Abstract
Background: Our group previously used a canine haemorrhagic shock model that removed a set volume of blood over a short period of time before comparing resuscitation interventions. We found wide variability in the degree of shock achieved. Removal of blood guided by blood pressure may be more synergistic with the individual’s ability to compensate and reduce variability across a group. Objectives: 1) To assess if a pressure-guided canine shock model (PGM) created more severe shock, as measured by a higher oxygen extraction ratio index (0ER), compared to a volume-guided model (VGM). 2) To determine if the PGM provided less variability in shock parameters. 3) To identify parameters contributing to 0ER variance in the PGM. Methods: The VGM model included 12 anaesthetised greyhound dogs whereby removal of 48 mL/kg of blood was targeted over 30 minutes. The PGM model included 24 anaesthetised greyhound dogs whereby blood was removed to maintain a mean arterial blood pressure (MAP) < 60 mmHg for 60 minutes. Cardiovascular parameters were measured at baseline (T0) and at the end of blood removal (T1). Linear regression was used to compare means. Variance of each parameter was compared at T1 using Levene’s test. Regression modelling was used to find the best explanatory variables for the variance in volume of blood removed (VOL)(variables at T0) and T1 OER (variables at T1) in the PGM. Significance was set at P < 0.05. Results: There was no significant difference in VOL between the PGM (mean 48.8 mL/kg, 95% confidence interval 44.9–52.7) and VGM (47.9, CI 47.5–48.3). At T1, the PGM had a significantly lower mean OER (0.45, CI 0.39–0.50) compared to the VGM (0.60, CI 0.49– 0.70). The PGM also had a significantly higher mean heart rate, higher mean MAP, higher mean systemic vascular resistance index (SVR), lower mean oxygen consumption, higher mean lactate and lower mean base excess. There was no significant difference in the variance of OER, or other parameters, between the two models. Stroke volume index (SV) explained most of the variance for both VOL and OER. For predicting VOL, a model including T0 variables SV, MAP and central venous pressure provided the highest R2 (0.45) while minimising C(p)(2.0). For predicting OER, the best model included T1 variables SV, VOL and SVR (R2 0.29, C(p) 0.59). Conclusion: Although there was no difference in mean VOL, the PGM showed evidence of better compensation, including lower OER and other parameters supportive of a sympathetic response. This may be due to the pressure-guided strategy or longer shock period. The change in model to a PGM did not reduce the variability of shock parameters. Individual SV best explained the variation in response to haemorrhage.