GB2430373A - Graft - Google Patents

Abstract

A graft comprises flow tubing defining a flow lumen, wherein the graft is pre-shaped such that the centre line of the flow lumen is curved in a single plane and has a change in the direction of curvature.

Description

-- 2430373 88942.628 The invention relates to grafts.

Vascular and cardiovascular disease, predominantly due to atherosclerosis, exert a high cost on the world population in terms of morbidity and mortality. While many risk factors have been identified, including smoking, obesity and diets high in saturated fats, these do not explain the focal nature or propensity for certain locations. For example, the left hand side of the aortic bifurcation experiences a higher incidence and severity of atheroma than the right, despite their similar circumstances and close proximity to each other.

Haemodynamic factors have long been under investigation with respect to their role in the pathogenesis of atherosclerotic disease. Principally, this work has focused on the relationship between atheroma and vascular wall shear stress. This can be defined as the product of the fluid viscosity and the velocity gradient at the vessel wall, termed the wall shear rate (Caro et al, 1978, "The Mechanics of the Circulation", Oxford Medical Publications).

Originally, atherosclerosis was assumed to be the result of flow-induced damage to the endothelium, the so-called high shear theory. Studies revealed denudation of the endothelium at very high shear stresses, presumed at the time to be physiological. This was supported by the frequent presence of atheroma at points of arterial bifurcation, and the high shear associated with these locations.

This view has since been superseded by the low shear theory, which conversely states that atherosclerotic plaques are correlated with areas of low shear stress. Supporting evidence for this theory was first reported in 1969 by Caro et al (Nature 223, 1159-1161 "Arterial wall shear and distribution of early atheroma in man"), having observed that the atheroma seen at bifurcations is principally found not on the flow divider, where areas of high shear are located, but on the opposite wall where flow separation results in low shear stresses.

Shear stress has been shown to effect the expression of more than two hundred endothelial cell genes. Many are involved directly or indirectly in processes involved in the pathogenesis of atherosclerosis. For example, low shear has been shown to upregulate the expression of Toll-like receptor 4 (TLR4), increasing NP- KB activity and monocyte adhesion force.

Transduction mechanisms are unknown, but thought to include deformable membrane proteins linked to the internal cell cytoskeleton. The model which is currently proposed involves mechanotranscjucers such as integrins and ion channels, intermediate signalling molecules including ras and protein kinase C. Mitogen activated protein kinases such as ERK 1/2 are also involved, with nitric oxide as the ultimate effector molecule. Similar concepts can be found in the current understanding of the pathogenesis of osteoarthritis. In this instance, chondrocyte function is regulated through the integrin-mediated response to mechanical stimulation and the subsequent activation of intracellular, autocrine and paracrine pathways. Undesirable mechanical circumstances or dysfunction of the processes involved can result in the structural and functional breakdown of the cartilage and the onset of osteoarthrjtjs.

It is now widely accepted that atheromatous lesions develop preferentially at locations of low average wall shear stress. As a result, the emphasis of the highly regarded response to injury' hypothesis of atherosclerosis has shifted from one of endothelial denudation to that of endothelial dysfunction.

Subsequently, attention has turned to the extent to which vascular geometry dictates the haemodynamic conditions within. A key related concept is planarity, a property which is present if all points in a curve or bifurcation lie on a single plane Thus, a simple horseshoe type bend is planar. Non-planar configurations have been shown to produce swirling flow (Doorly et al, 1997, "Helix and model graft flows: MM Measurement and CFD Simulations" ASME FED SM 97 (Bio-medjcal Fluids Engineering II, 1-8) and as a result fewer extremes of shear-stress The reduction in areas of stagnation prevents adverse interactions with the arterial wall, and increased mixing homogenizes concentrations of pro- atherogenic substances and metabolites.

A significant oxygen gradient commonly exists, with concentrations at the arterial wall often half that in the centre of the lumen. Increased mixing will therefore bring higher oxygen concentrations closer to the endothelium, reducing the potential for initiation of atherosclerosis through anoxia.

Non-planar geometry is seen throughout the vasculature, including the aortic arch. Taking the example of the aortic bifurcation, not only is it usually non -planar, but it is also frequently asymmetric, accounting for the asymmetric incidence of atheroma. The degree of non-planarity varies both between individuals and with skeletal and cardiac motion.

In WO 95/09585 it was proposed to provide a vascular prosthesis having a flow lumen with a curvature extending in three dimensions, i.e. a nonplanar curvature, so as to induce swirl flow in blood flowing through the prosthesis.

We have now recognised that a graft pre-shaped with a curve in a single plane, i.e. a planar curve, can become a non-planar graft when inserted in a patient.

This can be achieved by placing the graft so that it bends out of the plane in which it already has a pre-shaped curvature, for example by being placed on a curved anatomical part such as the surface of the heart.

Viewed from a first aspect, therefore, the invention provides a method of inserting a graft in a human or animal body, comprising using a graft comprising flow tubing defining a flow lumen, the graft being pre-shaped such that the centre line of the flow lumen is curved in a single plane, and placing the graft in the body so that it bends out of the plane of its initial pre-shaped curvature.

With such a method, the benefits of swirl flow can be obtained.

The swirling effect can be further improved if the graft is pre-shaped such that the centre line of the flow lumen which curves in a single plane has a change in the direction of curvature Thus, for example, as viewed along the direction of flow, the direction of curvature may change from curving left to curving right, i.e. there is a point or region of inflexion.

Viewed from a second aspect the invention provides a graft comprising flow tubing defining a flow lumen, wherein the graft is pre-shaped such that the centre line of the flow lumen is curved in a single plane and has a change in the direction of curvature The graft may be bent out of the plane when inserted in a patient, so creating a non-planar flow lumen. This, combined with the change of direction of the initial planar curvature, assists the generation of swirl flow.

Whilst a single change in the direction of curvature is beneficial, preferably the curved, planar centre line of the flow lumen has a plurality of changes in the direction of curvature. A tortuous graft, having more than one change in the direction of curvature, may thus be provided. There may be at least two, preferably three, more preferably four or five, changes of curvature.

In certain preferred embodiments the curvature of the centre line of the flow lumen is substantially sinusoidal.

The flow lumen of the graft is preferably substantially free of ribs or grooves. The cross-section of the flow lumen is preferably circular.

EXAMPLES

Recently, the effect of geometry in the coronary arteries has become a focus of research. Some individuals have more tortuous right coronary arteries, described as "Sigma" shaped, due to their appearance on angiogram as detailed in Figure 1.

Other individuals have less tortuous "C" shaped configurations, also shown in Figure 1.

The geometrical configuration of a coronary artery can be thought of as a two-dimensional pattern, whether tortuous or simple, superimposed upon a three- dimensionally curved surface. The comparison is therefore between one highly or moderately planar large curve in the case of a "C" shape, or the series of non-planar bends of a Sigma shaped right coronary artery (RCA).

Experiments were carried out to assess the protection against atherosclerosis potentially conferred by a Sigma' shaped right coronary artery and investigate the factors contributing to its reduced planarity when compared with a less tortuous "C" shaped RCA.

Methods The purpose of the laboratory methods used in this study was to evaluate the role of individual geometrical factors in the coronary circulation by separately controlling flow rate, tortuosity and curvature in the third d imension Three RCA imitations were formed from PVC tubing with exterior and lumenal diameters of 12 and 8mm respectively, set into shape through baking at temperatures below 100 C. In order to preserve a smooth and uniform surface and cross section during this process, silicone piping and incompressible metal springs were used to protect the lumen. All three models share the same amplitude of 3cm and a 30cm direct distance between the start and finish of the curved section. While all sinusoidal, they differ in their wavelengths of 60cm, a shallow curve akin to a C- shape, a far more tortuous 10cm simulating the Sigma shape, and an intermediate 20cm.

The shape and dimensions of the imitations were chosen to approximate the shape of a selection of actual coronary angiograms, but were also constrained by the limits to which the tubing could be manipulated. No consensus on a standard measure of tortuosity exists, and the sigma shaped phantom, as a series of curves with no distinct separation, produces a complex pattern of planarity. Therefore, for these reasons and due to the qualitative nature of this study, calculation of values for planarity or tortuosity was deemed to be unnecessary.

Investigation of the effect of curvature in the third dimension is achieved through the use of thin and flexible polycarbonate sheeting, flexed into half cylinders with radii of curvature of 5 and 10cm These are shown in Figure 2. As the lumenal diameter of the tubing used is in the order of twice that of a typical right coronary artery, these correlate with heart diameters of 5 and 10cm However, in vivo the coronary artery is largely embedded in the heart surface, whereas the phantom is laid upon the curved surface, resulting in a larger effective cardiac size.

For comparison, the models will be assessed while attached to both a flat surface and the above cylinders.

In order to correctly simulate the conditions of the right coronary artery, the flow within the phantom must correspond to in vivo levels A key measure of this is the Reynolds number (Re), a dimensionless figure representing the ratio between inertial and viscous forces within the fluid. A typical reported flow rate for the RCA is 1.5 ml/s, rising to 4 mI/s at the peak of the cycle This latter figure equates to a Reynolds number in the order of 500 Thus, Reynolds numbers of 100, 200, 300, 400 and 600 were chosen to represent the variety between individuals, activity level and phase of the cardiac cycle. To achieve these, the model is incorporated into a circuit fed by gravity from a raised water tank, filled to the same level for each cycle in order to maintain inlet pressure, with flow regulated through variation of resistance distal to the section being assessed. As identical tubing is used throughout, equivalent flow rates in mi/s for all configurations can be found in Table

Table 1

Reynolds Number Flow rate (ml/s) 100 0.625 1.25 300 1.875 400 25 600 3.75 While the coronary circulation in vivo is pulsatile, only steady flow is used.

However, recent research involving the application of computational fluid dynamics to magnetic resonance imaging acquired infrainguinal bypass graft reconstructions indicates that there is little difference in time averaged wall shear stress distribution between pulsatile and steady simulations. Therefore this simplification is considered acceptable and, in addition, vastly improves the opportunity for visual analysis of flow patterns.

Another possible discrepancy arises from the non-newtonian nature of the blood itself. The viscosity of the fluid is dependent upon a range of variables, primarily the vessel size due to the effect of individual red corpuscles. As the diameter of the right coronary artery is sufficiently large compared with the size of a red cell, it is not unreasonable to assume Newtonian behaviour in this circumstance Assessment of the flow within the models is aided by the injection of ink into the stream, 1/6t1 of the distance along the curved section of the tube, where the flow specific to that curve is likely to be established. This takes two forms, firstly the introduction of a stream of ink to enable visual evaluation of the presence and degree of swirling flow patterns. Differences are recorded by both video and still - frame digital photography. This is accompanied by the injection of a 0.5 ml bolus of ink, mixed with a small amount of ethanol and titrated against a beaker of water to assure equal density in order to reduce gravitational settling of the dye. The rate of clearance of the ink within the tube is documented with the use of a light emitting diode and an opposing photodiode, whose voltage is recorded by an analogue to digital converter (Pico ADC 11) and a personal computer running data- logging software (Picolog). Small variations in voltage at zero concentration are corrected for by converting the raw data into a percentage deviation from baseline deviation.

Calibration data providing sample voltages for known concentrations of ink can be

found in Table 2.

Table 2

Ink Concentration (%) Voltage (V) 0 2.1 0.05 1.5 0.1 1.37 0.2 0.93 0.5 051 1 0.26 5 024 0.24 0.24 While the relationship between ink concentration and voltage is not linear, this system does allow for direct comparison between the different geometrical configurations used. For the purposes of this comparison, the times in seconds for the percentage voltage decay between one hundred, fifty, twenty, five and one percent are calculated for every instance. These particular intervals were chosen for their similar times.

Results Visual Assessment A clear progression of flow patterns is evident within the different phantoms.

This ranges from there being negligible swirling flow visible in the planar C shaped phantom and a distinct area of high residence times adjacent to the inside wall of the curve, to a pattern of high swirl and low clearance times in the sigma tube when laid upon the highly curved cylinder. The photographs displayed in Figures 3 to 7 are intended to illustrate and support the above observations. Both bolus clearance and ink trail injections are included to convey an idea of the range of flow and clearance patterns produced throughout this study.

In the planar "C' shape case of Figure 3, no significant swirl can be seen, and an area of sluggish flow and long residence times is evident on the inner wall of the curvature.

Even when subjected to significant curvature (5 cm radius of curvature) in the third dimension (Figure 4), swirl is minimal in the "C" shaped phantom.

Figure 5 shows a planar sigma shape Despite the tortuous phantom, the planar geometry in this example retains the poor clearance of a bolus of the ink dye on the inner wall of each curve. An ink trail injected into the same configuration above shows no significant visible swirl, as shown in Figure 6.

Figure 7 shows a sigma shape on a 5 cm radius of curvature. The highly non-planar combination of the most tortuous phantom and a 5cm radius of curvature in the third dimension has produced a markedly swirling flow. Unlike the planar situation in Figure 5, little ink remains adjacent to the inner wall immediately distal to the bend and no particularly stagnant areas are visible Clearance Data Collection of sequential voltage recordings from the photodiode apparatus produced a voltage deflection as the ink passes the sensor, which decays with a similar shape for all configurations. An example is shown in Figure 8 Figure 9 shows the conversion of the raw data into a percentage value of the maximal voltage deviation This conversion both accounts for variation in baseline voltage between repetitions and reveals a more familiar decay curve.

Figure 10 displays the decay curves for a non-planar situation at all Reynolds numbers used. It clearly shows that the equipment is sensitive enough to discriminate between the flow rates used and that for this construction a similar shape is retained throughout.

Figure 11 displays a similar pattern to Figure 10 above, despite the simple and planar geometry, with distinct separation of the curves representing Reynolds numbers 300, 400 and 600. The anomalous decay of the example at Reynolds 200 is addressed in the "Discussion" text below.

Figure 12 compares the effect of the three curvatures used on the least tortuous phantom. A large difference is visible between the completely planar situation and the two non-planar instances, however little is seen between these two.

This is despite the data produced suggesting a greater distinction than in the two more tortuous cases, highlighting the inability of this method of data presentation to discriminate on a smaller scale, particularly in the tail of the curve. Analysis of this aspect of the data is therefore more appropriately achieved through the use of fixed decay interval times.

What a decay curve can display, however, is the variety of clearance patterns produced by the range of planarity that can reasonably be expected in vivo. Figure 13 shows a highly non-planar sigma phantom contrasted with the C shaped construct and less 3D curvature. This shows a distinct difference, suggesting the possibility of a clinically significant range of planarities in the general population The clearest discrepancy between the two lies in the tail portion of the graph.

Fixed Decay Interval Time Data The data in Table 3 details the time required for the predefined voltage percentage decays in all of the phantoms, curvatures and flow rates used. It is particularly important in discriminating differences occurring at low ink dye concentrations

Table 3

Fixed Decay Interval Time Data C-Shaped iianar 100--So 50--20 20--5 5--i i0() - - - - 17 50 90 58 300 21 62 63 40 400 16 48 50 36 600 10 18 43 37 1icm 100--SO 50--20 20--5 5--i

- - - -

5 26 78 79 300 11 16 43 46 400 12 9 33 43 600 7 7 20 28 5cm l00--50 50--20 20--S 5--i

- - - -

4 24 39 43 300 15 11 24 26 400 14 10 25 30 600 8 6 12 17 Intermediate Planar ioo--so 50--20 20--S 5--i 59 101 105 82 33 31 34 36 300 21 20 27 33 400 12 14 24 29 I 600 6 14 17 19 10cm 100--SO 50--20 20--S 5--I 35 50 62 53 18 21 33 28 300 13 11 22 24 400 11 9 19 16 600 6 6 11 12 5cm 100--50 50--20 20--5 s--i7 50 37 45 52 28 20 24 31 300 15 14 22 39 400 11 11 19 24 600 7 7 13 19 Sigma Shaped iinar 100-50 50-20 20--5 5--i 104 100 113 112 21 18 32 32 300 17 15 30 35 400 13 13 19 20 600 9 9 14 17 10cm i00--50 50--20 20--S 5--i 52 45 49 51 23 17 20 22 300 17 10 16 19 400 13 8 12 14 600 9 6 9 12 JScm iOO--50 50--20 20--5 5--i 55 38 36 37 27 24 26 26 300 16 18 18 17 400 9 9 11 13 600 7 11 11 11 For example, the ratio between the 5cm and 10cm curvature cases in Figure 12 in the third (20% to 5%) and fourth (5% to 1%) intervals are 25:33 and 30:43 respectively at a Reynolds number of 400. In the more tortuous examples, however, there is no apparent decrease in clearance times when radius of curvature is reduced from l0cmto5cm.

The flow rate also has a clear influence on decay times, as seen in Figure 10 Close scrutiny of the interval periods suggests a greater proportional decrease in the less tortuous imitations. For example, at a 10cm curvature, the percentage drop in time from Re=200 to Re=600 is 74% for the C shape, compared with only 55% for - 12 - the most tortuous. A similar picture is seen at a curvature of 5cm, where the corresponding figures are 69% and 58%.

Tortuosity also provides a uniform decrease in clearance times, with almost every reading from the sigma phantom being lower than the intermediate equivalent, which is in turn lower than the simple curve.

A final comparison between the extremes of planarity used (Sigma, 5cm and C, planar) shows a stark increase in clearance times, particularly in the second, third and fourth intervals, where the ratios are 48:9, 50:11 and 36:13. Differing by greater than a factor of five in some cases, these figures indicate the magnitude of the influence of coronary geometry on the flow within.

Discussion Before the experimental results are discussed in detail, it is appropriate to acknowledge the difficulties encountered and thus where potential for error and inaccuracy lies Firstly, despite the precautions taken, settling of ink dye did occur, most markedly in highly planar arrangements at low flow rates. In some situations, this was to such an extent that the majority passed below the line of sight between LED and photodiode. For this reason, clearance data with an insufficiently high voltage deflection as to make the correction for voltage fluctuations misleading was excluded, including all three C shaped configurations at Re= 100. All non-speed dependent assessment was therefore conducted at a Reynolds number of four hundred in order to minimize the effect of gravitational settling. The settling seen, while a technical challenge, does provide useful information in itself, for it corresponds to a lack of swirl and mixing. It is important to note that this will reduce the voltage decay times as a result. This is epitomised by the anomalous results seen in the planar C configuration at Re=200, as seen in Figure 11. Thus, any distinctions seen in these readings are likely to be less than the true values, instead of a misleading exaggeration of small differences or noise.

Despite the crude nature of the experimental setup and the scatter evident in the clearance times, clear and large distinctions are present, from which some

significant conclusions can be made.

- 13 - Visual evaluation of the flow patterns shows a gradation of swirling flow which is most marked in the tortuous configurations with a high curvature in the third dimension. The minimal swirl present in the planar cases can be explained as the result of the aforementioned gravitational settling and minor inconsistencies of geometry. Technically, by definition no swirl exists in a truly planar bend, instead two Dean vortices are present, symmetrical around the plane of curvature. It is only when the geometry moves out of plane that one vortex becomes larger than the other, thus creating swirl.

Turning to the clearance data, two distinct phases can be identified. Firstly, there is a peak as the majority of the ink dye passes the photodiode. Decay times depend mainly on Reynolds number, and relatively little with shape or curvature.

This is likely to represent the blood in the centre of the stream that does not interact with the vessel wall.

The second phase produces a more linear voltage decay, particularly evident in highly planar configurations at low flow rates such as the planar decay in figure 12. This does correspond to the clearance of ink from the wall and indicates the extent of regions of low flow and shear stress. The data shows differences, which are stark between the extremes of planarity, confirming geometry as a factor in the creation of the haemodynamic conditions within.

The effect of tortuosity is clear and uniform throughout the clearance data Regardless of flow rate or 3D curvature, the intermediate phantom performs better than the C shape in otherwise identical circumstances. The sigma imitation in turn has lower clearance times than the intermediate, emphasising the haemodynamic benefits of highly tortuous vascular geometry in both the coronary and peripheral circuits.

In addition to the shape of the artery itself, the radius of curvature in the third dimension is also an important factor. This is clearest in the case of the C shape, where a definite improvement in clearance times is seen with regard to all three variations. However, while the intermediate and sigma phantoms display benefit from the curvature, little difference is seen between a 5 or 10 centimetre radius.

Experimental error may account for some of this discrepancy, but it does point to a "planarity threshold", beyond which little haemodynamic improvement is seen.

- 14 - In vivo, curvature in the third dimension is determined by the ratio of the coronary artery and heart diameters. Thus, patients suffering from conditions that increase heart diameter, such as dilated cardiomyopathy, may be subject to less desirable coronary artery geometry. This is likely to be improved by the Dor procedure or the S more controversial Batista procedure, where heart size is reduced by the removal of aneurismal left ventricular wall or part of the homogenously diseased muscle respectively (Vitali et a!, 2003, Am J Cardiol. 91 (9A), 88F -94F "Surgical therapy in advanced heart failure").

Unsurprisingly, increased flow rates produce reduced clearance times in all situations, supporting the long established benefits of cardiovascular exercise.

However, the benefit of increased coronary blood flow does not appear to be uniform, with a greater proportional advantage in the planar configurations. It is possible that the higher velocity in the C shaped phantom emphasises what little swirl is present, creating flow patterns more akin to those seen in the more tortuous examples at all speeds.

There is clearly evidence to suggest that non-planar arterial geometry and the resulting flow patterns may be beneficial to the coronary circulation and in the vasculature as a whole. Such a conclusion is somewhat counter-intuitive, being that the longer and more tortuous RCA with a convoluted path is actually the preferable configuration The key to maximizing the utility of this information lies in its application to surgical procedures. Involving such considerations in both pre- operative planning and surgical device design has the potential to minimize morbidity, mortality and re-intervention rates.

A prime example of how surgical technique can determine vascular planarity can be found in the infrainguinal arterial bypass graft. Such grafts often occlude prematurely due to intimal hyperplasia, which primarily arises in areas of low shear.

This is the principal cause for the poor mid-term patency rates seen in patients undergoing the procedure, with as many as a quarter of all grafts occluding within one year of surgery, a proportion which can rise to a half at ten years. The bypass conduit can either be routed superficially or tunnelled deep, hence reducing the graft proximal angle (GPA) at the distal anastomosis. Recent research involving both in vivo data and computational fluid dynamics has shown there to be a significant - 15 difference between these methods, with a larger GPA being associated with a less planar anastomosis.

The concept of curvature in the third dimension also has potential relevance in peripheral applications. Whereas the straight edge and flat plane is the mainstay of human engineering, in vivo the curved surface is the norm rather than an exception. Thus, the use of such anatomical portions in the location of bypass conduits may "rescue" an otherwise highly planar configuration The conduit itself has a role to play in the creation of desirable haemodynamic flow patterns. Helical grafts, as disclosed in WO 2004/082534, whose helical shape promotes swirling flow, are currently under trial and have been shown to improve mixing and vastly reduce the extent of intimal hyperplasia.

Applications include the infrainguinal bypass graft and arteriovenous shunts, used in renal dialysis access. Providing such flow is likely to be particularly beneficial to the postoperative patient and others who are unable to take advantage of the benefits of cardiovascular exercise. The experimental results support the hypothesis that the possession of a

tortuous RCA confers protection against atheromatous plaque. The concepts discussed are also applicable to the vasculature as a whole. Therefore, planarity should be considered in future cardiovascular assessment and surgical intervention

Claims (7)

1. - 16 - Claims 1. A graft comprising flow tubing defining a flow lumen,

   wherein the graft is pre-shaped such that the centre line of the flow lumen is curved in a single plane and has a change in the direction of curvature
2. 2. A graft as claimed in claim I, wherein the curved, planar centre line of the flow lumen has a plurality of changes in the direction of curvature.
3. 3. A graft as claimed in claim I or 2, wherein the curvature of the centre line of the flow lumen is substantially sinusoidal.
4. 4. A method of inserting a graft in a human or animal body, comprising using a graft comprising flow tubing defining a flow lumen, the graft being pre -shaped such that the centre line of the flow lumen is curved in a single plane, and placing the graft in the body so that it bends out of the plane of its initial pre -shaped curvature.
5. 5. A method as claimed in claim 4, wherein the graft is pre-shaped such that the centre line of the flow lumen which curves in a single plane has a change in the direction of curvature.
6. 6. A method as claimed in claim 5, wherein the curved, planar centre line of the flow lumen has a plurality of changes in the direction of curvature.
7. 7. A method as claimed in claim 5 or 6, wherein the curvature of the centre line of the flow lumen is substantially sinusoidal.