U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Cover of Evidence review for the use of inferior vena caval filters in people with venous thromboembolism

Evidence review for the use of inferior vena caval filters in people with venous thromboembolism

Venous thromboembolic diseases: diagnosis, management and thrombophilia testing

Evidence review H

NICE Guideline, No. 158

London: National Institute for Health and Care Excellence (NICE); .
ISBN-13: 978-1-4731-3735-6

Inferior vena caval filters for people with venous thromboembolism (VTE)

Review question

What is the effectiveness of inferior vena caval filters to prevent PE in people with confirmed VTE?

Introduction

The inferior vena cava (IVC) drains the blood from the lower parts of the body and legs into the right atrium of the heart. If a DVT in the lower body or legs becomes dislodged, it will pass through the inferior vena cava and right heart into the pulmonary arteries, causing a pulmonary embolus which damages the lungs and may cause death. IVC filters are devices placed within the IVC to trap larger travelling thromboemboli and stop them reaching the pulmonary circulation.

An IVC filter is placed through a needle puncture operation in a groin or neck vein by an interventional radiologist. It is reserved for more serious cases of VTE. In recent years, IVC filters have seen increasing usage in certain groups of people with VTE, such as those people with VTE and a contraindication to anticoagulation or those people thought to be at particularly high risk of a PE (such as those people who have already had a PE, and people with iliac vein (proximal) DVTs). The 2012 NICE guideline recommended temporary inferior vena caval filters for people with proximal DVT or PE who cannot have anticoagulation treatment, with the filter being removed when the person becomes eligible for anticoagulation treatment. However, these recommendations were informed by relatively few studies and only one RCT assessing the use of filters, and concerns have been raised that in light of new evidence, these recommendations may no longer be appropriate.

This review aims to examine the current evidence for the use of IVC filters to prevent PE in people with VTE. It identified studies that fulfilled the conditions listed in Table 1. For full details of the review protocol, see appendix A.

Table 1. PICO for IVC filters for people with VTE.

Table 1

PICO for IVC filters for people with VTE.

Methods and process

This evidence review was developed using the methods and process described in developing NICE guidelines: the manual (2014). Methods specific to this review question are described in the review protocol in appendix A and the methods section in Appendix B.

Additional methodological issues were as follows:

1.

During protocol development. the committee identified six distinct clinical scenarios where filters may be used in people with VTE:

  • who cannot have anticoagulants, or
  • who have had a filter inserted because they could not take anticoagulants, but retain the filter once they start taking anticoagulants, or
  • who have a PE whilst taking anticoagulants, or
  • who have the filters inserted for prophylaxis before a potential provoking event (e.g. surgery), or
  • who can receive anticoagulants but are at high risk of poor outcomes if they had further PEs, or
  • who can receive anticoagulants but are at high risk of PE occurrence (for people with an initial event that was DVT) or recurrence (for people with an initial event that was PE). (See note 9 below.)

These groups were analysed in separate meta-analyses. In cases where there was a mixed population studies were excluded unless data could be extracted for the populations separately or if the majority of participants fell into a single population based on committee judgement. Downgrading for indirectness was considered in the latter case with the final decision being based on committee judgement.

2.

In addition, the committee identified people with VTE and cancer as a subgroup of interest for all of the above scenarios. Several studies were identified that included people with VTE and cancer, but it was unclear whether these participants fell under the group of people who can receive anticoagulants but are at high risk of PE (new or further PE) or whether having cancer put the participants at high risk of poor outcomes if they had further PEs. As a result, with committee agreement, these studies were analysed separately. No data for people with VTE and cancer was identified for the other scenarios.

3.

Outcomes of relevance to this review were reported at different time points:

  • during the procedure (this outcome is only relevant to those people undergoing surgery)
  • in-hospital
  • in the short term (up to 30 days)
  • at 3 months
  • in the long term (after 3 months).

When a study reported data on two time points within the same grouping (for example, 1 year and 8 years) the later time point was extracted.

4.

RCTs, prospective cohort studies and retrospective cohort studies were not pooled in the meta-analyses (see protocol deviation). However, they were shown on the same forest plots to facilitate visual comparison of the results.

5.

Risk of bias was assessed using the modified version of the first Cochrane risk of bias tool (Higgins, 2011) for RCTs and ROBINS-I for cohort studies. Risk of bias was assessed at the study (as opposed to the being conducted for each outcome). However, in instances where an additional risk of bias applied specifically to one outcome or group of outcomes (for example subjective versus objective outcomes), this was noted in the evidence table and reflected in the relevant GRADE table.

6.

The ROBINS-I risk of bias checklist has 5 possible overall ratings for risk of bias: low, moderate, serious, critical and no information. In the forest plots, cohort studies at moderate, serious and critical risk of bias were included in the same plots. However, sensitivity analyses were carried out as follows:

  • In cases where studies were at low to critical risk of bias, studies at serious and critical risk of bias were removed
  • In cases where studies were at low to critical risk of bias, studies at critical risk of bias were removed
  • In cases where studies were at serious to critical risk of bias, studies at critical risk of bias were removed.

7.

In the GRADE tables, cohort studies at low risk of bias from ROBINS-I were deemed to have no serious risk of bias, those at moderate risk of bias were deemed to be at serious risk of bias, while those at serious and critical risk were grouped as very serious risk of bias. There were no studies in the no information category. Pooled results were presented for each study type (RCT, prospective and retrospective cohort) separately and then for the sensitivity analyses.

8.

Data was available for subgroup analyses for people aged at least 80 years of age who were classed as being at high risk of poor outcomes in the event of a PE-occurrence. Due to a lack of data no other subgroup analyses were performed.

9.

DVT or PE-occurrence in this chapter refers to a DVT or PE developing in a person who already has confirmed VTE (DVT and/or PE). Unless otherwise stated, it is not clear whether the person had specifically a DVT or a PE as the index event. Subgroup analyses by index event were carried out where data was available.

10.

PE-occurrence in some studies was separated into symptomatic and asymptomatic PE-occurrence, whereas most studies simply reported the number of PE events without indicating whether they were symptomatic or asymptomatic. As these studies are retrospective, they are unlikely to have captured asymptomatic events as it is unlikely that routine imaging was used. Therefore, unless otherwise stated, PE-occurrence will be assumed to relate to symptomatic PE for the purposes of this review.

Declarations of interest were recorded according to NICE’s 2018 conflicts of interest policy.

Protocol deviation

Priority screening was not used for this review. All references returned by the search were screened at title and abstract level.

The committee decided that due to the serious/critical risk of bias associated with the included retrospective cohort studies, these should not be pooled with RCTs. Therefore, these study types were analysed separately and no sensitivity analysis by study type was necessary.

Clinical evidence

Included studies

This review was conducted as part of a larger update of the 2012 NICE VTE management guideline (CG144). A systematic literature search for randomised controlled trials (RCTs), cohort studies (retrospective or prospective) and systematic reviews (SRs) was conducted for this review and this returned 2,416 references (see appendix C for literature search strategy). Based on title and abstract, 2,373 references were excluded because they did not meet the review protocol, and 43 references were ordered for full text screening.

Of the 43 references screened as full texts, 21 references, reporting data on 20 unique studies met the inclusion criteria specified in the review protocol for this question (appendix A). The clinical evidence study selection is presented as a diagram in appendix D. Systematic reviews were used as a source of primary studies and were then excluded.

This review was carried out at the end of the update of the VTE management guideline and as a result, no rerun searches were carried out for this question.

Please see appendix E for the full evidence tables. The references of individual included studies are listed in appendix K.

Excluded studies

See Appendix J for a list of references for excluded studies, with reasons for exclusion.

Summary of clinical studies included in the evidence review

The 20 included studies were assigned to the categories as follows (see note below):

  • 2 retrospective cohort studies for people with VTE who cannot have anticoagulants (Table 2)
  • 2 cohort studies (1 retrospective, 1 retrospective analysis of a prospective cohort) for people with VTE who had a PE whilst taking anticoagulants (Table 3)
  • 4 studies (3 retrospective cohort studies, 1 RCT) for people with VTE who have filters inserted for prophylaxis before a potential provoking event (Table 4)
  • 7 retrospective cohort studies for people with VTE who are at high risk of poor outcomes in the event of PE-occurrence (Table 5)
  • 2 RCTs (3 papers) for people with VTE who are at high risk of PE-occurrence (Table 6)
  • 5 studies (4 retrospective cohort studies, 1 RCT) for people with VTE and cancer (Table 7)

Table 2. Studies including people with VTE who cannot have anticoagulants.

Table 2

Studies including people with VTE who cannot have anticoagulants.

Table 3. Studies including people with VTE who had a PE whilst taking anticoagulants.

Table 3

Studies including people with VTE who had a PE whilst taking anticoagulants.

Table 4. Studies including people with VTE who have the filters inserted for prophylaxis before a potential provoking event.

Table 4

Studies including people with VTE who have the filters inserted for prophylaxis before a potential provoking event.

Table 5. Studies including people with VTE who are at high risk of poor outcomes in the event of PE-occurrence.

Table 5

Studies including people with VTE who are at high risk of poor outcomes in the event of PE-occurrence.

Table 6. Studies including people with VTE who are at high risk of PE-occurrence.

Table 6

Studies including people with VTE who are at high risk of PE-occurrence.

Table 7. Studies including people with VTE and cancer.

Table 7

Studies including people with VTE and cancer.

Note: Decousus 1998 and PREPIC 2005 reported data on the same study. Stein 2018a and White 2016 each reported data for two distinct relevant populations and so are included in the tables below twice (and twice in the lists above).

See appendix E for full evidence tables.

Quality assessment of clinical studies included in the evidence review

See appendix E for the risk of bias assessments for individual studies, appendix F for forest plots and appendix G for full GRADE tables. Please refer to the evidence statement section for an overall summary of the evidence.

Economic evidence

Included studies

A systematic search was conducted to identify economic evaluations published since the 2012 update of the guideline. The search returned 233 records. In addition, 1 paper from the economic evidence review for the 2012 guideline was identified. Of the 234 records, 231 were excluded on the basis of title and abstract. The 3 remaining papers were screened in full and only the 1 study (from the 2012 guideline) met the criteria for inclusion in the evidence review.

Excluded studies

Details of excluded studies with reasons for their exclusion are provided in appendix J. For full references of excluded studies, please see appendix K.

Summary of economic studies included in the evidence review

Sarasin (1993) conducted a cost-utility analysis comparing the use of an IVC filter to observation (no treatment) in people with cancer and confirmed VTE. The analysis also considered a third strategy, immediate long-term anticoagulation but this strategy does not fall within the scope of the protocol for this review question. Separate analyses were conducted for the DVT and PE subpopulations. A decision tree was used to simulate the short-term impact of the interventions and a Markov model was used to capture the lifelong differences in recurrence, bleeding, mortality and filter-related complications. The Markov model was run over a lifetime horizon using monthly cycles. The analysis was carried out from the health provider perspective in the US.

The probabilities of VTE, bleeding, mortality and relative treatment effects were sourced from published observational studies or single arm trials. Because there was no study reporting the risk of VTE in people with cancer not receiving anticoagulation, the authors assumed that the probability of developing PE after a DVT was the same as in a population without cancer. It was also assumed that all embolic events would occur in the first month following index DVT or PE. Death was associated to specific events such as acute bleeding, long term consequences of bleeding, IVC filter insertion and PE-occurrence, or to excess cancer mortality. The model accounted for the costs of the interventions, complications (VTE, IVC filter insertion) and death, which were expressed in US dollars (1989/91). Utility values were obtained from consensus of clinicians’ opinions. Both costs and outcomes were discounted annually at 5%.

In the base case, the IVC filter option dominated the no treatment strategy as the IVC filter was both cheaper and generated more QALYs. IVC filter and anticoagulation remained more cost effective than no treatment when parameters were varied deterministically. The cost of filter insertion was estimated to be $1500 but the model assumed much higher costs ($3100 to $4600) in the event of a thromboembolic recurrence such that overall, costs for the IVC filter strategy were lower than no treatment. The analysis was strongly influenced by the short life expectancy of people with cancer, which reduced the likelihood of complications from the device.

The study was classified as being partially applicable because it was conducted from a non-UK NHS perspective. Full incremental cost-effectiveness results were reported only for people with lung cancer and costs were discounted at 5% annually. The study was categorised as having very serious limitations because utilities were estimated from expert opinion, the source of funding was not reported, it was unclear how studies for some clinical parameters were identified and probabilistic sensitivity analysis was not conducted.

Evidence statements

Clinical evidence statements
People with VTE who cannot take anticoagulation

Very low quality evidence from up to 2 retrospective cohort studies reporting data on up to 129,047 people with VTE who cannot take anticoagulation could not differentiate any-cause mortality (in the 30 days following admission) or PE-occurrence (in the year following admission) between in people with a filter compared to people without a filter.

Very low-quality evidence from 1 retrospective cohort study reporting data on 3,017 people with VTE who cannot take anticoagulation found a reduction in any-cause mortality (in the 3 months following admission) in people with a filter compared to people without a filter.

Very low-quality evidence from 1 retrospective cohort study reporting data on 3,017 people with VTE who cannot take anticoagulation found an increase in DVT-occurrence (in the 3 months following admission) in people with a filter compared to people without a filter.

People with VTE who have a PE whilst taking anticoagulants

Very-low quality evidence from 1 prospective cohort study reporting data on 139 people who experienced a PE within 3 months of an initial VTE found a reduction in all-cause mortality and PE related mortality at 30 days in people with a filter compared to people without a filter.

Very-low quality evidence from 1 retrospective cohort study reporting data on 814 people who experienced a PE within 3 months of an initial VTE found a reduction in all-cause mortality during hospital stay and at 3 months in people with a filter compared to people without a filter.

Very-low quality evidence from 1 prospective cohort study reporting data on 139 people who experienced a PE within 3 months of an initial VTE could not differentiate VTE-recurrence and major bleeding at 30 days between people with a filter and people without a filter.

People with VTE who have the filters inserted for prophylaxis before a potential provoking event

Low to moderate quality evidence from 1 RCT reporting data on 141 people with VTE undergoing surgery found a reduction in PE-occurrence (during the surgical procedure, during hospital stay and up to 2 years of follow-up) in people with a filter compared to people without a filter.

Very low quality evidence from 1 RCT reporting data 141 people with VTE undergoing surgery could not differentiate DVT-occurrence up to 2 years between people with a filter compared to people without a filter.

Moderate quality evidence from 1 RCT reporting data on 141 people could not estimate an effect for all-cause mortality (during the surgical procedure or in-hospital) and DVT-occurrence (during the surgical procedure or in-hospital) as both arms reported 0 events.

Very low quality evidence from up to 2 retrospective cohort studies reporting data on up to 1,787 people with VTE undergoing surgery found a reduction in in-hospital all-cause mortality, all-cause mortality at 3 months, PE-related mortality (in-hospital and at 3-months), in-hospital PE-occurrence (overall population and specifically in those people who received anticoagulation following the procedure) in people with a filter compared to people without a filter.

Very low quality evidence from up to 2 retrospective cohort studies reporting data on up to 1,787 people with VTE undergoing surgery could not differentiate the following outcomes between people with a filter compared to people without a filter:

  • in-hospital all-cause mortality (specifically in those people at least 80 years old)
  • in-hospital PE-occurrence (specifically in those with a contraindication to anticoagulation during their hospital stay)
  • all-cause mortality at 30 days, 3 months and 2 years.
  • PE-occurrence at 1 year
  • DVT-occurrence at 1 year

People with VTE who are at high risk of poor outcomes in the event of a PE

Very low quality evidence from up to 6 retrospective cohort studies reporting data on 446,762 people with VTE at high risk of poor outcomes in the event of a PE found a reduction in in-hospital all-cause mortality (overall, specifically in those at least 80 years of age and specifically in those with massive PE) and all-cause mortality at 3 months in people with a filter compared to people without a filter.

Very low quality evidence from 1 retrospective cohort study reporting data on 11, 218 people with VTE at high risk of poor outcomes in the event of a PE found an increase in in-hospital all-cause mortality in people with a filter compared to people without a filter.

Very low quality evidence from 1 prospective cohort study reporting data on 375 people with VTE at high risk of poor outcomes in the event of a PE found a reduction in all-cause mortality at 30 days in people with a filter compared to people without a filter.

Very low quality evidence from 2 retrospective cohort studies reporting data on 3,380 people with VTE at high risk of poor outcomes in the event of a PE could not differentiate PE-related mortality in-hospital or at 3 months between people with a filter and people without a filter.

Sensitivity analysis excluding studies at critical risk of bias

Very low quality evidence from 1 retrospective cohort study reporting data on 425,875 people with VTE at high risk of poor outcomes in the event of a PE found a reduction in all-cause mortality (in-hospital) in people with a filter compared to people without a filter.

People with VTE who are at high risk of PE-occurrence

High quality evidence from 1 RCT reporting data on 400 people with VTE at high risk of recurrence found a reduction in symptomatic PE-occurrence at 8 years in people with a filter compared to people without a filter.

High quality evidence from 1 RCT reporting data on 400 people with VTE at high risk of recurrence found no meaningful difference in rates of post-thrombotic syndrome at 8 years between people with a filter compared to people without a filter.

Very low to moderate quality evidence from up to 2 RCTs reporting data on up to 799 people with VTE at high risk of recurrence could not differentiate the following outcomes between people with a filter compared to people without a filter:

  • all-cause mortality at 12 days, 3 months, 6 months and 8 years.
  • VTE-recurrence at 3 months, 6 months and 8 years.
  • PE-related mortality at 3 months, 6 months and 8 years.
  • symptomatic PE-occurrence at 12 days, 3 months and 6 months.
  • DVT-occurrence at 3 months, 6 months and 8 years.
  • major bleeding at 12 days, 3 months, 6 months and 8 years.
  • asymptomatic or symptomatic PE-occurrence at 12 days.

People with VTE and cancer

Low quality evidence from 1 RCT reporting data on 64 participants could not differentiate all-cause mortality, PE-occurrence, major bleeding or IVC complications all at 3 months between people with a filter compared to people without a filter.

High quality evidence from 1 RCT reporting data on 64 people could not estimate an effect for DVT-occurrence at 3 months as both arms reported 0 events.

Very low quality evidence from 1 retrospective cohort study reporting data on up to 35,034 people with VTE and cancer found a reduction in in-hospital all-cause mortality (overall and specifically in those aged 80 years or older) and in all-cause mortality at 3 months, in people with a filter compared to people without a filter.

Very low quality evidence from 3 retrospective cohort studies reporting data on up to 15,374 people with VTE and cancer found an increase in long-term DVT-occurrence in people with a filter compared to people without a filter.

Very low quality evidence from 2 retrospective cohort studies reporting data on up to 15,270 people with VTE and cancer found an increase in all-cause mortality at 30 days, 3 months and 1 year and in VTE recurrence at 1 year in people with a filter compared to people without a filter.

Very low quality evidence from 3 retrospective cohort studies reporting data on up to 15,374 people with VTE and cancer could not differentiate the following outcomes between people with a filter compared to people without a filter:

  • PE-occurrence (long term)
  • major bleeding (long-term)

Sensitivity analysis excluding studies at critical risk of bias

Very low quality evidence from 1 retrospective cohort study reporting data on up to 1,270 people with VTE and cancer found an increase in DVT-occurrence at 1 year between people with a filter compared to people without a filter.

Very low quality evidence from 1 retrospective cohort study reporting data on up to 1,270 people with VTE and cancer could not differentiate the PE-occurrence at 1 year between people with a filter compared to people without a filter.

Economic evidence statements

A partially applicable study with very serious limitations (Sarasin et al.,1993) assessed the cost effectiveness of IVC filter versus no treatment in people with cancer and confirmed VTE. The IVC filter strategy was found to be dominant (more effective and less expensive). The results were robust to one-way sensitivity analysis. No probabilistic sensitivity analysis was conducted.

The committee’s discussion of the evidence

Interpreting the evidence
The outcomes that matter most

IVC filters are used to prevent thromboemboli from travelling into the pulmonary circulation in a number of clinical scenarios (see Table 1 and discussions below). VTE-recurrence (particularly PE-occurrence) is therefore one of the most important outcomes when assessing the effectiveness of IVC-filters. PE-occurrence can increase the risk of mortality, therefore VTE and all-cause mortality are also important outcomes. It may also lead to chronic thromboembolic pulmonary hypertension.

The committee agreed that the use of IVC filters is accompanied by a risk of filter complications IVC filter-site thrombosis may occur due to entrapment of an embolus which is the intended action of the filter or in situ thrombus which is an unintended action of the filter. It is usually impossible to determine which of these two scenarios has occurred. IVC filter-site thrombosis may result in IVC occlusion resulting in lower limb swelling. Paradoxically occluded filters may rarely act as a source of new or further pulmonary emboli. Filters may tip and angulate, either spontaneously or post-partum or after abdominal surgery, altering their filter effect and the ability to retrieve them. Filter struts may perforate the IVC wall. The clinical significance of this is uncertain unless there are adverse effects on adjacent structures which is uncommon. Other rarer complications are bleeding (access vessel or IVC) filter migration, fracture and embolisation to the lungs and filter infection. The committee agreed that these were also important outcomes to consider and could be linked to increased mortality in some cases.

The committee noted the importance of having results for outcomes in the short and long term. Short term outcomes (outcomes up 3 months) are important because IVC filters are often placed in people with an acute risk of thrombosis (such as people at high risk of PE and people undergoing surgery or other provoking events) and short term follow up will capture the effects of the filter in these situations. Additionally, complications may arise as a result of placing the filter or in the period immediately following insertion. Long-term evidence (outcomes occurring after 3 months) is important in people receiving filters that are expected to be left in for a longer duration (such as those people who had a recurrent event whilst taking anticoagulation). Filters have a risk of delayed complications such as filter angulation, migration, infection, fracture and IVC perforation or thrombosis which may occur at a differential rate in the long-term. It is therefore important that both time points are captured.

The quality of the evidence

The committee agreed that the best evidence available was from RCTs but that most of the available evidence came from retrospective cohort studies which compared groups of people with VTE who received a filter to those who did not. The committee agreed that the decision to place a filter is usually based on a variety of important clinical characteristics (including severity of the PE, general health etc.) and therefore a group of people who received a filter and a group who did not are likely to be very different populations. Some studies attempted to account for this disparity by identifying important clinical characteristics (such as indicators of PE severity) which contribute to making a filter more likely to be placed and adjusting for these factors.

Studies typically account for these confounders either by using propensity matching (matching the participants in the filter group to a pair in the no-filter group based on important clinical characteristics and excluding non-matched participants) or adjusting their analysis by propensity score (attributing a score to each participant for the likelihood of receiving a filter based on important clinical characteristics and applying this score as a coefficient in the analysis). However, the committee were concerned that differences between groups are likely to remain even after adjustment, and many studies either did not adjust their analysis or adjusted for only a few confounders.

All the cohort studies suffered from immortal time bias as follow-up began at the point of admission to hospital, but filters were placed at a later point in time. Therefore, the filter group cannot die in the period between admission and when the filter is placed (deaths in this period would have been placed in the no-filter group), however all deaths in the no-filter group were counted. This can lead to an overestimation of the benefit of filters in reducing mortality. Some studies attempted to account for immortal time bias by excluding all events occurring within 24 hours of admission, however as filters are often placed after 24 hours, this inadequately accounts for the bias. A more appropriate method (as used in White 2016) was to match participants across groups based on the propensity for receiving a filter, with people in the no-filter group having to be alive on the day their matched pair had their filter placed.

The committee agreed with the risk of bias assessments of the cohort studies and with marking down for risk of bias those studies that adjusted for a limited number of confounding factors or failed to adjust/adjusted poorly for immortal time bias. The committee were particularly concerned about studies judged to be at critical risk of bias because these studies typically did not adjust for confounders and /or immortal time bias. The committee agreed that it was important to include these studies in the review due to the scarcity of other studies, but that it was useful to conduct sensitivity analyses in which these studies were excluded from meta-analysis (along with additional sensitivity analyses excluding those studies at serious risk of bias).

There was high heterogeneity between the included cohort studies in the confounders the studies adjusted for, in the inclusion criteria and in the results obtained in some scenarios when meta-analysis was possible.

There were only two retrospective cohort studies looking at people with VTE who cannot take anticoagulation, both of which suffered from methodological issues. Data on actual use of (and contraindication to) anticoagulation was not captured by the sources these studies used to obtain their data. Instead, these studies used other available clinical characteristics to indicate a contraindication (such as active bleeding). The committee advised that the criteria used in these studies would very likely indicate a contraindication to anticoagulation but agreed that studies using active bleeding alone (White, 2016) are unlikely to capture all relevant participants.

Both of these studies adjusted for the likelihood of receiving a filter using various important clinical characteristics. The committee agreed that this improved their confidence in the evidence but advised that the inclusion of people with distal DVT (in Turner, 2018) is a risk of bias as these people are typically not candidates for a filter and will likely be over-represented in the no-filter group. Additionally, the evidence was inconsistent as the larger study (Turner, 2018) showed increased all-cause mortality at 30 days and White (2016) showed a reduction at 30 days and 3 months.

The committee were concerned with the very low quality of the limited evidence available for people with VTE who have had a PE whilst taking anticoagulants, noting that Stein (2019a) could not determine whether participants were on anticoagulation when the recurrent event occurred and did not adjust for confounders or immortal time bias. The committee agreed that there were fewer methodological concerns surrounding Mellado (2016) as this study used a prospective collected database and used propensity matching, however the sample size was small.

The evidence for people having filters inserted for prophylaxis before a potential provoking event came from people with VTE who were having a filter placed before surgery. The committee were concerned with the differences in study populations contained within this group as the type of surgery differed between studies. Additionally, three of the studies only included people undergoing a specific intervention (such as percutaneous endovenous treatment of acute symptomatic lower limb proximal DVT) whereas one study (White 2016) included all people with VTE undergoing major surgery. The committee advised that the nature of these interventions is very diverse. Consequently, it is difficult to generalise the available evidence to all people with VTE undergoing surgery. The committee agreed that there is a need for better evidence for specific types of surgery for which IVC filters may be more appropriate as a prophylactic measure (such as surgery in people with cancer).

The committee advised that there were considerable population differences in the studies looking at people at high risk of poor outcomes in the event of a PE (as the conditions that predisposed them to poor outcomes varied between studies) and studies looking at people with VTE who were having a filter inserted prior to surgery (as the type of surgery differed between studies). The evidence for this population subgroup was of very-low quality and came from retrospective cohort studies, the majority of which were at critical risk of bias. As such, it was hard to draw conclusions from the meta-analyses for these groups.

The committee noted that there were two RCTs looking at the use of filters in people at high risk of PE-occurrence; that these studies were at low risk of bias and reported outcomes into the long-term (PREPIC 2005 reported outcomes up to 8 years). However, these studies were unable to differentiate between filter and no-filter for the majority of outcomes. Additionally, the criteria for high risk of PE differed between studies: PREPIC (2005) defined risk according to physician’s judgement whereas Mismetti (2015) required that the person meet pre-specified criteria for risk. The committee noted that based on these criteria, Mismetti (2015) included a mixture of people at high risk of PE, at high risk of poor outcomes, and people with VTE who do not meet the inclusion criteria for this review. The committee agreed that this evidence should be downgraded for indirectness because of this, but that it should remain in this section of the review. In addition, they noted that despite the differences in the characteristics of the participants, the studies showed similar results for most outcomes.

There were 4 retrospective cohort studies at moderate to critical risk of bias and 1 RCT with low risk of bias reporting data on people with VTE and cancer. The committee were concerned that the evidence for people with VTE and cancer was inconsistent. Stein (2018c) noted a reduction in all-cause mortality associated with filter use, Coombs (2017) and Brunson (2017) found an increase (along with an increase in DVT-occurrence and VTE-recurrence) and the only RCT in this group (Barginear 2012) could not differentiate between filter and no filter for all of the outcomes reported.

Benefits and harms

The committee noted that IVC filters are used in a variety of circumstances in people with VTE, but there is a lack of consensus about their efficacy in most of these contexts. However, most of the evidence for the use of IVC filters in people with VTE that was identified in this review was of very low quality and the committee agreed that this limited their ability to make recommendations.

The committee discussed the potential harms associated with placing an IVC filter in a person with VTE. They agreed that there is an inherent harm associated with placing a filter due to the invasive nature of the procedure. In addition, the use of filters is also associated with a risk of filter complications, such as filter angulation, migration, infection, fracture and IVC perforation or thrombosis, and these potential harms must be considered when a decision is made to place a filter. The committee were aware of a review by Jia (2015) that reported high rates of IVC filter complications („15%), with major complications (typically requiring that the filter be retrieved) occurring in around 5% of people. However, the committee noted that many of the complications in this study are likely to be minor and there is uncertainty as to how important certain complications (such as IVC wall transgression by components of the filter) as clinically. The committee agreed that the evidence available does not adequately consider filter complications. The committee also noted that in some circumstances, filters may be placed prophylactically to reduce the perceived risk of PE-occurrence rather than based on clinical need. They agreed that there are financial costs to the health system and risks to the individual if filters are placed unnecessarily.

The committee were very concerned with the limited amount of long-term evidence available, leaving uncertainty surrounding treatment beyond 30 days, which is particularly important for certain groups of people (such as those people who had a recurrent event whilst taking anticoagulation) who typically receive long-term filters.

The committee discussed the benefits of using filters in the different scenarios. Firstly, the committee noted that the evidence for people at high risk of poor outcomes if they develop a PE was of very-low quality and that only two of the studies (Liang 2017 and Wadhwa 2018) attempted to adjust for confounders (and these were the only two that were not at critical risk of bias) produced conflicting results. Secondly, the committee agreed that as the evidence for people at high risk of a recurrent PE could not differentiate most outcomes at any point in time, they were unable to determine a benefit or harm associated with filter use in these people. Thirdly, they noted that the results from studies in people with VTE and cancer were contradictory with Stein (2018c) showing a reduction in all-cause mortality, while other studies (Coombs, 2017 and Brunson, 2017) showed an increase and the only RCT available for this group of people (Barginear, 2012) could not differentiate outcomes between people with a filter and those without. Finally, in people with VTE who are undergoing surgery, there was very-low quality evidence from a retrospective cohort study suggesting reduced all-mortality associated with the use of filters and low quality evidence from an RCT suggesting a reduction in PE-occurrence (due to a reduction in PEs occurring during surgery) in people with a filter, but this was in disagreement with another retrospective cohort study that could not differentiate between the filer and no filter groups for both outcomes. In addition, the committee were concerned with the heterogeneity of the studies contained within the grouping as each study involved a different type of surgery.

The committee agreed they were unable to recommend the use of filters for these groups due to the poor quality and contradictory or inconclusive nature of the evidence identified. However, they recognised there was a need for higher quality research to try to fill in the gaps in the evidence base and address the remaining uncertainty in these areas. They therefore made a recommendation to not offer filters to people with DVT or PE unless it is part of a clinical trial or was covered by their other recommendations for people in whom anticoagulation is contraindicated or who have a PE while taking anticoagulation treatment (see below for details of these recommendations). In addition, they made an accompanying research recommendation to try to determine the short and long term clinical and cost effectiveness of filters in people with VTE. The recommendation was for a large prospective study to follow-up people with VTE and capture information regarding IVC filter use. They envisaged that this cohort study would enrol everyone with VTE, with the intention of recording information regarding filter use and no filter use for each of the key population subgroups groups identified (see appendix Q for more details). They also included the option of an RCT to investigate all the population subgroups (with the exception of people who are at high risk of PE as this group was already covered by 2 RCTs). However, they noted that an RCT study design was likely to be less feasible than a prospective cohort study because it might be difficult to recruit sufficient people in the different subgroups to be able to detect a difference in outcomes between people who or do not have a filter.

The committee noted that people with VTE and a contraindication to anticoagulation are at a particularly high risk of VTE-recurrence. Therefore, IVC filters are typically seen as a viable and important alternative treatment (or secondary prophylactic measure) in these people. However, the committee agreed that in light of new evidence, which did not show a clear benefit to IVC filters in this group of people, and some evidence suggesting harm, the 2012 guideline recommendation - that IVC filters be offered to people with VTE with a contraindication to anticoagulants - was too strong. The committee noted that these recommendations were made by consensus and that the only evidence available at the time was from the PREPIC 2005 study (which contained a population of people receiving anticoagulation). Based on the limited evidence base and inconsistencies in the results between studies, the committee agreed that the recommendation should be downgraded to consider. They agreed that when filters are placed, they should be removed as soon as they are no longer needed (i.e. as soon as the individual is able to take anticoagulants and is on stable treatment with them).

The committee discussed the use of IVC filters in people who have a PE whilst taking therapeutic anticoagulation treatment. The committee noted a reduction in short term all-cause mortality (30 days) associated with the use of filters in the Mellado (2016) study. The committee agreed that the very serious risk of bias associated with Stein (2019a) limited their confidence in the findings of this study but noted that the trend was consistent with that of Mellado (2016). The committee recommended that IVC filters be considered in people with VTE who have a PE whilst taking therapeutic treatment, but only after various clinical checks are performed. They agreed that adherence to anticoagulation treatment should be checked, as a recurrent event associated with poor adherence may be more suitably treated by increasing the awareness of the person with VTE of the importance of taking the anticoagulants at the correct time and in the correct manner. If adherence to treatment is not an issue, then an increased dose of anticoagulant or alternative treatment regimen should be investigated as other anticoagulants may prove more effective for that individual. Finally, hypercoagulability should be assessed. The committee made a weaker ‘consider’ recommendation due to the limited and low quality of the evidence base.

The committee noted that the type of filter is the same, regardless of whether it is intended to be inserted temporarily or permanently. They agreed that the decision about whether a filter is planned to be temporary or permanent should be made on an individual basis but expected that in most cases the filter would be temporary. The committee therefore agreed that it is important that there is a strategy in place for removing a temporary IVC filter as soon as this is clinically advisable and that this plan is clearly documented. The committee agreed that they could not specify how frequently the strategy should be reviewed but agreed that a review should take place if the individual’s clinical situation changes. The committee made a recommendation to reflect these points to ensure that the filter is removed as soon as it is no longer needed.

Cost effectiveness and resource use

The committee discussed evidence from one published cost-utility study (Sarasin 1993) that compared the use of IVC filters to no treatment in people with cancer and confirmed VTE. The authors of the study concluded that using an IVC filter was more effective and less costly than no treatment. The study was classified as partially applicable with very serious limitations. The committee decided to include the study in the evidence review for transparency but felt it had limited value in informing the recommendations for several reasons. Firstly, the analysis was not conducted from a UK perspective. Secondly, the analysis was published in 1993 and therefore the evidence informing the effectiveness of the IVC filter strategy in the model did not reflect any of the more recent studies identified in the clinical evidence review and was unlikely to reflect the efficacy of IVC filters in current use. Finally, there were a number of methodological weaknesses in the modelling approach, including a lack clarity about the source of some model parameters, utility values that were elicited from healthcare professionals and an absence of probabilistic sensitivity analysis.

No published economic evidence on the cost effectiveness of IVC filters was identified for any of the other subgroups of interest and de novo modelling was not conducted for this review question. The committee reflected on the costs of placement and removal of IVC filters (£3,500 and £930 respectively based on 2017/2018 NHS Reference Costs YR22A – YR22C). By making more specific recommendations about the clinical situations in which to consider the use of IVC filters, the committee felt that this could lead to a reduction in the use of IVC filters in patients for whom there is no clear evidence of benefit and could potentially result in cost savings.

Appendices

Appendix A. Review protocol

Field (based on PRISMA-PContent
Review questionWhat is the effectiveness of inferior vena caval filters to prevent PE in people with confirmed VTE?
Type of review questionIntervention
Objective of the reviewTo determine the effectiveness of filters in people with VTE who cannot take anticoagulants or those people with VTE who might need filters in addition to anticoagulants.
Eligibility criteria – population/disease

Adults (18+ years) with:

  • VTE (DVT and/or PE) who cannot have anticoagulants or
  • VTE who have had a filter inserted because they could not take anticoagulants, but retain the filter once they start taking anticoagulants or
  • VTE who have a PE whilst taking anticoagulants or
  • VTE who have the filters inserted for prophylaxis before a potential provoking event (e.g. surgery) or
  • VTE who can receive anticoagulants but are at high risk of PE* or are at high risk of poor outcomes if they had further PEs**
*High risk was defined by the committee as people with free floating DVTs.

**This includes massive /sub-massive PE patients and those with severe pre-existing cardio-pulmonary disease.

Eligibility criteria – intervention(s)Vena cava filters with or without:
  • anticoagulation treatment and/or
  • mechanical interventions
Eligibility criteria – comparator(s)No filter with:
  • mechanical intervention and/or
  • anticoagulant treatment and/or
  • placebo or no treatment.
We will include studies that allow participants to have mechanical interventions, anticoagulation treatment or both, but these must be included in both arms of the trial so that the only difference in treatment between arms is the inclusion or exclusion of IVC filters.
Outcomes and prioritisation
  • Recurrent VTE (PE and DVT)
  • All-cause mortality
  • VTE-related mortality
  • Post-thrombotic syndrome
  • Pulmonary hypertension (PH)
  • Quality of life
    • Generic and disease-specific measures will be reported
    • Overall score will be reported (data on subscales will not be reported)
  • Adverse events
    • Total serious adverse events (as defined by the European medicines agency) will be reported if data is available.
    • Major bleeding (as defined by International Society on Thrombosis and Haemostasis)
    • Clinically relevant non-major bleeding (as defined by International Society on Thrombosis and Haemostasis)
    • Surgical complications at the time of placement and removal
    • Sepsis (or serious infections) for filters that are in place for longer periods
  • Resource use and costs
Eligibility criteria – study design
  • RCTs
  • Cohort studies (prospective or retrospective)
Other inclusion/exclusion criteriaInclusion:
  • English language papers only
Exclusion:
  • Filters in studies from before 1990 that are no longer used in clinical practice
Proposed sensitivity/sub-group analysisPopulations:
  • People with cancer
  • Very elderly people (defined as people over the age of 80)
  • Intravenous drug users
  • People with chronic liver disease
  • People with obesity (a BMI of over 30 kg/m2 or more).
Other factors:
  • Index event (DVT-only or PE with or without DVT)
  • Study type (RCT, prospective and retrospective cohort)
  • Filter type (retrievable or permanent)
  • Intervention type
Selection process – duplicate screening/selectio n/analysis

10% of the abstracts were reviewed by two reviewers, with any disagreements resolved by discussion or, if necessary, a third independent reviewer. If meaningful disagreements were found between the different reviewers, a further 10% of the abstracts were reviewed by two reviewers, with this process continued until agreement is achieved between the two reviewers. From this point, the remaining abstracts will be screened by a single reviewer.

This review made use of the priority screening functionality with the EPPI-reviewer systematic reviewing software. See Appendix B for more details.

Data management (software)See Appendix B
Information sources – databases and dates
  • Sources to be searched
    • Clinical searches - Medline, Medline in Process, PubMed, Embase, Cochrane CDSR, CENTRAL, DARE (legacy records) and HTA.
    • Economic searches - Medline, Medline in Process, PubMed, Embase, NHS EED (legacy records) and HTA, with economic evaluations and quality of life filters applied.
  • Supplementary search techniques
    • None identified
  • Limits
    • Studies reported in English
    • Study design RCT, SR and Observational filter will be applied (as agreed)
    • Animal studies will be excluded from the search results
    • Conference abstracts will be excluded from the search results
    • Date limit from August 2011 for search for RCTs, but no date limits for cohort studies search.
Identify if an update

This question is an update of a question in CG144. Original search date up to 01.08.2011 but only included RCTs.

Recommendations that may change as a result of this review:

1.2.10.

Offer temporary inferior vena caval filters to patients with proximal DVT or PE who cannot have anticoagulation treatment, and remove the inferior vena caval filter when the patient becomes eligible for anticoagulation treatment. [2012]

1.2.11.

Consider inferior vena caval filters for patients with recurrent proximal DVT or PE despite adequate anticoagulation treatment only after considering alternative treatments such as:

  • increasing target INR to 3–4 for long-term high-intensity oral anticoagulant therapy or
  • switching treatment to LMWH. [2012]

1.2.12.

Ensure that a strategy for removing the inferior vena caval filter at the earliest possible opportunity is planned and documented when the filter is placed, and that the strategy is reviewed regularly. [2012]

Author contacts https://www​.nice.org​.uk/guidance/indevelopment/gid-ng10087
Highlight if amendment to previous protocolFor details please see section 4.5 of Developing NICE guidelines: the manual
Search strategy – for one databaseFor details please see appendix C of the evidence review
Data collection process – formsA standardised evidence table format will be used, and published as appendix E (clinical evidence tables) or I (economic evidence tables) of the evidence review.
Data items – define all variables to be collectedFor details please see evidence tables in appendix E (clinical evidence tables) or I (economic evidence tables) of the evidence review.
Methods for assessing bias at outcome/study levelSee appendix B
Criteria for quantitative synthesisSee appendix B
Methods for quantitative analysis – combining studies and exploring (in)consistencySee appendix B
Meta-bias assessment – publication bias, selective reporting biasSee appendix B
Confidence in cumulative evidenceSee appendix B
Rationale/context – what is known

In CG144, temporary inferior vena caval filters were recommended for people with proximal DVT or PE who cannot have anticoagulation treatment, with the filter being removed when the person becomes eligible for anticoagulation treatment (2012 recommendations). The guideline committee raised concerns that this recommendation was no longer appropriate given the results from new clinical trials and that additional guidance is needed concerning how long a person should be unable to take anticoagulant treatment before an inferior vena caval filter (IVC) is fitted.

For more detail please see the introduction to the evidence review.

Describe contributions of authors and guarantor

A multidisciplinary committee developed the guideline. The committee was convened by the NICE Guidelines Updates Team and chaired by Susan Bewley in line with section 3 of Developing NICE guidelines: the manual.

Staff from the NICE Guidelines Updates Team undertook systematic literature searches, appraised the evidence, conducted meta-analysis and cost-effectiveness analysis where appropriate, and drafted the guideline in collaboration with the committee. For details please see the methods section of the evidence review in appendix B.

Sources of funding/supportThe NICE Guideline Updates Team is an internal team within NICE
Name of sponsorThe NICE Guideline Updates Team is an internal team within NICE
Roles of sponsorThe NICE Guideline Updates Team is an internal team within NICE
PROSPERO registration number[If registered, add PROSPERO registration number]

Appendix B. Methods

Priority screening

The reviews undertaken for this guideline all made use of the priority screening functionality with the EPPI-reviewer systematic reviewing software. This uses a machine learning algorithm (specifically, an SGD classifier) to take information on features (1, 2 and 3 word blocks) in the titles and abstract of papers marked as being ‘includes’ or ‘excludes’ during the title and abstract screening process, and re-orders the remaining records from most likely to least likely to be an include, based on that algorithm. This re-ordering of the remaining records occurs every time 25 additional records have been screened.

Research is currently ongoing as to what are the appropriate thresholds where reviewing of abstract can be stopped, assuming a defined threshold for the proportion of relevant papers it is acceptable to miss on primary screening. As a conservative approach until that research has been completed, the following rules were adopted during the production of this guideline:

  • In every review, at least 50% of the identified abstract (or 1,000 records, if that is a greater number) were always screened.
  • After this point, screening was only terminated if a pre-specified threshold was met for a number of abstracts being screened without a single new include being identified. This threshold was set according to the expected proportion of includes in the review (with reviews with a lower proportion of includes needing a higher number of papers without an identified study to justify termination), and was always a minimum of 250.
  • A random 10% sample of the studies remaining in the database when the threshold were additionally screened, to check if a substantial number of relevant studies were not being correctly classified by the algorithm, with the full database being screened if concerns were identified.

As an additional check to ensure this approach did not miss relevant studies, the included studies lists of included systematic reviews were searched to identify any papers not identified through the primary search.

Incorporating published systematic reviews

For all review questions where a literature search was undertaken looking for a particular study design, systematic reviews containing studies of that design were also included. All included studies from those systematic reviews were screened to identify any additional relevant primary studies not found as part of the initial search.

Quality assessment

Individual systematic reviews were quality assessed using the ROBIS tool, with each classified into one of the following three groups:

  • High quality – It is unlikely that additional relevant and important data would be identified from primary studies compared to that reported in the review, and unlikely that any relevant and important studies have been missed by the review.
  • Moderate quality – It is possible that additional relevant and important data would be identified from primary studies compared to that reported in the review, but unlikely that any relevant and important studies have been missed by the review.
  • Low quality – It is possible that relevant and important studies have been missed by the review.

Each individual systematic review was also classified into one of three groups for its applicability as a source of data, based on how closely the review matches the specified review protocol in the guideline. Studies were rated as follows:

  • Fully applicable – The identified review fully covers the review protocol in the guideline.
  • Partially applicable – The identified review fully covers a discrete subsection of the review protocol in the guideline (for example, some of the factors in the protocol only).
  • Not applicable – The identified review, despite including studies relevant to the review question, does not fully cover any discrete subsection of the review protocol in the guideline.

Using systematic reviews as a source of data

If systematic reviews were identified as being sufficiently applicable and high quality, and were identified sufficiently early in the review process (for example, from the surveillance review or early in the database search), they were used as the primary source of data, rather than extracting information from primary studies. The extent to which this was done depended on the quality and applicability of the review, as defined in Table 8. When systematic reviews were used as a source of primary data, and unpublished or additional data included in the review which is not in the primary studies was also included. Data from these systematic reviews was then quality assessed and presented in GRADE tables as described below, in the same way as if data had been extracted from primary studies. In questions where data was extracted from both systematic reviews and primary studies, these were cross-referenced to ensure none of the data had been double counted through this process.

Table 8. Criteria for using systematic reviews as a source of data

Evidence synthesis and meta-analyses

Where possible, meta-analyses were conducted to combine the results of quantitative studies for each outcome. For continuous outcomes analysed as mean differences, where change from baseline data were reported in the trials and were accompanied by a measure of spread (for example standard deviation), these were extracted and used in the meta-analysis. Where measures of spread for change from baseline values were not reported, the corresponding values at study end were used and were combined with change from baseline values to produce summary estimates of effect. These studies were assessed to ensure that baseline values were balanced across the treatment groups; if there were significant differences at baseline these studies were not included in any meta-analysis and were reported separately. For continuous outcomes analysed as standardised mean differences, where only baseline and final time point values were available, change from baseline standard deviations were estimated, assuming a correlation coefficient of 0.5.

Evidence of effectiveness of interventions

Quality assessment

Individual RCTs and quasi-randomised controlled trials were quality assessed using the Cochrane Risk of Bias Tool. Other studies were quality assessed using the ROBINS-I tool. Each individual RCT was classified into one of the following three groups:

  • Low risk of bias – The true effect size for the study is likely to be close to the estimated effect size.
  • Moderate risk of bias – There is a possibility the true effect size for the study is substantially different to the estimated effect size.
  • High risk of bias – It is likely the true effect size for the study is substantially different to the estimated effect size.

Each individual cohort study was classified into one of the following five groups:

  • Low risk of bias – The true effect size for the study is likely to be close to the estimated effect size.
  • Moderate risk of bias – There is a possibility the true effect size for the study is substantially different to the estimated effect size.
  • Serious risk of bias – It is likely the true effect size for the study is substantially different to the estimated effect size.
  • Critical risk of bias – It is likely the true effect size for the study is substantially different to the estimated effect size.
  • No information – There is no clear indication that the study is at serious or critical risk of bias and there is a lack of information in one or more key domains of bias.

Each individual study (RCT and cohort study) was also classified into one of three groups for directness, based on if there were concerns about the population, intervention, comparator and/or outcomes in the study and how directly these variables could address the specified review question. Studies were rated as follows:

  • Direct – No important deviations from the protocol in population, intervention, comparator and/or outcomes.
  • Partially indirect – Important deviations from the protocol in one of the population, intervention, comparator and/or outcomes.
  • Indirect – Important deviations from the protocol in at least two of the following areas: population, intervention, comparator and/or outcomes.

Methods for combining intervention evidence

Meta-analyses of interventional data were conducted with reference to the Cochrane Handbook for Systematic Reviews of Interventions (Higgins et al. 2011).

Where different studies presented continuous data measuring the same outcome but using different numerical scales (e.g. a 0-10 and a 0-100 visual analogue scale), these outcomes were all converted to the same scale before meta-analysis was conducted on the mean differences. Where outcomes measured the same underlying construct but used different instruments/metrics, data were analysed using standardised mean differences (Hedges’ g).

A pooled relative risk was calculated for dichotomous outcomes (using the Mantel–Haenszel method) reporting numbers of people having an event, and a pooled incidence rate ratio was calculated for dichotomous outcomes reporting total numbers of events. Both relative and absolute risks were presented, with absolute risks calculated by applying the relative risk to the pooled risk in the comparator arm of the meta-analysis (all pooled trials).

Fixed- and random-effects models (der Simonian and Laird) were fitted for all syntheses, with the presented analysis dependent on the degree of heterogeneity in the assembled evidence. Fixed-effects models were the preferred choice to report, but in situations where the assumption of a shared mean for fixed-effects model were clearly not met, even after appropriate pre-specified subgroup analyses were conducted, random-effects results are presented. Fixed-effects models were deemed to be inappropriate if one or both of the following conditions was met:

  • Significant between study heterogeneity in methodology, population, intervention or comparator was identified by the reviewer in advance of data analysis. This decision was made and recorded before any data analysis was undertaken.
  • The presence of significant statistical heterogeneity in the meta-analysis, defined as I2≥50%.

In any meta-analyses where some (but not all) of the data came from studies at high risk of bias, a sensitivity analysis was conducted, excluding those studies from the analysis. Results from both the full and restricted meta-analyses are reported. Similarly, in any meta-analyses where some (but not all) of the data came from indirect studies, a sensitivity analysis was conducted, excluding those studies from the analysis.

Meta-analyses were performed in Cochrane Review Manager V5.3, with the exception of incidence rate ratio analyses which were carried out in R version 3.3.4.

Minimal clinically important differences (MIDs)

The Core Outcome Measures in Effectiveness Trials (COMET) database was searched to identify published minimal clinically important difference thresholds relevant to this guideline. MIDs were assessed to ensure they had been developed and validated in a methodologically rigorous way, and were applicable to the populations, interventions and outcomes specified in this guideline. No MIDs were identified through this process. In addition, the Guideline Committee were asked to prospectively specify any outcomes where they felt a consensus MID could be defined from their experience. The committee agreed that any difference in mortality would be clinically meaningful, and therefore the line of no effect was used as an MID. The committee chose not to specify any other MIDs by consensus.

For continuous outcomes expressed as a mean difference where no other MID was available, an MID of 0.5 of the median standard deviations of the comparison group arms was used (Norman et al. 2003). For continuous outcomes expressed as a standardised mean difference where no other MID was available, an MID of 0.5 was used. For relative risks where no other MID was available, a default MID interval for dichotomous outcomes of 0.8 to 1.25 was used.

The ‘Evidence to Recommendations’ section of each review makes explicit the committee’s view of the expected clinical importance and relevance of the findings. In particular, this includes consideration of whether the whole effect of a treatment (which may be felt across multiple independent outcome domains) would be likely to be clinically meaningful, rather than simply whether each individual sub outcome might be meaningful in isolation.

GRADE for pairwise meta-analyses of interventional evidence

GRADE was used to assess the quality of evidence for the selected outcomes as specified in ‘Developing NICE guidelines: the manual (2014)’. Data from RCTs were initially rated as high quality, data from observational studies were originally rated as low quality. The quality of the evidence for each outcome was downgraded or not from this initial point, based on the criteria given in Table 9.

Table 9. Rationale for downgrading quality of evidence for intervention studies

Publication bias

Where 10 or more studies were included as part of a single meta-analysis, a funnel plot was produced to graphically assess the potential for publication bias.

Evidence statements

Evidence statements for pairwise intervention data are classified in to one of four categories:

  • Situations where the data are only consistent, at a 95% confidence level, with an effect in one direction (i.e. one that is ‘statistically significant’), and the magnitude of that effect is most likely to meet or exceed the MID (i.e. the point estimate is not in the zone of equivalence). In such cases, we state that the evidence showed that there is an effect.
  • Situations where the data are only consistent, at a 95% confidence level, with an effect in one direction (i.e. one that is ‘statistically significant’), but the magnitude of that effect is most likely to be less than the MID (i.e. the point estimate is in the zone of equivalence). In such cases, we state that the evidence showed there is an effect, but it is less than the defined MID.
  • Situations where the confidence limits are smaller than the MIDs in both directions. In such cases, we state that the evidence demonstrates that there is no meaningful difference.
  • In all other cases, we state that the evidence could not differentiate between the comparators.

For outcomes without a defined MID or where the MID is set as the line of no effect (for example, in the case of mortality), evidence statements are divided into 2 groups as follows:

  • We state that the evidence showed that there is an effect if the 95% CI does not cross the line of no effect.
  • The evidence could not differentiate between comparators if the 95% CI crosses the line of no effect.

Appendix C. Literature search strategies

A systematic search was conducted on 4th July 2019. The following databases were searched; Medline, Medline in Process, Medline epub ahead of Print, Embase (all via the Ovid platform), Cochrane Database of Systematic Reviews and the Cochrane Register of Controlled Trials (via the Wiley platform) and the Database of Abstracts of Reviews (via the Centre for Reviews and Dissemination platform). Date limits were applied to the date of the previous guideline for RCT and systematic review evidence. No date limits were applied for observational studies evidence. McMaster balanced RCT health-evidence.ca Systematic Review and NICE in house observational studies filters were used.

The Medline strategy is presented below. This was translated for other databases.

  1. Venous Thrombosis/
  2. (phlegmasia adj2 dolens).tw.
  3. (thrombo* adj2 (vein* or venous)).tw.
  4. (venous adj stasis).tw.
  5. (dvt or vte).tw.
  6. Venous Thromboembolism/ or Embolism, paradoxical/
  7. exp pulmonary embolism/
  8. ((pulmonary or lung) adj4 (embol* or thromboembo* or microembol*)).tw.
  9. (pulmonary adj infarction).tw.
  10. or/1-9
  11. Vena Cava Filters/
  12. vena cava, inferior/su or venae cavae/su
  13. ((((ivc or vena) adj (cava or caval)) or umbrella) adj2 (filter or filters)).tw.
  14. (ALN or Amplatz or Antheor or “Bird’s Nest” or Celect or Crux or Denali or G2 or Greenfield or “Gunther Tulip” or LGM or “Mobin-Uddin” or Ninitol or OptEase or Prolyser or Tempofilter or TrapEase or “Vena Tech” or Venatech).tw.
  15. or/11-14
  16. 10 and 15
  17. randomized controlled trial.pt.
  18. randomi?ed.mp.
  19. placebo.mp.
  20. or/17-19
  21. (MEDLINE or pubmed).tw.
  22. systematic review.tw.
  23. systematic review.pt.
  24. meta-analysis.pt.
  25. intervention$.ti.
  26. or/21-25
  27. 20 or 26
  28. 16 and 27
  29. limit 28 to ed=20110801-20190704
  30. Observational Studies as Topic/
  31. Observational Study/
  32. Epidemiologic Studies/
  33. exp Case-Control Studies/
  34. exp Cohort Studies/
  35. Cross-Sectional Studies/
  36. Controlled Before-After Studies/
  37. Historically Controlled Study/
  38. Interrupted Time Series Analysis/
  39. Comparative Study.pt.
  40. case control$.tw.
  41. case series.tw.
  42. (cohort adj (study or studies)).tw.
  43. cohort analy$.tw.
  44. (follow up adj (study or studies)).tw.
  45. (observational adj (study or studies)).tw.
  46. longitudinal.tw.
  47. prospective.tw.
  48. retrospective.tw.
  49. cross sectional.tw.
  50. or/30-49
  51. 16 and 50
  52. 29 or 51
  53. limit 52 to english language
  54. animals/ not humans/
  55. 53 not 54

Searches to identify economic evidence were run on 9th July 2019 in Medline, Medline in Process, Econlit and Embase (all va the Ovid platform), NHS EED and the Health Technology Database (via the Centre for Reviews and Dissemination platform). NICE inhouse economic evaluation and Quality of Life filters were attached to lines 1 to 29 of the core strategy in the Medline and Embase databases. The Medline version of the filters is displayed below.

Economic evaluations

  1. Economics/
  2. exp “Costs and Cost Analysis”/
  3. Economics, Dental/
  4. exp Economics, Hospital/
  5. exp Economics, Medical/
  6. Economics, Nursing/
  7. Economics, Pharmaceutical/
  8. Budgets/
  9. exp Models, Economic/
  10. Markov Chains/
  11. Monte Carlo Method/
  12. Decision Trees/
  13. econom$.tw.
  14. cba.tw.
  15. cea.tw.
  16. cua.tw.
  17. markov$.tw.
  18. (monte adj carlo).tw.
  19. (decision adj3 (tree$ or analys$)).tw.
  20. (cost or costs or costing$ or costly or costed).tw.
  21. (price$ or pricing$).tw.
  22. budget$.tw.
  23. expenditure$.tw.
  24. (value adj3 (money or monetary)).tw.
  25. (pharmacoeconomic$ or (pharmaco adj economic$)).tw.
  26. 26 or/1-25

Quality of Life

  1. "Quality of Life"/
  2. quality of life.tw.
  3. "Value of Life"/
  4. Quality-Adjusted Life Years/
  5. quality adjusted life.tw.
  6. (qaly$ or qald$ or qale$ or qtime$).tw.
  7. disability adjusted life.tw.
  8. daly$.tw.
  9. Health Status Indicators/ (22343)
  10. (sf36 or sf 36 or short form 36 or shortform 36 or sf thirtysix or sf thirty six or shortform thirtysix or shortform thirty six or short form thirtysix or short form thirty six).tw.
  11. (sf6 or sf 6 or short form 6 or shortform 6 or sf six or sfsix or shortform six or short form six).tw.
  12. (sf12 or sf 12 or short form 12 or shortform 12 or sf twelve or sftwelve or shortform twelve or short form twelve).tw.
  13. (sf16 or sf 16 or short form 16 or shortform 16 or sf sixteen or sfsixteen or shortform sixteen or short form sixteen).tw.
  14. (sf20 or sf 20 or short form 20 or shortform 20 or sf twenty or sftwenty or shortform twenty or short form twenty).tw.
  15. (euroqol or euro qol or eq5d or eq 5d).tw.
  16. (qol or hql or hqol or hrqol).tw.
  17. (hye or hyes).tw.
  18. health$ year$ equivalent$.tw.
  19. utilit$.tw.
  20. (hui or hui1 or hui2 or hui3).tw.
  21. disutili$.tw.
  22. rosser.tw.
  23. quality of wellbeing.tw.
  24. quality of well-being.tw.
  25. qwb.tw.
  26. willingness to pay.tw.
  27. standard gamble$.tw.
  28. time trade off.tw.
  29. time tradeoff.tw.
  30. tto.tw.
    or/1-30

Appendix D. Clinical evidence study selection

Image niceng158er8appdf1

Appendix E. Clinical evidence tables

Download PDF (724K)

Appendix F. Forest plots

Filter versus no filter in people who cannot have anticoagulants

Figure 1. All-cause mortality (30 days)

Filter versus no filter in people with VTE who have the filters inserted for prophylaxis before a potential provoking event

Figure 2. All-cause mortality (in-hospital)

Figure 3. PE-related mortality (in-hospital)

Appendix G. GRADE profiles

Filter versus no filter in people with VTE who cannot have anticoagulants

Quality assessmentNo. patientsEffectQuality
No. of studiesDesignRisk of biasInconsistencyIndirectnessimprecisionFilterNo filterRelative (95% CI)Absolute risk controlAbsolute risk intervention
All-cause mortality (30 days): HR “1 favours filter (Figure 1)
2Retrospective cohort studyVery serious2Very seriousNot seriousSerious3N/AN/A

HR 0.91

(0.53, 1.56)

N/AN/AVery low
All-cause mortality (3 months): HR “1 favours filter
1 White 2016Retrospective cohort studyVery serious1N/ANot seriousNot seriousN/AN/A

HR 0.73

(0.59, 0.90)

N/AN/AVery low
PE-occurrence (1 year): HR “1 favours filter
1 White 2016Retrospective cohort studyVery serious1N/ANot seriousSerious3N/AN/A

HR 1.04

(0.67, 1.61)

N/AN/AVery low
DVT-occurrence (1 year): HR “1 favours filter
1 White 2016Retrospective cohort studyVery serious1N/ANot seriousNot seriousN/AN/A

HR 2.35

(1.56, 3.53)

N/AN/AVery low
1

Study was at serious risk of bias for this population (people with contraindication for anticoagulants).

2

„33.3% of studies were at serious risk of bias.

3

95% CI crosses line of no effect.

Filter versus no filter in people with VTE who have a PE whilst taking anticoagulants

Quality assessmentNo. patientsEffectQuality
No. of studiesDesignRisk of biasInconsistencyIndirectnessimprecisionFilterNo filterRelative (95% CI)Absolute risk controlAbsolute risk intervention
All-cause mortality (in-hospital): RR “1 favours filter
1 Stein 2019aRetrospective cohort studyVery serious4N/ANot seriousNot serious18/60383/211

RR 0.08

(0.05, 0.12)

39.34 per 100

2.99 per 100

(1.84, 4.85)

Very low
All-cause mortality (30 days): RR “1 favours filter
1 Mellado 2016Retrospective analysis of prospective cohortSerious4N/ANot seriousNot serious1/4823/91

RR 0.08

(0.01, 0.59)

25.27 per 100

2.08 per 100

(0.29, 14.96)

Very low
All-cause mortality (3 months): RR “1 favours filter
1 Stein 2019aRetrospective cohort studyVery serious4N/ANot seriousNot serious18/60383/211

RR 0.08

(0.05, 0.12)

39.34 per 100

2.99 per 100

(1.84, 4.85)

Very low
VTE:-recurrence (30 days): RR “1 favours filter
1 Mellado 2016Retrospective analysis of prospective cohortSerious3N/ANot seriousVery serious12/482/91

RR 1.90

(0.28, 13.04)

2.20 per 100

4.17 per 100

(0.61, 28.66)

Very low
PE-related mortality (30 days): RR “1 favours filter
1 Mellado 2016Retrospective analysis of prospective cohortVery serious3N/ANot seriousSerious51/4816/91

RR 0.12

(0.02, 0.87)

17.58 per 100

2.08 per 100

(0.28, 15.24)

Very low
Major bleeding (30 days): RR “1 favours filter
1 Mellado 2016Retrospective analysis of prospective cohortSerious3N/ANot seriousVery serious12/483/91

RR 1.26

(0.22, 7.31)

3.30 per 100

4.17 per 100

(0.72, 24.09)

Very low
1

95% confidence interval crosses both ends of a defined MID interval.

2

I2 „66.6%.

3

The study was at Moderate risk of bias.

4

The study was at critical risk of bias.

5

95% confidence interval crosses one end of a defined MID interval.

Filter versus no filter in people with VTE who have the filters inserted for prophylaxis before a potential provoking event

Quality assessmentNo. patientsEffectQuality
No. of studiesDesignRisk of biasInconsistencyIndirectnessimprecisionFilterNo filterRelative (95% CI)Absolute risk controlAbsolute risk intervention
All-cause mortality (peri-procedure): RR “1 favours filter
1 Sharifi 2012RCTSerious1N/ANot seriousNot estimable70/700/71Not estimable7Not estimable7Not estimable7Moderate
All-cause mortality (in-hospital): RR “1 favours filter (Figure 2)
1 Sharifi 2012RCTSerious1N/ANot seriousNot estimable70/700/71Not estimable7Not estimable7Not estimable7Moderate
All-cause mortality (in-hospital): RR “1 favours filter (Figure 2)
1 Stein 2018aRetrospective cohort studyVery serious2N/ANot seriousNot serious20/24549/124

RR 0.21

(0.13, 0.33)

39.52 per 100

8.16 per 100

(5.09, 13.10)

Very low
All-cause mortality (in-hospital, „80 year olds only): RR “1 favours filter
1 Stein 2018aRetrospective cohort studyVery serious2N/ANot seriousSerious62/117/11

RR 0.29

(0.08, 1.08)

63.64 per 100

18.18 per 100

(4.80, 68.80)

Very low
All-cause mortality (30 days): HR “1 favours filter
1 White 2016Retrospective cohort studySerious1N/ANot seriousSerious6N/AN/A

HR 1.12

(0.71, 1.77)

N/AN/AVery low
All-cause mortality (3 months): RR “1 favours filter
1 Stein 2018aRetrospective cohort studyVery serious2N/ANot seriousNot serious21/24549/124

RR 0.22

(0.14, 0.34)

39.52 per 100

8.57 per 100

(5.39, 13.62)

Very low
All-cause mortality (3 months): HR “1 favours filter
1 White 2016Retrospective cohort studiesSerious1N/ANot seriousSerious6N/AN/A

HR 1.10

(0.76, 1.60)

N/AN/AVery low
All-cause mortality (2 years): RR “1 favours filter
1 Sharifi 2012RCTSerious1N/ANot seriousNot serious2/702/71

RR 1.01

(0.15, 7.00)

2.82 per 100

2.85 per 100

(0.42, 19.72)

Moderate
PE-related mortality (in-hospital): RR “1 favours filter (Figure 3)
2Retrospective cohort studyVery serious3Not seriousNot seriousNot serious2/103443/753

RR 0.03

(0.01, 0.11)

5.71 per 100

0.18 per 100

(0.05, 0.62)

Very low
PE-related mortality (3-months): RR “1 favours filter
1 Stein 2018aRetrospective cohort studyVery serious2N/ANot seriousNot serious3/21138/105

RR 0.04

(0.01, 0.12)

36.19 per 100

1.42 per 100

(0.45, 4.50)

Very low
DVT-occurrence (peri-procedure): RR “1 favours filter
1 Sharifi 2012RCTSerious1N/ANot seriousNot estimable70/700/71Not estimable7Not estimable7Not estimable7Moderate
DVT-occurrence (in-hospital): RR “1 favours filter
1 Sharifi 2012RCTSerious1N/ANot seriousNot estimable70/700/71Not estimable7Not estimable7Not estimable7Moderate
DVT-occurrence (1 year): HR “1 favours filter
1 White 2016Retrospective cohort studiesSerious1N/ANot seriousSerious6N/AN/A

HR 1.15

(0.57, 2.32)

N/AN/AVery low
DVT-occurrence (2 years): RR “1 favours filter
1 Sharifi 2012RCTSerious1N/ANot seriousVery serious42/702/71

RR 1.01

(0.15, 7.00)

2.82 per 100

2.86 per 100

(0.41, 19.72)

Very low
PE-occurrence (peri-procedure): RR “1 favours filter
1 Sharifi 2012RCTSerious1N/ANot seriousSerious51/708/71

RR 0.13

(0.02, 0.99)

11.27 per 100

1.43 per 100

(0.18, 11.12)

Low
PE-occurrence (in-hospital): RR “1 favours filter
1 Sharifi 2012RCTSerious1N/ANot seriousSerious51/708/71

RR 0.13

(0.02, 0.99)

11.27 per 100

1.43 per 100

(0.18, 11.12)

Low
PE-occurrence (in-hospital): RR “1 favours filter
1 Pan 2014Retrospective cohort studyVery serious2N/ANot seriousNot serious1/82311/648

RR 0.09

(0.01, 0.51)

1.70 per 100

0.15 per 100

(0.02, 0.87)

Very low
PE-occurrence (in-hospital, only those with contraindication to AC during stay): RR “1 favours filter
1 Pan 2014Retrospective cohort studyVery serious2N/ANot seriousSerious50/1321/26

RR 0.07

(0.00, 1.62)

3.85 per 100

0.26 per 100

(0.01, 6.22)

Very low
PE-occurrence (in-hospital, only those with AC following surgery): RR “1 favours filter
1 Pan 2014Retrospective cohort studyVery serious2N/ANot seriousNot serious1/69110/622

RR 0.09

(0.01, 0.70)

1.61 per 100

0.14 per 100

(0.02, 1.13)

Very low
PE-occurrence (1 year): HR “1 favours filter
1 White 2016Retrospective cohort studySerious1N/ANot seriousSerious6N/AN/A

HR 0.85

(0.35, 2.08)

N/AN/AVery low
PE-occurrence (2 years): RR “1 favours filter
1 Sharifi 2012RCTSerious1N/ANot seriousSerious51/708/71

RR 0.13

(0.02, 0.99)

11.27 per 100

1.43 per 100

(0.18, 11.12)

Low
1

Study was at moderate risk of bias.

2

Study was at critical risk of bias.

3

Both studies were at critical risk of bias.

4

95% confidence interval crosses both ends of a defined MID interval.

5

95% confidence interval crosses one end of a defined MID interval

6

95% CI crosses line of no effect.

7

Effect estimate not calculable as both arms have 0 events.

Filter versus no filter in people with VTE who are at high risk of poor outcomes in the event of PE-occurrence

The characteristics predisposing the populations to poor outcomes varied considerably between studies (see the corresponding forest plots and evidence tables for further detail).

Quality assessmentNo. patientsEffectQuality
No. of studiesDesignRisk of biasInconsistencyIndirectnessimprecisionFilterNo filterRelative (95% CI)Absolute risk controlAbsolute risk intervention
All-cause mortality (in-hospital): RR “1 favours filter (Figure 4)
6*Retrospective cohort studyVery serious4Very serious2Not seriousNot serious6986/7127345997/375489

RR 0.54

(0.39, 0.73)

12.25 per 100

6.59 per 100

(4.83, 8.98)

Very low
All cause mortality (in-hospital): HR “1 favours filter
1 Liang 2017Retrospective cohort studySerious6N/ANot seriousNot seriousN/AN/A

HR 1.24

(1.11, 1.38)

N/AN/AVery low
All-cause mortality (in-hospital, sensitivity analysis excluding studies at critical risk of bias): RR “1 favours filter
1 Wadhwa 2018Retrospective cohort studyVery serious5N/ANot seriousNot serious6541/6723743796/358638

RR 0.80

(0.78, 0.82)

9.73 per 100

7.78 per 100

(7.59, 7.98)

Very low
All-cause mortality (in-hospital, only in people aged 80 years or older): RR “1 favours filter (Figure 5)
3**Retrospective cohort studyVery serious4Not seriousNot seriousNot serious89/1145597/5012

RR 0.50

(0.41, 0.61)

11.91 per 100

5.96 per 100

(4.88, 7.27)

Very low
All-cause mortality (in-hospital, only in people with massive-PE): RR “1 favours filter
1 Wadhwa 2018Retrospective cohort studyVery serious5N/ANot seriousNot serious3815/1541121553/52708

RR 0.61

(0.59, 0.62)

40.89 per 100

24.76 per 100

(24.04, 25.49)

Very low
All-cause mortality (30 days): RR “1 favours filter
1 Tanabe 2014Retrospective cohort studyVery serious3N/ANot seriousNot serious14/18242/193

RR 0.35

(0.20, 0.63)

21.76 per 100

7.69 per 100

(4.35, 13.60)

Very low
All-cause mortality (3 months): RR “1 favours filter (Figure 6)
2***Retrospective cohort studyVery serious4Very serious2Not seriousNot serious336/15141472/3126

RR 0.35

(0.15, 0.82)

47.09 per 100

16.36 per 100

(6.91, 38.72)

Very low
PE-related mortality (in-hospital): RR “1 favours filter (Figure 7)
2***Retrospective cohort studyVery serious4Very serious2Not seriousSerious1193/1137951/2243

RR 0.12

(0.01, 2.24)

42.40 per 100

5.18 per 100

(0.28, 95.03)

Very low
PE-related mortality (3 months): RR “1 favours filter (Figure 8)
2***Retrospective cohort studyVery serious4Very serious2Not seriousSerious1215/11371009/2243

RR 0.15

(0.01, 1.89)

44.98 per 100

6.76 per 100

(0.54, 84.89)

Very low
1

95% confidence interval crosses the line of no effect.

2

I2 „66.6%

3

The study was at critical risk of bias

4

„33.3% of studies were at serious/critical risk of bias.

5

The study was at serious risk of bias.

6

The study was at moderate risk of bias.

*

The data for this analysis came from 5 individual studies, one of which (Stein 2018) contained data on two distinct populations.

**

The data for this analysis came from 2 individual studies, one of which (Stein 2018) contained data on two distinct populations.

***

The data for this analysis is from a single study (Stein 2018) containing data on two distinct populations.

Filter versus no filter in people with VTE who are at high risk of PE-occurrence

Quality assessmentNo. patientsEffectQuality
No. of studiesDesignRisk of biasInconsistencyIndirectnessimprecisionFilterNo filterRelative (95% CI)Absolute risk controlAbsolute risk intervention
All-cause mortality (12 days): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousSerious35/2005/200

RR 1.00

(0.29, 3.40)

2.50 per 100

2.50 per 100

(0.74, 8.50)

Moderate
All-cause mortality (3 months): RR “1 favours filter (Figure 9)
2RCTNot seriousNot seriousSerious5Serious330/40022/399

RR 1.36

(0.80, 2.32)

5.51 per 100

7.50 per 100

(4.40, 12.77)

Low
All-cause mortality (6 months): RR “1 favours filter
1 Mismetti 2015RCTNot seriousN/ASerious4Serious321/20015/199

RR 1.39

(0.74, 2.62)

7.54 per 100

10.50 per 100

(5.58, 19.77)

Low
All-cause mortality (8 years): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousSerious398/200103/200

RR 0.95

(0.78, 1.16)

51.50 per 100

49.00 per 100

(40.31, 59.56)

Moderate
VTE-recurrence (3 months): RR “1 favours filter
1 Mismetti 2015RCTNot seriousN/ASerious4Very serious17/2004/199

RR 1.74

(0.52, 5.86)

2.01 per 100

3.50 per 100

(1.04, 11.77)

Very-low
VTE-recurrence (6 months): RR “1 favours filter
1 Mismetti 2015RCTNot seriousN/ASerious4Very serious18/2006/199

RR 1.33

(0.47, 3.75)

3.02 per 100

4.00 per 100

(1.41, 11.32)

Very-low
VTE-recurrence (8 years): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousVery serious158/20055/200

RR 1.05

(0.77, 1.44)

27.50 per 100

29.00 per 100

(21.22, 39.64)

Low
PE-related mortality (3 months): RR “1 favours filter
1 Mismetti 2015RCTNot seriousN/ASerious4Very serious16/2002/199

RR 2.99

(0.61, 14.61)

1.01 per 100

3.00 per 100

(0.61, 14.69)

Very low
PE-related mortality (6 months): RR “1 favours filter
1 Mismetti 2015RCTNot seriousN/ASerious4Very serious16/2003/199

RR 1.99

(0.50, 7.85)

1.51 per 100

3.00 per 100

(0.76, 11.83)

Very low
PE-related mortality (8 years): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousVery serious12/2005/200

RR 0.40

(0.08, 2.04)

2.50 per 100

1.00 per 100

(0.20, 5.09)

Low
Symptomatic PE-occurrence (12 days): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousVery serious12/2005/200

RR 0.40

(0.08, 2.04)

2.50 per 100

1.00 per 100

(0.20, 5.09)

Low
Asymptomatic and symptomatic PE-occurrence (12 days): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousSerious22/2009/200

RR 0.22

(0.05, 1.02)

4.50 per 100

1.00 per 100

(0.22, 4.57)

Moderate
Symptomatic PE-occurrence (3 months): RR “1 favours filter (Figure 10)
2RCTNot seriousSeriousSerious5Very serious18/4009/399

RR 0.85

(0.15, 4.91)

2.26 per 100

1.92 per 100

(0.33, 11.09)

Low
Symptomatic PE-occurrence (6 months): RR “1 favours filter
1 Mismetti 2015RCTNot seriousN/ASerious4Very serious17/2004/199

RR 1.74

(0.52, 5.86)

2.01 per 100

3.50 per 100

(1.04, 11.77)

Very low
Symptomatic PE-occurrence (8 years): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousNot serious9/20024/200

RR 0.38

(0.18, 0.79)

12.00 per 100

4.50 per 100

(2.15, 9.44)

High
DVT-occurrence (3 months): RR “1 favours filter (Figure 11)
2RCTNot seriousNot seriousSerious5Very serious110/4007/399

RR 1.43

(0.55, 3.69)

1.75 per 100

2.50 per 100

(0.97, 6.48)

Very low
DVT-occurrence (6 months): RR “1 favours filter
1 Mismetti 2015RCTNot seriousN/ASerious4Very serious11/2002/199

RR 0.50

(0.05, 5.44)

1.01 per 100

0.50 per 100

(0.05, 5.47)

Very low
DVT-occurrence (8 years): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousSerious257/20041/200

RR 1.39

(0.98, 1.97)

20.50 per 100

28.50 per 100

(20.08, 40.45)

Moderate
Major bleeding (12 days) (RR “1 favours filter)
1 PREPIC 2005RCTNot seriousN/ANot seriousVery serious19/2006/200

RR 1.50

(0.54, 4.14)

3.00 per 100

4.50 per 100

(1.63, 12.41)

Low
Major bleeding (3 months) (RR “1 favours filter) (Figure 12)
2RCTNot seriousNot seriousSerious5Very serious119/40020/399

RR 0.95

(0.51, 1.75)

5.01 per 100

4.75 per 100

(2.58, 8.76)

Low
Major bleeding (6 months): RR “1 favours filter
1 Mismetti 2015RCTNot seriousN/ASerious4Very serious113/20015/199

RR 0.86

(0.42, 1.77)

7.54 per 100

6.50 per 100

(3.18, 13.30)

Very low
Major bleeding (8 years): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousVery serious126/20031/200

RR 0.84

(0.52, 1.36)

15.50 per 100

13.00 per 100

(8.02, 21.07)

Low
Post-thrombotic syndrome (8 years): RR “1 favours filter
1 PREPIC 2005RCTNot seriousN/ANot seriousNot serious109/200107/200

RR 1.02

(0.85, 1.22)

53.50 per 100

54.50 per 100

(45.48, 65.31)

High
1

95% confidence interval crosses both ends of a defined MID interval.

2

95% confidence interval crosses one end of a defined MID interval.

3

95% CI crosses line of no effect.

4

Study was partially applicable to the review question.

5

„33.3% of studies were partially applicable to the review question.

Filter versus no filter in people with VTE and cancer

Quality assessmentNo. patientsEffectQuality
No. of studiesDesignRisk of biasInconsistencyIndirectnessimprecisionFilterNo filterRelative (95% CI)Absolute risk controlAbsolute risk intervention
All-cause mortality (in-hospital): RR “1 favours filter
1 Stein 2018cRetrospective cohort studyVery serious1N/ANot seriousNot serious532/65893175/28445

RR 0.72

(0.66, 0.79)

11.16 per 100

8.07 per 100

(7.40, 8.82)

Very low
All-cause mortality (in-hospital, „80 year olds only): RR “1 favours filter
1 Stein 2018cRetrospective cohort studyVery serious1N/ANot seriousNot serious56/952469/3969

RR 0.50

(0.38, 0.65)

11.82 per 100

5.88 per 100

(4.50, 7.69)

Very low
All-cause mortality (30 days): HR “1 favours filter
1 Brunson 2017Retrospective cohort studySerious4N/ANot seriousNot seriousN/AN/A

HR 1.22

(1.15, 1.30)

N/AN/AVery low
All-cause mortality (3 months): RR “1 favours filter (Figure 13)
1 Barginear 2012RCTNot seriousN/ANot seriousSerious65/314/33

RR 1.33

(0.39, 4.51)

12.12 per 100

16.13 per 100

(4.76, 54.63)

Moderate
All-cause mortality (3 months): RR “1 favours filter (Figure 13)
1 Stein 2018cRetrospective cohort studyVery serious1N/ANot seriousNot serious1049/65894993/28445

RR 0.91

(0.85, 0.96)

17.55 per 100

15.92 per 100

(14.98, 16.92)

Very low
All-cause mortality (3 months): HR “1 favours filter
1 Brunson 2017Retrospective cohort studySerious4N/ANot seriousNot seriousN/AN/A

HR 1.26

(1.16, 1.37)

N/AN/AVery low
All-cause mortality (1 year): HR “1 favours filter
1 Coombs 2017Retrospective cohort studyVery serious2N/ANot seriousNot seriousN/AN/A

HR 1.26

(1.08, 1.46)

N/AN/AVery low
PE-occurrence (3 months): RR “1 favours filter
1 Barginear 2012RCTNot seriousN/ANot seriousVery serious51/311/33

RR 1.06

(0.07, 16.29)

3.03 per 100

3.23 per 100

(0.21, 49.37)

Low
PE-occurrence (long term)*: RR “1 favours filter (Figure 14)
2Retrospective cohort studyVery serious3Not seriousNot seriousVery serious511/35338/1021

RR 0.84

(0.45, 1.60)

3.72 per 100

3.14 per 100

(1.66, 5.96)

Very low
PE-occurrence (up to 1 year) sensitivity analysis excluding study at critical risk of bias: RR “1 favours filter
1 Coombs 2017Retrospective cohort studySerious4N/ANot seriousVery serious611/31733/953

RR 1.00

(0.51, 1.96)

3.72 per 100

3.14 per 100

(1.66, 5.96)

Very low
PE-occurrence (long-term): HR “1 favours filter
1 Brunson 2017Retrospective cohort studySerious4N/ANot seriousSerious6N/AN/A

HR 0.81

(0.52, 1.27)

N/AN/AVery low
DVT-occurrence (3 months): RR “1 favours filter
1 Barginear 2012RCTNot seriousN/ANot seriousNot estimable80/310/33Not estimable8Not estimable8Not estimable8High
DVT-occurrence (long-term)*: RR “1 favours filter (Figure 15)
2Retrospective cohort studyVery serious1Not seriousNot seriousNot serious34/35347/1021

RR 1.99

(1.30, 3.05)

4.60 per 100

9.18 per 100

(6.00, 14.03)

Very low
DVT-occurrence (up to 1 year) sensitivity analysis excluding study at critical risk of bias: RR “1 favours filter
1 Coombs 2017Retrospective cohort studySerious4N/ANot seriousSerious726/31740/953

RR 1.95

(1.21, 3.15)

3.72 per 100

3.14 per 100

(1.66, 5.96)

Very low
DVT-occurrence (long-term): HR “1 favours filter
1 Brunson 2017Retrospective cohort studySerious4N/ANot seriousNot seriousN/AN/A

HR 1.73

(1.31, 2.28)

N/AN/AVery low
VTE-recurrence (long term): RR “1 favours filter
1 Coombs 2017Retrospective cohort studyVery serious1N/ANot seriousSerious737/31773/953

RR 1.52

(1.05, 2.22)

N/AN/AVery low
Major bleeding (3 months): RR “1 favours filter
1 Barginear 2012RCTNot seriousN/ANot seriousVery serious51/312/33

RR 0.53

(0.05, 5.58)

6.06 per 100

3.23 per 100

(0.31, 33.82)

Low
Major bleeding (long-term, participants followed until death): RR “1 favours filter
1 Barginear 2009Retrospective cohort studyVery serious1N/ANot seriousVery serious50/369/68

RR 0.10

(0.01, 1.64)

13.24 per 100

1.30 per 100

(0.08, 21.70)

Very low
Major bleed (long-term): HR “1 favours filter
Brunson 2017 Retrospective cohort studyNot seriousN/ANot seriousSerious6N/AN/A

HR 1.11

(0.94, 1.31)

N/AN/AVery low
IVC complications (3 months): RR “1 favours filter
1 Barginear 2012RCTNot seriousN/ANot seriousVery serious52/310/33

RR 5.31

(0.27, 106.46)

0.00 per 100

0.00 per 100

(0.00, 0.00)

Low
1

The study was at critical risk of bias.

2

The study was at serious risk of bias.

3

„33.3% of studies were at serious or critical risk of bias.

4

The study was at moderate risk of bias.

5

95% confidence interval crosses both ends of a defined MID interval.

6

95% CI crosses line of no effect.

7

95% confidence interval crosses one end of a defined MID interval.

8

Effect estimate not calculable as both arms have 0 events.

*

Coombs (2017) had a follow-up of 1 year and Barginear (2009) had an unclear follow, lasting until the participants’ death.

Appendix H. Economic evidence study selection

Image niceng158er8apphf1

Appendix I. Economic evidence profiles

Download PDF (179K)

Appendix J. Excluded studies

Clinical studies

StudyReason for exclusion
Akhtar, O. S., Lakhter, V., Zack, C. J. et al (2018) Contemporary Trends and Comparative Outcomes With Adjunctive Inferior Vena Cava Filter Placement in Patients Undergoing Catheter-Directed Thrombolysis for Deep Vein Thrombosis in the United States: Insights From the National Inpatient Sample. Jacc: Cardiovascular Interventions 11(14): 1390–1397 [PubMed: 30025732]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Participants had DVT and were undergoing catheter directed thrombolysis.]

Billett, H. H., Jacobs, L. G., Madsen, E. M. et al (2007) Efficacy of inferior vena cava filters in anticoagulated patients. Journal of Thrombosis & Haemostasis 5(9): 1848–53 [PubMed: 17723124]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Population does not meet protocol as the study contained anyone with VTE who received filter]

Bikdeli, B., Chatterjee, S., Desai, N. R. et al (2017) Inferior Vena Cava Filters to Prevent Pulmonary Embolism: Systematic Review and Meta-Analysis. Journal of the American College of Cardiology 70(13): 1587–1597 [PMC free article: PMC8412839] [PubMed: 28935036] - Systematic review used as source of primary studies
Calligaro, K. D., Bergen, W. S., Haut, M. J. et al (1991) Thromboembolic complications in patients with advanced cancer: anticoagulation versus Greenfield filter placement. Annals of Vascular Surgery 5(2): 186–9 [PubMed: 2015191]

- Study does not contain a relevant intervention

[One arm were treated with anticoagulation and the other was not.]

Chen, M., Goodin, A., Xiao, H. et al (2018) Hospitalization metrics associated with hospital-level variation in inferior vena cava filter utilization for patients with venous thromboembolism in the United States: Implications for quality of care. Vascular Medicine 23(4): 365–371 [PubMed: 29781388]

- Does not contain a population of people fitting into one the categories of interest to this review, as outlined in the protocol

[Contained all people with VTE.]

Ghanim, A. J., Daskalakis, C., Eschelman, D. J. et al (2007) A five-year, retrospective, comparison review of survival in neurosurgical patients diagnosed with venous thromboembolism and treated with either inferior vena cava filters or anticoagulants. Journal of Thrombosis & Thrombolysis 24(3): 247–54 [PubMed: 17385008]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Participants had VTE and a significant risk of bleeding. Other studies have used this as a proxy for contraindication to anticoagulation however in the present study the majority of participants were noted to be receiving anticoagulation. Therefore, the population does not meet those outlined in the protocol.]

Isogai, T., Yasunaga, H., Matsui, H. et al (2015) Effectiveness of inferior vena cava filters on mortality as an adjuvant to antithrombotic therapy. American Journal of Medicine 128(3): 312.e23–31 [PubMed: 25446296]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[The study included people with PE on anticoagulation (without additional cha racteristics indicating risk level)]

Jiang, J.; Jiao, Y.; Zhang, X. (2017) The short-term efficacy of vena cava filters for the prevention of pulmonary embolism in patients with venous thromboembolism receiving anticoagulation: Meta-analysis of randomized controlled trials. Phlebology 32(9): 620–627 [PubMed: 27913756] - Systematic review used as source of primary studies
Leiderman, D. B. D., Zerati, A. E., Vieira Mariz, M. P. et al (2019) The Need for a Vena Cava Filter in Oncological Patients with Acute Venous Thrombosis: A Marker of a Worse Prognosis. Annals of Vascular Surgery 23: 23 [PubMed: 30802572]

- Study does not contain a relevant intervention

[Compared IVC without AC to AC alone]

Mismetti, P. (2013) Prevention of pulmonary embolism recurrences by retrievable vena cava filter: results of the randomized multicenter trial PREPIC 2. Journal of thrombosis and haemostasis : JTH 11(suppl2): 28 - Conference abstract
Mismetti, P. (2008) Randomized trial assessing the efficacy of the partial interruption of the inferior vena cava by an optional vena caval filter in the prevention of the recurrence of pulmonary embolism. PREPIC 2 trial: prevention of embolic recurrences by caval interruption (prospective, multicentric, randomised, open trial). Revue de pneumologie clinique 64(6): 328–331 [PubMed: 19084214] - Study not reported in English
Muriel, A., Jimenez, D., Aujesky, D. et al (2014) Survival effects of inferior vena cava filter in patients with acute symptomatic venous thromboembolism and a significant bleeding risk. Journal of the American College of Cardiology 63(16): 1675–83 [PubMed: 24576432]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Although significant bleeding risk has been used by other studies as being indicative of a contraindication to anticoagulation, the study identified that most participants received anticoagulation. Additionally, the level of anticoagulation usage was considerably different between intervention groups]

Olin, J. W., Young, J. R., Graor, R. A. et al (1987) Treatment of deep vein thrombosis and pulmonary emboli in patients with primary and metastatic brain tumors. Anticoagulants or inferior vena cava filter?. Archives of Internal Medicine 147(12): 2177–9 [PubMed: 3500686]

- Study does not contain a relevant intervention

[One arm received anticoagulation but the other did not.]

Rojas-Hernandez, C. M.; Zapata-Copete, J. A.; Garcia-Perdomo, H. A. (2018) Role of vena cava filters for the management of cancer-related venous thromboembolism: Systematic review and meta-analysis. Critical Reviews in Oncology-Hematology 130: 44–50 [PubMed: 30196911] - Systematic review used as source of primary studies
Senties, A. C.; Carrera, N. F.; Gordillo, O. G. (1977) Inferior vena cava ligation versus the Mobin-Uddin filter for prevention of recurrent pulmonary embolism. International Surgery 62(8): 420–5 [PubMed: 903202]

- Study does not contain a relevant intervention

[Study compared ligation versus filters]

Stein, P. D.; Matta, F.; Hughes, M. J. (2019) Usefulness of Inferior Vena Cava Filters in Stable Patients with Acute Pulmonary Embolism. American Journal of Cardiology 123(11): 1874–1877 [PubMed: 30952380] - Systematic review used as source of primary studies
Stein, P. D.; Matta, F.; Hughes, M. J. (2017) Inferior Vena Cava Filters in Elderly Patients with Stable Acute Pulmonary Embolism. American Journal of Medicine 130(3): 356–364 [PubMed: 27984007]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Study population does not match the groups outlined in the protocol]

Stein, P. D.; Matta, F.; Hughes, M. J. (2018) Inferior Vena Cava Filters in Stable Patients with Acute Pulmonary Embolism Who Receive Thrombolytic Therapy. American Journal of Medicine 131(1): 97–99 [PubMed: 28807710]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Contained all people with PE who underwent thrombolytic therapy.]

White, R. H., Zhou, H., Kim, J. et al (2000) A population-based study of the effectiveness of inferior vena cava filter use among patients with venous thromboembolism. Archives of Internal Medicine 160(13): 2033–41 [PubMed: 10888977]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Study is a general population of people with VTE. There is a subgroup analysis of people who have had previous VTEs however it is not possible to tell whether these occurred whilst taking anticoagulation.]

Yamashita, Y., Unoki, T., Takagi, D. et al (2016) Indications, applications, and outcomes of inferior vena cava filters for venous thromboembolism in Japanese patients. Heart & Vessels 31(7): 1084–90 [PubMed: 26135928]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Population did not meet protocol (included all people with VTE + IVC)]

Zektser, M., Bartal, C., Zeller, L. et al (2016) Effectiveness of Inferior Vena Cava Filters without Anticoagulation Therapy for Prophylaxis of Recurrent Pulmonary Embolism. Rambam Maimonides Medical Journal 7(3): 28 [PMC free article: PMC5001791] [PubMed: 27487310]

- Does not contain a population of people fitting into the categories of interest to this review, as outlined in the protocol

[Compared IVC without anticoagulation to group with anticoagulation]

Zuin, M., Rigatelli, G., Zonzin, P. et al (2019) Inferior Vena Cava Filters in Hemodynamically Unstable Patients with Acute Pulmonary Embolism: How Often are They Used? Data from Multicenter Prospective Registries on Acute Pulmonary Embolism. Cardiovascular & Interventional Radiology 42(8): 1073–1079 [PubMed: 31093719] - Systematic review used as source of primary studies

Economic studies

Author (year)TitleReason for exclusion
Raphael (2014) Pulmonary embolism after total joint arthroplasty: cost and effectiveness of four treatment modalitiesNot a cost-effectiveness analysis
Spangler (2010) Cost-effectiveness of guidelines for insertion of inferior vena cava filters in high-risk trauma patientsThe IVC filter strategy was not compared with an option without IVC filter

Appendix K. References

Included clinical studies

  • Barginear, M. F., Gralla, R. J., Bradley, T. P. et al (2012) Investigating the benefit of adding a vena cava filter to anticoagulation with fondaparinux sodium in patients with cancer and venous thromboembolism in a prospective randomized clinical trial. Supportive Care in Cancer 20(11): 2865–72 [PubMed: 22421738]
  • Barginear, M. F., Lesser, M., Akerman, M. L. et al (2009) Need for inferior vena cava filters in cancer patients: a surrogate marker for poor outcome. Clinical & Applied Thrombosis/Hemostasis 15(3): 263–9 [PubMed: 18385149]
  • Brunson, A., Ho, G., White, R. et al (2017) Inferior vena cava filters in patients with cancer and venous thromboembolism (VTE) does not improve clinical outcomes: A population-based study. Thrombosis Research 153: 57–64 [PubMed: 28334639]
  • Coombs, C., Kuk, D., Devlin, S. et al (2017) Outcomes after inferior vena cava filter placement in cancer patients diagnosed with pulmonary embolism: risk for recurrent venous thromboembolism. Journal of Thrombosis & Thrombolysis 44(4): 489–493 [PubMed: 28993967]
  • Decousus, H., Leizorovicz, A., Parent, F. et al (1998) A clinical trial of vena caval filters in the prevention of pulmonary embolism in patients with proximal deep-vein thrombosis. Prevention du Risque d’Embolie Pulmonaire par Interruption Cave Study Group. New England Journal of Medicine 338(7): 409–15 [PubMed: 9459643]
  • Group, Prepic Study (2005) 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 112(3): 416–22 [PubMed: 16009794]
  • Jha, V. M., Lee-Llacer, J., Williams, J. et al (2010) Adjunctive inferior vena cava filter placement for acute pulmonary embolism. Cardiovascular & Interventional Radiology 33(4): 739–43 [PubMed: 20526600]
  • Liang, N. L., Genovese, E. A., Avgerinos, E. D. et al (2017) Impact of Inferior Vena Cava Filter Placement on Short-Term Outcomes in Patients with Acute Pulmonary Embolism. Annals of Vascular Surgery 42: 71–77 [PMC free article: PMC5536973] [PubMed: 28341513]
  • Mellado, M., Pijoan, J. I., Jimenez, D. et al (2016) Outcomes Associated With Inferior Vena Cava Filters Among Patients With Thromboembolic Recurrence During Anticoagulant Therapy. Jacc: Cardiovascular Interventions 9(23): 2440–2448 [PubMed: 27838262]
  • Mismetti, P., Laporte, S., Pellerin, O. et al (2015) Effect of a retrievable inferior vena cava filter plus anticoagulation vs anticoagulation alone on risk of recurrent pulmonary embolism: a randomized clinical trial. JAMA 313(16): 1627–35 [PubMed: 25919526]
  • Pan, Y., Zhao, J., Mei, J. et al (2016) Evaluation of nonpermanent inferior vena cava filter placement in patients with deep venous thrombosis after lower extremity fracture: A single-center retrospective study. Phlebology 31(8): 564–72 [PubMed: 26249151]
  • Sharifi, M., Bay, C., Skrocki, L. et al (2012) Role of IVC filters in endovenous therapy for deep venous thrombosis: the FILTER-PEVI (filter implantation to lower thromboembolic risk in percutaneous endovenous intervention) trial. Cardiovascular & Interventional Radiology 35(6): 1408–13 [PubMed: 22271078]
  • Stein, P. D., Matta, F., Lawrence, F. R. et al (2018a) Usefulness of Inferior Vena Cava Filters in Unstable Patients With Acute Pulmonary Embolism and Patients Who Underwent Pulmonary Embolectomy. American Journal of Cardiology 121(4): 495–50 [PubMed: 29254678]
  • Stein, P. D., Matta, F., Lawrence, F. R. et al (2018b) Importance of Early Insertion of Inferior Vena Cava Filters in Unstable Patients with Acute Pulmonary Embolism. American Journal of Medicine 131(9): 1104–1109 [PubMed: 29906426]
  • Stein, P. D., Matta, F., Lawrence, F. R. et al (2018c) Inferior Vena Cava Filters in Patients with Acute Pulmonary Embolism and Cancer. American Journal of Medicine 131(4): 442.e9–442.e12 [PubMed: 29132839]
  • Stein, P. D., Matta, F., Lawrence, F. R. et al (2019a) Inferior Vena Cava Filters in Patients with Recurrent Pulmonary Embolism. American Journal of Medicine 132(1): 88–92 [PubMed: 30290192]
  • Stein, P. D.; Matta, F.; Hughes, M. J. (2019b) Inferior Vena Cava Filters in Stable Patients With Pulmonary Embolism and Heart Failure. American Journal of Cardiology 124(2): 292–295 [PubMed: 31097195]
  • Tanabe, Y., Obayashi, T., Yamamoto, T. et al (2014) Current status of the use of inferior vena cava filters in cases of pulmonary embolism in CCUs: From the Tokyo CCU Network. Journal of Cardiology 63(5): 385–9 [PubMed: 24239194]
  • Turner, T. E., Saeed, M. J., Novak, E. et al (2018) Association of Inferior Vena Cava Filter Placement for Venous Thromboembolic Disease and a Contraindication to Anticoagulation With 30-Day Mortality. JAMA Network Open 1(3): e180452 [PMC free article: PMC6324296] [PubMed: 30646021]
  • Wadhwa, V., Gutta, N. B., Trivedi, P. S. et al (2018) In-Hospital Mortality Benefit of Inferior Vena Cava Filters in Patients With Pulmonary Embolism and Congestive Heart Failure. AJR. American Journal of Roentgenology 211(3): 672–676 [PubMed: 30016144]
  • White, R. H., Brunson, A., Romano, P. S. et al (2016) Outcomes After Vena Cava Filter Use in Noncancer Patients With Acute Venous Thromboembolism: A Population-Based Study. Circulation 133(21): 2018–29 [PubMed: 27048765]

Other included clinical studies

  • Higgins, J.P., Altman, D.G., Gøtzsche, P.C., Jüni, P., Moher, D., Oxman, A.D., Savović, J., Schulz, K.F., Weeks, L. and Sterne, J.A., 2011. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJj, 343, p.d5928 [PMC free article: PMC3196245] [PubMed: 22008217]
  • Jia, Z., Wu, A., Tam, M., Spain, J., McKinney, J. M., & Wang, W. (2015). Caval penetration by inferior vena cava filters: a systematic literature review of clinical significance and management. Circulation, 132(10), 944–952. [PubMed: 26169756]

Excluded clinical studies

  • Akhtar, O. S., Lakhter, V., Zack, C. J. et al (2018) Contemporary Trends and Comparative Outcomes With Adjunctive Inferior Vena Cava Filter Placement in Patients Undergoing Catheter-Directed Thrombolysis for Deep Vein Thrombosis in the United States: Insights From the National Inpatient Sample. Jacc: Cardiovascular Interventions 11(14): 1390–1397 [PubMed: 30025732]
  • Bikdeli, B., Chatterjee, S., Desai, N. R. et al (2017) Inferior Vena Cava Filters to Prevent Pulmonary Embolism: Systematic Review and Meta-Analysis. Journal of the American College of Cardiology 70(13): 1587–1597 [PMC free article: PMC8412839] [PubMed: 28935036]
  • Billett, H. H., Jacobs, L. G., Madsen, E. M. et al (2007) Efficacy of inferior vena cava filters in anticoagulated patients. Journal of Thrombosis & Haemostasis 5(9): 1848–53 [PubMed: 17723124]
  • Calligaro, K. D., Bergen, W. S., Haut, M. J. et al (1991) Thromboembolic complications in patients with advanced cancer: anticoagulation versus Greenfield filter placement. Annals of Vascular Surgery 5(2): 186–9 [PubMed: 2015191]
  • Chen, M., Goodin, A., Xiao, H. et al (2018) Hospitalization metrics associated with hospital-level variation in inferior vena cava filter utilization for patients with venous thromboembolism in the United States: Implications for quality of care. Vascular Medicine 23(4): 365–371 [PubMed: 29781388]
  • Ghanim, A. J., Daskalakis, C., Eschelman, D. J. et al (2007) A five-year, retrospective, comparison review of survival in neurosurgical patients diagnosed with venous thromboembolism and treated with either inferior vena cava filters or anticoagulants. Journal of Thrombosis & Thrombolysis 24(3): 247–54 [PubMed: 17385008]
  • Isogai, T., Yasunaga, H., Matsui, H. et al (2015) Effectiveness of inferior vena cava filters on mortality as an adjuvant to antithrombotic therapy. American Journal of Medicine 128(3): 312.e23–31 [PubMed: 25446296]
  • Jiang, J.; Jiao, Y.; Zhang, X. (2017) The short-term efficacy of vena cava filters for the prevention of pulmonary embolism in patients with venous thromboembolism receiving anticoagulation: Meta-analysis of randomized controlled trials. Phlebology 32(9): 620–627 [PubMed: 27913756]
  • Leiderman, D. B. D., Zerati, A. E., Vieira Mariz, M. P. et al (2019) The Need for a Vena Cava Filter in Oncological Patients with Acute Venous Thrombosis: A Marker of a Worse Prognosis. Annals of Vascular Surgery 23: 23 [PubMed: 30802572]
  • Mismetti, P. (2013) Prevention of pulmonary embolism recurrences by retrievable vena cava filter: results of the randomized multicenter trial PREPIC 2. Journal of thrombosis and haemostasis : JTH 11(suppl2): 28
  • Mismetti, P. (2008) Randomized trial assessing the efficacy of the partial interruption of the inferior vena cava by an optional vena caval filter in the prevention of the recurrence of pulmonary embolism. PREPIC 2 trial: prevention of embolic recurrences by caval interruption (prospective, multicentric, randomised, open trial). Revue de pneumologie clinique 64(6): 328–331 [PubMed: 19084214]
  • Muriel, A., Jiménez, D., Aujesky, D., Bertoletti, L., Decousus, H., Laporte, S., … & RIETE investigators. (2014). Survival effects of inferior vena cava filter in patients with acute symptomatic venous thromboembolism and a significant bleeding risk. Journal of the American College of Cardiology, 63(16), 1675–1683. [PubMed: 24576432]
  • Olin, J. W., Young, J. R., Graor, R. A. et al (1987) Treatment of deep vein thrombosis and pulmonary emboli in patients with primary and metastatic brain tumors. Anticoagulants or inferior vena cava filter?. Archives of Internal Medicine 147(12): 2177–9 [PubMed: 3500686]
  • Rojas-Hernandez, C. M.; Zapata-Copete, J. A.; Garcia-Perdomo, H. A. (2018) Role of vena cava filters for the management of cancer-related venous thromboembolism: Systematic review and meta-analysis. Critical Reviews in Oncology-Hematology 130: 44–50 [PubMed: 30196911]
  • Senties, A. C.; Carrera, N. F.; Gordillo, O. G. (1977) Inferior vena cava ligation versus the Mobin-Uddin filter for prevention of recurrent pulmonary embolism. International Surgery 62(8): 420–5 [PubMed: 903202]
  • Stein, P. D.; Matta, F.; Hughes, M. J. (2018) Inferior Vena Cava Filters in Stable Patients with Acute Pulmonary Embolism Who Receive Thrombolytic Therapy. American Journal of Medicine 131(1): 97–99 [PubMed: 28807710]
  • Stein, P. D.; Matta, F.; Hughes, M. J. (2019) Usefulness of Inferior Vena Cava Filters in Stable Patients with Acute Pulmonary Embolism. American Journal of Cardiology 123(11): 1874–1877 [PubMed: 30952380]
  • Stein, P. D.; Matta, F.; Hughes, M. J. (2017) Inferior Vena Cava Filters in Elderly Patients with Stable Acute Pulmonary Embolism. American Journal of Medicine 130(3): 356–364 [PubMed: 27984007]
  • White, R. H., Zhou, H., Kim, J. et al (2000) A population-based study of the effectiveness of inferior vena cava filter use among patients with venous thromboembolism. Archives of Internal Medicine 160(13): 2033–41 [PubMed: 10888977]
  • Yamashita, Y., Unoki, T., Takagi, D. et al (2016) Indications, applications, and outcomes of inferior vena cava filters for venous thromboembolism in Japanese patients. Heart & Vessels 31(7): 1084–90 [PubMed: 26135928]
  • Zektser, M., Bartal, C., Zeller, L. et al (2016) Effectiveness of Inferior Vena Cava Filters without Anticoagulation Therapy for Prophylaxis of Recurrent Pulmonary Embolism. Rambam Maimonides Medical Journal 7(3): 28 [PMC free article: PMC5001791] [PubMed: 27487310]
  • Zuin, M., Rigatelli, G., Zonzin, P. et al (2019) Inferior Vena Cava Filters in Hemodynamically Unstable Patients with Acute Pulmonary Embolism: How Often are They Used? Data from Multicenter Prospective Registries on Acute Pulmonary Embolism. Cardiovascular & Interventional Radiology 42(8): 1073–1079 [PubMed: 31093719]

Included economic studies

  • Sarasin FP and Eckman MH (1993) Management and prevention of thromboembolic events in patients with cancer-related hypercoagulable states: a risky business. Journal Internal Medicine 8: 476–486 [PubMed: 8410419]

Excluded economic studies

  • Raphael IJ, Mckenzie JC, Zmistowski B et al (2014) Pulmonary embolism after total joint arthroplasty: cost and effectiveness of four treatment modalities. The Journal of Arthroplasty 29: 933–937 [PubMed: 24269095]
  • Spangler EL, Dillavou Ed and SmitheKJ (2010) Cost-effectiveness of guidelines for insertion of inferior vena cava filters in high-risk trauma patients. Journal of Vascular Surgery 52: 1537–45 [PubMed: 20843631]

Appendix L. Research recommendation

Research questionWhat is the short and long term clinical and cost effectiveness of inferior vena caval filters in people with VTE?
PopulationAdults (aged 18+) with confirmed VTE
Intervention(s)IVC filter with or without:
  • mechanical intervention and/or
  • anticoagulant treatment.
ComparatorNo filter with:
  • mechanical intervention and/or
  • anticoagulant treatment and/or
  • placebo or no treatment.
Studies can allow participants to have mechanical interventions, anticoagulation treatment or both, but these must be included in both arms of the trial so that the only difference in treatment between arms is the inclusion or exclusion of IVC filters.
Outcomes
  • Recurrent VTE (PE and DVT)
  • All-cause mortality
  • VTE-related mortality
  • Post-thrombotic syndrome
  • Pulmonary hypertension (PH)
  • Quality of life
    • Generic and disease-specific measures will be reported
    • Overall score will be reported (data on subscales will not be reported)
    • Adverse events
    • Total serious adverse events (as defined by the European medicines agency) will be reported if data is available.
    • Major bleeding (as defined by International Society on Thrombosis and Haemostasis)
    • Clinically relevant non-major bleeding (as defined by International Society on Thrombosis and Haemostasis)
    • Surgical complications at the time of placement and removal
    • Sepsis (or serious infections) for filters that are in place for longer periods
    • Resource use and costs
Outcome measures
  • Risk ratios
  • hazard ratios
  • Event data
Study designs
  • Randomised controlled trial or
  • Prospective cohort study
Subgroups of interestAdults (aged 18+) with confirmed VTE:
  • who cannot have anticoagulants or
  • who have a PE whilst taking anticoagulants or
  • who have the filters inserted for prophylaxis before a potential provoking event (e.g. surgery) or
  • who are at high risk of poor outcomes if they had further PEs or
  • who are at high risk of a PE (only for prospective cohort study as RCTs are available for this group) or
  • and cancer
Other subgroups of the above populations:
  • Type of surgery
  • People with chronic liver disease
  • Intravenous drug users
  • People who are obese (BMI ≥ 40 kg/m2)
Potential criterionExplanation
Importance to patients, service users or the population

IVC filters are currently placed in people with VTE in a range of different clinical scenarios (see population subgroups in PICO table above). However, there is continuing uncertainty about their clinical and cost-effectiveness in these different situations. More and higher quality evidence could help to establish in which clinical scenarios it would be beneficial to use filters in people with VTE.

Further RCT evidence will help to add to the evidence base and would be likely to provide the best quality evidence. However, the feasibility of carrying out RCTs in the above clinical scenarios may be limited by the number of people they can recruit and lack statistical power as a result. In contrast, a large prospective cohort study could provide a sufficiently large sample to capture enough events to clearly establish the efficacy of IVC filters in people with VTE and the subgroups specified above.

Relevance to NICE guidance

High priority: the research is essential to inform future updates of key recommendations in the guidance.

(The committee agreed that because of the lack of high quality evidence they could not make positive recommendations for the use of filters in several of the population subgroups they identified.)

Current evidence base

There is limited high quality evidence comparing the use of IVC filters to no filters in people with VTE. There are very few RCTs and most of the evidence came from retrospective cohort studies which have serious methodological limitations. The evidence base for each population subgroup is summarised below.

In people with VTE and a contraindication to anticoagulation, filters are being used as an alternative to anticoagulation. Evidence for the use of IVC filters in this population comes from two retrospective studies that have methodological shortcomings and report conflicting results for short-term all-cause mortality.

In people with VTE who are undergoing a provoking event, the evidence is very heterogeneous (the studies contain people undergoing different surgical procedures) and inconclusive. Several of the studies were at critical risk of bias. The only RCT suggests a benefit for PE-occurrence associated with filter use however the committee agreed that further evidence is needed to confirm this result.

There are two studies looking at filters in people with VTE who had a recurrent PE whilst taking anticoagulation. One study was at critical risk of bias due to very serious methodological problems. The other study suggested a benefit of filters at 30 days for all-cause and PE-related mortality. The committee again advised that further evidence is needed to clarify this effect.

Two RCTs were available for the use of filters in people with VTE who are at high risk of PE. However, for most outcomes these studies could not differentiate outcomes between the filter and no-filter group, and therefore a clear benefit or harm to filters is not established for this group of people.

In people with VTE who are at high risk of poor outcomes if they were to have a PE, evidence was again conflicting, and the committee felt unable to make recommendations without further research. When pooled together, evidence suggested a reduction for in-hospital all-cause mortality. However, the evidence was at serious-critical risk of bias, and the only study at moderate risk of bias suggested an increase in in-hospital all-cause mortality in those people given a filter.

In people with VTE and cancer, where the filters were placed specifically because the person had cancer, evidence was conflicting. The only available RCT could not differentiate any of the outcomes of interest to this review. Some of the cohort studies showed a benefit for filters and others showed a harm. Further evidence is needed to clarify the effectiveness of filters in this group of people.

EqualityNo specific equality concerns are relevant to this research recommendation.
Feasibility

There is a sufficiently large and well-defined population available that a high quality prospective cohort study should be possible.

A high quality RCT may be less feasible as there are a likely to be limited number of people in each of the specific population subgroups that are candidates for IVC filters.

Final

Evidence review underpinning recommendations 1.7.1 to 1.7.4 and the research recommendation in the guideline

This evidence review was developed by the NICE Guideline Updates Team

Disclaimer: The recommendations in this guideline represent the view of NICE, arrived at after careful consideration of the evidence available. When exercising their judgement, professionals are expected to take this guideline fully into account, alongside the individual needs, preferences and values of their patients or service users. The recommendations in this guideline are not mandatory and the guideline does not override the responsibility of healthcare professionals to make decisions appropriate to the circumstances of the individual patient, in consultation with the patient and/or their carer or guardian.

Local commissioners and/or providers have a responsibility to enable the guideline to be applied when individual health professionals and their patients or service users wish to use it. They should do so in the context of local and national priorities for funding and developing services, and in light of their duties to have due regard to the need to eliminate unlawful discrimination, to advance equality of opportunity and to reduce health inequalities. Nothing in this guideline should be interpreted in a way that would be inconsistent with compliance with those duties.

NICE guidelines cover health and care in England. Decisions on how they apply in other UK countries are made by ministers in the Welsh Government, Scottish Government, and Northern Ireland Executive. All NICE guidance is subject to regular review and may be updated or withdrawn.

Copyright © NICE 2020.
Bookshelf ID: NBK556697PMID: 32374569

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.1M)

In this Page

Other titles in this collection

Related NICE guidance and evidence

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Similar articles in PubMed

See reviews...See all...

Recent Activity

ClearTurn OffTurn On

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...