Gastro-oesophageal reflux disease (GORD) refers to the reflux of gastric contents into the oesophagus, leading to mucosal damage and/or symptoms of heartburn and regurgitation. Although the symptoms of GORD are common, many patients do not have erosive oesophagitis diagnosed on endoscopy.¹ Proton pump inhibitors seem to be equally effective in treating non-erosive reflux disease (NERD) and erosive oesophagitis, indicating that excessive gastric acid secretion may be associated with both conditions.² Reflux disease is underlined by a number of well described pathophysiological mechanisms but recent interest has focused on the role of Helicobacter pylori in this condition. Specifically, there has been debate over whether this infection and its associated divergent patterns of gastritis may impact against reflux disease and its complications. Some reports have indicated a lower prevalence of H pylori infection in patients with GORD,^3–7 and particularly low prevalence of the virulent cagA positive strains in patients with complicated GORD.^4,5,8 However, there are also some studies that have found no causal relationship between H pylori infection and GORD.^9,10 A number of factors, including lack of use of validated reflux symptom questionnaires, differences in the definition of reflux, method of H pylori detection, differences in H pylori genotypes, inclusion of inappropriate control groups, lack of documentation of patterns of gastritis (that is, antrum predominant versus corpus predominant or pangastritis), and differences in host genetic factors probably account for the discrepancies between the various published reports.

H pylori infection can have varying effects on gastric acid secretion, and the changes largely depend on the pattern and severity of gastritis. Antrum predominant gastritis is associated with hypergastrinaemia, gastric hypersecretion, and duodenal ulcer disease, while patients with corpus predominant gastritis and/or multifocal atrophic gastritis have decreased acid secretion and increased risk of gastric ulcer and gastric carcinoma.^11–13 Host and environmental factors, as well as bacterial virulence characteristics, determine the development of H pylori associated upper gastrointestinal diseases.^14,15 Host genetic factors are key determinants of risk for many benign and malignant diseases, particularly in the stomach.^15–19

A variety of pro- and anti-inflammatory cytokines are expressed in H pylori infected gastric mucosa. One of the key cytokines that is upregulated in the gastric mucosa by infection is interleukin (IL)-1β.²⁰ This cytokine is important in initiating and amplifying the inflammatory responses against the bacterium and is also a potent inhibitor of gastric acid secretion.^21–24 The IL-1B gene encoding IL-1β is highly polymorphic, and several diallelic polymorphisms have been reported. Two of these are in the promoter region at positions −511 and −31, representing C-T and T-C transitions, respectively. Several studies have shown that these two polymorphisms are in near total linkage disequilibrium.^16,25 These polymorphisms have been shown to significantly affect gastric mucosal IL-1β production in response to H pylori infection.^26,27 The effect of these polymorphisms on gastric acid secretion is therefore most likely mediated through higher production of IL-1β, which is a potent inhibitor of acid secretion.^21,23 In addition to the IL-1 gene cluster, proinflammatory genotypes of tumour necrosis factor α (TNF-A) and IL-10 were each associated with the risk of non-cardia gastric cancer.^18,28

Recently, Queiroz et al reported that IL-1B and IL-1RN proinflammatory genotypes were associated with a decreased risk of erosive oesophagitis.²⁹ The reflux group in that study excluded NERD patients. In the current study, we hypothesised that the proinflammatory IL-1B polymorphisms might decrease the risk of GORD symptoms and erosive oesophagitis by increasing the development of hypochlorhydria and gastric atrophy, especially among H pylori infected subjects. We examined prospectively the effect of the proinflammatory cytokine polymorphisms (IL-1B−511 C/T, IL-10−819T/C, and TNF-A−1031T/C) on H pylori induced gastric atrophy, reflux symptoms, and erosive oesophagitis.

METHODS

Study subjects

We studied consecutive outpatients with dyspeptic symptoms who underwent upper gastrointestinal endoscopy between June 2002 and January 2003 at the Gastroenterology Department of Nagoya University Graduate School of Medicine (table 1). Patients who were found to have peptic ulcers, cancers, polyps, or any other specific lesions in the upper gastrointestinal tract were excluded. A total of 320 patients were recruited and these comprised subjects with macroscopically normal upper gastrointestinal mucosa or with gastritis and/or oesophagitis. None of these patients had taken non-steroidal anti-inflammatory drugs, proton pump inhibitors, antibiotics, or bismuth compounds in the three months preceding upper gastrointestinal endoscopy. Two biopsy specimens per site were obtained from the antrum and corpus and were used for routine histological examination (haematoxylin-eosin and Giemsa stains). H pylori infection was assessed by ¹³C urea breath testing, histological examination by Giemsa stain, and the presence of serum IgG antibody against H pylori. All subjects gave written informed consent, and the study was approved by the ethics committee of the Nagoya University Graduate School of Medicine.

Evaluation of GORD symptoms

GORD symptoms in all patients were evaluated using the Carlsson-Dent reflux scale.³⁰ This scale uses seven items to evaluate the nature of the sensations experienced by patients and the temporal relationship of symptom occurrence to factors that are known to provoke (meals, bending, stooping, lifting), exacerbate (fatty or spicy food), or relieve (antacids) gastro-oesophageal reflux. Each response was assigned a positive, neutral, or negative score, and the score for each item was weighted so that the highest positive values were assigned to factors considered strongly indicative of the diagnosis of GORD. A score ranging from −7 to +18 was calculated by adding the individual positive and negative scores from the questionnaire. Patients were regarded as having reflux symptoms if their scores were higher than 6. This cut off point was estimated by the authors to have a sensitivity of 54 % and specificity of 60%.

Endoscopic study

A total of 320 patients with or without oesophagitis were included (table 1). Oesophagitis was assessed by endoscopic examination, and mucosal breaks were graded as A, B, C, or D according to the Los Angeles classification of oesophagitis³¹: grade A: one or more mucosal breaks confined to the mucosal folds, each no longer than 5 mm; grade B: at least one mucosal break more than 5 mm long confined to the mucosal folds but not continuous between the tops of two mucosal folds; grade C: at least one mucosal break continuous between the tops of the two or more mucosal folds but not circumferential; and grade D: circumferential mucosal break.

H pylori and CagA antibody

Fasting blood samples were analysed for H pylori antibody and CagA status. Serum was sampled before endoscopy and all samples were stored at −70°C until they were analysed in the same batch. Anti-H pylori IgG antibody tests, high molecular weight campylobacter associated protein with antigens extracted from clinically isolated Japanese H pylori strains (J-HM-CAP), and enzyme linked immunosorbent assay (ELISA) (Kyowa Medex, Tokyo, Japan) were used for identification of H pylori infected subjects. An ELISA value of 2.3 or over was regarded as positive for both tests. CagA serology was performed by ELISA, with horseradish peroxidase as the enzyme tracer (CagA IgG EIA; Sceti Co. Ltd, Tokyo, Japan). The anti-CagA IgG antibody concentration in standards and samples was measured in a spectrophotometer at 450 nm.

Histology

Glandular atrophy was assessed by a single expert gastrointestinal pathologist who used the updated Sydney system³² without knowledge of the experimental results. An assessment of H pylori was made after staining of serial tissue sections with Giemsa.^33,34 Two biopsies per site (antrum and corpus) were used and when the scores between the two biopsies differed, the more severe scores were selected.

Genetic analysis

DNA was extracted from 200 μl of buffy coat preserved at −40°C by the use of QIAamp DNA Blood Mini Kit (Qiagen Inc., Valencia, California, USA). Polymerase chain reaction (PCR) amplification was conducted using the primers listed in table 2. Genomic DNA (50 ng) was used in a volume of 25 μl with 0.1 mM dNTPs, 25 pmol of each primer, 0.5 units of “Takara Taq” (Takara Shuzo Co. Ltd, Otsu, Japan) or “Ampli Taq Gold” (Perkin-Elmer Corp., Foster City, California, USA), and 2.5 μl of 10× PCR buffer, including 15 mM MgCl₂. The PCR-restriction fragment length polymorphism method was used for genotyping the IL-1B−511 C/T polymorphism. For genotyping the IL-10−819T/C and TNF-A−1031T/C polymorphism, a PCR-confronting two pair primers method was applied. This method for single nucleotide polymorphism genotyping does not require a step to digest DNA products (table 2).³⁵ All PCR products were visualised on a 2% agarose gel with ethidium bromide staining.

Statistical analysis

Statistical analysis was performed using the χ² test, and an unconditional logistic model was applied for estimating odds ratios (ORs) and interaction terms by a computer program (STATA version 6; STATA Corp., College Station, Texas, USA). The product variable between gene and gene was included in the logistic model when evaluating the multiplicative interactive effects of selected gene pairs. ORs for erosive oesophagitis, GORD symptoms, and the presence of corpus glandular atrophy were adjusted for age and sex. A p value < 0.05 was considered statistically significant.

RESULTS

A total of 320 patients with or without oesophagitis were studied, with a mean age of 46 years (range 35–60). Of the 320 patients studied, 130 (41%) satisfied the criteria for GORD symptoms (table 1). A total of 208 patients (65%) were H pylori positive and 112 were negative. Two hundred and four (98%) of 208 H pylori infected patients were CagA positive. One hundred and thirty three (42%) of 320 patients had glandular atrophy in the gastric corpus.

Endoscopic findings for erosive oesophagitis

A total of 108 patients (34%) were found to have erosive oesophagitis endoscopically, according to the Los Angeles classification of oesophagitis: grade A: 78 (72%); grade B: 23 (21%); grade C: 6 (6%); and grade D: 1 (1%) (table 2). Erosive oesophagitis was significantly more common in H pylori negative compared with H pylori positive patients (59/112 = 53% v 49/208 = 24%; p<0.05) (tables 2, 3). GORD symptoms were reported in 83 (77%) of 108 patients with erosive oesophagitis compared with 47 (22%) of 212 without erosive oesophagitis (p<0.05).

Effect of H pylori infection on pattern of gastric atrophy and erosive oesophagitis

Glandular atrophy was assessed histologically in the gastric antrum and corpus, respectively. Subjects were divided into two groups: one with corpus atrophy and one without corpus atrophy. The latter group included subjects without atrophy and subjects with antrum atrophy only. A total of 133 of 320 (42%) patients had corpus atrophy and all of these patients also had antrum atrophy. H pylori positive patients were more likely to have corpus atrophy than H pylori negative patients. Of the 208 who were H pylori positive, 128 (62%) had corpus atrophy compared with only five of 112 H pylori negative patients (4%) (p<0.001) (table 3). Among H pylori positive patients, the presence of corpus atrophy was protective against erosive oesophagitis. Of the 159 H pylori positive patients who had no erosive oesophagitis, 110 (69%) had corpus atrophy. In contrast, of the 49 H pylori positive patients with erosive oesophagitis, only 18 (37%) had corpus atrophy (p<0.05).

Effect of the IL-1B−511C/T, IL-10−819T/C, and TNF-A−1031T/C polymorphisms on corpus gastric atrophy

Of the 320 patients participating in this study, 247 (190 H pylori positive and 57 H pylori negative) were approved for genetic studies and their samples were analysed for the IL-1B−511C/T, IL-10−819T/C, and TNF-A−1031T/C polymorphisms. The IL-1, IL-10, and TNF-A markers had no impact on risk of H pylori infection, with the different genotypes being equally prevalent among infected and non-infected subjects (table 4). Among H pylori positive patients, there were significantly more subjects with corpus gastric atrophy among those homozygous for the proinflammatory allele IL-1B−511T compared with those homozygous for the C allele (OR 4.85 (95% CI 1.86–12.6); p = 0.001). This suggests that a proinflammatory IL-1B profile is associated with an increased risk of gastric atrophy. None of the other polymorphisms had any significant effect on the risk of corpus gastric atrophy (table 5), and there were no interactions with the IL-1B−511 genotypes.

IL-1B−511C/T, IL-10−819T/C, and TNF-A−1031T/C polymorphisms and GORD

Among H pylori positive patients, subjects homozygous for the proinflammatory alleles IL-1B−511T had a significantly lower risk of erosive oesophagitis (OR 0.05 (95% CI 0.01–0.26); p<0.001) (table 6) and GORD symptoms (OR 0.14 (95% CI 0.03–0.64); p = 0.01) (table 7) compared with those homozygous for the −511C alleles. Comparing IL-1B−511T/T relative to C/C, age, sex, H pylori, and atrophy adjusted OR for GORD was 0.13 (95% CI 0.03-0.48). Taking gastric atrophy out of the model, age and H pylori adjusted OR for GORD comparing IL-1B−511T/T relative to C/C was 0.09 (95% CI 0.25–0.32). The results indicate that IL-1B, H pylori, and atrophy are independent markers of decreased risk of GORD.

None of the other cytokine gene polymorphisms had any significant correlation with erosive oesophagitis or GORD symptoms and there was no apparent interaction between the different genetic markers.

DISCUSSION

In the present study, we have shown that both endoscopic (erosive oesophagitis) and symptomatic (NERD) manifestations of GORD are significantly more common among H pylori negative subjects compared with those who are H pylori positive. We have also shown that subjects with evidence of H pylori induced gastric atrophy (documented by histological assessment) are protected from both erosive oesophagitis and NERD. The most novel finding in our study is the demonstration that both gastric atrophy and the associated protection from GORD complications are mediated by an increased prevalence of the proinflammatory IL-1B genetic profile. The work presented here complements the recent report by Queiroz et al who showed that the IL-1B−511/−31 proinflammatory alleles protect against erosive oesophagitis.²⁹ We have further expanded these findings by showing that the link between these proinflammatory IL-1B polymorphisms and protection from GORD is mediated through induction of gastric atrophy. This is the first demonstration of a direct link between H pylori infection, host genetic makeup, gastric atrophy, and protection from GORD and its complications. The findings are likely to have a significant impact on the understanding of the pathogenesis of GORD and the role of H pylori.

The most likely explanation for the protective role of these IL-1B genotypes against erosive oesophagitis in H pylori infected subjects is their association with increased production of IL-1β. Apart from being a very potent proinflammatory cytokine that initiates and amplifies the host’s immune response against H pylori infection, it is also a powerful acid inhibitor.²⁴ Thus high production of this cytokine in the gastric mucosa of infected subjects is associated with functional inhibition of gastric acid secretion, which is followed in time by permanent loss of acid secretory capacity through the destructive effects of chronic inflammation on parietal cells. Reduction in gastric acidity is the logical explanation for protection against GORD, a primarily acid related condition. There is evidence that the proinflammatory IL-1B genotypes are associated with hypochlorhydria in both Caucasian and Japanese populations.^16,19 Furuta and colleagues¹⁹ reported that Japanese subjects with the IL-1B−511 T/T genotype had a median gastric juice pH of 6.5 (range 4.2–7.3) (that is, almost neutral). The same genotype was associated with the highest atrophy scores and lowest pepsinogen I/II ratios. Thus it seems very likely that the reported protective effect of these same IL-1B genotypes in this study is also mediated through lower acid secretion due to corpus atrophy.

It is now well established that the damaging effects of the host proinflammatory polymorphisms on the gastric mucosa are significantly increased in the presence of virulent strains of H pylori. Figueiredo et al have shown that for each combination of bacterial/host genotype, the odds of having gastric carcinoma were greatest in those with both bacterial and host high risk genotypes.^17,18,27 The high risk combination is likely to not only increase the risk of gastric atrophy and cancer but also reduce the risk of developing GORD and its complications.

H pylori infection has divergent clinical outcomes that are in some instances mutually exclusive. Thus patients who develop duodenal ulcer disease are protected from gastric cancer.³⁶ Although both conditions are associated with H pylori, the pathophysiological consequences of the infection are diametrically opposite. Duodenal ulcers are associated with an antrum predominant pattern of gastritis, low prevalence of gastric atrophy, and very high acid secretion. In contrast, gastric cancer patients develop severe pan- or corpus predominant gastritis, multifocal atrophic gastritis, and hypo or achlorhydria. Thus it is apparent that an outcome characterised by low acid secretion protects from acid related disorders such as duodenal ulcers. This pattern also protects from GORD, a condition in which the lower oesophagus has increased exposure to gastric acid. There have been several reports in the literature indicating that infection with virulent strains of H pylori (for example, cagA⁺ strains) protects from the more severe outcomes of GORD, such as Barrett’s oesophagitis, oesophageal strictures, dysplasia, and oesophageal adenocarcinoma.^4,5,8 It has been speculated that the higher risk of developing gastric atrophy and its associated reduction in gastric acid secretion by cagA⁺ strains explain why these patients are protected from GORD and its complications. It is now apparent that a major factor in this sequence of events is the host genetic constitution. In our population (and in common with most Asian populations), all subjects were infected with cagA⁺ strains yet there was still a divergence of outcomes in relation to GORD. By examining the host genetic markers relating to IL-1β, we could clearly demonstrate that the proinflammatory acid inhibiting and atrophy inducing genotypes were crucial for development of the protective phenotype. Another important observation in our study was the high prevalence of gastric atrophy. Based on histological assessment, 42% of the examined subjects had evidence of atrophy in the corpus as well as in the antrum. The findings in this study confirm that progression of gastric atrophy may protect from erosive oesophagitis in H pylori positive individuals.

Demonstration that patients with a proinflammatory IL-1B genetic makeup have increased H pylori induced atrophy and a reduced risk of GORD must sound a cautionary note about the potential consequences of eradicating the infection in such subjects. It is entirely possible that eradication of the infection in those with a proinflammatory profile would allow the corpus mucosa to recover sufficiently so that acid secretion returns. In such circumstances, one might expect an increased incidence of GORD in those who have a predisposition to developing reflux. While this may be adequately handled with acid suppressive therapy, it is essential that health care providers are aware of the potential for the increased incidence of GORD and its possible complications. This point is particularly relevant if one contemplates eradication of the infection on a large population wide scale, particularly in areas of high gastric cancer incidence. There is currently no convincing evidence that H pylori eradication either increases the incidence of de novo GORD or worsens pre-existing GORD.³⁷ However, most of these studies have come from Western countries in which gastric atrophy and hypochlorhydria are much less common than in countries such as Japan, China, and Korea. In these Asian countries, the outcome might be different and it is essential that this be taken into consideration when planning a global eradication strategy. It is apparent and indeed very desirable that prospective studies addressing these issues should be undertaken in Asian countries.

In conclusion, we have shown that proinflammatory genotypes of the IL-1B gene increase the risk of gastric atrophy and reduce the risk of erosive oesophagitis and GORD symptoms, most probably through induction of gastric atrophy and gastric acid inhibition. These findings are very relevant to understanding the pathogenesis of GORD and the role of H pylori infection. They are also relevant to the ongoing debate about the long term consequences of H pylori eradication in the context of gastric cancer prevention.