This supplementary appendix provides detailed methodological information and statistical considerations for the cluster analysis of Sjögren's disease subgroups. The Ward method, an agglomerative hierarchical clustering technique, was used to group patients based on clinical and biological manifestations. The Euclidean distance metric was employed to assess the proximity between clusters, and the dendrogram was used to visually determine the number of clusters. In the Paris-Saclay cohort, all patients had complete data, while in the ASSESS cohort, missing values were handled using multiple imputations with the 'missMDA' package in R. Descriptive analyses included counts, percentages, medians, and interquartile ranges for categorical and continuous variables, respectively. Statistical tests such as Kruskal-Wallis rank sum test and Pearson's Chi-squared test were used for comparisons. No specific sample size calculation was conducted, but a minimum of 20 to 30 patients per subgroup was anticipated. The validity of statistical tests was assessed using Schoenfeld residual plots for Cox models and Q-Q plots and Mauchly's test for repeated measures ANOVA. Supplementary tables and figures provide additional details on treatment patterns, patient characteristics, and published cluster analyses for primary Sjögren's syndrome.This supplementary appendix provides detailed methodological information and statistical considerations for the cluster analysis of Sjögren's disease subgroups. The Ward method, an agglomerative hierarchical clustering technique, was used to group patients based on clinical and biological manifestations. The Euclidean distance metric was employed to assess the proximity between clusters, and the dendrogram was used to visually determine the number of clusters. In the Paris-Saclay cohort, all patients had complete data, while in the ASSESS cohort, missing values were handled using multiple imputations with the 'missMDA' package in R. Descriptive analyses included counts, percentages, medians, and interquartile ranges for categorical and continuous variables, respectively. Statistical tests such as Kruskal-Wallis rank sum test and Pearson's Chi-squared test were used for comparisons. No specific sample size calculation was conducted, but a minimum of 20 to 30 patients per subgroup was anticipated. The validity of statistical tests was assessed using Schoenfeld residual plots for Cox models and Q-Q plots and Mauchly's test for repeated measures ANOVA. Supplementary tables and figures provide additional details on treatment patterns, patient characteristics, and published cluster analyses for primary Sjögren's syndrome.