A population-based study of the effectiveness of inferior vena cava filteruse among patients with venous thromboembolism

Background: There are few population-based data regarding the effectiveness of inferior vena cava filter use in the prevention of symptomatic pulmonar y embolism. Objective: To determine the 1-year cumulative incidence of rehospitalizatio n for venous thrombosis or pulmonary embolism among patients with thromboem bolism treated with a vena cava filter compared with the incidence in a con trol population with thromboembolism. Patients and Methods: Population-based retrospective analysis of linked hos pital discharge abstracts in California. From January 1, 1991, through Dece mber 30, 1995, 3632 patients were treated with a filter and 64333 controls were admitted with a principal diagnosis of venous thromboembolism. Results: Filter-treated patients had significantly greater comorbidity, wit h a higher frequency of previous pulmonary embolism, recent major bleeding, malignant neoplasm, and stroke. Patients who initially manifested pulmonar y embolism were significantly more likely to be rehospitalized for pulmonar y embolism than patients with an initial diagnosis of venous thrombosis alo ne, among filter-treated patients (relative risk, 6.72; 95% confidence inte rval, 3.61-12.49) and controls (relative risk, 5.30; 95% confidence interva l, 4.61-6.10). Risk-adjusted proportional hazards modeling showed no signif icant difference between filter-treated patients and controls in the relati ve hazard of rehospitalization for pulmonary embolism. However, filter plac ement was associated with a significantly higher relative hazard of rehospi talization for venous thrombosis among patients who initially manifested pu lmonary embolism (relative hazard, 2.62; 95% confidence interval, 2.09-3.29 ), but not among those who presented with venous thrombosis (relative hazar d, 1.14; 95% confidence interval, 0.92-1.43). Conclusions: Insertion of a vena cava filter was not associated with a sign ificant reduction in the 1-year incidence of rehospitalization for pulmonar y embolism. Use of a filter was associated with a higher incidence of rehos pitalization for venous thrombosis, but only among patients who initially m anifested pulmonary embolism. A prospective clinical study is needed to det ermine the efficacy of filter use among patients with pulmonary embolism wh o do not meet strict guidelines for insertion of a vena cava filter.