Comparison of non-invasive cardiac output measurement by partial carbon dioxide rebreathing with the lithium dilution method in anesthetized dogs


  • Presented in part at the Spring Meeting of the Association of Veterinary Anesthetists, Doorwerth, The Netherlands, May 2003; and at the World Congress of Veterinary Anesthesia, Knoxville, TN, USA, October 2003.

Address correspondence and reprint requests to:
Dr. C. Gunkel, Department of Large Animal Clinical Sciences, College of Veterinary Medicine, University of Florida, Gainesville, FL 32610.


Objective: To compare the partial CO2 rebreathing method (non-invasive cardiac output [NICO]) and the lithium dilution method (lithium dilution cardiac output [LiDCO]) for cardiac output (CO) measurement in anesthetized dogs.

Design: Prospective study.

Setting: College of Veterinary Medicine, University of Florida.

Animals: Six adult dogs (weight range 22–25.4 kg).

Interventions: All animals were instrumented for CO determinations using the LiDCO and NICO methods. Direct blood pressure, heart rate, arterial blood gases, end-tidal isoflurane (ETI), and CO2 concentrations were monitored throughout the study. CO was manipulated with dobutamine and isoflurane to allow for intermediate, low, and high CO determinations in that order using LiDCO and NICO.

Measurements and main results: A 1.5% ETI produced the intermediate rate of CO, a constant-rate infusion of dobutamine (1–4 μg/kg/min) and 1.1% ETI, the highest rate, and 2.5–3% ETI, the lowest rate. Measurements were obtained in duplicate or triplicate for the LiDCO and continuously for the NICO method after achieving a stable hemodynamic plane for at least 15 minutes at each level of CO, allowing 5 minutes between measurements. Forty-seven comparisons were determined. The correlation coefficient (r) between the 2 methods was 0.888 for all determinations. The mean LiDCO and NICO from 47 measurements were 155.9±78.7 mL/kg/min (range, 49.6–303.2) and 146.6±62.9 mL/kg/min (50–290.3), respectively. The bias between LiDCO and NICO estimations was 9.3 (−60.7 to +79.4) mL/kg/min (mean and 95% confidence interval). The mean (mL/kg/min) of the differences of LiDCO–NICO was 1.11 × NICO. The relative error was 2.4±24.7%. As CO increased, the relative difference between the methods also increased.

Conclusions: The NICO is a viable non-invasive method for CO determination in the dog and compares well with the LiDCO.