MedBulletinsAura Celeste
Español
Vascular

Special edition: inferior vena cava filters — Retrieval and the worst complications — October 30, 2026

🎧 Listen to this issueAudio newscast · 6:01 min
Download MP3Send audio via WhatsApp

Special edition on inferior vena cava (IVC) filters. Period covered: September 25 to October 30, 2026. This issue brings four news items from the period: two randomized trials presented at VIVA 2026 (montelukast after endovascular treatment and transdermal oxygen for the surgical wound) and two US regulatory decisions (the approval of a vascular closure device for large-bore arterial access, and the clearance of a vacuum thrombectomy system for deep vein thrombosis). There were no filter-specific updates in the period: no new FDA safety communications, no filter recalls initiated in the period, and no new filter-specific guidelines. The 2026 studies that contribute most to the topic (retrieval programs and rescue techniques) were published before the period and are discussed as background in the deep dive, which this month covers retrieval and the worst complications.

Updates from the period

1. CADET-PAD: montelukast after endovascular treatment for peripheral artery disease

[Randomized trial · VIVA 2026 meeting · Poland]

First results of CADET-PAD, a prospective, randomized, double-blind trial registered by Jagiellonian University (Kraków), were presented at VIVA 2026 (Las Vegas, October 3 to 6). It enrolled 200 patients with Rutherford 3 and 4 peripheral artery disease undergoing endovascular treatment (mean age 64.7 years; 78.9% Rutherford 3; 39.0% with type 2 diabetes), randomized 1:1 to montelukast 10 mg daily for 12 months or placebo, on top of standard post-angioplasty treatment. At 12 months there were fewer major adverse limb events (MALE) with montelukast (16 vs 34; P < .001) and primary patency was higher (84.0% vs 66.0%; P < .001); quality of life (VascuQoL) and endothelial function parameters also improved.

Why it matters: montelukast is an inexpensive oral cysteinyl leukotriene receptor antagonist commonly used for asthma. If the effect is confirmed, it would open an anti-inflammatory route to reducing reocclusion after peripheral angioplasty.

Critical appraisal: these are meeting results, not yet peer-reviewed. The release does not report the number of centers; the sponsor is Jagiellonian University (registry NCT04277702). Nor does it define a formal primary endpoint (the registry describes the reocclusion rate as the objective). MACE (2 vs 6) and deaths (1 vs 3) are too few to draw conclusions, although the release assigns them significant P values. With 200 patients, this is a promising hypothesis that needs confirmation before it changes practice.

Sources

2. Continuous transdermal oxygen for the surgical wound after lower-extremity revascularization

[Randomized trial · VIVA 2026 meeting]

A prospective, randomized, single-center trial presented at VIVA 2026 enrolled 118 adults undergoing elective lower-extremity revascularization, randomized 1:1 to continuous transdermal oxygen therapy with the EPIFLO device for 28 days plus standard wound care, or standard wound care alone. Complete healing of the surgical incision at four weeks was 69% vs 50% (P = .01), with a median time to complete healing of 25 vs 40 days. Surgical site infections were 0 vs 3 (P = .09; not significant), and there were no device-related adverse events.

Why it matters: incision complications after lower-extremity revascularization surgery prolong care and can compromise the result; a noninvasive measure that speeds healing would have direct application.

Critical appraisal: a single center and 118 patients; the release does not describe blinding of wound assessment or funding, and there is no publication yet. The reduction in infections was not significant, so the demonstrated benefit is limited to the speed of healing.

Sources

3. FDA approves the PerQseal Elite vascular closure system for large-bore arterial access

[Regulator · Device · US]

FDA approved the PerQseal Elite vascular closure system for large-bore arterial procedures with 14–22F sheaths (arteriotomies up to 26F), Haemonetics announced on October 19, 2026. According to Vivasure, the premarket approval (PMA) application it filed in June 2025 built on the pivotal PATCH study. The approval release instead cites results of the European prospective, multicenter, single-arm ELITE study: per the company, 0% major device-related vascular access-site complications according to VARC-3 through 30 days and mean time to hemostasis under one minute.

Why it matters: percutaneous EVAR and TEVAR depend on reliable closure of large-bore femoral access; this is one more sutureless closure option for those sheaths.

Critical appraisal: the cited data come from the company release; the study described has no control group. The release does not specify which data the FDA relied on or the number of patients. The company says the clinical data will be presented at TCT 2026 (October 31 to November 3).

Sources

4. FDA clears Penumbra's Lightning computer-assisted vacuum thrombectomy for DVT

[Regulator · Device · US]

Penumbra announced on October 20, 2026, that the FDA cleared its Lightning computer-assisted vacuum thrombectomy (CAVT) technology for the treatment of deep vein thrombosis. The clinical support is BOLT, a prospective, multicenter IDE study presented at The VEINS 2026 (Las Vegas): 229 patients with acute iliofemoral DVT at 39 US sites. Complete or near-complete thrombus reduction (75% or more by Marder score, core-lab assessed) was achieved in 98.0% (198/202), with 48-hour major adverse events in 3.1% (7/229), a device time of 18 minutes, and single-session treatment in 99.1%. At 30 days, major adverse events were 7.6% and rethrombosis 3.9%. BOLT results were already summarized in the October 5 and October 19 issues; the news here is the clearance.

Why it matters: it adds a thrombolytic-free mechanical thrombectomy option for iliofemoral DVT; one of the co-principal investigators proposes it as first-line treatment.

Critical appraisal: the study describes no control group, and the reported follow-up reaches 30 days; planned follow-up is two years. The release does not explain why effectiveness was calculated on 202 of 229 patients. The long-term effect on post-thrombotic syndrome, which is what matters to patients, is not yet known.

Sources

IVC filters: no filter-specific updates in the period

There were no new filter-specific FDA safety communications and no filter recalls initiated between September 25 and October 30, 2026. For reference, the Class II recall of Argon Medical Devices' Option ELITE vena cava filter system (for increased resistance when advancing the dilator within the introducer sheath; 1,821 units, 1,794 in the US) was initiated on September 12, 2025 and posted by the FDA on November 24, 2025, outside the period.

Sources

Retrieval and the worst complications

The core problem: placed for a transient risk, left in forever

The rationale for a retrievable filter is to cover a window of risk when anticoagulation is not possible. Randomized evidence does not support adding a filter to anticoagulation: in PREPIC2 (retrievable filter plus anticoagulation vs anticoagulation alone in patients hospitalized for acute PE with lower-limb venous thrombosis and at least one severity criterion), recurrent PE at 3 months was 3% vs 1.5% (RR 2.00; 95% CI 0.51–7.89; P = .50; not significant) (Mismetti P et al., JAMA, 2015) [strong evidence: randomized trial]. Do not routinely add filters to anticoagulation; exceptional indications require specialist assessment.

The 2026 US multisociety acute PE guideline (Creager MA, Barnes GD, Giri J et al., Circulation, online February 19, 2026) holds that line: according to a summary by one of its authors, the clearest indication is acute VTE with a contraindication to anticoagulation, temporary filters are nearly always preferred, and retrieval should occur at the earliest clinically indicated time point [guideline recommendation; that summary does not give the class or level] (Giri J, Endovascular Today, April 2026).

When to retrieve

In its May 2014 update, the FDA suggested that once the transient PE risk has passed, the risk-benefit balance begins to favor removal between days 29 and 54 after implantation. That window comes from a decision analysis based on published literature (Morales JP et al., J Vasc Surg Venous Lymphat Disord, 2013), not from a trial [model, not direct clinical evidence]. In practice: retrieve when pulmonary embolism risk has sufficiently decreased and anticoagulation is tolerated, following individual assessment and planned follow-up.

Real-world retrieval rates and forgotten filters

Figures improve mainly where there is a program. In trauma, Abrantes et al. (Mayo Clinic; Eur J Trauma Emerg Surg, March 2026; n = 300) reported that 64% of prophylactic filters had been retrieved at last follow-up, but only 19% within 90 days (median dwell 117 days) [single-center retrospective cohort]. In lung transplant recipients with structured follow-up, retrieval reached 66.3% (63 retrieved of 95 recipients; Naser-Tavakolian K et al., Front Radiol, March 2026). In Brazil, an analysis of 21,630 procedures from 2015 to 2023 showed about a 140% increase in use, with a per-capita rate in the private system about four times that of the public system (4.25 times, per the conclusions), but no retrieval data because the databases lack follow-up (Bueno CS, Wolosker N et al., Clinics, 2026) [administrative data]. That is the scenario closest to ours: we know how many we place, not how many we remove.

Effective programs share features that are no longer anecdotal. In a qualitative study of seven US hospitals serving Medicare beneficiaries (four with risk-adjusted 1-year retrieval of 40–74% and three intermediate performers at 7–40%), barriers were the same everywhere; what distinguished the high performers was an empowered nonphysician champion, protocolized surveillance, presumptive communication with the treating physician (the filter is assumed to come out unless there is an objection), and extra visits or imaging reserved for complex patients; electronic dashboards helped workflow but did not distinguish the leaders (Trivedi P et al., J Am Coll Radiol, online July 18, 2026) [qualitative].

Techniques: standard and advanced

The standard technique is hook capture with a snare or cone via the jugular approach. When it fails, because of tilt, a hook embedded in the wall, or endothelialized struts, advanced techniques are used. The systematic review by Kethidi et al. (Diagn Interv Radiol, 2023; 19 studies, 758 patients, 770 attempts) found overall success of 92.6%, major complications of 2.8%, and no procedure-related deaths. By technique: forceps 95.5%, laser 98.8%, and modified loop snare 86.0% [systematic review of mostly retrospective series, with publication bias toward expert centers].

This year brought techniques for specific scenarios, all in small Chinese series: hook capture with pull-assisted traction for conical filters with the hook attached to the wall (37/44, 84.1%; Tian X, Liu J et al., Front Surg, January 2026), "en bloc" retrieval of thrombosed filters with the thrombus trapped inside (51/51, with minor asymptomatic PE in 7.8% on routine CT pulmonary angiography; Wu Z, Chen G et al., Front Radiol, March 2026), and laparoscopic removal in 28 patients, 23 of them after a failed endovascular attempt and 5 for anticipated technical difficulty on preoperative imaging (Jiang P, Liu J et al., Front Physiol, June 2026; see below). In Tian's series, filter deformation was associated with failure, whereas dwell time and tilt did not differ between successes and failures (only seven failures, imprecise estimates).

The worst complications: how frequent and what to do

Frequency figures depend heavily on how complications are sought: with routine CT, asymptomatic penetration is common; in clinical series, symptomatic complications are rare. In the Mayo Clinic trauma study, 4% had filter-related complications: filter thrombus (6), erosion (3), migration (2), fracture (1), and failed retrieval (1), with 23% symptomatic VTE despite the filter (Abrantes VB et al., 2026). In 2010 the FDA had already logged 921 adverse event reports (device migration, detachment of device components, fracture, and IVC perforation), and an 80-patient study cited at the time found fracture in 25% of Bard Recovery filters and 12% of G2 filters (MassDevice, 2010).

  • IVC and filter thrombosis: associated with dwell time and post-thrombotic syndrome. If there is thrombus in the filter, anticoagulate and, at experienced centers, retrieve with techniques that trap the thrombus.
  • Perforation with injury to adjacent organs: most cases are asymptomatic; symptomatic cases (pain, bleeding, duodenal, ureteral, or aortic injury) require CT angiography and a joint plan between interventional radiology and vascular surgery. Two filters with a strut inside the duodenum (a 9-year-old permanent Simon Nitinol and an 8-year-old Günther Tulip) were removed endovascularly without vascular or bowel injury by a single expert operator (Zhang B et al., AME Case Rep, 2026). By contrast, in the 28-patient laparoscopic series, removal was completed laparoscopically in 25 (89.3%), with 3 conversions to open surgery and overall technical success of 27/28 (96.4%), but there was one death (3.6%) from IVC injury with massive hemorrhage (Jiang P et al., Front Physiol, 2026): surgical removal is not a minor procedure. Keep an IVC rescue plan at hand (occlusion balloon, covered stent graft).
  • Fracture and fragment embolization: fragments in the heart or pulmonary artery are assessed individually by an expert team, according to symptoms, location, mobility, perforation, and the risks of removal versus observation; there is no automatic rule based on location (2020 SIR recommendation 12; context, not a new guideline in this period).
  • PE despite the filter and migration: these call for reassessing anticoagulation before placing a second filter.

Filters and anticoagulation

A filter does not anticoagulate. If the contraindication resolves, anticoagulate according to the underlying VTE and schedule retrieval. On direct oral anticoagulants with a filter, the meta-analysis by Wang G et al. (Front Cardiovasc Med, 2026; four studies, 532 patients) found no significant differences between rivaroxaban and vitamin K antagonists in recurrence, filter thrombosis, or bleeding [weak evidence: few studies, mostly observational; the only randomized trial included is still ongoing].

Clinical documentation

Documenting the indication, consent, follow-up and potential retrieval plan, and communication with the patient preserves a traceable record of care.

New devices

Absorbable filters remain investigational; the US pivotal trial of the Adient absorbable filter is registered (NCT05127915).

For practice in the Dominican Republic:

  • Every filter leaves the procedure room with a written planned retrieval date and a named owner; a nurse or coordinator with the list is worth more than software.
  • Presumptive communication: the filter is assumed to come out unless the treating physician documents an objection, once pulmonary embolism risk has sufficiently decreased and anticoagulation is tolerated.
  • In trauma, if the conditions for retrieval are already met at discharge, consider retrieval before discharge; if not, hand the patient the reassessment appointment: after discharge they are lost.
  • Consider pre-retrieval imaging based on dwell time, symptoms, anatomy, and local expertise; refer when complex retrieval is anticipated (2020 SIR recommendation 14; context, not a new guideline in this period).
  • Do not routinely add filters to anticoagulation; exceptional indications require specialist assessment. If a filter is placed, anticoagulate as soon as possible.
  • Create a minimal national registry (date, model, indication, retrieval): today, like Brazil, we do not know how many we remove.

References

Get every issue by email

Free. One email with each new issue, with a short summary and links to the page and the audio. No ads inside the content: any sponsor is identified and has no say in the selection. We do not share your data; unsubscribe with a single message.

Subscribe by email
← October 28, 2026