Gut, 1990, 31, 282-285

Lymphocytic gastritis - prospective study of itsrelationship with varioliform gastritis

J Haot, A Jouret, M Willette, A Gossuin, P Mainguet

AbstractLymphocytic gastritis is a new histopatho-logical entity characterised by a denselymphocytic infiltration of surface and pitgastric epithelium. Previous retrospectivework has suggested that lymphocytic gastritisis related to an endoscopic form ofgastropathycomprising enlarged folds, nodules anderosions, commonly denoted as varioliformgastritis. In the present prospective study, therelationship is clearly shown; nearly 82% (54/66) of the varioliform gastritis observed in fourdifferent endoscopy units correspond histo-logicaily to lymphocytic gastritis. The correla-tion is even better if cases showing strictlyantral localisation are excluded (53/55) - thatis, more than 96%. The histological concept oflymphocytic gastritis seems, however, toextend beyond varioliform gastritis as of 67cases of lymphocytic gastritis diagnosedduring the period under study, one third had noparticular endoscopic expression.

Departments ofPathological Anatomyand Gastro-Enterology,St Luc UniversityHospital, UCL, Brussels,BelgiumJ HaotP Mainguet

H6pital Notre Dame deFrameries, Brussels,BelgiumP Mainguet

IMC Peruwelz, H6pitalde Leuze, Brussels,BelgiumM Willette

IMC Beloeil, Brussels,BelgiumA Gossuin

H6pital Civil deTourmai, Brussels,BelgiumA JouretCorrespondence to: ProfessorJ Haot, Service d'AnatomiePathologique, CliniquesUniversitaires St Luc, AvenueHippocrate 10 - B 1200Bruxelles, Belgium.Accepted for publication15 May 1989

In 1947, Moutier and Martin' reported two casesof a distinctive inflammatory gastric diseaseshowing widespread mucosal nodules with orwithout central depressions or erosions associ-ated with an enlargement of the corporeal folds.The name varioliform gastritis is now widelyused in the continental scientific literature. Sincethen, similar endoscopic and radiologicalfeatures have been described under variousnames such as aphthous ulcers, chronic erosivegastritis, and 'octopus sucker' gastritis.2'More recently, Lambert et all reporting a

series of 90 cases, redefined the entity anddistinguished between two forms of varioliformgastritis: one involving the whole stomach wascalled diffuse varioliform gastritis while theother, being restricted to the antrum, was namedantral varioliform gastritis.

In 1985, we identified a new histopathologicalentity, characterised by dense intra-epitheliallymphocytic infiltration, called lymphocyticgastritis.6 Clinicopathological studies haveshown that lymphocytic gastritis corresponded,in one third of cases, to a clinical presentation ofweight loss and anorexia. Endoscopicallylymphocytic gastritis correlated with the findingof nodules, thickened folds and erosions pre-dominating in the corporeal region of thestomach.7` All these features appeared tocorrespond to varioliform gastritis. We haveshown in two recent retrospective studies, thatthe majority of cases diagnosed in the clinicalreports as varioliform gastritis revealed on histo-logical examination the typical dense intra-epithelial lymphocytic infiltration oflymphocyticgastritis. '° I

Taking into account, however, the oftenincomplete and imprecise descriptions in theendoscopic records, as early as 1985, a prospec-tive work was planned with the aim of betterdefining the correlation between the endoscopicdiagnosis of varioliform gastritis and the histo-logical diagnosis of lymphocytic gastritis. Theresults of this research are presented here.

MethodsThe present research was planned by the threeendoscopists and two pathologists taking part inthe study who agreed on the methodologyrequired to permit valid comparisons betweenthe entities of varioliform gastritis and lympho-cytic gastritis. These entities were defined asfollows.The endoscopical features of varioliform

gastritis are enlarged and thickened rugal foldsbearing erosions and widespread small nodulesfrequently surmounted by small roundederosions (aphthoid nodules). According to thetopography of the lesions, three forms of varioli-form gastritis can be distinguished: diffuse whenthe whole stomach is involved, corporeal whenthe anomalies are limited to the body and antralwhen they are only present in the antrum.

Histologically the diagnosis of lymphocyticgastritis rests upon the presence of an unusuallyhigh number of intraepithelial lymphocytes anddoes not take into consideration other micro-scopic features often present such as lympho-plasmocytic or neutrophilic infiltrate of thelamina propria.

ENDOSCOPYIn order to permit comparison between endo-scopic and histological data a rigorous pro-cedure was followed. All the patients (4840)endoscopied and biopsied between 1985 and1987 for a non-neoplastic condition of thestomach in four gastroenterology departments(Clinique Notre Dame de Frameries, IMCPeruwelz, Hopital de Leuze, IMC Beloeil) wereincluded in the study. The endoscopic protocolcovered all the anomalies of the gastric mucosa,their number and topography.

HI STOLOGYIn all patients, three to five biopsy specimenswere taken from the body as well as from theantrum sampling nodules or erosions andapparently uninvolved mucosa. They were fixedin Bouin's solution and embedded in paraffin.Five micron sections cut made and stained withhaematoxylin and eosin.The two pathologists jointly studied all the

282

on Novem

ber 2, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.3.282 on 1 March 1990. D

ownloaded from

Lymphocytic gastritis -prospective study ofits relationship with varioliform gastritis

gastric biopsies originating from the fourcentres. Without considering the clinical data,they systematically screened the biopsies insearch of cases of lymphocytic gastritis. Biopsieswere categorised of lymphocytic versus othertypes of gastritis.As already stated, the diagnosis of lympho-

cytic gastritis was based on the (subjective)observation of high densities of intraepitheliallymphocytes. Counts were made on all the cases.In order to avoid a bias due to a possiblepatchiness of distribution, two different biopsyspecimens from each patient were read atrandom. The lymphocytes were counted on anuninterrupted length of 200 surface epithelialcells. The results of the two counts were pooled.A comparison was made with an equivalentnumber of control cases of chronic atrophicgastritis studied under the same conditions. Thecases were chosen from the pool of patients seenin the present series. The controls were pairedwith each case of lymphocytic gastritis - that is,we chose from our files the first case of chronicatrophic gastritis after a case of lymphocyticgastritis.

Counts were also made on 20 cases where thegastric biopsies were considered histologicallynormal.

Results

ENDOSCOPIC DATA

Endoscopic diagnosisAmong the 4840 patients examined, 66 (1-4%)were diagnosed as varioliform gastritis. Thirtyfive were classified as diffuse varioliform gastritis(the sex ratio was 25 men/10 women - averageage 44 years; range: 16-82). Twenty were con-sidered as corporeal varioliform gastritis (15men/five women - average age 47 years; range:

31-82), finally 11 were antral varioliformgastritis: six men/five women - average age 45years; range: 27-64).

So far as diffuse varioliform gastritis is con-cerned, the features were consistent. Practicallyall cases exhibited irregularly thickened foldspersisting after distension with air. Theseextended through the whole body region, thin-ning progressively down towards the antrum;and were covered by a thick mucus secretion,crossed by large whitish serpiginous erosionslined with fibrin and punctuated at their top byulcerated nodules (aphthoid nodules) arrangedin a stringlike fashion. Flat erosions outside thenodules were rather scarce; they were most oftenseen in the antral region.The picture of corporeal and antral varioli-

form gastritis was, as a whole, characteristic butless pronounced than in the diffuse form. Thefolds were less prominent but nevertheless per-sisted after inflation; the nodules were morescarce, sometimes without ulceration and thelarge erosions on the folds were frequentlyabsent. Nevertheless, at least two features ofvarioliform gastritis were seen: either enlargedfolds and nodules, enlarged folds and erosions ornodules and erosions.

Correlation between endoscopy and histology(Table I)Of the 35 cases diagnosed as diffuse varioliformgastritis, all had lymphocytic gastritis on histo-logy. When the lesions were limited to the body,

TABLE I Correlation endoscopy - histology

Cases (n)Lymphocytic gastritis

Endoscopic diagnosis ofvarioliform gastritis

Diffuse Corporeal Antral Total

35 20 11 66;35 18 1 54

Figure 1: Low magnification (60x): Surface and pit lining epithelium is infiltrated by numerous lymphocytes. The laminapropria inflammatory infiltrate is mild.

283

on Novem

ber 2, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.3.282 on 1 March 1990. D

ownloaded from

Haot, Jouret, Willette, Gossuin, Mainguet284

F~~~~~~~~~~~~~~~~

.}

Figure 2: High magnification (480x): Dark nuclei of theintra-epithelial lymphocytes are surrounded by a clear rim or'halo'.

the correlation between endoscopy and histologywas less complete; nevertheless it remained veryclose because 18 of 20 cases corresponded tolymphocytic gastritis. The two exceptions were'subacute' gastritis, characterised by oedema ofthe lamina propria, regenerative features of theepithelium, and a mixed inflammatory infiltrate.

Antral varioliform gastritis behaved quitedifferently; although the number of cases wassmall, the difference was striking as 10 of 11cases were not lymphocytic gastritis. Seven casescorresponded to chronic active gastritis. In twocases, we again found oedema and acute inflam-mation analogous to those described in the twoanomalous corporeal cases. Finally, one caseshowed oedema and fibrosis of the laminapropria without inflammatory cells.

HISTOLOGICAL DATA

Histological study oflymphocytic gastritisIn all cases, the intraepithelial lymphocyteswere sufficiently numerous to give an easilyidentifiable picture even at low magnification.The epithelium appeared unusually basophilic,and was punctuated by numerous small cellswith little cytoplasm (Fig 1). Very frequently, inthe eroded areas, the picture was overshadowed

TABLE II Correlation histology - endoscopy

Endoscopic features corresponding tO the 67 cases ofIymphocyticgastritisVarioliform features Non-varioliform features

Typical 'Incomplete' Erosions Ulcers Non specific

54 2 3 3 5

by abundant oedema and a mixed interstitialcellular infiltrate. The foveolae were enlarged,corrugated and penetrated by polymorphs.Sometimes, the dilated lumens contained mucusplugs and polymorphs ('crypt' abscesses). Athigh magnification, however, the characteristiclymphocytic component could easily be found(Fig 2). The erosive features and the lympho-cytic infiltration were more evident in the bodythan in the antral region.Lymphocyte counts made on the surface

epithelium showed an average of 57 lymphocytesper 100 epithelial cells. The SD was 20-4 andthe range (31-138). In comparison the resultsyielded by counts in chronic atrophic gastritisgave an average of 3-4 (SD: 2 3) with a range of1-12-5. In histologically normal mucosa, theresults were 2 5 (SD: 2-4) with a range of 1-9.Statistical comparison (Student's t test) betweenlymphocytic gastritis and chronic atrophicgastritis was highly significant (p<0 00 1). More-over, even in this large series, comparison ofranges showed no overlap between lymphocyticgastritis and chronic atrophic gastritis. Asexpected, the difference in counts betweenlymphocytic gastritis and normal mucosa wasalso highly significant (p<0 001). Comparisonbetween chronic atrophic gastritis and normalmucosa was not significant.

Correlation between histology and endoscopy(Table II)During the period considered (1985-1987)among the 4840 gastric biopsies examined, weregistered 67 cases oflymphocytic gastritis. Fiftyfour corresponded to varioliform gastritis while13 cases had endoscopic features differing fromvarioliform gastritis. Among them, two couldpossibly be considered as having incompletefeatures of the disease. In one isolated corporealenlarged folds was found while in the other wereseen a bunch of nodules in a flat mucosa. The restof our observations comprised two ulcers andone ulcer scar, three chronic atrophic gastritiswith erosions on a flat mucosa and five patientswithout endoscopical lesions biopsied onsuspicion of chronic B type gastritis.

DiscussionWhen describing lymphocytic gastritis in 1985,6we were impressed by the similarity of particularclinical and endoscopic features. We havementioned in recent reports'" I that this picturewas very similar to that described by Lambert etal 19785 under the name of varioliform gastritis.As these early works were based on retrospec-

tive data, we have been reluctant so far to assessthe degree of correlation between the two condi-tions. The present prospective work allows us topositively state that the endoscopic featuresconstituting as diffuse varioliform gastritiscorrespond to the histological diagnosis oflymphocytic gastritis. This is true as long as thedefinition of diffuse varioliform gastritis isprecise. The disease, although predominating inthe body, must involve the whole stomach; itmust exhibit thick folds persisting after disten-sion with air. The folds bear at their top aphthoid

on Novem

ber 2, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.3.282 on 1 March 1990. D

ownloaded from

Lymphocytic gastritis -prospective study ofits relationship with varioliform gastritis 285

nodules, spreading along in a stringlike fashionand are often crossed by irregular serpiginouserosions. Erosions on flat mucosa are morescarce.

Moreover, it appears in our experience thatthere also exists a less florid form of varioliformgastritis limited to the body which correspondswith a diagnosis of lymphocytic gastritis innearly all cases, the rare exceptions being'subacute' gastritis of imprecise aetiology.On the contrary, the lesions limited to the

gastric antrum can present very different histo-logical appearances. Except for one case of 11,they do not correspond to lymphocytic gastritis.They can be either chronic active (erosive)gastritis of the B type or acute lesions having astheir histological counterpart oedema and super-ficial erosions. The divergent histological find-ings in antral varioliform gastritis have alreadybeen mentioned in the work of Lambert et al19785 who distinguished it clearly from diffusevarioliform gastritis. This point was recentlystressed again by Wyatt and Dixon"2 who refer toreports of Franzin et all3 and Nesland et all4 onantral erosions suggesting that the endoscopicaldiagnosis of antral varioliform gastritis did notcorrespond to a single histological entity butincluded examples of chronic (campylobacterpositive) gastritis, bile reflux gastritis andlymphocytic gastritis. This heterogeneity ofantral varioliform gastritis clearly shows in thepresent work and is in complete contrast withdiffuse and corporeal varioliform gastritis whichconstitute the macroscopical expression of ahomogenous histological entity: lymphocyticgastritis.While there appears to exist perfect concord-

ance between the endoscopic diagnosis of diffuseor corporeal varioliform gastritis and the histo-logical diagnosis of lymphocytic gastritis theconverse does not always apply as in our experi-ence nearly 20% of lymphocytic gastritis do notcorrespond to varioliform gastritis. It could beargued that some of these cases are in factincomplete varioliform gastritis that couldperhaps be diagnosed as such with better train-ing of the examiners. In the present series, weonly found two such cases, one showing nodulesand the other large folds without other lesions.Whether these cases are taken into account or

not, it is nevertheless clear that a part of ourlymphocytic gastritis cases cannot be correlatedwith any precise macroscopic entity. The mostlogical hypothesis is to assume that varioliform

gastritis is only a crude endoscopic expression ofa disease the characteristic feature of which islymphocytic intra-epithelial infiltration. In thishypothesis, the macroscopic disturbances couldonly appear at some periods in the evolution ofthe disease. In favour of this concept are theisolated results we have obtained in patientsfollowed by repeat biopsies. These showed thatin some cases, the macroscopic lesions candisappear while the histological stigmata oflymphocytic gastritis are still evident'" (unpub-lished data). Research currently in progress inwhich a large group of lymphocytic gastritispatients are under follow-up will give a morecomplete answer to this point.Very little is known about the aetiology of

lymphocytic gastritis. The dense intra-epitheliallymphocytic infiltrate is reminiscent of coeliacdisease.910 This suggests the possible involve-ment of alimentary antigens. New researchbased on clinical, epidemiological, and immuno-logical data are required to assess this point.The authors wish to thank Dr M Dixon, Leeds, for reviewing themanuscript and Dr J Bellassai, Assuncion, for technical help.

1 Moutier F, Martin J. Deux cas de gastrite varioliforme. ArchMalApp Digestif 1947; 36: 155-61.

2 Morgan G, Mc Adam WAF, Pyrah RD, Tinsley EGF.Multiple recurring gastric erosions (aphthous ulcers). Gut1976; 17: 633-9.

3 Elta GH, Fawez KA, Dayal Y, et al. Chronic erosive gastritis -a recently recognised disorder. Dig Dis Sci 1983; 28: 7-12.

4 Tsuneoka K, Takemoto I, Fukuchi S. Fiberoscopy of gastricdiseases. Stuttgart: Fischer, 1973.

5 Lambert R, Andre C, Moulinier B, Bugnon B. Diffusevarioliform gastritis. Digestion 1978; 17: 159-67.

6 Haot J, Wallez L, Jouret-Mourin A, Hardy N. La gastrite 'alymphocytes'. Une nouvelle entite? Acta Endosc 1985; 15:187-8.

7 Haot J, Delos M, Wallez L, Hardy N, Lenzen B, Jouret-Mourin A. Les lymphocytes intraepitheliaux en pathologiegastrique inflammatoire. Acta Endosc 1986; 16: 61-5.

8 Haot J, Jouret-Mourin A, Delos M, et al. Etude anatomo-cinique d'une serie de gastrites chroniques caracterisees parune infiltration lymphocytaire intra-epitheliale. Acta Endosc1986; 16:69-74.

9 Dixon MF, Wyatt JI, Burke DA, Rathbone BJ. Lymphocyticgastritis-Relationship to campylobacter pylori infection.JPathol 1988; 154: 125-32.

10 Haot J, Hamichi L, WalIez L, Mainguet P. Lymphocyticgastritis: a newly described entity, a retrospective endo-scopic and histological study. Gut 1988; 29: 1258-64.

11 Haot J, Berger F, Andre C, Moulinier B, Lambert R,Mainguet P. Lymphocytic gastritis versus varioliformgastritis. A historical series revisited. J Pathol 1989; 158:19-22.

12 Wyatt JL, Dixon MF. Chronic gastritis. A pathogeneticapproach. J Pathol 1988; 154: 113-24.

13 Franzin G, Manfrini C, Musola R, Rodella S, Fratton A.Chronic erosions of the stomach - a clinical, endoscopic andhistological evaluation. Endoscopy 1984; 16: 1-5.

14 Nesland AA, Berstad A, Serck-Hanssen A. Histologicalfindings in erosive prepyloric changes. ScandJ Gastroenterol1986; 21: 239-45.

15 Delos M, Jouret-Mourin A, Wallez L, Willette M, MainguetP, Haot J. Evolution histologique d'une serie de gastritescaracterisees par une infiltration lymphocytaire intra-epitheliale. Acta Endosc 1986; 16: 185-7.

on Novem

ber 2, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.3.282 on 1 March 1990. D

ownloaded from