Performance characteristics and estimation of measurement uncertainty of three plating procedures for Campylobacter enumeration in chicken meat

In this work, we present an intra-laboratory study in order to estimate repeatability (r), reproducibility (R), and measurement uncertainty (U) associated with three media for Campylobacter enumeration, named, modified charcoal cefoperazone deoxycholate agar (mCCDA); Karmali agar; and CampyFood ID agar (CFA) a medium by Biomerieux (R) SA. The study was performed at three levels: (1) pure bacterial cultures, using three Campylobacter strains; (2) artificially contaminated samples from three chicken meat matrixes (total it = 30), whereby samples were spiked using two contamination levels; ca. 103 Cfu Campylobacter1g, and ca. 1 04 Cfu Campylobacterl g; and (3) pilot testing in naturally contaminated chicken meat samples (n = 20). Results from pure culture experiment revealed that enumeration of Campylobacter colonies on Karmali and CFA media was more convenient in comparison with mCCDA using spread and spiral plating techniques. Based on artificially contaminated samples testing, values of repeatability (r) were comparable between the three media, and estimated as 0. 15 log(10) cfu/g for mCCDA, 0. 14 log(10) cfu/g for Karmali, and 0. 18 loglo cfu/g for CFA. As well, reproducibility performance of the three plating media was comparable. General R values which can be used when testing chicken meat samples are; 0.28 log(10), 0.32 log(10), and 0.25 log(10) for plating on mCCDA, Karmali agar, and CFA, respectively. Measurement uncertainty associated with mCCDA, Karmali agar, and CFA using spread plating, for combination of all meat matrixes, were +/- 0.24 log(10) cfu/g, 0.28 loglo cfu/g, and +/- 0.22 log(10) cfu/g, respectively. Higher uncertainty was associated with Karmalli agar for Campylobacter enumeration in artificially inoculated minced meat (0.48 log(10) cfu/g). The general performance of CFA medium was comparable with mCCDA performance at the level of artificially contaminated samples. However, when tested at naturally contaminated samples, non-Canipylobacter colonies gave similar deep red colour as that given by the typical Campylobacter growth on CFA. Such colonies were not easily distinguishable by naked eye. In general, the overall reproducibility, repeatability, and measurement uncertainty estimated by our study indicate that there are no major problems with the precision of the International Organization for Standardization (ISO) 10272-2:2006 protocol for Campylobacter enumeration using mCCDA medium. (c) 2007 Elsevier Ltd. All rights reserved.