Hemodialysis vascular access requires repetitive reliable access to the circulation1,2,3,4. This access to the circulation should meet three criteria. First, it should be suitable for repetitive circulatory access. Second, it should allow for a blood flow suitable to conduct modern high-efficiency dialysis. Third, the complication rate should be minimal.
Currently there are three types of hemodialysis vascular access: native arteriovenous fistulas (radiocephalic, brachiocephalic, transposed brachiobasilic, brachiobasilic fistulae), arteriovenous (AV) grafts, and central venous catheters1,2,3,4. Cuffed tunneled hemodialysis catheters are the most recent addition to the permanent vascular access armamentarium5. These vascular access devices do a creditable job performing the first two functions but do have a substantial complication rate1,2,3,4.
Complications of hemodialysis vascular access have emerged as a major cause of patient morbidity and a major cost to the end-stage renal disease (ESRD) program. In its latest report, the U.S. Renal Data System estimated the cost for access morbidity approaches $8000/patient/year at risk. Health care organizations planning for a capitated environment estimate as much as 25% of the total cost of the ESRD program may actually be spent on the maintenance of vascular access6,7,8. Thus, maintenance of access to the circulation has emerged not only as a major cause of patient morbidity, but is one of the single largest expenses in the United States ESRD system.
Hemodialysis access failure and morbidity have been carefully evaluated. The recent publication of the National Kidney Foundation Dialysis Outcome Quality Initiative represented the single largest compendium of literature dealing with hemodialysis vascular access2. This group of practice guidelines clearly identified that primary AV fistulae have superior patency compared with AV grafts, and that patency of all AV access was significantly improved by prospective monitoring and therapeutic intervention to resolve venous outflow stenoses. The repetitive measure of hemodialysis access flow, using various technologies, was identified as a sensitive and specific indicator of outflow venous stenosis.
Despite these significant new observations, our understanding of access flow criteria for intervention and the response to intervention have not been well delineated. The purposes of this pilot study, therefore, were (1) to evaluate an access flow threshold for access flow monitoring in regard to detection of venous outflow stenosis, (2) to confirm the relationship between AV access flow and the likelihood of AV access thrombosis, and (3) to define the impact of intervention with both percutaneous transluminal angioplasty (PTA) and surgery on access flow, and ultimately access thrombosis on of AV fistulae and AV grafts.
METHODS
This study represents a single-center pilot, prospective observational intervention study. AV access flows and elective AV access interventions were the evaluated events. Access thrombosis was used as the study ending event. Access flows were measured monthly and within one week of all interventions. Access flows were measured in the first hour of the hemodialysis treatment. Access flows were performed utilizing the ultrasound dilution technique9,10. Access flows measured in the first hour of hemodialysis have been found to be reproducible and relatively unaffected by the falls in cardiac output associated with ultrafiltration11. Elective access interventions were transluminal angioplasty and surgical revision. Prior to elective intervention, all AV accesses were evaluated by contrast fistulography. Radiocontrast fistulography was performed as previously described12. Radiocontrast fistulography was performed for any 20% decrement in access flow when the access flow was <1000 mL/min or when absolute access flow decreased to under 600 mL/min. These flow parameters were chosen based on results obtained in our preliminary trial and on previously published findings10. Any access stenosis> 50% of the lumen diameter (as determined by biplanar angiography) when associated with a significant flow decrement was initially treated with transluminal angioplasty (using our previously described technique)12 performed at the time of fistulogram. Patients were referred for surgical revision following failure of PTA to restore access flow by> 20% of the baseline. Intervention success for the purpose of the study was defined by an increase in access flow> 20% of the baseline measured within one-week post-procedure. This value of 20% was selected based on the concept that flow should be restored at least to the extent of the screening criteria to be considered successful. Data were collected on access flow and all access interventions during the study. Retrospective thrombosis data were compiled from this same dialysis center in the two years before the study for comparison. Comparison was performed using paired t-tests for paired data, unpaired t-test for unpaired data, and rate comparison analysis for historic and study rates.
Patients
Forty-two patients in a university-based hemodialysis unit with arteriovenous access were included in the study. The mean age of the patients was 61 
9 years (race, 29 black/13 white; gender, 26 female/16 male). Twenty-six of the 42 patients (62%) had diabetes mellitus as the cause of ESRD. Total follow-up was 64 patient-years of hemodialysis (range, 0.6 to 2.4 patient-years; median follow-up, 1.78 years). There were no eligible patients excluded from the study. Twenty-four patients (33%) had native AV fistulae (3 lower arm, 11 upper arm), while 28 patients dialyzed via AV grafts (67%; 6 lower arm, 18 upper arm, 2 axillary jugular, 1 axillary axillary, 1 thigh). Upper arm fistulae were brachial cephalic in six instances and transposed brachial basilic in five instances. Transposed brachial basilic fistulae were created by elevation and transposition of the basilic vein across the anterior surface of the upper arm. Patients receiving in-center hemodialysis who were training for home hemodialysis or peritoneal dialysis were deemed ineligible because of the short in-center follow-up (<3 months). During the course of the study, three patients died and three underwent transplantation. There were no patients in the facility receiving hemodialysis via an AV access who were maintained on Warfarin or other form of systemic anticoagulation. Monthly flow measurements were performed on all patients in the facility with an AV access as part of the facility access monitoring program when the facility converted from daily dynamic venous pressure monitoring. At the time of study entry, there were no patients with a detected access flow of less than 750 mL/min.
RESULTS
Figure 1 depicts the mean access flow at study entry and end for the study group as a whole. Figure 1 also shows access flow by type of access (AVF-AVG), for the intervention group pre- and post- all PTA (RX; by AVF-AVG) and at the study end for the intervention group alone. Thirty-seven referrals for diagnostic venogram of the AV access were done for study indications. All 37 referrals were performed because a falling access flow threshold was met (33 flow decrements, 4 absolute flow). Thirty-five diagnostic venogram evaluations of the AV access were performed. Two referrals did not keep evaluation appointments, and both thrombosed. All 35 successful referrals underwent venography of the access. All 35 were found to have a> 50% stenosis in the AV access and underwent PTA. There were no study venograms performed that did not detect a> 50% access stenosis. Stenoses in AV grafts occurred primarily at or within 6 cm of the vein graft anastomoses (74%). The remainder occurred more proximally in the venous circulation with one innominant vein stenosis. There were no arterial and only one intragraft stenoses detected. In AV fistulae, there was one central vein stenosis, with the remainder occurring in the venous outflow tract at sites of vein bifurcation and venous valves.
Figure 1.
Arteriovenous (AV) access flow measured in mL/min (y axis) at key study times in the study (x axis). Groups designated "All" represent all AV access active in the study. Groups designated RX represent intervention groups only. Values are reported SEM. P < 0.05 (study entry vs. pre-first rx); P < 0.05 (pre-first rx vs. post-first rx); P < 0.05 (post-first rx vs. pre-2nd rx); P < 0.05 (pre-2nd rx vs. post-2nd rx). Symbols are: (
) fistula; (
) graft.
Two patients were referred for venography by vascular surgery consultants for various indications independent of flow criteria (AV fistula aneurysm evaluation, pseudo aneurysm evaluation) independent of the study. In addition following completion of the study, seven patients (3 AVF and 4 AVG) were evaluated with biplanar venography for nonflow indications (access pain, bleeding post-needle removal, difficult fistula cannulation). There was one significant (50%) stenosis detected in these nine patients. PTA of this stenosis had no effect on access flow.
There was a total of 35 PTA (successful and unsuccessful) during the study. There were 15 primary, 11 secondary, and 2 tertiary PTAs in patients with AV grafts. There were six primary and one secondary PTA in the AVF group. Failure to increase flow by 20% post-PTA defined failure of PTA. This occurred in 14% of fistula PTA attempts (1 of 7 PTA) and in 21% of AV graft (6 of 28 PTA). Access flow results pre- and post-PTA are graphically represented in Figure 1. All PTA attempts are recorded in this figure, including those deemed successful and unsuccessful. Flow at the end of the study for all access that were still patent is also shown Figure 1. These end-study data represent all patients and include those with and without an intervention.
Mean time from first to second PTA was 5.8 months in AV grafts. This could not be calculated for AVFs because there was only one repeat intervention. Mean follow-up post-PTA for AVF stood at 11.4 months at study closure. Three patients (3 of 7 PTA failures) deferred surgical revision, as described later in this article. Elective surgical revisions occurred in four patients, all with AV grafts. Mean increase in flow post-surgical revision was 655 88 to 891 136 mL/min (P < 0.05).
During the study period that encompassed 64 patient-years of hemodialysis, there were 10 episodes of thrombosis Table 1; this calculates to a total thrombosis rate of 0.156 or on average 16% per patient-year. The thrombosis rate per patient-year of hemodialysis for AVF was considerable less than for AV grafts (0.216, AVG vs. 0.074, AVF; P < 0.05). Eight of the 10 thromboses were predicted in that they occurred in patients who had an unsuccessful PTA and who were unwilling to undergo elective surgical revision or in whom a fistulogram was recommended and who postponed or canceled their fistulogram appointment (AV grafts thrombosed within 4 months of diminished flow, AV fistulae within 1 year of diminished flow). Two thromboses occurred in patients in whom access flow did not detect a significant drop in flow. Their previous access history did not appear different from the group as a whole.
Percutaneous transluminal angioplasty was a common event in the study. There were 0.359 PTA per patient-year of hemodialysis (36% per patient-year; Table 1). The intervention rate was much higher in AVG than in AVF (0.75 compared with 0.26, P < 0.05; Table 1). The mean time between the first and second interventions in AVG was 5.8 months. There were not enough second interventions in AVF to calculate a rate. In all instances except one, the repeat interventions were performed on a recurrent and not a new stenosis.
When patients rather than episodes of referral were examined, PTA was performed in 15 of 28 patients with AV grafts (54%) and in 6 of 14 patients with AVF (43%).
The thrombosis rate for this study overall was 0.156 or 16% per patient-year. Our previous thrombosis rate for this same unit for the two years preceding the study was 0.251 or 25% per patient-year using dynamic venous pressure as the primary screening technique (P < 0.05). When compared by access type, the rate of thrombosis for AVF was 7% during the study and 16% during in the preceding two years (P < 0.05). The thrombosis rate for grafts was 22% during the study and 30% during the preceding two years (P = NS).
DISCUSSION
The concept of prospective monitoring with elective intervention to maintain patency of AV access has been substantiated in multiple clinical trials1,2,13,14,15,16,17,18,19. Our prospective, observational pilot study evaluated and extended the role of access flow monitoring in an overall hemodialysis access maintenance program, and suggests that monitoring with access flow has benefits beyond using dynamic venous pressure as a monitoring tool, especially in AV fistulae.
This study suggests that the trend of access flow is as important as the overall access flow. Decrements in access flow> 20%, when the flow was less than 1000 mL/min, were highly predictive of a significant AV access stenosis and when flow was not restored, predictive of access thrombosis. All patients who met these criteria and who underwent fistulogram had a significant stenosis either within the AV access or in the venous outflow tract. The vast majority of interventions occurred because of decrements in access flow and not for absolute flow criteria. The study also confirmed that failure to restore access flow led to a high likelihood of access thrombosis. There was a total of 10 access thromboses during the study period. Eight of these 10 were predicted by falling access flow measurements. In these eight instances, the patients failed to keep their appointment for prospective fistulogram or for access revision in the case of PTA failure. Only two episodes of thromboses occurred that were not expected based on access flow measurements. Thus, decrements openface> 20% in AV access flow, when the flows are <1000 mL/min, appear to be predictive of a hemodynamically significant access stenosis and an ultimate thrombosis. Patients without a fall in access flow were unlikely to have a hemodynamically significant stenosis. An alternate screening approach could have been to attempt to determine an absolute flow decrement rate as a monitor threshold rather than percentage decrement. We choose the percentage decrement for this pilot study based on our previous experience.
This study confirms that successful PTA or surgical access revision can restore flow. It also outlines that repetitive interventions become progressively less successful over time in AV grafts. It confirms that not all PTAs are successful despite the impression of physician performing the procedure. When follow-up flow measures were performed, 14% of AVF and 21% of AVG PTA failed to improve flow> 20%. The mean period of time between a successful intervention and the need for repeat intervention is approximately 5.8 months in AV grafts. Thus, intimal and fibromuscular hyperplasia proceeds regardless of the intervention. Repetitive transluminal angioplasty successfully returns lower and lower fractions of overall access flow as outlined in Figure 1. Thus, although prospective intervention significantly modifies vascular anatomy to allow continued flow, ideal long-term therapy should be prevention of this underlying biologic event. Nonetheless, as is clear from this observation, successful angioplasty and surgical revision continue to return flow back in a positive manner in the direction of the previous baseline. The article also shows that surgical revision is capable of salvaging PTA failures in AV grafts.
Interestingly, intervention in AV fistulae appears to be more durable than interventions in AV grafts (mean follow-up> 11 months at study end). The reason for this remarkably better durability is unclear (Table 1 and Figure 1). We speculate that either the PTA allows time for collaterals to develop in AVFs or intimal and fibromuscular hyperplasia occurs at a slower rate in AVFs.
Thrombosis rates overall and most marked thrombosis rates in AVF were improved in this prospective, access-flow pilot trial compared with retrospective controls screening with dynamic venous dialysis pressure. The improvement was most striking in AV fistula, where the thromboses rate improved from 16% percentage per patient-year to 7%. This is even more important when one considers that most thromboses in AV fistulae result in access loss because thrombectomy and thrombolysis are difficult to perform in AVF. Recent observations are changing this trend, but most U.S. centers, including ours, still have limited success in salvaging a fistula after thrombosis20. AVG thrombosis rates were also improved (30 to 22%) but did not reach statistical significance. Thus, in our pilot study, monitoring for flow had advantages over pressure monitoring as a monitoring technique. The significant new message is the value of flow monitoring and prospective PTA in AV fistulae. Not only does it appear that thromboses are avoided, but unlike AV grafts, the correction appears to be more durable.
An appropriate criticism of this concept is that monitoring especially of AV grafts merely replaces one intervention with another Table 1. The number of interventions, both transluminal angioplasty and hemodialysis vascular access revision, was 0.359 or 36% per patient year. When AVG grafts alone are examined, the intervention rate to maintain patency was 0.75 interventions per patient year. Nonetheless, the use of an elective outpatient intervention appears to extend access patency significantly and diminish the need for urgent thrombectomies and thrombolyses. The overall thrombosis rate reflects a significant improvement from our previous value of 0.25 thromboses per patient-year to 0.156 thromboses per patient-year when dynamic venous pressure was used as our primary monitoring technique. The case for monitoring and correcting native AV fistulae seems clear. Dramatic decrements in AVF thrombosis rates are obtained by flow monitoring and prospective PTA.
This prospective pilot study has several drawbacks. The size of the study population is small, and the follow-up period is limited. In addition, a retrospective rather than a concurrent control group is used. Nonetheless, the results are striking, and multiple new observations have emerged as outlined previously in this article. As with all small pilot trials, larger confirmatory studies will be required.
In summary, prospective monitoring with ultrasound dilution hemodialysis access flows is an excellent monitoring technique for detecting significant AV access outflow stenoses. When this screening technique is combined with PTA and surgical revision, decrements in thrombosis rate occur in both native AV fistulas and polytetrafluorethylene (PTFE) grafts. Angioplasty in both fistulas and grafts is not curative but, nonetheless, significantly extends the overall patency rate by increasing the flow. Repetitive angioplasties are required on any given outflow stenosis, especially in AV grafts, as intimal fibromuscular hyperplasia continues unabated despite the mechanical intervention. Thus, the ideal long-term therapy in the treatment of hemodialysis access is the prevention of fibromuscular and intimal hyperplasia rather than correction of an existing ongoing pathologic process.
References
- Schwab, SJ, Harrington, JT & Singh, A, et al: Vascular access for hemodialysis. Kidney Int 1999 55:2078–2090, 10.1046/j.1523-1755.1999.00409.x | Article | PubMed | ISI | ChemPort |
- National Kidney Foundation-Dialysis Outcomes Quality Initiative: NKF-DOQI clinical practice guidelines for vascular access. Am J Kidney Dis 1997 30:S150–S191,
- Fan, PY & Schwab, SJ: Vascular access: Concepts for the 1990s. J Am Soc Nephrol 1992 3:1–11, | PubMed | ISI | ChemPort |
- Windus, DW: Permanent vascular access: A nephrologist's view. Am J Kidney Dis 1993 21:457–471, | PubMed | ISI | ChemPort |
- Schwab, SJ & Beathard, G: The hemodialysis catheter conundrum: Hate living with them, but can't live without them. Kidney Int 1999 56:1–17, 10.1046/j.1523-1755.1999.00512.x | Article | PubMed | ISI | ChemPort |
- Feldman, HI, Kobrin, S & Wasserstein, A: Hemodialysis vascular access morbidity. J Am Soc Nephrol 1996 7:523–535, | PubMed | ISI | ChemPort |
- Feldman, HI, Held, PJ & Hutchinson, JT, et al: Hemodialysis vascular access morbidity in the United States. Kidney Int 1993 43:1091–1096, | PubMed | ISI | ChemPort |
- United States Renal Data System: The economic cost of ESRD, vascular access procedures, and Medicare spending for alternative modalities of treatment. USRDS. Am J Kidney Dis 1997 30:S160–S177,
- Krivitski, NM: Theory and validation of access flow measurement by dilution technique during hemodialysis. Kidney Int 1995 48:244–250, | PubMed | ISI | ChemPort |
- Depner, TA & Krivitski, NM: Clinical measurement of blood flow in hemodialysis access fistulae and grafts by ultrasound dilution. ASAIO J 1995 41:M745–M749, | PubMed | ChemPort |
- Neyra, NR, Ikizler, TA & May, RE, et al: Change in access blood flow over time predicts vascular access thrombosis. Kidney Int 1998 54:1714–1719, 10.1046/j.1523-1755.1998.00145.x | Article | PubMed | ISI | ChemPort |
- Schwab, SJ, Raymond, JR & Saeed, M, et al: Prevention of hemodialysis fistula thrombosis: Early detection of venous stenoses. Kidney Int 1989 36:707–711, | PubMed | ISI | ChemPort |
- Besarab, A, Sullivan, KL & Ross, RP, et al: Utility of intra-access pressure monitoring in detecting and correcting venous outlet stenoses prior to thrombosis. Kidney Int 1995 47:1364–1373, | PubMed | ISI | ChemPort |
- Kanterman, RY, Vesely, TM & Pilgram, TK, et al: Dialysis access grafts: Anatomic location of venous stenosis and results of angioplasty. Radiology 1995 195:135–139, | PubMed | ISI | ChemPort |
- Valji, K, Bookstein, JJ & Roberts, AC, et al: Pharmacomechanical thrombolysis and angioplasty in the management of clotted hemodialysis grafts: Early and late clinical results. Radiology 1991 178:243–247, | PubMed | ISI | ChemPort |
- Safa, AA, Valji, K & Roberts, AC, et al: Detection and treatment of dysfunctional hemodialysis access grafts: Effect of a surveillance program on graft patency and the incidence of thrombosis. Radiology 1996 199:653–657, | PubMed | ISI | ChemPort |
- Sands, JJ & Miranda, CL: Prolongation of hemodialysis access survival with elective revision. Clin Nephrol 1995 44:329–333, | PubMed | ISI | ChemPort |
- Turmel-Rodrigues, L, Pengloan, J & Blanchier, D, et al: Insufficient dialysis shunts: Improved long-term patency rates with close hemodynamic monitoring, repeated percutaneous balloon angioplasty, and stent placement. Radiology 1993 187:273–278, | PubMed | ChemPort |
- Besarab, A, Frinak, S & Sherman, RA, et al: Simplified measurement of intra-access pressure. J Am Soc Nephrol 1998 9:284–289, | PubMed | ISI | ChemPort |
- Turmel-Rodrigues, L, Pengloan, J & Baudin, S, et al: Treatment of failed native arterio-venous fistulae for hemodialysis by interventional radiology: Technique, results, and comparison with prosthetic grafts. Kidney Int 2000 57:1124–1140, 10.1046/j.1523-1755.2000.00940.x | Article | PubMed | ISI | ChemPort |
