Causal inference between serum bilirubin levels and juvenile idiopathic arthritis-associated uveitis: A bidirectional...

Several observational studies have suggested an association between low serum bilirubin levels and Behçet's disease uveitis. However, the causal inference between bilirubin level and juvenile idiopathic arthritis-associated uveitis (JIAU) remains ambiguous. We investigated the potential causal relationship between serum bilirubin levels and JIAU using a bidirectional two-sample Mendelian randomization (MR) framework.

We systemically integrated summary-level data from published large-scale genome-wide association studies on bilirubin level and JIAU in a Caucasian British population. To determine the causal effect of bilirubin level on JIAU, we constructed strong instrumental variables using 47 and 80 single-nucleotide polymorphisms (SNPs) specific to direct bilirubin and total bilirubin levels, respectively. For reverse causal inference, seven SNPs associated with JIAU were included in our study. Multiple complementary methods were further performed to evaluate the robustness of MR estimates.

The inverse-variance weighted estimate did not show any significant causal associations of genetically predicted serum direct or total bilirubin level with the risk of JIAU (odds ratio [OR]: 1.010, 95% confidence interval [CI]: 0.750-1.359, p = 0.947; OR: 0.867, 95% CI: 0.688-1.093; p = 0.227, respectively). MR-Egger and weighted median methods also obtained similar associations. Additionally, the results of reverse MR analyses using JIAU as exposure showed no associations of genetically predicted risk of JIAU with serum bilirubin levels (p > 0.05). In sensitivity analysis, the causal estimate between serum bilirubin levels and JIAU did not differ when SNPs associated with possible confounders were omitted.

Genetic evidence from our bidirectional analysis did not support a causal association between serum bilirubin levels and JIAU risk in the Caucasian British population. Future large-scale studies should be conducted to validate these findings and explore any causal effects on the disease process.