Walking a thin line: Thinning broiler houses with Salmonella contaminated material

Abstract

Salmonella Infantis and Salmonella Paratyphi B var. Java are important human pathogens that are mainly found on chicken meat and derivative products. Furthermore, both serovars are dominantly present in the Belgian broiler sector. In search of contamination routes in the sector, equipment used during partial depopulation or thinning was sampled at ten different broiler farms. All thinning materials, such as the trucks, crates and workwear of the catching crew were sampled before entering the broiler houses. To evaluate if the remaining birds became colonized by Salmonella strains introduced during thinning, the broiler houses were sampled twice using overshoes. Once immediately before thinning and approximately one week after thinning, immediately before final depopulation. The results show that during nine of the ten thinning events Salmonella was present on the equipment used. Especially the crates and loading surfaces of the trucks were frequently contaminated, 19.3 % (62/320) of the sampled crates and 50 % (11/22) of the sampled trucks were found contaminated with S. Infantis, Paratyphi B var. Java or Livingstone. The relative prevalences of S. Infantis, Java and Livingstone were, respectively, 31.3 %, 67.5 % and 1.3 %. Despite the presence of Salmonella during thinning, no cross-contamination to the remaining chickens was observed. A low abundance of Salmonella bacteria, a higher resistance of the older (five weeks), remaining chickens or a small sampling size can explain these results. Nevertheless, the results of this case report underscore the importance of appropriate biosecurity during thinning.

Introduction

A common cause of foodborne disease outbreaks is Salmonella. In 2023, 77486 cases of human salmonellosis were reported in the European Union (EFSA and ECDC, 2024), which is the second most prevalent zoonotic disease within the EU. Salmonella Infantis is one of the top five serovars frequently involved in human infections. This serovar is primarily transmitted through the consumption of contaminated broiler meat and its derivative products. Moreover, Infantis is the most common serovar found in broiler meat, accounting for 52.2 % of all positive broiler meat samples (EFSA and ECDC, 2024; Mughini-Gras et al., 2021). Furthermore, an increase in human cases of the serovar Paratyphi B var. Java (S. Java) was observed in 2023 in the EU. This serovar was especially isolated from broilers and belongs to the top 20 Salmonella serovars in food-animal sources. These findings are consistent with what is observed in Belgium, where S. Infantis was persistently the most prevalent serovar in broiler chickens from 2015 until 2023 (25 – 44 %) (DGZ 2024). Since 2023, S. Java has been the most prevalent serovar with 38 % of the Salmonella positive flocks. However, S. Infantis was still dominantly present with a prevalence of 29 %. These data highlight the importance of identifying effective control measures, particularly at farm level, to prevent further spread of these persistent serotypes (EFSA and ECDC, 2022; Newton et al., 2021). A first step in preventing further spread is to identify critical contamination risks. Feed, drinking water, rodents, pets and insects are known contamination risks (Machado et al., 2020). However, activities such as partial depopulation or thinning can also pose biosecurity risks. A risk analysis based study by Mughini-Gras et al. (2021) confirmed an increased risk of S. Infantis colonization when thinning was applied at Dutch broiler farms. The use of contaminated equipment during thinning can not only lead to external contamination of the birds during transport to the slaughterhouse, but also introduce contamination on a farm and lead to colonization of the remaining birds (Slader et al., 2002). To our knowledge, there is no epidemiological, more recent research then Slader et al. (2002) that have proven Salmonella is still present on cleaned and disinfected transport crates. Therefore, this study investigated the contamination risk of Salmonella (Infantis and Java) by partial depopulation at Belgian broiler farms. Accordingly, 24 broiler houses on 10 different broiler farms were sampled during partial depopulation. The used thinning equipment was sampled just before partial depopulation started and the samples were analyzed for the presence of Salmonella. To assess whether cross-contamination had occurred, the Salmonella status of the broiler chickens was evaluated by sampling two pairs of overshoes before partial depopulation and again, one week later, before final depopulation.

Materials and methods

Sampling

Between March 2022 and June 2024, ten Belgian broiler farms (A-J) were sampled during thinning to assess the role of used equipment and material (e.g. crates, truck, forklift, clothing catching crew etc.) in spreading Salmonella. Farmers took part in the study on a voluntary basis. On each farm all broiler houses were included in the study. In total, 24 flocks were sampled (Table 1), this is an average of 2 flocks per farm. If possible, the materials used for partial depopulation were sampled separately for each broiler house (e.g. the crates and trucks), as farmers sometimes cooperate for their individual broiler houses with different slaughterhouses. In total, it concerned eight different slaughterhouses. An average of 25 samples per broiler house were collected from the crates, catching crew, truck and forklift (Table 2) using 3 M Sponge Sticks (St. Paul, Minnesota, USA) premoistened with Buffered Peptone Water (BPW, Oxoid, Basingstoke, UK). A total of 16 crates per broiler house were sampled unless fewer crates were present. Per crate, eight sliders were sampled using one sponge stick. Per pair of boots one sponge stick was used and several gloves were sampled using one sponge stick. Four wheels of the truck were sampled with one sponge stick, as were four wheels of the forklift. All the equipment was sampled before it entered the broiler houses.

Table 1.

Sizes of the sampled flocks at the 10 different farms (A-J), the respective slaughterhouses, Salmonella status of the farms before partial depopulation and the number of days between the sampling during thinning and the sampling before final depopulation. Only the broiler houses with a Salmonella negative status were re-sampled before final depopulation.

Farm	Slaughterhouse(s)	Broiler house 1	Broiler house 2	Broiler house 3	Salmonella status before thinning	Days between samplings
A	Slaughterhouse 1	20.000	20.000	30.000	Positive	-
B	Slaughterhouse 2	28.350	28.350	N/A	Positive	-
C	Slaughterhouse 3	23.000	40.000	N/A	Positive	-
D	Slaughterhouses 1, 4, 5	23.000	40.000	40.000	Negative	7
E	Slaughterhouse 5	23.000	23.000	N/A	Negative	6
F	Slaughterhouses 1, 6	28.500	40.000	N/A	Negative	5
G	Slaughterhouse 7	37.500	37.500	75.000	Positive	-
H	Slaughterhouse 7	16.500	26.500	N/A	Negative	6
I	Slaughterhouse 8	36.000	36.000	N/A	Negative	8
J	Slaughterhouse 1	72.000	22.860	19.710	Positive	-

Table 2.

Overview and number of the samples collected during partial depopulation per farm. Per crate, eight sliders or the entire unit when less sliders were present (depending on the crate system) were sampled with one sponge stick. Only one sponge stick was used to sample several gloves of the catching crew.

Sample	Number of samples per farm
Crates (1 m²)	Max. 16 per broiler house
Catch crew: pairs of boots	Max. 6
Catch crew: gloves	1
Truck: loading surface (1 m²)	1 per broiler house
Truck: wheels (4)	1 per broiler house
Truck: tailgate (1 m²)	1 per broiler house
Forklift: forks (1 m²)	1
Forklift: wheels (4)	1
Metal drive-up piece	1
Pairs of overshoes	2 per broiler house

The Salmonella status of the broiler houses was determined by walking at least 100 steps through the first half of the broiler houses wearing one pair of overshoes and another pair for the second half. This status was determined before the start of thinning to confirm the Salmonella status of the broiler houses already determined by the farmers as set out in commission regulation (EU) No 200/2012. Five of the farms already had a Salmonella positive status, the other five had a Salmonella negative status which was confirmed by our sampling. After approximately one week, immediately before the final depopulation of the broiler houses, the Salmonella status was determined again. In that way, in case of Salmonella free broiler houses, it was possible to investigate if cross-contamination had occurred during thinning. All samples were transported in a refrigerated box and were analyzed the same day or kept overnight in the cooling until further processing in the laboratory the next day.

Microbiology

The sponge sticks and overshoes were supplemented with respectively 90 ml and 225 ml BPW and homogenized by a bag mixer. They were further analyzed based on the ISO 6579:2017 standard with some minor modifications. Briefly, the samples supplemented with BPW were incubated for 18 hours ± 2 h at 37°C. Afterwards, 100 µl of each sample was distributed on 3 sites of a Modified Semisolid Rappaport Vassiliadis (MSRV, Oxoid, Basingstoke, UK) agar plate and incubated for 24 and 48 hours ± 3 h at 41.5°C± 1°C. Suspected zones on the MSRV plate were sampled by dipping a 10 µl loop in the edge of the opaque growth, this was transferred to Xylose Lysine Deoxycholate (XLD, Oxoid, Basingstoke, UK). The XLD plates were incubated for 24 hours ± 3 h at 37°C. The colonies suspected as Salmonella were further purified and stored in brain heart infusion broth (Oxoid, Basingstoke, UK) with 15 % glycerol at −20°C.

Molecular analysis

Identification by PCR

Similar to De Witte et al. (2025), suspected isolates were lysed in 50 µl 0.25 % NaOH (Merck kGaA, Darmstadt Germany) and 50 µl 0.25 % SDS (Sigma-Aldrich, Saint Louis Missouri, USA) at 95°C for 10 minutes. These lysates were used in a Salmonella Infantis specific PCR to confirm the serovar Infantis (Kardos et al., 2007). Isolates testing negative for the serovar Infantis were then subjected to a Salmonella genus PCR to confirm the Salmonella genus (Aabo et al., 1993). Isolates confirmed as Salmonella, but not as the serovar Infantis, were further serotyped by Check&Trace Salmonella 2.0 (Check-Points BV, Wageningen, the Netherlands) by Lavetan (GBA Group, Turnhout, Belgium).

Fingerprinting by PFGE

Isolates confirmed as Salmonella were further typed by Pulsed Field Gel Electrophoresis (PFGE) according to the international Pulsenet protocol (Center for Disease Control and Prevention, 2013). A digital picture was taken using Bio-Rad GelDoc Go Imaging System. The obtained PFGE results were processed in BioNumerics 8.1.1 (Applied Maths NV). An unweighted-pair group method using arithmetic averages (UPGMA) dendrogram was constructed using the Pearson correlation, with an optimization of 2 %.

Results

Contamination of equipment

At nine of the ten farms, Salmonella was found on the materials used during partial depopulation (Table 3 and Fig. 1). In seven of the nine thinning events it concerned S. Infantis. Almost as frequently isolated (in six thinning events) was S. Paratyphi var. B Java. One exception found during only one thinning event was S. Livingstone. For the Salmonella positive samples, the total prevalence of the serovars Infantis, Paratyphi B var. Java and Livingstone was, 31.3 %, 67.5 % and 1.3 %, respectively. The materials most frequently contaminated, at eight of the ten thinning events, were the crates used from the slaughterhouses. Out of the 320 sampled crates 62 were found contaminated with Salmonella, 27.4 % (17/62) of these were the serovar Infantis and 72.6 % (45/62) Paratyphi B var. Java. Furthermore, the loading surfaces of the trucks were contaminated during five thinning events. In total, 11 of the 22 sampled loading surfaces were Salmonella positive. Here, 27.3 % (3/11) concerned S. Infantis, 63.6 % (7/11) S. Java and 9.1 % (1/11) S. Livingstone. Less frequently contaminated were the wheels of the trucks, the forks of the forklift and the boots of the catching crew. The wheels of the trucks were contaminated at two thinning events, 3 of the 24 samples were positive and these were identified as S. Infantis (2/3) and S. Java (1/3). The forks of the forklift were contaminated at one thinning event, 1 of the 13 samples was positive with S. Infantis. Finally, the boots of the catching crew were contaminated during two thinning events (3 of the 52 samples), the serovars found were Infantis and Java, at respectively 2/3 and 1/3. Gloves of the catching crew were never contaminated. The five farms that tested Salmonella negative before thinning, based on overshoe sampling, were still Salmonella free one week after thinning.

Table 3.

Overview of the sampling results during partial depopulation at ten broiler farms (A-J). The results are shown as number of Salmonella positive samples / total number of samples; the second row of each farm shows number of S. Infantis positive samples / number of Salmonella positive samples.

Farm		Crates	Truck		Forklift		Catching crew
			Surface	Wheels	Forks	Wheels	Gloves	Boots
A	Salmonella	5/8	0/2	0/4	0/2	0/2	0/1	0/4
	S. Infantis	2/5	-	-	-	-	-	-
B	Salmonella	13/32	2/2	0/2	0/1	0/1	-	-
	S. Infantis	0/13	1/2	-	-	-	-	-
C	Salmonella	1/32	0//2	2/2	0/1	0/1	0/1	2/5
	S. Infantis	1/1	-	2/2	-	-	-	2/2
D	Salmonella	5/48	3/3	0/3	0/2	0/2	0/2	0/12
	S. Infantis	1/5	1/3	-	-	-	-	-
E	Salmonella	1/32	2/2	0/2	0/1	0/1	0/1	0/6
	S. Infantis	0/1	0/2	-	-	-	-	-
F	Salmonella	1/32	0/2	0/2	0/2	0/2	0/1	0/6
	S. Infantis	1/1	-	-	-	-	-	-
G	Salmonella	0/32	0/2	0/2	0/1	0/1	0/1	0/4
	S. Infantis	-	-	-	-	-	-	-
H	Salmonella	0/32	0/2	0/2	0/1	0/1	0/1	1/6
	S. Infantis	-	-	-	-	-	-	0/1
I	Salmonella	10/32	1/2	0/2	1/1	0/1	0/1	0/3
	S. Infantis	10/10	1/1	-	1/1	-	-	-
J	Salmonella	26/40	3/3	1/3	0/1	0/1	0/1	0/6
	S. Infantis	2/26	0/3	0/3	-	-	-	-

Fig. 1.

The Salmonella positive samples at ten broiler farm (A-J). The different serovars found were Infantis, Paratyphi B var. Java and Livingstone. The total prevalence of the serovars are, respectively, 31.3 %, 67.5 % and 1.3 %. Farms A, B, C, G and J had a Salmonella positive status before thinning, the other farms (D, E, F, H, I) had a Salmonella negative status. This figure has been designed using resources from Flaticon.com.

Molecular characterization of salmonella

With the help of PFGE, an UPGMA dendrogram was constructed of the Salmonella isolates and the isolates were divided into pulsotypes based on their band patterns. For the serovar Paratyphi B var. Java, seven pulsotypes were distinguished. S. Infantis isolates could be divided into four pulsotypes and only one pulsotype of S. Livingstone was observed (Table 4). This shows that some Salmonella pulsotypes seem to be linked to only one slaughterhouse while others can be linked to multiple slaughterhouses. The other way around, it seems that some slaughterhouses deliver crates with various Salmonella serovars and/or pulsotypes.

Table 4.

The PFGE types of the S. Infantis and S. Java isolates found on thinning materials per farm and respective slaughterhouse. For the types marked with *, the isolates on the thinning material are identical to the pulsotypes isolated from the chickens on the Salmonella positive farms (A, B, C, G and J indicated in bold).

	Slaughter-house(s)	Salmonella Infantis				Salmonella Paratyphi B var. Java
		Type 1	Type 2	Type 3	Type 4	Type 1	Type 2	Type 3	Type 4	Type 5	Type 6	Type 7
Farm A	1		x*			x	x
Farm B	2			x*			x	x	x
Farm C	3	x*
Farm D	4	x				x
	1							x
	5	x
Farm E	5						x	x				x
Farm F	1	x
	6
Farm G	7
Farm H	7									x
Farm I	8		x
Farm J	1				x	x					x

On three farms with a Salmonella positive status before thinning (A, B and C), the S. Infantis pulsotype found on the crates during thinning had an identical pulsotype as one of the S. Infantis types found on the farms. This concerns type 2 for Farm A, type 3 for farm B and type 1 for farm C.

Discussion

During partial depopulation, crates were often found contaminated with Salmonella with a total contamination rate of 19.4 %. These results are not surprising since already in 1980, Rigby et al. (1980) referred to the transport crates as a contamination source. They also demonstrated that the C&D protocol of the crates was failing in eliminating Salmonella. Years later, multiple studies again demonstrated that crates washed at the slaughterhouse remain contaminated with Salmonella, posing a potential risk for contamination (Corry et al., 2002; Marin and Lainez, 2009; Rasschaert et al., 2007; Slader et al., 2002). Although no more recent research was identified, the present work provides evidence that contaminated crates continue to be of relevant concern.

Because of frequently contaminated crates and other materials, the thinning event poses a contamination risk for the broiler houses. For Campylobacter, this has already been thoroughly investigated and it was shown that Campylobacter is introduced during partial depopulation (Hertogs et al., 2021; Smith et al., 2016). However, for the impact of partial depopulation on Salmonella introduction in broiler houses no literature was found. The current research is the first study on the introduction of Salmonella during thinning. In contrast to Campylobacter, no introduction of Salmonella was observed based on the overshoe sampling approximately one week after thinning. One possible explanation is that a higher Salmonella dose is necessary to colonize the older broiler chicken. Another explanation is a higher abundance of Campylobacter in the crates after C&D compared to Salmonella. Campylobacter colonized broiler chickens excrete the bacteria in much higher titers and most likely all animals are colonized and excreting. Although no Salmonella colonization after thinning was observed in this study, attention for biosecurity during thinning is still of importance. Knowing the farmers participated in this study on a voluntary basis means they are probably already quite conscious about biosecurity. For example, it is the responsibility of the farmer that the catching crew enters the broiler houses with clean work- and footwear and that the forklift is disinfected before entering the broiler houses.

Furthermore, contaminated crates not only pose a risk for the broiler farms but also for the slaughterhouses. They can externally contaminate Salmonella free flocks before slaughter and therefore undermine the benefits of logistic slaughter in poultry slaughterhouses (Rasschaert et al., 2008; Zeng et al., 2021). Rigby & Pettit (1980) showed that under simulated transport conditions (5, 19 and 24 hours) chickens exposed to Salmonella in non-cleaned and disinfected crates, can become colonized within 24 hours and begin excreting the bacteria. However, Slader et al. (2002) tested colonization during transport of chickens under commercial conditions (2 hours), but could not demonstrate that this short time period was sufficient to colonize the chickens. McCrea et al. (2006) found no increased Salmonella prevalence after commercial transport compared with Salmonella prevalence in the broiler houses, except when the chickens were exceptionally kept in the crates for 6 to 8 hours before processing at the slaughterhouse. Thus, the duration of transport and the residence time in the crates seem both important factors.

The serovar most frequently isolated was Paratyphi B var. Java (67,5 %), which is consistent with the monitoring results in Belgium that mark Salmonella Java as the most dominant serovar in the broiler sector (DGZ, 2024). A possible explanation for this persistence is the strong biofilm-forming capacity of S. Java (Maes et al., 2020). Salmonella Infantis was isolated in 31.3 % of the positive samples, and it is also the second most abundant serovar in the monitoring results of Belgium (DGZ, 2024). The persistence of this serovar can be explained by the presence of the pESI-like mega-plasmid (plasmid of emerging Salmonella Infantis) that lends S. Infantis an increased resistance to antibiotics, better tolerance to biocides and environmental mercury, an improved biofilm-forming ability and enhanced virulence (Mughini-Gras et al., 2021).

Based on PFGE, the isolates were divided into pulsotypes. These results demonstrated that for S. Java three dominant pulsotypes were present on the crates of different slaughterhouses. For S. Infantis two dominant pulsotypes were found circulating on crates of different slaughterhouses. These results indicate that certain pulsotypes of both serovars are wildly circulating in the Belgian broiler sector. Crates coming from the slaughterhouses are in some cases also contaminated with various strains of S. Infantis and S. Java which indicates cross-contamination occurred at different steps in the production chain e.g. at the slaughterhouse, the truck, the farm, etc. It certainly indicates insufficient C&D of the crates between use for different flocks. Furthermore, some of the isolates found on the crates of different slaughterhouses had identical pulsotypes as the isolates found in the Salmonella contaminated broiler houses (A, B and C). This could imply that crates from the respective slaughterhouses contaminated the farms during a previous depopulation event, but there is not enough evidence to prove this.

Conclusion

From the current results can be concluded that crates pose an important contamination risk for Salmonella because they are frequently contaminated. Other materials used for thinning, such as the trucks, forklifts and boot of the catching crew, were also found to be occasionally contaminated and therefore pose a risk of contamination. Although no cross-contamination towards the remaining chickens was observed in five Salmonella negative farms, biosecurity during thinning is still very important because a link was observed between the S. Infantis types on the crates and the types on the Salmonella positive farms. Furthermore, two dominant S. Infantis pulsotypes and three dominant S. Java pulsotype were found on different farms and crates of different slaughterhouses.

Disclosures

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.