frontiers in gastrointestinal research reprint ad … · frontiers in gastrointestinal research...

Frontiers in Gastrointestinal Research ReprintEditor: P Rozen, Tel Aviv Publishers- S. Karger, Basel

Printed in Switzerland

AD-A237 725Dinart G, Rozen P. Bujanover , Lebenthal 1E (eds): JL0819Newser Tests and Procedures in Pediatric Gastroen'croogy.Front Gastrointest Res. Basel, Karger, 1989, vol 16, pp 169-219 C

Laboratory Investigation of Ac.*.siQAI Jr

Childhood Enteric Infections [i-iw i lVfl0 TAll C

Edgar C. Boedeker, Charles E. McQueen t .nan notnood CDepartment of Gastroenterology, Walter Reed Army Institute of Research, J itiOatio .

Washington, D.C.. USA

DIstributioon/.Avallabillty Cod&,

Introduction v"ai rnw i

This chapter will review current diagnostic methods for a ivariety offenteropathogens including bacteria, viruses and parasites. These methiodsmay include: (1) direct observation in stool; (2) culture and isolation fromstool or biopsy specimens, often through the use of selective or enrichmentmedia; (3) antibody.enhanced detection of organisms or their products instool and identification after culture; (4) identification of specific materialin stool or culture; (5) determination of host serum or secretory immuneresponses to the organisms or their products.

Major advances in diagnostic methods have come in the areas of:(1) improved methods for selective culture or isolation of "newly recog-nized" pathogens; (2) greater commercial availability of specific high titerantibodies particularly specific monoclonal antibodies, for novel, rapid dia-gnostic tests; (3) application of the techniques of molecular biology, inparticular genetic analysis and the development of specific genetic probesfor the identification of organisms or their virulence determinants.

In this chapter we will discuss enteropathogens of major importance inthe categories of bacteria, viruses and protozoal parasites. For eachpathogen we will briefly review clinical presentation, epidemiolgy and ;current taxonomy; we will then review current procedures for sample

5 The views expressed herein are those of the authors ana not necessarily those of theUnited States Army or the Department of Defense

91-04123

Boedeker/McQueen 170

collection, transport, isolation and identification of organisms withemphasis on new techniques. Where applicable we will also discussserodiagnosis. Finally, we will summarize our current recommendations forapproach to diagnosis.

The expanding application of nolecular biology techniques to theidentificulion of enteric pathogens deserves special attention at the begin-ning of this chapter on newer methods for diagnosis. The techniques ofplasmid profile analysis, restriction endonuclease analysis of plasmid andchromssomal DNA, and nucleic acid hybridization with specific DNAprobes are finding increasing application in epidemiologic studies, and, to alesser extent, in clinical microbiology [1, 2]. To date, these have beenapplied primarily to bacterial pathogens, and to a lesser extent to viruses.

Plasinid Analysis [2]

Since many of the virulence properties of bacterial enteropathogens are encoded onthese extrachromosomal DNA elements, the analysis of plasmid patterns after clectrophoresisin agarose gels and staining with ethiddium bromide may provide a recognizable "finger print"of an organism. Plasmid preparations sufficient for analysis can be extractec from smallvolumes of overnight broth cultures. Bacteria are generally lysed with combined treatmentvith EDTA, lysozyme and detergent at an alkaline pit which irreersibly denatureschromosomal DNA. After neutralization, only plasmid DNA is present in a form that remainsin the supernatant after brief centrifugation and is precipitated by ethanol in the cold,Electrophoretic analysis separates and displays plasmids by their molecular weight, Thisanalysis may be made more specific by first digcsting the plasmides with restrictionendonucleases. These enzymes cut the plasmid DNA at specific nucleotide sequences.yielding smaller fragments with very characteristic profiles on electrophoresis.

Bacterial Restriction Endonuclease DNA Analysis (BRENDA)

Endonuclease digestion can also be applied to the analysis of total chromoso,,Ial DNA,since not all entene pathogens have characteristic plasmids. DNA is again obtained ftora lysedcells and exiracted with phenol and cholorform to eliminate contaminants, Mfore cor'r,,. butequally characteristic, patterns are obtained on electrophoresis of digested chromosomalDNA than are seen with plasmid DNA [lj.

Construction and Utilization of Genetic Probes

Genetic probes are labelled, well-characterized, single-stranded DNA fragments Nohoenucleotide sequences are unique to a particular pathogen, These probes can be used to

Laboratory Investigation of Childhood Enterie Infections 171

recognize homologous, complementary, DNA sequenccs in cultures or clinical stool speci.mens through DNA/DNA hybridization.

To obtain a probe for avirulence determinant encoded on a plasmid. the genetic materialof interest must first be cloned [2). To do this, the native plasmid is isolated, digested with arestriction endonuclease, and the resulting fragments rehgated into a plasmid cloning vectorcontaining an antibiotic resistance marker (e.g. pB R322).The resultingrecombinant plasmidsare then used to transform a recipient strain (such as -113101). Transformed colonies areselected by antibiotic resistance, then screened for expression of the desired virulencedeterminant, Recombinant plasmids from colonies so identified are isolated and digestedwith the same restriction enzymes to identify the cloned fragments of interest. Fragments areseparated on agarose gels, and electroeluncd, before labelling with x-p by nick translation withT4 polynuclotide kinase. Alternaticly, if the nucleotide (or protein) sequence isknown, oligonuclotlide probes can be synthesized to conform to a known sequence andlabelled. Oligonucleotide probes may be more specific since they will contain no vectorDNA.

The.Uplabelled DNA is used to probe the total DNA in clinical specimens, or in isolatedcolonies, by hybridization by a modification of the method of Grunstem and Itogness [3).Stools samples, or colonies, are spotted onto nitrocellulose filters overlying nutrient agar andallowed to grow overnight, The filters are then treated with NaOIi to lyso the bacteria anddenature DNAThe DNA is fixed bybaking overnight at 65 'Cbefore probing, Preparation ofthe samples to this point can be performed in the field. Hybridization is performed inhybridization solution containing formamide, citrate, Flcoll PVP and BSA Labelledprobe DNA (l101cpm) is added in the presence of excess sheared carrier DNA foroxernight incubation. Detergent washed filters are autoradiographed by exposure to X-rayfilm with an intensification screen for 24 hat-70 1C, so that positive colonies are identified asdense spots,

Since it takes three days to complete these procedures, hybridization techniques cannotbe considered rapid diagnostic methods. Their utility lies in their ability to handle largenumbers of samples reproducibly. As a result they are usually used to simultaneouslyinmestigate a large number of samples from epidemiological studies rather than being appliedtodiagnosis of individual cases of diarrhea. Use of stable isotopes or nonradioactise detectionsystems should increase the utility of these probe systems since

5P.labelled probes hase a finite

half.life. Another limitation of gene probes is that they detect virulence determinants whichare encoded but which may not be expressed, thus they may yield false-positise results forpathogenicity,

Salmonella

Salmonellae cause both diarrheal disease and systemic infection, butthe most common manifestation of Salmonella infection, of whatevergroup, species, or serotype is gastroenetritis (4, 5]). Salmonella areclassified into three major groups (1, 11 and Ill) according to tile spectrum ofclinical disease they usually produce [5]. At present in the United States,three species are recognized which correspond to groups I to 11I. Group I


organisms (species S. enteritidis) generally cause diarrheal disease(although they may cause severe systemic disease in infants, the elderly,and the immunocompromised). Group I includes all but 2 of the more than2,000 serotypes which had been recognized by 1988, and which wereclassically named after the host, or location in which they were isolated(e.g. typhimurium, heidelberg, enteritidis, newport, infantis). Group IIorganisms (species S. cholerasuis) cause septicemias and are associatedwith localized suppurative infections in adults, but they often causegastroenteritis in children. Group III (species S. typhi) organisms causetyphoid (enteric) fever in humans, but can also have an initial presentationas a diarrheal illness, especially in children (6]. Clinical features cannot beused to differentiate Salmonella diarrhea from that produced by otherenteric pathogens. Isolation of the organism from stool or blood cultures isthe only way to confirm the diagnosis. Most patients have mild crampyabdominal pain with watery to paste-like stools, but a dysentery-like illnesswith small-volume bloody stools may occur (7]. Salmonellae pathogenic forhumans colonize the intestinal tracts of many animal species includingswine, poultry and cattle. Outbreaks of gastroenteritis due to Salmonellaoften develop after the consumption of contaminated food, in particulareggs, dry milk, and chicken. A DNA probe, such as that which detects theVi capsular antigen gene [8, 9], may eventually shorten the time icquiredfor diagnosis.

ProceduresSalmonella are present in the stool in appreciable numbers only during

the acute stage (first 3 days) of diarrheal disease, so specimens for cultureshould be obtained during this period if at all possible. Because manycommensals proliferate more rapidly than Salmonellac, specimens shouldalso be inoculated into transport media such as (Carey-Blair, 0.033,1lphosphate buffer mixed with equal parts of glycerol, Ames, or Stuart's) andthen onto appropriate culture media as soon as possible. Salmonella aredifferentiated in culture from other Enterobacteriaceae by their failure toferment lactose, ability to resist the inhibitory effect of citrate, and byproducing hydrogen sulfide in triple sugar iron agar (TSI) [4]. Miniaturizedmultiple-test systems that incorporate multiple biochemical reactions arenow in widespread use and have simplified and shortened the process ofidentification.

Once isolated, Salmonella can be scrotyped using the slide agglutina-tion method [101. This is of great importance in epidemiological studies,

Laboratoiy Investigation of Childhood Enterielnfections 173

such as to identify a common source for an outbreak, but is less useful forthe clinician. The large number of serotypes are grouped by the 0, orsomatic antigen, and typed by their H, or flagellar antigen. Becauseapproximately 65 different somatic antigens have been recognized, sub-grouping with polyvalent 0 antiserum is a necessity, and commerciallyprepared kits are available (Lederle Laboratories, Pearl River, N.Y.:Difco Laboratories, Detroit, Mich.; Baltimore Biological Laboratory,Cockeysville, Md.;, Lee Laboratories, Grayson, Ga.). From a practicalstandpoint, the cultures to be tested should be smooth and must notautoagglutinate in saline [10]. Also if no agglutination occurs with the 0diagnostic sera. the culture should be heated to 100 °C for 15-30min,cooled, and retested, since some Salmonellae possess K antigens and arenonagglutinable in the live or encapsulated form in 0 antisera. 0 nonagglu-tinable cultures must be sent to reference laboratories for completeserologic analysis.

Isolation of organisms in pure culture and performance of biochemicaland serological assays are cumbersome, time.consuming, and expensiveprocesses. Techniques of genetic analysis from the field of molecularbiology have begun to be applied to salmonella diagnosis.

An example of the use of plasmid profile analysis in the epidemiology of Salmonella isprovided by Taylor et al, (Ill. Marijuana was suspc:ted to be the most likely common sourceof infection for an outbreak of S. inuenchen, but this was difficult to prose, since isolatesstudied were phenotypically identical to other S. ,muenchen strains. The issue was clarified,Lowever, by the demonstration of both a common plasmid profile and a common restrictiondigest pattern for the plasmids recovered from patients, and drug isolates. The reliability ofrestriction enzyme analysis as an epidemiologic tool was demonstrated, since the restrictionpatterns remained virtually unaltered despite enteric passages in many naturally infectedindividuals.

A recently constructed DNA probe was validated as highly effectise at distinguishingS. ryphi from other gram-negative bacteria and also other strains of Salmonella in subcul.tures of diarrheal specimens collected in Peru and Indonesia [8). The vial) genetic locus ofS. typhlts involved in the synthesis of the capsular virulence (Vi) antigen expressed by virulentS. typhi strains, The DNA probe used to detect this locus is an 8.6-kb EcoRt fragment fromplasmid pWR14l, which was constructed by cloning the Vi antigen locus of a Ctrobacterfreundii strain [9). At present, this technique requires isolation of colonies from freshspecimens, and overnight culture on filters, to obtain sufficient concentrations of the bacterialDNA.

RecommendationsDiagnosis of salmunellosis requires isolation of the organism. Positive

blood, bone marrow, or urine cultures are confirmatory. Positive stool

Boedekcr/MfcQueen 174

cultures may indicate enteric infection, but they may also indicate a carrier

stat, presumably related to intermittent secretion of the organisms fromthe biliary tree. Culture is not cost-effective for patients with mildwatery diarrhea and who are afebrile with only mild systemic symptoms,since the illness often resolves before culture results are known.Culture may be indicated, however, to identify outbreaks of Sal-monellosis caused by common sources, particularly in nurseries or innursing homes. If dysentery, fever, or severe symptoms are present, cultureis necessary to guide therapy. Serology is not useful clinically. Techniquesof plasmid analysis re currently useful for epidemiological studies. DNAprobes currently a%,ilable are reliable for identifying subcultured organ-isms, and may be applied to direct detecti n of S. typhi directly in stool orblood specimens if methods to concentrate bacterial DNA can be de-veloped.

ShligellaShigella are highly virulent, locally invasive organisms which are the

major cause of dysentery [4, 12, 13]. Oral ingestion of as few as 100organisms can result in the full clinical syndrome [141. Stools are character-istically bloody and small in volume, with greater than 50 WBC/HPF onfecal smears. Nonintestinal manifestations may be severe, and includefever with convulsions, leukemoid reactions, hemolytic-uremic syndrome,and reactive arthritis. Watery diarrhea may also occur. The host range ofShigella species is limited to humans and primates [14]). Spread is fromperson.to-person or through fecal contamination of water supplies. Theclinical importance of Shigellae stems from their unique virulence andalso from their emerging antibiotic resistance. In the Asian subcontinent,drug-resistant S. dysenteriae I infections are currently achievingepidemic proportions. Four species (or groups) are identified on thebasis of biochemical and serological reactions: S. dysenteriae (group A),S. flexneri (group B), S. boydii (group Q), and S. sonnei (group D.)Diagnosis continues to depend on the identification ot the organismsin stool by culture. New developments in diagnosis include prepara-tion of a DNA probe based on virulence determinants clonedfrom the 140-Mdalton virulence plasmid which confers invasiveability upon Shigellae and enteroinvasive Escherichia coli [15-17].In addition, a high titer diagnostic antiserum, raised againstsurface antigens associated with virulence, has been utilized in an ELISA18].

Laboratory Investigation of Childhood Enteric Infections 175

ProceduresShigella are slender gram-negative nonmotile rod-shaped members of

the Enterobacteriaceae family. The standard laboratory diagnosis ofShigella infection has changed little over recent years. Stool specimens arecultured on differential and selective media and a screening battery ofbiochemical tests are performed on isolates. Stool specimens for Shigellaculture need to be obtained early in the course of the illness, and, ifpossible, should include bits of bloody mucus or epithelium which maycontain invading organisms. Immediate culturing of material obtainedduring proctoscopic examination increases the likelihood of obtainingpositive Shigella cultures [10, 19]; the organism quickly dies in acid stoolspecimens. By genetic studies, Shigella are extremely closely related to E.coli [13], although they are nonlactose fermenters. They are oxidasenegative, facultatively anaerobic, and non-spore-forming. In distinction toSalmonella, Shigella do not produce hydrogen sulfide, but have a fermenta-tion pattern similar to Salmonella on TSI agar. The overall scheme forserological identification of Shigella is similar to that for Salmonella.Because of the number of serotypes (group A, 10; group B, 6; group C, 15;group D, 1) subgrouping is done with polyvalent antisera, and commercialkits for slide agglutination tests are available (Lederle Laboratories; DifcoLaboratories; Baltimore Biological Laboratory; Lee Laboratories).

The shared virulence attributes of Shigella and enteroinvasive E. coilinclude their ability to invade epithelial cells, to multiply intracellulary andto spread intercellularly to adjacent cells. The genetic determinants of theseproperties have been the subject of intense research interest and several invivo and in vitro assays have been developed for their study. Theproduction of keratoconjuctivitis in the eye of a guinea pig (Sereny test) [20]has been a standard confirmatory test for invasive ability, but is impracticalfor testing large numbers of strains. Other tests include the demonstrationof HeLa cell invasiveness in tissue culture which relies on the ability ofinvading organisms to survive externally applied antibiotics. This test hasbeen refined into a plaque assay which reflects the intercellular spread ofthe organisms [211. These tests are not used for routine diagnosis, but haveprimarily been utilized to define the genetic and phenotypic determinantsof Shigella virulence which include the products of both chromosomal andplasmid loci.

An additional research method ii detect invasion-associated virulence factors but whosefuture importance in clinical diagnosis remains to be determined is an ELISA assay (18].


Invasive strains of Shigella or E. coll express a "virulence marker antigen" (VMA) on theirsurface, reflecting Sereny test positivity. The critical reagent for the ELISA is an absorbedantisera that detects the VMA This was raised in rabbits by immunization with an invasive E.coli strain, then absorbed with an organism of the identical serotype, biotype, antibioticsusceptibhlity, and plasmild pattern, but which had been rendered Sereny negative by serialpassage on nutrient agar. Mlicrotiter wells are coated with test bacteria, then incubated withthe anti-VNA antisera followed by an antirabbit IgG conjugate. Fifty-six Sereny test positivestrains gave high OD values (range 0.4-1.6) and 27 Sereny test negative strains gave a low OD(range 0-0.15). With one exception, no VMA-positive strain harbored the 140.Mdaltonplasmid [18]. This method requires that pure cultures of isolated organisms be used forsensitizing wells, and it has not been applicable to testing stool samples directly, making its usein diagnosis or for screening large numbers of isolates impractical.

Techniques of plasmid analysis have led to the discovery that large plasmids in the rangeof 140 Mdaltons (120 Mdaltons in the case of S, sonne! )confer most of the organism's invasiveability (15] and code for outer membrane proteins required in the recognition and penetrationprocess 116, 17, 22]. High degrees of homology have been demonstrated for the invasiveplasmids of the several Shigella species tested, perhaps reflecting their common origin [23].The relative ease of plasmid detection and the firm evidence that large plasmids are requiredfor invasion has made plasmid analysis an attractive alternative to other test for virulence.Detection of the invasion plasmid has been facilitated by application of DNA hybridizationtechniques. A 17-1,b rada labelled DNA probe, prepared from the 140-Mdalton virulenceplasmid of S. flexnei 5 has been shown to permit sensitive and specific screening of culturedstrains for the invasive genotype, thereby permitting analysis of many more specimens thanpossible byconventional assay (24,25]. Sensitivity in direct analysis of stool specimens directlyis less than optimal, however. The clinical need to more rapidly identify pathogens in cultureshas given impetus to the development of improved hybridization probes. Because radiolabel.led probes decay rapidly and have a limited useful life, a blotinylated 17.kb DNA probe hasbeen developed and shown to compare favorably with the autoradiographic assay in terms ofsensitivity and specificity [26]. In one study, both radioactive and blotinylated probeshybridized with stool blots from II of 13 Thai children with culture.proven shigellosisor CIECdiarrhea, All 43 children with cultures that were negative for Shigella or EIEC were alsonegative by both of the hybridization studies 126].

RecommendationsDiagnosis ofshigellosis is supported by the appropriate clinical presen-

tation and the demonstration of sheets of white cells on fecal smears, butultimately depends on identification of the organism in culture. The successof culture is enhanced by obtaining fresh specimens by proctoscopy early inthe course of the illness, and by the use of proper transport media if platingis delayed. Serotyping is useful in epidemiologic studies but not inindividual cases. Testing for the invasive virulence plasmid by DNAhybridization is in development. Use of biotinylated probes may simplifydiagnosis, and may greatly shorten the time required, if sufficient sensitivity

can be achieved to study stool directly without the need for prior isolation of


organisms. Low numbers of organisms in stool specimens, and their unevendistribution, may make clinical shigellosis less that ideal for the immediateapplication of these techniques,

Escherichia coliAt least four categories of E. coli strains cause diarrhea [27, 28]. These

can be classified according to their district pathogenic mechanisms, andclinical presentations, into the enteropathogenic E. coil (EPEC), theenterotoxigenic E. colt (ETEC), the enteroinvasive E. coli (EIEC), theenterohemorrhagic E. coli (EHEC) and perhaps a fifth, recently recog-nized category, the enteroadherent E. coil (EAEC) [28]. Since all of theseE. colt share growth and culture characteristics with the commensal E. coltstrains which colonize the distal intestine, identification of pathogenic E.colti strains from fecal specimens represents a particular diagnosticchallenge. Diagnosis depends on the recognition of particular serotypesassociated with disease, or the identification of their individual virulenceattributes. This is a difficult process, involving isolation and testing of arepresentative sample of 5-10 colonies from stool specimens. Diagnosticmethods for pathogenic E. colt in stool are usually reserved for populationand epidemiologic surveys, rather than for diagnosis of individual cases ofdiarrheal disease. E. coli serotype screening is performed less widely thanin the past, and tests for E. coli virulence attributes are not widely availablein routine laboratories. Molecular techniques have recently been applied tothe diagnosis of E. coii infections through the application of DNA probes.These may be applied to stool samples without the need forcolony isolation[1, 29-31].

Enteropathogenic E. coliThe enteropathogenic E. coli [32, 33) were the first E. coli to be

associated with diarrheal disease when strains of particular serotypes wererepeatedly associated with outbreaks of diarrhea in neonatal nurseries.EBEC diarrhea is now recognized more frequently in infants older than 30days than in neonates, and presents as severe, chronic diarrheal diseasewith weight loss. Watery stools usually contain no fecal leukocytes.

The value of serotyping to diagnose EPEC infection has beenrepeatedly confirmed as strains of characteristic serotypes have continuedto be associated with diarrheal disease over several decades. Serotypingincludes determination of heat-stable 0 (somatic lipopolysaccharide) andheat-labile H (flagellar) antigen. Major 0 serogroups (class 1) have been

Boedeer/McQueen 178

defined as 055, 086, 0111, 0119, 0125, 0126, 0127, 0128ab and 0142 128].Less frequently associated minor serogoups (class II) include 018,044,0112and 0114 (28]. Within these serogroups, disease is associated only withspecific serotypes (e.g. 055:H6, 0111:H5). Init:al screening of coloniesobtained from stool, small bowel aspirate or biopsy is performed usingcommercial pools of antisera (Difco) for several serotypes. Colonies whichscreen positive should be confirmed by testing with individual specific 0and H antisera. EPEC strains were initially reported to express unique heatlabile capsular (K) antigens (designated B type), antibodies to which wereabsorbed out with heated cultures. It now appears that these strains do notproduce true heat-labile acidic polysaccharide capsules; but instead pro-duce a surface oligosaccharide related to the 0 (LPS) determinant (33].Thus, older literature referred to B (or K) capsular determinants on thesestrains whereas current serotyping usually only designates 0 and Hdeterminants. Recently, EPEC attempts have been made to associate themajor and minor EPEC serogroups with particular virulence determinants[28].

EPEC Adherence Assays. EPEC strains are noninvasive but closelyadhere to intestinal epithelium producing a characteristic lesion describedas attaching, effacing enteroadherence [34). These lesions of the individualepithelial cell can be identified by electron microscopy of small bowel orrectal biopsies. In vitro assays examining characteristic patterns of adher-ence of EPEC strains to cell lines in tissue culture (HEp2, HeLa) or toorgan cultures of intestinal biopsies, have been developed 135, 36]. Inparticular, a pattern of localized adherence (LA) of EPEC microcolonies toIlEp2 and HeLa cells is now recognized as a property of the major EPECserotypes. Diffuse adherence (DA) over the entire surface of cultured cellsis a less-specific property of minor serotypes. Characteristic attachingeffacing EPEC lesions will develop within hours in intestinal organ culturesinfected with EPEC. These assays have at times been used forepidemiologic characterization of E. coli isolates, but their primary utility isas research tools to localize and clone the genetic loci for EPEC virulencedeterminants.

Plam mds and Probes. The determinants for llp 2 localized adherence of EPEC areencoded by a 60. to7O.kdaltonplasmid (termed pMAR2) in an 0127:116 EPEC strain 1371. ThepMAR2 plasmid (and homologous plasmids of smlar size) are also necessary for full EPECvirulence in volunteer challenge studies, They encode for a presumptoe adhesm termedEPEC adherence factor (EAF), A 23.kb Sal I fragment of pMAR2 has been cloned and used


as a probe for identification of EPEC [381. This probe has been successful in identifying EPECof the major serogroups listed above which show LA. Adherence genes in this fragment werecloned as two distinct plasmid regions of l-5Mb which confer the adherence phenotype onlywhen complementing each other [39). Analysis of data from several labs, however, suggests

that these plasmid encoded determinant(s) are neither necessary nor sufficient to produce thefull characteristic attaching and effacing lesion in vitro. They may promote an early stage ofattachment to the intact mucosal surface. The localization of the genes for EPEC effacingadherence and other virulence properties thus remains an area of intense research effort,EPEC strains do not produce LT or ST. Some strains previously considered to be EPEC.particularly those of scrotype 026:1111, produce moderate to large amounts of shiga-like toxin(vero-toxin), but such strains may more properly belong to the EHEC category describedbelow (28].

Enterotosigenic E. coli

The enterotoxigenic E. coli are responsible for the majority of

cases of traveller's diarrhea experienced within 1-2 weeks by 30-50% of

visitors from industrialized countries to developing countries. These

organisms are also a major cause of severe, and fatal, diarrheal illness in

infants and children in developing nations. They induce a watery diarrhea

without fecal leukocytes, which may be cholera-like in intensity. Trans-

mission is by the fecal-oral route and contaminated food is the usual

transmission vehicle. Although many animals (including pigs and cattle)

experience enterotoxigenic E. coli infection, the strains infecting humans

and animals are distinct in their expression of species-specific adherencedeterminants.

The ETEC strains fall into a limited number of serogroups which

correlate with the expression of particular combinations of colonization

factor antigens and toxins. Antisera against common ETEC 0 serogroupshave been suggested as a means of simplifying ETEC identification in theclinical laboratory. In one study, two pools of antisera against 12 totalserotypes had sensitivity of 64%, specificity of 96% and a predictiveaccuracy of 89% as compared to tissue culture enterotoxin assays 140].Orskov and Orskov [411 have proposed adding the F (or fimbrial) antigenexpression, as defined by crossed-line imnmunoelectorphoresis of salineextracts of fimbriated organisms, to the E. coli serotyping scheme. How-ever, this has been difficult to establish in practice since expression offimbrial antigens depends on growth conditions which are not standard forfinibrial types.

ETEC isolates from various sources produce both heat labile (LT) and

heat stable (ST) toxins. These include LT-I, LT-II, ST-I and ST-It.

Boedeker/,fcQueen 180

LT.L The E. colt heat-labile toxin (LT-I) is closely related to cholera toxin with which itshares 75% nucleotide and 70-0% amino acid homology. Like clolera toxin, it consists ofone 28kdalton A (active) subunit and five ll.5-kdalton B (or binding) subunits. The Asubunit activates adenylate cyclase through the ADP ribosylation of guanine nucleotidebinding proteins. The B subunit binds to G,11 ganghoside on the surface of intestinal epithelialcells. Assays for LT.I include those which measure activation of adenylate cyclase in tissueculture cells; measure bindingof f-I to G51i ganghoside; or depend on recognition of toxin byspecifie antibody. The genes for LT-I are located on plasmids, freqeuntly the same plasmidswhich encode for ETEC adherence factors. Recently, nucleotide probes have been developedfor LT-I genes,

The YI mouse adrenal (42,43) or Chinese hamster ovary (CHO) tests (44) utilize changesin tissue culture cell morphology (rounding) or steroidogenesis to detect E. coli LT activity inculture or stool filtrates, These tests were tially developed for cholera toxin activity but areequally applicable to LT. They provide the standard against which subsequent tests aremeasured for sensitivity and specificity. These tests have great advantages over rabbit iealloops, but they require a tissue culture facility and are not useful for field studies.

The GillELISA, as reported by Svennerholm and tlolmgren [45), is based on thespecific binding of LT.I to polystyrene-absorbed Gt1 ganglioside. Immunological detection ofLT uses commercially available cross-reactie anti-cholera toxin antisera in an EILISA. Thistest and the GII HRP-ELISA (46) require analysis of supernates of individual coloniescultured in broth, In the field, the color development can be satisfactorily interpreted with thenaked eye. An improved assay (47) optimizing culture conditions for toxin production(casamino acids) and toxin release from the penplasmic space (polymyxin B), achievedsensitivities similar to the YI assay. Direct culture of E. coil strains in the Git-coated wells ofmicrotliter plates (so.called DCM.G,\, ELISA) can be performed.

In the Elck/Biken immunodiffusion test (48) a precipitin line forms tn agar between asubeultured E, coil colony and anti-cholera toxin or anti-LT serum placed in an adjacent well,Colonies are grown for 48h, polymyxin B added to lyse cells, and then the immunodiffusionreaction developed for 24-48h. The test does not require isolation of culture filtrates andrequires no special equipment for reading; however, it takes 5 days and precipitin lines may besubtl. and hard to interpret,

Recently. Vadiveln et al. (49) described a membrane filter assay in which primaryisolation of bacteria is performed on a cellulose-acetate membrane filter, after which the filteris transferred to an agar medium containing anti-cholera toxin. Zones of precipitation in theagar are read at 8h. This assay permits examination of large numbers of strains without theneed for sibeulture, pooling, preparation of supernates or radiolabelling.

A modified staphylocccal co-agglutination test 150] (Phadebact ETECILT) has recentlybeen marketed by Pharmacia to detect LT extracted from isolated colonies of E. cols. This isbased on formaln-treated and heat.killed Staphylococcus arireas which have been sensitizedwith high-titer rabbit ati.LT serum, In the two-step test toxin is extracted from colonies whichmay be picked from initial isolation plates, then the extract is agglutinated with the sensitizedstaphylococcal organisms.

LT.I Plasmids and Probes. The genes for LT-I A and B subunits have been cloned (511on a 1.2-Mdalton fragment and subelones for A and B+A regions have been prepared for useas probes (31. 52). These have been used to test isolated colonies as well as stool blots frompatients with diarrhea. In general, the probes are highly sensitive for isolated colonies.


although they may give false positives (identifying strains which do not make entertoxin)because of contamination with ,ector DNA. They are less sensitive for stool blots, and inparticular for blots prepared in the field as opposed to in a central laboratory (presumablybecause of inadequate quality control in the fixation steps). A synthetic oligonucleolide probeof 20 bases 129] has also been prepared and evaluated for detecting LT genes in cultures andstool samples. The probe was sensitive for colonies but insensitive for testing sto. I samplesdirectly.

LT.H. This heat-labile toxin alsoactivates adenylate cyclase in YI adrenal cells via ADP-ribosylation and has A and B subunits similar in size to LT-I; howsever. this toxin isimmunologically distinct from LT-I and is chromosomally, rather than plasmid. encoded (531.The genes encoding this toxin have been cloned and a 0.8kb DNA probe prepared which doesnot hybridize with structural genes for LT-1 (54]. In a recent epideniological sursey of YIadrenal LT-positive E. colistrains from Southeast Asia using this probe, LT-II positive strains"ere commonly isolated from cattle, but rarely isolated from children with diarrhea (55].

Toxins designated ST-I (or STA) belong to a group, or family, of heat-stable, methanol-soluble peptides with molecular weights from 2-5 daltons that activate intestinal mucosal cellparticulate guanylate cyclase. The actise moiety is within a common Cterminal 18 amino acidsegment (56] which contains six cysteine residues and seems to require cross-linking foractivity.

The standard assay for the ST-I (STA) family is the in viNo suckling mouse in whichculture supernates are injected transabdominally into the stomaches of suckling mice [57].Intestines are later observed for fluid secretion which is quantitated as the ratio of the weight ofthe distended intestine to that of the carcass. One suckling mouse unit of toxin gises anintestmelcarass ratio of 0.08.The assay is cumbersome but reproducible. It is not sensitive fordetecting ST directly in stools.

Giannella ct at. 158] and Frantz and Robertson 159] developed radioimmunoassays forST-I (STA) utilizing high.titer antisera raised by conjugating the poorly immunogenic toxin asa hapten to carrier protein, and using highly purified radiolabelled toxin. Because of the needfor relatisely large amounts of purified and highly labelled toxin, these assays have not beengenerally available. The assays recognized secral types of ST-I (STA) but did not recognizeST-11 (STU).

The degree of relatedness of the members of the ST-! family have been shown by genecloning experiments of Moseley et al. 130, 60]. They first (60] prepared a radiolablledfragment of DNA (cstAl) encoding for an ST.I toxin designated ST-la from a pig isolate(hence also called ST-P). About hall (15/43) of the suckling mouse positive strains isolatedfrom humans hybridized with this probe under stringent conditions. Many more were positiseunder less-stringent conditions, suggesting related but not identical DNA sequences amongthe ST-I (suckling mouse positive) toxins produced by different E. coli strains. An additionalprobe (estA2) was prepared for an ST-I deignated ST-lb from a human isolate (also called ST-11) which hybridized with the remainder of the test strams (M], Some strains were detected byboth probes. The genes estA I and estA2 have 72% homology and their enterotoxins have 62%homology with highly consersed C-terminal regions. They do not hybridize with each otherunder stringent conditions. Thus, two probes are necessary to recognize the majority of ST-Iproducing isolates of human origin.

Since the abose probes are restriction fragments of plasmid sectors containing clonedgenes and they may theoretically give false-positie reactions due to contamination with sector

BoedekerAIcQueen 182

DNA, Murray et al. [29) evaluated small oligonucleotide probes of 22 bases synthesizediccording to the known sequences of ST-H and ST-P. They found good correlation betweencolony blots with the oigonucleotide probes and bioassy for ST-I; however, they found tests ofstool growth instead of individual colonies difficult to interpret.

ST-I genes have been identified as being in a transposable element, accounting for theirappearance on several different plasmids. In strains causing human diarrhea, STa and CFA/Iproduction were encoded by the same plasmid. whereas a separate plasmid in the strainsencoded LT1-1.

ST-l (ST). This methanol-insoluble heat-stable toxin is negative in the sucking mouseassay but induces secretion in the pig or rabbit intestinal loop. This toxin is produced by strainspathogenic for pigs and cows, but has not been implicated in strains causing human disease.Less is known about its genetic organisation although it may also be on a transposable element,

Colonization Factor Antigens. The recognized colonization factor antigens (CFAs) andputative colonization factors (PCF) of the ETEC strains include both fimbrinal (pilus)components (CPA/I, CSI and CS2 of CFA/tI, CS4 and CS5 of pcf8775) and finer fibrillarcomponents (CS3of CFA/II and CS6 of pcf8775). Expression of the several CFAs is associatedwith a limited number of srogroups (CFAIIW4, 01,025,063,078,090,0110,0126.0128.01531CFA/l 06, 08, 09, 078, 080, 0115.0139; pcf8775 025, 0115, 0167). The expression of thesesurface structures can be recognized because they confer the property of hemagglutination ofhuman and bovin erythroeytes. This hemagglutination of ETEC CFAs is not inhibited bymannose, thus distinguishing these fimbrial components from type I of common pili. Bothfimbriac and fibrillae are large macromotecular aggregates of smaller repeating proteinsubunits. Adhesive properties may reside in the major subunits, of in minor subunits. Theexpression of these adhesms depends on growth conditions and they are generally bestexpressed on the surface of agar, Orskov and Orskov 1411 have proposed a serotypiog schemefor fimbrial expression based on analysts of saline extracts of bacteria by crossed-lineelectrophoresis, but this has been difficult to implement because of the different growthconditions required for expression of different adhesins,

McConnell et al. 1611 hase developed an ELISA system for the expression of CFAs.They developed antisera specific for the several fimbrial and fibrillar components byimmunizing rabbits with strains expressing individual adhesms, and absorbing with homolog-ous adhesin-negative strains. Oserall, the ELISA was more sensitive than hemagglutmation ofimmunodiffusion. Identification of presently defined adhesins for ETEC will detect onlyabout 40% of clincial isolates shown to be toxin producers by other tests.

The genes for s-.veral CFA components have been cloned including CFA/I [621 and CS363, Iowever, these clones ha , not as yet been used to prepare nucleotide probes for clmncal

or epdemeologial studes.

Enteroisvasive E. coliThe enteroinvasive E. coli are genetically similar to Shigellae and

produce a similar clinical picture, dysentery or a watery diarrhea, but theydiffer from Shigellae in requiring a much higher inoculum to producedisease. They are nonmotile and slow or non-lactose fermenting. They arelysine decarboxylase negative and this has been suggested as a screening


test for selecting diarrheal isolates for further testing [64]. The characteris-tic 0 serogroups of EJEC are 028, 029, 0112, 0124, 0136, 0143, 0144, 0147,0152, 0164 and 0167. These antisera are not readily available and they mayidentify nonpathogenic E. coli of the same serogroups. Thus, it is desirableto confirm the invasive phenotype by other tests.

The tests for EIEC invasiveness are those previously discussed forShigellae. The Sereny test (20] (guinea pig keratoconjunctivitis) is defini-tive [64]. The indirect ELISA of Pal et al. [18] was developed using a hightiter antiserum to a virulent EIEC (0143) absorbed with the Sereny.negative derivative. It can detect invasive clinical E. coli and shigellaisolates, but it has not been useful to detect these organisms in stool samplesdirectly.

IIEC and Shigcllae possess large, 120- to 140-Mdalton plasmids which are assocatedwith invasion (16], Thus, one approach to identify ElIECis to screen isolates for large plasmrdsand then test them for invaseiveness by Sereny test or UlcLa cell assay. This is an impracticalway to handle large numbers of samples. A 17.1,b EcoRt fragment of thr 140.Mdalton plasmidof S. floxnerd S. used as a probe, will differentiate tIEC from other R. coil in culture 123).Recently, a blotinylated probe was developed which compares favorably with a previouslytested radiolabelled probe 1241.

Enterohemorrhagic E. coliThe EHEC strains of E. coli which are associated with outbreaks and

sporadic cases of hemorrhagic colitis are among the E. coli which producecytotoxins distinct from LT and ST. Community outbreaks have beenassociated with the ingestion of contaminated beef in fast food restaurants,and illness has also occurred in nursing homes and day care centers.Children and the elderly appear to be at greatest risk. These strains are alsotermed verocytotoxin (or shiga-like toxin) producing E. coli (VTEC). Theywere initially recognized as a novel serotype and then by the ability of theirculture supernates to induce toxic effects on Vero cells or HeLa cells intissue culture [651, effects which could be neutralized, or partially neu-tralized, by antibody to shiga toxin 1661. It is now known that theseorganisms produce two distinct shiga-like toxins (SLTs), called by someinvestigators verotoxins (VTs). SLT-I (VTI) is essentially identical to shigatoxin differing in only one amino acid in the A subunit. SLT-11 (VT2) is only60% homologous to shiga toxin based on nucleotide sequence. The genesfor these toxins are carried on bateriophages [67], from which they havebeen cloned. Strains producing these toxins are also associated withnonbloody diarrhea, and with the development of hemolytic uremic


syndrome (HUS) felt to be related to effects of absorbed toxin on vascularendothelium [68]. Although fewer than 10% of patients with EHECinfection develop HUS, infection with 0157:H7 strains is the predominantcause of HUS in the US.

Serotyping. Enterohemorrhagic E. colt of serotype 0157:117, the mostcommon serotype, do not ferment sorbitol, thus screening for organisms ofthis serotype can be facilitated by selecting sorbitol-negative colonies forconfirmatory serotyping and testing for SLT (VT) production after over-night incubation on sorbitol-MacConkey agar [69]. Additional, sorbitol.positive colonies should be selected for screening for SLT (VT) productionby EHEC of other serotypes (predominantly 026:K60:Hl and 0103:H2,but also 0111:K58:H- and 0145:H-) in a recent epidemiologic study ofsporadic cases in Canada [70].

Toxin production is determined 1651 by observing effects of culture supernates on Verocells or HeLa cells in culture and neutralizing with appropriate antisera. Such tests are usuallyavailable in reference laboratories or those with special interest. ELISA and latex agglutina.tion tests for demonstration of toxin production are being developed. The definition ofglobotriosylceramtde (Gb3) as the cell surface receptro for SLT- should permit the develop.ment of receptor-binding assays analogous to the Gss.bindivg assays for LT.I but these arecurrently less sensitive than cell culture assays,

These organisms can be demonstrated to be enteroadherent in an attaching and effacingmanner as seen with EPEC strains, Their adherence factors may include a new fimbrialantigen encoded by a 60-Mdalton plasmid 1711. A 3,4-kb DNA probe has been prepared fromthis plasmid which hybridized with 99% of 0157:117 strains and 77% of 026:H1l V-positaestrains 172),

Canipylobacter

Once considered rare agents of infections in humans, Campylobactersare now among the common causes of bacterial diarrhea in man, and areisolated as often as Shigella and Salmonella in both developed anddeveloping countries [73, 741. Diarrhea occurs 1-7 days following oralingestion of as few as 500 organisms and typically lasts less than 7 days,although persistent and relapsing symptoms can occur [75]. Diarrhea variesfrom watery stools to dysentery. Complications include Reiter's syndrome,reactive arthritis, and Guillain-Barrd syndrome. Sources of infectioninclude unpasteurized milk, improperly cooked poultry, and contaminatedwater. Campylobacters are commonly found as commensals in poultry,swine, sheep, cattle, dogs, and cats, and it is possible that the high optimum

Laboratoty Investigation of Childhood Enteric Infections 185

growth temperature of C. jejuni reflects an adaptation to birds, which havehigh body temperatures. The taxonomy has changed recently: Campy-lobacterfetus spp. jejuni has been classified into two separate species, C.jejuni and C. col. Although several species have been isolated fromanimals, C. jejuni and C. coli commonly cause diarrheal illness in humans,C. fetus spp. fetus is a relatively rare cause of human disease (often manifestonly as septicemia) which chiefly attacks debilitated patients. Two othergroups of spiral organisms are implicated in human disease and are includedin the Campylobacter genus but will not be discussed here: C. pylori hasbeen found colonizing the gastric mucosa of patients with gastritis. Theymay be more closely related to Wolinella than to other campylobacters [76].The second group of Campylobacter-like organisms (CLOS) include C.cinaedi and C. fennelliae [77] and have been isolated from rectal swabs orbiopsy specimens from symptomatic and asymptomatic homosexual men.

ProceduresCampylobacters are small (1.5-3.5 pmn in length, 0.2-0.4 jm in width),

nonspore-forming gram.negative organisms that have a characteristiccurved, S-shaped, or spiral morphology (hence their name - "curved rod"in Greek) that permits their distinction from other gram-negative roads onmicroscopic examination of fecal smears. The cells possess a singleflagellum at one or both poles, and they move with a characteristic rapidcorkscrew-like darting motion when observed with phase-contrast micro-scopy.

Transport media, including Campy-thio, Carey-Blair (thioglycolate),Ames or Stuart's modified thioglycolate, are useful and permit survival for1 week or longer especially if refrigerated [78]. Buffered glycerol saline isunsuitable [79]. Campy-Thio can be inoculated with 3-5 drops of liquidstool, or a rectal swab can be inserted about 1 cm into the medium andtwirled to allow inoculation of the organisms into a zone of reduced oxygen.After overnight refrigeration in Campy-Thio, the inoculum for subcultureshould be withdrawn by inserting a Pasteur pipette 2cn below the surfaceand continually aspirating during withdrawal. A few drops of this aspirateare then placed on a selective media and streaked for isolation.

C. jejuni in pure cultures grow best at 42 *C in microaerobic environ-ments (5-10% oxygen and 3-10% carbon dioxide) on a variety of basalenteric media including Mueller-Hinton and brucella agars. However,isolation from stool specimens requires direct inoculation of the specimensonto a selective antibiotic media supplemented with various nutrients (such

Boedeker/NicQueen 186

as Campy-BAP), or indirect inoculation into an enrichment broth (such asCampy-Thio) (thioglycollate broth with 0,16% agar and the antimicrobialsused in Campy-BAP) for overnight refrigeration with subsequent subcul-ture to a selective medium, since campylobacters are rapidly overgrown bythe other organisms in fecal specimens.

Incubation cultures at 42 *C allows for more rapid growth of C. jejutnilC. coli, and appears to inhibit growth of some fecal flora, but has thedisadvantage of inhibiting growth of C. fets. However, C. fetus is rarelyisolated in stools, and is basically an opportunistic organism associated withdisease in immune-compromised hosts 179].

Several different selective media that contain three or more antibioticsto retard the growth of other organisms are in common use. The three mostcommonly used media are campy-BAP, Skirrow's, and Butzler's mediumOxoid [79,80], and are available from BBL, Remel. Gibco, and Scott. Twoother described media are: Preston medium, reported to be as sensitive butmore selective and less expesive than Skirrow's, and Butzler's MediumVirion which is reported to be comparable to Butzler's Medium Oxoid in

sensitivity, but more effective at suppressing competing enteric flora andpermiting isolation at 37 *C. Two distinct colony types have been described:a flat, grayish, spreading nonhemolytic colony with an irregular shape, andround, convex colonies with an entire edge. The moisture content of themedium may determine which colonial morphology occurs [811.

An early selective method for isolating Campylobacter that is stilluseful, particularly in countries where the antibiotics forselcetive media arenot readily available, is based on filtration and depended on the small sizeof campylobacters for separation from other bacteria. Steele and McDer-mott 182] modified Dekeyser's filtration method [831 by directly applying a0.45-tum (Millipore) filter onto nonselective media. 8-10 drops of a heavysuspension of stool in saline is applied to the filter for 30min at roomtemperature. The filter is removed and the media incubated for 24-48 h at42 *C under microaerobic conditions. Other enteric organisms are removedwith the filter, but because of their small size, campylobacters pass through.Theoretical advantages of this method are that inhibition of Campylobactergrowth by antibiotics does not occur, and the large inoculum may permit anenrichment effect.

The classicl approach of serotyping gram-negative organisms basedon heat-stable 0 antigens and heat-labile H and K antigens has been lesssuccessful for Campylobacter than for other enterics because of autoag-glutination and non-specificity after heating. Complex serotyping schemes


based on either heat-labile or heat-stable antigens are not fully stand-ardized or available commercially. The Lauwers and Penner [84] systemswhich are both passive hemagglutination assays, attempt to avoid theproblems of nonspecificity by extracting heat-stable (presumably 0) anti-gens and sensitizing red blood cells with them. At least 50 differentserotypes have been identified in these systems. The Lior [85] system,which is a slide agglutination assay based on heat-labile antigens (probablyflagellar or capsular antigens), reduce the number of nonspecific reactionsby absorbing antisera with homologous heat-stable antigens. Thirty-sixserogroups have been identified by this system.

Several tests have been developed to detect the development of serumantibody to Campylobacter, including a serum bactericidal assay, aggluti-nation tests, indirect fluorescent antibody test, enzyme immunoassays, anda complement fixation test (86-88]. A limitation of most of these tests is thatthey react with strain-specific antigens, and therefore require that abacterial isolate from at least I patient involved in an outbreak be used as anantigen. The antigen for the complement fixation test, however, is pre-pared from a random mixture of serotypes and reacts with group-specificantigens, obviating the need for a homologous or related strain (Virion CFTest, Morristown, N.J.).

Plasmids have been detected in less than 20% of Campylobacters [89]. diminishing theirpotential importance for rapid identification or epidemiology. Tomplins et at, (90] havereported on the use of a variety of chromosomal DNA probes. including fragments of C. jejtnichromosomal DNA. specific endonuclease restriction fragments of C. jejuni chromosomalDNA cloned in pBR322, and single stranded synthetic oligonucteotide probes. A C. jejunichromosomal fragment radiolabeled by nick translation was shown to hybridize with C. jejuni,but not other Campylobacter species or other enteric organisms. This probe was used to detecteampylobacters in 299 diarrheal stools dotted directly to nitrocellulose filters (without growingthe colonies) and successfully identified 10 of 18 stools from which C. jejuni was isolated byculture. Although specific, sensitivity was low, requiring at least 1,007 organisms permilligram of stool.

RecommendationsFresh fecal specimens should be directly plated onto selective media or

inoculated into broths such as Campy-thio for transport and enrichment. Afecal smear showing typical morphology is helpful. A presumptive identifi-cation may be based on growth on selective media incubated at 42 "C undermicroaerobic conditions, typical colonial and gram stain morphology,motility, and positive oxidase and catalase tests. The filtration technique

may permit isolation if selective media are not available. The DNA


hybridization assay is insensitive, not commercially available, and presentsthe inherent problems involved in handling and maintaining radioactivereagents.

Yersinia

The importance of Y enterocolitica as an enteric pathogen wasemphasized in 1976, when an epidemic of gastroenteritis involved 222subjects and resulted in over a dozen children undergoing appendectomies[911. Yersinia is now consistently isolated in 8-10% of diarrheal stoolspecimens in some northern geographic areas [92]. Several clinical forms ofYersinia enteritis have been recognized, including acute and chronic(lasting longer than I week) diarrhea, acute abdominal pain resemblingappendicitis, localized infections, and septicemia [93]. The majority ofpatients with proven disease have been children. Children younger than 6years old tend to have diarrheal illness while older children may haveabdominal pain as the dominant feature. The diarrheal illnes may beginwith 1 or 2 days of watery stools that progresses to blood-streaked stoolswith tenesmus, right lower quandrant tenderness, fever, myalgias, andheadache. About 50% of children and 10% of adults also experiencepharyngitis. Postinfections complications that have been reported includethyroiditis, glomerulopathy, polyarthritis, Reiter's syndrome, erythemanodosum, erythema multiforme, and carditis. Children with chronicdiarrhea due to Yersinia are often suspected of having idiopathicinflammatory bowel dicease due to similarities in presentation. There arethree Yersinia species that are pathologic for man. Y enterocolitica and Ypseudotuberclosis are primarily intestinal pathogens, and Y pestis, theagent of bubonic plague, has a greater tendency to cause fulminantbacteremia and suppurative lesions of the lymphatic system. Two acceptedmethods for typing strains of Y enterocolitica are serotyping (based onbacterial agglutination with rabbit anto-0 antisera), and biotyping (definedby Wauters and based upon the use of indole, xylose, and trehalose and thepresence of lipase and DNAase) [93]. The predominance of humanserotypes varies with location. Serotypes 0:5, 0:8, and 0:9 are the usualisolates in the United States, but 0:3, 0:5, and 0:8 are most common inCanada and 0:3 and 0:9 in Sweden. Of the five biotypes, most human strainsare biotype four [93]. Y enterocolitica and V pseudotuberculosis have aworldwide distribution, but Y pseudotnbercnlosis is rarely isolated in theUnited States.


ProceduresStandard laboratory diagnosis involves culture of the organism,

followed by biochemical and serologic tests if necessary. Presumptiveidentification by culture on usual enteric media such as Hektoen enteric,MacConkey, eosin-methylene blue, and blood agar depend on an aware-ness of the pinpoint morphology of the colonies and temperature-depen-dent growth characteristics. They produce an acidlacid reaction on triplesugar iron slants, ferment glucose but not lactose, and procedure ornithinedecarboxylase. Human strains are usually sucrose positive, rhamnosenegative, and melibiose negative. The colonies are nonpigmented andorganisms are motile at 22 *C. Several selective and differential agar mediahave been developed to enhance recovery of Y. enterocolitica from stoolspecimens [94]. On Yersinia selective (Schiemann CIN) media, colonieshave a "bull's eye" appearance with a deep cherry-red center and transpa-rent margins and have a distinctive odor. Cold enrichment (4 *C) of stoolspecimens in phosphate-buffered saline with weekly subculture enhancesrecovery from stool specimens, and this is useful when organisms are shedin small numbers such as during the covalescence or in detecting carriers[931.

Serology has been useful as a diagnostic adjunct only in well circum-scribed outbreaks of Yersinia gastroenteritis [951. The use of serology inindividual cases is limited by a paucity of data about antibody prevalence, aswell as by the unavailability of Yersinia antigens for performing theseassays.

A rapid method to distinguish pathogenic Y. enterocolitica isolatesfrom nonpathogenic Yersinia strains would be useful in epidemiologicalsurveys, and one proposed virulence marker is a plasnid-independent24,000-dalton surface protein (96]. A coagglutination assay to detect thisprotein using staphylococci coated with specific antiserum has recentlybeen reported to correlate 100% with tested pathogenic biogroups [96].

The molecular genetics of Yersinia species hase not been widely exploited for diagnosticpurposes, although restriction maps hase been made for the virulence plasmids from seseralY. enterocolitica serotypes (0:3, 0'8, and 0.9), and the plasmids are reported to be closelyrelated, sharing functional simtlarities and high degrees of DNA homology, making thedevelopment of DNA probes conceivable 197).

RecommenelationsDiagnosis depends on isolation of the organisms in culture, and is

enhanced by the use of Yersinia-selective media. If pharyngitis is present,

Boedeer/McQueen 190

throat cultures should be obtained. Patients with diarrhea usually excretesufficient numbers of organisms so that cold enrichment is of marginalbenefit except to identify carriers. Recognition of typical colonial morpho-logy on enteric media is facilitated by the use of a stereomicroscope.Serotyping and biotyping can be performed by reference laboratories ifnecessary. Identification of virulence factors such as the 24,000-daltonsurface protein by coagglutination of antibody-coated staphylococci, orDNA probe hybridization are as yet restricted to the research laboratory.

Clostridiun difficile

C. difficile was not recognized as an enteric pathogen prior to the late1970s. C. difficde is now recognized as the most important cause ofantibiotic-associated colitis, and is one of the most commonly diagnosedenteric pathogens in hospital practice [981. The spectrum of presentationranges from a mild, watery diarrhea to hematochezia with high fever,severe abdominal pain, localized tenderness, and polymorphonuclearleukocytosis. Illness from C. difficile infection rarely occurs without a priorhistory of antibiotic administration (especially clindamycin, ampicillin orcephalosporins, but rarely vancomycin or parenteral aminoglycosides) andthe incidence of infection has been shown to correlate with the presence ofspores in the environment [991. Recent advances in laboratory diagnosishave been limited to the development of tests for toxins in stool samples.

ProceduresClinically, the diagnosis of C. difficile infection can be made with

remarkable certainty if typical yellow-white pseudomembranous plaquesare seen overlying the mucosa at proctosocopy, flexible sigmoidoscopy, orcolonoscopy 1981, since almost all (99%) cases of pseudonembranouscolitis thus recognized are caused by C. difficile. A failure to demonstratepseudomembranes on sigmoidoscopy does not rule out the diagnosis, sincethe distribution of these lesions may be patchy with rectal sparing [100].

Gram-stained fecal smears are not helpful for diagnosis because of theuneven distribution of organisms in specimens and difficulties in judging apredominance of grain-positive rods [101]. Wilson et al. 11021 detected C.difficile in smears by direct immunofluoresence in 26 (81%) of 32 stoolspecimens that were also positive by culture or toxin B assay, but wereunable to distinguish between pathogens and cross-reacting clostridialspecies.


Laboratory diagnosis depends on successful culture of C. difficile andor the detection of C. diffrcile toxins in stool filtrates. Isolation of C. difficilein pure culture was formerly difficult (hence its name), since it has fewcolonial morphological characteristics that distinguish it from other morenumerous anerobes on artificial media. The development of a commer-cially available selective anaerobic medium (CCFA) by George et al. [103]containing cycloserine, and cefoxitin has greatly improved the efficiency ofdetecting the organisms from patients who have more than 1,004 organismsper gram of stool. Rapid plating of specimens is required, since storage,freezing, or exposure of samples to air before plating diminishes recovery.On selective media the organisms grow in yellow, circular colonies with afilamentous edge, which emit a golden yellow fluorescence underultraviolet light. The addition of sodium taurocholate to selective agarimproves the detection of the spore forms, which is important in monitoringsoil or hospital environments [104, 105].

A number of difficulties hinder the interpretation of stool cultures,which are usually not qupntitative and are reported as positive or negative.The most important problem is that C. difficile can be detected in stoolsfrom healthy, asymptomatic carriers 1106]. 1-5% of adults are carriers, andeven higher isolation rates (10-60%) have been found in apparently healthyinfants in newborn nurseries [106]. These findings are partially explained bythe fact that not all C. difficile isolates are toxin producers and thereforepathogenic [107]. This is not the whole explanation, however, since hightoxin producers have been isolated from infants who are apparently well,suggesting that some infants may be resistant to the effects of toxin.Because of these difficulties, many laboratories will only perform cultureson diarrheal specimens if a clinical history of antibiotic usage is provided.Finally, a 48-hour delay is required for even a presumptive diagnosis byCCFA culture.

C. difficife produces at least two toxins: A, called an ent.'rotoxin (because of its affect onanimal loops), and B, called a cytotoxin (because of its effect on tissue cultures), Theimportance and exact role that each plays in the etiology of C. diffictle diarrhea has not beenfully defined, in part due to difficulties in separating the toxins i pure form. The detection oftoxin B in stools continues to be a standard laboratory test for C. diffcile Although mostadults with C. diffictle cytotoxin B in their stools have colitis and fewer than 2% withpseudomembranous colitis are negative 11051 for toxin B, its importance has come underquestion. Although toxin B induces rounding of tissue culture fibroblasts, probably bydepolymerization of actin-contanung thin microfilaments, it has no effect in intestinal loops,Toxin A. on the other hand, does cause secretion of fluid in ligated rabbit liea loops, ,here itelicits severe intestinal inflammation and alteration of epithelial permeability [10g), Toxin A


may prove to be more important than toxin B [1081, but assays for its detection are not easilyperformed.

Detection of toxin B is easier than anaerobic culture, and, when properly performed, isquite sensitive and specific (98). The stool specimen is homogenized, centrifuged, andsterilized by filtration. The filtrate is then added to a monolaer of tissue culture cells (usuallyfibroblasts) and the cells observed with an inverted phase microscope at 24 and 48 h. Picogramamounts of vytotoxin induce rounding of the cell bodies and thinning of the cytoplasmicprocesses that extend from the cell and attach it to the surface of the culture dish. Specificity isdemonstrated by neutralization of cytoxicity by pre.incubation the filtrate with C. difficileantitoxin. A commercial kit for detecting cytotoxin based on the above method ("Toxi-Titer"Bartels Immunodiagnostic Supplies, Inc., Bellevue, Wash.) has become available and allowslaboratories not routinely performing tissue culture assays to test for C. dtfficle toxin B [1091.The kit includes a toxin control, antisera against C. difficile, a diluent for preparing the stoolfiltrate, and a 48.well microtiter plate seeded with a 75-100% confluent growth of viablehuman foreskin fibroblast cells, ready for the cytotoxicity assaywithout additional processing.This kLit was recently compared to conventional tissue culture assays by two laboratories andgenerally found to be satisfactory 1109, 110].

Detection of toxin A has lagged behind assays for toxin B due to difficulties inpurification and isolation. A late\ agglutination kit was recently marketed for rapid detectionof toxin A (Culturette Brand Rapid Latex Test Mitsubishi Chemical Industries. Tokyo. Japan;distributed by Marion laboratories. Kansas City) [111). Hoever, further efforts withpurification and separation procedures indicated that the test reacted with an unidentifiedantigen and is not a test for enterotoxin A production, at least for some strains of C. difficile[1121, Paradoxically, the test has some clinical utility in that it was comparable in sensitivity tothe cytotoxin assay in patients meeting clinical criteria for C. diffict'e.related antibiotic.associated diarrhea and colitis 1113).

Several other laboratory tests for C. difficiet are reported but of undetermined clinicalusefullness, Initial reports of counterimmunoelectrophoresis (CIE) using C, dffieile culturefiltrates and detecting antigens with C. sordellil antitoxin suggested that this relatisely rapid,simple, and inexpensive test was more sensitive with fecal specimens than were cytotoxitassays (114, 115). ,-owever, impurity of antisera and cross.reactivity with nonpathogenicClostridial antigens resulted in too frequent false.negative and false-positive tests for the testto be useful in screening or diagnosis 1116-1181, Gas.liquid chromatography has bees used todetect a fatty acid metabolite of C. dtffleale in stools from patients and controls, and althoughspecificity was good. sensitivity was only 61% [119). Seotyping has not been useful. Abacteriophage and bacteriocin typing scheme is under development for use in epideuiologicalinvestigations, but is unlikely to be widelyused for routine diagnostic purposes [1201, Anotherpotentially useful tool for epidemiological studies involses a typing method based onpolyacrylamide gel electrophoresis of tI'tS.methionme.labeled proteins [1211.

RecontnendationsPatients with onset of diarrhea after starting a course of antibiotics

(especially clindamycin, ampicillin, ana cephalosporins) should have flex-

ible sigmoidoscopy if practical and specific therapy initiated if typicalpseudomeimbranes are seen. Full colonoscopy increases detection of

/


pseudomembranes from 30% of antibiotic-associated colitis patients to85% (122] but is not recommended in severely ill patients. Microscopicexamination of stools is not helpful. Specimens for culture should be

obtained during regular lab hours to permit rapid plating for enhancedrecovery. Satisfactory tests for cytotoxin B are commercially available in kitform. While the "Culturette Brand CDT" C. difficile test apparently doesnot detect enterotoxin A, it does appear to have clinical usefulness andpermits a presumptive diagnosis in hours instead of days.

Vibrio parahenolyticus

V paraheimolyticus are pleomorphic, halophilic grant negative rods ofthe family Vibrionaceac. These organisms are present in the marineenvironment and are responsible for outbreaks of gastroenteritis followingthe ingestion of raw or inadequately cooked fish or shellfish. Suchoutbreaks have been well recognized in Japan. In the US they have beenassociated with the ingestion of crabs or shrimp. The incubation period ofthe illness is 9-24 hours. The organism produces small superficial coloniculcerations and the presence of fecal leukocytes can be demonstrated.Illness can present as cholera-like watery diarrhea or as a dysenterysyndrome.

ProceduresV. parahemolyticis is isolated on TCBS (thiosulfate citrate bile salts

sucrose) where the colonies have a blue center. Enrichment can beaccomplished in Cary-Blair or alkaline peptone water with increased saltconcentration. The organisms differ from other vibrio species in theirlateral flagella, growth at 40 *C, utilization of putrescine and lack ofswarming on complex media. A serotyping scheme based on 0 and Kantigens has been developed. Pathogenic organisms (96% of clinicalisolates) have been associated with the production of a thernostablehemolysin which induces the Kanagawa phenomenon (KP) or hemolysis of

type 0 RBCs on Wagatsuma agar. 99% of environmental isolates lack thishemolysin.

A gene probe for this hemolysin, which is chromosomally determined, has beenprepared and tested in a dot colony hybndization test (DVII) 11231. DCH is more sensilinethan the Kanagawa test for detecting the hemolysin and compares fasvorably with immunologi-cal tests including the ELISA and modified ELEK (immunopreopltation) tests 1124).

Boedeker/Nieueen 194

Recently, strains hemolysin negative by gene probe and immunological testing (all s1rova4ant

03:K6) were isolated from an outbreak of Traveller's diarrhea in the Maldives 11251, thusother pathogenic preperties may actually be important for virulence. However, this studymust be interpreted it light of the demonstrated instability of the hemolysin gene i culture

Recommendations

This organism should be sought when there is a recent history ofseafood ingestion, particularly during the summer months. The presenta-tion may include evidence of active colonic inflammation. Tests for thethermostable hemolysin, including the gene probe, are not routinelyavailable in the US.

Aeromonas Species

The species of the genus Aeromonas are gram-negative, rod-shaped,oxidase-positive, facultative anaerobes with polyflagellae currently Ested inthe family Vibrionaceae (126]. These organisms are present in fresh waterand sewage and include a nonmotile group (salmonicida) which are fishpathogens. Species implicated in human disease are motile with single polarflagellae.

Isolation of Aeromonas from stool as predominant organisms has beenassociated with sporadic cases of diarrhea, but not with outbreaks. Theyhave been associated with Traveller's diarrhea in Thailand [1271, withsummer diarrhea in Australian children 1128] and in neonatal diarrhea.Taxonomy has been evolving. The current edition of Bergey's manual [126]divides the genus Aeromonas into the three species: hydrophila, caviae andsobria based on biotyping and analysis of DNA relatedness groups 1129].The species hydrophila and caviae produce potential virulence factors [130,131] but the relation of these factors to enteric pathogenesis remainsunresolved.

ProceduresAeromonads survive in hypotonie media, in buffered glycerol saline or

Cary-Blair transport media. They grow well on primary non-selectivemedia, such as sheep blood agar. Optimum growth is at 28 *C. Since moststrains are ampicillin resistant, ampicillin addition to blood agar platesprovides a selective medium. Enrichment can be achieved in peptonewater, or by maintenance at 4 °C for 2411 in buffered saline. These

Laboratory Inuestigation of Childhood Enteric Infections 195

techniques are of primary importance for isolation from environmentalsources, convalescent patients or possible carriers.

The organisms appear similar to E. coli on blood agar and on Hektoenenteric medium. Since they are oxidase positive, representative colonies onnonselective media should be screened for oxidase activity by floodingareas of the primary isolation plate with oxidase reagent turningAeromonas, Plesiononas and vibrios dark purple. This is the fastest test todifferentiate from E. colt, and it is also more practical than the use ofampicillin selection, since it gives an estimate of the predominance ofcolonies. Some labs only report Aeromonas if the organisms are heavy onnonselective media. Sensitivity testing on isolates representing fewer than5% of colonies will usually require a specific request front the clinician.

Oxidase.positie colonies can be confirmed by API 20E testing and by sensitivity tocompound 0/129 (2,4-diamino.6,7.disopropylpterndine). Kaper et al. 11321 described amultitest screening medium which combines manintol (postrive) and inostol (negaitne)fermentation, ornithine decarboxylation (negative) and deammatton (negative), indoleproduction (positive), motility (positive) and 1l1S production. Species determination ofconfirmed isolates can then be made by use of. Vosges-Postauer reaction (variable withsobfia. positive for hydrophita); IlS production from cysteine (positive for hydrophilia andsobria); gas from glucose (positisc for hydrophiia and sobria), esculin hydrolysis (positive forcaviae and hydrophitla) and salicin fermentation (positike with caviae and hydrophila), Thethree species have DNA homology %aloes of 35-50% Popoff ct al. 11291 have described 3DNA hybridization groups in A. hydrophila; 2 in A. raviae and 2 in A. sobria; representing70-90% relatedness amongthe groups in each species. At present, these groups within a singlespecies cannot be differentiated by phenotypic characteristics.

Aeromonas species produce a variety of potential virulence factors 1133 including; (a)hemolysms: demonstrable in culture supernates with 1% suspensions of rabbit erythroctcs inmicrotliter wtells, or by a plate method ilth 5% sheep blood in tryptic soy agar, (b) cytotoxin.demonstrable in cell free supernatants against Vero cells (1341. (c) heat-labile (56 'C. 10min)enterotoxin; actie against YI adrenal cells or demonstrable in the suckling mouse assaycomparing intestinal weight to remaining body weight after intragastrie injection (positivegreater than 0.08) (12S; or in a 16-hour rabbit ileal loop assay 11351.

In a recent study 11301, none of these factors were found with caviar strains, all ,erefound with hydrophila stratus, whereas sobria produced only hemolysin and cytotoxin. This isconsistent with the association of hydrophila and caviae isolates with diarrhea.

DiscussionWhether Aeromonas species represent true enter . pathogens has

been controversial. Epidemiologic studies of the association of Aeromonasspecies with diarrhea must be interpreted in light of whether they wereisolated from stool with enrichment techniques and selective media, or


whether they were predominant organisms on direct plating. Small num-bers of organisms in the stool may simply reflect the characteristics of thelocal water supply. In some areas of the United States, e.g. Michigan,Aeromonas are not found in normal stools even with selective media [1361.When rigorous isolation techniques were used in Thailand, Aeromonas wasisolated frequently (25-30%), and equally often from Thais with andwithout diarrhea [127); however, at the same time, they were isolated moreoften from Peace Corps volunteers with diarrhea than from those withoutdiarrhea.

Because of the uncertain clinical importance of isolating of smallnumbers of Aeromonas from the stool, we recommend that primary growthbe performed on nonselective media, such as sheep blood agar, withoutenrichment. This is in accord with the findings of Robinson et al. 1137) that89% of enterotoxigenie strains isolated from direct plating were associatedwith diarrhea: whereas only 21% of those isolated after enrichment wereassociated with diarrhea, even without ampicillin. In an Italian study ofchildhood gastroenteritis [138] ampicillin plates yielded no significantdifference in isolation rates between diarrheal (3.7%) and nondiarrhealstools (2.1%). We agree with Holberg at al. [1391 that the "determinationof the incidence of these bacteria as a possible cause of enteric disease willdepend on the ability and willingness ... to test predominant colonies onnonselective media for oxidase positivity".

Not all authors agree with the recommendation to use nonselectivemedia, Agger et al. [136] studied nonselective sheep blood agar vs. theselective medium sheep blood ampicillin (SBA) and concluded that SBAgave increased frequency of isolation but only ir specimens from patientswith diarrhea. Similarly, Kay et al. 1140], working in Peru, recommendedselective media and enrichment in peptone water for initial isolations. Suchrecommendations may depend on the low prevalence of Aeromonas inlocal water sources.

Interpretation of epidemiologic studies are complicated by the factthat, although a number of potential virulence factors have been describedin Aeromonas; their production has not been conclusively related todisease. In Australia during the summer, Gracey et al. [1281 reported thatenterotoxin-producing Aeromonas (as defined with a suckling mouseassay) were found much more often in stools from children with diarrhea(10.2%), than in those from well children (0.6%). At the same time, theisolation rate of all Aeromonas, using blood agar containingp-nitro phenylglycerine (and later plus 10jig ampicillinlml), was similar in the well and ill


group children. In contrast to this evidence suggesting a pathogenic role forenterotoxin, volunteer challenge studies with strains known to producevarious putative virulence factors, have failed to produce disease in healthyNorth American volunteers [141]. All strains tested in volunteers werepositive in rabbit ileal loops; all produced cytotoxin, hemolysin, enteroto-xin and DNAase; but all were Sereny test and hemagglutinin negative.Until our understanding of Aeromonas virulence factors is improved,epidemiologic studies of Aeromonas isolation rates are unlikely to deter-mine whether Aeromonas is an enteric pathogen. Moreover, tests forvirulence factors cannot currently be recommended to confirm pathogenic-ity of individual isolates.

Viruses

In the 1970s, identification of the medically important viral etiologiesof enteric disease relied on direct detection of virus particles in stoolspecimens by electron microscopy or immunoassays. Since that time, majoradvances have occurred in the development of highly sensitive, specific,and rapid solid-phase immunoassay techniques, several of which arecommercially available. It is now possible to identify a causative virus froma diarrheal stool within a few hours, a much shorter time than is required toidentify Salmonella, Shigella, or other common bacterial pathogens. Thishas practical significance, particularly for pediatric patients, in that apositive diagnosis eliminates the need for costly broad-spectrum IV antibio-tics, permits reassurance in terms of prognosis. and facilitates appropriatedisease spread-control measures.

Acute viral gastroenteritis is an extremely common illness (second onlyto the common cold in the US) that affects all age groups and occurs inepidemic and endemic forms [142-144]. The clinical syndrome varies, but isgenerally rapid in onset, self-limited, and characterized by various combi-nations of diarrhea, nausea, vomiting, and systemic symptoms such as low-grade fever, headache, abdominal cramps, myalgias, and malaise. Bloodydiarrhea is not a feature of viral enteritis. Thus, if a fecal smear revealssheets of inflammatory cells, a viral etiology is probably excluded. Threemajor groups of viruses, rotaviruses, Norwalk-like viruses, and enterieadenoviruses, are now unequivocally recognized as medically important.The reported frequency of occurence of the groups is influenced by age ofthe patients studied, the geographical location, and the season. Other

Boeded.er/McQueen 198

diarrhea-associated viruses, such as calicivirus, astrovirus, and corona-virus, can be identified, but their medical importance is uncertain.

Rotaviruses

Rotaviruses (family Reeviridae) primarily affect children, causingapproximately half the episodes of infantile diarrhea that require hospitali-zation, but also cause illness in adults 1145]. A new classification schemegroups Rotavirus into five groups (A-E) based on antigenic properties,genome profiles, and nucleotide sequences [146]. Thus far, human strainshave all been in groups, A, B, or C, with the usual agents of childhoodgastroenteritis being in group A [146]. Diagnostic tests include electronmicroscopy, solid-phase immune electron microscopy (SPIEM), counter-immunoelectrophoresis, polyacrylamide gel electrophoresis of viral RNA,radioimmunoassay (RIA), ELISAs, latex agglutination [142], and dothybridization 1147].

ProceduresRotaviruses were orginally discovered in biopsies of proximal small

intestine from hospitalized infants, and detection of virus in biopsy or stoolspecimens by electron microscopy remains the standard by which new testsare compared. Rotaviruses are 70nim in size, contain dsRNA, and survivelow pH (pH 3), heating (50 "C), and most detergents and solvents. Electronmicroscopy requires expensive equipment and highly trained technicians,and is impractical for screening large numbers of samples, but can detectatypical viruses using convalescent sera. Serologic assays are helpful inepidemiological studies, but are not helpful clinically since detectable pre-existing rotavirus antibodies are found in most individuals. Furthermore,the need for seroconversion prevents diagnosis early in the illness.Although 50% of field strains which have been detected can now success-fully be grown in cell culture, this is time consuming and cumbersome.Therefore, the several variations of solid-phase immunoassays are thediagnostic tests of choice for most situations [148-151].

A partial list of the commercially available assays includes: Rotazyme(Abbot laboratories, North Chicago, Ill.), Enzygnost (Behring Institut,Marburg, FRG), Rotolex (Orion Diagnostics, Helsinki, Finland) andRotoclone which uses a monoclonal antibody. The most critical factors forthe performance of solid-phase immunoassays are the antibodies, but


convalescent human sera are not of sufficiently high titer for practical use.This difficulty has been overcome by growing group A animal strains(which cross-react with human strains) to high titers in cell culture for thepreparation of hyperimmune diagnostic sera. Thus, commercially availableassays detect group A antigens. Few non-group A rotaviruses have beendetected in humans, but it should be noted that two outbreaks of group Brotaviral diarrheal illness (as determined by ELISA and genomic analysis)affected adults as well as children in China, indicating a need to determinethe global significance of the B viral group in humans [145, 152].

A nucleic acid dot hybridization technique applied to the detection ofrotavirus in stool specimens is reported to be specific and 10-100 timesmore sensitive than a conventional ELISA f 147].

RecommendationsAll assays are sensitive and specific, and the choice depends somewhat

on the needs of the laboratory. For example, the Rotazyme test isperformed with polystyrene beads as the solid phase in plastic test tubes andis read in a standard spectrophotometer, making it a convenient assay to usewhen small numbers of samples are processed. Enzygnost is performed inmicrotiter plates, permitting the use of automated equipment for thehandling of large numbers of specimens (153]. A DNA hybridization test isnot commercially available and not likely to replace the more convenientimmunoassays currently in clinical practice, since it has the disadvantagesof using short-lived radioisotopes that require special handling.

Norwalk-Like Viruses

The Norwalk-like viruses are a frequent cause of diarrheal illness(particularly in older children and adults), and usually occurs in epidemics.The Norwalk virus was the first virus to be convincingly associated withhuman gastroenteritis and was discovered by Kapikian et al. [154] in 1972by immune electron microscopy (IEM) studies on stool specimens obtainedduring a diarrheal epidemic in Norwalk, Ohio. These viruses are shed insmall amounts and for only a short time after the onset of diarrheal illness(3 days). Other viruses included in this group are the Hawaii. MarinCounty, Ditchling, "W", Snow Mountain, Parramatta, and cockle agent[155]. At least four distinct serotypes have been identified by immuneelectron microscopic studies of antigen relatedness.

Boedeker McQueen 200

ProceduresNorwalk-like viruses can only be visualized by the laborious and

technically difficult procedure of immunoelectron microscopy, and thedevelopment of other diagnostic tests has lagged far behind developmentsfor Rotavirus. These viruses are smaller (27 nm) than rotaviruses, round inshape, and the composition of their nucleic acids is uncertain, althoughstudies of the proteins of purified Norwalk virions indicate that this virus isrelated to RNA-containi'ig caliciviruses (155]. The virus has not beencultivated in vitro or in laboratory animals, and unlike Rotavirus, has noknown cross-reactive substitute antigens that can be produced in highlypurified forms, preventing production of hyperimmune serum reagents inanimals. Therefore, the critical reagents for Norwalk virus immunoassaysmust be derived from serum or stool samples from human volunteers whohave been infected with Norwalk virus, and this significantly limits theiravailability.

RecomnmendationsA radioimmunoassay is in use for detection of Norwalk virus in stools

and for the quantitation of antibodies to the virus in sera and intestinalfluids, but is limited to the few research laboratories that possess limitedamounts of the valuable human diagnostic reagents 1156, 157]. The recentdevelopment of a monoclonal antibody to the Snow Mountain agentindicates future promise for a practical diagnostic assay 1158].

Adenovirus

The importance of adenovirus as a cause of diarrheal illness has beendifficult to establish, since they have often been visualized in stools fromchildren who do not have enteritis. However, it has recently been estab-lished that two of the 41 known adenovirus types, adenovirus type 40(Ad40) and adenovirus type 41 (Ad41), are clearly implicated as etiologicagents of gastroenteritis. They cause 5-10% of diarrhea in children lessthan three years of age (although the total extent of disease probably hasnot yet been determined 11591, and clinically resemble Rotavirus infectionin many respects. In contrast to Rotavirus, Ad 40 or 41 associated diarrheais typically more persistent, with one-third having diarrhea for as long as 14days [160], and vomiting is milder. Respiratory symptoms are seen in aboutone-fifth of patients.


ProceduresAd40 or 41 can be visualized by EM during acute stages of the illness

when they are excreted in large numbers [161]. Ad40 and 41 are 70nm insize and contain DNA in their genome. They cannot be cultivated in celllines traditionally used for isolation of adenovirus. They can be isolated in ahuman embryonic kidney cell line (Graham 293) which has been trans-formed by adenovirus type 5-sheared DNA [1451, and these growthcharacteristics were initially used to identify EAd, but this has since beenshown to be unreliable. Ad 40 and Ad 41 are closely related antigenically,but they can be distinguished by a neutralization test in certain cell lines1161].

Se%,cral nonculture tests for detection of adenovtrus type 40 and 41 have been reported,including polclonal antibody-based enzmo immunoassay tests, direct genome profihng ofvirus in stools, nucleic acid hybridization techniques, and quantitative electron microscopy1162). A monoclonat antibody-based immunoassay has recently been developed and reportedto be both sensitive (compared to isolation by growth in Graham 293 cells), and specific asdetermined by genome profiles after digestion with Smal endonuclease (1621.

RecomtendationsBecause Ad40 and 41 cannot be distinguished morphologically from

nonpathogens, EM is of no use in diagnosis. The definitive means forisolation and identification of these viruses has been growth in cell culturefollowed by viral genome typing. Analysis of viral DNA or polypeptidesrequires large amounts of viruses and is therefore not practical for routinediagnostic work. The immunoassays are rapid, specific, and, presently, theoptimal method for detection of tle two EAd types. A combined commer-cial test kit based on specific latex agglutination for detection of Adenovirusand Rotavirus has recently been marketed (Adenolex and Rotalex, OrionDiagnostica, Espoo, Finland) and is reported to be greater than 90%sensitive and 98% specific for both Rotavirus and Ad [163], but is limited inusefullness because of difficulties and distinguishing pathogenic and non-pathogenic strains.

Other Viruses

CalicivirusCaliciviruses have been reported primarily in children in community-

wide and nosocomial outbreaks of gastroenteatis, as well as from sporadic


cases in various parts of the world. There appear to be at least four or fiveserological types of human calicivirus by IEM, none of which cross-reactwith known animal caliciviruses. By electron microscopy, caliciviruses havea scalloped border and cup-shaped indentations on their surface that canresult in a Star of David appearance. They are 31-35 nm in diameter andcontain a single, plus-stranded RNA genome [164]. Rises in sertm titers toNorwalk virus have been noted after calicivirus infection, and this inaddition to similarities in a structural protein indicate these viruses may berelated. Caliciviruses of human origin have only recently been cultivated invitro [165] and an RIA assay has been developed [166], so this shouldfacilitate additional studies of this agent.

AstrovirusAstroviruses have been detected in large numbers by IEM during

outbreaks of gastroenteritis from stools of young children, in pediatricwards, and in nursing homes. They seem to be of low pathogenicity, sincefew volunteers orally inoculated with stool filtrates have developed illnesseven though they shed the organisms in large numbers. At least fiveserotypes have been reported. They have a characteristic morphologyunder EM, having an electron dense center with triangular electron-lucentareas resulting in a five- or six-pointed star appearance. Cultivation hasbeen difficult, but serial passage in embryonic human kidney cells andsubsequent adaptation to growth in a continuous monkey kidney cell line,LLCMK2. has been reported [167].

Small Round VirusesLittle is known about this poorly defined group of viruses with names

such as Otofuke agent, Sapporo agent, minireovirus, and minirotavirus.For the most part, their significance as etiologic agents of gastroenteritisremains to be established. They have been reported from a variety ofoutbreaks but have also been detected in well individuals. They aregenerally small (20--30nm), and they lack a distinct ultrastructure. Little isknown about their biophysical or biochemical cnaracteristics. Cultivationand the development of specific assays has not been achieved.

Protozoal Parasites

Diagnosis of intestinal parasitic disease still depends on the directmicroscopic evaluation of stool specimens, in adequate numbers, by


experienced observers. To recognize the cyst and trophozoite forms, stoolsmust be appropriately collected, prepared and preserved. Concentratingmethods may be used to obtain larger numbers of cyst forms. Advanceshave been made in tests for host serologic responses, but these are usuallymore helpful forstudies of disease prevalence in population groups than fordiagnosis of disease in individual patients. One exception is the use ofparasite antigen specific IgM assays which may indicate recent infection.Parasites and their antigens can now be detected and identified in stool withmonoclonal immune reagents using fluorescence microscopy, counter-immunoelectrophoresis (CIE) and ELISA. This is true for Giardia andCryptosporidia. Although axenic culture has been achieved for bothamoebae and giardia, culture of protozoa remains a research tool which isnot routinely utilized for diagnostic purposes. Culture can be used to permitexamination of protozoal enzyme patterns (zymodemes) and DNA restric-tion digest patterns which may be related to pathogencity. No DNA probesare currently available for parasites. This section will review newermethods for the diagnosis of Entainoeba histolytica, Giardia laniblia andCryptosporidia.

E. histolyticaCurrent estimates of prevalence [168] place E. histolytica third among

the causes of protozoal-induced deaths (after malaria and schistosomiasis).Ameobiasis may present as invasive enteric disease (Dysentery), as liverabscess or as asymptomatic cyst passage.

Diagnosis is by stool examination. Even in dysenteric cases, the stoolscontain very few polymorphonuclear leukocytes (fewer than 50/hpf) incontrast to the findings in bacillary dysentery. Examination of rectalbiopsies has not proven more efficient than the careful examination of threeseparately collected stool specimens performed both in fresh saline suspen-sions and after concentration by the zinc sulfate or formal-ether method[169]. One cannot differentiate the casts of pathogenic and nonpathogenicamoebae, although motile pathogenic trophozoites can be distinguished bytheir phagocytosis of red cells (hematophagous). The observation ofscrapings from active ulcers for trophozoites in fresh saline suspensionmaintained at 37 *C may improve chances of diagnosis of active disease.Standard polyclonal antisera have been developed to detect E. histolyticaantigen in stool, but these are no longer commercially available [170].

Serologic tests for serum antibodies are most helpful when thereis difficulty in performing the stool exam because of interfering sub-

Boedeler/McQueen 204

stances such as barium, bismuth or clay; in cases of invasive disease withliver abscess; and in epidemiologic studies. Antibody can be sought byhemagglutination (HA) inhibition, ELISA or agarose gel diffusion.Stool antibody can be detected by these techniques as well, in par-ticular by counterelectrophoresis (CEP) and indirect immunofluorecence(IF).

An area of recent controversy has been the validity of distinguishingpathogenic from nonpathogenic forms of E. histolytica. Debate has cen-tered around the question of whether presumed nonpathogenic forms areonly phenotypic variants or whether they represent genetically distinctstrains. If the formercase obtains, then there may be no true nonpathogenicE. histolytica, since the nonpathogenic phenotypes might revert to viru-lence in response to as yet undetermined stimuli. Pathogenic and non-pathogenic variants have been described based on their isoenzyme patterns(zymodenes) of the 4 enzymes glucose phosphate isomerase, phopho-glucoconvertase, linolate NADP oxyreductase and hexokinase (two fastbands in pathogens). Twenty-two different zymodenes have been described[171,172]. The technique involves axenic culture of the isolates [173], lysis,and starch gel electrophoresis followed by visualization of enzyme bands.However, Mirelman et al. [174] have been able to achieve the convincing invitro conversion of nonpathogenic amoebae to pathogenic pheotypes bymanipulating the bacterial flora. Nevertheless, clinical outcome appears tocorrelate with zymodene pattern. Recently, Strachen et al. [175] reportedthe use of 2 monoclonal antibodies prepared against amoebal lysates of astrain with pathogenic zymodene, which appear capable of rapidly identify-ing cultures (but not cysts) of the fast hexokinase phenotype in animmunofluorescence assay. In vivo, one factor necessary for amoebicvirulence appears to be the presence of appropriate colonic bacteria.Putative amoebic virulence factors include a lectin which promotes attach-ment to epithelial cells and is mitogenic for T lymphocytes [176]. Amonoclonal antibody to this lectin has been produced but is not yet used forroutine diagnosis.

RecommendationDiagnosis of amoebiasis depends on the skilled examination of stool

specimens for cysts and trophozoites by multiple methods, using both freshsamples and stored, concentrated samples. One must recognize the pre-sence of agents which can interfere with the diagnosis by destroyingtrophozoites. Serology is most useful for invasive disease, but may be

= -- _ .L

Laboratory Investigation of Childhood Enten Infections 205

helpful in active disease if sto'1 examinations are inadequate. Methodsusing ELISA or IFA to detect antigen in stool are not commerciallyavailable. Culture methods remain a research tool. The use of zymodeneanalysis to distinguish pathogenic from nonpathogenic strains is also notgenerally available and it remains controversial whether the enzymepatterns identify genetically different strains or only reflect phenotypicvariation which may be induced by changes in the parasite's environment.The present weight of evidence appears to favor nontreatment of "non-pathogenic" phenotypes.

Giardia lamblia

G. lamblia is the most common pathogenic intestinal protozoa in theUS and in underdeveloped areas it is a nearly universal infection of infants.Populations at risk for this water-borne, flagellated protozoan includechildren in day-care centers, travellers to endemic areas where water

supplies are contaminated, backpackers, and immunocompromised hostsincluding those with IgA deficiency and common variable hypogamma-globulineinia. Water supplies may be contaminated by beavers and mus-krats. Age-specific incidence in day care centers indicate the possibility ofperson to person transmission. The parasite colonizes the uper small bowel,and diagnosis is complicated by the fact the parasites may not be detectablein the stool for 7-14 days after the onset of symptoms as shown in studies oftravellers returning from Russia to Finland 1177]. Thereafter, stool excre-tion of cysts may be intermittent. Onset of symptoms is insidious withabdominal bloating and distention accompanying diarrhea. The predilec-tion of the organisms for the upper small bowel appears to be related to arequirement for bile salt for optimal growth conditions. Pathogenesis of thediarrheal disease caused by giardia is not understood. Putative virulencedeterminants include a surface lectin which may be activated by trypsin andduodenal secretions [1781.

Diagnosis is by stool examination, duodenal fluid aspirate and smallbowel biopsy. Stool examination for ova and parasites should utilize iodinestained wet smears, trichrome stained cyst concentrates (performed byformalin ethyl acetate centrifugation or zinc sulfate concentration) andtrichrome stained polyvinyl alcohol (PVA) preserved stools. Three stoolsshould be examined. Sensitivity is 50-70%. Duodenal fluid samples may beobtained by the EnterotestO which is an absorptive sponge tethered on a


140 cm string. Small bowel biopsy yields a crush or smear preparation fordirect microscopic examination but may not be useful early in disease (1791.In the event of high clinical suspicion and negative diagnostic tests atreatment trial with metronidazole may be undertaken.

Immunodiagnostic reagents have been developed for the detection ofgiardia antigen in stool or intestinal fluid samples. CIE [180] and ELISA[181, 1821 have been used by research laboratories with 90% specificity andsensitivity. Alexon biomedical (Rolling Meadow, Ill.) recently marketed a"prospect/giardia" EIA test which utilizes tubes coated with antigiardiamonoclonal antibody which are used in a sandwich Elisa. This test is basedon the monoclonal antibody to giardia antigen (GSA 65) described byRosoff and Stibbs [183]. Meridian Diagnostics (Cincinatti, Ohio) hasmarketed an indirect immunofluorecent detection procedure (Merifluor/giardia) for the detection of giardia cysts in fixed fecal material or in watersupplies.

ScrodiagnosisAn IgM ELISA has been developed using trophozoites coated in the

wells which has good specificity at 1/200 and 11400 dilutions of serum foractive disease and recent infection. Antigen-specific IgM falls in 2-3 weeksafter active infection with treatment [184], whereas antigen-specific IgGremains elevated 1185].

Axenic culture of G. laniblia is now possible. This technique haspermitted differentiation among isolates with respect to surface proteinlabelling, isoenzymes, chromosomal banding and restriction endonucleasedigest patterns, as well as growth in culture and drug sensitivity. Recentstudies [186] of experimental human infection of volunteers have shownthat Giardia isolates have different virulence properties.

RecommelndationA high index of suspicion must be maintained in chronic illness in

travellers, children in day care centers and susceptible hosts. Repeatedanalysis of stools should yield diagnosis. Enterotest exam ;s probablynot more sensitive. The newer monoclonal antibody tests to detectgiardia antigen in stool appear to be highly sensitive and specific andmay replace direct microscopic observation of fecal specimens. IgMELISA has a role in detecting active infection. When suspicion is high, butdiagonstic test do not confirm the diagnosis, clinical trial of therapy maybe instituted.


Cryptosporidia

The high prevalence of chronic cryptosporidial infection and lifethreatening diarrhea in patients with HIV infection and AIDS hasincreased awareness of and diagnostic capability for this once seldomrecognized protozoa. It is now recognized as a frequent cause of transient(10-24 days) self-limited diarrheal illness in immunocompetent hosts, inchildren in day care centers, in water-borne outbreaks (where it isfrequently found together with giardia), in seasonal outbreaks during warmhumid months, and among travellers [1871. Carriage is extremely prevalentin young children in underdeveloped countries, although lower in breastfed infants. This was initially felt to be a genus with different species foreach host. However, when animal (calf)-to-human transmission was recog-nized [188, 189], it was felt to be a single species genus. At present, sixspecies are recognized, of which all human disease is caused by C. parvn.There are many reservoir hosts.

The organism has a unique life cycle within the intestine of a singlehost, in the lumen and at the apical surface of tle intestinal epitheliumwhere it is intracellular but extracytoplasmic in a parasitophorous vacuole[187]. Diagnosis was initially made by light or electron microscopicexamination of intestinal biopsies, where projecting organisms appear asbasophilic vesicles at the surface on H&E. It is now made by recognition ofrefractile, 2-1tm diameter circular oocytes in phase contrast microscopy ofconcentrated stool specimens obtained by Sheather's sugar flotation tech-nique. These can also be seen by iodine stain of wet mounts, or conve-niently sedimented in 10% formalin-preserved stool by Kinyoun's acid faststains where they stain red, or modified Ziehl-Nielsen stain [190]. Meridiandiagnostics (Cincinatti, Ohio) has recently marketed an indirect inmmuno-fluorescent detection procedure (merifluor/cryptosporidium) based on amurine monoclonal antibody against an oocyst determinant [191, 1921.

The organisms cannot presently be cultured, but complete develop-ment can be shown in cultured human fetal lung cells and chicken or porcinekidney cells infected with oocytes [1931.

SerologyCampbell and Current 1194] demonstrated serum antibody to crypto-

sporidal antigens by 111 in recovered immunocompetent hosts and AIDSpatients. Serum antibody could not be determined in patients withhypogammagiobulnemia. Their assay utilized intestinal tissues from mice

BoedekerfcQueen 208

infected with oocysts obtained from calves. This assay permitted determi-nation of antibody to life forms other than oocysts. Titers of 1/40 weresensitive and specific. 95% of those infected have both IgM and IgGresponses by ELISA except in AIDS (16% IgM, 90% IgG) [195]. This assaywvas used for seroepidemiology studies of infection in two latin Americanpopulations in Lperu and Venezuela. In these populations 15-20% hadserum IgG and IgM indicating recent infection (highest at age 2-3 years)and 64% had detectable IgG indicating endemic infection in these com-munities 1196].

RecommendationThis organism is not confined to patients with immunodeficiency, but

should be considered as a cause of transient illness in a range ofimmunoompetent hosts. Transmission from animals to man is now estab-lished. Detection of oocysts in concentrated stool specimens remains theprinciple means of diagnosis. This can be aided by immunoifluorescentantibody techniques. An ELISA technique for quantitative serology hasbeen developed, replacing the original technique which required infectedmouse tissue.

Referentces

I Echeveria, P.; Seriwatana, J.; Setbabutr,O0.; Taylor, D.N.: DNA hybridizationinithediagnosis of bacterial diarrhea, Clins Lab, IMed, 5:447-462 (1985).

2 Wachsmuth, IC.: Molecular epidemiology of bacterial infections, Examptes of methodol-08y and investigation of outbreaks. Rev, infect, Dis. 8. 682-689 (1986).

3 Grunstein, Nt., Hognnss, D.: Colony hybridization. A method for the isolation of ctonedDNA's that contain a specific gene. Proc. nain. Acad. Sci. USA 7: 3961-3965 (1975).

4 Soanenwirth, A.C.: The eateric baciMl and bacteroides; in Davis. Dutbecco, Elisen,Ginsberg, Microbiotogy; 3rd ed., pp, 646-672 (Harper & Row, Hagerstoswn 1980).

5 Hornick, Rtt.: Satmonelta; in Gorbach, infectious diarrhea, pp. 17-30 (Bltackwell,Basion 1986).

6 Coilon, AR.; Gross, D.R.; Tamer, M.Adz Typhoid fever in chitdren, Pediatrics,Springfietd 56: 6066W9 (1975).

7 Mandal, B.K.; Maui, V.: Colonic involvement in samonetosis. Lancei: v687 (1976),8 Rubin, F.A., IKopecko, D.J.; Sack, R.B.; Sudarmono. P.; Yi, A.; Mauna, D., Mecza.

R.; Mochltar. MIA.; Edman, D.C.; Hoffman, S. L.: Evatuation of a DNA probe foridentifying Salmnonrlla typht in Peruvian and Indonesian bacterial isolates J. infect. Dis.157: 1051-1053 (1988).

9 Rubin F. A., Kopecko, D. J., Noon, K. F., Baron, L S. .Development of aDNA probeto detect Salmonerlla typhL J. clin. Microbiol. 22. 600-605 (1985).


10 Martin, W.J.: Laboratory Studies; in Corbach, Infectious diarrhea, pp. 257-285(Blackwell, Boston 1986).

11 Taylor, D.N.; Wachumuth. I.K.; Shangkuan, Y.; Schmidt, E.V.; Barrett, T.J.;Schrader, I.S.; Scherach, 0.5.; McGee, I-lB.; Feldman, R.A.; Birenner, D.J.:Salmoncliosis associated with marijuana. A multistate outbreak traced by plasmidfingerprinting. New Engl. J. Med. 306. 1249-1253 (1982).

12 Taylor, D. N.: Echeverria, P.; Pal, T.; Sethabutr, 0.; Saiborisuch, S,; Sricharuaorn, S.;Rowe, B.; Cross, .:Terole of Shigella app., enteroinvasive Eschericliia coi. and otherenteroipathogens as causes of childhood dysentery in Thailand. J. infect. Dis 153:1132-I1138 (1986).

13 Rennels, M. B.; Levine, .MM.: Classical bacterial diarrhea: perspectives and update -Salmonella, Shigella, Eaclierichia coi. .4cconoana nd Plesiomonoa. Pediat. infect. Dis.& S91-S 100 (1986).

14 Keusch, G.T.; Sligella; in Gorbach, Infections diarrhea, pp. 31-50 (Black.well, Boston19S6).

15 Sansonetti. P.).; Kopecko, D.i.; Formal. S. B.: Ivolvement ofia plasinid in the invasiveability of Sliigella flextirri. Infect. Immunity 35:852-860 (1982).

16 Sansonetti. P.J.; Dflanterville, IL.: Formal, SAB: Toucas. MI.* Plasmid-mediatedinvasiveness of 'Shigella-liIke' Eschrridiia coi. Ann. Microbiol. 132: 351-355 (1982).

17 WVatanabe, H-.; Nakamura. A.: Large plasmids associated with virulence in shigellaspecies have a common function necessary for epithelial cell penetration. Infect.Immunity 48:260-262 (1985),

18 Pal, T.: Paean, A.S.; Emody. L.; Voros. S.: Selley, E.: Modified cnyme.Iiulkedimmunosorbent assay for detecting euleroinsasive Eachefichia coli and virulent Sliigellastrains. J. dlin. Microbiol. 21:415-418 (19S6).

19 Radetsky, M.: Laboratoiy evaluation of acute diarrhea. Pediat. infect. Dis. 5: 230-238(1986).

26 Serray, B.: Experimental Shigella keratoeonjuuctiviiis. Acta microbiol. Acad. Sci. 2:,193-196 (1955).

21 Oak~s, E2V.; Wingield, MX.; Formal, S.B.; Plaque formation by virulent Sliigellaflextieri. Infect. Immunity 48: 124-129 (1985).

22 Hlale, T. L. Sansonetti, P.; Schad, P. A., Austin. S.; Formal. S.B.: Characterization ofvirulence plasmids and plasmid-mediated outer membrane proteins in S/tigellajflextien,Siiigella sonuci. and Eachecichin coi. Infect. Immunity 40.-340-350 (1983).

23 Mosely, S.L.; Echeverria. P,. Seriwatana, L.; Tlrapat, C.; Chaicumpa, W.; Sakul.daipeara. T.; Fallow, S.: Identificationi of enterotoxigenic Eschericliia coi by colonyhybridization using three euterotonin gene probes. J. infect. Dii. 145: 863-869 (1982).

24 Boileau, C. I.; dlanteville. If. M.; Sansonetti, P.J.. DNA hybridization technique todetect Siiigelia species and enteroinvasive Eaclierichia coi. J. cin. Microbiol. 20.:959-961 (1984I).

25 Sethabutr, O.; eheverria, P.; Taylor, D, N., Pal, T.; Rone, B. : DNA hybridization inthe identification of euteroinvasive Eucheutchla coi and Sliigelia in children withdysentery; in Tzipori et 31., Infectious diarrhea in the young. pp. 350-356 (ElsesserBiomedical Division, Amsterdam 1985).

26 Sethabuir, 0.; Ilanchalay, S.;. Echeserria. P.; Taylor, D.N.; Leksomboon. U.: A non.radioactive DNA probe to identify Shigella and enteroinvasise Escherachia Coll in stoolsof children with diarrhoea. Lancet d., 1095-1097 (1995).

Boedeler/NfcQueen 210

27 Boedeker. 1.0.; Sherman, P.M.: Mechanisms of Escilerichia coli enterstis Front.gastrointest. Res., vol. 13, pp. 309-330 (Karger, Basel 1986).

28 Levine,.MM.: Eschericlin coil that cause diarrhea: enterotoxigenic, enteropatlingenic.enteroinvasive, enterohemorrhagic and enteroadherent. J. infect. Dis. 15:. 377-389(1987).

29 Murray, S. L.; Matthcwson, 3.3.; DuPont, H. L.; Hill, W. 1.: Utility of oligonacleotideprobes for detecting enterotoalgenic Escilerichia coi. J. infect. Dis. 155:809-811 (1987).

30 Mfoseley. S.L.; Hoq, L.; Ahm, A.R.M.A.: So. MI.:SamadpuMaaotalhi, M.;, Falkow.S.: Detection of enterotoxigenic Eschericlia coli by DNA colony hybrtdization.iJ. tnfect.Dis. 143: 863-869 (1982).

31 Moseley. S.L.; Echeserria, P.; Seriwatana, J.; Tirapat. C.: Chaicnmpa, W.; Saknl.daipeara, T.. Falkow, S.- Identification of enterotoxigenic Esclierichin colt by colonyhybridizatior4 using three enterotoain gene probes. J. infect. Dis. 15S. 809-811 (1987).

32 Robins-B~rowne, R.M.: Traditional enteropathogenie Eacerichin coi of infantilediarrhea. Rev, infect. Dis. 9: 28-52 (1987).

33 Levine, M.Ml.; Edelman, R.E.: Enteropathogenic Eacherichia coli of lassic serotypesassociated with infant diarrhea: epidemiology and pathogenesis, Epidemniol. Rev. 6:31-51 (1984).

34 Moon. 11W.; Whipp, S.C.; Argenzio, R.A.; Levine, M.M.; Giannella. R.A,:Attaching and effacing acttvities of rabbit and human enteropathogenic Eacherichia coiin pig and rahbit intestines. Infect. Immnnity 41.,1340-1351 (1983).

35 Nataro, JP.: Scaletsty. I.C.A.; Kuper,. B,;Levine, N.M]N.: Trablsi, L. R.: Plasmid.mediated factors confering diffnse and localired adherence of enteropathogenicEscliericliia coli. Infect. Immunity 48:378-383 (1985).

36 Scaletsky, I.C.; Lonrdcs. NI.; Trabulsi. L. R.; Distinctive patterns of adherence ofeniteropathogenic Escherichin coi to lIeI~a cells. Infect. Immunity 43: 534-536 (1984).

37 Ilaldini, M.M.: Kaper, J.11,: Lcvine, MM.; Candy. DC.A.; Moion. ILW.. Plasosidmediated adhesion of enteropathogenicEacliericiu coiJ. pediat. Gastroenterol. lNutr,2:534-538 (1983).

38 Ilaldini. MM:Nataro, j.P.; Kaper. J.11.: L.ocalization of a determinant for I-lip. 2

adherence by enteropathogenic Eacherichiu coli. Infect. Immnnity52. 334-336(1986).39 Nataro. J.P.; Ma3her. K.O.: Mfaclie, P. Kaper, 3.11.: Characterization of plasmiss

encoding the adherence factor of enteropathogenie Eacherielna coi. Infect. Immunity35: 2370-2377 (1987),

40 Merson, M. IL.: Rowe, BI.; Black, R. E3.: Ilnq, IL: Gross, R.J.; liusof, A,: Use of antiserafor identification of enterotoxigenic Escheriehin coi. Lancet di: 222-224 (1980).

41 Orsl,ov, 1., Orakov, F.: Serology of Eachiclua coil imbniae. Prog. Allergy vol. 3. pp.80-105 (Karger. Basel 1983).

42 Donta, S.T.; Moon, 11W.; Whippi, S.C.: Detection of heat-labile Eachesadnoa coienterotoxin with the use of adrenal cells in tissne culture. Science 183:334-335(1974).

43 Sack, D.A., Sack. R.It.: Test for enterotoaigenie Eachericlnn colt using Y-l adrenalcells in miniceltnre. Infect. Immunity 11: 334-336 (1975).

44 Guerrant, R.L.; Brunton, L.L.; Schnitman, TC,; Rebhn .1., Gilnian, A.G..Cyclic adenosine monophosphate and alteration of CHIO cell morphology: a rapid andsensitive in Nitro assay for V~ clnleioe and Esclicrichiacolt. Infect. Immunity 10. 334-336(1974).

45 Svennerholm, A.M., llolmgren, J.: Identification of Escliericlna coli heat-labile

Laboratory Investigation of Childisood Enteric Infections 211

enterotoxin by means of a gangioside irmunosorbent assay (GM-I ELISA) procedure.Curt. Microbiol. 1: 19-23 (1978).

46 Sack, D. A.; Huda, S.; Neogi, P. B.. Daniel. R. R.; Spira, W. NI.: Microtiter ganiglosideenzyme-linled immunosorbent assay for V'. chlsoerae and £ochericitia coi LU enteroto-alas aad antitoxins, J. clit. Microbial. 11., 35-40 (1980).

47 Ristaino. P.A.; Levine, M.M.; Yong, C.R.: Improved GMN-I enzyme-linkedimmanosorbent assay for detection of Esciirrichia coi heat-labile enterotoxin. J. jin.Microbiol. 18, 808-815 (1983).

48 Honda, T.: Akhtar, Q.; Glass. R.I.; Golam Ktbraya, A. K.. A simple assay to drtectEscizeric/tin coi producing brat labile enterotoxin, Results of a field study of the Bikenrest in Bangladesh. Lancet it'.609-610 (1981).

49 Vadivelu, J.: Lloyd, B.J.; Draser, B.S.: Peachem. R.G.; Con, N.P.: Harrison, T.J.:Membrane filterrassay for detection of enterotoxigenie Esc/zecca coiin epidermoogtestudies, Laitet 1. 1007-1008 (1986).

50 Ronabeeg, B.; Wadstrom. T.: Rapid detection by ceaggluttuation tesl of beat-labileeaiterotosin tocell lysates from bilood agar-grown scheric/iaucoil. . din. Mticrobiol. 17:1021-1025 (1983).

51 Dallas, W.S.; Gtll, D.M.; Falkow, S.: Cistrons encoding Eac/tericii coil brat labilerosin. J. Bact. 139: 850-858 (1979).

52 LanaitaC. K: Kaper..B.: Balldini,, MM.; Black, R. B.. Levine. NM.M.: Sensitivity andspecificity of DNA probes with the stool blot technique fordetection of Endueric/tin coildiarrhea, 3. infect. Dis. 152, 1057-1090 (1985).

53 Green, B.A.;. Neill, R.J.; Ruyechan, WV.T.; Holmses, R.K: Evidence that a newenterotoxin which activates adenylate cyclase in cucaryotie target cells is not plasmidmediated. Infect. Immunity 41:3834390 (1993).

54 Pick~ett. 0.1..: Taiddy, I3.M.; Beliste, B.WV.; Hlolmes, R.K. Clontng of genes thatencode a newv heat-labile enterotosin from Eac/icrictia coi. 3. Bact. 165:346-352 (1986),

55 Seriwatana. J,; Echeverna. P.; Taylor, D.N.; Rasrinaul. L.; Brown, J.C.;. Peirts,J.S.M.; Clayton, C.L.: 1T'pe 11 heat-labtle enlerntovin-productug Eschericliia coltisolated from animals and humans. Infect. Immunity 56: 1158-1161 (1988).

56 Chan, S.-K,; Gianuella, R.A.: Amino acid sequence of heat stable enteroroxinprosduced by Eachici coi pathogenic for man. J, biol. Chem. 2s6.-7744-7746(1981).

57 Dean, A.G.; Ching, Y..C.-; Williamas, R.G.; Hlarden, L.B.: Test for Eac/zerici coltetiterotonin using infant mice. Application in a stndy of diarrhea in children in H-onolulua.J. infect. Dis. 12S. 113-336 (1975).

58 Gianella, R, A.; Drake, K.W;., Luttrell, M.: Development of a radiominunassay forEnchericin coli heat stable entcrotosin: comparison wsith the suckling moose assay.Infect. Immunity j3: 186-192 (1981).

59 Frate, 3. C.; Robertson, D.C.: Immunological properties of Escheficlin coiheat stableearerotoxins: development of a radionmunassay specific for heat-siable enrerotoninwith suckling mouse activity. Infect. Immunity 33:193-198 (1981).

60 Moseley, S.L.; Hardy, J.N.; Ilnq, NI.I., Echevernia. P., Falkow, S.: Isolation andnucleoride sequence determination of a gene encoding a heat stable eurerotoxin ofEnchcricliia coi. Infect. Immunity 39, 1167-1174 (1983).

61 M~cConnell, NMh!., Thomas. L.V.* Day, N.I'; Rowe, B.: Einyme-linked immuno-sorbent assays for the detection of adhesion factor antigens of entcrotoxigroicEncherichin colt. J. infect. Dis. 132: 1120-1127 (1985).

Iloedcker/McQueen 212

62 Willishaw, G.A.; Smith. I1. R.; Rowe, B1.: Enzyme-linked immunosorbent assays forthe detection of adhesion factor antigens of enterotoxigenie Eschenichia co/i. J. infect.Dis. 152: 1120-1127 (1985).

63 Manning, P. A.: Timmis. K. N.; Stevenson, G.: Colonization factor antigen 11 (CFA/11)of enterotoxigenic Eschericuia coli, Molecular cloning of the CS3 component. Mol Gen.Genet. 200.:322-327 (1985).

64 Toledo, M.R.F.: Trabulsi, L.R.: Correlation between biochemical and serologicalcharacteristics of Eicheric/t/a co/i and results of the Sereay test. J. Clia. Microbial. 17:419-421 (1983).

65 IKonawalchati, J.: Speirs. 3.1.; Stavric, S.: Vero response to a cylotoxto of Esc/iericacoll. Infect. Immunity 18. 775-779 (1977).

66 O'Brien, A.D.; LaVeck. G.D.: Thompson. M.R.; Formal, S.B.: Production of S/age//adysen ser tpie cytotoxin by F-Wcieric/i/n co/i. J. infect. Dis, 146: 763-769 (1982).

67 Strockbine, N.A.; Marques, L.Rdl.; Newland, J.W.. Williams Smith, It.; Holmes.R.L.; O'Brien, A.D.* Two toxin converting phages from Esc/ieric/s/a co/i 0157:H7associated with hemorrhagic colitis. J. clin. Mlicrobial. 23:869-872(1986).

68 Karmali, M.A.: Petric. M.; Lim. C.; Fleming. P.C.; Arhus. 0.5.; Lior. H1.: Theassociation betwveen hemolytie uremie syndrome and infection by verotosin-peodncingEsc/iericl co/i. 3. infect. Pta. 151: 775-782 (1985).

69 March, S.B,. Ratma, S.: Sorbitol-MacConkey medium for detection of Esc/ier/c/i/a coil0157:117 associated with hiemorthagic colitis. J. clin. Microbiol. 23:869-872 (1986).

70 Psi, C.K; Ahmed, N.: Lior. II.; Johnson, W.M.: Sims. ll.V.: Woodt, Dl.3:Epidemiology of sporadic diarrhea due to verocytotoatn-peodncing Eariccia CO/i. Atoo year prospective study. J. infect. Dis. 157,,1054-1057 (1988).

71 Karch, 11.;Heeseman, J.; Laufs R.; O'Brien, A.D.; Tac),et, C..; Levine,M.M.: Aplasmid of enterohemorthagie Eseheri chin co/i 0157:117 is required fur expression of anew fimrbrial antigen and for adhesion to epihelial cells. Infect. Immunity 55: 455-461(1987).

72 Levine, M.M.; Xu, J..G.: Kaper, J.B.; Lior, If.; Prudo, B., Stoll, B.; Nataro, J.; Karch,I.; Wachsuth. K.: A DNA probe to identify enterohemorthagie Esclieric/tta co/i of0157:H17 and other aerotypes that cause hemorrhagic colitis and hemolytic uremiesyndrome. J. infect. Dis. 156 175-182 (1987).

73 Battler, J, P.; Stkirrow. M. B.: Cuoipy/ouncter eniicriris. Clin. Gastrocuterol. 8.:737-765(1979).

74 Blaser, M.J,; Reller. L. B.: Csnpy/nbncterenreruns. New Bug!. J. Mled. 305:-1444-1452(1981).

75 Walker, R.I.; Caldwell, M.D.; Lee, C.C.; Guecry, P.;, Trust, T.J.; Ruiz-Palacios,G.M.f: Pathuphysiology of Camtp) /obacter enterins. Microbial. Rev. 50., 81-94 (1986).

76 Romanink, P.J.; Zoltowuka, B.; Trust, T.3.; Lane, P.3.; Olsen, G.J.; Pace, N.R.;Stahl, P. A.: Camsy/obactrepy/ori, the spiral bacterium associa ted wi th human gastritis,is not a true Cnmpoy/obacter sp. J. Ilact. 169: 2137-2141 (1987).

77 Totten, P.A.; I'enuietl. C.L.; Tenoser, P.C.; Werenberg. J.M.; Periur, P.L.; Stamm,\V.E.; Hlolmes, K. K: Cauny/obacrer cinaed/ (sp. nov.) and Cuuipy/obacterfemieliae(up nov.). Peao new Campy/obacter species associated with enteric disease in homo-sexual man. 3. infect. Pus. 151: 131-139 (1985),

78 Luechtefeld, NW N.; WVang,. VL.; Itlaser, NM./.; Reller, L. B.: C'ampylobacter fetssubsp. jejuiin. Blackground and laboratory diagnosis. Lao. Mled. 12: 481-487 (1981).

Laboratorylnsestigation of Childhood Emtic Infections 213

79 Goosens, H.; De Btoek, M.; 1.aaduyt, H. van: Baizler. J.P.: Isolation of Campylobacterfejuni fronm human feces; in Buteler, Campylobacter infection in man and animals, pp.39-50 (CRC Press, Bocea Raton 1984).

80 Goosens. H.,. De Bock, NI.; Butzler, J.P.: A new selective medium for the Isolationof Campylobacter jejuaf from haman faecea. Eur, J. din. Microbiol. 2: 389-394(1983).

81 Buck, G. E.; Kelly, T.: Effect ot moisture content of the medium on colony morphologyof Carupylobacterfeais sabsp. iejauii. J. ella. Microbiol. 14. W8 (1981).

82 Steele. T.W.; McDermnott, S.N.; Technical note: the use of membrane filtees apptieddirectly to the aurf ace of agar plates foc the isolation of Campylobacterjejani from feces.Eur, J. din. Microbiol. 20.,636-640 (1984).

83 Delicyser. P.; Gossuia.Detrain, M.L; liutzler, J.P.; Sternon, J.; Acute enteritis due torelated Vibrio; first positie stool culture. J. infect, Dia. 125;:390-392 (1972).

84 Lausves, S.; Penuer.J.L.: Seroyping Cap~lotacterjeiaai and Canip3/obactcrcohionthe basis of thermoatabte antigens: in Ilntzter, Campylobacter infection in man andanimals. pp. 51-60 (CRC Press, B~oca Raton 1984).

85 Lior, H.;. Serotyping of Cninpylobacterjejuai by slide agglutination based on heat-labileantigenic factors; in Batzlee, Campylobacter infection in man and animals, pp. 61-76(CRC Press, Bosca Raton 1984).

86 WVatson, L.C.; Kerr, 18,3.; Comparison of agglutination, complement-fixation andimmanoflnorescence tests ind Campylobacter icjuai infections. J. Hyg. Camb, 88;165-171 (1982).

87 Svedhem, A.; Lauwers. S.; Kaijser, B.: Diagnostic serology for Camipylobacterjejuniinfections using the Dig.Etisa principle and a comparison with complement fi'.ation test;in Buizier, Campylobacter infection in man and animals, pp. 105-112 (CRC 1ress. Boca

88 Palmer, DF.; Whaley, S.D.; Complement fixation test; in Rose, Friedmen, Fahey,Manual of clinical immunology; 3rd ed. (American Society of Microbiology, Washing.ton 1986).

89 Austin, R.A.; Trust, T.J.; Detection of plasmids in the related group of the genusCumpylobacter. FEMIS Microbial. Lett,.S. 201 (1980).

90 Tompkins, L. S.; Mickelson, P.; McClure, J.;- Use of a DNA probe to detect Campy.lobiacter jelisni in fecal specimens in Caiapylosncter 1f; in Pearson. Skirrosv, Rowe.Davies, Jones Proceedings of the Second International Workshop on Catip)IobacterInfections (PHLS. London 1983).

91 Black., RE.; Jackson, RiJ.; Tasi. T.; Mledevesky. M.; Shayegani. M.; Feeley. J, C.,MacI~eod, K.I.; Wakelee, A.MI.: Epidemic Yerairiia eniierracolifica infection due tocontaminated chocolate milk. New UngI. J. Med. 298:76-79 (1978).

92 Simmonds, S.D.; Noble, MI.A.; Freeman, 11.3.; Gastrointestinal features of ciiltnre.positive Yerasiia enterocolifica infection. Gastroenterology 92: 112-117 (1987).

93 Gutman, L.T.: Yerainia eaterocoliticar and Yerssnia pseiidoaiberiilosis; in Gorbach,Infections diarrhea, pp 65-84 (Blac1.nell, Boston 1986).

94 Bottone, E.J.; Yersinia enrerocotica; in Ellmer, Infectious diariheal diseases, entreatconcepts and laboratory procedures (Dekcer, New York 1984).

95 Bottone, E.J.;Sheehas, D.J.; Yerainia enterocoliticar. Guidelines for serologic diagnosisof human infections. Rev, infect. Dis. 5, 898-906 (1983).

96 Toyos. J.; Diaz, R., Urra, E8.; Mforiyon. IL: Analysis by coaggintination of the

Boedeker/Mec~ueen 214

distribution of a24,000-dalton surface protein in Yersinia isolates. J. cin. Micirobiol. 23:804-805 (1986).

97 Pulkkinen, L.; Granbeth. L.; Grantors, K.: Toivanien. A.: Restricion map of virulenceplasmid in Yersinia enterocolitica 0:3. Plasmid 16:- 225-227 (1986).

98 Trnka, Y. M.; Lamont, J. T.: Clostridiumn difficile colitis. Adv. intern. Mled. 29: 85-107(1984).

99 Pekety, R.; Kim, K.-H.; Brown, D.; Baits, D.H.; Cudmore, M.; Silva, J.: Epidemio-logy of aniibiotic-associated eotitis, Am. J. Med. 70: 906-M0 (1981).

100 Tedesco, FJ.: Antibiotic-associated psendomembranous colitis with negative proc.tosignanidoseopy examination. Gasiroint. Endosc. 77: 295 (1979).

101 Shauholizer, C. .; Peterson, L. R.; Olson, .MN.; Gerdiog, D, N.: P'rospeettve Study ofgram-stained stool smears in diagnosis of Clostridiumi difficids cotitis. J. elmn. Microbial.17: 906-908 (1983).

102 Wilson, K.H.; Silva, 3.; Fekety, P.R.: Fluoresceni-antibody test for detection ofClosirkf lain diffi'cile in stool specimens. J. elmi. Micirebiol. 1&:464-468 (1982).

103 George, W.L.: Suter, V.L.; Citron, D.: Finegold. SM.: Selective and differentialmedium for isolation of Clostridiumz dificile. J. elmn. Microbiol. 9: 214 (1979).

104 Wilson, K.U.: Kennedy, M .;Fekety, .PR,: Use of sodium iaarochotate to enhancespore recoveryon a medium selectixe for C/ostriduanu difficile. 3. elm. Mlicrobial. 15:443(1982).

105 Kim, K.l-. etety, R.; Batts, D. H.; Brown, D.; Cudmore, M.; Silv'aJ.; Waters, D.:Isolation of Clostridnam iffflile from the environment and contacts of patients withaatibiotic-associated ealitis. J. infect. Dis. 143:42 (1981),

106 Gerdioig, D-N.; Olson, M.M.; Petersen, L.R.: Teaseley, DOG.: Gebliard, R,L.;Schwartz. M. L.: Lee. J.T.-: Clossidimn difficile-associaled diarrhea and colitis in adulis.A prospective case-controlled epidemiologic study. Archs intern, Mled. 146:, 95-100(1986).

107 rekety, R.: Antibiotic-associated enterocolitist in Ellmer, Infections diartbeal diseases,Current concepts and laboratory procedures, sal, 10, pp. 77-92 (Dekker, New York1984).

108 Tuiadafilapoutos. G.; Pothoulakis, C.; O'Brien. M.; Lamont, T.,. Differential effectsof Clostridium difficile toxins A and B on rabbit ileuin. Gastroenterology 93: 273-279(1987).

109 Wit. TC., Gerseb, S.: rvaluation of a commercial Lit for the routine detection ofClosiridai di/ffle cytotoxin by tissue culture. 3. cin. Mlicrobial. 23: 792-793 (1986).

110 Nachamtdn, L.; Lote-Nolan. L.; Skalmna, D,: Esaluation of a commercial cytotoxic-ty assay for detection of Clostridaa i/if/ie toxin. J. elmn. Microbial, 23: 954-955(1986).

Ill Peterson, L.R,; Hlolier, 3.3.; Shauholizer, M.T.: Garrett, C.R.; Gerding. DXN:Detection of Clostrimfisus difficite toxins A (enterotosin) and B (cytotonin) in clinicalspecimens. Evaluation of a latex agglutination test. Am. 3. cliu. Path. 86 208-211(1986).

112 Lylerly, D.M.; Wilkins, T.D.. Commercial latex test for Cloardin difficife toxin Adoes not detect toxin A. Am. J. eli, Microbial. 23, 622--623 (1986).

113 Peterson, L.R.; Halter, 3.3.; Shauholizer, CJ3.; Garret, C.R.; Gerding. O.NDetection of Cloairieliio difficife toxin A (euterotoxin). Am. J. elm. Path, 87: 298-299(1987).


114 Welch, D.F.; Me~nge, S.K.; Nlatsen, 3.Nf.: Identification of toxigemic Clostridiumndiffi'cile by eounterimmnsooeectrophoresis 3. clit. Microbiol. 11:470-473(1980),115 Ryan, R.W.; Kwasnik .; TiltonR C.: Rapid detection of Clostridiumn difficile toxin ihuman feces. 3. clin. Microbiol. 12: 776 (1980),

116 Poxton. R,; Byrne, M.D.: Detection of Clostridium difficile toxin by counter-immunoelectrophoresis: a note of eaution, 3. din. Microbiol. 14- 349 (1981).117 West, SI-.; Wilkins. T.D.: Problems associated with ceoterimmunoclextrophoressassays for detecting Clostrldhnn difficile toxin. 3. cin. Microbiol. 15, 347-349 (1982).118 Levine, 1-1G.; Kecnnedy, M.; Lamont. 3.?.: Counterimmunoelcctrophoresis vs.cyltooicity assay for the detection of Clostridiurn diffictle toxin J. infect. Dis. 145:398(1982).

119 Potvliege, C.; Labbe, NI.; Yoorassoxxsky, I.; Gas-liquid chromatography as screeningtest for Clostridhian dif9ie. Lancet U.' I 10O (198 1).120 Sell, . L;Sclsaberg. D). R.; Felsety, F, R.. Iacteriophage and bacteriocto typiugacheinefor Clostridiiim difficile. 3. din. icrobiol. 17: 1148-1152 (1983).121 Tabaqlchli, S.; O'Farrell. S.z Holland, D.; Silman, R.: Method for the typing ofClostridiain difficile based on polyacrylamide gel electrophoresis of [035S~metbionine-labeled proteins. 3. din. Microbiol. 23.:197-198 (1986),

122 Seppala. KC.; Hjclt, I..: Sipponen, R': Colonoscopy it the diagnosis of antibiotic.associated colitis. Scand. 3. Gastroent. 16: 405 (1981),12 Nishibnchl, MI, Ishibashi, NI.1, T1b1eda. Y., Kaper, J,1B,: Detection of the thenistabledirect hemotysin gene and related DNA sequences isn V. jarahionyic and other vibirospecies by the DNA hybridization test. Infect. Immunity 4.-9:481-486 (1986),124 Hlonda, T.; Chearslkul, S.; MlIcda, Y: Nliwataoi, '1', Immunological methods fordetection of Kanagawa phenomenon of I'ibris psarliemnolytcua. 3. clii. Microbic], 11,

6"0-63 (1980).12 Honda, S.: Goto, 1,*: Nliuematsu, L.; Ikeda, N.; Asano. N.:. Ishibashi1. t Kinoshita, Y,;Nishibuchi, NI.; Honda .: Nitwatani, T.: Gastroentcritis in Kanagawa negative Viriopaiaieiolyiicua. Lancet i. 331-332 (1987).126 Popoff, M.: dlrronionas, in K~rieg, Holt, Blergey's manual of systematic bacteriology,vol. I, pp. 54S.550 (1984),127 I'itaraugsi, C.; Echeverria, P.; Whlitmire, R.; Tirapat. C.; Formal, S.; Danimin, G.J.:Tingtatapong. NI.: Euteropathogenicity of Aeroioinas hydrophila and Plrisioiiionas

aliigelloides. Prevalence among individuals with and without diarrhea in Thlailansd,Infect. Immunity 35,666-673 (1982).128 Gracey, M.; Blurk~e, V.; Robinson, 3. Aeroiosiaassociated gasttoenteins. Lancet ii,

304-306 (1982).129 Popoff, NI.Y,; Coynault, C.; OKireityan, NI.; Lemelin, NI.: Polynucleotide relatednessamoag motile Aroionna species. Cut. Nlterobiol. 5., 109-114 (1981).130 Nfegraud. R.: Incidence and virulence of Aeroinonna species in fecees of children withdiarrhea. Eur. J. din. Nlicrobiol. S: 311-316 (1986).131 Nillership, S. .; Barer, MI F.; Tabaqchali, S.: Toxin production by Aeluvonas app.from different sources, 3. med. Nicrobiol. 22: 3'1!-314 (1986).132 Kaper. 3.; Seidler, R.J.; Lockman, H-.; Colwecll, R. R., A medium for the puesumptiseidentification of Aerosnonna Iiydophila and Iinterobactrriacea. AppI. eisvir. tcrbo

38: 1023-1026 (1979).133 Ljungh, A.; NVadstrom, T.; Aeromonas toxins. Pharmacol. Thee, 15: 339-354 (19S2)

Bctdeker/McQueen 2 16

134 Donia, S., Haddow, A.D.: Cytotoxie activity of Aeromonas Jsidrophitia. Infect.Immunity 21; 989-993 (1978)

135 Annapurna. E.; Sanyal, S.C.: Enterotoxigenitcity of Aeromonas hzsdrophiia. J. Mfed.Microbiol. 10,317-323 (1977).

136 Agger. %VA.; McCormick, J.D.; Gurwith. M.J.: Clinical and microbiological fea.tures of Aesomonaa Iiydrophla associated diarrhea. J. elmn. Microbiol. 21: 909-913(1985).

137 Robinson, 1.; Ileamon, J.z Wagener, L.; Burke, V.: Comparison of direct plating withthe use of enrichment culture for isotation of Acronionas spp. from feces. J. Mfed.Microbial. 22:315-317 (1986).

138 Figora, N.*: Maint. L.; Verdiant, S.; Ceechrii, C.; liarbert, A.: Prevalence, speciesdifferentiation and to'sigenicity of Aerosnonas strains in cases of chitdhood gastroen.teritis and in controls.15 cai. MIicrobiol. 23:595-599(1986).

139 Hotherg, S.D.; Schell, W.L.; Fanning. G.S.; Wachsinuth, 1.K.; Ifickmann, Blrenner,P. W.; Blake. P.WX.: Birenner, D.i., l'armer.J.J.: Aeromonas intestinal infections in theVisited States. Ann. intern. Med. 105: 683-4689 (1986).

140 Kay. B.A.: Guerrero, C.E.: Sack, R.l1.: Media for the isolation of AesoosnasIiydrophila. J. clia. Microbial. 22: 888-890 (1985).

141 Morgan. D. R.; Johnson. P. C.: DuPont. H.L.; Satterwvhite.T. K.; Wood, L. V.: Lack ofcorrelation between known virulence properties of Aerosuonas Iiydrsphifa andenteropathogenicity in humans, Infect. Immunity 59: 62-65 (1985).

142 Ciskor, G,; Illacklow, N. R.: Hunman viral gastroenteritis, Microbial. Rev. 48: 157-179(1984).

143 Dingle, SkI.; Badger, G.F.; Felter, A.1i.; Hodges. R.G.; Jordan, \V.S.; Rammel.kamp, C. H.: A study of itlness in a group of Clevelaud families, 1. Plan of study andcertain general observations. Am. J. Hyg. 58.,16-30 (1953).

144 Blacklow, N.R.; Cukor. G.: Viral gastroenteritis. New Engl. J. Med. 304: 397-406(1981).

145 Nakuta, S.; Estes, MNIX; Graham. D.Y.; Loosie, R.; Tao, HI.; Shusheng. W.; Saif,L.S.; Mecluick,S. L.: Antigeniecharactcriza'son and IiLISA detection of adult diarrhearotaviruses, J. infect. Dis. 154: 44&-455 (1986).

146 tlhnoo, 1L; Otding-Stenkvist, B.; Wadell, G.: Update of viral gastroenterilis. Dig. Dis.6: 102-116 (1985),

147 Flares. J,.: Boeggeman, E.; Purcell, RH.; Sereno, IM,; Perez. IL; W~hite, L.; Wayatt.R,G,; Chanock, R.M., Kapikian. A.Z.: A dot hybridication assay for detection ofrotavirus, Lancet it.555-559 (1883).

148 Graubalte, P.C.; Vestergaard, BAP; Meyliog, A.; Genner, J.: Optimized enzyme.linked immunosorbent assay for detection of human aiid bovine rotavirus in stools.Comparison with electron-microscopy, immunos'lectroosmophoresis. and fluorescentantibody techniques. J, med. Virol. 7. 29-40 (1981).

149 Yolken, R. IL; Wyatt. R. G.; Zissis, G.; Brandt, C. D.; Rodriguez. W. J.; Kim. 11.XV.,Patrotn, RHI.; Uirrutia, . J.., Main, L.; Greenberg, 11.; - Kapi kian, A. Z.; Chanock.R.M.: Epidemiology of human rotavirus types I and 2 as studied by eoeynse-liakedimmunosorbent assay. New Eingl. S. Mfed. 299:1156-1161 (1978).

150 Yolken, R.H.; Barbour, B.; Wyatt, R. G.; Kalica, A. R.; Kopikian, A. Z., Chanock,R.M.: Enzyme-linked immunosorbent assay for identification of rotairuses fromdifferent animal species, Science 201. *259-262 (1978).


151 Yolken, R.H.; Eiden, J.; Leister. FR: Self-rontained enzymie membrane immurro.assay for detection of rotavirus antigen in clinical samples. Lancet ii. 1305-1307(1986).

152 Tao, H.; Gnangmu, C.; Chiangan, W.; lHmnt, Y.; Zhaoying. F.: Timrgxin, C.: Zinyi, C.,Weiwe, Y.; Xuejian, C.; Shuasen, D.; Ziaoquang, L.; Weicheng, C.; Waterborneoutbreak of rotavirus diarrhoea in adults in China caused by a novel rotavirus. Lancet i:1139-1142 (1984).

153 Yok.en, R. If.; Leister, F.J.: Evaluation of enzyme immnnoassays for the detection ofhuman rotavirus. J. infect. Din. 144:379 (1981).

154 Kapikian, A.Z.; Wyatt, RG.; Dolin, R.; Thornhill, T.S.; Katica, A.R.; Chanock.R.M.: Visnalization by immune electron microscopy of a 27-nm particle associated withacnte infectious nonbacterial gastroenteritis. J. Virol. 10. 1075-1081 (1972).

155 Kapikian, AZ.; Greenberg. 1.LB.; Wyatt. R, G.; Kalica, A.R.; Channel., R. .: TheNorwalk. gronp of viruses - agents associated with epidemic viral gastboenteritis; inTyrrell, Kapikian, Virus infections of the gastrointestinal tract. Pp. 147-177 (Dekker,New York 1902).

156 ltlaclklow, N. R.; Csikor. G.: Virla gastroenteriis agents; in Lenneite, Italows, Ilausler,Truant, Manual of clinical microbiology: 3rd ed. p. 8914898. (American Society forMicrobiology, Washington 1980).

157 Greenberg, H-.B.; Wyatt. R.G.; Valdesuso, J.; Kalica. A.R.; 1.ondond. W.T.;Chanock, R.M.: Kapikian, A.Z.; Solid-phase microliter radinimmunoassay for detec-tion of the Norwalk strain of acute nonbacterial epidemic gastroenteritis virus and itsantibodies. J. med. Virol. 2:97-1089(1978).

158 Tretnor,1.,; Madore, 11. P.; Dotin, R.: Development of a monoclonal antibody to theSnow Mountain agent of gastroenieritis (absir. 109). Program and Abstracts of the 26thInteesience Conference on Antimicrobial Agents arid Chemotherapy (AmericanSociety for Microbiology, Washington 1986).

159 WVadell, G.: Molecular epidemiology of human adenostruses. in Doerfler, ed. Themolecular biology of adenoviruses. 2. 30 years of adenovirus research 1953-1983.Current topics in micobiology and immunology, vol. 110, pp. 191-220 (Springer, Berlin1984).

160 Uhno., L; Wadell, G.,, Svensson, W. Johausson, M1.: Importance of eoterieadenoviruses 40 and 41 is acute gastroenteriris in infants and young children. 3. clin.Microbiol. 20: 365-372 (1984).

161 do Jong,J. C.; WVrgand, R.; Kidd, A.I.; Wadell. G.; Kapsenbeig.J. G.; Marerie, C.J..;Wermenbol, A.G.; Firtzlaff, R..G,; Candidate adenoviruses 40 and 41. Fastidiousadenoviruses from human infant stool. J. med. Virol. 11: 215-231 (1983),

162 Hlerrmann, I.E.; Perron-Henry, D.M.; Itlacklow, N.R.: Antigen detection withmonoclonal antibodies for the diagnosis of adenoviruses gastroenteritis. J. infect. Dii.155, 1167-1171 (1987).

163 Haiva, V,.4.; H-yyrynen, 0.; Manainen, T.; Pale.. K.; The l7th Congress of theIAMLT, Stoclholm 1986. Pp. 149.

164 Dolin, R.; Treanor, J. J.; Madore, I.P.: Novel agents of viral enteritis in humans J.infect. Dis. 155: 365-376 (1987).

165 Cubit, W.D.; Barrett, A. D.: Propagation of human candidate calicivirus in cellculture. J. gen. Virol. 65: 1123-1126 (1984).

166 Naksla, S., Chiba, S.; Terashima, H., Sakuma, Y.; Kogasaka, R.; Nakao, T.: Microtiter


solid-phase radioimmunoassay for detection of homan caticivirus in stools. J. chn.Microbiol. 17.,198-201 (1983).

167 Lee, T. W.: Kurtz, J. B.; Serial propagation of astrovirus in tissue culture with the aid oftrypsin. J. gen. Viral. 57: 421-424 (1981).

168 Walsh, J.: Probtems in the recognition and diagnosis of amebiasis. Estimation of theglobal magnitude of morbidity and mortality. Rev, infect. Dii. 8:228-238 (1986),

169 Ridley. D. S.; Hawgood, B.C.: The value of formal-etlier concentrations of feral cystsand ova, J. cin. Path. 9: 74-76 (1956),

170 Root, D. M.; Cole, F.X.; Williamson. J.: The development and standardization of anELISA test for the detection of E. hiarolyiica antigen in fecal specimens. Arebs invest.Med. 9: snppt. 1, pp. 203-210 (1987).

171 Sargeant, P.G.; Williams, J.E.,; Greene, J. D.: The differentiation of invasive and non.invasive E. hisrolyiica by isoennyme electrophoresis. Trans. R. Soc. trap. Mfed. 1-yg. 72:519-521 (1987).

172 Krogstend, D.: Isoeneyme patterns and pathogenicity in amoebic infections. New Engl.J, Mfed. 31S: 390-391 (1986).

173 Diamond, L. S.; Harlow, D. R.; Cunniek, C. C.: A new mcdinm for axenir cultivationof E. /iistolytica and other Entnmoelia. Trans. R, Soc. trap. Mfed. 1-yg. 72: 431-432(1978).

174 Mirelman, D.; Brach, R.; Chalken, A.; Aust-lKettis. A.; Diamond, L.S.: E. hirio1lslica:effect of growth conditions and bacterial association on iscozyme patterns andvirulence. ExpI. Parasit. 62: 142-148 (1986).

175 Strachen, W.D.; Chiodini, P.L,; Spice, W.M.; Moody, Al-I.: Ackers, 3.P.:Immunological differentiation of pathogenic ni non-pathogenic isolates of Euranoebohiarolytica. Lancet k: 561-562 (1988).

176 Rardin. J.- Pathogenesis of disease cansed by . Iiistolyfica: studies of adherence,secreted toxins and contact dependent cytolysis. Rev infect. Dis. 8, 247-260 (1986).

177 Jolsipl, A.M.M.; Johipi, L.- Prepatency of giardiasis. Laurel i: 1095-1097 (1977).178 Lee. B.; Ward, H.; Kenich, G.T.; Pereira, N.M.E.A.: Lectin activation in Giarifin

lainbia by host proteases arouses host-parasite interactions. Science 232: 71-73 (1986).179 IKamath, K.R.; Mutugati. R.; A comparative study of four methods for detecting G.

lamblia in children with dinrrheal disease and malabsorption. Gastroenterology 66,16-21 (1974).

180 Craft, J.C.; Nelson, J.D.. Diagnosis of giardiasis by counterimmunoelectrophorcsis offeces. J. infect. Dis. 145: 499-504 (1982).

181 Ungar, B.O.P.; Yalken, R.1l.; Nash, T.E.; Quinn. T.C.: Enzyme-linked immune.sotbent assay for the detection of Ginrdia lasrblia in feral specimens. J. infect. Dis. 149:90-97 (1984).

182 Goha. A.K.-, Inge, R..; Farihing, N.J.G.: Immunoassay of Giardia antigen fordiagnosis of giardiasis. Gut 25: 1162 (1984).

183 Rosoff, J.D.; Stibbs. H.H.: Isolation and identification of n girdin lamblin-specificstool antigen (GSA 65) useful in coprodiagoosis of giargiasis. J. ella. Microbial. 231905-910 (1986).

184 Goha. A.K.; Ralston, D.D.K.; Mfathan, V.I.; Farihing, NI.S.I.: Diagnosis ofgiardiasis by IgM antibody enzyme-linked immunosorbent assay Lancei it, 184-186(1986).

185 Smith. P.D.; Gillin, F.D.; Brown, W.R., Nash, T.18.: lgG antibodies to Ginrdia

Laboratory Investigation of Cbstdhcxd E3nteric infectionts 219

Ianrbha detected by enzyme-linked immunosorbent assay. Gastroenterology 89:,1476-1480 (1981).

186 Nash. T.E.: Herrington, D.A.: LOSOusky. G.A., L.evine, N1,Isl: Iixpermentalliumartinfections wvitlt Giardia lainb/ia. 3. infect. Dis. 156: 974-984 (1987)187 payer. R.; Ungar. 3. L P.: Cryptospori'diaon pp. anidcryptospantdiosis. Mficrobiol. Ref.50: 458-483 (1986).

188 Tzipori, S.: Cryptosportda in animals and humans. Microbial. Rev. 42, 84-96 (1983),189 Current. S.L.: Hzaynes, T.1B.; 12it,i.V.. Blailey. W.S.; H-eyman, 13.13.: Weintraub,997*M.: H-umant cryptosporiosis in immunocompeomised and tmmunoideficientpatients. Studies of an outbreak and experimental transmission. Newrng). J. Med. 308:1252-1257 (1903).

190 Ma. P.: Soave. R.: Three stcpstool exatnination foreryptosporidiosas tn ten homosexualmen %ith protracted watery diarrhea. J, infect. Dis. 824:2S8(1983).191 Stibibs. RII:Ongerth. 3.13.: lnrmunfluorescricr detectinn of cryptosporidiom

oorysts in ferlsmears,. . in. MIicrobiol,24: 517-521 (1986).1921 McLaughlin, J,: Casenioxe. D.P.. Mamamoits. T.G.: Gerson. P.J,: Samuel, D.; Taylor.A.G0. Identification of cryptosporidinni oocytes by monoclonal antibody. Lancet i: I(1987),193 Current. W.L,: H-aynes. T.13.- Complete dc'elopmeiit of cryptospotidisis an cellculture. Science 224: 603-605 (19S7).194 Campbell. P'.N.,. Current. W.L,: Demoinstration of serums antibsodies to crypto.sporisliosis in normal anrd immanodefirienat humtants kwith coitfirmed infection, 3. clii.Microbial. 1&:165-169 (1983).195 Ungar, 13.L.". Soame. RI.:. Fa~er. R,: Nash. T.: Enzyme immunoassay detection ofinmunoglobultin Mi and G antibodies to cryptosporidisim in immunorompetent andiminunomproimised persons., J.infect. Dis, 153: 570-578 (19$6).196 Ungar . 3L: . ilman. R. ff.: Lanata. C. F.: lieres.Scbacl. I.: Seroepidemioogy ofcrypiospnridiin infection it too Latin American popalatins. J. infect. Dis. 157:-

S51-556 (1988).

Edgar C. Baoedeker, MD, Department of Gastroenterology,WValter Reed Army Institute of Research. Washington. DC 20307-5100 (USA)

frontiers in gastrointestinal research reprint ad … · frontiers in gastrointestinal research...

Documents