# Novel pneumonia score based on Systemic Immune-inflammation Index and Prognostic Nutritional Index in elderly patients

## TL;DR

A new score combining immune and nutritional indicators improves prediction of 30-day mortality in elderly ICU patients with pneumonia.

## Contribution

A novel SII-PNI score is developed to predict mortality in elderly ICU patients with community-acquired pneumonia.

## Key findings

- The SII-PNI score predicted 30-day mortality with an odds ratio of 2.19 (95% CI: 1.62–2.95, p < 0.001).
- Kaplan-Meier analysis showed higher mortality in patients with a score of 2 compared to scores of 0 or 1.
- The model demonstrated consistent performance in both internal and external validation cohorts.

## Abstract

The mortality rate for elderly patients with community-acquired pneumonia (CAP) admitted to intensive care units (ICU) is high. The combination assessment of Systemic Immune-inflammation Index (SII) and Prognostic Nutritional Index (PNI) can provide a more comprehensive evaluation of the patient’s immune response, systemic inflammatory burden, and nutritional metabolic status.

From the Medical Information Mart for Intensive Care IV database (MIMIC-IV, version 3.1), we selected 12457 patients with CAP admitted to the ICU. After exclusions, 634 patients were included and randomly split into training (n = 444, 70%) and internal validation (n = 190, 30%). Meanwhile, an independent external validation cohort comprised 149 patients admitted to The Affiliated People’s Hospital of Ningbo University (January 2024–March 2025) was collected. Optimal thresholds for SII and PNI were derived from receiver operating characteristic (ROC) analysis in the training cohort, which were subsequently used to calculate the SII-PNI score. Model performance was evaluated through net reclassification improvement, decision-curve analysis, logistic regression analysis, and Kaplan-Meier curves. Validation was performed in internal and external cohorts to assess the model’s predictive value in geriatric CAP patients.

Receiver operating characteristic analysis determined the optimal cutoff values for SII (2030.28; AUC = 0.573, 95% CI 0.517–0.628, p < 0.05) and PNI (29.07; AUC = 0.638, 95% CI 0.584–0.692, p < 0.001) in the training cohort. The SII-PNI scoring model was subsequently developed using these thresholds and demonstrated predictive value for 30-day in-hospital mortality [OR: 2.19 (95% CI: 1.62–2.95), p < 0.001]. Kaplan-Meier survival analysis confirmed consistent prognostic performance across all cohorts, that patients with a score of 2 on the SII-PNI scale had significantly higher 30-day mortality compared to those with scores of 0 or 1 (p < 0.05).

The SII-PNI may serve as a adjunct for evaluating the 30-day mortality rate among elderly ICU patients admitted with CAP.