The control of VTEC in the animal reservoir

Ž .International Journal of Food Microbiology 66 2001 71–78www.elsevier.nlrlocaterijfoodmicro

The control of VTEC in the animal reservoir

Dale Hancock), Tom Besser, Jeff Lejeune, Margaret Davis, Dan RiceField Disease InÕestigation Unit, Department of Veterinary Clinical Science, Washington State UniÕersity, Pullman, WA 99164-6610, USA

Abstract

A great diversity of VTECs exist but only in the case of Escherichia coli O157:H7, a common human foodbornepathogen, has sufficient research been done to allow generalizations about the ecology. The key features are as follows: lackof host specificity such that indistinguishable isolates can be found in a variety of species; near-ubiquitous distribution in

Ž .cattle and perhaps other ruminant farms; transient residence in the gastrointestinal flora of individual animals that is notassociated with clinical disease; temporal clustering at the population level such that most fecal shedding is confined to sharpbursts in a high percentage of animals separated by much longer periods of very low prevalence; a higher prevalence inyoung animals in comparison to older ones; a higher prevalence in animals with floral disturbance such as that caused bytransit, feed changes or antimicrobial dosing; and a markedly higher prevalence during warm months. Molecularepidemiological studies of E. coli O157:H7 have demonstrated that subtypes of the organism can persist on cattle farms foryears, thus supporting a conclusion that cattle farms represent a reservoir. Yet on such farms, common subtypes are oftenfound in environmental niches and in other species of animals; thus, it is not completely clear that cattle themselves are thereservoir. New subtypes are periodically observed on particular farms, and indistinguishable subtypes can be found on farmsthat are separated by hundreds of kilometers even in the absence of any obvious animal movements between them. The

Ž .number of subtypes found on a farm does not appear to be qualitatively correlated with cattle movements e.g., purchasesinto the farm. Commercial feeds are sometimes contaminated with E. coli O157:H7, and it seems likely that feeds representan important route of dissemination for this agent and other VTEC. Mixed feeds collected from feeding troughs arecommonly positive for E. coli O157:H7, as are cattle watering troughs, and feed and water likely represent the mostcommon means of infection. Environmental replication in feeds and in sediments of watering troughs occurs and mayaccount for the higher level of fecal shedding in the warm months. Since E. coli O157:H7 has been found to persist andremain infective for at least 6 months in water trough sediments, this may be an important environmental niche where theorganism survives during periods when it cannot be detected in cattle, especially during cold months. Traditional means ofcontrolling infectious agents, such as eradication or test and removal of carrier animals, do not appear to be feasible forVTECs. Nevertheless, certain farm management practices—especially those related to maintenance and multiplication of theagent in feed and water—may provide practical means to substantially reduce the prevalence of these agents in cattle onfarms and in those arriving at slaughter plants. q 2001 Elsevier Science B.V. All rights reserved.

Keywords: VTEC; Reservoir; E. coli

) Corresponding author. Tel.: q1-509-335-0711; fax: q1-509-335-0880.

Ž .E-mail address: [email protected] D. Hancock .

1. Introduction

The concept of controlling food borne diseaseagents, such as Escherichia coli O157:H7 and otherVTEC, in the livestock reservoir may seem foreign

0168-1605r01r$ - see front matter q 2001 Elsevier Science B.V. All rights reserved.Ž .PII: S0168-1605 00 00487-6

( )D. Hancock et al.r International Journal of Food Microbiology 66 2001 71–7872

to many in the research and public health communi-ties. Given the distinctly non-hygienic habits of live-stock, even the cleanest of farms seemingly has greatpotential for fecal–oral transmission. So many dif-ferent feeding, housing, and other variables exist insuch an endless variety on millions of farms that,even if one were inclined to seek pre-harvest inter-ventions, it would be difficult to know where tobegin. This was the quandary facing several groupsof researchers who, in the late 1980s or early 1990s,began to study E. coli O157:H7 and other VTEC onfarms.

2. Broad epidemiological features

On reflection, this immobilizing dilemma ofwhere-to-begin will be seen as no worse than thatfacing those who set out to solve any complexproblem. Until one has defined the broad character-istics of a problem, it is impossible to even formulatea cogent intervention hypothesis. Fortunately, theseveral research groups working on the epidemiol-ogy of VTEC on farms have now defined many ofthe major features of the epidemiology of E. coliO157:H7. These are summarized below and are dis-

Ž .cussed in detail elsewhere Hancock et al., 1998 .

v Near ubiquitous in cattle populations such that E.coli O157:H7 and other VTEC can be found onvirtually all farms, at least intermittently.

v Strongly seasonal with peak prevalence in summerand early fall.

v Transient residence in the gastrointestinal flora ofindividual animals that is not associated with clini-cal disease.

v Higher prevalence in immature animals.v Broad host range as manifested in isolations from

humans, cattle, sheep, deer, dogs, birds, horses,flies, and others.

v Temporal clustering at the population level suchthat most fecal shedding is confined to sharpbursts in a high percentage of animals separatedby much longer periods of very low prevalence.

v Complex molecular epidemiology such that sev-eral PFGE subtypes often exist on a farm simulta-neously with periodic additions and turnover, evenon farms that do not receive animals from outside.

v Capable of long distance transmission not associ-ated with animal movements, as indicated by in-distinguishable PFGE subtypes on closed farmsseparated by hundreds of kilometers.

These basic features of the epidemiology of E.coli O157:H7 place constraints on potential pre-harvest strategies as have been discussed in detail

Ž .elsewhere Hancock et al., 1998 . Eradication wouldseem to be futile for such a ubiquitous agent that isnot host specific. Efforts aimed at removing infectedanimals from a herd would seem pointless for anagent with such a short infection period and whichseems to do well in the environment. The concept ofpre-harvest testing of animals is complicated by thereality that an animal itself does not have to beintestinally colonized to have a contaminated hide,which is the immediate source of most carcass con-tamination. Any pre-harvest intervention wouldseemingly be limited to reducing prevalence ratherthan completely eliminating E. coli O157:H7 andother VTEC. A qualitative person would perhapsargue that it hardly seems worthwhile. A quantitativeone would point out that complex problems arenearly always solved incrementally and that the goalof achieving ‘‘mere’’ reductions in pathogen preva-lence at the pre-harvest level is not substantivelydifferent from pathogen reduction efforts at theslaughter, processing, and consumer levels of thefood chain. Simulation models have predicted thatpre-harvest reductions of E. coli O157:H7 preva-lence in cattle would result in substantial reductions

Žin contamination of beef Jordan et al., 1999;.USDA:FSIS, 2000 and consequent human disease

Ž .USDA:FSIS, 2000 . Consistent with this, a recentstudy found a moderately high correlation betweenprevalence of E. coli O157:H7 in pens of cattleŽ .fecal and hide at slaughter plants and the E. coliO157:H7 contamination rates of carcasses derived

Ž .from those pens of cattle Elder et al., 2000 .One can even interpret the aforementioned sea-

sonal pattern as evidence of the considerable poten-tial for pre-harvest interventions. The strong seasonalpattern of E. coli O157:H7 prevalence in cattle hasrecently been corroborated by research groups using

Žsensitive detection technologies Bonardi et al., 1999;Heuvelink et al., 1998; Van Donkersgoed et al.,

.1999 . A similar summer increase has been widely

( )D. Hancock et al.r International Journal of Food Microbiology 66 2001 71–78 73

Žreported in humans in temperate climates Wallace et.al., 2000; Michel et al., 1999; PHLS, 1999 . Though

previous speculation attributed the summer increaseof human E. coli O157:H7-associated disease to

Žseasonal consumption andror preparation e.g., out-.door cooking patterns, a recent British study re-

ported a seasonal pattern of E. coli O157:H7 con-tamination rates of retail meats that roughly agreeswith the seasonal pattern that has been observed in

Ž .cattle Chapman et al., 2000 . This harmony ofseasonal rates of E. coli O157:H7 in cattle, retailmeats, and human disease is evidence not only ofcausation but also of the potential for farm levelinterventions to be transmitted to the consumer in theform of lower risk. If rates of human disease declinein winter due to the cyclical decline in prevalence inthe reservoir, then it follows that a secular declinewould have a similar effect. The seasonal patternalso suggests certain intervention hypotheses relatedto replication of E. coli O157:H7 in livestock feedand water, as discussed below.

Another reason to consider on-farm interventionsis that a substantial fraction of total E. coli O157:H7in some regions seems to be attributable to directcattle contact andror to water or soil contaminationassociated with high cattle numbers within a vicinityŽCoia et al., 1998; MacDonald et al., 1996; Michel et

.al., 1999; Trevena et al., 1999 . It also seems proba-ble that many of the fruit- and vegetable-associatedoutbreaks are indirectly related to contamination withcattle feces.

3. Research on possible interventions

3.1. Feed and water hygiene

Although cattle and other livestock unavoidablyhave more exposure to fecal organisms than do adulthuman beings, most of this appears to occur via feed

Žand water. For example, the bunk feeds feed mixes.directly consumed by cattle on dairy farms in Wash-

ington State have generic E. coli concentrationsvarying from -101 CFUrg to greater than 104

Ž .CFUrg Lynn et al., 1998 . In an ongoing project,we have found that the bunk feed contaminationlevel is somewhat repeatable on a farm over time—

that is, some farms tend to have dirty feeds andothers clean. Considering that a high producing dairycow will commonly consume 35 kg or more of feedper day, at 104 CFUrg the total intake of generic E.coli from this source would be 3.5=108. Concentra-tions in water are also highly variable, but range

Žfrom -1 CFUrml to 100 CFUrml occasionally.much higher . It is difficult to estimate the amount of

fecal contamination that cows consume duringgrooming and social activities, but it seems unlikelythat it would approach the high-end numbers fromfeed and water. Hence, it seems probable that con-tamination levels of feed and water account for mostof the farm-to-farm variability in intake of fecalorganisms. The wide variability among farms indi-cates that feed and water contamination levels arepotentially subject to management control.

Ž .An earlier study by our group Lynn et al., 1998suggested that the high total E. coli counts on somefarms might be due to replication, and it was foundthat a variety of wet grain mixes and some silage-based mixes supported growth of generic E. coli andE. coli O157:H7. We hypothesized that this mightaccount for the seasonal pattern of E. coli O157:H7shedding. In our presently ongoing study, by con-trast, among )1200 bunk feed samples collected inwinter and summer from 20 farms, we have foundvery little evidence of replication of E. coli. Averagecounts in winter and summer are very similar. Silage,a near-universal bunk feed component in the north-western USA, appears to be consistently inhibitory togrowth of E. coli, E. coli O157:H7, and salmonella.The wide range of E. coli in bunk feeds on differentfarms appears to be due to contamination or replica-tion that occurs before purchase or during on-farmstorage and mixing of feeds. We are presently exam-ining these areas to determine key management vari-ables. Discerning the reasons for farm-to-farm vari-ability in counts of E. coli in water troughs has alsoproven more complex than one might think, but it isnoteworthy that, in contrast to bunk feeds, a signifi-cant increase in E. coli counts has been observed inwater troughs during summer months. E. coliO157:H7 has been relatively frequently found in

Ž . Ž .bunk feeds 1.8% and water troughs 3.8% . In thisŽstudy conducted in cooperation with the US Food

.and Drug Administration we will eventually follow30 herds for 1 year each to determine the correlation


between feed and water contamination levels andfecal prevalence of E. coli O157:H7 and salmonella.

Another aspect of feed hygiene that merits consid-eration relates to the finding of indistinguishable

Ž .subtypes even after two enzyme cuts on farms thatare separated by hundreds of kilometers and which

Ž .share no animal sources Rice et al., 1999 . ThoughE. coli O157:H7 has been reported in a variety ofwildlife species, it seems unlikely that wildlifemovement could account for the widespread occur-rence of common subtypes. A regionally dissemi-nated clonal subtype has been recently reported in

Ž .Scotland Allison et al., 2000 which also indicatessome means of regional dissemination. We hypothe-size that purchased feeds represent the primary vehi-cle for regional dissemination of E. coli O157:H7.Feeds are commonly contaminated with generic E.

Žcoli, indicating fecal contamination Lynn et al.,.1998 . In an ongoing project, we are collecting inter-

nal samples from feeds at the storage location whereŽthey are delivered to the study farms hence uncon-

.taminated by the farm . At present, 0.5% of suchŽ .samples 6 of 1100 have been found to contain E.

coli O157:H7. An additional positive sample wascollected at a feed mill. Although this prevalence ofcontamination may seem too low to play an impor-tant role in E. coli O157:H7 epidemiology, it isnoteworthy that even a small dairy farm typicallypurchases dozens of loads of feed in a year. Forlarger herds, the number easily reaches into thehundreds. PFGE subtyping of feed isolates and fecalisolates from the farms where they were collected isplanned and should shed light on the role of pur-chased feeds in the regional and farm-level epidemi-ology of E. coli O157:H7.

3.2. CompetitiÕe inhibition

In competitive inhibition, animals are fed or oth-erwise orally inoculated with bacteria that competewith or are inhibitory to the target organism. Theconcept is well established for reducing colonizationassociated with salmonellae in poultry and has been

Ž .discussed for other uses Nisbet, 1998 . At least oneŽgroup is working on a competitive inhibition or

.probiotic product for reducing E. coli O157:H7shedding in cattle and has published some promising

Ž .results Zhao et al., 1998 . Even if such a productreduced E. coli O157:H7 shedding by, say, only

50% in terms of number of days of shedding andrornumber of E. coli O157:H7 cells shed into theenvironment, it would likely have a marked effect onthe ecology of this agent in a population of cattle inthat it would reduce the level of environmental expo-sure to other animals. The ultimate decision onwhether such products will be effective will have tobe based on large field studies involving naturalexposure. Such studies have not yet been conductedto our knowledge.

3.3. Reductions in hide soiling

Since the immediate source of most bacteria onŽ .carcasses is the soiled hide Grau, 1987 , efforts to

reduce the level of hide soiling seem warranted forcontrol of E. coli O157 and other VTEC. Someslaughter companies in the USA have begun todiscriminate against heavily soiled animals. How-ever, two studies have presented data that lead todoubt about the degree to which visible soiling is auseful target for intervention aimed at reducing con-tamination of carcasses. Van Donkersgoed et al.Ž .1997 found little correlation between visible soilingand carcass bacterial counts. Based on best availableinformation and a stochastic simulation, Jordan et al.Ž .1999 estimated that the effects of industry-wide

Ž .reduction of visible hide soiling AtagsB would besmall. A recent study on pre-slaughter hide washingfound that power washing for 3 min was required tosignificantly reduce E. coli O157:H7 counts on con-taminated hides and that this procedure did not sig-nificantly reduce the concentration of E. coli

ŽO157:H7 transferred to the carcass Byrne et al.,.2000 .

3.4. Vaccination

Several groups are working on E. coli O157:H7vaccines for cattle, though we are aware of nobreakthroughs at this point. As with competitiveinhibition, even modest reductions in shedding couldhave major impacts on the dynamics of E. coliO157:H7 in a farm ecosystem.

3.5. Pre-slaughter hay feeding

The most widely publicized dietary interventionto date has been the recommendation to switch


grain-fed cattle to an all hay diet several days priorŽto slaughter Diez-Gonzalez et al., 1998; Russell et

.al., 2000 . According to the theory, this practice willŽ .result in lower levels of organic weak acids in the

colon and hence E. coli O157:H7 that are less acidresistant. Such E. coli O157:H7, according to thetheory, would be less likely to make it through thehuman gastric barrier. Among the possible pre-harvest interventions that are discussed here, this isthe only one for which the authors seemed to make arecommendation for widespread application in thecattle industry. Hence, in the interest of caution, it isworth considering some of the counter issues thathave been raised.

v ŽFirst, the aforementioned study Diez-Gonzalez.et al., 1998 did not directly examine acid resistance

of E. coli O157:H7 with and without hay feedingbut, rather, the acid resistance of generic E. coli. Toour knowledge, the only study to have directly exam-ined the effects of hay and grain feeding on acidresistance of E. coli O157:H7 in infected animals

Žfailed to find a difference associated with diet Hovde.et al., 1999 . This study also reported a markedly

increased average duration of shedding among cattleon the all hay diet.

v Secondly, most foodborne disease associatedwith E. coli O157:H7 is not thought to result fromconsumption of organisms immediately and directlyfrom the colon of cattle. The hide is thought to bethe immediate source of most bacterial contamina-

Ž .tion of carcasses Grau, 1987 , and, while bacteriaon the hide likely trace their ancestry to the gastro-intestinal tract, it is not clear that any effects ofcolonic environment would persist in an organismthat had survived, and possibly even replicated, forhours or days in the external environment. Even forthose E. coli that contaminate carcasses directlyfrom the colon, it is important to remember that beefis typically consumed 1 to 4 weeks after slaughterand that some replication is common, notably at the

Ž .retail level Gill and McGinnis, 1993 . Refrigerationhas been reported to increase acid tolerance of E.

Ž .coli O157:H7 Uljas and Ingham, 1998 . Thus, stor-age factors would seemingly tend to obviate anyinduced acid resistance acquired in the colonic envi-ronment. From the standpoint of evaluating the effectof hay feeding on human disease risk, the relevanttarget is not E. coli O157:H7 collected from the

colon before an animal is slaughtered but E. coliŽ .O157:H7 collected 2 weeks say after slaughter

from retail meats. Such a study has not been done, toour knowledge.

v Thirdly, most E. coli O157:H7 and other EHECare relatively acid resistant when in stationary phase

Žwithout being induced by weak acids Benjamin and.Datta, 1995 . We are aware of no epidemiological or

inoculation studies that support a conclusion that anyŽ .additional induced acid resistance decreases the

infectious dose of E. coli O157:H7 or otherwiseincreases the probability of infection. On the con-trary, even relatively acid sensitive E. coli are able

Žto make it through the gastric barrier Freter et al.,.1983 and are protected from the effects of acid by

Ž .association with food Watterman and Small, 1998 .ŽThe continual acquisition of new mainly commen-

.sal E. coli subtypes in food is evidently an ongoingŽ .phenomenon for humans Bettleheim et al., 1979 ,

and the notion that an E. coli, pathogenic or not,must possess extreme acid resistance to make itthrough the gastric barrier seems to be a conceptualerror.

v Fourthly, such a drastic change in diet immedi-ately before slaughter could possibly increase preva-lence of certain foodborne pathogens. Cattle are

Žcommonly exposed to salmonella in feed Kryten-.berg et al., 1998 and a variety of stresses can

Žincrease susceptibility to colonization Radostits et.al., 1994 . It seems possible to imagine that a switch

from a high grain diet to all forage might be such aŽ .stress. Furthermore, since volatile fatty acids VFAs

in the large intestine appear to be inhibitory toŽsalmonellae Que et al., 1986; Manning et al., 1994;

.McHan and Schotts, 1993 , the decline in colonicVFAs that accompanies the change to an all hay dietŽDiez-Gonzalez et al., 1998; Siciliano-Jones and

.Murphy, 1989 could result in increased susceptibil-ity to this agent. Certainly, the possibility wouldhave to be ruled out before the hay feeding methodcould be generally recommended.

4. Geographical differences

An area that has been little studied is the possibil-ity that differences in farming practices betweendifferent countries could account for some of the


wide differences in human disease associated withE. coli O157:H7 and other VTEC. Though one coulddismiss the variation in VTEC isolations as beinglaboratory and reporting artifacts, this seems unlikelyfor the 4-fold or greater variation in rates of hemor-

Žrhagic uremic syndrome seen in various regions De-cludt et al., 2000; Huber et al., 1998; Gianviti et al.,

.1994; Siegler et al., 1994; Waters et al., 1994 . Thelow rates of reported VTEC-associated disease inFinland, and seemingly other Scandinavian coun-tries, seems especially notable in comparison to some

Ž .other European countries Keskimaki et al., 1998 .Though one might be tempted to hypothesize aneffect of a colder climate, it seems noteworthy thatthe HUS rate in Italy is also lower than many other

Ž .European countries Gianviti et al., 1994 , thoughstill much higher than in Finland. Within the USA,northern states have historically reported a far higherrate of E. coli O157:H7-associated disease thansouthern ones, and the sentinel site surveillance pro-gram coordinated by the Centers for Disease Controlsuggests that the difference is not entirely due to

Ž .reporting bias Wallace et al., 2000 . It also seemsnoteworthy that a study employing identical methodsin feedlot cattle in northern and southern states foundno evidence of regional prevalence differencesŽ .Hancock et al., 1997 . Hence, it could be that theregional differences in human E. coli O157:H7-asso-ciated disease do not relate to regional farming dif-ferences. Yet, the regional differences in Europe arenot so easily dismissed. The one large study avail-able in Scandinavian cattle herds reported a much

Ž .lower rate of E. coli O157:H7 or H- in 197Norwegian herds than has been reported in other

Ž .European studies Vold et al., 1998 , hence it seemsvery possible that the low disease rates in that coun-try could be due to differences in the livestockreservoir. It seems noteworthy that Norway, Finland,and Sweden also have very low rates of domesticallyacquired human salmonellosis, and that this appearsto be due to low prevalence of salmonellae in domes-

Ž .tic livestock Kapperud et al., 1998 . It is interestingto speculate as to what aspects of farming mightaccount for differences in livestock prevalence andhuman disease incidence associated with E. coliO157:H7 between, say, Norway and Scotland. Un-covering the reasons for such differences presents anexcellent task for concerted action.

References

Allison, L.J., Carter, P.E., Thomson-Carter, F.M., 2000. Charac-terization of a recurrent clonal type of Escherichia coliO157:H7 causing major outbreaks of infection in Scotland. J.

Ž .Clin. Microbiol. 38 4 , 1632–1635, Apr.Benjamin, M.-M., Datta, A.-R., 1995. Acid tolerance of enterohe-

morrhagic Escherichia coli. Appl. Environ. Microbiol. 61,1669–1672.

Bettelheim, K.A., Cooke, E.M., O’Farrell, S., Shooter, R.A.,1979. The effect of diet on intestinal Escherichia coli. J. Hyg.Ž . Ž .London 79 1 , 43–45, Aug.

Bonardi, S., Maggi, E., Bottarelli, A., Pacciarini, M.L., Ansuini,A., Vellini, G., Morabito, S., Caprioli, A., 1999. Isolation ofverocytotoxin-producing Escherichia coli O157:H7 from cat-

Ž .tle at slaughter in Italy. Vet. Microbiol. 67 3 , 203–211, Jun.30.

Byrne, C.M., Bolton, D.J., Sheridan, J.J., McDowell, D.A., Blair,I.S., 2000. The effects of preslaughter washing on the reduc-tion of Escherichia coli O157:H7 transfer from cattle hides to

Ž .carcasses during slaughter. Lett. Appl. Microbiol. 30 2 ,142–145, Feb.

Chapman, P.A., Siddons, C.A., Cerdan Malo, A.T., Harkin, M.A.,2000. A one year study of Escherichia coli O157 in raw beef

Ž .and lamb products. Epidemiol. Infect. 124 2 , 207–213, Apr.Coia, J.E., Sharp, J.C.M., Campbell, D.M., Curnow, J., Ramsay,

C.N., 1998. Environmental risk factors for sporadic Es-cherichia coli O157 infection in Scotland: results of a descrip-

Ž .tive epidemiology study. J. Infect. 36 3 , 317–321, May.Decludt, B., Bouvet, P., Mariani-Kurkdjian, P., Grimont, F.,

Grimont, P.A., Hubert, B., Loirat, C., 2000. Haemolyticuraemic syndrome and Shiga toxin-producing Escherichia coliinfection in children in France. The Societe de Nephrologie

Ž .Pediatrique. Epidemiol. Infect. 124 2 , 215–220, Apr.Diez-Gonzalez, F., Callaway, T.R., Kizoulis, M.G., Russell, J.B.,

1998. Grain feeding and the dissemination of acid-resistantŽ .Escherichia coli from cattle. Science 281 5383 , 1666–1668,

Sep. 11.Elder, R.O., Keen, J.E., Siragusa, G.R., Barkocy-Gallagher, G.A.,

Koohmaraie, M., Laegreid, W.W., 2000. Correlation of en-terohemorrhagic Escherichia coli O157 prevalence in feces,hides, and carcasses of beef cattle during processing. Proc.

Ž .Natl. Acad. Sci. U. S. A. 97 7 , 2999–3003, Mar. 28.Freter, R., Brickner, H., Fakete, J., Vickerman, M.M., Carey,

K.E., 1983. Survival and implantation of Escherichia coli inthe intestinal tract. Infect. Immun. 39, 686–703.

Gianviti, A., Rosmini, F., Caprioli, A., Corona, R., Matteucci,M.C., Principato, F., Luzzi, I., Rizzoni, G., 1994.Haemolytic–uraemic syndrome in childhood: surveillance andcase-control studies in Italy. Italian HUS Study Group. Pedi-

Ž .atr. Nephrol. 8 6 , 705–709, Dec.Gill, C.O., McGinnis, C., 1993. Changes in the microflora on

commercial beef trimmings during their collection, distributionand preparation for retail sale as ground beef. Int. J. FoodMicrobiol. 18, 321–332.

Grau, F.H., 1987. Prevention of microbial contamination in theŽ .export beef abattoir. In: Smulders, F.J.M. Ed. , Elimination of


Pathogenic Organisms from Meat and Poultry. Elsevier, Ams-terdam, pp. 221–233.

Hancock, D.D., Rice, D.H., Thomas, L.A., Dargatz, D.A., Besser,T.E., 1997. Epidemiology of Escherichia coli O157 in feedlotcattle. J. Food Prot. 60, 462–465.

Hancock, D.D., Besser, T.E., Rice, D.H. et al., 1998. Ecology ofE. coli O157:H7 in cattle and impact of management prac-

Ž .tices. In: Kaper, J.B., O’Brien, A.D. Eds. , Escherichia coliO157:H7 and Other Shiga Toxin Producing E. coli strains.American Society for Microbiology Press, Washington, D.C.,USA, pp. 85–91.

Heuvelink, A.E., van den Biggelaar, F.L., Zwartkruis-Nahuis, J.,Herbes, R.G., Huyben, R., Nagelkerke, N., Melchers, W.J.,Monnens, L.A., de Boer, E., 1998. Occurrence of verocyto-toxin-producing Escherichia coli O157 on Dutch dairy farms.

Ž .J. Clin. Microbiol. 36 12 , 3480–3487, Dec.Hovde, C.J., Austin, P.R., Cloud, K.A., Williams, C.J., Hunt,

C.W., 1999. Effect of cattle diet on Escherichia coli O157:H7Ž .acid resistance. Appl. Environ. Microbiol. 65 7 , 3233–3235,

Jul.Huber, H.C., Kugler, R., Liebl, B., 1998. Infections with entero-

Ž .hemorrhagic Escherichia coli EHEC —results of an epi-demiologic survey in Bavaria for the April 1996 to May 1997

Ž .time frame. Gesundheitswesen 60 3 , 159–165, Mar.Jordan, D., McEwen, S.A., Lammerding, A.M., McNab, W.B.,

Wilson, J.B., 1999. Pre-slaughter control of Escherichia coliO157 in beef cattle: a simulation study. Prev. Vet. Med. 41Ž .1 , 55–74, Jun. 29.

Kapperud, G., Lassen, J., Hasseltvedt, V., 1998. Salmonella infec-tions in Norway: descriptive epidemiology and a case-control

Ž .study. Epidemiol. Infect. 121 3 , 569–577, Dec.Keskimaki, M., Saari, M., Heiskanen, T., Siitonen, A., 1998.

Shiga toxin-producing Escherichia coli in finland from 1990through 1997: prevalence and characteristics of isolates. J.

Ž .Clin. Microbiol. 36 12 , 3641–3646, Dec.Krytenburg, D., Hancock, D.D., Rice, D.H., Besser, T.E., Gay,

C.C., Gay, J.M., 1998. Salmonella enterica in cattle feedsfrom the Pacific Northwest. Anim. Feed Sci. Technol. 75,75–79.

Lynn, T.V., Hancock, D.D., Besser, T.E., Harrison, J.H., Rice,D.H., Stewart, N.T., Rowan, L.L., 1998. The occurrence andreplication of Escherichia coli in cattle feeds. J. Dairy Sci. 81Ž .4 , 1102–1108, Apr.

MacDonald, I.A., Gould, I.M., Curnow, J., 1996. Epidemiology ofinfection due to Escherichia coli O157: a 3-year prospective

Ž .study. Epidemiol. Infect. 116 3 , 279–284, Jun.Manning, J.G., Hargis, B.M., Hinton Jr., A., Corrier, D.E., De-

Loach, J.R., Creger, C.R., 1994. Effect of selected antibioticsand anticoccidials on Salmonella enteritidis cecal colonization

Ž .and organ invasion in Leghorn chicks. Avian Dis. 38 2 ,256–261, Apr.–Jun.

McHan, F., Shotts, E.B., 1993. Effect of short-chain fatty acids onthe growth of Salmonella typhimurium in an in vitro system.

Ž .Avian Dis. 37 2 , 396–398, Apr.–Jun.Michel, P., Wilson, J.B., Martin, S.W., Clarke, R.C., McEwen,

S.A., Gyles, C.L., 1999. Temporal and geographical distribu-tions of reported cases of Escherichia coli O157: H7 infection

Ž .in Ontario. Epidemiol. Infect. 122 2 , 193–200, Apr.

Nisbet, D.J., 1998. Use of competitive exclusion in food animals.J. Am. Vet. Med. Assoc. 213, 1744–1746, Dec. 15.

Ž .Public Health Laboratory Service of England and Wales PHLS ,1999. Disease facts: Escherichia coli O157. http:rrwww.phls.co.ukrfactsrGastrorecoliQua.htm.

Que, J.-U., Casey, S.W., Hentges, D.J., 1986. Factors responsiblefor increased susceptibility of mice to intestinal colonizationafter treatment with streptomycin. Infect. Immun. 53, 116–123.

Radostits, O.M., Blood, D.C., Gay, C., 1994. Veterinary Medicine.Balliere Tindall, London, 733.

Rice, D.H., McMenamin, K.M., Pritchett, L.C., Hancock, D.D.,Besser, T.E., 1999. Genetic subtyping of Escherichia coliO157 isolates from 41 Pacific Northwest USA cattle farms.

Ž .Epidemiol. Infect. 122 3 , 479–484, Jun.Russell, J.B., Diez-Gonzalez, F., Jarvis, G.N., 2000. Invited re-

view: effects of diet shifts on Escherichia coli in cattle. J.Ž .Dairy Sci. 83 4 , 863–873, Apr.

Siciliano-Jones, J., Murphy, M.R., 1989. Production of volatilefatty acids in the rumen and cecum-colon of steers as affectedby forage:concentrate and forage physical form. J. Dairy Sci.72, 485–492.

Siegler, R.L., Pavia, A.T., Christofferson, R.D., Milligan, M.K.,1994. A 20-year population-based study of postdiarrheal

Ž .hemolytic uremic syndrome in Utah. Pediatrics 94 1 , 35–40,Jul.

Trevena, W.B., Willshaw, G.A., Cheasty, T., Domingue, G.,Wray, C., 1999. Transmission of Vero cytotoxin producingEscherichia coli O157 infection from farm animals to humansin Cornwall and west Devon. Commun. Dis. Public Health 2Ž .4 , 263–268, Dec.

Uljas, H.E., Ingham, S.C., 1998. Survival of Escherichia coliO157:H7 in synthetic gastric fluid after cold and acid habitua-tion in apple juice or trypticase soy broth acidified with

Ž .hydrochloric acid or organic acids. J. Food Prot. 61 8 ,939–947, Aug.

Ž .United States Department of Agriculture USDA : Food SafetyŽ .Inspection Service FSIS , 2000. Risk assessment for E. coli

O157:H7 in ground beef. http:rrwww.fsis.usda.govrOPHSrecolriskr.

Van Donkersgoed, J., Jericho, K.W.F., Grogan, H., Thorlakson,B., 1997. Preslaughter hide status of cattle and the micro-biology of carcasses. J. Food Prot. 60, 1502–1508.

Van Donkersgoed, J., Graham, T., Gannon, V., 1999. The preva-lence of verotoxins, Escherichia coli 0157:H7, and Salmonellain the feces and rumen of cattle at processing. Can. Vet. J. 40Ž .5 , 332–338, May.

Vold, L., Klungseth Johansen, B., Kruse, H., Skjerve, E., Waste-son, Y., 1998. Occurrence of shigatoxinogenic Escherichiacoli O157 in Norwegian cattle herds. Epidemiol. Infect. 120Ž .1 , 21–28, Feb.

Wallace, D.J., Van Gilder, T., Shallow, S., Fiorentino, T., Segler,S.D., Smith, K.E., Shiferaw, B., Etzel, R., Garthright, W.E.,Angulo, F.J., 2000. Incidence of foodborne illnesses reported

Žby the foodborne diseases active surveillance network Food-. Ž .Net -1997. J. Food Prot. 63 6 , 807–809, Jun.

Waterman, S.R., Small, P.L., 1998. Acid-sensitive entericpathogens are protected from killing under extremely acidicconditions of pH 2.5 when they are inoculated onto certain


Ž .solid food sources. Appl. Environ. Microbiol. 64 10 , 3882–3886, Oct.

Waters, J.R., Sharp, J.C., Dev, V.J., 1994. Infection caused byEscherichia coli O157:H7 in Alberta, Canada, and in Scot-

Ž .land: a five-year review, 1987–1991. Clin. Infect. Dis. 19 5 ,834–843, Nov.

Zhao, T., Doyle, M.P., Harmon, B.G., Brown, C.A., Mueller,P.O., Parks, A.H., 1998. Reduction of carriage of enterohem-orrhagic Escherichia coli O157:H7 in cattle by inoculation

Ž .with probiotic bacteria. J. Clin. Microbiol. 36 3 , 641–647,Mar.

Documents

The control of VTEC in the animal reservoir