Postmarketing Surveillance of Inferior Vena Cava Filters Among US Medicare Beneficiaries: The SAFE-IVC Study.

Importance: Inferior vena cava filters (IVCFs) are commonly used to prevent pulmonary embolism in selected clinical scenarios, despite limited evidence to support their use. Current recommendations from professional societies and the US Food and Drug Administration endorse timely IVCF retrieval when clinically feasible. Current IVCF treatment patterns and outcomes remain poorly described.

Objectives: To evaluate temporal trends and practice patterns in IVCF insertion and retrieval among older US patients and report the incidence of periprocedural and long-term safety events of indwelling and retrieved IVCFs.

Design, setting, and participants: Prespecified, retrospective, observational cohort of Medicare Fee-for-Service (FFS) beneficiaries, leveraging 100% of samples of inpatient and outpatient claims data from January 1, 2013, to December 31, 2021.

Exposure: First-time IVCF insertion while insured by Medicare FFS.

Main outcomes and measures: The primary safety outcome was the composite of all-cause death, filter-related complications (eg, fracture, embolization), operating room visits following filter-related procedures, or new diagnosis of deep vein thrombosis (DVT). Events were considered periprocedural if they occurred within 30 days of IVCF insertion or retrieval and long-term if they occurred more than 30 days after.

Results: Among 270 866 patients with IVCFs placed during the study period (mean age, 75.1 years; 52.8% female), 64.9% were inserted for first-time venous thromboembolism (VTE), 26.3% for recurrent VTE, and 8.8% for VTE prophylaxis. Of these patients, 63.3% had major bleeds or trauma within 30 days of IVCF insertion. The volume of insertions decreased from 44 680 per year in 2013 to 19 501 per year in 2021. The cumulative incidence of retrieval was 15.3% at a median of 1.2 years and 16.8% at maximum follow-up of 9.0 years. Older age, more comorbidities, and Black race were associated with a decreased likelihood of retrieval, whereas placement at a large teaching hospital was associated with an increased likelihood of retrieval. The incidence of caval thrombosis and DVT among patients with nonretrieved IVCFs was 2.2% (95% CI, 2.1%-2.3%) and 9.2% (95% CI, 9.0%-9.3%), respectively. The majority (93.5%) of retrieval attempts were successful, with low incidence of 30-day complications (mortality, 0.7% [95% CI, 0.6%-0.8%]; filter-related complications, 1.4% [95% CI, 1.2%-1.5%]).

Conclusions and relevance: In this large, US real-world analysis, IVCF insertion declined, yet retrievals remained low. Strategies to increase timely retrieval are needed, as nonretrieved IVCFs may have long-term complications.