Advancing ICU mortality prediction in community-

Community-acquired pneumonia (CAP) is a leading cause of ICU admissions, with significant morbidity and mortality. Traditional risk stratification tools such as CURB-65, the Pneumonia Severity Index (PSI), and CT severity scores (CT-SS) are widely used for prognosis but could be improved by incorporating novel biomarkers. This retrospective study evaluated the fibrinogen-to-albumin ratio (FAR) as an additional predictor of 30-day mortality in ICU patients with CAP. A total of 158 CAP patients admitted to a tertiary care ICU were included. Baseline data encompassed demographic, clinical, laboratory, and radiological parameters, including FAR, CURB-65, PSI, and CT-SS. Logistic regression and ROC curve analyses were conducted to assess mortality predictors. The 30-day mortality rate was 70.88% (112/158). Higher FAR, PSI, CURB-65, CT-SS, and lactate levels were independently associated with increased mortality (p < 0.05). FAR demonstrated strong discriminatory power (AUROC: 0.704) and significantly improved the predictive accuracy of established models. Adding FAR to PSI increased the area under the receiver operating characteristic (AUROC) from 0.705 to 0.791 (p = 0.009), while combining FAR, CT-SS, and PSI yielded the highest predictive accuracy (AUROC: 0.844, p = 0.032). These findings suggest that FAR, which reflects both inflammation and nutritional status, complements traditional risk assessment tools by providing a dynamic perspective. Integrating FAR into existing models enhances the identification of high-risk patients, enabling timely interventions and more efficient resource allocation in the ICU.