This is the authors' abstract. We don't add key points for this paper.
INTRODUCTION AND OBJECTIVES: Acute-on-chronic liver failure (ACLF) is heterogeneous in the intensive care unit (ICU), and severity scores may not capture first-day physiology. This study aimed to derive first-day ACLF phenotypes in a 24-hour landmark cohort, characterize their clinical profiles, examine mortality associations, and assess preliminary external reproducibility. PATIENTS AND METHODS: Adults with cirrhosis meeting a database-specific operational CLIF-C OF-based definition of ACLF were identified in MIMIC-IV and eICU-CRD. In MIMIC-IV, each patient's first cirrhosis-associated ICU stay was selected before exclusions; patients alive and in the ICU at 24 h entered the landmark cohort. Fifteen variables collected during 0-24 h were clustered separately in 20 imputed datasets using partitioning around medoids, followed by label alignment and consensus assignment. Cox models assessed 30-day mortality, and Fine-Gray models assessed in-hospital death with discharge alive as a competing event. External assignment used MIMIC-IV-derived scaling parameters and phenotype centroids. RESULTS: Among candidate solutions, K = 2 showed the most favorable stability profile despite overlap in the principal-component projection. Phenotype 2 showed greater metabolic-renal disturbance and multiorgan dysfunction. After CLIF-C ACLF adjustment, Phenotype 2 was associated with 30-day mortality (hazard ratio, 1.72; 95% confidence interval, 1.35-2.17) and in-hospital death (subdistribution hazard ratio, 1.99; 95% confidence interval, 1.58-2.49). In eICU-CRD (n = 765), all 15 feature differences were concordant, and in-hospital mortality was 56.7% in Phenotype 2 versus 23.1% in Phenotype 1. CONCLUSIONS: Two first-day clinical patterns differed in physiology, organ-failure burden, and mortality after the 24-hour landmark. External concordance supports preliminary reproducibility; validation is required.
Answers are grounded in this paper and cite their sources.
No takes yet. Share an insight, caveat, or question.
Zhang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: