35 Journal of Food Protection, Vol. 71, No. 1, 2008, Pages 35–45 Copyright 䊚, International Association for Food Protection

Hygiene Indicator Microorganisms for Selected Pathogens on Beef, Pork, and Poultry Meats in Belgium Y. GHAFIR,1,2* B. CHINA,2,3 K. DIERICK,3 L. DE ZUTTER,4

AND

G. DAUBE2

1Belgian

National Reference Laboratory in Food Microbiology for the Federal Agency for the Safety of the Food Chain, and 2Department of Food Sciences, Microbiology, Faculty of Veterinary Medicine, University of Liege, Bat. B43b, Sart Tilman, 4000 Liege, Belgium; 3Institute of Public Health, rue Juliette Wytsman 14, 1050 Brussels, Belgium; and 4Department of Food Microbiology, Faculty of Veterinary Medicine, University of Gent, Salisburylaan 133, 9820 Merelbeke, Belgium MS 07-005: Received 3 January 2007/Accepted 28 May 2007

ABSTRACT Several bacterial indicators are used to evaluate hygiene during the meat slaughtering process. The objectives of this study were to assess the Belgian baseline data on hygienic indicators and the relationship between the indicators and zoonotic agents to establish hygiene indicator criteria for cattle, pig, and chicken carcasses and meat. The study used the results from the official Belgian surveillance plan from 2000 to 2003, which included the monitoring of Escherichia coli counts (ECC), Enterobacteriaceae counts (EC), aerobic colony counts (ACC), and Pseudomonas counts (PC). The sampling method was the wet and dry swabbing technique for cattle and pig carcasses and neck skin excision for broiler and layer chicken carcasses. The 75th and 95th percentiles of ECC were ⫺0.20 and 0.95 log CFU/cm2 for cattle carcasses, 1.20 and 2.32 log CFU/cm2 for pig carcasses, and 4.05 and 5.24 log CFU/g for chicken carcasses. The ACC were 2.1- to 4.5-log higher than the ECC for cattle, pigs, and chickens. For cattle and pig carcasses, a significant correlation between ECC, EC, and ACC was found. ECC for pork and beef samples and EC in pig carcasses were significantly higher in samples contaminated with Salmonella. In poultry samples, ECC were in general higher for samples containing Salmonella or Campylobacter. Thus, E. coli may be considered as a good indicator for enteric zoonotic agents such as Salmonella for beef, pork, and poultry samples and for Campylobacter in poultry samples.

Several bacterial indicators are used to evaluate hygiene levels during meat slaughtering processes. Fecal indicators are chosen because they may be easily detected and used as markers of pathogenic enteric zoonotic agents present in the processing environment or coming from the animals. The ideal fecal indicator should belong to a single species, have the same multiplication rate as the pathogens, and should be present in the feces in high numbers. It should be nonpathogenic and be easily, quickly, and economically detected, and its presence at a fixed level in food should be correlated with the presence of the enteric pathogens. This latter condition is difficult to achieve when the prevalence and distribution of the pathogen on the source animal matrix is low (36). Aerobic plate counts (ACC) frequently are used as indicators to monitor the hygiene of the entire meat production process. Enterobacteriaceae or Escherichia coli are used to assess enteric contamination; Pseudomonas is a psychrotrophic bacterium responsible for meat spoilage and is infrequently used as an indicator (14). In the United States, E. coli is the mandatory indicator for bovine, swine, and poultry carcasses and is used for the verification of the effectiveness of the hazard analysis critical control point (HACCP) plans (1). In the European Union, the new Reg* Author for correspondence. Present address: Federal Agency for the Safety of the Food Chain, bd S. Bolivar, 30, B-1000 Brussels, Belgium. Tel: ⫹32 2 208 36 25; Fax: ⫹32. 2 208 33 37; E-mail: [email protected].

ulation EC 2073/2005 on microbiological criteria for foodstuffs has established the surveillance of ACC and Enterobacteriaceae as hygiene indicator criteria for cattle, sheep, goat, horse, and pig carcasses (10). Between 2000 and 2003, an official Belgian surveillance plan was developed to assess the bacterial contamination (i.e., the presence of Salmonella and Campylobacter and enumeration of E. coli, Enterobacteriaceae, aerobic bacteria, and Pseudomonas) of the whole meat production process, from slaughter to retail. The samples were taken by well-trained personnel and were representative of the national meat production of pork, beef, and poultry (16, 21, 22). The objectives of this study were to assess the Belgian baseline data on hygienic indicators and the relationship between the indicators and zoonotic agents to establish hygiene indicator criteria for cattle, pig, and chicken carcasses and meat in Belgium. MATERIALS AND METHODS Sampling plan. From February 2000 to December 2003, the sampling plan was developed to be representative of the whole of Belgian meat production, and the establishments to be sampled were randomly chosen. Carcasses, trimmings, and minced meat from pig, cattle, and broiler and layer chicken carcasses, broiler prepared meat, and broiler fillets were sampled from slaughterhouses, processing plants, and/or retail establishments. Establishments licensed for export and low-capacity plants (plants slaughtering or producing reduced amounts for the Belgian domestic market) were chosen in accordance with their production capacity,

36

GHAFIR ET AL.

as described by Ghafir and coworkers (22). The sampling plan included approximately 110 cattle, 110 pig, and 40 chicken slaughterhouses. Of these slaughterhouses, 40 for pigs, 40 for cattle, and 25 for chicken were licensed. There were more than 1,000 production plants for beef and pork cuts, products, and ground meat and 125 cutting plants for poultry. Each year, all the licensed slaughterhouses and the largest processing plants were sampled, with a minimum of five samples per establishment. About 300 samples per type of product were chosen each year because this sample size was economically acceptable for the national public health and food authorities. Sample collection. The sampling was performed according to the procedure of Ghafir and coworkers (22). For pig and cattle carcasses, the wet and dry swabbing method was used because it allows the sampling of large areas (minimum of 100 cm2 for each zone), as determined by ISO standard 17604 (8) and as used by several authors (24, 29, 31, 33–35). Sampling was carried out in the chilling room, 2 to 4 h after slaughter, which was the most practical approach. The areas swabbed in this study were chosen because they were more susceptible to contamination due to the slaughtering method used, in accordance with ISO standard 17604 (8). Sterile cotton cosmetic pads of the type sold in supermarkets that were about 7 cm in diameter and without inhibitory substances were used to swab half-carcasses from cattle and pigs. Four zones on each half-carcass representing 1,600 and 600 cm2, respectively, were swabbed by well-trained veterinarians without using a surface template, and these swabbing bout constituted one sample. On cattle half-carcasses, the four 400-cm2 zones were the backlateral area of the thigh, the flank next to the midline, the brisket next to the midline at the level of the elbow, and the rear side of the forelimb. For pig half-carcasses, the four 100-cm2 zones were the internal muscular part of the ham, the back internal part of the pan, the back part of the forelimb, and a 300-cm2 area on the sternum. Entire broiler chicken carcasses and at least 200 g of minced meat, cut meat (small chops obtained at the end of the meat production process), fillet (boned breasts), or prepared meat (raw ground meat processed as sausages or patties) were sampled at the end of the production process (after chilling) or at the retail level and were placed in a sterile plastic bag. Layer chicken carcasses included those of reproductive hens and of hens slaughtered after egg production has ceased. Each sample was immediately placed in an insulated refrigerated box and transferred the same day to the laboratory. Microbiological analyses and expression of results. Three laboratories licensed by the Belgian Ministry of Public Health and accredited in accordance with the requirements of ISO standard 17025 (11) performed all the analyses. The three laboratories were located in the three Belgian regions and were responsible for analyzing all the types of meat and bacteria. The samples were distributed among the three laboratories according to their geographical location. In the laboratory, samples were stored chilled and examined within 24 h. The determination of E. coli counts (ECC), ACC, Enterobacteriaceae counts (EC), and Pseudomonas counts (PC) were made using the following validated standardized methods: the chromogenic rapid E. coli 2 method with an incubation period of 24 h at 44⬚C (Association Franc¸aise de Normalisation [AFNOR] validated SDP-07/1-07/93, Bio-Rad, Marnes La Coquette, France), the NF-V08-051 AFNOR standard method using plate count agar (CM0463 or CM0325, Oxoid, Basingstoke, UK; or ref. 3554459, Bio-Rad, Steenvoorde, France) at 30⬚C and incubation for 48 to 72 h (3), the NF-V08-054 method using violet red bile glucose agar (CM0485, Oxoid; or ref. 3554239, Bio-Rad, Steenvoorde) at 30⬚C for 24 h (4), and the NF-V04-504 method

J. Food Prot., Vol. 71, No. 1

using cetrimide-fucidine-cephalosporin agar medium (CFC; CM0559 with Pseudomonas CFC supplement SR0103E, Oxoid) at 25⬚C for 48 h with oxydase confirmation (2), respectively. Buffered peptone water (BPW; CM0225, Oxoid; or ref. 8532185, International Medical, Marche-en-Famenne, Belgium) or 0.1% peptone solution (CM0733, Oxoid; or bacteriological peptone LP0037, Oxoid with NaCl ref. 27810295, VWR International, Fontenay-sous-Bois, France) was used for initial suspension and for the 1:10 dilutions of the samples. Methods for the detection of Salmonella and Campylobacter have been previously described (21, 22). The methods used were the Belgian official methods using BPW and the semisolid Diasalm medium (Lab 537, Lab M, International Diagnostics Group PLC, Lancashire, UK) for Salmonella (SP-VG-M002) and the Preston medium (nutrient broth no. 2 CM0067, Campylobacter selective supplement SR0117E and lysed horse blood SR0048, Oxoid) and the modified charcoal deoxycholate agar (Campylobacter blood-free selective medium CM0739 and charcoal deoxycholate agar selective supplement SR0155, Oxoid) for Campylobacter (SP-VG-M003). Detection of Salmonella and counts of indicators were made on the same sample; however, detection of Campylobacter was made using a different sample from the same carcass or piece of meat. The results were recorded as CFU per square centimeter or per gram of the sample. Statistical analysis. All statistical analyses were carried out on the log-transformed values. When no colony was detected, a value of half of the limit of detection was used for the calculations (24). A permutation test was used for the determination of significant differences between the 50th, 75th, and 95th percentiles (17). The Mann-Whitney test was used to compare the medians for samples positive and negative for Salmonella and Campylobacter. The Spearman rank correlation test was used for determination of a correlation between different indicators (e.g., E. coli and Enterobacteriaceae) analyzed on the same samples during the 4year survey. Chi-square tests were used to determine the significance of differences in the proportion of samples above the detection limit for the samples obtained over several years. Fisher’s exact test was used for comparison of prevalence tables with two rows and two columns. The confidence intervals were calculated exactly using the binomial distribution.

RESULTS ECC in beef, pork, and poultry samples. The study carried out in 2000, 2001, 2002, and 2003 for E. coli revealed that beef carcasses, cuts, and minced meat had a median lower than or corresponding to the limit of detection. The 95th percentile of the samples had a maximum of 1.01, 2.93, and 2.79 log CFU/cm2 or log CFU/g, respectively. E. coli was found in only 32 to 62% of cattle samples. For pork carcasses and cuts, E. coli contamination was 1 log higher than that of beef samples. The cumulative results from 2000 (2002 for beef cuts) to 2003 revealed that beef and pork cuts were more contaminated with E. coli (with a 95th percentile of 2.77 and 3.64 log CFU/g, respectively) than was minced meat (with a 95th percentile of 2.64 and 2.91 log CFU/g, respectively). ECC were higher than the detection limit in 42 to 90% of pork samples. These results for ECC are shown in Table 1. A significant change in E. coli percentile results was observed for beef samples. A higher contamination rate was

J. Food Prot., Vol. 71, No. 1

37

HYGIENE INDICATOR CRITERIA FOR MEAT IN BELGIUM

TABLE 1. ECC results for 2000, 2001, 2002, and 2003 surveys of beef and pork carcasses, cuts, and minced meat Survey years: Sample type

Beef carcasses

Beef cuts

Beef minced meat

Pork carcasses

Pork cuts

Pork minced meat

E. coli result

2000

2001

2002

2003

2000–2003

Median (log CFU/cm2) Mean 75th percentile 95th percentile SD Prevalence (%)d n Median (log CFU/g) Mean 75th percentile 95th percentile SD Prevalence (%) n Median (log CFU/g) Mean 75th percentile 95th percentile SD Prevalence (%) n Median (log CFU/cm2) Mean 75th percentile 95th percentile SD Prevalence (%) n Median (log CFU/g) Mean 75th percentile 95th percentile SD Prevalence (%) n Median (log CFU/g) Mean 75th percentile 95th percentile SD Prevalence (%) n

⫺0.85b ⫺0.53 ⫺0.19 1.01 0.78 61 1,501

⫺1.10c ⫺0.62 ⫺0.25 0.88 0.75 47 1,653

⫺0.85b ⫺0.52 ⫺0.08 1.00 0.80 56 1,326 0.70c 1.13 1.30 2.74 0.70 41 223 0.70c 1.15 1.30 2.79 0.72 46 299 0.63 0.64 1.26 2.17 0.94 89 299 0.70c 1.26 1.63 3.03 0.90 47 224 1.00b 1.32 1.70 2.96 0.81 58 301

⫺1.15c ⫺0.78 ⫺0.38 0.79 0.81 55 1,485 0.70c 1.08 1.00b 2.93 0.74 34 293 0.70c 1.02 1.00b 2.40 0.63 35 298 0.46 0.51 1.05 2.40 1.07 90 288 0.70c 1.18 1.48 2.89 0.80 42 278 1.00b 1.25 1.60 2.73 0.76 55 299

⫺0.85 ⫺0.62 ⫺0.20 0.94 0.79 54

1.00b 1.21 1.48 2.64 0.69 62 489 0.48 0.55 1.21 2.31 1.00 88 431 1.48 1.86 2.47 4.12 1.16 78 291 1.00b 1.45 1.85 2.79 0.73 80 308

0.70c 1.09 1.00b 2.70 0.74 32 299 0.47 0.53 1.19 2.39 1.03 85 292 1.30 1.70 2.42 3.81 1.24 61 247 1.00b 1.41 1.97 3.45 0.94 56 300

Permutation testa

SS SS SS SS

0.70 1.10 1.30 2.77 0.72 37

NA

0.70 1.13 1.30 2.64 0.70 46

SS

0.49 0.55 1.20 2.32 1.01 88

NS

1.00 1.51 2.08 3.64 1.08 58

NS

1.00 1.36 1.78 2.91 0.81 63

NA

NS NS NS

S NS SS

NS NS NS

SS S SS

S S SS

a

Result of the permutation test to assess the significance of the difference between the median and 75th and 95th percentiles of results for 2000, 2001, 2002, and 2003. SS, P ⬍ 0.01; S, P ⬍ 0.05; NS, P ⬎ 0.05; NA, not applicable. b Result at the detection limit. c Result below the detection limit (all the results in this study have been taken into account, including the results below the detection limit by giving them the value of the detection limit divided by 2). d Percentage of the results above the detection limit, which differs according to the sample, the laboratory, and the year: ⫺1.22, ⫺1.15, ⫺0.85, or ⫺0.80 log CFU/m2 for beef carcasses; ⫺0.77, ⫺0.74, ⫺0.43, or ⫺0.37 log CFU/cm2 for pork carcasses; 1.00 log CFU/g for all other meat types.

detected in 2000 and 2002 than in 2001 and 2003, and this difference was significant for carcasses and minced meat (P ⬍ 0.05) but was less than 0.5 log CFU/cm2 or CFU/g. For pork, the lower E. coli percentile results observed in 2003 in comparison with 2000 and/or 2001 were significant for cuts and minced meat (P ⬍ 0.05).

For chicken, 81 to nearly 100% of the analyzed sample types contained detectable E. coli. The rate of this indicator was also higher in poultry than that in beef and pork samples. An increase in ECC occurred between 2000 and 2001, followed by a decrease on broiler carcasses, fillets, and prepared meat (P ⬍ 0.05 for the highest difference between

38

GHAFIR ET AL.

J. Food Prot., Vol. 71, No. 1

TABLE 2. ECC results for 2000, 2001, 2002, and 2003 surveys of broiler chicken carcasses, fillets, and prepared meat and layer chicken carcassesa Survey years: Sample type

Broiler carcasses

Broiler fillets

Prepared broiler meat

Layer carcasses

E. coli result (log CFU/g)

Median Mean 75th percentile 95th percentile SD Prevalence (%)c n Median Mean 75th percentile 95th percentile SD Prevalence (%) n Median Mean 75th percentile 95th percentile SD Prevalence (%) n Median Mean 75th percentile 95th percentile SD Prevalence (%) n

2000

2001

2002

2003

2000–2003

3.41 3.43 3.87 5.07 0.82 99 284 2.30 2.28 2.67 3.61 0.77 96 276

3.60 3.70 4.30 5.75 1.18 98 281 2.30 2.34 2.76 3.76 0.80 94 233

3.35 3.39 3.97 5.05 1.03 96 258 2.11 2.09 2.60 3.76 0.96 85 232 2.61 2.52 3.20 3.96 0.96 94 81 3.41 3.34 3.90 4.98 1.05 93 118

3.39 3.47 4.08 5.23 1.08 97 291 1.90 1.96 2.60 3.62 0.97 81 248 2.54 2.31 2.88 3.57 0.91 86 99 3.37 3.42 3.87 5.51 1.15 95 101

3.45 3.50 4.05 5.24 1.04 97

3.34 3.31 3.77 4.50 0.74 99 187

3.41 3.49 4.11 5.47 1.03 99 191

2.30 2.17 2.65 3.72 0.89 89 2.56 2.41 3.11 3.69 0.93 89 3.38 3.39 3.90 5.18 0.97 97

Permutation testb

S SS S NS SS NS NS SS NS S S NS NS NS NS SS

a

All the results in this study have been taken into account, including the results below the detection limit by giving them the value of the detection level divided by 2. b Result of the permutation test to assess the significance of the difference between the median and 75th and 95th percentiles of results for 2000, 2001, 2002, and 2003. SS, P ⬍ 0.01; S, P ⬍ 0.05; NS, P ⬎ 0.05. c Percentage of results above the detection limit (1.00 log CFU/g).

the percentile results for most of the chicken samples), but the difference was more than 0.5 log CFU/g only for broiler carcasses. The 95th percentile was 5.24 log CFU/g and 5.18 log CFU/g for broiler and layer carcasses, respectively, and 3.72 log CFU/g for broiler fillets and prepared meat. These results are shown in Table 2. EC, ACC, and PC on beef, pork, and poultry samples. In 2001, EC were determined on the beef and pork carcasses. In comparison with ECC, the EC percentiles were 0.2- to 1.2-log higher (Table 3). ACC were between 2.1 and 4.5 log CFU/cm2 or log CFU/g higher than those for ECC for cattle, pork, and poultry percentiles (beef and pork carcasses in 2001 and 2002; broiler carcasses, layer carcasses, and broiler prepared meat in 2003). The results for beef and pork carcasses decreased significantly between 2001 and 2002 (P ⬍ 0.05). On broiler and layer carcasses, Pseudomonas was enumerated only in 2002, and the contamination rate of 1.1 to 3.4 log CFU/g was higher than that of E. coli. The results for EC, ACC, and PC are shown in Tables 3 and 4. For beef and pork carcasses, a significant correlation

(P ⬍ 0.0001) was found between ECC, EC, and ACC. The Spearman correlation coefficient (rs) was the highest between ECC and EC (rs ⫽ 0.56 for beef and 0.75 for pork) followed by EC and ACC (rs ⫽ 0.46 and 0.51, respectively) and ECC and ACC (rs ⫽ 0.31 and 0.43, respectively). On broiler and layer carcasses, the Spearman correlation coefficient was not different from zero (P ⬎ 0.05) for the comparison between ECC and ACC (rs ⫽ 0.11 for broiler carcasses and ⫺0.09 for layer carcasses) or between ECC and PC on layer carcasses (rs ⫽ 0.17). A significant negative correlation was observed for broiler carcasses between ECC and PC (rs ⫽ ⫺0.15, P ⫽ 0.02). Relationship between ECC, EC, ACC, and PC and the prevalence of Salmonella and Campylobacter. Evaluation for the presence of Salmonella and Campylobacter was carried out on the same carcass or piece of meat that was used for the bacterial counts. The prevalence of Salmonella was 1.1 to 3.5% in beef, 11.1 to 18.9% in pork, and 9.5 to 25.6% in poultry samples. The prevalence of Campylobacter was 0.6% in beef minced meat, 2.5% in pork minced meat, and 18.7 to 46.9% in poultry samples

J. Food Prot., Vol. 71, No. 1

TABLE 3. EC and ACC results for the 2001 and 2002 surveys of beef and pork carcassesa

TABLE 4. ACC and PC results for surveys of broiler chicken carcasses, broiler prepared meat, and layer carcassesa

ACC (log CFU/cm2) in:

Sample type

Beef carcasses

Pork carcasses

Result

Median Mean 75th percentile 95th percentile SD Prevalence (%)c n Median Mean 75th percentile 95th percentile SD Prevalence (%) n

39

HYGIENE INDICATOR CRITERIA FOR MEAT IN BELGIUM

EC (log CFU/cm2) in 2001

2001

2002

Permutation testb

3.16 3.19 3.79

2.94 2.88 3.46

SS

2.10

5.04

4.40

S

1.05 84

1.16 98

0.93 99

2.61 1.00 98 288

Result

Broiler carcasses

⫺0.10 0.05 0.69

280 1.06 1.04 1.64

Sample type

280 184 3.64 3.02 3.62 3.16 4.14 3.68 4.97

5.28

SS

SS SS NS

0.91 1.12 100 100 288

299

a

EC, Enterobacteriaceae counts; ACC, aerobic colony counts. All the results in this study have been taken into account, including the results below the detection limit (detection limit differs according to the sample, the laboratory, and the year: ⫺0.85 or ⫺0.80 log CFU/cm2 for beef carcasses; ⫺0.43 or ⫺0.37 log CFU/cm2 for pork carcasses) by giving them the value of the detection level divided by 2. b Result of the permutation test to assess the significance of the difference between the median and the 75th and 95th percentiles of results of 2001 and 2002. SS, P ⬍ 0.01; S, P ⬍ 0.05; NS, P ⬎ 0.05. c Percentage of results above the detection limit, which differs according to the sample, the laboratory, and the year: ⫺1.22, ⫺1.15, ⫺0.85, or ⫺0.80 log CFU/cm2 for beef carcasses; ⫺0.77, ⫺0.74, ⫺0.43, or ⫺0.37 log CFU/cm2 for pork carcasses.

(21, 22). Comparisons of the median for ECC from the samples without Salmonella and samples with Salmonella revealed a higher number of E. coli when Salmonella was detected. This difference was significant (P ⬍ 0.01) for beef, pork, and broiler samples except broiler fillets. The relationship was similar for ECC and Campylobacter from broiler fillets and prepared meat and for layer carcasses (P ⬍ 0.01). For ACC, the median was higher for samples containing Salmonella or Campylobacter, but this difference was not significant (P ⬎ 0.05). On pork carcasses, EC were higher in samples positive for Salmonella (P ⬍ 0.01). For PC on broiler carcasses, the relationship was the opposite: the median was lower for samples with Campylobacter (P ⬍ 0.01). Among the significant results, the differences between the median values for samples positive and negative for Salmonella or Campylobacter were lower than 0.5 log CFU/cm2 or log CFU/g except for ECC and Salmonella in beef minced meat, pork cutting meat, and pork minced meat

Median Mean 75th percentile 95th percentile SD Prevalence (%)b n Broiler prepared meat Median Mean 75th percentile 95th percentile SD Prevalence (%) n Layer carcasses Median Mean 75th percentile 95th percentile SD Prevalence (%) n

ACC (log CFU/g) in 2003

PC (log CFU/g) in 2002

5.50 5.81 6.47 7.99 1.10 100 286 6.63 6.61 7.11 8.09 0.88 100 34 6.11 6.42 7.39 8.43 1.21 100 101

4.40 4.73 5.88 8.18 1.83 97 256

4.68 5.18 6.62 8.42 1.85 99 113

a

ACC, aerobic colony counts; PC, Pseudomonas counts. All detection limits were 1.00 log CFU/g. b Percentage of results above the detection limit.

and for PC and Campylobacter on broiler carcasses. The medians and means of ECC, EC, ACC, and PC in relation to the Salmonella and Campylobacter results are shown in Tables 5 and 6. The prevalence of Salmonella in pork and beef samples with an ECC ⱖ75th percentile was significantly higher (P ⬍ 0.01) than that in the samples with an ECC ⬍75th percentile. The differences were 4.6% for beef minced meat, 7.1% for pork minced meat, 9.0% for pork carcasses, and 27.7% for pork cutting meat (Fig. 1). For poultry samples, the prevalence of Salmonella was higher (but not significantly so, P ⬎ 0.05) for broiler carcasses and prepared meat samples with an ECC ⱖ75th percentile (Fig. 1). The difference was not significant for EC (7.9%) on pork carcasses (data not shown). The prevalence of Campylobacter in broiler prepared meat and layer carcasses was 20.6% (P ⬍ 0.05) and 12.8% (P ⬍ 0.01) higher in samples with an ECC ⱖ75th percentile, respectively (Fig. 2). DISCUSSION Several aspects were examined during interpretation of the results of the surveillance of hygienic indicators: (i) the sampling method for beef, pork, and poultry carcasses, (ii) the changes in ACC and ECC between 2000 and 2003, (iii) a comparison with other studies, and (iv) the indicators to be chosen for different types of samples. Sampling method for carcasses. The wet and dry swabbing sampling method was used in this study for pork and beef carcasses because this method, which is frequently

40

GHAFIR ET AL.

J. Food Prot., Vol. 71, No. 1

TABLE 5. ECC, ACC, PC, and EC median results for beef, pork, and poultry samples negative and positive for Salmonellaa Samples negative for Salmonella Sample type

Beef minced meat Pork carcasses

Pork cuts Pork minced meat Broiler carcasses

Broiler fillets Broiler prepared meat Layer carcasses

a

Indicator organism

n

ECC ECC ACC EC ECC ECC ECC ACC PC ECC ECC ECC ACC PC

1,334 968 481 229 858 1,075 1,004 250 238 858 134 461 101 89

Median (log CFU/g or cm2)

0.70 0.43 3.28 0.89 1.00 1.00 3.82 5.46 4.52 2.30 2.32 3.35 6.11 4.60

A C E G I K

M

Samples positive for Salmonella n

49 226 105 59 164 133 103 35 18 128 46 135 19 24

Median (log CFU/g or cm2) Survey year(s)

1.30 0.85 3.48 1.32 2.34 1.48 4.19 5.67 3.70 2.30 2.78 3.44 6.02 5.00

B D F H J L

N

2000–2003 2000–2003 2001–2002 2001 2000–2003 2000–2003 2000–2003 2003 2002 2000–2003 2002–2003 2000–2003 2003 2002

ECC, E. coli counts; ACC, aerobic colony counts; PC, Pseudomonas counts; EC, Enterobacteriaceae counts. Within the same row, medians that do not share a common letter are significantly different (P ⬍ 0.01).

used in Europe, is more practical and does not cause damage to the carcass as does the excision technique (14, 22, 23, 27, 31, 33). This sampling technique was in accordance with Belgian regulations and is usually used in Belgium by government authorities and operators (6). For poultry carcasses, the neck skin was sampled because this is the official sample technique used in Belgium (22, 27) and the method to be used according to the European Regulation EC 2073/2005 for Salmonella analysis (10). In comparison with the whole carcass rinsing method, the excision of neck skin is less reproducible but is more often used because the sampling and microbiological methods are easier (25). Changes in ACC and ECC between 2000 and 2003. Between 2000 and 2003, the survey revealed that significant changes (P ⬍ 0.05) in the median and the 75th or 95th percentiles occurred for all types of samples for ECC and ACC except for those of beef cuts, pork carcasses, and layer carcasses. However, the difference was significant and of a minimum of 0.5 log CFU/cm2 or log CFU/g for ECC in pork cutting meat, pork minced meat, and broiler carcasses and for ACC in beef and pork carcasses. Only these results

are considered significant in terms of biological process control. For beef carcasses, the 95th percentiles of ACC were higher in 2001 than in 2002 (P ⬍ 0.05). The carcasses were sampled in licensed and low-capacity slaughterhouses; 4.4% (in 2002) to 8.5% (in 2000) of the samples were collected in low-capacity slaughterhouses. The results are in agreement with those of a previous study in which a lower ACC was found in export than in domestic establishments and very small plants (34). However, the significant changes between the different years also were observed when only the licensed slaughterhouses were taken into account. Among these abattoirs, 73.9 to 100% were sampled each year, with a significant linear relationship (P ⬍ 0.001), i.e., there was no significant difference between the sampled establishments during the 4-year surveillance period. The changes between the years 2000 and 2002, and 2001 and 2003 seemed therefore not to be due to the repartition of the different types of sampled establishments and could be interpreted as a global decrease in ACC due to improvements in manufacturing hygiene standards following De-

TABLE 6. ECC, ACC, and PC median results for poultry samples negative and positive for Campylobactera Samples negative for Campylobacter Sample type

Broiler carcasses

Broiler fillets Broiler prepared meat Layer carcasses

a

Indicator organism

n

ECC ACC PC ECC ECC ECC ACC PC

756 202 166 793 94 470 101 88

Median (log CFU/g)

3.84 5.48 4.72 2.28 2.30 3.33 6.11 4.75

A C E G

Samples positive for Campylobacter n

346 79 89 183 83 117 13 24

Median (log CFU/g)

3.86 5.53 3.87 2.41 2.74 3.56 5.96 4.37

B D F H

Survey year(s)

2000–2003 2003 2002 2000–2003 2002–2003 2000–2003 2003 2002

ECC, E. coli counts; ACC, aerobic colony counts; PC, Pseudomonas counts. Within the same row, medians that do not share a common letter are significantly different (P ⬍ 0.01).

J. Food Prot., Vol. 71, No. 1

HYGIENE INDICATOR CRITERIA FOR MEAT IN BELGIUM

41

FIGURE 1. Prevalence of Salmonella in beef, pork, and poultry meat or carcasses for samples with ECC ⬍75th and ⱖ75th percentile. * Difference significant at P ⬍ 0.05; ** difference significant at P ⬍ 0.01.

cision 2001/471/EC (replaced by European Regulation EC 2073/2005 since 2006) and its Belgian implementation (5, 6) since the end of 2001. For minced beef, the sampled establishments were licensed (17.4%), low capacity (11.8%), butcheries (32.3%), and supermarkets (38.6%). The retail sampling places changed each year on a random basis. More than 70% of the samples were obtained from retail establishments. ECC in minced beef were higher in 2000 and 2002 than in 2001 and 2003, and the situation was the same for beef carcasses. The ECC in minced meat seems to be influenced by carcass contamination. For pork cuts and minced meat, as with the decrease in ECC from 2000 and/or 2001 to 2003, a decrease was observed in the prevalence of Salmonella, from 32.3 and 16.6% in pork cuts and minced meat, respectively, in 2000 to 6.1 and 6.4%, respectively, in 2003 (22). This improvement in the microbiological status of pork could be due to better application of hygienic practices and the HACCP plans. The constant ECC on pork carcasses between 2000 and 2003 indicates that improvements may still be needed to reduce fecal contamination. The decrease in ECC in minced meat is also in accordance with the decrease ob-

served for pork cuts, the raw material used for minced meat. For broiler meat, a decrease in ECC from 2000 or 2001 to 2003 was observed for all types of samples and could therefore be due to better hygienic practices. These changes were less marked in broiler samples (about a 0.3-log decrease) than in pork samples (up to a 1-log decrease). Except for broiler prepared meat (100% sampled at retail level), broiler meat was sampled in licensed slaughterhouses or cutting plants (57.2% of the samples), supermarkets (22.3% of the samples), butcheries (13.6% of the samples), and low-capacity establishments (7.0% of the samples), with a large variation in retail places from year to year chosen on a random basis. Hygienic indicators and comparison with other studies. For ECC, the proportion of results higher than the detection limit of the method was in general between 30 and 60% for beef and pork. It is not sufficient to use the mean of positive results for comparisons (this requires a minimum 80% of positive results, according to Gill and Jones (23)). It was therefore decided to take into account all the results in this study by using the value of the detec-

FIGURE 2. Prevalence of Campylobacter in broiler and layer carcasses and meat for samples with ECC ⬍75th and ⱖ75th percentile. * Difference significant at P ⬍ 0.05; ** difference significant at P ⬍ 0.01.

42

GHAFIR ET AL.

J. Food Prot., Vol. 71, No. 1

FIGURE 3. Satisfactory (S), acceptable (A), and nonacceptable (NA) zones delimited by the m and M limits on beef and pork carcasses (log CFU per square centimeter) according to the present study and European Regulation EC 2073/2005 for aerobic colony counts (ACC) and Enterobacteriaceae counts (EC). The differences in sampling method and in comparable recovery of microorganisms using the wet and dry swabbing method make interpretation of the results of the present study difficult in comparison with the European criteria.

tion limit divided by 2 for results below the detection limit (24). The European regulations regarding microbiological criteria (10) outline process hygiene criteria for ACC and EC on beef and pork carcasses sampled according to the destructive excision method. In comparison with this technique, bacterial recovery by swabbing ranges from 0.9 to 100%, according to various studies (13, 19, 23, 24, 31, 34, 42, 43). From 9 to 100% of the bacteria can be recovered using the wet and dry swabbing technique, even if no linear relationship is observed between carcass surface bacterial counts obtained by swabbing and those obtained by excision (14, 23, 24, 31, 33). Samples of cattle and pig carcasses were obtained 2 to 4 h after slaughtering, during the chilling process, because this was the most practical approach. The European regulations apply to carcasses before the chilling process, and the U.S. regulations apply to carcasses after the chilling process (1, 10). Several studies have compared the recovery of bacteria between dressed and chilled cattle and pig carcasses. For the excision sampling method, Yu and coworkers (43) found a recovery of 70% for ACC and 4% for total coliform counts (TCC), and Sofos and coworkers (38) observed a similar recovery for ACC, ECC, and TCC on chilled carcasses in comparison with dressed carcasses. With the sponge swab method, bacterial recovery was about 70% of ACC and TCC in one study and 43 to 77% (for ACC and TCC) or 53 to 93% (for ECC) in another study (42, 43). The irreversible attachment of bacteria and biofilms formed in superficial crevices of the skin decreases the number of bacteria recovered on the postchill carcasses 30 min to a few hours after dressing (14, 43). Most of the studies on cattle and pig carcass samples have been carried out after dressing (13, 18, 19, 24, 31) or after chilling (34, 41) or have been performed twice (23, 38). Comparisons between the two methods indicate that the results of the present study are more comparable to those from studies in which carcass samples were analyzing before rather than after chilling.

In 2002, the 75th and 95th percentiles of ECC and ACC obtained by the national surveillance plan were taken as a basis to determine the m (level of satisfaction) and M (limit of acceptability) for beef and pork carcasses for the Belgian regulations (6). To meet these mandatory criteria, corrective measures must be taken in the slaughterhouses that obtain ECC or ACC higher than the 75th percentile. Since 2001, the European Commission has recommended the destructive sampling method but allows use of the swabbing technique (5). The 75th and 95th percentiles (1, 40) of the results obtained in the present study (using swabs) were compared with the m and M limits of the European regulation criteria for the destructive sampling method (10) (Fig. 3). For ACC, the m and M limits showed a similar pattern for cattle and pig carcasses, with a maximal difference of 0.3 log CFU. The EC limits of this study were 0.4-log lower than the regulation limits, except for the m limit of beef carcasses, which was 0.8-log lower. These differences between the criteria of the present study (the 75th and 95th percentiles using the swabbing technique) and the European criteria (the m and M limits for the destructive method) are a maximum of 12% (Fig. 3). This difference could be due to the zones swabbed, the sampling method, or the contamination level in Belgium. In this study, the swabbed zones included areas on the internal side of the half-carcass, which is more susceptible to contamination due to the slaughtering technique. Following Decision 2001/471/EC (5), the areas to be sampled were on the external side of the cattle and pig half-carcasses. Another difference was the larger area swabbed on cattle (1,600 cm2) and pig (600 cm2) carcasses in comparison with the destructive method in the EC 2073/2005 regulation (20 cm2 before the chilling process) and the minimal area required for nondestructive methods (400 cm2, i.e., four swabs of 100 cm2) (8, 10). The materials used for obtaining the samples (swabs, sponges, and cloths) also are an important source of variation in the recovery of the microorganisms because of bacterial stress and attachment during the chilling process (14, 29, 32, 43). The differences in the

J. Food Prot., Vol. 71, No. 1

HYGIENE INDICATOR CRITERIA FOR MEAT IN BELGIUM

sampling method and in the comparable recovery of microorganisms using the wet and dry swabbing method make the interpretation of the present study difficult in comparison with the European criteria. A comparative study of the European destructive method and the Belgian swabbing technique would need to be carried out to explain the results. The ECC, EC, and ACC from carcasses and other meat types were the highest for poultry followed by pork and beef (Tables 1 through 4). This result is in accordance with those of other studies (14). Most of the reports concerning microbiological hygiene indicators have focused on mean total bacterial counts on cattle and pig carcasses. In comparison with studies using cotton or gauze swabs and in which samples were obtained before chilling, the differences of ACC in the present samples (3.2 log CFU/cm2 for cattle and 2.6 log CFU/cm2 for pig carcasses) were ⫺0.6 to ⫹3.5 log CFU/cm2; the differences of ECC (⫺0.7 log CFU/cm2 for cattle and 1.1 log CFU/cm2 for pig carcasses) were ⫺2.5 to ⫺0.8 log CFU/cm2 for cattle carcasses. For EC on cattle (0.1 log CFU/cm2) and pig (1.0 log CFU/cm2) carcasses, the differences were ⫺2.4 to ⫹0.9 log CFU/cm2 (13, 18, 19, 23, 24, 31). The mean ACC (3.1 log CFU/g) and ECC (1.1 log CFU/g) in beef trimmings were comparable to findings by Phillips and coworkers (34) (ACC of 3.5 log CFU/g) and Scanga and coworkers (37) (ACC of 3.3 log CFU/g and ECC of 1.1 log CFU/g). The ECC in ground beef (1.1 log CFU/g) were also similar to the results reported by Scanga and coworkers (1.0 log CFU/g) (37). In pork trimmings and ground meat, the mean ECC (1.5 and 1.4 log CFU/g) were a maximum of 0.5-log higher than the counts observed by Duffy and coworkers (between 1 and 1.2 log CFU/g) (20). For broiler carcasses, the mean values in the present study (ACC of 5.8 log CFU/g, ECC of 3.5 log CFU/g, and PC of 5.9 log CFU/g) were 3.2-log lower to 3.0-log higher than the observed counts in the following other studies on neck or breast skin. Mead and coworkers (30) reported 4.4 to 5.1 log CFU/g for ACC and 2.9 to 4 log CFU/g for PC, Berrang and coworkers (12, 28) reported 4.4 log CFU/g for ACC and 6.7 log CFU/g for ECC, and Cason and coworkers (15) reported 3.8 log CFU/g for ECC. For broiler fillets (ECC of 2.2 CFU/g), Berrang and coworkers (12) reported 1.6 log CFU/g. Choice of indicators. In the present study, counts of E. coli, which is considered a fecal indicator, were highly correlated with counts of Enterobacteriaceae, which are used as indicators of fecal and environmental contamination, especially at the level of the slaughterhouse. ACC, which are used as an indicator of overall hygienic conditions, were significantly related to counts of Enterobacteriaceae and E. coli. For pork cuts, ground pork, and ground beef, E. coli was detected at a significantly higher level (median of 0.5 to 1.3 log CFU/g, P ⬍ 0.01) in samples contaminated with Salmonella. The elimination of the establishments with ECC higher than the 75th percentile would allow a significant decrease in Salmonella contamination (of 4.6% for minced beef and from 7.1 to 27.7% for minced pork, pig

43

carcasses, and pork cuts). In poultry samples, ECC were in general higher for samples containing Salmonella (median, 0.4- to 0.5-log higher in broiler carcasses and minced meat, P ⬍ 0.01) or Campylobacter (median, 0.1- to 0.4-log higher in broiler fillets, layer carcasses, and broiler minced meat, P ⬍ 0.01). If the samples with ECC higher than the 75th percentile were deleted, the prevalence of Salmonella in broiler prepared meat would decrease by 16.3% (P ⬍ 0.05). Even the prevalence of Campylobacter would decrease by 12.8% (layer carcasses, P ⬍ 0.01) to 20.6% (prepared meat, P ⬍ 0.05) if the samples with an ECC ⱖ75th percentile were eliminated. Such a relationship between the presence of Salmonella and ECC has also been observed by Sofos and coworkers (39) on cattle carcasses. E. coli thus seems to be a good index for Salmonella and Campylobacter; this organism is mesophilic and originates from the intestinal tract of slaughtered animals. However, Enterobacteriaceae are less useful; many species are found naturally in the environment, including meat processing plants, and the interpretation of results were only feasible for pig carcasses. In comparison with EC, one limitation of the use of ECC as an indicator is its lower occurrence on cattle and pig carcasses. The solution used in this study was to take into account all the results, by using the value of the detection limit divided by 2 for the results below the detection limit (24). For ACC on pig and chicken carcasses, only small or nonsignificant differences in the 50th percentiles were observed between the samples contaminated or not contaminated with Salmonella and Campylobacter. In general, the changes in bacterial counts between samples containing or not containing Salmonella or Campylobacter were not significant when ACC were higher than 4 log units, as observed by Hutchison and coworkers (25). By contrast, a low Pseudomonas count was associated with a higher chance of recovering Salmonella or Campylobacter from broiler carcasses. Pseudomonas spp. were present in high numbers on most of the poultry carcasses. Because of their multiplication at refrigerated temperatures, Pseudomonas spp. have potential for use as an indicator that would give more information than ACC, but this study showed that the PC was not as useful. In a previous study, Pseudomonas was dominant in spoiled meat (26). The present study confirmed that this bacterium cannot be considered a good index for enteric pathogens. E. coli could be used as a hygiene indicator for beef, pork, and poultry to provide information on the strictness of slaughtering procedures and the efficiency of scalding, singeing or flaming, evisceration, and chilling and to ensure the avoidance of fecal contamination, which dramatically limits pathogen contamination of meat. For Salmonella and Campylobacter (21, 22), during the period of surveillance poultry was the meat most contaminated with aerobic bacteria and E. coli because no physical decontamination of the carcasses (such as flaming or removal of the skin) occurred. No chemical decontamination method is authorized in the EU during the slaughtering process. Lowering the contamination level of poultry is therefore difficult, but im-

44

GHAFIR ET AL.

provements in the processing chain could be monitored with bacterial indicators. In this study, E. coli was a reliable index for enteric zoonotic agents such as Salmonella for beef and pork. In poultry samples, E. coli was a less reliable index for Salmonella and Campylobacter. Enterobacteriaceae were counted at a higher level but did not provide more information than did E. coli when ACC were monitored. These ACC gave a better indication of the overall hygiene status of an establishment than did counts of E. coli or Enterobacteriaceae. In conclusion, E. coli counts remain very useful for tracking the control of fecal contamination and could be counted simultaneously with aerobic colonies to track global hygiene. If only one indicator must be chosen, Enterobacteriaceae should be preferred because of its correlation with E. coli and its presence in the environment. Thus, we propose the use of ECC and ACC as European hygiene indicator criteria and suggest establishment of the criterion values on the 75th and 95th percentiles obtained in each country after official national surveillance plans. These national criteria should be used as a basis for establishing European hygiene indicator criteria for the destructive and swabbing sampling methods. The criteria could be regularly revised according to the progress observed in subsequent surveillance programs so that hygiene status will continue to improve. These surveillance plans could be coupled with the zoonotic agent surveillance plan performed according to European regulations and directives on the control of Salmonella and other specified foodborne zoonotic agents, as already carried out in Belgium since 2000 (7, 9). ACKNOWLEDGMENTS The Belgian Federal Agency for the Safety of the Food Chain financially supported this study. The authors also acknowledge Marc Cornelis, Jean-Yves Franc¸ois, Martine Jouret and Franc¸ois Ruttens for their work on the surveillance plans and Fre´de´ric Farnir for his support in statistics.

J. Food Prot., Vol. 71, No. 1

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17. 18.

19.

REFERENCES 1.

2.

3.

4.

5.

6.

Anonymous. 1996. Pathogen reduction: hazard analysis and critical control point (HACCP) systems, final rule. U.S. Food Safety and Inspection Service. Fed. Regist. 61:38805–38989. Anonymous. 1998. NF-V04–504. Viandes et produits a` base de viande. De´nombrement des Pseudomonas. Association Franc¸aise de Normalisation (AFNOR), La Plaine Saint-Denis, France. Anonymous. 1999. NF-V08–051. Microbiologie des aliments. De´nombrement des microorganismes par comptage des colonies obtenues a` 30 degre´s Celcius—me´thode de routine. Association Franc¸aise de Normalisation (AFNOR), La Plaine Saint-Denis, France. Anonymous. 1999. NF-V08–054. Microbiologie des aliments. De´nombrement des ente´robacte´ries par comptage des colonies obtenues a` 30 degre´s Celcius—me´thode de routine. Association Franc¸aise de Normalisation (AFNOR), La Plaine Saint-Denis, France. Anonymous. 2001. Commission Decision 2001/471/EC of 8 June 2001 laying down rules for the regular checks on the general hygiene carried out by the operators in establishments according to Directive 64/433/EEC on health conditions for the production and marketing of fresh meat and Directive 71/118/EEC on health problems affecting the production and placing on the market of fresh poultry meat. Off. J. Eur. Union L165:48–53. Anonymous. 2002. Arreˆte´ royal du 28 aouˆt 2002 modifiant l’arreˆte´ royal du 4 juillet 1996 relatif aux conditions ge´ne´rales et spe´ciales

20.

21.

22.

23. 24.

25.

26.

d’exploitation des abattoirs et d’autres e´tablissements. Moniteur Belge 14/09/2002, 2:40882–40894. Anonymous. 2003. Directive 2003/99/EC of the European Parliament and of the Council of 17 November 2003 on the monitoring of zoonoses and zoonotic agents, amending Council Decision 90/ 424/EEC and repealing Council Directive 92/117/EEC. Off. J. Eur. Union L325:31–40. Anonymous. 2003. ISO 17604. Microbiology of food and animal feeding stuffs—carcass sampling for microbiological analysis. International Organization for Standardization, Geneva. Anonymous. 2003. Regulation (EC) no. 2160/2003 of the European Parliament and of the Council of 17 November 2003 on the control of Salmonella and other specified foodborne zoonotic agents. Off. J. Eur. Union L325:1–15. Anonymous. 2005. Commission Regulation (EC) no. 2073/2005 of 15 November 2005 on microbiological criteria for foodstuffs. Off. J. Eur. Union L338:1–26. Anonymous. 2005. ISO 17025. General requirements for the competence of testing and calibration laboratories. International Organization for Standardization, Geneva. Berrang, M. E., S. R. Ladely, and R. J. Buhr. 2001. Presence and level of Campylobacter, coliforms, Escherichia coli, and total aerobic bacteria recovered from broiler parts with and without skin. J. Food Prot. 64:184–188. Byrne, B., G. Dunne, J. Lyng, and D. J. Bolton. 2005. Microbiological carcass sampling methods to achieve compliance with 2001/471/ EC and new hygiene regulations. Res. Microbiol. 156:104–106. Capita, R., M. Prieto, and C. Alonso-Calleja. 2004. Sampling methods for microbiological analysis of red meat and poultry carcasses. J. Food Prot. 67:1303–1308. Cason, J. A., M. E. Berrang, R. J. Buhr, and N. A. Cox. 2004. Effect of prechill fecal contamination on numbers of bacteria recovered from broiler chicken carcasses before and after immersion chilling. J. Food Prot. 67:1829–1833. Chahed, A., Y. Ghafir, B. China, K. Dierick, L. De Zutter, D. Pie´rard, and G. Daube. 2005. Survey of the contamination of foodstuffs from animal origin by Shiga toxin producing Escherichia coli serotype O157:H7 in Belgium from 1999 to 2003. Euro Surveill. 10:33–36. Churchill, G. A., and R. W. Doerge. 1994. Empirical threshold values for quantitative trait mapping. Genetics 138:963–971. De Zutter, L., R. Abrams, and J. Van Hoof. 1982. Bacteriological survey of beef carcasses: correlation between swab and maceration method. Arch. Lebensmittelhyg. 33:33–56. Dorsa, W. J., C. N. Cutter, and G. R. Siragusa. 1996. Evaluation of six sampling methods for recovery of bacteria from beef carcass surfaces. Lett. Appl. Microbiol. 22:39–41. Duffy, E. A., K. E. Belk, J. N. Sofos, G. R. Bellinger, A. Pape, and G. C. Smith. 2001. Extent of microbial contamination in United States pork retail products. J. Food Prot. 64:172–178. Ghafir, Y., B. China, K. Dierick, L. De Zutter, and G. Daube. 2007. A seven-year survey of Campylobacter contamination in meat at different production stages in Belgium. Int. J. Food Microbiol. 116: 111–120. Ghafir, Y., B. China, N. Korsak, K. Dierick, J.-M. Collard, C. Godard, L. De Zutter, and G. Daube. 2005. Belgian surveillance plans to assess changes in Salmonella prevalence in meat at different production stages. J. Food Prot. 68:2269–2277. Gill, C. O., and T. Jones. 2000. Microbiological sampling of carcasses by excision or swabbing. J. Food Prot. 63:167–173. Hutchison, M. L., L. D. Walters, S. M. Avery, C. A. Reid, D. Wilson, M. Howell, A. M. Johnston, and S. Buncic. 2005. A comparison of wet-dry swabbing and excision sampling methods for microbiological testing of bovine, porcine, and ovine carcasses at red meat slaughterhouses. J. Food Prot. 68:2155–2162. Hutchison, M. L., L. D. Walters, G. C. Mead, M. Howell, and V. M. Allen. 2006. An assessment of sampling methods and microbiological hygiene indicators for process verification in poultry slaughterhouses. J. Food Prot. 69:145–153. Jay, J. M., J. P. Vilai, and M. E. Hughes. 2003. Profile and activity

J. Food Prot., Vol. 71, No. 1

27.

28. 29.

30.

31.

32.

33.

34.

35.

HYGIENE INDICATOR CRITERIA FOR MEAT IN BELGIUM

of the bacterial biota of ground beef held from freshness to spoilage at 5–7⬚C. Int. J. Food Microbiol. 81:105–111. Korsak, N., G. Daube, Y. Ghafir, A. Chahed, S. Jolly, and H. Vindevogel. 1998. An efficient sampling technique used to detect four foodborne pathogens on pork and beef carcasses in nine Belgian abattoirs. J. Food Prot. 61:535–541. Kotula, K. L., and Y. Pandya. 1995. Bacteria contamination of broiler chickens before scalding. J. Food Prot. 58:1326–1329. McEvoy, J. M., J. J. Sheridan, I. S. Blair, and D. A. McDowell. 2004. Microbial contamination on beef in relation to hygiene assessment based on criteria used in EU Decision 2001/471/EC. Int. J. Food Microbiol. 92:217–225. Mead, G. C., W. R. Hudson, and M. H. Hinton. 1993. Microbiological survey of five poultry processing plants in the UK. Br. Poult. Sci. 34:497–503. Miraglia, D., D. Ranucci, V. D’Ovidio, R. Branciari, and M. Severini. 2005. Comparison between carcass microbial load recovered by swabbing surfaces of different size and using the reference excision method. Vet. Res. Commun. 29(Suppl. 2):339–341. Murray, K. A., A. Gilmour, and R. H. Madden. 2001. Microbiological quality of chilled beef carcasses in Northern Ireland: a baseline survey. J. Food Prot. 64:498–502. Pepperell, R., C. A. Reid, S. N. Solano, M. L. Hutchison, L. D. Walters, A. M. Johnston, and S. Buncic. 2005. Experimental comparison of excision and swabbing microbiological sampling methods for carcasses. J. Food Prot. 68:2163–2168. Phillips, D., J. Sumner, J. F. Alexander, and K. M. Dutton. 2001. Microbiological quality of Australian beef. J. Food Prot. 64:692– 696. Ransom, J. R., K. E. Belk, R. T. Bacon, J. N. Sofos, J. A. Scanga,

36.

37.

38.

39.

40.

41.

42.

43.

45

and G. C. Smith. 2002. Comparison of sampling methods for microbiological testing of beef animal rectal/colonal feces, hides, and carcasses. J. Food Prot. 65:621–626. Ray, B. 2001. Indicators of bacterial pathogens, p. 409–417. In B. Ray (ed.), Fundamental food microbiology, 2nd ed. CRC Press, Boca Raton, Fla. Scanga, J. A., A. D. Grona, K. E. Belk, J. N. Sofos, G. R. Bellinger, and G. C. Smith. 2000. Microbiological contamination of raw beef trimmings and ground beef. Meat Sci. 56:145–152. Sofos, J. N., S. L. Kochevar, G. R. Bellinger, D. R. Buege, D. D. Hancock, S. C. Ingham, J. B. Morgan, J. O. Reagan, and G. C. Smith. 1999. Sources and extent of microbiological contamination of beef carcasses in seven United States slaughtering plants. J. Food Prot. 62:140–145. Sofos, J. N., S. L. Kochevar, J. O. Reagan, and G. C. Smith. 1999. Incidence of Salmonella on beef carcasses relating to U.S. meat and poultry inspection regulations. J. Food Prot. 62:467–473. Vanderlinde, P., I. Jenson, and J. Sumner. 2005. Using national microbiological data to set meaningful performance criteria for slaughter and dressing of animals at Australian export abattoirs. Int. J. Food Microbiol. 104:155–159. Vanderlinde, P. B., B. Shay, and J. Murray. 1998. Microbiological quality of Australian beef carcass meat and frozen bulk packed beef. J. Food Prot. 61:437–443. Ware, L. M., M. L. Kain, J. N. Sofos, K. E. Belk, and G. C. Smith. 1999. Comparison of sponging and excising as sampling procedures for microbiological analysis of fresh beef-carcass tissue. J. Food Prot. 62:1255–1259. Yu, S. L., P. H. Cooke, and S. I. Tu. 2001. Effects of chilling on sampling of bacteria attached to swine carcasses. Lett. Appl. Microbiol. 32:205–210.