Microbiological verification of the control of the processes of dressing, cooling and processing of beef carcasses at a high line-speed abattoir

Methods were described, and are offered to inspection agencies, to verify the microbiological adequacy of the processes of dressing, chilling and fabricating beef carcasses. The isolation rate of specific pathogens from the fabrication floor was also determined for pre-bagged subprimal beef cuts at the end of fabrication. The microbiological adequacy of the process on the slaughter floor was verified by inspection staff of Agriculture and Agri-Food Canada over an 11-month period at one high line-speed abattoir. Similar methods using hydrophobic grid membrane filter (HGMF) technology were applied by research projects at the same abattoir to verify the microbiological adequacy of the processes of chilling and fabricating beef carcasses. Ten excision samples (5 x 5 cm) were taken from each of 15-20 carcasses per evaluation of the slaughter and chill process. The range of the estimated mean log(10) of the most probable number of growth units per centimeter squared (LMPN) of 11 monthly evaluations (and a second evaluation in June 1993) of the slaughter floor was 1.64 +/- 0.06-2.0 +/- 0.07. The quality control programs, which were created by the management for the slaughter process at this abattoir, were verified to be working effectively for the 11 month period under study For a single evaluation of the chill floor the LMPN was 1.22. On the fabrication floor 140 subprimal cuts of carcasses were examined after bagging (pre-vacuum) at the end of five conveyer lines. Seventy surfaces freshly 'cut' and 70 surfaces 'not-cut' on the fabrication floor were sampled as above. Most probable numbers of aerobic bacteria, coliforms and Escherichia coli were determined per centimeter squared and log transformed (LMPN, LCOL and LEG, respectively). LMPN for bagged subprimal cuts (2.72 +/- 0.07) were more than 1 log(10) cm(-2) higher than for the single evaluation at the end of the chill floor. For LMPN and LCOL there was a significant effect of sampling date (P < 0.05) but no effect of conveyer line. LMPN were significantly lower (P < 0.05) from 'cut' than from 'not-cut' surfaces of bagged beef cuts. This difference was less than one log(10) cm(-2) and therefore, of questionable practical significance. There was no effect (P > 0.70) of 'cut' or 'not-cut' surfaces on LEC which had an overall mean level of 0.39. The associations among these variables was strongest for LCOL and LEC (r = 0.81, P < 0.001) and weak for the others. Listeria monocytogenes was isolated in four of 31 pooled samples (90 samples of bagged beef cuts, each pool two to three samples, 50-75 cm(2)) and verotoxigenic E. coli in five of these pools No E. coli O157:H7 was isolated.