general biosecurity measures

General hygienic and biosecurity measures should be part of the overall management plan of poultry producing companies. These measures are of utmost importance to control infections and all other measures lose their efficacy when the overall biosecurity plan is not functioning.

Introducing Salmonella-free animals is the crucial starting point to prevent infections. Incoming birds should have a high health status and should be purchased from reliable suppliers that have a quality assured breeding and hatching facility. Furthermore, Salmonella can be introduced on the broiler farm also by vehicles, people, clothing, footwear, equipment, water, feed, litter, insects, rodents, wild birds, pets, utensils and many more factors. Water and feed decontamination and insect and vector control are described below in detail. For all incoming living or non-living material, one needs to be aware of the risks for Salmonella entry on-farm. Controlled access is a key issue: only essential visitors, wearing protective clothing and disinfected boots should be allowed on-farm. Also, personnel should be aware of basic hygienic principles. Simple measures such as foot baths, hand hygiene, and minimizing movement between different animal houses need to be implemented properly. Hygienic barriers, including anterooms, should be present. In an overall management plan, cleaning and disinfection after each round must be routinely performed, as described below. The efficacy of the disinfection of the chicken houses needs to be tested by sampling of the floors, walls, drinking water, feed trays, and the environment.

cleaning and disinfection

The lack of a proper cleaning and disinfection program has been shown in multiple studies to constitute a risk factor for Salmonella contamination in poultry flocks. Cleaning and disinfection generally aims at reducing the number of micro-organisms in the poultry house. Basically, the procedure of a good cleaning and disinfection program contains the following elements (Morgan-Jones, Reference MORGAN-JONES1987): 1) cleaning (dry followed by wet cleaning with detergents) to remove all gross dirt and render surfaces visibly clean; 2) disinfection to kill micro-organisms; 3) rinsing to clear residues of disinfectants and 4) fumigation. Details of the procedure are out of the scope of this review but it is clear that inadequate cleaning and disinfection can lead to a build-up of pathogens in the poultry house. Most widely used disinfectants in broiler houses are aldehydes, peroxides, quaternary ammonium compounds (also in detergents) and phenolic compounds. Cleaning and disinfection have been documented to be able to reduce the Salmonella contamination in broiler houses significantly (Garber et al., Reference GARBER, SMELTZER, FEDORKA-CRAY, LADELY and FERRIS2003), but the efficacy is variable and depends on the accuracy of the procedure, technical issues and correct use of products (Davies and Breslin, Reference DAVIES and BRESLIN2003a). Wales et al. (Reference WALES, BRESLIN, CARTER, SAYERS and DAVIES2007) sampled 10 broiler houses, in which Salmonella positive flocks had been housed, after depopulation but before and after cleaning/disinfection. They showed that in nine out of ten houses, the prevalence of positive samples dropped significantly following cleaning and disinfection. Unfortunately many poultry producers either do not clean and disinfect or either just clean without disinfection or just disinfect without cleaning (Rose et al., Reference ROSE, MARIANI, DROUIN, TOUX, ROSE and COLIN2003). Cleaning without disinfection will not remove Salmonella and disinfecting without cleaning will not be efficient as the bacteria can be protected by the organic material. In a study involving 85 broiler flocks in France, it became apparent that dry cleaning (removal of manure before wet cleaning) was not a common practice (Rose et al., Reference ROSE, MARIANI, DROUIN, TOUX, ROSE and COLIN2003). Dust removal was not done in 64 of the 85 studied broiler farms, and cleaning of feeders was not done in 61 out of the 85 farms (while they were disinfected in 66 farms). Floors were not disinfected in 15 out of 85 farms (Rose et al., Reference ROSE, MARIANI, DROUIN, TOUX, ROSE and COLIN2003). In a study in the UK using floor sweepings, 25 and 42% of analyzed samples were found Salmonella positive in 2 Salmonella infected broiler farms (Davies et al., Reference DAVIES, BRESLIN, CORRY, HUDSIN and ALLEN2001). The number of positive samples was reduced to 1 and 6% after cleaning and disinfection, respectively. Also in the study of Wales et al. (Reference WALES, BRESLIN, CARTER, SAYERS and DAVIES2007), there was a high variation in effectiveness of the cleaning/disinfection programs between farms. It is suggested by the authors that a plausible explanation for this may be variation in the techniques of organic matter removal and in application of the disinfectants can be. Use of disinfectants at insufficient concentrations (Davies and Wray, Reference DAVIES and WRAY1996; Wales et al., Reference WALES, BRESLIN, CARTER, SAYERS and DAVIES2007) and poorly cleaned and inconsistently disinfected key areas, such as drinkers, were important shortcomings (Wales et al., Reference WALES, BRESLIN, CARTER, SAYERS and DAVIES2007). Also, failure to clean and disinfect air inlets and fans is an important factor for continued contamination of broiler houses (Higgins et al., Reference HIGGINS, MALO, RENÉ-ROBERGE and GAUTHIER1982; Rose et al., Reference ROSE, MARIANI, DROUIN, TOUX, ROSE and COLIN2003). These are hot spots for contamination of Salmonella in broiler houses (Davies et al., Reference DAVIES, BRESLIN, CORRY, HUDSIN and ALLEN2001). Not using a detergent for cleaning, and thus the inability to remove all organic material, was also identified as a risk factor for Salmonella infection of broiler flocks (Cardinale et al., Reference CARDINALE, TALL, GUEYE, CISSE and SALVAT2004).

insect control

The environment in broiler houses typically is very attractive to insects. A warm, moist atmosphere and litter supports multiplication of various insect species, including flies and beetles. Insects act as reservoirs and vectors for Salmonella in poultry farms. It has been known for many years that insects can carry Salmonella and other pathogenic bacteria (Harein et al., Reference HAREIN, DE LAS CASAS, POMEROY and YORK1970; Geissler and Kösters, Reference GEISSLER and KÖSTERS1972). Feeding S. typhimurium-inoculated poultry feed to lesser mealworms (Alphitobius diaperinus) led to external and internal colonization, and even transstadial transmission of the bacteria (McAllister et al., Reference MCALLISTER, STEELMAN and SKEELES1994). In Salmonella contaminated poultry farms, the organism can often be isolated from insects. S. enteritidis positive cultures were obtained from housefly (Musca domestica) pools from infected commercial chicken farms in some studies (Kinde et al., 1995; Mian et al., Reference MIAN, MAAG and TACAL2002). In caged-layer houses that had produced eggs implicated in human outbreaks, Salmonella was found in pools of houseflies, dump flies (Hydroteae aenescens) and lesser mealworms (Olsen and Hammack, Reference OLSEN and HAMMACK2000). On a free-range breeding farm which had been depopulated following identification of S. enteritidis in the chickens, adult and larval ground beetles and centipedes were found positive (Davies and Breslin, Reference DAVIES and BRESLIN2003b). Also the hairy fungus beetle (Typhaea stercorea) can carry Salmonella. In a Danish broiler farm in which a S. infantis infection already persisted for six rounds, the bacterium was isolated from the beetles in the period before stocking (Hald et al., Reference HALD, OLSEN and MADSEN1998). In the empty period, beetles were found positive for S. indiana in a broiler flock that was positive for the same strain in two consecutive rounds (Skov et al., Reference SKOV, SPENCER, HALD, PETERSEN, NAUERBY, CARSTENSEN and MADSEN2004). The insects can contract the infection from the environment and from the birds. Of all houseflies released into rooms housing S. enteritidis contaminated hens, 50% became contaminated within 48 h,, both on the exterior surface and in the gut (Holt et al., Reference HOLT, GEDEN, MOORE and GAST2007). Transmission from the insects to the birds requires ingestion of contaminated material or of the insects themselves by the birds. Whether insects can be the principal infection source, i.e. can be a single significant vector infecting the chickens, is difficult to investigate. Feeding contaminated houseflies and beetles to chickens can lead to colonization of the animals (Hald et al., Reference HALD, OLSEN and MADSEN1998; Holt et al., Reference HOLT, GEDEN, MOORE and GAST2007). Also feeding of poultry red mites (Dermanyssus gallinae), infected in vitro by S. enteritidis through a blood meal or cuticular contact, resulted in colonization of chicks (Moro et al., Reference MORO, FRAVALO, AMELOT, CHAUVE, ZENNER and SALVAT2007). Releasing Salmonella contaminated flies in a room containing uninfected hens, however, did not result in colonization (Holt et al., Reference HOLT, GEDEN, MOORE and GAST2007). Also release of contaminated beetles/larvae to a group of pullets did not result in infected birds (Davies and Breslin, Reference DAVIES and BRESLIN2003b). Although insects can carry Salmonella and contaminated insects can infect chickens when orally administered, the extent of transmission and the level of contamination caused by infected insects in the field is not completely clear. Conversely however, contaminated insects are certainly an indication of chicken house and environmental contamination by Salmonella.

Insect control nevertheless should be implemented in an overall management plan. For the control of flies, emphasis should be on avoiding wet manure, to make it an unattractive breeding place. Physical control includes screens for doors and windows, electrical insect traps and bait stations. When all other measures are not successful, chemical control using insecticides and larvicides can be applied. Given the potential of carrying and transmitting Salmonella to the animals (including flocks in neighbouring houses) insect control is to be part of any Salmonella control plan.

rodent control

Rodents can carry Salmonella and other pathogens. Chickens can be infected by contact with droppings from experimentally infected mice (Davies and Wray, Reference DAVIES and WRAY1995). On Salmonella contaminated poultry farms, rodents often are carriers of Salmonella. A proof of correlation between rodent contamination and chicken infection was given by Henzler and Opitz (Reference HENZLER and OPITZ1992), who sampled 715 mice and rats on 10 rodent-infested poultry farms, 5 of which were Salmonella positive based on analysis of environmental samples. While Salmonella was not detected in mice from non-contaminated farms, 24% of the mice on Salmonella-positive farms were positive. Observation of rodents on-farm was also identified as a risk factor for Salmonella persistence after decontamination in French broiler houses (Rose et al., Reference ROSE, BEAUDEAU, DROUIN, TOUX, ROSE and COLIN2000). Whether the presence of Salmonella in rodents caught in poultry houses where flocks with a Salmonella positive status are housed implies a causal relation is not clear. The overall hygienic status of rodent infested farms and the biosecurity management plan (e.g. general cleaning and disinfection) of these farms also need consideration, as it could be that farms with poor hygienic standards have more rodent problems.

Rodents typically are present in places where they can find plenty of feed and nesting spaces, as can occur on poultry farms. Although there are some differences between mice and rat prevention and control measures (and even between different rat species), some general principles can apply to all rodents. Preventing food and water supply and nesting places is crucial to rodent control. Preventing access of rodents to the building (e.g. holes in insulation, double walls et cetera), and maintenance of the building are important. Rodents like to hide away, and thus all materials should be properly stored in a separate room in the building, to avoid hiding places. The buildings should ideally be provided with lights on the outside walls and no materials or vegetation should be in close proximity of the outside walls. All these aspects are part of an integrated management plan of the poultry farm. Finally, when a rodent infestation is present despite all applied prevention measures, rodenticides can be applied to reduce the population.

drinking water decontamination

Litter, feed, dust and residual contamination of the drinkers can result in water contamination. Drinking water is a continuous risk for re-infection of Salmonella-contaminated flocks mainly due to the presence of faecal material in the drinkers. It is clear that organic material in the drinking water affects the efficacy of water decontamination treatments. Therefore frequent cleaning of the drinkers is a prerequisite for drinking water decontamination treatments to be effective. Acetic, lactic, formic and other organic acids are used as drinking water additives and these acids also have effects in the crop of the birds. It is therefore difficult to discriminate between in-water and in-bird effects. Since most acids exert antibacterial activity against Salmonella, it is clear that in clean drinking water, they can suppress Salmonella to some extent. Chlorination can also be used for drinking water sanitation. Although chlorination of drinking water gave a variable suppression of the number of Salmonella bacteria in the drinking water in chickens experimentally inoculated with Salmonella, this did not lead to a reduction in gut colonization (Poppe et al., Reference POPPE, BARNUM and MITCHELL1986). In a study where ‘natural’ infections were recorded, Al-Chalaby et al. (Reference AL-CHALABY, HINTON and LINTON1985) evaluated a commercial product containing propionic acid. The acid eliminated Salmonella from the drinking water, while more than 80% of the samples in control groups were positive. This did not influence the presence of Salmonella in litter samples, cloacal swab and caecal content samples.

feed decontamination

Commercial poultry feed can be contaminated with Salmonella. An early study of Hacking et al. (Reference HACKING, MITCHELL and CARLSON1978) analyzed 93 finished pelleted broiler feeds and found 4 of them contaminated with Salmonella. Recently, ten out of 111 pelleted broiler feed samples were found to be Salmonella positive by Bucher et al. (Reference BUCHER, HOLLEY, AHMED, TABOR, NADON, NG and D'AOUST2007). Feed ingredients are an important source of contamination of the feed. While Hacking et al. (Reference HACKING, MITCHELL and CARLSON1978) found animal protein to be a source of contamination, Jones and Richardson (Reference JONES and RICHARDSON2004) isolated Salmonella from corn, cottonseed meal, fish meal, soybean meal and wheat middlings. Veldman et al. (Reference VELDMAN, VAHL, BORGGREVE and FULLER1995) found 31% of 130 fish meal samples to be contaminated, as opposed to 4% of 83 meat and bone meal samples, 2% of 58 tapioca samples and 27% of 15 maize grit samples. Feed contaminated with Salmonella was regarded as an important source for Salmonella contamination in broilers as early as the 1950's, as reviewed by Williams (Reference WILLIAMS1981). The potential transmission of Salmonella strains from feed origin to the live animals and finally to the carcasses in the field was already proven in the late 70's (MacKenzie and Bains, Reference MACKENZIE and BAINS1976). Infection of broiler chickens may start from the immediate post-hatch period. Less than one colony forming unit per gram of feed has been shown to lead to colonization of young chickens in an experimental setup (Schleifer et al., Reference SCHLEIFER, JUVEN, BEARD and COX1984). Different methods can be used to eliminate Salmonella from feed. Steam pelleting is considered to be effective in killing Salmonella in feed. Although results are highly variable, depending on the temperature and moisture content of pellets, in most studies pelleted feeds are less contaminated than mash feeds. Cox et al. (Reference COX, BAILEY, THOMSON and JUVEN1983) detected Salmonella in 58% of mash feed samples compared to zero per cent in pelleted feed samples. Pelleting, however, does not preclude Salmonella contamination of the finished products. In a recent study of Veldman et al. (Reference VELDMAN, VAHL, BORGGREVE and FULLER1995), 21% of mash feed samples and 1.4% of pelleted feed samples were Salmonella positive. Jones and Richardson (Reference JONES and RICHARDSON2004) observed a decrease in contamination rate only from 8.8 to 4.2% after pelleting. These authors investigated different pelleting temperatures and concluded that at least 85°C should be used. The presence of Salmonella contaminated dust around the pellet mill can negate the positive effect of pelleting (Jones and Richardson, Reference JONES and RICHARDSON2004). It is suggested that dust within the pellet coolers is a major contamination source, and that pathogens can grow within the pellet coolers, a process correlated with moisture condensation in the coolers.

Finished poultry feeds as well as feed ingredients also can be decontaminated by irradiation. The ionizing radiations employed most frequently for irradiating feed components and finished poultry feed are gamma rays. The initial contamination influences the efficacy of the treatment. Elimination of Salmonella can be achieved at radiation doses of about 5 kGy (Leeson and Marcotte, Reference LEESON and MARCOTTE1993a). This dose apparently does not induce significant chemical changes in the feed. Leeson and Marcotte (Reference LEESON and MARCOTTE1993a) have reviewed irradiation of poultry feed to control pathogens and suggest that doses between 10-40 kGy are effective in killing micro-organisms when feed bags are treated, but the dose to eliminate Salmonella will, depending on the initial contamination, be between 5-25 kGy. Doses of about 25 kGy will lead to loss in potency of all fat-soluble vitamins, and to peroxidation of fats in the feed (Leeson and Marcotte, Reference LEESON and MARCOTTE1993b).

Another way to reduce Salmonella in feed is the inclusion of organic acids. Acidic compounds were first introduced for the control of Salmonella in the late 1960's, and mainly focused on decontamination of carcass meal. Khan and Katamay (Reference KHAN and KATAMAY1969) evaluated the efficacy of 32 different acid preparations to decontaminate bone meal, and showed that low molecular weight volatile fatty acids were the most promising. Feed inoculated with high doses of S. typhimurium showed a 1000-fold decrease in Salmonella count over 7 days when treated with a mixture of formic and propionic acid (Iba and Berchieri, Reference IBA and BERCHIERI1995). Mixtures of formic and propionic acid were also effective when feed was artificially inoculated with low doses of S. kedougou, and the decrease was most obvious after several weeks of storage (Hinton and Linton, Reference HINTON and LINTON1988). Humphrey and Lanning (Reference HUMPHREY and LANNING1988) showed that the number of Salmonella positive breeder feed samples decreased from 4.1 to 1.1% after the feed was supplemented with 0.5% formic acid. The antibacterial activity of organic acids is dependent on the temperature and moisture, and although the above mentioned studies show positive effects, the efficacy in dry feed can be questioned. The original concept of incorporating acids into feed was based on the notion that the acids would decontaminate the feed itself and prevent Salmonella uptake by the chickens. Later on, it was realized that addition of acids in-feed also exerts effects in the crop (Thompson and Hinton, Reference THOMPSON and HINTON1997) and even in the lower gastro-intestinal tract when the acids are coated on or impregnated in carriers. Although pelleting, irradiation and inclusion of acids can clearly help to control Salmonella in feed, one needs to be constantly aware of the possibility of recontamination of the finished feed during storage.

antibiotics

For many years, antibiotics have been added to broiler feed for their growth promoting effects. Antimicrobial growth promoters have been banned in poultry feed in the EU starting from 2005 (Regulation 1831/2003). In other countries, these substances are still used but their use in under debate. Prophylactic and curative use of antibiotics to control Salmonella is not recommended for different reasons. Firstly, antibiotic resistant Salmonella (and other) strains have emerged (Van Duijkeren et al., Reference VAN DUIJKEREN, WANNET, HOUWERS and VAN PELT2003; Bywater, Reference BYWATER2004). Secondly, there is a concern about the presence of antibiotic residues in meat. Thirdly, most antibiotics fail to eliminate Salmonella from the animals although shedding is (temporarily) decreased. This leads to the induction of carrier birds and thus infected flocks can erroneously be classified as Salmonella negative during bacteriological monitoring of faecal samples. This was clearly shown in an early study by Smith and Tucker (Reference SMITH and TUCKER1975), in chickens orally infected with S. typhimurium. The chickens were fed diets containing ampicillin, neomycin, chloramphenicol, oxytetracycline, polymyxin, spectinomycin, sulphadiazine or furazolidone. Shedding was strongly reduced but when the antibiotics were withdrawn, shedding increased to the same level as in chickens fed an unsupplemented diet (Smith and Tucker, Reference SMITH and TUCKER1975). In the EU, antimicrobials cannot be used as a specific method to control Salmonella in poultry (Regulation 1177/2006). It is clear that antibiotics are not the primary choice to control Salmonella in broiler flocks. Studies on the efficacy of antibiotics on Salmonella control will therefore not be discussed in this review. An overview of the effect of antimicrobial growth promoters on Salmonella colonization in chickens is given in an earlier paper (Van Immerseel et al., Reference VAN IMMERSEEL, CAUWERTS, DE VRIESE, HAESEBROUCK and DUCATELLE2002).

acidic compounds

Acidic compounds used as feed additives are short-chain and medium-chain fatty acids. As already mentioned above, the original concept of incorporating these acids into feed was based on the opinion that the acids would decontaminate the feed itself and prevent Salmonella uptake by the chickens. Later on, it was realized that addition of acids in-feed also exerts effects in the crop and the gastro-intestinal tract of the animal. Numerous studies have been carried out using different organic acids supplemented in the feed or drinking water, in infection models in which different Salmonella serotypes were used for direct inoculations in the crop or in feed inoculations (reviewed by Van Immerseel et al., Reference VAN IMMERSEEL, RUSSELL, FLYTHE, GANTOIS, TIMBERMONT, PASMANS, HAESEBROUCK and DUCATELLE2006). It is not the aim of the current review to summarize all these data but it is clear that, depending on the experimental protocol, organic acid incorporation in the feed or drinking water can have effects on colonization and shedding when the bacteria are inoculated in the crop. Since the effects of powder form acidic products are based on direct antimicrobial activity, one can choose a compound based on the anti-Salmonella activity, combined with some other considerations, such as cost, sensory characteristics (taste, smell) and effects on performance. At low environmental pH, organic acids can pass through bacterial membranes and dissociate in the cytoplasm to yield anions and protons. Anion accumulation in the bacterial cytoplasm has been shown to be at least partly responsible for the bactericidal effects of acids (Russell and Diez-Gonzalez, Reference RUSSELL and DIEZ-GONZALEZ1998). Factors such as chain length, side chain composition, pKa values and hydrophobicity all affect the antimicrobial activity. Medium chain fatty acids (C6 to C12; caproic, caprylic, capric and lauric acid) appear to be much more effective against Salmonella than the short-chain fatty acids (formic, acetic, propionic and butyric acid). As little as 25mM of the medium-chain fatty acids caproic, caprylic and capric acid were bacteriostatic to a S. enteritidis strain, but the same strain tolerated 100mM of short-chain fatty acids (Van Immerseel et al., Reference VAN IMMERSEEL, DE BUCK, MEULEMANS, PASMANS, VELGE, BOTTREAU, HAESEBROUCK and DUCATELLE2003; Reference VAN IMMERSEEL, DE BUCK, BOYEN, BOHEZ, PASMANS, VOLF, SEVCIK, RYCHLIK, HAESEBROUCK and DUCATELLE2004a). Some long-chain fatty acids (C14:0, C18:1 and C18:2) acids were shown not to be bacteriostatic (Sprong et al., Reference SPRONG, HULSTEIN and VAN DER MEER2001). Although medium chain fatty acids have the greatest antibacterial activity against Salmonella, this is not the only criterion to decide which organic acid to include in feed.

In powder form, organic acids (in feed or drinking water) have direct antimicrobial effects in the upper part of the gastro-intestinal tract (next to effects in feed and drinking water, see above). Indeed, due to absorption of the acids the lower gastro-intestinal tract is not targeted. Unfortunately, the lower intestinal tract is the primary colonization site of Salmonella (Desmidt et al., Reference DESMIDT, DUCATELLE and HAESEBROUCK1998). Recently, researchers have attempted to transport the organic acids further down the gastrointestinal tract by coating or micro-encapsulation, which should prevent absorption of the acids in the upper tract and ensure release further down in the gastro-intestinal tract. Different formulations are on the market, and depending on the coating and material used to produce the microbeads, the release of the acids takes place in different segments of the gastro-intestinal tract. Van Immerseel et al. (Reference VAN IMMERSEEL, FIEVEZ, DE BUCK, PASMANS, MARTEL, HAESEBROUCK and DUCATELLE2004b) studied the effect of microbeads containing formic, acetic, propionic and butyric acid on colonization of S. enteritidis in caeca, liver and spleen. Animals were infected (day 5 post-hatch) with 5 x 10³ cfu S. enteritidis and samples were taken 3 days post-infection. Caecal colonization was significantly increased when acetic acid was added to the feed, but decreased when butyric acid was added. When powder and coated butyric acid additives (0.63 g/kg butyric acid) were compared using the same infection protocol, the coated form decreased colonization of the caeca, but the powder form did not (Van Immerseel et al., Reference VAN IMMERSEEL, METHNER, RYCHLIK, NAGY, VELGE, MARTIN, FOSTER, DUCATELLE and BARROW2005). In an infection study using a seeder model, in which broilers were infected at day 5 post-hatch with 10⁵ cfu S. enteritidis and housed together with non-inoculated broilers, 0.63 g/kg coated butyric acid in the feed significantly reduced shedding of S. enteritidis in all broilers until slaughter age. A possible explanation of the effects of butyric acid on caecal colonisation is that invasion of Salmonella into epithelial cells after contact with butyric acid is decreased (Gantois et al., Reference GANTOIS, DUCATELLE, R, PASMANS, HAESEBROUCK, HAUTEFORT, THOMPSON, HINTON and VAN IMMERSEEL2006a). The effects of organic acids on epithelial cell invasion can be explained by changes in expression of genes of the Salmonella Pathogenicity Island I (SPI-1). Whole genome microarray analysis of both S. typhimurium and enteritidis has indicated that low doses of butyric acid downregulate SPI-1, while they do not alter metabolic gene expression (Gantois et al., Reference GANTOIS, DUCATELLE, R, PASMANS, HAESEBROUCK, HAUTEFORT, THOMPSON, HINTON and VAN IMMERSEEL2006a). S. typhimurium and enteritidis carrying a plasmid-borne hilA::luxCDABE transcriptional fusion confirmed the idea that butyrate down-regulated hilA, the key regulator of SPI-1 (Gantois et al., Reference GANTOIS, DUCATELLE, R, PASMANS, HAESEBROUCK, HAUTEFORT, THOMPSON, HINTON and VAN IMMERSEEL2006a). Expression of hilA has been shown to be crucial for caecal colonization (Bohez et al., Reference BOHEZ, DUCATELLE, PASMANS, BOTTELDOORN, HAESEBROUCK and VAN IMMERSEEL2006), and thus butyric acid mediated decreases in hilA expression appear to be responsible for the decreased colonization. Acetic acid is known to increase SPI-1 gene expression and it seems to be the case that the concentration ratio of acetic to butyric acid can be crucial for the invasive and colonizing characteristics of Salmonella (Lawhon et al., Reference LAWHON, MAURER, SUYEMOTO and ALTIER2002). Also effects of butyric acid on the gut morphology could play a role. Indeed, butyric acid increases villus length in pigs and decreases inflammation in the gut in humans (Wachtershauser and Stein, Reference WACHTERSHAUSER and STEIN2000; Kotunia et al., Reference KOTUNIA, WOLINSKI, LAUBITZ, JURKOWSKA, ROMÉ, GUILLOTEAU and ZABEILSKI2004; Kien et al., Reference KIEN, BLAUWIEKEL, BUNN, JETTON, FRANKEL and HOLST2007). Furthermore, effects of the released butyric acid on the gut microbiota in general could also play a role.

Studies on these effects of butyric acid in poultry are scarce. Studies on the use of medium-chain fatty acids as coated product are not well documented in the literature, but they could hold promise. These acids also decrease SPI-1 gene expression, at much lower concentrations compared to butyric acid (Van Immerseel et al., Reference VAN IMMERSEEL, DE BUCK, BOYEN, BOHEZ, PASMANS, VOLF, SEVCIK, RYCHLIK, HAESEBROUCK and DUCATELLE2004a). Whether these effects are specific or are the result of other disturbances in the Salmonella proteome is not known.

prebiotics

Prebiotics are non-digestible feed ingredients that beneficially affect the host by selectively stimulating the growth and/or activity of one or a limited number of bacterial species already resident in the colon, and thus attempt to improve host health (Gibson and Roberfroid, Reference GIBSON and ROBERFROID1995). For a dietary substrate to be classified as a prebiotic, at least 3 criteria are required: 1) the substrate must not be hydrolyzed or absorbed in the stomach or small intestine, and thus can reach the lower intestinal tract; 2) the prebiotic compound must be selective for beneficial bacteria in the lower gastro-intestinal tract and 3) fermentation of the substrate should induce beneficial luminal/systemic effects within the host (Manning and Gibson, Reference MANNING and GIBSON2004). By definition, all prebiotics are oligo- or polysaccharides. Indeed, monosaccharides such as glucose and fructose are digestible and thus not considered prebiotics according to the standard definition. Mannose is not used by the chicken host but cannot be considered to be a prebiotic since it is not fermented by the indigenous microbiota. Mannose, however, has been shown to decrease Salmonella colonization in chickens. Mannose is the saccharide ligand for bacterial type 1 fimbriae. These type 1 fimbriae are common surface projections of Salmonella, by which the bacteria can attach to the intestinal mucosa. Mannose binds type 1 fimbriae of Salmonella and thereby blocks adhesion of type 1 fimbriae bearing bacteria to epithelial cells and mucus (Craven et al., 1992; Dibb-Fuller et al., Reference DIBB-FULLER, ALLEN-VERCOE, THORNS and WOODWARD1999). Supplementing 2.5% mannose in the feed has been shown to reduce Salmonella colonization (Oyofo et al., Reference OYOFO, DELOACH, CORRIER, NORMAN, ZIPRIN and MOLLENHAUER1989). Addition to the feed of the more economically relevant 0.1% inclusion rate also reduced shedding, caecal and liver colonization after infection of 2 week old chickens with 2x10⁷ cfu of a S. enteritidis strain (Agunos et al., Reference AGUNOS, IBUKI, YOKOMIZO and MINE2007).

Amongst many available disaccharides, lactose has been used in multiple experimental studies. Several studies report that lactose administration increases the short-chain fatty acids and lactic acid concentrations in the gut (Hinton et al., Reference HINTON, CORRIER, ZIPRIN, SPATES and DELOACH1991; Hume et al., Reference HUME, KUBENA, BEIER, HINTON, CORRIER and DELOACH1992). Most studies use a combination of lactose and competitive exclusion cultures (Hinton et al., Reference HINTON, CORRIER, ZIPRIN, SPATES and DELOACH1991; Corrier et al., Reference CORRIER, HARGIS, HINTON and DELOACH1993; Nisbet et al., Reference NISBET, CORRIER and DELOACH1993; Reference NISBET, CORRIER, SCANLAN, HOLLISTER, BEIER and DELOACH1994). Lactose has been shown to successfully reduce Salmonella colonization mainly during feed withdrawal periods, i.e. during forced moult in layers or during the feed withdrawal period before slaughter in broilers. Reductions in Salmonella colonization in animals given lactose in drinking water are in all studies marginal or even non-existent (Barnhart et al., Reference BARNHART, CALDWELL, CROUCH, BYRD, CORRIER and HARGIS1999).

Oligosaccharides usually are obtained through enzymatic synthesis or through hydrolysis of polysaccharides. Fructo-oligosaccharides (FOS) are short-chain polymers of beta 1-2 linked fructose units, often produced by hydrolysis of inulin. Their use in broiler feed was shown to reduce colonization of the intestine by Salmonella in some studies, but in most studies the effects were marginal or absent (Bailey et al., Reference BAILEY, BLANKENSHIP and COX1991; Waldroup et al., Reference WALDROUP, SKINNER, HIERHOLZER and WALDROUP1993; Oyarzabal and Conner, Reference OYARZABAL and CONNER1996; Chambers et al., Reference CHAMBERS, SPENCER and MODLER1997; Fukata et al., Reference FUKATA, SASAI, MIYAMOTO and BABA1999). In vitro, batch cultures inoculated with faecal slurries show bifidogenic effects of FOS (Rossi et al., Reference ROSSI, CORRADINI, AMARETTI, NICOLINI, POMPEI, ZANONI and MATTEUZZI2005). Most bifidobacteria can also grow on FOS (Rossi et al., Reference ROSSI, CORRADINI, AMARETTI, NICOLINI, POMPEI, ZANONI and MATTEUZZI2005). The effects of Bifidobacteria on Salmonella colonization are indirect and are explained below. The most probable mechanism of the proposed anti-Salmonella activity of FOS is thus due to an alteration in the microbiota composition of the gut, and the production of metabolites, such as short-chain fatty acids. Galacto-oligosaccharides (GOS) are Glu α 1-4 [ß Gal 1-6]_n. They are not broken down in the stomach or small intestine. GOS also have been shown to selectively stimulate bifidobacteria in rats and pigs (Holma et al., Reference HOLMA, JUVONEN, ASMAWI, VAPAATALO and KORPELA2002; Tzortzis et al., Reference TZORTZIS, GOULAS, GEE and GIBSON2005). There are no studies in poultry using GOS. Mannan-oligosaccharides (MOS) derived from the yeast cell wall fragments of Saccharomyces cerevisiae are commercially available for use in poultry feed. Mixing 4000 ppm in the feed reduced caecal S. typhimurium concentrations after experimental infection of broiler chicks (Spring et al., Reference SPRING, WENK, DAWSON and NEWMAN2000). Caecal contents from hens fed MOS protected chicks from colonization with S. enteritidis (Fernandez et al., Reference FERNANDEZ, HINTON and VAN GILS2000). The mechanism of action may be through blockage of type 1 fimbriae mediated Salmonella adhesion to the mucosa and more importantly through modulation of the local (mucosal) immune system and preservation of intestinal wall integrity.

Recently a wide range of other oligosaccharides, including isomalto-oligosaccharides, soy-oligosaccharides and xylo-oligosaccharides were investigated for their prebiotic and health protecting effects in laboratory animal models and human volunteers (reviewed by Tuohy et al., Reference TUOHY, ROUZAUD, BRÜCK and GIBSON2005). Little is known regarding their potential for protection against Salmonella in poultry. A recent study however shows a dose-dependent decrease in shedding and gut colonization when arabinoxylo-oligosaccharides (AXOS) were added to broiler feed (Eeckhaut et al., Reference EECKHAUT, VAN IMMERSEEL, LOUIS, PASMANS, FIEVEZ, HAESEBROUCK, DUCATELLE and VANDAMME2008).

The most common polysaccharide prebiotics in chicken feed probably are guar gum (from the seeds of Cyamopsis tetragonolobus) and partially hydrolyzed guar gum (PHGG). Feed containing 250 ppm PHGG protects chicks against S. enteritidis infection from 14d onwards. This is thought to be due to improvements in the balance of the intestinal microbiota (Ishihara et al., Reference ISHIHARA, CHU, AKACHI and JUJENA2000). Inulin is a polymer of fructose moieties (see above) with a variable degree of polymerization ranging from 3 to 60, and is fermented to butyrate (Rossi et al., Reference ROSSI, CORRADINI, AMARETTI, NICOLINI, POMPEI, ZANONI and MATTEUZZI2005). It is believed to act in a similar way as FOS. Its use as feed supplement in chickens is limited. It is believed that the higher the degree of polymerization, the further the prebiotic effect will penetrate in the intestinal tract.

Taken together, most prebiotics with a proven protective activity against Salmonella act through a shift in the composition of the intestinal microbiota, or through a modification of the metabolic activity of some indigenous intestinal micro-organisms. An alternative approach thus can be to directly administer these beneficial micro-organisms. These feed additives containing the micro-organisms are often referred to as probiotics.

competitive exclusion products

Newly hatched chicks have little opportunity for rapid development of a normal intestinal microflora, due to the clean housing conditions in which the chicks are reared. Rapid transfer of normal flora from the hen to the chick is impossible in modern mass production systems and development of an intestinal microbiota is thus considerably delayed. The lack of an intestinal microbiota has been linked with a pronounced susceptibility for Salmonella infections (Gast and Beard, Reference GAST and BEARD1989). To overcome this problem, the concept of competitive exclusion was developed by Nurmi and Rantala (Reference NURMI and RANTALA1973). By oral administration of a saline suspension of the contents of the alimentary tract from adult birds to newly hatched chicks, an adult-type microflora is established, protecting against Salmonella infection (Nurmi and Rantala, Reference NURMI and RANTALA1973). This treatment is a prophylactic measure that aims at increasing the resistance of young chicks to Salmonella and other infections. It may also be used after antibiotic therapy to restore the microbiota (Seo et al., 2000). The first method of application was to add the preparation to the drinking water. Uptake of drinking water in the first 24h after hatch however is very variable and the viability of the anaerobic microbiota components in the water may be reduced. Moreover, there is a delay between hatch and placement in rearing houses (Mead, 2000). Spray application in the hatchery has been designed to overcome these problems (Goren et al., 1984; Schneitz et al., 1992). The birds naturally preen themselves after being sprayed and thereby ingest the microflora. This is commonly enhanced by using bright light. Combining spraying and administration through the drinking water of competitive exclusion products has been shown to be efficient in reducing Salmonella contamination (Blankenship et al., 1993). The efficacy of competitive exclusion products has not only been demonstrated in experimental trials (Nuotio et al., Reference SCHNEITZ C., NUOTIO, MEAD and NURMI1992) but also in the field. As an example, Palmu and Camelin (Reference PALMU and CAMELIN1997) studied 68 flocks, of which 50% were treated with a competitive exclusion product in the hatchery. Of all treated flocks, 13% were Salmonella positive on arrival at the farm, compared with 25% of the non-treated flocks. Furthermore, at slaughter age, the proportion of Salmonella-positive flocks was decreased to 6% in the treated flocks, while there was an increase to 42% in the untreated flocks (Palmu and Camelin, Reference PALMU and CAMELIN1997). Competitive exclusion products, however, are difficult to license for marketing, because the composition is undefined, always leaving the question of a possible risk of contamination with pathogens.

probiotics

Probiotics are living microbial food or feed ingredients that have a beneficial effect on human health (Salminen et al., Reference SALMINEN, BOULEY, BOUTROIN-RUAULT, CUMMINGS, FRANCK, GIBSON, ISOLAURI, MOREAU, ROBERFROID and ROWLAND1998). Probiotics are mostly used to correct an unbalanced indigenous gut microbiota, and thus not for treatment of disease- or bacteria-specific phenomena, especially in humans. Indeed, probiotic bacteria have been shown to positively affect the gut microbiota and symptoms in humans suffering from inflammatory bowel diseases, diarrhoea (infectious and antibiotic-associated) and even syndromes such as atopic allergy eczema and food allergies (Isolauri et al., Reference ISOLAURI, SALMINEN and OUWENHAND2004). Probiotics are administered not only in order to modify the microbiota composition, but also in order to improve gut wall integrity and to exert anti-inflammatory effects. Ideally, site-specific actions should be exerted by well-chosen probiotics, although this is difficult to achieve. In chickens, probiotics would be of particular value when they correct so-called dysbacteriosis, a vague entity related to wet litter. Also, increase in performance could be a valuable goal for probiotic use. Probiotics also may be administered for the purpose of Salmonella control. Most studies, however, show only limited effects of probiotics on colonization of Salmonella in chickens, with some exceptions. Most probiotic strains that are on the market are Bifidobacterium and Lactobacillus species.

Numerous reports have been published studying effects of Lactobacillus strains on Salmonella. It has been shown that metabolites produced by lactobacilli (including acetic acid, lactic acid, hydrogen peroxide and bacteriocins) can depress growth of Salmonella in vitro (Jin et al., Reference JIN, HO Y.W., ABDULLAH, ALI and JALALUDIN1996; Dembélé et al., Reference DEMBÉLÉ, OBDRZALEK and VOTAVA1998; Ocana et al., Reference OCANA, PESCE DE RUIZ HOLGADO and NADER-MACIAS1999; Miyamoto et al., Reference MIYAMOTO, HORIE, FUJIWARA, FUKATA, SASAI and BABA2000; Kizerwetter-Swida and Binek, Reference KIZERWETTER-SWIDA and BINEK2005). Lactobacilli were also shown to inhibit attachment of Salmonella to intestinal mucus and intestinal epithelial cells (Jin et al., Reference JIN, HO Y.W., ABDULLAH, ALI and JALALUDIN1996; Craven and Williams, Reference CRAVEN and WILLIAMS1997, Reference CRAVEN and WILLIAMS1998; Gusils et al., Reference GUSILS, PÉREZ CHAIA, GONZALEZ and OLIVER1999). Lactobacillus plantarum has been shown to possess mannose-sensitive receptors, an uncommon phenomenon in gram positive bacteria. With these receptors, it can compete for the same adhesion sites in the intestine as the gram negative pathogens (Bengmark, Reference BENGMARK1998). Lactobacilli also have been shown to possess immunomodulating properties. In vitro proliferation of spleen lymphocytes and induction of specific humoral responses to model antigens have been reported after contact with lactobacilli (Koenen et al., Reference KOENEN, KRAMER, VAN DER HULST, HERES, JEURISSEN and BOERSMA2004). Lactobacilli strains thus can affect Salmonella colonization due to their anti-bacterial, anti-adhesive and immunomodulating effects. Candidate probiotic strains can be chosen based on these properties, and in addition on their acid and bile resistance, and their colonization characteristics. In vivo trials indeed showed effects on Salmonella colonization, although most reports mostly mention minor effects. When 10⁸ cfu L. salivarius CTC2197 was dosed by oral gavage together with 10⁶ cfu S. enteritidis directly into the proventriculus in 1-day-old chicks, the Salmonella bacteria were completely cleared from the birds after 21 days (Pascual et al., Reference PASCUAL, HUGAS, BADIOLA, MONFORT and GARRIGA1999). The inclusion of this strain into a commercial feed mixture was proven a good way to supply it on the farm (Pascual et al., Reference PASCUAL, HUGAS, BADIOLA, MONFORT and GARRIGA1999). Also, in ovo inoculation has been proposed for administration of the probiotic strains to chickens. Inoculation of L. reuteri in ovo did not affect hatchability and it decreased the colonization by Salmonella post-hatch (Edens et al., Reference EDENS, PARKHURST, CASAS and DOBROGOSZ1997). More recently, Van Coillie et al. (Reference VAN COILLIE, GORIS, CLEENWERCK, GRIJSPEERDT, BOTTELDOORN, VAN IMMERSEEL, DE BUCK, VANCANNEYT, SWINGS, HERMAN and HEYNDRICKX2007) isolated 186 Lactobacilli from chickens and used 50 strains (based on taxonomy) for testing in vitro probiotic properties, such as acid production, inhibition of Salmonella, survival and growth at low pH and in the presence of bile salts and cell surface hydrophobicity (as a marker for adhesive properties). Based on the outcome of the in vitro trials, a selection of strains was used in experiments evaluating effects on colonization by Salmonella. However, only marginal effects on colonization of the gut and internal organs by Salmonella were detected using these Lactobacilli strains (Van Coillie et al., Reference VAN COILLIE, GORIS, CLEENWERCK, GRIJSPEERDT, BOTTELDOORN, VAN IMMERSEEL, DE BUCK, VANCANNEYT, SWINGS, HERMAN and HEYNDRICKX2007). Bifidobacteria also are believed to have effects on Salmonella colonization, although data from studies with chickens are scarce. Administration of a prebiotic that enhances numbers of Bifidobacteria significantly, did result in reduction of Salmonella colonization (Thitaram et al., Reference THITARAM, CHUNG, DAY, HINTON, BAILEY and SIRAGUSA2005). These data, however, do not prove a causal relationship between the number of Bifidobacteria and Salmonella colonization. Although methods have been developed to isolate Bifidobacteria from the chicken gut (Rada and Petr, Reference RADA and PETR2000), and to enrich Bifidobacteria in the chicken gut using different prebiotic compounds (Patterson et al., Reference PATTERSON, ORBAN, SUTTON and RICHARDS1997; Rada et al., Reference RADA, DUSKOVA, MAROUNEK and PETR2001; Thitaram et al., Reference THITARAM, CHUNG, DAY, HINTON, BAILEY and SIRAGUSA2005), the effect of inoculations with Bifidobacteria on Salmonella colonization in chickens has not yet been explored. Other lactic acid bacteria have been tested for their anti-Salmonella activity. Enterococcus faecium J96, isolated from the intestinal tract of a free-ranging chicken, inhibited the growth of S. pullorum, gallinarum, typhimurium and enteritidis in vitro, due to a combined effect of lactic acid and bacteriocin production (Audisio et al., Reference AUDISIO, OLIVER and APELLA1999). Lactic acid bacteria are believed to act by production of antibacterial metabolites, suppression of binding of pathogens to mucus or epithelial cells and immune stimulation. There is, however, another possible way of action, i.e. the stimulation of butyric acid production by anaerobic butyrate producing bacteria. Butyrate is known to decrease colonization of Salmonella, partly mediated by a specific suppression of the genes of the so-called Salmonella Pathogenicity Island I (Gantois et al., Reference GANTOIS, DUCATELLE, R, PASMANS, HAESEBROUCK, HAUTEFORT, THOMPSON, HINTON and VAN IMMERSEEL2006a; Van Immerseel et al., Reference VAN IMMERSEEL, METHNER, RYCHLIK, NAGY, VELGE, MARTIN, FOSTER, DUCATELLE and BARROW2005). Increases in lactic acid bacterial counts in the gut are correlated with increases in butyric acid concentrations in the rat model (Kleessen et al., Reference KLEESSEN, HARTMANN and BLAUT2001; Humblot et al., Reference HUMBLOT, BRUNEAU, SUTREN, LHOSTE, DORE, ANDRIEUX and RABOT2005). Although lactic acid bacteria do not produce butyric acid themselves, they increase butyric acid concentrations in the gut indirectly, as they stimulate proliferation of butyric acid producing bacteria. This mechanism is called cross-feeding. It has been shown that lactic acid, produced in vitro by Bifidobacterium adolescentis with starch as sole carbon source, is used by Anaerostipes caccae and Eubacterium hallii (in co-culture) for the production of high concentrations of butyric acid (Duncan et al., Reference DUNCAN, HOLTROP, LOBLEY, CALDER, STEWART and FLINT2004). In human gut samples, butyric acid producers are anaerobic bacteria belonging to the phylogenetic Clostridium clusters IV and XIVa (Pryde et al., Reference PRYDE, DUNCAN, HOLD, STEWART and FLINT2002). Random cloning and sequencing of 16SrDNA sequences isolated from chicken caeca revealed more than 85% of the clones belonging to Eubacteria and Clostridia spp. (Bjerrum, Reference BJERRUM2005). Approximately 10% of the clones have high similarity with Faecalibacterium prausnitzii, a species of Clostridium cluster IV that produces butyric acid in the human gut (Bjerrum, Reference BJERRUM2005). The same mechanisms of cross-feeding most likely are present in the chicken gut, as lactic acid bacteria are highly prevalent and anaerobic butyric acid producing bacteria in the chicken caecum, taxonomically classified as part of Clostridium clusters IV and XIVa, are present and have been recently isolated (Eeckhaut et al., Reference EECKHAUT, VAN IMMERSEEL, LOUIS, PASMANS, FIEVEZ, HAESEBROUCK, DUCATELLE and VANDAMME2008; personal communication).

Saccharomyces boulardii is a nonpathogenic yeast used in several countries as a living oral biotherapeutic agent to treat human patients suffering from diarrhoea. It has also been shown to reduce Salmonella colonization in broilers. Line et al. (Reference LINE, BAILEY, COX, STERN and TOMPKINS1998) administered Saccharomyces boulardii to day-old chicks by feed supplementation at a dose of 1 and 100 g/kg feed, after which the animals were challenged with 3.2x10⁸ cfu S. typhimurium at day 4. Of the positive control birds, 70% were colonized with Salmonella, whereas only 20 and 5% of the 1x and 100x yeast treated birds were colonized (Line et al., Reference LINE, BAILEY, COX, STERN and TOMPKINS1998).

Although it is clear that administration of certain bacterial species can suppress Salmonella colonization in broilers, it seems to be the case that combined cultures or competitive exclusion cultures are more efficacious. Most likely multiple microbiota components (and their substrates) can act synergistically.

bacteriophages

Lytic bacteriophages are viruses that invade bacterial cells and have the ability to lyse them. Phages have been proposed as antibacterial agents to combat pathogens in humans, animals and crops (Greer, Reference GREER2005). Although they have potential value, there are some limitations to their widespread use, including a) the narrow host range and thus the need to identify phages lytic for the pathogen to combat and b) the fear for transmission of virulence or resistance genes (Hermoso et al., Reference HERMOSO, GARCIA and GARCIA2007). In addition to the successful use in rinsing water for broiler carcasses (Higgins et al., Reference HIGGINS, ANDREATTI FILHO, HIGGINS, WOLFENDEN, TELLEZ and HARGIS2008), phages can also be used during the live phase. Recently, Atterbury et al. (Reference ATTERBURY, VAN BERGEN, ORTIZ, LOVELL, HARRIS, DE BOER, WAGENAAR, ALLEN and BARROW2007) used phages isolated from poultry farms in experimental infection studies, using Salmonella strains that were lysed in vitro by the phages. In general, oral administration of high titres of the phages resulted in a reduction of caecal colonization the first 2 days, but at day 3 post-infection the Salmonella titres were already increasing again. Interestingly, phage resistance was common, and the resistance level was correlated with the phage titre that was administered to the animals (Atterbury et al., Reference ATTERBURY, VAN BERGEN, ORTIZ, LOVELL, HARRIS, DE BOER, WAGENAAR, ALLEN and BARROW2007). Also Andreatti Filho et al. (Reference ANDREATTI FILHO, HIGGINS, WOLFENDEN, HIGGINS, TELLEZ and HARGIS2007) reported a decreased colonization of the caecal tonsils early post-infection that appeared to be lost at 48 hours post-infection. It is currently unknown whether low infection doses of Salmonella in experimental models would yield more promising data, and the use of phages in field studies has not yet been reported. Applying phages at 1 day pre-slaughter could be an option to prevent the appearance of bacteriophage insensitive Salmonella bacteria. Recently, endolysins, i.e. phage-encoded enzymes that break down bacterial peptidoglycan at the terminal stage of the phage reproduction cycle, have been used with success to control pathogens. References can be found in the review paper of Hermoso et al. (Reference HERMOSO, GARCIA and GARCIA2007). Endolysins or phage-derived proteins have not yet been tested in poultry.