- Email: [email protected]
Carditis is related to Helicobacter pylori infection in dyspeptic children and adolescents
Digestive and Liver Disease 39 (2007) 117–121
Alimentary Tract
Carditis is related to Helicobacter pylori infection in dyspeptic children and adolescents A.P. Carelli a , F.R.S. Patr´ıcio b , E. Kawakami c,∗ a c
Division of Pediatric Gastroenterology, Universidade Federal de S˜ao Paulo/Escola Paulista de Medicina, S˜ao Paulo/SP, Brazil b Department of Pathology, Universidade Federal de S˜ ao Paulo/Escola Paulista de Medicina, S˜ao Paulo/SP, Brazil Division of Pediatric Gastroenterology, Universidade Federal de S˜ao Paulo/Escola Paulista de Medicina, S˜ao Paulo/SP, Brazil Received 9 May 2006; accepted 24 October 2006 Available online 28 December 2006
Abstract Background. Etiology of gastric cardia inflammation is still controversial. Aims. To evaluate the association between carditis and Helicobacter pylori infection and the correlation among inflammatory changes observed in biopsies taken from cardia, corpus, and antrum in a well-defined group of patients. Patients. The mean age of 45 dyspeptic patients was 10.4 years (range 5.1–17.0 years); gender F/M rate: 1.6/1. Methods. A total of 450 specimens from esophagus (2), cardia (2), corpus (3), and antrum (4) were collected for biopsy. The presence of H. pylori was assessed by histology and a rapid urease test. The types of glandular epithelium of cardia found in specimens were identified and both inflammatory changes and H. pylori density were graded. Results. Carditis was present in specimens of 30/45 (66.7%) of the patients. Presence of H. pylori in specimens was detected in the antrum (26/45; 57.8%), in the corpus (19/45; 42.2%), and in the cardia (14/45; 31.1%). There was a strong association between carditis and presence of H. pylori infection (OR = 27.08) by multivariate analysis. The scores for inflammation and activity in the cardia, corpus and antrum have shown a relationship except for both cardia and antrum H. pylori density and corpus and cardia activity. The intensity of gastritis and degree of colonization with H. pylori were significantly higher in the antrum than in both the corpus and the cardia. Pangastritis was highly associated to H. pylori infection in 22/25 (88%) of the patients. Conclusions. 1. Carditis is associated to H. pylori infection in children with symptoms of dyspepsia; 2. The degrees of gastritis found at the cardia were correlated to those at the antrum and body except for both cardia and antrum H. pylori density and corpus and cardia activity. © 2006 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved. Keywords: Carditis; Dyspepsia; Gastric cardia; Helicobacter pylori
1. Introduction The term “carditis” has been widely adopted for inflammatory changes just below the squamocolumnar junction. There is no consensus on the length and histologic features of the most proximal portion of the stomach. Columnar mucosa containing pure mucous or a mixture of mucous and oxyntic glands was described [1]. However, pure oxyntic mucosa consisting only of chief and parietal cells was also described [2]. ∗ Corresponding author at: Rua Pedro de Toledo 441, 04040-032 S˜ ao Paulo/SP, Brazil. Tel.: +55 11 55795834; fax: +55 11 55795834. E-mail address: [email protected] (E. Kawakami).
The pathogenesis and clinical significance of chronic carditis are not yet clear. Some authors have reported a strong association between carditis and H. pylori infection [3,4]. In contrast, other authors argue that gastroesophageal reflux disease (GERD) [5–7] is the major mechanism of chronic injury of the cardia while some others state that carditis is related to both H. pylori and GERD [8–10]. Cardia intestinal metaplasia is not a marker of GERD but is related to H. pylori infection [1,11]. Genta et al. [3] have shown H. pylori in the cardia of 40/42 (95%) infected patients, and gastritis with similar intensity in both the cardia and the antrum was observed, although with a higher intensity than in the corpus.
1590-8658/$30 © 2006 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dld.2006.10.012
118
A.P. Carelli et al. / Digestive and Liver Disease 39 (2007) 117–121
The morphology of the gastric cardia and the pathogenesis of carditis have been poorly studied in pediatric patients. Glickman et al. [12], in a retrospective study, have concluded that carditis could be the result of both GERD and H. pylori infection. Borrelli et al. [13] have studied children with symptoms of acid peptic disease and showed that both carditis is more severe in children with H. pylori infection than in its absence (GERD) and the gastric cardia is a highly sensitive site for detection of H. pylori infection [13]. The purpose of this study was to evaluate the association between carditis and H. pylori infection, and to determine possible correlations among the inflammatory degrees in biopsy specimens from cardia, corpus and antrum in a welldefined group of patients.
2. Patients and methods 2.1. Patient population A total of 45 (F/M 1.6/1) patients suspected of acid peptic disease, referred to the Endoscopy Unit for 15 consecutive months, were included in the investigation. Barrett’s esophagus, presence of other digestive or extradigestive disease, previous H. pylori eradication, antibiotic and antisecretory treatment for at least 1 month before the study were exclusion criteria. Their ages ranged from 5.1 to 17 years (mean ± S.D.: 10.9 ± 2.9 years, median: 10.4 years), being 20 (44.4%) from 5 to 9 years, 19 (42.2%) from 10 to 14 years and 6 (13.3%) from 15 to 17 years. Twenty-four (53.3%) patients had a white skin phenotype. An informed clinical and demographic questionnaire was filled in by either the parents or the adolescent patient. The study protocol was previously approved by the local Ethics Committee. 2.2. Endoscopy and biopsy Upper gastrointestinal endoscopy was performed with an Olympus GIF-V videoendoscope (biopsy channel diameter: 2.8 mm) after intravenous sedation. Biopsy specimens were obtained from predetermined sites: antrum (four samples within 2–3 cm of the pylorus; two for a rapid urease test and two for histology), corpus (two samples about 8 cm from the cardia), distal esophagus (two samples 3–5 cm proximal to the Z-line), and gastric cardia (two samples) either in Uturn or in antegradely viewing position. The cardia specimens were taken just below the esophagogastric junction or near the Z-line. The esophagogastric junction was defined endoscopically as the proximal limit of the gastric folds or coincident with the distal limit of squamous-lined tubular esophagus as previously suggested [12]. All specimens were carefully opened and fixed in buffered formalin. Serial sections were stained with hematoxylin and eosin, and then with modified Giemsa stain to detect H. pylori. All specimens (450
biopsy specimens) were analyzed by the same pathologist, who was blinded to the clinical information. If a disparity between scores was observed in biopsies from the same anatomic level, the score of the most abnormal biopsy was recorded. 2.3. Pathology The glandular epithelium were classified as mucous, oxyntic and mucous-oxyntic (mixed) types. The overall degree of inflammation in the columnar mucosa was scored according to the modified Sydney system [14]. Polymorphonuclear neutrophil activity (simply recorded as “activity”), the degree of H. pylori colonization (recorded as H. pylori density) and the intensity of chronic inflammation (mononuclear cells), were graded as none (0), mild (1), moderate (2) or marked (3). The esophageal mucosa was also evaluated [15]. Lymphocytes and plasma cells were each quantified separately as follows: +: ≤10/hpf, ++: >10/hpf with the use of an eyepiece micrometer. Lymphoid aggregates were scored individually as either present or absent. H. pylori were identified in Giemsa-stained sections that provide the visualization of curved bacteria in the gastric mucous layer. The rapid urease test was performed with two fragments obtained from the antrum with a non-commercial solution [16]. Evidence of H. pylori in Giemsa-stained section was considered a diagnostic of H. pylori infection with or without a positive rapid urease test. 2.4. Statistical analysis Analysis of data was performed by using the SPSS software package. Comparison of clinical, endoscopic, and histologic variables between different groups was performed by using the χ2 analysis or Fisher’s exact test. Comparison of numerical variables was performed by using the Mann–Whitney U-test and Student’s t-test. Univariate and multivariate logistic regression analysis were applied. Any p-value <0.05 was considered statistically significant. The inflammatory changes from each gastric region were correlated with the Spearman rank correlation coefficient.
3. Results 3.1. Clinical and endoscopic features Epigastric pain duration was referred to range from 15 days (one patient with upper digestive bleeding) to 5 years (median 12 months). Both dyspeptic symptoms and endoscopic findings are shown in Table 1. Diagnosis of upper gastrointestinal endoscopy was normal in 17/45 (37.8%) and abnormal in 28/45 (62.2%) of the patients. Nodular gastritis was the most common endoscopic abnormality in more than half of the patients.
A.P. Carelli et al. / Digestive and Liver Disease 39 (2007) 117–121 Table 1 Clinical and endoscopic features of 45 patients
119
3.4. Cardia histology
Features
n
%
Clinical Vomiting Clocking Overloading Food related Heartburn Upper digestive bleeding
23 30 18 16 13 05
(51.1) (65.6) (40.0) (35.5) (28.9) (11.1)
Endoscopic Normal Abnormal Non-erosive esophagitis Non-erosive esophagitis + antrum nodular gastritis Erosive esophagitis + gastritis + duodenitis Antrum nodular gastritis Antrum nodular gastritis + erosive duodenitis Antrum nodular gastritis + non erosive duodenitis Erosive duodenitis
17 28 04 02 02 16 01 02 01
(37.7) (62.2) (14.2) (7.4) (7.4) (57.1) (3.5) (7.4) (3.5)
Presence of mucous cells in the surface epithelium was evidenced in all patients. Multilayered epithelium and goblet cells were absent. Regarding the glandular epithelium, 24 (53.3%), 14 (31.1%), and 7 (15.6%) presented oxyntic, mucous and mucous/oxyntic epithelium, respectively. In two specimens with the Z-line included, the glandular epithelium was oxyntic. Carditis was observed in 30/45 (66.7%) patients. Lymphocytes and plasma cells were (+) in 60% and (++) in 40% of the patients; lymphoid aggregates were observed in 51.1% of cardia specimens. 3.5. Chronic gastritis Inflammatory changes of cardia, corpus and antrum were more frequent and more severe in infected patients (Table 2). Pangastritis (cardia, corpus, and antrum inflammation) was observed in about half of patients (25/45: 55.5%), mainly H. pylori-infected patients (22/26: 84.6%) (p = 0.0000044). A correlation was found between the scores of histological variables for each gastric region (r = 0.50–0.69), except for antrum and cardia density (r = 0.43) and corpus and cardia activity (r = 0.47). The best correlation was found for cardia and antrum inflammation (r = 0.69) and a poor correlation was obtained for antrum and corpus density (r = 0.50). The intensity of gastritis was significantly higher in the antrum than in both corpus and cardia (p < 0.001). Intestinal metaplasia and gastric atrophy were not found in these patients. No association was found between carditis and age, gender, race, local pain intensity and frequency, as well as associated symptoms (dysphagia, heartburn, vomiting, regurgitation, clocking, upper digestive bleeding) in the carditis group (p > 0.05). The statistically significant variables related to carditis are summarized in Table 3, but only H. pylori infection was found to be related to carditis in a multivariate analysis (OR 27.08, 95% CI OR 4.78–153.53, p < 0.001).
3.2. Pathology Normal histologic features of esophagus, stomach, and duodenum mucosa simultaneously were found in the minority (17.4%) of the patients. Twenty-six (57.7%) patients had H. pylori in the antrum, 19 (42.2%) in the corpus, and 14 (31.1%) in the gastric cardia. The rapid urease test was shown to be positive in 23 (51.1%) and negative in 22 (48.9%) of the patients. 3.3. Esophagus histology Diagnosis of histologic analysis were normal in 34/45 (75.5%) and abnormal in 11/45 (24.5) of the patients. The histologic abnormalities were basal zone hyperplasia (4), basal zone hyperplasia and elongated stromal papillae (1), elongated stromal papillae (1) alone, polymorphonuclear cells in the epithelium (4) and polymorphs with epithelial defects (1). Table 2 Inflammatory changes in cardia, corpus, and antrum according to H. pylori status score
H. pylori +
H. pylori −
n (%)
n (%)
0
1
2
3
0
Cardia Inflammation Activity Density
3 (12) 8 (31) 12 (46)
10 (38) 11 (42) 6 (23)
6 (23) 7 (27) 5 (19)
7 (27) 0 (0) 3(12)
12 (63) 15 (79) 19 (100)
Corpus Inflammation Activity Density
1 (4) 6 (23) 7 (27)
12 (46) 15 (58) 10 (39)
11 (42) 3 (11) 5 (19)
2 (8) 2 (8) 4 (15)
Antrum Inflammation Activity Density
0 (0) 1 (4) 0 (0)
2 (8) 10 (38.5) 6 (23)
8 (31) 10 (38.5) 9 (35)
16 (61) 5 (19) 11 (42)
p
1
2
3
6(32) 4 (21) 0 (0)
1 (5) 0 (0) 0 (0)
0 (0) 0 (0) 0 (0)
<0.001 <0.002 <0.001
14 (74) 16 (84) 19 (100)
4 (21) 2 (11) 0 (0)
0 (0) 0 (0) 0 (0)
1 (5) 1 (5) 0 (0)
<0.001 <0.001 <0.001
8 (42) 18 (95) 19 (100)
10 (53) 1 (5) 0 (0)
1 (5) 0 (0) 0 (0)
0 (0) 0 (0) 0 (0)
<0.001 <0.001 <0.001
120
A.P. Carelli et al. / Digestive and Liver Disease 39 (2007) 117–121
Table 3 Logistic univariate regression analysis for carditis Variables
Category (n)
Esophagitis
A (35) P (11)
0.254
A (8) P (38)
0.003
A (18) P (28)
Antrum inflammation Antrum activity Antrum density Corpus inflammation Corpus activity Corpus density Cardia lymphoid aggregates
p
OR 1 2.66
95% CI OR 0.50–14.25
1 30
3.16–284.92
<0.001
1 27.08
3.55–78.32
A (19) P (27)
<0.001
1 27.08
4.78–153.53
A (15) P (31)
<0.001
1 18.56
3.93–87.73
A (22) P (24)
0.005
1 8.4
1.93–36.63
A (26) P (20)
0.009
1 9
1.73–46.90
A (31) P (15)
1 0.003
4.35
1.13–16.85
A: absent P: present.
Carditis was present in all infected children except for 3 (11.5%) while only 7/19 (36.8%) of the non-infected children had carditis.
4. Discussion This study shows that carditis is associated with H. pylori infection. Several reports support our results on both adults [3,4,17,18] and children [13]. Borrelli et al. [13] have observed carditis in 100% of 22 infected children and in 68% of the GERD group (p < 0.001). Moreover, the degrees of gastritis at the cardia are correlated with those at the antrum and body, except for both antrum and cardia density and corpus and cardia activity. These findings suggest that cardia inflammation may occur due to H. pylori-induced pangastritis. Contrarily to other report [3] we have observed gastritis in the cardia with a lower intensity. No similar studies were found in the literature comparing the inflammatory changes in these three gastric areas of children. It is likely that these findings are quite different in such patients. We have observed higher both prevalence and H. pylori density in the antrum (57.7%) than in the corpus (42.2%) and the cardia (31.1%), differently from other reports that describe colonizations that are similar in the antrum and in the cardia [3,4] or greater in the cardia [13]. This observation suggests differences in either the migration pattern from antrum to cardia or the focal bacteria distribution. This consecutive study had well defined both inclusion and exclusion criteria which reduced the selection and sample bias. Studies on gastric cardia in children are rare in the literature: two studies were on autopsy [2,19], one study was retrospective, and only one study was prospective [13].
Methodology errors may happen in autopsy studies due to the autolysis of gastric mucosa that occurs shortly after death. In the same way, endoscopic biopsy specimens present the inconvenience of the cardia anatomic landmarks, which are not always easy to establish and frequently do not correspond to the gastroesophageal junction [20]. Regarding carditis etiology, the disagreement between such findings could reflect the difficulty in standardizing the location where the mucosal specimens should be obtained [1]. We have observed that histologically defined gastric cardiac mucosa is not always present in the endoscopically defined cardiac region. However, the types of glandular epithelium found in the normal gastric cardia are controversial. We have observed pure oxyntic-type glands in 24 (53.3%) patients, including two patients whose biopsies showed a squamocolumnar epithelium. It is important to know how the anatomic extension of cardia is related to the patient’s size; as in children the cardia is shorter, good performances of biopsy in such patients are not frequent. Glickman et al. [12] have observed mucous glandular epithelium in 81% of the patients, mucous/oxyntic-type in 19% of them, and none presented exclusively oxyntic glandular epithelium. The same was observed by Wieczorek et ¨ al. [21]. In contrast, Oberg et al. [5] have found that only 88/334 (26%) patients showed mucous glandular epithelium in biopsy specimens from the esophageal gastric junction. In a retrospective study with 72 autopsies, including adults and children, the investigators have found oxyntic glandular epithelium in 29% of the cases. Another autopsy study performed in 30 pediatric patients report mucous-type epithelium in all cases, supporting the concept that the gastric cardia is present from birth as a normal structure [22]. Cardia mucosal features (epithelial cells, glandular epithelium, amount of lymphocytes and plasma cells in the mucosa, and presence of lymphoid aggregates) were graded according to two retrospective studies performed by the same group [12,21], one of them with cardia morphology analysis in children [12]. We have adopted this criterion due to a detailed description of the histologic features of the cardia in pediatric patients, which provide a richness of details to be analyzed, although this criterion is complex and extensive. A histologic review of the cardia epithelial cells in this study has shown the mucous type in the specimens of all patients. Multilayered epithelium was not observed in any of them. Wieczoreck et al. [21] have analyzed cardia histologic features in two patient groups (H. pylori-infected and GERD patients) and multilayered epithelium was found only in patients with carditis related to GERD (27%). The differences in both type of information (autopsy, retrospective and prospective studies) and the inclusion criteria, together with a lack of anatomic pattern in the cardia region, render the comparison of studies a hard task and contribute to the divergences found in the literature on the morphology of the cardia and the etiology of carditis. In conclusion, our study shows that carditis is strongly related to H. pylori infection and pangastritis is a frequent finding in infected patients.
A.P. Carelli et al. / Digestive and Liver Disease 39 (2007) 117–121
Furthermore, we have found a correlation among the intensities of inflammatory changes observed in cardia, corpus and antrum except for antrum and cardia density and corpus and cardia activity. Further studies on the anatomic region of cardia are needed. Practice points • Carditis is strongly related to H. pylori infection and pangastritis was a frequent finding in infected children and adolescents with dyspepsia. • The intensity of gastritis (inflammation, activity) and degree of colonization was significantly higher in the antrum than in the corpus and cardia.
[6]
[7]
[8]
[9]
[10] [11]
[12]
Research agenda • The lack of the anatomic and morphology pattern in the cardiac region render comparison between the studies of carditis difficult. • Further studies about the type of glandular epithelium in the cardia and the extension of this region are needed.
[13]
[14]
[15]
[16]
Conflict of interest statement None declared.
[17]
References [18] [1] Goldblum RJ, Richter JE, Vaezi M, Falk GW, Rice TW, Peek RM. Helicobacter pylori infection, not gastroesophageal reflux, is the major cause of inflammation and intestinal metaplasia of gastric cardiac mucosa. Am J Gastroenterol 2002;97:302–11. [2] Chandrasoma PT, Der R, Ma Y, Dalton P, Taira M. Histology of the gastroesophageal junction: an autopsy study. Am J Surg Pathol 2000;24:402–9. [3] Genta RM, Huberman RM, Graham DY. The gastric cardia in Helicobacter pylori infection. Human Pathol 1994;25:915–9. [4] Hackelsberger A, G¨unther T, Schultze V, Labez J, Roessner A, Malferttheiner P. Prevalence and pattern of Helicobacter pylori gastritis in the gastric cardia. Am J Gastroenterol 1997;92:2220–4. ¨ [5] Oberg S, Peters JH, DeMeester TR, Chandrosoma PT, Hagen JA, Ireland AP, et al. Inflammation and specialized intestinal metaplasia of
[19]
[20] [21]
[22]
121
cardiac mucosa is a manifestation of gastroesophageal reflux disease. Ann Surg 1997;226:522–32. Csendes A, Smok G, Burdiles P, Sagastume H, Rojas J, Puente G, et al. ‘Carditis’: an objective histological marker for pathologic gastroesophageal reflux disease. Dis Esophagus 1998;11:101–5. Der R, Tsao-Dei DD, Demeester T, Peters J, Groshen F, Lord RVN, et al. Carditis a manifestation of gastroesophageal reflux disease. Am J Surg Pathol 2001:25245–52. Voutilainem M, F¨arkkil¨an M, Mecklin JP, Juhola M, Sipponem P. Chronic inflammation at the gastroesophageal junction (carditis) appears to be a specific finding related to Helicobacter pylori infection and gastroesophageal reflux disease. Am J Gastroenterol 1999;94:3175–80. Lembo T, Ipponti AF, Ramers C, Weinstein WM. Inflammation of the gastro-oesophageal junction (carditis) in patients with symptomatic gastro-oesophageal reflux disease: a prospective study. Gut 1999;45:484–8. Wolf C, Sildenrijk CA, Timmer R, Breumelhof R, Smout AJPM. Does carditis have two different etiologies? Dig Dis Sci 2001;46:2424–32. Pieramico O, Zanetti MV. Relationship between intestinal metaplasia of the gastro-oesophageal junction, Helicobacter pylori infection and gastro-oesophageal reflux disease: a prospective study. Digest Liver Dis 2000;32:567–72. Glickman NJ, Fox V, Antonioli DA, Wang HH, Odze RD. Morphology of the cardia and significance of carditis in pediatric patients. Am J Surg Pathol 2002;26:1032–9. Borrelli O, Hassall E, D’Armiento F, Bosco S, Mancini V, DiNardo G, et al. Inflammation of the gastric cardia in children with symptoms of acid peptic disease. J Pediatr 2003;143:520–4. Dixon MF, Genta RM, Yardley JH, Correa P. Classification and grading of gastritis. The update Sydney Sistem. Am J Surg Pathol 1996;20:1161–81. Vandenplas Y. Reflux esophagitis in infants and children: a report from working group on gastro-oesophageal reflux disease of the European Society of Pediatric Gastroenterology and Nutrition. J Pediatr Gastroenterol Nutr 1994;18:413–22. Ogata SK, Kawakami E, Patricio FRS, Pedroso MZ, Santos AM. Evaluation of invasive and non invasive methods for the diagnosis of Helicobacter pylori infection in symptomatic children and adolescentes. S˜ao Paulo Med J/Rev Paul Med 2001;19:48–53. Morini S, Zullo A, Hassan C, Lorenzetti R, Stella F, Martini MT. Gastric cardia inflammation: role of Helicobater pylori infection and symptoms of gastroesophageal reflux disease. Am J Gastroenterol 2001;96:2337–40. Mcnamara D, Buckley M, Crotty P, Hall W, O’Sullivan M, O’Morain. Carditis: all Helicobacter pylori or is there a role for gastro-oesophageal reflux? Scand J Gastroenterol 2002;37:772–7. 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. Spechler JS. The role of gastric carditis in metaplasia and neoplasia at the gastroesophageal Junction. Gastroenterology 1999;117:218–28. Wieczorek J, Wang HH, Antonioli DA, Glickman JN. Pathologic features of reflux and Helicobacter pylori-associated carditis. Am J Surg Pathol 2003;27:960–8. Kilgore SP, Ormsby AH, Gramlich TL, Rice TW, Richter JE, Falk GW, et al. The gastric cardia: fact or fiction? Am J Gastroenterol 2000;95:921–4.
Alimentary Tract
Carditis is related to Helicobacter pylori infection in dyspeptic children and adolescents A.P. Carelli a , F.R.S. Patr´ıcio b , E. Kawakami c,∗ a c
Division of Pediatric Gastroenterology, Universidade Federal de S˜ao Paulo/Escola Paulista de Medicina, S˜ao Paulo/SP, Brazil b Department of Pathology, Universidade Federal de S˜ ao Paulo/Escola Paulista de Medicina, S˜ao Paulo/SP, Brazil Division of Pediatric Gastroenterology, Universidade Federal de S˜ao Paulo/Escola Paulista de Medicina, S˜ao Paulo/SP, Brazil Received 9 May 2006; accepted 24 October 2006 Available online 28 December 2006
Abstract Background. Etiology of gastric cardia inflammation is still controversial. Aims. To evaluate the association between carditis and Helicobacter pylori infection and the correlation among inflammatory changes observed in biopsies taken from cardia, corpus, and antrum in a well-defined group of patients. Patients. The mean age of 45 dyspeptic patients was 10.4 years (range 5.1–17.0 years); gender F/M rate: 1.6/1. Methods. A total of 450 specimens from esophagus (2), cardia (2), corpus (3), and antrum (4) were collected for biopsy. The presence of H. pylori was assessed by histology and a rapid urease test. The types of glandular epithelium of cardia found in specimens were identified and both inflammatory changes and H. pylori density were graded. Results. Carditis was present in specimens of 30/45 (66.7%) of the patients. Presence of H. pylori in specimens was detected in the antrum (26/45; 57.8%), in the corpus (19/45; 42.2%), and in the cardia (14/45; 31.1%). There was a strong association between carditis and presence of H. pylori infection (OR = 27.08) by multivariate analysis. The scores for inflammation and activity in the cardia, corpus and antrum have shown a relationship except for both cardia and antrum H. pylori density and corpus and cardia activity. The intensity of gastritis and degree of colonization with H. pylori were significantly higher in the antrum than in both the corpus and the cardia. Pangastritis was highly associated to H. pylori infection in 22/25 (88%) of the patients. Conclusions. 1. Carditis is associated to H. pylori infection in children with symptoms of dyspepsia; 2. The degrees of gastritis found at the cardia were correlated to those at the antrum and body except for both cardia and antrum H. pylori density and corpus and cardia activity. © 2006 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved. Keywords: Carditis; Dyspepsia; Gastric cardia; Helicobacter pylori
1. Introduction The term “carditis” has been widely adopted for inflammatory changes just below the squamocolumnar junction. There is no consensus on the length and histologic features of the most proximal portion of the stomach. Columnar mucosa containing pure mucous or a mixture of mucous and oxyntic glands was described [1]. However, pure oxyntic mucosa consisting only of chief and parietal cells was also described [2]. ∗ Corresponding author at: Rua Pedro de Toledo 441, 04040-032 S˜ ao Paulo/SP, Brazil. Tel.: +55 11 55795834; fax: +55 11 55795834. E-mail address: [email protected] (E. Kawakami).
The pathogenesis and clinical significance of chronic carditis are not yet clear. Some authors have reported a strong association between carditis and H. pylori infection [3,4]. In contrast, other authors argue that gastroesophageal reflux disease (GERD) [5–7] is the major mechanism of chronic injury of the cardia while some others state that carditis is related to both H. pylori and GERD [8–10]. Cardia intestinal metaplasia is not a marker of GERD but is related to H. pylori infection [1,11]. Genta et al. [3] have shown H. pylori in the cardia of 40/42 (95%) infected patients, and gastritis with similar intensity in both the cardia and the antrum was observed, although with a higher intensity than in the corpus.
1590-8658/$30 © 2006 Editrice Gastroenterologica Italiana S.r.l. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.dld.2006.10.012
118
A.P. Carelli et al. / Digestive and Liver Disease 39 (2007) 117–121
The morphology of the gastric cardia and the pathogenesis of carditis have been poorly studied in pediatric patients. Glickman et al. [12], in a retrospective study, have concluded that carditis could be the result of both GERD and H. pylori infection. Borrelli et al. [13] have studied children with symptoms of acid peptic disease and showed that both carditis is more severe in children with H. pylori infection than in its absence (GERD) and the gastric cardia is a highly sensitive site for detection of H. pylori infection [13]. The purpose of this study was to evaluate the association between carditis and H. pylori infection, and to determine possible correlations among the inflammatory degrees in biopsy specimens from cardia, corpus and antrum in a welldefined group of patients.
2. Patients and methods 2.1. Patient population A total of 45 (F/M 1.6/1) patients suspected of acid peptic disease, referred to the Endoscopy Unit for 15 consecutive months, were included in the investigation. Barrett’s esophagus, presence of other digestive or extradigestive disease, previous H. pylori eradication, antibiotic and antisecretory treatment for at least 1 month before the study were exclusion criteria. Their ages ranged from 5.1 to 17 years (mean ± S.D.: 10.9 ± 2.9 years, median: 10.4 years), being 20 (44.4%) from 5 to 9 years, 19 (42.2%) from 10 to 14 years and 6 (13.3%) from 15 to 17 years. Twenty-four (53.3%) patients had a white skin phenotype. An informed clinical and demographic questionnaire was filled in by either the parents or the adolescent patient. The study protocol was previously approved by the local Ethics Committee. 2.2. Endoscopy and biopsy Upper gastrointestinal endoscopy was performed with an Olympus GIF-V videoendoscope (biopsy channel diameter: 2.8 mm) after intravenous sedation. Biopsy specimens were obtained from predetermined sites: antrum (four samples within 2–3 cm of the pylorus; two for a rapid urease test and two for histology), corpus (two samples about 8 cm from the cardia), distal esophagus (two samples 3–5 cm proximal to the Z-line), and gastric cardia (two samples) either in Uturn or in antegradely viewing position. The cardia specimens were taken just below the esophagogastric junction or near the Z-line. The esophagogastric junction was defined endoscopically as the proximal limit of the gastric folds or coincident with the distal limit of squamous-lined tubular esophagus as previously suggested [12]. All specimens were carefully opened and fixed in buffered formalin. Serial sections were stained with hematoxylin and eosin, and then with modified Giemsa stain to detect H. pylori. All specimens (450
biopsy specimens) were analyzed by the same pathologist, who was blinded to the clinical information. If a disparity between scores was observed in biopsies from the same anatomic level, the score of the most abnormal biopsy was recorded. 2.3. Pathology The glandular epithelium were classified as mucous, oxyntic and mucous-oxyntic (mixed) types. The overall degree of inflammation in the columnar mucosa was scored according to the modified Sydney system [14]. Polymorphonuclear neutrophil activity (simply recorded as “activity”), the degree of H. pylori colonization (recorded as H. pylori density) and the intensity of chronic inflammation (mononuclear cells), were graded as none (0), mild (1), moderate (2) or marked (3). The esophageal mucosa was also evaluated [15]. Lymphocytes and plasma cells were each quantified separately as follows: +: ≤10/hpf, ++: >10/hpf with the use of an eyepiece micrometer. Lymphoid aggregates were scored individually as either present or absent. H. pylori were identified in Giemsa-stained sections that provide the visualization of curved bacteria in the gastric mucous layer. The rapid urease test was performed with two fragments obtained from the antrum with a non-commercial solution [16]. Evidence of H. pylori in Giemsa-stained section was considered a diagnostic of H. pylori infection with or without a positive rapid urease test. 2.4. Statistical analysis Analysis of data was performed by using the SPSS software package. Comparison of clinical, endoscopic, and histologic variables between different groups was performed by using the χ2 analysis or Fisher’s exact test. Comparison of numerical variables was performed by using the Mann–Whitney U-test and Student’s t-test. Univariate and multivariate logistic regression analysis were applied. Any p-value <0.05 was considered statistically significant. The inflammatory changes from each gastric region were correlated with the Spearman rank correlation coefficient.
3. Results 3.1. Clinical and endoscopic features Epigastric pain duration was referred to range from 15 days (one patient with upper digestive bleeding) to 5 years (median 12 months). Both dyspeptic symptoms and endoscopic findings are shown in Table 1. Diagnosis of upper gastrointestinal endoscopy was normal in 17/45 (37.8%) and abnormal in 28/45 (62.2%) of the patients. Nodular gastritis was the most common endoscopic abnormality in more than half of the patients.
A.P. Carelli et al. / Digestive and Liver Disease 39 (2007) 117–121 Table 1 Clinical and endoscopic features of 45 patients
119
3.4. Cardia histology
Features
n
%
Clinical Vomiting Clocking Overloading Food related Heartburn Upper digestive bleeding
23 30 18 16 13 05
(51.1) (65.6) (40.0) (35.5) (28.9) (11.1)
Endoscopic Normal Abnormal Non-erosive esophagitis Non-erosive esophagitis + antrum nodular gastritis Erosive esophagitis + gastritis + duodenitis Antrum nodular gastritis Antrum nodular gastritis + erosive duodenitis Antrum nodular gastritis + non erosive duodenitis Erosive duodenitis
17 28 04 02 02 16 01 02 01
(37.7) (62.2) (14.2) (7.4) (7.4) (57.1) (3.5) (7.4) (3.5)
Presence of mucous cells in the surface epithelium was evidenced in all patients. Multilayered epithelium and goblet cells were absent. Regarding the glandular epithelium, 24 (53.3%), 14 (31.1%), and 7 (15.6%) presented oxyntic, mucous and mucous/oxyntic epithelium, respectively. In two specimens with the Z-line included, the glandular epithelium was oxyntic. Carditis was observed in 30/45 (66.7%) patients. Lymphocytes and plasma cells were (+) in 60% and (++) in 40% of the patients; lymphoid aggregates were observed in 51.1% of cardia specimens. 3.5. Chronic gastritis Inflammatory changes of cardia, corpus and antrum were more frequent and more severe in infected patients (Table 2). Pangastritis (cardia, corpus, and antrum inflammation) was observed in about half of patients (25/45: 55.5%), mainly H. pylori-infected patients (22/26: 84.6%) (p = 0.0000044). A correlation was found between the scores of histological variables for each gastric region (r = 0.50–0.69), except for antrum and cardia density (r = 0.43) and corpus and cardia activity (r = 0.47). The best correlation was found for cardia and antrum inflammation (r = 0.69) and a poor correlation was obtained for antrum and corpus density (r = 0.50). The intensity of gastritis was significantly higher in the antrum than in both corpus and cardia (p < 0.001). Intestinal metaplasia and gastric atrophy were not found in these patients. No association was found between carditis and age, gender, race, local pain intensity and frequency, as well as associated symptoms (dysphagia, heartburn, vomiting, regurgitation, clocking, upper digestive bleeding) in the carditis group (p > 0.05). The statistically significant variables related to carditis are summarized in Table 3, but only H. pylori infection was found to be related to carditis in a multivariate analysis (OR 27.08, 95% CI OR 4.78–153.53, p < 0.001).
3.2. Pathology Normal histologic features of esophagus, stomach, and duodenum mucosa simultaneously were found in the minority (17.4%) of the patients. Twenty-six (57.7%) patients had H. pylori in the antrum, 19 (42.2%) in the corpus, and 14 (31.1%) in the gastric cardia. The rapid urease test was shown to be positive in 23 (51.1%) and negative in 22 (48.9%) of the patients. 3.3. Esophagus histology Diagnosis of histologic analysis were normal in 34/45 (75.5%) and abnormal in 11/45 (24.5) of the patients. The histologic abnormalities were basal zone hyperplasia (4), basal zone hyperplasia and elongated stromal papillae (1), elongated stromal papillae (1) alone, polymorphonuclear cells in the epithelium (4) and polymorphs with epithelial defects (1). Table 2 Inflammatory changes in cardia, corpus, and antrum according to H. pylori status score
H. pylori +
H. pylori −
n (%)
n (%)
0
1
2
3
0
Cardia Inflammation Activity Density
3 (12) 8 (31) 12 (46)
10 (38) 11 (42) 6 (23)
6 (23) 7 (27) 5 (19)
7 (27) 0 (0) 3(12)
12 (63) 15 (79) 19 (100)
Corpus Inflammation Activity Density
1 (4) 6 (23) 7 (27)
12 (46) 15 (58) 10 (39)
11 (42) 3 (11) 5 (19)
2 (8) 2 (8) 4 (15)
Antrum Inflammation Activity Density
0 (0) 1 (4) 0 (0)
2 (8) 10 (38.5) 6 (23)
8 (31) 10 (38.5) 9 (35)
16 (61) 5 (19) 11 (42)
p
1
2
3
6(32) 4 (21) 0 (0)
1 (5) 0 (0) 0 (0)
0 (0) 0 (0) 0 (0)
<0.001 <0.002 <0.001
14 (74) 16 (84) 19 (100)
4 (21) 2 (11) 0 (0)
0 (0) 0 (0) 0 (0)
1 (5) 1 (5) 0 (0)
<0.001 <0.001 <0.001
8 (42) 18 (95) 19 (100)
10 (53) 1 (5) 0 (0)
1 (5) 0 (0) 0 (0)
0 (0) 0 (0) 0 (0)
<0.001 <0.001 <0.001
120
A.P. Carelli et al. / Digestive and Liver Disease 39 (2007) 117–121
Table 3 Logistic univariate regression analysis for carditis Variables
Category (n)
Esophagitis
A (35) P (11)
0.254
A (8) P (38)
0.003
A (18) P (28)
Antrum inflammation Antrum activity Antrum density Corpus inflammation Corpus activity Corpus density Cardia lymphoid aggregates
p
OR 1 2.66
95% CI OR 0.50–14.25
1 30
3.16–284.92
<0.001
1 27.08
3.55–78.32
A (19) P (27)
<0.001
1 27.08
4.78–153.53
A (15) P (31)
<0.001
1 18.56
3.93–87.73
A (22) P (24)
0.005
1 8.4
1.93–36.63
A (26) P (20)
0.009
1 9
1.73–46.90
A (31) P (15)
1 0.003
4.35
1.13–16.85
A: absent P: present.
Carditis was present in all infected children except for 3 (11.5%) while only 7/19 (36.8%) of the non-infected children had carditis.
4. Discussion This study shows that carditis is associated with H. pylori infection. Several reports support our results on both adults [3,4,17,18] and children [13]. Borrelli et al. [13] have observed carditis in 100% of 22 infected children and in 68% of the GERD group (p < 0.001). Moreover, the degrees of gastritis at the cardia are correlated with those at the antrum and body, except for both antrum and cardia density and corpus and cardia activity. These findings suggest that cardia inflammation may occur due to H. pylori-induced pangastritis. Contrarily to other report [3] we have observed gastritis in the cardia with a lower intensity. No similar studies were found in the literature comparing the inflammatory changes in these three gastric areas of children. It is likely that these findings are quite different in such patients. We have observed higher both prevalence and H. pylori density in the antrum (57.7%) than in the corpus (42.2%) and the cardia (31.1%), differently from other reports that describe colonizations that are similar in the antrum and in the cardia [3,4] or greater in the cardia [13]. This observation suggests differences in either the migration pattern from antrum to cardia or the focal bacteria distribution. This consecutive study had well defined both inclusion and exclusion criteria which reduced the selection and sample bias. Studies on gastric cardia in children are rare in the literature: two studies were on autopsy [2,19], one study was retrospective, and only one study was prospective [13].
Methodology errors may happen in autopsy studies due to the autolysis of gastric mucosa that occurs shortly after death. In the same way, endoscopic biopsy specimens present the inconvenience of the cardia anatomic landmarks, which are not always easy to establish and frequently do not correspond to the gastroesophageal junction [20]. Regarding carditis etiology, the disagreement between such findings could reflect the difficulty in standardizing the location where the mucosal specimens should be obtained [1]. We have observed that histologically defined gastric cardiac mucosa is not always present in the endoscopically defined cardiac region. However, the types of glandular epithelium found in the normal gastric cardia are controversial. We have observed pure oxyntic-type glands in 24 (53.3%) patients, including two patients whose biopsies showed a squamocolumnar epithelium. It is important to know how the anatomic extension of cardia is related to the patient’s size; as in children the cardia is shorter, good performances of biopsy in such patients are not frequent. Glickman et al. [12] have observed mucous glandular epithelium in 81% of the patients, mucous/oxyntic-type in 19% of them, and none presented exclusively oxyntic glandular epithelium. The same was observed by Wieczorek et ¨ al. [21]. In contrast, Oberg et al. [5] have found that only 88/334 (26%) patients showed mucous glandular epithelium in biopsy specimens from the esophageal gastric junction. In a retrospective study with 72 autopsies, including adults and children, the investigators have found oxyntic glandular epithelium in 29% of the cases. Another autopsy study performed in 30 pediatric patients report mucous-type epithelium in all cases, supporting the concept that the gastric cardia is present from birth as a normal structure [22]. Cardia mucosal features (epithelial cells, glandular epithelium, amount of lymphocytes and plasma cells in the mucosa, and presence of lymphoid aggregates) were graded according to two retrospective studies performed by the same group [12,21], one of them with cardia morphology analysis in children [12]. We have adopted this criterion due to a detailed description of the histologic features of the cardia in pediatric patients, which provide a richness of details to be analyzed, although this criterion is complex and extensive. A histologic review of the cardia epithelial cells in this study has shown the mucous type in the specimens of all patients. Multilayered epithelium was not observed in any of them. Wieczoreck et al. [21] have analyzed cardia histologic features in two patient groups (H. pylori-infected and GERD patients) and multilayered epithelium was found only in patients with carditis related to GERD (27%). The differences in both type of information (autopsy, retrospective and prospective studies) and the inclusion criteria, together with a lack of anatomic pattern in the cardia region, render the comparison of studies a hard task and contribute to the divergences found in the literature on the morphology of the cardia and the etiology of carditis. In conclusion, our study shows that carditis is strongly related to H. pylori infection and pangastritis is a frequent finding in infected patients.
A.P. Carelli et al. / Digestive and Liver Disease 39 (2007) 117–121
Furthermore, we have found a correlation among the intensities of inflammatory changes observed in cardia, corpus and antrum except for antrum and cardia density and corpus and cardia activity. Further studies on the anatomic region of cardia are needed. Practice points • Carditis is strongly related to H. pylori infection and pangastritis was a frequent finding in infected children and adolescents with dyspepsia. • The intensity of gastritis (inflammation, activity) and degree of colonization was significantly higher in the antrum than in the corpus and cardia.
[6]
[7]
[8]
[9]
[10] [11]
[12]
Research agenda • The lack of the anatomic and morphology pattern in the cardiac region render comparison between the studies of carditis difficult. • Further studies about the type of glandular epithelium in the cardia and the extension of this region are needed.
[13]
[14]
[15]
[16]
Conflict of interest statement None declared.
[17]
References [18] [1] Goldblum RJ, Richter JE, Vaezi M, Falk GW, Rice TW, Peek RM. Helicobacter pylori infection, not gastroesophageal reflux, is the major cause of inflammation and intestinal metaplasia of gastric cardiac mucosa. Am J Gastroenterol 2002;97:302–11. [2] Chandrasoma PT, Der R, Ma Y, Dalton P, Taira M. Histology of the gastroesophageal junction: an autopsy study. Am J Surg Pathol 2000;24:402–9. [3] Genta RM, Huberman RM, Graham DY. The gastric cardia in Helicobacter pylori infection. Human Pathol 1994;25:915–9. [4] Hackelsberger A, G¨unther T, Schultze V, Labez J, Roessner A, Malferttheiner P. Prevalence and pattern of Helicobacter pylori gastritis in the gastric cardia. Am J Gastroenterol 1997;92:2220–4. ¨ [5] Oberg S, Peters JH, DeMeester TR, Chandrosoma PT, Hagen JA, Ireland AP, et al. Inflammation and specialized intestinal metaplasia of
[19]
[20] [21]
[22]
121
cardiac mucosa is a manifestation of gastroesophageal reflux disease. Ann Surg 1997;226:522–32. Csendes A, Smok G, Burdiles P, Sagastume H, Rojas J, Puente G, et al. ‘Carditis’: an objective histological marker for pathologic gastroesophageal reflux disease. Dis Esophagus 1998;11:101–5. Der R, Tsao-Dei DD, Demeester T, Peters J, Groshen F, Lord RVN, et al. Carditis a manifestation of gastroesophageal reflux disease. Am J Surg Pathol 2001:25245–52. Voutilainem M, F¨arkkil¨an M, Mecklin JP, Juhola M, Sipponem P. Chronic inflammation at the gastroesophageal junction (carditis) appears to be a specific finding related to Helicobacter pylori infection and gastroesophageal reflux disease. Am J Gastroenterol 1999;94:3175–80. Lembo T, Ipponti AF, Ramers C, Weinstein WM. Inflammation of the gastro-oesophageal junction (carditis) in patients with symptomatic gastro-oesophageal reflux disease: a prospective study. Gut 1999;45:484–8. Wolf C, Sildenrijk CA, Timmer R, Breumelhof R, Smout AJPM. Does carditis have two different etiologies? Dig Dis Sci 2001;46:2424–32. Pieramico O, Zanetti MV. Relationship between intestinal metaplasia of the gastro-oesophageal junction, Helicobacter pylori infection and gastro-oesophageal reflux disease: a prospective study. Digest Liver Dis 2000;32:567–72. Glickman NJ, Fox V, Antonioli DA, Wang HH, Odze RD. Morphology of the cardia and significance of carditis in pediatric patients. Am J Surg Pathol 2002;26:1032–9. Borrelli O, Hassall E, D’Armiento F, Bosco S, Mancini V, DiNardo G, et al. Inflammation of the gastric cardia in children with symptoms of acid peptic disease. J Pediatr 2003;143:520–4. Dixon MF, Genta RM, Yardley JH, Correa P. Classification and grading of gastritis. The update Sydney Sistem. Am J Surg Pathol 1996;20:1161–81. Vandenplas Y. Reflux esophagitis in infants and children: a report from working group on gastro-oesophageal reflux disease of the European Society of Pediatric Gastroenterology and Nutrition. J Pediatr Gastroenterol Nutr 1994;18:413–22. Ogata SK, Kawakami E, Patricio FRS, Pedroso MZ, Santos AM. Evaluation of invasive and non invasive methods for the diagnosis of Helicobacter pylori infection in symptomatic children and adolescentes. S˜ao Paulo Med J/Rev Paul Med 2001;19:48–53. Morini S, Zullo A, Hassan C, Lorenzetti R, Stella F, Martini MT. Gastric cardia inflammation: role of Helicobater pylori infection and symptoms of gastroesophageal reflux disease. Am J Gastroenterol 2001;96:2337–40. Mcnamara D, Buckley M, Crotty P, Hall W, O’Sullivan M, O’Morain. Carditis: all Helicobacter pylori or is there a role for gastro-oesophageal reflux? Scand J Gastroenterol 2002;37:772–7. 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. Spechler JS. The role of gastric carditis in metaplasia and neoplasia at the gastroesophageal Junction. Gastroenterology 1999;117:218–28. Wieczorek J, Wang HH, Antonioli DA, Glickman JN. Pathologic features of reflux and Helicobacter pylori-associated carditis. Am J Surg Pathol 2003;27:960–8. Kilgore SP, Ormsby AH, Gramlich TL, Rice TW, Richter JE, Falk GW, et al. The gastric cardia: fact or fiction? Am J Gastroenterol 2000;95:921–4.