A new design concept for knitted external vein-graft support mesh

journal of the mechanical behavior of biomedical materials 48 (2015) 125–133

Available online at www.sciencedirect.com

www.elsevier.com/locate/jmbbm

Research Paper

A new design concept for knitted external vein-graft support mesh Charanpreet Singha, Xungai Wanga,b,n a

Australian Future Fibres Research and Innovation Centre, Institute for Frontier Materials, Deakin University, Geelong, Victoria 3216, Australia b School of Textile Science and Engineering, Wuhan Textile University, Wuhan 430073, China

art i cle i nfo

ab st rac t

Article history:

Autologous vein-graft failure significantly limits the long-term efficacy of coronary artery

Received 23 January 2015

bypass procedures. The major cause behind this complication is biomechanical mismatch

Received in revised form

between the vein and coronary artery. The implanted vein experiences a sudden increase

31 March 2015

(10–12 fold) in luminal pressures. The resulting vein over-distension or ‘ballooning’

Accepted 1 April 2015

initiates wall thickening phenomenon and ultimate occlusion. Therefore, a primary goal

Available online 9 April 2015

in improving the longevity of a coronary bypass procedure is to inhibit vein overdistension using mechanical constriction. The idea of using an external vein-graft support

Keywords:

mesh has demonstrated sustained benefits and wide acceptance in experimental studies.

Knitted mesh

Nitinol based knitted structures have offered more promising mechanical features than

Vein-graft

other mesh designs owing to their unique loosely looped construction. However, the

Compliance Coronary artery bypass graft Segmented knitting

conventional plain knit construction still exhibits limitations (radial compliance, deployment ease, flexibility, and bending stresses) which limit this design from proving its real clinical advantage. The new knitted mesh design presented in this study is based on the concept of composite knitting utilising high modulus (nitinol and polyester) and low modulus (polyurethane) material components. The experimental comparison of the new design with a plain knit design demonstrated significant improvement in biomechanical (compliance, flexibility, extensibility, viscoelasticity) and procedural (deployment limit) parameters. The results are indicative of the promising role of new mesh in restoring the lost compliance and pulsatility of vein-graft at high arterial pressures. This way it can assist in controlled vein-graft remodelling and stepwise restoration of vein mechanical homoeostasis. Also, improvement in deployment limit parameter offers more flexibility for a surgeon to use a wide range of vein diameters, which may otherwise be rendered unusable for a plain knit mesh. & 2015 Elsevier Ltd. All rights reserved.

Abbreviations: CABG,

Coronary artery bypass graft; AVG,

Autologous vein graft; IH,

Intimal hyperplasia; PET,

Polyester;

NT, Nitinol; PCU, Polycarbonate urethane; LM, Low modulus; HM, High modulus; DL, Deployment limit n Corresponding Author at: Institute for Frontier Materials, Deakin University, Geelong, Victoria 3216, Australia. Tel.: þ61 3 5227 2894; fax: þ61 3 5227 2539. E-mail address: [email protected] (X. Wang). http://dx.doi.org/10.1016/j.jmbbm.2015.04.001 1751-6161/& 2015 Elsevier Ltd. All rights reserved.

126

1.

journal of the mechanical behavior of biomedical materials 48 (2015) 125 –133

Introduction

Coronary artery bypass graft (CABG) surgery is one of the most frequently performed surgical procedures in the United States, with over 400,000 procedures performed annually (Members et al., 2011). Autologous saphenous vein graft remains the graft of choice for surgeons performing CABG procedures (Vijayan et al., 2002). However, failure of autologous vein grafts (AVG)s still remains a major problem and has been reported as high as 25% within 12–18 months after surgery (Members et al., 2011). A number of factors contribute towards AVG failure namely, early thrombosis, atherosclerosis, intimal hyperplasia (IH), and compliance mismatch (Davies et al., 1992; Harskamp et al., 2013). However, IH has been a major investigative factor for decades owing to its significant contribution towards the failure (20–40% within the first 5 years) (Davies and Hagen, 1995). The initiation of IH is due to an abrupt exposure of AVG to the dynamic environment of arterial circulation (Davies and Hagen, 1995; Harskamp et al., 2013). Post-implantation, the AVG, which has experienced luminal pressures only up to 10 mmHg, is suddenly subjected to high arterial pressures (100–120 mmHg) (Dobrin et al., 1989; Powell and Gosling, 1998). In such conditions, circumferential wall stress in an AVG can increase 140-fold compared to that in a vein under normal circumstances (Liu and Fung, 1998). An uncontrolled wall thickening (or lumen narrowing) initiates as a response mechanism of AVG to return back to normal venous stress levels and normalize tangential wall stress (Dobrin et al., 1989; Vijayan et al., 2002). An AVG can become stiff beyond distending pressures of 75 mmHg owing to elastin fibre degeneration (Ozturk et al., 2013) and thus exhibit low compliance in much higher physiological pressure range of coronary artery. The formation of IH particularly at the heel and toe of the distal anastomosis initiates as a result of compliance mismatch between artery and vein (Kidson, 1983; Tiwari et al., 2003). An external mesh reinforcement has proven to be a promising technique in preventing over-distension of AVGs and IH development (Desai et al., 2010). Since its first trial in 1963 (Parsonnet et al., 1963), several studies have verified this effect using various mesh constructions (electrospun (ElKurdi et al., 2008), braided (Ben-Gal et al., 2013; Krejca et al., 2002; Zilla et al., 2008, 2009), woven (Jeremy et al., 2004), knitted (Moodley et al., 2013; Murphy et al., 2007; Schoettler et al., 2011; Zilla et al., 2011)) and different materials (polyurethane (El-Kurdi et al., 2008), polyester (Krejca et al., 2002; Longchamp et al., 2014; Murphy et al., 2007; Trubel et al., 1994), metal alloy (Ben-Gal et al., 2013), polyglactin (Jeremy et al., 2004; Vijayan et al., 2004), polytetrafluoroethylene (Kohler et al., 1989), nitinol (Zilla et al., 2008, 2009, 2011)). However, none of these trial meshes has been successfully included into regular clinical practice except the CE certified eSVSs mesh (Kips Bay Medical, Minneapolis, MN USA), a knitted nitinol mesh (Emery et al., 2012). The device is not yet U.S. Food and Drug Administration (FDA) approved and is currently undergoing clinical feasibility trial (eMESH trial). The manufacturers claim that the eSVSs mesh can restrict diameter of AVG and exhibit pulsatile flows similar to an artery. Till date there are very few studies which have reported the clinical performance of eSVSs mesh in animals (Moodley et al., 2013; Zilla et al., 2011) and humans (Genoni

et al., 2013; Schoettler et al., 2011). One of the first studies comparing a knitted nitinol with a braided nitinol mesh as femoral artery graft support was conducted on nonhuman primate models (Chacma baboons) (Zilla et al., 2011). The knitted mesh exhibited significantly better handling and biomechanical (radial compliance, bending stiffness, kinkfree radius, radial narrowing) properties than the braided mesh with nearly total suppression of IH and more physiologic remodelling of AVG media. Knitted structures are inherently flexible constructions and hence a promising design approach for an external support mesh. However, the limitations reported by the short term trials of eSVSs mesh cannot be ignored and can become an early call for further improvisation in its design. The kink-free configuration of knitted mesh allows easy adaptation to anatomic bends and curves (Zilla et al., 2011). However, nitinol wire in a looped knit configuration can undergo hemodynamic strain specifically at anatomical bends, and exhibit breakages (Moodley et al., 2013; Murphy et al., 2007; Zilla et al., 2011). A numerical study by van der Merwe and colleagues has attempted to provide solution to minimise loop breakage by suggesting the use of an even knit loop design and use of thinner nitinol wires (∅o0.05 mm) (van der Merwe et al., 2008). Longitudinal flexibility of knitted mesh provides high length and diameter stability during implantation procedure (Zilla et al., 2011). On the contrary, this property also makes the deployment stage a traumatising event for the AVG tissue as it involves an uneasy feeding of AVG from one end of an insertion straw and pulling it from other end (Zilla et al., 2011). This problem can be more prominent in longer AVGs and arises due to insufficient circumferential extensibility of the knitted mesh to accommodate a high calibre insertion straw. It was numerically computed that free movement of knitted loops allows diameter extension in a knitted nitinol mesh (∅¼3.34 mm) but only to limited pressure levels (up to 15 mmHg) which is far below the physiological arterial pressure (van der Merwe et al., 2008). As the pressure increases, the looploop interlocking jams the knit structure with no further increase in mesh diameter owing to inextensibility of the nitinol wire itself. The compliance of knitted nitinol mesh (∅¼ 3.37 mm, wire thickness¼0.05 mm) was less than half of the artery (Zilla et al., 2011), which may render such plain knit mesh design incapable of producing the required pulsatile flow and hence prone to long-term failure (Abbott et al., 1987; Trubel et al., 1994; Weston et al., 1996). The use of thinner nitinol wires (∅o0.05 mm) has been suggested to improve compliance in a computational study (van der Merwe et al., 2008) but not yet experimentally proved. A critical issue with a vein-graft is that it shows high compliance in low-pressure range (o10 mmHg) while remains inextensible in the high-pressure range (450 mmHg) (Stooker et al., 2003). An ideal external support mesh should be capable of enhancing the compliance property of an AVG even while acting as a mechanical constriction to the AVG wall. The combined use of a non-compliant AVG and a stiff mesh can further deteriorate the biomechanics of the treated artery (Trubel et al., 1994). Diameter mismatch in a compliant mesh is not as big an issue as a non-compliant mesh with matched diameter (Trubel et al., 1994). A non-complaint mesh graft with matched diameter will automatically become undersized after implantation as the diameters are generally matched at diastolic pressures (Weston et al., 1996).

journal of the mechanical behavior of biomedical materials 48 (2015) 125 –133

instead of a single thick filament was to achieve high flexibility while maintaining similar compression resistance as of a thick filament. PET allows for low temperature (100– 125 1C) annealing of mesh compared to a NT wire, while the latter provides higher shape retention property. On the other hand, a melt spun polycarbonate-urethane (PCU, ∅ ¼ 225 mm, Modulus¼ 10 MPa, Breaking strain¼ 850%) filament was selected to incorporate longitudinal flexibility, radial extensibility and pulse compliance property in the mesh.

The radial distensibility or compliance property of a vessel is proportional to the diameter-wall thickness ratio (D/h) and inversely proportional to Young's modulus (E) of the vessel wall material i.e. dD/D  dP¼ D/2h  E, where D is the initial diameter of vessel and dD is change in diameter for an increase in pressure (dP). Therefore, a possible way to achieve an improved compliance property is to reduce E while D/2h remains constant with increase in pressure. Since structural support is also a major criterion for mesh application, the E value cannot be very low for the whole mesh construction. The solution to this limitation forms the basis of the proposed segmented (high and low modulus segments) design concept, hypothesised to provide structural support and compliance property, simultaneously. The rationale behind this approach is that a complaint veingraft can undergo arterialisation in a more physiologic manner compared to mere atrophy of smooth muscle cells in a noncompliant vein-graft (Szilagyi et al., 1973; Zilla et al., 2011). The goal of this study is to enhance this capability in a knitted mesh design using a novel approach of segmented knitting and optimisation of segmentation parameters. The proposed design concept is inspired from the structural construction of a caterpillar cuticle and has previously demonstrated promising results in similar applications (Singh and Wang, 2014).

2.

Materials and methods

2.1.

Material

2.2.

1/1

Mesh construction

Two types of mesh designs (inner ∅¼ 5 mm) were manufactured: plain knit (PK) and segmented knit (SK). A customised eight needle (head ∅¼ 5.5 mm) knitting machine (Lamb Kmc Inc., MA) was used for manufacturing the samples. The PK sample was knitted using PET and NT only, while SK sample was developed by alternate feeding of PET/NT and PCU monofilaments for specific number of knit courses. A single knit course here denotes one complete row of knit loops around the mesh circumference. This type of knitting pattern generates a tubular structure with alternately arranged high modulus (HM) and low modulus (LM) segments (Fig. 1). Based on the number of knit courses in HM and LM segment, four types of SK samples were made; 1/1, 2/2, 3/3, and 4/4 denoting one, two, three and four courses in each HM and LM segment, respectively (Fig. 1). The PK sample is also referred to as 0/0 denoting absence of any segments. The theoretical volume fraction of each component material in all the support meshes is same. Fig. 1 (drawings) explains this in detail where 1/1, 2/2, and 3/3 mesh each have six HM courses and six LM courses along their length with exception to type 4/4 (72 courses). The type 4/4 thus also defined the maximum limit of design segmentation (i.e. going beyond to type 5/5). However, in an actual prototype, the fraction of each component can vary (75 courses per 10 cm length) due to inherent tendency of LM filament loops to relax after being taken off from the knitting machine.

Three types of materials in filament/wire form were used for mesh construction: polyester (Trevira GmbH, Germany), superelastic nitinol (Fort Wayne Metals, Ind) and ther moplastic-polycarbonate-urethane (DSM Medical, US). The polyester filament (PET, ∅ ¼ 200 mm, Modulus¼825 MPa, Breaking strain¼48%) and nitinol wire (NT, ∅ ¼120 mm, Modulus¼ 10 GPa, Breaking strain¼ 28%) were selected to provide structural stability and compression strength to the mesh structure. The rationale behind selecting two thin filaments

0/0

127

2/2

3/3

4/4

LM LM HM

LM HM

LM

HM

HM HM

LM HM

LM HM LM

LM HM

HM

LM

LM LM HM LM

HM HM HM

HM

Fig. 1 – Design pattern and structure of plain knit (0/0) and segmented knit (1/1, 2/2, 3/3, 4/4) mesh samples. HM: High modulus component, LM: Low modulus component.

128

2.3.

journal of the mechanical behavior of biomedical materials 48 (2015) 125 –133

Longitudinal flexibility

The mesh samples (length¼ 100 mm) were extended (100 mm/min) axially on a tensile tester machine (Instrons Model 5967). Load elongation curves were recorded for each sample and compared for slope of curve.

2.4.

Bending stiffness

A steel rod (length¼50 mm) was inserted in the mesh sample (length¼100 mm). Bending stiffness was measured as the bending moment (N mm) generated while bending free length of the sample by an angle of 901. The test setup was fixed with 100 N load cell attached on a tensile tester (Instrons Model 5967).

2.5.

SR% ¼ ðσ 0  σ 360 Þ  100=σ 0 where σ0 and σ360 denote initial peak stress and remaining stress after 360 s, respectively.

Radial compliance

Radial compliance was measured using an optical measurement setup (Fig. 2). A small diameter silicon balloon was inserted into the mesh sample (length¼ 100 mm) and pressurised using controlled compressed air inlet. The balloon dimensions (diameter¼ 5 mm, thickness¼0.3 mm) were chosen to simulate a tightly fitting AVG inside the mesh. The pressure inside the balloon was increased at intervals of 20 mmHg and measured using a pressure transmitter (PR-33X/80794, Keller AG.). The corresponding change in mesh diameter was measured (and averaged) over the entire mesh length with a portable laser micrometer (Metralight Inc, USA). The average reading was used to calculate radial compliance Cθ ¼ dD/(dP  Do), where Do is the diameter at initial pressure, and dD is change in diameter over a pressure increment of dP.

2.6.

cutting mesh tubes to obtain circular rings (width¼ 10 mm). The circular ring comprised of 8–9 courses across its width with equal contribution of both HM and LM sections. The care was taken to keep the HM loops on the outer edges to prevent the rings from fraying. The specimens were mechanically preconditioned at 12% strain with 20 loading and unloading cycles to ensure that the stress–strain curves start overlapping each other. Afterwards, circumferential strain was applied (Instrons tensile tester) by stretching the specimen to a fixed strain (10%) at 100 mm/min. The strain was kept constant for 360 s in stretched state. Normalised stress relaxation curves were plotted and relaxation stress (SR) was calculated as:

Static viscoelasticity

2.7.

Deployment limit

Deployment limit (DL) was the term used to determine the maximum AVG diameter which can be loaded into the mesh with sufficient ease. This parameter is dependent on the circumferential extension limit of mesh or the maximum diameter of insertion straw which can be pre-loaded into the mesh. Experimentally, DL was measured by manually inserting a gradually marked conical plastic tube inside the mesh samples. The maximum diameter limit was noted as the point beyond which the mesh sample was not able to slide forward. DL was expressed as the ratio of maximum diameter to initial mesh diameter.

2.8.

Viscoelastic behaviour of mesh samples in circumferential direction was evaluated using stress-relaxation experiments. Static stress-relaxation test is a good tool for comparing viscoelasticity of different materials (Hasegawa and Azuma, 1979; Zou and Zhang, 2011). Test specimens were obtained by

Statistical analysis

The test specimens (n¼ 10 per group) were kept same for all tests and tested in the order; (1) imaging, (2) bending stiffness, (3) longitudinal flexibility, (4) radial compliance, (5) viscoelasticity, and (6) deployment limit. All the characterisation results (n¼10 per group) were tested statistically for significance of differences

Laser micrometre Software interface Silicone balloon

Air inlet

Mesh sample

Pressure transmitter

Fig. 2 – Experimental setup for measurement of radial compliance of mesh samples.

129

journal of the mechanical behavior of biomedical materials 48 (2015) 125 –133

Results

The segmented mesh types (1/1, 2/2, 3/3, and 4/4) exhibited significantly better longitudinal extension property than the plain knit 0/0 mesh (Fig. 3A). At 20% elongation, the resistance force generated by the 1/1, 2/2, 3/3, and 4/4 mesh types was 2.4, 3.2, 3.5 and 6.5 times lower than the 0/0 mesh, respectively. The difference between SK and PK mesh behaviour was evident from the initial stages of the extension (Fig. 3B) indicating the role of PCU filament in bearing the extension strain. The difference in elongation behaviour within the SK meshes was insignificant until 10% strain. Beyond 15% elongation, the 4/4 mesh exhibited significantly (po0.05) lower stress response than the rest of the mesh types. The effect of mesh longitudinal extensibility was also observed in bending stiffness property (Fig. 4). The maximum (4.2 times) reduction in bending stiffness was observed in 3/3 and 4/4 mesh samples. The stiffness increased steeply beyond 2/2 with both 1/1 and 0/0 mesh types exhibiting significantly (po0.01) higher stiffness than the 2/2, 3/3 and 4/4 mesh samples. Radial compliance as expressed by pressure–diameter curves showed a significant improvement in SK compared to the PK mesh design (Fig. 5). The 0/0 mesh exhibited a maximum diameter increase of 4.6% at 140 mmHg compared to an average 17.8% increase observed in SK meshes. In the physiological pressure range (80/120 mmHg) the compliance of 0/0 mesh was very low at 1.45  10  4 mmHg  1 while that of meshes 1/1 (9.8  10  4 mmHg  1), 2/2 (8.9  10  4 mmHg  1), 3/3 (8.04  10  4 mmHg  1), and 4/4 (5.89  10  4 mmHg  1) was 6.8, 6.1, 5.5, and 4.1 times higher than 0/0 mesh, respectively. The stress–relaxation curves showed the effect of low modulus PCU segments and their pattern on degree and rate of stress relaxation (Fig. 6). The 0/0 mesh exhibited a stress relaxation of 18% after 360 s with meshes 4/4 and 3/3 exhibiting nearly similar (18.1% and 19%, respectively) results. However,

2.5

Bending moment (N.mm)

3.

the degree of stress relaxation was higher in 1/1 (21.1%) and 2/2 (23%, po0.05) samples. Also, the initial (o30 s) rate of relaxation was high in both these samples. Similar to the radial compliance results, the increasing order of diameter extensibility for SK mesh was observed in

2

1.5

1

0.5

0 0/0

1/1

2/2

3/3

4/4

MeshID

Fig. 4 – Bending resistance of mesh samples expressed as bending moment (Mean7SD, n ¼ 10 per group). 6.2

1/1

Diameter (mm)

using one way ANOVA analysis. Post-hoc t-tests were used for determining statistical difference between groups.

6

2/2

5.8

3/3 4/4

5.6

5.4

0/0

5.2

5

0

20

40

60

80

100

120

140

Pressure (mmHg)

Fig. 5 – Pressure–diameter curves depicting the radial compliance (Mean7SD) property of mesh samples. Each curve is representative average of 10 specimens per group.

0.6 0.5 0.4

14

0.3

12

0/0 1/1 2/2 3/3 4/4

0.2 10

0 0

1

Force (N)

0.1

8 6 4 2 0 0

2

4

6

8

10

12

14

16

18

20

Elongation %

Fig. 3 – Longitudinal extensibility of mesh samples expressed by load–elongation curves. Each curve is representative average of 10 specimens per group.

130

journal of the mechanical behavior of biomedical materials 48 (2015) 125 –133

100

the DL test experiment (Fig. 7). The maximum circumferential strain in 1/1 mesh was 1.4 times higher than that observed in the 0/0 mesh (DL ¼2.24 vs. DL¼1.6). The 2/2, 3/3, and 4/4 mesh demonstrated DL values of 2.04, 1.90, and 1.78, respectively.

Normalied stress (%)

95

90 0/0

4/4

3/3 1/1

85

4.

2/2

A new knitted external vein-graft mesh was developed utilising the concept of segmented knitting to improve its compliance and flexibility property. Compliance characteristic works in conjunction with hemodynamic and mechanical events to influence the graft patency (Abbott et al., 1987). An external mesh is beneficial for graft wall but can have a deteriorating effect on the host artery by initiating a severe intimal hyperplasia growth at the anastomosis (Trubel et al., 1994). In order to realise the real advantage of using a mesh constriction on a vein-graft, improvement in compliance property is an inevitable requirement. Compliance mismatch between constricted vein-graft and host artery produces low shear rates which favour intimal thickening (Weston et al., 1996). Therefore, it is very important for a mesh to match host artery diameter in the physiological pressure range rather than just matching in normal atmospheric conditions (0 mmHg). As observed in this study, the diameter strain in the physiological pressure range (80/120 mmHg) of the 0/0 mesh was 0.57% compared to 2.96% (5.2 times higher) in an artery (Kidson, 1983; Tiwari et al., 2003). Similar observation of low compliance of a plain knit nitinol mesh has been also reported elsewhere (Zilla et al., 2011). The two modes by which a tubular knitted mesh can extend circumferentially or exhibit radial compliance are: (a) increase in inter-loop space through loop straightening, and (b) elastic extension of loop material. The parameter which best describes contribution of each mode towards radial compliance is the inter-loop space (Fig. 8). Unlike the latter mode, the total length (number of loops  loop length) of a single knit course remains constant before and after circumferential extension in the

80

75 6

57

107

158

208

259

309

360

Time (sec)

Fig. 6 – Viscoelastic behaviour as expressed by stress relaxation curves of mesh samples. Each curve is representative average of 10 specimens per group and normalised to maximum stress value.

DL = 1.60

DL = 1.78

DL = 1.90

DL = 2.04

DL = 2.24

0/0

8 mm

8.9 mm

4/4

9.5 mm

3/3

2/2

10.2 mm

1/1

11.2 mm

Discussion

Fig. 7 – Deployment limit (DL) test setup showing the maximum achievable diameter by different mesh samples (initial diameter¼5mm) inserted on a conical plastic tube.

Mesh Axis

1 2

Loop junction space Loop length

3 fp

1

1

2

2

3

3

fp

PK mesh

fp

w

1

mmHg

Fc

ec

Fc fp

fp

2 SK mesh 3

1

Inter-loop space

2 Mesh width, w 3 w

ec

Fig. 8 – Circumferential extension behaviour of a knitted mesh comprised of three (1, 2, 3) knit courses. (mmHg: Luminal pressure, Fc: Circumferential force, fp: Pull force exerted by course-2 on courses-1 and 3, ec: circumferential extension, PK: Plain knit mesh, and SK: Segmented knit mesh).

journal of the mechanical behavior of biomedical materials 48 (2015) 125 –133

former mode. Circumferential extension by loop straightening mode involves an increase in inter-loop space with a corresponding decrease in loop length. The contribution of each mode towards mesh compliance is variable and dependent on elastic modulus of loops and their geometrical pattern. This has been explained by considering a section of a tubular knitted mesh (Fig. 8). The section is composed of three knit courses, each consisting of three loops. When the mesh is subjected to intraluminal pressure (mmHg), circumferential forces (Fc) are generated at the end of each course. As this force is equal on every course, the extension behaviour of only a single course (course-2) is considered here for simplicity. The force Fc acting on course-2 exerts a pull force fp on each loop junction created with courses-1 and 3. The resulting loop movement fills the loop junction space resulting in slight circumferential extension of the mesh. However, for an AVG mesh, this initial extension makes practically insignificant contribution to compliance as the circumferential stress (approximately 15 mmHg) corresponding to this level of circumferential extension is already reached during the deployment stage (van der Merwe et al., 2008). Therefore, radial compliance of mesh beyond this point is completely dependent on elastic modulus of the loop as explained by the extension behaviour comparison of a PK and SK knit mesh (Fig. 8). The knit courses-1, 2 and 3 in the PK mesh are constructed from a high modulus filament while courses-1 and 3 are replaced by a low modulus elastic filament in the SK mesh. The circumferential extension (ec) of the SK mesh is significantly higher than PK mesh and involves combined modes of loop straightening (course-2) and loop extension (courses-1 and 3). The low modulus elastic loops undergo high strains as well as allow the adjacent high modulus loops (course-2) to straighten even at low intraluminal pressures (Fig. 8). In the PK mesh, blood pressure force is too low for the high modulus loops (courses-1 and 3) to assist in loop straightening or elongation of course-2 loops. This marks the point (Figs. 8, 9) of structural jamming beyond which no further improvement in radial compliance is observed in the PK mesh. The difference in DL ratio of PK and SK meshes is also a direct indicator of the role of low modulus loops in extending the limits of mesh circumferential extension (Figs. 7, 8). A CABG is positioned on the surface of beating heart and hence subjected to repeated torsional, elongation and pendular movements (John, 2009). Therefore, longitudinal and bending flexibility are important physical properties required from an external support mesh. The experimental results of elongation and bending tests clearly demonstrate the effect of using a low modulus structural component (PCU filament) on stress response compared to a mesh with only high modulus components (PET and NT) (Fig. 3). At the initiation of extension (o2% strain), stress levels were 3.6 times higher in the PK sample than the average of all SK samples. Practically, this level of strain can be easily experienced during AVG insertion and can remain as a pre-stretch in the mesh. Another important observation from the extensibility experiment is that the stretch behaviour of the SK mesh was not only related to the simple relation of fractional content and individual elastic modulus of HM and LM sections but also to their axial arrangement i.e. 1/1, 2/2, 3/3, and 4/4 (Fig. 3). This observation indicates the ease of biomechanical customisation with which a SK design can

131

be matched closely to an arterial site. The bending stiffness property is dependent on both extension (outer curvature) and compression (inner curvature) behaviour of the mesh. The extension property is dependent on LM segment while the compression of inner curvature is dependent on the interlocking phenomenon of the adjacent HM segment loops which leads to a sudden increase in bending resistance force. The onset of interlocking is determined by the axial spacing of adjacent HM segments. As the spacing increases (i.e. 0/01/1-2/2-3/3-4/4) the interlocking gets delayed till higher curvatures. This was experimentally observed as decrease in bending stiffness for 1/1 and 2/2 samples in comparison to the 0/0 sample (Fig. 4). The reason behind a sudden drop in stiffness response for 3/3 and 4/4 samples is that the spacing between adjacent HM segments was high enough to completely avoid their interlocking. The mechanical properties of a polymeric filament: elastic modulus, stress relaxation time and SR% determine its fatigue resistance (Zhang et al., 2000). Higher SR% with shorter relaxation times is considered helpful in increasing the fatigue resistance. The role of AVG mesh wall viscoelasticity may thus be important during the mechanical homoeostasis of vein while it undergoes the ‘arterialisation’ process. Over the holding period of 360 s, the 1/1 and 2/2 mesh samples showed more significant stress relaxation than 0/0, 4/4, and 3/3 samples (Fig. 6). This suggests that nature of stress relaxation behaviour is also dependent on arrangement of LM and HM segments along the mesh axis. The improved SR% in shorter LM segment mesh (1/1 and 2/2) compared to 3/3 and 4/4 meshes is likely due to the availability of a LM loop adjacent to every HM loop in the former. The close proximity of a LM loop assists in early relaxation of HM loops which otherwise tend to remain under high stress levels for longer. Also, these designs (1/1 and 2/2) distribute circumferential compliance more evenly along the mesh length compared to widely spaced segmented designs (3/3 and 4/4). This can be an important characteristic considering the role of mesh support in improving vein-graft hemodynamics while reducing the chances of damage induced by mechanical oscillations. The importance of initial relaxation rate becomes more evident as explained by the Law of LaPlace (T p P.r) i.e. an increase in radius (during systole) requires larger wall tension to maintain the internal pressure and that at constant radius (end of systole), pressure will fall (start of diastole) if the tension decreases over time. An early stress-relaxation can prevent prolonged stressing of mesh wall when a compliant mesh expands (increase in r) to accommodate extra systolic blood volume. The geometrical arrangement of LM and HM segments or segmentation frequency showed a significant effect on all the mechanical properties. The theoretical fraction (number of courses per unit length) of each component (HM and LM) was nearly similar in all the SK samples (Fig. 1 drawings). The actual prototype undergoes relaxation after being taken off from the machine and thus varies in LM component fractions (75 courses per 10 cm length). However, this variation was insignificant compared to total number of courses (80–82) in the test mesh length. Different SK designs can relax to different extents and may vary in their fraction percentage slightly. High segmentation frequency (1/142/243/344/4)

132

journal of the mechanical behavior of biomedical materials 48 (2015) 125 –133

19.5 mmHg

Circumference (mm)

19.0

120 100

18.5

80

18.0

60

17.5 17.0

40

Structural jamming

20

16.5

Appendix A.

Supplementary data

Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.jmbbm. 2015.04.001.

r e f e r e n c e s

16.0 0

15.5 15.0 0/0

4/4

3/3

2/2

1/1

Mesh ID

Fig. 9 – Curves comparing the rate of increase in circumference at each pressure level for different mesh types (0/0, 1/1, 2/2, 3/3, 4/4). Dotted arrows represent the pressure limit at which the radial compliance dropped by 20% from its highest value (i.e. at 20 mmHg). leads to improved circumferential extensibility (Figs. 5–7), but has an interestingly opposite effect on longitudinal extensibility (Figs. 3 and 4). The effect of SK mesh pattern on mechanical behaviour was also observed as difference in the rate of increase in circumference or decrease in compliance with each incremental pressure step (Fig. 9). The dotted arrows (Fig. 9) indicate the pressure level at which radial compliance dropped by 20% from its highest value at 20 mmHg. On the basis of overall test results, this study completes the initial design optimisation stage and proposes the 2/2 mesh type as a suitable SK design configuration as it demonstrates equally optimised longitudinal and circumferential extensibility. An in-vitro long-term fatigue analysis is the next stage to complete the suitability justification of SK design for vein-graft application.

5.

Conclusion

This study reported the effect of utilising a low modulus elastic structural component (polyurethane filament) in a relatively stiff knitted vein support mesh (nitinolþpolyester filament). The resulting design was a segmented knit structure with alternately arranged low and high modulus segments. The observed improvement in mesh mechanical properties was not only related to material properties of both segments but also to their spatial organisation i.e. their frequency of occurrence along the mesh axis. The use of low modulus segments significantly improved the bending flexibility and longitudinal extensibility of the mesh. The enhancement in circumferential extensibility property was demonstrated in radial compliance and deployment limit experiments. The practical significance of both these properties is high. A low deployment limit requires the use of a large diameter mesh, which leads to rejection of most of small diameter veins. On the contrary, high deployment limit and radial compliance forms an optimum combination for allowing the use of a wide range of vein diameters during implantation as well as assisting in synchronised vein wall distension and recoil after implantation.

Abbott, W.M., Megerman, J., Hasson, J.E., L’Italien, G., Warnock, D. F., 1987. Effect of compliance mismatch on vascular graft patency. J. Vasc. Surg. 5, 376–382. Ben-Gal, Y., Taggart, D.P., Williams, M.R., Orion, E., Uretzky, G., Shofti, R., Banai, S., Yosef, L., Bolotin, G., 2013. Expandable external support device to improve Saphenous vein graft patency after CABG. J. Cardiothorac. Surg. 8, 122–130. Davies, A.H., Magee, T.R., Baird, R.N., Sheffield, E., Horrocks, M., 1992. Vein compliance: a preoperative indicator of vein morphology and veins at risk of vascular graft stenosis. Br. J. Surg. 79, 1019–1021. Davies, M.G., Hagen, P.O., 1995. Pathophysiology of vein graft failure: a review. Eur. J. Vasc. Endovasc. Surg. 9, 7–18. Desai, M., Mirzay-Razzaz, J., Von Delft, D., Sarkar, S., Hamilton, G., Seifalian, A.M., 2010. Inhibition of neointimal formation and hyperplasia in vein grafts by external stent/sheath. Vasc. Med. 15, 287–297. Dobrin, P.B., Littooy, F.N., Endean, E.D., 1989. Mechanical factors predisposing to intimal hyperplasia and medial thickening in autogenous vein grafts. Surgery 105, 393–400. El-Kurdi, M.S., Hong, Y., Stankus, J.J., Soletti, L., Wagner, W.R., Vorp, D.A., 2008. Transient elastic support for vein grafts using a constricting microfibrillar polymer wrap. Biomaterials 29, 3213–3220. Emery, R.W., Solien, E., Jamieson, S.W., 2012. Implantation of the eSVS mesh. Innov.: Technol. Tech. Cardiothorac. Vasc. Surg. 7, 65–67. Genoni, M., Odavic, D., Loblein, H., Dzemali, O., 2013. Use of the eSVS mesh: external vein support does not negatively impact early graft patency. Innov.: Technol. Tech. Cardiothorac. Vasc. Surg. 8, 211–214. Harskamp, R.E., Lopes, R.D., Baisden, C.E., De Winter, R.J., Alexander, J.H., 2013. Saphenous vein graft failure after coronary artery bypass surgery: pathophysiology, management, and future directions. Ann. Surg. 257, 824–833. Hasegawa, M., Azuma, T., 1979. Mechanical properties of synthetic arterial grafts. J. Biomech. 12, 509–517. Jeremy, J.Y., Bulbulia, R., Johnson, J.L., Gadsdon, P., Vijayan, V., Shukla, N., Smith, F.C.T., Angelini, G.D., 2004. A bioabsorbable (polyglactin), nonrestrictive, external sheath inhibits porcine saphenous vein graft thickening. J. Thorac. Cardiovasc. Surg. 127, 1766–1772. John, L.C.H., 2009. Biomechanics of coronary artery and bypass graft disease: potential new approaches. Ann. Thorac. Surg. 87, 331–338. Kidson, I.G., 1983. The effect of wall mechanical properties on patency of arterial grafts. Ann. R. Coll. Surg. Engl. 65, 24–29. Kohler, T.R., Kirkman, T.R., Clowes, A.W., 1989. The effect of rigid external support on vein graft adaptation to the arterial circulation. J. Vasc. Surg. 9, 277–285. Krejca, M., Skarysz, J., Szmagala, P., Plewka, D., Nowaczyk, G., Plewka, A., Bochenek, A., 2002. A new outside stent – does it prevent vein graft intimal proliferation?. Eur. J. Cardiothorac. Surg. 22, 898–903. Liu, S.Q., Fung, Y.C., 1998. Changes in the organization of the smooth muscle cells in rat vein grafts. Ann. Biomed. Eng. 26, 86–95.

journal of the mechanical behavior of biomedical materials 48 (2015) 125 –133

Longchamp, A., Alonso, F., Dubuis, C., Allagnat, F., Berard, X., Meda, P., Saucy, F., Corpataux, J.M., Se´bastien, D., Haefliger, J. A., 2014. The use of external mesh reinforcement to reduce intimal hyperplasia and preserve the structure of human saphenous veins. Biomaterials 35, 2588–2599. Members, W.C., Hillis, L.D., Smith, P.K., Anderson, J.L., Bittl, J.A., Bridges, C.R., Byrne, J.G., Cigarroa, J.E., DiSesa, V.J., Hiratzka, L. F., Hutter, A.M., Jessen, M.E., Keeley, E.C., Lahey, S.J., Lange, R. A., London, M.J., Mack, M.J., Patel, M.R., Puskas, J.D., Sabik, J.F., Selnes, O., Shahian, D.M., Trost, J.C., Winniford, M.D., 2011. 2011 ACCF/AHA guideline for coronary artery bypass graft surgery: a report of the american college of cardiology foundation/American heart association task force on practice guidelines. Circulation 124, e652–e735. Moodley, L., Franz, T., Human, P., Wolf, M.F., Bezuidenhout, D., Scherman, J., Zilla, P., 2013. Protective constriction of coronary vein grafts with knitted nitinol. Eur. J. Cardiothorac. Surg. 44, 64–71. Murphy, G.J., Newby, A.C., Jeremy, J.Y., Baumbach, A., Angelini, G. D., 2007. A randomized trial of an external Dacron sheath for the prevention of vein graft disease: the extent study. J. Thorac. Cardiovasc. Surg. 134, 504–505. Ozturk, N., Sucu, N., Comelekoglu, U., Yilmaz, B., Aytacoglu, B., Vezir, O., 2013. Pressure applied during surgery alters the biomechanical properties of human saphenous vein graft. Heart Vessels 28, 237–245. Parsonnet, V., Lari, A., Shah, I.H., 1963. New stent for support of veins in arterial grafts. Arch. Surg. 87, 696–702. Powell, J.T., Gosling, M., 1998. Molecular and cellular changes in vein grafts: influence of pulsatile stretch. Curr. Opin. Cardiol. 13, 453–458. Schoettler, J., Jussli-Melchers, J., Grothusen, C., Stracke, L., Schoeneich, F., Stohn, S., Hoffmann, G., Cremer, J., 2011. Highly flexible nitinol mesh to encase aortocoronary saphenous vein grafts: first clinical experiences and angiographic results nine months postoperatively. Interact. Cardiovasc. Thorac. Surg. 13, 396–400. Singh, C., Wang, X., 2014. A biomimetic approach for designing stent-graft structures: caterpillar cuticle as design model. J. Mech. Behav. Biomed. Mater. 30, 16–29. Stooker, W., Go¨k, M., Sipkema, P., Niessen, H.W.M., Baidoshvili, A., Westerhof, N., Jansen, E.K., Wildevuur, C.R.H., Eijsman, L., 2003. Pressure-diameter relationship in the human greater saphenous vein. Ann. Thorac. Surg. 76, 1533–1538. Szilagyi, D.E., Elliott, J.P., Hageman, J.H., Smith, R.F., Dall’olmo, C. A., 1973. Biologic fate of autogenous vein implants as arterial

133

substitutes: clinical, angiographic and histopathologic observations in femoro-popliteal operations for atherosclerosis. Ann. Surg. 178, 232–246. Tiwari, A., Cheng, K.S., Salacinski, H., Hamilton, G., Seifalian, A. M., 2003. Improving the patency of vascular bypass grafts: the role of suture materials and surgical techniques on reducing anastomotic compliance mismatch. Eur. J. Vasc. Endovasc. Surg. 25, 287–295. Trubel, W., Moritz, A., Schima, H., Raderer, F., Scherer, R., Ullrich, R., Losert, U., Polterauer, P., 1994. Compliance and formation of distal anastomotic intimal hyperplasia in Dacron mesh tube constricted veins used as arterial bypass grafts. ASAIO J. 40, M273–M278. van der Merwe, H., Daya Reddy, B., Zilla, P., Bezuidenhout, D., Franz, T., 2008. A computational study of knitted Nitinol meshes for their prospective use as external vein reinforcement. J. Biomech. 41, 1302–1309. Vijayan, V., Shukla, N., Johnson, J.L., Gadsdon, P., Angelini, G.D., Smith, F.C.T., Baird, R., Jeremy, J.Y., 2004. Long-term reduction of medial and intimal thickening in porcine saphenous vein grafts with a polyglactin biodegradable external sheath. J. Vasc. Surg. 40, 1011–1019. Vijayan, V., Smith, F.C.T., Angelini, G.D., Bulbulia, R.A., Jeremy, J.Y., 2002. External supports and the prevention of neointima formation in vein grafts. Eur. J. Vasc. Endovasc. Surg. 24, 13–22. Weston, M.W., Rhee, K., Tarbell, J.M., 1996. Compliance and diameter mismatch affect the wall shear rate distribution near an end-to-end anastomosis. J. Biomech. 29, 187–198. Zhang, X., Li, Y., Yeung, K.W., Yao, M., 2000. Viscoelastic behavior of fibers during woven fabric bagging. Text. Res. J. 70, 751–757. Zilla, P., Human, P., Wolf, M., Lichtenberg, W., Rafiee, N., Bezuidenhout, D., Samodien, N., Schmidt, C., Franz, T., 2008. Constrictive external nitinol meshes inhibit vein graft intimal hyperplasia in nonhuman primates. J. Thorac. Cardiovasc. Surg. 136, 717–725. Zilla, P., Moodley, L., Wolf, M.F., Bezuidenhout, D., Sirry, M.S., Rafiee, N., Lichtenberg, W., Black, M., Franz, T., 2011. Knitted nitinol represents a new generation of constrictive external vein graft meshes. J. Vasc. Surg. 54, 1439–1450. Zilla, P., Wolf, M., Rafiee, N., Moodley, L., Bezuidenhout, D., Black, M., Human, P., Franz, T., 2009. Utilization of shape memory in external vein-graft meshes allows extreme diameter constriction for suppressing intimal hyperplasia: a nonhuman primate study. J. Vasc. Surg. 49, 1532–1542. Zou, Y., Zhang, Y., 2011. The orthotropic viscoelastic behavior of aortic elastin. Biomech. Model. Mechanobiol. 10, 613–625.