Objective: Early detection and start of appropriate treatment are highly correlated with survival of sepsis and septic shock, but the currently available predictive tools are not sensitive enough to identify patients at risk.
Design: Linear (time and frequency domain) and nonlinear (unifractal and multiscale complexity dynamics) measures of beat-to-beat interval variability were analyzed in two mouse models of inflammatory shock to determine if they are sensitive enough to predict outcome.
Setting: University research laboratory.
Subjects: Blood pressure transmitter-implanted female C57BL/6J mice.
Interventions: IV administration of tumor necrosis factor (n = 11) or lipopolysaccharide (n = 14).
Measurements and main results: Contrary to linear indices of variability, unifractal dynamics, and absolute heart rate or blood pressure, quantification of complex beat-to-beat dynamics using multiscale entropy was able to predict survival outcome starting as early as 40 minutes after induction of inflammatory shock. Based on these results, a new and clinically relevant index of multiscale entropy was developed that scores the key features of a multiscale entropy profile. Contrary to multiscale entropy, multiscale entropy scoring can be followed as a function of time to monitor disease progression with limited loss of information.
Conclusions: Analysis of multiscale complexity of beat-to-beat dynamics at high temporal resolution has potential as a sensitive prognostic tool with translational power that can predict survival outcome in systemic inflammatory conditions such as sepsis and septic shock.