Predictive validity of SOFA-2 score for In-Hospital Mortality Among Patients With Suspected Infection Presenting to the Emergency Department

BACKGROUND: The Sequential Organ Failure Assessment (SOFA-1) score is widely used to quantify organ dysfunction in patients with suspected infection but does not reflect several modern critical care practices. The recently developed SOFA-2 score updates key organ dysfunction components; however, its performance in emergency department (ED) populations remains unclear. METHODS: We conducted a retrospective multi-institutional cohort study across six hospitals in Taiwan from January 2010 to December 2021. Adult patients (≥18 years) presenting to the ED with suspected infection were included. The primary outcome was in-hospital mortality. Predictive performance of SOFA-2 was compared with SOFA-1, quick SOFA (qSOFA), and Systemic Inflammatory Response Syndrome (SIRS) criteria. Discriminatory ability was evaluated using the area under the receiver operating characteristic curve (AUROC), and reclassification analysis was performed to assess differences between SOFA-2 and SOFA-1. RESULTS: Among 173,079 ED encounters included, in-hospital mortality was 9.3%. Compared with SOFA-1, SOFA-2 assigned lower scores in 62.3% of patients and higher scores in 1.5%. Patients reclassified to a higher SOFA-2 score had higher mortality (25.2%) compared with those reclassified to a lower score (9.0%) or with unchanged scores (9.1%). For predicting in-hospital mortality, SOFA-2 showed good discrimination (AUROC 0.78; 95% confidence interval [CI], 0.78-0.78), similar to SOFA-1 (AUROC 0.78; 95% CI, 0.77-0.78) and higher than qSOFA (AUROC 0.72; 95% CI, 0.72-0.73) and SIRS (AUROC 0.63; 95% CI, 0.63-0.64). CONCLUSIONS: SOFA-2 demonstrated good predictive validity for in-hospital mortality among ED patients with suspected infection, supporting its use for early risk stratification and potentially informing future revisions of the sepsis definition.