Use of inferior vena cava filters in the Medicare population

To examine the use of inferior vena cava (IVC) filters, we performed a population-based study using a 5% random sample of the United States Medicare population (1.25 million persons). Filter placement and its timing relative to diagnosis of venous thromboembolism (VTE) were determined using both hospital and physician Medicare billing codes after detailed review of large samples of complete individual claims records. Comorbid conditions and mortality were also noted. From July 1986 through June 1989, a total of 806 patients in the sample population received IVC filters. Mortality rates after filter placement were high: 16% during initial hospitalization, 32% within 6 months of filter placement, and 48% after 2 years. However, only 3 (1%) of 423 patients who underwent filter placement without the diagnosis of pulmonary embolism (PE) suffered PE within 1 year of filter placement. Use of IVC filters increased significantly over the study period (p < 0.001). Rates of filter placement among the nine census divisions of the United States differed significantly (p < 0.001). An East-West trend was notable with the highest rate (37 filters/100,000 enrollees/yr) in New England, as compared with 14 filters/100,000 enrollees/yr in the Pacific Division. No explanation for these regional differences was evident after patient demographics and comorbidity were examined. Vena cava filters prevent PE for at least 1 year after placement. The frequency of IVC filter use in elderly patients is increasing and varies substantially by region. Although reasons for these trends are unclear, these findings suggest the need for uniform practice guidelines for IVC filter placement.