Vena Cava Filter – Insertion

Permanent versus Optional Filter

The only proven benefit of vena cava filters is prevention of PE. Unfortunately, evidence-based recommendations for vena cava filter use are derived predominantly from nonrandomized data, with substantial differences existing among studies in terms of study populations, immediate and longterm endpoints, and duration of follow-up. Because of difficulties comparing data on different filters, several guidelines outlining reporting standards for filter devices have been published. Overall reported complications for vena cava filters include PE (2% to 5%), fatal PE (0.7%), death linked to filter insertion (0.12%), venous access site thrombosis (2% to 28%), filter migration (3% to 69%), vena cava penetration (9% to 24%), vena cava obstruction (6% to 30%), venous insufficiency (5% to 59%), filter fracture (1%), and guide wire entrapment (<1%).

Although it is unclear whether filters actually improve survival, they do provide protection against PE. Because the risk for PE in patients with proven venous thromboembolism and contraindications to anticoagulation is high, potential complications of filters must be balanced against the risk of no filter. Despite the large number of filters placed, only one randomized study on permanent filter use has been conducted, and it showed some correlation with increased potential problems with permanent filters over time. In the PREPIC (Prevention of Recurrent Pulmonary Embolism by Vena Cava Interruption) trial, 400 patients with proximal DVT with or without PE were randomized in a 2 × 2 factorial study design to filter placement versus no filter and unfractionated heparin versus enoxaparin. There was a significantly lower incidence of PE with filter protection during the first 12 days (1.1% versus 4.8%), but the filter group had a significantly increased incidence of recurrent DVT at 2 years (20.8% versus 11.6%). Eight-year follow-up data from the PREPIC trial confirmed the previous findings of cumulative recurrent PE (6.2% versus 15.1%, filter versus no filter, respectively) but increased recurrent DVT (35.7% versus 27.5%) and no difference in post-thrombotic venous insufficiency or survival. The authors concluded that although permanent filter use may be beneficial in patients at high risk for PE, systematic use in the general population with venous thromboembolism is not recommended. Unfortunately, no firm conclusions regarding filter efficacy in the prevention of PE can be drawn from the PREPIC trial given that the study design varies significantly from the wider application in current clinical practice, notably patients with documented venous thromboembolism in whom anticoagulation has failed or cannot be administered.

Although the potential for significant adverse outcomes with permanent filters is low, there are long-term complications that could be avoided with optional filters. However, given the paucity of randomized data, determining the best permanent versus optional filter design is difficult. Some studies have used meta-analysis of data to compare the different filter designs and document the efficacy of filters in the prevention of PE and complication rates. Table 78-1-1 summarizes data on different filter types; overall, most of the available filters are roughly equivalent in prevention of PE, but there is some variation in complication rates. Confounding any comparison is the fact that the more recently approved filters, though appearing to have lower PE rates, have smaller study populations and shorter follow-up intervals, thus making any comparative analysis imprecise.

Data on the efficacy and safety of retrievable filters are derived from small series with insufficient long-term data in comparison to other permanent filter designs to warrant permanent implantation. With the lack of extended outcome data for optional filters, the decision to use a retrievable filter instead of a permanent filter should be based on the intent to discontinue filtration. Factored into this decision should be the anticipated required duration of protection from venous thromboembolism versus the risk associated with anticoagulation. In patients with proven venous thromboembolism, anticoagulation should be resumed as soon as possible when the risk has diminished. For patients in whom filters are placed for prophylaxis of venous thromboembolism, anticoagulation should be restarted in accordance with published venous thromboembolism prophylaxis guidelines. When the risk of resuming anticoagulation is extended, permanent filtration would be preferable.

Further complicating the decision to use retrievable filters is poorly defined timing of possible retrieval. Based on FDA indications for use, retrieval of Günther Tulip filters is recommended within 20 days and OptEase filters within 14 days. The Recovery G-2 filter has a recommended retrieval window within a mean of 140 days and maximum of 300 days and the Celect filter within 52 weeks, a timing recommendation based on clinical trial data submitted to the FDA but not yet published. Further confusing optimal timing are case reports suggesting that potential retrieval at extended periods is possible for all filters but can be more problematic with a higher failure rate. Until further data are available, the optimal period for retrieval of currently available optional filters probably falls within a few months of placement, after which the technical success of retrieval will diminish.

Retrievable filters should be considered in the following clinical scenarios: (1) indications for permanent filters are not present, (2) the risk of clinically significant PE is acceptably low, (3) return to high risk for venous thromboembolism is not anticipated, (4) life expectancy is long enough that the benefit of filter removal will be realized, and (5) the filter can be removed safely or converted. Until further evidence is available, optional/retrievable filters should be used when there is a temporary need to prevent PE in a patient with documented venous thromboembolism and when there is a defined retrieval endpoint; these filters should not be used as a replacement for a permanent filter if permanent PE prevention is needed.