Because of the limitations of temporary IVC filters, retrievable IVC filters are likely to replace the temporary devices. Consequently, temporary devices are not discussed in this review.

Limitations of Available Data

The IVC filters, which are very effective in preventing pulmonary embolism, are used when anticoagulation is contraindicated or considered inadequate. In extensive review articles published in 2000¹ and 2005,² the authors identified indications and proposed indications for IVC filter use. Indications include the following:

• Contraindication to anticoagulation, such as serious bleeding, coagulopathy, cerebral metastasis, risk of falls, head injury, and significant thrombocytopenia (platelet count < 50 x 10³/mm³)
• Failure of anticoagulation
• Limited cardiopulmonary reserve
• Large free-floating clot loosely attached to wall of inferior vena cava

 Proposed indications include the following:

• Prophylaxis for pulmonary embolism in trauma patients
• Treatment of venous thromboembolism in patients with cancer
• Prophylaxis for pulmonary embolism in high-risk patients undergoing orthopedic surgery
• Pre- or postpulmonary embolectomy
• Prevention of pulmonary embolism in patients with extensive free-floating iliofemoral thrombi
• Prevention of pulmonary embolism in patients with chronic obstructive pulmonary disease and deep vein thrombosis
• Prevention of pulmonary embolism in patients with minimum cardiopulmonary reserve and deep vein thrombosis
• Treatment of venous thromboembolism in pregnant patients
• Treatment of venous thromboembolism in organ transplant recipients

In one of the review articles, the author also concluded that it is virtually impossible to quantify the rate of individual complications of these devices or to compare the safety and effectiveness of one device with that of another.¹ Even though reporting standards for vena cava filter placement and patient follow-up were published in 1999,³ and updated in 2005 to address retrievable filters,⁴ almost all of the available articles do not meet these standards. Specifically, almost all studies were not randomized, lacked controls, used various methods (or no methods) for surveillance of complications, included various types of patients, and used varied lengths of time for follow-up evaluations.

Consequently, event rates reported in most articles are probably an underestimate of the true short-term and long-term event rates. Even so, a number of complications have been identified that may result from IVC filter placement. Thrombosis at the insertion site has been reported in as many as 50% of patients, and 4–11% have IVC thrombosis.² Other problems include filter thrombosis, migration, tilting, fracture, and penetration of the IVC. The IVC filters appear to increase the risk of deep vein thrombosis (DVT) and may cause a pulmonary embolus, which may result from clot formation on the upstream side of the filter or from development of collateral vessels that may bypass the filter. 

Increasing Use of Inferior Vena Cava Filters

Although the data on the risks versus benefits of IVC filters are rather limited, the use of these devices has been increasing rapidly. From 1979–1999, the use of IVC filters in the United States increased 2000%.⁵ More recently, a registry that enrolled patients with DVT from 183 study sites in the United States between October 2001 and March 2002 found that 781 (14%) of 5451 patients received an IVC filter.⁶ It should be noted that this extensive use of IVC filters was documented before retrievable IVC filters were approved in the United States. The availability of retrievable IVC filters is likely to accelerate the use of these devices even though the risk and benefits of such use have not been well defined.

Anticoagulation for Permanent Filters

Although IVC filters are often used in situations in which anticoagulation is at least temporarily contraindicated, one of the few well-designed trials in this area (the Prévention du Risque d’Embolie Pulmonaire par Interruption Cave [PREPIC] study) has confirmed an IVC filter–related increase in DVT if such patients do not receive long-term anticoagulation.^{7, 8}In the PREPIC study, 400 patients with proximal DVT (with or without pulmonary embolism) received 3 months of standard anticoagulation and were randomly assigned to either receive or not receive an IVC filter. At 2 years of follow-up, signifi-cantly fewer pulmonary emboli were noted in the IVC filter group (1.1% vs 4.8%), but the IVC filter group had almost twice as many DVTs (20.8% vs 11.6%).⁷ After a total of 8 years of follow-up, the event rates had increased to 6.2% symptomatic pulmonary emboli in the IVC filter group versus 15.1% in the no filter group (p=0.008), and 35.7% DVTs in the filter group versus 27.5% in the no filter group (p=0.042).⁸ Mortality rate and the rate of postthrombotic syndrome were not different between the IVC filter and no filter groups.⁸ Because IVC filters reduce, but do not eliminate, the risk of pulmonary embolism, and because they increase the rate of DVTs, the authors concluded that the use of these devices may be beneficial in patients at high risk for pulmonary embolism, but that the systematic use in the general population of patients with DVT is not recommended.⁸ Others would suggest that the greater than 40% rate of combined pulmonary embolism and DVT in the IVC filter group is clear justification for long-term anticoagulation, if not otherwise contra-indicated.

Additional supportive data on the need for anticoagulation have been provided by a published case series.⁹ Over an 8-year period, the authors identified 30 patients who had symptomatic IVC filter thrombosis with perma-nent IVC filters at their institution. Indications for filter placement were a contraindication to anticoagulation in eight patients (27%), recurrent emboli despite adequate anticoagulation in three patients (10%), and prophylaxis for pulmonary embolism in 19 patients (63%). Anticoagulation therapy was used in 15 (50%) of the patients. Among the 30 patients, a filter thrombosis occurred during the first 6 months after filter placement in 15 patients. Of these 15 patients, 10 were not receiving an anticoagulant at the time filter thrombosis was diagnosed and the other 5 patients did not have therapeutic levels. Two patients had an activated partial thrombo-plastin time less than 1.5 times the normal control, and three patients had an international normalized ratio less than 1.5. In the other 15 patients, filter thrombosis occurred after 6 months from when the filter was placed, and none of these patients was receiving anticoag-ulation at the time filter thrombosis was diagnosed. In addition, in 10 (33%) of the 30 patients, thrombus developed on top of the filter, thereby creating a risk for pulmonary embolism.

Retrievable Filters

If permanent IVC filters require long-term anticoagulation, then it is reasonable to suspect that the availability of retrievable IVC filters might limit the need for anticoagulation to the length of time that the filter is in place. If the IVC filter is being used because of a transient period of high risk for pulmonary embolism, then it would seem logical that the filter could be removed after an initial period of anticoagulation had reduced the pulmonary embolism risk. Similarly, if the IVC filter is being used because of a transient contraindication to the use of anticoagulation, then the retrievable IVC filter may be used until standard anticoagulation may be implemented. Under such circumstance, being able to remove the retrievable IVC filter may eliminate the need for indefinite anticoag-ulation and reduce the risk for any complications from the IVC filter.

The introduction of retrievable IVC filters, therefore, provides flexibility of treatment that very likely will significantly increase the use of IVC filters. In a 1997 survey, respondents indicated that availability of a retrievable IVC filter would increase prophylactic use of IVC filters from 29% to 53% in patients with lower extremity trauma.¹⁰ The authors of a 2005 report found that the introduction of retrievable filters at their institution increased the rate of IVC filter placement from 1 in 44 months before the availability of retrievable filters to 27 in 37 months after retrievable filters became avail-able.¹¹ Ninety-three percent were retrievable.

Three retrievable IVC filters are available in the United States. The Recovery filter (Bard Peripheral Vascular, Inc., Tempe, AZ) was the first to be approved by the United States Food and Drug Administration, in July 2003.² Each of these devices is made so that it can be left in place permanently if necessary, or it can be retrieved if only temporary protection is warranted. Even so, one needs to consider the risks that may be associated with placement of such a device. Unfortunately, only limited data are available regarding the risks of retrievable IVC filters, and, as discussed earlier, no ran-domized controlled trials have been conducted with these new devices. Reporting standards have been modified and updated for retrievable IVC filters such that better data may be available in the future.⁴

One group of authors reported their experience in using the Günther Tulip temporary IVC filter (Cook, Inc., Bloomington, IN) for short-term protection against pulmonary embolism.¹²Eleven filters were placed in 10 patients. Indications for the temporary IVC filter were pulmonary embolectomy with DVT in three patients, pulmonary embolectomy without DVT in four patients, pulmonary embolus with DVT in two patients, and free-floating iliofemoral DVT in one patient. This study used anticoagulation in seven of the 10 patients receiving temporary IVC filters. Anticoagulation was given for 6 months in four of seven patients, and indefinitely in the other three patients. Of the seven patients who received anticoagulation, only one had thrombus present in the filter. Two of the remaining three patients (who did not receive anticoagulation) had a thrombus in the filter. Eight patients received systemic heparin immediately after filter placement. All filters were removed within 14 days after placement.

Another group described their experience in which 50 temporary IVC filters of different types were placed in 47 patients.¹³ Thrombi developed in 9 (18%) of 50 filters. Although the authors indicated that all 47 patients had a contra-indication to anticoagulation at the time of temporary IVC filter placement, they mention that anticoagulation therapy was discontinued 3 hours before filter insertion and was restarted as soon as possible with heparin. Therefore, it is difficult to determine which patients received anticoagulation and for how long, but there was still an 18% thrombotic event rate in their study population of patients who had received IVC filters.

Another unresolved issue of retrievable IVC filters is how long retrievable filters may be left in place. Initially, the recommendation was that the retrievable filters should be left in place for only 10–14 days (with differences dependent on each manufacturer’s recommendation). In a recently published study of temporary ALN filters (ALN Implants Chirurgicaux, Chisonaccia, France), the authors reported a 100% retrieval rate of the filters within 3 months of placement, but only a 50% retrieval rate after 3 months of filter place-ment.¹⁴ The main concern is that endothelialization of the attachment points develops and makes removal more difficult and dangerous because of possible damage to the vena cava. Some have suggested that periodic repositioning of the retrievable IVC filter may minimize such endothelialization and, thereby, allow for longer indwelling times.

Another group reported on 23 patients with Günther Tulip IVC filters who required repositioning of their filters after 14 days of implantation in order to extend the dwell time.¹⁵ The mean dwell time in this group was 34.8 days, and the mean number of repositionings/patient was 1.5. Attempted removal of one filter at 16 days proved to be impossible.

Others, however, have reported on successful delayed removal of retrievable filters without repositioning. In March 2005, a group of authors reported on the removal of a Günther Tulip IVC filter after 317 days in a trauma patient.¹⁶ Removal had been attempted at 21 and 25 days after placement but was aborted because of clot trapped in the filter. The patient then received warfarin therapy and was discharged. Approxi-mately 9 months later, the patient returned and the filter was found to be free of clot. The decision was made to attempt to retrieve the filter. During the removal attempt, the filter partially collapsed but could not be easily removed. Redeployment failed because the filter could not be reopened. Ultimately, the filter was removed with the use of additional force. Although a vena cavogram revealed mild stenosis of the IVC in the area where the filter had been, this stenosis had resolved 3 months later.

A few months later, another author described removal of a Günther Tulip IVC filter after 475 days.¹⁷ The patient had extensive DVT with multiple pulmonary emboli and a recent history of hemorrhagic pericarditis due to lupus erythematosus. Filter removal was considered at 8 days after placement but was delayed because of a large amount of thrombus within the filter. Anticoagulation was undertaken, and 15 months later when the IVC filter and leg veins were found to be patent, removal of the filter was again attempted. Attempted removal resulted in partial collapse of the filter, which could not be expanded. This resulted in the filter being removed with greater than usual force, and soft tissue was found attached to one of the filter legs. A computed tomographic scan with contrast material enhancement, however, showed no significant damage to the IVC.

Another author described the retrieval of a Recovery IVC filter at 224 days after implan-tation.¹⁸ In this case, the filter was placed in a comatose 27-year-old woman with extensive DVT, a subdural hematoma, and gastrointestinal bleeding after a vehicular accident. She later requested removal of the filter because of a desire to become pregnant. The filter was removed without difficulty at 224 days after implantation, and a vena cavogram showed no damage to the IVC.

Exactly how long various retrievable IVC filters can remain in place before being safely removed requires further study. The above case reports of filter removal more than a year after placement may be encouraging, but one must also wonder how often serious complications may result from attempted delayed removal. In one of the reports, one filter was unable to be removed after only 16 days of implantation.¹⁵ The importance of limitations on removal time may have con-tributed to the fact that only 35% of retrievable filters were removed in another report.¹¹

Conclusion

The use of IVC filters has increased dramatically in the past few years, and the recent introduction of retrievable IVC filters is likely to escalate the use of these devices. The available data on safety and effectiveness of these devices are very limited even though reporting standards for filter placement and follow-up have been published and updated. Recent data suggest that anticoagulation is warranted as long as IVC filters are in place, in order to reduce the increased risk of DVT that is associated with IVC filter use as well as the substantial risk of pulmonary embolism.

Retrievable filters offer flexibility in manage-ment that may limit the duration of anticoag-ulation. Although removal of retrievable IVC filters at more than a year after implantation has been described, one must question the wisdom of leaving these filters in long term because of the difficulty of removal reported in some cases and the fact that one filter was not removable at 16 days after implantation.

Although the retrievable IVC filters offer a valuable new alternative to the traditional permanent IVC filters, further work needs to be done to define the advantages, disadvantages, and limitations of these devices. In contrast to earlier reports, investigators should follow the reporting standards for IVC filter placement and follow-up.