|
|
| Journal of Vascular Access 2004; 5: 57 - 61 |
| The role of surveillance in mature arteriovenous fistula management |
N. Tessitore1, V. Bedogna1, A. Poli2
1Nephrology Division, Azienda Ospedaliera of Verona - Italy
2Department of Medicine and Public Health, University of Verona - Italy
|
|
|
|
ABSTRACT
The existing guidelines recommend arteriovenous fistulae (AVF) surveillance by access blood flow (Qa) measurement and the correction of hemodynamically significant stenoses to prolong access survival. Unfortunately, many studies supporting these recommendations are inadequate methodologically; therefore, both the optimal criteria for surveillance and the value of preventive stenosis repair in AVF remain controversial. Recent literature confirms that Qa measurement allows an accurate identification of both stenosis (area under the curve (AUC) ranging from 0.80-0.93) and access at risk of failure (AUC ranging from 0.82-0.98) in AVFs and suggests a Qa <700-1000 ml/min and/or a reduction in Qa >25% as optimal predictors for stenosis and a Qa <400 ml/min for incipient thrombosis. Recent prospective studies evaluated whether Qa surveillance could improve AVF patency rates compared to monitoring based on clinical and dialysis-related criteria alone. The majority of studies have historical, rather than concurrent, control groups and provide conflicting results, some showing a reduction and some showing no change in thrombosis rates by Qa monitoring. On the other hand, the few randomized controlled studies available show that Qa surveillance, when coupled with preemptive intervention, reduces the already low thrombosis rate in AVF and suggest that the functional access life can be prolonged. However, there is still the need for additional methodologically adequate studies to understand fully the role of surveillance in AVF management. (The Journal of Vascular Access 2004; 5: 57-61)
Key Words. Arteriovenous fistula, Surveillance, Access blood flow, Stenosis, Thrombosis
The native arteriovenous fistula (AVF) is regarded as the vascular access (VA) of choice for hemodialysis (HD) because of its superior patency and lower complication rates, once it has fully matured (1).
Mature AVFs are nevertheless prone to thrombosis, almost invariably due to the development of stenosis, although their thrombosis rates are much lower than that of grafts (1), ranging from 4.0-18.0% per year in the absence of any specific surveillance program other than bedside clinical monitoring, often associated with periodic measurement of delivered dialysis dose (2-9).
Despite the relatively low “spontaneous” AVF thrombosis rate, existing guidelines advocate surveillance programs for early stenosis detection and its correction before thrombosis onset to reduce morbidity and prolong access patency (1, 10).
The K/DOQI Guidelines-Update 2000 recommend that AVFs should be monitored for stenosis as outlined for grafts, preferably by monthly direct access blood flow (Qa) measurements and the guidelines indicate that an access with a Qa <600 ml/min or Qa <1000 ml/min that has declined by >25% over 4 months should be referred for a fistulogram. In addition, it is recommended that angiography should be performed in AVFs with elevated recirculation. The guidelines advocate the revision of hemodynamically significant stenoses (if >50% and a warning of inadequate dialysis dose delivery and impending thrombosis) to prolong the fistula life, while they do not support the elective repair of stenosis in well functioning VAs, since studies concerning efficacy have not been performed.
However, the K/DOQI approach has been questioned and considered premature due to the scarcity of methodologically adequate studies supporting its recommendations. Studies evaluating the diagnostic accuracy of Qa measurement to detect stenosis and thrombosis risk in AVF have small sample sizes, often report cumulative data for both AVF and grafts, and fail to provide consistent data indicating threshold Qa values ranging from <800 ml/min to <300 ml/min or a reduction in Qa over time >20% (5, 8). In addition, the majority of studies supporting preemptive stenosis repair are retrospective, or prospective with historical, rather than concurrent, control groups (1).
The recent Vascular Access Society (VAS) guidelines (10) take these considerations into account and, while confirming that Qa measurement is the best method to monitor VA function and predict access failure, state that angiography should be performed when Qa falls below certain cut-off values, but it remains unknown which cut-off value should be used. Therefore, it is recommended that in the absence of an established method for surveillance and prevention of VA failure, each unit should establish its own quality insurance program to monitor prospectively the VA by one or more monthly screening tests (ranging from physical examination to Qa and recirculation measurement, dialysis arterial and venous pressure monitoring and duplex ultrasound (DU) evaluation). The VAS guidelines also support preemptive stenosis correction and confirm that it should be treated if >50% and associated with abnormal physical findings, reduction in dialysis dose and access blood flow. However, they comment that the benefit of surveillance is not confirmed and conclude that there is no evidence that AVF surveillance is cost-effective.
Due to these uncertainties concerning the optimal criteria for the identification of AVF with stenosis and at risk of thrombosis, and the value of both surveillance and preemptive stenosis repair in AVF, several recent studies have addressed these unresolved issues.
Criteria for detecting stenosis and AVF at risk of thrombosis
The majority of studies have focused on the identification of optimal Qa thresholds for detecting AVF at risk of incipient failure.
In a prospective study including a small group of forearm AVFs followed-up with quarterly ultrasound dilution (UD)-based Qa measurements, we proposed a Qa <350 ml/min as an accurate threshold for predicting incipient thrombosis (within 7 months of Qa measurement) for its 100% sensitivity (SE) and 9.5% false positive rate (FPR) (5). A decline in Qa over time proved to be a less accurate predictor than absolute Qa measurement, while the presence of elevated urea-based access recirculation (R>5%) showed an acceptable predictive power for incipient thrombosis, although it was inferior to Qa because of a lower SE of 67% and comparable FPR. Predictive accuracy of Qa measurement was independent of the access location, being identical in wrist and mid-forearm AVFs.
In another observational study including a small group of forearm AVFs followed-up for a 4-yr period with yearly UD Qa measurements and no preemptive interventions, Basile et al (9) identified a Qa <700 ml/min as an accurate predictor of access thrombosis (89% SE and 69% specificity (SP)) and suggested this threshold as a reliable cut-off point at which to start close monitoring of the access. The thrombosis risk was much smaller when Qa was >700 ml/min, as shown by a cumulative 4-yr survival rate of 74%, compared to a 21% survival rate when Qa was below this threshold. However, it should be noted that as many as 20% of their patent AVFs had Qa levels consistently <500 ml/min, confirming that fistulae with low Qa can remain patent for a long time (1).
In the largest available study including forearm and upperarm AVFs followed-up with bimonthly UD Qa measurements, Tonelli et al (11) identified a Qa <500 ml/min as the most appropriate threshold for capturing clinically relevant lesions, defined as angiographically proven or clinically evident stenosis or access failure (i.e. the inability to deliver sufficient blood flow to permit HD treatment) within 6 months of the Qa measurement. In this study, the Qa <500 ml/min threshold showed a positive predictive value of 84% and the highest efficiency value (i.e. the percentage of agreement between test results and diagnosis) as the result of the combination of 70% SE and 93% SP. This cut-off value was preferred over both a Qa <600 ml/min and a Qa <400 ml/min, because the former was associated with a modest improvement in SE, but a large increase in FPR (and, therefore, of unnecessary angiograms) and the latter with a marked reduction in SE to 51%. In this study, predictive accuracy was also independent of access location, being similar in forearm or upperarm AVFs. However, the results of this study, should be considered with caution because verification bias (i.e. the lack of independence between the test under investigation (Qa) and the gold standard because many of the “failed” AVFs were identified by a Qa <500 ml/min) could have led to an overestimation of predictive accuracy. In addition, the choice of the composite outcome of stenosis and failure (which had to be used because angiography was not performed in all AVFs) was an additional drawback of the study, making the proposed Qa <500 ml/min threshold of uncertain utility in clinical practice, since it is known that stenotic AVFs do not necessarily thrombose (1). The investigators themselves acknowledged that access failure within 6 months was more likely when Qa was <400 ml/min, rather than when Qa was <500 ml/min.
In analogy with the previous study, Hoeben et al (12) identified UD Qa <500 ml/min as the most accurate predictor of a “vascular access impairment episode” (defined as angiographically proven >50% stenosis or thrombosis) in a small group of AVFs followed-up with bimonthly Qa measurements. The SE and SP of this threshold were, respectively, 57% and 92%. When Qa surveillance was combined with dynamic venous pressure monitoring SE improved up to 73%, although this diagnostic improvement was unassociated with any reduction in the thrombosis rate. However, the composite outcome measure of the study makes it uncertain whether the proposed threshold is the most accurate in predicting AVF failure.
Taking the results of these studies together, a UD Qa <400 ml/min appears to be the most reliable threshold for detecting AVF at risk of incipient thrombosis. This conclusion is also supported by the results of an earlier, larger study that identified a DU Qa <400 ml/min as the best predictor of AVF failure (13). These findings indicate that prompt intervention is warranted in AVF when Qa drops below the critical threshold of <400 ml/min. Higher Qa thresholds (ranging from a Qa <500 ml/min to a Qa <700 ml/min) could be more appropriate for angiography referral or for closer access monitoring.
Few studies have addressed the issues of the ability of Qa measurements to detect stenosis per se in AVFs.
In a large unselected group of forearm AVFs, we demonstrated that UD Qa measurement had high discriminative ability for stenosis (area under the curve (AUC) of 0.93) and identified a Qa <750 ml/min as the most accurate absolute threshold to detect significant (>50%) angiographically proven stenoses, with 84% SE and 8% FPR (5). A comparable diagnostic efficiency could be obtained by the either/or combination of a Qa <750 ml/min and a decline in Qa over time >25%, which could be preferable to the absolute Qa value of <750 ml/min because of its high 91% SE coupled with an acceptable 14% FPR. In this study, optimal Qa thresholds for stenosis appeared to be different depending on the anastomosis location, being a Qa <750 ml/min in wrist (90% SE and 13% FPR) and a Qa <1000 ml/min in mid-forearm AVFs (89% SE and 6% FPR). This preliminary observation requires confirmation, but it could be of interest, since the proportion of mid-forearm, elbow or upperarm AVFs is steadily increasing due to the high prevalence of HD patients with poor distal forearm vasculature.
Surrogate Qa markers, such as recirculation and monitoring dialysis blood pump flow (Qb), showed a significantly lower diagnostic accuracy than direct Qa measurement (AUCs of, respectively, 0.724 and 0.636), making these monitoring tools of uncertain utility in stenosis surveillance.
In a smaller group of forearm AVFs, Schwarz et al (14) compared the diagnostic accuracy for angiographically proven stenosis of the two most frequently used ultrasound-based Qa measurements, the UD and the DU techniques and showed that both performed equally well (AUCs of, respectively, 0.79 and 0.80). These investigators proposed a UD Qa of 465 ml/min and a DU Qa of 390 ml/min as the optimal thresholds for stenosis, because of their high SP (respectively, 89% and 79%) coupled with an acceptable SE of, respectively, 68% and 76%. However, re-analysis of their data showed that higher Qa thresholds performed equally well and provided a significant increase in SE (e.g. SE 86% for UD Qa <700 ml/min and 90% for DU Qa <500 ml/min), but unacceptably high FPR (respectively, 50% and 39%).
Although both studies demonstrated that ultrasound-based Qa measurements allow an accurate identification of stenosis in AVF (but often at the expense of unnecessary angiograms), the significance in identifying clinically silent stenoses is unknown, since randomized controlled studies evaluating the value of correcting such stenoses are unavailable (1).
The results of a recent prospective, controlled, quasi-randomized study, showing that prophylactic PTA of subclinical stenosis in well functioning forearm AVFs (i.e. able to provide an spKt/V>1.2) was associated with a significant 2.87-fold (95% confidence interval (95% CI) 1.21-6.80) reduction in the risk of failure and comparable reduction in access-related morbidity, indicate that stenosis correction before the onset of significant access dysfunction is beneficial in AVF (15) and strongly support the implementation of surveillance programs for early stenosis detection.
Qa surveillance and AVF patency rates
Several prospective studies have addressed the issue of whether or not Qa surveillance can improve AVF patency rates, but, unfortunately, they provide conflicting results and many are biased by the use of historical, rather than concurrent, control groups.
Two prospective studies with historical controls (2, 3) evaluated whether the implementation of the K/DOQI guidelines would affect AVF patency rates in comparison with dialysis dynamic venous pressure monitoring. Both studies showed that UD Qa surveillance approximately halved the thrombosis rate (from 16% (2) and 15% (3) to 7% per patient-year), although the drop in these rates was not always significant (3).
Another uncontrolled study by Lok et al (16) compared two different surveillance strategies and showed that thrombosis rates were much the same in AVF (10% and 12% per patient-year) with clinical monitoring coupled with monthly recirculation measurement and DU evaluation every 6 months as with bimonthly UD Qa surveillance (indications for fistulography in this study were a Qa <500 ml/min or a decline in Qa >15%).
On the contrary, Branger et al (8) documented that in AVF the implementation of monthly UD Qa surveillance (indication for DU evaluation of the access was a decline in Qa >25%) was associated with a 2-fold reduction (95% CI 1.02-3.78, p<0.05) in thrombosis rates (from 18-8.5% per patient-year) when compared to an historical period in which fistulography was performed based on clinical criteria alone.
A similar reduction in the thrombosis rate was obtained in a randomized controlled study by Sands et al (17), who showed that monthly UD Qa surveillance (indication for fistulography was a Qa <750 ml/min) reduced the thrombosis rate from 26-16% per patient-year compared with DU monitoring every 6 months. The results of this study should be considered with caution because the follow-up period was too short and the thrombosis rates in both subgroups were among the highest ever reported in AVFs.
We recently completed a randomized controlled study to evaluate whether Qa surveillance (indications for fistulography UD Qa <750 ml/min or decline in Qa >25% or recirculation >5%) coupled with preemptive stenosis correction (by angioplasty and open surgery) in well functioning forearm AVFs could reduce thrombosis rates and prolong the useful life of the access compared to the more traditional approach of waiting until dysfunction is clinically evident (as indicated by a reduction in delivered dialysis dose) before stenosis repair (18).
In our experience, preemptive subclinical stenoses correction was associated with a 3.35-fold (95% CI 1.44-7.78, p=0.003) reduction in the thrombosis rate and in a 2.67-fold (95% CI 0.98-6.85, p=0.055) increase in the useful life of the access. Subgroup analysis showed that preemptive intervention in stenotic AVFs with a Qa >350 ml/min was associated with the lowest thrombosis rate (3.7% per patient-year), which was significantly lower than the thrombosis rates observed in their matched controls and preemptively treated AVFs with a Qa <=350 ml/min (respectively, 13.9% and 17.8%). Preemptive intervention at a Qa >350 ml/min was also associated with the lowest AVF loss rate (2.4% per patient-year), which was significantly lower than the 13.2% loss rate in preemptively treated AVFs with a Qa <=350 ml/min (p=0.050), but similar to the 9.7% loss rate in control AVFs with a Qa >350 ml/min. These data strongly suggest that a superior outcome (i.e. reduction in the thrombosis rate and the prolongation in the useful life of the access) can be obtained when intervention is performed in AVFs with lesser hemodynamic impairment, although additional trials with larger sample sizes are needed to confirm these findings. In addition, our study suggested that late preemptive stenosis correction, i.e. in AVFs with a Qa <350 ml/min or with recirculation, had much the same access survival rates as stenosis correction after thrombosis, a finding consistent with that of a large retrospective study on forearm AVFs (19).
In conclusion, despite some study design drawbacks, the current data indicate that Qa surveillance allows an early and accurate identification of stenosis and access at risk of incipient thrombosis in mature AVFs and, when coupled with preemptive intervention, reduces the thrombosis rate and could prolong the functional life of the access.
However, there are many unresolved issues (Tab. I), and additional, methodologically adequate studies are needed to understand fully the role of surveillance in AVF management.
|
|
Table I -
Some unresolved issues in AVF surveillance
View |
|
|
Address for correspondence:
Nicola Tessitore, MD
Divisione di Nefrologia
Servizio Emodialisi Ospedale Policlinico
Piazzale LA Scuro, 10
37134 Verona - Italy
nicola.tessitore@azosp.vr.it
REFERENCES
1. National Kidney Foundation. K/DOQI Clinical Practice Guidelines for Vascular Access, Update 2000. Am J Kidney Dis 2001; 37 (suppl 1): S137-81.
2. Schwab SJ, Oliver MJ, Suhocki P, McCann R. Hemodialysis arteriovenous access: detection of stenosis and response to treatment by vascular access blood flow. Kidney Int 2001; 59: 358-62.
3. McCarley P, Wingard RL, Shyr Y, Pettus W, Hakim RM, Ikizler TA. Vascular access blood flow monitoring reduces access morbidity and costs. Kidney Int 2001; 60: 1164-72.
4. Konner K, Hulbert-Shearon TE, Roys EC, Port FK. Tailoring the initial vascular access for dialysis patients. Kidney Int 2002; 62: 329-38.
5. Tessitore N, Bedogna V, Gammaro L, et al. Diagnostic accuracy of ultrasound dilution access blood flow measurement in detecting stenosis and predicting thrombosis in native forearm arteriovenous fistulae for hemodialysis. Am J Kidney Dis 2003; 42: 331-41.
6. Elseviers MM, Van Waeleghem JP. Identifying vascular access complications among ESRD patients in Europe. Nephrol News Issues 2003; 17: 61-8.
7. Ravani P, Brunori G, Mandolfo S, et al. Cardiovascular comorbidity and late referral impact arteriovenous fistula survival: a prospective multicenter study. J Am Soc Nephrol 2004; 15: 204-9.
8. Branger B, Granolleras C, Dauzat M, et al. Frequence de thromboses de fistules arterio-veineuses pour hemodialyse: aport de deux methodes de surveillance: le Doppler et la diluition des ultrasons. Nephrologie 2004; 25: 17-22.
9. Basile C, Ruggieri G, Vernaglione L, Montanaro A, Giordano R. The natural history of autogenous radio-cephalic wrist arteriovenous fistulas of hemodialysis patients: a prospective observational study. Nephrol Dial Transplant 2004; 19: 1231-6.
10. Bakran A, Mickley V, Passlick-Deetjen J. Management of the renal patient: clinical algorithms on vascular access for hemodialysis: www.vascularaccesssociety.com 2003.
11. Tonelli M, Jhangri SH, Hirsch DJ, et al. Best threshold for diagnosis of stenosis or thrombosis within six months of access flow measurement in arteriovenous fistulae. J Am Soc Nephrol 2003; 14: 3264-9.
12. Hoeben H, Abu-Alfa AK, Reilly RF, Aruny JE, Bouman K, Perazella MA. Vascular access surveillance: evaluation of combining dynamic venous pressure and vascular access blood flow measurements. Am J Nephrol 2003; 23: 403-8.
13. Miranda CL, Sands JJ. Flow volumes as a predictor of hemodialysis access failure. J Vasc Technol 1998; 2: 73-6.
14. Schwarz C, Mitterbauer C, Boczula M, et al. Flow monitoring: performance characteristics of ultrasound dilution versus color Doppler ultrasound compared with fistulography. Am J Kidney Dis 2003; 42: 539-45.
15. Tessitore N, Mansueto G, Bedogna V, et al. A prospective controlled trial on effect of percutaneous transluminal angioplasty on functioning arteriovenous fistulae survival. J Am Soc Nephrol 2003; 14: 1623-7.
16. Lok CE, Bhola C, Croxford R, Richardson RMA. Reducing vascular access morbidity: a comparative trial of two vascular access monitoring strategies. Nephrol Dial Transplant 2003; 18: 1174-80.
17. Sands JJ, Jabyac PA, Miranda CL, Kapsick BJ. Intervention based on monthly monitoring decreases hemodialysis access thrombosis. ASAIO J 1999; 45: 147-50.
18. Tessitore N, Lipari G, Poli A, et al. Can blood flow surveillance and preemptive repair of subclinical stenosis prolong the useful life of arteriovenous fistulae? A randomized controlled study. Nephrol Dial Transplant 2004 (in press).
19. Turmel-Rodrigues L, Pengloan J, Baudin S, et al. Treatment of stenosis and thrombosis in hemodialysis fistulas and grafts by interventional radiology. Nephrol Dial Transplant 2000; 15: 2029-36.
|
|
N. Tessitore
Printable Format (PDF)