Background

Helicobacter pylori (H. pylori), formerly known as Campylobacter pyloridis, is a gram-negative, microaerophilic, helical, and motile (lophotrichous flagella) bacterium that colonizes the gastric sub-mucosa of more than 50% of the world’s population [1]. Infection with this pathogen generally occurs in childhood and may continue asymptomatic for life [2]. However in 15–20% of infected people, the primary infection progresses to worse conditions such as peptic ulcer disease (PUD), duodenal ulcer (DU), gastric ulcer (GU), gastric adenocarcinoma, and mucosa-associated lymphoid tissue (MALT) lymphoma; PUD rarely occurs in children [3,4,5]. According to the literature, interactions between bacteria, the host genome, and environmental conditions play a decisive role in the development of primary infection to severe clinical outcomes [6]. Although the main role of some virulence genes such as vacA and cagA is well known, the effect of other virulence factors on bacterial pathogenesis is unclear and needs further study [7,8,9,10]. The H. pylori genome encodes about 1,100 genes, of which 500–600 are strain-specific genes and in turn contribute in various clinical outcomes [11, 12]. Outer membrane proteins (OMPs) are among the most divergent proteins in this bacterium, encoded by 4–5% of the bacterial genome [13]. The hom genes are known as a small paralogous family of adhesion proteins and are distinguished from other OMPs by the signal sequence and hydrophobic motif located in the C-terminal domain [14]. The hom family consists of four classes homA, homB, homC, and homD, so that homA and homB are encoded by one locus, while each of homC and homD is encoded by a distinct locus [14, 15]. Despite 90% similarity between homA and homB, studies show that the distribution of each is specific in each geographic area, so that homA is a diagnostic marker for East Asian strains, and homB has global distribution [15, 16]. Recently, the relationship between homB and severe clinical outcomes has attracted much attention; the product of homB, HomB is a virulence factor that contributes to several bacterial activities such as biofilm formation, antibiotic resistance, delivery of CagA from type 4 secretion system (T4SS), induction of IL-8 production, gastritis, corpus atrophy, and persistent colonization [17,18,19,20]. Oleastro et al. first showed a significant association between the homb-positive strains and the progression of infection to PUD in Portuguese children [21]. In another study, they found that the ability of homB knockout mutant strains to bind to gastric epithelium was significantly reduced compared to homB-positive strains [22]. Jung et al. found a positive correlation between the simultaneous presence of cagA and homB genes in East Asian strains; they showed that having two copies of homB gene could increase the risk of PUD [23]. In contrast, the homA gene is correlated with non-ulcer dyspepsia (NUD), in other words, it appears that there is no significant relationship between homA and gastritis and corpus atrophy [16, 19]. In the present meta-analysis, we evaluated the association between homB-positive strains of H. pylori and several clinical outcomes.

Methods

Literature search strategy

At the first, a systematic search was performed using global databases such as Scopus, Web of Science, and PubMed to collect all the studies relevant to our purpose. All selective studies were related to the association between homB and clinical outcomes such as PUD, duodenal ulcer, gastric ulcer, and gastric cancer (GC). In this study, articles published up to December 2020 were retrieved separately by two authors (MK1 and MK2). Search terms were selected based on MeSH thesaurus including “Helicobacter pylori”, “H. pylori”, and “homB”; articles were search regardless of publication date and language.

Inclusion criteria

In this step, we used the full-text of all published case–control studies examining the relationship between the homB and severe clinical outcomes (patients with gastritis and NUD were considered as control group, while patients with PUD, GU, DU, and GC were assumed as case group). The inclusion criteria in adult populations included; (1) published researches on the presence of homB confirmed by polymerase chain reaction (PCR) or immunoblotting, (2) exploratory studies on the association between H. pylori homB gene and severe clinical outcomes, (3) articles published in English, and (4) articles about human adults. All relevant documents were independently reviewed by two authors (MK1 and MK2).

Exclusion criteria

Exclusion criteria included; (1) duplicate articles, (2) studies without raw data, (3) studies with control groups only (gastritis and NUD), (4) articles with repetitive samples and results, (5) conference abstracts, review articles, and case series, (6) in vitro or animal experiments, (7) articles with insufficient data for calculating OR with 95% confidence intervals (CIs), (8) studies on non-homB gene, and 9) (studies conducted in child populations.

Quality assessment and data extraction

Using the Newcastle–Ottawa Scale (NOS), 12 eligible studies were selected [21,22,23,24,25,26,27,28,29,30,31,32]. Required information such as first author, year of publication, country, population sample size, number of H.pylori strains, Frequency of homB in strains creating different clinical futures, diagnostic methods and reference number are listed in Table 1.

Table 1 Characteristics of included studies
Full size table

Statistical analysis

All statistical analysis were performed using Comprehensive Meta-Analysis (CMA) software version 2.2 (Biostat, Englewood, NJ, USA). The relationship between homB and clinical outcomes was estimated according to the summary odds ratio with 95% CIs. Heterogeneity between studies was assessed through parameters such as I2 index and Cochrane Q test, so that in cases of high heterogeneity (I > 25% and Cochrane Q test p value > 0.05) and non-significant heterogeneity, we used from random-effects model, and fixed-effects model, respectively. Finally, the publication bias of selected studies was measured using funnel plot asymmetry, Egger’s p value, and Begg’s p value test [33].

Results

Characteristics of selected studies

Following the initial systematic search, 138 related articles were collected and 126 articles were deleted according to inclusion criteria. The details of comprehensive search processing and study selection are summarized in Fig. 1. In the screening phase, we removed the irrelevant articles such as articles with unclear results and articles that did not meet our criteria. In total, out of 12 studies that met our inclusion criteria, 2930 patients and 2016 strains of H. pylori were evaluated. Of these, two studies compared the relationship between homB and clinical outcomes in both Western and Asian countries [24, 26], Six studies have been conducted on Western countries [21, 22, 25, 29,30,31], as well as four studies on the population of Asia [23, 27, 28, 32] (Table 1). Final results of some of eligible studies were contradictory and varied [22, 24, 26, 27].

Fig. 1
figure 1

The flowchart of study selection and included studies

Full size image

Association between homB and PUD

With the exception of the study by Casarotto et al., 11 studies had investigated the relationship between homB and PUD. The prevalence of homB in patients with GC and PUD was estimated at 54.4% [40.8–67.4 with 95% CIs; Effect size (with 95% CIs): 0.59 (0.18–1.00); I2: 96.1; Q-Value: 119.94; p value: 0.01; Begg’s p value: 0.21; Egger’s p value: 0.43]. However, the frequency of homB in patients with gastritis/NUD was estimated to be approximately 39.7% [27.6–53.2 with 95% CIs; Effect size (with 95% CIs): 0.43 (0.12–0.75); p value: 0.01; Begg’s p value: 0.26; Egger’s p value: 0.40]. We found that there was a significant relationship between the expression of homB gene and the progression of infection to PUD [OR: 1.36; 1.07–1.72 with 95% CIs; p value: 0.01; Effect size (with 95% CIs): 0.75 (0.28–1.38); p value: 0.01; I2: 81.41; Q-Value: 53.86; p value: 0.01; Begg’s p value: 0.87; Egger’s p value: 0.93]. According to the information, infection with homB-expressing strains appears to increase the risk of PUD (Fig. 2). Due to the high heterogeneity between studies, we used subgroup analysis to determine the role of homB in the development of primary infection to PUD in Western and Asian countries. Interestingly, a positive relationship was observed between the presence of homB gene and PUD in Western countries [OR: 1.61; 1.20–2.14 with 95% CIs; p value: 0.01; Effect size (with 95% CIs): 0.88 (0.44–1.43); p value: 0.01; I2: 71.65; Q-Value: 24.69; p value: 0.01; Begg’s p value: 0.90; Egger’s p value: 0.43] whereas, there was no meaningful relationship between this gene and PUD in Asian countries [OR: 0.89; 0.57–1.40 with 95% CIs; p value: 0.63; Effect size (95%CIs): 0.49 (0.23–0.65); p value: 0.01; I2: 83.60; Q-Value: 24.4; p value: 0.01; Begg’s p value: 0.11; Egger’s p value: 0.24]. Lack of access to raw data led to non-assessment the relationship between homB and duodenal and gastric ulcers.

Fig. 2
figure 2

Random-effects meta-analysis forest plot of the OR of peptic ulceration according to homB-positive H. pylori infection in eleven included studies

Full size image

Association between homB and GC

Five articles had evaluated the relationship between the homB gene and incidence of GC (low sample size), however, a positive association was observed between homB and GC [OR: 2.16; 1.37–3.40 with 95% CIs; p value: 0.01; Effect size (with 95% CIs): 0.54 (0.32–0.77); p value: 0.01; I2: 56.29; Q-Value: 9.15; p value: 0.05; Begg’s p value: 0.80; Egger’s p value: 0.77]. The summary of OR showed that the presence of homB gene significantly increases the incidence of GC (Fig. 3). In the subgroup analysis process, a positive relationship was observed between homB-positive strains and the risk of GC in Asian countries [OR: 3.71; 1.85–7.45 with 95% CIs; p value: 0.01; Effect size (with 95% CIs): 2.04 (0.33–3.72); p value: 0.01; I2: 57.12; Q-Value: 4.66; p value: 0.09]. Although a weak positive relationship was also observed between this gene and the incidence of GC in Western countries, but the threshold was not significant [OR: 1.42; 0.79–2.54 with 95% CIs; p value: 0.23; Effect size (with 95%CIs): 0.78 (0.27–0.83); p value: 0.01; I2: 0.00; Q-Value: 0.18; p value: 0.66]. Regarding the small number of included studies, many studies are needed to find the full relationship between homB gene and incidence of GC in patients infected with H. pylori.

Fig. 3
figure 3

Random-effects meta-analysis forest plot of the OR of gastric adenocarcinoma according to homB-positive H. pylori infection in five included studies

Full size image

Publication bias analysis

Publication bias was estimated based on both Begg’s p value and Egger’s p value tests, although no significant publication bias was observed. However, funnel plot asymmetry indicated a slight publication bias in the current meta-analysis.

Discussion

H. pylori is one of the most successful pathogens that colonizes the stomach of half the world's population. This bacterium can cause serious clinical consequences such as chronic gastritis, PUD, gastric atrophy and GC [34, 35]. According to documents, approximately 63% of GC cases worldwide are caused by H. pylori infection, and the bacterium is also responsible for 75% of gastric ulcers and 90% of duodenal ulcers [27]. The strains of this bacterium are genetically diverse and harbor different virulence genes [36, 37]. Studies in recent decades have shown that these genes are strain-specific (e.g. vacA, cagA, and omp) and play an important role in the immunopathogenesis of H. pylori and in the development of serious clinical outcomes [13, 17, 22, 27, 32, 38]. In several studies, the role of the homB gene in the pathogenesis of this pathogen was controversial; difference in results are related to differences in diet, environmental condition, hygiene status, age, socioeconomic level, and low sample size [29, 39]. Nevertheless, in the present study, we conducted a comprehensive literature review to assess the role of homB in the progression of primary infection to PUD and GC in Western and Asian countries. Oleastro et al. in their study showed that the presence of the homB gene is significantly higher than the homA gene in Portuguese children with PUD; “on” genotypes consistent cagA/vacAs1/ hopQI/oipA/homB strongly were associated with PU disease in children under four years of age [22]. In contrast, in studies on populations of Iraq, Turkey, and South Korea, none of homA and homB genes were correlated with PU disease [26, 40]. Interestingly, all studies in Western children have shown that the homB gene is associated with PUD, while the homA gene is more prevalent in the NUD [21, 22, 25]. In present study, frequency of homB gene in patients with affected to PUD and GC, severe clinical outcomes significantly was more prevalent than gastritis/NUD cases (54.4% and 39.7%, respectively). In addition, the summary OR showed that there was a significant relationship between homB-positive genotype and progression to PUD, especially in Western countries (OR: 1.16; 1.20–2.14 with 95% CIs; p value: 0.01), while in Asian countries there was no such relationship (OR: 0.89; 0.57–1.40 with 95% CIs; p value: 0.01). Therefore, our findings confirmed the results of previous studies. Also, strains isolated from Western countries contained two copies of the homB gene, but most infectious strains in Asian countries had only one copy of each of the homA and homB genes [19, 22, 31]. Related articles showed that the number of OMP copies also affects the status of bacterial compatibility and plays a role in the formation of clinical outcomes [14, 41, 42]. Recently, the role of homB as a cofactor in the increase of gastric adenocarcinoma in Asian countries has attracted much attention. The homB gene enhances the attachment of H. pylori to gastric epithelium, leading to dysregulation of normal signaling pathways and genetic instability [27, 40]. In addition, this gene increases the risk of GC through interferences such as inducing the inflammatory response, persistent infection, and gastric atrophy [23]. Abadi et al. showed that 78% of the strains isolated from GC patients contained the homB gene [27]. Jung et al. found that cagA-independent homB was associated with GC in Western countries [23]. However, in a study on the Chinese population, despite the presence of the homB gene in all isolated strains in patients with PUD and GC, no significant correlation was observed [39]. According to our results, a strong correlation was observed between the homB gene and the risk of GC in the Asian population (OR: 3.71; 1.85–7.45 with 95% CIs; p value: 0.01), whereas this correlation did not exist in Western countries (OR: 1.42; 0.79–2.54 with 95% CIs; p value: 0.66). Thus, depending on the geographical area, the homB gene appears to lead to PUD and GC in Western and Asian countries respectively. In several studies, correlation between homB and other virulence factors, especially vacA, cagA, oipA, hopQI, and babA in patients with PUD and GC was investigated. Sterbenc et al. in their study observed that there was no significant difference in the histopathological characteristics of PUD in both groups of children with and without the genotype profiles vacAs1m1/cagA/babA2/hompB [29]. Similar to this study, Oleastro et al. found that homB, independent of the cagA + /vacAs1 genotype profile, increases PUD risk in Western countries [24]. However, in other studies, it was shown that there is a significant correlation between homB and cagA, and homB also acts as a cofactor in complications such as PUD [23, 27, 40]. Due to the lack of raw data and uncertain results, we could not evaluate the relationship between homB and cagA in patients with PUD, but in GC cases, a weak correlation was observed (OR: 1.47; 0.95–2.28 with 95% CIs; p value: 0.79; I2: 80.15; Q-Value: 15.1; p value: 0.02; Egger’s p value: 0.07; Begg’s p value: 0.08). In the end it must be said, our study had several limitations such as low sample size, low number of included studies, inaccessibility to raw data, high heterogeneity in some studies, and also slight publication bias based on asymmetry of funnel plot. Hence, we need further studies to confirm the present findings.

Conclusions

Our results suggest that the presence of the homB gene in H. pylori strains contributes to the development of primary infection to severe clinical outcomes. We found that the homB gene could increase the risk of PUD in Western countries as well as GC in Asian countries.