ETEC and EPEC in nursery and weaning pigs: Clinical findings, diagnosis and control

Diseases caused by E. coli have been a recognised problem for as long as pigs have been raised. Early work in the 1960s and 1970s elucidated the pathogenic mechanisms of E. coli causing newborn diarrhea, and lead to the development of maternal vaccines which effectively control this form of the disease. However, maternal vaccination with these vaccines does not protect piglets against the diarrhea and edema disease that occur in the postweaning period. Postweaning diarrhea (PWD) is endemic in many farms and its prevalence fluctuates over time.

Recent advances in our understanding of how E. coli cause disease have led to a better classification of pathogenic strains based on presence of virulence factors, permitting much more accurate diagnostic strategies. Antimicrobial resistance has often resulted in a crisis situation for the pig producers because of limited treatment strategies and it is now being recognised that there is an increased public health danger due to potential transfer of resistance into the food chain. This situation has provided the impetus to find alternative control strategies such as novel vaccines for administration to weaned pigs.

E. coli is a gram-negative bacterial rod which inhabits the intestinal microflora or ecosystem of most mammalian and bird species, including the pig. E. coli is classified into 150 to 200 serotypes or serogroups based on somatic (O), capsular (K), fimbrial (F) and flagellar (H) antigens. Most E. coli are commensals, that is, they reside in the intestine but are not harmful for the host animal. Only a small proportion of strains are pathogenic, being classified into categories or pathotypes based on the production of broad classes of virulence factors and on the mechanisms by which they cause disease. Within each pathotype, strains are classified into virotypes, based on the combination of virulence factors. Strains of a particular pathotype belong to a restricted number of serotypes or clones. Molecular genotyping or fingerprinting techniques such as pulse-field gel electrophoresis (PFGE) and more recently, whole genome sequencing, are being increasingly used as an adjunct or instead of serotyping, for the epidemiological monitoring of the E. coli in pigs and their environment. 

Strains of the most important pathotype in pigs, the enterotoxigenic E. coli (ETEC), produce one or several of a class of toxins called enterotoxins, which act on the intestinal epithelial cells to induce the secretion of water and electrolytes into the intestinal lumen, causing the clinical signs of diarrhea. ETEC associated with neonatal diarrhea usually produce only the heat stable enterotoxin STa, whereas ETEC associated with diarrhea in older piglets produce one or more of the heat stable enterotoxins STa and STb, and heat labile LT. Most of these strains also produce the enteroaggregative heat stable enterotoxin EAST1, which was originally found in enteroaggregative E. coli associated with diarrhea in humans. The role of EAST1 in pig E. coli has not yet been demonstrated. ETEC must be able to adhere to and colonise the intestinal mucosa to permit the release of sufficient levels of enterotoxin to result in the development of diarrhea. This adherence is mediated by hair-like structures on the bacterial surface, called fimbriae or pili. ETEC associated with neonatal diarrhea may produce one or more of the fimbriae F4, F5, F6, and F41. The first three fimbriae are also known as K88, K99, and 987P. Whereas ETEC carrying F5(K99), F6(987P), or F41 are mostly restricted to pigs up to the first week of age, F4(K88)-positive ETEC may also be associated with diarrhea in piglets through to and following weaning. In fact, ETEC associated with postweaning diarrhea most commonly produce F4(K88) or F18 fimbriae. The relative importance in postweaning diarrhea of ETEC producing these 2 fimbrial types varies greatly from one country to another. Both fimbrial types have several variant subtypes, based on antigenic differences. In the case of F4(K88), the variants ab, ac, and ad have been found.

Almost all F4(K88) examined belong to F4ac (K88ac), and are often referred to simply as F4(K88). In the case of F18, there are 2 known variants, ab and ac. The latter is more commonly associated with postweaning diarrhea, whereas F18ab is associated with edema disease. Isolates producing the F4 (K88) or F18 adhesin and certain isolates producing F6(987P) demonstrate hemolysis on blood agar. All other ETEC from pigs are nonhemolytic. 

Some ETEC may also produce an adhesin involved in diffuse adherence (AIDA), originally detected in E. coli isolates from humans with diarrhea. The role of this adhesin in production of diarrhea in pigs is not yet known.

A second pathotype found in pigs with diarrhea is known as enteropathogenic E. coli (EPEC). EPEC were initially associated with diarrhea in children, especially in developing countries. These bacteria have a very intricate and complex secretion system which allows them to adhere very intimately to the intestinal epithelium and cause typical attaching and effacing lesions, giving rise to the name of a broad category known as attaching and effacing E. coli. EPEC from different animal species may have different virulence factors, but all possess a variant of the EPEC attaching effacing factor Eae or Intimin, a cell outer membrane protein which is responsible for the very intimate attachment. Hence, the presence of Eae is indicative of an EPEC.

Shiga toxin producing E. coli (STEC) produce one or more of a family of cytotoxins which are known collectively as shiga toxins (Stx) or verotoxins (VT), so-called because of their structural similarity with the Shiga toxins of Shigella, and their characteristic effect on Vero cells in culture. The 2 names are used interchangeably in the literature. Many STEC are probably not pathogenic, but are present in the normal intestinal microflora. However, certain STEC strains which possess additional virulence factors may be highly pathogenic. In pigs, the most important STEC are those which cause edema disease. These strains produce the toxin variant Stx2e (VT2e) and the fimbrial variant F18ab. Certain strains produce both Stx2e and enterotoxins, as well as the fimbrial variant F18ac. These strains are classified as ETEC rather than STEC, as they are associated more with postweaning diarrhea than edema disease.

In piglets during the first several days of life, a severe watery yellowish diarrhea, dehydration, and often death, may be observed due to infection with ETEC producing F4(K88), F5(K99), F6(987P), or F41 fimbrial adhesins. In some cases, the infection may progress so rapidly that death occurs before the development of diarrhea. One or more animals in a group are affected. The mortality can be very high in affected litters, representing up to half of all pre-weaning mortality. In pigs with diarrhea due to ETEC, common necropsy findings are gastric infarcts, fluid dilation and congestion of the small intestine. On histopathology, layers of E. coli are observed adhering to the mucosa of the jejunum and ileum. Bacteria colonise both the crypts and the apex of the villi. In many cases of diarrhea, no typical lesion is observed.

Less watery diarrhea may also be observed from the first week of age to weaning, with low mortality and often decreased weight gain. Such cases are often also associated with other common infectious causes of diarrhea in pigs of this age group, including Clostridium perfringens, transmissible gastroenteritis virus, rotavirus, and coccidia. Rapid death with symptoms of shock, including cutaneous cyanosis of the extremities, or less acutely, hyperthermia, diarrhea, and anorexia may be observed in unweaned and recently weaned pigs following infection with ETEC:F4. This disease has been called enteric colibacillosis complicated by shock. Typical microscopic lesions of haemorrhagic gastroenteritis, congestion, and microvascular fibrinous thrombi and villous necrosis may be observed in the mucosa of the stomach, small intestine, and colon. This phenomenon is probably due to the rapid release of large amounts of lipopolysaccharide (LPS) by the colonising ETEC bacteria.

Sudden death of one or several pigs, decreased feed consumption, and watery diarrhea are associated with infection by F4- or F18-positive ETEC, usually during the first week after weaning. Symptoms may be less severe than those observed at birth, and often result in decreased weight gain. Outbreaks may be observed in pigs through to the grower barns. 

Pathological findings and the mechanism for causing disease are similar to those observed in newborn pigs.

Diarrhea may also be associated with EPEC infections. Histopathological lesions range from mild and scattered through the large and small intestine, to severe involving mostly the caecum and colon. They include light to moderate inflammation of the lamina propria, enterocyte desquamation and some mild ulceration, and light to moderate villous atrophy in the small intestine. Extensive multifocal bacterial colonisation of the surface epithelium by a thin layer of dark-stained coccobacilli, often oriented in a palisade pattern, is observed. The contribution of intestinal lesions caused by EPEC to the presence of diarrhea has not yet been well established.

Mixed infections of ETEC positive for F4, F18, F5, or AIDA, ETEC possessing none of the known adhesins, and EPEC are being more commonly observed in cases of post-weaning diarrhea.

Mixed infections of F18-positive STEC and F4-positive ETEC may be observed. In these cases, the predominant clinical sign is often diarrhea, probably caused by the F4-positive ETEC, although histopathological evidence of edema may be present. Infections with F18- positive STEC which are also ETEC as they are positive for shiga toxin and enterotoxins, are usually associated with postweaning diarrhea rather than with edema disease.

In pigs with diarrhea, initial diagnosis is mostly based on the clinical picture which usually characterises enteric disease caused by E. coli. This includes age of the piglet, the circumstances of occurrence of the disease and manifestations such as fecal material around perineum, diarrhea, dehydration, death. A presumptive diagnosis may be made by the observation of an alkaline fecal pH due to the presence of secretory diarrheic fluid. Few specific pathological changes may be attributed to enteric disease caused by E. coli infection.

The characteristic smell of the small intestinal contents on necropsy is helpful in diagnosis of postweaning diarrhea caused by E. coli. If possible, it is preferable to perform a necropsy of one or more affected pigs, with a minimum of delay after euthanasia to minimise the effects of autolysis. Considerations for interpreting gross lesions should include whether the intestine is from euthanized or dead pigs, for example, whether the intestine is thin walled from fluid distension or from post-mortem gas distension. The presence of Gram-negative bacteria, usually closely adhering to the small intestinal mucosa, on histopathology is a strong indication of the presence of this disease when caused by ETEC. In pigs with EPEC infection, the presence of bacteria adherent to the mucosa is more commonly observed in the large intestine, particularly the colon.

Enteric disease caused by E. coli infection in young unweaned pigs must be differentiated from infection due to Clostridium perfringens, transmissible gastroenteritis virus, rotavirus, and coccidia. The differential diagnosis of postweaning diarrhea and enteric disease caused by E. coli infection complicated by shock, particularly when a high mortality rate is observed and at greater intervals following weaning, would include salmonellosis and transmissible gastroenteritis.

 

Gram staining of direct smears from rectal swabs or intestinal contents will often demonstrate a predominance of Gram-negative rods. As E. coli is part of the normal intestinal flora, of which only a small proportion are pathogenic, the diagnosis of E. coli enteric disease is strengthened by the isolation from rectal swabs or intestinal contents of pathogenic E. coli. Isolation of pathogenic E. coli Rectal swabs or, preferably, samples of intestinal contents should be inoculated onto blood and MacConkey agar or other media which are selective for Enterobacteriaceae and allow differentiation of lactose-fermenting from lactose-nonfermenting Gram-negative enteric bacilli. Use of transport medium such as alginate swabs or Stuart's medium should be considered if isolation cannot be done within 24 hours.

Morphology, lactose-fermentation on MacConkey agar, and odour of colonies are a first indication of the identity of bacteria involved in the infection. To identify the species as E. coli, it is essential to determine the capacity of colonies to transform indole, since 99% of E. coli strains are indole-positive. Identification can be completed by the citrate assay (E. coli are not able to use citrate as the only carbon source) and by the methyl red assay. Colonies of F4 (K88)- and F18-positive ETEC strains are almost always hemolytic on blood agar. Thus, presence of hemolytic colonies is often used as a rapid means for making a diagnosis of pathogenic E. coli as the causative agent. However, it should be remembered that F5 (K99)- positive and other ETEC E. coli involved in enteric disease are usually not hemolytic and we are now finding more and more non-hemolytic ETEC:F4 isolates. Also, certain extraintestinal pathogenic E. coli (ExPEC) are hemolytic. Diagnosis based only on the presence of hemolytic colonies would not discriminate between F4 and F18 ETEC and F18 STEC which may be an important consideration when putting into place prevention strategies, such as vaccination. In addition, in cases of diarrhea with mixed infections of hemolytic and nonhemolytic ETEC, the assumption that the presence of hemolytic colonies indicated that F4 or F18 ETEC was the only causative agent may result in the non-hemolytic ETEC remaining undetected. In cases of diarrhea mixed with sub-clinical edema disease, this assumption would result in the edema disease remaining undetected. These are important considerations in light of the increasing prevalence of mixed pathogenic E. coli infections associated with cases of diarrhea, particularly later in infection or in groups with a more endemic presentation Pathogenic E. coli may be identified by serotyping, since a small number of specific O groups have been associated with enteric disease. O serogroup identification may be carried out by slide and tube agglutination using specific O typing sera and bacterial suspensions heated at 100°C or autoclaved for 2.5 hours. Complete O and H serotyping can only be carried out in a few reference laboratories.

Virotyping, or determination of the virulence factors, is a more definitive way of identifying pathogenic E. coli, as not all strains of a given O serogroup are pathogenic. Slide agglutination, with or without latex particles, is a simple and easy method to determine the presence of fimbrial adhesins of ETEC, which are expressed in culture media. This method is commonly used for the identification of ETEC:F4. However, bacteria must be grown on the appropriate media for detection of F5 (K99) and F41, which are only produced when the bacteria are grown in special minimal glucose media. In addition, F6 (987P) and F18 are often poorly produced in culture conditions. Detection of expression of adhesins of EPEC by rapid testing is rarely carried out as these adhesins have been less well characterised.

A diagnosis of enteric infection caused by E. coli may also be confirmed by the detection of pathogenic E. coli adhering to the intestinal mucosa directly in infected animals by examination of frozen sections using indirect immunofluorescence or by examination of formalin-fixed, paraffin-embedded tissues using immunohistochemistry. 

Currently, genotypic analysis is more commonly used to define the virotypes involved in an infection. Most commonly, polymerase chain reaction (PCR) is used. These tests permit the detection of genes encoding for virulence factors such as toxins and adhesins. Primers recognising different genes related to toxins (STa, STb, EAST1 or LT) and adhesins (F4, F5, F18, F41, etc.) for ETEC strains; attachment and effacement, such as eae; and Stx for STEC strains are readily available and can be used to perform PCR. PCR may also be used to detect pathogenic E. coli directly in situ in formalin-fixed, paraffin-embedded tissues. 

As E. coli grows rapidly and easily in routine culture conditions, a simple, sensitive, and inexpensive approach for the detection of the presence of pathogenic E. coli in samples is to perform PCR on DNA extracted from bacteria grown in broth culture medium. For instance, multiplex PCR amplification may be used to detect the genes encoding for the enterotoxins of ETEC, shiga toxins of STEC, and Eae of EPEC, associated with diarrhea or edema disease.

Inhibitors may be present in clinical samples and even in certain culture media such as MacConkey, hence, Luria Bertani (LB) broth may be used for culture of clinical samples. This approach is rapid and indicates overnight the presence of pathogenic E. coli, identifying the pathotype(s) involved in a particular case. However, it does not permit the identification of specific virotypes, as is possible when colonies are tested. Isolates from pathotype-positive cases may be virotyped by PCR. Identification of the causative agent of disease will permit a more appropriate choice of isolate(s) for antimicrobial resistance testing. In many laboratories, hemolytic isolates are selected for further virotyping. In a particular case, ETEC positive for one of the adhesins F4, F5, F6, F41, or F18, or of STEC positive for F18, when present, could be considered the causative agent of the presenting diarrhea or edema disease, as these virotypes are rarely present in the intestinal microflora of normal pigs. On the other hand, EPEC or ETEC positive for AIDA, when present, would be considered as opportunistic agents of diarrhea, as strains of these virotypes have been demonstrated to cause diarrhea in pigs in challenge experiments but are also present in the intestinal microflora of normal pigs.

Similarly, ETEC or STEC possessing no known adhesins, when present, would be considered as possible agents of diarrhea or edema disease, as strains of these virotypes have not yet been demonstrated to cause disease in pigs, and are also found in the intestinal microflora of normal pigs. Finding of Gram-negative bacteria closely adhering to the small intestinal mucosa, on histopathology in pigs from which an opportunistic or possible agent has been isolated would be a strong indication that this E. coli is the cause of disease. More complete characterisation of isolates in reference laboratories by serotyping, and clonal typing using such techniques as pulse-field gel electrophoresis (PFGE) and Whole Genome Sequencing, will allow the monitoring of changing trends and the identification of new, emerging E. coli virulence determinants which could gain importance due to the pressure of vaccination and antimicrobial therapy. Similarly, in those laboratories not routinely carrying out molecular techniques such as PCR, potentially pathogenic isolates may be detected by F adhesin typing using slide agglutination, and sent to a reference laboratory for definitive identification.

In newborn piglets, treatment with antimicrobials may be on an individual or litter basis, by mouth or parenteral injection. Commonly used antimicrobials are ampicillin, apramycin, ceftiofur, gentamycin, neomycin, spectinomycin, furizolidone and potentiated sulfa drugs.

Bacteriophages have been used with success experimentally but not yet extensively applied in the field. Oral fluid therapy consisting of electrolyte replacement solutions containing glucose, is helpful for the treatment of acidosis and dehydration. Drugs inhibiting the secretory effects of enterotoxin, such as chlorpromazine and berberine sulfate, may be useful for the treatment of diarrhea and dehydration, although many of these drugs have undesirable side effects. The use of antisecretory drugs such as bencetimide and loperamide, alone or in combination with antimicrobial drugs, may be helpful.

In postweaning diarrhea, antimicrobial and electrolyte treatment is also required and may be initially administered orally or parenterally. Fluids can be injected intraperitoneally. Sick pigs eat and drink very little, even though they may stand close to the creep or the drinking nipple.

Antimicrobials may subsequently be added to the feed or water. Antimicrobials which reach the intestinal lumen must be selected, such as amoxicillin/clavulanic acid, fluoroquinolones, cephalosporins, apramycin, ceftiofur, neomycin, or trimethoprim.

Strategies to reduce pathogenic Escherichia coli levels and maintain a suitable environment for well-being of the pigs

Unweaned piglets should be maintained at 30 to 34oC, in an environment free of drafts and on a low-heat-conducting floor. Particular care should be taken of under average weight piglets, which tend to lose weight more rapidly than average piglets. Good hygiene in the farrowing area helps to reduce the numbers of E. coli being presented to the piglet to a level which is controllable by the piglet’s own defence mechanisms. A dry, warm environment,
attained by appropriate ventilation, reduces the moisture available to enhance the survival of E. coli. A desirable temperature for the sow is about 22°C.

Farrowing crate design is important in minimising the fecal contamination of the farrowing area. Crates should not be too long, to minimise contamination of available space, and should be adjustable. They should be on raised perforated floors to allow fecal material to drop through and away from the piglets. Drinking water and creep feed should be clean and fresh and not contaminated with fecal material.

Quarantine should be practised when introducing new pigs, to control the introduction of E. coli of new virotypes into the herd, as animals in the herd will have little immunity to E. coli fimbrial antigens with which they have not had contact. Farrowing crates should be thoroughly cleaned and disinfected between litters. An all-in/all-out farrowing system with thorough disinfection between batches will greatly reduce the E. coli population in the environment.

Management factors predisposing piglets to postweaning diarrhea should be addressed. Increasing the age or weight at weaning is helpful. The temperature in the weaning house should start at 28-32°C, with a minimum of drafts and temperature changes. A good quality creep feed should be made available, especially in piglets weaned at an older age. Weaner diets should be highly digestible, not contain high quantities of soybean meal, and ideally should be based on milk proteins, although the latter may not be economically possible. 

Restriction of feed intake, high-fibre diets, and ad libitum feeding of fibre have been reported as effective deterrents to the development of postweaning diarrhea and edema disease. Good hygiene should be maintained, with daily removal of feces and thorough cleaning and disinfection of pens between batches of pigs.

Several strategies may contribute to reducing the build-up of pathogenic E. coli in the intestine after weaning. Prophylactic antimicrobials are often used but should be avoided due to the alarming increase in drug resistance following their use. At present, preventive feed medication is practised in the majority of the affected herds in most countries. Antimicrobial resistance is often induced within days or a few weeks. In addition to the classes of antimicrobials mentioned above for parenteral therapy, various aminoglycosides and colistin are widely used. The latter has the advantages of high stability, low toxicity, absence of infectious resistance, and slow development of resistance. However, resistance is being more and more observed. Addition of organic acids to the water supply or weaner diets may reduce gastric acidity and minimise survival of ingested E. coli. ZnO dietary supplementation
effectively controls postweaning diarrhea in pigs, but does not seem to affect bacterial colonisation of the gut as fecal excretion is constant. Modification of the feed by adding soluble fibres seems to be effective. Administration of a probiotic Lactobacillus bacterial culture seems to inhibit ETEC strain adhesion. However, many of these measures tend to delay the onset of postweaning diarrhea rather than preventing it completely.

Immunity to enteric E. coli infections is humoral and is initially provided through the maternal colostrum, lactogenic antibodies in the milk of the sow, and subsequently by a local intestinal immune response. Specific antibodies inhibit bacterial adherence to receptors on the intestinal epithelial cells and neutralize the activity of the enterotoxins or cytotoxins produced by E. coli.

Colostrum from the sow contains high levels of immunoglobulin G (IgG), which rapidly decrease during lactation and IgA becomes the main immunoglobulin class. The latter protects the gut against E. coli infection.

The newborn piglet begins to synthesize specific immunoglobulin and develop intestinal immunity during the first week of life. At first, IgM predominates, but after 12 weeks, it is replaced by IgA as the most important immunoglobulin class in the intestine. Thus, during the first weeks of life, colostrum is the main source of immunologic protection for the piglet. Maternal vaccination has been one of the most effective ways of controlling neonatal ETEC diarrhea in piglets. Identification of virulence factors important in the pathogenesis of ETEC diarrhea and application of recombinant DNA technology have resulted in the production of more efficacious vaccines over the last several years.

Immunisation of the mother with a vaccine containing the appropriate adhesins will effectively protect young animals from ETEC infection following sufficient transfer of maternal antibodies. Vaccination of gilts is especially important. One of the earliest vaccination techniques consisted of taking the small intestinal contents from a piglet with diarrhea, culturing it and feeding the culture to pregnant sows, usually about one month before parturition. This method is effective and confers a lasting immunity. Commonly used commercially available vaccines are given parenterally and may be killed whole-cell bacterins or purified fimbrial subunit vaccines. Both types of vaccine appear to work equally well and are usually given approximately 6 weeks and 2 weeks prior to parturition. Bacterins usually contain strains representing the most important serogroups and producing the fimbrial antigens F4(K88), F5(K99), F6(987P), and F41. The latter type of vaccine usually contains the same four purified fimbrial antigens. In cases where use of commercially available vaccines appears to be ineffective, it would be important to identify the E. coli serotypes involved for possible inclusion in an autogenous bacterin. Maternal vaccination break-down may also result from any disease process causing agalactia in the sow, thus diminishing transfer of colostrum. Mastitis may also affect colostrum production. In such cases, passive treatment with hyperimmune gamma-globulin from pigs immunised with E. coli F4(K88), F5(K99), F6(987P), and F41 can reduce the severity of diarrhea in piglets. Finally, it should be noted that the sow has not been vaccinated or exposed to those ETEC present in the environment of the piglets, her colostrum will not contain specific antibodies necessary for protection against adherence and proliferation of ETEC.

Maternal immunisation for neonatal ETEC diarrhea, as described above, is not effective for the control of postweaning diarrhea. In addition, administration by injection of these vaccines directly to the piglets will stimulate mostly systemic rather than mucosal immunity, giving rise to circulating antibodies which do not reach intestinal bacteria in high enough levels to be very effective. Such vaccines may even suppress the mucosal immune response upon
subsequent oral infection with a pathogenic E. coli. 

Vaccination of piglets before or at the time of weaning by oral administration, usually in the drinking water, of live attenuated or non-toxigenic strains of F4(K88)- or F18-positive E. coli effectively protects pigs against infection. Such vaccines are now commercially available in Canada, Europe, and the United States. There are no commercially available vaccines for the protection of pigs against postweaning EPEC infections.

Passive protection against colonisation with F4(K88)- and F18-positive E. coli may be accomplished by feeding powdered egg yolk from hens immunised with F4(K88) or F18, although this approach may not be very cost-effective as the protection only occurs as long as the egg powder is fed.

This approach to prevention promises to be effective and economical in the long term to increase the presence of both F4 and F18 resistance loci in the pig population. It will be important to avoid co-selection of unwanted traits closely linked with loci coding for the F4 and the F18 receptors. It cannot be predicted if additional types of adhesive fimbriae or new variants of known types will emerge which could bind to yet unidentified receptors.

Availability of techniques for large-scale selection of resistant animals is lacking and will be the main challenge in the future. A PCR-RFLP test detecting FUT1 M307 polymorphism, correlated with the gene controlling expression of the E. coli F18 receptor, could be a simple and inexpensive method for large-scale selection of animals. Genotyping for the identification of pigs resistant to F4-positive ETEC is based on genetic polymorphisms, possibly in the porcine gene for MUC4 in the region of chromosome 13. 

 

 

 

 

Fairbrother, J.M., Gyles, C.L. Colibacillosis. In Diseases of Swine. Zimmerman J.J., Karriker L.A., Ramirez A., Schwartz K.J., Stevenson G.W., editors. Iowa State University Press. Ames, Iowa, USA. 10th edition, 2012, Chapter 53, pp. 723-749.

Fairbrother, J.M., Nadeau, E., Gyles, C.L. Escherichia coli in postweaning diarrhea in pigs: an update on bacterial types, pathogenesis, and prevention strategies. Animal Health Research Review, 2005, 6, 17-39.

Gyles, C.L., Fairbrother, J.M. Escherichia coli. In Pathogenesis of Bacterial Infections in Animals. Gyles C.L., Prescott J.F., Songer J.G., Thoen C.O., editors. Wiley-Blackwell. Ames, Iowa, USA. 4th edition, 2010, Chapter 15: 267-308.