BACKGROUND: Antibiotics are essential medicines and their effectiveness is under threat due to antimicrobial resistance. Guidelines are one way to conserve antibiotic effectiveness given that they are intended to modify clinician prescribing. Guidelines that provide antibiotic recommendations should make explicit contextual considerations that influence antimicrobial resistance and their downstream effects on resistance emergence. METHODS: We conducted a systematic review of tuberculosis, gonorrhoea, and respiratory tract infection guidelines and recommendations to examine how and to what extent they are considering contextual factors that influence antimicrobial resistance. We also investigated whether there are guidelines and recommendations that can be adopted or adapted to local contexts. RESULTS: We found that within 74 included guidelines, two thirds of recommendations considered antimicrobial resistance. Of which only five guidelines considered all factors required to consider local aspects such as values, resource use, acceptability, feasibility, and equity. As such, these five guidelines can be either adopted or adapted to Canadian and other contexts. We also found that 39% of guidelines met credibility scores of 60% or greater in AGREE II domains: scope and purpose, rigor of development, and editorial independence. CLINCAL IMPLICATIONS: There are very few Infectious disease guidelines for highly prevalent diseases that do not consider all important contextual factors may influence antimicrobial resistance. Our findings can support societies and organizations, public health policy, and health care stakeholders to develop and implement guidelines that are applicable to local contexts efficiently and resourcefully. Our antimicrobial resistance recommendation framework, used in addition to GRADE Evidence to Decision frameworks, is a start to having this come to fruition. / Thesis / Master of Public Health (MPH)

Master of Science / Department of Diagnostic Medicine/Pathobiology / Sanjeev K. Narayanan / The use of antimicrobials in animal feed has come under increasing scrutiny from the public and regulatory agencies. Direct-fed microbials (DFM) are considered valuable alternatives to antimicrobials in food animal nutrition. DFM are products containing live (viable microorganisms). Studies in Europe have reported antimicrobial resistance (AMR) in organisms used in DFM. This is of serious concern because of the potential for transferring resistance to pathogenic bacteria in the gut. The aim of the present study is to characterize phenotypic and genotypic AMR profiles for 20 different antimicrobials in bacterial strains isolated from 10 commercially available DFM used in. Two antimicrobial susceptibility testing methods, disc diffusion and broth micro-dilution based assay were performed. Enterococcus faecium isolates showed resistance towards metronidazole (n=9/9) with a MIC of > 32 µg/mL, erythromycin (n=5/9) with a MIC of ≥ 8 µg/mL, ciprofloxacin (n=2/9) with a MIC ≥ 4 μg/mL, ceftriaxone (n=6/9) with a MIC ≥ 0.25 μg/mL, rifampin (n=8/9) with a MIC of > 4 µg/mL, trimethoprim/sulfamethoxazole (n=4/9) with a MIC ≥ 1 μg/mL and clindamycin (n=5/9) with a MIC of > 0.5 µg/mL. A Propionibacterium freudenreichii isolate showed resistance towards kanamycin with a MIC of > 64 µg/mL. The same strain also had a MIC of 16 µg/mL for levofloxacin. Two Lactobacillus acidophilus were resistant to vancomycin (n=2/6) with a MIC ≥ 32 μg/mL. All the Lactobacillus species including L. acidophilus (n=6), L. casei (n=4) and L. plantarum (n=2) were resistant to metronidazole, MIC > 32 µg/mL. Two strains of Bacillus subtilis showed resistance to clindamycin, with an MIC of 4 µg/mL and erythromycin with an MIC of > 8 µg/mL, and one strain had no zone of inhibition for metronidazole (MIC > 32 µg/mL). Microarray analysis revealed resistance genes in E. faecium strains of 3 different DFM, including aminoglycoside resistance genes, ant(4’)-Ia, erythromycin resistance genes, ere(A2) and ermB, tetracycline resistance genes, tet39, tet31, tetK and tetC, and beta-lactam resistance gene, pbp5. Conjugation with filter mating showed erythromycin resistance gene transfer, msrC gene, from donor strains to a recipient strain (E. faecium 45-24). These studies show that AMR is prevalent among bacterial strains used as DFM in the cattle industry in the U.S., justifying further characterization, detection and observation of transferable antibiotic resistance between the same genus.

Enteric fever continues to affect people living in endemic settings substantially causing at least 20 million cases of febrile illnesses every year with 1% mortality. Over the last decade there has been considerable debate surrounding the burden and disease profile of enteric fever in the paediatric population. This is partially due to the similarity of the clinical features of paediatric enteric fever to most other febrile illness seen in endemic settings. The treatment of enteric fever is proving to be a challenge with the emergence of antimicrobial resistant strains, particularly the 4.3.1 genotype (H58 haplotype), which is spreading rapidly. Multi-drug resistant (MDR) enteric fever, defined as infection with typhoidal Salmonellae that exhibit a combined resistance to ampicillin, cotrimoxazole and chloramphenicol emerged in the 1990s and was mediated primarily via the 4.3.1 genotype population through the horizontal acquisition of antimicrobial resistance determinants. Subsequently, fluoroquinolones became the drug of choice and the treatment of enteric fever following which fluoroquinolone resistance emerged, again through the 4.3.1 genotype. However, these antimicrobial trends may not be uniform across endemic regions and an understanding of these differing patterns as well the temporal changes in these trends are important in planning treatment strategies. In the short and medium term work needs to be focused on achieving the greatest benefits from the prudent use of the recently WHO pre-qualified Vi-TT conjugate vaccine candidate. Whilst the long term vision towards eradicating enteric fever needs to focus on better understanding the underlying the biology of this disease through the use of contemporary technologies while simultaneously improving infrastructure for the provision of clean water, adequate sanitation and hygiene. This thesis aims to age-characterise the disease burden of typhoid fever in endemic regions of South and South-East Asia as well as the African continent. Following this, the molecular epidemiology of enteric fever in two endemic settings in the Indian subcontinent is delineated with a keen focus on the 4.3.1 genotype (H58) population as well the phenotypic patterns and molecular determinants of antimicrobial resistance. This thesis finally systematically reviews the global trends of antimicrobial resistance of S. Typhi isolates over time both from a phenotypic and molecular perspective. The key results from this thesis include; the age stratification of disease occurrence in endemic regions which showed a substantial proportion occurs in the youngest age group in both Africa and Asia, the uniform dominance of 4.3.1 genotypes conferring a high degree of fluoroquinolone resistance contrary to earlier suggestions of younger children being more susceptible to a broader range of infecting genotypes, the dissimilarities between the antimicrobial resistance carrying capabilities of lineage I and lineage II strains of the 4.3.1 genotype as well as novel AMR gene arrangements and finally the temporal trends of AMR in S. Typhi which were different between Asia an Africa. The high prevalence of lineage I strains in Africa and South-East Asia in contrast to the high prevalence of lineage II strains in the Indian subcontinent reflect the antimicrobial selection pressures as well the evolutionary characteristics of circulating pathogen populations in these regions. The implications of the data reported in this thesis have implications for treatment and prevention strategies. For the first time in history an opportunity has risen to effectively vaccinate the youngest age group (0-4 years) from typhoid through the Vi-TT conjugate vaccine. As highlighted in this thesis the youngest age group (0-4 years) have a high disease occurrence in endemic areas as seen in a meta-analysis as well as through data from two endemic sites collated and reported in this thesis. The older age groups also suffer greatly from this disease calling for a broad based vaccine strategy. The implications for treatment of enteric fever are however more relevant in the immediate term which suggest that in endemic regions in Asia, fluoroquinolones have little role to play in treatment protocols while fluoroquinolones are still relevant in the African setting. In Asia, reverting back to former first-line antimicrobials might be an option but the possibility of re-emergence of widespread resistance to these currently sensitive antimicrobials is very high exemplifying the ability of S. Typhi to adapt to changing antimicrobial pressures.

Fifty three Staphylococcus aureus isolates were obtained from three centres, two hospitals and a private pathology laboratory, and studied for susceptibility to bacteriophages, resistance to antimicrobial agents and plasmid contents.Results of bacteriophage typing revealed that they belonged to a variety of phage types. Eighteen were untypable by any of the International Set of Phages, 16 belonged to phage group 111, nine to group I, four to group 11, two to group IV and two to the miscellaneous group.The isolates were resistant to one or more of methicillin (Mc), benzyl penicillin (Pc), gentamicin (Gm) , kanamycin (Km) , neomycin (Nm) , streptomycin (Sm) , trimethoprim (Tp), sulphonamides (Su), tetracycline (Tc), minocycline (Mn), chloramphenicol (Cm), novobiocin (Nb) and fusidic acid (Fa). Resistance to Pc was due to the production of beta-lactamase (Bla). No resistance to vancomycin, spectinomycin and erythromycin was detected. Resistance was also found to heavy metals such as cadmium (Cd), mercury (Hg), phenyl mercuric acetate (Pma), arsenate (Asa) and to the nucleic-acid binding compounds propamidine isethionate (Pi) and ethidium bromide (Eb).All but one of the isolates harboured plasmids. The number of plasmids the isolates carried varied from one to six and their sizes ranged from < 1.0 kb to c.48 kb.Location of the resistance determinants was ascertained by curing and transfer experiments. Loss of resistance was tested after growth at 43.5°C and transfer of resistance determinants was attempted by transduction, mixed-culture transfer and conjugation. The results revealed that resistance to Mc, Gm, Tp, Mn and Fa was chromosomal in all the resistant isolates and in some isolates Bla and resistance to Sm and Cd were chromosomal as well as plasmid encoded. In the majority of cases Bla and resistance to Km, Nm, Sm, Tc, Cin, Cd, Hg, Asa, Pma, Pi and Eb was encoded ++ / by plasmids.Conjugation experiments led to the isolation of three unique conjugative plasmids which have not been found to confer resistance to antimicrobials or to produce haemolysins or diffusible pigment (Dip). The three plasmids, pWBG620, pWBG637 and pWBG661, were indistinguishable by restriction endonuclease analysis and DNA-DNA hybridisation. However pWBG620, unlike pWBG637 and pWBG661, was not detected in the cytoplasm of its host and was only detected in transconjugants after it mobilised a non-conjugative Sm-resistance (SmR) plasmid. Further analysis indicated that it is integrated into the chromosome of its host, excises during conjugation and mobilises the SmR plasmid.These plasmids were studied further using pWBG637 as a representative. It was compared with representatives of the two groups of conjugative plasmids which have been reported in the staphylococci. These are the plasmids which encode resistance to Gm, Km and Nm and those which code for the production of diffusible pigment. The three types of conjugative plasmids were compared by restriction endonuclease analysis and DNA- DNA hybridisation and were found to be different. A preliminary restriction map of pWBG637 has also been constructed.However since pWBG637 has no resistance phenotype direct selection for it was not possible in transfer experiments and for incompatibility (Inc.). To study it further it was necessary to construct resistant derivatives which could be selected for in transfer experiments. This was achieved by labelling pWBG637 with resistance transposons to generate two conjugative plasmids, pWBG636 carrying an insert of Tn3851 (Gm- resistance) and pWBG642 carrying an insert of Tn551 (hn- resistance). It was found that transposon labelling had not changed the incompatibility of pWBG637 and therefore pWBG636 and pWBG642 were used in further experiments in place of pWBG637. Inc. ++ / tests with the pWBG637 derivatives revealed that the pWBG637 type of plasmid is not only different from the other two types of conjugative plasmids but is different from any of the described staphylococcal Inc. groups and therefore the pWBG637 type of plasmids represent a new Inc. group 15. The pWBG637 type of plasmids were studied further using plasmids pWBG636 and pWBG642. They were able to transfer conjugatively to a capsulated S.aureus strain either by the polyethylene glycol method or on filter membranes. They also transferred by conjugation to S. epidermidis and Streptococcus faecalis and were able to transfer back from these strains to S.aureus indicating that they also replicate in these hosts. Consequently they have been used to mobilise non-conjugative plasmids from S.epidermidis and non phage typable S.aureus. Both plasmids failed to transfer conjugatively to Bacillus subtilis and Escherichia coli.pWBG637 transferred non-conjugative plasmids by mobilising them in a manner similar to mobilisation (donation) in E.coli or by recombining with them to form new resistance plasmids. In one case, pWBG628 which encodes Bla and resistance to Cd, Km, Nm and Sm and has no homology with pWBG637 recombined with it during conjugation to produce three new conjugative plasmids pWBG629, pWBG630 and pWBG631 carrying resistance determinants from pWBG628. One of these plasmids, pWBG629, was found to be pWBG637 which had acquired a 4.5 kb element encoding resistance to Km, Nm and Sm. This element was shown to be transposable in both rec+ and rec- backgrounds and has been designated Tn3854. It expressed Sm resistance in E.coli and differs on this account from the Gram-negative transposon Tn5 which expresses resistance to Km, Nm and Sm in non-enteric bacteria but only resistance to Km and Nm in E. coli.Where possible the non-conjugative plasmids encoding resistance to ++ / antimicrobial agents were compared with phenotypically similar plasmids isolated from other parts of the world. It was found that the Tc and Sm resistance plasmids were closely related to other plasmids with the same phenotype whereas the Cm resistance plasmids were different.Although the majority of the Bla plasmids belonged to Inc. group 1 they demonstrated significant restriction enzyme fragment length polymorphism when compared with other Bla plasmids.This study has provided the first data on the genetics of antimicrobial resistance in Nigerian S.aureus. Although many of the plasmids studied were found to be similar to those previously described the isolates also contained some unique and previously undescribed plasmids.

Prevalence of antimicrobial resistance in fecal Escherichia coli isolates from feedlot cattle was characterized. Tetracycline resistance in isolates from newly weaned, auction market derived calves on arrival at the feedlot in a clinical trial was 9.8% and resistance to three or more antimicrobials was 2.1% compared to 17.6 % and 5.9% in a cohort study. The prevalence of tetracycline resistance at 78.3% and resistance to three or more antimicrobials at 52.5 % in isolates from spring calves submitted to a regional diagnostic laboratory were higher than those found on arrival at the feedlot. Of isolates from composite feedyard pen samples late in the feeding period, 39.4% were tetracycline resistant and 7.6% were resistant to three or more antimicrobials, somewhat higher than on arrival. Use of oxytetracycline in the feed for disease prophylaxis and the metaphylactic use of long-acting injectable oxytetracycline were associated with increased proportions of cattle with one or more resistant E. coli isolates early in the feeding period, while the use of individual animal treatments was not. The proportion of animals with one or more tetracycline resistant E. coli isolates was not different between the control, metaphylactic treatment and prophylactic treatment groups preslaughter; however, there were significantly more resistant animals in all groups preslaughter than at arrival. There were also no associations found between the total volume of parenteral antimicrobials used for disease treatment in individual animals and antimicrobial resistance in the cohort study. In addition, no strong associations were found between pen-level prevalence of antimicrobial resistance antimicrobial use or other variables. There was no significant difference between the proportion of isolates per pen resistant to tetracycline, one or more, two or more antimicrobials, or three or more antimicrobials when using 20, 15, 10 or 5 isolates from composite pen-level fecal samples. Variance for isolates resistant to three or more antimicrobials was partitioned as 12.7% at the feedyard-level and 28.7% at the pen-level. The use of diagnostic laboratory data for AMR surveillance was also discussed, and alternatives to antimicrobial treatment in the feedlot were also investigated. Overall a significant contribution to our understanding of antimicrobial resistance in feedlot cattle was achieved.

This thesis summarizes an investigation of antimicrobial resistance (AMR) and antimicrobial use (AMU) in cow-calf herds. The specific objectives of this project were to describe common reasons for treatment and the types of antimicrobials used in cow-calf herds, to describe the frequency of AMR in generic fecal Escherichia coli isolated from various age groups commonly found on cow-calf farms, to determine risk factors associated with the occurrence of AMR, and finally to investigate the underlying molecular mechanisms of AMR in cow-calf herds. At least 86% of the herds treated one or more calves or cows during the study period; however, the overall proportion of both calves and cows reported as treated was less than 14% for calves and 3% for cows. The majority of antimicrobials reported as used in cow-calf operations were for individual therapeutic use rather than prophylaxis, metaphylaxis, or growth promotion. Injectable formulations were the most commonly reported method of antimicrobial administration on cow-calf farms. Cow-calf herds in Wetern Canada are not a significant reservoir for resistance to antimicrobials classified as very important to human medicine such as ciprofloxacin and ceftiofur. The three most common resistances detected were to tetracycline, sulphamethoxazole, and streptomycin regardless of age group. Young calves sampled in the spring of the year were more likely to be shedding AMR E. coli than older calves sampled in the fall of the year or than cows sampled in the spring of the year. The cow-calf pair relationship was not an important factor in transfer of AMR from the individual cow to her calf, but the presence of AMR in the general cow herd was associated with AMR in the calf population. The potential importance of co-selection for AMR at the molecular level was demonstrated by both the risk factor analysis and the molecular work. Phenotypic resistance to streptomycin, tetracycline, and sulphamethoxazole were each associated with the presence of resistance genes from all six families of antimicrobials examined in this study. Several statistically significant associations were also detected between the resistance genes considered. No significant associations were detected between any of the AMR phenotypes or genotypes and the STEC virulence factors stx1, stx2 and eae.

Prevalence of antimicrobial resistance in fecal Escherichia coli isolates from feedlot cattle was characterized. Tetracycline resistance in isolates from newly weaned, auction market derived calves on arrival at the feedlot in a clinical trial was 9.8% and resistance to three or more antimicrobials was 2.1% compared to 17.6 % and 5.9% in a cohort study. The prevalence of tetracycline resistance at 78.3% and resistance to three or more antimicrobials at 52.5 % in isolates from spring calves submitted to a regional diagnostic laboratory were higher than those found on arrival at the feedlot. Of isolates from composite feedyard pen samples late in the feeding period, 39.4% were tetracycline resistant and 7.6% were resistant to three or more antimicrobials, somewhat higher than on arrival. Use of oxytetracycline in the feed for disease prophylaxis and the metaphylactic use of long-acting injectable oxytetracycline were associated with increased proportions of cattle with one or more resistant E. coli isolates early in the feeding period, while the use of individual animal treatments was not. The proportion of animals with one or more tetracycline resistant E. coli isolates was not different between the control, metaphylactic treatment and prophylactic treatment groups preslaughter; however, there were significantly more resistant animals in all groups preslaughter than at arrival. There were also no associations found between the total volume of parenteral antimicrobials used for disease treatment in individual animals and antimicrobial resistance in the cohort study. In addition, no strong associations were found between pen-level prevalence of antimicrobial resistance antimicrobial use or other variables. There was no significant difference between the proportion of isolates per pen resistant to tetracycline, one or more, two or more antimicrobials, or three or more antimicrobials when using 20, 15, 10 or 5 isolates from composite pen-level fecal samples. Variance for isolates resistant to three or more antimicrobials was partitioned as 12.7% at the feedyard-level and 28.7% at the pen-level. The use of diagnostic laboratory data for AMR surveillance was also discussed, and alternatives to antimicrobial treatment in the feedlot were also investigated. Overall a significant contribution to our understanding of antimicrobial resistance in feedlot cattle was achieved.

This thesis summarizes an investigation of antimicrobial resistance (AMR) and antimicrobial use (AMU) in cow-calf herds. The specific objectives of this project were to describe common reasons for treatment and the types of antimicrobials used in cow-calf herds, to describe the frequency of AMR in generic fecal Escherichia coli isolated from various age groups commonly found on cow-calf farms, to determine risk factors associated with the occurrence of AMR, and finally to investigate the underlying molecular mechanisms of AMR in cow-calf herds. At least 86% of the herds treated one or more calves or cows during the study period; however, the overall proportion of both calves and cows reported as treated was less than 14% for calves and 3% for cows. The majority of antimicrobials reported as used in cow-calf operations were for individual therapeutic use rather than prophylaxis, metaphylaxis, or growth promotion. Injectable formulations were the most commonly reported method of antimicrobial administration on cow-calf farms. Cow-calf herds in Wetern Canada are not a significant reservoir for resistance to antimicrobials classified as very important to human medicine such as ciprofloxacin and ceftiofur. The three most common resistances detected were to tetracycline, sulphamethoxazole, and streptomycin regardless of age group. Young calves sampled in the spring of the year were more likely to be shedding AMR E. coli than older calves sampled in the fall of the year or than cows sampled in the spring of the year. The cow-calf pair relationship was not an important factor in transfer of AMR from the individual cow to her calf, but the presence of AMR in the general cow herd was associated with AMR in the calf population. The potential importance of co-selection for AMR at the molecular level was demonstrated by both the risk factor analysis and the molecular work. Phenotypic resistance to streptomycin, tetracycline, and sulphamethoxazole were each associated with the presence of resistance genes from all six families of antimicrobials examined in this study. Several statistically significant associations were also detected between the resistance genes considered. No significant associations were detected between any of the AMR phenotypes or genotypes and the STEC virulence factors stx1, stx2 and eae.

Campylobacter jejuni is a commensal microorganism of the poultry gastrointestinal tract. Broilers, layers, ducks, turkeys, and quails can be colonized by Campylobacter without illness occurring. The vast majority of human Campylobacter infections are recognized as being foodborne. For 2008, preliminary FoodNet data showed that the Campylobacter incidence of infection, 12.68 per 100,000 of the U.S. population, is the second highest, only behind Salmonella at 16.20 per 100,000. To further understand Campylobacter’s role as a foodborne pathogen, analysis at the molecular level is needed. Microbial molecular typing allows for identification and differentiation of bacterial strains beneath the species level. In this study, the “gold standard” method for molecular subtyping, Pulsed Field Gel Electrophoresis (PFGE), along with Diversilab® repetitive element Polymerase Chain Reaction (rep-PCR) and 16S-23S Internal Spacer Region Denaturing Gradient Gel Electrophoresis (ISR DGGE) were used for the molecular typing of Campylobacter jejuni isolates obtained during different stages of commercial broiler production and processing. In addition, the C. jejuni isolates were tested for resistance to antimicrobials commonly used in both veterinary and human medicine. Antimicrobial resistance testing was carried out using a broth dilution system. The majority of recovered isolates came from post-harvest carcass rinsates. Carcass rinses were obtained at post-evisceration, post-chill stages. All isolates (n = 46) were identified by the Polymerase Chain Reaction as Campylobacter jejuni. Three genotypes (n = 44, n = 1, n = 1) were identified by PFGE. The 46 rep-PCR products grouped into seven clusters and two outliers. Clustering of rep-PCR products by sample source was not observed. No relatedness trends were observed for isolates recovered from the same source. The combination of PFGE and Diversilab rep-PCR methods provides highly discriminatory molecular typing results. These results provide practical epidemiological information that shows postevisceration and post-chill stages are still important targets for intervention studies. The very high occurrence of C. jejuni isolates exhibiting genotype A suggests it may differentially express certain gene(s) that enable this strain to more favorably survive under the different harsh environmental conditions encountered during production and processing. In addition, phenotypic testing revealed all of the isolates were not resistant to the antimicrobials azithromycin, ciprofloxacin, erythromycin, gentamycin, tetracycline, florfenicol, nalidixic acid, telithromycin, and clindamycin at any of the concentrations tested. All the C. jejuni isolates exhibited an indistinguishable two-band 16S-23S ISR DGGE profile. Overall, these C. jejuni commercial broiler pre- and post-harvest isolates exhibited an extremely low degree of molecular and phenotypic variability.

Surgical site infections (SSIs) are an emerging cause of increased morbidity, mortality, and treatment cost, in veterinary medicine. Medical records were searched to evaluate for associations that could increase the risk of developing SSIs. Logistic regression was used to analyze the risk factors statistically, to determine their influence on SSI risk. An SSI incidence rate of 3.0% was found in this study for all small animal surgical procedures performed from September 2010 to July 2011, with implants, hypotension and surgical classification associated with increased likelihood of SSI. Active surveillance is crucial for the development of methods to prevent SSI’s. Biofilms contribute to the antimicrobial resistance properties commonly found in bacteria such as methicillin-resistant Staphylococcus pseudintermedius, which is found in canines. An enzyme known as DispersinB was studied to assess its effect on biofilm formation and degradation. DispersinB prevented the formation and eradicated biofilm in vitro. In vivo testing is required to further assess the effects of DispersinB. / Ontario Veterinary College Pet Trust, Canadian Institutes of Health Research, Kane Biotchech