Although several recent studies have shown that the eradication of Helicobacter pylori provokes reflux oesophagitis, the results are conflicting.
Although several recent studies have shown that the eradication of Helicobacter pylori provokes reflux oesophagitis, the results are conflicting.
To investigate the prevalence of reflux oesophagitis in patients after eradication of H. pylori and consider its association with hiatal hernia and corpus gastritis.
A total of 286 patients who underwent H. pylori eradication therapy and 286 age- and disease-matched H. pylori-positive controls who did not undergo eradication therapy were followed prospectively. All patients and controls underwent endoscopy at 1-year intervals or when upper gastrointestinal symptoms recurred. The presence of hiatal hernia and histology of the gastric corpus were evaluated at the time of initial endoscopy.
The estimated prevalence of reflux oesophagitis within 3 years was 18% after eradication therapy and 0.3% without therapy. Patients who developed reflux oesophagitis after therapy had a greater prevalence of hiatal hernia (P=0.0008) and more severe corpus gastritis (P=0.0005) before therapy. Cumulative prevalence of reflux oesophagitis was 26% in patients with hiatal hernia, 7.7% in those without hiatal hernia, 33% in those with corpus atrophic gastritis and 13% in those without corpus atrophic gastritis. When patients had both hiatal hernia and corpus gastritis, the prevalence of reflux oesophagitis was 37%. The newly developed reflux oesophagitis was classified as mild in 35 out of 36 (97%) patients who developed reflux oesophagitis after eradication therapy.
Eradication of H. pylori increased the prevalence of reflux oesophagitis in our patient group. The presence of hiatal hernia and corpus gastritis are closely related to the development of reflux oesophagitis after H. pylori eradication therapy.
Recent studies have clarified the association of Helicobacter pylori with chronic gastritis, peptic ulcer disease, and gastric carcinoma.1–8 One current area of interest is the relationship between H. pylori infection and reflux oesophagitis. Recent studies have found a negative association between H. pylori infection and reflux oesophagitis.9–11 In addition, development of reflux oesophagitis after H. pylori eradication therapy has been reported.12 Therefore, it is possible that disappearance of H. pylori may increase the risk for the development of reflux oesophagitis. It is important to clarify whether H. pylori eradication therapy contributes to reflux oesophagitis and then to determine whether H. pylori eradication therapy is indicated in certain individuals.
In this prospective study we evaluated the prevalence of reflux oesophagitis after eradication therapy in patients with gastritis or peptic ulcer disease as well as the factors affecting development of reflux oesophagitis in this population.
We studied 286 consecutive patients with gastritis or peptic ulcer disease undergoing eradication for H. pylori during a 3-year period (January 1994 to December 1996). In addition, 286 age-, sex- and disease-matched control subjects were also randomly selected from patients attending our hospital during the same period. Patients who had concomitant reflux oesophagitis, previous gastrectomy, or were taking H2-blockers or proton pump inhibitors during the 4 weeks before initial index endoscopy were not included in the study. H. pylori infection was confirmed in all patients and controls by rapid urease test, histological examination of antral and fundic biopsy specimens, and measurement of serum anti-H. pylori IgG antibodies. Subjects were defined as H. pylori-positive if three tests were positive or as H. pylori-negative if all three were negative. To avoid false positives for H. pylori infection, five individuals with differing results for the rapid urease test, histology and serology before eradication were not included in the study. Endoscopic findings, sex ratios and mean ages are shown in Table 1. No patient received antisecretory drugs during the study period. If patients medically required antisecretory drugs, they left the study and observation was stopped. Informed consent was obtained from each patient.
At initial endoscopy, after inspection of the oesophagus, stomach and proximal duodenum, and with the stomach inflated with air, the size of hiatal hernia (if present) was measured using the endoscope. The length of hiatal hernia was recorded as the distance from the oesophagogastric junction to the diaphragm. Without information of H. pylori status, hiatal hernia was diagnosed when this length was > 2 cm. At 4 weeks after completion of treatment, at 1-year intervals and when symptoms occurred, patients underwent endoscopy. At examination, endoscopic reflux oesophagitis was noted. Endoscopy was performed and photos for evaluation of reflux oesophagitis and hiatal hernia were taken. Reflux oesophagitis was graded according to the Los Angeles Classification System by one of three of the authors (HH, KH and MM).13 The end-point of this study was when reflux oesophagitis occurred.
At each endoscopic examination, four biopsy specimens, two from the antrum and two from the corpus, were taken. The specimens were fixed in 10% buffered formalin, embedded in paraffin, orientated edgewise, cut in sequential 4-μm thick sections and stained with haematoxylin and eosin for histological assessment and Giemsa stain for H. pylori. The biopsy specimens were scored without knowledge of the diagnosis, H. pylori status, or therapy regimen. Two parameters, inflammation and atrophy, were evaluated using the Updated Sydney System and graded on a 4-point scale of 0 (none) to 3 (severe).14 We regarded a subject to have corpus atrophic gastritis when the atrophic score of the corpus was 2 or 3, and also to have inflammation of the corpus when the inflammation score of the corpus was 2 or 3.
IgG antibody response against H. pylori was measured using a commercial enzyme-linked immunosorbent assay (ELISA) kit (Pyloristat kit; Whittaker Bioproducts, Inc, Walkersville, Maryland) as previously reported.7, 8 We regarded a subject as serologically negative when the titre was less than 1.0.
After initial endoscopic examination, patients were treated with omeprazole (20 mg b.d.), amoxycillin (750 mg b.d.), and clarithromycin (200 mg t.d.s.) for 1 week. Patients with gastric ulcers were then treated with omeprazole (20 mg o.d.) for 7 weeks, those with duodenal ulcers were treated with omeprazole (20 mg o.d.) for 5 weeks and those with gastritis were followed without medication. Every 4 weeks, patients visited our hospital and were questioned about their symptoms during the study period. Four weeks after completion of treatment, at 1-year intervals, and when symptoms occurred which we believed to have their origin in the upper gastrointestinal tract and require further treatment, endoscopy was repeated and antral and corpus biopsy specimens were taken for rapid urease test and histological examination. In addition, the 13C-urea breath test was performed for all patients 4 weeks after completion of treatment. Subjects were judged as H. pylori-negative if the results of all tests were negative. If the eradication therapy failed, subjects left the study and were treated accordingly (including re-eradication).
After initial endoscopy, patients with gastric ulcers were treated with omeprazole (20 mg o.d.) for 8 weeks, those with duodenal ulcers were treated for 6 weeks, and those with gastritis were followed without medication. Endoscopy was repeated with gastric antral and corpus biopsy specimens taken for histological examination at 1-year intervals or when symptoms recurred that were believed to have their origin in the upper gastrointestinal tract and required further treatment (including eradication therapy). Control patients visited our hospital every 4 weeks and were questioned about symptoms until the end-point. Informed consent was obtained from each patient. If control patients showed ulcer relapse and they wished to receive eradication therapy, we administrated eradication therapy to them.
Results are given as mean ± S.E.M. The statistical analyses shown were performed using the Mann–Whitney U-test for unpaired data and the χ2-test. Survival curves were constructed by the Kaplan–Meier method. Statistical significance between curves was tested by the log-rank test. A P-value < 0.05 was considered significant.
This study was approved by the ethics committee of the Gastrointestinal Research Unit of Hiroshima University Hospital.
A total of 286 patients comprised the eradication group; 223 patients were men and 63 were women. The mean age of the patients was 49.8 ± 0.7 years (range, 18–79 years). Of these patients, 75 were diagnosed endoscopically with duodenal ulcers, 55 with gastric ulcers, 23 with gastro-duodenal ulcers and 133 with gastritis. The mean follow-up period was 485 ± 20 days and 16 patients were followed for the full 3 years in the study group. Two hundred and ten (78%) patients were considered to have achieved successful H. pylori eradication (intention-to-treat analysis) 1 month after eradication therapy. In the control subjects, 48 (64%) patients with duodenal ulcers, 26 (47%) with gastric ulcers and 19 (83%) with gastro-duodenal ulcers had recurrent ulcers during the study. Within 1 year after completion of therapy, 30 (40%) patients with duodenal ulcers, 17 (13%) with gastric ulcers and 13 (57%) with gastro-duodenal ulcers had recurrent ulcers. The mean follow-up periods were 508 ± 27 days; 59 patients were followed for the full 3 years in the control group.
Before eradication, the atrophic scores of the corpus were similar between patients with and without hiatal hernia (0.67 ± 0.11 vs. 0.82 ± 0.13, P=0.88) and inflammation scores of the corpus were also similar between patients with and without hiatal hernia (1.41 ± 0.16 vs. 1.53 ± 0.16, P=0.54). There were no significant differences between patients with and without the occurrence of reflux oesophagitis regarding atrophic and inflammation scores of the corpus before the eradication for H. pylori ( Figure 1).
Thirty-six patients had new reflux oesophagitis after H. pylori eradication therapy, and only one patient in the control group had reflux oesophagitis during the 3 years of study. Using the Kaplan–Meier method, the estimated prevalence of reflux oesophagitis within 3 years was 18% after eradication therapy and 0.3% with no treatment (P < 0.01). Based on the Los Angeles Classification System, 24 cases of reflux oesophagitis were classified as grade A, 11 as grade B and 1 as grade C.
Reflux oesophagitis occurred in three out of 74 (4%) patients with duodenal ulcer, 10 out of 54 (18%) patients with gastric ulcer, 20 out of 133 (15%) patients with gastritis and three out of 23 (13%) patients with gastro-duodenal ulcer. The incidence of reflux oesophagitis after eradication of H. pylori was significantly higher in patients with gastritis, gastric ulcer or gastro-duodenal ulcer than in patients with duodenal ulcer (P < 0.01).
According to survival analysis, the presence of hiatal hernia significantly predicted the time to occurrence of reflux oesophagitis after H. pylori eradication. This is shown for study arm in Figure 2. Reflux oesophagitis occurred in 28 out of 141 (26%) patients with hiatal hernia and in eight out of 145 (7.7%) patients without hiatal hernia after eradication of H. pylori.
The severity of corpus atrophic gastritis also significantly predicted the time to occurrence of reflux oesophagitis (P < 0.01) ( Figure 3). Reflux oesophagitis occurred in 16 out of 71 (33%) patients with corpus atrophic gastritis and in 20 out of 215 (13%) patients without corpus atrophic gastritis after eradication of H. pylori. The severity of chronic inflammation of corpus did not predict the time to occurrence of reflux oesophagitis (P=0.54). Figure 4 shows the time to occurrence of reflux oesophagitis in patients with both hiatal hernia and corpus gastritis. Reflux oesophagitis occurred in 12 out of 34 (37%) patients with both hiatal hernia and corpus gastritis.
Age (P=0.49) and gender (P=0.49) were not associated with time to occurrence of reflux oesophagitis after eradication.
The present study demonstrates that H. pylori eradication therapy increases reflux oesophagitis and that the incidence was higher in patients with gastritis, gastric ulcers and gastro-duodenal ulcers than in patients with duodenal ulcers. The presence of hiatal hernia and corpus gastritis may contribute to the development of reflux oesophagitis after eradication therapy.
A recent study by Labenz et al. showed that reflux oesophagitis confirmed by endoscopy developed after H. pylori eradication therapy in 32 out of 244 (13%) H. pylori-negative patients with duodenal ulcers and in seven out of 216 (3%) H. pylori-positive patients.12 They reported the estimated incidence of reflux oesophagitis within 3 years of therapy to be 25.8% in patients with cured infections and 12.9% in patients with chronic infections. Di Mario et al. reported that gastro-oesophageal reflux disease (GERD) was noted in 44% of patients in their study of 447 patients, with duodenal ulcers during the 4-year period following eradication therapy.15 However, in a study of 275 non-ulcer dyspepsia (NUD) patients, Talley et al. found no significant difference in the severity of either regurgitation or heartburn between patients with and without eradication therapy.16 In our prospective study, the incidence of reflux oesophagitis was significantly higher in patients who underwent eradication therapy than in disease–age matched controls. Additionally, the incidence was higher in patients with gastritis or gastric ulcers than in those with duodenal ulcers.
Although the mechanism of reflux oesophagitis development after eradication is unknown, Labenz et al. showed that improvement in the quality of life after eradication leads to higher food intake and thus to weight gain.12 Obesity is a known risk factor for reflux oesophagitis.17 It has also been reported that the severity of corpus gastritis before eradication is an important factor predicting the occurrence of reflux oesophagitis.12 In our study, patients with corpus gastritis before eradication were also at high risk of developing reflux oesophagitis. The cumulative prevalence of reflux oesophagitis was 37% in patients with corpus gastritis during the 3 years following eradication. Corpus gastritis is known to be associated with gastric acid secretion. Sipponen et al. reported stimulated acid output to be decreased in relation to the severity of corpus gastritis.18 Therefore, it is thought that gastric acid secretion is lower in patients with corpus gastritis than in those without corpus gastritis. We previously reported that the eradication of H. pylori increases 24-h gastric acidity in patients with atrophic gastritis of the corpus.19 Several recent studies also showed that gastric acid secretion increases to normal or almost normal levels in patients with hypochlorhydria after successful therapy for H. pylori infection.20–23 Moreover, two recent reports showed that corpus gastritis is a protective factor against reflux oesophagitis. It was well known that patients with H. pylori of a cagA-positive strain have more severe gastritis than those with H. pylori of a cagA-negative strain. Vicari et al. reported that the prevalence of cagA-positive H. pylori strains decreased in patients with more severe complications of GERD.24 In addition, El-Serag et al. reported that corpus gastritis was less severe in patients with reflux oesophagitis than controls.25 Therefore, healing of gastritis and increased acid secretion after treatment for H. pylori might contribute to the development of reflux oesophagitis after eradication therapy in patients with gastritis or gastric ulcers, which are commonly accompanied by corpus gastritis.
On the other hand, antral gastritis (Type B) is common, and acid secretion is increased in patients with duodenal ulcers. H. pylori eradication therapy decreases acid secretion in patients with duodenal ulcers.26, 27 Therefore, the presence of corpus gastritis and the recovery of acid secretion after eradication did not affect the occurrence of reflux oesophagitis in patients with duodenal ulcers. This observation might be explained by the presence of ammonia from H. pylori in the gastric juice. It is known that ammonia protects against reflux oesophagitis. Ammonia has a pKa of 9.1 and acts as a powerful neutralizing substance when gastric pH is elevated but has no effect on physiological pH in the spontaneously secreting stomach. Therefore, ammonia could reduce the corrosive effect of gastric juice refluxing into the oesophagus. In vitro, ammonia leads to protective mucosal adaptation in the rat stomach.28 Eradication of H. pylori stops the production of ammonia. In our recent experimental rat study, injection of ammonia protected against the development of oesophagitis.29 The incidence of reflux oesophagitis was significantly lower in two out of seven (29%) rats treated with ammonia than in the seven out of seven (100%) control rats. Therefore, it is possible that ammonia plays a role in protecting against the development of reflux oesophagitis. However, the removal of ammonia production cannot explain the different rates of development of reflux oesophagitis among the patient groups.
A previous study demonstrated the contribution of hiatal hernia to clearance impairment in patients with oesophagitis.30 Patti et al. demonstrated the degree of oesophagitis to be greatest in the presence of a large hiatal hernia.31 A recent study by Kahrilias et al. showed that hiatal hernia reduces lower oesophageal sphincter pressure and alters the pressure topography of the gastro-oesophageal junction by reducing the maximal pressure.32 In multivariate analysis, Cadiot et al. showed that lower oesophageal sphincter pressure hypotonicity is an independent risk factor for both the presence and grade of reflux oesophagitis.33 In the present study, the presence of hiatal hernia also contributed significantly to the development of reflux oesophagitis after H. pylori eradication.
Our data showed that after eradication therapy, most patients with reflux oesophagitis had mild reflux oesophagitis (grade A or B according to the Los Angeles Classification System). A report by Labenz et al. showed oesophagitis after eradication to be uncomplicated (stage I or II) in 38 out of 39 patients and that only one patient with stenosis of the duodenal bulb developed reflux oesophagitis with circumferential ulceration.12 Therefore, most occurrences of reflux oesophagitis after H. pylori eradication are mild.
In conclusion, the eradication of H. pylori increased the prevalence of reflux oesophagitis in our patient group. The presence of corpus gastritis and hiatal hernia might contribute to the development of at least mild reflux oesophagitis after eradication therapy.
We wish to thank Tetsue Sarusawa for help with the manuscript preparation. This work was supported by a Grant-in-Aid for Cancer Research from the Ministry of Health and Welfare of Japan (9–8).