|Year : 2019 | Volume
| Issue : 2 | Page : 82-88
Ultrasound-assisted angioplasty for failing arteriovenous access
Shahzad Sarosh Bulsara, Vinit Paliwal, Govind Prasad, NR Manjubharath, Ashutosh Aher, Ashish Airen, Tapish Sahu, Virender Sheorain, Tarun Grover, Rajiv Parakh
Division of Peripheral Vascular and Endovascular Sciences, Medanta – The Medicity Hospital, Gurgaon, Haryana, India
|Date of Web Publication||6-Jun-2019|
Dr. Shahzad Sarosh Bulsara
Division of Peripheral Vascular and Endovascular Sciences, Medanta – The Medicity Hospital, Gurgaon, Haryana
Source of Support: None, Conflict of Interest: None
Introduction: Arteriovenous (AV) access maturity and patency is critical for patients on hemodialysis. Maintaining a functioning AV access for these patients is of paramount importance. Salvage of failing AV accesses has been traditionally done mainly with angiographic techniques requiring significant amounts of contrast agent and radiation exposure. Ultrasound-guided angioplasty has been described, but its utility needs to be further studied. Methods: Group 1: 42 patients undergoing ultrasound-assisted angioplasty (UAA) for AV access salvage were assessed prospectively for 9 months (October 2016 to June 2017). Group 2: Retrospective data of 47 patients who underwent angioplasty for failing AV access without ultrasound assistance, over 9 months was collected (January to September 2016). This group was matched with Group 1 in terms of age, sex, type of AV access, and cause of AV access dysfunction. Results: There was a statistically significant decrease in the average number of angiographic runs (2.5 vs. 7.9) and fluoroscopy time in minutes (3.8 vs. 13.1) recorded during UAA (P < 0.01). The average quantity of contrast agent used in milliliters (17.6 vs. 49.8) was also significantly less when ultrasound assistance was used (P < 0.01). Conclusion: UAA for failing AV access reduces the radiation exposure and the amount of contrast agent used in these patients who require multiple of these sessions and need dialysis to get the contrast agent out of their circulation. This technique is feasible and reliable; hence should be used more frequently and effectively. Our model of UAA for AV access salvage allows all cases including complex cases to be performed with the added advantage of reduced radiation exposure and contrast agent; serving as a boon to these already distressed patients. Immediate complication management and adjunctive open procedures also become possible because of the hybrid operating room setup with fluoroscopy facility.
Keywords: Arteriovenous access, fistuloplasty, hemodialysis, hybrid angiographic suite, ultrasound assistance
|How to cite this article:|
Bulsara SS, Paliwal V, Prasad G, Manjubharath N R, Aher A, Airen A, Sahu T, Sheorain V, Grover T, Parakh R. Ultrasound-assisted angioplasty for failing arteriovenous access. Indian J Vasc Endovasc Surg 2019;6:82-8
|How to cite this URL:|
Bulsara SS, Paliwal V, Prasad G, Manjubharath N R, Aher A, Airen A, Sahu T, Sheorain V, Grover T, Parakh R. Ultrasound-assisted angioplasty for failing arteriovenous access. Indian J Vasc Endovasc Surg [serial online] 2019 [cited 2022 May 28];6:82-8. Available from: https://www.indjvascsurg.org/text.asp?2019/6/2/82/259657
| Introduction|| |
The end-stage renal disease population is ever increasing, and hence, the demand for renal replacement therapy (RRT) keeps growing. Arteriovenous (AV) access maturity and patency is critical for RRT patients on hemodialysis. Maintaining a functioning AV access for these patients is of paramount importance. Lee et al. in their publication on “Standardized Definitions for Hemodialysis Vascular Access” have defined “Fistula Used Successfully for Hemodialysis–A fistula is deemed successfully used for dialysis if it can be used with two-needle cannulation for two-thirds or more of all dialysis runs for 1 month and deliver the prescribed dialysis within the prescribed time frame.”
The first step is to identify the failing AV access so that it can be salvaged. Hu et al. in their review article have described three types of AV access failures, namely, early thrombosis (<3 weeks), failure to mature (<6 months), and late failure (>6 months). Clinical definitions of a failing AV access are one or more of the following “absolute blood flow <500 mL/min, reduction of flow rate by >20% from baseline, >5% recirculation, dynamic venous pressures exceeding threshold levels three times in succession, or clinical signs and symptoms such as pulsatile AV access, prolonged bleeding from puncture site, arm pain, arm swelling, thrombosis, or stenosis anywhere along the AV access.”
AV access salvage procedures may include angioplasty with or without stenting of stenosis or obstruction, which may involve the inflow artery, anastomosis, outflow vein, and/or the central veins. Percutaneous thrombolysis and/or thrombectomies with or without angioplasty of the residual stenosis can also be performed for salvaging AV accesses. These procedures have been traditionally done mainly with angiographic techniques requiring significant amounts of contrast agent and radiation exposure. Ultrasound-guided (UG) angioplasty has been described, but its utility needs to be further studied.
Need and aims of the current study
Our institution recently published the feasibility of UG fistuloplasty in the office setting. However, we had excluded patients with suspected central vein stenosis (CVS) and fistulas that were thrombosed. We did have complications such as hematomas and the need for adjunctive fluoroscopy and stenting was required in two patients in our series of 43 patients performed in the office setting. To expand the advantages of ultrasound guidance in all cases of AV access salvage, we planned to assess the utility of ultrasound in the hybrid operating suite.
The impact of proximal vein stenosis on maturation failure and failing AV access is significant. The incidence of cephalic arch stenosis (CAS) cannot be underestimated. Sivananthan et al. in their review on CAS have described its incidence as high as 77% in failing upper arm/brachiocephalic (BC) AV fistulas (AVF's). In their review, one series has reported 15% incidence of CAS for combined BC and radiocephalic (RC) AVF, but it was 39% when only failing BC AVF were assessed. Other series also reported 55% incidence of CAS in the upper arm fistulas.
CVS causing AV access dysfunction needs further studies, nevertheless, its association with AV access failure should not be neglected. Unsuspected incidence of CVS in functioning grafts has been reported up to 29%. Gallagher et al., in their study on duplex-guided balloon-assisted maturation (DG-BAM), had 16% maturation failure because of cephalic arch and subclavian vein stenosis. Cho et al. reported 20.3% cases of central and arterial stenosis in their series of failing AV accesses.
Ehrie et al. showed that 39 patients returned within 3 months of angioplasty of their hemodialysis access with symptoms of CVS which they described as “unmasking” of CVS. They noted that 90% of unmasking was seen in the upper arm AV access. Cho et al. also encountered underlying CAS after correction of inflow stenosis. Computed tomography or magnetic resonance imaging to rule out CVS may not be always feasible and is also associated with significant amount of contrast exposure.
To overcome all these limitations and to be able to manage all types of AV accesses that are not fit for hemodialysis, we intended to perform this study. Our aim was to study the concept of a hybrid procedure, that is, the use of ultrasound assistance to compliment the traditional fluoroscopy based angioplasty for failing AV accesses with the added ability to manage complications simultaneously. The drive for office-based procedures is mainly governed by the health-care system. In the Indian health-care system, the ambulatory surgery or same-day procedure works equally effectively, hence the need to perform such a study. We also aimed to prove the efficacy of the ultrasound technique in accurately diagnosing the number of peripheral stenoses, and the accuracy of the technique to appropriately select the size of the balloon to be used for angioplasty.
| Methods|| |
- Group 1: 42 patients who underwent ultrasound-assisted angioplasty (UAA) for AV access salvage were assessed prospectively for a period of 9 months (October 2016 to June 2017)
- Group 2: Retrospective data of 47 patients who underwent angioplasty for failing AV access without ultrasound assistance, over a period of 9 months was collected (January to September 2016). This group was matched with Group 1 in terms of age, sex, type of AV access, and cause of AV access dysfunction.
All patients with the following problems with their dialysis AV access (AVF's and grafts) were included as follows:
- Non-maturing AV access
- Blood flow <400 ml/min
- High venous pressures (>200 mmHg) during three or more dialysis sessions 
- Dampened thrill or pulsatile flow on clinical examination
- Acute thrombosis of AV access
- Arm and/or facial swelling
- Stenosis (>50%) of the AV access on ultrasound examination
- Prolonged bleeding from the punctured site.
- Patients allergic to contrast agent
- Hemodynamically unstable and critically ill patients
- Technically unsuccessful procedures.
Hybrid angiographic suite setup
- The patient is placed in a supine position with the non-AV access arm tucked in on the side of the patient [Figure 1]
- The AV access arm is placed outstretched on a radiographically compatible arm board
- Ultrasound machine and fluoroscopic screen are placed on the side opposite to the AV access arm.
- Portable ultrasound machine
- Floor-mounted fluoroscopy machine
- Access sheath-5F up to 11F
- Iodinated contrast agent
- Multipurpose and support catheters
- Guide wires – 0.035” and 0.014” (J-tip, straight, regular and stiff)
- Angioplasty balloons from 4 mm to 16 mm (high pressure, cutting, and drug coated balloons).
- Informed consent was obtained from all patients
- All procedures were performed by the same team
- Sterile cleaning and draping done to include the entire hand, arm, shoulder, axilla, and chest in the sterile field
- Ultrasound probe was covered with a sterile probe wrap
- Local anesthesia was given with 2% xylocaine at the access site
- UG access was obtained either retrograde, antegrade, or dual either venous or arterial depending on the case
- Femoral vein access was required sometimes during central vein recanalization
- Sheath placed and heparin 30 units/kg given
- In the retrospective group, the procedure was performed as a routine angiography and angioplasty procedure with or without stenting
- In the prospective group of UAA, an screening angiography was performed to identify the location and number of peripheral and/or central stenosis
- Peripheral angioplasty was performed with ultrasound guidance as described by Banerjee et al.
- B-mode ultrasound clearly delineates the balloon, and the waist is seen which can be continuously monitored until the stenosis is broken, that is, the waist has disappeared on the ultrasound screen [Figure 2].
- CAS and CVS were treated with minimal use of fluoroscopy with high-pressure balloons with or without stenting
- Completion angiogram was performed
- Technical success was defined as residual stenosis of <30% and a hemodynamic improvement in the flow rate of >400 ml/min.
|Figure 2: Ultrasound-guided fistuloplasty – resolution of the waist – allowing continuous monitoring|
Click here to view
Data collection and statistical analysis
- The number of angiographic runs, fluoroscopy time, and quantity of contrast used was recorded for both the groups and compared
- t-test used to compare these parameters.
- The number of peripheral stenoses and vessel diameter postangioplasty measured by ultrasound and angiographic image were recorded and compared in the UAA group
- Pearson's correlation coefficient was used to compare these parameters.
- P < 0.05 was considered statistically significant
- Quantitative parameters are presented in terms of means and standard deviation
- Qualitative/categorical data are presented as absolute numbers and proportions
- Statistical testing was conducted with the Statistical Package for the Social Sciences system (SPSS) version 24 (IBM SPSS for Windows, Version 24.0; Armonk, NY: IBM Corp.).
| Results|| |
There was no significant difference between the two groups in terms of age and sex distribution, type of AV access (P = 0.61), and cause of AV access dysfunction (P = 0.83) [Table 1]. The mean age of UAA group was 58.6 ± 6.9 years and that of Group 2 was 57.4 ± 7.1 years (P = 0.33). There were 57.14% males in the UAA group, while Group 2 had 55.32% males. UAA group had 47.54% RC AVF, 42.78% BC AVF, 7.30% brachio-basilic (BB) fistula with transposition, and 2.38% AV grafts [Figure 3]. Group 2 had 42.55% RC AVF, 46.81% BC AVF, 6.38% BB AVF with transposition, and 4.26% AV grafts [Figure 4]. The percentage distribution of the causes of AV access dysfunction in the UAA group was 4.76% acute thrombosis, 42.86% outflow vein stenosis, 19.05% juxta-anastomotic (JA) stenosis, 2.38% arterial stenosis, 14.29% CAS, and 16.67% CVS [Figure 5]. Similarly, the percentage distribution of the cause of AV access dysfunction in the traditional fistuloplasty group was 4.26% acute thrombosis, 42.55% outflow vein stenosis, 21.28% JA stenosis, 4.26% arterial stenosis, 14.89% CAS, and 12.77% CVS [Figure 6].
|Figure 3: Type of arteriovenous access distribution in ultrasound-assisted angioplasty group (Group 1)|
Click here to view
|Figure 4: Type of arteriovenous access distribution in traditional fistuloplasty group (Group 2)|
Click here to view
|Figure 5: Cause of arteriovenous access dysfunction in ultrasound-assisted angioplasty group (Group 1)|
Click here to view
|Figure 6: Cause of arteriovenous access dysfunction in the traditional fistuloplasty group (Group 2)|
Click here to view
There was a statistically significant decrease in the number of angiographic runs and fluoroscopy time recorded during UAA (P < 0.01). The quantity of contrast agent recorded was also significantly less when ultrasound assistance was used (P < 0.01) [Table 2] and [Figure 7]. The average number of angiographic runs in the UAA group was 2.5, while it was 7.9 in the traditional fistuloplasty group (Group 2). The mean of total fluoroscopy time and contrast agent used in UAA group was 3.8 min and 17.6 ml compared to 13.1 min and 49.8 ml in the Group 2, respectively.
|Table 2: Mean of angiography runs, fluoroscopy time, and amount of contrast agent – comparison between the two groups|
Click here to view
|Figure 7: Graphical comparison showing the difference in the mean of angiography runs, fluoroscopy time, and amount of contrast agent between the two groups|
Click here to view
When data of UAA group were analyzed, there was a strong correlation between the number of stenosis recorded by ultrasound and by the pre-angioplasty angiography (R = 0.8) [Figure 8]. In this group, there was strong correlation also noted between the size of recovered stenotic segment recorded on ultrasound and post-angioplasty angiogram (R = 0.9) [Figure 9].
|Figure 8: Correlation to diagnose number of stenosis by ultrasound and by the pre-angioplasty angiography (UAA group)|
Click here to view
|Figure 9: Correlation to accurately measure the effect of angioplasty by ultrasound and by post-angioplasty angiogram (UAA group)|
Click here to view
| Discussion and Review of Literature|| |
UG angioplasty for failing or non-maturing AV access has been described for more than a decade. In a series of 10 patients, Marks et al. in 2007 showed the feasibility of ultrasonography (USG)-guided angioplasty for AV access stenosis and also for stenting.
USG-guided AVF angioplasty technique has been further studied since then and has been shown to be feasible, safe, and effective even in the office setting.,, However, all these studies have had hurdles or complications which could possibly be avoided with fluoroscopic backup. In the study of Ascher et al., they had one of the 25 patients requiring two procedures at different time intervals because of the presence of concurrent stenosis at shoulder level. Gorin et al., in their study of UG angioplasty of autogenous AVF's in the office setting, had performed 87% angioplasties for non-maturing fistulas. They were unable to cross the proximal stenosis under ultrasound guidance in two of the thirty patients. Similarly, Gallagher et al. concluded that DG-BAM of AVF's for hemodialysis was safe and effective in the office, but even in their series 7 of 44 patients required subsequent angioplasty of CAS and subclavian vein stenosis to achieve AVF maturation.
Bojakowski et al. have described in detail the technique of UG angioplasty and concluded that it is a safe and effective method. Until these studies, the spectrum of cases which underwent UG AV access angioplasty were mainly non-maturing fistulas.
Wakabayashi et al. though published their large series of 4896 cases in 2013, they have been using ultrasound guidance since 2004. They describe that ultrasound guidance was used as the first approach with 55 cases requiring fluoroscopy in addition. They even described the visualization of subclavian and brachiocephalic veins by the use of microconvex probe. Treatment of 443 cases of obstruction was performed by UG aspiration of thrombus and a success rate of 91.9% was achieved. Thus, it is clear from their study that ultrasound guidance is very useful, but backup or concomitant fluoroscopy is equally important.
Further studies by Kumar et al., Cho et al., and Garcia-Medina and García-Alfonso have expanded and further explored the utility of this technique. Kumar et al. included only patients with peripheral stenosis of AVF and they did suggest that a randomized controlled trial comparing fluoroscopy and ultrasound guidance in endovascular management of AVF stenosis is required.
Cho et al., in their 5-year study of UG-percutaneous angioplasty for AVF stenosis, only included JA stenosis or outflow vein stenosis. They excluded central lesions, CAS, arterial stenosis, thrombosis, perforating vein stenosis, and aneurysms. They also excluded the cases, in which concurrent fluoroscopy was required.
Garćia-Medina and García-Alfonso, in their study of UG angioplasty of dysfunctional vascular access for hemodialysis, studied the results of 189 UG balloon angioplasties. They were able to complete two-thirds (67%) of the procedures successfully without fluoroscopy. They had also excluded patients with CVS. In addition to the basic advantage of ultrasound of no radiation no contrast they claimed that ultrasound allowed better visualization of thrombus than fluoroscopy. The disadvantages of pure ultrasound guidance described by them were poor coordination, loss of panoramic view, loss of hemodynamic vision of the complete fistula, difficult assessment of fistula collateral circulation, risk of losing the guide wire, global result assessment difficulty, difficult traversing and progressing wire through acute anastomotic angles or aneurysms, and risk of entering a wrong way and dilating a collateral with risk of rupture. They concluded that combining ultrasound and fluoroscopy guidance could deliver a more efficient treatment.
One of the limitations mentioned by Garćia-Medina and García-Alfonso in their study was that the AV access was not mapped with fluoroscopy which is considered the gold standard. In our study, we have attempted to overcome this limitation by comparing the AV access extent of stenosis shown on fluoroscopy and that seen by ultrasound.
Leskovar et al. also studied the utility and showed the feasibility and safety of ultrasound guidance for AVF and AVG angioplasty. They performed their procedures in the operating room and performed open procedures and thrombectomies when needed. Stent placement when required was also performed under ultrasound guidance. They also used fluoroscopy when ultrasound visibility was poor or in cases of CVS. Their model of ultrasound use in AV access angioplasty is similar to our model, but they have not performed any comparison between fluoroscopy and details of what percentage of patients required fluoroscopy is not available.
Aurshina et al. recently also showed the feasibility of USG-guided salvage of acute AVF occlusions in their series of 14 patients. In our study also, we had 4.76% cases of acute thrombosis in the UAA group. These cases were managed with ultrasound-assisted thrombolysis, thrombus aspiration, and angioplasty of the residual stenosis, with minimal use of fluoroscopic guidance.
Huang et al. presented their study of 11 patients where contrast-enhanced ultrasound (CEUS) was used for treating failing or nonmaturing AVF's. They concluded that this technique may provide improved spatial resolution for narrow stenotic segments, would improve the safety of US-guided AVF interventions as it is sensitive to pick up potential complicating extravasations. CEUS has the added novel application of providing “sonographic angiogram” which was effectively used in their study pre- and post-angioplasty.
The ideal procedure for an AVF salvage and maturation assistance would be zero contrast zero radiation for all kinds of failing AV accesses and for all possible complications. This does not seem to be far from reality with the increased use of intravascular ultrasound (IVUS). IVUS was used in conjunction with fluoroscopy without contrast agent in a patient with CVS with contrast allergy. They successfully placed a stent in the central vein with the aid of IVUS, fluoroscopy, and pressure gradient measurements. Similarly, Casey et al. published a case report where they successfully used IVUS for angioplasty of an AV loop graft in a patient with contrast allergy. Larger studies with the use of IVUS are needed for exploring the utility of this modality in AV access salvage procedures.
Pure UG AV access angioplasty is highly beneficial for patients on hemodialysis with AV access dysfunction and becomes much more important for patients who are yet to start dialysis (non-maturing AV access), as they are at the highest risk for contrast agent-related complications. Hence, we recommend studies and research focusing more on pure UG AV access angioplasty for all kinds of AV access-related problems.
| Conclusion|| |
UAA for failing AV access reduces the radiation exposure and the amount of contrast agent used in these patients who require multiple of these sessions and need dialysis to get the contrast agent out of their circulation. This technique is feasible and reliable hence should be used more frequently and effectively. For properly selected AVF, stenosis pure USG-guided angioplasty in the office setting is feasible and cost-effective. When facilities are available currently, the ideal procedure would be an USG-assisted angioplasty (UAA) of any failing or nonmaturing AV access in an hybrid angiographic suite. Our model of UAA for AV access salvage allows all cases including complex cases to be performed with the added advantage of reduced radiation exposure and contrast agent; serving as a boon to these already distressed patients. Complications if any such as dissections or ruptures can also be immediately and more appropriately managed in the fluoroscopic setup. In the advent of a need of an adjunctive open procedure such as tributary ligation or management of a complication, it can be easily performed because of the hybrid operating room setup.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Lee T, Mokrzycki M, Moist L, Maya I, Vazquez M, Lok CE, et al.
Standardized definitions for hemodialysis vascular access. Semin Dial 2011;24:515-24.
Hu H, Patel S, Hanisch JJ, Santana JM, Hashimoto T, Bai H, et al.
Future research directions to improve fistula maturation and reduce access failure. Semin Vasc Surg 2016;29:153-71.
Sivananthan G, Menashe L, Halin NJ. Cephalic arch stenosis in dialysis patients: Review of clinical relevance, anatomy, current theories on etiology and management. J Vasc Access 2014;15:157-62.
Banerjee S, Patel HM, Sahu T, Sheorain VK, Grover T, Parakh R. Ultrasound-guided Fistuloplasty: A Novel Office-based Technique for Arteriovenous Fistula Salvage. Indian J Vasc Endovasc Surg 2017;4:58-62. [Full text]
Lumsden AB, MacDonald MJ, Isiklar H, Martin LG, Kikeri D, Harker LA, et al.
Central venous stenosis in the hemodialysis patient: Incidence and efficacy of endovascular treatment. Cardiovasc Surg 1997;5:504-9.
Gallagher JJ, Boniscavage P, Ascher E, Hingorani A, Marks N, Shiferson A, et al.
Clinical experience with office-based duplex-guided balloon-assisted maturation of arteriovenous fistulas for hemodialysis. Ann Vasc Surg 2012;26:982-4.
Cho S, Lee YJ, Kim SR. Clinical experience with ultrasound guided angioplasty for vascular access. Kidney Res Clin Pract 2017;36:79-85.
Ehrie JM, Sammarco TE, Chittams JL, Trerotola SO. Unmasking of previously asymptomatic central venous stenosis following percutaneous transluminal angioplasty of hemodialysis access. J Vasc Interv Radiol 2017;28:1409-14.
Marks N, Ascher E, Hingorani AP. Duplex-guided repair of failing or nonmaturing arterio-venous access for hemodialysis. Perspect Vasc Surg Endovasc Ther 2007;19:50-5.
Ascher E, Hingorani A, Marks N. Duplex-guided balloon angioplasty of failing or nonmaturing arterio-venous fistulae for hemodialysis: A new office-based procedure. J Vasc Surg 2009;50:594-9.
Gorin DR, Perrino L, Potter DM, Ali TZ. Ultrasound-guided angioplasty of autogenous arteriovenous fistulas in the office setting. J Vasc Surg 2012;55:1701-5.
Bojakowski K, Góra R, Szewczyk D, Andziak P. Ultrasound-guided angioplasty of dialysis fistula – Technique description. Pol J Radiol 2013;78:56-61.
Wakabayashi M, Hanada S, Nakano H, Wakabayashi T. Ultrasound-guided endovascular treatment for vascular access malfunction: Results in 4896 cases. J Vasc Access 2013;14:225-30.
Kumar S, Mahajan N, Patil SS, Singh N, Dasgupta S, Tejavath S, et al.
Ultrasound-guided angioplasty for treatment of peripheral stenosis of arteriovenous fistula – A single-center experience. J Vasc Access 2017;18:52-6.
García-Medina J, García-Alfonso JJ. Ultrasound-guided angioplasty of dysfunctional vascular access for haemodialysis. The pros and cons. Cardiovasc Intervent Radiol 2017;40:750-4.
Leskovar B, Furlan T, Poznič S, Potisek M, Adamlje A, Ključevšek T, et al.
Ultrasound-guided percutaneous endovascular treatment of arteriovenous fistula/graft. Clin Nephrol 2017;88:61-4.
Aurshina A, Ascher E, Hingorani A, Marks N. A novel technique for duplex-guided office-based interventions for patients with acute arteriovenous fistula occlusion. J Vasc Surg 2018;67:857-9.
Matthews R, Thomas J. Intravascular ultrasound-guided central vein angioplasty and stenting without the use of radiographic contrast agents. J Clin Ultrasound 2008;36:254-6.
Casey PE, Miranda CJ, Al-Khaffaf H, Woodhead PM. Intravascular ultrasound-guided angioplasty of hemodialysis loop graft in a patient with contrast allergy. J Vasc Access 2014;15:424-6.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9]
[Table 1], [Table 2]