Gastric cardia inflammation: role of helicobacter pylori infection and symptoms of gastroesophageal...

4
Gastric Cardia Inflammation: Role of Helicobacter pylori Infection and Symptoms of Gastroesophageal Reflux Disease Sergio Morini, M.D., Angelo Zullo, M.D., Cesare Hassan, M.D., Roberto Lorenzetti, M.D., Francesca Stella, M.D., and Maria Teresa Martini, M.D. Department of Gastroenterology and Digestive Endoscopy, “Nuovo Regina Margherita” Hospital; and Department of Pathological Anatomy, “San Giacomo” Hospital, Rome, Italy OBJECTIVE: Although high prevalences of both chronic in- flammation (carditis) and intestinal metaplasia at the gastric cardia have been reported, the pathogenesis is still unclear. This study assesses the role of Helicobacter pylori (H. pylori) infection and symptoms of gastroesophageal reflux disease (GERD) in these histological alterations. METHODS: Consecutive patients who underwent upper en- doscopy were enrolled in the study, irrespective of their symptoms. Patients previously treated for H. pylori infec- tion and those using proton pump inhibitors were excluded. Two biopsies were performed in the antrum, two in the gastric body, and two at the gastric cardia. All biopsies were used to look for H. pylori and for histological assessment. RESULTS: A total of 133 patients were enrolled. Carditis and intestinal metaplasia at the cardia were detected in 100 (75.2%) and in 18 (13.5%) patients, respectively. The H. pylori infection rate was significantly higher in patients with carditis than in those without it (87/100 vs 7/33; p , 0.0001), and was higher in those with intestinal metaplasia at the cardia than in those without it (17/94 vs 1/39; p 5 0.03). Conversely, the prevalence of GERD symptoms was not significantly different between patients with and without carditis (34/100 vs 16/33; p 5 NS), and between those with and without intestinal metaplasia (5/50 vs 13/83; p 5 NS). Interestingly, the prevalence of both H. pylori (64/94 vs 39/94; p 5 0.0005) and intestinal metaplasia (18/133 vs 4/133; p 5 0.0042) in the gastric cardia was significantly higher than that in gastric body. CONCLUSION: According to our study data, the gastric car- dia is frequently infected with H. pylori with consequent development of both carditis and intestinal metaplasia, whereas GERD does not seem to be involved in these histological changes. (Am J Gastroenterol 2001;96: 2337–2340. © 2001 by Am. Coll. of Gastroenterology) INTRODUCTION Chronic inflammation is frequently detected in the mucosa of the gastric cardia (i.e., “carditis”) (1), and is associated with intestinal metaplasia in 9 –23% of cases (2– 4). More- over, the development of dysplasia in patients with intestinal metaplasia in the gastric cardia has also been reported at follow-up (5, 6). The significance of these histological changes is of current interest, as epidemiological studies have shown an increase in the incidence of gastric cardia adenocarcinoma over the last decades (7). The pathogenesis of the chronic damage of the cardia is still unclear (8, 9). Some authors assert a major role for gastroesophageal reflux disease (GERD), assuming that increased acid reflux may lead to chronic injury of the cardia, including the develop- ment of metaplasia, similarly to what occurs in the distal esophagus (10, 11). Conversely, other studies found an association between Helicobacter pylori (H. pylori) infec- tion and the presence of both chronic inflammation and intestinal metaplasia in the cardia (4, 12). It has also been suggested that both H. pylori and GERD may independently cause chronic inflammation of the gastric cardia in different patients (13). On the other hand, according to other authors, neither H. pylori infection nor GERD play a role in the onset of carditis, with other factors such as age being associated with its development (14). Unraveling the pathogenic factors involved in carditis development could have a clinical impact on the manage- ment of such patients with respect to therapeutic approach and follow-up, and the conflicting results available suggest that further investigations are needed (9). The present study was performed to further evaluate the prevalence and the factors involved in the pathogenesis of chronic inflamma- tion and intestinal metaplasia development at the gastric cardia. MATERIALS AND METHODS Patients referred to our endoscopic unit for diagnostic upper GI endoscopy were consecutively selected to be enrolled in the study. Patients included in the study had at least one of the following symptoms: epigastric pain, discomfort in the upper abdomen, delayed digestion, nausea/vomiting, belch- ing, heartburn, or regurgitation. The presence of heartburn THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 96, No. 8, 2001 © 2001 by Am. Coll. of Gastroenterology ISSN 0002-9270/01/$20.00 Published by Elsevier Science Inc. PII S0002-9270(01)02601-6

Transcript of Gastric cardia inflammation: role of helicobacter pylori infection and symptoms of gastroesophageal...

Gastric Cardia Inflammation:Role of Helicobacter pyloriInfection andSymptoms of Gastroesophageal Reflux DiseaseSergio Morini, M.D., Angelo Zullo, M.D., Cesare Hassan, M.D., Roberto Lorenzetti, M.D.,Francesca Stella, M.D., and Maria Teresa Martini, M.D.Department of Gastroenterology and Digestive Endoscopy, “Nuovo Regina Margherita” Hospital; andDepartment of Pathological Anatomy, “San Giacomo” Hospital, Rome, Italy

OBJECTIVE: Although high prevalences of both chronic in-flammation (carditis) and intestinal metaplasia at the gastriccardia have been reported, the pathogenesis is still unclear.This study assesses the role ofHelicobacter pylori (H.pylori) infection and symptoms of gastroesophageal refluxdisease (GERD) in these histological alterations.

METHODS: Consecutive patients who underwent upper en-doscopy were enrolled in the study, irrespective of theirsymptoms. Patients previously treated forH. pylori infec-tion and those using proton pump inhibitors were excluded.Two biopsies were performed in the antrum, two in thegastric body, and two at the gastric cardia. All biopsies wereused to look forH. pylori and for histological assessment.

RESULTS: A total of 133 patients were enrolled. Carditis andintestinal metaplasia at the cardia were detected in 100(75.2%) and in 18 (13.5%) patients, respectively. TheH.pylori infection rate was significantly higher in patients withcarditis than in those without it (87/100vs 7/33; p ,0.0001), and was higher in those with intestinal metaplasiaat the cardia than in those without it (17/94vs 1/39; p 50.03). Conversely, the prevalence of GERD symptoms wasnot significantly different between patients with and withoutcarditis (34/100vs16/33;p 5 NS), and between those withand without intestinal metaplasia (5/50vs 13/83;p 5 NS).Interestingly, the prevalence of bothH. pylori (64/94 vs39/94; p 5 0.0005) and intestinal metaplasia (18/133vs4/133; p 5 0.0042) in the gastric cardia was significantlyhigher than that in gastric body.

CONCLUSION: According to our study data, the gastric car-dia is frequently infected withH. pylori with consequentdevelopment of both carditis and intestinal metaplasia,whereas GERD does not seem to be involved in thesehistological changes. (Am J Gastroenterol 2001;96:2337–2340. © 2001 by Am. Coll. of Gastroenterology)

INTRODUCTION

Chronic inflammation is frequently detected in the mucosaof the gastric cardia (i.e., “carditis”) (1), and is associated

with intestinal metaplasia in 9–23% of cases (2–4). More-over, the development of dysplasia in patients with intestinalmetaplasia in the gastric cardia has also been reported atfollow-up (5, 6). The significance of these histologicalchanges is of current interest, as epidemiological studieshave shown an increase in the incidence of gastric cardiaadenocarcinoma over the last decades (7). The pathogenesisof the chronic damage of the cardia is still unclear (8, 9).Some authors assert a major role for gastroesophageal refluxdisease (GERD), assuming that increased acid reflux maylead to chronic injury of the cardia, including the develop-ment of metaplasia, similarly to what occurs in the distalesophagus (10, 11). Conversely, other studies found anassociation betweenHelicobacter pylori(H. pylori) infec-tion and the presence of both chronic inflammation andintestinal metaplasia in the cardia (4, 12). It has also beensuggested that bothH. pylori and GERD may independentlycause chronic inflammation of the gastric cardia in differentpatients (13). On the other hand, according to other authors,neitherH. pylori infection nor GERD play a role in the onsetof carditis, with other factors such as age being associatedwith its development (14).

Unraveling the pathogenic factors involved in carditisdevelopment could have a clinical impact on the manage-ment of such patients with respect to therapeutic approachand follow-up, and the conflicting results available suggestthat further investigations are needed (9). The present studywas performed to further evaluate the prevalence and thefactors involved in the pathogenesis of chronic inflamma-tion and intestinal metaplasia development at the gastriccardia.

MATERIALS AND METHODS

Patients referred to our endoscopic unit for diagnostic upperGI endoscopy were consecutively selected to be enrolled inthe study. Patients included in the study had at least one ofthe following symptoms: epigastric pain, discomfort in theupper abdomen, delayed digestion, nausea/vomiting, belch-ing, heartburn, or regurgitation. The presence of heartburn

THE AMERICAN JOURNAL OF GASTROENTEROLOGY Vol. 96, No. 8, 2001© 2001 by Am. Coll. of Gastroenterology ISSN 0002-9270/01/$20.00Published by Elsevier Science Inc. PII S0002-9270(01)02601-6

and/or regurgitation at least twice weekly for$6 monthswas considered to be suggestive of GERD (15). Patientspreviously treated forH. pylori infection, those who hadtaken proton pump inhibitors and/or antibiotics in the monthbefore endoscopic examination, those with malignancy orprevious gastric surgery, and those with any contraindica-tion for biopsy at endoscopy were excluded from the study.

During endoscopy, three biopsies were performed in theantrum, three in the gastric body, and two additional biopsyspecimens were taken within 2 cm below a normal-lookingsquamocolumnar junction, as previously suggested (2, 16).All biopsy specimens were used for histological assessmentwith hematoxylin and eosin staining, and to look forH.pylori using Giemsa staining. Chronic inflammation andgastritis activity were graded according to the updated Syd-ney system in all gastric sites (17). Identification ofH. pyloriin any biopsy specimen from the cardia, corpus, or antrumwas considered diagnostic of infection. Cardiac mucosa wasidentified by its tubular or coiled racemose glands composedof mucus secreting cells, and intestinal metaplasia was de-fined by the presence of goblet cells in glandular mucosa(18). Alcian blue (pH 2.5)/periodic acid-Schiff staining wasused in selected cases in which there was uncertainty aboutthe presence of intestinal metaplasia. Intestinal metaplasiawas recorded as either absent or present. Slides were re-viewed by a single pathologist who was unaware of theclinical and endoscopic findings. Hiatal hernia at endoscopywas defined as pouch of gastric mucosa.2 cm long be-tween the endoscopic squamocolumnar junction and thediaphragmatic hiatus, as suggested previously (19). Erosiveesophagitis was defined according to the Savary-Miller clas-sification (20).

Statistical AnalysisStatistical analysis was carried out using thet, Fisher’sexact, andx2 tests as appropriate. Values ofp , 0.05 wereconsidered to be statistically significant.

RESULTS

A total of 140 patients were included in the study. Sevenpatients were excluded from the following analysis becausethe biopsy sampling at the cardia either was inadequate (twocases) or revealed oxintic (three cases) or esophageal (twocases) mucosa. Clinical and demographic characteristics ofthe remaining 133 patients are given in Table 1.

Histological assessment of the cardia showed chronicinflammation in 100 (75.2%) patients, with active inflam-matory infiltrate in 57 of them, and a normal histologicalfeature in the remaining 33 (24.8%) patients. There was nostatistically significant difference between the two groupsregarding age (566 12vs546 17 yr; respectively) and sexdistribution (male/female ratio, 44/56vs 16/17, respective-ly). The H. pylori infection rate was significantly higher inpatients with carditis than in those without gastric cardiainflammation (87/100vs 7/33; p , 0.0001), whereas theprevalence of both GERD symptoms (34/100vs16/33;p 5NS) and hiatal hernia (25/100vs 11/33;p 5 NS) were notsignificantly different between the groups. Specifically, ofthe 94H. pylori–positive patients, bacteria were found onlyin the gastric cardia in three (3.2%) patients, in both theantrum and cardia (but not in the gastric body) in 27 (28.7%)patients, and in all the gastric sites (antrum, gastric body,and cardia) in 34 (36.2%) patients. In the remaining 30patients,H. pylori was present only in the antrum in 18(19.1%) patients and in both the antrum and the body in 12(12.7%) patients. All 57 patients withH. pylori detected inthe cardia showed chronic active inflammation at this site.Interestingly, the overallH. pylori prevalence in gastriccardia was significantly higher than that in the gastric body(64/94vs39/94;p 5 0.0005) and was lower than that in theantrum (64/94vs 91/94;p , 0.0001).

As shown in Table 2, intestinal metaplasia in the gastriccardia was detected in 18 (13.5%) patients. The prevalenceof intestinal metaplasia in the gastric cardia was higher thanthat in the gastric body (18/133vs 4/133;p 5 0.0042) andwas similar to that in the antrum (18/133vs 29/133;p 5NS). In addition, a significant association between the pres-ence of intestinal metaplasia in the gastric cardia and in theantrum was observed (p 5 0.001). Intestinal metaplasia inthe gastric cardia was found in 17 of 94H. pylori–positivepatients and in one of 39H. pylori–negative patients (p 50.03). It is noteworthy that among theH. pylori–positivepatients with intestinal metaplasia in the gastric cardia,bacteria were found in this site in all cases. Conversely, theprevalence rate of intestinal metaplasia of the gastric cardiawas similar between patients with and without GERD symp-toms (5/50vs13/83;p 5 NS) as well as between those withand without hiatal hernia (8/36vs10/87;p 5 NS). No caseof dysplasia in the gastric cardia was observed in our study.

In the 39H. pylori–negative patients, histological featuresof carditis were detected in 13 (33.3%) cases, and only onecase of intestinal metaplasia was noted. Finally, the preva-

Table 1. Demographic and Clinical Characteristics of Patients

Male/female 60/73Age, yr (mean6 SD) 556 14Endoscopic finding

Duodenal ulcer 30Gastric ulcer 5Normal 54Hiatal hernia 36Erosive esophagitis 8

GERD symptoms 50

Table 2. Intestinal Metaplasia Distribution in All Gastric Sites

Gastric Site Number of Patients

Antrum 16Antrum 1 body 3Antrum 1 body 1 cardia 1Antrum 1 cardia 9Cardia 8

2338 Morini et al. AJG – Vol. 96, No. 8, 2001

lence of GERD symptoms was not statistically differentbetween patients with and without carditis (6/13vs 13/26;p 5 NS).

DISCUSSION

In recent last decades, epidemiological studies have shownan increased incidence of adenocarcinoma at the gastro-esophageal junction (7). This observation has triggered sev-eral studies aimed at evaluating mucosal damage of thegastric cardia, and a high prevalence (9–23%) of bothchronic inflammation and intestinal metaplasia has beenreported in different populations of patients undergoingupper endoscopy (2–16). To date, the pathogenesis of suchhistological alterations is still debated. At least two differentpathogenic factors have been suggested to play a role—namely, GERD andH. pylori infection (8, 9). The conflict-ing results observed may depend on different methodolog-ical strategies used in the various studies. For example,some studies enrolled patients without a sufficient washoutperiod after proton pump inhibitor therapy (11), whereasothers included in theH. pylori–positive group those pa-tients tested only with serology (2, 4). It is known that bothof these procedures could affect the true prevalence ofinfection (21). Moreover, other studies enrolled only pa-tients who wereH. pylori positive, without an adequatecontrol group (3). In another study, an unexpectedly lowprevalence ofH. pylori infection was observed in symptom-atic patients who underwent upper endoscopy (12).

The present study included patients with and withoutH.pylori infection and/or GERD symptoms to assess the im-portance of both factors in the pathogenesis of mucosaldamage at the gastric cardia. Our data have shown a veryhigh prevalence of chronic inflammation at this site, con-firming previous reports (2–16, 22). Carditis was found to bestrongly associated withH. pylori infection, whereas symp-toms of GERD did not seem to play an important role incarditis development. Moreover, the presence of intestinalmetaplasia in the gastric cardia was fairly high, being foundin about 15% of our patients, in agreement with anotherrecent study (23). As previously reported (2–4, 6), our datasuggest that intestinal metaplasia is strongly associated withH. pylori infection but not with GERD symptoms. Interest-ingly, bacteria were detected in cardia specimens from allbut one patient with intestinal metaplasia. These resultswere not unexpected, as the cardiac glands are greatly sim-ilar to those of the antrum (18), where bacteria are assumedto be the main cause of both gastritis and intestinal meta-plasia (24). Moreover, the degree of inflammatory responseto H. pylori in the cardiac mucosa has previously been foundto be similar to that in the antral mucosa (25). It is note-worthy that, in our study, intestinal metaplasia at the gastriccardia was observed in 26.6% of the 64 patients withH.pylori infection involving this gastric site, a rate very similarto that previously reported in the antrum ofH. pylori–positive patients (26). Intestinal metaplasia in the gastric

antrum is regarded as a risk factor for cancer in the distalstomach (27); it is not clear whether the development ofsuch metaplasia in the gastric cardia has the same patho-genic implications, although it may be possible (28). Indeed,the development of dysplasia in patients with intestinalmetaplasia in the gastric cardia has been reported in tworecent prospective studies (5, 6). Therefore, the strong as-sociation betweenH. pylori infection and intestinal meta-plasia in the gastric cardia that was found in several histo-logical studies (2–4, 6, 12, 23, 29) and that was clearlyconfirmed by our data is intriguing, although the associationbetween serological detection ofH. pylori infection andgastric cardia cancer has been regarded as weak by some(30). Thus, further investigations are warranted in this field.

Another relevant finding of our study is the observationthat the overall prevalence of infection in the gastric cardiais significantly higher than that in the gastric body. More-over, our data showed that the presence ofH. pylori can beconfined to the gastric cardia, as reported in another recentstudy (31). Indeed,H. pylori was detected exclusively in thegastric cardia specimens (but not in other gastric sites) inthree (3.2%) patients, all of whom harbored chronic activeinflammation and intestinal metaplasia. Such features wouldhave been missed if biopsies had not been performed at thegastric cardia, and therefore the infection in these patientswould not have been identified or treated. Moreover, it hasbeen reported that after eradication therapy,H. pylori maypersist in the cardia alone, while disappearing from theantrum and gastric body (31). In these cases, the bacteri-um—erroneously considered to be eradicated—continuesto cause chronic and potentially increasing damage (5, 6).Therefore, cardia biopsy may be useful for obtaining a morecomplete histological sampling (31), especially consideringthat, in our study, biopsies performed routinely in the gastricbody demonstrated a prevalence of both infection and meta-plasia that was significantly lower than that seen in thecardia.

Contrary to a previous report (11), carditis was infre-quently (,35%) found in ourH. pylori–negative patientgroup, and intestinal metaplasia was very rare in this group.Moreover, GERD symptoms did not seem to be involved inthe development of chronic inflammation. The significanceof carditis in this group remains controversial, and otherfactors could be advocated to explain its pathogenesis (9).

In conclusion, our data showed that the gastric cardia isfrequently infected byH. pylori, with the consequent devel-opment of both chronic inflammation and intestinal meta-plasia, whereas GERD does not seem to be involved in thesehistological changes.

Reprint requests and correspondence:Sergio Morini, M.D.,Ospedale Nuovo Regina Margherita, U.O. Gastroenterologia edEndoscopia Digestiva, Via E. Morosini, 30, 00153 Roma, Italy.

Received Sep. 19, 2000; accepted Apr. 2, 2001.

2339AJG – August, 2001 H. pylori and Gastric Cardia Pathology

REFERENCES

1. Riddell RH. The biopsy diagnosis of gastroesophageal refluxdisease, “carditis,” and Barrett’s esophagus, and sequelae oftherapy. Am J Surg Pathol 1996;20(suppl 1):S31–50.

2. Morales GT, Sampliner RE, Bhattacharyya A. Intestinal meta-plasia of gastric cardia. Am J Gastroenterol 1997;92:414–8.

3. Hackelsberger A, Gunther T, Schultze V, et al. Prevalence andpattern ofHelicobacter pylorigastritis in the gastric cardia.Am J Gastroenterol 1997;92:2220–4.

4. Golblum JR, Vicari JJ, Falk GW, et al. Inflammation andintestinal metaplasia of the gastric cardia: The role of gastro-esophageal reflux andH. pylori infection. Gastroenterology1998;114:633–9.

5. Sharma P, Weston AP, Morales T, et al. Relative risk ofdysplasia for patients with intestinal metaplasia in the distaloesophagus and in the gastric cardia. Gut 2000;46:9–13.

6. Morales TG, Camargo E, Bhattacharyya A, et al. Long-termfollow-up of intestinal metaplasia of the gastric cardia. Am JGastroenterol 2000:95:1677–80.

7. Pera M, Cameron AJ, Trastek VF, et al. Increasing incidenceof adenocarcinoma of the esophagus and esophagogastricjunction. Gastroenterology 1993;104:510–3.

8. Genta RM. Giving carditis back to the heart. Gut 1999;43:793–7.

9. O’Connor HJ. Gastro-oesophageal reflux disease,Helicobac-ter pylori and gastric cardia: A tale of two pathologies? DigestLiver Dis 2000;32:573–6.

10. Voutilainen M, Farkkila M, Juhola M, et al. Specialized co-lumnar epithelium of the esophagogastric junction: Prevalenceand association: The Central Finland Endoscopy Study Group.Am J Gastroenterol 1999;94:913–8.

11. Bowrey DJ, Clark GWB, Williams GT. Patterns of gastritis inpatients with gastro-oesophageal reflux disease. Gut 1999;45:789–803.

12. El-Serag HB, Sonnemberg A, Jamal MM, et al. Characteristicsof intestinal metaplasia in the gastric cardia. Am J Gastroen-terol 1999;94:622–7.

13. Hackelsberger A, Gunther T, Schultze V, et al. Intestinalmetaplasia at the gastro-oesophageal junction:Helicobacterpylori gastritis or gastro-oesophageal reflux disease? Gut1998;43:17–21.

14. Trudgill NJ, Suvarna SK, Kapur KC, Riley SA. Intestinalmetaplasia at the squamocolumnar junction in patients attend-ing for diagnostic gastroscopy. Gut 1997;41:585–9.

15. Klauser AG, Schindlbeck NE, Muller-Lissner SA. Symptomsin gastro-oesophageal reflux disease. Lancet 1990;335:205–8.

16. Wu JCY, Sung JJY, Ng EKW, et al. Prevalence and distribu-tion of Helicobacter pyloriin gastroesophageal reflux disease:A study form the east. Am J Gastroenterol 1999;94:1790–4.

17. Dixon FM, Genta RM, Yardley JH, Correa P. Classificationand grading of gastritis. The updated Sydney System. Am JSurg Pathol 1996;20:1161–81.

18. Owen DA. Normal histology of the stomach. Am J SurgPathol 1986;10:48–61.

19. Mittal RK, Lange RC, McCallum RW. Identification andmechanism of delayed esophageal acid clearance in subjectswith hiatus hernia. Gastroenterology 1987;92:130–5.

20. Savary M, Miller G. The oesophagus. Handbook and atlas ofendoscopy. Solothurn: Verlag Gassmann, 1978:119–205.

21. EuropeanHelicobacter pyloriStudy Group. Technical annex:Tests used to assessHelicobacter pyloriinfection. Gut 1997;41(suppl 2):S10–8.

22. Peck-Radosavljevic M, Puspok A, Potzi R, Oberhuber G.Histological findings after routine biospy at the gastro-oesoph-ageal junction. Eur J Gastroenterol Hepatol 1999;11:1265–70.

23. Pieramico O, Zanetti MV. Relationship between intestinalmetaplasia of the gastro-oesophageal junction,Helicobacterpylori infection and gastro-oesophageal reflux disease: A pro-spective study. Digest Liver Dis 2000;32:567–72.

24. Fontham ETH, Ruiz B, Perez A, et al. Determinants ofHel-icobacter pyloriinfection and chronic gastritis. Am J Gastro-enterol 1995;90:1094–101.

25. Genta RM, Huberman RM, Graham DY. The gastric cardia inHelicobacter pyloriinfection. Hum Pathol 1994;25:915–9.

26. Craanen ME, Dekker W, Block P, et al. Intestinal metaplasiaandHelicobacter pylori: An endoscopic bioptic study of thegastric antrum. Gut 1992;33:16–20.

27. Correa P. Human gastric carcinogenesis: A multistep andmultifactorial process–first American Cancer Society awardlecture on cancer epidemiology and prevention. Cancer Res1992;52:6735–40.

28. MacDonald WC, MacDonald JB. Adenocarcinoma of theesophagus and/or gastric cardia. Cancer 1987;60:1094–8.

29. Hirota WK, Loughney TM, Lazas DJ, et al. Specialized in-testinal metaplasia, dysplasia, and cancer of the esophagus andesophagogastric junction: Prevalence and clinical data. Gas-troenterology 1999;116:277–85.

30. Huang JQ, Sridhar S, Chen Y, Hunt RH. Meta-analysis of therelationship betweenHelicobacter pyloriseropositivity andgastric cancer. Gastroenterology 1998;114:1169–79.

31. Sheu BS, Lin XZ, Yang HB, Chien CH. Cardiac biopsy ofstomach may improve the detection ofH. pylori after dualtherapy. Hepatogastroenterol 1999;46:543–8.

2340 Morini et al. AJG – Vol. 96, No. 8, 2001