**Background:** Coeliac disease (CeD) is an immune-mediated enteropathy affecting 0.5–1% of Western populations. Observational studies have linked CeD to numerous comorbidities, including autoimmune diseases, liver disorders, certain cancers, and malnutrition, but whether these associations are causal remains unclear due to potential confounding and reverse causation. This study aimed to use Mendelian randomization (MR) to assess the causal relationships between genetic liability to CeD and a broad range of clinical outcomes.
**Methods:** The authors selected 58 independent single nucleotide polymorphisms (SNPs) associated with CeD at genome-wide significance (P < 5×10⁻⁸) from a meta-analysis of 12,041 CeD cases and 12,228 controls, excluding SNPs with linkage disequilibrium (R² > 0.001). All F-statistics exceeded 10 (mean = 233.3), indicating strong instruments. In the UK Biobank (N = 385,917 unrelated White British individuals), a weighted polygenic risk score (PRS) was constructed and tested for associations with 1060 clinical outcomes (defined by PheCODE schema, excluding those with <120 cases) using logistic regression adjusted for age, sex, assessment centre, and 10 principal components. False discovery rate (FDR) correction was applied. Replication was performed in the FinnGen study (N = 377,277) using two-sample MR with inverse variance weighted method, plus sensitivity analyses (weighted median, MR-Egger, MR-PRESSO). A secondary analysis removed 21 SNPs in the extended MHC region (chromosome 6, positions 25,726,291–33,368,421) to assess the role of HLA genes. Multivariable MR adjusted for genetically predicted abundance of genus Bifidobacterium was also conducted.
**Key Results:** In the UK Biobank, genetic liability to CeD was associated with 68 clinical outcomes after FDR correction. The strongest association was with CeD itself (OR 1.62, 95% CI 1.60–1.65, P < 9.67×10⁻⁸⁵), validating the instrument. Significant positive associations included: type 1 diabetes (OR 1.09, 95% CI 1.07–1.10, P = 5.45×10⁻²⁷), Graves' disease (OR 1.06, 95% CI 1.04–1.08, P = 2.46×10⁻¹¹), Sjögren syndrome (OR 1.10, 95% CI 1.07–1.13, P = 7.45×10⁻¹⁰), chronic hepatitis (OR 1.13, 95% CI 1.08–1.18, P = 2.40×10⁻⁸), systemic lupus erythematosus (OR 1.10, 95% CI 1.06–1.14, P = 8.22×10⁻⁷), sarcoidosis (OR 1.08, 95% CI 1.04–1.11, P = 1.22×10⁻⁶), cutaneous lupus (OR 1.12, 95% CI 1.05–1.18, P = 1.41×10⁻⁴), non-Hodgkin's lymphoma (OR 1.03, 95% CI 1.01–1.05, P = 2.89×10⁻⁴), osteoporosis (OR 1.01, 95% CI 1.00–1.02, P = 8.58×10⁻⁴), iron deficiency anaemia (OR 1.01, 95% CI 1.01–1.02, P = 4.92×10⁻⁴), and vitamin B-complex deficiencies (OR 1.03, 95% CI 1.01–1.04, P = 0.001). Inverse associations were observed for prostate cancer (OR 0.98, 95% CI 0.98–0.99, P = 4.53×10⁻⁹), prostatic hyperplasia (OR 0.98, 95% CI 0.97–0.98, P = 1.35×10⁻⁵), and other prostate disorders (OR 0.96, 95% CI 0.94–0.98, P = 3.31×10⁻⁴). Of the 57 outcomes available in FinnGen, 38 were replicated at P < 0.05, including all major autoimmune diseases, non-Hodgkin's lymphoma, osteoporosis, cataract, malnutrition, and prostate diseases. In the secondary analysis excluding MHC SNPs, associations for type 1 diabetes (OR 1.01, 95% CI 0.80–1.28) and non-Hodgkin's lymphoma attenuated, while other associations (e.g., Graves' disease OR 1.43, 95% CI 0.88–2.31) increased in magnitude but became non-significant due to wider confidence intervals. Sex-stratified and age-stratified analyses showed consistent results. Sensitivity analyses (MR-Egger, MR-PRESSO) did not reveal significant directional pleiotropy for most outcomes.
**Clinical Implications:** This study provides genetic evidence supporting causal relationships between CeD and multiple comorbidities, reinforcing current guidelines for screening CeD in patients with type 1 diabetes, autoimmune thyroid disease, iron deficiency anaemia, and osteoporosis. The attenuation of associations for type 1 diabetes and non-Hodgkin's lymphoma after removing MHC SNPs suggests shared genetic aetiology, but does not diminish the clinical need for comorbidity monitoring. The inverse association with prostate diseases is novel and warrants further investigation. These findings highlight the importance of multidisciplinary care for CeD patients, including monitoring for autoimmune conditions, haematological malignancies, and bone health. Limitations include the use of inpatient records which may miss milder cases, potential horizontal pleiotropy, and reduced power in the non-MHC analysis. Overall, this MR-PheWAS underscores the necessity of comorbidity surveillance in the CeD population.