Introduction: Extracorporeal membrane oxygenation (ECMO) utilization for acute respiratory distress syndrome (ARDS) remains controversial following equivocal trial results. Real-world effectiveness across diverse healthcare settings is unknown. We evaluated ECMO outcomes for ARDS using a nationally representative database, examining whether hospital ECMO volume modifies treatment effectiveness. Methods: We analyzed the National Inpatient Sample (2016-2022), identifying adult ARDS hospitalizations using validated ICD-10 codes. The primary outcome was in-hospital mortality. We employed propensity score matching, with >40 variables including demographics, comorbidities, organ failures, and hospital characteristics with interaction terms. Hospitals were stratified by annual ECMO volume (≥5 cases=high). Subgroup analyses examined ECMO type, cardiac arrest, and COVID-19 status. Economic outcomes included total charges and length of stay. Statistical analysis used survey-weighted regression for the NIS sampling design. Grammarly was used to edit the abstract. Results: Among 82, 948 ARDS hospitalizations (representing 414, 740 weighted cases nationally), 3, 011 (3. 6%) received ECMO. After propensity matching (2, 980 ECMO patients matched), overall mortality was similar (ECMO 45. 8% vs control 45. 4%, p=0. 78). However, hospital volume significantly modified the treatment effect. At high-volume centers (n=2, 200), ECMO improved survival (OR 1. 31, 95%CI 1. 08-1. 59, p=0. 005), while at low-volume centers (n=3, 329), ECMO worsened survival (OR 0. 87, 95%CI 0. 75-0. 99, p=0. 046). VA-ECMO showed higher mortality than VV-ECMO (OR 1. 87, 95%CI 1. 47-2. 38). Cardiac arrest patients experienced worse outcomes with ECMO (interaction p< 0. 001). ECMO increased costs by 700, 908 per patient and length of stay by 11. 2 days. COVID-19 diagnosis did not modify ECMO effectiveness (p=0. 77). Conclusions: ECMO demonstrated no overall survival benefit for ARDS in this large real-world analysis, with substantial costs. Critical effect modification by hospital volume suggests ECMO improves survival only at experienced centers while potentially causing harm at low-volume hospitals. These findings support regionalizing ECMO services to high-volume centers and emphasize careful patient selection, particularly avoiding ECMO in cardiac arrest patients.
Castillo et al. (Sun,) studied this question.