Response to Huang et al

We thank Huang and colleagues for their thoughtful comments on our study (1), which demonstrated that serial procalcitonin (PCT) measurements during the first 72 hours of admission provide clinically useful information for diagnosing bacterial infection and predicting short-term mortality in cirrhotic patients with systemic inflammatory response syndrome (SIRS) (2). Their observations raise important issues concerning the interpretation and application of PCT in this population. First, PCT values can be influenced by noninfectious processes, including hepatocellular injury in acute or chronic liver disease, systemic inflammation, and endotoxemia (3). These factors may complicate interpretation of a single PCT measurement. Although subgroup analyses by etiology would add insight, pooled evidence indicates that PCT has moderate-to-good diagnostic accuracy for bacterial infection in cirrhosis (sensitivity 73%, specificity 83%) and similar performance for spontaneous bacterial peritonitis (sensitivity 76%, specificity 87%) (4). However, interpretation of a single cutoff is limited by variability in thresholds across studies and the impact of liver dysfunction on baseline values. Evaluating serial changes, rather than relying solely on isolated measurements, may therefore better capture the clinical trajectory. In our cohort, dynamic changes were more informative than single values: A decline of more than 50% within 72 hours was associated with infection resolution, while a 72-hour/baseline ratio greater than 0.8 signaled persistent infection and increased mortality (1). Second, we agree that PCT should not be regarded as an independent predictor of outcome. While PCT kinetics correlated with 30-day mortality in our study, conventional prognostic scores, including MELD, CLIF-SOFA, and Child–Turcotte–Pugh remained strong determinants of prognosis (1). Additional clinical variables, including renal dysfunction, elevated C-reactive protein (CRP), multisite infections, acute-on-chronic liver failure (ACLF), and hepatocellular carcinoma are also established predictors of adverse outcomes and should be incorporated into comprehensive models. Prior studies have shown that adding PCT and/or CRP to MELD modestly improves prediction of 30–90 day mortality (5). Our findings suggest that serial PCT kinetics may further refine prognostic stratification, particularly when used in combination with MELD (1). Third, cytokines such as interleukin-6 (IL-6) may offer incremental value. Zhang et al reported that in HBV-related ACLF, presepsin and IL-6, rather than PCT, independently predicted bacterial infection (6). These data underscore the heterogeneity of cirrhotic populations and highlight the importance of multicenter investigations to validate biomarker performance across varying etiologies and clinical settings. Differences in baseline characteristics, liver disease severity, and comorbidity likely account for variability in predictive accuracy. Integrated approaches that combine established severity scores with multiple inflammatory markers (PCT, CRP, IL-6, leukocyte count) may ultimately provide greater prognostic precision than reliance on a single marker. However, such models have yet to be fully developed and externally validated. In summary, our findings support the role of serial PCT kinetics as an adjunctive tool that complements, but does not replace, established prognostic scores and other biomarkers. PCT should be interpreted in the broader clinical context and in conjunction with complementary markers. Future multicenter studies with longer follow-up are required to validate integrated biomarker-based models for diagnosing infection and predicting mortality in cirrhosis. CONFLICTS OF INTEREST Guarantor of the article: Phunchai Charatcharoenwitthaya, MD, MS. Specific author contributions: P.C.: conceptualization, formal analysis, supervision, manuscript writing. Financial support: This work was supported by the Faculty of Medicine Siriraj Hospital, Mahidol University, Grant Number R015633001. The funding sources had no involvement in the conduct of the research or preparation of the article. Potential competing interests: None to report.