Abstract
Background
A clear indication and strategy for placement of retrievable inferior vena cava filters (IVCFs) have not been established. This study was designed to evaluate the efficacy and disadvantages of the retrievable IVCF use particularly in venous thromboembolism (VTE) patients with malignancy.
Hypothesis
Retrievable IVCFs might be safe and useful in VTE patients with malignancy.
Methods
The study population consisted of 56 consecutive patients undergoing IVCF placement at our institution from January 1, 2008 to December 31, 2011. Prognostic data were retrospectively reviewed in April 2013.
Results
Mean follow‐up period was 584.6 (range, 1–1857) days. Twenty‐six of the 56 patients had a malignancy. In 16 of the 30 patients without malignancy, the filter was retrieved, whereas the other 14 patients eventually received permanent implantation. There was no significant difference in the survival rate between the retrieval group and the nonretrieval group in the nonmalignancy patients (1‐year survival rates, 94% vs 85%). In patients with malignancy, the nonretrieval group showed a significantly lower survival rate (P < 0.01). The 1‐year and 2‐year survival rates were 100% vs 46% and 100% vs 18%, respectively. There was no medical record of pulmonary thromboembolism occurrence or recurrence. All deaths in the patients with malignancy were malignancy related. In 4 of 5 malignancy patients who could undergo tumor resection surgery, adequate thrombus regression enabled us to retrieve the IVCF after surgery.
Conclusions
Permanent use of a retrievable IVCF is relatively safe in short‐ or midterm follow‐up regardless of malignancy status. Retrievable filter use might be reasonable in malignancy patients.
Introduction
As the recognition of venous thromboembolism (VTE) has spread, the number of cases diagnosed as VTE has increased. Therefore, the number of cases in which the indication of inferior vena cava filter (IVCF) placement is discussed has also increased. However, the use of an IVCF, particularly a retrievable IVCF, does not have a clear strategy with respect to its placement. A dwelling time limit of 14 days was described for the placement of Gunther Tulip filters, and a long‐term retrieval success rate for Gunther Tulip IVCFs has been reported.1, 2 Gunther Tulip retrievable IVCFs were previously used in 39 patients with malignancy (and 79 patients without malignancy) for management of VTE, and there was no difference in outcomes in these groups.3 But clinical research regarding retrievable filter use in malignancy cases was not fully evaluated, even though VTE is frequently encountered in patients with malignancy.
Complications associated with long‐term implantation of an IVCF such as vena caval penetration and post‐thrombotic syndrome following filter occlusion have been reported.4 Durack et al reported that in patients with long‐term implantation of an IVCF of more than 71 days, almost all cases showed filter protrusion through the vena cava wall.5 Many cases with filter protrusion were identified in imaging tests, such as computed tomography (CT) scan or venography, in an asymptomatic state. Although severe bleeding, pancreatitis, abscess formation, vertebral osteomyelitis, and perforation into peripheral organs were reported,6, 7, 8, 9, 10, 11 these cases are rare. Considering these symptomatic complications, it is perhaps better to avoid permanent implantation of an IVCF if possible.
Some VTE cases associated with malignancy need IVCF implantation. In many such cases, an anticoagulant treatment is not fully effective for a hypercoagulable state, and thus filter retrieval has to be abandoned. This study was designed to evaluate the efficacy and disadvantages of retrievable IVCF use particularly in VTE patients with malignant tumor. We investigated the complications of IVCF accompanied by permanent use and the prognosis after IVCF implantation in VTE patients with or without malignancy.
Methods
Study Population and Data Collection
The study population consisted of 56 consecutive patients who underwent new IVCF placement at the Department of Cardiovascular Medicine, University of Tokyo Hospital, from January 1, 2008 to December 31, 2011. Clinical information regarding baseline demographics, deep vein thrombosis (DVT) site, pulmonary thromboembolism (PTE) severity, comorbidities, placement and retrieval of the IVCF, and imaging tests including CT scans and venography was extracted from medical records. Prognostic data were retrospectively reviewed in April 2013. The study protocol was approved by the institutional ethical committee of the University of Tokyo (#2650). No extramural funding was used to support this work.
Definition of Diseases and Endpoints
DVT site was divided into 4 categories, inferior vena cava, iliac vein, femoral vein, and popliteal vein, or more distal depending on the most proximal site of the venous thrombus. PTE severity was classified into 4 categories as none/nonmassive, submassive, massive, and collapse.12 We placed an IVCF in VTE patients with DVT located proximal to the knee and tried to retrieve an IVCF if proximal DVT resolved within dwelling limit. Malignancy was defined as the presence of any kind of malignant neoplasm including a primary lesion, recurrence, or metastasis that was confirmed or treated within 5 years before IVCF placement. Filter penetration was defined as >3‐mm protrusion out of the vena cava wall on CT scan.3
In some cases, filter exchange was performed in view of the dwelling limit. Filter dwelling duration was calculated from the first filter insertion. Survival time was defined as the time difference between the first filter insertion and death or the last follow‐up. The duration to the latest CT scan or the diagnosis of filter penetration was calculated from the last filter insertion.
Statistical Analyses
All statistical analyses were performed using SPSS software version 12.0 (SPSS Inc., Chicago, IL). Categorical data were summarized as count and percentage. Continuous data were summarized as the mean ± standard deviation. Univariate comparisons between the 2 groups were made with the Pearson χ2 test and Mann‐Whitney U test. The Tukey‐Kramer method was used for comparisons of means between more than 3 groups. A Kaplan‐Meier survival curve was constructed to demonstrate the difference in survival, and comparisons between groups were assessed with the log‐rank test. All analyses were 2‐sided, and a P value of <0.05 was considered as statistically significant.
Results
We performed IVCF placement in 56 patients; 26 of these 56 patients had malignancy. The baseline characteristics are shown in Table 1. There were 13 females (43.3%) in the nonmalignancy group and 15 (57.7%) in the malignancy group. There was no significant difference between the 2 groups (P = 0.28). The mean age was 59.2 ± 17.4 years in the nonmalignancy group and 66.6 ± 13.8 years in the malignancy group. The malignancy group tended to be older than the nonmalignancy group (P = 0.08). The mean follow‐up period was 584.6 days (range, 1–1857 days).
Table 1.
Baseline Characteristics
| Nonmalignancy | Malignancy | |||
|---|---|---|---|---|
| Retrieval, n = 16 | Nonretrieval ,n = 14 | Retrieval, n = 7 | Nonretrieval, n = 19 | |
| Gender, female, n (%) | 8 (50.0) | 5 (35.7) | 5 (71.4) | 10 (52.6) |
| Age, y | 52.1 ± 18.4 | 67.3 ± 12.4 | 65.7 ± 14.3 | 66.9 ± 14 |
| DVT site, n (%) | ||||
| Popliteal | 0 (0) | 0 (0.0) | 1 (14.3) | 0 (0.0) |
| Femoral | 12 (75) | 8 (57.1) | 4 (57.1) | 8 (42.1) |
| Iliac | 3 (18) | 5 (35.7) | 0 (0.0) | 7 (36.8) |
| Inferior vena cava | 1 (6.3) | 1 (7.1) | 2 (28.6) | 4 (21.1) |
| PTE severity, n (%) | ||||
| None/nonmassive | 7 (43.8) | 9 (64.3) | 5 (71.4) | 15 (78.9) |
| Submassive | 7 (43.8) | 5 (35.7) | 1 (14.3) | 4 (21.1) |
| Massive | 1 (6.3) | 0 (0.0) | 1 (14.3) | 0 (0.0) |
| Collapse | 1 (6.3) | 0 (0.0) | 0 (0.0) | 0 (0.0) |
| Anticoagulation, n (%) | 16 (100) | 11 (78.6) | 6 (85.7) | 16 (84.2) |
| Retrieval failure, n (%) | 0 (0) | 3 (21.4) | 0 (0.0) | 0 (0.0) |
| Temporary dwelling duration, d | 21.3 ± 9.6 | — | 14.6 ± 7.0 | — |
| IVC filter, n (%) | ||||
| Neuhouse | 0 (0.0) | 0 (0.0) | 1 (14.3) | 0 (0.0) |
| Gunther tulip | 14 (87.5) | 9 (64.3) | 5 (71.4) | 12 (63.2) |
| Optease | 1 (6.3) | 4 (28.6) | 1 (14.3) | 3 (15.8) |
| ALN | 1 (6.3) | 1 (7.1) | 0 (0.0) | 4 (21.1) |
Abbreviations: DVT, deep vein thrombosis; IVC, inferior vena cava; PTE, pulmonary thromboembolism.
Placement and Retrieval of the IVCF
The types of IVCF used are shown in Table 1. All filters were inserted percutaneously under fluoroscopic control through a jugular or femoral vein (see Supporting Table 1 in the online version of this article for reasons for placement of an IVCF, and Supporting Table 2 for reasons for not retrieving an IVCF). A retrievable IVCF was inserted in all patients with no malignancy. In 16 of the 30 nonmalignancy patients, the filter was ultimately retrieved, and the mean dwelling period was 21.3 ± 9.6 days. They underwent filter exchange 0 to 2 times (mean, 0.50 times). The other 14 patients underwent placement of a retrievable IVCF for permanent use as a result. In 3 of these 14 patients, although anticoagulant therapy ameliorated the venous thrombus, IVCF retrieval failed due to adhesion of the hook to the vena cava wall or filter penetration. They underwent filter exchange 0 to 1 time (mean, 0.43 times).
Table 2.
Types of Comorbidities in Nonmalignancy Patients
| Nephrotic syndrome | 2 |
| Intracerebral hemorrhage | 1 |
| Amyotrophic lateral sclerosis | 1 |
| Fontan circulation | 1 |
| Systemic sclerosis | 1 |
| Systemic lupus erythematosus | 1 |
| Pneumoniae | 1 |
| Benign spinal tumor | 1 |
| Pelvic fracture | 1 |
| Hip surgery | 1 |
| Achilles tendon surgery | 1 |
| None | 18 |
Twenty‐five of 26 patients with malignancy received retrievable IVCF insertion, whereas the other patient underwent temporary IVCF insertion. In 7 of the 26 patients, the filter was ultimately retrieved, and the mean dwelling period was 14.6 ± 7.0 days. They underwent filter exchange 0 to 1 time (mean, 0.14 times). In 4 of these 7 retrieval patients, filter insertion was performed before surgery because it was impossible to administer anticoagulants in the perioperative period. The causes of VTE in the other 3 patients were central vein catheter placement, long‐term compression at the puncture site in cardiac catheterization, and postoperative status for uterine cervical carcinoma. The other 19 patients eventually received permanent implantation. They underwent filter exchange 0 to 1 time (mean, 0.16 times). Only 1 of these 19 patients underwent filter insertion before surgery. In the malignancy patients, the IVCF was retrieved after surgery due to substantial thrombus regression in 4 of the 5 patients who could undergo tumor resection surgery.
Follow‐up After IVCF Permanent Implantation
Nonmalignancy Patients
The types of comorbidities believed to be mainly related to VTE occurrence are listed in Table 2. In the nonmalignancy patients, the mean follow‐up period was 685.8 days (range, 1–1688 days). There was no significant difference in the survival rates between the retrieval group and the nonretrieval group (Figure, 1A). A patient died from thoracic empyema in the retrieval group, and 2 patients died from end‐stage renal disease and pneumonia in the nonretrieval group. CT scan follow‐up after filter placement was performed in 11 of 14 patients with permanent implantation. The latest CT follow‐up was 11 to 1176 days (mean, 467.1 days) after the last filter insertion.
Figure 1.
Kaplan‐Meier analyses for all‐cause death stratified with or without retrieval in nonmalignancy patients (A) and in malignancy patients (B). (A) Log‐rank test revealed no significant difference between the 2 groups stratified with (n = 16, blue) or without (n = 14, green) retrieval in nonmalignancy patients (P = 0.50). (B) In the malignancy patients, the nonretrieval group (n = 19, green) had a significantly lower survival rate than the retrieval group (n = 7, blue) (P < 0.01). Abbreviations: IVCF, inferior vena cava filter.
Filter penetration was observed in 4 patients, all of whom received Gunther Tulip filter implantation. The latest CT follow‐up was 11 to 1176 days (mean, 429.3 days) after implantation in the nonpenetration group and 12 to 960 days (mean, 533.3 days) days after in the penetration group (P = 0.64). Filter penetration was diagnosed 11 to 346 days (mean, 95.5 days) after insertion (Table 3). We did not observe any perforation to nearby organs. There were no cases with apparent aggravation or recurrence of PTE.
Table 3.
CT Follow‐up Period in Permanent Implantation Cases
| Nonmalignancy, n = 14 | Malignancy, n = 19 | |||||
|---|---|---|---|---|---|---|
| Penetration, n = 4 | Nonpenetration, n = 7 | Follow‐up (—), n = 3 | Penetration, n = 3 | Nonpenetration, n = 14 | Follow‐up (—), n = 2 | |
| The latest CT follow‐up, d | 533.3 ± 401.6 | 29.3 ± 447.5 | — | 345.3 ± 211.7 | 174.3 ± 243.8 | — |
| Detection of penetration, d | 95.5 ± 167.0 | — | — | 316.7 ± 189.4 | — | — |
Abbreviations: CT, computed tomography.
Malignancy Patients
The types of malignancy are listed in Table 4. The mean follow‐up period was 467.8 days (range, 20–1857 days). We found a significantly lower survival rate in the nonretrieval group (Figure, 1B; P < 0.01). The 1‐year and 2‐year survival rates were 46% and 18%, respectively. In 2 of 19 patients with malignancy, the retrievable IVCF was implanted with the goal of preventing fatal PTE for permanent use from the beginning in the light of a poor prognosis for the malignancy, and follow‐up CT examinations were not performed. In the other 17 patients, the latest CT follow‐up was 11 to 693 days (mean, 204.5 days) after insertion.
Table 4.
Types of Malignancies
| Pancreatic carcinoma | 5 |
| Ovarian carcinoma | 4 |
| Prostatic carcinoma | 3 |
| Endometrial carcinoma | 2 |
| Intracranial lymphoma | 1 |
| Gum carcinoma | 1 |
| Mammary carcinoma | 1 |
| Lung carcinoma | 1 |
| Hepatic carcinoma | 1 |
| Gallbladder carcinoma | 1 |
| Carcinoma of the ampulla of Vater | 1 |
| Colon carcinoma | 1 |
| Sarcoma of the iliopsoas muscle | 1 |
| Kidney carcinoma | 1 |
| Bladder carcinoma | 1 |
| Uterine cervical carcinoma | 1 |
Filter penetration was observed in 2 patients with a Gunther Tulip filter and in 1 patient with an ALN filter. The ALN filter was found to have perforated into the aorta, but no symptoms were observed. The latest CT follow‐up was 11 to 693 days (mean, 174.3 days) after in the nonpenetration group and 110 to 520 days (mean, 345.3 days) after in the penetration group (P = 0.13). Filter penetration was diagnosed 110 to 482 days (mean, 316.7 days) after; this duration tended to be longer compared to the nonmalignancy patients (P = 0.08) (Table 3). We did not observe exacerbation or recurrence of PTE in the malignancy patients. The causes of death were all the malignant neoplasms themselves.
Discussion
In this study, we found that malignancy cases in which we could not retrieve an IVCF showed a significantly lower survival rate than those in which we could. And the causes of death in the nonretrieval group were all related to the malignant neoplasms themselves. Moreover, they had relatively few complications with IVCF during the short‐ or midterm follow‐up period.
IVCFs are inserted for the purpose of preventing severe PTE following massive venous thrombus migration. An IVCF should be inserted only during the critical period under the presence of a massive venous thrombus with PTE and should be retrieved as soon as it is deemed unnecessary. However, in some cases, DVT cannot be sufficiently ameliorated with anticoagulants, and it is difficult to retrieve the IVCF. In cases of temporary filter use, additional filter insertion is sometimes needed beyond the filter dwelling limit. On the other hand, retrievable IVCFs are useful because they can also serve as permanent filters. An IVCF gradually adheres to the vena cava wall, so earlier retrieval is safe and easy. There are reports showing the safe retrieval of a Gunther Tulip filter after prolonged implantation.1, 2 Moreover, it is relatively simple to snare the hook and fold the filter in a Gunther Tulip filter; therefore, this type of filter was selected in most cases at our institution. However, conical filters have been reported to protrude through the vena cava wall.5, 6 Ota et al reported that 56.1% of Gunther Tulip filters for permanent use protruded more than 3 mm out of the IVC wall.3 In this study, we found filter protrusion of ≥3 mm in 6 of 20 cases of Gunter Tulip permanent use with CT follow‐up. More long‐term CT follow‐up will probably reveal a higher rate of filter penetration. We also found filter protrusion of ≥3 mm accompanied by perforation into the aorta in 1 of 3 cases of ALN permanent use with CT follow‐up.
Some VTE cases were associated with malignancy. In many of such cases, an anticoagulant is not fully effective, probably due to the hypercoagulable state, and it is difficult to ameliorate DVT. Thus, we abandoned trying to retrieve the IVCF and left it as a permanent filter. In our observation, only 3 (15%) of 19 patients with permanent implantation were intolerant to anticoagulant use because of gastrointestinal bleeding or other types of bleeding, and an anticoagulant such as heparin and warfarin was administered in the remaining 85% of the patients. A Kaplan‐Meier curve showed no significant difference in the survival rate depending on the presence or absence of filter retrieval in nonmalignancy patients. On the other hand, in the malignancy patients, we observed a significantly lower survival rate in the nonretrieval patients whose deaths were all due to malignancy, and 1‐year and 2‐year survival rates were 46% and 18%, respectively. This result indicates that filter retrieval tends to be impossible in more advanced malignancy cases, and that complications following permanent use, such as filter protrusion, are believed to be rare because of the shorter dwelling period. So we may need to pay less attention to the long‐term disadvantages after retrievable IVCF implantation for permanent use in malignancy cases. Moreover, there were no medical records of PTE occurrence or recurrence in malignancy patients. Verso et al reported the PTE occurrence rate was 0.6% patient‐years in DVT patients, and the recurrence rate was 2.7% patient‐years in PTE patients.13 Thus, permanent use of a retrievable IVCF for such cases is believed to be less harmful. In addition, 57% patients in the retrieval group and only 1 (5.3%) of 19 patients in the nonretrieval group underwent surgery for tumor resection. This indicates that it is difficult to successfully treat venous thrombi in inoperable malignancy cases. However, it should be kept in mind that the presence of malignant tumors with a shortened life expectancy tends to keep us from choosing retrieval.
In patients with retrievable IVCFs, the retrieval success rate is also important. In this study, we found 22 (88%) successful retrieval cases in which retrieval was attempted. Similar to our results, a retrieval success rate of approximately 90% was reported in some studies on the Gunther Tulip.1, 3 In some patients, the IVCF could not be retrieved because filter adhesion or penetration was manifested even after adequate thrombus regression. For permanent use of retrievable IVCFs, the features and designs of the filters themselves must be improved.
This study has several limitations. First, retrospective and observational analyses were used, and important confounding factors might not have been excluded. Second, there is an absence of clear definitions for IVCF placement and retrieval. For this reason, an indication of IVCF is significantly affected by general status and comorbidities, especially in malignancy cases. Third, we analyzed data from a small sample size from a single institution, so a large‐scale multicenter prospective study will be necessary to identify a definitive indication for retrievable IVCFs.
Conclusion
Permanent use of a retrievable IVCF is relatively safe in short‐ or midterm follow‐up regardless of malignancy status. Retrievable filter use might be also reasonable in malignancy patients.
Supporting information
Reasons for placement of an IVCF
Reasons for not retreiving an IVCF
This study was supported in part by Grants‐in‐Aid #17659229 (to Y.H.), Core Research for Evolutional Science and Technology from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.
The authors have no other funding, financial relationships, or conflicts of interest to disclose.
References
- 1. Smouse HB, Rosenthal D, Thuong VH, et al. Long‐term retrieval success rate profile for the Günther Tulip vena cava filter. J Vasc Interv Radiol. 2009;20:871–877. [DOI] [PubMed] [Google Scholar]
- 2. Binkert CA, Bansal A, Gates JD. Inferior vena cava filter removal after 317‐day implantation. J Vasc Interv Radiol. 2005;16:395–398. [DOI] [PubMed] [Google Scholar]
- 3. Ota S, Yamada N, Tsuji A, et al. The Günther‐Tulip retrievable IVC filter: clinical experience in 118 consecutive patients. Circ J. 2008;72:287–292. [DOI] [PubMed] [Google Scholar]
- 4. PREPIC Study Group . Eight‐year follow‐up of patients with permanent vena cava filters in the prevention of pulmonary embolism: the PREPIC (Prevention du Risque d'Embolie Pulmonaire par Interruption Cave) randomized study. Circulation. 2005;112:416–422. [DOI] [PubMed] [Google Scholar]
- 5. Durack JC, Westphalen AC, Kekulawela S, et al. Perforation of the IVC: rule rather than exception after longer indwelling times for the Günther Tulip and Celect retrievable filters. Cardiovasc Intervent Radiol. 2012;35:299–308. [DOI] [PubMed] [Google Scholar]
- 6. Sadaf A, Rasuli P, Olivier A, et al. Significant caval penetration by the celect inferior vena cava filter: attributable to filter design? J Vasc Interv Radiol. 2007;18:1447–1450. [DOI] [PubMed] [Google Scholar]
- 7. Chintalapudi UB, Gutierrez OH, Azodo MV. Greenfield filter caval perforation causing an aortic mural thrombus and femoral artery occlusion. Cathet Cardiovasc Diagn. 1997;41:53–55. [DOI] [PubMed] [Google Scholar]
- 8. Goldman HB, Hanna K, Dmochowski RR. Ureteral injury secondary to an inferior vena caval filter. J Urol. 1996;156:1763. [PubMed] [Google Scholar]
- 9. Sarkar MR, Lemminger FM. An unusual cause of upper gastrointestinal haemorrhage—perforation of a vena cava filter into the duodenum. Vasa. 1997;26:305–307. [PubMed] [Google Scholar]
- 10. Woodward EB, Farber A, Wagner WH, et al. Delayed retroperitoneal arterial hemorrhage after inferior vena cava (IVC) filter insertion: case report and literature review of caval perforations by IVC filters. Ann Vasc Surg. 2002;16:193–196. [DOI] [PubMed] [Google Scholar]
- 11. Parkin E, Serracino‐Inglott F, Chalmers N, et al. Symptomatic perforation of a retrievable inferior vena cava filter after a dwell time of 5 years. J Vasc Surg. 2009;50:417–419. [DOI] [PubMed] [Google Scholar]
- 12. Guidelines on diagnosis and management of acute pulmonary embolism . Task Force on Pulmonary Embolism, European Society of Cardiology. Eur Heart J. 2000;21:1301–1336. [DOI] [PubMed] [Google Scholar]
- 13. Verso M, Agnelli G, Ageno W, et al. Long‐term death and recurrence in patients with acute venous thromboembolism: the MASTER registry. Thromb Res. 2012;130:369–373. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Reasons for placement of an IVCF
Reasons for not retreiving an IVCF