Prognostic Value of 48-Hour Biomarker Reassessment Beyond Admission SOFA for 28-Day Mortality in Sepsis

Background/Objectives: Sepsis is clinically dynamic, and isolated admission biomarker values may insufficiently capture early biological evolution after treatment initiation. This study evaluated whether routine biomarker reassessment at approximately 48 h provides incremental prognostic information beyond admission Sequential Organ Failure Assessment (SOFA) score for 28-day mortality in sepsis. The analysis was framed as an exploratory 48 h landmark prognostic assessment among patients who were alive and had complete biomarker reassessment data at 48 ± 6 h. Methods: We conducted a prospective single-center observational cohort study including adult patients with sepsis. Clinical and laboratory data were collected at baseline (M1) and repeated 48 ± 6 h later (M2). The primary outcome was 28-day mortality. Candidate biomarkers included C-reactive protein (CRP), procalcitonin (PCT), lactate (LAC), and neutrophil-to-lymphocyte ratio (NLR). PCT clearance and NLR change were calculated as relative changes between M1 and M2, whereas 48 h CRP and 48 h lactate were evaluated as early reassessment values. Exploratory logistic regression models were constructed using admission SOFA as the clinical reference model. Model discrimination and fit were summarized using receiver operating characteristic analysis, likelihood-ratio testing, and Nagelkerke R²; the models were not intended as validated individual-level risk calculators. Results: The 48 h landmark analytical cohort included 126 patients, of whom 44 (34.9%) died within 28 days. Admission biomarker values showed limited prognostic signal. SOFA alone showed fair discrimination (AUC 0.740). Among the primary SOFA-augmented models, SOFA plus PCT clearance showed the highest discrimination and explanatory performance (AUC 0.810; Nagelkerke R² 0.332) and significantly improved model fit compared with SOFA alone. SOFA plus NLR change and SOFA plus 48 h lactate also provided incremental prognostic information, although their gains were more modest. In exploratory combined modeling, SOFA plus PCT clearance and NLR change provided the most coherent additional signal, with all predictors retaining independent associations with 28-day mortality. Conclusions: In this exploratory single-center 48 h landmark analysis, selected routine biomarker reassessment measures were associated with 28-day mortality beyond admission SOFA. PCT clearance provided the clearest incremental prognostic signal, while NLR change offered complementary information. Persistent 48 h lactate elevation was also informative, whereas lactate clearance was not. These findings should be interpreted as hypothesis-generating and require validation in larger cohorts, ideally including serial organ dysfunction measures such as 48 h SOFA or SOFA change.