Modeling cystic fibrosis patient prognosis: Nomograms to predict lung transplantation and survival prior to highly effective modular therapy.

Background: The duration of time a person with cystic fibrosis (pwCF) spends on the lung transplant waitlist is dependent on waitlist and post-transplant survival probabilities and can extend up to 2 years. Understanding the characteristics involved with lung transplant and survival prognoses may help guide decision making by the patient, the referring CF Center and the transplant team.

Methods: This study seeks to identify clinical predictors of lung transplant and survival of individuals with CF using 29,847 subjects from 2003-2014 entered in the Cystic Fibrosis Foundation Patient Registry (CFFPR).

Results: Predictors significant (p ≤ 0.05) in the final logistic regression model predicting probability of lung transplant/death were: FEV1 (% predicted), BMI, age of diagnosis, age, number of pulmonary exacerbations, race, sex, CF-related diabetes (CFRD), corticosteroid use, infections with B. cepacia, P. aeruginosa, S. aureus, MRSA, pancreatic enzyme use, insurance status, and consecutive ibuprofen use for at least 4 years. The final Cox regression model predicting time to lung transplant identified these predictors as significant FEV1 (% predicted), BMI, age of diagnosis, age, number of pulmonary exacerbations, race, sex, CF-related diabetes (CFRD), corticosteroid use, infections with B. cepacia, P. aeruginosa, S. aureus, MRSA, pancreatic enzyme use, and consecutive ibuprofen use for at least 4 years. The concordance indices were 0.89 and 0.92, respectively.

Conclusions: The models are translated into nomograms to simplify investigation of how various characteristics relate to lung transplant and survival prognosis individuals with CF not receiving highly effective CFTR modulator therapy.