**3. Discussion**

172 Gastrointestinal Endoscopy

atrophy of the body mucosa (Figure 2 A, C, D). This atrophy was mainly represented by partial or complete replacement of oxyntopeptic glands with glands with intestinal metaplasia and antral-type mucous glands known as pseudoantral or pseudopyloric metaplasia (Figs. 2 D and 3). In several small fragments with poor representation of gastric mucosa, the pseudoantral metaplasia seemed to be a factor responsible for the failure to distinguish between antral and body atrophic mucosa. Small biopsy fragments of atrophic

Fig. 2. Normal gastric mucosa of the antrum (A) and body (B) regions. Note that the foveolar region of the gastric antrum occupies approximately half the mucosa thickness while the mucus-secreting glands are distributed in the basal half of the mucosa. The connective tissue of the lamina propria is visible between the glands. Differently, the body mucosa (B) have short foveolas, and the oxintopeptic glands are numerous and justaposed occupy almost the entire space of conjunctive masking the lamina propria. C and D: body mucosa with severe glandular atrophy from a patients with ABG. Presence of glands with intestinal metaplasia (thin arrow) and pseudoantral metaplasia (thick arrow). The antral mucosa seen in A is from

the same patient with ABG seen in C and D

For many dyspeptic patients the diagnosis of ABG is the first step indicating the presence of autoimmune gastritis, associated or not with pernicious anemia. Since some of these patients should be monitored periodically from a clinical point of view and regarding the changes in their gastric mucosa, care should be taken not to overlook cases of gastritis with predominant or selective atrophy of the gastric fundus and body on the occasion of the first medical procedures they undergo. Knowing that the final diagnosis of this pathological entity usually depends on a close interaction between endoscopist and pathologist, an indication of the presence of these endoscopic changes to the pathologist directs more attention to the correct diagnosis. This correlation becomes even more necessary if we consider that the histologic processing artifacts, the absence of oxintopeptic glands in the histological sections plus the presence of extensive areas of intestinal and pseudoantral metaplasia, may complicate at first sight the differential diagnosis between ABG and atrophic multifocal gastritis.

Table 2 shows that standard endoscopic examination detected signs of gastric mucosa atrophy in 139 (70.9%) patients with ABG, although the affected gastric region was not specified. In the other 57 (29.1%) patients this change was not found although severe glandular atrophy of the body was found at histology. A more precise endoscopic report could direct the pathologist to a more conclusive histologic diagnosis. Most of the 55 patients confirmed as ABG cases among the 230 patients with varying degrees of mucosal atrophy at histology had an inaccurate endoscopy report.

Atrophic Body Gastritis: A Challenge for the

replaced or not with metaplastic glands (Dixon et al, 1996).

**3.2 ABG: Waiting for a pathological definition** 

2011).

Presumptive Endoscopic and Histologic Diagnosis of Autoimmune Gastritis 175

chronic gastritis the presence of few foci of intestinal metaplasia may correspond to the existence of established glandular atrophy is still an unresolved question. However, the presence of these foci in the gastric mucosa should be considered at least a signal of ongoing glandular loss, since in most cases intestinal metaplasia is associated with patent gastric atrophy. Some authors recommend that the histological diagnosis of moderate or severe atrophy of the gastric mucosa should be limited to cases of a glandular loss of at least 50%,

Fig. 3. Histological aspect of the body mucosa of a patient with ABG. Extensive areas of pseudoantral metaplasia (long arrows) and glands exhibiting intestinal metaplasia (short arrows). Oxyntopeptic glands are not observed. The picture may simulate atrophic antral mucosa and the diagnosis of multifocal gastritis, especially under conditions of limited interaction between endoscopists and pathologists and of inappropriate tissue processing

The designation of ABG has been extensively used without a specific definition of its limits. It is possible that the roots of this problem reside in the conceptual vagueness of what is called "atrophic gastritis" or even "gastric atrophy" (Genta, 1997). The criterion of ABG adopted for the present study involves only advanced cases of oxintopeptic mucosal atrophy with the antral mucosa showing no relevant histological changes. From a histological point of view, these patterns overlap those described for gastritis of autoimmune etiology. This definition of ABG is close to the one designated as "metaplastic autoimmune atrophic gastritis" (Park et al., 2010), and far from those defined by other authors (Vannella et al.

#### **3.1 Gastric mucosal atrophy: A problematic area of histopathology**

The histologic diagnosis of gastric mucosa atrophy remains a difficult area to handle. In addition to the subjectivity of the histologic diagnosis, other factors contribute to blurring the histological interpretation, such as: (*a*) representativeness of the limited endoscopic biopsy. This bias occurs more often in cases in which a small number of biopsies are obtained for this purpose. It should be remembered here that both major types of chronic gastritis and glandular atrophy of the gastric mucosa occur with multifocal distribution. Therefore, the number of gastric biopsies, while important, need not exceed 2 to 4 fragments per region of the stomach, except for special conditions. Consensus among pathologists and endoscopists advocates, in routine cases, two biopsies from the antral region, two from the body and one from the *incisura angularis* (angular notch) (Dixon et al, 1999); (*b*) important for a more accurate diagnosis of glandular atrophy of the gastric mucosa is the good representativeness of the tissue sections. Slices containing the entire thickness of the mucosa, from the more superficial epithelial lining to the *muscularis mucosa*, favor a more reliable histopathological analysis for the interpretation of the presence of glandular atrophy (Fig. 2 A, C, D); (*c*) endoscopic biopsies showing histologic processing distortions arising from inappropriate tangential sections of the mucosa limit the observation of tissue components; (*d*) prominent inflammatory changes and exuberant lymphoid follicles can cause distortion or separation of glands, giving false impression of glandular rarefaction; (*e*) endoscopic biopsies from only one region of the stomach, for example, only from the antral region, leave the mucosa of the body without proper histological analysis. In these cases, selective glandular atrophy of the mucosa of the body may be missed in cases where the body mucosa presents a false endoscopic appearance of normality.

Although highly subjective and dependent on many factors inherent in sampling, the diagnosis and grading of glandular atrophy of the gastric mucosa rests on the final decision of the pathologist. Moreover, in cases of severe atrophy histopathological diagnosis consistently achieves high sensitivity and reproducibility when the endoscopic and histotechnical procedures prior to histological analysis run with relative normality.

It should be remembered that the oxyntic mucosa usually presents a much higher glandular density than that of the antral mucosa. At first, connective tissue and small vessels of the *lamina propria* are not so evident because they are masked by the large number of juxtaposed oxintopeptic glands (Fig. 2 B). This fact becomes more relevant because of tissue shrinkage caused by fixatives commonly used, such as formaldehyde. Tissue fixatives retract the mucosal structures, especially the loose connective tissue of the *lamina propria*. Thus, when examining a histological section of normal oxyntic mucosa the first impression one gets is that it appears to be constituted by almost only glandular epithelial tissue.

Therefore, the loss of oxyntic glands and their replacement with connective tissue tends to be clearly identified by histology. In these cases the degree of subjectivity for the diagnosis of atrophy can be considered small. In contrast, the antral mucosa usually presents less dense glandular tissue and the *lamina propria* is more apparent with a consequent more problematic histologic interpretation of moderate degrees of atrophy in this region (Fig. 2 A). The presence of areas of intestinal metaplasia that occur with relative frequency in the antral and body mucosa in *H. pylori* chronic gastritis is an objective histological sign indicating the replacement of specialized glands of the gastric mucosa with glands of the intestinal type. This fact strengthens the diagnosis of evolving or established glandular atrophy. What the presence of intestinal metaplasia may represent regarding the existence and grading of gastric atrophy is still a debatable issue (Genta, 1996, 1997). Whether in some cases of

The histologic diagnosis of gastric mucosa atrophy remains a difficult area to handle. In addition to the subjectivity of the histologic diagnosis, other factors contribute to blurring the histological interpretation, such as: (*a*) representativeness of the limited endoscopic biopsy. This bias occurs more often in cases in which a small number of biopsies are obtained for this purpose. It should be remembered here that both major types of chronic gastritis and glandular atrophy of the gastric mucosa occur with multifocal distribution. Therefore, the number of gastric biopsies, while important, need not exceed 2 to 4 fragments per region of the stomach, except for special conditions. Consensus among pathologists and endoscopists advocates, in routine cases, two biopsies from the antral region, two from the body and one from the *incisura angularis* (angular notch) (Dixon et al, 1999); (*b*) important for a more accurate diagnosis of glandular atrophy of the gastric mucosa is the good representativeness of the tissue sections. Slices containing the entire thickness of the mucosa, from the more superficial epithelial lining to the *muscularis mucosa*, favor a more reliable histopathological analysis for the interpretation of the presence of glandular atrophy (Fig. 2 A, C, D); (*c*) endoscopic biopsies showing histologic processing distortions arising from inappropriate tangential sections of the mucosa limit the observation of tissue components; (*d*) prominent inflammatory changes and exuberant lymphoid follicles can cause distortion or separation of glands, giving false impression of glandular rarefaction; (*e*) endoscopic biopsies from only one region of the stomach, for example, only from the antral region, leave the mucosa of the body without proper histological analysis. In these cases, selective glandular atrophy of the mucosa of the body may be missed in cases where the body

Although highly subjective and dependent on many factors inherent in sampling, the diagnosis and grading of glandular atrophy of the gastric mucosa rests on the final decision of the pathologist. Moreover, in cases of severe atrophy histopathological diagnosis consistently achieves high sensitivity and reproducibility when the endoscopic and

It should be remembered that the oxyntic mucosa usually presents a much higher glandular density than that of the antral mucosa. At first, connective tissue and small vessels of the *lamina propria* are not so evident because they are masked by the large number of juxtaposed oxintopeptic glands (Fig. 2 B). This fact becomes more relevant because of tissue shrinkage caused by fixatives commonly used, such as formaldehyde. Tissue fixatives retract the mucosal structures, especially the loose connective tissue of the *lamina propria*. Thus, when examining a histological section of normal oxyntic mucosa the first impression one gets is

Therefore, the loss of oxyntic glands and their replacement with connective tissue tends to be clearly identified by histology. In these cases the degree of subjectivity for the diagnosis of atrophy can be considered small. In contrast, the antral mucosa usually presents less dense glandular tissue and the *lamina propria* is more apparent with a consequent more problematic histologic interpretation of moderate degrees of atrophy in this region (Fig. 2 A). The presence of areas of intestinal metaplasia that occur with relative frequency in the antral and body mucosa in *H. pylori* chronic gastritis is an objective histological sign indicating the replacement of specialized glands of the gastric mucosa with glands of the intestinal type. This fact strengthens the diagnosis of evolving or established glandular atrophy. What the presence of intestinal metaplasia may represent regarding the existence and grading of gastric atrophy is still a debatable issue (Genta, 1996, 1997). Whether in some cases of

histotechnical procedures prior to histological analysis run with relative normality.

that it appears to be constituted by almost only glandular epithelial tissue.

**3.1 Gastric mucosal atrophy: A problematic area of histopathology** 

mucosa presents a false endoscopic appearance of normality.

chronic gastritis the presence of few foci of intestinal metaplasia may correspond to the existence of established glandular atrophy is still an unresolved question. However, the presence of these foci in the gastric mucosa should be considered at least a signal of ongoing glandular loss, since in most cases intestinal metaplasia is associated with patent gastric atrophy. Some authors recommend that the histological diagnosis of moderate or severe atrophy of the gastric mucosa should be limited to cases of a glandular loss of at least 50%, replaced or not with metaplastic glands (Dixon et al, 1996).

Fig. 3. Histological aspect of the body mucosa of a patient with ABG. Extensive areas of pseudoantral metaplasia (long arrows) and glands exhibiting intestinal metaplasia (short arrows). Oxyntopeptic glands are not observed. The picture may simulate atrophic antral mucosa and the diagnosis of multifocal gastritis, especially under conditions of limited interaction between endoscopists and pathologists and of inappropriate tissue processing
