Phenotypic clusters of long-term damage in Sjögren's disease: a 15-year multicentre study

Objective: To identify phenotypic damage pattern at 15 years and predictors of long-term damage at diagnosis in a multicentre Sj & ouml;gren's disease (SjD) cohort. Methods: A SjD cohort with at least 15 years of regular follow-up was retrospectively included. Clinical and serological parameters, disease activity, comorbidities, disease damage and treatments were recorded every 5 years and changes compared. Hierarchical clustering analysis was performed building the model based on SjD disease damage index (SSDDI) and its domains at 15 years. Disease characteristics at diagnosis (baseline) were compared among clusters. Results: A total of 102 SjD patients were included. Damage significantly increased from 5 to 10 and 15 years (P < 0.001). European League against Rheumatism SS disease activity index (ESSDAI) score significantly reduced from baseline to 5, 10 and 15 years (P < 0.001). ESSDAI at baseline significantly predicted damage at 15 years (beta = 0.218, P = 0.001, CI: 0.103-0.336), independently from age at diagnosis and comorbidities. Among ESSDAI domains, lymphadenopathy independently predicted damage at 15 years (beta = 2.004, P = 0.011, CI: 0.565-3.411). The hierarchical clustering model identified five clusters with distinct damage patterns. Conclusions: In a multicentre Italian SjD cohort, the largest damage burden develops by year 5 and progressively builds up over at least 15 years with different progression pattern among damage domains. Disease activity at diagnosis and lymphadenopathy significantly predict damage. Five clusters with distinct damage patterns were identified, suggesting a tailored approach to the disease since diagnosis to reduce long-term damage accrual.