Objective: Cardiac outputs were determined with continuous thermodilution, bolus thermodilution, and the Fick method during pharmacologically varied hemodynamics.
Design: Prospective comparison of techniques.
Setting: University animal laboratory.
Participants: Swine.
Interventions: Swine were anesthetized, tracheally intubated, and instrumented to measure continuous (QTDC) and bolus (QTDB) thermodilution cardiac outputs and sample arterial and mixed venous blood. Continuous thermodilution of blood was facilitated by computer modulation of a thermal filament wrapped around the portion of the pulmonary artery catheter residing in the right atrium and ventricle. QTDC was computed from the thermodilution curve monitored by the thermistor. Bolus thermodilution was performed in triplicate by injecting 10 mL of 5% dextrose in water (0 to 4 degrees C). Oxygen consumption (VO2) was calculated as the averaged minute rate of disappearance of spirometer oxygen over a 6-minute steady state. Cardiac output was determined with the direct Fick method (QF) by dividing VO2 by the difference in arterial and mixed venous oxygen content. Basal QTDC was increased and decreased with an intravenous infusion of dobutamine or labetalol, respectively. Data are summarized as mean +/- SD or 95% confidence interval (CI 95%). Agreement between methods of determining cardiac output was assessed by calculating bias, percent bias, and percent coefficient of determination (100 r2).
Measurements and main results: Eighteen swine (38.9 +/- 1.2 kg) exhibited a range of QTDC from 2.2 to 14.8 L/min. Mean measurement variance of VO2, CaO2, CvO2, and QTDB was 1.5%, 1.5%, 2.0%, and 11.8%, respectively. Mean bias, percent bias, and 100 r2 was 0.004 +/- 1.05 L/min (CI 95%: 0.18 to 0.19 L/min), -0.37 +/- 13.8% (CI 95%: -2.75 to 2.01), and 89% between QTDC and QF, respectively. Bias, percent bias, and 100 r2 was 0.05 +/- 1.09 L/min (CI 95%: -0.14 to 0.23 L/min, 1.21 +/- 13.06% (CI 95%: -1.03 to 3.46%), and 91% between QTDC and QTDB, respectively. Bias, percent bias, and 100 r2 (Fig 6) was -0.04 +/- 0.69 L/min (CI 95%: -0.16 to -.08 L/min), -1.23 +/- 9.17% (CI 95%: -2.8 to 0.35%), and 94% between QTDB and QF, respectively.
Conclusion: Automatic cardiac output computed with continuous thermodilution appears accurate and reliable. Also, good agreement was confirmed between cardiac output derived by continuous and bolus thermodilution methods and bolus thermodilution and Fick methods.