|
| |||||||||||||||||||||||||||||||||||||||||
Background Helicobacter pylori infection plays an important part in the development of atrophic gastritis and intestinal metaplasia, conditions that predispose patients to gastric cancer. Profound suppression of gastric acid is associated with increased severity of gastritis caused by H. pylori, but it is not known whether acid suppression increases the risk of atrophic gastritis.
Methods We studied patients from two separate cohorts who were being treated for reflux esophagitis: 72 patients treated with fundoplication in Sweden and 105 treated with omeprazole (20 to 40 mg once daily) in the Netherlands. In both cohorts, the patients were followed for an average of five years (range, three to eight). After fundoplication, the patients did not receive acid-suppressive therapy. The presence of H. pylori was assessed at the first visit by histologic evaluation in the fundoplication group and by histologic and serologic evaluation in the omeprazole group. The patients were not treated for H. pylori infection. Before treatment and during follow-up, the patients underwent repeated gastroscopy, with biopsy sampling for histologic evaluation.
Results Among the patients treated with fundoplication, atrophic gastritis did not develop in any of the 31 who were infected with H. pylori at base line or the 41 who were not infected; 1 patient infected with H. pylori had atrophic gastritis before treatment that persisted after treatment. Among the patients treated with omeprazole, none of whom had atrophic gastritis at base line, atrophic gastritis developed in 18 of the 59 infected with H. pylori (P<0.001) and 2 of the 46 who were not infected (P = 0.62).
Conclusions Patients with reflux esophagitis and H. pylori infection who are treated with omeprazole are at increased risk of atrophic gastritis.
Proton-pump inhibitors such as omeprazole influence the normal acidic habitat of H. pylori and have direct activity in vitro against the bacteria. Nevertheless, monotherapy with these drugs does not lead to the eradication of H. pylori.4 On the contrary, the activity of corpus gastritis increases in H. pyloriinfected patients treated with omeprazole.5,6,7 A similar effect has been observed during H2-blocker therapy8 and after vagotomy,9,10 as well as during treatment of mice with omeprazole.11 One hypothesis is that the increased inflammation increases the speed of atrophic changes of the gastric mucosa during acid-suppressive maintenance therapy. We conducted a prospective histologic follow-up study to compare the effects of omeprazole maintenance treatment and fundoplication in patients with gastroesophageal reflux disease who were infected or not infected with H. pylori.
Methods
Patients
We studied two cohorts of patients in whom grade I to IV reflux esophagitis, defined according to the classification of Savary and Miller,12 had been diagnosed by endoscopy.
The first cohort consisted of 109 Dutch patients who were enrolled between 1986 and 1991 in an open, compassionate-use study to evaluate the long-term efficacy and safety of omeprazole (Losec) in patients with severe reflux disease. Patients were included if they had endoscopically verified persistent esophagitis despite at least three months of treatment with an H2 blocker in a dose equivalent to at least 600 mg of ranitidine daily. All eligible patients participated. During follow-up, three patients died of unrelated diseases and one emigrated. No other patients withdrew. The final analysis involved all biopsy specimens collected up to July 1994 from 105 patients (56 men and 49 women; mean age at the initial visit, 62 years [range, 16 to 83]). The patients were treated with 40 mg of omeprazole a day for 4 to 16 weeks until esophageal healing was endoscopically verified. Thereafter, maintenance treatment with 20 mg of omeprazole a day was started. The maintenance dose was raised to 40 mg in subjects with endoscopic evidence of a relapse. The mean follow-up period was five years (range, three to eight). During this treatment period, all patients underwent gastroscopy at one year and then every two years. During the pre-entry and follow-up endoscopies, four to six biopsy specimens were obtained with standard biopsy forceps from the area 10 cm below the cardia. The endoscopist was not made aware of previous histologic results. The study was performed in four referral centers in the Netherlands. The protocol was approved by the ethics committees of the participating hospitals and by the scientific review committee of the Free University Hospital, Amsterdam. The data on the efficacy of the omeprazole treatment obtained before April 1991 in 91 of these patients have been published previously.13
The second cohort consisted of 137 Swedish patients who were treated between 1987 and 1992 with either a NissenRosetti or a Toupet fundoplication procedure for recurrent, endoscopically verified reflux esophagitis. They participated in a randomized study of the clinical efficacy and side effects of these two types of fundoplication. The surgical results for this cohort are described elsewhere.14 All patients underwent gastroscopy with biopsy sampling preoperatively. Postoperatively, this procedure was repeated after one year and then every two years, according to the routine described for the Netherlands study. During follow-up, two patients died of unrelated diseases and three were lost to follow-up. Seven had a relapse of reflux disease requiring acid-suppressive medication. The series of biopsy specimens was incomplete for 53 patients. Thus, the histologic analysis involved 72 patients (28 men and 44 women; mean age at the initial visit, 53 years [range, 25 to 73]) all treated in the same hospital. The mean postoperative follow-up time was five years (range, three to eight). None of these 72 patients were treated with acid suppressors after surgery. The study was performed in one academic hospital in Sweden and was approved by the local ethics committee. Before inclusion in the study, all patients from both cohorts gave informed oral consent that was witnessed by two physicians.
Histologic Methods
Five-micrometer sections stained with hematoxylin and eosin were used for standard histologic evaluation. H. pylori was detected by a modified silver-impregnation method.15 SevierMunger silver staining,16 and sometimes Grimelius's staining as well,17 was used to detect argyrophil neuroendocrine cells (mainly enterochromaffin-like [endocrine-like] cells). Different staining techniques were used with adjacent sections. Slides from various visits were scrambled and stained together. All specimens from both cohorts were graded by one pathologist according to the Sydney classification.18 The pathologist was not aware of the clinical and endoscopic data. The following items were evaluated separately: H. pylori colonization, presence of an acute inflammatory component (especially the amount of neutrophilic infiltration), presence of a chronic inflammatory component (especially the amount of lymphoplasmacytic infiltration), mucosal atrophy, intestinal metaplasia, and argyrophil-cell changes. All these items were scored from 0 to 3. For the first five items, the scale ranged from absent (0) to slight (1), moderate (2), and severe (3). For the argyrophil-cell pattern, the classification of Solcia et al. was used.19 The scores ranged from normal (0) to diffuse (1), linear (2), and micronodular (3) hyperplasia. The results were scored blindly with the use of patient codes.
The presence of H. pylori infection at the first visit was determined by histologic evaluation in the cohort undergoing fundoplication and by histologic and serologic evaluation in the cohort receiving omeprazole. In the case of discrepant results, the histologic evaluation was considered definitive, unless no bacteria were observed in a sample of atrophic mucosa with a positive serologic result (a finding regarded as a sign of past infection).20 AntiH. pylori IgG serum antibodies were detected in the patients in the omeprazole cohort with an enzyme-linked immunosorbent assay.21
Statistical Analysis
Statistical analysis was performed with Fisher's exact test, McNemar's test, Wilcoxon's rank-sum test, and one-way analysis of variance by the unpaired Student's t-test.22 Two-tailed P values smaller than 0.05 were considered to indicate statistical significance.
Results
Omeprazole Cohort
At the initial visit, 46 of the 105 patients in the omep-razole cohort were negative for H. pylori and 59 were positive. The two groups did not differ in age (mean age in both groups, 62 years; P = 0.96 by Student's t-test). No new infections were observed, nor did any infected patient become negative for H. pylori. Before entry, all 46 H. pylorinegative patients had either no histologic signs of gastritis (39 patients) or only mild histologic signs (7 patients) (Table 1), and none had signs of atrophic gastritis (Table 2). After an average follow-up of five years, moderate-to-severe corpus gastritis had developed in only 2 (4 percent) of these 46 patients (Table 1). H. pylori infection could not be detected in them. In these two patients, histologic signs of mild-to-moderate atrophic gastritis developed concomitantly (Table 2). Thus, the prevalence of atrophic gastritis in this group increased from 0 to 4 percent (P = 0.62 by McNemar's test), corresponding to an annual increase of 0.8 percent (95 percent confidence interval, 0.2 to 3 percent). The prevalence of argyrophil-cell hyperplasia remained stable at 8.7 percent (Table 3), and there was no intestinal metaplasia.
|
|
|
|
The development of atrophic gastritis and that of argyrophil-cell hyperplasia were strongly associated with H. pylori infection (P<0.001 by Fisher's exact test; odds ratios, 9.9 [95 percent confidence interval, 2.2 to 45] and 35 [2.3 to 606], respectively).
Fundoplication Cohort
At the initial visit, 41 of the 72 patients in the fundoplication cohort were negative for H. pylori and 31 were positive. These two groups did not differ in age. Preoperatively, none of the 41 H. pylorinegative patients had histologic signs of active gastritis (Table 4) or atrophic gastritis (Table 5). After an average follow-up of five years, none of these patients had inflammation or atrophy of the gastric mucosa (Table 4 and Table 5), and none had become infected with H. pylori.
|
|
When the two cohorts were compared, no significant increase in the prevalence of atrophic gastritis was found with either treatment in the absence of H. pylori infection. However, among the infected patients, atrophic gastritis did not develop after fundoplication but did develop in 18 patients (31 percent) during omeprazole therapy (P<0.001 by Fisher's exact test; odds ratio, 28; 95 percent confidence interval, 1.6 to 484) (Figure 1).
Discussion
In recent years, acid-suppressive therapy has become of pivotal importance in the treatment of gastroesophageal reflux disease. Because of the high relapse rate, patients often receive maintenance therapy. Therefore, the long-term consequences of the treatment are relevant. In the past, attention was specifically given to hypergastrinemia and argyrophil-cell hyperplasia as side effects of treatment with proton-pump inhibitors. However, although elevated fasting serum gastrin levels (100 to 400 ng per liter) and focal linear hyperplasia of argyrophil cells occur regularly during treatment,13,23 complications, in particular carcinoid formation, have not to our knowledge been observed. Argyrophil-cell hyperplasia occurs predominantly in patients who have atrophic gastritis. Some investigators have regarded this as a normal consequence of aging and have considered the rate of progression to atrophic gastritis during omeprazole therapy to be no different from the rate in controls.23,24 However, the control study referred to by these investigators was based on a single cohort of patients with gastric ulcers.25,26 Such patients have reduced acid output, and the majority are infected with H. pylori. Therefore, these patients are not representative.
We investigated the role of H. pylori in the development of atrophic gastritis during acid-suppressive maintenance therapy. This agent is a frequent cause of chronic active gastritis and has an important role in the development of atrophic gastritis and intestinal metaplasia, conditions that substantially increase the risk of gastric cancer.
Our results demonstrate that gastritis is increased during long-term omeprazole therapy and suggest that this increase may have important clinical consequences. In the fundoplication cohort, the prevalence of atrophic gastritis was not significantly greater among patients with H. pylori infection than among those without such infection. The upper limit of the 95 percent confidence interval for the annual increase in the prevalence of atrophic gastritis was approximately 2 percent. In the omeprazole cohort, atrophic gastritis developed among noninfected patients at a similar rate each year (0.8 percent; 95 percent confidence interval, 0.2 to 3 percent). However, among H. pyloriinfected patients, the annual rate of development of atrophic gastritis was 6.1 percent (95 percent confidence interval, 3.8 to 8.8 percent); signs of atrophy developed in approximately one third of the infected patients after an average of five years of treatment.
The comparison between the cohorts suggests that H. pyloriinfected patients are at considerably higher risk than uninfected patients for the development of atrophic gastritis during profound acid suppression. We are aware that aspects of this comparison are problematic and that there may be various confounding variables. This was not a randomized study. The patients in the two cohorts were treated in different hospitals. Nevertheless, the biopsy specimens were obtained according to the same routine. All the specimens were processed at the same laboratory and analyzed by one pathologist. The mean age in the two cohorts differed. The fundoplication group was followed for an average of five years, during which the mean age of the cohort increased from 53 to 58 years. The omeprazole group was followed for five years, during which the mean age of the cohort increased from 62 to 67 years. Atrophic gastritis results from long-lasting gastritis, and the risk of atrophy may increase exponentially at older ages. However, we believe it unlikely that this moderate difference in age completely accounts for the observed discrepancy in the progression of atrophy.
Our results are in accordance with those of other studies. Four cohort studies focused on people without any specific disease or treatment.20,27,28,29 Three did not give separate data for H. pyloripositive and H. pylorinegative patients.27,28,29 They followed 137 people for 25 years,27 1422 people for 5 years,28 and 142 people for 6 years29; the annual increase in the prevalence of atrophic gastritis ranged from 1.2 to 3.3 percent. A fourth study, in which 107 people were followed for 11.5 years, specified results according to H. pylori status.20 Among 49 people without infection, the prevalence of atrophic gastritis increased by 0.3 percent annually (95 percent confidence interval, 0.04 to 1.2 percent). Among 58 people with H. pylori infection, the annual increase was 1.8 percent (95 percent confidence interval, 1.0 to 2.9 percent). Four other studies provided data on people with conditions affecting acid secretion, reporting on 39 patients with gastric ulcer disease followed for seven years,25 23 patients with peptic ulcer disease followed for three years after vagotomy,30 74 patients with peptic ulcer disease followed for five years during omeprazole maintenance treatment,23 and 195 patients with gastroesophageal reflux disease followed for one to two years during maintenance treatment with omeprazole.31 None provided separate results for subjects with and those without H. pylori infection. The annual increase in the prevalence of atrophic gastritis ranged from 3.8 to 8.7 percent.
In conclusion, long-term acid-suppressive therapy with omeprazole is not associated with the eradication of H. pylori in infected patients. However, such treatment is associated with a persistent increase in inflammation of the corpus and with the development of atrophic gastritis and argyrophil-cell hyperplasia in patients who are positive for H. pylori, but not in those who are negative. Future studies of the long-term effects of acid suppression should focus primarily on H. pylori rather than on gastrin and argyrophil cells. We suggest that patients with reflux esophagitis who require profound acid-suppressive maintenance therapy should be studied to determine whether they are infected with H. pylori. If they are infected, therapy to eradicate H. pylori should be considered. It remains to be seen, however, whether such a strategy can prevent atrophy and argyrophil-cell hyperplasia in these patients.
Supported by Astra Hässle, Sweden.
We are indebted to Agneta Dalväg, Astra Hässle, Mölndal, Sweden, and José van den Berg, Astra Netherlands, for expert assistance with data management and evaluation.
Source Information
From the Departments of Gastroenterology of the Free University Hospital, Amsterdam (E.J.K., E.C.K.-K., H.P.M.F., S.G.M.M.), University Hospital, Leiden (C.B.H.W.L.), University Hospital, Nijmegen (J.B.M.J.J.), Slotervaart Hospital, Amsterdam (P.S.), and Sophia Hospital, Zwolle (G.F.N.) all in the Netherlands; the Department of Surgery, Sahlgren's Hospital, University of Gothenburg, Gothenburg, Sweden (L.L., B.L., J.D.); and Astra, Södertälje, Sweden (N.H.). Presented in part at the annual meeting of the American Gastroenterology Association, San Diego, Calif., May 1417, 1995.
Address reprint requests to Dr. Kuipers at the Department of Gastroenterology, Free University Hospital, P.O. Box 7057, 1007 MB Amsterdam, the Netherlands.
References
| |||||||||||||||||||||||||||||||||||||||||
Related Letters:
Atrophic Gastritis and
Repucci A. H., Kuipers E. J., Klinkenberg-Knol E., Meuwissen S. G.M., Lundell L.
Extract |
Full Text
N Engl J Med 1996;
335:750-751, Sep 5, 1996.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | TERMS OF USE | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2009 Massachusetts Medical Society. All rights reserved. |