campylobacterpylor: clinical, histological, serological ... · musgrove,bolton, krypczyk,...

J Clin Pathol 1988;41:1316-1321

Campylobacter pylor: clinical, histological, andserological studiesC MUSGROVE,* F J BOLTON,t A M KRYPCZYK,* J M TEMPERLEY,4 S A CAIRNS,$W G OWEN,* D N HUTCHINSONt

From the Departments of*Histopathology and tMicrobiology, District Laboratory, Preston, and theDepartment ofIMedicine, Royal Preston Hospital, Preston

SUMMARY The presence of Campylobacterpylori, histologically diagnosed gastritis, and antibodiesto C pylori were determined in a series of 113 patients undergoing endoscopy. Paired biopsyspecimens from the fundus, body, and antrum were collected from 59 patients and from the antrum of54 patients. The presence of Cpylori was confirmed by either culture or silver stain in 30 of 59, 31 of59, and 54 of 103 biopsy specimens from the fundus, body, and antrum, respectively. Ofthe specimenswhich contained C pylori 20 of 30 (66%) from the fundus, 25 of 31 (80%) from the body, and 54(100%) from the antrum showed gastritis. Cpylori and gastritis were shown in seven of nine (78 1%)of patients with gastric ulcers and in nine of 11 (82%) of patients with duodenal ulcers. Using anenzyme linked immunosorbent assay (ELISA) technique to detect IgG antibody to C pylori, allpatients with histologically diagnosed gastritis and organisms present had titres of > 640; eight of 39(21 %) ofpatients without gastritis and without organisms gave similar titres. Hence the presence ofCpylori was associated with gastritis and with raised titres of IgG antibody.

Spiral organisms have been described in the stomachofman and other animals by several authors since the1920s.`5 Following the reports of Warren6 andMarshall7 there has been an increase in interest in theorganism which these authors named C pyloridis andwhich has since been redesignated C pylori.8 TheseCampylobacter-like organisms and others9 have beenisolated from gastric biopsy specimens using tech-niques which were developed for the isolation ofintestinal campylobacters.

Despite the profusion of short reports there havebeen few detailed substantiative prospective studies,especially in Great Britain. We therefore present theresults of our study in which 113 patients wereinvestigated endoscopically, histologically, bac-teriologically and serologically. Part of this study wasdesigned to determine the distribution of Campylo-bacter-like organisms in infected or colonisedstomachs, an aspect which has not clearly beenestablished. The survey was also undertaken so that wecould substantiate the association of these organismswith histologically diagnosed gastritis, gastric, andduodenal peptic ulceration. Furthermore, we inves-

Accepted for publication 11 July 1988

tigated the value of an enzyme linked immunosorbentassay (ELISA) technique for detecting the antibodyresponse in patients included in the survey.

Material and methods

A series of 113 patients with a clinical indication forupper gastrointestinal endoscopy, originating fromoutpatient and inpatient referrals over three months in1984 was studied. These patients comprised 71 menand 42 women with an age range of 19-91 years. Thepatients had a variety of symptoms relating to theupper gastrointestinal tract, most commonly, epigas-tric pain or dyspepsia, or both, and some presentedwith an acute gastrointestinal haemorrhage.

ENDOSCOPY AND SAMPLING PROCEDURESA 10 ml sample of blood was taken from each patientfor serological tests, immediately before the endo-scopy procedure. Biopsy specimens were obtainedusing biopsy forceps which had been disinfected in 2%glutaraldehyde and rinsed in distilled water. From thefirst 59 patients paired biopsy specimens from threesites were taken in the following order: gastric fundus,body, and antrum. In the remaining 54 patients pairedbiopsy specimens from the antrum only were

1316

copyright. on A

pril 26, 2021 by guest. Protected by

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.41.12.1316 on 1 D

ecember 1988. D

ownloaded from

http://jcp.bmj.com/

Clinical, serological, and histological studies ofC pyloriobtained. Examination of the duodenum was under-taken last to avoid contamination of the endoscopewith intestinal bacteria.

HISTOLOGYOne biopsy fragment from each site was fixedimmediately in 10% neutral buffered formalin, thenprocessed routinely, and embedded in paraffin wax.Paired 5 gm sections were cut and stained withhaematoxylin and eosin and by a silver impregnationmethod (Warthin-Faulkner).'°

Histological assessment of the stained sections wasmade independently by two histopathologists andincluded: (i) type and thickness of mucosa; (ii)presence or absence of chronic gastritis; and (a)whether quiescent or active in type, and (b) whethersuperficial or full thickness gastritis, using theguidelines of Whitehead"; (iii) glandular atrophy; (iv)intestinal metaplasia; (v) presence and distribution ofCampylobacter-like organisms.

BACTERIOLOGYThe other biopsy specimen ofany pair was placed intoa dry sterile bijoux bottle and transported to thelaboratory where it was cultured within two hours ofcollection. The biopsy specimens were bisected withsterile forceps and the freshly cut surface inoculated onto blood and chocolate agars. These plates wereincubated microaerobically at 37°C for up to five days.Isolates were confirmed as C pylori by colonialmorphology, Gram stain morphology, electronmicroscopy, positive catalase, oxidase and ureasereactions, negative nitrate and hippurate reactions,resistance to nalidixic acid (30 pg disc) and sensitivityto cephalothin (30 pg disc).

SEROLOGYThe antigen for the ELISA test was prepared from sixstrains of Cpylori grown on chocolate agar, incubatedmicroaerobically at 37C for three to four days. Thegrowth was harvested into 10 ml of sterile saline,mixed for one minute on a vortex mixer, centrifuged at3000 rpm for 15 minutes and the supernatant retained.Thiomersal was added to a final concentration of0-01% and the antigen stored at 4C. This antigen wasdiluted 1 in 400 and used to determine the titre ofantibodies to C pylori by a conventional ELISAtechnique."2

Results

HISTOLOGYOne hundred and eight of the 113 patients examinedgave satisfactory biopsy specimens for analysis andthese comprised 59 from whom multiple site specimenswere taken and 49 from whom antral specimens only

1317Table 1 Histologically assessed inflammation in gastricbiopsy specimens

Percentage ofspecimens showing gastritis

Fundus Body AntrumType ofgastritis (n = 59) (n = 59) (n = 103)

None 63 52 40Superficial gastritis

Active 7 5 13Quiescent 20 29 12

Full thickness gastritisActive 7 5 20Quiescent 3 7 10

Not assignable* 0 2 5

*Specimens showing gastritis but of insufficient thickness tosubclassify further.

were taken. Five ofthe 59 patients in the first series hadpreviously had a partial gastrectomy and therefore itwas not possible to collect biopsy specimens from theantrum. The 108 patients studied yielded a total of 59biopsy specimens from the fundus and body and 103from the antrum.Some specimens were of insufficient depth to

differentiate between superficial and full thicknessinflammation and were termed "not assignable". Fullthickness gastritis was not always accompanied byglandular atrophy or intestinal metaplasia. Activity ofinflammation implied high numbers ofpolymorphs inthe lamina propria or within the epithelium. Theresults (table 1) showed an increase in the prevalenceof gastritis, in particular full thickness gastritis andactive gastritis, towards the antrum. In the first series14 cases showed glandular atrophy and two casesshowed intestinal metaplasia at one or more sites. Inthe second series three cases showed intestinal meta-plasia.

Campylobacter-like organisms shown by silverstain were curved or spiral rods, and within anyindividual biopsy specimen the distribution of these

Fig 1 General view ofsuperficial gastric mucosa.Campylobacter-like organisms are seen within surface mucus((m), distributed along the surface epithelium (e), and withinlumen ofgastric pits (p). (Warthin-Faulkner).

copyright. on A


http://jcp.bmj.com

/J C


ecember 1988. D

ownloaded from

http://jcp.bmj.com/

Musgrove, Bolton, Krypczyk, Temperley, Cairns, Owen, Hutchinson

Lk

Fig 2 Demonstration ofCampylobacter-like organismswithin a gastric pit closely related to the surface ofepithelialcells (short arrows). Neutrophilpolymorphs (long arrows)are infiltrating the glandular epithelium. ( Warthin-Faulkner).

organisms was patchy. They were often seen lyingclose to the surface epithelium, within and beneathsurface mucus, and sometimes were seen around theintercellular margins of surface epithelial cells. Theywere most densely distributed within the lumen of thegastric pits (figs I and 2). Much less often theyextended into the lumen of deeper glands. There wereno convincing intracellular organisms, and none wasseen in the lamina propria.

DETECTION OF CAMPYLOBACTER-LIKEORGANISMS AND DENSITY OF COLONISATIONDetection of Campylobacter-like organisms in biopsyspecimens was determined by their presence in silverstained sections or by culture (table 2). Silver stainalone or culture alone failed to detect Campylobacter-like organisms in a few cases, possibly reflectingpatchy distribution ofthe organisms or poor quality ofthe biopsy specimen. When the results of silverstaining and culture are combined then all or none ofthe multiple site biopsy specimens (first series) from anindividual case contained organisms.

Qualitative assessment of Campylobacter-likeorganisms grown from biopsy specimens showed thatthere was a trend of increasing numbers of organismstowards the antrum (fig 3). Thirty per cent of culturepositive biopsy specimens from the fundus produced aheavy (+ + +) or moderate (+ +) growth ofCampylobacter-like organisms compared with 60%from the body and 78% from the antrum.

ASSOCIATION BETWEEN GASTRITIS ANDPRESENCE OF CAMPYLOBACTER-LIKE ORGANISMSOfthose biopsy specimens from the fundus, body, andantrum which contained Campylobacter-like organ-isms 20 of 30, 25 of 31, and 54 of 54, respectively,showed gastritis (table 2). Antral biopsy specimenscontaining Campylobacter-like organisms were three

Table 2 Association between histological gastritis andpresence ofCampylobacter-like organisms

Percentage ofspecimens containing Campylobacter-like-organisms

TotalBiopsy site Detected detected by Not detectedand by silver Detected by either by eitherinflammation stain culture method method

Fundus(n = 59)

Gastritis 30 29 34 3No gastritis 14 15 17 46

Body (n = 59)Gastritis 41 42 42 5No gastritis 10 9 10 43

Antrum(n = 103)

Gastritis 50 46* 52t 7$No gastritis 0 0 0 41

*Ninety eight specimens cultured.tlncludes four specimens that were not cultured.tlncludes one specimen that was not cultured.

times more likely to show full thickness gastritis thansuperficial gastritis and twice as likely to show activerather than quiescent inflammation. Campylobacter-like organisms were seen in the absence of gastritis in10 of 37 fundal biopsy specimens and six of 31 bodybiopsy specimens. In none of the three sites was thereany consistent relation between subjectively assesseddegree ofinflammation and density of colonisation byorganisms. Gastritis in the absence of organisms waspresent in two of 59 (3%), three of 59 (5%), and sevenof 103 (7%) of biopsy specimens from the gastricfundus, body, and antrum, respectively.

ASSOCIATION BETWEEN OTHER GASTRICFINDINGS AND PRESENCE OF CAMPYLOBACTER-LIKE ORGANISMSBiopsy specimens from the fundus which containedCampylobacter-like organisms were twice as likely toshow glandular atrophy as those without organisms;the corresponding figure was one and a half times for

Fig 3 Qualitative growth estimates ofCampylobacter-likeorganisms in biopsy specimensfrom the gastricfundus, body,and antrunm.

1318

copyright. on A


http://jcp.bmj.com

/J C


ecember 1988. D

ownloaded from

http://jcp.bmj.com/

Clinical, serological, and histological studies ofC pyloriTable 3 Comparison ofendoscopy findings with thepresence or absence of Campylobacter-like organisms in 108patients

Percentages ofpatients in each group

Campylobacter-like Campylobacter-likeEndoscopyfindings organisms detected organisms not detected

Normal (n = 29) 39 61Duodenitis 57 43

(n = 21)Bile reflux (n = 22) 50 50Gastritis (n = 52) 50 50Partial 80 10

gastrectomy(n = 5)

Gastric ulcer 78 22(n = 9)

Duodenal ulcer 82 18(n = 11)

the body; no excess was seen in antral biopsyspecimens. Intestinal metaplasia was present in sixspecimens (five subjects). All showed variable degreesof gastritis, and although one of these containedCampylobacter-like organisms, the organisms wereabsent from areas of intestinal metaplasia. Biopsyspecimens taken from sites remote from gastric andduodenal ulcers contained Campylobacter-like organ-isms in seven of nine and nine of 11 cases, respectively,and these also showed gastritis at all or most sites.Campylobacter-like organisms were present in five(100%) of fundal biopsy specimens and in four of five(75%) of body biopsy specimens from cases of partialgastrectomy and were associated with variable degreesof gastritis, of no specific type.

ASSOCIATION BETWEEN ENDOSCOPY FINDINGSAND PRESENCE OF CAMPYLOBACTER-LIKEORGANISMSOf the 108 patients investigated, five had a previouspartial gastrectomy, nine had a gastric ulcer, and 11had a duodenal ulcer. In these patients there was adefinite association between the endoscopy findingsand the presence of Campylobacter-like organisms(table 3). Patients with endoscopically normal gastricmucosa were more likely to have specimens withoutCampylobacter-like organisms. Patients with bilereflux or endoscopically diagnosed gastritis werefound equally likely to yield biopsy specimens with orwithout Campylobacter-like organisms.

SEROLOGYIgG titres determined by an ELISA method are shownin table 4. All of the patients with histologicallydiagnosed gastritis and Campylobacter-like organ-isms detected had IgG titres of > 640. Only eight ofthirty nine (21%) of patients without gastritis andwithout Campylobacter-like organisms gave similartitres. Similarly, all patients with duodenal or gastric

1319ulcers and Campylobacter-like organisms detectedhad IgG titres of > 640 whereas only one of fourpatients with ulceration but without Campylobacter-like organisms gave a similar titre.

Discussion

This study differed from other prospective studies inthat multiple biopsy specimens were taken from thethree different areas of the stomach. There wasexcellent correlation between histological identifica-tion and isolation of organisms from pairedspecimens, a finding in accord with previousstudies.'3 14 The few cases in which small numbers oforganisms were seen histologically and were notcultured may represent patchy distribution as des-cribed by Goodwin et al.'3 The morphology andlocation of the organisms was similar to the descrip-tions by Warren6 and Jones et al.'4 In agreement withthese workers, intracellular organisms were not seen;nor were organisms definitely shown within the mucusof epithelial cells as reported by Meyrick-Thomas etal.'5 The distribution around the luminal margin ofcells, however, was as shown by Phillips et al.'6The overall proportion of antral biopsy specimens

(52%) positive for Campylobacter-like organisms,determined by cultural and histological techniques,was of the same order as most other reports whichrange between 42% and 66%, although in two largerseries of 300 and 222 patients the respective identifica-tion rates were 39% and 35%.8"' In the present studyand in common with other workers we have found that

Table 4 IgG serum antibody titres to Cpylori in endoscopypatients

Percentage ofpatients withELISA IgG titres*

Patient category 4320 640 > 1280

GastritistCampylobacter-like 0 7 93

organisms detected(n = 59)

Campylobacter-like 44 33 23organisms not detected(n = 9)

Gastritis absentCampylobacter-like 100 0 0



Duodenal or gastric ulcerCampylobacter-like 0 12 88



*Titres expressed as reciprocalstHistologically diagnosed

copyright. on A


http://jcp.bmj.com

/J C


ecember 1988. D

ownloaded from

http://jcp.bmj.com/

1320 Musgrove, Bolton, Krypczyk, Temperley, Cairns, Owen, Hutchinsonthere was a clear association between the presence oforganisms and histological gastritis. In the antrumCampylobacter-like organisms were detected in 54(88%) ofthe 61 biopsy specimens showing gastritis butwere not detected in the 43 with normal morphology.This high correlation between gastritis and thepresence of organisms did not extend to the body andfundus, where 16 of 59 (27%) of fundal and 11 of 59(19%) of body biopsy specimens which were notinflamed contained Campylobacter-like organisms. Insimilar patients, however, Rathbone'8 and Goodwin etal'9 reported that gastritis was usually present some-where in the stomach, and in our series this was so inall but two subjects. One of these had had a partialgastrectomy: in the other subject, who had noserological evidence of infection, there was a verysmall number of organisms in the biopsy specimens,perhaps representing contamination during collectionor an early primary infection. In another seriesRollason et aP were unable to find any associationbetween the degree of colonisation by organisms andthe severity of gastritis, and our results accord withthat observation. Marshall and Warren2' have linkedactivity of gastritis with colonisation by Campylobac-ter-like organisms but their series comprised mainlyantral biopsy specimens. Our findings suggest that thislimited sampling may have led to an overestimation ofthe prevalence of active inflammation in associationwith colonisation by Campylobacter-like organisms.

In our series a group of histological features,described by Dixon et al,22 were seen in a few cases withendoscopically diagnosed bile reflux but only in theabsence of Campylobacter-like organisms as reportedby O'Connor et al.23 There was, however, refluxassociated with a very active or quiescent gastritis inalmost equal numbers of cases with and withoutCampylobacter-like organisms present. Attention hasbeen drawn to the absence of Campylobacter-likeorganisms in biopsy specimens showing intestinalmetaplasia,'3 19 and in all but one of the six cases in thepresent study organisms were not detected. Only asmall number of patients with peptic ulceration orpartial gastrectomy for previous peptic ulcerationwere included in this series but at 80% the associationof Campylobacter-like organisms and coexistent gas-tritis with peptic ulceration was of the same order as inother reports.2' 2425

In our study there was excellent correlation betweenthe presence of Campylobacter-like organisms,gastritis, and raised titres of IgG antibody (> 640 byELISA). A similar finding has also been shown byJones et al '` and McNulty et al 26 using complementfixation tests, and this would suggest that the applica-tion of serodiagnosis to gastritis associated withCampylobacter-like organisms could result in areduced need for endoscopy. In three of five patients

with IgG titres of > 640 and gastritis but in whomCampylobacter-like organisms were not shown, multi-ple biopsy specimens had been collected, and in eachsubject gastritis was observed only at one site. Thesefindings help to confirm the concept of patchy dis-tribution of gastritis and probably also patchy dis-tribution of organisms. In the present study IgG titresof > 640 were found in all patients with Campylobac-ter-like organisms, gastric or duodenal ulcers, andgastritis at some location in the stomach, whereas in asimilar study Marshall et al,' using a passive haemag-glutination test, 93% ofsuch patients had high titres ofantibody.Neither the symptoms of the patients investigated

nor the endoscopic findings showed any correlationwith the presence (or absence) of C pylori. Althoughthere is excellent correlation between the presence ofCpylori, gastritis, peptic ulceration and raised IgGantibody titres, the clinical importance of Cam-pylobacter-like organisms in the aetiology of gastricillness has not been defined. This subject has, however,been widely discussed in many papers and extensivelyreviewed.'9 To date, the most persuasive evidence ofanaetiological role is the attempt to fulfil Koch'spostulates by ingesting a culture of Campylobacter-like organisms after first inducing hypochlorhydria.28Under physiological conditions, however, both gas-tritis and Campylobacter-like organism colonisationmay be secondary to some other factor which is lessoperative at a distance from the antrum, thus account-ing for the variable distribution and density of thecolonisation and associated gastritis in other sites.

References

1 Lim RKS. A parasitic spiral organism in the stomach of the cat.Parasitology 1920;12:108-12.

2 Edkins JS. Spirella regaudi in the cat. Parasitology 1923;15:296-307.

3 Cowdrey EV, Scott GH. Effect on monkeys of small doses of aconcentrated preparation of viosterol. Archives of Pathology1936;1: 1-23.

4 Doenges JL. Spirochaetes in the gastric glands of Macacus rhesusand of man without related disease. Archives of Pathology1939;27:469-77.

5 Vial JD, Orrego H. Electron microscope observations on the finestructure of parietal cells. J Biophys Biochem Cytol 1960;7:367-72.

6 Warren JR. Unidentified curved bacilli on gastric epithelium inactive chronic gastritis. Lancet 1983;i:1273.

7 Marshall BJ. Unidentified curved bacilli on gastric epithelium inactive chronic gastritis. Lancet 1983;i: 1273-5.

8 Marshall BJ, Goodwin CS. Revised nomenclature of Campy-lobacter pyloridis. Int J Syst Bacteriol 1987;37:68.

9 Kasper G, Dickgiesser N. Isolation from gastric epithelium ofcampylobacter like bacteria that are distinct from "Campy-lobacter pyloridis'. Lancet 1985;i: 111-2.

10 Culling CFA. In: Handbook ofhistopathology techniques, 2nd ed.London: Butterworths. 1963:325.

11 Whitehead R. In: Mucosal biopsy of the gastro intestinal tract 3rded. Philadelphia: W B Saunders, 1984:41-58.

copyright. on A


http://jcp.bmj.com

/J C


ecember 1988. D

ownloaded from

http://jcp.bmj.com/

Clinical, serological, and histological studies ofCpylori 132112 Voller A, Bidwell DE, Bartlett A. The enzyme linked immuno-

sorbent assay (ELISA): a guide with abstracts of microplateapplications. Guernsey: Summerfield House, 1979.

13 Goodwin CS, Blincow ED, Warren JR, Waters TE, SandersonCR, Easton L. Evaluation of cultural techniques for isolatingCampylobacter pyloridis from endoscopic biopsies of gastricmucosa. J Clin Pathol 1985;38:1127-31.

14 Jones DM, Lessells AM, Eldridge J. Campylobacter-like organ-isms in the gastric mucosa: culture, histology and serologicalstudies. J Clin Pathol 1984;37:1002-6.

15 Meyrick-Thomas J, Poynter D, Gooding C, et al. Gastric spiralbacteria. Lancet 1984;ii:100.

16 Philips AD, Hine KR, Holmes GKJ, Woodings DF. Gastric spiralbacteria. Lancet 1984;ii:100-1.

17 Booth L, Holdstock G, MacBride H, et al. Clinical importance ofCampylobacter pyloridis and associated serum IgG and IgAantibody responses in patients undergoing upper gastrointes-tinal endoscopy. J Clin Pathol 1986;39:215-19.

18 Rathbone BJ, Wyatt JI, Worsley BW, Trijdosiewic LK, HeatlyRV, Locowsky MS. Immune Response to Campylobacterpyloridis. Lancet 1985;i:1217.

19 Goodwin CS, Armstrong JA, Marshall BJ. Cambylobacterpyloridis, gastritis and peptic ulceration. J Clin Pathol1986;39:353-65.

20 Rollason TP, Stone J, Rhodes JM. Spiral organisms in endoscopicbiopsies of the human stomach. J Clin Pathol 1984;37:23-6.

21 Marshall BJ, Warren JR. Unidentfied curved bacilli in the stomach

of patients with gastritis and peptic ulcer. Lancet 1984;i:311-4.22 Dixon MF, O'Connor HJ, Axon ATR, King RFGJ, Johnson D.

Reflux gastritis-a distinct histopathological entity? J ClinPathol 1986;39.524-30.

23 O'Connor HJ, WyattJl, Dixon MF, Axon ATR. Campylobacter-like organisms and reflux gastritis. J Clin Pathol 1986;39:531-4.

24 Marshall BJ, McGechie DB, Rogers PA, Glancy RJ. Pyloriccampylobacter infection and gastroduodenal disease. Med JAust 1985;142:439-4.

25 Lambert JR, Dunn KL, Eaves RL, Korman MG, Hansky J,Punkard K. Pyloric CLO in the human stomach. Med J Aust1985;143:174.

26 McNulty CAM, Crump B, Gearty J, et al. The distribution ofandserological response to campylobacter pyloridis in the stomachand duodenum. In: Pearson AD, Skirrow MB, Lior H, Rowe B,eds. Campylobacter III. PHLS: London, 1985: 174-5.

27 Marshall BJ, McGechie DB, Graham JF, Utley PJ. Pyloriccampylobacter serology. Lancet 1984;ii:281.

28 Marshall RI, Armstrong JA, McGechie DB, Glancy RJ. Attemptto fulfill Koch's postulates for pyloric campylobacter. Med JAust 1985;142:436-9.

Requests for reprints to: Dr D N Hutchinson, Director,Public Health Laboratory, Royal Infirmary, Meadow Street,Preston PRI 6PS, Lancs.

copyright. on A


http://jcp.bmj.com

/J C


ecember 1988. D

ownloaded from

http://jcp.bmj.com/

campylobacterpylor: clinical, histological, serological ... · musgrove,bolton, krypczyk,...

Documents