the effective dialysis time and dividing by either the anthropometric volume VWatson, the bioimpedance volume VBIS or the kinetic volume VUKM_C or VUKM_S. In order to avoid mathematical coupling when comparing (Kt/V)D with KOCM x t/VUKM_C or KOCM x t/VUKM_S (as the same blood samples are common to all calculations), we followed the procedure outlined by Di Filippo [9]. Therefore, values for VUKM_C or VUKM_S were taken from the study session for the previous month, giving one less comparison per patient for these two methods. Corrections to V caused by changes in post-weight since the previous dialysis were also taken into account. All results were calculated as mean ± 2 x SD (i.e. 95% confidence interval).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix A: Watson formulae...Appendix B: Calculation of...References |
|---|
Ten patients (six males, four females; median age 48 years, range 22–62; median time on renal replacement therapy 4.8 years, range 0.2–19.3 years) participated in the study. One patient was successfully transplanted after 4 months and was included in the analysis but not replaced. Technical problems occurred during three treatments so complete datasets were obtained in 65/68 study sessions.
The patients’ body mass index ranged from 17.6 to 30.8 kg/m2 (mean 25.0 ± 4.1 kg/m2) and post-dialysis weight ranged from 44.5 to 105.4 kg (mean 75.5 ± 18.5 kg). Only two patients had significant residual renal function (>150 ml of urine per day). The effective dialysis time recorded by the dialysis machine ranged from 170 to 250 min (mean 217 ± 24 min). Time-averaged blood flow, calculated from the total blood volume processed during the session divided by the dialysis time, ranged from 224 to 417 ml/min (mean 339 ± 35 ml/min).
The single-pool variable volume Kt/V determined from blood samples, spVV (Kt/V)D, averaged over the 6-month study, varied from 1.22 (patient 1) to 2.49 (patient 7). The post-dialysis urea distribution volume, again averaged over the study period, obtained from the methods under investigation (i.e. the anthropometric volume, VWatson, the bioimpedance volume, VBIS, or the kinetic volume, VUKM_C or VUKM_S) is shown in Table 1.
|
The discrepancy between the non-invasive Kt/V determined from the OCM clearance and the reference (Kt/V)D from blood samples is shown in Figure 1. By Bland–Altman analysis, use of V derived from the Watson formula to determine KOCMt/VWatson led to systematic underestimation of the dialysis dose by –0.36 ± 0.42 (P < 0.001). Use of the classical UKM V (determined from the previous session) to give KOCMt/VUKM_C led to a small overestimation of the spVV Kt/V of +0.04 ± 0.34 (NS), while in the case of the simple UKM V (VUKM_S) the difference to (Kt/V)D was found to be –0.01 ± 0.31 (NS). In this group of patients, the mean difference between KOCMt/VBIS and (Kt/V)D was –0.03 ± 0.25 (NS). Although there was no significant bias for the group, a systematic discrepancy was found between the methods for some patients as shown in Figure 2.
|
VUKM_C and VUKM_S are calculated from data collected the previous month.
|
During routine monitoring, VUKM_C is unlikely to be used due to the complex calculations required. In practice, staff using a V obtained from bioimpedance or simplified UKM (VBIS or VUKM_S), rather than the default anthropometric V, would most likely program the OCM module with a value determined from an earlier session. The effect of using a V obtained up to 6 months earlier, rather than in the same session, is shown in Figure 3a and b for methods VBIS or VUKM_S, respectively. Data for the patient who was transplanted during the study have been excluded. These figures reflect both the reproducibility of the method for determining V and actual variations of V in each patient over time throughout the study. In both methods (VBIS and VUKM_S), there was no significant difference between the initial Kt/V and the Kt/V at 6 months when the initial V was assumed constant.
|
Table 2 shows the change in post-dialysis weight and volume calculated by different methods over the 6-month study determined from a linear regression of the data. The mean weight change was a gain of 1.35 ± 1.34 kg. The standard error of the estimate (
est), which is based on the deviation of the measured or calculated values from the predicted regression line, provides a measure of the reproducibility of each method for determination of volume. Table 2 indicates that VBIS has better reproducibility than VUKM_S.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix A: Watson formulae...Appendix B: Calculation of...References |
|---|
This study confirmed that anthropometrically estimated volumes are significantly larger than the urea distribution volume determined from UKM or bioimpedance. This has been demonstrated by Daugirdas et al. in the HEMO study [5] and by Dumler [10]. A recent study undertaken by Moret et al. [11] has also concluded that volume determined by UKM or ionic dialysance is preferable to anthropometric volumes. If a more accurate value for V is not available, the OCM module will default to Watson's anthropometric volume if provided with the height, dry weight, age and gender of the patient. Although inaccurate, using VWatson is a ‘safe’ default as it will almost invariably lead to underestimation of the Kt/V.
Non-invasive measurements of Kt/V using OCM can assist in the supervision of dialysis adequacy in three ways:
- investigation of causes of a highly variable Kt/V (for example, use of different cannulation sites or heparin regimes);
- assessment of the effect of interventions to increase an inadequate Kt/V (such as increasing blood or dialysate flow);
- more frequent monitoring of Kt/V when there is concern about the delivered dialysis dose (for example, when using a catheter with poor flows, where a patient has unexplained high potassium levels).
When undertaking investigations and during assessment of changes that occur over a relatively short period, a fixed value of the volume used for KOCMt/V should be entered for each treatment monitored. With this approach, changes in K can be clearly distinguished making it easier to identify sources of dialysis inefficiency.
Our study indicated that bioimpedance volume offers good agreement with the blood measurements under examination. The small (5%), but systematic, discrepancy between KOCMt/VBIS and (Kt/V)D in some patients, as was shown in Figure 2, was unexplained. There was no obvious association with body composition, but errors in VBIS due to assumed body proportions and tissue composition may have contributed to the discrepancy. Assumptions in the calculation of reference (Kt/V)D, such as the urea generation rate, may also have introduced sources of error. However, it remains clear that where possible, bioimpedance should be the method of choice for determination of V. If bioimpedance is not available, the Watson volume can be used and the effect of successful interventions confirmed using blood samples.
When using OCM for longer term frequent monitoring, VBIS or VUKM_S can be used. Our data indicate that VBIS has higher reproducibility than VUKM_S. The reproducibility of VBIS depends on electrode positioning and contact, and various patient-related factors that are relatively easy to control, whilst for VUKM_S, reproducibility is affected by errors in blood sampling and variation in lab results, as well as inaccurate treatment time or weight loss. Averaging VUKM_S over several sessions would reduce the scatter in Figure 3b, and it has been shown by others that this procedure is necessary to obtain a sufficiently reliable estimate of V with blood sample-based methods [12].
A study by Di Filippo et al. [13] suggested that V might be sufficiently stable over the period of a year, despite a downward trend, to allow V to be determined once per year with minimal impact on Kt/V. In our study, there was no significant variation in Kt/V over a 6-month period when assuming constant V in an individual patient, regardless of the method used to determine V. This adds further credence to Di Filippo's proposal for the use of constant V. Table 2 shows changes in weight and volume occurred in either direction in 6 months even in our stable patient cohort. Although these changes transpired to be unimportant in terms of Kt/V in our study, we clearly cannot infer that true V can be assumed stable for prolonged periods in individual patients.
Ultimately the question of when to reassess the patient's V cannot be easily answered with a general guideline. In any situation that could influence a patient's target weight, such as illness, lifestyle changes or starting nutritional support, more frequent volume assessments are necessary. The advantage of BIS is that V can be determined with minimal overhead and therefore any doubts about the long-term stability of V in a patient can be easily addressed by more frequent monitoring, perhaps on a monthly basis. Furthermore, we have demonstrated that VBIS offers the lowest variability in measurement of V which makes the method better suited for frequent monitoring than more costly and time-consuming blood sampling.
In some studies, virtual Vs have been obtained from (eKt/V) and measurements of ionic dialysance (KID) [9]. The disadvantage of this approach is that the resulting V is influenced by dialysis efficiency, so care must be exercised when comparing this virtual V against reference methods. A virtual V can only be used as an input to OCM in subsequent treatments provided there is no change in treatment efficiency. By comparison, we used the single-pool Kt/V to calculate VUKM_S as it is closer to the TBW and provides more direct comparison with VBIS. Furthermore, the single-pool KOCMt/V can be converted to an eKt/V with the same formulae that are applied to convert the spVV Kt/V derived from blood samples to an equilibrated Kt/V [14]. As the only input parameter required is the treatment time, machines equipped with OCM could be modified to calculate an arterial- and venous-equilibrated Kt/V automatically.
In conclusion, OCM can be used to improve supervision and optimization of dialysis dose. When investigating poor or unstable Kt/V, a constant value for the urea distribution volume should be used and the method for determining V is not important. For more frequent monitoring of Kt/V, where consistency between KOCMt/V and the spVV Kt/V from blood samples is desirable, the volume obtained with bioimpedance spectroscopy is in good agreement with blood measurements and can be used for several months, depending on the stability of the patient. If BIS is not available, VUKM_S can be used, preferably as an averaged value from several sessions.
Conflict of interest statement. P.C. and A.W. are employees of Fresenius Medical Care. The study was supported by Fresenius Medical Care.
|
Appendix A: Watson formulae for estimating urea distribution volume [15] |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix A: Watson formulae...Appendix B: Calculation of...References |
|---|
For males, the urea distribution volume (total body water) is estimated from the subject's age, height and weight:
|
| 1 |
For females, only height and weight are required:
|
| 2 |
|
Appendix B: Calculation of the urea distribution volume using urea kinetic modelling |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix A: Watson formulae...Appendix B: Calculation of...References |
|---|
Two-sample urea kinetic modelling (UKM) provides a prediction of blood urea concentration throughout the weekly cycle based on the pre- and post-dialysis urea levels (C0 and Ct) and weight loss for the first session in the cycle [7]. For a patient in a steady state of urea generation and intradialytic fluid accumulation, the pre-dialysis blood concentration 1 week later will match the measured C0 (see Figure 4). In UKM, an iterative method is used to adjust the urea generation rate (G) and the modelled value of Vpost until the weekly urea concentration profile shows this consistency.
|
To model the weekly cycle, Equation (1) is used to describe session 1 for which pre- and post-dialysis blood samples were taken:
|
| 1 |
where 
is the dialysis time, C0 is the pre-dialysis urea concentration and Ct is urea concentration immediately post-dialysis. Keff is the effective clearance that can be the theoretical clearance calculated from the dialyzer specifications and the treatment parameters, or a measured clearance if one is available. For this study, KOCM was used. Kr is the residual renal clearance, and Qwl is water loss during dialysis.
The post-dialysis urea concentration is calculated from Equation (2). For session 1, Cpre is the measured pre-dialysis urea concentration. For sessions 2 and 3, Cpre is the calculated urea concentration at the end of the preceding interdialytic interval that is determined from Equation (3). The dialysis time, effective clearance and the rate of fluid removal are assumed to be constant for each session in the weekly cycle:
|
| 2 |
|
| 3 |
Cstart_int and Cend_int are the urea concentrations at the start and end of the interdialytic interval. For interval 1, Cstart_int is the measured post-dialysis urea concentration from session 1; for intervals 2 and 3, it is the calculated Cpost for sessions 2 and 3. Tint is the length of the interval and Qwg is the water gain in the interdialytic period.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionAppendix A: Watson formulae...Appendix B: Calculation of...References |
|---|
- Polaschegg HD. Automatic, noninvasive intradialytic clearance measurement. Int J Artif Organs (1993) 16:185–191.[Web of Science][Medline]
- Gotch FA, Panlilio FM, Buyaki RA, et al. Mechanisms determining the ratio of conductivity clearance to urea clearance. Kidney Int (2004) 66(Suppl_89):S3–S24.[Web of Science]
- Kuhlmann U, Goldau R, Samadi N, et al. Accuracy and safety of online clearance monitoring based on conductivity variation. Nephrol Dial Transplant (2001) 16:1053–1058.
[Abstract/Free Full Text] - Daugirdas JT. Second generation logarithmic estimates of single-pool variable volume Kt/V: an analysis of error. J Am Soc Nephrol (1993) 4:1205–1213.[Abstract]
- Daugirdas JT, Greene T, Depner TA, et al. Anthropometrically estimated total body water volumes are larger than modeled urea volume in chronic hemodialysis patients: effects of age, race, and gender. Kidney Int (2003) 64:1108–1119.[CrossRef][Web of Science][Medline]
- Wuepper A, Tattersall J, Kraemer M, et al. Determination of urea distribution volume for Kt/V assessed by conductivity monitoring. Kidney Int (2003) 64:2262–2271.[CrossRef][Web of Science][Medline]
- Moissl UM, Wabel P, Chamney PW, et al. Body fluid volume determination via body composition spectroscopy in health and disease. Physiol Meas (2006) 27:921–933.[CrossRef][Web of Science][Medline]
- Depner TA. Prescribing Hemodialysis: A Guide to Urea Modeling (1991) Boston: Kluwer.
- Di Filippo S, Andrulli D, Manzoni C, et al. On-line assessment of delivered dialysis dose. Kidney Int (1998) 54:263–267.[CrossRef][Web of Science][Medline]
- Dumler F. Best method for estimating urea volume of distribution: comparison of single pool variable volume kinetic modeling measurements with bioimpedance and anthropometric methods. ASAIO J (2004) 50:237–241.[CrossRef][Web of Science][Medline]
- Moret K, Beerenhout CH, Van Den Wall Bake AW, et al. Ionic dialysance and the assessment of Kt/V: the influence of different estimates of V on method agreement. Nephrol Dial Transplant (2007) 22:2276–2282.
[Abstract/Free Full Text] - Kloppenburg WD, Stegeman CA, Hooyschuur M, et al. Assessing dialysis adequacy and dietary intake in the individual hemodialysis patient. Kidney Int (1999) 55:1961–1969.[CrossRef][Web of Science][Medline]
- Di Filippo S, Andrulli A, Mangano S, et al. Reduction in urea distribution volume over time in clinically stable dialysis patients. Kidney Int (2006) 69:754–759.[CrossRef][Web of Science][Medline]
- Daugirdas JT. Simplified equations for monitoring Kt/V, PCRn, eKt/V, and ePCRn. Adv Ren Replace Ther (1995) 2:295–304.[Medline]
- Watson PE, Watson ID, Batt RD. Total body water volumes for adult males and females estimated from simple anthropometric measurements. Am J Clin Nutr (1980) 33:27–39.
[Abstract/Free Full Text]
Accepted in revised form: 16. 7.08
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||