Quantitative Determination of Campylobacter on Broilers along 22 United Kingdom Processing Lines To Identify Potential Process Control Points and Cross-Contamination from Colonized to Uncolonized Flocks.

ABSTRACT: As part of a program to reduce numbers of the human pathogen Campylobacter on retail chickens, 22 broiler processing lines, representing more than 90% of UK production, were characterized by enumerating Campylobacter on pooled neck skins after exsanguination, scalding, defeathering, evisceration, crop removal, inside-outside washing, and air-chilling stages of processing. Sixteen of the processing lines investigated showed significant (P < 0.05) reductions in Campylobacter numbers because of carcass scalding. However, in all of these lines, the following defeathering stage caused a significant increase in Campylobacter contamination that effectively negated the reductions caused by scalding. On four processing lines, primary chilling also caused a significant reduction in numbers of Campylobacter. On three lines, there was a significant microbiological benefit from inside-outside washing. The stages where Campylobacter numbers were reduced require further investigation to determine the specific mechanisms responsible so that the observed pathogen reductions can be optimized and then more widely implemented. The transfer of up to 4 log CFU Campylobacter per g of neck skin from a colonized flock to a following uncolonized flock was observed. Cross-contamination was substantial and still detectable after 5,000 carcasses from an uncolonized flock had been processed. Numbers of Campylobacter recovered from the uncolonized flocks were highest on the first of the uncolonized birds to pass along the line, and in general, the numbers declined as more uncolonized birds were processed. Air sampling recovered low numbers at the processing stages monitored, indicating that airborne transmission was unlikely to be the primary transfer mechanism operating for cross-contamination between flocks.

Consequences of Using Two Types of Skin Samples from Chilled Chicken Broiler Carcasses To Measure the Degree of Contamination by Campylobacter spp.

HIGHLIGHTS: Campylobacter levels on chicken neck and breast skin were compared. Neck skin was significantly more contaminated (P < 0.05) than breast skin. No relationship between the two skin types was found for Campylobacter levels. A UK government reduction target for highly contaminated chicken was not achieved.

Modelling-based identification of factors influencing campylobacters in chicken broiler houses and on carcasses sampled after processing and chilling.

AIMS: To identify production and processing practices that might reduce Campylobacter numbers contaminating chicken broiler carcasses. METHODS AND RESULTS: The numbers of campylobacters were determined on carcass neck skins after processing or in broiler house litter samples. Supplementary information that described farm layouts, farming conditions for individual flocks, the slaughterhouse layouts and operating conditions inside plants was collected, matched with each Campylobacter test result. Statistical models predicting the numbers of campylobacters on neck skins and in litter were constructed. Carcass microbial contamination was more strongly influenced by on-farm production practices compared with slaughterhouse activities. We observed correlations between the chilling, washing and defeathering stages of processing and the numbers of campylobacters on carcasses. There were factors on farm that also correlated with numbers of campylobacters in litter. These included bird gender, the exclusion of dogs from houses, beetle presence in the house litter and the materials used to construct the house frame. CONCLUSIONS: Changes in farming practices have greater potential for reducing chicken carcass microbial contamination compared with processing interventions. SIGNIFICANCE AND IMPACT OF THE STUDY: Routine commercial practices were identified that were correlated with lowered numbers of campylobacters. Consequently, these practices are likely to be both cost-effective and suitable for adoption into established farms and commercial processing.

Restoring the selectivity of Bolton broth during enrichment for Campylobacter spp. from raw chicken.

AIMS: When isolating Campylobacter spp. from retail raw chicken using BS EN ISO 10272-1:2006, contaminants frequently cause overgrowth on mCCDA plates. Therefore, these organisms proliferate in the enrichment medium, Bolton broth, indicating a lack of selectivity in this medium. This study sought to characterize the contaminant flora and to devise a modified Bolton broth to inhibit their growth. METHODS AND RESULTS: Contaminants (n=30) from separate samples were identified and antibiotic resistances determined. Most (93%) were extended spectrum ß-lactamase (ESBL) producing Escherichia coli, able to hydrolyse the cefoperazone present in Bolton broth and mCCDA. To inhibit these organisms, original formulation Bolton broth was supplemented with potassium clavulanate, at three concentrations, and recoveries of campylobacters from raw chicken were determined. Using standard Bolton broth, only 49% of samples (n=104) yielded campylobacters, but supplementation with 2 mg l(-1) potassium clavulanate increased this significantly (P<0.05), with 91% of samples positive. CONCLUSIONS: Potassium clavulanate can restore the selectivity of Bolton broth when isolating Campylobacter spp. from raw chicken. SIGNIFICANCE AND IMPACT OF THE STUDY: Raw chicken is often contaminated with the pathogen Campylobacter, but the ISO methodology for its detection is becoming compromised by the increasing presence of antibiotic-resistant bacteria. A simple modification ensures effective detection of this pathogen.

Comparison of disc diffusion and epsilometer (E-test) testing techniques to determine antimicrobial susceptibility of Campylobacter isolates of food and human clinical origin.

The antibiotic resistance profiles of 75 Campylobacter isolates of food and human clinical origin was determined by two agar diffusion susceptibility methods; disc diffusion and epsilometer-test (E-test). The most common therapeutic antimicrobials, erythromycin, ciprofloxacin and tetracycline were studied, along with chloramphenicol, ampicillin and naladixic acid. The resistance observed for each antimicrobial, as determined by both of methods, were statistically compared using Fisher two-tailed analysis. Of the six antimicrobials studied only two were shown to have statistically different patterns when resistance was compared by disc diffusion and E-test. The percentage of isolates resistant to clinically relevant antimicrobials using both techniques ranged from 6.6 to 21.3% for erythromycin, 25.3-26.6% for tetracycline and 33.3-36.0% for ciprofloxacin. The prevalence of multi-drug resistant (MDR) campylobacters (isolates resistant to 2 or more antimicrobials) for both disc diffusion and E-test was 44%. It can be concluded that, for four of the six antimicrobials assessed, antimicrobial resistance prevalences could be equally determined by either of the methods studied.

Factors affecting the recovery of Campylobacter spp. from retail packs of raw, fresh chicken using ISO 10272-1:2006.

AIMS: This study sought to determine the most effective protocol for the detection of Campylobacter spp. in retail packs of fresh, raw chicken based on ISO 10272-1:2006. METHODS AND RESULTS: Three sample preparation protocols were studied; two based on excision and one combining excision with a rinse of the remaining sample. Enrichment cultures were incubated both in closed bottles and microaerobically, and sub-cultured at 24 and 48 h. Packs of chicken (110) were analysed and only two yielded no Campylobacter spp. Subculturing enrichment broths at 24 h gave the same prevalence as at 48 h, P > 0.4 but microaerobic incubation yielded approximately 50% more positive samples than did incubation in closed bottles. Sampling based on excision plus rinsing gave the highest Campylobacter prevalence (92.7%). CONCLUSIONS: To isolate Campylobacter spp. from retail packs of chicken, enrichment cultures must be incubated in a microaerobic atmosphere and sub-cultured at 24 h and, possibly, 48 h. Sampling packs by excision plus rinsing maximized recoveries. SIGNIFICANCE AND IMPACT OF THE STUDY: ISO 10272-1:2006 permits the use of inefficient protocols which markedly underestimate the true prevalence of Campylobacter spp. in retail, fresh chicken. Equivalent results could be obtained 24 h earlier, with consequent savings. Its revision is essential.

Use of conductimetry to rapidly determine relative stress sensitivity in Salmonella isolates.

AIMS: To compare conventional plate counting and indirect conductimetry as techniques for ranking the resistance of Salmonella spp. to processing stressors. METHODS AND RESULTS: Forty Salmonella isolates were subjected to three separate stressors used in food processing; irradiation, heat and high hydrostatic pressure (HHP). Total viable counts (TVC) using conventional plate counts and time to detection (TTD) using indirect conductimetry were determined. A significant negative correlation between TVC and TTD was seen with irradiation (P < 0.01) and heat (P < 0.05) but not HHP. CONCLUSIONS: For a group of salmonellas, indirect conductimetry can rapidly determine a ranking of isolate sensitivity to irradiation and heat. However, for HHP, the results indicated that conventional plate counting alone cannot be used to determine sensitivity. SIGNIFICANCE AND IMPACT OF THE STUDY: The resistance of micro-organisms to processing systems must be ranked to allow the selection of appropriate isolates for process validation. TTD measurements allow rapid screening of salmonellas to rank isolates for resistance to irradiation and heat stress. However, following HHP, the TVC of survivors is independent of the time required for growth to a set cell density and therefore it cannot be used as the sole measure of relative stress resistance.

Use of Hugh and Leifson's medium as a simple screening test to aid in the differentiation of Arcobacter spp. from background flora during their isolation from foodstuffs.

AIMS: To investigate the suitability of Hugh and Leifson's medium (HLM) as the basis of a simple screening test to differentiate between contaminants and Arcobacter spp. during their isolation from foodstuffs. METHODS AND RESULTS: Characterized Arcobacter spp. were obtained from recognized culture collections. Wild-type isolates of Arcobacter spp. and contaminants were obtained using published isolation protocols. Retail packs of red meats were used as the source of the isolates. Eighteen defined Arcobacter spp. gave no reaction on HLM, as did 10 local wild-type isolates. Overall 163 contaminants were studied for oxidative reactions on HLM and 86% of isolates demonstrated this property. CONCLUSIONS: HLM can usefully serve as a simple and effective screening test to differentiate between Arcobacter spp. and contaminants. SIGNIFICANCE AND IMPACT OF THE STUDY: Arcobacter isolation procedures are still being developed, and no effective diagnostic media currently exist. Rapidly excluding most contaminants can markedly increase the efficiency of isolation procedures by removing the need for extensive biotyping or the requirement to isolate DNA and conduct PCR tests.

Diversity of Campylobacter coli genotypes in the lower porcine gastrointestinal tract at time of slaughter.

AIMS: To compare the genotypes of Campylobacter coli obtained from the rectal and ileal samples of pigs at the time of slaughter. METHODS AND RESULTS: Five animals were sampled following slaughter with ileal contents and anal swabs being taken post-evisceration. Swabs were directly plated onto charcoal cefoperazone desoxycholate agar (CCDA) while ileal contents were enriched in CCDA broth. Twenty isolates were picked from each site sampled and all 200 isolates were Camp. coli. Isolates were genotyped using random amplified polymorphic DNA (22 discrete types) and flaA (11 discrete types). Both methods found that 55% of the genotypes were unique to rectal samples. Only one animal yielded the same flaA type from ileal and rectal samples. CONCLUSIONS: Rectal sampling of pigs yielded a more diverse subset of Camp. coli genotypes than ileal contents, but failed to yield all of the genotypes carried by an individual animal. SIGNIFICANCE AND IMPACT OF THE STUDY: A small sample of pigs carried a very diverse population of Camp. coli genotypes; and sampling of a single site in the gut will recover only part of this population. Hence, any genotyping studies of Camp. coli in pigs must be interpreted with caution, and epidemiological studies could be confounded by the number of Camp. coli genotypes available.

A genotyping investigation of the colonization of piglets by Campylobacter coli in the first 10 weeks of life.

AIMS: To examine the Campylobacter genotypes colonizing a litter of piglets during the first 10 weeks of life and compare them with those of the sow. METHODS AND RESULTS: Campylobacters were isolated by direct plating of anal swabs. Piglets (n = 6) were sampled six times and five isolates per piglet obtained each time. The sow was also sampled but 20 isolates per sampling obtained. Isolates were genotyped by random amplification of polymorphic DNA, pulsed field gel electrophoresis and polymerase chain reaction/restriction fragment length polymorphism of the flagellin gene. Initially piglets were colonized by Campylobacter coli genotypes from the mother but after 66 days 33% of piglet isolates (based on a mean of the three-genotyping methods) were from other sources. The sow died after 14 days and the initial Campylobacter flora of the foster sow was subsequently replaced by genotypes from the piglets mother. However these constituted only a minor part of her flora after 52 days. Both foster sow and piglets carried multiple genotypes of Camp. coli: up to four in a single piglet sample and seven in one from the sow. SIGNIFICANCE AND IMPACT OF THE STUDY: Piglets are initially colonized by Camp. coli genotypes from their mother but later other genotypes displace them. Colonization is dynamic with the sow able to acquire genotypes from the piglets. CONCLUSIONS: The large numbers of Camp. coli genotypes carried by pigs, and frequent successions of dominant types, will render epidemiological studies problematic.

Carriage of four bacterial pathogens by beef cattle in Northern Ireland at time of slaughter.

AIMS: To determine the prevalence of four bacterial zoonotic pathogens in beef cattle at time of slaughter in Northern Ireland (NI), in order to assess their potential for reducing beef safety. METHODS AND RESULTS: Faeces were collected postmortem from beef cattle (n =220) at seven EU registered abattoirs. Standard enrichment culturing methods were employed, plus immunomagnetic enrichment in the case of Escherichia coli O157:H7. Campylobacter spp. were found in 52 samples (24.8%), Listeria monocytogenes in 10 (4.8%), E. coli O157:H7 in 2 (0.9%) whilst Salmonella spp. were isolated from six out of 200 samples (3.0%). Five salmonellas were Salmonella Chandans and one was Salmonella Liverpool. CONCLUSIONS: Campylobacter spp. were the most frequently isolated pathogen, despite being relatively rare in beef. Genotyping showed the campylobacters to be very diverse, indicating cattle encounter campylobacters from many sources. The remaining three pathogens, which are associated with meats, occurred at relatively low frequencies, especially E. coli O157:H7. The Salmonella serovars found rarely infect humans. SIGNIFICANCE AND IMPACT OF THE STUDY: The low prevalence of E. coli O157:H7 in NI beef cattle was confirmed and the reasons for this merit further study. The four pathogens should have little impact on beef quality.

Antibiotic resistance of retail food and human Campylobacter isolates on the island of Ireland from 2001-2002.

The antimicrobial resistance profiles of Campylobacter isolates recovered from a range of retail food samples (n=374) and humans (n=314) to eight antimicrobial compounds were investigated. High levels of resistance in food C. jejuni isolates were observed for ceftiofur (58%), ampicillin (25%) and nalidixic acid (17%) with lower levels observed for streptomycin (7.9%) and chloramphenicol (8.3%). A total of 80% of human C. jejuni isolates were resistant to ceftiofur, while 17% showed resistance to ampicillin and nalidixic acid, 8.6% to streptomycin and 4.1% to chloramphenicol. Resistance to clinically relevant antimicrobials such as erythromycin, ciprofloxacin and tetracycline was 6.7, 12, and 15% respectively for all food isolates and was similar to corresponding resistance prevalences observed for human isolates, where 6.4, 12 and 13% respectively were found to be resistant. Comparisons of C. jejuni isolates in each location showed a high degree of similarity although some regional variations did exist. Comparison of total C. jejuni and C. coli populations showed minor differences, with C. jejuni isolates more resistant to ampicillin and ceftiofur. Multidrug resistance patterns showed some profiles common to human and clinical isolates.

Evaluation of the use of conductimetry for the rapid and precise measurement of Salmonella spp. growth rates.

The growth rates of 14 Salmonella serovars in tryptone soy broth plus yeast extract (TSBYE) were estimated using conventional plating techniques and indirect conductimetry using a Don Whitley RABIT system. Both methods gave identical results for the maximum specific growth rate (mumax) P>0.05. However, using the conductimetric method, mumax for a single serovar was determined in less than 7 h, whereas the conventional method required an additional 24 h. In addition, the conductimetric method was considerably more precise, much less labour-intensive and required the use of considerably less consumables. Using conductimetry, a trained operator could accurately determine mumax for 24 serovars in 3 working days, but only one serovar using the conventional plate counting technique. Hence, the use of conductimetry can markedly increase the precision and accuracy of mumax determinations by allowing a very significant increase in the number of results obtained and in their precision. The data generated will allow the development of better mathematical growth models. The method can also be used to compare growth media and conditions and hence rapidly optimise detection protocols for this pathogen.

Occurrence of Campylobacter in retail foods in Ireland.

A surveillance study was carried out to determine the prevalence of Campylobacter in a range of retail foods purchased in three Irish cities over a 20-month period between March 2001 and October 2002. In total 2391 food samples were analysed during this period. Campylobacter was isolated from 444 raw chicken (49.9%), 33 turkey (37.5%) and 11 duck samples (45.8%). Lower isolation rates of 7/221 (3.2%), 10/197 (5.1%) and 31/262 (11.8%) were observed for raw beef, pork and lamb, respectively. One sample of pork paté from 120 samples analysed (0.8%) was Campylobacter-positive. A total of three shellfish samples (oysters) from 129 raw specimens examined (2.3%) were found to contain Campylobacter. Low prevalences of the organism (0.9%) were also isolated from fresh mushrooms. Of 62 raw bulk tank milk samples analysed, Campylobacter was recovered in a single sample (1.6%). Campylobacter was not detected in any of the comminuted pork puddings, prepared vegetables and salads, retail sandwiches or cheeses made from unpasteurised milk. In total, 543 Campylobacter were isolated from all of the food samples analysed, of which 453 (83.4%) were confirmed as Campylobacter jejuni and the remaining 90 (16.6%) as Campylobacter coli.

Comparison of 40 Salmonella enterica serovars injured by thermal, high-pressure and irradiation stress.

AIM: To investigate and compare the inherent resistance of 40 Salmonella serovars to heat, irradiation and high-pressure stress. METHODS AND RESULTS: D10 values for each of the three stresses were calculated for four serovars, chosen as representatives from a catalogue of 40. Based on these results, conditions for each stress were defined, which produced, on average, a three-log reduction in viability. Heat stress (57 degrees C for 13 min), high-pressure stress (350 MPa for 10 min at 20 degrees C) and irradiation stress (1.5 kGy at 20 degrees C) were applied to all 40 serovars in the collection. Injury and loss of viability for all serovars were determined. CONCLUSIONS: Cluster analysis identified five groupings of isolates in terms of resistance to the applied stresses. The independent response of each isolate to all three stresses suggests that there is no relationship between resistances. SIGNIFICANCE AND IMPACT OF THE STUDY: Each serovar is inherently different. For modelling of real-life food preservation processing the most resistant isolates for that process should be chosen. The results also emphasize the importance of including multiple stress resistant strains when food preservation systems apply multiple stresses.

Inter-laboratory evaluation of three flagellin PCR/RFLP methods for typing Campylobacter jejuni and C. coli: the CAMPYNET experience.

AIMS: To compare typeability, discriminatory ability, and inter-laboratory reproducibility of three flagellin PCR/RFLP (fla typing) methods previously described for Campylobacter. METHODS AND RESULTS: The sample set (n = 100) was diverse, including both C. jejuni (n = 85) and C. coli (n = 15). Two of the three flaA typing methods amplified flaA alone, whereas one, a multiplex assay, amplified flaB in addition to flaA. DdeI restriction enzyme was employed for all methods, but HinfI was also investigated. 98-100% typeability was obtained for flaA-based methods, but only 93% for the multiplex assay, due to inconsistent amplification of a non-specific product. In addition, there appeared to be selective amplification of flaA over flaB. More DdeI types were generated using a longer flaA PCR amplicon, whilst additional use of HinfI increased the number of types by ca 25%. Inter-laboratory reproducibility for both flaA-based methods was defined at 100%. CONCLUSIONS: Fla typing requires standardization with respect to PCR primers and restriction enzymes. This study identified an assay, employing the full flaA gene and DdeI digestion, as an appropriate method on which to standardize. 100% inter-laboratory reproducibility was demonstrated using that method. SIGNIFICANCE AND IMPACT OF THE STUDY: This work should facilitate progress towards inter-laboratory standardization of fla typing.

Presence of Listeria and Salmonella spp. in retail chicken in Northern Ireland.

AIMS: Retail packs of fresh chicken in Northern Ireland were sampled to determine the frequency with which they were contaminated with Salmonella and Listeria spp. METHODS: Packs of chicken were chosen from supermarkets ensuring a diverse range of EU producer codes were sampled. Salmonellas were isolated using BS EN 12824: 1998 methodology, biotyped and serotyped whilst Listeria spp. were isolated based on EN ISO 11290-1: 1996 procedures and identified using a multiplex PCR system utilizing genus and species specific primers. SIGNIFICANCE AND IMPACT OF THE STUDY: Only three of 205 samples yielded Salmonella spp. indicating that measures undertaken by the poultry industry to control this pathogen have apparently been successful. However, Listeria spp. were present in 38 of 80 samples tested (48%) and 14 (18%) yielded Listeria monocytogenes. Thus Salmonella controls do not markedly affect this pathogen and retail packs of raw chicken must be considered a potential source of L. monocytogenes, and appropriate precautions taken to prevent infection.

Molecular diversity of Campylobacter coli and C. jejuni isolated from pigs at slaughter by flaA-RFLP analysis and ribotyping.

A population of porcine isolates of Camplobacter jejuni (n = 11) and C. coli (n = 17) were examined for genotypic relatedness employing ribotyping, as well as polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) analysis of the flagellin (fla)A gene locus. PCR was employed to amplify a 533 bp fragment from the flaA gene, including the previously described short variable region (SVR), employing the novel primers, A2 and Al and successfully generated this amplicon for all wild-type strains examined (n = 28) of both C. jejuni and C. coli, as well as with both type strains, i.e. C. jejuni NCTC 11351 and C. coli NCTC 11366. Individual genotypes were assigned to each isolate typed employing the four typing methods (flaA-RFLP(Hae) III, flaA-RFLP(Pst) I ribotyping(Hae) III and ribotyping(Pst) I) and were assigned an arbitrary genotype code in ascending alphabetical order in comparison with a database of established genotypes for each of the methods employed. This study showed that several flaA-RFLP and ribopatterns existed within C. jejuni and C. coli, and demonstrated a heterogeneous diversity of strains occurring in the pigs examined. Ribotyping of strains with 16S and 23S rRNA with Pst I and Hae III digested chromosomal DNA allowed subdivision of strains into nine and eight groups, respectively. RFLP analyses with Pst I and Hae III digests probed with the flaA gene probe allowed subdivision of strains into eight and eleven subtypes, respectively. Employment of RFLP with the flaA nucleic acid probe and Hae III digests produced the greatest amount of variation of any genotyping scheme employed. Although there was a high degree of variability demonstrated by both typing methods, most isolates ( > 60%) clustered into four main genotypes, i.e. genotypes A-D. FlaA-PCR-RFLP typing demonstrated that the majority of isolates, 67.9 and 60.7%, were included in these four main genotypes for Pst I and Hae III restriction digests, respectively, although there was a high prevalence (7/11; 63.6%) of fla(Hae) III genotype A occurring within the C. jejuni isolates. Likewise, ribotyping studies demonstrated that most isolates were clustered into these four main genotypes, accounting for 81.5 and 60.7% of isolates for Pst I and Hae III restriction digests, respectively. This may indicate that the clonal population of campylobacters within this pig population is largely composed of persistent and dominant types, with a smaller number of hypervariable subtypes. Such data may useful in determining epidemiological routes of transmission of campylobacters from animal to animal, as well as helping to identify virulence determinants in persistent subtype populations.

Subspecies characterization of porcine Campylobacter coli and Campylobacter jejuni by multilocus enzyme electrophoresis typing.

Enteropathogenic Campylobacterjejuni, C. coli and C. lari are currently the most common causes of acute infectious diarrhoeal illness in the UK. Many domestic animals, including pigs, act as natural reservoirs of these organisms and infection may occur through the ingestion of contaminated foodstuffs. C jejuni and C. coli, isolated from the livers of bacon pigs, were examined at subspecies level by multilocus enzyme electrophoresis (MEE) typing with seven enzymic loci. Polymorphological variation was highest with indophenol oxidase, isocitrate dehydrogenase and L-phenylalanyl-L-leucine peptidase giving 5. 5 and 4 alleles at these loci, respectively. The 35 Campylobacter isolates examined in this study (12 C. jejuni and 23 C coli) represented 30 unique electrophoretic types (ETs). Of these ETs, 8 unique types were detected for the 12 C jejuni isolates and 19 unique ETs were detected for the 23 C coli isolates. In addition, 3 types (ETs 2, 5, 10) were shared in common among C. jejuni and C coli. The average number of alleles per enzyme locus was 3.28. The mean genetic diversity, i.e. arithmetic average over all loci assayed, including monomorphic values, was 0.5573 and 0.5350 for C jejuni and C coli. respectively. Alleles were shared by C jejuni and C coli, suggesting an exchange of genetic material between the species. MEE analyses of isolates showed that there was a wide range of subspecies types within both C. jejuni and C coli in porcine livers. In certain cases, up to four phenotypically different strains of C coli were isolated from one liver, indicating multiple infections.

Microbiological quality of chilled beef carcasses in Northern Ireland: a baseline survey.

To standardize the assessment of the hygienic quality of beef carcasses in Northern Ireland (NI) abattoirs, swabbing techniques were evaluated. Six materials, including two commercially produced swabs, were compared for their ability to recover spoilage and pathogenic bacteria and for their ease of use as carcass swabs. A sponge retailed for domestic use was selected on the basis of efficiency of recovery of microorganisms, ease of use, and cost. On sample carcasses, 1,000 cm2 of the brisket was swabbed, since this site is normally readily contaminated. For 9 months, 420 carcasses in seven of the nine European Union-approved abattoirs in NI were sampled while in the chiller (24 to 48 h after kill). Total viable count (TVC), yeasts and molds, and Enterobacteriaceae were enumerated after incubation at 22 (48 h) and 37 degrees C (48 h), and the results were expressed as log CFU/cm2. The mean TVC results at 22 and 37 degrees C were 2.80+/-0.70 and 2.75+/-0.64, respectively. Although 63% of samples had yeasts that grew at 22 degrees C, only 35% were positive at 37 degrees C. The respective mean yeast counts were 1.12+/-0.59 and 0.46+/-0.51. Enterobacteriaceae were present in 15% of samples at 22 degrees C and 21% of samples at 37 degrees C. The mean counts for positive samples were 0.41+/-0.37 and 0.40+/-0.30, respectively. Molds were found in less than 4% of samples. Given that the brisket is normally one of the most heavily contaminated parts of the carcass, these results suggest that good hygienic practices are in operation in NI abattoirs. The results also enabled the abattoirs with the cleanest carcasses to be identified, hence permitting best practices to be found.