Procalcitonin as a Specific Predictor of Clinical Outcomes in Acute-on-Cirrhosis Sepsis: A Retrospective Pilot Analysis

Background/Aims: Sepsis as an acute cause of liver dysfunction is associated with high mortality. Routine infection/inflammation markers—C-reactive-protein (CRP), procalcitonin (PCT), and leukocyte count (LeuC)—are frequently used for risk stratification in septic patients. This study aimed to evaluate these markers as predictors of short-term and 12-month mortality in septic patients with distinct liver dysfunction phenotypes. Methods: This single-center retrospective pilot analysis involved adults with sepsis and varying degrees of liver dysfunction—acute liver failure (ALF), acute-on-chronic liver failure (ACLF), or acute-on-cirrhosis (ACOC)—treated in intermediate or intensive care units between 2016 and 2017. At sepsis onset, patients were categorized into ACOC, ACLF, and ALF groups. Only patients with recorded CRP, PCT, and LeuC measurements 24 h before, on the day of, and 24/48 h after sepsis onset were included in the analysis. Associations with in-hospital and 12-month mortality were analyzed using Firth bias-reduced logistic regression, ROC analysis, and internal validation by bootstrapping and cross-validation. Results: 49 patients were included (ACOC n = 21; ACLF n = 20; ALF n = 8). In-hospital and 12-month mortality rates were 34.7% and 61.2%, respectively, with the highest long-term mortality observed in the ACOC group (76.2%). In the ACOC group, PCT 24 h before sepsis onset independently predicted in-hospital mortality (OR ~5 per PCT doubling; AUC 0.94), with an optimal rule-in cut-off of 1.0 ng/mL (specificity 1.00, PPV 1.00). PCT was not predictive in ACLF/ALF, and CRP/LeuC offered limited prognostic value. Conclusions: In this hypothesis-generating analysis, PCT 24 h before sepsis onset shows a phenotype-specific association with early mortality in ACOC. Larger, prospective multicenter studies are needed to validate PCT-guided risk stratification.