The fate of Escherichia coli O157 in soil and its potential to contaminate drinking water

The fate of Escherichia coli O157 in soil and its potential to contaminate drinking water

International Journal of Food Microbiology 66 Ž2001. 111–117 www.elsevier.nlrlocaterijfoodmicro The fate of Escherichia coli O157 in soil and its pot...

96KB Sizes 1 Downloads 29 Views

International Journal of Food Microbiology 66 Ž2001. 111–117 www.elsevier.nlrlocaterijfoodmicro

The fate of Escherichia coli O157 in soil and its potential to contaminate drinking water Iain D. Ogden a,) , David R. Fenlon b, Andrew J.A. Vinten c , Douglas Lewis c a

Applied Food Microbiology Group, Department of Medical Microbiology, UniÕersity of Aberdeen, Foresterhill, Aberdeen, AB25 2ZD, UK b Scottish Agricultural College, Microbiology Department, Craibstone, Aberdeen, AB21 9YA, UK c Scottish Agricultural College, Land Management Department, Kings Building, West Mains Road, Edinburgh, EH9 3JG, UK

Abstract The survival and transport of Escherichia coli and E. coli O157 after cattle slurry application were studied on drained plots in both grassland and arable stubble at three sites in Scotland. Leaching losses were between 0.2% and 10% of total E. coli and were dependent on rainfall. Recovery of E. coli in grass and soil declined with approximately first order kinetics. Residual numbers, in excess of background declined more slowly. The pattern was similar for both grass and arable plots. Laboratory incubations of soil cores, with applied slurry containing E. coli and E. coli O157 were performed in soils with different moisture contents at two temperatures for clay loam and sandy loam soils. Both E. coli populations were measured over a 4-week period. Using a dual population approach, the die off of the susceptible pool was linear with a half-life of 3–4 days, and was faster at the higher temperature and lowest moisture content. The resistant pool was not strongly affected by temperature or moisture and had a half-life for die off of between 18 and 24 days. After a 4-week period, - 100 cfu grsoil of E. coli and E. coli O157 remained. The die off rate of E. coli O157 was the same or slightly faster than that of the commensal E. coli population, indicating that the field behaviour of E. coli O157 can be studied by monitoring the total population of E. coli applied with slurry. The risk of significant pollution of water by E. coli is highest immediately after application of slurry, and the first increments of drainflow carry significant concentrations. Thereafter, the risk of pollution is very low. If weather conditions are dry after application on well-drained sandy soils, it is unlikely that any significant losses of organisms to drains will occur. Such data can be used to control and minimise the risk of E. coli O157 contaminating drinking water. q 2001 Elsevier Science B.V. All rights reserved. Keywords: E. coli O157; Survival; Soils; Slurry; Water

1. Introduction The disposal of both animal and human wastes of faecal origin by application to land has become routine in the UK. The method of treatment and )

Corresponding author. Fax: q44-122-468-5604. E-mail address: [email protected] ŽI.D. Ogden..

application of sewage sludges is laid down in the UK Code of Practice for the Agricultural Use of Sewage Sludge ŽAnonymous, 1996., the general principle being to limit the risk of spread of disease causing organisms. The annual amount of sewage sludge applied to land in the UK was recently estimated at 4.3 = 10 5 tonnes dry weight ŽDavis and Rudd, 1997.. They also estimated 2.1 = 10 7 tonnes dry weight

0168-1605r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 1 6 0 5 Ž 0 0 . 0 0 5 0 8 - 0

112

I.D. Ogden et al.r International Journal of Food Microbiology 66 (2001) 111–117

annually of stored and treated animal wastes disposed to land. In addition, faecal wastes are directly deposited onto the land by grazing animals. The application of animal wastes is subject to Codes of Practice ŽAnonymous, 1997, 1998., but it is more difficult to monitor the disposal of agricultural material than of sewage sludge. The disposal of waste of faecal origin entails a major risk of the spread of specific bacterial pathogens, as food animals are reservoirs of pathogens that can enter the food chain and cause human infection. The emergence of pathogens such as Campylobacter jejuni, Cryptosporidium parÕum and Escherichia coli O157, all of which have low infective doses, have changed our perceptions of the risks associated with faecal waste disposal. Such risks are of particular concern in rural areas of Scotland where intensive agricultural practices and continued use of untreated private water supplies have led to localised outbreaks of E. coli ŽLicence et al., 2000.. The disposal of agricultural waste has been related to cases of E. coli O157 in the human population. Chapman et al. Ž1997a. reported an outbreak of E. coli O157 linked to potatoes fertilised with manure from cattle. Jackson et al. Ž1998. reported E. coli O157 infection linked to the contamination of private water supplies, and infection has also be associated with water used for recreational activity ŽMoore et al., 1993.. Cross-infection of livestock by drinking water is well-documented ŽShere et al., 1998.. Animal products themselves are an obvious source of human infection, particularly meat and dairy products. An outbreak of E. coli O157 infection in central Scotland in 1996, caused by contaminated meat products, led to the deaths of 21 elderly people, emphasising the severe nature of this pathogen ŽAhmed and Donaghy, 1998.. E. coli O157 is reported to be excreted by a significant number of cattle in the UK ŽChapman et al, 1997b. and this level of report is likely to increase as improved detection techniques become more widely applied. A significant proportion of slurry and farmyard manure spread onto agricultural land will inevitably be contaminated with E. coli O157. Thus it is essential to obtain an accurate assessment of such potential risk posed by these animal wastes.

The objective of this study was to monitor and compare the survival in soil and water of faecal E. coli and E. coli O157 applied to land in cattle slurry. The sporadic and sparse nature of excretion of the E. coli O157 from cattle mean that direct determination of survival rates in field trials is not always possible. The experimental design has therefore been to measure the movement and survival of E. coli in the field scale studies, linked to laboratory studies using slurry inoculated soil cores to examine E. coli and E. coli O157 survival.

2. Materials and methods 2.1. Detectionr estimation of E. coli and E. coli O157 E. coli numbers in soils and waters containing high Ž) 10 3 . bacterial numbers were determined by plating 10-fold serial dilutions Ž0.1ml. on cellulose acetate membranes ŽWhatman 0.45 mm. placed on the surface of Minerals Modified Glutamate Agar ŽMMGB. ŽOxoid CM607.. Inoculated membranes were incubated at 378C for 4 h to allow recovery of stressed cells, transferred to tryptone bile agar ŽOxoid CM509., and incubated at 448C for 18–24 h. Presumptive E. coli colonies were identified by a pink colour change when membranes were placed onto filter paper soaked with indole reagent ŽAnderson and Baird-Parker, 1975.. E. coli were confirmed by growth and gas production in brilliant green bile broth ŽOxoid CM31. and by indole production from tryptone water at 448C. When E. coli numbers fell below the level of detection by the plating technique Ž- 10 3 ., a three-tube most probable number ŽMPN. method using minerals modified glutamate broth ŽMMGB. was performed ŽAnonymous, 1967.. Total coliforms were determined by acid and gas production in MMGB incubated at 378C for 48 h and the presence of E. coli were confirmed using the method above. Escherichia coli O157 were enriched from field lysimeter ŽVinten et al., 1991. samples of soil, water, slurry and Moore swabs Žabsorbent cotton pads located in drains, Moore, 1948. in buffered peptone water ŽOxoid CM509. containing 8 mgrl van-

I.D. Ogden et al.r International Journal of Food Microbiology 66 (2001) 111–117

113

Table 1 Details of lysimeter trials Site

Soil type

Slurry applied Žtrha.

Date applied

E. colirml

E. coli O157r100 ml

Glencorse Žgrass q arable plots. Dumfries Žgrass plots. Glencorse Žgrass q arable plots.

Clay loam Silty clay loam Sandy loam

40 50 30

8 Mar 99 17 Feb 99 9 Nov 99

5 = 10 4 2 = 10 4 1 = 10 5

30 33 Not detected

comycin incubated for 6 h at 378C, and recovered by immuno magnetic separation ŽIMS. using Captivatee O157 immunomagnetic beads ŽInternational Diagnostics Group, Bury, UK.. The beads were plated onto sorbitol MacConkey agar ŽSMAC, Oxoid CM813. containing cefixime Ž0.05 mgrl. and potassium tellurite Ž2.5 mgrl., ŽCTSMAC., and incubated for 18–24 h at 378C. Presumptive E. coli O157 were confirmed by latex agglutination with E. coli O157 latex kit ŽOxoid DR620.. E. coli O157 numbers in slurry were estimated by a three-tube MPN technique ŽAnonymous, 1967. with presencerabsence in each dilution determined by the above IMS procedure. 2.2. Lysimeter trial The main field trial was carried out on 9 March 1999 at Glencorse, Midlothian, East Central Scotland ŽTable 1. with the application of 19.6 m3rha dairy cattle slurry Ž5% dry matter. to four drained plots on clay loam soil, Žtwo spring barley stubble and two grass. ŽFig. 1.. Each drained plot had a nominal area

of 600 m2 and was instrumented with tipping bucket flowmeters and water sampling devices ŽVinten et al., 1991.. Drainage was sampled from one half of each plot and surface runoff was sampled from the other half of each plot. Water samples were collected 15, 12 and 6 days before application of slurry and 1, 3, 7, 10, 14, 17, 21, 29, 35 and 49 days after slurry application. Composite water samples, representative of the previous 2–4-day period prior to each collection date were collected. This gave complete coverage of the water quality during the 3 weeks after slurry application. Moore swabs were placed upstream in the line of flow from the plot drains and surface runoff collectors, which effectively sampled for the same period as the composite water samples. Moore swabs were sent for microbiological analyses and were replaced with sterile specimens. Simultaneous trials were carried out at Dumfries ŽSouth West Scotland. on a silty clay loam soil ŽTable 1.. Eight months later, a similar trial was conducted on a separate Glencorse site with sandy loam soil ŽTable 1.. Similar sampling protocols were employed on all plots. 2.3. Laboratory studies

Fig. 1. Layout of the Glencorse lysimeter plots.

Intact soil cores from the two Glencorse sites Žclay loam and sandy loam. were adjusted for moisture content Žmatric potentials of y5 Žwet., y20 and y100 Ždry. kPa. and transferred to the laboratory. Slurry was applied, equivalent to 50 tonnesrha containing natural levels of E. coli at 9.6 = 10 4rml supplemented with 3.5 = 10 4 E. coli O157rml Žone environmental strain plus one food strain from the Aberdeen University laboratory collection.. Cores were stored at either 68C or 158C. Total E. coli and E. coli O157 numbers were determined after destructive sampling of cores over a 4-week period by

114

I.D. Ogden et al.r International Journal of Food Microbiology 66 (2001) 111–117

Fig. 2. Drainflow and E. coli concentrations in drains from the Glencorse clay loam grass plot.

direct counting onto MacConkey and SMAC agars incubated at 428C.

3. Results 3.1. Lysimeter studies 3.1.1. Glencorse clay loam The slurry used on this site contained 5.3 = 10 4 E. colirg and its application to the Glencorse test

plots resulted in initial E. coli load of approximately 10 8rm 2 . The slurry also contained low numbers Ž30r100 ml. of naturally occurring E. coli O157 ŽPT2, VT2.. There was no significant difference between losses of E. coli from the arable and grassland plots. Fig. 2 gives an example of the water flow and total E. coli concentration measured in this water from one of the grassland plots. Although estimations of E. coli concentration are reported as discrete values linked to discrete sample recovery occasions, they each represent the mean value for the continuous period since the recovery of the previous sample. The duration of each sample period is indicated by a horizontal line. Fig. 3 shows the numbers of E. coli recovered in soil and water samples following slurry application Žnormalised for the numberŽs. of E. coli applied. and the associated drainflows. In the first 3 days following slurry application, the numbers of E. coli on soil and grass samples remained constant, thereafter declining to - 1% after 29 days. Transport to drains was mainly associated with rainfall events between days 3 and 7. The majority of E. coli leaching losses was in the first week of drainflow after application Ž5 mm drainflow.. Recovery of E.

Fig. 3. Normalised recovery of E. coli in soils and water from three lysimeter plots.

I.D. Ogden et al.r International Journal of Food Microbiology 66 (2001) 111–117

coli from grass and soil samples declined exponentially with time to - 1% of applied numbers of E. coli in 3 weeks. Thereafter, residual numbers Žin excess of background. declined more slowly. The decline was similar for both grass and arable plots. No E. coli O157 were detected in any of the Moore swabs located in the drains. Soil samples tested positive for E. coli O157 up to 7 days after slurry application, and negative thereafter. Grass samples contained 600–8800 E. colirg after 7 days and also contained low numbers E. coli O157 Ž100rg, absent by direct plating, qve by IMS.. 3.1.2. Dumfries silty clay loam Table 1 shows the details of the slurry application to this site and it is interesting to note that like the previous study, low numbers Ž33r100 ml. of naturally occurring E. coli O157 ŽPT8, VT1, VT2. were present. Almost 10% of applied E. coli were leached to the drains in the first 12 days after application. Numbers of E. coli in soil and grass declined to 14% of initial applied numbers in the first 12 days, but was still above 10% after 5 weeks. This recovery was higher than found in the Glencorse site and may have been due to additional faecal E. coli from flocks of grazing geese. Fig. 3 shows the decline of applied E. coli in soil and water following slurry application, normalised for the number of E. coli applied. 3.1.3. Glencorse sandy loam Slurry containing E. coli at 10 5rg but absent from naturally occurring E. coli O157 was applied. There was little drainflow in the 2 weeks following application and - 0.2% of the E. coli applied were leached to the drains in the 3 months after application Ž260 mm drainflow.. Recovery of E. coli in soil and grass declined at a similar rate to the Glencorse clay loam site. Fig. 3 shows the recovery of applied E. coli in soil and water following slurry application, normalised for the number of E. coli applied. 3.2. Laboratory studies Escherichia coli and E. coli O157 numbers in soil cores stored at 68C or 158C declined in a biphasic manner and were modelled using a dual popula-

115

Table 2 Decline of E. coli in soil cores Decay parameters

68C

158C

Susceptible pool half life Ždays. Resistant pool half life Ždays. Resistant pool fraction % R2

4.1 17.8 3.1 0.83

3.3 23.5 3.8 0.96

R 2 sGoodness of fit.

tion approach ŽTable 2., reflecting the presence of susceptible and resistant sub-populations. Decline in numbers in the susceptible sub-population Žwhich comprised ) 95% of total population. was linear with a half-life of approximately 3–4 days. The rates of decline in the susceptible sub-population were more rapid under higher temperature Ž158C. and low moisture Žy100 kPa. conditions. Decline in numbers in the resistant pool was also linear with a half-life of approximately 18 to 24 days. The rates of decline in the resistant population were not significantly different across the range of temperature and moisture contents applied to soil cores during the study. After 18–24 days storage, E. coli and E. coli O157 numbers in soil cores were - 100 organismsrg. The die off rate of E. coli O157 was the same or slightly quicker than that of the whole E. coli population.

4. Discussion The lysimeter studies showed that the risks of water pollution by E. coli is greatest immediately after an application of slurry, and the first increments of drainflow Žfrom rainfall events. carry significant bacterial concentrations. It is unclear from these studies what the effect of no rainfall Ž2–3 weeks. would be on E. coli numbers. The risk of contamination appears to diminish after the first rainfall event although it must be remembered that low numbers of E. coli O157 entering untreated drinking water Že.g. rural private water supplies. can still be hazardous due to its low infectious dose ŽWillshaw et al., 1994.. If weather conditions are dry after the application on well-drained sandy soils, it is unlikely that any significant losses of bacteria to drains will occur. Under these conditions, micro-organisms may remain on the soilrgrass surface where their numbers

116

I.D. Ogden et al.r International Journal of Food Microbiology 66 (2001) 111–117

will decline due to desiccation and exposure to natural UV radiation. It is significant to note that two of the three slurries used in this study were found to contain naturally occurring E. coli O157, and that these naturally occurring strains containing VT genes and were therefore potentially pathogenic. Although present at concentrations too low to accurately enumerate by the methods used in this study, their presence on soil and grass suggests that cross contamination of both people and food by contaminated slurries should not be over-looked. The laboratory studies using cores from sandy loam and clay loam soils inoculated with cocktails of E. coli and E. coli O157 showed that E. coli O157 numbers decline at similar or slightly faster rates that the general E. coli population. These results suggest the field behaviour of E. coli O157 applied in slurries can be safely and reliably monitored by estimating the total population of E. coli in such materials. This observation is particularly important as commensal E. coli occur much more frequently in agricultural and environmental samples and can be more easily detected by widely available and routine microbiological tests. The use of commensal E. coli will facilitate further research on the prediction of E. coli O157 persistence in different soil types and different climates Že.g. rainfall and temperature., as such commensals do not need the specific handling and containment requirements of the pathogenic strain.

5. Conclusions The results of this study indicate steps that could be taken to enhance the levels of safety offered by current agricultural guidelines in relation to the control of E. coli O157 dissemination and potential risks of drinking water contamination. The current UK recommendations to avoid grazing or harvesting for 3 weeks after application of slurry allows most added E. coli to die or be washed off the foliage. However, slurry derived E. coli are still detectable after 3 weeks and an extension of the ‘no grazing’ period to 4 weeks would allow a greater decline in surviving E. coli and E. coli O157 numbers. Furthermore, the information from this study in relation

to the decline in E. coli O157 numbers, and the relationship between such numbers and non-E. coli O157, should be of value in risk assessment exercises designed to reduce the possibility of cross contamination of E. coli O157 to the wider food chain.

Acknowledgements The authors thank the Scottish Executive Rural Affairs Department who wholly funded this work and the large number of co-workers involved.

References Ahmed, S., Donaghy, M., 1998. An outbreak of Escherichia coli O157:H7 in central Scotland. In: Kaper, J.B., O’Brien, A.D. ŽEds.., Escherichia coli O157:H7 and Other Shiga-Toxin-Producing E. coli Strains. Am. Soc. for Microbiol., Washington, DC, pp. 59–65. Anderson, J.M., Baird-Parker, A.C., 1975. A rapid and direct plate method for enumerating Escherichia coli Biotype 1 in food. J. Appl. Bacteriol. 39, 111–117. Anonymous, 1967. Multiple-tube fermentation technique. Standard Methods for the Examination of Water and Wastewater Including Bottom Sediments and Sludges. American Public Health Association, New York. Anonymous, 1996. Department of the Environment, Code of Practice for the Agricultural Use of Sewage Sludge. 2nd edn. DTER Publications, London. Anonymous, 1997. SOAEFD, Prevention of Environmental Pollution from Agricultural Activity. A Code of Good Practice. HMSO, Edinburgh. Anonymous, 1998. MAFF, Code of Good Agricultural Practice for the Protection of Water. MAFF Publications, London. Chapman, P.A., Siddons, C.A., Manning, J., Cheetham, C., 1997a. An outbreak of infection due to verocytotoxin-producing Escherichia coli O157 in four families: the influence of laboratory methods on the outcome of the investigation. Epidemiol. Infect. 119, 245–250. Chapman, P.A., Siddons, C.A., Cerdan-Malo, A.T., Harkin, M.A., 1997b. A 1 year study of Escherichia coli O157 in cattle, sheep, pigs and poultry. Epidemiol. Infect. 119, 113–119. Davis, R.D., Rudd, C., 1997. Investigation of the criteria for, and guidance on, the landspreading of industrial wastes. Wrc report No. 40088r5 to the Environment Agency and MAFF. Wrc Medmenham, Marlow, UK. Jackson, S.G., Goodbrand, R.B., Johnson, R.P., Odorico, V.G., Alves, D., Rahn, K., Wilson, J.B., Welch, M.K., Khakhria, R., 1998. Escherichia coli O157:H7 diarrhoea associated with well water and infected cattle on an Ontario farm. Epidemiol. Infect. 120, 17–20.

I.D. Ogden et al.r International Journal of Food Microbiology 66 (2001) 111–117 Licence, K., Oates, K.R., Synge, B.A., Reid, T., 2000. An outbreak of E. coli O157 infection with evidence of spread from animals to man through contamination of a private water supply. Epidemiol. Infect. in press. Moore, B., 1948. The detection of paratyphoid carriers in towns by means of sewage examination. Mon. Bull. Ministry Health Public Health Lab. Service 7, 241–248. Moore, A.C., Herwaldt, B.L., Craun, G.F., Calderon, R.L., Highsmith, A.K., Juanek, D.D., 1993. Surveillance for waterborne disease outbreaks United States 1991–1992. M.M.W.R. 42, 1–22.

117

Shere, J.A., Bartlett, K.J., Kasper, C.W., 1998. Longitudinal study of Escherichia coli O157:H7, dissemination on four dairy farms in Wisconsin. Appl. Environ. Microbiol. 64, 1390–1399. Vinten, A.J.A., Howard, R.S., Redman, M.H., 1991. Measurement of nitrate leaching losses from arable plots under reduced and AalternativeB N input regimes. Soil Use Manage. 7, 3–14. Willshaw, G.A., Thirlwell, J., Jones, A.P., Parry, S., Salmon, R.L., Hickey, M., 1994. Verocytotoxin-producing Escherichia coli O157 in beefburgers linked to an outbreak of diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome in Britain. L. Appl. Microbiol. 19, 304–307.