Characteristics of Inferior Vena Cava Filter Placement, Management, and Retrieval at a Multi-Center Institution

Abstract

Background: Retrievable inferior vena cava filters (IVCF) were developed to address the adverse events associated with permanent IVCF. However, this benefit cannot be fully appreciated if filters are not removed in a timely manner and when appropriate. Although there are clinical scenarios when permanent placement is desired, studies show that these situations are far outweighed by scenarios of medically inappropriate non-retrieval. There is substantial evidence describing adverse events related to IVCF such as filter thrombosis, DVT, fracture, perforation, and migration. Furthermore, evidence supporting the benefit of IVCF to prevent the development or progression of pulmonary embolism remains controversial. Therefore, indications for IVCF placement are not standardized at most centers. Specialty societies, such as radiology and cardiology, have differing recommendations for filter placement. IVCF placement is not uncommon (of 130,643 patients hospitalized at 263 hospitals with acute venous thromboembolism [VTE], 15% had IVCF placed) and retrieval rates of non-permanent filters remain low (18-41%). Methods: This study assesses IVCF indications, removal rate, complications, and practice patterns at a multi-center institution via retrospective chart review. Procedure codes and associated diagnosis codes were used to identify 626 adult patients who had IVCF placed over a span of 5 years by either interventional radiology, vascular surgery, or interventional cardiology. Of note, patients were excluded from analysis if they had a documented death within 60 days of filter placement. Descriptive statistics, chi-squared tests, and Wilcoxon rank-sum tests were used for statistical analysis. All IVCF evaluated were retrievable filters. Results: Most IVCF were placed by interventional radiology (77%) as expected. Retrieval rate was 27%. No significant difference was found in retrieval rates among the procedural specialties. Also, retrieval rate was not significantly different between specialty groups that followed the patient after filter placement, including radiology, vascular surgery, or hematology (p value: 0.5382). Overall, 58% of IVCF were unmanaged after placement and, as expected, this was associated with a higher rate of non-retrieval (p value: 0.0001). There was no significant difference in retrieval rate among inpatient or outpatient procedures (p value: 0.2067). Patients in the retrieval group had a lower Charlson Comorbidity Index score than those in the non-retrieval group (p value: 0.0001). The most common indications for placement were acute VTE with contraindication to anticoagulation (40%) and anticoagulation failure (21%). The most common filter complication was filter thrombus (2%). Primary care providers were the most common ordering provider group that requested filter placement (51% of filter placement requests). Interestingly, patients with Medicare/Medicaid had a lower rate of retrieval than patients with any other coverage, including no coverage (20% versus 35%, p value <0.0001). Recurrent VTE occurred in 44 patients (not including filter thrombus) and 9 of those events occurred after filter removal. 62% of patients received anticoagulation therapy during or after filter placement. Conclusion: Lack of follow-up and filter non-removal continues to remain a problem since the emergence and use of IVCF. Surprisingly, our study didn't show that retrieval rates were different based on inpatient or outpatient status, or by follow-up specialty. We did not expect to find that retrieval rates would be lowest in patients with Medicare/Medicaid, or that the most frequently ordering provider group would be primary care providers (over pulmonologists or hematologists, for example). Even after excluding patient deaths within 60 days of filter placement, we still found retrieval rates to be low. As a retrospective study, we can only show association and there may be unique characteristics of our centers that affect external validity. Also, in some clinical scenarios permanent filter placement may be indicated. Regardless, this analysis challenges assumptions regarding the factors involved with filter placement and persistently low rates of retrieval. It also proposes new targetable areas for collaboration and improvement. Disclosures No relevant conflicts of interest to declare.