Familial and genetic overlap between Sjögren’s disease and other autoimmune diseases

Purpose To evaluate familial clustering of Sjögren’s disease (SjD) with other autoimmune diseases and to characterize shared genetic architecture using an integrative genomic approach. Methods Meta-analysis identified studies assessing autoimmune disease incidence in probands and first-degree relatives (FDRs), and pooled relative risks (RRs) were calculated. Publicly available genome-wide association study (GWAS) summary statistics were analyzed within a hierarchical genetic architecture framework integrating genome-wide polygenic correlation (LDSC), locus-level overlap (Jaccard index), union-based susceptibility mapping, and SNP-level pleiotropy detection. Results Eighteen studies evaluated familial aggregation of SjD and other autoimmune diseases, of which nine were included in the pooled analysis. The RR of SjD was 10.54 when both proband and FDR were affected. Among discordant autoimmune probands, systemic lupus erythematosus (SLE) showed the highest RR for SjD (4.49), followed by systemic sclerosis (2.65). Genome-wide analyses demonstrated substantial polygenic sharing between SjD and systemic autoimmune diseases, positioning SjD as a bridging disorder. However, locus-level overlap was largely driven by the HLA region; after HLA exclusion, shared loci markedly decreased and were restricted to a limited number of immune regulatory hubs. SNP-level pleiotropy analyses similarly indicated predominantly HLA-dependent sharing with fewer non-HLA signals. Conclusion SjD shows strong familial aggregation and shared genetic susceptibility with multiple autoimmune diseases. These findings support a hierarchical model in which broad HLA-driven polygenic sharing coexists with selective non-HLA convergence. Strengths and limitations of this study This study integrates systematic familial meta-analysis with GWAS data to characterize shared autoimmune genetic architecture. However, inference regarding shared causal variants remains limited by reliance on summary statistics and locus-based resolution.