Introduction
Thermophilic Campylobacter are consistently the cause of most reported food poisoning outbreaks in the UK. In England and Wales for example Campylobacter were isolated on over 46,000 occasions in 2005 and 2006 (Health Protection Agency, Reference HEALTH PROTECTION AGENCY2008a) compared with salmonella which was isolated on 11,459 and 12,523 occasions respectively over the same time period (Health Protection Agency, Reference HEALTH PROTECTION AGENCY2008b). While Campylobacter can be found in a number of environments and are associated with a range of animals, including companion animals, the main route for Campylobacter entering the human food chain is considered to be chicken (ACMSF, Reference ACMSF2005). So whereas 5.7% of chicken purchased at retail outlets in the UK in 2001 were positive for Salmonella the comparable figure for Campylobacter was, depending on the country, up to 76% (FSA, Reference FSA2003). In part this is a reflection of the impact that the poultry industry has had on Salmonella through vaccination programmes and improved biosecurity and it highlights the inability of interventions to control Campylobacter.
A number of studies have considered the risk factors for Campylobacter infection of chicken and some have identified the water supply system as a route by which Campylobacter may enter the poultry house and/or be transported around the house. In this review the role that water supply systems have in the infection of chicken with Campylobacter is discussed as are potential control mechanisms.
Campylobacter
Campylobacter are zoonotic organisms that are often isolated from the digestive tract, and in particular the caeca of chickens, where they reside, normally without any apparent ill-effect (Corry and Atabay, Reference CORRY and ATABAY2001). However when established in the gut of man Campylobacter can cause acute enteric illness. C. jejuni is the main contaminant of poultry although C. coli and C. lari may also be present (Hald et al., Reference HALD, WEDDERKOPP and MADSEN2000).
Campylobacter have a curved, s-shaped or spiral morphology, ranging in length from 0.5–5.0 μm and stain Gram-negative. These motile organisms tend to be microaerophilic although they may tolerate higher levels of oxygen – leading to a transformation to a coccal morphology (which may also be seen with older cultures).
Campylobacter are considered capable of entering a viable but not-culturable phase (VNC) (Rollinson and Colwell, Reference ROLLINSON and COLWELL1986). The significance of the VNC phase is uncertain (e.g. is it infectious? Can it revert to a culturable phase? Is the coccoid form simply a degenerative form of the spiral?) however, as will be discussed later, the potential presence of VNC forms is one factor that can make the recovery and isolation of Campylobacter problematic.
Campylobacter do not normally grow outside the host but all 17 Campylobacter grow well at 37°C with the so-called thermophilic Campylobacter, C. jejuni, C. coli, C. lari and C. upsaliensis, growing optimally at 42-43°C. Campylobacter will however survive at temperatures down to or below freezing and, in environments such as those found on poultry units, for prolonged periods of time (Hoop and Ehrsam, Reference HOOP and EHRSAM1987).
Infection of chickens with Campylobacter
It is generally considered that Campylobacter infections in chicken flocks are the result of horizontal rather than vertical-transmission (Corry and Atabay, Reference CORRY and ATABAY2001). Typically flocks will remain negative for Campylobacter for up to three weeks of age (Bull et al., Reference BULL, ALLEN, DOMINGUE, G, JORGENSON, F, FROST, URE, WHYTE, TINKER, CORRY, GILLARD-KING and HUMPHREY2006) but thereafter it has been shown that more than 40% of the flocks can be positive by four weeks of age while by seven weeks of age this can increase to >90% (Evans and Sayers, Reference EVANS and SAYERS2000).
The reasons for the delay in the onset of infection are uncertain but probably reflect less a lack of exposure and more the inability of the organism to colonise the bird in significant numbers. Potential mechanisms advanced to account for this delay include the presence in the young chicken of maternal antibodies (most parent flocks in the UK will be positive for Campylobacter). The ACMSF (Reference ACMSF2005) were confident that ‘properly applied biosecurity’ would reduce the incidence of Campylobacter infections while acknowledging that this may be easier to achieve in Scandinavian countries than the UK (due to colder winters and, typically, the fewer number of poultry houses per farm in the former).
Campylobacter favour the relatively low oxygen tension that is found within the mucus layer in the caecal and cloacal crypts in particular (Doyle, Reference DOYLE1991). Once Campylobacter are established in the digestive tract, birds can shed large numbers in the faeces (up to 108 cfu/g, Stern et al., Reference STERN, JONES, WESLEY and ROLLINS1994) and the infection spreads, in often less than a week, through the flock (Jacobs-Reitsma et al., Reference JACOBS-REITSMA, VAN DE GIESSEN, N.M. and MULDER1995, Evans and Sayers, Reference EVANS and SAYERS2000).
While Campylobacter will not multiply in the environment the wide range of host species coupled with the ability to survive for prolonged periods means that it is distributed widely and can be considered to be ubiquitous. Various studies have failed to identify Campylobacter in the poultry house following cleaning so it is assumed that infection normally arises from a source outside the area used to house the birds (Corry and Atabay, Reference CORRY and ATABAY2001). However the complicated epidemiology of this organism, combined with the difficulty in recovering and culturing the organism (in particular the VNC form) means that attempts to identify sources and routes of infection of poultry flocks are often unsuccessful (Evans and Sayers, Reference EVANS and SAYERS2000).
A number of studies have however attempted to identify the main risk factors for the infection of chicken with Campylobacter (Table 1). A meta-analysis conducted by Adkin et al. (Reference ADKIN, HARTNETT, JORDAN, NEWELL and DAVISON2006) identified 14 sources of on-farm contamination and 37 contributing factors. The authors identified a depopulation event, another poultry house on the farm, on-farm staff and other animals on the farms as major sources of infection. These findings highlight the range of routes by which Campylobacter might enter a poultry house and, more specifically, to demonstrate that to date water or water supply equipment has not been routinely identified as a potential source of Campylobacter. In respect of the last point however it is pertinent to note that few studies have generated robust data on the contamination of water with Campylobacter.
Examples of risk factors associated with flock infection with Campylobacter.
*this counterintuitive finding was explained by the authors as possibly being an artefact of the way that the relevant question was posed and/or a reflection of poor boot-dipping (incomplete immersion of the boot and inadequate levels of disinfectant).
**results of a meta-analysis
Water as a source of Campylobacter infection in chickens
background
Pearson et al. (Reference PEARSON, GREENWOOD, HEALING, ROLLINS, SHAHAMAT, DONALDSON and COLWELL1993) identified drinking water as a source of infection, while Kapperud et al. (Reference KAPPERUD, SKJERVE, VIK, HAUGE, LYSAKER, AALMEN, OSTROFF and POTTER1993) found it to be a very significant risk factor. More recently Adkin et al. (Reference ADKIN, HARTNETT, JORDAN, NEWELL and DAVISON2006), using data from 74 studies in a meta-analysis, noted that the most frequently reported categories for sources of Campylobacter were vertical transmission, origin of water supply and sequential spread from the previous flock (i.e. carry-over). When the data were analysed however, of the 13 potential sources identified for Campylobacter, water supply came 11th on the list with a relevancy score of -8.41. It is also notable that the same study ranked water disinfection and water equipment as nine and 27 out of 32 potential contributing factors. As Campylobacter are more susceptible than, for example, E. coli to water chlorination (Lund, Reference LUND1996), the ranking of water disinfection as ninth highlights the importance of interpretation when assessing risk factors. Thus it is probable that water disinfection is not a contributing factor per se to infection but rather should be taken as a proxy indicator of microbial or disease problems on-farm – water disinfection being introduced in an attempt to control such conditions. Similar findings (i.e. an association between poor flock performance and water treatment) were reported by a major breeding company that investigated the performance of broiler breeder flocks across Europe (Nicholson pers comm.).
There is then limited evidence for water or water systems being a source or contributing factor for Campylobacter infection. but it is possible that the importance of water has not been properly recognised due to a failure to recover Campylobacter from water (e.g. due to organisms being in a VNC state, sampling of insufficient volumes of water, etc). A distinction also needs to be made between contamination of the incoming water supply (e.g. public water supply, borehole) and the subsequent contamination of water in the header tank or other parts of the water system – contamination that may cause the organism to be spread through the flock more rapidly that would otherwise be the case.
source of water as a risk factor for Campylobacter
When considering the source of Campylobacter infection it might be expected that so-called private water supplies (e.g. boreholes or wells) would pose a greater risk than public water supplies. This supposition is supported by Said et al. (Reference SAID, WRIGHT, NICHOLS, REACHER, M and RUTTER2003) who analysed human infections from UK water supplies over 30 years. The authors reported that while only 0.5% of the population obtain water from private water supplies, these accounted for 36% of enteric disease associated with drinking water outbreaks. It is also notable that Campylobacter were identified in 52% of all drinking water outbreaks. Similarly Schuster et al. (Reference SCHUSTER, ELLIS, ROBERTSON, CHARRON, ARAMINI, MARSHALL and MEDEIROS2005) reviewed outbreaks over 30 years of Canadian data on infectious human disease outbreaks linked to drinking water. Of the 288 outbreaks, 99 were linked to public water systems, 138 to semi-public systems and 51 to private systems, with Campylobacter coming second to Giardia in the list of causative organisms. Severe weather and proximity to animals were two of the key factors related to disease outbreaks.
However when considering Campylobacter outbreaks associated with poultry units the evidence against borehole water is equivocal. The meta-analysis conducted by Adkin et al. (Reference ADKIN, HARTNETT, JORDAN, NEWELL and DAVISON2006) found an effect in only 3 of 17 reports. Whether this represents a lack of robust data, the extensive use of sanitizers on borehole water prior to use or the true relative risks, is not clear.
Chlorination of water, as used to treat public water supplies, is an effective way of inactivating Campylobacter, however this is not to say that chlorinated water cannot be the source of infection as evidenced by an outbreak of C. jejuni infection on a residential estate in South Wales (Richardson et al., Reference RICHARDSON, THOMAS, SMITH, NEHAUL, RIBEIRO, BROWN and SALMON2007). On this occasion the public water supply was contaminated with surface water from a cracked holding reservoir. When chlorinated water is implicated as the source of a Campylobacter infection it is normally because chlorine has not been added at the concentration required, or more often, as in the South Wales outbreak, gross contamination overwhelms the protective effect of the chlorine, which occurs post-chlorination.
role of biofilms
The ability of Campylobacter to persist in water supply systems is affected by the cleanliness of the systems. Dissolved inorganic mineral salts will form what is commonly referred to as scale while organic contaminants give rise to biofilm.
‘Biofilm’ is a generic term used to describe the organic, primarily polysaccharide, layer that builds up on the inner surfaces of the water supply system due to microbial action, and is a potential feature of most water supply systems. Campylobacter, for example, can form biofilms on the materials used commonly in water supply systems (Reeser et al., Reference REESER, MEDLER, BILLINGTON, JOST and JOENS2007) or may occupy biofilms formed by other mechanisms. Interestingly, aflagellate forms of Campylobacter are unable to attach to solid surfaces and form a biofilm (Kalmokoff et al., 2006). Irrespective of source, the presence of biofilms, have been shown to significantly enhance the survival times of Campylobacter (Trachoo et al., Reference TRACHOO, FRANK and STERN2002, Lehtola et al., Reference LEHTOLA, PITKANEN, MEIBACH and MIETTINEN2006), the extent of the effect being influenced by the strain of Campylobacter. (Buswell et al., Reference BUSWELL, HERLIHY, LAWRENCE, McGUIGGAN, MARSH, KEEVIL and LEACH1998). The presence of biofilms will also decrease the efficacy of sanitizers (Trachoo and Frank, Reference TRACHOO and FRANK2002).
In order to remove biofilms from drinking systems it is normal commercial practice, when cleaning the system between flocks, to ‘shock’ water lines with a suitable sanitising or oxidising agent. The aim of this shock cleaning is to remove biofilm and dissolve any mineral deposits that have formed within the system. One approach to this would be as follows (Quiroz, Reference QUIROZ2008): drain the system, flush the system with 85% organic acid and then hold this solution in the system for 90 minutes, flush with 35% hydrogen peroxide and hold for 24-48h, flush again with organic acids.
Like biofilms, protozoa have also been demonstrated to protect ingested Campylobacter against the actions of disinfectants (King et al., Reference KING, SHOTTS, WOOLEY and and PORTER1988). More recently Snelling et al. (2006, 2007 and 2008) reported detecting protozoa and C. jejuni in poultry water supply systems and the ability of birds to become infected after having been challenged with an amoeba containing the C. jejuni.
survival and detection of Campylobacter in water
The influence of strain on the ability of Campylobacter to survive when inoculated into potable water has been noted by various authors, including Cools et al. (Reference COOLS, UYTTENDAELE, CARO, E., NELIS and DEBEVERE2003) who noted that when 19 strains of C. jejuni were tested (two reference strains, four water isolates, nine human isolates), the poultry isolates survived for the longest periods (32-50 days). More recently Ogden et al. (Reference OGDEN, MACRAE, JOHNSTON, STRACHAN, CODY, DINGLE and NEWELL2007), commented, when discussing the implications of the types of Campylobacter isolated from poultry drinking systems, that some Campylobacter may have developed 'stress response systems' which enable prolonged survival in water. The authors used multi-locus sequence typing (MLST) techniques to type isolates and recovered, among others, clonal complex no. 45, a type reported as being frequently associated with water sources (French et al., Reference FRENCH, BARRIGAS, BROWN, P, WILLIAMS, LEATHERBARROW, BIRTLES, BOLTON, FEARNHEAD and A.2005).
However while circumstantial evidence exists for water being a potential source or contributing factor to contamination of flocks, it is more difficult to find field studies that have clearly identify water as the source or route of infection. This may however, in part, reflect difficulties in recovering Campylobacter from water and water supply systems. For example Pearson et al. (Reference PEARSON, GREENWOOD, HEALING, ROLLINS, SHAHAMAT, DONALDSON and COLWELL1993) reported on a recurrent Campylobacter infection on a chicken farm that lasted for 18 months. Despite repeated sampling (185 samples) of various parts of the water supply system, including the bore hole, direct and enrichment methods failed to produce a positive result. It was only when Indirect Fluorescence Antibody microscopy was used that Campylobacter were identified in both the water source and the supply system.
More recently Zimmer et al. (Reference ZIMMER, BARNHART, U. and LEE2003) used a range of techniques, including, RAPD-PCR and immuno-fluorescence microscopy to study the role of the water supply system in Campylobacter infections in three sequential commercial flocks. Campylobacter were recovered from the biofilm associated with the nipples in the drinker line but the authors concluded that, based on the RAPD-PCR data, the Campylobacter found in the water system differed from those recovered from the chickens. In discussing the implications of their findings the authors noted that factors such as:
• the organism being present in water in a VNC state
• the difficulties of detecting low numbers of Campylobacter in water
• the possibility of the biofilm contributing to the VNC state
as all being possible explanations for the failure to confirm a link between Campylobacter in the water supply system and the birds.
Ogden et al. (Reference OGDEN, MACRAE, JOHNSTON, STRACHAN, CODY, DINGLE and NEWELL2007) were more successful in demonstrating a link between Campylobacter strains isolated from birds and from the water supply system when, using MLST, they matched strains isolated from the header tank with those isolated from the birds. However a link could only be demonstrated on one of 12 farms investigated and it was not possible to confirm whether the tank became contaminated before or after stocking with birds.
The temperature of the water can influence the ability of Campylobacter to survive, survival times being extended in the lower temperatures associated with the winter months (Hazeleger et al., Reference HAZELEGER, JANSE, KOENRAAD, BEUMER, ROMBOUTS and ABEE1995). Irrespective of temperature, drinking water will be depleted in nutrients compared with most other environments and as such will be a hostile environment for Campylobacter The ability of Campylobacter to survive in drinking water is dependent in part upon the ability to convert to the VNC form (Talibart et al., Reference TALIBART, DENIS, CASTILLO, CAPPELIER and ERMEL2000).
While the VNCs of Campylobacter can enable the organisms to survive for prolonged periods in water, the infectivity of the VNC form is open to debate. Various studies have shown that the VNC Campylobacter can, sometimes after passing through an animal, be infective (Saha et al., Reference SAHA, SAHA and SANYAL1991, Stern et al., Reference STERN, JONES, WESLEY and ROLLINS1994, Talibart et al., Reference TALIBART, DENIS, CASTILLO, CAPPELIER and ERMEL2000). However in terms of what may happen in the field, these studies may have little relevance (Newell and Wagenaar, Reference NEWELL and WAGENAAR2000). These authors argue that the even when high doses of the VNC form are administered, the colonisation potential of the organism is severely reduced and, in practice, the significance of the VNC form in causing infection is likely to be negligible. However this view may underestimate the importance of the VNC form in campylobacteriosis caused by contaminated water (Skelly, Reference SKELLY2003).
Water treatment on poultry farms
Irrespective of the data regarding Campylobacter, the on-farm treatment of water to be used as drinking water for poultry is increasing in popularity in Europe (Pattison, Reference PATTISON2001). However this is an area where practice tends to run ahead of the science with, given the extent of its use, relatively few peer-reviewed publications on the topic.
Drinking water supply systems used in commercial chicken units typically consist of:
• Source (e.g. public water supply, borehole, surface water)
• Header tank
• Supply line (some with an in-line filter)
• Delivery system (normally nipples but can be bell-type drinkers)
Numbers of bacteria found in water at the different locations within drinker systems will normally vary and the numbers isolated from similar locations in different systems can also vary significantly. Watkins (Reference WATKINS2005) reported that the number of aerobic bacteria isolated from the beginning and end of four drinker systems varied as follows (the first number being the colony forming units/ml at the source and the second the number at the end of the line): system one - 2,700/26,600, system two 203,000/2,340,000, system three – 600/282,000 and system 4 – 0/4,775,000.
The numbers, and potentially the profile of organisms in the water lines will also change during the life of the crop. When the building has been prepared to receive the day-old birds it would be normal practice to warm the air temperature to ~30°C and hold it at this temperature for 24 hours prior to the birds being placed in the building and for few days thereafter (after which time the temperature is lowered incrementally until a temperature of ~20°C is reached by 21 days of age). The combination of static water, followed by very low flow rates (chicks drinking small volumes of water relative to what will be consumed in later life), all within an environment held at or near to 30°C, allows the water temperature to rise significantly and, as a consequence, the microbial numbers to increase. Producers that have monitored microbial contamination of drinking water invariably find that contamination peaks during the first week of the bird's life (Pearson, pers comm).
Normally, systems would be cleaned out between crops using, as a minimum, a disinfectant product. The system would then be flushed and thereafter the system could be either dosed with sanitizer throughout the rearing period, dosed at the end of the rearing period or there may be no further treatment during rear. If a treatment is used during rearing it is normal to suspend treatment for a few days if vaccination is to be carried out through the drinking water.
A range of different product types can be used to clean out drinking systems between crops. To remove biofilm or scale products such as oxidisers (e.g. ozone, chlorine dioxide or hydrogen peroxide), acids or oxidisers combined with organic acids (e.g. peracetic acid) can be used. Terminal disinfectants can range from iodine based products to composite products containing peracetic acid, hydrogen peroxide, acetic acid and surfactant, to products containing peroxygen compounds and organic acids.
The types of products that are used to sanitise drinking water include: organic acids, chlorine, chlorine-dioxide, hydrogen peroxide, ozone and peroxygen compounds. The modes of action will differ with some treatments having the ability to reduce the microbial load in the water only while others can also modify the microflora in the bird.
acidifiers
Organic acid treatments are used to treat drinking water and feed for poultry. Whichever route is adopted the aim, in part, is to reduce and/or maintain a low pH in the crop and gizzard (which relative to more distal parts of digestive tract have a lower pH anyway). Knight et al. (Reference KNIGHT, HOFACRE, MATHIS, QUIROZ and DIBNER2006) reported that a commercial organic acid water treatment significantly reduced numbers of Salmonella in the crop, but did not affect the number of organisms in the caeca. Chaveerach et al. (Reference CHAVEERACH, KEUZENKAMP, LIPMAN and VAN KNAPEN2004) attempted to use organic acids to control Campylobacter in the growing chicken. While the drinking water remained free of Campylobacter throughout the study, the two groups of chickens, both inoculated at the outset with different levels of Campylobacter, remained positive for Campylobacter. The use of the fatty acid sodium caprylate did not affect the growth of Campylobacter in chickens (Metcalf et al., Reference METCALF, VENKITANARAYANAN, SANTOS, DONOGHUE, DIRAIN, REYES-HERRERA, AGUIAR, BLORE and DONOGHUE2007).
Organic acids tend to be more effective against Gram-negative organisms such as Campylobacter, although generally the activity of organic acids is a function of the tendency of the acid to disassociate (pK value) and the pH of the environment, the lower the pH the more effective the acid. The acid passes across the cell wall before it dissociates, releasing hydrogen ions, which in turn reduce the pH and impair DNA and protein synthesis. In the digestive tract, acidifiers tend to encourage the growth of organisms such as Lactobacilli, organisms which can make the gut environment less favourable for pathogens. However not all organic acids have inherent antimicrobial properties. Citric acid is a good example of a compound that lowers the pH effectively in water and which, in so doing, creates a favourable environment for say chlorine, but does not in itself have particularly good antimicrobial properties. In contrast, 2-hydroxy-4-methlythio butanoic acid has good antimicrobial proprieties in water and has been shown to be active against E. coli, Salmonella and Campylobacter (Quiroz, Reference QUIROZ2008).
Water standards for poultry often cite a pH below 5.9 as detrimental, however the basis for this has been questioned by Watkins (Reference WATKINS2005). Watkins, reporting work conducted at the University of Arkansas, noted that normal production parameters, such as feed intake and weight gain, were not affected by pH ranging from 3-6. Watkins highlighted two issues associated with the acidification of water. Firstly, in the Arkansas study, birds receiving water of pH 4 and 5 had a slightly higher prevalence of tibia dyschondroplasia. Secondly, when using organic acids (cf inorganic acids), there can, due to the weak nature of the acid and the resultant tendency not to readily release hydrogen ions, be palatability issues which result in birds reducing their water intake.
It is not uncommon (Watkins, Reference WATKINS2005) for those producing meat birds, in particular in the USA, to acidify the water intermittently (for periods of one to three days). A range of products can be used for this purpose, including acetic, citric, lactic, malic, proprionic, tartaric acids and blends of the same. These types of acids can lower the pH of drinking water to approximately three (Watkins, Reference WATKINS2005), a pH that approaches the pH of the crop (ca. pH 2). The mode of action of acidifiers has not been characterised fully, although Watkins notes that by reducing the microbial load of the crop they limit the type and number of pathogens that can establish themselves in the digestive tract.
chlorine
Chlorination is used widely as a means of reducing the microbial load in drinking water, be it for humans or livestock. When added to water chlorine forms a number of products including hypochlorous acid (HOCl) which is the most efficacious form in terms of inactivating micro-organisms, hypochlorite ions (OCl-) and hydrochloric acid (HCl). Hypochlorous acid breaks down into hydrochloric acid and oxygen, the efficacy of chlorine as a disinfectant being a product of the oxidising power of free oxygen atoms and the chlorine substitution reactions.
Chlorine can be supplied as a gas, although because of the problems associated with handling chlorine in this form it would not be usual to use this as a source or chlorine for water treatment on poultry units. More common is chlorine in a liquid form such as sodium hypochlorite (NaOCl) which typically has between 10-15% available chlorine, and a pH of 13.
Solid (tablet) forms of chlorine include products such as Trichlor (trichloro-s-triazinetrione) and Dichlor (sodium dichloro-s-triazinetrione). The former has 90% available chlorine and a pH of 3, while the latter has 62% available chlorine and a pH of 7. Typically UK farms would add (twice per day) chlorine in tablet form into the header tanks achieving a peak concentration of approximately 5ppm (Cameron, pers comm).
Chlorine's ability to exchange atoms with compounds is the key to its use as a disinfectant. For example chlorine can replace hydrogen atoms in enzymes causing molecules to change its shape and so lose functionality. However the pH of the water is important when considering the action of chlorine. When in a solution with a pH of 6 the ratio of hypochlorous acid to hypochlorite ions is 4:1 whereas at pH 8 the ratio is reversed (at pH 7.5 the ratio is 1:1). This is important because, compared with hypochlorite ions, hypochlorous acid is up to 100% more efficient at inactivating organisms – the negatively-charged cell wall of pathogens being penetrated more easily by the neutral hypochlorous than the negatively charged hypochlorite ion. Because chlorination is most effective when the pH of the water is between 6 and 7 it may be necessary to acidify the water. Because the mixing of acid and chlorine sources needs to be prevented (as it can result in the liberation of chlorine gas) it is recommended that a dual-injection pump is used to acidify and chlorinate the water.
The time taken to inactivate pathogens with chlorine is dependent not only on the pH of the water but also the organism itself (Table 2).
Disinfection time for different types of pathogenic microorganisms with chlorinated water (1 ppm free chlorine) at pH = 7.5 and 25 °C (from Lenntech, Reference LENNTECH2008).
While chlorine can reduce the microbial load in water it has been found, when included at rates up to 5 ppm, not to affect the frequency or level of colonisation of chickens with Campylobacter (Stern et al., Reference STERN, ROBACH, COX and MUSGROVE2002).
Acidified sodium chlorite has been used as a water treatment immediately prior to slaughter at concentrations up to 600 ppm to control levels of salmonella and Campylobacter in artificially infected birds (Mohyla et al., Reference MOHYLA, BILGILI, OYARZABAL, WARF and KEMP2007). The number of Salmonella in the upper digestive tract were significantly reduced by this treatment, although there was no effect on the number of Campylobacter isolated from the digestive tract.
There are a wide range of commercial products on the market but, in terms of free chlorine in the water, the recommendations tend to be consistent. For example, Carter and Sneed (Reference CARTER and SNEED1987) recommended levels of chlorine at 1ppm. Similarly a chlorine based product called ‘Aqualution’ (Forum Bioscience, Reference FORUM BIOSCIENCE2005) was recommended to be used at a rate that provides 1.5–2 ppm free chlorine in drinking water. The University of Florida Cooperative Extension Service (Reference UNIVERSITY OF FLORIDA COOPERATIVE EXTENSION SERVICE1998) observed that concentrations of free chlorine between 50 and 100ppm ‘are well tolerated’ but between 3-5ppm would be appropriate in most instances.
When liquid products are used, a dosing pump or similar is required if concentrations are to be delivered accurately. UK farms typically use a dosing pump set to deliver 0.2% from a stock solution, providing a free chlorine concentration of 3-5 ppm at the point of consumption. It is common practice to only dose the water in this way for two days in any week (Cameron, pers comm).
chlorine dioxide
Chlorine dioxide (ClO2) differs from chlorine in a number of ways. It is more soluble in water (approximately ten times the solubility of chlorine) where, rather than hydrolysing, it is present as a dissolved gas. The solubility of chlorine dioxide, relative to chlorine, in cold water is considered to be advantageous. Unlike chlorine, when dissolved in water, chlorine dioxide is present as a free radical which is a strong oxidiser (2.6 times that of chlorine, Du Pont, Reference DU PONT2008) which will disrupt amino acids and RNA. It has the added benefit of being active over a wide pH range and, compared to chlorine, is more efficacious at pHs higher than 7.
Dosing levels vary depending on whether it is applied pre- or post-disinfection. Typical values for pre-disinfection are between 0.5 and 2.0 mg/l of chlorine dioxide (contact time of between 15 and 30 minutes) and for post- disinfection concentrations of between 0.2 and 0.4 mg/l may be used.
In a comparison of chlorine and chlorine dioxide Baghurst (Reference BAGHURST2002) reported that, in potable water, infectious bursal disease virus (IBDV) activity was reduced by 90% when exposed to 3 ppm of free chlorine for 1 hour. This compared with 0.7 ppm of chlorine dioxide which resulted in >97% reduction in activity. Furthermore chlorine levels could only be maintained by adding chlorine throughout the 1 hour period. When the test was conducted in water sourced from a dam the maximum level of chlorine tested (10 ppm) did not inactivate IBDV whereas 4.4 ppm chlorine dioxide was, under the same conditions, able to inactivate the virus.
Unpublished data (Cameron, pers comm) show that adding chlorine dioxide to drinking water at a rate sufficient to ensure a concentration of 1 ppm has a significant and progressive affect on the number of viable aerobes recovered from the water, with values falling from 21,120 organisms/ml to <100 organisms per ml over a 6 week period. In a separate study (Cameron, pers comm) treatment of the water system (again with 1ppm chlorine dioxide) reduced the number of coliforms in the header tank and water lines over a period of 6 weeks from values ranging from 5.0 to 1.0 x 103 cfu/ml in the header tank and 2.0 x 103 to 1.5 x 105 cfu/ml to 1.0 cfu/ml in the tank and water line.
hydrogen peroxide
Hydrogen peroxide decomposes to releases free oxygen radicals and water, free radicals being capable of both oxidising and disinfecting, and may provide other residual action. It is not unusual to use hydrogen peroxide in combination with other products such as peracetic acid.
The use of a 50% hydrogen peroxide product, stabilised with silver nitrate, is, according to Watkins (Reference WATKINS2005), being used increasingly by American growers. Summers (Reference SUMMERS2006) cites a level of 30 ppm hydrogen peroxide as being safe in drinking water although levels up to 3000ppm were, based on limited tests that they conducted, considered showed not to adversely affect feed consumption.
Improvements in microbial quality of the drinking water might be expected to be reflected in reduced levels of contamination of the bird and/or improved bird performance. Wilson (Reference WILSON2008) noted that the treatment of the water in commercial growing houses with hydrogen peroxide (150 ppm for the first week to remove biofilm and 50 ppm thereafter) reduced mortality by 1% and increased body weight by 4%.
others
The treatment of drinking water with ozone is recognised as being suitable for poultry, although it carries the caveat that it may be corrosive to steel fixtures (Winchell, Reference WINCHELL2001).
UV light can be used in operations needing to sanitise large volumes of water however the investment required and the nature of the plant, combined with the absence of any residual activity has limited its use on farms for treating drinking water.
The use of ozone, while not widespread, has been used to treat drinking water on farms however although efficacious, the lack of residual action, combined with the practicalities of its use have limited its uptake.
Carter and Sneed (Reference CARTER and SNEED1987) cite iodine-based disinfectants as being effective, with, compared to some products, the potential advantage of having residual activity. Typical concentrations recommended for use in water systems would be ca. 15-18 ppm tritratable iodine.
Conclusions
There is evidence for drinking water being a potential source of Campylobacter infection in poultry units, the potential increasing if the water comes from an untreated source (such as bore holes or surface water) or if the water source is at risk of contamination from neighbouring operations (e.g. growing or slaughter units). The role that drinking water systems may play in spreading infection within a flock, through cross-contamination, is less well defined.
It is likely that the limited evidence for water as a source or risk factor for Campylobacter infection of poultry is a feature, in part, of difficulties that have been experienced in recovering the organism from the environment. The role, if any, for example of the VNC state remains uncertain.
Many producers treat the drinking water supplied to their poultry as a precautionary measure, the potential benefits being considered to go beyond the control of Campylobacter. To this end there are a range of established water sanitising treatments available to poultry producers. However optimum approaches to using these systems, i.e. reducing the microflora in the water and so improve performance and/or reduce the prevalence of potentially water borne infections, including Campylobacter, appear to be lacking. Thus many producers will only treat the water on one or two days in any seven day period. Others may limit treating the water towards the end of the production cycle.
If the full benefits of drinking water treatments are to be realised for the poultry industry then further evidence regarding the optimal use for the main treatments available to poultry producers is required, both in terms of the effects on the normal production parameters such as weight gain, feed conversion efficiency, flock uniformity, and more specifically in terms of any effects on the prevalence of infections such as Campylobacter.
Acknowledgements
This review was funded by the Food Standards Agency (project B15024). The author is grateful to E. Cameron and J. Johnson for advice on industry practice as well as to G. Mead and I. Smith for comments on the draft.
