Abstract P101: Comparison of Ventilation Rate Measured with Thoracic Bioimpedance and with Capnography during Out-of-Hospital Cardiopulmonary Resuscitation
Objective: To determine the ability of thoracic bioimpedance to measure ventilation rate during cardiac arrest and CPR.
Methods: Philips MRx devices monitored 32 patients during out-of-hospital cardiac arrest and CPR. The devices recorded chest compressions with an accelerometer, continuous 1-lead EKG, thoracic bioimpedance, and continuous capnography. Of the 32 files, 4 were not used in this study because of incomplete recording. Two reviewers manually annotated ventilation waveforms independently using Laerdal QCPR software, which also automatically annotated ventilation through the bioimpedance channel. Reviewers manually measured ventilation rate (number of breaths/min) recorded with capnography for each 1 minute epoch, which were matched and compared with those measured through bioimpedance for each patient file (N = 28). A total of 585 1-minute epochs were measured and compared. We assessed intra-class correlation for 2 individual raters for ventilation rates measured with capnography and with annotated bioimpedance to establish inter-user reliability of measurements. Ventilation rate measured with capnography vs. bioimpedance was compared with simple regression.
Results: The majority (60%) of ventilation rates measured with capnography and with automated software bioimpedance were within 2 breaths/min of each other. After manual annotation of the bioimpedance channel, 81% of 1-min epochs were within 2 breaths/min of rates measured with capnography. Ventilation rate measured with capnography had good correlation with bioimpedance (r = .82, p < .0001). Inter-rater agreement is estimated to be 0.96 for ventilation rate measured with capnography and 0.93 for rate measured with bioimpedance.
Discussion: The software occasionally missed obvious ventilation waveforms and occasionally annotated waveforms obviously caused by chest compressions. Manual review and annotation improved the accuracy of ventilation rates measured with bioimpedance. Approximately 75% to 90% of recordings made with the Philips MRx device are expected to be useful for measurement.
Conclusion: Ventilation rate measured with thoracic bioimpedance alone is acceptable using the Philips MRx device. Inter-rater agreement for measurements is excellent.