- Email: [email protected]
Experience with the Recovery Filter as a Retrievable Inferior Vena Cava Filter
Experience with the Recovery Filter as a Retrievable Inferior Vena Cava Filter William J. Grande, MD, Scott O. Trerotola, MD, Patrick M. Reilly, MD, Timothy W.I. Clark, MD, Michael C. Soulen, MD, Aalpen Patel, MD, Richard D. Shlansky-Goldberg, MD, Catherine M. Tuite, MD, Jeffrey A. Solomon, MD, Jeffrey I. Mondschein, MD, Mary Kate Fitzpatrick, CRNP, and S. William Stavropoulos, MD
PURPOSE: This study evaluates clinical experience with the Recovery filter as a retrievable inferior vena cava (IVC) filter. MATERIALS AND METHODS: One hundred seven Recovery filters were placed in 106 patients with an initial clinical indication for temporary caval filtration. Patients were followed up to assess filter efficacy, complications, eventual need for filter removal, time to retrieval, and ability to remove the filter. RESULTS: The patient cohort consisted of 62 men and 44 women with a mean age of 48 years (range, 18 –90 y). Mean implantation time was 165 days. Indications for filter placement in patients with deep vein thrombosis (DVT) and/or pulmonary embolism (PE) included contraindication to anticoagulation (n ⴝ 33), complications of anticoagulation (n ⴝ 8), poor cardiopulmonary reserve (n ⴝ 6), large clot burden (n ⴝ 3), and PE while receiving anticoagulation (n ⴝ 1). Indications for filter placement in patients without proven PE or DVT included immobility after trauma (n ⴝ 35); recent intracranial hemorrhage, neurosurgery, or brain tumor (n ⴝ 18); and other surgical or invasive procedure (n ⴝ 3). Three patients (2.8%) had symptomatic PE after placement of the Recovery filter. No caval thromboses were detected. No symptomatic filter migrations occurred. Recovery filter removal was attempted in 15 of 106 patients (14%) at a mean of 150 days after placement. The Recovery filter was successfully retrieved in 14 of 15 patients (93%); one removal was unsuccessful at 210 days after placement. Ninety-two filters (87%) currently remain in place. CONCLUSIONS: Although all the filters were placed with the intention of being removed, a large percentage of filters were not retrieved. The Recovery filter was safe and effective in preventing PE when used as a retrievable IVC filter. J Vasc Interv Radiol 2005; 16:1189 –1193 Abbreviations: DVT ⫽ deep vein thrombosis, IVC ⫽ inferior vena cava, PE ⫽ pulmonary embolism
VENOUS thrombosis and pulmonary embolism (PE) are major causes of morbidity and mortality (1,2). Initial
From the Section of Interventional Radiology, Department of Radiology (W.J.G., S.O.T., T.W.I.C., M.C.S., A.P., R.D.S.G., C.M.T., J.A.S., J.I.M., S.W.S.), and Department of Traumatology and Surgical Critical Care (P.M.R., M.K.F.), University of Pennsylvania Medical Center, 3400 Spruce Street, Philadelphia, Pennsylvania 19104. Received February 2, 2005; revision requested March 31; revision received May 5; accepted May 16. Address correspondence to S.W.S.; E-mail: [email protected] S.W.S. and S.O.T. have received grant support from C.R. Bard, Covington, GA. None of the other authors have identified a conflict of interest. © SIR, 2005 DOI: 10.1097/01.RVI.0000171689.52536.FD
therapy with unfractionated or lowmolecular-weight heparin followed by oral anticoagulation with warfarin remains the standard of care for venous thromboembolic disease (3). Vena cava filters have gained acceptance for use in patients with deep vein thrombosis (DVT) or PE in whom anticoagulation is contraindicated, who have had complications while receiving therapeutic anticoagulation, or who have had recurrent PE or new symptomatic DVT while receiving therapeutic anticoagulation. Additional indications for vena cava filters include free-floating iliofemoral or caval thrombus, DVT or PE with poor cardiopulmonary reserve, or as a means of prophylaxis in patient populations
at high risk for development of thromboembolic disease, such as trauma patients with multisystem injuries, patients scheduled for major surgical procedures, and patients with advanced cancer (4 – 6). Inferior vena cava (IVC) filters are safe to deploy and effective in preventing new PE; however, there are long-term risks associated with IVC filters. These risks include filter migration, filter fracture, caval perforation, and caval thrombosis, and the risk of these long-term complications of indwelling filters increases with time (5,7,8). Whereas the complication rate of IVC filters increases with dwell time, the risk of recurrent thromboembolic disease often decreases with time
1189
1190
•
September 2005
The Recovery Filter as a Retrievable IVC Filter
after the initial episode (9,10). It is therefore desirable to remove IVC filters from patients who are able to tolerate therapeutic anticoagulation or who are no longer at increased risk for thromboembolic disease. The use of temporary IVC filters is an attractive option in patients who require IVC filtration for a limited time. Nonpermanent IVC filtration devices include temporary and retrievable IVC filters. Temporary IVC filters are usually fixed by a wire or catheter to the skin or buried in subcutaneous tissue at the access site. Limitations of temporary IVC filters include mandatory removal within a specified time frame as well as increased infection risk. Retrievable IVC filters are deployed in the vena cava in the same manner as permanent filters, but may be removed percutaneously within a specified period of time. If necessary, retrievable IVC filters may be left in place as permanent IVC filters (5,11). There are currently three retrievable IVC filters available for use in the United States: the Optease filter (Cordis, Warren, NJ), the Gu¨nther Tulip filter (Cook, Bloomington, IN), and the Recovery IVC filter (C.R. Bard, Covington, GA). All three devices are approved by the Food and Drug Administration as retrievable and permanent IVC filters. This study reviews the initial experience with use of the Recovery filter as a retrievable IVC filter.
MATERIALS AND METHODS An institutional review board exemption was granted for this retrospective study. All patients who received the Recovery filter at our institution over a 14-month period were included in this study. Patients who had received a Recovery filter during this time period were identified with an interventional radiology quality assurance database (Hi-IQ; Conexsys, Woonsocket, RI). This database is checked daily against a daily written list of procedures performed to ensure that no cases were missed. Informed consent for placement of the Recovery filter was obtained from each patient. Per division protocol, a letter was given to the patient or family member describing the filter and the possibility of removal. This letter was also mailed to the patient 3 and 6 months after IVC filter placement. A
total of 106 patients received a Recovery filter with an initial clinical indication for temporary caval filtration based on a decision involving the referring physician and the interventional radiology attending physician placing the filter. During this period, all the other 161 patients referred to the interventional radiology division for IVC filter placement received permanent IVC filters. There were 62 male patients and 44 female patients, with a mean age of 48 years (range, 18 –90 y). Mean implantation time of the device was 165 days (range, 0 – 430 d). Each Recovery filter was placed with the intention of retrieval when the patient no longer required caval filtration or was able to tolerate anticoagulation. Medical and radiologic records were examined for indication for filter placement, placement complications, anticoagulation status after filter placement, need for filter removal, ability to retrieve the filter, complications of filter removal, PE, caval thrombosis, and filter migration. The Recovery filter is placed via a femoral approach through an introducer sheath with an internal diameter of 7 F. Maximum recommended caval size for the Recovery filter is 28 mm. The Recovery filter is constructed of a nickel–titanium (nitinol) alloy that has thermal–mechanical properties that allow the filter to assume a linear shape at room temperature and reform the designed filter shape at body temperature (5). The filter consists of twelve 0.014-inch nitinol wires attached to the filter tip, six of which are flexible legs that anchor the device to the caval wall and six of which are arms, creating a dual-level filtration system. Recovery filter retrieval was performed in patients who were no longer at significant risk of thromboembolic disease or who became able to tolerate or be treated with only therapeutic anticoagulation. Before filter removal, all patients were seen in the interventional radiology clinic and underwent computed tomographic (CT) venography or lower-extremity venous ultrasonography (US) study to evaluate for DVT per division protocol. When a clinical decision was made to remove the filter, the patient was scheduled for the removal procedure. An inferior vena cavogram was obtained before removal of all filters by passing a flush catheter inferior to the
JVIR
filter to assess for filter position and clot burden. Retrieval of the Recovery filter was performed via a right jugular approach with use of the 10-F proprietary retrieval device (Recovery Cone Removal System; C.R. Bard). The urethane-covered retrieval cone was used to engage the filter and the filter was retracted into the sheath (Figure). An inferior vena cavogram was obtained after removal of the Recovery filter to evaluate for residual thrombus or caval trauma. All filters were then inspected to ensure that the entire filter was removed and to examine for retained clot.
RESULTS One hundred seven Recovery filters were successfully placed in the infrarenal IVC in 106 patients. Fifty-one filters (48%) were placed as a result of a diagnosis of DVT or PE. Temporary caval filtration was indicated in these patients as a result of a contraindication to anticoagulation (n ⫽ 33), complication of anticoagulation (n ⫽ 8), poor cardiopulmonary reserve (n ⫽ 6), large clot burden (n ⫽ 3), or PE while receiving anticoagulation (n ⫽ 1). Fifty-six Recovery filters (52%) were placed in patients without documented DVT or PE. Indications cited for prophylactic temporary caval filtration included immobility after trauma (n ⫽ 35); recent intracranial bleeding, neurosurgery, or brain tumor (n ⫽ 18); and other surgical or invasive procedure (n ⫽ 3). In one instance, there was incomplete deployment of the Recovery filter in the patient’s IVC. This was believed to represent operator error in deploying the filter; specifically, the pusher rod was withdrawn during deployment, in violation of the instructions for use. This resulted in premature release of the leg hooks within the delivery sheath, in which the filter became stuck in a partially deployed position. The incompletely deployed filter was subsequently retrieved via a right jugular approach and a second filter was deployed in the infrarenal IVC without further complication. All other Recovery filter insertions were without immediate complications. Recovery filter removal was attempted in 15 of 106 patients (14%) a mean of 150 days after placement (range, 0 – 419 days). The Recovery fil-
Volume 16
Number 9
Grande et al
•
1191
Figure. The Recovery Cone is deployed above the Recovery filter before removal (a) and used to capture the tip of the filter (b) and pull it into the vascular sheath (c).
ter was successfully removed in 14 of the 15 patients in whom it was attempted (93%). Ten filters were removed because PE prophylaxis was no longer necessary. Two filters were removed because the patient was able to receive anticoagulation after surgery. Another filter was retrieved after the patient’s initially poor cardiopulmonary reserve was believed to have improved to a degree that it was safe to treat the patient solely with anticoagulation. One filter was retrieved as a result of incomplete deployment during placement, and a second filter was successfully deployed. One retrieval attempt failed in a patient in whom the Recovery filter had been in place for 210 days. In this patient, the Recovery filter tip had become apposed to the posterior wall of the vena cava and could not be engaged despite multiple passes with the Recovery retrieval device. Attempts to straighten the filter with use of a multipurpose angio-
graphic catheter, an Amplatz snare (Microvena, St. Paul, MN), and a tipdeflecting wire were also unsuccessful. After these attempts, the positions of two filter legs were altered. The filter has been left in place. No other complications occurred during filter retrieval. No significant thrombus was seen in any of the filters during retrieval, and no IVC abnormalities were noted on cavograms after retrieval. No patient developed a symptomatic PE after filter retrieval. This cohort includes the 13 patients who had their filters removed and did not have another filter placed. Mean follow-up in this group of patients was 97 days (range, 40 –206 d). Filters remain in place in 92 of 106 patients (87%). No patients developed symptomatic caval thrombosis after filter placement or retrieval. One patient was found to have chronic clot in the left common femoral vein with slow venous flow and lack of respira-
tory variation in either common femoral vein, indicating possible IVC thrombus. This patient was asymptomatic at the time of US and did not undergo further imaging. Another patient was noted on follow-up abdominal CT to have clot trapped in the IVC filter that was not occlusive. This patient has remained asymptomatic. Three patients (2.8%) developed symptomatic pulmonary emboli while the Recovery filter was in place. One patient who received a Recovery filter for bilateral DVT and recent surgery presented to an outside hospital 35 days after initial filter placement with symptoms suggestive of PE and was found to have multiple small emboli by CT. Another patient received a Recovery filter for a postpartum PE with intraperitoneal bleeding, was discharged after filter placement, and was subsequently readmitted 4 days after filter placement for new onset of dyspnea and pleuritic chest pain. The
1192
•
The Recovery Filter as a Retrievable IVC Filter
patient was subsequently discharged with a regimen of oral anticoagulation for a newly symptomatic PE. This same patient presented to the emergency department 3 months after filter placement with dyspnea, and CT imaging showed significant resolution of previous pulmonary emboli, but a new right middle lobe segmental embolus was also found. A third patient received a Recovery filter for large bilateral PE in the setting of a recent subdural hematoma and had documented right upper-extremity and right lower-extremity DVT. Six days after filter placement, the patient developed tachycardia and hypotension and required aggressive vasopressor and ventilatory support. Intravenous thrombolytic therapy was initiated for presumed PE, and the patient experienced a pulseless electrical activity cardiac arrest and was unable to be resuscitated. This event after Recovery filter placement may have been caused by additional acute PE from DVT in the upper or lower extremity or could have represented acute worsening of the patient’s clinical condition as a result of the known large bilateral PE that occurred before Recovery filter placement. Filter migration as a cause for the acute decompensation can be excluded because a chest radiograph taken during the patient’s acute event did not show any evidence of filter migration. None of the patients who had the Recovery filter retrieved had a symptomatic PE after retrieval. No patients who received the Recovery filter experienced symptomatic filter migration. Twenty-four patients (23%) had incidental follow-up abdominal imaging available at least 30 days after filter placement (mean, 116 days; range, 32– 419 d). Thirteen patients underwent imaging by abdominal plain radiography and 14 patients underwent abdominal CT imaging. No patients with follow-up imaging exhibited filter migration. Eleven patients who had their filters left in place (10%) subsequently died during the follow-up period. Ten of these patients died of causes unrelated to their filter or venous thromboembolic disease. One patient mentioned earlier died while undergoing thrombolytic therapy for a massive PE. The patient had a known large PE before the Recovery filter was placed and the patient’s death may or may
not have been caused by new PE, as discussed earlier.
DISCUSSION Anticoagulation remains the treatment of choice for patients with PE or lower-extremity DVT (3). IVC filters remain a safe alternative for PE prophylaxis in patients with contraindications to anticoagulation, complications, or recurrent thromboembolic disease while receiving anticoagulation, and as an adjunct to anticoagulation for patients with poor pulmonary reserve (5,6). Prophylactic filter placement may also reduce morbidity and mortality from thromboembolic disease in patients with multiple traumatic injuries or during the perioperative period in patients at high risk for such events (12). Given that increased risk for thromboembolic disease or contraindications to anticoagulation may be transient, whereas the risks of long-term complications of IVC filters increase with time, retrievable filters provide an attractive option for temporary caval filtration. Despite having an excellent safety profile, IVC filters are associated with some complications. Complications can occur during venous access, filter deployment, or after filter placement (8). It is the complications occurring after filter placement that are most common, and which could possibly be reduced by judicious use and retrieval of IVC filters. Caval occlusion may occur as a result of extension of thrombus trapped in the IVC filter or from spontaneous thrombosis of the IVC. The rate of symptomatic IVC occlusion has been reported to be between 2% and 10%, although asymptomatic caval occlusion can occur in as many as 30% of patients (13–16). Other complications related to long-term filter placement include filter migration and fracture, infection of the filter or trapped thrombus, caval perforation, and guide wire entrapment. Such complications are infrequent but are associated with significant morbidity (17–19). In a randomized prospective study, Decousus et al (20) showed that there was no decrease in incidence of PE or increased rate of survival at 2 years in patients who received a permanent IVC filter and anticoagulation compared with those who received antico-
September 2005
JVIR
agulation alone. This study did demonstrate a significantly increased risk of DVT at 2 years in the patients who received IVC filters. The patients who received IVC filters did have a lower incidence of PE at 12 days. This study suggests there may be a benefit to the use of a retrievable IVC filter that could be used to prevent PE early in the course of venous thromboembolic disease but could then be removed to prevent long-term complications. Our study provides evidence that the Recovery filter is safe and effective to use as a retrievable IVC filter. Recovery filters were safely placed in all 106 patients in whom placement was attempted, with one instance of premature filter deployment in the IVC and no other immediate complications. Symptomatic PE was detected in only 2.8% of patients in this study, which compares well with the results of other IVC filters currently in use (5). No complications such as symptomatic caval thrombosis, caval perforation, filter fracture, or filter migration were detected. Retrieval of the Recovery filter was uneventful in 14 of 15 retrieval attempts. Although all Recovery filters were placed with the intention of future retrieval, less than 15% of filters were retrieved. In an earlier study of the Recovery filter by Asch (21), 75% of filters were ultimately retrieved. Other studies of retrievable filters include the work of Millward et al (22) and Wicky et al (23), who quote retrieval rates with use of the Gu¨nther Tulip filter of 57% and 49%, respectively. The percentage of Recovery filters that were retrieved in our study is much lower than in all of these earlier studies, especially that by Asch (21). This was despite a concerted effort on our part to place these filters in patients who we thought would need only short-term IVC filtration. Our low retrieval rate also occurred despite our aggressive efforts to educate patients, their families, and referring physicians about the ability of these filters to be removed. This included mailing letters regarding filter removal to patients 3 and 6 months after IVC filter placement. An overestimation of the number of patients who would need only temporary IVC filtration probably contributed to our low retrieval rate. Increasing the stringency of patient selection criteria for Recovery filter
Volume 16
Number 9
placement would likely lead to a higher retrieval rate. The ability to retrieve the Recovery filter at time points remote from implantation should lead to increased rates of retrieval in patients with extended need for caval filtration. However, patient follow-up may decrease as time from initial filter placement increases, eroding some of the advantage gained by an extended window of retrievability. The significant number of retrievable filters left in place supports further study of retrievable filters when left in as permanent devices. Long-term studies of the safety and efficacy of the Gu¨nther Tulip, Optease, and Recovery filters when used as permanent filters are needed. There are limitations to our study. The study was a retrospective review that occurred early in our use of the Recovery filter. Patients were not routinely assessed by abdominal radiography for filter migration or fracture; therefore, not all patients were evaluated for these complications. Similarly, not all patients were screened for caval thrombosis after filter placement. Additional patients may have their filters removed in the future because the window of retrieval opportunity has not closed on many of the patients in this study. Therefore, the percentage of filters removed may increase in the future. This study confirms the feasibility, safety, and efficacy of the use of the Recovery filter as a retrievable filter. However, the study also illustrates the need to further evaluate retrievable filters for long-term placement and highlights the need for careful selection and follow-up of patients who receive nonpermanent IVC filters.
Grande et al
References 1. Kearon C. Natural history of venous thromboembolism. Circulation 2003; 107:I22–I30. 2. Bick RL. Hereditary and acquired thrombophilia: preface. Semin Thromb Hemost 1999; 25:251–253. 3. Hyers TM, Agnelli G, Hull RD, et al. Antithrombotic therapy for venous thromboembolic disease. Chest 2001; 119(suppl):176S–193S. 4. Stein PD, Kayali R, Olsen RE. Twenty-one year trends in the use of inferior vena cava filters. Arch Int Med 2004; 164:1541–1545. 5. Kinney TB. Update of inferior vena cava filters. J Vasc Interv Radiol 2003; 14:425–440. 6. Becker DM, Philbrick JT, Selby JB. Inferior vena cava filters: indications, safety, effectiveness. Arch Intern Med 1992; 152:1985–1994. 7. Crochet DP, Stora O, Ferry D, et al. Vena-Tech-LGM Filter: long-term results of a prospective study. Radiology 1993; 188:857–860. 8. Joels CS, Sing RF, Heniford BT. Complications of inferior vena cava filters. Am Surg 2003; 69:654–659. 9. Olin JW. Pulmonary embolism. Rev Cardiovasc Med 2002; 3(suppl 2):S68– S75. 10. Kaufman JA. Retrievable vena cava filters. Tech Vasc Interv Radiol 2004; 7:96–104. 11. Kercher K, Sing RF. Overview of current inferior vena cava filters. Am Surg 2003; 69:643–648. 12. Rogers FB, Shackford SR, Wilson J, et al. Prophylactic vena cava filter insertion in severely injured trauma patients: indications and preliminary results. J Trauma 1993; 35:637–642. 13. Ferris EJ, McCowan TC, Carver DK, et al. Percutaneous inferior vena cava filters: follow-up of seven designs in 320 patients. Radiology 1993; 188:851– 856. 14. Neuerburg JM, Gunther RW, Vorwerk D, et al. Results of a multicenter study of the retrievable Tulip vena
15.
16.
17.
18.
19.
20.
21.
22.
23.
•
1193
cava filter: early clinical experience. Cardiovasc Intervent Radiol 1997; 20: 10–16. Kinney TB, Rose SC, Weingarten KW, et al. IVC filter tilt and asymmetry: comparison of over-the-wire stainlesssteel and titanium Greenfield IVC filters. J Vasc Interv Radiol 1997; 8:1029– 1037. Wojtowycz MM, Stoehr T, Crummy AB, et al. The Bird’s Nest inferior vena cava filter: review of a single-center experience. J Vasc Interv Radiol 1997; 8:171–179. Friedell ML, Goldenkranz RJ, Parsonnet V, et al. Migration of a Greenfield filter to the pulmonary artery: a case report. J Vasc Surg 1986; 3:929–931. Kaufman JA, Thomas JW, Geller SC, et al. Guide-wire entrapment by inferior vena cava filters: in vitro evaluation. Radiology 1996; 198:71–76. Stavropoulos SW, Itkin M, Trerotola SO. In vitro study of guide wire entrapment in currently available inferior vena cava filters. J Vasc Interv Radiol 2003; 14:905–910. Decousus H, Leizorovicz A, Parent F, et al. A clinical trial of vena caval filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis. Prevention du Risque d’Embolie Pulmonaire par Interruption Cave Study Group. N Engl J Med 1998; 338:409–415. Asch MR. Initial experience in humans with a new retrievable inferior vena cava filter. Radiology 2002; 225: 835–844. Millward SF, Oliva VL, Bell SD, et al. Gunther tulip retrievable vena cava filter: results from the registry of the Canadian interventional radiology association. J Vasc Interv Radiol 2001; 12: 1053–1058. Wicky S, Doenz F, Meuwly JY, Portier F, Schnyder P, Denys A. Clinical experience with retrievable Gunther Tulip vena cava filters. J Endovasc Ther 2003; 10:994–1000.
PURPOSE: This study evaluates clinical experience with the Recovery filter as a retrievable inferior vena cava (IVC) filter. MATERIALS AND METHODS: One hundred seven Recovery filters were placed in 106 patients with an initial clinical indication for temporary caval filtration. Patients were followed up to assess filter efficacy, complications, eventual need for filter removal, time to retrieval, and ability to remove the filter. RESULTS: The patient cohort consisted of 62 men and 44 women with a mean age of 48 years (range, 18 –90 y). Mean implantation time was 165 days. Indications for filter placement in patients with deep vein thrombosis (DVT) and/or pulmonary embolism (PE) included contraindication to anticoagulation (n ⴝ 33), complications of anticoagulation (n ⴝ 8), poor cardiopulmonary reserve (n ⴝ 6), large clot burden (n ⴝ 3), and PE while receiving anticoagulation (n ⴝ 1). Indications for filter placement in patients without proven PE or DVT included immobility after trauma (n ⴝ 35); recent intracranial hemorrhage, neurosurgery, or brain tumor (n ⴝ 18); and other surgical or invasive procedure (n ⴝ 3). Three patients (2.8%) had symptomatic PE after placement of the Recovery filter. No caval thromboses were detected. No symptomatic filter migrations occurred. Recovery filter removal was attempted in 15 of 106 patients (14%) at a mean of 150 days after placement. The Recovery filter was successfully retrieved in 14 of 15 patients (93%); one removal was unsuccessful at 210 days after placement. Ninety-two filters (87%) currently remain in place. CONCLUSIONS: Although all the filters were placed with the intention of being removed, a large percentage of filters were not retrieved. The Recovery filter was safe and effective in preventing PE when used as a retrievable IVC filter. J Vasc Interv Radiol 2005; 16:1189 –1193 Abbreviations: DVT ⫽ deep vein thrombosis, IVC ⫽ inferior vena cava, PE ⫽ pulmonary embolism
VENOUS thrombosis and pulmonary embolism (PE) are major causes of morbidity and mortality (1,2). Initial
From the Section of Interventional Radiology, Department of Radiology (W.J.G., S.O.T., T.W.I.C., M.C.S., A.P., R.D.S.G., C.M.T., J.A.S., J.I.M., S.W.S.), and Department of Traumatology and Surgical Critical Care (P.M.R., M.K.F.), University of Pennsylvania Medical Center, 3400 Spruce Street, Philadelphia, Pennsylvania 19104. Received February 2, 2005; revision requested March 31; revision received May 5; accepted May 16. Address correspondence to S.W.S.; E-mail: [email protected] S.W.S. and S.O.T. have received grant support from C.R. Bard, Covington, GA. None of the other authors have identified a conflict of interest. © SIR, 2005 DOI: 10.1097/01.RVI.0000171689.52536.FD
therapy with unfractionated or lowmolecular-weight heparin followed by oral anticoagulation with warfarin remains the standard of care for venous thromboembolic disease (3). Vena cava filters have gained acceptance for use in patients with deep vein thrombosis (DVT) or PE in whom anticoagulation is contraindicated, who have had complications while receiving therapeutic anticoagulation, or who have had recurrent PE or new symptomatic DVT while receiving therapeutic anticoagulation. Additional indications for vena cava filters include free-floating iliofemoral or caval thrombus, DVT or PE with poor cardiopulmonary reserve, or as a means of prophylaxis in patient populations
at high risk for development of thromboembolic disease, such as trauma patients with multisystem injuries, patients scheduled for major surgical procedures, and patients with advanced cancer (4 – 6). Inferior vena cava (IVC) filters are safe to deploy and effective in preventing new PE; however, there are long-term risks associated with IVC filters. These risks include filter migration, filter fracture, caval perforation, and caval thrombosis, and the risk of these long-term complications of indwelling filters increases with time (5,7,8). Whereas the complication rate of IVC filters increases with dwell time, the risk of recurrent thromboembolic disease often decreases with time
1189
1190
•
September 2005
The Recovery Filter as a Retrievable IVC Filter
after the initial episode (9,10). It is therefore desirable to remove IVC filters from patients who are able to tolerate therapeutic anticoagulation or who are no longer at increased risk for thromboembolic disease. The use of temporary IVC filters is an attractive option in patients who require IVC filtration for a limited time. Nonpermanent IVC filtration devices include temporary and retrievable IVC filters. Temporary IVC filters are usually fixed by a wire or catheter to the skin or buried in subcutaneous tissue at the access site. Limitations of temporary IVC filters include mandatory removal within a specified time frame as well as increased infection risk. Retrievable IVC filters are deployed in the vena cava in the same manner as permanent filters, but may be removed percutaneously within a specified period of time. If necessary, retrievable IVC filters may be left in place as permanent IVC filters (5,11). There are currently three retrievable IVC filters available for use in the United States: the Optease filter (Cordis, Warren, NJ), the Gu¨nther Tulip filter (Cook, Bloomington, IN), and the Recovery IVC filter (C.R. Bard, Covington, GA). All three devices are approved by the Food and Drug Administration as retrievable and permanent IVC filters. This study reviews the initial experience with use of the Recovery filter as a retrievable IVC filter.
MATERIALS AND METHODS An institutional review board exemption was granted for this retrospective study. All patients who received the Recovery filter at our institution over a 14-month period were included in this study. Patients who had received a Recovery filter during this time period were identified with an interventional radiology quality assurance database (Hi-IQ; Conexsys, Woonsocket, RI). This database is checked daily against a daily written list of procedures performed to ensure that no cases were missed. Informed consent for placement of the Recovery filter was obtained from each patient. Per division protocol, a letter was given to the patient or family member describing the filter and the possibility of removal. This letter was also mailed to the patient 3 and 6 months after IVC filter placement. A
total of 106 patients received a Recovery filter with an initial clinical indication for temporary caval filtration based on a decision involving the referring physician and the interventional radiology attending physician placing the filter. During this period, all the other 161 patients referred to the interventional radiology division for IVC filter placement received permanent IVC filters. There were 62 male patients and 44 female patients, with a mean age of 48 years (range, 18 –90 y). Mean implantation time of the device was 165 days (range, 0 – 430 d). Each Recovery filter was placed with the intention of retrieval when the patient no longer required caval filtration or was able to tolerate anticoagulation. Medical and radiologic records were examined for indication for filter placement, placement complications, anticoagulation status after filter placement, need for filter removal, ability to retrieve the filter, complications of filter removal, PE, caval thrombosis, and filter migration. The Recovery filter is placed via a femoral approach through an introducer sheath with an internal diameter of 7 F. Maximum recommended caval size for the Recovery filter is 28 mm. The Recovery filter is constructed of a nickel–titanium (nitinol) alloy that has thermal–mechanical properties that allow the filter to assume a linear shape at room temperature and reform the designed filter shape at body temperature (5). The filter consists of twelve 0.014-inch nitinol wires attached to the filter tip, six of which are flexible legs that anchor the device to the caval wall and six of which are arms, creating a dual-level filtration system. Recovery filter retrieval was performed in patients who were no longer at significant risk of thromboembolic disease or who became able to tolerate or be treated with only therapeutic anticoagulation. Before filter removal, all patients were seen in the interventional radiology clinic and underwent computed tomographic (CT) venography or lower-extremity venous ultrasonography (US) study to evaluate for DVT per division protocol. When a clinical decision was made to remove the filter, the patient was scheduled for the removal procedure. An inferior vena cavogram was obtained before removal of all filters by passing a flush catheter inferior to the
JVIR
filter to assess for filter position and clot burden. Retrieval of the Recovery filter was performed via a right jugular approach with use of the 10-F proprietary retrieval device (Recovery Cone Removal System; C.R. Bard). The urethane-covered retrieval cone was used to engage the filter and the filter was retracted into the sheath (Figure). An inferior vena cavogram was obtained after removal of the Recovery filter to evaluate for residual thrombus or caval trauma. All filters were then inspected to ensure that the entire filter was removed and to examine for retained clot.
RESULTS One hundred seven Recovery filters were successfully placed in the infrarenal IVC in 106 patients. Fifty-one filters (48%) were placed as a result of a diagnosis of DVT or PE. Temporary caval filtration was indicated in these patients as a result of a contraindication to anticoagulation (n ⫽ 33), complication of anticoagulation (n ⫽ 8), poor cardiopulmonary reserve (n ⫽ 6), large clot burden (n ⫽ 3), or PE while receiving anticoagulation (n ⫽ 1). Fifty-six Recovery filters (52%) were placed in patients without documented DVT or PE. Indications cited for prophylactic temporary caval filtration included immobility after trauma (n ⫽ 35); recent intracranial bleeding, neurosurgery, or brain tumor (n ⫽ 18); and other surgical or invasive procedure (n ⫽ 3). In one instance, there was incomplete deployment of the Recovery filter in the patient’s IVC. This was believed to represent operator error in deploying the filter; specifically, the pusher rod was withdrawn during deployment, in violation of the instructions for use. This resulted in premature release of the leg hooks within the delivery sheath, in which the filter became stuck in a partially deployed position. The incompletely deployed filter was subsequently retrieved via a right jugular approach and a second filter was deployed in the infrarenal IVC without further complication. All other Recovery filter insertions were without immediate complications. Recovery filter removal was attempted in 15 of 106 patients (14%) a mean of 150 days after placement (range, 0 – 419 days). The Recovery fil-
Volume 16
Number 9
Grande et al
•
1191
Figure. The Recovery Cone is deployed above the Recovery filter before removal (a) and used to capture the tip of the filter (b) and pull it into the vascular sheath (c).
ter was successfully removed in 14 of the 15 patients in whom it was attempted (93%). Ten filters were removed because PE prophylaxis was no longer necessary. Two filters were removed because the patient was able to receive anticoagulation after surgery. Another filter was retrieved after the patient’s initially poor cardiopulmonary reserve was believed to have improved to a degree that it was safe to treat the patient solely with anticoagulation. One filter was retrieved as a result of incomplete deployment during placement, and a second filter was successfully deployed. One retrieval attempt failed in a patient in whom the Recovery filter had been in place for 210 days. In this patient, the Recovery filter tip had become apposed to the posterior wall of the vena cava and could not be engaged despite multiple passes with the Recovery retrieval device. Attempts to straighten the filter with use of a multipurpose angio-
graphic catheter, an Amplatz snare (Microvena, St. Paul, MN), and a tipdeflecting wire were also unsuccessful. After these attempts, the positions of two filter legs were altered. The filter has been left in place. No other complications occurred during filter retrieval. No significant thrombus was seen in any of the filters during retrieval, and no IVC abnormalities were noted on cavograms after retrieval. No patient developed a symptomatic PE after filter retrieval. This cohort includes the 13 patients who had their filters removed and did not have another filter placed. Mean follow-up in this group of patients was 97 days (range, 40 –206 d). Filters remain in place in 92 of 106 patients (87%). No patients developed symptomatic caval thrombosis after filter placement or retrieval. One patient was found to have chronic clot in the left common femoral vein with slow venous flow and lack of respira-
tory variation in either common femoral vein, indicating possible IVC thrombus. This patient was asymptomatic at the time of US and did not undergo further imaging. Another patient was noted on follow-up abdominal CT to have clot trapped in the IVC filter that was not occlusive. This patient has remained asymptomatic. Three patients (2.8%) developed symptomatic pulmonary emboli while the Recovery filter was in place. One patient who received a Recovery filter for bilateral DVT and recent surgery presented to an outside hospital 35 days after initial filter placement with symptoms suggestive of PE and was found to have multiple small emboli by CT. Another patient received a Recovery filter for a postpartum PE with intraperitoneal bleeding, was discharged after filter placement, and was subsequently readmitted 4 days after filter placement for new onset of dyspnea and pleuritic chest pain. The
1192
•
The Recovery Filter as a Retrievable IVC Filter
patient was subsequently discharged with a regimen of oral anticoagulation for a newly symptomatic PE. This same patient presented to the emergency department 3 months after filter placement with dyspnea, and CT imaging showed significant resolution of previous pulmonary emboli, but a new right middle lobe segmental embolus was also found. A third patient received a Recovery filter for large bilateral PE in the setting of a recent subdural hematoma and had documented right upper-extremity and right lower-extremity DVT. Six days after filter placement, the patient developed tachycardia and hypotension and required aggressive vasopressor and ventilatory support. Intravenous thrombolytic therapy was initiated for presumed PE, and the patient experienced a pulseless electrical activity cardiac arrest and was unable to be resuscitated. This event after Recovery filter placement may have been caused by additional acute PE from DVT in the upper or lower extremity or could have represented acute worsening of the patient’s clinical condition as a result of the known large bilateral PE that occurred before Recovery filter placement. Filter migration as a cause for the acute decompensation can be excluded because a chest radiograph taken during the patient’s acute event did not show any evidence of filter migration. None of the patients who had the Recovery filter retrieved had a symptomatic PE after retrieval. No patients who received the Recovery filter experienced symptomatic filter migration. Twenty-four patients (23%) had incidental follow-up abdominal imaging available at least 30 days after filter placement (mean, 116 days; range, 32– 419 d). Thirteen patients underwent imaging by abdominal plain radiography and 14 patients underwent abdominal CT imaging. No patients with follow-up imaging exhibited filter migration. Eleven patients who had their filters left in place (10%) subsequently died during the follow-up period. Ten of these patients died of causes unrelated to their filter or venous thromboembolic disease. One patient mentioned earlier died while undergoing thrombolytic therapy for a massive PE. The patient had a known large PE before the Recovery filter was placed and the patient’s death may or may
not have been caused by new PE, as discussed earlier.
DISCUSSION Anticoagulation remains the treatment of choice for patients with PE or lower-extremity DVT (3). IVC filters remain a safe alternative for PE prophylaxis in patients with contraindications to anticoagulation, complications, or recurrent thromboembolic disease while receiving anticoagulation, and as an adjunct to anticoagulation for patients with poor pulmonary reserve (5,6). Prophylactic filter placement may also reduce morbidity and mortality from thromboembolic disease in patients with multiple traumatic injuries or during the perioperative period in patients at high risk for such events (12). Given that increased risk for thromboembolic disease or contraindications to anticoagulation may be transient, whereas the risks of long-term complications of IVC filters increase with time, retrievable filters provide an attractive option for temporary caval filtration. Despite having an excellent safety profile, IVC filters are associated with some complications. Complications can occur during venous access, filter deployment, or after filter placement (8). It is the complications occurring after filter placement that are most common, and which could possibly be reduced by judicious use and retrieval of IVC filters. Caval occlusion may occur as a result of extension of thrombus trapped in the IVC filter or from spontaneous thrombosis of the IVC. The rate of symptomatic IVC occlusion has been reported to be between 2% and 10%, although asymptomatic caval occlusion can occur in as many as 30% of patients (13–16). Other complications related to long-term filter placement include filter migration and fracture, infection of the filter or trapped thrombus, caval perforation, and guide wire entrapment. Such complications are infrequent but are associated with significant morbidity (17–19). In a randomized prospective study, Decousus et al (20) showed that there was no decrease in incidence of PE or increased rate of survival at 2 years in patients who received a permanent IVC filter and anticoagulation compared with those who received antico-
September 2005
JVIR
agulation alone. This study did demonstrate a significantly increased risk of DVT at 2 years in the patients who received IVC filters. The patients who received IVC filters did have a lower incidence of PE at 12 days. This study suggests there may be a benefit to the use of a retrievable IVC filter that could be used to prevent PE early in the course of venous thromboembolic disease but could then be removed to prevent long-term complications. Our study provides evidence that the Recovery filter is safe and effective to use as a retrievable IVC filter. Recovery filters were safely placed in all 106 patients in whom placement was attempted, with one instance of premature filter deployment in the IVC and no other immediate complications. Symptomatic PE was detected in only 2.8% of patients in this study, which compares well with the results of other IVC filters currently in use (5). No complications such as symptomatic caval thrombosis, caval perforation, filter fracture, or filter migration were detected. Retrieval of the Recovery filter was uneventful in 14 of 15 retrieval attempts. Although all Recovery filters were placed with the intention of future retrieval, less than 15% of filters were retrieved. In an earlier study of the Recovery filter by Asch (21), 75% of filters were ultimately retrieved. Other studies of retrievable filters include the work of Millward et al (22) and Wicky et al (23), who quote retrieval rates with use of the Gu¨nther Tulip filter of 57% and 49%, respectively. The percentage of Recovery filters that were retrieved in our study is much lower than in all of these earlier studies, especially that by Asch (21). This was despite a concerted effort on our part to place these filters in patients who we thought would need only short-term IVC filtration. Our low retrieval rate also occurred despite our aggressive efforts to educate patients, their families, and referring physicians about the ability of these filters to be removed. This included mailing letters regarding filter removal to patients 3 and 6 months after IVC filter placement. An overestimation of the number of patients who would need only temporary IVC filtration probably contributed to our low retrieval rate. Increasing the stringency of patient selection criteria for Recovery filter
Volume 16
Number 9
placement would likely lead to a higher retrieval rate. The ability to retrieve the Recovery filter at time points remote from implantation should lead to increased rates of retrieval in patients with extended need for caval filtration. However, patient follow-up may decrease as time from initial filter placement increases, eroding some of the advantage gained by an extended window of retrievability. The significant number of retrievable filters left in place supports further study of retrievable filters when left in as permanent devices. Long-term studies of the safety and efficacy of the Gu¨nther Tulip, Optease, and Recovery filters when used as permanent filters are needed. There are limitations to our study. The study was a retrospective review that occurred early in our use of the Recovery filter. Patients were not routinely assessed by abdominal radiography for filter migration or fracture; therefore, not all patients were evaluated for these complications. Similarly, not all patients were screened for caval thrombosis after filter placement. Additional patients may have their filters removed in the future because the window of retrieval opportunity has not closed on many of the patients in this study. Therefore, the percentage of filters removed may increase in the future. This study confirms the feasibility, safety, and efficacy of the use of the Recovery filter as a retrievable filter. However, the study also illustrates the need to further evaluate retrievable filters for long-term placement and highlights the need for careful selection and follow-up of patients who receive nonpermanent IVC filters.
Grande et al
References 1. Kearon C. Natural history of venous thromboembolism. Circulation 2003; 107:I22–I30. 2. Bick RL. Hereditary and acquired thrombophilia: preface. Semin Thromb Hemost 1999; 25:251–253. 3. Hyers TM, Agnelli G, Hull RD, et al. Antithrombotic therapy for venous thromboembolic disease. Chest 2001; 119(suppl):176S–193S. 4. Stein PD, Kayali R, Olsen RE. Twenty-one year trends in the use of inferior vena cava filters. Arch Int Med 2004; 164:1541–1545. 5. Kinney TB. Update of inferior vena cava filters. J Vasc Interv Radiol 2003; 14:425–440. 6. Becker DM, Philbrick JT, Selby JB. Inferior vena cava filters: indications, safety, effectiveness. Arch Intern Med 1992; 152:1985–1994. 7. Crochet DP, Stora O, Ferry D, et al. Vena-Tech-LGM Filter: long-term results of a prospective study. Radiology 1993; 188:857–860. 8. Joels CS, Sing RF, Heniford BT. Complications of inferior vena cava filters. Am Surg 2003; 69:654–659. 9. Olin JW. Pulmonary embolism. Rev Cardiovasc Med 2002; 3(suppl 2):S68– S75. 10. Kaufman JA. Retrievable vena cava filters. Tech Vasc Interv Radiol 2004; 7:96–104. 11. Kercher K, Sing RF. Overview of current inferior vena cava filters. Am Surg 2003; 69:643–648. 12. Rogers FB, Shackford SR, Wilson J, et al. Prophylactic vena cava filter insertion in severely injured trauma patients: indications and preliminary results. J Trauma 1993; 35:637–642. 13. Ferris EJ, McCowan TC, Carver DK, et al. Percutaneous inferior vena cava filters: follow-up of seven designs in 320 patients. Radiology 1993; 188:851– 856. 14. Neuerburg JM, Gunther RW, Vorwerk D, et al. Results of a multicenter study of the retrievable Tulip vena
15.
16.
17.
18.
19.
20.
21.
22.
23.
•
1193
cava filter: early clinical experience. Cardiovasc Intervent Radiol 1997; 20: 10–16. Kinney TB, Rose SC, Weingarten KW, et al. IVC filter tilt and asymmetry: comparison of over-the-wire stainlesssteel and titanium Greenfield IVC filters. J Vasc Interv Radiol 1997; 8:1029– 1037. Wojtowycz MM, Stoehr T, Crummy AB, et al. The Bird’s Nest inferior vena cava filter: review of a single-center experience. J Vasc Interv Radiol 1997; 8:171–179. Friedell ML, Goldenkranz RJ, Parsonnet V, et al. Migration of a Greenfield filter to the pulmonary artery: a case report. J Vasc Surg 1986; 3:929–931. Kaufman JA, Thomas JW, Geller SC, et al. Guide-wire entrapment by inferior vena cava filters: in vitro evaluation. Radiology 1996; 198:71–76. Stavropoulos SW, Itkin M, Trerotola SO. In vitro study of guide wire entrapment in currently available inferior vena cava filters. J Vasc Interv Radiol 2003; 14:905–910. Decousus H, Leizorovicz A, Parent F, et al. A clinical trial of vena caval filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis. Prevention du Risque d’Embolie Pulmonaire par Interruption Cave Study Group. N Engl J Med 1998; 338:409–415. Asch MR. Initial experience in humans with a new retrievable inferior vena cava filter. Radiology 2002; 225: 835–844. Millward SF, Oliva VL, Bell SD, et al. Gunther tulip retrievable vena cava filter: results from the registry of the Canadian interventional radiology association. J Vasc Interv Radiol 2001; 12: 1053–1058. Wicky S, Doenz F, Meuwly JY, Portier F, Schnyder P, Denys A. Clinical experience with retrievable Gunther Tulip vena cava filters. J Endovasc Ther 2003; 10:994–1000.