Food microbiology proficiency testing unsing naturally contaminated matrices.

Since 1991, the REQUASUD network has been organizing food microbiology proficiency testing (PT) proposing authentic food matrices with or without artificial inoculation. In parallel, sterile food matrices and reference materials, artificially spiked at comparable levels, have been proposed as controls. ISO 22117 (Specific requirements and guidelines for food microbiology proficiency testing) specifies that the PT items should be as close as possible to routine samples, in order to accurately reflect the potential bias linked to the matrix. Most food microbiology PT organizers only provide sterile items (e.g. reference materials or irradiated samples), artificially inoculated. Although this protocol enables to produce high amounts of matrix, with a multi-parameters inoculation, a precise assigned value and a satisfying stability of the contamination levels over time, the artificial samples produced may not adequately represent the true challenges of natural matrices. Naturally contaminated food matrices raise huge challenges for the PT organizers in terms of standardization, processing and stabilization, but they are much closer to the laboratories real-life conditions. The comparison of REQUASUD PT results obtained on naturally and artificially contaminated PT samples enabled to highlight issues linked to the analysis of “authentic” food PT items. The intrinsic properties of those samples, like the presence of a background flora, are necessary technical burden to accurately evaluate the analytical performances of the laboratories in routine. The exclusive use of reference materials or sterile spiked matrices as PT samples might lead to a systematic underestimation of the inter-laboratory standard deviation and an inaccurate evaluation of the analytical performances of the participants