Sepsis is a life-threatening syndrome characterized by organ dysfunction caused by a dysregulated host response to infection. Sepsis-associated acute liver injury (SA-ALI) is a frequent and serious complication of sepsis that considerably impacts both short-term and long-term survival outcomes. In intensive care units (ICUs), the mortality rate of patients with SA-ALI remains high, mostly due to the absence of effective early diagnostic markers and suitable therapeutic strategies. Recent studies have demonstrated the importance of non-coding RNAs (ncRNAs) in the development and progression of SA-ALI. This review focuses on the critical roles of ncRNAs, including microRNAs (miRNAs), long non-coding RNAs (lncRNAs), and circular RNAs (circRNAs), in regulating "cytokine storms", oxidative stress, mitochondrial dysfunction, and programmed cell death in SA-ALI, and summarizes the current state and limitations of existing studies on lncRNAs and circRNAs in SA-ALI. By integrating advancements in high-throughput sequencing technologies, this review provides novel insights into the dual potential of ncRNAs as diagnostic biomarkers and therapeutic targets, offers new ideas for SA-ALI diagnosis and treatment research and highlights potential challenges in clinical translation.

Background: Sepsis, defined as life-threatening organ dysfunction caused by a dysregulated host response to infection, remains a leading cause of death worldwide. Its diagnostic criteria have evolved from Sepsis-1 (SIRS) to Sepsis-3 (SOFA). The recent introduction of the SOFA-2 score, an update to the original SOFA, warrants validation in specific patient populations and against other established scoring systems. This study aims to compare the performance of SIRS, NEWS2, SOFA, and the newly proposed SOFA-2 in the classification agreement and prognostic performance for sepsis in a cohort of patients with suspected infection. Methods: This retrospective study enrolled adults with suspected infection admitted to a tertiary emergency center (January 2024–February 2025). SIRS, NEWS2, SOFA, and SOFA-2 scores were calculated from admission data. Patients were stratified using established thresholds (SIRS ≥ 2, NEWS2 ≥ 5, SOFA ≥ 2, SOFA-2 ≥ 2). Concordance with Sepsis-3 (SOFA ≥ 2) and prognostic accuracy for 28-day mortality were evaluated using AUROC analysis. Score distributions and organ dysfunction patterns were compared. Results: Of 562 screened patients, 516 were included. For sepsis classification agreement, SOFA-2 showed excellent agreement with SOFA (kappa = 0.923) and higher specificity than SIRS and SOFA. For 28-day mortality prediction, SOFA-2 showed the numerically highest AUC (0.863, 95% CI: 0.830–0.892), demonstrating slightly better discrimination than SOFA (AUC:0.854, 95% CI: 0.820–0.883). Pairwise DeLong tests indicated no significant differences between SOFA-2 and SOFA (p = 0.160). At optimal cutoffs, SOFA-2 demonstrated higher specificity (89.08% vs. SOFA 78.59% vs. NEWS2 76.87% vs. SIRS 41.33%), while SOFA showed higher sensitivity (81.63% vs. SOFA-2 73.47%). Increasing SOFA-2 scores strongly correlated with higher in-hospital mortality and longer ICU stay (both p < 0.001). SOFA-2 reclassified respiratory and cardiovascular dysfunction with higher thresholds and greater granularity than SOFA. Conclusions: Based on our dataset, SOFA-2 demonstrates high diagnostic alignment with Sepsis-3 classification and higher specificity for mortality prediction, with slightly better discrimination compared to SOFA, NEWS2, and SIRS. While its slightly lower sensitivity may limit early risk stratification in some patients, its strong prognostic performance supports its utility for risk stratification. Multicenter studies are warranted to determine its role in future sepsis definitions.