Changes in Helicobacter pylori-induced gastritis in the antrum and corpus during 12 months of treatment with omeprazole and lansoprazole in patients with gastro-oesophageal reflux disease
meD. M.Stolte Institut für Pathologie, Klinikum Bayreuth, Preuschwitzer Straße 101, 95445 Bayreuth, Germany.
Several studies have shown that treatment with omeprazole leads to aggravation of Helicobacter pylori gastritis in the corpus. Whether this also applies to lansoprazole, and whether, in comparison with omeprazole, there are differences in therapy-induced gastritis parameter changes remains unclear.
In 111 patients infected with H. pylori and with gastro-oesophageal reflux disease we investigated the gastritis parameters in antral and corpus mucosa before and after 2, 6 and 12 months of treatment with 15 or 30 mg lansoprazole or 20 mg omeprazole/day.
In all groups the different treatments had a similar effect: in both regions of the stomach, suppression or partial elimination of H. pylori was seen. However, improvement in the inflammation was observed only in the antrum, while in the corpus most gastritis parameters worsened significantly. There was no increase in intestinal metaplasia or atrophy.
In common with omeprazole, lansoprazole aggravates the gastritis parameters in the corpus but improves them in the antrum. Treatment with proton pump inhibitors does not result in any increase in the incidence of atrophy/intestinal metaplasia. However, as gastritis predominating in the corpus seems to be associated with an elevated carcinogenic risk, consideration should be given to prophylactic H. pylori eradication therapy before initiating proton pump inhibitor treatment.
The therapeutic suppression of gastric acid secretion can lead to augmentation of the Helicobacter pylori gastritis in the fundus and corpus ranging even to atrophic gastritis.
This situation has already been described previously following vagotomy.1, 2 In a number of studies it has been shown that long-term treatment with omeprazole may have the same effect.3–5 The effect of pantoprazole on H. pylori has, to date, not been described, while the first report on the effect of lansoprazole also revealed a worsening of the gastritis parameters.6 The question as to whether the various proton pump inhibitors have different effects on H. pylori gastritis has not so far been investigated. To rectify this, we conducted a randomized double-blind study of the long-term treatment of gastro-oesophageal reflux disease (GERD), to investigate the effect of 12 months of treatment with 15 or 30 mg lansoprazole or 20 mg omeprazole on H. pylori gastritis in the antrum and corpus.
PATIENTS AND METHODS
Within the framework of a multinational randomized double-blind study, the efficacy of 15 and 30 mg lansoprazole and 20 mg omeprazole administered once daily in the long-term treatment of gastro-oesophageal reflux disease was compared. The initial results of an evaluation of the endoscopic findings, the symptomatology and the serum gastrin levels have already been published in a preliminary abstract.7 Within the framework of the German contribution to the study, we carried out the histopathological investigations of the biopsy specimens obtained from the antrum and corpus of the stomach.
A total of 111 patients infected with H. pylori and with gastro-oesophageal reflux disease in Savary–Miller stages I–IV were admitted to the study. The reflux disease was first treated for 8 weeks with a daily dose of 30 mg lansoprazole. Thereafter, the patients were randomized in double-blind fashion to three treatment groups for long-term treatment to the 12th month. The first group (n = 43) comprised the patients receiving 20 mg omeprazole/day, the second group (n = 33) contained those receiving 15 mg lansoprazole/day and the third group (n = 35) patients treated with 30 mg lansoprazole/day. In each case, the medication was taken at breakfast time.
Of the 111 patients initially participating in the study, 89 remained after 2 months, 74 after 6 months and 60 after 12 months. Adequate biopsy material was available for all patients remaining in the study at the respective follow-up examinations.
The endoscopy/biopsy investigations were carried out prior to the start of treatment, on day 56 (i.e. 8 weeks after initial treatment with lansoprazole 30 mg/day), after 6 months and after 12 months of treatment.
The gender distribution of the patients before therapy was 77.6% males and 22.4% females in the first group, 62.2% males and 37.8% females in the second group, and 51.4% males and 48.6% females in the third group (χ2 test: P = 0.041).
For the histopathological investigation of the gastric mucosa, two biopsy specimens were obtained from the antrum (3 cm proximal to the pylorus, lesser curvature) and two from the corpus (middle of the corpus, posterior wall), fixed in 4% buffered formalin solution, dehydrated in an increasing series of alcohols and xylol, and embedded in paraffin. Four-micron sections were cut, deparaffinized and stained with haematoxylin and eosin, and Warthin–Starry stain. The classification and grading of the gastritis was carried out in accordance with the Sydney system by one and the same pathologist (M. Stolte) who was blinded both to the treatment itself and to when it was applied.8 In addition to the grading parameters identified in the Sydney system—grade of gastritis, activity of gastritis, grade of atrophy and grade of intestinal metaplasia—we also assessed the grade of replacement of normal foveolar epithelium by regenerative epithelium, and the grade of mucus depletion, as well as recording the presence of lymphoid follicles and lymphatic aggregates.9
For analysing the effects of the individual drugs and doses, the gastritis parameters evaluated at each follow-up examination were compared with the baseline values. The Wilcoxon signed rank test for paired samples was employed for semi-quantitative analysis of the parameters grade of H. pylori colonization, grade of chronic gastritis, activity of gastritis, replacement of the foveolar epithelium by regenerative mucosa and mucus depletion. Owing to the small number of cases investigated, the assessment of the parameters atrophy, focal intestinal metaplasia and lymphoid follicles/lymphatic aggregates was based merely on the ‘gradings’‘present’ or ‘not present’, and then tested for uniformity during the course of the study using the McNemar test.
Finally, the Pearson χ2 test was used to investigate differences among the three treatment groups before therapy, and at 2, 6 and 12 months follow-up.
The study protocol was examined by the Ethics Commission in Freiburg, who raised no objections to its implementation.
(1) Differences in the grade and distribution of H. pylori gastritis between the three treatment groups
For grade of H. pylori colonization, degree and activity of gastritis, replacement of foveolar epithelium by regenerative epithelium, mucus depletion and gastric atrophy there were no significant differences between the three treatment groups throughout the study (i.e. at the start of treatment, and after 2, 6 and 12 months of therapy with each proton pump inhibitor) (χ2 test: P > 0.2). However, significant differences among the three groups were observed at the 2-month and 12-month follow-up examinations due to a comparatively high incidence of intestinal metaplasia in the antrum after 2 months of treatment with omeprazole (P = 0.020), and intestinal metaplasia in the corpus after 12 months of treatment with lansoprazole 30 mg/day (P = 0.027). At the 6 months follow-up examination, there was a significant difference between the three groups in terms of the incidence of lymphoid follicles in the antrum (P = 0.010).
(2) Changes in H. pylori gastritis under long-term treatment with omeprazole 20 mg/day
In the antrum, treatment with omeprazole provoked a significant decrease in the grade of H. pylori colonization, the density of infiltration with lymphocytes and plasma cells (grade of chronic gastritis), and the density of infiltration with neutrophils (activity of gastritis) (Tables 1–3 12 ). Other gastritis parameters, such as replacement of the normal foveolar epithelium by regenerative epithelium and mucus depletion, also decreased significantly during the course of treatment (all P < 0.02 at each of the follow-up examination time-points).
Table 1. .
Grade of Helicobacter pylori
Table 2. .
Density of infiltration with lymphocytes and plasma cells (degree of chronic gastritis)
Table 3. .
Density of infiltration with neutrophils (activity of gastritis)
In the corpus, although omeprazole resulted in a significant decrease in the grade of H. pylori colonization, there was a significant increase in the grade of replacement of foveolar epithelium and mucus depletion (all P < 0.01). Similar observations were made for the degree and activity of gastritis in the corpus, but statistical significance was only reached for the degree of chronic gastritis after 2 months of therapy ( Table 2).
Treatment with omeprazole had no effect on the incidence of atrophy and intestinal metaplasia in either antrum or corpus (Tables 5 and 6 5 ). However, there was a slightly significant decrease in the incidence of lymphoid follicles/lymphatic aggregates in the antrum after 12 months of treatment ( Table 4).
Table 5. .
Frequency of gastric atrophy
Table 6. .
Frequency of intestinal metaplasia
Table 4. .
Frequency of lymphoid follicles/lymphatic aggregates
(3) Changes in H. pylori gastritis under long-term treatment with lansoprazole 15 mg/day
Under long-term treatment with lansoprazole 15 mg/day, the grade of H. pylori colonization in the antrum and corpus decreased significantly. However, while in the antrum the degree and activity of gastritis, replacement of foveolar epithelium by regenerative epithelium, mucus depletion and frequency of lymphoid follicles all improved accordingly, there was a gradual increase in the gastritis parameters in the corpus mucosa (Tables 2 and 3 23 ). This increase was statistically significant for the gastritis parameters density of infiltration with lymphocytes and plasma cells (after 2, 6 and 12 months of treatment), and density of infiltration with neutrophils (after 2 and 6 months).
Other parameters, such as atrophy and intestinal metaplasia, did not change under treatment.
(4) Changes in H. pylori gastritis under long-term treatment with lansoprazole 30 mg/day
In common with the two other treatment regimens, a daily dose of lansoprazole 30 mg brought about a reduction in the grade of H. pylori colonization ( Table 1), aggravated the degree and activity of gastritis in the corpus but improved them in the antrum ( Tables 2 and 3 ). The aggravation of the grade of replacement of foveolar epithelium by regenerative epithelium, and the grade of mucus depletion in the corpus was highly significant (all P < 0.01 at each follow-up examination). In the antrum, all these parameters improved but were not statistically significant (P > 0.1).
During the course of treatment with lansoprazole 30 mg/day focal gastritis parameters, such as lymphoid follicles, atrophy, and intestinal metaplasia, did not change significantly.
Overall, this study confirms the results of earlier studies showing a worsening of inflammation in the corpus mucosa under treatment with omeprazole.3–5, 10, 11 In addition, our evaluation reveals that similar changes also occur under treatment with lansoprazole, and that these changes do not differ under 12 months of treatment whether with 15 or 30 mg lansoprazole/day, or 20 mg omeprazole/day. In contrast to other publications, we found no significant increase in the frequency of intestinal metaplasia or atrophy of the corpus mucosa during the course of this 12-month period of treatment with proton pump inhibitors.3, 4 An interesting aspect of the present study was the finding that under treatment with proton pump inhibitors the H. pylori load decreased in both regions of the stomach. In contrast to the improvement of the inflammation in the antrum, however, there was a deterioration in the gastritis parameters in the corpus mucosa. These data are in accordance with those of a previous study of ours,12 and also with the results reported by Graham et al.13 Hence, it may be concluded that lansoprazole and omeprazole both have in vivo activity against H. pylori—an effect that is not observed for treatment with antacids and H2-receptor antagonists14—but this suppression and partial elimination of H. pylori correlates with an improvement in gastritis only in the antrum.
In contrast to other studies, we were unable to detect any signs of atrophic gastritis under long-term treatment with either omeprazole or lansoprazole.3–5 This may be due to the fact that we tested the long-term effect of treatment with proton pump inhibitors for a period of only 12 months while, for example in a long-term study lasting 60 months, ‘subatrophic’ gastritis was observed in about 50% of the cases of gastritis.3 This, however, may also be due to the fact that in earlier studies atrophic gastritis was very poorly classified and graded. First, in all the previous studies considering the development of atrophy under long-term treatment with proton pump inhibitors, no information was provided as to whether the atrophy was due to an auto-immune gastritis caused by parietal cell antibodies, which can also be suspected histologically in earlier pre-atrophic stages.15 Nor was the question addressed as to whether atrophy might have been a consequence of the antigastric effect of H. pylori antibodies in the serum. H. pylori antibodies may act as antigastric antibodies either against membranes of the mucus-producing surface epithelium or against membrane substances of the secretory canaliculi in the parietal cells, and thus trigger an auto-aggressive form of gastritis.16, 17 Eradication of H. pylori also heals this type of gastritis (our own as yet unpublished observations). Thus, it might well be that in the studies reporting the occurrence of atrophic gastritis, the latter is due not to the proton pump inhibitor but to such cases with an antigastric effect of H. pylori antibodies.
The observed aggravation of H. pylori gastritis in patients on long-term treatment of GERD with proton pump inhibitors has prompted discussions about possible consequences. The first consideration is the question as to whether, in these cases, an attempt should be made to conduct long-term treatment of GERD with histamine receptor antagonists, or even antacids instead of proton pump inhibitors. However, not only do the former two substances achieve much poorer results than the proton pump inhibitors—both in terms of symptoms and endoscopic and histologically diagnosed reflux disease—they are also much inferior to the latter in terms of the changes to the gastric mucosa with H. pylori gastritis. In common with the proton pump inhibitors, antacids and histamine receptor antagonists also lead to an aggravation of the H. pylori gastritis in the acid-producing gastric mucosa.14, 18 The suppression or elimination of H. pylori occurring under proton pump inhibitor treatment leads to an improvement in antral H. pylori gastritis, but neither histamine receptor antagonists nor antacids can improve the gastritis parameters in the antrum.14
In our opinion, therefore, the better alternative is to provide H. pylori eradication treatment prior to starting on long-term treatment with proton pump inhibitors. Although we know of no studies providing evidence for an elevated risk of cancer associated with long-term antacid treatment,19, 20 and the studies on the gastric cancer risk following vagotomy are also inconclusive,21, 22 we would, for the sake of safety, nevertheless recommend that H. pylori eradication treatment should be given, because corpus-predominant gastritis is associated with an elevated carcinogenic risk23 and is also a risk factor for argyrophil cell hyperplasia.24 This is supported by our studies on the topography and the grading of gastritis in patients with early gastric carcinoma and in relatives of gastric cancer patients.25, 26
This paper is dedicated to our teacher, Professor Dr Kurt Elster, pioneer of gastroenterological pathology in Europe, on the occasion of his 80th birthday.