Association Between Helicobacter Pylori Infection and Graves' Disease: a Meta-Analysis

Background & Aims: Helicobacter pylori (H. pylori) infection is proposed to be related with autoimmune diseases, such as Graves' Disease (GD). This study aimed to assess the association between H. pylori infection and GD. Methods: A systematic literature review was conducted using Pubmed and Cochrane library. The quality of enrolled studies was assessed by the Critical Appraisal Skills Program Oxford. A fixed-effect model approach was used if there was no heterogeneity; otherwise, a random-effect model was used. Heterogeneity was assessed using I2. Publication bias was assessed by funnel plot. All data were analyzed using REVIEW MANAGER 5.3. Results: Six studies from Europe and Asia involving 983 patients were included. Overall H. pylori infection was significantly associated with GD (OR 2.7; 95% CI: 1.47-4.99; p < 0.001). In subgroup analysis of 3 studies using non-serological diagnostic method, the prevalence rate of H. pylori infection was higher in GD group (78.26% VS 42.42%) with significant relationship (OR 4.93; 95% CI: 3.16-7.69; p < 0.00001; I2 = 0%). The CagA antibody prevalence was significantly higher in GD group (46.57% VS 20.29%; OR 4.41; 95% CI: 2.65-7.33; p < 0.00001; I2 = 56%). No publication bias was observed. Conclusion: Our study showed association between H. pylori infection and GD. It might suggest the need of H. pylori examination in GD patients and the impact of H. pylori eradication in the treatment of GD.

Kata kunci: Helicobacter pylori, Graves' disease, meta-analisis references cited by the original published studies and relevant review articles. Article selection and assessment were done by reviewers. We contacted the authors via email to obtain the required information when relevant information was not available in the published article.
Studies were included for analysis based on the following inclusion criteria: (i) observational study having control group; (ii) the outcome was Graves' disease; (iii) Graves' disease was diagnosed by the presence of hyperthyroidism (suppressed TSH, elevated FT3, elevated FT4), diffuse goiter with positive antibody titers (TSI, TPOAb, TgAb), and, in some cases, opthalmopathy; (iv) the exposure was H. pylori infection; (v) the diagnosis of H. pylori infection was based on serological tests (antigenspecific enzyme-linked immunosorbent assay (ELISA) and Western blotting) or non-serological tests (rapid urease test, stool antigen test (SAT), 13 C-urea breath test (UBT)); (vi) studies had extractable data and sufficient information on the association between H. pylori infection and Graves' disease.
We also recorded the CagA serology examination result if it was done in the study. Since all parameter observed were objectively measured and having a written record, neither recall bias nor observer bias were occur in each study. We assessed the quality of each study by using the criteria from the critical appraisal skills program (CASP) Oxford United Kingdom consisting of 11 systematic questions for appraising case control study. The quality levels then were graded as good, fair, and poor. 10 Only studies with good quality were included in our final analysis review.
Discrepancies and disagreements were resolved by consensus.
Study characteristics were taken as follows: first author; year of publication; study design; country; H. pylori test method; subjects characteristics (case subjects, age in mean or median, sex, and matched

INTRODUCTION
Graves' disease, having thyrotoxicosis as clinical hallmark, is characterized by formation of autoantibody thyroid stimulating immunoglobulins (TSI) to the thyroid stimulating hormone receptor (TSH-R). Other antibodies, thyroid peroxidase antibody (TPOAb) and thyroglobulin antibody (TgAb), might also present in Graves' disease. The etiology of Graves' disease is considered as a complex combination of genetic and environment factors. 1,2 The discovery of Helicobacter pylori (H. pylori) by Marshall and Warren in 1982 has contributed significantly in understanding pathogenesis of diseases.
Globally, the prevalence of H. pylori infection is more than 50%. The impact of this microaerophilic, gram negative curved bacillus is not just limited at the gastrointestinal; moreover, it has been proposed to have extra-gastrointestinal manifestations, such as coronary heart disease, diabetes mellitus, autoimmune diseases. [3][4][5][6][7] The presence of seropositivity to the cytotoxin-associated gene A (Cag-A) is commonly used to identify the virulence of H. pylori. 6,7 The interaction of H. pylori as infectious environmental exposure and genetic susceptibility resulting in autoimmune disorder such as Graves' disease has become an appealing issue. 1,3,4,8 The present study was performed to evaluate the association between H. pylori infection and Graves' disease through review of existing studies.

METHOD
We conducted this study according to the metaanalysis PRISMA guideline (see PRISMA checklist). 9 We performed computerized literature search of Cochrane and PubMed database up to July 2015. For the search, we limited to humans and used the following keywords: "Helicobacter pylori" or "H. pylori", and "Graves' disease". We also searched manually related between two groups); CagA test availability; number of subjects with positive H. pylori test in each outcome group; number of subjects with positive CagA test in each outcome group.
We calculated the odds ratios (OR) with 95% confidence interval (CI) for H. pylori positivity. The Mantel-Haenszel method was used to weight the studies included. A fixed-effect model approach was use if there was no heterogeneity; otherwise, a random-effect model was used. Heterogeneity was assessed using I 2 . Negative value of I 2 was put equal to 0. I 2 values ranged from 0% (no observed heterogeneity) to 100%, and interpreted according to Cochrane Consumers and Communication Review Group. 11 For sub-analysis, we calculated the OR of CagA seropositivity. Moreover, to evaluate the possible bias due to difference in diagnostic method, we did sub-analysis of the OR of H. pylori positivity by serological diagnostic method and the OR of H. pylori positivity by non-serological diagnostic method (study using both serological and non-serological diagnostic method will be included here). Publication bias was assessed by funnel plot. All statistical analysis was performed using Review Manager 5.3.

RESULTS
Our literature search identified 16 studies, all published in English. We excluded studies which were not cohort, or case control studies (n = 5), had no available information on the association between H. pylori infection and Graves' disease (n = 2), had no control group (n = 1). After final-text screening, we decided to exclude 1 study due to different baseline profile of H. pylori infection and 1 study due to quality issue. Finally, a total of 6 studies involving 983 patients met our criteria. The flowchart showed the process of studies selection (Figure 1).
The studies were published between 1998 and 2013, and the characteristics of which are summarized in Table. Four studies were performed in Europe, and 2 studies were performed in Asia. 7,8,[12][13][14][15] Four studies had Graves' disease subjects as a subgroup of the case group (subjects with autoimmune thyroid disease) and 2 other studies had Graves' disease subjects as the whole subjects of case group. 7,8,12,15 All studies had greater number of female than male subjects in both Graves' disease and control groups. None of the studies performed endoscopy due to young age subjects and had no alarm symptoms. Three studies used serology method, 2 studies used SAT method, and 1 study used combination of serology and UBT method 12     Prevalance rate of H. pylori positivity showed statistically significant different between Graves' disease and control group in 4 studies. 7,8,13,14 One study showed higher prevalence of H. pylori positivity in Graves' disease group although it was not statistically significant. 12 One study, in contrary, showed higher prevalence of H. pylori positivity in control group but not statistically significant. 15      For the overall 6 studies, no evidence of publication bias was observed in the funnel plot ( Figure 5).
H. pylori cannot differentiate either past or current infection. Moreover, the background of H. pylori prevalence influences the positive predictive value (PPV) of this method, as discussed in the study by Soveid in Iran. [15][16][17][18][19][20][21][22] We found no significant association between H. pylori seropositivity and Graves' disease in this subgroup, as shown by the pooled OR of 3 studies using serology method, and moderate heterogeneity. In contrary, the association was statistically significant in the non-serology method subgroup with 0% of heterogeneity. This finding may underline the importance of ensuring the infection status (current infection) since the antibody itself can persist even after eradication therapy. Therefore, either UBT or SAT method is considered as the preferred non-invasive diagnostic method.
For the subgroup analysis of CagA seropositivity, overall studies showed that infection by the more virulent strain (CagA seropositive) group was significantly correlated with Graves' disease. The possession of CagA plays role in the development of Graves' disease through "antigenic mimicry" and cytokines expression. 17,19 Different diagnostic kit used in each study may contribute to the heterogeneity.
In this study, we had heterogeneous race of subjects, consisting of Caucasian (4 studies), Asian Chinese (1 study), and Asian Iranian (1 study). Genetic factors, which may be clustered more in people depending on their race might play role in the development of Graves' disease. However, there were just 2 studies specifically assessed the genetic factor (HLA alleles), one in Caucasian and one in Asian Chinese. Unfortunately, the result of both studies was not supporting each other. 8,14 Additional studies need to be done to confirm the specific alleles preventing or increasing risk developing of Graves' disease in correlation with H. pylori infection.
To our knowledge, this is the first meta-analysis study to investigate the association between H. pylori positivity and Graves' disease. This association might give an idea of the need H. pylori examination in

DISCUSSION
Based on meta-analysis of 6 studies, there was a significant association between H. pylori positivity and Graves' disease. Graves' disease, a member of autoimmune thyroid disease, arises from the complex interaction of genetic susceptibility and environmental factor. Infectious agent is one of the environmental factors receiving particular attention. Among infectious agents, H. pylori infection has been proposed to trigger the Graves' disease by several mechanisms. Earlier studies showed that the lipopolysaccharide (LPS) of H. pylori has a similar structure with human cell surface glycoconjugates Lewis X, Lewis Y. The "antigenic mimicry" expression of Lewis determinant in LPS is associated with CagA status of the H. pylori strain. [16][17][18] This "antigenic mimicry" might induce autoimmune response which influences the thyroid gland. Interestingly, H. pylori might play role through modulation of dendritic cells, T helper cells, resulting in expression of proinflammatory cytokines such as interferon (IFN)-gamma, interleukin (IL)-10, IL-17. [19][20][21] We performed subgroup analysis by separating the diagnostic methods into serologic and non-serologic. As we know, serological examination of antibody to Graves' disease patients and further investigation whether the treatment of H. pylori will positively influence the treatment of Graves' disease. However, this study has some potential limitations. First, since all studies collected the data of exposure and outcome in the same time, temporal relationship could not be clearly defined. Second, there were just six studies met the criteria and were included the meta-analysis.

CONCLUSION
Our study showed association between H. pylori infection and Graves' disease. We realize that more longitudinal studies with uniform diagnostic method, larger sample sizes, and multiethnic are needed for better understanding of H. pylori and Graves' disease relationship.