Development of a prediction model for progression of rheumatoid arthritis-associated interstitial lung disease using serologic and clinical factors: The prospective KORAIL cohort

Abstract

Objective: To develop a prediction model for rheumatoid arthritis-associated interstitial lung disease (RA-ILD) progression. Methods: We investigated predictors of RA-ILD progression in the Korean RA-ILD (KORAIL) cohort, a prospective study that enrolled patients with RA meeting ACR/EULAR criteria and ILD on chest computed tomography (CT) scans and followed for 3 years. Pulmonary function tests (PFTs) and chest CT scans were conducted annually. RAILD progression was defined as both physiological and radiological worsening, adapted from the 2023 ATS/ERS/ JRS/ALAT definition of progressive pulmonary fibrosis. Baseline factors included clinical factors and biomarkers (autoantibodies, inflammatory markers, and pulmonary damage markers). Results: We analyzed 138 RA-ILD patients (mean age 66.4 years, 30.4 % male, 60.1 % usual interstitial pneumonia [UIP] pattern). During a median follow-up of 2.9 years, 34.8 % (n = 48) had RA-ILD progression. Baseline associations with progression included: UIP pattern, ILD extent >10 %, DLCO %pred., anti-cyclic citrullinated peptide (anti-CCP), Krebs von den Lungen-6 (KL-6), and human surfactant protein D. We developed prediction models using UIP pattern, ILD extent, DLCO % pred., and anti-CCP titer with or without serum KL-6 levels. The models had areas under the curve (AUCs) of 0.73 and 0.75, respectively. The high-risk group had a positive predictive value for progression of 85.7 %, while the low-risk group had a negative predictive value of 94.7 %. Conclusion: In this prospective cohort, UIP pattern, ILD extent, lower DLCO, RA disease activity, anti-CCP levels, and pulmonary damage biomarkers were associated with RA-ILD progression. We developed prediction models that may be clinically useful to risk stratify once externally validated.