Predictors of systemic sclerosis patients with interstitial lung disease progression: A retrospective study

Abstract Background Lung involvement is one of the most common and serious internal organ damages in systemic sclerosis (SSc). The aim of this study was to establish a prediction model to assess the risk factors of progression in SSc patients with interstitial lung disease (ILD), thereby assisting clinicians in the early identification of high‐risk patients. Methods Eligible patients were registered in our database of adult patients hospitalized in the Rheumatology Department of the Nanjing Drum Tower Hospital from January 1, 2013 to January 1, 2022, and the overall cohort was divided into the derivation cohort and the internal validation cohort. ILD progression was defined as deteriorating symptoms and the extent of lung fibrosis on follow‐up high‐resolution computed tomography. A prediction nomogram for ILD progression was established in the derivation cohort by least absolute shrinkage and selection operator and multivariable logistic regression and validated in the internal validation cohort. The total points of the nomogram were calculated to delineate high‐risk and low‐risk groups. The Kaplan‐Meier curve was used to determine the association between SSc patients with or without ILD progression, as well as between high‐risk and low‐risk groups. Results We included 273 patients, 227 (83.2%) were females and 46 (16.8%) were males, the mean age was (45.1 ± 15.5) years, the disease duration was 3 (25th–75th percentiles: 1–8) years, 126 of whom developed ILD progression during follow‐up (1–5 years). The final SSc‐ILD progression prediction model was composed of six significant risk factors: male, type 2 diabetes mellitus, pericardial effusion, anti‐Ro52, hypoproteinemia, and elevated C‐reactive protein. An SSc‐ILD progression probability‐predictive nomogram was established. In two cohorts, the AUCs were 0.832 (95% confidence interval [CI]: 0.774–0.890) and 0.818 (95% CI: 0.722–0.914), respectively. The calibration curve, decision curve analysis, and clinical impact curve all showed good performance of the model. The Kaplan–Meier curve suggested that the occurrence of ILD progression was associated with poor prognosis, and the curve of high‐risk and low‐risk groups defined by our model demonstrated excellent prognostic utility. Conclusion This new model demonstrated good capability for assessing the risk in SSc‐ILD progression, which may be helpful for early identification and intervention.