INTRODUCTION — Vena cava filters can be placed into the inferior vena cava (IVC) or, much less commonly, the superior vena cava (SVC) as a treatment to prevent pulmonary embolism.
Two general types of vena cava filters are available: permanent and retrievable. When retrievable IVC filters are used, it is important to create a plan for removing the filter as soon as protection is no longer needed.
IVC filters are typically placed in an interventional suite using fluoroscopy to guide the final position of the filter. Alternatively, IVC filters can be placed at the bedside either using fluoroscopy or transabdominal or intravascular ultrasound. Knowledge of the normal and variant anatomy of the vena cava is important for successful placement of vena cava filters and prevention of complications.
Complications related to IVC filters can occur at the time of filter placement (technical issues, medication related) or at a later time related to the access site, the filter itself (eg, thrombosis), or as a consequence of filter retrieval.
The placement and complications associated with predominantly IVC filters are reviewed here. Other treatments for deep venous thrombosis and acute pulmonary embolism are discussed separately. (See "Prevention of venous thromboembolism in adults undergoing hip fracture repair or hip or knee replacement" and "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults" and "Approach to thrombolytic (fibrinolytic) therapy in acute pulmonary embolism: Patient selection and administration".)
FILTER UTILIZATION — Utilization of IVC filters varies widely [1-6]. Some of the variability may be related to differences in available guidelines and interpretation of these guidelines in various populations: in particular, what constitutes a contraindication to or failure of anticoagulation, and the appropriateness of prophylactic filter placement [7-11]. (See "Prevention of venous thromboembolic disease in adult nonorthopedic surgical patients" and "Venous thromboembolism risk and prevention in the severely injured trauma patient".)
The only widely accepted and validated indications for vena cava filter placement in patients with venous thromboembolism are an absolute contraindication to therapeutic anticoagulation, complications of anticoagulation, and failure of anticoagulation when there is acute proximal venous thrombosis. The effectiveness of filter placement even for these limited indications has been questioned. A retrospective study concluded that use of IVC filters in patients with venous thromboembolism and a contraindication to anticoagulation was associated with a higher 30 day mortality compared with treatment without IVC filter placement [12,13]. Whether or not this finding is true or is related to weaknesses inherent to observational trials, IVC filter effectiveness in patients with VTE remains uncertain in the absence of randomized clinical trials. (See "Overview of the treatment of proximal and distal lower extremity deep vein thrombosis (DVT)", section on 'Inferior vena cava filter'.)
Other possible treatment indications, and prophylactic filter placement, are even more controversial [14-21]. (See "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults", section on 'Inferior vena cava filters' and "Overview of the treatment of proximal and distal lower extremity deep vein thrombosis (DVT)", section on 'Supportive therapies' and "Venous thromboembolism risk and prevention in the severely injured trauma patient".)
Nevertheless, with the increasing ease of placement of vena cava filters, including bedside insertion under ultrasonographic guidance, their use has steadily increased [22], and the reported "indications for use" have been extended to include prophylactic filters in patients at high risk for venous thromboembolism, particularly following trauma [15,16,18-20]. In a study of 263 hospitals and 130,643 acute venous thromboembolism hospitalizations, vena cava filter placement occurred in 15 percent overall with high variability ranging from 0 to 39 percent [3]. Significant clinical predictors of vena cava filter use included acute bleeding at the time of admission (odds ratio [OR] 3.4, 95% CI 3.2-3.6), major operation after admission (OR 3.4, 95% CI 3.3-3.5), presence of metastatic cancer (OR 1.7, 95% CI 1.6-1.8), and more severe illness (OR 2.5, 95% CI 2.3-2.7). Hospital characteristics associated with lower vena cava filter use included a small number of beds (<100 versus >400 beds; OR 0.2, 95% CI 0.2-0.4) and rural location (OR 0.4, 95% CI 0.2-0.5). Use varied widely, even in geographically proximate areas.
For prophylactic filter placement, use also varies widely. Several series have reported that "prophylactic" filter placement is becoming more common than insertion for "therapeutic" purposes [18-20]. In a study from the Nationwide Inpatient Sample (NIS), prophylactic vena cava filter placement increased at a significantly higher rate compared with placement associated with deep vein thrombosis or pulmonary embolus (157 versus 42 percent) over the seven-year study period (1998 to 2005) [18]. Whether a prophylactic filter is placed may also depend on insurance status. In a study of the National Trauma Data Bank, significantly fewer IVC filters were placed in noninsured compared with insured patients (2.7 versus 4.9 percent) [23]. Rising concerns over the increase in prophylactic filter placement, poor retrieval rate, and adverse events prompted warnings from the US Food and Drug Administration in 2010 that were updated in 2014 [24]. Following these warnings, one study showed a decrease in the rate of IVC filter placement (all types) from 55.1 in 2010 to 39.1 in 2014 (per 100,000 US population) [22]. In a retrospective trauma cohort study, regardless of a declining use of IVC filters (>90 percent prophylactic), the rate of pulmonary embolism remained largely unchanged [25].
TYPES OF FILTERS — Surgical interruption of the inferior vena cava (IVC) to prevent pulmonary embolism (simple ligation, DeWeese clip) was performed with general anesthesia via a retroperitoneal incision until transvenous interruption of the IVC (via direct venous access) became clinically feasible in 1967 [26]. Although vena cava filters can be placed into the superior vena cava in select situations, the IVC is the standard location for filter placement. Since the introduction of the original Greenfield filter, which required a surgical cutdown until a percutaneous approach was later described, a number of percutaneous IVC filters have been developed. IVC filter devices are designed to optimize flow dynamics, maximize clot-trapping capacity, expedite ease of insertion, and, for some filters, allow future retrieval (figure 1) [27-30].
Nonretrievable filters — Several nonretrievable IVC filters are available for use (figure 1) [26]. Beyond differences in design and material, each differs in terms of the diameter of the introducer system. An important design feature involves the maximal IVC diameter that can be accommodated by the device. Among the currently available filters, most of the permanent filters can only be placed in an IVC with a diameter of 28 mm or less. The TrapEase can be used up to 30 mm, and only the Bird's Nest filter can be placed in a larger IVC up to a diameter of 40 mm. An alternative strategy for patients with an IVC too large to accommodate non-Bird's-Nest filters is the placement of bilateral retrievable filters into each of the common iliac veins [31].
Retrievable filters — A number of retrievable (or "optional") filters are also available (figure 1) [26]. Prophylactic filter placement appears to have no effect on mortality, but its use is associated with an increased risk for deep vein thrombosis and other complications [6,21].
The US Food and Drug Administration (FDA) issued an initial warning in 2010 updated in 2014 regarding timely removal of retrievable filters to minimize adverse events, stating that physicians and clinicians responsible for the ongoing care of patients with retrievable IVC filters should remove the filter as soon as protection from pulmonary embolism is no longer needed [24,32]. (See 'Filter retrieval' below and 'Complications' below.)
Most of the retrievable filters can be placed in an IVC up to 30 mm, but the ALN can be used up to 32 mm and Denali filters can only be placed in an IVC of 28 mm maximum diameter (figure 1). A novel retrievable IVC filter that is permanently attached to a central venous catheter is available and is inserted via a femoral access [33-36]. As is the case with any central venous catheter, the risk of catheter-associated infection or thrombotic complications should be weighed against the potential benefits of the device.
Whether retrievable IVC filters are cost effective is unknown. A theoretically based cost analysis that compared retrievable IVC filters placed for prophylactic versus therapeutic indications in high-risk trauma patients found that prophylactic filters were slightly more costly (USD $37,700 versus $37,300) and less effective (quality-adjusted life year [QALY] difference 0.139) in a base case analysis [37]. Prophylactic filter placement would become a preferred strategy for individuals never having a filter, who have either an annual probability of venous thromboembolism ≥9.6 percent (base case, 5.9 percent) or a very high annual probability of anticoagulation complications ≥24.3 percent (base case, 2.5 percent). A prophylactic filter strategy would also be favored for an annual probability of venous insufficiency after filter removal <7.7 percent (base case, 13.9 percent) or <1.90 percent with a retained filter (base case, 14.1 percent).
ANATOMIC CONSIDERATIONS — Knowledge of the normal and variant anatomy of the vena cava is important for successful placement of vena cava filters and prevention of complications.
The major veins of the upper and lower extremity drain into the central veins. In the upper extremity, the subclavian veins arch cephalad behind the medial clavicle and then slope caudally to join the internal jugular veins and form the brachiocephalic (innominate) veins (figure 2). The thyroid veins contribute to the left brachiocephalic vein, and, together with the right brachiocephalic vein and azygos vein, these join to form the superior vena cava (SVC), which drains into the right atrium.
In the lower extremity, the common femoral veins become the external iliac veins after passing below the inguinal ligament (figure 3). They are joined by the internal iliac veins to form the common iliac veins. The common iliac veins merge at the level of the umbilicus to form the inferior vena cava (IVC) at the L5 level. It continues superiorly to the right of the aorta and receives multiple lumbar tributaries. At the L1 to L2 level, the right and left renal veins join the IVC, and this junction divides the IVC into the infrarenal and suprarenal portions. Because the IVC is usually located to the right of the spine, the left common iliac vein is longer with a less vertical path to the IVC.
The tributaries to the IVC (caudal to cranial) include the median sacral vein, the paired lumbar veins, the right gonadal vein (the left usually drains into the left renal vein), the left and right renal veins, the right suprarenal vein (the left usually drains into the left renal vein), the hepatic veins (right, left, middle), and the inferior phrenic veins (figure 4).
Venous abnormalities — Venous abnormalities that may be encountered include narrowing of an access vein or the vena cava, which may impede the ability to successfully traverse it with the device. Narrowing can be due to intrinsic stricture or thrombus or to extrinsic compression and may require balloon dilation or even stent placement to facilitate successful deployment of the filter. The anatomy may dictate the most appropriate approach (ie, internal jugular, femoral). The possibility of congenital anomalies also needs to be entertained as they can impact the placement of vena cava filters. In one study, variants caused alteration in the placement or selection of IVC filters in 18 percent of patients [38]. Whether preprocedure imaging is necessary for all patients prior to vena cava filter placement to identify venous abnormalities is controversial. However, if computed tomography (CT) or magnetic resonance (MR) imaging of the abdomen/pelvis is available, the study should be evaluated for possible abnormalities or anomalies. (See 'Venous imaging' below.)
Congenital anomalies of the IVC and its tributaries have been identified more frequently in asymptomatic patients due to the increased use of cross-sectional imaging. It is important to be able to recognize the more common variations in order to interpret the possible aberrant course of the guidewire during vena cava filter placement, particularly if preprocedure imaging has not been performed. Anomalies occur in 5 to 15 percent of individuals [26]. Duplicated and left-sided vena cava, as well as a circumaortic left renal vein, are among the more common anomalies (figure 5A-B) [39,40]. Anomalies of the renal veins may lead to drainage at a more superior or inferior location. Multiple anomalies involving the SVC and thoracic vessels have been reported, often in association with complex cardiac or great vessel defects, and these anomalies can create confusion and difficulty with SVC filter placement but are rare.
PREPARATION AND PLANNING — Patients should undergo a general assessment prior to placement of a vena cava filter. Prior to placement of an inferior vena cava (IVC) filter, coagulation parameters including the prothrombin time (PT), partial thromboplastin time (PTT), and international normalized ratio (INR) should be obtained. For older patients, those with hypertension or diabetes, and any others at increased risk for contrast-induced renal insufficiency, we obtain renal function tests such as blood urea nitrogen and creatinine.
Antibiotic prophylaxis is not routinely recommended for IVC filter placement (table 1). If the patient has been treated for infection, then antibiotics should be continued.
Venous imaging — A venocavogram is usually a requirement prior to positioning and deploying the IVC filter, but whether other preprocedure imaging is needed is controversial.
At a minimum, sonography of the access site should be performed to ensure that it is free of thrombus. Some vascular interventionalists perform duplex ultrasound of the femoral veins, iliac veins, and IVC prior to all filter placements, but this may not be feasible in all patients depending on body habitus. Cross-sectional imaging is not necessary in most patients, but if the patient has had a recent computed tomography (CT) or magnetic resonance (MR) study of the abdomen, the interventionalist should take the opportunity to review it or any other available abdominal and pelvic imaging to detect any abnormalities that might interfere with or alter the placement strategy [41]. The presence of a duplicated IVC, the diameter of the IVC, and the position and anomalies of the renal veins can be seen on CT or MR imaging. Although the CT and/or MR imaging may be very helpful for evaluating normal and variant anatomy, unless imaging was very recent, these may not demonstrate thrombus within the IVC.
The venocavogram is evaluated for several important aspects: duplication or other anomalies of the IVC, diameter of the vena cava, positioning of the renal veins, or the presence of clot within the IVC. (See 'Filter placement' below.)
Bedside placement — IVC filters are most typically placed with fluoroscopy to guide the final position of the filter, which requires intravenous contrast and transport of the patient to the interventional suite, which may be more difficult in critically ill patients.
Placement of IVC filters can be performed using bedside fluoroscopy. This requires having an intensive care unit (ICU) bed that can accommodate an angiographic C-arm. Alternatively, bedside placement can be performed using transabdominal or intravascular ultrasound (IVUS). Ultrasound-guided IVC placement has been performed with technical success and without significant increases in complication rates [42-49]. In a review that included 398 patients who underwent attempted bedside IVC filter placement, overall technical success was 99.2 percent [42]. The filter was malpositioned in three cases. The overall procedure-related complication rate was 3 percent. Transabdominal ultrasound can be very difficult or impossible to do accurately in patients who are obese or who have significant abdominal wounds. In these cases, IVUS can be used by experienced operators to identify the position of the renal veins and the size of the IVC and to guide placement of the filter [42,45,46,48]. Additional training is needed to master ultrasound-guided techniques [44,47].
Venous access site — All vena cava filters can be placed using an internal jugular or femoral approach (table 2). Some can be placed via an upper extremity peripheral vein approach (usually the brachial or basilic vein). Placement from the upper arm may require a longer sheath for positioning of the filter. The selection of a central venous access site to place a vena cava filter for a given patient takes into account similar considerations as with placement of any central venous device such as prior access history, site of known venous stenosis or thrombosis, bleeding risk, body mass index, possibly the presence of any venous anomalies (if known), and, to some degree, the patient's preference or condition. As examples, if the patient has a condition that requires ongoing bedrest, then a femoral approach may be reasonable. If the patient has a need for head elevation (intracranial pressure), then a jugular approach may be more appropriate. (See "Central venous access: Device and site selection in adults", section on 'Access site' and "Central venous access: General principles", section on 'Device and site selection'.)
A right-sided approach, regardless of access vessel, is generally preferred over the left side given the more direct route to the IVC. As with any central venous access procedure, ultrasound-guided access is recommended. (See "Principles of ultrasound-guided venous access".)
Some authors, but not all, have suggested that an internal jugular vein approach may be preferred to decrease the risk for subsequent femoral vein thrombosis that could be related to femoral vein injury at the time of placement, particularly in patients receiving prophylactic filters. (See 'Acute and recurrent deep vein thrombosis' below.)
During the procedure, the femoral vein approach is somewhat more comfortable for the patient and is also somewhat easier for the operator, since the catheters and guidewires can be placed onto the draped angiographic table, making sterility a little easier to maintain. The right femoral vein has a very direct course to the IVC. There are no obstacles to placing guidewires, catheters, or the filter. The operator does need to be aware when using a right common femoral approach that a duplicated IVC can be missed, if not specifically sought. The easiest way to exclude a left-sided IVC is to determine that there is a left common iliac vein, which would not be present with a left-sided IVC. Placement of a filter from the left femoral vein has a tendency to cause some tilting of the filter as the sheath and filter tend to be directed at the right lateral wall of the IVC. Commonly, an ultrasound of the lower extremities has been previously performed to evaluate the patient for the presence of deep vein thrombosis (DVT). Presence of thrombus in the common femoral vein would change the access approach. The iliac veins may or may not be seen well on an ultrasound examination, so if a femoral vein approach is being chosen, some practitioners will perform an iliofemoral contrast injection from the femoral vein to demonstrate that there is no clot in the iliac vein. This is particularly important if a left common femoral vein approach is being used since there is an increased incidence of DVT involving the left iliofemoral system.
For a jugular approach, a separate sterile table is needed upon which to place instruments and from which to pass guidewires and sheaths. Also, many patients are uncomfortable having their face covered, but following the procedure, the patient can get out of bed, which tends to make their postprocedure recovery much more comfortable. The internal jugular approach has certain pitfalls that an operator must be aware of when placing a filter. There is the risk of pneumothorax, which should be minimized by using ultrasound guidance while obtaining access into the internal jugular vein. Occasionally, difficulty passing a sheath beneath the clavicle will occur. Also, access into the inferior vena cava is achieved by passing guidewires and sheaths through the right atrium and liver while avoiding the hepatic veins. Traversing the right atrium can cause arrhythmias, which usually resolve with changing the position of the catheter, but a few individuals may be very sensitive to catheter manipulation.
Lastly, it is important to realize that inadvertent cannulation of the right gonadal vein can occur when access is from above (jugular, peripheral upper extremity vein access). The right gonadal vein usually enters the IVC just inferior to the right renal vein and has a course that is parallel and just lateral to the IVC. (See 'Filter placement' below.)
FILTER PLACEMENT — The chosen access site should be prepared using standard techniques. As with any central venous access procedure, we recommend sterile technique, including full barrier precautions, and sterile drapes to cover the entire patient. Ultrasound imaging during needle placement reduces time to venous cannulation and reduces the risk of complications during central venous access. (See "Principles of ultrasound-guided venous access".)
Once venous access has been established, a guidewire is used to access the vena cava over which a catheter, or sheath and catheter, can be placed to perform a venocavogram. In patients with severe allergic reactions to contrast or with renal insufficiency, the inferior venocavogram and any selective renal injections can be performed with CO2. When using a jugular approach, the catheter should ideally be placed into the left common iliac vein for the injection of contrast to exclude a duplicated inferior vena cava (IVC). If the approach is from a femoral vein, contrast should be injected into the ipsilateral common iliac vein (not vena cava) to clearly delineate the origin of the contralateral common iliac vein. The venocavogram should first be inspected for evidence of thrombus. If occlusion is demonstrated on a femoral approach, then a jugular approach may be required to demonstrate the upper extent of the occlusion. If a clot is demonstrated, a change in the position of the filter may be needed, such as suprarenal positioning. If there is a chronic occlusion of the cava, then filter placement may not be needed.
When interpreting the venocavogram, the operator should be alert for any subtle findings suggesting common venous anomalies (figure 5A-B). Although there may be inflow of nonopacified blood from the renal veins, a circumaortic renal vein can be missed because the inflow from the lower branch is subtle. Many operators routinely selectively catheterize the renal vein to identify anomalies that would require a change in the position of the filter. The most common of these anomalies is a duplication of the left renal vein, usually with a circumaortic lower branch. However, accidental cannulation of retroperitoneal veins during venography for IVC filter placement is an infrequent occurrence and does not usually result in negative clinical outcomes [50]. Inadvertent cannulation of the right gonadal vein with access from above can appear similar to that of the IVC. Such inadvertent positioning would be readily determined by a test injection of contrast. The guidewire can also slip into the right gonadal vein during exchange of the catheter used for the inferior venocavogram. If this happens, the IVC filter sheath could easily be placed into the right gonadal vein, and if the operator is not aware of the slight change in the trajectory of the wire, deployment of the filter in the right gonadal vein will result. The filter will be constrained and will not expand, and if the filter cannot be retrieved safely, another IVC filter will need to be deployed.
Because the IVC is frequently oval (rather than round), the diameter of the IVC should be measured in both the anteroposterior and lateral projections and documented to ensure it does not exceed the upper limit for the planned filter. Many operators use a marker catheter to obtain an accurate measurement. If the diameter of the cava is too large on the anteroposterior (AP) projection, but the diameter obtained from the lateral projection is within the manufacturer's instructions for use (IFU), then the filter can be carefully placed. The IVC filter will first usually catch from the front to back of the cava (the narrowed end will tend to round out the cava, allowing the legs to attach right to left). Similarly, if the diameter in the lateral projection is too large, but the diameter obtained from the AP projection is within the IFU, then the IVC filter will catch from right to left of the cava, again tending to round it out, allowing the legs to attach from front to back.
The instructions for use for the individual device should be followed to position and deploy the device. Although not usually an issue during IVC placement, when fluoroscopy is being used, it is important to avoid advancing any guidewires beyond the filter once it has been deployed; there is a risk, particularly with J-type wires, for the guidewire to be entangled in the filter. More important is the problem of guidewire entrapment with subsequent placement of central vein catheters where the guidewire may be advanced into the IVC and become entangled in the filter [51]. (See 'Guidewire entrapment with central line placement' below.)
Following filter placement, the sheath is removed and pressure applied to the access site until hemostasis is achieved. A small piece of gauze held in place with self-adhesive film dressing is sufficient.
Filter positioning — The filter should be placed in a location that will reduce the risk of venous thromboembolism through all the likely venous pathways. With standard anatomy, the filter should be positioned with the tip of the filter just at the inflow of the renal veins, which minimizes the accumulation of thrombus above the filter in the event of filter thrombosis. If the filter thrombosis is substantially below the renal vein inflow, then the dead space between the thrombosed filter and the renal veins would allow a clot to form, potentially leading to pulmonary embolism. IVC filters placed across the renal veins may have no significant effect on renal function, but they may not be stable in a pararenal location due to the inability of fixation mechanisms to fully engage the IVC wall [52].
If a circumaortic left renal vein is present, then if possible, the filter should be placed inferior to the lower limb of the renal vein. The rationale is that the circumaortic vein can serve as a conduit for thrombus to pass from the lower extremities through the left renal system and into the lungs. This is particularly a concern if a filter above a circumaortic left renal vein thromboses. In some cases, the lower of the renal vein limb is close to the confluence of the iliac veins, and there is not enough room to put the filter below the lower limb and above the confluence. In this case, a decision needs to be made whether to place the filter spanning the lower limb but with the legs of the filter in the IVC above the confluence of the iliac veins (the more common approach) or to place IVC filters in the iliac veins bilaterally. Placement covering the renal veins potentially places veins at risk for damage/thrombosis, but there is no evidence of complications from this approach [52].
In the setting of duplication of the IVC, a filter should be placed in each of the left and right IVC segments.
Suprarenal or SVC placement
●Suprarenal filter placement – Suprarenal filter placement appears to be safe (although considered off-label use in the United States), when necessary [53]. Indications for placement of a suprarenal filter include renal vein thrombosis, gonadal vein thrombosis, IVC thrombus extending above the renal veins, and thrombus in the infrarenal IVC that does not provide sufficient room for the filter to be placed above the thrombus and below the renal veins. For those who are pregnant, the vena cava filter is usually positioned in a suprarenal position [54]. Suprarenal placement may increase stresses experienced by the filter, potentially increasing the risk for complications. (See "Deep vein thrombosis and pulmonary embolism in pregnancy: Treatment", section on 'Inferior vena cava filters' and 'Complications' below.)
●Superior vena cava placement – The placement of filters in the superior vena cava (SVC) for patients with upper extremity deep vein thrombosis and contraindications to anticoagulation has been well described, but there is much less experience with this indication and technique [55-57]. It is important to note that vena cava filters are not a routine part of the management of upper extremity deep venous thrombosis, which is discussed in detail elsewhere. One review of 154 patients undergoing SVC filter placement noted that over a mean follow-up of 628 days, 74 percent of patients died as a result of their chronic illness or from cancer complications [58]. Three cases of perforation leading to pericardial tamponade occurred. (See "Primary (spontaneous) upper extremity deep vein thrombosis" and "Catheter-related upper extremity venous thrombosis in adults".)
Documentation — All patients who have undergone placement of a vena cava filter should have proper documentation and have an alert placed on their chart [51]. The patient should also be made aware of potential issues related to its placement. Ideally, they should have identification to alert medical personnel of its presence to prevent future inadvertent venous cannulation that may dislodge the filter or lead to guidewire entrapment. If the filter is planned to be a temporary filter, then it is critical to make sure there is a tracking mechanism to ensure the patient comes back for removal. The patient and family and caregivers should be educated that the filter should be removed when possible. (See 'Guidewire entrapment with central line placement' below.)
FILTER RETRIEVAL — Manufacturer recommendations for timing of filter retrieval vary; retrieval success decreases as duration of placement increases; however, filter retrieval following a prolonged duration of time has been reported [59-62]. The success of late filter retrieval is dependent upon the degree of strut epithelialization that has occurred [62] but is less successful and generally should only be attempted by experienced operators. Retrieval of "permanent" filters has also been described [32,63].
In a systematic review, the overall retrieval rate among 6834 patients was 34 percent [64]. In spite of a high rate of technical success when filter retrieval is attempted, a high percentage of patients are not scheduled for filter retrieval for a variety of reasons [64-67]. In case series, up to 30 percent of patients are lost to follow-up; this is particularly true in trauma centers [65,68,69]. In one review at a large trauma center, an attempt to remove a previously placed retrievable filter was made in only 10.3 percent of the patients [66]. Among these attempts, 18.3 percent were unsuccessful. Many of the filters were placed because of a contraindication to anticoagulation; however, 24.9 percent of patients who received a retrievable inferior vena cava (IVC) filter were discharged on anticoagulant therapy, most with the filter still in place. In a review of retrievable filters among patients with and without active cancer, although cancer was not associated with filter-related complications, it was associated with a significantly lower rate of filter retrieval compared with those without cancer (28 versus 42 percent) [70]. Additional risk factors identified in a study of 605 filters, for which the retrieval rate was only 25 percent, were predominantly related to limited life expectancy (eg, age >80 years) or unresolved underlying conditions (eg, bleeding, pulmonary embolism, deep vein thrombosis) [71].
As a result of such reports, the US Food and Drug Administration has expressed concern over the relatively low frequency with which retrievable filters are removed and has recommended that "implanting physicians and clinicians responsible for the ongoing care of patients with retrievable IVC filters consider removing the filter as soon as protection is no longer needed" [24,32].
Several studies have found an increased effectiveness of physician education and systematic patient follow-up to assure removal of retrievable filters at the appropriate interval after placement [69,72-78]. With such a program in place, up to 75 percent of retrievable filters are successfully removed. But even when posthospital care can be tracked and there is an opportunity for removal, most patients do not have their filters removed primarily because of continued care in an extended care facility or an ongoing indication for the filter. However, one review demonstrated that the rate of filter retrieval was increased 14-fold after the development of a retrieval program in a level 1 trauma center [79]. The retrievable filters are all approved for permanent placement indications and have largely replaced the older filters that had permanent indications. However, the long-term safety of the retrievable filters compared with permanent filters is largely unknown. (See "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults", section on 'Inferior vena cava filters' and 'Types of filters' above.)
Embedded filters — Although filter retrieval is achieved in 80 to 90 percent of attempts [60,80-85], removal of the filter can be technically challenging and potentially unsuccessful [86-88]. Reasons for failed attempts include excessive tissue ingrowth, thrombus within the filter, and filter tilt. A period of anticoagulation may hasten resolution of the thrombus, allowing for subsequent filter removal [89]. There have been a number of techniques described in the literature to remove these filters other than with the prepackaged retrieval devices [90-99]. Such techniques include the use of rigid forceps, modified loop snare, or double loop techniques, use of balloon displacement, or use of Excimer laser sheath, among others [90-99]. Each of these advanced techniques carries the potential risk of hemorrhage, distortion, or breakage of the filter and other complications. These techniques should be performed only by experienced interventionalists, and if the filter is difficult to remove using standard techniques, it may be more prudent to leave the filter in place, depending on the need for removal, rather than using aggressive techniques that might damage the filter or injure the IVC.
POSTPROCEDURE CARE — Following vena cava filter placement, the patient is usually kept on bed rest; the length of time depends on the approach. For patients with jugular filter placement, the head of the bed should be elevated to 45 degrees or more for approximately two hours. If the patient has had the filter placed from a femoral approach, most patients will remain at bed rest for four hours with the access leg straight. If the patient is on anticoagulation at the time of the procedure, the time of bed rest may need to be increased.
Whether to anticoagulate the patient following vena cava filter placement if the contraindication to anticoagulation has resolved is controversial. Some patient populations are at increased risk for thrombosis (eg, patients with malignancy) such that initiating anticoagulation, if feasible, may reduce the rate of filter-related complications [100]. This decision should be based upon the clinical circumstances that indicated the need for filter placement. (See "Overview of the treatment of proximal and distal lower extremity deep vein thrombosis (DVT)", section on 'Inferior vena cava filter'.)
COMPLICATIONS — Complications of inferior vena cava (IVC) filters include those associated with filter placement, such as bleeding or infection at the puncture site; allergic reactions to contrast or other medications used during placement; malposition of the filter; entrapment of the guidewire within the filter; postprocedure complications related to the access site, such as acute venous thrombosis, hematoma, or arteriovenous fistula; and longer-term complications, such as filter erosion/migration or embolization, and chronic thrombosis/recurrent thromboembolism [60,66,101-103].
There are insufficient data to compare the safety and efficacy of specific types of IVC filters. The Predicting the Safety and Effectiveness of Inferior Vena Cava Filters (PRESERVE) trial, a trial initiated in 2014 and jointly sponsored by the Society of Interventional Radiology and the Society for Vascular Surgery, was designed to evaluate the safety and efficacy of six different commercially available IVC filters (permanent and retrievable) over the long term [104].
Complication rates may be higher with retrievable compared with nonretrievable filters. In a review of data from the US Food and Drug Administration Manufacturer and User Facility Device Experience (MAUDE), self-reported complication rates with retrievable filters were significantly higher compared with the self-reported complication rates for permanent filters [105].
Acute and recurrent deep vein thrombosis — Although filter placement may protect the pulmonary vascular bed, it does nothing to lessen the thrombotic predisposition or the incidence of lower extremity venous thrombosis. Acute insertion site thrombosis has been reported in up to 40 percent of patients undergoing filter placement; however, this appears to be more common in older literature when using both a femoral approach and a large introducer [106-110]. Among patients for whom IVC filters have been placed for a contraindication to anticoagulation, the rate of deep vein thrombosis is overall low (2 to 4 percent) [111-114].
A retrospective review identified 35 patients who had a duplex ultrasound study following IVC filter placement [110]. The indications for filter placement were deep vein thrombosis (DVT) in 16 patients, pulmonary embolus in 13 patients, DVT and pulmonary embolus in 3 patients, and prophylactic filter in 3 patients at high risk for thromboembolization. At a mean follow-up of 12±2 days (median 6 days), the incidence of proximal DVT in venous segments without evidence of thrombosis before filter insertion was 40 percent, of which 71 percent occurred in the common femoral vein. Three were located in the femoral vein, and one was in the external iliac vein. Smaller-diameter filters were not associated with a lower incidence of femoral thrombosis compared with larger-diameter filters. The incidence of thrombosis was higher in patients with preinsertion pulmonary emboli compared with those patients with DVT (50 versus 38 percent) and prophylactic insertion (50 versus 0 percent). However, these subgroups were too small to attain statistical significance.
In a retrospective review (data obtained from querying a California Patient Discharge database), among 80,697 noncancer patients with no contraindication to anticoagulation, patients with and without a vena cava filter were compared [114]. The incidence of DVT within one year was 5.4 percent among filter patients and 3.7 percent among no-filter patients. However, in a group of patients who had active bleeding and presumably a contraindication to anticoagulation, patients with a filter had a 32 percent reduction in the adjusted risk of death at 30 days compared with no filter (9.5 versus 11.5 percent). Filter use remained associated with a similar significant reduction in the risk of death at 90 days after admission. These findings are consistent with recommendations that IVC filters are more appropriately reserved for those patients with DVT/pulmonary embolism who have a contraindication to anticoagulation.
The incidence of venous thrombosis may increase with time. One trial randomly assigned 400 patients to receive a permanent IVC filter (VenaTech, Greenfield, Bird's Nest, or Cardial IVC) plus anticoagulation or anticoagulation alone [108,109]. Recurrent DVT was more common in the group that received an IVC filter. The incidence of DVT at two years was 21 percent in the IVC filter group compared with 12 percent in the no-filter group. At eight years, the incidence was increased at 36 versus 28 percent, respectively. However, patients with filters were less likely to experience a symptomatic pulmonary embolism (6 versus 15 percent). The overall mortality and incidence of post-thrombotic syndrome were similar in both groups.
A later study evaluated the risk of insertion-site thrombosis of the femoral vein after predominantly prophylactic filter placement using contemporary devices [106]. None of the 56 patients studied had thrombus formation at the insertion site. Filters were placed for prophylaxis in 73 percent of the patients, and the remaining filters were placed for therapeutic indications.
IVC thrombosis — IVC obstruction has been reported in 3 to 30 percent of patients following IVC filter placement (permanent or retrievable) related to new local thrombus formation, thrombogenicity of the device, trapped embolus, or extension of a more distal DVT cephalad [112,115-118]. Although caval obstruction can be asymptomatic, it may result in lower extremity edema, increase the risk of other sequelae of venous insufficiency (eg, stasis ulcers), and result in the development of collateral vessels that may serve as a route for recurrent embolization [8,119,120]. Endovenous techniques have been successfully used to recanalize IVC occlusions and reduce symptoms, but rethrombosis is common [120]. (See "Endovenous intervention for iliocaval venous obstruction".)
Recurrent pulmonary embolism — Small thrombi are capable of passing through patent filters or through collaterals around obstructed filters; furthermore, direct thrombus extension can occur through the filter itself leading to recurrent pulmonary embolism. However, data suggest that recurrent pulmonary embolism is unusual following filter insertion (2 to 4 percent in most series) [8,9,64,108,109,121-123]. Pharmacomechanical thrombolysis is an adjunct that has been used to treat patients with severe symptoms [124]. (See "Endovenous intervention for iliocaval venous obstruction".)
The thrombotic complications related to prophylactic IVC filter placement in at-risk populations such as trauma, neurosurgery, oncology, intensive care, and bariatric patients are poorly defined. Data come from diverse studies; in some, pulmonary embolism is specifically sought, and in others only clinically documented pulmonary emboli are reported.
One systematic review of retrievable IVC filters estimated an incidence of pulmonary embolism following filter placement of 1.3 percent [64]. A later randomized trial compared retrievable IVC filters plus anticoagulation with anticoagulation alone in patients with pulmonary embolism and a high risk of recurrence [125]. The rate of symptomatic recurrent pulmonary embolism was 3 percent in the filter group, which was not significantly different from the control group at 1.5 percent. Filters were retrieved overall in 85 percent of patients and in 93 percent in whom it was planned and attempted. This study demonstrates that if a patient can be anticoagulated, filter placement will not add any additional protection. In addition, it would suggest that once the patient can be anticoagulated, the filter should be removed since it is not adding any benefit.
A large number of reports on individual filters have been published; however, clearly defined endpoints and objective criteria for reporting recurrence and mortality rates are lacking [126,127].
Guidewire entrapment with central line placement — Guidewire entrapment, particularly with J-tipped wires, is a known complication of central line placement [128-132]. The guidewire can become entangled in the apex of the IVC filter. The placement of central venous devices in the intensive care setting has increased the number of incidents in which the guidewire becomes entrapped in a vena cava filter [130]. Since the majority of central lines are placed from the upper extremity or internal jugular veins, the incidence of guidewire entrapment is higher from an upper compared with lower extremity approach, and displacement of the filter into undesirable structures is more likely. However, entrapment of guidewires from the femoral approach can occur.
An in vitro study compared the risk of engaging and entrapping guidewires in eight IVC filters available in the United States, including two types of Greenfield filters, Simon Nitinol, Bird's Nest, VenaTech LGM, VenaTech LP, TrapEase, and Günther Tulip, using four different wires (15 mm, 3 mm, and 1.5 mm J-tipped, and straight) [132]. Engagement was defined as a filter/wire interaction causing the filter or wire to become deformed. The wires were passed through each filter 100 times, 50 from a jugular approach and 50 from a femoral approach. Although the straight wire did not engage any of the filters, each of the J-tipped wires engaged in all the filters tested. The VenaTech LP and Günther Tulip filters did not entrap any of the wires. The force required to disengage entrapped wires resulted in structural damage to the wire and/or filter.
If the operator placing a central line finds there is resistance in trying to remove the wire, the operator should immediately stop trying to remove the wire. If the wire is pulled, the IVC filter can become dislodged, distorted, or moved into a very undesirable position (right heart, superior vena cava, innominate, jugular veins). An abdominal plain film (KUB) should be obtained to assess the position of the wire. If it appears to be in contact with a filter, then a vascular interventionalist should be consulted for help in removing the wire. With careful maneuvering using fluoroscopy, the wires can usually be extracted.
Filter migration, erosion, fracture — Other late complications of IVC filter placement are uncommon. Most of these complications are reported in individual case studies, and there are otherwise few data to determine the factors that are responsible. These include filter migration [133-137], which can also occur acutely during filter placement; filter erosion and perforation through the IVC wall and into adjacent structures [103,138-153]; and filter fracture and fragment embolization [154-161]. Migration of the filter or fragments into the right atrium, pulmonary artery, right gonadal vein, lumbar veins, and even into the aorta have been reported [162]. Percutaneous image-guided removal of migrated or fractured filters and associated filter fragments can usually be accomplished if they are intravascular, but the success rate varies depending upon anatomic location [138,163].
MORTALITY — Mortality from filter placement appears to be quite low, regardless of which filter is used. As examples, three deaths (0.12 percent) were reported in a review of 2557 patients undergoing filter insertion [112], and a second series of 1765 insertions had a major complication rate of 0.3 percent [164]. However, patients who receive permanent IVC filters have high mortality with follow-up (50 percent at a mean follow-up of 36 months in one study [165]).
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Superficial vein thrombosis, deep vein thrombosis, and pulmonary embolism".)
SUMMARY AND RECOMMENDATIONS
●Inferior vena cava (IVC) filters may be used for the treatment of venous thromboembolic disease when a patient is not a candidate for anticoagulant therapy or has failed anticoagulant therapy or in those with a limited ability to tolerate subsequent pulmonary emboli (eg, patients with acute hemodynamically massive pulmonary embolism or chronic thromboembolic pulmonary hypertension). (See "Treatment, prognosis, and follow-up of acute pulmonary embolism in adults", section on 'Inferior vena cava filter' and 'Introduction' above.)
●Prophylactic IVC filter placement may be an option for carefully selected patients at high risk for venous thromboembolism in whom pharmacologic or mechanical prophylaxis is contraindicated or has proven ineffective. (See 'Introduction' above.)
●Two general types of vena cava filters are available: permanent and retrievable. When retrievable IVC filters are used, it is important to create a plan for removing the filter as soon as protection is no longer needed (ie, once the patient can be anticoagulated). In spite of a high rate of technical success when removal of a retrievable filter is attempted, a high percentage of patients are not scheduled for filter retrieval for a variety of reasons. Even when there is an opportunity, most patients do not have their filters removed primarily because of continued care in an extended care facility or an ongoing indication for the filter. (See 'Types of filters' above.)
●There are insufficient data to compare the safety and efficacy of specific types of IVC filters. Until there is more clinical experience with temporary or retrievable filters, clinicians should consider the decision to place a filter irrevocable and, given the absence of long-term follow-up studies, one that requires careful consideration. (See 'Mortality' above.)
●Complications of IVC filters can include deep vein thrombosis at the insertion site; IVC thrombosis; recurrent pulmonary embolism; and filter erosion, fracture, or migration. Migration can be due to guidewire entrapment and inadvertent dislodgement during central line placement. If an operator experiences resistance when trying to remove a guidewire in a patient with an IVC filter, abdominal imaging should be obtained to assess the position of the wire relative to the filter. Using fluoroscopy, the wire can usually be separated from the filter and extracted. (See 'Complications' above.)
