Clustering Acute Respiratory Distress Syndrome by Sex

Le résumé fourni par la source

Abstract RATIONALE. Sex differences have been observed in the incidence, severity, mortality, and outcomes of acute respiratory distress syndrome (ARDS) and COVID-19. However, the molecular basis for the differences between males and females remains unclear. METHODS. Bronchoalveolar lavage (BAL) samples collected from NIH-supported ARDSNet trials were provided by BioLINCC. To generate proteomics datasets, we purified extracellular vesicles (EVs) and then EV proteins, which were subsequently analyzed them using mass spectrometry. The proteomics datasets were precleaned and differentially expressed proteins were clustered to identify sex-specific endotypes. The prediction models for the identified endotypes were developed by training a univariate logistic regression model, and model performance was evaluated based on three-class accuracy.RESULTS. There were 57 ARDS patients and 49 healthy controls. We identified 6,616 proteins in BAL-EVs. Among these, the expression of 861 proteins was significantly upregulated or downregulated in patients compared to controls. K-means clustering of the demographics-correlated proteins revealed two distinct endotypes. These endotypes could be clearly differentiated based on gender, demographics, p/f ratio, smoking status, and various functional annotations. The prediction model developed based on proteomic datasets achieved an area under the curve (AUC) of 0.993 (sensitivity 94.59%, specificity 95.00%). When transformed into clinical settings, the new model derived from clinical variables showed an AUC value above 0.750. The prediction performance was improved if we combined omics data with clinical settings to develop the model (AUC 0.797), demonstrating the excellent performance of these three types of prediction models. CONCLUSION. Our study provides molecular mechanisms and prediction models for the sex differences in ARDS. Sex may serve as an independent predictor of the development, responses to treatments, mortality, and long-term outcomes associated with ARDS.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.