Identifying the molecular changes of severe COVID-19 infected lungs and the long-term outcomes of patients supported on VV- ECMO undergoing lung transplantation

We will perform an unbiased survival analysis using Kaplan-Meier of all COVID-19 and non-COVID- 19 patients undergoing transplants at our centers during the same study period. Additionally, we will perform a risk-adjusted analysis between the two cohorts. Finally, we will match our COVID-19 cohort with the SRTR database. The variables in common between the SRTR database and our COVID-19 transplant patients’ data at our centers will be selected and propensity score model will be created as shown (Figure 6). From the SRTR database, patients transplanted after June 2005, following the implementation of LAS score, will be included. We will choose a matching propensity score algorithm instead of a stratification or weighting approach because the control group of non-COVID-19 recipients in the SRTR will be considerably larger in our center. A logistic regression based nearest neighbor approach will be utilized to estimate the propensity score. The computations will made using RStudio (Version 1.4.1103) and MatchIt (4.1.0). To demonstrate feasibility of this approach, we performed a sample statistical modeling of the first 21 consecutive recipients undergoing double lung transplant for severe COVID-19 ARDS (Figure 6) and matched SRTR patients.