Choice of Methodology Impacts Outcome in Indirect Comparisons of Drugs for Idiopathic Pulmonary Fibrosis

Background and Objectives: Idiopathic pulmonary fibrosis (IPF) is a chronic condition leading to lung damage and deterioration in lung function. Following the availability of two new drugs, nintedanib and pirfenidone, a number of network meta-analyses (NMAs) of randomised controlled trials have been published which have conducted indirect comparisons on the two drugs. Differing recommendations from these studies are potentially confusing to clinicians and decision-makers. We aimed to systematically review published NMAs of IPF treatments, to compare their findings and summarise key recommendations. Materials and Methods: We systematically reviewed (PROSPERO: CRD42017072876) six eligible NMAs and investigated the differences in their findings with respect to key endpoints. We focused on differences in head-to-head comparisons between nintedanib and pirfenidone. Results: The NMAs were broadly consistent, with most differences being explained by model choice, endpoint definitions, inclusion of different studies, different follow-up durations, and access to unpublished data. A substantive difference remained, however, in the change from baseline forced vital capacity (FVC). One NMA favoured nintedanib, another found no statistical difference, whilst others did not conduct the analysis. These differences can be attributed to the choice of methodology, the use of the standardised mean difference (SMD) scale, and population heterogeneity. Conclusions: NMA methods facilitated the comparison of nintedanib and pirfenidone in the absence of a head-to-head trial. However, further work is needed to determine whether the trial populations are homogeneous and whether the SMD is appropriate in this population. Differences in patient characteristics may obscure the difference in treatment effects. To assist decision-makers, an exploration of efficacy in real-world populations may be prudent.