Prehospital loss of R-to-R interval complexity is associated with mortality in trauma patients

Background: To improve our ability to identify physiologic deterioration caused by critical injury, we applied nonlinear analysis to the R-to-R interval (RRI) of the electrocardiogram of prehospital trauma patients. Methods: Ectopy-free, 800-beat sections of electrocardiogram from 31 patients were identified. Twenty patients survived (S) and 11 died (NonS) after hospital admission. Demographic data, heart rate, blood pressure, field Glasgow Coma Scale (GCS) score, and survival times were recorded. RRI complexity was assessed via nonlinear statistics, which quantify entropy or fractal properties. Results: Age and field heart rate and blood pressure were not different between groups. Mean survival time (NonS) was 129 hours +/- 62 hours. NonS had a lower GCS score (8.6 +/- 1.7 vs. 13.2 +/- 0.8, p < 0.05). RRI approximate entropy (ApEn; 0.87 +/- 0.06 vs. 1.09 +/- 0.07, p < 0.01), sample entropy (SampEn; 0.80 +/- 0.08 vs. 1.10 +/- 0.05, p < 0.01) and fractal dimension by dispersion analysis (1.08 +/- 0.02 vs. 1.13 +/- 0.01, p < 0.05) were lower in NonS. Distribution of symbol 2 (Dis_2), a symbol-dynamics measure of RRI distribution, was higher in NonS (292.6 +/- 34.4 vs. 222 +/- 21.3, p < 0.10). For RRI data, logistic regression analysis revealed ApEn and Dis_2 as independent predictors of mortality (area under the receiver-operating characteristic curve = 0.96). When GCS(MOTOR) was considered, it replaced Dis_2 whereas ApEn was retained (area under curve = 0.92). When Injury Severity Score was considered, it replaced GCS(MOTOR); ApEn was retained. Conclusions: Prehospital loss of RRI complexity, as evidenced by decreased entropy, was associated with mortality in trauma patients independent of GCS score or Injury Severity Score.