Prevention of pulmonary embolisms associated with vena cava filter implantation

Abstract

Objective:

To investigate the prevention of pulmonary embolism (PE) by the implantation of inferior vena cava filters (IVCFs).

Methods:

Data of 1058 patients treated for lower extremity deep vein thrombosis (DVT) between January 2005 and January 2012 were analyzed retrospectively and divided into those with and without PE.

Results:

Of the 1058 patients, 34 had PEs (3.2%) and 1024 did not. Multivariable analyses showed that PEs were less likely to occur in patients with IVCFs than in those without IVCFs (odd ratio [OR] = 0.028, P < 0.001), and were more likely to occur in patients with vena cava thrombosis than in those without vena cava thrombosis (OR = 19.094, P < 0.001). Analyses stratified by DVT site showed that vena cava thrombosis was the only risk factor of PE for patients with a left DVT (crude OR = 12.814, P < 0.001), and IVCF implantation was the only protective factor for PE for patients with a right or bilateral DVT (crude OR = 0.028, P = 0.001). IVCF patency rate for follow-up was 98.7% with no filter migration, declination, or failure of expansion.

Conclusions:

IVCFs can reduce the occurrence of PEs in patients with DVTs.

Keywords

Deep vein thrombosis indications patency rate pulmonary embolism inferior vena cava filter

Introduction

The incidence of pulmonary embolism (PE) in patients with a deep vein thrombosis (DVT) is 50–70%, and the death rate as a result of PE is 0.7–4%.^1–3 An effective way to prevent PE in patients at high risk of thromboembolic events when anticoagulation therapy is not possible is the implantation of an inferior vena cava filter (IVCF).⁴ Although IVCFs have been in clinical use for 45 years, IVCF implantation is still controversial, in part because the indications for IVCF placement are not clear and the long-term risks of IVCF implantation remain uncertain.⁵ Permanent filters may decrease the risk of symptomatic PEs, but increase risk of DVTs occurring later (up to 8 years), and there may be no long-term benefit from filter implantation according to the results of the PREPIC Study Group.⁶

Complications after placement of an IVCF include filter migration, obstruction and fracture, vena cava thrombosis, access site thrombosis, hematoma, and infection.^4,7 Non-thrombotic complications are rare, but include retroperitoneal hematoma, upper gastrointestinal bleeding, sepsis from duodenal perforation, arteriovenous fistula, hydroureter, and pancreatitis, all stemming from filter strut erosion affecting structures near the IVCF.⁸ Guidelines for IVCF implantation have been developed by the National Institute of Health and Clinical Excellence (NICE) and The Society of Interventional Radiology Standards of Practice Committee,⁹ and may contribute to increasing IVCF usage. However, a review study in 2011 concluded that evidence of the effectiveness and risks of IVCF placement was limited by the small number of prospective studies.⁷

The purpose of the present study was to investigate prevention of PEs with IVCF implantation in patients with DVTs by analysis of over 2000 cases of DVTs treated at our institution.

Patients and methods

Patients

A total of 2115 patients with DVTs were treated at our institution between January 2005 and January 2012. The data of 1058 patients with lower extremity DVTs who met the inclusion criteria were included in this retrospective analysis. Inclusion criteria were (1) age > 18 years; (2) temporary or permanent risk of PE; (3) clinical indications for IVCF implantation (recurrent PE, contraindications for anticoagulation, DVT occurring in the course of anticoagulant therapy, and bleeding complications of anticoagulation therapy); (4) right lower extremity DVT or floating within inferior vena cava thrombosis; and (5) DVT requiring embolectomy or catheter thrombolysis. Patients with severe coagulation disorders, metastatic cancer, uncontrolled infection, recent surgery or trauma, allergies to contrast agents or vena cava filters, vena cava diameter >30 mm, or thromboembolism blocking filter implementation were excluded. Study subjects were divided into two groups: patients with and without PE.

PEs were diagnosed based on patient history, vital signs, pulmonary radioisotope scan (lung scintigraphy), and computed tomographic pulmonary angiography (CTPA) results. Right lower extremity DVT and lower extremity DVT extending to the inferior vena cava were chosen as the indications for IVCF implantation. All patients provided signed informed consent at the time of admission for their records to be anonymously evaluated for future research. The Internal Review Board of our institution reviewed and approved the study protocol.

Anticoagulation therapy

All included patients with a diagnosis of DVT received anticoagulation therapy for 7–10 days using subcutaneous, low-molecular-weight heparin (LMWH) 5000 U every 12 h, superimposed with oral warfarin for the first three days at a dosage of 6 mg, 6 mg, and 3 mg, respectively. Coagulation parameters were checked after three days, and the oral warfarin dosage was increased or decreased in order to maintain an international normalized ratio (INR) of 2.0–2.5. When the INR reached 2.0, LMWH was discontinued. Patients were discharged after INR values had stabilized, usually after three weeks. During follow-up, INR values were checked at least once every month and the warfarin dose was adjusted to maintain an INR between 2.0 and 2.5. If there were no side effects or adverse effects associated with anticoagulation therapy, warfarin was continued for at least three months.

IVCF implantation

Patients in whom IVCFs were implanted received the SegrCaretype Aegisy™-XJLX3260 vena cava filter (Lifetech Scientific Inc., Shenzhen, Guangdong, China). In addition to meeting the aforementioned inclusion/exclusion criteria, all patients were required to meet anatomic criteria established for filter placement; inferior vena cava (IVC) inner diameter < 30 mm, ideally between 17 and 28 mm, and no presence of thromboembolism blocking implementation. Filters were prepared following the manufacturer’s instructions for use, and either femoral or jugular access was obtained under ultrasonographic guidance as previously described by Funaki.¹⁰ Fifteen patients had a floating embolism in the wall of the inferior vena cava, and to prevent embolism migration in these cases, the IVCF was implanted between the liver and kidney from the right carotid vein.

Retrieval of IVCFs

Indications for IVCF retrieval included clearance of DVT by embolectomy or aggressive anticoagulant and thrombolytic therapy for three weeks. Contraindications for IVCF retrieval included thrombosis at the IVCF site and contraindications to anticoagulation. The SegrCaretype Aegisy™-XJLX3260 filter consists of quasi-lantern asymmetric structures with the remote end extended by a retrievable hook with an embedded spiral wireform by which the filter can be hooked for recovery. Retrieval was accomplished through either the femoral or jugular vein using a two-sheath system through which the outer sheath was passed over the filter and both were removed by means of a snare device.

Statistical analysis

Age was expressed as mean and standard deviation, and other categorical data were expressed as count and percentage. Comparisons between those with and without PE were performed by two independent samples t-test for age, and by Fisher’s exact test for other categorical data. Logistic regression analyses were performed to evaluate the factors associated with PE. Variables with P < 0.2 in univariable logistic regression analyses were entered into multivariable logistic regression analyses. A two-tailed P < 0.05 was considered statistically significant. Statistical analyses were performed using SPSS 15.0 (SPSS Inc, Chicago, IL, USA).

Results

Patient demographic and clinical characteristics

Of the 1058 patients with lower extremity DVTs included in the study, there were 621 males and 437 females (mean age, 53 years) and 34 had PEs (3.2%). Over 90% (n = 31) of PEs occurred within 10 days after the occurrence of the DVT. There were no significant differences in age, gender, or DVT type between the patients with and without PEs. Patient data are summarized in Table 1, and a summary flowchart is presented in Figure 1.

Figure 1.

Summary flowchart.

Table 1.

Patient characteristics.

		Total (n = 1058)	Pulmonary embolism		P-value
		Total (n = 1058)	No PE (n = 1024)	PE (n = 34)	P-value
Age (y)		53.0 (13.7)	53.1 (13.7)	52.2 (13.2)	0.711
Gender	Female	437 (41.3)	423 (41.3)	14 (41.2)	1.000
Gender	Male	621 (58.7)	601 (58.7)	20 (58.8)	1.000
Site of DVT	Right	115 (10.9)	98 (9.6)	17 (50.0)	<0.001*
	Left	931 (88.0)	918 (89.6)	13 (38.2)
	Bilateral	12 (1.1)	8 (0.8)	4 (11.8)
Type of DVT	Ileo-femoral	50 (4.7)	47 (4.6)	3 (8.8)	0.424
	Calf	100 (9.5)	97 (9.5)	3 (8.8)
	Thigh	908 (85.8)	880 (85.9)	28 (82.4)
Vena cava thrombosis	Yes	36 (3.4)	29 (2.8)	7 (20.6)	<0.001*
Vena cava thrombosis	No	1,022 (96.6)	995 (97.2)	27 (79.4)	<0.001*
IVCF implantation	Yes	171 (16.2)	169 (16.5)	2 (5.9)	0.068
IVCF implantation	No	887 (83.8)	855 (83.5)	32 (94.1)	0.068

DVT: deep vein thrombosis; PE: pulmonary embolism; VCF: vena cava filter.

Age was expressed as mean and standard deviation, and other categorical data were expressed as count and percentage.

*Indicates a significant association with the occurrence of PE.

Eighty-eight percent of DVTs were on the left side, 10.9% on the right side, and 1.1% were bilateral. Most patients had thigh DVTs (85.8%). The distribution of DVT sites was significantly different between PE and non-PE groups: 89.6% of patients without PEs had DVTs on the left side, while 38.2% of patients with PE had left-side DVTs. On the other hand, 9.6% of patients without PEs had DVTs on the right side, while 50% of patients with PEs had DVTs on the right side (P < 0.001). In addition, the frequencies of PEs in patients with right side DVTs (17/115, 14.8%) and bilateral DVTs (4/12, 33.3%) were significantly higher than those with left-side DVTs (13/931, 1.4%) (all, P < 0.001).

Among all patients, 171 had VCF implantation. Vena cava thrombosis was significantly associated with the occurrence of PE (P < 0.001), and was found prior to IVCF placement in 20.6% of patients in the PE group, but in only 2.8% of patients in the non-PE group (P < 0.001) (Table 1).

PE and mortality

Fifteen of the 1058 patients died for a mortality rate of 1.4%, and all 15 patients that died were in the PE group. Survival from DVT onset in these 15 patients ranged from 2 to 8 days (Figure 2). The follow-up time for surviving patients ranged from 20 to 58 months (median, 38 months).

Figure 2.

Plot of pulmonary embolism occurrence time (n = 34).

IVCF implantation

IVCFs were implanted in 171 patients (16.2%), including 151 implanted via the femoral vein and 20 implanted via the jugular vein. Of the 171 implanted IVCF, 139 were successfully retrieved. Comparison of patients with and without PE showed that 5.9% (n = 2) and 16.5% (n = 169) had IVCF implantation, respectively, but no statistically significant difference was found in univariable analysis (P = 0.063). A total of 149 patients out of 171 VCF patients had continued further follow-up for 39 ± 19 months (range, 1 month to 7 years). Of the 149 patients, 20 had ultrasonography and/or angiographic evidence of thrombus. Eighteen patients had no inferior vena cava obstruction after thrombolysis treatment, while thrombosis persisted in two patients resulting in obstruction of the inferior vena cava beyond the IVCF. The VCF cumulative patency rate during follow-up was 147/149 = 98.7%, with no filter migration, declination, or failure of expansion.

Independent factors affecting the occurrence of PE

Multivariable analyses (Table 2) showed that PEs were less likely to occur in patients with IVCF implantation than in those without IVCF implantation (odd ratio [OR] = 0.028, P < 0.001), and were more likely to occur in patients with vena cava thrombosis than in those without vena cava thrombosis (OR = 19.094, P < 0.001). In addition, PEs were more likely to occur in patients with right or bilateral DVT compared to those with left DVT (OR = 27.125, P < 0.001). However, analysis of the occurrences of PE stratified by DVT site and VCF implantation suggested that DVT site could be a confounder to the association between VCF implantation and PE, since the protective effect of VCF implantation was only observed in those with right and bilateral DVT, but not in those with left side DVT (Table 3). In fact, no significant difference in PE occurrence was observed between those with VCF implantation and without VCF implantation when the DVT was on the left side. However, for patients with a right or bilateral DVT, the rate of PE occurrence in those with VCF implantation was significantly lower than in those without VCF implantation (right: 1.6% vs. 30.2%, P < 0.001; bilateral: 0 vs. 80%, P = 0.01).

Table 2.

Factors affecting the occurrence of pulmonary embolism.

		Univariable analyses		Multivariable analyses
		OR (95% CI)	P-value	OR (95% CI)	P-value
Age (y)		0.995 (0.971, 1.020)	0.711
Gender	Female	0.995 (0.497, 1.991)	0.988
Gender	Male	Reference
Site of DVT	Right/bilateral	13.990 (6.807, 28.750)	<0.001*	27.125 (12.185, 60.379)	<0.001*
Site of DVT	Left only	Reference		Reference
Type of DVT	Ileo-femoral	Reference
	Calf	0.485 (0.094, 2.492)	0.386
	Thigh	0.498 (0.146, 1.699)	0.266
Vena cava thrombosis	Yes	8.895 (3.582, 22.090)	<0.001*	19.094 (4.514, 80.763)	<0.001*
Vena cava thrombosis	No	Reference		Reference
IVCF implantation	Yes	0.316 (0.075, 1.332)	0.117	0.028 (0.005, 0.161)	<0.001*
IVCF implantation	No	Reference		Reference

CI: confidence interval; DVT: deep vein thrombosis; OR: odds ratio; PE: pulmonary embolism; VCF: vena cava filter.

*Indicates a significant association with the occurrence of PE.

Table 3.

Occurrence of pulmonary embolism stratified by DVT site and VCF implantation.

		VCF implantation		P-value
		No	Yes	P-value
Site of DVT	Left	12/829 (1.4%)	1/102 (1.0%)	>0.999
	Right	16/53 (30.2%)	1/62 (1.6%)	<0.001*
	Bilateral	4/5 (80.0%)	0/7 (0.0%)	0.010*

DVT: deep vein thrombosis; VCF: vena cava filter.

*Indicates a significant difference in PE occurrence was observed between patients who received and did not receive VCF implantation.

Based on the aforementioned findings, further analyses stratified by DVT site were performed (Table 4). The analysis showed that vena cava thrombosis was the only risk factor of PE for the 937 patients with a left DVT, with a crude OR = 12.814 (P < 0.001), and IVCF implantation was the only protective factor for PE for the 127 patients with a right or bilateral DVT, a crude OR = 0.028 (P = 0.001).

Table 4.

Factors affecting the occurrence of PE for patients with a left DVT and for patients with a right or bilateral DVT.

		Left DVT only (n = 937)		Right or bilateral DVT (n = 127)
		OR (95% CI)	P-value	OR (95% CI)	P-value
Age (years)		0.984 (0.946, 1.024)	0.438	1.010 (0.974, 1.047)	0.592
Gender	Female	1.219 (0.406, 3.656)	0.724	0.867 (0.331, 2.269)	0.771
Gender	Male	Reference		Reference
Type of DVT	Ileo-femoral	Reference		Reference
	Thigh	0.271 (0.058, 1.276)	0.099	0.828 (0.087, 7.846)	0.869
	Calf	0.262 (0.023, 2.974)	0.280	0.533 (0.038, 7.487)	0.641
Vena cava thrombosis	Yes	12.814 (3.286, 49.968)	<0.001*	2.882 (0.781, 10.640)	0.112
Vena cava thrombosis	No	Reference		Reference
IVCF implantation	Yes	0.674 (0.087, 5.239)	0.706	0.028 (0.004, 0.216)	0.001*
IVCF implantation	No	Reference		Reference

CI: confidence interval; DVT: deep vein thrombosis; OR: odds ratio; PE: pulmonary embolism; VCF: vena cava filter.

*Indicates a significant association with the occurrence of PE.

Discussion

The results of this study demonstrated that PEs were less likely to occur in patients with IVC filter implantation than in those without it. Importantly, vena cava thrombosis was the only risk factor of PE in patients with a left DVT and IVCF implantation was the only protective factor for PE in patients with a right or bilateral DVT. The PE rate was relatively low (3.2%), with onset occurring most often within 10 days after DVT. The death rate also was low (1.4%), with deaths occurring in 15 patients who developed PEs within 10 days after DVT.

Indications for IVCF implantation

In the present study, the indications for IVCF implantation were recurrent PE, contraindications for anticoagulation, DVT occurring in the course of anticoagulant therapy, bleeding complications of anticoagulation, right lower extremity DVT or floating within inferior vena cava thrombosis, and patients with a DVT requiring embolectomy or catheter thrombolysis. In patients with thigh DVTs, which was the most prevalent type, no PEs were noted, which is possibly related to Cockett’s syndrome.¹¹ The incidence of PEs in patients with lower extremity DVTs associated with IVC embolism was much higher than that of patients with ileo-femoral DVTs, and the death rate was higher than that in the ileo-femoral-type group. The present study showed that PEs (including fatal PE) mostly occurred within 10 days after the onset of a DVT, with 35 days as the longest period. Other studies have reported that the longest duration of fatal PE was 21 days after onset of DVT.¹² Based on results of the present and prior studies, we suggest the following strategy for IVCF implantation. (1) If the patient has right leg DVT accompanied by distal IVC embolism, then IVCF implantation is necessary. (2) A double-purpose IVCF is needed if the patient has both right leg DVT and left leg DVT for young and elderly patients, and elderly patients should receive a permanent VCF. (3) IVFCs should be implanted when anticoagulation has failed to produce the desired results and PE is high risk. (4) IVCFs should be implanted if there is residual DVT after the first PE. (5) Anticoagulation should be continued after IVCF implantation. (6) IVCFs should be implanted within 10 days of DVT occurrence. It should be noted, however, that currently large DVTs in many cases are being managed with thrombolysis with or without thrombectomy. This management has the potential to change the strategy for IVCF usage.

Stringent indications for IVCF placement have been published by The Society of Interventional Radiology Standards of Practice Committee,⁹ and include evidence of PE or IVC/iliac/femoropopliteal DVT and one or more absolute or relative contraindications for anticoagulation therapy, complications or failure of anticoagulation therapy, or recurrent PE despite therapy. In the absence of thromboembolic disease, indications are temporary risk of PE in severe trauma without documented PE or DVT, closed head injury, spinal cord injury, long-bone or pelvic fractures, and high-risk situations such as prolonged immobilization.

Complications and effectiveness of IVCFs

A frequent problem with retrievable IVCFs is that the filter may adhere to the IVC wall, which can complicate or prevent removal. Leaving an IVCF in place causes long-term complications including chronic IVC thrombosis, lower extremity DVT, and development of post-thrombotic syndrome.^6,13 In the present study, post-implantation complications included emboli in the IVCF implantation site in 24 patients, and 15 of these occurred while the patients were hospitalized. Differentiation of embolisms caused by IVCFs themselves or by secondary embolisms is difficult, and the incidence of IVC embolisms in patients with IVCFs ranges from 2% to 30%.¹⁴ Considering that a PE occurred during hospitalization and was likely due to migration of the embolism, these patients were given anticoagulation treatment.

Though complications can occur with IVFCs, their effectiveness has been demonstrated. A recent study of the safety and effectiveness of the Crux IVCF in 125 patients reported three cases of PE, two cases of failure to deploy the filter, and no cases of filter migration, embolization, fracture or tilting of the filter, and the authors concluded that overall the Crux IVCF performed safely with a 98% clinical success rate.¹⁵ A study of perforation of the IVC by filter struts found that most filters involved in perforation were retrievable, and the authors advised that filter retrieval should be attempted.¹⁶ Another study found that filter tilt, malpositioning, and longer duration of filter placement are common causes of retrieval failure, but that retrievability is still preferable to placing permanent filters.¹⁷ IVCF placement has been associated with an increased incidence of proximal DVT, and the risk seems to be greater as duration of indwelling increases.⁴

This study has several limitations. This was a retrospective analysis which precludes determining causality. The lower PE rate observed in the filter group is an association based on case selection and practice within our institution, and does not indicate causation. This by its very nature is selection biased, and the data do not provide evidence that those who did not have PEs with a filter would have had a PE without a filter. The study lacks sufficient follow-up data to allow evaluation of long-term effects in all patients, including PE development and complications.¹⁸ The study also did not address placement of IVCFs in patients with PEs and no evidence of DVT. The identified PE's were not categorized as minimal, intermediate, and severe, which can influence the significance of IVCF placement. The effectiveness of the filters was not identified by filters with cloth. Lastly, there remains a great deal of controversy regarding patient selection and indications for IVCF implantation; if patient selection and indications for filer placement were different, the results and conclusions may be different.

Conclusions

The results of this retrospective review suggest that IVCFs can reduce the occurrence of PEs in patients with right and bilateral DVTs. Implantation of the IVCF used in this study was associated with a high patency rate for follow-up without filter migration, declination, or failed expansion.

Footnotes

Conflict of Interest

All the authors have no conflict of interest and nothing to disclose.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Sarani

Chun

Venbrux

. Role of optional (retrievable) IVC filters in surgical patients at risk for venous thromboembolic disease. Am J Surg 2005; 201: 957–964.

Allen

Carter

Morris

. Retrievable vena cava filters in trauma patients for high-risk prophylaxis and prevention of pulmonary embolism. Am J Surg 2005; 189: 656–661.

Stefanidis

Paton

Jacobs

. Extended interval for retrieval of vena cava filters is safe and may maximize protection against pulmonary embolism. Am J Surg 2006; 192: 789–794.

Spencer

Bates

Goldberg

. A population-based study of inferior vena cava filters in patients with acute venous thromboembolism. Arch Intern Med 2010; 170: 1456–1462.

Harvey

Hopkins

McCafferty

. Inferior vena cava filters: What radiologists need to know. Clin Radiol 2013; 68: 721–732.

The PREPIC Study Group. Eight-year follow-up of patients with permanent vena cava filters in the prevention of pulmonary embolism: the PREPIC (Prevention du Risque d'Embolie Pulmonaire par Interruption Cave) randomized study. Circulation 2005; 112: 416–422.

Angel

Tapson

Galgon

. Systematic review of the use of retrivable inferior vena cava filters. J Vasc Interv Radiol 2011; 22: 1522–1530.

Malek

Kwolek

Conrad

. Presentation and treatment outcomes of patients with symptomatic inferior vena cava filters. Ann Vasc Surg 2013; 27: 84–88.

Caplin

Nikolic

Kalva

. Quality improvement guidelines for the performance of inferior vena cava filter placement for the prevention of pulmonary embolism. J Vasc Interv Radiol 2011; 22: 1499–1506.

10.

Funaki

. Inferior vena cava filter insertion. Semin Intervent Radiol 2006; 23: 357–360.

11.

Dong

. Surgical treatment of Cockett's syndrome in patients with deep vein thrombosis of lower extremities. Zhonguo Yi Xue Yuan Xue Bao 2007; 29: 51–54. (In Chinese).

12.

Sing

Camp

Heniford

. Timing of pulmonary emboli after trauma: implications for retrievable vena cava filters. J Trauma 2006; 60: 732–735.

13.

Kuo

Tong

Hwang

. High-risk retrieval of adherent and chronically implanted IVC filters: Techniques for removal and management of thrombotic complications. J Vasc Interv Radiol 2006; 20: 1548–1556.

14.

Athanasoulis

Kaufman

Halpern

. Inferior vena cava filters: review of a 26-year single-center clinical experience. Radiology 2006; 216: 54–66.

15.

Smouse

Mendes

Bosiers

. The RETRIEVE trial: safety and effectiveness of the retrievable crux vena cava filter. J Vasc Interv Radiol 2012; 24: 609–621.

16.

Wood

Malgor

Gasparis

. Reporting the impact of inferior vena cava perforation by filters. Phlebology 2013; May 28. (Epub ahead of print).

17.

Dixon

Stavropoulos

. Improving retrieval rates for retrivable inferior vena cava filters. Expert Rev Med Devices 2013; 10: 135–141.

18.

Hidaka

Hachisuga

Tsukimori1

. Intrapartum placement of an inferior vena cava filter for a woman with hereditary antithrombin III deficiency: Its role in the prevention of fatal pulmonary embolism. J Obstet Gynaecol Res 2008; 34: 95–99.