Secondary Logo

Journal Logo

Dialysis & Kinetics

Comparison Between Two On-Line Reversed Line Position Hemodialysis Vascular Access Flow Measurement Techniques: Saline Dilution and Thermodilution

Wijnen, Edwin*; Essers, Stig*; van Meijel, Ger*; Kooman, Jeroen P.*; Tordoir, Jan*†; Leunissen, Karel M.L.*; van der Sande, Frank M.*

Author Information
doi: 10.1097/01.mat.0000227680.67901.01
  • Free


Numerous studies have pointed out that periodical access flow measurements can predict the development and presence of vascular access flow-limiting stenosis and subsequent thrombosis. Preemptive intervention (either radiological or surgical) extends the duration of access sites and may even reduce health care costs,1–5 although, because of recent negative trials, this issue has become more controversial.6

The Kidney Disease Outcome Quality Initiative (K/DOQI) Clinical Practice Guidelines for Vascular Access strongly recommend surveillance of vascular access monitoring by periodical flow measurements.7 The most commonly used technique to measure access flow is the saline dilution technique, as introduced by Krivitski.8 Recently, it has become possible to perform access flow measurements during dialysis treatment with devices that can be integrated into the dialysis machine itself. The Blood Temperature Monitor (BTM, Fresenius Medical Care, Bad Homburg, Germany), integrated into the 4008H dialysis machine (Fresenius Medical Care, Bad Homburg, Germany), measures access flow on the basis of temperature dilution.

A great practical advantage of the BTM compared with the HD01 (Transonic Systems Inc., Ithaca, NY) is the integration with the dialysis machine, which offers the opportunity to measure numerous patients simultaneously. However, except for two studies, the agreement between thermodilution and saline dilution and variability of thermodilution has not been widely studied in a larger population.

The aim of the present study was, therefore, first, to analyze the agreement between the results of both measurement techniques, and, second, to determine the reproducibility of each separate measurement technique.



Measurements were performed during the first hour of dialysis treatment in 40 patients with end-stage renal disease. Forty vascular accesses (16 forearm fistula, 13 forearm grafts, 10 upper arm fistula, 1 upper arm graft) were studied. For each study, a series of measurements was performed: (1) access recirculation (normal blood line position) and (2) access flow (reversed blood line position) were measured by using the saline dilution technique; (3) access flow was measured by using the BTM. Access flow measurement using the BTM consists of two recirculation measurements, one with correct placement of bloodlines and one with reversed bloodlines. With the use of an access flow measurement protocol supplied by the manufacturer, a second recirculation measurement with correct placement of bloodlines was performed. A spreadsheet document was used to calculate access flow with the help of an embedded formula (Equation 3). The spreadsheet data presented two access flow values that were averaged. Before each access flow measurement, blood pressure was measured. With each HD01 and BTM measurement, blood pump speed was set at 300 ml/min.

Twenty of the above described accesses were studied twice, at weekly intervals.

Analysis Techniques

Both techniques to measure access flow are based on reversed line position. The purpose of reversed line position is to enable delivery of the indicator into the venous dialyzer line upstream of the access and then to be able to sample downstream of the access (after the indicator has mixed) in the arterial line.

Measuring access flow with the use of the Transonic HD01, a bolus of isotonic saline (indicator) is administrated in the venous bubble trap after line reversal. The administration time has to be less than 6 seconds to prevent cardiopulmonary recirculation. Two ultrasound dilution sensors are clamped onto the bloodlines, one on the arterial and one on the venous. The venous saline dilution sensor will first sense the diluted blood (ultrasound velocity blood: 1560 to 1590 m/s, isotonic saline 0.9%: 1533 m/s), which is a reference value to calculate the actual recirculation, which is what is left of the dilution in the arterial line after passing the access (Figure 1).

Figure 1.
Figure 1.:
HD01, reversed lines access flow measurement setup.

Besides sensing dilution, the saline dilution sensors simultaneously measure blood flow in the bloodlines. The obtained recirculation fraction (R) and measured extracorporeal blood flow (Qb) provide the possibility of calculating access flow (Qa), according to Equation 1.

Recirculation measurement with the saline dilution technique is performed with normal extracorporeal line position. The administered bolus of isotonic saline in the venous bubble trap will disappear upstream of the access. However, when vascular access recirculation appears, a fraction of the administered saline will be sensed in the arterial line.

Instead of fluid dilution the BTM uses temperature as indicator through dialysate temperature to heat up or cool down the returning blood to the patient. Temperature is registered by two temperature sensors, which are placed around the venous and arterial bloodlines, both located 1 meter from the access. When the blood temperature downstream of the artificial kidney drops or rises, the change in temperature will be registered by the sensor placed around the venous bloodline. This same temperature change will affect the temperature of the central blood volume when it enters the patient's bloodstream. Through cardiopulmonary recirculation, the extracorporeal induced temperature change will be sensed by the temperature sensor clamped around the arterial bloodline (Figure 2). At the end of the measurement, the difference between the venous and arterial bloodline temperatures is displayed on the BTM monitor as a relative value and stands for the temperature recirculation (RBTM,n), caused by cardiopulmonary recirculation. The relative recirculation value obtained by executing this measurement with reversed bloodlines (RBTM,x) includes both the recirculation over the access and the cardiopulmonary recirculation, because of measurement time, which takes approximately 5 minutes and is technique-dependent. To separate the cardiopulmonary-induced temperature recirculation from the recirculation fraction over the access, the ratio of access flow to cardiac output must be defined by using the “double recirculation technique”9 and stands for the actual cardiopulmonary recirculation (CPR):

Figure 2.
Figure 2.:
BTM cardiopulmonary recirculation (CPR) measurement setup. Qa = access flow, Qb = extracorporeal blood flow, Qdia = dialysate flow, Tv = temperature in the venous bloodline, Ta = temperature in the arterial bloodline, TQdia = temperature of the dialysate, BTM,A = arterial temperature sensor head, BTM,V = venous temperature sensor head. This figure shows the measurement of the CPR measurement. The dialysate heater initiates a drop in temperature (TQdia). Following the red arrows, this temperature drop has an effect on the temperature in the venous line (Tv). Through CPR, the extracorporeal-induced temperature drop is reflected at the BTM,A (Ta). Finally, the relation of the temperature drops measured by arterial and venous sensor heads is equivalent to the percentage recirculation of blood flow. Executing this measurement with reversed bloodlines includes both the recirculation over the access (comparable with the saline dilution technique) and the CPR as the result of long measurement time.

where Qb,n is the extracorporeal blood flow with normal line position and Qb,x is the extra corporeal blood flow with reversed line position. Qb is displayed on the 4008H dialysis machine main display and represents the operator's setup extracorporeal pump speed corrected for the arterial pressure. Afterward, Qb is also corrected for the ultrafiltration rate (Ufrate/60). Eventually, access flow is calculated by using Equation 3, which is based on Equation 1, embedded with the above described corrections.

Dialysis Strategy

Thirty-one patients were treated with bicarbonate hemodialysis with low flux polysulfone membranes (F8HPS; Fresenius; Bad Homburg; Germany). Nine patients were treated with on-line postdilution hemodiafiltration (HDF) with high-flux membranes (APS-1050, Asahi Medical Co., Tokyo, Japan). Postdilution HDF probably has no effect on BTM measurement because the substitution fluid is from the same source as the dialysate. Changing the dialysate temperature also affects the substitution fluid temperature in the same way. With the use of the Transonic HD01, infusion of substitution fluid was temporarily stopped because the infusion fluid during HDF was infused in postdilution mode. The venous ultrasound signal is disturbed as the result of pulsatile fluid infusion, which yields errors in the measurement. Sodium concentration of the dialysate was 138 or 140 mmol/l; calcium concentration was 1.5 mmol/l; and temperature of the dialysate was 36 or 36.5 °C. Ultrapure dialysate was used, achieved by double reverse osmosis, electric deionization, ozone sanitization, and filtration through Diasafe® (Fresenius; Bad Homburg; Germany).

Statistical Analysis

Statistical analysis was performed with the use of SPSS 12.0 software. The relation between Qa (HD01) and Qa (BTM) was studied by using regression analysis and the Bland Altman plot. The reproducibility of both techniques was assessed by analyzing the relative difference of weekly subsequent measurements: Δ xrel = √([x2/x1–1]2). Data are given as mean ± SD values. A value of p < 0.05 is considered significant.


Agreement Between Both Techniques

Average access flow measured with the saline dilution technique and the thermodilution technique was 1053 (±495) ml/min and 1034 (±527) ml/min, respectively, (p = 0, ns) (n = 40). Correlation between access flow measurements by both techniques expressed in r2 was 0.79 (r = 0.89) (Figure 3 and Figure 4).

Figure 3.
Figure 3.:
Scatterplot. Qa results, saline dilution (HD01)/thermodilution (BTM).
Figure 4.
Figure 4.:
Bland Altman Qa results, saline dilution (HD01) vs thermodilution (BTM).

Reproducibility Results of Each Separate Technique

Reproducibility of saline and thermodilution subsequent measurements with a weekly interval, expressed in relative difference, (Δ xrel) was 13 (±11)% and 24 (±14)%, respectively (p < 0.01) (n = 20). Reproducibility results of saline and thermodilution subsequent measurements are also displayed with the help of a Bland Altman analysis in Figure 5 and Figure 6, respectively. Saline dilution technique correlation, expressed in r2 of subsequent measurements, was 0.82 (r = 0.91). The thermodilution technique r2 of subsequent measurements was 0.79 (r = 0.89) (Figure 7).

Figure 5.
Figure 5.:
Bland Altman HD01 Qa measurement 1 vs HD01 Qa measurement 2.
Figure 6.
Figure 6.:
Bland Altman BTM Qa measurement 1 vs BTM Qa measurement 2.
Figure 7.
Figure 7.:
Scatterplot. Repeated BTM Qa results at weekly intervals.

Access Recirculation

The measured access recirculation fraction was 0% in all separate recirculation measurements done with the saline dilution method (n = 40). Access recirculation obtained with the thermodilution technique had to be calculated with help of Equation 4, which is described elsewhere.9

Thermodilution mean access recirculation result was 0.74 (±0.46)% (n = 40).

Additional Results

The mean relative difference between Qb values measured with both techniques was 5.1 (±3)% (n = 40). The mean relative difference of BTM recirculation measurements with normal line position (RBTM,n), measured before and after each reversed line recirculation measurement, was 27 (±25)% (n = 40). Average absolute values for the first and second BTM recirculation measurements with normal line position were 8.2 (±2.6) and 7.0 (±2.7) (n = 40), respectively (p < 0.001). The intertreatment reproducibility for recirculation with normal blood line position was 30.3 (±21.4)% (n = 20). Average absolute values for these measurements were 8.4 (±2.3) and 7.0 (±2.8) (n = 20), respectively (p = 0.065). Average measurement time to calculate Qa with the HD01 took 4 (±0.5) minutes. Average measurement time to calculate Qa with the BTM, based on only one RBTM,n measurement, took 29 (±7) minutes.

The mean arterial pressure (MAP) values at which the reproducibility testings were performed for saline dilution Qa measurement were 88.8 (±18.0) mm Hg (week 1) and 90.2 (±17.7) mm Hg (week 2), respectively (p = ns) (n = 20). The MAP values at which the reproducibility testings were performed for temperature dilution access flow measurements were 84.4 (±15.2) mm Hg (week 1) and 86.1 (±17.0) mm Hg (week 2), respectively, (p = ns) (n = 20).


In this study, we showed a good correlation but wide limits between the thermo dilution technique and the saline dilution technique. However, reproducibility was less with the thermodilution technique compared with the saline dilution technique. To the best of our knowledge, only two papers studied correlation between both techniques: The Schneditz et al.10 validation report of the thermodilution technique showed a similar correlation (r2 = 0.84, n = 52 in 17 patients) compared with this study (r2 = 0.79, n = 40 in 40 patients). Lopot et al.11 found a higher correlation between both techniques, expressed as r = 0.9543 (r2 = 0.91, n = 54, number of patients unknown).

Reproducibility of the separate techniques is also described in the Lopot study: a correlation coefficient of 0.9197 (r2 = 0.85) and 0.9702 (r2 = 0.94) in 40 subsequent thermodilution measurements and in 58 subsequent saline dilution measurements respectively, compared with this study, r = 0.89 (r2 = 0.79, n = 20) and r = 0.91 (r2 = 0.82, n = 20).

Ragg et al.12 studied within-treatment reproducibility of the thermodilution technique (n = 189 in 56 patients). Correlation between repeated Qa measurements was rather low (r = 0.68). They concluded that reproducibility lessened considerably with increasing magnitude of flow rate (≥600 ml/min). In our study, the variation between subsequent thermodilution measurements with a weekly interval appeared to be present in every flow range (Figure 6). Possibly, the long thermodilution measurement time, compared with the saline dilution method, may play a role in reduced reproducibility, considering that variable hemodynamic conditions are present during dialysis treatment.13,14 To prevent variable hemodynamic conditions in this study, reproducibility measurements were executed in two separate dialysis sessions at the very same time (start of measurement series 5 to 10 minutes after start of dialysis treatment). The MAP values at which the reproducibility testings were performed were identical between week 1 and week 2 (88.8 [±18.0] mm Hg versus 90.2 [±17.7] mm Hg for saline dilution measurements and 84.4 [±15.2] mm Hg versus 86.1 [±17.0] mm Hg for BTM measurements). Thus, differences in MAP do not appear to have played a role in apparent differences in reproducibility of the measurements. However, the small difference in MAP values between saline dilution and temperature dilution measurements might have played some role in the lesser agreement between both methods.

Of course, the differences in measurement techniques might also influence the separate reproducibility results. The two most important differences can be illustrated by using Equation 1. To calculate Qa, two independent values must be measured: Qb and the reversed lines access recirculation (Rx,access). Qb is measured by using the HD01, whereas the BTM technique must estimate true Qb as described before. However, the described mean relative difference between both Qb #x00B4; s (5.1 [±3]%), is not significant.

Rx,access is also directly measured by using the HD01 and must be calculated when using the BTM technique, which needs two separate recirculation measurements (normal and reversed lines) to correct for the cardiopulmonary recirculation. The frailty of this calculated approximation of Rx,access (BTM) can be illustrated, showing the mean relative difference of subsequent RBTM,n results, measured before and after each reversed line recirculation measurement, and the intertreatment subsequent RBTM,n results, which was 27 (±25)% and 30.3 (±21.4)%, respectively. Lopot et al.11 described that when CPR is neglected, it will significantly underestimate actual Qa values up to 30%.

Access Recirculation

Access recirculation measured with the saline dilution technique was 0% in all patients. Mean access recirculation obtained with the thermodilution technique was practically comparable: 0.74 (±0.46)% (n = 40). Wang et al.15 described that a BTM recirculation with normal blood lines above a threshold of 15% has been found as highly sensitive and specific for access recirculation. In this study, only one patient had a recirculation value larger than 15% (15.2%). In this patient, access flow was normal: 1150 ml/min for the saline dilution method and 565 ml/min for the temperature dilution method, and no access recirculation was found by the saline and the temperature dilution technique.

Implications for Clinical Practice

The reproducibility of the saline dilution and thermodilution techniques mentioned in both previously described studies11,12 are not expressed in relative difference (Δ xrel).

The average relative difference of the thermodilution technique (24 [±14]%) described in our study almost identifies with the call to intervene, knowing the 25% access flow decline as advised by K/DOQI Clinical Practical Guidelines for Vascular Access7 and the European Guidelines.16 This could lead to unnecessary intervention, based on false measurement, or missing the severe access flow decline (Figure 5).

Of course, we should not forget the relative difference of the saline dilution technique (Δ xrel = 13 [±11]%), which, however, less compared with the relative difference of the thermodilution technique, is still significant. With the saline dilution technique, it is possible though, to measure Qa three times in a row within a short moment of time (5 to 7 minutes), hereby minimizing the effect of variable hemodynamics. Averaging these three results will partly reduce the measurement variation of the saline dilution technique. To measure Qa once with the BTM technique, using the described protocol, took an average time of 29 (±7) minutes.


Considering that the BTM was initially designed to obtain cardiovascular stability during dialysis, the thermodilution measurements correlate well with the saline dilution measurements. However, access flow is measured to prevent vascular access failure and, of course, accuracy is an issue here. The better reproducibility results of the saline dilution technique and the shorter measurement time should be considered when access flow measurement plays a vital role within the local vascular surveillance program.


The authors thank Piet Claessens and Paul Bocken for technical support.


1.McCarley P, Wingard RL, Shyr Y, et al: Vascular access blood flow monitoring reduces access morbidity and costs. Kidney Int 60: 1164–1172, 2001.
2.Schwab S, Oliver M, Suhocki P, McCann R: Hemodialysis arteriovenous access: detection of stenosis and response to treatment by vascular access blood flow. Kidney Int 59: 358–362, 2001.
3.Valj K: Prophylactic angioplasty: Is it worthwhile? In: Gray RJ, Sands JJ (eds). Dialysis Access: A Multidisciplinary Approach. Philadelphia, Lippincott Williams & Wilkins, 2002, pp153–156.
4.Sands J, Jabyac P, Miranda C, Kapsick B: Intervention based on monthly monitoring decreases hemodialysis access thrombosis. ASAIO J 45: 147–150, 1999.
5.Sands J: Vascular access monitoring improves outcomes. Blood Purification 23: 45–49, 2005.
6.Paulson WD: Access monitoring does not really improve outcomes. Blood Purification 23: 50–56, 2005.
7.NKF-K/DOQI: Clinical practice guidelines for vascular access: Update 2000. Am J Kidney Dis 37: S137–S181, 2001.
8.Krivitski NM: Theory and validation of access flow measurement by dilution technique during haemodialysis. Kidney Int 48: 244–250, 1995.
9.Schneditz D, Fan Z, Kaufman AM, Levin NW: Measurement of access flow during dialysis using the constant infusion approach. J Am Soc Artif Intern Org 44: 74–81, 1998.
10.Schneditz D, Wang E, Levin W: Validation of haemodialysis recirculation and access blood flow measured by thermodilution. Nephrol Dial Transplant 14: 376–383, 1999.
11.Lopot F, Nejedly B, Sulkova S, Blaha J: Comparison of different techniques of hemodialysis vascular access flow evaluation. Int J Artif Organs 26: 1056–1063, 2003.
12.Ragg JL, Treacy JP, Snelling P, et al: Confidence limits of arteriovenous fistula flow rate measured by the ‘on-line' thermodilution technique. Nephrol Dial Transplant 18: 955–960, 2003.
13.Besarab A, Lubkowski T, Vu A, et al: Effects of systemic hemodynamics on flow within vascular accesses used for hemodialysis. ASAIO J 47: 501–506, 2001.
14.Rehman SU, Pupim LB, Shyr Y, et al: Intradialytic serial vascular access flow measurements. Am J Kidney Dis 34: 471–477, 1999.
15.Wang E, Shcneditz D, Kaufman Am, Levin NW: Sensitivity and specificity of the thermodilution technique in detection of access recirculation. Nephron 85: 134–141, 2000.
16.Bakran A, Volker M, Passlick-Deetjen J: Management of the renal patient: clinical algorithms on vascular access for haemodialysis. Lengerich; Berlin; Bremen; Miami; Riga; Viernheim; Vienna; Zagreb: Pabst Science Publishers, 2003.
Copyright © 2006 by the American Society for Artificial Internal Organs