Disinfection of housing systems, with special reference to Salmonella

Kim O. Gradel, Danish Veterinary Institute, Hangøvej 2, 8200 Århus N, Denmark.

Phone: + 45 89 37 24 58. Fax: + 45 89 37 24 70. e-mail: kog@vetinst.dk.

Introduction

In Denmark, the Salmonella control programmes have effectively reduced new infections in the poultry sector (Wegener et al. 2003). However, Salmonella is often detected on the same farms and on these frequently in the same houses in successive crops (Gradel and Rattenborg 2003), i.e. persistently Salmonella-infected premises.

Various guidelines on effective decontamination procedures have been compiled (Linton et al. 1987; Engvall 1993; Svedberg 1993; Meroz and Samberg 1995). These correctly emphasize the importance of a thorough cleansing prior to disinfection, but this is difficult in many poultry houses having surfaces with cracks and crevices that are water pervious and equipment with bends and inaccessible areas. Battery cage houses are especially difficult to cleanse and disinfect properly, and several of these have had persistent Salmonella infections for years in spite of apparently effective decontamination procedures.

Most reports on disinfection of animal houses come from disinfectant companies, but the scientific literature on this topic is sparse. Disinfection research has focused on food enterprises and hospitals in which other disinfectants than those in the agricultural sector are often used, disinfection is normally performed daily, and surfaces and equipment are easier to clean, so it is generally difficult to extrapolate results to the agricultural sector.

Therefore, a disinfection project under the Danish Salmonella Control Programme for Poultry was implemented to obtain documentation for the elimination of persistent Salmonella infections in poultry houses. An important aspect is controlled studies in which worst-case scenarios in badly cleaned poultry houses are simulated. The project has two main pillars, one focusing on heating (“Heat disinfection as a sanitation method for Salmonella infections in poultry houses”), the other on chemical disinfectants (“Development of microbiological monitoring models in broiler houses: assessment of the impact of cleaning and disinfection procedures on Salmonella persistence”). Heating studies have been performed and most results published, whereas only some data from chemical disinfectant studies have been reported.

Heat disinfection, laboratory tests

The aim was to find a temperature-humidity-time treatment that would kill all Salmonella and E. coli (possible indicator bacteria) under worst-case scenarios mimicking badly cleaned poultry houses.

Survival of three Salmonella strains (S. Enteritidis, S. Typhimurium, S. Infantis) and naturally occurring E. coli was assessed in faeces and pelleted feed in their original state or pre-equilibrated for 10 days to 30% relative humidity (RH). Final heating temperatures were 50, 55, 60, 65 or 70 ^oC and final RH was 16-30 or 100%. Heating was performed, starting at 20 ^oC and increasing with 1 ^oC per hour until the final heating temperature, which was maintained for 48 hours. For a temperature-time-humidity scheme that killed all bacteria tests were repeated, using egg yolks and finely ground feed as well as faeces and pelleted feed.

Generally, there was a higher survival in pelleted feed than in faeces, and in dried than in non-dried faeces, whereas the survival in dried and non-dried feed was similar. There were no noticeable differences in survival between the three Salmonella strains. Heating at 100% RH was significantly more effective than at 16-30% RH. There were high correlations between results for Salmonella and E. coli. After 24 hours, all bacteria were killed at 60 ^oC and 100% RH.

More detailed descriptions have been published (Gradel 2002; Gradel et al. 2003a).

Heat disinfection, field tests

A temperature-humidity-time treatment of 60 ^oC and 100 %RH during at least 24 hours was the gold standard (cf. above) for heating naturally Salmonella-infected poultry houses. This was achieved by steam heating, often supplemented with 30 ppm formaldehyde at the beginning of the process. Altogether seven houses (two barn and five battery cage houses), distributed on six farms, were steam treated. Moreover, on one farm with three Salmonella-infected barn houses, one was steam treated, one was surface disinfected, and one was pulse fogged.

Several monitoring methods were used. A number of 100-300 Salmonella samples was taken both before and after the treatments and analysed by traditional qualitative procedures. In each house, temperatures were logged at 5-minute intervals at 12 sites. At each site, challenge samples (either feed spiked with E. coli or E. faecalis, or faeces with naturally occurring E. coli and enterococci) were placed during the heat treatment. Air humidity was also logged in some houses. It was difficult to achieve the desired 60 ^oC near the floor, so additional temperature loggers were placed at different heights near the floor to see at which height 60 ^oC was achieved.

A temperature of 60 ^oC was easily achieved 10 cm over the floor and above, while a 5-10 ^oC lower temperature was achieved at floor level. A relative air humidity of 100 was achieved within half an hour after commencing the steam treatment. In general, satisfactory bacteriological results at the gold standard were seen regardless of monitoring method. Moreover, formaldehyde seemed to lower the lethal temperature by 2-5 ^oC (comparison of challenge samples between sites with or without formaldehyde). Thus, steam heating of the poultry house combined with a chemical surface disinfection of the floor is recommended.

These studies have been described in more detail (Gradel et al. 2002; Gradel et al. 2003b).

Chemical disinfection, MIC tests

In Danish broiler houses, some Salmonella serotypes tend to persist for years whereas others are generally eliminated in one or a few crops; the reasons for these differences have not been elucidated. A few disinfectants are commonly used in Danish poultry houses, and theoretically this could favour resistance development, but with regard to animal houses little has been published on this.

The aims of this study were:

· To find minimum inhibitory concentrations (MICs) of five disinfectants for “non-persistent” and “persistent” Salmonella serotypes commonly isolated from Danish broiler houses and to relate these to serotype, persistence and use of disinfectants.

· To find MICs of five disinfectants for other Salmonella serotypes mainly isolated from poultry enterprises.

· To perform adaptation and de-adaptation studies with the five disinfectants for selected strains having high or low MICs to see if disinfectant resistance was developed and maintained.

MIC tests in replicate were performed on 286 Salmonella isolates (269 from Danish poultry, including 256 from broiler houses, and 17 from England, mainly from poultry) against five disinfectants used in the poultry industry (glutaraldehyde/benzalkonium chloride compound, formaldehyde, oxidising compound, tar oil phenol, iodophor). Generally, the small variations in MIC could not be associated with previous use of disinfectants, persistence or serotypes. Adaptation studies involving the five disinfectants did not alter the MICs beyond the normal biological variation, i.e. within one doubling dilution. Thus, these studies indicate that there is no reason to alternate between different disinfectants. It must, however, be emphasized that MIC studies mainly illustrate properties of bacterial strains, as they are far from real-life conditions, amongst other things because a bacterial killing, and not merely an inhibition, is the end goal in these.

Details can be read elsewhere (Gradel 2003; Gradel and Randall 2003).

Chemical disinfection, worst-case scenario carrier tests

According to the Danish Salmonella data base for broiler flocks (Angen et al. 1996), a glutaraldehyde/benzalkonium chloride compound, formaldehyde and an oxidising compound are used in ca. 39%, 32% and 15%, respectively, of the download periods. All these disinfectants are probably effective in properly cleaned houses under relatively high temperatures, but less is known about their efficacy under worst-case scenarios, e.g. materials difficult to clean, much organic matter and low temperatures that are often encountered, e.g. in winter periods.

Therefore, surface/carrier disinfection tests that mimicked worst-case scenarios encountered in badly cleaned poultry houses, often at low temperatures, were performed (Table 1).

Table 1: Factors used in worst-case scenario carrier tests

Factor (outcome)	Comments
Bacteria (Salmonella Enteritidis, S. Senftenberg, Enterococcus faecalis)	In the MIC-studies (cf. above) the S. Enteritidis and S. Senftenberg strain had low and high MICs, respectively, so it was tested how this related to simulated real-life conditions. Enterococcus faecalis is considered more resistant to outer detrimental conditions than Salmonella (i.e. possible indicator bacteria).
Bacteria concentration in organic matter (low, high)	Cf. Tables 2-5.
Materials (concrete flags, steel feed chain links, jute egg belts, wooden dowels)	Commonly found in poultry houses.
Organic matter (feed for layers, egg yolk, fat)	Feed represents the input of proteins, fats and carbohydrates (i.e. components that protect bacteria) to the animal house. Egg yolk is probably more protective for bacteria than egg white. Fat is often seen in badly cleaned feed troughs.
Weight of organic matter	Only used on the surface of concrete flags.
Temperature during a 24-hour period before disinfection (5, 10, 20, 30 ^oC)	A range of temperatures often encountered in empty poultry houses.
Disinfectant (formaldehyde (24.5% v/v), Bio Komplet Plus^®, Virkon S^®, water (control))	The three disinfectant types used most commonly in the Danish poultry sector (cf. above), all in 1% concentrations. WHO standard hard water was used for all disinfection solutions and the control.
Disinfection time (5, 15, 30 minutes)	Disinfection times that simulate those seen in poultry houses, although it is difficult to measure this, e.g. on vertical surfaces.
Temperature during a 25-hour period after disinfection (5, 10, 30 ^oC)	A range of temperatures often encountered in empty poultry houses.

Only the following combinations of materials and organic matter were tested: concrete flags/feed, feed chain links/feed, feed chain links/fat, jute egg belts/egg yolks, wooden dowels/feed, wooden dowels/fat, as we focused on combinations between materials and organic matter found commonly in poultry houses. Although faeces are commonly found in badly cleaned poultry houses, they were not used as organic matter, because the literature and our heating laboratory studies indicate it is easier to eliminate bacteria in faeces than in feed. Results are seen in Tables 2-5.

Table 2: Results for concrete flags inoculated with feed for layers

(disinfection conditions deteriorating towards the bottom of the table)

CFU¹

GPF²

Tb³

Ta⁴

Dt⁵

S. Enteritidis

S. Senftenberg

F⁶

B⁶

V⁶

W⁶

Low

10.0

20.2

10.9

N⁷

Low

20.0

20.2

10.9

ND⁸

Low

20.0

10.9

Low

20.0

5.9

Low

20.0

5.9

High

20.0

5.9

High

20.0

5.9

¹Colony forming units; low = ca. 4 x 10⁵-6 x 10⁶ g^-1 organic matter; high = ca. 4 x 10⁶-6 x 10⁷ g^-1 organic matter.

²Gram organic matter per flag.

³Mean temperature (^oC) during 24-h period before disinfection.

⁴Mean temperature (^oC) during 25-h period after disinfection.

⁵Disinfection time (minutes).

⁶Disinfectant; F = formaldehyde; B = Bio Komplet^® Plus, V = Virkon S^®, W = WHO standard hard water (control).

⁷N = no Salmonella detected; Y = Salmonella detected. Replicate results written in the same row.

⁸Not done because MSRV plates crystallized during incubation; no growth was seen either on Rambach agar plates when streaking from MSRV plates.

Table 3: Results for feed chain links inoculated with either fat or feed for layers

(disinfection conditions deteriorating towards the bottom of the table)

Organic

matter

S. Enteritidis

S. Senftenberg

Enterococcus faecalis

Fat

30.0

5.9

Feed

10.9

5.9¹

5.9

¹The tests in this row were performed with only one feed chain link pair per 250 ml disinfectant (the tests in the other rows were performed with two chain link pairs per 250 ml disinfectant).

Legend, cf. Table 2.

Table 4: Results for wooden dowels inoculated with either fat or feed for layers

(disinfection conditions deteriorating towards the bottom of the table)

Organic

matter

S. Enteritidis

S. Senftenberg

Enterococcus faecalis

Fat

30.0

YYN

YYY

NYN

YYY

NNN

YNN

NNN

YNY

YYY

YNN

YYY

5.9

YYY

Feed

10.9

NNN

YYN

YYY

NNN

YNN

YYY

NNN

YYY

5.9

NNN

YYY

NNN

YYY

YNN

YYY

5.9

NNY

NYN

YYY

NNN

YYY

NYY

NNY

YYY

5.9

YYN

NYN

YYY

NNN

YNN

YYY

Legend, cf. Table 2.

Table 5: Results for jute egg belt pieces inoculated with egg yolk

(disinfection conditions deteriorating towards the bottom of the table)

CFU¹

S. Enteritidis

S. Senftenberg

Enterococcus faecalis

Low

10.9

NNN

YYY

Low

5.9

Low

5.9

High

10.9

High

5.9

High

5.9

¹Colony forming units; low = ca. 2.9 x 10⁵-4.6 x 10⁶ g^-1 organic matter; high = ca. 2.9 x 10⁶-4.6 x 10⁷ g^-1

organic matter.

Legend, cf. Table 2.

For both S. Enteritidis (SE) and S. Senftenberg (SS), formaldehyde (F) was more effective (i.e. p < 0.05) than Bio Komplet Plus (B) (SE: p = 4.4 x 10^-5; SS: p = 5.8 x 10^-4), Virkon S (V) (SE: p < 10^-7; SS: p = 5.0 x 10^-7) and WHO water (W) (p < 10^-7 for both serotypes), and B was more effective than W (SE: p = 9.8 x 10^-5; SS: p = 9.0 x 10^-7). B was more effective than V for SE (p = 0.012), but not for SS (p = 0.075), and V was more effective than W for SS (p = 6.1 x 10^-4), whereas they did not differ for SE (p = 0.056). For E. faecalis (EF), there were no differences between any disinfectants when these were compared pair wise, maybe because EF is less susceptible than salmonella, maybe because it was mainly used under conditions that yielded a high protection to the bacteria (e.g. fats as organic matter and low temperatures). Therefore, it was also appropriate to compare the three bacteria two by two for the series in which they were tested. Here, the only differences between SE vs. EF was for F (p = 0.044), whereas all three disinfectants differed when SS and EF were compared (F: p = 1.9 x 10^-4; B: p = 5.6 x 10^-3; V: p = 2.6 x 10^-3); all these differences were to the benefit of EF. When SE and SS were compared, only V was more effective against the latter (p = 3.3 x 10^-3).

All tested combinations of poultry house materials and organic matter generally supported the statistical tendencies, except when feed chain links with fat were tested at 30 ^oC before and after disinfection (cf. Table 3), where V was better than the aldehydes, both for SE and EF. During the disinfection procedure, a seething was seen in the fat immersed in V, but not in F, B or W. However, the same conditions were reiterated with wooden dowels, and here no seething was seen for any of the disinfectants. Thus, it seems that the metal corrosive properties of V might enhance a bacterial killing.

It is often postulated that glutaraldehydes are effective down to ca. 5 ^oC whereas formaldehyde needs at least 16 ^oC to be efficient (Anonymous 2002), although the scientific documentation for this is sparse. It was therefore conspicuous that F was more effective than B in spite of the fact that temperatures around 5 ^oC were prevailing in many of the test series.

Overall, the efficacy of the tested disinfectants was (best first): formaldehyde > Bio Komplet Plus > Virkon S > WHO water (control), with the exception of feed chain links at 30 ^oC where Virkon S seemed to be the most effective. Generally, there were few differences between the two Salmonella serotypes. Enterococcus faecalis was often less susceptible than the two tested Salmonella serotypes, but more laboratory tests are needed before it should be used finally as an indicator bacterium, e.g. in field studies.

General discussion and conclusions

There are no general harmonised rules for the approval of disinfectants or disinfection methods. Only a few countries (France, Germany, Netherlands, UK, USA) have official disinfection tests, but all these are suspension tests from which results are difficult to extrapolate to real-life conditions, and each country has its own test, so results cannot be compared. Tests simulating real life conditions are more difficult to standardise, even in the laboratory. In addition, field tests have many uncontrollable factors, and controls are often impossible to include. We tried to compensate for this lack of a rigorous standardisation by increasing the number of tests and using many different conditions, and general tendencies were still consistent.

The above studies have elucidated applied aspects related to the disinfection of Salmonella infected poultry houses. Moist heat can conveniently be applied in the field, and it seems to be effective, especially when combined with formaldehyde. Formaldehyde also seems to be the most effective when tested in worst-case scenario surface disinfection studies. At this stage, this advocates the use of steam heating with formaldehyde, supplied with formaldehyde surface disinfection of the floor, where it is difficult to achieve sufficient temperatures.

Much more work needs to be done to gain a more comprehensive view of conditions prevailing in animal houses. Other aspects, such as the occurrence and elimination of biofilms and other kinds of poultry house materials and organic matter, should be investigated. Moreover, a standardised use of putative indicator bacteria, both naturally occurring and prepared in the laboratory (e.g. spiked organic matter on surfaces), should be considered.

References

Angen, Ø., Skov, M.N., Chriél, M., Agger, J.F. and Bisgaard, M. (1996) A retrospective study on Salmonella infection in Danish broiler flocks. Preventive Veterinary Medicine 26, 223-237.

Anonymous (2002) Desinfektion i Husdyrbruget . Landsudvalget for Svin - Danske Slagterier - Primærproduktion - Forsøg og udvikling, Denmark.

Engvall, A. (1993) Cleaning and disinfection of poultry houses. In International Course in Salmonella Control in Animal Production and Products - a Presentation of the Swedish Salmonella Programme, Malmö, Sweden, ed. Bengtson, S.O. pp. 155-159.

Gradel, K.O. (2002) Varmedesinfektionsprojektet: Resultater fra laboratorieforsøg. Dansk Erhvervsfjerkræ 7, 208-211.

Gradel, K.O. (2003) Desinfektionsmiddeltolerance og stationære salmonellainfektioner i fjerkræstalde: Er der en sammenhæng? Dansk Erhvervsfjerkræ 6, 181-183.

Gradel, K.O., Jørgensen, J.C., Andersen, J.S. and Corry, J.E.L. (2003a) Laboratory heating studies with Salmonella spp. and Escherichia coli in organic matter, with a view to decontamination of poultry houses. Journal of Applied Microbiology 94, 919-928.

Gradel, K.O., Jørgensen, J.C., Andersen, J.S. and Corry, J.E.L. (2003b) Monitoring the efficacy of steam and formaldehyde treatment of naturally Salmonella-infected layer houses. Submitted.

Gradel, K.O., Nielsen, B.L., Jessen, B. and Knudsen, V. (2002) Varmedesinfektionsprojektet: Resultater fra feltforsøg. Dansk Erhvervsfjerkræ 8, 238-244.

Gradel, K.O. and Randall, L.P. (2003) Minimum inhibitory concentrations of and adaptation to five disinfectants commonly used towards Salmonella in the poultry industry. In XI International Congress in Animal Hygiene, Mexico City, Mexico, ed. Saltijeral, J. pp. 339-344.

Gradel, K.O. and Rattenborg, E. (2003) A questionnaire-based, retrospective field study of persistence of Salmonella Enteritidis and Salmonella Typhimurium in Danish broiler houses. Preventive Veterinary Medicine 56, 267-284.

Linton, A.H., Hugo, W.B. and Russell, A.D. (1987) Practical aspects of disinfection and infection control. In Disinfection in Veterinary and Farm Animal Practice ed. Linton, A.H., Hugo, W.B. and Russell, A.D. pp. 144-167. London: Blackwell Scientific Publications.

Meroz, M. and Samberg, Y. (1995) Disinfecting poultry production premises. Revue scientifique et technique/Office international des épizooties 14, 273-291.

Svedberg, J. (1993) Salmonella sanitation in poultry farms - practical guidelines. In International Course on Salmonella Control in Animal Production and Products - A Presentation of the Swedish Salmonella Programme, Malmö, Sweden, ed. Bengtson, S.O. pp. 161-183.

Wegener, H.C., Hald, T., Wong, D.L., Madsen, M., Korsgaard, H., Bager, F., Gerner-Smidt, P. and Molbak, K. (2003) Salmonella control programs in Denmark. Emerging Infectious Diseases 9, 774-780.