The updated Sequential Organ Failure Assessment score, SOFA-2, demonstrates good discrimination for short-term in-hospital mortality in older emergency department patients with sepsis but does not improve prognostic performance compared with the original SOFA score, according to a retrospective cohort study.
Investigators noted that SOFA-2 was developed to update the conventional SOFA score for contemporary clinical practice. However, its prognostic performance among older adult patients presenting to the emergency department with sepsis had remained uncertain prior to this evaluation.
The single-center retrospective study analyzed data from 330 hospitalized emergency department patients aged 65 years or older diagnosed with sepsis. Researchers calculated SOFA-2, the original SOFA score, the National Early Warning Score (NEWS), NEWS with lactate (NEWS-L), and the Modified Early Warning Score (MEWS) using measurements obtained during each patient’s initial emergency department assessment.
Primary study outcomes were defined as 1-day and 7-day in-hospital mortality. The research team evaluated discriminatory capacity using the area under the receiver operating characteristic curve (AUC), applying DeLong’s test alongside Holm–Bonferroni correction for pairwise comparisons, and additionally assessed threshold-based sensitivity and specificity.
Within the cohort, mortality occurred in 43 patients (13.0%) within 1 day and in 141 patients (42.7%) within 7 days of hospital presentation. For 1-day mortality, SOFA-2 yielded an AUC of 0.83 (95% CI 0.77–0.89) compared with 0.82 (95% CI 0.76–0.89) for original SOFA, a difference that was not statistically significant after adjustment (p = 0.526). For 7-day mortality, SOFA-2 achieved an AUC of 0.76 (95% CI 0.70–0.81) versus 0.75 (95% CI 0.70–0.80) for SOFA, which was also not statistically significant (adjusted p = 0.756). SOFA-2 did not differ significantly from NEWS-L for either time point; it outperformed MEWS for both 1-day and 7-day mortality and outperformed NEWS for 7-day mortality, while the difference from NEWS for 1-day mortality was not significant following correction.
When evaluated at the intermediate-or-high-risk threshold, SOFA-2 exhibited lower sensitivity than the original SOFA score for 1-day mortality (86.0% versus 100.0%) and 7-day mortality (67.4% versus 92.9%), alongside higher specificity (59.2% versus 20.6% at 1 day and 68.8% versus 25.9% at 7 days). The authors emphasized that the lower sensitivity of SOFA-2 at this cutoff is a direct function of the non-validated thresholds applied rather than evidence of diminished clinical utility, concluding that larger multicenter studies are required to provide external validation.
Regarding discriminative capacity, SOFA-2 achieved an area under the receiver operating characteristic curve of 0.83 (95% CI 0.77–0.89) for 1-day in-hospital mortality and 0.76 (95% CI 0.70–0.81) for 7-day mortality. The original SOFA score yielded an area under the curve of 0.82 (95% CI 0.76–0.89) for 1-day mortality and 0.75 (95% CI 0.70–0.80) for 7-day mortality. Pairwise comparisons confirmed that these differences were not statistically significant, with an adjusted p-value of 0.526 at 1 day and 0.756 at 7 days.
Beyond original SOFA, the study evaluated SOFA-2 alongside early warning scores derived from the initial emergency department assessment. SOFA-2 significantly outperformed MEWS for both 1-day and 7-day mortality. Compared with NEWS, SOFA-2 demonstrated statistically superior discrimination for 7-day mortality, whereas the difference for 1-day mortality was not statistically significant after Holm–Bonferroni correction. Furthermore, pairwise testing revealed no statistically significant difference between SOFA-2 and NEWS-L for either outcome window.
The single-center retrospective cohort included 330 hospitalized emergency department patients aged 65 years or older with sepsis. All five scoring instruments, including SOFA-2, SOFA, NEWS, NEWS-L, and MEWS, were calculated from each patient’s initial emergency assessment. Overall mortality was documented in 43 patients (13.0%) within 1 day and in 141 patients (42.7%) within 7 days. Discrimination was evaluated using receiver operating characteristic curves, with pairwise comparisons performed by DeLong’s test and Holm–Bonferroni correction.
Threshold-based analysis at the intermediate-or-high-risk cutoff showed clear trade-offs between sensitivity and specificity across both scores. For 1-day mortality, SOFA-2 demonstrated lower sensitivity than original SOFA (86.0% vs. 100.0%) but higher specificity (59.2% vs. 20.6%). Similarly, for 7-day mortality, SOFA-2 exhibited lower sensitivity than SOFA (67.4% vs. 92.9%) alongside higher specificity (68.8% vs. 25.9%).
In their conclusion, the study authors stated that while SOFA-2 demonstrated good discrimination for short-term mortality, it did not improve prognostic performance compared with the original SOFA score in this cohort. They stressed that the lower sensitivity observed for SOFA-2 at the intermediate-or-high-risk threshold is a direct function of applying non-validated cutoffs and should not be interpreted as evidence of reduced clinical utility. Because performance in older adult emergency department patients with sepsis had remained uncertain, the investigators emphasized that larger multicenter studies are needed to provide external validation.
Subject of Research: Comparative performance of SOFA-2 for short-term mortality prediction in older emergency department patients with sepsis
Article Title: SOFA-2 vs. SOFA and early warning scores for short-term mortality in older adults with sepsis in the emergency department
Article References: Satici, M. O., Gündüz, F. N. E., Küka, B., Satici, C., & Özdemir, S. (2026). SOFA-2 vs. SOFA and early warning scores for short-term mortality in older adults with sepsis in the emergency department. BMC Emergency Medicine. https://doi.org/10.1186/s12873-026-01807-1
Image Credits: AI Generated
DOI: 10.1186/s12873-026-01807-1
Keywords: SOFA-2, SOFA score, sepsis, emergency department, older adults, mortality prediction, early warning scores, in-hospital mortality, SOFA, early, warning, scores
News Source: Ophelia Keating. (November 9, 2026). Updated SOFA-2 Score Matches Original for Predicting Short-Term Sepsis Mortality. Scienmag.



