Excimer Laser–Assisted Removal of Embedded Inferior Vena Cava Filters
A Single-Center Prospective Study
Background—Although chronically implanted inferior vena cava filters may result in filter-related morbidity, there is currently no routine option for removing such filters when they become firmly embedded along the vena cava endothelium.
Methods and Results—During a 3-year period, 100 consecutive patients were prospectively enrolled in a single-center study. There were 42 men and 58 women (mean age, 46 years; limits, 18–76 years). Retrieval indications included filter-related acute inferior vena cava thrombosis, chronic inferior vena cava occlusion, and pain from retroperitoneal or bowel penetration. Filter retrieval was also performed to prevent risks from prolonged implantation and to potentially eliminate the need for lifelong anticoagulation. After standard methods failed, photothermal tissue ablation was attempted with a laser sheath powered by a 308-nm xenon chloride excimer laser. Applied forces were recorded with a digital tension meter before and during laser activation. Laser-assisted retrieval was successful in 98.0% (95% confidence interval [CI], 93.0%–99.8%) with mean implantation of 855 days (limits, 37–6663 days; >18 years). The following filter types were encountered in this study: Günther-Tulip (n=34), Celect (n=12), Option (n=17), Optease (n=20, 1 failure), TrapEase (n=6, 1 failure), Simon-Nitinol (n=1), 12F Stainless Steel Greenfield (n=4), and Titanium Greenfield (n=6). The average force during failed standard retrievals was 7.2 versus 4.6 pounds during laser-assisted retrievals (P<0.0001). The major complication rate was 3.0% (95% CI, 0.6%–8.5%), the minor complication rate was 7.0% (95% CI, 0.3%–13.9%), and there were 4 adverse events (2 coagulopathic hemorrhages, 1 renal infarction, and 1 cholecystitis; 4.0%; 95% CI, 1.1%–9.9%) at mean follow-up of 500 days (limits, 84–1079 days). Scar tissue ablation was histologically confirmed in 96.0% (95% CI, 89.9%–98.9%). Successful retrieval allowed cessation of anticoagulation in 30 of 30 (100%) patients and alleviated morbidity in 23 of 24 patients (96%).
Conclusions—Excimer laser–assisted removal is effective in removing embedded inferior vena cava filters refractory to standard retrieval and high force. This method can be safely used to prevent and alleviate filter-related morbidity.
WHAT IS KNOWN
Prolonged inferior vena cava filter implantation may result in filter-related morbidity and thrombotic complications; therefore, the US Food and Drug Administration has urged all physicians caring for patients with inferior vena cava filters to consider removing the filter as soon as protection from pulmonary embolism is no longer needed.
Over time, many retrievable-type inferior vena cava filters become embedded and are refractory to standard retrieval methods, and patients with implanted permanent-type filters currently have no routine options for filter removal.
WHAT THE STUDY ADDS
The study supports a new indication for excimer laser use in the venous system to remove embedded vena cava filters, regardless of the dwell time.
With appropriate expertise, laser-assisted retrieval may be used to remove a variety of embedded inferior vena cava filters, including permanent-type filters, refractory to standard retrieval and high force.
Laser-assisted filter retrieval may be used to alleviate filter-related morbidity, to prevent further risks associated with long-term implantation, and to eliminate the need for filter-related lifelong anticoagulation.
In the United States, the number of implanted inferior vena cava (IVC) filters has doubled during the past decade,1,2 and >250 000 IVC filters were implanted in the United States alone in 2012.2 Rising filter use has led to an increase in filter-related morbidity and recognition of potential complications from indwelling filters.3 As a result, the US Food and Drug Administration (FDA) has issued a safety communication2 alerting all physicians caring for patients with IVC filters to consider removing the filter as soon as protection from pulmonary embolism is no longer needed. Despite this heightened awareness, it is estimated that as many as 40% to 60% of retrievable-type filters cannot be removed using standard methods alone because they have become firmly embedded.1,4 In addition, many patients have undergone placement of a permanent-type filter designed to permanently embed into the vessel wall. Although these patients may also develop filter-related morbidity after chronic implantation, there is currently no routine option for removing permanent-type filters. The purpose of this study was to evaluate the safety and efficacy of photothermal tissue ablation with an excimer laser sheath technique to facilitate removal of both retrievable-type and permanent-type filters that have become embedded within the vena cava endothelium. We present results from our single-center prospective experience on endovascular laser-assisted filter retrieval.
Editorial see p 498
During a 3-year period, 100 consecutive patients undergoing attempted IVC filter retrieval using an endovascular laser-assisted sheath technique, after failure of standard methods either outside or within our hospital, were prospectively enrolled in an institutional review board–approved study funded by our department, and no such patients were excluded from the study. All data were collected using case report forms, and enrollment was capped at 100 patients for this analysis. The study end points were defined as follows: retrieval success versus failure, immediate procedure-related complications, long-term complications at clinical follow-up, resolution of symptoms in patients with filter-related morbidity, and further need for prescribed filter-related lifelong anticoagulation.
Ninety-two of 100 patients (92%) were referred or self-referred from other institutions. There were 42 men and 58 women (mean age, 46 years; limits, 18–76 years). In all patients, IVC filtration was no longer clinically required, and the specific indications for filter removal were classified into 3 categories: (1) symptomatic patients experiencing filter-related morbidity, (2) asymptomatic patients wishing to avoid filter-related complications and risks associated with prolonged implantation, and (3) anticoagulated patients among groups 1 and 2, with no underlying thrombophilia, to eliminate the potential need for ongoing filter-related anticoagulation, which had been prescribed by a hematologist to reduce thrombotic risks associated with long-term filter implantation.5 All patients were evaluated and counseled before the procedure in either the Interventional Radiology Clinic or the inpatient ward where each patient underwent an extensive informed consent process explaining the risks and benefits of undergoing alternative retrieval maneuvers, including the use of laser tissue ablation versus the risks and benefits of permanent filter implantation. Before retrieval attempts, acute or recurrent deep venous thrombosis was excluded in each patient using lower extremity Doppler ultrasound of the deep veins or computed tomographic venography or conventional venography of the iliofemoral segments. In 3 patients, acute IVC thrombus was identified and removed using overnight catheter-directed infusion of alteplase (tissue-type plasminogen activator; Genentech, South San Francisco, CA) at 0.5 to 1.0 mg tissue-type plasminogen activator/h before attempted filter removal.
All procedures were performed percutaneously under moderate sedation except in 4 patients who either refused moderate sedation because of anxiety or required general anesthesia because of severe pain from filter component penetration, identified during previous retrieval attempts at outside institutions. Patients with radiographic evidence of bowel penetration were given 1 g IV cefotetan before the procedure for bowel prophylaxis. All patients received intraprocedural therapeutic anticoagulation to minimize thrombotic risk per prior protocol.6 For patients already receiving therapeutic anticoagulation, the following regimens were continued during the procedure: enoxaparin (1 mg/kg SC, BID), warfarin (target international normalized ratio, 2.0–3.0 or 2.5–3.5), or fondaparinux (7.5 mg SC, QD for a 50- to 100-kg adult). For patients not receiving prior anticoagulation, 50 U/kg of unfractionated heparin was administered intravenously during the procedure. In patients with tilted, tip-embedded filters, the devices were first untilted and the filter apex engaged using described methods.7,8 After capturing the filter tip, attempts to sheath the filter were made, and the diagnosis of an embedded filter was confirmed when any portion of the filter was refractory to standard sheathing and substantial force as determined by the operator or as objectively measured using a digital force gauge (McMaster-Carr, Santa Fe Springs, CA). After the first 15 patients, routine use of the force gauge was added to all 85 subsequent cases. Failure of standard retrieval was then defined based on preliminary data9 as an inability to remove the filter, despite applying ≥6 pounds of tension to the filter. The force gauge was also used to prevent overexertion (>8–9 pounds) and potential device damage based on early experience.9 After failure of standard retrieval methods, a laser sheath (Spectranetics, Colorado Springs, CO) was calibrated to 60 mJ/mm2 and inserted to the point of resistance along the embedded filter components. Before laser activation, this point of resistance was confirmed again while applying tension against the inactive laser sheath to reconfirm the presence of significant scar tissue. Treatment escalation was then attempted by activating the laser-tipped sheath powered by a 308-nm xenon chloride excimer laser generator (CVX-300, Spectranetics) to perform photothermal tissue ablation per prior protocol9 (Figures 1 and 2). During laser activation, the applied tension was recorded while advancing the laser sheath through scar tissue around the filter.
A postprocedure venogram was performed in all patients. If filter retrieval was successful, subsequent IVC recanalization, angioplasty, and stent placement were performed if necessary to restore venous outflow in patients with symptomatic caval occlusion, and stent-graft placement was performed using previously described methods10–12 if significant caval injury was identified. During preprocedure preparation, large-diameter stent grafts (Gore, Flagstaff, AZ) and an occlusion balloon (AGA Medical, Plymouth, MN) were made readily available in the procedure room. Major and minor procedural complications were defined according to the clinical practice guidelines of the Society of Interventional Radiology.13 All retrieved filter specimens were inspected for device integrity. The filters were then submitted to surgical pathology for histological evaluation of adherent tissue harvested from the filter attachment sites, and elastic van Gieson staining was also performed to assess the presence of native vessel wall tissue. After retrieval attempts, therapeutic anticoagulation was continued in those with an underlying thrombophilia (factor V Leiden, prothrombin gene mutation, protein C or S deficiency) or a retained filter from retrieval failure. If retrieval was successful, an attempt was made by a hematologist to discontinue anticoagulation in patients with no thrombophilia who had been prescribed filter-related lifelong anticoagulation. All patients underwent routine clinical follow-up to assess the resolution of any procedure-related complications, to assess the improvement in any filter-related morbidity (pain, lower extremity edema, exercise intolerance), and to evaluate for signs of new venous thromboembolism. If underlying caval stenosis or postprocedure caval injury was identified requiring bare-stent or stent-graft placement, follow-up cross-sectional imaging with computed tomography or MRI was also obtained within 6 months to confirm lesion resolution and stent patency. Differences in applied forces between failed standard retrieval attempts and laser-assisted retrieval attempts were analyzed with a paired Wilcoxon test in 85 consecutive patients. A P<0.05 was considered to represent statistical significance. Confidence intervals (CIs) were computed using the exact binomial method, and all statistics were calculated using Stata Release version 11.2 software (StataCorp LP, College Station, TX).
The indications for filter retrieval among the 3 groups were as follows: (1) 24 symptomatic patients (24%) experiencing filter-related morbidity, (2) 76 asymptomatic patients (76%) wishing to avoid filter-related complications and risks associated with prolonged implantation, and (3) 30 anticoagulated patients (30%) among groups 1 and 2, with no underlying thrombophilia, to eliminate the potential need for prescribed filter-related lifelong anticoagulation. Among group 1, the following filter-related complications were observed before attempted filter removal: recurrent acute caval thrombosis (n=6; 2 complicated by acute pulmonary embolism, 2 with chronic IVC stenosis at the filter implantation site), chronic IVC occlusion with lower extremity edema and exercise intolerance (n=9; 1 complicated by caval rupture from venous pressure elevation, requiring prior open surgical repair), abdominal pain from intestinal penetration (n=4; 1 with concomitant retroperitoneal penetration), and pain from retroperitoneal penetration (n=5; 1 with concomitant filter fracture and central component embolization).
Laser-assisted filter retrieval was successful in 98 of 100 patients (98.0%; 95% CI, 93.0%–99.8%) with mean implantation of 855 days (limits, 37–6663 days; >18 years). The following filter types were encountered in this study: Günther-Tulip (Cook, Bloomington, IN; n=34), Celect (Cook, Bloomington, IN; n=12), Option (Argon Medical, Athens, TX; n=17), Optease (Cordis, Miami Lakes, FL; n=20, 1 failed retrieval), TrapEase (Cordis, Miami Lakes, FL; n=6, 1 failed retrieval), Simon-Nitinol (Bard, Tempe, AZ; n=1), 12F Stainless Steel Greenfield (Boston Scientific, Natick, MA; n=4), and Titanium Greenfield (Boston Scientific, Natick, MA; n=6). Filter types and implantation lengths are summarized in Table 1. All filters failed standard retrieval methods with an average maximum force of 7.2 pounds (limits, 5.0–9.0 pounds) before laser activation, and the average maximum force applied during laser-assisted retrievals was 4.6 pounds (limits, 3.0–8.5 pounds), which was significantly lower (P<0.0001) (please see Appendix in the online-only Data Supplement). Two cases failed laser-assisted filter retrieval: 1 Optease implanted for 188 days and 1 TrapEase implanted for 1124 days. Both retrieval failures were associated with a cylindrical-shaped filter and chronic thrombotic occlusion of the IVC that impeded laser sheath advancement, despite laser activation.
In 4 cases (4%), before routine use of the force gauge, overexertion on the laser sheath resulted in focal cracking of the outer sheath polymer, outside the patient, and this required exchange of the damaged laser sheath for a new one. After initiating routine use of the force gauge to prevent overexertion on the sheath, there were no further device events in the 85 subsequent patients. The major procedure-related complication rate was 3.0% (95% CI, 0.6%–8.5%): 1 patient with an underlying thrombophilia developed acute IVC thrombus after removal of an embedded Optease filter, requiring catheter-directed thrombolysis that was successful in resolving the clot. Two patients developed procedure-related acute caval injury resulting in major hemorrhage requiring stent-graft placement that occurred during the first half of our study: 1 involved a chronically embedded Celect IVC filter (conical shaped) that had been complicated by severe multifocal penetration through the IVC and 1 involved a permanent-type TrapEase filter (cylindrical-shaped) complicated by chronic IVC occlusion that was performed before routine force gauge use. In both cases, the patients experienced immediate tachycardia with hypotension (systolic pressure <90 mm Hg), and focal contrast extravasation along the IVC was observed. In both cases, a readily available 24-mm compliant balloon was quickly inserted and inflated to tamponade the IVC, resulting in hemodynamic stabilization. Both complications were then successfully treated with urgent endovascular stent-graft placement, with no need for open surgery, and there was stent patency with no extravasation on follow-up cross-sectional imaging within 6 months and no complications at long-term clinical follow-up (mean, 638 days; limits, 408–867 days). The minor complication rate was 7.0% (95% CI, 0.3%–13.9%). Three patients developed small (<5 mm) focal hemorrhage that was successfully treated with temporary inflation of an occlusion balloon (n=2) or was self-limited (n=1), and resolution was confirmed by follow-up venography. Four patients developed a small (<2 cm) postprocedure pseudoaneurysm that was self-limited, completely resolved on follow-up cross-sectional imaging, and required no further treatment. Among 4 patients with bowel penetration, there were no clinical signs of postprocedure bowel injury or infection. All procedure-related complications are summarized in Table 2.
After removal, examination of the filter specimens showed no procedure-related filter fractures. Sufficient adherent tissue was identified and harvested from 98 of 100 retrieved specimens, and evidence of photothermal ablation through filter-adherent scar tissue was histologically confirmed on elastic van Gieson stain in 94 of 98 patients (96.0%; 95% CI, 89.9%–98.9%). From this group, the ablation margins consisted of neointimal hyperplasia and fibrosis in 88 of 94 (93.6%; 95% CI, 77.89%–98.11%) cases, with substantial cautery extension into overlying media in the remaining 6 cases (6.4%; 95% CI, 4.11%–16.11%). Among these 6 cases, ablation of vessel media adjacent to scar tissue was noted along points of prior filter component penetration (4 Günther-Tulip) or high surface area contact (2 Optease), but there was no radiographic evidence of procedure-related caval injury. In addition, significant amounts of nonablated caval tissue were identified in 9 of 98 cases (9.2%; 95% CI, 4.3%–16.7%). Among these 9 cases, there was also prior filter component penetration through caval wall from a conical-shaped filter (2 Günther-Tulip, 2 Celect) or high surface area contact from a cylindrical-shaped filter (5 Optease). From this subgroup, 4 of the 9 cases (44%) had the following complications: in 3 patients (2 Optease, 1 Günther-Tulip), a small self-limited focal pseudoaneurysm (<2 cm) and in 1 patient with severe multifocal penetration from a Celect filter, major hemorrhage requiring stent-graft repair as mentioned above. There were 4 adverse events (4.0%; 95% CI, 1.1%–9.9%) at mean clinical follow-up of 500 days (limits, 84–1079 days). Two patients developed coagulopathic retroperitoneal hemorrhage after resuming therapeutic anticoagulation for underlying thrombophilia, and both were successfully managed with temporary cessation of anticoagulation. One patient developed an idiopathic right renal infarct, incidentally discovered after 4 months, which did not require further treatment. One patient developed acute cholecystitis after 1 month, which was successfully treated with cholecystectomy. Among 9 patients with symptomatic chronic IVC occlusion, revascularization with angioplasty alone (n=2) or angioplasty and stent placement (n=7) was successful in alleviating venous obstructive symptoms and exercise intolerance in 8 of 9 patients. One patient experienced persistent symptoms from chronic post-thrombotic syndrome, despite revascularization with angioplasty and stent placement. For patients treated with placement of an IVC stent (n=7) or stent graft (n=2), caval patency and vessel integrity were confirmed in all patients on follow-up cross-sectional imaging within 6 months. Among patients with prior filter-related thrombotic events (n=6), there were no further VTE episodes after successful filter removal. Among patients with prior pain from filter component penetration (n=9), the pain resolved in all patients after filter removal. Overall, filter-related morbidity was alleviated in 23 of 24 patients (96%) with no long-term complications, and there were no recurrent thrombotic events in any patient at mean clinical follow-up of 500 days (limits, 84–1079 days), including 21 patients who remained on therapeutic anticoagulation for an underlying thrombophilia. In addition, 30 patients (30%; 95% CI, 21.2%–40.0%) with no thrombophilia were candidates to have filter-related anticoagulation discontinued, and successful laser-assisted retrieval allowed cessation of anticoagulation in all 30 (100%) of these patients.
In a randomized study,14 patients with acute deep vein thrombosis receiving IVC filters were found to have reduced mortality from acute pulmonary embolism at 12 days after implantation. However, at 8-year follow-up of the same study group, patients with permanent IVC filters experienced an increased risk of chronic deep vein thrombosis with no benefit in mortality from pulmonary embolism versus the nonfilter group.5 Additional thrombotic risks, such as caval occlusion and post-thrombotic syndrome from long-term filter implantation, have been confirmed in recent studies,15,16 and significant nonthrombotic risks, such as filter fracture, component embolization, and caval penetration (with pain and organ injury), have also been reported.1,3,17 Furthermore, many patients with permanent filters are routinely managed with ongoing anticoagulation, solely to reduce thrombotic risks associated with chronic filter implantation, subjecting them to the inconvenience and bleeding risk from lifelong anticoagulation.
An increase in filter use,2 combined with historically low retrieval rates and limited clinical follow-up,1,18 has resulted in the current epidemic of filter-related complications. Indeed, the FDA reported a 5-year increase2 in the number of reported adverse events from prolonged filter implantation, and this prompted release of an FDA Safety Alert in 2010 urging all physicians responsible for the ongoing care of patients with retrievable IVC filters to consider removing the filter as soon as protection from pulmonary embolism is no longer needed.2 However, despite heightened awareness and closer follow-up to remove these devices, it is estimated that up to 40% to 60% of retrievable-type filters cannot be removed using standard methods,1,3,4 especially after 1 year of implantation.3 For instance, if 40% of the 250 000 retrievable-type filters placed annually2 cannot be removed, then 100 000 patients per year could develop an embedded filter in the United States alone. In addition, thousands of additional patients undergo placement of a permanent-type filter each year, and these are never routinely followed and have no standard option for filter retrieval even if complications arise or if ongoing filtration is no longer required.
Experimental use of a laser-assisted sheath technique for removal of embedded filters was described in an animal model19 and recently reported in human studies.9,20 Based on early results, the current study was conducted to further investigate this technique and protocol in a larger patient cohort containing a variety of embedded filter types. Successful laser-assisted retrieval was achieved in 98% of patients refractory to standard retrieval and aggressive force. The use of laser tissue ablation also permitted removal of these embedded devices using a significantly lower average force versus the high forces applied during standard retrieval attempts that failed. Although success was high in this study, retrieval failure was observed in 2 cases involving attempted removal of 1 Optease and 1 TrapEase, and these devices could not be removed despite applying laser energy and an upper force limit of 8.5 pounds. Because both filters were cylindrical-shaped with larger surface area contact along the caval wall, we believe there was substantial scar tissue formation along the filter struts. Both these filters were also associated with chronic caval thrombosis that also hindered sheath advancement. Despite failed retrieval, there were no procedure-related complications in these 2 cases because a force gauge was used to avoid overexertion. We believe the use of a force gauge is important to minimize risk of damage and deformity to the underlying vein, embedded filter, and retrieval apparatus during these procedures. In this study, force measurements also helped to confirm that laser-assisted retrieval was effective in achieving filter removal when high force alone was insufficient or that it reduced the average necessary force required to achieve successful retrieval. Interestingly, all recorded forces in our study exceeded the maximum recommended force of 2 pounds previously described21 for routine filter retrievals, and this further emphasizes the degree of difficulty of the embedded filters we encountered. The force gauge was also used to ensure that we did not exceed an upper limit of safety established earlier,9 and we believe this contributed to a low rate of major complications with no procedure-related filter fractures. Overall, procedure safety was further supported by the histological data that confirmed ablation through filter-adherent scar tissue in 96%, with sparing of underlying native vein in 94%. This finding coincides with prior research on laser tissue interaction using the xenon chloride excimer laser system.22–25 A combination of photothermal ablation with photochemolysis disintegrates tissue into particles ≤5 μm in diameter. Because 308-nm wavelength penetrates only 50 to 100 μm through vascular tissue, precise tissue ablation is achieved immediately adjacent to the laser tip, with minimal risk of collateral tissue damage.26
Although use of an excimer laser sheath to retrieve embedded IVC filters is a novel application not currently approved by the FDA, the method of endovascular laser tissue ablation to free adherent venous pacemaker leads is a well-established technique. Earlier studies such as the Pacemaker Lead Extraction with the Excimer Sheath (PLEXES) trial showed that a laser sheath method significantly improved the efficiency of transvenous lead extraction, particularly through ablation of dense fibrous tissue, reducing the magnitude of necessary force.27 Based on these data, we sought to apply this technology to patients with embedded IVC filters with similar pathophysiology of adherent fibrous tissue formation along a chronic device implant. Indeed, we achieved successful laser-assisted filter removal in 98% of our study group, despite a mean implantation of 855 days or >2.3 years, with an upper dwell time that exceeded 18 years. Successful removal alleviated filter-related morbidity in 23 of 24 patients (96%), and among asymptomatic patients, the procedure served to completely remove the source of potential risk from prolonged implantation. Furthermore, successful filter retrieval allowed cessation of prescribed filter-related anticoagulation in 30 of 30 (100%) patients, sparing them further inconvenience and risks from lifelong anticoagulation.
The current study has important limitations. The results were acquired by applying a set of endovascular skills and expertise from years of experience with complex and laser-assisted filter retrieval that may be specific to our department. Consequently, the overall safety and efficacy of removing embedded filters as described in our study may not translate into similar outcomes when performed at institutions with less experience. Although complication rates were low in this study, major injury still occurred in 2 cases that required urgent stent-graft repair. Therefore, it is imperative that a thorough assessment of the risk-to-benefit ratio be made. Important considerations such as the presence and severity of filter-related morbidity, the filter shape, and the existence of anatomic complications should be carefully weighed. Specifically, the technique should be used with extreme caution or perhaps be contraindicated in cases with severe multifocal filter penetration or when there is high surface area contact between a cylindrical-shaped filter and the vena cava. Another limitation is the off-label use of the laser sheath apparatus that is currently not designed, intended, or FDA-approved for IVC filter removal. Similarly, the permanent-type filters we attempted to retrieve were not designed or FDA-approved for removal. Therefore, additional large-scale studies are warranted to resolve these issues.
In conclusion, this study supports a new indication for excimer laser use in the venous system to remove embedded vena cava filters. In a patient cohort refractory to standard retrieval and high force, endovascular laser-assisted retrieval was overall safe and successful in removing a variety of filter types, regardless of the dwell time. This procedure has the potential to achieve successful IVC filter removal in thousands of patients to alleviate filter-related morbidity, to prevent further risks associated with long-term implantation, and to eliminate the need for filter-related lifelong anticoagulation.
The abstract from this study was presented at the 2013 Society of Interventional Radiology Annual Scientific Meeting on April 17, 2013, in New Orleans, LA.
The online-only Data Supplement is available at http://circinterventions.ahajournals.org/lookup/suppl/doi:10.1161/CIRCINTERVENTIONS.113.000665/-/DC1.
- Received May 6, 2013.
- Accepted August 19, 2013.
- © 2013 American Heart Association, Inc.
- 2.↵U.S. Food and Drug Administration. Removing Retrievable Inferior Vena Cava Filters: Initial Communication. http://www.fda.gov/MedicalDevices/Safety/AlertsandNotices/ucm221676.htm. Accessed May 2013.
- 5.↵The PREPIC Study Group. Eight-year follow-up of patients with permanent vena cava filters in the prevention of pulmonary embolism: the PREPIC (Prevention du Risque d’Embolie Pulmonaire par Interruption Cave) randomized study. Circulation. 2005;112:416–422.
- Kuo WT,
- Odegaard JI,
- Louie JD,
- Sze DY,
- Unver K,
- Kothary N,
- Rosenberg JK,
- Hovsepian DM,
- Hwang GL,
- Hofmann LV
- Castelli P,
- Caronno R,
- Piffaretti G,
- Tozzi M
- Hommes M,
- Kazemier G,
- van Dijk LC,
- Kuipers EJ,
- van Ijsseldijk A,
- Vogels LM,
- Schipper IB
- Decousus H,
- Leizorovicz A,
- Parent F,
- Page Y,
- Tardy B,
- Girard P,
- Laporte S,
- Faivre R,
- Charbonnier B,
- Barral FG,
- Huet Y,
- Simonneau G