Stent-Graft Design and Engineering

Timothy A.M. Chuter, D.M.
 UCSF Division of Vascular Surgery

In the absence of well validated design parameters or realistic models for pre-clinical testing, most advances in stent-graft design have come from clinical experience. With a few notable exceptions, stent-grafts have undergone convergent evolution, as successive generations of devices have incorporated advantageous features. The causes of late failure have been the most persistent, because, by their nature, they take longer to identify.

Stent and Barb Fracture
Self-expanding stents are made from either Nitinol or 300 series stainless steel. Nitinol, a nickel titanium alloy, is versatile, but fragile. Its superelasticity and thermal memory confer advantages in manufacture and use. In blood, the surface of Nitinol passivates by forming a layer of Titanium oxide, which fractures and separates when stent deformation exceeds strict limits. Recent improvements in Nitinol-based stent-grafts have involved electropolishing and the avoidance of welds (1). Nevertheless, stent fractures are a common finding on explanted Nitinol-based stent-grafts. High rates of stent breakage affect both Talent (1) and AneuRx (2) devices. Self-expanding stainless steel-based stent-grafts have also been seen to fracture after long-term implantation. One problem has been corrosion of solder at junctions in the wireform. Another is a stent-graft configuration or stent design that produces locations of high stress and strain. For example, the large stent just above the bifurcation of the Zenith stent-graft was found to be fractured in a high proportion of long-term implants.

Fabric Erosion
The combination of stiff, hard, metal stents and a softer, more yielding, fabric graft in a constantly moving prosthesis is a potential cause of structural failure. Where the two meet the stent may damage the fabric, especially if the junction permits movement, or the apex of the stent impinges on the fabric (3). One way to minimize potentially damaging movement of stent on graft is to suture the two components together at multiple points. However, this can also damage fabric structure by disrupting spacing in the weave and creating small holes. AneuRx stent-grafts, for example, have an average of 1.3 holes larger than 0.2mm per explant (2). This porosity may explain why aneurysms treated with the AneuRx stent-graft shrink at a lower rate than with other stent-grafts (4).

Migration
Distal migration of the stent-graft has become the commonest cause of late failure in the United States. If left untreated, migration leads to secondary type I endoleak, aneurysm pressurization and rupture. Not surprisingly, retrospective analysis of Eurostar data identified migration as a strong risk factor for subsequent aneurysm rupture. Factors that help to maintain stent position include: friction, barb penetration, suprarenal attachment, column strength and arterial ingrowth. Barbs, in particular, have been shown to greatly increase the pull-out force in studies of cadaveric stent-graft implantation (5). The clinical role of barb-mediated stent-graft attachment is illustrated by the difference in migration rates between AneuRx, and Ancure, both of which are implanted entirely below the renal arteries.

Component Separation
Many of the forces that cause proximal stent migration also act on the junction between components of a modular system. All of the current designs depend on friction in the overlap zone to stabilize the connection. Strong stents and long overlap reduce the risk. The risk of component separation also depends on the relative lengths of the trunk and limbs. Short limbs have a restricted range of movement, which lowers their potential for dislocation. None of the current devices has any mechanisms for active fixation of inter-component junctions, but protoypes with this feature are in development.

Kinking
In patients with aortic aneurysms, the iliac arteries are often tortuous or irregular. Both features predispose the unsupported graft limbs of stent-grafts like the Ancure to thrombosis. Wallstent implantation seems to be an effective remedy. All other stent-grafts have some form of limb support and rates of limb thrombosis are generally in the 3-5% range.

Blood flow and pressure apply large, caudally directed forces to aortic stent-grafts (6). The calculated effects of blood pressure and stent-graft size correlate well with risk factors for migration.  However, published analyses ignore temporal and spatial effects, and rest on questionable assumptions, such an external (sac) pressure of zero. More sophisticated computational flow dynamics (CFD) and observed effects, such as stent-graft movement are required for realistic finite element analysis (FEA) and accelerated durability testing (ADT).

Pulsatile Movement of the Stent-Graft
It is somewhat surprising that nothing has been published on temporal variations in stent-graft shape, size or position through the cardiac cycle, given that all stent-grafts are inserted under fluoroscopic guidance and even Nitinol stents are radioopaque enough to be seen with a standard C-arm. These movements are more than just an interesting phenomenon. Standard methods of pre-clinical durability assessment model the effect of a hypothetical aortic pulse on stent diameter. Yet the actual extent of pulsatile expansion and contraction are unknown. Currently, the only way to assess the validity of these tests is by their ability to predict structural failure of the stent-graft in clinical use. By that time, it is a little late for the unfortunate patient.

Computational Flow Dynamics
A rapid increase in available computer processing power has aided the development of computational methods that map temporal variations in flow, wall shear, and pressure in different phases of the cardiac cycle. These analyses are even better when they are based on measured pressures and real geometries. The results provide explanations of the distribution of thrombus on the wall of the stent-graft and clinical observations of stent movement and stent fracture.

Design of New Stent-Grafts
New stent-grafts help to deal with challenging anatomic features, such as tortuosity and implantation site compromise, and challenging locations, such as the proximal thoracic aorta. Fenestrations and side branches maintain flow to vital arteries, while excluding the aneurysm from the circulation.

Descending Thoracic Aorta
The closer a stent-graft lies to the aortic valve, the higher the hemodynamic forces, and the more difficult it becomes to achieve durable endovascular aneurysm repair. Thoracic aortic implantation sites tend to be large, compliant, angulated and widely separated.

Fenestrated and Branched Stent-Grafts
These two technologies are evolving along convergent paths, as fenestrated devices are combined with bridging stent-grafts, not just bridging stents. We have used modular multi-branched stent-grafts in the aortic arch, thoracoabdominal aorta, and iliac bifurcation (7-9). Short term issues of device insertion, and long term issues of durability raised different design considerations.

All aspects of endovascular aneurysm repair are device specific and some are now predictable, based on the observed behavior of certain design features. In the case of new applications, modularity simplifies the approach and allows greater predictability, since the new designs are often built by combining previously tested components, and the new procedures developed by combining previously used techniques.