A comparison between transposed brachiobasilic arteriovenous fistulas and prosthetic brachioaxillary access grafts for vascular access for hemodialysis

A comparison between transposed brachiobasilic arteriovenous fistulas and prosthetic brachioaxillary access grafts for vascular access for hemodialysis Andy Robert Weale, MRCS, Paul Bevis, MRCS, William D. Neary, MRCS, Paul A. Lear, MS, FRCS, and David C. Mitchell, MS, FRCS, Bristol, United Kingdom Objective: Patients requiring access for hemodialysis in whom radiocephalic or brachiocephalic arteriovenous fistulas cannot be formed or have failed present a significant clinical challenge. We compare outcomes in patients undergoing transposed brachiobasilic arteriovenous fistulas (BBAVF) with expanded polytetrafluoroethylene brachioaxillary access grafts in a single European center. Methods: We identified all patients undergoing a first upper limb tertiary-access procedure—that is, either BBAVF or brachioaxillary access graft for hemodialysis— between January 1, 2000, and December 31, 2005. The median follow-up was 18.1 months (interquartile range, 8.7-34.9 months). Successful use for dialysis, primary patency, secondary patency, and patient survival was assessed. Results: A total of 185 patients were identified; 71 had a BBAVF, and 114 had an access graft. The median age was 64.3 years (interquartile range, 50.7-74.4 years). The groups were well matched for age, sex, ethnicity, diabetes, and number of prior access procedures. Significantly fewer BBAVFs were successfully used for dialysis: 69.0% BBAVFs compared with 89.4% access grafts (P ⴝ .001; ␹2). One- and two-year primary patency rates were 45.3% and 40.0%, respectively, for BBAVF and were 56.4% and 43.2% for access grafts (P ⴝ .579; log rank). Furthermore, there was no significant difference in secondary patency between the two procedures (P ⴝ .868; log rank). We found that surgeons in training had no influence on the primary patency of either BBAVF or access grafts. However, infective complications necessitating an operation were significantly higher in the access graft group (6.2% vs 0%; P ⴝ .031; Fisher exact test). Conclusions: Although more difficult to establish, BBAVFs provide patency at least equivalent to that of brachioaxillary access grafts. However, infective complications are fewer in the BBAVF group. As such, we believe that BBAVF should be the first choice of the vascular access surgeon when fistulas using the cephalic vein are not possible or have failed. ( J Vasc Surg 2007;46:997-1004.)

Patients who are unable to undergo dialysis by using a radiocephalic or brachiocephalic arteriovenous fistula (AVF), because of either unsuitable anatomy or access failure, present a common challenge.1,2 If suitable, the basilic vein may be anastomosed to the brachial artery. To form an easy-to-needle autogenous brachiobasilic AVF (BBAVF), the vein may be elevated or transposed either at the time of fistula formation (stage 1) or after arterialization of the vein some weeks after the initial fistula formation (stage 2).3,4 Primary failure rates vary widely in the literature, with between 0%5,6 and 38%7 of fistulas not maturing to use. Primary patency rates range from 23%8 to 90%5,9 at 1 year, with secondary patency rates of 47% to 96%.5,7 Although minimally invasive techniques have been described,10 most patients require a long medial arm incision to isolate the basilic vein. Given this extensive dissection and the disappointing patency of BBAVF in some series, From the Department of Vascular and Transplant Surgery, Southmead Hospital, North Bristol NHS Trust, Westbury on Trym. Competition of interest: none. Presented in part at the Annual Scientific Meeting of the Vascular Society of Great Britian and Ireland, Edinburgh, UK, Nov 24, 2006. Reprint requests: Andy Robert Weale, MRCS, Department of Surgery, Southmead Hospital, North Bristol NHS Trust, Westbury on Trym, Bristol, UK (e-mail: [email protected]). 0741-5214/$32.00 Copyright © 2007 by The Society for Vascular Surgery. doi:10.1016/j.jvs.2007.07.023

there remains enthusiasm for the readily useable prosthetic access graft inserted in the forearm loop11 or the straight brachioaxillary configuration.12 Comparisons between isolated series of prosthetic access grafts and BBAVF inevitably suffer from center effects. To date, only four studies have compared BBAVF with upper limb prosthetic access grafts9,13-15 (Table I). Although the first two studies showed clear advantages in terms of patency over prosthetic grafts,9,13 the more recent studies have found little difference in patency when analyzed on an intention-to-treat basis.14,15 Oliver et al14 had to exclude the 21% of BBAVFs that failed primarily before a statistical difference in patency could be demonstrated over prosthetic grafts. Nevertheless, all studies have clearly demonstrated that autogenous BBAVFs have lower infection rates than prosthetic grafts. Of the four studies, three were from North American centers, and one was from Taiwan. Thus, no study has been published from a European center comparing the outcomes of BBAVF with upper limb prosthetic grafts. Given the different patterns of access between Europe and North America,16,17 we hypothesized that differences between the two procedures may be exaggerated in centers with higher overall autogenous access placement. As such, we embarked on this retrospective comparison of patency and complication rates between BBAVF and prosthetic brachioaxillary grafts in a single European center. 997

JOURNAL OF VASCULAR SURGERY November 2007

998 Weale et al

Table I. Studies comparing BBAVF and vascular access grafts

9

Coburn 1994 Rhode Island, USA Matsuura 199813 Georgia, USA Oliver 200114 North Carolina, USA Lee 200415 Taipei, Taiwan

Type

N

BBAVF PTFE BBAVF PTFE BBAVF PTFE BBAVF PTFE

59 47 30 68 59 80 20 50

1 year patency 90% (primary) 70% (primary)

2 year patency 86% (primary) 49% (primary) 70% (primary) 46% (primary)

64% (cumulative) 62% (cumulative) 70% (cumulative) 69% (cumulative)

PTFE, Polytetrafluoroethylene, BBAVF, brachiobasilic arteriovenous fistula.

METHODS Patients. We identified every patient undergoing his or her first surgical procedure for vascular access in our center between January 1, 2000, and December 31, 2005. Within this cohort we studied all patients who went on to require an access procedure in the upper arm (tertiary access; that is, BBAVF or a straight brachioaxillary prosthetic access graft). Thus, any patient who had undergone any surgical access procedure before January 1, 2000 was excluded from the analysis. Only the first tertiary access procedure in the arm was considered in the analysis. It should be noted that because of surgical preference, only a few (n ⫽ 4) forearm loop grafts11 were inserted during the time period; because this group is not large enough to perform a meaningful analysis, these patients were not considered. Cases were identified by using the hospital’s computerized renal patient records and cross-checked with operating room records. Patients’ age, sex, ethnic group, diabetic status, access history, and date first seen by a nephrologist were recorded. The main surgeon was identified by using the operative record. Operations. Before surgery all patients were assessed clinically by one of two experienced consultant vascular access surgeons (P.A.L. and D.C.M.). During the study, preoperative duplex scanning or venography was performed selectively. We identified all patients who had a preoperative duplex scan and recorded the minimum diameter of the basilic vein identified. All operations were performed with patients under general anesthesia. All BBAVFs were formed as described by Dagher et al,3 and most were a one-stage procedure. Briefly, a single long incision or incision with a skin bridge was made in the medial aspect of the arm and deepened to expose the basilic vein from antecubital fossa to axilla. Tributaries were ligated, and the vein was isolated. Once free, it was disconnected in the antecubital fossa. The vein was then tunneled subcutaneously and anastomosed end to side to the brachial artery with 6-0 Prolene (Ethicon, Inc, Somerville, NJ), in or just proximal to the antecubital fossa. No completion angiography was performed. Brachioaxillary access grafts required exposure of the axillary vein and brachial artery via small incisions in the axilla and antecubital fossa, respectively. Six-millimeter expanded polytetra-

fluoroethylene was tunneled subcutaneously in a straight configuration, and the graft was anastomosed to the vein and artery end to side by using 6-0 Prolene. Routine postoperative anticoagulation was not used. Postoperative duplex investigation and surveillance were performed only in patients in whom fistulas or grafts failed to mature or in whom dialysis was poor. Outcome. Renal patient records were accessed in October 2006 to identify the outcome of each fistula. Median follow up was 18.1 months (interquartile range [IQR], 8.7-34.9 months). A successful fistula/graft was defined as a fistula or graft that had been successfully needled for dialysis. The presence or absence of a thrill was not used to define technical success, because the simple presence of a thrill does not determine whether a fistula will mature to become usable or can be used for dialysis.18 Primary functional patency was defined according to the recommended reporting standards25—ie, the percentage of all BBAVFs/ grafts attempted that were still being used for dialysis that had not required any intervention designed to maintain or re-establish patency at 1 and 2 years.19 Interventions included angioplasty, thrombectomy, and surgical revision of the fistula. Secondary functional patency was defined as the percentage of all BBAVF/grafts attempted that were being used irrespective of interventions to re-establish or maintain patency at 1 and 2 years. The time to first use and the dates of death or transplantation were recorded as relevant. Patients who underwent transplantation, died, or never proceeded to long-term hemodialysis were considered as lost to follow-up and were censored for all survival analysis. Differences in patient survival were also investigated. Infections necessitating operative intervention (incision and drainage of collection related to graft or BBAVF) or loss or excision of access were identified. Operations performed for steal syndrome were also noted (access ligation or distal revascularization and interval ligation20 procedure). Statistical analysis. Patient demographic factors were compared by using the Mann-Whitney U test for continuous variables and ␹2 tests for categorical values. For fistula success, proportions were compared by using a ␹2 test. Fistula patency, fistula survival, and patient survival were analyzed by using life tables and Kaplan-Meier survival analysis. The two groups were compared by using a log-

JOURNAL OF VASCULAR SURGERY Volume 46, Number 5

Weale et al 999

Table II. Patient demographics

Age Females Diabetes Ethnic group: Caucasian Black Asian Chinese Previous operations (median (IQR)) Previous successful dialysis using either a radio-cephalic or brachiocephalic fisutla First seen by nephrologist ⬍3 months prior commencing hemodialysis

Access graft (n ⫽ 114)

BBAVF (n ⫽ 71)

P value

66.3 (52.9-74.5) 64 (56.1%) 50 (43.9%)

61.5 (48.6-73.0) 38 (53.5%) 33 (46.5%)

0.225 (Mann-Whitney) 0.728 (␹2) 0.727 (␹2) 0.787 (␹2)

101 (88.6%) 2 (1.7%) 10 (8.8%) 1 (0.9%) 1 (0-2)

66 (93.0%) 1 (1.4%) 4 (5.6%) 0 1 (0-1)

0.904 (␹2)

34 (29.8%)

16 (22.5%)

0.262 (␹2)

33 (28.9%)

13 (18.3%)

0.104 (␹2)

BBAVF, Brachiobasilic arteriovenous fistula; IQR, interquartile range.

rank test. Factors (age, sex, diabetes, number of prior access attempts, prior functioning AVF, and grade of surgeon) potentially influencing the primary patency of BBAVF and access grafts were analyzed with a Cox proportional hazards regression analysis. In patients who had a preoperative duplex scan, we assessed the effect of minimum basilic vein diameter on BBAVF primary patency. Statistical analysis was performed with SPSS (version 12.0 for Windows; SPSS Inc, Chicago, Ill). A P value of ⬍.05 was considered statistically significant RESULTS During the study period, 716 patients underwent 1007 procedures to create vascular access for hemodialysis. A total of 361 patients had a radiocephalic fistula formed, and 297 had a brachiocephalic fistula formed as the first access procedure. Twenty (2.8%) had a BBAVF, and 38 (5.3%) had a vascular access graft as the primary procedure. Patient demographics are shown in Table II. Of the 658 patients who had a radiocephalic or brachiocephalic fistula as the primary surgical access procedure during the study period, 126 patients required either a BBAVF (n ⫽ 51) or a brachioaxillary access graft (n ⫽ 76). Thus we were able to compare 71 BBAVF with 114 brachioaxillary access grafts. It is worthy of note that during the study, 253 patients underwent 319 tertiary access procedures (103 BBAVFs and 216 access grafts). Sixty-eight patients who had undergone a surgical access procedure before 2000 where excluded. Thirty-four of the 253 had more than one tertiary access procedure; only the first access procedure was considered. Thus, although the comparison is not a consecutive series of all access grafts and BBAVFs, it considers only the first tertiary access procedure and does not suffer from double inclusion of patients. Significantly fewer BBAVFs were ever used for dialysis than access grafts; 49 (69.0%) BBAVFs compared with 102 (89.4%) access grafts (P ⫽ .001; ␹2). Of the 22 patients with a BBAVF that was never used for dialysis, 4 did not actually ever proceed to dialysis. Although a good bruit was

documented in all of these patients, we have not classified these patients as having a successful fistula, because it is well recognized that the presence of a thrill does not determine whether a fistula can be used for dialysis.18 Thrombosis occurred in 18 (25.3%) BBAVFs before dialysis: 6 patients had a BBAVF that failed within 30 days of the procedure, whereas of the 12 BBAVFs that thrombosed later without being successfully used for dialysis, only 2 patients presented to the hospital for an attempted thromboembolectomy. One of these had a procedure that successfully restored blood flow. Of the 12 patients who had an access graft that was never needled, only 1 did not actually proceed to dialysis. Two patients required removal of the graft within the first week as a result of steal, whereas a further five patients had graft thrombosis within 30 days. Three of these patients attended the hospital and underwent successful thromboembolectomy. However, reclotting of these grafts occurred within weeks, and further access creation was required in all cases. Of the four patients with failure after 30 days and nonuse, three had graft revisions before thrombosis and loss of access. The median time to first use for BBAVF was 71 days (IQR, 45-107 days). The median time to first use for access grafts was significantly shorter (P ⬍ .001; Mann-Whitney U test), at 14 days (IQR, 13-14 days). The primary functional patency of BBAVF was 45.3% at 1 year and 40.0% at 2 years. The 1- and 2-year primary patency rates of access grafts were 56.4% and 43.2%, respectively. There was no difference in primary functional patency between BBAVF and access grafts (P ⫽ .579; logrank test), as illustrated in the survival curve in Fig 1. The secondary functional patency of BBAVF was 53.6% at 1 year and 50.9% at 2 years. Fifteen patients with a BBAVF underwent surgical revision (n ⫽ 4), thromboembolectomy (n ⫽ 4), angioplasty (n ⫽ 6), or both surgical and radiologic intervention (n ⫽ 1) to maintain patency of the fistula. One- and two-year secondary patency rates of access grafts were 61.7% and 41.1%, respectively. Twenty-four patients with an access graft underwent surgical revision (n ⫽ 5), thromboembolectomy (n ⫽ 15), angioplasty (n ⫽ 1),

JOURNAL OF VASCULAR SURGERY November 2007

1000 Weale et al

100%

100%

80%

80%

Access survival

Primary patency

BBAVF

60%

BBAVF 40%

p=0.579

60%

p=0.2853

Access graft

40%

Access graft 20%

20%

0%

0% 0

6

12

18

24

30

36

42

48

54

60

66

72

Months following creation of access

At risk BBAVF

71

43

22

15

10

5

5

5

2

Access graft

114

60

33

23

12

8

4

3

3

2

1

Fig 1. Primary patency of brachiobasilic arteriovenous fistula (BBAVF) and polytetrafluoroethylene access grafts. The dashed bar illustrates the point at which the Kaplan-Meier curve becomes unreliable. 100%

Secondary patency

80%

60%

BBAVF

p=0.9553

40%

Access graft

20%

0% 0

6

12

18

24

30

36

42

48

54

60

66

72

Months following creation of access

At risk BBAVF

71

48

24

16

12

6

6

5

2

Access graft

114

66

38

24

12

8

5

4

4

3

1

Fig 2. Secondary patency of brachiobasilic arteriovenous fistula (BBAVF) and polytetrafluoroethylene access grafts. The dashed bar illustrates the point at which the Kaplan-Meier curve becomes unreliable.

or both surgery and angioplasty (n ⫽ 3) to maintain patency. There was no significant difference between the need for interventions (P ⫽ .990; ␹2), the number of interventions (P ⫽ .962; Mann-Whitney), or overall secondary functional patency (Fig 2; P ⫽ .868; log-rank test) between the two procedures. Even when primary failures were excluded and only the patency of access that had gone on to function (needled for dialysis) was considered, there remained no difference between the two access procedures in terms of access survival (Fig 3). We observed that age, female sex, late referral to nephrology (dialysis within 3 months of first seeing a specialist), diabetes, number of previous access procedures, previous successful dialysis with a fistula, and grade of operator (surgeon in training) had no effect on loss of primary patency of either access grafts (Table III) or BBAVF (Table IV).

0

6

12

18

24

30

36

42

48

54

60

66

72

Months following first use of access

At risk BBAVF

49

31

19

13

8

5

5

4

1

Access graft

101

60

36

22

10

8

5

4

4

2

1

Fig 3. Access survival of brachiobasilic arteriovenous fistula (BBAVF) and polytetrafluoroethylene access grafts. The dashed bar illustrates the point at which the Kaplan-Meier curve becomes unreliable.

Just less than 60% (n ⫽ 43) of the BBAVF group underwent preoperative imaging of the upper limb with duplex scanning before undergoing the procedure. Although this incomplete subgroup should be interpreted with caution, the median minimum measure size of the basilic vein on duplex scanning in the 26 BBAVFs that were successfully needled was 3.3 mm (IQR, 2.5-3.8 mm), compared with 3.1 mm (IQR, 2.6-3.3 mm) in the 17 that were never used (P ⫽ .456; Mann-Whitney U test). Eleven BBAVFs were attempted in patients with a minimum measured vein diameter of 2.5 mm or less; in seven (63.6%) of these patients, the fistula successfully matured to use. Of the 32 patients who had a basilic vein measured at greater than 2.5 mm, 19 (59.4%) veins were successfully used (P ⫽ .803; ␹2). Although the numbers are small and the results should be interpreted cautiously, Fig 4 shows that there is no difference in primary patency between BBAVF created when vein diameters are smaller than or larger than 2.5 mm (P ⫽ .9631; log rank). Infective complications necessitating operative intervention were more common (P ⫽ .031; Fisher exact test) with access grafts (n ⫽ 7 removed) than BBAVFs (n ⫽ 0). The incidence of steal syndrome was low in this series, with no statistical difference in steal syndrome necessitating intervention and with 2 grafts removed and no BBAVF necessitating intervention (P ⫽ .378; Fisher exact test). One BBAVF was tied off because it was implicated in the patient’s cardiac failure; however, the patient’s symptoms did not improve with fistula ligation. Patients who had undergone a BBAVF had a better survival rate than patients who only had an access graft inserted (P ⫽ .001; log rank). This is shown in Fig 5. DISCUSSION This study is important in the context of changing patterns of vascular access17 and in comparing outcomes from upper arm prosthetic access grafts with transposed

JOURNAL OF VASCULAR SURGERY Volume 46, Number 5

Weale et al 1001

Table 3. Impact of factors on loss of primary patency of brachio-axillary access grafts (Cox regression analysis)

Age Number of access procedures Diabetes Female sex Dialysis within 3 months of referral Previous dialysis with a fistula Non-consultant operator

Exp(B)

95% CI

P value

0.989 1.112 0.839 0.782 1.461 0.704 1.656

0.973-1.005 0.752-1.645 0.463-1.523 0.435-1.408 0.758-2.817 0.386-1.287 0.943-2.910

0.190 0.595 0.568 0.412 0.257 0.254 0.079

Table 4. Impact of factors on loss of primary patency of BBAVF (Cox regression analysis)

Age Number of access procedures Diabetes Female sex Dialysis within 3 months of referral Previous dialysis with a fistula Non-consultant operator

Exp(B)

95% CI

P value

0.997 0.792 1.498 0.739 0.921 0.707 1.356

0.975-1.020 0.525-1.194 0.777-2.891 0.384-1.423 0.375-2.259 0.296-1.688 0.661-2.784

0.787 0.265 0.228 0.366 0.857 0.436 0.406

BBAVF, Brachiobasilic arteriovenous fistula.

80%

80%

Primary patency

100%

Patient survival

100%

60%

BBAVF 40%

60%

<2.5mm

40%

p=0.9631

>2.5mm

p=0.001 20%

Access graft

20%

0%

0%

0

0

6

12

18

24

30

36

BBAVF Access graft

42

48

54

60

66

72

Months following creation of access

At risk 71 114

65 85

54 61

44 44

33 35

26 25

23 19

14 14

11 8

4 6

4

2

6

12

18

24

30

36

42

48

54

60

Months following creation of access

At risk Greater than 2.5mm

32

19

8

6

Less than 2.5mm

11

5

3

1

4

1

1

1

1

2

Fig 4. Patient survival of patients with a brachiobasilic arteriovenous fistula (BBAVF) and polytetrafluoroethylene access grafts. The dashed bar illustrates the point at which the Kaplan-Meier curve becomes unreliable.

autogenous fistulas in a European center, and it describes one of the largest series of BBAVFs in the literature to date.21 We found that primary and secondary functional patency is equivalent between the BBAVF and prosthetic access grafts. This is in keeping with the two most recent comparisons.14,15 The primary functional patency of BBAVF at 1 year of 45.3% is at the lower end of that found in the worldwide literature (Table V), but it is similar to reports published within the last 5 years.7,8,15,22,23 It is interesting to note that the patency rates of BBAVF are very similar to the patency rates of radiocephalic and brachiocephalic fistulas we have previously published from our center.24 The patency rates were nonetheless disappointing,

Fig 5. Effect of basilica vein size on the primary patency of brachiobasilic arteriovenous fistulas. The dashed bar illustrates the point at which the Kaplan-Meier curve becomes unreliable.

and we require changes to our access service in light of new evidence. Early failure rates in both grafts and BBAVF were also higher than expected. This may relate to the low use of preoperative imaging. We do not believe that the use of general anesthesia has had a detrimental effect on early outcome. Although it is recognized that both BBAVFs and access grafts can be inserted by using regional blockade and, indeed, that this may be preferable in terms of patient risk and venodilatation,25 there is little evidence that regional blockade improves patency in tertiary access. When access procedures that did not result in a functioning graft or fistula were excluded, we did not find that BBAVFs provided greater longevity, although this may relate to the follow-up time.

JOURNAL OF VASCULAR SURGERY November 2007

1002 Weale et al

Table V. Literature review of results of BBAVF Author Hatjibaloglou39 Rivers40 Coburn9 Hakaim6 Matsuura13 El Mallah41 Humphries42 Murphy43 Oliver14 Ascher30 Tsai5 Taghizadeh44 Segal23 Hossny4 Lee15 Rao7 Hill10 Keuter22 Wolford8

Year of publication

Number of BBAVF

1992 1993 1994 1998 1998 1998 1999 2000 2001 2001 2002 2003 2003 2003 2004 2004 2005 2005 2005

22 65 59 26 30 40 67 74 59 63 54 75 99 70 20 56 32 31 100

Failure to mature or technical failure

1 year primary patency

1 year secondary patency 81% 58%

90% 0% 0% 10%-40% 32% 21% 9% 0% 8% 23% 5.7%

79% 85% 50-80% 84% 73% 64% 69% 89%

95.9% 66%

47% 84.2%-90%

38%

80% 35%

47%

3% 21%

58% 23%

90% 47%

BBAVF, Brachiobasilic arteriovenous fistula.

During the study we did not assess fistula maturation in the early postoperative course to identify failing AVFs, which may have been salvageable by secondary procedures.26 Recent randomized studies, published after the inclusion dates of this study have also suggested that surveillance may improve the patency of both access grafts and autogenous fistulas.27,28 We have modified our practice and now institute a 2-week review with a vascular access nurse specialist followed by duplex surveillance. As such, we anticipate that with increased familiarity with the technique and a rigorous postoperative assessment before and during dialysis, failures can be reduced and results significantly improved. This is a retrospective series, within which selection for either procedure was influenced by surgeon preference, anatomic suitability, and the urgency for access for dialysis due to late referral or access failure. Relative to many centers within the United Kingdom, we have a high prosthetic graft use (5% of primary access in this cohort and 14% of prevalent dialysis population), thus reflecting a preference for definitive surgically created access to long-term dialysis catheters.29 In such a retrospective series, it is difficult to objectively measure urgency and thus assess the effect that may have on choosing a particular access procedure. Furthermore, not all patients are anatomically suitable for BBAVF; however, within this series, approximately 40% of patients were deemed appropriate to attempt the procedure. Most studies to date have defined that the basilic vein should be at least 3 mm8,30,31 or 4 mm15 in diameter before proceeding to BBAVF; however, the evidence for this rationale is limited.1,32 In the subgroup of patients who had a preoperative duplex scan in this study, we found that even preoperative vein diameters of less than 2.5 mm produced proportions of functional BBAVF equivalent to those in patients with larger veins.

Another important finding within our study is the effect of surgeons in training on outcomes in tertiary access. Although it has been suggested that the surgeon has a major effect on the outcome of AVF formation,33 we have previously shown that trainee surgeons can attain outcomes as good as their trainers in primary access surgery.24 In this study we found that in both the BBAVF group and access graft groups, surgeons in training did not have a statistically worse outcome than consultants. However, we acknowledge that selection of appropriate cases for the trainee is required to obtain such results. We have no explicit guidelines to define cases suitable for a trainee in our center, but we do not doubt that selection bias may undermine our findings. We strongly believe than vascular access surgeons should be encouraged to involve surgeons in training in this expanding specialty. Similar to the four other studies9,13-15 comparing the two access procedures, we found significant differences in infective complications, with no BBAVF lost to infection, compared with 6% of access grafts. Avoiding such serious complications and associated patient morbidity provides a forceful argument in favor of autogenous access. However, we found no differences in the rates of intervention for steal syndrome or cardiac failure in our series, and the reported prevalence of these complications was low.14 We found that patients who had a BBAVF had an improved survival compared with those who had undergone insertion of an access graft. One explanation is that we may be inserting access grafts into patients whom we do not expect to have lengthy survival. However, this finding is in keeping with previously published data suggesting that patients with autogenous fistulas have a lower mortality than those with prosthetic grafts.34,35 The possible reasons for this are complex but undoubtedly include morbidity associated with access-related complications and differ-

JOURNAL OF VASCULAR SURGERY Volume 46, Number 5

ences in the patient population in terms of age and the urgency for need for dialysis access. Patients urgently requiring access for dialysis due to acute or late referral or due to access failure are generally more unwell. To avoid lengthy use of central venous dialysis catheters, these patients in our unit often have an access graft inserted to allow a short lead-in time to dialysis via a surgically created access. However, in this study we found no statistical difference in age, in the proportion of patients in each group who were referred late, or who failed to establish dialysis on a radiocephalic or brachiocephalic fistula before having tertiary access. Like the four previous studies, the results of an isolated single-center retrospective series may not be viewed as representative. The lack of randomized controlled trials in vascular access surgery is often noted.36 However, some clinicians may be of the opinion that that a randomized study comparing BBAVF and brachioaxillary access grafts is unnecessary given the available nonrandomized controlled trial data relating to infective complications. We hypothesized we may find greater differences between BBAVF and access grafts than previous series because of the high use of autogenous access in our center. The fact that we have found that in terms of 1-year primary and secondary patency, access grafts appear slightly better, although not statistically significantly so, is worthy of further discussion. The unit policy has always been to base access placement on clinical assessment, with little use of duplex scanning. In this series, of those who went on to need tertiary access, approximately a quarter had been established on dialysis by using a wrist or brachiocephalic fistula (Table II). Indeed, with placing radiocephalic and brachiocephalic AVFs, a policy of selective duplex scanning is supported by the literature.37 However, for patients requiring tertiary access, we have persevered with selective use of duplex scanning, on the basis of surgical preference at the time of assessment. As such, our BBAVF outcomes may be worse than expected because of differences in case selection and worse than many of those series from North America, where autogenous access placement is traditionally low,17 and rely heavily on preoperative imaging.38 As previously discussed, we have subsequently changed our unit policy and await the effect on our outcomes. CONCLUSION We have shown that BBAVFs provide patency statistically equivalent to brachioaxillary access grafts. However, infective complications are less and patient survival seems greater in the BBAVF group. As such, we believe that BBAVFs should be the first choice of the vascular access surgeon when fistulas using the cephalic vein are not possible or have failed. AUTHOR CONTRIBUTIONS Conception and design: ARW, PAL, DCM Analysis and interpretation: ARW Data collection: ARW, PB, WDN Writing the article: ARW

Weale et al 1003

Critical revision of the article: ARW, PAL, DCM Final approval of the article: ARW, PB, WDN, PAL, DCM Statistical analysis: ARW Overall responsibility: ARW REFERENCES 1. Allon M, Lockhart ME, Lilly RZ, Gallichio MH, Young CJ, Barker J, et al. Effect of preoperative sonographic mapping on vascular access outcomes in hemodialysis patients. Kidney Int 2001;60:2013-20. 2. Rooijens PP, Tordoir JH, Stijnen T, Burgmans JP, Smet de AA, Yo TI. Radiocephalic wrist arteriovenous fistula for hemodialysis: meta-analysis indicates a high primary failure rate. Eur J Vasc Endovasc Surg 2004; 28:583-9. 3. Dagher FJ, Gelber RL, Ramos EJ, Sadler JH. Basilic vein to brachial artery fistula: a new access for chronic hemodialysis. South Med J 1976;69:1438-40. 4. Hossny A. Brachiobasilic arteriovenous fistula: different surgical techniques and their effects on fistula patency and dialysis-related complications. J Vasc Surg 2003;37:821-6. 5. Tsai YT, Lin SH, Lee GC, Huen GG, Lin YF, Tsai CS. Arteriovenous fistula using transposed basilic vein in chronic hypotensive hemodialysis patients. Clin Nephrol 2002;57:376-80. 6. Hakaim AG, Nalbandian M, Scott T. Superior maturation and patency of primary brachiocephalic and transposed basilic vein arteriovenous fistulae in patients with diabetes. J Vasc Surg 1998;27:154-7. 7. Rao RK, Azin GD, Hood DB, Rowe VL, Kohl RD, Katz SG, et al. Basilic vein transposition fistula: a good option for maintaining hemodialysis access site options? J Vasc Surg 2004;39:1043-7. 8. Wolford HY, Hsu J, Rhodes JM, Shortell CK, Davies MG, Bakhru A, et al. Outcome after autogenous brachial-basilic upper arm transpositions in the post-National Kidney Foundation Dialysis Outcomes Quality Initiative era. J Vasc Surg 2005;42:951-6. 9. Coburn MC, Carney WI Jr. Comparison of basilic vein and polytetrafluoroethylene for brachial arteriovenous fistula. J Vasc Surg 1994;20: 896-902. 10. Hill BB, Chan AK, Faruqi RM, Arko FR, Zarins CK, Fogarty TJ. Keyhole technique for autologous brachiobasilic transposition arteriovenous fistula. J Vasc Surg 2005;42:945-50. 11. Rooijens PP, Burgmans JP, Yo TI, Hop WC, de Smet AA, van den Dorpel MA, et al. Autogenous radial-cephalic or prosthetic brachialantecubital forearm loop AVF in patients with compromised vessels? A randomized, multicenter study of the patency of primary hemodialysis access. J Vasc Surg 2005;42:481-6. 12. Huber TS, Carter JW, Carter RL, Seeger JM. Patency of autogenous and polytetrafluoroethylene upper extremity arteriovenous hemodialysis accesses: a systematic review. J Vasc Surg 2003;38:1005-11. 13. Matsuura JH, Rosenthal D, Clark M, Shuler FW, Kirby L, Shotwell M, et al. Transposed basilic vein versus polytetrafluorethylene for brachialaxillary arteriovenous fistulas. Am J Surg 1998;1998:219-21. 14. Oliver MJ, Mccann RL, Indridason OS, Butterly DW, Schwab SJ. Comparison of transposed brachiobasilic fistulas to upper arm grafts and brachiocephalic fistulas. Kidney Int 2001;60:1532-9. 15. Lee CH, Ko PJ, Liu YH, Hsieh HC, Liu HP. Brachiobasilic fistula as a secondary access procedure: an alternative to a dialysis prosthetic graft. Chang Gung Med J 2004;27:816-23. 16. Pisoni RL, Young EW, Dykstra DM, Greenwood RN, Hecking E, Gillespie B, et al. Vascular access use in Europe and the United States: results from the DOPPS. Kidney Int 2002;61:305-16. 17. Rayner HC, Besarab A, Brown WW, Disney A, Saito A, Pisoni RL. Vascular access results from the Dialysis Outcomes and Practice Patterns Study (DOPPS): performance against Kidney Disease Outcomes Quality Initiative (K/DOQI) Clinical Practice Guidelines. Am J Kidney Dis 2004;44:22-6. 18. Won T, Jang JW, Lee S, Han JJ, Park YS, Ahn JH. Effects of intraoperative blood flow on the early patency of radiocephalic fistulas. Ann Vasc Surg 2000;14:468-72. 19. Sidawy AN, Gray R, Besarab A, Henry M, Ascher E, Silva M Jr, et al. Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg 2002;35:603-10.

1004 Weale et al

20. Berman SS, Gentile AT, Glickman MH, Mills JL, Hurwitz RL, Westerband A, et al. Distal revascularization-interval ligation for limb salvage and maintenance of dialysis access in ischemic steal syndrome. J Vasc Surg 1997;26:393-402. 21. El Sayed HF, Mendoza B, Meier GH, LeSar CJ, DeMasi RJ, Glickman MH, et al. Utility of basilic vein transposition for dialysis access. Vascular 2005;13:268-74. 22. Keuter XH, van der Sande FM, Kessels AG, de Haan MW, Hoeks AP, Tordoir JH. Excellent performance of one-stage brachial-basilic arteriovenous fistula. Nephrol Dial Transplant 2005;20:2168-71. 23. Segal JH, Kayler LK, Henke P, Merion RM, Leavey S, Campbell DA Jr. Vascular access outcomes using the transposed basilic vein arteriovenous fistula. Am J Kidney Dis 2003;42:151-7. 24. Weale AR, Barwell J, Chant H, Lear PA, Mitchell DC. The impact of training on outcomes in primary vascular access surgery. Ann R Coll Surg Engl 2004;86:275-80. 25. Hingorani AP, Ascher E, Gupta P, Alam S, Marks N, Schutzer RW, et al. Regional anesthesia: preferred technique for venodilatation in the creation of upper extremity arteriovenous fistulae. Vascular 2006;14: 23-6. 26. Berman SS, Gentile AT. Impact of secondary procedures in autogenous arteriovenous fistula maturation and maintenance. J Vasc Surg 2001; 34:866-71. 27. Malik J, Slavikova M, Svobodova J, Tuka V. Regular ultrasonographic screening significantly prolongs patency of PTFE grafts. Kidney Int 2005;67:1554-8. 28. Tessitore N, Lipari G, Poli A, Bedogna V, Baggio E, Loschiavo C, et al. Can blood flow surveillance and pre-emptive repair of subclinical stenosis prolong the useful life of arteriovenous fistulae? A randomized controlled study. Nephrol Dial Transplant 2004;19:2325-33. 29. Ansell D, Feest D, Byrne C, Ahmed A. The Renal Association: UK Renal Registry (the eighth annual report). 2005 http://www.renalreg.com/ Report%202005/Cover_Frame2.htm. 30. Ascher E, Hingoran A, Gunduz Y, Yorkovich Y, Ward M, Miranda J, et al. The value and limitations of the arm cephalic and basilic vein for arteriovenous access. Ann Vasc Surg 2001;15:89-97. 31. Gormus N, Ozergin U, Durgut K, Yuksek T, Solak H. Comparison of autologous basilic vein transpositions between forearm and upper arm regions. Ann Vasc Surg 2003;17:522-5.

JOURNAL OF VASCULAR SURGERY November 2007

32. Robbin ML, Chamberlain NE, Lockhart ME, Gallichio MH, Young CJ, Deierhoi MH, et al. Hemodialysis arteriovenous fistula maturity: US evaluation. Radiology 2002;225:59-64. 33. Prischl FC, Kirchgatterer A, Brandstatter E, Wallner M, Baldinger C, Roithinger FX, et al. Parameters of prognostic relevance to the patency of vascular access in hemodialysis patients. J Am Soc Nephrol 1995;6: 1613-8. 34. Weale AR, Usher J, Lear PA, Mitchell DC. Mortality associated with vascular access surgery for haemodialysis. Br J Surg 2003;90:501-2. 35. Dhingra RK, Young EW, Hulbert-Shearon TE, Leavey SF, Port FK. Type of vascular access and mortality in U.S. hemodialysis patients. Kidney Int 2001;60:1443-51. 36. Huber TS, Buhler AG, Seeger JM. Evidence-based data for the hemodialysis access surgeon. Semin Dial 2004;17:217-23. 37. Wells AC, Fernando B, Butler A, Huguet E, Bradley JA, Pettigrew GJ. Selective use of ultrasonographic vascular mapping in the assessment of patients before haemodialysis access surgery. Br J Surg 2005;92:143943. 38. Allon M, Robbin ML. Increasing arteriovenous fistulas in hemodialysis patients: problems and solutions. Kidney Int 2002;62:1109-24. 39. Hatjibaloglou A, Grekas D, Saratzis N, Megalopoulos A, Moros I, Kiskinis D, et al. Transposed basilic vein-brachial arteriovenous fistula: an alternative vascular access for hemodialysis. Artif Organs 1992;16: 623-5. 40. Rivers SP, Scher LA, Sheehan E, Lynn R, Veith FJ. Basilic vein transposition: an underused autologous alternative to prosthetic dialysis angioaccess. J Vasc Surg 1993;18:391-6. 41. El Mallah S. Staged basilic vein transposition for dialysis angioaccess. Int Angiol 1998;17:65-8. 42. Humphries AL Jr, Colborn GL, Wynn JJ. Elevated basilic vein arteriovenous fistula. Am J Surg 1999;177:489-91. 43. Murphy GJ, White SA, Knight AJ, Doughman T, Nicholson ML. Long-term results of arteriovenous fistulas using transposed autologous basilic vein. Br J Surg 2000;87:819-23. 44. Taghizadeh A, Dasgupta P, Khan MS, Taylor J, Koffman G. Long-term outcomes of brachiobasilic transposition fistula for haemodialysis. Eur J Vasc Endovasc Surg 2003;26:670-2. Submitted Apr 27, 2007; accepted Jul 19, 2007.