The RETRIEVE Trial: Safety and Effectiveness of the Retrievable Crux Vena Cava Filter

CLINICAL STUDY

The RETRIEVE Trial: Safety and Effectiveness of the Retrievable Crux Vena Cava Filter H. Bob Smouse, MD, Robert Mendes, MD, Marc Bosiers, MD, Thuong G. Van Ha, MD, and Tami Crabtree, MS, for the RETRIEVE Investigators

ABSTRACT Purpose: To evaluate the safety and effectiveness of the Crux vena cava filter in patients at risk for pulmonary embolism (PE). Materials and Methods: The Crux Biomedical Evaluation of the Crux Inferior Vena Cava Filter System trial was an international prospective, multicenter, single-arm clinical trial in 125 patients implanted with the Crux filter between June 2010 and June 2011. Follow-up was 180 days after filter placement and 30 days after filter retrieval. The primary objective was to determine whether the clinical success rate was at least 80%. Clinical success was defined as technical success of deployment and freedom from definite PE, filter migration, and device-related adverse events requiring intervention. Results: The clinical success rate was 96.0% (120 of 125), with a one-sided lower limit of the 95% confidence interval of 91.8%. The rate of technical success was 98.4% (123 of 125). There were three cases of definite PE (2.4%), two cases of deployment failure, and no cases of device migration, embolization, fracture, or tilting. Investigators observed nine cases of thrombus (all nonocclusive) in or near the filter (six during retrieval evaluation vena cavography, two during computed tomography [CT] scans for PE symptoms, and one during CT for cancer management) and 13 cases of deep vein thrombosis. Device retrieval was attempted at a mean of 84.6 days ⫾ 57.6 (range, 6–190 d) after implantation and was successful for 98.1% of patients (53 of 54). All deaths (n ¼ 14) were determined to be unrelated to the filter or PE. Conclusions: The Crux vena cava filter performed safely, with high rates of clinical, technical, and retrieval success.

ABBREVIATIONS CI = confidence interval, DVT = deep vein thrombosis, ePTFE = expanded polytetrafluoroethylene, FDA = Food and Drug Administration, IVC = inferior vena cava, PE = pulmonary embolism, RETRIEVE = Crux Biomedical Evaluation of the Crux Inferior Vena Cava Filter System [trial], VTE = venous thromboembolism

Permanently deployed inferior vena cava (IVC) filters have been effective in reducing the short-term and long-term risk of pulmonary embolism (PE), but they have been From the Department of Vascular and Interventional Radiology (H.B.S.), OSF St. Francis Medical Center, 530 NE Glen Oak, Peoria IL 61637; Department of Vascular Surgery (R.M.), Rex Hospital, Raleigh, North Carolina; Department of Vascular Surgery (M.B.), A.Z. St. Blasius Hospital, Dendermonde, Belgium; Department of Radiology and Section of Vascular and Interventional Radiology (T.G.V.H.), University of Chicago Medical Center, Chicago, Illinois; and independent biostatistician (T.C.), Santa Rosa, California. Received August 21, 2012; final revision received January 11, 2013; accepted January 16, 2013. Address correspondence to H.B.S.; E-mail: [email protected] From the SIR 2012 Annual Meeting. The RETRIEVE trial was sponsored by Crux Biomedical (Menlo Park, California). H.B.S., R.M., and T.C. are consultants for Crux Biomedical. None of the other authors have identified a conflict of interest. & SIR, 2013 J Vasc Interv Radiol 2013; 24:609–621 http://dx.doi.org/10.1016/j.jvir.2013.01.489

associated with increased incidence of proximal deep vein thrombosis (DVT), with the risk appearing to increase with the indwelling duration (1–3). Other complications include venous access site thrombosis, hematoma, and infection; filter misplacement; penetration of the vessel wall; and filter migration, tilting, obstruction, and fracture (4,5). In 2003 and 2004, the United States Food and Drug Administration (FDA) approved changes in the instructions for use of three then-available permanent filters to allow for percutaneous retrieval, without modification of the indications for placement or the addition of indications for retrieval (6,7). The subsequent development of devices specifically designed to offer the option of permanent retention or appropriately timed removal has led to increased use of vena cava filters, with application expanded to patients with clearly short-term indications (eg, in a setting of trauma). The potential long-term complications associated with the permanent devices may be avoided by filter retrieval if and when the risks of

610

’

Safety and Effectiveness of the Retrievable IVC Filter

bleeding or PE resolve or substantially decrease (7,8). Although several retrievable filters have been approved by the FDA, there has been limited publication of supportive data from prospective randomized trials (9–14). Concerns have been raised related to reported low rates of actual retrieval of the devices, and about the occurrence of timedependent and device-dependent serious complications including filter migration or fracture and IVC perforation or occlusion (8,15–17). Although the complication rates of the retrievable devices may increase with prolonged indwelling time, the risk of retrieval failure may also increase (8,16,18). The Crux Vena Cava Filter (Crux Biomedical, Menlo Park, California) is a new retrievable device with a nonconical, opposing-helix design including a thrombustrapping web of expanded polytetrafluoroethylene (ePTFE) filaments and atraumatic, radiopaque retrieval tails at both ends to allow the option of jugular or femoral access for retrieval as well as for delivery. The Crux Biomedical Evaluation of the Crux Inferior Vena Cava Filter System (RETRIEVE) trial was undertaken to assess the safety and effectiveness of the Crux filter as both a permanent and a retrievable device.

MATERIALS AND METHODS

Smouse et al

’

JVIR

regulations. Analyses of primary and secondary endpoints were performed by an independent statistician and the trial principal investigator. An independent physician served as the medical monitor.

Patient Selection The trial inclusion and exclusion criteria were consistent with previous prospective studies with retrievable filters (11–14). Enrolled patients were older than 18 years of age, had temporary or permanent risk of PE, and had clinical indications for filter placement as a result of contraindication to or failure of anticoagulation or the presence of temporary risk factors such as trauma or planned surgical procedures (eg, bariatric or pelvic surgery) or medical conditions known to increase the risk of venous thromboembolism (VTE). The indications for enrollment were consistent with the therapeutic (ie, documented thromboembolic disease) and prophylactic (ie, no current thromboembolic disease) categories in the 2011 Quality Improvement Guidelines for IVC filter placement from the Society of Interventional Radiology (SIR) Standards of Practice Committee (Table 1) (19). To be eligible, patients were required to have a documented infrarenal IVC diameter of 17–28 mm and venous anatomy and access vessels adequate for infrarenal IVC placement of the filter. Patients were excluded if they were pregnant, if they

Study Design and Conduct The RETRIEVE trial was a prospective, multicenter, singlearm, nonrandomized clinical trial conducted in patients with temporary or permanent risk of PE, and in whom IVC filtration was clinically indicated as a result of contraindication to or failure of anticoagulation. The RETRIEVE trial outcomes reported here are pooled from three clinical trials that were conducted during the same time period: RETRIEVE 2 (clinicaltrials.gov identifier NCT01120509), conducted in the United States with an FDA-approved investigational device exemption; RETRIEVE 3 (clinicaltrials.gov identifier NCT01120522), conducted in Belgium with Federal Agency for Medicines and Health Products notification; and RETRIEVE 4 (clinicaltrials.gov identifier NCT01120535), conducted in Australia and New Zealand with Therapeutic Goods Agency notification. RETRIEVE 1 (clinicaltrials.gov identifier NCT00605332) was an earlier trial of a different design of the filter. For the three current RETRIEVE trials, the filters implanted, the site training methods, the subject eligibility criteria, the case report forms, the Data and Safety Monitoring Board (DSMB), the study definitions, the monitoring plan, and the database were all identical, and procedures to ensure data quality were applied with equal rigor. Each RETRIEVE trial site (see Appendix) obtained ethics committee or institutional review board approval, and all patients provided written informed consent before trial enrollment. The RETRIEVE trial was conducted in accordance with the ethical principles of the Declaration of Helsinki, and in compliance with local and national

Table 1 . Society of Interventional Radiology Standards of Practice Committee Classification of Indications for IVC Filter Placement (19) Therapeutic indications (documented thromboembolic disease) Evidence of PE or IVC/iliac/femoropopliteal DVT and one or more of the following: Absolute or relative contraindication to anticoagulation Complications of anticoagulation Failure of anticoagulation Recurrent PE despite adequate therapy Inability to achieve/maintain adequate anticoagulation Propagation/progression of DVT during therapeutic anticoagulation Massive PE with residual DVT in a patient at risk for further PE Free-floating ileofemoral or IVC thrombus Severe cardiopulmonary disease and DVT (eg, cor pulmonale with pulmonary hypertension) Prophylactic indications (no current thromboembolic disease) Temporary risk of PE in one of the following settings: Severe trauma without documented PE or DVT Closed head injury Spinal cord injury Multiple long-bone or pelvic fractures High-risk situations (eg, patient immobilized or in ICU) DVT ¼ deep vein thrombosis, ICU ¼ intensive care unit, IVC ¼ inferior vena cava, PE ¼ pulmonary embolism.

Volume 24

’

Number 5

’

May

’

2013

already had a filter in place, if they required suprarenal filter placement, if they had a condition inhibiting IVC radiographic visualization, or if they had a known allergy or intolerance to polytetrafluoroethylene or nitinol. They were also excluded if they had renal vein thrombosis, duplication of the IVC, gonadal vein thrombosis, uncontrolled infectious disease, or uncontrollable coagulopathy. Anatomic eligibility criteria were assessed with angiography at the time of filter placement. Any patient who did not meet these criteria could receive a commercially available filter. Any patient who was enrolled but did not receive the investigational device was only followed through a 30-day visit and then considered to have completed the trial. Patients meeting the trial eligibility requirements continued to receive any concomitant therapies including chemotherapy, surgical procedures, and initiation or reinitiation of antithrombotic agents throughout the trial period.

Device Description The Crux Vena Cava Filter System (Fig 1) consists of a self-expanding nitinol filter delivered from a single-use disposable 9-F catheter. The filter is composed of two opposing self-expanding nitinol spiral wireforms. One end of each wireform is formed into a sinusoidal-shaped retrieval tail to aid in retrieval of the filter by means of femoral vein or jugular vein access. Five tissue anchors,

611

oriented to reduce cranial migration, are attached to the spiral wireforms. The thrombus-trapping portion of the filter is formed from a web of ePTFE filaments attached to the caudal half of the filter. The ePTFE is radiolucent. To ensure that the filter is always deployed with the web oriented caudally, the filter system comes in two distinct preloaded configurations, depending on whether access will be femoral or jugular. One filter size is available to treat IVC inner diameters of 17–28 mm.

Filter Placement and Retrieval Technique After preparation of the filter in accordance with the instructions for use, femoral or jugular vein access was obtained with standard endovascular techniques. A vena cavogram was obtained before filter insertion. Deployment of the filter could proceed if the IVC measured between 17 and 28 mm at its widest diameter and the patient met all other anatomic eligibility criteria. The radiopaque marker band on the delivery catheter was positioned at the iliac vein bifurcation when the approach was via the jugular vein, and just below the renal veins when the approach was via the femoral vein. Hemostasis was achieved by using standard techniques. Removal of the filter could be performed from the femoral vein or a jugular vein. The use of a two-sheath coaxial system (6-F inner sheath, 10-F outer sheath) with a snare device was recommended but not required; the final choice of retrieval system was left to operator discretion. With use of a two-sheath coaxial system, the snare and inner sheath were advanced to capture the filter retrieval tail within the inner sheath, the outer sheath was advanced over the filter, and both retrieval sheaths were removed from the patient with the filter and snare within them (Fig 2).

Trial Follow-up and Evaluation for Filter Retrieval

Figure 1. The Crux vena cava filter.

Patients were evaluated at 30, 90, and 180 days after implantation. At any time after filter implantation and before the 180-day follow-up visit, the investigator could evaluate the patient for filter retrieval. The predefined criteria for filter retrieval were satisfied when all the following were true: the patient’s clinical status no longer required the filter in situ, there was no acute swelling of the legs, a duplex ultrasound (US) examination was negative for lowerextremity DVT, and the retrieval vena cavogram confirmed no evidence of clinically significant thrombus. When filter retrieval took place, the patient returned for a 30-day postretrieval visit to complete trial participation. All other patients, including those who were evaluated for filter retrieval but did not undergo retrieval and those in whom retrieval attempts were not successful, were continued on the schedule of regular follow-up through the 180-day visit at study exit. All follow-up and retrieval visits included clinical assessment for PE and DVT, x-ray image review, and assessment for adverse events between visits (eg, shortness of breath, leg swelling). Filter stability and integrity were assessed by x-ray evaluation (anteroposterior and both

612

’

Safety and Effectiveness of the Retrievable IVC Filter

Smouse et al

’

JVIR

Figure 2. Retrieval of the Crux vena cava filter. (a) The 10-F outer sheath of the retrieval snare device has been advanced over the 6-F sheath, which has captured the Crux filter retrieval tail. (b) The 10-F outer sheath of the retrieval snare device has been advanced over the Crux filter.

oblique projections) at each follow-up visit, images from which were compared with baseline images. Any changes from baseline were assessed by a DSMB member.

Trial Endpoints and Data Analysis The filter was evaluated in accordance with the SIR reporting standards (20,21). The predefined primary endpoint for the RETRIEVE trial was clinical success, a composite of technical success of deployment, freedom from definite PE, freedom from filter migration, and freedom from filterrelated adverse events requiring intervention. Technical success was defined as implantation of the filter in the infrarenal IVC at the desired location in a functional orientation. Definite PE was defined as thrombus within the pulmonary arteries documented by pulmonary arteriography, CT cross-sectional imaging, ventilation/perfusion lung scan, or autopsy. Filter migration was defined as a shift in position, either caudal or cranial, by more than 20 mm from the original implant location. Predefined secondary endpoints included retrieval success, filter migration, thrombus in or near the filter, and device integrity. Retrieval success was defined as percutaneous removal of the filter without a serious device-related or procedure-related complication through 30 days after retrieval. Adverse event information was collected throughout the trial. An adverse event was considered serious if it was fatal or life-threatening, if it required or prolonged hospitalization, if it resulted in persistent or significant disability, or if it resulted in medical or surgical intervention to prevent permanent injury. The classification of serious adverse events was similar to that for ‘‘major’’ complications in the SIR clinical practice guidelines (22). All adverse events were assessed by the trial independent

reviewer, by a member of the DSMB, or, in some cases, by the full DSMB. The primary objective was to determine whether the onesided lower limit of the 95% confidence interval (CI) for the observed clinical success rate was at least 80%. With a onesided a of 0.05, and allowing for 15% attrition, 104 patients were required to provide 80% power to test the primary endpoint. The 95% one-sided lower confidence limit was calculated by using the Clopper–Pearson exact method. The analysis included all available data. Statistical analyses were performed by using SAS version 9.3 (SAS, Cary, North Carolina) and StatXact (Cytel, Cambridge, Massachusetts). The final enrollment of 125 was used to meet a prospective regulatory objective for retrieval cases. Post hoc sensitivity analyses were conducted to determine differences in the primary endpoint when patients who did not complete the trial per protocol and therefore had incomplete data were excluded from the analysis and when patients with clinical events without imaging confirmation (and therefore not meeting the protocol definition of definite PE) adjudicated as cases of suspected PE were included. In another post hoc analysis, all enrolled patients were categorized as having therapeutic or prophylactic indications for treatment, according to the 2011 Quality Improvement Guidelines for IVC filter placement from the SIR Standards of Practice Committee (Table 1) (19), and the rate of filter retrieval for each of these categories was determined.

RESULTS The RETRIEVE trial enrolled 125 patients between June 2010 and June 2011. Study patient disposition is detailed in Figure 3. There were 105 patients eligible for 30-day

Volume 24

’

Number 5

’

May

’

2013

613

Figure 3. Disposition of the 125 enrolled patients through the course of the RETRIEVE trial.

follow-up, 70 eligible for 90-day follow-up, and 49 eligible for 180-day follow-up. There were two technical failures, 53 successful retrievals with no subsequent deaths, 14 deaths in patients who retained their filters, two patients lost to follow-up, and six patients who withdrew from the trial. Enrollment was 88 patients at the 17 sites in the United States, 17 patients at the two sites in Belgium, and 20 patients at the three sites in Australia and New Zealand. The baseline characteristics, medication use, and thrombotic risk factors are detailed in Table 2, and the indications identified as primary for filter placement are summarized in Table 3, with breakouts in both tables for the patients who eventually underwent successful filter retrieval compared with those who retained their filters. The mean patient age was 59.1 years ⫾ 17.2, 58.4% were male, and 79.2% were white. Most of the patients had a history of cardiovascular, pulmonary, or trauma conditions, and most patients had multiple thrombotic risk factors, including active DVT at baseline in 58.4%, history of DVT in 49.6%, contraindication to anticoagulation in 37.6%, and history of PE in 36.8%. The majority (85.6%) had more than one of these risk factors. At trial baseline, some form of antithrombotic medication was prescribed for 72 of the 125 patients enrolled (57.6%). At the time of the 30-day visit following device retrieval, 34 of the 52 patients (65.4%) who had the follow-up were

receiving some form of antithrombotic medication. Of the 49 patients still in the trial with their filters left in place at the end of the 180-day visit, 42 (85.7%) were receiving some form of antithrombotic medication.

Filter Placement Technical success of deployment was achieved in 123 of the 125 enrolled patients (98.4%). Two cases were categorized as technical failures. In one of these cases, with an infrarenal IVC length of 10 cm, the filter was deployed with the cranial loop across the renal vein ostia in a ‘‘high’’ position. The device was retrieved immediately and replaced with a commercially available filter rather than an investigational device at the discretion of the investigator, who was not aware of the option of implanting another investigational device. In the second case, with an infrarenal IVC length of 9 cm, there were two attempts at implanting a filter, and, in each attempt, the filter was inadvertently moved caudally when the delivery system was withdrawn without fluoroscopic guidance. After the second attempt, a commercially available filter was implanted. Neither case resulted in any device-related or procedure-related adverse events. The second technical failure was also considered a device malfunction. Case analysis revealed that the filter may have moved as a result of a component on the delivery

614

’

Safety and Effectiveness of the Retrievable IVC Filter

Smouse et al

’

JVIR

Table 2 . Baseline Characteristics, Medication Use, and Thrombotic Risk Factors Variable

Filter Retained (n ¼ 70)n

Filter Retrieved (n ¼ 53)

All Patients (N ¼ 125)

62.6 ⫾ 17.1

55.5 ⫾ 16.5

59.1 ⫾ 17.2

27–99

19–86

19–99

Baseline demographics Age (y) Mean ⫾ SD Range BMI (calculated) Mean ⫾ SD

30.3 ⫾ 8.1

32.9 ⫾ 10.0

31.6 ⫾ 9.1

Range

17.1–57.6

19.1–73.6

17.1–73.6

Male sex White race

40 (57.1) 55 (78.6)

32 (60.4) 42 (79.2)

73 (58.4) 99 (79.2)

Baseline antithrombotic medication†

45 (64.3)

25 (47.2)

72 (57.6)

Warfarin Unfractionated heparin

15 (21.4) 13 (18.6)

6 (11.3) 8 (15.1)

22 (17.6) 22 (17.6)

Low molecular weight heparin

24 (34.3)

10 (18.9)

34 (27.2)

9 (12.9) 2 (2.9)

7 (13.2) 1 (1.9)

16 (12.8) 3 (2.4)

Neurological Cardiovascular

29 (41.4) 60 (85.7)

22 (41.5) 43 (81.1)

53 (42.4) 105 (84.0)

Pulmonary

49 (70.0)

31 (58.5)

82 (65.6)

Gastrointestinal Genitourinary

48 (68.6) 41 (58.6)

32 (60.4) 28 (52.8)

82 (65.6) 71 (56.8)

Aspirin Clopidogrel Medical history

Musculoskeletal

50 (71.4)

38 (71.7)

90 (72.0)

Dermatologic Other

17 (24.3) 59 (84.3)

15 (28.3) 42 (79.2)

33 (26.4) 103 (82.4)

Active DVT History of DVT

47 (67.1) 38 (54.3)

25 (47.2) 23 (43.4)

73 (58.4) 62 (49.6)

History of PE

27 (38.6)

18 (34.0)

46 (36.8) 58 (46.4)

Thromboembolic disease

Thrombotic risk factors Inactivity/immobilization

34 (48.6)

23 (43.4)

Anticoagulation contraindicated

27 (38.6)

20 (37.7)

47 (37.6)

Recent surgery Malignancy

15 (21.4) 29 (41.4)

25 (47.2) 10 (18.9)

40 (32.0) 39 (31.2)

Recent trauma Hypercoagulability disorder Respiratory/cardiac failure Hormonal therapy Anticoagulation noncompliance

7 (10.0)

14 (26.4)

22 (17.6)

13 (18.6) 12 (17.1)

6 (11.3) 4 (7.5)

20 (16.0) 16 (12.8)

3 (4.3) 1 (1.4)

3 (5.7) 2 (3.8)

6 (4.8) 3 (2.4)

Postpartum state

0

1 (1.9)

1 (0.8)

Other

7 (10.0)

9 (17.0)

16 (12.8)

Values in parentheses are percentages. BMI ¼ body mass index, DVT ¼ deep vein thrombosis, PE ¼ pulmonary embolism, SD ¼ standard deviation. n Patients who did not have their filters retrieved through the trial cutoff at the 180-day visit, or at the time of death or withdrawal from the trial before 180 days, not including the two patients with technical failure, for whom no investigational filter remained implanted at the initial procedure and there was no filter indwell time. † Medication use within 7 days before device implantation.

catheter used to align the filter onto the catheter in manufacturing that, during the withdrawal of the catheter, could have engaged with the filter. Within the course of the trial, this component was removed from the design of the delivery catheter. The mean procedure time for all enrolled patients was 22 minutes ⫾ 12, and the mean time for deployment was 5 minutes ⫾ 5, with the approach by femoral access in

84.8% and by jugular access in 15.2% of cases. The mean infrarenal IVC diameter was 21.4 mm ⫾ 2.8, and no patient who received a filter had an IVC diameter outside the prespecified range of 17–28 mm; the mean infrarenal IVC length was 10.9 cm ⫾ 1.7. Two patients had successful filter placement with the use of intravascular US solely. Table 4 presents details for the implantation procedures, with a breakout for the patients who eventually underwent

Volume 24

’

Number 5

’

May

’

2013

615

Table 3 . Indications Identified as Primary for Filter Placement Filter Retained (n ¼ 70)n

Filter Retrieved (n ¼ 53)

20 (28.6)

24 (45.5)

45 (36.0)

Abdominal surgery Bariatric surgery

7 (10.0) 3 (4.3)

5 (9.4) 3 (5.7)

12 (9.6) 7 (5.6)

Orthopedic surgery

6 (8.6)

11 (20.8)

17 (13.6)

4 (2.0) 14 (20.0)

5 (9.4) 4 (7.3)

9 (7.2) 18 (14.4)

Primary Indication Surgery with PE risk

Other surgery Anticoagulation contraindicated Anticoagulation noncompliance

0

All Patients (N ¼ 125)

1 (1.8)

1 (0.8)

4 (5.7) 32 (45.7)

12 (21.8) 12 (21.8)

18 (14.4) 44 (35.2)

12 (17.1)

7 (13.2)

19 (15.2)

3 (4.3) 7 (10.0)

0 3 (5.7)

3 (2.4) 10 (8.0)

Unable to receive anticoagulation

4 (2.0)

1 (1.8)

5 (4.0)

Cancer status (ongoing) Recurrent PE despite adequate anticoagulation

5 (7.1) 1 (1.4)

0 0

5 (4.0) 1 (0.8)

Recent trauma Other risk DVT Coagulopathy Current PE/pulmonary thrombus

Surgical risk

0

4 1 Risk factor/indication

60 (85.7)

1 (1.8)

1 (0.8)

47 (85.5)

107 (85.6)

DVT ¼ deep vein thrombosis, PE ¼ pulmonary embolism. Patients who did not have their filters retrieved through the trial cutoff at the 180-day visit, or at the time of death or withdrawal from the trial before 180 days, not including the two patients with technical failure, for whom no investigational filter remained implanted at the initial procedure and there was no filter indwell time. n

Table 4 . Filter Implantation Procedure Details Variable

Filter Retained (n ¼ 70)n

Filter Retrieved (n ¼ 53)

All Patients (N ¼ 125)

23 ⫾ 13

19 ⫾ 9

22 ⫾ 12

3–83

2–44

2–83

5⫾5

4⫾4

5⫾5

1–30

0–26

o 1 to 30

21.6 ⫾ 2.7

21.1 ⫾ 2.9

21.4 ⫾ 2.8

17–28

17–28

17–28

Overall procedure time (min) Mean ⫾ SD Range Filter deployment time (min) Mean ⫾ SD Range Infrarenal IVC diameter (mm) Mean ⫾ SD Range Infrarenal IVC length (cm) Mean ⫾ SD

10.6 ⫾ 1.6

11.2 ⫾ 1.7

10.9 ⫾ 1.7

Range Jugular approach

7–15 11 (15.2)

7–17 8 (15.1)

7–17 19 (15.2)

Femoral approach

59 (84.3)

45 (84.9)

106 (84.8)

IVC patent Technical success

70 (100) 68 (97.1)

53 (100) 53 (100)

125 (100) 123 (98.4)

IVC ¼ inferior vena cava, SD ¼ standard deviation. Patients who did not have their filters retrieved through the trial cutoff at the 180-day visit, or at the time of death or withdrawal from the trial before 180 days, not including the two patients with technical failure, for whom no investigational filter remained implanted at the initial procedure and there was no filter indwell time. n

successful filter retrieval compared with those who retained their filters; there were no notable differences between these two groups.

Filter Retrieval Through the course of the trial to the 180-day follow-up visits, there were 60 retrieval evaluation procedures with vena cavograms performed. Retrieval was attempted in 54 of 123

cases (43.9%) and was successful in 53 (98.1%), with 19 of the 22 study sites (86.4%) performing at least one retrieval procedure. Investigators documented multiple indications for filter retrieval; the primary ones were resolution of the baseline condition in 37 patients (68.5%), treatment with anticoagulants in 24 patients (44.4%), a negative duplex US finding in 21 patients (38.9%), and the absence of acute swelling of the legs in 21 patients (38.9%). The mean time

616

’

Safety and Effectiveness of the Retrievable IVC Filter

Smouse et al

’

JVIR

Figure 4. Distribution of patients by time to successful retrieval after filter implantation. During the actual trial period, through the 180-day follow-up visit, retrieval was attempted in 54 patients and was successful in 53, with the one unsuccessful attempt occurring on day 167 after implantation. Beyond the actual trial follow-up period, through day 584 after implantation, there were an additional six retrieval attempts, all of which were successful.

from filter implantation to attempted retrieval during the trial timeframe was 84.6 days ⫾ 57.6 (range, 6–190 d; median, 78 d), with the one unsuccessful retrieval attempt occurring at day 167 after implantation (Fig 4). The mean procedure time for the 54 retrieval attempts was 24 minutes ⫾ 15 (range, 4–79 min), and the mean time for the actual retrieval (from venous insertion of the snare to removal of the filter) was 7 minutes ⫾ 8 (range, 1–48 min), with the approach by femoral access in 38 cases (70.4%) and by jugular access in 16 (29.6%). In all patients who underwent retrieval attempts, there were no signs of new DVT. The ease of retrieval was not found to differ between the femoral and jugular approaches, although the sample sizes are small. No cases of PE were reported before any retrieval evaluation, and no signs of filter migration or fracture were observed. There were two patients with continuing DVT, identified at earlier follow-up visits, who had successful filter retrievals (one patient was receiving aspirin and one was receiving enoxaparin at the time of retrieval). Among the 53 patients who had their filters retrieved, there were no deaths. Six patients underwent retrieval evaluations but did not have retrieval attempted. The reasons cited were thrombus in or near the filter in five patients and the presence of iliocaval DVT in one patient. In three of these cases, later reevaluations were scheduled to allow for thrombus resolution, and two of those planned reevaluations within the trial resulted in successful filter retrievals (at postimplantation days 82 and 129). After each filter retrieval, the investigator visually inspected the filter for thrombus and device integrity, and, in all cases, the filter and web were intact with no evidence of fracture or loss of device integrity. Minimal strands of thrombus, averaging 2  6 mm in size, were observed on seven of the 53 explanted filters (13.2%). None had been detected on preretrieval imaging, and all were considered clinically insignificant. At the 30-day postretrieval evaluations undergone by 52 of the 53 patients (one patient withdrew from the trial shortly after successful retrieval), there were no new events or unanticipated findings,

including clinical signs of new DVT or PE. The two patients with continuing DVT at the time of filter retrieval had persistent DVT at the 30-day postretrieval evaluation. The one case of retrieval failure was in a 36-year-old woman with a history of PE and DVT along with other comorbidities. The filter had been implanted 167 days earlier because of ongoing surgical procedures secondary to trauma. Via a femoral approach, the retrieval hook was captured with a snare, but then the 10-F outer sheath could not be advanced over the filter without the use of unreasonable force, and the investigator chose to leave the filter in place. Postprocedure imaging showed the filter remaining in an appropriate position and orientation. There were no adverse clinical sequelae and no further interventions required or attempted. The patient exited the trial per protocol after the 180-day visit, which included an x-ray image demonstrating an intact filter. In a separate retrieval case, 118 days after implantation, the study site reported a finding of self-limiting pseudoaneurysm. The filter was intact when removed, but the postretrieval vena cavogram revealed an outpouching, approximately 1 cm in size, from the right side of the IVC at the previous position of the filter. No free extravasation of contrast medium was noted, but some spasm was seen, with vessel narrowing of approximately 35%. With repeat vena cavograms at 10 and 15 minutes, the outpouching became smaller, there was less narrowing of the IVC, and the patient was discharged the same day without hospitalization. There were no clinical sequelae through the 30-day postretrieval follow-up. Beyond the 180-day follow-up visits that formally ended the trial, there were six additional attempts and successful filter retrievals. These occurred at days 212, 230, 260, 279, 311, and 584 after implantation. The retrieval at day 311 occurred in one of the three patients who had been found to have new PE within 30 days after filter implantation.

Primary and Secondary Endpoint Results Of the 125 patients enrolled in the trial, five did not meet the criteria for clinical success (three with definite PE and

Volume 24

’

Number 5

’

May

’

2013

617

Table 5 . Primary Endpoint Results for the RETRIEVE Trial Filter Retained (n ¼ 70)n

Filter Retrieved (n ¼ 53)

–

–

123 (98.4)

3/61 (4.9) 0

0/38 0

3/102 (2.9) 0

Device-related AE requiring interventiony

0

0

0

Clinical success composite99 One-sided lower limit of 95% CI (%)

– –

– –

120 (96.0) 91.8

Endpoint Technical success† New PE‡ Filter migrationy

All Patients (N ¼ 125)

Values in parentheses are percentages. AE ¼ adverse event, CI ¼ confidence interval, PE ¼ pulmonary embolism, RETRIEVE ¼ Crux Biomedical Evaluation of the Crux Inferior Vena Cava Filter System. n Patients who did not have their filters retrieved through the trial cutoff at the 180-day visit, or at the time of death or withdrawal from the trial before 180 days, not including the two patients with technical failure, for whom no investigational filter remained implanted at the initial procedure and there was no filter indwell time. † The two technical failures occurred when the study device was implanted but immediately retrieved because of investigator dissatisfaction with filter position; in each case (different trial sites), the investigator chose to implant an alternative commercially available filter. ‡ All three new PE cases were observed on imaging at the 30-day follow-up visit. y No events were observed at any follow-up visit. 99 The components of clinical success were technical success and freedom from PE, filter migration, and device-related adverse events requiring intervention. Of the 125 patients in the full cohort, two were technical failures (no filter implant after enrollment), and three experienced nonfatal PE (all three of these patients had filters still in place at 180 d, whereas one of these three patients eventually underwent successful retrieval at postimplantation day 311, after completion of the trial).

two with technical deployment failures; Table 5). There were no cases of device migration of more than 20 mm, and there were no device-related adverse events requiring intervention. The rate of clinical success was therefore 96.0% (120 of 125), with the one-sided lower limit of the 95% CI being 91.8% and therefore meeting the performance objective of at least 80%. The two cases of technical failure have been described here earlier. The three cases of definite PE were all documented between baseline and the 30-day visit. One of these three patients was an 87-year-old man with a history of PE and DVT, with DVT present at baseline. The patient had PE at day 16 after implantation, 5 days after a planned surgical procedure. The PE was diagnosed by a perfusion lung scan and treated with anticoagulant agents. The PE resolved 17 days later, and the patient continued in the trial through the 180-day follow-up visit with no further devicerelated adverse events and no filter migration or fracture. A second PE occurred in a 73-year-old woman with a history of DVT and intracranial bleeding. She had DVT at baseline. At 25 days after implantation, a PE was diagnosed by a chest CT that had been ordered because of shortness of breath. An abdominal CT also demonstrated iliocaval thrombus extending to the level of the lower portion of the filter. The patient was provided oxygen therapy and pain medication for the PE. No ventilator was required, and she was not a candidate for thrombolytic therapy as a result of a recent hemorrhagic stroke. The PE resolved at 41 days after implantation, and the patient continued in the trial through the 180-day visit with no further device-related adverse events and no filter migration or fracture. The third definite PE occurred in a 32-year-old man with a history of multiple recent PEs (including 4 d before trial enrollment) and DVT at baseline. He had a planned surgical procedure at 3 days

after implantation, and then, at 5 days after implantation, he was diagnosed by CT with a PE following syncope and bradycardia, indicating massive PE (23). The CT revealed thrombus at the level of the filter, which was nonocclusive and asymptomatic, and a large saddle embolus into the right and left pulmonary arteries. In addition, there was a caudal shift of the filter that did not meet the definition of migration per DSMB review. After several surgical interventions for his underlying disease, the patient exited the trial after the 180-day visit with the filter in place. There had been no further filter movement, and there was no filter fracture. The investigator later reported that the filter had been successfully retrieved at 311 days after implantation. During the adverse event review process described here later, two cases of suspected PE were identified. The first case was in a 58-year-old man with massive PE diagnosed 2 days before filter implantation. At 103 days after implantation, he experienced chest pain and shortness of breath, which was relieved by oxygen. Chest radiography and electrocardiography findings were normal. Because no CT or lung perfusion scan was performed, the occurrence of a PE could not be ruled out. The second case of suspected PE was in a 72-year-old man with a history of PE and mesothelioma and a contraindication to anticoagulant agent use as a result of scheduled surgery. At 2 days after surgery and index filter implantation, following an episode of hypotension and tachycardia, the patient was diagnosed by chest radiography with a pleural effusion. The following day, respiratory failure was diagnosed. Again, because no CT or lung perfusion scan was performed, this event was categorized as a suspected PE. The patient recovered and was discharged with a prescription for anticoagulation at 9 days after filter implantation.

618

’

Safety and Effectiveness of the Retrievable IVC Filter

Smouse et al

’

JVIR

Table 6 . Secondary Endpoint Results for the RETRIEVE Trial Filter Retained (n ¼ 70)n

Filter Retrieved (n ¼ 53)

All Patients (N ¼ 125)

30-d visit 90-d visit

5/61 (8.2) 1/53 (1.9)

1/38 (2.6) 2/16 (12.5)

6/102 (5.9) 3/69 (4.3)

180-d visit

3/49 (6.1)

–

3/49 (6.1)

0/4

1/2 (50.0)

1/6 (16.7)

30-d visit

3/61 (4.9)

0/38

3/102 (2.9)

90-d visit 180-d visit

0/53 0/49

0/16 –

0/69 0/49

Endpoint New lower-limb DVT

Retrieval evaluation† New IVC thrombus in/near filter

Retrieval evaluation†

4/4 (100)

1/2 (50.0)

5/6 (83.3)

Retrieval Filter embolization‡

– 0

1/53 (1.9) 0

1/53 (1.9) 0

Filter migration‡

0

0

0

Filter fracture‡

0

0

0

Values in parentheses are percentages. DVT ¼ deep vein thrombosis, IVC ¼ inferior vena cava, RETRIEVE ¼ Crux Biomedical Evaluation of the Crux Inferior Vena Cava Filter System. n Patients who did not have their filters retrieved through the trial cutoff at the 180-day visit, or at the time of death or withdrawal from the trial before 180 days, not including the two patients with technical failure, for whom no investigational filter remained implanted at the initial procedure and there was no filter indwell time. † At retrieval evaluation only, in cases with no attempt to retrieve the filter. In six instances, patients underwent retrieval evaluations but did not then have retrieval attempted; two of those six patients did eventually undergo successful retrievals. ‡ No events were observed at any follow-up visit.

There were 13 cases of DVT diagnosed during the course of the trial in 12 of the 125 enrolled patients (9.6%; Table 6). All were diagnosed with duplex US except for one iliocaval DVT diagnosed based on a vena cavogram at a retrieval evaluation (duplex US findings had been negative). In one patient, DVT was diagnosed before the 30-day visit, was resolved after anticoagulant therapy, but then recurred before the 180-day visit. There were nine cases in which thrombus was observed in or near the filter (Table 6). The thrombus was nonocclusive and asymptomatic in all cases. The thrombus was detected on retrieval evaluation vena cavograms in six cases, during CT scans for PE symptoms in two cases, and during CT for cancer management in one case. Seven of these nine patients (77.8%) had active DVT or PE at baseline, and the remaining two (22.2%) had family or patient history of VTE. At the 180-day follow-up visits, six of the nine patients (66.7%) were receiving antithrombotic medications.

Adverse Events A total of 275 adverse events were reported, most related to concomitant medical conditions, and 98 (35.6%) of those adverse events were considered serious. Five of the serious adverse events (in four patients) were categorized as device-related or potentially device-related: two of the cases of definite PE, the one asymptomatic case of thrombus at the level of the filter coincident with PE, and the two cases of suspected PE. There were 14 deaths during the trial follow-up period, none of which were related to the investigational device or

to a new PE. Among the six patients who withdrew during the course of the trial, three cited deteriorating health, and none reported a relationship to the investigational device or procedure as the cause.

Post Hoc Analyses The first of two post hoc sensitivity analyses addressed the impact of including in the denominator for the primary clinical success endpoint the 22 patients who did not complete the trial per protocol and thus had incomplete data. When these patients (14 deaths, two patients lost to follow-up, and six patients who withdrew from the trial) were excluded from the analysis, the rate of clinical success was 95.1% (98 of 103), with the one-sided lower limit of the 95% CI being 90.1%, which met the performance objective of 80% or greater. The second sensitivity analysis addressed the impact of including in the numerator for the clinical success endpoint the two cases of suspected PE. As described earlier, these cases did not meet the protocol definition for definite PE in that they lacked confirmatory imaging but, in the subsequent adverseevent review, they were adjudicated as suspected PE. When these two cases were included in the analysis, the rate of clinical success was 94.4% (118 of 125), with the one-sided lower limit of the 95% CI being 89.7%, which met the performance objective of 80% or greater. In another post hoc analysis, when all enrolled patients were categorized according to the 2011 Quality Improvement Guidelines for IVC filter placement from the SIR Standards of Practice Committee (Table 1) (19), 57.6% (72

Volume 24

’

Number 5

’

May

’

2013

of 125) were determined to have had therapeutic indications and 42.4% (53 of 125) were determined to have had prophylactic indications. The patients with prophylactic indications had a higher rate of filter retrievals within 6 months of follow-up than the patients with therapeutic indications (53.8% vs 36.6%).

DISCUSSION Through 180 days of follow-up after implantation of the Crux filter, the composite primary endpoint of clinical success was achieved in 96.0% of all enrolled patients (120 of 125), with the one-sided lower limit of the 95% CI being 91.8%, which met the primary trial performance objective of 80% or greater. The five failures on the composite clinical success endpoint were in the three patients with definite PE and the two cases counted as technical failures because the study device was immediately retrieved as a result of investigator dissatisfaction with filter position. The 98.4% rate of implant technical success exceeded the 97% threshold value established by the SIR Standards of Practice Committee (19). The three cases of definite PE all occurred in patients with complicated medical histories and multiple thrombotic risk factors. All three exited the trial after the full term of follow-up with their filters intact. In fact, one of the three underwent successful filter retrieval at postimplantation day 311, well beyond the completion of the trial follow-up. The 2.4% rate of definite PE in the RETRIEVE trial was within the established threshold value of 5% (24). Other similar filter studies have recently reported rates of definite PE ranging from 1% to 6% (11–14). In the RETRIEVE trial, there were two additional cases (without imaging) adjudicated by the independent reviewer as suspicious for PE, but, even with these counted, the revised 4.0% rate of PE (five of 125) was still below the 5% threshold value, and the rate of clinical success was recalculated in the post hoc analysis as 94.4%, with the one-sided lower limit of the 95% CI being 89.7%, which also still met the performance objective of 80% or greater. There were no instances of filter migration, embolization, fracture, or tilting in the present study. There were 13 new DVT cases diagnosed, and there were nine observations of new thrombus in the IVC in or near the filter, all of which were nonocclusive and asymptomatic. This incidence of IVC thrombus was well below the previously suggested thresholds and compares well with the outcome ranges for other devices (7,8,13,14). Future studies will be needed to provide more direct comparisons among device types with the use of consistent definitions and protocols. In the RETRIEVE trial, the overall rate of retrieval success was 98.1%, with 70.4% of the retrievals achieved by the femoral route. Although the ease of retrieval was not found to differ between jugular and femoral approaches, the relatively small sample size of jugular retrievals limits analysis. Expanded study will be required to determine if the cranial orientation of the device anchoring hooks

619

negatively impacts retrieval from the jugular route. The 43.2% rate of retrieval attempts in enrolled patients through the 6 months of the trial is at the high end of the range reported for seven different retrievable filters in a recently published systematic device review (8). In August 2010, an FDA ‘‘initial communication’’ expressed concern that various adverse events could be related to retrievable filters being retained for long periods, beyond the time when the risk of PE has subsided (17). For example, in a multicenter study by the American Association for the Surgery of Trauma (25), 446 patients received retrievable IVC filters, with the majority (76%) placed for prophylactic indications, and the retrieval rate was 22%, with half the patients having documented postimplantation follow-up. In the RETRIEVE trial, 53 of the 123 implanted filters (43.1%) were retrieved within 6 months of trial followup, at a mean time from filter implantation of 83.0 days. Just more than 60% of the retrieval attempts occurred within the first 3 months after implantation (the mean time to retrieval was just less than 3 mo), whereas the remainder were distributed more or less evenly over time to well beyond the 6-month trial cutoff. Taking into account the six additional successful retrievals that occurred beyond the 6month trial window, as late as 584 days after implantation, the overall retrieval rate was increased to 48.0% (59 of 123) for the entire trial cohort. A majority of the RETRIEVE trial patients had VTE at the time of enrollment. However, a substantial number of patients underwent filter implantation for temporary indications (eg, surgery with PE risk, recent trauma) classifiable as prophylactic per the SIR guidelines (19). Post hoc analysis determined that, of the 125 enrolled patients, 72 (57.6%) had therapeutic indications (documented thromboembolic disease), whereas 53 (42.4%) had prophylactic indications. As expected, those patients with prophylactic indications had a higher rate of retrieval within 6 months of follow-up compared with patients with therapeutic indications (53.8% vs 36.6%, respectively). Limitations of the present study include the nonrandomized single-arm design, the follow-up extending to only 6 months, and the potential for biases associated with the manufacturer funding of the regulatory device trial. With the reported data having been acquired in only 125 patients, it is possible that more clinical complications or technical failures (at deployment or retrieval) could be revealed as the use of the device expands. Because the trial was designed to assess clinically significant events, with pulmonary or vena cava angiography or CT scans performed only when clinically indicated, more extensive study will be required to assess the potential for subclinical events related to use of the device. Regarding the trial endpoint definitions, ad hoc analysis was required to adjust for events that were eventually adjudicated as suspicious for PE but that lacked imaging confirmation. Another shortcoming in terms of consistency with reporting standards was that symptomatic caval thrombus was not included as a component of the primary clinical success

620

’

Safety and Effectiveness of the Retrievable IVC Filter

endpoint, although in fact no such instances were reported. In addition, ad hoc analysis was also required to categorize the enrolled patients in terms of the recently established distinction between therapeutic and prophylactic indications for placement of retrievable filters, and then to determine the relative impact of the indications on filter retrieval rates. The investigators acknowledge the uncertainty involved in adjudicating the device relatedness of caval thrombus. Ongoing investigation is warranted concerning the role of the new nonconical design of the device in inhibiting and/or promoting thrombus generation, and concerning the potential risk of entrapment of devices that are passed through the radiolucent thrombus-trapping ePTFE web. Particularly in light of the FDA expression of concern regarding the potential for adverse events related to retrievable filters being retained for long periods (17), further study of the device in a larger cohort with longer dwell times will be important to optimize patient safety. Before initiation of the international RETRIEVE trial, a previous design of the Crux filter was evaluated (RETRIEVE 1). Enrollment in the RETRIEVE 1 trial was suspended after treatment of 88 patients as a result of separate instances of embolization (n ¼ 3) and filter fractures (n ¼ 3). The filter was redesigned to address IVC wall fixation and device integrity. The results of the present study provide a preliminary indication that the issues have been addressed. Ongoing data collection and further, largescale trials are now needed to determine the long-term durability of the device and to ensure optimal patient safety. In summary, in the present cohort of 125 patients with temporary or permanent risk of PE and contraindication to or failure of anticoagulation, the implantation and retrieval of the Crux vena cava filter was technically and clinically successful, within the effectiveness and safety thresholds suggested by SIR, and with complication rates well within the ranges reported for other retrievable IVC filters.

ACKNOWLEDGMENTS The authors acknowledge the contribution of Griff Tully, MD (Menlo Park, California), an independent medical reviewer experienced in clinical trials and regulatory requirements, who served as the independent medical monitor for the trial and was responsible for the review and validation of all reported adverse events that were considered serious and/or device-related and/or procedurerelated. The authors also acknowledge the contribution of Thomas Kinney, MD, University of California, San Diego, a board-certified radiologist and member of the DSMB, who was also retained to perform overreading of radiographic images if and when follow-up examinations noted any device changes. T.C. served as the independent biostatistician for analysis of primary and secondary endpoints and overall descriptive data review. The DSMB

Smouse et al

’

JVIR

consisted of four practicing physicians knowledgeable about vena cava filters: Thomas Kinney, MD; John Kaufman, MD; Christopher Morris, MD; and Robert McLafferty, MD. Editorial and administrative support for the trial report manuscript was provided by representatives of Galen Press and paid for by the trial sponsor, Crux Biomedical (Menlo Park, California).

APPENDIX The following physicians and institutions participated in the Crux Biomedical Evaluation of the Crux Inferior Vena Cava Filter System trial: Robert Mendes, MD, and Houman Tamaddon, MD, University Hospital, Augusta, Georgia; Marc Bosiers, MD, A.Z. St. Blasius Hospital, Dendermonde, Belgium; Anthony Comerota, MD, Jobst Vascular Institute/University of Toledo, Toledo, Ohio; Andrew Holden, MD, Auckland City Hospital, Auckland, New Zealand; Thuong G. Van Ha, MD, University of Chicago Medical Center, Chicago, Illinois; Joseph J. Fulton, MD, University of North Carolina Medical Center, Chapel Hill, North Carolina; John R. Laird, MD, University of California, Davis, Medical Center, Sacramento, California; John D. Martin, MD, Anne Arundel Medical Center, Annapolis, Maryland; Kwame Amankwah, MD, Upstate Medical University Hospital, Syracuse, New York; Rabih A. Chaer, MD, University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania; John Clouston, MD, Royal Brisbane Hospital, Brisbane, Australia; Ken Thomson, MD, Alfred Hospital, Melbourne, Australia; H. Bob Smouse, MD, St. Francis Hospital, Peoria, Illinois; Mark J. Garcia, MD, Christiana Care Medical Center, Newark, Delaware; Allan Herr, MD, Albany Medical Center Hospital, Albany, New York; Kurt Openshaw, MD, St. Joseph’s Hospital of Orange, Orange, California; Patrick Peeters, MD, Imelda Hospital, Bonheiden, Belgium; Wael Saad, MD, University of Virginia Health System, Charlottesville, Virginia; Robert J. Feezor, MD, University of Florida Medical Center, Gainesville, Florida; Frank C. Lynch, MD, Milton S. Hershey Medical Center, Hershey, Pennsylvania; John S. Lane, MD, University of California, Irvine, Medical Center, Irvine, California; and David Rosenthal, MD, Atlanta Medical Center, Atlanta, Georgia.

REFERENCES 1. 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. 2. 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. 3. Spencer FA, Bates SM, Goldberg RJ, et al. A population-based study of inferior vena cava filters in patients with acute venous thromboembolism. Arch Intern Med 2010; 170:1456–1462. 4. Kinney TB. Update on inferior vena cava filters. J Vasc Interv Radiol 2003; 14:425–440.

Volume 24

’

Number 5

’

May

’

2013

5. Chung J, Owen RJ. Using inferior vena cava filters to prevent pulmonary embolism. Can Fam Physician 2008; 54:49–55. 6. Kaufman JA, Kinney TB, Streiff MB, et al. Guidelines for the use of retrievable and convertible vena cava filters: report from the Society of Interventional Radiology multidisciplinary consensus conference. J Vasc Interv Radiol 2006; 17:449–459. 7. Martin MJ, Blair KS, Curry TK, Singh N. Vena cava filters: current concepts and controversies for the surgeon. Curr Probl Surg 2010; 47:524–618. 8. Angel LF, Tapson V, Galgon RE, Restrepo MI, Kaufman J. Systematic review of the use of retrievable inferior vena cava filters. J Vasc Interv Radiol 2011; 22:1522–1530. 9. Kaufman JA. Optional vena cava filters: what, why, and when. Vascular 2007; 15:304–313. 10. Kaufman JA, Rundback JH, Kee ST, et al. Development of a research agenda for inferior vena cava filters: proceedings from a multidisciplinary research consensus panel. J Vasc Interv Radiol 2009; 20:697–707. 11. Johnson MS, Nemcek AA Jr, Benenati JF, et al. The safety and effectiveness of the retrievable Option inferior vena cava filter: a United States prospective multicenter clinical study. J Vasc Interv Radiol 2010; 21:1173–1184. 12. Smouse HB, Rosenthal D, Van Ha T, et al. Long-term retrieval success rate profile for the Gunther Tulip vena cava filter. J Vasc Interv Radiol 2009; 20:871–877. 13. Lyon SM, Riojas GE, Uberoi R, et al. Short- and long-term retrievability of the Celect vena cave filter: results from a multi-institutional registry. J Vasc Interv Radiol 2009; 20:1441–1448. 14. Binkert CA, Drooz AT, Caridi JG, et al. Technical success and safety of retrieval of the G2 filter in a prospective, multicenter study. J Vasc Interv Radiol 2009; 20:1449–1453. 15. Nicholson W, Nicholson WJ, Tolerico P, et al. Prevalence of fracture and fragment embolization of Bard retrievable vena cava filters and clinical implications including cardiac perforation and tamponade. Arch Intern Med 2010; 170:1827–1831. 16. Cipolla J, Weger NS, Sharma R, et al. Complications of vena cava filters: a comprehensive clinical review. Opus 12 Scientist 2008; 2:11–24.

621

17. US Food and Drug Administration. Inferior vena cava (IVC) filters: initial communication: risk of adverse events with long term use. Posted August 9, 2010. Available at http://www.fda.gov/Safety/MedWatch/ SafetyInformation/SafetyAlertsforHumanMedicalProducts/ucm221707. htm. Accessed May 31, 2012. 18. Ray CE Jr, Mitchell E, Zipser S, Kao EY, Brown CF, Moneta GL. Outcomes with retrievable inferior vena cava filters: a multicenter study. J Vasc Interv Radiol 2006; 17:1595–1604. 19. Caplin DM, Nikolic B, Kalva S, et al. Quality improvement guidelines for the performance of inferior vena cava filter placement for the prevention of pulmonary embolism. J Vasc Interv Radiol 2011; 22: 1499–1506. 20. Participants in the Vena Caval Filter Consensus Conference. Recommended reporting standards for vena caval filter placement and patient follow-up. J Vasc Interv Radiol 2003; 14(suppl):S427–S432. 21. Millward SF, Grassi CJ, Kinney TB, et al. Reporting standards for inferior vena caval filter placement and patient follow-up: supplement for temporary and retrievable/optional filters. J Vasc Interv Radiol 2009; 20(suppl):S374–S376. 22. Sacks D, McClenny TE, Cardella JF, Lewis CA. Society of Interventional Radiology clinical practice guidelines. J Vasc Interv Radiol 2003; 14(suppl):S199–S202. 23. Jaff MR, McMurtry MS, Archer SL, et al. Management of massive and submassive pulmonary embolism, iliofemoral deep vein thrombosis, and chronic thromboembolic pulmonary hypertension: a scientific statement from the American Heart Association. Circulation 2011; 123: 1788–1830. 24. Grassi CJ, Swan TL, Cardella JF, et al. Quality improvement guidelines for percutaneous permanent inferior vena cava filter placement for the prevention of pulmonary embolism. J Vasc Interv Radiol 2003; 14(suppl): S271–S275. 25. Karmy-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; 62:17–24.