ARTICLE IN PRESS

FOOD MICROBIOLOGY
Food Microbiology 21 (2004) 203212

www.elsevier.nl/locate/jnlabr/yfmic

Prevalence and numbers of Escherichia coli O157:H7 in minced beef and beef burgers from butcher shops and supermarkets in the Republic of Ireland
C. Cagneya, H. Crowleya, G. Duffya,*, J.J. Sheridana, S. OBriena, E. Carneya, W. Andersonb, D.A. McDowellc, I.S. Blairc, R.H. Bishopc
a b

The National Food Centre, Teagasc, Dunsinea, Castle Knock, Ashtown, Dublin 15, Ireland The Food Safety Authority of Ireland, Abbey Court, Lower Abbey Street, Dublin 1, Ireland c The University of Ulster, Newtownabbey, Co. Antrim, BT37 0QB, Northern Ireland, UK Received 23 April 2003; accepted 15 May 2003

Abstract This study investigated the prevalence and numbers of Escherichia coli O157:H7 in minced beef and beef burgers in supermarkets and butcher shops in the Republic of Ireland. Fifteen samples were collected quarterly from each of 26 counties over a 13-month period. All samples (n=1533) were (1) directly plated on SMAC, and (2) enriched in mTSB with novobiocin, extracted by immunomagnetic separation (IMS), plating onto SMAC-CT agar and nally conrmed by PCR. Overall, E. coli O157:H7 was recovered from 43 samples (2.80%) with counts ranging from o0.524.03 log10 cfu g1. Of the positive samples, 2.70% (32/1183) were purchased from supermarkets and 3.14% (11/350) from butcher shops. Only one product type (fresh unpacked burgers from supermarkets) was negative for E. coli O157:H7. Of the products containing the pathogen, fresh packaged burgers from supermarkets had the highest prevalence of 4.46% (7/157) while fresh unpackaged mince purchased from supermarkets had the lowest prevalence of 2.01% (6/299). Of the 43 isolates recovered, 41 possessed verotoxin-producing genes (vt1 and vt2), E. coli attaching and effacing gene (eaeA), haemolysin gene (hlyA), 60-MDa plasmid or rfb gene cluster that encodes for the biosynthesis of the O-antigen (pO157) and agellar H7 antigen encoding gene (iCh7). The remaining 2/43 isolates contained only one of the verotoxin-producing genes (vt1 or vt2) and all the other genes named. r 2003 Elsevier Ltd. All rights reserved.
Keywords: Escherichia coli O157:H7; Minced beef; Beef burgers; Enumeration; Detection

1. Introduction Escherichia coli O157:H7 has emerged as an important foodborne pathogen of considerable public health concern, because of the severity of infection which it causes and an infectious dose which may be as low as 10 organisms (Coia, 1998). In an outbreak study reported by Willshaw et al. (1994) contamination levels in an implicated product were reportedly as low as 2 cells per 25 g. This pathogen has been implicated in a number of high-prole outbreaks in the USA (Bell et al., 1994), Scotland (Ahmed and Donaghy, 1998) and Japan
*Corresponding author. Tel.: +353-1-8059500; fax: +353-1805-9550. E-mail address: gduffy@nfc.teagasc.ie (G. Duffy). 0740-0020/03/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0740-0020(03)00052-2

(Michino et al., 1998) as well as in many sporadic cases of infection. Cattle are still regarded as one of main reservoirs of E. coli O157:H7, with the pathogen occurring in the faeces (Chapman et al., 1997), rumen (Van Donkersgoed et al., 1999), hide (Elder et al., 2000), and derived carcasses (Elder et al., 2000). At the other end of the human food chain, numerous studies have reported the presence of E. coli O157:H7, usually at low prevalences, on retail meats. A recent study in France (Vernozy-Rozand et al., 2002) reported 0.12% (4/3450) samples positive for E. coli O157:H7 in large-scale processed minced beef, while studies of butchers shops in a range of countries have reported higher values i.e. 3.8% (6/160) in Argentina (Chinen et al., 2001); 2.3% (5/211) in Switzerland (Fantelli and Stephan, 2001); and 1.1% (36/3216) in the UK (Chapman et al., 2000). In the

ARTICLE IN PRESS
204 C. Cagney et al. / Food Microbiology 21 (2004) 203212

Netherlands, Heuvelink et al. (1999) reported an overall prevalence of 1.1% (6/571) E. coli O157:H7 in minced beef products. Other studies found very different results, ranging from 16.8% (50/296) E. coli O157:H7 samples in Washington State, USA (Samadpour et al., 2002) to a study by Tarr et al. (1999) which did not recover the pathogen from 1400 retail ground minced beef samples from six stores in Seattle, USA. With the exception of the study of Vernozy-Rozand et al. (2002) the above studies merely determined presence or absence of E. coli O157:H7, usually after enrichment, and in general, little data are available on the numbers of this pathogen in retail meat products. In one study, E. coli O157:H7 numbers in beef burgers and in minced beef were estimated using a most probable number (MPN) technique and ranged from o0.34 cfu g1 (0.50.6 log10 cfu g1) (Bolton et al., 1996), although one sample was reported to contain 2300 cfu g1 (3.36 log10 cfu g1). The general pattern observed by Bolton et al. (1996) was conrmed by Tuttle et al. (1999) who used a MPN technique to determine that samples (n=76) of ground beef patties from a slaughter house epidemiologically linked to an outbreak at a fast-food outlet, contained the pathogen at concentrations ranging from o0.3 to 15 cfu g1 (0.521.18 log10 cfu g1). Studies in other foods have established counts of o102 cfu ml1 in raw milk, and o104 cfu g1 in raw-milk cheeses (Coia et al., 2001). Because of the low infectious dose of the pathogen, these reported concentrations of the pathogen are of potential public health signicance. A number of small scale or localised investigations of E. coli O157:H7 numbers in meat in the Republic of Ireland (ROI) have been previously reported. These include a study in the southern region of the ROI (Munster), which detected only 1/586 meat samples positive for E. coli O157:H7 (Walsh et al., 1997). Another small study of retail fresh and processed meats (n =120) from the Dublin area, recovered the pathogen in 3/56 salami samples (at concentrations of approximately 1.5 log10 cfu g1) and in 1/31 cooked ham products (at a concentration of approximately 0.5 log10 cfu g1) (Duffy et al., 1998). The numbers of clinical cases of E. coli O157:H7 in the ROI have remained relatively small in recent years with 76, 51, 42, and 52 cases from 1998 to 2001, respectively (NDSC, 2001). Most of these cases were sporadic, involving children between 1 and 4 years old. A number of suspect foods were reported by cases but as the majority of these were sporadic cases it was not possible to epidemiologically link most of them to an infection source (NDSC, 2001). Considering the severity of the disease caused by this organism, and the population section most frequently infected, it is important to establish the role of beef in the transmission of E. coli O157:H7 in Ireland. Such investigation

necessitates a nation-wide study, to obtain qualitative and quantitative data on the presence of E. coli O157:H7 in Irish retail meat products, which would provide an accurate baseline for accurate risk analysis and assessment, to underpin the development of adequate means of controlling this dangerous pathogen in such products. The aim of this study was to determine the incidence and numbers of E. coli O157:H7 in minced beef (unpackaged or packaged by modied atmosphere packaging (MAP) or air), and beef burgersfrozen and fresh (unpackaged or packaged by MAP or air) sold in Irish butcher and supermarket outlets over a period of 13 months in the 26 counties in the ROI.

2. Materials and methods 2.1. E. coli O157:H7 Two toxigenic strains of E. coli O157:H7 (Ent C9490 and ATCC 43895, ProtectTM Stock Culture Bead) and a non-toxigenic strain of E. coli O157:H7 (NCTC 12900) were maintained on ProtectTM Stock Culture Beads (Protect, Technical Consultants Limited, UK) at 18 C. Beads with each strain were resuscitated, cultured on Sorbitol MacConkey agar (SMAC) (Oxoid) and included as controls in the analysis of each batch of samples examined. 2.2. Sample plan, locations and types Samples (n=1533) were collected (15 samples from each of 26 counties, every 3 months between March 21st 2001 and April 12th 2002). Samples were drawn on a rolling basis within each period so that counties were sampled and resampled in a xed order. During each visit to each county, 15 samples were collected from ve different premises (three supermarkets/two butcher shops). In each county, the ve sampled premises were in two large towns (four premises) and one small town (one premise). Not all sample locations could provide all sample types. The numbers of each type of sample and their sources are presented below. Beef burgers (fresh, packaged, brand name, supermarket) (n=157) Minced beef (fresh, unpackaged, butcher shop) (n=211) Beef burgers (frozen, packaged, brand name, supermarket) (n=206) Minced beef (fresh, packaged, brand name, supermarket) (n=457) Beef burgers (fresh, unpackaged, butcher shop) (n=140) Minced beef (fresh, unpackaged, supermarket) (n=299)

ARTICLE IN PRESS
C. Cagney et al. / Food Microbiology 21 (2004) 203212 205

Beef burgers (fresh, unpackaged, supermarket) (n=63) At each location, three samples were randomly purchased from the range available, immediately placed into cooler boxes at 471 C, transported back to the laboratory, within 32 h, held at 171 C, and examined within 64 h. 2.3. Sample analysis E. coli O157:H7 were (1) enumerated in samples by direct plating and/or (2) detected by enrichment, IMS recovery and plating. 2.3.1. Enumeration Samples of minced beef or beef burgers (10 g) were stomached (230 rpm) for 1 min in 90 ml of EC Broth (Becton Dickinson Microbiology Systems, USA). Aliquots (0.1 or 1.0 ml), of undiluted, 101, or 102 dilutions in maximum recovery diluent (MRD) (Oxoid) were plated in duplicate and in triplicate onto SMAC agar. All plates were incubated at 37 C for 22 h and examined for typical E. coli O157:H7 colonies, i.e. colourless, smooth, circular and entire edge colonies with brown centres. Five typical (suspect) colonies from each duplicate plate were identied, subcultured onto Nutrient agar (Oxoid), and subjected to conrmatory tests as described below. 2.3.2. Enrichment and IMS recovery of E. coli O157:H7 Samples were analysed using the method described in ISO 16654. In brief, samples (25 g) of minced beef or beef burgers were added to 225 ml volumes of modied Tryptone Soya Broth (mTSB, Medical Supply Company, (MSC) Mulhuddart, Dublin) supplemented with 1.0 ml of a 0.45% novobiocin solution (Sigma, Tallaght, Dublin). The samples were stomached (230 rpm) for 2 min and incubated for 24 h at 41.5 C. Immunomagnetic separation (IMS) was conducted after 6 and 24 h using immunomagnetic beads coated with an anti-E. coli O157 antibody (Dynabeadss anti-E. coli O157, Dynal A.S., Oslo, Norway). Each recovered IMS bead complex was spread in duplicate onto Sorbitol MacConkey agar (SMAC) (Oxoid, Basingstoke, UK) supplemented with Cexime tellurite SelectavialTM (MSC) using a sterile cotton swab, and incubated at 37 C for 22 h. Five colonies exhibiting typical presumptive positive characteristics from each duplicate plate were identied, subcultured onto Nutrient agar (Oxoid), and subjected to conrmatory tests as described below. 2.4. Conrmatory tests Typical colonies from SMAC and Nutrient agar plates were directly streaked onto Levines Eosin Methylene Blue (EMB) agar (Oxoid) and Phenol Red

Sorbitol agar with 4-methylumbelliferyl-B-d-Glucuronide (PRS-MUG) (Oxoid) agar, and incubated overnight at 37 C. Colonies displaying a green metallic sheen on EMB and no uorescence on PRS-MUG illuminated with UV light, were identied as colonies exhibiting typical characteristics of E. coli O157 species. These colonies were emulsied in 5 ml of sterile saline solution, inoculated onto an API 20E strips ! (bioMerieux, Basingstoke, UK), and incubated at 37 C for 24 h. Comparison with the results obtained, with the API numerical database, along with a number of additional biochemical tests, i.e. sorbitol, indole and oxidase tests allowed identication of E. coli O157 colonies. 2.5. O157 and H7 antigen determination All sorbitol non-fermenting, indole positive, MUGnegative colonies were examined by latex agglutination (Wellcolex, Merseyside, UK). These beads are coated with antibodies which bind to any O157 or H7 antigens on the test organisms, forming a visible antigenantibody precipitate (De Boer and Heuvelink, 2000). Colonies giving a precipitation reaction were conrmed as E. coli O157:H7 positive. 2.6. Detection of virulence factor genes by polymerase chain reaction (PCR) analysis Conrmed E. coli O157:H7 positives colonies were subject to PCR analysis to detect the presence of the verotoxin genes (vt1, vt2), the attaching and effacing gene (eaeA), enterohaemorrahagic E. coli hlyA gene and the O-antigen encoding region of pO157 gene. DNA was extracted from latex positive colonies using a DNeasy extraction kit (QiagenTM, Crawley, UK). The DNeasy protocol for animal tissues was used, and all DNA extractions were performed according to the manufacturers instructions, with minor adjustments as follows. Each positive colony was inoculated in 10 ml of Brain Heart Infusion (BHI) broth (Oxoid) and incubated for 24 h at 37 C. Bacterial cells were recovered from each culture by centrifugation of 1.5 ml of inoculum at 7500 rpm for 10 min. The supernatant was discarded and procedure repeated. The recovered cells were resuspended in 180 ml of Tissue Lysis Buffer (ATL) and 20 ml of proteinase K was added instead of 40 ml and lysed by incubation at 55 C in a water bath for 3 h. Bacterial DNA was recovered by a series of further washes as described in manufacturers handbook. The primers and PCR conditions used in this study (Assays IIII) are detailed in Table 1. Assays I and II were modied versions of the (Paton and Paton, 1998) multiplex PCR (Fitzmaurice, 2003). Assay III is a modied version of the (Fratamico et al., 2000) multiplex PCR assay. Assay IV, which used all the above

ARTICLE IN PRESS
206 Table 1 Oligonucleotide primers sequences Primer Sequence (50 30 ) Reference Amplication product size (bp) 180 C. Cagney et al. / Food Microbiology 21 (2004) 203212

Assay I vt1F vt1R vt2F vt2R eaeAF eaeAR hlyAF hlyAR Assay II 0157F 0157R Assay III iCh7-F iCh7-F bpbase pairs.

ATAAATCGCCATTCGTTGACTAC AGAACGCCCACTGAGATCATC GGCACTGTCTGAAACTGCTCC TCGCCAGTTATCTGACATTCTG GACCCGGCACAAGCATAAGC CCACCTGCAGCAACAAGAGG GCATCATCAAGCGTACGTTCC AATGAGCCAAGCTGGTTAAGCT

Paton and Paton (1998) modied by Fitzmaurice (2003) Paton and Paton (1998) modied by Fitzmaurice (2003) Paton and Paton (1998)

255

384

Paton and Paton (1998)

534

CGGACATCCATGTGATATGG TTGCCTATGTACAGCTAATCC

Paton and Paton (1998)

259

GCGCTGTCGAGTTCTATCGAGC CAACGGTGACTTTATCGCCATTCC

Fratamico et al. (2000)

625

primers in a single PCR assay, was only used if the other assays (I, II or III) yielded unclear or ambiguous results. In Assay I, 10 ml of genomic DNA was used in each 50 ml PCR reaction. The primer concentrations were as follows: vt1 10 pmol/ml; vt2 10 pmol/ml; eaeA 12.5 pmol/ml and hlyA 25 pmol/ml (MWGBiotech AG, Ebersberg, Germany). For Assay I, 5 ml of genomic DNA was used in each 50 ml PCR reaction. The primer concentration for pO157 was 12.5 pmol/ml. For each assay the PCR mixture consisted of 50 mm KCl (Promega, Madison, Wisconsin, USA), 2.0 mm MgCl2 (Promega), 400 mm (each) of the four deoxynucleoside tri-phosphates (dNTPs) (Promega), 2.5 U Taq DNA polymerase, thermophilic DNA polymerase 10X buffer, magnesium free (Promega). Samples from Assays I and II were subjected to 35 PCR cycles, each consisting of 1 min denaturation at 95 C and 2 min of annealing at 65 C for the rst 10 cycles. Then decreasing to 60 C by cycle 15; and 1.5 min of elongation at 72 C, increasing to 2.5 min from cycles 2535. In Assay III, 5 ml of genomic DNA was used in each 50 ml PCR reaction. The primer concentration of iCh7 (MWGBiotech AG) in a 50 ml reaction was 25 pmol. The primer mixture consisted of 50 mm KCL (Promega), 3.0 mm MgCl2 (Promega), 400 mm (each) of the four deoxynucleoside tri-phosphates (dNTPs) (Promega), 2.5 U Taq DNA polymerase, thermophilic DNA polymerase 10X buffer, magnesium free (Promega). The PCR samples were then subjected to 94 C for 2 min and

then 35 cycles of denaturation at 94 C for 20 s, annealing at 57 C for 1 min, and extension at 72 C for 1 min following the 35 cycles a nal extension for 10 min at 72 C was carried out. In Assay IV, the PCR mixture per 50 ml reaction consisted of 3 ml of genomic DNA, 1 ml of each primer pair at a nal concentration of 0.10.5 mm and 25 ml of Taq PCR Master Mix (QiagenTM). The PCR samples were then subjected to an initial denaturation period at 94 C for 2 min and then an additional cycle of 94 C for 1 min, followed by annealing at 62 C for 1 min and elongation at 72 C for 3 min. Excluding the initial denaturation the programme was repeated for 35 cycles with a nal extension of 72 C for 8 min. All assays included negative controls (in which isolate DNA was replaced with sterile distilled water) and positive controls (in which isolate DNA was replaced with DNA from known positive control organisms). 2.7. Electrophoretic analysis PCR products were detected by gel electrophoresis, on a 2.0% agarose gel containing 2.0 ml of ethidium bromide (Gibco, UK). A volume of 10 ml of each PCR mixture was supplemented with 2 ml tracking dye (Sigma) and loaded onto a well of the gel. A 100-bp DNA ladder molecular weight marker (Pharmica Biotech, UK) was included in each electrophoretic run to allow identication of the amplied products. PCR products were visualized under UV illumination and

ARTICLE IN PRESS
C. Cagney et al. / Food Microbiology 21 (2004) 203212 207

catalogued with a gel documentation system (Stratagene EE 2, Germany). The sizes of the PCR products were compared with the 100-bp marker DNA ladder (Promega) and the PCR products of the required size indicated a presence of the target gene. 2.8. VTEC screening of isolates Those isolates conrmed by PCR as possessing the verotoxin genes (vt1, vt2), the attaching and effacing gene (eaeA), enterohaemorrahagic E. coli hlyA gene and the O-antigen encoding region of pO157 gene were cultured onto Brain Heart Infusion agar slopes. Then treated with Polymyxin B (Bacilius cereus supplement, Oxoid), to release membrane bound toxins, and examined for the presence of VT1 and VT2 toxins using the VTEC-RPLA reverse passive latex agglutination test (Oxoid). 2.9. Analysis of gas atmosphere in packaged minced beef and burger samples A representative number (n=7) of each type of presumptive MAP packaged product was sampled to ascertain the gas composition in the packaging environment. Gas composition of each pack was analysed by gas chromatography (GC) (Gow-Mac Spectra 250, Gow-Mac Instrument Company) using a normalization method (Beggan, 2001). A CI-4100 Integrator (Milton Roy, Ireland) was used to calculate and plot the chromatographs and calculate the proportions of specic gases in the pack headspace from the areas under the curves. 2.10. Statistical analysis The proportion of positive samples that were found in a particular product type was compared to determine if they were signicantly different from each other. This was carried out by calculating the standard error of the differences between the proportions for each pair of products using the formula sdifference
between proportions

product types. The presence or absence of E. coli O157:H7 in a sample can be considered to be a binomial process. Using a stochastic approach, the proportion of positive samples can be represented by a Beta distribution. Montecarlo software, @RiskTM was used to t a Beta distribution to the positive proportion and the 95th percentile was calculated. This represented the maximum likely prevalence of E. coli O157:H7 at the 95% condence level.

3. Results This survey detected E. coli O157:H7 in 43/1533 (2.80%) samples. The effect of the type of minced beef products tested on the prevalence of positive samples was not signicant at the 95% condence level. Given the survey size and the stochastic analysis of the aggregate prevalence data, the mean prevalence of E. coli O157:H7 in minced beef products on retail sale in the ROI is likely to be less than or equal to 3.6% at the 95% condence level. There were period differences observed, with higher levels of positive samples 4.8% (17/356) in period 4, (January to April 2002) than in Period 1 (March to June 2001 and April 2002), period 2 (June to September 2001) or period 3 (October 2001 to January 2002) which, yielded 3.2% (13/404); 2.4% (9/383); and 1.0% (4/390) positive samples respectively. Fig. 1 shows the monthly breakdown of the 43 positive samples, 4.65% (8/172) in January, 4.68% (18/385) in March/April, and 4.35% (5/115) in August while the remaining (1.39%) 12/861 positive samples were detected over the remaining 8 months. Tables 25 detail the products, location, and date on which the E. coli O157:H7 isolates were detected in each period and detail the number of the pathogen present and its virulence prole. The number of E. coli O157:H7 in 21 of the 43 positive samples ranged from 0.524.03 log10 cfu g1 while in the remaining 22 positive samples, the pathogen was detectable by enrichment only i.e. o0.52 log10 units. The range of counts for positive samples was similar in all four periods with counts in period 1 (o0.523.04 log10 cfu g1), period 2 (o0.523.04 log10 cfu g1), period 3 (o0.520.81 log10 cfu g1) and period 4 (o0.524.03 log10 cfu g1), respectively. E. coli O157:H7 was detected at numerous locations throughout the ROI with positives reported in 19 of the 26 counties sampled. There was some location and time grouping of positive samples in each sampling period generally around large towns or cities. In period 1 (April) a large group of ve positive samples was recorded in two towns in county Meath [Navan town (n=4) and Kells (n=1)]. In period 2 (August) a group of three positive samples was recorded in county Cavan

Op1 q1 =n1 p2 q2 =n2 ;

where p1=proportion positive in product 1, q1=1p1 and n1=number of samples taken of product 1 and p2=proportion positive in product 2, q2=1p2 and n2=number of samples taken of product 2. At the 95% condence level proportions were considered signicantly different if they were separated by more than twice the calculated standard error of the difference between proportions. In addition, the likely population prevalence of E. coli O157:H7 in all retail minced beef products sold at retail in Ireland was calculated from the total number of positive samples in the 1533 samples taken of all ve

ARTICLE IN PRESS
208 C. Cagney et al. / Food Microbiology 21 (2004) 203212

9.00%

8.00%

116*

7.00% Percentage of positive samples

6.00%

5.00%

172*

91*

115*

4.00% 269* 3.00% 114* 174* 2.00% 44*

1.00% 128* 0.00% jan feb mar apr may jun Months *Number of products sampled each month.
Fig. 1. Annual distribution of E. coli O157:H7 (n=43) isolated from Irish mince beef.

48* jul aug sep

141* oct

121* nov dec

Table 2 Summary of location, month, premises, product type and E. coli O157:H7 count (log10 cfu g1) and virulence prole for isolates recovered from Irish mince beef in Period 1 County Town Month Premises type Product type Count Verocytotoxin genes vt1 Wicklow Tipperary Meath Meath Meath Meath Meath Louth Dublin Sligo Sligo Mayo Mayo Greystones Thurles Navan Kells Navan Navan Navan Dundalk Ballyfermot Sligo town Tubbercurry Westport Castlebar March March April April April April April April April May May May May Butcher Supermarket Supermarket Butcher Butcher Supermarket Supermarket Supermarket Butcher Supermarket Butcher Butcher Supermarket Mince fresh (u) Mince fresh (map) Mince fresh (u) Mince fresh (u) Mince fresh (u) Mince fresh (map) Mince fresh (map) Burgers frozen (p) Mince fresh (u) Burgers fresh (map) Mince fresh (u) Burger fresh (u) Mince fresh (u) 122 1.18 E E E 1.18 1.00 E 1.89 3.04 2.40 0.90 2.69 + + + + + + + + + + + + vt2 + + + + + + + + + + + + + Verotoxin production VT1 +++ +++ +++ +++ VT2 +++ +++

p=packaged; u=unpackaged; map=modied atmosphere packaging; E=detectable only by enrichment; ++=strong agglutination reaction; +++=very strong agglutination reaction. All isolates contained hylA, eaeA, pO157 and iCh7.

[Cavan town (n=3)] and in period four (March) a cluster of four positive samples was recorded in county Donegal [Letterkenny (n=3) and Bundoran (n=1)]. Of the positive samples 2.70% (32/1183) were purchased from supermarkets and 3.14% (11/350) from butcher shops.

Table 6 summarizes by product type the samples containing E. coli O157:H7. Only one product type (fresh unpacked burgers from supermarkets) had no detected E. coli O157:H7. Of the products containing the pathogen, fresh packaged burgers from supermarkets had the highest prevalence of 4.46% (7/157)

ARTICLE IN PRESS
C. Cagney et al. / Food Microbiology 21 (2004) 203212 209 Table 3 Summary of location, month, premises, product type and E. coli O157:H7 count (log10 cfu g1) and virulence prole for isolates recovered from Irish mince beef in Period 2 County Town Month Premises type Product type Count Verocytotoxin genes vt1 Kilkenny Kilkenny Kildare Dublin Leitrim Leitrim Cavan Cavan Cavan Kilkenny city Kilkenny city Kildare town Ballyfermot Carrick-on-shannon Ballinamore Cavan town Cavan town Cavan town July July July July August August August August August Supermarket Supermarket Butcher Supermarket Supermarket Butcher Supermarket Supermarket Supermarket Mince fresh (map) Mince fresh (map) Mince fresh (u) Burgers fresh (map) Mince fresh (map) Mince fresh (u) Burgers fresh (map) Mince fresh (map) Mince fresh (map) E 3.04 2.30 2.07 0.60 1.65 0.95 E E + + + + + + + + + vt2 + + + + + + + + Verotoxin production VT1 +++ +++ +++ +++ VT2 ++

u=unpackaged; map=modied atmosphere packaging; E=detectable only by enrichment; ++=strong agglutination reaction; +++=very strong agglutination reaction. All isolates contained hylA, eaeA, pO157 and iCh7.

Table 4 Summary of location, month, premises, product type and E. coli O157:H7 count (log10 cfu g1) and virulence prole for isolates recovered from Irish mince beef in Period 3 County Town Month Premises type Product type Count Verocytotoxin genes vt1 Sligo Sligo Limerick Cork Sligo town Sligo town Limerick city Cork city January January January December Supermarket Supermarket Supermarket Supermarket Mince fresh (map) Mince fresh (map) Burger frozen (p) Burger fresh (map) E 0.81 E E + + + + vt2 + + + + Verotoxin production VT1 +++ VT2 +++ ++

p=packaged, map=modied atmosphere packaging; E=detectable only by enrichment; ++=strong agglutination reaction; +++=very strong agglutination reaction. All isolates contained hylA, eaeA, pO157 and iCh7.

Table 5 Summary of location, month, premises, product type and E. coli O157:H7 count (log10 cfu g1) and virulence prole for isolates recovered from Irish mince beef in Period 4 County Town Month Premises type Product type Count Verocytotoxin genes vt1 Wexford Waterford Waterford Kilkenny Kilkenny Clare Meath Roscommon Donegal Donegal Donegal Donegal Tipperary Tipperary Cork Cavan Offaly Wexford town Waterford city Waterford city Thomastown Thomastown Shannon Ashbourne Ballahadreen Letterkenny Letterkenny Letterkenny Bundoran Thurles Thurles Cork city Bailieborough Birr January January January January January February February February March March March March March March March April April Supermarket Supermarket Supermarket Supermarket Supermarket Supermarket Supermarket Supermarket Supermarket Supermarket Supermarket Butcher Supermarket Supermarket Supermarket Butcher Butcher Mince fresh (u) Mince fresh (map) Mince fresh (map) Burger frozen (p) Mince fresh (u) Burger frozen (p) Burger frozen (p) Mince fresh (u) Burgers fresh (map) Mince fresh (map) Burgers fresh (map) Burgers fresh (u) Mince fresh (u) Burgers fresh (map) Burgers frozen (p) Mince fresh (u) Mince fresh (u) 3.43 4.03 2.17 0.51 1.69 E E E E E E E E E E E E + + + + + + + + + + + + + + + + + vt2 + + + + + + + + + + + + + + + + + Verotoxin production VT1 +++ +++ +++ +++ +++ +++ +++ +++ +++ VT2 +++ +++ +++ +++ +++ ++ ++ ++ ++

p=packaged; u=unpackaged; map=modied atmosphere packaging; E=detectable only by enrichment; ++=strong agglutination reaction; +++=very strong agglutination reaction. All isolates contained hylA, eaeA, pO157 and iCh7.

ARTICLE IN PRESS
210 C. Cagney et al. / Food Microbiology 21 (2004) 203212 Table 6 Summary for the number of products sampled and the number of products positive for E. coli O157:H7 in 12-month period of study Product type Premises type Quarter 1 TS Burgers fresh (map) Fresh mince (u) Burgers frozen (p) Fresh mince (map) Burgers fresh (u) Fresh mince (u) Burgers fresh (u) Supermarket Butchers Supermarket Supermarket Butchers Supermarket Supermarket Total 37 54 52 119 36 79 27 404 TP 1 5 1 3 1 2 0 13 Quarter 2 TS 45 53 51 117 28 73 16 383 TP 2 2 0 5 0 0 0 9 Quarter 3 TS 42 52 52 117 38 75 14 390 TP 1 0 1 2 0 0 0 4 Quarter 4 TS 33 52 51 104 38 72 6 356 TP 3 1 4 3 2 4 0 17 Total TS 157 211 206 457 140 299 63 1533 TP 7 8 6 13 3 6 0 43 Percentage % 4.46 3.79 2.91 2.84 2.14 2.01 0 2.80

p=packaged, u=unpackaged, map=modied atmosphere packaging; TS=total number of products sampled, TP=total number of products positive for E. coli O157:H7.

while fresh unpackaged mince purchased from supermarkets had the lowest prevalence of 2.01% (6/299). Of the 43 E. coli O157:H7 positive samples, 20 were modied atmosphere packaged (MAP) in approximately 80% O2 and 20% CO2. This group included the samples with the highest E. coli O157:H7 counts in each period i.e. (Period (P) 1, fresh packaged burgers (3.04 log10 cfu g1); P2, fresh packaged mince (3.04 log10 cfu g1); P3, fresh packaged mince (0.81 log10 cfu g1) and P4, fresh packaged mince (4.03 log10 cfu g1) (Tables 25). Tables 25 presents the genes detected in, and the toxins expressed by the 43 E. coli O157:H7 positive samples. All of these samples contained the eaeA, hylA, pO157 and iCh7 genes. Forty-one of the samples contained both genes encoded for verotoxins (vt1 and vt2) with two samples missing one of the verotoxin genes. Of the 43 isolates, 20 produced VT1 or VT2 toxins, or both.

4. Discussion Of the 1533 minced beef samples collected over a 13-month period, a total of 43 (2.80%) were positive for E. coli O157:H7. This result is higher than noted in previous studies i.e. 0.17% (1/586, Walsh et al., 1997) and 1.4% (4/282, Duffy et al., 1998). However, taking sample size into account, estimates of the population mean prevalence at the 95% condence level was 3.6% in this current study and 3.2% and 0.8% in the studies of Duffy et al. (1998) and Walsh et al. (1997), respectively. Direct comparison of results is difcult due to differences in the study methodologies, such as sampling over a full year and the sampling of all regions in the country. While there is some evidence that E. coli O157:H7 may be increasingly common in beef production systems (McDowell and Sheridan, 2001), the detection of higher proportions of E. coli O157:H7 in more recent studies is more probably associated with the

wider use of more sensitive detection methods such as IMS (Chapman et al., 2001). In comparison to other countries, the prevalence reported in this study is higher than values reported in studies conducted in France (0.12%) (Vernozy-Rozand et al., 2002) and the UK (1.1%) (Chapman et al., 2000), similar to a Swiss study (2.3%) (Fantelli and Stephan, 2001) and lower than reported in a study in Argentina (3.8%) (Chinen et al., 2001). In approximately half (21/43) of the positive samples, E. coli O157:H7 was present in numbers which were below the detection limit of direct plating methods i.e. 0.52 log10 units. In 22/43 samples, direct counts, ranging from 0.524.03 log10 cfu g1 were obtained. Four out of 43 positive samples, had counts >3.00 log10 cfu g1. Such high values suggest point source contamination during primary meat production and processing, and/or subsequent temperature abuse of the meat and meat products within the production/retail chain. The counts obtained in this study are on average higher than the o0.34 cfu g1 (0.50.6 log10 cfu g1) reported in a previous study in the UK by Bolton et al. (1996) although these authors did report one sample to contain, 2300 cfu g1 (3.36 log10 cfu g1). Considering the very low infective dose of this pathogen, its detection at such concentrations in retail beef products poses signicant public health risks. While it is to be hoped that most retail beef products will be adequately cooked before consumption, leading to the destruction of the pathogen, the presence of contaminated meats at retail and consumer levels places consumers at risk. E. coli O157:H7 may persist in undercooked minced beef or beef burgers (Bell et al., 1994) or be transferred from such raw meats to cooked products, or products that do not receive heat treatments prior to domestic consumption, e.g. salad items, by cross contamination of hands, utensils or surfaces (Little and de Louvois, 1998; Field et al., 1977) during domestic food preparation. E. coli O157:H7 was detected in samples from the majority (19/26) of the counties of the ROI. The

ARTICLE IN PRESS
C. Cagney et al. / Food Microbiology 21 (2004) 203212 211

proportions of positive samples showed some grouping in relation to location and sampling occasion as detailed in Tables 25. However, further investigations (not shown) failed to establish any links between positive samples, and individual abattoirs and suppliers in the beef supply chain and this may be due to the transient and sporadic nature of this pathogen. There was a seasonal effect observed in this study. The monthly breakdown of the 43 positive samples showed that 4.65% (8/172) in January, 4.68% (18/385) in March/April, and 4.35% (5/115) in August while the remaining (1.39%) 12/861 positive samples were detected over the remaining 8 months (Fig. 1). A previous study in the ROI on the prevalence of E. coli O157:H7 on beef carcasses also reported the highest prevalence of the pathogen in spring and late summer (McEvoy et al., 2001). In the UK, Chapman et al. (2000) also observed seasonal distribution in the prevalence of E. coli O157:H7 on retail meats with the prevalence highest in summer (July) at 4.0%, and lowest in winter (December) at 0.2%. Of the positive samples 2.70% (32/1183) were purchased from supermarkets and 3.14% (11/350) from butcher shops. This is similar to that reported by Heuvelink et al. (1999). E. coli O157:H7 was found in almost all products types examined including fresh unpackaged mince and burgers, fresh packaged mince and burgers and also frozen burgers as detailed in Table 6. These results conrm reports in the literature that neither freezing (Ansay et al., 1999) or MAP (Uyttendaele et al., 2001) prevent the survival of E. coli O157:H7 in minced beef and burgers. This conrms previous studies (Schluter et al., 1994) which reported that MAP can extend the shelf life of meat products but has little effect on pathogen survival. All 43 positive samples contained the eaeA, hylA, pO157 and iCh7 genes. The genes encoding for verotoxins (vt1 and vt2), that determine the virulence potential of the organism which are essential in the establishment of the disease (Schmidt et al., 2001), were present in 41 of the samples tested. These ndings are supported by a study in an Irish abattoir on the presence of E. coli O157:H7 on beef carcasses (McEvoy et al., 2001) which noted that more than 97% of isolates contained both vt genes, with a very small proportion of isolates carrying only one vt gene. Isolates from a study by Chinen et al. (2001) contained both the vt1 and vt2 genes, but none produced the verotoxins. The prevalence of E. coli O157:H7 in retail minced beef products in the ROI as calculated in this study (2.8%) and the observation of gross levels of contamination (>3.00 log10 cfu g1) in a number of samples pose a signicant risk of transmission of verocytotoxigenic E. coli to Irish consumers. However, while this study has focused on the detection and/or enumeration of E. coli O157:H7 at

near consumer levels, effective action to reduce or eliminate the risks posed by this organism will involve diverse and co-ordinated actions at a number of stages of the food chain. These include the incorporation and consistent application of Good Agricultural practice (GAP), Good Manufacturing practice (GMP), and Hazard Analysis of Critical Control Points (HACCP) at every stage of the beef supply chain, from the farm, through the abattoir, to the retailer, and those involved with the handling and processing of such raw meat products in the home environment (Attenborough and Matthews, 2000). In addition, suitable intervention measures may be necessary to eliminate the pathogen in food reaching the consumer.

Acknowledgements This research was supported by the Food Safety Authority of Ireland (FSAI).

References
Ahmed, S., Donaghy, S., 1998. An outbreak of Escherichia coli O157:H7 in Central Scotland. In: Kaper, J.P., OBrien, A.D. (Eds.), Escherichia coli O157:H7 and Other Shiga Toxin-producing E. coli Strains. ASM Press, Washington DC, pp. 5965. Ansay, S.E., Darling, K.A., Kaspar, C.W., 1999. Survival of Escherichia coli O157:H7 in ground beef patties during storage at 2, 2, 15 and then 2 C, and 20 C. J. Food Prot. 62, 12431247. Attenborough, M., Matthews, K.R., 2000. Food safety through the meat supply chain. J. Appl. Mirobiol. Symp. Suppl. 88, 144s148s. Beggan, M., 2001. Shelf life extension of retail beef stored in a lowoxygen environment. Ph.D. Thesis, Department of Agriculture, National University of Ireland, University College Dublin. Bell, B.P., Goldoft, M., Grifn, P.M., Davis, M., Gordon, D.C., Tarr, P.L., Bartleson, C.A., Lewis, J.H., Barrett, T.J., Well, J.G., 1994. A multistate outbreak of Escherichia coli O157:H7-associated bloody diarrhoea and haemolytic uraemic syndrome from hamburgers. The Washington experience. J. Am. Vet. Med. Assoc. 272, 13491353. Bolton, F.J., Crozier, L., Williamson, J.K., 1996. Isolation of Escherichia coli from raw meat products. Lett. Appl. Microbiol. 23, 317321. Chapman, P., Siddons, C.A., Cerdan Malo, A.T., Harkin, M.A., 1997. A one year study of Escherichia coli O157 in cattle, sheep, pigs and poultry. Epidemiol. Infect. 119, 245250. Chapman, P., Siddons, C.A., Cerdan Malo, A.T., Harkin, M.A., 2000. A one year study of Escherichia coli O157 in raw beef and lamb products. Epidemiol. Infect. 124, 207213. Chapman, P.A., Ellin, M., Ashton, R., 2001. A comparison of immunomagnetic separation and culture, RevealTM and VIPTM for the detection of E. coli O157 in enrichment cultures of naturally contaminated raw beef, lamb and mixed meat products. Lett. Appl. Microbiol. 32, 171175. Chinen, I., Tanoro, J.D., Miliwebsky, E., Lound, L.H., Chillemi, G., Ledri, S., Baschkier, A., Scarpin, M., Manfredi, E., Rivas, M., 2001. Isolation and characterisation of Escherichia coli O157:H7 from retail meats in Argentina. J. Food Prot. 64, 13461351.

ARTICLE IN PRESS
212 C. Cagney et al. / Food Microbiology 21 (2004) 203212 caused by Escherichia coli O157:H7 in Japan. In: Kaper, J.B., OBrien, A.D. (Eds.), Escherichia coli O157:H7 and Other Shiga Toxin-Producing E. coli strains. ASM Press, Washington D.C, pp. 7381. NDSC (National Disease Surveillance Centre), 2001. The epidemiology of verocytotoxigenic E. coli O157:H7 in Ireland 2001. Annual Report. The National Disease Surveillance Centre, 2527 Middle Gardiner Street, Dublin 1, Ireland, pp. 4853. Paton, A.W., Paton, J.C., 1998. Detection and characterisation of shiga toxigenic Escherichia coli by using multiplex PCR assays for stx1, stx2, eaea, enterohaemorrhagic E. coli hlya, rfbO111, and rfbO157. J. Clin. Micro. 36, 598602. Samadpour, M., Kubler, M., Buck, F.C., Depavia, G.A., Mazengia, E., Stewart, J., Yang, P., Al, D., 2002. Prevalence of shiga toxin producing Escherichia coli in ground beef and cattle faeces from King County, Washington. J. Food Prot. 65, 13221325. Schluter, A.R., Miller, M.F., Jones, D.K., Meade, M.K., Ramsey, C.B., Patterson, L.L., 1994. Effects of distribution packaging methods and storage time on the physical properties and retail display characteristics of pork. Meat Sci. 37, 257269. Schmidt, H., Karch, H., Bitzan, M., 2001. Pathogenic aspects of STEC infection in humans. In: Duffy, G., Garvey, P., McDowell, D. (Eds.), Verocytotoxigenic E. coli. Food and Nutrition Press, INC, Trumbell, CT, USA, pp. 241262. Tarr, P.I., Tran, N.T., Wilson, R.A., 1999. Escherichia coli O157:H7 in retail ground beef in Seattle: results of a one-year prospective study. J. Food Prot. 62, 133139. Tuttle, J., Gomez, T., Doyle, M.P., Wells, J.G., Zhao, T., Tauxe, R.V., Grifn, P.M., 1999. Lessons from a large outbreak of Escherichia coli O157:H7 infections: insight into the infectious dose and method of widespread contamination of hamburger patties. Epidemiol. Infect. 122, 185192. Uyttendaele, M., Jozwik, E., Tutenel, A., De Zutter, L., Uradzinski, J., Pierard, D., Debevere, J., 2001. Effect of acid resistance of Escherichia coli O157:H7 on efcacy of buffered lactic acid to decontaminate chilled beef tissue and effect of modied atmosphere packaging on survival of Escherichia coli O157:H7 on red meat. J. Food Prot. 64, 16611666. Van Donkersgoed, J., Graham, T., Gannon, V., 1999. The prevalence of verotoxins, Escherichia coli O157:H7 and Salmonella in the faeces and rumen of cattle at processing. Can. Vet. J. 40, 322338. Vernozy-Rozand, C., Ray-Gueniot, S., Ragot, C., Bavai, C., Mazuy, C., Montet, M.P., Bouvet, J., Richard, Y., 2002. Prevalence of Escherichia coli O157:H7 in industrial mince beef. Lett. Appl. Microbiol. 35, 711. Walsh, L., Dooge, D., Hill, C., 1997. Screening for Escherichia coli O157:H7 in Irish ground beef using two commercial detection systems. Irish Vet. J. incorporating Irish Vet. Times 50, 111115. Willshaw, G.A., Thirwell, J., Jones, A.P., Parry, S., Salmon, R.L., Hickey, M., 1994. Verocytotoxin producing E. coli O157 in beef burgers linked to an outbreak of diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome. Lett. Appl. Microbiol. 19, 304307. Coia, J.E., 1998. Clinical, microbiological and epidemiological aspects of Escherichia coli O157 infection. FEMS Immunol. Med. Microbiol. 20, 19. Coia, J.E., Johnston, Y., Steers, N.J., Hanson, M.F., 2001. A survey of the prevalence of Escherichia coli O157:H7 in raw meats, raw cows milk and raw-milk cheeses in Southeast Scotland. Int. J. Food Microbiol. 66, 6369. De Boer, E., Heuvelink, A.E., 2000. Methods for the detection and isolation of Shiga toxin-producing E. coli. J. Appl. Microbiol. Symp. Suppl. 88, 13351435. Duffy, G., McEvoy, J.M, Sheridan, J.J., Kilbride, B., Garvey, P., 1998. Unpublished Data from The National Food Centre, Teagasc, Ashtown, Dublin 15, Ireland. Elder, R.O., Keen, J.E., Siragusa, G.R., Barkocy-Gallagher, G.A., Koohmaraie, M., Laegreid, W.W., 2000. Correlation of enterohaemorrhagic Escherichia coli O157 prevalence in faeces, hides and carcasses of beef cattle during processing. Proc. Natl. Acad. Sci. 97, 29993003. Fantelli, K., Stephan, R., 2001. Prevalence and characteristics of shigatoxin-producing Escherichia coli and Listeria monocytogenes strains isolated from minced meat in Switzerland. Int. J. Food Microbiol. 70, 6369. Field, R.A., Smith, F.C., Deane, D.D., Thomas, G.M., Kotula, A.W., 1977. Sources of variation of the retail level in bacteriological condition of ground beef. J. Food Prot. 40, 385388. Fitzmaurice, J., 2003. Molecular diagnostic assay for Escherichia coli O157:H7. Ph.D. Thesis Department of Microbiology, National University of Ireland, University College Galway, Ireland. Fratamico, P.M., Bagi, L.K., Tiziana, P., 2000. A multiplex polymerase chain reaction assay for rapid detection and identication of Escherichia coli O157:H7 in foods and bovine faeces. J. Food Prot. 63, 10321037. Heuvelink, A.E., Zwartkruis-Nahuis, J.T.M., Beumer, R.R., De Boer, E., 1999. Occurrence and survival of verocytotoxin-producing Escherichia coli O157 in meats obtained from retail outlets in The Netherlands. J. Food Protect. 62 (10), 11151122. Little, C.L., de Louvois, J., 1998. The microbiological examination of butchery products and butchers premises in the United Kingdom. J. Appl. Microbiol. 85, 177186. McDowell, D.A., Sheridan, J.J., 2001. Survival and growth of VTEC in the environmental. In: Duffy, G., Garvey, P., McDowell, D. (Eds.), Verocytotoxigenic E. coli. Food and Nutrition Press, INC. Trumbell, CT, USA, pp. 279304. McEvoy, J.M., Doherty, A.M., Sheridan, J.J., Thomson-Carter, F.M., Garvey, P., McGuire, L., Blair, I.S., McDowell, D.A., 2001. The incidence and spread of Escherichia coli O157:H7 at a commercial beef abattoir. In: Duffy, G., Garvey, P., Coia, J., Wasteson, Y., McDowell, D. (Eds.), Verocytotoxigenic E. coli O157 in Europe, 5. Epidemiology of Verocytotoxigenic E. coli. Teagasc, The National Food Centre, Dublin, pp. 1227. Michino, H., Araki, K., Minami, S., Nakayama, T., Ejima, Y., Hiroe, K., Tanaka, H., Fujita, N., Usami, S., Yonekawa, M., Sadomoto, K., Takaya, S., Sakai, N., 1998. Recent outbreaks of infections