J Clin Pathol 1992;45:959-963

Aeromonas spp as a potential cause of diarrhoeain children

M H Wilcox, A M Cook, A Eley, R C Spencer

AbstractAims: To determine the prevalence ofAeromonas spp in the faeces of childrenand the association with symptoms ofgastroenteritis.Methods: Faecal specimens (n = 1026)were cultured for Aeromonas spp usingthree selective media and an enrichmentbroth at both 30°C and 37°C. The isolationof Aeromonas spp was correlated withsymptoms of gastroenteritis, previousantibiotic use, and environmental tem-perature.Results: Aeromonas spp (n = 28) from 26(2-5%) patients were recovered. Bile salt,hrgasan, and brilliant green agar was themost efficient selective culture medium.Eleven of the patients had symptoms ofgastroenteritis, usually mild diarrhoea oftwo to three days' duration, in the absenceof other recognised enteropathogens. Acaviae was a particularly frequent isolate(nine out of 11 cases) in symptomaticindividuals. Only one out of seven Aero-monas spp recovered by enrichment cul-ture alone was possibly associated withsymptoms of gastroenteritis. There was aclose correlation between the environ-mental temperature and isolation ofAero-monas spp.Conclusions: Aeromonas spp and partic-ularly A caviae may cause gastroenteritisin children, most often during warmermonths of the year. Culture for thesepotential enteropathogens could be con-fined to summer and autumn months.Bile salt, Irgasan, and brilliant greenselective agar, but not alkaline peptonewater enrichment, is an efficient culturemedium for recovering possibly clinicallyimportant isolates.

( Clin Pathol 1992;45:959-963)

Department ofExperimental andClinical Microbiology,The University ofSheffield MedicalSchool, Beech HillRoad, SheffieldS10 2RXM H WilcoxA M CookA EleyR C SpencerCorrespondence to:M HWilcox

Accepted for publication6 May 1992

The potential role of mesophilic aeromonadsas a cause of gastroenteritis remains con-

troversial.' Geographically, there seems to be a

pronounced variation in the isolation rate ofAeromonas spp from human faeces. For exam-

ple reports from Australia,2 Djibouti,' India,4Peru' and the USA6 note isolation rates of11%, 3%, 2%, 58%, and 1% from diarrhoealstools, respectively. Aeromonas spp were detec-ted in the faeces of controls in the same studiesin 1%, 0%, 0%, 9%, and 0% of cases.

However, the use of varied isolation proce-dures and different study populations (chil-

dren, adults, community and hospital patients,for example) makes comparison of these ratesdifficult. Furthermore, as the detection ofAeromonas spp has been shown to be seasonallyassociated,2 56 studies should ideally reportisolation rates over 12 month periods, ormultiples thereof.Of the three commonly recognised pheno-

types, A hydrophila andA sobria have generallybeen considered to be the most virulent, withAcaviae rarely being implicated as a pathogen.7However, recent studies have indicated that Acaviae may indeed be an enteropathogen,particularly in children.5 "" One of these"noted an association between high stool pHand the isolation ofA caviae, possibly related tothe poor tolerance (suicide phenomenon) ofacidic conditions by this species.'2We report the results of a hospital-based

study lasting 12 months to determine theisolation rate ofAeromonas spp from the faecesof children in Sheffield. Clinical evidence ofenteropathogenicity for each of the majorspecies was determined by recording symp-toms and signs of gastroenteritis. Evidence tosupport the association between stool pH andA caviae isolation was also sought. To maximisethe isolation rate and to determine the mostefficient means of detecting Aeromonas spp infaeces, we used three different selective mediaat both 30°C and 37°C, and also an enrich-ment medium.

MethodsA total of 1026 faecal specimens from childrenwith gastrointestinal symptoms, submitted tothe Bacteriology Department at the SheffieldChildren's Hospital between November 1990and October 1991, were examined. This repre-sented 60% of all the faecal specimensreceived, as only those samples that could beprocessed within 24 hours of receipt wereincluded in the study. Eighty five per cent ofthe specimens were from inpatients with theremainder equally divided between attendersat the Accident and Emergency Departmentand outpatients.

Routine investigations for enteropathogenswere performed on all the study specimens.These included wet film microscopy, a modi-fied Ziehl-Neelsen stain for cryptosporidia,culture on MacConkey's, deoxycholate, bril-liant green and charcoal-cefaperazone agars,and selenite broth enrichment. Specimenswere also examined by electron microscopy ifvirological investigation was requested by theclinician; 40% of the study specimens were

959

on July 8, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.11.959 on 1 N

ovember 1992. D

ownloaded from

Wilcox, Cook, Eley, Spencer

Aeromonas

Others

Shigelsa

Calicivirus

Giardia

Campylobacter

Adenovirus

Salmonella

Rotavirus

10 20 30

Number of positive resultsFigure 1 Enteropathogens and aeromonads isolated from 1026 faecal specimens.*Aeromonas isolated together with one campylobacter and three salmonellae.

examined as such.For the isolation of Aeromonas spp a sterile

swab was used to inoculate faeces on to twowhole plates of each of the following media:ampicillin (30 g I`) blood agar, comprising 7%horse blood and Oxoid Columbia agar base(ABA) 13; Difco Aeromonas agar (bile salt,Irgasan, brilliant green agar) (DAA) 14; andxylose deoxycholate agar (XDCA) (Life Tech-nologies Ltd, Scotland)."5 One of each pair ofplates was incubated overnight at 30°C and theother at 37°C. A loopful of faeces was alsoinoculated into alkaline peptone water(pH8-6), which was incubated overnight at37°C and then subcultured on to half-plates ofABA, DAA, and XDCA.

After incubation, potential Aeromonas col-onies (translucent colonies on DAA and paleyellow colonies on XDCA) were screened foroxidase production using filter paper impreg-nated with tetramethyl-p-phenylenediamine.All oxidase positive colonies were subculturedfor purity, and a Hugh and Leifson (oxidation/fermentation) test was done. Fermentative,oxidase positive isolates were tested for resist-ance to the vibriostatic agent 0129 by discdiffusion using 10 ,ug and 150 pg discs(Oxoid). Those isolates resistant to' the higherlevel of 0129 were then phenospeciated bythree methods: conventional biochemistryaccording to the criteria of Janda et al 16; thesuicide phenomenon and aesculin produc-tion 12; API 20 NE strips and haemolysinproduction (Wilcox MH, Cook AM, Geary I,Spencer RC, abstract 729; 17th InternationalCongress of Chemotherapy 1991).The pH of each study specimen was meas-

ured by direct contact between Whatman pHpaper (BDH) and faeces.Maximum and minimum temperatures for

each day of the study period were obtainedfrom the local meteorological office. This issituated immediately opposite the SheffieldChildren's Hospital. The data were convertedto mean weekly temperatures by averaging 14daily figures.

Patients' details, including an antibiotic his-tory, were collected to determine the potentialsignificance of specimens positive for Aeromo-

nas spp. The former data were collated withoutknowledge of the culture results. Diarrhoeawas defined as multiple loose stools of morethan one day's duration. Gastroenteritis wasconsidered to be present if there was diarrhoeaplus or minus additional symptoms such asnausea, vomiting, and abdominal pain.The x2 test with Yates' correction for small

numbers was used to compare the numbers ofA caviae in symptomatic and asymptomaticpatients; the Wilcoxon rank sum test was usedto compare their ages.

ResultsA total of 89 accepted enteropathogens wererecovered from faecal specimens, equivalent toan isolation rate of 8-7% (fig 1). Twenty sixspecimens (2 5%) contained 28 Aeromonas sppisolates (two each containing two differentisolates). In four of these 26 cases Aeromonasspp were isolated together with known entero-pathogens (three salmonellae and one campy-lobacter). The 28 Aeromonas spp comprised 17

DAA

Primary

ABA

XDCA

Enrichment

3TC 300C'

5 1iA :4

Figure 2 Isolation of aeromonads according to primarymedia type (A), primary or enrichment culture (B), andtemperature ofprimary culture (C). Figures refer tonumber of isolates.

I -- I

960

on July 8, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.11.959 on 1 N

ovember 1992. D

ownloaded from

Aeromonas spp as a potential cause of diarrhoea in children

.4 + *++ t.++. o

IEach + represents new positiveI

0 4 8 12 16 20 24 28 32 36 40 44 48 52

Week number (November 1990 - October 1991)

Figure 3 Correlation between mean weekly atmospheric temperature over 12 months andisolation ofAeromonas spp from faeces.

(61%) A caviae, eight (29%) A hydrophila, two(7%)A sobria, and one isolate that could not beidentified. Apart from the latter isolate, thethree different methods of phenospeciationgave concordant results.

Figure 2 shows the relation between culturemedia and temperature conditions with theisolation of Aeromonas spp. Of the 21 isolatesrecovered on primary culture plates, 19 grewon DAA, 1 1 on ABA, and seven on XDCA. Allthe isolates grew on either DAA or ABA,although a quarter were recovered only afteralkaline peptone water enrichment. Notably,only two of the seven isolates recovered byenrichment alone came from patients withsymptoms of gastroenteritis (see below); in oneof these cases an accepted enteropathogen wasalso isolated. Incubation of primary cultures at37°C increased the yield of positive results byfive isolates compared to 30°C. The routinelaboratory methods failed to detect Aeromonasspp in any of the study faecal specimens.

Interestingly, of the 28 isolates, 17 were laterfound to be lactose positive on MacConkey'sagar; only three out of 17 A caviae isolates werelactose negative. The pHs of faeces from whichA caviae was isolated (range 6-0-8-5, median7-0) were similar to those containing otherAeromonas spp (6-0-8-5, median 7.0), and alsoto negative specimens (6-0-8-5, median 6-5).The isolation of Aeromonas spp was more

common during periods of warmer weather

Relation ofAeromonas spp isolated to clinical setting

No of cases positive Antibiotic use within Identity ofAeromonasClinical settng forAeromonas spp previous two weeks spp

Diarrhoea (no 11 3 9 A caviaeaccepted 2 A hydrophilaenteropathogen)

Diarrhoea (with 4 0 3 A caviaeaccepted 1 A sobriaenteropathogen)

No diarrhoea 11* 6 5 A caviae6 A hydrophila1 A sobria1 Unknown

* In two cases both A caviae and A hydrophila were isolated; one of the isolates could not bespeciated.

(fig 3). There were only two positive faecalspecimens when the mean weekly environmen-tal temperature was less than 7°C. After acluster of isolations in March-April 1991(weeks 20-24), two months (weeks 25-33)elapsed without any positive specimens. InMarch 1991 (weeks 17-21) the mean tem-perature was 8°C compared with a 30 year(1951-1980) historical average of 5 5°C. Con-versely, the month ofJune (weeks 30-34) saw amean temperature of 12-2°C which was 2 2°Clower than the historical average.The age range (and median age) of all study

patients and of those with Aeromonas spp intheir faeces was 1 week to 15 years (11months) and 1 month to 9 years (16 months),respectively. Symptomatic Aeromonas positivepatients were generally younger (1 month-3years; median 1 year) than asymptomaticindividuals (5 months-9 years; median 2years); 0 05 < p < 0 1. Of the 26 patients fromwhom Aeromonas spp were isolated, 1 1 had nodiarrhoea nor any other symptoms of gastero-enteritis (table 1). Six of them, however, wereeither receiving antibiotics at the time ofisolation or had done so within the previoustwo weeks; three patients received amoxycillin,two had penicillin and flucloxacillin, and onewas given cefuroxime and metronidazole. Fae-cal specimens were sent to the laboratory infive of these 11 cases because of the passage ofa single loose/offensive stool. Three of the 11Aeromonas positive patients who did havediarrhoea (but no accepted enteropathogen)had a recent history of antibiotic treatment-one with amoxycillin, one with penicillin, andone with erythromycin. The duration of diar-rhoea ranged from two to 21 days, with amedian of three days. Two patients had asso-ciated abdominal pain and three had vomiting,of which one (a 15 month old baby with anileostomy) became clinically dehydrated.Although the numbers are small, the isolationofA caviae was particularly common in symp-tomatic (nine out of 11) compared withasymptomatic patients (five out of 11);p > 0-1.

DiscussionAccumulating evidence suggests that in at leasta proportion of cases Aeromonas spp can causegastroenteritis. l 2 5 There are several weak-nesses in this hypothesis: a lack of a suitableanimal model; inconsistent results in humanvolunteer challenge studies"7; and an absenceof reported outbreaks of infection. As themajor virulence determinants ofAeromonas spphave still to be clarified, it is difficult toperform adequate animal and human chal-lenge studies. Strains used in such work maylose their pathogenicity during subculture andstorage. A recent report documented for thefirst time that an identical strain ofA hydrophilawas isolated from a patient with acute gastro-enteritis and from recently consumedprawns. 18

This study is the second largest UnitedKingdom report of the prevalence of Aero-monas spp in the faeces of paediatric inpatients.

4.+ + + + +20 -

18*16-14 -12-10 -8-6-4.2-0-.

U

0

-

20cLE0

c0 o I-%

-2 i I I I I I -I I I 9 t

961

on July 8, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.11.959 on 1 N

ovember 1992. D

ownloaded from

Wilcox, Cook, Eley, Spencer

An earlier report noted that 4-4% of specimenscontained Aeromonas spp, but did not correlateisolation with clinical details. The isolation rateof Aeromonas spp from routine faecal speci-mens in this study was 2 5%, similar to that forthe accepted enteropathogens salmonellae(2-2%) and rotavirus (2 5%). We consideredthe possibility of collecting specimens from acontrol population of asymptomatic children,but following earlier unsuccessful results in apilot study this proved impracticable. Our useof three culture media, an enrichment broth,and two culture temperatures to isolate Aero-monas spp makes it unlikely that we had falsenegative specimens. Several different commer-cially available selective media and ampicillinblood agar have been suggested for isolatingAeromonas spp. '9 In our hands DAA per-formed very well, and in only two out of 21cases did Aeromonas spp fail to grow on it onprimary culture; these two isolates were recov-ered on ABA. DAA contains brilliant green,bile salts, and the disinfectant Irgasan as

selective agents. The usefulness of brilliantgreen has been noted earlier,2' and Irgasan hasalso been incorporated successfully intoanother selective medium (CIN; cefsulodin,irgasan, novobiocin).20 DCA inhibits com-mensal flora by virtue of bile salts, but itperformed poorly in the present study and inthis formulation cannot be recommended.We confirmed the findings of others that the

use of alkaline peptone water enrichmentincreases the yield of faecal specimens positivefor Aeromonas Spp.2' 22 Studies that do notincorporate such enrichment techniques arelikely to underestimate significantly the preva-lence ofAeromonas spp in faeces. However, theclinical importance of isolates recovered byenrichment alone must be considered to beextremely doubtful. Robinson et al found thatthe use of enrichment culture frequently detec-ted Aeromonas spp in asymptomatic individ-uals.23 Similarly, we observed that only one outof seven isolates detected by alkaline peptonewater enrichment alone came from a patientwith gastroenteritis, in the absence of a recog-nised pathogen.We were prompted to compare isolation at

30°C and 37°C, because preliminary studieswith a collection of 30 Aeromonas spp hadshown two strains that failed to grow on at leastone of the media at 37°C. However, no isolateswere recovered from faeces on media incu-bated at 30°C compared with those incubatedat 370C.

Clinical assessment of Aeromonas spp posi-tive individuals not otherwise infected with aknown enteropathogen suggested that half ofthe cases had symptoms attributable to thesemicro-organisms. Undoubtedly, Aeromonasspp are carried asymptomatically by someindividuals.5 24 However, this is also true forwell recognised enteropathogens such as sal-monellae. In this study children with gastro-enteritis and Aeromonas spp in their faecesgenerally had non-profuse diarrhoea of shortduration (median three days). Vomiting andabdominal pain were present in only a fewcases. The affected children were 3 years old or

younger, while four (36%) of those carryingAeromonas spp asymptomatically were at least5 years old. An association between antibioticuse and isolation of Aeromonas spp notedelsewhere was most apparent in asymptomaticchildren in the present study.25 A caviaepredominated among the clinical isolates, par-ticularly in those from symptomatic individuals(82%). These results support the recent find-ings that A caviae seems to be a relativelycommon isolate from the faeces of youngchildren with gastroenteritis.5 l" We wereunable to confirm an association betweenalkaline faecal pH and isolation ofA caviae. "

However, the patients positive for A caviaereported by Namdari and Bottone" weremostly less than 1 year old. In our series thethree infants of less than 1 year old withgastroenteritis possibly caused byA caviae hadstool pHs of 8O0 (twice) and 7 5; all were bottlefed. The alkaline faecal pH found in bottle fedbabies may indeed help to select for A caviaeand explain the relatively high isolation rate ofthis species in infants.Aeromonas spp are widely distributed in fresh

water, sewage, marine environments close tofreshwater outlets, and drinking water.26-28Maximal levels of these bacteria in fresh wateroccur during summer and early autumn. Aero-monas spp have also been suggested as poten-tial foodborne pathogens (see above), which issupported by their high prevalence in super-market meat samples (58-74%) and raw milk(27%).29 Our data confirm the close relationbetween environmental temperature and isola-tion ofAeromonas spp from faeces.5 625 Labor-atories could focus their efforts to isolate thesebacteria during the warmer months to limit theadditional costs of the extra media required.We would encourage attempts to isolate

Aeromonas spp from well defined populationsto determine more closely their prevalence andpotential to cause gastroenteritis. Selective(DAA) but not enrichment media are recom-mended. Uncertainties concerning the tax-onomy ofAeromonas spp and the virulence ofgiven species will not be resolved without suchstudies.

We thank Difco Laboratories Ltd and Life Technologies Ltd forkindly supplying their respective selective media for use in thisstudy.

I Wadstrom T, Ljungh A. Aeromonas and Plesiomonas asfood- and waterborne pathogens. Int J Food Microbiol1991;12:303-12.

2 BurkeV, Gracey M, Robinson J, Peck D, Beaman J, BundellC. The microbiology of childhood gastroenteritis: Aero-monas species and other infective agents. J Infect Dis1983;148:68-74.

3 Mikhail IA, Fox E, Haberberger RL, Ahmed MH, AbbatteEA. Epidemiology of bacterial pathogens associated withinfectious diarrhoea in Djibouti. J Clin Microbiol 1990;28:956-61.

4 Deodhar LP, Saraswathi K,Varudkar A. Aeromonas spp andtheir association with human diarrheal disease. J ClinMicrobiol 1991;29:853-6.

5 Pazzaglia G, Sack RB, Salazar E, et al. High frequency ofcoinfecting enteropathogens in Aeromonas-associateddiarrhea of hospitalised Peruvian infants. 7 Clin Microbiol1991;29:1151-6.

6 AggerW A, McCormick J D, Gurwith M J. Clinical andmicrobiological features of Aeromonas hydrophila-asso-ciated diarrhea. J Clin Microbiol 1985;21:909-13.

7 Millership SE, Barer MR, Tabaqchali S. Toxin productionby Aeromonas spp from different sources.JMed Microbiol1986;22:311-14.

8 Watson I M, Robinson JO, BurkeV, Gracey M. InvasivenessofAeromonas spp in relation to biotype, virulence factors,

962

on July 8, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.11.959 on 1 N

ovember 1992. D

ownloaded from

Aeromonas spp as a potential cause of diarrhoea in children

and clinical features. J Clin Microbiol 1985;22:48-51.9 Janda JM, Clark RB, Brenden R. Virulence of Aeromonas

species as assessed through mouse lethality studies. CurrMicrobiol 1985;12:163-8.

10 Namdari H, Bottone EJ. Cytotoxin and enterotoxin produc-tion as factors delineating enteropathogenicity of Aero-monas caviae. J Clin Microbiol 1990;28: 1796-8.

11 Namdari H, Bottone EJ. Microbiologic and clinical evi-dence supporting the role of Aeromonas caviae as apediatric pathogen. J Clin Microbiol 1990;28:837-40.

12 Namdari H, Bottone EJ. Suicide phenomenon in meso-philic aeromonads as a basis for species identification. JClin Microbiol 1989;27:788-9.

13 George WL, Nakata MM, Thompson J, White ML. Aero-monas-related diarrhoea in adults. Arch Intern Med1985;145:2207-1 1.

14 Hunt GH, Price EH, Patel U, Messenger L, Stow P, SalterP. Isolation of Aeromonas spp from faecal specimens. JClin Pathol 1987;40:1382-91.

15 Shread P, DonovanTJ, Lee JV. A survey of the incidence ofAeromonas in human faeces. Soctety for MicrobiologyQuarterly 1981;8: 184.

16 Janda MJ, Reitano M, Bottone EJ. Biotyping ofAeromonasisolates as a correlate to delineating a species-associateddisease spectrum. J Clin Mcrobiol 1984;19:44-7.

17 Morgan DR, Johnson PC, DuPont HL, Satterwhite TK,Wood LV. Lack of correlation between known virulenceproperties of Aeromonas hydrophila and enteropatho-genicity for humans. Infect Immun 1985;50:62-5.

18 Altwegg M, Lucchini G M, Luthy-Hottenstein J, RohrbachM. Aeromonas-associated gastroenteritis after consump-tion of contaminated shrimp. Eur J Clin Microbiol InfectDis 1991;10:44-5.

19 Robinson J, Burke V, Worthy PJ, Beaman J, Wagener L.Media for isolation ofAeromonas spp. from faeces. J MedMicrobiol 1984;18:405-1 1.

20 Altorfer R, Altwegg M, Zollinger-Iten J, von Graevenitz A.Growth of Aeromonas spp. on celsulodin-irgasan-novo-biocin agar selective for Yersinia enterocolitica. J ClinMicrobiol 1985;22:478-80.

21 Millership SE, Chattopadhyay B. Methods for the isolationof Aeromonas hydrophila and Plesiomonas shigelloidesfrom faeces. J Hyg (Camb) 1984;92:145-52.

22 von Gravenitz A, Bucher C. Evaluation of differentialselective media for isolation ofAeromonas and Plesiomo-nas spp from human faeces. J Clin Microbiol 1983;17:16-21.

23 Robinson J, Beaman J, Wagener L, Burke V. Comparison ofdirect plating with the use of enrichment culture forisolation of Aeromonas spp. from faeces. J7 Med Microbiol1986;22:315-7.

24 Pazzaglia G, Escalante JR, Sack PB, Rocca C, Benavides V.Transient intestinal colonisation by multiple phenotypesof Aeromonas species during the first week of life. J ClinMicrobiol 1990;28: 1842-6.

25 Moyer NP. Clinical significance of Aeromonas speciesisolated from patients with diarrhoea. Jf Clin Microbiol1987;25:2044-8.

26 Burke V, Robinson J, Gracey M, Peterson D, Partridge K.Isolation of Aeromonas hydrophila from a metropolitanwater supply: correlation with clinical isolates. ApplEnviron Microbiol 1984;48:361-6.

27 Rippey SR, Cabelli VJ. Occurrence of Aeromonas hydro-phila in limnetic environments: relationship of the organ-ism to trophic state. Microbiol Ecol 1980;6:45-54.

28 Monfort P, Baleux B. Distribution and survival of motileAeromonas spp. in brackish water receiving sewagetreatment effluent. Appi Environ Microbiol 1991 ;57:2459-67.

29 Ibrahim A, Mac Rae IC. Incidence of Aeromonas andListeria spp. in red meat and milk samples in Brisbane,Australia. IntJ Food Microbiol 199 1;12:263-70.

963

on July 8, 2020 by guest. Protected by copyright.

http://jcp.bmj.com

/J C

lin Pathol: first published as 10.1136/jcp.45.11.959 on 1 N

ovember 1992. D

ownloaded from