Sunteți pe pagina 1din 11

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225744225

Validation of Innovative Food Microbiological Methods According to the EN


ISO 16140 Standard

Article  in  Food Analytical Methods · June 2011


DOI: 10.1007/s12161-010-9154-4

CITATIONS READS

8 4,177

2 authors:

Bertrand Lombard Alexandre Leclercq


Agence Nationale de Sécurité Sanitaire de l'Alimentation, de l'Environnement et d… Institut Pasteur
53 PUBLICATIONS   930 CITATIONS    119 PUBLICATIONS   2,122 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Surveillance of Listeria View project

Taxonomy of Listeria genus View project

All content following this page was uploaded by Bertrand Lombard on 16 May 2014.

The user has requested enhancement of the downloaded file.


Food Anal. Methods
DOI 10.1007/s12161-010-9154-4

Validation of Innovative Food Microbiological Methods


According to the EN ISO 16140 Standard
Bertrand Lombard & Alexandre Leclercq

Received: 7 April 2010 / Accepted: 27 May 2010


# Springer Science+Business Media, LLC 2010

Abstract Reference methods in food microbiology are Keywords Food Microbiology . Method Validation .
mostly based on conventional microbiology. Innovative Innovative Methods . Alternative Methods
methods have been developed and commercialised, with
several advantages, but their users need guarantees on their
performance. Validation schemes have thus been estab- Introduction
lished to assess whether these methods perform at least as
well as the corresponding reference methods. In addition, a Reference methods in food microbiology, for the analysis of
European and International Standard, EN ISO 16140, has food, animal feeding stuffs, samples from primary produc-
been developed to provide a common reference protocol for tion and food processing and handling, are standardised at
the validation of alternative methods, as well as to the international level by International Organization for
determine general principles for their possible subsequent Standardization (ISO) and at the European level by
certification. The content of this standard is summarised, a European Committee for Standardization (CEN) (Lombard
technical protocol is defined separately for qualitative and et al. 1996; Lombard 2004; Leclercq 2002; Leclercq et al.
quantitative methods and the validation study, which 1999; Lombard and Leclercq 2010). The standardisation
comprises two phases, a comparative study and an inter- process consists of establishing reference documents in
laboratory study, is described. This standard is currently agreement with all parties and countries concerned. These
being revised and the main directions taken are presented. reference methods are mostly based on conventional
This paper then briefly introduces the validation/certifica- microbiology, enabling the major characteristic of a
tion systems before addressing several aspects related to the standard to be met: the largest consensus and the broadest
links with laboratory accreditation and European regula- applicability in food microbiology worldwide, including
tions on microbiological criteria, as well as the use of developed countries themselves. Application of reference
alternative methods. methods requires skilled staff and is labour-intensive and
time-consuming, due to the time needed by microorganisms
to grow in or on culture media.
Innovative methods in food microbiology, so called
B. Lombard (*)
AFSSA-LERQAP (French Food Safety Agency—Laboratory ‘rapid methods’, have been developed to achieve results
for Study and Research on Food Quality and Food Processing), within a shorter time and sometimes with higher perfor-
23 avenue du Général de Gaulle, mance than ‘classical’ methods based on conventional
94706 Maisons-Alfort, France
microbiology. They can be classified into three main types:
e-mail: b.lombard@afssa.fr
(1) solid culture media other than those prescribed in the
A. Leclercq standard reference methods, mostly based on chromogenic
WHO Collaborative Center for Foodborne Listeriosis, substrates so as to enable easier visual recognition of the
National Reference Center for Listeria, Institut Pasteur,
target bacteria on the plates, (2) immunoenzymatic kits,
28 rue du Docteur Roux,
75724 Paris cedex 15, France such as ELISA, and (3) molecular-based methods, such as
e-mail: alexlec@pasteur.fr hybridisation probes or PCR, either conventional or real-
Food Anal. Methods

time. They are often based on proprietary components (e.g. easy to carry out and/or automation, analytical perform-
the chromogenic substrates, the ELISA antibodies, the ances at least as good as those of the reference method. In
DNA probes or DNA primers). Their main advantage is other words, an alternative method is designed and intended
to achieve quicker, more specific and sensitive results than to be used instead of the corresponding reference method.
with the conventional methods. They are also designed to The term ‘alternative method’ refers to the combination of
be more convenient to use, requiring less labour, in product, equipment and complete operating procedure,
particular when they are automated. Automation also from sample preparation to final result. It includes all the
enables large sets of samples to be analysed at the same elements necessary for implementing the alternative meth-
time. They are also better suited to hazard control in the od: equipment, culture media, software for analysis of
food chain (HACCP systems), to the control of food results, etc.
products with short shelf life, to the demand of legal In general, validation of a method consists of confirm-
experts or insurance companies in the event of litigation ing, by examination and provision of objective evidence,
and finally to the management of food crises (Leclercq et that the particular requirements for a specific intended use
al. 1999). are fulfilled (Leclercq 2002). For a specific intended use, a
Some new reference methods under development, such laboratory has to reply to the question: why am I analysing
as detection of verotoxigenic Escherichia coli, or some the samples and for what purpose are the resulting data to
recently revised CEN/ISO reference methods, such as on be used? For objective evidence, laboratories need to have
the detection and enumeration of Listeria monocytogenes or refer to validation studies. For confirmation, laboratories
(EN ISO 11290-1 and 2), use innovative products (here, have to compare information obtained from validation
respectively, real-time PCR and a chromogenic medium) studies with customer's requirements.
when conventional microbiology alone is unable to provide In order to have a common reference protocol to conduct
satisfactory solution for analysing the given target: conven- the validation of these innovative methods, CEN has
tional enrichment/isolation procedures perform poorly or do developed a standard on the validation of alternative
not allow identification of the targeted microorganisms. An methods in food microbiology, EN ISO 16140 (Anonymous
essential requirement for the standardisation of these 2003), a standard adopted in parallel by ISO. According to
products is that they should be ‘open’ in their description the standard on general requirements and guidance for
(composition of culture media, sequence of the DNA microbiological examinations (EN ISO 7218 (Anonymous
probes and primers, etc.), so as to avoid giving exclusivity 2007) Clause 14.2), the validation of alternative methods
in the standard to one manufacturer. But the main drawback shall be performed according to EN ISO 16140.
in standardising innovative methods is the slow stand- This paper will mainly present the protocol for the
ardisation process, especially at the European and interna- validation of alternative methods according to the EN ISO
tional levels, which takes at least 3 years, a pace which is 16140 Standard, then will give the revision status of this
not easily compatible with the development of new standard and introduce validation/certification systems, as
technologies. well as their relation with laboratory accreditation and
Given this situation, systems for the validation of regulatory aspects, and the use of validated methods.
innovative methods against reference methods were estab-
lished in the 1990s, to enable quicker recognition of these
alternative methods and at the same time to protect the Protocol for the Validation of Alternative Methods
know-how of the test kit manufacturers (the method can be According to the EN ISO 16140 Standard
validated as a ‘black box’) (Lombard et al. 1996). In
addition, these validation schemes, operated by third-party Origin
organisations independent from the manufacturers, meet the
needs of the users of these methods (food industry, private The EN ISO 16140 Standard originated from the ‘Micro-
and public laboratories, public authorities, accreditation Val’ project of the European Eureka R&D programme
bodies) who require guarantees on their performance. Such (Rentenaar 1996). This project, which ran from 1993 to
innovative methods are proprietary in the sense that they 1998, aimed at setting up a European protocol for the
are given a trade name, sold by their manufacturer or validation of alternative (proprietary) methods in food
distributors and are usually protected by licences. They are microbiology. It included 20 partners: test kit manufac-
often called ‘alternative methods’. An alternative method turers, food manufacturers, food testing laboratories and
can be defined as a method which analyses, for a given standardisation/certification bodies (AFNOR and NEN,
scope, the same target analyte as that measured by the respectively, the French and Dutch standardisation bodies,
corresponding reference method and which meets the the latter having managed the project). Based on an
following requirements: rapid analysis and/or response, experimental phase, it developed two reference documents,
Food Anal. Methods

the first one defining general rules and a certification this definition, the word ‘comparable’ can be understood as
scheme, the second one technical rules. ‘at least equivalent’. A note then explains that the word
‘comparable’ is defined later in the standard in a technical
Standardisation Process protocol adapted to each type of method (qualitative or
quantitative).
Based on the MicroVal technical protocol, CEN/TC 275/ In the first part of the standard, the principles of the
WG 61, the group in charge of European standardisation in validation protocol are established. It comprises two
food microbiology, decided to develop a standard on the phases: (1) a comparative study of the alternative method
validation of alternative methods in food microbiology. For (AM) with the corresponding reference method (RM),
this purpose, in 1996, it established a group, Task Group 2 conducted by one laboratory, and (2) an inter-laboratory
‘Validation of alternative methods’ of WG 6, chaired by study on both methods in parallel, organised by the
Roy HOLBROOK (UK) and comprising 25 members from laboratory in charge of the first phase. The standard defines
MicroVal and WG 6. After a rather long process (7 years), technical rules for these two phases, according to each type
due to the complexity of the topic which was being of method, either qualitative or quantitative.
standardised for the first time in food microbiology, the If certification of methods is sought in addition to their
EN ISO 16140 Standard was published in 2003. It has been technical validation, general principles for this certification are
adopted in parallel as an ISO Standard by ISO/TC 34/SC 92, given. We may recall at first a definition of certification,
the group in charge of international standardisation in food derived from ISO/IEC 17000 (Anonymous 2004) and adapted
microbiology. to our context: it is a process through which a recognised
body acting independently, the certification body, with no ties
Presentation of EN ISO 16140 to the parties involved, gives written assurance that a product
(here test kits) meets the baseline requirements set out in a
Scope reference standard (here the Standard EN ISO 16140). The
certification principles of proprietary methods set out in EN
The standard's objectives are to define: ISO 16140 consist of (1) requirements for the quality
management system of the test kit's production line, based
(a) The general principles and technical protocol for the
on EN ISO 9001 (Anonymous 2008), and (2) regular audits
validation of alternative (mainly proprietary, innova-
of the test kit manufacturer's quality system and of the test
tive) methods, with the idea of allowing the use of
kit's production control requirements. Details on how certifi-
alternative methods instead of the corresponding
cation is organised (management of the two phases of the
reference methods, for the reasons given in the
validation protocol, all the different partners involved includ-
introduction
ing the expert laboratory, the reviewers) are provided by the
(b) The general principles for third-party certification of
certification bodies in their certification rules (see for example
alternative methods
the certification rules of AFNOR Validation (http://www.
It applies to the microbiological analysis of any food afnor-validation.com/afnor-validation-food-industry/food-
intended for human and animal consumption, as well as to industry.html; MicroVal, http://www.microval.org/rules.html).
the analysis of samples from primary production (the The standard then details the validation technical rules,
breeding environment) and from the food production and which consist of the selection of a set of analytical
handling environment. performance criteria. For each criterion, the standard gives
a definition, an experimental design (type/number of
Overview samples to be tested), calculations and an interpretation.
If the AM has already been validated prior to the
As for any standard, a clause of definitions can be found, implementation of the standard, previous results may be
including what can be regarded as the main one, ‘validation taken into account if they have been obtained according to
of an alternative method’. According to the standard, this the same experimental design. Annex A of the standard
consists in ‘demonstrating that results obtained with the gives specific rules for the acceptance of such external
alternative method are comparable to the results which results, dealing with two scenarios:
would have been obtained with the reference method’. In
(a) The method has already been validated against a CEN/
ISO Standard RM. If the method has been modified
1
Working Group 6 “Microbiology” of Technical Committee 275
since its initial validation, the degree of change to the
‘Food analysis—Horizontal methods’.
2
Sub-Committee 9 “Microbiology” of the ISO Technical Committee method needs to be assessed, and if minor, the results
34 ‘Food Products’. can be accepted.
Food Anal. Methods

(b) The method has already been validated against a RM of positive results for both methods divided by the number
which is not a CEN/ISO Standard. The results can be of positive results for the RM. Confidence intervals can be
accepted if the study was performed by a recognised calculated, as described in Annex E of the standard. The
validation system (see below) and if the differences degree of difference between the two methods is assessed
between the RM used and the corresponding CEN/ISO by a χ2 test (McNemar) or by comparison to a binomial
Standard are judged by the validation/certification law, as detailed in Annex F of the standard. Only these tests
body to be minor. in the standard enable the equivalence of qualitative
methods to be assessed.
An overview of the technical protocol is now given.
The limit of detection (called ‘relative detection level’ in
the standard) is defined as the smallest number of cultivable
Technical Protocol for Qualitative Methods microorganisms which can be detected on 50% of occa-
sions. It is determined using artificially contaminated
Method Comparison Study samples from one food matrix chosen from each category
tested (see above), preferably at five (minimum three)
The first phase, called the ‘methods comparison study’ in contamination levels, including a negative control, using
the standard (Clause 5.1), enables three different sets of one target microorganism per category and six replicates
criteria to be assessed, trueness, limit of detection and per method. The calculation is quite basic: the limit of
inclusivity/exclusivity, so as to characterise the performance detection is characterised by a range between the two
of the AM in comparison to the RM. contamination levels giving, for the first one, less than 50%
The trueness (called ‘relative accuracy’ in the standard) of positive results (i.e. less than three positives out of six
is defined as the degree of correspondence between the AM replicates) and for the second one more than 50% of
and RM on identical samples. The experimental design positive results (i.e. more than three positives out of six
gives priority to naturally contaminated samples, from a replicates). The ranges for the AM and RM are then
large variety of matrices likely to be contaminated by the compared.
microorganism of interest. If not enough naturally contam- The inclusivity and exclusivity are combined criteria;
inated samples are available, artificial contamination can be they assess the AM's ability to detect the target microor-
performed, according to a protocol defined in Annex C of ganism and its lack of interference with a range of non-
the standard, with the purpose of mimicking natural target microorganisms. Pure strains are tested: at least 50
contamination. If a method is intended to be validated for target strains (except for Salmonella: 30) and 30 non-target
all foods, five food categories are tested (these categories strains. Criteria for selecting the strains are given in Annex
are defined in Annex B of the standard, and consist of the G of the standard. Inoculation methods are also described.
main food families, such as milk and dairy products, meat
and meat products). A method can also be validated for Inter-laboratory Study
specific food categories (less than five). Samples from the
primary production and food production environments are The second phase, the inter-laboratory study (Clause 5.2 of
briefly treated as separate categories. For each category and the standard) aims at assessing the variability of the AM in
each method, 60 test portions are analysed, belonging to an inter-laboratory context (‘transferability’ of the method).
three food types defined in Annex B (they generally The experimental design requires at least ten collabora-
correspond to different processes, such as heat treatment, tive laboratories having obtained results which can be taken
fermentation, etc.). The results are presented in the usual into account for precision and trueness estimation. The
2×2 table for qualitative methods, except that the terms samples are prepared with one matrix, contaminated at
‘false positives’ and ‘false negatives’ have been replaced by three levels (including a negative control). Eight blind
‘positive deviations’ and ‘negative deviations’. It indicates replicates are prepared at each level, to be analysed by both
that the AM has been compared exactly to the RM, without the AM and RM. Annex H of the standard provides detailed
attempting to assess whether the results of the RM are true, guidance on the organisation of the inter-laboratory study.
i.e. the positives (negatives) of the AM are true or false Laboratory results are excluded only in the event of
positives (negatives), in the event of deviation from the observed technical problems: samples damaged, use of
RM. Three criteria are calculated: (1) relative accuracy, i.e. inappropriate culture media, transport conditions (time,
the number of pairs of identical results for both methods temperature) not respected, acceptability rules for colony-
divided by the total number of results, (2) relative counting/enumeration values not respected and deviations
specificity, i.e. the number of pairs of negative results for from SOP. These are detailed in Annex K of the standard.
both methods divided by the number of negative results for The trueness criteria (accuracy, sensitivity and specific-
the RM and (3) relative sensitivity, i.e. the number of pairs ity) are calculated for each contamination level, and their
Food Anal. Methods

values are compared to those obtained in the first phase. RM, refers for the experimental design and calculation to
Annex L of the standard gives further optional criteria linearity.
based on the works of Langton et al. (2002), which attempt Inclusivity and exclusivity are defined as for qualitative
to better address the random variability of a qualitative methods; the experimental design is similar except that the
method by introducing parameters (accordance, concor- minimum number of target strains to be tested is set at 30,
dance and concordance odds ratio) equivalent to repeat- and 20 for non-target strains.
ability and reproducibility for quantitative methods.
Inter-laboratory Study
Technical Protocol for Quantitative Methods
The inter-laboratory study (Clause 6.3 of the standard)
Method Comparison Study addresses the inter-laboratory precision and trueness of the
AM against the RM, by estimating the well-known parameters
This first phase of the study (Clause 6.2 of the standard) of repeatability (minimum random variability), reproducibility
enables four sets of criteria to be assessed: linearity and (maximum random variability) and bias (systematic error).
trueness, limit of detection/quantification, sensitivity and The experimental design prescribes at least eight
inclusivity/exclusivity. laboratories having obtained results without outliers. The
Linearity is defined as the ability of the AM, for a given samples are prepared from one matrix, contaminated at four
matrix, to yield results that are in proportion to the amount of levels, including a negative control, and analysed in blind
analyte present in the sample. For trueness, the term ‘relative duplicates by both the AM and RM.
accuracy’ is also used with the same definition as for As for the qualitative methods, results are excluded only
qualitative methods (see above). The standard clarifies that, on the basis of reported technical deviations. The laboratory
for instrumental methods, which are commonly used as AMs results are at first log10 transformed, a common step in
in quantitative microbiology (especially for total flora or quantitative microbiology in order to approach a Normal
hygienic indicators), the calibration curve between the distribution of the data and to stabilise the variance over
response/signal of the instrumental method and the unit of contamination levels (Lombard et al. 2005).
the RM (colony-forming units) is not part of the validation The precision characteristics are then calculated with
process; it should be established in a preliminary stage. robust statistics, based, for the repeatability standard
The experimental design requires a minimum of five deviation on the median of the duplicate standard devia-
contamination levels with priority given to natural contam- tions, and for the reproducibility standard deviation on the
ination. At least duplicates (preferably five to ten replicates) Rousseuw recursive median Sn (Rousseeuw and Croux
are prepared. The rules defined above for qualitative 1993). Robust statistics are less sensitive to outlying data
methods on the categories to be studied also apply here. than parametric statistics; they accommodate with these
Once the results have been obtained, a regression extreme data and make it possible to avoid the use of
analysis is conducted. Two methods for linear regression statistical tests to exclude outliers (Lombard et al. 2005).
are suggested in Annex R of the standard: orthogonal and For each contamination level, the bias (relative accuracy)
least-squares linear regressions. The linearity of the AM of the AM against the RM is estimated by the median of the
against the RM is then assessed (lack-of-fit test), as well as laboratories' differences between duplicate means for the
its trueness (bias in the linear equation). AM and the RM.
The limit of detection (LOD) is defined as the smallest
amount of analyte which can be detected but not quantified
with a high probability, i.e. 95%, whereas the limit of Revision of EN ISO 16140
quantification (LOQ) is the smallest amount of analyte
which can be measured and quantified with defined In 2005, both CEN/TC 275/WG 6 and ISO/TC 34/SC 9
precision and trueness. decided to launch the revision of EN ISO 16140, based on
As for analytical chemistry, the experimental design and its implementation in particular by AFNOR Certification.
calculation for indirect instrumental methods refer to blank The main topics for revision are : (1) to cover more
sample analyses and their standard deviation (sO): LOD= comprehensively the topic of method validation in food
3.3sO and LOQ=10sO. For colony-count or most probable microbiology (not only the case of alternative/proprietary
number techniques, the standard suggests an approach for methods), (2) to improve the statistical approaches, and (3)
LOD derived from the approach for qualitative methods, to include acceptability criteria for the analytical perfor-
but which is not detailed. mance characteristics, most of them only being calculated
The sensitivity, i.e. the AM's ability to detect two in the current version. Acceptability criteria are a require-
(slightly) different amounts of analyte measured by the ment of technical guidelines or rules of accreditation bodies
Food Anal. Methods

in the field of food microbiology, in the context of Clause pating laboratories (less than eight or ten). It is
5.4.5 of the EN ISO 17025 Standard (Anonymous 2005a). envisaged to have several participants in the same
In addition, the scope of EN ISO 16140 was to be enlarged. laboratory/geographical site, with a minimum of five
Indeed, since the publication of the existing version of this different laboratories.
standard, the respective scopes of ISO TC34/SC9 and CEN – For qualitative methods, it is intended to combine the
TC275/WG6 have been enlarged to include the analyses of experimental design for trueness and LOD. A protocol for
food-borne viruses and parasites: the new version of EN artificial contamination better mimicking the natural stress
ISO 16140 should also cover them, and not be restricted to conditions of the targeted bacteria would be introduced:
bacteria, yeasts and moulds. In addition, details would be instead of a spiking technique with evaluation of injury
needed on the category of primary production samples. The efficiency given in a normative annex of the current
validation of PCR methods is also expected to be explicitly standard, a seeding protocol, without injury efficiency
covered. evaluation, would be provided in an informative annex,
Given the CEN/ISO cooperation agreement and in order thus, a less stringent requirement which may introduce
to have an international standard on validation of AMs fully noticeable differences between the technical protocols of
recognised by all certification/validation bodies worldwide, validation/certification systems. It is intended to report the
the leadership for the revision of EN ISO 16140 has been AM results on two forms: non-confirmed (screening) and
transferred to ISO/TC 34/SC 9, which has set up a new confirmed: compared to the existing version, the estima-
group, the WG 3 ‘Method Validation’, chaired by Paul In’t tion of trueness would not be limited to a strict comparison
Veld (Netherlands), who is also the current project leader of the AM and RM. It is envisaged to assess the
for the dormant TAG 2 at the CEN TC275 WG 6 level. performance of the qualitative methods on the basis of
Statistical aspects related to the revision of the standard are two parameters: (1) mainly relative LOD (RLOD, a ratio
covered by another group of ISO/TC 34/SC 9, the WG 2 between the LODs of the AM and RM), at 50% and 95%
‘Statistics’, which is providing recommendations on these levels, with calculations based on a complementary log
aspects to WG 3. −log model (a generalised linear model) developed by
The revised standard will comprise several parts, ISO/TC 34/SC 9/WG 2 (Wilrich and Wilrich 2009) and
covering different aspects of method validation, and at (2) inclusivity/exclusivity, as in the current standard. An
least the following ones: acceptability limit would be defined for RLOD, to assess
equivalence between the AM and RM. RLOD would be
– Part 1 on terminology
a performance characteristic for both the method com-
– Part 2 on validation of proprietary methods
parison study and the inter-laboratory study. The other
– Part 3 on in-house method validation, distinguishing in
parameters of the existing standard (accuracy, sensitivity
particular the case where the method to be validated is
and specificity) are considered less valuable than RLOD
compared or not to a RM, and the case of cultivable/
since their values are dependent upon the contamination
non-cultivable microorganisms (e.g. viruses)
level of the samples used in the experimental design, and
– Part 4 on method verification, i.e. how a laboratory can
they would be maintained for information only. The
verify that it is correctly implementing a validated
RLOD approach is currently suited to unpaired results
method, in particular in the context of laboratory
(results obtained from two different test portions, each
accreditation
analysed by AM and RM), but investigations are under
Because the case of alternative/innovative methods way within ISO/TC 34/SC 9/WG 2 to expand its use to
(called ‘proprietary’ methods in the draft revision of the paired results (results obtained from one test portion, the
standard) involves the most comprehensive validation first step of the analysis—generally pre-enrichment—
process, with an in-house study and an inter-laboratory being common to AM and RM). In case of a successful
study, part 2 will represent the basis for the other parts of outcome of this work, the McNemar test, specific for use
the revised standard. with paired results, would be removed.
Compared to the existing standard presented earlier, part – For quantitative methods, a new approach would be
2 will include the following key modifications: introduced for the performance characteristics: the
accuracy profile, already implemented in the pharma-
– The table defining the classification according to food ceutical field (Hubert et al. 2003, 2006a, b), which
categories and food types (Annex B of the current jointly interprets precision and trueness. It avoids one
standard) will be reviewed and restructured. of the drawbacks of the usual approach to validation of
– The experimental design for the inter-laboratory study the AM against a RM: the test for bias significance
for both qualitative and quantitative methods would may conclude, for a given bias value, an acceptable
allow the possibility of a reduced number of partici- bias with an AM having poor precision, and a
Food Anal. Methods

significant bias with an AM having good precision. Official Methods scheme of AOAC International, Perfor-
The accuracy profile approach would replace the mance Tested Methods scheme of AOAC Research Institute
current approach to linearity/trueness based on regres- in the USA and NordVal of NMKL in the European Nordic
sion analysis, keeping only a graphical representation. countries. The Official Methods scheme and NordVal
The LOQ would be maintained, but with a new follow a technical protocol similar but not identical to that
estimation, directly derived from the graphical repre- of EN ISO 16140. In contrast, the technical protocol of the
sentation of the accuracy profile. Performance Tested Methods scheme does not include an
inter-laboratory study. In addition, the organisations
For any calculations, WG 2 is to provide calculation
concerned do not operate as certification bodies, lacking
tools which could be easily implemented by microbiolog-
in particular the control of the manufacturer's quality
ical laboratories, without the need to ask a statistician to
assurance.
perform them. An added value would be to improve the
In the future, official reference laboratories, such as
standardisation of the data analysis. This goal should be
Community or National Reference Laboratories in Europe,
met by developing Excel™ spread-sheets or by providing
may become new actors in the field of validation of AMs,
calculation tools on the ISO website.
depending on possible new regulatory duties given by their
A draft for part 1 and part 2 was circulated in July 2009
prescribers, the Directorate-General for Health and Con-
to ISO/TC 34/SC 9 and CEN/TC 275/WG 6 for an initial
sumers (DG SANCO) of the European Commission, or
vote to include these two parts in the official work
national competent authorities in charge of official food
programme. The outcome was positive, and comments
control.
received have been considered by ISO/TC 34/SC 9/WG 3
Another point is that certification bodies contract with
to prepare a new draft to be submitted to the next vote (ISO
expert laboratories to conduct the two parts of the
Committee Draft vote) in the course of 2010.
validation study described in EN ISO 16140: the methods
comparison study and the inter-laboratory study. The two
certification schemes (AFNOR Validation and MicroVal),
Validation/Certification Systems Based on EN ISO
as well as NordVal, require that the expert laboratory be
16140
accredited according to EN ISO 17025 for the relevant RM,
whereas this is not the case with AOAC. However, such an
A test kit's certification based on the EN ISO 16140
accreditation does not appear to cover an expert laboratory's
Standard obviously requires that the technical validation
entire field of competence necessary to implement the
follow the entire protocol defined in the standard, providing
experimental designs, calculations and interpretation of
the assessment of a kit's performance at a given time. The
the different phases of the validation study. Such a
added benefit of certification is that the validation study is
competence includes (1) the organisation of the inter-
performed and evaluated by a certification body, a third
laboratory study, which is particularly difficult in food
party independent from the test kit's manufacturer. It also
microbiology, given the living nature of the bacteria, and
ensures that the method's performance remains constant
(2) the quality assurance of the expert laboratory's strain
over time, with quality assurance requirements and audits
collection (traceability, characterisation), a key element to
on the production site. The organisational aspects of
conduct inclusivity and exclusivity testing, as well as to
certification are defined by the certification body itself
perform artificial contamination. It is true that a single
and described in documents called certification rules.
reference document for validation of AMs, the EN ISO
Today, only two certification systems are known to apply
16140, is crucial, but the guarantee of its correct application
EN ISO 16140 in full:
with reliable results is no less important. It would require in
– AFNOR Validation scheme managed by AFNOR the future the development of the accreditation of expert
Certification, which has long experience in test kit laboratories for the application of EN ISO 16140, in the
certification in food microbiology (since 1989) and has same way that proficiency testing (PT) scheme providers
certified to date around 80 methods (http://www.afnor- can be accredited according to EN ISO 17043 (Anonymous
validation.com/afnor-validation/afnor-validation.html) 2010) to organise these PT schemes.
– MicroVal, managed by NEN, currently comprising one
certification body, Lloyd's Register QA (http://www.
microval.org), and derived from the MicroVal project Validation of Confirmation/Identification Step
(see above)
The EN ISO 16140 Standard is intended to be used to
Other systems to validate innovative methods in food validate a complete method (from sample preparation to
microbiology can be found, the main ones being the final result) but not to validate a separate step of an AM,
Food Anal. Methods

such as the identification or confirmation of the targeted product conformity with these criteria. When a regulatory
bacteria by biochemical galleries, nucleic probes, PCR 16 S text defines a microbiological criterion, it should also
DNA and other molecular biology methods. To our define the RMs and AMs to be used.
knowledge, no protocol is currently available to validate As regards the situation in Europe, Commission Regu-
an identification/confirmation step, and we consider that the lation (EC) 2073/2005 (Anonymous 2005b) gathers and
development of such a protocol, with the subsequent updates microbiological criteria for food in the European
validation/certification of commercial kits, is necessary. Union (EU). This regulation is formally applicable to food
According to the EN ISO 7218 Standard (Anonymous business operators, along the whole food chain from
2007), dealing with the possible use of biochemical production to distribution, but it is also indirectly applicable
galleries (Clause 12.4) or nucleic probes (Clause 12.5) as to official controls performed by public authorities of EU
alternatives to the biochemical confirmation tests described Member States. It defines, for each microbiological
in standard RMs, an evaluation file established by the criterion, a RM, generally an EN ISO Standard, but states
manufacturer, based on studies published in the interna- in an article of the main text (Article 5) that other methods
tional scientific literature, or the recommendations of a can be applied, under the following conditions: ‘when the
reference laboratory for this microorganism is sufficient, methods are validated against the Reference Method in
but guidance on the content of this validation file is still Annex I and if a proprietary method, certified by a third party
lacking. in accordance with the protocol set out in EN ISO standard
16140 or other internationally accepted similar protocols’.
From this text, it is clear that, if an AM has been validated
Relations Between EN ISO 16140 and Laboratory according to EN ISO 16140 or if a proprietary method has
Accreditation been validated and certified according to this standard, in both
cases in comparison to the corresponding RM cited in the
According to the EN ISO 17025 (Anonymous 2005a) regulation (mostly CEN/ISO Standards), it can be used under
Standard (Clause 5.4.5.1 on laboratory quality manage- this regulation for the operator's own checks and for official
ment), accredited laboratories shall validate the non- controls, so as to assess whether food products meet the
standardised methods they use, such as test kits. If an AM microbiological criteria defined in the regulation.
has been validated/certified according to EN ISO 16140, it Risk managers should be aware that, if an innovative
is obvious that the EN ISO 17025 requirement has been method has been validated, it offers at least the same
met and the laboratory does not need to conduct any more performance as the corresponding RM, and it may even
validation itself when using this method. Method validation perform better. A difference in performance between the
can be considered as one of the three pillars of quality two methods may be a source of litigation between two
assurance of test results, together with proficiency testing countries. For example, a product testing positive for
and reference materials. Salmonella with a PCR method by the importing country
The EN ISO 17025 Standard (Clause 5.4.2) also requires may be rejected, whereas it may have been tested negative
that laboratories verify that the methods they use, once by the exporting country using the EN ISO 6579 reference
validated, are correctly implemented in the particular culture method.
context of their laboratory, and in particular that pre-set
performance criteria are met. The new part of EN ISO
16140 on method verification, under development, will Relations Between EN ISO 16140, Its Revision and New
allow us to have a common reference document to define in Standards
food microbiology the notion of method verification, with
the purpose to harmonise the practices and requirements of The rapid development of techniques, such as molecular
accreditation bodies. methods, and their proprietary nature will make it difficult
in a growing number of cases to fully describe the methods
in ISO/CEN Standards. With the current revision of EN
Relations Between EN ISO 16140 and European ISO 16140, a new type of standard could be developed in
Regulation 2073/2005 on Microbiological Criteria food microbiology, which is already the case for food
chemistry: standards composed, apart from a precise scope,
According to Codex Alimentarius (Anonymous 1997), a of analytical performance criteria. In this case, any
microbiological criterion shall be based in particular on a innovative methods could be applied if they met the scope
detection or quantification method which is relatively easy and performance criteria described in these standards. For
to implement but also reliable and validated. Microbiolog- example, a standard for the detection of staphylococcal
ical criteria are strictly linked to the methods used to check enterotoxins is being developed according to this new
Food Anal. Methods

approach, which appears to be suited to the detection step innovative method has been validated, the certification
which currently needs to rely upon proprietary ELISA kits: body issues a certificate. Laboratories using such validated
setting performance criteria will avoid the need to refer to methods often do not read these certificates carefully and
specific kits in the standard. think that AMs are systematically fully validated; this is not
Another impact of the EN ISO 16140 revision on always the case. For example, the certificate includes the
European and International Standards is the preparation, validation scope which underlines the food matrices tested:
by the same WG 3 of ISO/TC 34/SC 9, of a new CEN ISO not all food matrices are covered and the method may
publication on the validation of RMs. To achieve full perform differently with specific matrices than with the one
recognition at worldwide level and to better play its role of reported in the validation study. Another aspect is the
anchor method for the validation of AMs, a RM should be possible restriction of use of the AM, to be clearly
validated during its development process and before its stipulated on the certificate, for example, stating that the
approval. A RM should be also validated if used outside its validated method only detects motile Salmonella, and not
intended scope or if technically amplified or modified, with all Salmonella spp. as in the EN ISO 6579 Standard RM.
major changes. It is intended that the new CEN ISO Validation certificates also describe performance criteria,
publication will define the technical requirements (type of accompanied by comments which should be read carefully:
studies and performance criteria) to standardise a new RM some limitations or values of performance criteria may be
or to revise an existing standard method. Validation of RMs acceptable for the technical board of the certification body
is a big challenge, since these methods are of common but not for customers of the laboratory. For example, in the
interest to all countries, but their validation requires exclusivity criterion, some strains (species or serotypes or
financial support at either national or regional levels. types) may not be detected by the AM, but this is not
stipulated in the restrictions of use.
When a food microbiology laboratory, especially if
Different Users of EN ISO 16140 accredited according to EN ISO 17025 (Anonymous
2005a), conducts a contract review with its customers, it
The main users of EN ISO 16140 (Part 2 for its revision) should mention all these restrictions. In this case, the
are the validation/certification bodies and their partners, in performance of a method matches the criteria agreed
particular the expert laboratories, which conduct validation between the analyst and the end-user of the data (Thompson
of test kits at the request of kit manufacturers. The second et al. 2002). Moreover, EN ISO 17025 requires that the range
type of user is expected to be public authorities who are to and accuracy of the values obtained from a validated method
refer to this standard in order to accept AMs, which could be relevant to the customer's needs.
be used for official controls. Regarding the other parts of A laboratory's choice of an AM should not only be based
EN ISO 16140 under revision, users would be (1) on its technical validation. Other aspects should be taken
laboratory networks, such as in Europe the networks of into account: (1) logistic and laboratory aspects (sample
National Reference Laboratories, each coordinated by a preparation, staff competence, availability and safety of
European Union Reference Laboratory for different bacteria reagents, time to achieve the final result, productivity,
or viruses, or (2) individual laboratories, which have automation, laboratory space needed, quality assurance,
developed their own methods and intend to validate them verification and metrology of the equipment and technical
against the corresponding RMs, or verify their correct assistance provided by the manufacturer), (2) financial
implementation in the specific context of the laboratory, for aspects (purchase cost of the equipment and/or reagents)
accreditation purposes. Depending on the method and its and (3) strategic aspects (recognition by customers, distrib-
scope, the estimated cost of implementing EN ISO 16140 is utors, public authorities and accreditation bodies).
expected to lie between 50,000 and 150,000 Euros for a test Once the proprietary method has been selected, the
kit manufacturer in the context of a validation/certification laboratory needs to verify the correct implementation of this
scheme. validated method: this is a requirement for accredited
laboratories and for any method used, according to Clause
5.4.2. of EN ISO 17025. Verification is the confirmation,
Choice and Use of Validated/Certified Commercial through the provision of objective evidence, that specified
Methods requirements have been fulfilled (Anonymous 2008). It
differs from validation since objective evidence is obtained
According to Thompson et al. (2002), to validate a method not to fulfil the intended use but the specified requirements.
is to investigate whether the analytical purpose of the It is an additional procedure that confirms that the
method is achieved, that is the acquisition of analytical laboratory has achieved the performance characteristics of
results with an acceptable uncertainty level. Once an the method. Given the lack of harmonised rules on method
Food Anal. Methods

verification, it has been decided that a separate part of the ries in its implementation and allowing a harmonised
revised EN ISO 16140 would be dedicated to this topic (see assessment by accreditation bodies.
above). To further ensure the reliability of results obtained with
Finally, during the use of a given method, some method these innovative methods, ISO and CEN should develop
parameters may have to be changed or adjusted if the general requirements and guidance to use these methods in
values of the performance characteristics appear to fall food microbiology laboratories, such as the EN ISO 22174
outside their acceptance limits. The question is whether Standard (Anonymous 2005c) for PCR.
such changes require revalidation. For accredited laborato-
ries, according to Clause 5.4.5.2 Note 3 of EN ISO 17025,
when changes are made to validated non-standard methods, References
the influence of such changes should be documented and, if
appropriate, a new validation should be carried out. In order
Anonymous (1997) CAC/GL 21 Principles for the establishment and
to deal with this question, the laboratory should define an application of microbiological criteria for foods, Codex Alimen-
operating range for each method, either based on its tarius, Rome
experience with similar methods or investigated during Anonymous (2003) EN ISO 16140: Microbiology of food and animal
feeding stuffs—Protocol for the validation of alternative meth-
method development. This range should be verified during ods, ISO, Geneva
method validation in trueness studies and should be part of Anonymous (2004) ISO/IEC 17000: Conformity assessment—Vocab-
the method characteristics. Revalidation is also required if ulary and general principles, ISO, Geneva
the scope of the method has been changed or extended. Anonymous (2005a) EN ISO 17025: General requirements for the
competence of testing and calibration laboratories, ISO, Geneva
This question will be covered in the forthcoming part 2 of
Anonymous (2005b) Commission Regulation (EC) 2073/2005 of 15
the revised EN ISO 16140. November 2005 on microbiological criteria for foodstuffs.
Official Journal of the European Union, L338/1-26
Anonymous (2005c) EN ISO 22174: Microbiology of food and
animal feeding stuffs—Polymerase chain reaction (PCR) for the
Conclusion
detection of food-borne pathogens—General requirements and
definitions, ISO, Geneva
In numerous cases, laboratories performing routine food Anonymous (2007) EN ISO 7218: Microbiology of food and animal
microbiological analyses do not use the RMs based on feeding stuffs—General requirements and guidance for microbi-
ological examinations, ISO, Geneva
conventional microbiology but increasingly often choose
Anonymous (2008) EN ISO 9001: Quality management systems—
proprietary methods, based on innovative technologies, Requirements, ISO, Geneva
which provide results in a shorter amount of time, are Anonymous (2010) EN ISO 17043: Conformity assessment—General
easier to perform, require fewer resources in terms of requirements for proficiency testing, ISO, Geneva
Hubert Ph et al (2003) STP Pharma Pratiques 13:3
skilled staff, and are more automated. The validation of
Hubert Ph et al (2006a) STP Pharma Pratiques 16:1
these proprietary methods according to EN ISO 16140, and Hubert Ph et al (2006b) STP Pharma Pratiques 16:2
their subsequent certification, give assurance to the labora- Langton SL, Chevennement R, Nagelkerke N, and Lombard B (2002)
tories and their customers (in the food industry, food International Journal of Food Microbiology 79
Leclercq A (2002) Accredit Qual Assur 7:299
distribution, public control authorities) that the results
Leclercq A, Lombard B, Mossel DAA (1999) Sciences des Aliments
obtained with these methods are valid and at least as good, 20
if not better, than those obtained with the corresponding Lombard B (2004) ISO Focus 1:8
RMs, the ‘gold standards’. Lombard B, and Leclercq A (2010) The role of standardization bodies
in the harmonization of analytical methods in Food Microbiol-
In certain cases, laboratories may also use in-house
ogy. In: Global Issues in Food Science and Technology, IUFoST,
methods that they have developed. The revised EN ISO Elsevier ed (in press)
16140 will provide a tool to the laboratories to validate Lombard B, Gomy C, Catteau M (1996) Food Control 7:1
these methods and to give confidence in their reliability, Lombard B, Feinberg MH, Lahellec C (2005) J AOAC Int 88:3
Rentenaar IMF (1996) Food Control 17:1
both to their customers and to accreditation bodies.
Rousseeuw PJ, Croux C (1993) J Am Stat Assoc 88:424
The revised EN ISO 16140 will also define what method Thompson M, Ellison SLR, Wood R (2002) Pure Appl Chem 74:835
verification in food microbiology is, guiding the laborato- Wilrich C, Wilrich PT (2009) J AOAC Int 92:6

View publication stats

S-ar putea să vă placă și