Although traditional stents consist of nonocclusive metal scaffolds, the treatment of many disease processes requires the use of a “covered stent.” These stents are able to open vessels and provide a circumferentially occlusive boundary between the stent and vessel. Thus, covered stents are ideal for reestablishing the integrity of aneurysmal vessels or for minimizing the risk of in-stent stenosis.1–3 The potential applications of such covered stents include the treatment of coronary artery disease, aortic and central nervous system vascular aneurysms, carotid artery or pulmonary artery stenoses, and even treatment of ruptured vessels.1–7 In the palliation of congenital heart disease, specifically, the appropriate covered stent would be of tremendous value in stenting the ductus arteriosus, coarctation of the aorta, or in the stenting of pulmonary veins, an intervention often plagued by in-stent stenosis. Various materials have been used to cover stents, including silicone, polyurethane, and polytetrafluoroethylene.1,2,8 To date, the production of a highly flexible, durable, and thrombus-resistant material has not been achieved for all applications. Here we explore the use of a new thin film nitinol material for covering a stent.
Nitinol has many desirable attributes ideal for medical applications, especially in the construction of transcatheter devices. A biologically inert nickel-titanium (NiTi) alloy, nitinol exhibits both shape memory and super-elastic properties. The thermally induced phase transformation from its martensite to its austenite phase underlies nitinol’s shape memory capabilities. In the low-temperature martensite phase, nitinol is extremely malleable and can be compressed into catheters. Upon heating (in many cases simply to body temperature), nitinol transforms into its austenite phase and recovers from the deformation induced in the martensite state. In many applications (e.g., covered stents), it may be unnecessary to rely on the phase transformation but solely rely on the malleability of the Nitinol. In other words, the stent would produce the restoring deformation. In addition to its unique structural properties, nitinol is remarkably biocompatible.9,10 In physiological solutions, a titanium oxide layer forms on the surface, which prevents corrosion of the bulk material.11 Furthermore, nitinol is resistant to thrombus formation and does not calcify.12,13
Although fabrication of thin film nitinol (thickness of about 8 microns) has been attempted since the early 1990s using flash and vacuum evaporation, ion beam sputtering, and laser ablation, most of these methods have been unsuccessful in producing high-quality uniform film required for medical applications.14–17 DC Magnetron sputter deposition under ultra-high vacuum is a preferred method for the production of medical quality thin film nitinol because it allows for high levels of process controllability and “batch-to-batch” consistency.18 In brief, sputter deposition involves ejecting atoms from a target material and directing them to form a thin film on a substrate. Our group has previously demonstrated that target heating during sputtering creates films of uniform thickness and composition not achieved with conventional sputtering processes.19,20 This allows for precise process control of film composition. The effects of target composition and annealing temperatures on thin film nitinol transition temperatures have been previously described.19,20 The biocorrosive properties of thin film nitinol have also been previously studied.9,10,21–24
We hypothesize that thin film nitinol represents a superior material for manufacturing covered stents. The present study sought to determine both the feasibility and efficacy of thin film nitinol covered stents using both in vitro and in vivo models.
Methods and Materials
Thin Film Sputter Deposition
Nitinol films were deposited using a novel hot-target sputtering method.19 As residual gases can deplete the amount of titanium reaching the substrate, a residual gas analyzer (Stanford Research Systems, Sunnyvale, CA) was used to monitor residual gas contamination levels prior to sputtering. The combined pressure of water, carbon dioxide, and carbon monoxide gases were maintained below 10−9 Torr during sputtering. An argon scrubber further cleaned the argon to 99.999% purity. Sputtering of thin film nitinol onto a silicon wafer with 500-nm thick wet thermal oxide was accomplished with a 3-inch DC magnetron gun (MeiVac, San Jose, CA). The target consisted of bulk nitinol cut from a 3-inch boule of nitinol containing 48% nickel and 52% titanium (SCI Engineering, Columbus, OH). All films were deposited at base pressures below 5 × 10−8 Torr. The substrate-to-target distance was 4 cm and a sputtering power of 300 Watts was used. During deposition, the substrate was translated back and forth in relation to the target at 45° arcs and 80 mm length to minimize compositional variation. The deposited amorphous film was crystallized by heating to 500°C for 120 minutes before removal from the sputtering system.
Differential Scanning Calorimetry
A Shimadzu DSC-50 (Shimadzu, Kyoto, Japan) differential scanning calorimeter (DSC) was used to determine the transformation temperatures of the thin film nitinol. Thin film nitinol was removed from the wafer and a sample weighing 19 mg was cut from the freestanding film. The film was cut into small sections to reduce internal stresses during DSC testing. The specimen was heated to 150°C and then cooled to −20°C at a constant rate of 10°C/min. Transformation temperatures were determined from the endothermic and exothermic peaks of the heating and cooling curves.
Stress-Strain Curve Generation
To characterize the stress-strain and shape memory properties of the thin film nitinol, an MTS Tytron (MTS, Eden Prairie, MN) was used. The MTS Tyron has a displacement resolution of 0.1 μm and a minimum force of 0.01 N. Thin film nitinol was removed from the wafer using a crack and peel method to produced a free-standing film. Tensile samples were fabricated using a razor blade and a straight edge to produce strips of thin film nitinol with dimensions of 3 mm by 20 mm. The specimens were arranged in grips such that the length of the specimen was 10 mm. All tests were conducted at room temperature. Before testing, a small load (0.01 lbs) was applied to eliminate slack. The load on the thin film nitinol sample was ramped from 0.22 to 15.5 N (0.05–3.5 lbs) at a rate of 0.35 N/s (0.08 lbs/s). The load was then returned to 0.22 N and the film was heated to above the austenite finish temperature in order to record strain recovery.
Covered Stent Construction
Standard photolithography and etching techniques were used to generate precise two-dimensional shapes required to produce thin film nitinol sleeves for covering the stents. First, a positive photoresist (Clariant AZ 4620, Muttenz, Switzerland) was spin coated onto an 8-micron thin film nitinol coated silicon oxide wafer. The photoresist was exposed through a patterned glass mask (Computer Circuit Inc., Gardena, CA) and developed, leaving the desired PR pattern on the nitinol film. The unprotected portions of the thin film nitinol (areas without PR) were etched away in 1:1:15 HNO3:HF:H2O solution, and the remaining PR was removed with acetone. The fabricated thin film sleeves were mechanically removed from the silicon oxide wafer. This photolithography approach reduced the number of imperfections on the edges of the thin films, thereby reducing/eliminating the incidence of tearing.
Both self-expanding nitinol stents and balloon-inflatable stainless steel stents were mechanically covered with thin film nitinol. Specifically, balloon-inflatable PG1910B and PG2910B stents (19 and 29 mm in length, respectively) from Cordis (Johnson and Johnson, Miami, FL), and prototype pfm (pfm AG, Homburg, Germany) nitinol self-expanding stents (20 mm in length) were used for laboratory and animal testing.
Several nitinol cover designs were examined. The simplest “wrap” cover was ultimately employed for laboratory and animal testing (Figure 1A). In short, the “wrap” design employed a two-dimensional piece of thin film nitinol which was wrapped circumferentially around the stent. A tab-and-notch design was used to attach the thin film nitinol to the stent (Figure 1B). Tabs were fashioned on one end of the thin film nitinol used to cover the stent. These tabs were then woven through the lumen of the stent, around the stent strut, and back out to the exterior of the stent. In this manner, one end of the thin film nitinol was mechanically coupled with the stent.
Covers were affixed to balloon-inflatable Cordis stents as described above. To cover the self-expanding pfm stents, each stent was cooled with liquid Nitrogen to −195.95°C and compressed into 5 French sheaths. The thin film cover was then wrapped around this sheath. The 5 French sheath wrapped in nitinol was subsequently placed into a larger 7 French sheath. Finally, the 5 French sheath was then pulled back, leaving the thin film cover in place while simultaneously exposing the stent. The stent became covered by self-deploying into the thin-film nitinol wrap. The final 7 French sheath was then attached to a modified pfm stent delivery system.
In Vitro Testing
In vitro testing of the thin film nitinol covered stents was designed primarily to demonstrate the feasibility of successful stent deployment, and to test the success of different covered stent designs. A pulsatile flow loop was constructed using a Harvard Apparatus Pulsatile Blood Pump (Harvard Bioscience, Holliston, MA) and clear PVC tubing (10 mm in diameter). A 7 French sheath was used for the insertion of catheters and balloons for stent deployments. Results were recorded on both digital video and still images with pulsatile flow of 1.5 L/min.
In Vivo Testing
After obtaining approval from the Office for Protection of Research Subjects, a swine animal model was used for in vivo testing of thin film nitinol covered stents. Four animals were used.
After induction of adequate anesthesia, thin film nitinol covered stents were percutaneously implanted into appropriately sized vessels of each swine. All arterial implantations were performed in the descending aorta. Venous implants were performed in either the superior vena cava or inferior vena cava. In all animals, percutaneous vascular access was obtained via the Seldinger technique and cineangiograms were performed with machine injections of contrast solution via marker calibrated pigtail catheters. Fifty units per kilogram of heparin was administered before stent deployment. Animals were not further anticoagulated after recovery.
In the first animal implantation, one thin film nitinol covered Genesis PG1910B stent was deployed in the descending aorta (DAo). This animal was recatheterized and euthanized 2 weeks later. The three subsequent animals had a combination of balloon expandable and self-expanding thin film nitinol covered stents placed in the DAo, superior vena cava, and/or inferior vena cava. Control stents without thin film nitinol coverings were also implanted in each swine.
Swine were recatheterized 2, 3, 4, and 6 weeks after stent implantations for cineangiographic evidence of stent appearance at each of these time-points. All animals were euthanized with an intravenous injection of pentobarbitol while under general anesthesia.
After euthanization, stent-containing vessels were harvested, as well as specimens from major organs (heart, lungs, liver, stomach, kidney, pancreas, and spleen). Light microscopy was employed to analyze the covered and uncovered portions of the harvested stents. Gross examination, trichrome staining, and hematoxylin and eosin staining (H&E) were used to analyze the recovered organ specimens and stent implantation sites.
Scanning Electron Microscopy
Scanning electron microscopy was used to determine the surface characteristics of the thin films following postmortem explantation. The samples were air dried, sputtered with gold, and examined using a Cambridge Stereoscan 250 (Cambridge Instruments, Cambridge, MA).
Thin Film Characterization
A stress-strain curve quantifying the ductility and shape memory behavior of the thin films is shown in Figure 2. The modulus of the film was calculated to be 17.8 GPa and the transformation stress occurs at 136 MPa. The film withstood tensile forces above 425 MPa and was strained to above 5%. Upon unloading and heating, the thin film nitinol exhibited complete strain recovery showing excellent shape memory behavior. The DSC results confirmed the shape memory behavior of the thin film nitinol (Figure 3). The start and finish transition temperatures for the martensite (Ms, Mf) and austenite (As, Af) phases were determined from the exothermic and endothermic peaks of the cooling and heating curves. The Af, As, Mf, and Ms temperatures were found to be 86.8, 69.3, 26.6, and 44.3°C, respectively. A rhombohedral phase is also observed during cooling.
Results of In Vitro and In Vivo Deployments
Thin film nitinol covered stents deployed successfully in our flow loop in multiple consecutive trials. Upon deployment, all stents expanded with ease and remained in place throughout each trial. After both the self-expanding and balloon-expandable thin film nitinol covered stents were shown to deploy in a consistent fashion in the laboratory flow loop, animal studies were begun. Four animal trials were performed, resulting in the successful implantation of 10 thin film nitinol covered stents (4 Palmaz and 6 pfm). In angiograms taken 2–6 weeks after implantation, the thin-film nitinol covered stents in the descending aorta showed absolutely no in-stent stenosis. At 2 and 3 weeks, minimal neointimal obstruction was seen in the thin film nitinol covered stents placed in the venous circulation, whereas moderate-to-severe neointimal obstruction was observed in the uncovered portions of the stents (Figure 4). However, at 4 and 6 weeks, relatively severe neointimal in-stent stenoses were observed within the lumen of thin film nitinol covered stents placed in the venous circulation. This degree of stenosis was comparable to that observed with uncovered stent portions.
Upon postmortem dissection, the covered portions of all stents were easily removed, whereas uncovered portions were adherent to the vessel walls (Figure 5). Postmortem analyses correlated with the angiographic findings. No significant in-stent neointimal hyperplasia was found on thin film nitinol covered stents placed in the arterial circulation. Moderate-to-severe neointimal proliferation was seen in all venous stent implants after 3 weeks. Notably, several of the stent covers were found to have defects that were likely caused during deployment.
Scanning electron microscopy of nitinol coverings removed from stents placed in the arterial circulation revealed rapid endothelialization, with a thin film of endothelial cells coating the entire surface of the nitinol film by 21 days postimplantation (Figure 6). On microscopic examination, no endothelial injury was noted in vessel walls previously housing covered stent portions, compared with significant endothelial injury and neointimal proliferation observed in vessel walls housing uncovered stent portions (Figure 7). Microscopic examinations of organ tissues showed no evidence of device-related abnormalities.
Our short-term results with the first prototype thin film nitinol covered stents lend credence to the notion that thin-film nitinol will be an attractive material for producing novel, low profile, covered stents.
In addition to its shape-memory and super-elastic properties, thin film nitinol possesses remarkably high tensile strength. Tensile testing of the thin film nitinol used for this study showed not only immense strength but also significant elasticity (Figure 2). Of note, in this metallic system, there can be variability for different loading scenarios, but for a given sample from a given wafer the variability is minimal. Thus, although the mechanical analyses presented here are based on testing of a single wafer, these results are adequate representations of this material’s mechanical characteristics.
Furthermore, DSC analysis revealed few impurities, which likely contribute to both its strength and biologic compatibility. These properties make thin film nitinol particularly amenable for use in transcatheter devices. Moreover, thicknesses on the order of 5–10 microns allow for the construction of extremely low profile covered stents. Because this material is manufactured as a thin film, there is little room for fluctuations in surface texture, resulting in an extremely smooth surface. In contrast to bulk nitinol currently used in multiple biomedical applications, the DC sputtering process used in our laboratory produces thin films of nitinol which are completely free of contaminants. Thin film nitinol has also been shown by our group to be resistant to corrosion in a biological environment.
As this study was a short-term feasibility study, it is difficult to draw definitive conclusions about the suitability of thin film nitinol for covering stents. Nonetheless, it is certainly noteworthy that all stents were able to be delivered with their recommended delivery sheathes (i.e., the thin film nitinol coverings did not significantly increase the device size). Additionally, rapid endothelialization was observed, as all stents (up to 29 mm in length) became covered in a thin layer of smooth neointima within the first month after implantation. Furthermore, there were no indications of thrombosis or embolic phenomenon following stent implantation. There was no device-related injury noted on histology of vessel walls previously containing covered stent portions.
Defects were noted in some stent coverings. These tears were likely caused during deployment, and these findings indicate that the thin film nitinol used for this study were not resistant to perforation. Because thin film nitinol is known to have poor pierce resistance, potential designs of thin film nitinol covered stents must incorporate strategies to avoid piercing of the thin film by the stent scaffolding. Unfortunately, small defects in thin film nitinol can easily propagate and cause large tears.
Thin-film nitinol covered stents could theoretically prevent in-stent stenosis secondary to direct neointimal hyperplasia. Unlike the newly marketed drug-eluting stents, covered stents are able to prevent vessel growth into the lumen of a stent by forming a mechanical barrier to directly prevent neointimal proliferation into the stent’s lumen. Unfortunately, the intima can still grow into a covered stent’s lumen from its edges. In the arterial circulation, the thin film nitinol cover did prevent in-growth of neointima and supported growth of endothelial cells over the first month after implantation. In the venous system, there was angiographic evidence of less in-stent stenosis in the first 3 weeks after implantation (when compared with an uncovered stent). However, by 4 and 6 weeks all venous stents (covered and uncovered) had significant in-stent stenosis from neointimal growth. The vastly different hemodynamic environments encountered in the arterial versus the venous circulation likely contributed significantly to this differential result. The pulsatile arterial circulation, reaching peak flow velocities of approximately 1 m/s, is likely to have had a significant influence on the type and extent of neointimal proliferation when compared with the continuous low flow state present in the venous circulation. It is likely that the covered stents placed in the venous circulation prevented direct growth of the vessel wall through the stent but not onto and around the device. It may be possible to obtain improved results by covering both the outside and inside of the stent with thin film nitinol. Because thin film nitinol seems to promote a very thin layer of endothelialization, an inner covering may improve our results by shielding the lumen of the vessel from exposure to the bulk nitinol or stainless steel present in the stent scaffolding. As thin film nitinol is much smoother and contains many less contaminants than bulk nitinol, this inner covering could allow for a more favorable biological response when compared with bulk nitinol.
The present study is a preliminary attempt at evaluating the feasibility of using thin film nitinol to produce covered stents. The results of our in vitro and in vivo studies show that thin film nitinol is a suitable material for the production of covered stents. However, design improvements to reduce thin film tearing and in-stent stenosis in the venous circulation are necessary. Thus, further studies are warranted in determining whether thin film nitinol will prove to be a superior material for the covering of stents compared to PTFE. To reduce venous intimal proliferation into the stent, future generations of thin film nitinol covered stents will be produced with a layer of thin film nitinol on both the inside and outside of the stent. It is our hope that this technology can be used to produce a superior stent for use in a wide range of medical applications.
1. Saatci I, Cekirge HS, Ozturk MH, et al
: Treatment of internal carotid artery aneurysms with a covered stent: experience in 24 patients with mid-term follow-up results. Am J Neuroradiol
25: 1742–1749, 2004.
2. Lee KH, Won JY, Lee DY, Choi D, Shim WH, Chang BC: Elective stent-graft treatment of aortic dissections. J Endovasc Ther
11: 667–675, 2004.
3. Fischer B, Palkovic S, Wassmann H, et al
: Endovascular management of tandem extracranial internal carotid artery aneurysms with a covered stent. J Endovasc Ther
11: 739–741, 2004.
4. Dundar Y, Hill RA, Bakhai A, Dickson R, Walley T: Angioplasty and stents in coronary artery disease: a systematic review and meta-analysis. Scand Cardiovasc J
38: 200–210, 2004.
5. Rabe K, Sievert H: Carotid artery stenting. J Interv Cardiol
17: 417–426, 2004.
6. Cheatham JP: Stenting of coarctation of the aorta. Catheter Cardiovasc Interv
54: 112–125, 2001.
7. Magoufis GL, Vrachliotis TG, Stringaris KA: Covered stents to treat partial recanalization of onyx-occluded giant intracavernous carotid aneurysm. J Endovasc Ther
11: 742–746, 2001.
8. Repici A, Conio M, De Angelis C, et al
: Temporary placement of an expandable polyester silicone-covered stent for treatment of refractory benign esophageal strictures. Gastrointest Endosc
60: 513–519, 2004.
9. Shabalovskaya SA: On the nature of the biocompatibility and on medical applications of NiTi shape memory and superelastic alloys. Biomed Mater Eng
6: 267–289, 1996.
10. Shabalovskaya SA: Surface, corrosion and biocompatibility aspects of Nitinol as an implant material. Biomed Mater Eng
12: 69–109, 2002.
11. Firstov GS, Vitchev RG, Kumar H: Surface oxidation of NiTi shape memory alloy. Biomaterials
23: 4863–4871, 2002.
12. Tepe G, Duda SH: Thrombogenicity of various endovascular stent types: An in-vitro evaluation. J Vasc Interv Radiol
13: 1029–1035, 2002.
13. Stepan LL, Levi DS, Carman GP: A Thin Film Nitinol Heart Valve. J Biomech Eng
127: 915–918, 2005.
14. Sekiguchi Y, Funami K, Funakubo H: Deposition of NiTi shape memory alloy thin film by vacuum evaporation. Proceedings of the 32nd Meeting of Japan Society of Materials, pp. 65–67, 1983.
15. Makino E, Uenoyama M, Shibata T: Flash Evaporation of NiTi shape memory thin film for microactuators. Sensors Actuators A Phys A
71: 187–192, 1998.
16. Lehnert T, Tixier S, Boni P, Gotthardt R: A new fabrication process for Ni-Ti shape memory thin films. Mat Sci Eng A: Struct Mater Properties Microstruct Process A
273: 713–716, 1999.
17. Ikuta K, Hayashi M, Matsuura T, Fujishiro H: Shape memory alloy thin film fabricated by laser ablation. Proceedings IEEE Micro Electro Mechanical Systems, pp. 355–360, 1994.
18. Ming H, Wu K: Fabrication of nitinol materials and components. Proceedings of the International Conference on Shape Memory and Superelastic Technologies, Kunming, China, pp. 285–292, 2001.
19. Ho K, Carman G: Sputter deposition of NiTi thin film shape memory alloy using a heated target. Thin Solid Films
370: 18–29, 2000.
20. Woolman J, Carman G: Composition and annealing effects on the mechanical properties of superelastic thin film nickel titanium. PhD thesis, UCLA, Los Angeles, CA, 2003.
21. Armitage DA, Parker TL, Grant DM: Biocompatibility and hemocompatibility of surface-modified NiTi alloys. J Biomed Mater Res
66: 129–137, 2003.
22. Ryhanen J, Niemi E, Serlo W: Biocompatibility of nickel-titanium shape memory metal and its corrosion behavior in human cell cultures. J Biomed Mat Res
35: 451–457, 1997.
23. Rocher P, El Medawar L, Hornez JC: Biocorrosion and cytocompatibility assessment of NiTi shape memory alloys. Scripta Materialia
50: 255–260, 2004.
24. Stepan LL, Levi DS, Gans E, Panduranga MK, Ujihara M, Carman GP: Biocorrosion investigation of two shape memory nickel based alloys: Ni-Mn-Ga and thin film nitinol. J Biomed Mat Res
82: 768–776, 2007.