- Email: [email protected]
Previous Hemodialysis Access Improves Functional Outcomes of the Proximal Radial Artery Fistula in Males
Previous Hemodialysis Access Improves Functional Outcomes of the Proximal Radial Artery Fistula in Males Michael F. Amendola,1,2 John Pfeifer,1 Francisco Albuquerque,1,2 Luke Wolfe,2 Mark M. Levy,1,2 and Ronald K. Davis,1 Richmond, Virginia
Background: The proximal radial artery fistula (PRA) has been established as an early viable surgical option for arteriovenous fistula creation. The overall assisted primary patency reported in the literature approaches 100% at 1 year. We hypothesize that this excellent patency does not represent a functional result when seen in light of successful cannulation and fistula utilization. Methods: We retrospectively queried our Veterans Administration Hospital operative database to identify 284 male patients who had 571 access procedures performed by a senior vascular surgeon attending (R.K.D.) from January 1, 2003, to December 31, 2008. Operative details, patient comorbidities, fistula maturation time (time to first cannulation), functional patency (date of access to abandonment, revision to another fistula type, conversion to a prosthetic graft, thrombosis of the fistula, conversion to peritoneal dialysis, renal transplant, or patient death), and total duration (creation of the fistula to the end of its functional patency) were collected and analyzed. Results: A total of 144 PRAs were placed during the study period. In all, 87 patients underwent primary proximal radial artery fistula (P-PRA) placement in a limb without previous access; 57 patients had a secondary proximal radial artery fistula (S-PRA) after a failed previous fistula or graft in the same limb. There were no differences between the 2 groups in terms of age, comorbidities, and operative details. A total of 91 patients (63.2%) were receiving hemodialysis at the time of P-PRA or S-PRA placement. Outcomes of P-PRA and S-PRA populations on hemodialysis were examined. There was increased cannulation success (33% vs. 55%; P ¼ 0.00354, Fisher’s exact test), functional patency (755.2 ± 661.2 days vs. 405.4 ± 531.9 days; P ¼ 0.0220, Wilcoxon two-sample test), and total duration (859.5 ± 650.7 days vs. 516.8 ± 547.2 days; P ¼ 0.0361, Wilcoxon two-sample test) of S-PRA over P-PRA. There was no difference in endovascular interventions between the 2 groups (1.6 ± 1.0 interventions per access versus 1.1 ± 0.7 interventions per access; P ¼ 0.2109, Wilcoxon two-sample test). Subgroup analysis (analysis of variance) of the S-PRA group indicated that a patent but failing previous access in the same arm was not superior in terms of successful cannulation, functional patency, or total duration when compared with a thrombosed previous access. Conclusions: The PRA remains a viable first access procedure undertaken at our institution. Compared with the reported 12-month assisted primary patency of this fistula type, we found a small percentage of PRAs actually being accessed for successful hemodialysis treatment. The S-PRA appears to have a significantly higher successful cannulation rate, functional
1
Department of Surgery, McGuire VA Medical Center, Richmond, VA. Division of Vascular Surgery, Virginia Commonwealth University Health System, Richmond, VA. 2
Correspondence to: Michael F. Amendola, MD, Department of Surgery, Hunter Holmes McGuire VA Medical Center, Surgical Services, 1201 Broad Rock Boulevard, Richmond, VA 23249, USA; E-mail: [email protected]
920
Ann Vasc Surg 2015; 29: 920–926 http://dx.doi.org/10.1016/j.avsg.2014.12.021 Published by Elsevier Inc. Manuscript received: August 18, 2014; manuscript accepted: December 24, 2014; published online: March 7, 2015.
Vol. 29, No. 5, July 2015
Previous hemodialysis access improves outcomes of PRA 921
patency, and total duration time when compared with the P-PRA in patients receiving hemodialysis treatments. The mechanism of these improved outcomes is not known; considering patency or thrombosis of a previous access in the S-PRA group did not predict future access success in the same extremity.
INTRODUCTION It has been estimated that approximately 19.2 million people in the United States are affected with some form of chronic renal disease.1 This poses an ever-mounting public health risk with a cost in excess of 40 billion dollars and includes an estimated 398,861 patients receiving hemodialysis treatments on a periodic basis.2 In spite of national awareness and emphasis for fistula placement, a large percentage of patients remain on hemodialysis via central venous catheterization.1 In 1997, the National Kidney Foundation Kidney Dialysis Outcomes Quality Initiative clinical practice guidelines recommended native arteriovenous fistulas as the best hemodialysis access in patients with end-stage renal disease.3 This was based on known negative outcomes of patients who received hemodialysis from central catheterization or via arteriovenous grafts. Creation of an arteriovenous fistula using the cephalic vein to the radial artery (Cimino Fistula) was first described in 1966.4 The Cimino fistula has become the preferred first line access site for most patients receiving hemodialysis treatments. This access type, however, has a known high incidence of primary failure because of early thrombosis or maturation failure.5 The proximal radial artery fistula (PRA) has been promoted as an alternative inflow site for arteriovenous fistula creation. This fistula type was first described in 1977 in 20 patients, half of whom were diagnosed with diabetes.6 This initial report used the cephalic vein as the outflow source of the fistula. There was variable follow-up in this study with a reported primary patency rate of 90%. The PRA was reintroduced in a seminal study by Bruns and Jennings in 2003.7 Their work detailed the creation of 73 PRAs in 71 patients. The overall assisted primary patency approached 100% at 1 year. This study was followed with an additional report of 105 PRAs, of which, 34 were reported as a secondary fistula procedure after a previously failed access in the same limb.8 Again, the overall assisted patency was excellent at 97%. In each of these reports, patency was defined as the percentage of fistulas that had flow at some future time point, not necessarily the percent of PRAs being successfully accessed for hemodialysis treatments. The outflow
vein was described as basilic, cephalic, median cubital, lateral antecubital, or perforating deep vein. Patency is an important measurement of arteriovenous fistula function; however, most would agree that a more accurate marker of a success should be the amount of time that the fistula is needle cannulated for hemodialysis treatments. This definition serves as a functional measure of the usefulness of the arteriovenous fistula in the treatment of renal failure. We hypothesize that previous access in the same extremity before the creation of the PRA will favorably influence outcomes secondary to dilation of the venous circuit in that limb with a planned future access.
METHODS Study Population We retrospectively queried our institutional operative database from January 1, 2003, to December 31, 2008, to identify all vascular access procedures that were performed by a senior vascular surgery attending (R.K.D.) with the assistance of either a second or third year general surgery resident. Primary proximal radial artery fistula (P-PRA) was defined as a fistula that was constructed in an extremity that did not have any prior fistula or graft in the same extremity. Secondary proximal radial artery fistula (S-PRA) was defined as PRA construction in an extremity that did have a prior fistula or graft either thrombosed or nonfunctional. Data Collection All cases were collected and stored in an electronic clinical database per institutional guidelines. The operative records, date of each surgery, side of operation, indications, anastomosis type, date and type of subsequent operative revisions, and date and type of subsequent interventional procedures were identified. The patient’s electronic medical record was also examined to extract the following variables: date of birth, date of death, central venous catheterization, history of diabetes mellitus and hypertension, gender, ethnicity, and statin medicine use. Hemodialysis treatment records were reviewed for date of first cannulation and date of access abandonment.
922 Amendola et al.
Annals of Vascular Surgery
All endovascular procedures were noted and captured in the study database. These included all fistulograms for diagnostic and interventional purposes. Ligations of tributary veins with a patent fistula and elevation of an existing fistula without revision of anastomosis were not considered conversion to another fistula type. Functional Assessment Successful cannulation was defined as accessing the fistula with 1 or 2 needles for a complete hemodialysis treatment. Time to maturation referred to the number of days from fistula creation to successful cannulation. Functional patency was defined as the number of days from successful cannulation to access abandonment, revision to another fistula type, conversion to a prosthetic graft, thrombosis of the fistula, conversion to peritoneal dialysis, renal transplant, or patient death. Total duration was the number of days from the creation of the fistula to the end of its functional patency. Censored end points for patency included patient death or a patent fistula at the end of the study period.
Fig. 1. Histogram of arteriovenous access types and total for each year studied. Cimino, Cimino fistula; AVG, arteriovenous graft; Revised AVG, revisions of arteriovenous grafts; Revised AVF, revisions of arteriovenous fistula.
Statistical Analysis All statistical analyses were performed using the Windows based SAS System version 9.3 (SAS Institute Inc., Cary, NC). All tests were 2-tailed with a 0.05 significance level. Continuous factors were compared using the Wilcoxon rank test. Discrete factors were compared using the Fisher’s exact test. Logistic regression was used to identify independent predictors of improved patency in S-PRA subgroup.
Procedural Details
Follow-up
PRA construction was routinely undertaken with regional block and local anesthesia for augmentation as needed. All patients had their venous anatomy examined after regional block and tourniquet application to delineate what type of fistula would be created or modified. Ultrasound was not used to preoperatively assess venous anatomy. Once sterile drape was established and the decision to undergo PRA was made, local exploration of the antecubital fossa was undertaken through a longitudinal incision. Outflow veins were identified at this time as either originating from the cephalic, basilica, or a combination thereof. The proximal radial artery was identified an in-flow source for the fistula. The outflow vein was then dissected from its subcutaneous tissue. The patient was then given systemic heparin (100 U/kg), after 3 min, the proximal radial artery was clamped and then anastomosed to the outflow vein in either a side-to-side (n ¼ 18) or end-to-side (n ¼ 126) manner. The anastomosis was undertaken with either 6-0 or 7-0 Gore-TexÒ suture (W. L. Gore and Associates Inc., Flagstaff, AZ). When possible and under direct vision, the first distal vein valve was dilated with a coronary dilator. Patency was confirmed with hand-held Doppler probe after the flow was restored.
All patients had routine follow-up at 2 weeks after PRA placement and then at the discretion of the examining surgeon. Patient outcomes were tracked until December 31, 2008, for a total maximum of 1,095 days. Institutional Board Review The experimental protocols and the waiver of informed consent were approved by the Institutional Review Board (IRB # 01836) at the McGuire Veterans’ Affairs Medical Center in Richmond, Virginia before the start of this study.
RESULTS Population as a Whole During the study period from January 1, 2003, to December 31, 2008, we identified 284 male patients who had 571 access procedures. A total of 144 (25.2%) of the procedures were PRAs. The remaining population underwent a variety of other fistulas including Cimino arteriovenous fistula placement. Revisions of existing fistulas and arteriovenous grafts were also common during the study period. A yearby-year breakdown as a percentage of the total access placed is depicted in Figure 1. The percentage of PRA
Vol. 29, No. 5, July 2015
Previous hemodialysis access improves outcomes of PRA 923
Table I. Patient demographics among 144 patients undergoing PRA construction Demographic
P-PRA (n ¼ 87)
S-PRA (n ¼ 57)
P
Hemodialysis treatment Age, years, mean ± SD Alive at study end Left sided Hypertension African American Caucasian Diabetes Statin use
52.8% (n ¼ 46) 69.7 ± 10.7 31.03% (n ¼ 27) 39.08% (n ¼ 53) 89.66% (n ¼ 78) 64.37% (n ¼ 56) 35.63% (n ¼ 31) 64.37% (n ¼ 56) 53.49% (n ¼ 46)
78.9% (n ¼ 45) 68.8 ± 11.2 31.58 (n ¼ 18) 29.82% (n ¼ 40) 92.98% (n ¼ 53) 66.67% (n ¼ 38) 29.82% (n ¼ 17) 57.89% (n ¼ 33) 38.60% (n ¼ 22)
0.0015a 0.7023b 1.0000a 0.2884a 0.5659a 0.3301c 0.3301c 0.4849a 0.0898a
a
Fisher’s exact test. Wilcoxon two-sample test. c Chi-squared analysis. b
placements varied during the period ranging from 14% to 34% in any given year. The percentage of grafts started at 24% of our total access procedures in 2003 and dropped to 7% in 2008. The percentage of revisions of arteriovenous fistulas increased dramatically from 17% in 2003 to 41% in 2008. Patient Demographic Information Patient demographic information was collected and compared between the P-PRA and S-PRA groups as detailed in Table I. In all, 52.8% (n ¼ 46) of the P-PRA population was receiving hemodialysis treatments compared with 78.9% (n ¼ 45) of the S-PRA with a significant P value of 0.001. Demographics including age, death rates, history of hypertension and diabetes, race, and statin medication usage were similar in the P-PRA and S-PRA groups studied. PRA Population Of the 144 total PRAs placed during the period, 63.2% (n ¼ 91) were actively receiving dialysis treatments with the remaining 36.8% (n ¼ 47) in various stages of renal failure but not in need of hemodialysis treatments. The total number of successful cannulation of all PRAs was 43% (n ¼ 63). The mean functional patency was 645.7 ± 592.4 days, and total duration was 840.8 days ± 609.6 standard deviation. At the time of P-PRA or S-PRA placement, patients were stratified into receiving or not receiving hemodialysis treatments. The breakdown of the S-PRA as compared with the P-PRA groups and the total number of patients receiving hemodialysis treatments are detailed in Figure 2. PRA Operative Findings Most PRAs were placed in the left extremity (n ¼ 93, 64.6%). The majority anastomosis type was end-to-
Fig. 2. Stratification of patients on hemodialysis and not on hemodialysis at the time of PRA into P-PRA and S-PRA placements.
side (n ¼ 126) with the remaining being side-to-side (n ¼ 18). Outflow vein was not consistently reported in the operative record; thus, this variable was not used in direct statistical comparison of the 2 groups. In looking at the revisions of the cannulated PRAs (n ¼ 62), we found 10 patients who underwent a ‘‘revision’’ to another access type (i.e., superficialization, ligation, patch angioplasty, anastomosis revision, and so forth). PRA Complications Examination of our P-PRA and S-PRA populations revealed a small rate of arterial steal in 5 patients (5.7%) and 1 patient (1.7%), respectively. Arm edema was only reported in 1 patient in our SPRA group for a rate of 1.1%. PRA Outcomes in Patients Receiving Hemodialysis Treatments The outcomes comparing P-PRA and S-PRA for patients receiving hemodialysis treatments are summarized in Table II. S-PRA versus P-PRA
924 Amendola et al.
Annals of Vascular Surgery
Table II. Patients receiving hemodialysis treatments Outcome
P-PRA (n ¼ 46)
S-PRA (n ¼ 45)
P
Successful cannulation Time to maturation (mean ± SD) Functional patency (mean ± SD) Total duration (mean ± SD) Interventions per access (mean ± SD)
15 patients (33%) 111.4 ± 59.1 days 405.4 ± 531.9 days 516.8 ± 547.2 days 1.1 ± 0.7
25 patients (55%) 104.0 ± 95.9 days 755.2 ± 661.2 days 859.5 ± 650.7 days 1.6 ± 1.0
0.0354a 0.3212b 0.0220b 0.0361b 0.2109b
a
Fisher’s exact test. Wilcoxon two-sample test.
b
Table III. Patients not receiving hemodialysis treatments Outcome
P-PRA (n ¼ 41)
S-PRA (n ¼ 12)
P
Successful cannulation Functional patency (mean ± SD) Total duration (mean ± SD) Interventions per access (mean ± SD)
18 Fistulas (43.9%) 672.7 ± 550.1 daysb 1,085 ± 543.6 days 2.3 ± 2.9
4 Fistulas (33.3%) 740.5 ± 445.5 days 837.3 ± 458.1 days 3.0 ± 2.0
0.7404a 0.7017b 0.5232b 0.3200b
a
Fisher’s exact test. Wilcoxon two-sample test.
b
Table IV. Previous AV access type constructed in extremities receiving S-PRA Outcome
Fail to mature (n ¼ 25)
Thrombosis (n ¼ 32)
P
Cimino fistula Other fistula type Arteriovenous graft Days from initial access to S-PRA (mean ± SD)
20 Fistulas (80%) 5 Fistulas (20%) Grafts (0%) 482.5 ± 522.91 days
13 Fistulas (40.6%) 10 Fistulas (31.2%) 9 Grafts (28.1%) 639.56 ± 850.21 days
0.0023a 0.9487b
a
Fisher’s exact test. Wilcoxon two-sample test.
b
population with hemodialysis treatments was shown to have improved successful cannulation, functional patency, and total duration. There was no difference in the time to maturation between the 2 groups. In addition, there was no difference in the endovascular interventions between the 2 groups (1.6 ± 1.0 interventions per access versus 1.1 ± 0.7 interventions per access; P ¼ 0.2109). PRA Outcomes in Patients Not Receiving Hemodialysis Treatments The outcomes comparing P-PRA and S-PRA for patients not yet receiving hemodialysis are summarized in Table III. Statistical significance could not be found between the P-PRA and S-PRA groups in terms of success of cannulation, functional patency, total duration, nor endovascular interventions per access. Maturation time could not be directly compared between the P-PRA and S-PRA groups because of varying levels of renal failure affecting the eventual time that a patient would present to a hemodialysis unit for cannulation attempts of his
fistula. This variability would act as a mitigating factor to confound maturation time. Multivariate Analysis Multivariate logistical regression with stepwise and user directed factors were undertaken on all PRAs placed to predict successful cannulation. Ipsilateral central venous catheter placement, contralateral central venous catheter placement, age, race, history of hypertension and diabetes, dialysis dependence, and race were not found to significantly affect successful cannulation. Subgroup Analysis Examination of previous access types of the 57 patients who underwent S-PRA is detailed in Table IV. There were no grafts in the failure to mature group of the total failed accesses before S-PRA placement. The time between the initial fistula or graft placement and the subsequent S-PRA placement was not statistically significant.
Vol. 29, No. 5, July 2015
DISCUSSION We set out at the start of our analysis to reveal a more accurate representation of the functional outcomes of the PRA. Primary patency and assisted primary patency have been reported for this fistula type previously.6,7 Previous PRA studies implied all enrolled patients had successful cannulation of their fistula although not strictly delineated. The reporting of patency of an arteriovenous fistula is important, however, the more relevant measure for successful dialysis access is needle cannulation. Measuring only patency and not cannulation could incorrectly imply an improved outcome with any fistula type. Our overall cannulation rate was an abysmal 43%; however, we believe this analysis of our patient population is a real-world true functional outcome as it relates to successful use of this fistula type. Because of the limitations of the data collected, we were not able to clearly delineate reasons for such a low cannulation rate. We recognize that our hemodialysis access practice has changed during the study period. Most dramatically, this has been seen with the shift toward increased fistula utilization based on national guidelines.3 In addition, we experienced a reduction in the placement of arteriovenous grafts and a commensurate increase in the number of arteriovenous fistula revisions. The proximal radial artery and Cimino fistula were a sizable percentage of our total access procedures in any given year. The demographics of the P-PRA and S-PRA were similar with the noted exception of the S-PRA group having a greater number of patients receiving hemodialysis treatments (n ¼ 45, 78.9%). This is an intuitive finding considering most patients undergoing a secondary fistula procedure often have either failed primary fistula placement and are currently receiving hemodialysis treatments or were previously successfully cannulated and went on to access failure. Patient’s comorbidities between P-PRA and S-PRA groups were found to be comparable in terms of diabetes to previous studies. Diabetes was found in 64.4% of our P-PRA population and 57.9% of our S-PRA with a combined total incidence of 61.8%. This is slightly less than the initial percentage Jennings found in his reintroduction of the PRA,7 however, comparable with his follow-up studies.8 Most of our population undergoing PRA placement were African American. Previous PRA experiences did not delineated ethnicities; however, others9,10 have suggested worst outcomes in African Americans undergoing a variety of arteriovenous fistula placements. This has been attributed to
Previous hemodialysis access improves outcomes of PRA 925
significantly smaller arm vein diameters in this population. Subgroup analysis of the ethnic differences in our population was not powered enough to uncover any differences. Hypertension was found to be more prevalent in our population at around 90% compared with other studies with the PRA. This factor has been implicated as a major cause of renal failure in African American patients.11 Analysis of our population in terms of statin medication usage failed to find any effect on cannulation, functional patency, nor total duration. Statin medication usage has been shown in an induced diabetic rat model to have positive effects in arteriovenous fistulas hemodynamics.12 We can draw only limited conclusions from our population in terms of statin medication usage considering that our data were not randomized and several statin medication types were used during the study period. Our S-PRA group appears to have increased success with cannulation compared with P-PRA patients for those receiving hemodialysis treatments. The S-PRA group had increased success in cannulation at 55% compared with P-PRA of 33%. The S-PRA also had increased functional and total duration compared with P-PRA. It is not clear why the secondary nature of the PRA after either a failed or failing access improved outcomes. Given that the previous access would potentially dilate the venous system in the same extremity thus making the outflow vein at a future date more likely for successful cannulation and improved patency. This hypothesis might explain the improved outcomes in most of our patients being African American and the recent assertion that this population has smaller arm vein diameters. With this thought in mind, we examined the subgroup of patients who had a previous access and went on to thrombose or remain patent but unable to be cannulated before the S-PRA placement. Our logistical analysis failed to uncover a statistically significant affect of the patency of the previous access. It is not clear if this was because of the small number of patients in this subgroup analysis. We did find failing to mature S-PRA were statistically more likely to have had a previous Cimino fistula in the same extremity. In addition, most of thrombosed S-PRAs had a previous arteriovenous graft placed. Our data set did not capture the amount of time that a previous access in the S-PRA group was patent. This amount of time that a previous access was patent could have a dose-related effect on the future success of S-PRA. Again, we can only speculate and leave this for possible future analysis of our data.
926 Amendola et al.
An admitted limitation of our study was of our 100% male population studied compared with 39e50% in other studies regarding the PRA.7,8 It has been conjectured that female patients on average would have smaller caliber arteries thus would have increased difficulties with arteriovenous fistula placements. This has been found to be the case with dedicated diameter measures13 and in meta-analysis.5 We cannot comment on this directly in our patient population considering it was 100% male; however, dismal cannulation rates in our population might indicate a possible gender bias against PRA placement in males. It is well known from meta-analysis that cephalic vein diameters greater than 2 mm at the wrist predict success in Cimino fistula placement.14 This analysis has not been extended to the PRA. As part of our practice, we did not routinely measure vein diameter pre, intra, or postoperatively for fistula placement. We instead relied on our intraoperative assessment of vein adequacy after regional block to determine what anatomic location we would place a fistula and which vein wound be selected for venous outflow. This clearly represents a bias in our approach; however, this was our practice during the time period studied. Future directions for our population will focus on increased enrollments into our study group. We believe this will increase our ability to analyze patients who are not currently receiving hemodialysis treatments in regard to PRA outcomes.
CONCLUSIONS The PRA used in our daily practice has yielded patency rates are much lower than the reported primary and secondary patency rates for this fistula type. Our additional reporting of successful cannulation is a real-world reality of this fistula type. It appears that from our retrospective analysis that the S-PRA has an increased success of cannulation, increased functional patency, and increased total duration when compared with a P-PRA in male patients receiving hemodialysis treatments. The exact mechanism for this finding is not clearly deducible from our study; however, the increased number of African American patients in our study compared with results with the PRA might indicate a possible subset population that have poorer outcomes with PRA access.
Annals of Vascular Surgery
Conclusions for patients not receiving hemodialysis treatments have yet to be drawn because of the underpowered nature of this subgroup.
The authors would like to acknowledge the unending devotion that Marcela Woogen-Fisher FNP shows to our hemodialysis patients on a daily basis.
REFERENCES 1. Coresh J, Astor BC, Greene T, et al. Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third National Health and Nutrition Examination Survey. Am J Kidney Dis 2003;41:1e12. 2. USRDS 2012 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, 2010 and 2011. 3. National Kidney Foundation Kidney Disease Outcomes and Quality Initiative (K/DOQI), 2006 update. National Kidney Foundation, Inc.. Accessed at, www.kidney.org; 2006. 4. Brescia MJ, Cimino JE, Appel K, et al. Chronic hemodialysis using venipuncture and a surgically created arteriovenous fistula. N Engl J Med 1966;275:1089e92. 5. Rooijens P, Tordoir J, Stijnen T, et al. Radiocephalic wrist arteriovenous fistula for hemodialysis: meta-analysis indicates a high primary failure rate. Eur J Vasc Endovasc Surg 2004;28:583e9. 6. Toledo-Pereyra LH, Kyriakides GK, Ma K, et al. Proximal radial artery-cephalic. Arch Surg 1977;112:226e7. 7. Jennings WC, Bruns SD. Proximal radial artery as inflow site for native arteriovenous fistula. J Am Coll Surg 2003;197:58e63. 8. Jennings WC. Creating arteriovenous fistulas in 132 consecutive patients exploiting the proximal radial artery arteriovenous fistula: reliable, safe, and simple forearm and upper arm hemodialysis access. Arch Surg 2006;141:27e32. 9. Ishaque B, Zayed MA, Miller J, et al. Ethnic differences in arm vein diameter and arteriovenous fistula creation rates in men undergoing hemodialysis access. J Vasc Surg 2012;56:424e31. discussion 431e2. 10. Biuckians A, Scott EC, Meier GH, et al. The natural history of autologous fistulas as first-time dialysis access in the KDOQI era. J Vasc Surg 2008;47:415e21. discussion 420e1. 11. Freedman BI. End-stage renal failure in African Americans: insights in kidney disease susceptibility. Nephrol Dial Transplant 2002;17:198e200. 12. Roan JN, Fang SY, Chang SW, et al. Rosuvastatin improves vascular function of arteriovenous fistula in a diabetic rat model. J Vasc Surg 2012;56:1381e9. 13. Caplin N, Sedlacek M, Teodorescu V, et al. Venous access: women are equal. Am J Kidney Dis 2003;41:429e32. 14. Glass C, Johansson M, DiGragio W, et al. A meta-analysis of preoperative duplex ultrasound vessel diameters for successful radiocephalic fistula placement. J Vasc Ultrasound 2009;33:65e9.
Background: The proximal radial artery fistula (PRA) has been established as an early viable surgical option for arteriovenous fistula creation. The overall assisted primary patency reported in the literature approaches 100% at 1 year. We hypothesize that this excellent patency does not represent a functional result when seen in light of successful cannulation and fistula utilization. Methods: We retrospectively queried our Veterans Administration Hospital operative database to identify 284 male patients who had 571 access procedures performed by a senior vascular surgeon attending (R.K.D.) from January 1, 2003, to December 31, 2008. Operative details, patient comorbidities, fistula maturation time (time to first cannulation), functional patency (date of access to abandonment, revision to another fistula type, conversion to a prosthetic graft, thrombosis of the fistula, conversion to peritoneal dialysis, renal transplant, or patient death), and total duration (creation of the fistula to the end of its functional patency) were collected and analyzed. Results: A total of 144 PRAs were placed during the study period. In all, 87 patients underwent primary proximal radial artery fistula (P-PRA) placement in a limb without previous access; 57 patients had a secondary proximal radial artery fistula (S-PRA) after a failed previous fistula or graft in the same limb. There were no differences between the 2 groups in terms of age, comorbidities, and operative details. A total of 91 patients (63.2%) were receiving hemodialysis at the time of P-PRA or S-PRA placement. Outcomes of P-PRA and S-PRA populations on hemodialysis were examined. There was increased cannulation success (33% vs. 55%; P ¼ 0.00354, Fisher’s exact test), functional patency (755.2 ± 661.2 days vs. 405.4 ± 531.9 days; P ¼ 0.0220, Wilcoxon two-sample test), and total duration (859.5 ± 650.7 days vs. 516.8 ± 547.2 days; P ¼ 0.0361, Wilcoxon two-sample test) of S-PRA over P-PRA. There was no difference in endovascular interventions between the 2 groups (1.6 ± 1.0 interventions per access versus 1.1 ± 0.7 interventions per access; P ¼ 0.2109, Wilcoxon two-sample test). Subgroup analysis (analysis of variance) of the S-PRA group indicated that a patent but failing previous access in the same arm was not superior in terms of successful cannulation, functional patency, or total duration when compared with a thrombosed previous access. Conclusions: The PRA remains a viable first access procedure undertaken at our institution. Compared with the reported 12-month assisted primary patency of this fistula type, we found a small percentage of PRAs actually being accessed for successful hemodialysis treatment. The S-PRA appears to have a significantly higher successful cannulation rate, functional
1
Department of Surgery, McGuire VA Medical Center, Richmond, VA. Division of Vascular Surgery, Virginia Commonwealth University Health System, Richmond, VA. 2
Correspondence to: Michael F. Amendola, MD, Department of Surgery, Hunter Holmes McGuire VA Medical Center, Surgical Services, 1201 Broad Rock Boulevard, Richmond, VA 23249, USA; E-mail: [email protected]
920
Ann Vasc Surg 2015; 29: 920–926 http://dx.doi.org/10.1016/j.avsg.2014.12.021 Published by Elsevier Inc. Manuscript received: August 18, 2014; manuscript accepted: December 24, 2014; published online: March 7, 2015.
Vol. 29, No. 5, July 2015
Previous hemodialysis access improves outcomes of PRA 921
patency, and total duration time when compared with the P-PRA in patients receiving hemodialysis treatments. The mechanism of these improved outcomes is not known; considering patency or thrombosis of a previous access in the S-PRA group did not predict future access success in the same extremity.
INTRODUCTION It has been estimated that approximately 19.2 million people in the United States are affected with some form of chronic renal disease.1 This poses an ever-mounting public health risk with a cost in excess of 40 billion dollars and includes an estimated 398,861 patients receiving hemodialysis treatments on a periodic basis.2 In spite of national awareness and emphasis for fistula placement, a large percentage of patients remain on hemodialysis via central venous catheterization.1 In 1997, the National Kidney Foundation Kidney Dialysis Outcomes Quality Initiative clinical practice guidelines recommended native arteriovenous fistulas as the best hemodialysis access in patients with end-stage renal disease.3 This was based on known negative outcomes of patients who received hemodialysis from central catheterization or via arteriovenous grafts. Creation of an arteriovenous fistula using the cephalic vein to the radial artery (Cimino Fistula) was first described in 1966.4 The Cimino fistula has become the preferred first line access site for most patients receiving hemodialysis treatments. This access type, however, has a known high incidence of primary failure because of early thrombosis or maturation failure.5 The proximal radial artery fistula (PRA) has been promoted as an alternative inflow site for arteriovenous fistula creation. This fistula type was first described in 1977 in 20 patients, half of whom were diagnosed with diabetes.6 This initial report used the cephalic vein as the outflow source of the fistula. There was variable follow-up in this study with a reported primary patency rate of 90%. The PRA was reintroduced in a seminal study by Bruns and Jennings in 2003.7 Their work detailed the creation of 73 PRAs in 71 patients. The overall assisted primary patency approached 100% at 1 year. This study was followed with an additional report of 105 PRAs, of which, 34 were reported as a secondary fistula procedure after a previously failed access in the same limb.8 Again, the overall assisted patency was excellent at 97%. In each of these reports, patency was defined as the percentage of fistulas that had flow at some future time point, not necessarily the percent of PRAs being successfully accessed for hemodialysis treatments. The outflow
vein was described as basilic, cephalic, median cubital, lateral antecubital, or perforating deep vein. Patency is an important measurement of arteriovenous fistula function; however, most would agree that a more accurate marker of a success should be the amount of time that the fistula is needle cannulated for hemodialysis treatments. This definition serves as a functional measure of the usefulness of the arteriovenous fistula in the treatment of renal failure. We hypothesize that previous access in the same extremity before the creation of the PRA will favorably influence outcomes secondary to dilation of the venous circuit in that limb with a planned future access.
METHODS Study Population We retrospectively queried our institutional operative database from January 1, 2003, to December 31, 2008, to identify all vascular access procedures that were performed by a senior vascular surgery attending (R.K.D.) with the assistance of either a second or third year general surgery resident. Primary proximal radial artery fistula (P-PRA) was defined as a fistula that was constructed in an extremity that did not have any prior fistula or graft in the same extremity. Secondary proximal radial artery fistula (S-PRA) was defined as PRA construction in an extremity that did have a prior fistula or graft either thrombosed or nonfunctional. Data Collection All cases were collected and stored in an electronic clinical database per institutional guidelines. The operative records, date of each surgery, side of operation, indications, anastomosis type, date and type of subsequent operative revisions, and date and type of subsequent interventional procedures were identified. The patient’s electronic medical record was also examined to extract the following variables: date of birth, date of death, central venous catheterization, history of diabetes mellitus and hypertension, gender, ethnicity, and statin medicine use. Hemodialysis treatment records were reviewed for date of first cannulation and date of access abandonment.
922 Amendola et al.
Annals of Vascular Surgery
All endovascular procedures were noted and captured in the study database. These included all fistulograms for diagnostic and interventional purposes. Ligations of tributary veins with a patent fistula and elevation of an existing fistula without revision of anastomosis were not considered conversion to another fistula type. Functional Assessment Successful cannulation was defined as accessing the fistula with 1 or 2 needles for a complete hemodialysis treatment. Time to maturation referred to the number of days from fistula creation to successful cannulation. Functional patency was defined as the number of days from successful cannulation to access abandonment, revision to another fistula type, conversion to a prosthetic graft, thrombosis of the fistula, conversion to peritoneal dialysis, renal transplant, or patient death. Total duration was the number of days from the creation of the fistula to the end of its functional patency. Censored end points for patency included patient death or a patent fistula at the end of the study period.
Fig. 1. Histogram of arteriovenous access types and total for each year studied. Cimino, Cimino fistula; AVG, arteriovenous graft; Revised AVG, revisions of arteriovenous grafts; Revised AVF, revisions of arteriovenous fistula.
Statistical Analysis All statistical analyses were performed using the Windows based SAS System version 9.3 (SAS Institute Inc., Cary, NC). All tests were 2-tailed with a 0.05 significance level. Continuous factors were compared using the Wilcoxon rank test. Discrete factors were compared using the Fisher’s exact test. Logistic regression was used to identify independent predictors of improved patency in S-PRA subgroup.
Procedural Details
Follow-up
PRA construction was routinely undertaken with regional block and local anesthesia for augmentation as needed. All patients had their venous anatomy examined after regional block and tourniquet application to delineate what type of fistula would be created or modified. Ultrasound was not used to preoperatively assess venous anatomy. Once sterile drape was established and the decision to undergo PRA was made, local exploration of the antecubital fossa was undertaken through a longitudinal incision. Outflow veins were identified at this time as either originating from the cephalic, basilica, or a combination thereof. The proximal radial artery was identified an in-flow source for the fistula. The outflow vein was then dissected from its subcutaneous tissue. The patient was then given systemic heparin (100 U/kg), after 3 min, the proximal radial artery was clamped and then anastomosed to the outflow vein in either a side-to-side (n ¼ 18) or end-to-side (n ¼ 126) manner. The anastomosis was undertaken with either 6-0 or 7-0 Gore-TexÒ suture (W. L. Gore and Associates Inc., Flagstaff, AZ). When possible and under direct vision, the first distal vein valve was dilated with a coronary dilator. Patency was confirmed with hand-held Doppler probe after the flow was restored.
All patients had routine follow-up at 2 weeks after PRA placement and then at the discretion of the examining surgeon. Patient outcomes were tracked until December 31, 2008, for a total maximum of 1,095 days. Institutional Board Review The experimental protocols and the waiver of informed consent were approved by the Institutional Review Board (IRB # 01836) at the McGuire Veterans’ Affairs Medical Center in Richmond, Virginia before the start of this study.
RESULTS Population as a Whole During the study period from January 1, 2003, to December 31, 2008, we identified 284 male patients who had 571 access procedures. A total of 144 (25.2%) of the procedures were PRAs. The remaining population underwent a variety of other fistulas including Cimino arteriovenous fistula placement. Revisions of existing fistulas and arteriovenous grafts were also common during the study period. A yearby-year breakdown as a percentage of the total access placed is depicted in Figure 1. The percentage of PRA
Vol. 29, No. 5, July 2015
Previous hemodialysis access improves outcomes of PRA 923
Table I. Patient demographics among 144 patients undergoing PRA construction Demographic
P-PRA (n ¼ 87)
S-PRA (n ¼ 57)
P
Hemodialysis treatment Age, years, mean ± SD Alive at study end Left sided Hypertension African American Caucasian Diabetes Statin use
52.8% (n ¼ 46) 69.7 ± 10.7 31.03% (n ¼ 27) 39.08% (n ¼ 53) 89.66% (n ¼ 78) 64.37% (n ¼ 56) 35.63% (n ¼ 31) 64.37% (n ¼ 56) 53.49% (n ¼ 46)
78.9% (n ¼ 45) 68.8 ± 11.2 31.58 (n ¼ 18) 29.82% (n ¼ 40) 92.98% (n ¼ 53) 66.67% (n ¼ 38) 29.82% (n ¼ 17) 57.89% (n ¼ 33) 38.60% (n ¼ 22)
0.0015a 0.7023b 1.0000a 0.2884a 0.5659a 0.3301c 0.3301c 0.4849a 0.0898a
a
Fisher’s exact test. Wilcoxon two-sample test. c Chi-squared analysis. b
placements varied during the period ranging from 14% to 34% in any given year. The percentage of grafts started at 24% of our total access procedures in 2003 and dropped to 7% in 2008. The percentage of revisions of arteriovenous fistulas increased dramatically from 17% in 2003 to 41% in 2008. Patient Demographic Information Patient demographic information was collected and compared between the P-PRA and S-PRA groups as detailed in Table I. In all, 52.8% (n ¼ 46) of the P-PRA population was receiving hemodialysis treatments compared with 78.9% (n ¼ 45) of the S-PRA with a significant P value of 0.001. Demographics including age, death rates, history of hypertension and diabetes, race, and statin medication usage were similar in the P-PRA and S-PRA groups studied. PRA Population Of the 144 total PRAs placed during the period, 63.2% (n ¼ 91) were actively receiving dialysis treatments with the remaining 36.8% (n ¼ 47) in various stages of renal failure but not in need of hemodialysis treatments. The total number of successful cannulation of all PRAs was 43% (n ¼ 63). The mean functional patency was 645.7 ± 592.4 days, and total duration was 840.8 days ± 609.6 standard deviation. At the time of P-PRA or S-PRA placement, patients were stratified into receiving or not receiving hemodialysis treatments. The breakdown of the S-PRA as compared with the P-PRA groups and the total number of patients receiving hemodialysis treatments are detailed in Figure 2. PRA Operative Findings Most PRAs were placed in the left extremity (n ¼ 93, 64.6%). The majority anastomosis type was end-to-
Fig. 2. Stratification of patients on hemodialysis and not on hemodialysis at the time of PRA into P-PRA and S-PRA placements.
side (n ¼ 126) with the remaining being side-to-side (n ¼ 18). Outflow vein was not consistently reported in the operative record; thus, this variable was not used in direct statistical comparison of the 2 groups. In looking at the revisions of the cannulated PRAs (n ¼ 62), we found 10 patients who underwent a ‘‘revision’’ to another access type (i.e., superficialization, ligation, patch angioplasty, anastomosis revision, and so forth). PRA Complications Examination of our P-PRA and S-PRA populations revealed a small rate of arterial steal in 5 patients (5.7%) and 1 patient (1.7%), respectively. Arm edema was only reported in 1 patient in our SPRA group for a rate of 1.1%. PRA Outcomes in Patients Receiving Hemodialysis Treatments The outcomes comparing P-PRA and S-PRA for patients receiving hemodialysis treatments are summarized in Table II. S-PRA versus P-PRA
924 Amendola et al.
Annals of Vascular Surgery
Table II. Patients receiving hemodialysis treatments Outcome
P-PRA (n ¼ 46)
S-PRA (n ¼ 45)
P
Successful cannulation Time to maturation (mean ± SD) Functional patency (mean ± SD) Total duration (mean ± SD) Interventions per access (mean ± SD)
15 patients (33%) 111.4 ± 59.1 days 405.4 ± 531.9 days 516.8 ± 547.2 days 1.1 ± 0.7
25 patients (55%) 104.0 ± 95.9 days 755.2 ± 661.2 days 859.5 ± 650.7 days 1.6 ± 1.0
0.0354a 0.3212b 0.0220b 0.0361b 0.2109b
a
Fisher’s exact test. Wilcoxon two-sample test.
b
Table III. Patients not receiving hemodialysis treatments Outcome
P-PRA (n ¼ 41)
S-PRA (n ¼ 12)
P
Successful cannulation Functional patency (mean ± SD) Total duration (mean ± SD) Interventions per access (mean ± SD)
18 Fistulas (43.9%) 672.7 ± 550.1 daysb 1,085 ± 543.6 days 2.3 ± 2.9
4 Fistulas (33.3%) 740.5 ± 445.5 days 837.3 ± 458.1 days 3.0 ± 2.0
0.7404a 0.7017b 0.5232b 0.3200b
a
Fisher’s exact test. Wilcoxon two-sample test.
b
Table IV. Previous AV access type constructed in extremities receiving S-PRA Outcome
Fail to mature (n ¼ 25)
Thrombosis (n ¼ 32)
P
Cimino fistula Other fistula type Arteriovenous graft Days from initial access to S-PRA (mean ± SD)
20 Fistulas (80%) 5 Fistulas (20%) Grafts (0%) 482.5 ± 522.91 days
13 Fistulas (40.6%) 10 Fistulas (31.2%) 9 Grafts (28.1%) 639.56 ± 850.21 days
0.0023a 0.9487b
a
Fisher’s exact test. Wilcoxon two-sample test.
b
population with hemodialysis treatments was shown to have improved successful cannulation, functional patency, and total duration. There was no difference in the time to maturation between the 2 groups. In addition, there was no difference in the endovascular interventions between the 2 groups (1.6 ± 1.0 interventions per access versus 1.1 ± 0.7 interventions per access; P ¼ 0.2109). PRA Outcomes in Patients Not Receiving Hemodialysis Treatments The outcomes comparing P-PRA and S-PRA for patients not yet receiving hemodialysis are summarized in Table III. Statistical significance could not be found between the P-PRA and S-PRA groups in terms of success of cannulation, functional patency, total duration, nor endovascular interventions per access. Maturation time could not be directly compared between the P-PRA and S-PRA groups because of varying levels of renal failure affecting the eventual time that a patient would present to a hemodialysis unit for cannulation attempts of his
fistula. This variability would act as a mitigating factor to confound maturation time. Multivariate Analysis Multivariate logistical regression with stepwise and user directed factors were undertaken on all PRAs placed to predict successful cannulation. Ipsilateral central venous catheter placement, contralateral central venous catheter placement, age, race, history of hypertension and diabetes, dialysis dependence, and race were not found to significantly affect successful cannulation. Subgroup Analysis Examination of previous access types of the 57 patients who underwent S-PRA is detailed in Table IV. There were no grafts in the failure to mature group of the total failed accesses before S-PRA placement. The time between the initial fistula or graft placement and the subsequent S-PRA placement was not statistically significant.
Vol. 29, No. 5, July 2015
DISCUSSION We set out at the start of our analysis to reveal a more accurate representation of the functional outcomes of the PRA. Primary patency and assisted primary patency have been reported for this fistula type previously.6,7 Previous PRA studies implied all enrolled patients had successful cannulation of their fistula although not strictly delineated. The reporting of patency of an arteriovenous fistula is important, however, the more relevant measure for successful dialysis access is needle cannulation. Measuring only patency and not cannulation could incorrectly imply an improved outcome with any fistula type. Our overall cannulation rate was an abysmal 43%; however, we believe this analysis of our patient population is a real-world true functional outcome as it relates to successful use of this fistula type. Because of the limitations of the data collected, we were not able to clearly delineate reasons for such a low cannulation rate. We recognize that our hemodialysis access practice has changed during the study period. Most dramatically, this has been seen with the shift toward increased fistula utilization based on national guidelines.3 In addition, we experienced a reduction in the placement of arteriovenous grafts and a commensurate increase in the number of arteriovenous fistula revisions. The proximal radial artery and Cimino fistula were a sizable percentage of our total access procedures in any given year. The demographics of the P-PRA and S-PRA were similar with the noted exception of the S-PRA group having a greater number of patients receiving hemodialysis treatments (n ¼ 45, 78.9%). This is an intuitive finding considering most patients undergoing a secondary fistula procedure often have either failed primary fistula placement and are currently receiving hemodialysis treatments or were previously successfully cannulated and went on to access failure. Patient’s comorbidities between P-PRA and S-PRA groups were found to be comparable in terms of diabetes to previous studies. Diabetes was found in 64.4% of our P-PRA population and 57.9% of our S-PRA with a combined total incidence of 61.8%. This is slightly less than the initial percentage Jennings found in his reintroduction of the PRA,7 however, comparable with his follow-up studies.8 Most of our population undergoing PRA placement were African American. Previous PRA experiences did not delineated ethnicities; however, others9,10 have suggested worst outcomes in African Americans undergoing a variety of arteriovenous fistula placements. This has been attributed to
Previous hemodialysis access improves outcomes of PRA 925
significantly smaller arm vein diameters in this population. Subgroup analysis of the ethnic differences in our population was not powered enough to uncover any differences. Hypertension was found to be more prevalent in our population at around 90% compared with other studies with the PRA. This factor has been implicated as a major cause of renal failure in African American patients.11 Analysis of our population in terms of statin medication usage failed to find any effect on cannulation, functional patency, nor total duration. Statin medication usage has been shown in an induced diabetic rat model to have positive effects in arteriovenous fistulas hemodynamics.12 We can draw only limited conclusions from our population in terms of statin medication usage considering that our data were not randomized and several statin medication types were used during the study period. Our S-PRA group appears to have increased success with cannulation compared with P-PRA patients for those receiving hemodialysis treatments. The S-PRA group had increased success in cannulation at 55% compared with P-PRA of 33%. The S-PRA also had increased functional and total duration compared with P-PRA. It is not clear why the secondary nature of the PRA after either a failed or failing access improved outcomes. Given that the previous access would potentially dilate the venous system in the same extremity thus making the outflow vein at a future date more likely for successful cannulation and improved patency. This hypothesis might explain the improved outcomes in most of our patients being African American and the recent assertion that this population has smaller arm vein diameters. With this thought in mind, we examined the subgroup of patients who had a previous access and went on to thrombose or remain patent but unable to be cannulated before the S-PRA placement. Our logistical analysis failed to uncover a statistically significant affect of the patency of the previous access. It is not clear if this was because of the small number of patients in this subgroup analysis. We did find failing to mature S-PRA were statistically more likely to have had a previous Cimino fistula in the same extremity. In addition, most of thrombosed S-PRAs had a previous arteriovenous graft placed. Our data set did not capture the amount of time that a previous access in the S-PRA group was patent. This amount of time that a previous access was patent could have a dose-related effect on the future success of S-PRA. Again, we can only speculate and leave this for possible future analysis of our data.
926 Amendola et al.
An admitted limitation of our study was of our 100% male population studied compared with 39e50% in other studies regarding the PRA.7,8 It has been conjectured that female patients on average would have smaller caliber arteries thus would have increased difficulties with arteriovenous fistula placements. This has been found to be the case with dedicated diameter measures13 and in meta-analysis.5 We cannot comment on this directly in our patient population considering it was 100% male; however, dismal cannulation rates in our population might indicate a possible gender bias against PRA placement in males. It is well known from meta-analysis that cephalic vein diameters greater than 2 mm at the wrist predict success in Cimino fistula placement.14 This analysis has not been extended to the PRA. As part of our practice, we did not routinely measure vein diameter pre, intra, or postoperatively for fistula placement. We instead relied on our intraoperative assessment of vein adequacy after regional block to determine what anatomic location we would place a fistula and which vein wound be selected for venous outflow. This clearly represents a bias in our approach; however, this was our practice during the time period studied. Future directions for our population will focus on increased enrollments into our study group. We believe this will increase our ability to analyze patients who are not currently receiving hemodialysis treatments in regard to PRA outcomes.
CONCLUSIONS The PRA used in our daily practice has yielded patency rates are much lower than the reported primary and secondary patency rates for this fistula type. Our additional reporting of successful cannulation is a real-world reality of this fistula type. It appears that from our retrospective analysis that the S-PRA has an increased success of cannulation, increased functional patency, and increased total duration when compared with a P-PRA in male patients receiving hemodialysis treatments. The exact mechanism for this finding is not clearly deducible from our study; however, the increased number of African American patients in our study compared with results with the PRA might indicate a possible subset population that have poorer outcomes with PRA access.
Annals of Vascular Surgery
Conclusions for patients not receiving hemodialysis treatments have yet to be drawn because of the underpowered nature of this subgroup.
The authors would like to acknowledge the unending devotion that Marcela Woogen-Fisher FNP shows to our hemodialysis patients on a daily basis.
REFERENCES 1. Coresh J, Astor BC, Greene T, et al. Prevalence of chronic kidney disease and decreased kidney function in the adult US population: Third National Health and Nutrition Examination Survey. Am J Kidney Dis 2003;41:1e12. 2. USRDS 2012 Annual Data Report: Atlas of Chronic Kidney Disease and End-Stage Renal Disease in the United States, National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD, 2010 and 2011. 3. National Kidney Foundation Kidney Disease Outcomes and Quality Initiative (K/DOQI), 2006 update. National Kidney Foundation, Inc.. Accessed at, www.kidney.org; 2006. 4. Brescia MJ, Cimino JE, Appel K, et al. Chronic hemodialysis using venipuncture and a surgically created arteriovenous fistula. N Engl J Med 1966;275:1089e92. 5. Rooijens P, Tordoir J, Stijnen T, et al. Radiocephalic wrist arteriovenous fistula for hemodialysis: meta-analysis indicates a high primary failure rate. Eur J Vasc Endovasc Surg 2004;28:583e9. 6. Toledo-Pereyra LH, Kyriakides GK, Ma K, et al. Proximal radial artery-cephalic. Arch Surg 1977;112:226e7. 7. Jennings WC, Bruns SD. Proximal radial artery as inflow site for native arteriovenous fistula. J Am Coll Surg 2003;197:58e63. 8. Jennings WC. Creating arteriovenous fistulas in 132 consecutive patients exploiting the proximal radial artery arteriovenous fistula: reliable, safe, and simple forearm and upper arm hemodialysis access. Arch Surg 2006;141:27e32. 9. Ishaque B, Zayed MA, Miller J, et al. Ethnic differences in arm vein diameter and arteriovenous fistula creation rates in men undergoing hemodialysis access. J Vasc Surg 2012;56:424e31. discussion 431e2. 10. Biuckians A, Scott EC, Meier GH, et al. The natural history of autologous fistulas as first-time dialysis access in the KDOQI era. J Vasc Surg 2008;47:415e21. discussion 420e1. 11. Freedman BI. End-stage renal failure in African Americans: insights in kidney disease susceptibility. Nephrol Dial Transplant 2002;17:198e200. 12. Roan JN, Fang SY, Chang SW, et al. Rosuvastatin improves vascular function of arteriovenous fistula in a diabetic rat model. J Vasc Surg 2012;56:1381e9. 13. Caplin N, Sedlacek M, Teodorescu V, et al. Venous access: women are equal. Am J Kidney Dis 2003;41:429e32. 14. Glass C, Johansson M, DiGragio W, et al. A meta-analysis of preoperative duplex ultrasound vessel diameters for successful radiocephalic fistula placement. J Vasc Ultrasound 2009;33:65e9.