Background/Purpose: Interstitial lung disease (ILD) is a leading cause of premature mortality in systemic sclerosis (SSc). Immunosuppression is used for treatment of established disease. However, little is known about whether immunosuppression might prevent ILD in SSc. The aim of this study was to determine if, in SSc patients without ILD, immunosuppression (given for other manifestations) was associated with a lower risk of new onset SSc-ILD.

Methods: A retrospective cohort of 887 SSc patients without ILD at baseline was studied, using data from the Canadian Scleroderma Research Group registry. The primary exposure of interest was immunosuppression with methotrexate, cyclophosphamide, mycophenolate and/or azathioprine. The primary outcome variable was new onset ILD defined using a published algorithm (Steele, ACR 2012). Time to primary outcome was compared between exposed and unexposed subjects, modeled as a time-dependent exposure variable, using an unadjusted Kaplan-Meier model and a marginal structural model incorporating inverse probability of treatment weights (IPTW) to account for confounding. Weights were constructed using variables most likely to influence a decision to initiate immunosuppression, namely age, sex, ethnicity, disease duration, anti-centromere, anti-topoisomerase I, and anti-RNA polymerase III status, modified Rodnan skin scores, C-reactive protein levels, forced vital capacity, presence of inflammatory arthritis, presence of inflammatory myositis, and exposure to immunosuppressive therapy in the past. Subjects were censored at the visit when ILD was first recorded, or time of death, permanent study drop-out or last study visit. Both robust asymptotic and non-parametric bootstrap confidence intervals were constructed.

Results: The study included 218 subjects exposed to immunosuppression at baseline or during follow up and 669 unexposed subjects. Baseline characteristics of exposed and unexposed subjects are presented in Table 1. In unadjusted Kaplan Meier analysis, the observed risk of ILD was higher in exposed compared to unexposed subjects (log rank p <.0001), consistent with confounding. However, in a multivariate analysis incorporating IPTW, subjects exposed to immunosuppression had a lower estimated risk of developing ILD compared to unexposed subjects: weighted hazard ratio (HR) 0.50 (95% CI 0.28, 0.90, p=0.021), and, after bootstrapping, 0.50 (95% CI 0.17, 0.99).

Conclusion: To our knowledge, this is the first study to demonstrate a role for immunosuppression in preventing SSc-ILD in an observational cohort using modern causal statistical methods. In subjects perceived to be at increased risk for developing ILD, there may be value in initiating immunosuppression early, as this may alter disease course and potentially outcomes. Table 1. Baseline characteristics of the cohort separated by exposure status (N=887)