Microbial and anaemiaand hypogammaglobulinaemiaVolatile fatty acids were detected by use of Chromosorb 101 (Phase Separations Ltd) pre-treated bythe methodofBarrettetal.'8 Thecolumn

J Clin Pathol 1985;38:946-953

Microbial and metabolic profile of achlorhydricstomach: comparison of pernicious anaemia andhypogammaglobulinaemiaSP BORRIELLO, PJ REED, JM DOLBY, FE BARCLAY, ADB WEBSTER*

From the Divisions of Communicable Diseases and Immunological Medicine, * Clinical Research Centre,Watford Road, Harrow, Middlesex

SUMMARY The microbial flora and some of its metabolites and enzymes in the stomach werecompared in patients with achlorhydria, pernicious anaemia, and primary hypogammaglo-bulinaemia and in patients with dyspepsia with normal gastric acidity. Detailed analysis of theflora of the gastric juice and of the mucosa from the antrum, body, and fundus in six patients withhypogammaglobulinaemia (mean pH 8.2), seven patients with pernicious anaemia (mean pH7.3), and five patients with dyspepsia (mean pH 1.9) yielded 22 different genera of bacteria,mainly from the patients with achlorhydria, the most common being streptococci, micrococci,staphylococci, veillonella, and lactobacilli. A similar flora was found associated with the mucosaat all three sites. Various metabolites were also looked for. , Glucoronidase and C,4 lipase werefound in patients with hypogammaglobulinaemia but not in those with pernicious anaemia or dys-pepsia. Volatile fatty acids were not found. Relatively high concentrations of ethanol were foundin the patients with hypogammaglobulinaemia compared with those with pernicious anaemia (p= 0.02). Similar concentrations of dimethylamine were found in all three groups, but the con-centrations of trimethylamine were much higher in patients with pernicious anaemia andhypogammaglobulinaemia.The high concentrations of some microbial enzymes and ethanol differentiated the group with

hypogammaglobulinaemia from the rest, and these may bear some relation to the high incidenceof gastric cancer in patients with hypogammaglobulinaemia.

Considerable interest in the microbial flora of theachlorhydric stomach,'-' followed the suggestionthat carcinogenic nitrosamines may be formed atnear neutral pH by the bacterial reduction of dietarynitrates to nitrites, with subsequent N nitrosation ofsecondary amines.9 This theory may explain theincreased incidence of gastric cancer in patients withachlorhydria, especially those with perniciousanaemia or hypogammaglobulinaemia. Most of thework has been concentrated on analysis of the bac-terial flora and nitrate and nitrite concentrations ofgastric juice' 7-9; little attention has been paid to themicrobial flora associated with the mucosa in differ-ent regions of the stomach. Possibly, metabolitesproduced at the mucosal surface are more importantthan those produced in the lumen. In addition, thepresence of amines, cocarcinogens, or other possible

Accepted for publication 8 May 1985

markers of disease have rarely been looked for.We performed a detailed analysis of the microbial

flora of the gastric juice and gastric mucosal flora atthree different sites in patients with achlorhydria andhypogammaglobulinaemia or pernicious anaemia.The findings were compared with those obtained ina group of subjects with low gastric pH values. Thegross microbial findings are presented in briefelsewhere.' In addition, the juices were analysed forthe presence of various enzymes, short chain fattyacids, amines, and alcohols. The aim of these laterinvestigations was to identify and measure thoseamines that might affect the formation of nitro-samines and to determine whether there were othermicrobial metabolites present that could allow us todifferentiate between the groups of patients.

Material and methods

Three groups of patients, described in detail else-946

on June 14, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.38.8.946 on 1 A

ugust 1985. Dow

nloaded from

http://jcp.bmj.com/

Microbial and metabolic profile of achlorhydric stomachwhere,' were investigated. Six adult patients withprimary hypogammaglobulinaemia of late onset,known to have pentagastrin fast achlorhydria, werestudied. Five of these patients were receiving weeklyreplacement treatment with gammaglobulin. Onlyone patient had taken antibiotics (amoxycillin andflucloxacillin) within a week of the investigation.The second group comprised seven patients withtypical pernicious anaemia responsive to vitaminB,2 who were also unable to produce stomach acidafter an injection of pentagastrin. Both these groupsunderwent routine endoscopy as a screening proce-dure for early gastric carcinoma; they had giveninformed consent for the procedures outlined below.The third group comprised five patients who under-went endoscopy for dyspepsia but who were subse-quently considered not to have significant disease.

COLLECTION OF GASTRIC JUICE AND BIOPSYMATERIALSamples of gastric juice were collected at endoscopyas described previously' and stored frozen at - 30°Cas a 1:10 dilution in glycerol broth for bacteriologi-cal analysis'0 and neat for biochemical analysis.Biopsies from the antrum, body, and fundusobtained during this investigation were individuallyplaced into 4-5 ml volumes of glycerol broth, trans-ferred to an anaerobe cabinet (Forma Scientific,Ohio) within 10 minutes of collection, washed threetimes with 5 ml of a brain heart infusion broth:physiological saline mix (1:1), and then placed in apreweighed bijou containing 2 ml glycerol broth.These were reweighed, the weight of the biopsy(about 10 mg) was determined, and they were thenrapidly frozen and stored at -30°C.

BACTERIOLOGICAL INVESTIGATIONSBiopsies and juice were thawed in the anaerobecabinet. Tissue was homogenised with sterile glassgrinders in the suspending fluid, and these and thejuice diluted 10-fold over six steps in prereducedbrain heart infusion (Oxoid) with added reducingagents and growth factors." Quantitative estimateswere made according to the method of Miles andMisra.'2 To detect small numbers a spread plate of0*1 ml of the lowest dilution was made. The follow-ing range of media were used: reinforced clostridialagar (Oxoid) with 7-5% horse blood for anaerobes;Columbia agar (Oxoid) with 5% horse blood foraerobes; selective media for bacteroides, veillonellaand bifidobacteria," lactobacilli,'3 streptococci'4;and MacConkey medium (Oxoid) for enterobac-teriaceae. For the specific isolation of clostridiatreatment with ethanol was used to select forspores'5 and the treated suspensions seeded on to amedium containing egg yolk.'6

All agar media for the isolation of anaerobes wereprereduced and after inoculation incubated in anatmosphere of 10% carbon dioxide, 10% hydrogen,and 80% nitrogen, at 37°C for five days in anincubator housed in an anaerobe cabinet. MacCon-key agar medium was incubated at 37°C in air over-night. All other media were incubated at 37°C incandle jars for two days. Organisms were identifiedby standard criteria, including analysis of the volatilefatty acid end products of metabolism by gas-liquidchromatography.'7 Triplicate mean counts weredetermined and the results recorded as the log,0number of organisms/ml or g.

GAS CHROMATOGRAPHIC ANALYSIS OF GASTRICJUICEUndiluted samples were analysed on a Varian 3700chromatograph with dual flame ionisation detectors.In all cases glass columns five feet long with aninternal diameter of 2 mm were used. To detectamines the column was soaked overnight in 2Mpotassium hydroxide, dried without rinsing, packed(leaving a 7 cm gap at the top) with 4% Carbowax20M10 8% potassium hydroxide coated on to Car-bopak B (Supelco Inc, Chromatography Services,Carr Lane Industrial Estate, Hoylake, Wirral,Merseyside), and plugged at both ends with glasswool that had been soaked in potassium hydroxide.The column was conditioned by heating to 220°Covernight, then injecting 30 x 10 ,ul aliquots of dis-tilled water at intervals of one minute before reduc-ing the temperature. Working conditions were:injector 220°C, detector 250°C, column 70°C (forfive minutes), rising at 7°C/min to 135°C (for 20minutes), carrier gas nitrogen 23 ml/min, andattenuation 8 x 10-'2/amps/mv. Volumes of 0-1 IlIwere coinjected on to the column with equalvolumes of 0-2M potassium hydroxide. Standardsolutions of dimethylamine (10-8 uglml), retentiontime 1-98 minutes, and trimethylamine (18-5 ,ug/ml),retention time 3*29 minutes, were also run and theamine content of the samples estimated by compar-ing peak heights with these external standards.Ethanol was detected by use of Chromosorb 103

(Phase Separations Ltd, Deeside Industrial Estate,Queensferry, Clwyd). The column was washed with10% hydrochloric acid and 2M sodium hydroxidebefore packing. Working conditions were: injector110°C, detector 180°C, column 105°C (for fiveminutes), rising at 10°C/min to 165°C, carrier gasnitrogen 23 mlmin, attenuation 32 x 10-'2/amps/mv. Volumes of 1 ,ul were coinjected on to thecolumn with an equal volume of 0-05M sodiumhydroxide. The ethanol content of samples wasestimated by comparing integrated peak areas withthose of a standard solution of ethanol (71 ,ug/ml).

947

on June 14, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.38.8.946 on 1 A

ugust 1985. Dow

nloaded from

http://jcp.bmj.com/

948

Standard solutions of ethanol gave a linear responsefrom 10 ,g/ml to 1000 ,ug/ml.

Volatile fatty acids were detected by use ofChromosorb 101 (Phase Separations Ltd) pre-treated by the method of Barrett et al.'8 The columnwas washed with 10% hydrochloric acid beforepacking. Working conditions were: injector 180°C,detector 200°C, column 170°C, carrier gas nitrogen23 m/min, and attenuation 4 x 10- "/amps/mv.One ,ul of sample was coinjected on to the columnwith an equal volume of 0 1% formic acid.

MEASUREMENT OF GASTRIC JUICE pHImmediately after sampling the pH of the gastricjuice was measured with a microglass electrode (PyeUnicam model 290) calibrated against standard buf-fers. In one of the patients with pernicious anaemiathe pH was measured after storage.

DETECTION OF GASTRIC JUICE ENZYMESThe undiluted gastric juices were screened for thepresence of acid and alkaline phosphatase, C4 ester-ase, C8 esterase lipase, C,4 lipase, leucine, valine andcystine arylamidase, trypsin, chymotrypsin, phos-phoamidase, a and ,8 galactosidase, ,3 glucuronidase,a and /8 glucosidase, N acetyl glucosaminidase, amannosidase, and a fucosidase by use of the APIZym system (API Laboratory Products Ltd, Basing-stoke, Hampshire) according to the manufacturer'sinstructions.

Borriello, Reed, Dolby, Barclay, Webster

Statistical analysis The data were analysed bythe method of Hill and Peto,'9 which is suited toanalyses of small numbers of results.

Results

pH VALUES OF GASTRIC JUICEThe pH of the gastric juice was alkaline in all six ofthe patients with hypogammaglobulinaemia, with amean value of 8-2 (range 7-8 to 8.4). All of thepatients with pernicious anaemia had achlorhydriawith pH values close to neutral, with a mean value of7-3 (range 6-8 to 7.5). The five patients with dyspep-sia in the group for comparison had acid gastric juiceof mean pH 1-9 (range 1-3 to 3.4). The gastric pHvalues of the patients with hypogamma-globulinaemia were significantly higher than those ofthe patients with pernicious anaemia (p = 0-001).

BACTERIAL FLORAA great variety of different micro-organisms wereisolated from the juices and mucosal biopsies of thestomachs of the patients with achlorhydria (Table1). In general, the most commonly found bacteriawere veillonella, lactobacilli, and facultative Grampositive cocci such as streptococci, micrococci, andstaphylococci (Table 1). Klebsiella were found inthe stomachs of five of the seven patients with per-nicious anaemia, but in only one with hypogamma-globulinaemia (Table 1). Escherichia coli was not

Table 1 Micro-organisms isolated from all sites in stomach (figures are numbers ofpatients)Micro-organism Patients with hypogammaglobulinaemia Patients with pernicious anaemia Group with low pH

(n =6) (n =7) (n= 5)Gram positve cocci:

Aerobic streptococci 6 7 3Anaerobic streptococci 1 1 0Micrococci 6 6 2Staphylococci 6 5 5

Gram negatve cocci:Neisseria 2 1 0Veillonella 6 7 1

Gram positve rods:Lactobacilli 5 6 2Diptheroides 2 4 1Bifldobacteria 4 6 1Eubacteria 1 1 0Clostridia 0 1 0Bacilli 0 1 0Propionibacteria 1 2 0

Gram negative rods:Klebsiella 1 5 0Pseudomonas 1 3 0Haemophilus 1 1 0Bacteroides 3 0 0Fusobacteria 0 0 0Citrobacter 0 1 0Alcaligenes 0 0 0Proteus 1 0 0Hafnia 1 0 0

Yeasts 0 2 0

on June 14, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.38.8.946 on 1 A

ugust 1985. Dow

nloaded from

http://jcp.bmj.com/

Microbial and metabolic profile of achlorhydric stomachTable 2 Micro-organisms isolated from gastric juice (figures are numbers ofpatients and mean log,, No oforganismslmljuice)

Micro-organism Patients with hypogammaglobulinaemia Patients with pernicious anaemia Group with low pH(n = 6) (n = 7) (n = 5)

Gram positive cocci:Aerobic streptococci 5 (6-6) 7 (7-1) 2 (3-3)Anaerobic streptococci 1 (4-7) 1 (6.6) 0Micrococci 2 (6.3) 5 (5-8) 1 (2.6)Staphylococci 5 (4-3) 3 (5-0) 3 (3.2)

Gram negative cocci:Neisseria 1 (5 0) 0 0Veillonella 3 (6-1) 6 (4.7) 1 (2-0)

Gram positive rods:Lactobacilli 3 (3 8) 5 (4-6) 1 (3 0)Diptheroides 1 (6-0) 2 (6-5) 0Bifidobacteria 1 (7-3) 6 (5-9) 0Clostfidia 0 0 0Bacilli 0 1 (3-0) 0Propionibacteria 0 2 (5-2) 0

Gram negative rods:Klebsiella 1 (4-1) 5 (4.2) 0Pseudomonas 1 (6 0) 3 (4 2) 0Haemophilus 1(8-0) 0 0Bacteroides 1(4-7) 0 0Fusobacteria 0 0 0Citrobacter 0 1 (4-0) 0Alcaligenes 0 0 0Proteus 1 (3 0) 0 0Hafnia 1(4-4) 0 0

Yeasts 0 2 (3-2) 0

isolated from any of the groups. Far fewer types ofbacteria were found in the group with low pH; theseconsisted mainly of staphylococci, which were foundin all of the patients, and streptococci (Table 1).A similar pattern of colonisation was found when

analysing the microbial flora of only the gastric juiceof these patients (Table 2), although micrococci andveillonella were not always detected in the juice inpatients with hypogammaglobulinaemia despitetheir presence in the gastric mucosa (Tables 3, 4,and 5). When present, streptococci tended to be thedominant organism (Table 2), although bifidobac-

teria and haemophili attained high concentrations inone of the patients with hypogammaglobulinaemia(Table 2). In general, from a combined qualitativeand quantitative aspect, the patients with perniciousanaemia had the most complex gastric juice flora. Asexpected, the group with low pH had a much simplerflora and harboured much lower concentrations ofthese bacteria (Table 2).

Tables 3, 4, and 5 show the types of bacteria foundin association with the gastric mucosa of the antrumbody and fundus. Within any one group the patientswith hypogammaglobulinaemia and those with per-

Table 3 Micro-organisms associated with gastric mucosa ofantrum (figures are numbers ofpatients and mean log1, No oforganisms/g tissue)

Micro-organism Patients with hypogammaglobulinaemia Patients with pernicious anaemia Group with low pH(n = 6) (n = 7) (n = 5)

Gram positive cocci:Aerobic streptococci 6 (5 6) 7 (5.7) 0Anaerobic streptococci 0 1 (3-4) 0Micrococci 3 (4-6) 5 (4.3) 0Staphylococci 3 (4-4) 2 (4.0) 2 (3.8)

Gram negative cocci:Neisseria 1(6.-1) 1(4-4) 0Veillonella 3 (5.1) 6 (4-2) 0

Gram positive rods:Lactobacilli 1 (3 3) 1 (4-1) 0Diptheroides 1 (5-4) 3 (4.0) 0Bifidobacteria 2 (7-0) 5 (4-1) 0Eubacteria 1 (4-2) 1 (3-6) 0

Gram negative rods:Haemophilus 1 (5-7) 1 (3-6) 0Bacteroides 2 (5-5) 0 0Proteus 1 (4-0) 0 0Hafnia 1 (4-6) 0 0

949

on June 14, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.38.8.946 on 1 A

ugust 1985. Dow

nloaded from

http://jcp.bmj.com/

Borriello, Reed, Dolby, Barclay, Webster

Table 4 Micro-organisms associated with gastric mucosa ofstomach body (figures are numbers ofpatients and mean log,,No oforganismslg tissue)

Micro-organism Patents with hypogammaglobulinaemia Patients with pernicious anaemia Group with low pH(n = 6) (n = 7) (n = S)

Gram positve cocci:Aerobic streptococci 6 (59) 7 (65) 3 4Micrococci 5 (4-6 5 46 1 40Staphylococci 2 (4.4 2)522 38

Gram negative cocci:Neisseria 1 (55) 1 (4-7) 0Veillonella 4 (4-7) 6 (4 9) 0

Gram positive rods:Lactobacilli 1 (4 6) 3 (3-8) 1 (4-6)Diptheroides 0 3 (4-4) 0Bifidobacteria 1 (2-7) 5 {4.2) 1 (3-6)Propionibacteria 0 1 (60) 1 (40)

Gram negatve rods:Haemophilus 1 (5-3) 0 0Bacteroides 1 (44) 0 0Hafnia 1 4.1) 0 0

nicious anaemia were colonised as commonly at theantrum as at the body and fundus, whereas thegroup with low pH had a more complex flora local-ised in the body of the stomach rather than theantrum and fundus. In keeping with the findings ingastric juice, the bacteria most commonly associatedwith the gastric mucosa in all three groups werestreptococci, micrococci, staphylococci, and veil-lonellae. Interestingly, klebsiella, though present inthe gastric juice of five of the patients with pernici-ous anaemia, were not associated with the gastricmucosa. Overall, however, the flora associated withthe mucosa did not differ significantly from that ofthe gastric juice.

GASTRIC JUICE ENZYMESTable 6 shows the results. All the gastric juicestested contained phosphoamidase activity, and noneof them contained chymotrypsin, a galactosidase, or

a mannosidase. All the juices tested from thepatients with hypogammaglobulinaemia and per-nicious anaemia contained alkaline and acid phos-phatase, C4 and C8 esterase, and C8 esterase lipase.The only enzymatic activity noted in the group withlow pH was phosphoamidase and C4 and C8 ester-ase. The presence of /8 glucuronidase and C,4 lipase,which was present in the same three patients withhypogammaglobulinaemia, seemed to differentiatethis group of patients from the others.

GASTRIC JUICE AMINES, ETHANOL, AND FATTYACIDSTable 7 shows the results for gastric juice aminesand ethanol, with the corresponding pH values ofgastric juice and degree of bacterial colonisation.Short chain fatty acids were not detected in any ofthe samples analysed, and the only alcohol detectedwas ethanol. Neither of the phenolics, p-cresol or

Table 5 Micro-organisms associated with gastic mucosa offundus (figures are numbers ofpatients and mean log10No oforganisms/g tissue)

Micro-organism Patents with hypogammaglobulinaemia Patents with pernicious anaemia Group with low pH(n = 6) (n = 7) (n = 5)

Gram positive cocci:Aerobic streptococci 6 (6-0) 7 (6-0) 1 (5.2)Anaerobic streptococci 0 1 (4-1) 0Micrococci 3 (4-9) 6 (4.6) 0Staphylococci 2 (6-4) 3 (4-2) 0

Gram negatve cocci:Neisseria 1 (5.0) 0 0Veillonella 5 (4-7) 5 (4-5) 0

Gram positive rods:Lactobacilli 0 3 (3.8) 0Diptheroides 1 (5.2) 1 (3-5) 1 (3-3)Bifidobacteria 3 (5-2) 6 (5-3) 0Propionibacteria 1 (3-6) 0 0

Gram negative rods:Haemophilus 1 (4-2) 0 0Bacteroides 1 (4-1) 0 0Hafnia 1 (3-5) 0 0

950

on June 14, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.38.8.946 on 1 A

ugust 1985. Dow

nloaded from

http://jcp.bmj.com/

Microbial and metabolic profile ofachlorhydric stomachTable 6 Enzymatic activity present in gastric juice (figures are numbers ofpatients)

Enzyme -- Patients with hypogammaglobulinaemia Patients with pernicious anaemia Group with low pH(n =4) (n= 5) (n = 4)

Alkaline and acid phosphatase4 5 0C4 and C8 esterase 4 5 2C,4 lipase 3 1 0C. esterase lipase 4 5 0Leucine arylamidase 2 5 0Valine arylamidase 2 1 0Cystine arylamidase 1 1 0Trypsin 2 1 0Chymotrypsin, a galactosidaseand a mannosidase 0 0 0

,/ Galactosidase and,8 Glucosidase 1 0 0

/3 Glucuronidase 3 0 0a Glucosidase 3 4 0N acetyl-/8-glucosaminidase 3 2 0a Fucosidase 0 1 0Phosphoamidase 4 5 4Mean pH of juice 8-1 7-3 1-6Mean log,0 No of bacterialml juice 5-8 6-6

952

seven patients who underwent endoscopy for sus-pected carcinoma. The three positive samples werefrom patients with confirmed gastric cancer.

There was little overall difference in the gastricflora between different sites in the stomach in eitherpatients with hypogammaglobulinaemia or thosewith pernicious anaemia. This suggests that thedegree of gastric atrophy has little influence on theattachment of bacteria to mucosa, as most of thepatients with hypogammaglobulinaemia had a pan-gastritis affecting the antrum, unlike most of thosewith pernicious anaemia in whom the antral mucosawas normal.2'

Despite the attention that the theory that nitro-samines and gastric cancer are related has received,most of the work has concentrated on the presenceof bacteria that reduce nitrate' 5 7 and the concen-trations of nitrites in gastric juice.' 7-9 Most workershave assumed that amines from foodstuffs22-24would be present, and this has been confirmed bythe indirect finding of nitrosatable amines8 and thedirect finding of N nitrosamines25 in gastric juice.We directly screened for, identified, and quantifiedsome of the amines present. Of straight chainamines up to C6, only dimethylamine andtrimethylamine were found. All three groups ofpatients had similar concentrations ofdimethylamine, and it is therefore unlikely that thedimethylamine was the product of microbialmetabolism in the stomach. The most likely sourceof this amine is the diet. In contrast, the concentra-tion of trimethylamine, which reacts only poorlywith nitrite, was higher in the patients withhypogammaglobulinaemia and pernicious anaemiathan in those with dyspepsia. We are unable toexplain this finding.

Although all of the ingredients required for theproduction of nitrosamines are present in the ach-lorhydric stomachs of these patients-that is, bac-teria that reduce nitrate, nitrite, and nitrosatableamines-the procarcinogenic nitrosamines requiremetabolic activation to exert their carcinogeniceffect. It has been shown that explant cultures ofhuman colonic cells can metabolise dimethylnit-rosamine to the active carcinogen,26 and gastric cellswould probably be equally active. An alternativemethod of this activating nitrosamine would be amicrobial mediated nitroso group exchange betweenan amide and a nitrosamine to form a highly reactivenitrosamide. This has been shown in vitro for vari-ous amides, dimethylnitrosamine, and E coli.27Although we did not measure the concentrations ofN nitroso compounds in gastric juice in this study,concentrations of about 0 5 uM/kg have been foundin the gastric juice of two patients with achlorhydriaand hypogammaglobulinaemia (Smith and Webster,

Borriello, Reed, Dolby, Barclay, Webster

unpublished findings).The gastric juices were also analysed for microbial

metabolites and gastric juice enzymes in an attemptto find markers that could differentiate states of dis-ease. A surprising finding was the presence ofethanol in the gastric juice of these subjects. Noexogenous source for this ethanol could be found;the patients had not been drinking, and it wasassumed that its presence was the result of microbialactivity. A similar finding and conclusion werereached by others who detected ethanol in thestomachs of patients taking antacids or cimetadine.28We found no correlation, however, between con-centrations of ethanol and the degree of microbialcolonisation, the presence or absence of yeasts, orthe pH of gastric juice. The most striking finding wasthe low concentrations of ethanol found in thepatients with pernicious anaemia compared with thetwo other groups. The patients with hypogamma-globulinaemia had particularly high concentrations.Alcohol may have a role in the aetiology ofoesophageal cancer as it facilitates the penetrationof nitrosamines into mucosa,29 and a similar situa-tion may occur in the stomach.The total activity of various enzymes in gastric

juices such as lactic dehydrogenase and ,3glucuronidase30 is raised in gastric cancer. Possibly,enzymes derived from microbes, such as /3glucuronidase, could contribute to these changes.We found a wide range of enzymatic activity in thegastric juices of the patients with achlorhydria andlesser activity in the group with normal acid pH dys-pepsia. In view of the findings of an association be-tween high levels of 8 glucuronidase and gastriccancer, it is interesting that this enzyme seemed todifferentiate the patients with hypogamma-globulinaemia, who have a high risk of gastriccancer, from the other patients.3' These raisedenzyme activities, however, may merely reflect thegreater extent of gastritis in these patients.2' Anadditional differentiating marker, Cl4 lipase, wasalso higher in the patients with hypogamma-globulinaemia, and it may prove worth while toinclude this enzyme in any further studies on thevalue of screening gastric juice enzymes in gastriccancer.

In conclusion, this study shows that there is acomplex microbial flora in the gastric juice andassociated with the gastric mucosa of the achlor-hydric stomach. The metabolic potential of thisflora is high, and several possible microbial metabo-lites and enzymes were detected. There is no evi-dence, however, that these bacteria are responsiblefor producing carcinogens or for the much higherincidence of gastric carcinoma in patients withhypogammaglobulinaemia compared with those suf-

on June 14, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.38.8.946 on 1 A

ugust 1985. Dow

nloaded from

http://jcp.bmj.com/

Microbial and metabolic profile of achlorhydric stomach

fering from typical pernicious anaemia. The particu-larly high concentrations of ethanol in these patientsremains unexplained, and further investigation intotheir origin might uncover further differences in thegastric environment between patients deficient inantibodies and those with typical perniciousanaemia.

References

Dolby JM, Webster ADB, Borriello SP, Barclay FE, Barth-olomew BA, Hill MJ. Bacterial colonization and nitrite con-centrations in the achlorhydric stomachs of patients with prim-ary hypogammaglobulinaemia or classical pernicious anaemia.Scand J Gastroenterol 1984; 19:105-10.

2 Drasar BS. Shiner M, McLeod GM. Studies on the intestinalflora. 1. The bacteria of the gastrointestinal tract in healthyand achlorhydric persons. Gastroenterology 1969; 56:71-9.

3 Gray JDA, Shiner M. Influence of gastric pH on gastric andjejunal flora. Gut 1967;8:574-81.

4Hillman KM, Riordan T, O Farrell SM. Tabaqchali S. Coloniza-tion of the gastric contents in critically ill patients. Crit CareMed 1982; 10: 444-7.

Ruddell WSJ, Axon ATR, Findlay JM, Bartholomew BA,Hill MJ. Effect of cimetidine on the gastric bacterial flora.Lancet 1980;i:672-4.

6 Snepar R, Poporad GA, Romano JM, Kobasa WD, Kaye D.Effect of cimetidine and antacid on gastric microbial flora.Infect Immun 1982;36:518-24.

7Stockbrugger RW, Cotton PB, Eugenides N, Bartholomew BA,Hill MJ, Walters CL. Intragastric nitrites, nitrosamines, andbacterial overgrowth during cimetidine treatment. Gut1982;23: 1048-54.

8 Ruddell WSJ, Bone ES, Hill MJ, Walters CL. Pathogenesis ofgastric cancer in pernicious anaemia. Lancet 1978;i:521-3.

Ruddell WSJ, Bone ES, Hill MJ, Blendis LM, Walters CL.Gastric-juice nitrite. A risk factor for cancer in the hypoch-lorhydric stomach? Lancet 1 976;ii: 1037-9.

'° Crowther JS. Transport and storage of faeces for bacteriologicalexamination. J Appl Bacteriol 197 1;34:477-83.

Borriello SP, Hudson MJ, Hill MJ. Investigations of the gastroin-testinal bacterial flora. Clin Gastroenterol 1978;7:329-49.

12 Miles AA, Misra SS. The estimation of the bactericidal power ofthe blood. J Hyg 1938;38:732-49.

13 Rogosa M, Mitchell JA, Wiseman RF. A selective medium forthe isolation of oral and fecal lactobacilli. J Bacteriol1951;62: 132-3.

4 Donnelly LS, Hartman PA. Gentamycin-based medium for theisolation of group D streptococci and application of themedium to water analysis. Applied and Environmental Micro-biology 1978;35:576-81.

'5 Koransky JR, Allen SD, Dowell VR. Use of ethanol for selectiveisolation of sporeforming microorganisms. Applied andEnvironmental Microbiology 1978;35:762-5.

16 Willis AT, Hobbs G. Some new media for the isolation andidentification of clostridia. Journal ofPathology and Bacteriol-ogy 1959;77:511-21.

7 Holdeman LV, Cato EP, Moore WEC. Anaerobe laboratorymanual. 4th ed. Blacksburg: Virginia Polytechnic Institute,1977.

18 Barrett E, Lynam G, Trustey S. Gas liquid chromatography fordetection of bacteriuria: examination for volatile acidic andneutral compounds. J Clin Pathol 1978;31:859-65.

9 Hill ID, Peto R. Probabilities derived from finite populations.Applied Statistics 1971;20:99-105.

20 Nelson DP, Mata 14. Bacterial flora associated with the humangastrointestinal mucosa. Gastroenterology 1970;58: 56-61.

21 Webster ADB. Immune deficiency disorders and the gastrointes-tinal tract. In: Triger DR, ed. Clinical immunology ofthe liverand gastrointestinal tract. Bristol, John Wright and Sons (inpress).

22 Kawamura T, Sakai K, Miyazawa F, Wada H, Ito Y, TanimuraA. Distribution of seconday amines in food. Journal of theFood Hygienic Society of Japan 1971; 12:192-7.

23 Kawamura T, Sakai K, Miyazawa F, Wada H, Ito Y, TanimuraA. Distribution of seconday amines in food (2). Journal oftheFood Hygienic Society ofJapan 1971; 12:394-8.

24 Ayukawa I, Sakai A, Tanimura A. Relation of alkalinity torecovery in the determination of secondary amines. Journal ofthe 'Food Hygienic Society ofJapan 1973; 14:100-4.

25 Reed PL, Smith PLR, Haines K, House FR, Walters CL. Gastricjuice N-nitrosamine in health and gastroduodenal disease.Lancet 1982;ii:550-2.

26 Autrup H, Harris CC, Trump BF. Metabolism of acyclic andcyclic N-nitrosamines by cultured human colon (40294). ProcSoc Exp Biol Med 1978; 159:111-5.

27 Mandel M, Ichinotsubo D, Mower H. Nitroso group exchange asa way of activation of nitrosamines by bacteria. Nature1977;267: 248-9.

28 Bode JC, Rust S, Bode C. The effect of cimetidine treatment onethanol formation in the human stomach. Scand J Gastroen-terol 1984; 19:853-6.

29 Tuyns AJ, Gricuite LL. Carcinogenic substances in alcoholicbeverages. Excerpta Medica International Congress Series1980;484: 130-5.

30 Rogers K, Roberts GM, Williams GT. Gastric-juice enzymes-an aid in the diagnosis of gastric cancer? Lancet 1980;i: 1124-6.

31 Kinlen U, Webster ADB, Bird AG, et al. Prospective study ofcancer in patients with hypogammaglobulinaemia. Lancet1985;i:263-6.

Requests for reprints to: Dr SP Borriello, Division ofCommunicable Diseases, Clinical Research Centre, Wat-ford Road, Harrow, Middlesex, HAl 3UJ, England.

953

on June 14, 2021 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.38.8.946 on 1 A

ugust 1985. Dow

nloaded from