Antimicrobial resistance in bacteria in animals is a health concern globally, with the role of companion animals, particularly dogs and cats, in the spread of antimicrobial-resistant bacteria to humans being poorly understood.1–3 Results of studies4–7 have begun to shed light on the occurrence of antimicrobial-resistant bacteria in pets; however, there is a lack of studies investigating pet-related risk factors for carriage of antimicrobial-resistant bacteria. Antimicrobial resistance in zoonotic pathogens, particularly Salmonella spp and Campylobacter spp, in companion animals is of great importance because pets can be important risk factors for infections in humans caused by these pathogens.8–11 Their close relationship and frequent contact with humans may make these animals a potential source of antimicrobial-resistant bacteria and represent an important public health issue.1,12,13 Multidrug-resistant Escherichia coli and Enterococcus faecalis have been transmitted from a dog to its owner through a penetrating bite wound.14 In addition, E coli with similar antimicrobial resistance patterns and genes were found in dogs and cats sharing a household and in humans and dogs sharing another household.15
Dog ownership is common in Canada. It has been estimated that there are > 6 million pet dogs living in approximately 32% of Canadian households,16 and a survey in the region of Waterloo, ON, Canada, found that 43% of survey participants owned a dog.17 In North America, dogs are often considered family members rather than merely pets, which can be attributed to an increasingly strong human-animal bond. Because of this close relationship, concerns have been raised about the potential transmission of zoonotic antimicrobial-resistant bacteria between dogs and humans. Studies18–21 have identified previous treatment with antimicrobials as a risk factor for the carriage of antimicrobial-resistant Salmonella spp and E coli in pet dogs. Hospitalization of dogs has also been reported as a significant risk factor for the shedding of antimicrobial-resistant E coli.4 Furthermore, a potential association between dog age and tetracycline resistance in E coli has been reported.18
There appear to be a limited number of published reports of studies conducted to evaluate pet-related management factors and their association with carriage of antimicrobial-resistant Salmonella spp and generic E coli in pet dogs in Canada. The purpose of the study reported here was to determine associations of pet-related management factors, including type of diet fed (eg, raw, cooked, or processed), exposure or access to other pets and livestock, veterinary treatments, and pet-related demographics (age, breed, and sex) with the carriage of antimicrobial-resistant Salmonella spp and E coli in a population of pet dogs from volunteer households in Ontario, Canada.
Materials and Methods
Sample
From October 2005 through May 2006, a convenience sample of households in Ontario was recruited to participate in a study22 that investigated the presence of Salmonella spp, Clostridium difficile, and E coli for the purpose of evaluating antimicrobial resistance in household dogs. In brief, 138 dogs from 84 households were recruited through advertisements in brochures, emails, and email list services that targeted the University of Guelph, the Ontario Veterinary College, pet therapy organizations in southwestern Ontario, local veterinary conferences and meetings, and veterinary clinics in Ontario that agreed to display brochures related to the project. The data used in the study reported here were a subset of a larger environmental household study,22 and no sample size calculation was performed for the outcomes that were subsequently investigated. The study was approved by the University of Guelph Research Ethics Board, and owner consent was obtained for all participating dogs.
Sample collection and questionnaire
Study homes were visited by trained technicians who collected environmental samples, administered the questionnaire, and provided a fecal collection kit to the primary caregiver of each household dog (the caregiver was to collect and return 1 fecal sample/d for 5 consecutive days).22 The questionnairea was used to collect information on each dog's primary diet and whether additional animal products were added to the diet; the presence and type of other pets in the household; people living in the household; the dog's activities; veterinary care, including deworming; any contact with livestock; and the use of any probiotics during the preceding month.
Isolation of Salmonella spp
All fecal samples were shipped in prepared kits via express post by the dog owners on a daily basis and were received at the University of Guelph Ontario Veterinary College Department of Population Medicine. Samples were not submitted in transport media or in temperature-controlled containers. A complete description of the isolation of Salmonella spp from the fecal samples can be found elsewhere.22 In brief, 10 g of fresh feces were combined with saline (0.85% NaCl) solution, homogenized, and then enriched with buffered peptone water. Two parallel isolation methods were used. In the first, the buffered peptone water mixture was inoculated into modified semisolid Rappaport-Vassiliadis agar and subsequently subcultured on MacConkey agar; presumptive nonlactose fermenting colonies were then inoculated on full tryptic soy agar. In the second method, the buffered peptone water solution was inoculated into Rappaport-Vassiliadis broth, which was then added to tetrathionate broth and inoculated onto each of full xylose lysine tergitol 4, brilliant green sulfa, and bismuth sulfate agars. Two typical colonies from each agar plate were subcultured onto MacConkey agar, and nonlactose fermenting colonies then were plated onto tryptic soy agar. Biochemical testing for both methods was conducted with Christensen's urea, triple sugar iron, and Salmonella O antiserum Poly A-I & Vi agglutination tests. A sample was considered Salmonella positive if it had positive test results for either isolation method. Up to 3 Salmonella-positive isolates/dog/d were submitted to a laboratoryb for serotyping and phagetyping and to another laboratoryc for antimicrobial susceptibility testing.
Isolation of E coli
All fecal samples from the participating dogs were submitted for E coli culture at the Canadian Research Institute for Food Safety laboratory, University of Guelph. A complete description of the isolation of E coli from the fecal samples can be found elsewhere.23 In brief, a loopful of fecal slurry from the Salmonella spp isolation procedure was inoculated onto MacConkey agar; presumptive E coli colonies were transferred to a secondary purification stage of MacConkey agar and then transferred onto tryptic soy agar. Confirmation testing for E coli was conducted with Kovac indole spot reagent and Simmon citrate agar. Three E coli isolates/dog were collected from the first fecal sample that had positive results for E coli and forwarded to the laboratoryc for antimicrobial susceptibility testing.
Antimicrobial susceptibility testing
Antimicrobial susceptibility testing for the Salmonella and E coli isolates was conducted with an automated broth microdilution system.d The National Antimicrobial Resistance Monitoring System antimicrobial susceptibility panel CMV1AGNF was used for E coli and Salmonella spp in accordance with methods described by the Canadian Integrated Program for Antimicrobial Resistance Surveillance.24,25 Antimicrobials and their MIC breakpoints used for susceptibility testing of E coli and Salmonella spp were summarized25–28 (Appendix 1). Breakpoints for resistance were those used by the Canadian Integrated Program for Antimicrobial Resistance Surveillance and National Antimicrobial Resistance Monitoring System, which were derived from information from the Clinical Laboratory Standards Institute, when available.24,25 A lower breakpoint for ceftriaxone26 was used in the present study. Antimicrobials were also classified on the basis of their importance to human medicine because many of the chemical classes of antimicrobial drugs used in domestic animals are also used in humans and some of these antimicrobials are essential for treatment of serious life-threatening infections in humans.27
Statistical analysis
Data were entered in and analyzed with statistical software.e All tests were 2 tailed, with significance set at α = 0.05. Isolates were considered susceptible or resistant on the basis of the MIC breakpoints (Appendix 1). Isolates with intermediate susceptibility were considered susceptible.
Multilevel logistic regression models were used to screen each variable in the questionnaire 1 at a time for association with the following outcomes: resistance to any antimicrobial, multiclass resistance (ie, resistance to antimicrobials of ≥ 2 classes), and specific antimicrobial resistances. Pet-related variables were examined in univariable models (Appendix 2). Multilevel models were created with a modulef for fitting multilevel models in a software program.g All models contained random intercepts for dog and household to account for clustering because multiple isolates were tested for each dog and multiple dogs were tested for each household. Random intercept models were created by means of reweighted iterative generalized least squares with predictive quasi-likelihood or marginal quasi-likelihood and the first-order derivative of the Taylor series expansion for linearization.29,30 Only antimicrobials and resistance categories for which antimicrobial-resistant isolates of E coli or Salmonella spp were found in > 25 dogs were examined by means of multilevel logistic regression. Variables with values of P ≤ 0.20 were examined for collinearity on the basis of Pearson φ correlation coefficients before further multivariable models were built.
If model convergence could be achieved, models with 2 explanatory variables (combining potential confounders with significant variables from the univariable models in a pairwise manner) were used to test the impact of potential biological confounders. Potential confounders included breed size (small, medium, large, or mixed); age category (young [< 1 year old], adult [1 to 7 years old], or senior [> 7 years old]); sex (male or female); and neuter status (sexually intact or neutered). A variable was considered a confounder if it was a nonintervening variable and its removal from the model resulted in a change of > 20% in the coefficient of another variable.29,31 Normality of best linear unbiased predictors was assessed with normal quantile plots to determine model fit for all multilevel models.29 Because of the limited effective sample size of the dataset and model instability, full multivariable models were not reported. Variance estimates from each model were used to estimate the percentage variance at each level (ie, isolate, dog, and household) in the model, in accordance with the latent variable technique.29
Results
Isolates and antimicrobial susceptibility testing
In total, 138 dogs from 84 households were recruited for the study and provided 515 bacterial isolates for antimicrobial susceptibility testing. Of the 138 dogs enrolled in the study, 32 (23.2%; 95% CI, 16.4% to 31.1%) had at least 1 fecal sample that had positive results for Salmonella spp, and they provided 120 Salmonella isolates. Escherichia coli was recovered from 133 (96.4%; 95% CI, 91.8% to 98.8%) dogs, and they provided 395 E coli isolates for antimicrobial susceptibility testing. Of the 120 Salmonella isolates, 24 (20.0%; 95% CI, 13.3% to 28.3%) were resistant to at least 1 antimicrobial, and 17 (14.2%; 95% CI, 8.5% to 21.7%) were multiclass resistant. Of the 395 E coli isolates, 77 (19.5%; 95% CI, 15.7% to 23.8%) were resistant to at least 1 antimicrobial, and 41 (10.4%; 95% CI, 7.6% to 13.8%) were multiclass resistant. Of the 515 isolates tested for antimicrobial susceptibility, resistance to ampicillin was the most common resistance, with 70 (13.6%; 95% CI, 10.8% to 16.9%) isolates resistant to that antimicrobial.
Univariable analyses
In total, 48 pet-related variables, including type of diet fed (eg, raw, cooked, or processed), exposure or access to other pets and livestock, veterinary treatments, and pet-related demographics (Appendix 2), were examined in univariable multilevel logistic models for associations with isolates being considered resistant, multiclass resistant, and ampicillin resistant on the basis of MIC breakpoints. Breed, age, sex, and neuter status were collected for 129 of 138 (93.5%) dogs. Dog identification number and household identification number were used as random intercepts in all multilevel logistic regression models. Variables with values of P ≤ 0.20 and in which model convergence could be achieved were retained in the univariable multilevel logistic regression models for isolates considered resistant, multiclass resistant, and ampicillin resistant on the basis of MIC breakpoints (Tables 1–3). Multiple comparison corrections were not conducted because of the preliminary nature of the study and the intention to identify all potentially important risk factors.
Univariable associations between pet-related factors and the risk of antimicrobial resistance for 515 Salmonella spp or Escherichia coli isolates, 136 dogs, and 83 households* for pet dogs from volunteer households in Ontario from October 2005 through May 2006.
| Variation | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Household | Dog | ||||||||
| Variable | OR | 95% CI† | P value‡ | Value (SE)† | % | Value (SE)† | % | Exposed§ | Unexposed§ |
| Bacterial species isolated (E coli vs Salmonella spp) | 2.34 | 1.06–5.19 | 0.036 | 2.39 (0.94) | 35.05 | 1.15 (0.70) | 16.77 | 395/133/82 | 120/32/21 |
| Fed homemade diet or homemade food product added to diet | 3.50 | 1.18–10.37 | 0.024 | 2.13 (0.92) | 31.95 | 1.26 (0.72) | 18.80 | 204/39/19 | 311/97/64 |
| Fed raw diet or raw food product added to diet | 3.28 | 1.00–10.74 | 0.049 | 2.22 (0.93) | 32.90 | 1.24 (0.72) | 18.35 | 160/28/13 | 355/108/70 |
| Fed raw beef, raw chicken, raw pork, or raw eggs in past week | 3.11 | 0.93–10.39 | 0.065 | 2.22 (0.93) | 32.92 | 1.23 (0.72) | 18.28 | 157/27/12 | 358/109/71 |
| Fed homemade raw diet | 4.43 | 1.24–15.91 | 0.022 | 2.16 (0.92) | 32.27 | 1.24 (0.72) | 18.57 | 140/22/10 | 375/114/73 |
| Fed raw chicken in past week | 5.17 | 1.38–19.32 | 0.015 | 2.01 (0.91) | 30.43 | 1.32 (0.74) | 19.88 | 120/18/9 | 395/118/74 |
| Diarrhea in past month | 0.38 | 0.10–1.50 | 0.169 | 2.25 (0.92) | 33.46 | 1.19 (0.71) | 17.69 | 98/28/18 | 417/108/65 |
| Owner worked in health-care or veterinary-related field | 0.33 | 0.10–1.07 | 0.065 | 2.06 (0.90) | 31.20 | 1.25 (0.72) | 18.94 | 146/41/27 | 369/95/56 |
| Annual vaccination | 0.30 | 0.10–0.90 | 0.031 | 2.19 (0.92) | 32.69 | 1.22 (0.71) | 18.26 | 331/99/65 | 184/37/18 |
| Received heartworm preventive in past 6 months | 0.31 | 0.10–0.93 | 0.036 | 2.21 (0.93) | 32.82 | 1.24 (0.72) | 18.38 | 365/99/62 | 150/37/21 |
| Fed commercial dry or canned diet | 0.31 | 0.06–1.56 | 0.155 | 2.26 (0.93) | 33.40 | 1.21 (0.71) | 17.93 | 104/17/8 | 411/119/75 |
| Fed homemade cooked diet | 2.47 | 0.66–9.16 | 0.177 | 2.25 (0.92) | 33.36 | 1.20 (0.71) | 17.81 | 94/22/12 | 421/114/71 |
| Diet bought in grocery store | 2.34 | 0.74–7.46 | 0.149 | 2.07 (0.89) | 31.37 | 1.24 (0.72) | 18.73 | 124/29/19 | 391/107/64 |
| Fed rawhide chews in past week | 2.51 | 0.84–7.53 | 0.101 | 2.31 (0.92) | 34.32 | 1.12 (0.69) | 16.73 | 117/38/25 | 398/98/58 |
| Allowed to run freely in dog park | 2.36 | 0.77–7.21 | 0.132 | 2.27 (0.92) | 33.80 | 1.16 (0.70) | 17.29 | 334/84/53 | 181/52/30 |
| Received an herbal product in past month | 2.86 | 0.98–8.33 | 0.054 | 2.18 (0.93) | 32.46 | 1.24 (0.73) | 18.46 | 191/45/24 | 318/91/59 |
Each variable of interest (P ≤ 0.20) was examined 1 at a time in a multilevel model with random effects for dog and household
Value calculated by use of multilevel logistic regression
Associations were considered significant at P ≤ 0.05.
Values reported are number of isolates/number of dogs/number of households.
Univariable associations between pet-related factors and risk of multiclass resistance for 515 Salmonella spp or E coli isolates, 136 dogs, and 83 households* for pet dogs from volunteer households in Ontario from October 2005 through May 2006.
| Variation | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Household | Dog | ||||||||
| Variable | OR | 95% CI† | P value‡ | Value (SE)† | % | Value (SE)† | % | Exposed§ | Unexposed§ |
| Fed raw eggs in past week | 4.26 | 0.64–28.59 | 0.135 | 1.82 (0.95) | 29.95 | 0.96 (0.78) | 15.82 | 54/7/4 | 461/129/79 |
| Owner worked in health-care or veterinary-related field | 0.32 | 0.09–1.18 | 0.087 | 1.67 (0.92) | 28.13 | 0.98 (0.79) | 16.47 | 146/41/27 | 369/95/56 |
| Annual vaccination | 0.42 | 0.13–1.31 | 0.135 | 1.91 (0.97) | 31.13 | 0.94 (0.77) | 15.29 | 331/99/65 | 184/37/18 |
| Fed rawhide chews in past week | 2.18 | 0.71–6.67 | 0.171 | 1.80 (0.94) | 29.96 | 0.92 (0.77) | 15.27 | 117/38/25 | 398/98/58 |
| Received an herbal product in past month | 3.37 | 1.13–10.02 | 0.029 | 1.70 (0.94) | 28.13 | 1.04 (0.80) | 17.29 | 191/45/24 | 318/91/59 |
| Infant (< 1 year old) in household | 5.55 | 0.41–74.40 | 0.195 | 1.70 (0.93) | 28.37 | 1.00 (0.80) | 16.73 | 32/3/2 | 483/133/81 |
See Table 1 for key.
Univariable associations between pet-related factors and risk of ampicillin resistance for 515 Salmonella spp or E coli isolates, 136 dogs, and 83 households* for pet dogs from volunteer households in Ontario from October 2005 through May 2006.
| Variation | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Household | Dog | ||||||||
| Variable | OR | 95% CI† | P value‡ | Value (SE)† | % | Value (SE)† | % | Exposed§ | Unexposed§ |
| Fed homemade diet or homemade food product added to diet | 2.98 | 0.95–9.39 | 0.061 | 2.26 (0.99) | 35.02 | 0.90 (0.70) | 13.90 | 204/39/19 | 311/97/64 |
| Fed homemade raw diet | 2.85 | 0.75–10.83 | 0.124 | 2.28 (0.99) | 35.28 | 0.89 (0.70) | 13.77 | 140/22/10 | 375/114/73 |
| Fed commercial dry or canned diet | 0.33 | 0.06–1.70 | 0.183 | 2.13 (0.96) | 33.62 | 0.91 (0.72) | 14.43 | 104/17/8 | 411/119/75 |
| Fed homemade cooked diet | 2.71 | 0.73–10.10 | 0.138 | 2.20 (0.98) | 34.46 | 0.90 (0.71) | 14.10 | 94/22/12 | 421/114/71 |
| Fed raw chicken in past week | 3.06 | 0.76–12.22 | 0.114 | 2.20 (0.98) | 34.29 | 0.92 (0.72) | 14.42 | 120/18/9 | 395/118/74 |
| Fed raw eggs in past week | 4.08 | 0.56–29.49 | 0.164 | 2.13 (0.96) | 33.60 | 0.92 (0.72) | 14.50 | 54/7/4 | 461/129/79 |
| Diet bought in grocery store | 2.37 | 0.73–7.74 | 0.152 | 1.99 (0.93) | 31.92 | 0.94 (0.73) | 15.16 | 124/29/19 | 391/107/64 |
| Diet prepared at home | 2.94 | 0.70–12.41 | 0.141 | 2.13 (0.96) | 33.65 | 0.91 (0.71) | 14.37 | 104/17/8 | 411/119/75 |
| Diarrhea in past month | 0.27 | 0.05–1.36 | 0.112 | 2.15 (0.96) | 33.98 | 0.89 (0.71) | 14.06 | 98/28/18 | 417/108/65 |
| Annual vaccination | 0.42 | 0.13–1.36 | 0.148 | 2.24 (0.98) | 34.76 | 0.91 (0.71) | 14.13 | 331/99/65 | 184/37/18 |
| Confined to fenced yard | 0.46 | 0.15–1.38 | 0.164 | 1.93 (0.92) | 31.39 | 0.93 (0.73) | 15.17 | 334/84/53 | 181/52/30 |
| Received an herbal product in past month | 3.49 | 1.17–10.42 | 0.025 | 1.96 (0.96) | 31.18 | 1.04 (0.76) | 16.49 | 191/45/24 | 318/91/59 |
| Infant (< 1 y old) in household | 6.08 | 0.41–89.76 | 0.189 | 1.98 (0.94) | 31.78 | 0.97 (0.74) | 15.51 | 32/3/2 | 483/133/81 |
| Other pets in household | 2.84 | 0.93–8.65 | 0.067 | 2.18 (0.96) | 34.45 | 0.86 (0.70) | 13.54 | 150/42/25 | 365/94/58 |
See Table 1 for key.
Results for the univariable multilevel logistic regression models for isolates considered resistant on the basis of MIC breakpoints revealed that bacterial species (E coli vs Salmonella spp), feeding a homemade diet or adding homemade food to the diet, feeding a raw diet or adding a raw food product to the diet, feeding a homemade raw food diet, and feeding raw chicken in the past week were significant risk factors for isolates being resistant to at least 1 antimicrobial (Table 1). Annual vaccination and treatment with a heartworm preventive in the past 6 months were significant sparing factors for resistance to at least 1 antimicrobial. Many of the significant variables for the resistance models were highly correlated (ρ = 0.68 to 0.99) and related to feeding of a raw diet or raw food products.
Results for the univariable multilevel logistic regression models for isolates considered multiclass resistant revealed that a dog receiving an herbal product was the only significant risk factor for isolates being resistant to ≥ 2 antimicrobials (Table 2). Results for the ampicillin-resistant univariable multilevel logistic regression models revealed that a dog receiving an herbal product was again the only significant risk factor for isolates being resistant to ampicillin (Table 3).
Confounder analyses for antimicrobial resistance
When breed size, age category, sex, and neuter status were separately examined for potential confounding effects on the significant univariable associations, all had appreciable effects on most independent variables (Table 4). The greatest confounding effects were for bacterial species (E coli vs Salmonella spp), being fed rawhide treats in the past week, and being allowed to run freely in a dog park. Feeding rawhide treats in the past week and allowing dogs to run freely in a dog park became significant variables with the inclusion of each of the examined confounders; however, the direction of the associations remained the same. Two variable models, and consequently the impact of potentially confounding and extraneous variables, could not be assessed for associations with isolates being considered multiclass resistant and ampicillin resistant because of small effective sample sizes.
Risk factors for the carriage of antimicrobial-resistant Salmonella spp or E coli isolates and their unadjusted and adjusted* ORs on the basis of multilevel models in which controlling for a confounder resulted in a > 20% change in the model coefficient.
| Confounder | Variable | Unadjusted OR* | 95% CI*† | P value*† | Adjusted OR‡ | 95% CI‡ | P value†‡ |
|---|---|---|---|---|---|---|---|
| Breed size | Bacterial species isolated (E coli vs Salmonella spp) | 2.34 | 1.06–5.19 | 0.036 | 7.96 | 2.39–26.51 | 0.001 |
| Owner worked in health-care or veterinary-related field | 0.33 | 0.10–1.07 | 0.065 | 0.45 | 0.13–1.55 | 0.206 | |
| Received heartworm preventive in past 6 months | 0.31 | 0.10–0.93 | 0.036 | 0.23 | 0.07–0.74 | 0.014 | |
| Fed commercial dry or canned diet | 0.31 | 0.06–1.56 | 0.155 | 0.51 | 0.06–4.20 | 0.532 | |
| Fed homemade cooked diet | 2.47 | 0.66–9.16 | 0.177 | 3.42 | 0.88–13.35 | 0.076 | |
| Diet bought in grocery store | 2.34 | 0.74–7.46 | 0.149 | 2.62 | 0.74–9.21 | 0.134 | |
| Fed rawhide chews in past week | 2.51 | 0.84–7.53 | 0.101 | 3.90 | 1.22–12.50 | 0.022 | |
| Allowed to run freely in dog park | 2.36 | 0.77–7.21 | 0.132 | 4.22 | 1.18–15.11 | 0.027 | |
| Age category | Bacterial species isolated (E coli vs Salmonella spp) | 2.34 | 1.06–5.19 | 0.036 | 7.58 | 2.28–25.19 | 0.001 |
| Fed homemade diet or homemade food added to diet | 3.50 | 1.18–10.37 | 0.024 | 4.51 | 1.36–14.91 | 0.014 | |
| Diarrhea in past month | 0.38 | 0.10–1.50 | 0.169 | 0.27 | 0.06–1.24 | 0.092 | |
| Fed commercial dry or canned diet | 0.31 | 0.06–1.56 | 0.155 | 0.47 | 0.06–3.84 | 0.483 | |
| Fed homemade cooked diet | 2.47 | 0.66–9.16 | 0.177 | 3.78 | 0.98–14.59 | 0.053 | |
| Fed rawhide chews in past week | 2.51 | 0.84–7.53 | 0.101 | 3.89 | 1.21–12.48 | 0.022 | |
| Allowed to run freely in dog park | 2.36 | 0.77–7.21 | 0.132 | 3.79 | 1.08–13.32 | 0.038 | |
| Sex | Bacterial species isolated (E coli vs Salmonella spp) | 2.34 | 1.06–5.19 | 0.036 | 7.92 | 2.40–26.17 | 0.001 |
| Owner worked in health-care or veterinary-related field | 0.33 | 0.10–1.07 | 0.065 | 0.42 | 0.13–1.36 | 0.148 | |
| Annual vaccination | 0.30 | 0.10–0.90 | 0.031 | 0.33 | 0.10–1.02 | 0.054 | |
| Received heartworm preventive in past 6 months | 0.31 | 0.10–0.93 | 0.036 | 0.22 | 0.07–0.68 | 0.009 | |
| Fed commercial dry or canned diet | 0.31 | 0.06–1.56 | 0.155 | 0.48 | 0.06–3.67 | 0.483 | |
| Fed homemade cooked diet | 2.47 | 0.66–9.16 | 0.177 | 3.11 | 0.86–11.20 | 0.083 | |
| Fed rawhide chews in past week | 2.51 | 0.84–7.53 | 0.101 | 3.94 | 1.29–12.00 | 0.016 | |
| Allowed to run freely in dog park | 2.36 | 0.77–7.21 | 0.132 | 3.41 | 1.01–11.51 | 0.048 | |
| Neuter status | Bacterial species isolated (E coli vs Salmonella spp) | 2.34 | 1.06–5.19 | 0.036 | 7.63 | 2.32–25.08 | 0.001 |
| Fed commercial dry or canned diet | 0.31 | 0.06–1.56 | 0.155 | 0.48 | 0.06–3.64 | 0.479 | |
| Fed homemade cooked diet | 2.47 | 0.66–9.16 | 0.177 | 3.32 | 0.93–11.88 | 0.065 | |
| Diet bought in grocery store | 2.34 | 0.74–7.46 | 0.149 | 2.49 | 0.76–8.20 | 0.133 | |
| Fed rawhide chews in past week | 2.51 | 0.84–7.53 | 0.101 | 3.74 | 1.23–11.44 | 0.021 | |
| Allowed to run freely in dog park | 2.36 | 0.77–7.21 | 0.132 | 3.68 | 1.10–12.27 | 0.034 |
Represents results for 457 bacterial isolates, 127 dogs, and 78 households of dogs for which breed size, sex, neuter status, and age category were provided.
In the adjusted models, 1 potential confounder (ie, breed size, age category, or sex and neuter status) was examined at a time because of a limited effective sample size. Initial multilevel logistic regression models were performed with random intercepts for household and dog
Values were considered significant at P ≤ 0.05
Value obtained with 2-variable multilevel logistic regression models, including the independent variable and potential confounder and random intercepts for household and dog.
Residuals for the multilevel models that included confounding variables were examined for outliers and influential covariate patterns. There were a small number of observations with large residuals, which were examined for mistakes in data entry and for the impact of their removal from the models; because of a lack of errors in data entry, all observations were retained in the models. The best linear unbiased predictors for each of the multilevel models for isolates considered resistant on the basis of MIC breakpoints appeared to be normally distributed.
Estimation of variance
Multilevel random effects in the univariable models for isolates considered resistant, multiclass resistant, and ampicillin resistant on the basis of MIC breakpoints allowed for comparison of variance within and between households and dogs. Approximately 28% to 35% and 14% to 20% of the variance in the models for isolates considered resistant, multiclass resistant, and ampicillin resistant on the basis of MIC breakpoints were explained at the household and dog levels, respectively (Tables 1–3).
Discussion
In the study reported here, several potential pet-related risk factors for carriage of antimicrobial-resistant Salmonella spp or E coli by pet dogs from volunteer households in Ontario were examined. This is one of the few studies in which potential risk factors for antimicrobial resistance in Salmonella spp and E coli in pet dogs in Canada have been examined. The prevalence of antimicrobial resistance in Salmonella spp or E coli isolates was approximately 20%, and the prevalence of carriage of an antimicrobial-resistant isolate by a dog (ie, the percentage of dogs sampled that were carrying an antimicrobial resistant isolate) was approximately 28%. Multilevel logistic regression models were used to identify several potentially important pet-related management factors associated with an increased risk of antimicrobial resistance in Salmonella spp or E coli in this population of pet dogs. Many of the associations were related to feeding raw food, which highlights the potential public health risk of adding raw animal products to pet dog diets. Specifically, bacterial species (E coli vs Salmonella spp), feeding a homemade diet or adding homemade food to the diet, feeding a raw diet or adding a raw food product to the diet, feeding a homemade raw food diet, and feeding raw chicken in the past week were all significant risk factors for carriage of antimicrobial-resistant Salmonella spp or E coli. On the basis of variance estimates for dog and household from the multilevel models for antimicrobial-resistant bacteria, it was found that approximately 50% of the variance was explained at the household and dog levels, with the residual variance at the isolate level.
Results for the univariable and 2-variable models for antimicrobial resistance generally revealed that bacterial species (E coli vs Salmonella spp) and several pet-related management factors associated with feeding raw food remained significant even after controlling separately for confounding variables (breed size, age category, sex, and neuter status). Antimicrobial-resistant Salmonella spp and E coli are commonly recovered from raw animal products in Canada, particularly chicken and turkey.32 Consequently, raw animal products or raw diets made with these products could be a potential source of antimicrobial-resistant Salmonella spp and E coli.33,34 In 2011, the Canadian Integrated Program for Antimicrobial Resistance Surveillance recovered E coli with resistance to at least 1 antimicrobial from approximately 75% of retail chicken meat samples collected from across Canada; Salmonella spp with resistance to at least 1 antimicrobial were recovered from approximately 56% of the same retail chicken meat samples.32 Ampicillin resistance in E coli was common in Canadian retail meat samples, with approximately 47% of the E coli recovered from chicken meat samples having a decreased susceptibility to ampicillin.32 Ampicillin resistance was also common among Salmonella isolates recovered from retail meat samples, with approximately 41% of isolates recovered from retail chicken meat being resistant to this drug, depending on the province where the meat was purchased.32 The exact ingredients of the raw and homemade diets fed to the pet dogs in the present study were not identified, so a clear connection with the meat source could not be made. However, antimicrobial-resistant Salmonella spp and E coli have been recovered from commercially prepared raw diets available in Canada and the United States,33,35 and raw meat consumption has been significantly associated with carriage of multidrug-resistant E coli in therapy dogs.5
Through univariable analyses, administration of herbal products to dogs was identified as a significant risk factor for isolates considered multiclass resistant and ampicillin resistant (Tables 2 and 3). In the present study, 45 dog owners indicated that they were giving their dogs an herbal product. This question was intended to identify naturopathic products used by dog owners. Of these, information on the herbal product was available for 20 dogs; glucosamine was the product most commonly recorded. To our knowledge, this association has not been reported in other studies of pet dogs, and it may be acting as a proxy for other pet-related factors; however, because of the large number of variables investigated, it may be a type I error. In addition, complete information regarding the herbal products given to the dogs in this study was not available, which makes it difficult to reach a conclusion on the basis of this association. Multidrug-resistant Salmonella spp and E coli have been recovered from dogs in other studies,1,2,36,37 but an association with medications or herbal products that the dogs had been given was not reported.
Annual vaccination and treatment with a heartworm preventive in the previous 6 months were found to be sparing factors in the models for isolates considered resistant on the basis of MIC breakpoints. To our knowledge, this association has not been reported in previous studies, and these variables could be acting as proxies for consistent veterinary care. In addition, we speculate that the variables could be acting as a proxy for feeding processed dry or canned diets because owners who follow these particular preventive health measures may be more likely to purchase these diets. Consequently, these pets may be less likely to be fed raw diets or raw animal products, which reduces their risk of carrying Salmonella spp.34
Controlling for confounders (breed size, age category, sex, and neuter status) in the 2 variable multilevel models for antimicrobial-resistant bacteria revealed that feeding a rawhide treat in the past week and allowing a dog to run freely in a dog park became significant variables, assuming the examined confounders were not intervening variables. These associations have not been reported previously. However, our research group previously reported22 that being fed a rawhide treat was associated with a decreased risk for carriage of Salmonella spp, but this association was nonsignificant when breed was introduced into the model. Antimicrobial-resistant Salmonella organisms have been recovered from rawhide treats,34 and it is possible they could act as a source for antimicrobial-resistant Salmonella infection in dogs. The association of allowing a dog to run freely in a dog park makes sense biologically in that such dogs would be exposed to many other dogs and other sources of bacteria.
The point estimate for species of bacteria (E coli vs Salmonella spp) increased when we controlled for confounders, with E coli isolates being > 7 times as likely to be antimicrobial resistant as Salmonella isolates. Escherichia coli of various types are commonly used in antimicrobial-resistance studies38–40 because they readily acquire and maintain resistance genes from other bacteria. However, the isolation of E coli has not been found to be a reliable predictor of resistance in Salmonella spp at the animal or sample level.23,38
The high amount of variation explained at the dog and household level for the various antimicrobial resistance outcomes was expected. Clustering of infection status within a herd or household is typical in most infectious diseases and must be accounted for in the analysis if multiple subjects are from a common household or herd.29,41 Analysis of the results of the present study indicated that interventions targeted at the household level, reflecting general pet-related management of a dog (eg, diet or veterinary care), could be effective in limiting antimicrobial-resistant bacteria in pet dogs.
The design of the present study needs to be considered when interpreting the results. Furthermore, caution must be used when extrapolating these findings to other dog and pet populations. Dogs were recruited through convenience sampling (the data used were a subset from a larger study that could not feasibly use random sampling to recruit dogs), and this group of dogs may not have been representative of the typical canine population in Ontario. In addition, if dogs were related through common social groups (eg, therapy or agility programs), there may have been some unrecorded clustering that was not accounted for in the analyses. Salmonella spp carriage was abnormally high in this population of dogs, which may have been a result of the large number of dogs fed a raw diet,22 but it may also have been attributable to the fact that > 1 method of isolation was used with 3 selective media, which improved test sensitivity and recovery of Salmonella organisms. However, given that potential clustering by dog and household was accounted for in the analyses and that a large number of pet-related factors were examined, the results are important for the development of guidelines for safe pet ownership and the development of antimicrobial-resistance surveillance programs for companion animals.
Because of the large number of variables examined in the study reported here, it was possible that some significant associations may simply have been attributable to chance. Selection bias may also have been an issue and would have affected the estimated ORs if the recruitment process affected both infection status and exposure, but not either alone.29,42 This bias has been discussed extensively elsewhere22 and would have been an issue only if the owners’ decisions to participate were based on bacterial status of their dogs and antimicrobial susceptibility of the bacteria isolated from their dogs. Although recruitment could have been affected by exposure, it is unlikely that owner participation was affected by bacterial carriage. Finally, given that this study was cross-sectional in nature and that prevalence is a function of incidence and duration, we cannot determine which of the factors caused carriage of antimicrobial-resistant Salmonella spp and E coli and which of the factors prolonged carriage.29,42 However, controlling management factors related to prevalence would be useful for protecting public health, if not necessarily suitable for understanding the causal relationship between exposure and carriage of pathogens or antimicrobial resistance.
A major limitation of the present study was that previous antimicrobial use was not investigated through the survey because the original project was not specifically designed for that purpose. Antimicrobial administration is an important risk factor for the development of antimicrobial resistance in bacteria and has been identified as a risk factor in several studies of dogs.18–21,36 However, it should be noted that these studies often were designed to examine E coli from clinical infections and dogs admitted to veterinary hospitals, which may not be comparable to E coli isolated from healthy dogs.2,43
Finally, given the small sample size in the present study, caution should be used when interpreting the nonsignificant results. Potential risk factors for carriage of antimicrobial-resistant Salmonella spp and E coli may have been missed because of the large effect or the small amount of variation needed in small studies to detect significant associations.29 Weaker associations may not have been detected, especially for the models used to examine isolates considered multiclass resistant and ampicillin resistant. In addition, we were restricted in our ability to control for confounding variables. However, it should be noted that the sample size in this study was comparable to or larger than those in many previous studies of pet dogs,7,18,20,21,44 which ranged from < 10 to almost 200 dogs.
The study reported here identified several potentially important risk factors for carriage of antimicrobial-resistant Salmonella spp and E coli in pet dogs. Therefore, it can be concluded that pet dogs are a potential source of antimicrobial-resistant Salmonella spp and E coli. These results may warrant a change in the current methods for surveillance of antimicrobial resistance and support the recommendation that companion animals be included in antimicrobial-resistance surveillance programs.45 In addition, the associations with feeding raw food diets and raw animal products highlighted the potential health risk of adding raw or inadequately cooked animal products to diets of pet dogs. The protective associations related to vaccination and other veterinary care require further investigation. Given the close relationship that most owners have with their dogs, further studies into pet-related health management factors that may increase or decrease a dog's risk for carriage of antimicrobial-resistant bacteria should be performed. Identification of potential risk factors could be valuable as preliminary data for future, more targeted epidemiological studies to address antimicrobial resistance in pet dogs. Such studies would be crucial for the development of evidence-based guidelines for safe dog ownership and responsible pet management to protect the public. Such information is particularly important for pet owners who are immunocompromised.
Acknowledgments
This manuscript represents a portion of a thesis submitted by Dr. Leonard to the University of Guelph Ontario Veterinary College as partial fulfillment of the requirements for a Doctor of Philosophy degree. The study was completed at the University of Guelph.
Supported by a grant from the Canada Foundation for Innovation and the Ontario Research Fund awarded to Dr. Pearl. Support for Dr. Leonard was provided by the Blake Graham Fellowship from the Ontario Veterinary College. Support was provided by Public Health Agency of Canada for sample collection and testing.
Presented in abstract form at the Conference for Research Workers in Animal Disease, Chicago, December 2012.
ABBREVIATIONS
| CI | Confidence interval |
| MIC | Minimum inhibitory concentration |
Footnotes
The questionnaire is available from the corresponding author on request.
Office Internationale des Epizooties Reference Laboratory for Salmonellosis, Laboratory for Foodborne Zoonoses, Public Health Agency of Canada, Guelph, ON, Canada.
LFZ Susceptibility Testing Laboratory, Laboratory for Foodborne Zoonoses, Public Health Agency of Canada, Guelph, ON, Canada.
Sensititre, Trek Diagnostic Systems Ltd, Independence, Ohio.
Intercooled Stata/MP, version 11.1, StataCorp LP, College Station, Tex.
Leckie G, Charlton C. runmlwin: Stata module for fitting multilevel models in the MLwiN software package, Centre for Multilevel Modelling, University of Bristol, Bristol, England.
MLwiN, version 2.1, Centre for Multilevel Modelling, University of Bristol, Bristol, England.
References
1. Guardabassi L, Schwarz S, Lloyd DH. Pet animals as reservoirs of antimicrobial-resistant bacteria. J Antimicrob Chemother 2004;54:321–332.
2. Weese JS. Antimicrobial resistance in companion animals. Anim Health Res Rev 2008;9:169–176.
3. Lloyd DH. Reservoirs of antimicrobial resistance in pet animals. Clin Infect Dis 2007;45(suppl 2):S148–S152.
4. Ogeer-Gyles J, Mathews KA, Sears W, et al. Development of antimicrobial drug resistance in rectal Escherichia coli isolates from dogs hospitalized in an intensive care unit. J Am Vet Med Assoc 2006;229:694–699.
5. Lefebvre SL, Reid-Smith R, Boerlin P, et al. Evaluation of the risks of shedding salmonellae and other potential pathogens by therapy dogs fed raw diets in Ontario and Alberta. Zoonoses Public Health 2008;55:470–480.
6. Acke E, McGill K, Quinn T, et al. Antimicrobial resistance profiles and mechanisms of resistance in Campylobacter jejuni isolates from pets. Foodborne Pathog Dis 2009;6:705–710.
7. Murphy C, Reid-Smith RJ, Prescott JF, et al. Occurrence of antimicrobial resistant bacteria in healthy dogs and cats presented to private veterinary hospitals in southern Ontario: a preliminary study. Can Vet J 2009;50:1047–1053.
8. Cherry B, Burns A, Johnson GS, et al. Salmonella Typhimurium outbreak associated with a veterinary clinic. Emerg Infect Dis 2004;10:2249–2251.
9. Goossens H, Vlaes L, Butzler JP, et al. Campylobacter upsaliensis enteritis associated with canine infections. Lancet 1991;337:1486–1487.
10. Acha PN, Szyfres B. Campylobacteriosis. In: Acha PN, Szyfres B, eds. Zoonoses and communicable diseases common to man and animals. 3rd ed. Washington, DC: Pan American Health Organization, Pan American Sanitary Bureau, Regional Office of the World Health Organization, 2001;67–78.
11. Acha PN, Szyfres B. Salmonellosis. In: Acha PN, Szyfres B, eds. Zoonoses and communicable diseases common to man and animals. 3rd ed. Washington, DC: Pan American Health Organization, Pan American Sanitary Bureau, Regional Office of the World Health Organization, 2001;233–246.
12. Boerlin P, White DG. Antimicrobial resistance and its epidemiology. In: Giguère S, Prescott JF, Baggot JD, et al, eds. Antimicrobial therapy in veterinary medicine. 4th ed. Ames, Iowa: Blackwell Publishing, 2006;27–43.
13. Jackson CR, Fedorka-Cray PJ, Davis JA, et al. Prevalence, species distribution and antimicrobial resistance of enterococci isolated from dogs and cats in the United States. J Appl Microbiol 2009;107:1269–1278.
14. Johnson JK, Perencevich EN, Lincalis DP, et al. Dog bite transmission of antibiotic-resistant bacteria to a human. Infect Control Hosp Epidemiol 2006;27:762–763.
15. Johnson JR, Clabots C. Sharing of virulent Escherichia coli clones among household members of a woman with acute cystitis. Clin Infect Dis 2006;43:e101–e108.
16. Perrin T. The business of urban animals survey: the facts and statistics on companion animals in Canada. Can Vet J 2009;50:48–52.
17. Stull JW, Peregrine AS, Sargeant JM, et al. Household knowledge, attitudes and practices related to pet contact and associated zoonoses in Ontario, Canada. BMC Public Health [serial online]. 2012;12:553. Available at: www.biomedcentral.com/1471-2458/12/553. Accessed May 17, 2015.
18. Rantala M, Lahti E, Kuhalampil J, et al. Antimicrobial resistance in Staphylococcus spp., Escherichia coli and Enterococcus spp. in dogs given antibiotics for chronic dermatological disorders, compared with non-treated control dogs. Acta Vet Scand 2004;45:37–45.
19. Mentula S, Virtanen T, Kanervo-Nordstrom A, et al. Relatedness of Escherichia coli strains with different susceptibility patterns isolated from Beagle dogs during ampicillin treatment. Int J Antimicrob Agents 2006;27:46–50.
20. Damborg P, Sorensen AH, Guardabassi L. Monitoring of antimicrobial resistance in healthy dogs: first report of canine ampicillin-resistant Enterococcus faecium clonal complex 17. Vet Microbiol 2008;132:190–196.
21. Grønvold AM, L'abée-Lund TM, Sørum H, et al. Changes in fecal microbiota of healthy dogs administered amoxicillin. FEMS Microbiol Ecol 2010;71:313–326.
22. Leonard EK, Pearl DL, Finley RL, et al. Evaluation of pet-related management factors and the risk of Salmonella spp. carriage in pet dogs from volunteer households in Ontario (2005–2006). Zoonoses Public Health 2011;58:140–149.
23. Leonard EK, Pearl DL, Finley RL, et al. Comparison of antimicrobial resistance patterns of Salmonella spp. and Escherichia coli recovered from pet dogs from volunteer households in Ontario (2005–06). J Antimicrob Chemother 2012;67:174–181.
24. Canadian Integrated Program for Antimicrobial Resistance Surveillance. Preliminary results. In: 2007 CIPARS annual report. Guelph, ON, Canada: Public Health Agency of Canada, 2008. Available at: www.phac-aspc.gc.ca/cipars-picra/pdf/2007pr-eng.pdf. Accessed May 26, 2013.
25. US FDA. Enteric bacteria. In: National Antimicrobial Resistance Monitoring System 2007 executive report. Rockville, Md: US FDA, 2010. Available at: www.fda.gov/downloads/AnimalVeterinary/SafetyHealth/AntimicrobialResistance/NationalAntimicrobialResistanceMonitoringSystem/UCM209494.pdf. Accessed May 26, 2013.
26. Clinical Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. M100–S20. 20th informational supplement. Wayne, Pa: Clinical Laboratory Standards Institute, 2010.
27. Veterinary Drugs Directorate. Proposed categorization of antimicrobial drugs. In: Current thinking on risk management measures to address antimicrobial resistance associated with the use of antimicrobial agents in food-producing animals. Ottawa: Health Canada, 2005.
28. Clinical Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing. M100–S16. 16th informational supplement. Table 2A. M7–A6-MIC testing section. Wayne, Pa: Clinical Laboratory Standards Institute, 2010.
29. Dohoo IR, Martin W, Stryhn H. Veterinary epidemiologic research. 2nd ed. Charlottetown, PE, Canada: University of Prince Edward Island, 2009;865.
30. Goldstein H. Discrete response data. In: Multilevel statistical models. 3rd ed. London: Arnold, 2003;98–104.
31. Thrusfield MV. Veterinary epidemiology. 3rd ed. Ames, Iowa: Blackwell Science Ltd, 2005;624.
32. Canadian Integrated Program for Antimicrobial Resistance Surveillance. CIPARS short report 2011. Guelph, ON, Canada: Public Health Agency of Canada, 2012. Available at: publications.gc.ca/collections/collection_2013/aspc-phac/HP2-4-2-2011-eng.pdf. Accessed May 28, 2014.
33. Finley R, Reid-Smith R, Ribble C, et al. The occurrence and antimicrobial susceptibility of salmonellae isolated from commercially available canine raw food diets in three Canadian cities. Zoonoses Public Health 2008;55:462–469.
34. Finley R, Reid-Smith R, Ribble C, et al. The occurrence and anti-microbial susceptibility of salmonellae isolated from commercially available pig ear pet treats. Zoonoses Public Health 2008;55:455–461.
35. Strohmeyer RA, Morley PS, Hyatt DR, et al. Evaluation of bacterial and protozoal contamination of commercially available raw meat diets for dogs. J Am Vet Med Assoc 2006;228:537–542.
36. Gibson JS, Morton JM, Cobbold RN, et al. Risk factors for multidrug-resistant Escherichia coli rectal colonization of dogs on admission to a veterinary hospital. Epidemiol Infect 2011;139:197–205.
37. Shaheen BW, Boothe DM, Oyarzabal OA, et al. Antimicrobial resistance profiles and clonal relatedness of canine and feline Escherichia coli pathogens expressing multidrug resistance in the United States. J Vet Intern Med 2010;24:323–330.
38. Varga C, Rajic A, McFall ME, et al. Comparison of antimicrobial resistance in generic Escherichia coli and Salmonella spp. cultured from identical fecal samples in finishing swine. Can J Vet Res 2008;72:181–187.
39. Caprioli A, Busani L, Martel JL, et al. Monitoring of antibiotic resistance in bacteria of animal origin: epidemiological and microbiological methodologies. Int J Antimicrob Agents 2000;14:295–301.
40. Martel JL, Tardy F, Sanders P, et al. New trends in regulatory rules and surveillance of antimicrobial resistance in bacteria of animal origin. Vet Res 2001;32:381–392.
41. Diez Roux AV, Aiello AE. Multilevel analysis of infectious diseases. J Infect Dis 2005;191(suppl 1):S25–S33.
42. Rothman KJ, Greenland S, Poole C, et al. Basic concepts. In: Rothman KJ, Greenland S, Lash TL, eds. Modern epidemiology. 3rd ed. Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins, 2008;5–86.
43. Prescott JF. Antimicrobial use in food and companion animals. Anim Health Res Rev 2008;9:127–133.
44. Gibson JS, Cobbold RN, Kyaw-Tanner MT, et al. Fluoroquinolone resistance mechanisms in multidrug-resistant Escherichia coli isolated from extraintestinal infections in dogs. Vet Microbiol 2010;146:161–166.
45. Day MJ, Breitschwerdt E, Cleaveland S, et al. Surveillance of zoonotic infectious disease transmitted by small companion animals. Emerg Infect Dis 2012;18. [serial online]. Available at: wwwnc.cdc.gov/eid/article/18/12/12-0664_article. Accessed Oct 6, 2013.
Appendix 1
Resistance breakpoints for antimicrobial susceptibility testing of Salmonella spp and generic Escherichia coli isolates cultured from fecal samples collected from pet dogs of volunteer households in Ontario from October 2005 through May 2006.
| Antimicrobial | Resistance breakpoint (μg/mL) | Class | Category* |
|---|---|---|---|
| Amikacin† | ≥ 64 | Aminoglycoside | II |
| Amoxicillin–clavulanic acid† | ≥ 32 | Penicillin–β-lactamase inhibitor combination | I |
| Ampicillin† | ≥ 32 | Penicillin | II |
| Cefoxitin† | ≥ 32 | Second-generation cephalosporin | II |
| Ceftiofur† | ≥ 8 | Third-generation cephalosporin | I |
| Ceftriaxone‡ | ≥ 4 | Third-generation cephalosporin | I |
| Chloramphenicol | ≥ 32 | Chloramphenicol | III |
| Ciprofloxacin† | ≥ 4 | Fluoroquinolone | I |
| Gentamicin† | ≥ 16 | Aminoglycoside | II |
| Kanamycin† | ≥ 64 | Aminoglycoside | II |
| Nalidixic acid† | ≥ 32 | Quinolone | II |
| Streptomycin§ | ≥ 64 | Aminoglycoside | II |
| Sulfisoxazole† | ≥ 512 | Sulfonamide | III |
| Tetracycline† | ≥ 16 | Tetracycline | III |
| Trimethoprim-sulfamethoxazole* | ≥ 4 | Sulfonamide combination | II |
Categories were defined as follows27: I = antimicrobials of very high importance in human medicine, essential to the treatment of serious bacterial infections; (no alternatives for infections attributable to resistant bacteria); II = antimicrobials of high importance in human medicine, used to treat a variety of infections; (an alternative for bacteria resistant to category III antimicrobials); and III = antimicrobials of medium importance in human medicine, used as first-line drugs; (alternatives for resistant bacteria are generally available).
Resistance determined by use of criteria reported elsewhere.28
Resistance determined by use of criteria reported elsewhere.26
Resistance determined by use of criteria reported elsewhere.25
Appendix 2
Variables (as modified from a previous report22) investigated for an association with antimicrobial resistance in Salmonella spp and E coli isolated from fecal samples collected from pet dogs of volunteer households in Ontario from October 2005 through May 2006.
| Category | Variable |
|---|---|
| General diet information | Fed raw beef in past week, fed raw chicken in past week, fed raw pork in past week, fed raw eggs in past week, fed dried pig ears in past week, fed bones in past week, fed rawhide chews in past week, fed cooked table scraps in past week, fed other pet treats in past week, primary type of diet fed (commercial dry or canned, commercial raw diet, homemade raw diet, homemade cooked diet, commercial home-cooked diet, or other), anything added to commercial dry or canned diet, and combination variables (fed raw meat or eggs in past week, fed raw food [raw diet or raw food product added to diet], or homemade food [fed homemade diet or fed homemade food added to diet]) |
| Dog health information and demographics | Received a probiotic in past month, received an herbal product in past month, diarrhea in past month, vomiting in past month, age, age category (puppy [< 1 year old], adult [1 to 7 years old], or senior [> 7 years old]), breed size (small pure breed [< 11.4 kg], medium pure breed [11.4 to 27.3 kg], large or giant pure breed [> 27.3 kg], or mixed breed [all body weights]), sex (male or female), neuter status (sexually intact or neutered-spayed), received annual vaccination (yes or no), received heartworm preventive in past 6 months (yes or no), and retriever-type breed (yes or no) |
| Activities and dog information | Purchase of dog's diet (grocery store, pet store, or prepared at home), > 1 dog in household (yes or no), cats in household (yes or no), other pets in household (yes or no), allowed to run freely in park (yes or no), confined to fenced yard (yes or no), contact with livestock (cattle, sheep, goats, pigs, or horses [yes or no]), and contact with cats other than those in household (yes or no) |
| Household information | Number of children (< 18 years old) in household, infant (< 1 year old) in household (yes or no), person > 65 years old in household (yes or no), immunocompromised person in household (yes or no), and owner's occupation |
| Bacterial species isolated | Salmonella spp or E coli |
