The study tests the hypothesis that noninvasive cardiac output monitoring based upon bioreactance (Cheetah Medical, Portland, OR) has acceptable agreement with intermittent bolus thermodilution over a wide range of cardiac output in an adult porcine model of hemorrhagic shock and resuscitation.Design:
Prospective laboratory animal investigation.Setting:
Preclinical university laboratory.Subjects:
Eight ~ 50 kg Yorkshire swine with a femoral artery catheter for blood pressure measurement and a pulmonary artery catheter for bolus thermodilution.Interventions:
With the pigs anesthetized and mechanically ventilated, 40 mL/kg of blood was removed yielding marked hypotension and a rise in plasma lactate. After 60 minutes, pigs were resuscitated with shed blood and crystalloid. Noninvasive cardiac output monitoring and intermittent thermodilution cardiac output were simultaneously measured at nine time points spanning baseline, hemorrhage, and resuscitation.Measurements and Main Results:
Simultaneous noninvasive cardiac output monitoring and thermodilution measurements of cardiac output were compared by Bland-Altman analysis. A plot was constructed using the difference of each paired measurement expressed as a percentage of the mean of the pair plotted against the mean of the pair. Percent bias was used to scale the differences in the measurements for the magnitude of the cardiac output. Method concordance was assessed from a four-quadrant plot with a 15% zone of exclusion. Overall, noninvasive cardiac output monitoring percent bias was 1.47% (95% CI, −2.5 to 5.4) with limits of agreement of upper equal to 33.4% (95% CI, 26.5−40.2) and lower equal to −30.4% (95% CI, −37.3 to −23.6). Trending analysis demonstrated a 97% concordance between noninvasive cardiac output monitoring and thermodilution cardiac output.Conclusions:
Over the wide range of cardiac output produced by hemorrhage and resuscitation in large pigs, noninvasive cardiac output monitoring has acceptable agreement with thermodilution cardiac output.