Abstract
The objectives of this study were to investigate the associations between antimicrobial resistance patterns in generic Escherichia coli and Salmonella spp. isolates recovered from identical pen pooled fecal samples, and to evaluate potential clustering of multiple isolates of these organisms within identical fecal samples. Up to 5 generic E. coli (n = 922 isolates) and Salmonella spp. (n = 922 isolates) isolates were obtained from each of 188 pen pooled fecal samples that had been collected from 45 finishing swine farms in Alberta in 2000, and tested for susceptibility to 15 antimicrobials. No isolates of either organism were resistant to 3rd generation cephalosporins or fluoroquinolones, which in Canada are considered antimicrobials of very high importance to human health. Approximately twice as many generic E. coli isolates as Salmonella spp. isolates were resistant to at least 1 antimicrobial. In addition, E. coli isolates showed more multidrug-resistance patterns. No significant association was observed between the resistance phenotypes of Salmonella spp. and E. coli at the fecal sample level. More clustering at the sample level was observed for proportions of antimicrobial resistance (AMR) in Salmonella spp. isolates than E. coli indicating that in future studies it might be sufficient to test fewer than 5 Salmonella spp. isolates per sample.
Résumé
Les objectifs de la présente étude étaient d’évaluer les associations entre les patrons de résistance aux antimicrobiens d’isolats d’Escherichia coli génériques et de Salmonella spp. provenant de pools d’échantillons de fèces prélevés dans des parcs identiques, et d’évaluer le potentiel de regroupement d’isolats multiples de ces micro-organismes à l’intérieur d’échantillons de fèces identiques. Un maximum de 5 isolats d’E. coli génériques (n = 922) et de Salmonella spp. (n = 922) ont été obtenus de chacun des 188 échantillons de fèces provenant des parcs qui avaient été amassés en 2000 dans 45 troupeaux de porcs en finition en Alberta, et testés pour leur sensibilité à 15 agents antimicrobiens. Aucun isolat des deux micro-organismes n’était résistant aux céphalosporines et fluoroquinolones de 3e génération, qui au Canada sont considérés comme des agents antimicrobiens de très haute importance pour la santé humaine. Environ deux fois plus d’isolats d’E. coli générique que de Salmonella spp. étaient résistants à au moins 1 agent antimicrobien. De plus, les isolats d’E. coli montraient plus de patrons de résistance multiple. Aucune association significative n’a été notée entre les phénotypes de résistance de Salmonella spp. au niveau de l’échantillon de fèces. Plus de regroupements au niveau de l’échantillon étaient observées pour les proportions de résistance aux antimicrobiens (AMR) chez les isolats de Salmonella spp. que ceux d’E. coli indiquant ainsi que dans des études ultérieures il serait suffisant de tester moins que 5 isolats de Salmonella spp. par échantillon.
(Traduit par Docteur Serge Messier)
Introduction
The increasing incidence of resistant bacteria in humans, animals, and the environment is a major concern globally (1–4) and is the subject of increased monitoring. Most active AMR monitoring programs utilize selected zoonotic pathogens and sentinel (indicator) bacteria isolated from animals, humans, and the environment (5–11). This approach allows comparisons of the prevalence of resistance in bacteria from animals, humans, and the environment, as well as detection of resistance genes transferable among different populations (12).
Generic E. coli are frequently used as indicator bacteria to monitor trends in AMR patterns (13) because they are prevalent commensal enteric bacteria in humans and animals, are cultured easily and inexpensively (14), and they can acquire and preserve resistance genes from other organisms in the environment and in animal populations (15). Escherichia coli are also considered good indicators of the selective pressure imposed by antimicrobial use (AMU) in food animals (16,17).
Multi-resistant Salmonella Typhimurium DT104 (2) and DT193 (18,19), associated with pork and other foods of animal origin, are examples of specific Salmonella that have caused outbreaks of food-borne diseases in humans. These organisms may also cause disease in animals and thus, Salmonella spp. are often monitored through routine public health and veterinary diagnostic systems (20), including AMR surveillance programs (5–11,12,21).
There is, however, a lack of information on the relationship between E. coli and Salmonella spp. with respect to AMR. Thus, the objectives of this study were to investigate potential associations between the AMR patterns in isolates of generic E. coli and Salmonella spp. recovered from 188 pen pooled fecal samples and to evaluate possible clustering of multiple isolates of these organisms within identical samples.
Materials and methods
Sample collection and Salmonella spp. isolation
Pen pooled fecal samples were collected in the year 2000, as part of a cross-sectional study, described in detail elsewhere (22). In brief, a convenience sample of 90 Alberta finishing swine farms producing ≥ 2000 pigs per year were selected by 10 veterinarians from their client list, based on the producers’ willingness to participate in this study. The veterinarians visited the farms 3 times at approximately 4- to 6-week intervals. At each visit, the veterinarians randomly selected 5 pens, and from each pen they collected a 25 g pooled sample of fecal material (5 g samples of feces were taken from 5 different spots within the pen), for an average total of 15 pooled samples per farm. Agri-Food Laboratory Branch, Food Safety Division, Alberta Agriculture and Food (Edmonton, Alberta) conducted all laboratory tests. All pen pooled fecal samples were cultured for Salmonella as described in the following text, and up to 5 colonies from each Salmonella-positive sample were obtained for antimicrobial susceptibility testing. Two 5 g fecal aliquots were selected for specific enrichment methods; one was inoculated into 45 mL of selenite cystine broth (SCB) and incubated at 35°C for 24 h. The SCB was then inoculated onto XLT4 and Rambach agar plates, incubated at 35°C and evaluated for typical colonies of Salmonella species at 24 and 48 h. The 2nd 5 g aliquot was added to 45 mL of buffered peptone water (BPW) and incubated at 35°C for 24 h. The BPW solution was inoculated into 10 mL of tetrathionate broth (TB) supplemented with 0.2 mL of iodine and incubated for 24 h at 35°C. The TB was then inoculated onto XLT4 agar plates, to be incubated at 35°C for 24 h, and onto modified semisolid Rappaport-Vassiliadis (MSRV) plates, to be incubated at 41.5°C for 20–24 h. The “halos” of growth that occurred on the MSRV plates were subsequently streaked onto XLT4 and Rambach plates and incubated at 35°C for 24 h. Suspected Salmonella colonies from the XLT4 and Rambach plates were screened using lysine iron agar plates and triple sugar iron and urease agar slants. Salmonella-suspect colonies were screened using Salmonella poly O and O1 antisera agglutination (Denka Seiken, Tokyo, Japan) or a Salmonella latex agglutination kit (Oxoid, Basingstoke Hampshire, UK). At least one Salmonella-positive pooled fecal sample was detected in 45 out of the 90 farms studied; thus, the number of farms represented in this study was 45. A 2-g aliquot of each Salmonella — positive fecal sample was frozen for subsequent culture of generic E. coli. Five presumptive Salmonella colonies per sample were harvested and frozen at −70°C for subsequent susceptibility testing.
Generic E. coli isolation
All pen pooled fecal samples (n = 188) from which Salmonella spp. had been recovered in the previous study (22), were individually cultured for generic E. coli. A swab from each sample was streaked onto a MacConkey’s agar plate and incubated at 35°C for 24 h. Suspect colonies were plated onto blood and MacConkey’s agar plates to harvest 5 presumptive colonies from each pen pooled fecal sample. These colonies were confirmed to be E. coli using EC-4-methylumbelliferyl-β-D-glucuronide and indol, methyl red, Voges-Proskauer, and citrate tests (23).
Susceptibility testing of Escherichia coli and Salmonella spp. isolates
Generic E. coli isolates were tested for susceptibility to 15 antimicrobials using a broth microdilution technique following Clinical and Laboratory Standards Institute (CLSI) guidelines (24,25). The Sensititre NARMS Gram-negative minimum inhibitory concentration (MIC) plate (CMV1AGNF, Sensititre; TREK Diagnostic Systems, Westlake, Ohio, USA) was used.
Salmonella spp. isolates were tested for antimicrobial susceptibility as a part of a previously described study (28) in which the Sensititre NARMS Gram-negative MIC plate (CMV6CNDC, Sensititre; TREK Diagnostic Systems) was used. The antimicrobials tested and MIC breakpoints used for resistance for Salmonella spp. (26) and E. coli (27) are described elsewhere.
Data analysis
Antimicrobial susceptibility data were transferred to a spreadsheet (Microsoft Excel 2000; Microsoft Corporation, Redmond, Washington, USA) and reviewed for missing values, proper coding, and distribution of values. The analysis was limited to the 15 antimicrobials that were present in both susceptibility plates. Equal numbers of generic E. coli and Salmonella spp. isolates were selected from each fecal sample, to a maximum of 5 per sample. Isolates with intermediate susceptibility to the tested antimicrobials were considered “susceptible” for analysis purposes. Resistance to ≥ 2 antimicrobial classes was defined as a multidrug-resistant pattern.
A statistical software package (Intercooled Stata 9.1; Stata Corporation, College Station, Texas, USA) was used for the analyses. Isolate level prevalence was computed for each antimicrobial by dividing the total number of resistant isolates by the total number of isolates tested and 95% confidence intervals (CI) for binomial proportions were calculated using an exact approach. Only antimicrobials for which the frequency of resistance in both E. coli and Salmonella isolates was ≥ 5% were included in the analysis.
To assess the utility of AMR in E. coli as an indicator of AMR in Salmonella spp. in this study, the AMR prevalence among Salmonella spp. and E. coli from identical fecal samples was compared. For this, a 4-level random intercept logistic regression model was used with adaptive quadrature, exchangeable correlation structure, and 16 integration points using the Generalized Linear Latent and Mixed Model (GLLAMM procedure) (28–30). The outcome variable was resistance (yes/no) to individual antimicrobials and the predictor variable was a dichotomous variable indicating the bacterial species (E. coli = 1/Salmonella spp. = 0). In order to account for clustering, the pooled fecal samples, farm visits, and farms were included in the model as random effects.
The prevalence of AMR at the farm level was also calculated for both organisms. The farm-level status was considered to be resistant to the studied antimicrobial if at least one of the isolates from that farm was resistant. To evaluate the clustering effect at the pen pooled fecal sample level, estimates of variance components for the unexplained variation were calculated for each antimicrobial for both E. coli and Salmonella spp. Random intercept logistic regression models were generated using the aforementioned GLLAMM procedure. The dichotomous outcome variable was whether the isolate was resistant to a specific antimicrobial. Based on the estimated variance components for the unexplained variation at the sample and isolate levels of the model, intra-class correlation coefficients (ICC) were computed for correlation between samples by assuming that level 1 (isolates) variance on the logit scale was π2 ÷ 3 (30,31).
The association between resistance in E. coli and Salmonella spp. from the same sample was evaluated using a 4-level random intercept logistic regression model using the aforementioned procedures. To account for clustering, the samples, farm visits, and farms were included in the model as random effects. The dichotomous outcome variable represented the resistance (yes/no) of each Salmonella spp. isolate, and the predictor variable was a dichotomous variable representing the resistance (yes/no) of any E. coli isolate from the same sample. Odds ratios were calculated as a measure of the strength of association.
Results
Data from 1844 E. coli and Salmonella spp. isolates (922 isolates of each) were used in the analysis. The frequencies of resistance to 15 individual antimicrobials for E. coli and Salmonella spp. isolates are shown in Table I. Of the 922 E. coli isolates, 826 (89.6%) showed resistance to at least 1 antimicrobial, and 647 (70.2%) isolates were resistant to 2 or more antimicrobial classes. Of the 922 Salmonella spp. isolates, 408 (44.3%) showed resistance to at least 1 antimicrobial and 258 (28.0%) were resistant to 2 or more antimicrobial classes. The frequency of resistance at the isolate level to individual antimicrobials was in higher in E. coli than Salmonella, except for kanamycin (Table I). Higher ICC values were observed at the sample level for prevalence of resistance in Salmonella spp. (0.86–0.99) than in E. coli (0.33–0.57).
Table I.
Frequency of antimicrobial resistance in Salmonella spp. (n = 922) and generic E. coli (n = 922) recovered from 188 pen pooled fecal samples collected on 45 finishing swine farms in Alberta
| Antimicrobialsa,b (Breakpoints) | Percent of resistant E. coli isolates (number) | 95% CI | Percent of resistant Salmonella spp. isolates (number) | 95% CI |
|---|---|---|---|---|
| AMP (≥ 32) | 30.1 (276) | 27.1–33.2 | 6.1 (56) | 4.6–7.9 |
| CHL (≥ 32) | 11.2 (103) | 9.3–13.5 | 4.5 (41) | 3.2–6.0 |
| KAN (≥ 64) | 9.3 (85) | 7.5–11.3 | 14.4 (132) | 12.2–16.8 |
| STR (≥ 64) | 48.6 (446) | 45.4–51.9 | 26.5 (243) | 23.7–29.5 |
| FIS (≥ 512) | 40.0 (367) | 36.8–43.3 | 21.9 (201) | 19.3–24.7 |
| TET (≥ 16) | 79.4 (732) | 76.5–81.9 | 43.4 (400) | 39.9–46.4 |
| SXT (≥ 4) | 4.8 (44) | 3.5–6.3 | 0.2 (2) | 0.02–0.8 |
| GEN (≥ 16) | 0.1 (1) | 0.002–0.5 | 0.5 (5) | 0.1–1.2 |
| AUG (≥ 32) | 0.2 (2) | 0.02–0.8 | 0 | N/A |
| FOX (≥ 32) | 0.3 (3) | 0.06–0.9 | 0 | N/A |
AMP — ampicillin; CHL — chloramphenicol; KAN — kanamycin; STR — streptomycin; FIS — sulfisoxazole; TET — tetracycline; SXT — trimethoprim/sulfamethoxazole; GEN — gentamicin; AUG — amoxicillin/clavulanic acid; FOX — cefoxitin.
No E. coli isolates were resistant to amikacin, ceftiofur, ceftriaxone, ciprofloxacin, and nalidixic acid; and no Salmonella spp. isolates were resistant to amikacin, amoxicillin/clavulanic acid, cefoxitin, ceftiofur, ceftriaxone, ciprofloxacin, and nalidixic acid.
The differences in resistance prevalence among Salmonella spp. and E. coli from identical fecal samples, accounting for pooled sample, farm visit, and farm level clustering, is shown in Table II. At the isolate level, significantly higher resistance frequency was observed in E. coli compared to Salmonella for all tested antimicrobials where resistance was observed, except for kanamycin, where the opposite was observed (Table II). At the farm level, the frequency of AMR was significantly higher for E. coli than Salmonella for all antimicrobials (Figure 1). At least 1 E. coli isolate was resistant to sulfisoxazole on each farm, and almost all farms included at least 1 isolate resistant to tetracycline and streptomycin (Figure 1). The proportion of multi-drug-resistance (MDR) patterns among E. coli and Salmonella isolates is shown in Figure 2. On visual inspection, the proportion appears to decline as the number of drugs in a resistance pattern increased. The frequencies of the most common resistance patterns among the 922 E. coli and Salmonella isolates are shown in Table III.
Table II.
Association between bacterial species (E. coli compared with Salmonella spp.) and AMR prevalence after accounting for clustering by sample, visit, and farm for antimicrobials with ≥ 5% frequency of resistance using the generalized linear latent and mixed model (GLLAMM) procedure
| Differences between AMR prevalence in E. coli and Salmonella |
Variance estimates (% of total variance)
|
||||||
|---|---|---|---|---|---|---|---|
| Antimicrobialsa | Coefficient | ORb | 95% CI | P | Samples | Farm visits | Farms |
| AMP | 2.70 | 14.88 | 9.87–22.20 | < 0.001 | 1.04 (17.1) | 0.85 (14.0) | 1.65 (27.1) |
| CHL | 1.20 | 3.32 | 2.19–5.00 | < 0.001 | 0.85 (12.6) | 0.74 (10.9) | 1.89 (27.9) |
| KAN | −0.75 | 0.47 | 0.33–0.66 | < 0.001 | 0.93 (12.0) | 2.23 (28.7) | 1.32 (16.9) |
| STR | 1.28 | 3.60 | 2.84–4.54 | < 0.001 | 0.72 (13.7) | 0.22 (4.1) | 1.01 (19.2) |
| FIS | 1.30 | 3.66 | 2.86–4.66 | < 0.001 | 0.81 (14.9) | 0.40 (7.3) | 0.94 (17.3) |
| TET | 2.79 | 16.25 | 11.70–22.47 | < 0.001 | 1.46 (14.4) | 0.76 (7.5) | 4.63 (45.7) |
AMP — ampicillin; CHL — chloramphenicol; KAN — kanamycin; STR — streptomycin; FIS — sulfisoxazole; TET — tetracycline.
OR — odds ratio — the odds of antimicrobial resistance in E. coli compared to Salmonella isolates from the same sample.
Figure 1. Percentage of farms (N = 45) with at least one resistant isolate of E. coli and Salmonella.
AMP — ampicillin; CHL — chloramphenicol; KAN — kanamycin; STR — streptomycin; FIS — sulfisoxazole; TET — tetracycline
95% Binomial Confidence Interval E. coli — AMP (62.91, 88.80); CHL (27.66, 57.85); KAN (29.64, 60.00); STR (81.73, 98.60); FIS (92.13, 1.00); TET (84.85, 99.46)
95% Binomial Confidence Interval Salmonella — AMP (9.58, 34.60); CHL (9.58, 34.60); KAN (16.37, 44.31); STR (35.77, 66.30); FIS (31.66, 62.13); TET (44.33, 74.30)
Figure 2. Percentage of isolates with multi-drug resistance (resistance to ≥ 2 antimicrobial classes) for generic fecal E. coli (n = 922) and Salmonella spp. (n = 922) recovered from 188 pen pooled fecal samples from 45 finishing swine farms in Alberta.
AMR2–AMR6: numbers of antimicrobials in the resistance pattern
95% Binomial Confidence Interval E. coli — AMR2 (25.77, 31.73); AMR3 (23.77, 29.60); AMR4 (11.58. 16.14); AMR5 (0.68, 2.27); AMR6 (0.03, 0.79)
95% Binomial Confidence Interval Salmonella — AMR2 (11.07, 15.56); AMR3 (6.68, 10.38); AMR4 (2.95, 5.64); AMR5 (1.17, 3.08); AMR6 (0.12, 1.11)
Table III.
Frequencies of the most common multidrug-resistance (resistance to ≥ 2 anti-microbial classes) patterns among generic fecal E. coli and Salmonella spp. isolates obtained from 188 pen pooled fecal samples from 45 finishing swine farms in Alberta
| Multidrug-resistance pattern E. coli | Percent of AMR pattern (number) | Multidrug-resistance pattern Salmonella spp. | Percent of AMR pattern (number) |
|---|---|---|---|
| AMP-STR-FIS-TET | 6.4 (59) | KAN-STR-FIS-TET | 2.4 (22) |
| AMP-KAN-STR-TET | 2.3 (21) | STR-FIS-TET | 5.7 (52) |
| KAN-STR-FIS-TET | 1.4 (13) | KAN-STR-FIS | 1.5 (14) |
| CHL-KAN-STR-TET | 1.20 (11) | STR-FIS | 6.4 (59) |
| STR-FIS-TET | 9.9 (91) | STR-TET | 2.3 (21) |
| AMP-STR-TET | 7.2 (66) | AMP-STR | 1.8 (17) |
| AMP-FIS-TET | 3.6 (33) | KAN-TET | 1.3 (12) |
| CHL-STR-TET | 1.8 (17) | N/A | N/A |
| CHL-FIS-TET | 1.3 (12) | N/A | N/A |
| STR-TET | 10.8 (99) | N/A | N/A |
| FIS-TET | 7.7 (71) | N/A | N/A |
| AMP-TET | 5.4 (50) | N/A | N/A |
| STR-FIS | 1.5 (14) | N/A | N/A |
| CHL-TET | 1.3 (12) | N/A | N/A |
AMP — ampicillin; CHL — chloramphenicol; KAN — kanamycin; STR — streptomycin; FIS — sulfisoxazole; TET — tetracycline.
Although initial screening among univariate logistic regression models without random effects suggested potentially significant associations between generic E. coli and Salmonella resistance to certain antimicrobials, resistance in generic E. coli was not significantly associated at the sample level with resistance in Salmonella in any of the multilevel random effects logistic regression models (Table IV). Attempts were made to evaluate the association at the farm level, but the models did not converge due to little variability of AMR among Salmonella and E. coli at this level.
Table IV.
Associations between the resistance of E. coli and Salmonella spp. from simple logistic regression models and multilevel logistic regression models, accounting for clustering by sample, visit, and farm
| Simple logistic regression models
|
Multilevel logistic regression modelsb |
||||||||
|---|---|---|---|---|---|---|---|---|---|
| Antimicrobialsa | Coefficientb | 95% CI | OR | P | Coefficientb | 95% CI | OR | P | |
| AMP | 1.60 | 0.74, 2.46 | 4.95 | < 0.001 | — | — | — | — | |
| CHL | −1.13 | −2.08, −0.18 | 0.32 | 0.02 | −0.14 | −3.31, 3.04 | 0.87 | 0.93 | |
| KAN | 0.27 | −0.11, 0.66 | 1.31 | 0.17 | −1.84 | −5.43, 1.76 | 0.16 | 0.32 | |
| STR | 0.70 | 0.21, 1.19 | 2.01 | 0.005 | 0.43 | −2.56, 3.43 | 1.54 | 0.78 | |
| FIS | 0.55 | 0.10, 0.99 | 1.73 | 0.02 | −0.04 | −2.89, 2.82 | 0.96 | 0.98 | |
| TET | 1.54 | 0.67, 2.42 | 4.66 | < 0.001 | 1.15 | −5.78, 8.08 | 3.16 | 0.75 | |
AMP — ampicillin; CHL — chloramphenicol; KAN — kanamycin; STR — streptomycin; FIS — sulfisoxazole; TET — tetracycline.
Association between the resistance of E. coli and Salmonella spp.
Discussion
In this study, the odds of the prevalence of resistance were consistently and statistically significantly higher in E. coli than Salmonella for almost all antimicrobials except kanamycin, for which the opposite was observed, at least at the isolate level. This suggests that under the conditions of this study, E. coli may be more susceptible to antimicrobial selection pressures than Salmonella for many antimicrobials. If true, this could be a function of any number of agent, host, or environmental factors, such as, uptake of transmissible resistance determinants, duration of colonization in pigs, or persistence at the farm level. Other researchers (32) have suggested that E. coli might acquire resistance more easily and rapidly than Salmonella. In E. coli, higher prevalences of resistance were observed to ampicillin, tetracycline, streptomycin, and sulphisoxazole than other antimicrobials in the panel. The findings were similar in Salmonella, except that kanamycin resistance was also comparatively more prevalent. Antimicrobial selection pressure is one factor that may at least partially explain these findings. Several antimicrobials were reportedly used on study farms, and of these, penicillin and tetracycline were among the most frequently used antimicrobials individually or in combination with other antimicrobials, mostly through feed (33). The associations between AMU and AMR in E. coli on these farms are described in more detail elsewhere (29). Mathew et al (34) reported that in swine, AMR development in E. coli isolates might be more influenced by AMU than in Salmonella isolates. The relatively high number of observed multidrug-resistance patterns in E. coli supports the view that E. coli may be an important reservoir of resistance genes (15).
Under the conditions of this study, resistance to specific antimicrobials in E. coli isolates was not significantly associated with resistance to those antimicrobials in Salmonella at the isolate level in regression models. It is possible there was insufficient statistical power to detect small but meaningful associations, if they indeed existed. This is noteworthy, because despite the differences in AMR prevalence in E. coli and Salmonella discussed previously, there are remarkable overall similarities in the observed rank order of resistance prevalence in these bacteria (resistance to kanamycin is the only inconsistency in the top 7 for both bacteria as shown in Table I). Other researchers (34) have suggested that E. coli might not be a good indicator organism to represent the effects of selective pressure of antimicrobials on Salmonella, which is consistent with the findings herein at the sample level. It is still possible that associations may exist at higher levels of organization (barn, farm, or region levels); however, it was not possible to estimate these due to problems with model convergence. Future studies, therefore, should re-examine this issue at the farm and higher population levels, but to do so may require a larger sample of farms than was used in this study, in order to achieve reasonable statistical power.
Researchers (35,36), have demonstrated the transmission of genes conferring resistance among bacteria of the Enterobacteriaceae family, including both generic E. coli and Salmonella spp. If transmission of resistance genes was rapid and occurred regularly between these organisms, more similar resistance patterns in E. coli and Salmonella recovered from the same fecal samples would be expected. The findings in this study, however, suggest that such transfer does not occur frequently enough to detect similarities by the approach used. Further molecular epidemiological studies are warranted to better understand this phenomenon.
The high ICC values (range, 0.86–0.99) for prevalence of resistance among Salmonella spp. isolates within pen pooled fecal samples (correlation of resistance within samples was higher than between samples) might be explained either by a tendency for 1 clone to dominate within pens of pigs, or alternatively, perhaps the enrichment procedure enhanced the expansion or overgrowth of 1 clone, which resulted in 5 identical isolates for most of the samples (32). Multiple Salmonella isolates from each sample were used in order to match the number of E. coli isolates to maximize statistical power; however, to reduce laboratory costs in future studies and avoid unnecessary duplication, one Salmonella isolate per sample should be sufficient. If resources are available, testing fewer isolates from more samples may be beneficial from a statistical power standpoint. Moderate clustering in resistance observed in generic E. coli (range, 0.33–0.57) suggests that testing up to 3 isolates per sample might be sufficient to understand the prevalence and major trends in resistance.
In summary, the results suggest that E. coli is not a good indicator of Salmonella spp. resistance at the pooled finisher pig fecal sample level. Future studies and programs need to be carefully planned in order to balance the number of samples and the number of generic E. coli and Salmonella isolates per sample, to reduce unnecessary costs and to meet targeted research and surveillance objectives.
Acknowledgments
The authors thank the swine veterinarians and producers from Alberta who participated in this study, and the technical staff of the Agri-Food Laboratory Branch, Food Safety Division, Alberta Agriculture and Food (AAF) for laboratory support. Financial and in-kind support was provided by the Food Safety Division, AAF, Edmonton, Alberta, and the Laboratory for Foodborne Zoonoses, Public Health Agency of Canada, Guelph, Ontario.
References
- 1.Aarestrup FM. Veterinary drug usage and antimicrobial resistance in bacteria of animal origin. Basic Clin Pharmacol Toxicol. 2005;96:271–281. doi: 10.1111/j.1742-7843.2005.pto960401.x. [DOI] [PubMed] [Google Scholar]
- 2.Threlfall EJ, Ward LR, Frost JA, Willshaw GA. The emergence and spread of antibiotic resistance in food-borne bacteria. Int J Food Microbiol. 2000;62:1–5. doi: 10.1016/s0168-1605(00)00351-2. [DOI] [PubMed] [Google Scholar]
- 3.Anon . American Society for Microbiology. American Academy of Microbiology; Washington, DC: 1999. Antimicrobial Resistance: An Ecological Perspective. [PubMed] [Google Scholar]
- 4.World Health Organization Report on Infectious Diseases 2000. Overcoming Antimicrobial Resistance. World Health Organization (WHO/CDS/2000.2); Geneva, Switzerland: 2000. [Google Scholar]
- 5.Bager F. DANMAP: Monitoring antimicrobial resistance in Denmark. Int J Antimicrob Agents. 2000;14:271–274. doi: 10.1016/s0924-8579(00)00135-7. [DOI] [PubMed] [Google Scholar]
- 6.SVARM. Swedish Veterinary Antimicrobial Resistance Monitoring. National Veterinary Institute; Uppsala, Sweden: 2001. [Google Scholar]
- 7.Moreno MA, Dominguez L, Teshager T, Herrero IA, Porrero MC. Antibiotic resistance monitoring: The Spanish programme. The VAV Network. Red de Vigilancia de Resistencias Antibióticas en Bacterias de Origen Veterinario. Int J Antimicrob Agents. 2000;14:85–290. doi: 10.1016/s0924-8579(00)00138-2. [DOI] [PubMed] [Google Scholar]
- 8.United States Department of Agriculture, Agricultural Research Service. Annual Reports. Atlanta, Georgia: FDA/USDA/CDC; [Last accessed January 23, 2007]. National Antimicrobial Resistance Monitoring System. Annual Veterinary Isolates Data. [homepage on the Internet]. Available from http://www.ars.usda.gov/Main/docs.htm?docid=6750. [Google Scholar]
- 9.Government of Canada. Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) 2005. Guelph, Ontario: Public Health Agency of Canada, 2007; [Last accessed January 10, 2007]. [homepage on the Internet]. Available from http://www.phac-aspc.gc.ca/cipars-picra/pdf/cipars-picra-2005_e.pdf. [Google Scholar]
- 10.Kijima-Tanaka M, Ishihara K, Morioka A, et al. A national surveillance of antimicrobial resistance in Escherichia coli isolated from food-producing animals in Japan. J Antimicrob Chemother. 2003;51:447–451. doi: 10.1093/jac/dkg014. [DOI] [PubMed] [Google Scholar]
- 11.NORM/NORM-VET. Usage of Antimicrobial Agents and Occurrence of Antimicrobial Resistance in Norway. Tromso/Oslo; 2004. ISSN 1502–20307. [Google Scholar]
- 12.Aarestrup FM. Monitoring of antimicrobial resistance among food animals: Principles and limitations. J Vet Med B Infect Dis Vet Public Health. 2004;51:380–388. doi: 10.1111/j.1439-0450.2004.00775.x. [DOI] [PubMed] [Google Scholar]
- 13.World Health Organization. WHO global principles for the containment of antimicrobial resistance in animals intended for food. Report of a WHO consultation; 5–9 June; Geneva, Switzerland. 2000. [Google Scholar]
- 14.van den Bogaard AE, London N, Stobberingh EE. Antimicrobial resistance in pig faecal samples from the Netherlands (five abattoirs) and Sweden. J Antimicrob Chemother. 2000;45:663–671. doi: 10.1093/jac/45.5.663. [DOI] [PubMed] [Google Scholar]
- 15.Murray BE. Antibiotic resistance. Adv in Intern Med. 1997;42:339–367. [PubMed] [Google Scholar]
- 16.van den Bogaard AE, Stobberingh EE. Epidemiology of resistance to antibiotics. Links between animals and humans. Int J Antimicrob Agent. 2000;14:327–335. doi: 10.1016/s0924-8579(00)00145-x. [DOI] [PubMed] [Google Scholar]
- 17.Aarestrup FM, Bager F, Jensen NE, Madsen M, Meyling A, Wegener HC. Resistance to antimicrobial agents used for animal therapy in pathogenic, zoonotic, and indicator bacteria isolated from different food animals in Denmark: A baseline study for the Danish Integrated Antimicrobial Resistance Monitoring Programme (DANMAP) APMIS. 1998;106:745–770. doi: 10.1111/j.1699-0463.1998.tb00222.x. [DOI] [PubMed] [Google Scholar]
- 18.Maguire HC, Codd AA, Mackay VE, Rowe B, Mitchell E. A large outbreak of human salmonellosis traced to a local pig farm. Epidemiol Infect. 1993;110:239–246. doi: 10.1017/s0950268800068151. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Pontello M, Sodano L, Nastasi A, et al. A community based outbreak of Salmonella enterica serotype Typhimurium associated with salami consumption in Northern Italy. Epidemiol Infect. 1998;120:209–214. doi: 10.1017/s095026889800870x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Johnson JM, Rajić A, McMullen LM. Antimicrobial resistance of selected Salmonella isolates from food animals and food in Alberta. Can Vet J. 2005;6:141–146. [PMC free article] [PubMed] [Google Scholar]
- 21.Bywater RJ. Veterinary use of antimicrobials and emergence of resistance in zoonotic and sentinel bacteria in the EU. J Vet Med B Infect Dis Vet Public Health. 2004;51:361–363. doi: 10.1111/j.1439-0450.2004.00791.x. [DOI] [PubMed] [Google Scholar]
- 22.Rajić A, Keenliside J, McFall ME, et al. Longitudinal study of Salmonella species in 90 Alberta swine finishing farms. Vet Microbiol. 2005;105:47–56. doi: 10.1016/j.vetmic.2004.10.005. [DOI] [PubMed] [Google Scholar]
- 23.Koneman EW. Color atlas and textbook of diagnostic microbiology. 5. Philadelphia, Pennsylvania: Lippincott-Raven; 1997. p. 135. [Google Scholar]
- 24.CLSI. Performance standards for Antimicrobial Disk and Dilution Susceptibility Tests for Bacteria Isolated from Animals; Information Supplemenet. Updated tables for the NCCLS antimicrobial susceptibility testing standard. 2004:M31–A2. [Google Scholar]
- 25.CLSI. Performance standards for Antimicrobial Susceptibility Testing. 15th Informational Supplement. 2005:M100–S15.25. 102–106. [Google Scholar]
- 26.Rajić A, McFall M, Deckert A, et al. Antimicrobial resistance of Salmonella isolated from finishing swine and the environment of 60 Alberta swine farms. Vet Microbiol. 2004;104:190–196. doi: 10.1016/j.vetmic.2004.09.013. [DOI] [PubMed] [Google Scholar]
- 27.Varga C. Prevalence and risk factors for antimicrobial resistance in generic fecal Escherichia coli isolated from Alberta finishing swine [MSc thesis] Guelph, Ontario: University of Guelph; 2003. [Google Scholar]
- 28.Dohoo I, Martin W, Stryhn H. Veterinary Epidemiology Research. Vol. 337. AVC Inc. Charlottetown; Prince Edward Island: 2003. pp. 500–503. [Google Scholar]
- 29.Skrondal A, Rabe-Hesketh S. Generalized latent variable modeling: Multilevel, longitudinal and structural equation models. Boca Raton, Florida: CRC Press; 2004. [Google Scholar]
- 30.Rabe-Hesketh S, Skrondal A. Multilevel and Longitudinal Modeling using Stata. College Station, TX: Stata Press; 2005. [Google Scholar]
- 31.Snijders TAB, Bosker RJ. Multilevel Analysis: An Introduction to Basic and Advanced Multilevel Modelling. London: Sage; 1999. p. 266. [Google Scholar]
- 32.Mathew AG, Jackson F, Saxton AM. Effects of antibiotic regimens on resistance of Escherichia coli and Salmonella serovar Typhimurium in swine. J Swine Health Prod. 2002;10:7–13. [Google Scholar]
- 33.Rajić A, Reid-Smith R, Deckert AE, Dewey CE, McEwen SA. Reported antibiotic use in 90 swine farms in Alberta. Can Vet J. 2006;47:446–452. [PMC free article] [PubMed] [Google Scholar]
- 34.Mathew AG, Beckmann MA, Saxton AM. A comparison of antibiotic resistance in bacteria isolated from swine herds in which antibiotics were used or excluded. J Swine Health Prod. 2001;9:125–129. [Google Scholar]
- 35.Blake DP, Hillman K, Fenlon DR, Low JC. Transfer of antibiotic resistance between commensal and pathogenic members of the Enterobacteriaceae under ileal conditions. J Appl Microbiol. 2003;95:428–436. doi: 10.1046/j.1365-2672.2003.01988.x. [DOI] [PubMed] [Google Scholar]
- 36.Poppe C, Martin LC, Gyles CL, et al. Acquisition of resistance to extended-spectrum cephalosporins by Salmonella enterica subsp. enterica serovar Newport and Escherichia coli in the turkey poult intestinal tract. Appl Environ Microbiol. 2005;71:1184–1192. doi: 10.1128/AEM.71.3.1184-1192.2005. [DOI] [PMC free article] [PubMed] [Google Scholar]