Aim

To develop a new scoring model for patients with cardiogenic out-of-hospital cardiac arrest (OHCA) to facilitate neurological prognosis prediction upon hospital arrival by using prehospital resuscitation features alone.

Methods

Between 2005 and 2019, we enrolled 942,891 adult patients with OHCA of presumed cardiac aetiology from the All-Japan Utstein Registry. Scoring models applied prehospital resuscitation features a priori from the variables the American College of Cardiology algorithm including age, duration to return of spontaneous circulation (ROSC) or hospital arrival, no bystander cardiopulmonary resuscitation (CPR), unwitnessed arrest, and nonshockable rhythm (R-EDByUS score) to predict unfavorable neurological outcomes defined as Cerebral Performance Category 3, 4, or 5 at 1 month. We created nomograms as a "Regression-based model," and created a "Simplified model" in which points were assigned by category for predicting unfavorable neurological outcomes for both the prehospital ROSC cohort (67,064 patients) and the ongoing CPR cohort (875,827 patients). For internal validation, bootstrap optimism-corrected estimates of predictive performance were calculated.

Results

A total of 46,971 (70.0%) and 870,991 (99.4%) patients in the prehospital ROSC and ongoing CPR cohorts, respectively, had unfavorable neurological outcomes. In the prehospital ROSC cohort, the C-statistics of the Regression-based and Simplified models were 0.851 and 0.842, and the bootstrap-validated C-statistics were 0.852 and 0.841, respectively. In the ongoing CPR cohort, the C-statistics of the Regression-based and Simplified models were 0.872 and 0.865, and the bootstrap-validated C-statistics were 0.852 and 0.841, respectively.

Conclusions

The R-EDByUS score accurately predicted the neurological prognosis of cardiogenic OHCA upon hospital arrival.