Genetic liability to systemic lupus erythematosus and autoimmune thyroid diseases: evidence from a bidirectional...

Previous observational studies imply a potential link between systemic lupus erythematosus (SLE) and autoimmune thyroid disease (AITD), although they are susceptible to confounding factors. This study aims to assess the genetic evidence for a potential causal association between SLE and AITD using a bidirectional and multivariable two-sample Mendelian randomization (MR) analysis.

We retrieved datasets for systemic lupus erythematosus (SLE), autoimmune thyroiditis (AT), autoimmune hyperthyroidism, and autoimmune hypothyroidism from European population-based databases, and data for Graves' disease (GD) from Biobank Japan (BBJ). Instrumental variables (IVs) related to exposure were selected for the Mendelian randomization (MR) analysis. Sensitivity tests, including horizontal pleiotropy and heterogeneity assessments, were conducted to ensure result reliability, and the analysis outcomes were visualized. Bonferroni correction was applied to the 8 primary exposure-outcome pairs (corrected significance threshold p < 0.00625). The use of a Japanese GWAS dataset for GD alongside predominantly European datasets for other phenotypes represents a methodological limitation due to potential cross-ancestry differences in linkage disequilibrium structure and allele frequencies.

Genetic variants linked to SLE, AT, GD, autoimmune hyperthyroidism, and autoimmune hypothyroidism were utilized as IVs. MR analysis was performed using data from 5201 SLE cases and 9066 controls; 244 AT cases and 187,684 controls; 2176 GD cases and 210,277 controls; 962 autoimmune hyperthyroidism cases and 172,976 controls; 22,997 autoimmune hypothyroidism cases and 175,475 controls. In the primary IVW analysis, genetic liability to SLE was associated with increased risks of AT (OR = 1.136, 95% CI 1.006-1.284, p = 0.040) and autoimmune hypothyroidism (OR = 1.036, 95% CI 1.004-1.070, p = 0.025). Genetically predicted autoimmune hyperthyroidism is associated with an increased risk of SLE in both univariable (OR = 1.905, 95% CI 1.195-3.037, p = 0.007) and multivariable (OR = 1.381, 95% CI 1.042-1.830, p = 0.024) MR analyses. However, these associations did not remain statistically significant after Bonferroni correction for 8 tested exposure-outcome pairs (corrected significance threshold: p < 0.00625). Sensitivity analysis showed limited evidence of horizontal pleiotropy, though significant heterogeneity was detected for several exposure-outcome pairs (see Fig. 5).

Mendelian randomization analysis suggests that genetic liability to SLE is associated with a potential increased risk of AT and autoimmune hypothyroidism, and genetic liability to autoimmune hyperthyroidism is associated with a possible increased risk of SLE in European populations. These findings are based on genetic instruments and should be interpreted as genetic evidence consistent with potential causal effects rather than definitive proof of causality. The observed associations have relatively modest effect sizes and reflect lifelong genetic liability; they should not be equated with the clinical epidemiological risk among patients with established disease. Their clinical relevance requires further validation in large-scale prospective clinical studies. The mixed ancestry of GWAS datasets and the presence of significant heterogeneity are important limitations. Key Points • This study systematically evaluated the bidirectional genetic associations between SLE and distinct AITD subtypes via a comprehensive bidirectional and multivariable two-sample MR design. • Genetic liability to SLE was genetically associated with an increased risk of AT and autoimmune hypothyroidism in European populations. • Autoimmune hyperthyroidism was identified as a potential genetic risk factor for SLE, validated by both univariable and multivariable MR analyses. • The findings provide genetic evidence for understanding the shared genetic architecture and potential causal associations between SLE and AITD subtypes. However, given the modest effect sizes and the distinction between genetic liability and clinical disease, they do not currently support routine targeted thyroid comorbidity screening in all SLE patients without further clinical validation.