Inability to easily predict individual deaths in a hemorrhage and resuscitation model

Abstract

While working on a hemorrhage with low-pressure stabilization protocol, we examined the possibility of identifying individual animals’ short term outcomes prior to occurrence using the hemorrhage patterns of variables currently suggested to identify patients and animals in severe hemorrhagic shock. Methods: 8 anesthetized, ventilated dogs were hemorrhaged (H, 90min mean arterial pressure [MAP]=35-40mmHg or MAP<30mmHg>10min or MAP<25mmHg>1min), hypotensively stabilized (S, 120min enalaprilat 0.01mg/kg/hr + hemoglobin based oxygen carrier [HBOC] or 7.8% hypertonic saline dextran 70 [HSD] for MAP=40-45mmHg), resuscitated (60min lactated Ringer’s for MAP=75-80mmHg), monitored (60min no fluids), and euthanized. Results: 2 non-survivors: NS1 (1 of 5 HBOC, died 117min into S) and NS2 (1 of 3 HSD, died 9min into S). Patterns of continuous, non-invasive variables: Non-survivors had the highest and 2nd lowest heart rates. All animals’ mean arterial pressures and end-tidal CO2s (related to cardiac output) were intermingled. Foreleg muscle StO2 was highest in NS1 and intermingled with the 4 lowest survivors in NS2. Patterns of discontinuation and/or invasive variables: Shed blood in NS1, NS2, and 2 survivors were equivalent (38-40mL/kg) and less than the 4 other survivors (45-56mL/kg). SvO2 values were intermingled. Start H base excess values were intermingled. End H base excesses were NS1 = -18.8, NS2 = -19.2, 2 survivors each = -19.0, other survivors –13.9 to –17.5 mmol/L. Conclusions: Hemorrhage values and obvious patterns of the monitored variables failed to distinguish individual animals that would undergo cardiovascular collapse despite receiving the same treatment from those that would survive.