Index and recurrent decompensation have different impacts on the course of acute decompensation (AD). Contemporary biochemical liver failure definitions for acute-on-chronic liver failure (ACLF) do not distinguish between index and prior decompensation. We evaluated whether prior decompensation defines a clinically distinct phenotype and examined whether established biochemical liver failure criteria retain consistent prognostic performance across these phenotypes

This multicentre study evaluated 966 adults hospitalized with AD across four centres in India, stratified into first-presentation (n = 364) and prior-decompensation (n = 602) cohorts. Biochemical thresholds from APASL (bilirubin ≥5 mg/dL and INR ≥1.5), CLEARED (≥7.5 mg/dL and INR≥1.5), and EASL-CLIF (bilirubin ≥12 mg/dL) were evaluated against 30-day mortality.

Patients with prior decompensation were older, more often had MASLD-related cirrhosis (21.3% vs 10.7%), and showed greater extra-hepatic organ failures. Thirty-day mortality was higher in patients with prior decompensation independent of MELD and baseline infection (40.5% vs 30.5%; adjusted RR 1.31, 95% CI 1.10–1.56, p=0.003), largely attributable to hepatic encephalopathy, which was twice as common (60% vs 29%). In the overall cohort, APASL and CLEARED modestly predicted mortality (OR 1.36 and 1.41, respectively), whereas EASL-CLIF threshold did not. In decompensation phenotype-stratified analysis, prognostic performance varied; biochemical thresholds were associated with 30-day mortality only among prior decompensation (APASL: OR 1.57; CLEARED: OR 1.67; CLIF-C: OR 1.71), but not in the first-presentation phenotype. The bilirubin-only CLIF-C OF liver-failure component captured the fewest deaths and showed significant phenotype-dependent prognostic behaviour.

Prior decompensation defines a clinically distinct, high-risk phenotype mediated by encephalopathy. Liver-failure criteria perform discordantly across phenotypes with different prognostic signals in first and prior decompensations.