Share this article on:

Effective Ionic Dialysance/Blood Flow Rate Ratio: An Indicator of Access Recirculation in Arteriovenous Fistulae

Mohan, Sumit; Madhrira, Machaiah; Mujtaba, Muhammad; Agarwala, Rajesh; Pogue, Velvie; Cheng, Jen-Tse

doi: 10.1097/MAT.0b013e3181e743eb
Adult Circulatory Support

Effective ionic dialysance (EID) is an online measure of hemodialysis (HD) effective urea clearance that is calculated using changes in dialysate sodium conductivity. Effective ionic dialysance is blood flow (Qb) dependent. The presence of significant (≥5%) access recirculation (sAR) during dialysis lowers EID at a given Qb, thereby lowering EID/Qb. We propose using EID/Qb as a useful chairside tool for detection of sAR in arteriovenous fistulae (AVF). Data were collected from 47 patients with AVF (72% men, mean age 49 ± 11.8 years, duration on dialysis 3.78 ± 3.4 years, duration of fistula use 3.35 ± 3.42 years) dialyzed with an high-efficiency dialyzer with a mass transfer area coefficient (KoA) of 1714 ml/min. Effective ionic dialysance were measured at regular intervals by the Gambro Phoenix dialysis system during treatments. The access recirculation (AR) and access blood flow (Qa) were measured using the reference standard saline dilution technique (Transonic HD-02 monitor). Among the 323 HD sessions where Qb, EID, AR, and Qa were available, we identified 17 instances of sAR. The performance of EID/Qb as indicator of sAR was assessed by a receiver operator characteristic (ROC) curve (Stata version 10.1). The area under the ROC curve was 0.935 (95% confidence interval 0.869–1.000), which demonstrated a sensitivity of 76.5% and specificity of 96.4% at an EID/Qb ≤50% with a positive likelihood ratio of 21, negative likelihood ratio of 0.24, positive predictive value of 54.2%, and negative predictive value of 98.7%. We found similar test performance in patients who received HD with dialyzers with smaller surface areas and lower KoAs. The high specificity of EID/Qb makes it an excellent yet simple and early chairside indicator of AVF recirculation.

From the Division of Nephrology, Department of Medicine, Columbia University College of Physicians and Surgeons, Harlem Hospital Center, New York, New York.

Submitted for consideration April 2010; accepted for publication in revised form May 2010.

Presented at the Meeting of the American Society of Nephrology, Philadelphia, PA; November 2008.

Reprint Requests: Sumit Mohan, MD, MPH, Division of Nephrology, Department of Medicine, Columbia University College of Physicians and Surgeons, Harlem Hospital Center, 506 Lenox Avenue, Room 12101, New York, NY 10037. Email:

The higher extracorporeal blood flow rates used in modern high-efficiency hemodialysis (HD) require a well-functioning vascular access.1 Access dysfunction is associated with significant morbidity and accounts for as many as 25% of all hospitalizations of HD patients.2 The preferred access for incident HD patients is an arteriovenous fistula (AVF), but fistula recirculation during HD decreases the efficiency of the procedure. Recirculation is often a sign of fistula stenosis or a sign of inadequate access flow rates.1 Significant access recirculation (sAR) can occur in the low-pressure system of an AVF without an associated increase in venous pressure making it more difficult to detect early.3,4 An organized monitoring and surveillance approach for assessment of the HD access have been shown to be beneficial and is now recommended in the 2006 Kidney Foundation Disease Outcomes Quality Initiative clinical practice guidelines for vascular access.3,5

In addition to monitoring of access flow, the increased probability of recirculation at high blood flow rates requires a convenient and accessible method for the detection of access recirculation (AR). Monitoring AR is considered an acceptable surveillance technique for stenoses in AVF.3 Several methods have been proposed, often requiring the use of additional equipment, manpower, with specific operator interventions such as infusion of various substances, and carefully timed multiple blood draws, thus limiting their implementation as part of a regular monitoring protocol.6–12 These methods are unlikely to detect inadvertent reversal of lines during dialysis because of the relative infrequency with which these monitoring modalities are performed.13,14 Some of these methods are also limited in their ability to distinguish cardiopulmonary recirculation and AR.9

Recent advances have allowed the measurement of real-time online “effective ionic dialysance” (EID) as a measure of the dose of dialysis delivered.15 Effective ionic dialysance can be used in the place of effective urea clearance because NaCl diffusivity approximates urea diffusivity across the dialyzer membrane and the sodium osmotic distribution volume is equal to the volume of distribution of urea, that is, the total body water.16 Sodium chloride clearance can be calculated from dialysate conductivity measurements.16 In the presence of AR, effective urea clearance decreases and is reflected in a decrease in EID.15 Effective ionic dialysance is also dependent on the blood flow (Qb) and dialysate flow rates during high flux HD. Attempts to use EID to assess access flow have been described but have been of limited practical application.17–21 We propose using EID corrected for the blood flow (=EID/Qb) as an indicator of AR during HD. We demonstrate the utility of using EID/Qb ≤50% in the detection of recirculation among patients with AVF receiving dialysis with high-flux, high- efficiency dialyzers. In addition, we assess the performance of our proposed ratio for detecting recirculation in patients receiving dialysis using a dialyzer with a smaller surface area and 30% lower mass transfer area coefficient (KoA). Obtaining this ratio is potentially simple, noninvasive, and uses widely available standard equipment with no additional cost.

Back to Top | Article Outline

Materials and Methods

Data were obtained from patients undergoing maintenance HD using a single use high-flux cellulose triacetate dialyzer with a surface area of 2.1 m2 (Exeltra 210; Baxter, Deerfield, IL; KoA 1714 ml/min) and 1.9 m2 (Exeltra 190, KoA 1214 ml/min) at the Harlem Hospital Hemodialysis Unit. Effective ionic dialysance was measured by the Diascan module at regular intervals (30 minutes) in the Gambro Phoenix (Lakewood, CO) dialysis machine during each dialysis session and recorded by the nursing staff. We used the compensated blood flow rate (Qb) that is reported by the Phoenix machine and is an estimate of the actual rate of blood flow through the dialyzer determined from the operator-set blood pump speed and the negative arterial pressure.22,23 Saline dilution is the reference standard for vascular access monitoring and is performed routinely on all HD patients using the Transonic HD-02 system at our center. Access monitoring with saline dilution is performed during the first 90 minutes of a dialysis session. We reviewed 323 instances in 47 patients with AVF where the nursing staff recorded the EID and the corresponding Qb in addition to the simultaneously measured access flow and recirculation using the Transonic HD-02 monitor (Figure 1).

We constructed a receiver operator characteristic (ROC) curve for EID/Qb for patients in our study using the roctab function in Stata (version 10.1, Stata Corp, College Station, TX) to assess the diagnostic utility of the EID/Qb ratio in the detection of AR (Figure 2). The area under the ROC curve, sensitivity, specificity, predictive values, and likelihood ratios for each cutpoint were reported by Stata. The Youden index was used to determine the optimal cutpoint on the ROC curve and was calculated as (sensitivity + specificity) − 1 (Figure 3).24,25 The Youden index accounts for both rates of false positives and false negatives while being independent of both the relative and absolute sizes of the control and affected groups.25 A Bland-Altman plot was used to demonstrate the limits of agreement and the error and bias between blood flow rates measured by saline dilution and by the Gambro Phoenix dialysis machine (Figure 4).26 The different measures of blood flow were also compared using linear regression (Figure 5).

Back to Top | Article Outline


A total of 323 instances where EID and AR were assessed and recorded simultaneously were identified in 47 patients with AVF receiving dialysis with an Exeltra 210 dialyzer over a 12-month period. Characteristics of our patient cohort are described in (Table 1). AR ≥5% using the previously validated saline dilution method (Transonic) was considered significant. By using this threshold, we found 17 instances (5.3% prevalence) of sAR (Figure 1) of which 2 (0.6%) were instances of inadvertent line reversal.

Patients with sAR had similar Qb (408 ± 57 vs. 420 ± 33 ml/min, p = ns) but significantly lower Qa (369 ± 215 vs. 1606 ± 933 ml/min, p < 0.0001), EID (117 ± 49 vs. 260 ± 30 ml/min, p < 0.0001), and EID/Qb ratio (43.4 ± 10.6 vs. 62 ± 6.5, p < 0.0001) compared with those patients who did not. We identified 2 instances of inadvertent line reversal, which did not differ significantly from the other 17 instances of recirculation with respect to Qb or EID. Of the remaining 15 instances, we were able to measure the access flow rate (Qa) in 11, all of which had a flow rate <500 ml/min. In these 11 instances of AR, we found a significant negative linear relationship (y = 53.79 − 0.398x, r = 0.59, p = 0.02) between the EID/Qb ratio and the degree of recirculation (Figure 6), and a significant inverse quadratic relationship (y = 109.65 − 0.41x + 0.0004x2, r = 0.98, p < 0.0001) between recirculation and access flow measured by saline dilution (Figure 7).

A ROC curve was used to assess the utility of EID/Qb as a predictor of sAR (≥5%). The area under the curve was 0.935 ± 0.033 (95% confidence interval [CI] 0.869–1.000; Figure 2). The Youden index was calculated for each EID/Qb ratio, and the optimal specificity and sensitivity was found at an EID/Qb ratio of 53.7% with a Youden index of 0.80 (Figure 3). However, we chose to use an EID/Qb ratio of 50% as our cutoff point because this represents a more practical ratio with a marginally higher diagnostic accuracy and specificity (Table 2). The EID/Qb ratio of 50% has a sensitivity of 76.5% and specificity of 96.4% (Youden index of 0.73), a positive likelihood ratio of 21, and a negative likelihood ratio of 0.18, resulting in a positive predictive value of 54.2% and a negative predictive value of 98.7%.

We also identified 45 instances among 8 patients being dialyzed with the smaller dialyzer surface area (1.9 m2), where all the data required had been recorded simultaneously and identified only 2 instances with AR. The ROC curve constructed for the EID/Qb ratio for these patients had an area under the curve of 0.721% ± 0.259% (95% CI 0.213–1.000). Among these patients, the EID/Qb ratio of 50% had a sensitivity of 50% and specificity of 83.7% (Youden index of 0.34), a positive likelihood ratio of 3.1, and a negative likelihood ratio of 0.6, resulting in a positive predictive value of 12.5% and a negative predictive value of 97.3%. The highest Youden index of 0.47 was noted at an EID/Qb ratio of 56.6% with a sensitivity of 100% and specificity of 47.1%.

A solitary instance of AR (among 18 instances available with all the data) in a patient being dialyzed with an even smaller dialyzer (1.7 m2, KoA 1103 ml/min) was noted incidentally. The EID/Qb ratio in this instance was 49%—a level that would have been detected by our recommended threshold of ≤50%.

Linear regression analysis showed excellent correlation between the flow rate measured by the Gambro Phoenix machines' compensated flow rate and the Transonic HD-02 monitor (y = 117.48 + 0.74x, r = 0.79, p < 0.0001; Figure 4). The compensated blood flow overestimated the Transonic blood flow by 14.2 ml/min on average (95% CI 10.96–16.1 ml/min), and the agreement between these two measures of pumped blood flow rate through the dialyzer is shown using a Bland-Altman plot (Figure 4).

Back to Top | Article Outline


Vascular access function and patency are essential for optimal management of HD patients. Vascular access dysfunction is associated with an increase in morbidity and mortality. Arteriovenous fistulae are the preferred choice access for HD patients.3 They have the lowest rate of thrombosis, infection, cost of implantation, and maintenance, as well as morbidity and mortality when compared with other types of permanent vascular access.3 As a result, there is an ongoing effort to increase the prevalence of AVF among HD patients in the United States. In the era of high-flux, high-flow HD, a functioning AVF capable of providing high pump blood flow rate without AR is required to ensure maximal efficiency of HD.

Arteriovenous fistulae develop stenoses over time, which if detected and corrected early, can help to prevent inadequate dialysis and other complications related to access dysfunction.5,17 Arteriovenous fistula recirculation occurs in the presence of hemodynamically significant stenoses or when the pumped blood flow rate through the dialyzer (Qb) exceeds the access flow rate (Qa).3,21,27 However, the monitoring of access flow alone is not an adequate measure of access function as significant recirculation can occur with adequate access flow as a result of inappropriately close positioning of needles, inadvertent line reversal, or cardiopulmonary recirculation and in fistulae with branches and collaterals.13,14,28

The Transonic HD-02's measure of the blood flow rate (QbT) through the dialyzer is considered the reference standard to which we compared the Phoenix's compensated blood flow rate (Qb). Analyses using the EID/QbT ratio were also performed (data not shown). The area under the ROC curve constructed with EID/QbT (data not shown), although marginally lower (0.914, 95% CI 0.830–0.998), did not differ significantly (χ2 = 1.88 p = ns) from the ROC curve constructed using EID/Qb. However, the EID/Qb ratio was preferred as it uses the more readily available measurement of blood flow rate (Qb), resulting in a more accessible and practically applicable ratio.

Dialysis adequacy is adversely affected by recirculation and is reflected in a lower EID. The dialyzer surface area, dialysate flow rate (Qd), and the blood flow rate (Qb) also influence the EID.16 As a result, a low EID is also seen in situations that are associated with low Qb in the absence of access dysfunction such as in patients with a new AVF. The EID/Qb ratio ≤50% identified AR among patients who received dialysis using dialyzers with a range of surface areas and KoAs. Use of the EID/Qb ratio instead of EID alone prevents the inadvertent identification of cases of low EID resulting from a low Qb as cases of AR. This simple ratio can be readily calculated at the chairside and can be used to identify patients with inadvertent line reversal, needle malposition, a smaller dialyzer surface area, or true AVF dysfunction when used early during a dialysis session. As a result, once inadvertent line reversal and needle malposition have been ruled out, EID/Qb ratio ≤50% can be used to identify patients who need further evaluation of their AVF.

Using a higher threshold for the EID/Qb ratio may allow this to be a useful screening tool for AR. For example, an EID/Qb ratio of ≤60% had a sensitivity of 88.2% and specificity of 66%, performance characteristics expected of a good screening test. The EID is influenced by several dialysis prescription factors such as the blood flow rate, dialysate flow rate, dialyzer surface area, and reported KoA. However, the EID/Qb ratio to detect AR seems to be similar even with the use of dialyzers of varying surface area and KoA. Our current analysis, however, does not allow us to assess the impact of lower blood or dialysate flow rates on the utility of this ratio and requires further study.

Back to Top | Article Outline


The EID/Qb ratio is an excellent online chairside test for identifying AR during HD in patients with AVF. A ratio of ≤50% suggests the presence of AR, thereby identifying patients who need further evaluation of their vascular access. The EID/Qb ratio is a potentially simple, noninvasive method that can be used during every dialysis session at no additional cost allowing for easy monitoring of AR.

Back to Top | Article Outline


1. Collins DM, Lambert MB, Middleton JP, et al: Fistula dysfunction: Effect on rapid hemodialysis. Kidney Int 41: 1292–1296, 1992.
2. Feldman HI, Kobrin S, Wasserstein A: Hemodialysis vascular access morbidity. J Am Soc Nephrol 7: 523–535, 1996.
3. Vascular Access Work Group: Clinical practice guidelines for vascular access. Am J Kidney Dis 48: S248–S273, 2006.
4. Daniels ID, Berlyne GM, Barth RH: Blood flow rate and access recirculation in hemodialysis. Int J Artif Organs 15: 470–474, 1992.
5. Tessitore N, Lipari G, Poli A, et al: Can blood flow surveillance and pre-emptive repair of subclinical stenosis prolong the useful life of arteriovenous fistulae? A randomized controlled study. Nephrol Dial Transplant 19: 2325–2333, 2004.
6. Bosc JY, LeBlanc M, Garred LJ, et al: Direct determination of blood recirculation rate in hemodialysis by a conductivity method. ASAIO J 44: 68–73, 1998.
7. Brancaccio D, Tessitore N, Carpani P, et al: Potassium-based dilutional method to measure hemodialysis access recirculation. Int J Artif Organs 24: 606–613, 2001.
8. Magnasco A, Alloatti S, Bonfant G, et al: Glucose infusion test: A new screening test for vascular access recirculation. Kidney Int 57: 2123–2128, 2000.
9. Schneditz D, Wang E, Levin NW: Validation of hemodialysis recirculation and access blood flow measured by thermodilution. Nephrol Dial Transplant 14: 376–383, 1999.
10. Sombolos K, Christidou F, Bamichas G, et al: HCO3 increment in arterial line can reveal significant vascular access recirculation in high-flux hemodialysis: A preliminary report. Hemodial Int 10: 88–93, 2006.
11. Wang E, Schneditz D, Kaufman A, Levin NW: Sensitivity and specificity of the thermodilution technique in detection of access recirculation. Nephron 85: 134–141, 2000.
12. Wang E, Schneditz D, Ronco C, Levin NW: Surveillance of fistula function by frequent recirculation measurements during high efficiency dialysis. ASAIO J 48: 394–397, 2002.
13. Krisper P, Aschauer M, Tiesenhausen K, et al: Access recirculation in a native fistula in spite of a seemingly adequate access flow. Am J Kidney Dis 35: 529–532, 2000.
14. Shapiro W, Gurevich L: Inadvertent reversal of hemodialysis lines—A possible cause of decreased hemodialysis (HD) efficiency. J Am Soc Nephrol 8: 172A–173A, 1997.
15. Lindsay RM, Bene B, Goux N, et al: Relationship between effective ionic dialysance and in vivo urea clearance during hemodialysis. Am J Kidney Dis 38: 565–574, 2001.
16. Gotch FA, Panlilio FM, Buyaki RA, et al: Mechanisms determining the ratio of conductivity clearance to urea clearance. Kidney Int 66: SS3–SS24, 2004.
17. Levy SS, Sherman RA, Nosher JL: Value of clinical screening for detection of asymptomatic hemodialysis vascular access stenoses. Angiology 43: 421–424, 1992.
18. Lindsay RM, Sternby J, Olde B, et al: Hemodialysis blood access flow rates can be estimated accurately from on-line dialysate urea measurements and the knowledge of effective dialyzer urea clearance. Clin J Am Soc Nephrol 1: 960–964, 2006.
19. Lopot F, Nejedly B, Sulkova S, Blaha J: Comparison of different techniques of hemodialysis vascular access flow evaluation. J Vasc Access 5: 25–32, 2004.
20. Mercadal L, Du Montcel ST, Jaudon MC, et al: Ionic dialysance vs urea clearance in the absence of cardiopulmonary recirculation. Nephrol Dial Transplant 17: 106–111, 2002.
21. Mercadal L, Hamani A, Bene B, Petitclerc T: Determination of access blood flow from ionic dialysance: Theory and validation. Kidney Int 56: 1560–1565, 1999.
22. Gambro Phoenix Operator's Manual. 6980197A: 2005.
23. Nilsson L-G, Sternby J: KT measurements using ionic dialysance are independent of monitor. Am J Kidney Dis 52: 1027–1028, 2008.
24. Ruopp MD, Perkins NJ, Whitcomb BW, Schisterman EF: Youden index and optimal cut-point estimated from observations affected by a lower limit of detection. Biom J 50: 419–430, 2008.
25. Youden WJ: Index for rating diagnostic tests. Cancer 3: 32–35, 1950.
26. Altman D, Bland J: Measurement in medicine: The analysis of method comparison studies. Statistician 32: 307–317, 1983.
27. Besarab A, Lubkowski T, Frinak S, et al: Detection of access strictures and outlet stenoses in vascular accesses. Which test is best? ASAIO J 43: M543–M547, 1997.
28. Schneditz D: Recirculation, a seemingly simple concept. Nephrol Dial Transplant 13: 2191–2193, 1998.
Copyright © 2010 by the American Society for Artificial Internal Organs