Abstract
Background: Prior phenotyping in out-of-hospital cardiac arrest (OHCA) relied predominantly on prehospital indicators, which neither describe nor grade the underlying ischemia-reperfusion injury. We aimed to derive laboratory-based phenotypes and describe their clinical characteristics and outcome associations.
Methods: We conducted a retrospective cohort study of adult resuscitated OHCA patients of presumed cardiac etiology between 01/02/2011 and 01/01/2022, linking Singapore's national OHCA and cardiovascular databases. Unsupervised latent class analysis (LCA) was applied to 14 laboratory variables within six hours post-return of spontaneous circulation (ROSC). Stability was confirmed through split-sample validation. Logistic regression and Cox proportional hazards model, adjusted for age, sex and initial cardiac rhythm, assessed poor neurological recovery at hospital discharge, 30-day and 3-year mortality.
Results: Among 1514 eligible patients (median age 65.0 years, 73.6% males), LCA identified three distinct phenotypes: multi-organ dysfunction (MOD, N=286), homeostatic (HOM, N=967), and immune-dysregulated (IMD, N=261), independently reproduced in derivation (entropy=0.817) and replication (entropy=0.892) subsamples. The most discriminative variables were creatinine, eosinophil and lymphocyte. Compared to MOD, both HOM (aOR: 0.50, 95% CI 0.30-0.78, p=0.004) and IMD (aOR: 0.46, 95% CI 0.26-0.78, p=0.005) phenotypes had significantly lower odds of poor neurological recovery. This pattern was consistent for 30-day mortality (HOM aHR: 0.71, 95% CI 0.61-0.83; IMD aHR: 0.78, 95% CI 0.64-0.94) and 3-year mortality (HOM aHR: 0.70, 95% CI 0.60-0.82; IMD aHR: 0.74, 95% CI 0.62-0.90), all p≤0.01.
Conclusion: Early post-ROSC laboratory data identified three prognostically distinct phenotypes among resuscitated OHCA patients. These exploratory findings support external validation and investigation of phenotype-guided therapies.