Background: Despite growing evidence of complications, inferior vena cava filters are being used increasingly in patients with venous thromboembolism based on limited efficacy data. In such a controversial situation, the effectiveness of filter use on mortality as an adjuvant to antithrombotic therapy for pulmonary embolism remains uncertain.
Methods: Using the Diagnosis Procedure Combination database in Japan, we identified patients hospitalized with pulmonary embolism who received anticoagulation or thrombolytic therapy from the day of admission. We compared the in-hospital mortality between patients who received a filter and patients who did not, using propensity score and instrumental variable methods.
Results: Of 13,125 eligible patients, 3948 received a filter, and 9177 did not receive a filter. The propensity score-matched analysis showed that filter use was significantly associated with lower in-hospital mortality than nonuse (2.6% vs 4.7%, P < .001; risk ratio 0.55; 95% confidence interval [CI], 0.43-0.71; risk difference -2.1%; 95% CI, -3.0% to -1.2%; number needed to treat, 48; 95% CI, 34-82). We obtained similar results in the inverse probability of treatment-weighting analysis. The instrumental variable analysis confirmed that filter use was associated with a decreased risk of in-hospital mortality with adjustment for all measured variables (risk difference -2.5%, 95% CI, -4.6% to -0.4%).
Conclusions: This study suggested that filter use was potentially effective for reducing in-hospital mortality in patients with pulmonary embolism. Prospective studies are needed to confirm the effectiveness observed in our results and to define which subpopulations of patients would benefit most from filters.
Keywords: Deep vein thrombosis; In-hospital mortality; Inferior vena cava filter; Pulmonary embolism; Venous thromboembolism.
Copyright © 2015 Elsevier Inc. All rights reserved.