| | Computational Modeling of Blood Flow in the TrapEase Inferior Vena Cava FilterReceived 17 February 2008; received in revised form 31 January 2009; accepted 4 February 2009. published online 30 April 2009. PurposeTo evaluate the hemodynamics of the TrapEase vena cava filter (Cordis, Miami Lakes, Florida) by using three-dimensional computational fluid dynamics, including simulated thrombi of multiple shapes, sizes, and trapping positions. The study was performed to identify areas of stagnant and/or recirculating flow that may have an effect on intrafilter thrombosis. Materials and MethodsThree-dimensional computer models of the TrapEase filter, various thrombi shapes and sizes, and a 23-mm-diameter cava were constructed. The hemodynamics of steady-state flow were examined for the unoccluded and partially occluded filter. ResultsFlow in the unoccluded TrapEase filter experienced minimal disruption. Spherical thrombi in the downstream trapping position induced stagnant and/or recirculating flow downstream of the thrombus. The volume of stagnant flow and the peak wall shear stress increased with thrombus volume. For spherical thrombi trapped upstream, disruption of flow was observed along the cava wall ipsilateral to the thrombus and within the filter. Peak wall shear stress was greatest with conical thrombi, less with spherical thrombi, and least with ellipsoidal thrombi. ConclusionsThe authors have designed a computer model to study the hemodynamics of the TrapEase filter with various thrombi and trapping positions. The model offers advantages over in vitro techniques, specifically improved resolution and easy adaptation for new filter designs, thrombus morphologies and/or sizes, and flow parameters. The results agree with those of previous bench experiments that suggest the upstream trapping position of the TrapEase filter leads to a potentially thrombogenic region of stagnant and/or recirculating flow with low shear stress. These findings are supported by clinical studies showing an increased incidence of occlusive and/or nonocclusive thrombus within the TrapEase filter and the retrievable, nearly structurally identical, OptEase filter. INFERIOR vena cava (IVC) filters have played a crucial role in the prevention of pulmonary embolism from deep venous thrombosis for more than 30 years. In the United States alone, it is estimated that more than 100,000 filters are placed each year (1). During the past several years, multiple generations of filters have flooded the market, offering retrievability, lower-profile delivery systems, and ease of deployment. Despite these changes and advances, successful clinical IVC filtration still hinges on a filter design that traps emboli without being inherently thrombogenic. The PREPIC Study Group (2) noted an increase in recurrent deep venous thrombosis 2 years after placement of a permanent IVC filter and suggested that this observation may be related to thrombosis at the filter site. A relationship between filters and deep venous thrombosis is unclear, but there are two possible reasons why filters could cause deep venous thrombosis. First, filters may cause progressive damage to the vein wall with secondary IVC occlusion at the filter or narrowing and eventual stenosis that contributes to deep venous thrombosis at or below the filter. Second, IVC filters may induce a local and inherently prothrombotic state either by their design or after trapping emboli. Our study is designed to analyze the latter relationship. The TrapEase IVC filter (Cordis, Miami Lakes, Florida) is a low-profile, symmetric, 6-F permanent nitinol filter. Several clinical studies have evaluated the safety and efficacy of the TrapEase filter, with more than 1,000 patients evaluated in these reports (3, 4, 5, 6). One study has raised concern over a possible increase in IVC thrombosis with this filter design (6). The TrapEase design incorporates a “dual filtration” system with two trapping positions: one upstream, which traps emboli along the wall of the IVC, and one downstream, which traps emboli in the central cone (Fig 1). Recent investigators have evaluated the hemodynamics of the TrapEase filter in an in vitro setting by using the photochromic flow visualization technique. This technique uses a laser and photochromic dye in solution to evaluate flow dynamics in a laboratory setting. The findings of those studies suggest that trapping emboli in the upstream position may lead to a prothrombotic state (7, 8, 9). In this study, we have constructed computer models of the hemodynamics of the TrapEase IVC filter both free of thrombus and with various thrombus volumes, shapes, and trapping positions. With use of these models, we have computed the wall shear stress and velocity contours under various conditions and identified regions of flow stagnation and recirculation. This information may provide valuable insight into the relationship between filter design and the potential for filter-induced deep venous thrombosis. Materials and Methods  Three-dimensional computer models of the IVC, TrapEase filter, and simulated thrombi were constructed to study the flow dynamics. Flow around the unoccluded and partially occluded filter was simulated by using three-dimensional computational fluid dynamics. All computations were performed on one processor of a supercomputer at the Lawrence Livermore National Laboratory. IVC Model A schematic diagram of the flow configuration is shown in Figure 1. The IVC was modeled as a rigid, straight pipe. The diameter of the pipe was 23 mm, per the average IVC diameter described by Kaufman et al (10). The length of the pipe was 90 mm; this length provided adequate distance for the parabolic inflow velocity profile to develop fully before reaching the filter, that is, Poiseuille flow. The sides of the pipe and the solid surface of the filter were no-slip boundaries (ie, zero velocity). Filter Model Computer-aided design was used to construct a geometrically accurate model of the TrapEase filter. First, the filter was inserted into a 23-mm-inner-diameter glass test tube, which compressed the filter as though it were in an IVC of the same diameter. Then, high-resolution photographs were taken of the filter by using a Dimage Xt digital camera (Minolta, Osaka, Japan). Measurements of the filter were also obtained with a Cen-Tech 6-inch digital caliper (Harbor Freight Tools, Camarillo, California). The photographs and measurement data were then imported into the GNU Image Manipulation Program (GIMP) (available at http://www.gimp.org), where spatial geometry of the filter was extracted on the basis of pixel color and location. The geometric specifications were then imported into the Overture software framework (Lawrence Livermore National Laboratory, Livermore, California, available for download at http://computation.llnl.gov/casc/Overture) (11, 12), where a computer-aided design model was constructed. For ease in modeling, the extraluminal barbs on the filter were excluded from the model because their presence does not alter the characteristics of the flow. Simulations The flow was modeled as an incompressible (constant density), Newtonian fluid whose motion is described mathematically by the Navier-Stokes equations (13). Other computer studies of IVC flow have assumed a similar flow model (14). The Navier-Stokes equations were solved by using the incompressible flow solver within the Overture software framework. Overture uses the method of overset grids to discretize the equations by using a collection of curvilinear meshes and then solves the equations by using finite difference approximations (15). The spatial resolution of the solution was 2 mm or less. The steady-state solution (from Overture) was obtained by using a pressure-based equation solver (16, 17). Postprocessing was performed by using tools provided by Overture, custom scripts written in Matlab (MathWorks, Natick, Massachusetts), and the GIMP. Simulations of the partially occluded filter were conducted with 0.25-, 0.50-, 1-, 1.5-, 1.875-, and 2.0-mL spherical thrombi in the downstream trapping position and 0.25- and 0.50-mL spherical thrombi in the upstream trapping position (see Fig 1). Conical and ellipsoidal clots of 1.875 mL were also placed in the downstream trapping position. Thrombus volumes were selected on the basis of a visual scale described by Wang et al (18). The thrombi were incorporated into the simulations by using Overture's computer-aided design capabilities. The mean inlet velocity was 6.45 cm/sec, which corresponds to a flow rate of 1.6 L/min in the 23-mm cava and a Reynolds number (Re) of 600 (Re = ρU D/μ, where ρ is the density of blood [1,040 kg/m3], U the mean inlet velocity, D the diameter of the vena cava [23 mm], and μ the viscosity of blood [2.57e-3 kg/msec]). A setting of Re = 600 has been used in previously published reports involving vena cava experiments (7, 19, 20). Results The three-dimensional Navier-Stokes equations were solved for the velocity (three components) and the pressure. Contour plots of the axial velocity (ie, along the axis of the pipe) were normalized by the corresponding theoretical value for fully developed pipe flow (ie, Poiseuille flow). Transverse velocities (ie, perpendicular to the axis of the pipe) were small in comparison to the axial velocity and are not discussed. The wall shear stress, which is the product of the viscosity and velocity gradient at the wall, was also normalized by the theoretical value for pipe flow. In all figures, flow is left to right. Unoccluded Filter The velocity contours and wall shear stress for the unoccluded filter are shown in Figure 2. A peak normalized velocity of 2.05 occurred upstream of the filter, and flow inside the filter had a normalized velocity in the range 0.95–1.85. Furthermore, the velocities at the cava wall and the surface of the filter were zero, as required by the no-slip boundary condition. Disruption of the flow due to the filter was minimal, except immediately downstream of both filter tips, which have hollow cores that allow some fluid to pass through. Normalized peak wall shear stresses of 2.15 and 2.1 were observed near the downstream and upstream “hips” of the filter, respectively. Both stresses are within 15% of those reported in photochromic experiments by Leask et al (7). Between the upstream and downstream hips, the wall shear stress relaxes toward that of Poiseuille flow as the flow inside the filter increased momentum and decreased the velocity gradient near the wall. Partially Occluded Filter Downstream thrombi Velocity contours and wall shear stresses for the partially occluded filter with 0.50-, 1-, and 1.875-mL spherical thrombi are shown in Figure 3, Figure 4, Figure 5, Figure 6. Simulations were also performed with thrombus volumes of 0.25, 1.5, and 2 mL, but contour plots are not shown. The results demonstrate, as anticipated, that larger thrombi generate higher peak flow velocities. In all cases, the peak velocity occurred in the narrow passage between the thrombus and the wall of the cava. The increase in peak velocity was due to the reduction in cross-sectional area through which the flow could pass. Larger thrombi also produced larger regions of low velocity and recirculating flow downstream of the thrombus, and all thrombi disrupted the flow inside the filter. The peak wall shear stress increased steadily with thrombus volume and always occurred downstream of the peak velocity. This was because flow around the clot forced fluid toward the wall of the cava, thereby producing a steep velocity gradient due to mass conservation. An increase in wall shear stress was also observed near the upstream hips of the filter, but the peak value was nearly independent of thrombus volume. Figure 5 compares profiles of wall shear stresses for all spherical thrombus volumes (including those for which velocity contours are not shown); the peak stress moved upstream in proportion to the increase in thrombus volume. Figure 6 compares velocity contours and wall shear stresses for 1.875-mL spherical, conical, and ellipsoidal thrombi. As demonstrated, the spherical and conical thrombi produced peak velocities of 2.5, whereas the ellipsoidal thrombus had a peak velocity of 2.12. The ellipsoidal thrombus, due to its streamlined shape, disrupted the flow the least of the three shapes as exhibited by relatively small and well-confined regions of disturbance both upstream and downstream of the thrombus. The wall shear stresses for all three shapes were qualitatively similar and are compared in Figure 6. The peak stress for the ellipsoidal thrombus was 4.8, whereas the spherical and conical thrombi produced peak stresses of almost 7.5 and 8.5, respectively. The location of the peak stresses was similar for all three thrombi. Upstream thrombi Figure 3 shows velocity contours and wall shear stresses for a 0.50-mL spherical thrombus in the upstream trapping position (similar results, not shown, were observed for a 0.25-mL thrombus). Regions of recirculating flow were observed near the cava wall ipsilateral to the thrombus, but velocity contours near the downstream tip were similar to those of the unoccluded filter. The shear stresses along the walls opposite and ipsilateral to the thrombus are also shown in Figure 3. Upstream of the thrombus, the shear stress decreased to zero as the thickness of the boundary layer increased, and a region of flow reversal was observed that resulted in a point of zero wall shear stress upstream of the thrombus. The shear stress immediately upstream and downstream of the thrombus exhibited an oscillatory pattern due to the thrombus forcing flow away from the ipsilateral cava wall, leading to a highly three-dimensional and complex flow structure. Downstream of the thrombus, the normalized wall shear stress remained less than 0.75 until the downstream hips were encountered and the flow regained momentum along the ipsilateral wall. A corresponding increase in wall shear stress was observed, although the stress remained lower than that observed on the opposite wall. Discussion The TrapEase filter has two trapping positions, an upstream position that traps thrombus between the filter and the cava wall, and a downstream position that traps thrombus centrally and away from the wall (Fig 1). For all thrombi considered, the partially occluded filter develops low-velocity flow immediately downstream of all thrombi and areas of recirculation are present. In the case of a 0.5-mL spherical thrombus (Fig 3), disruption to the flow is observed approximately 2 cm downstream of the thrombus for both trapping positions. Furthermore, for the downstream position, the low-velocity flow is located in the central portion of the cava away from the walls; for the upstream trapping position, the low-velocity flow is near the cava wall. The shear stress along the wall opposite the thrombus was significantly higher for the downstream thrombus than for the upstream thrombus. The downstream findings described earlier were observed for all three thrombus shapes: sphere, cone, and ellipsoid. For the ellipsoid, however, flow disruption was reduced due to its streamlined shape. The differences in shear stress and locations of low-velocity and/or recirculating flow between the two trapping positions may be clinically important. High shear stresses can activate platelets near the vessel wall (though typically at arterial shear stress levels, not venous shear stress levels seen in the IVC) (21), resulting in platelet adhesion and primary hemostasis; however, high shear stresses can enhance the removal of thrombin and fibrin, thereby reducing the likelihood of secondary hemostasis (22). High levels of shear stress may also stimulate endothelial cells that secrete tissue plasminogen activator, reducing the risk of hemostasis (21). Areas of recirculating blood flow under low shear stresses can lead to venous thomboembolism (22). In our model, with thrombus in the upstream trapping position, low shear stress along the cava wall may allow enough stagnation that thrombin and fibrin could accumulate, and secondary hemostasis may be initiated. Conversely, with high shear stress surrounding the thrombus in the downstream, central trapping position, secondary hemostasis would be more difficult to initiate as thrombin and fibrin are cleared and mechanical lysis can occur. Leask et al (7) performed an in vitro evaluation of the hemodynamic effects of clot entrapment of the TrapEase IVC filter by using a photochromic technique. Our velocity contour maps and shear stress calculations for the unoccluded filter are similar, thereby corroborating our methods. Our computer model offers several advantages over the photochromic technique and/or in vitro testing in general: it has improved spatial resolution (0.2 mm vs 0.5 mm); it can evaluate different sizes, shapes, locations, and configurations of thrombi without rerunning expensive and time-consuming bench experiments; flow characteristics such as inflow velocity and viscosity of blood are easily modified; quantities that are difficult to evaluate experimentally are accurately computed; and our computer model's platform is readily adaptable to evaluate other filter designs without physically constructing the filter and performing time- and/or labor-intensive laboratory experiments. Our computer model is founded on simplifying yet realistic assumptions. The model assumes that flow through the IVC is steady. Compared to the aorta, IVC flow has less pulsatility due to a damping of the pulse pressure by the time blood reaches the venous system (23). The wall of the cava is modeled as smooth and rigid, and inflow from renal, lumbar, and other venous tributaries is excluded. In vivo, the vena cava is dynamic, maintaining a complex topology specific to each patient. Our simplification allows direct model validation with Leask et al (7) and provides a foundation on which to construct more detailed computer models. The simulated spherical, elliptical, and conical thrombi are solid; in vivo clots assume random shapes with variable elasticity and porosity. As noted by Swaminathan et al (14), however, spherical thrombi represent, in some sense, a statistical average of irregular shapes. Finally, our model assumes that the flow is laminar and blood is Newtonian: viscosity is a function of the local shear rate. Although blood is generally thought to be non-Newtonian, Swaminathan et al (14) found that non-Newtonian effects for flow in the IVC are minimal. Our modeling infrastructure is capable of addressing the issues described above by incorporating additional realism based on in vitro or in vivo data (eg, flow velocities, computed tomographic [CT] scans). A preliminary clinical study of the TrapEase IVC filter raised questions of an increased risk of cava thrombosis (24). In that study (n = 189), three patients were noted to have IVC thrombosis (1.5%). A current review of the literature shows four clinical studies of the TrapEase IVC filter with a total of 1,047 patients (24, 25, 26, 27). Taken as a whole, eight patients (0.76%) demonstrated total or near-total IVC occlusion. Among all IVC filters, reported ranges of filter-related IVC occlusion vary widely from 0% to 28% (28); moreover, 0.76% is considerably less than the average reported rates of cava occlusion for many other IVC filters (29). However, it should be noted that imaging follow-up of permanent IVC filters is often lacking and that more reliable follow-up imaging has been seen with retrievable IVC filters. The OptEase filter (Cordis) is a retrievable version of the TrapEase filter and is structurally identical except that it has unidirectional barbs (as opposed to bidirectional barbs in the TrapEase) and a tiny hook at the upstream end of the filter. In a multi-institutional study of 446 patients, Karmy-Jones et al (30) found a statistically significant (P < .05) increased rate of caval thrombosis with the OptEase filter compared to the Günther Tulip filter (Cook, Bloomington, Indiana) and the Bard Recovery filter (C.R. Bard, Tempe, Arizona). In that study, the rates of caval occlusion were 11% for the OptEase, 1% for the Recovery, and 0% for the Günther Tulip. Beyond caval occlusion, there is an increased rate of intrafilter, nonocclusive thrombus with the TrapEase and OptEase filters. The largest clinical study of the TrapEase filter reported a total of 751 patients, 270 of whom underwent follow-up abdominal CT (25). Within this group of 270 patients, 68 (25.2%) had nonocclusive thrombus noted within the filter. Three recent publications on the OptEase filter have reported rates of nonocclusive thrombus in the 26%–46% range (30, 31, 32). In contrast, two recent reports on the Günther Tulip retrievable filter show nonocclusive intrafilter thrombus rates of 12%–19% (33, 34). Reporting standards of nonocclusive thrombus are not uniform, which limits study comparisons; however, a single-center, retrospective case control comparison of the TrapEase and Günther Tulip filters has been published (31). In that study, patients were treated with either the Günther Tulip filter (n = 92) or the OptEase filter (n = 80). Rates of nonocclusive intrafilter thrombus 4 weeks after placement were 21% for the Günther Tulip filter and 39% for the OptEase filter. In particular, smaller thrombi (<25% of the IVC diameter) were seen in 31% of OptEase filters and only 8% of Günther tulip filters (P = .002) (31). This difference could be attributed to an increased level of trapping of thrombi by the OptEase filter, but rates of symptomatic pulmonary emboli with both filter groups were similarly low (1.2%) (31). Thus, an inherent thrombogenic effect of the TrapEase and OptEase filters may exist. The increased clinical rates of occlusive and nonocclusive intrafilter thrombus with the TrapEase and OptEase filters lend correlational support to our experimental finding: that the upstream trapping position results in recirculating and/or stagnant flow with low shear stress that may promote intrafilter thrombosis. Acknowledgments This work was performed under the auspices of the U.S. Department of Energy (DOE) by Lawrence Livermore National Laboratory (LLNL) in part under Contract W-7405-Eng-48 and in part under Contract DE-AC52-07NA27344 and by DOE contracts from the ASCR Applied Math Program. Computer time on LLNL's Yana cluster was provided under Livermore Computing's Multiprogrammatic & Institutional Computing Initiative. LLNL is operated by Lawrence Livermore National Security, LLC, for the DOE, National Nuclear Security Administration under Contract DE-AC52-07NA27344. References 1. 1Lee P, Raizada A, Ciocca R. Growing utilization of IVC filter placement from 2001-2005: analysis of NIS (Abstract for the Society for Clinical Vascular Surgery 2009 Meeting). http://scvs.vascularweb.org/SCVS_Contribution_Pages/Abstracts/2009/MP26.html. 2. 2The PREPIC Study Group. Eight-year follow-up of patients with permanent vena cava filters in the prevention of pulmonary embolism. Circulation. 2005;112:416–422.
CrossRef
3. 3Rousseau H, Perreault P, Otal P, et al. The 6-F-nitino TrapEase inferior vena cava filter: results of a prospective multicenter trial. J Vasc Intervent Radiol. 2001;12:299–304. 4. 4Liu WC, Do YS, Choo SW, et al. The mid-term efficacy and safety of a permanent nitinol IVC filter (TrapEase). Korean J Radiol. 2005;6:110–116. MEDLINE |
CrossRef
5. 5Kalva SP, Wicky S, Waltman AC, et al. TrapEase vena cava filter: experience in 751 patients. J Endovasc Ther. 2006;13:365–372. MEDLINE |
CrossRef
6. 6Schutzer R, Ascher E, Hingorani A, et al. Preliminary results of the new 6F TrapEase inferior vena cava filter. Ann Vasc Surg. 2003;17:103–106. Abstract | Full Text |
Full-Text PDF (486 KB)
|
CrossRef
7. 7Leask RL, Johnston KW, Ohja M. Hemodynamic effects of clot entrapment in the TrapEase inferior vena cava filter. J Vasc Intervent Radiol. 2004;15:485–490. 8. 8Harlal A, Ohja M, Johnston KW. Vena cava filter performance based on hemodynamics and reported thrombosis and pulmonary embolism patterns. J Vasc Intervent Radiol. 2007;18:103–115. 9. 9Couch GG, Johnston KW, Ojha M. Full-field flow visualization and velocity measurement with a photochromic grid method. Meas Sci Technol. 1996;7:1238–1246. 10. 10Kaufman JA, Waltman AC, Rivitz SM, Geller SC. Anatomical observations on the renal veins and inferior vena cava at magnetic resonance angiography. Cardiovasc Intervent Radiol. 1995;18:153–157. MEDLINE 11. 11Brown DL, Henshaw WD, Quinlan DJ. An object oriented framework for solving partial differential equations. In: Scientific Computing in Object-Oriented Parallel Environments, Springer Lecture Notes in Computer Science. 1343:1997;p. 177–194. 12. 12Henshaw WD, Schwendeman DW. Moving overlapping grids with adaptive mesh refinement for high-speed reactive and non-reacting flow. J Comp Phys. 2006;216:744–779. 13. 13Panton RL. Incompressible flow. In: Incompressible flow. 2nd ed.. New York, NY: John Wiley & Sons; 1996;p. 228–253. 14. 14Swaminathan TN, Hu HH, Patel AA. Numerical analysis of the hemodynamics and embolus capture of a Greenfield vena cava filter. J Biomed Eng. 2006;128:360–370. 15. 15Chesshire GS, Henshaw WD. Composite overlapping meshes for the solution of partial differential equations. J Comp Phys. 1990;90:1–64. 16. 16Henshaw WD. A fourth-order accurate method for the incompressible Navier-Stokes equations on overlapping grids. J Comp Phys. 1994;113:13–25. 17. 17Henshaw WD, Petersson NA. A split-step scheme for the incompressible Navier-Stokes equations. In: Hafez MM editors. Numerical simulation of incompressible flows (World Scientific). 2003;p. 108–125. 18. 18Wang SL, Timmermans HA, Kaufman JA. Estimation of trapped thrombus volumes in retrievable inferior vena cava filters: a visual scale. J Vasc Interv Radiol. 2007;18:273–276. Abstract | Full Text |
Full-Text PDF (829 KB)
|
CrossRef
19. 19Leask RL, Johnston KW, Ojha M. In vitro hemodynamic evaluation of a Simon nitinol vena cava filter: possible explanation of IVC occlusion. J Vasc Interv Radiol. 2001;12:613–618. Abstract | Full Text |
Full-Text PDF (2081 KB)
|
CrossRef
20. 20Harlal A, Ojha M, Johnston KW. Vena cava filter performance based on hemodynamics and reported thrombosis and pulmonary embolism patterns. J Vasc Interv Radiol. 2007;18:103–115. Abstract | Full Text |
Full-Text PDF (3066 KB)
|
CrossRef
21. 21Kroll MH, Hellums JD, Mcintire LV, Schafer AI, Moake JL. Platelets and shear stress. Blood. 1996;88:1525–1541. MEDLINE 22. 22Lowe G. Virchow's triad revisited: abnormal flow. Pathophysiol Haemost Thromb. 2003/2004;33:455–457. MEDLINE |
CrossRef
23. 23Cheng CP, Herfkens RJ, Taylor CA. Inferior vena cava hemodynamics quantified in vivo at rest and during cyclic exercises using magnetic resonance imaging. Am J Physiol Heart Circ Physiol. 2003;284:1161–1167. 24. 24Schutzer R, Ascher E, Hingorani A, Jacob T, Kallakuri S. Preliminary results of the new 6F Trapease inferior vena cava filter. Ann Vasc Surg. 2003;17:103–106. Abstract | Full Text |
Full-Text PDF (486 KB)
|
CrossRef
25. 25Kalva SP, Wicky SW, Waltman AC, Athanasoulis CA. TrapEase vena cava filter: experience in 751 patients. J Endovasc Ther. 2006;13:365–372. MEDLINE |
CrossRef
26. 26Liu WC, Do YS, Choo SW, et al. The mid-term efficacy and safety of a permanent nitinol IVC filter (TrapEase). Korean J Radiol. 2005;6:110–116. MEDLINE |
CrossRef
27. 27Rousseau H, Perreault P, Otal P, et al. The 6-F nitinol TrapEase inferior vena cava filter: results of a prospective multicenter trial. J Vasc Interv Radiol. 2001;12:299–304. Abstract | Full Text |
Full-Text PDF (266 KB)
|
CrossRef
28. 28Streiff MB. Vena caval filters: a comprehensive review. Blood. 2000;95:3669–3677. MEDLINE 29. 29Kinney TB. Update on inferior vena cava filters. J Vasc Interv Radiol. 2003;14:425–440. Abstract | Full Text |
Full-Text PDF (2280 KB)
30. 30Karmy-Jones R, Jurkovich GJ, Velmahos GC, et al. Practice patterns and outcomes of retrievable vena cava filters in trauma patients: an AAST multicenter study. J Trauma. 2007;62:17–25. MEDLINE 31. 31Keller IS, Meier C, Piffner R, et al. Clinical comparison of two optional vena cava filters. J Vasc Interv Radiol. 2007;18:505–511. Abstract | Full Text |
Full-Text PDF (93 KB)
|
CrossRef
32. 32Oliva VL, Szatmari F, Giroux M, et al. The Jonas Study: evaluation of the retrievability of the Cordis OptEase inferior vana cava filter. J Vasc Interv Radiol. 2005;16:1439–1445. Abstract | Full Text |
Full-Text PDF (744 KB)
33. 33Hoppe H, Nutting CW, Smouse HR, et al. Gunther Tulip filter retrievability multicenter study including CT follow-up: final report. J Vasc Interv Radiol. 2006;17:1017–1023. Abstract | Full Text |
Full-Text PDF (871 KB)
34. 34Wicky S, Doenz F, Meuwly J, et al. Clinical experience with retrievable Gunther Tulip vena cava filters. J Endovasc Ther. 2003;10:994–1000. MEDLINE |
CrossRef
a Center of Applied Scientific Computing, Lawrence Livermore National Laboratory, Livermore, California b Division of Vascular and Interventional Radiology, Kaiser Permanente Santa Clara Medical Center, 700 Lawrence Expressway, Santa Clara, CA 95051  Address correspondence to S.L.W.
From the 2009 SIR annual meeting. None of the authors have identified a conflict of interest. PII: S1051-0443(09)00201-2 doi:10.1016/j.jvir.2009.02.015 © 2009 SIR. Published by Elsevier Inc. All rights reserved. | |
|
FULL TEXT
