Multiscale Complexity Analysis for Vasopressor Treatment Effect Among Sepsis Patients

Rationale In complex biological systems, fluctuations occur over a wide range of time scales, from milliseconds to hours. Multiscale Entropy (MSE) quantifies the complexity of these signals across multiple time scales, offering valuable insights that standard entropy measures, often used for heart rate variability, may miss. Sepsis, characterized by multiscale disruptions in physiological processes, presents a compelling context for MSE application. However, its potential to evaluate treatment efficacy remains underexplored. This study is the first to investigate the impact of vasopressor therapy and concurrent interventions on MSE physiomarkers, accounting for sepsis severity and polytherapy effects, to elucidate treatment-specific signatures and inform personalized therapeutic strategies.Methods This retrospective cohort study analyzed intensive care unit patients who met Sepsis-3 criteria and were admitted to Emory University Hospital between January 2016 and December 2018. Two groups, those receiving norepinephrine and those not treated with vasopressors, were compared using propensity score matching based on age, gender, Acute Physiologic Assessment and Chronic Health Evaluation (APACHE) II and Sequential Organ Failure Assessment (SOFA) score within the first 24 hours of admission. The norepinephrine cohort was subdivided for comparison based on the use of single adjunct therapies and second-line treatments, specifically vasopressin and phenylephrine. Continuous electrocardiograms (ECG) recorded from bedside monitors were utilized to extract MSE features. We investigated the impact of single and combined vasopressor therapies on MSE using a multifaceted analytical approach, comprising: (1) two-sample t-tests to compare treatment effects, (2) ElasticNet modeling to elucidate feature importance, and (3) odds ratio analysis to quantify treatment efficacyResultsA total of 979 patients were included in the study (norepinephrine group: n=445; no vasopressor group: n=534). The average age was 59.9 years, and the mean SOFA was 2.5. Significant (p<0.001) MSE scales for mortality prediction, were identified through the Elastic Net method. Odds ratio calculations were performed for the same scales. The MSE mortality analysis under vasopressor treatment revealed that Scales 9 and 13 had odds ratios above 1, suggesting a potential association with higher mortality. Conversely, Scales 2, 4, 6, 10, 15, 16, 17, and 20 had odds ratios below 1, indicating a possible association with improved survival. These findings underscore the complex role of specific MSE scales in influencing mortality outcomes during vasopressor therapy.Conclusion The findings of this study indicate that medications and interventions may affect the interpretation of MSE markers, underscoring the need for further research on MSE in the context of sepsis progression and clinical decision-making.

Duke Scholars

Author Rishi Kamaleswaran Trauma, Acute, and Critical Care Surgery