Abstract Body (Do not enter title and authors here): Background: Previously described algorithms evaluated defibrillator biosignals during ongoing CPR to characterize the underlying rhythm and its physiologic phenotype. We hypothesized that a “smart” resuscitation strategy incorporating these algorithms could reduce CPR interruptions and better align rescuer actions with patient-specific physiology.
Methods: In a cohort study of ventricular fibrillation OHCA from 2017-2019, rescuer actions (rhythm analysis, shock delivery, pulse check and medication administration) were extracted from EMS, audio, and defibrillator recordings. Previously validated algorithms were combined to assess both cardiac rhythm and physiologic measures. The combined smart algorithm was compared to observed clinical practice with respect to rhythm accuracy and CPR interruption. The frequency of potentially misdirected drug therapy, defined as epinephrine use with >50% probability of spontaneous pulse or antiarrhythmic use with a predicted rhythm of asystole or bradycardia, was determined. Low vitality physiologic phenotype was defined as a shockable rhythm with low probability of post-shock ROSC or a non-shockable rhythm with low probability of a pulse during a clinical pulse check. Vitality phenotypes were compared with respect to post-shock ROSC following shock delivery and pulse prevalence during pulse checks.
Results: Of 390 VF-OHCA cases, median age was 64, and 46% survived to hospital discharge. There were a median of 5 rhythm analyses, 3 shocks, and 2 pulse checks per case. The smart strategy achieved comparable shock accuracy (95% sensitivity, 98% specificity) to observed care while decreasing the median CPR interruption from 12 to 6 seconds. Of 597 epinephrine doses, the algorithm identified 17% (n=99) with predicted probability of ROSC over 50%. Of 248 antiarrhythmic doses, the algorithm predicted the rhythm was asystole or bradycardic organized in 9% (n=23). Following 1334 VF shocks, post-shock ROSC differed by phenotype: 4% (9/217) with low vitality versus 22% (244/1117) with high vitality (p<0.01). During 1088 pulse checks, a pulse was present in 9/171 (5%) with low vitality compared to 325/917 (35%) with high vitality (p<0.01).
Conclusion: A smart resuscitation strategy that assesses the specific rhythm and physiologic phenotype could positively affect hands-on time, medication administration, shock delivery, and pulse checks. It has the promise to improve OHCA survival but requires rigorous comparison to current protocols.