BMC Pulm Med. 2026 Jul 11. doi: 10.1186/s12890-026-04422-9. Online ahead of print.
ABSTRACT
BACKGROUND: Community-acquired pneumonia (CAP) constitutes a significant cause of hospital admissions and mortality globally. Accurate assessment of CAP severity and prognosis is crucial for optimizing treatment strategies. This study investigates the utility of the Systemic Immune-Inflammation Index (SII), Pan-Immune Inflammation Value (PIV), and modified Glasgow Prognostic Score (mGPS) in evaluating CAP severity and predicting patient outcome.
METHODS: Conducted a retrospective analysis of our medical records of 1879 patients with CAP who were admitted between January 2021 and December 2023. The patients were categorized into a mild disease group (n = 595, 31.66%) and a severe disease group (n = 1284, 68.33%) using the Pneumonia Severity Index (PSI). Additionally, based on the prognosis during the first 28 days, the patients were divided into a good prognosis group (n = 1618, 86.11%) and a poor prognosis group (n = 261, 13.89%) based on their prognosis within the first 28 days. The mean values of the SII, PIV, and mGPS were calculated and statistically analyzed to determine their association with disease severity and prognosis. Model comparison with the PSI and assessment of incremental prognostic value were performed using hierarchical logistic regression and receiver operating characteristic curve analyses.
RESULTS: In our study, we found that the levels of SII and PIV are much higher in patients with severe CAP. These two indices have a positive correlation with bad outcomes (SII: rho = 0.198, P < 0.001; PIV: rho = 0.154, P < 0.001). Multivariate logistic regression identified mGPS as the strongest independent predictor of severe CAP (OR = 18.342, P < 0.001), along with respiratory rate, diastolic blood pressure, and PIV. For poor prognosis (28-day mortality), independent predictors included BMI, cardiovascular disease, PSI score, mGPS (OR = 26.512, P < 0.001), heart rate, respiratory rate, body temperature, diastolic blood pressure, SII, and PIV. ROC analysis showed that SII (AUC = 0.665), PIV (AUC = 0.628), and mGPS (AUC = 0.677) exhibited moderate individual predictive accuracy for 28‑day mortality; however, adding these indices to the PSI improved the AUC from 0.672 to 0.712.
CONCLUSION: SII, PIV, and mGPS correlate with CAP severity and 28‑day outcomes. Although their standalone discriminative ability is only moderate, they provide incremental prognostic information when combined with established scores such as the PSI. These inexpensive, routinely available biomarkers may serve as adjunctive tools to enhance risk stratification, and their integration with clinical variables modestly improves prediction of adverse outcomes.
PMID:42436459 | DOI:10.1186/s12890-026-04422-9