Page 78 - 2013 ISPAM Mid-Year Mtg Book

Therefore verification is divided into two steps:
- Demonstration of the competence of the laboratory to perform the method in
accordance with the parameters established during the validation for a defined matrix.
Use of at least? one relevant matrix is required.
- Verification of the method on other matrices. The types of matrices to be tested
depend on the diversity of the matrices tested routinely by the laboratory. The
selected matrices shall be representative of the types tested by the laboratory and
shall also be relevant to the target microorganisms of the method. However, for
laboratories testing a broad range of products (foods, feeds, environmental and
primary production stage samples) this approach is not practicable. In such cases,
the laboratory shall verify a selection of relevant matrices with different properties.
The number of matrices to be tested must be defined.
7 Protocol for verification of qualitative methods
7.1 General
The relevant performance characteristic for verification of a qualitative microbiological
method is the level of detection. The level of detection is “the lowest concentration of
microorganisms that may be determined with a given probability under the experimental
conditions of the method under evaluation” (see ISO 16140-1).
The theoretical level of detection for microbiological methods is 1 CFU per test portion.
However, in practice, the level of detection may be affected by the sample matrix. For
example, in the presence of high numbers of competitive flora it is not always possible to
detect a single and often sub-lethally damaged cell.
For microbiological methods the level of detection is expressed as LOD 50.
NOTE For current ISO standard qualitative methods there is no validation data for LOD 50.
For alternative (proprietary) methods, evaluation of the level of detection shall be performed
by calculating the relative level of detection (RLOD) which is defined as the ratio of the LOD
50 of the alternative method and the reference method.
Use of RLOD will be discussed with WG2.
7.2 Determination of LOD 50
In this study, the laboratory shall is encouraged use Certified Reference Material (CRM) with
a concentration of 1-5 CFU per test portion. The number of samples for each matrix is 10
with a recommended acceptance criterion of 100 %.
When no CRM is available, the laboratory should produce low level inocula as described in
Annex A (to develop).
8 Protocol for verification of quantitative methods
8.1 General
Comment [M1]: Paul will provide altrenate
language or clarification.