Selection of materials for implants, including self-expanding stents, is essential for their clinical performance and safety. This article presents an overview of the known facts concerning the fundamental characteristics of the materials for the manufacturing of metal self-expanding stents, especially for implantation into the gastrointestinal tract (GIT). Facts are presented from the perspective of the manufacturer and therefore take into account the data purely technical as well as medical. For producers are also important regulatory principles (European directives and standards), which are not mentioned in the article.

A general definition of the stent presents it as a tube-shaped device that can be inserted into a narrowed or weak anatomical passageway to hold it open. Stents became widely used in many anatomical areas, organs or organ systems (central and peripheral vascular system, extravascular organ systems?GIT; respiratory tract, urinary tract, etc.). There are also stents, the use of which is beyond the general definition, e.g., stent supporting transjugular intrahepatic portosystemic shunt (TIPS), stent supporting drainage of the pancreatic cyst into the stomach or stent stopping the acute bleeding from the esophageal varices.

So widespread use means that stents are used in dozens of environments that differ in chemical composition, loads and speed of their change.

Duerig et al¹ distinguish two principal body fatigue environments: the strain-controlled and the stress-controlled. In the strain controlled environment a device is alternately deformed between two set shapes while in the stress controlled it is subjected to cyclic loading. Although the majority of fatigue environments in the body combines both above mentioned, the authors believe that the predominant is strain controlled environment due to the compliant nature of biological materials.

The GIT consists of hollow organs, in which the lumen represents different fatigue and biological environment. To avoid extensive discussion, we try to describe the environment in the esophagus and its effect on the implanted stent including interactions with stent material.

Factors of biological environment in the esophagus that affect implanted stent may be divided into the following groups: anatomy including its pathological change; esophageal peristalsis; esophageal propulsive force (EPF) and substances passing through the esophagus.

The esophagus is the muscular tubular structure 25 cm long in adults. In the resting state it is collapsed. It opens to accommodate swallowed or regurgitated material. It is approximately 30 mm in its longer lateral diameter and 19 mm in its shorter anterior-posterior diameter, although it has a capability of distending to greater dimensions. The above mentioned dimensions and their pathological changes (esophageal stenosis either straight or tortuous, stable or developing; esophageal dilation, etc.) are crucial to determine the dimensions of esophageal stents and their delivery systems. Dimensions of luminal diameter of the esophagus limit deformation of the stent. Severity of the esophageal stricture determined by its internal diameter, length and ease or difficulty of its dilation correlates with the mechanical stress of the stent and thus affects its properties, including material selection.

The wavelike contractions (peristalsis) is the next important factor defining the environment of the esophageal lumen. In the body of the esophagus, swallowing is associated with an initial wave of muscular inhibition followed by a sequential activation of longitudinal and circular muscles. This is primary peristalsis (the stripping emptying wave). The peristaltic velocity is of 3 to 5 cm/sec. Due to the action of the lumen, flexible stents mimic to some extent the movement of the lumen. This reduces the risk of injury to the esophageal wall but increases the risk of migration because the stent with constant diameter along its entire axial length does not follow expansion of the esophageal diameter in full. Rudney et al² determined the average number of swallows per hour as of 122, with a wide range of swallows per hour from 18 to 400.

Leaving aside the secondary peristalsis, the EPF is another important factor determining mechanical stress on the stent. Hsiao et al³ presents it as resulting from a large distensions of the esophagus due to an immovable bolus?stent. It consists of a single, sustained force developed by circular esophageal muscles which is significantly larger than that resulting from primary or secondary peristalsis. With respect to the esophageal lesion (no stenosis, benign and malignant stenosis), they define three types of esophageal responses to stent placement. In all three situations, the authors assume no peristalsis at upper and lower flare of the stent. In case of no stenosis they assume the EPF at the upper edge of the stent and at the base of the lower stent flare plus secondary peristalsis along the stent body. In case of benign stenosis they assume the EPF at the beginning of the stent body and at the base of the lower stent flare plus weak or no peristalsis along the stent body. In case of malignant stenosis they assume the EPF at the upper edge of the stent and at the base of the lower stent flare plus primary peristalsis along the stent body. The above mentioned model of forces developed by the esophageal wall demonstrates how difficult is to create an ideal esophageal stent.

Substances passing through the esophagus represent another factor determining internal environment interfering with implanted stent. In healthy individuals those substances are drinks, food, saliva and transient, physiological reflux of gastric contents. People consume many different foods and beverages with a large range of pH. Citrus fruits like oranges, lemons and limes are quite acidic (pH = 2.0?4.0). Carbonated drinks such as cola have a pH 4.0 to 4.5 while egg whites are slightly alkaline, with a pH of 8.0. The pH of tap water should be between 6.5 and 9.5. Recommended daily fluid intake in adults is 2.1 to 2.3 L/day. The amount of secreted saliva is 1 to 2 L/day, pH = 6.0?7.0.⁴ It contains 99.5% water and small amounts of other important substances such as electrolytes, mucus, antibacterial compounds, enzymes, epithelial, and bacterial cells.

Gastroesophageal (acid) reflux is in most subjects a benign physiologic process that occurs most commonly following meals. In these subjects, even repeated daily contact of the esophageal epithelium with acidic gastric contents with pH lower than 4 does not produce damage to the tissue.⁵ Patients with gastroesophageal reflux disease experience more episodes of reflux. Many of them are prolonged because of a failure of acid clearance. The above mentioned facts demonstrate high corrosivity of the environment in the esophageal lumen that attacks the implanted stent. Thus, material resistance to corrosion is one of the basic characteristics required for esophageal stents.

Materials that have been used in the past

Due to the extensive history of stents we concentrate on the materials used for the manufacture of GIT stents, especially esophageal. The history of self-expandable metallic stents in human GIT starts from Wallstent (Medinvent SA, Lausanne, Switzerland) that was implanted in esophageal cancer by Domschke et al⁶ in 1990. The stent was uncovered, woven in the form of a tubular mesh from surgical grade stainless steel alloy filaments. Five years later Gagneja et al⁷ mentioned in their overview 26 published studies of esophageal stents. Three main stent types differing in material as well as design were used: uncovered and covered Wallstent (Microvasive; Boston Scientific, Natick, MA, USA), covered Z-stent (Wilson-Cook, Winston-Salem, NC, USA) and uncovered or covered Ultraflex (Boston Scientific). The authors published a table presenting the following composition of the stent mesh: Wallstent, a biomedical grade cobalt-chromium-iron-nickel-molybdenum alloy (Elgiloy Specialty Metals, Sycamore, IL, USA); Z-stent, noble stainless steel; Ultraflex, nitinol. In 2011, Martinez et al⁸ published the article with the table listing 11 esophageal stents, which were then on the market in the US, Europe, and Asia. Four of them were made of stainless steel, six of nitinol and one (Polyflex; Boston Scientific) of polyester. Only one stent was uncovered, others were covered partially or fully. This demonstrated increasing trend of using covered stents made from nitinol. In addition, the first stent made from polyester thread and covered by silicone foil appeared on the market. Even more important position of nitinol is presented with the table published in 2013 by Song et al⁹ that lists the self-expandable metal and plastic, esophageal stents marketed in the US, Europe, and Asia.

It starts to show a trend towards the use of biodegradable materials (BD-Stent made from polydioxanone fiber; ELLA-CS, Hradec Kr?lov?, Czech Republic).

The way to ideal GIT stent (material, design, and mechanical performance)

The way to an ideal GIT stent is thorny. Laasch¹⁰ describes efforts to achieve the aforementioned ideal through various designs of self-expanding enteral stents (segmented, laser-cut, woven, braided and knitted, mostly covered) and materials (metal, plastic, and biodegradable polymers).

Hirdes et al¹¹ measured radial and axial force developed by 12 self-expanding esophageal stents commercially available in 2013 (i.e., 10 nitinol stents, 1 polyester stent, and 1 polydioxanone stent). The radial force, i.e., radial resistive force (RRF) and chronic outward force (COF), was measured with respect to the method presented by Duda et al.¹² For axial force measurements, the samples were fixed in a set-up comparable to the previous report by Isayama et al.¹³ Hirdes et al¹¹ distinguished two groups with quite different radial force curves.

The first group comprises braided, nitinol self-expanding stents (Wallflex, Evolution, and Niti-S stents) and the braided biodegradable, polydioxanone stent?BD-Stent (ELLA-CS). These stents develop low initial radial force (< 150 N) that gradually decreased to 0 N at full expansion.

Ultraflex, Alimaxx, Hanaro, and Polyflex stents form the second group of stents characterized by a high radial force (300?400N) when contracted followed by a steep drop in radial force during expansion. The high radial force may be attributed to the fact that these stent designs prevent elongation during contraction. The other factors also contribute to a higher radial force, e.g., stiffer stent material, thicker wire diameter or fully covered design.

Hirdes et al¹¹ demonstrated that majority of the stents with the high radial force develop the low axial force and vice versa. The Polyflex stent was the only stent demonstrating the high axial as well as radial force. Hirdes et al¹¹ consider the combination of high radial and low axial force to be optimal with respect to mechanical performance of an esophageal stent (Ultraflex and Alimaxx-ES). However, the authors admit a significant role of several other variables (material used to construct the stent and its covering, diameter of the wires, the size of each cell and the angle of the crossover wires, stent flares, uncovered parts of the stent, etc.).

Properties of different materials and the impact on stent manufacturing

As stated before, the materials most commonly used for the production of GIT stents include metals and their alloys (stainless steel, nitinol, Elgiloy-Phynox), non-degradable polymers (polyester), and degradable polymers (polylactide, polyglycolide, polydioxanone).

Stoeckel²³ summarized in his review seven reasons for nitinol’s success in construction of medical devices: elastic deployment, thermal deployment, kink resistance, constant stress, dynamic interference, stress hysteresis (biased stiffness), and temperature dependence of stress. Nitinol exhibits also other important properties, such as magnetic resonance imaging compatibility, biocompatibility and corrosion resistance.

Thermal deployment is not but a usage of the thermal shape-memory of nitinol. A device with a transition temperature (Af) of 30°C can be compressed and inserted into the delivery system at room or lower temperature. It start to expand to its preset shape when the temperature is raised to > 30°C, i.e., when the body heat warms it up. This feature of nitinol reduces the mechanical load of a delivery system with preloaded stent and enables to reduce its profile. Kink resistance prevents localization of the strain beyond the loading plateau level which is the essence of the wire kinking and stent crush. This is why the braided esophageal nitinol stents can be bent under the large angle without kinking.

Dynamic interference refers to the long-range nature of nitinol stresses. Unlike balloon-expandable stainless steel stents, self-expanding nitinol stents will always expand to their preset diameters without recoil. Although Stoeckel²³ presents nitinol vascular stents as an example of the dynamic interference, it can be reasonably assumed that nitinol GIT stent will also continue to gently push outward against the wall of implanted anatomic structure and follow its movements.

Biased stiffness (stress hysteresis) is a feature of superelastic nitinol, which has the most significant effect on the behavior of self-expanding nitinol stents. The hysteresis, stress-strain curve as it is described by Stoeckel²³ for vascular stents may be also applied to GIT stents. Once the stent reaches the diameter of the lumen of given anatomy, it starts to develop a low outward force (COF) against it. This force remains nearly constant even if the lumen would increase in diameter (dynamic interference). If it is compressed from the outside, the stent resists deformation with a higher force (RRF). The biased stiffness means that the stent exerts only low outward force, but resists deformation with a much higher force.

Temperature-dependent stiffness means that the plateau stresses are strongly temperature dependent above the transition temperature of the alloy. This feature allows to modify the stiffness of a superelastic nitinol stent at body temperature by adjusting the transition temperature of the nitinol alloy. Lowering the transition temperature makes the device stiffer at body temperature.

Duerig²⁴ compared physical properties of implanted metals (stainless steel, cobalt-chrome, titanium, and nitinol). He presented nitinol as super material in many ways (high strength, very low stiffness, fatigue good in strain control, excellent resistance against corrosion and shape memory as a bonus).

Biocompatibility is one of the most important parameters of materials intended for the manufacture of implants. Its evaluation is governed by the testing according to the standard EN ISO 10993?1²⁵ and associated standards. In spite of well-known toxicity of nickel, nitinol is considered to be high biocompatible material.

Ryh?nen²⁶ presented the study determining the nitinol being safe material. He tested the primary cytotoxicity of corroded nitinol and compared with that exhibited by samples of 316LVM stainless steel and pure titanium. Despite the higher initial nickel dissolution, NiTi induced no toxic effects, decrease in cell proliferation or inhibition in the growth of cells in contact with the metal surface.

The general soft tissue responses to NiTi were compared to corresponding responses to 316LVM stainless steel and Ti-6Al-4V alloy in rats during a follow-up of 26 weeks. The muscular tissue response to NiTi was clearly non-toxic and non-irritating. Ryh?nen²⁶ concluded the biocompatibility of nitinol seems to be similar to or better than that of 316LVM stainless steel or Ti-6Al-4V alloy. He stressed importance of optimal surface treatment of nitinol in long-term implants.

Sullivan et al²⁷ studied the effects of oxide layer composition and radial compression on nickel release in nitinol stents. They proved that additional Ni release due to stent radial compression is highly dependent on surface processing and that thicker oxide layers may be susceptible to cracking and exposing nickel-rich regions when compressed.

Sullivan et al²⁷ referred to the US Pharmacopeia that suggested a permissible daily exposure for nickel as a metallic impurity in drug products as of 0.5 μg/kg/day. This is greater than the highest Ni release (9 μg/day) for peripheral endovascular stents in the Sullivan’s study. However, the authors admit that larger stents or stents in an overlapped condition may generate higher nickel release rates than those reported in this study.

With respect to the above mentioned features the nitinol seems to be an ideal material for stents. However, the real situation of its clinical use is much more complicated.

The properties of the stent, including its resistance to fatigue and corrosion, are determined not only by the material (nitinol), but also by original material form (e.g., wire, tubing), fabrication technology (e.g., laser machining, braiding, knitting) its design-stent geometry (rings, helix, coil, unconnected, open cell, etc.) as demonstrated by Stoeckel et al²⁸ with “stent design pyramid”. If inappropriately chosen for a given application or body environment, each of these parameters may weaken the performance or safety of the stent.

Leaving aside the complications that generally accompany the clinical use of esophageal nitinol stents (migration, bleeding from aroded artery, pain, over-growth, etc.), we experienced a single but significant complication that really surprised us. One patient experienced massive destruction of wire mesh in central portion of the self-expanding nitinol esophageal stent (Esophageal Stent; ELLA-CS) after a 6-month period of implantation (Fig. 1, 2).

We conceived the suspicion that the damage was due to the chemical action of reflux liquid with a low pH, which came between layers of polyethylene coating of the stent, where it can not interfere with any self-cleaning mechanism. ELLA-CS organized extensive testing of wire samples taken from the failed stent. Laboratory testing was conducted by prominent Czech and foreign technical institutions and universities (The Institute of Physics Czech Academy of Science in Prague, The Institute of Chemical Technology in Prague; Materials Evaluation and Engineering Inc., Plymouth, MN, USA). We believe that the most accurate evaluation of failure mechanism was provided by The Institute of Physics Czech Academy of Science in Prague. Racek et al²⁹ proposed the mechanism for fatigue failure of superelastically cycled nitinol wires in simulated biofluids. The mechanism involves the eight steps starting from fracture of the TiO₂ surface layer by dense network of thin oxide cracks. These cracks are periodically open and allow repeated exposure of bare nitinol to the electrolyte (in our case to the reflux fluid). The next steps are hydrogen generation at the nitinol/liquid interface, its uptake by the bare nitinol surface and transport into bulk of the wire. The mechanism is continuing by local loss of the strength of the nitinol matrix due to the hydrogen embrittlement, transfer of the oxide cracks into the embrittled matrix, propagation of selected cracks along the hydrogen traps causing ultimately the brittle fracture of the wire. Fig. 3 and 4 demonstrate starting corrosion of the wire associated with the damage of TiO₂ layer and final damage to the wire through brittle and plastic fractures.

Khara et al³⁰ published a case report and review of the literature dealing with esophageal stent fracture. The authors reported 8 published cases of complete esophageal self-expandable metallic stents fracture. These were all nitinol stents from different manufacturers and the timing of stent fracture was anywhere from 8 to 40 weeks after initial stent placement. Half of these cases involved the Esophacoil stent, which is no longer available at the market. Partial nitinol esophageal stent fracture has been reported more commonly (33 patients in 6 references). However, most of these cases did not need any intervention. Only six cases required placement of a new stent through the lumen of the damaged stent. Khara et al³⁰ referred to diverse factors reported as potential cause of the stent fracture: the use of balloon catheter to dilate nitinol stent immediately after its deployment and laser application leading to thermal overstrain of the stent. However, other authors mention defective stent material or evaluated the fracture as spontaneous.

To prevent similar acute massive corrosion of its nitinol esophageal stents, ELLA-CS developed a method of covering the stent mesh by a silicone layer. Fig. 5 and 6 demonstrate uniform density and integrity of the silicone covering between nitinol stent arms, coats the stent wire mesh even at wire crossings. The images were taken by the laboratory of Institute of Materials Chemistry, Faculty of Chemistry of Brno University of Technology to the order of ELLA-CS.³¹

We do not think the hydrogen embrittlement to be the only cause of nitinol esophageal stent fractures. We believe that the nitinol stent breakage, particular the partial one, is not rare event. In our opinion it occurs occasionally, but the damage to the stent produces little or no clinical symptoms and therefore remains undetected. A lot of questions in this field leaves unanswered.

We trust in the future of non-degradable polymer materials, e.g., polyester which currently used in the stent Polyflex, as well biodegradable polymer materials like poly-lactic acid, poly-glycolic acid, poly-caprolactone, poly-dioxanone, poly-lactide-co-glycolide and polydioxanone (Esophageal BD-Stent; ELLA-CS). Lorenzo-Z??iga et al³² published the overview of GIT biodegradable stents concluded by conclusion on promising results of polydioxanone stent, however some aspects should be improved, e.g., premature losing of radial force, stent-induced mucosal or parenchymal injury, etc.

However, the way to widely used, standard esophageal biodegradable stent is difficult, and probably, long. This difficulty is caused by production problems such as the necessity of ensuring a high homogeneity of the starting polymer with a low content of impurities, monomer and catalyst residues. The rate of degradation must be suitably adjusted, particularly through proportion of crystalline vs. amorphous component, etc. Ensuring the biocompatibility is very complicated and expensive, especially in relation to degradation products, the speed of their formation, place of their accumulation (tissue, organ), elimination rate, etc.

There are many other problems such as the protection of the wire mesh of the stent against the unwanted degradation both during manufacture and storage. It is necessary to carefully determine the shelf life of the product. Also irreversible plastic deformation during prolonged compression of the stent is also a hard nut to crack.

It seems that the life is, as usual, ahead of regulatory activities because the majority of current regulatory documents is aimed at ensuring stability of implanted materials. A few of guidance documents issued by the U.S. Food and Drug Administration is an exception, e.g., Guidance Document for Testing Biodegradable Polymer Implant Devices.³³

Song et al⁹ list 12 self-expandable metal and plastic, esophageal stents marketed in the US, Europe, and Asia. Only two of them are presented as uncovered. The most frequently used materials are polyurethane, silicone or permalume silicone.¹¹ We consider polyurethane the most questionable material because of its predisposition to early degradation. Kim et al³⁴ questioned biodurability of polyurethane-covered biliary stent using a flow phantom. They demonstrated that stents exposed to bile remained intact for 2 weeks. Cracks were observed at 4, 6, and 8 weeks. Their size increased gradually. After 8 weeks several large holes in the polyurethane membrane were evident.

Current options

With respect to our experience we consider silicone of appropriate quality being the best option for GIT stent covering (mechanical properties, resistance to degradation/aging, biocompatibility including possibility of long-term implantation).

Advantages and disadvantages

The covering provides the stent with resistance against obstruction due to an ingrowth of malignant tissue and facilitates its retrieval. At the same time, however, it produces a series of production problems. Above all, once it is placed on the mesh of the braided stent, the covering changes considerably its properties, particularly mechanical performance (dimensions, expansion force). Such changes further progress over time. It is therefore necessary to subject the stent coating to accelerated aging, determine the changes and correct them by changing the settings of the stent original properties.

Future materials for GIT stent covering

We are developing a biodegradable polyurethane covering that should cover the biodegradable esophageal polydioxanone stent. We also assume that significant progress could produce drug-releasing coatings, e.g., releasing the drugs with anti-inflammatory effect.