Measurement error of cardiac output determined by nitrous oxide rebreathing and impedance cardiography in healthy adults

Cardiac output (Q̇) is a fundamental physiological variable but remains challenging to measure. Exercise Q̇ is commonly measured by inert gas rebreathing using acetylene or nitrous oxide (Q̇_N₂O-IGR) and impedance cardiography (Q̇_IC), but device measurement error has not been assessed at different workloads or in females. This study determined the precision of Q̇_N₂O-IGR (Innocor, COSMED Inc., USA) and Q̇_IC (Physioflow Enduro, Manatec Biomedical, France) in duplicate in 60 adults (30 females; 22 ± 5 years; V̇O_2max: 41.2 ± 8.6 mL^.kg^-1.min^-1) during upright rest and cycling at 50 W and 90% peak power output (PPO) (277 ± 71 W). Measurement variance was higher for Q̇_N₂O-IGR vs. Q̇_IC (<i>p</i> < 0.0001), at 90% PPO compared to 50 W at rest (<i>p</i> < 0.001), and in males vs. females at 50 W with Q̇_N₂O-IGR (<i>p</i> = 0.005). At rest the typical error (TE), coefficient of variation (CV), and intraclass correlation (ICC) were 0.6 L/min (12.4%), 11.6 ± 8.5%, and 0.75 [0.62-0.82] for Q̇_N₂O-IGR and 0.4 L/min (6.9%), 7.5 ± 8.4, and 0.87 [0.79-0.92] for Q̇_IC. At 50 W, the TE, CV, and ICC were 0.8 L/min (7.7%), 5.4 ± 5.8%, and 0.79 [0.68-0.87] for Q̇_N₂O-IGR and 0.5 L/min (5.6%), 3.7 ± 3.8%, and 0.85 [0.76-0.91] for Q̇_IC. At 90% PPO, TE, CV, and ICC were 1.2 L/min (7.5%), 5.4 ± 4.3%, and 0.89 [0.82-0.93] for Q̇_N₂O-IGR and 0.7 L/min (4.0%), 5.0 ± 3.2%, 0.96 [0.93-0.98] for Q̇_IC. Duplicate Q̇_N₂O-IGR differed up to ∼4 L/min (38%) and Q̇_IC ∼3 L/min (24%). In conclusion, group-level measurement precision was generally better for Q̇_IC than Q̇_N₂O-IGR but depended on workload and sex. Duplicate Q̇_N₂O-IGR and Q̇_IC differed substantially; therefore, repeat measures are important.