Using Out-of-Hospital Cardiac Arrest (OHCA) and Cardiac Arrest Hospital Prognosis (CAHP) Scores with Modified Objective Data to Improve Neurological Prognostic Performance for Out-of-Hospital Cardiac Arrest Survivors

This study aimed to determine whether accuracy and sensitivity concerning neurological prognostic performance increased for survivors of out-of-hospital cardiac arrest (OHCA) treated with targeted temperature management (TTM), using OHCA and cardiac arrest hospital prognosis (CAHP) scores and modified objective variables. We retrospectively analyzed non-traumatic OHCA survivors treated with TTM. The primary outcome was poor neurological outcome at 3 months after return of spontaneous circulation (cerebral performance category, 3-5). We compared neurological prognostic performance using existing models after adding objective data obtained before TTM from computed tomography (CT), magnetic resonance imaging (MRI), and biomarkers to replace the no-flow time component of the OHCA and CAHP models. Among 106 patients, 61 (57.5%) had poor neurologic outcomes. The area under the receiver operating characteristic (AUROC) curve for the OHCA and CAHP models was 0.89 (95% confidence interval (CI) 0.81-0.94) and 0.90 (95% CI 0.82-0.95), respectively. The prediction of poor neurological outcome improved after replacing no-flow time with a grey/white matter ratio measured using CT, high-signal intensity (HSI) on diffusion-weighted MRI (DWI), percentage of voxel using apparent diffusion coefficient value, and serum neuron-specific enolase levels. When replaced with HSI on DWI, the AUROC and sensitivity of the OHCA and CAHP models were 0.96 and 74.5% and 0.97 and 83.8%, respectively (100% specificity). Prognoses concerning neurologic outcomes improved compared with existing OHCA and CAHP models by adding new objective variables to replace no-flow time. External validation is required to generalize these results in various contexts.

Automated summary

This study found that replacing the no-flow time component in OHCA and CAHP models with new objective variables, such as CT, MRI, and biomarkers, can improve the accuracy and sensitivity in predicting the neurological prognosis of cardiac arrest patients. Among 106 patients, approximately 57.5% had poor neurological outcomes. By adding new variables to replace the no-flow time, the prediction of poor neurological outcomes was improved.