Relationship between intestinal metaplasia of the gastro-oesophageal junction, Helicobacter pylori infection and gastro-oesophageal reflux disease: a prospective study

ALIMENTARYTRACT

DlGESlllVER OIS 2000;32:567-72

Relationship between intestinal metaplasia of the gastro-oesophageal junction, Helicobacter pylori infection and gastro-oesophageal reflux disease: a prospective study 0. Pieramico M.V. Zanettil

Background. The role of Helicobacter pylori infection and/or gastro-oesophageal reflux disease in pathogenesis of intestinal mataplasia in gastric cardia is still unclear. Aims. To prospectively evaluate prevalence of inflammation and intestinal metaplasia of cardia in relationship to Helicobacter pylori infection in patients with gastro-oesophageal reflux disease and in healthy controls. Patients. A total of 122 consecutive patients with gastro-oesophageal reflux disease and 49 control subjects were included. Methods. During endoscopy a total of six biopsies were taken from antrum, corpus and cardia. Helicobacter pylori infection was assessed by histology and rapid uraase test. Degree of chronic gastritis, inflammatory activity and Helicobacter pylori colonization were scored from 0 to 3. Results. No difference in prevalence was observed between gastro-oesophageal reflux disease patients and controls as far as concerns Helicobacter pylori (4 1% vs 38%], inflammation of cardia 159.5% vs 70%) and intestinal metaplasia of cardia [I 8% vs IS%]. Inflammation of cardia was significantly [pp
Digest

Liver

Dis 2000;32:567-72

Key words: gastro-oasophagaal /ori; intestinal metaplesia

junction;

gastro-oesophageal

reflux

disease;

Helicobacter

py-

Castro-oeropha@eal

junction in 60RD

Introduction The gastric cardia consists of an area of a few centimetres localized immediately below the squamocolumnar junction. Inflammatory changes and intestinal metaplasia (IM) in this area have recently been afforded increased interest due to the hypothesis that these changes may potentially represent a precursor lesion of cardia neoplasms, the incidence of which has risen in the last few decades lm3. A relationship between inflammation of cardia and gastro-oesophageal reflux disease (GORD) has been shown by some investigators 4 5, but was not confirmed by others who have found a relationship between IM and inflammation of cardia and Helicobacter pylori (H. pylori) 6 l. A possible role of H. pylori infection in the pathophysiology of GORD is still debated. Considering that gastric acid plays an important role in the development of GORD and that the gastric corpus is the main acid-producing region of the stomach, H. pylori-related corpus gastritis has been associated with reduced gastric acid secretion indicating a protective role of H. pylori for GORD *. Several studies have investigated the association between H. pylori and GORD with controversial results g-17.Some authors observed no difference in the prevalence rate of H. pylori between patients with oesophagitis and controls g-‘1. In other studies, however, a lower prevalence rate of the infection was reported in GORD leading to the hypothesis that H. pylori infection may protect from GORD I2 13. IM of the cardia may be a marker of malignant potential although it is not known whether it arises as a long-term consequence of acid exposure, as in GORD or as a result of gastritis associated with H. pylori infection. A prevalence rate of IM of the cardia has been reported in 9% to 36% of patients undergoing endoscopy lgmz4.This finding was associated with histologic oesophagitis 18, or an increased frequency of GORD symptoms lg. On the contrary, other investigators 7 2o observed no relationship between GORD and IM but an association to H. pylori infection. This study was, therefore, carried out in patients with GORD and in asymptomatic subjects to prospectively investigate the prevalence of inflammation and IM of the cardial region and their relationship with H. pylori infection.

Patients and Methods Patient population A total of 122 consecutive patients referred for GORD symptoms to the Endoscopy Unit over a seven-month 568

period from June 1998 to January 1999, were included in the investigation. The protocol of the study was approved by the Local Institutional Board and all patients gave informed consent to take part. On a clinical basis, GORD was defined by the presence of typical symptoms such as heartburn and/or acid regurgitation at least twice a week for at least 6 months. Clinical data regarding age, sex, past history, medications, frequency and duration of symptoms were collected prior to endoscopy. Exclusion criteria included age ~18 and >75 years, previous eradication therapy, use of antibiotics in the last 2 months, anticoagulant therapy, oesophageal or gastric surgery, past or current history of peptic ulcer disease, gastric or oesophageal cancer and upper gastrointestinal bleeding. In addition, patients showing any length of columnar-appearing mucosa in the tubular oesophagus at endoscopy with the histological presence of IM in this area (Barrett’s oesophagus) were also excluded. In patients with GORD, only antacids were allowed in the two weeks prior to endoscopy in order to correctly assess the extent of oesophagitis. The control population consisted on 49 subjects (30 males, 19 females; mean age 50+15 SD years) who underwent endoscopy in the same period as the study for reasons other than GORD symptoms, Barrett’s oesophagus, active or previous peptic ulcer disease, gastric or oesophageal neoplasms, or dyspepsia. This group included patients with chronic diarrhoea requiring small bowel biopsy and subsequently no histological evidence of coeliac disease (7 patients), mild iron deficiency anaemia (haemoglobin > 10 g/dl) (19 patients), prior to laparoscopic cholecystectomy (11 patients), to exclude evidence of neoplasm in the upper gastrointestinal tract (12 patients). Endoscopy and histology All endoscopies were performed by the same investigator (OP) in a standard fashion, with visualization of the oesophagus, stomach, and duodenum. The presence of reflux oesophagitis was identified and graded according to the Savary-Miller classification as follows: grade 1 - simple erosions; grade 2 - confluent erosions; grade 3 - circumferential erosions; and grade 4 - oesophageal ulcers or strictures. The gastro-oesophageal junction was identified by the junction of the tubular oesophagus with the most proximal gastric folds and the position of the lower oesophageal sphincter. The cardia biopsy specimens were obtained within 1 cm below the squamo-columnar junction using a standard biopsy forceps under direct vision with the gastroscope in the conventional anterograde position in the distal oesophagus. In addition, two gastric biopsy specimens were obtained from the antrum and corpus. The antrum biopsies were taken from within 2 cm of

0. Pieramico, MU. Zanelti

the pylorus. The corpus biopsies were obtained from the largest folds on the gastric greater curvature. Specimens were stained with haematoxylin-eosin to evaluate inflammatory changes and with Giemsa stain to assess H. pylori colonization. The presence of gastritis, activity of gastritis, IM and H. pylori colonization was assessed by a pathologist unaware of patient symptoms. Chronic gastritis, defined by the presence of mononuclear cells within the lamina propria, was graded as: grade 0 - absent; grade 1 - mild; grade 2 moderate, and grade 3 - severe. The activity of gastritis, as assessed by the presence of polymorphonuclear cells within the lamina propria, was scored in a similar manner. If any of the specimens from the same gastric region showed different scores, only the higher score was considered for analysis. Finally, specimens were evaluated for the presence or absence of IM. Complete IM was defined as the presence of goblet cells without acidic alcian-blue acid Schiff positive material in columnar-type cells. Incomplete IM was defined as the presence of goblet cells, with acidic mucins in goblet and adjacent columnar-appearing cells.

no significant difference in the H. pylori colonization rates when analysis was performed after stratifying patients according to the grade of oesophagitis: grade 0 (27/68, 40%), grade I (21/44, 48%), and grade II-III (3/10, 33%) (Fig. 1).

Statistical analysis The prevalence rates, estimated as proportions and differences between groups were determined using X2 and Fisher’s exact test. A Mann-Whitney U-test (twotailed) was used to compare the mean ages between groups. A p value co.05 was considered as significant. Results

A total of 142 patients with GORD were clinically and endoscopically assessed and initially enrolled in the study. Barrett’s oesophagus was found in 20 of these patients, 14 of whom presented with a short segment <3 cm, and 6 with a long segment, all of whom were excluded from further analysis. A total of 122 patients with GORD (7 1 males, 5 1 females; mean age 49 years, range 21-73) were, therefore, available for final analysis. Of these, 54 presented with endoscopic evidence of reflux oesophagitis, while the remaining 68 patients with GORD had no endoscopic signs of reflux oesophagitis. There was no significant difference in sex or age between GORD patients and control subjects. Prevalence of H. pylori infection The overall prevalence of H. pylori infection was 41% (70/171 patients) being 41.8% in the GORD group (51022 patients) and 38% in the control population (19149 subjects) (Table I). The difference was not statistically significant. In patients with GORD, there was

As far as concerns patients with H. pylori infection (n=70), bacteria were identified in all three biopsy sites (antrum, corpus and cardia) in 64170 patients (91%) being 46/51 in those with GORD (90%) and 18/19 in controls (95%) (ns). When present, bacteria were identified in all antrum and corpus specimens. The density of bacterial colonization in antrum and in corpus did not differ between GORD and controls, however, in the cardia it was significantly (p=O.O3) less marked in controls than in GORD patients. Inflammation and IM of cardia Inflammation of the cardia was detected in 107 of all 171 subjects (61%) and no difference was found between the control population (34/49, 69%) and GORD patients (73/122, 60%) (Table II). 569

Grstro-oesopbs~d

junction in 60RO

not. Twenty-two subjects out of the 32 (69%) with cardial IM had GORD and 10 were control subjects (3 1%). Thirteen subjects with cardial IM had GORD and were positive for H. pylori infection (4 1%). IM of the cardia was significantly more frequently associated (p=O.OOl) with inflammation of the cardia (27/107,25%) than with a normal histological appearance of the cardia (5/64, 8%). Overall, 27/32 (84%) patients with IM of the cardia also had carditis and 5/32 (16%) had not.

Inflammation of the cardia was significantly (pcO.001) more frequent in subjects with H. pylori infection (65/70, 93%) than in those without infection (42/101, 42%) but no difference was observed between the two subgroups of GORD and controls with H. pylori infection (92% vs 94%, respectively). Overall, 65/107 (61%) patients with carditis were colonized with H. pylori and 42/107 (39%) were not (p=O.O02) (Fig. 2).

The grade of chronic inflammation of the cardia was similar in all patients with GORD and the control population and similarly, did not differ between H. pyloripositive GORD and controls (Table I). In patients with GGRD, inflammation of the cardia was detected in 34/54 (63%) patients with erosive oesophagitis and in 42/68 (62%) patients without erosive oesophagitis. IM of the cardia was identified in 32 of all 171 subjects (19%) and showed no difference between patients with GGRD (22/122, 18%) and controls (10/49, 20%). However, the frequency of lM of the cardia in all subjects with H. pylori infection (19/70, 27%) was about twice (p=O.O3) that in the uncolonized population (13/101, 13%), independently of GORD symptoms. Overall, 19/32 (59%) patients with IM of the cat&a were colonized with H. pylori and 13/32 (41%) were 570

Discussion Results of this study show that inflammation of the cardia is detected in about two-thirds of the patients irrespective of the presence of GORD symptoms but is more than twice as frequent in H. pylori infection whilst the frequency is comparable in patients with GORD with or without erosive oesophagitis. IM of the cardia is significantly associated with chronic inflammation in that region and the frequency in patients with H. pylori infection is twice that in uncolonized patients without any relationship with GORD symptoms or with the presence of oesophageal erosions. A number of studies have attempted to assess the prevalence of H. pylori infection in CORD with controversial results g-17,but the overall prevalence of the infection in published studies varies widely and it does not appear to be significantly different between patients with GORD and the general population. The prevalence of H. pylori infection in GORD has been reported to range between 20% to 75% with a mean prevalence of about 35% and a similar wide range has also been shown in the control population being between 36% and 82% 25. Several explanations may account for the differences in prevalence of H. pyZori in GORD between studies. First, the selection of patients with GORD differs between the various studies. Most studies have included only patients with erosive oesophagitis g-12whereas other authors l4 studied patients with and without oesophagitis. A second critical point is the lack of a prospective evaluation of the prevalence of the infection in most studies and an inadequate number of patients upon which to draw conclusions. Third, the selection of a control population was very heterogeneous including patients with a normal appearance at endoscopy irrespective of abdominal symptoms or patients with irritable bowel or unexplained dyspepsia. Finally, the age of subjects is a further critical point. This is especially important, since the prevalence of the infection increases with advancing age to more than 50% in subjects over 60 years of age. In fact, some studies have reported a high prevalence of the infection in patients with GORD, mainly

0. Pienmico,

because elderly patients had been included l5 l6 and analysis of the data is even more difficult if the age of the control population is not adjusted. In our study, we evaluated prospectively patients with clinically assessed GORD including, therefore, patients with or without oesophagitis and excluding patients with Barrett’s oesophagus. The control population included subjects with a normal appearance of the mucosa of the upper gastrointestinal tract at endoscopy. The prevalence of H. pylori infection did not differ in our study populations of patients with GORD and controls and closely resembles the mean prevalence of infection reported in other studies on this issue. Moreover, the prevalence of H. pylori infection in our control population was the same (38%) as that reported in a larger study population of asymptomatic blood donors previously tested in our area 25.However, we cannot exclude that a possible difference may have been missed due to the relatively small sample size. Other results on the relationship between H. pylori infection and the severity of reflux oesophagitis are also conflicting 9 lo l2 27-30.It has been reported that the severity of reflux oesophagitis is positively 28 or negatively correlated with H. pylori infection 30. Other authors, however, have not found any relationship 9 lo l2 27. Confirming our previous report of a retrospective study 29, we found that H. pylori infection is not related to the severity of reflux oesophagitis. It may, therefore, be concluded that it is unlikely that H. pylori plays an important role in the pathogenesis of GORD. In infected patients, bacteria have been found in all gastric regions, including cardia, in over 90% of the subjects. This finding is in agreement with data reported in the literature 6 3’. However, we could not confirm the finding of a lower bacterial density in the cardia with respect to antrum and corpus in patients with GORD, reported by others 31. In our patients with GORD, the grade of bacterial colonization was similar in antrum, corpus and cardia. Therefore, our data do not support the hypothesis that luminal acid in GORD could reduce H. pylori density in the cardia. It has been suggested that inflammation of the cardia (carditis) may be a histological marker of GORD 4 and that carditis may be the only abnormal histological finding in GORD 5, however, a relationship between H. pylori infection and carditis has also been reported 7. Goldblum et al. 7 have not shown any relationship between GORD and carditis but a close association with H. pylori infection. Other authors 27, however, have found that carditis is less closely linked to H. pylori than mucosal inflammation in other gastric regions. Inflammation of the gastric cardia was found in 61% of subjects in our study population which is in agreement with other studies showing a prevalence ranging from 41% to 79% 7 27.We did not find any relationship between carditis and GORD since

M.V. Zanetti

the prevalence of carditis in asymptomatic subjects and patients with GORD was similar. On the contrary, we found that over 90% of patients colonized with H. pyZori had carditis irrespective of GORD symptoms. This prevalence was more than twice that in uncolonized subjects. We can, therefore, conclude that inflammation of the cardia is not associated with GORD and is not even a marker of it. Carditis shows a close relationship to H. pylori infection and is, in most cases, associated with mucosal inflammation of the other gastric regions. IM of the cardia is a finding of increasing potential importance since it may be associated with adenocarcinoma of the cardia the incidence of which, in contrast to other tumours located elsewhere in the stomach, has increased in the last few decades le3. Other studies have attempted to investigate the prevalence of IM in the cardia and have reported a prevalence ranging from 5% to 23% I8 24. Differences in patient population and number of biopsies may probably account for the variability of prevalence of these findings. Morales et al. 23 found a 23% prevalence of IM of the cardia, however, the mean age of the study population was 61 years which is over 10 years more than patients in our study and included subjects with a range up to 85 years old. Indeed, Voutilanen et al. 22have demonstrated that age is an independent risk factor for IM of the cardia. In our study, 19% of all subjects had IM of the cardia, a result which is within the range of data reported in the literature. A possible relationship between H. pylori infection and IM in the cardia has been challenged. Two studies performed in patients with or without GORD 732 found a significant association between H. pylori infection and IM of the cardia. This finding, however, was not confirmed by other authors. El-Serag et al. *I observed that 52% of patients with IM of the cardia were colonized with H. pylori and 42% of patients without. Bowrey et al. 27 found no significant difference in the prevalence of IM of the cardia between patients colonized or not with H. pylori. In our study, H. pylori infection was about twice more frequently associated with IM of the cardia than in uncolonized subjects and 59% of patients with IM of the cardia had H. pylori. IM of the cardia was about three times more frequently associated with inflammation of the cardia. We were not able to find any association between GORD and inflammation and IM of the cardial region. Our data, therefore, strongly support the concept that IM in the cardia is related to H. pylori infection rather than to GORD. In conclusion, H. pylori infection does not seem to play an important role in the pathogenesis of GORD. Inflammation and IM of the cat-dial region are not markers of GORD but are related to H. pylori infection. Whether H. pylori infection may be involved in the pathogenesis of adenocarcinoma of the cardia remains to be elucidated. 571

Castro-oesophageal

junction

in

60RO

References ’ Clark GWB, Smyrk TC, Burdiles P, Hoeft SF, Peters JH, Kiyabu M. Is Barrett’s metaplasia the source of adenocarcinoma of the cardia? Arch Surg 1994;129:609-14. 2 Powell J, McConkey CC. Increasing incidence of adenocarcinoma of the gastric cardia and adjacent sites. Br J Cancer 1990;62:440-3. 3 Blot WJ, Devesa SS, Kneller RW, Fraumeni JF. Rising incidence of adenocarcinoma of the esophagus and gastric cardia. J Am Med Ass 1991;265:1287-9. 4 Oberg S, Peters JH, DeMeester TR, Chandrasoma P, Hagen JA, Ireland AP Inflammation and specialized intestinal metaplasia of cardiac mucosa is a manifestation of gastroesophageal reflux disease. Ann Surg 1997;226:522-32. 5 Riddell RH. The biopsy diagnosis of gastroesophageal reflux disease, “carditis” and Barrett’s esophagus, and sequalae of therapy. Am J Surg Path01 1996;2O(Suppl 1):31-50. 6 Genta RM, Huberman RM, Graham DY. The gastric cardia in Helicobacter pylori infection. Hum Path01 1994;25:915-9. 7 Goldblum JR, Vicari JJ, Falk GW, Rice TW, Peek RM, Easley K, et al. Inflammation and intestinal metaplasia of the gastric cardia: The role of gastroesophageal reflux and H. pylori infection. Gastroenterology 1998; 114:633-9. 8 El-Serag HB, Sonnenberg A, Jamal JJ, Inadomi JM, Crooks LA, Feddersen RM. Corpus gastritis is protective against reflux oesophagitis. Gut 1999;45:181-5. 9 Newton M, Bryan R, Bumham WR, Kamm MA. Evaluation of Helicobacter pylori in reflux esophagitis and Barrett’s mucosa. Gut 1997;40:9-13. I0 O’Connor HJ, Cunnane K. Helicobacter pylori and gastroesophageal reflux disease: a prospective study. Ir J Med Sci 1994;163:369-73. I’ Befrits R, Grandstrom M, Rylander M. Helicobacter pylori in 205 consecutive endoscopy patients. Stand J Infect Dis 1993;25:185-91. I* Werdmuller BFM, Loeffeld RJLE Helicobacter pylori infection has no role in the oathoeenesis of reflux esonhagitis. Dirr Dis Sci _ I 1997;42:103-5. A I3 Vicari JJ, Peek RM, Falk GW. The seroprevalence of tag-A-positive Helicobacter pylori strains in the spectrum of gastroesophageal reflux disease. Gastroenterology 1998; 115:50-7. I4 Csendes A, Smok G, Cerda G. Prevalence of Helicobacter pylori infection in 190 control subjects and in 236 patients with gastroesophageal reflux, erosive esophagitis or Barrett’s mucosa. Dis Esophagus 1997;10:38-42. I5 Liston R, Pitt MA, Banerjee AK. Reflux oesophagitis and Helicobacter pylori infection in elderly patients. Postgrad Med J 1996;72:221-3.

572

I6 Ohara S, Sekine H, Iijima K. Gastric mucosal atrophy and prevalence of Helicobacter pylori in reflux esophagitis of the elderly. Jpn J Gastroenterol 1996;93:235-9. I7 O’Connor HJ. Helicobacter pylori and gastro-oesophageal reflux disease: clinical implications and management. Aliment Pharmaco1 Ther 1999;13:117-27. I8 Spechler SJ, Zeroogian JM, Antonioli DA. Prevalence of metaplasia in the gastroesophageal junction. Lancet 1994;344: 1533-6. I9 Johnston MH, Hammond AS, La&in W, Jones M. The prevalence and the clinical characteristics of short segments of specialized intestinal metaplasia in the distal esophagus on routine endoscopy. Am J Gastroenterol 1996;9 1: 1507- 11. 2o Trudgill NJ, Suvama SK, Kapur KC. Intestinal metaplasia at the squamocolumnar junction in patients attending for diagnostic gastroscopy. Gut 1997;41:585-9. 21 El-Serag HB, Sonnenberg A, Jamal MM, Kunkel D, Crooks L, Feddersen RM. Characteristics of intestinal metaplasia in the gastric cardia. Am J Gastroenterol 1999;94:622-7. 22 Voutilanen M, Farkkill M, Juhola M, Mecklin JP, Sipponen P and the Central Finland Endoscopy Study Group. Complete and incomplete intestinal metaplasia at the oesophagogastric junction: prevalences and associations with endoscopic erosive oesophagitis and gastritis. Gut 1999;45:644-8. 23 Morales TG, Sampliner R, Bhattacharyya A. Intestinal metaplasia of the gastric cardia. Am J Gastroenterol 1997;92:414-8. 24 Hackelsberger A, Gunther T, Schultze V, Manes G, DominguezMunoz JE. Roessner A, et al. Intestinal metaplasia at the pastroesophageal junction: Helicobacter pylori gastritis or &troesophageal reflux disease? Gut 1998;43:17-21. 25 Lee JM, O’Morain CA. Different management for H. pylori positive and negative patients with gastro-oesophageal reflux disease? Gut 1998;43(Suppl l):S14-20. 26 Pieramico 0, Zanetti MV, Franceschini I. Seroprevalence of H. pylori infection in healthy people, hospital patients and patients with dyspepsia in South Tyrol. Ital J Gastroenterol 1996;28(Sup~12): 196-7. 27 Bowrey DJ, Clark GWB, Williams GT. Patterns of gastritis in patients with gastro-oesophageal reflux disease. Gut 1999;45:798-803. 28 Sekiguchi T, Shirota T, Horikoshi. Helicobacter pylori infection and severity esophagitis. Gastroenterology of reflux 1996;110:A755.

29 Pieramico 0, Zanetti MV. Role of H. pylori infection in gastroesophageal reflux disease. Gut 1997;41(Suppl l):A24. 30 Grebenev AL. Helicobacter pylori and reflux esophagitis. Am J Gastroenterol 1994;89:A1372. 31 Hackelsberger A, Gunther T, Schultze V, Labenz J, Roessner A, Malfertheiner P Prevalence and pattern of Helicobacter pylori gastritis in the gastric cardia. Am J Gastroenterol 1997;92:2220-4.