Skip to main content
NCBI home page
Science Advances logoLink to Science Advances
. 2024 Jul 31;10(31):eadi8646. doi: 10.1126/sciadv.adi8646

Use of bidirectional Mendelian randomization to unveil the association of Helicobacter pylori infection and autoimmune thyroid diseases

Kai Wang 1,†, Qinnan Zhang 1,†, Panpan Zhang 1,†, Qian Yang 1, Fanfan Pan 1, Bingbing Zha 1,*
PMCID: PMC11290481  PMID: 39083605

Abstract

Previous observational studies found associations between Helicobacter pylori infection and autoimmune thyroid diseases (AITDs), but the causal nature of this association is still uncertain. We investigated the causal effect of six crucial antibodies against H. pylori on AITDs using a bidirectional Mendelian randomization (MR). We found that anti–H. pylori outer membrane protein (OMP) significantly increased the risk of hyperthyroidism and Graves’ disease (GD). In addition, our reverse MR analysis indicated that hyperthyroidism could increase the levels of cytotoxin-associated gene A and OMP antibodies. We also observed causal roles of GD on anti–H. pylori OMP. Our analyses indicate the mutual effects of H. pylori infection and AITDs, suggesting the existence of a gut-thyroid axis. These results also provide evidence of the bidirectional causal association between anti–H. pylori OMP with hyperthyroidism and GD, resulting in a vicious circle.

Helicobacter pylori infection and autoimmune thyroid diseases could reinforce each other.

INTRODUCTION

Helicobacter pylori is considered one of the most successful human pathogens. Globally, more than half of the population is colonized with H. pylori, which is identified as the most common bacterial infection (1). The prevalence estimates of H. pylori infection are higher in developing countries with the reported rate up to 80% (2–4). H. pylori infection causes chronic gastritis without clinical symptoms in most cases, but infected individuals could develop peptic ulcer, gastric mucosa-associated lymphoid tissue lymphoma, and gastric cancer (5, 6). Furthermore, H. pylori was designated a class I bacterial carcinogen by the World Health Organization International Agency for Research on Cancer as early as 1994 (2, 7). In addition, there is a large quantity of proteins playing crucial roles in H. pylori infection, particularly cytotoxin-associated gene A (CagA), outer membrane protein (OMP), immunoglobulin G (IgG), UREA, vacuolating cytotoxin (VacA), and catalase.

Autoimmune thyroid diseases (AITDs) are regarded as the most common human autoimmune disorders (8). Characterized by lymphocytic infiltration of the thyroid, AITDs consist mainly of Graves’ disease (GD) and Hashimoto’s thyroiditis (HT), affecting approximately 5% of the population worldwide. The clinical hallmarks of GD and HT are thyrotoxicosis and hypothyroidism, respectively (9). GD is the most common cause of hyperthyroidism, and HT, also known as chronic lymphocytic or autoimmune thyroiditis, is the most common thyroiditis.

It is now generally believed that AITDs have a complex relationship with environmental factors and genetic susceptibility. Of the environmental factors infection, diet, iodine, and smoking probably are the most important ones (10, 11). H. pylori infection could also be implicated in AITDs. It has been found clinically that there are relations between H. pylori and thyroid autoimmunity by some studies. A meta-analysis including 862 individuals suggested a relationship between H. pylori infection and the risk of developing AITDs. In a large-scale study involving a total of 5502 patients, more frequent thyroid peroxidase antibody (TPO-Ab) positivity was confirmed among patients with H. pylori infection than those without. Besides, a previous meta-analysis conducted by Hou et al. (12), contained 1716 AITDs and 1330 controls, indicating a higher incidence of H. pylori infection in AITDs (63.17%) versus controls (45.41%). It also concluded that the eradication of H. pylori infection could reduce serum levels of thyroid autoantibodies. However, it remains unclear whether the observed associations are causal, as observational studies are often prone to selection bias, residual confounding, and reverse causality.

In this study, we aim to analyze causal relationships between H. pylori infection and AITDs-related phenotypes. Mendelian randomization (MR) is a widely used epidemiological method for potential causal inference between exposure and outcome, which uses genetic variants as instruments. Since alleles are randomly assigned at meiosis, genetic instruments are relatively independent of environmental variables and unmodified by disease processes, thereby minimizing potential biases of residual confounding and reverse causation in traditional observational studies (13). Routine MR approaches comprise inverse-variance weighted (IVW), weighted median, MR-Egger, as well as sensitivity analyses, which are applied to keep the results robust. Compared with conventional MR, generalized summary data–based MR (GSMR) has obvious strengths: Accounting for linkage disequilibrium (LD) among genetic variants makes the analysis more efficient. In addition, it adopts the Heterogeneity in Dependent Instruments (HEIDI) test to detect instrumental outliers and eliminate pleiotropic single-nucleotide polymorphisms (SNPs) (14). Given that the previous epidemiological studies did not clearly establish a causal association between H. pylori infection and AITDs, we leveraged genotypic and phenotypic data from large-scale genome-wide association studies (GWASs) to apply both bidirectional two-sample MR and GSMR, disentangling the complicated causal relationship between them.

RESULTS

Study overview

The design of this study consists of two steps. We first aimed to explore whether H. pylori infection (six anti–H. pylori antibodies traits) casually affects AITDs (Table 1), including autoimmune thyroiditis, hyperthyroidism, GD, autoimmune hypothyroidism, and hypothyroidism (step 1). Second, using the summary-level data of FinnGen, we then investigated the reverse relationships, i.e., whether AITDs causally affect H. pylori infection (step 2).

Table 1. Baseline characteristics for participants. Participants from the FinnGen consortium and in Butler-Laporte et al. (30) are included in the study.

Exposure/outcome Data source Phenotype Population Sample size*
Autoimmune thyroid diseases (AITDs) The FinnGen consortium Autoimmune thyroiditis European 244/187,684
Hyperthyroidism European 962/172,976
Graves’ disease European 2311/6,691
Autoimmune hypothyroidism European 22,997/175,475
Hypothyroidism European 26,342/59,827
Antibodies against H. pylori Butler-Laporte et al. (30) CagA European 985
OMP European 2640
IgG European 8735
UREA European 2251
VacA European 1571
Catalase European 1558
Open in a new tab

*Sample size shown as a total number for quantitative traits and cases/controls for binary traits.

The causal effect of H. pylori infection on AITDs risk

Anti–H. pylori CagA antibodies

We first investigated the causal effect of H. pylori infection on AITDs (Table 2, table S1, and Fig. 1). In our main analyses, six antibodies against H. pylori were analyzed in relation to aforementioned five thyroid traits using GSMR. We observed that a one-SD increase in anti–H. pylori CagA caused an increased risk of autoimmune thyroiditis [odds ratio (OR): 1.33, 95% confidence interval (CI): 1.02 to 1.72, P = 0.035].

Table 2. Main MR results when antibodies against H. pylori are considered as exposure.
Exposure Outcome GSMR (main analysis) Inverse-variance weighted Weighted median MR-Egger
OR (95% CI) P* OR (95% CI) P* OR (95% CI) P* OR (95% CI) P*
CagA Autoimmune thyroiditis 1.33 (1.02, 1.72) 0.035 1.32 (0.90, 1.95) 0.156 1.08 (0.73, 1.58) 0.704 1.06 (0.45, 2.50) 0.895
OMP Autoimmune thyroiditis 1.50 (1.02, 2.22) 0.041 1.56 (0.88, 2.76) 0.125 2.00 (1.01, 3.98) 0.048 0.99 (0.14, 7.14) 0.993
Hyperthyroidism 1.60 (1.29, 1.98) 1.3 × 10−5 1.93 (0.95, 3.91) 0.069 1.11 (0.65, 1.90) 0.689 1.43 (0.12, 17.6) 0.788
Graves’ disease 1.70 (1.46, 1.98) 7.2 × 10−12 1.73 (1.07, 2.80) 0.027 1.16 (0.83, 1.62) 0.381 1.62 (0.27, 9.54) 0.614
Open in a new tab

*Raw (unadjusted) P values are given. Results from the main GSMR analysis presented in this table satisfied the criteria: P < 0.05. For GSMR analyses, a Bonferroni-corrected threshold of P = 8.3 × 10−4 was considered significant.

Fig. 1. Forest plot for the GSMR effect of antibodies against H. pylori on AITDs.

Fig. 1.

Open in a new tab

GSMR analyses of H. pylori infection with AITDs. For GSMR analyses, a Bonferroni-corrected threshold of P = 8.3 × 10−4 was considered significant.

Anti–H. pylori OMP antibodies

We then tested whether genetically increased level for anti–H. pylori OMP has causal effect on AITDs susceptibility. Our analyses demonstrated a positive correlation between genetically raised anti–H. pylori OMP levels and autoimmune thyroiditis (OR: 1.50, 95% CI: 1.02 to 2.22, P = 0.041) as well as hyperthyroidism (OR: 1.60, 95% CI: 1.29 to 1.98, P = 1.3 × 10−5). These positive relationships also exist for anti–H. pylori OMP with GD (OR: 1.70, 95% CI: 1.46 to 1.98, P = 7.2 × 10−12).

The reverse MR analyses of AITDs on H. pylori liability

Anti–H. pylori CagA antibodies

To understand the causal relationships among these traits, we implemented MR analyses to test the existence of the reverse or bidirectional causal relationships between AITDs and H. pylori infection (Table 3, table S2, and Fig. 2). We detected a positive association of autoimmune thyroiditis with increased anti–H. pylori CagA levels (OR: 1.05, 95% CI: 1.01 to 1.10, P = 0.016). As for other tested thyroid traits, hyperthyroidism (OR: 1.16, 95% CI: 1.07 to 1.26, P = 2.1 × 10−4) significantly increased the level for anti–H. pylori CagA.

Table 3. Main MR results when AITDs are considered as exposure.
Exposure Outcome GSMR (main analysis) Inverse-variance weighted Weighted median MR-Egger
OR (95% CI) P* OR (95% CI) P* OR (95% CI) P* OR (95% CI) P*
Autoimmune thyroiditis CagA 1.05 (1.01, 1.10) 0.016 1.06 (1.00, 1.13) 0.051 1.02 (0.96, 1.09) 0.492 0.99 (0.89, 1.09) 0.827
IgG 1.01 (1.00, 1.01) 0.018 1.01 (1.00, 1.01) 0.128 1.01 (1.00, 1.02) 0.320 1.00 (0.99, 1.02) 0.512
Hyperthyroidism CagA 1.16 (1.07, 1.26) 2.1 × 10−4 1.19 (1.07, 1.31) 0.001 1.23 (1.08, 1.40) 0.001 1.42 (1.14, 1.78) 0.038
OMP 1.11 (1.06, 1.17) 3.2 × 10−5 1.15 (1.05, 1.27) 0.003 1.13 (1.04, 1.22) 0.003 1.16 (0.90, 1.48) 0.313
UREA 1.07 (1.02, 1.13) 0.006 1.06 (0.99, 1.14) 0.098 1.07 (0.99, 1.16) 0.086 1.21 (1.04, 1.40) 0.072
Graves’ disease CagA 1.10 (1.00, 1.20) 0.049 1.25 (1.06, 1.47) 0.007 1.44 (1.21, 1.71) 3.2 × 10−5 1.45 (0.98, 2.15) 0.121
OMP 1.15 (1.09, 1.22) 1.4 × 10−6 1.19 (1.09, 1.31) 1.4 × 10−4 1.23 (1.11, 1.37) 1.4 × 10−4 1.42 (1.18, 1.71) 0.013
IgG 1.02 (1.00, 1.03) 0.029 1.02 (1.00, 1.04) 0.111 1.02 (0.99, 1.04) 0.134 1.01 (0.97, 1.06) 0.654
UREA 1.10 (1.03, 1.17) 0.003 1.12 (1.03, 1.22) 0.008 1.14 (1.03, 1.27) 0.013 1.22 (1.00, 1.48) 0.108
Autoimmune hypothyroidism OMP 1.09 (1.03, 1.16) 0.003 1.03 (0.96, 1.12) 0.403 1.00 (0.89, 1.13) 0.949 1.03 (0.86, 1.25) 0.723
Hypothyroidism OMP 1.08 (1.00, 1.17) 0.043 1.08 (0.99, 1.18) 0.084 1.04 (0.91, 1.18) 0.571 1.04 (0.82, 1.32) 0.743
Open in a new tab

*Raw (unadjusted) P values are given. Results from the main GSMR analysis presented in this table satisfied the criteria: P < 0.05. For GSMR analyses, a Bonferroni-corrected threshold of P = 8.3 × 10−4 was considered significant.

Fig. 2. Forest plot for the GSMR effect of AITDs on antibodies against H. pylori.

Fig. 2.

Open in a new tab

GSMR analyses of AITDs with H. pylori infection. For GSMR analyses, a Bonferroni-corrected threshold of P = 8.3 × 10−4 was considered significant.

Anti–H. pylori OMP antibodies

We found that genetic predictors of hyperthyroidism exert positive effects on anti–H. pylori OMP levels (OR: 1.11, 95% CI: 1.06 to 1.17, P = 3.2 × 10−5). The GD also causally increased the level for anti–H. pylori OMP (OR: 1.15, 95% CI: 1.09 to 1.22, P = 1.4 × 10−6). Similar relationships were found between genetically raised autoimmune hypothyroidism, hypothyroidism with increased anti–H. pylori OMP (for autoimmune hypothyroidism, OR: 1.09, 95% CI: 1.03 to 1.16, P = 0.003; and for hypothyroidism, OR: 1.08, 95% CI: 1.00 to 1.17, P = 0.043).

Anti–H. pylori IgG antibodies

We observed a causal association of a one-SD increase in autoimmune thyroiditis with an increase in anti–H. pylori IgG (OR: 1.01, 95% CI: 1.00 to 1.01, P = 0.018). Consistent with the above result, there was evidence of positive effects of GD on increased levels for anti–H. pylori IgG (OR: 1.02, 95% CI: 1.00 to 1.03, P = 0.029).

Anti–H. pylori UREA antibodies

Our analysis next demonstrated a causal role of genetically driven hyperthyroidism on raised anti–H. pylori UREA (OR: 1.07, 95% CI: 1.02 to 1.13, P = 0.006). Again, the genetically instrumented GD aggravated the condition of anti–H. pylori UREA (OR: 1.10, 95% CI: 1.03 to 1.17, P = 0.003).

DISCUSSION

The relationship between H. pylori infection and AITDs has been unsettled for years. In bidirectional MR analyses, we found that genetically predicted anti–H. pylori OMP antibodies were causally involved in AITD development, and there were significant causal effects of AITDs on the risk of H. pylori infection, supporting a pathway involving the gut-thyroid axis. In addition, bidirectional effects between anti–H. pylori OMP and AITDs might give rise to a vicious circle.

We demonstrated that genetically raised levels of OMP antibodies to H. pylori are causal risk factors for hyperthyroidism and GD, whereas there was a lack of clinical evidence so far. Composed of 32 members, OMPs of H. pylori play a pivotal role in the attachment and colonization of gastric mucosa. Yamaoka et al. (15) have identified that oipA, hopZ, hopO, and hopP (genes of OMP families) loss of function strongly decreased both H. pylori density and colonization ability in mice. In addition, previous studies reported that H. pylori antigen epitopes were similar to the local amino acid sequence of thyroid endogenous proteins, including the segments of thyrotropin receptor, thyroid autoantigen, and sodium iodide symporter (16–18). These findings support the triggering role of OMPs in AITDs through molecular mimicry, which agrees with our results (18). Besides, the observation that eradication of H. pylori is followed by a gradual decrease in the levels of thyroid auto-antibodies further supports our analyses (19, 20). Moreover, there are some mechanisms suggested that end up activating T helper 1 (TH1) autoreactive lymphocytes, associating H. pylori infection with AITDs. First, dendritic cells present the shared epitopes of H. pylori to naïve T cells, and in the absence of peripheral tolerance, a TH1-driven autoreactive clone is activated. Second, H. pylori stimulates the proliferation of CD4+ T lymphocytes that recognize epitopes of H. pylori structurally similar to H+/K+–adenosine triphosphatase (ATPase) on the gastric parietal cells. In addition, the H+/K+-ATPase is present on the thyroid cells, then TH1 autoreactive lymphocytes activate versus H+/K+-ATPase, which induces apoptosis and synthesis of proinflammatory cytokines such as tumor necrosis factor–α (TNF-α) and interferon-γ (INF-γ) in thyroid gland. Meanwhile, TH1-specific cytokine INF-γ stimulates major histocompatibility complex class II expression of follicular thyroid cells, which then can be allowed to function as antigen-presenting cells and activate TH1 autoreactive lymphocytes (21). All of the above immune cascades further support our findings.

For CagA, it was unexpected to find no significant causal effect of anti–H. pylori CagA on AITDs. None of the GSMR results met the Bonferroni-corrected significance threshold. However, a meta-analysis of seven observational studies indicated that the presence of anti-CagA antibodies in the peripheral blood of H. pylori–infected patients enhances the risk for AITDs (22). CagA is an immunoreactive protein encoded by the H. pylori cagA, which maps to its related pathogenicity island (cagPAI) linked to enhanced pathogenicity of the bacterium. Infection with H. pylori strains having CagA is associated with an increased risk of developing peptic ulcer and gastric cancer (12, 22). Regrettably, as for the relationships among anti-CagA and AITDs, the conclusion has remained controversial, with some research observing positive correlation and others finding null or inverse relationships (23). In addition, no direct evidence of these associations has been found to date, leading to different hypotheses tossed around. It has been demonstrated that OMPs are essential to CagA translocation into gastric cells. What is more is that some researchers contended CagA can come into play only when OMPs adhere to the epithelium (24). Hence, we speculate that CagA plays an indirect role in the complex relation of H. pylori infection and AITDs, causing the negative results, where vitro and vivo experiments are warranted to examine the underlying mechanisms.

In our reverse MR analysis, we observed that genetically driven hyperthyroidism is a causal risk factor for an increase in levels of CagA antibodies to H. pylori. Previous meta-analysis reported that patients with hyperthyroidism had a 1.5-fold risk of developing H. pylori infection than those without (12). In addition, the increased prevalence of H. pylori infection in patients with hyperthyroidism was confirmed by higher anti-CagA levels. Besides, Bassi et al. (25) found that CagA-positive H. pylori infections occurred more frequently in adult patients with GD, which supported our results. We also observed causal roles of hyperthyroidism, GD on the levels of anti–H. pylori OMP. Unfortunately, there is no relevant clinical evidence for OMP so far. Studies have shown that the secretion of cytokines induced by humoral immunity could modify profile of the adhesion molecules expressed on gastric mucosa and may increase H. pylori colonization in patients with GD. Given the pivotal role of OMP in gastric epithelial attachment and colonization, we suppose that OMP plays a key role in H. pylori infection of patients with AITD. While a causal effect between GD and anti-OMP antibodies was observed, there was no significant causal relationship between genetic autoimmune thyroiditis and anti-OMP. In patients with AITDs, the lymphocytes differentiate into TH1/TH2 subtypes. It is known that cellular autoimmunity with the TH1 profile is predominant in HT, whereas humoral autoimmunity with the TH2 profile is prevalent in GD (25, 26). Besides, these different activated profiles in AITDs induce the expression of diverse panels of cytokines, such as interleukin-2 (IL-2), TNF-α, and INF-γ in HT and IL-4, IL-5, IL-6, and IL-10 in GD. All of the above factors may lead to the gastric environment favorable or unfavorable to H. pylori colonization, causing the disparate results (27).

Together, our analyses reveal that H. pylori infection interacts with AITDs and endorse the earlier observational studies, indicating the pathophysiologic interactions between H. pylori infection and thyroid gland. First, as the thyroid gland is, embryogenetically and phylogenetically, derived from the primitive gut, we could consider the thyroid cells as primitive gastroenteric cells. Second, the thyroid and stomach share iodide-concentrating ability. In addition, there have been many morphological and functional similarities such as cell polarity, apical microvilli, secretion of glycoproteins, and peptide hormones. Besides, the two organs are alike in the H+/K+-ATPase cell surface expression (28). Given these resemblances, we combined our analysis results to hypothesize the existence of a gut-thyroid axis. However, the concrete regulation mechanism of H. pylori infection and the thyroid gland or series of AITDs has not yet been comprehensively illustrated. Hence, a thorough study of the mechanism is desperately needed to elucidate the alterations on thyroid gland as well as AITDs. Continued research will pave the way for identifying high-risk patients in early phase, which is vital to achieving more optimized surveillance and earlier diagnosis. Moreover, we will devote ourselves to preventing and treating AITDs or other thyroid dysfunction in patients with H. pylori effectively.

There are several major advantages in our study. First, the inclusion of MR eliminates reverse causation and circumvents residual confounding relatively. Second, to explore the direction of causality between two tightly correlated biological pathways, we conducted a bidirectional MR analysis. Furthermore, the application of GSMR method could mitigate the effects of pleiotropy and confounding. Last, all the GWAS data were derived from participants of European ancestry, indicating that the population stratification is unlikely to affect our results. However, our study has limitations to be noted. First, the samples selected in our study were of European descent, which might limit the generalization of our findings in other ancestry groups. Second, our findings only report the mutual causal relationship between H. pylori infection and AITDs, but the underlying mechanisms warrant further exploration. Third, it is incapable of achieving full non-overlap between exposure and outcome in view of the publicly available GWAS database (29).

In conclusion, our integrative MR study combining bidirectional and GSMR driven methods indicate the mutual effects of H. pylori infection and AITDs, suggesting the existence of a gut-thyroid axis. These results also provide evidence of the bidirectional causal association between anti–H. pylori OMP with hyperthyroidism and GD, resulting in a vicious circle.

MATERIALS AND METHODS

Data sources

The study samples were of European ancestry. We used the summary GWAS data for six antibodies against H. pylori presented in Butler-Laporte et al. (30), comprehensively covering multiple phenotypes of H. pylori infection. With regard to levels of anti–H. pylori CagA, OMP, UREA, VacA and catalase antibodies, summary-level data were derived from GWAS analyses in 985, 2640, 2251, 1571, and 1558 participants separately, available at MRC-IEU UK Biobank (UKB) OpenGWAS under the GWAS ID ebi-a-GCST90006911, ebi-a-GCST90006914, ebi-a-GCST90006915, ebi-a-GCST90006916, and ebi-a-GCST90006912. For anti–H. pylori IgG seropositivity, the GWAS summary statistics is made up of 8735 from European ancestry (ebi-a-GCST90006910).

We extracted instrumental variables (IVs) of AITD traits from FinnGen (https://r5.finngen.fi/). All cases were required to meet International Classification of Diseases (ICD-10). For autoimmune thyroiditis GWAS, they included 244 cases and 187,684 controls. For hyperthyroidism GWAS, the participants were 173,938 Europeans, with 962 cases and 172,976 controls. For autoimmune hypothyroidism, we used the European ancestry GWAS study of 198,472 participants including 22,997 cases. The hypothyroidism GWAS included 86,169 individuals of European ancestry (26,342 cases/59,827 controls). Furthermore, summary data for GD were derived from FinnGen (https://r7.finngen.fi/pheno/E4_GRAVES_STRICT) in up to 2311 cases and 6691 controls of European ancestry.

This MR study was performed using GWAS summary statistics, and ethical approval was obtained by each GWAS. Therein, Neale Laboratory received approval from the Ethics Advisory Committee of the UKB to perform the GWAS. The release of summary statistics pertaining to UKB has been approved by the UKB, and these data are publicly downloadable from the Neale Laboratory website. The FinnGen Biobank GWAS was performed by the FinnGen team and was approved by the FinnGen Steering Committee. The summary statistics are publicly downloadable in the website. All of these data are deidentified, freely downloadable, and can be used without restriction.

MR analysis

Summary data–based MR

We applied three additional MR methods to all relationships between antibodies and diseases from the main analysis-GSMR, including the IVW random effects (IVW-RE), MR-Egger regression, and weighted median approaches. SNPs passing genome-wide significance level (P < 5 × 10−6) were included as IVs for anti–H. pylori antibodies. To study causality of the reversed association, we used the following selection criteria to choose the genetic instruments: a set of SNPs that reached the GWAS significance threshold (P < 5 × 10−8) in association with thyroid disease traits, except for the autoimmune thyroiditis, for which the GWAS significance threshold was set at P < 5 × 10−6 as there were too few SNPs (<2) for the procedure to work properly. Then, we clumped the GWAS significant SNPs from each split’s GWAS with a clumping window of 10,000 kb and an r2 threshold of 0.001. All nonmatching alleles were switched (A1 = A2 and A2 = A1), and the sign of the beta estimates changed. For quality control, we checked for reciprocal strand alleles (e.g., C = G and T = A). We calculated the F statistics of the selected SNPs to detect the strength of the IVs at a threshold of F > 10, which is typically recommended in MR analysis (31–33). Then, we removed SNPs associated with confounders that interfere with the pathway between H. pylori infection and AITDs. We considered three potential confounders, including iodine intake, alcohol intake, and smoking behavior. These three traits have been reported by previous studies to influence AITDs (34). Two of them, alcohol intake and smoking behavior, are considered to be risk factors for H. pylori infection (35, 36). We used the PhenoScanner V2 database (http://phenoscanner.medschl.cam.ac.uk/) to remove SNPs that were significantly (P < 1 × 10−5) associated with confounders in European participants. According to the PhenoScanner, no SNP was associated with iodine intake. We identified one SNP for alcohol intake (rs61938963) as well as two SNPs for smoking behavior (rs597808 and rs7310615). Last, three confounder SNPs were removed (tables S3 and S4). The between-instrument heterogeneity Cochran’s Q-statistic was used to quantify heterogeneity across the instruments. We used the MR-Egger intercept test and MR-PRESSO to evaluate possible bias from horizontal pleiotropy. Last, the leave-one-out analysis was performed to evaluate the robustness of the MR results.

GSMR approach

In our main analyses, we performed GSMR, a flexible approach that performs an MR analysis with multiple near-independent instruments to test for causal association between a risk factor (or phenotype) with a disease using summary-level GWAS data from independent studies (37). In genetic instruments for anti–H. pylori antibodies, we applied the clumping algorithm to select genome-wide significant SNPs for each trait (r2 threshold = 0.05 and P value threshold = 5 × 10−6) using the 1000 Genomes (1000G) phase 3 European samples as the reference for LD estimation. In reverse analysis, for GWAS summary statistics of thyroid traits, we selected independent SNPs setting an LD threshold of r2 < 0.05 with a P value threshold of 5 × 10−8, whereas for GWAS of autoimmune thyroiditis, we used the genome-wide P value threshold of 5 × 10−6 with regard to a low number of SNPs. Then, the HEIDI-outlier approach was applied to remove instruments with strong putative pleiotropic effects, and we specified 0.01 as the P value threshold for the HEIDI-outlier filtering analysis (37).

General criteria for assessing the significance of results

To address the issue of multiple testing, we applied a Bonferroni-corrected significance threshold, which was computed as 8.3 × 10−4 (0.05/60, for the 60 tests). We considered as significant if the directions of the estimates by four methods (GSMR, IVW, MR-Egger, and weighted median) were consistent, GSMR approach passed the significance threshold, and no significant pleiotropy tested by MR-PRESSO global test and modified Q-statistics. MR analyses were performed with “TwoSampleMR” and “gsmr” packages in R version 4.2.0 (http://r-project.org/).

Acknowledgments

We acknowledge all the participants and staff involved in FinnGen and UKB for valuable contribution. We are extremely grateful to all the authors of the Butler-Laporte et al. GWAS for sharing summary statistics, which made analyzes of risk for H. pylori infection in this study possible.

Funding: This work was supported by the Scientific Research Project funded by Shanghai Municipal Science and Technology Commission 22ZR1448700 (B.Z.) and Scientific Research Project funded by Shanghai Fifth People’s Hospital, Fudan University 2022WYZD01 (B.Z.).

Author contributions: B.Z. and K.W. designed the study. K.W., Q.Z., and P.Z. performed the data analysis and generated the figures. K.W. performed the statistical analysis. K.W. and Q.Z. wrote the manuscript. B.Z., K.W., Q.Z., P.Z., Q.Y., and F.P. interpreted the data and contributed to and reviewed the manuscript.

Competing interests: The authors declare that they have no competing interests.

Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials. The GWAS summary data on anti–H. pylori antibodies used for this study can be downloaded from the MRC-IEU UK Biobank OpenGWAS (gwas.mrcieu.ac.uk/datasets/). The data on thyroid diseases are available for bona fide researchers from FinnGen (www.finngen.fi/en/access_results) and can be accessed by filing an application to the FinnGen.

Supplementary Materials

This PDF file includes:

Figs. S1 to S30

Tables S1 and S2

Legends for tables S3 and S4

sciadv.adi8646_sm.pdf (1.4MB, pdf)

Other Supplementary Material for this manuscript includes the following:

Tables S3 and S4

sciadv.adi8646_tables_s3_and_s4.zip (141.4KB, zip)

REFERENCES AND NOTES

  • 1.Hathroubi S., Servetas S. L., Windham I., Merrell D. S., Ottemann K. M., Helicobacter pylori biofilm formation and its potential role in pathogenesis. Microbiol. Mol. Biol. Rev. 82, e00001–e00018 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Aziz F., Khan I., Shukla S., Dey D. K., Yan Q., Chakraborty A., Yoshitomi H., Hwang S. K., Sonwal S., Lee H., Haldorai Y., Xiao J., Huh Y. S., Bajpai V. K., Han Y. K., Partners in crime: The Lewis Y antigen and fucosyltransferase IV in Helicobacter pylori-induced gastric cancer. Pharmacol. Ther. 232, 107994 (2022). [DOI] [PubMed] [Google Scholar]
  • 3.Aziz F., Qiu Y., The role of anti-LeY antibody in the downregulation of MAPKs/COX-2 pathway in gastric cancer. Curr. Drug Targets 15, 469–476 (2014). [DOI] [PubMed] [Google Scholar]
  • 4.Rawla P., Barsouk A., Epidemiology of gastric cancer: Global trends, risk factors and prevention. Prz. Gastroenterol. 14, 26–38 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Salama N. R., Hartung M. L., Müller A., Life in the human stomach: Persistence strategies of the bacterial pathogen Helicobacter pylori. Nat. Rev. Microbiol. 11, 385–399 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Stein M., Ruggiero P., Rappuoli R., Bagnoli F., Helicobacter pylori CagA: From pathogenic mechanisms to its use as an anti-cancer vaccine. Front. Immunol. 4, 328 (2013). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.IARC working group on the evaluation of carcinogenic risks to humans, Schistosomes, Liver Flukes and Helicobacter pylori (International Agency for Research on Cancer, Lyon, 1994). [PMC free article] [PubMed] [Google Scholar]
  • 8.Conigliaro P., D'Antonio A., Pinto S., Chimenti M. S., Triggianese P., Rotondi M., Perricone R., Autoimmune thyroid disorders and rheumatoid arthritis: A bidirectional interplay. Autoimmun. Rev. 19, 102529 (2020). [DOI] [PubMed] [Google Scholar]
  • 9.Antonelli A., Ferrari S. M., Corrado A., Di Domenicantonio A., Fallahi P., Autoimmune thyroid disorders. Autoimmun. Rev. 14, 174–180 (2015). [DOI] [PubMed] [Google Scholar]
  • 10.Tomer Y., Huber A., The etiology of autoimmune thyroid disease: A story of genes and environment. J. Autoimmun. 32, 231–239 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Tomer Y., Mechanisms of autoimmune thyroid diseases: From genetics to epigenetics. Annu. Rev. Pathol. 9, 147–156 (2014). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Hou Y., Sun W., Zhang C., Wang T., Guo X., Wu L., Qin L., Liu T., Meta-analysis of the correlation between Helicobacter pylori infection and autoimmune thyroid diseases. Oncotarget 8, 115691–115700 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Choi K. W., Chen C. Y., Stein M. B., Klimentidis Y. C., Wang M. J., Koenen K. C., Smoller J. W., Major Depressive Disorder Working Group of the Psychiatric Genomics Consortium , Assessment of bidirectional relationships between physical activity and depression among adults: A 2-sample Mendelian randomization study. JAMA Psychiatry 76, 399–408 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Lin B. D., Alkema A., Peters T., Zinkstok J., Libuda L., Hebebrand J., Antel J., Hinney A., Cahn W., Adan R., Luykx J. J., Assessing causal links between metabolic traits, inflammation and schizophrenia: A univariable and multivariable, bidirectional Mendelian-randomization study. Int. J. Epidemiol. 48, 1505–1514 (2019). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Yamaoka Y., Kita M., Kodama T., Imamura S., Ohno T., Sawai N., Ishimaru A., Imanishi J., Graham D. Y., Helicobacter pylori infection in mice: Role of outer membrane proteins in colonization and inflammation. Gastroenterology 123, 1992–2004 (2002). [DOI] [PubMed] [Google Scholar]
  • 16.Wang X. S., Xu X. H., Jiang G., Ling Y. H., Ye T. T., Zhao Y. W., Li K., Lei Y. T., Hu H. Q., Chen M. W., Wang H., Lack of association between Helicobacter pylori infection and the risk of thyroid nodule types: A multicenter case-control study in China. Front. Cell. Infect. Microbiol. 11, 766427 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Benvenga S., Santarpia L., Trimarchi F., Guarneri F., Human thyroid autoantigens and proteins of Yersinia and Borrelia share amino acid sequence homology that includes binding motifs to HLA-DR molecules and T-cell receptor. Thyroid 16, 225–236 (2006). [DOI] [PubMed] [Google Scholar]
  • 18.Benvenga S., Guarneri F., Molecular mimicry and autoimmune thyroid disease. Rev. Endocr. Metab. Disord. 17, 485–498 (2016). [DOI] [PubMed] [Google Scholar]
  • 19.Bertalot G., Montresor G., Tampieri M., Spasiano A., Pedroni M., Milanesi B., Favret M., Manca N., Negrini R., Decrease in thyroid autoantibodies after eradication of Helicobacter pylori infection. Clin. Endocrinol. 61, 650–652 (2004). [DOI] [PubMed] [Google Scholar]
  • 20.Papamichael K. X., Papaioannou G., Karga H., Roussos A., Mantzaris G. J., Helicobacter pylori infection and endocrine disorders: Is there a link? World J. Gastroenterol. 15, 2701–2707 (2009). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Cellini M., Santaguida M. G., Virili C., Capriello S., Brusca N., Gargano L., Centanni M., Hashimoto's thyroiditis and autoimmune gastritis. Front. Endocrinol. 8, 92 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Shi W. J., Liu W., Zhou X. Y., Ye F., Zhang G. X., Associations of Helicobacter pylori infection and cytotoxin-associated gene A status with autoimmune thyroid diseases: A meta-analysis. Thyroid 23, 1294–1300 (2013). [DOI] [PubMed] [Google Scholar]
  • 23.Wang Y., Zhu S., Xu Y., Wang X., Zhu Y., Interaction between gene A-positive Helicobacter pylori and human leukocyte antigen II alleles increase the risk of Graves disease in Chinese Han population: An association study. Gene 531, 84–89 (2013). [DOI] [PubMed] [Google Scholar]
  • 24.Matsuo Y., Kido Y., Yamaoka Y., Helicobacter pylori outer membrane protein-related pathogenesis. Toxins 9, 101 (2017). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Bassi V., Marino G., Iengo A., Fattoruso O., Santinelli C., Autoimmune thyroid diseases and Helicobacter pylori: The correlation is present only in Graves's disease. World J. Gastroenterol. 18, 1093–1097 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Roura-Mir C., Catálfamo M., Sospedra M., Alcalde L., Pujol-Borrell R., Jaraquemada D., Single-cell analysis of intrathyroidal lymphocytes shows differential cytokine expression in Hashimoto's and Graves' disease. Eur. J. Immunol. 27, 3290–3302 (1997). [DOI] [PubMed] [Google Scholar]
  • 27.Höcker M., Hohenberger P., Helicobacter pylori virulence factors—One part of a big picture. Lancet 362, 1231–1233 (2003). [DOI] [PubMed] [Google Scholar]
  • 28.Venturi S., Venturi M., Iodide, thyroid and stomach carcinogenesis: Evolutionary story of a primitive antioxidant? Eur. J. Endocrinol. 140, 371–372 (1999). [DOI] [PubMed] [Google Scholar]
  • 29.Chen X., Kong J., Pan J., Huang K., Zhou W., Diao X., Cai J., Zheng J., Yang X., Xie W., Yu H., Li J., Pei L., Dong W., Qin H., Huang J., Lin T., Kidney damage causally affects the brain cortical structure: A Mendelian randomization study. EBioMedicine 72, 103592 (2021). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Butler-Laporte G., Kreuzer D., Nakanishi T., Harroud A., Forgetta V., Richards J. B., Genetic determinants of antibody-mediated immune responses to infectious diseases agents: A genome-wide and HLA association study. Open Forum Infect. Dis. 7, ofaa450 (2020). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Chen L., Peters J. E., Prins B., Persyn E., Traylor M., Surendran P., Karthikeyan S., Yonova-Doing E., Di Angelantonio E., Roberts D. J., Watkins N. A., Ouwehand W. H., Danesh J., Lewis C. M., Bronson P. G., Markus H. S., Burgess S., Butterworth A. S., Howson J. M. M., Systematic Mendelian randomization using the human plasma proteome to discover potential therapeutic targets for stroke. Nat. Commun. 13, 6143 (2022). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Pierce B. L., Ahsan H., Vanderweele T. J., Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int. J. Epidemiol. 40, 740–752 (2011). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Palmer T. M., Lawlor D. A., Harbord R. M., Sheehan N. A., Tobias J. H., Timpson N. J., Smith G. D., Sterne J. A., Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat. Methods Med. Res. 21, 223–242 (2012). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Brent G. A., Environmental exposures and autoimmune thyroid disease. Thyroid 20, 755–761 (2010). [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.Leja M., Grinberga-Derica I., Bilgilier C., Steininger C., Review: Epidemiology of Helicobacter pylori infection. Helicobacter 24, e12635 (2019). [DOI] [PubMed] [Google Scholar]
  • 36.Mentis A., Lehours P., Mégraud F., Epidemiology and diagnosis of Helicobacter pylori infection. Helicobacter 20, 1–7 (2015). [DOI] [PubMed] [Google Scholar]
  • 37.Zhu Z., Zheng Z., Zhang F., Wu Y., Trzaskowski M., Maier R., Robinson M. R., McGrath J. J., Visscher P. M., Wray N. R., Yang J., Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat. Commun. 9, 224 (2018). [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figs. S1 to S30

Tables S1 and S2

Legends for tables S3 and S4

sciadv.adi8646_sm.pdf (1.4MB, pdf)

Tables S3 and S4

sciadv.adi8646_tables_s3_and_s4.zip (141.4KB, zip)

Articles from Science Advances are provided here courtesy of American Association for the Advancement of Science

RESOURCES