Advanced Food Microbiology

Module 9: Course assessment

Microbiological sampling and testing

Overview

You may wonder why we undertake microbiological testing on food products in the first place. There are many reasons for doing so in modern food production and it is important to have an understanding of why sampling and testing is a key part of the food safety system.

Even if we have a kill step in place and sufficient controls to ensure that products do not become contaminated, end product testing can throw up issues post kill step – where contamination is introduced to the product once it has been cleared of other microbes.

If microbial contamination occurs at this stage, the lack of competition will mean that the new organisms introduced will be free to grow without competition for nutrients. Laboratory testing is a useful verification step that all food safety controls and processes are working correctly.

Occasionally, food safety controls may not work properly, due to operator error or issues with equipment – for example, a cook step may not reach the desired time and temperature to kill vegetative cells, or rapid cooling of cooked product may take too long due to issues with the blast chiller.

Because of this, lab testing is a requirement of food regulations globally and part of any contractual arrangement with retail customers – who will often set their own specifications for microbiological quality that are more stringent than the regulatory ones.​

Sample collection is just as important as the testing and there are a number of methods employed, depending on the circumstances and requirements of the testing.

Sampling plans

Hazard groupReduceNo changeIncrease
UtilityCase 1: n = 5, c = 3Case 2: n = 5, c = 2Case 3: n = 5, c = 1
IndicatorCase 4: n = 5, c = 3Case 5: n = 5, c = 2Case 6: n = 5, c = 1
ModerateCase 7: n = 5, c = 2Case 8: n = 5, c = 1Case 9: n = 10, c = 1
SeriousCase 10: n = 5, c = 0Case 11: n = 10, c = 0Case 12: n = 20, c = 0
SevereCase 13: n = 15, c = 0Case 14: n = 30, c = 0Case 15: n = 60, c = 0
ICMSF risk based sampling plans

The table above is based on the cases sampling plans developed by the International Commission on Microbiological Specifications for Food (ICMSF) – a global body of experts who provided scientific evidence for adoption of various microbiological controls, HACCP and food safety objectives.

The following key can be used to help interpret the sampling plan:

Utility: testing for spoilage organisms (total plate count, yeasts and mould) and reduced shelf life – no public health concern

Indicator: measures of process hygiene controls, including organisms such as Enterobacteriaceae, coliforms, E. coli

Moderate hazard: organisms that are not life threatening, illness is of short duration, self limiting – includes organisms such as Staphylococcus aureus, Bacillus cereus, Clostridium perfringens

Serious hazard: illness is incapacitating, but generally not life threatening – organisms such as Salmonella, Shigella flexneri, Yersinia enterocolitica

Severe hazard: illness can be life threatening, of long duration and with chronic sequelae – organisms include E. coli O157:H7, Clostridium botulinum toxin or Cronobacter in infants

The above is an example of 2-class attributes sampling plans. In food safety we also use 3-class attributes sampling plans.

A 2-class attributes sampling plan has a single limit for each organism and has 2 variable attributes that effect the acceptance criteria. Looking at this table, you will see that for Cases 1 and 4, n = 5 and c = 3. What this tells us is that we test 5 samples and out of those 5 samples tested 3 are permitted to be out of specification.

We would also have a level “m” which separates acceptable quality from marginally acceptable or unacceptable, as below:

TestncLimit (m)​
Standard plate count​5​1​<1,000cfu/g​
Enterobacteriaceae​5​1​<100cfu/g​
Yeast/Mould​5​1​<100cfu/g​
Listeria monocytogenes​5​0​ND in 25g​

For 3-class attributes sampling plans, the 3rd variable introduced is the “M” limit. With this type of sampling plan, the number of samples to be taken is again (n), the number of non-conforming samples allowed is (c) and any sample that exceeds the second limit (M) is unacceptable, along with the whole batch or lot.

Therefore, less than (m) is acceptable; small (m) to big (M) is marginal and over (M) is unacceptable for the lot from which the samples are taken. 3-class sampling plans are preferred in regulations where the results are quantitative (a count), rather than qualitative (presence/absence). Regulatory limits generally use 3-class sampling plans, except where the risk of a pathogen is higher and a presence/absence test provides more robust safety acceptance criteria.​

TestncmM
Standard plate count​5​2​<1,000cfu/g​≥10,000cfu/g​
Enterobacteriaceae​5​2​<10cfu/g​≥100cfu/g​
Yeast/Mould​5​2​<100cfu/g​≥1,000cfu/g​

Microbiological testing

There is a wide range of lab testing methods that can be applied to both products and environment that will provide information on the presence or absence of bacteria, yeasts and moulds.

From general settle plate analysis to gather information about the quality of the air in the facility, to product specific analyses such as that for wild yeasts in the brewing process, it is possible to build a clear picture of the levels of contamination that are or could be present in your products.

Microbiological testing is a valuable part of the verification process for HACCP or food safety plans, rather than a substitute for such plans. The risk assessment element of HACCP is vital for gaining accurate information on product characteristics and the risks associated with it in your particular facility.​

Types of microbial tests

Presence/absence testing is appropriate where there is a need to quickly find out whether a particular pathogen is present in a product – and its presence would mean rejection of the product. For presence/absence tests, the specification is usually “Not Detected in 25g”of product and is generally a regulatory requirement

For example, in ready to eat (RTE) foods that support the growth of Listeria, you would need to use a presence/absence test.

Enumeration tests are useful when a pathogen can be present in the food, but at levels that are too low to cause illness because the formulation of the food will prevent widespread growth of the organism. Taking Listeria again – in foods that do not support the growth of Listeria, you would do an enumeration test and the specification for this would be <100cfu/g. These specifications can be found in the regulations for a range of foods (Food Standards Code, Schedule 27 – Microbiological limits in food).​

 

Laboratory testing is along established method for determining or enumerating levels of microorganisms in food and beverage products. Different media (the plates in the image above) are used to select for specific organisms that favour growth in that particular medium.

Examples of the range of testing available are:

  • Total yeast and mould enumeration of swabs​
  • Yeast and mould – foods with high and reduced Aw​
  • Yeast and mould – osmophilic enumeration​
  • Wild yeasts – products from brewing process, filtration and spread plate methods​
  • Settle plate analysis​ for air quality assessment
  • Spoilage bacteria and yeasts in the brewing process​
  • Preservative resistant yeast – enumeration and presence/absence​
  • Hygienic state of surfaces – contact plate​
  • Enumeration of spoilage organisms in acidic products​
  • Pathogen presence/absence
  • Pathogen enumeration
  • Anaerobic plate count
  • Spore formers enumeration
  • Anaerobic spore formers

Each test has a Method that is either accredited by the National Association of Testing Authorities (NATA) or an in house method that is not formally accredited by NATA.

NATA accredited methods are based on internationally recognised standards and are consistent across laboratories. Ensuring that your testing laboratory is NATA accredited means that you can be confident that all processes, methods and training is audited on a regular basis and complies with international and national quality standards (such as ISO 17025).

Laboratories will also offer a choice of rapid and traditional methods for testing foods. The graphic below summarises some of the key differences between the ELISA and Australian Standard methods for pathogens such as Listeria and Salmonella. Both methods are NATA accredited, but ELISA testing will take half the time and is less resource intensive than the Australian Standard. 

Results interpretation

Food microbiology is a complex area, within which there is an element of uncertainty when it comes to interpreting results

The particulate nature of microorganisms and their ability to reproduce rapidly under favourable conditions has an impact on accuracy when it comes to testing.

Microbiological specifications for a range of food products have been established using robust scientific methods and captured in international standards and regulations where appropriate (Australian specifications are available in the Food Standards Code, Schedule 27 and in the FSANZ “Compendium of Microbiological Criteria for Food“.

For Qualitative methods (Detected or Not Detected), a positive (ie Detected) result in many cases means that the sample/batch is unacceptable and further action must be taken. The extent of the action will depend on a number of factors, including: regulatory requirements, industry standards, customer specifications and company HACCP plan.

Quantitative methods provide a count of microorganisms in a given amount of sample – for example, Standard Plate Count 150 cfu/g.

What is acceptable here will also depend on specifications derived from a combination of the above factors.

For Environmental Monitoring, there is less guidance and no internationally recommended specifications. This is simply because all facilities are different, with different risks.

It is therefore crucial that, prior to any testing commencing, a thorough risk analysis is conducted, including sampling points, sampling frequency, target organisms, specifications, control measures and corrective actions.

Occasionally, you may receive notification of a presumptive or suspect result – for example, Listeria. This means that the generic first part of the test has detected a potential positive which requires further confirmation prior to issuing a Certificate of Analysis.

Sometimes the suspect organism is the target one – for example Listeria monocytogenes. Other times, it is another similar organism (there are now 20 species of Listeria identified, one of which is Listeria monocytogenes): therefore, if doing a presence/absence for Listeria species, we would need to test further in order to confirm that it is Listeria monocytogenes.

Presumptive results are therefore those where there is a two stage confirmatory process to a method and the first stage has indicated that this second stage is required.

A “presumptive” positive warning is issued to the client prior to the confirmation stage being undertaken.

Definitions/terms

ND – not detected – for qualitative methods where there is a simple detected or not detected response

CFU – colony forming unit – used to estimate the number of viable bacteria or fungal cells in a sample, expressed in mL (for liquids) and g (solids)

CoA – Certificate of Analysis – a formal, controlled document showing the test method, sample name and test result, signed by an official signatory

Less than (<) 

“When reporting a result where no organisms are detected, you cannot report zero, so would need to report <1 (in the case of liquids) or <10 (in the case of solids – due to a dilution taking place).”

ISO 7218 (Microbiology of food and animal feedingstuffs — General requirements and guidance for microbiological examinations)

Subsequent dilutions would affect how the results were reported (for example, <100 or <1000).

Greater than (>) – a reported result of >x indicates that the number of viable organisms exceed the maximum amount that can be reported for that particular method (also can be expressed as TNTC – Too Numerous To Count) 

E or ~ – Estimated – this is used if an actual count lies outside of the reportable range of results for a particular method.

For example, if a method has a reporting range from 50 – 500 and 40 cfus were counted, the result would be reported as 40E or ~40