CONVENTIONAL VIDEOENDOSCOPY CAN IDENTIFY HELICOBACTER PYLORI GASTRITIS?

ABSTRACT Background: Studies with latest technologies such as endoscopy with magnification and chromoendoscopy showed that various endoscopic aspects are clearly related to infection by Helicobacter pylori (HP). The description of different patterns of erythema in gastric body under magnification of images revived interest in identifying these patterns by standard endoscopy. Aim: To validate the morphologic features of gastric mucosa related to H. pylori infection gastritis allowing predictability of their diagnosis as well as proper targeting biopsies. Methods: Prospective study of 339 consecutive patients with the standard videoendoscope image analysis were obtained, recorded and stored in a program database. These images were studied with respect to the presence or absence of H. pylori, diagnosed by rapid urease test and/or by histological analysis. Were studied: a) normal mucosa appearance; b) mucosal nodularity; c) diffuse nonspecific erythema or redness (with or without edema of folds and exudate) of antrum and body; d) mosaic pattern with focal area of hyperemia; e) erythema in streaks or bands (red streak); f) elevated (raised) erosion; g) flat erosions; h) fundic gland polyps. The main exclusion criteria were the use of drugs, HP pre-treatment and other entities that could affect results. Results: Applying the exclusion criteria, were included 170 of the 339 patients, of which 52 (30.58%) were positive for HP and 118 negative. On the positive findings, the most associated with infection were: nodularity in the antrum (26.92%); presence of raised erosion (15.38%) and mosaic mucosa in the body (21.15%). On the negative group the normal appearance of the mucosa was 66.94%; erythema in streaks or bands in 9.32%; flat erosions 11.86%; and fundic gland polyps 11.86%. Conclusion: Endoscopic findings are useful in the predictability of the result and in directing biopsies. The most representative form of HP related gastritis was the nodularity of the antral mucosa. The raised erosion and mucosa in mosaic in the body are suggestive but not specific to the infection. The other forms were not conclusive of the presence of HP.

The diagnosis of the infection requires at least two tests in accordance with the european guidelines 22 . The most used are the rapid urease test and histological analysis 25 . The rapid urease test has a sensitivity of 92% and specificity of 95% 5 . Laine et al. 18 found histological sensitivity variation in the identification of HP bacteria according to the bacterial density in the sample. The H & E staining showed 70% to 98% of sensitivity and specificity of 89% to 98% in the identification of HP and Giemsa sensitivity of 64% to 96% and specificity of 98% to 100% 18 . HP is distributed irregularly in gastric mucous epithelial surface and the relatively low density of bacteria in various groups of patients can lead to false-negative results in methods of biopsy 15 .
There are few reports in the literature regarding endoscopic patterns of related HP gastritis using conventional endoscopy. In the initial works in 1995, some authors concluded that it was not possible to establish this diagnosis based on only endoscopy 4,28 . However, newer technologies such as magnification and chromoendoscopy 3,27 showed that there endoscopic aspects that are clearly associated with HP infection, while others relate to uninfected or eradication 11 . Yagi et al. 32,33 described the characteristics of endoscopic findings with magnification in the gastric body with normal appearance and negative HP: enanthema in tiny streaks or spots in "pinhole" aspect, which correspond to sub-epithelial capillaries and venules networks called RAC (regular arrangement of collecting venules). Anagnostopoulos et al 15 demonstrated that enanthema Mosaic or speckled in the gastric body is more related to infection by HP, as also the enanthema in association with swelling of folds and exudate, indicating intense active inflammatory process. Enanthema in streaks or bands (gastropathy) and the appearance of tiny red spots corresponding to subepithelial venules coletantes (normal condition) are associated with the absence of infection by HP. These authors suggest that, using this technique to perform the biopsy pathology is not required 1 .
However, magnification and chromoendoscopy it is not available in most diagnostic centers, but also demand more time for execution and learning and does not seem to be practical in daily routine examinations. If specific patterns of HP related gastritis can be identified using conventional endoscopy, these standards could be applied to predict and select patients and biopsies could be directed to areas suspected of being infected by HP.
This study aims to verify the validity of the recognition of morphological patterns of gastritis associated with HP using conventional endoscopy, which would be helpful to favor the targeting of biopsies for the most affected areas.

METHODS
This is an observational cross-sectional study approved by the ethics committee in local research. In this study was analyzed, prospectively, the endoscopic findings of 339 consecutive patients from May 27 2015 until July 10, 2015 in Endoclinic, SP, Brazil. Free and informed consent was obtained from all participants.
The included underwent endoscopy unit with Fujinon Pentax EPM 4400 or 3500 and all tests performed by a single professional. Sedation was carried out with 25 to 50 ug fentanolamina and midazolam 2-5 mg 19 . Images were recorded and saved in the database (OCRAM® system, SP) being obtained of 12-20 images per patient in all cases. These were selected for this review 6-8 images. Endoscopic aspects that were evaluated were: normal mucosa appearance, nonspecific diffuse erythema of antrum and body, erythema in streaks or bands (red streaks), mosaic mucosal pattern in the gastric body, flat erosions (minor surface defects 5 mm and flat edges), elevated (raised) erosions, nodularity of the mucosa and fundic gland polyps (Figures 1 and 2).  The diagnosis of H. pylori infection was done by the urease tests (Uretest Renylab®, MG) performed with at least two fragments from the antrum and two from the body. The positive histological fragments were subjected to histological examination by hematoxylin and eosin (HE) staining and Giemsa stain to identify the HP, made by a pathologist who was blinded to the other results.
The following exclusion criteria were: patients with anemia, liver cirrhosis, gastric cancer, gastrectomy, renal failure, congestive heart failure, recent use of antiinflammatory drugs, aspirin, antithrombotics, use of proton pump inhibitors or H2-receptor antagonists in past two months, prior history of eradication of HP; extensive gastric mucosal atrophy.
The data were studied in frequency tables and contingency being used the Fisher test and chi-square association for nominal data and Mann-Whitney and unpaired t test for association of numerical data. The significance used was 5% (p = 0.05). The calculations were made with the Graph Pad Prism version 5.0 software.

RESULTS
The positive HP endoscopic findings are shown in Table  1 where it can be seen that the most frequent finding was erythema. In Table 2 are the frequency of the findings in the negative group. In about 8/52 (15.38%) infected patients, endoscopic examination was normal while normality was present in 79/118 (66.94%) of the uninfected. Comparing the endoscopic findings in Table 3, it is possible to note that patients with HP infection have more antral nodularity, mosaic pattern in the gastric body and redness of antrum and body.

DISCUSSION
In the present study we sought to identify endoscopy findings related to HP infected or uninfected gastric mucosa. The selected endoscopic findings for this research have clear association with HP related gastritis and have been described in previous papers. Edema plies, with or without exudate, diffuse or patchy erythema are regarded as mucosal inflammatory process and are good indicators of the presence of HP 30 .
Atrophic gastritis areas were avoided for biopsies in this study because they are hostile regions of the mucosa to the colonization of HP, causing false-negative test results and are present in older patients with longtime HP chronic infection 7 .
Magnifying studies have shown that the numerous tiny lines or red dots in the gastric body, seen with conventional endoscopy, were regular arrangement of collecting venules (RAC), characteristic finding in normal stomach without infection by HP with 100% sensitivity and 90% especificidade 32 .
In the current study the following imaging findings showed a positive association with HP: antral nodularity, mosaic pattern in the body, diffuse redness and raised erosion.
According to the literature, the antral nodularity is significantly associated with chronic active gastritis and follicular gastritis and showed high specificity (98.5%) and high positive predictive value (91.7%) but low sensitivity (32%) for the diagnosis of HP 4,17 . In this research those images shown to be 89.2 times more frequent in infected individuals (OR = 89.26 and 95% CI = 5.19 -153) demonstrating that this finding is valuable in the diagnosis of HP infection and the endoscopic finding that best showed this association.
The raised erosions are mucosal elevations on gastric folds of the antrum and distal body containing fibrin exudation and sometimes hematin. Denote chronic inflammation and besides being frequent in patients with HP, appear also in individuals with chronic use of antiinflammatory drugs 2 . In this research the findings were nearly 22 times more common in infected patients and none in the negative group (p = 0.0081, OR = 21, 99; 95% CI = 1.16 to 416.6).
The flat erosions are mucosal continuity solutions, associated with erythema, fibrin and sometimes hematin. In general it is smaller than 5 mm in diameter and less than 1 mm depth 2 . These images were not useful for the diagnosis in question. They represented 9.61% of HP positive group and 11.86% of HP negative group, with p = 0.33.
The redness of the mucosa was the most common finding. For this search, this kind of image was divided into diffuse redness in antrum and body, red strikes (according to the literature is found most negative HP cases) and mosaic pattern (more related in the HP positive cases) 30 . In this study diffuse redness was found in 30 patients in positive group HP (57.69%) and in negative group 25 (21.18%), being 5.7 times more common in infected individuals (p <0.0001, OR = 5.07, 95% CI 2.50 to 10.27). The mosaic pattern was found in 11 patients in positive group (21.15%) and in only three patients (2.54%) in the negative group (p = 0.0002; OR = 10.28; 95% CI 2.73 to 38.7 in).
Although red streaks findings have shown negative association with infection, this association could not be confirmed by analyzing the confidence interval obtained. On the negative HP group were found 11 patients with this aspect (9.32%) and none in HP positive group (p = 0.01; OR = 0.089 95% CI = 0.005 to 1.54).
Fundic gland polyps, according to literature 10,29 are associated only with uninfected cases. In this work all detected polyps are fundic gland and in all cases HP was negative (n = 14; 11.86%). No hyperplastic polyp (associated with the presence of HP and with congestive gastropathy) was found and also no adenomatous polyp (associated with intestinal metaplasia).
Polyps fundic gland, red streaks and normal mucosal appearance correlate with the negativity of HP infection as other studies 7,10,29 , but such associations could not be here demonstrated when analyzing the confidence interval obtained.
With the selective collection of fragments for histological study in the supposedly positive cases 5,6,9,12 it avoids the routine submission for pathology in cases of morphological patterns not related to infection and with HP negative urease test.

CONCLUSION
Endoscopic findings are useful predictability of location and direction of biopsies in the HP research. The most representative form of HP related gastritis was the nodularity of the antral mucosa. The raised erosion and mucosa in mosaic in the body are suggestive but not specific to the infection. The other forms were not conclusive of the presence of HP.