Long-term retrievability of IVC filters: should we abandon permanent devices?

Thromboembolic disease produces a considerable disease burden, with death from pulmonary embolism in the UK alone estimated at 30,000-40,000 per year. Whilst it is unproven whether filters actually improve longevity, the morbidity and mortality associated with thromboembolic disease in the presence of contraindications to anticoagulation is high. Thus complications associated with filter insertion, and whilst they remain in situ, must be balanced against the alternatives. Permanent filters remain in situ for the remainder of the patient's life and any complications from the filters are of significant concern. Filters that are not permanent are therefore attractive in these circumstances. Retrievable filters, to avoid or decrease long-term filter complications, appear to be a significant advance in the prevention of pulmonary embolism. In this review, we discuss the safety and effectiveness of both permanent and retrievable filters as well as the retrievability of retrievable inferior vena cava (IVC) filters, to explore whether the use of permanent IVC filters can be abandoned in favor of retrievable filters. Currently four types of retrievable filters are available: the Recovery filter (Bard Peripheral Vascular, Tempe, AZ, USA), the Günther Tulip filter (Cook, Bloomington, IN, USA), the OptEase Filter (Cordis, Roden, The Netherlands), and the ALN filter (ALN Implants Chirurgicaux, Ghisonaccia, France). Efficacy and safety data for retrievable filters are as yet based on small series, with a total number of fewer than 1,000 insertions, and follow-up is mostly short term. Current long-term data are poor and insufficient to warrant the long-term implantation of these devices into humans. The case of fractured wire from a Recovery filter that migrated to the heart causing pericardial tamponade requiring open heart surgery is a reminder that any new endovascular device remaining in situ in the long term may produce unexpected problems. We should also bear in mind that the data on permanent filters are much more robust, with reports on over 9,500 cases with follow-up of up to 8 years. The original implantation time of 10-14 days has been extended to more than 100 days as the mean implantation time with some of the filter types. Follow-up (preferably prospective) is necessary for all retrievable filters, whether or not they are retrieved. Until these data become available we should restrict ourselves to the present indications of permanent and retrievable filters. If long-term follow-up data on larger numbers of cases confirm the initial data that retrievable filters are as safe and effective as permanent filters, the use of the retrievable filters is likely to expand.