Abstract
Background: Elderly patients infected with SARS-CoV-2 are at higher risk of developing cytokine storm and severe outcomes; however, specific immunological and proteomic biomarkers for early prediction remain unclear in this vulnerable group. Methods: We enrolled 182 elderly COVID-19 patients from the Chinese PLA General Hospital between November 2022 and April 2023, categorizing them based on progression to respiratory failure requiring mechanical ventilation (defined as severe progression). Olink proteomic analysis was performed on admission serum from 40 propensity score-matched samples, with differentially expressed proteins (DEPs) validated by cytometric bead array (CBA) in 178 patients. To predict severe progression, a model was developed using a 70% training set and validated on a 30% validation set. LASSO regression screened features followed by logistic regression and receiver operating characteristic (ROC) analysis to optimize the model by incrementally incorporating features ranked by random forest importance. Results: Elderly patients progressing to severe COVID-19 exhibited early immune dysregulation, including neutrophilia, lymphopenia, monocytopenia, elevated procalcitonin (PCT), C-reactive protein (CRP), interleukin-6 (IL-6), neutrophil-to-lymphocyte ratio (NLR), and systemic immune-inflammation index (SII), as well as coagulation dysfunction and multi-organ injury. Proteomics identified a set of biomarkers, including tumor necrosis factor-related apoptosis-inducing ligand (TRAIL), and revealed disruptions in signaling pathways, including the mTOR and VEGF signaling pathways. The optimal predictive model, which incorporated PCT, IL-6, monocyte percentage, lymphocyte count, and TRAIL, achieved an area under curve (AUC) of 0.870 (0.729-1.000) during validation. TRAIL levels negatively correlated with fibrinogen (p < 0.05). Conclusions: Elderly COVID-19 patients with severe progression demonstrate early immune dysregulation, hyperinflammation, coagulation dysfunction, and multi-organ injury. The model we proposed effectively predicts disease progression in elderly COVID-19 patients, providing potential biomarkers for early clinical risk stratification in this vulnerable population.