serovars are among the most important agents of food-borne outbreaks throughout the world. Poultry and poultry products are important sources for food-related salmonellosis in humans [1
, with its great diversity, is adapted to a variety of hosts. Environmental niches can be filled by several adapted serovars [2
Java (correctly known as S
. Paratyphi B d-tartrate positive) first appeared in poultry during the 1990s in the Netherlands and Germany, presumably introduced via feed or parent flocks [3
Since then, there has been an increase in the prevalence of S
. Java. In the EU in 2010, S
. Java accounted for 4.6% of the serovars found in broiler meat. It became the third most prevalent serovar in the EU in 2010 due to high prevalences in the Netherlands (53.5%) and Germany (20.7%) [4
]. Furthermore, non-European countries have also reported the incidence of S
. Java in poultry (e.g., Saudi-Arabia [5
] and Bangladesh [6
]). In humans, this serovar leads to gastroenteritis and, to a minor degree, to invasive disease [7
]. In the EU in 2014 it was the 11th most frequently reported serovar of human salmonellosis [8
]. Outbreaks were described by Desenclos et al. [9
] due to contaminated goat’s milk cheese or due to owning an aquarium along with a recent purchase of tropical fish or contaminated alfalfa sprouts [10
]. Human infections in Germany and the Netherlands attributable to the consumption of poultry products were described by Toboldt et al. [12
] and van Pelt et al. [13
] as well as by Brown et al. in Scotland due to consumption of imported poultry [14
]. It is a public health concern because of high occurrence rates of antimicrobial resistance, notably also to antimicrobials considered as critically important by joint experts of the Food and Agriculture Organization (FAO), the World Health Organization (WHO), and the World Animal Health Organization (OIE) (e.g., fluorquinolones or cephalosporines) [15
Barns once infected with S
. Java often become persistently infected with this serotype for the following rearing periods which is contrary to other serovars including S
. Enteritidis or S
. Typhimurium, as the author’s own observations in the poultry veterinary practice have shown. This is a potential risk factor for introducing S
. Java into the downstream food processing plants and to the final consumer, but also to other farms via the same equipment or the same personnel employed for prethinning [18
In 2004, van Immersel et al. [19
] investigated the bacteria-host interaction of S.
Java in vitro and in vivo in the poultry host. S
. Java efficiently invades chicken epithelial cells and macrophages and can be found in caeca, the liver, and the spleen seven days post infection and up until slaughter, thus indicating its good adaptation to the poultry host.
This present study elucidates the infection dynamics and the antimicrobial resistance pattern of S. Java in a broiler flock under field conditions over three consecutive rearing periods. The acquired knowledge will help to better understand the bacteria-host interaction in the field and with it, benefits to develop intervention strategies as well as monitoring systems for the control of this serovar, thereby having an impact on public health.
The data indicate a great variation per time-point in the load of S.
Java in the caecal content of broilers sampled within a persistently infected flock as shown in the great range of 95% confidence intervals. This results in a heterogeneous colonisation of chickens with S.
Java regarded per time-point, except in the case of antibiotic treatment prior to sampling (compare results of time-point 5 in rearing period one when there is no range of 95% confidence interval). The data suggest a peak of colonisation at time-point 4 compared to time-points 3 and 5, both confirmed by direct plating and qPCR. Chickens have a greater chance of being positive at time-point 4 which results in more S.
Java positive chickens slaughtered at prethinning (around day 30) than at 40 days of age. This poses a greater risk of contamination of the slaughtering line with S
. Java. If we further consider that chickens slaughtered at prethinning are intended to be merchandised unprocessed, this might additionally increase the exposure of the final consumer. Furthermore, during prethinning, equipment and personnel are brought to the farm, thus posing a risk of additional S
. Java contamination on this farm as well as on other farms, as equipment and personnel are often employed on more than one farm [18
]. The variation in the individual Salmonella
status of broilers should be considered when samples are taken at slaughter for risk assessment, as was previously shown by Hansson et al. [25
] for Campylobacter
Van Immerseel et al. [19
] demonstrated that one week after experimental infection all chickens were positive for S.
Java, this being confirmed with cloacal swabs directly plated onto BG agar. Animals infected by a seeder model were all positive after two weeks post infection, this also being confirmed by cloacal swabs that were directly plated onto BG agar. At slaughter age (six weeks) all caecal samples were positive on BG agar, but only after enrichment. These results are in accordance with our data which also suggest a decrease in positive results by direct plating the older the animals become.
Day-old chicks are always free of S
. Java. It is therefore of utmost importance to house chickens in a S
. Java-free environment. The studied barn was cleaned and disinfected according to standard protocols which were insufficient to eradicate S
. Java. Therefore, there is always a risk of newly housed chickens becoming infected. It is well known that S.
Java persists in broiler houses for a long time and is therefore the source of infection of day-old chicks in the following rearing period. The data in Table 2
indicate that the studied broiler house was still contaminated with S.
Java after cleaning and disinfection before arrival of the new birds. The fact that 7 out of 15 sampling points were positive for S
. Java indicates a high level of contamination, although the level of contamination might vary between different rearing periods. We demonstrate in another paper that only routine cleaning and disinfection alone would not be appropriate for eradication of S
. Java from the farm [20
]. At least two disinfection steps were required for successful eradication of S
. Java on the farm. We implemented a feasible risk-orientated hygiene analysis examining individual critical control points of the barn for successful eradication of the pathogen. New chickens were not housed before all sampling points had been tested negative in qPCR for Salmonella
occurrence. Prethinning, as a risk factor for reintroduction of S
. Java, was abolished.
The studied farm has been positive for S. Java for more than five years. For that reason it is more than likely that the birds in the following rearing period become Salmonella positive. The detected differences in the caecal colonisation of the birds in the different rearing periods might be due to the different levels of contamination in the broiler houses when the birds arrived at the farm. On the other hand, young chickens are very susceptible to infection with Salmonella in general. At two weeks of age, when we first performed a quantitative examination of S. Java, incidences of S. Java might be high due to the excretion of S. Java by birds being first infected due to contamination of the environment. This fact results in further infection of other birds, comparable to a seeder model. The detailed quantitative level of contamination in the houses was not tested as the extra costs are not justifiable for routine testing. Presently, the most important aim to further reduce the contamination of poultry products with Salmonella organisms, and therefore to reduce the risk for humans, is to transfer day-old broilers to a Salmonella-free environment, as in most cases newly hatched chicks are free of Salmonella organisms. This demand is absolutely essential for further control of Salmonella spp. in poultry.
During rearing period one, animals received a combined antimicrobial therapy with amoxicillin and colistin which was stopped three days prior to sampling. The therapy seemed to result in a uniform colonisation of chickens with S.
Java (compare Table 4
and Figure 1
) at a low level, but was not capable of eliminating the pathogen from the population. Antimicrobial therapy can reduce the transmission and excretion of S.
Java, perhaps below the level of detection, but sustains the persistence of S
. Java. This might even result in a reduced diagnostic sensitivity of microbiological methods without enrichment used in this study when samples are taken after antimicrobial treatment and might therefore interfere with the diagnosis of S
. Java. Antimicrobial treatment is no alternative to good hygienic practices and biosecurity measures that might lead to the elimination of S.
Java. Antimicrobial treatment of broilers also poses a risk of carcass contamination with resistant S.
Java as well as resistant commensal bacteria [26
After antimicrobial treatment and subsequent sampling in trial 1, positive results were only obtained with qPCR as there had been no detection of bacterial growth on BG agar. On the other hand, there might have been a high level of competing commensal flora that may complicate microbiological diagnosis. Therefore, we used qPCR to obtain a qualitative result concerning the S
. Java status of animals. For qPCR, it has been proven by using artificial culture mixes and after enrichment that background microorganisms have no influence on the detection limit (3 CFU/mL) of Salmonella
During recent years there has been a rise in awareness concerning the selection of antimicrobial resistance among food-producing animals and its potential risk concerning public health [27
]. Between 1960 and 1993 there was a great diversity among isolates of S.
Java. However, since 1994 multidrug-resistance and the predominance of one single clone have been obvious [28
]. Strains of this group emerged due to the increased use of antimicrobials to manage the S.
Enteritidis crisis at the beginning of the 1990s. They have no integron class 1, but multidrug-resistance is due to the acquisition of class-2 integrons in the bacterial chromosome. The core spectrum of resistance is directed against trimethoprim, spectinomycin, and streptomycin and is mediated by the gene cassette dfrA1-sat1-aadA1 carried by transposon Tn7
]. This gene cassette also seems to be expressed by the studied isolates shown by antimicrobial susceptibility testing but needs further confirmation by molecular methods. Occurrence of BlaTEM
-genes in S.
Java have been described by Hasman et al. [29
] in Dutch poultry and meat thereof as well as occurrence of BlaCTX
-genes and BlaTEM
-genes by Rodriguez et al. [30
] in ceftiofur-resistant Salmonella
enterica isolates from chickens and chicken meat in Germany. They state that extended spectrum β-lactamase (ESBL) activity as well as AmpC-β-lactamase activity increase in S.
Java due to a variety of genetic mobile elements capable of horizontal gene transfer leading to a new onset of a public health problem. Isolates investigated in Belgium from 2008–2010 showed a plasmid-borne resistance towards ESBL antimicrobials. Even though these antimicrobials are not indicated for use in poultry there was a resistance selection as plasmids conferring ESBL resistance often have other resistance genes (e.g., against sulphonamides or trimethoprim) and these antimicrobials are frequently used in the veterinary poultry sector [17
]. Van Pelt et al. [13
] found that S
. Java becomes less sensitive to ciprofloxacin, this also being confirmed in our study. It is important for the therapy of invasive Salmonella
infections in adults [31
]. The interpretation of occurring antibiotic resistance is somewhat difficult as ECOFFs are not always established for S.
Java. For example, the minimum inhibitory concentration for cefotaxime is exactly the same as the published ECOFF for Salmonella
spp. It is a sensitive substrate for confirming CTX-M enzymes, which are the most common ESBL enzymes. On the other hand, cefoxitin is a sensitive, but less specific parameter for confirming AmpC-producing enterobacteriaceae [27
]. One point to consider is the intermediate resistance profile of S.
Java to ampicillin in the agar diffusion test that was not confirmed in the microdilution test, but might be further evidence for the expression of AmpC enzymes. Thus, additional molecular tests for the presence of ESBL and/or AmpC enzymes are needed to affirm their occurrence in the studied isolates. Resistance profiles might be misinterpreted because of missing established ECOFFs for S
. Java or varied results in the agar diffusion and microdilution, test as is the case with sulfamethoxazole/trimethoprim in this study. Further molecular methods and epidemiological data for interpreting and establishing ECOFFs are required for precise interpretation of the resistance profile of the studied isolate. In the agar diffusion test, S
. Java was found to be resistant to sulfamethoxazole/trimethoprim as well as to trimethoprim in the microdilution test. For sulfamethoxazole, the three tested isolates ranged from minimal inhibitory concentrations of 16µg/mL, 32µg/mL, and 64µg/mL. No ECOFF for S. Java and sulfamethoxazole has been established so far, making interpretation of results difficult. Schroeter et al. [32
] used a breakpoint of 512 µg/mL to assess resistance of Salmonella
species to sulfamethoxazole in their study. In this case, the studied isolate was not resistant to sulfamethoxazole. Additionally, resistance to sulfamethoxazole is mediated by Salmonella
Genomic Island I (SGI I) [33
Java comprises two different lineages that either express the O:5 antigen or do not [34
]. O:5 positive strains possess further virulence genes that allow it to interact with a greater range of hosts, whilst O:5 negative strains have the ability to persist in poultry in the Netherlands, Germany, and Belgium and possess less virulence, but are multiresistant. O:5 positive strains express SGI I that usually confers resistance to sulfamethoxazole. The studied isolates seem to belong to the group of O:5 negative less invasive strains which persist in poultry and exhibit a range of antimicrobial resistance, but do not express SGI I [34
Another serovar persistent in poultry in France, S
. Senftenberg, shows a low capacity to invade chicken enterocytes and to pass the intestinal wall [35
]. In contrast, S
. Java invades chicken enterocytes and macrophages. It can be found in caeca, the liver, and the spleen of infected broilers and is able to spread efficiently within chickens [19
]. Nonetheless, the invasiveness of Salmonella
and its virulence are not always correlated [36
]. The ability to colonise a particular host is diverse and involves the host (e.g., age and breeding differences), the serovar, and extrinsic pressures (e.g., antimicrobial therapy, stress, coccidiosis), and this interplay leads to the observed variety in host ranges observed in different countries [37