A108-05 The Impact of Extracorporeal Membrane Oxygenation and Reduced Inspired Oxygen Fraction on Lower Airway Microbiome Dysbiosis in Severe Covid-19

Abstract Rationale Extracorporeal membrane oxygenation (ECMO) enables lung-protective ventilation and lower inspired oxygen fraction (FiO2) in ARDS. High FiO2 and mechanical ventilation are known to alter the lower-airway microbiome, yet it is unclear whether FiO2 reduction through the use of ECMO mitigates such effects. We evaluated whether reduced FiO2 exposure among ECMO-treated patients was associated with greater microbial diversity and enrichment of facultative anaerobes. Methods We conducted a prospective longitudinal cohort study of 175 critically ill adults with COVID-19 pneumonia requiring mechanical ventilation at NYU Langone and Bellevue Hospitals (2020-2022). Forty patients received ECMO and 135 did not. Lower-airway samples were collected via bronchoalveolar lavage (BAL) or tracheal aspirate; only baseline samples (one per patient) were analyzed. Metatranscriptomic RNA sequencing enabled taxonomic classification and diversity analyses. The median FiO2 over the preceding 48 hours was used, and patients were stratified above or below the cohort median of 50%. Differential expression was assessed with edgeR (FDR 0.2). Results ECMO patients exhibited lower FiO2 requirements during sampling across weeks 1-2 (median FiO2 40% 40-57.5% vs non-ECMO 51.25% 40-67.5%), consistent with current trends in oxygenation support during ECMO. Across all patients, neither β-diversity nor Shannon diversity differed significantly between those above and below the median FiO2 of 50%. However, differential expression analyses revealed compositional trends between ECMO and non-ECMO patients. In non-ECMO patients, higher FiO2 exposure was associated with enrichment of aerobic and opportunistic taxa (Haemophilus influenzae, Proteus spp., Aggregatibacter, Serratia, and Rockefellervirus IPLA5), whereas lower FiO2 samples showed greater representation of facultative anaerobes and oral commensals (Cupriavidus, Corynebacterium, and Parvimonas). In contrast, among ECMO patients, FiO2-associated patterns were directionally consistent with those observed in non-ECMO patients—such as enrichment of Serratia and other opportunistic taxa under higher FiO2—but with attenuated magnitude of change and overall more balanced community composition. These findings suggest that ECMO, likely through a reduction in FiO2, mitigates oxygen-driven microbial shifts observed in conventionally ventilated patients. Conclusion Among critically ill COVID-19 patients, lower alveolar oxygen exposure—characteristic of ECMO support—was associated with preservation of lower-airway microbial composition and attenuation of FiO2-linked dysbiosis. High FiO2 favored aerobic and opportunistic taxa in non-ECMO patients, while ECMO maintained a more stable, commensal-dominant microbiome. Future analyses will use propensity matching to compare high-FiO2 non-ECMO patients with low-FiO2 ECMO counterparts, emulating the impact of oxygen reduction in comparable clinical contexts. This abstract is funded by: None