Purpose of the study: Real-time capnometry is a valuable monitoring tool during cardiopulmonary resuscitation (CPR). However, waveform capnography is often distorted by chest compressions (CC) [1] and CC artefact might compromise the end-tidal CO2 (ETCO2) measurements provided by capnometers [2]. We quantified the differences in ETCO2 measurements by two CO2 detector devices caused by CC artifact during manual CPR.
Materials and methods: We retrospectively analysed out-of-hospital cardiac arrest monitor-defibrillator recordings from two different agencies (n=30 from agency A and n = 14 from agency B). Each agency used a different side-stream capnography technology and monitor-defibrillators. Capnograms were reviewed in 3-min segment and each segment was classified as distorted or non-distorted by CC artefact. Ventilations were identified using both the transthoracic impedance signal and the capnogram. ETCO2 level was manually annotated as the maximum CO2 level in each ventilation, and compared to values extracted from the monitor-defibrillators.
Results: Eighty capnogram segments were analysed, containing a total of 2,087 ventilations with a 43.5% of them distorted by CC artifact. Segments from agency A had a total of 1,125 ventilations (50.5% distorted), whereas those from agency B had 962 ventilations (35.1% distorted). Globally, ETCO2 median (IQR) unsigned error in percentage for non-distorted and distorted intervals was 3.2 (2.1-4.6)% and 9.4 (6.9-12.9)%, respectively. In device A the median unsigned error was 2.1 (1.3-3.1)% and 8.4 (6.4-10.4)%, while in device B the median error was 3.9 (3.3-6.0)% and 13.9 (10.6-15.8)%, respectively.
Conclusion: Presence of CC artefact increased ETCO2 measurement errors in the assessed CO2 detector devices. CC artefact suppression techniques are needed to enhance waveform capnography, allowing reliable and clinically useful ETCO2 analysis during CPR. 

[1] Leturiondo M et al. Resuscitation 2018;124:63-68.
[2] Raimondi M et al. Resuscitation 2016;103:e9-e10.

2 - Chest compression artefact compromises real-time feedback capnometry: quantification of differences in end-tidal measurements by two capnometers