Advertisement

Outcomes comparison of HeRO and lower extremity arteriovenous grafts in patients with long-standing renal failure

Open ArchivePublished:January 14, 2013DOI:https://doi.org/10.1016/j.jvs.2012.09.040

      Objective

      The Hemodialysis Reliable Outflow (HeRO) graft is becoming a recognized alternative to lower extremity arteriovenous grafts (LEAVGs) as an option for patients who have exhausted traditional upper extremity access; however, which should be applied preferentially is unclear.

      Methods

      A retrospective review of LEAVG and HeRO implants from January 2004 to August 2010 was performed. Patient demographics, medical history, procedural data, and outcomes were evaluated.

      Results

      Within the time periods, 60 HeROs were placed in 59 patients and 22 LEAVGs were placed in 21 patients. Demographics were similar between the two groups for many factors; however, the patients who underwent HeRO placement had significantly higher body mass index compared with the LEAVG group. Mean follow-up was 13.9 months for the HeRO group and 11.9 months for the LEAVG group. The HeRO patients underwent a mean of 6.3 previous tunneled dialysis catheter insertions and 3.1 previous AVG/arteriovenous fistula placements. The LEAVG patients underwent placement of a mean of 4.1 previous tunneled dialysis catheters and 2.6 previous AVG/arteriovenous fistulas. The principal difference was the number of interventions to maintain patency, which was 2.21 per year in the HeRO group and 1.17 per year in the AVG group (P = .003) Secondary patency at 6 months was 77% for the HeRO patients and 83% for the LEAVG patients (P = .14). The HeRO and LEAVG groups had no difference in infection rate per 1000 days (0.61 vs 0.71; P = .77) or mortality rate (22% vs 19% respectively; P = .22) at 6 months.

      Conclusions

      In access challenged patients, LEAVG and HeRO offer similar rates of secondary patency, infection, and all-cause mortality. The LEAVG required fewer interventions to maintain patency, and the HeRO maintains the benefit of utilizing the upper extremity site of venous drainage. In our practice, we prefer the HeRO to LEAVG, especially in patients with peripheral arterial disease and in the obese population, because it preserves lower extremity access options.
      Central venous stenosis and central vein occlusion have become pandemic in the renal dialysis population. For the more than 350,000 patients in the United States who depend on hemodialysis (HD), >70% start dialysis with a catheter and 18% are chronically dependent on catheters for access. These catheters are fraught with complications and have been associated with an increased need for replacement, infection of the access site, and a 1.6 times increase in sepsis than in patients who are dialyzed through an arteriovenous fistula or graft.
      • Collins A.J.
      • Foley R.N.
      • Herzog C.
      • Chavers B.M.
      • Gilbertson D.
      • Ishani A.
      • et al.
      US Renal Data System 2010 Annual Data Report.
      The risk is clearly tied to the catheters, as it has been shown that removal of catheter dependency can reduce the relative risk of mortality from 3.4 to 1.4.
      • Allon M.
      • Daugirdas J.
      • Depner T.A.
      • Greene T.
      • Ornt D.
      • Schwab S.J.
      Effect of change in vascular access on patient mortality in hemodialysis patients.
      The Kidney Disease Outcomes Quality Initiative (KDOQI) guidelines provide an algorithm for placement of various fistulas and grafts, but it does not provide definitive instructions for patients with central venous obstruction (CVO). These guidelines support “chest wall” or “necklace” prosthetic graft or lower extremity fistula or graft”; however, the guidelines also report that “all upper-arm sites should be exhausted” before catheter dependence.
      Vascular Access 2006 Work Group
      Clinical practice guidelines for vascular access.
      The Society of Vascular Surgery has also made recommendations regarding vascular access strategies but has not solidified an algorithm for the management of patients with CVO.
      • Sidawy A.N.
      • Spergel L.M.
      • Besarab A.
      • Allon M.
      • Jennings W.C.
      • Padberg Jr., F.T.
      • et al.
      The Society for Vascular Surgery: clinical practice guidelines for the surgical placement and maintenance of arteriovenous hemodialysis access.
      Although catheter dialysis has been shown to be suboptimal, no comparison to date has determined the next best alternative in patients with CVO. The options explored in this study are the Hemodialysis Reliable Outflow (HeRO) graft (Hemosphere Inc, Minneapolis, Minn) and the lower extremity arteriovenous graft (LEAVG).
      The HeRO vascular access device was developed for use in maintaining upper extremity vascular access for HD patients who have CVO. Katzman et al
      • Katzman H.E.
      • McLafferty R.B.
      • Ross J.R.
      • Glickman M.H.
      • Peden E.K.
      • Lawson J.H.
      Initial experience and outcome of a new hemodialysis access device for catheter-dependent patients.
      reported that the HeRO device had function and patency rates similar to those of conventional AVGs but was suitable for patients with impaired venous outflow that would preclude upper extremity subcutaneous vascular access. Long-term data on the HeRO device are scant, and its role as an option for long-term dialysis has not been thoroughly described.
      An LEAVG has been the standard option in patients with impaired upper extremity venous drainage. The LEAVG has been demonstrated to carry a risk of infection (9%-41%) and ipsilateral limb ischemia (9%), but it is still in widespread use.
      • Miller C.D.
      • Robbin M.L.
      • Barker J.
      • Allon M.
      Comparison of arteriovenous grafts in the thigh and upper extremities in hemodialysis patients.
      • Cull J.D.
      • Cull D.L.
      • Taylow S.M.
      • Carsten 3rd, C.G.
      • Snyder B.A.
      • Yourkey J.R.
      • et al.
      Prosthetic thigh arteriovenous access: Outcome with SVS/AAVS reporting standards.
      Even with these reported risks, this type of access is chosen for many patients due to less morbidity and mortality during access creation compared with other exotic types of hemoaccess, including axillary–axillary arteriovenous straight access (necklace graft),
      • Morsy M.A.
      • Khan A.
      • Chemla E.S.
      Prosthetic axillary-axillary arteriovenous straight access (necklace graft) for difficult hemodialysis patients: a prospective single-center experience.
      superficial femoral vein transposition,
      • Chemla E.S.
      • Morsy M.
      • Anderson L.
      • Makanjuola D.
      Complex bypasses and fistulas for difficult hemodialysis access: a prospective, single-center experience.
      and accesses that require body cavity incisions, such as an axillary artery to right atrial graft.
      • Kopriva D.
      • Moustapha A.
      Axillary artery to right atrium shunt for hemodialysis access in a patient with advanced central vein thrombosis: a case report.

      Methods

       Device description

      The HeRO device has been classified by the United States Food and Drug Administration as a graft and has been commercially available since April 2008. The HeRO, when constructed, consists of a standard 6-mm expanded polytetrafluoroethylene (ePTFE) graft, which is attached to a 5-mm nitinol-reinforced silicone outflow component. An anastomosis is created between the ePTFE arterial component and an arterial inflow source, either to the brachial or axillary artery. The outflow component is placed into a central vein and does not rely on a patent extremity outflow venous tract. Using endovascular techniques, the outflow component can even be placed across occluded venous segments to a patent central vein (eg, occluded innominate vein).
      The LEAVGs in this study were placed in a loop (femoral artery to femoral vein) or straight (popliteal artery to femoral vein) configuration after a preoperative workup to exclude significant peripheral arterial disease of the lower extremity. A combination of 6-mm straight and tapered ePTFE as well as 6-mm bovine mesenteric vein (BMV) were used as conduits at the surgeon's discretion.

       Study design

      A retrospective review of LEAVG and HeRO grafts implanted within a single health network from January 1, 2004, to August 31, 2010, was performed. The health network was involved in premarket testing of the HeRO; however, all patients in this study had the HeRO placed after commercial release of the device. Patients were identified by current procedural terminology codes for HD graft placement. Exclusion criteria included patients younger than 18 years or older than 89 years at the time of access procedure. Patient charts were reviewed for demographics and preoperative comorbidities. At the time of index access insertion, the technique of insertion was recorded as well as any adjunctive maneuvers necessary for successful implantation. After insertion, the charts were examined for procedures performed on the access at one of the medical center sites or the outpatient endovascular center. Follow-up was recorded until July 31, 2011, by search of the electronic medical record, which is used for hospital and office-based encounters.
      The access modality was selected by the operating surgeon, who generally adhered to the KDOQI guidelines as an access strategy. As such, LEAVG and HeRO were only used as options after traditional fistula and graft options were exhausted. All of the surgeons had extensive experience in percutaneous and open techniques for treatment of central venous stenosis or occlusion; therefore, it was assumed that every reasonable effort was made to avoid nontraditional access. Often, treatment of a central venous occlusion with angioplasty resulted in a functional access, and an alternate modality was not necessary. Central venograms were obtained both to attempt to treat failing upper extremity access and to plan subsequent hemoaccess.
      Primary outcomes were length of time that the grafts were capable of providing access for HD or secondary patency. The need for intervention performed throughout the life of the graft was also analyzed and compared as a ratio to the length of time the graft maintained secondary patency. Secondary outcomes were morbidity related to the device insertion process, all-cause mortality, and device infection. Patency definition was based on standardized nomenclature as previously outlined.
      • Sidawy A.N.
      • Gray R.
      • Besarab A.
      • Henry M.
      • Ascher E.
      • Silva Jr., M.
      • et al.
      Recommended standards for reports dealing with arteriovenous hemodialysis accesses.
      Graft infections were identified as need for treatment with excision of a portion or the entire graft at the discretion of the operating surgeon.
      Prior to data collections, the Institutional Review Board approved study design and scientific method. Statistical analysis was performed to compare the two groups using two-sample Wilcoxon test, the Fisher exact test, and Poisson regression. Kaplan-Meier survival curves were compared with the log-rank test. Dotted lines were used to represent the time of the standard error exceeding 10% of the mean. P < .05 was considered statistically significant. An independent statistician corroborated the statistical analysis.

      Results

      During the 6.5 years studied, 22 LEAVGs were placed in 21 patients, and 60 HeRO grafts were placed in 59 patients. One patient had one HeRO and one LEAVG placed. In total, 97% (58/60) of the HeROs were placed in the upper extremity using the brachial artery as inflow and the outflow placed centrally. Two HeROs were placed in the lower extremity in a loop configuration from the femoral artery directly to the vena cava by a retroperitoneal exposure.
      Of the LEAVGs, 91% (20/22) were placed in a loop configuration from either the common femoral or the superficial femoral artery to the common femoral or femoral vein. The remaining 9% (2/22) were constructed from the popliteal artery to femoral vein in a straight configuration. Thirty-two percent (7/22) of grafts were constructed of BMV (Procol; Hancock Jaffe, Irvine, Calif), and 68% (15/22) were ePTFE. Of the ePTFE grafts, seven were 6 mm, five were a tapered graft, and in three procedures the diameter of the ePTFE could not be determined.
      The preoperative clinical characteristics of the patients prior to access placement are summarized in the Table. Patients who underwent a HeRO insertion had similar age, sex, height, and race as patients who received an LEAVG. Not surprisingly, patients selected to receive an LEAVG were significantly more slender, with a mean weight of 160 lb compared with a mean weight of 193 lb for HeRO patients (P = .0178). The mean body mass index, which is derived from weight but takes into account height, was significantly increased in the HeRO cohort at 32 kg/m2 compared with 26 kg/m2 in the LEAVG group (P = .0248). The two groups were similar for many comorbidities, including hypertension, congestive heart failure, coronary artery disease, cerebrovascular disease, chronic obstructive pulmonary disease, hyperlipidemia, types of diabetes, depression. and history of deep vein thrombosis. A history of positive blood culture was documented in 51% of patients before receiving a HeRO compared with 10% of patients who received an LEAVG (P = .0013).
      TablePreoperative clinical characteristics
      HeRO graft (N = 59)Thigh AVG (N = 20)P value
      Age, mean (SD; range), years58.2 ([14.2]; range, 26.8-86.0)53.2 ([17.0]; range, 2.5-79.3).1854
      Male49.2% (29)30% (6).1935
      Female50.9% (30)70% (14)
      Height, mean (SD; range), inches65.7 ([6.0]; range, 38-77)65.4 ([3.2]; range, 60-70) (n = 19).5452
      Weight, mean (SD; range), pounds192.7 ([56.3]; range, 112-451) (n = 59)160.4 ([24.9]; range, 128-197) (n = 19).0178
      Body mass index, mean (SD; median; range), kg/m232.0 ([10.0]; median, 31.6; range, 19.7-75.0) (n = 57)26.4 ([4.0]; median, 26.6; range, 18.4-32.4) (n = 19).0248
      Race.1969
       Other3.4% (2)10% (2)
       Black88.1% (52)90% (18)
       White8.5% (5)0% (0)
      History of bacteremia50.9% (30)10% (2).0013
      Hypertension91.5% (54)95% (19)1
      Congestive heart failure23.7% (14)15% (3).5375
      Coronary artery disease/myocardial infarction33.9% (20)20% (4).2771
      Cerebrovascular disease/cerebrovascular accident18.6% (11)25% (5).5346
      Chronic obstructive pulmonary disease13.6% (8)10% (2)1
      Hyperlipidemia/cholesterolemia39.0% (23)15% (3).0577
      Diabetes.0788
       None39.0% (23)60% (12)
       Type I44.1% (26)40% (8)
       Type II17.0% (10)0% (0)
      Current tobacco use13.6% (8)15% (3)1
      Depression17.0% (10)5% (1).2732
      Deep vein thrombosis30.5% (18)20% (4).5644
      AVG, Arteriovenous graft; HeRO, Hemodialysis Reliable Outflow; SD, standard deviation.
      Patients had undergone multiple previous HD access procedures before the index event. The mean number (standard deviation [SD]) of previous AVGs in the HeRO group was 1.8 (SD, 1.3; median, 2.0; range, 0-5) and 2.1 ([1.3]; median, 1.5; range, 1-4) in the LEAVG group (P = .51). The mean arteriovenous fistula creation in the HeRO group was 1.3 ([0.9]; median, 1.0; range, 0-4) and in the LEAVG group was 0.5 ([0.5]; median, 0.5; range, 0-1; P = .02). Catheters were a large part of the history of these patients, with the HeRO group having received a mean of 6.3 ([5.0]; median, 5.0; range, 1-27) and the LEAVG group having received a mean of 4.1 ([3.3]; median, 3.0; range, 0-11; P = .09).
      The mean length of time from insertion to either loss of secondary patency or completion of study was 11.8 months (range, 0.1-59.5 months) for the LEAVG group and 13.9 months (range, 0-37.5 months) for the HeRO group (P = .94). For determining all-cause mortality, follow-up concluded at date of death or July 31, 2011, if the patient was still alive. There was an increased incidence in the number of procedures necessary to maintain patency in the HeRO group compared with the LEAVG group. Each HeRO required a mean of 2.21 interventions per patient per year compared with 1.17 interventions per patient per year for the LEAVG group (P = .003).
      Of the 40 HeROs that underwent reintervention, 37 of the first procedures (93%) were a thrombectomy. Of the others, two (5%) were excised for infection and one (2%) for treatment of steal, one titanium connector (2%) became dislodged, and one (2%) failed thrombectomy. Of the eight LEAVGs that underwent reintervention, five of the first procedures (63%) were a thrombectomy. All three other LEAVGs (37%) were excised for infection.
      Previous studies of LEAVG have reported an increased risk of infection; however, compared with the HeRO devices in this study the infection rates are statistically similar. The LEAVG became infected at a rate of 0.71 per 1000 days, whereas HeROs became infected at a rate of 0.61 per 1000 days (P = .77). Overall, 13 of 59 HeROs (22.0%) and 6 of 20 LEAVGs (30%) became infected over the study period.

       Perioperative complications

      Within the HeRO group, one patient died 3 days after a femoral HeRO placement due to an uncontrolled retroperitoneal hematoma. One patient developed a hematoma at the site of the brachial arterial anastomosis, which caused compression and thrombosis. This patient underwent a thrombectomy on the first postoperative day. A third patient developed upper extremity steal syndrome following HeRO insertion and on the first postoperative day was treated with a 4-mm ePTFE interposition at the arterial anastomosis. The only similar complication of the LEAVG group was one patient with a steal syndrome; however, this occurred 2.8 years after insertion and was treated with graft excision.

       Survival analysis

      Raw data were inspected, and survival plots were performed to analyze patency over the study period. At 12 months, only 15% of HeROs maintained primary unassisted patency compared with 51% of LEAVGs. However, due to the small sample size at 1 year, these percentages are not reliably accurate. Fig 1 shows the Kaplan-Meier curve of primary patency, which is truncated very early due to small sample size. The standard error of the mean exceeds 10% at approximately 20 days for the LEAVG and 60 days for the HeRO and therefore is not reliable beyond those points. No meaningful comparison can be made from these primary patency curves. At 12 months, 57% of HeROs maintained secondary patency compared with 69% of LEAVGs (P = .14; Fig 2). The statistical validity of this curve exists through 6 months, showing patency of 77% for HeRO and 83% for LEAVG.
      Figure thumbnail gr1
      Fig 1Primary patency. AVG, Arteriovenous graft; HeRO, Hemodialysis Reliable Outflow (graft).
      Figure thumbnail gr2
      Fig 2Secondary patency. AVG, Arteriovenous graft; HeRO, Hemodialysis Reliable Outflow (graft).
      Kaplan-Meier analysis was also performed for survival free of device infection and all-cause mortality with grafts (Figs 3 and 4). At 6 months, 84% of HeRO grafts and LEAVGs required surgical treatment for infection (P = .76; Fig 3). All-cause mortality at 12 months was 33% in the HeRO group and 20% in the LEAVG group (P = .22; Fig 4).
      Figure thumbnail gr3
      Fig 3Survival free of device infection. AVG, Arteriovenous graft; HeRO, Hemodialysis Reliable Outflow (graft).
      Figure thumbnail gr4
      Fig 4Kaplan-Meier curve for survival. AVG, Arteriovenous graft; HeRO, Hemodialysis Reliable Outflow (graft).

      Discussion

      This study describes the longest follow-up for the HeRO device and the only direct comparison with LEAVG from a single center. The KDOQI guidelines provide a widely accepted strategy for vascular access in the majority of the HD population; however, the guidelines provide no clear algorithm in patients who have exhausted traditional upper extremity access options. The preferred “unconventional” access has not been elucidated in the literature. This study shows in a retrospective manner that the outcomes of LEAVG and HeRO devices are similar in terms of length of secondary patency, freedom from infection, and all-cause mortality; however, HeROs require more interventions to maintain patency.
      The patients involved in this study had low rates of upper extremity fistula insertion (mean, 1.3 for HeRO; mean, 0.5 for LEAVG) before requiring an alternate access option despite the responsible surgeon adhering to a fistula-first access strategy. The most significant contributing to the low fistula rate is likely related to the patency and quality of the superficial venous system in the arms of these patients. Patients who progress to an alternative access modality either have poor venous conduits or have required dialysis for a length of time that has exhausted the acceptable options that are available.
      Katzman et al
      • Katzman H.E.
      • McLafferty R.B.
      • Ross J.R.
      • Glickman M.H.
      • Peden E.K.
      • Lawson J.H.
      Initial experience and outcome of a new hemodialysis access device for catheter-dependent patients.
      previously reported on 36 HeRO devices over a mean of 8.6 months and described an intervention rate of 2.5 per patient per year, which is comparable to the rate of 2.2 per year in this study. Their article also reports an infection rate of 0.70 per 1000 days, similar to our rate of 0.61 per 1000 days.
      • Katzman H.E.
      • McLafferty R.B.
      • Ross J.R.
      • Glickman M.H.
      • Peden E.K.
      • Lawson J.H.
      Initial experience and outcome of a new hemodialysis access device for catheter-dependent patients.
      Gage et al
      • Gage S.M.
      • Katzman H.E.
      • Ross J.R.
      • Hohmann S.E.
      • Sharpe C.A.
      • Butterly D.W.
      • et al.
      Multi-center experience of 164 consecutive hemodialysis reliable outflow [HeRO] graft implants for hemodialysis treatment.
      reported better outcomes of 164 HeROs inserted at four centers with mean follow-up of 12.8 months, with an intervention rate of 1.5 per year and an infection rate of 0.14 per 1000 days. Those outcomes could not be duplicated in this study. Additionally, the study reports a 1-year primary patency of 49% and a 1-year secondary patency of 91%.
      • Gage S.M.
      • Katzman H.E.
      • Ross J.R.
      • Hohmann S.E.
      • Sharpe C.A.
      • Butterly D.W.
      • et al.
      Multi-center experience of 164 consecutive hemodialysis reliable outflow [HeRO] graft implants for hemodialysis treatment.
      The Dialysis Access Consortium report from a multicenter prospective study published in the New England Journal of Medicine could not duplicate these findings in 649 AVGs. This study reports primary patency rate of 23% to 28% at 1 year, and another Dialysis Access Consortium study reports an AVG secondary patency rate of 62% to 70%.
      • Dixon B.S.
      • Beck G.J.
      • Vazquez M.A.
      • Greenberg A.
      • Delmez J.A.
      • Allon M.
      • et al.
      Effect of dipyridamole plus aspirin on hemodialysis graft patency.
      • Dixon B.S.
      • Beck G.J.
      • Dember L.M.
      • Vazquez M.A.
      • Greenber A.
      • Delmez J.A.
      • et al.
      Use of aspirin associates with longer primary patency of hemodialysis grafts.
      Cull et al
      • Cull J.D.
      • Cull D.L.
      • Taylow S.M.
      • Carsten 3rd, C.G.
      • Snyder B.A.
      • Yourkey J.R.
      • et al.
      Prosthetic thigh arteriovenous access: Outcome with SVS/AAVS reporting standards.
      previously reviewed a series of 125 LEAVGs followed for a mean of 20 months. Their report produced a life-table analysis that approximates this study's prediction for primary patency (34% at 1 year), secondary patency (68% at 1 year), survival (74% at 1 year), and intervention rates (1.68 per patient per year) for LEAVG.
      • Cull J.D.
      • Cull D.L.
      • Taylow S.M.
      • Carsten 3rd, C.G.
      • Snyder B.A.
      • Yourkey J.R.
      • et al.
      Prosthetic thigh arteriovenous access: Outcome with SVS/AAVS reporting standards.
      There are similarities and differences that can account for incongruent data between the previously published reports on HeRO, LEAVG, and this study. Many of these reports had a small sample size, which is partially due to the few number of patients who have undergone HeRO and LEAVG implantation. To some extent this is because the HeRO was only released commercially in 2008. In addition, there is a very small subset of the population that will require an exotic HD access. The small sample size in this report is reflected in the truncated Kaplan-Meier curves at the point when the data become statistically unreliable.
      Determining the best access in this patient population should not be based purely on patency data. Preserving options for further access as well as a reduction in catheter dependency should be strong considerations for most patients. A HeRO may preserve an upper extremity venous outflow that currently is being stented open with an indwelling catheter. If the upper extremity venous outflow is utilized, this could preserve the lower extremities for future dialysis access. A key advantage of the HeRO is that additional access sites can be used and others preserved for subsequent access options.
      A cost analysis would be helpful in comparing these two modalities of hemoaccess. The expense of the HeRO is related not only to the device but also to fluoroscopy and adjunctive endovascular tools that may be necessary for insertion. The LEAVG has a range of charges that are related to the AVG conduit used. The reinterventions are another expense that needs to be factored into the choice of access. An in-depth cost analysis is beyond the scope of this study; however, the cost of treating a catheter-related bloodstream infection likely would surpass the cost of both device insertion and maintenance.
      There are multiple limitations to this study. The LEAVG patients may not have had a patent upper extremity venous outflow and therefore may not have been candidates for an upper extremity HeRO. The LEAVG patients were selected at the discretion of the operating surgeon and did not have standard exclusion or inclusion criteria. Ankle-brachial indexes were not assessed in all cases, and often the quality of the femoral pulse was the only arterial examination performed. The two groups were similar in the factors identified but only were controlled as much as possible for a retrospective study. Patients who received HeROs had a greater mean weight and body mass index. This is likely due to the surgeon's preference for avoiding LEAVG and the accompanying groin incisions in patients with a large abdominal pannus for fear of graft infection.
      There is some heterogeneity to the HeRO population because two devices were placed in a femoral configuration. One patient experienced a mortality related to the procedure; however, the other maintained secondary patency at last follow-up of approximately 6 months. The other patient previously had been relying on a tenuous transhepatic catheter for access and was believed to be better served with the femoral HeRO.
      Another limitation was the length of the Kaplan-Meir curves. The validity of the life-table analysis can only extend until the standard error exceeded 10% of the mean. The length of this time period was negatively influenced by the relatively small sample sizes at the onset and the high proportion of grafts that lost patency during the study period. Once the grafts lost patency, they were excluded from this calculation, which effectively reduced the sample size at that time point.
      The medical record reviewed and hospital system provided treatment for the vast majority of patients in the area. Although it is possible that a patient could have seen another surgeon to care for a complication, it is less likely considering the complexity of these patients' operative course. It is thought that the overwhelming majority of procedures performed after the index procedures were captured in the chart review.
      The LEAVG group was heterogeneous with regard to the material used for the conduit. The BMV has been adopted by a portion of the surgeons and accounted for one third of the grafts implanted. The ePTFE was used for the remaining two thirds of the LEAVGs. The BMV may have a better patency than ePTFE; however, the groups were too small to produce any significant or meaningful conclusions. This is clearly an area where further studies are needed.
      The rate of reintervention in the HeRO group likely is related to the “mode of failure” of the HeRO. The most common failure mode of a traditional AV access is intimal hyperplasia at the venous anastomosis causing stenosis and occlusion. Because the HeRO does not rely on a venous anastomosis, its primary mode of failure cannot be extrapolated from graft failures and has yet to be discovered. When a HeRO experiences thrombosis, it normally can undergo successful thrombectomy because the venous outflow, the vena cava, usually remains patent.
      There are limitations to these data and the analysis due to the retrospective nature of the study. A prospective, randomized comparison could yield guidelines for management of HD patients with CVO that precludes traditional upper extremity AV access. Until that information is available, this analysis could provide guidance for access in these difficult patients by displaying the results of one health network's approach.

      Conclusions

      In our experience, the advantage of the HeRO over the LEAVG resides in the ability to maintain an upper extremity access site with superior vena cava venous drainage. The advantage of the LEAVG over the HeRO is the reduction in the number of interventions necessary to preserve the access. The HeRO is equal to the LEAVG in secondary patency, rate of infection, and all-cause mortality. In our practice, we prefer the HeRO to the LEAVG, particularly in patients with peripheral arterial disease and in the obese population, because it preserves the lower extremities for future access options.

      Author contributions

      Conception and design: SS, JH, JMP
      Analysis and interpretation: SS, JW, JH, JMP, MG
      Data collection: SS, JW, CK, JP, MG
      Writing the article: SS, MG
      Critical revision of the article: SS, JMP, MG
      Final approval of the article: SS, JW, JH, CK, AM, JP, JMP, MG
      Statistical analysis: SS, JW, JMP, MG
      Obtained funding: Not applicable
      Overall responsibility: SS

      References

        • Collins A.J.
        • Foley R.N.
        • Herzog C.
        • Chavers B.M.
        • Gilbertson D.
        • Ishani A.
        • et al.
        US Renal Data System 2010 Annual Data Report.
        Am J Kidney Dis. 2010; 57: A8
        • Allon M.
        • Daugirdas J.
        • Depner T.A.
        • Greene T.
        • Ornt D.
        • Schwab S.J.
        Effect of change in vascular access on patient mortality in hemodialysis patients.
        Am J Kidney Dis. 2006; 47: 469-477
        • Vascular Access 2006 Work Group
        Clinical practice guidelines for vascular access.
        Am J Kidney Dis. 2006; 48: S176-S247
        • Sidawy A.N.
        • Spergel L.M.
        • Besarab A.
        • Allon M.
        • Jennings W.C.
        • Padberg Jr., F.T.
        • et al.
        The Society for Vascular Surgery: clinical practice guidelines for the surgical placement and maintenance of arteriovenous hemodialysis access.
        J Vasc Surg. 2008; 48: S2-25
        • Katzman H.E.
        • McLafferty R.B.
        • Ross J.R.
        • Glickman M.H.
        • Peden E.K.
        • Lawson J.H.
        Initial experience and outcome of a new hemodialysis access device for catheter-dependent patients.
        J Vasc Surg. 2009; 50: 600-607
        • Miller C.D.
        • Robbin M.L.
        • Barker J.
        • Allon M.
        Comparison of arteriovenous grafts in the thigh and upper extremities in hemodialysis patients.
        J Am Soc Nephrol. 2003; 14: 2942-2947
        • Cull J.D.
        • Cull D.L.
        • Taylow S.M.
        • Carsten 3rd, C.G.
        • Snyder B.A.
        • Yourkey J.R.
        • et al.
        Prosthetic thigh arteriovenous access: Outcome with SVS/AAVS reporting standards.
        J Vasc Surg. 2004; 39: 381-386
        • Morsy M.A.
        • Khan A.
        • Chemla E.S.
        Prosthetic axillary-axillary arteriovenous straight access (necklace graft) for difficult hemodialysis patients: a prospective single-center experience.
        J Vasc Surg. 2008; 48: 1251-1254.e1
        • Chemla E.S.
        • Morsy M.
        • Anderson L.
        • Makanjuola D.
        Complex bypasses and fistulas for difficult hemodialysis access: a prospective, single-center experience.
        Semin Dial. 2006; 19: 246-250
        • Kopriva D.
        • Moustapha A.
        Axillary artery to right atrium shunt for hemodialysis access in a patient with advanced central vein thrombosis: a case report.
        Ann Vasc Surg. 2006; 20: 418-421
        • Sidawy A.N.
        • Gray R.
        • Besarab A.
        • Henry M.
        • Ascher E.
        • Silva Jr., M.
        • et al.
        Recommended standards for reports dealing with arteriovenous hemodialysis accesses.
        J Vasc Surg. 2002; 35: 603-610
        • Gage S.M.
        • Katzman H.E.
        • Ross J.R.
        • Hohmann S.E.
        • Sharpe C.A.
        • Butterly D.W.
        • et al.
        Multi-center experience of 164 consecutive hemodialysis reliable outflow [HeRO] graft implants for hemodialysis treatment.
        Eur J Vasc Endovasc Surg. 2012; 44: 93-99
        • Dixon B.S.
        • Beck G.J.
        • Vazquez M.A.
        • Greenberg A.
        • Delmez J.A.
        • Allon M.
        • et al.
        Effect of dipyridamole plus aspirin on hemodialysis graft patency.
        N Engl J Med. 2009; 360: 2191-2201
        • Dixon B.S.
        • Beck G.J.
        • Dember L.M.
        • Vazquez M.A.
        • Greenber A.
        • Delmez J.A.
        • et al.
        Use of aspirin associates with longer primary patency of hemodialysis grafts.
        J Am Soc Nephrol. 2011; 22: 773-781