Microbiological profile and 28-day mortality prediction model for bloodstream infection in patients with acute-on-chronic liver failure.

BACKGROUND

Acute-on-chronic liver failure (ACLF) is characterized by systemic immune dysfunction, which predisposes patients to severe infections. Bloodstream infection (BSI) is a common and life-threatening complication in individuals with ACLF; however, pathogen epidemiology and validated prognostic tools specific to this population remain inadequately defined.

METHODS

This retrospective cohort study enrolled 106 patients with ACLF and microbiologically confirmed BSI between January 2020 and December 2024. A predictive nomogram was developed using multivariable logistic regression and internally validated via bootstrap resampling.

RESULTS

The overall incidence of BSI was 11.9%, with corresponding 14-day and 28-day mortality rates of 35.8% and 49.1%, respectively. Of the 106 BSI episodes, 59.4% (n = 63) were hospital-acquired. Gram-negative bacteria were the predominant pathogens (60.4%), with Klebsiella pneumoniae (25.5%) and Escherichia coli (14.2%) being the most frequently isolated species. Among Gram-positive isolates (32.1%), Enterococcus spp. (11.3%) and coagulase-negative staphylococci (10.9%) were the most common. Multidrug-resistant organisms (MDROs) were identified in 32.1% of all isolates and were significantly more prevalent among non-survivors than survivors (42.3% vs. 22.2%; p = 0.027). Four independent predictors of 28-day mortality were identified: concurrent pneumonia, leukocytosis (> 10.7 × 10⁹/L), MELD score > 20.7 and Pitt bacteremia score≥ 2. The predictive nomogram demonstrated strong discriminative ability (AUC = 0.872), excellent calibration, and favorable clinical net benefit.

CONCLUSIONS

BSI in ACLF patients is associated with high short-term mortality and a challenging antimicrobial resistance landscape dominated by gram-negative pathogens. The nomogram-based on four routinely available clinical variables- enables individualized 28-day mortality risk stratification to guide clinical management.