Multiplying flow and pressure: detecting respiratory phases in intra-arrest ventilation

Area of Science:

• Biomedical Engineering
• Critical Care Medicine
• Respiratory Physiology

Background:

• Detecting respiratory phases during cardiopulmonary resuscitation (CPR) is challenging due to chest compressions.
• Accurate respiratory phase detection is crucial for optimizing ventilation strategies during advanced cardiac life support.

Purpose of the Study:

• To develop and validate an automated algorithm for detecting respiratory phases and their precise onsets.
• To enable scalable, objective analysis of ventilation during intra-arrest conditions with ongoing chest compressions.

Main Methods:

• An algorithm was developed utilizing the product of airflow and airway pressure, and their slopes.
• Ventilatory data from 13 pigs were analyzed, including periods of regular ventilation and intra-arrest ventilation with chest compressions.
• Algorithm performance was assessed against investigator-validated respiratory phase onsets.

Main Results:

• The algorithm achieved perfect classification of inspiratory and expiratory phases during both regular and intra-arrest ventilation.
• An F1-score of 1 was obtained for regular ventilation phase onset detection.
• An F1-score of 0.971 was achieved for intra-arrest ventilation phase onset detection.

Conclusions:

• A robust algorithm for detecting respiratory phases and onsets during chest compressions was developed.
• The algorithm leverages the relationship between airflow and airway pressure to distinguish ventilation from compression artifacts.
• This method shows excellent performance and potential for automated respiratory monitoring in critical care.