Objective: Retrievable inferior vena cava (IVC) filters were first approved for use in the United States in 2003 to address the long-term complications of migration, thrombosis, fracture, and perforation observed with permanent IVC filter implantation. Although Food and Drug Administration approval of retrievable IVC filters includes permanent implantation, the incidence of complications from long-term implantation appears to be greater than that reported with existing permanent IVC filters. Also, only a small fraction of such retrievable IVC filters are ever retrieved. The purpose of the present study was to determine the threshold retrieval rate at which the use of retrievable IVC filters could be justified.
Methods: A Markov decision tree was constructed comparing retrievable and permanent IVC filters regarding their effectiveness and cost. A review of the reported data provided outcome probabilities, and the Tufts Medical Center Cost-Effectiveness Analysis Registry was the source of the utility values for the various potential outcomes. Medicare reimbursement rates served as a proxy for costs. A sensitivity analysis was performed for various parameters, primarily to determine the retrieval rate threshold at which the use of retrievable IVC filters would outperform the use of permanent IVC filters.
Results: Base case analysis demonstrated a greater predicted effectiveness for permanent compared with retrievable IVC filter implantation (5.41 quality-adjusted life-years [QALY] vs 5.33 QALY) at a lower cost ($2070 vs $4650). Monte Carlo simulation at 10,000 iterations confirmed the expected utility (5.4 ± 3.0 QALY vs 5.3 ± 3.0 QALY; P = .0002) and cost ($1900 ± $7400 vs $4800 ± 9900; P < .0001) to be statistically superior for permanent IVC filters. A sensitivity analysis for the filter retrieval rate demonstrated that the strategy of using a retrievable IVC filter was never preferable for utility or cost. The superiority of permanent IVC filter placement for effectiveness and cost persisted, regardless of anticipated patient-predicted annual mortality. A two-way sensitivity analysis for both IVC filter removal rate and annual patient mortality confirmed the superiority of permanent IVC filter placement at all levels.
Conclusions: The predicted effectiveness of permanent IVC filters was greater and the predicted cost lower than those for retrievable IVC filters, regardless of the IVC filter retrieval rate. This interpretation of existing reported data using Markov decision analysis modeling supports the argument that unless the long-term complication rate of retrievable IVC filters can be significantly improved, their use should be abandoned in favor of currently available permanent IVC filters.
Keywords: Decision analysis; Decision tree; Vena cava filter.
Published by Elsevier Inc.