Comparison between mainstream (Capnostat 5) and a low-flow sidestream capnometer (Capnostream) in mechanically ventilated, sevoflurane-anesthetized rabbits using a Bain coaxial delivery system.

OBJECTIVE: To evaluate agreement between end-tidal carbon dioxide (Pe'CO2) and PaCO2 with sidestream and mainstream capnometers in mechanically ventilated anesthetized rabbits, with two ventilatory strategies. STUDY DESIGN: Prospective experimental study. ANIMALS: A total of 10 New Zealand White rabbits weighing 3.6 ± 0.3 kg (mean ± standard deviation). METHODS: Rabbits anesthetized with sevoflurane were intubated with an uncuffed endotracheal tube (3.0 mm internal diameter) and adequate seal. For Pe'CO2, the sidestream capnometer sampling adapter or the mainstream capnometer was placed between the endotracheal tube and Bain breathing system (1.5 L minute-1 oxygen). PaCO2 was obtained from arterial blood collected every 5 minutes. A time-cycled ventilator delivered an inspiratory time of 1 second and 12 or 20 breaths minute-1. Peak inspiratory pressure was initially set to achieve Pe'CO2 normocapnia of 35-45 mmHg (4.6-6.0 kPa). A total of five paired Pe'CO2 and PaCO2 measurements were obtained with each ventilation mode for each capnometer. Anesthetic episodes were separated by 7 days. Agreement was assessed using Bland-Altman analysis and linear mixed models; p < 0.05. RESULTS: There were 90 and 83 pairs for the mainstream and sidestream capnometers, respectively. The mainstream capnometer underestimated PaCO2 by 12.6 ± 2.9 mmHg (proportional bias 0.44 ± 0.06 mmHg per 1 mmHg PaCO2 increase). With the sidestream capnometer, ventilation mode had a significant effect on Pe'CO2. At 12 breaths minute-1, Pe'CO2 underestimated PaCO2 by 23.9 ± 8.2 mmHg (proportional bias: 0.81 ± 0.18 mmHg per 1 mmHg PaCO2 increase). At 20 breaths minute-1, Pe'CO2 underestimated PaCO2 by 38.8 ± 5.0 mmHg (proportional bias 1.13 ± 0.10 mmHg per 1 mmHg PaCO2 increase). CONCLUSIONS AND CLINICAL RELEVANCE: Both capnometers underestimated PaCO2. The sidestream capnometer underestimated PaCO2 more than the mainstream capnometer, and was affected by ventilation mode.