Applications and Use of Criteria and Other Tests

2011
7.1 Introduction
As discussed in Chap. 1, it is widely recognized that application of prerequisite programs at
preharvest,
harvest and postharvest level (e.g., Good Agricultural Practices (GAP), Good Farming
Practices (GFP), Good Veterinary Practices (GVP), Good Hygiene Practices (GHP), Good
Manufacturing Practices (GMP), etc.) and Hazard Analysis Critical Control Point (HACCP) program
is the most effective food safety management strategy. Effective control of undesirable microorganisms

in foods is best addressed at appropriate steps in the food chain through targeted and synergistic
application of these approaches. Microbiological testing of process hygiene can play an important
role in verifying the effectiveness of food safety management programs (prerequisite programs and
HACCP) when used in a thoughtful, well-planned manner. In some cases, microbiological testing of
the end product may also be used if no prior history of the product is available (e.g., at port of entry).
Consistent with previous ICMSF considerations (2002), testing should be required only when the
following two conditions exist:
1. The product group has been implicated in foodborne disease or may have an inadequate shelf life
or other microbiological issues if effective controls are not applied.
2. The application of testing will reduce the health risk or quality issues associated with a food or will
effectively assess adherence to microbiological control measures or process controls.
This chapter provides background on the considerations that the Commission used to propose microbiological
criteria for some commodities and not others. It also indicates how the criteria should be
interpreted and applied.
The recommendations for end product testing in the following chapters replace those given in
Microorganisms in Foods 2: Sampling for Microbiological Analysis: Principles and Specific
Applications (ICMSF 1986). Significant advances in the understanding of food production and processing,
risk management, and the statistics of sampling have made changes necessary. Additionally,
the following chapters provide recommendations not only for end product testing, but also other tests
that may provide useful information for microbiological safety and quality management.
Although considerable effort was given to develop appropriate, risk-based criteria, ICMSF recommendations
have no official status. Promulgation of official national microbiological standards is the
responsibility of governments and articulation of international food safety guidelines is the province
of intergovernmental standards setting bodies such as the Codex Alimentarius Commission, which is
organized under the Food and Agriculture Organization (FAO) of the United Nations and the World
Health Organization (WHO).
7.2 Format of Product Chapters
The product groupings generally follow those used in Microorganisms in Foods 6: Microbial Ecology
of Food Commodities (ICMSF 2005), which provides details on the microbial ecology of specific
commodities and products. The following chapters focus on practical application of testing in the
production of microbiologically safe and wholesome foods rather than on the microbial ecology of the
products. Each chapter briefly discusses the relevant microbial hazards and spoilage concerns for
each sub-commodity and, based on their significance, may recommend tests and criteria for the various
stages of production and distribution, as described below.
7.2.1 Primary Production
For some commodities, such as fruits, vegetables, spices, meat, poultry, and fish products, primary
production practices can have a major influence on the microbiological quality of the product. Where
appropriate and where information is available, recommendations for irrigation or seafood harvest
waters, fertilizer, vaccination programs, feeding regime and other on-farm practices may be provided
or referenced to national standards.
7.2.2 Ingredients
Many foods are composed of a number of different ingredients. The microbiological quality and
safety of some ingredients may be critical to the safety and stability of the final product. Control of
a microbiological concern at the ingredient level may be essential for products when there is no subsequent
kill step (e.g., cocoa powder in chocolate that has no heat treatment, beef intended for production
of unheated fermented salami). For other foods, ingredients may be subjected to a kill step
during processing and therefore microbiological criteria are less important (e.g., cocoa powder in ice
cream mix that is subsequently pasteurized, beef intended for production of cooked meat products).
Anticipated initial levels or criteria for such ingredients discussed in other chapters may be crossreferenced,
as appropriate. Testing is generally recommended for an ingredient if the answer to either
of the following questions is “Yes” for the commodity under consideration:
1. Is control at the ingredient step necessary for safety or quality?
2. Is testing necessary to verify the acceptability of the ingredient?
7.2.3 In-Process
In this book, the term “in-process” testing is used to describe testing to (1) verify a kill step or (2)
monitor whether the product is likely to become contaminated. The concept of HACCP emphasizes
the importance of applying validated and verified process controls for the production of safe food.
Certain tests may be used to verify that processes are performing as anticipated (e.g., initial in-plant
validation to assess the performance of a control measure at certain production step). For example,
testing for indicator organisms such as coliforms or Enterobacteriaceae on in-process product
emerging from cooking equipment may be useful to verify adequacy of the cooking process.
Sampling intermediate product (e.g., from conveyors, filler heads, holding tanks or vats, etc.) and
processing line samples (e.g., process wash water, sifter tailings, fines, line residues, and scrapings)
offers an alternative or complimentary approach to the use of swabs or sponge samples to monitor
for contamination with microorganisms of concern to public health or spoilage. In-process product
7.2 Format of Product Chapters 65
or product residues that accumulate on equipment may represent a worst case when such materials
accumulate under conditions that support microbial growth throughout a production period.
In-process testing may provide more useful information about potential microbiological concerns
than end product testing, particularly when the data are used in a process control system as discussed
in Chap. 3 of this book and in Microorganisms in Foods 7: Microbiological Testing in Food Safety
Management (ICMSF 2002).
In-process testing is generally recommended if the answer to all of the following questions is
“Yes” for the commodity under consideration:
1. Does the process need to be controlled to prevent increase, ensure decrease, maintain current level,
or prevent spread of a microbial concern?
2. Is testing needed to verify (a) the process is functioning as intended or (b) contamination is not
occurring in the process?
3. Are there locations in the process where accumulated product residue may provide a representative
or worst case sample that predicts the safety or quality of the final product?
7.2.4 Processing Environment
Maintenance of a hygienic processing environment is essential for the production of safe and wholesome
food; however, microbiologically relevant considerations will vary for different food commodities.
This section generally addresses the use of swabs or sponges for sampling sites on equipment or
in the environment. This type of testing is very useful and effective for verifying that the environment
is under appropriate hygienic control for the specific commodity. General guidance on environmental
sampling can be found in Chap. 4 of this book and in Microorganisms in Foods 7: Microbiological
Testing in Food Safety Management (ICMSF 2002). As with in-process sampling, a well designed
environmental testing program based on a predetermined clear objective may provide more useful
information about potential microbiological concerns than end product testing, particularly when the
data are used in a process control system as discussed in Chap. 3 of this book and in Microorganisms
in Foods 7 (ICMSF 2002).
Environmental testing is generally recommended with potential tests to consider, if the answer to
the following two questions is “Yes” for the commodity under consideration:
1. Does the environment need to be controlled to prevent contamination of the product with a microbial
concern?
2. Will testing be beneficial to verify control of the microbial concern in the environment?
7.2.5 Shelf Life
The shelf life of food commodities is influenced by deleterious changes to product quality that
develop over time, many of which are nonmicrobial, such as enzymatic activity, oxidation, structural
changes, staleness etc. However, microbial activity can play an important role in the safety or spoilage
of some food commodities. Shelf life testing is discussed only if microbial activity is relevant to
a particular commodity. For certain products (e.g., bulk shipments) shelf life testing may not be feasible.
Shelf life testing is generally recommended if the answer to the following two questions is
“Yes” for the commodity under consideration:
1. Is shelf life limited by a microbiological safety or quality concern?
2. Is shelf life testing feasible?
66 7 Applications and Use of Criteria and Other Tests
7.2.6 End Product
End product criteria are recommended if they can be used to demonstrate the successful application
of food safety controls or to assess the microbiological status of a lot when insufficient information
exists to assess its status. For a limited number of foods, available prerequisite programs and HACCP
may be inadequate to provide consumer protection. For such foods end product testing may be a
necessary step to provide additional protection to consumers.
The determination of the relative importance of end product testing must be made on a product by
product basis (see Sect. 7.2.7), and end product testing may be used for lot acceptance when there is
insufficient access to other process or testing information. The suggested criteria for lot acceptance
are based on baseline data, experience, industry practice, relative risk when ICMSF cases are considered,
or existing microbiological criteria that have been developed internationally as a result of the
risk analysis process established by The Codex Alimentarius Commission (see Sect. 7.4). Other
sampling plans may be appropriate in certain situations. For example, reducing the number of samples
may be acceptable for on-going surveillance activity; whereas increasing the number of samples
may be prudent when investigating significant process deviations or outbreaks. Testing is generally
recommended if the answer to one of the following questions is “Yes” for the commodity under
consideration:
1. Is end product testing necessary to verify control of the overall manufacturing process?
2. Is end product testing relied upon for ensuring the safety or quality of the lot?
7.2.7 Relative Importance of Tests Recommended
Tables within each commodity chapter include a rating (i.e., low, medium, high) for the relative
importance of the tests recommended. These ratings reflect the level of importance for routine testing
during operation under GHP/GMP and HACCP using processes that have been validated to consistently
deliver product that is acceptable from the perspective of safety and quality. In assigning ratings,
the Commission attempted to identify the types of samples that would provide the most useful
information to evaluate the microbiological status of the product being manufactured. It is important
to note that the relative importance of a test must be evaluated in the context of the overall microbiological
testing program. For example, if ingredient, in-process, and environmental monitoring are
routinely conducted in a diligent manner, on a routine basis, in a stable processing environment, with
the intent to use the information for trend analysis and process improvement, then the relative importance
of finished product testing is likely to be low. However, if upstream testing is done occasionally
or in a manner that would not provide confidence that the process is under control, then the relative
importance of finished product sampling may increase.
The relative importance and recommended sampling plans may change in nonroutine situations.
For example, when validating a new process, qualifying a new ingredient or supplier, performing
corrective action for a significant process deviation or investigating a foodborne illness outbreak,
more extensive testing is generally warranted. Previous chapters on corrective action, process validation
and customer–supplier relationships provide guidance in these areas.
7.3 Choice of Microorganisms or Products Thereof
Recommendations for tests are included for microbes or their products (e.g., mycotoxins) that are most
important in respect to hazard/risks or noncompliance with GHP/GMP. This choice is based on a hazard
analysis and risk categorization (i.e., a qualitative risk assessment) that considers epidemiologic
evidence,
public health impact, the scientific literature and expert opinion, in-process experimental
7.4 Selection of Limits and Sampling Plans 67
validation, and recognizes the limitations of current methodologies. Quality issues are also considered
in recommending tests. Detailed discussion of microbial concerns for each commodity is provided in
Microorganisms in Foods 6: Microbial Ecology of Food Commodities (ICMSF 2005).
7.4 Selection of Limits and Sampling Plans
Limits and sampling for in-process and environmental tests are greatly influenced by the site, process,
geographic region and other factors, therefore it is not possible to specify limits that are universally
applicable in all situations. Typical guidance levels or typical levels encountered may be provided for
these tests, but these are not intended to be applied universally. Accordingly, no methods, number of
samples, or sample size are specified in most instances. It is important to emphasize that a typical
level encountered does not indicate a limit. It is expected that levels will periodically exceed a typical
level encountered.
For end product testing, the following questions were asked sequentially to help identify the
appropriate basis for recommended sampling plans and criteria:
1. Does a risk assessment exist?
2. Has an appropriate level of protection (ALOP) been established that would enable determination
of a Food Safety Objective or a Performance Objective?
3. Are sufficient data available to define typical values likely to be encountered that are consistent
with safe food, or food of good quality, and do data exist to estimate the variability in values
encountered, e.g., within and between batches?
The availability of a risk assessment, dose-response data, consumer exposure data and defined ALOP
or FSO/PO, and data on microbial levels typically encountered in a food facilitates development of
microbiological criteria that have a link to public health goals. ICMSF (2002) and van Schothorst
et al. (2009) reviewed this process in some detail. However, the availability of formal risk assessments
for many food types is limited (e.g., qualitative and quantitative). In the absence of a risk
assessment, the Commission used the ICMSF cases (ICMSF 2002), generally accepted international
regulations (e.g., Codex, national regulations, industry guidelines) or expert opinion to recommend
sampling plans and limits.
ICMSF cases, summarized in Table 7.1, consider both the severity of the hazard, the susceptibility
of the intended consumer and the potential for the risk to decrease, remain the same, or increase
between the time of sampling and when the food is consumed. Sampling plans become increasingly
more stringent with increased severity. The following terms are used:
n = the number of sample units to be analyzed
c = the maximum number of sample units allowable with marginal but acceptable results (i.e., between
m and M)
m = concentration separating good quality or safety from marginally acceptable quality
M = concentration separating marginally acceptable quality from unacceptable quality or safety
Limits (m and M) recommended for utility, indicator, and moderate hazards (Cases 1–9) are typically
reported on a per gram basis, and quantitative methods are generally used. The c criterion included
in Cases 1–9 recognizes that statistical variation may occasionally contribute to results above m.
Specifying a maximum limit M helps to prevent acceptance of product that may greatly exceed quality
or safety indicators without further investigation or action.
For serious and severe hazards (Cases 10–15), when c = 0, the maximum acceptable level is m = M.
For Cases 10–15, test results are greatly influenced by sample size because they are typically reported
as being present (positive) or absent (negative) in the sample tested. For this book,
the analytical unit for each sample, n, for Cases 10–15 is 25 g unless otherwise specified. Thus, for
68 7 Applications and Use of Criteria and Other Tests
Table 7.1 Sampling plan stringency (case) in relation to degree of risk and conditions of use
Degree of concern relative
to utility and health hazard Examples
Conditions under which food is expected to be
handled and consumed after sampling in the usual
course of eventsa
Reduce risk No change in risk May increase risk
Utility: General
contamination, reduced
shelf life, incipient
spoilage
Aerobic colony count, yeasts
and molds
Case 1 Case 2 Case 3
n = 5, c = 3 n = 5, c = 2 n = 5, c = 1
Indicator: Low,
indirect hazard
Enterobacteriaceae,
generic E. coli
Case 4 Case 5 Case 6
n = 5, c = 3 n = 5, c = 2 n = 5, c = 1
Moderate hazard: Not
usually life threatening,
usually no sequelae,
normally of short
duration, symptoms
self-limiting, can be
severe discomfort
S. aureus, B. cereus,
C. perfringens,
V. parahaemolyticus
Case 7 Case 8 Case 9
n = 5, c = 2 n = 5, c = 1 n = 10, c = 1
Serious hazard:
Incapacitating but not
usually life threatening,
sequelae are rare,
moderate duration
Salmonella,
L. monocytogenes
Case 10 Case 11 Case 12
n = 5, c = 0 n = 10, c = 0 n = 20, c = 0
Severe hazard: For the
general population or
in foods targeted for
susceptible populations,
causing life threatening
or substantial chronic
sequelae or illness of
long duration
For the general
population,
E. coli O157:H7, C.
botulinum neurotoxin;
for restricted
populations, Salmonella,
Cronobacter spp.;
L. monocytogenes
Case 13 Case 14 Case 15
n = 15, c = 0 n = 30, c = 0 n = 60, c = 0
a More stringent sampling plans would generally be used for sensitive foods destined for susceptible populations
Case 10, n = 5, five individual 25 g samples are analyzed. Statistical considerations underlying the
sampling plans recommended in this book are discussed in Appendix A and explained in greater
detail with examples by van Schothorst et al. (2009), Whiting et al. (2006) and ICMSF (2002).
7.4.1 Comparing ICMSF Cases to Codex Criteria for L. monocytogenes
The following example evaluates the relative stringency of ICMSF cases, which use a qualitative risk
assessment approach for groups of microorganisms, to the Codex Alimentarius Commission criteria
for L. monocytogenes in ready-to-eat (RTE) foods, which was based on quantitative risk
assessments.
7.4.1.1 Stringency of Sampling Plans Using ICMSF Cases
The relative stringency of selected ICMSF cases is compared in Table 7.2, using various hypothetical
values for m and M. The mean concentration that would be correctly rejected with a probability of 95%
is provided using the calculations of van Schothorst et al. (2009). A standard deviation of 0.8 and a log
7.4 Selection of Limits and Sampling Plans 69
normal distribution is assumed. As the severity of hazard increases, the stringency of the cases increases
and the mean concentration that can be reliably detected decreases (from top to bottom).
The mean concentration also decreases as the potential for the hazard increases from left to right.
7.4.1.2 Stringency of Codex L. monocytogenes Criteria
The criteria for L. monocytogenes in RTE food recommended in this book where developed through
the step-wise consensus process within the Codex Alimentarius Committee for Food Hygiene. FAO/
WHO (2004) conducted a risk assessment on L. monocytogenes in RTE foods to address questions
on the risk of serious illness in relation to the level of L. monocytogenes in food for different
susceptible
populations relative to the general population, as well as the risk of serious illness from
L. monocytogenes in foods that support and do not support its growth at specific storage and shelf
life. The risk assessment indicated that the vast majority of listeriosis cases were associated with the
consumption of foods that do not meet current standards for L. monocytogenes (not detected in 25 g
or <100 CFU/g) and that the greatest benefit to public health would be to effect a significant reduction
in the number of servings contaminated with high numbers of L. monocytogenes (FAO/WHO 2004).
Therefore, control measures that reduced the frequency of contamination would be expected to have
a proportional reduction in the rates of illness.
The risk assessment used a worst case scenario, where it was assumed that all servings had the
maximum level being considered, as well as a more realistic approach that allowed for a distribution
of the levels of L. monocytogenes to be considered. Both scenarios demonstrated that as the frequency
or level of contamination increased the risk and the predicted number of cases also increased. It was
assumed that ingestion of a single cell could potentially cause illness. According to the risk assessment
and assuming a fixed serving size, if all RTE foods went from having 1 to 1,000 CFU/serving,
the risk of listeriosis would increase 1,000-fold (see Table 7.3).
In contrast, the risk associated with introducing 10,000 servings of food that were contaminated
with 1,000 L. monocytogenes CFU/g into the food supply would, theoretically be compensated by
removing a single serving contaminated at a level of 107 CFU/g from the food supply. In interpreting
these results and the actual effect of a change in the regulatory limits for L. monocytogenes in RTE
foods, one also has to take into account the extent to which noncompliance with established limits
occurs. Based on data available for the US, where the limit for L. monocytogenes in RTE foods was
Table 7.2 Relative performance of ICMSF cases in terms of the mean concentrations (in bold
text) that will be rejected with at least 95% probability, assuming hypothetical criteria
and a
standard deviation of 0.8
Type and likely change
to level of hazard Reduce No change
May
increase
Indicator, indirect hazard;
m = 1,000/g, M = 10,000/g
Case 4 Case 5 Case 6
n = 5, c = 3 n = 5, c = 2 n = 5, c = 1
5,100 CFU/g 3,300 CFU/g 1,800 CFU/g
Moderate hazard; m = 100/g,
M = 10,000/g
Case 7 Case 8 Case 9
n = 5, c = 2 n = 5, c = 1 n = 10, c = 1
2,600 CFU/g 1,100 CFU/g 330 CFU/g
Serious hazard; m = 0/25 g Case 10 Case 11 Case 12
n = 5, c = 0 n = 10, c = 0 n = 20, c = 0
1 CFU/55 g 1 CFU/100 g 1 CFU/490 g
Severe hazard; m = 0/25 g Case 13 Case 14 Case 15
n = 15, c = 0 n = 30, c = 0 n = 60, c = 0
1 CFU/330 g 1 CFU/850 g 1 CFU/2,000 g
70 7 Applications and Use of Criteria and Other Tests
0.04 CFU/g, the estimated number of cases for listeriosis for that population was 2,130, using the
baseline level in the US Listeria risk assessment (FDA-FSIS 2003). If a level of 0.04 CFU/g was
consistently achieved, one could expect <1 case of listeriosis/year. This, in combination with available
exposure data, suggested that a portion of RTE food in the US contains a substantially greater number
of the pathogen than the limit of 0.04 CFU/g and that the public health impact of
L. monocytogenes is almost exclusively a function of the foods that greatly exceed that limit.
Therefore it could be asked if a less stringent microbiological limit for RTE foods could be beneficial
in terms of public health if it simultaneously fostered the adoption of control measures that resulted
in a substantial decrease in the number of servings that greatly exceeded the established limit. The
results of the risk assessment illustrated that the potential for growth of L. monocytogenes strongly
influences risk, though the extent to which growth occurs depends on the characteristics of the food
and the conditions and duration of refrigerated storage. Using selected RTE foods, their ability to
support the growth of L. monocytogenes appears to increase the risk of listeriosis 100- to 1,000-fold
on a per-serving basis. In order to reflect the difference in relative risk different criteria were developed
depending on whether the product will support the growth (Table 7.4).
The criterion for products that do not support the growth of L. monocytogenes (i.e., 5 samples with
a limit of 102 CFU/g) would reject a lot of food, with a probability of 95%, when the geometric mean
concentration was 80 CFU/g, assuming a standard deviation of 0.8 (see Appendix A). This criterion
reflects the conclusion from the risk assessment that the vast majority of listeriosis cases result from
the consumption of high numbers of L. monocytogenes and also the desire to use a level that helps
promote compliance within the industry. In contrast, the criterion for products that may support
growth is much more stringent. This criterion also uses 5 samples but has a much more stringent limit
of absence in 25 g for each analytical unit. This would be able to reject a lot with a geometric mean
concentration of 1 CFU in 55 g with 95% confidence (assuming a standard deviation of 0.8). It should
be noted that in this example a standard deviation of 0.8 was used to calculate the relative stringency
of the ICMSF cases, whereas a standard deviation of 0.25 was used for calculations in the Codex
Annex (Codex Alimentarius 2009). The effect of using different standard deviation values from 0.25
to 1.2 on the relative performance of different criteria is given in Appendix A. The risk assessment
estimated that products that support growth represent a 100- to 1,000-fold increase in risk per serving.
This relative difference in stringency and also comparison to existing ICMSF cases is illustrated in
the Fig. 7.1. This criterion provides a higher degree of confidence that L. monocytogenes will not be
present in foods that represent the greatest risk from illness and is therefore approximately 1,000
times more stringent than the criterion for products that do not support growth.
In this book, the Codex criteria for L. monocytogenes are used in place of ICMSF cases.
Table 7.3 Relative risk of illness and estimated number of cases/year in the United
States if all RTE meals were contaminated at that level. Relative risk used the risk
from a dose of 1 CFU (FAO/WHO 2004)
Level (CFU/g) Dose (CFU) Relative risk
Estimated number
of cases/year
<0.04 1 1 0.54
0.1 3 2.5 1
1 32 25 12
10 316 250 118
100 3,160 2,500 1,185
1,000 31,600 25,000 11,850
7.5 Limitations of Microbiological Tests 71
Table 7.4 Codex criteria for L. monocytogenes in RTE foods (Codex Alimentarius 2009) and relative performance in
terms of the log mean concentration (in bold text) that will be rejected with at least 95% probability, assuming a standard
deviation of 0.8
Product Microorganism Analytical methoda Case
Sampling plan and limits/g
n c m M
Ready-to-eat foods that do
not support growth
L. monocytogenes ISO 11290-2 NAb 5 0 102 NA
Log mean concentration rejected = 80 CFU/g
Sampling plan and limits/25 g
n c m M
Ready-to-eat foods support
growth
L. monocytogenes ISO 11290-1 NA 5c 0 0 NA
Log mean concentration rejected = 1 CFU in 55 g
a Alternative methods may be used when validated against ISO methods
b NA = not applicable as Codex criterion used in place of ICMSF cases
c Individual 25 g analytical units (see Sect. 7.5.2 for compositing)
7.5 Limitations of Microbiological Tests
When used properly and combined with validated process controls, testing can provide actionable
information that helps to assure the safety and stability of the products produced. However, testing
cannot guarantee the safety of the product. Microbiological testing alone may convey a false sense of
security due to the statistical limitations of sampling plans, particularly when the hazard presents an
unacceptable risk at low concentrations and has a low and variable prevalence. This is because microorganisms
are not homogeneously distributed throughout food and therefore, testing may fail to
Increasing Plan Stringency
2-Class Sampling Plans
1/10kg 1/kg 1/100g 1/10g 1/g 10/g 102/g 103/g 104/g 105/g
Concentration (CFU)
ICMSF
Cases
13-15
ICMSF
Cases
10-12
ICMSF
Cases
7-9
ICMSF
Cases
4-6
Codex standard for
L. monocytogenes
products that do not
support growth
in
Codex standard for
L. monocytogenes in
products that support
growth
3-Class Sampling Plans
Fig. 7.1 Geometric mean concentrations of hazard rejected with at least 95% probability for Codex L. monocytogenes
standards and ICMSF Cases 4–6 (m = 103/g, M = 104/g), Cases 7–9 (m = 102/g, M = 104/g), and Cases 10–15 (m = 0/25 g),
assuming a standard deviation of 0.8
72 7 Applications and Use of Criteria and Other Tests
detect organisms present in the batch when the sample is taken from an acceptable portion of the
batch. Food safety is always a result of several factors and it is ensured primarily through appropriate
preventive, proactive measures applied along the food chain (e.g., primary production, ingredients,
in-process and processing environment) and not through a microbiological testing alone. End product
testing alone is reactive and deals only with consequences and not with the causes of the problems.
7.5.1 Analytical Method
To be complete, it is important to identify the analytical method that is associated with a microbiological
criterion because variation can exist between the results generated by different methods.
Considerations in assessing and assuring the performance of microbiological analytical methods are
discussed in Appendix A, Sampling Considerations and Statistical Aspects of Sampling Plans.
Estimates for the performance of sampling plans presented in this book do not take into account any
errors that might occur from the microbiological methods used to determine either the presence or
concentration of microorganisms in foods. For consistency, with the Codex Alimentarius Commission,
International Standards Organization (ISO) methods are used for most of the criteria identified in this
book. Appendix C provides a list of the ISO methods referenced in the product chapters. Other methods
may be used if validated against the ISO methods identified.
7.5.2 Analytical Units and Compositing
For serious and severe hazards, enrichment methods are generally recommended to increase the likelihood
that contamination can be detected. Enrichment methods rely on growth of the pathogen to a
level that can be detected in the enrichment medium and the level of detection can vary dependingon
the analytical method used. In most instances, this book recommends use of 25 g analytical units
for enrichment methods. Each 25 g analytical unit should be selected individually. However, for
analysis, multiple units (e.g., 5, 10, 15, 20 etc.) may be composited and run as one test if the method
has been validated to detect growth of a single cell after the period of enrichment. Jarvis (2007)
reviewed the practical considerations to ensure that testing composited samples is as sensitive
as testing
individual units.