A Combined Inflammatory-Nutritional Index Model for Early Prognosis Prediction in Sepsis: Evidence From Logistic Regression and Decision Curve Analysis
Ein kombiniertes Entzündungs-Nährstoff-Index-Modell zur frühen Prognoseabschätzung bei Sepsis: Evidenz aus logistischer Regression und Decision-Curve-Analyse
British Journal of Hospital Medicine · 03.02.2026 · 87 (2)
DOI: 10.31083/bjhm50658 ↗Auf einen Blick
Bei 324 Patient:innen mit Sepsis waren höherer SII und NPS mit höherer 28-Tage-Mortalität verbunden. Ein Nomogramm aus beiden Parametern zeigte eine AUC von 0,846.
Fragestellung
Wie gut sagen SII und NPS das 28-Tage-Überleben bei Sepsis voraus, einzeln und kombiniert?
Methode
Retrospektive Studie mit Patient:innen mit Sepsis aus der ICU eines Krankenhauses; Analyse klinischer Daten von 2021 bis 2023; logistische Regression, ROC-Kurven, Kalibrierung und DCA.
Ergebnisse
324 Fälle. Höherer SII und NPS ≥2 waren mit erhöhter 28-Tage-Mortalität assoziiert und unabhängige Risikoprädiktoren. Das kombinierte Modell erreichte eine AUC von 0,846 und war den Einzelmodellen überlegen.
Schlussfolgerung der Autor:innen
SII und NPS sind robuste, unabhängige Prädiktoren der kurzfristigen Mortalität bei Sepsis. Das daraus entwickelte Nomogramm ermöglicht eine praktische, datenbasierte individuelle Risikovorhersage.
Kernaussagen
- Höherer SII und NPS ≥2 waren mit höherer 28-Tage-Mortalität verknüpft
- Kombiniertes Modell: AUC 0,846
- Nomogramm zeigte gute Übereinstimmung und klinischen Nutzen
Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.
Kurz gefasst
Fachartikel in der Zeitschrift British Journal of Hospital Medicine, online erschienen am 03.02.2026. Verfasst von Xianyuan Chen und 3 weiteren Autor:innen. Beteiligte Einrichtung: Department of Critical Care Medicine, The People's Hospital of Cangnan, Wenzhou Medical University, 325800 Wenzhou, Zhejiang, China. Hauptthemen laut MEDLINE: Sepsis, Nutrition Assessment, Inflammation. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Basic Research Project of Wenzhou City.
Themen: Sepsis
Abstract (Originalsprache)
Aims/Background: Sepsis remains a challenging condition with high short-term mortality despite advances in intensive care, emphasizing the significance of early prognostic assessment in guiding treatment decisions. Inflammatory and nutritional-immune indices are well-established predictors of clinical outcomes in critically ill patients. This study aims to evaluate the prognostic significance of the systemic immune-inflammation index (SII) and the Naples Prognostic Score (NPS) in predicting 28-day survival outcomes among individuals diagnosed with sepsis. The study also aims to construct and assess a visual predictive tool—specifically, a nomogram—that incorporates both of these biomarkers.
Methods
This retrospective study analyzed clinical data recorded from patients with sepsis who were treated in the intensive care unit (ICU) of The People’s Hospital of Cangnan between January 2021 and December 2023. Applying pre-determined inclusion-exclusion criteria, 324 cases were ultimately included in the final analysis. Comprehensive baseline data, including clinical features and laboratory findings, were systematically retrieved from the electronic health record system. Mortality-associated markers were identified within 28 days using univariate analyses followed by multivariable logistic regression, with SII included as a continuous variable and NPS as a categorical variable. The predictive performance of SII and NPS, both individual and in combination, was assessed using receiver operating characteristic (ROC) curve. A predictive nomogram was developed, and the accuracy and clinical utility of the model were then evaluated using calibration plots and decision curve analysis (DCA).
Results
The analysis revealed that higher SII values (p < 0.001) and NPS scores (≥2, p < 0.001) were closely associated with increased 28-day mortality. Multivariate logistic regression analysis identified SII (p < 0.001) and NPS (≥2, p < 0.001) as independent risk predictors. A predictive model was developed by combining both SII and NPS, demonstrating a superior area under the ROC curve [area under the curve (AUC): 0.846] compared to models utilizing either indicator alone. Furthermore, the nomogram that incorporated these two parameters exhibited high consistency between predicted probabilities and actual outcomes, while also demonstrating strong net clinical benefit in DCA.
Conclusion
SII and NPS are robust and independent predictors of short-term mortality in sepsis. The nomogram developed from these indicators offers a practical, data-driven approach to individualized risk prediction. This study highlights the clinical utility of integrating inflammatory and nutritional-immune indices in prognostic evaluation.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Substanzen
Autor:innen und Institutionen
- Department of Critical Care Medicine, The People's Hospital of Cangnan, Wenzhou Medical University, 325800 Wenzhou, Zhejiang, China.
Förderung
- Basic Research Project of Wenzhou City
Zitieren
APA
Chen, X., Li, K., Xue, S., & Liang, Y. (2026). A Combined Inflammatory-Nutritional Index Model for Early Prognosis Prediction in Sepsis: Evidence From Logistic Regression and Decision Curve Analysis. British Journal of Hospital Medicine, 87(2). https://doi.org/10.31083/bjhm50658
Vancouver
Chen X, Li K, Xue S, Liang Y. A Combined Inflammatory-Nutritional Index Model for Early Prognosis Prediction in Sepsis: Evidence From Logistic Regression and Decision Curve Analysis. British Journal of Hospital Medicine. 2026;87(2). doi:10.31083/bjhm50658
BibTeX
@article{Chen2026,
author = {Chen, Xianyuan and Li, Ke and Xue, Shengdong and Liang, Yan},
title = {A Combined Inflammatory-Nutritional Index Model for Early Prognosis Prediction in Sepsis: Evidence From Logistic Regression and Decision Curve Analysis},
journal = {British Journal of Hospital Medicine},
year = {2026},
volume = {87},
number = {2},
doi = {10.31083/bjhm50658}
}
EUVIMED