`i
`University of Wisconsin
`1305 Uilden Dr.. Madison, Wi::>. 53706
`
`... · ...... "'_~~.l.:::,:'"~.' .... ;.~.~.-~..
`
`•
`
`.
`
`American
`. Journal of
`Roentgenology
`
`September 1983
`
`Edwards Exhibit 1005, p. 1
`
`
`
`601
`
`A New Percutaneous Vena
`Cava Filter
`
`A new, percutaneous vena cava filter was developed and tested in 11 dogs. Nitinol
`wire (0.045 cm diam) spitals were straightened and passed through an 8 French Teflon
`catheter into the inferior vena cava. The filter resumed its original spiral shape on
`warming to body temperature and was left in place up to 6 weeks. Long-term patency
`and capture of injected radiopaque clots were demonstrated. This new filter design
`may be a simple alternative to currently used implants.
`
`Pulmonary thromboembolism has been treated surgically by both ligation (1]
`and plication [2, 3] of the inferior vena cava. During the last 10 years, transven(cid:173)
`ously placed caval filters [4,5] and'occluding balloons [6] have been developed.
`While most of these techniques are effective in trapping emboli, a surgical
`cutdown of a peripheral vein is required and long-term patency is inconsistent
`[7]. These drawbacks have prompted recent investigators to search for a more
`easily placed and effective vena cava filter [8, 9].
`Recently, 8im.on arid associates (8, 10] developed a filter using nitinol wjre.
`Nitinol, a unique alloy with a heat-dependent memory, can be passed as a
`straight wire through relatively small angiographic catheters. When extruded and
`warmed to body temperature, the wire resumes its original complex shape.
`We recently developed a new technique for the nonsurgical placement of
`arterial endoprostheses using nitinol helices [11]. We believed that this same
`basic shape could be modified to act as an effective percutaneous caval filter.
`The design is' simpler than those previously reported and may be less thrombo(cid:173)
`genic since less wire is introduced into the vena cava. Its method of introduction,
`ability to capture emboli, and long-term patency are the subject of this report.
`
`Materials and Methods
`
`Filter Desi!J;,
`
`. The characteristics of nitinol wire have been described in detail [8, 11). Briefly, nitinol is
`a nickle-titanium alloy with·a heat-sensitive memory. If the wire is given a certain shape and
`annealed at greater than 500°C, it will memorize that shape. On cooling in ice water, the
`wire becomes soft and can be straightened without destroying its memory. When the wire
`is heated to its transition temperature (about 30°C for the alloy used in this study), it rapidly
`resumes its original shape, becoming more rigid than stainless steel.
`Vena cava filters were made by wrapping nitinol wire (0.045 cm diam) on a specially
`designed steel mandrel (fig. 1 A). The cone-shaped mandrel was threaded so that the
`distance between the threads was 2.75 mm. In this fashion, a helical spiral was produced
`with constant openings of 2.75 mm between each turn of the wire (fig. 1B). At the apex of
`the spiral, the wire was drawn out to form a large stabilizing loop that kept the cone of the
`filter centered in the middle of the caval lumen. The first filter in the study embolized to the
`lung within 2 weeks of placement. In all other filters, therefore, a small, 2 mm hook was
`made at the end of the stabijizifig ring, which penetrated the caval wall and securely
`
`L
`i'
`II ~, ;
`
`'I
`I
`.1
`
`Andrew Cragg 1
`Gunnar Lund
`Erich Salomonowitz
`Joseph Rysavy
`Flavio Castaneda
`Wilfrido Castaneda-Zuniga
`Kurt Amplatz
`
`Received Aprif 13, 1983; accepted May 13,
`1983.
`1 .AII authors: Department of Radiology, UnIver(cid:173)
`sity of Minnesota Hospitals, Box 701 Mayo, 428
`Delaware St. S.E., Mi[1neapolis, MN 55455. Ad(cid:173)
`dress reprint requests to A. Cragg.
`AJR 141:601-604, September 1983
`0361-803X/83/1413-0601
`© American Roentgen Ray Society
`
`Edwards Exhibit 1005, p. 2
`
`
`
`'I I
`
`602
`
`"
`
`I
`
`i
`I;
`
`CRAGG ET Al.
`
`\.
`
`.)
`
`"
`
`AJR: 141, September 1983
`
`Fig. i.-A, Stainless steel filter man(cid:173)
`drel. Nitlnol wire is wrapped on mandrel
`and annealed in BunseQ burner flame.
`Open area of filter can be, varied by
`changing tightnel>s of spiral. B, Nitinol
`spiral filter. Delivery wire is screwed to
`adapting plug (closed arrow) and filter is
`drawn into catheter. When extruded in
`body, filter resumes its shape. Stabiliz(cid:173)
`ing ring (open arrow) centers filter cone
`in lumen of vena cava. C, Filter viewed
`from below. Spiral design provides ef(cid:173)
`fective sieve without restricting blood
`flow. Anchoring hook (arrow) extendl> 2
`mm from stabilizing ring.
`
`,. ,. r
`
`"
`
`A
`
`c
`
`anchored the filter (fig. 1 C).
`Most investigators have stated that nitinol wire should be an(cid:173)
`nealed at 525°C for 20-30 min [8]. We found that equally good
`resu'its are obtained by heating the mandrel and wire in the flame of
`a Bunsen burner until the wire changes from its unannealed straw
`color to a deep blue. This process Is simpler and can be accom-
`plished in less than 5 min.
`. \
`After annealing, the mandrel is allowed to cool and the wire is
`removed in ice water. Usually, an "oversized" filter is constructed
`by winding extra turns of the wire on the mandrel. In this way the
`filter could be fitted exactly to a given vena cava size by trimming
`one or more of the spiral turns, thus decreasing the overall ~iameter
`of the filter.
`!
`
`Animal Studies
`
`Eleven adult mongrel dogs (15-30 kg) were anesthetized with
`intravenous sodium pentobarbital (30 mg/kg) and maintained on
`spontaneous respiration. An 8 French introducing sheath was
`placed percutaneou?ly in a femoral vein and connected to a hepa(cid:173)
`rinized (2 U/ml) saline solution. An inferior vena cavogram was
`obtained by injecting 15 ml of 60% diatrizoate contrast medium at
`10 mil sec through the sheath.
`The diameter of the infrarenal part of the vena cava was measured
`and the filter trimmed accordingly. An adapting plug [11] was
`attached to the trailing wire of the filter, and a delivery wire was
`screwed to the filter. The filter was then immersed in sterile iced
`saline and drawn into a 60-cm. 8-French Teflon catheter. When
`fully straightened, the filter was about 20 cm long.
`The catheter with the loaded nitinol wire was Inserted through
`the introducer and passed under .fluoroscopic control to the desired
`level in the inferior, vena cava. The filter was deposited in the vena
`cava by withdrawing the catheter over the adapting wire. As the
`
`straightened nitinol wire came into contact with the blood it rapidly
`resumed its original filter shape. Once extruded, the filter was pulled
`down to the desired level and then quickly advanced by pushing on
`the delivery wire, This caused the hook on the filter to engage the
`wall of the cava thus anchoring the filter. A venogram was obtained
`immediately after extrusion of the filter to document its position and
`anchoring. If the position of the filter was not optimal, It could be
`withdrawn into the catheter and positioned again. To release the
`filter, the delivery wire was unscrewed and withdrawn together with
`the catheter from the vena cava.
`The ability of the fiiter to capture emboli was tested in five animals
`acutely and in one chronic animal at the time of sacrifice. Emboli
`were prepared by drawing fresh blood into a thin-walled polyethyl(cid:173)
`ene catheter (3 mm diam, 10 cm long). A 'small amol!nt of tantalum
`powder was added to the blood for opacific<:ition. After 10 min, the
`clot was injected into a femoral vein. This was monitored fluoro(cid:173)
`scopically and plain films of the abdomen and ches~ were made to
`document the effectiveness of the filter.
`Long-term patency studies were performed in six animals. Ven(cid:173)
`ograms were repeated at 2 weeks and again at 6 weeks before
`sacrifice (fig. 2). The animals were killed by an ove~dose of pento(cid:173)
`barbital, and the abdomens were opened and examined for evi(cid:173)
`dence of retroperitoneal bleeding. The venae cavaerwere isqlated
`and the filters examined for evidence of thrombosis.
`None of the animals received long-term anticoagulation.
`
`Results
`
`Filters were placed successfully, in all animals. (n two
`animals, the original placement of the filters was not optimal.
`These filters were withdrawn into the catheter and success(cid:173)
`fully repositioned.
`Venograms performed
`
`immediately after ,Placement
`
`Edwards Exhibit 1005, p. 3
`
`
`
`AJR:141, September 1983
`
`NEW VENA CAVA FILTER
`
`603
`
`Fig. 2.-A, Vena cavogram 6 weeks
`after percutaneous placement of nltlnol
`spiral filter. Vena cava is widely patent
`and anchoring hook has penetrated wall
`(arrow). At autopsy, hook was com(cid:173)
`pletely covered by mound of fibrous tis(cid:173)
`sue. B, Same filter after injection of ra(cid:173)
`diopaque autologous clot. Clot (3 mm
`x Hl em) (arrow) was completely cap(cid:173)
`tured by filter.
`
`A
`
`B
`
`showed patent filters in all cases without evidence of throm(cid:173)
`bosis. Of the six chronic animals, four had widely patent
`filters at 2'and 6 weeks (fig. 2A). In one animal, the lumen
`was 60% stenosed at 2 weeks and occluded at 6 weeks.
`This filter: was the only one placed with the cone pointing
`down instead of up. The unanchored filter that embolized to
`the lung was found lodged in the left pulmonary artery at 6
`weeks. The pulmonary artery was still widely patent and the
`filter was free of thrombus.
`The capture of 3 mm x 10 cm autologous clots was
`complete in all six filters studied (fig 28). In several In(cid:173)
`stances, the clot fragmented into several smaller pieces on
`extrusion from the catheter; however, even small 1-2 cm
`emboli appeared to be captured by the filter. Opacified clot
`could not be found on postembolization chest radiographs.
`It is possible that very small fragments passed through .the
`filter and were not detected on the chest film.
`On postmortem examination, the anchoring hook pro(cid:173)
`truded through the wall of the vena cava 1-3 mm and was
`covered by a small mound of fibrous tissue. No pericaval
`bleeding was noted. The largest rings of the filter were
`incorporated into the wall of the vena cava by 6 weeks. In
`most cases, small fibrous bands extended between the
`spirals of the filter. (
`
`Discussion
`Surgical interruption of the vena cava in patients prone to
`pulmonary thromboembolism largely has been supers~ded
`
`by transvenous mechanical vena cava filters. The main
`disadvantage of these filters is the need for a surgical
`cutdown in a usually very sick patient. A percutaneous
`introduction technique of standard filters has been de(cid:173)
`scribed, but because of the large size of the delivery cap(cid:173)
`sule, this technique has not gained 'popularity (12, 13]. The
`development of an effective percutaneous filter therefore
`would be a great improvement over present techniques.
`T~e' nitinol spiral filter developed in our laboratory pos(cid:173)
`sesses several advantages over other vena cava filters: (1)
`It can be introduced percutaneously using an 8 french
`catheter; (2) the filter is easily trimmed to be adapted to
`different· vena cava sizes; (3) it can be withdrawn and
`repositioned easily; (4) proper orientation'of the filter cone
`is maintained by a stabilizing loop; and (5) the open area of
`the filter can be adjusted by winding a tighter or looser spiral
`on the mandrel.
`The disadvantage of the present design is the mode of
`fixation. While one anchoring hook was adequate for normal
`animals, we believe that a more secure system of fixation
`needs to be developed before human trials can be consid(cid:173)
`ered.
`While only a small number of long-term filters was tested,
`the thrombogenicity and biocompatibifity of the filter seem
`comparable to those in present use [14]. In vivo thrombo(cid:173)
`genicity studies have not been performed on the nitinol filter
`developed by Palestrant et at. [10]; therefore, we cannot
`compare the relative thrombogenicity of the two filter de-
`
`-',1
`
`"
`
`Edwards Exhibit 1005, p. 4
`
`
`
`604
`
`CRAGG ET AL . .
`
`AJR: 141 , September 1983
`
`signs. Undoubtedly, however, the most important factor in
`determinit:lg thrombogenicity of a venous prosthesis is the
`degree to which it obstructs blood flow. Thus an inverse
`relation exists between thrombogenicity and the open area
`of the filter. The more wire, the better the filter action but
`the greater the tendency to thrombose. In humans, long(cid:173)
`term anticoagulation may favorably influence this balance.
`We believe that the spiral filter shape used in this study may
`provide the most effective sieve with the least amount of
`wire, thereby decreasing thrombogenic'ity.
`It is interesting to note that the thrombosed filter was the
`only one oriented with the apex of the cone down. It is
`possible that the orientation of a filter affects its thrombo(cid:173)
`genicity. Greenfield et al. [15] claim that the optimal orienta(cid:173)
`tion of a cone-shaped filter is with its apex directed upward.
`It is claimed that with this orientation a, large embolus
`occludes less of the lumen and has a self-anchoring effect
`by distending the filter against the caval wall. For this
`reason, we chose to orient our filter with the apex directed
`upward.
`In summary, an effective vena cava filter has been devel(cid:173)
`oped, which overcomes many of the disadvantages of cur(cid:173)
`rently used filters. The filter is introduced percutaneously
`through a small catheter and effectively stops propagated
`emboli. With the development of a more effective anchoring
`system, the design may be tested in humans.
`
`REFERENCES
`
`1. O'Neill EE. Ligation of the inferior vena cava in the prevention
`and
`treatment of pulmonary embolism. N Engl J Med
`1945;232: 641-646
`2. DeWeese MD, Hunter DC. A vena caval filter for the prevention
`of pulmonary emboli. Bull Soc tnt Chir 1958;17: 117 -125
`3. Adams JT, DeWeese JA. Experimental and clinical evaluation
`
`c
`.'
`of partial vein interruption in the prevention of pulmonary
`emboli. Surgery 1965;57: 82-1 02
`:
`4. Mabin-Uddin K, McLean R, Jude JR. A new catheter technique
`of interruption of inferjor vena cava for prevention of pulmonary
`embolism. Am Surg 1969;35:889-894
`5. Greenfield LJ, McCurdy JR. Brown pp'. Elkins RC. A new
`intracaval filter permitting continued flow and resolution of
`emboli. Surgery 1973;73: 599-606
`6. Hunter JA. Dye WS, Javid H, Najafi H. Goldin MD, Serry C.
`P.ermanent transvenous balloon occlusion of the inferior vena
`cava. Ann Surg 1977;186:491-499
`7. Cimochowski GE, Evans RH, Zarins CK, Lu C, DeMeester MD.
`Greenfield filter versus Mobin-Uddin umbrella.· J. Thorac Car(cid:173)
`diovasc Surg 1980;79: 358-365
`8. Simon M. Kaplow R, Salzman E, Freiman D. A vena cava filter
`using thermal shape memory alloy: experimental aspects. Ra(cid:173)
`diology 1977;125~89-94
`9. Gianturco C, Anderson ~H. Wallace S. A new vena cava filter:
`experimental animal evaluation. Radiology 1980; 137: 835-
`837
`10. Palestrant AM. P~ince M, Simon M. Comparative in viiro eval(cid:173)
`uation of the nitinol inferior vena cava filter. Radiology
`1982;145:351-355
`.
`11 .. Cragg A, Lund G, Rysavy J, Castaneda F. Castaneda-Zuniga
`W, Amplatz K. Nonsurgical placement of arterial endo(cid:173)
`prostheses: a new technique using nitinol wire. Radiology
`1983;147: 261-263
`12. Rizk GH, Amplatz K. A percutaneous method of introducing
`the caval umbrella. AJR 1973;117: 903-909
`13. Formanek A, Castaneda-Zuniga W, Knight L, Amplatz K. Three
`years experience with percutaneous introduction of the inferior
`vena cava filter. Rev tnteram Radiot1977;2: 171-175
`14. Edie JW. Andreasen GF, Zaytoun MP. Surface corrosion of
`nilinol and stainless steel under clinical conditions. Angle Or(cid:173)
`thodontist 1981;51 :319-324
`15. Greenfield LJ. Payton MD, Brown PP, Elkins RC. Transvenous
`management of pulmonary embolic disease. Ann Surg
`1978;180:461-468
`
`. .-'
`
`/
`
`I
`!
`
`~I
`
`I,
`
`.,
`~ i
`
`ill
`
`I
`
`;:
`
`1:1
`;:1
`;'1
`I:.
`I .,
`" ii I!
`Ii
`
`Edwards Exhibit 1005, p. 5
`
`
`
`Edwards Exhibit 1005, p. 6
`
`Edwards Exhibit 1005, p. 6
`
`