Development and validation of a prediction model for pulmonary infection in elderly patients with traumatic brain injury.

Objectives: This study aimed to investigate risk factors associated with pulmonary infection (PI) in elderly patients with traumatic brain injury (TBI). Additionally, this study sought to develop and validate a predictive model for PI in elderly patients with TBI using clinical data obtained upon admission.

Methods: The study retrospectively analyzed elderly patients (≥ 65 years) with TBI at Tangdu Hospital between January 2011 and December 2021. These patients were randomly allocated to training and validation sets in a 7:3 ratio. A nomogram model was developed to predict the risk of PI in elderly patients with TBI. Internal validation was conducted using a verification set, while external validation was performed using patient data from a different hospital.

Results: A total of 592 elderly patients with TBI were included. The Glasgow coma scale score on admission, chest injury, hemoglobin, albumin, C-reactive protein, procalcitonin, B-type natriuretic peptide, troponin, and surgery was found to be independent predictors of PI in elderly patients with TBI. The nomogram demonstrated good discrimination ability, with a consistency index of 0.918 (95% confidence interval (CI): 0.891-0.944), which was verified to be 0.848 (95% CI: 0.786-0.910). The area under the curve for the external validation cohorts was 0.836 (95% CI: 0.770-0.903).

Conclusions: This study developed and validated a prediction model for PI in elderly patients with TBI. The nomogram model demonstrated a favorable discriminatory and predictive capacity for predicting PI in elderly patients with TBI.