BackgroundObservational studies have indicated an association between Hashimoto's thyroiditis (HT) and the risk of thyroid cancer (TC); however, the causality and direction of these effects in genetics remain unclear. Therefore, our study aims to investigate the relationship between genetic susceptibility to HT and thyroid cancer through a bidirectional Mendelian randomization (MR) approach using twin samples. MethodsMR analysis was conducted using genetic instruments associated with HT, selected from a comprehensive genome-wide association meta-analysis involving a total of 756,000 individuals of European and East Asian descent. The dataset for TC comprised 1.486 million individuals, including both European and East Asian populations. Single nucleotide polymorphisms (SNPs) closely linked to HT were derived from genome-wide association studies (GWAS). Two-sample, bi-directional MR analyses were applied to assess the causal association between HT and thyroid cancer, using inverse-variance weighted (IVW) methods, and MR Egger, weighted median, simple mode, and weighted mode. Furthermore, sensitivity analyses were conducted employing the MR-Egger regression model, weighted median method, MR-PRESSO, and leave-one-out technique. ResultsMR analyses revealed no evidence of a causal relationship between HT and TC in either European or East Asian populations (all P > 0.1). However, bidirectional MR analysis demonstrated a causal relationship between TC and HT in the European population (OR=1.0838,95%CI:1.0346, 11354, P=0.000686). ConclusionThis MR analysis indicates a lack of evidence supporting an association between genetically predicted HT and the risk of developing TC. In contrast, there is evidence of a causal relationship between genetically predicted TC and HT.