If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To examine the longitudinal migratory force required to cause disconnection of the bifurcated distal body component from the tubular proximal body of a fenestrated stent-graft.
Methods
Using a previously reported mathematical model distal distraction forces were calculated prior to performing in vitro pullout testing. The top end of the proximal body and the iliac limbs of the distal body were attached to the grips of a tensile tester via plastic sealing plugs and pneumatic clamps. Channels within the plugs allowed pressurisation of the inside of the stent-graft. Pullout tests were conducted in the vertical plane. Force and displacement data were recorded and tests repeated 8 times at room temperature with the stent-grafts either dry or wet and unpressurized, at 100 mmHg or at 120 mmHg.
Results
The median maximum pullout force was 2.9 N (2.6–4.1) when dry, 3.9 N (3.5–5.4) when wet and unpressurized, 6.3 N (4.8–8.3) when wet and pressurized at 100 mmHg and 6.5 N (4.8–7.2) when wet and pressurized at 120 mmHg. There was a significant difference between pressurized and unpressurized conditions (P < 0.01).
Conclusions
The force required to distract the distal bifurcated component of a fenestrated stent graft is much lower than the reported proximal fixation strength of both a standard and fenestrated Zenith stent graft. Although this helps protect the fenestrated proximal body from the effects of longitudinal migration forces in vivo the current strength of the body overlap zone may actually be unnecessarily weak and requires careful surveillance in follow up.
The modular design of fenestrated stent-grafts aims to protect the stented fenestrations in the proximal body from the full effects of the longitudinal migration forces. The current overlap zone is investigated and found to be potentially too weak to prevent modular distraction of components in the long term. This study highlights the need for careful surveillance of this zone in follow up.
Introduction
The fenestrations of a fenestrated stent-graft are designed and deployed to be precisely matched up with the ostia of their target vessels.
Due to the size of the target vessels, longitudinal and rotational migration of as little as 2–3 mm can cause significant compromise, risking target vessel loss. In the early Australian experience, several unstented renal fenestrations led to target vessel loss resulting in the recommendation that all renal fenestrations should be stented.
In addition, the modular design of the fenestrated stent-graft body, with overlap, aims to place the drag component of the longitudinal migration forces on the distal bifurcated body, thereby reducing the potential for the fenestrations in the proximal tubular body to migrate.
Modular components risk separation leading to type III endoleaks and repressurisation of the aneurysm. There has been one reported death from a ruptured aneurysm secondary to complete component separation in a patient previously treated with a fenestrated stent-graft.
Resch et al have reported a patient presenting to another hospital at 27 months after implantation with complete separation resulting in rupture of the aneurysm and the need for urgent open conversion.
This study investigates the fixation between the modular components of a fenestrated stent-graft's proximal and distal bodies. These modular components rely upon the radial force of stents to provide secure fixation to one another.
Methods
Mathematical model
Mohan et al reported a simple model of the haemodynamic forces acting on a bifurcated stent-graft.
This was developed based on an idealized bifurcated device in which the bifurcation is planar and symmetrical, and the blood flow is distributed equally through the iliac limbs (Fig. 1). Using this mathematical model it was possible to calculate the longitudinal distraction force on the stent-grafts used in this study and how this varies with stent-graft geometry and dimensions, blood pressure and blood flow.
(1)
Figure 1Mathematical model for studying the longitudinal displacement forces in a bifurcated stent-graft. The idealized bifurcated endograft used in this model is planar and symmetrical; blood flow is distributed equally through the iliac limbs. The momentum equation is used to calculate the forces on the device. U1 = proximal velocity of blood; U2 = distal velocity of blood; A1 = proximal cross-sectional area; A2 = distal cross-sectional area; P1 = proximal blood pressure; P2 = distal blood pressure; fx = longitudinal displacement force; fy = transverse force; θ = iliac bifurcation half angle; Q = volume flow rate (A1U1).
The longitudinal displacement forces were calculated for a range of stent-graft diameters and blood pressures (Fig. 2). The flow was fixed at 1.5 L/min.
Figure 2Calculated distraction force on the distal bifurcated body of a fenestrated stent-graft with an inlet diameter (D1) of 22 mm or 24 mm. Outlet diameters of both iliac limbs fixed at 12 mm. The effect of a change in bifurcation angle (θ°) and increasing blood pressure is shown.
The inlet diameter of the proximal body can range from 24 to 30 mm and in the current design, the outlet diameter is fixed at 22 mm. For the distal bifurcated body the dimensions for the inlet were 22 mm (current design). The iliac limbs (outlet diameter) were fixed at 12 mm. Angulation of the limbs (2θ) were varied.
In vitro testing
A fenestrated stent-graft (Cook Europe, Bjaeverskov, Denmark) that had been custom made for a patient who died (from rupture of their aneurysm) just after manufacture of the stent-graft, but before insertion, was used for the experiments in this study. The stent-graft is made of woven polyester fabric and Gianturco stainless-steel stents.
The barbed bare metal anchor stent and the fenestrated proximal sealing stents were amputated off the tubular proximal body. The distal bifurcated body was deployed within the proximal fenestrated body such that there was a two stent overlap (35 mm length) between the proximal and distal bodies following deployment. The distal diameter of the proximal body stents was 22 mm. The distal bifurcated body had two proximal sealing stents with a diameter of 24 mm. A minimum of a two stent overlap is the recommendation of the manufacturer.
The top end of the proximal body and the iliac limbs of the distal body were attached to a tensometer (Model M5, NENE Instruments Ltd, Wellingborough, UK) via plastic sealing plugs and pneumatic clamps (Fig. 3). The plugs had custom cut channels to allow pressurisation of the inside of the stent-graft. The pullout tests were conducted in the vertical plane and the components were moved apart to a maximum displacement of 35 mm (where they became fully separated). The tensile tester was set to move at a constant rate of 100 mm/min. Force and displacement was recorded simultaneously with data being collected on a personal computer. The test was repeated 8 times dry, wet and pressurized.
Figure 3Distal body of a fenestrated stent-graft deployed within the proximal body and held in the clamps of the mechanical tester (Dry run, unpressurized). a) Two stent overlap at the start, b) Only one stent overlap, c) Complete distraction.
Following each distraction run the distal component was redeployed within the proximal component manually ensuring the full two stent overlap was achieved as well as the same alignment (Fig. 3a). No balloon moulding of the overlap zone was performed. A single operator performed all of the testing.
Under pressure the stent-graft fabric leaks and therefore the stent-graft components were coated with Gelatin (Sigma–Aldrich Company Ltd). This was achieved by injecting an aqueous gelatin solution (10% by weight) through the stent-graft until it had permeated fully through the pores of the fabric. To stabilise the gelatin, it was fixed/crosslinked by immersing the device in a glutaldehyde solution (1% by weight). This made the fabric water tight. For the wet run experiments the components were soaked for 20 min in normal saline solution before starting this set of experiments, and rinsed in saline between runs. Pressurisation of the stent-graft was achieved by a reservoir of normal saline, open to the atmosphere, being placed at the required height to give a pressure at the stent-graft level of 100 mmHg and 120 mmHg.
(2)
The pressure P in the stent-graft was calculated using equation (2) where h is the height of the reservoir with respect to the height of the stent-graft overlap, ρ is the density of normal saline and g the acceleration due to gravity. For 120 mmHg the reservoir was placed at 162 cm and for 100 mmHg it was placed at 135 cm.
All testing was performed at room temperature. Graft components were inspected after each test run to look for fabric and or stent damage.
Statistical analysis
Due to the relatively small sample size the data was assumed to be non-normally distributed. As a consequence summary statistics are presented as the median along with the interquartile range given in parentheses. The Kruskal–Wallis test was used to investigate possible differences between all four testing conditions and the Mann–Whitney test was used post-hoc to investigate differences between pairs of testing conditions. Statistical analyses were performed using SPSS software (Version 15; SPSS Chicago, IL, USA).
Results
Fig. 2 shows the calculated in vivo forces that the distal component is likely to be subjected to at various bifurcation angles and blood pressures. The bifurcation angle in the in vitro model (Fig. 3) is zero and was not varied.
In vitro a similar pattern is seen under all 4 testing conditions in that the force required rises to the peak force. It then drops off after approximately 15 mm which corresponds to there being only one stent overlap from this point (Fig. 4). As the amount of overlap decreased from this point onwards the distraction force became progressively less. The first part of these curves consists of two “humps” (peak followed by trough). The peak force of the first hump could be interpreted as the static friction force that had to be overcome before the components started to displace and the trough as dynamic friction force. The increase in force to form the second hump is difficult to explain. It is probably due to the inhomogeneous construction of the endograft (stents and fabric) that causes non-uniform friction forces.
Figure 4Median force to cause modular distraction of distal body. Vertical Lines indicate the range of forces (Min–Max).
Variability between runs under the same conditions was observed (Fig. 4). However, there was not a progressive increase or decrease in the force required between the first and eighth runs under each testing condition. An averaged force was not used because it smoothes out the curves that show important peaks representing component movement. There was no evidence of macroscopic damage to the stent-grafts observed between runs.
The median peak force required to distract the distal body from the proximal body dry was 2.9 N (2.6–4.1), wet and unpressurized 3.9 N (3.5–5.4), at 100 mmHg 6.3 N (4.8–8.3) and at 120 mmHg 6.5 N (4.8–7.2). There was a significant difference between pressurized and unpressurized conditions (P < 0.01).
Discussion
The incidence of fenestrated stent-graft body migration is unknown. In the authors' own institution two patients have required a cuff extension to prevent a modular disconnection between the proximal and distal bodies (Fig. 5). There have been two reports of actual complete body dislocation documented in the literature.
Figure 5Complications of Modular Distraction. A & B) Initially a two stent overlap but 4 years later a reducing seal zone risking a Type III endoleak with re-pressurisation of the aneurysm sac and the risk of aneurysm rupture.
Migration of a stent-graft occurs when the distal displacement forces exceed the strength of fixation at the proximal attachment zone. The decrease in cross sectional area at the bifurcation coupled with the usual posterior angulation in the aorto-iliac region, results in bifurcated devices in this region being subject to a considerable caudal displacement force.
A mathematical model to predict the in vivo pulsatile drag forces acting on bifurcated stent grafts used in endovascular treatment of abdominal aortic aneurysms (AAA).
This force tends to cause stent-grafts to migrate distally.
Originally the distal diameter of the proximal body stents were 24 mm and the distal bifurcated body had proximal sealing stents with a diameter of 24 mm as well. The manufacturer changed this to the current design of 22/24 mm due to reports of, and the risk of, modular distraction. Both patients in the authors' series that have required a cuff extension had received devices with the 24/24 mm design.
All the pullout tests in this study had similar characteristics (Fig. 4). The degree of variability in maximum distraction force under similar conditions was also observed by Hinchliffe et al during testing on iliac limb fixation strength in modular aortic stent-grafts.
They suggested this variability may be related to the sutures used to secure the stents to the graft fabric, i.e. sutures may snag on each other causing increased resistance.
The force required to distract the wet devices (unpressurized) was greater than when dry. This is likely to be due to the fabric absorbing water (both because of water permeating through the microporous/microfibrous structure of the fabric as well as water absorption by the polyester fibres). As a result the fabric becomes softer creating greater apposition and increased friction.
that assumed non-pulsatile, turbulent flow in a rigid, symmetrical graft, with an inlet diameter of 30 mm and outlet diameters of 10 mm, calculated a downward displacement force of 7–9 N on bifurcated grafts. Morris et al. reported that for an iliac angle of 30°, a proximal diameter of 24 mm and an iliac diameter of 12 mm, the drag force acting on bifurcated stent grafts varied, over the cardiac cycle, between 3.9 and 5.5 N in the axial direction.
A mathematical model to predict the in vivo pulsatile drag forces acting on bifurcated stent grafts used in endovascular treatment of abdominal aortic aneurysms (AAA).
but are similar to the forces measured in the pressurized studies above that caused modular disconnection. Fenestrated stent-grafts have been shown to have increased proximal fixation strength compared to standard infrarenal stent-graft
and the modular design of the fenestrated stent-graft body, with overlap, aims to place the drag component of the longitudinal migration forces on the distal bifurcated body, thereby reducing the potential for the fenestrations, in the proximal tubular body, to migrate. These estimated in vivo forces would suggest that the distal body should be expected to migrate.
is based on a number of simplifying assumptions to allow the axial force acting on a bifurcated device to be calculated. Because of the assumed symmetry of the bifurcation, the lateral force on the device is zero. The model does assume nonpulsatile flow and ignores the influence of gravity in order to simplify the equation. The forces shown in Fig. 2, displaying a linear increase in the axial force with blood pressure, can therefore only be regarded as rough estimates. An experimental validation of Mohan's model with measurements made under pulsatile flow, using a fluid with viscosity similar to blood, reported forces higher than predicted but only by 6–18%.
Fig. 2 does demonstrate that with very high pressures and large angulation, the haemodynamic forces in vivo may exceed the in vitro measured distraction force.
Pressure caused the distraction force required to be greater (Fig. 4) and this is a result of the components being “pushed” together by the luminal pressure, thereby increasing the frictional force between them. The current study only looked at the junction between the two stent-graft components. The situation is more complex than this in vivo. One needs to consider the effect of pressure on the fixation force of the proximal tubular component against the aortic neck wall as well as at the overlap of the two body components. The greater the blood pressure, the greater the forces acting on the stent-graft.
This force is due to the pressure acting on the cross-sectional area of the stent-graft and is resisted by the fixation force at the aortic neck and the overlap zone.
The proximal body distraction force has been shown to be dependent on oversizing.
When pressure is increased, the aorta expands (at least in a healthy elastic aorta) thus causing a reduction in the oversize and hence a reduction in the fixation force. Increasing pressure will also cause greater apposition of the graft against the aorta due to increased radial force and this will lead to greater frictional force, i.e. greater fixation force. These two effects work against each other and it is not clear what the net effect is. At the body component overlap zone, the effects will be similar except that the elasticity effect will be negligible because the two components are made of identical materials and have the same construction. Therefore oversizing at this region will not change significantly with pressure. The inner graft, however, will exert a greater radial force against the outer graft with increasing pressure. The friction in this region is probably greater (stent-graft on stent-graft) than at the aortic neck (stent-graft on aorta). The increased radial force against two high friction surfaces accounts for the greater distraction force with pressure.
This may have clinical implication in that potentially patients with fenestrated stent-graft who are hypertensive may be more likely to pass on more of the longitudinal distraction force to the proximal fenestrated component.
Large degrees of overlap (>2 stent overlap) are now being built into the interbody overlap of fenestrated stent-grafts to ensure modular disconnection does not occur.
Although the aneurysm sac (in the absence of sac expansion) should limit excessive migration of the distal body in vivo large amounts of overlap may have some disadvantages. It may allow excessive kinking to develop at the limb-body junction (Fig. 5) risking limb thrombosis/loss.
There is also the potential for increasing overlap to increase the frictional force, i.e. fixation strength, between the two bodies such that the fenestrated region is exposed to higher displacement forces that may risk target vessel patency. The optimal length of overlap and whether too much overlap adversely affects the proximal fenestrated component is an area for further study.
The force required to distract the distal bifurcated body of a fenestrated stent graft with two stent overlap is much lower than the proximal fixation strength of a standard infrarenal Zenith stent-graft. This was the manufacturer's aim with the intention of trying to protect the fenestrated proximal body from the full effects of the longitudinal migration forces in vivo. The current strength of the body overlap zone may actually be unnecessarily weak. This zone therefore requires careful surveillance in follow up. Whether the strength of this zone should be increased is an area for further study.
An important clinical message is demonstrated in Fig. 4 – the force required to cause further distraction reduces with diminishing overlap. This is relevant to the clinician when trying to decide if and when to intervene in follow up if movement is detected at the modular overlap.
Limitations
The in vitro model employed in this study relied on cranial movement of the proximal body relative to the distal body rather than caudal movement of distal body relative to the proximal tubular stent-graft body but this is unlikely to have materially affected the results. The effect of a gelatin coating on the fabric was not determined. The same stent-grafts were re-used in the tensometer for all testing. Graft fabric and stent damage sustained during the repeated deployments and distraction runs is a potential source of bias. In addition, there could potentially have been a manufacturing fault with the stent-grafts used. The cost of these customized stent-grafts meant that the procurement of additional devices for testing was not an option.
The effect of pulsatile blood flow was not assessed and the presence of a pseudointima may also have a significant effect on the distraction force required in vivo. These were not assessed within the current study, but Liffman et al. have investigated movement and dislocation of modular stents-grafts due to pulsatile flow in the thoracic aorta.
This in vitro study has demonstrated that the force required to distract the distal bifurcated body of a fenestrated stent graft is much lower than the proximal fixation strength (reported by others
) of a Zenith stent-graft. This was the manufacturer's aim with the intention of trying to protect the fenestrated proximal body from the full effects of longitudinal migration forces in vivo. Significant modular distraction caused the manufacturer to change the design from a 24/24 mm to a 22/24 mm. The strength of the overlap zone between the proximal and distal bodies may still be too weak to prevent long term modular distraction. In addition, the optimal length of overlap is yet to be defined.
Acknowledgements
Mr Geoffrey Gilling-Smith FRCS was part of the research team but sadly died before completion of the manuscript.
A mathematical model to predict the in vivo pulsatile drag forces acting on bifurcated stent grafts used in endovascular treatment of abdominal aortic aneurysms (AAA).
To submit a comment for a journal article, please use the space above and note the following:
We will review submitted comments as soon as possible, striving for within two business days.
This forum is intended for constructive dialogue. Comments that are commercial or promotional in nature, pertain to specific medical cases, are not relevant to the article for which they have been submitted, or are otherwise inappropriate will not be posted.
We require that commenters identify themselves with names and affiliations.
00406-6&id=gr1.jpg){kind=link}
00406-6&id=gr2.jpg){kind=link}
00406-6&id=gr3.jpg){kind=link}
00406-6&id=gr4.jpg){kind=link}
00406-6&id=gr5.jpg){kind=link}