GASTROENTEROLOGY 1997;U3:S51-555

SECTION III:
ROLE OF HELICOBACTER PYLORI INFECTION IN CANCER

Helicobacter pylori, Inflammation, Mucosal Damage, and

Apoptosis: Pathogenesis and Definition of Gastric Atrophy

ROBERT M. GENTA
Veterans Affairs Medical Center and Baylor College of Medicine, Houston, Texas

Much of what is currently accepted on the natural
history of Helicobacter pylori-induced gastritis and its
relationship with gastric adenocarcinoma rests on the
assumption that atrophic gastritis can be correctly
identified and reproducibly recognized. Recently, sev-
eral stUdies have indicated that pathologists have a
low level of agreement on this topic, and the terms
"gastric atrophy" and "atrophic gastritis" remain impre-
cisely defined and, therefore, poorly understood. Fur-
thermore, the genesis and progression of the atrophic
changes taking place in the gastric mucosa of some,
but not all, subjects infected with H. pylori are incom-
pletely characterized. This review has three aims: (1) to
briefly reexamine our current knowledge of the mecha-
nisms involved in the injury and repair of gastric glands;
(2) to present a hypothesis on the development of
gastric atrophy; and (3) to propose a new, stringent
definition of gastriC atrophy that may be usefully
applied in the clinical research arena.
T he recognition of the carcinogenic potential of
Helicobacter pylori rests on the acceptance of the
paradigm that H. pylori-induced gastritis causes atrophic
gastritis, a condition believed to represent a precursor of
the intestinal-type gastric adenocarcinoma. l The founda-
tion of this model, originally proposed by Correa in 1975,
well before H. pylori was identified as a human pathogen,
is the assumption that atrophy and atrophic gastritis can
be correctly identified and reproducibly recognized. 2
Recently, several studies have indicated that even
experienced gastrointestinal pathologists have a low level
of agreement on the assessment of gastric atrophy. In
1994, 20 sets of glass slides chosen to represent a wide
variety of gastric inflammatory lesions were circulated to
an international group of gastric pathologists who later
convened in Houston, Texas, to consider an updated
version of the Sydney system. Each pathologist was asked
to grade the histopathologic features of each case accord-
ing to the published guidelines of the Sydney system 3 and
to generate a diagnosis that would reflect the individual
practice and preferences. Although a modest degree of
agreement was achieved on the degree of such features as
intensity of H. pylori infection, chronic inflammation, and
intestinal metaplasia, the level of agreement for atrophy,
as calculated by the K statistics, fell into the poor range. 4
Another study performed among dedicated gastrointestinal
pathologists and trainees in our institution showed an even
more dismal degree of interobserver concordance on atrophy,
whereas there was only minimal variation on the assessment of
intensity of infection and inflammatoty features. 5
In both studies summarized above, pathologists were
given carefully selected gastric biopsy specimens, chosen
not only because they represented certain features of
gastritis, but also because they were topographically
defined, well-oriented, and properly stained samples. Yet,
experienced pathologists interested in gastritis failed to
achieve a satisfactory level of agreement. A great many
published studies on gastritis and its relationship with
gastric cancer are based on the examination of archival
biopsy specimens, often of uncertain topographical prov-
enance, suboptimal size, inadequate orientation, and dubious
staining. Some of these studies have relied on the histopatho-
logic diagnoses rendered by the pathologist assigned to the
case (the so-called routine diagnosis). One does not have to be
particularly critical to suspect that data collected in this
fashion may suffer from profound observer bias, and, therefore,
some of the conclusions reached may be questionable.
Thus, for reproducible and profitable research on
atrophic gastritis to be carried out, two ctucial problems
need to be addressed and solved: (1) the formulation of a
universally accepted definition of gastric atrophy and (2)
the discovery of methods of evaluation that would ensure
minimal interobserver variation.
Abbreviation used in this paper: IL, interleukin.
© 1997 by the American Gastroenterological Association
0016-5085/97/$3.00
S52 ROBERT M. GENTA GASTROENTEROLOGY Vol. 113, No. 6

Two Views of Atrophy A

The recent publication of the updated Sydney
system for the classification and grading of gastritis 4 may
eventually help pathologists achieve a greater degree of

. .... " " ..... .. ..

interobserver agreement. Pilot studies have shown that
the use of visual analogue scales (drawings in which the
B ;. ::, ' . .~

..~lf~ ..U~~t:~:f.~:.~:
i:-·~~mtWI1
different scores of a histopathologic characteristic are

....~:-.......... .. ,-. .
schematically represented; originally designed by T.

-.... .........
.;...... ..... :.......
Karttunen of the University of Oulu, Finland) simplifies
the grading of biopsy specimens and fosters concordance ,.,
:,
,:
'"
.
"' .
;
',:
',,',.;.,.......... .
":,"
::

among pathologists. However, in the case of atrophy, the

visual analogue scales must be used in conjunction with a
stringent definition: the drawings depicting antral and
c
corpus atrophy show the progressive decrease of glandular
structures as the mucosa changes from normal to severely
atrophic. However, what takes the place of the missing
glands is not shown and, therefore, one of the major Figure 1. (A)Schematic representation of the normal gastric mu-
sources of confusion in the assessment of atrophy is not cosa, in which the glands occupy most of the lamina propria. If a
removed. dense mononuclear cell infiltrate occupies the mucosa, as happens
in a subset of subjects infected with H. pylori, (B)the glands seem to
As seen in a microscopic section, the normal gastric be sparse or completely "lost." Treatment ofthe infection will cause
mucosa consists of mucous and/or oxyntic glands sepa- the mucosa to return to normal; neither fibrosis nor intestinal
rated by little or no extracellular matrix with few metaplasia are present. Gastric mucosal atrophy, represented in G,
is characterized by (1) loss of glands, (2) fibrosis and thinning of the
mononuclear cells (Figure lA). If the mucosa in which
lamina propria, and/or (3) replacement of variable portions of the
the glands reside is densely infiltrated by inflammatory functional gastric epithelium with intestinalized metaplastic epithe·
cells, the glands are separated and pushed aside and may lium.
become invisible, or lost (Figure IB). Of course, when
glands have been destroyed and substitured with some
other tissue (e.g., metaplastic epithelium or fibroblasts step of atrophic gastntls has stimulated a practical
question: can the cure of H. pylori infection reverse
and extracellular matrix) they are also invisible, or lost
atrophic gastritis and, consequently, reduce the risk of
(Figure lC). Because most definitions of gastric atrophy
gastric carcinoma?9 If the histopathologic appearance
have emphasized gland loss without specifying what
depicted in Figure IB is accepted as atrophy, then the
replaces the missing glands, mucosa showing the histo-
answer is yes; the slow disappearance of the mononuclear
pathologic features depicted in either Figure IB or C has
cell infiltrate after the eradication of H. pylori is paralleled
been called atrophic. Gastric function studies do not help
by a reorganization of the mucosal structure, the reappear-
clarify the issue; inflammation in the oxyntic mucosa
ance of glands, and a slow return to normal acid
interferes with the acid production,6-S and, of course, the
destruction of oxyntic glands results in hypochlorhydria. production. However, if fibrosis and metaplasia are
considered essential elements of atrophy, then the cure of
As a consequence, these two types of gland loss (one
H. pylori will not result in glandular regeneration or
apparent and one real, but both with similar pathophysi-
functional recovery. Gastric mucosal fibrosis, similar to
ological consequences) have not been considered as
pulmonary fibrosis or cirrhosis of the liver, is an irrevers-
separate entities in studies that have addressed the
ible process.
relationship between structure and function of the gastric
The practical implications of this distinction are
mucosa. However, only atrophy characterized by fibrosis
substantial. For example, an article published in 1996 in
and intestinal metaplasia, and, therefore, genuine loss of
glands, has been convincingly associated with the devel- The New England Journal of Medicine concluded that
long-term therapy with proton pump inhibitors pro-
opment of gastric cancer.
motes the development of atrophy of the gastric corpus in
patients with H. pylori infection. 1o Because the term
Why It Is Crucial to Distinguish
atrophic gastritis (used but not defined in the article)
Atrophy From Corpus Gastritis evokes images of increased gastric cancer risk,l)l the
The acceptance of the paradigm that H. pylori possibility that antisecretory maintenance therapy might
predisposes to gastric cancer through the intermediate increase the risk of cancer caused sufficient concern for
December Supplement 1997
regulatory agencies in the United States to convene an ad
hoc panel to examine the issue. Prolonged and intense
acid suppression alters the topographical distribution of
H. pylori in the gastric mucosa; in these patients, larger
numbers of bacteria infect the corpus than the antrum
and increase the severity of corpus gastritis while antral
gastritis subsides. 12 Data and photomicrographs from
other works by the same investigators 13 suggest that a
broad definition of atrophy was used, one that accepts
that glands are to be separated and obliterated by
inflammation, but not necessarily destroyed and replaced
by fibrous tissue or metaplastic epithelium. In this case,
as in many other instances of discrepant data among
investigators, the lack of a commonly agreed upon
unequivocal definition of atrophy resulted in unnecessary
misunderstandings.
Injury and Repair:
The Determinants of Atrophy
H. pylori infection causes visible damage both to
the superficial epithelial cells and to the cells that form
pits and glands. 14 Cellular injury may be inflicted
directly by the bacteria, or by mechanisms mediated by
inflammation or apoptosis. Direct damage is believed to
occur in surface epithelial cells as a result of the
attachment of H. pylori, which induces actin polymeriza-
tion within the cells. Bacterial adhesion also stimulates
the production of tumor necrosis factor ex, interferon
gamma, interleukin (IL)-l, IL-6, and IL_8. 15 - 17 Both
histological and ultrastructural changes (flattening, mu-
cin depletion, cell dropout) have been well documented
in surface epithelial cells colonized by H. pylori, even in
the absence of inflammation. IS However, it is unknown
whether similar inflammation-independent mechanisms
are operative at the level of oxyntic or chief cells. Bacterial
attachment and superficial cell activation results in the
rapid recruitment of neutrophilic polymorphonuclear
cells into the gastric mucosa, soon followed by the
recruitment of lymphocytes, and later by plasma cells,
macrophages, and eosinophils. 15 - 24 Neutrophils are acti-
vated by IL-8 and may also be recruited by chemotactic
factors released directly by H. pylori that further augment
the oxidative burst and promote their migration across
the epithelium into the lumen. This constellation of
inflammatory changes results in injury to both the
superficial and the glandular mucosa, including the
oxyntic glands. Increased apoptosis has been described
recently in H. pylori-infected gastric mucosa and has been
proposed as a possible mechanism of epithelial dam-
age. 25 •26 However, methodological differences among the
studies and the paucity of dah concerning the apoptotic
GASTRIC ATROPHY: PATHOGENESIS AND DEFINITION S53
Gland
•..
Damage
'.
'. .........
Figure 2. Pathogenesis of atrophic gastritis. During H. pylori infec-
tion, gastric epithelium and glands are damaged by direct bacterial
toxicity, by the effects of inflammation, and possibly by increased
apoptosis. The development of atrophy depends on the predominant
pathway of repair, which can consist either of a continuing process
of restitutio ad integrum, during which the epithelium regenerates
and reacquires its normal function and structure, or of a replace-
ment by a mixture of intestinal metaplasia and fibrosis.
activity in glandular cells currently hinder our understand-
ing of the role of apoptosis in gastric mucosal damage and
in the genesis of atrophy.
Irrespective of the mechanism of damage, the gastric
mucosa may either regenerate and return to normal or
undergo adaptive reparative processes that lead to the
replacement of the normal functional epithelium. When
the destroyed glands fail to regenerate, the space they
previously occupied in the lamina propria may be
replaced by fibroblasts and extracellular matrix. The end
result of this process is an irreversible loss of functional
structure, which can be called atrophy. In another
pathway, often overlapping with the first one, the
specialized gastric glandular epithelial cells (oxyntic,
chief, and mucous cells) are substituted by an intestinal-
type epithelium containing goblet cells, intermediate,
so-called absorptive cells, and in most cases are lined by a
brush border. This replacement is known as intestinal
metaplasia. During chronic H. pylori infection, all three of
these types of repair occur, the respective proportion of
each probably being modulated by environmental, ge-
netic, and bacterial factors. Figure 2 represents schemati-
cally two possible scenarios of damage and repair.
In the first scenario, the stomach responds to injury
with a continuing process of restitutio ad integrum, the
gastric mucosa remains hospitable for H. pylori, and
chronic, active, nonatrophic gastritis persists indefinitely.
In the second scenario, the progressive replacement of the
glands by fibrosis and intestinalized epithelium results in
the ablation of the gastric acid production, the gastric
S54 ROBERT M. GENTA
t.
Figure 3. In a portion of infected subjects, the size of which varies
greatly in different populations, atrophy and metaplasia develop in
patches on a background of pangastritis. Over a period of years, the
patches may extend and become confluent. This constellation of
features has been called multifocal atrophic gastritis, and repre·
sents the nosological entity that is biologically and epidemiologi-
cally associated with intestinal-type adenocarcinoma of the stom-
ach .
mucosa becomes increasingly more inhospitable for H.
pylori, and the inflammation in the fibrotic and metaplas-
tic areas progressively wanes. The process continues in
areas of persisting infection and inflammation, and the
stomach acquires a mixed pattern consisting of architec-
turally normal inflamed areas alternating with expanding
patches of atrophy and metaplasia (Figure 3). This
constellation of features characterizes multi focal atrophic
gastritis, the nosological entity that is biologically and
epidemiologically associated with intestinal-type adeno-
carcinoma of the stomach. 1,4
A Strict Definition
The main purpose of accurately diagnosing atro-
phic gastritis is to assess the potential risk for the
development of adenocarcinoma. Because only replace-
ment of glands with fibrotic and metaplastic tissue is
associated with gastric cancer, and the other type of gland
loss is likely the result of a reversible inflammatory
phenomenon, a useful definition of atrophy would incor-
porate all these elements. By defining atrophy as "the
irreversible loss of gastric glands with replacement by
metaplastic epithelium or fibrosis," we would avoid
confusion with the apparent gland loss caused by the
expansion of the inflammatory component of the lamina
propria. The latter phenomenon, as with other inflamma-
tory processes, can be reversed by the removal of the
causative agent.
How can reversibility be predicted? When a large
proportion of the glandular and surface epithelium is
metaplastic, the lamina propria is replaced by extracellu-
lar matrix and fibroblasts, and there is scanty inflamma-
tion, atrophy can be diagnosed confidently using the
definition provided. If, on the other hand, a dense
inflammatory infiltrate occupies the lamina propria,
irreversibility can only be diagnosed a posteriori, by
GASTROENTEROLOGY Vol. 113, No.6
exammmg gastric biopsy specimens at least 6 months
after H. pylori infection has been cured. After the
inflammation has abated, it becomes possible to assess the
status of the glandular component; it may be normal,
reduced, or completely absent and replaced by fibrous
tissue. With the advantage of a view unobstructed by
inflammation, it will then be possible to apply the visual
analogue scales to evaluate the extent of atrophy.
This approach has the disadvantage that one has to
wait several months after treatment before being able to
evaluate atrophy. However, it is impossible to have a
satisfactory appreciation of the architecture of the gastric
mucosa unless the inflammation is first removed. A
similar situation exists in the assessment of epithelial
dysplasia in inflammatory bowel disease. Usually, the
absence of dysplasia and the presence of high-grade
dysplasia can be confidently diagnosed in a biopsy
specimen, irrespective of the intensity of accompanying
inflammation. However, when the aegree of cellular
atypia is mild or moderate and the colonic mucosa is
severely inflamed, it is virtually impossible to determine
whether the epithelial cellular changes are due to regen-
eration and, therefore, are reversible, or whether they are
due to an incipient neoplastic process. In such instances,
the pathologist can only issue a tentative diagnosis of
"indefinite for dysplasia" and request that a new set of
biopsy specimens be sent after the patient has received
appropriate anti-inflammatory therapy. 27 A similar strat-
egy could be used for the diagnosis and grading of
atrophy.
Conclusion
Once pathologists agree on a set of criteria for the
diagnosis of atrophy, the histological characteristics are
easy to detect. Both reticulin and trichrome stains
improve the visualization of fibrosis; intestinal metapla-
sia can be easily detected by adding alcian blue at pH 2.5
to the routine H&E stain . An even better method to
reliably assess atrophy is provided by the triple stain
known as the Rubin stain,28 which combines H&E, alcian
blue at pH 2.5, and saffron; fibrous tissue stains light
okra, intestinal metaplasia bright blue, and the other
features of the gastric mucosa are stained as in a usual
H&E.
The definition of atrophy proposed above has the
advantage of being simple and straightforward. Its
usefulness needs to be tested, as does its predictive
component. Future research may suggest that small
rectifications or radical changes are necessary. However,
even if its only function should be that of stimulating an
December Supplement 1997
interest in reaching a workable definition of atrophy, it
will have achieved its major goal.
References
1. Correa P. Helicobacter pylori and gastric carcinogenesis. Am J
Surg Pathol 1995;19(Suppl 1):S37-S43.
2. Genta RM. Recognizing atrophy: another step toward a classifi-
cation of gastritis. Am J Surg Pathol 1996;20(Suppl 1):S23-
S30.
3. Price AB. The Sydney System: histological division. J Gastroen-
terol HepatoI1991;6:209-222.
4. Dixon MF, Genta RM, Yardley JH, Correa P, for the Houston
International Workshop on Gastritis. Classification and grading
of gastritis: the updated Sydney system. Am J Surg Pathol
1996;20:1161-1181.
5. EI-Zimaity HMT, Graham DY, AI Assi MT, Malary H, Karttunen TJ,
Graham DP, Huberman RHM, Genta RM. Interobserver variation
in the histopathological assessment of Helicobacter pylori
gastritis. Hum PathoI1996;27:35-41.
6. Sipponen P, Hyvarinen H, Siurala M. H. pylori corpus gastritis:
relation to acid output. J Physiol PharmacoI1996;47:151-159.
7. Cal am J. The somatostatin-gastrin link of Helicobacter pylori
infection. Ann Med 1995;27:569-573.
8. EI-Omar EM, Penman 10, Ardill JE, Chittajallu RS, Howie C,
McColl KE. Helicobacter pylori infection and abnormalities of
acid secretion in patients with duodenal ulcer disease. Gastroen-
terology 1995;109:681-691.
9. Genta RM, Segura AM. Will curing Helicobacter pylori eliminate
gastric cancer? Adv Anat PathoI1996;3:228-232.
10. Kuipers EJ, Lundell L, Klinkenberg-Knol EC, Havu N, Festen HP,
Liedman B, Lamers CB, Jansen JB, Dalenback J, Snel P, Nelis
GF, Meuwissen SG. Atrophic gastritis and Helicobacter pylori
infection in patients with reflux esophagitis treated with omepra-
zole or fundoplication. N Engl J Med 1996;334:1018-1022.
11. Correa P, Miller MJ. Helicobacter pylori and gastric atrophy:
cancer paradoxes. J Natl Cancer Inst 1995;87:1731-1732.
12. Graham DY, Genta RM, Evans DG, Reddy R, Clarridge JE, Olson
AC, Edmonds AL, Siepman N. Helicobacter pylori does not
migrate from the antrum to the corpus in response to omepra-
zole. Am J GastroenteroI1996;91:2120-2124.
13. Kuipers EJ, Uyterlinde AM, Pena AS, Hazenberg HJ, Bloemena E,
Lindeman J, Klinkenberg-Knol EC, Meuwissen SG. Increase of
Helicobacter pylori-associated corpus gastritis during acid
suppressive therapy: implication for long-term safety. Am J
GastroenteroI1995;90:1401-1406.
14. Genta RM. Helicobacter pylori infection in gastric pathology. In:
Weinstein RS, ed. Advances in pathology and laboratory medi-
cine. Volume 7. St. Louis, MO: Mosby-Year Book, 1994:443-
465.
15. Smoot DT, Resau JH, Naab T, Desbordes BC, Gilliam T, Bull-
Henry K, Curry SB, Nidiry J, Sewchund J, Mills-Robertson K.
Adherence of Helicobacter pylori to cultured human gastric
epithelial cells. Infect Immun 1993;61:350-355.
16. Dytoc M, Gold B, Louie M, Huesca M, Fedorko L, Crowe S,
Lingwood C, Brunton J, Sherman P. Comparison of Helicobacter
GASTRIC ATROPHY: PATHOGENESIS AND DEFINITION S55
pylori and attaching-effacing Escherichia coli adhesion to eukary-
otic cells. Infect Immun 1993;61:448-456.
17. Fiocca R, Luinetti 0, Villani L, Chiaravalli AM, Capella C, Solcia
E. Epithelial cytotoxicity, immune responses, and inflammatory
components of Helicobacter pylori gastritis. Scand J Gastroen-
teroI1994;205(Suppl):11-21.
18. Chen XG, Correa P, Offerhaus J, Rodriguez E, Janney F, Hoff-
mann E, Fox J, Hunter F, Diavolitsis S. Ultrastructure of the
gastric mucosa harboring Campylobacter-like organisms. Am J
Clin PathoI1986;86:575-582.
19. Mooney C, Keenan J, Munster 0, Wilson I, Allardyce R, Bagshaw
P, Chapman B, Chadwick V. Neutrophil activation by Helicobac-
ter pylori. Gut 1991;32:853-857.
20. Fan XG, Chua A, Fan XJ, Keeling pw. Increased gastric produc-
tion of interleukin-8 and tumour necrosis factor in patients with
Helicobacter pylori infection. J Clin PathoI1995;48:133-136.
21. Sharma SA, Tummuru MK, Miller GG, Blaser MJ. Interleukin-8
response of gastric epithelial cell lines to Helicobacter pylori
stimulation in vitro. Infect Immun 1995;63:1681-1687.
22. Karttunen R, Karttunen T, Ekre HP, MacDonald n. Interferon
gamma and interleukin-4 secreting cells in the gastric antrum in
Helicobacter pylori-positive and -negative gastritis. Gut 1995;
36:341-345.
23. Crabtree JE, Shallcross TM, Heatley RV, Wyatt JI. Mucosal
tumour necrosis factor alpha and interleukin-6 in patients with
Helicobacter pylori-associated gastritiS. Gut 1991;32:1473-
1477.
24. Mai UE, Perez-Perez GI, Allen JB, Wahl SM, Blaser MJ, Smith
PD. Surface proteins from Helicobacter pylori exhibit chemotac-
tic activity for human leukocytes and are present in gastric
mucosa. J Exp Med 1992;175:517-525.
25. Mannick EE, Bravo LE, Zarama G, Realpe JL, Zhang XJ, Ruiz B,
Fontham ET, Mera R, Miller MJ, Correa P. Inducible nitric oxide
synthase, nitrotyrosine, and apoptosis in Helicobacter pylori
gastritis: effect of antibiotics and antioxidants. Cancer Res
1996;56:3238-3243.
26. Moss SF, Calam J, Agarwal B, Wang S, Holt PRo Induction of
gastric epithelial apoptosis by Helicobacter pylori. Gut 1996;38:
498-501.
27. Riddell RH, Goldman H, Ransohoff DF, Appelman HD, Fenoglio
CM, Haggitt RC, Ahren C, Correa P, Hamilton SR, Morson BC.
Dysplasia in inflammatory bowel disease: standardized classifi-
cation with provisional clinical applications. Hum Pathol 1983;
14:931-968.
28. Levine DS, Surawicz CM, Ajer TN, Dean PJ, Rubin CEo Diffuse
excess mucosal collagen in rectal biopsies facilitates differen-
tial diagnosis of solitary rectal ulcer syndrome from other
inflammatory bowel diseases. Dig Dis Sci 1988;33:1345-
1352.
Received June 30, 1997. Accepted July 22, 1997.
Address requests for reprints to: Robert M. Genta, M.D., Depart-
ment of Pathology - 113, Veterans Affairs Medical Center, 2002
Holcombe Boulevard, Houston, Texas 77030. Fax: (713) 794-7810;
e-mail: rmgenta@bcm.tmc.edu.