Evaluation of Better Staining Method among Hematoxylin and Eosin, Giemsa and Periodic Acid Schiff-Alcian Blue for the Detection of Helicobacter pylori in Gastric Biopsies

Background This study was undertaken to evaluate the preferred method (Giemsa or periodic acid Schiff-Alcian blue [PAS-AB] stains) of detecting Helicobacter pylori (H. pylori) in gastric mucosal biopsies in terms of sensitivity, specificity and applicability. To the best of my knowledge, this is the first report comparing Giemsa and PAS-AB staining for the detection of H. pylori in such biopsies. Methods The formalin-fixed paraffin-embedded blocks of 49 gastric biopsies from different patients were collected from the archive of anatomical pathology at King Abdulaziz Medical City, National Guard, Riyadh, Saudi Arabia. From each block, three slides were prepared and analysed using the hematoxylin and eosin (H&E), Giemsa and PAS-AB stains to detect the presence/absence of H. pylori, and the results were compared in terms of sensitivity, specificity and applicability. Results The majority of the biopsies in this study showed antrum-type gastric mucosa. Only 15 biopsies showed active gastritis, whereas the rest showed chronic gastritis. Three biopsies showed intestinal metaplasia. All were detected by PAS-AB stain, but only two-thirds were detected by H&E stain. Fifteen gastric biopsies showed H. pylori infection in general and in 13 of them, active gastritis cases were discovered. Fourteen out of these 15 H. pylori infection cases were detected by Giemsa stain, whereas only 13 cases were detected by H&E stain. PAS-AB stain showed the worst results since it demonstrated only 40% sensitivity and 67.65% specificity in H. pylori detection. Conclusion Giemsa stain has better sensitivity and specificity in gastric H. pylori infection detection than PAS-AB. Therefore, using PAS-AB stain to detect H. pylori infection is not recommended.


Introduction
Helicobacter pylori (H. pylori) is a welldefined, spiral-shaped, gram-negative bacterium responsible for the onset of several gastric pathologies ranging from mild gastritis to gastric malignancies (1). This microorganism was first discovered and described in a 1985 article by Marshall et al. (2). Since then, many studies have been conducted to explore this kind of bacterium preferred method among Giemsa and PAS-AB stains in detecting H. pylori in gastric biopsies. The specific aim is to compare the sensitivity, specificity and applicability of Giemsa and PAS-AB stains in detecting H. pylori in gastric biopsies.

Methods
After receiving ethical approval from the regional ethical committee as well as obtaining the patients' informed written consents, 56 formalin-fixed paraffin-embedded blocks from gastric biopsies and their glass slides from different patients were collected. These represented all the gastric biopsy cases from 15 May 2018-31 May 2018 in the archive of the Anatomical Pathology Department at King Abdulaziz Medical City, Ministry of National Guard, Riyadh City, Saudi Arabia. From each block, three glass slides were prepared and stained with H&E, Giemsa and PAS-AB stains, according to the standard protocols. Each glass slide was examined blindly by a pathologist for the presence/absence of H. pylori, without knowing the results from other stains of the same biopsy. At the same time, two other pathologists examined the slides in combination, to be controls for the study. That is, they looked at the H&E, Giemsa and PAS-AB glass slides together for each case, to determine the presence/absence of the H. pylori.
The main inclusion criterion was the presence of gastritis and the primary exclusion criterion included cases with dysplasia or gastric carcinomas, since the presence of these kinds of lesions could be indirect hints of the presence of H. pylori infection in the biopsy. Therefore, 7 cases were excluded and we finished with only 49 paraffin blocks.

Results
A total of 49 patients participated in this study, ranging in age from 22-63 years. Eighteen patients were male and 31 patients were female. The majority of the histological types of the biopsies showed antrum-type gastric mucosa (Table 1).
Fifteen biopsies showed chronic active gastritis (Figure 1), whereas the remaining 34 showed only chronic gastritis (Table 1). Three cases showed intestinal metaplasia ( Table 2). All of these were detected by PAS-AB stain examined and tested histologically and cultured, with special stains, immune stains, polymerase chain reaction (PCR) tests and a rapid urase test (Campylobacter-like organism [CLO] test). The non-invasive methods include many tests such as the urea breath test (UBT), serological tests and detection of the H. pylori antigen in urine, blood and stool samples (3,(9)(10)(11)(12)(13). Each method has its own advantages and disadvantages. Detection of H. pylori in gastric biopsies is the gold standard method of detecting H. pylori infection, but exactly what should be used as a panel of tests in these biopsies remains controversial (1,9,(11)(12)(13)(14).
Several histopathological staining panels have been used for many years to detect H. pylori in gastric biopsies, including the combination of the hematoxylin and eosin (H&E) stain and other special stains, such as the Giemsa, periodic acid Schiff-Alcian blue (PAS-AB), methylene blue and Warthin-Starry silver stains (3,13,15). Other detection methods, such as immunohistochemistry (IHC) and fluorescent in-situ hybridisationbased staining methods are also available for H. pylori detection (14,(16)(17)(18). Molecular biology techniques, such as PCR tests, are also used to diagnosis H. pylori infection (13,19).
It is now well documented that chronic gastritis-related disorders are one of the most prevalent causes of death in humans and their incidence rate is continuously on the rise globally, including in Saudi Arabia (20)(21). Despite this, the prognosis for advanced gastric disorders is still limited and poor. Therefore, early detection of these disorders is extremely important for better treatment of patients with gastric disorders (1). The Giemsa and PAS-AB stains are the most common panels used routinely in gastric biopsies. The former is used to detect H. pylori infection, whereas the latter is used to highlight the presence of intestinal metaplasia in these biopsies.
The role of PAS-AB stain (compared to Giemsa stain) in the detection of H. pylori still needs further exploration, as some pathologists claimed that Giemsa is a better stain in H. pylori detection compared to other stains. Others claimed that the PAS-AB stain is as good for this purpose as the Giemsa stain. Hence, the selected panel of staining for detecting H. pylori in gastric biopsies appears controversial and confusing. Therefore, this study is designed to investigate the detection methods and to determine the     ( Figure 2). In contrast, only two-thirds of them were detected by routine H&E stain (Figure 3). The reason was that the lesion was focal in the biopsy. Fifteen gastric biopsies were found to be positive for H. pylori infection (through the combination of H&E stain, Giemsa stain, PAS-AB stain, the Urea breath test and the CLO test). These cases were distributed as follows: 13 cases showed chronic active gastritis and 2 cases showed only chronic gastritis. The sensitivity and specificity of H&E, PAS-AB and Giemsa stains in the detection of H. pylori are shown in Table 3. From this table, it is clear that Giemsa stain was superior to both H&E and PAS-AB stains in the detection of H. pylori with 93.33% sensitivity and 100% specificity ( Figure 4). It is also clear that among the three stains, PAS-AB performed the worst since it had high false positivity as well as false negativity. This is mostly because of the PAS-AB dirty background of the slide, which made it difficult to differentiate between true H. pylori infection and the dirty background, as in Figures 5 and 6. In H. pylori detection, H&E stain had good specificity; however, its sensitivity was relatively low (Figure 1).

Discussion
The detection of H. pylori infection through endoscopic features alone is not appropriate for a diagnosis. Therefore, these must be combined with the histopathological report (22,23). This is necessary because: i) the H. pylori microorganism cannot be seen endoscopically; ii) there are many types of mucosal alteration that can be associated with H. pylori infection (such as gastritis, mucosal atrophy, ulceration, erosion, polyps and neoplasms) (24); and iii) there is no correlation between endoscopic findings and histopathological findings in patients with H. pylori infection (25).
Many researchers have concluded that the most common microscopic mucosal changes associated with H. pylori infection are active gastritis and/or the presence of lymphoid follicles with germinal center formation (16,(24)(25)(26). This is also supported by our research, since 86.67% of the H. pylori infection cases were associated with active gastritis. Yet if one considers the sensitivity, specificity, cost, reproducibility and rapidity of any test, there is no single gold standard test to detect H. pylori infection that has all these parameters (13).
However, gastric biopsies are the gold standard method to detect this infection (1,9,(11)(12)(13)(14). The question is which panel to use in these biopsies. In addition to the standard H&E stain, many panels of stains have been proposed for this purpose, including Giemsa, PAS-AB, Warthin-Starry and IHC stains (13). However, some authors still suggest that ancillary studies should be done only for cases in which there is a high suspicion of the presence of H. pylori infection that cannot be visually seen by H&E (such as cases with active gastritis or with germinal centre formation) (13,23,(28)(29)(30)(31)(32). The advantages of H&E alone are that: i) it is inexpensive; ii) it is a rapid test; and iii) it allows the evaluation of the presence of H. pylori as well as its complications (such as intestinal metaplasia and neoplasms) (13,30,32,34). The disadvantages of this method are that: i) it will not detect H. pylori in cases with mild gastritis and a small quantity of bacteria; and ii) it overlooks this bacterium in patients who have been partially treated with proton pump inhibitors and/or antibiotics (13,(31)(32). According to our research, the specificity of H&E in H. pylori detection is high (91.18%); however, its sensitivity is low (66.67%).
Giemsa stain is a simple, rapid and inexpensive stain that has good sensitivity, specificity and consistency in the detection of H. pylori infection (13,27,33,34). Therefore, it is routinely used in some institutions. According to many scientific papers, Giemsa stain is superior to H&E in the detection of H. pylori (13,27,30,33,34). Based on our findings, the sensitivity of Giemsa stain in the detection of H. pylori is 93.33% while its specificity is 100%, which makes it better than H&E and PAS-AB stains.
The data from this study showed that PAS-AB stain is the worst choice in H. pylori detection (in comparison with H&E and Giemsa stains) with 40% sensitivity and 67.65% specificity. However, this stain is best in the detection of intestinal metaplasia, since it detects all cases (even those with a small focal focus) and it produces no false positives. Although some authors recommend the routine use of IHC stains in detecting H. pylori (14,35), the majority recommend their use only if necessary, as in cases of ambiguous H&E or Giemsa stain morphology because their use is time-consuming and expensive (13,36,37).