Causal role of uric acid in oral lichen planus independent of key metabolic factors: multivariable Mendelian randomization with multi-omics analyses

Abstract Background Emerging evidence links oral lichen planus (OLP) with hyperuricemia. However, whether this reflects a genuine causal relationship or is confounded by metabolic conditions remains unclear, and the underlying biological mechanisms are incompletely understood. Methods Two-sample Mendelian randomization (MR) with inverse-variance weighted (IVW) estimation and multivariable MR (MVMR) were used in genome-wide association study (GWAS) summary data to assess causal effects of UA and five key metabolic traits on OLP. All genome-wide association study summary statistics were derived from European-ancestry populations. MR pleiotropy residual sum and outlier (MR-PRESSO) and Causal Analysis Using Summary Effect estimates (CAUSE) controlled for pleiotropy. Linkage disequilibrium score regression (LDSC) evaluated genetic correlations. Pleiotropic analysis under composite null hypothesis (PLACO) and cross-phenotype association (CPASSOC) with Bayesian colocalization identified shared loci, further evaluated by summary-data-based MR (SMR) and joint-tissue transcriptome-wide association MR (MR-JTI). Results Genetically predicted higher UA levels were associated with increased OLP risk in the primary MR analysis (OR IVW-RE = 1.64 [95% CI: 1.10–2.45] per 1-SD inverse-rank normal transformed UA, P = 0.016); the five secondary metabolic phenotypes showed no significant associations. Multivariable MR confirmed that the UA–OLP association was independent of the metabolic traits included in the model. Nominally significant genetic correlations were observed for OLP with UA (r g = 0.066 [95% CI: 0.009, 0.123], P = 0.021) and high-density lipoprotein (HDL) (r g = − 0.096 [95% CI: − 0.180, − 0.012], P = 0.025). Bayesian colocalization identified FGF5 and BIRC6 as shared loci; SMR and MR-JTI provided supportive evidence for FGF5 and BIRC6 as candidate genes. Conclusion The findings provide genetic evidence consistent with a causal effect of elevated UA on OLP susceptibility, independent of the metabolic traits included in the multivariable MR model and under the assumptions of Mendelian randomization. Colocalization, SMR and MR-JTI further implicate FGF5 and BIRC6 as shared-locus candidate genes linking UA and OLP; these signals require experimental validation.