Impact of Advanced Technique on Improvement in the Retrievable Inferior Vena Cava Filter Retrieval Rate

Abstract

Background: It is recommended to remove retrievable inferior vena cava filters (r-IVCFs) when they are no longer needed because their presence may give rise to serious complications related to prolonged placement of the filter. An advanced filter retrieval technique may help improve the retrieval rate.

Methods and Results: 107 consecutive patients (mean age; 61±18 years, male 53%) in whom r-IVCF retrieval was attempted were prospectively enrolled between April 2012 and December 2018. The frequently used advanced techniques were sling technique and biopsy forceps dissection technique. Retrieval success was 75% with standard retrieval technique alone; however, the overall retrieval success rate improved to 98% with advanced techniques. We observed few serious complications related to the retrieval procedure. Logistic multivariate analysis identified prolonged indwelling time (P=0.0011) and embedded hook in the caval wall (P=0.0114) as independent predictors, and the cutoff value for the indwelling time for requirement of advanced technique was 80 days.

Conclusions: Advanced retrieval techniques helped improve the retrieval rate without serious complications. We may need to consider the referral of patients to centers with expertise in advanced retrieval techniques when the indwelling time is >80 days, and pre-retrieval CT image shows a hook embedded in the vessel wall.

Since the early 2000 s, when the Food and Drug Administration (FDA) approved the use of the retrievable inferior vena cava filter (r-IVCF), the number of IVCF implantations has risen significantly.¹ However, most r-IVCFs were left in the body for an extended period despite prophylactic use, leading to multiple complications associated with the filter, such as fracture, perforation, migration, and occlusion.² Although the FDA issued a statement in 2010 encouraging all physicians involved in the treatment and follow-up of IVCF recipients to remove the r-IVCF when no longer needed or clinically indicated, the attempted retrieval rates for r-IVCF remain low.³

Although there are many possible reasons for not retrieving r-IVCFs, retrieval failure is a major reason.⁴ Recent reports have described the use of so-called “advanced techniques” after failure of r-IVCF retrieval with the standard technique alone, and have highlighted the resultant improvement in the retrieval rate.^5,6 However, the excimer laser that is considered a crucial device for retrieval in more difficult cases is yet to be approved for this purpose in Japan. In this regard, we expect the retrieval success rate with advanced techniques in Japan to be relatively lower than that reported by studies in other countries.

In the present study, we used our clinical experience to determine the retrieval success rate with advanced techniques, the independent predictors for r-IVCF retrieval failure with the standard retrieval technique, and the appropriate timing for performing retrieval.

Methods

Study Population and Protocol

Patients who underwent r-IVCF retrieval procedure at Ota Memorial Hospital between April 2012 and December 2018 were consecutively enrolled. The timing of the retrieval was at the discretion of each attending physician, considering the clinical condition of each patient, including low ongoing risk of venous thromboembolic event, possibility of being able to undergo anticoagulation, improved ambulatory capability, and acceptable medical profile; retrieval attempt within 3 months after implantation was generally recommended.⁷

In all patients, pre-retrieval computed tomography (pre-CT) with contrast was scheduled within 2 weeks before the retrieval procedure to evaluate filter characteristics, amount of residual thrombus in the pulmonary artery and the pelvic and limbs veins, and trapped thrombus in the r-IVCF.

Pre-CT Measurements

The pre-CT image was used to assess the integrity of the r-IVCF and the relationship of each leg strut to the caval wall.⁸ We also calculated each leg strut penetration (LSP) length outside of the caval wall on a cross-sectional view. We calculated the distance from the vessel wall to the apex of each strut as the LSP length. LSP length were calculated for all struts of each filter and the following indices were derived: maximum LSP length; longest LSP length of each strut in each filter, and mean LSP length; average LSP length of each strut in each filter. These indices were obtained for all types of r-IVCF except OptEase (Cordis, Warren, NJ, USA) because its configuration is a box-shape without leg struts, whereas the Günther-Tulip (Cook Medical Inc., Bloomington, IN, USA) has 4 leg struts, and the Denali (Bard Peripheral Vascular, Inc., Tempe, AZ, USA) and ALN (ALN Implants Chirurgicaux, Ghisonaccia, France) have 6 leg struts. Thus, the mean LSP length was calculated by dividing the sum of each LSP length in each filter by 4 for the Günther-Tulip and by 6 for the Denali and ALN.

The filter tip was considered to be embedded if the tip appeared tightly apposed to the caval wall. Although the OptEase has tips at the top and the bottom of the filter, in this study we only assessed the bottom tip with a hook.⁹

Filter Retrieval Procedures

Technical retrieval success was defined as successful endovascular removal of the r-IVCF. Standard retrieval techniques were initially attempted for all procedures, using the retrieval system dedicated for each type of r-IVCF, which consisted of a coaxial vascular sheath and a snare system. When standard techniques failed, adjunctive advanced techniques, including single and double sling wire technique, balloon displacement technique, forceps dissection technique, and mechanical sheath dedicated for pacemaker lead retrieval (Byrd sheath: Cook Medical Inc.) were used. The procedure was discontinued when there was a serious complication or available options for retrieval failed, even if the r-IVCF was yet to be removed. After filter removal, venography was performed to confirm whether any complications, including vessel rupture had occurred. CT with contrast was scheduled for the following day to assess the status of the IVC.

All the retrieval procedures were performed by 3 board-certified interventional cardiologists.

Advanced Retrieval Techniques

We describe each advanced technique used in our procedures.⁵ (1) The single sling technique: a loop just below the hook or apex of the filter is made with a 0.035 guidewire and a 12/20 mm Trilobe snare (EnSnare: Merit Medical, South Jordan, UT, USA). This technique is useful for straightening a tilted filter and enhancing traction force. There are 2 approaches. The first is to make a loop under the filter’s apex. A 0.035 guidewire is placed under the filter. Another 0.035 guidewire is passed through separate interstices (ideally on the opposite side), exchanging it for an EnSnare. The tip of the 0.035 guidewire is snared by the EnSnare, being pulled up and withdrawn from the sheath. The other approach is to make a loop above the filter’s apex. We place a 4Fr JL1 diagnostic catheter with small-radius reverse-curve below the apex, advancing a 0.035 guidewire to above the filter. We introduce an EnSnare to catch the 0.035 wire and externalize it. We chose our approach based mainly on the configuration of the r-IVCF: conical or box. Essentially, we chose the first approach for the conical type and the second approach for the box type. (2) Forceps dissection technique: a 7Fr endomyocardial biopsy forceps (Cordis, Santa Clara, CA, USA) is directed towards the embedded filter and used to dissect the adherent tissue around the filter’s apex. The apex is then grasped and released from the caval wall. (3) Double sling technique: this approach was used if the single sling technique failed. The concept is similar to the bidirectional approach in the coronary intervention field.¹⁰ In the double sling technique, another loop is created around one of the leg struts from the femoral vein (conical type) or at the top of the filter from the internal jugular vein (box type) with a 0.035 guidewire from the femoral vein. Two loops are created on both sides of the filter, providing more powerful traction force, and keeping the r-IVCF position at the same level, thus preventing intussusception of the IVC.¹¹ (4) Balloon displacement technique: a curved-tip catheter is used to direct the wire between the filter’s apex and the caval wall. An angioplasty balloon is then placed at the level of the apex. Balloon inflation leads to displacement of the filter from the caval wall. (5) Byrd mechanical sheath dedicated for pacemaker lead retrieval: when removing severely embedded filters, more powerful devices are needed to remove adherent tissue around the filter. This sheath is made of polypropylene and is so rigid that it can remove embedded filter struts from the caval wall. The excimer laser (Spectranetics Corp, Colorado Springs, CO, USA) is a promising device for this purpose,⁶ but as it is not approved for filter retrieval in Japan, the Byrd mechanical sheath is an alternative, as described here.

Data Collection

The patients in this study were enrolled consecutively, and data for the following parameters were collected from a prospectively acquired database: patient age, sex, filter dwelling time, technical retrieval success, advanced retrieval techniques used, and procedure-related adverse events. Pre-CT measurements were evaluated by 2 interventional cardiologists with considerable experience in performing IVCF procedures. These cardiologists were blinded to information about the clinical background of the patient and procedure details.

Statistical Analysis

Categorical variables are reported as numbers and percentages, and were compared using Chi-squared test or Fisher’s exact test. Continuous variables are reported as mean±standard deviation and compared with Student’s t-test. Receiver-operating characteristics (ROC) curve analysis was used to identify the cutoff indwelling time for retrieval failure with the standard retrieval technique, and we calculated the area under the curve (AUC) for accuracy. Multiple logistic regression analysis was used to determine the independent factors for necessity of advanced retrieval technique. Log odds with 95% confidence intervals (CI) are reported. A P value <0.05 was considered to indicate statistical significance. All statistical analyses were performed with EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria).¹²

Ethical Considerations

The institutional review board approved the study protocol, and the study procedures were performed as per the Declaration of Helsinki and the Good Clinical Practice guidelines. Patient information was anonymized and de-identified before the analysis. The need for written informed consent from each patient was waived because clinical information was obtained from routine practice records, and no patient refused to participate in this study when contacted for follow-up.

Results

Patient Demographics (Table 1)

Table 1. Demographics of the Study Patients

n	107
Age (years)	61±18 (15–90)
Sex (male)	57 (53%)
PE and DVT/DVT alone	45/62
DVT location (proximal site)
IVC	22 (21%)
Iliac/CFV	63 (59%)
FV/PV	21 (20%)
Infrapopliteal vein	1 (1%)
Bilateral/left/right	14/64/26
Active cancer	18 (17%)
IVC filter indication (SIR)
Classical	23 (21%)
Extended	84 (79%)
Prophylactic	0 (0%)
Type of filter
Günther-Tulip	84 (79%)
Denali	13 (12%)
ALN	8 (7%)
OptEase	2 (2%)
Indwelling time (days)	94±176 (2–1,279)
Suprarenal implantation	11 (10%)
Retry case	12 (11%)

Data are expressed as mean±standard deviation or number (percentage). CFV, common femoral vein; DVT, deep venous thrombosis; FV, femoral vein; IVC, inferior vena cava; PV, popliteal vein; SIR, Society of Interventional Radiology.

A total of 107 patients (mean age: 61±18 years, male 53%) were enrolled from April 2012 to December 2018. All patients developed acute deep venous thrombosis (DVT), and 45 had accompanying acute pulmonary thromboembolism (APE). Thrombus extension of the DVTs was as follows: in the IVC for 22 patients (21%), in the iliac vein in 63 patients (59%), in the superficial femoral vein in 21 patients (20%), and in the infrapopliteal vein in 1 patient (1%). There were 18 patients (17%) with active cancer.

The indication for r-IVCF was classified into 3 groups based on the Society of Interventional Radiology (SIR) guidelines: classic indication, extended indication, and prophylaxis.¹³ In detail, classic indications include absolute contraindication to anticoagulation, complication of anticoagulation resulting in therapy discontinuation, and anticoagulation failure. Extended indications include iliocaval DVT or large, free-floating proximal DVT, difficulty in establishing therapeutic anticoagulation, massive APE treated with thrombolysis/thrombectomy, catheter-directed thrombolysis for iliocaval DVT, VTE with limited cardiopulmonary reserve, poor compliance with anticoagulation, and high risk of complication of anticoagulation. Prophylactic indications were absence of venous thromboembolism (VTE) but risk of developing VTE and inability to take anticoagulants. In total, 23 patients (21%) and 84 patients (79%) were classified as classical and extended indications, respectively. There were no cases of prophylactic indication in this population.

Currently in Japan, 4 r-IVCF devices are officially approved. Of the r-IVCF we tried to retrieve, the Günther-Tulip was the main type of r-IVCF, which accounted for 79% (84 patients), while the Denali, ALN, and OptEase accounted for 12% (13 patients), 7% (8 patients), and 2% (2 patients), respectively.

The average indwelling time of the r-IVCF was 94±176 days, and the range was wide (2–1,279 days). Among the 107 patients, 11 r-IVCFs were deployed in the suprarenal position, and 12 patients had been referred after failure of the first retrieval attempt; an advanced technique was the initial choice in those 12 patients.

Procedural Outcomes

Of the 107 patients who underwent r-IVCF retrieval, the procedure was successful in 80 (75%) with the standard technique alone. Of the 27 patients (25%) in whom retrieval failed, we successfully removed the filters in 25 patients (23%) with an advanced technique, resulting in a total retrieval success rate of 98% (Figure 1).

Figure 1.

IVC filter retrieval success rates. With only the standard technique, the IVC filter retrieval success rate is 75% (80/107). However, the retrieval success rate improved to 98% with advanced techniques. IVC, inferior vena cava.

Figure 2 shows the number and distribution of the different advanced techniques used in this study. The sling technique was most frequently used (22 patients), followed by the biopsy forceps dissection technique (17 patients). The other techniques, such as double sling technique, Byrd mechanical sheath usage, and balloon displacement technique, were used in 5, 5, and 4 patients, respectively. It is noteworthy that a combined approach with more than one advanced technique was used in 18 patients (67%); the most commonly used combined approach was sling technique with biopsy forceps in 10 patients (37%). In contrast, we successfully removed filters with the sling technique alone in 8 patients (30%) and using biopsy forceps alone in 1 patient (4%).

Figure 2.

Advanced techniques used in this study. Total number of patients who required advanced techniques was 27. The sling technique was the most frequently used technique. Combined therapy indicates that more than one advanced technique was used in the same procedure.

Retrieval Failures and Adverse Events Related to Filter Retrieval

We encountered 3 complications associated with the retrieval procedure. The first case was a 60-year-old man with a history of cancer who had developed acute DVT and APE. A Günther-Tulip r-IVCF was implanted before continuous anticoagulation. Two years later, he was referred for filter retrieval after failure of the first attempt at another hospital. The main reason for filter retrieval was the patient’s desire to discontinue anticoagulants. The pre-CT image revealed a small amount of residual thrombus in the left superficial femoral vein and there was no embedded hook in the caval wall. Two leg struts had penetrated the caval wall, protruding beyond the caval wall by 1.3 and 2.1 mm, respectively. The other 2 struts were within the caval wall. The initial procedure was the sling technique under a 14Fr sheath via the right internal jugular vein. We felt strong resistance when pulling the filter against the caval wall and an intussusception developed, leading to occlusion of the IVC with large hematoma around the caval wall. Fortunately, the patient’s condition remained stable during the procedure, but filter retrieval failed. The subsequent clinical course was uneventful but there was bilateral leg edema with disappearance of the hematoma on follow-up CT. The second case was a 30-year-old woman who developed acute DVT during pregnancy and underwent r-IVCF (Günther-Tulip) implantation. A few months after delivery, she presented with bilateral leg edema. Although CT showed thrombotic occlusion of the filter with bilateral iliac vein thrombotic occlusions, filter retrieval failed and she had been followed up with anticoagulation. She was referred to our hospital for filter retrieval after 2.5 years. We successfully removed the filter by combining all devices and approaches. However, a tiny fractured strut was left in the caval wall, which we failed to pick up. She had an uneventful course without symptoms related to residual fractured strut. The final case was a 30-year-old obese woman who presented with acute DVT. She was treated with catheter-directed thrombolysis (CDT) under r-IVCF protection (Günther-Tulip). The massive DVT was cleared after 1 week of CDT, and we attempted to retrieve the filter. We determined there was extreme filter tilting with the hook in the right renal vein. Although the filter was successfully retrieved with the sling technique plus biopsy forceps dissection technique, a large hematoma in the vessel wall around the right renal vein ostium developed after vessel wall injury. The hematoma was absorbed in 1 week. Overall, there were no fatal events in this study population.

There was one more case of failure to retrieve the r-IVCF. A 30-year-old woman developed DVT and underwent r-IVCF (OptEase) placement in the infrarenal position at another hospital. They tried to remove the r-IVCF twice within 2 months after implantation, but were unsuccessful. Three years later, she was referred to us for filter retrieval because of her desire to have a baby. After insertion of a 14Fr long sheath from the right internal carotid vein and the right femoral vein, the double sling technique was used with 0.035 guide wires. Even though a Byrd mechanical sheath was used to dissect the adhesion between the filter struts and the caval wall, we were unable to retrieve the r-IVCF owing to the strong resistance and the patient’s pain.

Indwelling Time and Advanced Techniques

In patients who needed an advanced technique for filter retrieval, the indwelling time was much longer than in patients who underwent the standard technique alone (224±312 days vs. 47±25 days, P<0.0001; Figure 3A). Figure 3B shows the proportion of advanced technique use per the indwelling time. The proportion of use of an advanced technique after 3 months of filter deployment significantly increased in comparison with before 3 months (85% vs. 11.4%, P<0.0001). Of note, we needed to use an advanced technique in some patients even when attempting retrieval before 3 months after filter implantation.

Figure 3.

(A) Comparison of the indwelling time of the IVC filter between advanced technique and standard technique alone. (B) Proportions and numbers of standard and advanced techniques according to indwelling time period (≤30 days, 31–90 days, 91–365 days, >365 days).

Predictive Factors and Cutoff Indwelling Time for the Use of Advanced Techniques

Table 2 shows the comparison of patients with the standard technique and advanced techniques regarding various factors that could be associated with the necessity of using an advanced technique for successful filter retrieval. Patients undergoing an advanced technique were younger than those with the standard technique (P<0.0001). The number of each r-IVCF type we attempted to retrieve varied widely (P=0.015), although the Günther-Tulip accounted for 78.5% of total r-IVCFs. As previously noted, the indwelling time was strongly associated with the necessity of using an advanced technique, and was much longer in patients undergoing an advanced technique (P<0.0001). With respect to the CT measurements, mean LSP length of patients with the standard technique was shorter than that of patients with an advanced technique (P=0.043). The patients with advanced technique showed a tendency towards longer maximum LSP length than those with the standard technique (P=0.074). Therefore, the degree of penetration of strut legs was probably more severe in patients with the advanced techniques. Moreover, the CT finding of an embedded hook in the caval wall was associated with using an advanced technique (P<0.0001). The average radiation time for procedures with the advanced techniques was longer than that with the standard technique (P<0.0001), reflecting the complexity of the advanced techniques. Logistic multivariate analysis revealed that the indwelling time and embedded hook in the caval wall on CT image were independent predictors for the necessity of an advanced technique (Table 3). We estimated the cutoff value for indwelling time of the r-IVC filter for the necessity of an advanced technique with ROC curves. The cutoff value was 80 days (sensitivity 0.741, specificity 0.912) and the AUC was 0.816 (95% CI: 0.697–0.935) (Figure 4).

Table 2. Comparison of Standard vs. Advanced Technique

	Standard technique (n=80)	Advanced technique (n=27)	P value
Age (years)	65±15	49±19	<0.0001
Sex (male)	44 (55%)	13 (48%)	0.69
Type of filter G/D/A/O	64/12/4/0	20/1/4/2	0.015
Indwelling time (days)	47±25	234±312	<0.0001
Mean leg strut penetration length (mm)*	1.24±0.75	1.64±1.0	0.043
Max leg strut penetration length (mm)*	2.53±1.72	3.36±2.42	0.074
Embedded hook in caval wall	22 (28%)	20 (74%)	<0.0001
Radiation time during procedure (min)	8.9±5.8	40.8±32.1	<0.0001

Data are expressed as mean±standard deviation or number (percentage). *Calculated only in patients with Günther-Tulip (G), Denali (D) and ALN (A) filters. O, OptEase.

Table 3. Multivariate Analysis for the Use of Advanced Techniques

	OR	95% CI	P value
Age (years)	0.99	0.95–1.03	0.59
Indwelling time (days)	1.04	1.01–1.06	0.0011
Max penetration length (mm)*	0.893	0.36–2.21	0.807
Mean penetration length (mm)*	1.21	0.131–11.1	0.867
Embedded hook into caval wall	7.81	1.59–38.4	0.0114

*Calculated only in patients with Günther-Tulip, Denali and ALN filters. CI, confidence interval; OR, odds ratio.

Figure 4.

Receiver-operating characteristic curve and cutoff value for indwelling time of the r-IVC filter for necessity of deidentified advanced technique. The cutoff value is 80 days (sensitivity 0.741, specificity 0.912). AUC, area under the curve; CI, confidence interval; r-IVCF, retrievable inferior vena cava filter.

Discussion

Recent studies have demonstrated a significant decline in filter placements after the 2010 FDA safety communication regarding potential long-term complications of filter implantation. Several studies have demonstrated a tendency of increased retrieval rates,^14,15 but they remain lower than expected,^4,16 therefore, potential benefit of this device designed for limited use may be offset by the increased risk of complications associated with prolonged or often indefinite implantation. The most important points we must consider for improving the filter retrieval rate include reiterative follow-up or dedicated IVCF clinics.¹⁷ However, it is evident that we need to use advanced techniques for filter retrieval when the routine standard approach fails.

One distinguishing feature of our study is that we focused on prolonged dwelling time of r-IVCF as the time when standard retrieval techniques fail and advanced retrieval techniques are required, although previous research on prolonged dwelling time has been limited to retrieval failure.

In our study, the retrieval success rate was 75% with the standard technique alone, but this relatively low rate improved to 98% with use of advanced techniques. Previous studies have reported similar retrieval success rates.^18–20 We noted a significant difference between the rate at ≤3 months and that at >3 months after implantation in terms of the proportion of use of advanced techniques. Dasei et al²¹ reported the results of 648 retrieval procedures with various types of r-IVCF in their institution. In their study, advanced techniques were necessary in 95 procedures (14.7%). Although only 52 filter retrieval procedures were performed after 6 months, an advanced technique was more likely to be used in the retrieval procedure performed after 6 months of deployment than in those performed before 6 months (67% vs. 11%, P<0.001). They also reported that advanced techniques were significantly associated with technical success (odds ratio (OR), 3.97 (95% CI, 1.31–11.95); P=0.01). It is easy to understand that filter retrieval after a prolonged dwelling time will more often require complex procedures mainly because of the aggressive endothelialization and adhesion around the hook and struts. However, it is noteworthy that the proportion of needing an advanced technique was distinctively different in our study, with a big difference before and after 3 months of filter deployment. The ROC analysis yielded a cutoff value of 80 days after filter implantation for the necessity of using an advanced technique for retrieval. Desai et al evaluated the time point when the risk of standard retrieval technique failure increases significantly and concluded that 7 months of indwelling time was the inflection point.²² The exact reason for the discrepancy with our study is unclear; however, there are some possible reasons. Firstly, they enrolled more types of IVCF, with the main type being Celect (Cook Medical Inc.), which has easier removal than the Günther-Tulip, the filter used most often in the patients in our study.²³ Another reason was that the mean dwelling time was shorter than in our study (1.9 vs. 3 months), which may be associated with the larger cutoff value for indwelling time in their study. In any case, we need to be cautious about the timing of retrieval and ensure not to extend filter placement beyond the cutoff value. Moreover, we may consider referring patients to centers with expertise in advanced techniques when the indwelling time is >80 days.

The following advanced techniques were used in our practice: biopsy forceps dissection technique, single and double sling technique, balloon displacement technique, and Byrd mechanical sheath usage. In total, two-thirds of the cases of advanced techniques needed more than one technique; the most frequently used combination was the single sling technique with biopsy forceps. Thus, we should be familiar with all the advanced techniques in order to ensure that the retrieval success rate improves. As mentioned earlier, the excimer laser is yet to be approved for filter retrieval in Japan. Several studies have suggested that excimer laser-assisted removal is effective in removing embedded IVCFs refractory to standard retrieval and could be safely used to prevent and alleviate filter-related morbidity.^6,24,25 In our series, we used the Byrd mechanical sheath as an alternative to the excimer laser. Based on our experience, the Byrd sheath can provide as much traction power as the excimer laser to tear the severely adherent tissue.

Generally speaking, aggressive techniques are often associated with increasing levels of complications. The advanced retrieval techniques are more invasive than the simpler standard technique, resulting in possible complications such as vessel perforation and rupture. The Manufacturer and User Facility Device Experience (MAUDE) database contained 111 reports of complications that occurred during r-IVCF removal, which accounted for 13% of the total complications in the short and long term after filter implantation.²⁶ Al-Hakim et al reported that the rate of complications associated with filter retrieval using advanced techniques was significantly higher than when not using an advanced technique (5.3% vs. 0.4%; P<0.05).¹⁹ In our study, we observed 3 complications associated with advanced technique usage. With respect to the huge hematoma in the caval wall, we intentionally injured the caval wall while extracting the tilted hook in the right renal vein using the sling technique with biopsy forceps. The hematoma spontaneously resolved without sequel. One case of intussusception of IVC occurred after strongly pulling out the filter with the single sling technique. After that event, we introduced the double sling technique to maintain the level of the filter when withdrawing it with high traction, thus preventing intussusception. Thereafter, we have not had similar complications because we used the double sling technique from the outset when we expected more difficulty in retrieval. It is noteworthy that no fatal complications occurred in the study population. In summary, filter retrieval using advanced techniques were conducted safely, although the radiation time was significantly longer than with the standard technique.

We sought to identify the predictive factors that help determine whether an advanced technique is needed, including the clinical background and pre-CT images. Previous studies^8,19 have yielded similar results that prolonged dwelling time, increased tilt angle, hook embedment, and strut leg penetration are independent predictors. In our study, logistic multivariate analysis identified prolonged indwelling time and embedded hook as independent predictors; however, leg penetration was not a factor. There are several possible reasons for this inconsistency. The first is that most of the pre-CT images were 2D, which is not as accurate as 3D and may be responsible for leg penetration not being driven as an independent predictor. The length of the leg struts outside of the vessel wall should be calculated on the 3D image, but because most of the pre-CT images were 2D, we had to use them to measure the distance from the vessel wall to the apex of each leg strut. Although this length did not accurately reflect the actual distance, we assumed a positive correlation. As the second reason, Go et al reviewed 262 post-filter CT findings in their institution and found that penetrations, defined as the filter strut being entirely outside the IVC or the filter strut interacting with an adjacent organ outside the IVC, occurred with 49.0% of retrievable filters.²⁷ Oh et al retrospectively assessed 64 pre-CT images before IVCF retrieval and penetrations as defined above occurred in 64% of all cases.⁸ Therefore, penetration of the filter strut is a common phenomenon and may be associated with the filter type and length of indwelling time. Moreover, Oh et al reported that IVCFs with obvious filter strut penetration, such as entirely outside the IVC or interacting with adjacent organs outside the IVC, could be removed safely. It is our belief that the appearance of filter strut penetration on CT images should not be a contraindication for filter retrieval.

Study Limitations

This was single-center study, and the sample size was relatively small compared with previous trials that dealt with IVCF retrieval using advance techniques and related complications. Another limitation is that there was no rule for when and how to use advanced techniques; the decision to use advanced techniques was completely based on operator preference. The threshold of transitioning from the standard technique to advanced technique might vary for each operator. The final limitation is that Günther-Tulip was the dominant IVCF type (79%) in this study. The technical challenges encountered during the retrieval procedure vary by filter type. Therefore, the cutoff value for the necessity of advanced retrieval techniques may change by the device subtype. Therefore, the result may differ according to the IVCF type.

Despite these limitations, to our knowledge this is the first study to provide a detailed report on the advantages of advanced techniques for complicated r-IVCF retrieval procedure in Japan. Thus, this report could contribute to the improvement in the filter retrieval rate, resulting in decreased complications related to unnecessary prolonged r-IVCF implantation.

Conclusions

Advanced filter retrieval techniques were necessary in one-quarter of all filter retrieval cases and use of these advanced techniques helped improve the r-IVCF retrieval rate without serious complications, especially for IVCF with prolonged indwelling time even under the peculiar circumstance of the excimer laser being unapproved for filter retrieval in Japan. Further, the cutoff value of indwelling time for retrieval failure with standard technique that required the use of advanced techniques was 80 days. Therefore, we may need to consider referral of patients to centers with expertise in advanced retrieval techniques when the indwelling time is >80 days and the pre-CT image shows the hook embedded in the vessel wall.

Disclosure

The authors have no conflicts of interest to disclose.

IRB Information

The present study was approved by the committee of Institutional Review Board of Ota Memorial Hospital (reference no. OR20025).

References

• 1.   Smouse B, Johar A. Is market growth of vena cava filters justified?: A review of indications, use, and market analysis. Endovasc Today 2010; 74–77.
• 2.   Grewal S, Chamarthy MR, Kalva SP. Complications of inferior vena cava filters. Cardiovasc Diagn Ther 2016; 6: 632–641.
• 3.   Tao MJ, Montbriand JM, Eisenberg N, Sniderman KW, Roche-Nagle G. Temporary inferior vena cava filter indications, retrieval rates, and follow-up management at a multicenter tertiary care institution. J Vasc Surg 2016; 64: 430–437.
• 4.   Jia Z, Fuller TA, McKinney JM, Paz-Fumagalli R, Frey GT, Sella DM, et al. Utility of retrievable inferior vena cava filters: A systematic literature review and analysis of the reasons for nonretrieval of filters with temporary indications. Cardiovasc Intervent Radiol 2018; 41: 675–682.
• 5.   Iliescu B, Haskal ZJ. Advanced techniques for removal of retrievable inferior vena cava filters. Cardiovasc Intervent Radiol 2012; 35: 741–750.
• 6.   Kuo WT, Odegaard JI, Rosenberg JK, Hofmann LV. Excimer laser-assisted removal of embedded inferior vena cava filters: A single-center prospective study. Circ Cardiovasc Interv 2013; 6: 560–566.
• 7.   Mismetti P, Laporte S, Pellerin O, Ennezat PV, Couturaud F, Elias A, et al. Effect of a retrievable inferior vena cava filter plus anticoagulation vs. anticoagulation alone on risk of recurrent pulmonary embolism: A randomized clinical trial. JAMA 2015; 313: 1627–1635.
• 8.   Oh JC, Trerotola SO, Dagli M, Shlansky-Goldberg RD, Soulen MC, Itkin M, et al. Removal of retrievable inferior vena cava filters with computed tomography findings indicating tenting or penetration of the inferior vena cava wall. J Vasc Interv Radiol 2011; 22: 70–74.
• 9.   Dinglasan LAV, Oh JC, Schmitt JE, Trerotola SO, Shlansky-Goldberg RD, Stavropoulos SW. Complicated inferior vena cava filter retrievals: Associated factors identified at preretrieval CT. Radiology 2013; 266: 347–354.
• 10.   Sumitsuji S, Inoue K, Ochiai M, Tsuchikane E, Ikeno F. Fundamental wire technique and current standard strategy of percutaneous intervention for chronic total occlusion with histopathological insights. JACC Cardiovasc Interv 2011; 4: 941–951.
• 11.   Grewal S, Chamarthy MR, Kalva SP. Complications of inferior vena cava filters. Cardiovasc Diagn Ther 2016; 6: 632–641.
• 12.   Kanda Y. Investigation of the freely-available easy-to-use software “EZR” (Easy R) for medical statistics. Bone Marrow Transplant 2013; 48: 452–458.
• 13.   Young ED, Minocha J. Inferior vena cava filters: Guidelines, best practice, and expanding indications. Semin Intervent Radiol 2016; 33: 65–70.
• 14.   Wadhwa V, Trivedi PS, Chatterjee K. Decreasing utilization of inferior vena cava filters in post-FDA warning era: Insights from 2005 to 2014 nationwide inpatient sample. J Am Coll Radiol 2017; 14: 1144–1150.
• 15.   Ahmed O, Patel K, Patel MV. Declining national annual IVC filter utilization. Chest 2017; 151: 1402–1404.
• 16.   Ahmed O, Wadhwa V, Patel K , Patel MV, Turba UC, Arslan B. Rising retrieval rates of inferior vena cava filters in the United States: Insights from the 2012 to 2016 Summary Medicare Claims Data. J Am Coll Radiol 2018; 15: 1553–1557.
• 17.   Simon TE, Walker PE, Daab LJ, White JM, White PW. A quality improvement project to improve inferior vena cava filter retrieval. J Vasc Surg Venous Lym Dis 2017; 5: 42–46.
• 18.   Minocha J, Idakoji I, Riaz A, Karp J, Gupta R, Chrisman HB, et al. Improving inferior vena cava filter retrieval rates: Impact of a dedicated inferior vena cava filter clinic. J Vasc Interv Radiol 2010; 21: 1847–1851.
• 19.   Al-Hakim R, Kee ST, Olinger K, Lee EW, Moriarty JM, McWilliams JP. Inferior vena cava filter retrieval: Effectiveness and complications of routine and advanced techniques. J Vasc Interv Radiol 2014; 25: 933–939.
• 20.   Glocker RJ, Novak Z, Matthews TC, Patterson MA, Jordan WD, Pearce BJ, et al. Factors affecting Cook Gunther Tulip and Cook Celect inferior vena cava filter retrieval success. J Vasc Surg Venous Lym Dis 2014; 2: 21–25.
• 21.   Desai KR, Lewandowski RJ, Salem R, Mouli SK, Karp JK, Laws JL, et al. Retrieval of inferior vena cava filters with prolonged dwell time: A single-center experience in 648 retrieval procedures. JAMA Intern Med 2015; 175: 1572–1574.
• 22.   Desai KR, Laws JL, Salem R, Mouli SK, Errea MF, Karp JK, et al. Defining prolonged dwell time: When are advanced inferior vena cava filter retrieval techniques necessary?: An analysis in 762 procedures. Circ Cardiovasc Interv 2017; 10: e003957.
• 23.   Lyon SM, Riojas GE, Uberoi R, Patel J, Lipp MEB, Plant GR, et al. Short- and long-term retrievability of the Celect vena cava filter: Results from a multi-institutional registry. J Vasc Interv Radiol 2009; 20: 1441–1448.
• 24.   Kuo WT, Cupp JS. The excimer laser sheath technique for embedded inferior vena cava filter removal. J Vasc Interv Radiol 2010; 21: 1896–1899.
• 25.   Kuo WT, Odegaard JI, Louie JD, Sze DY, Unver K, Kothary N, et al. Photothermal ablation with the excimer laser sheath technique for embedded inferior vena cava filter removal: Initial results from a prospective study. J Vasc Interv Radiol 2011; 22: 813–823.
• 26.   Angel LF, Tapson V, Galgon RE, Restrepo MI, Kaufman J. Systematic review of the use of retrievable inferior vena cava filters. J Vasc Interv Radiol 2011; 22: 1522–1530.
• 27.   Go MR, Keller-Biehl L, Starr JE. Penetration of the inferior vena cava and adjacent organs after filter placement is associated with retrievable filter type and length of time in place. J Vasc Surg Venous Lym Dis 2014; 2: 174–178.