Food Protection Trends - June 2010 - 337

calculated. The results of this study are presented in this report.

MATERIALS AND METHODS
Samples were collected in January and February 2008 from five abattoirs. Sampling was carried out on 14 beef carcasses that had been held under active refrigeration for 16 to 24 h. On each carcass side, an area measuring 10 cm × 10 cm (100 cm2) was marked at each of the rump, flank and brisket sites, using a knife sanitized by immersion in water at 82°C. One side of a polyurethane sponge (Whirlpak speci-sponge, NASCO, USA) moistened with Butterfield’s solution (25 ml; bioMérieux) was used to sample each site. A total of 10 samplers were used, all of whom had undergone a training course in sponge sampling and were either industry/government inspectors or research staff. The ability of each operator to sponge each site was assessed according to whether the operator conformed with the technique prescribed in the Microbiological Guidelines that accompany the Australian Standard for production and transportation of meat and meat products for human consumption (2): “Wipe the sponge over the sampling area (10 cm × 10 cm) approximately 10 times in the vertical and 10 times in the horizontal directions. The pressure of sponging is important and should be as if you are removing dried blood from the carcase. However, the pressure should not be so hard as to crumble or destroy the sponge.” Using a knife sanitized by immersion in water at 82°C, the previouslysponged and demarcated area was carefully excised by slicing approximately 2 mm below the surface, and the sample was placed in a sterile Stomacher bag. All samples were packed in insulated containers with chiller packs for transportation to the laboratory, a journey never longer than 3 h. At the laboratory, samples were held at 2–4°C until examination within 1 h of arrival. After the sponge had been manually squeezed several times and the fluid had been stripped from it, serial dilutions of the fluid were prepared in 0.1% peptone water (Oxoid, Hampshire, England). Aliquots (1 ml) were trans. ferred to APC Petrifilm® (3M, Sydney, Australia). To each excision sample, 30 ml of peptone water (0.1%) containing

2% (v/v) Tween 80 (Merck Pty Ltd, Victoria, Australia) was added and the tissue homogenized in a stomacher (Colworth Stomacher 400, A.J. Seward & Co. Ltd, London, UK) for 2 minutes. Aliquots were diluted and plated as already described. Duplicate Petrifilm® plates were incubated at 25°C for 72 h, after which time colonies were counted. The limit of detection for sponge samples was 0.25 CFU/cm2 and for excised samples 0.33 CFU/cm2. The total number of bacteria recovered was defined as the number obtained by sponging plus the number from the excised tissue. All counts were converted to counts per square centimeter and analyzed using analysis of variance to test for differences between sites (rump, brisket, flank). Plant and Operator were also included in the model, so that the variability between plant and the variability between operators could be evaluated. The variability between operators within each plant was confounded with carcass, as each operator sponged only a single carcass. Only Operator 10 sponged carcasses at each of the five plants.

RESULTS AND DISCUSSION
Bacterial numbers recovered when ten experienced operators sponged sites of beef carcasses chilled overnight are presented in Table 1. At each establishment, operators sampled adjacent carcass sides, each sponging an area marked by Operator 10, who also excised each site after sponging was completed. The mean proportion of bacteria removed at each site was 39.1% (rump), 39.9% (flank) and 33.7% (brisket) and the standard deviation at each site was relatively high (28.5, 21.3 and 17.4%, respectively), reflecting the wide variation of recovery among operators (2.3 – 93.1%). There was no significant difference among sites, on average, for the proportion of bacteria removed (P = 0.67) or aerobic plate counts (APCs) (P = 0.19). Although all operators had received training some years previously in carcass sponging, conforming exactly to procedures set out in the Australian Standard, their technique in this 2008 study varied considerably. Important departures from the standard method were doublingover the sponge (effectively halving the area available for removing bacteria) and sponging “lightly,” differences which might be expected to reduce recovery of

bacteria, or using more than 10 up-anddown strokes and exceeding the marked area, which might be expected to increase recovery. This is in agreement with a recent study in which total viable counts were also shown to be significantly different depending on the person sampling the carcass, as well as the animal species tested and the bacterial load (7). However, unlike the present study, operator variability using sponging and excising methods on different carcass sites was compared. To determine the proportion that could be removed by continued sponging, a single operator (Operator 10) undertook sponging and excision on two carcass sides. Numbers of bacteria recovered from each site after use of five separate sequential sponges are presented in Table 2, together with the numbers remaining on the sponged site measure by excision. The proportion of bacteria recovered by sequential sponging varied between 11.1 and 97.4% and, with one exception, larger proportions were recovered from the first sponge. As indicated in Table 1, sponges were capable of removing relatively large numbers of bacteria, with 26.2 and 28.6% of sponges removing more than 50% of the surface load and containing more than 100,000 colony-forming units (CFU), respectively. In contrast, the sponge used by Operator 10 removed only 2.3% of the 1.1 million bacteria from a rump site, indicating that there may be other factors, apart from operator technique, that influence removal of bacteria. One factor, which was noted during the study, was fat cover, with the possibility that the pores of the sponge become occluded, thereby reducing removal of bacteria. This is in addition to variables such as rates of bacterial attachment to the meat surface, uneven distribution of bacteria on the carcass, whether carcasses are sampled “warm” or chilled, abrasiveness of the swab/sponge, and the vigor with which the sponge is applied to the site (4, 5, 6, 9). In addition, it is well recognized that the sponge itself retains bacteria. A further variable confounding comparison of non-destructive and excision sampling methods is that previous studies have compared bacterial populations on different sample sites, either on the same or different carcasses (3, 4, 6, 7, 8, 9). By contrast, the present study limited the number of variables listed above, first because the sponged area was the

JUNE 2010 | FOOD PROTECTION TRENDS

337

Food Protection Trends - June 2010

Table of Contents for the Digital Edition of Food Protection Trends - June 2010