Gary W. Procop MD
Franklin Cockerill III MD
Essentials of Diagnosis
The Enterobacteriaceae are a diverse family of bacteria that, in nature, exist in soil, on plant material, and in the intestines of humans and other animals. Another ecological niche in which these organisms thrive is the hospital. Many of these organisms cause a wide variety of extraintestinal diseases that are often nosocomial and commonly present in debilitated or immunocompromised hosts. These manifestations have been examined previously in this volume and include urinary tract infections (Chapter 16), skin and soft tissue infections (Chapter 13), lower respiratory infections and pneumonia (Chapter 10), and infective endocarditis and sepsis syndrome (chapters 11 and 17). The genera most frequently associated with extraintestinal disease are Escherichia, Enterobacter, Klebsiella, Proteus, Citrobacter, and Serratia..
Some members of the Enterobacteriaceae, however, cause primarily enteric disease or enteric-associated systemic disease (Box 53-1). These are Shigella and Salmonella species and particular strains of E coli and Yersinia enterocolitica. This chapter is devoted to enteritis and enteric-associated systemic disease caused by diarrheagenic strains of E coli and of Shigella and Salmonella species. Enteritis and associated mesenteric lymphadenitis caused by Y enterocolitica are covered separately (see Chapter 60).
The family Enterobacteriaceae consists of at least 27 definitive genera and numerous enteric groups. The final classification of the enteric groups has yet to be resolved. Although there is diversity within this group, members have several common, characteristic features. All Enterobacteriaceae are gram-negative, facultative anaerobic bacilli that have the ability to reduce nitrate to nitrite. They produce catalase, ferment D-glucose with or without gas production, and do not produce oxidase. This latter characteristic is useful in differentiating this group of organisms from other nosocomial gram-negative pathogens, such as Pseudomonas aeruginosa, which is oxidase positive.
Traditionally, phenotypic characterization has been the method of choice for differentiating the members of the Enterobacteriaceae, but, with increasing frequency, genotypic methods are being used to detect particular genera and pathogenic varieties. Phenotypic methods, however, remain the standard for routine identification of these organisms in the clinical microbiology laboratory. The characteristics that are most useful for identification include colony morphology (particularly distinctive for some Proteus strains), indole production from tryptophan, reactions of the methyl red and Voges-Proskauer tests, motility, the ability to produce hydrogen sulfide, and the ability to use various organic substances. Several pathogenic Enterobacteriaceae, particularly many of the diarrheagenic E coli, fail to demonstrate unique phenotypic characteristics and continue to pose a challenge for microbiologists.
Essentials of Diagnosis
BOX 53-1 Enterobacteriaceae Syndromes
In a normal host, exposure to ETEC leads to the development of mucosal immunity, presumably secondary to secretory immunoglobulin A (IgA) antibodies directed toward the fimbriae-associated colonization factor antigens (CFAs). In endemic areas, natural resistance is usually well developed in adults. These individuals serve as carriers and shed large numbers of toxigenic E coli into the environment. Disease occurs in those lacking mucosal immunity, principally neonates and travelers from nonendemic areas.
The EAggEC are also a cause of diarrheal disease in both developing and developed countries and have caused serious outbreaks in nurseries. Infants are most commonly affected, and growth retardation may be caused by persistent diarrhea. In the United States, EAggEC have been shown to cause persistent diarrhea in patients infected by the human immunodeficiency virus (HIV).
Although not extensively studied, the DAEC seem to infect children > 1 year of age, rather than neonates.
Many of the diarrheagenic E coli, namely the ETEC, EPEC, EAggEC, DAEC, and non-O157:H7 EHEC, demonstrate colony morphology and biochemical reactions in standard tests (biophysical profile) that are identical to normal-flora E coli. For this reason, the identification of these organisms is a problem for clinical microbiologists. If these diarrheagenic varieties are not suspected clinically and additional testing is not performed, they will be erroneously dismissed as normal flora. Fortunately, in most instances, enteritis caused by these agents is self-limited and resolves without antimicrobial therapy. The most commonly encountered EHEC strain, E coli O157:H7, and the EIEC have distinctive features that may be used to screen for these organisms.
Animal models and cell culture assays have been used to detect the heat-labile (LT) and heat-stable (ST) enterotoxins, but those methods have largely been replaced by immunoassays. Commercial immunoassay kits are now available for both the LT and ST toxins. Molecular assays, designed for the detection of the genes that encode for these toxins, are another method of detection. Signal amplification assays with nucleic acid probes have been used directly on stool and on colony blots of suspect isolates. Polymerase chain reaction (PCR)-based assays have also been used to detect these genes. If PCR is attempted directly on stool specimens, a preamplification treatment may be necessary to remove PCR inhibitors that are inherently present in stool. Better results may be obtained if PCR is performed on isolated colonies, rather than stool.
In South America and Australia, the non-O157:H7 EHEC serotypes cause more hemorrhagic colitis (HC) and hemolytic-uremic syndrome (HUS) than the O157:H7 serotype. These serotypes are more difficult to detect, because they are able to ferment sorbitol and are therefore phenotypically similar to normal-flora E coli on the SMAC screening agar. The identification of these strains is more complicated and relies on methods to detect the shigalike toxin or verotoxin (VT), the VT gene, or other markers of pathogenicity.
Cultured Vero or HeLa cells may be used to detect the presence of VT, but this test is costly and labor intensive, and it has largely been replaced by immunoassays. A wide variety of molecular assays have been developed for the detection of EHEC. These methods include PCR-based assays and nucleic acid probes for the detection of the VT gene or other markers of pathogenicity, such as the pO157 plasmid and the eae gene encoding intimin, which is necessary for attachment to enterocytes.
Although these detection methods are not currently used in most clinical microbiology laboratories, they may be more widely used in the future. Many of these methods offer greater sensitivity than the SMAC agar, and they detect non-O157:H7 serotypes. However, problems with these technologies do exist. Substances present in stool may inhibit PCR. Some non-nucleic acid amplification methods, such as the measurement of free fecal cytotoxin or culture enhancement by using O157 immunomagnetic beads, may have sensitivities that are superior to PCR-based assays. Determining the significance of the detection of either VT or the VT gene is another problem. The presence of VT or the VT gene is supportive evidence, but not definitive proof, that the isolate is the cause of disease. There are > 200 serotypes of E coli that can express VT and only a fraction of these have been associated with HC. The serotypes O26:H11, O111:H-, O103:H2, and O113:H21 are other EHEC that have been associated with HC and/or HUS outbreaks. Therefore, when VT or the VT gene is detected in stool or stool isolates, the detection of other virulence factors or particular O:H serotypes would be necessary to further characterize the isolate.
The principal disadvantage of most molecular detection technologies is cost. In many instances, an expensive screening test is not cost effective, especially when the pathogen is infrequently encountered and less expensive screening methods exist. These technologies, however, are attractive in outbreak situations or when the incidence of disease is high.
The stool from patients with EIEC enteritis is indistinguishable from the stool of patients with shigellosis and may be watery or dysenteric. Shigella antigen testing is of no use in differentiating these bacteria, because EIEC may cross-react with Shigella antiserum. In a similar manner, tests of invasiveness, in cell culture or animal model (Sereny test), are positive with both EIEC and Shigella spp. Molecular methods, such as enzyme-linked immunosorbent assays, nucleic acid probes, or PCR, have been used for the detection of Shigella/EIEC species, but they are less useful for differentiation.
The biophysical profiles of the enteroadherent E coli subtypes are indistinguishable from normal-flora E coli in routine bacteriologic studies. These varieties, which do represent distinct pathogenic subtypes, adhere differently from one another to HEp-2 cells in cell culture. The EPEC adhere to the HEp-2 monolayer in a localized pattern, the EAggEC produce three-dimensional aggregates, and the DAEC adhere diffusely. Although these adherence patterns are reliably produced, the HEp-2 assay requires cell culture capability and interpretation experience.
The EPEC can also be identified by the production of an attaching and effacing lesion (A/E) and absence of verotoxin (VT) (EHEC produce the A/E lesion, but also produce VT). If the phenotype is used for identification, it is important to demonstrate the absence of VT, because EHEC also produces an A/E lesion. Like the other diarrheagenic E coli, the EPEC have been associated with particular somatic antigens. Many of the EPEC possess specific O:H profiles that may be used for organism identification. Antiserum directed against some of these O antigens is commercially available and may be used to screen suspect colonies. A positive screening test then requires confirmation, which may include titration, complete O:H typing, or molecular analysis (see below).
Both direct DNA probes and PCR-based assays may be used for the detection of genes that encode pathogenic factors of the enteroadherent E coli. The presence of the eae gene, which is associated with the A/E phenotype, and the EAF plasmid, which contains the bfp gene cluster, are important for pathogenesis of EPEC; detection of these genes and the absence of the VT gene are important for molecular identification of the EPEC. Molecular detection of EAggEC may be accomplished by identifying a 65-MDa plasmid by either DNA probe or PCR technology. The molecular detection of the DAEC has been accomplished through the detection of a particular fimbria (F1845)-associated gene, but false positive reactions may occur.
All of the ancillary tests, both phenotypic and genotypic, used to identify the enteroadherent E coli subtypes are impractical and cost prohibitive for most clinical microbiology laboratories. These tests are most efficiently and cost-effectively performed by reference or public health laboratories that have expertise in identifying these organisms. In most instances, however, the complete identification of the enteroadherent E coli subtype is unwarranted, because most patients have a self-limited course with appropriate supportive therapy. If diarrhea from an infectious etiology becomes persistent, antimicrobial therapy is warranted and should be based on the organism's particular antimicrobial susceptibility profile.
One of the E coli ST toxins, STa, results in osmotic diarrhea, similar to the LT toxin, but through activation of guanylate cyclase and increased cytologic cyclic GMP (cGMP) levels. The other ST toxin, STb, does not cause alterations in cAMP or cGMP levels, but increases intracellular calcium, promotes the secretion of bicarbonate, and stimulates serotonin and prostaglandin E2 release. In addition, unlike the other toxins, STb damages enterocytes and causes epithelial cell loss and partial villous atrophy. These changes also contribute to diminished absorption and osmotic diarrhea.
The EHEC are a subset of E coli that produce disease through the combination of a variety of virulence factors. The most important virulence factor is a lysogenic bacteriophage-encoded toxin that is typical of Shigella dysenteriae type 1. This shiga toxin is also known as a VT because of its toxicity to Vero cells. VT binds to G3b, a glycolipid receptor. After binding, a portion of the VT enters the cell and disrupts protein synthesis by enzymatically altering the 28S ribosomal subunit. The high concentration of G3b on intestinal villous tip cells and renal endothelial cells may partially explain the damage to the intestine and kidneys in HC and HUS, respectively.
Although important, the presence of VT alone is probably insufficient for the production of disease, because there are VT-producing E coli, which do not produce HC/HUS. Another virulence factor of the EHEC is a chromosomal 35-kilobase locus of enterocyte effacement (LEE), as also found in the EPEC. This locus contains the eae gene, which codes for an outer membrane protein, an intimin, that mediates adherence between the EHEC and the enterocyte. This locus also confers the A/E phenotype that is typical of EHEC and EPEC. Finally, plasmid-encoded hemolysins are also probably virulence factors.
The EPEC, like the EHEC, contain a chromosomal LEE gene cluster. This locus includes the eae gene, which encodes for an intimin outer membrane protein (OMP). This OMP mediates adherence between the bacterium and the enterocyte. The LEE locus also contains the genes responsible for the attaching and effacing lesion (A/E lesion). After attaching, the bacterium causes changes that result in the effacement of the microvilli of the cell membrane, which can only be fully appreciated when viewed by electron microscopy. However, a portion of the A/E lesion, the aggregated intracytoplasmic, filamentous actin, may be detected in cell culture or intestinal biopsy by using immunohistochemical stains. Some strains may produce an enterotoxin and may contain a cell entry gene product. The precise relationship, however, between these products and the production of disease remains to be determined.
The EAggEC adhere to enterocytes via a fimbria designated aggregative adherence fimbria I or AAF/I, a 38-kDa OMP. Another fimbria (AAF/II) has also been implicated in cytoadherence. It is currently thought that, after colonization, the EAggEC promote enhanced mucus secretion. This mucus forms a protective biofilm, which contains the EAggEC and may diminish nutrient absorption. Finally, the production of enterotoxin may cause enterocyte damage and diarrhea.
Enterocyte surface adherence by the DAEC probably also occurs through the fimbriae (F1845) and/or outer membrane proteins. The precise mechanism of disease, however, remains to be determined.
The diarrhea produced by these organisms is variable and may be watery, mucoid, dysenteric, or bloody, and it may be associated with serious and even fatal systemic sequelae. The presentation and course of disease are largely dependent on the infecting E coli subtype, as well as the age and nutritional status of the patient. Infants and children, especially if malnourished, are particularly susceptible to dehydration and may succumb rapidly. The mechanism of disease production also varies with the E coli subtype. The enteritis caused by these organisms results from toxin production, enterocyte invasion, intimate bacterial/enterocyte adherence, or a combination of these mechanisms.
Patients with ETEC enteritis usually have an abrupt onset of watery diarrhea that does not contain blood, pus, or mucus (ie, is nondysenteric). The diarrhea is usually mild to moderate in severity, but some patients may have severe fluid loss, like that seen in patients with cholera. A low-grade fever, nausea, and abdominal pain may also be present. Dehydration may become severe and life threatening in neonates and children, necessitating aggressive fluid and electrolyte replacement. A self-limited course, with resolution in 2–5 days, is most common in adult travelers who acquire the disease.
Disease caused by the EHEC usually follows the ingestion of contaminated food or beverage. The incubation phase averages 3–4 days, with a range from 1 to 8 days. Individuals remain asymptomatic during the incubation phase. Early in the course of disease, patients develop watery diarrhea that is usually not bloody. Accompanying symptoms may include nausea and vomiting, abdominal cramping, and a low-grade fever. The diarrhea may then become noticeably bloody within a few days. It is interesting that fecal leukocytes are characteristically few and are detected in only ~ 50% of patient's stools. In most patients, the disease is self-limited. However, < 10% of children and a lesser number of adults may develop HUS.
HUS is a severe systemic disease with a significant mortality. It consists of the triad of microangiopathic hemolytic anemia, renal failure, and a thrombocytopenia that may be part of a consumptive coagulopathy. The kidneys are particularly susceptible to damage (see below). In general, the kidneys may have a “flea-bitten” appearance secondary to punctate cortical hemorrhages that result from multifocal occlusion of afferent arterioles. Biopsies demonstrate microvascular deposition of immunoglobulins, complement components, and fibrin by immunofluorescence. Arteriolar and intimal hyperplasia and subintimal fibrin deposits may be seen in histologic sections. Additional findings may include microinfarcts, acute tubular necrosis, and interstitial edema. Renal failure commonly develops, with resultant oliguria, azotemia, and hematuria. Patients who survive HUS suffer morbidity caused by central nervous system and renal sequelae.
Volunteer studies and occasional, well-documented outbreaks have helped to establish the EIEC as a cause of enteritis. Patients infected by EIEC have moderate-to-severe diarrhea that begins watery, but may become dysenteric with sheets of leukocytes, blood, and mucus. The watery diarrhea is similar to that produced by ETEC infection, and the dysentery is indistinguishable from that produced by Shigella species infection. Fever and abdominal cramping are also frequently present.
The presentation and course of disease caused by the enteroadherent E coli (EPEC, EAggEC, and DAEC) are variable and depend to some degree on the infecting E coli subtype. The EPEC primarily cause acute, profuse, watery diarrhea, which rarely may become persistent. Stools are typically not bloody, mucoid, or dysenteric. Low-grade fever with nausea and vomiting may be present. Microscopic examination of the stool may disclose rare fecal leukocytes.
The EAggEC produce an acute, secretory diarrhea that is usually watery to mucoid and may also become prolonged. In some instances, gross blood may be present. A low-grade fever is common, but vomiting is infrequent.
The DAEC seem to produce a watery diarrhea, usually without blood or fecal leukocytes, but too few studies have been performed to adequately characterize this disease.
The differential diagnosis of diarrheal diseases is extensive and includes infectious etiologies, inflammatory-bowel disease, irritable-bowel syndrome, postsurgical dumping syndromes, and even some neoplasms, such as hypersecretory villous adenomas or vasointestinal peptide-producing neuroendocrine tumors. This extensive differential is successfully narrowed through the examination of the patient (history and physical exam), laboratory studies, and often endoscopy with biopsy.
The physical examination of the patient, as well as exposure/travel history, past medical history, and the duration and presentation of current disease, yields clues to the cause of the diarrhea.
Laboratory tests, radiologic studies, and special procedures, such as endoscopy with biopsy are useful in delineating the cause of disease. A colonoscopy with biopsy is often necessary to determine the cause of persistent diarrhea, particularly when microbiologic studies are negative. Inflammatory bowel disease, ischemic colitis, lymphocytic/collagenous colitis, and neoplasia are often suspected clinically and confirmed with histopathologic studies.
The stool of patients with fever and new-onset diarrhea should be cultured for common bacterial enteric pathogens. Fecal leukocyte testing is not useful, because fecal leukocytes may be present in the stool of patients who have enteritis caused by a wide variety of pathogens, as well as in the stool of patients who have noninfectious enteritis. Clinical microbiologists use selective and differential agar media to screen the stool for Salmonella, Shigella, and Campylobacter species and for E coli O157:H7. In most laboratories, cost-effective screening for E coli O157:H7 is performed only when patients have bloody stools or a history of bloody stools. In a similar manner, the culture for even rarer bacterial enteric pathogens, such as Y enterocolitica and Vibrio species, is most often performed upon request and usually after exclusion of more typical pathogens. If bacterial agents that resemble normal flora on routine bacterial media are suspected (ETEC, EPEC, EAggEC, DAEC, or EHEC other than O157:H7), the laboratory should be notified so that confirmatory testing may be performed.
Other methods used to detect infectious agents in the stool include the direct examination, immunofluorescence enzyme-linked immunosorbent assays and cell culture-based assays. The microscopic examination of stool, often by using special stains (ie, trichrome or modified acid-fast stain), is used to detect ova and parasites. Immunoassays and ELISAs are available for the detection of Giardia lamblia, Cryptosporidium parvum, rotavirus, and Clostridium difficile toxins. Some laboratories use cell culture-based assays to detect C difficile toxin. Similar assays may be used to detect the VT of EHEC and shiga toxin of Shigella dysenteriae 1.
Individually, these microbiologic assays may be relatively inexpensive, but when numerous tests are ordered nonjudiciously, costs rapidly accrue. Microbiologic assays, like any laboratory assay, should be used to answer specific clinical questions. In patients with enteritis, the most likely causes of disease should be explored first. Patients who develop diarrhea after 3 d of hospitalization and who have been treated with antimicrobial agents probably do not warrant bacterial stool cultures or a stool examination for ova and parasites. These patients should be tested for C difficile toxin, because they are more likely to have pseudomembranous colitis.
All of the diarrheagenic E coli may produce dehydration and electrolyte abnormalities. The EAggEC may cause chronic disease and EIEC, like Shigella species, may cause a protein-losing enteropathy if dysentery is produced. Disease secondary to these agents is usually most pronounced in children, who may become dehydrated rapidly.
The most severe complication of EHEC infection is the development of HUS. This devastating disease most commonly occurs in children and has a significant mortality rate (3%–10%). Patients develop microangiopathic hemolytic anemia, thrombocytopenia, and renal failure. Patients that do not succumb may suffer significant morbidity secondary to renal and central nervous system dysfunction.
Enteritis caused by the ETEC, EPEC, EIEC, DAEC, and EAggEC is usually self-limited and may be associated with travel or a particular outbreak. In most instances, the presence of one of these organisms is assumed, and the diagnosis is based on history, physical examination, and the exclusion of other enteric pathogens. A definitive diagnosis cannot be made, unless specific assays are performed. These confirmatory assays are costly and usually clinically unnecessary. In outbreak situations, however, complete microbiologic characterization of clinical isolates, with strain analysis, may be warranted for public health purposes.
The severe sequelae that may occur secondary to EHEC infections necessitate the identification of suspicious isolates on SMAC screening agar. The diagnosis of EHEC is strongly associated with bloody diarrhea or a history of bloody diarrhea and is often associated with the ingestion of undercooked beef. The presence of E coli O157:H7 is determined by culture and serotyping.
Treatment of fluid and electrolyte loss is usually achieved through oral rehydration. The use of the World Health Organization Oral Rehydration Salts (ORS) solution is recommended. Intravenous rehydration may be necessary for infants, individuals with excessive vomiting, or those with severe dehydration.
Bismuth subsalicylate, 1 oz of liquid or two (262.5-mg) tablets taken every 30 min for 4 h, may decrease the amount of diarrhea and the duration of disease. Antimicrobial therapy is generally not indicated, because of the self-limited nature of most of these diseases. Some contend that empiric therapy may decrease symptomatology and shorten the clinical course. It is a concern, however, that the overuse of antimicrobial agents will promote resistant strains. If chronic or persistent diarrhea develops in patients infected by one of the enteroadherent E coli strains, specific antimicrobial therapy should be used. If available, the antimicrobial susceptibility profile should be used to guide therapy. If unavailable, a trial of empiric therapy for E coli is warranted (Box 53-2). Antimicrobial therapy has not been shown to decrease the morbidity/mortality of patients with HC/HUS. Antimicrobial therapy may worsen the clinical course, possibly by decreasing competitive enteric flora. A synthetic analog of G3b and diatomaceous earth, SYNSORB-Pk (Synsorb Biotech, Inc.), holds promise as a treatment of HC/HUS. Taken orally, this agent should absorb the VT and, it is hoped, prevent HUS.
As previously noted, disease caused by the EIEC is very much like shigellosis. It is unclear, however, if duration of disease and shedding of viable organisms by patients infected with EIEC are diminished by antimicrobial therapy, as occurs in patients with shigellosis. The diagnosis of EIEC infection requires the isolation of the organism; therefore, an antimicrobial susceptibility profile should be available to guide therapy.
BOX 53-2 Empiric Therapy for Diarrheagenic, Non-EHEC, E coli Infection1
Infections by the EHEC are severe and sometimes fatal. Antimicrobial therapy and antimotility agents may worsen the clinical course. Fluid and electrolyte replacement should be used as needed to treat dehydration. Transfusion, dialysis, and other supportive measures may be required for patients with HUS. Antimotility agents should not used by patients with severe infectious enteritis regardless of the etiology, but should be especially avoided by patients infected by EHEC. These drugs increase the duration and severity of disease by inhibiting the passage of the pathogenic bacteria and their toxins.
Prevention & Control
The diarrheagenic E coli, with the notable exception of EHEC, are transmitted in a human fecal-oral cycle. Most of these organisms thrive in underdeveloped countries secondarily to poor living conditions, ineffective sanitation, and unsafe drinking water. Improvements in sanitation and the quality of drinking water, as well as raising the standards of living, would greatly diminish the prevalence of these diseases (Box 53-3). In endemic areas, political instability, war, and a weak socioeconomic infrastructure contribute to the persistence of these organisms. Prophylactic antibiotics are not recommended for most travelers, but individuals at high risk for severe disease may benefit from antimicrobial prophylaxis. Travelers should drink bottled water and avoid eating raw, locally washed vegetables. Bismuth subsalicylate, 2 oz or two tablets four times daily, provides some prophylactic benefit, but should not be used as a substitute for other preventive measures. There are currently no vaccines approved for human use against the diarrheagenic E coli..
The EHEC inhabit the intestinal tracts of cattle and other animals. Control of this organism in its natural reservoir is currently impractical. Antimicrobial agents should not be used for its suppression in animals, because these practices provide selective pressures that promote antimicrobial resistance. Infections by the EHEC can be diminished by thoroughly cooking food, consuming only clean water and pasteurized juices, using good food preparation techniques, and maintaining good personal hygiene.
A wide variety of foods have served as vehicles for the transmission of EHEC. These include beef and beef products, dried salami, yogurt, and fresh-pressed apple cider. Of these, ground beef is especially prone to contamination and should never be eaten unless thoroughly cooked. When cooking or reheating meats, all parts of the meat should reach at least 70°C. Children, who are at higher risk for HUS, should never be given undercooked hamburger.
Only pure, clean water should be consumed and used in food preparation. Wells, especially if they are near farms, should be periodically checked for enteric pathogens. Only pasteurized milk and juices should be consumed. Pasteurization is an effective means of eliminating EHEC, but care must be taken to avoid inadvertent postpasteurization contamination.
BOX 53-3 Prevention & Control of Bacterial Gastroenteritis/Enteric Fever
Good food preparation practices can also diminish the likelihood of EHEC infection. Hands should be washed thoroughly before food preparation and whenever raw meats have been touched. In addition, great care must be taken to avoid cross-contamination of foods that may not be thoroughly cooked before ingestion. Vegetables and fruits should be rinsed thoroughly under clean, free-flowing water. Cutting boards, knives, and other cooking utensils should be washed after they have been in contact with raw meat.
The infectious dose of E coli O157:H7 is so low that person-to-person transmission may occur. Anyone with EHEC HC must thoroughly wash their hands to avoid transmitting the bacteria. Children with a diarrheal disease should be carefully monitored for good hand washing. Finally, children with a diarrheal disease or history of recent diarrheal disease, especially bloody diarrhea, should avoid contact with other children, particularly contact during swimming.
Essentials of Diagnosis
Shigella species are unique among bacterial enteric pathogens in that < 200 and possibly ≥ 10 organisms may transverse the gastric acid barrier and cause disease. For this reason, person-to-person transmission is common. Person-to-person transmission results in increased frequencies of shigellosis in day care centers, schools, and custodial-care facilities. Disease is most common in infants and young children and frequently occurs in family members of patients. Peak incidence occurs in the summertime, and common houseflies are thought to contribute to the spread of disease. Outbreaks have also occurred from fecally contaminated food. Transmission through contaminated water is most common in developing countries that lack adequate sewage and water treatment facilities.
In the United States, S sonnei is the most commonly encountered Shigella species, whereas S boydii has a worldwide distribution. The prevalence of shigellae appears to be cyclic, with replacement of the predominant strain approximately every 20 years. This cycling of prevalence is presumably secondary to slowly acquired herd immunity in a given host population. Epidemic shigellosis, caused by S dysenteriae and S flexneri, is prevalent in underdeveloped countries, but may develop anywhere that poverty, overcrowding, or conditions of war exist.
The numbers of shigellae present in the stool vary with the course of disease. Early in the watery-diarrhea phase, shigellae are abundant and number 103–109 shigellae/g of feces. During this phase of disease, shigellae are easily recovered on MacConkey or eosin methylene blue (EMB) agar, where they appear as lactose nonfermenting colonies. Later in the course of disease, in the dysentery and postconvalescent phases, bacterial stool counts decline to 102–103 shigellae/g of feces. Furthermore, the recovery of shigellae is inversely proportional to specimen transport time, especially in stool specimens with a low number of shigellae. During the latter phase of disease, culture is best accomplished by rapid specimen transport or bedside medium inoculation, combined with the use of enrichment broth and moderately to highly selective media, such as xylose-lysine-desoxycholate medium and Shigella-Salmonella medium.
In many laboratories, suspect colonies, lactose nonfermenters, are screened by using a three-tube set: (i) one tube containing triple sugar iron (TSI) or Kligler iron agar (KIA), (ii) the second containing lysine iron agar (LIA), and (iii) the third containing Christensen's urea agar (CU) (also see the Salmonella Microbiology section below). On the TSI and KIA, shigellae characteristically produce an alkaline slant and acid butt without the production of gas. Rare isolates may produce gas. Negative reactions are produced on the LIA and CU, because shigellae do not decarboxylate lysine or hydrolyze urea. In addition, shigellae do not produce hydrogen sulfide, which is detected by the TSI, KIA, and LIA systems. An attempt to agglutinate organisms that are thought to represent Shigella species may be performed by using group antisera. Isolates with a suggestive screen profile are further characterized by additional biochemical reactions in either traditional or automated systems.
Useful clues in the identification of shigellae include the following: the majority of shigellae cannot ferment mucate, cannot use acetate, and are negative for indole and ortho-nitrophenyl-β-galactopyranoside.
The pathogenesis of the watery-diarrhea phase of bacillary dysentery is caused by a combination of lumenal bacterial replication and superficial mucosal invasion in the small intestine. During this phase of disease, large numbers of shigellae are present in the lumen of the small intestine. This phase of the disease is correlated with the onset of cramping abdominal pain, fever, and toxemia.
Within days, the lumenal contents of the small intestine do not contain shigellae, and the site of infection is the colon. The shigellae invade colonic mucosa and occasionally invade to the level of the submucosa. Factors that are important for invasion are present on the bacterial chromosome, as well as on a 140-MDa plasmid. Eventually, epithelial cell death occurs, and the mucosa sloughs, possibly secondarily to shigatoxin production. The loss of mucosa evokes an intense inflammatory response and allows for the introduction of coliform bacteria. Microabscesses, epithelial ulcerations, and pseudomembranes that consist of sloughed epithelial cells, bacteria, fibrin, and inflammatory cells may be seen. This phase of the disease correlates with tenesmus and fractionated stools that contain blood, mucus, and inflammatory debris.
Noninfectious causes of diarrhea must also be considered. The differential diagnosis of noninfectious colitis is extensive and includes inflammatory-bowel disease, lymphocytic/collagenous colitis, neoplasia, and numerous other disorders. Patients with inflammatory-bowel disease may also have fecal leukocytes, limiting the usefulness of this test. An accurate diagnosis may be achieved through a thorough history and physical examination, excluding enteric pathogens through appropriate microbiologic studies, and by obtaining and reviewing gastrointestinal biopsies via endoscopy and histopathologic studies.
Patients with acute diarrhea, which may be watery to dysenteric; fever; abdominal pain; and systemic symptomatology/toxemia may have shigellosis. A history of exposure to individuals with shigellosis, travel to endemic areas, and exposure to a high-risk population, such as persons in a custodial-care facility, should raise the index of suspicion. The presence of leukocytes in the stool, although supportive, is by no means definitive for shigellosis. Fecal leukocytes may be present in the stools of patients with other bacterial enteritides, amoebic dysentery, pseudomembranous colitis, and noninfectious disease, such as inflammatory-bowel disease. The definitive diagnosis requires the microbiologic identification of a Shigella species.
Shigellae are particularly susceptible to some environmental changes, and they die rapidly in transport. Therefore, it is imperative to rapidly transport the stool of patients suspected of having shigellosis to the laboratory. This is especially important for patients in the latter stages of disease, in whom the number of shigellae in the stool are relatively few.
BOX 53-4 Treatment of Shigella Gastroenteritis1
Fluid and electrolyte replacements are necessary for patients with dehydration. In most instances, this is readily accomplished by oral rehydration. Unlike in many other bacterial enteritides, antibiotic therapy is important in the treatment of shigellosis (Box 53-4). Antibiotic therapy limits the clinical course of the disease, may decrease the likelihood of intestinal complications, and decreases the fecal excretion of viable pathogenic organisms, which in turn diminishes transmission. Fluoroquinolones are the treatment of choice for adults. TMP/SMX is the treatment of choice for children. Alternatives are ampicillin, chloramphenicol, and nalidixic acid. In areas of known resistance to TMP/SMX, such as parts of Southeast Asia, Africa, and South America, quinolones should be used for adults, and one of the above mentioned alternatives for children with shigellosis. When available, the antimicrobial-susceptibility profile should guide therapy.
Antimotility agents, such as diphenoxylate, should not be used. The inhibition of diarrhea increases the contact between the intestinal mucosa and the pathogenic organisms and their toxins and may cause more fulminant disease.
The prognosis is generally good for patients with endemic or sporadic shigellosis. Infants and the elderly, especially if malnourished, suffer the highest mortality. Epidemic shigellosis caused by S dysenteriae, however, is a severe and often life-threatening disease with mortality rates from 5% to 20%. This disease must be treated aggressively with antimicrobial and rehydration therapies.
Prevention & Control
The development and refinement of sewage disposal and drinking water treatment systems are important in developing countries. In both developed and developing countries, personal hygiene, good hand washing practices, and clean living conditions are important preventive measures, particularly in custodial-care facilities (see Box 53-3). Fly control and hygienic food preparation practices should also diminish the incidence of disease.
Essentials of Diagnosis
For decades, phenotypic studies have been used to identify and categorize the Salmonella species. These organisms have been categorized by antisera directed against particular bacterial somatic (O) and flagellar (H) antigens. Serologic stratification has resulted in the identification of > 2000 Salmonellaserotypes. These serotypes have traditionally been treated as individual species. More recently, molecular analysis has revealed significantly less variability among the salmonellae than serologic studies have implied. Newer taxonomy recognizes two Salmonella species, S enterica and S bongori, and six subspecies of S enterica. Although this taxonomy is more precise, many healthcare providers have a greater familiarity with the previous nomenclature. For this reason, reference to previous nomenclature is appropriate and should avoid confusion, facilitate communication, and ensure optimal patient care. Therefore, for the remainder of this chapter, the Salmonella serotypes will be treated traditionally as species.
Enteric fever is distinguished from enteritis if systemic manifestations predominate and there is bacterial dissemination throughout the body and extensive involvement of the reticuloendothelial system. Classically, the enteric fevers are typhoid fever or paratyphoid fever, caused by S typhi or S paratyphi, respectively. Infrequently, other strains may cause enteric fever. S typhi is found exclusively in humans, and S paratyphi is found predominantly in humans. Contraction of disease, therefore, requires the presence of individuals either recovering from enteric fever or harboring the organism. Enteric fever remains a major public health problem in much of the world, and in some areas it ranks among the top five causes of death.
Enteric fever is a severe, life-threatening disease that, despite antimicrobial therapy, still causes significant mortality. Some individuals who have had enteric fever become chronic carriers. These individuals are usually asymptomatic and most frequently harbor the bacteria in the gallbladder, which contains calculi. In areas where Schistosoma haematobium is endemic, the bacteria may also be harbored in the bladder, associated with schistosome eggs. Chronic carriers are public health threats, because they continue to shed pathogenic bacteria into the environment over long periods of time and may substantially contaminate local water supplies, particularly in areas lacking appropriate treatment facilities. If carriers are food handlers and poor hand washers, epidemics may occur through contaminated food.
Although the number of cases of typhoid fever has diminished dramatically over the past 100 years in the United States, the number of cases of nontyphoid salmonellosis continues to increase. These organisms colonize the intestinal tracts of a wide variety of animals and are also transmitted by the fecal-oral route, through the ingestion of contaminated meats and animal products. Some Salmonella serotypes occur in many different animals, whereas others tend to occur in particular animals. For example, S typhimurium and S enteritidis are often associated with chickens, whereas Salmonella arizonae is associated with reptiles.
Person-to-person transmission of Salmonella species among food handlers and healthcare workers is possible but is not a common mode of transmission. Infants and neonates, however, are at an increased risk for infection if exposed to an infected mother or other family members.
When plating stool for the detection of Salmonella and/or Shigella species, a moderately or highly selective agar is often used in conjunction with the MacConkey or EMB plates. Moderately selective agars for the isolation of Salmonella and Shigella species include Hektoen-enteric and xylose-lysine-desoxycholate agars. Highly selective media, such as Salmonella-Shigella agar, brilliant green agar, and bismuth-sulfate agar, are most effectively used in outbreak situations. If used appropriately, an enrichment broth, such as selinite broth, may also be used to increase the recovery of Salmonella species from stool.
Bacteria that are suspected to be Salmonella or Shigella species are often then tested in the three-tube set as described above. In the TSI/KIA system, the majority of salmonellae produce an alkaline slant over an acid butt and produce gas. However, strains of S paratyphi A can ferment lactose, which produces an acid over acid reaction. Although most salmonellae produce hydrogen sulfide from sodium thiosulfate metabolism, ~ 90% of S paratyphi A and 50% of S cholerasuis isolates do not. Hydrogen sulfide production is detected by the TSI, KIA, and LIA systems and, when present, reacts with ferric ammonium citrate to form a black precipitate. It is significant that S paratyphi A, which produce an A/A TSI reaction and do not produce hydrogen sulfide, may be discarded as non-Salmonella Enterobacteriaceae.
The LIA system detects lysine decarboxylation, deamination, and the production of hydrogen sulfide. The hydrogen sulfide-producing bacteria genera Morganella, Proteus, and Providencia may be differentiated from Salmonella by their ability to deaminate phenylalanine. Except for S paratyphi A, the salmonellae generate a positive reaction for the decarboxylation of lysine. Salmonella species, like Shigella species, do not produce urease and generate no reaction on CU agar. This medium is included in the three-tube set to help differentiate Salmonella species from some of the other hydrogen sulfide producers and to detect Y enterocolitica.
Isolates with an appropriate biochemical profile in the three-tube set are often serotyped and definitively identified by traditional or automated biochemical testing.
For patients with enteric fever, the percent yield from blood or bone marrow culture vs stool culture varies during the course of disease (see Clinical Findings). Clinicians must be aware of which sites render the highest yield during each phase of the disease and submit appropriate specimens for culture.
The incubation phase begins after ingestion of S typhi and usually lasts between 10 and 14 days. The incubation time, however, is variable and to a certain extent is inversely proportional to the size of the inoculum. Normal stomach acid functions as a physiologic barrier to infection, and ~ 105organisms are required to survive gastric passage. A smaller inoculum may cause disease in patients with achlorhydria or decreased gastric acid production.
The next phase of the disease is characterized by bacterial invasion of the mucosa. The salmonellae adhere to the lumenal surface of the enterocytes, particularly specialized enterocytes, termed M cells, which overlie the Peyer's patches. The low number of organisms present in the stool at this stage of disease explains the common occurrence of negative stool cultures.
After adhesion, the enterocyte cell membrane becomes ruffled, and there are cytoskeletal alterations. The bacteria are then internalized by endocytosis and transmigrate through the enterocyte. They exit the basilar aspect of the cell through the basement membrane and are free in the lamina propria. The organisms are phagocytosed by mononuclear phagocytes in Peyer's patches and/or are drained by terminal lacteals to regional lymph nodes. This is followed by lymphatic and hematogenous dissemination of the bacteria to organs with fixed tissue histiocytes, such as the liver, spleen, and bone marrow. The salmonellae are able to alter the environment within macrophages and survive phagocytosis. This intracellular location also protects the bacteria from phagocytosis by polymorphonuclear leukocytes, to which they are susceptible, and to aminoglycoside antibiotics, which have poor intracellular penetration. Another virulence factor that these organisms possess is a capsule that has antiphagocytic properties and protects them from complement and antibody-mediated killing.
Bacteria, which involve the liver, may be subsequently passed into the bile. These organisms thrive in bile and may colonize the gallbladder, especially if gallstones are present. A bacterial enterohepatic circulation occurs wherein bacteria that have been shed into the bile pass into the small bowel and again adhere to and invade the small intestinal mucosa.
Salmonella species survive poorly at low pHs and are killed rapidly by stomach acid. Unlike Shigella species, a substantial number of salmonellae are required to successfully cross the gastric acid barrier and cause disease. Any decrease in gastric acidity increases the likelihood of Salmonellasurvival. Anyone, such as persons that have had gastric surgery, with hypo- or achlorhydria are at increased risk for salmonellosis. Undercooked foods that are contaminated with salmonellae buffer gastric acid and thereby facilitate the passage of bacteria into the small intestine. Neonates and infants that may be relatively hypochlorhydric are at increased risk for infection. Persons with an impaired immune response are also at risk for nontyphoidal salmonellosis.
Intestinal perforation usually occurs during the latter portion of the middle stage or the early portion of the late stage, when symptoms are beginning to wane. Perforation may present as worsening abdominal pain or rapidly progressive hypotension if hemorrhage occurs. In the absence of rapidly progressing hypotension, abdominal imaging is useful in advanced typhoid fever if perforation is suspected. Emergency surgery is necessary for survival.
The development of a chronic carrier state is a complication of enteric fever and is of more importance to the public health system than to the patient. Chronic carriers shed viable, pathogenic organisms into the environment and serve as sources for subsequent outbreaks. S typhi is the serotype that is most often associated with the chronic carrier state. Much more rarely, patients infected with other serotypes may develop a chronic carrier state. Most often, chronic carriers harbor S typhi in the gallbladder in association with calculi. Urinary tract carriage is associated with Schistosoma haematobiuminfection or urinary tract calculi. Chronic carriers require long-term therapy, follow-up cultures, and occasionally surgery for the eradication of these organisms.
In the appropriate clinical setting, the definitive diagnosis of enteric fever requires isolation and biochemical characterization of the etiologic agent. Febrile patients who have visited endemic areas may have enteric fever. Some patients with enteric fever may have an insidious onset, but eventually become severely ill. The Widal antigen/antibody agglutination is a presumptive test that is still used for the diagnosis of typhoid fever. This test, however, lacks both sensitivity and specificity. Blood and/or bone marrow culture, followed by empiric therapy, is a more reliable method of diagnosing enteric fever. Early in the course of disease, blood and bone marrow cultures yield the highest recovery of organisms, while later in the course of disease, stool and sometimes urine cultures are more likely to become positive. The submission of appropriate specimens for culture is important, so that rapid isolation of the causative agent and susceptibility testing can be performed. Antimicrobial susceptibility testing is necessary to optimally direct therapy, because of possible antimicrobial resistance.
Chloramphenicol, TMP/SMX, ampicillin, third-generation cephalosporins, and quinolones have been used successfully for the treatment of enteric fever (Box 53-5). Unfortunately, antimicrobial resistance has emerged to each of these agents. Some Salmonella isolates are multidrug resistant. For this reason, whenever possible, antimicrobial therapy should be based on an individual isolate's susceptibility profile, obtained by standard methods. Until such data are available, empiric therapy should be used, based on known antimicrobial-resistance profiles.
Chloramphenicol was the first drug used for the treatment of typhoid fever. However, increasing resistance, high relapse rates, bone marrow toxicity, and the promotion of a chronic carrier state have limited its usefulness. If isolates are susceptible to chloramphenicol, advantages include its high efficacy, low cost, and oral administration.
Ampicillin and TMP/SMX were used to treat enteric fever after chloramphenicol resistance emerged. For susceptible isolates, these drugs are effective, easily administered, and do not have the high rate of relapse associated with chloramphenicol.
The third-generation cephalosporin, ceftriaxone, is highly effective for the treatment of typhoid fever in adults and children. Third-generation cephalosporins are especially useful as empiric therapy in areas in which multiple-drug resistance has been reported. When the susceptibility profile of a particular isolate is known, the antimicrobial agent may be changed to a drug with a narrower spectrum of activity. Hopefully, this will also diminish the selection of organisms with resistance to third-generation cephalosporins.
Currently, ciprofloxacin is the drug of choice for adults from India, Asia, or the Middle East. In these areas, S typhi strains that are resistant to chloramphenicol, ampicillin, and/or TMP/SMX have been reported. Chromosomally mediated quinolone resistance has emerged, however, but it is hoped that it will not spread as rapidly as the plasmid-mediated resistance to chloramphenicol, ampicillin, and TMP/SMX. First- and second-generation cephalosporins and aminoglycosides should not be used to treat S typhi infections, regardless of the in vitro susceptibility profile. Patients with typhoid fever who develop mental status changes may benefit from a short course of dexamethasone.
Atherosclerotic plaques may also become infected during Salmonella bacteremia. These are serious infections, because infected plaque material is almost impossible to sterilize with antibiotics alone. These plaques become an intravascular focus of infection and continually seed the bloodstream. In the laboratory, intravascular infections, including plaque infections, are detected as high-grade bacteremia, wherein > 50% of three or more blood cultures are positive. The abdominal aorta is a common site for atherosclerotic disease and is the vessel most commonly infected. Abdominal aortic plaque infections may result in a life-threatening aortoduodenal fistula or in mycotic aneurysms. Intravascular infections usually require combined medical and surgical therapy. If the patients are not surgical candidates, they may require long-term, suppressive antimicrobial therapy.
Endocarditis is not common in either Salmonella enteritis-associated bacteremia or typhoid fever. Patients with structural cardiac anomalies, such as previous rheumatic heart disease or ventricular aneurysm, are at higher risk for developing endocarditis, regardless of the etiologic agent.
BOX 53-5 Treatment of Enteric Fever1
Patients with Salmonella enteritis have positive stool cultures and occasionally positive blood cultures. Apart from the agents of enteric fever, Salmonella cholerasuis is more likely to result in bacteremia than other Salmonella species. After isolation, Salmonella species are rapidly grouped by using antisera and are differentiated by various biochemical reactions.
Over the past decade, several molecular methods have been explored for the rapid detection of Salmonella species, including enzyme-linked immunosorbent assays, indirect immunofluorescence, and polymerase chain reactions. Advances in molecular diagnostics should permit the rapid detection of antigens or nucleic acids that are unique to these organisms. Many of these tests, however, are not commercially available and may be cost prohibitive compared with culture.
Uncomplicated enteritis caused by Salmonella species should not be treated with antimicrobial agents. Antimicrobial agents may prolong bacteria shedding and promote antimicrobial resistance. Fluid and electrolyte replacement should be used to treat dehydration. In most instances, oral rehydration is sufficient, but, if severe electrolyte anomalies exist, intravenous rehydration may be necessary. Antimotility agents should not be used, because these agents inhibit the clearance of pathogenic bacteria and their toxins. Patients at risk for or with systemic disease/complications should be treated. A third-generation cephalosporin, such as ceftriaxone, may be used for patients with severe disease, until antimicrobial susceptibility data are available (see Box 53-5). Prophylactic therapy is appropriate for individuals at increased risk for severe disease (Box 53-6). In the absence of an antimicrobial-susceptibility profile, empiric therapy should be used.
Systemic infections should be treated with antimicrobial agents based on the particular isolate's susceptibility profile. Intravascular infections and osteomyelitis may require long-term antimicrobial therapy, possibly combined with surgery to effect a cure.
Prevention & Control
The prevention of all Salmonella infections requires an interruption of the fecal-oral cycle (see Box 53-3). Developed countries with adequate sewage disposal, clean drinking water, and an effective public health service have dramatically lowered the prevalence and incidence of enteric fever. Infections by S typhi have been particularly affected, because humans are the only known reservoir for this organism. Diminishing infections caused by other Salmonella strains are more difficult, because these organisms are commensals in the intestinal tracts of a wide variety of animals. Infections by these organisms are usually associated with contaminated or incompletely cooked foods. The control of these infections may be effected through education and efforts that stress the importance of rinsing meats when appropriate, avoiding cross-contamination of food preparation utensils, avoiding raw-egg–containing products, and thoroughly cooking meats.
Of particular public health interest is the association of S enteritidis with chicken eggs. Ovarian infections in chickens may result in the transovarial passage of Salmonella species. Therefore, even thoroughly washed eggs with intact shells may transmit S enteritidis. Foods that require uncooked egg whites and/or yolks should be prepared by using pasteurized egg products. For this reason, raw eggs and foods or beverages that contain raw eggs should not be ingested. Additionally, cracked eggs should never be consumed, because these are even more likely to be contaminated with Salmonella species. The thorough cooking of eggs renders them safe for consumption.
BOX 53-6 Prophylaxis for Salmonella Enteritis1
Beef and poultry may become contaminated with feces during the slaughter process. If possible, these products should be washed with free-flowing water and cooked thoroughly before consumption. Knives, cutting boards, and other utensils become contaminated during contact with uncooked foods. These fomites may cross-contaminate other foods and thereby transmit Salmonella species. Utensils and food preparation surfaces that have been used to process uncooked food should be washed thoroughly before they are used to process other foods. The thorough cooking of food and pasteurization of milk and other liquids substantially reduce the risk of Salmonella enteritis.
Vaccinations are available for individuals at risk for typhoid fever; prophylactic antimicrobial agents are not recommended. Current recommended candidates for vaccination include travelers to endemic areas and people who are household contacts of infected persons. In the United States, the traditional vaccine consists of heat-killed, phenol-treated organisms. This vaccine offers 55%–77% protection, but its usefulness is limited by side effects. The minor side effects include fever, headache, and local pain at the site of injection and may last from hours to several days, but more severe reactions may occur. This vaccine requires two injections 4 weeks apart. The minimum age for vaccination is 6 months, and a booster is required every 3 years.
A newer, capsular polysaccharide vaccine, ViCPS, appears to offer similar protection after a single injection and has fewer side effects. This vaccine has a minimum age requirement of 2 years and requires a booster every 2 years.
An attenuated, live-bacteria, oral vaccine has been licensed in the United States and may rapidly become the vaccine of choice for many individuals. This vaccine, Ty21a, appears to be safe and effective, with no serious side effects. It relies on the development of natural immunity by using an attenuated S typhistrain. Therefore, patients on antimicrobial therapy should not be given this vaccine concomitantly. As with other “live” vaccines, this vaccine should not be given to persons that are immunocompromised. It is not recommended for children < 6 years old. Because of these limitations, the oral vaccine will not completely replace the injectable vaccines. A booster dose is needed every 5 years.
None of the typhoid vaccines offer long lasting protection, and booster doses are required to maintain protective antibody levels. In addition, none of the vaccines are 100% effective and should be used in conjunction with other preventive measures. Information regarding typhoid fever prevention and the prevention of other travel-associated diseases is available from the Centers for Disease Control and Prevention toll free at 1-888-232-3228 or on the internet at http://www.cdc.gov/travel.
Future directions for vaccine development include vaccines that elicit longer lasting immunity and confer immunity to more than one enteric pathogen. Such vaccines have been created, using molecular techniques. The insertion into Salmonella strains of genetic material that encodes for the somatic antigens of S sonnei and V cholera has allowed for the genetic construction of bacteria with multiple somatic antigens. It is hoped that vaccination with these genetically engineered bacteria will confer protection to several enteric pathogens.
Dupont HL: Shigella species (bacillary dysentery). In Mandell GL et al: Principles and Practice of Infectious Diseases, 4th ed. Churchill Livingstone, 1995.
Ericsson CD: Travelers' diarrhea: epidemiology, prevention, and self-treatment. Infect Dis Clin North Am 1998; 12:285.
Miller SI, Hohmann EL, Pegues DA: Salmonella (including Salmonella typhi). In Mandell GL et al: Principles and Practice of Infectious Diseases, 4th ed. Churchill Livingstone, 1995.
Nataro JP, Kaper JB: Diarrheagenic Escherichia coli. Clin Microbiol Rev 1998;11:142.
Ryan ET, Kain KC: Health advice and immunizations for travelers. N Engl J Med 2000;342:1716.
Spangler BD: Structure and function of cholera toxin and the related Escherichia coli heat-labile enterotoxin. Microbiol Rev 1992;56:622.
Wolf MK: Occurrence, distribution, and associations of O and H serogroups, colonization factor antigens, and toxins of enterotoxigenic Escherichia coli. Clin Microbiol Rev 1997;10:569.