Cardiac output (CO) during exercise can be determined noninvasively by using the indirect Fick CO2-rebreathing technique. CO2 measurements for this technique are usually performed with an infrared analyzer (IA) or mass spectrometer (MS). However, IA CO2 measurements are susceptible to underreading in the face of high O2 concentrations because of collision broadening. We compared an IA (Ametek model CD-3A) with a MS (Marquette model MGA-1100) to see the effect this would have on mixed venous PCO2 (PVCO2) and CO measurements. After calibration with room air and a gas mixture of 5% CO2–12% O2–83% N2, both devices were tested with three different gas mixtures of CO2 in O2. For each gas mixture, IA gave lower CO2 values than did the MS (4.1% CO2: IA, 3.85 +/- 0.01% and MS, 4.13 +/- 0.01%; 9.2% CO2: IA, 8.44 +/- 0.07% and MS, 9.19 +/- 0.01%; 13.8% CO2: IA, 12.57 +/- 0.15% and MS, 13.82 +/- 0.01%). Warming and humidifying the gases did not alter the results. The IA gave lower values than did the MS for eight other medical gases in lower concentrations of O2 (40–50%). Equilibrium and exponential rebreathing procedures were performed. Values determined by the IA were > 10% higher than those determined by the MS for both rebreathing methods. We conclude that all IAs must be checked for collision broadening if they are to be used in environments where the concentration of O2 is > 21%. If collision broadening is present, then either a special high O2-CO2 calibration curve must be constructed, or the IA should not be used for both arterial PCO2 and PVCO2 estimates because it may produce erroneously low PVCO2 values, with resultant overestimation of CO.
Improving the Reliability of the Rebreathing Method for Measuring Mixed Venous Pco2