Antimicrobial resistance of bovine Salmonella enterica ssp. enterica isolates from the Alberta Agriculture and Forestry Disease Investigation Program (2006–2014)

Abstract

The objectives of this study were to describe the antimicrobial susceptibility and serotypes of clinical Salmonella spp. isolates from Alberta cattle, to inform antimicrobial stewardship decisions for Alberta bovine veterinarians and to provide data for national surveillance. Isolates were collected from cattle and serotyped by Alberta Agriculture and Forestry from 2006 to 2014. Susceptibility testing was completed using Canadian surveillance breakpoints. There were 81 unique Salmonella isolates from 72 visits to 27 farms. The majority of isolates were S. Typhimurium (66.7%) and S. Dublin (19.8%). The prevalence of multidrug resistance was high in S. Typhimurium (89.1%) and S. Dublin (93.8%), including ceftiofur resistance (43.6% and 68.8%, respectively), while there was no resistance in other serotypes. As ceftiofur is a recommended treatment option for enteritis and septicemia caused by Salmonella in cattle, these results reinforce that obtaining bacterial culture and susceptibility results is critical for suspected cases of bovine salmonellosis in Alberta.

Introduction

Antimicrobial resistance (AMR) is an international, One Health concern that adversely impacts animal and human health. Bovine salmonellosis can have severe economic implications for producers due to calf mortality and loss of production, and its zoonotic potential is a public health concern (1). Parmley et al (2) suggested that antimicrobial resistance can limit treatment options for bovine and human Salmonella infections in Canada and internationally. It is imperative to develop a better understanding of the prevalence of AMR in clinical bovine Salmonella infections in Canada to identify effective treatment options and to provide guidance for antimicrobial stewardship decisions in cattle that reduce AMR selection pressure on foodborne Salmonella that may be transmitted to humans (3–5).

Bovine salmonellosis is caused by many serotypes of Salmonella enterica ssp. enterica. In Alberta Salmonella Typhimurium and Salmonella Dublin are particularly important in cattle due to their zoonotic potential and clinical impact in cattle herds. As a result, they are reportable diseases in cattle in Alberta under the Reportable and Notifiable Diseases Regulation of the Animal Health Act (6). Bovine salmonellosis has a morbidity of over 50% in calves, with a death rate reaching 100% without treatment (7). A study from 2003 in Alberta found that 4 of 50 dairy herds sampled had 1 or more cows shedding Salmonella (8).

In the past, ampicillin and trimethoprim-sulfamethoxazole have been the first line antimicrobial treatment for bovine salmonellosis (7). The most recent data for Alberta and Canada indicated that the most commonly identified resistance pattern for S. Typhimurium is to ampicillin, chloramphenicol, streptomycin, sulfamethoxazole, and tetracycline (ACSSuT), though resistance to ceftiofur, ceftriaxone, and/or trimethoprim-sulfamethoxazole has been identified (8–12). This coincides with findings from other parts of the world (5,13). As such, therapy is increasingly dependent on the use of extra-label dosage regimens of antimicrobials that are important to human medicine (14), namely third generation cephalosporins (e.g., ceftiofur) and fluoroquinolones (15). There is concern that resistance to ceftiofur, a common antimicrobial for bacterial infections in cattle, including enteritis, may limit treatment options in both cattle and humans (16–18). Resistance to third generation cephalosporins in zoonotic serotypes such as S. Typhimurium or S. Dublin is concerning as these drugs (e.g., ceftriaxone) are used to treat human infections (4,18,19).

The objectives of this study were to: i) describe the antimicrobial susceptibility and serotype profiles of Salmonella spp. isolates from Alberta cattle obtained through a provincial disease investigation program, representing a clinical population; ii) provide antimicrobial stewardship information to bovine veterinarians in Alberta; and iii) provide AMR surveillance information that complements national surveillance.

Materials and methods

Salmonella enterica ssp. enterica isolates collected from cattle operations from 2006 to 2014 under the Alberta Agriculture and Forestry (AF) Disease Investigation Program (DIP) were tested for antimicrobial susceptibility. Cattle operations with suspected salmonellosis must be reported to the Office of the Chief Provincial Veterinarian (OCPV). When S. Typhimurium or S. Dublin infections were confirmed or suspected, AF veterinarians worked with herd veterinarians to conduct on-farm investigations to assist in controlling the outbreak. Multiple fecal and/or environmental samples were taken during each farm visit and selectively cultured and tested for Salmonella spp. at AF’s Agri-Food Laboratory (AFL).

For each farm visit, AFL stored 1 Salmonella spp. isolate at −70°C that was previously cultured as part of the investigation. These isolates were originally serotyped by the Alberta Provincial Laboratory for Public Health. Any isolates with inconclusive results were submitted to the Public Health Agency of Canada’s National Microbiology Laboratory in Winnipeg, Manitoba for serotype confirmation. Stored isolates were reconstituted and tested for antimicrobial susceptibility. The susceptibilities of each serotype were tested in cases in which farms had multiple serotypes.

Susceptibility testing was conducted on these isolates for 14 antimicrobials using the CMV3AGNF plate for the Sensititre automated broth micro-dilution system (TREK Diagnostic Systems, Westlake, Ohio, USA) according to the manufacturer’s procedures. Clinical Laboratory Standards Institute (CLSI) volume M100-25 for Enterobacteriaceae (20) guidelines were used for interpretation of minimum inhibitory concentrations (MICs). Antimicrobials tested included: amoxicillin-clavulanate, ceftiofur, ceftriaxone, ciprofloxacin, ampicillin, azithromycin, cefoxitin, gentamicin, nalidixic acid, streptomycin, trimethoprim-sulfamethoxazole, chloramphenicol, sulfisoxazole, and tetracycline. This commercially available plate was chosen to align with susceptibility testing protocols for bovine Salmonella isolates from the Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS). This commercial plate includes chloramphenicol, even though this drug is no longer used in food animals, because it may represent resistance to florfenicol and allow detection of the ACSSuT pattern.

Antimicrobials that did not have CLSI MIC breakpoints were interpreted using the values selected by CIPARS, which are based on those used by the US National Antimicrobial Resistance Monitoring System (21). Salmonella spp. isolates were classified as susceptible, intermediate susceptibility or resistant based on these MIC breakpoints. Multidrug resistance (MDR) was defined as resistance to 3 or more drug classes. Data were analyzed in Excel (Microsoft Corporation, Redmond, Washington, USA) and STATA (Intercooled Stata 14; Stata Corporation, College Station, Texas, USA).

Results

There were 81 unique Salmonella spp. isolates obtained from 76 unique laboratory submissions from cattle operations in Alberta through the AF DIP from 2006 to 2014. These 76 submissions represent 72 visits to 27 farms. Of the isolates, 59.3% (48/81) were from 17 dairy farms, 24.7% (20/81) were from 3 feedlots, 9.9% (8/81) were from 3 cow-calf operations, and 6.2% (5/81) were from 3 veal calf grower operations. There were 14 Alberta veterinary clinics engaged in the investigations with 63 isolates coming from southern Alberta compared to 18 from central Alberta. The breakdown of serotypes by type of cattle operation is shown in Table 1. The most common serotypes were S. Typhimurium (67.9%, 55/81 isolates) and S. Dublin (19.8%, 16/81), with 9 other serotypes present as 1 or 2 isolates. The 55 S. Typhimurium isolates came from 16 farms (10 dairy farms, 3 cow-calf operations, 2 feedlots, and 1 veal calf grower). The 16 S. Dublin isolates came from 10 farms (7 dairy farms and 3 veal calf growers).

Table 1.

Salmonella enterica ssp. enterica serotypes isolated from cattle operations in Alberta from the Agriculture and Forestry Disease Investigation Program (2006–2014).

	Number of isolates
Salmonella serotype	Dairy	Feedlot	Cow-Calf	Calf grower	Total
Typhimurium	28	18	8	1	55
Dublin	13	0	0	3	16
Enteritidis	2	0	0	0	2
Kiambu	0	1	0	0	1
Mbandaka	1	0	0	0	1
Montevideo	0	1	0	0	1
Newport	0	0	0	1	1
Ohio	1	0	0	0	1
Schwarzengrund	1	0	0	0	1
Senftenberg	1	0	0	0	1
Uganda	1	0	0	0	1
Total isolates	48	20	8	5	81

There were 28 herd investigations for bovine Salmonella infection on 27 farms (1 feedlot had 2 investigations more than 2 y apart with different serotypes). Typically, each of the laboratory submissions represented a unique farm visit during which samples were collected. The number of farm visits per investigation, based on unique sampling dates, ranged from 1 to 12. There was 1 feedlot that was sampled on 12 different dates, resulting in 16 different submissions and all isolates were S. Typhimurium.

In samples from 5 separate visits to different farms there were 2 unique Salmonella serotypes identified at the same visit. In 4 of these, all dairy farms, 1 serotype was Typhimurium and the others were Senftenberg, Ohio, Mbandaka, or Enteritidis. The fifth was from a veal calf grower that yielded serotypes Dublin and Newport. Another veal calf grower had S. Dublin identified on the first visit and S. Typhimurium identified on a subsequent visit 9 mo later. Interestingly, the dairy farm that had Typhimurium and Senftenberg identified during 1 farm visit was under investigation for Dublin identified at the farm 3 mo earlier.

The AMR prevalence estimates and resistance patterns for the S. Typhimurium and S. Dublin isolates are shown in Tables 2 and 3. Resistance in S. Typhimurium and S. Dublin was detected in every drug class except for macrolides (azithromycin). There was no resistance detected to ciprofloxacin; however, 7.3% of S. Typhimurium isolates had intermediate susceptibility. These same S. Typhimurium isolates were resistant to nalidixic acid. There were moderate to high levels of resistance in S. Typhimurium and S. Dublin to the critically important (category I) beta-lactam antimicrobials ceftiofur, ceftriaxone, and amoxicillin-clavulanic acid. Of the S. Typhimurium isolates, 44% were resistant to all 3 of these drugs. The highest prevalence of resistance in S. Typhimurium was to tetracycline (89.1%) and sulfisoxazole (90.9%) compared to only 3.6% for trimethoprim-sulfamethoxazole (Table 2). Resistance in S. Dublin was highest for ceftiofur and ceftriaxone (68.8%), amoxicillin-clavulanic acid (87.5%), tetracycline (93.8%), and chloramphenicol (93.8%), while all isolates were susceptible to ciprofloxacin, nalidixic acid, azithromycin, sulfisoxazole, and trimethoprim-sulfamethoxazole (Table 2).

Table 2.

The prevalence of resistance to different antimicrobials in Salmonella Typhimurium and Salmonella Dublin isolates from cattle operations in Alberta from the Agriculture and Forestry Disease Investigation Program (2006–2014).

	Salmonella Typhimurium n = 55, proportion (95% CIc)			Salmonella Dublin n = 16, proportion (95% CIc)
Antimicrobiala	Sb	I	R	S	I	R
Id
AUG2	45.4 (32.5–59.0)	10.9 (4.9–22.7)	43.6 (30.9–57.3)	12.5 (2.6–43.0)		87.5 (57.0–97.4)
XNL	56.3 (42.7–69.1)		43.6 (30.9–57.3)	12.5 (2.6–43.0)	18.8 (5.3–48.6)	68.8 (40.2–87.8)
AXO	56.3 (42.7–69.1)		43.6 (30.9–57.3)	12.5 (2.6–43.0)	18.8 (5.3–48.6)	68.8 (40.2–87.8)
CIP	92.7 (81.7–97.3)	7.3 (2.7–18.3)		100
II
AMP	10.9 (4.9–22.7)		89.1 (77.3–95.1)	12.5 (2.6–43.0)		87.5 (57.0–97.4)
AZI	100			100
FOX	56.3 (42.7–69.1)		43.6 (30.9–57.3)	25.0 (8.6–54.3)	12.5 (2.6–43.0)	62.5 (34.8–83.9)
GEN	100			68.8 (40.2–87.8)		31.2 (12.2–59.8)
NAL	92.7 (81.7–97.3)		7.3 (2.7–18.3)	100
STR	30.9 (19.8–44.7)		69.0 (55.3–80.1)	37.5 (16.1–65.2)		62.5 (34.8–83.9)
SXT	96.4 (86.1–99.1)		3.6 (1.0–13.9)	100
III
CHL	60.0 (46.2–72.4)		40.0 (27.6–53.8)	0.06 (1.0–39.3)		93.8 (60.7–99.3)
FIS	9.1 (3.7–20.5)		90.9 (79.5–96.3)	100
TET	10.9 (4.9–22.7)		89.1 (77.3–95.1)	0.06 (1.0–39.3)		93.8 (60.7–99.3)

^aAUG2 — amoxicillin/clavulanic acid; XNL — ceftiofur; AXO — ceftriaxone; CIP — ciprofloxacin; AMP — ampicillin; AZI — azithromycin; FOX — cefoxitin; GEN — gentamicin; NAL — nalidixic acid; STR — streptomycin; SXT — trimethoprim/sulfamethoxazole; CHL — choramphenicol; FIS — sulfisoxazole; TET — tetracycline.

^bS — susceptible; I — intermediate susceptibility; R — resistant; Assigned using Canadian Surveillance Breakpoints (21).

^c95% CI — confidence interval.

^dAntimicrobials are grouped by their importance to human medicine as categorized by Health Canada Veterinary Drugs Directorate (14).

Table 3.

Antimicrobial susceptibility patterns and prevalence estimates of Salmonella Typhimurium and Salmonella Dublin isolates from cattle operations in Alberta from the Agriculture and Forestry Disease Investigation Program (2006–2014).

Antimicrobiala susceptibility pattern	Frequency	Proportion	95% CIb
Salmonella Typhimurium isolates (n = 55) — *MDRc 89.1% (95% CIb 77.3 to 95.1)
AUG2-XNL-AXO-AMP-FOX-NAL-STR-CHL-FIS-TET*	3	5.4	(1.7–16.1)
AUG2-XNL-AXO-AMP-FOX-STR-CHL-FIS-TET*	18	32.7	(21.4–46.5)
AUG2-XNL-AXO-AMP-FOX-STR-FIS-TET*	1	1.8	(0.2–12.5)
AUG2-XNL-AXO-AMP-FOX-FIS-TET*	2	3.6	(0.8–13.9)
AMP-NAL-STR-CHL-FIS-TET*	1	1.8	(0.2–12.5)
AMP-STR-SXT-FIS-TET*	1	1.8	(0.2–12.5)
AMP-SXT-FIS-TET*	1	1.8	(0.2–12.5)
AMP-STR-FIS-TET*	13	24.6	(14.0–37.0)
AMP-FIS-TET*	9	16.4	(8.6–29.0)
STR-FIS	1	1.8	(0.2–12.5)
Pan-susceptible	5	9.1	(3.7–20.5)
Salmonella Dublin isolates (n = 16) — *MDRc 93.8% (95% CIb 60.7 to 99.3)
AUG2-XNL-AXO-AMP-FOX-STR-CHL-FIS-TET*	5	31.3	(12.2–59.8)
AUG2-XNL-AXO-AMP-FOX-GEN-CHL-FIS-TET*	5	31.3	(12.2–59.8)
AUG2-XNL-AXO-AMP-STR-CHL-FIS-TET*	1	6.3	(0.7–39.3)
AUG2-AMP-STR-CHL-FIS-TET*	3	18.8	(5.3–48.6)
STR-CHL-FIS-TET*	1	6.3	(0.7–39.3)
FIS	1	6.3	(0.7–39.3)

^aAUG2 — amoxicillin/clavulanic acid; XNL — ceftiofur; AXO — ceftriaxone; AMP — ampicillin; FOX — cefoxitin; GEN — gentamicin; NAL — nalidixic acid; STR — streptomycin; SXT — trimethoprim/sulfamethoxazole; CHL — choramphenicol; FIS — sulfisoxazole; TET — tetracycline. Resistance determined using Canadian surveillance breakpoints.

^b95% CI — 95% confidence interval.

^cMDR — multi-drug resistance (resistance to ≥ 3 drug classes), denoted with *.

Table 3 shows the prevalence of resistance patterns and MDR in S. Typhimurium and S. Dublin isolates. The prevalence of MDR was high for both S. Typhimurium (89.1%, 49/55 isolates) and S. Dublin (93.8%, 15/16), with most isolates displaying resistance to 4 or more drug classes. There were no S. Dublin isolates and only 4 S. Typhimurium isolates that were susceptible to all antimicrobials. In contrast, all 10 isolates of other serotypes were pan-susceptible. Of these 10 isolates, only 2 (Senftenberg and Enteritidis) had intermediate susceptibility to chloramphenicol.

There were 11 resistance patterns for S. Typhimurium (Table 3). The ACSSuT resistance pattern was present in 40.0% (22/55) of isolates. Three of these isolates were resistant to 6 drug classes. Of these ACSSuT-resistant isolates, 95.5% (21/22) were also resistant to amoxicillin-clavulanic acid, cefoxitin, ceftiofur, and ceftriaxone (AUG2-FOX-XNL-AXO) and 14.3% (3/21) were resistant to nalidixic acid. There were 6 resistance patterns for S. Dublin (Table 3). The ACSSuT resistance pattern was present in 56.3% (9/16) of isolates and 87.5% (14/16) were resistant to 5 drug classes. Five of the nine ACSSuT-resistant isolates were also resistant to AUG2-FOX-XNL-AXO.

Discussion

This study describes the antimicrobial susceptibility and serotype profiles of clinical Salmonella spp. isolates from bovine operations in Alberta from 2006 to 2014 obtained through the provincial DIP. In Alberta, Salmonella infection in cattle was typically only diagnosed when there was a clinical presentation of disease as there was no ongoing surveillance in healthy cattle until the Alberta sentinel site for FoodNet Canada began sampling feedlots in 2016 (data not yet available). Antimicrobial resistance was detected in the S. Typhimurium and S. Dublin isolates from Alberta cattle in this study, including resistance to third generation cephalosporins (ceftiofur/ceftriaxone), tetracycline, and chloramphenicol. Though controversial, case management of bovine salmonellosis often includes antimicrobial therapy for sick calves and sometimes cows. This is in part due to the lack of a definitive etiologic diagnosis at the beginning of an outbreak and the need for rapid therapy for animals with severe diarrhea, dehydration, shock, and septicemia before a definitive bacterial culture result can be obtained (7).

The Canadian Veterinary Medical Association (CVMA) Antimicrobial Prudent Use Guidelines (2008) recommend various antimicrobial treatment options for different ages and production types of cattle for enteritis caused by Salmonella spp. or Escherichia coli (15). All of the recommended antimicrobials or the dose regimens for bovine salmonellosis in Canada represent extra-label drug use (15). Canadian drugs with label claims for bovine salmonellosis tend to be injectable trimethoprim-sulfas or oral tetracycline antimicrobials, but the recommended doses of the former for calves with septicemia are still extra-label. The CVMA guidelines recommend either trimethoprim-sulfadoxine or ceftiofur for neonatal beef and dairy calves or adult dairy cattle with acute enteritis caused by Salmonella spp. (15). They provide an additional treatment option of danofloxacin or enrofloxacin (fluoroquinolones) for beef calves. Though chloramphenicol is banned for use in food animals, resistance is a concern because at least 1 of the possible genetic mechanisms for resistance, the floR gene, also confers resistance to florfenicol in the same drug class (22,23). Florfenicol resistance and resistance determinants were not evaluated in this study. Florfenicol is not included in the CVMA guidelines for treatment of Salmonella, but clinical experience of the authors suggests that it is still used to treat undifferentiated enteritis in newborn calves in Alberta. Two recent US studies in Wisconsin (24) and Minnesota (25) reported florfenicol resistance in veterinary diagnostic laboratory collections of bovine Salmonella spp. isolates from 2006 to 2015. In both studies, S. Dublin was the most common serotype, with florfenicol resistance ranging from 96% to 100% in the S. Dublin isolates (Wisconsin), and 37% to 55% in all Salmonella isolates (Minnesota). We are not aware of data on whether newer generation macrolides (e.g., tulathromycin) are used to treat bovine salmonellosis in Alberta, but the Wisconsin study speculates that it is commonly used in the dairy industry in that state (24).

The study herein indicates that ceftiofur or florfenicol may not be the best choices for empirical treatment of suspected bovine Salmonella in Alberta prior to obtaining culture and susceptibility results. Given that S. Typhimurium and S. Dublin are the most common serotypes detected, the potential for resistance to ceftiofur and possibly florfenicol is a cause for concern. The frequency of multidrug resistance in S. Typhimurium and S. Dublin, especially isolates resistant to both ceftiofur and chloramphenicol, makes empirical decisions for antimicrobial therapy even more difficult. The fact that this isolate collection is largely susceptible to trimethoprim-sulfamethoxazole suggests that this may still be a good empirical treatment choice. However, these findings reinforce the need to obtain a definitive etiologic diagnosis that includes a bacterial culture and an antimicrobial susceptibility profile to make appropriate clinical and antimicrobial stewardship treatment decisions for calves with enteritis or septicemia as the etiology and antimicrobial susceptibility can be unpredictable (3,4).

The treatment recommendations from the CVMA guidelines represent only 1 source of information available to veterinarians to make clinical treatment decisions. Therefore, they do not necessarily represent what is happening in practice. There is no current, ongoing antimicrobial use surveillance system that tracks antimicrobial use data at the veterinary practice or producer level for beef and dairy cattle in Canada. Antimicrobial use data have been collected by individual veterinary practices for use on their client’s operations (most commonly feedlot practices), sometimes in collaboration with CIPARS as part of pilot projects (26). However, the authors on this paper can confirm from veterinary field experience that antimicrobials such as ceftiofur or florfenicol are frequent first choice drugs for treatment of undifferentiated enteritis or septicemia in newborn calves, with fluoroquinolones (e.g., enrofloxacin) being used in more difficult cases. One author led an investigation that provided isolates to this study, during which there were anecdotal reports of enrofloxacin use for enteritis caused by Salmonella in newborn beef and dairy calves as recommended by practicing veterinarians in the area. It is also important to note that the CVMA Guidelines were published in 2008 and the recommended treatments and available drug formulations may have changed in the past decade. The CVMA is currently in the process of updating these guidelines for food and companion animals (27).

Frequencies of resistance in this collection of bovine Salmonella isolates from Alberta are similar to those reported in other jurisdictions in that resistance is most common to tetracycline, streptomycin, and sulfisoxazole (28). The presence of resistance to third-generation cephalosporins in our study mirrors that from the US from national surveillance (18) and the veterinary diagnostic collections in Wisconsin and Minnesota (24,25). Rao et al (10) found that resistance to tetracycline, streptomycin, and sulfisoxazole was common in Salmonella isolates from Alberta feedlot cattle that were collected in 2004, but there were only 20 Salmonella isolates out of 2100 samples. They detected ceftiofur resistance (5/20), but all isolates were susceptible to ciprofloxacin. They also found moderate levels (8/20 isolates) of penta-drug resistance (ACSSuT), sometimes in combination with resistance to amoxicillin-clavulanate, cefoxitin, ceftiofur, and/or trimethoprim-sulfa (10). The ACSSuT resistance pattern is commonly reported in the literature for Salmonella from cattle in Canada (8–12,23,29). Most recently, an Alberta feedlot study conducted from 2007 to 2010 found only 2 Salmonella isolates in 933 composite pens samples, making susceptibility testing moot (26).

Older data from Alberta (1996 to 1999) showed similar resistance profiles for Salmonella isolates from food animals and food, including 42 isolates from cattle (11). Two other studies from Alberta dairy cattle (1999) (8) and beef cattle (1998 to 1999) (12) found resistance, but at lower frequencies. Based on the MIC breakpoints used in their study, Johnson et al (11) found 1.0% resistance to ceftiofur in 209 Salmonella isolates, with no resistance to ceftriaxone or ciprofloxacin, but it was not clear whether these isolates came from cattle or other food animals or food. Sorenson et al (12) did not find any third-generation cephalosporin or fluoroquinolone resistance in beef cattle Salmonella isolates. It was not clear from the dairy cattle study whether or not susceptibility to these antimicrobials was tested, but no resistance was reported (8).

Salmonella Typhimurium and S. Dublin dominated the serotype distribution of this clinical isolate collection. This is not surprising in Alberta and mirrors clinical isolate collections in Wisconsin and Minnesota (24,25). It is not known if laboratory methods have the ability to preferentially detect certain serotypes. It is uncertain as to why isolates from dairy farms are more frequent in this collection from Alberta. Dairies and veal calf grower operations were the only operation types that identified S. Dublin. Veal calf growers have a high throughput of calves sourced from dairy farms in western Canada. These isolates may represent Salmonella infections from dairy farms in the region. There were isolates from only 3 feedlots in this collection. The majority of feedlot isolates (18/20) were Typhimurium, with 16 coming from 1 feedlot over a 100-day period. The MICs varied by isolate for some antimicrobials, but the resistance phenotypes based on breakpoints were the same for all 16 isolates and 14 antimicrobials.

Previously reported data from Alberta assessed the serotype distribution of livestock Salmonella isolates from passive surveillance databases held by AF from 1990 to 2001 (30). In bovine isolates, S. Typhimurium and S. Dublin were the most common, but there was a higher frequency of other serotypes not identified in this study. This difference may in part be explained by the fact that from 1990 to 1998, AF offered routine diagnostic services to cattle producers and veterinarians. The differences could also be explained by different source populations in that the 1990 to 2001 study represented incidental findings from a passive surveillance database compared to our isolates from investigations of clinical disease outbreaks. Recent surveillance projects in healthy Canadian cattle tend to detect Salmonella species at low frequencies, particularly in feedlot cattle. In 2004, an Alberta study found Salmonella spp. in only 1% (20/2100) of cattle manure samples from 21 feedlots of greater than 5000 head (31). Another Alberta study collected 933 composite pen samples from 4 feedlots from 2007 to 2010 and found only 2 to contain Salmonella from 2 separate feedlots (26). A 2013 study that integrated Salmonella data from multiple Canadian surveillance programs reported prevalence results of 10% (45/432) for cow-calf farms and 12% (80/643) for dairy farms in the C-EnterNet (now FoodNet Canada) sentinel surveillance site in Ontario between 2005 to 2010 (2). The FoodNet Canada sentinel site in Alberta started collecting samples from feedlot operations in 2016, but data are not yet available.

These Canadian studies are in stark contrast to data reported from the United States. The US National Animal Health Monitoring System Dairy study found 30.9% and 39.7% of dairy operations were positive for Salmonella in 2002 and 2007, with 13.8% of animals being positive in 2007 (32). A study in US feedlot cattle found that 64.7% of animals were positive for Salmonella on arrival at the feedlot (33). The serotype distributions in these 2 US studies were much different than in this isolate collection. Comparatively, S. Dublin was the most common serotype in clinical bovine Salmonella isolates (23% and 31.8%, respectively) in the 2 US veterinary diagnostic studies in Wisconsin (24) and Minnesota (25); detection of S. Typhimurium was much less common. The bovine isolates in both studies came predominantly from dairy operations in these States. It is not known why the prevalence of Salmonella is higher in US dairy and feedlot cattle compared with Canadian cattle.

Despite the relatively low prevalence of Salmonella in Canadian cattle from available data, the clinical impact of bovine salmonellosis can be severe in an individual herd experiencing an outbreak (7). Clinical infections in cow-calf and dairy operations tend to target newborn calves, resulting in severe diarrhea and septicemia, often with high mortality. Pregnant cows can experience abortion, while lactating dairy cattle may experience decreased milk production and enteritis (34). These are the typical cases that are reported to the Alberta OCPV. Management of such cases has 4 main objectives: to minimize the spread of Salmonella within the herd, to prevent transmission to farmers and their families, to prevent spread off the farm, and to assist the herd veterinarian in managing the clinical outbreak.

It is important to note that the isolate collection in this study is likely not representative of the cattle population of Alberta. It is more representative of a clinical case population of cattle herds experiencing outbreaks of salmonellosis in the province. Isolates from the same farm on different visit dates were included in the collection, representing clustering of isolates at the farm, veterinary clinics, and sometimes farm visit levels. There were only 3 feedlots included in this collection. Farms with resistant isolates may be more likely to be reported to the OCPV, to be under investigation, and to influence these resistance data. Farms with resistant strains may be more likely to experience antimicrobial treatment failure, translating to higher morbidity and mortality, increased likelihood of contacting a veterinarian, and ultimately seeking a bacterial culture. Lastly, there is selection bias in the isolate collection towards S. Typhimurium and S. Dublin in the province (6). Due to the small sample numbers of isolates and farms, it was not possible to evaluate differences in resistance by cattle operation type. It is also important to note that there were no linked farm data from veterinarians or producers for antimicrobial use in these investigations. Such data are important to interpret subsequent resistance profiles to understand antimicrobial selection pressure (35).

The findings from our study provide valuable data to guide antimicrobial stewardship for cattle producers and veterinarians by providing information on the extent of antimicrobial resistance from clinical cases of bovine salmonellosis in Alberta. These findings reinforce that definitive etiologic diagnosis including a bacterial culture and antimicrobial susceptibility profile must be sought when Salmonella is suspected, as the susceptibility profile may vary widely depending on the serotype. It is desirable to collect and submit samples for bacterial culture and susceptibility at the outset of an outbreak to ensure that antimicrobial therapy is appropriate (3,4). Findings from our study reinforce the importance of considering bovine Salmonella within national antimicrobial resistance surveillance, despite the low prevalence in recent feedlot studies (26). More importantly, it brings into question the current recommended antimicrobial treatments for bovine enteritis and septicemia from the CVMA (15). Data from our study should be considered during the update of these guidelines. This isolate collection demonstrates that older drugs of relatively lower importance to human medicine may still be useful to treat some cases of bovine salmonellosis in Alberta. Veterinarians should be wary of selecting drugs of high importance to human medicine, such as ceftiofur, as first line therapy.