Background:Progression of interstitial lung disease (ILD) is a candidate for long-term mortality in patients with systemic sclerosis (SSc). It is to date unclear, how the new definitions of progressive ILD predict mortality.

Objective:To estimate the prevalence of ILD progression applying different definitions and test their performance of predicting mortality.

Methods:We included all SSc-ILD patients from the Oslo and Zurich cohorts and defined disease progression as: (A)Absolute FVC decline >5% over 12 months;(B)PPF ATS guideline criteria;(C)INBUILD PF-ILD criteria. We assessed the prevalence of ILD progression using these competing definitions, applied multivariable cox regression models and compared the performance using Harrels c-index.

Results:In total, 231 SSc-ILD patients were included, with 31% showing FVC decline >5%, 19% fulfilling the PPF guideline and 39% the PF-ILD criteria. The number of deaths (47%) over mean 7.7 years (SD 3.9) were comparable in the different groups. The progression definitions performed differently in multivariable cox models adjusted for age, sex, disease duration, SSc subtype, extent of lung fibrosis, baseline FVC and treatment using FVC decline>5% (HR1.87, 1.10-3.17 95%CI; p=0.020; c-index 0.7331), PPF guideline (HR1.42, 0.79-1.84 95%CI; p=0.231; c-index=0.7156) and INBUILD PF-ILD (HR 2.38, 1.40-4.04 95%CI: p<0.001; c-index=0.7338). The models discriminating ability was not significantly different (p=0.138).

Conclusion:The prevalence of ILD progression varies depending on which definition was applied. FVC decline alone and PF-ILD criteria predicted mortality but not the PPF guideline criteria.