External Validation of Updated Prediction Models for Neurological Outcomes at 90 Days in Patients with Out-of-Hospital Cardiac Arrest

Abstract Background The accurate prediction of neurological outcomes in patients with out-of-hospital cardiac arrest (OHCA) with post-cardiac arrest syndrome is crucial for determining the optimal treatment or termination of resuscitation efforts. Hence, this study aimed to externally validate updated prediction models for OHCA outcomes using a large nationwide dataset. Methods Existing prediction models for adult patients with non-traumatic OHCA who achieved return of spontaneous circulation were refined using data obtained from the CRITICAL study, a multicentre registry in Osaka, Japan, between January 2013 and December 2019. The primary outcome was a dichotomised 90-day Cerebral Performance Category score. The model was updated using logistic regression with least absolute shrinkage and selection operator regularisation. External validation was performed using data from the JAAM-OHCA registry between January 2014 and December 2019. This is a nationwide multicentre registry in Japan that represents a geographically distinct population from the derivation set. The model performance was evaluated using a validation set. Results Two models (Model 1 included patient demographics, pre-hospital information, and the initial rhythm upon hospital admission; Model 2 included information obtained in the hospital immediately after the return of spontaneous circulation) were updated using the derivation set (n = 3337). In the validation set (n = 4250), Model 1 and 2 exhibited a C-statistic of 0.945 (95% confidence interval (CI): 0.935–0.955) and 0.958 (95% CI: 0.951–0.960), respectively. Both models were well-calibrated to the observed outcomes. Model 2 demonstrated higher net benefits at all risk thresholds according to the decision curve analysis. A web-based calculator was developed to estimate the probability of poor outcomes (https://pcas-prediction.shinyapps.io/90d_lasso/). Conclusions In the validation set, the updated model demonstrated excellent performance in predicting neurological outcomes at 90 days in patients with OHCA. The model, enhanced by incorporating hospital-available information as a predictor, reduced reliance on ambiguous predictors and improved prediction accuracy.