Abstract 4: Physiological Response Space: Mapping Resuscitation Response in a Swine Model of Traumatic Shock
Introduction: Efficacy of resuscitation fluids is traditionally assessed by multiple physiological responses studied independently. However responses differ between systems and subjects, resulting in unpredicted synergisms. To identify these synergisms, we propose the concept of a physiological response space (PRS) to represent the complexity of response to resuscitation. PRS can be visualized as a multi-dimensional polygon, with dimensions defined by clinically relevant physiological variables, and size and shape determined by the magnitude of response of each variable. We hypothesized that PRS better discriminates treatment effects compared to univariate analyses in a swine model of severe shock-trauma and resuscitation.
Methods: Male swine (n=28) were hemorrhaged to a target oxygen debt O2D (80 ml/kg) and then resuscitated over 20 min with a 500 mL bolus of Hetastarch (HTS), hypertonic saline (HS), a combination hemoglobin-based oxygen carrier (HBOC)-HS solution (AfterShock, AS), or whole blood (WB). Response variables (VO2, MAP, cardiac output CO, O2 delivery DO2, lactate) were assessed every 30 min after fluid bolus for 3 hr. Data were analyzed by traditional repeated-measures mixed model ANOVA on each variable; PRS was modeled by quadratic regression and canonical discriminant analysis, and visualized by color-coded contour plots.
Results: Except for MAP, there were no statistical differences between treatments (p > 0.3). However, PRS plots discriminated among fluid treatments as early as 30 min post-resuscitation. PRS revealed that the apparent early MAP advantage of WB and HTS was modulated by increasing O2D and lactate. Repayment of O2D and CO support were similar in AS and HS groups, but a significant and sustained MAP increase was displayed by AS, whereas HS showed modest delayed increases in MAP coupled with a large VO2 response.
Conclusions: PRS provided better discrimination of fluid resuscitation treatments compared to individual outcome analysis. Analysis of isolated physiological components failed to inform of important synergisms that were quickly apparent with PRS analysis. Multivariate PRS analysis has great potential for rapid assessment of physiological state in a clinical setting.
- © 2010 by American Heart Association, Inc.