Long-term acid suppression is believed to accelerate atrophic gastritis in Helicobacter pylori-positive patients. The influence of long-term therapy with lansoprazole has not been examined.
Long-term acid suppression is believed to accelerate atrophic gastritis in Helicobacter pylori-positive patients. The influence of long-term therapy with lansoprazole has not been examined.
To study the clinical and endoscopic efficacy and histological evolution of gastric mucosa during 5 years of maintenance treatment with lansoprazole, 30 mg.
Seventy-eight patients with endoscopically proven oesophagitis were followed for 5 years. Biopsies taken at the start of the study, during follow-up and after 5 years were available for 73 patients.
The total endoscopic relapse rate was 14.1%. At the start of the study, 34 patients were Helicobacter pylori negative and 39 were Helicobacter pylori positive (two atrophy, 25 antral gastritis, 12 pangastritis). At 5 years, no histological changes had occurred in Helicobacter pylori-negative patients. In the Helicobacter pylori-positive group, 20 patients developed pangastritis, six had normal histology and one had antral gastritis. Ten of the 12 patients with pangastritis had reduced antral activity. There was no increase in intestinal metaplasia, but there was a tendency towards regression of atrophy in the antrum and towards increased atrophy in the body of the stomach.
Maintenance treatment with lansoprazole, 30 mg, is efficacious. The development of glandular atrophy and intestinal metaplasia was not accelerated in Helicobacter pylori-positive patients. Helicobacter pylori eradication must be considered only because of the higher cancer risk associated with chronic Helicobacter pylori-related gastritis.
Although Helicobacter pylori is probably not implicated in the pathogenesis of reflux oesophagitis, usually one-third to one-half of adult patients with reflux oesophagitis or endoscopy-negative reflux disease are infected with H. pylori.1 This figure is comparable with the prevalence of the infection in the same age group.2 Two major patterns of H. pylori infection have been recognized: the low acid route, with pangastritis, atrophy and metaplasia, and the elevated acid route, with antral gastritis and duodenal ulcer.3 The low acid route is believed to give rise to an increased risk of cancer.3 Atrophy and intestinal metaplasia are long-term sequelae of H. pylori infection.4, 5 They are considered to be important precursors for gastric carcinoma.6–8 Type I (complete — small intestinal type) and type II (incomplete — sialomucin-containing goblet cells scattered among gastric-type neutral mucin-secreting cells) metaplasia are believed to carry the lowest risk. Type III metaplasia (incomplete form, with goblet cells secreting sialomucins and/or sulphomucins and showing large intestinal characteristics) has been linked more closely to the dysplasia–carcinoma sequence.9, 10 Long-term acid suppression in H. pylori-positive patients with chronic gastritis is believed to be an additional risk factor which might promote the development of mucosal atrophy and intestinal metaplasia, and hence eventually lead to dysplasia and malignancy.11 The influence of long-term treatment with lansoprazole, a proton pump inhibitor of the substituted benzimidazole class, which markedly inhibits basal and stimulated gastric acid secretion, on the gastric mucosa in H. pylori-positive and H. pylori-negative patients has not yet been studied extensively. In studies of 1 year of maintenance treatment, an increase in body gastritis and atrophy has been shown to occur in H. pylori-infected patients.12, 13 Similar results were obtained in a 5-year follow-up study of 42 patients with peptic disorders resistant to H2 blockers.14 In a Dutch multicentre study, the efficacy of lansoprazole, 30 mg daily, for the prevention of recurrent reflux oesophagitis was examined, and the clinical and histological evolution of the patients at 3 and 5 years under this maintenance therapy was followed. In the present paper, we report the data of the histological evolution for the patients included in this study. A better knowledge of this evolution is important as it may have an impact on the therapeutic strategy and may help to answer the question of whether eradication of H. pylori is indicated before starting long-term acid suppression.
One hundred and seventy-five patients, 18 years or older, and presenting with endoscopically proven ulcerative, non-bleeding, reflux oesophagitis, grade II or more according to the Savary–Miller classification, were initially included in the study (Figure 1). Ninety-three patients (53%) were H. pylori negative by histology. All patients gave oral and written informed consent prior to entering the study. One hundred and forty-eight patients completed the first year, and 97 patients continued with long-term therapy. Follow-up endoscopy was performed at 3 years after the start of the initial treatment and at 5 years. Twenty-one of the 97 patients stopped prematurely (seven had adverse events, six refused further cooperation, four were withdrawn by the investigators, three were lost to follow-up and one had a relapse despite treatment with lansoprazole, 60 mg). After 5 years of therapy, a final endoscopy was performed in 75 patients. Adequate gastric biopsies taken at the start of the study before administration of treatment and at 5 years were available for 73 patients (48 male and 25 female; mean age, 61 ± 6 years). The H. pylori status at the start of the study was assessed in these patients by histology (n=73) and serology (n=60).
At the start of the study, five samples were taken from the gastric mucosa. These included three biopsies of the corpus, obtained from the middle area of the posterior wall, and two antral biopsies, obtained at 3 cm from the pylorus at the posterior and anterior wall.15 At 5 years, a similar number of biopsy specimens were obtained, but from three regions including the corpus (n=2), the antrum (n=2) and the incisura angularis, according to the updated Sydney system.16 Orientation of the biopsies using a Millipore filter was advised at the start of the study in order to enable more precise analysis. The samples were fixed in Bouin’s solution and routinely processed. The start biopsies were initially assessed following the Sydney system, but reassessed together with the biopsies obtained at 5 years according to the updated Sydney system. All sections were coded and stained with haematoxylin and eosin for routine analysis, and with a Sirius red technique for the assessment of fibrosis. The latter staining involves deparaffinization of the section and treatment for 45 min with a Sirius red solution (1 L of distilled water saturated with picric acid in which 1 g of Sirius red, chrom 1A280 (Vel, Leuven, Belgium), is dissolved). After treatment with the Sirius solution, the specimen was rinsed for 3 min with tap water and counterstained with Harris’ haematoxylin. For each sample, 10 routinely stained sections obtained at different levels were examined in order to reduce intrabiopsy variability.17 Analysis was performed blindly by one pathologist, without a knowledge of the clinical data and the pre- or post-treatment status of the sample. The presence of H. pylori was initially assessed on haematoxylin and eosin sections and, in cases of doubt, immunohistochemical staining using antibodies directed against H. pylori (rabbit anti-Helicobacter pylori, immunoglobulin fraction, Dako, dilution, 1:5) was performed after microwave heating for antigen retrieval. A patient was considered to be positive for H. pylori when the biopsy was positive, the serology was positive, or both. The graded histological variables included H. pylori density, activity, chronic inflammation, intestinal metaplasia and glandular atrophy. They were evaluated according to the guidelines of the updated Sydney system.
For the assessment of atrophy, defined as the loss of glandular tissue, the thickness of the glandular compartment was compared with the length of the crypts, and a Sirius red stain was performed to evaluate the presence of fibrosis. The depth of the pits was used for comparison because the pits are short in the normal body mucosa, taking up about one-quarter of the total mucosal thickness, while in the antrum they are longer, taking up about one-half of the mucosal thickness. Therefore, they allow an assessment of the thinning of the mucosa. The Sirius red stain helps to distinguish between oedema and fibrosis by the dense dark red fibrillar staining of fibrous tissue. Atrophy was diagnosed when the pits were longer than the glandular compartment in the antrum and when a positive bright red basal staining was present either in the antrum or in the body in the glandular compartment with the Sirius stain.
Intestinal metaplasia and atrophy were assessed on the samples obtained before treatment and at week 4 in order to minimize sampling error. Intestinal metaplasia was diagnosed following staining with Alcian blue (pH 2.5)–periodic acid–Schiff, Alcian blue (pH 0.5) and high iron diamine, allowing a distinction to be made between complete and incomplete intestinal metaplasia. The different samples were compared and the worst appearance was used for evaluation. In addition, the effect of acid suppression on the configuration of the glands and of the lining cells in the body mucosa was assessed. The latter was defined by the presence of enlarged parietal cells and cystic dilatation of the fundic glands.18, 19
One hundred and seventy-five patients were included in the study. One hundred and forty-eight patients completed the first year, 132 of whom were in remission. During the first year, 16 patients developed reflux oesophagitis grade I and two patients developed reflux oesophagitis grade II. The total relapse rate during the first year was 11.5% (18 of 156; 19 patients refused a final endoscopy). Ninety-seven patients continued with long-term therapy after 1 year of treatment. During the subsequent 4 years, 78 patients had one or more intervening endoscopies. Twelve relapses of reflux oesophagitis were observed in 11 patients, seven with reflux oesophagitis grade I, four with reflux oesophagitis grade II and one with reflux oesophagitis grade III. At 5 years, a final endoscopy was performed in 75 of 76 patients, with only one patient showing a relapse of reflux oesophagitis (grade I). During the second year and up to and including the fifth year, the total relapse rate was 14.1% (11 of 78), excluding two patients with a relapse at 1 year who continued study treatment (protocol violation).
The population at the start of the study was divided into two groups according to histology and serology. Group I was composed of 34 H. pylori-negative patients (47%) (26 male and eight female; mean age at the end of the study, 59 years; range, 34–80 years). Group II comprised 39 H. pylori-positive patients (53%) (22 males and 17 females; mean age at the end of the study, 63.6 years; range, 40–84 years). In group I, the histology of the stomach was normal at the start of the study in 24 of 34 H. pylori-negative patients. Mild atrophy with fibrosis was present in the antrum in five patients and atrophy with intestinal metaplasia was present in one patient. Four patients had inactive, mildly chronic, non-atrophying, antral gastritis. At 5 years, 31 of the 34 H. pylori-negative patients had a normal histology of the stomach and three had developed pangastritis. The latter became H. pylori positive during follow-up at week 52 and months 36 and 60. Histological signs of acid suppression were present in eight patients.
In group II, antral gastritis was present at the start of the study in 25 patients and pangastritis in 12 patients. Disease activity, chronicity and atrophy are summarized in Tables 1–3. Intestinal metaplasia of the antrum (n=15) or body (n=1) was present at the start of the study in 11 of the 25 patients with antral gastritis and in five of the 12 patients with pangastritis. The metaplasia was extensive in three patients and mild to moderate in 13 patients (Table 4). At the end of the study, pangastritis was observed in 30 patients, two of whom had atrophy, 18 of whom had antral gastritis and 10 of whom had pangastritis at the start of the study. The gastritis in the body mucosa was mildly active in 20 of the 39 patients (51%) and moderately active in eight (21%). There was a two-step difference in seven patients when compared with the activity before treatment. Six of the 25 patients with antral gastritis had a normal histology, while one patient retained antral gastritis. The evolution of activity in the antrum is shown in Table 1. The disease activity was reduced in the antrum in all 10 patients with persisting pangastritis (already present at start). The evolution of corpus gastritis is shown in Table 2. The evolution of atrophy and intestinal metaplasia for all H. pylori-positive patients is shown in Tables 3–5. In the antrum, a regression of atrophy from moderate to mild or no atrophy was observed. Intestinal metaplasia was still present in the final set of biopsies of the antrum in three of the 11 patients with antral gastritis and intestinal metaplasia at the start of the study and in one of the five patients with pangastritis and initial intestinal metaplasia. In addition, it appeared in the antrum in one patient with pangastritis and in the corpus in two patients (Table 4). The metaplasia was extensive in all of these patients and of the complete type in four patients (57%).
The biopsies of the antrum and the incisura angularis showed an identical picture in 68 of 73 cases. In two patients, the biopsy of the incisura angularis was normal and the antral biopsies were abnormal, in one patient the incisura angularis biopsy was abnormal and the antral biopsies were normal and in two patients the biopsies of the incisura angularis and antrum were all abnormal but there were differences in the presence of chronic inflammatory cells.
Histological signs of acid suppression, characterized by enlarged parietal cells and cystic dilatation of glands in the body mucosa, were present in six of the 39 patients. No dysplasia or neoplasia was observed.
The total relapse rate after 1 year of maintenance treatment with lansoprazole was 11.5% (18 of 156 patients). Most of these relapses were of reflux oesophagitis grade I. Other maintenance studies with lansoprazole, 30 mg, have shown comparable results with relapse rates of 10–15%.20–22 When all withdrawals were counted as failures, the total relapse rate was 21.1% (37 of 175 patients). A similar relapse rate (20%; n=75) was found in another intention-to-treat analysis after 1 year on lansoprazole, 30 mg daily.23 Maintenance studies with other proton pump inhibitors, all probably based on intention-to-treat analysis, showed slightly higher relapse rates. One year of treatment on omeprazole, 10 mg daily, omeprazole, 20 mg daily, and pantoprazole, 20 mg daily, resulted in the following relapse rates: 32–38%, 28% and 25–29%, respectively.23–27
Maintenance therapy on lansoprazole was not continued in all patients. During the subsequent 4 years, 14.1% (11 of 78 patients) had a relapse of reflux oesophagitis, and for the intention-to-treat population the endoscopic relapse rate was 33.0% (32 of 97 patients). Most of these cases developed reflux oesophagitis grade I. Only one patient had reflux oesophagitis grade III and none of the patients had a relapse of reflux oesophagitis grade IV. Most of the relapses occurred in the first 3 years of maintenance treatment. There are no comparable long-term studies with lansoprazole, following patients for 5 years on lansoprazole maintenance therapy, in the literature.
In H. pylori-negative patients, treatment with lansoprazole, 30 mg, had no major influence on the histology. In H. pylori-positive patients, pangastritis developed in 20 patients having either antral gastritis or antral atrophy at the start of the study. Overall, disease activity increased in the body of the stomach in some patients and decreased in the antrum in most patients. These data confirm earlier findings.13, 28 A mild increase of atrophy was noted for the body mucosa, but there was no increase in intestinal metaplasia (Table 2). This result seems different from the findings reported earlier in patients treated with acid suppressive therapy. In a series of 59 patients infected with H. pylori and treated with omeprazole, the development of atrophic gastritis was observed in 18 patients.11 A significant association between H. pylori infection and the development of atrophic gastritis and intestinal metaplasia has also been noted in individuals not undergoing systematic treatment.5 A reliable diagnosis of atrophy in the gastric mucosa is, however, difficult, and important interobserver variability has been reported in the literature.3, 29, 30 Special stains for the detection of collagens, such as the Sirius red stain, and comparison with the length of the crypts can help to distinguish between a genuine loss of glands accompanied by fibrosis and oedema. A diagnosis of atrophy, characterized by the replacement of normal glands with metaplastic glands, is largely dependent on the number of biopsy samples because intestinal metaplasia is a multifocal process. The detection of both atrophy and intestinal metaplasia also depends on the number of sections examined. Our data, with biopsies obtained at the start of the study following the guidelines of the Sydney system and at the end of the study according to the updated Sydney system, show no increase in intestinal metaplasia at 5 years. The addition of biopsies from the incisura angularis did not increase the diagnostic yield, and in the large majority of patients these biopsies were identical to those obtained in the antrum. Three more series of biopsies obtained during follow-up (week 4, week 52 and month 36) showed similar data. There was also no increase in the number of patients with incomplete intestinal metaplasia, the subtype believed to carry a higher risk of gastric cancer.31 Intestinal metaplasia was present in 23% of the total number of patients and in 41% of H. pylori-positive patients at the start of the study. This is comparable with data from other studies of the same geographical area, where intestinal metaplasia was found in the antrum in 25% of all patients and in 34% of H. pylori-positive patients.4 Our study therefore does not show the progressive increase in atrophy and intestinal metaplasia previously described, but confirms the extension of gastritis towards the corpus.5
Part of the difference can be explained by the different age of the study group. In the study by Kuipers et al., the mean age at the first visit was 49.0 years (range, 15–80 years).5 In a Finnish adult population with a mean age of 46 years, 32% of patients had normal mucosa, 56% had non-atrophic gastritis and 12% had atrophic gastritis.32 Comparable figures were obtained in a French population.33 In our population, the mean age at the first visit was 58.6 years for the H. pylori-positive patients, which is more comparable with the mean age of the patients studied by Kuipers et al. at the end of their study.5 In the present series, there was a tendency towards regression of atrophy in the antrum and towards an increase in atrophy in the corpus. Our data for the corpus are comparable with those reported in another series of patients with a similar age distribution, showing atrophy of the corpus in 12% of patients at baseline (20% in our series) and in 39% at follow-up (36% in our series).34 The increase of atrophy in our patients indicates an annual evolution of 3.4%, a figure which lies in between the natural annual progression towards atrophy in chronic gastritis (which varies between 1 and 2%) and the accelerated evolution attributed to acid suppression (which varies between 4 and 6%).32–34 The number of patients in our series is, however, probably too small to draw firm conclusions. The reasons for the difference between the antrum and the corpus in our series are unclear, but may be related to the dosage of the drug, the severity of the activity of gastritis at the start of the study or sampling. Sampling seems an unlikely reason because the series of biopsies obtained at 36 months provided similar results.35
Thus far, limited and contradictory literature data are available on the potential reversibility of atrophy. In a follow-up study of duodenal ulcer patients with either successful or failed H. pylori eradication, glandular atrophy and intestinal metaplasia remained unchanged.36 Similar results were obtained in a prospective study with 122 patients, following the evolution after eradication in relation to cagA-positive H. pylori strains, and in a study comparing the effect of eradication on chronic gastritis during omeprazole therapy.37, 38 Mild regression of atrophy was found in 7% of infected subjects in the series of Kuipers et al.5 We speculate that the decrease in antral atrophy could be explained by the trophic action of gastrin and the decrease in disease activity observed on treatment with lansoprazole.
In conclusion, our data do not confirm an accelerated evolution of gastric mucosal atrophy and intestinal metaplasia with maintenance treatment of lansoprazole, 30 mg, in H. pylori-positive patients with reflux oesophagitis. However, they confirm the extension of antral gastritis towards pangastritis, although there is no increase in disease activity. This might partially explain why atrophy is not progressive because glandular atrophy is likely to be the result of inflammation. On the other hand, persisting H. pylori-positive chronic gastritis might carry an increased cancer risk because of persistent epithelial cell damage. Such damage and reparative proliferation are substantially reduced following H. pylori eradication.39 Therefore, eradication of H. pylori should still be considered before starting maintenance acid suppressive therapy in H. pylori-positive patients with reflux oesophagitis.
The authors would like to thank the following investigators who participated in this study: A. J. van Bork, Boxmeer; J. G. S. Breed, Weert; C. P. M. Dekkers, Breda; R. J. X. M. van Etten, Oss; W. H. Ezendam, Gorinchem; A. A. M. Geraedts, Amsterdam; G. H. de Groot, Beverwijk; G. den Hartog, Arnhem; J. R. Juttmann, Tilburg; K. Te Velde, Den Haag; J. N. Verhagen, Utrecht; B. D. Westerveld, Zwolle; A. M. H. Wetzels, Stadskanaal; A. A. C. Wolff, Geldrop.