Risk factors for deep vein thrombosis and pulmonary embolism after traumatic injury: A competing risks analysis.

BACKGROUND: Venous thromboembolism, including deep vein thrombosis (DVT) and pulmonary embolism (PE), is typically reported as a composite measure of the quality of trauma center care. Given that recent data suggesting postinjury DVT and PE are distinct clinical processes, a better understanding may result from analyzing them as independent, competing events. Using competing risks analysis, we evaluated our hypothesis that the risk factors and timing of postinjury DVT and PE are different. METHODS: We examined all adult trauma patients admitted to our Level I trauma center from July 2006 to December 2011 who received at least one surveillance duplex ultrasound of the lower extremities and who were at high risk or greater for DVT. Outcomes included DVT and PE events, and time-to-event from admission. We used competing risks analysis to evaluate risk factors for DVT while accounting for PE as a competing event, and vice versa. RESULTS: Of 2,370 patients, 265 (11.2%) had at least one venous thromboembolism event, 235 DVT only, 19 PE only, 11 DVT and PE. Within 2 days of admission, 38% of DVT cases had occurred compared with 26% of PE. Competing risks modeling of DVT as primary event identified older age, severe injury (Injury Severity Score, ≥ 15), mechanical ventilation longer than 4 days, active cancer, history of DVT or PE, major venous repair, male sex, and prophylactic enoxaparin and prophylactic heparin as associated risk factors. Modeling of PE as the primary event showed younger age, nonsevere injury (Injury Severity Score, < 15), central line placement, and prophylactic heparin as relevant factors. CONCLUSION: The risk factors for PE and DVT after injury were different, suggesting that they are clinically distinct events that merit independent consideration. Many DVT events occurred early despite prophylaxis, bringing into question the preventability of postinjury DVT. We recommend trauma center quality reporting program measures be revised to account for DVT and PE as unique events. LEVEL OF EVIDENCE: Epidemiologic, level III.

The differential associations of preexisting conditions with trauma-related outcomes in the presence of competing risks.

INTRODUCTION: Pre-existing chronic conditions (PECs) pose a unique problem for the care of aging trauma populations. However, the relationships between specific conditions and outcomes after injury are relatively unknown. Evaluation of trauma patients is further complicated by their discharge to care facilities, where mortality risk remains high. Traditional approaches for evaluating in-hospital mortality do not account for the discharge of at-risk patients, which constitutes a competing risk event to death. The objective of this study was to evaluate associations between 40 PECs and two clinical outcomes in the context of competing risks among older trauma patients. METHODS: This retrospective study evaluated blunt-injured patients aged 55 years and older admitted to a level I trauma centre in 2006-2012. Outcomes were hospital length of stay (HLOS) and in-hospital mortality. Survivors were classified as discharges home or discharges to care facilities. Competing risks regression was used to evaluate each PEC with in-hospital mortality, accounting for discharges to care facilities as competing events. Competing risk estimates were compared to Cox model estimates, for which all survivors to discharge were non-events. Analyses were stratified using injury-based mortality risk at a 50% cutpoint (high versus low). RESULTS: Among 4653 patients, 176 died in-hospital, 3059 were discharged home, and 1418 were discharged to a care facility. Most patients (98%) were classified with a low mortality risk. Only haemophilia and coagulopathy were consistently associated with longer HLOS. In the low-risk subgroup, in-hospital mortality was most strongly associated with liver diseases, haemophilia, and coagulopathy. In the high-risk group, Parkinson's disease, depression, and cancers showed the strongest associations. Accounting for the competing event altered estimates for 12 of 19 significant conditions. CONCLUSIONS: Excess mortality among patients expected to survive their injuries may be attributable to complications resulting from PECs. Discharges to care facilities constitute a bias in the evaluation of in-hospital mortality and should be considered for the accurate calculation of risk. In conjunction with injury measures, consideration of PECs provides physicians with a foundation to plan clinical decisions in older trauma patients.