LAUSR

Background Rheumatoid arthritis (RA)-associated interstitial lung disease (ILD) is associated with significant morbidity and is a critical cause of mortality in patients with RA. Objective Our aim was to evaluate predictive and prognostic factors for RA-ILD and to describe the therapeutic management of the condition from a large China cohort. Methods This was a retrospective cohort study. We collected data of 1121 RA patients who underwent chest HRCT from 2008 to 2017. Patients without ILD at RA diagnosis were included in the analysis. The development and evolution of ILD in RA patients were followed up. Determinants of ILD development and progression were identified through multivariable logistic analysis. Cox hazards analysis was used to determine significant variables associated with survival. Results A total of 923 patients without ILD at RA diagnosis were identified and enrolled. Among them, 278 cases (30.12%) were diagnosed as ILD during follow-up. Logistic regression analysis showed that advanced age (> 60 years old) at RA onset (OR: 1.485), male (OR: 1.882), short duration of RA (0~5 years) (OR: 2.099), RF positive (OR: 1.728), elevated lactate dehydrogenase (LDH) (OR: 3.032), and no medication (OR: 1.833) were closely correlated to the development of RA-ILD. No correlation was found between ILD development and traditional DMARDs such as methotrexate and leflunomide. According to the follow-up data, 83 RA-ILD patients were identified as interstitial lung disease (ILD) progression, and 102 participants were stable. Logistic regression modeling demonstrated that DLCO% < 45% (OR: 3.025) and UIP possible pattern on HRCT (OR: 3.476) were independent risk factors for the ILD progression. No correlation was found between ILD progression and traditional DMARDs such as methotrexate and leflunomide. A total of 53 RA-ILD deaths occurred during follow-up. Cox hazards analysis revealed that advanced age (> 60 years old) at RA-ILD diagnosis (HR: 3.181) and extensive lung involvement on HRCT (HR: 2.401) were associated with worse survival. Treatment with cyclophosphamide (HR: 0.210) was associated with better survival. Conclusions Advanced age, male, short duration of RA, RF positive, elevated LDH, and no medication are closely correlated with RA-ILD. No correlation was found between traditional DMARDs and ILD development. DLCO% < 45% and UIP possible pattern are predictive factors for ILD progression. No correlation was found between traditional DMARDs and ILD progression. Advanced age and extensive lung involvement on HRCT independently predict mortality; cyclophosphamide treatment helps to improve the prognosis of RA-ILD. Key Points • We designed this study to investigate the predictive and prognostic factors for RA-ILD and to explore the potential role of DMARDs in the evolution of RA-ILD from the development to progression and death. • Patients without ILD at RA diagnosis were enrolled and followed up retrospectively. • Our results showed that no correlation was found between traditional DMARDs and the development and progression of ILD, and regular treatment may improve the development of RA-ILD. • Our results revealed that clinical variables appeared predictive implications for the diagnosis of ILD and physiological and radiological variables appeared predictive implications for the prognosis of ILD, which can provide reference to rheumatologists and help to improve poor prognosis of RA-ILD.