The Relationship Between Helicobacter pylori Infection and Open-Angle Glaucoma: A Meta-Analysis

Abstract

Purpose: Previous studies assessed the associations between Helicobacter pylori infection and open-angle glaucoma (OAG) and produced inconsistent results. Therefore, we performed a meta-analysis to evaluate the effects of H. pylori infection on OAG risk.

Methods: Relevant studies that reported the associations between H. pylori infection and OAG were identified through an extensive search of the Exerpta Medica Database (EMBASE), Web of Science, and PubMed databases in English and Chinese National Knowledge Infrastructure, VIP, and Wan Fang databases in Chinese, and by reviewing the reference lists of the key articles. The summary odds ratio (OR) and 95% confidence interval (CI) were calculated using the random-effects model.

Results: This meta-analysis involved 695 glaucoma patients and 1580 control individuals. The overall combined ORs showed a significant correlation between H. pylori infection and OAG (OR = 2.08, 95% CI = 1.42–3.04). The subgroup analysis showed the association of H. pylori infection with primary open-angle glaucoma (POAG; OR = 3.06, 95% CI = 1.27–2.46) and normal tension glaucoma (NTG; OR = 1.77, 95% CI = 1.76–5.34), but not with pseudoexfoliation glaucoma (PXFG; OR = 1.46, 95% CI = 0.40–5.30).

Conclusions: The result of this meta-analysis suggested a statistically significant association between H. pylori infection and OAG. Further analysis showed that this positive relation is observed only in POAG and NTG patients, but not in the PXFG patients.

Open-angle glaucoma (OAG) is a common form of glaucoma that is expected to affect approximately 79.6 million people worldwide in 2020 and is the second-leading cause of blindness, which affects 6.7 million people.¹ It is a chronic neurodegenerative disease characterized by an increased cupping of the optic disk.² The pathogenesis of OAG is not completely understood. However, several factors such as elevated IOP, familial history of glaucoma, advanced age, diabetes mellitus, and myopia have been considered risk factors of OAG.^3–7 

Recently, several studies have provided insights into the role of Helicobacter pylori infection in the pathogenesis of OAG. Kountouras et al.⁸ first reported the relationship between H. pylori infection and OAG, and since then, other studies had focused on this relationship. Helicobacter pylori has been proved to be associated with the genesis of upper gastrointestinal diseases.^9,10 Helicobacter pylori infection also has been considered to be associated with ischemic heart disease, migraine, and cerebrovascular diseases,^11,12 all which are associated with OAG.^13,14 The hypothesis of the correlation between H. pylori infection and OAG originated from the role of H. pylori in inducing a systemic autoimmune response and the release of a large number of inflammatory substances,¹⁵ and in influencing the trabecular meshwork cell apoptotic process.¹⁶ More importantly, H. pylori antibodies are considered to cross-react with ciliary body epithelial antigens.¹⁷ 

Many studies have shown significant interest in investigating the relationship between H. pylori infection and OAG.^18,19 However, the findings to support this relationship are conflicting and inconclusive. For example, some studies reported a statistically significant association between H. pylori infection and OAG.^19–21 Conversely, other studies reported a lack of association.^17,22,23 Considering the inconsistent reports on H. pylori infection as a risk factor for OAG, conducting a meta-analysis to combine the results of all relative studies to assess the correlation of H. pylori infection with OAG would be informative. Thus, we performed this systematic review and meta-analysis to assess the association between H. pylori infection and OAG susceptibility. 

Methods

Literature Search

We performed this systematic review and meta-analysis based on the meta-analysis of Observational Studies in Epidemiology guidelines.²⁴ A systematically electronic literature search was conducted using the databases of PubMed, Exerpta Medica Database (EMBASE), and ISI Web of Science in English and the Chinese National Knowledge Infrastructure (available in the public domain at http://www.cnki.net/), Wan Fang (available in the public domain at http://www.wanfangdata.com.cn), and VIP (available in the public domain at http://www.cqvip.com/) in Chinese (up to March 2015). The following search terms were used: glaucoma or ocular hypertension in combination with HP or Helicobacter pylori or H. pylori. We also examined the references of these studies for potential additional relevant articles. If several articles published by the same population were found, the recent study would be included. If more than one type of OAG was reported in one article, each type of OAG combined with the control group was extracted as one independent dataset, respectively. 

Inclusion and Exclusion Criteria

In the present of meta-analysis, the following inclusion and exclusion criteria were required to be met. The inclusion criteria were designed as case control or cohort, evaluated H. pylori infection and OAG, and odds ratio (OR) and the corresponding 95% confidence interval (CI) were provided or enabled for calculation based on the raw data presented in the article. The exclusion criteria were duplicate data, editorials, letters, review articles, case reports, and article not reporting the outcomes of interest. 

Data Extraction and Quality Assessment

For each study in the analyses, two reviewers (JZ and HL) independently extracted the following information: publication year, first author, country of origin, population ethnicity, mean age of study subjects, type of glaucoma, source of controls, method of diagnosis of H. pylori infection, outcome measured with 95% CIs, and numbers of cases and controls. Any disagreement was resolved by a joint reevaluation of the article by a third reviewer (CD). If data were not available, we would contact the study's authors to request for the missing data. The methodological quality of each included study was assessed independently by two reviewers using the Newcastle–Ottawa Scale (NOS).²⁵ Three major parameters (selection, comparability, and exposure) of the NOS were evaluated. The NOS assigns a maximum of 4, 2, and 3 points for selection, comparability, and exposure/outcome, respectively. Therefore, scores range from 0 to 9 points. Articles with scores higher than 7 were considered of high quality.^26,27 Moreover, quality assessment was independently performed by two reviewers. 

Statistical Analyses

We conducted this meta-analysis using the Stata software package (Version 12.0; Stata Corp., College Station, TX, USA). The strength of the associations between H. pylori infection and OAG was estimated using the OR and 95% CI. The presence of between-study heterogeneity was evaluated using the I² statistic and Cochran's Q test. For the Q statistic, P < 0.05 indicated significant heterogeneity. For I², a value >50% indicated the existence of heterogeneity.²⁸ We chose a random effect model if I² was greater than 50%.²⁹ Subgroup analyses were conducted on the basis of ethnicity (Asian and Caucasian), type of glaucoma (primary open-angle glaucoma [POAG], pseudoexfoliation glaucoma [PXFG], normal tension glaucoma [NTG], ocular hypertension [OHT]), method of diagnosis of H. pylori infection (serum H. pylori IgG antibody test, histological H. pylori test, and ¹³C-urea breath test), and source of control (cataract patients, participants without glaucoma, healthy controls, and anemic control participants). Sensitivity analysis was performed by omitting an individual study each time to assess the reliability of the outcomes of the meta-analysis. Finally, publication bias was statistically assessed by Begg's and Egger's tests, and assessed using Begg's funnel plots.^30,31 Two-sided P value less than 0.05 was considered to be statistically significant in the test results for overall effect. 

Results

Literature Search

Database searches yielded 496 potentially relevant entries. After the removal of 95 duplicate publications, 401 studies remained. Among the 401 articles that qualified for title and abstract review, 386 were excluded because of their apparent irrelevance. In total, 15 articles were assessed further for full-text review. Among these, six were excluded because of the following reasons: one study⁸ was a duplicate report of another study³² on the same population, two articles were reviews,^33,34 and three articles were letters.^35–37 One additional study retrieved from reference also was included in this meta-analysis.³⁸ Thus, 10 articles, including 9 English^{17–23,32,38} and 1 Chinese,³⁹ met the inclusion criteria and were included in this meta-analysis.^{17–23,32,38,39} Of these 10 eligible articles, the study by Galloway et al.¹⁷ involved four types of glaucoma (POAG, NTG, PXFG, and OHT), the study by Kountouras et al.²¹ and Deshpande et al.²³ involved two types (POAG and PXFG), and the study of Kurtz et al.²² involved three types (POAG, NTG, and PXFG). We grouped the different types of glaucoma into different datasets. In another article,¹⁸ the authors included one group of patients and control participants from an outpatient glaucoma clinic, and another group of patients and healthy controls from a primary health care center. We also assumed that the patients from the outpatient glaucoma clinic and those from the primary health care center were the subjects in two datasets. Overall, 18 datasets were selected for our meta-analysis. Our search process is presented in Figure 1. 

Figure 1

 

Flow diagram outlining the selection process for inclusion of studies in the systematic review and meta-analysis.

Characteristics and Quality of the Included Studies

The main characteristics of the included studies are presented in Table 1. The included studies were published between 2002 and 2014, and conducted in Greece, Canada, China, India, Israel, Korea, and Iran. Among the studies, 15 diagnosed the H. pylori infection on the basis of the serum IgG anti–H. pylori antibodies test, two on the basis of the histological H. pylori test, and one on the basis of the ¹³C-urea breath test. Overall, 695 glaucoma patients and 1580 control individuals were included in this meta-analysis. With regard to the quality of the included studies, 90% were of high quality (NOS score >7), with an average NOS score of 7.3, indicating that the methodological quality generally was good. 

Table 1

 

Main Characteristics of Studies Included in This Meta-Analysis

 

Meta-Analysis of H. pylori Infection and OAG

We identified 18 datasets that reported results on H. pylori infection and OAG incidence. The results of the pooled ORs and the heterogeneity test of the meta-analysis are presented in Table 2 and Figure 2. The random effects meta-analysis yielded a pooled OR for H. pylori infection and OAG susceptibility of 2.08 (95% CI, 1.42–3.04). Among the studies, statistically significant heterogeneity was detected (P < 0.001, I² = 63.6%). 

Table 2

 

Meta-Analyses of the Association Between H. pylori Infection and OAG Risk

 

Figure 2

 

Forest plot for the association between H. pylori infection and OAG risk.

Subgroup Analysis

The result of this meta-analysis showed a significant heterogeneity among the studies, and, thus, we conducted a set of subgroup meta-analyses based on ethnicity, type of glaucoma, method of diagnosis of H. pylori infection, and source of control. Table 2 shows the detailed results of the stratified analysis. In the stratified analysis by type of glaucoma, the correlation between H. pylori infection and OAG was significant in POAG patients (OR = 3.06, 95% CI = 1.76–5.34; P < 0.001) and NTG patients (OR = 1.77, 95% CI = 1.27–2.46; P = 0.001), but not in PEXG patients (OR = 1.46, 95% CI = 0.40–5.30; P = 0.562). The subgroup analyses were conducted on the basis of the method of diagnosis of H. pylori infection. The pooled ORs with 95% CIs were 1.73 (95% CI = 1.18–2.54) for studies using the serum H. pylori IgG antibody test, 6.64 (95% CI = 2.96–14.88) for studies using the histological H. pylori test, and 4.49 (95% CI = 1.26–16.01) for the study using the ¹³C-urea breath test. Similarly, stratification by ethnicity and source of control revealed a significant association between H. pylori infection and OAG in all subgroups. However, significant heterogeneity was found among these studies in some subgroups. 

Sensitivity Analysis and Publication Bias

To evaluate the stability of the correlation between H. pylori infection and OAG incidence, we performed a sensitivity analysis that deleted one study at a time and calculated the pooled OR for the remaining studies. The result showed that when we excluded any one study, the estimated pooled OR was similar as before. This finding reflects the high reliability and stability of the results of the present meta-analysis (Table 3). However, the heterogeneity among studies was not significantly reduced by the sensitivity analysis. Publication bias was tested using Begg's funnel plot and Egger's test. The P values for Begg's and Egger's tests were 0.405 and 0.358, respectively, which indicated a low probability of publication bias (Fig. 3). 

Table 3

 

Sensitivity Analysis of the Meta-Analysis Results for H. pylori and the OAG Risk

 

Figure 3

 

Funnel plot for studies of the association of H. pylori infection and OAG risk.

Discussion

In recent decades, H. pylori infection has been considered an important factor involved in OAG risk. However the current literature provides inconsistent conclusions because of their small study sample. Therefore, to systematically elucidate the correlation between H. pylori infection and OAG risk, we summarized all the available relevant evidence. In this meta-analysis, 10 articles with 18 datasets were identified. The pooled results from this meta-analysis supported a significant association between H. pylori infection and OAG incidence. The findings showed that H. pylori infection increased the risk of OAG incidence. To obtain convincing and reliable results, we conducted sensitivity analysis and publication bias analysis. By excluding a single study at a time and rerunning the model to determine the effect on the pooled results, we found that this procedure did not greatly change the pooled result, and it confirmed the stability and reliability of our results. Furthermore, no significant publication bias was observed in the pooled results, thus, indicating the robustness of our study. 

Several possible mechanisms can support the fact that H. pylori infection increases the risk of OAG. First, H. pylori could release various inflammatory factors, such cytokines, C-reactive protein, and nitric oxide.⁴⁰ As glaucoma has been proved to be associated with inflammation,⁴¹ it could cause glaucomatous changes in the globe. Second, H. pylori infection could stimulate the platelet and platelet–leukocyte aggregation, which leads to the decrease in ocular blood flow and results in ocular ischemia.²⁰ Third, studies proved that glaucoma patients have a common genetic factor that makes them more susceptible to H. pylori infection.³⁸ In addition, the toxic materials secreted by H. pylori may influence the glaucoma and cause antibody-induced apoptosis, which is attributed to inflammation in the retrobulbar area.¹⁸ However, the exact mechanism of this correlation remains unclear. Further research is needed to better understand these phenomena. 

In the present meta-analysis, several studies investigated the effect of H. pylori infection on the subtypes of OAG that include POAG, NTG, PXFG, and OHT. The pooled result showed that H. pylori infection increased the risk of POAG and NTG incidence, and that the risk of POAG is greater than that of NTG. Nevertheless, we failed to detect the association between H. pylori infection and PXFG incidence. Pseudoexfoliation glaucoma is a secondary open-angle glaucoma characterized by exfoliation material blocking the trabecular meshwork and resulting in high IOP.⁴² The pathomechanisms of PXFG are different from those of POAG and NTG. Thus, we speculate that PXFG is less likely to be directly related to H. pylori infection. In a separate analysis of Caucasian and Asian ethnicities, the data showed that H. pylori infection increased the OAG risk in both populations, consistent with the overall result of this meta-analysis. Note that the studies included in this meta-analysis used different diagnosis methods of H. pylori infection (i.e., histological H. pylori test, serum H. pylori IgG antibody test, and ¹³C-urea breath test). Histological H. pylori test is the gold standard for diagnosing H. pylori infection.¹⁸ The different diagnosis methods may have different positive rates to test H. pylori, and they might have affected the result of this association. However, none of the studies included in this meta-analysis provided sensitivity/specificity information for the particular method used. Therefore, we cannot comment on their qualitative accuracy in diagnosing H. pylori. The different diagnostic methods used to detect the presence of H. pylori might be a cause of heterogeneity; however, the subgroup analysis showed that the different diagnostic methods did not affect the pooled results. Furthermore, the controls were not uniformly defined in this meta-analysis, and some inevitable selection bias might exist in the results. Thus, further subgroup analysis by source of control (cataract patients, participants without glaucoma, healthy controls, and anemic control participants) was conducted. The pooled result of the different subgroups was consistent with the overall result. 

The present meta-analysis shows a positive association between H. pylori infection and OAG risk, but the significant heterogeneity among the studies limits the ability to draw stable conclusions. To investigate the sources of heterogeneity, we performed several stratified analyses based on ethnicity, type of glaucoma, source of control, and method of diagnosis of H. pylori infection. We found only slight changes, and heterogeneity remained in some subgroups. Next, we conducted a sensitivity analysis, and the results of which showed that no single study was the main source of heterogeneity. The heterogeneity was expected given the between-study variation, such as inconsistent diagnosis methods, different sources of control, and different sample sizes. Thus, the pooled results should be considered with caution. 

There are several advantages of the present meta-analysis. First, to the best of our knowledge, this study is the first analysis to date that explores the association between H. pylori infection and OAG risk. It provides pooled data on a substantial number of cases and controls for the better understanding of the association between H. pylori infection and OAG. Second, the quality of the included studies in this meta-analysis was relatively high. Third, we performed a series of analysis, namely, sensitivity analysis and stratification analysis. The result showed the robustness and the reliability of the pooled result. 

Despite the strengths of this meta-analysis, its limitations should be considered. First, all the studies exhibited great heterogeneity. Significant variability was found in terms of exclusion criteria, inclusion criteria, OAG definitions, type of glaucoma, and method of diagnosis of H. pylori infection. All these factors could be the source of the high heterogeneity among the studies. Despite the use of the random-effect model in summarizing the data to minimize the effects of heterogeneity, we could not account for these differences. Therefore, our findings should be interpreted with caution. Second, different studies had different control participants. Note that in two studies, the control subjects were anemic control participants; selection bias also could have occurred in these studies. However, the stratification analysis suggested that the subgroups of the different sources of control showed similar results. Third, the pooled results were based on unadjusted estimates. They were not adjusted for other known and unknown confounding factors, such as age, sex, family history and IOP, because sufficient information could not be extracted from the original studies. Fourth, visual inspection of the Begg funnel plot identified a little asymmetry. The possible cause of this asymmetry is heterogeneity between studies. Finally, despite the comprehensive search of all relative articles, publication bias still was inevitable because research with null results were not reported, and this bias might have led to the overestimation of the correlation between H. pylori infection and OAG risk. However, the Begg's and Egger's tests did not provide evidence for such bias. 

In conclusion, the present meta-analysis suggests that H. pylori infection has an increased incidence of OAG. However, further analysis shows that this positive relation is limited to POAG and NTG patients, but not in PEXG patients. Despite our rigorous methodology, the limitations of this study should be considered, and the conclusions should be interpreted with caution. In the future, larger and more rigorous studies are required to clarify the association between H. pylori infection and OAG risk. 

Acknowledgments

Supported by a grant from the National Natural Science Foundation of China (No. 81300758). The authors alone are responsible for the content and writing of the paper. 

Disclosure: J. Zeng, None; H. Liu, None; X. Liu, None; C. Ding, None