Abstract Aims This study aimed to develop Metabolic Derangement score and validate its prognostic performance in patients with out-of-Hospital Cardiac Arrest (OHCA) and also its discriminative performance in guiding the application of extracorporeal life support (ECLS). Methods and Results From 5,100 patients (average age 67 years, 69% male), Metabolic Derangement score was developed to predict primary outcome (death or poor neurologic outcome, cerebral performance category ≥ 3 at day 30) from initial blood laboratory test including pH, PaCO2, PaO2, potassium, haemoglobin, lactate, and creatinine. It was created using machine learning algorithms and a training set (60%), and validated with a test set (40%). The risk of primary outcome consistently increased from 34% to 100% as the Metabolic Derangement score ranged from 0 to 7. Patients with Metabolic Derangement score ≤ 2 showed no difference in the primary outcome risk between those with or without ECLS. Conversely, those with a Metabolic Derangement score ≥3 had a lower risk with ECLS (difference of restricted mean survival time=6.5 days, 95% CI=3.3-9.6 days, ratio of restricted mean time lost=0.76, 95%CI=0.65-0.88, all p<0.001). Clinical subgroup analyses showed generally consistent results. Metabolic Derangement score also outperformed OHCA-, CAHP-, and NULL-PLEASE-scores in prediction the primary outcome (p<0.05, all). Conclusions The severity of metabolic derangement predicted the clinical outcome of OHCA, and could also identify patients likely to benefit from ECLS.Metabolic Derangement score