Pulmonary embolism in Veterans Affairs Medical Centers: is vena cava interruption underutilized?

Veterans with venous thrombosis or pulmonary embolism (PE) were evaluated using Veterans Affairs patient treatment file (PTF) data from fiscal years 1990-1995, inclusive, to define outcomes for those with PE. The specific aims of the study were to define how often those with PE underwent vena cava interruption (VCI) and whether VCI affected in-hospital mortality rates. Outcomes were defined using PTF data and Patient Management Category (PMC) software for 26,132 veterans discharged from all Veterans Affairs Medical Centers (VAMCs) with venous thromboembolism, which included a subset of 4,882 patients identified by both PTF data and PMC software to have PE. PMC software also generated measures of illness severity, patient complexity (PMC count), and resource utilization (called resource intensity scale) for each hospital admission. The in-hospital mortality rate for those with PE was 15.9 per cent (775 of 4882). Only 157 VCIs were performed in those with PE which constituted 3.2 per cent of the latter group. Those with PE who had VCI experienced a 13.4 per cent unadjusted in-hospital mortality rate (21 of 157) versus a 16 per cent unadjusted mortality rate without VCI (754/4725; not significant). In a logistic regression model of in-hospital mortality in those with PE, increasing age and illness severity were directly related to mortality, whereas VCI was independently associated with reduced mortality. The odds of death were reduced by 0.482 (0.287-0.807, 95% limits) for patients with PE who underwent VCI (P<0.005). Utilization of VCI varied among VAMCs: the hospital rates that VCI were performed in those with PE ranged from 0 to 16.7 per cent. Mortality associated with PE was substantial in VAMCs, and VCI was independently associated with reduced in-hospital mortality. The low percentage of veterans with pulmonary embolism who underwent VCI was surprising. VCI may be underutilized in veterans with PE.