Abstract

The present study was undertaken to determine whether there is a histologic pattern consistent with Helicobacter Pylori (H.Pylori) colonization in the stomach and the first part of duodenum, to correlate individual parameters like surface epithelial and crypt degenerative changes, mucin content, nature and grade of inflammatory infiltrate and the presence of the organism. To compare the efficacy of histological examination hematoxylin and eosin (H and E) cold carbol fuschin stain, Gram’s stained smears and the rapid urease test (10% urea solution) in the detection of the organism. Correlation of endoscopic appearance and histology was also attempted, as well as comparison of the efficacy of multiple site biopsies with single site biopsies, in the detection of the organism. The age and sex distribution of the cases was reviewed and the nature and duration of upper abdominal symptoms was correlated with H. Pylori colonization.

Introduction

Helicobacter Pylori (H. Pylori) is a spiral, gram negative organism exclusively adapted to gastro-duodenal type of epithelium. The major issue at stake is whether the organism is definitely a pathogen or whether it is a commensal, with no role in disease causation. Further the mechanism by which the organism incites mucosal inflammation is still in the realm of hypothesis.

Material and Methods

97 biopsies were studied, 31 prospective and 66 retrospective. In the prospective study, the patients presenting with upper GI symptoms were randomly studied and at least six biopsies, two from the stomach body, two from the antrum and two from the first part of duodenum were taken. An extra antral biopsy was used to make smears from Gram’s stain and for the rapid urease test (RUT). The Gram’s stain was the Burk’s modification of Gram’s method and for the RUT 0.5 ml of 10% urea solution in de-ionized H2O at pH 6.8 with 1% phenol red as pH indicator was used. The test was observed over a 20 minute period. A positive test was indicated by a change in yellow colour of the solution to pink. All positive tests appeared within 5 minutes. The other 6 biopsies were all stained for H and E and the cold carbol fuschin (CCF) method, using the Ziehl Nielsen carbol fuschin for one minute and counterstaining with 0.8 % malachite green for one minute. Alcian blue (pH 2.5)/Periodic Acid Schiff (AB/PAS) stain was also done to evaluate mucin depletion and detect metaplasia and Gomori’s reticuline stain to confirm gastric atrophy when suspected. In the retrospective study, 66 cases were examined by H and E, CCF, AB/PAS and Gomori’s reticulin stain. Most of these biopsies however were single site biopsies, the Sydney system was used for the classification of gastritis. The Whitehead classification for duodenitis and the density grading system of Hazell et al for the density grading of the organisms.

Results

The association of H.Pylori with the various morphological and topographical forms of gastritis was examined in the prospective and retrospective studies. The commonest morphological form of gastritis encountered in both the studies was chronic quiescent gastritis (CQG), this was followed by chronic active gastritis (CAG) (Fig.1). In the prospective study CQG accounted for 54.8% cases and CAG for 16.1% cases. In the prospective study the percentage of H. Pylori positivity was much higher in the retrospective study (Tables 1 and 2). This is because multiple sites were biopsied in the prospective study, whereas most of the biopsies of the retrospective study were single

Fig. 1 : Photomicrograph showing chronic gastritis (antral biopsy). Degenerative changes in surface epithelium. Grade III lamina propria inflammatory infiltrate with lymphoid aggregate. (H and E , x100).

Table 1 and2 : Morphological diagnosis (Stomach) and H.Pylori status

site biopsies either of the antrum or body. The percentage positivity was also higher in the biopsies that were within normal limits, in the prospective study – 50% compared to the retrospective study – 20%. The density grading of H. Pylori in both the types of gastritis was higher in most cases, grade II – III (Fig. 2) compared to the cases that were within normal limits Grade I – II.

Fig. 2 : Photomicrograph showing helicobacter pylori in surface mucin (antral biopsy). Grade III. Cold Carbol Fuchsin stain x 1000.

Amongst the cases in the prospective study with gastritis 80.6% the commonest topographical form of gastritis was pan gastritis. 64% followed by antral gastritis 36%. In pan gastritis, the percentage positivity of H. Pylori was 75% and in antral gastritis it was 77.8%. This indicates the possibility of almost equal colonization of both the body and the antrum. Though significance of the association of H. Pylori with the various morphological and topographical forms of gastritis was evaluated using chi-square test. The association was not significant in any instance (P > 0.05) in both the studies.

In the prospective study duodenal biopsies were assessed in 31 cases, in 5 cases the biopsies were inadequate Whitehead’s grading system was utilized. Of the remaining 26 cases, 16.6% had grade I duodenitis, 7.6% grade II and 30.8% grade III duodenitis (Fig. 3). Among those with grade I duodenitis 75% were positive for H. Pylori in the antrum, with grade II, 62.5% cases with grade III duodenitis (Table 3). However the association

Table 3 : (prospective study) Gastritis H. Pylori status in the antrum and duodenitis

Fig. 3: Photomicrograph showing Grade III duodenitis. (H&E , x100).

of antral H. Pylori was not considered statistically significant with any of the grades of duodenitis (p > 0.05). Gastric metaplasia was found in the duodenum in 22.5% cases, out of which 51.5% had grade III duodenitis and 28.6% grade I duodenitis. In the cases of grade III duodenitis and gastric metaplasia both the duodenum and antrum were colonized with H. Pylori in 75% cases and 25% cases had no H. Pylori. Gastric metaplasia appears to be common in the higher grades of duodenitis. Moreover in the cases of severe duodenitis, there is a concurrence between H. Pylori in the antrum and duodenum.

Methods of detection

In the prospective study, cold carbol fuschin stain detected the organism in 74.2% cases (Table 4). Gram stain smear of a single antral biopsy, detected the organism in 32.3% cases, all of which were positive on histology. The rapid urease test, was performed on a single antral biopsy in 17 cases, detected the organism in five cases 29.4%, all of which were also positive on histology. There were no false positives by the gram’s stain or RUT method, but the sensitivity of either method was much lower than histology. In the prospective study only a single antral biopsy was used for Gram’s stain, and the RUT, which could account for the lower sensitivity of the methods even though both were highly specific. The organism was detected in both body and antrum in 20 cases, i.e. a concurrence rate of 87% in the retrospective study where most of the biopsies were single site biopsies; H. Pylori was detected at a lower rate of 31.8% cases. The yield was better in antral biopsies 43.3% compared to body biopsies 35.3% cases.

In the prospective study 93.54% cases presented with non-ulcer dyspepsia with 72.4% positive for H. Pylori. The total number of males was 74.2% and females 25.8% with a mean age for males 39 years and females 34 yrs. 73.9% males and 75% females were positive for H. Pylori. In the retrospective study 74.2% males and 25.5% females were biopsied. The mean age for males was 50

Table 4 : Histologically assessed Biopsy sites in the stomach and H. pylori detection rates (Cold carbol fuschin stain on paraffin embedded tissue)

years and females 41 years. 28.6% males and 11.8% females were positive for H. Pylori. A comparison of the two studies revealed a higher prevalence rate for the organism in the 30 – 60 year age group with male predominating. In the prospective study there appeared to be no correlation between age and density of colonization, keeping in mind, the fact that H. Pylori prevalence rates increases with age. The duration of symptoms in the prospective study was more than 4 weeks, in 91.3% cases, positive for H. Pylori. No relationship was apparent between the duration of symptoms and density of colonization. In the prospective study a history was obtained according to a definite protocol, listing symptoms most predictive of ulceration, epigastric pain was the commonest symptom in H. Pylori positive cases 35.5%. The commonest morphological diagnosis of gastritis associated with the above symptoms was chronic quiescent gastritis with all the grades of duodenitis, epigastric pain was again the commonest symptom since almost all the cases also had gastritis, it is not possible to attribute symptoms to pathology at any particular site. The commonest morphological form of duodenitis was grade I duodenitis. A statistically significant association between the symptoms and these morphological forms was not obtained. A correlation between endoscopic abnormalities and morphological form of gastritis was poor. By contrast endoscopic findings in the duodenum, correlated with the histological grades of duodenitis.

Discussion

Most studies reported in western literature beginning with Warren et al1-,4 have found H. Pylori to be associated with chronic active gastritis. Eastern references also mention a higher percentage of gastritis4 and chronic active gastritis,5 these include Indian references,5 Jones et al,6 however found the organism to be similarly correlated with active and inactive gastritis. In our study the commonest morphological form of gastritis encountered was chronic quiescent gastritis (CQG) this was followed by chronic active gastritis (CAG) (Tables 1 and 2) (Fig.1).

Stolte and Eidt,7 found the degree of colonization by H. Pylori to be identical in the body and antrum in 49.3% cases with the density of colonization correlating with the activity of gastritis this is supported by other authors as well.4 Though some authors feel that the evaluation of bacterial burden by histological score seems only accurate for the most severe density.10 In our prospective study, out of 23 cases positive for H. Pylori the body and antrum were both colonized to a similar degree in 20 cases-86.9% concurrent rate, though the activity of gastritis varied from site to site. The degree of H. Pylori positivity was much higher compared to the retrospective study since multiple sites were biopsied in the prospective study (Tables 1,2 and Fig. 2).

Bayerdorffer et al,9 emphasized the fact that the area of colonization appeared to be larger than the area of CAG. The reported prevalence rate of the organism in histologically normal mucosa ranges from 10% to 43%. In our study the percentage positivity was higher in the biopsies that were within normal limits, in the prospective study-50% compared to the retrospective study 20%. The density grading of H. Pylori in both the types of gastritis active and quiescent was higher in most cases Grade II and III compared to the cases that were within normal limits Grade I and II. The commonest reported topographic form of gastritis is antral gastritis3,10 though there are reports of a higher percentage of body gastritis11 as well. In our study, the commonest topographic form of gastritis was pan gastritis 64% followed by antral gastritis 36%.

The significance of the association of H. Pylori with the various morphological and topographical forms of gastritis was evaluated using the chi-square test. The association was not significant in any instance (p > 0.05) in both the studies Villako et al,12 felt that H. Pylori may be involved in the early stages of chronic gastritis but is unrelated to its progression. The failure to demonstrate a significant association between the organism and the morphological or topographical forms of gastritis in our study may reflect the already evolved state of gastritis in these cases. In the prospective study almost all the cases with H. Pylori positivity had symptoms of more than 4 weeks duration, when biopsied, suggesting that if H. Pylori is responsible for initiating gastritis the present biopsies are not representative of the initial change in the mucosa. Alternatively the lack of significant association may be due to strain variation , organisms of low virulence and genetically determined variation in the mucosal responses to the organisms in the population studied.7

Marshall and Warren13 found spiral bacteria in the antrum of 77% patients with gastric ulcers, others11 failed to demonstrate a definite association between antral spiral bacteria and gastric ulcer. In the present retrospective study the organism was found in one out of 5 biopsies for gastric ulcer.

Wyatt and Dixon stated that epithelial degeneration, irregularities of surface epithelium, with gaps between cells and tufting is often seen in H. Pylori gastritis. especially in areas where colonization is dense and H. Pylori appears to insinuate in between epithelial cells. Bonvincini et al14 in 1989, felt that H. Pylori associated gastritis had no distinctive features. This suggested that the organism does not directly damage the epithelium. The responsible factor may be neutrophils and inflammatory mediators. In the present study surface and crypt features like degenerating changes, intra epithelial leucocytes, erosions and mucin content of body and antral biopsies, were correlated with the presence of and density of colonization of H. Pylori, the presence of density of colonization to be appear in a associated to be statistically significant way with the changes there was also no statistically significant reduction in mucin content [assessed by the Alcian blue (pH 2.5)/PAS stain] in the surface epithelium or crypts, as the density grade of H. Pylori increased. A reduction in mucin content however has been found in association with H. Pylori colonization by some authors.15

The activity of gastritis and the presence of plasma cells and neutrophils14 has been associated in a statistically significant was with the presence of H. Pylori.5 Though some authors felt that was accurate only in the most severe density of colonization. In the present study the nature and density of inflammatory infiltrate was correlated with the presence and density of H. Pylori colonization. When the density of colonization was grade III the infiltrate density was also grade III in 71.4% and 50% of the antral biopsies in the prospective and retrospective studies respectively, with a grade III colonization in body biopsies a grade III infiltrate was seen in 66.7% cases in the prospective study and more in the retrospective study.

Stolte and Eidt7 found mucosal lymphoid follicles in 54% of H. Pylori associated gastritis, the prevalence of which correlated with the degree and activity of gastritis in the antrum and also with the degree of H. Pylori colonization. In the prospective study 32.2% cases had mucosal lymphoid follicles, in the gastric biopsies, out of which 25.8% had H. Pylori associated gastritis and 6.4% had normal mucosa. 5/8 cases of H. Pylori positivity had a grade II infiltrate.

In the prospective study, duodenal biopsies were assessed in 31 cases, in 5 cases the biopsies were inadequate. Whitehead’s grading system was utilized, in the remaining 26 cases, 16.6% had Grade I duodenitis, 7.6%-grade II and 30.8% grade III duodenitis (Table 3 and Fig. 3). Among those with grade I duodenitis 75% were positive for H. Pylori in the antrum, one case with grade II and 62.5% cases with grade III duodenitis, however the association of antral H. Pylori was not considered statistically significant with any of the grades of duodenitis (P>0.05). Other authors have found duodenitis occurring in various proportions some high 86% and some low4 -9.3%. The association with H. Pylori has also been found to be significant.5 Gastric metaplasia was found in the duodenum in 22.5% cases, of which 51.1% had a grade III duodenitis and 28.6% had a grade I duodenitis. In the case of grade III duodenitis and gastric metaplasia both the duodenum and antrum were colonized with H. Pylori in 75% cases. Gastric metaplasia appears to be commoner in the higher grades of duodenitis. More over in the cases of severe duodenitis there was a concordance between H. Pylori in the antrum and duodenum. Graham16 1989 was of the opinion that H. Pylori seems to form a common link between gastric metaplasia and duodenitis and H. Pylori duodenitis is always found in areas of gastric metaplasia. Wyatt et al10 however found gastric metaplasia and gastric H. Pylori to be independent predictors of acute duodenitis. The latter may depend on different causal factors, not related to H. Pylori infection as some authors19 have not found any significant difference between H. Pylori positive and negative cases in the extent of gastric metaplasia or severity of duodenitis.

In the prospective study a cold carbol fuschin stain detected the organism in 74.2% cases (Table 4, Fig. 2) Gram’s stained smear of a single antral biopsy detected the organism in 32.3% cases all of which were positive on histology. The rapid urease test (RUT) was performed on a single antral biopsy in 17 cases, detected the organism in 29.4 %, all of which were also positive on histology. There were no false positives by the Gram’s stain or RUT methods but the sensitivity of either method was much lower than histology. There were no false positives by the Grams stain or RUT methods but the sensitivity of either method was much lower than histology. Culture and histology have been found to be of equally high sensitivity.18 For all methods results were better when both antrum and fundal mucosal were sampled.18 In the prospective study only a single antral biopsy was used for Gram’s stain and the RUT which could account for the lower sensitivity of the methods even though both were highly specific. The organism was detected in both body and antrum in 20 cases i.e. a concordance rate of 87% in the retrospective study where most of the biopsy were single site biopsies, H. Pylori was detected at a lower rate in 31.8% cases. The yield was better in antral biopsies in 43.3% cases, compared to body biopsies-35.3% cases. Bayerdorffer et al9 1989, concluded that the number of biopsies, greatly influenced the diagnosis of gastritis and H. Pylori detection, the more the sites biopsied, the better the yield.

Prevalence rates for H. Pylori increases with age and is 50% in persons more than 50 years of age.6 This parallels the rising incidence of chronic gastric with age. Ahmed et al18 1991 found a male predominance in non ulcer dyspepsia (NUD) cases positive for H. Pylori however other authors,8 have found a female predominance. In the prospective study 93.54% cases presented with NUD with 72.4% positive for H. Pylori, the total number of males was 74.2% and females 25.8% with a mean age for males 39 yrs. and females 34 yrs. 73.9 %males and 75% females were positive for H. Pylori. In the retrospective study 74.2% males and 25.8% females were biopsied. The mean age for males was 50 years and for females 41 yrs. 28.6% males and 11.8% females were positive for H. Pylori. A comparison of the two studies records a higher prevalence rate for the organism in the 30-60 year. age group with males predominating. In the prospective study there appeared to be no correlation between age and density of colonization, keeping in mind the fact that H. Pylori prevalence rates increase with age.

The duration of symptoms in the prospective study was more then 4 weeks in 91.3% cases positive for H. Pylori, no relationship was apparent between the duration of symptoms and the density of colonization. Marshall and Warren13 failed to observe typical symptomatology in association with antral H. Pylori. Ahmed et al18 and Velanovich et al3 however felt that H. Pylori is a cause of dyspeptic symptoms, the most common being epigastric pain, heart burn and burping. In the prospective study the history was obtained according to a defined protocol listing symptoms most predictive of ulceration, epigastric pain was the commonest symptom in H. Pylori positive cases-35.5%. The commonest morphological diagnosis of gastritis associated with the above symptoms was CQG with all grades of duodenitis, epigastric pain was again the commonest symptom since almost all the cases also had gastritis, and it is not possible to attribute symptoms to pathology at any particular site. The commonest morphological form of duodenitis was grade I. A statistically significant association between the symptoms of these morphological forms were not obtained. Morson and Dawson19 state that the correlation between the clinical symptoms, gastroscopic appearance and gastric histology is poor in chronic gastritis. Shousha et al20 found a good correlation between endoscopies findings and histological diagnosis of duodenitis. In the prospective study also, the correlation between endoscopies abnormalities and morphological form of gastritis was poor, in contrast endoscopic finding in the duodenum correlated well with the histological the grades of duodenitis.

Conclusions

A histological pattern consistent with H. Pylori colonization could not be demonstrated, and H. Pylori does not seem to be directly responsible for surface and crypt changes, or the nature and grade of luminal infiltrate. The density and extent of colonization did not seem to have an effect on the severity of morphological changes in the stomach and duodenum. Histological examination with cold carbol fuschin stain is the most sensitive method for the detection of the organism when compared to RUT or Gram’s stained smear of mucosal biopsies. No conclusions are possible regarding the role of the organism in the evolution of chronic gastritis. Strain variation, gastric factors and environmental factors may all contribute. No definite conclusion was possible about the role of the organism in gastric / duodenal ulceration and gastric carcinoma, as the number of cases was small, multiple biopsies from different sites in the stomach improves the rate of detection of the organism. Correlation between endoscopic findings and histological changes in the gastric mucosa is poor. However duodenal histology correlates well with endoscopic appearance of the mucosa. Both the antrum and body of the stomach are colonized to a similar extent in most cases, so a biopsy, from either site is likely to yield a positive result. The presence of H. Pylori in the gastric mucosa may not be directly responsible for symptom causation. The prevalence of the organism exists in a wide age range viz 30-60yrs. males out numbered females among cases with non-ulcer dyspepsia. The clinical significance of the presence of the organism in the gastro-duodenal mucosa and a definite cause and effect relationship to disease states can only be elucidated on a large scale, long term population based prospective study with emphasis on the evolution of histological changes in the stomach and duodenum.

References

1. Warren JR,Marshall BJ. Unidentified curved bacilli in gastric epithelium in active chronic gastritis. Lancet 1983;1 :1273-5.
2. Brigic E, Hodzic L, Zildzic M. Helicobacter pylori and gastroduodenal disease in our patients: A 2 year experience. Med Arh 2000; 54 : 313-6.
3. Velanovich V. The spectrum of H pylori infection in upper gastro intestinal disease. Ann Surg 1996; 62 : 60-63.
4. Al-Mueilo SH. Gastroduodenal lesions and Helicobacter pylori infection in hemodialysis patients. Saudi Med J 2004; 25 : 1010-4.
5. Bansal D, Patwari AK, Logani KB, Malhotra VL, Anand VK. Study of diagnostic modalities and pathology of H pylori infection in children. Indian J Pathol Micrology 1999; 42 : 311-15.
6. Jones DM, Lessells AM, Eldridge J. Campylobacter like organisms on the gastric mucosa:culture, histological, and serological studies. J Clin Pathol 1984; 37 : 1002-6.
7. Stolte M,Eidt S. Difference H Pylori associated gastritis in the antrum and body of the stomach. J Gastroentrol 1990; 28 : 229-23.
8. Auroux J, Lamarque D, Tankovic J, Benamouzig R, Mahe S, Chaumette MT, et al.Comparison of quantifying Helicobacter pylori gastric infection by culture, histology and C13 urea breath test. Gastroenterol Clin Biol 1998; 22 : 405-6.
9. Bayerdorffer E, Oertel H, Lehn N, Kasper G, Mannes GA, Sauerbruch T and Stolte M. Topographic association between active gastritis and Campylobacter pylori colonization. J Clin Pathol 1989; 42 : 834-39.
10. Wyatt JI, Rathbone BJ, Heatley RV. Local immune response to gastric Campylobacter in non ulcer gastritis. J Clin Pathol 1986; 39 : 863-70.
11. Shankar RR, Vikram K, Ananthakrishnan N, Harish BN, Jayanthi S. Erosive gastroduodenitis and Helicobacter pylori infection. Med Sci Monit 2003; 9 : CR222-4.
12. Villako K,Maards H,Tammur R. et al. Helicobacter (Campylobacter) pylori infestation and the development and progression of chronic gastritis: results of long-term follow up examination of a random sample. Endoscopy 1990; 22 : 114-17.
13. Marshall BJ, Warren JR. Unidentified curved bacteria in the stomach of patients with gastritis and peptic ulceration. Lancet 1984; 1 :1311-15.
14. Bonvicini F, Versura P, Pretolani S, Gasbarrini G, Laschi R. Scanning electron microscopy in the study of Campylobacter pylori associated gastritis Scanning Microsc 1989;3 (1) : 355-65; discussion 365-8.
15. Sarosiek J, Marshall BJ, Peura DA, Hoffman S, Feng T, McCallum RW. The impact of H.Pylori colonization in thickness of the gastroduodenal mucosa layer. Gastro Supplement 1991; 100 : A155.
16. Graham DY. Campylobacter pylori and peptic ulcer disease. Gastroenterology 1989; 96 : 615-25.
17. Suriani R, Venturini I, Actis GC, Rocca G, Rizzetto M, Cerutti E, et al. Effect of Helicobacter pylori eradication on bulbitis and duodenal gastric metaplasia. Hepatogastroenterology 2004; 51 : 176-80.
18. Ahmed N, Malur PR, Sankapal MN, Nagalotimath SJ.Prevalence of Campylobacter in H.Pylori non ulcerative dyspepsia. Indian J Pathol Microbiol 1991; 34 : 247-52.
19. Morson BC,Dawson IMP. In Gastrointestinal pathology 2nd Ed, stomach Blackwell scientific publication,1979.
20. Shousha S,Spiller RC,Parkins RA. The endoscopically abnormal duodenum in patients with dyspepsia: biopsy findings in 60 cases. Histopathology 1983; 7 : 23-24.

*Asso.Prof.; ***Resident Pathologist, Department of Pathology, Grant Medical College, Byculla, Mumbai. **Consultant Pathologist, West Bengal.