Gut, 1990,31,940-945

REVIEW ARTICLE

Helicobacter pylori: bridging the credibility gap

AW McKinlay, R Upadhyay, C G Gemmell, R I Russell

The earliest descriptions of bacteria colonisingthe stomach date from the 19th and early 20thcentury,' but it was not until 1983 that success-ful culture was achieved.4 As is often the case, itwas a combination of scientific acumen, luck,and determination that led to the isolation ofHelicobacter pylorn. The organism is slowgrowing and early cultures were discardedprematurely. Only during the Easter holiday wassufficient time left for colonies to develop.5

In retrospect, the intervention of a religiousfestival seems entirely appropriate. H pyloriattracts great loyalty from its devotees.'Pylorites' accept that the organism causesgastritis, duodenal ulcer, and non-ulcer dyspep-sia. In contrast 'Schwartzians,' who attribute alldisease to acid, remain sceptical; the undisputedefficacy of the H2 receptor antagonists being thecornerstone of their credo. Currently, there isthe need for a middle ground to bridge thecredibility gap. What facts are accepted and inwhich areas should we focus research?

EpidemiologyH pylon has been found in all human popula-tions studied to date."'0 In western countries thecarriage rate is low among children, but risesprogressively through life. At least 50% of thoseover the age of 50 are positive for H pyloi.6 Incontrast the limited data from the third worldsuggest that a greater percentage of the popula-tion may be colonised in childhood.8' "

The source of infection is unknown. Onereport suggests that the carriage rate amongabattoir workers in close proximity to animals isgreater than in their clerical counterparts." Todate the only animal, other than humans, fromwhich H pylori has been isolated is the rhesusmonkey. 12 There are virtually no reports ofisolations of helicobacter-like organisms fromdomesticated animals.'3 We think that it is morelikely that the source of the organism will proveto be other humans.

Gastroenterology Unitand Department ofMicrobiology, GlasgowRoyal Infirmary, GlasgowA W McKinlayR UpadhyayC G GemmellR I RussellCorrespondence to:Dr AW McKinlay,Gastrointestinal Unit,Woodend Hospital, Aberdeen,AB9 2YS.Accepted for publication23 October 1989

Diagnostic methodsContrary to the expectations of many clinicians,H pylori is not difficult to diagnose. Mostmethods use endoscopic gastric biopsy speci-mens taken from two sites within the stomach,one of which should be the antrum.'4 Theorganism can be readily identified by histopatho-logists provided they are familiar with itscharacteristic appearance. H pyloni can be seenin haematoxylin and eosin stained sections, but a

variety of stains can improve detection.'5-'9Biopsy urease tests, which can be carried out inthe endoscopy room, are easily applied andprovide a useful addition to histology.20 2'Although the combination of positive

histology and a positive biopsy urease test ishighly specific for the organism, we would stillfavour the routine culture of gastric biopsyspecimens. The culture of H pylori involvesmethods that are essentially identical to thoseused in the culture of Campylobacter.22 AsCampylobacterjejuni is now the most commonlyisolated enteric pathogen, the culture ofH pylon'should not be beyond the capability of anycompetent microbiological laboratory.More recently, the urea breath test has been

developed, although this remains confined to anumber of research centres.23 24 Two isotopes ofcarbon have been used. Carbon 14 is radioactiveand easily detected, but is not recommended foruse in children or women of childbearing age.Carbon 13 is not radioactive and has no suchrestrictions, but requires a mass spectrometer forits measurement.25 The advantage of breath testsis that they are non-invasive and allow repeatedmeasurements on the same subject.

Serological diagnostic tests were describedremarkably quickly after the discovery ofH pyloni,2627 and are the subject of continuingdevelopment. A number of different laboratorymethods have been used including haemaggluti-nation26 and complement fixation,27 but theenzyme linked immunosorbent assay (ELISA)has been most widely adopted for routine use.28 29The more exacting technique of Western blot-ting'" has given invaluable insights into thediversity of the human systemic antibodyresponse but is impracticable for mass screen-ing.The choice of antigen for routine diagnostic

use remains problematic. Ideally the antigenshould be specific but highly immunogenic,easily prepared, and stable on storage.3'A number of the earlier antigenic prepara-

tions, such as whole cell sonicates, encompasseda broad range of antigens but often includedflagella proteins, which are highly immunogenicbut are shared with other species, most notablyC jejuni.32 Antibodies to this organism arecommon in the general population, and the netresult was tests that were sensitive but lackedspecificity. Some early epidemiological datamay, therefore, need to be re-examined withmore specific techniques.An acid extractable preparation33 or the ultra-

centrifuged whole cell supernatant are cleaner

940

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.8.940 on 1 August 1990. D

ownloaded from

Helicobacter pylori: bridging the credibilitygap

preparations and produce more acceptableresults when used in ELISA.3'An alternative approach is to use a protein

unique to H pylori, such as its urease, as thebasis for the assay.34 This produces the requiredspecificity but illustrates a second problem.Western blotting shows a remarkably diverseresponse among human subjects to each indi-vidual antigen, which in turn implies that not allsubjects may respond strongly to a single antigentest. Such an assay may be highly specific buthave a lower sensitivity than those using multipleantigens.3'The solution will probably involve combining

a series of highly specific antigens to produce therequired sensitivity.While serology has been invaluable in eluci-

dating the epidemiology of H pylori, its role inclinical screening is at present uncertain andsubject to much speculation.35 36 It has beensuggested that ELISA screening could reduceendoscopy by about one quarter. Patients posi-tive forH pylori by serology would be started ontreatment, and only submitted to endoscopy ifsymptoms persisted. Those negative on sero-logical testing, or in whom gastric carcinoma issuspected, would undergo endoscopy beforetreatment. We believe that this approach hasa number of weaknesses. Firstly, gastriccarcinoma is primarily found in patients over theage of 50 years, 50% ofwhom will be positive forH pylon, and it is frequently not suspected untilendoscopy is performed. A few early carcinomaswill be missed, and while this will not alter theconsequences for most patients, many cliniciansmay be uneasy about this prospect. Secondly,the strategy entails treating some H pylori posi-tive patients with reflux oesophagitis or irritablebowel syndrome with an unnecessary course oftripotassium dicitratobismuthate. Is this justifi-able? Finally, we suspect that the saving inendoscopy time produced by serological screen-ing has been overestimated. In our practicemuch of our endoscopy time is spent checkingfor peptic ulcer healing, in sclerotherapy forvarices, or investigating anaemia or dyspepsia inpatients taking non-steroidal anti-inflammatorydrugs. H pylori status is irrelevant under thesecircumstances.

Serological tests are of limited value aftertreatment for H pylori has been started,although there is some evidence that titres fallafter successful treatment." 37We believe that an attempt to make a positive

diagnosis should be made before the use of anti-microbials is considered.

Clinical associationsCHRONIC GASTRITISThere is a strong association between type Bchronic gastritis and the presence of H pylori,which is found in between 70% and 92% ofpatients with this. 1438 As unbelievers are quick topoint out, however, this is not proof of a causalrelation. In the case of type B gastritis, however,there are good reasons for believing thatH pyloriis more than an innocent bystander.39 Firstly, theassociation has been found in all populationsstudied to date. Secondly, there is a relation,

although admittedly a weak one, between thenumber of organisms and the intensity of theinfiltrate, particularly of the polymorph com-ponent.40"4 Thirdly, a number of groupshave shown that successful eradication of theorganism is accompanied by resolution of thegastritis, and in turn that recurrence ofH pyloriis associated with relapse of the gastritis.4243Finally, two human volunteers have taken livecultures of the organism, and both developed anacute gastritis. In one this proved to be atransient phenomenon and the organism waseliminated before seroconversion occurred.'The second was less fortunate and developed achronic gastritis that was accompanied by sero-conversion.45 These experiments are importantbecause they show a clear temporal relationbetween acquisition of the organism and thesubsequent development of gastritis. Bothvolunteers had normal gastric mucosa before theintroduction ofH pylori.Once the gastric mucosa is colonised by

H pylori, what mechanisms might explain theonset of gastritis? Hazel and Lee' have arguedthat urease may act as an important pathogenicmechanism by creating ammonium ions withinthe mucus-bicarbonate barrier, allowing backdiffusion ofhydrogen ions. Unfortunately, thereis little evidence to support the pivotal role ofurease in this respect. Other urease positiveorganisms found in animals are not associatedwith inflammatory changes.47 H pylon is nowknown to produce a variety of other extracellularproducts,48 including at least one cytotoxin49which is, as yet, poorly characterised. It seemsunlikely that H pylori is entirely innocuous. Theaim for the future must be to clarify its aggressivefactors and place them in perspective against themore established luminal agents such as pepsin,bile, and gastric acid itself.

DUODENAL ULCERDuodenal ulcer shows an even stronger associa-tion withH pylon, with isolation of the organismoccurring in most patients with this disorder. Insome reports, up to 100% of all patients withduodenal ulcer have been found to be positivebut in a review of 10 published studies the meancarriage rate was 86%.5° The association, how-ever, is between duodenal ulcer and carriage ofthe organism in the gastric antrum, rather thanin the duodenum itself. H pylon can be found inthe duodenum, but only on areas of gastricmetaplasia.51 The association between H pyloriand duodenal ulcer is found worldwide. Less isknown about the temporal relation betweencontracting the infection and the development ofduodenal ulcer, and this is potentially of greatimportance. If the organism is causative it mustbe acquired before the ulcer develops. More-over, any attempt to link H pylori and ulcero-genesis must also embrace the established dataOn acid secretion.

Patients with duodenal ulcer have a higherstimulated mean acid output, parietal cell mass,and nocturnal acid secretion than normal people.It has been recognised for many years, however,that there is considerable overlap between ulcerpatients and normal subjects. Wyatt and her

941

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.8.940 on 1 August 1990. D

ownloaded from

McKinlay, Upadhyay, Gemmell, Russell

colleagues have suggested that abnormal acidsecretion induces gastric metaplasia within theduodenal cap, allowing the bacterium to movefrom its natural habitat in the gastric antrum intothe duodenum. This in turn, initiates duo-denitis, which either alone or in conjunction withother factors leads to duodenal ulceration.50The relation between gastric metaplasia and

acid secretion is interesting. Metaplasia is said tooccur only in patients with a low fasting pH, tobe extensive in those with the Zollinger-Ellisonsyndrome, and to correlate with maximal acidoutput.50 Animal experiments also indicate a linkwith acid secretion. Some of the data are old,however, and the subject deserves prospectivere-evaluation.

Gastric metaplasia is essentially a microscopicdiagnosis, and while the determination of itsextent within an individual biopsy specimen isstraightforward, it is difficult to extrapolate thisto the duodenal cap as a whole. The problem ofsampling error will always remain.More importantly, can the hypothesis explain

some of the known differences between theepidemiology of duodenal ulcer and the carriageofH pylori? Duodenal ulcer is more common inthe north of the United Kingdom, affectsyounger people, and shows a male preponder-ance. There is no evidence to date that H pyloriis more common in the north or among men, andits incidence rises progressively with age. Gastricmetaplasia is, however, more common in men.52

Levi and colleagues53 have suggested thatplasma gastrin values are higher in patientscolonised with H pylon than in those who arenot. It is suggested that the release of ureasewithin the mucous layer leads to a rise in pH andpromotes gastrin release. There are importantreservations about these data. The presence ofH pylon was determined by the biopsy ureasetest and 19% of the duodenal ulcers were nega-tive for the organism. It has been suggested thatthis figure is higher than might have beenexpected,54 but a similar proportion has beenreported previously, admittedly by the samegroup, using culture and histology as diagnosticcriteria.55 More importantly, a small study sug-gests that both basal and stimulated gastrinvalues fall after treatment designed to eradicateH pylon from the stomach, although intragastricand nocturnal acid secretion values did notchange.56 Only nine patients were investigated,but the findings suggest that the gastrin linkshould be pursued further. Even more surpris-ingly, it has been suggested that healing resistantduodenal ulcers with omeprazole can clear theorganism from the antral mucosa.57 Again, onlynine patients were studied and endoscopy wasperformed immediately after the end of treat-ment. Confirmation of this finding in largerprospective studies will be required, but it raisesfurther questions about the interaction betweenH pylon and its peculiar environmental niche.Why should an organism that is incapable ofgrowth below pH 4 find omeprazole detrimental,and if the changes in plasma gastrin are genuineare they capable of producing significant altera-tions in acid output? To add further confusion,there is evidence that acute infections withH pylon may lead to hypochlorhydria, although

this is probably only a transient state.58 Finally, itis not clear whether H pylon associated gastritisinfluences mucosal prostaglandin values. It hasbeen suggested that 6-keto F2(, concentrationsare lower in H pylori positive patients than intheir negative counterparts.59 In contrast, thereare contradictory findings regarding prosta-glandin E2, a prostaglandin with a more estab-lished role in mucosal protection.5960 Theinteraction between the infection and host isobviously complex and requires further elucida-tion.

It has been known for many years that bis-muth salts can heal ulcers. The suggestion thatduodenal ulcer relapse occurs less frequentlyafter treatment with tripotassium dicitratobis-muthate (DeNol) was given greater credibility in1981 with the publication of a study comparing itand H2 blockers.6' The discovery thatH pylon issensitive to bismuth in vitro prompted newquestions regarding the mechanism of action oftripotassium dicitratobismuthate and the role ofH pylon in ulcer relapse. A number of studiesnow suggest that in patients with duodenal ulcer,in whom the organism is successfully eradicatedand who remain free of H pylon for longer thanone month, the ulcer relapse rate is 1O-20%.62This is significantly lower than in control groupstreated with H2 blockers. Recurrence of theorganism seems to predict relapse. The problemremains, however, that H pylon may be acting asa marker for other factors operating within thegastric environment. For example, it could beargued that healing the ulcer itself may alterconditions in such a way as to make the uppergastrointestinal tract less hospitable to thebacterium. This seems unlikely because thecarriage rate for H pylon remains unchanged inpatients whose ulcers are healed by H2 block-ers.63 More importantly, the low ulcer recurrencerate may be directly related to tripotassiumdicitratobismuthate itself, which is known tohave actions unrelated to its bactericidal activity.Tripotassium dicitratobismuthate has beenshown to form complexes with mucus,6519making it more effective as a diffusion barrier tohydrogen ion. It reduces pepsin output andactivity70 and promotes prostaglandin synthe-sis."1 Some bismuth is absorbed during treat-ment72 and this may act as a depot within thegastric mucosa. What is less clear, however, iswhether plasma or urine concentrations trulyreflect bismuth values within the gastric mucosa,and whether these concentrations would besufficient to eradicate or suppress the growth ofthe organism. This seems unlikely in view of therelative inefficiency of tripotassium dicitratobis-muthate alone in the eradication of H pylori,with rates as low as 5-33% one month aftertreatment.4373 It cannot be taken as proved,therefore, that H pylon is responsible for thelower duodenal ulcer relapse rate for duodenalulcer after treatment with tripotassium dicitrato-bismuthate, but it is certainly possible that it hasan important role in this respect.

GASTRIC ULCERThe association with gastric ulcer is less strongwith only 65% ofthese patients being positive for

942

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.8.940 on 1 August 1990. D

ownloaded from

Helicobacter pylori: bridging the credibility gap

H pylori.74 Less information is available regard-ing the outcome of treatment and H pyloristatus. Interestingly, bismuth has been shown tobe as effective in healing gastric ulcer as it is induodenal ulcer despite their differing associa-tions withH pylori.75

NON-ULCER DYSPEPSIAAttempting to investigate the relation betweenH pylori and non-ulcer dyspepsia is difficultbecause there is no universally accepted defini-tion of the condition. Retrospective studies haveoften classified non-ulcer dyspepsia as occurringin those patients with dyspepsia who have under-gone a negative upper gastrointestinal endo-scopy. Frequently, the precise symptomatology,the age range of the patients, and details of otherinvestigations are omitted from reports. Notsurprisingly, therefore, there is considerablevariation in the H pyloni carriage rates betweenstudies. Roccas and colleagues foundH pylorn in45% of patients with predominantly upperabdominal pain, which was three times higherthan in asymptomatic controls.76 Similarly, in astudy from Leeds, the incidence of antibody toH pylori among patients with non-ulcer dys-pepsia was twice that ofage matched blood donorcontrols, for every age group.77 A number ofcontrolled studies have been published in whichpatients with defined symptoms have receivedtreatment designed to eradicate H pylori, withsubsequent reassessment.78"88 Elimination ofH pyloni seems to improve symptoms, with theprovisos that these patients may represent onlyone part of a wider spectrum of non-ulcerdyspepsia and that follow up to date has beenshort.

Helicobacter pylori in childrenSerological studies suggest that H pylori coloni-sation is rare among asymptomatic children.6Retrospective studies indicate that H pylori isassociated with both duodenal ulcer and antralgastritis,8' 82 but the actual numbers of positivepatients is small. Prospective studies on childrenwith non-specific abdominal pain suggest thatantral gastritis and H pylori are found morecommonly than would be expected in assympto-matic children of a similar age.8 83 There is someevidence that treatment will improve symptoms,but numbers again are small.84

Children represent a particularly importantpopulation group who may help to clarify ourunderstanding of H pylori infection. Theirupper gastrointestinal tract is subjected to lessself poisoning, with low rates of non-steroidalanti-inflammatory drug usage, little smoking,and only modest consumption of alcohol incomparison to adults. A prospective study of acohort of children, to determine their rate ofacquisition of the organism and subsequent riskof developing peptic ulcer, may help to answersome of the outstanding questions. The 13Cbreath test and serology are minimally invasiveand regular monitoring for H pylori is, there-fore, feasible.

Treatment ofHpylori infectionInitially, it seemed likely that H pylori infectionwould prove an easy target for treatment becausethe organism is sensitive to a wide range ofantimicrobials, including bismuth salts such astripotassium dicitratobismuthate. Surprisingly,McNulty's study in 1986 showed that one of thestrongest contenders, erythromycin, was littlebetter than placebo.85 Moreover, the efficacy ofbismuth containing compounds was almost cer-tainly overestimated in early reports becauseinsufficient time was allowed between treatmentand repeat biopsy specimen examination. Theclearance rate seen immediately after treatmentis high, but falls rapidly over the following fourweeks as the infection recurs.86 This effect maybe due to a declining bismuth depot, but mayalso be related to the organism's ability toproduce metabolically inactive 'coccoid' forms.

Amoxycillin is also disappointing when usedalone,87 and tinidazole produces unacceptablyhigh resistance when employed as a single

88agent. Both agents are more effective whencombined with tripotassium dicitratobismuthate(De-Nol), and this also applies to metronidazolewhere promising eradication rates have recentlybeen described.89 Triple treatment is also effec-tive, but the cost may be greater side effects.909'No regime described to date has proved 100%effective.

This prompts the question: which patientsshould be treated? Duodenal ulcer healing rateswith any of the above combinations are broadlycomparable with those obtained with H2 block-ers, but the drug combinations may be lessacceptable to patients. Eradication of H pylorishould now be considered in patients sufferingfrom duodenal ulcer who have one or morerecurrences after treatment with H2 blockers,and may represent an alternative to long termmaintenance treatment. To achieve low relapserates the organism must be eradicated, neces-sitating repeat biopsy at least four weeks aftertreatment has stopped. It is not clear what actionshould be taken if eradication has not beenachieved, but culture and sensitivity testing maybe helpful if a further course of treatment isbeing considered. Elimination of H pylori maybe another line oftreatment worthy of considera-tion in patients with resistant duodenal ulcers,before surgery. Little data are available, how-ever, on this subject.Some patients with non-ulcer dyspepsia may

also benefit from treatment. It is essential toexclude other organic causes of dyspepsia and toconfirm that the patient has H pylori associatedgastritis, before beginning treatment.From the preceding discussion it can be seen

that many uncertainties remain, and we wouldurge clinicians interested in treating H pyloninfections to enter their patients into clinicaltrials.

SummaryIn summary, therefore, there are interestingassociations between H pylon, duodenal ulcer,and non-ulcer dyspepsia. In type B gastritis theremay be enough evidence to suggest a causal role.The relation between gastritis and upper gastro-

943

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.8.940 on 1 August 1990. D

ownloaded from

944 McKinlay, Upadhyay, Gemmell, Russell

intestinal symptomatology, however, remainscontentious. The relation between H pylon' andacid secretion may be more intimate than waspreviously thought. 'Pylorites' must tempertheir enthusiasm and provide hard data;'Schwartzians' must broaden their horizons.

1 Bottcher. DorpaterMed Z 1874; 5: 148.2 Kasai K, Kobayashi R. Stomach spirochaetes occurring in

mammals. JParasitol 1919; 6: 1-10.3 Doenges JL. Spirochaetes in the gastric glands of Macacus

rhesus and of man without related disease. Arch Pathol 1939;27: 469.

4 Warren JR, Marshall BJ. Unidentified curved bacilli on gastricepithelium in active chronic gastritis. Lancet 1983; i: 1273.

5 Marshall BJ, Royce H, Annear D, et al. Original isolation ofCampylobacter pylonrdis from human gastric mucosa.Microbios Lett 1984; 25: 83-8.

6 Jones DM, Eldridge J, Fox AJ, Sethi P, Whorrell PJ.Antibody to the gastric campylobacter-like organismCampylobacter pyloridis - clinical correlations and distribu-tion in the normal population.J MedMicrobiol 1986; 22: 57-62.

7 Graham DY, Klein PD, Opekun AR, et al. Epidemiology ofCampylobacter pyloridis infection. Gastroenterology 1987; 92:1411.

8 Ramirez-ramos A, Gilman RH, Recaverren S, et al. Campylo-bacter pylonrdis in a developing country. Gastroenterologv1987; 92: 1588.

9 Dwyer B, Nanxiong S, Kaldor J, et al. Antibody response toCampylobacter pylon in an ethnic group lacking pepticulceration. ScandJ Infect Dis 1988; 20: 63-8.

10 Wyatt JI, De Caestecker JS, Rathbone BJ, Heatley RV.Campylobacter pylonrdis in tropical Africa [Abstract]. Gut1987; 28: A1409-10.

11 Vaira D, D'Anastasio CD, Holton J, et al. Campylobacter pylonrin abattoir workers: Is it a zoonosis? Lancet 1988; ii: 725-6.

12 Baskerville A, Newell DG. Naturally occurring chronicgastritis and C. pylon infection in the Rhesus monkey: apotential model for gastritis in man. Gut 1988; 29: 465-72.

13 Goodwin CS, Armstrong J, Bronsdon M, Sly L. Is Campylo-bacter pylorn a zoonosis? Lancet 1988; ii: 968.

14 Marshall BJ, Mcgechie DB, Rogers PA, Glancy RJ. Pyloriccampylobacter infection and gastroduodenal disease. MedJAust 1985; 142: 439-44.

15 Gray SF, Wyatt JI, Rathbone BJ. Simplified techniques foridentifying Campylobacter pyloridis. J Clin Pathol 1986; 39:1279-80.

16 Walters LL, Budin RE, Paull G. Acridine orange to identifyCampylobacter pylonrdis in formalin fixed, paraffin-embedded gastric biopsies. Lancet 1986; i: 42.

17 McMullen L, Walker MM, Bais LA, Karim GN, Baron JH.Histological identification of campylobacter using Gimeneztechnique in gastric antral mucosa. J Clin Pathol 1987; 40:464-5.

18 Trowell JE, Yoong AKH, Saul KJ, Gant PW, Bell GD. Simplehalf-Gram stain for showing presence of Campylobacterpyloridis in sections. J Clin Pathol 1987; 40: 702.

19 Burnett RA, Brown IL, Findlay J. Cresyl fast violet stainingmethod for campylobacter-like organisms. J Clin Pathol1987; 40: 353.

20 McNulty CAM, Wise R. Rapid diagnosis of Campylobacterassociated gastritis. Lancet 1985; i: 1443-4.

21 Marshall BJ, Warren JR, Francis GJ, Langton SR, GoodwinCS, Blincow ED. Rapid urease test in the management ofCampylobacter pyloridis-associated gastritis. Am _7Gastroenterol 1987; 82: 200-10.

22 Goodwin CS, Blincow ED, Warren JR, Waters TE, SandersonCR, Easton L. Evaluation of cultural techniques for isolatingCampylobacter pylonrdis from endoscopic biopsies of thegastric mucosa. J Clin Pathol 1985; 38: 1127-3 1.

23 Marshall BJ, Surveyor I. Fifteen minute ura-C14 breath testfor the diagnosis of Campylobacter associated gastritis.Gastroentrology 1987; 92: 1518.

24 Bell GDA, Weil J, Harrison G, et al. 14-C urea breath analysis,a non-invasive test for Campylobacter pylon' in the stomach.Lancet 1987; i: 1367-8.

25 Graham DY. Klein PD, Evans DJ, et al. Campylobacter pyloridetected non invasively by the 13C-urea breath test. Lancet1987;i: 1174-7.

26 Marshall BJ, Mcgechie DB, Francis Gi, Utley PJ. PyloricCampylobacter serology. Lancet 1984; ii: 281.

27 Jones DM, Lessells AM, Eldridge J. Campylobacter-likeorganisms on the gastric mucosa: culture, histology andserological studies. J Clin Pathol 1984; 37: 1002-6.

28 Kaldor J, Tee W, McCarthy P, Watson J, Dwyer B. Immuneresponse to Campylobacter pylon'dis in patients with pepticulceration. Lancet 1985; i: 921.

29 Rathbone Bi, Wyatt JI, Worsley BW, et al. Systemic and localantibody responses to gastric Campylobacter pyloridis in non-ulcer dyspepsia. Gut 1986; 27: 642-7.

30 Burnie JP, Lee W, Dent iC, McNulty CAM. Immunoblotlfi9n8gerp2rinti5ng of Campylobacter pylori. .7 Med Microbiol

31 Newell DG, Stacey AR. The serology of Campylobacter pyloniinfections. In: Rathbone Bi, Heatley RV, eds. Campylo-bacter pylon and gastroduodenal disease. Oxford: BlackwellScientific, 1989: 74-82.

32 Newall DG. Identification of the outer membrane proteins ofCampylobacter pyloridis and antigenic cross-reactivitybetween C. pyloridis and C. jejuni,. Gen Microbiol 1987;

33 Newell DG. Human antibody response to the outer surfaceantigens of Campylobacter pyloridis. SerodiagnosisImmunother 1987; 1: 209-17.

34 Dent JC, McNulty CAM, Uff JS, Gear MWL, Wilkinson SP.Campylobacter pylon urease: a new serological test. Lancet1988;i: 1002.

35 Loffeld RJLF, Stobberingh E, Flendrig JA, Spreeuwel JPvan, Arends JW. Diagnostic value of an immunoassay todetect anti Campylobacter pylon' antibodies in non-ulcerdyspepsia. Lancet 1989; i: 1182-5.

36 Graham DJ, Evans DJ, Evans DG. Detection of Campvlo-bacterpylori infection. Lancet 1989; ii: 569.

37 Oderda G, Vaira D, Holton J, Ainley C, Altare F, Ansaldi N.Amoxycillin plus tinidazole for Campylobacter pylori gastritisin children: Assessment by serum IgG antibody, PepsinogenI, and gastrin levels. Lancet 1989; i: 690-2.

38 Dixon MF. Campylobacter pylori and chronic gastritis. In:Rathbone BJ, Heatley RV, eds. Campylobacter pylori andgastroduodenal disease. Oxford: Blackwell Scientific, 1989:106-16.

39 McKinlay AW, Upadhyay R, Russell RI. Campylobacterpyloridis, peptic ulcer and gastritis. Scott Med J 1987; 32:68-9.

40 Steer HW. The gastro-duodenal epithelium in peptic ulcera-tion. JPathol 1985; 146: 355-62.

41 Wyatt JI, Rathbone BJ, Heatley RV. Local immune responseto gastric Campylobacter in non-ulcer dyspepsia.. ClinPathol 1986; 39: 863-70.

42 Lambert JR, Dunn KL, Turner H, Korman MG. Effect onhistological gastritis following eradication of Campvlobacterpylondis. Gastroenterology 1986; 90: 1509.

43 Rauws AJ, Langenberg W, Houthoff HJ, Zanen HC, TytgatGNJ. Campylobacterpylonrdis associated chronic active antralgastritis, a prospective study of its prevalence and the effectsof antibacterial and anti-ulcer treatment. Gastroenterology1988; 94: 33-40.

44 Marshall BJ, Armstrong JA, Mcgechie DB, Glancy RJ.Attempt to fulfil Koch's postulates for pyloric campylo-bacter. MedjAust 1985; 142: 436-9.

45 Morris A, Nicholson G. Ingestion of Campylobacter pvloridiscauses gastritis and raised fasting gastric pH. Am Gastro-enterol 1987; 82: 192-9.

46 Hazell SL, Lee A. Campylobacter pvloridis, urease, hydrogenion back diffusion and gastric ulcer. Lancet 1986; ii: 15-7.

47 Tompkins DS, Wyatt JI, Rathbone BJ, West AP. Thecharacterisation and pathological significance of gastriccampylobacter-like organisms in the ferret: a model forchronic gastritis? Epidemiol Infect 1988; 101: 269-78.

48 Megraud F. Campylobacter pylori: Enzymes. In: Rathbone BJ,Heatley RV, eds. Campylobacter pylon' and gastroduodenaldisease. Oxford: Blackwell Scientific, 1989: 39-47.

49 Leunk RD, Johnson PT, David BC, Kraft WG, Morgan DR.Cytotoxic activity in broth-culture filtrates of Campvlobacterpylori.J Med Microbiol 1988; 26: 93-9.

50 Wyatt JI. Campylobacter pylon', duodenitis and duodenalulceration. In: Rathbone BJ, Heatley RV, eds. Campylo-bacter pylon' and gastroduodenal disease. Oxford: BlackwellScientific, 1989: 117-24.

51 Wyatt JI, Dixon MF. Chronic Gastritis - a pathogeneticapproach. ] Pathol 1988; 154: 113-24.

52 Wyatt JI, Rathbone BJ. Gastric metaplasia in the duodenumand Campylobacter pylon. Gastroenterol Clin Biol 1989; 78:78B-82B.

53 Levi S, Beardshall K, Haddad G, Playford R, Ghosh P, CalamJ. Campylobacter pylon' and acid secretion. Lancet 1989; i:1167-8.

54 Smith T, Vaira D, Ainley C, Holton J. Campylobacter pylon'and gastrin and duodenal ulcer. Lancet 1989; ii: 389.

55 Price AB, Levi J, Dolby JM, et al. Campylobacter pylondis inpeptic ulcer disease: microbiology, pathology and scanningelectron microscopy. Gut 1985; 21: 1183-8.

56 McColl KEL, Fullarton GM, Nujami AMI, MacDonald AM,Brown IL, Hilditch TE. Lowered gastrin and gastric acidityafter eradication of Campylobacter pylorn in duodenal ulcer.Lancet 1989; ii: 499-500.

57 Mainguet P, Delmee M, Debongnie J-C. Omeprazole,Campylobacter pylon' and duodenal ulcer. Lancet 1989; ii:389-90.

58 Hunt RH. Campylobacter pylon' and spontaneous hypo-chlorhydria. In: Rathbone BJ, Heatley RV, eds. Campylo-bacter pylori and gastroduodenal disease. Oxford: BlackwellScientific, 1989: 176-84.

59 Goren A, Fotherby KJ, Shorthouse M, Wight DGD, HunterJO. Campylobacter pylori and acid secretion. Lancet 1989; ii:212.

60 Taha AS, Boothman P, Holland P, et al. Duodenal ulcer andCampylobacter pylonr. Lancet 1989; i: 333-4.

61 Martin DF, Hollander D, May SJ, Ravenscroft MM, TweedleDEF, Miller JP. Difference in relapse rates of duodenalulcer after healing with cimetidine or tripotassium dicitratobismuthate. Lancet 1981; i: 7-10.

62 Coghlan JG, Gilligan D, Humphries H, et al. Campylobacterpylon and recurrence of duodenal ulcers: A 12-monthsfollow-up study. Lancet 1987; ii: 1109-11 .

63 Marshall BJ, Goodwin CS, Warren JR, et al. Prospectivedouble-blind trial of duodenal ulcer relapse after eradicationof Campylobacter pylon. Lancet 1988; ii: 1437-42.

64 Smith AC, Price AB, Borriello P, Levi AJ. A comparison ofranitidine and tripotassium dicitratobismuthate TDB inrelapse rates of duodenal ulcer. The role of Campylobacterpylon [Abstract]. Gut 1988; 29: A71 1.

65 Wilson TR. The pharmacology of tripotassium dicitratobis-muthate (TDB). Postgrad Med Jf 1975 51 (suppl 5): 18-

66 Coghill SB, Hopwood D, McPherson S. The ultrastructural

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.8.940 on 1 August 1990. D

ownloaded from

Helicobacter pylori: bridgingthe credibility gap 945

localisation of De-Nol (colloidal bismuth subcitrate - TDB)in the upper gastrointestinal tract of man and rodentsfollowing oral and instrumental administration. J Pathol1983; 139: 105-14.

67 Lee SP. A potential mechanism of action of colloidal bismuthsubcitrate: diffusion barrier to hydrochloric acid. Scand JGastroenterol (suppl) 1982; 80: 17-21.

68 Hollanders D, Morrissey SM, Mehta J. Mucus secretion ingastric ulcer patients treated with tripotassium dicitratobis-muthate (De-Nol). BrJ Clin Pract 1983; 37: 112-4.

69 Koo J, Ho J, Lam SK, Wong J, Ong GB. Selective coating ofgastric ulcer by potassium dicitratobismuthate in the rat.Gastroenterology 1982; 82: 864-70.

70 Baron JH, Barr J, Batten J, Sidebotham R, Spencer J. Acid,pepsin and mucus secretion in patients with gastric andduodenal ulcer before and after colloidal bismuth subcitrate(De-Nol). Gut 1986; 27: 486-90.

71 Konturek SJ, Radecki T, Piastucki I, Drozdowicz D.Advances in understanding of the mechanism of cytoprotec-tive action by colloidal bismuth subcitrate (De-Nol). ScandJ7Gastroenterol 1986; 21 (suppl 122): 6-10.

72 Gavey CJ, Szeto ML, Nwokolo C, Pounder R. Does oraltripotassium dicitrate bismuthate (TDB) provide a depotpreparation of Bismuth? [Abstract]. Gut 1988; 29: A1495.

73 McKinlay AW, Upadhyay R, Boothman P, Gemmell CG,Russell RI. Study to investigate the clearance of Campylo-bacter pyloin by five preparations containing bismuth. In:Megraud F, Lamouliatte H, eds. Gastroduodenal disease andCampylobacter pylon: proceedings of The European Campylo-bacter Study Group, Bordeaux. Amsterdam: Elsevier, 1989.

74 Blaser MJ. Gastric Campylobacter-like organisms, Gastritisand peptic ulcer disease. Gastroenterology 1987; 93: 371-83.

75 Tytgat GNJ. Colloidal bismuth subcitrate in peptic ulcer - Areview. Digestion 1987; 37 (suppl 2): 31-41.

76 Rokkas T, Pursey C, Uzoechina E, et al. Campylobacter pyloriand non-ulcer dyspepsia. AmJ7 Gastroenterol 1987; 82: 1149-52.

77 Shallcross TM, Rathbone BJ, Heatley RV. Campylobacterpylori and non-ulcer dyspepsia. In: Rathbone BJ, HeatleyRV, eds. Campylobacter pylori and gastroduodenal disease.Oxford: Blackwell Scientific, 1989: 155.

78 Rokkas T, Pursey C, Uzoechina E, et al. Non-ulcer dyspepsiaand short term De-Nol therapy: a placebo controlled trialwith particular reference to the role of Campylobacter pylori.Gut 1988; 29: 1386-7.

79 Borody T, Hennessy W, Dascapoulos G, Carrick J, Hazell S.Double-blind trial of De-Nol in non-ulcer dyspepsia associ-ated with Campylobacter pylori gastritis. Gastroenterology1987; 92: 1324.

80 Lambert JR, Borromeo M, Korman MG, Hansky J. Role ofCampylobacter pyloridis in non-ulcer dyspepsia - a random-ised controlled trial. Gastroenterology 1987; 92: 1488.

81 Mahony MJ, Wyatt JI, Littlewood JM. Campylobacterpylorndis-associated gastritis in children. Arch Dis Child1988; 63: 654-5.

82 Drumm B, Sherman P, Cutz E, Karmali M. Association ofCampylobacter pylori on the gastric mucosa with antralgastritis in children. N EnglJ Med 1987; 316: 1557-61.

83 Thomas JE, Eastham EJ, Elliott TSJ, Nelson R. Campylo-bacterpylori associated gastritis in children - a common causeof symptoms [Abstract]. Gut 1988; 29: A707.

84 Eastham EJ, Elliott TSM. Campylobacter pyloridis in children.Arch Dis Child 1987; 62: 657.

85 Mcnulty CAM, Gearty JC, Crump B, et al. Campylobacterpyloridis and associated gastritis: investigator blind placebocontrolled trial of bismuth salicylate and erythromycinethylsuccinate. BrMedJ7 1986; 293: 645-9.

86 Langenberg W, Rauws EAJ, Widjojokusumo A, Tytgat GNJ,Zanen HC. Identification of Campylobacter pylori isolates byrestriction endonuclease DNA analysis. J Clin Microbiol1986; 24: 414-7.

87 Glupczynski Y, Burette A, Labbe M, Deprez C, De Reuck M,Deltenre M. Campylobacter pylon'-associated gastritis: adouble-blind placebo-controlled trial with amoxycillin. AmJ Gastroenterol 1988; 83: 365-72.

88 Goodwin CS, Marshall BJ, Blincow ED, Wilson DH,Blackbourn S, Phillips M. Prevention of nitroimidazoleresistance in Campylobacter pylori by coadministration ofcolloidal bismuth subcitrate: clinical and in-vitro studies.J Clin Pathol 1988; 41: 207-10.

89 Weil J, Bell GD, Gant P, Jones P, Trowell J, Harrison G. Highsuccess rate for eradication of Campylobacter pylon infectionusing a colloidal bismuth subcitrate (CBS) metronidazolecombination [Abstract]. Gut 1989; 30: A733-4.

90 Borody T, Cole P, Noonan S, et al. Longterm Campylobacterpylon recurrence post-eradication [Abstract]. Gastro-enterology 1988; 94: A43.

91 Borsch G, Mai U, Opferkuch W. Oral triple therapy mayeffectively eradicate Campylobacter pylon in man: a pilotstudy [Abstract]. Gastroenterology 1988; 94: A44.

on Septem

ber 27, 2020 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.31.8.940 on 1 August 1990. D

ownloaded from