A Simple Platelet-Centered Coagulation–Inflammation Combined Score for Predicting In-Hospital Mortality in Sepsis

OBJECTIVE: Accurate mortality prediction in sepsis requires assessing coagulation and inflammatory disturbances, yet existing scores are complex or incompletely capture these interconnected processes. We developed and internally validated a simple score combining inflammatory and platelet-related indicators with clinical parameters to predict in-hospital mortality in sepsis. METHODS: This retrospective study included 6175 adults with sepsis from the MIMIC-IV database. Patients were randomly divided into training (n=4322) and validation (n=1853) cohorts. Candidate predictors were screened by least absolute shrinkage and selection operator regression, and final variables entered multivariable logistic regression to construct the Coagulation-Inflammation Combined Score (CICS). Performance was assessed by discrimination, calibration, internal bootstrap validation, and decision curve analysis. RESULTS: Of 6175 patients, 807 (13.1%) died during hospitalization. The final CICS incorporated five routinely available variables: lactate, Glasgow Coma Scale score, age, neutrophil-to-lymphocyte ratio, and platelet count. Discrimination was good, with area under the receiver operating characteristic curve 0.768 (95% CI, 0.747-0.788) in training and 0.806 (95% CI, 0.776-0.835) in validation cohorts, comparable to SOFA (DeLong P=0.526 and P=0.180). Calibration was acceptable (Brier scores 0.102 and 0.091). Bootstrap validation indicated minimal optimism, and decision curve analysis showed exploratory findings across clinically relevant thresholds. CONCLUSIONS: The CICS is a parsimonious prognostic score using five routinely available variables from the first 24 hours after ICU admission. It showed discrimination comparable to SOFA, acceptable calibration, and favorable decision-curve performance. CICS should be interpreted as a separate prognostic score for early ICU stratification once first-24-hour data are available, rather than for emergency department triage or ICU admission decisions.