Occurrence and Characterization of Salmonella Isolated from Table Egg Layer Farming Environments in Western Australia and Insights into Biosecurity and Egg Handling Practices

The aim of this study was to investigate the occurrence and distribution of Salmonella in commercial layer farming environments of 26 flocks belonging to seven egg businesses (free-range and barn-laid) in Western Australia (WA). Between November 2017 and June 2018, a total of 265 environmental samples of dust, feed, water, pooled feces, and boot swabs were tested for detection of Salmonella according to standard culture-based methods. Isolates were assayed for serovar and subtyped by multilocus sequence typing (MLST). Salmonella spp. were recovered from 35% (93/265) of all tested samples. Dust (53.8%, 28/52) and pooled fecal (54.5%, 18/33) samples provided the highest Salmonella recovery rates. Nine different Salmonella serovars were characterized across the positive (n = 93) environmental samples, of which S. Typhimurium (60/93, 64.5%) and S. Infantis (21/93, 22.5%) were the most prevalent. MLST revealed that all S. Typhimurium isolates were of sequence type ST-19. Microbiological screening of Salmonella was not routinely practiced in any of the surveyed egg businesses. Some of the egg businesses exhibited variable levels of compliance with basic biosecurity measures as well as high-risk egg handling practices. Egg businesses in WA should be encouraged to adopt a voluntary program of environmental sampling and verification testing for Salmonella. Such voluntary programs will aid in supporting solutions for the management of this pathogen in the human food chain.


Introduction
Non-typhoidal Salmonella species are among the major foodborne pathogens in Australia [1], with many traced human outbreaks attributed to consumption of eggs and egg-based products [2,3]. S. Typhimurium is the key endemic serovar in the Australian egg production industry and is the most frequent serovar isolated from human cases [4]. In recent years, there has been a significant expansion of non-cage (free-range and barn-laid) egg production systems around Australia [5] arising from an increase in consumer demand and concerns regarding caged bird welfare [6]. Nevertheless, some researchers hypothesize that non-cage egg production systems may lead to a higher exposure of layer flocks to the outdoor environment, pests, and wildlife vectors [4,7,8]. The growth of non-cage egg production systems, along with the inherent biosecurity and environmental challenges in such systems, have raised concerns regarding the likelihood of pathogens such as Salmonella accessing consumer retail eggs in contaminated farm environments [9]. In Western Australia (WA), the number of human salmonellosis cases has dramatically increased and has been higher than in any other Australian state in recent years [10]. In 2017, the number of cases in WA was more than double the previous five-year average, with eggs and egg-based dishes emerging as the key culprits in several Salmonella foodborne outbreaks [10]. At present, there is no published research on Salmonella status in table egg layer farm environments in WA, a state that occupies the entire western third of Australia. To understand the epidemiology of salmonellosis events in WA, it is important to generate a baseline status for Salmonella contamination in the layer farm environment. The aims of this study were: (1) to establish a baseline survey of the occurrence of Salmonella in egg layer farming environments in WA; (2) to characterize identified Salmonella isolates based on their serovars and multilocus sequence type (MLST) diversity; and (3) to collect industry profile information on common primary production practices and identify areas in which egg businesses may require further assistance in meeting biosecurity and compliance requirements.

The Occurrence of Salmonella in the Layer Farm Environments
Salmonella was isolated from the environments of all seven egg businesses and from 23 of the 26 (88.5%) sampled flocks. Salmonella was recovered from all (11/11) of the free-range flocks as well as from 80% (12/15) of the barn-laid flocks (Table 1).
Results presented in Table 2 demonstrate that although S. Typhimurium was the only detected serovar in the environmental samples collected from egg businesses D and E, there was more than one serovar detected in most of the other egg businesses. S. Typhimurium and S. Infantis co-existed in flock houses sampled from egg businesses F and G. Moreover, egg businesses A and B harbored four and three different serovar in their flocks, respectively ( Table 2). Thirty-three S. Typhimurium isolates recovered from different flocks were further subtyped using MLST. All 33 S. Typhimurium isolates were characterized as ST-19.

Production Management, Biosecurity, and Egg Handling Practices
A descriptive overview of the questionnaire results obtained from the farm management teams of the seven participating egg businesses is summarized in Table 3. All of the businesses had in-house quality assurance systems, and almost all had adopted a written food safety management statement. A written/documented Hazard Analysis and Critical Control Point (HACCP) system was also present in most of the farms. Nevertheless, few farm managers and staff followed any kind of either formal or in-house training on aspects of HACCP, food safety principles, or biosecurity. The results of the questionnaire also indicated that the egg businesses did not always seek vaccinations against Salmonella for their flocks (Table 3). Further, the summary of answers indicated a variable level of compliance with basic biosecurity measures. Areas observed as needing improvement included systematic cleaning and sanitization of sheds between production cycles, standards for disinfection of incoming vehicles to farms, and regular change of workers' clothes while moving between sheds or instructing farm workers to wear personal protective equipment (Table 3). Other biosecurity gaps included inadequate or absent handwashing facilities and boot sanitization dips in many of the farms, poor maintenance of pest control stations and baits, and inadequate practices to restrict rodent infestations and access to feed and sheds.
Some high-risk egg handling practices were also observed. For instance, three of the seven (42.9%) egg businesses indicated that the storage period prior to grading was as long as 48 h. Moreover, in four of the seven egg businesses (57.1%), pre-graded eggs were typically stored at ambient temperatures (rather than being refrigerated or kept at 15 • C). Finally, the sanitizing frequency of egg handling equipment was variable in terms of frequency and procedure. Daily sanitization of egg handling equipment occurred in four (57.1%) of the visited farms, while weekly (14.3%, 1/7) and even monthly (14.3%, 1/7) sanitization was reported by the remainder.

The First Insight on the Status of Salmonella in Egg Businesses in WA
Microbiological surveillance of farm environments may be beneficial for predicting potential Salmonella introduction and colonization of layer flocks and subsequent contamination of table eggs [11]. This is the first study to investigate the occurrence of Salmonella in commercial egg businesses in WA. The results of the present study supply new knowledge on the diversity of Salmonella serovars and Sequence Types (STs) and provide industry profile information on common primary production practices in WA egg businesses.
All of the visited egg businesses in the present study in WA were positive for Salmonella (100%, 7/7). Compared with our results, farm-level surveys conducted in the Australian states of Queensland (Qld) [12] and New South Wales (NSW) [13] found a lower occurrence of Salmonella (33.3% and 47.0%, respectively). Similarly, lower rates were also found in countries such as New Zealand (20%, 4/20) [14], Korea (59.3%, 19/32) [15], and Argentina (43.3%, 13/30) [16]. Additionally, our results indicate that S. Typhimurium was present in five of the seven (71.5%) surveyed farms, which is considerably higher compared with findings from the Australian states of Qld and NSW, which reported detection of environmental S. Typhimurium in 13.5% and 20% of surveyed farms, respectively [12,13]. Our results were also higher than those of a New Zealand survey of S. Typhimurium in the environment of layer farms (16.6%, 3/18) [14]. The higher rates of Salmonella, particularly S. Typhimurium, detected in the present study compared with those found in other parts of Australia and New Zealand may be attributable to the climatic conditions in different geographical areas, as well as to Salmonella infection status of birds and the effectiveness of cleaning and other hygiene and biosecurity measures [17]. S. Typhimurium has been the dominant serovar in recent human salmonellosis cases in Australia and has been frequently implicated in causing egg-borne human salmonellosis outbreaks, which is an increasing public health concern [18,19]. In WA, the reported number of human salmonellosis cases has increased dramatically compared with other Australian states, and S. Typhimurium was the major species responsible for most of the egg-linked outbreaks in WA [10].
According to our results (Table 1), Salmonella was recovered from all of the free-range egg businesses flocks and 80% of the barn-laid egg businesses flocks. These results contrast with investigations conducted in Qld (free-range [30%, 3/10] and barn-laid [50%, 1/2]) [12] and NSW (free-range [40%, 12/30] and barn-laid [100%, 4/4]) [13]. There is conflicting evidence on the effect of housing as a risk factor for Salmonella shedding. A previous study in Belgium suggested that there is no consistent conclusion on the interaction of stress with different housing systems [20]. In contrast, a study in the United States found that free-range hens are exposed to a higher number of environmental stressors, including weather variations, heat stress, and predation, as well as contact with wild birds, which leads to a high chance of Salmonella shedding in environment [8]. However, more studies are required to understand the epidemiology of S. Typhimurium in free-range egg businesses; especially that such production system has seen growth driven by consumer demands, focus on animal welfare, as well as the environmentally friendly production of eggs in Australia and other countries [9].

Salmonella Recovery from Environmental Samples
According to our results, pooled fecal (54.5%, 18/33) and dust (53.8%, 28/52) samples had the highest rate of Salmonella recovery. Similar to our findings, investigations in South Australia (SA) also found that dust samples had the highest occurrence of Salmonella, with recovery rates ranging from 14.1% (10/70) [9] to 41% (29/70) [4], which are lower than our results. A recent study in New Zealand [14] found that dust samples had the highest rate of Salmonella recovery (28.3%, 19/67), which aligns with our findings even though the rates were lower. Compared with our findings, lower rates of Salmonella were found in fecal samples originating from layer farm environments in SA (11.7%, 5/40) [9] and NSW (17%, 15/90) [13] as well as in other parts of the world, including New Zealand (10.4%, 7/67) [14], Korea (41.8%, 28/67) [15], and Kosovo (11%, 22/200) [21]. For routine verification and voluntary self-monitoring of egg businesses in WA, fecal and dust samples are highly recommended as the most significant indicators of present and prior Salmonella infection in flocks, respectively. Reintroduction of Salmonella to flocks has been attributed to residual Salmonella contamination on the layer farm [22,23]. It is also hypothesized that Salmonella has a greater ability to persist in dust than in other environmental samples [9]. Therefore, dust samples provide an excellent verification of the effectiveness of cleaning and disinfection. In free-range flocks, it is challenging to clean and disinfect ranging areas, however, effective sanitation by removing dust from different surfaces in sheds, especially between production cycles, is shown to be valuable in reducing the level of Salmonella contamination [24].
According to our results, the level of Salmonella recovered from inside-shed feed samples (30.4%, 14/46) was higher than that found in similar samples in a farm environment survey in NSW (17%, 17/101) [13]. Further, Salmonella was also recovered from the tested silo feed samples (26.9%, 7/26) in the present study. Compared with our results, a lower rate of Salmonella recovery was reported for stockfeed samples in investigations conducted in Qld (0%, 0/21) [12], NSW (11%, 3/27) [13], and New Zealand (3%, 1/33) [14]. In the present study, S. Typhimurium was the serovar detected in 57.1% (4/7) of the stockfeed samples found to be culture-positive for Salmonella. However, this serovar was not detected in any of the Salmonella-positive stockfeed samples in the NSW investigation [13]. A recent 2019 survey from New Zealand reported that only 1 of 33 tested feed samples was positive for Salmonella, which was serotyped as S. Thompson [14]. The limited number of local stockfeed supplier companies for egg businesses in WA along with the abundance of S. Typhimurium recovery from in-silo feed samples warrants a trace-back verification of the status of Salmonella in primary stockfeed production lines. Contaminated feed ingredients such as animal protein sources or the use of contaminated vehicles for distributing feed may play a significant role in the spread of Salmonella on farms [21,25]. Protein sources in poultry feed have been frequently reported to be reservoirs of many Salmonella serovars [26][27][28].

Diversity of Salmonella Subtypes Recovered from Environmental Samples
According to our findings, S. Typhimurium (64.5%, 60/93) and S. Infantis (22.5%, 21/93) were the most common isolates from farm environments in WA. Similar to our findings, S. Typhimurium (30%, 39/130) and S. Infantis (19%, 25/130) were also reported as predominant serovars in a previous study in NSW [13]. Other Australian investigations have also indicated that these two serovars are among the predominant Salmonella serovars in Australian egg farms [9,12,[29][30][31]. The results of a recent Salmonella survey in layer farms in New Zealand revealed that S. Infantis was also the most detected serovar, while S. Typhimurium was the third most detected [14]. The majority (84%) of reported foodborne Salmonella spp. outbreaks in Australia between 2001 and 2016 were attributed to S. Typhimurium, while only 1% of these outbreaks were related to S. Infantis [32]. Although human cases associated with S. Infantis are not as common as those associated with S. Typhimurium, S. Infantis has emerged as the fourth most common serovar causing human salmonellosis in Europe in 2014 [33].
In the present study, S. Typhimurium ST-19 was identified as the common multilocus sequence type circulating in layer farms in WA. Numerous common gastroenteritis-causing S. Typhimurium isolates have been attributed to ST-19 [34][35][36]. ST-19 was also the most common in S. Typhimurium isolates recovered from chicken carcasses in Iran and India [37] as well as in clinical Salmonella strains in Brazil [38].
Our results stress the important role of the egg production sector in the epidemiology of Salmonella outbreaks in WA; the abundant recovery of S. Typhimurium of ST-19 in layers farming environments as well as the predominance of such subtype in human cases calls for prioritization of Salmonella management efforts at the primary egg production sector.

Biosecurity and Production Practices-Points of Attention
Adherence to basic biosecurity measures is crucial for protecting laying hen farms against the introduction of Salmonella, as has been reported in several previous investigations [39,40]. Findings from the questionnaire (Table 3) indicate a variable level of biosecurity training for farmers and other staff members in the surveyed egg businesses. It is important to adopt either formal or in-house training on biosecurity and food safety aspects at the level of farms operating in WA. Such training will provide reassuring refreshment on basic principles, and will attain new workers to acquire the best required practices. A Canadian study on poultry farms indicated that biosecurity compliance is closely related to responsibility, work experience, and education [41].
Routine Salmonella screening of layer flocks is not mandatory in Australia. However, monitoring for S. Enteritidis for commercial egg producers exporting eggs to overseas markets is required in some Australian states, including NSW, Qld, and Victoria [42]. Thus, not surprisingly, Salmonella screening was not routine on the visited farms in the present study in WA. Having a regular monitoring/surveillance program for environmental sampling and testing for Salmonella will not only help detect and respond to Salmonella infection before it spreads, but will also generate knowledge for developing solutions to control this pathogen at the farm level [40].
Various cleaning and sanitization methods were used by business owners in WA. However, most (71.4%, 5/7) did not consider practicing microbiological verification for cleaning and sanitization between successive flocks. The results of a study conducted in five countries across Europe (Belgium, Germany, Greece, Italy, and Switzerland) indicated that 31.8% of farms with Salmonella-positive flocks conducted cleaning and disinfection of henhouses, although there was no information about microbiological verification after cleaning [39]. Given that Salmonella can survive in empty sheds for several months, especially in organic material such as dust and manure [43], it may transfer between consecutive production cycles. Therefore, ineffective dry-cleaning procedures such as the mechanical removal of organic material (manure, dust, and feed spills) may play a significant role in the persistence of Salmonella on farms [39].
According to our questionnaire survey, farmers noted that rodents and birds had access to the vast majority of the visited sites. Open and unsecured shed doors may also provide easy access for unwanted insects, pests, and predators. Similar to our findings, previous investigations in Australia have also demonstrated the role of environmental vectors in the introduction of S. Typhimurium, particularly in free-range layer farms [4]. Another study conducted in three Caribbean countries also revealed that 90% of contaminated farms had problems with rodents [44]. High rodent density in layer flocks leads to a higher rate of Salmonella shedding and infection [45]. Compared with most other sources of Salmonella, rodents have a higher ability to spread this pathogen from one flock to another [46]. Rodents can maintain asymptomatic latent infections and shed Salmonella through their fecal pellets, which can infect chickens through the consumption of mouse fecal pellets in feed [47]. Therefore, rodent control programs in poultry premises, including the repairing of holes and other potential access points for rodents into henhouses, the removal of vegetation around sheds, the application of useful baits and bait placement, may help to control Salmonella contamination in layer flocks [48].
In the present study, some high-risk egg handling practices were observed at the visited farms in WA. For instance, three of the seven farms indicated that the storage period before grading was as long as 48 h. Our findings also pointed that in four of the seven farms, pre-graded eggs were typically stored at ambient temperatures (rather than at 15 • C, as recommended by the code of Australian Eggs -a producer-owned corporation). According to the Australian Eggs recommendation, eggs should be stored below 15 • C (±3 • C) on farms, during transport and at the retail outlet [49]; apparently such recommendation is not met in most of the visited farms in WA. Salmonella can survive on the surface of the eggs for several weeks [50], thus reducing storage temperature can decrease the multiplication and penetration of Salmonella during storage at farms. Egg-collecting areas are one of the most important reservoirs for Salmonella cross-contamination. The results of a study in Belgium indicated that S. enteritidis was common on equipment and surfaces in egg packing areas on farms [51]. Contamination may occur during and after egg collection via contact with workers, surfaces, and equipment [52]. Therefore, adherence to good practices, as those recommended by the code of Australian Eggs [49], is necessary during egg handling, including sorting, candling, grading, and packaging of eggs, to minimize cross-contamination and eggshell surface damage [52].
Further, our results also revealed that farmers did not consider vaccinating their flocks against Salmonella. Vaccination is one of a widely accepted means worldwide of preventing or reducing Salmonella shedding [53]. Poor adoption of Salmonella vaccination in WA egg businesses is not surprising, as the vaccination culture varies between and within countries-for instance, a 2012 study conducted by Sasaki et al. [54] in Japan showed that one-third of the visited layer flocks had been vaccinated against Salmonella. In Australia, there is only one approved live attenuated vaccine against S. Typhimurium; the long-term efficacy of this vaccine in commercial layer flocks has been evaluated in previous Australian studies but is still questionable [55,56]. However, it has been shown that vaccination is most effective when applied alongside other biosecurity and sanitation procedures [52]. While vaccination is usually effective to one or a low number of serovars, a suitable microbiota could be effective for a broad range of serovars. The first attempt to find such a suitable microbiota for chicks has been reported recently [57].

Recruitment of Egg Businesses
In the present study, we targeted non-cage (free-range and barn-laid) egg businesses because this sector is the most rapidly emerging in the WA table egg industry (promoted by the state government's plan for a 10-year phase-out of conventional cages for egg-laying hens) [58]. Field work for this study was conducted between November 2017 and June 2018. An invitation package was sent by mail and email to listed egg businesses (n = 27) across WA prior to the commencement of the project. The list of businesses was supplied by the WA Department of Health (Food Unit). The aim of the invitation packages was to communicate the objectives of our research to egg businesses and local governments as well as to assure data confidentiality of the participating businesses. In total, seven egg businesses (7/27 (26%)) voluntarily agreed to participate in the present study. The geographical distribution of the surveyed businesses included metropolitan and non-metropolitan regions across WA. The epidemiological sampling unit in the present study was 'a flock', defined in the scope of this study as a group (or batch) of laying hens raised in the same shed. The research team sampled 26 flocks from the participating egg businesses. Inclusion criterion was flocks with a production age of >20 weeks (because the laying capacity is less than 50% at younger ages) [12].

Environmental Sampling
A total of 265 environmental samples were collected from 26 flocks belonging to seven egg businesses. The methodology for collection of each sample type is described in Table 4. The samples consisted of dust (n = 52), boot swabs from inside sheds (n = 40), boot swabs from outside sheds (n = 22), feed from inside sheds (n = 46), feed from outside sheds (n = 26), and water from inside sheds (n = 46). Flock sampling methodology adopted in this study was based on guidelines from the European Union (EU) monitoring program [59] and the United Kingdom National Control Program (UK-NCP) (Commission Regulation (EC) no. 1168/2006) for Salmonella in layer hens with some modifications. The EU and UK-NCP guided protocols recommend that pooled fecal samples be collected from the environment of cage-laid flocks; however, in the present study, we included pooled fecal samples (n = 33) from barn-laid sheds to assess the added value of this sample type in the setting of non-caged flocks. Disposable sterile gloves, boot swabs, and bags were used during sample collection. All collected samples from the various flock sheds were labelled and aseptically transferred to the Veterinary Public Health Research Laboratory at Murdoch University under chilled conditions. On arrival at the laboratory, samples were refrigerated at 5 • C, and microbiological testing started on the day of sampling.

Isolation, Identification, and Serotyping of Salmonella
Preparation of the collected samples was performed with some modifications to UK-NCP and EU recommendations as described by Carrique-Mas et al. [60]. Briefly, the bag containing fecal sample was weighed and a corresponding volume of buffered peptone water (BPW) (Oxoid UK) was added to maintain a sample to diluent ratio of 1:1 (1 in 1). The feces and BPW were then homogenized in a stomacher for 1 min to enable thorough mixing of the sample with the diluent. Subsequently, 50 mL of this mixture was homogenized for 1 min with 200 mL of BPW. For each boot swab, the sample (pair) was also homogenized for 1 min in a stomacher containing 225 mL of BPW. Further, Whirl-Pak sample bags containing dust and feed samples were weighed separately and filled with a corresponding volume of BPW to maintain a sample to diluent ratio of 1:9 (1 in 10). Finally, each water sample was filtered using Rapid-Flow™ sterile disposable filter units with a nylon membrane (Nalgene) before the nylon membrane was transferred and vortexed with 45 mL of BPW in sterile falcon tubes.
Isolation of Salmonella was performed according the 2017 recommendations of the ISO 6579-1 standard [61]. In the pre-enrichment stage, each homogenized sampling unit was incubated at 37 • C for 36 h. Following pre-enrichment, 1 mL was sub-cultured in Muller Kauffmann tetrathionate novobiocin broth (Oxoid UK) and incubated for 48 h at 37 • C, and 0.1 mL was sub-cultured in modified semi-solid Rappaport Vassiliadis (MSRV) (Oxoid UK) and incubated for 48 h at 41.5 • C. MSRV plates were checked after 24 h for a migration zone (turbid, white halo, with a radius larger than 10 mm), and plates with no migration zones were rechecked after an additional 24 h. Following pre-enrichment of both media, streak cultures were conducted on xylose lysine deoxycholate agar and brilliant green agar (Oxoid UK), which were then incubated at 37 • C for 24 h. Suspected colonies (up to five) were selected from both selective media and transferred onto nutrient agar (Oxoid UK) plates. After incubating nutrient agar at 37 • C for 24 h, the purified suspected colonies were identified to species level by matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF MS) using the microflex instrument (Bruker Diagnostics, Germany). All confirmed Salmonella isolates (up to five isolates per positive sample) were serotyped (Kauffmann-White scheme) at a nationally accredited reference laboratory (PathWest) in Perth, WA. Isolates from confirmed positive samples were stored at −80 • C for further surveys.

MLST of S. Typhimurium Isolates
DNA was extracted from 33 S. Typhimurium isolates recovered from positive samples using the Bioline DNA extraction kit (ISOLATE II, Genomic DNA Kit) according to the manufacturer's instructions. According to the University of Warwick (http://mlst.warwick.ac.uk/mlst/), seven housekeeping genes (aroC, dnaN, hemD, hisD, purE, sucA and thrA) were used for the molecular typing of Salmonella isolates. Sequence types (STs) were assigned according to the S. enterica MLST database (http: //mlst.warwick.ac.uk/mlst/dbs/Senterica).

Biosecurity and Production Practices Questionnaire
A comprehensive questionnaire including approximately 120 variables was developed to benchmark the current production practices of surveyed egg businesses in WA. Egg businesses were asked to voluntarily complete the questionnaire which covered several aspects, including information about locality, production system, quality management system, feed and water supply, layer management, aspects of biosecurity management, and egg handling practices. The key variables of the questionnaire are summarized in Table 3. The present study protocol was approved by the Murdoch University Human Research Ethics Committee (Permit No. 2017/196), and the research team obtained written consents from the egg businesses.

Conclusions
The results of the present study illustrate the high occurrence of Salmonella in the surveyed egg business environment in WA. The predominant serovar (S. Typhimurium) identified in this investigation was also reported in several human cases of salmonellosis in recent years in WA. The high level of Salmonella recovery from egg business environments in WA highlights the need for effective management practices and biosecurity measures, including cleaning and disinfection of sheds, regular Salmonella testing programs, rodent control, vaccination, and provision of uncontaminated feed. Insights from the questionnaire may assist the state authority and local governments to develop industry assistance and regulatory (audit/inspection) practices that may prove useful to egg businesses.

Conflicts of Interest:
The authors declare no conflict of interest.