Abstract
BACKGROUND: Mortality for patients receiving extracorporeal membrane oxygenation (ECMO) for COVID-19 increased over the course of the pandemic. We investigated the association between immunomodulators and mortality for patients receiving ECMO for COVID-19. METHODS: We retrospectively analysed the Extracorporeal Life Support Organisation registry from 1 January, 2020, through 31 December, 2021, to compare the outcomes of patients who received no immunomodulators, only corticosteroids, only other immunomodulators (selective interleukin blockers, janus-kinase inhibitors, convalescent plasma, and intravenous immunoglobulin), and a combination of corticosteroids and other immunomodulators administered either before or during ECMO. We used Cox regression models to estimate survival time until 90 days. We estimated the propensity score of receiving different immunomodulators using multinomial regression, and incorporated these scores into the regression models. RESULTS: We included 7181 patients in the final analysis; 6169 patients received immunomodulators either before or during ECMO. The 90-day survival was 58.1% (95%-CI 55.1-61.2%) for patients receiving no immunomodulators, 50.7% (95%-CI 49.0-52.5%) for those receiving only corticosteroids, 62.2% (95%-CI 57.4-67.0%) for those receiving other immunomodulators, and 48.5% (95%-CI 46.7-50.4%) for those receiving corticosteroids and other immunomodulators. Compared to patients without immunomodulators, patients receiving either corticosteroids alone (HR: 1.13, 95%-CI 1.01-1.28) or with other immunomodulators (HR: 1.21, 95%-CI: 1.07-1.54) had significantly shorter survival time, while patients receiving only other immunomodulators had significantly longer survival time (HR: 0.79, 95%-CI: 0.66-0.96). The receipt of immunomodulators (across all three groups) was associated with an increase in secondary infections. CONCLUSIONS: In this cohort study, we found that immunomodulators, in particular corticosteroids, were associated with significantly higher mortality amongst patients receiving ECMO for COVID-19, after adjusting for potential confounding variables and propensity score. In addition, patients receiving corticosteroids with or without other immunomodulators had longer ECMO runs, which has potential implications for resource allocation. While residual confounding likely remains, further studies are required to evaluate the timing of immunomodulators and better understand the possible mechanisms behind this association, including secondary infections.