Multi-Omics Integrative Analysis of Serum and Sputum to Uncover Key Determinants of Prognosis in Patients With ARDS

Multi-Omics Integrative Analysis of Serum and Sputum to Uncover Key Determinants of Prognosis in Patients With ARDS

This study is a prospective observational investigation designed to systematically characterize key microbial signatures, metabolite profiles, and gene-expression features in serum and sputum from patients with acute respiratory distress syndrome (ARDS), and to evaluate their associations with response to invasive mechanical ventilation and clinical outcomes. A total of 411 adult ARDS patients meeting the Berlin definition and receiving invasive mechanical ventilation will be enrolled; individuals with significant pre-existing pulmonary disease or a history of immunosuppression will be excluded. Blood and sputum specimens will undergo high-throughput sequencing and metabolomics, including 16S rRNA-based microbiome profiling, whole-transcriptome RNA sequencing, and targeted/untargeted metabolite quantification by LC-MS/MS. Ventilator-related parameters (e.g., tidal volume, positive end-expiratory pressure [PEEP], and respiratory system compliance) and clinical endpoints (e.g., 28-day mortality) will be integrated to establish a comprehensive multi-omics analytical framework.

Data preprocessing will include batch-effect correction, normalization, and multiple-testing adjustment with false discovery rate control (FDR < 0.05). Differential microbes, metabolites, and transcripts will be functionally interpreted using KEGG pathway and Gene Ontology (GO) enrichment analyses. Spearman correlation analyses will be performed to examine associations between omics features and ventilator parameters. Key features will be selected using LASSO regression, followed by development of random forest and support vector machine (SVM) models to predict mechanical ventilation response and the risk of ventilator-induced lung injury (VILI). Model performance will be assessed using ROC curves, area under the curve (AUC), calibration plots, and decision curve analysis.

By integrating serum and sputum multi-omics data, this study aims to identify molecular biomarkers that influence the effectiveness of mechanical ventilation and prognosis in ARDS, thereby providing evidence to support precision ventilatory strategies and individualized clinical management to improve patient outcomes. The findings are expected to deepen mechanistic understanding of ARDS pathobiology and lay a foundation for future development of multi-omics-guided diagnostic and therapeutic approaches.

Inclusion Criteria:

• Written and dated informed consent provided.
• Age ≥18 years.
• ARDS diagnosed according to the 2012 Berlin definition.
• Clinically stable
• eligible for enrollment after 12-24 hours of stability.
• Baseline oxygenation index (PaO2/FiO2 ≤300 mmHg)
• No history of major chronic lung disease and no history of immunosuppression.
• Able to comply with study procedures, including collection of blood/plasma and sputum samples and clinical data.

Exclusion Criteria:

• Severe autoimmune disease or active systemic inflammatory disease.
• Severe hepatic dysfunction or renal failure requiring dialysis.
• Invasive mechanical ventilation for >36 hours prior to enrollment, or death prior to enrollment.
• Expected death within 24 hours of admission or inability to complete study-related assessments due to imminent death.
• Presence of a cardiac pacemaker or other implanted electronic device that may affect respiratory function assessment.
• Pregnant or breastfeeding women.
• Severe allergy to materials required for sample collection or other relevant contraindications.
• Severe psychiatric illness or inability to cooperate with study procedures.
• Prior participation in other interventional clinical trials that could affect interpretation of study results.
• Refusal of consent or withdrawal of informed consent.