Predictors of long-term survival in patients with idiopathic pulmonary fibrosis: data from the IPF-PRO Registry

PURPOSE: We used data from the IPF-PRO Registry of patients with idiopathic pulmonary fibrosis (IPF) to identify characteristics that predicted survival for a further > 5 years. METHODS: Participants had IPF that was diagnosed or confirmed at the enrolling center in the previous 6 months. Patients were followed prospectively. A Classification And Regression Tree (CART) was used to identify predictors of survival > 5 versus ≤ 5 years following enrollment. The following variables, assessed at enrollment, were considered: age; body mass index (BMI); former smoker; current smoker; time from first imaging evidence, symptoms, or diagnosis of IPF to enrollment; forced vital capacity (FVC) % predicted; diffusing capacity of the lungs for carbon monoxide (DLco) % predicted; antifibrotic drug use; supplemental oxygen use; history of cardiac disease; pulmonary hypertension; COPD/emphysema; and rural location. RESULTS: The analysis cohort comprised 819 patients, of whom 278 (33.9%) survived > 5 years. DLco % predicted, supplemental oxygen use and FVC % predicted were the most important variables for predicting survival > 5 versus ≤ 5 years after enrollment. The importance of these variables (scaled such that the most important had an importance of 100%) was 100%, 78.2% and 74.2%, respectively. The optimism-corrected area under the curve (AUC) of the CART was 0.72, with an accuracy of 0.72. CONCLUSION: Among patients enrolled in the IPF-PRO Registry, a decision tree that included DLco % predicted, oxygen use and FVC % predicted facilitated the prediction of survival > 5 years. Understanding predictors of longer-term survival may facilitate conversations with patients about their prognosis and treatment.