: Terminology, Etiology, and Clinicopathological Correlations: Another Biased View

HENRY D. APPELMAN, MD

Thehistological approach to gastritis, especially the chronic forms, these, in any combination or in no combination at all, has undergone a series of re-evaluations by different experts over may or may not have anything to do with the gastritis the past decade, mainly because of the recognition of individual of. the patient and the gastritis of the endoscopist.‘-” To disease patterns that have specific clinical and epidemiological impli- be perfectly frank the correlation between histological cations. The most spectacular of these was the discovery of Helico- and endoscopic gastritis is terrible! When the endoscop- backrpylmi and its common gastritis, its relation to almost all duode- ist sees spectacular inflammatory changes, even with nal peptic ulcers and to most gastric peptic ulcers, its potential as a precursor of first multifocal atrophic gastritis and later tubule-form- erosions, a biopsy specimen of such an area, even of an ing gastric carcinomas, and its status as a cause of gastric mucosal erosion, is almost as likely to be normal as inflamed.“,” lymphomas. During this same decade other classes of gastric reaction Whatever it is that the endoscopists see often has no and have been recognized, including chemical injury histological counterpart. Furthermore, a biopsy speci- and lymphocytic gastritis. Also in the same decade the importance of men from an endoscopically normal antrum may be non-steroidal anti-inflammatory drugs (NSAIDs) has emerged as a intensely and actively inflamed. In contrast, the endo- cause of gastric mucosal injuries. To add emphasis to all these discov- scopic histological correlation for lumps, masses, or big eries, biopsies are being performed on stomachs in almost epidemic ulcers is very good. No wonder there have been a bewil- numbers and each biopsy specimen has the potential of having the dering number of classifications of gastritis, some of features of one or more of these injuries as well as injuries that which attempt to relate endoscopic and histological al- have yet to be described. To cope with tbis rapidly expanding gastric terations as well as others that make no pretense at inflammatory informational extravaganza, pathologists need some way of dealing with the various entities comfortably and some method correlation. Some classifications are named after peo- of cataloging them in ways that are understandable both to them and ple, such as Whitehead, whereas others are named after to the endoscopists with whom they work. However, if emerging data cities, such as Sidney.‘.’ Some use letters to designate about the chronic gastritides are correct, it is conceivable that the types of gastritis, giving us types A. B, AB, and C to need to diagnose them, from a strictly clinical standpoint, is limited. play with, whereas others skip letters and use words Either we may know what is in the biopsy specimen before we see it instead.“,” It is also no wonder that any clinical classifi- or what we see may not be important, although it may be intellectually cation of gastritis is likely to be of little value, consider- challenging. HUM PATHOL 25:1006-1019. Copyright 0 1994 by W.B. ing the lack of correlation between symptoms, endo- Saunders Company scopic findings, and histological features. Nevertheless, we pathologists must contend with “Gastritis” should not be a nasty word, yet it has gastric inflammatory disease, much as we must deal with become one. It is a word that often seems mysterious, inflammations of the and colon. Our clinical a word with too many conflicting meanings, each of colleagues continue to perform biopsies on stomachs, which is dependent on who defines it. For the patient hoping that we will offer them some help with their gastritis conjures up images of acute epigastric peptic patients who have unpleasant upper gut symptoms or ulcer-type pain, sometimes accompanied by striking endoscopic findings. To begin, we have to de- and vomiting, perhaps after an episode of overindul- mystify the word gastritis. This is best accomplished by gence with food and drink. Other patients’ gastritis ac- developing a classification system that is simple and counts for their constant low grade upper abdominal easily applicable to the types of inflammations that we gnawing pain that only temporarily goes away with over- see regularly in our gastric biopsy specimens. In the the-counter antacids and then comes back. There is past 6 years several classifications for gastritis have ap- endoscopic gastritis, the erythematous, edematous, fria- peared in independent articles, chapters in books, and ble, slightly eroded mucosa found in stomachs of pa- gastrointestinal pathology texts (Table 1). All of them tients who have the symptoms referred to above and, include the common gastritides, although the names peculiarly, also in the stomachs of those who are asymp- vary.!‘-“L Furthermore, much of the descriptive work in tomatic. Finally, for the pathologist gastritis is an histological gastritis is still being published and, in addi- annoying collection of pits, glands, acute and chronic tion, it is critical that a classification scheme recognize forms of , metaplasias, and bacteria, and that there are inflammations that do not fit into prede- termined categories; in other words, there are gastriti- des that have not yet been named. Therefore, any classi- From the Department of Pathology, Univeuity of Michigan, Ann Arhw, MI. Accepted fol- publication April 18, 1994. fication system must allow for periodic revisions. Such Kq 7ourrl,:aastl-itis, helicohactel-. atrophic. lymphorytic. chemical. a system also should recognize that more than one type Address car-rcspontlcr~c and reprint xquests to Hemy D. Appel- of gastritis will occasionally occur in stomachs, so that man. MD, Dcpxtmrnt of. Patholc~h~. Univenity of. Michigan Hospi- there may be double or triple gdstritides. tals, I.500E Mdical Center- Dr. Ann Arbor. MI 48109-0054. Copyright 61 1994 hy W.B. Saundcr~ Company We have known for years that there are acute gas- 0046817~,‘94/25 IO-000.5$5.00/l, tric injuries and inflammations that include small super-

1006 GASTRITIS: TERMINOLOGY AND CORRELATIONS (Henry D. Appelman)

TABLE 1. Recent Classifications of Most of the Common Chronic Gastritides

owen”~ Yardlw”’ Dixon”’ Goldman” I.ewin et al” 1989 1990 1990 1992 I992

(Chronic nonspecific H pylori gastritis Type B gastritis (:hronic antral Nonerosivr, nonspecific gantritia. tvpe B gastritis gastritis, H pvlori- associated gastritis whtypc (:hronic nonspecific Mrtaplastic atrophic No specific categon (:hronic antral Notxxc6ivr. nonspwitic gastritis. tape B gastritis. tvpe B gastritis gastritis. no specific [environmental] cuht~pe (%ronic nonspecific Mrtaplastic atrophic Typr A chronic <:hronic ftmdic N~mer&ive. nonspecific gastritis. type A gastritis. type A gastritis gastritis gastritis. severe [autoimmune] atrophic fundic gland gdstritit wht\pe No \pccitic categon Lytnphocytic Lymphocytic gastritis Lymphocytic gastritis Lvmphocvtic gnstritis gastritis [a type of chronic erosive gastritis] (Ihrmical qastritis Type (: gastritis No specific categon

ficial ulcers. or erosions, with no plasma cells, scars, or exist, although new discoveries seem to relate some of’ granulation tissue to indicate chronicity. Sometimes, them. Unfortunately, the more we know about the these injuries lead to extensive superficial mucosal ne- chronic gastritides, the more names we invent for the crosis accompanied by hemorrhage. These reactions same diseases. The practicing pathologist must choose are the acute types of gastritis, and they have been given which names he or she will use routinely, and that pathol- names like “acute erosive gastritis” or “acute hemor- ogist must ensure that the endoscopists understand what rhage gastritis.” Precipitating injuries include alcohol, the names mean. It is silly to adopt a nomenclature that steroids, .aspirin, and other non-steroidal anti-inflam- has no meaning for the very people on whose business matory drugs (NSAIDs). Such acute processes have not we depend for our livelihood! The following discussion been part of the gastritis mystery unless theywere super- is based on a simple, functional classification of gastritis, imposed ton a chronic inflammation. These will be dis- concentrating mostly on the chronic fijrms, that has cussed in more detail below. proven useful in a busy endoscopic biopsy practice in a It is the variety of inflammations with overtones of large teaching hospital (Table 2). It incorporates many chronicity that is the problem. In the past it was as- attributes of other classifications, hopefully the hest ones, sumed that chronic gastritis was a single disease with a and thus is not a fancifi4 invention of the author. In collectioir of- histological characteristics that were pres- fact, it is really a simplified version of the classifications c’nt in various combinations. Different forms of chronic proposed by Correa and Yardley.“,“’ It affirms the exis- gastritis differed only in the intensities of these histolog- tence of several distinct entities, but it offers several ical characteristics and in their mix.’ Furthermore, the choices of names for most entities, recognizing that dif- was looked on as a single organ: yet we know that thcrr arc several distinctive mucosae, as different Irom each other as the mucosa of the TABLE 2. A Simplified Classification of Gastritis differs from that of’ the colon. This approach was aided for 1994 by the fact that no studies identified a precursor lesion and svstematically followed it with sequential biopsies Acutr through intermediate stages to the full-blown end-stage Erosive. hemorrhag-ic: some of thesr are 11,,I inllnmmations hur Gmplv hemorrhages chronic gastritis. As a matter of fact. this giant hole in (:hronic our knowledge and understanding persists because we H&cobnctrr Pylon qpe (1,000 synonyms. includitlg chronic activr. still have no proof of how each chronic gastritis evolves chronic n&specific. diffuse’ antral) and what the precursor or precursors are. A frequent Atrophic suggestion implicates a ubiquitous mild inflammation Type A. atttoimtnune, dithw body. ditlusr uxporal Type B, tlotl-autoimmune, multifocal. ~rlvir,,llrlrctlt,tl known simply as “superficial gastritis” as the common Lymphocytic. Fprue-like. Menetrier-like precursor. hut there is no proven support for that con- Focal and miscellaneous, diagnostic and Iluntliqnostlc tention, Ionlv some questionable evidence based on pop- (Irohn’s disease of the stomach ulation stuciies.‘~“~” However, this should not deter us. (;r;~n~~lomatous. not (:rohn’c M’c also havcb no idea of the histological evolution of Allergic Specific infectious for which agents I~aw twen idwtitivd most otl1c.r chronic inflammations of the gut, including Miscellaneous. including pimples: thaw gastritides that d(~ not ulcerative . , and Crohn’s dis- fit into the other categories. The numhvr of thew i\ in ease. Recent discoveries, aided by the intense use of direct proportion to the number of Kastric biopul upper endoscopy and the frequent biopsy specimens specimens. (Xemical gastropathies: those induced hy hilcvrfluu. NSAID,. obtained by our endoscopic colleagues. have given ~1s and other direct surface-damaging agent>. insights into the intricacies of the chronic gastritides. We now recognize that several separate chronic entities NOTE. Thr rntitirs are important: thr ndme~ WC ncqotiablt~

1007 HUMAN PATHOLOGY Volume 25, No. 10 (October 1994)

FIGURE 1. and its gastritis. (A) The organisms are the slightly curved rods clustered close to the epithelium in this superficial pit. (Warthin-Starry: original magnification x825.) (B) The typical low power appearance of the active and chronic antral gastritis of H pylori. The mucosa appears dark because of an infiltrate of plasma cells in the superficial and mid lamina propria that spread apart the tubules. There is a small crushed lymphoid nodule at the base. The pits are elongated and tortuous. (HE; original magnification x53.) (C) Prominent lymphoid hyperplasia is an integral part of this gastritis. (HE; original magnification x83.) (D) The mid mucosa of intense H pylori gastritis has dense plasmacytosis. The tubule with neutrophils migrating through its epithelium is a neck, (HE; original magnification x330.)

ferent names are used by different pathologists, all of gastritis that began approximately a week after inges- whom are capable and accomplished. This classification tion.” At this time we are in the midst of an information system can be used easily when more than one type of explosion about H pylon’ and its epidemiological, histo- gastritis occurs in a single stomach, each of which may logical, genetic, and cancer-associated aspects. There have an independent cause or epidemiological associa- are studies of prevalence in different populations and tion. The following discussion analyzes each item, not in different diseases, some using histological diagnosis necessarily in its order in the Table but in order of its as proof of infection, whereas others use serum antibod- importance in the evolution of the classification system. ies as evidence of prior exposure, Whole books have been dedicated to the organism and its praises have been sung during week-long symposia. Obviously, the HfL/CO5ACTfR PYLO/?/ AND ITS GASTRITIS information presented here must be an abbreviated ver- sion of what is known. However, even this shortened For decades chronic gastritis was more of a pain version of the H pylon’ story must by necessity dominate in the neck than in the stomach. Then approximately this entire article because of the importance of the or- 10 years ago Marshall and Warren gave us the great ganism. gastritis gift, their discovery that one type of chronic Briefly, H pyhri is a worldwide pathogen that has gastritis was an infectious disease, and the responsible different infection rates in different populations, the organism was a bacillus that now goes by the name of greatest rates occurring in developing countries where Helirobarter pylo?-i (Fig 1A) “~“’ In fact. Dr Marshall even infection usually occurs in childhood.” In developed infected himself with H p$on’ and developed an acute countries the infection rate increases with age, is more

1008 GASTRITIS: TERMINOLOGY AND CORRELATIONS (Henry D. Appelman) common in lower socioeconomic populations, and is men, and in such cases some pathologists have resorted greater in men than in women. Thus, there is a huge to performing either Giemsa or Wdrthin-Starrv stains population of asymptomatic infected individuals who routinely on every gastric biopsy specimen. not only h:ave the infection but also have its associated The: lymphoid nodules and the plasma cells are gastritis.“’ Recent studies have cast doubt on adult ac- probably the result of chronic antigenic stimulation by quisition of the infection and suggest that now in West- the organisms. Genta et al”” found that all individuals ern societises H fgh~ri is usually an infection acquired in with Hpyloti infection, with or without gastric 01 duode- childhood and that it is decreasing in frequency.“’ nal ulcers, had lvmphoid nodules that were more com- It was well known that duodenal peptic ulcers were mon in the antrum than in the bodv, on the lesser almost always accompanied by a significant chronic gas- rather than the greater curvature, and in ulcer patients tritis that involved the antral mucosa.21’” Furthermore, rather than in those without ulcers. Treatment of the this gastril.is was characterized by three components: H j?llloriinfection resulted in gradual shrinkage of the ( 1) diffuse plasmacytosis of the superficial mucosa, (2) nodules. It has been our experience that the density of Iymphoid nodllles deep in the mucosa, and (3) neutro- the plasma cells also diminishes and there is gradual phils invading the proliferative zone of the mucosa, loss of those plasma cells deep in the mucosa.‘” the neck Iregion (Figs 1B to lD).“.‘)‘) In addition, pit The neutrophils generally invade epithelium that hyperplasia in which the pits were longer than normal is mid-zone in the mucosa, at the level of the necks, and often mucin-depleted was common. This complex whereas the infection is colonizing the surface and SW was called “hypersecretory gastritis” by,Correa.” It also perficial pits. This fact is repeatedly- illustrated in essen- involved the body mLlCOS& especially if the ulcers were tially every publication in which there is a microscopic in the stomach instead of the . In such cases picture. The epithelium covered by the organisms is the gastritis occurred proximally to the ulcer as well as never inflamed, whereas inflamed epithelium is never in the antrum so that the more proximal the ulcer, the covered by organisms. This discrepancy in location be- more extensive the gastritis.‘,” It even can involve the tween bacteria and neutrophils has not been adequatcl) cardia that has a mucosa much like antral. Subse- explained. Soluble chemotactic factors, either released qtlently, this gastritis has been given several different by the bacteria or by their interactions with host tissues, names, including “diffuse antral gastritis,” “type B have been suspected. In patients infected with H p$ori chronic gastritis.” “chronic active gastritis,” and tumor necrosis factor [TNF]-alpha production bv antral “chronic nonspecific gastritis.““,“,” mucosa was higher when there was active gastriiis than The organisms are found on the surface epithelial when the gastritis was inactive.‘” Tumor necrosis factor- cells and less commonly on the superficial foveolar cells. alpha can activate neutrophils. Other neutrophilic che- In both locations they lie beneath the muco~ls coat, motactic factors have been identified that appear to mostlv over intercellular junctions. They appear as come from the bacterium itself.‘7.‘x slightly curved short rods. Ultrastructurally thev have There are modifications of the typical changes. For four flagella at one pole. In almost all cases the brgan- instance, the plasmacytosis may be accompanied by lym- ihms are visible with hematoxylin-eosin (HE). The phocytosis and even ?.eosinophils, althongh when the) (iiernsa stain will darken them and silver stains, such as ale present the possibility of a second insult might be the W’arthin-Starry, will both darken and thicken them considered. The infiltrate in the lamina propria also (Fig 1) . A. tissue &am stain also will make them more commonly extends into the glandular compartment fol obvious. If all else fails there are commerciallv available unpredictable distances, sometimes reaching the mns- antibodies for immunohistochemical de&on. How cularis mucosae. Perhaps this deep extension is an in- fhstidious a pathologist should be in looking for H&hi dex of severity. The combination of a deeply extending is a matter of‘ individual choice. Our approach has been or transmucosal infiltrate plus deep lymphoid nodules to use HE only because it has been our experience results in a histological lesion in which the glands arc that a careful search of the HE-stained slide will detect either spread apart or else are destroyed bv the in- bacteria in almost every case; special stains add little, flammation: in other words, there is loss of glands 01 perhaps increasing the Yield by no more than 1%. If atrophv. Unfortunately, we do not know which of these the gastritis is totally typical and if we cannot find the is the correct interpretation. It’ this is indeed destruc- bacteria, then we use the Warthin-Starry stain. This situ- tion of glands then we should have identified a gland- ation occllrs less than once a month, and we resort to destructive phase of the inflammation, yet such a phase the special stain mostly out of f-rustration at not finding has not been found. The active epithelial damage with the bugs in the expected setting. Usually, the Warthin- neutrophils involves the necks. Because this is the gen- Starry stain is also negative in these cases. In contrast, erative zone of the from which both pits some studies have reported a significant infection rate and glands arisr, it is possible that darnnge to this zone in patients with normal antral biopsy specimens using results in selective loss of regenerative capabilities for the specinl stains or culture or both.“’ We have no idea the glandular compartment. If such is the case then what it means to have H p$ori with no disease, but gland loss may be attributed to attrition without replacc- perhaps it indicates first infektion or a nonpathogenic ment. strain. In addition. pressures from outside the pathol- H&cohmrtu?-infection also has been reported to be 0~ department may dictate the intensity with which wt. associated with damage to the colonized surface and look for .Hpvlori. The gastroenterologists with whom we foveolar cpithelial cells in patients with both duodenal work mav feel it is important for them to know whether and gastric peptic ulcers or with the tvpical active H @!cJ~? ii; present or absent in every gastric biopsy speci- chronic gastritis without ulcers.-“‘.“” Thc,se changes in-

1009 HUMAN PATHOLOGY Volume 25, No. 10 (October 1994)

elude irregularities of the apical cytoplasm with indis- reported in Barrett’s mucosa and in ectopic gastric mu- tinct cell borders and loss of apical mucin in almost cosa in Meckel’s diverticula and in the .‘“.‘” all cases. Collections of such cells with loss of apical When Barrett’s epithelium is infected the gastric mu- cytoplasm produce surface mucosal depressions or pits. cosa is always infected as well, suggesting that the Bar- How such changes relate to the ulcers is not clear be- rett’s epithelium became colonized as a result of reflux. cause many of the ulcers are in the duodenum. The status of the gastric mucosa in the Meckel’s cases The evolution of the typical chronic and active dif- was not known, but in a report of H pylon’ in ectopic fuse antral gastritis that results from H pyloti infection gastric mucosa in the rectum the stomach also was in- is not known. As with other forms of chronic gastritis, fected.‘* a superficial inflammatory phase has been postulated The role of Hpylori in atrophic gastritis and intesti- in which the superficial mucosa, that part including the nal metaplasia has been studied in several diverse envi- pits and necks and surrounding lamina propria but not ronments. Data are accumulating rapidly regarding the the glands, contains excess plasma cells and/or lympho- associations of Hpylori with carcinoma of the stomach.“” cytes. Supposedly, this reaction extends deeper into the It has been shown that the risk of gastric carcinoma is mucosa and at some point the lymphoid nodules de- increased in populations with high infection preva- velop, but this theory has no proof. The acute infection, lences, and this relationship is almost linear.“” This ma) such as the one Marshall gave himself, may evolve into explain, at least in part, the high incidence of gastric the typical chronic gastritis if the bacteria are not eradi- cancer in Japan where H pvloti infection seems to be cated or cleared.g’ unusually common in olderadults. Scores of published Helicobncter pylori is a strange gastroenteric organ- studies are appearing monthly linking H pvlori with gas- ism. It does not like stagnant stomachs, but it seems to tric cancer, and possibly with certain histological types, prefer those stomachs in which there is normal motility. but this may depend on the country in which the can- In a study of , that is, stomachs with slow cers were studied. For instance, in a study from Japan emptying, H pylori was rarely found compared with culture-proven H pylori infection occurred in 79% of stomachs that emptied normally.?” It does not seem to patients with differentiated, or tubule-forming, early thrive in less acidic environments, such as in bile reflux. gastric carcinomas but only in 29% of patients with In one study 83% of patients with peptic ulcers had H undifferentiated, including signet ring cell. early carci- pylori before gastric operations, mostly partial gastrecto- nomas.“” In another studv from the United States H mies. After the operations that led to bile reflux the pyloriwas found in nonmalignant gastric mucosa in 890/o infection rate decreased to 54%, but after Roux-en-Y of patients with tubule-forming carcinomas but in onl) biliary diversion that prevented bile reflux the infection 32% of patients with diffLlse carcinomas.“’ Further- rate increased to 920/o.“’ more, it appears that Hpvlori is associated with noncar- Is H pylon infection and its resultant gastritis a disc carcinomas rather than with those that arise in the cause of symptoms, such as nonulcer dyspepsia? The cardia.‘” In contrast, in a study from Singapore there data are conflicting. In one study published before was no difference in the rates of histological H pylon’ much was known about H pylon’, active inflammation, infection between patients with tubule-forming and-dif- that is, neutrophils, in biopsy specimens of either gastric fuse carcinomas.“” Similar results were reported from body, antrum, or duodenum was more likely to be Finland where there was no difference in frequency of found in dyspeptic compared with nondyspeptic pa- antibodies to Hpvlori and in Hpylori-associated gastritis tients.” In another study published after H pyloti was in patients with both patterns of gastric carcinoma.“’ a well-known pathogen, nonulcer dyspepsia, whether As of the time of publication of this article, it is safe to associated with H pylori gastritis or with no histological say that the exact role of H pvlori in causing gastric inflammation and no bacteria, responded equally to a carcinoma is still not settled. Piesumably it is through regimen directed at eradication of the organism.” In its relation with multifocal atrophic gastritis that Hp~j1or-i still a third study of patients with nonulcer dyspepsia and gastric cancer also are related. This is discussed who were infected with H pvlori, specific treatment di- in more detail in the discussion of multifocal atrophic rected against the organism eradicated it and improved gastritis. the associated gastritis but was not significantly better Chronic Hpyloti infection with its induction of gas- than placebo in improving symptoms.“’ tric lymphoid tissue has been implicated as a precursor Helicobncter pylori colonizes metaplastic gastric sur- of gastric lymphoma of the unique B-cell type that arises face cells in the duodenum, and this rnav be important from mucosd-associated lymphoid tissue (MALT) .” In if it is a cause of duodenal peptic ulcer.“’ However, the fact there have been attempts to treat some cases of the exact mechanism by which H pylon‘ infection leads to low grade variant with anti- Hpyloti regimens, and there duodenal ulcers is not established. Nevertheless, the has been some success, at least for the limited follow- association is clear. Close to 1OOYo of patients with duo- up period.” Nevertheless, the possibility of’ treating a denal peptic ulcers have H pylon’ and its gastritis in the malignancy with antibiotics is exciting but also unset- antrum. Eradication of the organism dramatically re- tling! duces ulcer recurrence. Why only a tiny fraction of indi- Hdicobactrr pyhri and NSAIDs may interact or po- viduals with this incredibly common infection develop tentiate each other’s injury, although the published peptic ulcers is not known, but rrhaps it is related to data are not in total agreement. For instance, among specific strains of the organism:” nonulcer patients taking NSAIDs who developed endo- Hdirobnrterp$oti also colonizes gastric type surface scopic mucosal damage, H pvlo~i infection was much epithelial cells almost anywhere else. Infection has been more common than among those who had normal en-

1010 GASTRITIS: TERMINOLOGY AND CORRELATIONS (Henry D. Appelman) doscopic findings. “’ Furthermore, the more severe the scattered deep lymphoid nodules. This complex in- endoscopic gastropathy, the more likely there was to be volves the body mucosa diffusely but spares the antral H pylo~i inf’ection. This study suggested that Hpyloti may mucosa, and many cases also have autoantihodies, par- put NSALDs users at risk for endoscopic gastropathy. ticularly those against parietal cells and . On the other hand, another study found that H pylon’ The names “autoimmune gastritis” and “‘autoimmune infection in patients with rheumatoid arthritis taking chronic atrophic gastritis” clearly are appropriate fi)r NSAIDs was only approximately 60% as common as in this combination. It also goes by the name of “tvpe A patients not taking the drugs, suggesting that WXLDs chronic atrophic gastritis” and “diffuse corpus gastri- may protect against H p$ori.” tis. “‘).I’ This disease is most common in Scandinavia Bile reflux may have some effect in eradicating H and Northern Europe. p$ori.” In one stlldy patients with How this gastritis evolves is not known, but there treated with antrectomy, all of whom had Hpylori infec- is some suggestion from a Swedish study that early in tion initially, had more persistent infection if the antrec- its evolution the lymphocytic and plasma cell inflam- tomy was accompanied by an antibile reflux procedure mation is mostly in the deep mucosa rather than the than by a proreflux procedure. The association between superficial mucosa.” Possibly progression of this deep N /~rloti, atrophic gastritis, intestinal metaplasia, and inflammation leads to destruction or atroph\, of the gas&c cancer will be discussed later. glands. In contrast, other investigators have implicated Hdicohncttr @loti is not the only cause of chronic an early phase of superficial inflammation.” Prcsum- active gastritis. There are biopsy specimens that we en- able deeper extension of this inflammation leads to counter from time to time that have all the characteris- loss of body glands. We have no idea why the tnetapla- tics of H ,t$ori infection, yet have no Hrlicohncter, and sias occur. no H Pylori are found by culture or by ancillary tests. Patients with autoimmune chronic atrophic gastri- There is one other organism that is capable of causing tis obviously produce no acid and no intrinsic tactor, a gastritis resembling H pylori gastritis. It is a tightly both of‘which are manufactured by parietal cells. They spiraled bacteria that has heen given the name of Cns- also have veIT little H pylori gastritis, suggesting that the tmpitillunr homir~i.s.5’~“4This is a spirochete that has yet organisms prefer an acid environment, although the\ to be cultured. 1~ seems to be transmitted from house- attempt to bluffer themselves from that same acid.““.“’ hold pets, dogs and cats, and it is a rare human patho- This type of diffuse body atrophic gastritis also oc- gen. In one monumental study the investigators exam- curs without the pernicious anemia and without autoan- ined more than 15,000 antral biopsy specimens and tibodies. It is not clear if the anemia- and autoantibody- found only Xl cases of Gc~strospidlum infection. Two negative diffuse body gastritis is truly the same as the other smaller studies have found these bacteria in pernicious anemia-associated type or if it is an unusualI) roughly the same frequency, with six cases in 1,650 pa- difftise form of the multifocal or type B ,itt-ophic gastri- tients in the first study and two cases in approximately tis that is discussed next. 700 patients in the second study.‘4 The inflammation Because this type of gastritis results in total destruc- induced by this bacterium is similar to that induced by tion of the parietal cells. there is no arid to turn of’f‘ 1 irlicobnctcl; hut the plasmacytosis, activity, and production by the antral <; cells. Thus, gasrrin l\,mphoid hyperplasia all tend to be less intense. production goes on unimpeded and at increased levels. and this results. in turn, in stimulation of‘ the. gastrill- AUTOIMMUNE CHRONIC ATROPHIC dependent cells in the hod\ mucosa, spccificallv the, GASTRITK enterochromaffinlike (ECL,) population of endocrine cells at the base of’ the mucosa. Then proliferate, first We have known for years that pernicious anemia, forming small basal nodules of’ endocrine cells, a fe;~- ;I megdloblastic anemia caused by vitamin B-12 defi- ture that is totally predictable in this tape of gastritis. ciency, is accompanied by a gastritis characterized by Some of these nodules apparently enla;ge. resulting in total or near total loss of oxyntic mucosa and replace- progressively larger endocrine cell lumps, some of ment of that mucosa by elongated pits and metaplastic which invade as tumors. In addition, this ppe changes in the pits and in the remaining glands (Fig of gastritis has an increased risk for dysplasias and carci- “A). The metaplasias include several types of intestinal noma. The risk for carcinoma has been estimated at metaplasia with goblet cells either scattered among al- approximately O..5LToper year, so that at the end of’ 20 tered gastric-type surf&e epithelium, incomplete intes- years approximately 10% of the patiems will have devel- tinal metaplasla. or mixed with other small intestinal oped a carcinoma. This risk is still small enough and cells, including absorptive cells and endocrine cells, the detection svstems, including upper endoscopy with complete intestinal metaplasia. Furthermore, another biopsy and cytology, still expensive enough that there c’ommon meraplasia occurs in which the body-type is no agreement about the need for a stuveillance pro- glands at-c replaced by glands that resemble antropy- gram to detect the precancerous dysplasias and the C~I-- loric or c,drdiac glands, the glands in thejunctional type cinemas in their very low stages. of Barrel-t’s mucosa, and the metaplastic pyloric-type To diagnose this gastritis with confidence the hi- glands that occur in the small bowel in segments of opsy specimens must come from areas normallv covered Crohn’s (disease. These epithelial changes often are ac- by body mucosa. there must bc no parietal cells or al- compani8rd by inflammatory changes that include plas- most none, and there must be no comparable gastritis tnacytosis and lymphocytosis in the lamina propria and in the antral mucosa. The characteristic antiparictal cell

1011 HUMAN PATHOLOGY Volume 25, No. 10 (October 1994)

FIGURE 2. The atrophic gastritides. (A) Type A or autoimmune chronic atrophic gastritis involving body mucosa. This is thin mucosa. There are no body-type glands. The only glands at the base of the mucosa are mucous glands that resemble pyloric glands, The pits vary in length, and some of them contain goblet cells, indicating intestinal metaplasia. The lamina propria is sparsely cellular. This biopsy specimen came from a patient with pernicious anemia. (HE: original magnification x 103.) (B) Type B or multifocal or environmental chronic atrophic gastritis involving the antrum. In the center of the field is relatively normal mucosa with clusters of antral mucous glands at the base, On either side the glands have been lost, and intestinalized tubules with goblet cells extend from the surface to the base. (HE: original magnification x83.)

and anti-intrinsic factor antibodies are helpful, as is the the antrum or antral-bodyjunction, especially along the presence of the basal endocrine cell nodules. lesser curvature at the incisura region (Fig 2B).i8 This Finally, as is true for all the other gastritides, there process has been designated as “multifocal atrophic is imperfect correlation between what the endoscopist gastritis” to emphasis its distribution or as “environ- sees and refers to as atrophic gastritis and what biopsies mental gastritis” to suggest its epidemiological associa- of such a stomach will yield. In general, endoscopic tions or as “type B chronic atrophic gastritis” to sepa- atrophy is represented by flat mucosa with few or no rate it from the type A atrophic gastritis that involves folds and a prominent vascularity, thought to be submu- the body diffusely and spares the antrum.“.“’ It is possi- cosal vessels more easily viewed due to an unusually ble that every patch of intestinal metaplasia is a tiny thin mucosa. In truth, this appearance often does corre- manifestation of this multifocal disease. However, this late with atrophic gastritis, but we have observed a num- disease, in spite of its worldwide dominance, is omitted ber of cases of diffuse body atrophic gastritis in stom- from some chronic gastritis classifications, which ap- achs with normal folds or even exaggerated folds. pear to bury it in other categories that are likely to include H pylon’ gastritis.““” This is the type of gastritis that seems to be the most common precursor of tubule- MULTIFOCAL ATROPHIC GASTRITIS forming adenocarcinomas and its geographic distribu- tion is the same as that of .” In an analy- Quite frequently a biopsy specimen, especially one sis of precancerous lesions associated with 1,900 cases from the antral mucosa, has a small patch of intestinal of early gastric cancer from Japan, atrophic gastritis of metaplasia or even a single metaplastic pit. The ques- this type was found in almost 95% of the cases.59 It tion arises as to whether or not this is a forme fruste of shares many histological features with Barrett’s esopha- atrophic gastritis. Clearly no one knows the answer with gus and diffuse body gastritis, including glandular atro- certainty, so the answer becomes a philosophical mat- phy; intestinal metaplasias of all varieties (both com- ter. If we assume that any intestinal metaplasia, no mat- plete and incomplete); pyloric gland metaplasia when ter how small the focus, is the result of injury and if it involves body mucosa, lymphocytes and plasma cells that patch of intestinal metaplasia is accompanied by in the lamina propria, and little, if any, neutrophilic diminution in the number of adjacent or underlying activity. glands, then this patch should qualify as a focus of The causes of this gastritis have not been clearly atrophic gastritis. Such patches occur in stomachs that elucidated. It is almost endemic in some parts of the have Hpyloti gastritis in the next microscopic field and world, such as in Japan, whereas it is much less prevalent others occur in otherwise normal mucosa. It is likely in other places, such as in the United States. Obviously. that they are manifestations of the most common gastric this suggests that there are distinct environmental in- inflammation worldwide, other than H pyloti gastritis. fluences, such as diet, but the specific information is This is a multifocal gastritis that has glandular atrophy lacking, although excess salt and lack of certain fresh and intestinal metaplasia as its histological components fruits and vegetables have been implicated.” Recently and that can involve any part of the stomach, but most H pylon’ has been implicated as a potential cause. The commonly the antral and/or transitional mucosae of evidence for this association is circumstantial, but it is

1012 GASTRITIS: TERMINOLOGY AND CORRELATIONS (Henry D. Appelman)

FIGURE 3. Chemical gastropathy. (A) This biopsy specimen came from a gastric antrum that was bathed by bile according to the endoscopist. The pits are lengthened, are tortuous, and have darker than normal epithelium caused by mucous depletion. The lamina propria is expanded, mainly by edema. (B) This is a bias cut of a similar mucosa. The entire biopsy specimen consists of pits in cross and tangential sections. In most of them the epithelium appears darker than normal because of mucous depletion and the presence of enlarged nuclei, both the result of pit hyperplasia. Such a mucosa might be misdiagnosed as low grade dysplasia. The lamina propria is edematous and slightly cellular. (A and B: HE; original magnification x52.)

compelling nevertheless. In an Italian study the preva- cluded the following: (1) foveolar hyperplasia, (2) de- lence of antral intestinal metaplasia increased with age creased mucin in the foveolar cells, (3) superficial and with change in the distribution of H p~hi from edema, (4) increased smooth muscle fibers in the lam- antral alone to antral and body to body alone.“” This ina propria, and (5) little inflammation (Fig 3) .“I In suggested that H j?ylurifirst infected the antrum; with other words, this was not really a gastritis but a gastropa- time the body became involved and atrophic gastritis thy. unless we regard the edema as an inflammatory developed in the antrum. perhaps secondary to the an- exudate. All or any subset of these changes occur with trdl infection. regularity in stomachs that are anastotnosed to the small In spite of its importance as the ma$or gastric can- bowel, but the most common are the pit hyperplasia. cer precursor worldwide, in most Western societies gas- the mucin depletion. and the absence of intense in- tric carcinoma is no longer one of the important can- flammation. Such changes occur not only in the gastric c(‘rs numerically. Thus, the finding of this gastritis in a pouch but at the anastomotic site where they are often biopsy specimen is no indication for placing any patient accompanied by prolapse changes that include pit dis- on a surveillance program to detect the dysplastic can- tortion and even more smooth muscle in the lamina cer precursors. In contrast, in those societies in which propria. In the stump one additional alteration is 11s~ gastric cancel- is a numerically important disease, ally present, namely, glandular atroph!,. This is inde- screening programs for early detection are active, but pendent of the reflux gastropathy and IS the result of’ such programs are not based on the finding of this the antrertomy that removed the gastrin-producing gastritis because it is assumed that this gastritis is en- cells, resulting in hypogastrinemia and subsequent atro- demic in ‘the population being screened. phy of the body glands, especially the parietal cells that are dependent on normal gastrin secretion for their CHEMICAL GASTROPATHIES, INCLUDING maintenance. BILE REFL.UX Bile reflux is best diagnosed either by the endo- scopic finding of bile bathing the gastric mucosa or by A co~nplex of histological changes in stomachs of a low luminal pH in the intact stomach that has proven patients who had gastroenteric anastomoses that were normal acid secretion. Although the same histological thought to be the result of reflux of bile into the stom- changes occur in intact stomachs without gastroenteros- ach was describecl. These patients were likely to have tomies, in such cases it appears that they are rarely bile reflux because the gastroenterostomy almost always caused by bile reflux but instead by other surface-dam- had been performed for peptic ulcer disease and thus aging agents, such as NSAIDs.“:! In hct, NSAIDs-related was accompanied by a distal gastrectomy, the purpose changes in the stomach have been described. consisting of which was to remove the antrum and all the gastrin- of the bile reflux complex plus a deep inflammation in producimg cells. To remove the antrum completely the the lamina propria that includes an infiltrate of eosino- with its sphincter also was excised. The gastro- phils.“’ Thus, the changes of. pit hyperplasia and mucin enterostomy reattached the stomach to the small intes- depletion with little superficial inflammation form the tine, but Ithis attachment had no sphincter, so that bile basis for a group of gastric reactions to superficial injur- could easily rcflllx into the gastric pouch. ies. The causes of many such injuries are considered to The changes that were blamed on bile reflux in- be chemical agents and the basic group of histological

1013 HUMAN PATHOLOGY Volume 25, No. 10 (October 1994)

FIGURE 4. Lymphocytic gastritis involving antral mucosa. (A) Low magnification: plasma cells fill the lamina propria, and the surface and pit epithelium contain many extra nuclei. (HE; original magnification x165.) (B) High magnification: the surface epithelium and that of the pit are disorganized and infiltrated by lymphocytes. In fact, it is difficult to determine which nuclei are lymphocytes and which are epithelial cells. (HE; original magnification x330.)

alterations has been called “chemical gastropathies.““‘“” dromes or sprues. In fact, some cases occur in patients Non-steroidal anti-inflammatory drugs are emerg- with the sprues, both gluten-sensitive and -resistant.‘“’ ing as the dominant cause of this reaction, although as At times the deeper pit epithelium also may be full of with everything else in gastritis the data are conflicting. lymphocytes and there may be an accompanying pit In one study there was no difference in the histological hyperplasia. In one of the early reports it was thought findings of chemical gastropathy among patients with that this peculiar inflammation was a response to a local either rheumatoid or osteoarthritis taking NSAIDs com- antigen, possibly H p$ori, but this has never been pared with a control group that did not take the drugs.“” proven.“’ When it accompanies one of the sprues we In contrast, in another study histological chemical gas- suspect that the gastric epithelium is being damaged by tropathy occurred about four times as often in NSAIDs the same mechanisms that damage the surface epithe- users compared with those not taking the drugs:” lium in the small intestine. However, other than this The recognition of this family of chemically in- association with we have no idea of its duced abnormalities has greatly helped to simplify the cause or causes. classification of gastritis, although they are not always Endoscopically, lymphocytic gastritis tends to pro- inflammations. The changes of pit expansion, mucous duce a complex of abnormal patterns that have been depletion, and superficial edema are common to sev- grouped into the category of “varioliform gastritis.” eral diseases, including the gastritis of H @ori, some This includes thickened folds, often topped by small phases of the atrophic gastritides, the overhanging bumps or nodules with central depressions that may edges of peptic ulcers, the tips of hyperplastic polyps, look like tiny erosions. However, there are no histologi- and the tops of the folds in some of the giant fold cal erosions, so this is still another example of poor diseases. Finding them in isolation, unassociated with endoscopic-histological correlation. We have observed bacteria, atrophy, metaplasias, ulcers, polyps, or big lymphocytic gastritis produce small endoscopic polyps, folds, should raise the possibility of chemical injuries sometimes single, but other times as one of several pol- as potential instigators. yps, the others usually being hyperplastic polyps. In fact, we have even observed gastric hyperplastic polyps in which the surface and superficial pit epithelia were in- LYMPHOCYTIC GASTRITIS filtrated by lymphocytes in a pattern identical to lymphocytic gastritis. We now recognize a chronic gastritis with intense Lymphocytic gastritis usually does not produce lymphocyte infiltration of the surface epithelium and symptoms. However, one type of giant fold disease, or the superficial pits, accompanied by lymphocytosis and hypertrophic gastropathy, is a diffuse form of lympho- plasmacytosis of the lamina propria (Fig 4). This has cytic gastritis associated with pit hyperplasia and accom- been given the name of “lymphocytic gastritis.“““,” We panied by many of the clinical features of Menetrier’s can assume that this gastritis existed before it was disease, including protein loss in about a third of the named less than a decade ago. We probably placed it patients as well as nausea, vomiting, and abdominal in some other category, such as “superficial gastritis.” pain.“* The lymphocytes are small, mature T cells. The infil- trated epithelium is disorganized, it has diminished mu- MISCELLANEOUS CHRONIC GASTRITIDES tin content, and the nuclei are stratified. These changes are virtuallv identical to those observed in the surface There are many gastritides that do not differ sig- epithelium’in the primary intestinal malabsorption syn- nificantly from similar inflammations elsewhere, includ-

1014 GASTRITIS: TERMINOLOGY AND CORRELATIONS (Henry D. Appelman)

FIGURE 5. Two examples of the broad group of miscellaneous gastritides. (A) Focal inflammation (focal gastritis) in which the cluster of glands in the center is surrounded by a dense mantle of inflammatory cells, mostly lymphocytes in this case, and one of the glands is invaded by the same cells and is damaged. This biopsy specimen might have come from a patient with Crohn’s disease, in which case it is a helpful diagnostic hint, or it may be an isolated finding, in which case it is an unexplained annoyance, (B) Granulomas like the tight non-necrotizing one in the center might be part of any disease that has granulomas, from sarcoidosis to Crohn’s disease to granulomatous infections to the isolated granulomatous disease that only involves the stomach. In this case the patient had Crohn’s disease. (A and B: HE: original magnification x 165.)

tions,““ 70 ing those that occur in syphilis, mycobacterial and cyto- Possibly as many as a third of patitbnts with megalovirus infections, histoplasmosis, candidiasis, and small intestinal and colonic (;rohn’s disease have upper other opportunistic fungi. Even cryptosporidiosis in the gastrointestinal involvement, but the changes are more immunosuppressed patient, such as in those infected likely to be microscopic with no associated endoscopic by the human immunodeficiency virus, involves the findings. The typical histological appearances of gastric stomach.” These are not unique gastric inflammations Crohn’s disease are comparable with those of’ Crohn’s and will not be covered in this article because of space disease in other sites, namely. patchy or ti~nl inflam- limitations. mation, including such diverse changes as lvmphoid In any surgical pathology practice affiliated with a aggregation, pititis or glanditis mediated rithrr bv neu- busy endoscopy service, there are likelv to be a Sew trophils or lvmphocytes. plasmacytosis, tiny ulcers, and annoying gastric biopsy specimens containing inflam- granulomas.’ Unfortunately, these are the changes de- mations that do not fit neatly into any established cate- scribed above in the unnamed focal lesions, which are gory. Such processes,justify the inclusion in the gastritis found periodically in random gastric biopsy specimens classification system of a category that recognizes un- in individuals without CL-ohn’s disease. Nevertheless, a usual inflammations and outliers, that is, inflammations set of biopsy specimens that contains mllltiplta focal lc- that partly satisfy the diagnostic criteria for one of the sions as described above should raise the possibilitv of c,stablished gastritides but that lack one or more criteria Crohn’s disease. or have unusual or unique features (Table 2). Included There is a f’ami,lv of granulomatous reactions ok in this category are the focal gastritides. often character- granulomatous gastrltldes (Fig 5B). Some* of‘ these arc ized by patches of intense inflammation in an otherwise part of a systemic granulomatoua disease, such as sar- normal I-LILICOS~ (Fig 5%). Usually these are foci of lvm- coidosis, and some have been described as part of a phocytosis and/or plasmacytosis associated with systemic vasculitic syndrome or as an accompaniment lymphocytic or neutrophilic infiltration of pits, glands, of Whipple’s disease. Others arc the gastric tnanifesta- or necks. Most of these have no names and they ma) tions of focal gut diseases (usuallv (:rohn’s disease), and IX viewed as comparable to the focal inflammatory le- in such cases the granulomas arc almosr always small sions, or pimples, that dermatologists often biopsy and and solitaIT.7’.” Still others are not associated with an\ that challenge classification by dermatopathologists. In other disease and such cases have been designated as general, except for Crohn’s disease, microscopic focal “isolated granulomatous gastritis.” Finally, occasional intlammations anywhere in the granulomas occur next to ruptured pits. apparently re- lack diagnostic specificity, but as we see more of them actions to extravasated mucin comparablt* with the rup- we will gradually begin to name them. tured crypt granulomas that are f’ound in thtt colon in Crohn’s diiease affects the stomach, although duo- many inflammatory diseases. denal irlvolvement is more common and more im- .Allergic gastritis is usually pal-t of’ a niorc extensive portant clinically. The prevalence of gastric involve- allergic .7’3 It mainlv affi.cts children, es- ment probably depends on whether patients with newly pecially young children, who pre&Jlt with vomiting, di- diagnosed (Zrohn’s disease of the small bowel and/or arrhea, weight loss, peripheral blood ~osinophilia. ele- the colon have careful upper endoscopic. examina- vated serlml immunoglobulin E. and a histon. of‘aller~

1015 HUMAN PATHOLOGY Volume 25, No. 10 (October 1994)

FIGURE 6. Two examples of acute gastritides. (A) In this acute erosion the superficial mucosa is necrotic and sparsely cellular, whereas the underlying pits are regenerative with dark epithelium. Very little inflammation is present considering the intensity of the destruction. (B) Acute hemorrhagic gastritis. The entire superficial mucosa is necrotic with extension into the mid-mucosa on the left The dark, granular appearance in the necrotic areas is caused by large numbers of extravasated erythrocytes. (A and B: HE; original magnification x 103.)

either in the patient, the family, or both. Biopsy speci- defect with no surface epithelium and a cell-poor granu- mens of the gastric antrum have a high yield of findings lar lamina propria that may have a few extravasated that include eosinophilia of the lamina propria that erythrocytes in it. Considering the intensity of the ne- may be diffuse and intense as well as eosinophilic infil- crosis there is remarkably little inflammation, usually tration of the superficial and pit epithelium with accom- no more than a few neutrophils at the junction between panying epithelial damage and concurrent regenera- the viable and necrotic tissues. Intense epithelial regen- tion. eration usually is present at the same time with regener- Weird, one-of-a-kind gastritides also occur, some ating tubules often extending to the base of the mucosa of which even have published names. Thus, there is a as if the regenerative zone became so active that it en- single case report of “collagenous gastritis,” character- croached on the glandular compartment. The total pic- ized by a thick subepithelial band of collagen and a ture is a reaction to injury that looks much like acute mild underlying inflammation.74 It is only fitting that ischemia or other surface necroses, such as occurs in we now have collagenous gastritis because collagenous the colon in the acute colitis of verotoxin-producing colitis and collagenous sprue in the small bowel already Escherichia coli. These days most such acute erosions are exist. Collagenous cannot be far behind! caused by ingestion of one of the NSAIDs, including aspirin. These erosive lesions develop as a result of de- ficient mucosal protective capabilities and may result ACUTE GASTRITIS from lowered prostaglandin synthesis by the gastric mu- cosa caused by the drugs. Although this article has concentrated on the Alcohol in actively drinking alcoholics has been chronic gastritides because they are by far the most historically implicated as another cause of such ero- common diagnostic problems and the processes that sions, but although tiny superficial microscopic ero- are currently studied most intensely, there are few truly sions may occur, alcohol seems to induce mainly a histo- acute inflammations of the gastric mucosa that deserve logical chemical gastropathy with reactive surface and recognition. When biopsies are performed on gastric pit epithelium. However, in some alcoholics there are mucosa they add to the diagnostic problems of the sur- subepithelial hemorrhages in an edematous mucosa, gical pathologist. Some of them may even be superim- both without much necrosis or inflammation.“.7”.7” posed on the chronic gastritides and. as best as we can Occasional debilitated patients develop a diffusely tell, they do not result in chronic sequellae, except for necrotizing and hemorrhagic injury involving the su- occasional chronic ulcers. Thus, they are not precursors perficial mucosa (Fig 6B). This is possibly caused by of the chronic gastritides as far as we know. Acute H ischemia in some cases, but the cause is often not pylori gastritis was mentioned earlier, but this is not a known.“-” On occasion this can be life threatening. biopsy issue because biopsies virtually never are per- leading to uncontrollable hemorrhage requiring gas- formed on acute infectious gastritides, and there is very trectomy. little published information about them or experience with them. Many of the important, noninfectious acute in- MULTIPLE COEXISTENT GASTRITIDES flammatory lesions of the stomach are small superficial erosions in which the epithelium of the surface and If each of these types of gastritis is accepted as a superficial pits and the superficial lamina propria are separate entity and because many of them are common. necrotic (Fig 6A). This results in a superficial mucosal it should not be surprising when more than one of

1016 GASTRITIS: TERMINOLOGY AND CORRELATIONS (Henry D. Appelman) thtm is foLtnd in a given stomach at the same time, so mars or screen for dysplasia; however, in nei- that a biopsy specimen contains multiple gastritides. As ther situation is the biopsy for the gastritis. examples, we have observed antral biopsy specimens in (4) Virtually every patient with tubule-forming gas- which there was H l?ykori gastritis as well as patches of tric adenocarcinoma has multifocal type R atrophic gastritis with intestinal metaplasia and loss of chronic atrophic gastritis and except for intel- glands. We have even observed this combination com- lectual gratification or for purposes of epide- plicated by an acute erosion in a patient treated with miological analysis we are not interested in the long-term NSAIDs. Sometimes patients with diffuse nontumorous mucosa. atrophic gastritis of the body, the type associated with (5) If we perform biopsies on enough patients pernicious anemia, also have antral biopsy specimens with sprue we shall find that some of them with multifocal atrophic gastritis. Thus, even the two have lymphocytic gastritis, but the real disease types of atrophic gastritis exist in the same stomachs is in the proximal small bowel and not in the f&,m time to time. As mentioned earlier. it is unusual stomach. for H p$oti gastritis to be found in the antrum when (6) Perhaps every patient who is a long-term autoimmune gastritis is present in the body. NSAIDs user will have microscopic gastric changes. However, their dyspeptic symptoms will tell us that they have drug toxicity and whether or not we find chemical gastropath) CONCLUSlON in the biopsy specimen will not matter. (7) In patients with known Crohn’s disease involv- 1s there a preferable, functional, completely useful, ing various parts of the small intestine or co- intellectually satisvi,ng classification of gastritis. espe- lon, careful examination of the stomach with cially chronic gastritis? Of course there is, but it is prob- many biopsy specimens will pick up focal in- ably diffe,r-ent for different practitioners who work in flammations in some patients, so why even per- different countries and in different practice settings. form a biopsy? The Crohn’s disease that is the Any useful classification system probably depends on target of treatment is the large symptomatic the types ~of’stomachs from which biopsy specimens are segment, not the incidental focus that is onl) obtained, and this is likelv to be population-based. Any detected by screening. usetill classification system also depends on the needs All of these statements are true and each is sup- of‘ the clinicians with whom we work because they are ported by extensive literature. If what we know is ab- on the receiving end of our diagnoses. We pathologists plied to the way we practice and if we accept that there must not work in a vacuum. The best approach to the are very few indications for histological diagnoses in classification of gastritis, especially that for the gamut the nontumorous stomachs, eventually we may put his- of chronic gastritis, must consider what it is our clinical tological chronic gastritis to rest. In the meantime we colleagues want us to tell them. pathologists shall continue to face the intellectual and Furthermore, it is critical that we pathologists know professional challenges that require that we make only whether it is even necessary and important to ever make accurate diagnoses, and we shall continue the search a histological diagnosis of any kind of gastritis, because for some satisfying classification scheme for the collec- if it is not then the next issue becomes whether it is tion of pits, necks, glands, plasma cells, lymphoid nod- necessary or important ever to perform a biopsy on the ules, neutrophils, eosinophils, distortion, atrophy, stomach except to detect neoplasms or specific infec- metaplasias, and epithelial damage that will allow us to tions. There are facts about the different forms of gastri- place each histological gastritis into a predetermined tis. most of which have been presented earlier in this niche. article, that may make most pathological diagnoses of gastritis obsolete in the coming years: (1) L’irtually every duodenal peptic ulcer is accom- panied by H pylori antral gastritis. The only reason to perform a biopsy of the antrum in Ahbrf~ialions: NSAlDs, non-steroidal anti-inflammatory drugs: such cases is to determine if it is normal or if HE, hematoxylin-eosin; TNF, tumor necrosis f:dctor; MALT, it has chemical gastropathy, so that the clini- mucosa-associated lymphoid tissue; ECL, entel-ochromaf- cian cm be notified that the patient may be finlike. an NSAIDs user and the ulcer may be drug- rather than bacterial-induced. REFERENCES It is a toss-up whether a given patient with non- U~CCI-dyspepsia will have H pylori gastritis, and 1. Cronstrdt JL, Simson IW: Correlation bctwcrn gastroscopic some studies indicate that it does not matter and direct vision biopsy findings. Gastrointrst E,ndosc lY:l-2. 1973 if they do because approximately two thirds of 2. Elta GH. Appelman HD. Brhler EM. et al: A study of thr all patients with this set of symptoms respond correlation between endoscopic and histologic diagnose\ in gastrodw to antibacterial treatment. odenitis. Am J Gastroenterol 82:749-753, 1987 Every patient with pernicious anemia has type 3. Fung MT’. Papadimitriou JM. Matz 1.R: Endoscopic, histologi- cal and ultrastructural correlations in cht-onic gastr-iti\. Am J Gas- .A diffuse chronic atrophic gastritis of the body troenterol 71:269-279. 1979 or corpus, so what is the point of biopsy except 4. Gad A: Erosion: A correlative rndoscopi(- histological multi- perhaps to chop out a few tiny carcinoid tu- centric study. Endoscopy 18:7&7!3. 19Xti

1017 HUMAN PATHOLOGY Volume 25, No. 10 (October 1994)

5. Rubin CE: Histological classification of chronic gastritis: An sponse. gastric mucosal histology. and gastric juice ascorbic acid con- iconoclastic view. 102:360-361. 1992 centrations. Gut 32:1415-1418, 3993 6. Toukan ,%U. Kamal MF, Amr SS, et al: Gastroduodenal in- 32. Barnett JL. Behler EM, Appelman HD, et al: Camp~lobudrr flammation in patients with non ulcer dyspepsia. Dig Dis Sci 30:616- f;t$i is not associated with gastroparesis. Dig Dis Sci 34:1677-1680, 620. I985 7. Whitehead R: Etiolog)i and significance of simplr chronic gas- 33. O’Connor HJ, Newbold KM. Alexander-Williams J, et al: Ef- tritis. in Mucosal Biopsy of the (;astr-ointestina] Tract, <:h 6. 7. (cd fect of Roux-en-Ybiliary diversion on Campylobacturp$mi.&astroenter- 3). Philadelphia, PA, Saunders, 19X5, pp 41-64 ology 97395%964, 1989 8. Misiewicl IJ, Tvtgat GFJ. Goodwin CS, vt al: The Sydney Sys- 34. I.offeld RJLF. Potters HVJP, Stobberingh. et al: ~;clmp~lobnctrr tem: A new classificaiion of gastritis. in Working Party Reports of associated gastritis in patients with non-ulcer dyspepsia: A double the 9th M’orld Congress of Gastroenterology. Melbourne, .kustralia, blind placebo controlled trial with colloidal bismuth subcitrate. (;ut Blackwell Scientific, 1990. pp l-10 .30:120ci121~~-, 1989 9, (Zorrra P: Chronic gastritis: A clinico-pathological classifica- 35. Carrick J, Lee A, Hue11 S, et al: Cum@obnctrr,tyloon’. duodenal tion. Am J Gastroenterol 83:504-509. 1988 ulcer. and gastric metaplasia: Possible role of functional heterotopic 10. Dixon MF: PI-ogress in the pathology of gastritis and duodc- tissue in ulcerogenesis. Gut 303790-797, 1989 nitis. in Williams GT (ed): Gastrointestinal Pathology. Berlin, (;cr- 36. Graham DY, Go MF: Hulzrobactrrpy1m-i:Currrn t status. Gastro- many, Springer-\‘erlag, 1990. pp l-?7 enterology 105:279-282, 1993 11. Goldman H: Gastritis, in Ming SC, Goldman H (eds): Pa- 37. De Cothi GA, Newbold KM, O’Connor HJ: C~ampylobartwlike thology of the Gastrointestinal Tract, Ch 20. Philadelphia, PA, Saun- organisms and heterotopic gastric murosa in Meckel’s diverticula. J ders, 1992, pp 481-516 Clin Path01 42: 132-l 34, 1989 12. Lewin KL. Riddell RH, Weinstein WM: Gastrointestinal Pa- 38. Dyr KR. Marshall BJ, Frierson HF. Jr. et al: (:amp$obnrtrr thology and Its Clinical Implications. New York, NY, Igaku-Shoin, p?ilon colonizing heterotopic gastric tissue in the rectum. Am ,J Clin 1992, pp .508-537 Pathol 93:144147, 1990 13. Owen DA: Stomach, in Sternberg SS (ed); Diagnostic Surgi- 39. Gruppo Operative per lo Studio delle Precancerosi Eso- cal Path&g). Ch 29. New York, NY, Raven Press. 1989_ pp 939-948 fagee: Helicobacter @x-i in Barrett’s orsophagus. Histopatholog! 14. Yardley JH: Pathology of cht-onic gastritis and . l&568-570, 1991 in Goldman H, Appelman HD. Kaufman N (eds): Gastrointrstinal 40. Paull G, Yardley ,JH: Gastric and esophageal Cclmfilobadrr Pathology, Ch 3. Baltimore, MD. U’illiams % Wilkins, 1990, pp 69. pylori in patients with Barrett’s esophagus. Gastroentrrology 9.5:21& 121 218, 1988 15. W;an-en JR: Unidentified curved bacilli on gastric cpithelium 41. Talley NJ, Cameron AJ, Shorter KG, et al: Cwn~lobactrrp~lon in active chronic gastritis. I,ancet 1:1273-1275. 1983 (letter) and Barrett’s esophagus. Mayo Clin Proc 63:1176-I 180. 1988 42. The Eurogast Study Group: An international association be- 16. Marshall BJ, Warren JR: L’nidentified curved bacilli in the tween Helicobartrr~1m-iinfection and gastric cancer. 1;ancet 341:1359- stomachs of patients with gastritis and peptic ulceration. Lancer 1362. 1993 1:1311-131.5, 1984 43. Tatsuda M, lishi H, Okuda S. et ai: The association of H&u- 17. Marshall E%J,Armstl-ongJA, McGechie DB, et al: Attempt to hncte-rpylon with differentiated-type early gastric cancer. Cancer fulfill Koch’s postulates for pyloric Cwnpylobartrr. MedJ Aust 142:43& 72:1841-1845, 1993 439. 1985 44. Parsonnet J, Vdndersteen D. Goatrs J, ct al: Helicobacter 18. Mendall MA, Noggin PM, Molineaux N. et al: Childhood pylori infection in intestinal- and diffuse-type gastric adrnocarcino- living conditions and I~&cr~hn&r/@ti seropobitivity in adult life. Lan- mas. J Nat1 Cancer Inst 83:640-643, 1991 cet 339:89&897, 1992 45. Talley NJ, Zinsmeister AR, Weaver A, et al: Gastric adenocal-- 19. Dooley CP. Cohen H. Fit/-gibbons PI., et al: Prevalrnce of cinema and Hrlicobnctpr,@vti infection. .J Nat1 Cancer Inst 83: 1734 Hdirobnctn pylon’ infection and histologic gastritis in asymptomatic 1739, 1991 persons. N Engl J Med 321:1562-1566. 1989 46. Wee A, Kang JY. Teh M: Helicobacter pylori and gastric 20. Cullen DJE, Collins BJ, Christiansen KJ, et al: When is H&u cancer: Correlation with gastritis. intestinal metaplasia. and tumour bnc/Prpylotiinfection acquired? (:ut 341681-1682. 1993 histology. Gut 33:1029-1032. 1992 21. Correa P: Chronic gastritis and gastric cancer, in Ming 47. Sipponen P, Kosunen TU, \‘alle J, et al: Hdzcobnctrr /#VT S-C (cd) Precursors of Gastric Cancer. Ch 9, New York, NY, Praeger infection and chronic gastritis in gastric cancer. J Clin Pathol 45:319- Publishers, 1984, pp 105-l 16 323, 1992 22. Johnston HD: Blind gastric and duodenal mucosal biopsies 48. Wotherspoon AC, Ortid-Hildago C. Falzon MR, et al: Hdico- in patients with duodenal ulcers and those with an ulcer-like syn- barter #$&associated gastritis and primary &cell gastric lymphoma. drome. South Med J 57:79-84, 1964 Lancet~338:1175-1176, 1991 “3. I.offeld RJLF, Potters H\‘PJ, Arends JW. et al: (knpylobadpr 49. Wotherspoon AC. Doglioni C, Diss TC, et al: Regression associated gastritis in patients with non-ulcer dyspepsia.,] Clir; Path01 of primary low-grade Bqell gastric lymphoma of mucosa-associated 41:85-8X, 1988 lymphoid tissue type after eradication of Hdicobact~ pylori. Lancet 24. Genta KM. Hamner HW, Graham DYz Gastric lymphoid folli- 342:575-577. 1993 cles in Hrlicobnckr pylon infection: Frequenq. distribution, and re- 50. Heresbach D. Raoul IL. Bretagns JF, et al: Helirobu~trrp~loric sponse to triple th&apv. HLrct P:\I.I~~II. 24:577-583, 1993 A risk and severity factor of non-steroidal anti-inflammatoq drug 25. Eita GH. Murphy R. Behler EM, et al: Gzmp$obarterpyloti in induced gdstropathy. Gut 33: 1608-I 611, 1992 patients with dyspeptic symptoms and endoscopic evidence of ero- 51. Taha AS, Nakshabendi I. I.ee FD, et al: Chemical gastritis sion [s]. Am J Gastrornterol R4:643-636, 1989 and HdicohnctPrp?rlorirelated gastritis in patients receiving non-steroi- 26. Crabtree JE. Shallcross TM, Heatley R\‘f, et al: Mucosal tll- da1 anti-inflammatov drugs: Comparison and correlation with peptic mom necrosis factor alpha and interleukin-‘6 in patients with H&o- ulceration. J Clin Pathol 45:135-139, 1992 hnctfl p$ori associated gastritis. Gut 32: 1473-1477, 1991 52. Offerhaus GJA, Rieu PNMA, Jansen ,JBM,J, et al: Prospective 27: Craig PM, Territo MC, Karnes WE, et al: Helicobackr /@ori comparative study of the influence of postoperative bile reflux on seclctes a chrmotactic factor for monocytes and neutrophils. Gut gastric mucosal histology and Campylobarter p$cm infection. Gut 33: 1020-1023. 1992 30:15.52-1557. 1989 2X. Nielsen H. Andersen LP: Chemotactic activity of Hdirobnctw 53. Heilmann KL, Borchard F: Gastritis due to spiral shaped pdon sonicate for human polymorphonuclear leucocytes and mono- bacteria other than He1icobactPrpylor-i:Clinical. histological. and ultra- cytes. Gut 33:738-742, 1992 structural findings. Gut 32:137-140, 1991 29. Chan WY, Hui PS Ghan JIX, et al: Epithelial damage by 54. Morris A. Ali MR, Thomsen L. et al: Tightly spiral shaped Hdicoburkrp?lori in gastric ulcers. Histopathology 19:47-53, 1991 bacteria iu the human stomach: Another cause of active chronic 30. Hui PK, Chan WY, Cheung PS, et al: Pathologic changes of gastritis. Gut 31:139-143, 1990 gastric mucosa colonired by Hrlirobactrrpylon. HUM PAT~~OL.23:.548- 55. Burman P, &mpe 0, Kraa/ W, et al: A study of autoimmune 556. 1992 gastritis in the postpartum period and at a 5-year follow-up. <>astroen- 31. Sobala GM. CrabtreeJE, Dixon MF. et al: Acute Hdicobnck-r terology 103:934942. 1992 p&-i infection: Clinical features, local and systemic immune re- 56. FlejouJ-F. Bahame P. Smoth AC, et al: Pernicious anaemia

1018 GASTRITIS: TERMINOLOGY AND CORRELATIONS (Henry D. Appelman)

68. Woll\on H(:, (krpentcr HA. ‘T;rllc\, NJ: hlc11rtricr’\ diseaw: ,A for-111 of hypertrophic gawopath\ 01. g;r&tiC (;a~troc11terotoa 104:1510-1319. 1993 69. Daneah I%JZ, Park KHR, Up,tdhyy K, CI xl: l)iagn~~stic yirtd in upper g,1st1-[1inteatitlal rndoscopic hiopsie\ it1 paticnc\ with (:I-oh11’\ distaw. Castroe1lterolo~ Yfi:AlOX, 1989 (ahstr) i0. Lenaertr CKC(:. Vailtancourt M, b’rtwr ,451. t’t at: 1Ggh inci- dc11cc of uppc1- ,gastrointestiml tract invol\enw1t i11 children with (:roh11’s diseasr. Pediatrics X3:777-%1, IWI 71. F.rtors Nl.. Dixon MF. Gehocs &J, ct ~1: (;r;c111~t~)rn,rto11s gab- rritis: A morphologicnl and diagnostic app1-

/5. I.ainr 1,. Weinstein UXI: Hiaolog\’ of .dcoh~1ht hcmo1-rhagic “gasti-itis”: A pi-ospcctivr e~iluation. (kstrorntcr-0loF 9-t: I251l?C?. 198X 76. l.aiur I.. Weinstein WY: The hktology of trorive “gastritis” i1l alcoholic patimtx X prospcctivc \tutlv. (;a\t1-(,rntcrl,l(~~, 92: 148X. lYX7(ahslr) 77. Owen D.4: Gastritis and duodenitis, in Xppvl111an HD (rd): I’atholofl of the Esophagus. Stomach. nnd Duodc11run. (:h 2. NW, Yor-k. NY. Churn hill IAingstone, 19X4. pp 4L’-l:3

1019