Study based on “real world” US database sheds light on inferior vena cava perforations by filters

Emily A Wood

Emily A Wood, Division of Vascular Surgery, Stony Brook University Medical Center, USA, spoke at the European Venous Forum Annual Meeting (28–30 June, Florence, Italy) on the results of a study that set out to quantify the incidence of inferior vena cava filter perforations of the inferior vena cava and surrounding structures and to assess the outcomes of filter perforations. The retrospective study found that the most common filter involved in perforating the inferior vena cava was the Bard G2. This was followed by the Bard Recovery and Cook Celect filters.

Wood clarified that the study was based on the US FDA MAUDE (Manufacturer and User Facility Device Experience) database which was not universal and voluntary— this meant that it possibly did not entirely represent the complications that occurred during these years. She told delegates that inferior vena cava filter placement has more than doubled over the last decade (30,756 to 65,041) and that while uncomplicated filter placement is a relatively straightforward ambulatory procedure, the setting up of inferior vena cava filter registries and reporting outcomes are not mandatory.


Perforation of the inferior vena cava and its surrounding structures by filter struts is a known complication. “Our purpose was to develop a sense of how many inferior vena cava perforations develop each year in the United States and then to understand how this complication, once it is realised, is managed by the physician,” Wood told delegates.

The impact of inferior vena cava perforation by filters was gauged based on a “real world” open database provided by users, facilities and manufacturers.

The investigators reviewed 3,311 adverse events from inferior vena cava filters reported in the US FDA MAUDE database from January 2000 to June 2011. Outcomes of interest included clinical presentation, inferior vena cava perforation, type of device utilised and management (including retrievability rates). Wood clarified that the database was not “all encompassing” as it was not mandatory, and that it was not international, since data were from the United States alone.

Wood clarified that the team excluded adverse event descriptions that did not specifically involve inferior vena cava perforation such as device malfunction, failure to deploy device properly, migration/tilting and cardiopulmonary system involvement. This was done to ease data collection as migration was difficult to define. “Perforation of inferior vena cava is an easily quantifiable event as it either occurred or not,” she said.


Wood told delegates that 391 (11.1%) adverse events of inferior vena cava perforation were reported. A five-fold increase in the number of adverse events related to inferior vena cava filters has been noted since 2004; but the accrual numbers of perforation have not significantly changed over the years. A five-fold increase in adverse events related to inferior vena cava perforation occurred from 2003 to 2004. The annual distribution was 35 cases (11.8%) which ranged from seven (5.6%) to 70 (15.7%).

The most common filter involved in perforating the inferior vena cava was the Bard G2 (168 cases), followed by the Bard Recovery (59 cases) and Cook Celect in 46 cases. Trapease, Greenfield, Eclipse, Gianturco-Roehm were some of the other filters which were involved in perforations.

“An important thing to note about inferior vena cava perforations is that a large number of the patients are asymptomatic. Our study noted that in 171 (47%) patients, the vein wall perforation was an incidental finding,” said Wood. Surrounding organ involvement was found in 121 (33%) cases with the aorta involved in 40 (33%) and the duodenum in 26 (21%) cases. “With regard to retrieval, there was 268 adverse events reported that were confined to perforation of the wall alone; of these 28% (n=76) were retrieved by endovascular means and 7% (n=20) were retrieved by open surgery. Fifty five per cent (n=148) were not retrieved and retrieval was unspecified in 9% (n=24) of cases. When perforation involved the surrounding structures, equal numbers of endovascular and open retrieval procedures was carried out. Neither major bleeding requiring further intervention nor mortality was reported secondary to filter retrieval.

“Inferior vena cava filter placement is not free from complications and the indications for filter placement must be carefully reviewed prior to any procedure. Our study has shown that retrieval filters are commonly involved in inferior vena cava perforation and appropriately-timed filter retrieval is critical and should be performed whenever possible in order to decrease the number of filter-related complications,” said Wood. She also made the point that inferior vena cava perforations by filters remain stable over the past decade despite increasing numbers of adverse events reported. “The majority of filters involved in perforation were retrievable and had a multiprong design for better attachment to the vein wall,” she said.