SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. GENERAL CONSIDERATIONS AND SUMMARY
  8. Acknowledgements
  9. References

Background:

Long-term acid suppression may accelerate the development of atrophic gastritis in Helicobacter pylori-positive subjects. The pathogenetic mechanism remains unclear.

Aim:

To test the hypothesis that gastric double infection with H. pylori and non-H. pylori bacterial species—during acid suppression—may result in an enhanced inflammatory response, contributing to the development of atrophic gastritis.

Patients and methods:

A consecutive series of patients with gastro-oesophageal reflux disease undergoing treatment with proton pump inhibitors (n=113) or histamine2-receptor antagonists (H2-RAs) (n=37), and 76 non-treated dyspeptic controls were investigated. Gastric mucosal H. pylori and non-H. pylori bacteria, histological gastritis, H. pylori serology, and circulating interleukin (IL)-1β, IL-6, and IL-8 were examined.

Results:

Patients on acid suppression with either proton pump inhibitors or H2-RAs had a similar prevalence of H. pylori infection to the controls, but a higher prevalence of non-H. pylori bacteria (61% and 60% vs. 29%, P < 0.0001 and P < 0.002). Both the presence of H. pylori and non-H. pylori bacteria were independent risk factors of atrophic gastritis (antrum: relative risks (RRs), 10.1 and 5.07; corpus: RRs, 11.74 and 6.38). A simultaneous presence of H. pylori and non-H. pylori bacteria was associated with a markedly increased risk of atrophic gastritis (antrum: RR, 20.25; corpus: RR, 20.38), compatible with a synergistic effect. Furthermore, the simultaneous presence of both types of bacteria was associated with higher cytokine levels than in patients without any type of bacteria. This increase was also greater than in patients with H. pylori infection alone (P < 0.001, for both IL-1β and IL-8).

Summary and conclusions:

H. pylori-positive patients on long-term acid inhibition displayed three features: non-H. pylori bacterial growth; increased cytokine levels; and a higher risk of atrophic gastritis. We suggest that double infection with H. pylori and non-H. pyloribacteria is a major factor in the development of atrophic gastritis during gastric acid inhibition.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. GENERAL CONSIDERATIONS AND SUMMARY
  8. Acknowledgements
  9. References

The identification of Helicobacter pylori has considerably improved our understanding of the pathogenesis of chronic atrophic gastritis.1, 2 Infection with H. pylori causes chronic active inflammation of the gastric mucosa, which may progress towards atrophic gastritis and intestinal metaplasia, conditions that increase the risk of gastric cancer.3, 4 The clinical outcome of the infection is driven by a complex interplay between bacterial pathogenic factors, host response, and environmental factors.5, 6

Current evidence indicates that the anatomical distribution and severity of H. pylori gastritis are strongly influenced by the individual gastric acid secretory status.7, 8 Subjects with a normal or high level of acid output are prone to develop antral predominant gastritis and duodenal ulcer disease. In contrast, subjects with a low level of acid output are prone to develop a corpus (body) predominant gastritis, leading to glandular atrophy and the risk of gastric cancer. Such corpus predominant gastritis is frequently seen when H. pylori-infected subjects are treated with acid-suppressive medication.9[10][11]–12 At present, the mechanism underlying the accelerated development of atrophic gastritis in H. pylori-infected patients treated with acid inhibition is still unclear. It has been suggested that an increase in antral pH impairs local growth conditions for H. pylori, causing a ‘shift’ of the bacterium to the corpus mucosa, where the acid secretion is still preserved.9 Furthermore, decreased gastric acid secretion may impair the dilution and washout of potentially cytotoxic and pro-inflammatory agents induced by H. pylori, which in turn may cause mucosal damage.11, 13

On the other hand, acid suppression alters the intragastric milieu, favouring the colonization by non-H. pylori bacterial species; aerobic and to a minor degree anaerobic, gram-positive and gram-negative bacterial strains have been consistently isolated.14[15][16][17]–18 The intragastric bacterial overgrowth and its potential consequences have received considerable attention in the past.14 So far, it is not known whether such bacteria could play a role in the development of atrophic gastritis during acid-suppressive therapy.

Infection with H. pylori induces mucosal over-expression of various pro-inflammatory cytokines, such as interleukin (IL)-1β, IL-6, IL-8, and tumour necrosis factor-α; this is accompanied by gastric morphological changes.19[20]–21 Experimental evidence shows that commensal bacteria and their by-products can also elicit cytokine production.22, 23 In this context, the following question emerged: Does the gastric double infection with H. pylori and non-H. pylori bacteria cause an enhanced inflammatory response, leading to increased risk for the development of atrophic gastritis?

In a large series of patients on continuous acid-suppressive therapy and of non-treated controls, we tested this hypothesis by studying intragastric mucosal bacterial flora in relation to the profile of circulating pro-inflammatory cytokines (IL-1β, IL-6, and IL-8) and to gastric mucosal histology. Special attention was paid to the development of corpus atrophic gastritis in this clinical setting.

PATIENTS AND METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. GENERAL CONSIDERATIONS AND SUMMARY
  8. Acknowledgements
  9. References

Study population

Between January 1997 and June 1998, 226 consecutive dyspeptic patients referred for upper gastrointestinal endoscopy to an open-access unit (University Hospital Maastricht, the Netherlands) were investigated in a cross-sectional design. Case subjects were patients on continuous acid-suppressive therapy for gastro-oesophageal reflux disease (GERD), with or without Barrett’s oesophagus. Continuous acid-suppressive therapy was defined as at least one daily dose of either a proton pump inhibitor or a histamine2-receptor antagonist (H2-RA), taken for either 6 weeks to 1 year (medium-term) or longer than 1 year (long-term), respectively. Control subjects were dyspeptic patients with a similar distribution of age and gender, who did not receive any acid-suppressive medication or antacids before referral, and who had normal endoscopic findings.

The exclusion criteria for all subjects were: present or past history of peptic ulcer disease; previous gastric surgery and/or vagotomy; prior H. pylori eradication therapy; treatment with antimicrobial agents or prokinetic drugs within 30 days before endoscopy; chronic use of NSAIDs; and any condition suspected or confirmed to influence the cytokine production, such as an acute infection in the previous 30 days, non-infectious inflammatory diseases, neoplasms, liver or renal diseases, and use of immuno-suppressive therapy. The study was approved by the Ethical Review Board of the University Hospital Maastricht and each subject gave informed consent before being enrolled into the study.

Study design

General clinical records and a structured interview concerning the history of reflux symptoms, type, daily dose, and length of antisecretory therapy, any co-morbidity/co-medication, and family history of gastric cancer were obtained from all patients. Endoscopies were performed after an overnight fast, using a carefully disinfected Pentax EG-2901 instrument. During endoscopy, fasting gastric juice pH was measured using pH paper strips with grading steps of 0.5 from pH 0 to pH 14 (Schleicher & Schüll GmbH; Dassel, Germany). Two antral biopsies (2 cm proximal to the pylorus) and four corpus biopsies (10 cm below the gastro-oesophageal junction, along the greater curvature) were sampled for histological examination. Fasting blood samples were obtained for the determination of H. pylori status and circulating cytokines. All blood samples were drawn between 09.00 hours and 11.00 hours.

Serological tests

The H. pylori status of the patients was determined using an immunoblot method (Helicoblot 2.0, Imphos BV, Zambon Group, Amersfoort, the Netherlands). Strips were incubated with serum samples, diluted 1 : 100. After washing, they were treated with antihuman IgG antibody conjugated with alkaline phosphatase and the substrate BCIP/NBT. Bands were present for 19.5, 26.5, 30, 35, 89, and 116 kDa (CagA) proteins. H. pylori positivity was assessed according to the manufacturer’s instructions. H. pylori-positive and -negative control sera were included each time the test was performed.

Circulating IL-1β, IL-6, and IL-8 levels were determined by ELISA (Cytoscreen human IL-1β ELISA kit, BioSource International, California, USA; IL-6 ELISA kit IBL-Hamburg, Germany, and Genzyme IL-8 ELISA kit, Cambridge, USA, respectively). The detection limit for all these assays is 1 pg/mL. The intra-assay variations of IL-1β, IL-6 and IL-8 are less than 5%, 4.5% and 6.5%, whilst the inter-assay variations of these cytokines are less than 7.5%, 5.5% and 10.5%. All samples were examined in duplicate and without knowledge of the patients’ clinical and histological data.

Histology

Serial sections (4 μm) from formalin-fixed and paraffin-embedded specimens were prepared for haematoxylin-eosin stain (H&E), modified Giemsa stain (MG), and an immunohistochemical stain (IMM).

For the detection and histological classification of the current intragastric bacterial flora, the IMM and MG were used in combination, as previously detailed.24[25]–26 IMM with a polyclonal antiserum (DAKO B471, ITK diagnostics BV, Uithoorn, the Netherlands) in a dilution of 1 : 100 was applied to detect H. pylori.24 The presence of non-H. pylori bacteria was screened in the MG stain—on the basis of morphology and location—and ascertained by a negative IMM stain of an adjacent section, on the same position, as detailed elsewhere.25, 26 At least two adjacent sections of all biopsies were examined completely at 400 × magnification; if necessary a 1000 × magnification was used. This technique was validated against culture methods, in groups of patients with different gastric conditions, as well as in the context of acid inhibitory treatment.24, 26

The histomorphologic status of the gastric mucosa was evaluated according to the Sydney classification.27[28][29]–30 The presence of active inflammation (i.e. neutrophilic cell infiltration), chronic inflammation (i.e. mononuclear cell infiltration), glandular atrophy (i.e. loss of appropriate glands), and intestinal metaplasia was graded as absent, mild, moderate or severe changes. The density of non-H. pylori bacterial colonization was scored semi-quantitatively, as follows: grade 0 (no bacteria detected); grade 1 (occasionally bacteria detected); grade 2 (scattered bacteria identified in several high power fields); and grade 3 (large amounts detected in many high power fields).28

For the purpose of this study, the following sub-types of chronic gastritis were considered: diffuse non-atrophic gastritis, defined as active and/or chronic inflammation of either antral, corpus or both antral and corpus mucosa, without any signs of glandular atrophy; antral predominant atrophic gastritis, defined as glandular atrophy detected only in the antrum or in both antrum and corpus, but with a higher degree of active or chronic inflammation in the antrum; and corpus predominant atrophic gastritis, defined as glandular atrophy detected only in the corpus or in both the antrum and corpus, but with a higher degree of active or chronic inflammation in the corpus. All biopsy specimens were reviewed by one experienced gastroenterological pathologist (A.B.) and by the main investigator (S.S.), without knowledge of the patients’ clinical and serologic data. In the event of discordant results (less than 5% of all cases), the specimens were re-examined by a second pathologist and were discussed until agreement was reached.

Statistical evaluation

Differences in dichotomous variables were examined using the χ2-test or Fisher’s exact test, when appropriate. Differences in continuous variables were examined using the Kruskal–Wallis test for multiple comparisons and the Mann–Whitney U-test for two group comparisons. Correlations were performed with the Spearman rank test. The concordance between histological (IMM) and serological evidence of H. pylori was tested with k statistics (k index). Univariate analyses using, as the outcome variable, the presence of atrophic gastritis were computed for each potential risk factor: age, gender, duration of acid-suppressive therapy, CagA positivity, presence of H. pylori alone, of non-H. pylori alone, and of both types of bacteria. Multiple logistic regression analyses were carried out to determine the independent contribution of each factor to the presence of atrophic gastritis.31 Interaction was studied with standard methods.32 The relative excess risk due to an interaction was calculated as:

inline image

where A and B are the factors whose interaction is being studied. The synergy index was calculated as:

inline image

In the multiple regression analysis, a relative excess risk > 0 and a synergy factor > 1 were regarded as indices of interaction between two risk factors. Two-sided P-values < 0.05 were assumed to indicate statistical significance.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. GENERAL CONSIDERATIONS AND SUMMARY
  8. Acknowledgements
  9. References

Table 1 shows the demographic, clinical, and endoscopic characteristics of the 226 patients, classified according to the type of acid-suppressive therapy they were undergoing. The age distribution and sex ratios were comparable among the three patient groups: non-treated controls, H2-RA group, and proton pump inhibitor group.

Table 1.   Demographic, clinical and endoscopic characteristics of the 226 patients enrolled into the study Thumbnail image of

Median gastric juice pH levels (range in brackets) in H. pylori-positive patients were 3.1 (2.5–3.8) in non-treated controls, 4.0 (2.9–4.9) in the H2-RA group, and 5.6 (4.8–6.5) in the proton pump inhibitor group. The corresponding median gastric juice pH levels in H. pylori-negative patients were 2.2 (1.8–2.7), 3.0 (2.3–3.6) and 4.8 (4.2–5.2), respectively.

Prevalence and distribution of intragastric bacteria in the gastric mucosa

Figure 1 illustrates the prevalence of H. pylori and non-H. pylori bacteria in the gastric mucosa. The patients on acid suppression with either proton pump inhibitors or H2-RAs had a similar prevalence of H. pylori infection to non-treated controls, but a higher prevalence of non-H. pylori bacteria (61% and 60% vs. 29%, P < 0.0001 and P < 0.002). No significant difference in this regard was found between the proton pump inhibitor group and the H2-RA group.

image

Figure 1.  Prevalence of H. pylori and non-H. pylori bacteria in the gastric mucosa of 226 subjects enrolled into the study.

Download figure to PowerPoint

The distribution of H. pylori between antral and corpus mucosa was similar in patients treated with H2-RAs and controls, whilst patients treated with proton pump inhibitors had a lower prevalence of H. pylori colonization in the antral than corpus mucosa (27% vs. 35%, P=0.033; Table 2). There were no significant differences in the prevalence of non-H. pylori bacteria between the antral and corpus mucosa in each of the three patient groups.

Table 2.   Distribution pattern and density of colonization by H. pylori and non-H. pylori bacteria Thumbnail image of

The patients on acid suppression with either proton pump inhibitors or with H2-RAs also had higher median scores (grade 1–3) of non-H. pylori bacterial density—in both the antrum and corpus mucosa—than the controls (antrum: P=0.012 and P=0.003; corpus: P=0.05 and P=0.028, respectively; Table 2). No significant differences in this regard were found between the proton pump inhibitor group and the H2-RA group, or between the antral and corpus mucosa within each of the three patient groups.

Intragastric bacteria in relation to chronic gastritis

Overall, 39 (17%) patients had diffuse non-atrophic gastritis, 44 (20%) patients had antral predominant atrophic gastritis, and 28 (12%) had corpus predominant atrophic gastritis, whilst the remaining 115 (51%) patients had normal gastric histology. In patients with antral predominant atrophic gastritis, antral gland atrophy was mild in 22 cases, moderate in 14 cases, and severe in eight cases. In patients with corpus predominant atrophic gastritis, corpus gland atrophy was mild in six cases, moderate in 12 cases, and severe in 10 cases. Of the 44 patients with antral predominant atrophic gastritis, 38 had intestinal metaplasia, which in 28 cases was mild, in six cases was moderate, and in the remaining four was severe. Of the 28 patients with corpus predominant atrophic gastritis, 20 had intestinal metaplasia, which in 10 cases was mild, in six cases was moderate, and in four cases was severe.

In patients with antral predominant atrophic gastritis, H. pylori only, non-H. pylori only and both types of bacteria were detected in 15 (34%), 10 (23%), and 17 (39%) of the 44 cases. The corresponding figures in patients with corpus predominant atrophic gastritis were nine (32%), seven (25%), and 11 (39%) of the 28 cases, respectively.

Overall, there was a strong concordance between histology (IMM) and serology for the presence of H. pylori (controls, 95% agreement, k=0.89; H2-RAs, 94% agreement, k=0.84; proton pump inhibitors, 90% agreement, k=0.81). CagA antibodies were found in 62 (70%) of the histologically H. pylori-positive subjects (22 out of 32, 69%, controls; 11 out of 15, 73%, patients on H2-RAs, and 29 out of 42, 69%, patients on proton pump inhibitors). In histologically H. pylori-positive subjects, antral predominant atrophic gastritis was more common in CagA-positive than CagA-negative subjects (P=0.024), but no such relationship was found between CagA positivity and corpus predominant atrophic gastritis (P=0.42).

In Table 3, the results of the multiple logistic regression analysis examining independent risk factors for the development of atrophic gastritis are detailed. Age was associated with a minor, but significant, risk of atrophic gastritis, whilst gender did not influence this outcome. In contrast to antral atrophic gastritis, the risk of corpus atrophic gastritis increased with the duration of acid inhibition, as expressed per year of treatment. The corresponding relative risk after 10 years of therapy can be estimated at 3.16 (95% CI: 2.79–3.58). CagA positivity was a risk factor for antral atrophic gastritis, but not for corpus atrophic gastritis.

Table 3.   Results of multivariate analyses to determine the independent risk factors of atrophic gastritis in the antrum and corpus Thumbnail image of

H. pylori infection alone was an independent risk factor of atrophic gastritis in both the antrum and corpus. The presence of non-H. pylori bacteria alone independently increased the risk of antral atrophic gastritis, but only as a trend in the gastric corpus. The simultaneous presence of H. pylori and non-H. pylori bacteria was associated with the highest risk of atrophic gastritis in the antrum and corpus when compared to the risk by either type of bacteria alone. The relative excess risk resulting from the interaction between H. pylori and non-H. pylori bacteria was 6.08 for antral predominant atrophic gastritis and 3.26 for corpus predominant atrophic gastritis, compatible with a synergistic effect (synergy index of 1.46 and 1.26, respectively). Analysis of the data based on serological (instead of histological) evidence of H. pylori provided similar results (data not shown).

Circulating cytokines in relation to intragastric bacteria and chronic gastritis

In the entire series, four sub-groups of patients were distinguished according to the presence of H. pylori and of non-H. pylori bacteria in the gastric mucosa: patients without any type of bacteria (64 cases); those with H. pylori alone (49 cases); those with non-H. pylori bacteria alone (73 cases); and those with both types of bacteria (40 cases). Patients infected with H. pylori alone had higher levels of IL-1β and IL-8 (Figures 2A and B) than those without any type of bacteria (P < 0.0001, for both cytokines). Patients with non-H. pylori bacteria alone had slightly higher IL-1β and IL-8 levels than those without any type of bacteria, but differences reached statistical significance for IL-8 only (P < 0.05). Patients harbouring both H. pylori and non-H. pylori bacteria had threefold higher levels of circulating IL-1β and IL-8 than those without any type of bacteria (P < 0.0001, for both cytokines); this increase was also greater than that in patients with H. pylori infection alone (P < 0.001, for both cytokines).

image

Figure 2.  Individual values of circulating IL-1β (A) and IL-8 (B) in subjects with H. pylori alone, non-H. pylori alone, or with both types of bacteria, compared to those without bacteria in the gastric mucosa. Median values are shown by horizontal lines. Numbers of subjects in parentheses.

Download figure to PowerPoint

Table 4 describes the relationship between cytokine levels and different types of chronic gastritis. Patients with diffuse non-atrophic gastritis, patients with antral predominant atrophic gastritis, and especially those with corpus predominant atrophic gastritis had significantly higher circulating IL-1β, IL-6, and IL-8 levels than those with normal gastric histology. As illustrated in Figure 3, in patients with antral predominant atrophic gastritis or with corpus predominant atrophic gastritis, cytokine levels (particularly IL-8) were highest in patients with double infection and severe atrophic gastritis. As shown in Figure 4, in patients with corpus predominant atrophic gastritis, cytokine levels positively correlated with the histological degree of glandular atrophy. Additionally, CagA-positive subjects had significantly higher IL-8 levels than CagA-negative subjects (3.6 pg/mL [range 0–12.9 pg/mL] vs. 2.4 pg/mL [0–10.4 pg/mL], P=0.012), but no such relationship was found between CagA positivity and IL-1β (P=0.12) or IL-6 (P=0.26) levels.

Table 4.   Circulating cytokines in relation to different types of chronic gastritis Thumbnail image of
image

Figure 3.  Relationship between intragastric bacteria, circulating cytokines (IL-8 level), and the histological degree of atrophic gastritis. Median IL-8-values are shown by columns. Numbers of subjects are shown in parentheses. Patients with antral predominant atrophic gastritis and with corpus predominant atrophic gastritis are shown together.

Download figure to PowerPoint

image

Figure 4.  Correlation between circulating IL-1β (□), IL-6 (○), IL-8 (▵), and the histological degree of corpus predominant atrophic gastritis. Numbers of subjects in parentheses.

Download figure to PowerPoint

Figure 5 shows the relationship between the presence of non-H. pylori bacteria and of corpus predominant atrophic gastritis (left scale), cytokine levels (right scale), and duration of treatment with proton pump inhibitors, in H. pylori-negative (A) and -positive (B) subjects on acid inhibition vs. H. pylori-negative and -positive non-treated controls. Five sub-groups of patients were distinguished according to the length of treatment they had undergone (from ‘none’ to ‘> 36 months’). The age distribution was similar across the five sub-groups of patients with or without H. pylori infection. In H. pylori-negative patients treated with proton pump inhibitors, the prevalence of non-H. pylori bacteria increased with the longer duration of treatment; this was paralleled by a low prevalence of corpus atrophic gastritis (affecting one out of 44 [2.2%] controls, and one out of 71 [1.4%] patients on proton pump inhibitors), and by nearly stable cytokine levels. In contrast, in H. pylori-positive patients treated with proton pump inhibitors, the prevalence of non-H. pylori bacteria increased with the longer duration of treatment. This was paralleled by an increasing prevalence of corpus atrophic gastritis (affecting nine out of 10 [90%] patients who had been taking proton pump inhibitors for longer than 36 months), and by increasing cytokine levels. Similar figures applied to both types of acid-suppressive therapy (proton pump inhibitors and H2-RAs). However, these reached statistical significance for the proton pump inhibitor group only, probably because of the small number of patients in the H2-RA group (figure not shown).

image

Figure 5.  Relationships between the presence of non-H. pylori bacteria (░), of corpus predominant atrophic gastritis (▮) (left scale), cytokine levels (right scale), and duration of treatment with proton pump inhibitors in H. pylori-negative (A) and -positive (B) subjects on acid inhibition vs. H. pylori-negative and -positive non-treated controls. Median levels of circulating IL-1β (□), IL-6 (○), and IL-8 (▵) are represented.

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. GENERAL CONSIDERATIONS AND SUMMARY
  8. Acknowledgements
  9. References

This study addressed the question of whether the double gastric infection with H. pylori and non-H. pylori bacteria—in patients treated with acid-suppressive medication—may result in an enhanced inflammatory response, contributing to an increased risk of developing atrophic gastritis. We have shown that: (i) H. pylori-infected patients on long-term acid inhibition displayed three features: non-H. pylori bacterial growth, increased cytokine levels, and a high prevalence of corpus atrophic gastritis; (ii) the development of atrophic gastritis during acid-suppressive therapy was associated with a simultaneous gastric infection with H. pylori and non-H. pylori bacteria; (iii) the type of interaction between H. pylori and non-H. pylori bacteria was synergistic. Taken together, these findings indicate that the double gastric infection with H. pylori and non-H. pylori bacteria may play a major role in the development of atrophic gastritis.

Intragastric bacteria, atrophic gastritis, and circulating cytokines

Intragastric growth of non-H. pylori bacteria commonly occurs at an elevated intragastric pH, during the administration of either H2-RAs or proton pump inhibitors.15[16][17]–18 Our results confirm and further extend this observation, demonstrating that in H. pylori-positive subjects treated with acid inhibition, the overgrowth of non-H. pylori bacteria was accompanied by a high rate of corpus atrophic gastritis. However, this latter phenomenon was not observed in the H. pylori-negative patients treated in a similar fashion. As such, it can be argued that infection with H. pylori is a prerequisite for the development of atrophic gastritis in this setting.10, 11

An important finding of this study was that cytokine levels correlated with the anatomical distribution and the severity of chronic gastritis. Patients with antral predominant atrophic gastritis and those with diffuse non-atrophic gastritis displayed a slight, but significant, elevation in circulating cytokines compared to patients with normal mucosa. Noteworthy, the predominant involvement of the gastric body and fundus, representing a (twofold) larger anatomical area than the gastric antrum, was associated with the greatest elevation in circulating cytokines.

Our results also indicated that increasing age was associated with a higher risk of atrophic gastritis. Yet, its impact was marginal compared to that of the H. pylori infection, thereby supporting the contention that the development of atrophic gastritis is merely H. pylori-dependent rather than age-dependent.33, 34 Additionally, we found that infection with CagA-positive H. pylori strains increases the risk of atrophic gastritis in the antrum but not in the corpus, a finding in line with recently published data.35

Non-H. pylori bacteria: a cause or a consequence of atrophic gastritis?

Two possible explanations can be suggested for the demonstrated link between non-H. pylori bacteria and atrophic gastritis:

1 Non-H. pylori bacterial overgrowth could be an epiphenomenon. Previous studies, including ours, have shown that in subjects infected with H. pylori, acid-suppressive therapy leads to a greater elevation in gastric juice pH, predisposing to enhanced colonization of the stomach by non-H. pylori bacterial species.8, 36[37]–38 This observation has been confirmed in the present study. Furthermore, recent evidence linked genetic factors (such as polymorphisms in the genes encoding IL-1β) to an enhanced cytokine production in response to H. pylori infection, resulting in the development of atrophic gastritis, and subsequent overgrowth of non-H. pylori bacteria.39

2 Non-H. pylori bacterial flora could be a co-factor in the development of atrophic gastritis. The following arguments come in favour of this hypothesis: first, as shown in Figure 5(B), H. pylori-positive patients on long-term proton pump inhibitors were characterized by non-H. pylori bacterial growth and a high prevalence of corpus atrophic gastritis; this was in sharp contrast to the H. pylori-positive non-treated controls, who had similar demographic features, but a very low rate of corpus atrophic gastritis. Second, in H. pylori-positive patients treated with acid suppression, the presence of non-H. pylori bacteria preceded the development of corpus atrophic gastritis, a sequence supporting the hypothesis that bacterial infection was a causal factor rather than a consequence of atrophy. Third, the magnitude and ranking of cytokine increase were strongly concordant with the risk factors for atrophic gastritis emerging from the multiple regression analysis: patients harbouring the double infection had the highest cytokine levels, paralleled by the highest risk of atrophic gastritis.

Atrophic gastritis during acid-suppressive therapy: possible mechanism(s)

It remains to be explained how the gastric co-infection with non-H. pylori bacteria could cause aggravation of the H. pylori-induced chronic gastritis. A study by Mowat et al. has indicated that acid inhibition with omeprazole profoundly alters the intragastric milieu in subjects infected with H. pylori, leading to enhanced bacterial colonization, elevated intragastric nitrites, and depletion of intragastric vitamin C.38 They suggested that such changes may facilitate the development of gastric mucosal atrophy. Additionally, it is possible that commensal bacteria and their by-products act as a persistent antigenic stimulus, which enhances and perpetuates the inflammatory response initiated by the H. pylori infection. In this sense, it is worthwhile to note that the lipopolysaccharides produced by gram-negative aerobic bacteria, such as Escherichia coli, may evoke substantially (> 1000-fold) greater pro-inflammatory response than the H. pylori lipopolysaccharides.40, 41 The pro-inflammatory cytokines (mainly IL-1β) are potent inhibitors of the gastric parietal cell secretion.42[43]–44 As such, the increased production of IL-1β, driven by a double infection, may further promote the overgrowth of non-H. pylori bacteria, in a self-perpetuating process. Such explanations could underlie the observed synergistic interaction between H. pylori and non-H. pylori bacteria.

Figure 6 suggests a possible sequence of events leading to the accelerated development of atrophic gastritis during acid-suppressive therapy: In the presence of H. pylori infection, long-term acid-suppressive therapy leads to a corpus predominant pattern of gastritis or pangastritis.9[10]–11, 45 Chronic hypochlorhydria and subsequent super-infection with non-H. pylori species may occur. This double gastric infection with H. pylori and non-H. pylori bacteria may then elicit a greater inflammatory response, as reflected by the elevated cytokines, leading to an accelerated development of atrophic gastritis.

image

Figure 6.  Development of atrophic gastritis during acid-suppressive therapy: a model.

Download figure to PowerPoint

Methodological aspects

A few comments have to be made concerning methodological aspects of this study: First, the patients were consecutively recruited from a primary-care population, homogenous with regard to risk factors for the development of atrophic gastritis. Hence, the results presented and conclusions are representative for an everyday clinical situation. Second, this study was cross-sectional in design. Nevertheless, the role of the time factor in the development of atrophic gastritis could be estimated by patient histories and documentation concerning the length of antisecretory treatment, and by comparison with a sizeable control group. Further longitudinal studies are needed, however, to fully elucidate the time-course of gastric morphological changes caused by a double bacterial infection. Third, in the present study no special attention was paid to further identifying the non-H. pylori bacterial species in the gastric mucosa, because this has been done in earlier publications by our own group;24, 26 we have consistently found a predominance of the oropharyngeal flora (e.g. Streptococci spp., Neisseria spp. and Corynebacterium spp.), but faecal-type bacteria (e.g. Escherichia coli, Klebsiella spp.) were also isolated.

GENERAL CONSIDERATIONS AND SUMMARY

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. GENERAL CONSIDERATIONS AND SUMMARY
  8. Acknowledgements
  9. References

The present study may have clinical implications for the treatment of H. pylori-positive patients subjected to long-term acid-suppressive medication for acid-related disorders, especially for gastro-oesophageal reflux disease. An H. pylori‘test-and-treat’ strategy seems recommendable in this setting, in order to prevent the potential cumulative damage to the gastric mucosa resulting from a double bacterial infection. The non-H. pylori infection alone—unavoidable during gastric acid suppression—seems to do less harm to the gastric mucosa, and probably does not need major concern.

In summary, we have shown that in H. pylori-positive patients, the gastric co-infection with non-H. pylori bacteria (secondary to acid inhibition) seems to elicit a more vigorous inflammatory response. This may evolve to corpus atrophic gastritis, which is a precursor lesion for gastric cancer. The demonstrated relationship between gastric acid inhibition, H. pylori, and non-H. pylori bacteria suggests that closer attention should be paid to the interaction between these factors when investigating the mechanisms underlying the development of atrophic gastritis and gastric cancer.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. GENERAL CONSIDERATIONS AND SUMMARY
  8. Acknowledgements
  9. References

The authors are grateful to Ms R. Bretveld for assistance with data analysis.

This study was supported by an educational grant from AstraZeneca B.V., the Netherlands.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. PATIENTS AND METHODS
  5. RESULTS
  6. DISCUSSION
  7. GENERAL CONSIDERATIONS AND SUMMARY
  8. Acknowledgements
  9. References
  • 1
    Rauws EA, Langenberg W, Houthoff HJ, Zanen HC, Tytgat GNJ. Campylobacter pyloridis-associated gastritis: a prospective study of its prevalence and the effects of antibacterial and antiulcer treatment. Gastroenterology 1988; 88: 3340.
  • 2
    Kuipers EJ, Uyterlinde AM, Peña AS, et al. Long-term sequelae of Helicobacter pylori gastritis. Lancet 1995; 345: 15258.
  • 3
    Correa P, Haenszel W, Cuello C, Tannenbaum S, Archer M. A model for gastric cancer epidemiology. Lancet 1975; 2: 5860.
  • 4
    Sipponen G. Gastric cancer—a long-term consequence of Helicobacter pylori infection? Scand J Gastroenterol 1994; 201: 247.
  • 5
    Correa P. Human gastric carcinogenesis: a multistep and multifactorial process—first American Society Award lecture on cancer epidemiology and prevention. Cancer Res 1992; 52: 673540.
  • 6
    Graham DY. Helicobacter pylori infection in the pathogenesis of duodenal ulcer and gastric cancer: a model. Gastroenterology 1997; 113: 198391.
  • 7
    Lee A, Dixon MF, Danon SJ, et al. Local acid production and Helicobacter pylori: a unifying hypothesis of gastroduodenal disease. Eur J Gastroenterol Hepatol 1995; 7: 4615.
  • 8
    El-Omar EM, Oien K, El-Nujumi A, et al. Helicobacter pylori infection and chronic gastric acid hyposecretion. Gastroenterology 1997; 113: 1524.
  • 9
    Logan RP, Walker MM, Misiewicz JJ, Gummett PA, Karim QN, Baron JH. Changes in the intragastric distribution of Helicobacter pylori during treatment with omeprazole. Gut 1995; 36: 126.
  • 10
    Kuipers EJ, Lundell L, Klinkenberg-Knoll E, et al. Atrophic gastritis and Helicobacter pylori infection in patients with reflux esophagitis treated with omeprazole or fundoplication. N Engl J Med 1996; 334: 101822.
  • 11
    Eissele R, Brunner G, Simon B, Solcia E, Arnold R. Gastric mucosa during treatment with lansoprazole: Helicobacter pylori is a risk factor for argyrophil cell hyperplasia. Gastroenterology 1997; 112: 70717.
  • 12
    Lundell L, Miettinen P, Myrvold HE, et al. Lack of effect of acid suppression therapy on gastric atrophy. Gastroenterology 1999; 117: 31926.
  • 13
    Suzuki M, Miura S, Suematsu M, et al. Helicobacter pylori-associated ammonia production enhances neutrophil-dependent gastric mucosal injury. Am J Physiol 1992; 263: G71925.
  • 14
    Stockbrügger RW, Cotton PB, Menon GG, et al. Pernicious anaemia, intragastric bacterial overgrowth, and possible consequences. Scand J Gastroenterol 1984; 19: 35564.
  • 15
    Stockbrügger RW, Cotton PB, Eugenides N, Bartholomew BA, Hill MJ, Walters CL. Intragastric nitrites, nitrosamines and bacterial overgrowth during cimetidine treatment. Gut 1982; 23: 104854.
  • 16
    Sharma BK, Santana IA, Wood EC, et al. Intragastric bacterial activity and nitrosation before, during and after treatment with omeprazole. Br Med J 1984; 289: 7179.
  • 17
    Verdu E, Viani F, Armstrong D, et al. Effect of omeprazole on intragastric bacterial counts, nitrates, and N-nitroso compounds. Gut 1994; 35: 45560.
  • 18
    Thorens J, Froehlich F, Schwizer W, et al. Bacterial overgrowth during treatment with omeprazole compared to cimetidine: a prospective randomized double blind study. Gut 1996; 39: 549.
  • 19
    Crabtree JE, Shallcross TM, Heatley RV, Wyatt JI. Mucosal tumour necrosis factor and interleukin-6 in patients with Helicobacter pylori associated gastritis. Gut 1991; 32: 14734.
  • 20
    Ando T, Kusugami K, Ohsuga M, et al. Interleukin-8 activity correlates with histological severity in Helicobacter pylori-associated antral gastritis. Am J Gastroenterol 1996; 91: 11506.
  • 21
    Gionchetti P, Vaira D, Campieri M, et al. Enhanced mucosal interleukin-6 and interleukin-8 in Helicobacter pylori-positive dyspeptic patients. Am J Gastroenterol1994; 883–7.
  • 22
    Rath HC, Herfarth HH, Ikeda JS, et al. Normal luminal bacteria, especially Bacteroides species, mediate chronic colitis, gastritis, and arthritis in HLA-B27/human beta2 microglobulin transgenic rats. J Clin Invest 1996; 98: 94553.
  • 23
    Sartor RB, Rath HC, Sellon RK. Microbial factors in chronic intestinal inflammation. Curr Opin Gastroenterol 1996; 12: 32733.
  • 24
    Jonkers D, Stobberingh E, De Bruine A, Arends JW, Stockbrügger RW. Evaluation of immunohistochemistry for the detection of Helicobacter pylori in gastric mucosal biopsies. J Infection 1997; 35: 14954.
  • 25
    Jonkers D, Gisbertz I, De Bruine A, et al. Helicobacter pylori and non-Helicobacter pylori bacterial flora in gastric mucosal and tumour specimens of patients with primary gastric lymphoma. Eur J Clin Invest 1997; 27: 88592.
  • 26
    Jonkers D, Houben P, Hameeteman W, et al. Differential features of gastric cancer patients, either Helicobacter pylori positive or Helicobacter pylori negative. Ital J Gastroenterol Hepatol 1999; 31: 83641.
  • 27
    Price AB. The Sydney System: Histological division. J Gastroenterol Hepatol 1991; 6: 20922.
  • 28
    Dixon M, Genta R, Yardley J, Correa P. Classification and grading of gastritis. The updated Sydney system. Am J Surg Pathol 1996; 20: 116181.
  • 29
    Genta RM. Recognizing atrophy: another step toward a classification of gastritis. Am J Surg Pathol 1996; 20(Suppl. 1): S2330.
  • 30
    Andrew W, Wyatt JI, Dixon MF. Observer variation in the assessment of chronic gastritis according to the Sydney system. Histopathology 1994; 25: 31722.
  • 31
    Kleinbaum DG, Kupper LL, Muller KE. Applied Regression Analysis and Other Multivariate Methods. Boston: PBS-Kent publ. Co, 1988.
  • 32
    Rothman K. Modern Epidemiology. Boston: Little, Brown, 1986.
  • 33
    Correa P, Haenszel W, Cuello C, et al. Gastric precancerous process in a high-risk population: cohort follow-up. Cancer Res 1990; 50: 473740.
  • 34
    Valle J, Kekki M, Sipponen P, Ihamäki T, Siurala M. Long-term course and consequences of Helicobacter pylori gastritis. Scand J Gastroenterol 1996; 31: 54650.
  • 35
    Oksanen A, Sipponen P, Karttunen R, et al. Atrophic gastritis and Helicobacter pylori infection in outpatients referred for gastroscopy. Gut 2000; 46: 4603.
  • 36
    Verdu E, Armstrong D, Fraser R, et al. Effect of Helicobacter pylori on intragastric pH during treatment with omeprazole. Gut 1995; 36: 53943.
  • 37
    Sanduleanu S, Jonkers D, De Bruine A, Hameeteman W, Stockbrügger RW. Non-Helicobacter pylori bacterial flora during acid-suppressive therapy: differential findings in gastric juice and gastric mucosa. Aliment Pharmacol Ther 2001; 15(3): 37988.DOI: 10.1046/j.1365-2036.2001.00888.x
  • 38
    Mowat C, Williams C, Gillen D, et al. Omeprazole,Helicobacter pylori status and alterations in intragastric milieu facilitating bacterial N-nitrosation. Gastroenterology 2000; 119: 33947.
  • 39
    El-Omar EM, Carrington M, Chow WH, et al. Interleukin-1 polymorphisms associated with increased risk of gastric cancer. Nature 2000; 404: 398402.DOI: 10.1038/35006081
  • 40
    Birkholz S, Knipp U, Nietzki C, et al. Immunological activity of lipopolysaccharide of Helicobacter pylori on human peripheral blood cells in comparison to lipopolysaccharides of other intestinal bacteria. FEMS Immunol Med Microbiol 1993; 6: 31724.
  • 41
    Perez-Perez GI, Sheperd VI, Morrow JD, Blaser MJ. Activation of human THP-1 cells and rat bone marrow-derived macrophages by Helicobacter pylori lipopolysaccharide. Infect Immun 1995; 63: 11837.
  • 42
    Baume PE, Nicholls A, Baxter CH. Inhibition of gastric acid secretion by a purified bacterial lipopolysaccharide. Nature 1967; 215: 5960.
  • 43
    Kondo S, Shinomura Y, Kanayama S, et al. Interleukin-1 beta inhibits gastric histamine secretion and synthesis in the rat. Am J Physiol 1994; 267: G96671.
  • 44
    Yasunaga Y, Shinomura Y, Kanayama S, et al. Mucosal interleukin-1 beta production and acid secretion in enlarged fold gastritis. Aliment Pharmacol Ther 1997; 11: 8019.
  • 45
    McColl KE, El-Omar EM, Gillen D. Interactions between H. pylori infection, gastric acid secretion and anti-secretory therapy. Br Med Bull 1998; 54: 12138.