Retrievability and Device-Related Complications of the G2 Filter: A Retrospective Study of 139 Filter Retrievals

Abstract

To evaluate the retrievability and safety of the G2 filter. A retrospective study of all G2 filter retrievals at a single institution was conducted. Hospital records and imaging studies were reviewed for complications, and factors affecting retrieval were analyzed. From 2005 to 2009, a total of 139 patients presented for retrieval of their G2 filter, and 131 pairs of pre- and post-placement cavagrams and 39 computed tomography scans were available for analysis. The following findings were recorded: limb penetration (n = 33), tilt greater than 15° (n = 22), local migration greater than 2 cm (n = 17), retained thrombus within the filter (n = 16), deformity (n = 10), inferior vena cava (IVC) occlusion (n = 3), fracture (n = 2), and pulmonary embolism breakthrough (n = 2). A total of 118 filters were removed, with a mean indwelling time of 131.8 days (range, 3-602 d). Indwell time (< 90, 90-180, or > 180 d) did not affect retrieval (P = .4). There were 21 filters (15.1%) left in situ as a result of severe tilt (n = 9), significant thrombus in the filter (n = 5), IVC occlusion (n = 3), filter incorporation into the caval wall (n = 3), or lack of central venous access (n = 1). There was a strong relationship between penetration and caudal migration (P < .0001). Severe tilt was associated with prolonged fluoroscopic times for retrieval (P = .003). The majority of G2 filters can be removed without difficulty. The most common factor affecting retrieval was severe tilting. The indwelling time had no impact on retrieval. G2 filter-related complications were frequent but most, including fractures, were clinically insignificant.