[Skip to Content]
[Skip to Content Landing]
Views 1,648
Citations 0
Biotech Innovations
August 13, 2019

Smart Devices Detect Agonal Breathing in Cardiac Arrest

JAMA. 2019;322(6):495. doi:10.1001/jama.2019.11648

Roughly half of people who experience a cardiac arrest demonstrate agonal breathing, a brainstem reflex during severe hypoxia. Researchers at the University of Washington in Seattle are teaching smart devices to detect this audible biomarker, often described as gasping breaths. Agonal breathing sounds from Seattle-area 911 cardiac arrest calls were used to train and validate a detection system recently introduced in npj Digital Medicine.

The system was tested on a commercially available smart speaker and smartphone in a proof-of-concept study involving 164 hours of sleep sounds from 35 bedroom environments and 82 hours of sleep lab sounds, including apneas, hypopneas, and snoring events that can mimic agonal breathing. The technology detected agonal breathing around 97% of the time and was effective at a range of 6 m. The system’s already low false-positive rate fell to 0% when agonal breathing frequency was factored in.

To further refine the program, the researchers plan to obtain more audio examples of agonal breathing from 911 databases around the world and test it in environments with above-average risks of cardiac arrest, such as elder care facilities and certain hospital wards.

With additional training, the software could be deployed on smart speakers, smartphones, or smartwatches, said Jacob Sunshine, MD, the study’s senior author. In the event of a detected cardiac arrest, the system would summon a cardiopulmonary resuscitation–trained individual, activate the emergency response system, or both, after providing users a chance to cancel any false alarms, he said.