Introduction: The course of SSc-ILD is difficult to predict.

Aim: To identify factors associated with progression of SSc-ILD over 52 weeks in the SENSCIS trial.

Methods: Using logistic regression models, we analysed associations between baseline characteristics and absolute decline in FVC % predicted >5% or death over 52 weeks in all patients with SSc-ILD in the placebo group of the SENSCIS trial and in the subgroup who had risk factors for rapid FVC decline at baseline (<18 months since first non-Raynaud symptom, elevated inflammatory markers [C-reactive protein ?6 mg/L and/or platelets ?330 x 10⁹/L], or modified Rodnan skin score [mRSS] >18). In multivariate models, variables were selected via stepwise selection based on p-values. Mycophenolate use was a fixed variable. Model performance was assessed using the area under the receiver operating characteristic curve (AUC).

Results: In the placebo group overall (n=288), 86 (29.9%) had an absolute decline in FVC % predicted >5% or died over 52 weeks. The performance of the final model for predicting this outcome was weak (AUC of 0.635). In the subgroup with risk factors for rapid FVC decline at baseline (n=155), the final model for predicting this outcome had an AUC of 0.749 and included anti-topoisomerase I antibody (ATA) status (positive vs negative odds ratio 1.63 [95% CI 0.71, 3.75]), mycophenolate use (yes vs no 0.71 [0.33, 1.55]), DLco % predicted (0.95 [0.92, 0.98]), mRSS (1.06 [1.02, 1.10]), time since first non-Raynaud symptom (0.74 [0.58, 0.93]).

Conclusions: SSc-ILD is a heterogeneous disease and prediction of progression may require different approaches in different subgroups.