Real-Time Calculation of System-Level Complexity During Trauma/Hemorrhage: Can We Do It?

Abstract : Objectives: We previously showed that sample entropy (SampEn), and other nonlinear measures of the complexity of the ECG time series, decrease in response to hypovolemia and/or injury. These measures characterize only one signal, and thus refer to signal-level complexity. In contrast, system-level complexity quantifies the amount of interaction among the signals arising from all available sensors attached to a patient (ECG, blood pressure, oxygen saturation, etc.). We found that system-level complexity (OSC) calculated with OntoSpace software (Ontonix S.r.l., Como, Italy) fluctuates during critical events such as asphyxia. Relatedly, system robustness (OSR) represents the margin between current OSC and maximal or minimal possible OSC. OSR thus is greatest when OSC is in the middle range; it decreases when OSC approaches either extreme, where the system becomes unstable and is prone to crash. Here, we present data calculated in real time from both signal-level and system-level complexity in a model of acute respiratory distress syndrome (ARDS) due to trauma, hemorrhagic shock and resuscitation. Methods: Nine swine were anesthetized with ketamine and midazolam and underwent baseline measurements (BL), right-chest pulmonary contusion (PC), hemorrhage of 12 mL/kg (Bleed), resuscitation with lactated Ringer s (LR), transfusion of shed blood (Tx), and post-resuscitation observation (Post-Resus). Data were collected continuously and analyzed in 15-min datasets. We calculated heart rate (HR, bpm); mean arterial pressure (MAP, mmHg); PaO2-to-FiO2 ratio (PFR); ECG SampEn (unitless); ECG multiscale entropy (MSE, unitless), and percentage of normal-to-normal RRIs differing by more than 50 ms (pNN50). We calculated OSC and OSR (both unitless) from 56 different channels of single-sensor data. Results: see table. Means SEMs are reported. Statistics: one way ANOVA with Tukey s adjustment.