Histological and endoscopic features of the stomachs of patients with Chagas disease in the era of Helicobacter pylori

Introduction: Most studies that have evaluated the stomachs of patients with Chagas disease were performed before the discovery of Helicobacter pylori and used no control groups. This study compared the gastric features of chagasic and non-chagasic patients and assessed whether gastritis could be associated with Chagas disease. Methods: Gastric biopsy samples were taken from patients who underwent endoscopy for histological analysis according to the Updated Sydney System. H. pylori infection was assessed by histology, 16S ribosomal ribonucleic acid (rRNA) polymerase chain reaction (PCR), serology and the 13C-urea breath test. Patients were considered H. pylori-negative when all of these diagnostic tests were negative. Clinical and sociodemographic data were obtained by reviewing medical records and using a questionnaire. Results: The prevalence of H. pylori infection (70.3% versus 71.7%) and chronic gastritis (92.2% versus 85%) was similar in the chagasic and non-chagasic groups, respectively; such as peptic ulcer, atrophy and intestinal metaplasia. Gastritis was associated with H. pylori infection independent of Chagas disease in a log-binomial regression model. However, the chagasic H. pylori-negative patients showed a signifi cantly higher grade of mononuclear (in the corpus) and polymorphonuclear (PMN) (in the antrum) cell infi ltration. Additionally, the patients with the digestive form of Chagas disease showed a signifi cantly lower prevalence of corpus atrophy than those with other clinical forms. Conclusions: The prevalence of H. pylori infection and of gastric histological and endoscopic features was similar among the chagasic and non-chagasic patients. Additionally, this is the fi rst controlled study to demonstrate that H. pylori is the major cause of gastritis in patients with Chagas disease.

Chagas disease, which is a zoonosis caused by the protozoan Trypanosoma cruzi, affects approximately 8 million people in Latin America 1 . In endemic areas, T. cruzi is commonly transmitted to humans through triatomine vectors; however, in non-endemic areas where the vector is lacking, T. cruzi can be transmitted horizontally through blood transfusions or through organ and tissue transplantation or transmitted vertically from infected mothers to their offspring. Currently, thousands of people with Chagas disease are estimated to live in the USA, Europe and Asia due to the migration of infected individuals from endemic areas 2,3 .
The acute phase of Chagas disease is usually asymptomatic, and this disease progresses to a chronic phase, called the indeterminate form, which generally starts with a long period characterized by the absence of symptoms 4 . Several years after the initial infection, approximately 20-30% of infected people will develop lesions in the heart and gastrointestinal tract that are characteristic of the other clinical forms of Chagas disease 1 . The primary clinical manifestations of the chronic cardiac form of Chagas disease are heart failure, cardiac arrhythmia, and thromboembolism 4 . In the digestive form of this disease, the clinical manifestations are due the destruction of the intramural plexuses, primarily the myenteric plexus, leading to the incoordination of peristalsis and sphincter METHODS achalasia of the esophagus and resulting in megaesophagus and megacolon 5,6 . Some patients can simultaneously develop lesions in the heart and digestive tract, which are characteristic of the cardiodigestive form of chronic Chagas disease.
Previously, the chronic gastritis commonly observed in chagasic patients primarily with the digestive form of Chagas disease was attributed to biliary duodenal-gastric refl ux, as well as gastric hypomotility and hypochlorhydria caused by enteric nervous system injury due to chronic T. cruzi infection [7][8][9][10][11] . However, these studies were performed before the recognition that infection with the bacterium Helicobacter pylori is the main cause of chronic gastritis throughout the world and plays an important role in peptic ulcers, gastric carcinoma and mucosaassociated lymphoid tissue (MALT) lymphoma 12 .
After the discovery of H. pylori, two uncontrolled studies that evaluated a small group of patients with the digestive 13 and indeterminate 14 16,17 demonstrated a signifi cantly higher seroprevalence of H. pylori infection among chagasic patients independent of the clinical form of the disease; however, the gastric mucosa of the patients was not evaluated in either study or in the study by Pinazo et al. 15 .
Therefore, the aim of this study was to compare the prevalence of H. pylori infection and the gastric histological and endoscopic features of chagasic and non-chagasic patients and of chagasic patients with the digestive form of Chagas disease and those with other clinical forms because data regarding alterations in the gastric mucosa of chagasic patients, especially after the discovery of H. pylori, are lacking. In addition, whether gastritis could be associated with Chagas disease was assessed.

Patients and clinical samples
From July 2007 to November 2011, gastric biopsy specimens were obtained from consecutive chagasic (cases) and non-chagasic (controls) outpatients who underwent upper gastrointestinal endoscopy at the Endoscopic Service of Universidade Federal do Triângulo Mineiro (UFTM) for an evaluation of dyspeptic symptoms or as a follow-up for chronic Chagas disease. These patients were from the same geographical region and lived in a rural or urban area of Uberaba and other nearby cities west of the State of Minas Gerais, Brazil in a region called Triângulo Mineiro, which was previously considered an endemic area for Chagas disease. During endoscopy, the biopsy specimens were obtained from the antrum and corpus of the stomach of each patient for histological and molecular analysis. The exclusion criteria were previous therapy for the eradication of H. pylori infection, pregnancy, the presence of coagulation disorders, complications such as perforation or gastric bleeding, anatomical impediment to endoscopy, esophageal varices, and concomitant severe illness. Additionally, 5mL of blood was collected from each patient for the serological diagnosis of Chagas disease and for the detection of anti-H. pylori immunoglobulin G (IgG).

Ulcers and esophagitis
Ulcers were diagnosed from endoscopic fi ndings and further stratifi ed according to the location into gastric or duodenal ulcers 18 . Esophagitis was defi ned as the presence of one or more mucosal breaks and was graded according to the Los Angeles classifi cation system 19 .

Chagas disease diagnosis
Enzyme-linked immunosorbent assay (ELISA) (Chagatest ELISA, Wiener Laboratories, Rosario, Argentina), hemagglutination (HA) (Chagatest HAI, Wiener Laboratories, Rosario, Argentina) and immunofluorescence (Immuno-Con Chagas, WAMA Diagnostica, São Carlos, Brazil) tests were performed according to the manufacturers' instructions. The chagasic patients showed positive reactions to at least two serological tests, and the non-chagasic patients displayed negative reactions to all three tests used. The patients with only one positive test were excluded. The chagasic patients were classifi ed according to the clinical form of the disease (i.e., indeterminate, digestive, cardiac or cardiodigestive) based on the results of a conventional electrocardiogram and of a contrast X-ray of the esophagus and colon.

Histological analysis
The endoscopic biopsy samples from the gastric mucosa of the antrum and corpus were fi xed overnight in buffered formalin, and embedded in paraffi n. Then, each paraffi n block was sliced to 4-µm thickness and stained with hematoxylin and eosin for histological examination.
All of the samples were examined by the same experienced pathologist who was blinded to the patients' clinical information.
The presence of mononuclear cell infi ltration indicated chronic gastritis, and polymorphonuclear (PMN) cell infiltration signifi ed gastritis activity. The degrees of gastric mucosal infl ammation (mononuclear cells) and of gastritis activity (PMN cells) were assessed in accordance with the Updated Sydney System and were classifi ed using the following four grades: 0: none; 1: mild; 2: moderate; and 3: severe 20 . The gastritis location was classifi ed according to the presence of mononuclear cells in the antrum, corpus or both. Glandular atrophy and intestinal metaplasia were classifi ed as absent or present.
Helicobacter pylori infection was diagnosed histologically by direct light microscopic observation of the organism using Warthin-Starry staining. The infection was considered negative if the bacterium was absent from all of the biopsy sites and positive if the bacterium was present in at least one of the biopsy sites. The location of H. pylori was classifi ed according to the presence of bacteria in the antrum, corpus or both.
Helicobacter pylori status The histological analysis, urea breath test, 16S rRNA PCR, and anti-H. pylori IgG (serology) detection were performed to determine the H. pylori status. The patients for whom two of the tests were positive were classifi ed as H. pylori-positive, and the patients for whom all of the tests were negative were considered H. pylori-negative. The patients using a proton pump inhibitor (PPI) and/or antibiotics with gastric histological alterations were considered H. pylori-positive even if they were positive by the serological testing alone.

C-Urea breath test
The urea breath test (UBT) was performed after 6h of fasting. A fi rst basal breath sample was collected, and then, an oral dose of 75mg of 13 C-urea dissolved in 200mL of orange juice was administered. An additional breath sample was collected after 30min. The breath samples were analyzed using an isotope-selective non-dispersive infrared spectrometer (NDIRIS; Wagner Analytical Systems, Bremen, Germany). A change in the delta over baseline (DOB) value of > 4‰ was considered H. pylori-positive 21 .

16S RNA gene detection
Antrum or corpus tissue deoxyribonucleic acid (DNA) was extracted using a commercial kit (QIAamp DNA Mini Kit; Qiagen, Hilden, Germany), according to the manufacturer's instructions with minor modifi cations. The presence of H. pylori DNA was detected by the amplifi cation of the 16S rRNA gene by PCR using specifi c primers described previously by Riley et al. 22 . The amplifi cation products were visualized by 1.5% agarose gel electrophoresis. The gels were stained with ethidium bromide and observed under UV light.

Anti-Helicobacter pylori IgG detection
IgG antibodies specifi c to H. pylori were detected using a commercial ELISA kit (Pyloriset EIA-GIII; Orion Diagnostica, Espoo, Finland). The antigen employed in this kit is a cell surface protein of H. pylori strain NCTC 11637. The assay was performed according to the manufacturer's instructions. The concentrations of the IgG antibodies in the serum samples were determined by interpolation from a standard curve constructed by plotting the absorbance values obtained for each of the 4 calibrator sera samples, which are provided by the manufacturer, against the corresponding anti-H. pylori concentrations in U/mL.

Statistical analysis
The data were analyzed using the Statistical Package for the Social Sciences (SPSS) software (version 16.0; SPSS Inc., Chicago, IL, USA). The association of variables such as H. pylori infection, chronic gastritis, active gastritis, atrophy, metaplasia, gastric ulcers, and esophagitis with Chagas disease and with the digestive clinical form was assessed by a prevalence ratio (PR), chi-square and Fisher's exact test as appropriate. An adjusted analysis for age and sex (confounding factors) was also performed. In addition, a model of log-binomial regression for potentially confounding factors, such as gender, mean age, alcohol consumption, and PPI and non-steroidal antiinfl ammatory drug (NSAID) usage, was constructed to evaluate the association between chronic gastritis and Chagas disease or H. pylori infection. The prevalence ratios (PRs) and the 95% confi dence intervals (CIs) were determined. The signifi cance level was set at 5%. The intensity (mononuclear cell) and activity (polymorphonuclear cell) of gastritis were compared by the Mann-Whitney U test.

Ethical considerations
This study was approved by the Ethics Committee of the Universidade Federal do Triângulo Mineiro (UFTM), Uberaba, Minas Gerais (MG), Brazil and was conducted in accordance with the Declaration of Helsinki. All patients provided their signed written informed consent forms and answered a questionnaire to obtain socio-demographic data. The clinical information was obtained by reviewing the patients' medical records.
In total, 184 patients were included in the study. Of these patients, 64 (34.8%) were chagasic (32 male and 32 female; mean age 62.5 ± 9.4 years) and 120 (65.2%) were non-chagasic (38 male and 82 female; mean age 47.6 ± 14.3 years). The prevalence rates of H. pylori infection and the gastric features of chagasic and non-chagasic patients are summarized in Table 1.
No signifi cant differences were observed between the chagasic and non-chagasic groups regarding H. pylori-positive status, chronic gastritis, chronic active gastritis, atrophy, intestinal metaplasia, esophagitis, and gastric ulcers. Although duodenal ulcers were observed only among the non-chagasic patients (p=0.01), the difference was not signifi cant after adjusting for age and sex (p=1.0). All of the chagasic patients with gastric ulcers and the majority of the non-chagasic patients with gastric ulcers (8 of 10; 80%) and duodenal ulcers (10 of 11; 90.9%) were H. pylori-positive. Higher rates of H. pylori positivity among the chagasic and non-chagasic patients were detected by serology (67.2% and 75.8%, respectively) compared with detection by 16S rRNA PCR (62.5% and 61.7%, respectively), histology (64.1% and 57.5%, respectively), and the UBT (52.7% and 62.1%, respectively). The bacterium was in both the antrum and corpus gastric regions in 71.1% of the chagasic and 57% of the non-chagasic H. pylori-positive patients, similar to the results for the location of gastritis (57.8% and 51.2%, respectively).
Chronic gastritis was associated with H. pylori infection (p<0.01) independent of Chagas disease and of the other covariables that were evaluated (Table 2). However, we found H. pylori-negative patients with gastritis in both the chagasic and non-chagasic groups, as shown in Table 3. Six (31.6%) of the 19 chagasic and one of the 34 (2.9%) non-chagasic H. pylori-negative patients showed PMN cell infiltration (chronic active gastritis) in the antrum or corpus, whereas higher numbers of the H. pylori-negative chagasic (14 of 19; 73.7%) and the non-chagasic (17 of 34; 50%) patients showed mononuclear cell infiltration (chronic gastritis). However, no H. pylori-negativ e patients who presented PMN cell infi ltration were graded as moderate or severe, and none that presented mononuclear cell infi ltration was graded as severe. Notably, the group of chagasic H. pylori-negative patients exhibited higher grades of mononuclear (p<0.01) and    Table 4. No signifi cant differences were observed between the chagasic patient subgroups regarding H. pylori status, chronic gastritis, chronic active gastritis, gastric ulcers and esophagitis. Intestinal metaplasia and atrophy were detected in the antrum region in a higher percentage of the patients with the digestive form of Chagas disease than among those patients with other clinical forms. In contrast, intestinal

DISCUSSION
Approximately half of the world's population is estimated to be infected with H. pylori, which is a Gram-negative bacterium that is now considered the most common cause of gastritis. All H. pylori-infected individuals develop chronic gastritis due a local immune response to the bacterium; however, most infected individuals remain asymptomatic. Long-term gastric infl ammation leads to more serious outcomes, such as peptic ulcer disease, gastric atrophy (a precursor of gastric carcinoma) and mucosaassociated lymphoid tissue lymphoma, in a small percentage of infected individuals. We found a high prevalence of gastritis and H. pylori infection in both groups of chagasic and non-chagasic patients in the present study. Although two previous studies have reported similar fi ndings for a small group of chagasic patients 13,14 , this is the fi rst controlled study to demonstrate that H. pylori is the major cause of gastritis in patients with Chagas disease.
Moreover, we found some chagasic and non-chagasic patients with gastritis who were not infected with H. pylori. This phenomenon of chronic active gastritis with no detectable H. pylori seems to be increasingly common in the Western world, and some possible explanations for this phenomenon include antibiotic therapy for the treatment of other infections, the widespread use of PPIs, and inadequate sampling or suboptimal staining techniques as suggested previously [23][24][25] . However, we used multiple diagnostic tests in the present study to avoid false H. pylori-negative results because no single test is considered the gold standard for the diagnosis of H. pylori infection. The tests in the present study included a commercial serological test that has high sensitivity and specifi city 26 and that is not infl uenced by a temporary or recent decrease in the bacterial load caused by PPI usage and by nonspecifi c antibiotic therapy. Taken together, we believe that the chance of having patients with false H. pylori-negative gastritis in the study population is low. Therefore, other causes of gastritis should be considered in patients with H. pylori-negative gastritis, especially among those patients with Chagas disease because, interestingly, the intensity (in the corpus mucosa) and the activity (in the antral mucosa) of gastritis were signifi cantly higher in this group of patients than in the non-chagasic group, as demonstrated in Table 3. Based on these fi ndings, one may speculate that factors inherent to T. cruzi infection could induce the development of gastritis in H. pylori-negative patients; this hypothesis merits further investigation.
Other causes of gastritis that may mimic H. pylori infection include reactive (or chemical) gastritis with focal activity, autoimmune gastritis, lymphocytic gastritis, and other infections 20 . Notably, approximately 29% of chagasic patients with H. pylori-negative gastritis showed corpus-predominant immune cell infi ltration, which is a common characteristic of autoimmune gastritis. In contrast, no control patients with H. pylori-negative gastritis presented immune cell infi ltration predominantly in the corpus mucosa. Bile-associated chemical metaplasia and atrophy of the corpus mucosa were more frequent in the chagasic patients with other clinical forms than among those patients with the digestive form; however, a signifi cant difference was observed only for atrophy even after adjusting for gender (p=0.02).
gastritis may be another cause of gastritis, especially among chagasic patients, because earlier studies demonstrated that duodenogastric refl ux is common in chagasic patients with megaesophagus 11 .
Chronic H. pylori infection exerts profound and diverse effects on gastric acid secretion that depend on the gastritis degree and location and that are related to the disease outcome. In subjects with predominantly antral gastritis, H. pylori increases acid secretion, predisposing patients to duodenal ulcers, whereas in other patients with predominantly corpus gastritis or pangastritis, acid secretion is impaired, resulting in gastric atrophy, and these subjects have an increased risk of gastric cancer 27 . In Chagas disease, although the most signifi cant digestive manifestation is the development of megaesophagus and/or megacolon, a partial denervation of the stomach at the level of the intramural plexuses has been reported, inducing several gastric motor and secretory disturbances, including delayed gastric emptying of solids, hypergastrinemia and reduced gastric acid secretion 6,11,28 . However, the gastric physiological changes and the effects on the clinical outcome that could be caused by long-term co-infection with H. pylori and T. cruzi are unknown.
Before the discovery of H. pylori, hypochlorhydria in patients with Chagas disease was believed to be a protective factor for duodenal ulcers, and hypomotility along with gastric stasis was thought to cause an increase of gastric ulcers. In an earlier study, Ceneviva et al. 29 evaluated patients with peptic ulcers undergoing surgery and reported a higher prevalence of gastric ulcers than duodenal ulcers among those patients with Chagas disease. In the present study, the prevalence of gastric ulcers was also higher than duodenal ulcers among the chagasic patients; however, no signifi cant difference was observed compared with the control group. In contrast, two uncontrolled studies of patients with megaesophagus 11 and with Overall, the prevalence of gastric atrophy and intestinal metaplasia in the antrum and corpus was similar between the chagasic and non-chagasic patients. However, the reasons why patients with other clinical forms of Chagas disease showed a signifi cantly higher prevalence of corpus-predominant atrophy than those patients with the digestive clinical form are unknown. Nonetheless, we believe that this fi nding is not associated with PPI usage because the percentages of the chagasic patients with the digestive form and with other clinical forms taking this medication were similar. Two other studies have also assessed the gastric histological features of chagasic patients, although only the antral mucosa was examined. Barbosa et al. 13 reported antrum atrophy in 47% of patients with the digestive form, which is a higher prevalence than that found in our present study (15%). In contrast, Oliveira et al. 14 did not fi nd atrophy or intestinal metaplasia among patients with the indeterminate form of Chagas disease.
The high prevalence of co-infection with T. cruzi and H. pylori in our study (approximately 70%) and in other studies is not surprising because the transmission of both of these pathogens is associated with poor socioeconomic conditions [13][14][15][16][17] . Moreover, in Brazil, which is considered a developing country, the overall prevalence of H. pylori is high and varies from 55% to 87% according to the geographical region 30,31 . Notably, no signifi cant differences in the prevalence of infection with H. pylori among the clinical forms of Chagas disease were observed in the present study or in other previous studies 16,17 .
This fi nding indicates that patients with the digestive form are not at high risk for H. pylori infection, although these patients have hypochlorhydria 7,8,10 , which a condition that can favor gastric colonization by H. pylori 32,33 .
In conclusion, the prevalence of H. pylori infection and gastric histological and endoscopic features was similar among the chagasic and non-chagasic patients. Additionally, our fi ndings indicate H. pylori infection as the major cause of chronic gastritis in patients with Chagas disease, although the impact of Chagas disease on gastritis cannot be excluded, especially in H. pylori-negative patients. Furthermore, considering the high prevalence of H. pylori among chagasic patients irrespective of the clinical form of the disease, we believe that the investigation and eradication of H. pylori infection may result in an improvement of non-specifi c upper gastrointestinal symptoms among patients with Chagas disease, in addition to preventing the development of more severe gastric diseases associated with H. pylori.