POST-WEANING DUFUEEA AND MORTALITY CAUSED BY ESCHERICElIA

COLI- INVESTIGATION OF RISK FACTORS AND CONTROL METHODS

A Thesis

fresented to

The Faculty of Graduate Students

of

The University of Guelph

by

MARIA DEL ROC10 AMEZCUA

In partial fuifilment of requirements

for the degree of

Master of Science

April, 2001

O Maria del Rocio Amezcua, 2001

I*I ofNational
Canada
Library Bibliothèque nationale
du Canada
Acquisitions and Acquisitions et
Bibliographie Services services bibliographiques
395 Wellington Street 395, rue Wellington
Ottawa O N K IA ON4 Ottawa ON K I A ON4
Canada Canada
Your file Votre rélérenai

Our file Norra reldrence

The author has granted a non-
exclusive licence allowing the
National Library of Canada to
reproduce, loan, distribute or sell
copies of this thesis in microfonn,
paper or electronic formats.
L'auteur a accordé une licence non
exclusive permettant a la
Bibliothèque nationale du Canada de
reproduire, prêter, distribuer ou
vendre des copies de cette thèse sous
la forme de microfiche/nlm, de
reproduction sur papier ou sur format
électronique.

The author retains ownership of the L'auteur conserve la propriété du

copyright in this thesis. Neither the
thesis nor substantial extracts fkom it
may be printed or otherwise
reproduced without the author's
permission.
droit d'auteur qui protège cette thèse.
Ni la thèse ni des extraits substantiels
de celle-ci ne doivent être imprimés
ou autrement reproduits sans son
autorisation.
 ABSTRACT

POST-WEXVNG A
-
D AND MORTALWY CAUSED BY E S C E I U C Z 4

COLI-INVESTIGATION O F RISK FACTORS AND CONTROL METHODS

Maria del Rocio Amezcua Moreno Advisor:

University of Guelph, 1999 Professor R Friendship

Post-weaning Escherichia coli diarrhea (PWECD)in Ontario was investigated using a

case-control study and a vaccination trial.The E. coli serogroups involved in PWECD
were identified in a study involving 50 Ontario nurseries. The clinical signs, the impact on
productive parameters and epidemiologicd risk factors were deterrnined by means of a
producer sunrey. The E. coZi serogroups involved in disease were resistant to multiple
antibiotics. PWECD was shown to be an economicdy important disease in Ontario.
Factors such as the use of detergent in cleaning procedures, feed manufacturing, and the
Porcine Reproductive and Respiratory Syndrome (PRRS)vaccination status for nursery
pigs and sows, were found to be associated with the disease. These relationships need to
be investigated further. The efficacy of a live oral and killed intramuscular E. coli vaccine
for the control of disease were evaluated on two commercial fanris with PWECD. Neither
vaccine provided a solution against PWECD in the herds tested.
Dedication

To Angel, Santiago and Montserrat

To my dad and mom

To my sisters and nieces

1 would Like to thank Ontario Pork and Ontario Minisiry of Agriculture, Food and Rural

M a i r s for fimding this project, and dl the producers and practitioners who made this

project possible.

1 would like to thank rny advisors Dr Robert Friendship, Dr Cate Dewey and Dr Carlton

Gyles.

1 also would like to acknowledge and thank the following people for al1 their help

throughout my Masters program and my research:

Bryan Bloomfield, Muriei Burke, Dr Anne Deckert, Jackie Gallant, Dr Gaylan Josephson,

Dr Jutta Hamrnerrnueller, Dr. Beverly McEwen, Dr John Fairbrother, Dr Valena Parreira-

Pinto, Dr Peter Physick-Sheard, Anna Pietruszkiewicz, Anne Valliant and Steve Wolfgrarn

1would specidy like to thank Angel, Santiago and Montserrat de Grau, and my family in

Mexico for their love and support.

 Table of Contents

Dedication
Acknowledgements
Table of Contents
Guide of Tables vii
Guide of Figures

Chapter 1: Post-weaning diarrhea and mortality caused by Escherichia coli-

Investigation of risk factors and con.trol methods: Introduction and literature
review
1- 1. Introduction
1-2. Charactenstics and classification of Escherichia coli strains that cause
diarrhea in pigs
1.2.1. SeroSrpe
1.2.2. Pathotype
1.2-2.1. Enterotoxigenic E. coli (ETEC)
1.2.2.2. Attaching and Effacing E. coli (AEEC)
1.2.2.3. Verotoxigenic E coli (VTEC)
1.2.3. V i r o ~ p e
1.2.3.1. Fimbriae
1.2.3.2. Toxins
1.2.4. Serogroups associated with neonatal colibaciilosis
1.2.5. Serogroups associated with post-weaning colibacillosis
1-3. Escherichia coli diarrhea
1.3- 1. Neonatal E. coli diarrhea
1-3-2. 'Tuew" post-weaning E. coli diarrhea (P WECD)
1.3-2.1. Epidemiology
1.3.2.2. Pathogenesis
1.3-2.3. Clinical signs
1.3.M. Lesions
1.3.2.5. Diagnostic techniques
1.4. Risk factors associated with post-weaning E. coli diarrhea (PWECD)
1.4.1. Management factors
1-4.1- 1. Weaning
 1-4-1.2.Temperature
1.4.1.3. Sanibtion
1.4.1.4. Density
1.4.2. Nutritional factors
1A.2- 1.Carbohydrates
1.4.2.2. Protein
1-4.2-3. Minerais
1.4.2.4. Fibre
1.4.2.5. Feed manufactming
1-4.2.7. Feed management
1.4.3. Water
1-4.4. Changes in the intestinal flora, physiology and morphology
1-4.5. Genetic factors
1.4.6. Concurrent infections
1.4.6.1. Porcine Reproductive and Respiratory Syndrome
1.4.6.2. Transmissible Gastroenteritis
1A.6.3. Rotavirus
1-5 Control and preïention of post-weaning E. coli diarrhea
1-5.1. Control of factors that predispose pigs to diarrhea
1.5.2, Factors directed at E. coli
152.1. Growth promoters
152.2. Probiotics
1.5.2.3. Zinc oxide
1-52.4. Organic acids
152.5. Bromelain
1.5.2.6- Egg yoik
1-5.3. Treatment of post-weaning E. coli diarrhea
1.5.4. Antibiotic resistance
1-6. Vaccination as an aid in controlling post-weaning E. coii diarrhea
1.6.1. ImmUILity system of the porcine gut
1.6.1.1. Antibody mediated defense
1.6.1.2. Cell mediated defense
1.6.2. Immunization
1-6.3. Passive immunity
1.6.4. Active immun@ of the weaned pig for the control of PWECD
1.6.4.1. Oral irnmunization
 1-6.4.2, Parenteral immuni7iition
1.7. References

Chapter 2: A study învestigating the presentation of post-weaning E. coZi diarrhea in

Southern Ontario, prevalence of E. coZi serogroups involved and their
antimicrobial resistance patterns. 88
2.1 Introduction 88
2.2. Matenal and Methods 90
2.3. Results 92
2.4. Discussion 97
2.5 References 112

Chapter 3 : A study ùivestigating epidemiological risk factors associated with post-weaning

E. coli diarrhea in Southern Ontario 114
3.1 Introduction 114
3.2 Material and Methods 115
3.3 Results 118
3-4 Discussion 121
3 -5 References 135

Chapter 4: Study of the efficacy of two vaccines: intramuscular autogenous and oral live
K12-K88 vaccines for the control of post-weaning E. coli diarrhea in pigs. 137
4.1 Introduction 137
4.2 Materials and Methods 138
4.2.1. E. coli strains and vaccine preparation
4.2.2. The farms
4.2.3. Vaccination protocol
4.2.4. Statistical analysis
4.3 Results
4.4 Discussion
4.5. References

Chapter 5: Post-weaning diarrhea and mortality caused by Escherichia coli - investigation

of nsk factors and control methods: General Discussion 159
Appendices

1 Producer survey.

II Raw data of a study investigating post-weaning E. coli diantiea in

Ontario.

III Raw data of a s h d y investigating epidemiological nsk factors for

post-weanùig E.coli diarrhea in Ontario.

IV SAS analysis output

 Guide of Tables

1.1. E. cuir' pathotypes, serogroups and virulence factors associated with

diarrhea in pigs

1.2, Total cases and herds positive for K88 (F4) (1992-1998) (AHL) 19

1.3. Common diagnostic techniques used for the determination of different 34

virulence factors produced by strains of E. coli which cause diarrhea in post-
weaning pigs

2.1. Proportion of positive and negative isolation of E. coli K88 (F4) between 102
case and control farms as part of a post-weaning E.coli diarrhea study
involving 50 Ontario farms in 1999

2.2 Proportion of diffsrent toxin genes identified by the Polymerase Chain

Reaction (PCR)between Kg8 Q?4) positive case and K88 (F4) positive
control farms as part of a study involvuig 50 Ontario f m s in 1999

2.3. Resistance patterns of 68 K88 (F4) positive isolates f?om 17 case farms
and 8 F4 positive isolates fkom 3 control farms and resistance patterns
according to serogroups of E. coli

2.4. Proportion of positive and negative isolation of E. coli K88 (F4) between 105
case and control f m s , as part of a survey study of post-weaning E. coli
diarrhea involving 48 Ontario farms in 1999.

2.5. Average daily gains and tota1 nursery mortality rate between case and
control f m s before and afier a post-weaning E. coZi diarrhea problem.

2.6. Clinical problems associated with the post-weaning E. coli diarrhea

problem, estimated fkorn the odd ratios, between case and control farms.

2.7. In feed antibiotics used in 12 case f m s which change antibiotic after the 108
problem of diarrhea occurred, compare to the in-feed antibiotic used in the
k s t feed for control fanns

vii
 Management of diarrhea between case and control farms as part of a study 109
involving 48 Ontario f m s in 1999 (1).

Management of diarrhea between case and control farms as part of a study 110
involving 48 Ontario f m s in 1999 (2).

Proportion of positive and negative E. coZÏ K88 (F4) isolates between 124
case and control f m s as part of a study invoIving 48 Ontario farms in 1999.

Quantitative variables tested for association with an E, coli diarrhea

problem in 48 Ontario nurseries.

Use of specific management systems between case and control farms, as

part of a study investigating risk factors associated with post-weaning
E. coli diarrhea.

Quantitative risk factors tested for association with an E. coli diarrhea

problem in Ontario nurseries.

Cleaning and disinfection procedures between case and control herds as

part of a study investigating risk factors associated with post-weaning
E. coli diarrhea.

Sources per pen and room between case and control herds as part of a study
investigating risk factors associated with post-weaning E. coli diarrhea.

Porcine Reproductive and Respiratory Syndrome (PRRS)and Traasmissible

Gastroenteritis (TGE) information, as part of a study investigating nsk
factors associated with post-weaning E. coli diarrhea.

Assessing for confounding between sow vaccinated against PRRSV and

nursery pigs vaccinated against PRRSV.

Final multi-variate analysis for the post-weaning E. coli diarrhea study

3.10. The average score code of rnicrobial contamination in the nursery facilities 133
@=IO) d e r disinfection, for total number of living bacteria, and the average
water coliform contamination between case and control fanns.

4.1. Results of K88 (F4)Escherichia coli vaccination with kiUed autogenous 149
and a live attenuated vaccine in a herd with endemic post-weaning diarrhea
(Fm1)

4.2. Results of the scores of diarrhea for three treatrnents: control, live
attenuated vaccine and killed vaccine in a herd with endemic diarrhea
( F m 1)

4.3. Results of a killed autogenous vaccine used as post-weaeing E. coli

diarrhea prevention in a herd with endemic post-weaning diarrhea
( F m 1)

4.4. Results of the scores of diarrhea for the injectable killed autogenous
vaccine in a herd with endemic scours (Farm 1)
 Guide of Fimres

2.1 Procedure followed by Gallants Laboratories Inc., for the serogrouping of 11 1

E. coli samples

3.1 Procedure followed by Gallants Laboratories Inc., for the serogrouping of 134
E. coli samples

4.1, Proportion of animals within the four different scores used to measure
severity of diarrhea for the three treatment groups (controls, live ord
vaccine, and killed intramuscular vaccine) in a herd with endemic scours
( F m 1)

4.2 Proportion of morbidity and mortality between the groups vaccinated 154
with an autogenous kilIed vaccine in a field trial and a case study foliow up

4.3. Proportion of animals within the four different scores of diarrhea, between 155
the groups vaccinated with the autogenous killed vaccine in a field trial
and a case study follow up to field trial

5.1. Distribution of 28 case and 22 control farms involved in a study of post- 166
weiuiing E. coli diarrhea in Southem Ontario in 1999.
Chapter 1: Post-weaning diarrhea and mortality caused by Escirerichia eoli-

Investigation of risk factors and control methods: Introduction and literature review

1.1. Introduction

The Animal Health Laboratory (AHL)in Guelph reported a sudden increase in post-

weaning pig diarrhea due to K88 positive E. coZi (P W E C D ) in the fa11 and winter of 1997

(Josephson and Smart, 1998a,b). What made this disease remarkable was the severity of

clinical signs reported and the fact that K88 E. coli was being isolated fkom significantly

older animals than previously. The disease has been characterized by acute omet of

diarrhea, someîimes occurring so rapidly that pigs are found dead before clinical signs are

observed. Less acute cases resuit in severe watery dianhea and dehydration (Josephson

and Smart, 1998a,b; Gyles, 2000).

The disease is caused by enterotoxigenic strains of Escherichia coli (ETEC), mainly of

the 0 149:Kg 1:K88ac (F4ac) serogroup with genes for one to three enterotoxins (LT, STa

and STb); although, enteropathogenic E. coli (EPEC) strains may also occasionally be

involved (Josephson and Archambault, 2000). An increase in the prevalence of post-

weaning colibacillosis in pigs has also been reported in the United States and Quebec

(Fairbrother, 1999a). Most outbreaks have occurred in segregated early weaning

nurseries (SEW), although traditional farrow-to-finish herds are also aected. These E.

coli isolates are often resistant to a wide range of antimicrobials. The factors contributhg

to the increased number of outbreaks of this more severe form of enteric E. coli infection

are not yet known (Fairbrother, 1999a).

 Various control methods have been attempted, hcluding changes in the diet,

management, and housing. Some practitioners have also used m e r e n t vaccination

protocols, such as autogenous bacterins injected into sucklixig piglets, additional

vaccination of sows to enhance passive immunity, or orally immunizing pigs with modified

live vaccines. Producers have often experienced short penods of success in combatting

the disease, only to have the disease reappear a few weeks Iater.

In this literature review, classification of Escherichia coli straius that cause diarrhea in

pigs and the different serogroups involved in neonatai and post-weaning diarrhea wifi be

described. Characteristics of the neonatal and the "new" post-weaning E. coli diarrhea

(PWECD), and diagnostic techniques will be discussed. Also, risk factors associated with

PWECD presentation, and possible rnethods of prevention and treatrnent will be

reviewed. Finally, the use of oral andior parenterai vaccines in suckling piglets as a

possible control measure of PWECD wili be described,

1.2. Characteristics and classification of Escherichia coli strains that cause diarrhea

in pigs

Escherichia coli are classified within the farnily Enterobacteriaceae, which are Gram

negative unicellular rods capable of facultative anaerobic growth (Gyles, 198 6;

Bertschinger et al, 1992; Woodward, 1997). Escherichia coli are an important

constituent of the normal intestinal flora of animals and are found in the environment

following fecal contamination. Most of the E. coli found in the intestine are non-

pathogenic, however a limited number of strains are pathogenic (Woodward, 1997;

Fairbrother, 1999a). These strains are grouped into several categories based on the

clinicd syndrome and pathological lesions that they cause, and on the presence of certain

virulence factors which are responsible for these lesions (Smith, 1992; Fairbrother, 1999a).

Curent practice is to use classification schemes which include serotype, pathotype and

virotype (Woodward, 1997).

1.2.1. Serotype

Perhaps the most widely used method of typing strains of E. coli is that of serologicai

charactenzation. Serotypes are based on four antigens: the somatic (O) antigens are

lipopolysaccharide (LPS) complexes in the ce11 wall, which are pyrogenic and have been

associated with toxic shock (van Beers Schreurs et al, 1992). Capsular (K) antigens are

polysaccharides which constitute a sIime layer or capsule; the presence of the K antigen

may result in altered hydrophobicity, adherence to inanimate surfaces to form biofilms,

resistance to desiccation, and increased tolerance to chemical disinfectants. The flagella

(H) antigens are proteins contained in the flagellum which are appendages involved in

motility of the bactena. Another important group of surface proteins are the fimbrial (F)

antigens, fmbriae are hair-like appendages involved in adhesion (Lior, 1994).

Serotyping is of value because some serotypes are more comrnonly associated ~ 4 t h

certain micro-environments than others and they may be linked to particular vinilence

factors (Gyles, 1986; Lior, 1994; Woodward, 1997; MacKimon, 1998).

 1.2.2. Pathotype

Escherichia coli types are also classified according to their Wulence properties. The

curent classification on this basis for E. coli that produce enteric disease in pigs, is as

follows:

1.2.2.1. Enterotoxigenic E. coli (ETEC)

Enteric diseases due to strains of ETEC are the most commonly occurring form of

colibacillosis in pigs. The vinilence charactenstics of ETEC are strongly dependent on the

production of adhesins and enterotoxins. These strains when îngested, stick to specific

receptors in the pig intestinal epithelium by means of h b r i a l or pilus adhesins. The

attached bactena multiply and then produce one or both ST and LT enterotoxins, which

act on the epithelial cells and cause them to secrete large volumes of fluids and electrolytes

resulting in a cholera-like diarrhea (MacKinnon, 1998; Fairbrother, 1999a; Nagy and

Fekete, 1999).

1.2.2.2. Attaching and Effacing E. coli (AEEC)

Al1 AEEC bactena have a chromosomal pathogenicity cassette, called LEE (for locus

of enterocyte effacement), which is a 35-kb region of DNA that encodes the gene

products necessary for the formation of the attaching and effacing (AE) lesion: these

products include proteins that mediate attachent. One of the factors that promotes

intimate adherence is the product of eaeA, a 94-to 97kDa outer membrane protein (OMP)

known as "intimin" (Zhu et al, 1994, 1995; Woodward, 1997; Agin and Wolf, 1997).

Close adhesion leads to the effacement of the rnicrovilli and to lesions similar to those

observed for enteropathogenic (EPEC) strains in human infantile diarrhea (Bertschinger et

al, 1992). The AE lesions have a charactenstic pattern of effacement of microvilli,

pedestal formation and intimate adherence of bacteria to the epithelial cell membrane as

observed by electron microscopy (Wada et al, 1996; Agin and Wolf, 1997).

The process of i n h a t e adherence also follows a series of biochemical changes

including rearrangement of the enterocyte cytoskeleton, tyrosine kinase activity and

increased levels of intracellular calcium (Smith, 1992; Woodward, 1997; Bertschinger,

1999; Fairbrother, 1999a). The attaching and effacing (AE) lesions have been Observed

mostly in pigs between four and six weeks of age, but rarely in pigs less than four weeks

of age (Zhu et al, 1994).

Attaching and Effacing coZi are associated with 045, 0103 and 0108 serogroups

which cause diarrheas in weaned pigs (Zhu et al, 1994; MacKinnon, 1998; Fairbrother,

1999a). E. coli 045 have been fiequently isolated fiom cases of PWECD in Quebec and

have been associated with AE lesions (Zhu et al, 1994).

1.2.2.3. Verotoxigenic E coli (VTEC)

These strains are associated with oedema disease of pigs and are typically groups

0138,0139 and 0141 (Ganon et al, 1988). Some PWECD isolates produce shiga-like

toxin type II variant S L n e , either with or without the production of enterotoxins. SLT2e

is rnainly associated with classical oedema disease serogroups, although not exclusively. It

has been suggested that the presence of this toxin, could play a role in PWECD (Ganon et

al, 1988; Hampson, 1994; MacKinnon, 1998; Fairbrother, 1999a).

 1.2.3. Virotype

Because many strains of E. coli isolated fiom animals are not pathogenic, it is

important to i d e n w the virulence factors produced by ETEC, EPEC or VTEC strains, to

establîsh the etiology of diarrhea

Most vinilence factors are located in transmissible plasmids (Smith and Lingood, 1971;

van Beers-Schreurs et ai, 1992; Hampson, 1994). Plasmids are circular double strands of

DNA, which replicate autonomously in the bacterial cytoplasm. It has been shown that

they encode for vinilence detenninants that can be transferred fiom one E. coli to another,

allowing for more rapid change in the Wulence of strains of E. coli (Woodward, 1997;

Charbonneau, 1999').

Structures such as flagella, capsule, cell wall, and products such as colicüis, cytotoxins,

and hemolysin have a potential role in vinilence in the gut of the pig (Gyles, 1994), but the

major classes of virulence factors identified for enterotoxigenic E. coli (ETEC) are:-pili or

h b r i a e and enterotoxins (Gyles, 1986; Gyles, 1994).

1.2.3.1. Fimbriae

The ability of adhesion of ETEC to intestinal epithelial ceils is mainly due to the

production of thin (3-7nm) rod-like protein structures that project fiom the surface of the

bacteria, with an adhesin either at the tip or along their whole length, which can be

morphologically, biologicaliy and antigenicdy different in various strains (Nagy and

Fekete 1999).

'Personal Cornmunication
 Adherence to, and colonizattion of the small intestines is an important virulence

attribute of enterotoxigenic E. coli that cause diarrhea in neonatai and weaned pigs. The

adhesin binds the fimbriae to specific receptors sites on the brush border of the enterocytes

(Gyles, 1994; MacKinnon, 1998). Five distinct h b r i a e are found on porche ETEC

isolates: K88 (F4), Kg9 (F5), 987P(F6), F41 (van Beers-Schreurs et al, 1992; Hampson,

1994; Gyles, 1994), and recently demonstrated F 18 (Dean-Nystrom et al, 1993;Nagy et

al, 1997). Almost all post-weanïng colibacillosis cases in pigs are caused by ETEC

expressing K88 (F4) or F 18ac (Josephson and Archambault, 2000).

K88 (F4) pili

There are three antigenic variants of K88: K88ab (F4ab), K88ac (F4ac), and K88ad

0;4ad). ETEC colonize the length of the jejunurn and the ileum. Fimbriae adhere to

specific receptors on the celi wall of the intestinal epithelial cells and to specific receptors

in the mucus coating the epithelium. The arnount of the receptor on the intestinal waIl has

been shown to be age dependent (Metcaif et al, 1991;van den Broeck et al, 2000). A

strong correlation exists between the expression of glycoprotein receptor complexes

associated with the brush border membrane and the development of disease symptoms

(Jeyasingham et al, 1999). Edfors-Lilja et al (1995), found that loci encoding the porcine

intestinal receptors for E. coli K88ac were assigned to chromosome 13.

It has been reported that certain pigs do not have receptors for the K88 (F4) adhesin

and are thus resistant to infection (Gyles, 1986; Bertschinger et al, 1992). Jeyasingham et

al (1 999), revealed the presence of an analogous glycoprotein complex in the K88 (F4)

resistant pigs, which did not bind the K88 fimbriae antigen, concluding that genetic
differences in glycosyl moieties of the receptor complex provide the basis for disease

susceptibility to K88 positive E-coli.

FI8 Piii

During the 1 s t few years, three new colonization factors or adhesive fimbriae were

described for groups of E. coZi involved in PWECD or edema disease: F 107 on edema

strains (Nagy et al, 1992; Dean-Nystrom et al, 1993), 2 134P and '8813' on ETEC strains

(Salajka et al, 1992; Nagy et al, 1997). Rippinger et al (1995), investigated the

relatedness of the three fimbriae: piii 2134P, fïmbriae FlO7, and colonization factor '88 lY,

and found that al1 were morphologically similar and shared a common antigenic

determinant, in addition to a variant-specific determinant. It was suggested that the

symbol 'a' should be used for the cornmon determinant and syrnbols 'b' and 'c' for the

specific determinants. Two serological variants were identified: F 18ab (for F 1O7), and

F18ac (for 2134P and 8813) (Rippinger et al, 1995).

Further studies demonstrated that F 18ab and F 18ac are biologically distinct. F 18ab,

fist called F 107, is fiequently linked with the production of SLT2e and serogroup 0 139,

producing edema disease, while Fl8ac (2134P and 8813) is most often linked with

enterotoxin (STa, S m ) production, and serogroups 0141 and 0157 producing diarrhea in

post-weaning pigs (Nagy et al, 1997).

Intestinal receptors of fimbriai FI 8ab have been investigated. Genetic data support the

hypothesis indicating that the property of brush border adhesion by F 18ab E. coli is

dominant and the resistance is a recessive trait, resulting fiom a lack of brush border
receptor to the F18ab h b r i a l . The receptor gene is localîzed on the porcine chromosome

6 (Nagy and Fekete, 1999).

Receptors for F18ac are different fiom those of K88 (F4) in that the F18 receptors are

not present on the microviili of newborn pigs and that the F 18ac and F4 receptors are not

always present simultaneously on the microvilli of older pigs either (Nagy et al, 1997;

Jeyasingham et al, 1999; van den Brock et al, 2000).

1.2.3.2, Toxins

Extracellular protein toxins are the main detenninants of illness in E. coli infections.

Enteritis and diarrhea in pigs are highly correlated with serogroups that produce

enterotoxins (Smith, 1992; MacKinnon, 1998). Enterotoxins are secreted by the E. coli

and work extemal to the bacteria causing a combination of deleterious effects on the

function of intestinal cells (Smith, 1992; Charbonneau, 1999').

E. coli Enterotoxins

Two classes of E. coli enterotoxins have been recognized:

Heat-stable enterotoxin (ST): ST represents a family of low molecular weight toxins,

capable of inducing diarrhea. There are iwo main types of heat stable enterotoxins, STI

and STII (Gyles, 1994; Nagy and Fekete, 1999). These toxins are classified on the basis

of methanol soiubility and biological activity (Nagy and Fekete, 1999):

STI (also known as STa) is a small peptide of 18 or 19 arnino acids, non-immunogenic,

methanol soluble toxin, which induces intestinal secretion in infant mice and neonatal pigs

(Gyles, 1986). This toxin is subdivided into two different types, STIa and STb. The
toxins bind to specific glycoprotein receptors on the b m h borders of the intestinal

epithelial cells and activate a trammembrane guanyl cyclase, which leads to raised levels of

cyclic gustnosine monophosphate (cGMP) in intestinal epithelial cells. This causes

increased fluid secretion by the osmotic effect of the inhibition of absorption of sodium

(Na+) ions. A role has also been suggested for arachidonic acid, prostaglandin and

leucotrienes in upsetting of the host fluid balance caused by STI (Gyles, 1992; Gyles,

1994; MacKinnon, 1998). Plasmids with genes for dmg resistance and STI have been

reported f?om ETEC of human and animal origins (Gyles, 1992).

STII (also known as STb) is a methanol insoIuble peptide of 48 amino acids, that

induces intestinal secretion in weaned and neonatal pigs, but does not affect infant mice- It

appears to be restricted to porcine ETEC and be involved primarily in post-weaning rather

than neonatal diarrhea (Gyles, 1992; Nagy and FeketeJ999). Dubreuil et al (1991),

dcmonstrated the immunogen capacity of purified STb (Dubreuil et al, 1991).

Although the effect of STII has not been fully elucidated, it has been suggested that it

produces mild histological damage to the intestinal epithelium (Rose et ai, 1987; Gyles,

1992), which causes impaired intestinal absorption. Net active secretion of bicarbonate

(HC03-) ion, has also been noted. Prostaglandin E2 may be the mediator of the fluid

excretion. A role for the G-protein-linked calcium channel has also been suggested

(Gyles, 1994; MacKinnon, 1998; Nagy and Fekete, 1999). This toxin is usually found in

association with STa, LT, or STa and LT (Gyles, 1994).

The genes that encode STII on plasmids are heterogenous and may also determine

other properties including LT, STI, colonization factors, dmg resistance, colicin
production, and transfer functions (Smith and Linggood, 1971; Gyies, 1992; Nagy and

Fekete, 1999).

Heat-labile enterotoxin (LT): this toxin is produced by some serogroups of

enterotoxigenic E, coli and closely resembles cholera toxin (CT) (Nagy and Fekete,

1999). LT is a large molecule of immunogenic protein made up of one A and 5 B

fragments. The A subunit consists of an Al fragment which contaîns the active site, and

an A2 fragment, which links the A l h g m e n t to the B subunits. The B subunit contains

the site for binding to intestinal epithelitim, which recognizes specific cell-membrane

receptors containing the oligosaccharides of the GM1 ganglioside. Once the B subunits

have fixed the toxin molecule to the ceil s d a c e , the Al Fagments will translocate into the

ce11 where they activate the adenylate-cyclase system which trïggers the release of very

high levels of cyclic adenosine monophosphate (CAMP),causing massive secretion of

chlonde (Cl-) ions and impaked absorption of sodium ( Na+) ions. The steep osmotic

gradient, so established, results in profuse watery diarrhea (van Beers-Schreurs et al,

1993; Gyles, 1994; MacKinnon, 1998; Nagy and Fekete, 1999).

E. coii Shiga-like toxins (SLT) (synonym :verocytotoxin, VT)

There are two antigenic types of VT: VT1 and VT2 or SLT-1 and SLT-II, different

antigenic types related with SLT-II have been identified: SLTIIv (VT2e); SLTIIvha

(VT2v-a); SLTIIvhb (VT2v-b) and SLT-IIva (VTev) (Gyles, 1992).

E. coli responsible for oedema disease belong to the verotoxigenic category. VTEC

colonize the intestinal mucosa by means of a fimbrial adhesin, usually F18ab, and produce

the cytotoxin SLTnv (VT2e), which targets endothelial cells and induces the typical
lesions. Certain isolates produce both enterotoxins and VT2e, and cary the F4 6 8 8 ) or

F18ac adhesin. These isolates may be associated with oedema disease, but more

fiequently with diarrhea (Nagy,et al, 1997; Fairbrother, 1999a).

1.2.4. Serogroups associated with neonatal colibacillosis

The most important O serological groups associated with neonatal diarrhea in North

Amenca are 0 8 , O I38,Ol4l7 0147 and 0 157 al1 with F4 and producing LT and ST

enterotoxins.(Alexander, 1994; Nagy and Fekete, 1999). Other groups such as 09, 064

and O 1O 1 have been increasingly isolated producing fimbrial antigen Kg9 ( F S ) , F41 or

987P (F6) and STa and or STb enterotoxins (Alexander, 1994) (see Table 1.1 .).

Receptors for K88 (F4) are abundant in newborn pigs and decrease with age.

Receptors for Kg9 (F5) gradually decrease with age (Nagy et al, 1992; Nagy and Fekete,

1999). In contrast, receptors for 987P (F6) in fact increase with age, but swine develop an

innate resistance against the strain by three weeks of age (Dean et aI, 1989). The nature

of the aging of receptors for F41 is unknown but they may dl be produced through the

weauing age (Nagy and Fekete, 1999).

1.2.5. Serogroups associated with post-weaning colibaciIlosis

The E. coli which cause post-weaning diarrhea are hemolytic and belong to a limiîed

nurnber of serogroups. The strains that are most comrnonly implicated in PWECD belong

to serogroups 0 l38,O l39,Ol4l, 0 149 and 0 157. Of particular interest is 0149, since

it only emerged as an important group in the late 1960's and early 1970's. Strains of

0249K91 with or without the K88ac pili are, since the 80's, the most fi-equently isoIated

ETEC Çom pigs with diarrhea in several European countries and parts of North America
(Gyles, 1986). Serogroup 0157 also appears to have been more fiequently encountered in

the Iate 1970's and 1980's (Hoblet et al, 1986). Regionai ciifferences dso seem to occur,

for example 0 4 5 has been only frequently reported fiom Quebec, Canada (Gyles, 1986;

Hampson, 1994).

The main adhesive virulence factor of ETEC PWECD strains is the K88 p 4 ) (mainly

K88ac) fïmbriae (Nagy et al, 1997; Dean-Nystrom et al, 1993). Some ETEC strains may

produce multiple adhesins such as K88, FI 8ac or K88, F41 or even K88, F18ac and F41

(Nagy and Fekete, 1999). PWECD are characterized by production of either LT, STa, or

STb or combinations (STa is almost always accompanied by STb and /or LT). The most

fiequently observed enterotoxin combinations are LT and STb, or LT, STa and STb

(Celemin et al, 1995). Porcine ETEC strains producing LT, aimost always produce K88

and hemolysin (Nag and Fekete, 1999), and these isolates usuaily belong to the O 149

group (Wilson and Francis, 1986; Soderlind et al, 1988; Hampson, 1994; Fairbrother,

1999a; Osek, 1999). F18ac isolates usuaily produce STa or STb, or STa and STb (Nag

and Fekete, 1999; Osek, 1999).

Receptors for K88 (F4) are produced, although to a somewhat reduced extent, al1

through the nursery pig stage. M e r 35 days of age these receptors start to decline

(Conway et al, 1WO), while receptors for the variants of F 18 (F 18ab and F 18ac) are

increasingly produced up to weanllig age (Nagy et al, 1992; Vogeli et al, 1996).

Pigs ùifected with strains (ETECNTEC) producing SLT2e verotoxin and LT, usually

die from diarrhea and dehydration, whereas, those infected with strains producing SLT2e

and STa a d o r STb may die with signs of either diarrhea, sudden death or oedema
disease. Small intestines become transiently more susceptible to heat stable (ST)

enterotoxins immediately afier weaning (Bertschinger and Gyles, 1994; Bertschinger,

1999).

Isolates belonging to the AEEC category are also observed in about 6% of piglets with

diarrhea in the post-weaning period (Fairbrother, 1999a). There is an association of eaeA

gene and attaching and effacing (AE) ability among porcine 045 E. coli fiom PWECD

(Zhu et al, 1994, 1995). Intimin mediates a close attachment of the bacteria to the apical

surface of epithelial cells, and is required for the f d l pathogenesis of the bacteria (Zhu et

ai, 1995).
Table 1.1. E coCi pathotypes, serogroups and vinilence factors associated with
diarrhea in pigs

Classification of E. coli Age of pig Serogroups Virulence factors

affected

Enterotoxigenic (ETEC) Neonatal 08,09,064,0138, Fimbrial: K88(F4),

0 101,0141,0147 K99(F5), 987P(F6),

and O157 F41

Enterotoxigenic (ETEC) Post-weaning 0138,O 139, Fimbrial: K88 (F4),

Toxin: STa,STb/LT
Attaching Effacing (AEEC) Post-weaning 045,O 103,O108 Products of the Locus

for Enterocyte

Effacement (LEE)
Verotoxigenic (VTEC) Fimbrial: F 18ac

(2134P and 8813)

Toxin: STx2e
 1.3. Escherichia coli diarrhea

Diarrhea is an economically important disease in pigs. It may be classified into three

main entities: neonatal diarrhea (1 to 3 days of age), young piglet diarrhea (f?om fust week

of birth to weaning), and post-weaning diarrhea Whilst there are many possible causes of

diarrhea in the young pig, the most important etiologic agent is E. coli p i e h l and

Hoefling, 1986; Taylor, 1996; MacKinnon, 1998; Francis, 1999). Several serologic types

of ETEC are capable of causing disease; the dominant S p e s Vary over tirne, possibly in

response to changes in natural and artificial protective immwity (Gyles, 1986).

Although pigs rnay suffer fiom E. coli diarrhea from birth to 12 weeks of age the first

week of life and the first 2 weeks after weaning constitute two distinct peaks in occurrence

of the disease (Gyles, 1986; Biehl and Hoefling, 1986; Taylor, 2996)-

1.3.1. Neonatal E. coli diarrhea

The e s t few hours of life are highly hazardous for the newbom piglet, in part, because

of the multiplication of E. coli throughout the intestine (Alexander, 1994). Pre-weaning

colibacillosis is caused by either hemolytic or non-hemolytic E. coli. The normal E. coli

flora in the nursing pig is rarely hemolytic; therefore, hemolytic E. coli isolated fiom

nursing pigs are nearly always pathogenic (Francis and Moxley, 1991).

In its most severe acute form, pre-weaning E. coli diarrhea affects piglets in the first 1-

2 hours of life, morbidity and mortality are high. If the onset of diarrhea is delayed for

several hous or days after bitth, the disease is then usually less severe.

The very first clinical signs in individual pigs may be raised hair and a slight shiver.

The watery appearance of the faeces is easily overlooked. In pens with minimal bedding,
pools of liquid feces on the solid parts of the pen floor can be noted. Diarrhea may be

brown, brightly coloured, creamy-white, grey or fawn. When voided by the piglet the

liquid faeces tend to dribble d o m the perineum, staining it and sometimes making it red

and sore.

Enterotoxins lead to loss of nuid and electrolytes, culminating in severe dehydration

and acidosis. Mected pigs rapidly lose condition, becoming hairy and weak, with

inelastic discoloured or sometimes dirty skui. As the diarrhea progresses the eyes become

du11 and sunken, the emaciation progresses and the ribs, spinal bones and pelvis bones

become increasingly prominent. If the pig survives, it will recover and thrive (Alexander,

1994; Bertschinger, 1996). Piglets continue to suckle until shortly before death. Post-

mortem signs include a stomach filled with clotted milk and only small amounts of fluid, or

in a later stage, pasty contents in the large intestine. Severe dehydration is also prominent

in dead piglets (Bertschinger, 1996).

Neonatal diarrhea occurs sporadically, sometimes af5ecting only one or two pigs in a

litter but commonly involving the whole Litter; or it may occur more frequently building up

until successive litters are almost al1 afTected. This variability is a consequence of genetic

resistance, and of variable contents of gammaglobulins in the sow's milk (Alexander,

1994; Bertschinger, 1996).

Many trials have confirmed the excellent protection conferred by sow vaccination

(Bertschinger, 1996). Passively acquired materna1 antibodies play a vitally important role

in the prevention of neonatal colibacillosis. Colostrum contains high levels of the

immunoglobulins GgG, IgM and IgA). The main barrier to attachment is the IgA
(Alexander, 1994; MacKinnon, 1998) which blocks adhesion of ETEC to the enterocytes;

with the aid of complement and lysozyme, it rnay bring about bacteriolysis.

Immmoglobulin IgA also neutralizes LT but not STa The lactogenic IgA immunity, lasts

throughout lactation but seems to be most effective in the hrst week, with its effectiveness

apparently waning at 10- 14 days postpartum (Alexander, 1994).

For affected piglets, fluid and electrolyte replacement would be helpful but is less

practical in this age group than antibiotic treatment. A test of susceptibility to antibiotics

is recommended. A w a m nest area suppoas medical treatments because diarrhetic piglets

are deficient in energy (Bertschinger, 1996).

1.3.2. cbNew''post-weaning E. coli diarrhea (PWECD)

Traditionally K88+ (F4) E. coZi organisms have been identified as a cause of diarrhea in

neonatal pigs, which has been kept under control by farrowing room cleaniiness and sow

vaccination. At weaning, the rich source of protective antibodies provided fiom the sow's

m i k rapidly declines, and the pig becomes dependent on it's own active immunity for

protection. (Alexander, 1994; Charbonneau, 1999'). Since the early seventies, together

with intensification of the pig industry, PWECD has become more important in wemed

pigs (van Beers-Schreurs et al, 1992). In a survey in East Anglia, UK, the disease was

considered to be endernic in 42.2% of f m s s w e y e d (MacKinnon, 1998), and in

Switzerland the prevalence was 3 5% (Bertschinger, 1999).

'Persona1 communication
1.3.2.1. EpidernioIogy

An apparent increase in the number of cases of PWECD and sudden death was noted in

the falL/winter of 1997 in Ontario. Most of the intestinal samples yielded pure cultures of

strongly hemolytic E. coli ( Josephson and Smart, 1998qb; Fnendship and Dewey, 1999;

Charbonneau, 1999'). A search of the files fiom the Animal Health Laboratory, University

of Guelph, and fiom the Veterinary Laboratory SeMces Branch, OMAEXA, revealed the

following information:

Table 1.2 Total cases and herds positive for K88 (F4) (1992-1998) (AHL)

Total cases 28 52 47 41 47 114 221

Total herds 27 48 41 40 46 103 187

In addition to a marked increase in prevalence of PWECD, the disease is often more

severe than it used to be and is now being increasingly seen in otder pigs (Gyles, 2000).

The mean age of piglets affected in 1998 was 25.6 days as compared to the average of

13.6 days in 1995, substantiating the reports fkom producers that the post-weaning pigs

are the most oflen affected. In some cases pigs are normal when leaving the sow barn and

are found dead on arrival at the nursery d e r periods of transit longer thm 24 h o m .

However, PWECD mainly occurs in the first week after weaning (van Beers-Schreurs et

al, 1992) but some authors report that the disease can affect pigs after two to three weeks

'Personal cornmunicatin
following weaning, or in some cases, following transfer of pigs to the grower units (Moon

and Biinn, 1993; Fairbrother, 1999a).

Why this organism has become a severe pathogen in post-weaning pigs is not known.

It has been suggested that the popularity of larger group sizes has led to an increased

abiiity for the disease to spread within pen groups. PWECD can however be seen in pen

groups as small as 10 pigs per pen. It has also been suggested that specialization of

labour in hog production, has led to improved management of the farrowing house, with

lowered incidence of farrowing house diarrhea, the piglets may be exposed to less E. coli

during the suckluig penod ailowing little opportunity to develop active immunity prior to

w e e g (Charbonneau, 1999').

Most outbreaks have occurred in early weaned piglets and in these operations, the

condition recurs sporadically (Josephson and Smart, 1998b; Charbonneau, 1999 ',
Josephson et al, 1999). Losses in segregated early weaning (SEW) operations usually

occur closely following the change in diet fkom Phase I to Phase II, or from Phase II to D I

but can occur at any tirne. It has also been suggested that the early weaned pig has an

immature immune system (Charbonneau, 1999 l ), and therefore these animals have

reduced resistance to E. coli challenge. On the other hand K88+ (F4) E. coli diarrheas

have also become a major problem on smaller conventional farrow-to-finish operations

(Josephson et al, 1999; Fairbrother, 1999a).

' Persona1 Communication

 The E. coli serogroup 0149:K88 is commody isolated fiom the affected fanns.

Cultures fiom clinical cases produce pure growths of the organism which, in vitro, are

hemolytic. This serogroup seems more v i d e n t in individual animals and more persistent in

herds than other E. coZi strains. It is not known whether this group has more or different

vinilence factors than other E. coli, nor why it persists in m i t s for many years when other

serogroups could be expected to be replaced by less v i d e n t forms (Wood, 1992).

Losses are attributed to: i) actual death of individual animals, some farms experience

mortalities as high as 20% in weaned pigs (Bertschinger, 1999); ü) the cost of medication;

iii) the labor costs involved in treating individual pigs, and iv) a reduced average daily

gain (ADG). Studies have revealed a decrease in ADG fiom 430-450 g r a m per day, to

400-410 grams per day, following the appearance of K88-positive E. coli in the nursery

(Josephson et al, 1999).

In a study made in Danish herds, where consequences of post-weaning diarrhea were

examined, a mortality rate of 8.5% was observed in litters affected with post-weaning E.

coli. Litters had reduced weight gains; were 2.3 days older at 25 kg body weight than

non-diarrhetic litters and had a four-fold increase in the incidence of subsequent

respiratory disease compared to pigs with no diarrhea (Svensmark et al, 1989).

E. coli isolates responsible for the disease are often resistant to a wide range of

antimicrobials. The fiequency of resistance has been steadily increasing over the last

several years (Fairbrother, 1999b). A wide range of antibiotics given in feed andor water,

as well as other chemotherapeutic agents such as zinc oxide, have given only partial

success and some times at a high cost (Wood, 1991). The disease has also persisted in
spite of attention to the u d environmental and hygiene factors, such as the rise in lower

critical temperzture at weaning and rigorous cleaning between batches.

1.3.2.2. Pathogenesis

It is convenient to discuss the disease process in two parts: colonization of the small

intestine and production and action of enterotoxins by the ETEC (Gyles, 1986).

C o l o h t i o n requires mucosai adhesion and proliferation. The degree of colonization

determines whether or not disease will result fiom infection (Bertschinger, 1999).

The number of E. coli present in the mid-jejunum in normal animals is of the order of

1o4. Post-weaning diarrhea is associated with an enormous proliferation of bacteria,

followed by colonization of the small intestine of weaned piglets by predominantly

hemolytic strains of E. coli. Such proliferation is necessary because a large concentration

of bacteria ( 10' to 101°) is required to produce diarrhea This concentration of bacteria is

not usually present in the piglet's environment @&Allister et al, 1979; Gyles, 1986; van

Beers-Schreurs et al, 1992).

When the protective milk supply stops at weaning, the disease may develop in

susceptible animals. It is apparent that the intestinal environment created by weaning

favors multiplication of certain strains of E. coZi (McAllister et al, 1979). Host factors

associated with prevention of colonization d e r the pig is weaned, include: age, genetic

resistance resulting fiom lack of receptors, gastnc pH, changes in the morphology, flora

and function of the intestine, and the presence in the intestine of specific antibodies against
surface antigens of the ETEC (Gyles ,1986; van Beers-Schreurs et al, 1992; Nabuurs,

1998; Bertschinger, 1999).

Weaning of piglets is associated with villus atrophy and a subsequent diminishing of

natural defenses, however this alteration in the villus, still enable E. coli to colonize

(Hoblet et al, 1986; van Beers-Schreurs et al, 1992).

Colonization takes place by adherence of the bacteria to the intestine. It is known that

pili play a critical role in attachent of ETEC to the surface of intestinal epithelial ceUs,

and that the capsular polysaccharide may contribute in strengthening the bonds between

organisms and epithelium, thereby permitting formation of micro-coIonies. These enable

the organisms to proliferate high in the small intestine, where the gut flora is minimal and

the intestine is particularly sensitive to the effects of enterotoxin (Gyles, 1986; van Beers-

Schrews et al, 1999).

Once attached, ETEC straùis begin to secrete one or more of the enterotoxins, the

crypt cells secrete fluids and electrolytes into the lumen of the intestine. An osmotic

gradient is generated which causes a vast outpouring of body water into the gut. These

enterotoxins may also impair absorption of fluid and electrolytes. Fluid loss may be so

severe that 30 to 40% of the pigs body weight rnay be lost into the lumen of the intestine

within hours (MacKinnon, 1996; Charbo~eau,1999').

1.3.2.3. Clinical signs

Tn spontaneous outbreaks caused by 0 149 E. coli, the f i s t manifestation reported is

sudden death of one or several pigs with no prior indication of illness, (Josephson et al,

'Personal Communication
1999) as early as one to two days after weaning. A bIuish-red discoloration, especially

over the extremities such as nose, ears, and the abdomen, is generally seen, and the

animais are often in good body condition. In many cases, where sudden death has been a

problem, the F18 or the gut edema strains of E-coli have also been suspected

(Charbonneau, 1999').

Other animals show a more "benign" form of PWECD, characterized by a marked

decline in feed consumption and the development of mild to severe diarrhea varying in

consistency fiom pasty to watery. The colour of the diarrhea may be white, brown, green

or very clear, it infrequently contains feed particIes but never contains fia& bfood (van

Beers-Schreurs et al, 1992; Hampson, 1994; Fairbrother, 1999a; Fnendship and Dewey,

1999).

Diseased anirnals become emaciated with a bony appearance, and may remain permanently

stunted. The rectal temperature is normal and vomiting occurs occasionaiiy

(Bertschinger, 1999). The skin may take on the appearance of dry parchment, as

dehydration progresses. The anus and perineum often have a red, irritated Look due to

skin contact with the alkaline feces (Friendship and Dewey, 1999; Bertschinger, 1999).

Pigs appear depressed, develop a rough hair coat and become pot-bellied. Their appetite

is reduced, but even in the terminal stage of the disease, they try to drink (Taylor, 1996;

'
Hampson, 1994; Bertschinger, 1996; Charbonneau, 1999 ). Also, severely a e c t e d pigs

try to move around with staggering and uncoordinated movements (Bertschinger, 1999).
 In the chronic stages of the disease pigs are severely dehydrated, cachectic, weak and

may be recumbent. Endotoxic shock may be seen in some cases as the E. coli invades the

blood strem late in the course of the disease (Bertschinger, 1996; Charbonneau, 1999';

Knif5en and Neumann, 2000). The severity of clinical signs depends on the pig's age,

immune status and the virulence factors of the E. coli involved.

Most of the pigs in a group become affected to a greater or lesser extent (Hampson,

1994). During an epidemic, greater than 70% of al1 pigs may show clinical signs ( M e n

and Neumann: 2000). The group mortaIity rate can reach 25% in the absence of adequate

medication (Wood, 1991; Hampson, 1994, Bertschinger, 1999). The peak of moaality

occurs six to ten days after weaning. Some surviving pigs recover well and some pigs

remain completely spared fiom the disease (Bertschinger, 1999).

Wada et al (1996), reported the first dual infection of AEEC and ETEC in weaned

pigs. Such dual infection in the small and large intestine might increase the severity of

diarrhea.

1.3.2.4. Eesions

Gross lesions

The carcasses are generally cyanotic and dehydrated. Pigs in the peracute stages of the

disease generally have ernpty stomachs but have intestines that are edematous, thin walled

and contain varying amounts of fluid of normal color. Large intestines are distended and

may or may not include gas. These lesions are most prominent in the posterior part of the

jejunum. The contents Vary fiom watery to mucoid, with a characteristic smell.

'Personal communication
Occasiondy blood-tinged contents may be seen (van Beers-Schreurs et al, 1992;

Hampson, 1994; Josephson et al, 1999).

Piglets in the acute and subacute stages of the disease frequently have empty small

intestines and variably filled large intestines. Extensive hyperemia of the gastric mucosa is

often noted in the fundic region. The mesentery is aIso heavily congested, as are the

rnesenteric lymph nodes. Contents of the large intestine most often look light-greenish or

yellowish and are mucoid to watery. The epithelium of al1 sections of intestine appear

normal and intact, but rnay be covered with a muco-purulent discharge. Lungs look pale

and dry (Kniffen and Neuman, 2000).

Pigs dying late in an outbreak look emaciated and exhibit a strong srnell of ammonia.

There are irregularly shaped superficial ulcerations in the fundic region of the stomach and

similar lesions of smaller size in the Iarge intestine (Bertschinger, 1999).

Some authors use the terms "hemorrhagic gastroenteritis" or ''colibacillary shock" to

describe a f o m of E. coli diarrhea characterized by severe congestion of the gastric

fundus and small intestine, with or without blood-tinged contents of the small intestine and

some tirnes the upper Iarge intestine, but only exceptionally with bloody feces

(Bertschinger, 1999). Josephson et al (1999), reported a coincidental presence of colitis; a

finding not identified previousIy in K88-positive E. coli induced diarrheas.

When enteropathogenic E. coli are involved, the smail intestine may be inflamed with

reddened contents and loss of villi and there may be necrotic material on the mucosa

surface. In some cases the large intestine is d s o afTected with reddish or brownish watery
contents and necrotic material and adherent contents on the caecal and colonic mucosa

(Taylor, 1996)-

Microscopic lesions

Microscopic examination of sections of intestines fiom affected, sacrificed pigs reveal

the presence of large numbers of bactena adherent to the mucosal surface, vascdar

congestion, and some hemorrhage (Hampson, 1994; Friendship and Dewey, 1999). The

bacterial layers are restncted to the villi and look patchy. The mucosa and the epitheliurn

appear intact (Bertschinger, 1999).

White blood cells cari be seen rnigrating into the lamina propia and lumen of the

intestine, especially in the jejunum and ileum, but fkequently the colon is also involved

(Friendship and Dewey, 1999) In pigs with so-called hemorrhagic gastroenteritis, severe

congestion of the gastric and small-intestinal mucosa is commonly associated with micro

vascular fibrinous thrombi. Necrosis of Mlli with marked infiltration of neutrophils occurs

in severe cases (Bertschinger, 1999).

Presence of micro-thrombi in the organs of some pigs, also suggests that some agent

which causes vascular damage may be contributhg to disease in at least some pigs (Gyles,

2000).

In pigs infected with attaching effacing isolates, a multi-focal coIonization of the brush
border of mature enterocytes with degradation and Iight to moderate inflammation of the

lamina propria is obsexved, mostly in the ileum. On transmission electron microscopy,

bacteria are intimately attached to the cytoplasmic membrane of mature enterocytes and

arranged in regular palisades, parallel to the microvilli, with effacement of adjacent villi.
The bacterial ce11 wall and the apical cell membrane of the enterocyte are separated by a

narrow reguIar gap of 10 nm at the cupping pedestai, and apical dense regions are seen at

attachment sites (Fairbrother, 1992; Wada et al, 1996).

Diagnostic techniques

hemoIytic cultures of weaned pigs should be considered significant. However,

the intestines of weaned pigs fkequently harbor hemolytic non-pathogenic E. coli, so one

should not assume that al1 hemolytic organisms isolated fiom those animals are

diarrheagenic (Francis and Moxley, 1991). Hence, for enteric disease it is necessary to use

methods that distinguish the pathogenic fiom the nonpathogenic E. coli.

Diagnosis of post-weaning E. coli diarrhea (PWECD) is relatively straightforward,

being based on the hridings of diarrhea in groups of weaned pigs, with minimal gross

changes in the intestinal mucosa, and with presence of gram negative organisms adhering

to the intestinal mucosa in enteric infections (Fairbrother, 1999a).

The E. coli isolates can be M e r identified by serotyping, and/or examining for

specific vinilence detemiinants. These may be detected using a variety of techniques,

including imrnunological assays, biological assays, and methods for identification of

specific DNA sequences. Although sophisticated methods for detecting the presence of

v i d e n t strains of E. coli are now available, some care must be exercised in ùiterpretation

since these bacteria rnay also be found in healthy pigs (Hampson, 1994; Nagy and Fekete,

1999).
 Culture of E. coli

A simple approach for diagnosis is to culture rectal swabs, or small intestinal or fecal

samples, and to use transport medium if the samples c m not be processed withui 2-3 hours

of collection (Gyles, 1986; Alexander, 1994; Fairbrother, 1999a; Nagy and Fekete, 1999).

The bacteriological analysis of fecal samples for ETEC is more difficult because the

bacteria present in the feces may not reflect the microbial status of the small intestine

@avis, 1989; Nagy and Fekete, 1999), and may not necessarily be the causal agent,

although if it is in pure profüse culture it is more likely that it is. The reliability of the

diagnosis is greatly improved if one or more untreated live early cases or, failing that,

fkeshly dead untreated piglets with diarrhea are available (Alexander, 1994).

Altematively, E. coli can also be estirnated by culturing of the mucosa of the ileum on

blood agar and/or a differential medium agar such as Tergitol-7 or MacConkey (Francis,

1999). The specimens are cultured using blood agar plates, where E. coli appear as gray

to white colonies and may exhibit hemolysis, or Tergitol-7 plates where E. coli colonies

usually ferment lactose and appear as yellow colonies (smooth, intermediate, rough, or

mucoid). If a predominant colony type is observed in high numbers, it may be signifïcant

and can be tested for sensitivity and serotyping. Because blood agar is the enhancement

medium for the K88 antigen, when swabs are streaked directly onto this medium, the

colony selection can corne directly fiom the blood agar plate. If the initial isolation utilized

Tergitol-7, then the selected colony should be subcuitured onto blood agar (Davis, 1989).
 Serological assays

Diagnosis of ETEC infection is based on the detection of known virulence factors of

the suspected ETEC. To test if the isolates are ETEC,the fimbiral antigens K88 (F4), Kg9

(FS), F41 and 987P (F6) can be detected by slide agglutination using specific absorbed

sera or by latex agglutination for which there are monoclonal antibody-based kits

available (Wray and Woodward, 1994; Taylor, 1996; Nagy and Fekete, 1999;

Fairbrother, 1999a) (Table 1.3 .).

The swabs are cultured ovemight (12h), classically on sheep blood agar and

MacConkey or Tergitol-7 agar. The next day, suspect E.coli colonies are picked off and

tested in slide or tube agglutination tests with appropriate antisera to determine 0, K, and

pilus antigens (Gyles, 1986; Davis, 1989; Alexander, 1994). If they belong to a pathogenic

serogroup, their susceptibility to antimicrobial dnigs commonly used for piglet diarrhea

may be tested (Alexander, 1994). The need to determine antibiotic resistance patterns,

makes culture and test of bacterid attributes in-vitro an accepted routine for diagnostic

laboratories (Nagy and Fekete, 1999).

ImmunoIogical assays

Immunofluorescence or irnrnunocytochemistry, is one of the better approaches to

determine the existence of ETEC infection fkom ileal impression smears or histologic

sections for bacteria expressing adhesive fimbnae (Francis, 1999). As there may be ETEC

strains without known (or detectable) adhesive virulence factors, it is advisable to perform

tests for enterotoxins as well.

 Enzyme-linked immunosorbent sandwich assay (ELISA) has been developed to detect

and quanti@ both the K88-positive fimbrial antigen of E. coZi, and the concentration of

bacterial ceils in fecal sampies uçing anti-fhnbrial antibodies. The assay is sensitive and

specific and provides a good estirnate of the amount of Çnbrial protein or the number of

K88-positive ETEC in the fecal sample (Mills et al, 1983; Wray and Woodward, 1994).

The assay can be completed Inone day, and resdts c m be relayed rapidly; and give the

possibility for doing the assay in smail laboratories with little equipment (Miils et al,

1983).

LT and ST toxins can d s o be identified by ELISA. Although STs are not

imrnuogenic, antibodies have been raised to them by coupling to a protein carrier such as

bovine IgG, bovine semm albumin or haemocyanin. A cornpetitive enzyme immunoassay

is comrnercidiy available for STa,and for the detection of STb (Wray and Woodward,

19943.

ELISA test for the detection of VT-producing cultures or for direct detection of VTEC

in feces has been used by a number of workers. The receptor for VTl and VT2 is

globotriosyl ceramide (Gb3) which occurs in hydatid cyst fluid. Porcine VTZe, in

contrast, binds to globotetrasoyl ceramide (Gb4). Consequently, ELISA'S using hydatid

cyst fluid to coat the plate and bind the toxin have been developed. Conjugated polyclonal

or monoclonal antibodies may then be used to indicate the presence of the toxin (Wray

and Woodward, 1994) (Table 1.3.).

 Genetic methods

Another approach to the characterization of E. coli strains isolated fiom diarrhetic pigs

is genetic anaiysis (Francis, 1999). DNA probes are used for detection of DNA sequences

consistent with those that encode virulence determinants of interest. DNA-based tests

could include assays for the genes of adhesive h b r i a e , enterotoxins (Nagy and Fekete,

1999), STx2e (VTEC), and eae (AEEC) infections. An example of a DNA-based test for

genes of E. coli Wulence detenninants is the multiplex polyrnerase chain reaction (PCR).

This test overcomes the problems associated with poor expression of some virulence

detemiinants under in-vitro conditions. However, this test does not determine whether a

gene is actually encoding a specific virulence factor, only whether a specific segment of

DNA is present in the E. coli strain being tested. DNA segments fiom penes containing

mutations that render them firnctionally inactive, or that are similar in sequence but

different in function, can lead to fdse positive results (Francis, 1999)-

Studies made by Osek et al, (1999) in Poland, demonstrated that PCR for detection of

h b r i a e and toxin genes is a valuable and sensitive method for identification of virulence

markers and gives good results in epidemiological investigations of diarrhed E. d i strains

isolated from pigs. Especially F18 fimbriae and toxin (LT, STa and STb) genes rnay be

detected using this method (Gyles, 2000). This is more important since serological tests

for LT, ST and VT2e toxins are based on specific antibodies which are usuaily difficult to

produce (especially antibodies to STI and to VT2e may show cross-reactivities and their

use is often restricted to specialized laboratories) (Osek et al, 1999) (Table 1.3.).
 The ability of PCR to detect very low quantities of an infectious agent at high

specscity allows investigators to apply this technique for the direct identification of the

vinilence factors in clinical specimens, including stool and intestinal samples within a

single working day (7 h) (Yavzori et ai, 1998; Fairbrother, 1999a).

Biological assays

There are some biologicd assays to determine the presence of enterotoxins. The

ligated intestine test rnay be used to detect the LTysability to induce accumulation of fluid

in ligated segments of s m d intestine in rabbits or pigs. The suckling mouse test is used to

detect the presence of STa toxin. Finally, rat or mouse inoculation to deterrnine the

presence of STb toxin can be perforrned (Table 1.3). These tests may be necessary in

investigations of enterotoxicity of bacterial products, but there can be no justification for

their use in the diagnostic Iaboratory (Wray and Woodward, 1994).

An accurate diagnosis of the type of E. coZi is essential, because it will determine

appropriate measures for treatment and controI (MacKinnon, 1996; Nagy and Fekete,

1999). However, diagnosis of PWECD diarrhea also requires careful consideration of the

predisposing factors and bactenological results. In addition, there is almost always a need

for differential diagnostic investigations (such as virus detection) as well (MacKinnon,

1996).
Table 1.3 Common diagnostic techniques used for the determination of different
virulence factors produced by strains of E. coCi which cause diarrhea in post-
weaning pigs

Virulence factor Serologic assay Immunologie Genetic Biologie assay

assay assay

Fimbriae (F4) Slide Agglutination ELISA PCR

Latex particle

Agglutination
STa toxin - ELISA PCR Suckling mouse
STb toxin - ELISA PCR Ligated gut (rat)
LT toxin - ELISA PCR Ligated gut

(rabbit, pigs);

ce11 culture
eae (AEEC) - ELISA PCR -
VT2e(VTEC) - ELISA PCR Ceil culture
1.4. Risk factors associated with post-weaning E. coli diarrhea (PWECD)

In modem veterinary science, the causality of disease is considered to be multifactorial.

A single, necessary etiological agent is often active, but several other factors may idluence

the occurrence and impact of disease, positively or negatively (Halgaard, 1981).

Causes of post-weaning diarrhea are often various and complex and the simple

presence of enteric pathogens is not always sunicient to produce clhical disease (Biehl

and Hoefling, 1986; Fairbrother, 1992; Dewey et a[, 1995). Therefore, in the case of

PWECD, it is strongly advised not to be content with a possible bactenological resdt

detecting some types of ETEC, but it is also necessary to consider other physiological,

environmental, dietary and viral factors that may sometimes be as important as the given

ETEC bacteria themselves (Dewey et al, 1995; Nagy and Fekete, 1999).

1.4.1. Management factors

Enteritis in the young pig caused by E. coli is generally considered to be a disease of

poor or inappropnate management. Experiments show that inspite of the presence of

ETEC 0149 in the intestines of weaned pigs in a group, when they were managed more

carefully during the post-weaning period than would be expected on a commercial pig

unit, none of them developed diarrhea (MacKinnon, 1998).

1.4.1.1. Weaning

Pigs are commonly weaned abruptly between three and four weeks of age. The

problems facing the newly-weaned pigs are: "immunity gap", deficiencies in

thermoregulation, cornpetition and aggression fiom others in the group which precipitate
stress-related physiological change, profound changes in alirnentary physiology, and the

challenge of diseases (Valery, 1995; MacKinnon, 1998).

The piglet is able to develop intestinal immunity during the fist week of Iife

(Fairbrother, 1992). At first IgM predominates, but after two to three weeks this

immunoglobulin is replaced by IgA as the most important Ig class in the intestine. Immune

incornpetence is still demonstrable at four weeks of age, when an c'imrnunological gap"

exists, that leaves the piglet vuluerable to infectious diseases. This penod of Iife

corresponds to peaks of exposure to enteric pathogens, especially in conditions of poor

hygiene. Tmmunïty is m e r reduced by competition and aggression leading to stress and

to elevated circulating cortisol levels (Svensmark et al, 1989; Valery, 1995; MacKinnon,

1998). The elevated corticosteroids after weaning have a profound influence on the

Mmune system and both humoral and cell-mediated iwnunity are compromised. This has

a negative effect on growth, feed efficiency and health status. Several stress-related

syndromes associated with weaning are now recognized in the pig and are associated with

depressed immune function (Valery, 1995).

The pH of gastric contents increases after weaning to non-bactencidal values and this

change allows hemolytic E. coli to gain access to the smail intestine. Survival rate of

potentiaüy pathogenic E. coli in the stomach of pigs is reduced at pH 3.6 and below

(Thomlinson and Lawrence, 198l), although, some other studies have shown that the

range of pH values and coliform counts varied widely between individuals, fiom 1.4 to 5.8

and log 1.O to 6.7 per g, respectively, in individual post-weaning pigs (Hampson, 1985).
 Studies made in intensively managed Danish herds (1989), found that among different

factors, the most marked influence upon occurrence of post-weaning diarrhea were age

and weight at weaning. Litiers weaned below two weeks of age had a two-fold risk of

developing diarrhea after weaning and a 2.4-fold higher mortality rate than did litters

weaned at six to seven weeks (Svensmark et al, 1989). Similady, litters weaned at an

individual piglet weight below 3 kg body weight had a three-fold higher risk of developing

diarrhea after weaning and a five-fold higher mortality rate than did pigs £rom Litters

weaned at a body weight of 7-8 kg (Svensmark et al, 1989; Skirrow et al, L 997).

Fangman et al (1998), demonstrated that pigs weaned at 16 to 21 days of age had a

significantly better daily gain in weight than the pigs weaned at 11 to 16 days of age

(Fangman et al, 1998). Weaning piglets younger than five weeks caused physiological

changes detrimental to cellular immunity, which can alter disease susceptibility in young

pigs (Valery, 1995).

Other studies showed that weaning in itself does not necessarily influence the coliform

counts in the intestine and that weaning per se is not required for induction of diarrhea

(Hampson et al, 1985; Sarmiento et al, 1988).

Another management factor that has been found associated with post-weaning diarrhea

is that: litters with diarrhea during the suckling period have increased nsk of post-weaning

diarrhea (Skirrow et al, 1997). Furthemore the incidence of the disease hcreases with

litter size at weaning but decreases with increasing herd size, probably due to better

management and production control compared to srnall traditional'herds (Svensmark et al,

1989). In addition, large populations of pigs in limited space, mixing with UnfanUlliar pigs,

37
rough handling received from stockmen, and contact with excreta, may contribute to the

presence of PWECD (Svensmark et al, 1989; Valery, 1995).

1.4.1.2. Temperature

In the absence of other risk factors, pigs in thermoneutrality are unlikely to suffer fiom

post-weaning diarrhea even when carryuig ETEC. Piglets at weaning are often weaned

into an environment below their lower critical temperature. It has been suggested that

chilling and overfeeding can elicit post-weaning diarrhea. Low temperatures reduce

peristaltic rnovement of the intestines, decrease the rate of passage of food through the

intestinal tract, encourage increased food consumption, and reduce antibody-mediated

immune responses, dowing time for multiplication and adherence of ETEC (MacKinnon,

1998; Charbonneau, 1999'). Any or al1 of these effects codd interact in the predisposition

to PWECD (Hampson, 1994).

On perEorated flooring, the weaning temperature should be 30-32°C for the fïrst 2-3

days to cornpensate for reduced feed intake. Temperatures shodd be lowered by 1OC

every two days until22OC is reached approximately three weeks post-weaning

(MacKinnon, 1998).

1.4.1.3. Sanitation

Piglets are constantly challenged by fecal pathogens, either fiom their own littermates

or fiom the rnixed groups of pigs of different litters at weaning. Without adequate

disinfection programs, this provides ideal exposure conditions to these pathogens, which

'Persona1 Communication
are enhanced by mechanical transfer on handler's equipment, clothing, hands and footwear

and by vectors such as fies and rodents (MacKinnon, 1998).

AU-in/all-out facilities provide the opportunity to clean and disinfect the premises

between groups of pigs. This will reduce the reservoir load of pathogenic entenc

organisms. This management style typically includes smaIier rooms with a narrow age

group of piglets in one room (Dewey et ai, 1995).

Environment of the weaner house appears to be the most Iikely source of hemolytic E.

coli infecting weaned pigs (Hampson et al, 1987). In a study where contamination in farm

animal buildings was measured, it was concluded that enterobacteria account for a high

level of enwonmental contamination, and cleaning and disinfecting operations must not be

considered as minor tasks and should clearly be integrated into the process of pig

production (Madec et al, 1999).

E. coli are able to survive drying for months, and they c m retain viability in pig slurry

for 5-1 1 weeks. The incidence of post-weaning diarrhea in general is greater in pigs

housed on solid floors, with or without bedding or in pens where the flooring does not

stay clean (MacKinnon, 1998). In addition, it has been shown that smooth surfaces harbor

a Iower bacterial load, that the storage of slurry close to the floor contributes to bacterial

pollution despite an dl-in/alI-out type of management, and that the reduction of the

contact between animds and slurry are the principal constituents of good husbandry

(Madec et al, 1999). It is important that weaning accommodations are designed to allow
separation of the animais fiom their feces and to d o w ease of cleansing and disinfection

(MacKinnon, 1998; Charbonneauy1999').

Even where the weaning house is not contaminated, however, unweaned pigs can

acquire infection in the farrowing house fkom previous environmental contamination, and

carry it undetected into the weaning house. Routine cleaning and disuifection of the

piggery might not prevent infection (Hampson et al, 1987).

Skirrow et al, (1997) demonstrated that not only the water source itself but aiso the

actual nipple drinker in weaner pens can each occasionally be highly contamhated with

coliforms. Cleaning and disinfection of the drinking nipple is rarely performed, but it

should be part of the routine pen cleaning procedure.

1.4.1.4. Density

Van den Broeck et al (1999a), demonstrated in studies made in Belgium, an increase of

F4-seropositivity with increasing pig farm density, indicating the importance of the swine

population density on the prevalence of the infection.

On the other hand, producers appear to have become locked into a circle of increasing

stocking densities in weaner accommodations to keep up their net margins (Vdery, 1995).

It has been s h o w that high stocking density of weaners (less than 0.42 d p i g ) can also

predispose to PWECD (Skirrow et al, 1997).

'Persona1 Communication
 1.4.2. Nutritional factors

Diet plays a role in weaning diarrhea because there is a drastic change in dietary

regimen at weaning which coincides with the initiation of diarrhea. At weaning, piglets are

rnoved into a contaminated environment, the gut is not bathed by protective antibodies

f?om sow's m i k , but possibly the most important factor is that the composition of the diet

is drastically changed (Lecce et al, 1983). At weaning, the small intestine undergoes a

reduction in villus height and an increase in crypt depth with a concornikt decrease in

enzyme activity and absorption in the small intestine. The potential for dehydration,

exacerbated by hyper-secretory diarrhea, is high. This effect varies with the type and

quantity of food eaten, so careful attention to diet formulation at weaning is essential. The

incidence of post-weaning diarrhea is considerably reduced when a high digestible diet is

fed (MaclCimon, 1998).

There are thought to be many feed factors contributhg to PWECD, and therefore feed

changes have been often instituted as part of control programs.

1.4.2.1.Carbohydrates: The presence of carbohydrates in the Iarge intestines enables

saccharolytïc flora (E. coli and other bacteria) to proliferate and produce acids.

Specifically, if they are poorly digested or absorbed may provide a substrate for E. coli

proliferation (van Beers-Schreurs et al, 1992;Friendship and Dewey, 1999).

1.4.2.2. Protein: Excess cmde protein c m act as a substrate to enhance E. co2i growth.

The presence of proteins or amino acids in the terminal ileum and large intestines allows

the proliferation of saccharo-proteolytic flora such as E. coli, resulting in acid and amine

production. These amines are also thought to increase the intake of endo- and exotoxins in

41
the intestine. Concurrent with the change in colonic bacterial flora, intraluminal osrnotic

pressure increases, leading to M e r increase in secretion and osmotic diarrhea (van Beer-

Schreurs et al, 1992).

Soybean meal has been shown to cause a transient d e r g i c reaction that damages the

intestine, contributing to malabsorption (Hampson, 1994; Friendship and Dewey, 1999).

Many researchers believe the mechanism for soybean-med hypersensitivity in pigs is a

result of ceil mediated immunity (CMI). It has been found that type IV hypersensitivity

responses to dietary antigens can bring about increases in mitotic rate in the enterocytes.

Diarrhea may result fiom this increased ce11 turnover in two ways: mild diarrhea occurs

due to the malabsorption syndrome aione reducing water absorption in the small intestine.

Tn addition, these changes may produce a great susceptibility to E.coli enterotoxin (Miller
et al, 1983).

Other studies do not support a cell-mediated component. Most recently, the

hypersensitivity response to soybean meal proteins in piglets has been classified as a Type

III hypersensitivity. The transient hypersensitivity and poor performance would be

attributed to an increase in IgG and soy protein precipitating as an immune complex in the

intestine with fixation of complement (Engie, 1994).

1.4.2.3. Minerals: High calcium levels might act as a b a e r and lead to relatively high

gastrointestinal pH which favors E. coli growth (Friendship and Dewey, 1999). The

recornmended levels of calcium according to the NRC is between 0.6 and 0.8%. High

minerai levels such as zinc level of 1500-2000 ppm c m be bactenostatic and prevent E.

coli infection.
 1.4.2.4. Fibre: Increasing the level of fibre has been reported to decrease the incidence of

colibacillosis, because it reduces the amount of substrate available in the intestinal tract,

which would otherwise be selectively utilized by hemolytic E. coli. Growth rates may be

reduced if the ration is so bullcy that the pig is unable to increase feed intake adequately to

compensate for a lower energy density (Hampson and Pethick, 1998). Increased fibre may

also af3ect pellet quality or flowability of feed through feed bins and feeders

(Charbonneau, 1999').

1.4.2.5. Feed manufacturing: Heat treatment of feed including peueting, expansion, or

extrusion has been suggested to increase the proliferation of E. coli by increasing the

amount of available nutrients required by the E. colt growth (Lawrence, 1985). The

reductions that normally occur in height of vUi, crypt ce11 production rates and brush

border enzyme activities after weaning are apparently tnggered by exposure to dry

pelleted meal at weaning (Hampson, 1994). In an increasing order of importance, pelleting

expansion and extrusion will increase the relative risk of enterobacterial infection

(Charbonneau, 1999 '). The least severe diarrhea problems have been seen in piglets fed

low solid diets with one-third the nutrient intake either hourly or three tirnes a day (Lecce

et al, 1983).

1.4.2.7. Feed management: Pigs that consume feed poorly for the first few days after

weanuig may have a tendency to engorge on feed at some later tirne, when they begin to

fïnd feed more palatable. A pig that eats a large rneal will have more feed in its stomach

'Personal Communication
than its stomach can adequately acidw. M e n the quantity of acid is not adequate the pH

of the stomach and intestinal contents will rise and create an environment more suitable

for the growth of E. coli (Charbonneau, 1999 '). Hampson and Smith (1986), found that a

major factor predisposing to diarrhea after weaning was over consumption of feed

immediately after weaning.

Feed consmption improves when fiesh feed is supplied on a more fiequent basis.

Newly weaned pigs should be fed a m;nimum of three times a day to ensure that fiesh feed

is aiways available. It is beneficial to gradually change pigs fiom one feed to another;

gradually increasing the proportion of the new phase of feed and decreasing the

proportion of the previous phase of feed to reduce the nsk of intestinal upset (van Beers-

Schreurs et al, 1998; Charbonneau, 1999').

It is suggested that pigs that consume smail amounts of creep feed pre-weaning, are

more Iikely to become sensitized to solid feed and subsequently to develop post-weaning

diarrbea than are those that consume large amounts of creep feed (Miller et al, 1984;

Skirrow et al, 1997). Adequate amounts of creep feed consumed before weaning result in

lower gastric pH values after weaning, stimulating acid production by the gastric mucosa,

a function which is nomally Iimited in young pigs (Hampson, 1985). Unfortunately, as the

sow's milk yield peaks at three weeks post-p-, there is Little incentive for piglets to

consume large arnounts of creep feed. It is important that creep feed is highly palatable

' Personal Co11110.unication

(Valery, 1995) and that it is offered in an appropriate manner (MiIler et al, 1983).

However in early weaning systems, it may be best not to offer feed prior to weaning.

Some researchers have demonstrated that withholding creep feed before three-week-

weaning has a protective effect against diarrhea after weaning (Hampson and Smith,

1986). However, other authors state that creep feeding is not required for the production

of diarrhea and that creep feeding does not induce any rnorphologic changes characteristic

of an allergic reaction in the small intestine (Sarmiento et ai, 1990).

1.4.3. Water

Elevated iron, nitrates and suiphates have been suggested to contribute to the

proliferation of E. coli. An accepted sulfate level for the pig is not adequately deked.

Some investigators believe that levels of 500 ppm or above innuence diarrhea, while

others show no concem until levels reach 1500 ppm (Davis, 1989).

Iron has been shown to nourish E. coli in ligated gut loops when the iron binding

capacity of lactoferrin is exceeded. In problern herds, iron should be evaluated in the

rations, and water supply @avis, 1989).

Water sources and delivery systems may become contarninated with disease causing

o r g ~ s m ssuch
, as E. coli. There is clear evidence that high bacterial counts in water c m

be a cause of post-weaning diarrhea. Quality of water in many herds, based on counts of

fecal contaminants and total bacteria, is poor. Producers have reported a marked decrease

in mortality due to diarrhea after installing an ultra-violet water sterilization unit. Total
colifoms and fecal coliforms should.be assessed at least yearly (Skirrow et al, 1997;

Charbonneau, 1999').

1.4.4. Changes in the intestinal flora, physiology and morphology

The consumption of mille by the nursing pig results in the growth of mainly

Lacto bacillus spp. bacteria group in the stomach and intestinal tract. These bacteria use

some of the lactose in milk to produce lactic acid, which lowers the pH in the stomach.

This acidic condition aids the digestion process and prevents the growth of other

microorganisms. When miik products are eliminated fiom the starter diet, these bactena

are quickly reduced in number, whereupon other bactena become established in the

intestinal tract. If pathogenic bactena are allowed to predominate during this transition

period, diarrhea c m occur (Hill et al, 1998).

McAlliester et al (1 979), demonstrated that non-hemolytic E. coli and Sn-eptococciare

the major aerobes in the entire intestine ofnursing pigs. Anaerobes were ten fold more

numerous than aerobes in the lower intestine. After weaning, the population of aerobes,

Lactobacilli and Bacteroides-Clostridia decline to about one tenth the pre-weaning level in

most segments of the intestine. In cases of diarrhea, the population of hemolytic E. coli

increases throughout the srnall intestine. They are the only organisms that proliferate

markedly beyond the pre-weaning population and c m respond to the new environment

created by dietary change (McAlliester et al, 1979).

'Personal Communication
 Nutrients are absorbed in the small intestine through elongated villi. Upon weaning, the

villi become shorter in length, and the digestive enzymes fkom the villi are temporarily lost.

Consequently, when the villi length becomes shortened, the absorptive area and the overall

digestive capacity of the weaned pig is reduced (Hill et a., 1998). Many authors agree that

changes such as villous atrophy, increased crypt depth and increased epithelial ce11 rnitosis

after weaning are involved in the pathogenesis of diarrhea (van Beers-Shreurs et al, 1992;

Nabuurs, 2998). It has also been observed, that the immature gut exhibits an increased

sensitivity to toxins, and that bacterial toxin receptor and effector responses depends on

the stage of maturity of the epitheliai cells (Nabuurs, 1998). Other authors associate the

changes of the intestine m a d y to a change of diet (van Beers-Schreurs et al, 1992).

1.4.5. Genetic factors

Genetic susceptibility also is a factor in post-weaning E. coli infection. Not al1 pigs are

susceptible to K88-positive E. coli. Innate resistance to infection caused by these

organisms is inherited via an autosomal recessive gene, and has been correlated with lack

of receptors on their epithelial cell bnish border, to which the fimbriae bind. Resistance to

colonization is not comrnon, since the presence of the receptors is controlled by an

autosomal dominant gene, Therefore based on the principle of Mendelian ïnheritance, only

the genotype ss pigs are resistant to K88 (F4) isolates (Davis, 1989; Hampson, 1994;

Michael et al, 1994; Baker et al, 1997).

Binding specificity for each K88 (F4) phenotype has been described: A (adhesive to d l

three variants), B (adhesive to K88ab and K88ac), C (adhesive to K88ab and K88ad), D

(adhesive to K88ad), E (nonadhesive) (van den Broeck et al, 2000)and F (adhesive to

47
K88ab) (Gyles, 2001). Baker et al (1997), reported that the first five phenotypes are

distributed among breeds of swioe in the United States. As only the E. coli expressing the

K88ac (F4ac) variant appear to be of significance in PWECD, o d y pigs of phenotype A

and B appear to be at nsk to infection. In this study, 48% of the pigs were phenotypes A

and B, suggesting that susceptible population must be very common (Baker et ai, 1997).

Pigs of the Chinese Meishan breed have been reported to be uniformly of the non-adhesive

(E) phenotype (Michaels et al, 1994; Baker et al, 1997).

1.4.6. Concurrent infections

Infection caused by other agents in the herd tend to intensify in weaner

accommodation. Gut infections caused by rotavim, Transmissible Gastroenteritis (TGE)

and bacteria other than E. coli will be important pre-disposing factors in post-weaning

diarrhea. Lung infections caused by Porcine Reproductive and Respiratory Syndrome

"Ns(PRRSv), Mycoplasma hyopneumoniae, influenza virus and various other vinises

and bacteria will also add to the problems of homeostasis (MacKinnon, 1998).

1.4.6.1. Porcine Reproductive and Respiratory Syndrome: An increase in secondary

bacterial infections on PRRS prevalent f m s has been noticed. Pasteurella multocida,

Huernophilus parasuis, Streptococcus suis, Sheptococcous spp, Salmonella cholerasuis,

Salmonella spp, Actinobacillus plenropneumonia and Mycoplasma hyorhinis have been

reported as the microorganisms which caused severe secondary infections (Zimmerman et

al, 1998). E. coli has aiso been regarded as one of the important agents involved in

secondary infections (Done and Patron, 1995). A dual infection with ETEC and PRRS

virus in weaning pigs that died suddenly was reported by Nakamine et al (1998). PRRS

48
virus was demonstrated in the lungs of every sudden death pig that showed dianhea and

was proved to be active among those pigs. In this study it was suggested that following a

decline of passive immunity, infection with PRRS was activated and ETEC became

systemic, causing the peracute death in nursery pigs.

UnpubIished observations made by Nagy et al, have shown that if PRRS virus is

introduced into a hitherto uninfected herd, neonatal andor post-weaning colibacillosis WU

be triggered. In such herds, prevention measures should include vaccination and

management techniques not oniy against ETEC but also against PRRS to suppress this

viral infection on a herd basis (Nagy and Fekete, 1999).

1.4.6.2. Transmissible Gastroenteritis: An endemic form of TGE disease affecting

sucklùig and recently weaned pigs, which may be difficult to diagnose because of the

production of mild clinical signs and low mortality rate, has been reported. Among

endemic TGE cases, only 25% of pigs display the typical histologie lesions. Often the

postmortem findings are more suggestive of enterotoxigenic E. coli than TGE. In fact,

some studies have demonstrated a dual enterotoxigenic E. coli and TGE infection, that

probably contributes to a more severe diarrhea problem in individual pigs (Pritchard, 1987;

Dewey et al, 1999).

Gradzki et al (1996), also reported a concurrent infection with TGE, rotavirus and E.

coli in post-weaning pigs in Poland. In routine bacteriological examination, a pure culture

of hemolytic E. coli was isolated, which was resistant to most of the antibiotics used in

veterinary therapy. Infection with TGE was also confimed.

1.4.63. Rotavirus: The v i n s has been implicated in some outbreaks of PWECD,

producing a more severe diarrhea than that resulting fkom infection with each agent

separately (Gyles, 1986; Bertschinger, 1999). In this concurrent infection, age at weaning

may also be important since piglets weaned at two weeks of age have a much higher

prevdence of rotavirus excretion than do those weaned at four to five weeks of age

(Hampson, 1994).

Rotavirus not only cause villus atrophy but aiso interferes with the normal clearing

mechanisms of the gut either by non-specifically inhibithg protective mechanisms against

gut coionization or by having a specific effect, such as altering the binding sites on

enterocytes so that enterotoxigenic E. coZi can proliferate (van Beers-Schreurs et al, 1992;

Bertschinger, 1999).
1.5 Control and prevention of post-weaning E, coli diarrhea

Having stated that E. coli infections are generdy diseases of management, it is

essential that good husbandry practices are foilowed to control the disease (Mackinnon,

1998). There are two broad approaches for the prevention of PWECD. The first method

attempts to minimize factors that predispose to diarrhea, and the second approach is more

specifïcaily directed at the E. d i .

1.5-1. Control of factors that predispose pigs to diarrhea

Factors that need to be taken into consideration to prevent PWECD, are:

1. The target of average weaning weight should be 6-8 kg at 24 to 30 days (Davis, 1989;

Skirrow et al, 1997; MacKinnon, 1998).

2. Handle pigs with care during weaning and do not wean weak pigs (MacKuuion, 1998).

3. Thermoregulation plays an important role in the prevention of colibacillosis. The

temperature in the weaner house at weaning should start in the range 28-32OC, with

minunal temperature fluctuations, and no cirafts (Fairbroder, 1992; Harnpson, 1994;

MacKinnon, 1998). Recording the temperature twice a day can ofien reveal problems

overlooked by casuai observation. Temperature shoutd be monitored in confined housing

using at least one maximum/minimum thermometer in each nursery space (Davis, 1989).

4. Stress responses should be minimized by weaning pigs into small groups with as little

mixing as possibIe (Valery, 1995).

5. Pigs should have fiee access to clean water which implies that drinkers should not be

prone to contamination by excreta and they should be supplied by a seded system. The

use of sanitizers is recommended @avis, 1989; MacKinnon, 1998).

6.Pigs should be fed little and ofien for 2-3 days post-weaning to mirnic the communal

feeding habits of the suckling period (Davis, 1989; MacKinnon, 1998).

7. Weaning accommodations should be operated on an ail-in ail-out basis with adequate

tune allowed for washing and drying between batches (Davis, 1989; MacKinnon, 1998).

8. House the pigs in small groups of eight to ten pigs (Davis, 1989).

9. It may be necessary to reseict dietary intake in the f i t few days d e r weaning and

provide fibre in the diet (Hampson, 1994). Dietetic prevention is difficult to practice

because the degree of protein reduction in weaner feeds can be detrimental. If restriction

goes too far, spontaneous spread of the organism is reduced to a degree where active

immunity does not develop. A low nutrient content of the feed is expensive since one

additional week fiom weaning to market weight is required (Bertschinger, 1996;

Hampson, 1994).

1.5.2. Factors directed at E. coli

There are a number of strategies available that are aimed directly at reducing the build-

up of numbers of hernolytic E. coli in the intestine after weaning (Hampson, 1994). A

variety of ingredients have been added to the creep and starter diets to control enteritis

caused by post-weaning proliferation of E. coli, including: probiotics, antibiotics,

inorganic chemicals and organic acids (MacKuinon, 1998).

1.5.2.1. Growth promoters: Low levels of prophylactic antimicrobials are used in many

piggeries, which have been shown to reduce the adhesive effect of E. coli in-vitro by

alterhg the shape of the bacterial cell, releasing the adhesins fkom the ce11 surface,

inducing the formation of functionally deficient adhesins or suppressing the synthesis of

adhesins (MacKinnon, 1998). This routine encourages developrnent of drug resistance,

and indeed, the spectnim of resistance in many PWECD strains is alanning.

1.5.2.2. Probiotics: The value of probiotics is so far limited to specdation (Bertschinger,

1996). The exact working mechanism of probiotics is stil unclear, though different

assumptions have been made: inactivation of bacterial toxins or metabolites; production of

bacteriocines; reduction of the intestinal pH by lactic acid production; cornpetition for

intestinal receptors and a stimulation of the non-specific irnmunity of the host are said to

play a major role (Cupere et al, 1992; Conway, 2999).

In an experiment done by Cupere et al (1992), the use of LactobaciZZus spp..

Streptococczisfaecium or Bacillus cereus, did not proved to be helpful when

supplemented in the feed of recently weaned piglets. In this study, the use of these

probiotics did not prevent mortality and clùucal symptoms nor reduce the fecal excretion

of hemolytic E. coli.

Spencer and Chesson (1 994) showed that Lactobacillus strains, selected for their

capacity to adhere to jejunal enterocytes, were unable to prevent or reduce the adherence

of 0149K88 E. coli strains. On the other hand, studies made in-vitro, demonstrated that

Lactobacillus strains release a proteinaceous component which affects mucus, such that

the adhesion of the K88ab (F4ab) and K88ac (F4ac) was reduced (Blomberg et al, 1993).

53
It has been suggested that probiotics to be effective, must be given eariy in life, must have

the ability to colonize the gut and must remain there as a barrier. Results obtained in older

pigs are much less satisQing. Probiotics are often compromised by the use of therapeutic

levels of antibiotics to control a variety of PWECD (Cupere et al, 1992; MacKinnon,

1998).

1.5.2.3. Zinc oxide: Zinc oxide is widely used to control post-weaning colibacillosis at an

inclusion level that provides up to 2500 to 3 100 ppm zinc- The mode of action is unclear,

but it has been suggested to be rnediated through the inhibition of the bacterial respiratory

c h a h oxidase system or by disrupting enzyme activity in the cell wall of enteric bacteria

(Wood, 1991; MacKirinon, 1998; Bertschinger, 1999).

Huang et al (1999), demonstrated that supplementing the nursery diet with 3000 ppm

Zn0 reduced bacterial translocation fiom the small intestine to the corresponding lymph

nodes, and subsequently to the blood and interna1 organs in weaned pigs. The mechanism

involved in achieving this effect was not determined. It was suggested that zinc has a

primary effect on tissues with a high tumover rate as in those of the gastrointestinal and

immune system. Zinc is needed in these tissues for DNA and protein synthesis. Zinc

stabilizes the membrane structure and may modie membrane functions, and protect

membranes fi-om the effect of infectious agents (Huang et al, 1999).

Another study showed that feeding p h m a c o logical concentrations of zinc oxide

(3,000 ppm) enhanced growth in traditionally weaned (>2 1d) and early weaned (44d)

pigs, during at least the first two weeks after weaning. Pharmacological concentrations of
zinc stimulate metallothionein synthesis in mucosal ceils, which regdate zinc uptake into

the body, resulting in the irnproved growth observed in weaning pigs (Carlson et al, 1999).

Zinc oxide is virtualiy insoluble at the pHs encountered in the gut, is not absorbed fkom

the alimentary tract, and so, appears eventually in the slwry. Zinc oxide at the high level

advocated is also unpalatable and has a limiting effect on feed intake. It is claimed that the

levels recommended do not give rise to any toxicity in the pig or the consumer of pig meat

(Wood, 199 1;MacKinnon, 2 998; Bertschinger, 1999).

The danger of using zinc oxide is the potential damage to the environment caused by

depositing a heavy metal onto land in spreading effluent fiom such pigs. The application of

manure containing zinc oxide to g r a s and arable land will not immediately cause any

problems; but zinc remains in the topsoil and may eventually reach levels that affect plant

growth. Its elimination fiom the soi1 is by plant uptake, which is a very slow process. Zinc

in zinc-laden sluny is not likely to reach ground water. However, there is a concem that

surface run-off might contaminate water courses and rivers (Wood, 199 1).

1.5.2.4. Organic acids: Little evidence in general has been forthcoming about the mode

of action or the efncacy of in-feed or in-water organic acids (MacKinnon, 1998).

Reduction in gasûic acidity and survival of ingested E. coli are believed to be the main

benefits f?om their use (Thomlinson and Lawrence, 1981). This procedure has variable

results, and may delay the onset of PWECD rather than prevent it completely.

The pH of the stomach contents falls after weaning but several investigators have

found that the pH of the jejunum cannot be reduced by acidification of the feed. The pH

close to the jejunal brush border is slightly acid and highly regulated. It is not influenced

55
by the pH of the chime (Bertschulger, 1999). Acidifïed diets have so far not been proven

to reduce mortality (Bertschinger, 1996).

1.5.2.5. Bromelain: Bromelain is a proteolytic extract obtained fkom pineapple stems.

Oral administration of the protease bromelain has shown to inhibit the F4ac receptor

activity in a dose dependent manner. It can inactivate ETEC receptor activity in vivo and

protect against ETEC induced diarrhea. Bromelain rnay therefore be an effective

prophylactic agent against ETEC infection (Mynott et al, 1996; Chandler and Mynott,

1998).

1.5.2.6. Egg y o k chicken: Egg yolk containing specific ETEC antibodies fed to post-

weaning pigs has been suggested to be good for prophylactic use (Nagy and Fekete,

1999). Antibodies are obtained by immunking hens with purîfied F4 or FI8 fknbnae.

Antibodies are M e r isolated fkom diluted egg yolks by ammonium sulfate precipitation,

and are added to the feed of weaned piglets. It has been shown that egg y o k antibodies

duninish the cases of diarrhea and death in animals infected with F4 and F18 positive E.

coli (Imberechts et al, 1997).

1.5.3. Treatment of post-weaning E. coli diarrhea

Treatment of PTNECD is based on oral antibiotics (preferentially through water) and on

fluid/electrolyte replacement, together with attention to management, hygiene and

immunity (MacKinnon, 1998; Nagy and Fekete, 1999). When choosing feed as a drug

vehicle, it should be rernembered that pigs eat less after weaning, and feed intake may Vary

greatly among pigs within the same group (Nagy and Fekete, 1999).
 Parenteral treatment pennits individual medication. Substances must be selected which

reach the intestinal lumen, such as cephaIosporins, or trimethroprim- Testing bacterial

resistance is indispensable if there is a herd probiem (Bertschinger, 1999).

Either commercial or clinic prepared electrolyte solutions are relatively inexpensive and

can be used for an initiai complaint or on a routine basis to prevent metabolic acidosis,

dehydration, and the resulting circulatory failure (Davis, 1989). Rehydration nuid should

be offered for spontaneous intake or injected intraperitoneally ifthe pig is anorectic. Such

fluids may contain glucose, glycine, citric acid, and potassium dihydrogen phosphate in an

isotonic solution. Uptake should be equal to the loss (Bertschinger, 1999).

1.5.4. Antibiotic resistance

Preventive antimicrobial treatments should be Iimited to herds where al1 other methods

have failed (Bertschinger, 1999). Bacteria use several different mechanisms to resist the

action of antibiotics. Once a resistance mechanism is established it may easily spread to

other bacteria, even to other species of bacteria. The most important mechanisms for this

dissemination of resistance genes between bacteria are plasmids and transposons.

Plasmids, DNA molecules existing more or less independently of the chromosome in the

bacterial cell, often carry antibiotic resistance genes or other genes which may confer a

selective advantage on the bacterium. A transposon which is a small, highly mobile DNA

element located on a plasrnid or on the chromosome, may carry al1types of antibiotic

resistance genes as well (White, 2000). Recent research has shown that after many

generations, a bacterial population adapts to maximize its ability of survival, keeping their

resistance genes.
 Isolates fiom PWECD and edema disease (ED) show the highest rate of resistance

within porcine E. coli (MacKinnon, 1998; Bertschinger, 1999). However, it is not possible

to give universal data on resistance, because the situation varies in different pig

populations depending on the antirnicrobial agents preferentially used (Bertschinger,

1999).

Patterns of antibiotic resistance are mainly dependent on age of pig and level of

antibiotic used on the f m s . An increased incidence of resistance is noted in nursery pigs.

This may be a reflection of an increase in antibiotic use at that time, but aiso it may reflect

an increased colonization by pathogens that occurs post-weaning, in which resistance may

occur more commonly than in commensal organisms and because bacteria in the intestinal

tract of post-weaning pigs have the increased potential for resistance transfer (Mathew et

al, 1998, 1999). On f m s where routine sub-therapeutic feed-based antibiotics and or

injectable antibiotics are used, the incidence of multiple resistance for E. coli in pigs at 35

days of age is greater (Mathew et al, 1999).

Dunlop et al (1996) found that, in general, increased risk of resistance among E. coli

was associated with the use of various in-feed antimicrobials, while individual animal

treatment was not, except for the use of gentamicin. In addition, he showed that there is a

consistent selection of resistant organism when animals are exposed to other

antimicrobials, concluding that there is an effect of indirect selection of resistance fkom

other antirnicrobial uses (Dunlop et al, 1996). A substance continuously used in a weaner

facility exerts a tremendous selective pressure. As a consequence, categories of

antimlcrobials, have already lost or are on the point of losing their activity against porcine

E. coli (Bertschinger, 1999).

In Canada, a wide variety of antibiotics are included routinely in weaner diets at sub-

therapeutic levels, including carbadox, penicillin, spectinomycin, sulfonarnides and

tetracyclines (Dunlop et al, 1998). Studies made in Ontario, have shown that

antimicrobial usage in post-weaning pigs varied enormously among swine operations,

suggesting that producers rnay be able to efficiently raise swine with fewer antibiotic

treatments (Dunlop et al, 1996, 1998).

In-vitro sensitivity of E. coli isolates to a wide range of antimicrobial agents has greatly

increased over the last several years. Over a decade ago most E. coli isolates were

sensitive to gentamicin and the trimethoprim/sulfamethoxazolecombination, however

recent studies in Quebec have shown a substantial Ievel of resistance. Resistance to

cephalothin and gentamicin fias steadily increased over the last four years. Most isolates

showed sensitivity to amikacin and apramycin (Fairbrother, 1999a). Apramcyn (Apralan)

remains the antimicrobial of choice for use in feed or in water. Enrofloxacin is the

treatment of choice in different countries for use parenterally after weaning, but it is not

commonly used in Canada (MacKinnon, 1998).

In a survey study developed in the UK (1993), a high resistance was found to

apramycin in pigs of al1 ages and humans, including hospital patients and also a pig

worker. Resistance to apramycin provides cross-resistance with other arninoglycosides

such as gentamicin and tobramycin. In addition, apramycin resistant E. coli were found to
persist in a dry environment in a pig pen which had been empty for 10 months (Hmter,

1993).

From more than 100 E. coli isolates recovered fiom piglets with neonatal and post-

weanulg diarrhea in North Dakota, antimicrobid susceptibility revealed that all E. coli

isolates were susceptible to amikacin, ceftriaxone, and ciprofloxacin. However, increased

levels of resistance to ampicillin, chloramphenicol, kanamycin, streptomycin,

sulfamethoxazole, and tetracycline were observed. Multiple antimicrobial resistance was

observed in 94% of E. coli isolates assayed (White, 2000).

1.6. Vaccination as an aid in controlling post-weaning E. coli diarrhea

Immunization of the sow using comrnerciaLly available vaccines may effectively control

neonatal diarrhea but not the post-weaning diarrhea or edema disease. Development of

multiple bacterial resistance to a wide range of commonly used antibiotic and a recent

increase in the prevalence of the post-weaning syndrome will necessitate the use of

alternative measures, such as piglet vaccination, for their control (Fairbrother, 1999b).

1.6.1. Immunity system of the porcine gut

The gastrointestinal tract is presented with a vast array of antigens fiom potential

pathogens to dietary antigens, and the type of response they elicit is critical, since failure

to respond to the former may lead to infectious disease whilst an inappropnate response to

the latter wiU be wastefbl and may lead to immuno-pathological damage. However, the

response to either type of antigen may be influenced by the other (Stokes, 1988).

Local intestinal immune system operates largely independently of systernic immunity

and which is best stunulated by local application of antigen. In order to prevent disease,

the prirnary role of mucosal defense is to exclude foreign material. The mucosal immune

system provides a number of rnechanisms whereby pathogenic microorganisms can be

expelled or eliminated, this is accomplished with the aid of number of non-specïfîc defense

mechanisms such as, gastric pH, mucus, peristalsis, intestinal flora, lactoferrin and

interferon (Stokes, 198 8).

1.6.1.1. Antibody mediated defense

Olsson et al (1986), reported that in pigs orally infected on the first day of life with a

virulent E. coli 0149X88 strain, anti-lipopolysaccharide antibodies of the IgA and IgG

classes were detected at day 7 of life in most of the piglets infected, whereas anti-K88

antibodies of the IgG and IgM classes predominated. In 21 day old pigs, antibodies of all

immunoglobulh classes were produced. Oral challenge did not stimulate systemic immune

response. In piglets of greater than three weeks of age, challenge with E. coli produced an

antibody response mainly associated with antibodies of the IgA class (Olsson et al, 1986).

The most important protective humoral immune factors at mucosal levels are locally

produced by antibodies of the secretory IgA isotope. This isotope constitutes 80% of al1

antibodies produced in mucosa-associated tissues (Laval, 1996). In most species it is the

predorninant immunoglobulin in milk and other secretions (Stokes, 1988). Ingested

antigens can be taken up into specialized lymph-reticular tissues; known as gut-associated

lymphoid tissues (GALT) (Laval, 1996; Nagy and Fekete, 1999). GALT is collectively

represented by the Peyer's patches, the appendix and srnall solitary lymphoid nodules.

Payer's patches contain a dome region enriched for lymphocytes, macrophages and plasma

celis. This dome is covered by a unique epithelium with cells called " M cells, which take

up the antigen fiom the lumen. Antigen uptake does not result in degradation, but of

delivery of intact antigen into the underlying lymphoid tissue. B ce11 follicles occur beneath

the dome. Adjacent to these follicles are the T-cell dependent areas. The development of

lymphocytes within the follicles of the Peyer's patches is followed by a migratory phase.

During the migration, maturing lymphocytes travel via the circulation to return either to
the gut or, in the pregnant or Iactating sow, to the mammary gland. Stimuiated B

lymphocytes secrete antibody, principaily IgA, which is selectively transported across the

gut epithelium into the mucus layer. IgA acts by trapping antigen within the mucus layer

and preventing its absorption and colonization of the gut surface. It enhances the

stimulation of more active weapons of the immune system. These mechanisms explain how

vaccination of the sow induces an immune response and subsequently increases the

production of IgA by the udder at farrowing. Piglets will thus be protected by mik

throughout the suckling period. This passive protection stops on the f i s t day of weaning

and must not be overlooked in the case of early weaning (Aizpunia and Russell-Jones,

1988; Laval, 1996).

IgA is relatively resistant to proteolysis and becomes associated with the mucus layer.

Secretory IgA has also an anti-idammatory effect, and a bacteriostatic effect, reducing

the efficiency with which the bacteria can utilize iron and so increasing the effectiveness of

lactoferrin-mediated iron sequestration. IgA may also block the specific receptor sites for

enterotoxins and hence neutralize their activity (Stokes, 1988; Laval, 1996; Fairbrother,

1999b).

Bertschinger et al (2000), demonstrated that in mice, rats and rabbits, specific IgA in

the intestinal lumen enhances the binding of antigens to M cells and the transfer to

contiguous cells of the immune system. A similar mechanism may facilitate development of

mucosal immunity in oral vaccinated pigs in an endemically infected herd (Bertschinger et

al, 2000).
 IgM and IgG antibodies also appear in intestinal secretions to a lesser extent (Laval,

1996). In swine, IgM can play an important role in mucosal immunity as secretory IgM

(Olsson et aI, 1986; Van den Broeck et al, 1999~).In the pig, IgG is only a minor

component in gut secretions where it appears as a result of passive transudation rather

than active transport (Laval, 1996).

1.6.1.2, CeU mediated defense

Ce11 mediated mechanisms may also play a significant role in local immune defense.

Peyer's patches also supply T cells which populate the intestinal mucosa (Stokes, 1988).

Functional activities including cytotoxicity, delayed type hypersensitivity, and natural killer

activity have been ascribed to them. Following infection, T-lymphocytes are capable of

producing lymphokines and these may contribute to gut defense increasing the rate of

division of crypt cells and recruiting non-specïfic effector cells such as macrophages,

natural killer cells and mucosal mast cells (Stokes, 1988). Large numbers of neutrophils

have also been demonstrated in the intestinal lumen in response to specific immune

challenge, and they have been shown to be capable of killing bacteria. Rose and Moon,

(1985), showed that when the small intestine of a neonatal piglet is colonized with ETEC

E. d i , enteroluminal neutrophils are seen in significant numbers.

1.6.2. Immunization

Passive imrnunity derived fkom colostrum prevents the immune system of a pre-weaned

pig fiom being overwhelmed by more pathogens than it can handle at one time, but the

development of active immunity is essential for the survival of older animals (Fangman et

al, 1998).
 The major class of virulence factors that have been exploited as candidates for

vaccination against E. coli are: pihs-adhesins and enterotoxins (Isaacson, 1994). As these

antigens play an important role in the pathogenesis of ETEC-induced neonatal and post-

weaning diarrhea, they constitute major components in vaccines (Van den Broeck et al,

1999~).Bianchi et al (1996), showed that in mice, oral immunization with a live bacteria

induced an enteric immune response against the E. coli K88ac-positive fimbriae. More

recently, Van den Broeck et al (1999c), demonstrated that purified F4 fimbriae is a strong

antigen for oral as well as systemic immunizations of pigs.

1.6.3. Passive immunity

PigIets are born without matemal, tramplacental immunity and they have not mature

immu11oglobulins of their own in the serum (Bijlsma, et al., 1987). Since neonatal E. coli

diarrhea occurs at such an early age, it is not possible to induce active immunity in the

animals at risk and the vaccination strategy that has been adopted is to induce passive

ïmmunity through the stimulation of lacteal immunity in the sow (Bijlsma et al, 1987;

Isaacson, 1994; van den Broeck et al, 1999 b,c).

One of the early approaches consisted of feeding the sow at the end of gestation with a

bacterial culture onginating from the intestinal contents of piglets with diarrhea. This

method, which is being used less and less, results in a very good protection during the

suckling penod (Fairbrother, 1999b).

Currently,comrnercially available vaccines are used, and are directed against serovars.

These vaccines consist of inactivated whole bacteria (bacterins) or of fimbriae which have

been more or less purified (sub-unit vaccines). The use of recombinant DNA technology

65
has resulted in the production of greater quanttities of fïmbriae. (Moon and Bunn, 1993;

Fairbrother, 1999b; Nagy and Fekete, 1999). Osek et al (1995), in a study where

evaluation of different vaccines to control pig colibacillosis was made, demonstrated that

the best results are obtained when pregnant sows are immunized with a vaccine containing

F4, F5, F6 and the subunit B of the LT toxin.

Vaccines are administered by the parenteral (intra-muscular or sub-cutaneous) or oral

route at the end of the gestation. Two administrations, at six and two weeks before

parturition, are necessary for immunization of the gilt, and a booster is necessary at each

subsequent gestation. Piglets suckiing sows given parenterai vaccines receive, transiently,

levels of IgG antibodies. In contras, animals suckling orally immunized sows receive a

continuing supply of IgA protective antibodies in both colostrum and milk (Moon and

Bunn, 1993; Morona, 1994; Fairbrother, 1999b).

1.6.4. Active irnmunity of the weaned pig for the control of PWECD

After being weaned, the pigs are deprived of the passive protection and becorne

susceptible to ETEC infections. At that moment, an active mucosal immunity is required

to evoke a protective imrnunity in which the production of secretory IgA antibodies plays

an important role (Van den Broeck et al, 1999 b,c).

It has been demonstrated that oral administration of solubilized purined F4 b b r i a e

induces an intestinal mucosal immune response in F4 receptor-positive (F4R+) piglets

(Van den Broeck et al, 1999b). However, for the success of the development of an

effective vaccine against weaning diarrhea, some considerations need to be taken:

1. Identification of the important adhesins and vinilence factors for the strains involved in

PWECD. ETEC fimbriae apparently do not have cross-protective epitopes and therefore,

vaccines are more usefûi if they contain the h b r i a e antigen prevalent in the target host

species (Moon and Bunn, 1993).

2. A means to stimulate a local immune response. Since the disease is caused by strains

that colonize the mucosal surface of smail intestines, and secrete an enterotoxin, it is

necessary to stimulate a local immune response (Isaacson, 1994).

3. The ability to stimulate this response during the narrow window of time between birth

and weaning (Isaacson, 1994). Pigs are generally weaned three to four weeks after birth,

with the trend towards earlier weaning. This limits the period of time to elicit a protective,

active immune response in these animais.

4. In order to elicit a protective immune response against organisrns that cause weaning
9

diarrhea, vaccination would have to occur shortly after birth. If the dam has been exposed

to ETEC that cause weaning diarrhea, it is likely that ETEC specific antibodies will be

present in the colostrum and rnilk. The presence of these antibodies may suppress an

immune response in the weaning pigs. This may be due to the reaction of vaccine antigen

with matemal antibodies and rapid clearance fiom the intestines, thereby preventing

processing by cells of the immune system. In additon, exposure to neonates to antigens in

the vaccine during the first day or so of life may further suppress the ability to elicit an

appropriate antibody response to these antigens and result in immune tolerance (Isaacson,

1994; Laval, 1996).

1.6.4.1, Oral immunization

Oral vaccination of the piglet stimulates the mucosal immune system to produce

secretory antibodies (Stokes, 1988; Isaacson, 1994, Lavai, 1996),and can be performed

with live ETEC or inactivated antigens. Oral imrnunization with killed bacteria induces

very little mucosal immune response. Attenuated strains or live recombinant vaccines may

take a variety of foms and have been used against a number of agents (Stokes, 1988).

Novel vaccines are currently available for enteric diseases. The development of these

vaccines has been focused on the production of antigens through genetic engineering

(Laval, 1996; Fairbrother, 1999b; Turnes et al, 1999).

Inactivated oral vaccines

The response to inactivated oral vaccines is highly dose dependent and feeding

extremely large numbers of organisms for several days is required to induce protection.

The vaccine is very expensive to produce and is not able to stimulate imrnunological

memory (Stokes, 1988; Moon and Bunn, 1993; Laval 1996; Fairbrother, 1999b).

Furthemore, it has been demonstrated that antigen presented in this way via normal villus

epithelial cells in the intestine, triggers a predomïnantly suppressive response.

Consequently, to generate an effective and protective immune response, the use of

powerful adjuvants is advised (Laval, 1996). Many authors agree that oral vaccines

consisting of whole inactivated bacteria do not seem to be effective for imrnuiization of

piglets (Moon and B u , 1993; Fairbrother, 1999b).

However, studies have reported positive results. Porter et al (1974), compared an oral

0149 killed strain incorporated in feed and an intramuscular heat killed 0149 strain. Pig

68
feeding on the antigen supplemented diet showed improved weight gains and feed

conversion ratios. In addition, signifîcantly fewer animals were observed with diarrhea

and there was also a reduction in the requirement for antibiotic therapy post-weaning, as

well as a reduction in numbers and proportion of hemolytic E. coli in the feces during the

two weeks po st-weaning. Parenteral immunization did not provide comparable benefits

(Porter et al, 1974).

Live oral vaccines

Live vaccines are more effective as they are able to induce an active and durable immune

response in three-week-old piglets (Moon and Bunn, 1993; Lavai, 1996). However, it

needs to be considered that they are potentially dangerous as vaccinated aoimals might

spread the live vaccine strain to unvaccinated contact animals. In addition, the genetic

stability of the strains must be strictly evaluated to avoid reversion to virulence (Laval,

1996). Recently, efforts have focused on development of fimbnai based vaccines to

prevent post-weaning ETEC infections in swine. Such vaccine would ideally:

1. Temporarïly colonize the small intestine of the s u c k h g pig in spite of the antifimbrial

antibody ingested with miik.

2. Not produce enterotoxin or other substances that would adversely affect the health and

productivity of the colonized piglet.

3 Stimulate the mucosal immune system of the (3 week old pig to secrete protective levels

of antifimbrial IgA into the small intestine.

4. Reflect the h b r i a i antigens prevalent among ETEC associated with post-weaning

diarrhea (Moon and Bunn, 1993).

In Australia, experiments have had success in controlling, and reducing the severity of

PWECD with the used of live oral autogenous E. coli vaccines given before weaning

(Hampson, 1994). Further experiments have provided evidence about the protection of

pigs against PWECD and edema disease by a iive oral vaccine containing F18 h b r i a e

(Sarrazin and Bertschinge- 1997; Bertschinger et aI, 2000).

Van den Broeck et al (1999b), demonstrated that the presence of the antigen-specific

receptors on brush borders of villous enterocytes is a prerequisite for the induction of an

immune response following oral immuniation with purified F4 h b r i a e . Oral F4

vaccination of F4R+ (receptor positive) pigs induced an immune reaction with a clear

increase of F4-specific antibodies in s e m . Moreover, both IgG and IgA responses were

enhanced by a second vaccination 16 days Iater. In contrast F4-specific antibodies were

not detected in semm or feces of F4R- (receptor negative) animals following the two

vaccinations. In the same study, oral vaccination with purihed F4 prevented colonization

of the intestine by virulent F4+ ETEC infections (Van den Broeck et al, 1999b).

In another study, the same group found that after four days of an oral immunization to

pigs with purified F4 b b r i a e , without previous experimental priming, a large number of

antigenic-specific antibody-secreting cells (ASC) were present in Peyer's patches.

Subsequently, the amount of ASC increased in the mesenteric lymph nodes, migrating

subsequently into the blood. After recirculating, a decrease of ASC in blood and an

increase in different lyrnphoid tissues is recognized. So it is assumable that ASC leave the

70
circulation and home into lymphoid tissues. In this case, lamina propia serves as a mucosal

effector site, but ASC are also present in spleen and bone marrow, which are systemic

effector sites. At that moment, also a high number of F4 specific IgM, IgA and IgG ASC

are seen in the Peyer's patches (Van den Broeck et al, 1999~).

Genetically altered vaccines

Oral imrnunization of one to two week old piglets with the so-cailed genetic

immunization are currently being investigated (Fairbrother, 1999b). The use of these live

avinilent E. coli with vectors carrying genes that code antigenic proteins, appears as the

most novel approach in vaccinoIogy. These vaccines show the advantages of replicating

antigens (live vaccines), but not their disadvantages. Genetic vaccines have proved to be

immunogenic against severai vinises, parasites, intraceliular bacteria and bacterial toxins

(Laval, 1996; Turnes et al, 1999; Fairbrother, 1999b). This techniques also gives the

opportunity to produce attenuated antigens in large amounts (Laval, 1996).

In a study where an experimental attenuated vaccine for ETEC in humans was used, a

substantial protection against experimental challenge inoculation with pathogenic ETEC

was observed. Francis and Wiilgohs (1991), tested various avirulent plasmid constnicted

E. coli vaccine strains on four-week-old pigs and showed that pigs vaccinated with an

LT+,K88- remained unprotected, as did the control group, and that pigs vaccinated with

and LTb-,K88+ appeared less protected than those vaccinated with an LT+,K88+. It was

demonstrated that protection appeared largely attributable to immunologie stimulation by

K88 fimbriae.
 In Australia, Faghy et al (1987), had success in controliing and reducing the seventy of

PWECD in pigs. In this country, these vaccines are commercidly available and used on

regular b a i s in many f m s with very good resuits.

1.6.4.2, Parenteral immunization

Vaccination of weaned piglets is usually designated for parenteral administration

before weaning. It is generally assumed that parenteral vaccines do not stimulate the

production of IgA antibodies and intestinal imrnunity, but stimulate systemic IgG- Their

delivery fiom blood to the intestinal lumen is poor, which iimits the efficacy of vaccination

as the bacteria rernain on the epithelial surface (Francis, 1991; Moon and BUM, 1993;

Laval, 1996; Van den Broeck et al, 1999~).

Parenteral immunization of suckling piglets using vaccines designed for sow

imrnunization to protect against neonatal diarrhea are not highly effective in preventing

post-weaning colibacillosis, since the adhesins of ETEC associated with PWECD are often

different fiom those associated with neonatal diarrhea (Fairbrother, 1999b).

There is a general agreement that parenterai vaccines are not enicacious in suckling

with circulating Iactogenic antibodies and their efficacy against PWECD has not been

convincingly demonstrated (Bertschinger et al, 2000).

However, there is a study where an injectable vaccine was tested on 7 farms, resulting

in a significant reduction in mortaiïties and incidence of clinical disease, and improved

growth rates in vaccinated pigs (Connaughton et al, 1992). A combined (live oral plus

killed parenteral) vaccine against PWECD has dso been tested to prevent PWECD,

resulting in a great protection against losses due to the disease (Alexa et al, 1995).
72
 Interestingly, Van den Broeck, et al, (1999~)demonstrated that systemic

imrnunization, performed by intramuscular injection, evokes a higher IgG but also induces

an IgA response. IgA senun antibody titres decline much faster than IgG titre which

suggests their secretion. However, they demonstrated that F4 specific antibody secreting

cells (ASC) upon IM immunization are homing preferentiaiiy into systemic and peripheral

lymphoid tissues (retropharyngeal lymph node and spleen) whereas d e r oral infection,

they home into mucosal lymphoid tissues (mesenteric lymph nodes) and spleen. The

difference in the migration patterns of ASC and the ratio of IgMgG semm antibodies

indicate that IM immUIlj.zation stimulates preferentially the systemic rather the mucosai

immune system (Van den Broeck et al, 1999~).

In another experimen~Bianchi et al (1.996)showed that parenteral immunization

induced a state of suppression that was reflected by the lack of an enteric immune

response upon a subsequent oral infection with live bacteria, and concluded that parenteral

vaccination of piglets with E. coli F4ac antigen is ineffective in inducing protective

immunity at the mucosal level against post-weaning diarrhea and that might be detrimental

(Bianchi et al, 1996).

1.7. References

Agin, T.S.and Wolf, M.K.: IdentXcation of a family of intjmins common to Escherichia

coli causing attaching-effacing lesions in rabbits, humans, and swine. (1997) Infection and
T m m ~ t y 65:
, 320-326.

Aizpurua, H.J. and RusseiLJones, J.G.: Oral vaccination: identircation of classes of

proteins that provoke an immune response upon oral feeding.(l988) J Exp Med, 167: 440-
451.

Alexa, P., Salajka, E., Salajkova, Z., and Machowa, A.: Combined parenterd and oral
immunization against enterotoxigenic Escherichia coli diarrhea in weaned piglets. (1995)
Vet Med (Praha), 12: 365-370.

Alexander, T.J.L.: Neonatal diarrhoea in pigs. In: Escherichia coli in domestic animais
and humans (1994) ed. Gyles, C.L. CAB International: 151-169.

Baker, D.R., Billey, L.O. and Francis, D.H.: Distribution of K88 Escherichia coli-
adhesive and nonadhesive phenotypes among pigs of four breeds. (1997) Vet Microb, 54:

Bertschinger, H.U., Nief, V. and Tschape, H.: Active oral immunization of suckling
piglets to prevent colonization after weaning by enterotoxigenic Escherichia coli with
fîmbriae F18. (2000) Vet Microb, 71: 255-267.

Bertshinger, H.U.: Post-weaning Escherichia coli diarrhea and edema disease In: Disease
of Swine ed. (1 999): 441-454.

Bertschinger, H.U.: Enteric diseases caused by Escherichia coli. In: Pigs Misset Special
Ed. Entenc Diseases (May, 1996): 14-16.

Bertschinger, H.U and Gyles, CL.: Oedema disease of pigs. (1994) In: Escherichia coli in
Domestic Animals and Humans. Ed. Gyles, C.L., CAB International: 193-219.
Bertschger, H.U., Fairbrother, J.M., Nielsen, N.O. and Pohlenz, J.F.: Escherichia coli.
In: infections Disease of Swine. 7~ ed. (1992): 487-5 11.

Bianchi, A.T., Scholten, J. W., van Zijderveld, A.M., van Zijderveld, F.G. and Bokhout,
B.A.: Parenteral vaccination of mice and piglets with F4+ Escherichia coli suppresses the
enterïc anti-F4 response upon oral infection. (1996) Vaccine, 14: 199-206.

Biehl, L.G. and Hoefling, D.C.: Diagnosis, treabnent, and prevention of diarrhea in 7- to
14-day-old pigs, (1986) JAVMA, 188: 1144-1 146.

Bijlsma, I.G.W., Van Houten, M; Frik, J.F. and Ruitenberg, E.J.: K88 variants K88ab,
K88ac and K88ad in oral vaccination of different porcine adhesive phenotypes.
Immunological aspects. (1987) Vet Immun01 and Tmmunopathol., 16: 235-250.

Blomberg, L., Henriksson, A. and Conway, P.L.: Inhibition of adhesion of Escherichia

coli K88 to piglet ileal mucus by Lactobacillus spp. (1993) Appl Environ Microbiol, 59:
34-39.

Carlson, M.S., Hill, G.M. and L a , J.E.: Phase-feeding pharmacological concentrations of

zinc oxide: effect on rnetallothioneixl and mineral concentrations. (1999) J Anim Sci, 77:
1199-1207.

Celemin, C., Rubio, P., Echevema, P. and Suarez, S : Gene toxin patterns of Escherichia
coli isolated fiom diseased and heahhy piglets. (1995) Vet Microb, 45: 121- 127.

Chandler, D.S. and Mynott, T.L.: Bromelain protects piglets fiom diarrhea caused by oral
challenge with K88 positive enterotoxigenic Escherichia coli. (1998) Gut, 43: 196-202.

Charbonneau, G.R.: Post-weaning colibacillosis. Diseases-bacterial. (1999) Personal

co~~~fllunication.
Connaughton, I.D., Driesen, S.J., Fahy, V.A. and Sammons, N.G.: Field trials with an E.
coli bacterin (weanavac) to prevent post-weaning coliform enteritis. In: 12" International
Pig Veterinary Society Congress, ~ r o c e é d i n ~August
s, 17-20, 1992: 254

Conway, P.L.: Specifïcally selected probiotics c m irnprove health and performance of

pigs. In: Manipulating Pig Production VIL Symposium: Antibiotic in pig production. Ed
Cranwell PD. Nov-Dec (1999) ~ k t r a l i a n
Pig Science Ass. Conference 7h Pig Research
and Development Corporation: 220-224.

Conway, P.L., Welin, A. and Cohen, P.S.: Presence of K88-specific receptors in porcine
ileal mucus is age dependent. (1990) Infection and Immunity, 58: 3 178-3 182.

Cupere D.F., Deprez, P., Demeulenaere, D.and MuyUe, E.: Evaluation of the effect of 3
probiotics on experirnental Escherichia coli enterotoxaemia in weaned piglets. (1992)
Zentralbl Veterinarmed @), 39: 277-284.

Davis, A. P.: Post-weaning Escherichia coli control. (1989) AASP Newsletter, 1: 1-6.

Dean, E.A., Whipp, S.C. and Moon, H.W.: Age-specific colonization of porcine intestinal
epithelium by 987P-palliated enterotoxigenic Escherichia coli. (1989) Tnfection and
Immunity, 57: 82-87.

Dean-Nystrom, E.A., Casey, T.A., Schneider, R.A. and Nagy, B.: A monoclonal antibody
identifies 2 134P fimbriae as adhesins on entertoxigenic Escherichia coli isolated fkorn
post-weaning pigs.(1993) Vet Microb 37: 20 1-114.

Dewey, C.E., Carman, S., Hazlett, M., van Dreumel, T., and Smart, N.: Endemic
transmissible gastroenteritis: Difficulty in diagnosis and attempted confîrmation using a
transmission trial. (1999) Case Report. Swine Health and Production, 7: 73-78.

Dewey, C.E., Wittum, T.E., Hurd, S., Dargatz, D.A. and Hill. G.W.: Herd- and litter-level
factors associated with the incidence of diarrhea morbidiw and mortality in piglets 4-14
days of age. (1 995) Swine Health and Production, 3: 105-112.
Done, S.H. and Patron, D.J.: Porcine reproductive and respiratory syndrome: clinicd
disease, pathology and immmology suppression. (1995) Vet Record, 136: 32-35.

Dubreuil, J.D., Fairbrother, J.M., Laiiier, R. and Lariviere, S.: Production and purification
of heat-stable enterotoxin b fiom a porcine Escherichia coli strain. (199 1) Infection and
Inimunity, 59: 198-203.

Dunlop, R.H., McEwen, S.A., Meek, AH., BIack, W.D., Clarke, R.C. and Fnendship,
R.M.: Lndividud and group antimicrobid usage rates on 3 4 fmow-to finish swine farms in
Ontario, Canada. (1998) Prev Vet Medicine, 34: 247-264.

Dunlop, R.H.: Antimicrobial treatments and antimicrobid resistance of fecal Escherichia

coZi of swine in Ontario, Canada. (1996) PhD. Thesis. University of Guelph.: 275-285.

Edfors-Lilja, I., Gustafsson, U., Duval-Iflah, Y., Eilergren, H., Johansson, M., Juneja,
R.K-, Marklund, L. and Andersson, L.: The porcine intestinal receptor for Escherichia
coli K88ab, K88ac: regional localization on chromosome 13 and influence of IgG
response to the K88 antigen. (1995) Animal Genetics, 26: 237-242.

Engle, M.J.: The role of soybean meai hypersensitivity in post-weaning lag and diarrhea in
piglets- (1994) Swine Health and Production, 2: 7-10.

Fahy, V.A., Comaughton, I.D., Driesen, S.J. and Spicer, E.M.: Post-weaning
colibacillosis. (1987) In: (eds) Manipulating Pig Production. Australia Pig Science
Association, Wembee, Victoria: 187-20 1.

Fairbrother, J.M.: Identification, nomenclature, and diagnosis of pathogenic Escherichia

coli. (1999a). Western Canadian Association of Swine Practitioners. Anrual Meeting.
Proceedings, October 15 &16, Saskatoon: 2 1-31,

Fairbrother, M.J.: Approaches to the control of Escherichia coli section.

( 1 999b)Western Canadian Association of Swine Practitioners. Annual Meeting.
Proceedings, October 15 & 16, Saskatoon: 65-68
Fairbrother, J.M.: Enteric ColibaciIiosis. In: Disease of Swine ed (1992): 489-497.

Fangman, T.J., Ostiund, E.N., Tubbs, RC.and He~ingsen-Dyer,K.: Effect of vaccine-

induced immune activation on the performance of early-weaned pigs. (1998) Vet Record,
143: 327-330.

Francis, D.H.: Colibacillosis in pigs and its diagnosis. (1999) Swine Health and
Production, 7: 241-244.

Francis, D.H. and Moxley, R.A.: Diagnostic notes. Porcine colibaciliosis. American
Association of Swine Practitioners AASP, Nov1991: 18-20.

Francis, D.H. and Willgohs, J.A.: Evaluation of a live avident Escherichia coli vaccine
K88+, LT+ enterotoxigenic colibacillosis in weaned pigs. (199 1) J Vet Res, 52: 1051-
1055.

Friendship, R.M. and Dewey, C.E.: Post-weaning Escherichia coli diarrhea (PWECD). Ln:
Health Management for Swine. University of Guelph (1999): 87-90.

Ganon, V.P. J., Gyles, C.L., and Friendship R.M.: Characteristics of verotoxigenic
Escherichia coli from pigs (1988) Can J Vet Res 52: 3 3 1-33 7.

Gradzki, S., Winiarczyk, B. and Rutkowska-Pejsak.: An outbreak of severe diarrhea

caused by concurrent infection with TGE, Rotavirus and E. coli in a large scale pig farm in
Poland. July 1996. Proceedings of the 14&IPVS Congress, Bologna, Italy.

Gyles, C.L.: Escherichia coli in diseases of weaned pigs: Biological aspect. In: Enteric
diseases of nursery pigs. (2001) 32ndh u a l Meeting U S V , Nashviile, Tenn: 29-41.

Gyles, CL.: Post-weaning E. coli Diarrhea: An Old Problem in new guises. (2000),
Printing in process.
Gyles, C.L-: Escherichia coli Enterotoxins. In: Escherichia coli in domestic animals and
humans. Ed. Gyles, C.L. Cab International (1994): 337-364.

Gyles, C.L.: Escherichia coli cytotoxins and enterotoxins- (1992) Can. J. Microb, 38:
734-746,

Gyles, C.L.: Escherichia coli. In: Pathogenesis of Bacterial Infections in Aaimals. Gyles,
C.L. and Thoen, (1986) Iowa State University Press Ames: 1 14-1 3 1.

Halagaard, C.: Epidemiologic factors in piglet diarrhea. (1981) Nord Vet Med, 33: 403-
412.

Hampson D.J. and Pethick, DW: C m enteric bacterial infections be controlled by diet?.
Proceedings of the 19&Western Nutrition Conference. New Ideas and new lcnowledge.
Saskatoon, Saskatchewan (1998): 53-59.

Hampson, D.J.: Post-weaning Escherichia coli diarrhea in pigs. In: Escherichia coli in
domestic animals and humans. Ed. Gyles, C.L. Cab international (1 994): 17 1-1 9 1.

Hampson, D.J., Fu, A.F. and Robertson, I.D.: Investigation of the source of hemolytic
Escherichia coli infecting weaned pigs. (1987) Epidem Infect, 99: 149-153.

Hampson D.J. and Smith, W.C. : Influence of creep feeding and dietary intake after
weaning on malabsorption and occurrence of diarrhea in the newly weaned pig. (1 986)
Res Vet Sci, 41 : 63-69.

Hampson, D.J., Hinton, M. and Kidder, D.E.: Coliform numbers in the stomach and smali
intestine of healthy pigs foilowing weaning at three weeks of age. (1985) J Comp Path,
95:353-362.

Hill, G., Rozeboom, De, Trottier, N., Mahan, D., Adeoli, L., C h e , T., Forsyth, D.,
Richert, B.: The started pig. In: Tri-State Swine Nutrition Guide. The Ohio State
University. (1998): 26-3 1.
Hoblet, K.H., Kohler, E.M., Saif, L.J., Theil, K.W- and Ingalls, W.L.: Study of porcine
post-weaning diarrhea involving K88(-) hemolytic Escherichia coli. (1986) Am J Vet Res,
47: 1910-1912.

Huang, S.X., McFall, M., Cegielski, A C , and Kirkwood, RN.: Effect of dietary zinc
supplernentation on Escherichia coli septicemia in weaned pigs. (1999) Swine Heaith
Production, 7: 109-1 11.

Hunter, J.: Some studies on multiple-resistant E. coli and the use of antibiotic in the
treatment of diarrhea in pigs. (1993) Pig Veterïnary Journal, 3 1: 143- 15 1.

Imberechts, H., Deprez, P., Van Drïessche, E. and Pohl, P.: Chicken egg yolk antibodies
against F l8ab fhabriae of Escherichia coli inhibit shedding of F 18 positive E. coli by
experimentally infected pigs. (1997) Vet Microb, 54: 329-34 1.

Isaacson, R.E.: Vaccines against Escherichia coli diseases. In: Escherichia coli in
domestic animals and humans. Ed Gyles, CL. Cab International (1994): 629-647.

Jeyasingharn, M.D., Butty, P., King, T.P., Begbie, R. and Denise Kelly.: Escherichia coli
K88 receptor expression in intestine of disease-susceptible weaned pigs. (1999) Vet
Microb, 68:2 19-234.

Josephson, G. and Archambault, M.: ColibaciUosis in pigs in 1999. (2000) Animal Health
Laboratory Newsletter, 4: 8.

Josephson, G., Smart, N., McEwen, B. and Gough, J.: K88+ E. coli diarrhea in post
weaning pigs. Animai Health Laboratory, University of Guelph. In: 18" Annual Centralia
Swine Research Update, January 27 (1999): 38-39.

Josephson, G. and Smart, N.: K88 strains of E. coli. (1998a) Animal Health Laboratory
Newsletter June, 2: 4
Josephson, G. and Smart, N.: Update on K88 positive E. coli. (1998b) Animal Health
Laboratory Newsletter Sept, 2: 2.

Kniffen, T. and Neumann, E.: Field experience with E. coli K88. In: Emerging Nursery
Pig Diseases and Management Solutions. 3 lnh u a l Meeting Am. Ass. Of Swine
Practitioners. Indianapolis (May, 1996): 20-21.

Laval, A.: A review of new vaccines for enteric diseases. In: Pigs Misset Special Ed.
Enteric Diseases (May, 1996): 20-21.

Lawrence, T.L.J.: Processing and preparation of cereais for pigs diets. (1985) In: Recent
Developments in Pig Nutrition. Ed. Cole, D.J.A. and Haresign, W: 230-245.

Lecce, J-C., Clare, D.A., Balsbaugh, R.K., Collier, D.N.: Effect of dietary regirnen on
rotavirus-Escherichia coli weaniing diarrhea of piglets. (1983) J Clin Microbiol, 17: 689-
695.

Lior, H.: Classification of Escherichia coli. In: Escherichia coli in domestic animals and
humans. Ed. Gyles, C.L. CAB International (1994):3 1-71.

MacKinnon, J.D.: Enteritis in the young pig caused by Escherichia coli. (1998) The Pig
Journal, 41: 227-255.

MacKinnon, J.D.: What causes gastro-intestinal disease?. (May, 1996) In: Pigs Misset
Special Ed. Enteric Diseases: 4-5.

McAllister, J.S., Kurtz, H.J. and Short, E.C.Jr.: Changes in the intestinal flora of young
pigs with post-weaning diarrhea or edema disease. (1979) 5 Anim Sci, 49: 868-879.

Madec, F., Humbert, F., Salvat, G. and Maris, P.: Measurement of the residual
contamination of post-weaning facilities for pigs and related risk factors. (1999) J Vet
Med, 46: 37-46.
Mathew, A.G., Saxton, A.M., Upchurch, W.G. and Chattin, S.E-:Multiple antibiotic
resistance patterns of Escherichia coZi isolates form swine f m s . (1999) Appl Environ.
Microb,, 65: 2770-2772.

Mathew, A.G., Upchurch, W.G. and Chattin, S.E.: Incidence of antibiotic resistance in
fecd Escherichia coli isolated fiom commercial swine f m s . (1998) J Anim Sci, 76: 429-
434.

MetcaK, J.W., Krogfelt, K.A., Krivan, H.C., Cohen, P. S. and Laux, D .C. :
Characterization and identification of a porcine small intestine mucus receptor for the
K88ab nmbnd adhesin. (1991) Infection and Immunity: 9 1-96.

Michaels, RD., Shannon, C., Whipp, and Rothschild, M.F.: Resistance of Chinese
Meishan, Fengjing, and Minzhu pigs to the K88ac+ strain of Escherichia coli. (1994) Am
J Vet Res, 55: 333-337.

Miller, B.G., Newby, T.J., Stokes, C.R. and Bourne, F.J.: Influence of diet on post-
weaning maiabsorption and diarrhea in the pigs. (1984) Res Vet Sci, 36: 187-193.

Miller, B., Newby, T.J., Stokes, C.R., Hampton, D. and Bourne, F.J. :The role of dietary
antigen in the etiology of post-weaning diarrhea. (1983) Ann Rech Vet, 14: 487-492.

Mills, K.W., Tietze, K.L. and Phillips, R.M.: Use of enzyme-linked immunosorbent assay
for detection of K88 pili in fecd specimens fiom swine. (1983) Am J Vet Res, 44: 2188-
2189.

Moon, H.W. and Bunn,T.O.: Vaccines for preventing enterotoxigenic Escherichia coli
infections in farm animals. (1993) Vaccine, 1 1 : 213-220.

Morona, EL, Morona, J.K., Considine, A., Hackett, J.A., van den Bosch, L., Beyer, L. and
Attridge, S.R-: Construction of K88-and K99-expressing clones of Salmonella
typhirnurium G30: immunogenicity following oral administration to pigs. (1994) Vaccine.
 Mynott, T.L., Luke, R.K.J. and Chandler, D.S.: Orai administration of protease inhibits
enterotoxigenic Escherichia coZi receptor activity in piglet smail intestine. (1996) Gut, 38:
28-32.

Nabuurs, M.J: Weaning piglets as a model for studying paîhophysiology of diarrhea.

(1998) Vet Q, 20: Suppl3: S42-45.

Nagy, B. and Fekete, P.Z.: Entertoxigenic Escherichia coli (ETEC) in f

m animals.
(1999) Vet Res, 30: 259-284.

Nagy, B., W p p , S.C., Imberechts, H., Bertschinger, H.U., Dean-Nystrom, E.A., Casey,
T.A. and Salajka, E.: Biological relationship between F 18ab and F 18ac fimbnae of
entertotoxigneic and vertotoxigenic Escherichia coli f?om weaned pigs with oedema
disease or diarrhea. (1997) Microb Pathogen, 22: 1-1 1.

Nagy, B., Casey, T.A., Whipp, S.C. and Moon, H.W.: Susceptibility or porcine intestine
to pilus-mediated adhesion by some isolates of palliated enterotoxigenic Escherichia coli
increases with age. (1992) Infection and Iinmunity, 60: 1285-1294.

Nakamine, M., Kono, Y., Abe, S., Hoshino, C., Shiraï, J., and Ezaki, T. : Dual infection
with enterotoxigenic Escherichia coli and porcine reproductive and respiratory syndrome
virus observed in weaning pigs that died suddenly. (1998) J Vet Med Sci, 60: 555-561.

Olsson, E., Smyth, C.J., Soderlind, O., Svemerholm, A.M. and Molby, R.: Development
of intestinal antibodies against Escherichia coli antigens in piglets with experirnental
neonatal E. coli diarrhea. (1986) Vet Microb, 12: 119-133.

Osek, J.: Prevalence of vinilence factors of Escherichia coli strains isolated fiom
diarrhetic and healthy piglets after weaning (1999) Vet Micro. 68: 209-2 17.

Osek J., Gallien, P., Truszcq-ibki, M and Protz, D.: nie use of polymerase chah reaction
for determination of virulence factors of Escherichia coli strains isolated fiom pigs in
Poland. (1 999) Comm Immun Micro & Infect Dis, 22: 163-174.
Osek, J., Truszczyilski, M., Tarasiuk, K. and Pejsak, 2.: Evaluation of dserent vaccines
to control of pig colibacillosis under large-scale f m conditions. (1995) Comp Immun
Microb Infect Dis, 18: 1-8.

Porter, Kenworthy, R. and Allen, W.D.: EEect of oral immunization with E. coZi antigens
on post-weaning enteric infection in the young pig. (1974) Vet Rec, 95: 99-104.

Pritchard, G.C.:Transmissible gastroenteritis in endernicdy infected breeding herds of

pigs in East Anglia, 1981-1985. (1987) Vet Rec, 120: 226-230.

Rippinger, P., Bertschuiger, H.U., Imberechts, H-, Nagy, B., Sorg, I., Stamm, M., Wild,
P., and Wittig, W.: Designations F 18ab and F 18ac for the related fimbrial types F 107,
2134P and 88 13 of Escherichia coli isolated fiom porcine post-weaning diarrhea and fiom
edema disease. (1995) Vet Microb, 45: 28 1-295.

Rose, R. and Moon, H.W.: Elicitation of entroliiminal neutrophils by enterotoxigenic and

nonenterotoxigenic strains of Escherichia coli in swine. ( 1 985) Infection and Immunity,
48: 812-823.

Rose, R., Whipp, S.C. and Moon, H. W.: Effects of Escherichia coZi heat-stable
enterotoxui on small intestinal villi in pigs, rabbits, and Iambs (1987) Vet Pathol, 24:71-
79.

Salajka, E., Salajkova, Z., Alexa, P. and Homich, M.: Colonization factor different from
K88, K99, F41 and 987P in enterotoxigenic Escherichia coli strains isolated fiom post-
weaning diarrhea in pigs. (1992) Vet Microb, 32: 163-175.

Sarmiento, J.I., Runnels, P.L. and Moon, H. W.: Effects of pre-weaning exposure to a
starter diet on enterotoxigenic Escherichia coli- induced post-weaning diarrhea in swine.
(1990) Am JVet Res, 51: 1180-1183.

Sarmiento, J.I., Dean, E.A. and Moon, H.W.: Effects of weaning on diarrhea caused by
enterotoxigenic Escherichia coli in three-week-old pigs. (1988) Am J Vet Res, 49: 2030-
2033.
Sarrazin, E. and Bertschinger, H.U.: Role of fimbnae F 18 for actively acquired immunity
against porcine enterotoxigenic Escherichia coli. (1997) Vet Microb, 54: 133 - 144.

Skirrow, S.Z., Buddle, J.R, Mercy, A.R., Madec, F. and Nicholls R.R.:Epidemilogical
studies of pigs diseases: 2. Post-weaning diarrhea and performance in Western Australian
pigs. (1997) Aust Vet J, 75: 282-288.

Smith, H.: Virulence determïnants of Escherichia coli: Present knowledge and questions.
(1992) C m J Microb, 38: 747-752.

Smith, W.H. and Linggood, M.A.: Observations on the pathogenic properties of the K88,
hIy and ent plasrnids of Escherichia coli with particular reference to porcine diarrhea.
(1971) J Med Microb, 4: 467-485.

Soderlind, O., Thafielin, B. and Mollby, R.: Virulence factors in Escherichia coli strains
isolated from swedish piglets with diarrhea (1988) J Clin Microb: 879-884.

Spencer, R.J. and Chesson, A.: The effect of Lactobacillus spp. on the attachment of
enterotoxigenic Escherichia c d i to isolated porcine enterocytes. (1994) J Appl Bacetriol,
77: 215-220.

Stokes, C.R.: Immune systems in the porcine gut. (1988) Veterinary Journal, 20: 19-30.

Svensmark, B., Nielsen, K., Willeberg, P. and Jorsal, S.E.: Epidemiological studies of
piglet diarrhea in intensively managed Danish sow herds. II. Post-weaning diarrhea. (1989)
Acta Vet Scand. 30: 55-62.

Taylor, D.J.: Piglet enteritis fiom 10 days to weaning. (1996) The Pig Journal-Proceedings
Section, 36: 158-174.

Thomlinson, J.R. and Lawrence, T.L. Jr: Dietary manipulation of gastric pH in the
prophylaxis of enteric disease in weaned pigs: some field observations. (198 1) Vet
Record: 120-122.
Tunies, C.G., Aleixo, J.A., Monteiro, A.V. and Deliagostin, O.A.: DNA inoculation with
a plasmid vector carrying the faeG adhesin gene of Escherichia coZi K88ab induced
immune responses in mice and pigs. (1999) Vaccine, 17: 2089-2095.

Vdery, M.A.: Behaviourial patterns of the weaned piglet. (1995) The Pig Journal, 34 : 71-
97.

Van Beers-Schreurs, H.M., Vellenga, L., Wensing T, Breukink, H.J.: The pathogenesis of
the post-weaning syndrome in weaned piglets: a review. (1992) Vet Q, 14: 29-34.

Van den Broeck, W., Cox, E., Oudega, B. and Goddeeris, B.M.: The F4 fimbrial antigen
of Escherichia coli and its receptors. (2000) Vet Microb, 7 1:223-244.

Van den Broeck, W., Cox, E. and Goddeeris, B.M.: Seroprevalence of F4+
enterotoxigenic Escherichia coZi in regions with different pig farm densities. (1999a) Vet
Microb, 69: 207-2 16.

Van den Broeck, W., Cox, E. and Goddeeris, B.M.: Receptor-dependent immune
response in pigs after oral imrnunization with F4 fimbriae. (1999b) Infection & Immunity,
67: 520-526.

Van den Broeck, W., Cox, E.and Godeeris, B.M.: Induction of immune responses in pigs
following oral administration of purified F4 funbriea. (1999~)Vaccine, 17: 2020-2029.

Vogeli, P., Bertschinger, H.U., Stamm, M., Stncker, C., Hagger, C., Fries, R., Rapacz, J.,
and Stranzinger, G.: Genes speciQing receptors for FI8 h b r i a t e d Escherichia coli,
causing oedema disease and post-weaning diarrhea in pigs, map to chromosome 6. (1996)
Animal Genetics, 27: 32 1-328.

Wada, Y., Nakaoka, Y., Kondo, H., Nakazawa, M., and Kubo, M.: DuaI infection with
attaching and effacing Escherichia coZi and enterotoxigenic Escherichia coli in post-
weaning pigs. (1996) J Comp Pathol, 114: 93-99.
White, D.G.: Characterization of antimicrobid resistance among pathogenic swine
Escherichia coli. (2000) In: Emerging Nursery Pig Diseases and Management Solutions.
Am. Ass of Swine Practitioners, Indianapolis, Indiana: 23-29.

Wilson, R.A. and Francis, D.H.: Fimbriae and enterotoxins associated with Escherichia
coli serogroups isolated fiom pigs with colibaciIlosis. (1986) Am J.Vet Res, 47: 2 13-217.

Wood, E.N.: Fashionable and füture diseases. (199 1) Vet Journal, 27: 193-197.

Woodward, ML: An old fiiend revisited .(1997) The Pig Joumal, 39: 54-62.

Wray, C. and Woodward, M.J.: Laboratory diagnosis of E. coli infections. In: Escherichia
coli in domestic animals and hans.(1994) ed. Gyles, C.L. CAB International: 595-628.

Yavzon, M., Porath, N., Ochana, O., Dagan, R., Orini-Wasserlauf, R. and Cohen, D.:
Detection of enterotoxigenic Escherichia coli in stool specimens by polymerase chain
reaction. (1998), Diagn Microb Lnfect Dis, 3 1: 503-509.

Zhu, C., HareI, J., Jacques, M. and Fairbrother, J.M.: Interaction ulth pig ileal explants of
Escherichia coli 045 isolates fiom swine with post-weaning diarrhea. (1995) Can J Vet
Res, 59: 118-123.

Zhu, C., Harel, J., Jacques, M., Desautels, C., Domenberg, M.S., Beaudry, M. and
Fairbrother, J.M.: Virulence properties and attachhg-effacing activity of Escherichia coZi
045 form swine post-weaning diarrhea. (1994) Infection and Immunity, 62: 4153-4159.

Zimmerman, J., Yoon, K.J., Stevenson, G. and Dee, S.A.: Infection, disease, and
economics. In: PRRS Compendium. A comprehensive reference on Porcine Reproductive
and Respiratory Syndrome for pork producers, veterinary practitioners, and researchers.
National Pork Producers Council& National Pork Board. (1998): 17-26.
Chapter 2: A study investigating the presentation of post-weanhg E. coli diarrhea

in Southern Ontario, prevalence of E. c d i serogroups involved and their

antimicrobial resistance patterns.

2.1. Introduction

Enterotoxigenic E. coli (ETEC) strains that express K88 (F4) fimbriae are a major

cause of diarrhea and death in neonatd and newly weaned pigs. Animals af3ected by

ETEC typicdly show signs of severe, watery diarrhea during the fist few days of Me or a

few days d e r weaning. ETEC adhere to the smdl intestinal microvilli and produce

enterotoxins acting locally on enterocytes. This action results in the hypersecretion of

water and electrolytes, and reduced absorption (Mackinnon, 1998; Nagy and Fekete,

1999; Fairbroîher, 1999).

Post-weaning diarrhea and mortality caused by K88 (F4) E. coli have recently emerged

as an important disease problem in Ontario. The infection progresses so rapidly that pigs

of two to eight weeks of age are sometimes found dead before clinical signs are observed.

The provincial veterïnary diagnostic laboratory records revealed that this was a rare

problem prior to 1996. A three-fold increase in the isolation of enterotoxigenic E. coli

(ETEC), specincally of 0l49:K88 (F4) serogroup, was noted in 1997 (Josephson and

Smart, 1998qb; Josephson et al, 1999). Pure cultures of the ETEC organisms were grown

froom intestinal swabs of post-weaning diarrhea cases. In-viîro, these strains of E. coli

produce extensive hemolysis. It has been suggested that the newly ernerging problem in
weaned pigs is caused by E. coii strains that are more v i d e n t in individual anùnals and

more persistent within a herd than previous E. coli strains (Wood, 1991).

Some post-weaning E. coli ETEC strains may produce multiple adhesins such as K88,

Fl8ac or K88, F18 and F41 (Nagy and Fekete, 1999). Post-weaning E. coli diarrhea

(PWECD) are characterized by production of either LT, STa,or STb or combinations.

The most fkequently observed enterotoxin combinations are LT and STb, or LT, STa, and

STb (Celemin et al, 1995). Isolates belonging to the Attaching Effacing (AEEC) category,
are also observed in about 6% of piglets with diarrhea in the post-weaning period

(Fairbrother, 1999).

Economic costs of PWECD have been attributed to death of individual ;inimals

(Bertschinger, 1999); cost of medication; labour costs involved in treating individuai pigs

and reduced average daily gain (ADG) (Josephson et al, 1999). Some f m s expenence

mortalities as high as 20% and a decrease of ADG fkom 430-450gram per day, to 400-

410 g r a m per day, following the appearance of K88-positive E. coli in the nursery.

Antimicrobial therapy is usually initiated in an attempt to control morbidity and

mortality associated with these bacterial pathogens, but E coli isolates responsible for the

disease are often resistant to a wide range of antimicrobials (Dunlop, 1996; Fairbrother,

1999). The fiequency of resistance has been steadily increasing over the last several years

(Fairbrother, 1999). A wide range of antibiotics given in feed &or water, as well as

other chemotherapeutic agents such as zinc oxide, have provided only partial success at a

high cost (Wood, 1991).

 The objectives of this study were to describe the comrnon F4 E. coli serogroups

isolated f?om nursery UIllts in Ontario in the summer of 1999, to examine antimicrobial

resistance patterns arnong these isolates; to evaiuate the impact of the disease on certain

productivity parameters and to characterize the presentation of post-weaning E. coli

diarrhea on pig f m s in Ontario.

2.2. Material and Methods

A total of 50 farms were visited in the summer of 1999 as p a of

~ a~case-control shidy.

Twenty-five case and twenty-five control herds were selected fkom the records of the

Animal Heaith Laboratory, Guelph, Ontario, and of several swine practitioners. From the

1998 and 1999 Animal Heaith Laboratory records, f m s were selected based on the

presence or absence of K88-positive E. coli, in pigs that were two weeks of age or older,

and a history of diarrhea andor sudden death. The swine practitioners identified clients

who had pigs either with or without K88-positive E. coli and who would be willing to

participate in the study. Farms were visited shoaly after consent was acquired. A standard

protocol was followed on each fann.

A case farm met the following criteria: pigs that were at least two weeks of age, that

had clinical signs of post-weaning E. coli diarrhea and mortality as well as positive

cultures of K88 E. coli. A control f m was one where pigs were at least twc weeks of

age, without a history of clinical signs of post-weanhg E. coli diarrhea and no positive

cultures of K88 E. coli.

 During each f m visit the producer was asked to answer a survey (Appendix 1).

Information on average da* gain (ADG), mortality, presentation and management of

diarrhea problems, and antimicrobial usage was collected. ADG was obtained either fiom

farm records, estimates of producers or by subtractîng the weaning weight from the

weight of the pigs when they were moved out of the nursery, divided by the number of

days in the nursery. Pre-weaning ADG was calculated subtracting an average birth weight

of 1.5 kg fiom the average weight at weaning, divided by the average weaaùig days.

Clinicai problems which have been associated with a diarrhea problem were coliected in

the case and control farms.

Rectd swabs of ten weaned pigs were collected fiom each of the f m s visited to

ensure a consistent attempt to culture K88-positive E. coli at the time of the visit. The

samples were taken fiom pigs which showed clinicd signs of diarrhea, in case farms, and

fiom pigs which looked unthrifty or with dirty tails in the control f m s . These samples

were sent to Gallant Custom Laboratories Inc, Cambridge, Ontario, where slide

agglutination tests were performed. Isolates that were hemolytic in antisenun pool 2 N

were reported in the results, but only hemolytic 0149: K9 1K88 isolates were tested for

mtimicrobial sensitivity (Figure 2.1).

Sensitivity testing was conducted for the followuig antibiotics: ampicillin, carbencillin,

cefadroxil, erythromycin, gentamich, amikacin, tobrarnycin, enrofloxacin, neomycui,

ciprofloxacin, polymyxin B, spectinomycin, sulfamethoxazole, tetracycline, ceftiofur, and

aprarnycin.
 Subsequently, blood agar plates with a single colony type f?om each of the positive

farms, were submitted to the School of Veterinary Medicine at the University of Montreal.

Isolates were tested for the presence of specinc K88 (F4) and FI8 b b n a e and for genes

for toxins associated with PWECD (STa,STb, LT, and VT) via eçtablished Polymerase

Chain Reaction (ETR) protocols.

Clinical signs attributed to the PWECD in the case farms were cdculated based on the

odds of disease and the estimated attributable fiaction procedures according to the

foilowing formula (Martin et al, 1988):

(OR- 1)/OR

The simple association between case and control herd status and putative disease

factors was determined using Chi-square for qualitative variables and Student's T test for

quantitative variables. Numerical variables which were not normally distributed, according

to the Wilk-Shapiro test, were tested with the Mann-Whitney test. Variables signifiant at

a p-value 50.05, were considered significant. Variables with a p-value between 0.06 and

0.1 were considered numerically reportable as potential trends. Statistical analyses were

completed in Statistix (using version 1.O)(Statistix 0 Analytic Software. Tallahassee,

Florida).

2.3. Results

Three control herds developed diarrhea problems and were diagnosed with K88-

positive E. coli prior to the faxm visit. These three famis were considered cases. One

s w e y fiom a case fxm was not completed at the time of the visit. The f m e r was asked
92
to complete and return the survey but failed to do so. Another case farm tumed out to be

a grower-finisher operation. Bo& these f m s were dropped fkom the survey anaiysis,

however, rectal swabs were taken fiom each f m , and these were cultured, typified, and

tested for sensitivity (Table 2.1).

There were 28 case farms and 22 control f m s included in the anaiysis for the

serogroup and antibiotic resistance study. In the slide agglutination test, 17 case farms and

three control f m s were 0149:K91:K88 positive! The three positive control f m s were

left in the control group because no diarrhea was observed at the time of the study and

there was no history of an E coli dianhea problem pnor to the study. Three K88-positive

case f m s also had isolates of serogroup 0139:K82 and one O 149:K91 :K88 negative

case f m was positive for 0138K8 1. Among the negative control f m s , one was

positive for 08:KX105.

The results of the PCR test are summarized in Table 2.2. From 17 K88-positive case

farms, 14 were positive for STa, STb, and LT (82.3 5%), two were positive to STb and LT

(1 1.76%) and one farm (5.88%) was positive to STb, STa,LT, as well as for VT and F18.

Case herds were more likely than control f m s to have STb, STa,and LT @= 0.01). Two

K88-negative case f m s , positive for 0:138:K81 and 0139:KS2, tumed out to be F 18

positive. Isolates fkom the three control f m s which were positive to K88 were positive to

STb and LT, and among the K88-negative control farms, two were eae-positive (E. coli

attaching and effacing).

A total of 68 isolates fiom 17 K88-positive case f m s and eight isolates fiom three

K88 positive control farrns were assayed for antibiotic sensitivity. The E. coli organisms

93
were resistant to multiple antibiotics. Multiple antimicrobial resistance as typined by

resistance to at least two distinct antimicrobial classes, was observed in 100% of the K88-

positive case and control f m s assayed (Table 2.3).

From the 16 antibiotics tested, the K88-positive E. coli isolates fiom case farms were

resistant to more antibiotics (75%) than E. coli fiom control farms (44%). Isolates of E.

coZi fiom case and control K88-positive f m s were 90400% resistant to spectinomycin,

tetracyclines, and erythromycin. The proportion of case f m s , resistant to ampicillin,

carbencillin, and neomycin was 53%; for sulfamethoxazole, 4 1.17%; for cefadroxil,

gentamicin, apramycin and tobramycin, 23.52%, and for ceftiofur, 5.88%. For the K88-

positive control f m s , 66.7% of these isolates were resistant to neomycin, and 3 3 -3% for

ampicillin, carbencillin, apramycin and sufamethoxazole. Al1 isolates were susceptible to

amikacin, enrofloxacin, cyprofloxacin and polymyxin B. In addition, isolates fiom the

three control K88-positive farms were susceptible to cefadroxil, gentamicin, tobramycin

and ceftiofur (Table 2.3).

Antibiotic patterns according to the specific toxins isolated from the K88-positive

farms were analyzed, and results are shown in Table 2.3. E. coli with genes for al1three

enterotoxins or just two enterotoxins, were resistant to spectinomycin, tetracyclines, and

erythromycin (90-100%). Serogroups with three toxins varied in percentage of resistance

as follows: 53.3% to ampicillin, carbencillin and neomycin; 20% resistance to cefadroxil,

gentamicin, tobrarnycin and apramycin, and 6.7% resistance to ceftiofirr. Serogroups with

two tohs had 60% resistance to neomycin; 40% to ampicillin, carbencillin,

sulfamethoxazole, and 20% resistance to cefadroxil, gentamicin, tobramycin and

apramycin. In the case of three and two toxins respectively, 40 and 80% of the isolates

were intemediately sensitive for cefadroxïl.

For the survey analysis, 26 f m s were designated as cases and 22 f a m s were

designated as controls. K88-positive E. coli was isolated fkom 15 of the 26 case farms.

Although K88-positive E. coli was cultured fiom 3 contrd farms, no diarrhea problems

were apparent (Table 2.4).

In 23 case fanris (88.5%), the diarrhea and mortality began within a week of weaning,

just one negative F4 control f m reported severe diarrhea problems within this period

(p<0.001). Three case f m s (11.5%) had the worst problern of diarrhea at two to three

weeks after weaning, and four case f m s (15.38%) d s o reported diarrhea problems in the

pre-weaning period.

There was no signûicant differences in growth rate of pigs in the pre-weaning period

between case and control farms. After weaning, growth rate in the nursery tended to be

better on control f m s than on case f m s (452 per day vs 414 g per day respectively) @=

0.07). There was no difference in mortality rate before the E. coli outbreak occurred

between the two groups. Afier the problem, the mortality in case f m s (7.7%) was higher

than on control farms (1-8%) @ <0.001) (this p value was obtauied fiom the Mann-

Whitney test). The mean mortality in case herds ranged fiom 0.5% to 10-30% (Table 2.5).

The most common clinical signs associated with an outbreak of PWECD were, sudden

death (97.1%), profuse watery diarrhea (98.2%), a bony or mthrifty appearance (90%),

vomition @O%), and dehydration (97%). More control farms reported rnild diarrhea

problems than case f m s . Ml the clinical problems measured were significantly different

95
between case and control farms (Table 2.6). Other clinical signs reported by the f m e r s

were: pigs with purple coloration of the body before death, and the impression that certain

pens were more affected than others.

Management and feed changes to control diarrhea problems were more commonly

reported in case f m s than in control f m s @ c0.001). The most common management

changes were: increased age at weaning; better control of temperature and ventilation;

creation of sick pens; decreased density in pens; and reduced mWng of pigs. The most

common feed changes reported were: purchasing feed fiom a different supplier; changing

the feed medication; offering a limited amount of feed four or five times per day; blending

of feed between phases; adding high levels of zinc oxide; decreasing the level of protein;

and increasing the level of fibre.

Case f m s commonly switched feed antibiotics when diarrhea problems were present

(46.2%). The most common antibiotics used in fanns which had switched in-feed

antibiotics were: carbadox (30.8%), chlortetracycline (23%) and a combination of

chlortetracycline, sulfamethazine, and procaine penicillïn (ASP) (15.4%). However, not

simiificant differences were found between the use of these antibiotics and the ones used in

control famis (Table 2.7). Case f m s were more likely to use antibiotics in water (73.1%)

than control f m s (13.6%) @<0.00 l), with apramycin being the first water medication of

choice in case farms. Case farms were more likely to change to a second choice of

antibiotic in water when no resuIts were observed with the first choice of antibiotic.

Neomycin was the antibiotic of choice in these cases (p= 0.05) (Table 2.8).
 The case farms were more likely to use injectable antibiotics of individual pigs than

control fanns (63.4 and 18.2% respectively) @= 0.001). The injectable antibiotic most

cornrnonly used to treat PWECD was trimethroprïm with sulfadoxine (41-2%) (p= 0.O 1)

(Table 2.8)-

The use of E. coZi vaccines in nursery pigs, probiotics, acidifiers in water or feed were

other forms of prevention measures employed by a few of the case f m s . Sanitizers in

water were used on 19.2% of the case f m s and just in one (4.5%) of the control f m s .

More case farms (61.5%) tended to use high levels of zinc oxide (22.5 kg/T) compared to

control f m s (36.4%) @= 0.08). Case farms (42.3%) tended to be more likely to use

electrdytes than control farms (18.2%) (p= 0.07) (Table 2.9).

2.4. Discussion

This study demonstrates that post-weaning E. coli diarrhea (PWECD) is an

economically important disease in pigs. Moaality increased fiom 2 to 7% on average

following an outbreak. If it is assurned that a newly weaned pig is worth approximately

$40, then this level of mortality on a 500-sow herd would equal a loss of about $20,000

dolars annually. However, some f m s experienced mortalities as high as 20-30%, over a

one to two month time span. Average daily gain (ADG) tended to be lower in f m s with

the problem compared with the farms which do not have the disease (452 grams per day

and 414 grams per day, respectively). Similar decreases in growth rates and mortality

were reported by Josephson et al (1999). In addition, case farms used more antibiotics and
other treatments such as vaccines, acidifiersi, probiotics, high levels of zinc oxide and

injection of individual pigs, which reflect an increased production cost and labour cost.

In this study it was noted that PWECD occurs most commody in the first week &er

weaning, but in agreement with Fairbrother (1999), the disease was dso observed to affect

pigs after two to three weeks foilowing weaning, or in some cases, foilowing transfer of

pigs to the grower-fisher units.

It has been reported that the condition often recurs sporadically. This study supports

this in the fact that some case farms were free of the problem at the t h e of the visit and

no K88-positive E. coli were found. Producers have often experienced short periods of

success in combating the disease, only to have the disease reappear a few weeks later. The

clinical signs observed by the f m e r s on this study are similar to those described else

where (Hampton, 1994; Charbonneau, 1999'; Fairbrother, 1999; Bertschinger, 1999,

Fnendship and Dewey, 1999).

The 0l49:Kg 1:K88 ETEC E. coli serogroups were the most commody isolated in

pure cultures, fÏom pigs with post-weaning E. coli diarrhea problems (PWECD). This

agrees with previous reports, as this group has been the most commoniy isolated fiom pigs

with the problem in Ontario (Josephson and Archambault, 2000). The 0 lD:K82

serogroup was isolated fkom some farms, this serogroup is the most commody isolated

fkom cases of edema disease (Gyles, 2001). The most fiequently observed enterotoxin

combination was STa, STb,LT, which agrees with reports made by Celemin et al (1995).

The emergence of this E. coli with three toxllis has prompted Fairbrother (1999) to

'~ersooalcommunication
suggest this may indicate the emergence of a new and more virulent bactenum.

Combination ofjust two toxins (Sm, LT) was observed less commody in case f m s , but

were present in all positive control farms, where ctinical problems were not apparent. The

presence of multiple adhesins and toxins was observed in at least one farm which had a

serious diarrhea problem (STa, STb, LT F4 and VT+F 18). Nagy and Fekete (1999),

reported that some ETEC strains may produce more than one adhesin factor. Other

authors have suggested that the presence of Shiga-like toxin (SLT), specifically SLT2e are

associated with oedema disease, but that they can play a role in post-weaning E. coli

diarrhea (Hampson 1994; MacKinnon, 1998; Fairbrother, 1999). Attaching-effacing

isolates were found in contcol f m s which had cuitures positive in Pool 2N. These isolates

were not associated with a diarrhea problem. This is in contrast with Zhu et al (1994) and

Fairbrother (1 999), where isolates belonging to this serotype were commonly associated

fiom pigs with PWECD .

In agreement with others studies (Dunlop, 1996; MacKinnon, 1998; Bertschinger,

1999; Fairbrother, 1999), antibiotic resistance of ETEC isolates fiom pigs with post-

weaoing diarrhea, showed a high rate of resistance to multiple antibiotics. hcreased risk of

resistance arnong E. coli has been associated with the use of various antimicrobials

(Dunlop 1996, Bertschinger, 1999). The use of various antibiotics via feed, water or

injection, was a cornmon finding on f m s which had PWECD problems. In addition, a

substance continuously used in a weaner facility exerts a tremendous selective pressure, as

a consequence, categories of antimicrobials have lost or are on the point of loshg their
activity against porcine coli (Bertschinger, 1999). A good example found in this study,

was the hi& rate of resistance to tetracycline, neomycin and sulfimethoxazole.

More antibiotics are used on case f m s , and therefore it is not surprising, that the E.

coli bacteria on case f m s showed more resistance in general. Antibiotics appear to be a

short term solution and as fanriers move from one antibiotic to the next, as illustrated in

this study, antibiotic resistance is likely to develop.

Dunlop et al (1 998) reported that in Canada, the medications most commonly added to

creep and starter rations were; a combination of chlortetracycline-sulfamethazine-

penicillin, carbadox, and tylosin plus furazolidone. In agreement with Dunlop's findings,

except for fürazolidone that has been withdrawn as an approved in-feed medication for

swine rations, f m s on this study tended to use these antimicrobial dmgs as growth

promoting agents and as therapeutic dmgs to treat diarrhea

According to Fairbrother (l999), apramycin (Apralan TM) remains the antimicrobial of

choice for use in water. In the Fairbrother study, most isolates showed sensitivity to

apramycin and amikacin. In this present study, apramycin was a commonly used treatment

and the resistance of E. d i isolates to the antibiotic was found to be 23.5%,

demonstrating the development of resistance. In contrast, no resistance was observed for

amikacin (a h g not used in swine production) in any of the isolates fiom this study. In a

survey study developed in the UK (1993), a high resistance was also found to apramycin

in pigs of al1 ages and humans, including hospital patients and one pig worker. Also

apramycin resistant E. coZi were found to persist in a dry environment in a pig pen which

had been empty for 10 months (Hunter, 1993). Resistance to apramycin provides cross-

100
resistance with other arninoglycosides such as gentamich and tobrarnycin (Hunter, 1993);

as weIl as neomycin, which makes this d m g a curious second choice on f m s that have

switched from apramycin. There were some other antibiotics included in the study which

are not commoniy used in the swine industry, however is interesting to know their

resistance patters as these antibiotics are used also for human medicine. It is known that

bacterias easily transfer, through plasmids, information that codes for antibiotic resistance

to other bacterias.

This study suggests that PWECD is a serious and costly problem in some f m s . The E.

coli involved tend to persist in spite of attention to the usual management, environmental

and hygiene factors. In other f m s , the disease tends to disappear, just to reappear a few

months later. The antibiotic resistance patterns varied among f m s , similar to the E. coli

serogroups. This suggests that the bacteria have been around on the farms for a while, not

the pattern one would expect if this was an E. coli spreading rapidly fiom farm-to-fm.

This was the first study of PWECD in Ontario. In order to discover the true prevalence

and economic impact of the disease in Ontario, a larger study utiliPng a random sample of

farms is required. Ln addition, there is a need to conduct careful field trials with controls to

examine the true efficacy of the many treatrnents and control measures instituted on case

farms as revealed by this study.

 Table 2.1 Proportion of positive and negative isolation of E. coli K88 (F4) between
- case and control farms as part of a post-weaning E. col..diarrhea study involving 50
Ontario farms in 1999

Cases (28) a Controls (22) P value

positive 60.7 17 13.6 3

negative 39.3 11 86.3 19 <0.001
a Grower finisher fmand fmwhich did not complete the survey are included in these data
Table 2.2 Proportion of dinerent toxïns genes identified by the Polymerase Chain
Reaction (PCR) between K88 (F4) positive case and KSS (F4)positive controi farms
as part of a study involving 50 Ontario farms in 1999

Toxias and fimbriae Positive Cases (17) a Positive Controls (3) P

value

STb LT F4 L1.8 2 100 3 0.0 1

STa STb LT F4 82.3 14 O 0.0 1
STa STb LT F4+VT F18 5.9 L O NS
Overall P value 0.01
Cases (28) Controls (22)
F18
EAE
a Grower fmisher farm and farm where producer did not complete the survey are included in these data
Table 2.3 Resistance patterns of 68 K88 (F4) positive isoIates from 17 case farms
and 8 Kt38 (F4)positive isolates from 3 control farms and resistance pattern
according to serogroups of E. coCi

Carbencillin

Neomycin
Suifamethoxazole
Cefadroxil

Tobrarnycin
Apramycin
CeRiofür

Amikacin
Emofloxacin
Cyprofloxacin
PolmVXifl B
d d O O
Grower finisher farm and f m where producer did not complete the survey are included in these data

R= Resistance

I= Intermediate
Table 2.4 Proportion of positive and negative isolation of E. coli K88 (374) between
case and control farms, as part of a survey study of post-weaning E. coli diarrhea
uivolving 48 Ontario farms in 1999

Cases (26) a Controls (22) P value

positive 57.7 15 13.6 3

negative 42.3 11 86.3 19 0.007
a Grower hisher fann and f m which did not cornplete the survey are wt included in these data
Table 2.5 Average daily gains and total nursery mortality rate between case and
control farms before and after a post-weaning E. coii diarrhea problem

Cases Controls P value

Pre-weaning average d d y gain 21 229 36 19 238 35 NS

W ~ Y )
Average daily gain of nursery 22 4 15 76 20 452 54 0.07
Wday)
Mortality before E. coli 26 2.0 1.1 20 1.8 0.69 NS
problem
Mortality after E. coli problem 26 7.7 7.3 20 1.8 0.69 <0.001
SD= standard deviation
Table 2.6 Clinical problems associated with the post-weaning E. coli diarrhea
problem, estimated from the odd ratios, between case and control farms

Clinical problems Attributable Cases (26) Controls (22) P value

observed fraction*
Nursery pig orm mat ion
Sudden death 97.1 22 3 <O.OO 1

Mild diarrhea 9 13 0.08

Profirse diarrhea 98.2 19 I <O.OO 1
Bony appearance 90 13 2 0.002

Dehydration 97 20 2 <O.OO 1
Majority of pigs affected 84.4 6 O 0.026
* Estirnated fkom odds ratio
Table 2.7 In feed antibiotics used in 12 case farms which change antibiotic after the
problem occurred, comparedd to the in-feed antibiotic used in the control farms

Cases (12) Controls (19) P value

Antïbiotic chan ed in case

f
f m s compare to antibiotic
used in control fhms
Bacitracin
Carbadox

ASP a

Tylosin,
Sulfamethazine
Chlortetracycline
Lincomycin,
Spectinomycin

a
Aureomycin, SulfamethaWie, procaine PenicilIin
Table 2.8 Management of diarrhea between case and control farms as part of a
study involving 48 Ontario farms in 1999 (1)

Cases (26) Controls (22) P value

Farms which used antibiotic

a
in water
b
Antibiotics used in water

Tiamulin O 33.3 1 NS

Tylosin 5-3 1 O NS
Nernoycin 15.7 3 33.3 1 NS
Second antibiotic used in
water
Type fantibiotic used in
water
2 P;tc~ine

nic'llin
&epiom9cin O 33.3 I NS
Neornycin 71.4 5 O 0.05

F ~ which
S used injectable
antr 1otics
Type of injectable antibiotic Ceftiofür
b
used
Lincomycin O 25 1 NS

Tiamulin 11-7 -
7 O NS

Tylosui
Enrofloxacin

a Percentage based on total of cases and control f m s

Percentage based on famis which used antibiotics
Aureomycin, Sulfarnethazine,procaine Penicillin
Table 2.9 Management of diarrhea between case and control farms as part of a study
involving 48 pig Ontario farms in 1999(2)

Vaccine against E coli diarrhea in nursery

pigs a

Acidifiers in feed a

Acidifiers in water a

Probiotics a

Sanitizers in water

Use of electrolytes when required a

Zinc Oxide r2500ppm

Management changes at the moment of
the problem a

Feed changes at the moment of the

problem
a
Percentage based on total of cases and control f m s .
Percentage based on fanns which used antibiotics
Figure 2.1 Procedure foiïowed by Gallants Laboratories Inc., for the serogrouping of
E. coli samples

Sample Entry

E coli set up on Blood agar and MxConkey plates = hemolysis

I
MacConkey plates :Lactose fermentation @ink colonies)
I
Subculture to Blood agar - 4 colonies

Serotype (agglutination) Biochemical testing

F4,lN a and 2~ Pools (pure sarnples)
I I I
+ 2N Pool = 0149:Kgl E. coli Not E. coli

Subcdture to Tryptic Soy Broth (TSB) + antibiotic sensitivity disk

Cuiture storage = bac te^ production

a
Pool 1N is for detection by slide agglutination of E. coli serogroups most commonly associated with
neonatal diarrhea in pigs and calves (08:K7'S16", 08:K25,09:KZ3,09:K28,09:K30,09:K35,09:K103,
09:Kyy79-416",020:KIOI, 064:K"v142", 08:K+)

Pool 2N is for detection by slide agglutination of E coli serogroups most commonly associated with
diarrhea in older and post-weaning pigs and with edema disease in pigs (O 138:K81, O 139:K82,
0141:K85abY0141:K85ac, 045ac:K"E65", 0157:K"V17", 0 1 15:K"V17", O 1 15:K"V165", 08:K"xlOS",
O?:K48,0 I49:Kg 1)
2.5 References

Bertshinger, H.U.: Post-weaning Escherichia coli diarrhea and edema disease Ln: Disease
of Swine 8" ed. (1999): 441-454.

Celemin, C.,Rubio, P., Echeverria, P. and Suarez, S: Gene toxin patterns of Escherichia
coZi isolated fiom diseased and healthy piglets. (1995) Vet Microb, 45: 121-127.

Dunlop, R.H., McEwen, S.A., Meek, A.H., Black, W.D., Clarke, R.C. and Friendship,
R.M.: Individual and group antimicrobial usage rates on 34 farrow-to-finish swine fanns in
Ontario, Canada. (1998) Prev Vet Medicine, 34: 247-264.

Dunlop, R.H.: Antirnicrobial treatments and antimicrobial resistance of fecai Escherichia

coli of swine in Ontario, Canada. (1996) PhD. Thesis. University of Guelph.: 275-285.

Fairbrother, J.M.: Identification, nomenclature, and diagnosis of pathogenic Escherichia

coli. (1999). Western Canadian Association of Swine Practitioners. Annual Meeting.
Proceedings, October 15 & 16, Saskatoon: 2 1-3 1.

Friendship, R.M. and Dewey, C.E.: Post-weaning Escherichia coli diarrhea (PWECD). In:
Health Management for Swine (Course notes) University of Guelph (1999): 87-90.

Gyles, C.: Escherichia coli in diseases of weaned pigs: biological aspect. h:Enteric
Diseases of Nursery Pigs. Am Ass. Swine Pract., Febr 200 1: 29-42.

Hampton, D.J.: Post-weaning Escherichia coli diarrhea in pigs. In: Escherichia coli in
domestic animais and humans. Ed. Gyles, C.L. Cab hternational(1994): 171-19 1.

Hunter, J.: Some studies on multiple resistant E. coli and the use of antibiotic in the
treatment of diarrhea in pigs. (1993) Pig Veterinary Jouniat, 3 1: 143-151.

Josephson, G. and Archambault, M.: Colibacillosis in pigs in 1999. (2000) Animai Health
Laboratory Newsletter,4: 8.
Josephson, G., Smart, N., McEwen, B. and Gough, J.: K88+ E. coli diarrhea in post
weaning pigs. Animal Health Laboratory, University of Guelph. In: 18h Annual Centralia
Swine Research Update, January 27 (1 999): 3 8-39.

Josephson, G. and Smart, N.: K88 strains of E. coZi. (1998a) Animal Health Laboratory
Newsletter, June, 2: 4

Josephson, G. and Smart, N.: Update on K88 positive E. coli. (1998b) Animal Health
Laboratory Newsletter, Sept, 2: 2.

MacKinnon, J.D.: Enteritis in the young pig caused by Escherichia coli. (1998) The Pig
Journal, 41: 227-255.

Martin, S.W., Meek, A.H. and Wïlleberg, P.: Disease Causation. In: Veterinary
Epiderniology P ~ c i p l e and
s Methods. Iowa State University Press Ames.: 121-248.

Nagy, B. and Fekete, PZ.:Entertoxigenic Escherichia coli (ETEC) in farm animds.

(1999) Vet Res, 30: 259-284.

Wood, E.N. : Fashionable and future diseases. (199 1) Vet Journal, 27: 193- 197.

Zhu, C., Harel, J., Jacques, M- and Fairbrother, J.M: Interaction with pig ileal explants of
Escherichia coli 045 isolates fiom swine with post-weaning diarrhea. (1995) Can J Vet
Res, 59: 118-123.
Chapter 3: A study investigating epidemiological risk factors associated with post-

weaning E. coli diarrhea in Southern Ontario

3.1 Introduction

An increase in the nuinber of cases of post-weaning diarrhea and rnortality associated

with K88 (F4)E. coli has been reported in Ontario since the F d of 1997 (Josephson and

Smart, 1998).The pathogenesis and lesions associated with the disease have been weU

described in the literature but information regarding nsk factors and epidemiology is sparse

and largely anecdotal (Kniffen and Neumann, 2000).

There is no doubt that toxigenic E. coZi is involved in the pathogenic process of post-

weaning digestive disorders (van Beers-Schreurs et al, 1992), but these pathogens can aIso

be found in healthy piglets raised on farms with no history of post-weaning diarrhea

problems (Fairbrother, 1999). In enzootic health disorders, such as post-weaning

colibacillosis, the presence of an infectious agent is not enough to give rise to the typical

clinical manifestations of the disease. This means that such diseases are strongly dependent

on environmental conditions, although, sufEcient infection pressure is needed for full

disease expression. Hence it can be concluded that additional factors are required for the

onset of clinical signs of disease (Fairbrother, 1999; Madec and Leon, 1999). The

conditions on many farms are such that a combination of harmful risk factors c m occur

(Skirrow et al, 1997).

The objective of this study was to investigate management, housing/nutrition factors, as

well as potential viral Sections (such as Porcine Respiratory and Reproductive Syndrome
114
(PRRS) or Transmissible Gastroenteritis (TGE) that could be associated with this new

post-weaning E.coli diarrhea in Ontario.

3.2 Material and Methods

A total of 50 f m s were visited in the summer of 1999 as part of a case-control study.

Twens-five case and twenty-five control herds were selected fiom the records of the

Animal Health Laboratory, Guelph, Ontario, and several private swine v e t e ~ a r y

practitioners. From the 1998,1999 Animal Health Laboratory records, the selection of the

farms was based on the presence or absence of K88 (F4)E. coli, in pigs that were two-to

four-weeks-old, with a history of diarrhea or sudden death. The swine practitioners

identified clients who had pigs either with or without K88 (F4)E. coli and who would be

willing to participate in the study. Farms were visited shortly after consent was acquired. A

standard protocol was followed on each f m ,

A case f m met the following criteria; pigs that were at least bvo weeks of age, had

sudden death andfor diarrhea occuning in pigs, and have a previous positive culture of K88

(F4) E. coli. A control group farm was one where pigs were at least two weekç of age, do

not have sudden death and have a previous negative culture to K88 (F4) E. c o k

During each f m visit the producer was asked to answer a survey (Appendix 1).

Information on f m hygiene, nursery management, nursery facilities, management of the

diarrhea, PRRSV status and feed management in the nursery, was collected.

Rectal swabs of ten weaned pigs were collected fiom each of the famis visited, to check

the status of K88 (F4)E. coli at the moment of the visit. The sarnples were taken fiom pigs
115
which showed clinical signs of diarrhea in case f m s and fiom pigs which looked unthrifs

or with dirty tails in the control f m s . These samples were sent to Gallant Custom

Laboratorïes Inc, Cambridge, Ontario, where culturing and slide agglutination tests were

performed. Hemolytic 0149:Kg1:K88 isolates were considered positive and were tested

for antimicrobial sensitivity. The procedures are summarized in Figure 3.1.

S e m samples were obtained fiom ten pigs (3-5 weeks of age) using the orbital sinus

method for blood collection (Huhn et al, 1969). Blood samples were aUowed to cool and

then refigerated ovemight. The blood samples were centrifuged the following day and al1

semm samples were fiozen at -70°C.These samples were tested for TGE and PRRSV, and

results were compared to the post-weaning E. c d i diarrhea status. TGE was tested using

the coronaWus differential ELISA on positive semm samples for PRCV-TGE obtained

fiom previous Wus-neutralization tests. PRRSV was tested using the IDEXX ELISA test.

Both TGE and PRRSV tests were performed by the Animal Health Laboratory, University

of Guelph, Ontario. The following were w d to group farms according to their PRRSV

status:

O S:P ratio (0.4 in al1 sampIes

1 S:P ratio between 0.4 and 2.5 in at least one sample

2 S:P ratio >2.5 in at least one sample

A score of one is usually a result of matemal antibodies, past infection, or vaccination.

A score of two implies recent field infection, usually referred to as active PRRSV infection.

Environmental and water samples were taken fiom f m s which had an empty,

disinfected nursery room, in order to monitor rnicrobial loads in the facility and sanitation
116
quality, using the Microenviro Kits, Microlab Diagnostic Inc, Guelph Ontario. The kit

consists of an agar slide with two surfaces, one with Tryptone Glucose Extract Agar

TGEA (plate count aga) for the test of the total number of living bacteria per area tested,

and the other with MacConkey Agar, which is a selective medium for gram negative

bacteria. Colonies capable of fermenthg lactose, such as KlebsieZZa and E. coli, are red in

color in this media. Samples were incubated at 37°C for 24 hours, and the counting of

bacterial colonies was compared to specific score codes included in the kit. The counting

was done as follows:

Score code TGEA MacConkey

(0) No bacteria isolated No bacteria isolated

Excellent (1) 1-10 1-3

G O O ~ (2) 11-30 4-8

Fair (3) 3 1-100 9- 15

Poor (4) lOO+ 15+

The simple associations between case and control herd statu and management and

disease factors were determined using a Chi-square and Fisher's exact test for qualitative

variables and Student's T-test for quantitative variables. The Kruskd-Wallis one-way non-

parametric test, was used to compare categorical data. The prevalence of each factor

measured was cdculated for case and control fanris. A p<0.05 was considered as

significant, whereas p- values between 0.06 and 0.1 were considered numerically reportable

as potential trends. Significant variables were tested for interaction with a Mantel-Haenszel
117
procedure, Risk factors related to the case and control class~cation(PC0.25) were re-

examined in a multivariate model using logistic regression. Models were built using

backwards elimination; those with the highest p-value were removed one at a time until ail

factors left in the model were statisticaliy sigaincant at p< 0.10. Statistical analysis was

completed in Statistix (using version 1.O) (1996) (StatistixB Analytic Software,

Tallahassee, Florida).

3.3 Results

On three control herds, pigs developed diarrhea problems and were diagnosed with K88

E. coli just before the fanri was visited. These three faniis were reclass5ed as cases. On

one case herd the survey was not completed. Another case farm was found to be a grower-

finisher barn. Both f m s were dropped fiom the analysis. Twenty-six f m s were

designated as cases and twenty-two f a m s were designated as controls. Within the 26 case

f m s , 15 were K88 (F4) E. coli culture positive. Within the 22 control farms, 3 resulted in

positive identification of K88 (F4) E. coli, but diarrhea problems were not apparent. Case

f m s tended to be more likely to yield positive resdts to the culturing of K88(F4) E. coli

than control herds @ =0.007) (Table 3.1).

Case f m s had more feeder spaces per pen (7.7 SD=3.5) than control f m s (5.2

SD=2.3) (p= 0.008) (Table 3.2). No significant differences were found for density, average

weaning days, weight at weaning, level of protein, and fat or fibre in the fkst feed offered

to the pigs in the nursery.

 No significant differences were found between case and control f m s using specific

management systems, however, 46.1%of the case f m s were farrow-to-finish compared

with 27.3% of contrd f m s , and 54.5% of the control f m s were off-site nurseries,

compared with 34.6%of c a s e - f m s (Table 3 -3). Interestïngly, a grower-finisher barn was

affected with the E. coli problem. No s i g d c a n t Merences were found regarding type of

flooring and whether or not creep feed was offered before weaning between case and

control f m s , however, 76.9 % of the case farms offered creep feed before weaning

compared to 59.1% of control f m s (Table 3.4).

A higher number of case f m s (57.7%) blended the nursery feed between phases of

feed than control f m s (36.3%); more control famis (95.4%)offered full access to feed

when pigs were moved into the nursery than case farms (88.4%);and a highly significant

difference was found between case and control f m s with regard to changes in feeding and

management @ <0.001). The most common feed changes were: purchasing feed fiom a

different supplier; changing the feed medication; offering a limited amount of feed four or

five times per day; blending of feed between phases; decreasing the level of protein and

increasing the level of fibre. The most common management changes were: increased age

at weaning, better control of temperature and ventilation; creation of sick pens; decreased

density in pens; and reduced inixing of pigs.

Case famis were more likely to use pelleted and crumbled feed (88%) in the first ration

offered in the nursery, compared to control f m s (59%), and control farms were more

likely to use mashed feed (41%) than case f m s (12%) @= 0.02) (Table 3 -4).
 For data related with the cleaning and disinfection procedures, it was found that more

control farms (54.5%) used detergent than case farms (26.9%) @= 0.05) (Table 3 -5)-No

significant ciifferences were found between case and control fanris regarding the sources

per pen and room in the nurseries (Table 3.6). The use of mashed feed versus pellet and the

use of detergent were tested for interaction within the PRRS positive and negative status of

the f m . No interaction was found between PRRS status and the use of detergent and

mashed feed versus pellet.

Within the 26 case farms, 7 were negative for PRRSV. The remainder consisted of 17

f m s with low titres (possibly resulting fiom vaccination or passive immunity) (S:P ratio

between 0.4 and 2.5) and two f m s with hi& titres indicative of recent PRRSV infection

(S:P ratio >2.5). Within the 22 control farms, 6 were negative for PRRSV, 16 tiad low

titres, suggestive of vaccination or passive immunity and no control farms presented titres

indicative of active PRRSV infection- There was no association between PRRSV status

and post-weaning E. coli diarrhea @= 0.4). However, more case farms vaccinated their

nursery pigs against PRRSV (40%) than control farms (13.63%0) @= 0.04). Vaccination

of gestating sows against PRRSV, was also more likely to occur in case farms ( 56.52%)

than in control f m s (25%) (p= 0.03) (Table 3.7). Within the PRRS positive status, it was

found that after controlling for the sow vaccination, nursery vaccination was not important

as a risk factor for E. coli. After controlling for nursery vaccination, sow vaccination was

not also important as a risk factor for E. coli. These two variables were highly associated

between each other @=0.00S) (Table 3.8), which indicate that might be confounding.
 The final multi variate mode1 included 'PRRSV vaccine use in sows7@= 0.05); 'use of

pellet feed versus mashed' (p- 0.03) and 'use of more feeder space per pen' @= 0.02)

(Table 3.9).

Within the 26 case f m s , two were positive for TGE with the coronavinis deferential

ELISA test and one of the 22 control farms was positive for TGE. No association was

found between post-weaning E. coli diarrhea and TGE @- 0.6) (Table 3.7).

Summary of the results for the assessrnent of bacterial contamination of nursery

facilities and water contamination, are presented in Table 3.10. No significant differences

were found for the total living bacteria per area between case and control f m s (p= 0.4).

Case famis were more likely to have high levels of water contamination with colifomis

(Score 4 = poor) compared with control f m s (Score 1= excellent) (p= 0.01).

3.4 Discussion

There were no differences between case and control farms for a nurnber of factors

commonly thought to be associated with nursery pig diseases (Josephson and Smart, 1998;

Josephson et al, 1999). In particular, early weaning age and the commhgling of pigs fkom

multiple farms into one nursery room, practices commonly associated with segregated early

weaning (SEW) production systems, were not found to be associated with a higher

likelihood of PWECD. In this study al1 types of farrns were represented in both case and

control in approximately equal proportions, and weaning weights and ages were similar

between case and control farrns. Weaning at an early age has been considered a risk factor
for PWECD, because the intestinal tract in very young pigs has not produced proper

enzymes and active immunity (Valery, 1995; MacKinnon, 1998).

It was expected that some factors associated with the case farms were more likely a

response to the diarrhea problem rather than causative factors; for example, case f m s

were more likely to blend rations to reduce the effect of a sudden feed change compared to

control farms, and case f m s were more likely to limit-feed pigs- Similarly, case farms

more fiequently reported recent feed changes and management changes compared to

contro1farms.

Differences in feeding between case and control famis which may t d y reflect a risk

factor include: pellet versus mashed feeds, and number of feeduig places. Several

researchers have suggested that heat treated cereds such as pelleted feed, might provide a

better environment for E. coli proliferation (Lawrence, l98S), by increasing the amount of

available nutrients required by the E. coli growth plus increased consumption and speed of

feed passage. This study suppoas this concem. Feeder space availability might ailow or

encourage higher levels of feed consumption and lead to diarrhea.

It was expected that farms with E. coli diarrhea problems, may not do as good a job

cleaning as a f m that didn't see the problem; however, there was little evidence of this.

Most fârms cleaned pens between groups of pigs whether or not E. cdi was a problem.

More control f m s used a detergent as well as a disinfectant in the cleaning process and in

addition, water lines of control f m s were cleaner than case f m s . These two h d i n g s

suggest that possibly cleaning was better on control fxms compared to case f m s , but in

general there was little obvious difference between the two groups as far as sanitation and

122
cleaniiness. In al1 likelihood, pigs carry the E. coli into the nursery in intestinal flora

(Hampson et ai, 1987) and the cleaning procedure rnay not be very relevant.

There is clear evidence that hi& bacterial counts in water in itself can be a cause of

post-weaning diarrhea (Skirrow et al, 1997). In o u .study, it was observed that case farms

had higher coliform counts in water than conbol farms, however, the number of farms

tested is very limited and information cannot be extrapolated to all Ontario f m s .

The relationship between PRRSV vaccine use in gestating sows and in nursery pigs and

the presence of E. coZi is interesting. PRRSV vaccine is a live vaccine and has been

associated with the birth of weak viremic pigs (Mengeling et al, 1998,1999; Dewey et al,

1999). There are reports that PRRSV may exacerbate E. coli problems (Done and Patron,

1995; Nakamine et al, 1998). There is some biological b a i s to believe that vaccination of

pregnant sows or weanling pigs could contribute to an E. coli diarrhea and therefore thk

association needs to be further investigated.

The most important fïndings of this study were, that many of the factors most

commonly considered to be risk factors such as weaning age and mixing of pigs fiom

multiple sources did not appear to be important contributhg factors, and that previously

unstudied factors such as PRRSV vaccination practices appeared to be related to PWECD.

This study would suggest that there is not a single, obvious risk factor responsible for

tnggering E. coli outbreaks in Ontario nurseries.

Table 3.1. Proportion of positive and negative E. coli Ka8 (F4)isolates between case
and control farms as part of a study invoIving 48 Ontario fanns in 1999.

Cases (26) a Controls (22) P value

positive 57.7 15 13.6 3

negative 42.3 11 86.3 19 0.007

a
Grower fmisher fmand farm where producer did not complete the s w e y are not included in this data
Table 3.2 Quantitative variables tested for association with an E. coli diarrhea
problems in 48 Ontario nurseries in 1999

Cases Controls P value

Number of feeders per pen

Total square feet per pig
Average weanhg weight
Average weaning days
Level of protein of the £ k t
feed %
Level of fat of the first
feed %
Level of fibre in the fkst
feed %
SD= standard deviation
Table 3.3 Use of specific management systems between case and control farms, as
part of a study investigating risk factors associated with post-weaning E. coli
diarrhea

Cases Controls P value

- -

Farrow-finish 26 46.1 12 22 27.3 6 NS

Farrow-partly f i s h 26 3 -8 1 22 9.1 2 NS
Farrow-to-feeder 26 15.4 4 22 9.1 2 NS
Off-site 26 34.6 9 22 54.5 12 NS
Table 3.4 Qualitative risk factors tested for association with an E. coCi diarrhea
problem in 48 Ontario nurseries

Cases Controls P OR

Total solid floor

Total sIotted floor
Mixed floor
Pellet feed vs mashed in
first nursery feed
Farms which offered creep
feed before weaning
Nurseries which were fdl
feed access
Farms which blended feed
between phases
Farms which made
management changes
Fanns which made feed
changes
Table 3.5. Cleaning and disinfection procedures between case and control herds as
part of a study investigating risk factors associated with post-weaning E. coli
diarrhea

Cases (26) Controls (22) P value OR

Scraping
High pressure
Hot water
Detergent
Disinfect the nursery
Disinfect the feeders
Disinfect the diinkers
Dry 12 hours
Dry 24 hours
Disùi£ect the water lines
Table 3.6 Sources per pen and room between case and control herds as part of a
study investigating risk factors associated with post-weaning E. coli diarrhea

Sources per From one litter 19.2 5 4.5 1 NS

Pen
More than one Iitter 46.2 12 45.4 10 NS
More than one farrowing 19.2 5 18.2 4 NS
roorn
More than one sow farm 15.4 4 3 1.2 7 NS
Sources per From one farrowing roorn 53-8 14 3 6.3 8 NS
room
More than one farrowing 26.9 7 27.3 6 NS
room
More than one sow farm 19.2 5 36.3 8 NS
Table 3.7. Porcine Reproductive and Respiratory Syndrome (PRRS) and
Transmissible Gastroenteritis (TGE) information, as part of a study investigating
risk factors associated with post-weaning E. coli diarrhea

Cases Controls P value OR

PRRSV status at the O 26 26.9 7 22 27.2 6 NS

time of visit,
according to S:P 1 26 65.3 17 22 72.7 16 NS
ratios a ,using
ELISA 2 26 7.6 2 O NS

Nursery pigs vaccinated 25 40 10 22 13.6 3 0.04 2.9

against PRRSV
Gestating sows vaccinated 23 56.5 13 20 25 5 0.03 3.9
against PRRSV
TGE status at time of visit 26 7.6 2 22 4.5 1 NS
tested by ELISA

a O S:P ratio ~0.4

in d sarnples

I S:P ratio between 0.4 and 2.5 in at least one sample

2 S:P ratio >2.5 in at least one sample

Table 3.8 Relationship between sows vaccinated against PRRSV and nursery pigs
vaccinated against PRRSV.

Nursery pig Nursery pig Total

vaccinated not vaccinated
Sows vaccinated 8 10 18
Sows not vaccinated 2 21 23
Total 10 31 41
p= 0.008
Table 3.9 Final mulai-variate analysis for the post-weaning E. coli diarrhea study

SE P-value Odds ratio

Gestating sow vaccinated against PRRSV 0.9 0-05
Pellet feed versus mashed 1.4 0.03
Number of feeder space per pen 0.2 0.02
SE= standard error
Table 3.10 The average score code of microbial contamination in the nursery
facilities @=IO) after disinfection, for total number of living bacteria, and the
average of water coliform contamination between case and control farms

E. coli isolates a Average score code Water score code

Cases
1.85 4
2 4
1.9 4
Total average score 1.9 (good) 4 (poor)
Controls

Total average score 1.98 (good) 1 (excellent)

a 1= positive

O= negative
 Figure 3.1. Procedure foiïowed by GaIIants Laboratories Inc., for the serogrouping of

E. cuii samples

E coli set up on Blood agar and MacConkey plates : hemolysis

MacConkey plates = Lactose fermentation (pink colonies)

I
Subcdhue to Blood agar - 4 colonies

Serotype (agglutination) Biocbemicai testing

F4, INa and 2~ Pools bure samples)
I I I
+ ZN Pool 0149:Kgl
: E. coli Not E. coli

Subcdture to Tryptic Soy Broth (TSB) + antibiotic sensitivïty disk

Incubate 37°C 18-24h

I
Culture storage : Bacterin production

a
Pool 1N is for detection by slide agglutination of E. coli serogroups most cornmonly associated with
neonatal diarrhea in pigs and calves (08:K"Sl6", 08:K25,09:K28,09:K28, 09:JK30,09:K35, 09:K103,
Og:K"79-4 l6", 020:KlO 1,064:KV l42,08:K+)
b
Pool 2N is for detection by slide agglutination of E. coli serogroups most commonly associated with
diarrhea in older and post-weaning pigs and with edema disease in pigs (0138:KS lY0139:K82,
0141 :K85ab, 0141:K85ac, 045ac:K"E65", 0157:K"V17", 0 115:K"V17", 0 115:K"V165", 08:KX105,
O?:K48,0149:K91)
3.5 References

Dewey, C.E.Wilson, S., Buck, P. and Leyenaar, J.K.: The reproductive performance of
sows after PRRS vaccination depends on stage of gestation. (1999) Prev Vet Med., 40:
333-241.

Done, S.H. and Patron, D.J.: Porcine reproductive and respiratory syndrome: clinical
disease, pathology and irnrniuiology suppression. (1 995) Vet Record, 136: 32-3 5.

Fairbrother, J.M. :Identification, nomenclature, and diagnosis of pathogenic Escherichia

coli. (1 999). Western Canadian Association of Swine Practitioners. Annual Meeting.
Proceedings, October 15 &16, Saskatoon: 2 1-3 1.

Huhn, RG., Osweiller, GD. and Switzer, W.P.: Application of the orbital sinus bleeding
technique to swine. (1969) Lab Animal Care, 19: 403-405.

Hampson, D.J., Fu, -A.F. and Robertson, I.D.: Investigation of the source of hemolytic
Escherichia coli infectkg weaned pigs. (1987) Epidem Infect., 99: 149-153.

Josephson, G. and Smart, N.: Update on K88 positive E. coli. (1998) Animal Health
Laboratory Newsletter Sept, 2: 3.

Josephson, G., Smart, N., McEwen, B and Gough, J.: K88+ E. coli diarrhea in post-
weanhg pigs. Animal Health Laboratory, University of Guelph. In: 18" Annual Centralia
Swine Research Update, January 27, 1999: 38-39.

-en, T. and Neumann, E.: Field experience with E. coli K88. In: Emerging Nursery Pig
Diseases and Management Solutions. 3 1%Annual Meeting Am. Ass. of Swine Practitioners.
Indianapolis, Indiana, USA (2000):3 1-34.

Lawrence, T.L.J.: Processing and preparation of cereals for pigs diets. (1985) In: Recent
Developments in Pig Nutrition. Ed. Cole, D.J.A. and Haresign, W: 230-245.
MacKinnon, ID.: Enteritis in the young pig caused by Escherichia coli. (1 998) The Pig
Jounial, 41:227-255.

Madec, F. and Leon, E.: The role of management and husbandry in pig health, with
emphasis on post-weaning enteric disorders. In: ManipuIating Pig Production VI1
Symposium. Ed. Cranweii P.D. Australian Pig Science Ass. Conference 7U. Pig Research
and Development Corporation. Nov-Dec, (1999): 200-209

Mengeling, W.L., Lager, K.M. and Vorwald, A.C.: Safety and efficacy of vaccination of
pregnant gilts against porcine reproductive and respiratory syndrome. (1999) Am Jour Vet
Res, 60: 796-80 1.

Mengeling, W.L., Lager, K.M. and Vorwald, AC.: Clinical consequences of exposing
pregnant gilts to strains of porcine reproductive and respiratory syndrome (PRRS) virus
isolated 6 o m field cases of "atypical" PRRS.(1998) Am Jour Vet Res, 59: 1540-1544.

Nakamine, M., Kono, Y., Abe, S., Hosino, C., Shùai, J. and Ezaki, T.: Dual infection with
enterotoxigenic Escherichia coli and porcine reproductive and respiratory syndrome virus
observed in weaning pigs that died suddenly. (1998) J Vet Med Sci, 60: 555-56 1.

Skirrow, S.Z., Buddle, J.R, Mercy, A.R, Madec, F. and Nicholls R.R.: Epidemiological
studies of pigs .diseases: 2. Post-weaning diarrhea and performance in Western Australian
pigs. (1997) Aust Vet Jour, 75: 282-288.

Valery, M.A.: Behaviourial patterns of the weaned piglet. (1995) The Pig Journal, 34: 71-
79

Van Beers-Schreurs, H.M., Vellenga, L., Wensing T, Breukink, H.J.: The pathogenesis of
the post-weaning syndrome in weaned piglets: a review. (1992) Vet Q, 14: 29-34.
Chapter 4: Study of the efficacy of two vaccines: intramuscular autogenous and oral

h e K12-K88vaccines for the control of post-weaning E. coli diarrhea in pigs.

4.1 Introduction

Post-weaning diarrhea caused by K88 (F4)E. coli is an important disease problem in

confinement swine operations (Nagy and Fekete, 1999; Fairbrother, 1999). The disease is

responsible for economic losses due to m o r t d i ~decreased

, growth rate, and cost of

medication (Josephson et al, 1999; Fairbrother, 1999; Bertschinger, 1999).

A variety of control measures have been employed including, the use of antibiotics or other

antibacterial agents in the feed such as zinc oxide, probiotics, acidifiers, improved hygiene,

changes in feed composition and changes in feeding practices (Hainpson, 1994;

MacKinnon, 1998). Producers have often experienced short periods of success in

combating the disease, o d y to have the disease reappear a few weeks later.

Some veterinary practitioners have used vaccination strategies. Vaccination against this

diseâse has not shown much success, although the theoretical basis is clear and the need is

unquestionable (Francis and Willgohs, 1991;Nagy and Fekete, 1999).

It is known that irnmunization of sows with cornmercially available vaccines for

neonatal E. coli is not effective for the prevention of post-weaning E. coli diarrhea, since

specific lactogenic and colostrum antibodies are no longer available to the weaned piglets

(Fairbrother, 1999; Van den Broeck et al, 1999qb). Active immunity induced by

vaccination of the piglets before weaning might be more effective to protect them against

post-weaning E. coli diarrhea (PWECD).

137
 It has been shown that parenteral imrnunization stimulates littie or no local irnmunity in

the piglet intestine (Francis and Willgohs, 1991;Moon and Bunn, 1993; Laval, 1996, van

den Boeck et al, 2999b) and efficacy fiom parenteral vaccines against PWECD has not

been convincingly demonstrated. However, autogenous parenteral bacterins are being used

in Ontario for control of PWECD.

The most promising experiments have been done in the area of live oral vaccines applied

before weaning. The development of these vaccines h a . focused on the production of

antigens through genetic e n g i n e e ~ (Laval,

g 1996; Bianchi et ai, 1996).

The objectives of this trial were to determine the usefulness of an oral geneticaliy

altered K12-K88 vaccine and an autogenous intramuscular killed vaccine as methods of

controllhg post-weaning diarrhea and mortality on two farms in Ontario.

4.2 Materials and Methods

A total of 50 Ontario swine farrns were visited in the summer of 1999 as part of a case-

control study. From the group of 28 case fanns, two f m s were chosen as sites for

vaccination trials. The farms were selected based on previous history of post-weaning K88

(F4) E. coli diarrhea, including laboratory isolation of K88 (F4) positive E. coli fiom the

f m s ' pigs and for the farmers' willingness to cooperate and maintain accurate records.

4.2.1. E. coli strains and vaccine preparation

Preparation of the geneticalZy altered vaccine. An Escherichia coZi KI2 contalliing a

plasmid which encodes the K88ac antigen (Kl2-K88ac strain) was used for the preparation

of the vaccines. The strain was grown overnight at 37°C on brain heart infusion (Difco,
138
Detroit, MI) agar in Iarge petn plates, each containing ZOO ml of medium. The next

moming the bacterial growth was harvested with a bent g l a s rod, and resuspended in

sterile phosphate-b&ered saline (PBS) at the rate of 50 ml of PBS for each agar plate.

Following centrifùgation at 10,000 x g for 10 minutes, the supernatant was decanted and

the pellet was resuspended in pasteurized whole cow's m i k at the rate of 20 ml of milk for

each 50 ml of PBS suspension. Dilutions of the bacterial suspension were plated in order to

determine the concentration of bacteria These were approximately 10" per ml. The

bacterial suspension was stored in sterile plastic tubes at 5 OC for no longer than 24 hours,

allowed to reach room temperature, and shaken thoroughly before use.

Preparation of fie autogenous killed vaccine: The bacterin was prepared at Gallant

Custom Laboratories, Inc, Cambridge, Ontario. E. coZi sîrains (0l49:Kgl K88ac) used for

the vaccine were f m specific. Bacteria were cultured overnight (18h) in Tryptic Soy

Broth (TSB). Sarnples of the suspension were cultured on blood agar pIates for the test of

puris.. Preservatives and bactericides were added to the broth and subsequently, bacteria

were killed by heat treatment (6hrs at 56°C). After k a t treatment, reculture on blood agar

was performed to test for sterility, Suspension with the killed bacteria was blended with an

adjuvant (aluminum hydroxide) and tested again for sterility on blood aga. Concentration

of the bacterin per dose was 5x10 .The vaccines were stored in sterile plastic bottles at

5 OC until they were used.

 4-2.2. The farms

Farm 1 was a 250 sow farrow-to-finish operation, weaning pigs at 4 weeks of age. The

f m had 3 nursery rooms with 8 pens per room which did not allowed direct pen-to-pen

contact. The pens were 100% slatted and held an average of 2 2 pigs, resulting in a density

of 0.18 square meters per pig. The f m was diagnosed with E. coli in the fall of 1998.

Pure cultures of a K88 (F4) E coli with genes for three toxins (STa,STb,LT) were

consistently isoIated fkom pigs with diarrhea. Post-weaning mortality ranged fiom 5- 10%

and diarrhea affected about 70% of the pigs. The f m e r had unsuccessfùlly tried multiple

antibiotic treatments, zinc oxide in feed, acidifiers in water, feed changes, and management

changes.

Farm 2 was a 500 sow fmow-to-feeder pig operation with a similar history of

persistent post-weaning dimhea and losses associated with a K88 (F4) E. coli (STa, STb,

LT). The farm was diagnosed with E. coli in December of 1998. MortdiSr ranged fiom 8-

10%. This operation had 6 nursery rooms with 6 pens each îhat held 50 pigs, having a

density of 0.27 square meterdper pig. Pens were 100% slatted and direct contact between

pigs was prevented. The average age at weaning was 18 days, with a minimum of 13 days.

Because the farmer filled just four pens each time, he weaned 200 pigs, there was one point

where different ages of pigs were present in the same room.

4.2.3. Vaccination protocol

One week prior to weaning, litters were assigned to one of three treatment groups:

-Autogenous intramuscular killed bacterin ( I d )

-Live oral attenuated vaccine (1ml)

140
 -Control (untreated)

The oral vaccine was instilled deeply into the mouth by means of a 5 cm long plastic

tube fitted to a syringe. The intramuscular vaccine was administered into the neck of the

piglets.

Treatments were given one week before weaning, and were assigned by litter since the

orally vaccinated pigs would be shedding the bacteria in their feces and there was a danger

of spread among Litter mates. At this time they were tagged, weighed, and given the

appropriate vaccine. At weaning, a second dose of the oral vaccine was given to the pigs,

and they were moved into one nursery room. As was standard practice on the f m s ,

weaned pigs were sorted according to size, mixing different litters in one pen, but keeping

the live vaccinated piglets in separate pens. The feed offered to the piglets in the farrowing

room and for the first week in the nursery was unmedicated. This was necessary for the

survival of the bacteria in the oral vaccine, but was used for al1 three treatment groups.

Over the course of the study, f m e r s were asked to keep records of mortality due to

diarrhea and of treatments given to diarrhetic pigs. Proportion of pigs treated was used to

analyze the Ievel of morbidity in the nursery pigs. Average daily gain and mortality due to

E. coli diarrhea and scores of diarrhea, after one week of weanhg, were also recorded. The

following diarrhea scoring system was used (Sarmiento et al, 1988):

O = Çn feces

1 = soft feces

2 = fluid feces

3 = projectile diarrhea
 Scores of two and three were considered as diarrhea, and zero and one were considered

normal.

On f m 1, the trial was conducted using four replicates, wiîh 1 13, 130 and 123 pigs

assigned to the control, killed autogenous and Iive attenuated treatment, respectively. In

this f m , the pigs were weighed again one and three weeks after weaning. Over the course

of the study, the farmer kept records of mortality due to diarrhea and of treatments given

to pigs with diarrhea. The farmer treated diarrhetic pigs with lcc of gentamicin given

intramuscularly and 2 ml of neomycin given orally. A total of 56 rectal swabs were taken

£iom three to five-week old pigs with diarrhea during the trial.

Based on the results obtained after the field trial, it was arranged that the f m e r wodd try

the vaccine with the best results, as a routine procedure. As a follow up the first two

groups were to be monitored in a similar fashion to the originai trial wiîh the exception that

pigs were not to be identified and weighed individually.

O n f m 2, the trial was conducted using two replicates with 110, 134 and 148 pigs

assigned to the control, killed autogenous and live attenuated treatment respectively. In this

case, the pigs were weighed for the second time at the day of weaning (7 days post-

vaccination), and for the oral vaccine, a different vaccine strain was used for the second

dose of the &st batch of pigs and the first dose of the second batch of pigs. The strain used

was an M 189 (08-060:F4+:LT-; STb-;STa-; VT 1-;VT2-;CNF-;HIy-) which was

considered to be a more robust strain than the K12-K88straïn that had been used on the

first f m . This strain was previously used by Dr. Fairbrother in another trial, with positive

results.
 4.2.4. Statistical analysis

Proportion of pigs treated, and scores of diarrhea were used to analyse the level of

morbidity in the nursery pigs. Mortality and average daily gain were also analysed

statisticaily to evaluate the severity of ihess.

Average daily gains were analyzed by analysis of variance of the SAS program

(PC/SAS@ SAS Enstitute Inc. Cary, North Carolina) and in Statistix @ (Analytic SoEtware,

Tallahassee, Florïda). Descriptive statistics for numerical data and test for normality were

performed with the PROC UNIVANATE in SAS program (PC/SAS@ SAS Institute hc.

Cary, North Carolina). Scores of diarrhea by treatment, were anaiyzed by the Kruskal-

Wallis one-way non-pararnetric test with Statistix @ (Analytic Software, Tallahassee,

Florida). Treatment groups were tested for association with mortality due to E. coli, and

used of treatments against E. coli diarrhea, with Chi-square test. Mortality due to diarrhea

was compared among the different groups with two by two tables. A p value lower than

0.0 17 was considered as signincant, after Bonferroni correction was performed (Norman

and Streiner, 2000). Mortality was aiso analyzed controlling for sow, pen, room, with

Logistic Regression in a Generalized Esthnating Equations (GEEs) in a GENMOD

procedure in SAS program.

There was no association between total average daily gain and vaccination status (p= 0.44)

(Table 4.1). No significant Ievel of difference was detected between the three groups in the

average daily gain fiom the one week pre-weaning to the one week post-weaning period

143
(p= 0.63), or fiom the one week post-weaning to the three week post-weaning penod (p=

0.104). However, when comparing between treatrnent groups in the last period, the pigs in

the killed autogenous vaccine group showed a lower average daily gain compared to the

pigs in the live attenuated vaccine group (p= 0.03).Vaccination was not associated with

morbidity as measured by the number of pigs that were treated for diarrhea For the scores

of diarrhea, the control group showed more diarrhea than either the two vaccinated groups

(p= 0.0 16) (Table 4.2, Figure 4.1).

The mortality rate due to diarrhea for pigs given the injectable vaccine were lower than

that for either the control pigs or the pigs given the oral vaccine k-0.02) (Table 4.1).
After controlling for the sow, pen, and room of origin, pigs given the killed intramuscular

vaccine tended to have a lower mortality due to E. coli @= 0.08) than the control group

and the pigs given the oral vaccine.

Cultures of rectal swabs showed variable results among the treatments. Some samples

contained non-hemoiytic E. col&others showed heavy growth of hemolytic E coli and

others the strain identified as 0149:Kg 1K 8 8 . Some cultures were subsequently sent to the

University of Montreal for their categorization by the PCR test. The presence of an VT

F 18 strain was also dernonstrated fiom this f m .

Based on the low death rate and the low presence of diarrhea for the autogenous

intramuscular vaccinated pigs, the farmer chose to try this bacterin as a routine procedure

dong with zinc oxide and antibiotic medicated feed. Two batches of pigs (n= 2 10) were

followed after the field trial.The results of the case study follow-up to the field trial are

shown in Table 4.3. The number of pigs treated for diarrhea was 19 (9.04%), and the

144
number of pigs that died due to E. coli was 12 (5.71%). Scores of diarrhea were aiso

determined for these two groups (Table 4.4). Results of the autogenous vaccine on îhis

case study with the resuits of the use of the same vaccine for the field trial were compared

(Fig 4.2). Mortaîity and morbidity increased in this case study follow-up compared to the

previous vaccine comparison study (Fig 4.3). Mortality between the parentedly vaccinated

pigs in the follow-up study increased signifïcantly compared to the previous field trial @=

0.02). The vaccine program was discontinued because it did not appear to prevent

mortality against E. coli diarrhea.

On Farm 2, severe diarrhea began within two or three days of weaning, with more than

50% of the pigs affected in each of the treatments. The farmer started to medicate with

apramycin in water (100mgA)six days after weaning. Within that week, eight pigs died

fiorn diarrhea, but unfortunately, identification was not recorded by treatment. One sick pig

was taken to the Animal Health Laboratory (AHL) at the University of GueIph for anaiysis.

Gross pathology and bactenology was performed, and post-weaning colibacillosis was

diagnosed. For the second vaccinated group, the picture was similar to the first group of

vaccinated pigs. Although the farmer started to medicate the fust day when diarrhea was

observed, six pigs died within that week. At the same time the farm started to have severe

diarrhea problems in the fanowing rooms, losing complete pens of pigs. The vaccination

aial was quickiy halted as a result of the severe diarrhea and high mortality. In this case,

four pre-weaned pigs were sent to AHL in Guelph, and coliform enteritis was diagnosed.

Histopathology was also performed in al1 pigs because Transmissible Gastroenteritis (TGE)

145
was suspected, but no histological evidence of virai infection was observed in any of the

pigs. One week after the trail was stopped, the f m e r was still losing complete Litters of

pigs, and at this point, three pre-weaned pigs were sent again to AHL.Blood samples were

taken fi0111 four (eight-to-ten-day old pigs) three (23-25-day-old pigs), and five (seven-to-

eight week-old pigs). Although histopathology of the three pre-weaned pigs showed no

evidence of viral infection, eleven of the twelve serum samples were positive to (TGE) with

the PRCWTGEV differentiating ELISA test.

4.4 Discussion

There are some reports of farms which successfully control PWECD with the use of a

vaccine consisting of killed cultures of 0l49:K8 8 given parenteraliy to pre-weaning pigs

(Wood, 1991; Connaughton et al, 1992). A number of farms in Ontano are using such

products and are satisfÏed with the result. For a brief period the owner of farm 1 was

convinced that a killed autogenous vaccine was working. Subsequently failure of the

autogenous intramuscular vaccine may have been predictable. There is little scientific bais

to support the use of an intramuscular vaccine against E. coli diarrhea. These vaccines,

tend to stimulate systemic rather than rnucosaI immunity. In fact, Bianchi et al, (1996)

suggested that parenteral vaccination suppresses the local anti-F4 response upon infection

of the intestine with E. coli F4 strains. In this study, one possible scenario of the apparent

efficacy of the intramuscular vaccine in the trial is that the vaccine stimulated a generalized

immunity. However it is more likely a result of mere chance.

 A live oral vaccine is more Uely to be effective. In Australia, Faghy et al (1987,1992)

had success in controlhg and reduchg the severity of PWECD in several farrns. Australian

vaccines containing live E. coli are commercially available and used on a regular bais in

many farms with reportedly good results. However, in our study, results were

disappointing. Possibly the (K12-K88) strain could not successfÛUy colonize the intestine,

or perhaps conditions in the small intestinal tract were not appropriate for expression of the

fïxnbriae of the vaccine organism. Et has been suggested that suckling piglets are abIe to

produce intestinal antibody in spite of the presence of specific antibodies in both, colostrum

and milk of the mother (Olsson et al, 1986). In addition, Francis and Willgohs (1991)

reported that antibodies in the sow's milk may in fact prevent the vaccine strain fiom

colonizing suniciently to stimulate protective immunity. Even though, in this study, sow's

m i k was not tested for antibodies against fimbriae expressed by the vaccine strain, it is

probable that some were present.

In f m one, F 18 serogroup was also demonstrated in some cultures, leading to another

possible potential problem with the live oral vaccine, which just included a K12X88 strainain

In Canada, fimbrïae F 18 has been shown to play an important role in PWECD of pig

(Fairbrother, 1999), and they should be considered as potential vaccine antigens.

Specific circumstance, that should be taken into consideration for the vaccine failure,

was that the diarrhea occurred too soon after weaning, not giving the pig an opportunity to

build sufncient immunity. Also, the presence of other diseases (such as TGE) may have

confounded the results. In addition, a negative aspect of the live vaccine is that antibiotics

need to be taken out of feed, and therefore, conducting the trial under these circumstances

147
may make the PWECD problem worse. T h i s codd also be the reason why the second farm

had a severe probIem of diarrhea during the trial.

The concurrent TGE and K88 (F4) E, coli outbreak and the absence of ântibiotics in the

feed, was likely responsible for the drastic outcome on the second f m . Concurrent

endemic transmissible gastroenteritis and colibaciiiosis infection has been described by

Pritchard (1987) and Dewey et al (1999). In agreement with their findùlgs, in this study,

c h i c a l signs, postmortems and bacteriological hdings were often more suggestive of

colibacillosis than a viral etioiogy.

The results of our study showed that vaccination either with a killed intratlluscular

bac te^ or live oral vaccine did not appear to provide a solution against post-weaning E.

coli diarrhea in these two farms. It is possible under Werent circumstances, vaccines may

have offered some protection. Further studies with a larger number of f m s and animals,

are needed.
Table 4.1 Results of Kg8 (F4) Escherichia coli vaccination with kiIled autogenous and
a live attenuated vaccine in a herd with endemic post-weaning scours (Farm 1)

Control Injectable Vaccine Oral Vaccine

(Kiiied &ive
Autogenous) Attenuated)
Number of pigs at start of trial 113 130 124
Total average daily gain 315 d= 91 308 d= 79 323 * 107
(ADG)*S .D. (grams/day)
Morbidity @igs treated for 10.5% 7.7%
diarrhea)
Mortality due to E. coli 6.2% a-x 0.76% bvy 8.1%

Mortality due to something 1.8% 0.8%

other than diarrhea
a,b within row differs p c0.02
x,y withïn row diffen pCO.0 17*
SD=standard deviation

'Bonferroni correction
Table 4.2. Results of the scores of diarrhea for three treatments: control, live oral
vaccine and killed vaccine in a herd with endemic scours (Farm 1)

Score of Diarrhea Control (%) Injectable Vaccine Oral Vaccine

(Killed Autogenous) (%) @ive Attenuated) (%)
O 20 (1 8.9) 42 (3 3 -3) 38 (33)
I 57 (53.8) 63 (50)
2 24 (22.6) 15 (1 1.9)

3 5 (4.7) 6 (4.8)
Total of animals 106 126

Scoring system:

O = firm feces

1 = soft feces

2 = fluid feces

3 = projectile diarrhea

2 and 3 were considered clinïcal diarrhea

Figure 4.1. Proportion of animais within the four different scores used to measure
severity of diarrhea, for the three treatment groups (controis, live oral vaccine and
killed intrarnuscular vaccine) in a herd with endemic scours (Farm 1)

Scores of diarrhea
Table 4.3. Results of a kiiled autogenous vaccine used as post-weaning E. coli
diarrhea prevention in a herd with endemic post-weaning diarrhea @'am 1)

First Group Second Group Total

- -

Number of pigs at start the trial

Number and proportion of pigs 5 (4.76) 14 (13.33) 19 (9-04)
treated for diarrhea
Number and proportion of pigs that 5 (4.76) 7 (6.66) 12 (5.71)
died due to diarrhea
Average daily gain (ADG)
(grams/day
Table 4.4. Results of the scores of diarrhea for the injectable killed autogenous
vaccine in a herd with endemic diarrhea (Farm 1)

Score of Diarrhea First Group Second Group Total of Scores

(%) Wo) (%)

Total number of pigs 102

scored
Figure 4.2. Proportion of morbidity and mortality between the groups vaccinated
with an autogenous kiiied vaccine in a M d trial and a case study foilow up

Field Trial Case Study

Figure 4.3. Proportion of animals within the four different scores of diarrhea,
between the groups vaccinated with the autogenous kilied vaccine in a field hial and
a case study foiiow up to field trial

Field Trial mCase Study

4.5. References

Bianchi, A.T., Scholten, J.W., van Zijderveld, A.M., van Zijderveld, F.G. and Bokhout,
B.A.: Parenteral vaccination of mice and piglets with F4+ Escherichia coli suppresses the
enteric anti-F4 response upon oral infection. (1996) Vaccine, 14: 199-206.

Bertshinger, H.U.: Postweaning Escherichia coli diarrhea and edema disease In: Disease of
Swuie gLhed. (1999): 441-454.

Connaughton, I.D., Driesen, S.J., Fahy, V.A. and Samrnons, N.G.: Field tnals with and E.
coli bacterin (Weanavac) to prevent post-weaning colifom enteritis. (1992) In:
International Pig Veterinary, Proceedings. (Eds) Scientific Committee of the 12" IPVS
Congress, Netherlands: 254.

Dewey, C.E., Carman, S., Hazlett, M., van Dreumel, T., and Smart, N.: Endemic
transmissible gastroenterïtis: Diniculty in diagnosis and attempted confïrrnation using a
transmission ûial. Case Report. (1999) Swine Health and Production, 7: 73-78.

Fahy, V.A., Comaughton, I.D., Driesen, S.J. and Spicer, E.M. : Field trials with an oral E.
coli vaccine (Autovac) to prevent post-weaning coliform ententis. (1992) In: International
Pig Veterinary, Proceedings. (Eds) Scientific Committee of the 1 2 IPVS
~ Congress,
Netherlands: 255.

Fahy, V.A., Connaughton, ID., Driesen, S.J. and Spicer, E.M.: Post-weaning
colibacillosis. (1987) In: Australia Pig Science Association Committee (Eds) Manipulating
Pig Production. Australian Pig Science Association, Werribee, Victoria: 187-20 1.

Fairbrother, M.J.: Approaches to the control of Escherichia coli infection. (1999) Western
Canadian Association of Swine Practitioners. Annual Meeting. Proceedings, October 15
& 16, Saskatoon: 65-68.

Francis, D.H. and Willgohs, J.A.: Evaluation of a live avirulent Escherichiu coli vaccine
K88+, LT+ enterotoxigenic colibacillosis in weaned pigs. (199 1) 5 Vet Res, 52: 1051-
1055.

Harnpson, D.J.: Post-weaning Escherichiu coli diarrhea in pigs. In: Escherichia coli in
domestic animals and humans. Ed. Gyles, C.L. Cab International (1994): 17 1- 191.
Josephson, G., Smart, N., McEwen, B. and Gough, J.: K88t E. coli diarrhea in post
weaning pigs. Animal Heaith Laboratory, University of Guelph. In: 18" Annual Centralia
Swine Research Update, January 27 (1999): 3 8-39

Laval, A.: A review of new vaccines for enteric diseases. In: Pigs Misset Special Ed.
Entenc Diseases (May, 1996): 20-2 1.

MacKinnon, J.D.: Enteritis in the young pig caused by Escherichia coli. (1998) The Pig
Journal, 42: 227-255.

Moon, H.W. and Bunn, T.O.: Vaccines for preventing enterotoxigenic Escherichia coli
infections in farm animds. (1993) Vaccine, 11:2 13-220.

Nagy, B. and Fekete, P.Z.: Entertoxigenic Escherichia coli (ETEC) in farm animals.
(1999) Vet Res, 30: 259-284.

Norman, R.G. and Streher, L.D.: Comparing more than two groups. One way ANOVA.
In: Bioestatistics, The Bare Essentials. BC Decker Inc, Hamilton London. 2nded (2000):
71-72.

OIsson, E., Smyth, C.J., Soderlind, O., Svennerholm, A.M. and Molby, R.: Development of
intestinal antibodies against Escherichia coli antigens in piglets with experimental neonatai
E. coli diarrhea. ( 1 986) Vet Microb, 12: 12 9- 133.

Pritchard, G.C.: Transmissible gastroenteritis in endemically infected breeding herds of pigs

in East Anglia, 1981-1985. (1987) Vet Rec., 120: 226-230.

Sarmiento, J.I., Dean, E.A. and Moon, H. W.: Effect of weaning on diarrhea caused by
enterotoxigenic Escherichia coli in three-week-old pigs. (1998) Am J Vet Res, 49: 2030-
2033.

Van den Broeck, W., Cox, E. and Goddeeris, B.M.: Receptor-dependent immune response
in pigs after oral immunization with F4 fknbriae. (1999a) Infection & Immunity, 67: 520-
526.
Van den Broeck, W., Cox, E. and Godeeris, B.M.: Induction of immune responses in pigs
following oral administration of purified F4 h b n a e . (1 999b) Vaccine, 1 7: 2020-2029.

Wood, E.N.: Fashionable and future diseases. (1 99 1) Vet Journal, 27: 193- 197.
Chapter 5: Post-weaning diarrhea and mortality caused by Escherichia coli -
investigation of risk factors and control methods:General Discussion

Traditionally K88 (F4) E. coli organisms have been identified as a cause of diarrhea in

neonatal pigs. More recently in Ontario and elsewhere in North America, post-weaning E.

coli dimhea (PWECD) has become an important problem (van Beers-Schreurs et ai,

1992). A sudden increase in the number of cases of PWECD and mortality associated with

0149K88 (F4) was reported in the fall/winter of 1997 in Ontario, and there is evidence

fkom the Animal Heaith Laboratory at the University of Guelph, that the prevalence of the

disease is increasing. The clinical signs of disease are more severe than previously

described, and most of the intestinal samples yield pure cultures of strongly hemolytic E.

coli, which are resistant to multiple antibiotics (Josephson and Smart, 1998a,b). It has been

suggested these strains are more v i d e n t in individuai animals and more persistent within

the herds than previous isolates (Wood, 1991).

Why this disease has changed to become a more severe problem in post-weaning pigs is

not known. The pathogenesis and lesions associated with the problem have been described

elsewhere, but information regarding risk factors, impact of the disease on productivity

parameters and epidemiology is sparse.

Various control methods have been attempted, inciuding changes in diet, management,

and housing, sometimes with limited results. Some practitioners have also used a number of

vaccination protocols, such as autogenous bacterins injected into suckling piglets,

additional vaccination of sows to boost passive imrnunity or oral immunization of pigs

with live vaccines. Producers have often experienced shoa periods of success in

combatting the disease, only to have the disease reappear a few weeks later.

The objectives of this study were to described the common K88-positive E. coli

serogroups isolated fiom nursery units in Ontario and examine their antimicrobial

resistance patterns, to evaluate productivity on afYected famis, to determine

epidemiological nsk factors associated with PWECD in Ontario, and to determine the

effectiveness of two different vaccines applied before weaning for the control of the

disease.

A total of 50 farms were visited in the summer of 1999 as part of a case-control study.

Twenty-five case and twenty-five control herds were selected fiom multiple sources.

During each fann visit, the owner or f m manager completed a survey. Idormation on

f m hygiene, nursery management, nursery facilities, clinical signs and management of

diarrhea, feed, and Porcine Reproductive and Respiratory Syndrome (PRRS) was

collected. Rectal swabs of ten weaned pigs were taken from each farm, to check the statu

of K88-positive E. coli at the time of the visit. These samples were serotyped using a slide

agglutination test. Cultures positive for O l49:KgI :K88 were tested for antimicrobial

sensitivity.

Blood samples were obtained fiom ten pigs (3-5 weeks of age) on each farm. The

samples were centrifuged and the sera were tested for antibody to Transmissible

Gastroenteritis (TGE) by a coronavinis differential ELISA, and for PRRSV antibody by

the IDE= ELISA test.

 Three control herds developed dimhea problems and were diagnosed with K88-

positive E. coli before the fami visit. These three f m s were reclassified as cases. One case

herd was elùninated due to incomplete survey collection. Another case f m was eliminated

because it was a grower-fhïsher operation. Both of these f m s were dropped fiom the

survey analysis, however, rectal swabs were taken fkom each f m , and these were cultured

and identified, and tested for sensitivity. These two f m s were included in the analysis for

the serotype and antibiotic resistance patterns.

Of the 28 case fanris, 0149:K91:K88 E. coli was cultured from 17 fanns. Three of the

22 control farms were positive for K88 E. coli but no diarrhea problems were apparent at

the time of the vis&and during the study period, or in the history taken during the survey.

Two case fanns were chosen as sites for vaccination trials. The fanns were selected

based on previous history of PWECD and for the famer's willingness to cooperate and

maintain accurate records. This trial was completed to determine the usefulness of a K12-

K88 E. coli and an autogenous intramuscular killed vaccine as methods of controlling this

disease. Average daily gain and proportion of pigs treated for diarrhea were used to assess

the level of morbidity. Mortality due to E. coli diarrhea and diarrhea scores, according to

Sarmiento et a1 (1 988), were aiso recorded.

Although PWECD has a high prevalence, the nurnber of f m s was probably Iimited, for

the large number of variables measured. The distribution of the farms visited in thïs study is

shown in Figure 5.1.

This study demonstrates that post-weaning E. coli diarrhea (PW E C D ) is an

econornically important disease in Ontario nurseries. Mortality increased fiom 2 to 7% on

161
average following an outbreak. However, some farms expenenced mortalities as high as

20-30%, over a one to two month time span. A significant reduction in average daily gain

(ADG) is also observed in f m s with the problem compared with f m s without the

disease (452 grams per day and 414 grams per day, respectively). Similar decreases in

growth rates and mortality are reported by Josephson et a1 (1999). In addition case farms

used signincantly more antibiotics and other treabnents such as acidifiers, probiotics, high

levels of zinc oxide, and injections of antibiotics to individual pigs, which reflects an

increase in labour cost.

The 0l49:Kg 1:K88 ETEC serogroups were the most commonly isolated bacteria in

pure cultures fiom pigs with PWECD in this study. This was expected as this s e r o w e has

been the most cornmonly isolated in Ontano (Josephson and Archambault, 2000). The

most fiequently observed enterotoxin combination was STa,STb, LT. The presence of this

E. coli with 3 toxins has been suggested as the emergence of a new and more virulent

bacterium (Fairbrother, 1999). The combination of just two toxins (STb,LT) was observed

less commonly in case f m s , but were present in three control farms, where clinical

problems were not apparent. The presence of multiple adhesins and toxins were observed

in one case f m which had severe diarrhea (0149:STa,STb,LT7F4and VTF18).

The disease occurred most commonly in the first week &er weaning but it was also

observed to affect pigs after two to three weeks following weaning, or in some cases,

following transfer of pigs to the grower-fïnisher units. This study, demonstrated that the

disease often recurs sporadically, as indicated by the observations that some case farms

were £kee of the problem at the tirne of the visit, and no E. coli was found. Producers have
often experienced short periods of success in combatting the disease only to have the

disease reappear a few weeks later. However, in other fanns, the problem tends to persist

in spite of attention to the mual management, environmental and hygiene factors.

Ln this study, ETEC isolates fiom pigs with post-weaning diarrhea showed a high rate

of resistance to multiple antibiotics, and the use of rndtiple antibiotics via feed, water, and

injection was a comrnon finding in f m s which had PWECD problems. It appeared that

farmers wili move fiom one antibiotic to another antibiotic to try to solve the problem,

raising a concern regarding resistance to multiple antibiotics.

Surprisingly there were no dflerences between case and control farms for a number of

factors commody thought to be associated with PWECD problem, such as off-site

Segregated Early Weaning (SEW), particularly when pigs are mixed fiom multiple f m s

(Josephson et al, 1999). Contrary to common belief, case f m s were more likely to blend

rations to reduce the effect of a sudden feed change compared to control f m s , and case

farrns were more Iikely to limit-feed pigs. These factors Iikely reflect management attempts

to control diarrhea rather than identifying causative factors.

According to the find'mgs of this study, a possible nsk factor for PWECD was the number

of feeder spaces available, which might encourage higher levels of feed consumption and

lead to diarrhea.

Using detergent during the cleaning procedure appears to be a sparing factor. Also

control fanns used more disinfectant and water lines were cleaner than case f m s . These

h d i n g s suggest that possibly cleaning procedures were better on control farms compared

to case f m s , but in general there was little obvious difference between the two groups in

163
sanitation and cleanliness. It is suggested that pigs c a r y the E. coli into the nursery in their

intestinal flora (Hampson et al, 1987) and cleaning procedures rnay not be very relevant.

The use of pellet versus mashed feeds was found to be significant in this study. Case

f m s were more Iikely to use pelleted feed and control f m s were more likely to use mash

feed. Several researchers have suggested that heat treated cereals such as peIleted feed,

might provide a better environment for E. coli proliferation (Lawrence, 1985;

Charbonneau, 1999').

PRRS vaccine status of gestating sows and nursery pigs was also found to be associated

with PWECD. Confounding was found in these two variables. PRRS vaccine is a live

vaccine and has been associated with the birth andior presence of post-weaning weak

viremic pigs (Mengelhg et al, 1998, 1999; Dewey et al, 1999). Also there are some reports

that PRRS rnay exacerbate E. coli problems (Done and Patron, 1995; Nakamine et al,

1998). This needs to be investigated M e r using controlled studies.

Farms represented a wide geographicd distribution and diverse management-housing

types. Sample size of this study may not allow extrapolation of results to all herds in

Ontario, but it certainly suggests that E. coli is not a disease restricted to a single genetic

line, a single feeding company's starter product, a single management style such as early

weaning, or is caused by a single "hot" strain of E. coli. This study proved that PWECD is

a widespread problem without a clear set of risk factors to trigger the disease.

The results of the field trial showed that vaccination either with a killed intramuscdar

product or Iive oral vaccines, did not appear to provide a solution against post-weaning E.
coli diarrhea in the two f m s tested. The number of f m s in this study was very limited

and a farm effect might be involved in the negative results obtained. Further studies with a

larger nuaber of f a m s and anirnals is needed to identw the factors that resulted in the

vaccine failures.

Post-weaning E. coli diarrhea is a very complex problem with many factors involved

and therefore, the solution to controlling this emerging disease is unlikely to be easily

solved-
Figure 5.1. Distribution of 28 case and 22 control farms involved in a study of post-
weaning E .coli diarrhea in Southeni Ontano in 1999.

Cases

Controls
References

Dewey, C.E., Wilson, S., Buck, P. and Leyenaar, J.K.: The reproductive performance of
sows d e r PRRS vaccination depends on stage of gestation. (1999) Prev Vet Med, 40:
233-241.

Done, S.H. and fatron, DL: Porcine reproductive and respiratory syndrome: clinical
disease, pathology and immunology suppression. (1995) Vet Record, 136: 32-35.

Fairbrother, LM.: Identification, nomenclature, and diagnosis of pathogenic Escherichia

c d i . ( 1 999). Western Canadian Association of Swine Fractitioners. Amiral Meeting.
Proceedings, October 15 & 16,Saskatoon: 21-3 1.

Hampson, D L , Fu, A.F. and Robertson, I.D.: Investigation of the source of hemolytic
Escherichia co2i infecting weaned pigs. (1987) Epidem Infect, 99: 149-153.

Josephson, G. and Archambauit, M.: Colibacillosis in pigs in 1999. (2000) Animal Health
Laboratory, Newsletter 4:8

Josephson, G., Smart,N., McEwen, B and Gough, J.: K88+ E. coli diarrhea in post-
weaning pigs. Animal Health Laboratory, University of Guelph. In: 18" Annual Centralia
Swine Research Update, Janii;lrv 27, 1999: 38-39.

Josephson, G. and Smart, N.: K88 strains of E. coli. (1998a) Animal Health Laboratory
Newsletter June, 2: 4

Josephson, G. and Smart, N.: Update on K88 positive E. coli. (1998b) Animal Health
Laboratory Newsletter Sept, 2: 2.

Lawrence, T.L.J.: Processing and preparation of cereals for pigs diets- (1985) In: Recent
Developments in Pig Nutrition. Ed. Cole, D.J.A. and Haresign, W: 230-245.
Mengeling, W.L., Lager, K.M. and Vorwald, AC.: Safety and efficacy of vaccination of
pregnant gilts against porcine reproductive and respiratory syndrome. (1999) Am J Vet
Res, 60: 796-80 1,

Mengeling, W.L., Lager, K.M. and Vorwald, A C : Clinical consequences of exposing

pregnant gilts to strains of porcine reproductive and respiratory syndrome (PRRS) virus
isolated fiom field cases of "atypical" PRRS. (1998) Am Vet Res, 59: 1540-1544.

Nakamine, M., Kono, Y., Abe, S., Hoshino, C., Shh-ai, J., and Ezaki, T.: Dual infection
with enterotoxigenic Escherichia coli and porcine reproductive and respiratory syndrome
virus observed in weaning pigs that died suddenly- (1998) J Vet Med Sci, 60: 555-56 1.

Sarmiento, J.I., Dean, E.A. and Moon, H.W.: Effects of weaning on diarrhea caused by
enterotoxigenic Escherichia coli in three-week-old pigs. (1988) Am J Vet Res, 49: 2030-
2033.

Van Veers-Schreurs, H.M., Vellenga, L., Wensing T. and Breukink, H.J.: The
pathogenesis of the post-weaning syndrome in weaned piglets: a review. (1992) Vet Q, 14:
29-34.

Wood, EN.: FashionabIe and future diseases. (199 1) Vet Journal, 27: 193-197.
 Appendk 1

Producer Survey
 Date

Survey for Post-weaning Diarrhea caused by E- coli

2. Herd phone

3. Herd manager phone

4. Farm Address

5. Lot Township Couniy

6. The Swine Unit is (choose one)

1. Farrow to fmish
2. Farrow to partly finish
3. Farrow to feeder pigs
4,Off-site nursery
5. Other (please describe)

7. Please describe the genetics of your sow and boar herd

Farrowing

8. How many farrowing rooms are on your fann? Rooms

9. How many crates per room are on your farrn? Crates

10. Are fmowing rooms routinely emptied, cleaned and disinfected following weaning?

Always Usually Seldom Never

11. Doyou cross-foster between fmowing rooms? Yes No

12. Do you offer creep feed to nursing pigs, before they are weaned?
Yes No

If yes, when do you start to offer creep feed days

Survey #

Nursery Facility and Management

13. Are the nursery rooms in the same building as the farrowing rooms?
Yes No

14. What is the proportion of floor that is slotted or solid?

SIotted % SoIid YO

15. What is the material of the siotted floor?

16. What is the rnateriat of the solid floor?

17. Do pen walls allow pigs direct contact with pigs in the next pen?
Yes No

18.1s the manure storage for the weaner rooms under the slats?
Yes No

19. Which of the following husbandry practices are used in the nursery? Check al1 that apply.

AIways If Required Never

1, Feeding boards

2. EIectrolytes
3 Antibiotic injection of
individual pigs
4. Sorted by size
-~

5. Sorted by sex
6 . Sick pens
7. Floor mats
8. Zone heating (heat Iamps)
9. Water medication
10. Gruel feeding
Survey #

20. Please describe the facility used for nursery (weaner) pigs.

1. Total inventory
2. Number o f sites
L1st. stage nursery

3. Buildings per site

4. Rooms per building

5. Pens per room

6.AU-in by (select one)
pen, roorn, building, site

1 7. All-out by (select one)

pen, roorn, building, site

1 8. Continuos Flow (YIN)

8. Width of average pen (fi)
9. Length of average pen(ft)
10. Average nurnber of pigs per pen

1 1. Maximum number of pigs per pen

12. FIoor feed (YM)

13. Wet/Dry feeders

14. Nurnber of feeder spaces per pen

15. Number of waterers per pen

16.Check al1 that apply)

Trough watering system .............
Nipple waterers (single).............
Nipple waterers (double) ..............
Bowl drinkers............................
17. Chlorinated water Yesl No (YM)

18. Average daily gain

19. Average mortality
Survey #

2 1, What are the current weaning ages and weight?

Age (days) Weight ( Kg / Ib)
Average
Minimum
Maximum

22, How often do you move pigs into the nursery? (Check one)
1, Once a week
2. Twice a week
3. Three times a week
4. Other (please explain)

23. How many pigs are moved into the nursery at one time? Pigs

24. How many sources are used to fil1 the nursery? (Check one)
1. One sow herd
2. Two sow herds
3. Three or more sow herds

25. When pigs are moved into nursery pens, (check only one)
1. Pigs from only one litter go into each pen
2. Pigs fiom more than one litter go into each pen
3. Pigs fiom more than one farrowing room go into each pen
4. Pigs fkom more than one sow fann go into each pen

26. When pigs are moved into nursery rooms, (check only one)
1. Pigs fiom one farrowing room go into one nursery room
2. Pigs fiom more than one farrowing roorn go into each room
3. Pigs fiom more than one sow faim go into each room

27. On average, what percent of pigs are too Iight to be moved to nursery and have to be culled?
%

28. How long does it take to fil1 a,

Pen days
Room days
Building days
Site days
Survey #

29. How long do pigs stay in the nursery and how heavy are they when they move out?

Length of stay
I Weight (kg/lb)

Average

30. Do you remove slurry fiom the pit bellow the slotted floor before you clean and disinfect the
room?
Always Usually Seldom

3 1. What is the procedure of cleaning the nursery? (Check al1 that apply)
1. Scraping residual feed and rnanure off the pens
2. High pressure hose
3. Use of hot water
4. Use of detergent solution
5. Stearn
6. Disinfect the nursery
7. Wash and disinfect feeders
8. Wash and disinfect drinkers
9. Allowed to dry (12 hrs)
10. Allowed to dry (24 hrs)
11. Other (please expiain)

32. What system of disinfection do you use?

Sprayer Washing pump Foam delivery machine

33. What kind of disinfectant do you use for the nursery?

Amount

34. Do you disinfect water lines between batches of pigs?

Yes No

Disease Status

Porcine Reproductive and Respiratory Syndrome

35. Has your herd been diagnosed with PRRS? Yes No

If no, please skip to question 37

36. When was the most recent diagnosed outbreak of PRRS in your herd?
Month Year
 g latest outbreak, which clinical signs did you see in nursing and nursery pigs
37. D u ~ the
1. Weak new bom pigs
2. Increase pre-weaning mortality
3. Respiratory disease in nursery pigs
4- Diarrhea in nursery pigs
5. lncreased mortaIity in the nursery barn
6- Sudden Death
7. Other please explain

38. Have you vaccinated for PRRS (check al1 that apply)

1. Replacement gilts

2- Open Sows I I
3. Lactating ~ o w s I I
4. Gestating Sows 1 1
6. Nursery pigs 1 1

Post-weaning E. coli Diarrhea

39. Has your herd been diagnosed with K88 E-colz"! Yes No
if no, please skip to question 42
If yes, when was the most recent diagnosed outbreak of K88 E colz?
Month Year

40. How long after the pigs were placed in the nursery did the problems begin?
1. 1-2 weeks
2.2-3 weeks
3 . 3 - 4 weeks
4, >4weeks
4 1. When the original outbreak occurred, which c h i c a l signs did you see?
1, Sudden death with no previous clinical signs
2. Mild diarrhea in weaned pigs
3. Profiise watery diarrhea in weaned pigs
4. Pigs with ccbony"appearance
5. Vomiting
6. Watery diarrhea in suckling pigs
7. Al1 pigs of one pen were affected
8. Mortality %
9. Other please explain

42. What have you tried to treat this problem? (Check al1 that apply)

Treatment Product Amount/Dosage Results

1. Antibiotics in feed
2. Antibiotics in water
3. Injectable antibiotics
4. Vaccines
5. Acidifiers in water
6. Acidifiers in feed
--

8. Probiotics

9. Zinc Oxide 1 1 1
IO. Sanitiser in water I
43. Have you tried any feed changes to solve the problem?
Yes No
If yes, please explain

44. Have you tried any management changes to solve the problem?
Yes No
If yes, please explain
45. Do you have any of the following problems in your nursery? (Check al1 that apply)
1. Respiratory Diseases Currently Never Not Evident now
Coughing
Sneezing
Nasal discharges
Abdominal breathing
Pneumonia
2. Other types of scouring
TGE
Rotavirus
Salmonella
Ileitis(Campy1obacter)
Edema disease
Coccidia
3. Nervous signs
Strept. suis
Edema disease
H. parasuis (Glassers)
4. Other, please explain

Vaccination Protoc01

46. Please describe the entire Nursing pig (pre-weaning) vaccination protocol:
Vaccine Age at fist vaccination Age at booster

47. Please describe the entire weaner (nursery) pig vaccination protocol:
Vaccine Age at first vaccination Age at booster

48. Do you vaccinate sows for E- coli?

Yes No

If yes, which vaccine do you use

-
49. General Nutrition

Size ( a g e h t ) of pigs when begin eating feed

Size (ageht) of pigs when taken offfeed .......

Amount feed fed (kg/lb) ......
Days feed fed (days) ........
1 Amount feed consurnecl/ pig (kg/lb)
Feed contains Spray dried plasma (yesho)
Purchased complete feed .....
Mixed on farm with supplement.....
Mixed on farm with pre-mix .....
- - -
- - --
Cree
~ e e z
Nursery
Feed # I

Estimated nutrient analysis (unit~)

Protein.......
Fat .......
Fibre ......
Mashed feed.......
Cntmbled...........
Pelleted ..............
50, Please k t the medications used in each of the above listed feeds and the amount of each product
that is added or stick, if it is possible, the feed tag
- -

Feed Medications Amount (measurement)

Creep Feed

Nursery
Feed # I
Nursery
Feed # 2
Nursery
Feed # 3

51. Are the pigs Iimit fed or full access? Limit fed Full

If limit fed, how often are they fed? Tirnedday

52. When you change between diets, do you blend the current phase of feed with the next phase?
Yes No

If yes, how long does it take to make the transition between both diets days

Thank you for cornpleting this survey, your assistance is greatly appreciated.
To ensure confidentiality, only summary information will be presented.
 Appendix II

Raw data of the information not inciuded in the body of the thesis, for a

study investigating post-weaning E. coCi diarrhea in Ontario

Table 1: Antibiotic usage on creep feed, between case and control farms

Cases(26) Controk (22) P value

Medicated creep feed

Type of antibiotics used in Carbadox
creep feed
Chfortetracycline

ASP a

Lincornycin
Lincornycin,
Spectinomycin

Tylosin,
Sulfarnettiazine
Overall P value
Farms which used more
than one creep medication
in feed
Antibiotic used Carbadox 66.6 2 O NS

ASP a

Overatl P value NS
Table 2: Antibiotic usage on the first nursery feed, between case and control farms

Cases (26) Controls (19) P-

value
YO n+ Oh n+

Medication of fmt Carbadox

nursery feed used in the
nursery Chlortetracycline

ASP a
Lincomycin 3.8 1 O NS
Lincomycin, Spectinomycin 3 -8 1 15.8 3 NS

Tylosin, Sulfarnethazine O 5.3 1 NS

Flavomycin 3.8 1 O NS

Overall P value NS
Second choice of F m s which used a second 23.1 6 26.3 5 NS
antibiotic used in the choice
first nursery feed in the
nursery Carbadox 7.7 2 O NS
Chlortetratcyline O 5.3 1 NS

Penicillin 3.8 1 15.8 3 NS

Overall P value NS
Third choice of Farms which used a third choice 7.7 2
antibiotic used in the
Fim nursery feed in the NeOmycin 3 -8 1
nursery
TiamuIin 3.8 1
Overall P value

a
Aureomycin, Sulfamethazine, procaine Penicillin
Table 3: Antibiotic usage on the second nursery feed, between case and control farms

Cases (25) Controls (20) P value

YO n+ OA nf
Antibioric used in the Carbadox 24 6 35 7 NS
second nursery feed in
the nursery Chiortetracycline 16 4 15 3 NS

ASP
Lincomycin 8 2 O NS
Lincomycin, Spectinomycin O 10 2 NS
Oxytetracycline 8 2 25 5 NS

Tiamulin O 5 1 NS

Tylosin 4 I O NS
Flavomycin 4 1 O NS
Overall P value NS
Second choice of Farms which used a second 28 7 40 8
antibiotic used in the choice
second nursery feed on
the nursery Carbadox 8 2 O NS

Nemoycin 4 1 O NS
Tiamulin 8 2 1O 2 NS
Penicillin 8 2 25 5 NS
Overall P value NS
 Appendix III

Raw data of the information not included in the body of the thesis, for a

study investigating epiderniological risk factors for post-weaning E. coli

diarrhea in Ontario
 Table 1: Cases and controis using various husbandry techniques for nursery pigs

Cases (26) Controls ( 22) P value

Feeding board never 65.4 17 59.1 13 NS

if required 15.4 4 22.7 5 NS
always 192 5 18.2 4 NS
Electrolytes never 42.3 11 59.1 13 NS
if required 42.3 11 18.2 4 0.07

always 15.4 4 22-7 5 NS

Antibiotic injection of never 3.8 1 4.5 1 NS
individual pigs
if required 69.2 18 72.7 16 NS
always 26.9 7 22.7 5 NS
Sorted by size never 11.5 3 9.1 2 NS
if required 15.4 4 18.2 4 NS

always 73.1. 19 72.7 16 NS

Sorted by sex never 57.7 15 36.3 8 NS

if required 3-8 1 9.1 2 NS

always 38.5 1O 54.5 12 NS
Table 2: Cases and controls using various husbandry techniques for nursery pigs

Cases (26) Controls ( 22) P value

- --

Sick pens never 46.1 12 40.9 9 NS

if requüed 30.8 8 22.7 5 NS

always 23.1 6 36.4 8 NS
Heat floor mats or floors never 61.5 16 54.5 12 NS

if required 19.2 5 22.7 5 NS

always 19.2 5 22.7 5 NS
Zone Heating (heat lamp) - never 46.1 12 54.5 12 NS

if required 23.1 6 13.7 3 NS

always 30.8 8 31.8 7 NS
Water medication never 19.2 5 40.9 9 NS

if required 53.8 14 40.9 9 NS

aiways 26.9 7 18.2 4 NS
Gruel Feed never 80.8 21 63 -6 14 NS

if required 19.2 5 3 1.8 7 NS

always O 4.5 1 NS
Table 3: Management in nursery and nursery faciiity, analysis of numerical data

Cases Controts P value

Total nurnber of anirnaIs

Rooms per building
Pens per room
Time to fil1 a pen
Time to fil1 a room
Number of pigs moved to
nursery at one tune
Total feet sq per pig
Average animals per pen
Maximum animals per pen
Number of feeders per pen
Number of waterers per pen
Average weaning days
Minimum weaning days
Average weaning weight
Minimum weaning weight
Average days in the nursery
Table 4: Management in the nurseiy and nursery facilities. Analysis for categorical
data
Cases(26) Controls (22) P value

Pen
room
site
Al1 out Pen (23)
room (23)
site (23)
Continues flow
Nursery is in the sarne
building
Direct contact

Type of dridcer Watenng system

Nipple single
Nipple double
bowl drinker
Chlorinate the water
Wet feed
Floor feed
Table 5: System of disinfection and manure handling in case and control farms

Cases (26) ControIs (22) P value

System of disinfection none 3.84 1 4.5 1 NS

sprayer 57.7 15 59.1 13 NS

wash pump 26.9 7 27.3 6 NS

foam deIivery 11.5 3 9.1 2 NS

Manure under the slatts 77 20 72.7 16 NS

Remove sluny never 11.5 3 4.5 1 NS

seldom 19.2 5 27.3 6 NS

usually 15.4 4 9.1 2 NS

always 53.8 14 59.1 13 NS

Table 6: PRRS vaccination programs

Cases (25) Controls (22) P value

Nursery pigs vaccinated with PRRSV 40 10 13.6 3 0.04

Cases (23) Controls (20)

Oh n+ YO n+
Gestating sows vaccinated with PRRSV 56.5 13 25 5 0.03

Lactating sow vaccinated with PRRSV 43 -5 10 45 9 NS

Open sows vaccinated with PRRSV 47.8 11 40 8 NS

Suckling pigs vaccinated with PRRSV 8.7 2 15 3 NS

Table 7: Acquisition of the k t and second nursery feed

Cases (26) Controts (22) P value

Acquisition of the first Purchased complete feed 96.1 25 90.9 20

feed in the nursery
Mixed with suppiement in the O 9.1 2
f m
Mixed with pre-mix in the 3.8 1
fami

Overd1 P value

Acquisition of the second Purchased cornplete feed 692 18 77.3 17

feed in the nursery
Mixed with supplement in the 7.7 2 9.1 2
fm
Mixed with pre-mix in the 23.1 6 13.6 3
farm
Overall P value
Table 8: Texture of feed used in first and second nursery feed

Cases (25) Controis (22) P

value
YO n O h n
-

Texture of the fmt feed in the nursery Mashed feed

Cnunbled
Pelleted
Overall P value
Texture of the second feed in the Mashed feed
nursery

Overall P value
Farms which switch texture of feed
from the first feed to the second feed
Table 9: When creep, first and second feed was offered to nursery pigs

Cases Controb P value

Day when creep feed is offered for 25 10.9 4.39 22 12.9 5.9 NS
the first time
Days when Fust feed is offered in 26 20.9 4-17 22 21 4.57 NS
the nursery

Days when second feed is offered 22 27.7 5.32 16 29.1 7.2 NS

in the nursery
 Appendîx I V

SAS analysis output

E- c o l i mortality mode1

The GENMOD Procedure

Mode1 I n f o r m a t i o n

Description Value label

Data Set WORK. TEMP

Distribution BINOMIAL
L i n k Function LOGIT
Dependent V a r i a b l e ECOLIDEA ECOLIDEATH
Observations Used 281
Number Of Events 13
Number O f T r i a l s 281
M i s s i n g Values 86

Class L e v e l I n f o r m a t i o n

Class Levels Values

SOWIDG 28 16 20 27 28 29 32 45 46 64 65
67 68 101 103 108 109 138 139
140 250 272 276 518 577 593
644 656 15404

Parameter I n f o r m a t i o n

f arameter Effect

PRM1 INTERCEPT
PRM2 ORAL
PRM3 IM
PRM4 R
PRM5 P
 C r i t e r i a F o r Assessing Goodness Of F i t

Criterion DF Value Value/DF

Deviance 276 96.0740 0.3481

Scaled Deviance 276 276.0000 1 .O000
Pearson Chi-Square 276 263.1880 O. 9536
Scaled Pearson X 2 276 756.0824 2.7394
Log L i k e l i h o o d - 138.0000

A n a l y s i s O f I n i t i a l Parameter Estimates

Parameter DF Estimate Std E r r Chisquare Pr>Chi

INTERCEPT 1 -1.1901 0.6592 3.2598 0.0710

ORAL 1 -0.2014 0.4573 0.1940 0.6596
IM 1 -1 .9578 O. 6581 8.8508 0.0029
R 1 -1 -5896 0.7660 4.3064 0.0380
P 1 O. 1389 0.0909 2.3336 0.1266
SCALE O 0.5900 0.0000

GE€ Mode1 I n f o r m a t i o n

Description Value

Correlation Structure Exchangeable

Sub j e c t E f f e c t SOWIDG (37 l e v e l s )
Number o f C l u s t e r s 37
C l u s t e r s W i t h M i s s i n g Values 12
C o r r e l a t i o n M a t r i x Dimension 12
Maximum C l u s t e r S i z e 12
Minimum C l u s t e r S i z e O
 A n a l y s i s Of GE€ Parameter E s t i m a t e s
E m p i r i c a l Standard E r r o r E s t i m a t e s

E m p i r i c a l 95% Confidence L i m i t s
Parameter Estimate Std E r r Lower Upper Z Pr>lZI
INTERCEPT - 1 .1693 1 .Il65 -3.3576 1.0190 -1.047 0.2950
ORAL -0.1969 O. 8889 - 1 -9391 1 .5454 - .2215 O. 8247
IM - 1 .9577 1 .1374 -4.1870 0.2716 -1.721 0.0852
R - 1 .6363 1 .3941 -4.3687 1.0960 -1.174 0.2405
P 0.1436 0.1779 -0.2051 0.4922 0.8070 0.4196
Scale 0.5900