Abstract
Recent guidelines have recommended performing native arteriovenous fistulae (AVF) in hemodialysis patients rather than synthetic grafts whenever possible. However, early failure of AVF may reach up to 50%. The purpose of this study was to assess the factors associated with early failure of such procedures in hemodialysis patients.
A prospective study was performed on all patients with end-stage renal disease who had an AVF between June 2003 and March 2005. Data including patient characteristics and the type of AVF were recorded. The internal diameter of the vein and artery and intraoperative blood flow were measured. Patients were followed up for 3 months.
One hundred twenty-six AVF were included in this study. Early failure was in 14 (9%) patients. The internal diameter of the vein and artery and intraoperative blood flow were significantly lower in the failure group than in the patent group. The failure rate was not significantly related to other parameters.
Our data showed that intraoperative blood flow is a reliable parameter that determines the early failure of an AVF. Careful selection of the vein and the artery may reduce the rate of failure.
Several studies have shown that native arteriovenous fistulae (AVF) have higher long-term primary and secondary patency rates when compared with prosthetic grafts. 1,2 Recent guidelines published by the National Kidney Foundation Dialysis Outcome Quality Initiative (DOQI) on vascular access have recommended that a minimum of 50% of all new hemodialysis accesses should be AVF. 3 In Europe, an overall 80% of hemodialysis patients have AVF, whereas only 24% of patients in the United States use an AVF. 2 However, up to 50% of fistulae thrombose directly after operation or do not function adequately owing to failure of maturation. 4,5 Successful AVF depend on identifying and avoiding risk factors and surgical techniques associated with thrombosis or failed maturation after fistulae creation. The aim of this study was to evaluate the risk factors associated with early failure of AVF.
Patients and Methods
A prospective study was performed on all patients with end-stage renal disease (ESRD) referred to the Vascular Unit of King Fahd University Hospital between June 2003 and March 2005 for permanent vascular access. Careful examination of the arteries and superficial veins of the upper extremity of all patients was performed. All patients with signs of diseased arteries (absent or diminished pulses) underwent preoperative duplex ultrasonography. Preoperative upper limb venography was indicated in the following: (1) patients in whom superficial veins were not visible such as children, the obese, and those with limb edema; (2) before attempting vascular access for the second time in the same arm; (3) patients in whom a central vein line had been inserted previously. Most of these indications follow the DOQI guidelines. 3 Venography is also preferred over duplex scanning because the accuracy of duplex ultrasonography depends on the operator's skills and because approximately one-third of these cases need venography in addition to duplex ultrasonography. 6 Based on the preoperative examination and venography, the final planned vascular access was decided. The vessels were considered adequate for construction of a fistula if the diameter of the artery was more than 1.5 mm and that of the vein was greater than 2.0 mm. We followed the DOQI guidelines 3 for the order in which vascular access should be attempted: (1) radiocephalic AVF, (2) brachiocephalic AVF, and (3) transposed brachial basilic AVF. We tried to construct the AVF in the nondominant arm. If this was not possible, then the dominant arm was used. If there were no adequate peripheral veins in both upper limbs, access was performed using an arteriovenous graft. All patients who were considered unsuitable for AVF were excluded from the study. Data including age, gender, body mass index, smoking, diabetes, hypertension, hyperlipidemia, a history of access surgery, and the type of AVF were recorded.
Operative Procedure
All patients received a preoperative prophylactic single dose of antibiotic (coamoxiclav 1.2 g). The construction of the AVF was performed under local anesthesia, regional axillary block, or, rarely, general anesthesia. All patients were operated on by the same vascular surgeon or his senior resident, under his supervision.
After exposure of the artery and sufficient mobilization of the vein, the internal diameter of the vein and artery was measured using a coronary dilator or a ruler. If the diameter was considered adequate, an end-to-side vein-to-artery anastomosis was performed using 6-0 polypropylene. All patients received intravenous heparin during surgery to prevent thrombosis. In cases of brachial basilic vein fistula, superficialization of the basilic vein was performed before doing the vascular anastomosis. About 5 minutes after completion of the anastomosis, intraoperative blood flow (in mL/min) was measured using a handheld flow probe (Transonic System Inc. HT207, Ithaca, USA). This flowmeter uses transit-time ultrasonography. It measures the flow rate directly without the need for additional calculations related to cross-sectional area, which is needed when measuring the flow rate by conventional ultrasonography. The flow probe was placed around the venous outflow tract, and repeated measurements were made until a consistent reading was obtained. Postoperative evaluation was done by palpation and auscultation.
Follow-Up
Patients were followed up by both the nephrologists and the vascular surgeon. Dialysis was usually started by means of central vein catheters until cannulation of the fistula became possible. First cannulation was performed when the vessels had matured adequately, usually after 6 to 8 weeks. Follow-up was continued for 3 months after the procedure or earlier if there was failure of the fistula or the patient died. Failure of the fistula was defined as fistula thrombosis or an inability to cannulate both arterial and venous needles or to obtain sufficient dialysis blood flow (> 350 mL/min) within 8 weeks after fistula creation.
Statistical Analysis
The influence of individual parameters on the result of AVF creation and maturation was analyzed. Patients' characteristics were analyzed using the chi-square test with Yates correction for continuity. If there were two cells with expected counts less than 5, the Fisher exact test was used. Continuous variables, the diameter of the artery and vein, and intraoperative flow of failed and functioning AVF, were reported as mean ± standard deviation and compared using Student's t-test. Significance was set at p < .05 for all comparisons.
Results
During the study period, 125 patients were referred for access surgery. In six patients (5%), there were no adequate veins in both upper limbs and access was established using arteriovenous grafts. Nine patients had two consecutive fistulae. Therefore, 128 AVF were performed. One patient died after a kidney transplant; another was lost to follow-up. Thus, the study was performed on 126 AVF. Most of the patients were middle age (range 12–75 years), with a high incidence of diabetes mellitus (Table 1). Brachiobasilic fistula was the most commonly used AVF (44%) (Table 2).
Baseline Patient Characteristics
Operative Data
Early failure was in 11 fistulae (9%). Three fistulae (2%) failed to mature, and eight fistulae clotted (7%); four of these were in the immediate postoperative period, and the other four were during the follow-up period. All of the failed group underwent upper limb venography. No stenosis was detected in the veins. Nine cases had a second AVF, but two refused operations and were lost to follow-up. Thus, 115 (91%) fistulae developed adequately to be used for dialysis with a blood flow of at least 350 mL/min. There was no significant difference between successful and failed AVF in the baseline patients' characteristics (Table 3).
Effect of Baseline Patient Characteristics on the Success of Arteriovenous Fistulae
There was no statistically significant difference between types of fistulae. The intraoperative blood flow was significantly lower in the failure group than in the patent group. All fistulae with flow < 90 mL/min failed to maintain patency. The internal diameter of the vein and artery was also significantly lower in the failure group (Table 4).
Operative Data in Successful and Failed Arteriovenous Fistulae
Discussion
The recent recommendation of the DOQI may generate an aggressive approach for the creation of an AVF rather than a graft for hemodialysis patients. Some authors have attempted to develop objective criteria for the ideal hemodialysis access fistula. There is controversy about the type of patients who might be at risk of failed AVF.
Early failure is reported to be around 20 to 53%, 4,7–10 whereas we experienced a failure rate of only 9%. Early failure of an AVF may result from nonmaturation or thrombosis. The former cause has been the most common reason of early failure in some recent series (33–43%) 4,5 compared with only 7% in our series. It is possible that this difference is accounted for by the smaller number (5%) of older patients in our series than reported by others (30–40%). 7,10
The optimal preoperative preparation and surgical technique were used to reduce the incidence of AVF failure. Routine venous mapping has been reported to improve the maturation rate 10–12 and increase the creation of an AVF. 13 It has also been shown that the use of intraoperative heparin reduced the incidence of early AVF thrombosis. 8,14 The technique of end-to-side vascular anastomosis has also been shown to be associated with improved AVF success. 8
Our rate of AVF (95%) exceeds the 50% target of the DOQI guidelines 3 and compares favorably with published reports. 15,16 Most of the superficial veins of our hemodialysis patients usually have segmental occlusions as a result of multiple intravenous infusion during repeated hospital admissions. This may be the reason for the high incidence (44%) of transposed brachiobasilic AVF in the present study compared with 30 to 39% in other reports. 15,16
Few recent studies have been published about predictors of failure of AVF. Whereas the present study, like others, 17,18 has not identified gender as a risk factor for failure of AVF, some reports have shown that female gender was significantly correlated with failure. 4,9,19 In contradiction to some reports, 9,17 the present study and other reports 4,11,20,21 did not find any significant effect of diabetes mellitus on early failure of AVF. Old age has been reported to increase the risk of failure. 10,18,21 This is probably due to the loss of elasticity of the vessels in this age group. In our series, the older group failed to show a statistically significant difference compared with the younger group. Smoking and hypertension were also found to have no significant effect on early failure of AVF in the present series and others. 9 Overweight was another significant risk factor for failure in some studies. 4,18 However, it failed to show any correlation in our patients and other reports. 11,22 The effect of the site of the fistula on the failure rate has been studied. Some series, including ours, did not show any significant effect. 4,11,15 Others showed that distal fistulae were at higher risk of failure. 19
The use of an artery with a large internal diameter was associated with a high success rate in the present study and in others. 5,23 In the present study, the internal diameter of the artery below which there would be a high failure rate was 1.8 mm. This observation is comparable with others, 10,19 except the study by Tordior and colleagues, who reported no significant impact of radial arterial diameter on AVF outcome. 4 The same observation applied to the size of the veins. The present series and others showed that a venous diameter less than 2.3 mm was associated with a high failure rate. 4,12,24 The fact that most of our cases had the optimal arterial and venous diameters explains the low incidence of failure in our patients.
The role of access blood flow measurement for predicting vascular access failure has been studied. There are several methods to measure vascular blood flow, including Doppler ultrasonography, indicator dilution technique, or, more recently, transit-time ultrasonography, which we used in our study. Its advantage over Doppler ultrasonography is that it is not dependent on the angle of isonation and is diameter independent. 25,26 Access blood flow proved to be a good predictor of early failure of AVF in our series and in those of others. 18,27 Only one study reported measuring intraoperative blood flow in 27 instances and did not find a relationship between flow rates and outcome. 23 In the present series, the flow rate below which the fistula should be abandoned immediately was 90 mL/min, which is comparable with other reports. 16,27 A flow rate of 160 mL/min was reported to have a high risk of early failure in both our study and others. 27 In these borderline cases, the fistula should be watched carefully and allowed longer time to mature.
In conclusion, intraoperative blood flow rate is a reliable parameter that determines early failure of an AVF. Careful selection of the vein and the artery is likely to improve the outcome.
Footnotes
Presented in part at the 4th International Congress of the Vascular Access Society, Berlin, Germany, May 2005.