Predicting PGD after Lung Transplantation

Purpose Primary graft dysfunction (PGD) is the leading cause of early morbidity and mortality after lung transplantation. Prior studies highlight the role of donor smoking in poor post-transplant outcomes. We therefore aimed to identify clinical risk factors for PGD after lung transplantation, more accurately assess the impact of donor smoking, and develop an accurate, generalizable PGD prediction model. Methods We performed a 10-center prospective cohort study of patients in the Lung Transplant Outcomes Group from 2012-2018. PGD was defined as grade 3 PGD at 48 or 72 hours after lung reperfusion. We defined donor smoking history by clinical documentation, UNOS-reported donor smoking, or donor cotinine >9 mg/ml from urine collected at organ procurement. Building on our prior PGD prediction models, multivariable logistic regression was used to identify PGD risk factors and develop regularized predictive models with a simple user interface (R shiny). Results Of 1180 patients, 26% developed PGD and 633 (53%) received an organ from a previously smoking donor. The PGD predictive model included novel predictors of center and donor distance plus recipient age, lung allocation score (LAS), body mass index, pulmonary artery pressure, sex, and indication for transplant; donor age, sex, mechanism of death, and donor distance; and interaction terms for LAS and donor distance. The interface allows for real-time assessment of PGD risk for any donor/recipient combination. The predictive model offers decision making benefit in the PGD risk range of 0.15-0.5. Donor smoking was associated with increased risk of PGD but not patient mortality (Figure 1). Conclusion We developed a clinically applicable PGD predictive algorithm based on modern lung transplant practices to aid in transplant decision making, post-transplant care, and to facilitate development of enriched PGD treatment trials. Donor smoking is a risk factor for PGD development but is not associated with increased mortality, potentially because of a lower risk phenotype of PGD occurring in smoking donor recipients.