Trends in vena caval interruption.

OBJECTIVE

This study examined the national use of vena cava filters (VCFs) from 1998 to 2005.

METHODS

Methods for complex surveys were used to examine hospital discharge data from the Nationwide Inpatient Sample (NIS) to determine the use of VCFs for the years 1998 to 2005. VCF placement in the absence of deep venous thrombosis (DVT) or pulmonary embolus (PE) was categorized as prophylactic.

RESULTS

During the study period, the estimated rate of hospitalizations per year with a diagnosis of DVT (odds ratio [OR], 1.025; 95% confidence interval [CI], 1.019-1.032; P < .01) or PE (OR, 1.076; 95% CI, 1.069-1.083; P < .01) rose significantly. The estimated weighted frequency of VCF placement increased from 52,860 procedures in 1998 to 104,114 procedures in 2005 (0.15% and 0.27% of all discharges, respectively), representing an 80% increase. VCF placement significantly increased during hospitalizations with any diagnosis of DVT or PE, or both, and no DVT or PE (P < .01 for each). Logistic regression models revealed that the rate of prophylactic VCF placement increased at a significantly higher rate than VCF placement associated with DVT or PE (157% vs 42%; P < .01), after adjusting for age, gender, and hospital characteristics. Prophylactic VCF placement in the setting of morbid obesity (P < .01) and head injury (P = .03) rose significantly over time.

CONCLUSIONS

From 1998 to 2005, the estimated rates of prophylactic VCF placement increased at a significantly higher rate than VCF placement in the setting of DVT or PE. Significant increases in the use of prophylactic VCFs were seen in the setting of morbid obesity and head injury.