Noninvasive Cardiac Arrest Recognition Using Wearable Sensors in a Porcine Model of Ventricular Fibrillation

Avatar
Poster
Voice is AI-generated
Connected to paperThis paper is a preprint and has not been certified by peer review

Noninvasive Cardiac Arrest Recognition Using Wearable Sensors in a Porcine Model of Ventricular Fibrillation

Authors

Englhardt, Z.; Bin Saeed, M.; Razi, A.; Ayoub, I.; Baetiong, A.; Patel, S.; Gazmuri, R. J.; Sunshine, J. E.; Iyer, V.

Abstract

Background: Wearable detection of out-of-hospital cardiac arrest (OHCA) could improve survival, given the prognostic importance of time to resuscitation and that 50% to 75% of events are unwitnessed. Existing systems may be limited by their form factor and by specificity achieved at the cost of sensitivity. We investigated whether the raw photoplethysmography (PPG) signal from low-cost wearable sensors contains a distinct signature of circulatory arrest. Methods: Ventricular fibrillation (VF) was electrically induced in 9 invasively monitored pigs while raw PPG was recorded with a wearable-class sensor. PPG dynamics after VF onset were related to aortic and right atrial pressures. A heuristic detection algorithm was evaluated for sensitivity and latency by leave-one-out cross-validation in the porcine data and for false positives in 57,857 hours of free-living PPG from 115 human participants. Results: After VF onset, the raw PPG showed a reproducible exponential decay (time constant 10.53 s; coefficient of variation 8.9% to 17.0% over the first 10 s) that tracked rising right atrial pressure (r = 0.90), reflecting arterial-to-venous blood redistribution. The algorithm detected VF in 8 of 9 animals (sensitivity 0.889; 95% CI, 0.518-0.997) with a median latency of 57.2 s. In the human cohort, it produced 2.73 false positives per user-year (95% CI, 1.62-4.31), comparable to current systems. Conclusions: Raw wearable PPG encodes circulatory collapse during VF, as supported by invasive hemodynamic measurements. While preliminary, these results suggest the potential for wearable-based OHCA detection with improved sensitivity, reduced latency, and comparable specificity.

Follow Us on

0 comments

Add comment