Factors predicting failure of retrieval of inferior vena cava filters

Abstract

ObjectiveInferior vena cava (IVC) filters have been commonly used to prevent pulmonary embolism in patients with deep vein thrombosis. However, IVC filters have been associated with risks, including IVC perforation, filter migration, fracture, and thrombosis. Filter retrieval has not always been successful. Our objective was to identify the factors associated with failure of retrieval of IVC filters. MethodsThe present study was an institutional review board-approved retrospective medical record review of patients who had undergone IVC filter retrieval attempts at the Department of Vascular Surgery at Cleveland Clinic from 2011 to 2018. The patients were identified by International Classification of Diseases code query, and data were gathered regarding demographics, filter position, procedure details, and patient outcomes. Computed tomography imaging and venography was used to determine the IVC filter location before retrieval. ResultsWe identified 295 filter retrieval attempts in 294 patients. No procedural IVC ruptures, morbidity, or mortality occurred. Retrieval was successful for 249 filters (84.4%). The median filter dwell time was 196 days for the successful retrievals compared with 375 days for the failed retrieval attempts (P = .004). Penetration of the filter tines through the caval wall occurred in 291 filters (98.6%). However, the hook/apex (HA) of 31 filters (10.5%) had become embedded or had penetrated through the caval wall. The hook/apex and collar (HA+C) of 33 filters (11.2%) were embedded or had penetrated through the caval wall. The failure rate of filter retrieval with the HA embedded was 48.4% (15 of 31). The failure rate with the HA+C embedded was 66.7% (22 of 33). Finally, the failure rate for filters without these issues was 3.9% (9 of 231). The failure rate for HA and HA+C was greater than that for those without these issues (P < .001) but did not differ between the two issues (P = .14). Among those with computed tomography scans, the association of any portion of the filter with other adjacent retroperitoneal structures was not related to an increased rate of retrieval failure (P = .16). Complex retrieval methods involving endobronchial forceps, ballooning, or snaring the collar of the filter was associated with increased retrieval failure compared with simple retrieval involving snaring the hook of the filter (P < .001). The failure rates decreased over time (P = .004). Of the patients with failed retrieval attempts, 8% experienced subsequent venous thromboembolism. ConclusionsRetrieval should be attempted for all IVC filters, irrespective of the chronicity and complexity, given the procedural safety. Tine penetration was nearly ubiquitous; however, IVC filters with the HA or HA+C imbedded into or penetrating through the caval wall was a predictor of retrieval failure.