METHODS

Study subjects

A total of 438 patients with H pylori infection were recruited from those visiting the endoscopy unit at our hospital between May 1995 and August 1996. The patients, comprising 282 men and 156 women, age range 18 to 72 years (mean age, 41.1; SD, 10.1), were divided into three groups: (1) group 1 consisted of 185 dyspeptic patients with chronic gastritis (105 superficial chronic gastritis and 80 atrophic gastritis confirmed by endoscopy and histopathology); (2) group 2 consisted of 69 patients with endoscopically confirmed gastric ulcers; (3) group 3 consisted of 184 patients with endoscopically confirmed duodenal ulcers. Fifty H pylori negative dyspeptic patients with endoscopically and histologically confirmed normal gastric antral mucosa were selected as the control group.

Endoscopy and biopsy sampling

All patients underwent upper gastrointestinal endoscopy before H pylori eradication treatment in our hospital. Those patients with chronic gastritis underwent repeat endoscopy after treatment. According to the mapping principle of our study, during each endoscopic examination, six antral mucosal biopsies were taken from the anterior and posterior wall, and the lesser and the greater curvature within 2–3 cm from the pylorus. Additional biopsy specimens were taken from lesional areas when necessary. Two antral biopsy specimens were used for microbiological assessment (rapid urease test and culture), the others were used for histological assessment.

Histopathology

All antral biopsy specimens for histological examination were fixed in 10% formalin, embedded in paraffin wax on the oriented edge, and cut into 5 μm thick sequential sections. All tissue sections were stained with haematoxylin and eosin for histological examination and with a Giemsa stain for H pylori assessment.

The severity and activity of gastritis, atrophy, and intestinal metaplasia, in addition to H pylori density, were evaluated according to the updated Sydney system.⁵ The infiltration of gastric mucosa by mononuclear cells and polymorphonuclear leucocytes, atrophy, and intestinal metaplasia were graded as follows: 0, none; 1, mild; 2, moderate; 3, severe. Atrophy was defined as the loss of inherent glandular tissue, with or without replacement by intestinal-type epithelium. For optimal histological evaluation, all gastric biopsy specimens included surface epithelium and muscularis mucosae, and small pinch biopsies were excluded from our study. A total of 1660 gastric biopsy specimens were investigated and analysed. A lymphoid follicle was defined as an aggregate of lymphocytes with a germinal centre. Lymphoid aggregates were defined as accumulations of lymphocytes and plasma cells without a germinal centre. To assess the degree of lymphoid tissue hyperplasia, lymphoid follicles and aggregates in each antral biopsy specimen were counted and tabulated. The density of lymphoid follicles and aggregates (the number of biopsy specimens with lymphoid tissue hyperplasia/the number of biopsy specimens examined) represents the intensity of lymphoid tissue hyperplasia in the antral gastric mucosa.

All gastric specimens, including those from the normal control group, were investigated and scored independently by two pathologists (XYC and YS). The pathologists were blinded to any clinical information. Before starting the study, the two pathologists reached consensus about the methods of grading the features of gastritis through interactive sessions at a multiheaded microscope. Disagreements were resolved by consensus.

Helicobacter pylori assessment

The presence of H pylori was determined by the rapid urease test, culture, and histopathology. Patients were positive for H pylori when H pylori was detected by at least two of the above three methods. The eradication of H pylori infection was defined as the absence of microorganisms in both culture and tissue sections four to six weeks after cessation of treatment.

Treatment and follow up study

Eligible patients with H pylori positive chronic gastritis were given a seven day course of bismuth triple treatment. The patients again underwent endoscopy and biopsy at four to six weeks after completion of the treatment. Complete sets of gastric antral biopsy specimens were obtained from 35 and seven patients, respectively, who were H pylori negative 12 and 24 months after eradication of H pylori. Histopathological changes and lymphoid tissue hyperplasia were reinvestigated in those biopsy specimens.

Statistical analysis

Statistical analysis was performed using the statistical package for social science (SPSS) program. Two tailed χ² and Fisher exact probability and 95% confidence intervals (CI) were used when appropriate. Significance was set at p < 0.05.

RESULTS

The prevalence and density of lymphoid follicles and aggregates in gastroduodenal diseases

Figure 1 shows the prevalence of lymphoid follicles and aggregates in various gastroduodenal diseases. Overall, lymphoid follicles and aggregates were present in the antral mucosa of 350 (79.9%; 95% CI, 0.76 to 0.84) patients. The prevalences of lymphoid follicles and aggregates in various gastroduodenal diseases were significantly different. The highest number of lymphoid follicles and aggregates (89.9%; 95% CI, 0.83 to 0.97) occurred in those patients with gastric ulcers. The lowest prevalence of lymphoid follicles and aggregates (74.6%; 138 of 185; 95% CI, 0.68 to 0.81) was found in the patients with chronic gastritis. The prevalence of lymphoid follicles and aggregates was significantly different in patients with chronic gastritis and those with gastric ulcer (p = 0.0082). The presence of lymphoid follicles and aggregates in the gastric antral mucosa was higher in patients with gastric ulcer than in those with duodenal ulcers, although the difference was not significant (p = 0.109).

A total of 1476 eligible gastric antral biopsy specimens from different groups (669 chronic gastritis, 239 gastric ulcer, and 568 duodenal ulcer) were analysed, and the number of lymphoid follicles and aggregates in each sample of gastric antral mucosa was graded from 0 to 6 (average, 0.7). The overall density of the lymphoid follicles and aggregates was 0.71 (1031 of 1476) in infected antral mucosa. The density of lymphoid follicles and aggregates in gastric antral mucosa was also significantly different in the various gastroduodenal diseases, and correlated with the prevalence of lymphoid follicles and aggregates. The highest (0.82; 196 of 239) and lowest (0.56; 318 of 568) density of lymphoid follicles and aggregates occurred in the patients with gastric ulcers and duodenal ulcers, respectively (fig 2). There were significant differences in the density of lymphoid follicles and aggregates between patients with chronic gastritis and duodenal ulcers (p = 1.6 × 10^-15), and between those with gastric ulcers and duodenal ulcers (p = 2.24 × 10^-12). Although the density of lymphoid follicles and aggregates was higher in the patients with gastric ulcers than in those with chronic gastritis, the difference was not significant (p = 0.105).

There were very few scattered lymphocytes and no lymphoid tissue hyperplasia in the antral lamina propria in the normal control group.

Relation between lymphoid follicles and aggregates and pathological features in the chronic gastritis group

Figures 3–5 show the relation between lymphoid follicles and aggregates and pathological features in the chronic gastritis group. The prevalence of lymphoid follicles and aggregates was higher in the patients with atrophic gastritis (71 of 80; 88.7%; 95% CI, 0.82 to 0.96) than in those with superficial gastritis (67 of 105; 63.8%; 95% CI, 0.55 to 0.73). The prevalence and density of lymphoid follicles and aggregates were closely correlated with the severity and activity of inflammation. With an increasing degree of chronic inflammation in the gastric antral mucosa, higher numbers and a greater density of lymphoid follicles and aggregates were detected (fig 4). Numbers of lymphoid follicles and aggregates were significantly higher in the patients with active gastritis than in those without active gastritis (p < 0.0001; fig 3). Correlations between the detection of lymphoid follicles and aggregates and atrophy or intestinal metaplasia were evaluated in fig 5; a higher prevalence was detected in atrophic and intestinal metaplasia mucosa (p < 0.001). However, a lower prevalence and density of lymphoid follicles and aggregates was found with increasing degree of atrophy and intestinal metaplasia.

Prevalence and density of lymphoid follicles and aggregates after H pylori eradication in chronic gastritis

Figures 6 and 7 show the changes in the prevalence and density of lymphoid follicles and aggregates in the gastric antral mucosa before treatment and four to six weeks, 12 months, and 24 months after H pylori eradication treatment. Eradication of H pylori infection resulted in a decrease of lymphoid tissue hyperplasia in the gastric antral mucosa. Eradication of H pylori was seen in 132 of the 185 (71.4%; 95% CI, 0.65 to 0.78) patients after triple antibiotic treatment. The prevalence of lymphoid follicles and aggregates before treatment was 74.6% (138 of 185; 95% CI, 0.68 to 0.81) and decreased to 57.3% (106 of 185; 95% CI, 0.50 to 0.64) four to six weeks after treatment (p = 0.000447). The number of lymphoid follicles and aggregates in the antral mucosa significantly declined in those patients without H pylori infection (54.5%; 95% CI, 0.41 to 0.68; p = 0.000198), whereas in those where H pylori infection remained there was only a slight decrease (95% CI, 0.56 to 0.72; p = 0.134) after treatment. The density of H pylori colonisation decreased significantly in 53 patients in whom H pylori infection remained four to six weeks after treatment. Helicobacter pylori colonisation was seen in only 62 of 188 biopsy specimens in those patients.

The prevalence and density of lymphoid follicles and aggregates decreased to 22.9% (95% CI, 0.09 to 0.37) and 0.12, respectively, in 35 patients who were H pylori negative one year after treatment. Furthermore, a small but steady decline in the frequency (14.3%; 95% CI, −0.12 to 0.40) and density (0.11) of lymphoid follicles and aggregates was observed in the antral mucosa in seven patients who were H pylori negative two years after treatment.

DISCUSSION

Wyatt and Rathbone¹¹ investigated 419 pairs of antral and corporal mucosal biopsy specimens and found lymphoid follicles in 27.4% of patients with H pylori associated gastritis. Eidt and Stolte⁸ studied “two to three antral specimens” from 2692 patients with H pylori infection and detected lymphoid follicles and aggregates in 54% of those patients. Genta and Hammer⁹ found that the prevalence of lymphoid follicles was 63.8% (110 of 174) in patients with chronic gastritis and 100% in patients with H pylori infection. In our study, lymphoid follicles and aggregates were found in the gastric antral mucosa of 76.0% (350 of 438) of patients with various H pylori associated gastric diseases. These differences in prevalences between other previously published data and our study might result from several factors: (1) Biopsy sites may have been different in the various studies. It has been known for some time that the frequency of lymphoid follicles and aggregates is lower in oxyntic mucosa (14.8–44.0% lower) than that seen in antral mucosa.^8,10 (2) The number of biopsy specimens taken varied in the different studies. In many studies where a lower prevalence of lymphoid follicles and aggregates was reported a small number (two to three) of antral biopsy specimens may have been taken. The higher rate of detection of lymphoid follicles and aggregates in the antrum in our study might be explained by the larger number (four) of antral specimens.

The highest prevalence (89.9%) and density (0.82) of lymphoid follicles and aggregates was found in those patients with gastric ulcers in a “background” of chronic gastritis. By increasing the number of biopsy specimens taken, a higher prevalence and density of lymphoid follicles and aggregates may be detected in gastric antral mucosa. Lymphoid follicles and aggregates in gastric antral mucosa might be found in 100% of H pylori positive patients if sufficient gastric biopsy specimens are examined, in accordance with the finding of Genta and Hammer.⁹ It is widely believed that the presence of lymphoid follicles and aggregates in gastric mucosa is a conspicuous pathological feature of H pylori associated chronic gastritis.

“Lymphoid tissue hyperplasia is a specific immunological reaction to H pylori infection”

This is remarkable considering the lack of lymphoid tissue in the normal gastric mucosa, and H pylori infection seems to be the major determinant in the development of gastric lymphoid tissue.^4,5,12 The eradication of H pylori infection results in a decrease in lymphoid follicles and aggregates. In our study, only very few scattered lymphocytes were detected and no lymphoid tissue hyperplasia was found in the antral lamina propria in 190 biopsy specimens in the normal control group. Lymphoid follicles and aggregates in the gastric mucosa decreased rapidly after the eradication of H pylori infection, which corresponded well with previous data.⁶ No lymphoid tissue hyperplasia was detected in the gastric antral mucosa in 77.1% and 85.7% of patients who were H pylori negative one and two years after treatment in our present study. Our data confirmed that the presence and degree of colonisation with H pylori strongly correlates with the development and disappearance of lymphoid tissue in the stomach. After colonisation of the gastric mucosa by H pylori, close and persistent contact between the bacterium and the mucosa leads to a specific immune response. Lymphoid tissue hyperplasia is a specific immunological reaction to H pylori infection. Previous studies have shown that the proliferation of T cells and macrophages is induced by cytotoxin and H pylori, resulting in the release of several cytokines (interleukin 2 (IL-2) and IL-6), which leads to the proliferation of B cells and the development of lymphoid follicles and aggregates in the infected gastric mucosa. The infiltration of T cells decreased significantly when H pylori were eradicated or reduced in number after treatment, resulting in a reduced stimulation of B cell proliferation and the disappearance of lymphoid tissue hyperplasia in the gastric mucosa.^13,14

Our study confirmed the finding of Stolte and Eidt^7,8 that there was a significant correlation between lymphoid follicles and aggregates and the severity and activity of inflammation in the gastric mucosa. Lymphoid follicle and aggregate development was maximal in the infected antral mucosa in patients with severe active chronic gastritis (fig 8). There was a progressive decline in the infiltration of chronic inflammatory cells in the antral mucosa with increasing degree of atrophy and intestinal metaplasia. With the regression of inflammation in the advanced stage of chronic gastritis, fewer numbers of lymphoid follicles and aggregates were detected. The main reasons for this loss of lymphoid follicles and aggregates are: (1) Because H pylori only colonised the gastric epithelium, the organisms were absent from the areas of intestinal metaplasia that are usually present in atrophic stomach. (2) The hypochlorhydric environment is an unfavourable one for the growth of H pylori. (3) Acidic glycoproteins secreted by metaplastic epithelium may provide a more hostile environment for H pylori than the natural glycoproteins of the normal mucosal layer. (4) The replacement of gastric epithelium and glandular tissue by intestinal-type epithelium leads to lower numbers of specific receptors for H pylori.¹⁵ (5) Intestinal epithelium is more resistant to H pylori infection than gastric epithelium because of the relatively higher concentration of specific IgA.¹⁶ All the above reasons might prohibit H pylori from attaching to intestinal-type cells.

In conclusion, there was a close correlation between lymphoid tissue hyperplasia and morphological changes, such as the severity and activity of chronic gastritis, atrophy, intestinal metaplasia, and colonisation by H pylori. Lymphoid tissue hyperplasia caused by H pylori infection decreased quickly after eradication of the infection, declining at a slow and steady rate, but did not disappear during the two years of follow up.