Background/Purpose: To describe the characteristics and morbidity and mortality impact of interstitial lung disease in rheumatoid arthritis (RA).

Methods: A retrospective review was performed at 2 high volume tertiary medical centers identifying patients with clinician determined RA associated interstitial lung disease (RA-ILD) in 1998-2012.  Data regarding demographics and disease characteristics of RA, including serologic status and erosions, were collected. Features of pulmonary disease, including successive radiographic findings and pulmonary function test (PFT) results, were abstracted.  Kaplan-Meier methods and Cox models were used to estimate survival rates and examine risk factors for mortality.

Results: A total of 119 patients were identified with RA-ILD (55% women; mean age at RA diagnosis 56.6 [range: 13.1-84.8] years; 63% former/current smokers; 75% rheumatoid factor positive).  Cyclic citrullinated antibodies were tested in 92 (79%) and present in 71/92 (77%).  Erosive disease was present in 32 patients (30%).  The mean time from RA diagnosis to ILD diagnosis was 9.4 [-10.9-43.0] years.  The mean follow-up time after ILD diagnosis was 3.5 [0.0-15.0] years.  Over follow-up, 51 patients (44%) required supplemental oxygen.  On repeat PFT, the mean percent predicted were 76.1% [40-148] for total lung capacity (TLC), 70.5% [32-141] for forced vital capacity (FVC), 70.0% [25-135], for forced vital capacity in one second (FEV1) and 51.9% [6-95] for diffusion capacity for carbon monoxide (DLCO).  33 patients died with cause of death attributed to pulmonary disease in 21%.  One-year survival rate was 78% (95% confidence interval (CI) 71-86%) with 5 year survival only 41% (95% CI 32-52%).  The most common types of ILD were usual interstitial pneumonia (UIP; 44 [37%]) and nonspecific interstitial pneumonia (NSIP; 42 [35%]).  Survival rates did not differ for those with UIP compared to those without (5 year survival rate 44% vs 40%; p=0.45; Figure). In patients with UIP compared to those without, the initial percent predicted DLCO (p=0.02) and the last available percent predicted TLC, FVC, FEV1, and DLCO were all significantly lower (p=0.003, 0.005, 0.02, 0.02, respectively).  In patients without UIP, the first FEV1, both absolute volume and percent predicted, were associated with mortality (p=0.027 and p=0.007, respectively, adjusted for age and sex).

Conclusion: In this large cohort, ILD is an important complication of RA with a very high mortality rate in the first 5 years following diagnosis.  The most common RA-ILD diagnoses were UIP and NSIP, and contrary to previously published data, there was no survival difference between these two groups.  Earlier in disease, a lower diffusion capacity may identify UIP RA-ILD compared to non UIP RA-ILD.  Further research is required to identify risk factors and biomarkers for this extraarticular manifestation of RA.