Abstract
Recent interest in growth differentiation factor 15 (GDF15) has been further heightened by its association with rheumatic diseases. GDF15 has been extensively studied as a potential diagnostic and prognostic biomarker. However, no direct evidence currently links GDF15 to primary Sjögren's syndrome (pSS). In this study, we measured serum GDF15 levels in 151 pSS patients and 61 healthy controls (HCs). GDF15 levels were significantly elevated in pSS patients compared to HCs. Receiver operating characteristic (ROC) curve analysis demonstrated that GDF15 effectively distinguished pSS patients from HCs. Notably, GDF15 levels were higher in pSS patients with extra-glandular manifestations and positively correlated with the number of such manifestations. Specifically, elevated GDF15 levels were observed in patients with pulmonary, hematological, renal, or nervous system involvement. Furthermore, GDF15 levels were significantly increased in patients with active disease compared to those in non-active stage, and ROC analysis confirmed its discriminative capacity for disease activity. Using advanced machine learning techniques, we developed an adaptive boosting model that combined GDF15 with albumin, red blood cell count, and C-reactive protein, which showed reliable predictive value for disease activity. These findings suggest that serum GDF15 may serve as a valuable biomarker for assessing disease severity and guiding treatment in pSS.