Survival of severe blunt trauma patients treated with resuscitative endovascular balloon occlusion of the aorta compared with propensity score-adjusted untreated patients

BACKGROUND: Despite a growing call for use of resuscitative endovascular balloon occlusion of the aorta (REBOA) for critically uncontrolled hemorrhagic shock, there is limited evidence of treatment efficacy. We compared the mortality between patients who received a REBOA with those who did not, adjusting for the likelihood of treatment and injury severity, to measure efficacy. METHODS: We analyzed observational prospective data from the Japan Trauma Data Bank (2004-2011) to compare the mortality between adult patients who received a REBOA with those who did not. To adjust for potential treatment bias, we calculated the likelihood of REBOA treatment via a propensity score (PS) using available pretreatment variables (vital signs, age, sex, as well as anatomic and physiologic injury severity) and matched treated patients to up to five similar PS untreated patients. We compared survival to discharge between treated and untreated groups using conditional logistic regression and Cox proportional hazards regression. RESULTS: Of 45,153 patients who met inclusion, 452 patients (1.0%) received REBOA placement. These patients were seriously injured (median Injury Severity Score [ISS], 35) and had high mortality (76%). Patients who did not receive a REBOA had significantly lower injury severity (median ISS, 13; p < 0.0001) and lower mortality (16%). After matching REBOA patients with controls with similar PSs for treatment, the crude conditional odds ratio of survival by REBOA treatment was 0.30 (95% confidence interval, 0.23-0.40). CONCLUSION: REBOA treatment is associated with higher mortality compared with similarly ill trauma patients who did not receive a REBOA. The higher observed mortality among REBOA-treated patients may signal last ditch efforts for severity not otherwise identified in the trauma registry. Copyright (C) 2015 Wolters Kluwer Health, Inc. All rights reserved.