Trends and Predictors of Vena Cava Filter Use (2004–2023): A Swiss Nationwide Epidemiological Study

Abstract

Background

Recent data on inferior vena cava filter (IVCF) placement mainly originate from the United States, with limited evidence from Europe.

Methods

We used nationwide, patient-level data from 2004 to 2023, including hospitalizations with mention of IVCF placement or venous thromboembolism (VTE). Placement was classified as secondary prevention if VTE was documented during the hospitalization, otherwise as primary prevention. We assessed IVCF placement rates overall and in patients with pulmonary embolism (PE), and studied in-hospital case fatality rates, retrieval rates, and length of hospitalization. Analyses were stratified by prevention type and sex.

Results

IVCF placement was recorded in 5,123 (81.1% secondary prevention) patients. The age-standardized rate of IVCF placement increased from 3.4 (95% confidence interval [CI]: 2.0; 4.8) per 1,000 PE-related hospitalizations in 2004 to 19.2 (95% CI: 16.7; 21.6) per 1,000 PE-related hospitalizations in 2023. The IVCF-related in-hospital case fatality rate increased from 7.1% (95% CI: 1.1; 13.2) in 2004 to 10.2% (95% CI: 7.2; 13.3) in 2023 and was higher after IVCF placement for secondary prevention than for primary prevention. The estimated retrieval rate within an average of 6 months was 31.9% (95% CI: 29.9; 34.0), peaking at 34.6% (95% CI: 32.4; 36.8) after excluding deaths during index hospitalization. The retrieval rate decreased progressively over time. IVCFs were less likely to be retrieved among older patients and in patients with intracranial hemorrhage or cancer, whereas the removal rate was higher among patients with recent trauma.

Conclusion

We showed an increasing trend of IVCF placement procedures. The IVCF retrieval rate decreased over time, emphasizing the need for improved follow-up protocols.

Keywords

inferior vena cava filter - venous thromboembolism - retrieval

Introduction

Inferior vena cava filters (IVCFs) are devices placed via endovascular venous access to prevent pulmonary embolism (PE), PE recurrence, and PE-related death. Their efficacy is controversial, and their use is suggested for secondary prevention in patients with venous thromboembolism (VTE) if an anticoagulant therapy is contraindicated or has proven to be ineffective in this patient. Additionally, IVCFs are used for primary prevention in patients who are at a high risk of VTE, but anticoagulant treatment is contraindicated.[1] [2] [3] In the United States, the usage rate of IVCF placement was approximately 160 per 1,000 PE-related hospitalizations in 2010, showing an increasing trend.[4] The case fatality rate of patients receiving IVCF remains high at 7 to 8%.[5] Although the early removal of IVCF is recommended to decrease the risk of filter-related complications, such as deep vein thrombosis (DVT), only about 20% are ultimately retrieved.[4]

Knowledge gaps remain regarding measures of occurrence, current nationwide retrieval rates, and the impact of IVCF placement on in-hospital mortality and the length of hospital stay (LOS).[5] [6] Published data on the retrieval rate of IVCFs are outdated or were adopted from randomized controlled trials.[4] [7] [8] Finally, there is a lack of data from European countries, hampering the evaluation of this treatment option in European health care systems.

In this study, we described IVCFs use and their clinical course in Switzerland over the past 20 years. This includes IVCF placement rate, retrieval rate, in-hospital case fatality rate, LOS, and predictors for placement and retrieval.

Methods

Data Source

We conducted a patient-level retrospective analysis of the Swiss Medical statistics collected by the Swiss Federal Statistical Office, covering all hospitalizations in Switzerland from 2004 to 2023. The Swiss Medical Statistics is a collection of all case-related hospital discharge data. By law, health care institutions are required to collect discharge data and report them annually. Data are anonymized by the Swiss Federal Statistical Office to comply with personal data protection regulations. Diagnoses are collected in accordance with the International Classification of Diseases, 10th revision with German Modifications (ICD-10-GM). Performed procedures are collected as the Swiss classification of surgical interventions (CHOP) codes.

Patients were selected based on the mention of ICD-10-GM codes for VTE and of CHOP codes of IVCF placement. Specific codes for IVCF placement and retrieval have been available since 2011; prior to this, more generic codes were used ([Supplementary Table S1]). Therefore, the analysis of retrieval focuses on data after 2011.

Patients were classified according to the indication for IVCF placement. Primary prevention was defined as placement of an IVCF without any recorded ICD-10-GM codes for VTE ([Supplementary Table S1]) during the same hospitalization, whereas secondary prevention was defined as placement of an IVCF with mention of at least one ICD-10-GM code for VTE ([Supplementary Table S1]) during the same hospitalization.

ICD-10-GM codes for concomitant diseases, selected based on previous literature and clinical relevance, are presented in [Supplementary Table S2]. ICD-10-GM codes and CHOP codes for the definition of PE with high-risk features (cardiopulmonary resuscitation, use of vasopressor or thrombolytic substances, extracorporeal membrane oxygenation [ECMO], mechanical ventilation, surgical embolectomy, shock, cardiac arrest) are presented in [Supplementary Table S3]. At the start of data analysis, data up to the year 2023 were available. ICD-10-GM codes are reviewed for each data year X by May 31 of the year X + 2.[9]

In Switzerland, health care facilities are categorized into general hospitals and specialized clinics, based on the number of medical specialties, teaching competencies across medical specialties, and the reported inpatient volume. For this study, we grouped general hospitals in three categories: (1) University hospitals, (2) major hospitals, and (3) regional hospitals. Specialized clinics were combined into a single category labelled as “Others” ([Supplementary Fig. S1]).[10]

This study was conducted in accordance with the REporting of studies Conducted using Observational Routinely-collected Data (RECORD) standards.[11] Due to the use of anonymized data, approval by an institutional review board was not necessary.

Statistical Analysis

The primary analysis was performed for all VTE-related cases, therefore accounting for VTE reported as primary or concomitant disease during hospitalization or at discharge. We calculated the age-standardized frequency of IVCF as the number of IVCF placements per 10,000 hospitalizations using the 2013 age distribution as standard data. We calculated the age-standardized frequency of IVCFs in patients with PE as the number of patients with PE and IVCF placement per 1,000 PE cases (PE as primary or concomitant diagnosis). As a sensitivity analysis, we additionally calculated this frequency in patients with PE as primary diagnosis or as concomitant diagnosis with diagnostic codes for computed tomography (CT) of the thorax or CT angiography ([Supplementary Table S1]). Furthermore, the annual proportion of deaths in all cases with IVCF placement (case fatality rate) was calculated.

We calculated the annual rate of in-hospital IVCF retrieval as the number of retrieved IVCF per 100 IVCF placements. The grouping of hospitalizations with IVCF retrieval was according to the indication group assigned for the hospitalization with IVCF placement. For the analysis of IVCF retrieval, we only included patients with IVCF placement between January and June, due to the lack of traceability of patients after the end of the calendar year. We chose this timeframe due to IVCF retrieval usually being attempted within 6 months after placement.[7] [8]

We used univariable and multivariable logistical regression models reporting odds ratios (ORs) with 95% confidence intervals (95% CIs) to study factors associated with the placement of IVCF and the retrieval of IVCF. Variables included in the models were identified based on clinical plausibility and relevance.

Data are presented either as count and percentage or, for continuous variables, as median and interquartile range (IQR). The data distribution was visually inspected. Rates are presented as crude annual nationwide rates and, if applicable, as age-standardized rates. For age standardization, we used the age distribution in 2013. To visualize trends, figures are presented with locally estimated scatter point smoothing, and underlying values are presented for each year. The statistical analysis was performed using R version 4.4.1.[12]

Role of the Funding Source

The authors are solely responsible for the content of this work. No external funding was obtained for this study. The corresponding author had full access to all the data and agrees to be responsible for the submission for publication.

Results

As of 2023, Switzerland had approximately 9.0 million inhabitants. During the study period (2004–2023), a mean year-end population of 8,531,876 (50.6% female) inhabitants lived in the country. Out of 323,679 hospitalizations with VTE, IVCF placement was reported in 1.3%. IVCF placement was documented in 5,123 individual hospitalizations: In 4,157 (81.1%) cases, IVCF placement was used for secondary prevention in patients with VTE, whereas filters were placed for primary prevention in the remaining 966 (18.9%) cases. Multiple hospitalizations with recorded IVCF placement or IVCF retrieval occurred in <0.1% of patients.

[Table 1] presents the baseline characteristics of 5,123 patients receiving IVCF for primary or secondary prevention and 319,522 patients with VTE who did not receive IVCF. The prevalence of women was 47% in the group receiving IVCF and 51% in the group without IVCF. The median age was 68 (IQR: 55–77) years in the group receiving IVCF opposed to 71 (IQR: 58–80) years in patients without IVCF. Patients with (vs. without) IVCF placement presented with a higher prevalence of cancer, accidents or trauma, and coagulation defects. Cardiovascular disorders were frequent in all groups. More than half of the IVCF placements were performed at university hospitals, accounting for 55% in the primary prevention group and 53% in the secondary prevention group (vs. 22% of patients with VTE who did not receive IVCF).

Table 1 Baseline characteristics of hospitalizations with inferior vena cava filter placement or venous thromboembolism

Characteristic	Indication for IVCF placement			VTE cases without IVCF placement, N = 319,522	Overall, N = 324,645
	Any, N = 5,123	Primary prevention, n = 966	Secondary prevention, n = 4,157
Age, median (Q1, Q3)	68 (55, 77)	65 (51, 76)	68 (56, 77)	71 (58, 80)	71 (58, 80)
Sex
Women, n (n/N)	2,394 (47%)	440 (46%)	1,954 (47%)	161,640 (51%)	164,034 (51%)
Comorbidities
Cancer, n (n/N)	1,913 (37%)	344 (36%)	1,569 (38%)	88,098 (28%)	90,011 (28%)
Cardiovascular diseases, n (n/N)	3,023 (59%)	464 (48%)	2,559 (62%)	178,311 (56%)	181,334 (56%)
Ischemic heart disease, n (n/N)	565 (11%)	98 (10%)	467 (11%)	39,533 (12%)	40,098 (12%)
Cerebral infarction, n (n/N)	487 (9.5%)	29 (3.0%)	458 (11%)	12,969 (4.1%)	13,456 (4.1%)
Respiratory diseases, n (n/N)	1,951 (38%)	266 (28%)	1,685 (41%)	110,571 (35%)	112,522 (35%)
Accidents and trauma, n (n/N)	3,079 (60%)	594 (61%)	2,485 (60%)	89,218 (28%)	92,297 (28%)
Pregnancy, n (n/N)	75 (1.5%)	9 (0.9%)	66 (1.6%)	4,863 (1.5%)	4,938 (1.5%)
Coagulation defects, n (n/N)	1,482 (29%)	232 (24%)	1,250 (30%)	31,723 (9.9%)	33,205 (10%)
Hospital type
University hospital, n (n/N)	2,731 (53%)	534 (55%)	2,197 (53%)	68,814 (22%)	71,545 (22%)
Major hospital, n (n/N)	2,076 (41%)	368 (38%)	1,708 (41%)	146,890 (46%)	148,966 (46%)
Regional hospital, n (n/N)	304 (5.9%)	60 (6.2%)	244 (5.9%)	73,019 (23%)	73,323 (23%)
Other, n (n/N)	12 (0.2%)	4 (0.4%)	8 (0.2%)	30,799 (9.6%)	30,811 (9.5%)

Abbreviations: IVCF, inferior vena cava filter; VTE, venous thromboembolism.

Percentages are calculated for each column.

[Table 2] presents an overview of the clinical presentation, VTE localization, hemodynamic features, and site of bleeding among patients considered in this study. PE was the most prevalent presentation of VTE, diagnosed in 2,863 patients with IVCF placement, corresponding to 56% of all patients with IVCF and 69% of those with IVCF for secondary prevention, as opposed to 53% among those without IVCF. Bleeding complications were more frequently reported for patients with (vs. without) IVCF placement, notably intracranial hemorrhage (9.1% vs. 1.1%), gastrointestinal hemorrhage (11% vs. 4.3%), and intrathoracic hemorrhage (2.1% vs. 0.4%), respectively. The in-hospital case fatality rate among patients with IVCF placement was 9.1% (6.2% for primary prevention and 9.7% for secondary prevention) versus 9.4% among patients with VTE without IVCF, although the prevalence of high-risk features (mechanical ventilation, use of vasopressors, shock, cardiopulmonary resuscitation, and ECMO) was overall higher among patients with IVCF. A possible indication for IVCF placement (cancer, accidents or trauma, coagulation defects, or hemorrhage) for primary prevention was present in 91.3% of patients.

Table 2 Clinical presentation of hospitalizations with inferior vena cava filter placement or venous thromboembolism

Characteristic	Indication for IVCF placement			VTE cases without IVCF placement, N = 319,522	Overall, N = 324,645
	Any, N = 5,123	Primary prevention, n = 966	Secondary prevention, n = 4,157
Length of hospitalization, median (Q1, Q3)	20 (12, 33)	17 (11, 29)	21 (12, 34)	10 (5, 18)	10 (5, 18)
VTE
PE, n (n/N)	2,863 (56%)	n.a.	2,863 (69%)	183,604 (57%)	186,467 (57%)
DVT without PE, n (n/N)	1,159 (23%)	n.a.	1,159 (28%)	93,923 (29%)	95,082 (29%)
Thrombosis of the vena cava or other thoracic veins, n (n/N)	267 (5.2%)	n.a.	267 (6.4%)	5,829 (1.8%)	6,096 (1.9%)
Thrombosis of the renal veins, n (n/N)	25 (0.5%)	n.a.	25 (0.6%)	1,270 (0.4%)	1,295 (0.4%)
Other VTE without DVT or PE, n (n/N)	88 (1.7%)	n.a.	88 (2.1%)	40,119 (13%)	40,241 (12%)
High-risk features, n (n/N)	1,553 (30%)	246 (25%)	1,307 (31%)	31,130 (9.7%)	32,683 (10%)
Mechanical ventilation, n (n/N)	699 (14%)	136 (14%)	563 (14%)	13,018 (4.1%)	13,717 (4.2%)
Use of vasopressor or thrombolytic substances, n (n/N)	658 (13%)	73 (7.6%)	585 (14%)	13,233 (4.1%)	13,891 (4.3%)
Shock, n (n/N)	647 (13%)	80 (8.3%)	567 (14%)	11,048 (3.5%)	11,695 (3.6%)
Cardiopulmonary resuscitation, n (n/N)	184 (3.6%)	17 (1.7%)	167 (4.0%)	6,084 (1.9%)	6,268 (1.9%)
ECMO, n (n/N)	53 (1.0%)	1 (0.1%)	52 (1.3%)	1,047 (0.3%)	1,100 (0.3%)
Hemorrhage	2,752 (54%)	501 (52%)	2,251 (54%)	42,945 (13%)	45,697 (14%)
Intracerebral hemorrhage, n (n/N)	466 (9.1%)	58 (6.0%)	408 (9.8%)	3,620 (1.1%)	4,086 (1.3%)
Gastrointestinal hemorrhage, n (n/N)	576 (11%)	57 (5.9%)	519 (12%)	11,490 (3.6%)	12,066 (3.7%)
Intrathoracic hemorrhage, n (n/N)	106 (2.1%)	32 (3.3%)	74 (1.8%)	1,198 (0.4%)	1,304 (0.4%)
Peritoneal hemorrhage, n (n/N)	84 (1.6%)	16 (1.7%)	68 (1.6%)	566 (0.2%)	650 (0.2%)
Other hemorrhage, n (n/N)	2,161 (42%)	420 (43%)	1,741 (42%)	33,441 (10%)	35,602 (11%)
Death	464 (9.1%)	60 (6.2%)	404 (9.7%)	25,716 (8.0%)	26,180 (8.1%)
University hospital, n (n/N)	239 (4.7%)	38 (3.9%)	201 (4.8%)	6,280 (2.0%)	6,519 (2.0%)
Major hospital, n (n/N)	205 (4.0%)	20 (2.1%)	185 (4.5%)	13,050 (4.1%)	13,255 (4.1%)
Regional hospital, n (n/N)	20 (0.4%)	2 (0.2%)	18 (0.4%)	5,232 (1.6%)	2,252 (1.6%)
Other, n (n/N)	0	0	0	1,154 (0.4%)	1,154 (0.4%)

Abbreviations: DVT, deep vein thrombosis; ECMO, extracorporeal membrane oxygenation; IVCF, inferior vena cava filter; n.a., not applicable; PE, pulmonary embolism; VTE, venous thromboembolism.

Percentages are calculated for each column.

[Supplementary Table S4] presents an overview of the characteristics of patients with IVCF placement for the period 2004–2013 and 2014–2023. In the second period, patients were older, IVCF was less frequently placed without a VTE diagnosis, and PE was more prevalent. No major sex differences were observed.

Inferior Vena Cava Filter Placement Rate

We found an increasing trend in the age-standardized prevalence of IVCF placement in hospitalized patients rising from 0.5 (95% CI: 0.4; 0.6) per 10,000 hospitalizations in 2004 to 2.4 (95% CI: 2.1; 2.6) per 10,000 hospitalizations in 2023, which was mainly driven by an increasing number of IVCF placements for secondary prevention, specifically in patients with recorded PE. The frequency of placement for primary prevention peaked in 2009, subsequently decreasing ([Fig. 1] and [Supplementary Fig. S2]). The overall frequency of IVCF placement in hospitalized patients with PE increased from 3.4 (95% CI: 2.0; 4.8) per 1,000 PE cases in 2004 to 19.2 (95% CI: 16.7; 21.6) in 2023. With increasing hospital size, the frequency of IVCF placement in patients with PE increased ([Fig. 2]). Among men, the overall placement rate and the placement rate in patients with PE were slightly higher, irrespective of indication; however, observed time trends were similar between sexes ([Supplementary Figs. S3] and [S4]). In a sensitivity only including PE as primary diagnosis or as concomitant diagnosis with diagnostic codes for CT of the thorax or CT angiography, we found similar frequencies and time trends ([Supplementary Fig. S5]).

Fig. 1 Age-standardized prevalence of inferior vena cava filter (IVCF) placed per 10,000 hospitalizations stratified by the indication for placement, 2004 to 2023. Age standardization was performed using the age distribution of hospitalizations in 2013. DVT, deep vein thrombosis; PE, pulmonary embolism.

Fig. 2 Age-standardized prevalence of inferior vena cava filter (IVCF) in pulmonary embolism (PE) patients stratified by hospital type, 2004 to 2023. Age standardization was performed using the age distribution of PE hospitalizations in 2013.

Trends of In-Hospital Case Fatality Rate

The overall in-hospital case fatality rate of patients with IVCF placement increased from 7.1 (95% CI: 1.1; 13.2) deaths per 100 patients with IVCF placement in 2004 to 10.2 (95% CI: 7.2; 13.3) deaths per 100 patients with IVCF placement in 2023. In most years, the rate was higher for IVCF placement for secondary (vs. primary) prevention, peaking at 14.1 (95% CI: 10.2; 17.9) deaths per 100 patients with IVCF placement for secondary prevention in 2022. The case fatality rate was slightly higher in major hospitals than in university hospitals (9.9 [95% CI: 8.6; 11.2] vs. 8.8 [95% CI: 7.7; 9.8]). Females tended to have a lower in-hospital case fatality rate; the observed time trends were similar. [Supplementary Figure S6] presents an overview of the case fatality rate stratified by indication and sex.

Trends in Length of Hospitalization

Overall, the median LOS of patients with IVCF remained steady at around 18 to 20 days. Starting in 2005, patients with IVCF as secondary (vs. primary) prevention presented with a longer LOS. The median LOS for IVCF placement for secondary prevention increased slightly from 14 (IQR: 10–28) days in 2004 to 18 (IQR: 11–31) days in 2023. Patients with IVCF placement for primary prevention presented with a decreasing LOS (29 [IQR: 11–41] days in 2004, 13.5 [IQR: 9–22.75] days in 2023; [Supplementary Fig. S4]). Women tended to have shorter LOS compared to males; however, the observed time trends were similar ([Supplementary Fig. S7]).

Retrieval Rate

In the period 2011 to 2023, retrieval was mentioned in 1,347 cases (959 during index hospitalization, 388 during a later hospitalization), corresponding to an overall retrieval rate of 26.3%. The rate was 31.9% (95% CI: 29.9; 34.0) in the subset of patients with IVCF placement between January and June. The overall retrieval rate was higher among patients with IVCF placement for primary prevention (38.6%; 95% CI: 33.3; 44.0) versus secondary prevention (30.6%; 95% CI: 28.5; 32.9). Only considering patients who survived the index hospitalization, the rate was 40.0% (95% CI: 34.5; 45.5) and 33.5% (95% CI: 31.1; 35.9), respectively. Retrieval within 3 months after placement was recorded in 38.7% (95% CI: 33.2; 44.2) and 27.9% (95% CI: 25.6; 30.2), respectively ([Supplementary Table S5]).

The retrieval rate over time for patients who received an IVCF between January and June decreased from 37.1% (95% CI: 28.3; 45.9) in 2011 to 31.9% (95% CI: 24.7; 39.1) in 2023. In most years, the rate was higher among patients with IVCF placement for primary (vs. secondary) prevention. [Figure 3] and [Supplementary Table S6] present the annual IVFC retrieval rate for patients with IVCF placement between January and June, stratified by the indication for placement. Conditional on survival of the index hospitalization, the retrieval rate was higher both for placement for primary and secondary prevention ([Supplementary Tables S7] and [S8]). In fact, retrieval was rarely attempted in patients who died during the index hospitalization. Observed time trends in the retrieval rate were similar in both sexes ([Supplementary Fig. S9]).

Fig. 3 Retrieval rate of inferior vena cava filter (IVCF) placed between January and June stratified by the indication for placement, 2011 to 2023.

Predictors for Placement and Retrieval

The odds for IVCF placement increased with increasing hospital size. This was evident in univariate and multivariable models with adjustment for age, sex, and comorbidities. Further factors associated with the placement of an IVCF were intracerebral hemorrhage, coagulation defects, PE, and cancer. Age and sex were associated with the placement of IVCF before adjustment; following adjustment, no association was found ([Supplementary Table S7]).

Conditional on survival of the index hospitalization with IVCF placement, age, intracranial hemorrhage, and cancer were negatively associated with retrieval in univariable logistical regression models. In contrast, a positive association was found for accidents or trauma. After adjustment, the negative association remained for age, intracranial hemorrhage, and cancer, with an additional negative association for coagulation defects. Similarly, the positive association of accidents or trauma persisted ([Table 3]).

Table 3 Logistical regression model for inferior vena cava filter removal

	Univariable model (removal vs. no removal), OR (95% CI)	Multivariable model (removal vs. no removal), OR (95% CI)	Multivariable model (removal vs. no removal), OR (95% CI)	Multivariable model (removal vs. no removal), OR (95% CI)
Age	0.98 (0.98; 0.99)	0.98 (0.98; 0.99)	0.98 (0.98; 0.99)	0.98 (0.98; 0.99)
Female sex	0.92 (0.75; 1.11)	0.94 (0.77; 1.15)	0.93 (0.77; 1.14)	0.97 (0.79; 1.18)
Hospital size	1.07 (0.92; 1.24)	–	1.02 (0.87; 1.19)	0.96 (0.82; 1.13)
Intracranial hemorrhage	0.68 (0.47; 0.97)	–	0.70 (0.48; 1.01)	0.69 (0.47; 1.0)
Cancer	0.68 (0.55; 0.84)	–	–	0.76 (0.61; 0.94)
Pulmonary embolism	0.86 (0.71; 1.05)	–	–	0.87 (0.71; 1.06)
Coagulation defects	0.80 (0.65; 1.0)	–	–	0.75 (0.60; 0.94)
Accidents or trauma	1.59 (1.30; 1.96)	–	–	1.51 (1.22; 1.88)
Cardiovascular diseases	0.93 (0.76; 1.13)	–	–	1.10 (0.88; 1.38)
Respiratory diseases	1.06 (0.87; 1.29)	–	–	0.99 (0.79; 1.23)
High risk features	1.22 (0.98; 1.51)	–	–	1.05 (0.83; 1.34)

Abbreviations: OR, odds ratio; 95% CI, 95% confidence interval.

Discussion

We presented a comprehensive overview of the trends and predictors of IVCF use in a European country based on administrative vital registration data. To date, only North American data have been made available. Our results revealed (a) an increasing usage of IVCF placement over time that was mainly driven by patients with placement for secondary prevention, (b) an increasing in-hospital case fatality rate of patients with IVCF over time, and (c) a declining rate of IVCF retrieval over time.

As expected, the frequency of contraindications for anticoagulation was higher in patients with IVCF placement compared to the frequency in VTE patients who did not receive an IVCF, pointing to the indication for IVCF placement.[1] [2] We found a higher frequency of accidents or trauma as concomitant diseases in patients who received an IVCF. Both present possible risk factors for VTE and potential sequelae, especially in PE patients.[13] [14] At the same time, they might lead to bleeding complications, constituting contraindications for anticoagulation and thus a potential deciding factor for IVCF placement.[15] The rate was higher among patients with IVCF placement for primary prevention, which might be explained by accidents or trauma being the underlying cause for the hospitalization or advanced cancer associated with a high risk of VTE. Especially in patients undergoing surgery, physicians face challenges in how to optimize anticoagulation management, which might lead to a more liberal use of IVCF.[1] [16]

Overall, we found an increasing prevalence of IVCF placement per 10,000 hospitalizations. Up to 2010, both the prevalence of placement for primary and secondary prevention showed an increasing trend. Thereafter, the increase was driven by placement for secondary prevention. Evidence on the advantages of IVCF placement for primary prevention is controversial,[17] [18] and the majority of recently published studies do not support a benefit due to the placement for primary prevention and describe poorer outcomes with an increased risk for complications.[19] [20] [21] Thus, the decrease found is in-line with current evidence and indicates patient selection according to guidelines.[22]

In patients with PE, we found large differences in the prevalence of IVCF placement, with a positive correlation with hospital size. This is most likely due to the more frequent availability of departments of interventional radiology, cardiology, or angiology in larger hospitals, and the treatment of more complex patients.[23] [24] In Swiss university hospitals, we found an increasing prevalence of IVCF placement up to 2009, the year in which prevalence peaked at almost 70 IVCF placed per 1,000 PE patients. Subsequently, the rate stabilized at roughly 40 per 1,000 PE patients. In comparison, in 2010, the IVCF placement rate in patients with PE in the United States was roughly 160 per 1,000 patients.[4] Data on IVCF complications have been studied more extensively in the United States compared to Europe. In 2010 and 2014, the U.S. Food and Drug Administration (FDA) issued device safety warnings, leading to a subsequent reduction in the placement rate with a decreasing time trend.[25] [26] Nevertheless, the rate of placement remained high at 39.1 per 100,000 general U.S. population in 2014, possibly much higher than the observed rate in Switzerland.[26]

The in-hospital case fatality rate of patients with IVCF placement for secondary prevention was double the rate in patients with placement for primary prevention. This difference might result from the indication itself; patients with a high risk of death might not be considered for IVCF placement for primary prevention. The rate in patients with placement for secondary prevention was comparable to the in-hospital case fatality rate found for hospitalized PE patients in Switzerland.[27] The effect of IVCF on mortality is controversial.[28] [29] [30] [31] [32] In the United States, a decrease in the in-hospital case fatality rate was observed among stable patients with PE with IVCF who received thrombolytic therapy, as well as among unstable PE patients with IVCF, regardless of thrombolytic therapy, compared to patients without IVCF.[5]

The median LOS of patients with IVCF placement for secondary prevention remained steady at around 18 to 20 days over the study period. In patients with placement for primary prevention, a decrease from 29 in 2004 to 13.5 in 2023 was observed. This decrease may be seen in the context of a progressive decrease in LOS for any pathology observed at a national level. Indeed, patients with IVCF placement tended to be hospitalized for a longer duration compared to VTE patients without IVCF placement. This is most likely driven by the higher rate and severity of comorbidities. The difference between primary and secondary prevention may be explained by the indication itself, as placement for primary prevention could potentially prevent severe sequelae that would necessitate a longer hospitalization. This is reinforced by past findings showing that IVCF for primary prevention in the first 48 hours of hospitalization decreases the length of stay in intensive care and overall LOS.[6]

Over time, we found a decreasing IVCF retrieval rate from 37% in 2011 to 32% in 2023 for filters placed between January and June. The retrieval rate among patients with IVCF placement was higher than the rate reported in the United States, where an overall 1-year retrieval rate of 20 to 25% has been observed. Notably, following the 2010 device safety warning by the U.S. Food and Drug Administration, an increasing time trend was observed.[33] [34] [35] The rate remained far lower than the rates found in the Prevention of Recurrent Pulmonary Embolism by Vena Cava Interruption 2 (PREPIC2) study, a randomized controlled trial that compared retrievable IVCF plus anticoagulation with anticoagulation alone.[7] In PREPIC2, retrieval was attempted in 90% of patients with a success rate of 90% (overall 80% successful retrieval). Furthermore, for Switzerland, we found that most IVCFs were retrieved during the index hospitalization, indicating inadequate reevaluation of the indication during follow-up. The Predicting the Safety and Effectiveness of Inferior Vena Cava Filters (PRESERVE) study demonstrated that a plan for IVCF retrieval at placement and frequent reevaluation of indication might increase the rate of retrieval to 50%.[8] The low retrieval rates and the inconsistencies between various study settings and administrative data may be partly attributed to underreporting. Nonetheless, our findings underscore the need for structured programs to ensure timely retrieval of IVCF,[36] [37] to minimize the risk of complications associated with long-term filter placement.

Device retrieval among patients who survived the initial hospitalization was negatively associated with age, intracranial hemorrhage, cancer, and coagulation defects. The only factor associated with a higher OR for removal was the mention of accidents or trauma. Our results are in accordance with previous findings on predictors of filter retrieval.[38] Maintaining a cava filter might be justifiable under certain circumstances, such as ongoing contraindications to anticoagulation, cava filter thrombosis, and poor prognosis.[39] Increasing age, intracranial hemorrhage, and cancer might represent markers for poor prognosis, and thus, a negative association is to be expected. Accidents or trauma present transient indications for filter placement; therefore, retrieval should be attempted once the acute situation is overcome.

International guidelines are heterogeneous, with varying indications for placement, particularly concerning the use of permanent versus retrievable filters, and the role of placement for primary prevention. Moreover, multiple guidelines omit recommendations on retrieval altogether, whereas those that address retrieval emphasize close patient follow-up with periodic reevaluation of indication and timely retrieval. Therefore, further research is warranted to generate high-quality evidence on appropriate indications for IVCF placement and retrieval, as well as optimal follow-up strategies.[1] [2]

Limitations

This study carries the typical limitations of studies conducted with administrative vital registration data. First, the codes used only consider inpatient hospitalizations; thus, patients discharged without formally being admitted (usually at least one overnight stay) have not been included. This could have lowered the retrieval rate since retrieval might have been performed in an outpatient setting. Second, the ICD-10-GM coding of diagnoses, even if attributed by physicians, could have been inaccurately assigned, leading to possible misclassification of diseases. Procedural codes are more robust overall; nevertheless, inaccuracies in assignment might influence our study. Third, due to working with anonymized data, no manual accuracy checks of ICD-10-GM and CHOP codes were possible. Fourth, the Swiss Medical Statistics does not allow tracing patients beyond the calendar year. Therefore, we only considered patients with filter placement between January and July in the calculation of time trends in the retrieval rate. Nonetheless, this could have led to an underestimation of the retrieval rate. Fifth, no information on the timing of events was available; thus, it is unknown whether cava filter placement happened before or after the thromboembolic event. Therefore, cases of primary prevention might have been classified in the category of secondary prevention. Lastly, due to the limited availability of longitudinal data, no analysis of complications or adverse events of IVCF placement was conducted.

Conclusion

We found an increasing trend of IVCF placement for secondary prevention and a decreasing trend for IVCF placement for primary prevention over time. An IVCF was retrieved in slightly more than one-third of patients, with decreasing trends over time and lower retrieval rates in the elderly, in patients with intracranial hemorrhage, and cancer.

Conflict of Interest

L.H. reports lecture/consultant fees from Boston Scientific, INARI Medical, Penumbra, MSD, and Johnson & Johnson, outside the submitted work.

Authors' Contributions

P.M. and S.B. contributed to the concept and design of the study, interpretation of the results, writing of the manuscript, and final approval of the article. S.W. contributed the concept and design of the study, statistical analysis, interpretation of the results, writing of the manuscript, and final approval of the article. S.C., B.B., L.H., K.K., and N.K. contributed to the interpretation of the results, critical revision of the manuscript, and final approval of the article.

Data Availability Statement

Statistical source code used to generate population estimates and calculation of rates can be obtained from the study statistician (S.W., email: [e-mail]). The Swiss Medical Statistics can be purchased from the Swiss Federal Statistics Office. Demographic data can be downloaded from the webpage of the Swiss Federal Statistics Office.

^* These authors share co-first authorship for this article.

• References

• 1 DeYoung E, Minocha J. Inferior vena cava filters: Guidelines, best practice, and expanding indications. Semin Intervent Radiol 2016; 33 (02) 65-70
• 2 Visconti L, Celi A, Carrozzi L. et al. Inferior vena cava filters: Concept review and summary of current guidelines. Vascul Pharmacol 2024; 155: 107375
• 3 Zuin M, Bikdeli B, Ballard-Hernandez J. et al. International clinical practice guideline recommendations for acute pulmonary embolism: Harmony, dissonance, and silence. J Am Coll Cardiol 2024; 84 (16) 1561-1577
• 4 Stein PD, Matta F, Keyes DC, Willyerd GL. Impact of vena cava filters on in-hospital case fatality rate from pulmonary embolism. Am J Med 2012; 125 (05) 478-484
• 5 Elkbuli A, Ehrhardt JD, Kinslow K, McKenney M. Timing of prophylactic inferior vena cava filter placement and trauma outcomes: Does early placement matter?. Am Surg 2020; 86 (12) 1741-1747
• 6 Bikdeli B, Wang Y, Minges KE. et al. Vena caval filter utilization and outcomes in pulmonary embolism: Medicare hospitalizations from 1999 to 2010. J Am Coll Cardiol 2016; 67 (09) 1027-1035
• 7 Mismetti P, Laporte S, Pellerin O. et al; PREPIC2 Study Group. 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 (16) 1627-1635
• 8 Johnson MS, Spies JB, Scott KT. et al. Predicting the Safety and Effectiveness of Inferior Vena Cava Filters (PRESERVE): Outcomes at 12 months. J Vasc Interv Radiol 2023; 34 (04) 517-528.e6
• 9 Swiss DRG. Coding Revision. Accessed July 22, 2024 at: https://www.swissdrg.org/de/rehabilitation/st-reha-system-302025/kodierrevision
• 10 Swiss Federal Statistic Office. Sections Health Services. Statistik der stationären Betriebe des Gesundheitswesens. Krankenhaustypologie 2016; 5: 2
• 11 Benchimol EI, Smeeth L, Guttmann A. et al; RECORD Working Committee. The REporting of studies Conducted using Observational Routinely-collected health Data (RECORD) statement. PLoS Med 2015; 12 (10) e1001885
• 12 R Core Team (2025). _R: A Language and Environment for Statistical Computing_. R Foundation for Statistical Computing, Vienna, Austria. <https://www.R-project.org/>.
• 13 Wolf S, Barco S, Di Nisio M. et al. Epidemiology of deep vein thrombosis. Vasa 2024; 53 (05) 298-307
• 14 Barco S, Ende-Verhaar YM, Becattini C. et al. Differential impact of syncope on the prognosis of patients with acute pulmonary embolism: A systematic review and meta-analysis. Eur Heart J 2018; 39 (47) 4186-4195
• 15 Xiao N, Desai KR. Venous thromboembolism in trauma: The role of anticoagulation and inferior vena cava filters. Semin Intervent Radiol 2021; 38 (01) 40-44
• 16 Leon L, Rodriguez H, Tawk RG, Ondra SL, Labropoulos N, Morasch MD. The prophylactic use of inferior vena cava filters in patients undergoing high-risk spinal surgery. Ann Vasc Surg 2005; 19 (03) 442-447
• 17 Woodward CC, Eby M, Gandhi R. et al. Do prophylactic inferior vena cava filters affect the rates of venous thromboembolism and pulmonary embolism in patients undergoing major spine surgery?. Global Spine J 2023; 13 (07) 1909-1917
• 18 Minocha J, Smith AM, Kapoor BS. et al; Expert Panel on Interventional Radiology. ACR Appropriateness Criteria^® radiologic management of venous thromboembolism-inferior vena cava filters. J Am Coll Radiol 2019; 16 (5S): S214-S226
• 19 Kelkar AH, Rajasekhar A. Inferior vena cava filters: A framework for evidence-based use. Hematology (Am Soc Hematol Educ Program) 2020; 2020 (01) 619-628
• 20 Kelkar AH, Rajasekhar A. Do prophylactic inferior vena cava filters in trauma patients reduce the risk of mortality or pulmonary embolism?. Hematology (Am Soc Hematol Educ Program) 2020; 2020 (01) 629-633
• 21 Bikdeli B, Chatterjee S, Desai NR. et al. Inferior vena cava filters to prevent pulmonary embolism: Systematic review and meta-analysis. J Am Coll Cardiol 2017; 70 (13) 1587-1597
• 22 Konstantinides SV, Meyer G, Becattini C. et al; ESC Scientific Document Group. 2019 ESC Guidelines for the diagnosis and management of acute pulmonary embolism developed in collaboration with the European Respiratory Society (ERS). Eur Heart J 2020; 41 (04) 543-603
• 23 Hayirli TC, Warinner CB. Hospital characteristics associated with the availability of interventional radiology facilities and services. Radiology 2023; 307 (01) e221189
• 24 Sharma N, Schwendimann R, Endrich O, Ausserhofer D, Simon M. Variation of daily care demand in Swiss general hospitals: Longitudinal study on capacity utilization, patient turnover and clinical complexity levels. J Med Internet Res 2021; 23 (08) e27163
• 25 Van Ha TG. Complications of inferior vena caval filters. Semin Intervent Radiol 2006; 23 (02) 150-155
• 26 Reddy S, Lakhter V, Zack CJ, Zhao H, Chatterjee S, Bashir R. Association between contemporary trends in inferior vena cava filter placement and the 2010 US Food and Drug Administration Advisory. JAMA Intern Med 2017; 177 (09) 1373-1374
• 27 Wolf S, Valerio L, Fumagalli RM. et al. Acute pulmonary embolism with and without hemodynamic instability (2003-2022): A Swiss nationwide epidemiologic study. J Thromb Haemost 2025; 23 (04) 1340-1351
• 28 Stein PD, Matta F, Hughes PG, Hughes MJ. Adjunctive therapy and mortality in patients with unstable pulmonary embolism. Am J Cardiol 2020; 125 (12) 1913-1919
• 29 Muriel A, Jiménez D, Aujesky D. et al; RIETE Investigators. Survival effects of inferior vena cava filter in patients with acute symptomatic venous thromboembolism and a significant bleeding risk. J Am Coll Cardiol 2014; 63 (16) 1675-1683
• 30 Stein PD, Matta F, Hughes MJ. Prophylactic inferior vena cava filters in patients with fractures of the pelvis or long bones. J Clin Orthop Trauma 2018; 9 (02) 175-180
• 31 Lee SJ, Fan S, Guo M. et al. Prophylactic IVC filter placement in patients with severe intracranial, spinal cord, and orthopedic injuries at high thromboembolic event risk: A utilization and outcomes analysis of the National Trauma Data Bank. Clin Imaging 2022; 91: 134-140
• 32 Sarosiek S, Rybin D, Weinberg J, Burke PA, Kasotakis G, Sloan JM. Association between inferior vena cava filter insertion in trauma patients and in-hospital and overall mortality. JAMA Surg 2017; 152 (01) 75-81
• 33 Brown JD, Raissi D, Han Q, Adams VR, Talbert JC. Vena cava filter retrieval rates and factors associated with retrieval in a large US cohort. J Am Heart Assoc 2017; 6 (09) e006708
• 34 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 (11) 1553-1557
• 35 Ferro EG, Mackel JB, Kramer RD. et al. Postmarketing surveillance of inferior vena cava filters among US Medicare beneficiaries: The SAFE-IVC Study. JAMA 2024; 332 (24) 2091-2100
• 36 Inagaki E, Farber A, Eslami MH. et al. Improving the retrieval rate of inferior vena cava filters with a multidisciplinary team approach. J Vasc Surg Venous Lymphat Disord 2016; 4 (03) 276-282
• 37 Stevens H, Bortz H, Chao S. et al. Improving the rate of inferior vena cava filter retrieval through multidisciplinary engagement. Res Pract Thromb Haemost 2023; 7 (01) 100040
• 38 Mission JF, Kerlan Jr RK, Tan JH, Fang MC. Rates and predictors of plans for inferior vena cava filter retrieval in hospitalized patients. J Gen Intern Med 2010; 25 (04) 321-325
• 39 Millward SF, Oliva VL, Bell SD. et al. Günther Tulip Retrievable Vena Cava Filter: Results from the Registry of the Canadian Interventional Radiology Association. J Vasc Interv Radiol 2001; 12 (09) 1053-1058

Publication History

Received: 04 September 2025

Accepted: 11 September 2025

Article published online:
25 November 2025