Idiopathic Pulmonary Fibrosis (IPF) is a progressive, scarring lung disease with a poor prognosis and median survival of 3 years. It is a heterogeneous disorder with varying rates of progression, which presents a challenge for accurate prognostic prediction. The CPI¹ and the GAP² score are validated scoring systems for prognostic determination in IPF. However, it remains unclear which is the optimal method for accurate prediction of mortality in IPF. Therefore, a comparison of the predictive ability of the GAP score and CPI was undertaken in a cohort of IPF patients.

Methods Baseline data were collected retrospectively from 213 IPF patients (in accordance with ATS/ERS criteria and MDT consensus) from a single centre in the UK between 19^th April 2007 and 14^th July 2014. Thirty-eight patients were excluded, either because pulmonary function test results could not be obtained or the patient received a lung transplant during follow-up. Gender, age, FVC and DLCO were used to calculate the GAP score whilst FVC, FEV1 and DLCO were used for the CPI. Spearman’s correlation was used to analyse the relationship between GAP score or CPI and survival time. The ability of the scoring systems to predict survival at 1 and 3 years was assessed using ROC curve analysis.

Results Of the 175 patients, 131 (75%) were male with a mean age of 71 ± 8 years (mean ± SD) at presentation. Overall 3-year mortality was 50%. The CPI demonstrated a better correlation with survival (r² = 0.37, p < 0.01) compared to the GAP score (r² = 0.24, p < 0.01). ROC curve analysis for 1-year mortality found that area under curve (AUC) was 0.726 for the GAP score and 0.783 for the CPI. For 3-year mortality AUC was 0.749 for the GAP score and 0.805 for the CPI.

Conclusion These data show that the CPI more accurately predicts survival at presentation and 3-years than the GAP score in a UK IPF cohort.

• Download figure
• Open in new tab
• Download powerpoint

Abstract S44 Figure 1

Correlation of GAP score (1a) and CPI (1b) and survival in IPF

References 1 Wells AU, et al. Am J Resp Crit Care Med. 2003;167:962

2 Ley B, et al. Ann Intern Med. 2012;156:684