Several guidelines/recommendations have been published by different professional societies regarding the appropriate indications for IVC filter insertion [4, 5, 7]. Recently, efforts were made to revise collaboratively these guidelines, particularly by the American College of Chest Physicians (ACCP) and the Society of Interventional Radiology (SIR). Despite these efforts, significant differences exist, mainly regarding prophylactic indications. For clarity, indications could be divided in three categories: absolute indications, relative indications, and prophylactic indications.

Filter insertion can be either performed as outpatient or inpatient procedure. Most filter insertion occurred, however, in inpatient population due to indications and/or the underlying disease. The local anesthesia is basically used for both insertion and retrieval procedures. Depending on the patency of the venous access, filter insertion is performed by femoral venous access (preferably right side), jugular access (usually the right side), or brachial vein.

Prior to placement, the IVC should be assessed. Transcatheter venography prior to placement assesses IVC patency and potential variants that may require specific approaches [26]. The maximum diameter of the IVC and the number and position of renal veins should also be evaluated. If available, recent noninvasive imaging (CT or MRI) may be used to obtain the required informations. With most devices, the maximum diameter that could accept a filter placement ranged from 18 to 31 mm. Only one filter could be inserted in a larger vein (Bird’s Nest filter, Cook) up to 40 mm.

The standard technique of filter placement uses fluoroscopy guidance (Fig. 13.2) in dedicated facilities of interventions; however, ultrasound can alternatively be used for placement for bedside placement in unstable patients or nonmobilized patients [27, 28].

The optimal position of the filter insertion is the infrarenal segment immediately below the renal veins according to the manufacturer’s information for use. However, in some circumstances other target positions may be acceptable. Indications for suprarenal placement include the presence of IVC thrombus in the infrarenal segment or extrinsic compression, gonadal vein thrombosis, and filter placement during pregnancy, prior to pelvic or abdominal surgery that potential expose to IVC mobilization.

The technical success rate of filter placement is set at 97 % in a trained hand. Procedure-related complications are rare. Death and filter embolization are reported in 0.12 % [29] and 0.1 % [30, 31], respectively. Placement outside the target region occurred from 1 % to 9 % [28]. Deployment failure and vena cava perforation have also been reported as procedure-related complications [32, 33]. More frequent and less important is the thrombosis of the access site reported from 3 % to 10 % [30, 34–36]. If the low profile latest generations (introducer size less than 6 F) should significantly reduce the risk of site access thrombosis, a systematic use of imaging to assess the site of insertion in the postoperative course might show a relatively high rate even with modern devices [37].

The procedure of retrieval could be done in outpatient basis under local anesthesia. Depending on the filter, venous access is femoral or jugular. Usually, a dedicated set for retrieval is commercially available for each filter. However, for the filter designed with a hook, any other materials could be used provided that the size of the catheter/introducer to be inserted over the filter is respected. While the filter placement could be done on bedside basis, the retrieval procedure requires fluoroscopy guidance. After a cavogram, the catheter/introducer (8–12 F) is positioned close to the filter. A snare or a dedicated cone (grasping device) is then advanced over the filter. After catching and securing the hook in the snare, the introducer is advanced over the filter, which will collapse during the maneuver. The filter is then pulled back through the introducer and carefully inspected. A final cavogram is obtained to assess the IVC integrity and patency. When the filter is correctly centered with a limited penetration into the wall, the procedure is simple. Figure 13.3 summarizes steps for VCF retrieval procedure.

In cases when the filter is not centered or with exaggerated tilt or deep penetration, more aggressive approaches have been reported including double venous access, curved catheters, double guidewires, balloons, forceps, or laser extraction sheaths [38–41]. The most common procedure is to use a curved catheter to insert a guidewire around the filter distal to the hook and then insert the guidewire tip into a snare to externalize the guidewire tip and form an in situ homemade guidewire-based snare, which is used to trap the filter hook.

The skills needed for retrieval procedure are higher than for placement procedure and the learning curve longer. Technical success is approximately 85 % of attempts and most successful when performed within 12 weeks after placement [19, 42, 43]. In the CIRSE VCF Retrieval Registry, technical success rate for retrieval was 92 % [24]. Major complications of VCF retrieval include IVC thrombosis (4.3 %) [44] and IVC laceration.

Despite few reports that could not draw firm statements regarding efficiency of filters in preventing PE [45, 46], there are large data using recent technologies that support this objective. In the systematic review of the use of retrievable filters reported by Angel et al. in 2011 [44] and collecting 11 prospective clinical trials, the rate of recurrent PE was as low as 1.7 %. In a prospective randomized controlled trial (PREPIC [47]), a significant reduction of recurrent PE was found in the group who received VCF (1.1 % versus 4.8 %, p = 0.03), but a significant increase in symptomatic DVT was observed in the same group (28.8 % versus 11.6 %, p = 0.02). No benefit on the mortality was observed in the VCF group at short term as well as after 8 years follow-up [48]. It is of interest to notice that in unstable patients who had urgent VCF, a significant benefit on the mortality was observed [49].

Several device-related complications have been reported. These complications were the major reason for developing optimal filter and aggressive strategies for retrieval. They are appropriate to be recorded, and they might drive the quality improvement program. The Manufacturer and User Facility Device Experience (MAUDE), based on a voluntary reporting, accumulated only 842 complications from 2000 to 2010 [44]. Complications seem to vary in their rate and profile between filters. Furthermore, existing data suggest that long-term complications are correlated to the time after insertion [47, 50]. Accordingly, recent guidelines advocate filter retrieval [4, 7]. In 2010, the US Food and Drug Administration issued a statement for filter retrieval.

These complications include IVC occlusion with a rate ranging from 2 % to 30 % [51–54] and IVC penetration with a rate ranging from 0 % to 86 % [29, 51–53, 55, 56]. The IVC occlusion was probably underestimated as can sometimes be asymptomatic [57]. Also of note is that the presence of thrombus inside the filter (Fig. 13.4), which might be considered as a precursor of the IVC thrombosis, is a proof of filter effectiveness. Similarly, the IVC penetration (Fig. 13.5) is also a part of the mechanism of VCF stabilization at the target region of placement. When the penetration is too deep, it becomes a complication. The definition of threshold to becoming a complication is unclear and varied from 3 to 10 mm. Furthermore, the large majority of IVC penetrations are asymptomatic [55], and most of them did not limit retrieval procedure. Although the majority is asymptomatic, several reports pointed out induced injuries to the surrounding structures [58–60]. Recent technological advances have focused on the wall penetration control.

There are two major issues in managing patients with IVCF. The first one is certainly administration of anticoagulation and the second one is the time of the filter retrieval.

Whether filters are an indication of anticoagulation, once the preexisting contraindications subsided, is still a matter of debate. Theoretically, prolonging patients with a filter in place might prevent filter-related thrombosis and reduce recurrence of PE with, however, an increased risk of bleeding complication. Practically, there is no strong data supporting the use of anticoagulation in this population in terms of benefits but also in terms of intensity and duration of anticoagulation [61, 62]. The most common strategy regarding the filters and anticoagulation is to treat the underlying disease without considering the presence of the filter. However, since the extensive use of optional filters, the presence of thrombus inside the filter at the time of retrieval becomes a new issue. There is no clear recommendation in how to manage these patients and when to consider the presence of thrombus a matter of concern that could postpone the retrieval procedure. Usually, thrombus smaller than 10 mm or 25 % of the filter volume (Fig. 13.4) is suitable for retrieval. In case of larger thrombus, patient should receive anticoagulation for at least 3 weeks and reassessed for retrieval. This option is compromised if short-term optional filter is used. Exceptionally, thrombectomy devices are used to help filter extraction.

The second issue of filters in the era of optional filters is the filter retrieval. It is believed that early retrieval prevents filter-related complications [4, 7, 48, 50, 63]. Long-term dwell time increases the risk of adherence to the caval wall resulting in a decrease of the likelihood of successful retrieval and an increase in associated complications [42]. Filter retrieval requires, however, that there is no longer contraindication for anticoagulation and/or no longer indication for vena cava filtration, there is no significant thrombus that could compromise the procedure, and finally there is no deep wall penetration. Theoretically, it was reported that upward 85 % of optional filter should be retrieved [32]. Actually, less than 35 % of optional filters are effectively retrieved [23, 44]. The lack of following up patients with filters was reported as the main reason [64, 65], especially in those patients with long-term optional filters. In response to this limitation, several approaches were developed to improve patient’s follow-up [66–68]. Figure 13.6 summarizes our approach based on a systematic patient visit at 5 months after filter placement planned at the time of insertion. This strategy helps increase the retrieval rate by at least 50 %. Other developments to improve the retrieval rate are advanced techniques for filter retrieval and the extended use of new generations of filters that efficiently prevent deep penetration in the caval wall. Furthermore, standardized strategies through consensual recommendations should facilitate not only harmonization in the filter use but also in the filter removal strategies [25].