Multi-modal data to identify key factors influencing lung injury in ARDS patients undergoing invasive mechanical ventilation: A prospective multi-center observational study protocol

Abstract

Background Patients with moderate to severe acute respiratory distress syndrome (ARDS) exhibit extremely poor prognoses following mechanical ventilation, with mortality rates as high as 40% to 55%. Despite extensive research into ARDS classification and prognostic assessment, the disease’s pathogenesis remains incompletely understood, and there remains a critical lack of specific biomarkers and effective therapeutic targets for its prevention and management. The core challenges lie in two key areas. First, ARDS demonstrates marked heterogeneity in etiology, pathophysiology, and pathogenesis. Existing research, predominantly reliant on population-level average data, fails to capture inter-individual variability, hindering the precise identification of patient subgroups responsive to specific therapeutic regimens. Second, current definitions of ARDS phenotypes are often confined to clinical symptoms and routine diagnostic indices, lacking integrated analysis of deeper mechanistic indicators, such as key biomarkers and respiratory mechanics parameters, thereby limiting the stability and clinical utility of existing classification systems. Methods/Design We designed a prospective multicenter cohort study incorporating multi-omics analyses. This research aims to investigate the mechanisms underlying the development and progression of ARDS during mechanical ventilation, providing a theoretical foundation and practical guidance for future ARDS therapies. The study plans to enroll over 165 patients with moderate to severe ARDS receiving mechanical ventilation across 10 medical centers. Peripheral blood and bronchoalveolar lavage fluid (BALF) samples will be collected on the first 24 hours after enrollment and at extubation for metagenomic/meta-transcriptomic sequencing, bulk RNA sequencing, single-cell RNA sequencing, proteomics detection, and metabolomics analyses. Concurrently, comprehensive monitoring of physiological indices, electrical impedance tomography, transpulmonary pressure, pulmonary ultrasound findings, and other relevant parameters will be conducted during the enrollment. Study participants will be stratified by survival and mortality outcomes to analyze the dynamic trends of all measured indices and their underlying molecular mechanisms. Biomarkers derived from multi-omics data and clinical baseline characteristics will be evaluated and integrated, followed by multidimensional dimensionality reduction. Predictive models will be subsequently constructed via early or late fusion to identify core prognostic markers, with performance validated using standardized metrics. Discussion Through comparative analysis of multi-omics data, we aim to identify specific markers and risk factors associated with distinct clinical trajectories of ARDS, further clarifying the key determinants of lung injury. Ultimately, this research will reveal critical immune cell subtypes that govern ARDS onset and prognosis, offering novel insights and therapeutic targets to advance precision medicine for ARDS. Study protocol registration ClinicalTrials.gov NCT05922826 .