A novel laboratory-based nomogram for assessing infection presence risk in acute-on-chronic liver failure patients

Abstract

AbstractAccurate assessment of infection presence risk level, timely diagnosis, and effective control are critical for decreasing mortality of Acute‑on‑chronic liver failure (ACLF). We aimed to develop and validate a novel diagnostic model to accurately assess infection presence risk level in ACLF patients. 185 ACLF patients with/without infection were enrolled, and their demographic, physical findings, immune-inflammatory, hepatic function, metabolism, and coagulation-fibrinolysis indicators were analyzed. Regression analysis was performed to identify the independent diagnostic parameters, which were further used to establish diagnostic models with a nomogram for visual. An area under receiver operating characteristic curve (AUROC), calibration plots, clinical impact curves, decision curve analysis, and net reclassification index were used to evaluate and identify the best model. An external validating cohort was introduced to verify the diagnostic accuracy. We screened out white blood cell (WBC) count, LYM%, blood urea nitrogen (BUN), and D-dimer for assessing infection presence risk levels in ACLF patients. WBD (WBC + BUN + D-dimer) was established and proposed as a novel diagnostic model for infection presence risk levels assessment in ACLF patients with an AUROC of 0.803 (95%CI 0.723–0.883), 0.885 (95%CI 0.786–0.984) in training and external cohorts, respectively. In stratification analysis by ACLF etiology and stages, WBD achieved an AUROC of 0.791 (95%CI 0.691–0.891) and 0.873 (95%CI 0.78–0.966) in HBV-related and early-stage patients, respectively. Whereas a higher AUROC of 0.905 (95%CI 0.807–1.00) in the early-stage of HBV-related ACLF patients indicated its optimum application scope. WBD, a novel laboratory-based nomogram, can serve as a decision-making support tool for clinicians to assess infection presence risk levels in ACLF patients.