Clinical Prediction Models for Prognostication After Out-of-Hospital Cardiac Arrest: A Systematic Review and...

Area of Science:

• Neurology
• Cardiology
• Critical Care Medicine

Background:

• Accurate neuroprognostication is crucial for guiding clinical decisions after out-of-hospital cardiac arrest (OHCA).
• Existing clinical prediction models (CPMs) are used to predict outcomes, but their performance requires evaluation.

Purpose of the Study:

• To summarize the prognostic performance of current clinical prediction models (CPMs) for neuroprognostication in adult patients following out-of-hospital cardiac arrest (OHCA).

Main Methods:

• A systematic literature search was conducted in Medline and Embase databases up to June 1, 2025.
• Studies evaluating CPMs for predicting poor functional outcome in adult OHCA survivors were selected, excluding derivation cohorts and models with insufficient external validation.
• Sensitivity, specificity, and area under the receiver operating characteristic curve (AUROC) were pooled where possible; risk of bias and certainty of evidence were assessed.

Main Results:

• 39 observational cohorts (95,037 patients) evaluated 11 CPMs, most commonly the OHCA and Cardiac Arrest Hospital Prognosis (CAHP) scores.
• The OHCA score showed pooled sensitivities and specificities ranging from 64.9% to 81.8% and 74.2% to 89.5% for poor outcome (low certainty).
• The CAHP score (≥150) demonstrated pooled sensitivity of 81.3% and specificity of 77.0% (moderate certainty), with pooled AUROCs across models ranging from 0.75 to 0.88, indicating substantial heterogeneity.

Conclusions:

• Clinical prediction models for neuroprognostication after OHCA exhibit moderate accuracy.
• Significant heterogeneity across validation cohorts limits the clinical utility of these CPMs.
• The current limitations restrict the reliable use of CPMs for irreversible decisions, such as withdrawal of life-sustaining therapy.