Real-time sample entropy predicts life-saving interventions after the Boston Marathon bombing

Abstract

Purpose: Identifying patients in need of a life-saving intervention (LSI) during a mass casualty event is a priority. We hypothesized that real-time, instantaneous sample entropy (SampEn) could predict the need for LSI in the Boston Marathon bombing victims. Materials and methods: Severely injured Boston Marathon bombing victims (n = 10) had sample entropy (SampEn) recorded upon presentation using a continuous 200-beat rolling average in real time. Treating clinicians were blinded to real-time results. The correlation between SampEn, injury severity, number, and type of LSI was examined. Results: Victims were males (60%) with a mean age of 39.1 years. Injuries involved lower extremities (50.0%), head and neck (24.2%), or upper extremities (9.7%). Sample entropy negatively correlated with Injury Severity Score ( r = 0.70; P = .023), number of injuries ( r = 0.70; P = .026), and the number and need for LSI ( r = 0.82; P = .004). Sample entropy was reduced under a variety of conditions.SampEn (mean ± SD)PAmputation, n = 50.7 ± 0.3No amputation, n = 51.9 ± 0.8.027Transfusion, n = 50.7 ± 0.3No transfusion, n = 51.9 ± 0.8.027Intubation, n = 60.8 ± 0.3No intubation, n = 42.1 ± 0.7.027Vasopressors, n = 70.8 ± 0.3No vasopressors, n = 32.4 ± 0.3.004. Conclusions: Sample entropy strongly correlates with injury severity and predicts LSI after blast injuries sustained in the Boston Marathon bombings. Sample entropy may be a useful triage tool after blast injury. © 2013 Elsevier Inc.