- Email: [email protected]
Initial clinical experience with a polytetrafluoroethylene vascular dialysis graft reinforced with nitinol at the venous end
Initial clinical experience with a polytetrafluoroethylene vascular dialysis graft reinforced with nitinol at the venous end Filippo Benedetto, MD, PhD,a Domenico Spinelli, MD,a Narayana Pipitò, MD, PhD,a Giambattista Gagliardo, MD,b Alberto Noto, MD, PhD,c Simona Villari, MD,a Antonio David, MD,c and Francesco Spinelli, MD,d Messina, Palermo, and Rome, Italy
ABSTRACT Objective: The purpose of this study was to examine the outcomes of a vascular hybrid polytetrafluoroethylene (PTFE) graft, provided with a nitinol-reinforced section (NRS) on one end, in hemodialysis vascular access placement. Methods: A retrospective study was conducted including all the consecutive patients who underwent Gore Hybrid Vascular Graft (GHVG; W. L. Gore & Associates, Flagstaff, Ariz) implantation for hemodialysis access placement between October 2013 and November 2015. A propensity-matched control group was obtained from consecutive patients who underwent standard PTFE arteriovenous graft implantation between January 2010 and July 2013. The selection criteria were inadequate venous material for autogenous arteriovenous fistula placement, patent deep venous circulation, and vein diameter of 4 to 8.5 mm. The implantation technique involves the insertion of the NRS some centimeters into the target vein. Fluoroscopic guidance helps deploy the device in the desired landing zone (ie, position of the proximal end of the NRS), based on anatomic landmarks. Survival, functional patency rates, and complications were compared with a propensity-matched historical control group. Vein diameter, previous vascular access placement, and diabetes were tested as predictors of reintervention with a logistic regression analysis. Results: There were 32 patients (14 men; mean age, 69 6 14 years) who received the GHVG graft. The historical control group included 43 patients. Technical success was 100%. The graft configuration was brachial-axillary (n ¼ 22 [69%]), brachial-basilic loop (n ¼ 5 [16%]), brachial-antecubital loop (n ¼ 3 [9%]), axilloaxillary loop (n ¼ 1 [3%]), and femoralfemoral loop (n ¼ 1 [3%]). Mean NRS oversize was 20% 6 7% (range, 3%-34%; median, 19%). Perioperative complications requiring revision included acute limb ischemia treated with thrombectomy (n ¼ 1 [3%]) and graft infection requiring explantation (n ¼ 2 [6%]). Two patients (6%) died in the hospital of unrelated causes. The mean follow-up was 15 6 11 months (range, 0-33 months; median, 15.5). The propensity-matched groups included 25 patients each. Survival estimates at 24 months for the GHVG and standard PTFE groups were 91% 6 6% and 82% 6 9% (P > .05), respectively. The 12-month patency estimates were as follows: functional primary patency, 66% 6 10% vs 51% 6 10% (P > .05); functional assisted primary patency, 75% 6 9% vs 51% 6 10% (P > .05); and functional secondary patency, 79% 6 9% vs 67% 6 10% (P > .05). Reduction in vein diameter was associated with reintervention. Conclusions: The GHVG is a safe and effective alternative to standard PTFE in hemodialysis access surgery. Careful planning for the landing zone is advisable, especially for small outflow veins. Larger studies and randomized trials are needed to define the role for this device. A study including a greater number of centers experienced with this device is currently under way. (J Vasc Surg 2016;-:1-9.)
The life expectancy of hemodialysis-dependent patients has been increasing in the last few decades.1 Autogenous arteriovenous fistula (AVF) is the “gold From the Department of Biomedical Sciences and Morphological and Functional Imaginga and Department of Human Pathology of Adult and Developmental Age “Gaetano Barresi”,c University of Messina, Messina; the Vascular Surgery Unit, Noto-Pasqualino Clinic, Palermob; and the Vascular Surgery Unit, Campus Bio-Medico University, Rome.d Author conflict of interest: none. Correspondence: Domenico Spinelli, MD, Department of Biomedical Sciences and Morphological and Functional Imaging, University of Messina, Via Consolare Valeria 1, Messina 98125, Italy (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2016.07.133
standard” for hemodialysis access for its good longterm patency.2,3 After repeated failure of an AVF and exhaustion of venous material, the use of a prosthetic arteriovenous graft (AVG) is the best alternative. Many types of prostheses have been used so far, with 1-year patency rates generally lower than for AVF. A review including 34 studies reported a cumulative 18-month primary patency of 51% for AVF and 33% for AVG.4 Intimal hyperplasia (IH) is considered to be the main cause of graft occlusion.1 The increasing need for long-lasting vascular accesses has fostered the research on IH. The turbulent flow, strictly related to the width of the anastomosis angle, produces shear stress on the vein wall.5 The subsequent endothelial injury promotes smooth muscle cell migration and proliferation and eventually is responsible for platelet recruitment and coagulation cascade activation.6-10 The next step is graft thrombosis. In light 1
2
Benedetto et al
Journal of Vascular Surgery ---
of these findings, a new concept of vascular hybrid prosthesis has been developed, whereby the vein-side anastomosis is replaced by a nitinol-reinforced section (NRS), to be inserted and deployed into the target vein, to form an end-to-end junction. This design, based on a straight geometry, is intended to keep the flow laminar, reducing shear stress and thus preventing IH.11 Nonetheless, it is still unclear whether the straight geometry actually allows laminar flow in AVGs. Evidence that a nitinol stent at the venous end of a standard polytetrafluoroethylene (PTFE) graft improved patency of a failing AVG compared with simple angioplasty supports the validity of the hybrid concept.12 The idea of replacing the distal anastomosis of a graft by a nitinol-covered stent was also described by Lachat et al.13 The aim of the study was to report a 2-year experience with the hybrid graft in hemodialysis access and to analyze outcomes and complications.
METHODS Patient cohort. A multicenter retrospective cohort study was conducted including all consecutive patients who underwent implantation of the Gore Hybrid Vascular Graft (GHVG; W. L. Gore & Associates, Flagstaff, Ariz) for hemodialysis access between October 2013 and November 2015 at the University Hospital Gaetano Martino, Messina, Italy, and at Noto-Pasqualino Clinic, Palermo, Italy. A propensity-matched control group was obtained from consecutive patients who underwent standard PTFE AVG implantation between January 2010 and July 2013 at the same centers. Patient demographics and clinical and procedure data were collected in a dedicated prospectively maintained Excel (Microsoft, Redmond, Wash) database. This study is in agreement with the principles outlined in the Declaration of Helsinki. The protocol and the informed consent were approved by the Institutional Review Board. All patients gave a written informed consent to participate in a study. This study did not undergo editorial review, nor was it supported by representatives of the graft manufacturer. Preoperative protocol. All patients underwent preoperative routine blood tests and careful physical examination, evaluating the presence and validity of distal pulses, skin inflammation, and any evidence of central venous obstruction, such as edema and collateral veins of the upper extremity or chest wall. Allen test was routinely performed. Duplex ultrasound (Vivid 4; GE Healthcare, Chalfont St Giles, United Kingdom) arterial and venous preoperative mapping was routinely performed. Target artery and vein diameters were measured on the transverse ultrasound scan, with the patient supine at room temperature, by experienced operators. Vein compressibility, distensibility, thickness, and continuity were assessed with a rubber tourniquet in place. Arterial patency and presence of triphasic flow were
2016
ascertained by duplex ultrasound with a 7 MHz linear probe at 60-degree insonation angle. Patients were deemed unsuitable for autogenous access placement if a superficial vein, including the upper arm cephalic and basilic vein, at least 2.5 mm in diameter was lacking or was present for a short track before the confluence into the deep veins, in both the dominant and nondominant arms.14 Antecubital veins, basilic veins, deep veins at the elbow, axillary veins, and femoral veins were explored and considered a possible target vein if unsuitable for AVF placement. When central vein stenosis or occlusion was suspected or in patients known to have had a previous central vein catheter or pacemaker, venography or magnetic resonance angiography was performed.15 A complete cardiac evaluation, including echocardiography, was part of the preoperative protocol. The selection criteria were unsuitability for AVF placement, presence of patent deep venous circulation, appropriate diameters of inflow artery (3 mm) and target vein (4-8.5 mm), absence of local and systemic infections, and absence of upper limb ischemia. If the anatomic requirements were met, the GHVG was preferred to a traditional PTFE graft. Device characteristics. The GHVG is approved by the U.S. Food and Drug Administration and the European Union (CE Mark). It is commercially available in the United States, Europe, Russia, and a few Middle Eastern countries.16 It consists of a PTFE graft with an NRS at one end. The continuous luminal surface provides additional resistance to thrombosis and seroma formation thanks to covalently bonded heparin of porcine origin and a low permeability film. The NRS is conceived to be inserted directly through a venotomy or over the wire through a peel-away sheath into the target vessel, to be deployed so as to replace the venous anastomosis with an end-to-end junction. The deployment is carried out by pulling a constraining wire. The endoluminal position of the proximal end of the NRS is defined as the landing zone. The idea behind this design is to reduce turbulent flow to prevent IH and, at the same time, to allow the treatment of stenoses of the target vein.11 The NRS is available in five diameters (5-9 mm) and two lengths (5 and 10 cm). The nonreinforced end is anastomosed to the inflow artery with a running suture like a standard AVG. It has standard 6-mm diameter and is tapered to adapt to the size of the NRS. It uses stretch technology and should be put under tension before placement of the suture. Technique. Four surgeons performed the procedures. All procedures were carried out in a surgical room equipped with a mobile fluoroscopic C-arm (Philips BV Pulsera; Philips Medical Systems, Amsterdam, The Netherlands) because contrast angiography is required. Vascular accesses were placed in both the upper and lower limbs. The selection of graft size and access
Journal of Vascular Surgery Volume
-,
Number
Benedetto et al
3
-
Fig 1. Brachial-axillary hybrid arteriovenous graft (AVG). A, Venous anastomosis between nitinol-reinforced section (NRS) and axillary vein. B, Angiogram.
configuration was made by the surgeon.17 Because the graft by design occludes the distal venous outflow, the NRS was placed as distal as reasonable, based on vein diameter, aiming to safeguard the chance for future access placements and preserving the confluence of venous branches whenever possible. Optimal NRS oversize was considered to be 5% to 20%, as suggested in the instructions for use. The technique of brachial-axillary graft implantation involved the surgical incision in the armpit with dissection of the axillary vessels. The brachial artery was then exposed through an incision proximal to the elbow. The nonstented part of the device was stretched and tunneled in the subcutaneous tissue. The NRS was inserted into the axillary vein through a venotomy or a peel-away sheath. The NRS was then advanced inside the vein except for about 2 cm, deployed, and secured to the vein with two 6-0 polypropylene stitches (Fig 1, A). These sutures, required per instructions for use, prevent graft dislocation during the procedure. The arterial anastomosis was performed in an end-to-side fashion, as for traditional grafts. As a final step, a balloon dilation of the stent was performed through direct puncture of the prosthesis. Any additional endovascular procedure addressing outflow stenoses was performed at this time. Alternatively, if doubt persisted about graft sizing, angiography was performed before the selection of the GHVG. Completion control was routinely performed at the end of the procedure by contrast angiography when it was not contraindicated (Fig 1, B). Patients not yet under hemodialysis were controlled with duplex ultrasound only. For the brachialantecubital forearm loop, the brachial artery and the antecubital vein were exposed through a single surgical incision. The prosthesis was then tunneled in a loop
configuration. A comparable approach has been used for femoral-femoral looped inguinal access placement. Through an inguinal incision, the femoral vessels were exposed, and the superficial femoral vein served as target vein (Fig 2). For axilloaxillary loop AVG, the axillary vessels were exposed through an axillary incision. All procedures were performed under local anesthesia (mepivacaine hydrochloride 2%), locoregional plexus block, or spinal anesthesia, except axilloaxillary loop AVG, which was performed under general anesthesia. Antibiotic prophylaxis with cephalosporin was performed before the surgical incision. Postoperative management. Patients were discharged with antiplatelet therapy (clopidogrel, 75 mg/d). Required recovery time before cannulation was 2 to 4 weeks. The vascular access was evaluated with duplex ultrasound through hospital discharge, at 30 days, and every 3 months after the procedure. Flow measurements were performed monthly. When dynamic venous pressure >200 mm Hg with flow of 300 mL/min, Kt/V <0.8, or recirculation >25% was present, duplex ultrasound or angiographic control was performed.18 Outcomes and statistical analysis. Interventions, complications, and outcomes were defined according to the reporting standards on arteriovenous accesses developed by the Society for Vascular Surgery and the American Association for Vascular Surgery.19 Primary end points were survival, functional primary patency, assisted primary patency, and secondary patency. Secondary end points were vascular access-related complications: postoperative hematoma, limb edema, graft infection, anastomosis or venous stenosis, steal syndrome, and other complications. Continuous variables are reported as
4
Benedetto et al
Journal of Vascular Surgery ---
2016
Fig 2. Femoral-femoral loop arteriovenous graft (AVG). A, Arterial and venous anastomoses are carried out on superficial femoral vessels. B, Angiogram.
mean 6 standard deviation or median with range. Categorical variables are reported as count (percentage). Statistical analysis was performed using SPSS software (IBM Corp, Armonk, NY). Patients in the GHVG group were matched to patients in the standard PTFE group, using the propensity score matching (PSM) methodology of Rosenbaum and Rubin.20 Our chosen methodology used caliper matching on the propensity score without replacement to create a balanced data set. The following variables were used for PSM: AVG configuration, sex, age, hypertension, diabetes, coronary artery disease, chronic obstructive pulmonary disease, renal failure stage, vein diameter, and previous vascular access surgery. The analyses of matched data sets were conducted with Wilcoxon signed rank test for continuous variables and McNemar test for dichotomous ones. Functional patency rates were analyzed with Kaplan-Meier curve with Cox proportional hazards model stratified on matched pairs. A logistic regression model was performed to ascertain the effects of target vein diameter, previous vascular access surgery on the same limb, and diabetes on the likelihood of GHVG-related reintervention.
RESULTS Patient population. Thirty-two patients received a GHVG for hemodialysis access placement. The historical control group included 43 patients. Demographics and baseline characteristics for both groups are shown in Table I. Among patients who received a GHVG, 28 patients (88%) were under hemodialysis treatment for end-stage renal disease, and in 4 patients (13%), the
anticipated start of hemodialysis treatment was within 6 weeks. Four patients (13%) received the GHVG as a first choice because they lacked superficial veins >2.5 mm in diameter and long enough for an effective AVF to be placed. The remaining had at least one previous intervention of vascular access placement. The mean target vein diameter was 5.8 6 1.0 mm (range, 3.8-8.0 mm; median, 5.8 mm). Treatment characteristics. The mean duration of interventions was 78 6 27 minutes (range, 40-155; median, 70). The mean amount of contrast agent used was 30 6 8 mL, (range, 20-54; median, 28). The mean NRS oversize was 20% 6 7% (range, 3.4%-34.6%; median, 19.7%). The NRS size of the implanted GHVG ranged from 5 to 9 mm in diameter, and both lengths were used (Table II). The graft configuration was brachialaxillary, brachial-basilic loop, brachial-antecubital loop, axilloaxillary loop, and femoral-femoral loop in 22 cases (69%), 5 cases (16%), 3 cases (9%), 1 case (3%), and 1 case (3%), respectively. In six cases, additional planned endovascular procedures were carried out to treat stenoses of the outflow vein or central veins: three cases were treated with a covered stent in the right innominate vein, axillary vein, and brachial vein, respectively; three cases were treated with simple balloon angioplasty of the right innominate vein, axillary vein, and superficial femoral vein, respectively. Perioperative outcomes. Technical success was achieved in all 32 cases; 29 (91%) GHVGs have been successfully used for dialysis, whereas 3 (9%) have failed
Journal of Vascular Surgery Volume
-,
Number
Benedetto et al
Table I. Demographics and baseline before propensity score matching (PSM)
Patients Age, years Vein diameter, mm Previous surgery Male
characteristics
Standard PTFE graft
32
43
70 6 11
71 6 8
.75
5.8 6 1.0
5.5 6 0.9
.15
Vein diameter, mm
261
161
.27
Previous surgery
14 (43.8)
32 (74.4)
Male
P value Patients Age, years
<.01 .95
Brachial-axillary
22 (68.8)
Brachial-basilic
5 (15.6)
11 (25.6)
Femoral-femoral
1 (3.1)
Axilloaxillary
1 (3.1)
Brachial-antecubital
Table III. Demographics and baseline characteristics after propensity score matching (PSM)
GHVG
Graft type
GHVG
Standard PTFE graft
25
25
71 6 11
72 6 9
.77
5.8 6 1.0
5.6 6 0.8
.17
261
261
13 (52)
14 (56)
P value
.84 1
Graft type
28 (65.1)
.79
Brachial-axillary
18 (72)
13 (52)
Brachial-basilic
3 (12)
8 (32)
1 (2.3)
Femoral-femoral
0 (0)
1 (4)
2 (4.7)
Axilloaxillary
1 (4)
2 (8)
3 (9.4)
1 (2.3)
3 (12)
1 (4)
Hypertension
21 (65.6)
25 (58.1)
.51
Hypertension
15 (60)
17 (68)
Diabetes
18 (56.3)
28 (65.1)
.44
Diabetes
14 (56)
13 (52)
1
Coronary artery disease
10 (31.3)
11 (25.6)
.59
Coronary artery disease
9 (36)
8 (32)
1
.65
Renal failure type
28 (87.5)
36 (83.7)
22 (88)
20 (80)
Renal failure type Dialysis Uremic state COPD
4 (12.5)
7 (16.3)
7 (21.9)
14 (32.6)
Brachial-antecubital
Dialysis Uremic state .31
COPD
COPD, Chronic obstructive pulmonary disease; GHVG, Gore Hybrid Vascular Graft; PTFE, polytetrafluoroethylene. Categorical variables are presented as count (%). Continuous data are presented as mean 6 standard deviation.
Table II. Nitinol-reinforced section (NRS) Size, mm
No.
5 50
1
6 100
3
6 50
9
7 100
5
-
1
7 50
10
8 50
5
9 50
3
before starting cannulation. Mean time to cannulation was 20 6 6 days (range, 13-37; median, 20). Early complications (within 30 days) were as follows: surgical site hematoma in three cases (9%) of brachial-axillary configuration, none of which required surgical revision; limb edema in one patient (3%) due to a focal stenosis of superior vena cava that was addressed with a balloon-expandable bare-metal stent; acute lower limb ischemia in one patient (3%) with femoral-femoral configuration, successfully treated with thrombectomy; and graft infection in two cases (6%), treated with graft explantation and central venous catheter placement. Two patients (6%), both in poor general condition, died in the hospital. Follow-up. Three patients were lost to follow-up. The mean follow-up time was 15 6 11 months (range,
.77
.68 3 (12)
5 (20)
7 (28)
8 (32)
1
COPD, Chronic obstructive pulmonary disease; GHVG, Gore Hybrid Vascular Graft; PTFE, polytetrafluoroethylene. Categorical variables are presented as count (%). Continuous data are presented as mean 6 standard deviation.
0-33 months; median, 15.5). During follow-up, two other patients died of unrelated causes. During follow-up, two outflow vein stenoses were detected and treated with covered stent implantation and simple balloon angioplasty, respectively. Six occlusions occurred and were treated with simple thrombectomy, except for three cases that required adjunctive endovascular procedures on the outflow vein (two stenting and one balloon angioplasty). One patient presented with graft infection at 7 months and underwent graft explantation. A histologic analysis revealed IH proximal to the end of the NRS. No steal syndrome was observed during follow-up. Mean access flow was 1202 6 436 mL/min (range, 650-2200; median, 1150). After PSM, two groups of 25 patients each were created with correction of the differences in demographics and baseline characteristics (Table III). Twoyear survival estimates for hybrid and standard PTFE groups were 91% 6 6% and 82 6 9% (P > .05), respectively (Fig 3). One-year functional primary, functional assisted primary, and functional secondary patency estimates for the hybrid group vs the standard PTFE group were 66% 6 10% vs 51% 6 10% (P > .05), 75% 6 9% vs 51% 6 10% (P > .05), and 79% 6 9% vs 67% 6 10% (P > .05; Figs 4 to 6). The probability of functional primary, assisted primary, and secondary patency during 24 months for the GHVG group was 40%, 50%, and 38% more than for the PTFE group. However, no statistical
6
Journal of Vascular Surgery
Benedetto et al
---
Fig 3. Kaplan-Meier plot showing survival of Gore Hybrid Vascular Graft (GHVG) vs standard polytetrafluoroethylene (PTFE). Standard error was <10% at all time points.
Fig 4. Kaplan-Meier plot showing functional primary patency estimates of Gore Hybrid Vascular Graft (GHVG) vs standard polytetrafluoroethylene (PTFE). Curves exceeding 10% standard error are dashed.
significance was reached (Table IV). Secondary outcomes showed no statistically significant difference between the groups. The logistic regression model was statistically significant, c2 (3) ¼ 10.984 (P < .001). The model explained 43.0% (Nagelkerke R2) of the variance of GHVG-related reintervention and correctly classified 87.5% of cases. Of the three predictor variables (target vein diameter, previous vascular access surgery on the same limb, and diabetes), only one was statistically significant: target vein diameter. Reducing target vein diameter was associated with an increased likelihood of GHVG-related reintervention (Table V).
DISCUSSION AVF is currently the first choice in patients requiring access for hemodialysis treatment. An AVG is required when native vein anatomy is deemed disadvantaged or after repeated AVF failures. Reported 1-year primary
2016
Fig 5. Kaplan-Meier plot showing functional assisted primary patency estimates of Gore Hybrid Vascular Graft (GHVG) vs standard polytetrafluoroethylene (PTFE). Curves exceeding 10% standard error are dashed.
Fig 6. Kaplan-Meier plot showing functional secondary patency estimates of Gore Hybrid Vascular Graft (GHVG) vs standard polytetrafluoroethylene (PTFE). Curves exceeding 10% standard error are dashed.
Table IV. Effect of Gore Hybrid Vascular Graft (GHVG) implantation on functional patency rates during 24 months Hazard ratio (95% CI)
P value
Primary
0.595 (0.25-1.4)
.23
Primary assisted
0.498 (0.20-1.23)
.13
Secondary
0.625 (0.22-1.72)
.36
Functional patency
CI, Confidence interval.
and secondary patency rates for standard PTFE AVGs range from 40% to 60% and from 50% to 70%, respectively.21-24 In this study, we report on 32 patients treated with GHVG, with 1-year functional primary and secondary patency rates of 66% 6 10% and 79% 6 9%, respectively. Although the usefulness of the AVG is undeniable, its Achilles heel
Journal of Vascular Surgery Volume
-,
Number
Benedetto et al
7
-
Table V. Predictors of Gore Hybrid Vascular Graft (GHVG)-related reintervention
Target vein diameter
B
SE
Wald
df
1.997
.85
5.51
1
.02
P
0.14 (0.03-0.72)
Odds ratio (95% CI)
Previous vascular access
.705
.51
1.88
1
.17
2.02 (0.74-5.54)
Diabetes
.517
1.05
.24
1
.62
1.68 (0.22-13.07)
Constant
8.478
4.33
3.83
1
.050
4807.47
B, Regression coefficient; CI, confidence interval; df, degrees of freedom; SE, standard error of coefficient; Wald, Wald c2.
is the venous anastomosis. The main cause of failure of AVGs is considered to be stenosis at the venous anastomosis and subsequent thrombosis. This stenosis is mainly due to IH, which is triggered by turbulent flow, shear stress, and compliance mismatch between graft and vein.10,25 Turbulence is a well-established factor in the development of IH. Fillinger et al26 demonstrated a significant correlation between Reynolds number and intimalmedial thickening at the venous anastomosis. Sivanesan et al27 showed that the value of perianastomotic wall shear stress falls within the range that can cause endothelial injury (35-40 N/m). Fluid dynamics and geometry of the anastomosis have been addressed by several in vitro and in vivo models in an effort to understand and to prevent the development of IH.28-30 The influence of the anastomotic angle on the flow was studied using a porcine model.5 This consisted of an aortic graft interposition with end-to-side configuration. Distal anastomoses were performed with angles of 90, 45, or 15 degrees. The anastomoses of both 90 and 45 degrees showed a recirculation zone, whereas no turbulence of the flow was documented in the 15-degree ones. Hakaim et al31 showed that a sharp anastomotic angle, along with the incorporation of a flow diffuser, increased patency in brachial-axillary AVGs compared with a control group matched for age and gender. The hallmark of GHVG is the absence of a sewn venous anastomosis, replaced by a 5-cm or 10-cm NRS. This feature, besides offering a technically easier surgical approach to the target vein, allows a straight venous anastomotic junction, which might theoretically reduce the amount of turbulent flow. Nonetheless, the finding of IH is not totally excluded, as attested to in the single histologic analysis available. More anatomic samples are necessary for any conclusion to be drawn in this regard. The duration of intervention is comparable to that of standard AVG implantation. Time spared by sutureless anastomosis offsets the additional angiographic time. Clopidogrel might account for our three hematomas. However, it was reported to reduce AVF thrombosis without increasing bleeding events.32 Few reports on the use of GHVG are present in the literature to date. Anaya-Ayala et al33 reported on a series of 25 patients who received GHVG for hemodialysis access, with 1-year primary and secondary patency of 66% 6 8%
and 69% 6 7%, respectively. We observed a slightly higher secondary patency in our series. However, the selection criteria adopted by the two surgical groups are different. Indeed, Anaya-Ayala et al included only patients with a target vein <3 mm in diameter, whereas we excluded these patients and considered eligible only patients with a target vein at least 4 mm in diameter. Unlike Anaya-Ayala et al, we did not use the percutaneous approach, but we exposed the target vein in all cases. Moreover, our group implanted the graft in a wider variety of configurations compared with Anaya-Ayala et al, who reported exclusively brachial-axillary and axilloaxillary loop configurations. In our series, we observed five stenoses of the outflow vein, detected after AVG malfunction or occlusion. Most of the stenoses were located in the axillary or subclavian vein, next to a venous valve, and were mainly treated by positioning a Viabahn stent (W. L. Gore & Associates). The treatment of outflow vein stenoses with covered stents was previously described for standard AVGs, both as a planned procedure and during reintervention.33-35 Since the availability of a graft with 10-cm NRS, we started to perform intraoperative angiography routinely before the selection of the graft.11 Indeed, in our current practice, starting from the latest cases, the finding of a venous valve close to the expected landing zone of the NRS is a criterion to prefer the longer length. The 5-cm length, dominant in our series, was the only one used during the first period of our experience. As argued by Anaya-Ayala et al, the sutureless venous anastomosis may give rise to additional difficulties in case of reintervention.33 In our experience, of six cases of thrombectomy, one case required landing zone balloon angioplasty and two cases required angioplasty and stenting as adjunctive procedures, as a venous stenosis was present proximal to the end of the graft. Actually, most thrombectomies failed to result in a satisfactory lengthening of hybrid access survival as reocclusion occurred within a few weeks in four cases. The reason for this unsatisfactory outcome was unclear, and we believe a deterioration of proximal outflow may have played a role. On the other hand, maintenance procedures, carried out for access malfunction with a patent graft, were successful and were not followed by further reinterventions or occlusion. For this reason, in our opinion, a strict compliance with the follow-up controls is critical for this type of graft.
8
Benedetto et al
Journal of Vascular Surgery ---
Our multivariate analysis provided an insight into possible strategies to improve outcomes. In effect, it has shown that a reduction of target vein diameter was associated with an increased likelihood of GHVGrelated reintervention. We acknowledge that this finding is limited because of the small size of the sample and needs to be confirmed with further studies. Nonetheless, in our opinion, additional caution is prompted in the selection of the target vein whenever the diameter is in the lower end of the recommended range. In these cases, a possible solution could be to achieve a more proximal and thus larger in diameter landing zone by selecting the 10-cm NRS. This study presents some important limitations. The sample is heterogeneous in graft configurations, its size is limited, and the follow-up is relatively short. Larger cohorts of patients, a longer observation period, and possibly a multicenter randomized trial may help in defining the role for GHVG. A multicenter observational study, including a greater number of participating centers experienced with GHVG, is currently being conducted under the guidance of the first author. Whether the in-line graft-vein junction effectively produces laminar flow, in patients’ anatomies, has not yet been proved. In silico and in vitro models could be designed to better understand the fluid dynamics of real-anatomy geometries.
CONCLUSIONS The GHVG is a safe and effective alternative to standard PTFE AVG. Based on our results, a careful planning of the landing zone is advisable when the target vein is small because a decrease in vein diameter was associated with higher risk of reintervention. Larger studies and randomized trials are needed to better delineate the indications for GHVG.
AUTHOR CONTRIBUTIONS Conception and design: FB, GG, AD, FS Analysis and interpretation: DS, NP, AN, SV Data collection: DS, NP, GG, SV Writing the article: DS, NP, AN, SV Critical revision of the article: FB, DS, NP, GG, AN, AD, FS Final approval of the article: FB, DS, NP, GG, AN, SV, AD, FS Statistical analysis: DS, NP, AN Obtained funding: Not applicable Overall responsibility: FB
REFERENCES 1. Akoh JA. Prosthetic arteriovenous grafts for hemodialysis. J Vasc Access 2009;10:137-47. 2. Palder SB, Kirkman RL, Whittemore AD, Hakim RM, Lazarus JM, Tilney NL. Vascular access for hemodialysis. Patency rates and results of revision. Ann Surg 1985;202: 235-9. 3. Besrab A, Work J, Brouwer D. Clinical practice guidelines for vascular access. Am J Kidney Dis 2006;48(Suppl 1):S176-247.
2016
4. Huber TS, Carter JW, Carter RL, Seeger JM. Patency of autogenous and PTFE upper extremity arteriovenous hemodialysis accesses: a systematic review. J Vasc Surg 2003;39:491-6. 5. Staalsen NH, Ulrich M, Winther J, Pedersen EM, How T, Nygaard H. The anastomosis angle does change the flow fields at vascular end-to-side anastomoses in vivo. J Vasc Surg 1995;21:460-71. 6. Hofstra L, Bergmans DC, Leunissen KM, Hoeks AP, Kitslaar PJ, Daemen MJ, et al. Anastomotic intimal hyperplasia in prosthetic arteriovenous fistulas for hemodialysis is associated with initial high flow velocity and not with mismatch in elastic properties. J Am Soc Nephrol 1995;6:1625-33. 7. Hofstra L, Bergmans DC, Leunissen KM, Hoeks AP, Kitslaar PJ, Tordoir JH. Prosthetic arteriovenous fistulas and venous anastomotic stenosis: influence of a high flow velocity on the development of intimal hyperplasia. Blood Purificat 1996;14: 345-9. 8. Roy-Chaudhury P, Kelly BS, Narayana A, Desai P, Melhem M, Munda R, et al. Hemodialysis vascular access dysfunction from basic biology to clinical intervention. Adv Ren Replace Ther 2002;9:74-84. 9. Lemson MS, Tordoir JH, Daemen MJ, Kitslaar PJ. Intimal hyperplasia in vascular grafts. Eur J Vasc Endovasc Surg 2000;19:336-51. 10. Hofstra L, Tordoir JH, Kitslaar PJ, Hoeks AP, Daemen MJ. Enhanced cellular proliferation in intact stenotic lesions derived from human arteriovenous fistulas and peripheral bypass grafts. Does it correlate with flow parameters? Circulation 1996;94:1283-90. 11. Jones RG, Inston NG. Commentary: the hybrid vascular graft: a new frontier in arteriovenous access for dialysis or just another graft? J Endovasc Ther 2015;22:786-8. 12. Haskal ZJ, Trerotola S, Dolmatch B, Schuman E, Altman S, Mietling S, et al. Stent graft versus balloon angioplasty for failing dialysis-access grafts. N Engl J Med 2010;362: 494-503. 13. Lachat M, Mayer D, Criado FJ, Pfammatter T, Rancic Z, Genoni M, et al. New technique to facilitate renal revascularization with use of telescoping self-expanding stent grafts: VORTEC. Vascular 2008;16:69-72. 14. Kim JJ, Koopmann M, Ihenachor E, Zeng A, Ryan T. The addition of ultrasound arterial examination to upper extremity vein mapping before hemodialysis access. Ann Vasc Surg 2016;33:109-15. 15. Hemodialysis Adequacy 2006 Work Group. Clinical practice guidelines for hemodialysis adequacy, update 2006. Am J Kidney Dis 2006;48(Suppl 1):S2-90. 16. Chiesa R, Kahlberg A, Mascia D, Tshomba Y, Civilini E, Melissano G. Use of a novel hybrid vascular graft for sutureless revascularization of the renal arteries during open thoracoabdominal aortic aneurysm repair. J Vasc Surg 2014;60: 622-30. 17. Wuamett JC, Glickman MH. Arteriovenous graft configuration in hemodialysis: does it matter? J Vasc Access 2015;16: 78-81. 18. Jiménez-Almonacid P, Gruss E, Jimenez-Toscano M, Jiménez-Toscano M, Lasala M, Rueda JA, et al. Multidisciplinary approach to hemodialysis graft dysfunction and thrombosis. Nefrologia 2013;33:692-8. 19. Sidawy AN, Gray R, Besarab A, Henry M, Ascher E, Silva M, et al. Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg 2002;35: 603-10. 20. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41-55.
Journal of Vascular Surgery Volume
-,
Number
Benedetto et al
9
-
21. Katzman HE, Glickman MH, Schild AF, Fujitani RM, Lawson JH. Multicenter evaluation of the bovine mesenteric vein bioprostheses for hemodialysis access in patients with an earlier failed prosthetic graft. J Am Coll Surg 2005;201:223-30. 22. 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. 23. Gibson KD, Gillen DL, Caps MT, Kohler TR, Sherrard DJ, Stehman-Breen CO. Vascular access survival and incidence of revisions: a comparison of prosthetic grafts, simple autogenous fistulas, and venous transposition fistulas from the United States Renal Data System Dialysis Morbidity and Mortality Study. J Vasc Surg 2001;34:694-700. 24. Hodges TC, Fillinger MF, Zwolak RM, Walsh DB, Bech F, Cronenwett JL. Longitudinal comparison of dialysis access methods: risk factors for failure. J Vasc Surg 1997;26:1009-19. 25. Loth F, Jones SA, Zarins CK, Giddens DP, Nassar RF, Glagov S, et al. Relative contribution of wall shear stress and injury in experimental intimal thickening at PTFE end-to-side arterial anastomoses. J Biomech Eng 2002;124:44-51. 26. Fillinger MF, Reinitz ER, Schwartz RA, Resetarits DE, Paskanik AM, Bredenberg CE. Beneficial effects of banding on venous intimal-medial hyperplasia in arteriovenous loop grafts. Am J Surg 1989;158:87-94. 27. Sivanesan S, How TV, Black RA, Bakran A. Flow patterns in the radiocephalic arteriovenous fistula: an in vitro study. J Biomech 1999;32:915-25. 28. Heise M, Schmidt S, Krüger U, Pfitzmann R, Scholz H, Neuhaus P, et al. Local haemodynamics and shear stress in
29.
30.
31.
32.
33.
34.
35.
cuffed and straight PTFE-venous anastomoses: an in-vitro comparison using particle image velocimetry. Eur J Vasc Endovasc Surg 2003;26:367-73. Krueger U, Huhle A, Krys K, Scholz H. Effect of tapered grafts on hemodynamics and flow rate in dialysis access grafts. Artif Organs 2004;28:623-8. Hughes PE, How TV. Effects of geometry and flow division on flow structures in models of the distal end-to-side anastomosis. J Biomech 1996;29:855-72. Hakaim AG, Nalbandian MN, Heller JK, Chowla AC, Oldenburg WA. Improved patency of prosthetic arteriovenous grafts with an acute anastomotic angle and flow diffuser. J Vasc Surg 2003;37:1032-5. Dember LM, Beck GJ, Allon M, Delmez JA, Dixon BS, Greenberg A, et al. Effect of clopidogrel on early failure of arteriovenous fistulas for hemodialysis: a randomized controlled trial. JAMA 2008;299:2164-71. Anaya-Ayala JE, Davies MG, El-Sayed HF, Peden EK, Naoum JJ. Early experience with a novel hybrid vascular graft for hemodialysis access creation in patients with disadvantaged anatomy. J Endovasc Ther 2015;22:778-85. Vogel PM, Parise C. Comparison of SMART stent placement for arteriovenous graft salvage versus successful graft PTA. J Vasc Interv Radiol 2005;16:1619-26. Carmona J, Rits Y, Jones B, Dowers L, Bednarski D, Rubin JR. Patency of the Viabahn stent graft for the treatment of outflow stenosis in hemodialysis grafts. Am J Surg 2016;211:551-4.
Submitted Mar 19, 2016; accepted Jul 26, 2016.
ABSTRACT Objective: The purpose of this study was to examine the outcomes of a vascular hybrid polytetrafluoroethylene (PTFE) graft, provided with a nitinol-reinforced section (NRS) on one end, in hemodialysis vascular access placement. Methods: A retrospective study was conducted including all the consecutive patients who underwent Gore Hybrid Vascular Graft (GHVG; W. L. Gore & Associates, Flagstaff, Ariz) implantation for hemodialysis access placement between October 2013 and November 2015. A propensity-matched control group was obtained from consecutive patients who underwent standard PTFE arteriovenous graft implantation between January 2010 and July 2013. The selection criteria were inadequate venous material for autogenous arteriovenous fistula placement, patent deep venous circulation, and vein diameter of 4 to 8.5 mm. The implantation technique involves the insertion of the NRS some centimeters into the target vein. Fluoroscopic guidance helps deploy the device in the desired landing zone (ie, position of the proximal end of the NRS), based on anatomic landmarks. Survival, functional patency rates, and complications were compared with a propensity-matched historical control group. Vein diameter, previous vascular access placement, and diabetes were tested as predictors of reintervention with a logistic regression analysis. Results: There were 32 patients (14 men; mean age, 69 6 14 years) who received the GHVG graft. The historical control group included 43 patients. Technical success was 100%. The graft configuration was brachial-axillary (n ¼ 22 [69%]), brachial-basilic loop (n ¼ 5 [16%]), brachial-antecubital loop (n ¼ 3 [9%]), axilloaxillary loop (n ¼ 1 [3%]), and femoralfemoral loop (n ¼ 1 [3%]). Mean NRS oversize was 20% 6 7% (range, 3%-34%; median, 19%). Perioperative complications requiring revision included acute limb ischemia treated with thrombectomy (n ¼ 1 [3%]) and graft infection requiring explantation (n ¼ 2 [6%]). Two patients (6%) died in the hospital of unrelated causes. The mean follow-up was 15 6 11 months (range, 0-33 months; median, 15.5). The propensity-matched groups included 25 patients each. Survival estimates at 24 months for the GHVG and standard PTFE groups were 91% 6 6% and 82% 6 9% (P > .05), respectively. The 12-month patency estimates were as follows: functional primary patency, 66% 6 10% vs 51% 6 10% (P > .05); functional assisted primary patency, 75% 6 9% vs 51% 6 10% (P > .05); and functional secondary patency, 79% 6 9% vs 67% 6 10% (P > .05). Reduction in vein diameter was associated with reintervention. Conclusions: The GHVG is a safe and effective alternative to standard PTFE in hemodialysis access surgery. Careful planning for the landing zone is advisable, especially for small outflow veins. Larger studies and randomized trials are needed to define the role for this device. A study including a greater number of centers experienced with this device is currently under way. (J Vasc Surg 2016;-:1-9.)
The life expectancy of hemodialysis-dependent patients has been increasing in the last few decades.1 Autogenous arteriovenous fistula (AVF) is the “gold From the Department of Biomedical Sciences and Morphological and Functional Imaginga and Department of Human Pathology of Adult and Developmental Age “Gaetano Barresi”,c University of Messina, Messina; the Vascular Surgery Unit, Noto-Pasqualino Clinic, Palermob; and the Vascular Surgery Unit, Campus Bio-Medico University, Rome.d Author conflict of interest: none. Correspondence: Domenico Spinelli, MD, Department of Biomedical Sciences and Morphological and Functional Imaging, University of Messina, Via Consolare Valeria 1, Messina 98125, Italy (e-mail: [email protected]). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2016.07.133
standard” for hemodialysis access for its good longterm patency.2,3 After repeated failure of an AVF and exhaustion of venous material, the use of a prosthetic arteriovenous graft (AVG) is the best alternative. Many types of prostheses have been used so far, with 1-year patency rates generally lower than for AVF. A review including 34 studies reported a cumulative 18-month primary patency of 51% for AVF and 33% for AVG.4 Intimal hyperplasia (IH) is considered to be the main cause of graft occlusion.1 The increasing need for long-lasting vascular accesses has fostered the research on IH. The turbulent flow, strictly related to the width of the anastomosis angle, produces shear stress on the vein wall.5 The subsequent endothelial injury promotes smooth muscle cell migration and proliferation and eventually is responsible for platelet recruitment and coagulation cascade activation.6-10 The next step is graft thrombosis. In light 1
2
Benedetto et al
Journal of Vascular Surgery ---
of these findings, a new concept of vascular hybrid prosthesis has been developed, whereby the vein-side anastomosis is replaced by a nitinol-reinforced section (NRS), to be inserted and deployed into the target vein, to form an end-to-end junction. This design, based on a straight geometry, is intended to keep the flow laminar, reducing shear stress and thus preventing IH.11 Nonetheless, it is still unclear whether the straight geometry actually allows laminar flow in AVGs. Evidence that a nitinol stent at the venous end of a standard polytetrafluoroethylene (PTFE) graft improved patency of a failing AVG compared with simple angioplasty supports the validity of the hybrid concept.12 The idea of replacing the distal anastomosis of a graft by a nitinol-covered stent was also described by Lachat et al.13 The aim of the study was to report a 2-year experience with the hybrid graft in hemodialysis access and to analyze outcomes and complications.
METHODS Patient cohort. A multicenter retrospective cohort study was conducted including all consecutive patients who underwent implantation of the Gore Hybrid Vascular Graft (GHVG; W. L. Gore & Associates, Flagstaff, Ariz) for hemodialysis access between October 2013 and November 2015 at the University Hospital Gaetano Martino, Messina, Italy, and at Noto-Pasqualino Clinic, Palermo, Italy. A propensity-matched control group was obtained from consecutive patients who underwent standard PTFE AVG implantation between January 2010 and July 2013 at the same centers. Patient demographics and clinical and procedure data were collected in a dedicated prospectively maintained Excel (Microsoft, Redmond, Wash) database. This study is in agreement with the principles outlined in the Declaration of Helsinki. The protocol and the informed consent were approved by the Institutional Review Board. All patients gave a written informed consent to participate in a study. This study did not undergo editorial review, nor was it supported by representatives of the graft manufacturer. Preoperative protocol. All patients underwent preoperative routine blood tests and careful physical examination, evaluating the presence and validity of distal pulses, skin inflammation, and any evidence of central venous obstruction, such as edema and collateral veins of the upper extremity or chest wall. Allen test was routinely performed. Duplex ultrasound (Vivid 4; GE Healthcare, Chalfont St Giles, United Kingdom) arterial and venous preoperative mapping was routinely performed. Target artery and vein diameters were measured on the transverse ultrasound scan, with the patient supine at room temperature, by experienced operators. Vein compressibility, distensibility, thickness, and continuity were assessed with a rubber tourniquet in place. Arterial patency and presence of triphasic flow were
2016
ascertained by duplex ultrasound with a 7 MHz linear probe at 60-degree insonation angle. Patients were deemed unsuitable for autogenous access placement if a superficial vein, including the upper arm cephalic and basilic vein, at least 2.5 mm in diameter was lacking or was present for a short track before the confluence into the deep veins, in both the dominant and nondominant arms.14 Antecubital veins, basilic veins, deep veins at the elbow, axillary veins, and femoral veins were explored and considered a possible target vein if unsuitable for AVF placement. When central vein stenosis or occlusion was suspected or in patients known to have had a previous central vein catheter or pacemaker, venography or magnetic resonance angiography was performed.15 A complete cardiac evaluation, including echocardiography, was part of the preoperative protocol. The selection criteria were unsuitability for AVF placement, presence of patent deep venous circulation, appropriate diameters of inflow artery (3 mm) and target vein (4-8.5 mm), absence of local and systemic infections, and absence of upper limb ischemia. If the anatomic requirements were met, the GHVG was preferred to a traditional PTFE graft. Device characteristics. The GHVG is approved by the U.S. Food and Drug Administration and the European Union (CE Mark). It is commercially available in the United States, Europe, Russia, and a few Middle Eastern countries.16 It consists of a PTFE graft with an NRS at one end. The continuous luminal surface provides additional resistance to thrombosis and seroma formation thanks to covalently bonded heparin of porcine origin and a low permeability film. The NRS is conceived to be inserted directly through a venotomy or over the wire through a peel-away sheath into the target vessel, to be deployed so as to replace the venous anastomosis with an end-to-end junction. The deployment is carried out by pulling a constraining wire. The endoluminal position of the proximal end of the NRS is defined as the landing zone. The idea behind this design is to reduce turbulent flow to prevent IH and, at the same time, to allow the treatment of stenoses of the target vein.11 The NRS is available in five diameters (5-9 mm) and two lengths (5 and 10 cm). The nonreinforced end is anastomosed to the inflow artery with a running suture like a standard AVG. It has standard 6-mm diameter and is tapered to adapt to the size of the NRS. It uses stretch technology and should be put under tension before placement of the suture. Technique. Four surgeons performed the procedures. All procedures were carried out in a surgical room equipped with a mobile fluoroscopic C-arm (Philips BV Pulsera; Philips Medical Systems, Amsterdam, The Netherlands) because contrast angiography is required. Vascular accesses were placed in both the upper and lower limbs. The selection of graft size and access
Journal of Vascular Surgery Volume
-,
Number
Benedetto et al
3
-
Fig 1. Brachial-axillary hybrid arteriovenous graft (AVG). A, Venous anastomosis between nitinol-reinforced section (NRS) and axillary vein. B, Angiogram.
configuration was made by the surgeon.17 Because the graft by design occludes the distal venous outflow, the NRS was placed as distal as reasonable, based on vein diameter, aiming to safeguard the chance for future access placements and preserving the confluence of venous branches whenever possible. Optimal NRS oversize was considered to be 5% to 20%, as suggested in the instructions for use. The technique of brachial-axillary graft implantation involved the surgical incision in the armpit with dissection of the axillary vessels. The brachial artery was then exposed through an incision proximal to the elbow. The nonstented part of the device was stretched and tunneled in the subcutaneous tissue. The NRS was inserted into the axillary vein through a venotomy or a peel-away sheath. The NRS was then advanced inside the vein except for about 2 cm, deployed, and secured to the vein with two 6-0 polypropylene stitches (Fig 1, A). These sutures, required per instructions for use, prevent graft dislocation during the procedure. The arterial anastomosis was performed in an end-to-side fashion, as for traditional grafts. As a final step, a balloon dilation of the stent was performed through direct puncture of the prosthesis. Any additional endovascular procedure addressing outflow stenoses was performed at this time. Alternatively, if doubt persisted about graft sizing, angiography was performed before the selection of the GHVG. Completion control was routinely performed at the end of the procedure by contrast angiography when it was not contraindicated (Fig 1, B). Patients not yet under hemodialysis were controlled with duplex ultrasound only. For the brachialantecubital forearm loop, the brachial artery and the antecubital vein were exposed through a single surgical incision. The prosthesis was then tunneled in a loop
configuration. A comparable approach has been used for femoral-femoral looped inguinal access placement. Through an inguinal incision, the femoral vessels were exposed, and the superficial femoral vein served as target vein (Fig 2). For axilloaxillary loop AVG, the axillary vessels were exposed through an axillary incision. All procedures were performed under local anesthesia (mepivacaine hydrochloride 2%), locoregional plexus block, or spinal anesthesia, except axilloaxillary loop AVG, which was performed under general anesthesia. Antibiotic prophylaxis with cephalosporin was performed before the surgical incision. Postoperative management. Patients were discharged with antiplatelet therapy (clopidogrel, 75 mg/d). Required recovery time before cannulation was 2 to 4 weeks. The vascular access was evaluated with duplex ultrasound through hospital discharge, at 30 days, and every 3 months after the procedure. Flow measurements were performed monthly. When dynamic venous pressure >200 mm Hg with flow of 300 mL/min, Kt/V <0.8, or recirculation >25% was present, duplex ultrasound or angiographic control was performed.18 Outcomes and statistical analysis. Interventions, complications, and outcomes were defined according to the reporting standards on arteriovenous accesses developed by the Society for Vascular Surgery and the American Association for Vascular Surgery.19 Primary end points were survival, functional primary patency, assisted primary patency, and secondary patency. Secondary end points were vascular access-related complications: postoperative hematoma, limb edema, graft infection, anastomosis or venous stenosis, steal syndrome, and other complications. Continuous variables are reported as
4
Benedetto et al
Journal of Vascular Surgery ---
2016
Fig 2. Femoral-femoral loop arteriovenous graft (AVG). A, Arterial and venous anastomoses are carried out on superficial femoral vessels. B, Angiogram.
mean 6 standard deviation or median with range. Categorical variables are reported as count (percentage). Statistical analysis was performed using SPSS software (IBM Corp, Armonk, NY). Patients in the GHVG group were matched to patients in the standard PTFE group, using the propensity score matching (PSM) methodology of Rosenbaum and Rubin.20 Our chosen methodology used caliper matching on the propensity score without replacement to create a balanced data set. The following variables were used for PSM: AVG configuration, sex, age, hypertension, diabetes, coronary artery disease, chronic obstructive pulmonary disease, renal failure stage, vein diameter, and previous vascular access surgery. The analyses of matched data sets were conducted with Wilcoxon signed rank test for continuous variables and McNemar test for dichotomous ones. Functional patency rates were analyzed with Kaplan-Meier curve with Cox proportional hazards model stratified on matched pairs. A logistic regression model was performed to ascertain the effects of target vein diameter, previous vascular access surgery on the same limb, and diabetes on the likelihood of GHVG-related reintervention.
RESULTS Patient population. Thirty-two patients received a GHVG for hemodialysis access placement. The historical control group included 43 patients. Demographics and baseline characteristics for both groups are shown in Table I. Among patients who received a GHVG, 28 patients (88%) were under hemodialysis treatment for end-stage renal disease, and in 4 patients (13%), the
anticipated start of hemodialysis treatment was within 6 weeks. Four patients (13%) received the GHVG as a first choice because they lacked superficial veins >2.5 mm in diameter and long enough for an effective AVF to be placed. The remaining had at least one previous intervention of vascular access placement. The mean target vein diameter was 5.8 6 1.0 mm (range, 3.8-8.0 mm; median, 5.8 mm). Treatment characteristics. The mean duration of interventions was 78 6 27 minutes (range, 40-155; median, 70). The mean amount of contrast agent used was 30 6 8 mL, (range, 20-54; median, 28). The mean NRS oversize was 20% 6 7% (range, 3.4%-34.6%; median, 19.7%). The NRS size of the implanted GHVG ranged from 5 to 9 mm in diameter, and both lengths were used (Table II). The graft configuration was brachialaxillary, brachial-basilic loop, brachial-antecubital loop, axilloaxillary loop, and femoral-femoral loop in 22 cases (69%), 5 cases (16%), 3 cases (9%), 1 case (3%), and 1 case (3%), respectively. In six cases, additional planned endovascular procedures were carried out to treat stenoses of the outflow vein or central veins: three cases were treated with a covered stent in the right innominate vein, axillary vein, and brachial vein, respectively; three cases were treated with simple balloon angioplasty of the right innominate vein, axillary vein, and superficial femoral vein, respectively. Perioperative outcomes. Technical success was achieved in all 32 cases; 29 (91%) GHVGs have been successfully used for dialysis, whereas 3 (9%) have failed
Journal of Vascular Surgery Volume
-,
Number
Benedetto et al
Table I. Demographics and baseline before propensity score matching (PSM)
Patients Age, years Vein diameter, mm Previous surgery Male
characteristics
Standard PTFE graft
32
43
70 6 11
71 6 8
.75
5.8 6 1.0
5.5 6 0.9
.15
Vein diameter, mm
261
161
.27
Previous surgery
14 (43.8)
32 (74.4)
Male
P value Patients Age, years
<.01 .95
Brachial-axillary
22 (68.8)
Brachial-basilic
5 (15.6)
11 (25.6)
Femoral-femoral
1 (3.1)
Axilloaxillary
1 (3.1)
Brachial-antecubital
Table III. Demographics and baseline characteristics after propensity score matching (PSM)
GHVG
Graft type
GHVG
Standard PTFE graft
25
25
71 6 11
72 6 9
.77
5.8 6 1.0
5.6 6 0.8
.17
261
261
13 (52)
14 (56)
P value
.84 1
Graft type
28 (65.1)
.79
Brachial-axillary
18 (72)
13 (52)
Brachial-basilic
3 (12)
8 (32)
1 (2.3)
Femoral-femoral
0 (0)
1 (4)
2 (4.7)
Axilloaxillary
1 (4)
2 (8)
3 (9.4)
1 (2.3)
3 (12)
1 (4)
Hypertension
21 (65.6)
25 (58.1)
.51
Hypertension
15 (60)
17 (68)
Diabetes
18 (56.3)
28 (65.1)
.44
Diabetes
14 (56)
13 (52)
1
Coronary artery disease
10 (31.3)
11 (25.6)
.59
Coronary artery disease
9 (36)
8 (32)
1
.65
Renal failure type
28 (87.5)
36 (83.7)
22 (88)
20 (80)
Renal failure type Dialysis Uremic state COPD
4 (12.5)
7 (16.3)
7 (21.9)
14 (32.6)
Brachial-antecubital
Dialysis Uremic state .31
COPD
COPD, Chronic obstructive pulmonary disease; GHVG, Gore Hybrid Vascular Graft; PTFE, polytetrafluoroethylene. Categorical variables are presented as count (%). Continuous data are presented as mean 6 standard deviation.
Table II. Nitinol-reinforced section (NRS) Size, mm
No.
5 50
1
6 100
3
6 50
9
7 100
5
-
1
7 50
10
8 50
5
9 50
3
before starting cannulation. Mean time to cannulation was 20 6 6 days (range, 13-37; median, 20). Early complications (within 30 days) were as follows: surgical site hematoma in three cases (9%) of brachial-axillary configuration, none of which required surgical revision; limb edema in one patient (3%) due to a focal stenosis of superior vena cava that was addressed with a balloon-expandable bare-metal stent; acute lower limb ischemia in one patient (3%) with femoral-femoral configuration, successfully treated with thrombectomy; and graft infection in two cases (6%), treated with graft explantation and central venous catheter placement. Two patients (6%), both in poor general condition, died in the hospital. Follow-up. Three patients were lost to follow-up. The mean follow-up time was 15 6 11 months (range,
.77
.68 3 (12)
5 (20)
7 (28)
8 (32)
1
COPD, Chronic obstructive pulmonary disease; GHVG, Gore Hybrid Vascular Graft; PTFE, polytetrafluoroethylene. Categorical variables are presented as count (%). Continuous data are presented as mean 6 standard deviation.
0-33 months; median, 15.5). During follow-up, two other patients died of unrelated causes. During follow-up, two outflow vein stenoses were detected and treated with covered stent implantation and simple balloon angioplasty, respectively. Six occlusions occurred and were treated with simple thrombectomy, except for three cases that required adjunctive endovascular procedures on the outflow vein (two stenting and one balloon angioplasty). One patient presented with graft infection at 7 months and underwent graft explantation. A histologic analysis revealed IH proximal to the end of the NRS. No steal syndrome was observed during follow-up. Mean access flow was 1202 6 436 mL/min (range, 650-2200; median, 1150). After PSM, two groups of 25 patients each were created with correction of the differences in demographics and baseline characteristics (Table III). Twoyear survival estimates for hybrid and standard PTFE groups were 91% 6 6% and 82 6 9% (P > .05), respectively (Fig 3). One-year functional primary, functional assisted primary, and functional secondary patency estimates for the hybrid group vs the standard PTFE group were 66% 6 10% vs 51% 6 10% (P > .05), 75% 6 9% vs 51% 6 10% (P > .05), and 79% 6 9% vs 67% 6 10% (P > .05; Figs 4 to 6). The probability of functional primary, assisted primary, and secondary patency during 24 months for the GHVG group was 40%, 50%, and 38% more than for the PTFE group. However, no statistical
6
Journal of Vascular Surgery
Benedetto et al
---
Fig 3. Kaplan-Meier plot showing survival of Gore Hybrid Vascular Graft (GHVG) vs standard polytetrafluoroethylene (PTFE). Standard error was <10% at all time points.
Fig 4. Kaplan-Meier plot showing functional primary patency estimates of Gore Hybrid Vascular Graft (GHVG) vs standard polytetrafluoroethylene (PTFE). Curves exceeding 10% standard error are dashed.
significance was reached (Table IV). Secondary outcomes showed no statistically significant difference between the groups. The logistic regression model was statistically significant, c2 (3) ¼ 10.984 (P < .001). The model explained 43.0% (Nagelkerke R2) of the variance of GHVG-related reintervention and correctly classified 87.5% of cases. Of the three predictor variables (target vein diameter, previous vascular access surgery on the same limb, and diabetes), only one was statistically significant: target vein diameter. Reducing target vein diameter was associated with an increased likelihood of GHVG-related reintervention (Table V).
DISCUSSION AVF is currently the first choice in patients requiring access for hemodialysis treatment. An AVG is required when native vein anatomy is deemed disadvantaged or after repeated AVF failures. Reported 1-year primary
2016
Fig 5. Kaplan-Meier plot showing functional assisted primary patency estimates of Gore Hybrid Vascular Graft (GHVG) vs standard polytetrafluoroethylene (PTFE). Curves exceeding 10% standard error are dashed.
Fig 6. Kaplan-Meier plot showing functional secondary patency estimates of Gore Hybrid Vascular Graft (GHVG) vs standard polytetrafluoroethylene (PTFE). Curves exceeding 10% standard error are dashed.
Table IV. Effect of Gore Hybrid Vascular Graft (GHVG) implantation on functional patency rates during 24 months Hazard ratio (95% CI)
P value
Primary
0.595 (0.25-1.4)
.23
Primary assisted
0.498 (0.20-1.23)
.13
Secondary
0.625 (0.22-1.72)
.36
Functional patency
CI, Confidence interval.
and secondary patency rates for standard PTFE AVGs range from 40% to 60% and from 50% to 70%, respectively.21-24 In this study, we report on 32 patients treated with GHVG, with 1-year functional primary and secondary patency rates of 66% 6 10% and 79% 6 9%, respectively. Although the usefulness of the AVG is undeniable, its Achilles heel
Journal of Vascular Surgery Volume
-,
Number
Benedetto et al
7
-
Table V. Predictors of Gore Hybrid Vascular Graft (GHVG)-related reintervention
Target vein diameter
B
SE
Wald
df
1.997
.85
5.51
1
.02
P
0.14 (0.03-0.72)
Odds ratio (95% CI)
Previous vascular access
.705
.51
1.88
1
.17
2.02 (0.74-5.54)
Diabetes
.517
1.05
.24
1
.62
1.68 (0.22-13.07)
Constant
8.478
4.33
3.83
1
.050
4807.47
B, Regression coefficient; CI, confidence interval; df, degrees of freedom; SE, standard error of coefficient; Wald, Wald c2.
is the venous anastomosis. The main cause of failure of AVGs is considered to be stenosis at the venous anastomosis and subsequent thrombosis. This stenosis is mainly due to IH, which is triggered by turbulent flow, shear stress, and compliance mismatch between graft and vein.10,25 Turbulence is a well-established factor in the development of IH. Fillinger et al26 demonstrated a significant correlation between Reynolds number and intimalmedial thickening at the venous anastomosis. Sivanesan et al27 showed that the value of perianastomotic wall shear stress falls within the range that can cause endothelial injury (35-40 N/m). Fluid dynamics and geometry of the anastomosis have been addressed by several in vitro and in vivo models in an effort to understand and to prevent the development of IH.28-30 The influence of the anastomotic angle on the flow was studied using a porcine model.5 This consisted of an aortic graft interposition with end-to-side configuration. Distal anastomoses were performed with angles of 90, 45, or 15 degrees. The anastomoses of both 90 and 45 degrees showed a recirculation zone, whereas no turbulence of the flow was documented in the 15-degree ones. Hakaim et al31 showed that a sharp anastomotic angle, along with the incorporation of a flow diffuser, increased patency in brachial-axillary AVGs compared with a control group matched for age and gender. The hallmark of GHVG is the absence of a sewn venous anastomosis, replaced by a 5-cm or 10-cm NRS. This feature, besides offering a technically easier surgical approach to the target vein, allows a straight venous anastomotic junction, which might theoretically reduce the amount of turbulent flow. Nonetheless, the finding of IH is not totally excluded, as attested to in the single histologic analysis available. More anatomic samples are necessary for any conclusion to be drawn in this regard. The duration of intervention is comparable to that of standard AVG implantation. Time spared by sutureless anastomosis offsets the additional angiographic time. Clopidogrel might account for our three hematomas. However, it was reported to reduce AVF thrombosis without increasing bleeding events.32 Few reports on the use of GHVG are present in the literature to date. Anaya-Ayala et al33 reported on a series of 25 patients who received GHVG for hemodialysis access, with 1-year primary and secondary patency of 66% 6 8%
and 69% 6 7%, respectively. We observed a slightly higher secondary patency in our series. However, the selection criteria adopted by the two surgical groups are different. Indeed, Anaya-Ayala et al included only patients with a target vein <3 mm in diameter, whereas we excluded these patients and considered eligible only patients with a target vein at least 4 mm in diameter. Unlike Anaya-Ayala et al, we did not use the percutaneous approach, but we exposed the target vein in all cases. Moreover, our group implanted the graft in a wider variety of configurations compared with Anaya-Ayala et al, who reported exclusively brachial-axillary and axilloaxillary loop configurations. In our series, we observed five stenoses of the outflow vein, detected after AVG malfunction or occlusion. Most of the stenoses were located in the axillary or subclavian vein, next to a venous valve, and were mainly treated by positioning a Viabahn stent (W. L. Gore & Associates). The treatment of outflow vein stenoses with covered stents was previously described for standard AVGs, both as a planned procedure and during reintervention.33-35 Since the availability of a graft with 10-cm NRS, we started to perform intraoperative angiography routinely before the selection of the graft.11 Indeed, in our current practice, starting from the latest cases, the finding of a venous valve close to the expected landing zone of the NRS is a criterion to prefer the longer length. The 5-cm length, dominant in our series, was the only one used during the first period of our experience. As argued by Anaya-Ayala et al, the sutureless venous anastomosis may give rise to additional difficulties in case of reintervention.33 In our experience, of six cases of thrombectomy, one case required landing zone balloon angioplasty and two cases required angioplasty and stenting as adjunctive procedures, as a venous stenosis was present proximal to the end of the graft. Actually, most thrombectomies failed to result in a satisfactory lengthening of hybrid access survival as reocclusion occurred within a few weeks in four cases. The reason for this unsatisfactory outcome was unclear, and we believe a deterioration of proximal outflow may have played a role. On the other hand, maintenance procedures, carried out for access malfunction with a patent graft, were successful and were not followed by further reinterventions or occlusion. For this reason, in our opinion, a strict compliance with the follow-up controls is critical for this type of graft.
8
Benedetto et al
Journal of Vascular Surgery ---
Our multivariate analysis provided an insight into possible strategies to improve outcomes. In effect, it has shown that a reduction of target vein diameter was associated with an increased likelihood of GHVGrelated reintervention. We acknowledge that this finding is limited because of the small size of the sample and needs to be confirmed with further studies. Nonetheless, in our opinion, additional caution is prompted in the selection of the target vein whenever the diameter is in the lower end of the recommended range. In these cases, a possible solution could be to achieve a more proximal and thus larger in diameter landing zone by selecting the 10-cm NRS. This study presents some important limitations. The sample is heterogeneous in graft configurations, its size is limited, and the follow-up is relatively short. Larger cohorts of patients, a longer observation period, and possibly a multicenter randomized trial may help in defining the role for GHVG. A multicenter observational study, including a greater number of participating centers experienced with GHVG, is currently being conducted under the guidance of the first author. Whether the in-line graft-vein junction effectively produces laminar flow, in patients’ anatomies, has not yet been proved. In silico and in vitro models could be designed to better understand the fluid dynamics of real-anatomy geometries.
CONCLUSIONS The GHVG is a safe and effective alternative to standard PTFE AVG. Based on our results, a careful planning of the landing zone is advisable when the target vein is small because a decrease in vein diameter was associated with higher risk of reintervention. Larger studies and randomized trials are needed to better delineate the indications for GHVG.
AUTHOR CONTRIBUTIONS Conception and design: FB, GG, AD, FS Analysis and interpretation: DS, NP, AN, SV Data collection: DS, NP, GG, SV Writing the article: DS, NP, AN, SV Critical revision of the article: FB, DS, NP, GG, AN, AD, FS Final approval of the article: FB, DS, NP, GG, AN, SV, AD, FS Statistical analysis: DS, NP, AN Obtained funding: Not applicable Overall responsibility: FB
REFERENCES 1. Akoh JA. Prosthetic arteriovenous grafts for hemodialysis. J Vasc Access 2009;10:137-47. 2. Palder SB, Kirkman RL, Whittemore AD, Hakim RM, Lazarus JM, Tilney NL. Vascular access for hemodialysis. Patency rates and results of revision. Ann Surg 1985;202: 235-9. 3. Besrab A, Work J, Brouwer D. Clinical practice guidelines for vascular access. Am J Kidney Dis 2006;48(Suppl 1):S176-247.
2016
4. Huber TS, Carter JW, Carter RL, Seeger JM. Patency of autogenous and PTFE upper extremity arteriovenous hemodialysis accesses: a systematic review. J Vasc Surg 2003;39:491-6. 5. Staalsen NH, Ulrich M, Winther J, Pedersen EM, How T, Nygaard H. The anastomosis angle does change the flow fields at vascular end-to-side anastomoses in vivo. J Vasc Surg 1995;21:460-71. 6. Hofstra L, Bergmans DC, Leunissen KM, Hoeks AP, Kitslaar PJ, Daemen MJ, et al. Anastomotic intimal hyperplasia in prosthetic arteriovenous fistulas for hemodialysis is associated with initial high flow velocity and not with mismatch in elastic properties. J Am Soc Nephrol 1995;6:1625-33. 7. Hofstra L, Bergmans DC, Leunissen KM, Hoeks AP, Kitslaar PJ, Tordoir JH. Prosthetic arteriovenous fistulas and venous anastomotic stenosis: influence of a high flow velocity on the development of intimal hyperplasia. Blood Purificat 1996;14: 345-9. 8. Roy-Chaudhury P, Kelly BS, Narayana A, Desai P, Melhem M, Munda R, et al. Hemodialysis vascular access dysfunction from basic biology to clinical intervention. Adv Ren Replace Ther 2002;9:74-84. 9. Lemson MS, Tordoir JH, Daemen MJ, Kitslaar PJ. Intimal hyperplasia in vascular grafts. Eur J Vasc Endovasc Surg 2000;19:336-51. 10. Hofstra L, Tordoir JH, Kitslaar PJ, Hoeks AP, Daemen MJ. Enhanced cellular proliferation in intact stenotic lesions derived from human arteriovenous fistulas and peripheral bypass grafts. Does it correlate with flow parameters? Circulation 1996;94:1283-90. 11. Jones RG, Inston NG. Commentary: the hybrid vascular graft: a new frontier in arteriovenous access for dialysis or just another graft? J Endovasc Ther 2015;22:786-8. 12. Haskal ZJ, Trerotola S, Dolmatch B, Schuman E, Altman S, Mietling S, et al. Stent graft versus balloon angioplasty for failing dialysis-access grafts. N Engl J Med 2010;362: 494-503. 13. Lachat M, Mayer D, Criado FJ, Pfammatter T, Rancic Z, Genoni M, et al. New technique to facilitate renal revascularization with use of telescoping self-expanding stent grafts: VORTEC. Vascular 2008;16:69-72. 14. Kim JJ, Koopmann M, Ihenachor E, Zeng A, Ryan T. The addition of ultrasound arterial examination to upper extremity vein mapping before hemodialysis access. Ann Vasc Surg 2016;33:109-15. 15. Hemodialysis Adequacy 2006 Work Group. Clinical practice guidelines for hemodialysis adequacy, update 2006. Am J Kidney Dis 2006;48(Suppl 1):S2-90. 16. Chiesa R, Kahlberg A, Mascia D, Tshomba Y, Civilini E, Melissano G. Use of a novel hybrid vascular graft for sutureless revascularization of the renal arteries during open thoracoabdominal aortic aneurysm repair. J Vasc Surg 2014;60: 622-30. 17. Wuamett JC, Glickman MH. Arteriovenous graft configuration in hemodialysis: does it matter? J Vasc Access 2015;16: 78-81. 18. Jiménez-Almonacid P, Gruss E, Jimenez-Toscano M, Jiménez-Toscano M, Lasala M, Rueda JA, et al. Multidisciplinary approach to hemodialysis graft dysfunction and thrombosis. Nefrologia 2013;33:692-8. 19. Sidawy AN, Gray R, Besarab A, Henry M, Ascher E, Silva M, et al. Recommended standards for reports dealing with arteriovenous hemodialysis accesses. J Vasc Surg 2002;35: 603-10. 20. Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika 1983;70:41-55.
Journal of Vascular Surgery Volume
-,
Number
Benedetto et al
9
-
21. Katzman HE, Glickman MH, Schild AF, Fujitani RM, Lawson JH. Multicenter evaluation of the bovine mesenteric vein bioprostheses for hemodialysis access in patients with an earlier failed prosthetic graft. J Am Coll Surg 2005;201:223-30. 22. 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. 23. Gibson KD, Gillen DL, Caps MT, Kohler TR, Sherrard DJ, Stehman-Breen CO. Vascular access survival and incidence of revisions: a comparison of prosthetic grafts, simple autogenous fistulas, and venous transposition fistulas from the United States Renal Data System Dialysis Morbidity and Mortality Study. J Vasc Surg 2001;34:694-700. 24. Hodges TC, Fillinger MF, Zwolak RM, Walsh DB, Bech F, Cronenwett JL. Longitudinal comparison of dialysis access methods: risk factors for failure. J Vasc Surg 1997;26:1009-19. 25. Loth F, Jones SA, Zarins CK, Giddens DP, Nassar RF, Glagov S, et al. Relative contribution of wall shear stress and injury in experimental intimal thickening at PTFE end-to-side arterial anastomoses. J Biomech Eng 2002;124:44-51. 26. Fillinger MF, Reinitz ER, Schwartz RA, Resetarits DE, Paskanik AM, Bredenberg CE. Beneficial effects of banding on venous intimal-medial hyperplasia in arteriovenous loop grafts. Am J Surg 1989;158:87-94. 27. Sivanesan S, How TV, Black RA, Bakran A. Flow patterns in the radiocephalic arteriovenous fistula: an in vitro study. J Biomech 1999;32:915-25. 28. Heise M, Schmidt S, Krüger U, Pfitzmann R, Scholz H, Neuhaus P, et al. Local haemodynamics and shear stress in
29.
30.
31.
32.
33.
34.
35.
cuffed and straight PTFE-venous anastomoses: an in-vitro comparison using particle image velocimetry. Eur J Vasc Endovasc Surg 2003;26:367-73. Krueger U, Huhle A, Krys K, Scholz H. Effect of tapered grafts on hemodynamics and flow rate in dialysis access grafts. Artif Organs 2004;28:623-8. Hughes PE, How TV. Effects of geometry and flow division on flow structures in models of the distal end-to-side anastomosis. J Biomech 1996;29:855-72. Hakaim AG, Nalbandian MN, Heller JK, Chowla AC, Oldenburg WA. Improved patency of prosthetic arteriovenous grafts with an acute anastomotic angle and flow diffuser. J Vasc Surg 2003;37:1032-5. Dember LM, Beck GJ, Allon M, Delmez JA, Dixon BS, Greenberg A, et al. Effect of clopidogrel on early failure of arteriovenous fistulas for hemodialysis: a randomized controlled trial. JAMA 2008;299:2164-71. Anaya-Ayala JE, Davies MG, El-Sayed HF, Peden EK, Naoum JJ. Early experience with a novel hybrid vascular graft for hemodialysis access creation in patients with disadvantaged anatomy. J Endovasc Ther 2015;22:778-85. Vogel PM, Parise C. Comparison of SMART stent placement for arteriovenous graft salvage versus successful graft PTA. J Vasc Interv Radiol 2005;16:1619-26. Carmona J, Rits Y, Jones B, Dowers L, Bednarski D, Rubin JR. Patency of the Viabahn stent graft for the treatment of outflow stenosis in hemodialysis grafts. Am J Surg 2016;211:551-4.
Submitted Mar 19, 2016; accepted Jul 26, 2016.