Abstract
The evidence supporting the use of Resuscitative Endovascular Balloon Occlusion of the Aorta (REBOA) in severely injured patients is still debatable. Using the ABOTrauma Registry, we aimed to define factors affecting mortality in trauma REBOA patients. Data from the ABOTrauma Registry collected between 2014 and 2020 from 22 centers in 13 countries globally were analysed. Of 189 patients, 93 died (49%) and 96 survived (51%). The demographic, clinical, REBOA criteria, and laboratory variables of these two groups were compared using non-parametric methods. Significant factors were then entered into a backward logistic regression model. The univariate analysis showed numerous significant factors that predicted death including mechanism of injury, ongoing cardiopulmonary resuscitation, GCS, dilated pupils, systolic blood pressure, SPO2, ISS, serum lactate level and Revised Injury Severity Classification (RISCII). RISCII was the only significant factor in the backward logistic regression model (p < 0.0001). The odds of survival increased by 4% for each increase of 1% in the RISCII. The best RISCII that predicted 30-day survival in the REBOA treated patients was 53.7%, having a sensitivity of 82.3%, specificity of 64.5%, positive predictive value of 70.5%, negative predictive value of 77.9%, and usefulness index of 0.385. Although there are multiple significant factors shown in the univariate analysis, the only factor that predicted 30-day mortality in REBOA trauma patients in a logistic regression model was RISCII. Our results clearly demonstrate that single variables may not do well in predicting mortality in severe trauma patients and that a complex score such as the RISC II is needed. Although a complex score may be useful for benchmarking, its clinical utility can be hindered by its complexity.