Predicting mortality in idiopathic pulmonary fibrosis. Which parameters should be used to determine eligibility for treatment? Analysis of a UK prospective cohort

Abstract

Introduction: The burden of mortality from Idiopathic Pulmonary Fibrosis (IPF) is significant (1), however treatment with Pirfenidone remains restricted to patients with an FVC of 50-80% by National Institute for Health and Care Excellence (NICE) (2). Aim: An analysis of established prognostic tools such as single lung function parameters, the Gender Age Physiology (GAP) index and Composite Physiologic Index (CPI) will be validated in a cohort of IPF patients to determine which parameter best predict mortality. Method: 209 patients with IPF were enrolled in to a prospective cohort study at a UK tertiary centre. Baseline GAP Score, CPI, TLCO, FVC and FEV1 were analysed using Cox Regression and calculation of C-Statistic. Results: Multivariate analysis demonstrated the C Statistic was highest for GAP (C=0.7596) then CPI (C=0.7559). However hazards ratio (HR) demonstrated the TLCO to be the only significant parameter in each model (p= Univariate analysis found Sex and Age, the distinguishing variables of GAP, to be poor predictors of mortality, C=0.5137 and C=0.5795 respectively. TLCO was the best predictor of mortality C=0.7518 followed by FVC (C=0.6765) and FEV1 (C=0.6522). Conclusions: Evidence suggests TLCO can predict mortality more accurately. Comorbidities such as emphysema and PHT can preserve lung volume whilst depreciating TLCO beyond that which can be attributed to fibrosis. IPF patients can therefore be ineligible for treatment with Pirfenidone on the NHS, with current NICE guidance, whilst having a higher risk of mortality. TLCO or CPI should be evaluated especially if there is evidence of emphysema or PHT.