Antimicrobial resistance and recovery of Salmonella, Campylobacter, and Escherichia coli from chicken egg layer flocks in Canadian sentinel surveillance sites using 2 types of sample matrices

Abstract

Eggs are important to the diet of Canadians. This product is one of the supply-managed commodities in Canada, but unlike other commodities, where food safety risks are extensively explored and reported, information on the prevalence of enteric organisms (e.g., Salmonella, Campylobacter) and antimicrobial resistance (AMR) in layers in Canada are limited. This study was conducted to determine the prevalence of select bacteria and the associated AMR patterns in layer flocks using 2 sample matrices. Farms were located within FoodNet Canada and the Canadian Integrated Program for Antimicrobial Resistance Surveillance sentinel sites (SS). Fecal samples (Ontario: ON_SS1a, ON_SS1b) and environmental sponge swabs (British Columbia: BC_SS2a) were collected. Salmonella prevalence was 29% and 8% in ON_SS1a and ON_SS1b, respectively, and 7% in BC_SS2a. S. Kentucky and S. Livingstone were the most frequently isolated serovars and no S. Enteritidis was detected. Campylobacter was not detected in the BC sponge swabs but was isolated from 89% and 53% of Ontario fecal samples (ON_SS1a and ON_SS1b, respectively). Seven C. jejuni from Ontario were ciprofloxacin-resistant. Escherichia coli prevalence was high in both sample types (98%). Overall, tetracycline resistance among E. coli ranged from 26% to 69%. Resistance to ceftiofur (n = 2 isolates) and gentamicin (n = 2) was relatively low. There were diverse resistance patterns (excludes susceptible isolates) observed among E. coli in Ontario (10 patterns) and British Columbia (14 patterns). This study revealed that fecal samples are more informative for farm-level monitoring of pathogen and AMR prevalence. Without further validation, sponge swabs are limited in their utility for Campylobacter detection and thus, for public health surveillance.

Introduction

In Canada, the per capita consumption of eggs in 2018 was 21 dozen per person, indicating that this commodity is important to the diet of Canadians (1). Historically, in Canada, the consumption of ungraded, Grade B, or undercooked eggs and the handling of raw eggs have been identified as risk factors for salmonellosis in humans (2–4). An industry-led initiative, Start Clean-Stay Clean (SC-SC) is the Canadian egg industry’s on-farm food safety program designed to ensure the quality and safety of table eggs and egg products. It has a protocol for the reduction of Salmonella prevalence, in particular, S. Enteritidis, which is a pathogen associated with egg-related salmonellosis outbreaks (5,6). However, there are no publicly accessible national surveillance programs showing trends in prevalence of Salmonella, Campylobacter, and other pathogens and their antimicrobial resistance (AMR) profiles in laying hens or table eggs.

Campylobacter (447.23 cases per 100 000 population) and non-typhoidal Salmonella (269.26 cases per 100 000 population) ranked 3rd and 4th, respectively, among the domestically acquired foodborne illnesses in Canada (7). Within the Public Health Agency of Canada, FoodNet Canada (FNC) and Canadian Integrated Program for Antimicrobial Resistance Surveillance (CIPARS) conduct surveillance of foodborne bacteria (e.g., Salmonella, Escherichia coli, and Campylobacter), from farm to fork. The FNC program is a multi-partner, sentinel-based comprehensive surveillance program that collects data in multiple sectors (e.g., agriculture and human health units) for the effective evaluation and development of policies for the safety of food and water in Canada (8). The CIPARS program monitors trends in AMR and antimicrobial use (AMU) in humans and animals and has active farm AMU and AMR surveillance programs in broiler chickens, turkeys, swine, and beef cattle (9). The collection of AMU information at the farm level complements AMR data, but there is a knowledge gap in terms of AMU in the layer sector in Canada. Although, we hypothesize that AMU is relatively low in the major egg-producing provinces such as Ontario (10) and Quebec (11), due to the relatively stable number of diagnoses of reproductive disorders in layers that have bacterial etiology, including bacterial peritonitis or salpingitis, and early systemic bacterial infections.

The CIPARS program previously conducted 2 research studies in the layer sector to determine Salmonella baseline prevalence and explore potential areas in the egg production chain where routine national monitoring of Salmonella and AMR could be implemented (12). The first study involved cecal sample collection from spent hens at the processing plant level (2009 to 2011) in Ontario, where 42% (117/279) of samples tested Salmonella-positive; of these 117 isolates, 15 S. Enteritidis were identified. The second study involved sampling liquid whole eggs at the Ontario egg-breaking stations, where 2% (5/300) of samples tested S. Enteritidis-positive (12). Similarly, a research project in Alberta layer flocks detected Salmonella in 57% and 60% of samples from barns and flocks, respectively (13).

Farm-level monitoring programs in layers such as Canada’s SC-SC program and the United States Food and Drug Administration’s 21 CFR Parts 16 and 118 “Prevention of Salmonella Enteritidis in Shell Eggs During Production, Storage and Transportation; Final Rule” (14) aim to eliminate S. Enteritidis to offset the burden of illness associated with the consumption of egg and egg products, an important dietary source of protein in North America. In Canada, the annual incidence rate for endemic S. Enteritidis cases has remained stable, between 8.6 and 8.3 cases per 100 000 population between 2014 and 2017 (8). While human cases attributable to chicken meat, chicken-derived products, and manure have been extensively investigated by both CIPARS and FNC, data related to the egg layer sector is sparse. Globally, up to 24% of non-typhoidal Salmonella spp. cases have been attributed to the consumption of eggs (16). Similarly, for Campylobacter, up to 4.5% of human cases have been attributed to the consumption of eggs (16,17). The public health significance described above and the potential co-occurrence of these 2 major foodborne pathogens in layer flocks (18) emphasizes the importance of including the egg sector in comprehensive surveillance programs with One Health themes such as CIPARS and FNC.

The SC-SC program utilizes environmental sponge swabs (5), but the collection of manure in addition to egg belt swabs or dusts were cited as the preferred sample type in the United States Final Rule (14), as informed by routine monitoring protocols (15), and in the European Union member states (18), to enhance the diversity of the sample matrix for improved bacterial recovery. In the context of the global and Canadian public health implications of Salmonella and Campylobacter described earlier, and in an effort to align with international surveillance programs for pathogen recovery and AMR, the objectives of this study were to i) evaluate the utility of 2 different sample matrices [fecal droppings (i.e., standard farm samples used by FNC and CIPARS) and environmental sponge swabs (i.e., SC-SC program)] for the simultaneous detection of Salmonella, Campylobacter, and E. coli (the latter as an indicator organism routinely used for AMR monitoring); and ii) describe the resistance of these bacteria to antimicrobials used in veterinary and human medicine. This study will inform the development of surveillance protocols for layer AMR and pathogen prevalence and explore the utility of an existing monitoring platform, the SC-SC program, as a mechanism for collecting AMU data and samples for pathogen recovery and AMR testing.

Materials and methods

To complement Salmonella prevalence and AMR information previously collected by CIPARS and FNC, a farm-level pilot study in layers was conducted from sentinel sites: i) Ontario sentinel site 1a (ON_SS1a) in 2013; ii) British Columbia sentinel site 2a (BC_SS2a) between 2013 and 2014; and iii) Ontario sentinel site 1b (ON_SS1b) between 2016 and 2017. The sampling of layer flocks was part of the larger FNC sentinel site program, which has human, environment/water, and animal commodity components (8). The priority animal commodities sampled were based on the agricultural profile of the province where the sentinel site is located. Layer flocks constituted the agricultural profile of the FNC sentinel sites above. With the support of the 2 provincial egg-marketing boards, a notification was sent to all producers explaining the study and encouraging them to support the pilot study. Participation to the study was voluntary.

The farm sampling methods are summarized in Table I. All samples were collected during the laying phase that coincided with the sampling/flock audit schedule according to the Egg Farmers of Canada’s SC-SC program (5). The SC-SC program requires 2 sampling visits during the laying phase: 1 early lay (19 to 35 wk) and 1 late lay (36 to 60 wk).

Table I.

Summary of farm sampling protocol and microbiological methods in 2 sample matrices.

	Ontario sentinel sites (ONSS1a and ONSS1b)	British Columbia sentinel site (BCSS2a)
Sample matrix	Fecal samples	Environmental sponge swabs
Farm selection and protocol	FNC/CIPARS farm protocol, > 20 wk of age	EFC SC-SC program, > 20 wk of age
Sampling methods	Fecal samples, 1 pooled sample taken from 1 manure belt or any accessible areas	Specified surfaces (egg conveyor belts, manure belts, dusts)
Total samples	4 pooled fecal samples	Pooled swabs from 1 to 7 individual swabs, depending on farm size
Sample preparation	Fecal sample + BPW (1:10 ratio), incubated at 35°C ± 1°C for 24 h	Pooled sponges + 100 mL BPW, incubated at 35°C ± 1°C for 24 h
Salmonella	0.1 mL BPW mixture onto MSRV; 42°C ± 1°C for 24 to 72 h incubation Migration ≥ 20 mm were streaked onto MacConkey agar Further testing: TSI and urea agar slants inoculation, indole test (negative), slide agglutination with Salmonella Poly A-I and Vi antiserum All isolates sent to OIE Salmonella Reference Laboratory at Guelph	1 mL BPW to 9.9 mL RVS broth; 42°C ± 2°C for 20 to 24 h incubation Streaked to XLT4 and BGA; 35°C ± 2°C for 18 to 48 h incubation Subculture: 4 Salmonella-like colonies to BAP; 35°C ± 2°C for 18 to 48 h incubation Further testing: indole (neg), oxidase (pos), Gram stain; confirmed one suspect colony using API 20E followed by serogrouping 2 isolates stored in cryovials (−80°C) sent to OIE Salmonella Reference Laboratory at Guelph
	OIE Salmonella Reference laboratory (all isolates): Detection of O or somatic antigens via slide agglutination; H or flagellar antigens identified with a microtitre plate well precipitation method (9).
Escherichia coli	0.1 mL of the BPW mixture onto MacConkey agar; 35°C ± 1°C for 18 to 24 h incubation Lactose-fermenting colonies were screened for purity (Luria-Bertani agar); Presumptive E. coli colonies were screened for purity (TSA with 5% sheep blood) Further testing: Simmons citrate and indole tests; confirmed using a bacterial identification test kit (API 20E)	0.1 mL of the BPW mixture onto MacConkey agar; 35°C ± 1°C for 18 to 24 h incubation 2 lactose-fermenting colonies from MacConkey were subcultured in BAP Stored in cryovials (−80°C) and sent to AMR Laboratory, NML at Guelph
Campylobacter	1 mL BPW mixture mixed with 9 mL of HEB; 35°C ± 1°C for 4 h microaerophilic incubation 36 μL sterile cefoperazone added to HEB tubes; 42°C ± 1°C for 20 to 24 h microaerophilic incubation Culture inoculated onto mCCDA plate; 42°C ± 1°C for 24 to 72 h microaerophilic incubation Further testing: Gram stain, oxidase, and catalase	1 mL BPW mixed with 9 mL Enriched Bolton Broth; 35°C ± 2°C for 4 h then 42°C ± 2°C overnight incubation Streaked onto CCDA; 42°C ± 2°C for 48 h microaerophilic incubation Subculture: at least 2 suspect colonies to Columbia BAP; 42°C ± 2°C for 24 h microaerophilic incubation Further testing: Gram stain, oxidase, and catalase
	NML at St.Hyacinthe: Speciation using multiplex PCR (31) with QIAxcel Advanced. This system uses capillary electrophoresis to automate analysis of DNA. It provides accurate data regarding their sizes and quantities. Any contaminating DNA fragments such as non-specific amplicons, primer–dimers, or undigested DNA are detected (9).
	Amplification of specific genomic targets from bacterial lysates: CC18F: 5′-AGG TAT GAT TTC TAC AAG CGA-3′ CC519R: 5′-ATA AAA GAC TAT CGT CGC GTG-3′ Hippo-F: 5′-GAC TTC GTTG CAG ATA TGG ATG CTT-3′ HipO-R: 5′-GCT ATA ACT ATC CGA AGA AGC CAT CA-3′ 16S-F: 5′-GGA GGC AGC AGT AGG GAA TA-3′ 16S-R: 5′-TGA CGG GCG GTG AGT ACA AG-3′
AMR susceptibility testing	NML at St. Hyacinthe and NML at Guelph: Minimum inhibitory concentrations using an automated broth microdilution and the CLSI methodology (20) (for CMV2AGNF panel) standards and breakpoints when available (9).
	E. coli and Salmonella: 14 antimicrobials (public health panel) using the CMV2AGNF plate for 2013. CMV3AGNF used for 2013–2014 isolates; CMV4AGNF used for 2016–2017 isolates; CLSI (21) for breakpoints. Campylobacter isolates: 9 antimicrobials using the NARMS CAMPY plates.

BAP — blood agar plate; BGA — brilliant green agar; BPW — buffered peptone water; CIPARS — Canadian Integrated Program for Antimicrobial Resistance Surveillance; CLSI — Clinical Laboratory Standards Institute; EFC SC-SC — Egg Farmers of Canada’s Start Clean-Stay Clean program; FNC — FoodNet Canada; HEB — Hunt’s enrichment broth; mCCDA — modified charcoal-cefoperazone deoxycholate agar; MSRV — modified semi-solid Rappaport Vassiliadis; NML — National Microbiology Laboratory; RVS — Rappaport-Vassiliadis; TSA — tryptic soy agar; TSI — triple sugar iron; XLT4 — xylose, lactose, and tergitol 4 agar.

FNC and CIPARS sampling methodology

In ON_SS1a and ON_SS1b, FNC and CIPARS requested the collection of extra fecal samples during the regular SC-SC farm visits by the field worker/Egg Farmers of Ontario staff. To harmonize with the other poultry farm surveillance programs, in which one pooled fecal sample is required for one quadrant of a barn (or sampling unit), the layer collection approach was modified depending on the structure of the laying barn (e.g., one pooled fecal sample collected from one tier of cages, one manure belt or the equivalent of a quarter of the barn population). Approximately 60 g or half of the 129 mL standard fecal containers per pooled sample was collected. For each farm, one sampling unit (barn, floor, pen) was randomly selected for testing so that each farm sample submission (n = 4 per farm) represented a single age group of layers. Field workers were instructed to collect samples from safe and accessible areas where there was obvious pooling of fecal material such as the end of the manure conveyor belts. All Ontario samples were sent to the National Microbiology Laboratory (NML) in Saint-Hyacinthe for bacterial isolation.

SC-SC protocol

In BC_SS2a, the environmental swab samples collected for the SC-SC program were pooled in sterile Whirl-Pak bags and labelled to indicate that they were for the FNC and CIPARS enhanced testing. One pooled environmental sponge swab sample represented multiple surface swab locations within the barn (e.g., manure belts, egg conveyor belts, dust from walls and floors). The number of samples submitted was proportional to the bird population and varied from one (collected from ≤ 5 000 birds of the same age on the same floor) to 7 samples (collected from ≥ 35 000 birds of different ages) per farm. Flocks of the same age but from different barns were sampled separately. Each pooled sample represented one age group, thus, some of the farm submissions constituted multiple flocks. Samples were submitted to the British Columbia Ministry of Agriculture Animal Health Center (AHC) in Abbotsford, for bacterial isolation.

In both provinces, a 1-page survey sheet was included with each sampling kit/submission in order to collect farm information [e.g., date of sample collection, age of the flock(s) sampled, and total bird population] for both FNC and CIPARS. No farm identifiers were included in the submission form; the codes were kept by the Egg Farmers of Ontario and the British Columbia Egg Marketing Board staff.

Microbiological methods

Table I outlines the respective primary isolation steps. The British Columbia AHC forwarded all isolates to either NML in Saint-Hyacinthe (Campylobacter) or NML in Guelph (Salmonella and E. coli) for further characterization and susceptibility testing, per routine CIPARS protocol (9). Briefly, for Salmonella and E. coli, isolates were tested using the CMV2AGNF (2013 isolates) or CMV3AGNF (2014 to 2017 isolates) (Sensititre; Trek Diagnostic Systems, West Sussex, England), a public health panel comprised of 14 antimicrobials designed by the United States National Antimicrobial Resistance Monitoring Program (NARMS). For Campylobacter, isolates were tested using the CAMPY AST Plate (Sensititre; Trek Diagnostic Systems) comprised of 9 antimicrobials.

Data analysis

Analysis was performed with SAS 9.3 (SAS Institute, Cary, North Carolina, USA) as per routine CIPARS analysis described elsewhere (9). Data were dichotomized for each organism at the isolate level, into susceptible (including intermediate susceptibility) or resistant, using Clinical Laboratory Standards Institute (CLSI) breakpoints (20,21). In the absence of CLSI interpretative criteria, breakpoints were based on the distribution of the minimum inhibitory concentrations and harmonized with those of the NARMS (9). As previously described, multiple samples per flock were accounted for, clustering in the calculation of prevalence estimates using generalized estimating equations with a binary outcome, logit-link function, and exchangeable correlation structure. Null binomial response models were run for each antimicrobial and organism, and from each null model, the intercept (β0) and the 95% confidence interval (CI) were used to calculate population-averaged prevalence estimates using the formula:

$${[1+\text{exp}(-{\mathrm{\beta }}_0)]}^{-1}$$( 9).

Limitations of the study

We caution our readers that the total flocks sampled were lower than the number of flocks per sentinel site routinely sampled in the CIPARS and FNC programs. Sampling timeframes also varied and coincided with the establishment of the FNC sites. The intent of this paper was to provide a landscape of farm sample collection, laboratory methodologies, and prevalence of foodborne bacteria and their AMR profile and explored available resources from which a routine AMR surveillance could be built. Some isolates were not tested for AMR (Ontario, 2013 to 2014) as these were collected prior to the harmonization of farm-level surveillance programs between CIPARS and FNC. A more detailed questionnaire and standardized farm sampling methodology are required to complement these initial data collected for national-level estimates of Salmonella and Campylobacter recovery rates and AMR prevalence.

Results

Flock characteristics

The average flock population had 12 612 layer birds and ranged from 3200 to 32 000 layer birds. The average age of sampling was 37 wk and ranged from 20 to 64 wk of age.

Recovery

There were 38 pooled fecal samples from 9 layer farms in ON_SS1a, 60 fecal samples from 15 layer farms in ON_SS1b, and 54 environmental sponge swab samples from 26 layer farms in BC_SS2a. The E. coli recovery rate was comparable, 98% for both fecal and environmental sponge swabs (Table II). The Salmonella and Campylobacter recovery rate varied: 29% in ON_SS1a and 8% in ON_SS1b fecal samples compared to 7% in BC_SS2a for Salmonella; 89% (ON_SS1a) and 53% (ON_SS1b) for the fecal samples compared to 0% for the environmental sponge swabs (BC_SS2a) for Campylobacter.

Table II.

Recovery of Escherichia coli, Salmonella, and Campylobacter in layers from Canadian Integrated Program for Antimicrobial Resistance Surveillance, and FoodNet Canada sentinel sites (2013 to 2017).

Sample matrix and sentinel site	Year(s)	Escherichia colia % (positive/total submitted)	Salmonella % (positive/total submitted), serovars (n)	Campylobacter % (positive/total submitted), species (n)
Fecal samples (ONSS1a)c	2013	Not done	29% (11/38)	89% (34/38)
			Braenderup (3)	C. coli (n = 20)
			Heidelberg (1)	C. jejuni (n = 14)
			Kentucky (5)
			Ohio var. 14+ (2)
Fecal samples (ONSS1b)	2016 to 2017	98% (59/60)	8% (5/60)	53% (32/60)
			Kentucky (1)	C. coli (n = 12)
			Livingstone (4)	C. jejuni (n = 20)
Environmental sponge swab (BCSS2a)	2013 to 2014	98% (53/54)b	7% (4/54)	0/54
			Infantis (1)
			Liverpool (1)
			Livingstone (1)
			Mbandaka (1)

^aGeneric E. coli isolates were not further characterized. All isolates were nonhemolytic.

^b2 isolates per pooled environmental sample were submitted for susceptibility testing, except for 1 sample where only 1 isolate was available for testing (N = 107).

^cAntimicrobial susceptibility testing was not done on these isolates (transition year for the methodological and operational harmonization between CIPARS and FoodNet Canada farm components).

Overall, 8 different Salmonella serovars were detected (Table II), the 2 most frequent being S. Kentucky (6 isolates) and S. Livingstone (5 isolates). There was no S. Enteritidis recovered.

Antimicrobial resistance

Escherichia coli data for the ON_SS1a isolates were unavailable (collected before the harmonization of FNC and CIPARS farm surveillance methods) but full data were available from ON_SS1b and BC_SS2a. Percentage of resistance among E. coli and 95% CI were organized according to the Veterinary Drugs Directorate (VDD), Health Canada’s categorization of antimicrobials according to their importance in human medicine (22) (Figure 1). Resistance to ceftriaxone, a VDD Category I antimicrobial, was relatively low (< 1% or 1 isolate in each sentinel site). Three isolates (5%) from ON_SS1b and 4 isolates (4%) from BC_SS2a were resistant to gentamicin. In BC_SS2a, resistance to all other antimicrobials was relatively low (0% to 9%, depending on the antimicrobial) except for tetracycline at 26% (95% CI: 18% to 38%). In ON_SS1b, 69% (95% CI: 42% to 82%) of E. coli isolates were resistant to tetracycline.

Figure 1.

Resistance of Escherichia coli (n = 166) in layer flocks from Canadian Integrated Program for Antimicrobial Resistance Surveillance and FoodNet Canada sentinel sites. Roman numerals II to IV indicate categories of importance in human medicine, as outlined by the Veterinary Drugs Directorate.

There were E. coli isolates that exhibited resistance to one or more of the antimicrobial drugs tested. In ON_SS1b, there were 11 resistance patterns observed (including susceptible isolates) and the prevalence for each pattern ranged from 2% (5 various resistance patterns) to 37% (tetracycline) with a corresponding low percentage of susceptible isolates (27%) (Figure 2). In contrast, diverse resistance patterns (n = 15) were detected from BC_SS2a isolates and the prevalence ranged from 1% (11 various patterns) to 16% (tetracycline) with a corresponding high percentage of susceptible isolates (69%) (Figure 3). In ON_SS1a, 2% of isolates exhibited resistance to 7 and 8 antimicrobials (4 to 6 classes) and in BC_SS2a, 2% of isolates exhibited resistance to 8 antimicrobial drugs (5 classes). All of the Salmonella isolates were susceptible to all the antimicrobials tested.

Figure 2.

Drug resistance patterns of Escherichia coli in layer flocks (n = 166) from Canadian Integrated Program for Antimicrobial Resistance Surveillance and FoodNet Canada sentinel sites. Drug patterns occurring less frequently (1% each) include: AcAmCeCfCxStSuTe, AmGeNaStSuTeTm, AmSt, AmStTe, AmSu, AmTe, ChNaSuTe, GeStTe, GeTe, Na, Su.

Ac — amoxicillin-clavulanic acid; Am — ampicillin; Ce — ceftiofur; Cf — cefoxitin; Ch — chloramphenicol; Cx — ceftriaxone; Ge — gentamicin; Na — nalidixic acid; St — streptomycin; Su — sulfisoxazole; Te — tetracycline; Tm — trimethoprim-sulfamethoxazole.

Figure 3.

Resistance of Campylobacter in layer flocks from Canadian Integrated Program for Antimicrobial Resistance Surveillance and FoodNet Canada sentinel sites. Roman numerals II to IV indicate categories of importance in human medicine, as outlined by the Veterinary Drugs Directorate.

Campylobacter isolated from ON_SS1a (n = 4) and ON_SS1b (n = 3) exhibited resistance to ciprofloxacin, a fluoroquinolone class of antimicrobials belonging to VDD Category I. In both sentinel sites, C. jejuni ciprofloxacin resistant isolates originated from the same farm. In ON_SS1a, 38% (95% CI: 16% to 66%) and in ON_SS1b, 78% (95% CI: 42% to 94%) of Campylobacter isolates exhibited resistance to tetracycline (Figure 3). The 7 ciprofloxacin resistant isolates from Ontario (both sites) exhibited the ciprofloxacin-nalidixic acid-tetracycline resistance pattern.

Discussion

This study evaluated the prevalence of E. coli, Salmonella, and Campylobacter using 2 different sample matrices, serotype (i.e., S. Enteritidis), species (i.e., Campylobacter spp.), and resistance of isolates to the range of antimicrobials routinely tested by CIPARS. Our analysis indicates that E. coli and Salmonella can be isolated from either of the sample matrices, fecal samples, or environmental sponge swabs, using routine culture methodology. In the Ontario sites, the recovery of Salmonella varied depending on the surveillance year, but overall, Salmonella prevalence rates in our study were lower (8% and 29% in Ontario and 7% in British Columbia) compared to previous studies in layers conducted by CIPARS (42% of spent hens) (12) and in Alberta (57% to 60% barn and flock level prevalence, respectively) (13). These initial findings provided a descriptive landscape of current monitoring efforts, laboratory methods, and the potential contribution of chicken egg layers in the ecology of foodborne zoonotic bacterial pathogens in Canada.

With limited data collected so far by both CIPARS and FNC, temporal patterns of Salmonella prevalence could not be determined as the studies varied in terms of sampling design, age of the flock at sampling, geographical locations, study timeframe (i.e., sampling coincided with the establishment of the FNC site and coordination of sampling activities with the layer industry), and the sample matrix used (swabs versus fecal samples). As for Campylobacter, the organism was detected from the fecal samples and not from the environmental sponge swabs. Although there were slight variations in primary culture techniques, both sample types were subjected to Campylobacter isolation techniques according to routine methods suggested for poultry specimens (23).

The overall intent of the SC-SC program is to improve the quality and microbiological safety of eggs and to reduce S. Enteritidis entering the food chain. The absence of egg-associated outbreaks in Canada in recent years indicates that this food safety program plus the intervention, which involves the eradication of positive flocks and a compensation program(6), and post-eradication measures (cleaning and disinfection) are working. However, in the context of a comprehensive and harmonized public health surveillance program, such as CIPARS and FNC, the ability of the current SC-SC methodology for detecting other enteric pathogens such as Campylobacter appears to be limited. Previous studies in layers that have used voided fecal/cecal samples found high Campylobacter recovery rates (24,25), which are directly reflective of bird-level Campylobacter compared to other sample matrices such as environmental swabs, which are likely more reflective of the barn or farm environmental flora. The low recovery rates from the sponge swabs could be due to the presence of environmental flora or the lack of transport media, thus requiring further validation of the technique. By contrast, although there was some variability, voided fecal/cecal samples performed consistently with the methods used by CIPARS and FNC in other poultry sectors such as broiler chickens and turkeys (9). Exploration of other farm-level methods similar to a study in the United States, which used a combination of environmental swabs taken from surfaces with visible fecal contamination including system wires, nest boxes, scratch pads, and fecal swabs plus shell pools (26–28), could be an option to enhance Campylobacter recovery. It is known that layer flocks could be colonized internally with Campylobacter (29) and regardless of the housing structure (cage and modifications of the cage systems), the organism could persist in the barn environment (26–28). As such, other than methodological validation of the sponge swab technique, the public health impact of layer flock exposures or egg consumption associated Campylobacter illnesses in people and the contamination pathways leading to human illness (i.e., dissemination via the food chain, via egg handling or environment) also warrant further research.

The antimicrobial resistance observed in E. coli was relatively infrequent compared to other poultry sectors. The moderate to high prevalence of resistance to certain antimicrobials (e.g., tetracycline and streptomycin), resistance to multiple drugs, and resistance to cephalosporins and fluoroquinolones, although comparatively lower than broiler chickens and turkeys, are somewhat unexpected. It is generally supposed that during the laying phase, exposure to antimicrobials in laying flocks is limited because of potential drug residues in table eggs (30); however, there are no published data to corroborate that assumption. In general, AMR prevalence and AMR profiles of isolates from Ontario differed from British Columbia. Bacterial diseases continue to be diagnosed in laying hens in major egg-producing provinces (10,11) including salpingitis/egg peritonitis and early systemic bacterial infections, but without animal health/operational farm-level information, it is impossible to assess the factors or drivers for AMR in layers. Surveillance timeframes and provincial variations in drug use could have played a role in the variations in AMR. The sample matrix used in the surveillance design could also have had an impact, as freshly voided fecal samples (or cecal samples) may be more reflective of a relatively recent AMU selection pressure. The CIPARS farm program (9) routinely uses fecal samples, as do many other food safety and AMR surveillance programs (31,32), because of the consistency in AMR profiles observed and the ability to detect changes in AMR prevalence over time. In contrast, the sponge swab, as previously described, may be reflective of the microbial flora within the farm/barn environment, where AMR dynamics distinct from the gut flora are potentially occurring. External to the bird’s immediate environment, exposure of the organisms to disinfectants and other cleaning agents and the presence of vectors could be playing a role.

The AMR findings warrant monitoring of AMU at all phases of egg production (e.g., hatchery, pullet, and laying phases) to understand the emergence of AMR in layer flocks in Canada. To the best of our knowledge, this is the first evidence that Campylobacter spp. and E. coli resistant to multiple antimicrobials are present in layer flocks in Canada. Although the environmental sponge swabs yielded comparable recovery rates for E. coli and comparable Salmonella prevalence rates, the use of fecal samples appears to be the most appropriate matrix for public health surveillance that includes Campylobacter and to inform a comprehensive foodborne pathogen intervention and AMU stewardship in the layer sector. In light of the changes in AMU regulations in Canada (33) and AMU reduction called for by the larger poultry industry group (34,35), the utility of fecal samples could also be explored for animal health surveillance (e.g., Clostridium perfringens, Brachyspira spp., Eimeria spp.). This approach could allow across-species harmonization and methodological harmonization of farm level surveillance for layers, such as those used in the United States (14) and in the European Union Member States (19). However, the CIPARS and FNC programs are voluntary in nature and have limited flock coverage compared to the SC-SC, which is mandatory for all layer flocks; thus, opportunities for synergies could be explored to enhance the sustainability of these surveillance and monitoring programs.

In summary, inclusion of layer flocks in a surveillance program with a public health theme would contribute to the understanding of the ecology of AMR in the poultry industry and enable comprehensive source attribution and risk mitigation of foodborne illnesses and AMR using a One Health approach.