External validation of MIRACLE2, rCAST, and TTM scores for neurological outcomes in comatose out-of-hospital cardiac arrest survivors: A multi-center study.
Journal
Heart and Lung
Journal Volume
78
Start Page
Article Number : 102776
ISSN
1527-3288
Date Issued
2026-04-01
Author(s)
Huang, Sih-Shiang
Chang, Wei-Lun
Huang, Chun-Hsiang
Ho, Yi-Ju
Chen, Ching-Yu
Lien, Chun-Ju
Chen, Yun-Chang
Wang, Tzu-Hsueh
Abstract
Background: The MIRACLE2, rCAST, and TTM scores are established tools for predicting poor neurological outcomes in out-of-hospital cardiac arrest (OHCA) survivors. However, studies concurrently validating these scoring systems within a single cohort are lacking. Additionally, their relative applicability and optimal cutoff values remain uncertain. Objectives: This study aimed to externally validate and compare the predictive performance of the MIRACLE2, rCAST, and TTM scores and to identify clinically relevant cutoff values for predicting poor neurological outcomes. Methods: This study included adult OHCA survivors with a presumed cardiac origin (external validation cohort) who underwent post-cardiac arrest care at three tertiary medical centers in Taiwan between January 2016 and December 2023. Poor neurological outcomes at 6 months post-arrest were defined as a Cerebral Performance Category score of 3–5. Discriminative ability was assessed using the area under the receiver operating characteristic curve (AUROC), and calibration was evaluated using calibration plots. Performance metrics were derived for the optimal cutoff thresholds. Results: A total of 693 patients were assessed, with 35.5 % presenting favorable neurological outcomes. The AUROC values for MIRACLE2, rCAST, and TTM were 0.799 (95 % confidence interval [CI]: 0.766–0.833), 0.812 (95 % CI: 0.779–0.845), and 0.815 (95 % CI: 0.784–0.846), respectively. A MIRACLE2 score cutoff of 8 achieved a specificity of 99.2 % and a positive predictive value (PPV) of 94.4 %. A cutoff of 16.5 for rCAST yielded a specificity of 98.4 % and a PPV of 94.8 %. A TTM score cutoff of 21 produced a specificity of 99.2 % and a PPV of 98.3 %. rCAST exhibited the highest applicability for predicting poor neurological outcomes. Conclusion: Among the three validated models, the rCAST score demonstrated the highest clinical applicability for identifying patients at high risk of poor neurological outcomes.
Subjects
External validation
MIRACLE2 score
Neurological outcome
Out-of-hospital cardiac arrest
TTM score
rCAST score
Publisher
Elsevier Inc.
Type
journal article