Degradation phenomena on last generations of polyethylene terephthalate knitted vascular prostheses

Abstract

Objectives

The aim of this study was to analyze a series of new generations of explanted knitted polyethylene terephthalate (PET) vascular grafts (VGs) presenting nonanastomotic degradations according to preoperative computed tomography angiography (CTA) when available in order to better understand the mechanisms leading to rupture.

Methods

Explanted knitted PET VGs were collected as part of the Geprovas European Collaborative Retrieval Program. VGs implanted after 1990 presenting a nonanastomotic rupture of the fabric were included. Clinical data and pre-explantation CTA data when available were retrieved for each VG. The ruptures were characterized by macroscopic examination and optical microscopy according to a standardized protocol.

Results

Nineteen explants were collected across 11 European centers, 13 were implanted as infrainguinal bypasses, 3 at the aortic level, and 1 as an axillobifemoral bypass. The mean implantation duration was 9.2 years. Pre-explantation CTA data were available for 8 VGs and showed false aneurysms at the adductor canal level on 4 VGs, at the inguinal ligament level on 2 VGs, and in the proximal or middle third thigh level on 3 VGs. Examination revealed longitudinal ruptures on 9 explanted VGs (EVGs), transversal ruptures on 15 EVGs, 45°-oriented ruptures on 5 EVGs, V-shaped ruptures on 7 EVGs, and punctiform ruptures on 2 EVGs. Ruptures involved the remeshing line on 11 EVGs, the guideline on 10 EVGs, and the crimping valley on 15 EVGs.

At the microscopic level, two main degradation phenomena could be identified: a decrease in the density of the meshing and local ruptures of the PET fibers. Fourteen EVGs presented a loosening of the remeshing line and 17 EVGs an attenuation of the crimping.

Conclusions

New-generation PET VG degradation seems to result from both anatomic constraints and intrinsic textile structure phenomena.

Clinical Relevance

In this study, the analysis of explanted new-generation polyethylene terephthalate grafts presenting nonanastomotic lesions allowed the identification of two main degradation phenomena: a decrease in the density of the meshing and local ruptures of the polyethylene terephthalate fibers. This highlights that both anatomic constraints and intrinsic textile structure phenomena are involved in degradation phenomena. This study highlights the need to keep studying the degradation mechanisms in order to develop compliant and durable vascular substitutes.

Article Highlights.

• •Type of Research: Multicenter retrospective cohort study
• •Key Findings: The analysis of 19 explanted new-generation polyethylene terephthalate grafts presenting nonanastomotic degradations allowed the identification of two main degradation phenomena: (1) a decrease in the density of the meshing and (2) local ruptures of the polyethylene terephthalate fibers.
• •Take Home Message: Degradation seems to result from both anatomic constraints and intrinsic textile structure phenomena, reinforcing the need to keep studying degradation mechanisms in order to develop a substitute for vascular repair that would be performant in terms of both compliance and durability.

One of the major steps of vascular surgery was the introduction of synthetic vascular grafts (VGs) 70 years ago. Expanded polytetrafluoroethylene and polyethylene terephthalate (PET) are the main materials used to build the current generations of VGs.¹ If synthetic VGs provide acceptable results for the reconstructions involving middle-sized and large-diameter vessels, there is still no strong recommendation concerning which synthetic material to favor in the absence of a suitable autologous material for small-diameter vessels.^2,3

Dilatation and ruptures of VGs were described as major consequences of long-term VG degradation.4, 5, 6 One could fall in the trap of summarizing that PET VG degradations only addressed old VG generations. Ruptures on warp-knitted PET VGs have however still been observed, highlighting the need to carry on with explanted VG investigations.^7,8

New generations of PET VGs have been proposed to better mimic arterial mechanical properties to enhance their compliance by introducing the concept of thin-wall VGs. We previously demonstrated that they did not fulfill expectations in terms of compliance,⁹ but the long-term durability of these thinner structures remains unclear. Previous reports concluded that new generations of PET VGs are still subjected to long-term degradation mechanisms.¹⁰ Accordingly, the role of material surveillance still remains essential to better understand the overall mechanisms of degradation and improve next generations of devices.

We already know that VG degradation is multifactorial associating the design of the VG as well as postimplantation factors such as chemical and mechanical factors. Retrospective analysis is difficult, but data such as imaging results provided during the life duration of the VG could help for analysis. This is the reason why we tried to get computed tomography angiography (CTA) data because it could provide objective and analyzable data allowing a better understanding of these phenomena of VG degradation.

The aim of this study was to analyze a series of new generations of explanted knitted PET VGs presenting nonanastomotic degradations according to preoperative CTA when available, in order to better understand the mechanisms leading to rupture.

Methods

The explants

The explanted PET VGs were collected as part of the Geprovas European Collaborative Retrieval Program. The analysis focused on explanted knitted VGs presenting a rupture of the fabric. VGs implanted before 1990 were excluded as they were not considered as part of the new VG generations.

Clinical data were retrieved for each explant, indexing the surgeon who explanted the VG, the model of the VG, the reason for implantation, the site and duration of implantation, and the reason for explantation. When available, pre-explantation CTA was systematically analyzed.

Explant processing and analysis

The VGs were fixed in a formalin solution right after explantation in the operating room. All specimens were studied according to our ISO 9001-certified standard protocol¹⁰ including the following steps:

• (1)Indexation of the explant on reception and storage of all clinical features provided by the surgeon.
• (2)Macroscopic analysis before cleaning using a Nikon D5100 camera (Nikon France).
• (3)If needed a sample was collected for histological analysis.
• (4)Cleaning of the VG according to our standardized protocol that was chemical using sodium hypochlorite until 2017¹¹ and became papain-based enzymatic after.¹²
• (5)After cleaning, VGs were once again photographed and analyzed macroscopically to determine the area where the ruptures occurred.
• (6)The analysis was completed by carrying out a microscopical examination using a Keyence VHX-600 digital microscope (Keyence France).

Classification of the observed ruptures

Each rupture was classified according to its shape and localization along the tubular structure (remeshing line, guideline, and crimping). Five different rupture shapes were considered in the frame of this work: longitudinal, transversal, 45° oriented, V-shaped, and punctiform. When a rupture presented combined mechanisms, each observed rupture shape was classified individually. Independently of the ruptures, signs of aging degradation of each textile knitted structure were identified (local mesh opening, remeshing line loosening, and loss of crimping). Two investigators performed the analysis on all the explants.

CTA analyses

When pre-explantation CTA was available, a 3D reconstruction was performed to understand the VG in vivo conformation and indirectly the eventual mechanical stresses that could have been applied on the VGs. For each CTA, the implantation site, the location of the false aneurysm, the presence of small dilatation along the graft, and remarkable events were evaluated. Therefore, it was possible to study the observed ruptures on the VG in correlation with the in vivo conditions.

Results

Explant characteristics

From January 2011 to July 2020, 336 PET VGs were collected as part of the European Collaborative Retrieval Program. Twenty-one of them were knitted VGs presenting a rupture of the fabric. Two of them were implanted before 1990 and therefore excluded. In total, 19 explanted VGs (EVGs) collected from 11 European centers were analyzed.

The clinical characteristics are presented in Table I.

Table I.

Clinical characteristics and structural properties of explanted vascular grafts (EVGs)

Duration of implantation years (months)	Reason for implantation	Reason for explantation	Site of implantation	Available CTA	Model of prosthesis	Thin wall	Velour structure	Coating	Crosslinking
13.1 (157)	Intermittent claudication	Nonanastomotic rupture	Above-the-knee femoropopliteal	No	Model 1	No	External	Collagen	Glutaraldehyde
2.3 (28)	Intermittent claudication	Nonanastomotic false aneurysm	Above-the-knee femoropopliteal	Yes	Model 2	Yes	None	Collagen	NA
6.8 (82)	Aneurysm	Nonanastomotic false aneurysm	Above-the-knee femoropopliteal	Yes	Model 2	Yes	None	Collagen	NA
4.8 (57)	Chronic limb ischemia	Nonanastomotic false aneurysm	Above-the-knee femoropopliteal	Yes	Model 2	Yes	None	Collagen	NA
7.0 (84)	Intermittent claudication	Nonanastomotic rupture	Above-the-knee femoropopliteal	Yes	Model 2	Yes	None	Collagen	NA
8.2 (98)	Chronic limb ischemia	Nonanastomotic rupture	Above-the-knee femoropopliteal	Yes	Model 2	Yes	None	Collagen	NA
7.9 (95)	NA	Nonanastomotic rupture	Femorofemoral	Yes	Model 2	Yes	None	Collagen	NA
7.9 (95)	Intermittent claudication	Nonanastomotic rupture	Above-the-knee femoropopliteal	Yes	Model 2	Yes	None	Collagen	NA
3.2 (38)	NA	Nonanastomotic rupture	Above-the-knee femoropopliteal	No	Model 1	No	External	Collagen	Glutaraldehyde
5.5 (66)	Intermittent claudication	Thrombosis	Above-the-knee femoropopliteal	No	Model 1	No	External	Collagen	Glutaraldehyde
3.5 (42)	Chronic limb ischemia	Anastomotic false aneurysm	Crossed femorofemoral	Yes	Model 3	Yes	None	Gelatin	Formaldehyde
19.8 (238)	Aneurysm	Anastomotic false aneurysm	Aortobiiliac	No	Model 4	No	External	Gelatin	Formaldehyde
21.0 (252)	Intermittent claudication	Nonanastomotic rupture	Above-the-knee femoropopliteal	No	Model 5	No	Double	Collagen	Formaldehyde
NA	Chronic limb ischemia	NA	NA	Yes	Model 2	Yes	None	Collagen	NA
8.7 (104)	Chronic limb ischemia	Infection	Above-the-knee femoropopliteal	No	Model 1	No	External	Collagen	Glutaraldehyde
7.4 (89)	NA	Nonanastomotic false aneurysm	Axillobifemoral	No	Model 4	No	External	Gelatin	Formaldehyde
6.4 (77)	NA	Nonanastomotic rupture	NA	No	Model 6	No	None	Collagen	NA
22.8 (273)	Aneurysm	Nonanastomotic rupture	Aortobifemoral	No	Model 7	No	External	None	None
9.3 (111)	Intermittent claudication	Nonanastomotic false aneurysm	Aortofemoral	No	Model 4	No	External	Gelatin	Formaldehyde

CTA, Computed tomography angiography; NA, not available.

We collected seven different models of VGs from Perouse Medical, InterVascular, Vascutek, C.R. Bard, and Baxter International. Graft models were anonymized and substituted with numbers 1 through 7 in the table.

Nine of them were thin-wall VGs (wall thickness <0.4 mm). Eighteen VGs were coated (4 with gelatin and 14 with bovine collagen). Eight VGs had an external velour surface only with a nonvelour inner surface, and one VG was double velour. Nine VGs were glutaraldehyde and formaldehyde free.

The mean duration of implantation was 9.2 years (range: 2.3-22.8 years).

Thirteen VGs were implanted as infrainguinal bypasses (11 above-the-knee femoral-popliteal bypasses, 1 femorofemoral bypass, and 1 crossed femorofemoral bypass), three VGs were implanted at the aortic level (one aortobifemoral bypass, one aortobiiliac bypass, and one aortofemoral bypass), and one VG was implanted as an axillobifemoral bypass.

The main reason for explantation was a nonanastomotic rupture with false aneurysm (n = 16), infection (n = 1), and thrombosis requiring the redo procedure (n = 1). The reason for explantation was missing in one case. When VGs were explanted for nonanastomotic rupture or thrombosis requiring the redo procedure, infection was excluded based on the clinical reports provided by the surgeon as well as the CTA data when available.

CTA reconstructions

Preexplantation CTA images were available for eight VGs: six femoral-popliteal bypasses, one aortofemoral bypass, and one cross-femorofemoral bypass. The results of CTA analysis are listed in Table II.

Table II.

Conclusions drawn from the analysis of pre-explantation computed tomography angiography (CTA) of the vascular grafts (VGs)

VG	Site of implantation	False aneurysm location	Presence of small dilatation along the bypass	Remarkable event
VG2	Above-the-knee femoral-popliteal	Adductor canal	No	No
VG3	Above-the-knee femoral-popliteal	Adductor canal	No	The distal anastomosis seemed to pull on the popliteal artery
VG4	Above-the-knee femoral-popliteal	Proximal third thigh	Yes (adductor canal)	No
VG6	Above-the-knee femoral-popliteal	Adductor canal	Yes (adductor canal)	No
VG8	Above-the-knee femoral-popliteal	Junction between the proximal and middle third thigh	No	No
VG9	Above-the-knee femoral-popliteal	Mid thigh	No	No
VG11	Femorofemoral	Left inguinal ligament	Yes (right inguinal area and middle of the graft)	Graft kinking
VG19	Above-the-knee femoral-popliteal	Inguinal ligament	No	No

False aneurysms were observed at the adductor canal level on four EVGs (Fig 1, A), at the inguinal ligament level on two EVGs (Fig 1, B), and at the proximal or middle third thigh level on three EVGs. We observed on three CTAs more than one false aneurysm along the EVGs.

Fig 1.

A, Reconstruction of the pre-explantation computed tomography angiography (CTA) of vascular graft 3 (VG3): above-the-knee femoral-popliteal bypass presenting a false aneurysm at the middle third thigh level in the adductor canal. The distal anastomosis seems to pull on the popliteal artery. Duration of implantation: 6.8 years. B, Reconstruction of the pre-explantation CTA of VG11: cross-femorofemoral bypass presenting a left false aneurysm, a smaller right femoral false aneurysm, and a kinking in the middle of the prosthesis. Side-by-side comparison of the explanted graft before cleaning. Duration of implantation: 3.5 years.

On one CTA it seemed that the distal anastomosis of the VG exerted a traction on the native artery (Fig 1, A). One CTA displays a kinking on the VG that was also observed on the corresponding EVG (Fig 1, B).

Types of degradations

After cleaning, all EVGs revealed one or several nonanastomotic ruptures. Observations on degradations are presented in Table III.

Table III.

Characteristics of observed degradations on the explanted vascular grafts (VGs) after cleaning, macroscopical and microscopical analysis

VG	Picture	Number of ruptures	Length, mm	Minimal break initiating rupture	Type of ruptures								Localization of the rupture					Aging degradation
					Longitudinal	Transversal	45°	V-shaped	Punctiform	Combined mechanisms	Due to endoprosthesis inside the graft	Due to explantation	Anastomotic	Remeshing line	Guideline	Crimping	Other	Crimping fatigue	Loosed meshes	Loosed remeshing line
VG1		8	167	Yes	2	3	2	0	1	Yes	0	0	0	1	3	2	2	Yes	No	Yes
VG2		7	253	Yes	2	2	2	0	0	Yes	0	1	0	0	1	2	4	Yes	No	Yes
VG3		8	89	Yes	3	4	1	0	0	Yes	0	0	0	2	1	4	1	Yes	No	Yes
VG4		5	80	Yes	0	3	0	1	0	Yes	0	1	0	1	1	3	0	Yes	No	Yes
VG5		25	162	Yes	6	17	0	2	0	Yes	0	0	0	2	4	17	2	Yes	No	Yes
VG6		3	173	Yes	1	2	0	0	0	No	0	0	0	1	0	2	0	Yes	No	Yes
VG7		1	51	Yes	0	0	0	1	0	Yes	0	0	0	0	1	0	0	Yes	No	Yes
VG8		3	151	Yes	1	1	0	0	0	Yes	0	1	0	1	0	1	1	Yes	No	Yes
VG9		1	55	No	0	0	0	0	1	No	0	0	0	0	0	0	1	Yes	No	No
VG10		3	310	Yes	0	1	0	1	0	No	0	1	0	1	0	1	1	Yes	No	No
VG11		3	130	Yes	3	1	0	0	0	Yes	0	0	1	0	1	1	0	Yes	No	Yes
VG12		5	117	No	0	0	0	0	0	No	4	1	0	0	0	0	5	Yes	No	Yes
VG13		7	44	Yes	0	1	5	1	0	No	0	0	0	1	1	1	4	Yes	Yes	Yes
VG14		8	197	Yes	2	5	0	1	0	Yes	0	0	0	1	2	5	0	Yes	No	Yes
VG15		4	41	No	2	0	0	1	0	Yes	0	0	0	1	1	0	1	Yes	No	Yes
VG16		3	70	No	0	2	0	0	0	No	0	1	0	0	0	2	1	No	No	No
VG17		1	10	No	0	1	0	0	0	No	0	0	0	0	0	1	0	Yes	No	Yes
VG18		3	35	Yes	0	2	1	0	0	Yes	0	0	0	1	0	2	0	No	No	No
VG19		1	51	Yes	0	1	0	0	0	No	0	0	0	0	0	1	0	Yes	No	No
Total number of VGs (n = 19)	14	9	15	5	7	2	11	1	6	1	11	10	15	11	17	1	14

The mean length of the collected EVGs was 115 mm. Fifteen EVGs presented more than one rupture. Fourteen EVGs presented additional initiations of ruptures outside of the ruptured areas. The macroscopic analysis revealed longitudinal ruptures on 9 EVGs, transversal ruptures on 15 EVGs, 45°-oriented ruptures on 5 EVGs, V-shaped ruptures on 7 EVGs, and punctiform ruptures on 2 EVGs. Ruptures with multiple mechanisms were observed on 11 EVGs. Six EVGs presented ruptures due to the explantation process. One EVG presented ruptures due to the insertion of an endoprosthesis inside the graft.

Ruptures involved the remeshing line on 11 EVGs, the guideline on 10 EVGs, and the crimping valley on 15 EVGs. An anastomotic rupture was observed on one EVG.

All the transversal ruptures observed occurred on the crimping (Fig 2).

Fig 2.

A, Microscopic examination (×20) of a transversal rupture on the crimping of vascular graft 6 (VG6) (above-the-knee femoral-popliteal bypass with a duration of implantation of 8.2 years) with the schematic representation of potential traction forces exerted on the damaged area. B, Microscopic examination (×50) of VG3 (above-the-knee femoral-popliteal bypass with a duration of implantation of 6.8 years) showing broken filaments along the guideline. C, Microscopic examination (×30) of VG4 (above-the-knee femoral-popliteal bypass with a duration of implantation of 4.8 years) showing broken filaments forming a transversal rupture after the crimping valleys.

Microscopic examinations

At the microscopic level, two main degradation phenomena could be identified: (1) a decrease in the density of the meshing (Fig 2, A); (2) local ruptures of the PET fibers (Fig 2, B). Aging degradations were seen in majority on the crimping valleys and the remeshing line, with 14 EVGs presenting a loosening of the remeshing line (Fig 3) and 17 EVGs presenting an attenuation of the crimping. One EVG also presented loosed meshes outside of those areas.

Fig 3.

A, A V-shaped rupture on vascular graft 7 (VG7) (femorofemoral bypass with a duration of implantation of 7.9 years) with simulation of the graft aspect before the rupture and schematization of the combination of potential traction, flexion, and torsion forces exerted on the textile. B, Macroscopic examination (×30 and ×50) showing loosened meshes on the edges of the rupture, mostly on the remeshing line.

Specific cases

One explant (VG9) presented only one punctiform rupture with clean edges (Fig 4). The rupture was circular. Its diameter included five rows of meshes, and the center of the hole was located on top of the crimping. After cleaning, the borders were characterized by yarn fuzzing. Broken filaments appeared as pulled out through the external surface. The tongue-like pattern of some broken filaments suggested local ruptures due to fatigue.

Fig 4.

A, Macroscopic observation of vascular graft 9 (VG9) (above-the-knee femoral-popliteal bypass with a duration of implantation of 3.2 years) after cleaning, presenting a single circular rupture with clean regular edges. B, Microscopic examination (×50) showing broken filament in a tongue-like pattern all around the edges of the rupture.

One explant (VG12) presented an endoprosthesis implanted inside the graft (Supplementary Fig, online only). The ruptures observed on the graft’s textile were attributed to the stents of the endoprosthesis.

Discussion

The main findings presented in the frame of this study point out that new generations of VGs can still undergo degradations despite improvement efforts performed in the manufacturing technology. Previous reports of graft degradations led to modifications in textile design such as the selection of warp-knitted and woven structures,^4,5 or the incorporation of textured fibers with cylindrical filaments.¹³ Improvement of porosity led to the development of impregnated protheses with different types of sealing bioresorbable molecules.¹⁴ Newer polyester knitted VGs implanted after 1990 in this study are all warp-knitted prostheses. They introduce the concept of thin-wall VGs, which are supposed to offer better compliance and conformability with a better resemblance to native arteries. Other new-generation characteristics are the absence of a velour structure or the limitation to an external velour structure and a coating agent (either bovine collagen or hydrolyzable gelatin).

Previous VG generation failures have been described as a limited number of longitudinal failures.¹⁵ Observed degradations in this study ranged from a majority of transversal ruptures, to transversal, V-shaped, and 45° ruptures, and two punctiform ruptures. Frequent areas of degradations were identified as remeshing lines, guidelines, and crimping valleys. The mechanisms of degradation at the microscopic level were either a decrease in the density of the meshing or local rupture of textile fibers. As several models have been included in the analysis, one can conclude that this issue can concern all the currently available models.

One key finding here is that the degradation tends to occur preferentially in some graft zones where the textile VG was already weakened at the manufacturing level like remeshing lines or crimping peak or valley. A remeshing line is a zone that presents a discontinuity regarding the way yarns are crossed in the knitted construction. Although the line is necessary to obtain a tubular knitted shape, the local structural change induces stress concentration.

With respect to crimp peaks and valleys, the yarns located in these zones have been pressed against the crimp molding device and heated above vitreous transition temperature. This has locally fragilized the structure. These observations have already been made in the past. We previously demonstrated that ruptures observed on explanted knitted PET VGs occurred on these areas that could be qualified as areas of weakness.¹⁵ The guideline and the remeshing line have therefore been identified as weak spots potentially leading to longitudinal ruptures. Further chemical and mechanical analyses concluded that these areas of weakness could stem from the manufacturing steps, inducing alterations in the PET yarns and resulting in premature aging of the VG structure.^16,17

No difference was seen between the grafts that could explain the tendency of graft degradation according to the duration of implantation. However, it is always difficult to identify the precise cause of the degradation, as several factors can be involved such as VG deformation stress (extension, flexure, and torsion), biodegradation by tissue fluids and enzymes, or textile wear when the VG is in contact with an abrasive zone, or locally compressed by a neighbor vessel or outgrowing body part.¹⁸ However, imaging can help interpreting the degradation cause. It can help to establish a correlation between the degradations observed and the position of the degraded zone in the body. This information is of great interest for VG manufacturers and clinicians in order to improve the quality of the devices at the manufacturing level and the way they are used in clinical practice.

The rupture patterns observed in the frame of this work pointed out several rupture mechanisms, which can be partly explained by the stress applied to the VG in vivo. The main stress patterns applied to a prosthesis are extension, flexure, and torsion.

First, the transverse ruptures can be typically caused and explained by the involvement of a flexure stress. When the VG is bent, fibers located at the level of the VG external curvature undergo locally high extension stress. As the loading is cyclic under pulsed cardiac flow, the mesh starts to become loose and the fibers tend to break presenting a tongue-shaped fatigue rupture facies. The rupture initiated at the top of the curvature then tends to propagate radially along the fragilized crimp valleys or peaks, as can be seen in Fig 2. Second, if some additional stress like extension is applied to the VG, the direction of the initial radially oriented rupture is modified and becomes inclined. At last, the V-shaped ruptures can suggest the involvement of torsion stress because these are not symmetrically shaped, as seen in Fig 3.

One key finding of this work is that this combination of stress applied to the VG can be partly explained from the CTA images. As shown in the analysis of pre-explantation CTA, a recurrent location for false aneurysms of femoral-popliteal bypasses tends to be the adductor canal. The adductor canal is an anatomic area known for the strong mechanical constraints it applies on the superficial femoral artery.¹⁹ There, the VGs could undergo axial compression and extension, radial compression, flexion, and torsion. Unlike the natural superficial femoral artery, VGs are lacking elastic fibers, thus resisting less efficiently to mechanical forces. The same process can be hypothesized for the inguinal ligament area.²⁰ Anatomic constraints could then partly explain the contrived aging process of the textile leading to rupture.²¹ Moreover, some of the VGs presented a local creasing that could have been present before explantation as can be seen in some CTA reconstruction. If the VG was bent in vivo because of anatomic disposition, the blood flow inside the VG would circulate in a curved line, producing a pulsed traction at both extremities of the curve. The knitted textile would have been strongly solicited in these areas (especially on the outward of the curve where the traction constraint is the most prominent), creating a fatigue of the PET fibers and interfilament friction leading to rupture. The curved flow would also create a cyclic flexion inside the creasing that would generate more flexion forces in the fibers.

Understanding the mechanisms of VG degradation is mandatory. Investigating these mechanisms with textile engineers, manufacturing engineers and physicians could then allow late VG rupture in the future. First, carrying out mechanical strength tests on virgin VGs of different models, before and after the compaction process and thermic treatment, would allow a better understanding of the mechanisms of rupture related to the weaknesses induced by the manufacturing steps of VGs. Another prospect would be to carry out histological analysis of these types of EVGs in order to assess in vivo interactions with biological tissues, the graft encapsulation and healing,²² and the interaction with the coating and with the textured fibers, especially in the area of the textile rupture. Although experience allows us to rule out degradations induced by explantation techniques and surgical handling of the VG, some lesions can be of difficult interpretation. In the case of the punctiform rupture on VG9, however, macroscopic and microscopic analyses alone do not allow us to conclude toward a specific mechanism of degradation.

The prevalence of VG failures is still underestimated to this day and difficult to truly evaluate as only few cases are reported, highlighting the need to keep studying vascular explants. The main interest of our retrieval program is to allow the analysis of a significant number of VGs to be able to point out mechanisms of aging of different models of devices and to learn about the concept itself. However, a study based on a retrieval program has unavoidable limitations that must be taken into consideration when interpreting the results because one may consider that the probability of finding structural lesions onto VGs for a complication is higher than on uncomplicated VGs. Moreover, the analysis must considered that lesions could have been created during the explantation of the VG. However, our expertise and the fact that all analyses of EVGs are performed using a multistep standardized protocol systematically associating clinicians and engineers can exclude the risk of misinterpretation.

Conclusions

Explant analysis tends toward the conclusion that VG degradation results from both anatomic constraints and intrinsic textile structure. The combination of stress in traction, flexion, and torsion with the cyclic arterial flow exerts strong pressures on the warp-knitted textile in the localized area of weaknesses, producing a global loosening of the mesh and the loss of the crimping, and then leading to rupture. These degradations still exist in today’s generation of VGs, reinforcing the need to keep studying their mechanisms in order to develop a substitute for vascular repair that would be performant in terms of both compliance and durability. Preoperative imaging can become a great tool in both the manufacturing and clinical process as it is shown to help in the interpretation of degradation mechanisms.

Author Contributions

Conception and design: NC, FH, AL

Analysis and interpretation: AB, SK, DD, LS

Data collection: AB, SK

Writing the article: AB, NC, FH, AL

Critical revision of the article: SK, DD, LS

Final approval of the article: AB, NC, SK, DD, LS, FH, AL

Statistical analysis: Not applicable

Obtained funding: Not applicable

Overall responsibility: AL

Appendix

Collaborators: Samuel Bouttier (Clinique Claude Bernard, Ermont, France), Pierre Carlier (Centre Médico-Chirurgical Floréal, Bagnolet, France), Jacques Chevalier (Hôpital Saint-Philibert, Lomme, France), Sébastien Deglise (Centre Hospitalier Universitaire Vaudois, Lausanne, Switzerland), Denis Garnier (Clinique générale d'Annecy, France), Adrien Hertault (Centre Hospitalier de Valenciennes, France), Rémi Laurent (Hôpital Saint-Philibert, Lomme, France), Boris Postaire (Centre Hospitalier Universitaire de Nantes, France), Pascal Vernon (Centre Hospitalier de Saint-Quentin, France), Renaud Vidal (Hôpital La Casamance, Aubagne, France).

Supplementary Fig (online only).

Macroscopic observation of vascular graft 12 (VG12) (aortobiiliac bypass with a duration of implantation of 19.8 years) after cleaning, presenting four ruptures facing the stents of the endoprosthesis.