2 Sampling For Microbiological Analysis: Principles and Specific Applications

MICRO
ORGANISMS
IN FOODS 2
Sampling for
microbiological
analysis:
Principles and
specific
applications
Second edition
ICMSF
Blackwell Scientific Publications
Contents (continued)
PART II. SPECIFIC PROPOSALS FOR

SAMPLING AND SAMPLING PLANS
10 INTRODUCTION: THE APPLICATION AND USE OF CRITERIA
127
A. Format of commodity chapters
127
B. Selection and application of criteria
128
C. Choice of organisms
129
D. Selection of limits
129
11 SAMPLING PLANS FOR RAW MEATS

A. Reasons for sampling
130
130
B. Sampling plans
132
C. Sampling procedures
135
D. Sample collection
136
E. Test procedures
137
12 SAMPLING PLANS FOR PROCESSED MEATS

B. Sampling plans
13 SAMPLING PLANS FOR POULTRY AND POULTRY PRODUCTS
139
139
140
148
148
B. Sampling plans
149
154
D. Test procedures
154
14 SAMPLING PLANS FOR FEEDS OF ANIMAL ORIGIN AND PET

FOODS
157
A. Feeds
157
B. Pet foods
158
C. Choice of microbiological criteria
159
15 SAMPLING PLANS FOR MILK AND MILK PRODUCTS
161
A. Relatively perishable products
161
B. Relatively stable products
162
16 SAMPLING PLANS FOR EGGS AND EGG PRODUCTS
169
175
A. Sampling plans
176
B. Sampling procedures
179
17 SAMPLING PLANS FOR FISH AND SHELLFISH
181
A. Public health aspects of fish as human food
181
B. The basis for specific sampling proposals
184
C. Sampling plans
187
D. Sampling procedures
193
18 SAMPLING PLANS FOR VEGETABLES, FRUITS, AND NUTS
197
A. Raw vegetables and fruits
197
B. Frozen vegetables and fruits
198
C. Canned vegetables and fruits
199
D. Dried vegetables and fruits
199
E. Yeasts and single-cell proteins
201
F. Nuts
201
G. Sampling procedures
202
19 SAMPLING PLANS FOR SOFT DRINKS, FRUIT JUICES,

CONCENTRATES, AND FRUIT PRESERVES
20 SAMPLING PLANS FOR CEREALS AND CEREAL PRODUCTS
203
206
A. Grains and flours
206
B. Frozen and dried pasta products
210
C. Frozen filled and topped bakery products
210
D. Sampling procedures and methods
21 SAMPLING PLANS FOR SPICES, CONDIMENTS, AND GUMS
211
213
A. International commerce
213
B. Collection of spices and gums
214
C. Microbiology of spices, gums, and herbs
214
D, Sampling plans
215
22 SAMPLING PLANS FOR FATS AND OILS
217
23 SAMPLING PLANS FOR SUGAR, COCOA, CHOCOLATE, AND

CONFECTIONERY
219
I. Sugar and related sweeteners
219
219
B. Sampling plaiis
220
II. Cocoa and chocolate
221
221
B. Sampling plans
222
III. Confectionery
222
223
B. Sampling plans
223
IV. Sampling and testing procedures for cocoa, chocolate liquor,

coatings containing cocoa, chocolate, and other confectionery
24 SAMPLING PLANS FOR FORMULATED FOODS

A. Foods for infants and children
223
226
228
B. Sampling plans for foods for infants and children and for certain
categories of dietetic foods
230
232
25 SAMPLING PLANS FOR NATURAL MINERAL WATERS, OTHER

BOTTLED WATERS, PROCESS WATERS, AND ICE
234
A. Definitions and economic importance of waters in international trade
234
B. Reasons for sampling
236
C. Sampling plans
237
D. Sampling procedures
241
E. Test procedures
26 SHELF-STABLE CANNED FOODS
241
244
A. Recommended hygienic practices for canned foods
245
B. Integrity of containers
245
C. Cooling water
245
D. Incubation tests
246
E. Process validation and records
246
F. Sampling
247
GLOSSARY
275
PART II
SPECIFIC PROPOSALS FOR SAMPLING
AND SAMPLING PLANS
10
Introduction:
The application and use
of criteria
This chapter provides background information on the considerations

which have led the Commission to propose microbiological criteria for
some commodities and not others. It also indicates how the criteria
should be interpreted and applied.
A. FORMAT OF COMMODITY CHAPTERS
In the first edition of this book foods were grouped on the basis of
commodities (e.g., milk and milk products) or processes (e.g., frozen
foods, dried foods). In this edition the commodities grouping used in
Microbial Ecology of Foods, vol. 2 (ICMSF 1980), has been followed with
two exceptions. These are formulated foods, comprising raw materials
from several commodity groups, and low-acid canned products. Criteria
for formulated foods will depend on conditions of manufacture, the types
of raw materials used, the process, the intended distribution system, and
shelf-life. Hence criteria are proposed only where a need has been
demonstrated and such information is available (see Chapter 24,
Formulated Foods). Sampling plans and microbiological tests are not
relevant to the safety of shelf-stable canned foods and are therefore not
proposed. Extensive cross-reference is made to the volume referred to
above (ICMSF 1980) so that information pertaining to the need for criteria
and the appropriate cases may readily be found.
Each chapter discusses the microbiological hazards associated with a
commodity group and, based on a consideration of their relevance, may
propose criteria. General sampling procedures are dealt with in Chapter
9, but if a commodity requires special sampling procedures
Reprinted from: Microorganisms in Foods 2. Sampling for microbiological analysis: Principles and specific
applications. 2nd Ed. International Commission on Microbiological Specifications for Foods.
128
Proposals for sampling and sample plans
these are detailed within the chapter. For methods, reference is made to
ICMSF, 1978, or to other appropriate sources.
B. SELECTION AND APPLICATION OF CRITERIA
Ideally the control of microorganisms in foods is at the point of production, processing, or preparation for consumption (see Chapter 7,
Control at Source). However, for much food in international trade there is
no knowledge of control at source or of the conditions used during
processing and handling. Therefore, there remains a need for criteria to
assess the acceptability of foods at port of entry.
Before recommending a criterion for a product, the Commission
decided that each product must meet the following conditions:
1 The product must be in international trade.
2 There must be good epidemiological evidence that the product, or
product group, has been implicated in food-borne disease and/or may
have an inadequate shelf-life if Good Commercial Practice (GCP) has
not been followed.
3 There must be good evidence, or good reason to believe, that the
application of a criterion will reduce the health risk in food and/ or
effectively assess adherence to MCP.
It is important to consider some of the difficulties in expecting a
microbiological examination to portray the true microbiological condition
of food. For example, a relatively few sample units may not accurately
reflect the true microbiological status of a consignment, particularly if the
consignment is composed of several batches of food. In this case, if only
one of the batches is defective, sampling a portion of the consignment
may not reveal the unacceptable part. The resulting decision, if applied to
the whole consignment, would be quite unsatisfactory. The sampling
plans proposed in this book should whenever possible be applied to
identifiable lots (see definition of a lot, Chapter 3, Section A).
If at port of entry the intention is to safeguard against staphylococcal
intoxication, a criterion for Staphylococcus aureus would not detect the
hazard if the viable cells have died. If epidemiological evidence shows
that this product constitutes a hazard, a more direct criterion would then
be needed, such as one for thermonuclease or enterotoxin. If a criterion is
applied to a finished product it may be inadequate if a destructive process
has been applied (e.g., destruction of Staph. aureus, but with enterotoxin
persisting).
The application and use of criteria
129
The microbiological status of some foods changes continually during

storage and distribution and is taken into account when selecting the
case. Thus, control at the source is the ideal, and criteria applied at the
port of entry will never be as effective, though there are many instances
where their application is effective (e.g., Salmonella in dried milk).
C. CHOICE OF ORGANISMS
The ICMSF has attempted to include in criteria those microbes most

important in respect of hazard and/or non-compliance with GCP. This
choice was based on epidemiological evidence and expert opinion,
recognizing limitations of current methodologies.
D. SELECTION OF LIMITS
For 3class attributes plans it was necessary to establish m values

(associated with GCP) as well as M values (related to the safety/quality
limit). The latter are based on expert opinion as to the acceptable limit,
but the former should be based on firm data obtained from producers
and retailers operating according to Good Manufacturing Practice (GMP)
and GCP.
Although ICMSF has collected a large amount of data, there is not
sufficient information for certain commodities to establish m values on
this basis. When information was not available, expert opinion was
relied upon to establish m values for 3class plans. Where such values
could not be derived, 2class plans are adopted. Appropriate values of m
for 3class plans must await the collection of further microbiological
data.
The ICMSF has now implemented a pilot computer-based collection
and retrieval system for microbiological data and is anxious to obtain
data to expand this data base. Readers having relevant data are invited to
contact the secretary of the ICMSF, who will be pleased to provide printed
forms for confidentially recording data in a format suitable for the
system.
REFERENCES
ICMSF. 1978. Microorganisms in Foods. 1. Their Significance and Methods of
Enumeration, 2nd ed. University of Toronto Press, Toronto
1980. Microbial Ecology of Foods. Vol. 2. Food Commodities. Academic Press, New York
11
Sampling plans for raw meats
World trade in raw meats, both chilled and frozen, continues to be

significant in terms of volume despite their potential hazard to health.
Total world red meat production in 1979 was more than 78 million
tonnes, with about 9% being traded internationally.
A. REASONS FOR SAMPLING
Raw meats are important sources of Salmonella and Clostridium perfringens, which are often incriminated in outbreaks of food-borne disease
(ICMSF, 1980, p. 354). They are also sources of Staphylococcus aureus,
Campylobacterfetus subsp. jejuni, and Yersinia enterocolitica. Salmonellae are found more often in pork and in veal from young calves than in
other meats (ICMSF, 1980, p. 347). In the first edition of this book criteria
for salmonellae were included as guidelines to assist control of this
organism in raw meat. It was hoped that such guidelines would lead to a
reduction in the contamination rate and a drop in the incidence of foodborne disease. However, outbreaks of salmonellosis due to meat have
continued with little evidence of diminution. In some instances
salmonellosis has been caused by the consumption of raw or inadequately
cooked meat products but a more common hazard arises through crosscontamination from raw to cooked meat or other foods, and subsequent
time-temperature abuse.
Clostridium perfringens is ubiquitous, and although it occurs on
carcass meat, usually in low numbers, it cannot be controlled by any
known means. The majority of outbreaks of C. perfringens gastroenteritis attributable to meat result from inadequate storage of the cooked
product. Prevention involves attention to the time/temperature conditions
applications. 1986. 2nd Ed. International Commission on Microbiological Specifications for Foods.
Raw meats
131
of cooking and, more important, to hot holding, cooling, and reheating

before consumption.
Staphylococcus aureus may occur on raw meat although usually only
in low numbers. Contamination by animal strains of Staph. aureus, which
are thought to have a low enterotoxin-forming potential, is probably of
less consequence than contamination from human sources.
Staphylococcus aureus competes poorly with the normal microbial flora
of raw meat and constitutes a health hazard only when this competing
flora is restricted and there is temperature abuse of the product.
Campylobacter fetus subsp. jejuni is often present in the intestinal
flora of healthy animals used for food production (Blaser, 1982). However, the numbers present on red meat are generally low and the organism
has only a limited potential for growth or survival on refrigerated or
cooked meat. Food-borne outbreaks of Campylobacter enteritis appear to
arise almost exclusively from eating raw or undercooked food (Blaser,
1982).
The use of microbiological criteria as a control measure for reducing
the incidence of disease arising from the consumption of meat carrying
Salmonella, C. perfringens, and Staph. aureus was considered by a
FAO/WHO Working Group (FAO/WHO, 1979) in light of the Codex General
Principles for the Establishment and Application of Microbiological
Criteria for Foods (Codex, 1981). According to these principles a
microbiological criterion should be established and applied only where
there is a definite need and where it is both practical and likely to be
effective. Such criteria should form part of, or be related to, a Codex
Code of Practice for the particular product. The presence of Salmonella
and C. perfringens on raw meats is generally more a reflection of their
incidence in the live animal than of a breakdown of hygiene. In present
circumstances the occurrence of salmonellae and the other pathogens in
raw meat cannot be entirely prevented by the application of codes of
hygienic practice (ICMSF, 1980, p. 347). Also, the extreme variability of
distribution of pathogens such as salmonellae in meats prevents the
establishment of practical sampling plans which would ensure with any
degree of confidence the absence of salmonellae in meats. Consequently
the Working Group concluded that the application of limits for pathogenic microorganisms at port of entry was not appropriate in terms of the
Codex General Principles.
In view of the importance of Salmonella as a cause of food-borne
disease and meat as a vehicle for its transmission there is much to be
gained from regular in-plant monitoring to establish the extent of the
problem in particular areas and to detect increases in prevalence.
132
Proposals for sampling and sampling plans
Products that have a history of being implicated as causes of salmonellosis should be sampled at appropriate points during production and
distribution to determine the prevalence of contamination and to trace its
source on the farm as well as to look at the effects of travel, holding, and
processing of stock. The aim is to collect information that will help to
motivate the producer to eliminate salmonellae in the live animal and
thus control the problem at its source.
The ultimate solution to the Salmonella problem lies in producing
Salmonella-free animals. The provision of Salmonella-free foodstuffs for
animals would be a major advance towards this aim (see Chapter 14,
Feeds of Animal Origin and Pet Foods). Until this is achieved treatments
that destroy salmonellae in the product (e.g., irradiation) may provide
some protection for the consumer. In the absence of effective control
measures, environmental and line sampling are more useful than testing
the end-products. Although improvement of slaughter hygiene is incapable of eliminating contamination with salmonellae, neglect of
hygiene can make the situation much worse. A comprehensive monitoring program in the processing plant can provide a check on the
prevalence of infection in livestock and the effectiveness of measures of
hygiene in controlling spread of contamination.
The FAO/WHO Working Group (FAO/WHO, 1979) concluded also that
the number of indicator organisms in meat neither reflects adherence to a
code of hygienic practice nor indicates presence or absence of pathogens.
Hence criteria based on indicator organisms were not justified for raw
meat. These conclusions, and those concerned with pathogens, have been
B. SAMPLING PLANS
In the virtual absence of studies relating health risks to surface counts of

aerobic organisms on carcass meats it is not appropriate to reject
consignments of product for public health reasons solely on the basis of
high aerobic plate counts. However, aerobic plate counts (APC) can be
used to monitor Good Commercial Practice (GCP), and criteria based on
such examinations are a valuable aid to in-plant quality control.
Microbiological quality control of meat-processing involves development
and use of processing methods designed to keep microbial numbers low
by reducing contamination and preventing growth. Microbiological
monitoring of product and plant at critical control points can be used
Raw meats
133
TABLE 19
Sampling plans and recommended microbiological limits for raw meata
Limit per cm2
or gram
Productb
Carcass meat, before
chilling
Carcass meat, chilled
Edible offal, chilled
Carcass meat, frozen
Boneless meat, frozen
(beef, veal, pork,
mutton)
Comminuted meat,
frozen
Edible offal, frozen
Test
Method
referencec
Case
Plan
class
m
5
M
106
APC
118
10
APC
APC
118
118
118
118
1
1
1
1
3
3
3
3
5
5
5
5
3
3
3
3
106
106
5x105
5x105
107
107
107
107
APC
118
106
107
APC
118
5x105
107
APC
APC
a Not for use at port of entry but for in-plant quality control. Refer to Section B for
appropriate applications.
b Unfrozen carcasses and primal cuts, swab counts per cm2; other meats and meat products,
per gram.
c Refers to page number in ICMSF 1978.
to assess the effectiveness of sanitation factors in limiting microbial

growth. These counts are compared with APC criteria established for
particular products prepared under specific conditions. Sampling plans
and microbiological criteria for certain raw meat commodities are listed
in Table 19.
The proposals for raw meats are based on limited data collected
mainly from production and under a restricted range of circumstances.
The criteria have been established as guidelines and some tolerance
must be allowed in their application; for example, meat marginally
acceptable, or even rejected, by these guidelines could be used in
heated products, but would have very limited shelf-life as fresh meat,
even under good refrigeration. Meat prepared according to the Recommended International Code of Hygienic Practice for Fresh Meat
(CAC/RCP 11-1976, FAO, Rome) may on occasion exceed these limits,
which should not be used as legal standards. No attempt has been made
to distinguish between meat from different species although it has been
noted that sheep carcasses normally have a slightly higher level of
initial
134
contamination than those of beef and numbers of bacteria tend to be

higher again on pig carcasses (ICMSF, 1980, p. 342).
Although there are considerable quantities of horse meat in international trade, criteria have not been included because of lack of data.
The FAO/WHO Working Group (FAO/WHO, 1979) considered that the
only point at which APC could be used for chilled meat, to evaluate the
hygienic conditions under which it was produced, was at the slaughterhouse, and because of its perishable nature it would not be possible to
set APC limits for use in guidelines or specifications for chilled meats at
port of entry. This consideration applies to meat that is vacuum-packaged
in an oxygen-impermeable plastic film as well as to chilled meat
packaged in other ways. Vacuum-packaging and modified-atmosphere
packaging bring about changes in the bacterial flora, and the storage life
depends more on the nature of the flora that develops than on the
numbers of bacteria present after processing.
Case 1 with a 3class plan for the APC is appropriate for utility tests
for general contamination of a product that is normally cooked before
consumption. The values for m reflect current commercial attainment
based on the examination of meat from different sources, mainly at point
of production, and are supported by experience that meat with plate
counts lower than the value set for m usually has not been subject to
excessive contamination or undue faulty handling. A normal shelf-life
would be anticipated. Overall data from a large number of samples have
been analysed but not all products have been tested to a similar degree.
The M values for APC are based on experience that meat with APC
values in excess of 107 is either grossly contaminated or has been
exposed to conditions permitting microbial growth to a level not far short
of the point at which incipient spoilage can be detected. Immediately after
slaughter an M value of 106 is justified on the basis that counts below this
level can be readily achieved under GCP. At this level some tolerance is
allowed for further processing which could involve slight increase in
microbial numbers.
Microbiological criteria applied to frozen meat should be similar to
those applied to chilled meat. The criteria proposed for frozen carcass
meat and frozen cuts assume that the meat is sampled in the frozen state
and thawed quickly before examination of samples. Spoilage occurs at
about the same level; hence the M values are the same. Reflecting the
small decrease in numbers due to freezing and storage m values are 5 x
105 instead of 106. Commercially thawed meat usually has a
Raw meats
135
higher count than the frozen product, and the criteria for chilled carcass
meat could apply.
The microbiological quality of comminuted meat reflects the situation in the meat from which it is prepared. Counts in general are 10 to
100fold greater in commercial minced meat than on the equivalent
carcass (ICMSF, 1980, p. 370). If scrap meat and trimmings from the outer
surface of carcasses are used, the counts in the mince will be higher than
if whole cuts are minced. Mechanically deboned meat derived from
good-quality raw material can be microbiologically equivalent to or
better than minced, manually boned meat provided strict temperature
control is exercised, but the process needs careful control in respect to
the material being boned and sanitation of the equipment. There have
been many proposals for microbiological criteria for minced meat but
there is no evidence that a criterion has any relevance to health hazard.
Accordingly criteria similar to those for carcass meat are proposed.
There is little information available on the microbiology of edible
offals but the considerations are similar to those for carcass meat and the
criteria proposed are the same.
C. SAMPLING PROCEDURES
(a) Carcasses and primal cuts

For chilled or frozen carcass meat and large bone-in or boneless cuts, as
n = 5 for all criteria suggested, take subsamples from 5 individual carcasses or cartons of cuts (sample units). If the lot is distributed in several
shiploads or freightcars take the sample units from more than one
transportation unit.
Because contamination of meat is often very uneven, a number of
subsamples should be taken from different parts of the carcass or cut,
including areas known to be subject to contamination or particularly
favourable for microbial growth. The subsamples from a carcass or
carton may be pooled and thoroughly mixed to form a composite
analytical unit, or each may be treated as a separate analytical unit.
Bacterial contamination of carcasses or cuts is generally on the
surface, and this should be recognized in the collection of subsamples.
Because defacing of carcasses or cuts is often economically undesirable,
the surface swab technique is recommended for the APC on unfrozen
meat. Samples should be taken at sites likely to be contaminated (see,
136
e.g., Roberts et al., 1980). Lamb and pork carcasses should be swabbed at
at least two points (leg and brisket), and beef and horse at three (leg,
flank, and neck). Sampling of cuts should include the original exterior
surface and a cut surface.
For a comparison of swabbing and excision methods see ICMSF, 1980,
p. 359.
(b) Boneless bulk cuts, comminuted meat, and edible offals
Except for the offals it is usually not possible to identify the original
surface. From each of 5 cartons or packages remove a sample unit of
about 200 g of tissue, which should comprise subsamples from a number
of different parts of the pack. This composite sample unit should be
mixed thoroughly and the appropriate analytical unit withdrawn.
(c) Retail packages
These include chilled or frozen consumer-portion packed cuts and offals
as well as comminuted meat. The sampling procedure entails the examination of 5 packages from the lot.
D. SAMPLE COLLECTION
Remove any wrappings from the carcass or package carefully without

handling the meat. Use sterile swabs and templates for taking surface
swab samples. Mark the sampling area with sterile metal guides (e.g., 5,
50, or 100 cm2). Use two sterile swabs to sample each area. Moisten the
first swab with sterile peptone water and rub firmly across the exposed
area several times in all directions. Use the second swab dry and rub over
the same area. Introduce both swabs into a bottle containing 3 or 4 glass
beads and an appropriate known volume of diluent (e.g., 0.1 % peptone,
0.9% NaCl). Shake vigorously.
To avoid cross-contamination when removing portions of meat, use
sterile instruments (knives for unfrozen meat; saws, cleavers, and special
drills for frozen meat), and transfer the piece aseptically to sample
containers.
Samples of frozen meat should remain frozen during transport to the
laboratory. Samples from chilled carcasses should not be frozen.
Raw meats
137
E. TEST PROCEDURES
(a) Aerobic plate counts

Comminute, blend, or stomach the various portions of meat
composing the sample unit and withdraw the analytical unit. Combine
the swabs to form an analytical unit, or treat each separately. Prepare
dilutions and perform the APC test according to the directions given in
ICMSF, 1978, p. 118. Spread plates are preferable to pour plates when
examining raw meats.
The temperature used for incubation of plates should be related to
the material being examined. To estimate numbers of mesophilic contaminants on freshly dressed carcasses it is advisable to incubate plates
at 35 or 37C. For meat that has been held under chilled conditions,
plates should be incubated at 25C to include psychrotrophs and thus
give a better indication of the microbial growth that has occurred.
Counts at 20C give similar results after a longer incubation time.
(b) Salmonella
The test for salmonellae requires an analytical unit of 25 g and is based
on positive or negative results using the method described in ICMSF,
1978, p. 163. For frozen products use a non-selective enrichment
before selective enrichment and plating on selective agar media. Sample units should be taken from appropriate parts of carcasses, cuts of
meat, and offals.
REFERENCES
Blaser, M.J. 1982. Campylobacter jejuni and food. Food Technology, 36 (3), 89-92
Codex. l981. Codex Alimentarius Commission, Fourteenth Session, 1981: Report of the
17th Session of the Codex Committee on Food Hygiene, Washington, D.C.
17-21 November 1980. ALINORM 81/13. Appendix ii, General Principles for the
Establishment and Application of Microbiological Criteria for Foods. Codex
Alimentarius Commission, FAO, Rome
FAO/WHO (Food and Agricultural Organization / World Health Organization). 1979.
Microbiological Criteria for Foods. Report of a Joint FAO/WHO Working Group on
Microbiological Criteria for Foods, Geneva, 20-26 February 1979. (Document
WG/Microbiol/79/1) WHO, Geneva
ICMSF.1978. Microorganisms in Foods. 1. Their Significance and Methods of
138
- 1980. Microbial Ecology of Foods. Vol. 2. Food Commodities. Academic Press, New
York
Roberts, T.A., MacFie, H.J.H., and Hudson, W.R. 1980. The effect of incubation
temperature and the site of sampling on the assessment of the number of bacteria on
red meat carcasses at commercial abattoirs. Journal of Hygiene, Cambridge, 85, 371380
12
Sampling plans for processed meats
Processed meats comprise a number of meat products, manufactured by

many different technologies, including heat-processing, curing, drying,
fermenting, acidifying, and packaging (ICMSF, 1980).
Processing changes the eating quality of meat. It also changes the
composition of the microflora, and some processes kill or prevent growth
of pathogens. Basically processing is intended to prolong the shelf-life of
the products. The increase in shelf-life may range from a few days to
several years. Processed meats with only a slight increase of shelf-life
would not be candidates for international trade, while those given some
degree of heat-treatment or curing combined with an adequate storage
temperature move in international trade.
According to the FAO (1981), international trade during 1980 of
processed meat and meat preparations amounted to about 1 million
tonnes, of which sausages comprised about 115,000 tonnes. These FAO
figures relate mainly to three groups of products: (a) shelf-stable cured
canned meats; (b) perishable cured canned meats; and (c) cooked or
uncooked, fermented, dried, or semi-dried sausages.
In the following section processed meats will be treated according to the

outline in ICMSF, 1980, pp. 378-409. Discussion will deal with hazards
(public health or spoilage), international trade of the products, and the
possibility and feasibility of instituting means of control other than by
microbiological criteria.
140
TABLE 20
Sampling plans and recommended microbiological limits for processed meatsa
Method
referenceb
Case
Plan
class
Limit per
gram
m
M
Product
Test
Dried blood,
plasma, and
gelatin
Staph. aureus
C. perfringens
Salmonella
220
264
163c
8
8
11
3
3
2
5
5
10
1
1
0
102
102
0
Roast beefd
Salmonella
163c
12
20
Pt
Salmonella
163c
12
20
104
104
a Refer to text for appropriate applications.

b Refers to page numbers in ICMSF, 1978.
c See also ISO 6579 (1981).
d Includes beef cooked in water baths; see Section B(f).
B. SAMPLING PLANS
(a) Dried meats and dried animal products

Dried meats like jerky, charqui, and biltong are mainly produced and
consumed locally. Most dried meats found in international trade are
intended for further processing. Further processing would normally
consist of dry-blending with other food ingredients (e.g., dried soups) or
rehydration and canning. In both cases the dried meats are subjected to
conditions that either would not change or would reduce the hazards.
In international trade no evidence of hazard has been demonstrated,
and no microbiological criteria are proposed.
Dried animal products such as dried whole blood, blood plasma, and
gelatin are used as ingredients for formulated foods. When these
ingredients are used in further processed products that are not subjected
to adequate heat-treatment it is proposed to include sampling plans for
Staphylococcus aureus, Clostridium perfringens, and Salmonella (Table
20).
For Staph. aureus and C. perfringens 3class sampling plans, case 8
with n = 5, c = 1, m = 102, M = 104, are proposed. Salmonella should be
considered in case 11 with a 2class sampling plan with n = 10, c = 0.
Processed meats
141
(b) Raw cured meats, high aw

High aw meat products are defined as those with an aw above 0.92,
such as occurs in bacon produced by a Wiltshire or similar process
(ICMSF, 1980, pp. 383-388).
In Wiltshire curing sides or parts of sides of pork, with or without
bones, are injected with brine before either tank- or dry-curing. Smoke
may be applied, but the product is not heated. There are many other
types of manufacture of bacon, but this is the most common for bacon in
international trade.
The microbiology of bacon is characterized by a rapid change from
the initial gram-negative flora to gram-positive salt-tolerant micrococci
and lactic acid bacteria. Only where curing is inadequate or the pH of
the meat (usually the collar) is high will putrefactive spoilage dominate.
Otherwise, spoilage is characterized by sliming.
Staphylococci could be considered to be of concern but are not able
to compete with the normal saprophytic flora. Applying microbiological
criteria at manufacture can make little contribution to the control of
spoilage.
Salmonella is not a major concern because the product is cooked
immediately before consumption.
(c) Raw, cured, or salted meats and natural casings, low aw
Low aw meat products are defined (ICMSF, 1980) as those having a
water activity of 0.92 or below. This group includes whole cuts of meat
(e.g., Bundener Fleisch, Parma ham, Westphalian ham, Smithfield ham,
Prosciutti), non-fermented dried sausages, and natural casings, all of
which are in international commerce.
Although some of the raw cured or salted meats have been implicated in outbreaks of food-borne illness, particularly botulism, the
incidents have usually involved home-prepared products. Commercial
products have had a very good record of safety. Growth of nonpathogenic bacteria occurs during the process of salting, curing, and
drying of many of these products and can contribute to their flavour
characteristics. Hence, microbiological criteria need not be applied to
these commercial products. Cured, dried non-fermented sausages are in
international trade but the only spoilage problem appears to be mould
formation on the casing. No public hazards have been identified, and no
microbiological criteria are proposed.
Natural casings from international commerce are generally packed
142
in salt and have not been identified as a vehicle of food-borne illness.

Survival and growth of some salt-tolerant bacteria occurs in salted
casings, but Salmonella indigenous to the casings does not survive for
more than two or three weeks after salting. Microbiological criteria are
unnecessary after salting.
(d) Fermented sausages

Cured fermented sausages include both semi-dry and dry types. Both
types undergo a fermentation process which should rapidly lower the
product pH to <5.5. At this point semi-dry products may or may not be
heated but if the pork used contains Trichina heating to 58.3C may be
mandatory. Other semi-dry sausages may only receive light heating (e.g.,
Thuringer). Dry sausage is placed in dry rooms with controlled
humidity for an extended period after fermentation. During this ageing
process, the product loses moisture and the aw falls to c. 0.88, which is
sufficiently low to make the heat-treatment to control Trichina
unnecessary.
In both semi-dry and dry fermented sausages the main hazard is the
growth of Staph. aureus during the initial stages of the fermentation
before the pH of the mix falls sufficiently to inhibit its growth.
It has been suggested that microbiological criteria for Staph. aureus
(not more than 100,000 per gram) be applied at the conclusion of the
fermentation process or at the beginning of the drying process. More
benefit could be obtained by a controlled acidulation to below pH 5.4
(using starter cultures or chemical acidulants such as glucono-deltalactone) than by microbiological testing. Salmonella derived from the raw
ingredients may survive processing but is unlikely to multiply. Total
counts criteria are inappropriate because of growth of fermenting
organisms.
The American Meat Institute has recently formulated Good Manufacturing Practices in connection with the production of fermented
sausages (AMI, 1982). A maximum pH of 5.3 is prescribed in both dry
and semi-dry fermented sausages.
(e) Cured meats packed under vacuum
Some of these products support the growth of Clostridium botulinum
(Pivnick and Barnett, 1965) and Salmonella (Davidson and Webb, 1972)
at abuse temperatures. The developing flora in vacuum-packaged cured
meats is composed almost entirely of lactic acid bacteria. Large
Processed meats
143
numbers of these organisms may develop without objectionable organoleptic changes. Despite the fact that these products are often
abused through inadequate refrigeration, outbreaks of staphylococcal
foodpoisoning are seldom traced to them, probably because
staphylococci fail to grow in competition with other microorganisms.
Salmonellae are capable of survival during protracted storage of these
products at 5C and can grow when the meats are incubated at ambient
temperature (Goepfert and Chung, 1970), but these products have not
proved to be a significant Salmonella hazard.
It is assumed that sliced bacon will be heated prior to consumption
and, in addition, that the growth of lactic acid bacteria will afford
protection against growth of Staph. aureus. There is no evidence of
food-borne disease from sliced bacon and thus criteria are not
proposed.
Ham is sometimes involved in food-borne illness due to Staph.
aureus, usually as a consequence of contamination during slicing or
handling, followed by temperature abuse. Vacuum-packed sliced hams
are rarely implicated. During storage of vacuum-packed meats, including hams, under refrigeration a large population of lactic acid
bacteria develops in whose presence staphylococci fail to grow. There
is no evidence that criteria at manufacture or at port of entry would
reduce the hazard.
Frankfurters and bologna have also been implicated occasionally in
food-borne illness, but control of the production process affords greater
assurance of safety than would the application of criteria.
(f) Cooked uncured meats
Cooked uncured meats (e.g., roast-beef joints and cuts) are increasingly
important in international trade. Their involvement in food-borne salmonellosis could be avoided by proper heat-processing, prevention of
post-process contamination, rapid cooling, and subsequent appropriate
time-temperature control in the home or food-service establishment. A
criterion for Salmonella is proposed: case 12, n = 20, c = 0 (Table 20).
There is no evidence that aerobic plate count (APC) limits would serve a
useful purpose.
Considerable international trade exists in pasteurized comminuted
meat products such as pates. At port of entry they frequently have high
APCs and are occasionally associated with food-borne illness. Control is
by adequate heat-processing, prevention of recontamination, and
appropriate chill storage. Microbiological data are inadequate, but these
products may contain Salmonella. If there is reason for concern,
applications. 1986. 2nd Ed. International Commission on Microbiological Specifications for
144
the appropriate sampling plan for Salmonella would be case 12, n = 20,
c = 0 (Table 20).
(g) Perishable cured canned meats
The most important meat products of this kind in international trade
are pasteurized canned cured hams and pork shoulders. A high proportion of these products are heated to centre temperatures of only 65 to
75C. They should be transported under refrigeration, and should be
labelled perishable, keep under refrigeration or with words to similar
effect. While the importance of testing such products is recognized, the
high price of canned hams and similar foods is a serious obstacle to
intensive examination. For these products, data on thermal processing,
water supply, seam inspection, and chemical composition should be
available, together with records of temperature during shipping. If all
such data are satisfactory, no testing is necessary.
The Codex Alimentarius Committee for Processed Meat and Poultry
Products is currently considering sampling plans for this category of
products. Until the results become available the following is offered for
guidance.
1. Sampling procedure at port of entry. If adequate production
records are not available, sampling should be done according to the
following scheme (see also Table 21).
a
b
Measure the air temperature in the space between containers, pref

erably with an electronic measuring device.
Take 10 containers at random from the lot. Identification of in
dividual containers at this point is unnecessary. Examine the 10
containers for swells and defects (see footnote, Table 21).
If the temperature does not exceed 10C or any specified lower
temperature and if no swells or defective containers are found
accept the lot.
If one or more defective or swollen containers are found, hold the lot
and determine the cause by more extensive sampling (see Chapter 5,
Section N, Investigational Sampling).
If the temperature exceeds 10C or any specified lower temperature,
or if there is reason to suspect that these temperatures have been exceeded
during transport, proceed to step d.
Processed meats
145
TABLE 21
Sampling plans for perishable canned cured meats (to be used when processing data are not
available or are unsatisfactory)
Test
Plan
Case class
Limit per
gram
m M Acceptance
Step 1
Visual inspection
for swells and
defective
containers*
10
Reject if 1 or more defective

cans are found. Proceed to
step 2 if no defective cans are
found.
Step 2
Measurement of
air temperature
between cans
10
Accept if temperature is
below 10C. If higher
proceed to step 3.
Step 3
Determination of
aerobic plate
count in centre
and periphery
d
e
103 104 See Section G of this chapter

for interpretation of the
presence of microorganisms.
A defective container is a container which is of faulty manufacture, is improperly closed,

or has been damaged in such a manner that may permit contamination of the contents
following the heat-process (post-processing contamination).
Take 5 containers from the warmer places in the lot and hold the
lot. Proceed to step e.
Identify the 5 containers selected under d and send them to a
laboratory for microbiological examination. Transport should take
place under refrigeration, at 10C or less.
2. Laboratory analysis
a In the laboratory draw sample units from the 5 containers aseptically so as to obtain one unit from the centre and one from the
periphery of each container.
b Examine for APC (ICMSF, 1978, p. 118). A direct Gram stain and
an anaerobic count may sometimes be helpful. Rarely, the only
bacteria present in large numbers are anaerobes.
146
Reject if an analytical unit from one or more of the 5 containers has a

viable count exceeding 104 per gram. Also reject if one or two
analytical units from 3 or more containers show a viable count higher
than 103 per gram. Otherwise, accept.
When rejection is indicated an investigation into the cause of the

problem might be appropriate and the supplier informed of the
results. It should be emphasized that the values for m and M in this
context do not necessarily reflect any spoilage or health hazard, but
merely indicate degrees of undesirable commercial practice. The
limits have been established from data derived from very large
numbers of analyses over several years and are known to be attainable under GCP. It has been found that at export counts are
normally less than 102 per gram. If the lot does not meet this criterion,
the decision on the action to be taken should be one of the options
given in Chapter 2.
(h) Shelf-stable cured or uncured canned meats

Shelf-stable meats have undergone a treatment sufficient to guarantee the
subsequent safety and stability of the product at normal ambient
temperatures. Canned cured and canned uncured meat products are
considered with other low-acid canned foods in Chapter 26.
REFERENCES
AMI (American Meat Institute). 1982. Good Manufacturing Practices: 1. Voluntary
Guidelines for the Production of Dry Fermented Sausages. 11. Voluntary Guidelines for
the Production of Semi-dry Fermented Sausages. AMI, Washington, DC
Davidson, C.M., and Webb, G. 1972. The behaviour of salmonellae in vacuum-packaged
cooked cured meat products. Proceedings of the 18th Meeting of Meat Research Workers,
Guelph, Canada. University of Guelph, Guelph, Ontario
FAO (Food and Agriculture Organization) 1981. Food Trade Yearbook, Vol. 34.
FAO Statistics Series No. 35, Rome
Goepfert, J.M., and Chung, K.C. 1970. Behaviour of salmonellae in sliced luncheon meats.
Applied Microbiology, 19, 190-192
ICMSF.1978. Microorganisms in Foods. 1. Their Significance and Methods of
ISO (International Organization for Standardization) 1981. Microbiology - General
Guidance on Methods for the Detection of Salmonella. International Standard
6579. ISO, Geneva
Reprinted from: Microorganisms in Foods 2. Sampling for microbiological analysis: Principles and specific applications. 1986.
2nd Ed. International Commission on Microbiological Specifications for Foods.
Processed meats
Pivnick, H., and Barnett, H. 1965. Effect of salt and temperature on toxinogenesis by
Clostridium botulinum in perishable cooked meats vacuum-packaged in airimpermeable plastic pouches. Food Technology, 19 (5), 140-143
147
13
Sampling plans for poultry
and poultry products
There has been a steady increase in trade on the world poultry market. In
1980, over 1.4 million tonnes of poultry were exported from 45 poultryproducing countries (FAO, 1981). Imports by the industrialized and nonindustrialized countries were nearly equal. Most of the product is frozen;
however, there is some trade in fresh poultry between closely situated
countries. There is also some trade in heat-processed and dried poultry
products.
Raw poultry products are frequently contaminated by food-borne pathogens (e.g., Salmonella, Staphylococcus aureus, Clostridium perfringens,
Campylobacter fetus subsp. jejuni, and Yersinia enterocolitica; ICMSF,
1980, pp. 419-50). Disease surveillance reports frequently identify
poultry as vehicles in outbreaks of salmonellosis, staphylococcal foodpoisoning, C. perfringens enteritis, and other enteric illnesses (Bryan,
1980; Hepner, 1980; Horwitz and Gangarosa, 1976; Todd, 1978;
FAO/WHO, 1979). The prevalence of salmonellae in raw fresh and frozen
poultry is of great concern and approaches 80% in some countries.
Contaminated chickens and turkeys bring salmonellae into kitchens
and give rise to human salmonellosis, either directly or through crosscontamination to other foods (ICMSF, 1980, pp. 838-61; Bryan, 1981).
Such outbreaks commonly occur from foods served in food-service
establishments, hospitals, and other institutions. Deaths are infrequent,
but they do occur in debilitated persons and infants.
Programs for decreasing Salmonella infection to low levels in fowl
have been known for many decades; these are control of Salmonella
Reprinted from: Microorganisms in Foods 2. Sampling for microbiological analysis: Principles and specific applications.
1986. 2nd Ed. International Commission on Microbiological Specifications for Foods.
Poultry and poultry products
149
in breeding flocks, prompt fumigation of eggs for hatching, rigid sanitary control in hatcheries, and provision of Salmonella-free feed. Where
practised, these measures can greatly reduce the prevalence of Salmonella-contaminated carcasses, leading to low incidence of human
infection from poultry. Most countries, however, have few or ineffective
programs to combat the Salmonella problem, increased cost of poultry
products being the most common reason given. An effective, but at
present commercially unavailable, treatment could be the irradiation of
packaged poultry products (ICMSF, 1980, pp. 455-457).
Spores of C. perfringens survive in some heat-processed poultry
products. Non-spore-forming pathogens may survive if heat-processing
is inadequate, and they may also survive curing, smoking, or drying
processes. Poultry products are subject to post-processing contamination from staphylococci when sliced, wrapped, or otherwise handled
by workers in processing plants and food-service establishments, and by
persons who prepare foods in homes.
B. SAMPLING PLANS
(a) Fresh and frozen raw poultry

At the time of preparing the previous edition of this book, it was
thought that the sampling plan for Salmonella in poultry of case 10, n =
5, c = 0, m = 0 was not feasible; hence the c value was set at 1 (ICMSF,
1978b, pp. 137-139). Even this has proved to be unrealistic under
current conditions of intensive poultry production and processing in
most countries (FAO/WHO, 1979; ICMSF, 1980, pp. 438-442).
The proportion of salmonellae-contaminated poultry carcasses is
determined mainly by the quantity of incoming infected or contaminated
live fowl and by subsequent spread during defeathering and other
processing (ICMSF, 1980, pp. 410-50). Unfortunately, even good manufacturing practices in a plant do not appreciably affect the situation
(FAO/WHO, 1979). Until changes are made to prevent or minimize the
infection and contamination of fowl on farms, or until a method of
decontamination (e.g., irradiation of packaged carcasses or treatment
with an acceptable chemical) is routinely applied, it is questionable
whether practicable port-of-entry microbiological criteria can be set for
Salmonella on frozen raw poultry without risk of eliminating poultry as
a food. Educational programs are needed to maintain food safety in
kitchens. Training must emphasize thorough cooking, prevention of
cross-contamination, rapid cooling of cooked products that are not
150
to be eaten immediately, and thorough reheating, particularly for poultry

being prepared in medical care and geriatric institutions.
Investigational sampling plans (see Chapter 5, Section N) may be
used to (1) determine the extent of contamination of a feed supply or a
product with salmonellae; (2) identify or monitor critical control points of
hatchery, farm, or processing operations; (3) trace a source of
contamination; or (4) evaluate effectiveness of control measures. Such
data may help to convince producers and others that a serious problem
exists and that steps should be taken towards the elimination of
Salmonella in live fowl, thereby controlling the problem at source.
Initially, a lenient sampling plan for Salmonella will be needed (e.g., n =
5, c = 3, m = 0); but as control measures (such as the use of Salmonellafree chicks and feeds) are initiated and maintained, the plan should be
made more stringent by reducing c.
Although Campylobacter fetus subsp. jejuni and Y. enterocolitica are
frequent contaminants of raw poultry products (deBoer et al., 1982;
Leistner et al., 1975; Norberg, 1981; Simmons and Gibbs, 1979), and
poultry-associated outbreaks of campylobacteriosis have been reported
(Cunningham, 1982), criteria are not considered necessary. Thorough
cooking, care to avoid cross-contamination, and proper storage of cooked
products will minimize the risks associated with these bacteria.
Clostridium perfringens is a part of the normal intestinal flora of
poultry and is shed in their faeces, and is widespread in their environment.
Its presence on poultry carcasses cannot be avoided, but it will not
multiply on chilled carcasses. While cooking destroys vegetative cells of
C. perfringens, spores survive and may even be activated. Also the redox
potential of carcasses is lowered by cooking. Temperature abuse of
cooked poultry can promote spore germination and multiplication of the
resulting vegetative cells. A criterion for C. perfringens in or on raw
poultry for port-of-entry sampling is not recommended. Prevention of
illness must therefore be based on timetemperature control after cooking.
Staphylococcus aureus is a part of the normal flora of poultry, and
even adherence to good processing practices cannot totally prevent
carcass contamination. However, staphylococci do not compete well with
normal spoilage flora of poultry and are not likely to multiply on chilled
carcasses. They become a problem only if cooked poultry is contaminated
and subsequently subjected to time-temperature abuse. A criterion for
Staph. aureus in raw poultry is not recommended for port-of-entry
sampling. Poultry eaten shortly after cooking or not handled after cooking
and rapidly cooled is not associated with staphylococcal food-poisoning.
151
A sampling plan for Staph. aureus is useful to evaluate carcass

contamination from picking (plucking) machines in which staphylococci
can accumulate particularly and contaminate carcasses passing through.
A sampling plan of case 4 (staphylococci as an indicator of a source of
contamination), n = 5, c = 3, m = 103, M = 104 per cm2 of swabbed
surface might be initially considered for in-plant quality control.
It is also considered unnecessary to have microbiological criteria for
other indicator organisms (Enterobacteriaceae, coliforms, faecal
coliforms, Escherichia coli, and enterococci) because they are part of
the normal intestinal flora of poultry. Some will be spread among
carcasses even under good processing practices (FAO/WHO, 1979). Some
strains of species that fall into the indicator organisms categories (such
as certain coliforms and Enterobacter) are psychrotrophic and can
multiply on refrigerated raw poultry carcasses and products.
Microorganisms responsible for spoilage of poultry multiply during
refrigeration, even at recommended temperatures of 1 to 4C (ICMSF,
1980, p. 443). A commonly recommended microbiological method for
estimating the shelf-life of chilled poultry, an aerobic plate count (APC)
at 0 to 5C, is impracticable for use at port-of-entry because it requires
incubation for 10 to 14 days.
In the previous edition of this book, a criterion for APC (20C) of
case 1, n = 5, c = 3, m = 5 x 105, M = 107, was suggested (ICMSF, 1978b,
p. 138). This criterion is achievable for processed chickens, but not
necessarily turkeys, in modern processing plants that follow good
processing practices (CEC, 1979) and is a useful guideline. It is not
recommended, however, as a port-of-entry microbiological criterion or
as a basis for estimating shelf-life.
Irradiation could eliminate Salmonella from processed raw and
frozen poultry (ICMSF, 1980, p. 455). A criterion of case 10, n = 5, c =0
m = 0, would be applicable to Salmonella in raw or frozen irradiated
poultry. It is premature to propose other microbiological criteria for
irradiated poultry before these products enter international trade, and
until additional microbiological data and experience become available.
(b) Frozen, heat-processed poultry products
There is increasing international trade in heat-processed poultry products (i.e., turkey rolls and entree items such as fried chicken, chicken
la king, and pot pies). (See Chapter 20, Cereals and Cereal Products, for
suggested microbiological criteria and further discussion of pot pies
TABLE 22
Sampling plans and recommended microbiological limits for poultry and poultry products
Limit per gram

Method
referencea
218
163
Case
8
10
Plan
class
3
2
n
5
5
c
1
0
m
103
0
M
104
3
3
2
5
10
10
1
1
0
103
103
0
104
104
104
Product
Cooked poultry meat, frozen;
to be reheated before eating
(e.g., prepared frozen meals)
Test
Staph. aureusb
Salmonella
Cooked oultry meat, frozen,

ready-to-eat (e.g., turkey
rolls)
Staph. aureusb
218
Salmonella
163
8c
9c
11
Cured and/or smoked poultry

meat
Staph. aureusb
Salmonella
218
163
9
11
3
2
10
10
1
0
103
0
Dehydrated poultry products
Salmonella
163
11
10
Raw chicken (fresh or frozen),

during processing
APC
118 (20C)
5x105
107d
a Refers to page numbers in ICMSF 1978a, where the methods are described. Use analytical unit sizes recommended in the methods, except
where otherwise indicated.
b If either packaged or repackaged after processing; not for products processed in packages that are kept closed until time of final
preparation.
c Case depends on whether (case 9) or not (case 8) subsequent time-temperature abuse of the products is likely.
d Not for port-of-entry sampling; for use to evaluate production and operations and critical control points in processing plants
only.
Reprinted from: Microorganisms in Foods 2. Sampling for microbiological analysis: Principles and specific applications. 1986. 2nd Ed.
International Commission on Microbiological Specifications for Foods.

15
3
and other blended products with pastry shells or covers.) Presence of

Salmonella may be due to failure of heat processes or cross-contamination from the raw to the cooked product (ICMSF, 1980, pp. 850-859).
A microbiological criterion for Salmonella in heat-processed poultry
products is given in Table 22.
Staphylococci and faecal indicators are useful monitors of contamination of frozen cooked poultry products. Such monitoring is unnecessary when products are adequately cooked in plastic casings or bags,
but it is essential when these products are handled, cut, repackaged after
heat-processing, or otherwise subjected to any operation that may
permit contamination. Although the data base is limited, microbiological criteria for Staph. aureus are recommended. Case 8 applies,
because enterotoxins of Staph. aureus withstand the time-temperature
exposure of reheating. If there is likelihood of time-temperature abuse,
case 9 would apply (see Table 22). Staphylococci survive freezing
better than the Gram-negative organisms (such as E. coli) and can be a
useful indicator of post-heat-process contamination.
(c) Smoked, cured, and dried poultry products
Although there is limited international trade in smoked, cured, and dried
poultry products other than dried eggs, and the need for microbiological
criteria has not clearly been established, these products should be free
from Salmonella. The presence of Salmonella constitutes a health
hazard and case 11 applies (Table 22). Dried products are shelf-stable,
as long as packages are intact, until rehydrated.
Monitoring for APC, E. coli, and Staph. aureus would give some
indication of inadequate processing (heating, curing, smoking), postprocess contamination, or storage at temperatures above those recommended for frozen or refrigerated distribution and storage. These
products are likely to become contaminated during slicing or wrapping.
The aw level may inhibit common spoilage organisms but not necessarily Staph. aureus.
Because only limited microbiological data on these commodities are
available, criteria for APC or E. coli are not recommended. However, a
criterion is suggested for Staph. aureus on smoked or cured poultry
products (Table 22). Case 9 is applicable because these products are
frequently distributed or stored at temperatures conducive to multiplication of staphylococci.
154
Take 5 or 10 sample units, as appropriate, from each lot. Sample units of

fresh or frozen raw poultry should be taken from more than one carton
where possible. Sterile disposable plastic gloves are useful for holding
carcasses. Use sterile instruments to cut portions, if necessary, and to
collect sample units (usually 200 g) of comminuted, cooked, dehydrated,
smoked, or cured poultry, and aseptically transfer them to sterile
containers. (See Chapter 9, Collecting and Handling Sample and
Analytical Units, for more information on aseptic sampling, identifying
samples, and transporting samples to the laboratory.)
Use the shake and rinse procedure for sampling fresh or thawed raw
poultry carcasses to be tested for Salmonella. In this procedure, the
carcass is placed aseptically into a heavy-duty, plastic bag (which has
not been previously used) containing 300 ml buffered peptone water.
The bag is closed, and the enclosed carcass shaken vigorously, vertically
and horizontally, for 30 seconds. The entire surface and cavity of the
carcass should come into contact with the water during the shaking.
Pour the rinse into a sterile container and keep it at 4C until it is used in
the laboratory. If raw turkey carcasses are to be tested, use 500 ml
buffered peptone water. This procedure is difficult to use for large
turkeys because they are heavy and it is difficult to shake them
effectively. Swabbing is a suitable alternative method (see below).
The shake and rinse samples are not usually practicable for on-theline sampling. In such situations, rub a swab (which has been
moistened in sterile peptone water) over the skin and, if practicable,
internal regions of the carcass. Include both neck and vent in the
swabbed areas. Alternatively aseptically cut portions of neck skin and
take interior scrapings. When these procedures are used, however, the
likelihood of detecting salmonellae may be less than with the shake and
rinse procedure.
The recommended procedure for sampling raw carcasses for APC is
to collect samples of neck skin. Use sterile instruments to grasp the
carcass and to cut sample units from carcasses and aseptically put them
into a sterile container.
D. TEST PROCEDURES
Carcass-rinse and swab sample units become the analytical units for the
Salmonella or other test. Thaw (drip) water from raw poultry can
155
also be used as analytical units. Rinse and thaw waters can also be used
for tests for other organisms. Incubate each as prescribed in ICMSF,
1978a, p. 163. When testing skin sample units for Salmonella, put cut
portions of skin into pre-enrichment broth according to procedures in
ICMSF, 1978a, p. 164. For APC, weigh 20-50 g of neck skins and
homogenize this analytical unit with 0.1% peptone saline (ratio 1:9) in a
stomacher or blender.
Sample units from frozen poultry should be thawed at 1 to 5C,
preferably within 18 hours. Comminute and blend a sample unit, and
from this composite weigh out analytical units for the APC, Staph. aureus,
and other counts and for the Salmonella test according to directions
given in ICMSF, 1978a.
REFERENCES
Bryan, F.L. 1980. Foodborne diseases in the United States associated with meat and
poultry. Journal of Food Protection, 43, 140-150
1981. Current trends of salmonellosis in Canada and the United States. Journal of Food
Protection, 44, 394-402
CEC (Commission on the European Communities). 1979. Microbiological Methods for
Control of Poultry Meat, Study P203, VI/5021/79-EN, EEC, Brussels
Cunningham, F.E. 1982. Microbiological aspects of poultry and poultry products - An update.
Journal of Food Protection, 45, 1149-1164
deBoer, E., Hartog, B.J., and Oosterom, J. 1982. Occurrence of Yersinia enterocolitica in
poultry products. Journal of Food Protection, 45, 322-325
FAO. 1981. 1980 FAO Trade Yearbook. Vol. 34. FAO, Rome
FAO/WHO. 1979. Microbiological Criteria for Foods. Report of a Joint FAO/WHO
Working Group on Microbiological Criteria for Foods, Geneva, 20-26 February, 1979
(Document WG/Microbiol/79/1). WHO, Geneva
Hepner, E. 1980. Food poisoning and Salmonella infection in England and Wales, 19761978. An analysis of reports to the Communicable Disease Surveillance Centre of the
Public Health Laboratory Service. Public Health, 94, 337-349
Horwitz, M.A., and Gangarosa, E.J. 1976. Foodborne disease outbreaks traced to poultry,
United States, 1966-1974. Journal of Milk and Food Technology, 39, 859-863
ICMSF. 1978a. Microorganisms in Foods. 1. Their Significance and Methods of
1978b. Microorganisms in Foods. 2. Sampling for Microbiological Analysis: Principles
and Specific Applications. University of Toronto Press, Toronto
Leistner, L., Hechelmann, H., Kashiwazaki, M., and Albertz, R. 1975. Nachweis von Yersinia
enterocolitica in Faeces and Fleisch von Schweinen, Rindern and Gefiugel.
Fleischwirtschaft, 11, 1599-1602
Norberg, P. 1981. Enteropathogenic bacteria in frozen chicken. Applied and Environmental
Microbiology, 42, 32-34
156
Simmons, N.A., and Gibbs, J.F. 1979. Campylobacter spp. in oven-ready poultry.
Journal of Infection, 1, 159-162
Todd, E.C.D. 1978. Foodbome disease in six countries - a comparison. Journal of
Food Protection, 41, 559-565
14
Sampling plans for
feeds of animal origin
and pet foods
Large quantities of feeds destined for animals and increasing quantities

of dehydrated pet foods are in international trade. Feeds and pet foods
have constituents of vegetable and animal origin, both of which can be
contaminated with many kinds of microorganisms or their metabolites,
including those of animal or human health significance. Although the
absence of other pathogens in feeds cannot be guaranteed, the main
organisms of concern are salmonellae. Fungal metabolites are
potentially hazardous to animals and to human health through residues
in foods of animal origin (ICMSF, 1978).
A. FEEDS
Animals ingesting feeds contaminated with Salmonella may contract

clinical disease or subclinical infections and excrete salmonellae for
long periods. This fact has become important with the current increase
in world trade of feeds, and fish meal and meat-bone meal have been
responsible for the spread of certain Salmonella serotypes, such as
Salmonella agona, world-wide (Clark et al., 1973).
Animals excreting salmonellae may endanger public health either by
direct contact (e.g., turtles in the home) or indirectly when foods of
animal origin are contaminated as in slaughtering or milking. For
discussions of factors in addition to contaminated feeds, such as the
environment, see ICMSF, 1980, pp. 459-469; Riemann and Bryan, 1979;
WHO, 1981; Silliker, 1980.
It is accepted that elimination of Salmonella from feeds will not
guarantee the absence of Salmonella in animals. However, to assist in
the reduction of the hazard, efforts should be made to reduce the
contamination of animal feeds either by producing Salmonella-free
158
feed components of animal and vegetable origin or by subjecting the

compounded feed to a decontaminating treatment like certain methods
of pelleting, heating, or irradiation (Statutory Instruments, 1981a, b; van
Schothorst and Brooymans, 1982). However, criteria for livestock feed
have not been proposed because they are unlikely to reduce the
prevalence of Salmonella in animals effectively unless other measures
are taken at the same time. Feed subjected to heating, irradiation, and
some pelleting processes should be free from Salmonella. For such
feeds a sampling plan n = 10, c = 1, or more desirably c = 0, is
recommended. Feeds subjected to other pelleting processes, not primarily intended to destroy salmonellae, cannot be expected to meet this
criterion.
Heating during the fat-extraction stage of production of proteinrich
feed components of animal origin (e.g., fish meal) or of vegetable origin
(e.g., soya flour, cotton-seed meal) yields products free from
Salmonella and other heat-sensitive contaminants. However, postprocess contamination frequently occurs, sometimes reintroducing
Salmonella by cross-contamination with raw material and from
common machinery as well as aerosols from grinders, dust, flies, and
excrement of birds and rodents.
During processing good hygienic housekeeping, keeping dry areas
of the factory dry, completely separating raw material and finished
product, and preventing recontamination are the most important measures to ensure Salmonella-free products. Environment and line sampling
are more effective in Salmonella control than testing end-products
(ICMSF, 1980, pp. 459-469; van Schothorst and Oosterom, 1984). At port
of entry, testing for the presence of Salmonella may identify
contaminated consignments.
B. PET FOODS
Dehydrated, extruded, pelleted, and intermediate-moisture pet foods can

be produced free from Salmonella even though some of the raw
materials contain it. In contrast pet foods produced by blending such
raw ingredients cannot be produced free from Salmonella because no
bactericidal treatment is applied (ICMSF, 1980; van Schothorst and
Brooymans, 1982). Irradiation might be considered in this case (WHO,
1981).
Canned pet foods are not dealt with in this chapter because methods for
control do not differ in principal from those described in Chapter 26
(Shelf-Stable Canned Foods).
Feeds of animal origin and pet foods
159
TABLE 23
Sampling plans and recommended microbiological limits for pet foodsa
Product
Intermediate moisture
pet foods
Test
Salmonella
Method
Referenceb
163172
Dry pet foods not to be

reconstituted
Salmonella
Dehydrated pet food to be

reconstituted
Salmonella
Case
11
Plan
2
n
10
c
0
m
0
163172
11
10
163172
12
20
a Case 11 could also be used for monitoring animal feeds for Salmonella (see this
chapter).
b Refers to page numbers in ICMSF, 1978; see also ISO 6579 (1981).
Frozen raw meat of origins not covered in Chapter 11 (e.g., horse

and kangaroo meat) and meat offals are shipped as pet food to many
countries of the world. These pet foods may contain Salmonella and may
be a source of contamination of the kitchen as well as sources of
infection for animals. These pet foods are occasionally eaten by humans,
though this may be regarded as misuse of the product. Taking into
account the present status of the Salmonella problem (FAO/WHO, 1979),
criteria for these pet foods at the port of entry are not warranted (WHO,
1981; cf. Chapter 11, Raw Meats; Chapter 13, Poultry). However,
monitoring of product for investigational purposes would provide
information on the magnitude of possible problems.
C. CHOICE OF MICROBIOLOGICAL CRITERIA
If the same considerations are applicable to the selection of cases for

Salmonella in animal feeds as for human foods, case 11 and case 12
would be chosen for the various categories of feeds or pet foods. This
choice is reasonable because the ranges of minimal infective doses which
provoke clinical disease in humans and in animals are similar.
Case 11 was previously chosen for animal feeds and dry and
intermediate-moisture pet foods, because their low levels of water
activity prevent multiplication of salmonellae (ICMSF, 1980, p. 469). For
dehydrated pet foods which are reconstituted and left at room
temperature, case 12 applies (ICMSF, 1980, p. 469) (Table 23). Feed
components of animal origin are not included.
160
The criterion for Salmonella in animal feeds should therefore, in

principle, be case 11, n = 10, c = 0, m = 0. However, many feeds in
international trade cannot meet this criterion. Decontamination (e.g., by
irradiation, heating, or pelleting) would increase costs; and adequate
facilities for these processes are not yet widely available. Therefore, it is
recommended that consignments be monitored for the presence of
Salmonella for epidemiological or investigational purposes.
In Chapter 9, guidance on sample collection, preparation, and
analysis is given. The method of ICMSF, 1978, pp. 163-172, is recommended. It is, in principle, in agreement with ISO Standard No. 6579.
REFERENCES
Clark, G. McC., Kaufmann, A.F., Gangarosa, E.J., and Thompson, M. 1973.
Epidemiology of an international outbreak of Salmonella agona. Lancet, ii,
490-493
Microbiological Criteria for Foods. Report of a Joint FAO/WHO Working Group on
Microbiological Criteria for Foods, Geneva, 20-26 February 1979. (Document
WG/Microbiol/79/1) WHO, Geneva
ISO (International Organization for Standardization). 1981. Microbiology - General Guidance
on Methods for the Detection of Salmonella. International Standard
6579. ISO, Geneva
Riemann, H., and Bryan, F.L. (eds). 1979. Food-Borne Infections and Intoxications.
Academic Press, New York
Silliker, J.H. 1980. Status of Salmonella - Ten years later. Journal of Food
Statutory Instruments. 1981a. The Diseases of Animals (Protein Processing) Order
1981. Statutory Instruments No. 676. HMSO, London
1981b. The Importation of Processed Animal Proteins Order 1981. Statutory
Instruments No. 677. HMSO, London
van Schothorst, M. and Brooymans, A.W.M. 1982. Effect of processing on microbial
contaminants in feeds. In Handbook of Nutritive Value of Processed Food.
Vol. 2. Animal Feedstuffs, ed. M. Rechcigl, Jr, 371-385. CRC Press, Inc., Boca Raton, USA
van Schothorst, M. and Oosterom, H. 1984. Enterobacteriaceae as indicators of
Good Manufacturing Practices in rendering plants. Antonie van Leeuwenhoek,
50, 1-6
WHO (World Health Organization). 1981. Report of the WHO/WAVFH Round Table
Conference on the Present Status of the Salmonella Problem (Prevention and Control).
Bilthoven, The Netherlands, 6-10 October 1980. (Document VPH/81.27) WHO, Geneva
15
Sampling plans for milk
and milk products
For purposes of applying microbiological criteria, dairy products may be

divided into two broad groups: (a) the more perishable (fresh) products
such as milk, cream, flavoured milk and skim milk drinks, fresh cheese
(cottage cheese), and fermented milks; and (b) the relatively stable
products having extended shelf-life under appropriate conditions of
storage, such as hard cheese, butter, dried milk products, ice cream
mixes, evaporated (canned) milk, and sterilized or ultrahigh-temperature
(UHT) milk (for fluid consumption).
A. RELATIVELY PERISHABLE PRODUCTS
It is impracticable to apply end-product microbiological criteria to detect

defective products in this category, which will have been widely
distributed and probably consumed before microbiological examinations
have been completed. However, periodic analyses (see below) are
commonly made by local authorities to decide whether or not to
continue to accept a product. Judgment is based on the performance
record of the producer or processor over a period of time, such as the 3
out of 5 rate of compliance with the standard as specified in the U.S.
Grade A Pasteurized Milk Ordinance (USPHS, FDA, 1978).
Health problems to be anticipated from pasteurized fluid milk and
related products are likely to be of local concern, and control is normally
in the hands of public health authorities. Pasteurized fluid milk rarely
crosses frontiers, so agreements between exporting and importing
countries appear as the best solution to the problem of control. Therefore, microbiological criteria are not proposed for pasteurized milk in
international trade. The report of the joint FAO/WHO Expert Committee on
Milk Hygiene (FAO/WHO, 1970) and ICMSF (1980b, pp. 482-486)
162
provide useful information on milk control practices and disease hazards.

Hygienic codes and methods of sampling and analysis are described in
publications of the International Dairy Federation (IDF, 1973, 1980b,
1980c) and the American Public Health Association (APHA, 1978), and in
the Grade A Pasteurized Milk Ordinance (USPHS, FDA, 1978).
The microbiological tests most often used are the aerobic plate count
(APC) and the coliform test. In addition, the phosphatase test is used
world-wide. If this enzyme is detected in allegedly pasteurized milk, it is
almost certain either that the milk was underpasteurized or that raw milk
had somehow entered the product.
B. RELATIVELY STABLE PRODUCTS
Milk products of this group are usually subject to test before being
distributed and consumed. Those that are currently common in international trade and are of concern because of microbiological hazards they
may present are dried milk and ripened cheese. While others may be
important in international trade, either they are unlikely to present a
microbiological health hazard or their safety and quality are not amenable
to measurement by end-product microbiological criteria. For example,
butter is an important commodity in international trade, but modern
methods of manufacture ensure a relatively long shelf-life under normal
commercial practice, and only rarely has illness been attributed to butter
(see ICMSF, 1980b, pp. 775-776). The safety of UHT milk, as of other lowacid canned foods (including evaporated milk), depends primarily upon
ensuring that the heat-processing is adequate and that container integrity
is maintained. Routine sampling and testing of end-product will not
provide adequate assurance of safety (see Chapter 10 and ICMSF, 1980a,
pp. 32-36). However, it is recommended that industry monitor
microbiologically, wherever applicable, critical control points in
processing operations. To some extent icecream mixes occur in
international trade. Ice cream itself is generally subject to control by local
authorities. If such controls do not exist it is recommended that the WHO
proposals (Christian, 1983, pp. 14-16) be followed for monitoring
purposes.
(a) Dried Milk
Several types and grades of dried milk-occur in commerce. Few outbreaks of food-borne disease due to dried milk have been reported in
Milk and milk products
163
recent years. This fact undoubtedly reflects the effectiveness of (a) the
extensive Salmonella testing of dried milk and subsequent action to
prevent distribution of contaminated product and (b) the critical control
measures now commonly practised in the industry (see ICMSF, 1980b, pp.
492-494; and Chapter 7, Control at Source) introduced following the
outbreaks of staphylococcal food-poisoning and salmonellosis that
spanned the years 195356 and 196465, respectively (Anderson and
Stone, 1955; Armijo et al., 1957; Collins et al., 1968). Nevertheless,
considering its intended use, dried milk is a sensitive product and may
present serious health hazards if not manufactured and used according to
good hygienic practices.
The various cases for concern are given in Tables 6 and 10 (pp. 43
and 74). The m and M values are given in Table 24. Case 2 for APC and
case 5 refer to indicator measurements and are recommended in
association with 3class plans (see Table 10). Coliform organisms in
dried milk may die out during storage. They should, nevertheless, be
absent or present only in low numbers in dried milk, even when the
product is freshly made. Thus, control of coliforms is necessary but their
absence is not sufficient to qualify a product as acceptable. Ideally the m
value for coliforms should be low (c. 1 per g) and, indeed, an m value of
<3 was suggested in the previous edition of this book. Since the current
method (3tube MPN) is imprecise in estimating such a low number, a
different value is indicated. It was desired to change this previous plan to
one having essentially the same consumers and producers risks but
using a higher m value. The adjustment can best be achieved by
decreasing c from 2 to 1 but increasing m to 10. n remains at 5 and M at
100 (see Table 24). It is anticipated that this case 5 sampling plan for
coliforms in dried milk will provide essentially the same assurance as the
plan previously proposed (n = 5, c = 2, m = <3, M = 100).
Enterotoxigenic Staphylococcus aureus is a potential hazard in dried
milk and should be absent or be present only in very low numbers. The
hazard of enterotoxin being present in the dried product is likely only
when deviations from Good Manufacturing Practice (GMP) provide
opportunities for sufficient growth of the organism either in the milk
before condensing or in the condensed milk prior to drying. The likelihood that conditions would permit such extensive growth is remote in
modern dry-milk manufacture because of industrys awareness of the
problem and of the measures taken to prevent it. There are, however,
situations where the product might become contaminated after drying.
Thus, if reconstituted milk is exposed to times and temperatures
TABLE 24
Sampling plans and recommended microbiological limits for dried milk and cheese
Limit per
gram
m
M
Product
Testd
Method
referencea
Dried milk
APC
118
3x104
3x105
Coliforms
126131
10
102
Salmonella,e
normal routine
163
10
11
12
2
2
2
5
10
20
0
0
0
0
0
0
Salmonella,e
for high-risk
population
163
10
11
12
2
2
2
15
30
60
0
0
0
0
0
0
Staph. aureus
220
2f
104
Cheese,
hard and
semi-soft
types b,c
Case
Plan
class
a Refers to page number in ICMSF, 1978, where the method is described. Use sample unit sizes
recommended in the methods, except where otherwise indicated.
b In addition, such cheese should be aged for 60 days at not less than 4.4C (40F) when made
from unpasteurized or unheated milk or curd (see Section B(b) 1).
c See also discussion on thermonuclease, Section B(b) 3.
d APC, see also IDF, 1970; coliforms, see also IDF, 1971; Salmonella, see also ISO, 1981.
e The case for Salmonella is to be chosen in accordance with whether the expected use of the
product will reduce, cause no change in, or increase concern (see Table 10, p. 74).
f See Section B(b) 3 for discussion of sampling plan.
sufficient to permit growth of staphylococci for example, temperature

abuse during mass feeding operations in hospitals, schools, institutions,
or in relief or emergency conditions the hazard of enterotoxin could
exist. For discussion of sampling plans for products used in this manner,
see Chapter 24, Formulated Foods.
The presence of Salmonella in dried milk is a serious hazard. While
current GMPS are effective in excluding Salmonella from dried milk,
contaminated lots are occasionally found (see ICMSF, 1980b, p. 494).
End-product testing for Salmonella is widely practised and is an
important factor in identifying contaminated lots and in preventing them
165
from entering trade. For routine sampling, in accordance with expected

use of the product cases 10, 11, and 12 would apply (see Table 10, p. 74).
Where the intended use is not known case 12 also would apply. More
stringent sampling plans for cases 10, 11, and 12 are indicated when the
product is intended for consumer groups of increased susceptibility, such
as infants, the aged, and infirm (see Table 8, p. 55; Chapter 6, Section B;
and Table 12, p. 78). The 2class plans which are recommended are
given in Table 24.
Recently, the Codex Committee on Food Hygiene proposed endproduct specifications to be included in a Code of Hygienic Practice for
Dried Milk (Codex 1981, 1983). Their proposed criteria are intended to
apply only when the product is used for other than high-risk populations
and are similar to those shown in Table 24 except that only one
specification for Salmonella (n = 15, c = 0, m = 0) is recommended
regardless of product use.
Microbiological guidelines were not recommended for inclusion in
the Code of Hygienic Practice for Dried Milk. It was concluded that
guidelines should be formulated by manufacturers taking into account the
need to meet limits prescribed in end-product criteria. ICMSF believes this
conclusion was prudent. It leaves the mechanism of compliance with the
code flexible and at the option of producers or manufacturers.
Foods prepared especially for infants and children may consist wholly
or in part of dried milks. These products are included in the Codex Code
of Hygienic Practices for Foods for Infants and Children (Codex, 1979)
and are considered in Chapter 24 (Formulated Foods).
(b) Cheese
Throughout the world diverse procedures are used to produce various
types of cheese from milk, which may be raw, pasteurized (phosphatasenegative), or heat-treated. For many types of cheese data on their
bacteriological quality, from the hygienic point of view, are few and often
have not been related to GMP. More data are necessary before meaningful
bacteriological limits can be recommended for every cheese variety.
Nevertheless, for some cheeses common in international trade, present
knowledge and experience are sufficient to warrant proposals for
microbiological criteria.
Although cheese has been the vehicle in outbreaks of food-borne
illness, epidemiological evidence places it among the relatively infrequent vehicles of transmission.
166
1. Aged cheese made from raw milk

Many varieties of cheese are made from milk which has not been heated
before the cheese-making process begins. Certain types, however, require
heat-treatment of the curd at some point during the cheesemaking
process, such as during cooking.
Several pathogenic bacteria, including Salmonella, Campylobacter,
and Listeria, can survive in cheeses made from raw milk. Some protection against the transmission of infectious disease, especially brucellosis, is achieved by requiring cheese made from other than
pasteurized milk to be aged for at least 60 days at a temperature of not
less than 4.4C.
2. Aged cheese made from milk heated during the cheese-making
process
Some types of cheese (e.g., Gruyere, Emmental, and Grana) are made
from raw milk, but the curd undergoes a heat-treatment at approximately
5355C for about 1 hour during the cheese-making process. These
heating processes reduce the bacteriological risk in such cheese.
3. Aged cheese made from pasteurized or heat-treated milk
Cheese is made also from pasteurized milk (phosphatase-negative) or
milk heated to a degree which may not provide a negative phosphatase
test but does substantially reduce the number of microorganisms. The
latter process is sometimes referred to as heat-treating or thermization
of milk; the heat-treatment approximates 65.5C for 15 to 20 seconds.
Such heat-treatments destroy or greatly reduce the number of heat-labile
organisms, including staphylococci and other pathogens (Zottola and
Jezeski, 1969; Zottola et al., 1969). The presence of excessive numbers of
such organisms in the final product results either from contamination of
the milk after heating and subsequent growth or from some faulty practice
or accident during the cheese-making process (e.g., starter failure).
The most serious health risk presented by cheese shipped in international trade is from staphylococcal enterotoxin. Available information
about staphylococci in cheese and factors influencing their growth and
production of enterotoxins is limited primarily to hard (Cheddar and
similar types, and Swiss) and semi-soft (Roquefort, Blue, Gouda, Brick,
and Mozzarella) varieties (ICMSF, 1980b, pp. 508-510; Devoyod et al.,
1968, 1969). The maximum growth of staphylococci occurs when whey
drainage is complete and generally prior to salting or brining. Later,
during the ageing period, they die slowly and only
167
a relatively small proportion of the earlier numbers will be recovered

when cultured. However, any enterotoxins produced will remain.
Growth of enterotoxigenic staphylococci to c. 15 million per gram
of cheese during cheese-making is necessary to produce amounts of
enterotoxin detectable by current assay methods (Tatini et al., 1971a,
1971b, 1973). Heat-treatments less severe than those indicated above
(65.5C for 15 to 20 seconds), when applied to milk before or after the
cheese-making process begins, may leave a residual population of
coagulase-positive staphylococci. In addition, insanitary equipment may
recontaminate the milk after heat-treatment. However, the most
important factor permitting the growth of staphylococci, and of many
other undesirable bacteria, during cheese-making is failure of the lactic
acid starter culture to develop normally (starter failure). This usually
results from infection of the starter culture by bacteriophage, or from the
presence of small amounts of antibiotics or other inhibitory substances
in the milk. Staphylococci found in ripened cheese probably represent
only a fraction of the population present in the unripened cheese. Thus,
the reliability of criteria for staphylococci in ripened cheese is limited,
and this fact must be taken into account when test results are interpreted.
The varieties of cheese that have been associated with outbreaks of
staphylococcal food-poisoning are Cheddar and similar types; recently
Swiss-type cheese was involved in a Canadian outbreak (Todd et al.,
1981). Outbreaks have been infrequent since 1965, perhaps because of a
better understanding of the behaviour of Staph. aureus in milk before
and during cheese-making, which has led to the application of effective
preventive measures (ICMSF, 1980b, pp. 510-513).
A viable count of 10,000 Staph. aureus per gram of cheese appears
to be well within technological achievement, and numbers below that
level in ripened cheese in international trade are likely to indicate that
numbers had not reached a level associated with detectable toxin (see
above comment on reliability). The sampling plan indicated for case 8
would seem to apply (see Table 10, p. 74). However, because of the
different rates of die-off of staphylococci in different cheese varieties
during ripening, specifying an m value in a 3class plan, which would
assure GMP, is not possible. Accordingly, a 2class plan is recommended
for case 8 as shown in Table 24.
The thermonuclease (TNase) test for assessment of staphylococcal
growth and possible presence of enterotoxins in cheese is recommended. The utility of this test has been shown by many authors (Cords
and Tatini, 1973; Tatini et al., 1975, 1976; Park et al., 1978; van
168
Schouwenburg-van Foeken et al., 1978; Stadhouders et al., 1980; Todd et

al., 1981; Ibrahim and Baldock, 1981). In some instances, however,
cheese containing enterotoxin did not contain detectable TNase (Todd et
al., 1981), which prompted concern whether cheese suspected of
containing enterotoxins, or of having contained large numbers of Staph.
aureus, should be released for consumption if TNase is not detected.
However, Ibrahim and Baldock (1981) and van Schouwenberg-van
Foeken et al. (1978) always detected TNase in extracts of cheese
containing enterotoxin. Further, Ibrahim and Baldock (1981) detected
TNase in cheese whenever the count of Staph. aureus reached c. 106 per
gram.
Although the reasons for the above differences are not clear, the
TNase assay used by Ibrahim and Baldock (1981) was more sensitive
than that used by others (perhaps as much as 30 times greater than the
procedure used by Todd et al., 1981). Differences in sensitivity of
enterotoxin assay procedures used and other factors may also contribute.
Nevertheless Ibrahim and Baldock (1981) concluded that TNase-positive
cheese should be considered unfit for further processing for human
consumption unless proved to be free from staphylococcal enterotoxins,
and that TNase-negative cheese containing Staph. aureus may be
processed for human consumption. Similarly, van Schouewenburg-van
Foeken et al. (1978) previously concluded that the presence of
enterotoxin is unlikely in cheese negative for TNase and that it can be
declared safe for consumption. However, cheese positive for TNase may
not always contain detectable enterotoxin. Therefore, TNase-positive
cheeses must always be subjected to enterotoxin assay before they are
declared unsafe for consumption. For further information on
staphylococci in cheese and measures for their control, reference to the
following should be helpful: IDF, 1973, 1980a; NCI, 1980.
During the latter part of 1971 and early 1972 numerous outbreaks of
gastroenteritis occurred in the United States caused by imported soft
cheese which contained pathogenic strains of Escherichia coli (enteropathogenic, enteroinvasive, or enterotoxigenic) (Marier et al., 1973). A
few earlier outbreaks in other countries had been reported (Costin et al.,
1964; Gaines et al., 1964). While the coliform problem in cheese is well
known, presence of these organisms in many cheese varieties is extremely
difficult to prevent completely. With some varieties, if coliforms are
present initially, it is virtually impossible to prevent their growth during
manufacture or during the ripening period. In several types of cheese E.
coli can even be considered characteristic. With the
169
exception of some strains of E. coli high populations of coliforms are

unlikely to present a health hazard. There is ample evidence that if
pathogenic strains of E. coli (PEC) are present early in the cheesemaking
process their numbers may increase to hazardous levels (Mourgues et al.,
1977; Frank et al., 1977 , 1978; Park et al., 1978) . However, in view of
the scarcity of evidence of recurring outbreaks due to PEC in cheese, and
the high cost of routine testing, it is doubtful that establishment of endproduct criteria for either coliforms or E. coli would be justified.
Accordingly, no sampling plan is proposed.
Nevertheless, every effort should be made to adhere to manufacturing
practices that will prevent coliform, especially E. coli, contamination and
subsequent growth (see ICMSF, 1980b, pp. 507, 510, 511). Toward this
end the direct plate method (DPM) of Anderson and BairdParker (1975) for
the enumeration of E. coli and as modified by Holbrook et al. (1980) to
include a resuscitation step should be useful. With this method results are
obtained in 24 hours rather than the 4 days by traditional procedures.
Thus, it would be feasible to examine on a routine basis soft and semisoft cheese for E. coli should subsequent developments deem it necessary.
Certainly, the application of guidelines for E. coli at critical control points
of cheese manufacture is advisable, and industry should be encouraged to
establish and use them as part of an internal microbiological safety and
quality assurance system.
For general instructions on collecting and handling field samples, see

Chapter 10. For specific procedures on sampling milk and milk products
consult the IDF Standard 50A proposed by the International Dairy
Federation (IDF, 1980c). IDF procedures not included in the general
procedures in Chapter 10 are outlined below, but it is recommended that
the full IDF text be studied by all persons who collect samples of dairy
foods.
(a) Dried milk
1 Where feasible, take original unopened packages.
2 When sampling bulk containers in the field or packages in the
laboratory, use the precautions described in Chapter 9, Section B (c),
when opening the containers. Take the sample from a point near the
170
centre of the container if possible. First remove the surface layer of

powder with a sterile instrument (e.g., a broad-bladed knife) and then
draw the sample with a sterile spoon or probe.
3 In the event of dispute concerning the bacteriological condition of the
surface powder in a package, take a sample of the product from this
area.
4 Preferably use brown glass sample containers, to exclude light.
(b) Cheese
1 Use one of the following three techniques, depending on the shape,
weight, and type of cheese: (a) sampling by cutting out a sector; (b)
sampling by means of a trier; (c) taking a complete cheese as a sample
unit. When a choice must be made between (a) and (b), method (b) is
often more practicable, especially with hard cheese of large size.
(i) Sampling by cutting out a sector. Using a knife with a pointed
blade, make two cuts radiating from the centre of the cheese. The size of
the sector thus obtained should be such that after removal of any
inedible surface layer, the remaining edible portion is more than twice
as much as will be required for analysis. Use this method for Edam and
Gouda cheese. It can also be used for semi-hard and soft cheese, and in
general for all cheese when sampling with a trier is impracticable.
(ii) Sampling by means of a trier. The trier may be inserted
obliquely towards the centre of the cheese once or several times into one
of the flat surfaces at a point not less than 10 to 20 cm from the edge.
From the plug or plugs thus obtained cut off not less than 2 cm of the
extremity containing the rind and use this piece to close the hole made
in the cheese. The remainder of the plug or plugs constitutes the sample
unit.
To prevent avoidable loss of product close the plug holes with great
care, especially with a large cheese, and if possible seal over with an
approved compound. This method is most suitable for hard and semihard cheese, such as Emmental and Cheddar.
The trier may be inserted perpendicularly into one face and passed
through the centre of the cheese to reach the opposite face. This method
is suitable for Provolone, Caciocavallo, and similar varieties.
The trier may be inserted horizontally into the vertical face of the
cheese, midway between the two plane faces, towards the centre of the
cheese. This method is suitable for cheese such as Tilsit, Cantal,
Roquefort, Pecorino, and Romano.
In the case of cheese transported in barrels, boxes, or other bulk
171
containers, or which is formed into large compact blocks, sampling may

be performed by inserting a 10 to 15 inch trier obliquely through the
contents of the container from the top toward the base. This method is
suitable for processed cheese and cheese foods.
(iii) Sampling by taking an entire or substantial portion of cheese.
Use this method for fresh cheese (for example, cottage, cream, or double
cream cheese), for soft cheese of small size, and for wrapped portions of
cheese packaged in small containers (e.g., some processed cheese and
various soft cheeses).
2 Immediately after sampling, place each sample unit (plug, sector, entire
small cheese) in a sterile container of suitable size and shape, and seal.
The sample unit may be cut into pieces for insertion into the container,
but do not compress or grind it.
3 Send or transport the sample units immediately to the laboratory, and
conduct the analysis as soon as possible, preferably on the same day. If
either dispatch or analysis must be delayed, place the containers in a
refrigerator at a temperature of 5 to 8C. This is especially important
with perishable cheese, such as soft cheese.
Precaution. If bacteriological, chemical, and/or organoleptic analyses are
to be made on the same sample unit, always remove the analytical units
for the bacteriological tests first.
REFERENCES
APHA (American Public Health Association). 1978. Standard Methods for the
Examination of Dairy Products, 14th ed. APHA, Washington, DC
Anderson, J.M., and Baird-Parker, A.C. 1975. A rapid and direct plate method for enumerating
Escherichia coli biotype 1 in foods. Journal of Applied Bacteriology, 39, 111-117
Anderson, P.H.R., and Stone, D.M. 1955. Staphylococcus food poisoning associated with
spray dried milk. Journal of Hygiene, 53, 387-397
Armijo, R., Henderson, D.A., Timothee, R., and Robinson, H.B. 1957. Food poisoning
outbreaks associated with spray-dried milk. American Journal of
Public Health, 47, 1093-1100
Christian, J.H.B. 1983. Microbiological Criteria for Foods. Summary of Recommendations
of FAO/WHO Expert Consultations and Working Groups 1975-1981. Document
VPH/83.54, WHO, Geneva
Codex. 1979. Codex Alimentarius Commission, Thirteenth Session, 1979: Report of the 15th
Session of the Codex Committee on Food Hygiene, Washington, D.C. 18-22 September
1978. ALINORM 79/13 Appendix v, Code of Hygienic Practice for Foods for Infants and
Children. Codex Alimentarius Commission, FAO, Rome
1981. Codex Alimentarius Commission, Fourteenth Session, 1981: Report of
172
17th Session of the Codex Committee on Food Hygiene, Washington, D.C.

17-21 November 1980. ALINORM 81/13. Appendix v, Extract from the Report of the
FAO/WHO Working Group on the Establishment and Application of Microbiological
Criteria for Dried Milk Products and Natural Mineral Waters. Codex Alimentarius
Commission, FAO, Rome
1983. Codex Alimentarius Commission, Fifteenth Session, 1983: Report of 18th i Session
of the Codex Committee on Food Hygiene, Washington, D.C. 22-26
February 1982. ALINORM 83/13. Appendix II, Draft Code of Hygienic Practice for
Dried Milk. Codex Alimentarius Commission, FAO, Rome
Collins, R.H., Trager, M.D., Goldsby, J.B., Boring, J.R., Cohoon, D.B., and
Bar, R.N. 1968. Interstate outbreak of Salmonella newbrunswick infection traced to
powdered milk. Journal of American Medical Association, 203, 838-844
Cords, B.R., and Tatini, S.R. 1973. Applicability of heat stable deoxyribonuclease assay
for assessment of staphylococcal growth and the likely presence of enterotoxin in
cheese. Journal of Dairy Science, 56, 1512-1519
Costin, I.D., Voiculescu, D., and Gorcea, V. 1964. An outbreak of food poisoning in adults
associated with Escherichia coli serotype 86:B7:H34. Pathologia Microbiologia, 27,
68-78
Devoyod, J.J., Bret, G., and Auclair, J.E. 1968. The microbial flora of Roquefort cheese. I.
Its evolution during cheesemaking and ripening. Le Lait, 48 (479/480), 613-629
1969. The microbial flora of Roquefort cheese. II. Staphylococci and micrococci. Le Lait,
49 (481/482), 1-20
FAO/WHO (Food and Agricultural Organization/World Health Organization). 1970. Third
Report of the Joint FAO/WHO Expert Committee on Milk Hygiene. WHO Technical
Report Series 453. WHO, Geneva
Frank, J.F., Marth, E.H., and Olson, N.F. 1977. Survival of enteropathogenic Escherichia
coli during manufacturing of Cambembert cheese. Journal of Food Protection, 40, 835842
1978. Behavior of enteropathogenic Escherichia coli during manufacture and ripening
of Brick cheese. Journal of Food Protection, 41, 111-115
Gaines, S., Achavasmith, Y., Thareesarvat, M., and Dungamani, C. 1964. Types and
distribution of enteropathogenic Escherichia coli in Bangkok, Thailand. American
Journal of Hygiene, 80, 388-394
Holbrook, R., Anderson, J.M., and Baird-Parker, A.C. 1980. Modified direct plate method
for counting Escherichia coli in foods. Food Technology in Australia,
32 (2), 78-83
Ibrahim, G.F., and Baldock, A.K. 1981. Thermostable deoxyribonuclease content and
enterotoxigenicity of Cheddar cheese made with sub-normal starter activity. Journal of
ICMSF. 1978. Microorganisms in Foods. 1. Their Significance and Methods of Enumeration, 2nd ed. University of Toronto Press, Toronto
1980a. Microbial Ecology of Foods. Vol. 1. Factors Affecting Life and Death of
Microorganisms. Academic Press, New York
1980b. Microbial Ecology of Foods. Vol. 2. Food Commodities. Academic Press, New
York
IDF (International Dairy Federation). 1970. Colony Count of Dried Milk and Whey
Powder. IDF Standard No. 49. IDF, Brussels
173
1971. Dried Milk and Dried Whey Count of Coliforms. IDF Standard No. 64.
IDF, Brussels
1973. Code of Hygienic Practice for the Manufacture of Cheese. Document No. 70. IDF,
Brussels
1980a. Behaviour of Pathogens in Cheese. IDF, Brussels
1980b. General Code of Hygienic Practice for the Dairy Industry. IDF, Brussels
1980c. Milk and Milk Products - Guide to Sampling Techniques. IDF Standard 50A. IDF,
Brussels
ISO (International Organization for Standardization). 1981. Microbiology - General
6579. ISO, Geneva (Also appears as BSI Document: BS5763: Part 4: 1982,
Methods for Microbiological Examination of Food and Animal Feeding Stuffs.
Part 4. Detection of Salmonella)
Marier, R., Wells, J.G., Swanson, R.C. Callahan, W., and Mehlman, I.J. 1973. An outbreak of
enteropathogenic Escherichia coli foodborne disease traced to imported French cheese.
Lancet, ii, 1376-1378
Mourgues, R., Vassal, L., Auclair, J., Mocquot, G., and Vandeweghe, J. 1977. Origin et
developpement des bacteries coliformes dans les fromages a pate molle. Le Lait, 57, 131149
NCI (National Cheese Institute). 1980. Major Policy Recommendations. Microbiological
Levels. NCI, Chicago
Park, C.E., El Derea, H.B., and Rayman, M.K. 1978. Evaluation of the staphylococcal
thermonuclease (TNase) assay as a means of screening foods for growth of staphylococci
and possible enterotoxin production. Canadian Journal of Microbiology, 24, 1135-1139
Stadhouders, J., Hassing, F., and Galesloot, T.E. 1980. A rapid and simple method for the
detection of Staphylococcus aureus thermonuclease in cheese. Netherlands Milk and Dairy
Journal, 34, 199-204
Tatini, S.R., Cords, B.R., and Gramoli, J. 1976. Screening for staphylococcal enterotoxins in
food. Food Technology, 30 (4), 64, 66, 70, 72, 73
Tatini, S.R., Jezeski, J.J., Olson, J.C. Jr., and Casman, E.P. 1971a. Factors influencing the
production of staphylococcal enterotoxin A in milk. Journal of Dairy Science, 54, 312-320
1971b. Production of staphylococcal enterotoxin A in Cheddar and Colby cheese. Journal
of Dairy Science, 54, 815-825
Tatini, S.R., Soo, H.M., Cords, B.R., and Bennett, R.W. 1975. Heat stable nuclease for
assessment of staphylococcal growth and likely presence of enterotoxins in foods. Journal
of Food Science, 40, 352-356
Tatini, S.R., Wesala, W.D. Jezeski, J.J., and Morris, H.A. 1973. Production of staphylococcal
enterotoxin in Blue, Brick, Mozzarella and Swiss cheese. Journal of Dairy Science, 56,
429-435
Todd, E., Szabo, R., Robern, H., Gleeson, T., Park, C., and Clark, D.S. 1981. Variation in
counts, enterotoxin levels and TNase in Swiss type cheese contaminated with
Staphylococcus aureus. Journal of Food Protection, 44, 839-848
USPHS, FDA (United States Public Health Service, Food and Drugs Administration). 1978.
Grade A Pasteurized Milk Ordinance, 1978. Recommendations of the US Public Health
Service / Food and Drug Administration. US Government Printing Office, Washington,
DC
174
van Schouwenburg-van Foeken, A.W.J., Stadhouders, J., and Jans, J.A. 1978. The
thermonuclease test for assessment of the growth of coagulase positive staphylococci in
Gouda cheese with a normal acidity development. Netherlands Milk and Dairy Journal,
32, 217-231
Zottola, E.A., and Jezeski, J.J. 1969. Comparison of short-time holding procedures to determine
thermal resistance of Staphylococcus aureus. Journal of Dairy Science, 52, 1855-1857
Zottola, E.A., Jezeski, J.J., and Al-Dulaimi. 1969. Effect of short-time subpasteurization
treatment on the destruction of Staphylococcus aureus in milk for cheese manufacture.
Journal of Dairy Science, 52, 1707-1714
16
Sampling plans for
eggs and egg products
Eggs and egg products are important commodities in international trade.

Exports of shell eggs exceed 20 million boxes per annum, and exports
of egg products approach 100,000 tonnes. Of the latter, 8590% is in
liquid or frozen form and of this some 60% is whole egg (IEC, 1980).
For the purposes of this chapter, eggs and egg products are subdivided
into (I) shell eggs and (2) liquid, frozen, and dried eggs. Mention is also
made of foods containing eggs.
The introduction to the subject in ICMSF, 1980, p. 521, describes the
properties of the whole egg in the shell and the methods of preservation
for the three categories of egg products. The whole egg in the shell is
self-protective unless abused by excessive exposure to contamination at
the laying stage and gross temperature and humidity changes during
storage. Immediately after the shell is broken, by whatever means, the
liquid egg is exposed to contamination from hands, equipment, and the
external surface of the shell itself.
The most important pathogen likely to be present in liquid, frozen,
and dried eggs is Salmonella. This organism may occasionally be
present within the egg at the time of laying, but much more commonly it
contaminates the liquid egg from the external surface of the shell during
breaking of the shell. There are then the dangers of (a) eating the
contaminated raw product, (b) cross-contamination from raw or
unpasteurized egg products to foods prepared in the bakery and confectionery trades or in food-service kitchens, (c) direct contamination
when an unpasteurized dried-egg product is incorporated into blended
foods, and (d) inadequate heating of egg products (e.g., those used for
infant feeding) before consumption.
Mandatory pasteurization of liquid egg should eliminate these hazards. However, any salmonellae surviving pasteurization or contami-
176
nating the product subsequently may be expected to survive freezing or

drying. Thus, salmonellae remain a potential hazard in egg products,
because of the opportunities for multiplication that may arise from
time/temperature abuse after thawing or rehydration.
Manufactured foods containing egg, such as certain pastas, puddings, and frostings, may also constitute a Salmonella hazard. Although
prepared with egg of acceptable bacteriological quality, a poorly
controlled process may permit substantial multiplication of contaminating salmonellae before the aw falls below the level limiting their
multiplication.
Pasta may be a vehicle for the growth of staphylococci and the
formation of enterotoxin. Growth may occur during mixing if wet spots
develop and are allowed to persist in the mixer, resulting in detectable
levels of enterotoxin in commercial products.
A. SAMPLING PLANS
(a) Shell eggs

Sampling plans are not proposed for shell eggs.
(b) Liquid, frozen, and dried eggs
The cases for the sampling plans are discussed in ICMSF, 1980, with
respect to liquid and frozen eggs (p. 560) and dried eggs (p. 565).
The aerobic plate count (APC) may be regarded as an indicator of the
quality of the raw material before processing and also of the possibility of
contamination after processing; it may be usefully applied at port-ofentry, manufacturing, and retail levels. In the criteria proposed, the APC is
omitted for dried egg albumen because (a) there is always extensive
growth of Group D streptococci during desugaring, even when the
enzymic method is used; these bacteria are more heatresistant than
organisms of concern such as Salmonella and will initiate growth at the
pH of egg albumen (c. 9.3); and (b) the pasteurization process that can be
applied to liquid albumen must be less severe than that used for liquid
whole egg and yolk.
The criteria for APC listed in Table 25 differ slightly from those in
ICMSF, 1978, pp. 111-115, 120. A major report (FAO/WHO, 1975), which
included a survey of national criteria for egg products, concluded that
there was no justification for setting different plans and limits for
Eggs and egg products
177
dried and frozen egg and that the ICMSF (1978) m value of 104 was too
strict. The APC criteria in Table 25 conform with the recommendation of
the above report, which is now part of the Code of Hygienic Practice for
Egg Products (Codex, 1978), and with the best estimate based on the data
available.
The use of coliform counts is intended for evidence of (a) insufficient
heat-treatment or (b) post-pasteurization contamination. This test is more
applicable at the point of manufacture than at port of entry, particularly
when there is no record of manufacturing procedures. In ICMSF, 1978,
coliform tests were included in criteria for dried (p. 111), but not for
frozen (p. 120), egg products. They have been included in Table 25 for
both products at the level suggested previously for dried egg, as they
serve the same purpose in both types of product. This is consistent with
Codex, 1978.
The Salmonella criteria suggested in ICMSF, 1978, were stricter for
frozen egg whites (case 11) (p. 120) than for frozen whole egg (case 10)
(p. 120) because the former are often not adequately cooked to destroy
Salmonella. However, all of the egg products considered in this chapter
may have a range of food-ingredient applications that will influence the
degree of concern in different ways. Thus the case chosen should be
carefully considered in relation to the use intended for each product. For
normal routine sampling for Salmonella in accordance with intended use,
case 10, 11, or 12 would apply (see Table 10, p. 74). The age and
condition of the intended consumer should also be considered. The very
young, the elderly, and the infirm are more susceptible to infection so
more stringent sampling plans should be applied to egg products destined
for such customers (see Table 12, p. 78).
The criteria suggested in Table 25 taking the Salmonella sampling
plan as case 11 are consistent with those of the Codex Alimentarius
Commission (Codex, 1978).
Occasionally small numbers of staphylococci may be found in salted
yolks, presumably because of the selective nature of the highsalt
menstruum. However, no criterion is suggested for staphylococci in this
product because there is no evidence of hazard. Few countries include
staphylococcal counts in their criteria for egg products (FAO/WHO, 1975).
Physical measurements of efficacy of pasteurization, such as the
destruction of a-amylase in whole-egg melange, are practised with good
effect in some countries and in particular in the United Kingdom.
178
TABLE 25
Sampling plans and recommended microbiological limits for pasteurized liquid, frozen, and
dried egg products
Limit per
gram
m
Testa
Method
referenceb
APC
118
5x104
106
Coliforms
126131
101
103
Salmonella,d
normal routine
163
10
11
12
2
2
2
5
10
20
0
0
0
0
0
0
Salmonella,d
for high-risk
population
163
10
11
12
2
2
2
15
30
60
0
0
0
0
0
0
Case
Plan
class
NOTE: SAMPLING PLANS AND RECOMMENDED LIMITS FOR FOODS CONTAINING EGG:
Sampling plans for indicator organisms need to take account of microbial

contributions of non-egg ingredients. Where salmonellae are considered a hazard,
the above criteria for salmonellae are appropriate.
a APC, see also ISO 4833 (1978c); coliforms, see also ISO 4831, 4832 (1978a, b); Salmonella,
see also ISO 6579 (1981).
b Refers to page numbers in ICMSF, 1978.
c Omit APC for dried egg white.
d The case for Salmonella is to be chosen in accordance with whether the expected use of the
product will reduce, cause no change in, or increase concern (see Table 10 on p. 74 and
ICMSF, 1980, Chapter 19, pp. 521-566).
e For food intended for high-risk populations, the case remains as indicated in Table 10,
but n is increased to increase stringency.
However, it should be emphasized that, for the amylase test to be a

reliable indicator of the adequacy of the process, the pasteurization
temperature must reach 64.5C (for 2.5 minutes) or higher (UK Ministry of
Health, 1963).
Some unpasteurized eggs are shipped between countries, and represent a serious hazard as vehicles for the dissemination of salmonellae.
The international shipment of unpasteurized egg products should be
discouraged and their sale prohibited. No recommended limits are
proposed for these products.
Eggs and egg products
179
(c) Foods containing egg

There is a wide range of manufactured foods which contain egg products (ICMSF, 1980, p. 524). It is not possible to generalize in terms of
sampling plans for indicator organisms, because of the major contributions that non-egg ingredients may make to these populations. The
most important pathogens in such foods are likely to be salmonellae
and, where they are considered to be a hazard, the criteria for salmonellae in Table 25 are appropriate, with due regard being given to
the case. Streptococcus pyogenes and Staphylococcus aureus are also
of concern, particularly in foods prepared in food-service establishments and homes (Riemann and Bryan, 1979).
B. SAMPLING PROCEDURES
Reference should be made to Chapter 9, and in particular to those

sections dealing with the sampling of frozen, liquid, and powdered
products. The sample will usually consist of unopened containers, such
as hermetically sealed cans or consumer packs (e.g., dehydrated egg
products). Select the required number of sample units, which may be
individual or composited sample units from individual cans or packs.
(a) Frozen egg
Drill diagonally through the frozen material in the can after opening
with aseptic precautions. Use a drill or brace and sterile bit, and scrape
about 50 g of the drillings into a sterile container with a sterile spoon;
then close container and can. Alternatively a modified plastic funnel
may be used with the electric drill to collect samples (Adams and
Busta, 1970). Thaw the sample unit for not more than 90 minutes at
room temperature, or place the required amount of analytical unit
directly into an homogenizer with pre-enrichment medium in the case
of Salmonella testing, or into appropriate diluent for other tests.
(b) Dried egg
Open the original container or laboratory container (waxed carton or
plastic bag) taking aseptic precautions, and spoon the appropriate
amount of analytical unit directly into pre-enrichment medium in the
case of Salmonella testing, or into appropriate diluent for other tests.
180
REFERENCES
Adams, D.M., and Busta, F.F. 1970. Simple method for collection of samples from a
frozen food. Applied Microbiology, 19, 878 (only)
Codex (Codex Alimentarius Commission). 1978. Recommended International Code of
Hygienic Practice for Egg Products. Addendum 1-1978 to CAC/RCP 15-1976.
FAO, Rome
Microbiological Specifications for Foods. Report of a Joint FAO/WHO Expert
Consultation, Geneva, 7-11 April 1975. (Document EC/Microbiol/75 Report 1) FAO,
Rome
ICMSF. 1978. Microorganisms in Foods. 2. Sampling for Microbiological Analysis:
Principles and Specific Applications. University of Toronto Press, Toronto
IEC (International Egg Commission). 1980. Six-Monthly Statistics Bulletin, October 1980,
London
ISO (International Organization for Standardization). 1978a. Microbiology - General
Guidance for the Enumeration of Coliforms - Most Probable Number Technique at 30C.
ISO 4831-1978. ISO, Geneva
1978b. Microbiology - General Guidance for Enumeration of Coliforms - Colony Count
Technique at 30C. ISO 4832. ISO, Geneva
1978c. Microbiology - General Guidance for Enumeration of Microorganisms -Colony
Count Technique at 30C. ISO 4833-1978. ISO, Geneva
1981. Microbiology - General Guidance on Methods for the Detection of Salmonella.
International Standard 6579. ISO, Geneva
Riemann, H., and Bryan, F.L. (eds). 1979. Food Borne Infections and Intoxications, 2nd ed.
Academic Press, New York
UK Ministry of Health. 1963. The Liquid Egg (Pasteurization) Regulations. HMSO, London
17
Sampling plans for
fish and shellfish
Fish and shellfish are taken in nearly all the salt and fresh waters of the
world. The annual catch is estimated at about 71 million tonnes, of
which 72% is for direct human consumption (FAO, 1979). The catch is
made up of a great diversity of species (800 species are listed by FAO),
which range over several phyla from primitive molluscs to marine
mammals. A large proportion of the fish catch enters international trade,
moving world-wide and involving products from both developing and
industrialized countries. This is exemplified by the fact that 162
countries in 1978 exported 32% of their total production of 17.63
million tonnes of seafood (FAO, 1979).
A. PUBLIC HEALTH ASPECTS OF FISH AS HUMAN FOOD
Despite the fact that fishery products are of diverse origin and include
many species of animals, the composition of the naturally occurring
(autochthonous) microflora is similar in most cases. The microorganisms most commonly encountered are typical of the free-living forms
found in water and sediment and rarely include any species of mammalian pathogens (ICMSF, 1980, pp. 575-578). Hence fish caught in
waters not polluted by human or animal wastes are free from intrinsic
microbiological hazard when handled according to good commercial
practice.
Indeed fish and other free-swimming marine animals do not usually
carry those organisms generally considered to be typical of the mammalian microflora, including Escherichia coli, the faecal coliforms,
and enterococci. The presence of human enteric organisms on marine
food products is clear evidence of contamination from a terrigenous
source.
182
Fish and shellfish products are a minor source of bacterial foodborne disease in North America, the United Kingdom, and Australia
(Todd, 1978), but there is a continuing high relative incidence of
bacterial food-borne disease from fish products in Japan and probably
in Southeast Asian countries where fish are commonly eaten raw or
with little cooking (Sakazaki, 1979). In most cases the aetiologic agent
is a vibrio though some botulism due to improperly prepared products
also occurs in Japan, Poland, and the USSR and in other North Pacific
rim countries including northern Canada and the USA (Hobbs, 1976).
A number of different Vibrio species have been implicated in foodborne illness resulting from eating seafood. These include V. parahaemolyticus, V. vulnificus, V. cholerae, V. fluvialis, and V. alginolyticus. Only V. alginolyticus occurs consistently on the living animals,
and the actual involvement of this organism in human gastroenteritis is
somewhat doubtful. Vibrio parahaemolyticus is relatively common in
warm-water areas and is a cause of gastroenteritis in coastal regions of
India and in Japan and countries of Southeast Asia. Disease due to
Vibrio species is usually associated with situations where large populations of living Vibrio species are present on Seafoods, either in the
raw state where fish and shellfish are eaten uncooked or in recontaminated cooked products which are stored at temperatures that permit
growth. However, infective population levels of vibrios can occur on
freshly harvested marine animals, particularly where water quality is
poor and temperatures are high. Experiments in the USA in connection
with the investigation of cholera from ingestion of cooked crabs suggest
that normal cooking practices may not always destroy vibrios (CDC,
1981).
Non-proteolytic strains of Clostridium botulinum types E, B, and F
occur in small numbers in the intestines of fish and may be transferred
to flesh during evisceration or other primary processing. Botulism has
been caused by ingestion of botulinal toxin produced in improperly
fermented products in Japan, in marine products in arctic North
America, in hot smoked fish held at improper temperatures, and in
imperfectly canned seafoods (Sakaguchi, 1979).
Raw or processed seafoods are in general excellent substrates for
the growth of most common bacterial agents of food-borne disease if
held at improper temperatures. It is important therefore to avoid contamination of these foods during preparation and storage and to hold
them at chill temperatures.
There is a particular problem in the case of sessile molluscs such as
oysters, mussels, and clams which grow in and are harvested from
Fish and shellfish
183
inshore waters most likely to be contaminated by human or animal

wastes. These organisms feed by passing large volumes of water over
selective mouthparts which strain out small particles from the water.
They select and concentrate bacteria from the environment and when
held in polluted waters may become dangerous vehicles of both pathogenic bacteria and viruses. Contaminated shellfish may be a source of
hepatitis A virus as well as of more generally recognized bacterial
enteric pathogens (ICMSF, 1978a, p. 65).
At present the only practical control measure is through bacteriological testing of growing-area waters for coliforms and E. coli, supplemented by testing of harvested shellfish for aerobic plate count (APC),
coliforms, and E. coli; shellfish harvesting is closed when bacterial
counts rise unusually (FAO/WHO, 1978). This system has proved to be
very effective in countries that use it; and food-borne disease, including
viral hepatitis, caused by eating shellfish is almost invariably due to
animals taken from unregulated or closed beds. However, poor
correlation has been reported between virus incidence in shellfish and
indicator bacteria counts in water, and a direct test method for hepatitis
A virus would be desirable. In cases where the level of contamination is
low, relaying in clean waters or active depuration will effectively clean
bivalves of undesirable bacteria although virus and some vibrios may
persist. The most effective control is mandatory heat-treatment of
bivalve molluscs to destroy bacteria and viruses, but this is often
impracticable. In any case shellfish in commerce should be tested for
enteric indicator bacteria as an index of bacterial quality although this
will not assure safety if the animals are collected from unregulated beds
since the test method cannot be expected to give reliable indications of
the presence or absence of viruses.
Fish and shellfish poisonings are due to toxins in the flesh of
animals either intrinsically present or derived from the food they eat.
This is true for puffer fish poisoning (caused by tetradotoxin), ciguatera
(caused by ciguatoxin and possibly other toxins), and paralytic shellfish
poisoning (PSP) (caused by saxitoxin and other toxins). Scombroid toxin
is an exception and is apparently due to products of bacterial growth on
tunas, mackerels, and related genera after their death. Histamine is most
commonly identified as the toxin. These conditions generally require
chemical or biological analysis for toxic substances, and therefore no
recommendations for criteria are included. However, individuals and
agencies concerned with the safety of seafoods in international
commerce should be aware of the potential hazard of the toxicants and
should include toxin tests where appropriate.
184
B. THE BASIS FOR SPECIFIC SAMPLING PROPOSALS
The number of outbreaks of food-borne disease from consumption of

seafood during the years since the first volume of the book was published has, apart from chemical fish poisoning, shown little change. A
higher incidence of bacterial and viral disease from fish and shellfish
continues to occur in India, Japan and Southeast Asian countries than in
Europe or North or South America. The microorganisms of principal
concern on a world-wide basis appear to be vibrios and C. botulinum,
with hepatitis A virus and enteric bacterial pathogens important in
molluscan shellfish. The more common food-borne pathogens such as
Salmonella, Staphylococcus aureus, and C. perfringens, which have
been reported in recent years as causes of fish-borne disease, are not
typical environmental contaminants but generally gain access to the
food during processing or food-service operations.
Most seafood products are handled and stored under conditions
which preclude growth and toxin production by C. botulinum. Botulism
has been particularly associated with improperly processed canned fish,
hot smoked fish, and certain pickled and fermented products (Hobbs,
1976). The increasing use of modified-atmosphere storage to extend the
shelf-life of fish may increase the risk of growth of C. botulinum
(Eklund, 1982). In the case of products at risk, proper processing (e.g.,
heat-treatments to destroy the organism) and control of storage conditions to prevent germination and outgrowth of C. botulinum spores are
necessary to ensure safety of the product. The types of C. botulinum of
greatest concern are the non-proteolytic psychrotrophic types B, E, and
F. While C. botulinum is recognized to be a significant pathogen
associated with fish, no criteria are proposed. The detection of C.
botulinum on a food does not signal an immediate hazard but only
indicates potential risk. Microbial counts or simple determination of the
presence of C. botulinum spores serve no useful purpose. Even in
botulism outbreaks toxin tests, not counts of cells or spores, are
necessary to confirm involvement of the suspect food.
For the purposes of applying microbiological criteria, seafood
products are grouped as follows:
(a) Fresh, frozen, and cold smoked fish, including fresh and frozen
(including shipboard frozen) marine and freshwater fish, fish
blocks, comminuted fish blocks, scallops, cold smoked fish.
(b) Precooked breaded fish, including breaded fillets, portions, fish
sticks, and fish cakes.
Fish and shellfish
185
(c) Frozen raw crustaceans, including frozen raw shrimp, prawns, and
lobster tails.
(d) Frozen cooked crustaceans, including frozen cooked shrimp (peeled
or unpeeled), prawns, and lobster tails.
(e) Cooked crabmeat: cooked or pasteurized, chilled or frozen, whole
crab, or picked meat.
(f) Molluscs: fresh and frozen mussels, clams, oysters in shell or shucked.
Note that seafoods not included in these categories are omitted because
insufficient microbiological data are available. This grouping differs from
that in the previous edition of this book (ICMSF, 1978b) in that some
product groups have been combined because the criteria were the same.
Escherichia coli has been accepted as a better criterion of potentially
hazardous contamination than faecal coliforms, and is now used here
for fish and shellfish. This change conforms with criteria used for other
commodities and makes possible the use of rapid and more objective
analytical methods based on membranes (Holbrook et al. 1980; Sharpe et
al. 1981).
Vibrio parahaemolyticus, Salmonella, and Staphylococcus in (a), (b)
and (c) have been made optional criteria to be applied in special
circumstances, and their appropriate use is indicated in footnotes to
Tables 26 and 27 below.
In general there has been little change in the assignment to cases
within the new groupings except that V. parahaemolyticus is allocated to
case 7, 8, or 9 rather than to case 10, 11, or 12 in accordance with the
revision on Table 6 (p. 43). A 3class plan is now proposed for V.
parahaemolyticus.
Neither hot smoked fish nor semi-conserved fish products have been
included. They have been associated with a number of outbreaks of foodborne illness including botulism, staphylococcal intoxication, and
salmonellosis, and in the case of hot-smoked mackerel scombroid
poisoning, but there is insufficient evidence that microbiological criteria
would contribute significantly to preventing outbreaks attributed to those
products.
Dried fish are important to world trade. These are produced in both
temperate and tropical regions of the world, often under conditions of
poor hygiene, and are frequently shipped with minimum protection
against casual contamination. Unfortunately, little has been published on
the microbiology of dried fish or their involvement in food-borne illness.
Therefore, microbiological criteria are not proposed.
186
The selection of criteria for seafoods has been made with due
attention to variability factors such as goegraphical source and handling, storage, and processing practices (ICMSF, 1980, pp. 567-605;
FAO/WHO, 1976, 1978; WHO, 1974) and with consideration of existing
national and international criteria (Codex, 1977).
Microbiological criteria for molluscs are dependent on proper control of the growing waters and harvesting process. A list of microbiological criteria used in a number of countries is presented in the
Codex Alimentarius Commissions Recommended International Code
of Hygienic Practice for Molluscan Shellfish (FAO/WHO, 1978), which
includes the U.S. standards.*
In selecting criteria in Table 27 we have been guided by actual
levels of counts reported in seafood products in international trade
particularly products moving in significant quantity and the reported
occurrence of food-borne disease from various fish products.
A realistic sampling plan should allow the inspecting agency to
make the necessary examinations without excessive expenditure of
manpower and expense. Concern with expenditure is particularly important with fish and fishery products where there is a large and continuing flow of yearly shipments with a great diversity in the size,
weight, and type of package. Often the history of handling, processing,
packaging, storage, and transportation is unknown. In addition, the
perishability of many fish products requires that tests be completed
quickly in order to minimize delay within the usual trade channels. With
fish and fishery products, counts at 20 to 25C are probably the most
useful indicators of incipient spoilage (utility) because of the
psychrotrophic nature of the spoilage flora. Hence we recommend that
the APC be carried out at 25C. The APC at 35C is frequently of the
order of one-tenth of the count at 25C (Liston, 1957; Shewan, 1977).
* Maximum APC levels are 100,000 per gram in Denmark and 500,000 per gram in the
USA; there are no APC criteria in France, Italy, Netherlands, and the UK.
DENMARK APC (20C): n = 10, m = 100,000 per gram (2class); E. coli I: n = 10, absent in
1/5 ml; Salmonella: n = 10, absent in 1/5 ml.
FRANCE E. coli: n = 510, absent in 1 ml macerated shellfish or up to 2 E. coli per ml
for shellfish consumed cooked; Salmonella: n = 510, 25 g tested, no Salmonella
in 25 g.
ITALY E. coli: n = 10, <6 E. coli per gram (by MPN).
NETHERLANDS E. coli: n = 10, for hold pending investigational sampling, m = 4 per ml,
for rejection m = 15 per ml.
UK A commonly operated guideline <2 E: coli per ml.
USA APC (35C): <500,000 per gram; faecal coliform: <2.3 per gram.
Fish and shellfish
187
Although in certain countries Good Commercial Practice (GCP) is

commonly attained by plants processing Seafoods, no information is
available for a significant proportion of producers whose products are in
international trade. Although the sampling procedures that are recommended attempt to take cognizance of this fact, the first consideration
has been the safety of the consumer. It is, however, important that safety
tests (e.g., for Salmonella) should not be confused with indications of
shelf-life or quality such as the APC. It should be recognized that the
bacterial counts on raw fish vary very widely both geographically and
because of differences in the duration of shipboard storage, which
greatly affects counts on processed products and makes the application
of criteria with unreasonably low counts with little scope for variation
undesirable. Several items of fish and fishery products in international
trade have high APCs, but there is no evidence that they have given rise
to any corresponding health hazard. Accordingly, the recommended APC
limits reflect present accepted practice in the fish industry. The faecal
coliforms and E. coli are particularly useful as indicators of
contamination and mishandling of seafoods since the organisms are
absent from fish and crustaceans at the time of capture (except in grossly
polluted waters). Moreover, fish and shellfish should be held at
temperatures below those which support growth of E. coli or faecal
coliforms when GCP is followed (see FAO/WHO, 1976, 1978). Thus the
presence of detectable numbers of these organisms indicates
contamination while relatively large numbers suggest temperature abuse
in product handling.
In view of the increasing contamination of inshore marine waters
with domestic sewage and hence of the increased probability of contamination of fish and crustaceans taken in such waters, examination for
enteric viruses might be as important as for Salmonella (Clarke and
Chang, 1959; Murphy et al., 1979; Bryan, 1980). However, the lack of
appropriate methodology prevents this (Cliver, 1979).
C. SAMPLING PLANS
The sampling plans selected for seafood products are outlined in Table
26. Criteria for APC and E. coli counts are recommended in all instances.
Staphylococcal criteria are recommended for crabmeat and frozen
cooked crustaceans because staphylococcal food-poisoning is a
recognized hazard in these products. Additional criteria for V. parahaemolyticus, Salmonella, and Staphylococcus are recommended for
breaded or precooked products, warm-water products, and molluscs.
188
This procedure is recommended to avoid unnecessary routine application

of expensive and time-consuming tests except where the known or suspected origin of the seafood products suggests a hazard may exist.
For economic reasons the number of sample units to be examined, n,
is limited to five except for Salmonella and, in one instance, for V.
parahaemolyticus. Even five samples would tax the laboratory facilities
for routine testing available in many areas which import large quantities
of fish. Testing for salmonellae is recommended for some crustaceans
and fish harvested from warm-water areas, but should be applied on a
selective basis taking into account the origin of and processing applied to
the products.
Similarly, while V. parahaemolyticus is of widespread occurrence in
inshore marine waters, sediments, and marine animals, it is usually
present in significant numbers only on seafoods taken from tropical or
subtropical waters or from temperate-zone waters in summertime. Routine testing for V. parahaemolyticus enumerates both non-pathogenic and
enteropathogenic strains, but it is unusual, even in foods implicated in
outbreaks of food-borne illness, to find high numbers of the pathogenic
serotype.
Except for crabmeat, testing for Staph. aureus is also recommended
only in special circumstances, because most seafood products carry few
staphylococci. Where testing is indicated the recommendations are based
on demonstrated hazard for that category of food under the conditions
specified. It should be noted that counts on selective media for both
staphylococci and V. parahaemolyticus are often erratic below, or near,
102 per gram (see Table 27, footnote f). Reproducibility improves at
higher numbers.
An aerobic plate count is recommended for all products because of
its usefulness as an indicator of utility and the condition and length of
storage of products prior to stabilizing processes such as freezing. Most
aquatic animals at the point of harvest have counts in the region of 102
105 organisms per gram. There are exceptions including some tropical
shrimp, molluscs, and freshwater fish. An increase in APC to levels in
excess of 106 per gram is usually indicative of long storage at chilling
temperatures or temperature abuse prior to freezing. Thus APC is
indicative of general quality and to a lesser extent of handling and
storage procedures.
Escherichia coli testing is also included in all cases, and it is noted
that faecal coliform counts may be used instead where this is the
preferred method. The usefulness of this test for seafoods was indicated
earlier and is based on the fact that these organisms should be absent
TABLE 26
189
Allocation of fish and shellfish products to case

Product
Condition
Test
Case
Fresh and frozen fisha
Usually cooked before eating; hazard

reduced or low
APC
E. coli
1
4
Precooked breaded fish

productsb
Usually heated before eating but

heating may be insufficient to destroy
pathogens; low health hazard
APC
E. coli
2
5
Frozen raw crustaceansc
Usually cooked prior to eating; hazard

reduced
APC
E. coli
1
4
Frozen cooked
crustsceansd
Little or no cooking before eating;

frequently held under conditions after
thawing which cause increased
moderate hazard
APC
E. coli
Staph. aureus
2
5
8
Cooked, chilled, frozen

crabmeate
Not cooked prior to consumption;

known health hazard; hazard
increased
APC
E. coli
Staph. aureus
3
6
9
Fresh and frozen bivalve

molluscsf
Often eaten raw or with light cooking

only
APC
E. coli
3
6
a For fish known to derive from inshore or inland waters of doubtful bacteriological quality,
particularly in warm-water areas and where fish are to be eaten raw, it may be desirable to
test for Salmonella by case 10 and V. parahaemolytlcus by case 7. Smoked fish may be
tested for Staph. aureus by case 7.
b Staph. aureus may also be tested for by case 8, for cooked fish products likely to be
mishandled.
c Staph. aureus may be tested in breaded products using case 7. Salmonella and V.
parahaemolyticus tests may be applied to products from warm-water areas and V.
parahaemolyticus tests in products from temperate regions in summer using case 8, if
likely to be eaten raw.
d In view of the storage conditions cases 3, 6, and 9 should apply. However such a load of
microbiological testing would be impracticable for most fish-testing laboratories. Also,
some survey data indicate that cases 3, 6, and 9 would reject an unacceptably large
proportion of the product in international trade. Experience suggests that the criteria
recommended above would assist in maintaining hygienic standards. In the case of
cooked crustaceans originally harvested from waters and processed in regions of known
high environmental hazard, Salmonella and V. parahaemolyticus may be tested for
according to cases 11 and 8 respectively.
e Crabmeat from animals harvested from waters above 15C may be tested for V. parahaemolyticus using case 9.
f I. These criteria are to be used only for molluscs from approved harvesting areas where
waters are free from enteric bacteria or virus contamination and there is no significant
contamination by toxic metals or toxic and carcinogenic chemicals which may be
accumulated by animals. Molluscs from non-approved areas should enter trade only after
processing by a treatment to destroy enteric bacteria and viruses and may be tested for
dangerous chemicals before distribution.
II. Salmonella should be tested for when there is concern for bacteriological safety using
case 12, and V. parahaemolyticus should be tested for in molluscs from endemic areas
and harvested from warm waters using case 9.
TABLE 27
Sampling plans and recommended microbiological limits for seafoods
Product
Testa
Fresh and frozen fish and cold-smoked
APC
E. coli
APC
E. coli
Frozen raw crustaceans
APC
E. coli
Frozen cooked crustaceansc
APC
E. coli
Staph. aureus
Cooked, chilled, and frozen crabmeat d
APC
E. coli
Staph. aureus
Fresh and frozen bivalve molluscs
APC
E. coli
Case
Plan
Limit per gram

or per cm2
m
M
115
133e
1
4
3
3
5
5
3
3
5x105
11
107
500
115
133e
2
5
3
3
5
5
2
2
5x105
11
107
500
115
133e
1
4
3
3
5
5
3
3
106
11
107
500
115
133e
220
2
5
8
3
3
2f
5
5
5
2
2
0
5x105
11
103
107
500
115
133e
220
2g
6
9
3
3
2f
5
5
5
2
1
0
105
11
103
106
500
115
133e
3
6
2
2
5
5
0
0
5x105
16
Method
referenceb
a For additional tests, conditions of election are given in footnotes to Table 26, p. 189. These tests are described in Table 27
(supplement), p. 191.
b Refers to page numbers in ICMSF, 1978a.
c Note that although Staph. aureus is a moderate hazard, it is infrequently found at high levels and a 2class plan n = 5, c = 0 is
recommended to reduce analytical work.
d Refer to note c for explanation of Staph. aureus case assignment. APC for frozen products only.
e See text.
f A 3class plan would require mvalues of about 101, which are not detectable with current methodologies.
g In this product APC is not related to hazard.
TABLE 27 (supplement)
Additional tests to be carried out when appropriate (refer to footnotes, Table 26, and see text)
Product
Test
Method
reference
Fresh and frozen fish and cold-smoked
Salmonella
V. parahaemolyticus
Staph. aureus
Limit per gram

or per cm2
m
M
Case
Plan
163
202
220
10
7
7
2
3
3
5
5
5
0
2
2
0
102
103
103
104
Staph. aureus
220
103
104
Frozen raw crustaceans
Salmonella
V. parahaemolyticus
Staph. aureus
163
202
220
10
8
7
2
3
3
5
5
5
0
1
2
0
102
103
103
104
Frozen cooked crustaceans
Salmonella
V. parahaemolyticus
163
202
11
8
2
3
10
5
0
1
0
102
103
Cooked, chilled, and frozen crabmeat
V. parahaemolyticus
202
10
102
Fresh and frozen bivalve molluscs
Salmonella
V. parahaemolyticus
163
202
12
9
2
3
20
10
0
1
0
102
103
103
192
from the freshly caught raw material. They also provide inferential
evidence of temperature abuse in handling of the product.
Attention is drawn to the warning footnote in Table 26 concerning
standards for molluscs. The hazard associated with these seafoods when
unprocessed and of unknown origin is unpredictable and yet potentially
so high that no simple set of criteria can ensure a safe product for the
consumer. It is essential that the condition of the growing waters from
which the animals are harvested be known before hazard can be assessed.
Also be aware of the possibilities of paralytic shellfish poisoning and
poisoning from other shellfish toxins in certain areas of the world.
The sampling plans and bacteriological criteria are listed in Table 27.
For fresh and frozen fish, m for APC has been reduced to 500,000 per
gram from 1,000,000 per gram because this value is well within the reach
of GCP products in most situations. In a few cases where fish are taken in
tropical waters or inshore waters heavily contaminated with bacteria this
level may not be attainable. For such situations it is recommended that, if
the other bacterial safety indicators are within limits, a high m may be
accepted, since this criterion relates primarily to utility. The m for cooked
fish has also been set at 500,000 per gram and it is recognized that this is
slightly high since most regulatory agencies set their limit for cooked
products at 100,000 per gram. Nevertheless, this figure recognizes the
actual situation in international trade, and it has not been shown that
products with higher counts are associated with food-borne illness. The m
for APC for raw crustaceans is set at 1,000,000 per gram, recognizing the
high count on much of the shrimp in international trade. However, small
pink shrimp from cold northern and southern waters generally have much
lower counts, and an m of 500,000 per gram might be applicable for
them. Cooked crustaceans have been allocated an m for arc of 500,000
per gram. Again this is somewhat high for a cooked product but more
representative of the world trade situation than 100,000 per gram or less.
This and other considerations have led to cases 2, 5, and 8 being
proposed (see Table 26, footnote d). The values for molluscs represent a
compromise among the standards listed for various countries in the
Recommended International Code of Hygienic Practice for Molluscan
Shellfish (FAO/WHO, 1978). They are probably excessively strict for
products to be cooked before eating but, since molluscs are often eaten
raw or lightly cooked, the more stringent view has been accepted.
Fish and shellfish
193
D. SAMPLING PROCEDURES
For general directions on collecting sample units and handling analytical units, see Chapter 9. In routine analysis, collect 5 sample units
(see Table 27; 10 are needed in certain cases) from each lot (or consignment) and in each case take at least twice as much food as will be
required for the laboratory analysis. Use 0.1% peptone water with salt
(ICMSF, 1978a, p. 320) as the diluent for all tests; for fish and fish
products, experience has shown that this is the best diluent.
(a) Iced or chilled raw fish
1 Whole fish in international trade is normally destined for further
processing and is not usually tested bacteriologically. Where whole fish
are to enter directly into the retail market, bacteriological tests are
normally directed at obtaining estimates of bacterial populations on skin
surface for large fish and total bacterial count per unit weight for small
fish. Sample units usually consist of single fish in the case of large fish
and one or more fish in the case of small fish. Surface samples may
consist of swabs taken over a prescribed surface area (e.g., 200 cm2), or
of an aseptically excised area of skin including a minimum of
underlying tissues. Shake the swab or skin tissue thoroughly in 10 ml of
peptone diluent and proceed with analyses as described in ICMSF, 1978a,
p. 115 or 118. Express results in counts per unit area (e.g., per cm2).
2 For chilled or iced fillets of small fish, take a whole fillet for each
sample unit; for large fish such as halibut take a representative portion
(e.g., three sections from different parts of the fillet). In the laboratory
weigh out an analytical unit of 100 g from each sample unit and
proceed with blending, dilution, and analysis as in 1. The initial dilution
in blending can range from 1:3 to 1:10.
3 For scallops in ice, collect one sample unit from each of five
containers. Weigh out 25 g analytical units from each sample unit and
proceed with blending, dilution, and analysis as in 1.
(b) Frozen fish
Frozen fin fish products are normally shipped as blocks of whole gutted
fish, fillet blocks, comminuted fish blocks, or consumer packages of
fillets or other prepared material.
1 Large fish such as salmon, tuna, swordfish, or halibut may be
194
shipped as frozen whole fish (usually gutted) with or without heads.

Sampling and bacteriological testing of such products is difficult and
generally is not done. In certain cases imported products of this type are
detained until examination of a sample from subsequent processing
(e.g., steaking operation) has been shown to be satisfactory. Examination of the processed sample follows the methods given in 2 below. 2
Sample units of wholesale lots of frozen fish may consist of individual
fish or entire blocks, cartons, or institutional packages; they may be
pieces cut, sawn, or drilled from these larger units (preferably during a
primary processing operation in the receiving plant). In the laboratory,
thaw sample units for 3 hours at room temperature (20-25C), overnight
in a chilled room (2-5C), or in sealed plastic pouches under running
water, and weigh out 100 g analytical units. Whenever possible the
analytical unit should be blended with diluent before becoming
completely defrosted. The temperature of the product should not rise
above 5C prior to analysis. Conduct blending, dilution, and analysis as
in 1. The initial dilution can range from 1:3 to 1:10.
3 For sampling retail packages of frozen fish, use 100 g analytical
units, and treat as in 2 above.
4 Sample and analyse frozen raw or cooked shrimp, prawns, and
lobster tails, and frozen raw breaded shrimp (items 3, 4, 5 in Table 27)
as described in 2 and 3 above.
(c) Cooked and cold smoked fish
Sample and analyse cold smoked fish as described for raw (fresh or
frozen) fish above. Sample cooked fish by cutting portions and blending
at least 100 g with diluent. Then proceed as for raw fish.
(d) Molluscs
For shellstock. Clean off the surface using a stiff brush and running
water of potable quality. Open the shell using a sterile instrument to cut
the adductor muscle and transfer the shell contents (liquor and tissue) to
a sterile container using aseptic technique. Dilute with an equal weight
of peptone diluent (ICMSF, 1978a, p. 320) and proceed according to
ICMSF, 1978a, p. 115. For the E. coli count follow the procedures in
ICMSF, 1978a, p. 126 et seq.
For shucked meats. Dilute with an equal weight of peptone diluent and
proceed as described above.
Fish and shellfish
195
REFERENCES
Bryan, F.L. 1980. Epidemiology of foodborne diseases transmitted by fish, shellfish and
marine crustaceans in the United States 1970-1978. Journal of Food Protection, 43, 859868, 873-876
CDC (Centers for Disease Control). 1981. Foodborne Disease Outbreaks: Annual Summary
1978 revised, pp. 49-50
Clarke, N.A., and Chang, S.L. 1959. Enteric viruses in water. Journal of the American
Waterworks Association, 51, 1299-1317
Cliver, D.O. 1979. Viral infections. In Foodborne Infections and Intoxications, 2nd ed., ed. H.
Riemann and F.L. Bryan, 299-342. Academic Press, New York
Codex (Codex Alimentarius Commission). 1977. Draft Code of Practice for Shrimps
or Prawns. (Document CX/FFP7/7). FAO, Rome
Eklund, M. 1982. Significance of Clostridium botulinum in fishery products preserved
short of sterilization. Food Technology, 36 (12), 107-112
FAO (Food and Agriculture Organization). 1979. Yearbook of Fisheries Statistics, 1978, 47,
11
FAO/WHO (Food and Agriculture Organization / World Health Organization). 1976.
Recommended International Code of Practice for Fresh Fish. (Document CAC/RCP 91976). FAD, Rome
1978. Recommended International Code of Hygienic Practice for Molluscan Shellfish.
(Document CAC/RCP 18-1978) FAO/WHO, Rome
Hobbs, G. 1976. Clostridium botulinum and its importance in fishery products. Advances
in Food Research, 22, 135-185
Holbrook, R., Anderson, J.M., and Baird-Parker, A.C. 1980. Modified direct plate method for
counting Escherichia coli in foods. Food Technology in Australia, 32, 78-83
ICMSF. 1978a. Microorganisms in Foods. 1. Their Significance and Methods of
1978b. Microorganisms in Foods. 2. Sampling for Microbiological Analysis: Principles
and Specific Applications. University of Toronto Press, Toronto
Liston, J. 1957. The occurrence and distribution of bacterial types on flatfish. Journal of
General Microbiology, 16, 205-216
Murphy, A.M., Grohmann, G.S., Christopher, P.J., Lopez, W.A., Davey, G.R., and Millsom,
R.H. 1979. An Australia-wide outbreak of gastroenteritis from oysters caused by Norwalk
virus. Medical Journal of Australia, 2, 329-333
Sakaguchi, G. 1979. Botulism. In Foodborne Infections and Intoxications, 2nd ed., ed. H.
Riemann and F.L. Bryan, 389-442. Academic Press, New York
Sakazaki, J.R. 1979. Vibrio infections. In Foodborne Infections and Intoxications, 2nd ed., ed.
H. Riemann and F.L. Bryan, 173-209. Academic Press, New York
Sharpe, A.N., Peterkin, P.L, and Rayman, M.K. 1981. Detection of Escherichia coli in foods:
Indole staining methods for cellulosic and polysulfone membrane filters. Applied and
Environmental Microbiology, 41, 1314-1315
Shewan, J.M. 1977. The bacteriology of fresh and spoiling fish and the biochemical changes
induced by bacterial action. In Proceedings of the Conference on the Handling, Processing
and Marketing of Tropical Fish, 51-66. Tropical Products Institute, London
196
Todd, E.C.D. 1978. Foodborne disease in six countries - A comparison. Journal of Food
WHO (World Health Organization). 1974. Fish and Shellfish Hygiene. WHO Technical
Report Series No. 550. WHO, Geneva
18
Sampling plans for
vegetables, fruits, and nuts
Vegetables and fruits are important in international commerce mainly as

fresh, frozen, dried, or canned products. Ground and tree nuts are also
shipped internationally. The range of products is enormous, and with
the continued introduction of new varieties it is impossible to discuss all
hazards presented by them and all requirements of Good Manufacturing
Practice (GMP). Therefore only the broad classes of these commodities
are considered, together with yeasts, and single-cell proteins. For further
information on these commodities and a detailed discussion of their
microbiology see ICMSF, 1980, pp. 606-642.
A. RAW VEGETABLES AND FRUITS
Many vegetables and fruits are not cooked before eating. There is an
increasing market for fresh vegetables which are chopped, packaged,
then chilled or frozen, and intended to be consumed uncooked in the
form of salads. Such preparations are subject to more contamination
than are unchopped raw vegetables.
Human and animal enteric pathogens (except soil-borne sporeformers such as Clostridium perfringens and Bacillus cereus) are usually absent from fresh vegetables at harvest, unless they have been
fertilized with human or animal wastes or irrigated with water containing such wastes. A few outbreaks of human disease as a result of
eating raw vegetables contaminated by waste water have been reported
(Bryan, 1977). Recently an outbreak of listeriosis was caused by contaminated cole-slaw (cabbage) (HWC, 1981). Unclean containers may
cause additional contamination as demonstrated by the detection of
salmonellae on produce packed in wooden crates formerly used to ship
198
iced poultry (NCDC, 1968). Further details on the pathogens of concern

and the products involved are given by ICMSF, 1980.
Raw vegetables intended to be cooked prior to eating are less
hazardous, because the pathogens contaminating them, other than bacterial spores, are mostly heat-sensitive. Spores are of little concern as a
health hazard in the raw product but may become a problem in the
cooked product if it is subjected to time-temperature abuse. Nevertheless, raw vegetables and fruits may introduce pathogens into processing plants and kitchen environments, and contaminate other foods.
Good agricultural practices in growing crops, combined with acceptable hygienic methods during harvesting, packing, and transporting
of vegetables and fruits, are far more important than microbiological
testing. Therefore, adequate knowledge of fertilization, irrigation, and
harvesting and washing practices in the production area is essential.
Vegetables likely to be eaten raw should not be obtained from areas
where these practices are known, or suspected, to be unhygienic. While
the use of potable water to wash and freshen the harvested vegetables is
useful both in commercial practice and in the home, the effects of a
water rinse on the level of contamination should not be overestimated.
Even thorough commercial washing can remove only c. 90% of the
surface flora, which means that 105 organisms per gram or cm2 will
often remain on the surface of produce after washing. Residual water
may support extensive microbial growth during further storage.
Several appropriate codes of hygienic practice and good manufacturing procedure are available (e.g., Codex, 1983; FDA, 1969).
Routine microbiological examination of raw vegetables and fruits is
unlikely to reduce hazards to any great extent and is not recommended.
However, tightened inspection may be appropriate on lots suspected of
being implicated in food-borne illness, or of having been produced
under hygienically unsatisfactory conditions (see Chapter 5 Section O:
Tightened Inspection).
B. FROZEN VEGETABLES AND FRUITS
Prepacked frozen vegetables and fruits continue to be popular convenience foods and are in international trade. These products have a
commendable microbiological safety record. Blanching, a commonly
used procedure, intended primarily to inactivate degradative enzymes,
usually kills any vegetative form of pathogens contaminating surfaces
of vegetables and fruits. Therefore, the bacteriological condition of such
products is primarily a reflection of the hygienic practices employed
Vegetables, fruits, and nuts
199
after blanching. In many instances, however, recontamination of the

blanched products occurs.
Although sampling plans for aerobic plate counts (APC) and Escherichia coli were recommended in the first edition of this book, the
Commission no longer supports the need for the APC recommendation.
Without information on the history of a product and its processing
conditions, it is impossible to interpret the results of an APC performed
at a port of entry. If this information is available (e.g., to industry)
appropriate limits can be established and case 1 or 2 would be applicable depending on whether or not the product is to be cooked.
Escherichia coli is usually an indicator of faecal contamination so a
criterion (case 5) is proposed for both blanched and unblanched
products having a pH value greater than 4.5 (Table 28). Although
blanching may diminish the level of contamination, recontamination can
occur during subsequent handling. Products having a pH value of 4.5 or
less are unlikely to cause problems because the low pH accelerates dieoff during frozen storage and inhibits growth in thawed products. A
sampling plan for Salmonella is not recommended because there is
neither epidemiological nor laboratory evidence that Salmonella occurs
in these frozen products.
C. CANNED VEGETABLES AND FRUITS
Both acid and low-acid canned vegetables and fruits are shelf-stable
products whose safety is ensured by a heat process, providing that all
canning procedures are in accordance with GMP. No sampling plans are
proposed for these products (see Chapter 26: Shelf-Stable Canned
Foods).
D. DRIED VEGETABLES AND FRUITS
Information on the microbiology of dried vegetables and fruits is scarce

(ICMSF, 1980, pp. 606-642). The flora depends largely on the conditions
of growing and harvesting of the particular commodity (see Section A
of this chapter) and the processing prior to drying (washing), or other
cleaning and blanching. Several products, mainly fruits, such as apricots
and prunes, are treated with sulphur dioxide. Although such gassing,
spraying, or dipping is primarily intended to stabilize colour, it also
considerably reduces the microbial load. The method of drying affects
the microflora of products but cannot be relied on to eliminate all
pathogenic organisms. Drying of vegetables and fruits is mostly
TABLE 28
Sampling plans and recommended microbiological limits for vegetables, fruits, nuts, and yeast
Product
Test
Method
referencea
Frozen vegetables and fruits

(pH > 4.5)b
E. coli
Dried vegetables
Limit per
gram
m
Case
Plan
126/131
102c
103
E. coli
126/131
102c
103
Coconut(desicated)
Growth not anticipated
Growth anticipated
Salmonella
160d
11
12
2
2
10
20
0
0
0
0
Yeast
Salmonella
12
20
160d
a Refers to page number in ICMSF, 1978, where methods are described. Use analytical unit sizes recommended in the methods.
b pH measured at the time of sampling. Commodities with pH 4.5 or less are not likely to represent a major hazard and criteria are
not recommended.
c In the absence of systematic data m values are estimates.
d See also ISO 6579 (1981).
Reprinted from: Microorganisms in Foods 2. Sampling for microbiological analysis: Principles and specific applications. 1986. 2nd Ed. International Commission on
Microbiological Specifications for Foods.
Vegetables, fruits, and nuts
201
combined with heating, but the internal temperature of the product

during drying rarely exceeds 35 to 40C because of evaporation of
water. Microbial growth can occur during drying in those internal
tissues which still contain sufficient moisture. In the final dried product
the water activity is usually below 0.70, and only some moulds and
osmotolerant yeasts can multiply.
For the reasons explained in Section B, the Commission recommends that a criterion (case 5) for E. coli be applied to dried vegetables
and fruits, primarily to monitor GMP (Table 28).
To the best of our knowledge dried vegetables and dried fruits have
not caused salmonellosis. Therefore a criterion for Salmonella is not
recommended. If routine monitoring detects E. coli, there may be
reason to monitor for Salmonella.
E. YEASTS AND SINGLE-CELL PROTEINS
Dried yeast and yeast extract have caused outbreaks of salmonellosis.

These products are used as ingredients for dietary foods which are
often eaten without prior heating. Therefore limits are recommended
for Salmonella; case 12 applies (Table 28).
Single-cell proteins that is, proteins derived from microorganisms
are potentially important replacements of traditional proteins in manufactured foods. However, because their usage is limited, criteria are
not recommended. If international trade becomes significant the same
criteria as for dried yeasts would apply.
F. NUTS
(a) Ground nuts

Peanuts are the only important commodity in this category. The only
hazard is the presence of mycotoxins. This is controlled by sorting out
damaged nuts by visual inspection or electronic devices. Lots are
checked for aflatoxins by chemical procedures or ultraviolet light. No
microbiological criteria are suggested. Peanut butter is dealt with in
Chapter 22 (Fats and Oils).
(b) Tree nuts
Though many varieties of tree nuts move in international trade, only
coconut gives cause for concern. The hazard is contamination with
202
Salmonella. Case 11 is appropriate for products where multiplication of

Salmonella does not occur; in products where growth may occur case 12
is recommended (Table 28). For other tree nuts, testing for E. coli has
proved an effective monitor of GMP.
G. SAMPLING PROCEDURES
Appropriate methods for sampling frozen or dried foods are given in

Chapter 9 (Collecting and Handling Sample and Analytical Units). For
sampling most raw fruits and vegetables a rinse technique should be used
as the organisms of concern will be present mainly on the surface. Make
a 1/10 dilution (w/v) of the commodity in 0.1 % peptone water containing 1 % tergitol; shake for 1 minute. Care should be taken not to
break the skin of fruits.
If acid fruits are blended, the acids released may destroy the
organisms being sought. Counts should be expressed on a per gram basis
recognizing that data obtained using the rinse technique for differentsized fruits (e.g., watermelon and red currant) are not comparable
because of the large difference in the surface/volume ratios.
REFERENCES
Bryan, F.L. 1977. Diseases transmitted by foods contaminated by wastewater. Journal of
Codex (Codex Alimentarius Commission). 1983. Recommended International
Code of Practice: General Principles of Food Hygiene (CAC/vol. A - Ed. 1), First
Revision (1979). FAO/WHO, Rome
FDA (United States Food and Drug Administration). 1969. Human foods; current good
manufacturing practice. Federal Register, 34, 6977 (CRF, pt. 128)
HWC (Health and Welfare Canada). 1981. Listeriosis - Atlantic Provinces. Canada Diseases
Weekly Report 7 (52), 257-259
ICMSF. 1978. Microorganisms in Foods. I. Their Significance and Methods of Enumeration,
2nd ed. University of Toronto Press, Toronto
on Methods for the Detection of Salmonella. International Standard
6579. ISO, Geneva
NCDC (United States National Communicable Disease Center). 1968. Salmonella
Surveillance. Report No. 72, National Communicable Disease Center, Atlanta, Georgia
19
Sampling plans for soft drinks,
fruit juices, concentrates, and
fruit preserves
The microbiological safety and stability of this diverse group of products depend on their formulation (including the use of chemical preservatives, carbonation, low pH values, and pasteurization). For details
of preservation requirements see ICMSF, 1980, pp. 643-668. Control
should be by application of the HACCP system to raw materials, formulations, processing conditions, and packaging (see Chapter 7: Control at Source).
Examples of such controls would be: (a) measurement of temperature during heat-treatment of pasteurized products; (b) measurement of
carbonation levels by routine monitoring of C02 pressures in carbonated
products; (c) measurement of pH, aw, product viscosity, and preservative
levels in products preserved by the use of chemical preservatives in
combination with other formulation parameters; (d) maintenance of
good hygienic practices to prevent the build-up of spoilage organisms
on the plant equipment and in the production areas.
Most microbiological problems arise because of poor quality of raw
materials (e.g., fruit concentrates, sugars, and syrups) and poor process
hygiene, which lead to overcoming of the preservation system applied
during manufacture by the spoilage organisms. Thus there is a need for
routine sampling of raw materials, preventive maintenance of
equipment, and monitoring of those parts of the process where
microbiological contamination can occur (e.g., during filling and in
holding tanks). For the routine monitoring of raw materials simple
microbiological tests such as the use of the direct microscopic count, a
yeast and mould count, or a standard plate count (using a medium
capable of supporting lactic acid bacteria) are most generally applicable.
For certain materials more specialized tests, such as the determination of
diacetyl and acetylmethylcarbinol in citrus and apple juices
204
and their concentrates, may give a better index of microbiological quality

(Murdock, 1968).
It is impossible to recommend specifications for specific types of
microorganisms in raw materials as these will very much depend on how
they are processed and how they are to be used in the final product. For
instance, for a pasteurized juice it is important to control levels of heatresistant moulds in raw materials whereas in a product preserved by the
use of chemical preservatives it is more important to know the level of
preservative-resistant microbes. Thus, Gluconobacter and related
bacteria such as Acetobacter are important spoilage organisms in
products that contain benzoate and/or sorbate and are distributed in
plastic containers. If the differentiation of these organisms is important,
the method described by Cirigliano (1982) may be used. A case 1
sampling plan would be appropriate for the examination of materials.
Although routine microbiological testing of end-products is not
useful for the direct control of manufacturing operations it may provide
useful trend data on the performance of a particular operation over a
period of time. For this purpose, once a production line has been
established to perform to within commercially acceptable spoilage levels,
the numbers of containers that should be examined from a production
plant on a daily basis should be between 0.01 and 0.1 per cent of the
production units taken from production lines.
Unless prior knowledge indicates that sampling of containers should
be concentrated on particular parts of a production period (e.g., during
start-up, change in product type, or adjustment of packaging equipment),
containers should be sampled at random.
Samples should be incubated appropriately and examined for evidence of microbial spoilage. In most products simple visual inspection
for haze, sediment, gas, or mould clumps will suffice. If microbiological
examinations are made, the methods used should be the most suitable for
the likely spoilage organisms. Most spoilage problems result from the
growth of yeasts (principally Saccharomyces, but often Brettanomyces in
carbonated beverages) and/or acid-tolerant bacteria such as Acetobacter,
Gluconobacter, Lactobacillus, and Leuconostoc. The complete identification of spoilage organisms is often useful in identifying the cause of a
particular microbiological problem.
REFERENCES
Cirigliano, M.C. 1982. A selective medium for the isolation and differentiation of
Gluconobacter and Acetobacter. Journal of Food Science, 47, 1038-1039
Soft drinks, fruit juices, concentrates, and preserves

ICMSF. 1980. Microbial Ecology of Foods. Vol 2. Food Commodities. Academic Press,
New York
Murdock, D.I. 1968. Diacetyl test as a quality control tool in processing frozen
concentrated orange juice. Food Technology, 22, 90-94
205
20
Sampling plans for cereals
and cereal products
Cereal grains and their products, such as flours, grits, and meal, are
important commodities in international trade and form the staple foods of
many communities. There are many cereal-based products and a wide
variety of confectionery and bakery products with a cereal as a main
component. Some of these products are potential microbial foodborne
disease risks by virtue of their cereal components but many more may be
risks because of the various animal and other vegetable components they
contain. Those products for which the use of microbiological criteria
would appear to be of value are listed in Table 29. The microbiology of
cereals and their products is discussed in detail in ICMSF, 1980, pp. 669730.
A. GRAINS AND FLOURS
Cereals include wheat, maize, barley, oats, rye, durum, rice, millet, and
sorghum. Protein-rich oil seeds, such as soya beans, and the flours and
protein isolates and concentrates derived from them are included in this
chapter as their storage, handling, and use pose risks that are somewhat
similar to those of cereal grains.
Grains, provided they are harvested when in good condition and
rapidly dried to a water activity level preventing microbial growth, and
then stored under conditions such that the excessive ingress or movement
of water is avoided, have virtually no microbiological risks. However, in
practice, these conditions are not always met and mould growth may
result. The principal organisms of concern are moulds which produce
mycotoxins. Many of the field and storage fungi found on grains and in
flours derived from them are capable of producing
Cereals and cereal products
207
mycotoxins in incorrectly dried or stored cereal products (see ICMSF,

1980, p. 685).
Their control should be principally at source, that is, by controlling
conditions during and after harvesting. It is not recommended that these
products be routinely monitored for mycotoxins because routine
analytical methods for most mycotoxins are not available. However,
where the presence of a mycotoxin of concern (e.g., aflatoxin) has been
identified in a particular production area, extensive monitoring of the
product may be indicated. For methods of analysis of products for
mycotoxins, see Horwitz, 1980.
Levels of mould propagules found in food-quality grains and flours
vary between approximately 102 and 105 per gram depending on the
type of grain and the growing, harvesting, drying, and storage conditions (including time of storage). Grains and flours may be variously
treated, and this circumstance will affect the levels of moulds found on
examination. Thus grains may be washed in chlorinated waters and
flours may be heat-treated to change functional properties or treated
with a bleaching agent such as chlorine dioxide. All these treatments
will reduce mould levels, although none will have a significant effect on
mycotoxins, if present. Patent flours, made from part of the endosperm
of the grain, will generally contain only low levels of moulds. Wholemeal flours will often contain much higher levels as these include the
surface layers of the grain, which are those parts principally
contaminated with moulds. Thus when deciding on an acceptable limit
for moulds it is necessary to know the conditions to which the flour or
grain has been subjected prior to sampling. Without such information,
results of analyses cannot be properly interpreted. Providing the
processors have this knowledge, mould counts of grains in flours may
provide useful information concerning the suitability of these products
in various formulations. For this purpose a 3class sampling plan (case
5) with an M value of 105 mould propagules per gram would be
applicable. In place of a single criterion, a range of m values is
proposed because of the wide range of grains, flours, and related
products involved for which neither data are available nor a single
Good Commercial Practice (m value) would be applicable.
Salmonellae are recognized contaminants of oil seeds and flours,
and isolates and concentrates derived from them (e.g., soya beans,
cotton seed, and sunflower seeds). Although these commodities are
invariably cooked, or heat-processed, before eating, they may be a
source of contamination, and levels of salmonellae should thus be
208
controlled. Because the major problem is with soya products a 2class

sampling plan (case 10) for Salmonella is recommended for them in
addition to the 3class plan for moulds (Table 29).
Bacillus cereus is common. It is a soil organism and an unavoidable
contaminant of grains and flours. While levels of B. cereus can be
controlled by correct storage and may be reduced somewhat by cleaning
grains, it cannot be eliminated. Two cereal products, fried rice and
custard powder, have been frequently associated with food-poisoning,
although on occasion other cereal-based products have been incriminated (Gilbert, 1979). Bacillus cereus food-poisoning associated with
fried rice is the consequence of holding cooked rice for several hours at
warm (room) temperatures, or in large containers in refrigerators, prior
to frying. The problem with custard also results from temperature abuse
of the prepared custard prior to consumption. Such incidents are best
prevented by education and training of food-handlers and informative
labelling of products (e.g., refrigerate after preparation for consumption). They will not be controlled by monitoring products for B. cereus
contamination. Thus criteria for B. cereus are not recommended except
for frozen entres containing rice or cornflour as a main ingredient,
where there is a risk of abuse during manufacture. For this purpose a 3
class sampling plan (case 8) is applicable but because of very limited
data on a narrow range of products only a tentative m can be proposed.
The B. cereus selective agar of Holbrook and Anderson (1980) is
recommended.
Flours and particularly starches are frequently contaminated with
thermophilic spore-forming bacteria which can cause spoilage of
canned foods stored above c. 37C. Thus it is recommended that levels
of aerobic thermophilic spores in starches and other cereal products to
be used in canned products should not exceed 150 spores per 10 g with
an average of no more than 125 spores per 10 g based on the method of
analysis specified by the National Food Processors Association
(Maunder, 1976). This criterion is not applicable to these commodities
when intended for other purposes.
Flours used in the manufacture of bread and pastry products may be
a source of specific spoilage microorganisms. Of particular concern in
breads are the rope-causing bacteria (mucoid variants of B. subtilis and
B. lichenformis). Various microbiological methods have been proposed
for monitoring flours for such organisms but none has been found to be
useful in practice. Because the sensitivity of bread to rope formation
depends on its formulation and expected storage condition it is useful to
carry out a practical baking test to determine whether a
209
TABLE 29
Sampling plans and recommended microbiological limits for cereals and cereal
products
Product
Test
Method
referencea
Case
Plan
class
Limit per
gram
m
2
4f
105
Cereals
Moulds
158
10 -10
Soya flours,
concentrates,
and isolatesb
Moulds
Salmonellac
158
163d
5
10
3
2
5
5
2
0
102-104f
0
105
Frozen bakery Staph. aureus 220

products
Salmonellac 163d
(ready to eat)
with low-acid
or high-aw
fillings or
toppings
9
12
3
2
5
20
1
0
102g
0
104
Frozen bakery Staph. aureus 220

products (to
Salmonella,c 163d
be cooked)
with low-acid
or high-aw
fillings or
toppings (e.g.,
meat pies,
pizzas)
8
10
3
2
5
5
1
0
102g
0
104
274e
103g
104
Frozen and
Staph. aureus 220
dried products Salmonella,c 163d
8
10
3
2
5
5
1
0
102g
0
104
Frozen entres B. cereus

containing rice
or corn flour
as a main
ingredient
a Refers to page number in ICMSF, 1978.

b Soya flours, concentrates, and isolates should in addition be examined for
Salmonella (case 10; plan case 2; n = 5; c = 0; m = 0).
c 25 g sample units should be used; these may be pooled for analysis.
d See also ISO 6579 (1981).
e Preferably the method of Holbrook and Anderson (1980) should be used.
f Best estimate based on limited data. Value will depend on type of grain. In case
of flours, value will depend on conditions used for their manufacture.
g Best estimate based on limited data; should be checked against products
produced under GCP.
210
particular flour is suitable for bread manufacture; that is, prepare a small
batch of bread in the usual way, store the bread under expected
conditions of storage, and check by visual inspection whether or not
rope develops (see ICMSF, 1980, p. 717).
Refrigerated doughs and other raw pastry products may sour as a
result of growth of lactic acid bacteria present in cereal components.
Such organisms occur in flours used for making doughs and grow to
high numbers on the dough-making equipment, which is difficult to
clean. Microbiological methods have been proposed for checking levels
of lactic acid bacteria in cereal products (Hesseltine et al., 1969).
However, the potential for souring depends on the formulation and
storage conditions, and the numbers of lactic acid bacteria that can be
tolerated in particular products can be determined only by practical tests.
Problems are avoided by strict attention to process hygiene.
B. FROZEN AND DRIED PASTA PRODUCTS
Pastas are raw dough products made from wheat flour, semolina, farina,
water, and other ingredients (e.g., egg) to form a stiff dough usually
containing about 30% moisture. The dough is extruded into various
shapes and for certain types (e.g., ravioli) may be formed around a meatcontaining ingredient. Because of ingredients used, conditions of storage
prior to drying, temperatures used for drying, and the slow rate of drying
which is necessary to prevent damage to certain types of pasta, these
may contain salmonellae and Staphylococcus aureus. As these products
are to be cooked prior to consumption, case 10 is recommended for
testing for salmonellae, and case 8 for Staph. aureus (Table 29). A limit
of M = 104 per gram is proposed for Staph. aureus. However, if Staph.
arueus is found in pasta or pasta products in excess of 103 per gram
additional tests such as the thermonuclease test (or ideally tests for
enterotoxins) should be used as Staph. aureus may be destroyed during
processing but any enterotoxins may not be inactivated. For a suitable
method see ICMSF, 1978, p. 242).
Pasta may also be used in made-up dishes such as frozen lasagne,
and for such products the same limits should apply.
C. FROZEN FILLED AND TOPPED BAKERY PRODUCTS
Most frozen bakery products are not a food-borne disease risk. Those of
potential concern are filled or topped with sensitive ingredients such
211
as products of animal origin (for example, meat pies, dairy-cream-filled

doughnuts, pies and cakes, pizzas, and flans) or contain coconut. Abuse
may occur after these products leave processing plants. For example, the
incidents of botulism in California during 1975 and 1976 involving
frozen pot pies resulted from temperature abuse to which the frozen
products where subjected after thawing, such as the leaving of unheated
pies in ovens overnight. Such gross mishandling of a product cannot be
taken into account by any criterion. Limits for Staph. aureus and
salmonellae are considered appropriate for certain types of frozen, filled,
and topped bakery products (see Table 29).
The potential for Staph. aureus and salmonellae to grow in nonfrozen bakery products depends on the formulation of the fillings and
toppings. The presence of these organisms depends mainly on whether
or not the sensitive ingredients have been pasteurized and hygienically
handled prior to use. Of particular concern are those bakery products that
contain unpasteurized ingredients, are ready to eat, and are eaten without
heating. No criterion is proposed for products containing fillings and
toppings that have water activity levels below 0.85, pH values below 4.5,
or with a combination of aw, pH, and preservatives that prevents growth
of these organisms. Such products include jams, jellies, certain artifical
creams, acid cheeses, spices, and certain fermented meat products.
The hazard also depends on conditions of use. For products likely to
be abused after preparation, stricter limits should be applied (i.e., case 9
for Staph. aureus and case 12 for salmonellae).
D. SAMPLING PROCEDURES AND METHODS
Appropriate sampling procedures are described in Chapter 9 (Collecting

and Handling Sample and Analytical Units), and the relevant microbiological methods are given in Table 29.
REFERENCES
Gilbert, R.J. 1979. Bacillus cereus gastroenteritis. In Food-borne Infections and
Intoxications, 2nd ed., ed. H. Riemann and F.L. Bryan, 495-518. Academic Press, New
York
Hesseltine, C.W., Graves, R.R., Rogers, R., and Burmeister, H.R. 1969. Aerobic and
facultative microflora of fresh and spoiled refrigerated dough products. Applied
Holbrook, R., and Anderson, J.M. 1980. An improved selective and diagnostic medium for
the isolation and enumeration of Bacillus cereus in foods. Canadian Journal of
212
Horwitz, W. 1980. Official Methods of Analysis of the Association of Official

Analytical Chemists, 13th ed., AOAC, Washington, DC
on Methods for the Detection of Salmonella. International Standard 6579, ISO, Geneva
Maunder, D.T. 1976. Thermophilic flat sour sporeformers. In Compendium of Methods
for the Microbiological Examination of Foods, ed. M. L. Speck, 254-262. American
Public Health Association, Washington, DC
21
Sampling plans for spices,
condiments, and gums
Spices are dried plants or parts of plants which are used primarily for
flavouring, seasoning, and imparting aroma or colour to foods and
beverages. For the purposes of this discussion herbs, the aromatic leaves
and flowers of certain plants, are included as spices. Condiments are
spices and blends of spices which may be formulated with flavour
accentuators and potentiators to enhance the flavour of foods. Spices
and condiments is the official nomenclature adopted by the International
Organization for Standardization (ISO, 1972).
Gums include a wide range of plant and microbial polysaccharides,
or their derivatives, which hydrate in either cold or hot water to form
viscous solutions or dispersions. In addition to tree exudates, actual gums
include seaweed extracts (e.g., agar, alginates, and carrageenan), gums
from seed (e.g., locust bean, guar), and microbial gum (e.g., xanthan). As
hydrocolloids, gums usually exhibit suspending, dispersing, and
stabilizing properties, or they may function as emulsifiers, impart gelling
properties, or be mucilaginous. Some gums serve as binders, flocculating
agents, film formers, foam stabilizers, mould release agents, or lubricants
(Peppier and Guarino, 1984).
A. INTERNATIONAL COMMERCE
Spice-bearing plants are chiefly indigenous to the Asian tropics but are
grown commercially in many tropical and some temperate climates (e.g.,
in Hungary). Herbs are grown in more temperate climates of the
Mediterranean, Middle East, North Africa, and North America and
widely throughout Europe. India, one of the major spice producers and
exporters, accounts for more than 50% of the international trade now
estimated at 400 million dollars annually. The United States spice
214
consumption, alone, is around 500 million pounds per year. The spice
map of the world continues to change as some sources decline, others
are revitalized, and new areas of spice production are developed
(Peppier and Guarino, 1984).
The major sources of gums include India, Sudan, North Africa, Iran,
Asia Minor, Pakistan, and Mediterranean countries (Peppier and
Guarino, 1984).
B. COLLECTION OF SPICES AND GUMS
Spices are produced from botanically diverse plants grown in a wide

variety of soils and climates. Depending on the spice, various parts of
the plant are used, such as the seed (mustard and nutmeg), the berry
(pepper), the fruit (celery seed and nutmeg), the leaves (oregano and
bay), the root (horseradish), or the bark (cinnamon) (ICMSF, 1980, pp.
731-751).
As with grains, the primary consideration after harvesting is to dry
the plant portion to a water activity level sufficiently low to prevent
spoilage. Subsequently, further processing may involve, depending on
the spice, such treatments as grinding, pulverizing, cleaning, and sorting. Some of these operations may be carried out at or near the point of
harvest; others may be done at a remote site after shipment in
international commerce. The microflora of spices reflects the diversity
of raw materials from which they are derived, together with the varied
and climatic conditions under which the plants are grown. These factors
are further influenced by conditions of drying and storage, both of which
may be extremely primitive (ICMSF, 1980, pp. 731-751).
Processing methods for many gums are proprietary, but patents and
other literature provide some insight. Most tree gums are collected by
hand and sorted, graded, packed, and shipped from the growing area.
The processor further grades, cleans, mills, and blends the gum. As with
spices, the water activity of the crude material is reduced to a level to
preclude microbial growth. For example, gum ghatti is sundried and
pulverized. Carrageenan is extracted from red algae with hot water,
filtered and concentrated to about 3% gum content, precipitated with
ethanol, dried, and milled (Peppier and Guarino, 1984).
C. MICROBIOLOGY OF SPICES, GUMS, AND HERBS
These products commonly carry large numbers of bacteria and moulds,

in the case of herbs mainly of soil origin. Conditions of handling after
Spices, condiments, and gums
215
harvest often permit extensive contamination and microbial growth

although drying with heat somewhat reduces microbial numbers. The
predominating flora is generally composed of aerobic spore-forming
bacteria. Non-spore-forming bacteria, indicator organisms, and pathogens
may be found.
The numbers of microbial contaminants can be reduced by treatment
with ethylene oxide and to a lesser extent with propylene oxide.
Irradiation is also effective in reducing microbial populations (ICMSF,
1980, p. 746).
Bacterial spoilage subsequent to harvesting and drying constitutes no
problem, but spoilage by fungi may occur during storage and shipping.
Of major concern is the role of contaminating microorganisms as spoilage
agents in foods where spices and gums are used as ingredients. Spices
contaminated with excessive numbers of spore-forming bacteria may be
responsible for the spoilage of canned foods and processed meats (ICMSF,
1980, pp. 747-748). Though spore-forming organisms capable of causing
gastroenteritis are found in small numbers in spices, they present no
public health concern unless the foods into which the spices are
incorporated as ingredients are mishandled to the extent that the small
number of Bacillus cereus and Clostridium perfringens, for example, are
permitted to multiply extensively. Salmonella is found infrequently, but
in a wide variety of spices, and on rare occasions has been responsible for
outbreaks of human salmonellosis (Laidley et al., 1974; Lystad et al.,
1982; Gustavsen and Breen, 1984).
D. SAMPLING PLANS
In the first edition of this book (p.115) sampling plans and microbiological limits were suggested for spices and gums. In retrospect, it is
felt that these recommendations were inappropriate. A considerable
portion of these commodities in international commerce would not meet
the suggested limits; furthermore, failure to meet the limits might or
might not have borne a relationship to food quality or safety.
It is now our recommendation that spices and gums be treated as raw
agriculture commodities. Accordingly, the ultimate use of such products
will be the determinant. A spice or gum which is unsatisfactory for one
purpose may be quite adequate for another. For example, black pepper
containing large numbers of aerobic spore-forming bacteria may cause
spoilage when used as an ingredient in canned foods or cooked sausage,
but it could be used as a condiment on the table
216
for seasoning cooked foods. A manufacturer using spices and gums

must determine whether the microbiological quality of these products
is satisfactory for the use intended (NCA, 1968). If it is not, his alternatives are to seek another source or to render the available product
satisfactory. Treatment with ethylene oxide has been practised extensively for many years but there is some concern over its continued use.
Gamma-irradiation may be an effective alternative provided flavours
are not impaired.
In commercial practice the microbial and colour problems, and to
some extent variability, are overcome by using extracts of the spices.
For convenience in use these are often diluted on a carrier such as salt
or a finely ground cereal (usually rusk, i.e., ground non-sweetened
biscuit). Another approach is to use microencapsulated extracts using
gelatin or starch as the encapsulant. Use of these forms is increasing at
the expense of raw spices.
REFERENCES
Gustavsen, S. and Breen, O. 1984. Investigation of an outbreak of Salmonella oranienburg
infection in Norway, caused by contaminated black pepper. American Journal of
Epidemiology, 119, 806-812
ICMSF. 1980. Microbial Ecology of Foods. Vol. 2. Food Commodities. Academic Press,
New York
ISO (International Organization for Standardization). 1972. Spices and Condiments.
Nomenclature finalized draft proposal. TC-34/SC-7, ISO, Budapest
Laidley, R., Handzel, S., Severs, D., and Butler, R. 1974. Salmonella weltevreden outbreak
associated with contaminated pepper. Epidemiology Bulletin (Department of National
Health and Welfare, Ottawa), 18 (4), 62
Lystad, A., Aasen, S., and Lassen, J. 1982. Food-borne disease surveillance. Outbreak of
Salmonella oranienburg infection (Norway). Weekly Epidemiological Record, 57, 329330
NCA (National Canners Association). 1968. Laboratory Manual for Food Canners and
Processors, vol. 1, 3rd ed. AVI Publishing Co. Westport, Conn.
Peppler, H.J., and Guarino, P.A. 1984. Spices and condiments. In Compendium of Methods
for the Microbiological Examination of Foods, 2nd ed., ed. M.L. Speck. American Public
Health Association, Washington, DC
22
Sampling plans for
fats and oils
Four categories of foods were discussed under fats and oils in ICMSF,
1980, pp. 752-777, namely mayonnaise and salad dressings, peanut
butter, margarine, and butter. Product criteria would serve no useful
purpose for any of these products except nut butters, including peanut
butter.
The safety of mayonnaise and salad dressings is related directly to
the pH and acetic acid content of the moisture (continuous) phase of the
emulsions. The safety and stability of butter and margarine is primarily
dependent on heat-treatment of the ingredients and the stability on
emulsification to ensure uniformly small water droplets within the fat
(continuous) phase. While microbiological criteria may be effectively
applied at various stages in the production of these foods, product
criteria, for application at ports of entry, would be of extremely limited
value. Other spreads of different composition, however, may be less
stable microbiologically.
With peanut and other nut butters microbiological stability is related
to reduced water activity in the moisture phase of the product. The
greatest hazard is the possibility of contamination with salmonellae, and
if these organisms are present they will survive for long periods in a
finished product although they would not grow in it. Peanut butter is
generally eaten without either heating or a treatment that would destroy
microorganisms. Therefore, for Salmonella, case 11 would apply (see
Table 30), unless peanut butter is to be used as an ingredient in a highmoisture food, in which instance case 12 would apply. Thus 2class
plans are suggested with n = 10, c = 0 for case 11 and n = 20, c = 0 for
case 12.
218
TABLE 30
Sampling plans and recommended microbiological limits for peanut butter and other nut
butters
Product use
Test
Method
referencea
Case
Plan
class
Limit per
gram
m
M
Consumed without Salmonella

heating or other
treatment to
destroy microbes
160b
11
10
Used as ingredient Salmonella

in high-moisture
food
160b
12
20
a Refers to page numbers in ICMSF, 1978.

b See also ISO 6579 (1981).
REFERENCES
1980. Microbial Ecology of Foods. Vol. 2. Food Commodities. Academic Press,
New York
ISO (International Organization for Standardization). 1981. Microbiology - General
6579. ISO, Geneva
23
Sampling plans for
sugar, cocoa, chocolate,
and confectionery
1. SUGAR AND RELATED SWEETENERS

Sugar (sucrose) is obtained from sugar cane and sugar beets. It is sold for
commercial use in both crystalline and liquid form. The liquid form may
be sucrose or invert sugar obtained by hydrolysing sucrose. Corn
sweeteners (glucose syrups) are obtained by acid and/or enzymatic
hydrolysis of corn starch to dextrose, maltose, and higher polymers.
They are sold both in solid form and as syrups. An increasingly
important sweetener is high-fructose corn syrup (HFS, also known as
isoglucose or isosyrup) made by enzymic conversion of glucose syrup to
fructose.
The world production of sucrose and other sweeteners mentioned
above in 1980 was 113 million tonnes (mt) of which 5 mt was HFS
(Earley, 1980). It is predicted that by 1990 about 150 mt of sweeteners
will be produced of which 8.2% will be from hydrolysed corn starch, and
most of the corn-based sweeteners will be HFS. Sucrose obtained from
various fruits and fructose by hydrolysis of inulin contributes only a
minor portion of the world supply of sweeteners.
The concentrations of the liquid sweeteners most commonly used in
commerce are listed in Table 31 (Marignetti and Mantovani, 1979/80;
Minifie, 1980). Sugar syrups with concentrations of 67% w/w or higher,
if properly prepared and stored, are usually not subject to spoilage by
surface moulds or osmophilic yeasts.
Dry refined sugar and dried glucose syrups are remarkably stable, have
not been vehicles of food-borne pathogens, and for many industrial uses
(e.g., jam, confectionery, bakery products) do not contain microbes
220
TABLE 31
Concentrations of commercial liquid sweeteners
Liquid sweetener (syrup)
Percentage (w/w)
Sucrose
6768
Invert sugar
7677
Glucose (2665 DE)*
7582
High fructose (42%, 55%, and less frequently up to 90% of
7677
sugar is fructose)
* DE = dextrose equivalent = the percentage of reducing sugars, on a dry basis,
calculated as dextrose
that will spoil products in which they are used. Some liquid sweeteners
may become mouldy or ferment unless stored to preclude condensation
on the surface and formation of microenvironments permitting growth of
surface moulds or osmophilic yeasts. Similarly, contamination with
osmophilic yeasts must be avoided to prevent their contaminating
sensitive products (Coleman and Bender, 1957; ICMSF, 1980, pp. 778821). Dry and liquid sugars used in certain canned foods, and for soft
drinks, may introduce sufficient numbers of spoilage microorganisms for
some to survive the process and spoil the product. For this reason, sugars
used in the canning and soft drinks industries are usually subjected to
microbiological testing. Both the microbial limits and the analytical
methods have been defined (NCA, 1968; 1972; NSDA, 1975; AOAC, 1972).
B. SAMPLING PLANS
It has been suggested (ICMSF, 1980, pp. 802-803) that case 2 be used for
dry sugar that is for purposes other than canning or bottling, and case 3
for both dry and liquid sugar to be used for canning or bottling. The basis
for this suggestion was the potential for such sweeteners to introduce
spoilage organisms into these classes of products. Additionally, liquid
sugars (and other liquid sweeteners) may be spoiled by osmophilic
yeasts. On further consideration, however, sampling plans and
microbiological criteria are not recommended for the following reasons:
(a) sugars and other sweeteners have not been involved in microbial
food-borne disease; (b) the microbial content of sugars and other
sweeteners is mainly the concern of manufacturers using them as
ingredients in other products (see NCA, 1972 for limits suitable for
purchase specifications).
Sugar, cocoa, chocolate, and confectionery
221
II. COCOA AND CHOCOLATE

Cocoa beans are the source of material for a wide variety of cocoa and
chocolate products. The annual crop is relatively stable at about 1.5
million tonnes, but the price is sometimes subject to enormous changes
within a few years (Minifie, 1980). The combined annual per capita
consumption of cocoa and chocolate among 19 countries that consume
the bulk of the product varies from 0.6 kg in Portugal to 8.9 kg in
Switzerland; the median is 4.3 kg (Minifie, 1980).
Cocoa beans are harvested, fermented, and dried mostly in technically

underdeveloped or in developing countries, frequently under primitive
jungle conditions. Most of the dried beans have in the past been processed to cocoa, cocoa butter, and chocolate in more technically advanced countries, where most of the product is consumed. However,
there is a trend towards processing an increasing percentage of the crop
to intermediate products such as chocolate liquor, cocoa press cake, and
cocoa butter in the country of origin to increase local economic return.
This practice may cause some concern because one step in processing,
roasting, is the main germicidal barrier between a raw agricultural
product and the consumer. Unless roasting is adequate and handling of
roasted beans well controlled, the intermediate products may be
contaminated with Salmonella.
Salmonellae are not readily destroyed in chocolate liquor or in
confectionery made with cocoa or chocolate liquor because of their
increased resistance to heat at the low aw of such products. In cocoa and
in chocolate confectionery, salmonellae may persist for years. Chocolate
even appears to protect these bacteria against destruction by acid in the
stomach. Small numbers (<1 per gram) of salmonellae in chocolate have
caused at least one large outbreak of salmonellosis (Craven et al., 1975;
DAoust et al., 1975) and larger numbers (about 25 per gram) have
caused another outbreak (Gill et al., 1983). Salmonella-contaminated
cocoa has been associated with a similar large outbreak (Gastrin et al.,
1972). Even when no illness has been reported from Salmonellacontaminated products, the risk has on occasions prompted their recall
(Silliker, 1967; Lennington, 1967; Anonymous, 1978).
Moulds are frequently found in cocoa beans, and it is not uncommon
to find mycotoxin-producing moulds and occasionally low levels of
mycotoxins in cocoa. However, raw cocoa beans do not readily
222
TABLE 32
Sampling plans and recommended microbiological limits for cocoa, chocolate, and
confectionery
Product
Test
Method
referencea
Cocoa
Salmonella
Chocolate and other Salmonella

confectionery
Case
Plan
class
Limit
m
163
11
10b
163
11
10b
a Refers to page number in ICMSF, 1978. Methods as described (pp. 163-172) using non-
selective enrichment medium made with skim milk and brilliant green (ICMSF, 1978, p.
324).
b The 25 g analytical units may be composited.
support the production of mycotoxins although they do support abundant

growth of mycotoxin-producing fungi (Llewellyn et al., 1978).
Therefore, there is always a threat of mycotoxins in cocoa products, but
toxin has been found only infrequently and only at low levels.
B. SAMPLING PLANS
The ICMSF (1980, p. 818) has suggested that case 2 be used for aerobic
plate counts (APC) and mould counts for cocoa, chocolate, and compound
coatings containing cocoa, and that case 5 be applied for coliforms and
Escherichia coli, and case 11 with n = 10 and c = 0 for Salmonella.
However, on further consideration, only Salmonella, as the main foodborne hazard in cocoa and cocoa products, need be tested for at ports of
entry. The total microbial population and the presence of indicator
oganisms are generally controlled by suppliers as part of purchase
specifications. Based on the inability of these products to support
microbial growth, and their non-involvement in food-borne disease other
than salmonellosis, there appears little point in testing them for microbes
other than Salmonella (see Table 32).
III. CONFECTIONERY
Sugar confectionery is eaten in most countries. Minifie (1980) reported
that the annual per capita consumption in 17 countries varied from 1.2 kg
in Portugal to 6.6 kg in Ireland, with a median of 3.6 kg.
223
Confectionery, because of its low aw is generally free from microbial

hazards and microbial spoilage problems. The main hazards, when they
exist, may come from such ingredients as cocoa, coconut, milk powder,
and egg albumen, all of which sometimes contain Salmonella, and nuts
and cocoa, which may contain mycotoxins. Cocoa has been discussed
above and the other ingredients in other chapters. The main microbial
spoilage problems are (1) explosion of enrobed preparations of fruit
and fondant as a result of growth and gas production by osmophilic
yeasts introduced with the fruit, (2) explosion of marzipan from gas
produced by osmophilic yeasts introduced with nuts or from the environment, and (3) Wasserflecken (globular areas with watery appearance) in marzipan resulting from growth of moulds (Windisch and
Neumann, 1965; Windisch et al., 1978). Moulds may also cause spoilage as a result of condensation of moisture in water-impermeable
packages (Jarvis, 1982).
B. SAMPLING PLANS
The ICMSF (1980, p. 821) suggested that case 3 was appropriate for
spoilage organisms and case 11 for Salmonella in confectionery. Spoilage organisms are mainly the concern of the manufacturer. At port-ofentry sampling for Salmonella only, case 11 is recommended (see
Table 32).
IV. SAMPLING AND TESTING PROCEDURES FOR
COCOA, CHOCOLATE LIQUOR, COATINGS
CONTAINING COCOA, CHOCOLATE, AND
OTHER CONFECTIONERY
For general directions on collecting and handling samples, see Chapter
9 and ISO, 1981.
Among the many factors to be considered when examining food for
Salmonella, the anti-Salmonella activity in some cocoa which is
negated by casein is especially pertinent when isolating this organism
from cocoa and combinations of cocoa with milk (e.g., milk chocolate)
(Park et al., 1976; Zapatka et al., 1977). In the interest of uniformity,
we recommend that reconstituted skim milk with brilliant green be used
as the non-selective enrichment medium for cocoa-containing products
whether or not they already contain milk. This method appears
224
to be most useful for the products of concern, although it differs in

some details from several other recommendations (IOCC, 1977; ICMSF,
1978; FDA, 1978; Wilson et al., 1980).
REFERENCES
Anonymous. 1978. FDA seeks injunction against candy firm. Food Chemical News, 3 April,
21
AOAC (Association of Official Analytical Chemists). 1972. Detecting and estimating
numbers of thermophilic bacterial spores in sugars - Official first action. Journal of the
Association of Official Analytical Chemists, 55, 445-446
Coleman, M.C., and Bender, C.R. 1957. Microbiological examination of liquid sugar using
molecular membrane filters. Food Technology, 11 (7), 398-403
Craven, P.C., Baine, W.B., Mackel, D.C., Barker, W.H., Gangarosa, E.J., Goldfield, M.,
Rosenfeld, H., Altman, R., Lachapelle, G., Davies, J.W., and Swanson, R.C. 1975.
International outbreak of Salmonella eastbourne infection traced to contaminated
chocolate. Lancet, i, 788-793
DAoust, J.Y., Aris, B.J., Thisdele, P., Durante, A., Brisson, N., Dragon, D., Lachapelle, G.,
Johnston, M., and Laidley, R. 1975. Salmonella eastbourne outbreak associated with
chocolate. Canadian Journal of Food Science and Technology, 8, 181-184
Earley, T.C. 1980. High fructose corn syrup in the USA, Europe and Japan. In Proceedings
of the International Sweetener and Alcohol Conference, 62-69. International Business
Conferences and Seminars, Financial Times Conference Organization, London
FDA (United States Food and Drug Administration). 1978. Bacteriological Analytical
Manual. 5th ed. Food and Drug Administration, Bureau of Foods, Division of
Microbiology. Association of Official Analytical Chemists, Washington, DC
Gastrin, B., Kampe, A., Nystrom, K., Oden-Johanson, B., Wessel, G., and Zetterberg, B.
1972. An epidemic of Salmonella durham caused by contaminated cocoa.
Lakartidningen, 69 (46), 5335-5338 (original in Swedish)
Gill, O.N., Sockett, P.N., Bartlett, C.L.R., Vaile, M.S.B., Rowe, B., Gilbert,R.J., Dulake, C.,
Murrell, H.C., and Salmaso, S. 1983. Outbreak of Salmonella napoli infection caused by
contaminated chocolate bars. Lancet, i, No. 8324, 574-577
ICMSF. 1978. Microorganisms in Foods. 1. Their Significance and Methods of Enumeration, 2nd ed. University of Toronto Press, Toronto
1980. Microbial Ecology of Foods. Vol. 2. Food Commodities. Academic Press, New
Yorkiocc (International Office of Cocoa and Chocolate). 1977. Microbiological
Examination of Chocolates and Other Cocoa Products. Page 101/B/E-1977. The IOCC
and the International Sugar Confectionery Manufacturers Association, Brussels
ISO (International Organization for Standardization). 1981. Agriculture Food
Products - Layout for a Standard Method of Sampling. Draft International
Standard ISO/DIS 7002. ISO, Geneva
Jarvis, B. 1982. The occurrence of mycotoxins in UK foods. Food Technology in
Australia, 34, 508-514
225
Lennington, K.R. 1967. FDA looks at the Salmonella problem - Ever-present threat to the
industry. The Manufacturing Confectioner, 48 (July), 31-33
Llewellyn, G.C., Benevides, J., and Eadie, T. 1978. Differential production of aflatoxin on
natural and heat-treated cocoa beans. Journal of Food Protection, 41, 785-787
Marignetti, N., and Mantovani, G. 1979/80. Liquid sugar. Sugar Technology Reviews,
7, 3-47
Minifie, B. 1980. Chocolate, Cocoa and Confectionery: Science and Technology, 2nd ed.
AVI Publishing Co., Westport, Conn.
NCA (National Canners Association). 1968. Laboratory Manual for Food Canners and
Processors. Vol. 1. Microbiology and Processing. AVI Publishing Co., Westport, Conn.
1972. Bacterial Standards for Sugar. NCA, Washington, DC
NSDA (National Soft Drink Association). 1975. Quality Specifications and Test Procedures
for Bottlers Granulated and Liquid Sugar. NSDA, Washington, DC
Park, C., Rayman, K., Stankiewicz, Z., and Hauschild, A.H.W. 1976. Toxic effect of cocoa on
Salmonella during the pre-enrichment period. Proceedings of the Canadian Federal
Biological Society, 162
Silliker, J.S. 1967. The Salmonella problem at the candy manufacturers level. The
Manufacturing Confectioner, 47, (Oct.), 31-33
Wilson, C.R., Andrews, W.H., and Poelma, P.L. 1980. Recovery of Salmonella from milk
chocolate using a chemically defined medium and five non-defined broths. Journal of
Food Science, 45, 310-313, 316
Windisch, S., Kowalski, S., and Zander, I. 1978. Demonstration of osmotolerant yeasts in
almonds. CCB Review for Chocolate, Confectionery and Bakery, 3 (2), 28-29
Windisch, S., and Neumann, I. 1965. Ober die Wasserfiecken des Marzipans and ihre
Entstehung. Zeitschrift fur Lebensmittel-Untersuchung and -Forschung 129, 9-16
Zapatka, F.A., Varney, G.W., and Sinskey, A.J. 1977. Neutralization of the bactericidal effect
of cocoa powder on salmonellae by casein. Journal of Applied Bacteriology, 42, 21-25
24
Sampling plans for
formulated foods
Commercially prepared ready-to-cook or ready-to-eat foods have been

widely accepted throughout the world. When such a product consists of
major ingredients from more than one commodity category, it is termed
a formulated food. Examples of the multitude of such products are meat
and vegetable pot pies, seafood and meat salads, ice cream, infant
formulae, dried soups, dessert pies, and certain ethnic dishes including
certain pasta products, enchilada, and eggrolls. New products are being
developed and marketed, and others discontinued. Consequently it is not
possible to provide a meaningful list of all formulated foods.
A wide range of processes are used to prepare these foods, which
may be offered for sale to the public as perishable, semi-preserved, or
shelf-stable products. Relatively minor changes in formulation,
especially the post-processing addition of condiments such as grated
cheese, sesame seed, ground spices, or chocolate frosting, may alter the
microfiora of the product to a degree that different microbiological
criteria would be needed for apparently similar products.
For the reasons noted above, the ICMSF makes no pretence of
covering the entire range of formulated foods. The purpose of this
chapter is to provide guidance and examples for those food industries
and regulatory agencies that perceive the need to apply microbiological
criteria to particular formulated foods. For information about the basic
concepts and principles involved, reference should be made to Part I of
this book.
Several commodity chapters in Part II of this book contain examples
of formulated foods that are traditionally associated with a particular
commodity which is a major ingredient of the product (e.g., ice cream
with milk products, pasta with cereal products). Some of the formulated
Formulated foods
227
foods mentioned in the first edition of this book have been omitted from
this revision because there is insufficient information to determine
whether microbiological criteria are useful. Accordingly the types of
products included in this chapter are dried soup and gravy mixes, meat
pies, salads, selected precooked frozen foods and certain dietetic foods,
and foods specifically for infants and children. Other products such as
salads and pies are described in ICMSF, 1980, pp. 822-833. For most of
those foods the quality of the raw materials is of paramount importance
for the quality and safety of the end-product. In such cases, setting
criteria for the end-products for the purpose of reducing the eventual
hazard to the consumer will be less effective than testing the raw
ingredients.
For many formulated foods, the number of microorganisms as
determined by the aerobic plate count (APC) is not indicative of adherence
to Good Manufacturing Practice (GMP). Moreover, the food will be
cooked just prior to eating, thus killing most microorganisms, or will be
eaten shortly after preparation, thus avoiding any extended opportunity
for microbial growth. For example, certain dried soups and ready-tocook foods, which will be cooked before eating, contain raw ingredients
that may have unavoidably large numbers of microorganisms including
coliforms and occasionally Salmonella. As the process of manufacture
does not always reduce the number of microorganisms in the raw
material, their presence in retail samples does not indicate a lack of GMP.
A few commercially processed formulated foods, such as meat pies
that are eaten cold, have been incriminated in outbreaks of food-borne
disease due to lack of GMP in the factory. Most outbreaks have occurred
because of post-processing time-temperature abuse, improper storage, or
mishandling by the preparer before serving. Commercial food preparation introduces no hazard that is not present in the domestic situation.
However, the magnitude of the hazard is much greater because of the
greater number of persons exposed to the lot of the commercial product.
In general the need to establish microbiological criteria for most
formulated foods has not been demonstrated. Indeed, in the case of dried
soups, the Codex Alimentarius Committee on Food Hygiene decided
specifically that such criteria are not justified. However, there are some
products that require special attention either because they may contain
pathogenic organisms that are not killed before eating, or because they
are typically eaten by highly susceptible persons such as those debilitated
by age or illness. The latter include foods for infants
228
and children who are in the high-risk category, and the costs involved in
controlling the microbial populations mentioned below seem justified.
Microbiological end-product specifications for these products were
established by the Codex Alimentarius Commission (Codex, 1981).
A. FOODS FOR INFANTS AND CHILDREN
Commercially processed foods for infants and children were not considered in detail in ICMSF, 1980, and some comments concerning their
manufacture are necessary before considering sampling plans and limits.
Canned baby foods, however, will not be discussed here because they
must meet the same requirements for safety as other shelf-stable,
sterilized products (see Chapter 26, Shelf-Stable Canned Foods, for
discussion of these foods). Three other types of foods are dealt with in
the Codex Alimentarius Commissions microbiological specifications
for foods for infants and children (Codex, 1979). These are dried
biscuit-type products, dried instant products, and dried products
requiring heating before consumption.
Plain biscuit-type foods are usually thoroughly baked, and no hazard should exist for these because of the heat process and their low aw.
Those with a coating or a filling require more sophisticated manufacturing practices, but the prior heat-treatment of fillings and coatings
should be such that the numbers of vegetative forms of microorganisms
are sufficiently reduced. The aw of the product prevents microbiological
multiplication during storage and distribution; moreover, since these
products are not reconstituted in water or milk before consumption,
risks of mishandling of product are minimal.
Dried instant products range from simple milk powders to very
complex mixtures. The milk, with or without additional ingredients, is
heat-treated (but not sterilized), concentrated, and then spray-dried, or
roller-dried when cereals have been added. The heat-treatment is as mild
as possible to maintain high nutritive quality. Dried instant foods
include certain dry-blended products which are not heated after
blending.
Good manufacturing practices should include careful selection of
raw materials, hygienic design of processing equipment, adequate control of the critical points during manufacture, and, of course, continuous
good housekeeping. Adherence to these good manufacturing practices
will normally lead to products containing very few vegetative
microorganisms. The APC may, however, be 103 to 104 per gram,
Formulated foods
229
consisting usually of spores that have survived heat-treatments applied

during preparation of the product or its ingredients.
Bacillus cereus, a normal contaminant of milk, can be expected to
be present in low numbers. Its mere presence does not constitute a
hazard, but if the product is sufficiently abused to permit growth to high
numbers, a significant hazard could result.
In principle, there should be very little risk that the product contains
heat-sensitive pathogenic organisms. Nevertheless, contamination by
Salmonella occasionally occurs, sometimes from the use of contaminated ingredients added during dry-mixing or from the environment of
the production line. The severity of the hazard presented by these
organisms depends on subsequent treatment of the product. If adequately heated the hazard from Salmonella is removed. If not heated,
this hazard remains and will increase if subsequent abuse of the product
occurs.
Mishandling that is, leaving the reconstituted product at room
temperature for periods of more than 1 to 2 hours may present a risk
to the young, old, or debilitated consumer. It should also be taken into
account that uncontaminated infant formula may become contaminated
during preparation. If prepared in advance, such products should be
cooled rapidly in order to prevent bacterial growth.
A third category of foods for infants and children is dried foods
which require heat-treatment before consumption. The production of
such foods by dry-mixing of non-heat-treated raw materials or with less
severe heat processes means that most of these products contain higher
numbers of microorganisms than do dry instant foods, and even some
heat-sensitive bacteria may be present. However, the presence of these
microorganisms in these products is accepted with the full knowledge
that cooking will reduce their numbers to non-hazardous levels, making
the food safe. Instructions for heating such products should appear on
the label.
The advisory end-product specifications of Codex Alimentarius are
too strict to be met by all the above categories of products in international commerce (Collins-Thompson et al., 1980). These criteria were
established to give guidance on what is considered good microbiological quality. However, only at the factory, when the microbiology
of raw material and ingredients and the influence of the process are
known, will it be possible to judge if divergence from the endproduct
specifications truly represents some fault in GMP. Products are present
on the market, made according to GMP, which contain higher
230
numbers of aerobic bacteria than stated in the Codex criteria (for dried
and instant products, APC: n = 5, c = 2, m = 103, M = 104; coliforms: n =
5, c = 1, m = 3, M = 20). Such products may have been made with raw
materials containing higher spore counts. Moreover, a few coliforms
can sometimes be present without indicating a risk to the health of the
consumer. Thus, in order to avoid unnecessary rejection of safe foods in
retail markets, or at a point of entry where the history of the product is
unknown, the criteria for many of these products will have to be more
lenient than the end-product specifications of Codex Alimentarius.
B.
SAMPLING PLANS FOR FOODS FOR INFANTS AND

CHILDREN AND FOR CERTAIN CATEGORIES OF DIETETIC
FOODS
The following considerations are applicable to foods for infants and

children and certain categories of dietetic foods (e.g., those intended for
consumption by debilitated, or otherwise sensitive, consumers).
The microorganisms to be considered are the mesophilic aerobes,
coliforms, staphylococci, and salmonellae. The APC may be used to
determine the number of bacterial spores in the raw materials, growth of
bacteria at or between critical points, the effect of the applied heattreatment, growth in accidentally rehydrated products during storage
and distribution, and, in general, adherence to Good Commerical Practices (GCP). Since all coliforms should be killed during the heat-treatments applied, this group of microorganisms can be used as indicators
for post-process contamination and growth; in short, they can be used
for the same purposes as the APC to check adherence to GCP. In drymix
processes, when no heat-treatment is applied, the coliform count is
useful to indicate adherence to GCP in the production, storage, and
distribution of the raw materials.
Staphylococci may occasionally be present, but usually only in low
numbers, as a result of contamination after drying or during dryblending
(see Chapter 15, Milk and Milk Products). In the production of foods for
infants and children rigid quality-assurance programs should be applied
to meet the end-product specifications. A massive contamination with
staphylococci without an alert during production or an increase in APC or
coliforms is highly improbable, while their presence in low numbers
does not indicate a direct public health hazard. Such a hazard would
arise only after mishandling of the product after
Formulated foods
231
reconstitution. In this case the situation is not different from that

described above for B. cereus.
In the light of these considerations there is no need for routine
examination for Staphylococcus aureus in these foods. A similar conclusion was reached in Codex, 1979. However, when excessive contamination with Staph. aureus is suspected, case 9 (n = 10, c = 1, m =
10, M = 102) would apply. When test results or other information
indicate that a build-up of Staph. aureus has occurred, the product
should be considered as unsuitable for the use intended. Salmonella is
considered a hazard because the minimal infective dose might be low for
the intended consumer group, and for products requiring reconstitution
the hazard will be increased if subsequent mishandling occurs. For these
products sampling plans for Salmonella are recommended (Table 33).
For plain dried biscuit no microbiological criteria are suggested
because there is no hazard which can be limited by setting such criteria.
For coated and filled biscuits, tests for coliforms and Salmonella would
suffice. Since the condition of use of this product would not increase the
hazard, case 5 is selected for the coliform criterion. Case 11 is chosen
for Salmonella because of (a) the possible presence of an ingredient
contaminated with Salmonella and (b) the conditions of use, which lead
to no increased hazard.
For dried and instant products, a criterion for APC is recommended in
addition to those for coliforms and Salmonella. Directions for use of
these products do not include any microbial destructive treatment (e.g.,
heating) which can be relied on. As these products must be rehydrated,
the possibility of subsequent bacterial growth due to mishandling cannot
be excluded. Case 6 is selected for the APC and the coliform counts. For
Salmonella case 12 remains but, because of the high-risk consumer
group, n is increased to 60 (see Table 12, p. 78).
For dried products that require heating before consumption case 4 is
chosen for APC and coliforms and case 10 for Salmonella, because
cooking will reduce the hazard.
The limits are based on available published and unpublished data,
and take into consideration that maintaining the nutritive quality of such
products is more important than reducing the numbers of mesophilic
bacteria of little public health significance. In the last two categories
values of m and M have been modified from those in Codex, 1981, as a
consequence of bacteriological surveys and commercial experience.
232
TABLE 33
Sampling plans and recommended microbiological limits for foods for infants and
children and certain categories of dietetic foods
Product
Test
Coated or filled,
dried shelf-stable
biscuits
Coliforms
Salmonellaef
Dried and instant APCg

products requiring
reconstitution
Coliforms
Salmonellae
Dried products
requiring heating
to boiling before
consumption
APC
Coliforms
Salmonellae
Method
referencea
126-131
(Method 3)b
163-172c
115-118
(Method 1)d
126-131
(Method 3)b
163-172c
115-118
(Method 1)d
126-131
(Method 3)b
163-172c
Class
Case plan
Limit per
gram
m
M
10
102
11
30
104
105
10
102
12
60
105
106
10
102
10
15
a Refers to page number in ICMSF, 1978.

b See also ISO 4831 (1978a).
c See also ISO 6579 (1981).
d See also ISO 4833 (1978b).
e Use 25 g analytical units which may be pooled.
f Applies only to products containing a Salmonella-sensitive ingredient.
g Not applicable to fermented products.
Reference should be made to those sections of Chapter 9 (Collecting and

Handling Sample and Analytical Units) dealing with the sampling of
dried and powdered products.
In the examination of dried products for Salmonella, rapid rehydration by shaking the powder with the pre-enrichment fluid should be
avoided (van Schothorst et al., 1979; Andrews et al., 1983).
REFERENCES
Andrews, W.H., Wilson, C.R., and Poelma, P.L. 1983. Improved Salmonella species recovery
from non-fat dry milk pre-enriched under reduced rehydration. Journal of Food Science,
48, 1162-1165
Formulated foods
233
Codex. 1979. Codex Alimentarius Commission. Thirteenth Session, 1979: Report of the 15th
Session of the Codex Committee on Food Hygiene, Washington, D.C. 18-22 September
1978. ALINORM 79/13. Appendix V, Code of Hygienic Practice for Foods for Infants and
Children (up to three years). Codex Alimentarius Commission, FAO, Rome
1981. Codex Alimentarius Commission, Fourteenth Session, 1981: Report of the 17th
Session of the Codex Committee on Food Hygiene, Washington, D.C. 17-21 November
1980. ALINORM 81/13. Appendix VII. Microbiological Specifications for Foods for Infants
and Children. Codex Alimentarius Commission, FAO, Rome
Collins-Thompson, D.L., Weiss, K.F., Riedel, G.W., and Charbonneau, S. 1980.
Microbiological guidelines and sampling plans for dried infant cereals and powdered infant
formula from a Canadian national microbiological survey. Journal of Food Protection, 43,
613-616
ICMSF. 1978. Microorganisms in Foods. 1. Their Significance and Methods of Enumeration,
ISO (International Organization for Standardization). 1978a. Microbiology - General
Guidance for the Enumeration of Coliforms - Most Probable Number Technique at
30C. ISO 4831-1978. ISO, Geneva
1978b. Microbiology - General Guidance for Enumeration of Microorganisms - Colony
Count Technique at 30C. ISO 4833-1978. ISO, Geneva
1981. Microbiology - General Guidance on Methods for the Detection of Salmonella.
International Standard 6579. ISO, Geneva
van Schothorst, M., van Leusden, F.M., de Gier, E., Rijnierse, V.F.M., and Veen, A.J.D.
1979. Influence of reconstitution on isolation of Salmonella from dried milk. Journal of
25
Sampling plans for
natural mineral waters,
other bottled waters,
process waters, and ice
A. DEFINITIONS AND ECONOMIC IMPORTANCE OF WATERS IN

INTERNATIONAL TRADE
Natural mineral water comes from underground water-bearing strata

and is characterized by its mineral salt and trace element constituents. It
is subjected to only minimal treatment such as carbonation, aeration,
sedimentation, and filtration. A more complete definition is cited by
Codex, 1978; a portion of this definition is given below (extracted from
the Draft European Regional Standard for Natural Mineral Waters).
2.1 Definition of Natural Mineral Water. Natural mineral water is a water
clearly distinguishable from ordinary drinking because: (a) it is characterized by
its content of certain mineral salts and their relative proportions and the
presence of trace elements or of other constituents; (b) it is obtained directly
from natural or drilled sources from underground water bearing strata; (c) of
the constancy of its composition and the stability of its discharge and its
temperature, due account being taken of the cycles of natural fluctuations; (d) it
is collected under conditions which guarantee the original bacteriological
purity; (e) it is bottled close to the point of emergence of the source with
particular hygienic precautions; (f) it is not subjected to any treatment other
than those permitted by this standard; (g) it is in conformity with all the
provisions laid down in this standard.
2.2 Supplementary Definitions
2.2.1 Naturally carbonated natural mineral water. A naturally carbonated
natural mineral water is a natural mineral water which, after possible treatment
in accordance with sub-section 3.1.1 and replacement of gas and after
packaging, has the same content of gas from the source as at emergence of the
natural mineral water taking into account the usual technical tolerance.
Bottled waters, process waters, and ice
235
2.2.2 Non-carbonated natural mineral water. A non-carbonated natural mineral water is a natural mineral water which, by nature and after possible
treatment in accordance with sub-section 3.1.1 and after packaging, does not
contain free carbon dioxide in excess of the amount necessary to keep the
hydrogen carbonate salts present in the water dissolved.
2.2.3 Decarbonated natural mineral water and natural mineral water fortified
with carbon dioxide from the source. A decarbonated natural mineral water or
a natural mineral water fortified with carbon dioxide from the source is a
natural mineral water which, after possible treatment in accordance with subsection 3.1.1 and after packaging, does not have the same carbon dioxide
content as at emergence.
2.2.4 Carbonated natural mineral water. A carbonated natural mineral water
is a natural water which, after possible treatment in accordance with subsection
3.1.1 and after packaging, has been made effervescent by the addition of carbon
dioxide from another origin.
3.1.1 Treatments permitted include separation from unstable constituents by
decantation and/or filtration, if necessary accelerated by previous aeration.
3.1.2 The treatments provided for in sections 2.2.1, 2.2.2, 2.2.3, 2.2.4 and 3.1.1
above may only be carried out on condition that the mineral content of the water
is not modified in its essential constituents, which give the water its properties.
Natural mineral water may be either non-carbonated or carbonated.

The latter may be either naturally carbonated at its origin or by adding
carbon dioxide (C02) from its own source or from other sources. Hence,
the C02 concentration may vary considerably. The pH of carbonated
mineral waters is often rather low, between 3 and 4. Natural mineral
waters without, or with very little, carbon dioxide are known in Europe
as flat waters (eaux plates).
Although specific figures are not available, bottled mineral waters
are shipped internationally in considerable quantities. Water from similar sources, but not having any particular specified amount of mineral
content, and potable* water from other sources are sometimes also
bottled. These bottled waters may be sold in large quantities in certain
*
Potable means that water must be free from microorganisms capable of causing disease
and free from either minerals or organic substances that can produce adverse
physiological effects. It must also be aesthetically acceptable and thus be free from
apparent turbidity, colour, odour, and objectionable taste and have a reasonable
temperature for its intended purpose. Hence, it can be ingested or used in food-processing
in any amount without concern for adverse effects on health (Borchardt and Walton,
1971).
236
countries or regions of a country, but are at present of limited importance in international trade.
Water is used widely in the food industry. It is used to move
products, to wash vegetables, fruits, fish, and poultry, to freshen raw
vegetables after harvesting and during distribution, and as an ingredient
in some foods. Virtually all frozen foods carry a glaze of ice which is
often derived from process water, and for certain frozen foods (such as
fish and shellfish) a glaze is added as a protective measure. This process
water is thus an integral part of some processed foods that are shipped
internationally. For example, fish is typically shipped in ice; poultry may
be cooled in water and slush ice and transported in ice.
B. REASONS FOR SAMPLING
Even the most protected and best drawn waters are never sterile and
possess a certain natural microbial (autochthonous) flora which is rather
constant for a particular supply. Subsequent contamination during
catching, storing, transporting, or bottling accounts for the introduction
of microbial species (allochthonous flora) normally present (ICMSF,
1980b, pp. 834-837).
(a) Bottled waters
Contaminated surface or ground (well or spring) waters have frequently
been identified as vehicles in outbreaks of water-borne diseases
(Holden, 1970; Mossel et al., 1973, 1977; WHO, 1979, 1980). Particular
care, therefore, is needed to obtain water from sources that are potable
and to protect the sources and water in distribution from contamination.
This is particularly important for natural mineral waters because, by
definition, they cannot be disinfected before or after bottling and can be
subjected only to certain treatments.
Water ingested between meals passes rapidly through the stomach
because the pyloric area is relaxed at that time, and thereby avoids the
bactericidal effect of intragastric retention (Levine and Nalin, 1976;
Mossel and Oei, 1975). Thus, water containing even small numbers of
pathogens can cause disease.
Bottled waters have been implicated in outbreaks of waterborne
disease, but there is little published information. Bottled, noncarbonated
mineral water was a vehicle of Vibrio cholerae in a cholera outbreak in
Portugal (Blake et al., 1977). Buttiaux (1960) attributed outbreaks of
gastroenteritis in Hungary and of typhoid fever among
237
Polish children to ingestion of bottled water. Persons from countries

with a high standard of hygiene who travel to countries where standards
are lower often drink bottled water in the hope of avoiding illness.
However, they frequently develop travellers diarrhoea, and sometimes
the bottled water comes under suspicion (Gangarosa et al., 1980).
Bottled natural mineral water is sometimes used for domestic preparation of either infant formulas or dietetic foods on the assumption that
it is safer than other waters. Consumers of these particular products are
often highly susceptible to disease. Pseudomonas aeruginosa, for
example, may be a hazard for infants (Panero et al., 1970; Mossel,
1977). An epidemic of P. aeruginosa infection occurred among infants
in a nursery when well water was contaminated by seepage of sewage
and by infiltration of stream water (Weber et al., 1977).
Thus, it is essential that the initial bacteriological quality of the water
be high and maintained by strict control at its source and by monitoring
of the bottling operation. In the absence of information on the efficacy of
control at source, importing countries should seek assurance that bottled
water is safe by appropriate microbiological monitoring of incoming
lots.
(b) Process waters
Processing water and ice can be important sources of spoilage organisms
and, if from non-potable sources or subsequently contaminated, potential
sources of pathogens. They can introduce these organisms to foods
during processing, storage, or preparation. If illness occurs as a result of
such contamination, the contaminated food is considered as the vehicle,
although the initial source of pathogens was process water. For example,
raw untreated river water was used for washing fish that were later
identified as a vehicle in an outbreak of salmonellosis (Gangarosa et al.,
1968). Contaminated cooling water was implicated as the source of
Salmonella typhi in canned corned beef responsible for several outbreaks
of typhoid fever (Howie, 1968). Water used to keep tofu (oriental soyabean curd) moist in packages was the source of Yersinia enterocolitica
that was responsible for an outbreak of gastroenteritis (Nolan et al.,
1982).
C. SAMPLING PLANS
Unfortunately, no single organism, or group of organisms, serves as an

ideal indicator of pollution of water. Tests for the following are
sometimes used to evaluate the safety of water: APC (42C, 37C, and
238
20C), coliforms, faecal coliforms, Escherichia coli, Enterobacteriaceae, faecal streptococci, sulphite-reducing clostridia, E. coli phages,
and specific pathogens. Many of these tests are used to examine bottled
water in several countries (Morisetti, 1979). The decision as to which
organisms should be sought to evaluate the potability of bottled waters
is complicated because unique advantages and disadvantages are
associated with each indicator organism or group. These have been
reviewed by Cominazzini (1978), Geldreich (1978), Hoadley and Dutka
(1976), and Mossel (1976).
(a) Non-carbonated waters
Although water-testing laboratories in some countries use the E. coli
test for routine examination of water for the protection of public health,
others, as recommended by WHO (1971) and an FAO/WHO Working
Group (1981), use the coliform test. The coliform test, even with its
disadvantages, is the method chosen by the Codex working group for
examining natural mineral water (Codex, 1978). Microbiological data
available to the Commission indicate that bottled natural mineral water
can be produced without contamination by coliforms. Examination for
coliforms provides a more stringent microbiological rejection criterion
than testing for E. coli. Routine testing of bottled waters for both
coliforms and E. coli offers little advantage and increases the cost of
analysis.
Cases 4 to 6 apply when indicator organisms are sought. Although
certain microorganisms multiply in bottle-stored water, the organisms
usually employed as indicators decrease in number with time of storage
(Schmidt-Lorenz, 1976). Because time of bottling is often unknown,
case 5 seems the most appropriate. This case calls for a sampling plan
of n = 5 and c = 2 (Table 10, p. 74) but such a plan is too lenient for a
commodity of special public health concern. These concerns are: (1) the
sampling plan employed should give at least the same protection of
public health as those required for community water supplies; (2)
bottled water is usually stored longer than water in community water
supplies, and during the storage period certain indicators of faecal
pollution (such as coliforms) may die off more rapidly than some
pathogens, while autochthonous or other allochthonus organisms multiply; and (3) home-prepared formulas for infants are sometimes reconstituted with bottled mineral water. These formulas are sometimes
subject to temperature abuse between preparation and feeding (e.g.,
when a child falls asleep during feeding and later wakes and feeds
239
again). A 2class sampling plan for coliforms with n = 5 and c = 0 and a

limit of m = 0 is recommended for examination of bottled noncarbonated waters including natural mineral water (Table 34). Analytical
units of 250 ml should be examined for coliforms by the membrane filter
method. Decreasing the c value from 2 to 0 and increasing the size of
the sample unit to 250 ml make a stringent plan to serve the abovementioned considerations.
Pseudomonas aeruginosa, a frequent contaminant of the environment, may cause water-borne disease (Hoadley, 1976, Young, 1977,
Weber et al., 1977). After adaptation, it can multiply in a mineral water
of extremely low nutrient level (Schmidt-Lorenz, 1976). Indicator
organisms are of little value in detecting the presence of this
opportunistic pathogen. The finding of P. aeruginosa in natural mineral
water should be cause for concern, particularly if the water is to be
consumed by highly susceptible persons such as infants. Therefore, it is
recommended that natural mineral water for such a purpose be tested for
P. aeruginosa. Case 8 applies, and a 2class plan with m = 0 is
suggested (Table 34).
Aerobic plate counts (at 3537C for 24 hours and at 2022C for
72 hours) at source and within 12 hours after bottling may be considered, according to the general principles of FAO/WHO, as guidelines useful
in interpreting water quality at its source and its place of bottling. To be
acceptable, counts should not exceed those set for natural mineral water
by EEC (1980a). APC at 20C may be useful for quality control within
certain food-processing plants.
(b) Carbonated waters
Pathogenic organisms (other than spores) should succumb rapidly (usually within hours) after exposure to a pH value of 3.5 or lower (ICMSF
1980a, pp. 92-111). Therefore, it is not necessary to test carbonated
natural mineral waters, or other carbonated bottled waters having a pH
value of 3.5 or lower, for indicator organisms. To determine whether
testing is necessary, a sampling plan that includes testing the pH of five
bottles (n) from a lot is recommended. If the pH of any sample unit
exceeds 3.5, the sampling plans for testing non-carbonated bottled
waters should be followed (see Table 34). If it is not possible to test the
pH of the water, the sampling plan for non-carbonated waters should be
followed.
TABLE 34
Sampling plans and recommended microbiological limits for bottled water
Method
reference
Product
Test
Non-carbonated natural mineral waters

(eaux plates) and bottled non-carbonated
waters, not classified as mineral waters
Coliforms
MFa
Pseudomonas aeruginosab See text
Carbonated waters (natural mineral or

other bottled waters
pH
See text
Limit per 250 ml

Case
Plan
5
8
2
2
5
5
0
0
0
0
3.5c
a MF = membrane filter method (see Greenberg et al., 1981). N. B. Resuscitate prior to recovery on selective media (Morisetti, 1979).
b This test is used only if it is anticipated that bottled waters are to be used in infant formulae or for other highly susceptible persons.
c If any sample unit is greater than pH 3.5, proceed with the above sampling plans for non-carbonated waters.
Reprinted from: Microorganisms in Foods 2. Sampling for microbiological analysis: Principles and specific applications. 1986. 2nd Ed.
International Commission on Microbiological Specifications for Foods.
241
(c) Process waters and ice

Guidelines for microbiological criteria of process water and ice must
be applied at the processing or manufacturing plant, at their sources,
and/or at appropriate distribution points. Process water and ice should
be made from potable water, and should meet the microbiological
criteria cited for potable water by EEC (1980b), Greenberg et al. (1981),
or WHO (1971). Ice glaze is considered as part of the frozen food itself
and will be tested during analysis of the food.
D. SAMPLING PROCEDURES
Sample units of bottled water should be collected randomly from the

lot as described in Chapter 9 and brought to the laboratory within 1
hour of collection. If this is not practicable, they should be cooled to
10C or below, iced, and transported to the laboratory in insulated
containers. Water should be collected at its source or during
processing, distribution, or storage as described by Bryan et al. (1979).
When sampling waters at their source, samples should be stored chilled
and APCs made within 12 hours. Ice should be brought frozen to the
laboratory, either as unopened bags of ice or as ice sampled in bulk
with a sterile pick, scoop, or tongs and put into sterile bags or jars.
E. TEST PROCEDURES
The coliform test by the membrane filter technique is recommended

because it facilitates sampling large volumes of water. Laboratories not
able to use this technique should test for coliforms using the multiple
tube dilution (MPN) method. Test procedures for water and ice are not
described in ICMSF, 1978, but those given in Standard Methods for the
Examination of Water and Wastewater (Greenberg et al., 1981) are
applicable. Carbonated waters should be degassed before microbiological testing. The pH should be measured by the method specified
in Greenberg et al. (1981). There is no standard method for detection of
P. aeruginosa in water, but methods are given by Greenberg et al.
(1981), Mossel et al. (1976), and WHO (1980).
REFERENCES
Blake, P.A., Rosenberg, M.L., Costa, J.B., Ferreira, P.S., Guimaraes, C.L. and Gangarosa,
E.J. 1977. Cholera in Portugal, 1974. 1. Modes of transmission. American Journal of
Epidemiology, 105, 337-343.
Borchardt, J.A., and Walton, G. 1971. Water quality. In Water Quality and Treatment:
242
A Handbook of Public Water Supplies, 3rd ed., ed. American Water Works Association.
McGraw-Hill, New York
Bryan, F.L., Anderson, H.W., Baker, K.J., Craun, G.F., Duel, W., Lewis, K.H., McKinley,
T.W., Robinson, R.A., Swanson, R.C., and Todd, E.C.D. 1979. Procedures to Investigate
Waterborne Illness. International Association of Milk, Food, and Environmental Sanitarians,
Ames, Iowa
Buttiaux, R. 1960. La surveillance bacteriologique des eaux minerales en bouteilles et en
boites. Annales de lInstitut Pasteur de Lille, 11, 23-28
Codex (Codex Alimentarius Commission). 1978. Report of the 10th Session of the
Coordinating Committee for Europe, Vienna, 13-17 June 1977. ALINOM 78/19. FAO,
Rome
Cominazzini, C. 1978. Escherichia coli and coliforms. In Faecal Bacterial Indicators in
Foods and Drinking Water (in Italian). Societ Italiana de Microbiologia Applicata, Milan
EEC. 1980a. Council Directive for Natural Mineral Waters. 80/777-15 July 1980.
European Economic Community, Brussels
1980b. Council Directive for Potable Waters. 80/778-15 July 1980. European Economic
Community, Brussels
FAO/WHO. 1981. Report of the Fourth FAO/WHO Working Group on the Establishment
and Application of Microbiological Criteria for Foods Dried Milk Products and
Natural Mineral Water, Washington, DC 10-14 November 1980. (Document VPH/81,
32) WHO, Geneva
Gangarosa, E.J., Bisno, A.L., Eichner, E.R., Treger, M.D., Goldfield, M., DeWitt, W.E.,
Fodor, T., Fish, S.M., Dougherty, W.J., Murphy, J.B., Feldman, J., and Vogel, H. 1968.
Epidemic of febrile gastroenteritis due to Salmonella java traced to smoked whitefish.
American Journal of Public Health, 58, 114-121
Gangarosa, E.J., Kendrick, M.A., Loewenstein, M.S., Merson, M.H. and Mosley, J.W. 1980.
Global travel and travelers health. Aviation, Space and Environmental Medicine, 51, 265270
Geldreich, E.E. 1978. Bacterial populations and indicator concepts in feces sewage, storm
water, and solid wastes. In Indicators of Viruses in Water and Food, ed. G. Berg. Ann
Arbor Science Publication, Ann Arbor, Michigan
Greenberg, A.E., Conners, J.J., Jenkins, D., and Franson, M.A.H. 1981. Standard Methods
for the Examination of Water and Wastewater, 15th ed. American Public Health
Association, Washington, DC
Hoadley, A.W. 1976. Potential health hazards associated with Pseudomonas aeruginosa in
water. In Bacterial Indicators / Health Hazards Associated with Water, ed. A.W.
Hoadley and B.J. Dutka. American Society for Testing and Materials, Philadelphia
Hoadley, A.W., and Dutka, B.J. (eds). 1976. Bacterial Indicators / Health Hazards
Associated with Water. American Society for Testing and Materials, Philadelphia
Holden, W.S. 1970. Water Treatment and Examination. Williams and Williams Company,
Baltimore
Howie, J. W. 1968. Typhoid in Aberdeen, 1964. Journal of Applied Bacteriology, 31,171178
ICMSF. 1978. Microorganisms in Foods. 1. Their Significance and Methods of Enumeration,
Bottles waters, process waters, and ice
243
1980a. Microorganisms in Foods. Vol. 1. Factors Affecting Life and Death of

Microorganisms. Academic Press, New York
1980b. Microorganisms in Foods. Vol. 2. Food Commodities. Academic Press, New York
Levine, R.J., and Nalin, D.R. 1976. Cholera is primarily waterborne in Bangladesh. Lancet,
ii, 1305
Morisetti, M.D. 1979. Microbiological standards for natural waters. Nuovi Annali dIgiene a
Microbiologia, 30 (No. 3-4), 253-264
Mossel, D.A.A. 1976. Various taxo- and ecogroups of bacteria as index organisms for the
enteric contamination of bottled waters: Their significance and enumeration. Annali
dellIstituto Superiore di Sanitd, 12, 177-190
1977. Microbiological quality assurance of water in relation to food hygiene. Archiv fr
Lebensmittelhygiene, 28, 1-2
Mossel, D.A.A., and Oei, H.Y. 1975. Person-to-person transmission of enteric bacterial
infection. Lancet, i, 751
Mossel, D.A.A., Harrewijn, G.A., and van Sprang, F.J. 1973. Microbiological quality
assurance for weaning formulae. In The Microbiological Safety of Food, ed. B.C. Hobbs
and J.H.B. Christian, 77-88. Academic Press, London
Mossel, D.A.A., de Vor, H., and Eelderink, I. 1976. A further simplified procedure for the
detection of P. aeruginosa in often contaminated aqueous substrata. Journal of Applied
Bacteriology, 41, 307-309
Mossel, D.A.A., van Ekeren, A.J.W.M., and Eelderink, 1. 1977. A simplified procedure for the
examination of drinking water for bacteria of public health significance: The differential
hydrobacteriogramme. Zentralblatt far BakteriologieHygiene, i. Abt. Orig. B 165, 498-516
Nolan, C., Harris, N., Ballard, J., Allord, J., and Kobayashi, J. 1982. Outbreak of Yersinia
enterocolitica - Washington State. Morbidity and Mortality Weekly Report (U.S.), 31,
562-564
Panero, C., Grasso, A., La Canza, C., Raguzzini, R., and Bosco, G. 1970. Pollution of drinking
water and infections caused by Pseudomonas aeruginosa in a paediatric ward for immature
babies (in Italian). Giornale di Batteriologia, Virologia, Immunologia, 63, 517-527
Schmidt-Lorenz, W. 1976. Microbiological characteristics of natural mineral water. Annali
dellIstituto Superiore di Sanitd, 12, 93-112
Weber, G., Werner, H.B. and Matschnigg, H. 1977. Pseudomonas aeruginosa in Trinkwasser
als Todesursache bei Neugeborenen. Zentralblatt Bakteriologie Parasitenk. Infektionskr.
Hygiene, t. Abt. Orig. 216, 210-214
WHO (World Health Organization). 1971. International Standards for Drinking Water,
3rd ed. WHO, Geneva
1979. Human Viruses in Water, Wastewater, and Soil. Technical Report Series 639. WHO,
Geneva
1980. Guidelines for Drinking Water Quality Microbiological Aspects (Background
Information). WHO, Geneva
Young, V.M. (ed.) 1977. Pseudomonas aeruginosa: Ecological Aspects and Patient
Colonization. Raven Press, New York
26
Shelf-stable canned foods
Shelf stable canned foods are packed in hermetically sealed containers

and are commercially sterile. Commercial sterility of thermally processed
food means the condition achieved by the application of heat, alone or in
combination with other treatments, to render the food free from
microorganisms capable of growing in the food under normal conditions
of distribution and storage. Such processes include: (a) the botulinum
cook used for low-acid foods (i.e., those having an equilibrium pH
above 4.6); (b) the lesser heat-treatment applied to products containing
curing ingredients such as sodium chloride or sodium nitrite; (c) the
relatively mild heat-treatment given to products of pH 4.6 and below
(acid foods); and (d) the mild heat-treatment given to foods in which the
water activity, or combination of water activity with other preservation
factors such as pH, inhibits growth from bacterial spores.
Many canned foods have the potential to cause botulism. However,
Clostridium botulinum is present in such low numbers, and so infrequently, in such products that no realistic sampling plan would be
adequate as a direct measure of its presence.
In practice the controls applied consist of in-plant control to ensure
that (i) the containers are sound and correctly seamed or sealed, (ii) the
correct thermal process has been applied, (iii) the cooling water is of
suitable microbiological quality, either by chlorination and chemical
monitoring or by microbiological monitoring, and (iv) hygienic practices
during and after cooling are satisfactory.
245
A. RECOMMENDED HYGIENIC PRACTICES FOR CANNED

FOODS
Within recent years a number of codes of practice for commercial

processing of foods in hermetically sealed containers have been published. For example, the United States Food and Drug Administration has
issued a code of good manufacturing practice for low-acid canned foods
(FDA, 1979) as has the UK Department of Health and Social Security
(DHSS, 1981). The Codex Alimentarius Commission has drafted a
Recommended International Code of Hygienic Practice for Low-Acid
and Acidified Low-Acid Canned Foods (Codex, 1983).
B. INTEGRITY OF CONTAINERS
Even when an adequate heat-process has been applied, the integrity of

hermetically sealed containers used for canned foods is critical to safe
processing and requires constant surveillance by the can manufacturer
and canner. A defective or unsound container is one that is of faulty
manufacture, is improperly closed, or has been damaged in such a manner
as to permit recontamination of the can contents following the heatprocess (post-processing contamination). If the contaminant is a pathogen
and the food is capable of supporting its growth, a health hazard exists, as
demonstrated by type E botulism from canned tuna (Johnston et al.,
1963), typhoid from canned corned beef (Howie, 1968); and
staphylococcal food-poisoning from canned peas (Bashford et al., 1960).
Even a sound container, if mishandled, may leak, and if hygiene is poor,
contamination of the contents may occur (Put et al., 1972; Stersky et al.,
1980). To minimize the chances of contamination caused by defective
containers, the container manufacturers, as well as the canners, should
maintain a strict quality assurance program including pressure-testing and
can seam measurements to ensure that predetermined tolerances are met.
Minimum inspection programs for seam and seal checks are detailed in
FDA, 1978, and in guidelines issued by the NFPA (1979). More detailed
guidance on can inspection is available in HPB, 1983; OSU 1982; and
Thorpe and Barker, 1984, which also includes information on containers
with welded side seams.
C. COOLING WATER
In addition to container-integrity checks and process controls, the canner

should use water of suitable microbiological quality for cooling
246
the retorted cans. If appropriate, the water should be chlorinated or

otherwise sanitized. Examination of the cooling water is the responsibility of the canner and should involve regular microbiological examination and, if a sanitizer is used, frequent analysis for the concentration
of the sanitizer present.
D. INCUBATION TESTS
Opinions differ concerning the value of incubating and testing a small

number of cans immediately after processing, but there is no doubt that
such incubation tests are of no value in ensuring the safety of shelf-stable
canned foods. It must also be borne in mind that such limited incubation
tests will reveal only gross underprocessing and/or extensive postprocessing contamination. However, limited incubation tests are still
required by some agencies for certain low-acid canned foods which do
not receive a full botulinum cook, including some cured meats which are
shelf-stable as a result of the presence of curing agents. Incubation tests
may also be of use to the manufacturer to monitor spoilage trends over
long periods of time. It is accepted practice to examine all cans subjected
to incubation tests for swelling, but it must be realized that a variable
proportion of contaminated cans produce swells. Hence, the contents of
as many of the incubated cans as practicable should be examined for pH
change and other evidence of spoilage. The practice of deliberate storage
(or incubation) of all cans after thermal processing, and removal of
swollen cans prior to labelling or distribution, is to be discouraged. It can
lead to the marketing of an unsatisfactorily processed product, which may
be a health hazard as some pathogens (e.g., staphylococci and Salmonella
typhi) do not produce gas in many canned products. It should be
emphasized that no sampling procedure ensures adequate safety with
respect to the botulism hazard. Industrial experience suggests that greater
reliance should be placed on the following: thermal-processing records,
constancy of product formula (solid-to-liquid ratios), volume of fill, the
use of automated instrumentation such as time-temperature recording
devices, and alarm warning systems to signal any interruption of the
specified thermal process.
E. PROCESS VALIDATION AND RECORDS
In addition to data on seam inspection (e.g., tear-down and visual

inspection), thermal processing, and cooling-water quality, it is
247
recommended that information on appropriate chemical composition

(e.g., salt and nitrite content), pH, or any other critical factors be
recorded and made available to the controlling agency. The data on
chemical composition should give, at least, mean values and the standard
deviation or range. If this information gives cause for concern, or is
lacking or inadequate, further can examinations appear advisable (see
Chapter 5, Section N: Investigational Sampling).
F. SAMPLING
For reasons stated above, a sampling plan for the routine surveillance of
shelf-stable canned foods to determine microbiological safety is not
recommended.
REFERENCES
Bashford, T.E., Gillespie, T.G. and Tomlinson, A.J.H. 1960. Staphylococcal Food Poisoning
Associated with Processed Peas. Campden Food Preservation Research Association,
Chipping Campden, Gloucestershire, England
Codex (Codex Alimentarius Commission). 1983. Recommended International Code of Practice
for Low Acid and Acidified Low-Acid Canned Foods, 1st ed. CAC/RCP 23-1979, FAO,
Rome
DHSS (Department of Health and Social Security). 1981. Code of Hygienic Practice for LowAcid Canned Foods, HMSO, London
FDA (United States Food and Drug Administration). 1978. Bacteriological Analytical Manual,
5th ed. Form FD2905 (4/76), Chapter 22. Examination of metal container integrity (R.C.
Lin, P.H. King, and M.R. Johnston)
1979. Rules and Regulations 21 CFR Part 113 - Thermally Processed Low-Acid
Foods Packaged in Hermetically Sealed Containers Current Good Manufacturing Practice.
Federal Register, 44. No. 53 16209-16238
Howie, J.W. 1968. Typhoid in Aberdeen 1964. Journal of Applied Bacteriology, 31, 171-178
HPS (Health Protection Branch). 1983. Draft Document MFA-25. Methods for the Evaluation
of Canned Foods and Their Containers. HPB, Health and Welfare Canada, Ottawa
Johnston, R.W., Feldman, J., and Sullivan, R. 1963. Botulism in canned tuna fish. Public
Health Reports, 78, 561-564
NFPA (National Food Processors Association). 1979. Guidelines for Evaluation and
Disposition of Damaged Canned Food Containers, 2nd ed. Bulletin 38L, NFPA,
Washington, DC
OSU (Oregon State University, Extension Service of the Department of Food Science and
Technology). 1982. Top Double Seaming Manual, 5th printing. OSU, Corvallis, Oregon
Put, H.M.C., Van Doren, H., Warner, W.R., and Kruiswijk, J.T.H. 1972. The mechanism of
microbiological leaker spoilage of canned foods: A review. Journal of Applied
Bacteriology, 35, 7-27
248
Stersky, A., Todd, E., and Pivnick, H. 1980. Food poisoning associated with post-process
leakage (P.P.L.) in canned foods. Journal of Food Protection, 43, 465-467
Thorpe, R.H., and Barker, P.M. 1984. Visual Can Defects. The Campden Food Preservation
Research Association, Chipping Campden, Gloucestershire, England
Glossary
acceptance sampling The application of a predetermined sampling plan to decide

whether or not the lot meets defined criteria of acceptance.
aerobic plate count (APC) The number of colony-forming units of aerobic mesophilic microorganisms present per gram or per ml in the analytical unit as determined by a standard method.
aliquot The portion of food that is inoculated into a container of bacteriological
medium in accordance with a specified method. For example, from a 1 kg sample
unit, a 25 g analytical unit may be blended with 225 ml of diluent to give a 1:10
dilution. A 1 ml portion of this suspension would be an aliquot of 0.1 g.
analytical unit That amount of product withdrawn from a sample unit for an
analysis.
attributes plan A sampling plan in which each selected sample unit is classified
according to the quality characteristics of the product and in which there are only
two or three grades of quality: for example, acceptable, defective; absent, present;
acceptable, marginally acceptable, defective; low count, medium count, high
count.
c The maximum allowable number of defective sample units (2class plan) or marginally acceptable sample units (3class plan). When more than this number are
found in the sample, the lot is rejected. case A set of circumstances related to the
hazard a food represents and to its anticipated subsequent treatment.
consignment A quantity, large or small, of food destined, in commerce, to a particular recipient, usually consisting of multiple containers of food from one or
more lots.
consumers risk The probability that a lot will be accepted on the basis of results of
testing analytical samples when in reality the lot as a whole is substandard relative
to the stated criteria of acceptability.
defective Not suitable for the intended purpose.
frame That portion of the lot from which the sample units are drawn. Ideally it
should be the whole lot; in practice it may be the accessible portion of the lot.
Good Commercial Practice (GCP) A general term defining conditions of Good
Manufacturing Practice combined with acceptable conditions of distribution and
storage.
276
Glossary
Good Manufacturing Practice (GMP) Those procedures in a food-processing

plant which consistently yield products of acceptable quality. GMP is usually
described in a code defining processes, equipment, plant layout, sanitation,
hygiene, and laboratory tests; but, as used here, it does not apply to any
particular code now in use.
hazard The potential to cause (i) disease when eaten or (ii) organoleptic spoilage
in a period shorter than the normal shelf-life of the food.
Hazard Analysis Critical Control Point (HACCP) A system in which observations
and/or tests are made to identify actual or potential hazards in operations and to
identify critical control points in a process. Control measures are designed and
implemented, and the control points are monitored to ensure that control is
maintained.
indicator Historically, an organism itself non-pathogenic, but often associated with
pathogens, used to portray a risk of the presence of pathogens for which feasible
methods of detection were not generally available (sometimes called index organisms). This usage is currently expanded to denote groups or species of organisms
whose presence in a food reveals exposure to conditions that might introduce
hazardous organisms and/or allow their growth. Specific indicators are now used
to reveal excessively contaminated raw materials; unsanitary manufacturing
practice; contamination from faecal, nasopharyngeal, or suppurative sources;
unsuitable time-temperature conditions of storage; or failure of a process. The
indicator groups and their methods of enumeration are described inICMSF,
Microorganisms in Foods, I. Their Significance and Methods of Enumeration
(1978).
lot In the commercial sense, a lot is a quantity of food supposedly produced under
identical conditions, all packages of which would normally bear a lot number that
identifies the production during a particular time interval, and usually from a
particular line, retort, or other critical processing unit (see Chapter 3, Section A).
Statistically, a lot is considered as a collection of units of a product from which a
sample is to be drawn to determine acceptability of the lot.
m A microbiological limit which, in a 2class plan, separates good quality from
defective quality or, in a 3class plan, separates good quality from marginally
acceptable quality. In general, values equal to m, or below, represent an acceptable
product and values above it are either marginally acceptable or unacceptable.
M A microbiological limit which, in a 3class plan, separates marginally acceptable quality from defective quality. Values above M are unacceptable.
microbiological criterion A criterion concerning the presence or quantity of
microorganisms or their toxins per unit of mass, volume, or area, established by
use of defined procedures and applied in acceptance sampling of foods (see
Appendix 6: General Principles for the Establishment and Application of Microbiological Criteria for Foods).
microbiological guidelines A microbiological criterion used by a manufacturer or
regulatory agency to monitor a food, ingredient, process, or system; often used
also to describe a microbiological criterion where no standard has been prescribed. The Codex Alimentarius definition is more restrictive (see Appendix 6:
General Principles for the Establishment and Application of Microbiological
Criteria for Foods).
microbiological purchasing specification The microbiological criterion or criteria
Glossary
277
conditional to acceptance of a specific food or food ingredient by a food manufacturer or other private or public purchasing agency.
microbiological standard A microbiological criterion in a law or regulation controlling foods produced, processed, or stored in the area of jurisdiction of a regulatory agency, or imported into the area.
moderate hazard The hazard associated with ingestion of a food containing a
pathogen or toxin which can cause a short-term disease not critically severe in its
manifestations and normally without sequelae: for example, salmonellosis, staphylococcal food-poisoning.
Most Probable Number (MPN) A statistically estimated number (per unit of mass
or volume) of the test organism present in a sample unit, based upon its presence
or absence in replicate aliquots (see also ICMSF, Microorganisms in Foods, 1.
Their Significance and Methods of Enumeration [1978]).
n The number of sample units which are examined from a lot to satisfy the
requirements of a particular sampling plan.
population (epidemiological) The number of people estimated to be at risk from a
particular food.
population (microbiological) The number of microorganisms, total or of a species
or group(s), dispersed within a defined quantity of food and detectable by the
methods used. Hence, a total microbial population or a population of staphylococci, coliforms, etc.
population (statistical sampling) The aggregate of units about which an inference
is to be made from analytical results. In this book, population usually relates to
the totality of hypothetical individual portions within a lot, each identical in
quantity with the defined sample unit.
probability of acceptance The likelihood that a lot of known quality would be
accepted on the basis of results of testing analytical samples relative to the stated
criteria.
producers risk The probability that a lot of acceptable quality will be rejected on
the basis of results of testing analytical samples relative to the stated criteria.
random sample of size n A collection of n sample units obtained in such a way that
every sample of size n has an equal chance of selection, hence excluding bias.
Reference to a table of random numbers is usually involved (see Chapter 3,
Section C).
sample The total number (one or more) of individual sample units drawn from a
lot (ideally at random) which will be tested in accordance with a specific sampling plan and method(s).
sample unit The individual portion or container of food taken as part of the sample
and from which one or more analytical units may be drawn for analysis.
sampling plan Used synonymously in this text with lot acceptance sampling plan,
and meaning a statement of the criteria of acceptance to be applied to a lot, based
on examination of a required number of sample units by defined analytical
methods.
severe hazard The hazard associated with the presence of a pathogen or

toxin in a food which when ingested is likely to cause severe disease, either
in a healthy population or in a particularly susceptible group to which the
food in question is often destined. Severity of hazard is largely determined
by clinical severity of the disease induced, but many ancillary factors are
pertinent (see Chapter 5 and Tables 6 and 7). Pathogens of concern in this
category include Brucella spp., Clostridium botulinum,
278
Glossary
Cl. perfringens type A, Salmonella typhi, Shigella dysenteriae, Vibrio cholerae O1

(Table 7); they are normally sought by application of investigational sampling (see
Chapter 5, Section N).
shelf-life The period after processing during which a food remains in acceptable
condition under appropriate conditions of storage.
stratification A technique for dividing consignments into logical groupings such as
size of cans, grades, box, cases, which controls known sources of variation. It may
be used where prior knowledge exists that the consignment is potentially not of
uniform quality.
stringency The probability that a lot of food of defined microbiological quality will
be rejected by a prescribed sampling plan, relative to other sampling plans.
subsample The several individual swabs or portions of food taken from a single
carcass or carton and which make up a single sample unit.
variables plan A sampling plan in which the criteria employ statistics of a sample,
such as the mean and standard deviation, rather than the segregation of sample
units into attribute classes (see attributes plans). In the context of this book the
variables plans referred to are based on the application of the mean (x) and standard deviation (s) of the sample log counts. These parameters are combined with a
constant or constants (k1, k2) selected for the appropriate number of sample units
(n) for comparison with a limit value (v or V).

2 Sampling For Microbiological Analysis: Principles and Specific Applications

Uploaded by

Copyright:

Available Formats

2 Sampling For Microbiological Analysis: Principles and Specific Applications

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2 Sampling For Microbiological Analysis: Principles and Specific Applications

Uploaded by

Copyright:

Available Formats

MICRO

Blackwell Scientific Publications

PART II. SPECIFIC PROPOSALS FOR

A. Format of commodity chapters

B. Selection and application of criteria

11 SAMPLING PLANS FOR RAW MEATS

12 SAMPLING PLANS FOR PROCESSED MEATS

13 SAMPLING PLANS FOR POULTRY AND POULTRY PRODUCTS

A. Reasons for sampling

14 SAMPLING PLANS FOR FEEDS OF ANIMAL ORIGIN AND PET

C. Choice of microbiological criteria

15 SAMPLING PLANS FOR MILK AND MILK PRODUCTS

A. Relatively perishable products

B. Relatively stable products

16 SAMPLING PLANS FOR EGGS AND EGG PRODUCTS

17 SAMPLING PLANS FOR FISH AND SHELLFISH

A. Public health aspects of fish as human food

B. The basis for specific sampling proposals

18 SAMPLING PLANS FOR VEGETABLES, FRUITS, AND NUTS

A. Raw vegetables and fruits

B. Frozen vegetables and fruits

C. Canned vegetables and fruits

D. Dried vegetables and fruits

E. Yeasts and single-cell proteins

19 SAMPLING PLANS FOR SOFT DRINKS, FRUIT JUICES,

20 SAMPLING PLANS FOR CEREALS AND CEREAL PRODUCTS

A. Grains and flours

B. Frozen and dried pasta products

C. Frozen filled and topped bakery products

D. Sampling procedures and methods

21 SAMPLING PLANS FOR SPICES, CONDIMENTS, AND GUMS

B. Collection of spices and gums

C. Microbiology of spices, gums, and herbs

22 SAMPLING PLANS FOR FATS AND OILS

23 SAMPLING PLANS FOR SUGAR, COCOA, CHOCOLATE, AND

I. Sugar and related sweeteners

A. Reasons for sampling

II. Cocoa and chocolate

A. Reasons for sampling

A. Reasons for sampling

IV. Sampling and testing procedures for cocoa, chocolate liquor,

24 SAMPLING PLANS FOR FORMULATED FOODS

25 SAMPLING PLANS FOR NATURAL MINERAL WATERS, OTHER

A. Definitions and economic importance of waters in international trade

B. Reasons for sampling

26 SHELF-STABLE CANNED FOODS

A. Recommended hygienic practices for canned foods

E. Process validation and records

This chapter provides background information on the considerations

Proposals for sampling and sample plans

The application and use of criteria

The microbiological status of some foods changes continually during

The ICMSF has attempted to include in criteria those microbes most

For 3class attributes plans it was necessary to establish m values

World trade in raw meats, both chilled and frozen, continues to be

of cooking and, more important, to hot holding, cooling, and reheating

Proposals for sampling and sampling plans

In the virtual absence of studies relating health risks to surface counts of

to assess the effectiveness of sanitation factors in limiting microbial