Predictive food microbiology as a tool in risk ... - ProSafeBeef
Predictive food microbiology as a tool in risk ... - ProSafeBeef
Predictive food microbiology as a tool in risk ... - ProSafeBeef
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
<strong>Predictive</strong> <strong>food</strong> <strong>microbiology</strong> <strong>as</strong> a <strong>tool</strong> <strong>in</strong><br />
<strong>risk</strong> <strong>as</strong>sessment<br />
Kost<strong>as</strong> Koutsoumanis<br />
Assistant Professor<br />
Lab of Food Microbiology and Hygiene Dpt. Of Food<br />
Science and Technology<br />
Aristotle University of Thessaloniki,
Presentation Outl<strong>in</strong>e<br />
1.Food Safety Management<br />
1.1 The Risk Analysis process<br />
1.2 Risk Assessment<br />
1.3 Risk Management<br />
2.<strong>Predictive</strong> Microbiology<br />
2.1 K<strong>in</strong>etic Models<br />
2.1.1 Primary<br />
2.1.2 Secondary<br />
2.1.3 Tertiary (software)<br />
2.2 Growth/no growth boundary models<br />
2.3 Stoch<strong>as</strong>tic Models<br />
3.C<strong>as</strong>e Studies<br />
3.1 Exposure <strong>as</strong>sessment of L monocytogenes <strong>in</strong><br />
p<strong>as</strong>teurized milk<br />
3.2 Risk-b<strong>as</strong>ed approach for evaluat<strong>in</strong>g compliance of<br />
RTE <strong>food</strong> to the safety criteria for L. monocytogenes
Food Safety Management<br />
Traditional Food Safety Management approach w<strong>as</strong> b<strong>as</strong>ed<br />
on end-product test<strong>in</strong>g<br />
End Product Sampl<strong>in</strong>g<br />
Is the product<br />
safe?<br />
Yes<br />
No<br />
Food Safety<br />
Qualitative (Discrete)<br />
variable
Food Safety Management<br />
Traditional Food Safety Management approach w<strong>as</strong> b<strong>as</strong>ed<br />
on end-product test<strong>in</strong>g<br />
Food Safety Crisis of 90’s<br />
Antibiotics<br />
Listeria<br />
BSE<br />
SARS<br />
Diox<strong>in</strong>s<br />
E. coli<br />
Campylobacter<br />
Acrylamide<br />
Salmonella<br />
Need for change
Food Safety Management<br />
Development of new <strong>tool</strong>s<br />
New Tools<br />
<strong>Predictive</strong> Microbiology<br />
Risk Assessment
Risk Analysis<br />
Risk Assessment is a component of Risk Analysis<br />
(WHO/FAO, 1995):<br />
Risk<br />
Assessment<br />
Risk<br />
Communication<br />
Risk<br />
Management
Risk Analysis<br />
Risk Assessment is a component of Risk Analysis
Risk Analysis<br />
The Risk Analysis Process - CAC
Risk Assessment<br />
Risk Assessment Stages<br />
Hazard Identification: what biological, chemical and<br />
physical agents are we deal<strong>in</strong>g with and with which <strong>food</strong>s is<br />
it <strong>as</strong>sociated?<br />
Hazard Characterization: what illness can be caused,<br />
<strong>as</strong>sociated <strong>in</strong> relation to dose and population?<br />
Exposure Assessment: how likely it is that an <strong>in</strong>dividual or<br />
a population will be exposed to a microbial hazard and what<br />
numbers of organisms are likely to be <strong>in</strong>gested?<br />
Risk Characterization: the <strong>in</strong>tegration of the above<br />
result<strong>in</strong>g <strong>in</strong> the probabilities of illness
Risk Assessment<br />
Exposure Assessment<br />
Exposure <strong>as</strong>sessment provides an estimate of the<br />
occurrence and level of the pathogen <strong>in</strong> a specified portion<br />
of a certa<strong>in</strong> <strong>food</strong> at the time of consumption <strong>in</strong> a specified<br />
population by tak<strong>in</strong>g <strong>in</strong>to account uncerta<strong>in</strong>ty and variability<br />
Exposure = f(Contam<strong>in</strong>ation ; Consumption)<br />
Prevalence (proportion of<br />
contam<strong>in</strong>ated units)<br />
and levels of contam<strong>in</strong>ation<br />
(<strong>in</strong> contam<strong>in</strong>ated units)<br />
at the time of consumption<br />
Consumption rates<br />
and serv<strong>in</strong>g sizes
Exposure Assessment<br />
Data and <strong>tool</strong>s required for Exposure Assessment<br />
‣Food product flows<br />
-Farm-to-fork<br />
-Import and export<br />
‣Pathogen dynamics at start<strong>in</strong>g po<strong>in</strong>t<br />
-Prevalence<br />
-Numbers<br />
‣Process<strong>in</strong>g<br />
-Conditions<br />
-Effects (models)<br />
‣Distribution, retail and domestic storage<br />
-Conditions (Time-Temperature)<br />
-Effects (models<br />
‣Consumption patterns<br />
-Intake<br />
-Preparation
Exposure Assessment<br />
Variability and Uncerta<strong>in</strong>ty (Def<strong>in</strong>itions)<br />
Variability represents a true heterogeneity of the<br />
population that is a consequence of the physical system and<br />
irreducible (but better characterized) by further<br />
me<strong>as</strong>urements.<br />
‣Variability between sub-populations<br />
Examples: differences <strong>in</strong> serv<strong>in</strong>g sizes between<br />
<strong>in</strong>fants/children/teenagers/adults, male versus female…<br />
‣Variability with<strong>in</strong> a (sub-) population<br />
Examples: variability of serv<strong>in</strong>g sizes from one person to<br />
another,from one serv<strong>in</strong>g (cocktail) to another (ma<strong>in</strong> meal)…
Exposure Assessment<br />
Variability and Uncerta<strong>in</strong>ty (Def<strong>in</strong>itions)<br />
Uncerta<strong>in</strong>ty represents our lack of knowledge and may<br />
<strong>in</strong>clude:<br />
‣scenario uncerta<strong>in</strong>ty<br />
‣model uncerta<strong>in</strong>ty<br />
‣parameter uncerta<strong>in</strong>ty
Exposure Assessment<br />
Variability (Example)<br />
We all want to move to the 5 th floor us<strong>in</strong>g the elevator <strong>in</strong><br />
groups of 5 (randomly selected) people<br />
The weight limit of the elevator is 480 kg<br />
Estimate the chance of exceed<strong>in</strong>g the weight limit<br />
Determ<strong>in</strong>istic method (variability is not taken <strong>in</strong>to account)<br />
Average <strong>in</strong>dividual weight=70 kg<br />
5 persons x 70 =350 kg
Exposure Assessment<br />
Variability (Example)<br />
Stoch<strong>as</strong>tic method (variability is taken <strong>in</strong>to account)<br />
25<br />
Normal (70,8 kg)<br />
20<br />
%Frequency<br />
15<br />
10<br />
5<br />
0<br />
50 55 60 65 70 75 80 85 90 95 100<br />
Individuals average Weight
Variability (Example)<br />
Exposure Assessment<br />
Stoch<strong>as</strong>tic method (variability is taken <strong>in</strong>to account)<br />
%Frequency<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
50 55 60 65 70 75 80 85 90 95 100<br />
Individuals average Weight<br />
Random selection of<br />
5 values<br />
Repeat 100000<br />
(iterations)<br />
Sum the 5 values<br />
0,1<br />
Monte Carlo<br />
Simulation<br />
Probability<br />
0,08<br />
0,06<br />
0,04<br />
0,02<br />
Limit<br />
P(>480)=3x10 -4<br />
0<br />
240<br />
260<br />
280<br />
300<br />
320<br />
340<br />
360<br />
380<br />
400<br />
420<br />
440<br />
460<br />
480<br />
500<br />
Total Weight of 5 persons
Exposure Assessment<br />
Uncerta<strong>in</strong>ty (Example)<br />
Stoch<strong>as</strong>tic method (variability is taken <strong>in</strong>to account)<br />
Uncerta<strong>in</strong>ty: We don’t know the weight limit of the<br />
elevator<br />
Expert Op<strong>in</strong>ion: M<strong>in</strong>:450, Max:550 Most likely:500<br />
Triang(450; 500; 550)<br />
2,5<br />
X
0<br />
0<br />
2<br />
1<br />
4<br />
2<br />
3<br />
6<br />
8<br />
0<br />
2<br />
4<br />
0<br />
6<br />
8<br />
1<br />
2<br />
3<br />
0<br />
1<br />
2<br />
3<br />
0,40<br />
0,35<br />
0,30<br />
0,25<br />
Expon(2,5) Shift=-2,5<br />
Exposure Assessment<br />
0,40<br />
Triang(-2,5; 0; 2,5)<br />
0,20<br />
0,35<br />
0,15<br />
0,30<br />
0,10<br />
0,25<br />
0,05<br />
0,20<br />
0,00<br />
-4<br />
-2<br />
90,0%<br />
-2,37 4,99<br />
5,0%<br />
Contam<strong>in</strong>ation<br />
at t=0<br />
10<br />
><br />
0,15<br />
0,10<br />
0,05<br />
0,00<br />
-3<br />
-2<br />
-1<br />
5,0% 90,0%<br />
5,0%<br />
-1,709 1,709<br />
Consumption<br />
data<br />
MODELS (growth<strong>in</strong>activation-survival<br />
Contam<strong>in</strong>ation at<br />
consumption time<br />
Normal(0; 1)<br />
0,40<br />
Triang(-2,5; 0; 2,5)<br />
Monte<br />
Carlo<br />
0,45<br />
0,40<br />
0,35<br />
0,35<br />
0,30<br />
0,30<br />
0,25<br />
0,20<br />
0,15<br />
0,25<br />
0,20<br />
0,15<br />
Exposure<br />
0,10<br />
0,10<br />
0,05<br />
0,00<br />
-3<br />
-2<br />
-1<br />
0,05<br />
0,00<br />
Lognorm(2,5; 2,5) Shift=-2,5<br />
< 5,0% 90,0%<br />
5,0% ><br />
-1,645 1,645<br />
-3<br />
-2<br />
-1<br />
5,0% 90,0%<br />
5,0%<br />
-1,709 1,709<br />
0,40<br />
0,35<br />
Process<strong>in</strong>g,<br />
distribution,<br />
storage<br />
parameters<br />
0,40<br />
0,35<br />
0,30<br />
0,25<br />
0,20<br />
0,15<br />
0,10<br />
0,05<br />
0,00<br />
-4<br />
-2<br />
Lognorm(2,5; 2,5) Shift=-2,5<br />
90,0%<br />
-2,05 4,45<br />
5,0%<br />
10<br />
12<br />
><br />
0,30<br />
0,25<br />
0,20<br />
0,15<br />
0,10<br />
0,05<br />
0,00<br />
-4<br />
-2<br />
0<br />
2<br />
4<br />
90,0%<br />
-2,05 4,45<br />
6<br />
8<br />
5,0%<br />
10<br />
12<br />
>
Risk Characterization<br />
< 5,0% 90,0%<br />
5,0% ><br />
-1,645 1,645<br />
Exposure<br />
Dose-Response<br />
Lognorm(2,5; 2,5) Shift=-2,5<br />
100<br />
0,40<br />
0,35<br />
0,30<br />
0,25<br />
0,20<br />
0,15<br />
0,10<br />
0,05<br />
0,00<br />
-4<br />
-2<br />
0<br />
2<br />
4<br />
90,0%<br />
-2,05 4,45<br />
6<br />
8<br />
5,0%<br />
10<br />
12<br />
><br />
Frequency of Infection, %<br />
90<br />
80<br />
70<br />
60<br />
50<br />
40<br />
30<br />
20<br />
10<br />
0<br />
0 1 2 3 4 5<br />
Dose (Log10 cfu)<br />
6<br />
7<br />
Data<br />
Gompertz-Log<br />
Probability of illness<br />
Normal(0; 1)<br />
0,45<br />
0,40<br />
0,35<br />
0,30<br />
0,25<br />
0,20<br />
0,15<br />
0,10<br />
0,05<br />
0,00<br />
-3<br />
-2<br />
-1<br />
0<br />
1<br />
2<br />
3
Food Safety Management<br />
Microbiological Risk Assessment estimates<br />
Number of c<strong>as</strong>es of (a certa<strong>in</strong>) illness per year per<br />
(e.g.) 100.000 persons <strong>in</strong> a given population caused<br />
by a certa<strong>in</strong> micro-organism or group of<br />
microorganisms <strong>in</strong> a particular <strong>food</strong> or <strong>food</strong> type<br />
The output of Risk Assessment<br />
is a probability (Risk)<br />
e.g 10 -6
Food Safety Management<br />
Food Safety Management approach<br />
Food Safety<br />
Is the product safe? Yes/No<br />
Qualitative (Discrete)<br />
variable<br />
Risk Assessment<br />
Quantitative (Cont<strong>in</strong>ues)<br />
variable<br />
What is the degree of product safety? Risk
Food Safety Management<br />
Food Safety Management approach<br />
ACCEPTABLE RISK?<br />
100% Safety (zero <strong>risk</strong>) does not<br />
exists and should not be expected
Food Safety Management<br />
Food Safety Management approach<br />
1996: WTO/SPS agreement<br />
Appropriate Level of Protection (ALOP)<br />
Level of protection deemed appropriate by a<br />
member (country) establish<strong>in</strong>g a sanitary or<br />
phytosanitary me<strong>as</strong>ure to protect human, animal or<br />
plant life or health with<strong>in</strong> its territory.
Food Safety Management<br />
Food Safety Management approach<br />
Appropriate Level of Protection (ALOP)<br />
Degree of <strong>risk</strong> that a society is will<strong>in</strong>g to<br />
tolerate/accept<br />
The “costs” that society is will<strong>in</strong>g to bear to<br />
achieve a specific degree of control over a<br />
Hazard<br />
“Costs” <strong>in</strong>cludes: human, quality, nutritional,<br />
economic, ethical, medical, legal, etc<br />
ALOP example: less than 0.25 c<strong>as</strong>es of<br />
Listeriosis per 100,000 people per year <strong>in</strong> USA
Food Safety Management<br />
Food Safety Management approach<br />
Need for conversion of ALOP <strong>in</strong>to actions. To<br />
implement actions, an objective must be def<strong>in</strong>ed,<br />
expressed <strong>in</strong> terms of a level of a hazard <strong>in</strong> a<br />
<strong>food</strong><br />
Food Safety Objective (FSO)<br />
The maximum frequency and/or concentration of<br />
a microbial hazard <strong>in</strong> a <strong>food</strong> at the moment of<br />
consumption that provides the ALOP
Food Safety Management<br />
Food Safety Management approach<br />
Establishment FSO b<strong>as</strong>ed on ALOP<br />
Dose-Response relationship
Food Safety Management<br />
Food Safety Management approach<br />
ICMSF Safety Equation<br />
H 0 - ΣR+ ΣG+ ΣC < FSO.<br />
H 0 : Initial contam<strong>in</strong>ation<br />
ΣR: Total Reduction (log CFU/g)<br />
ΣG: Total Growth (log CFU/g)<br />
ΣC: Total Contam<strong>in</strong>ation (log CFU/g)
Food Safety Management<br />
Conversion of FSO to practical criteria for the<br />
<strong>food</strong> <strong>in</strong>dustry<br />
Performance Objective: The maximum frequency and/or<br />
concentration of a (microbial) hazard <strong>in</strong> a <strong>food</strong> at a<br />
specified step <strong>in</strong> the <strong>food</strong> cha<strong>in</strong> before time of consumption<br />
that still provides or contributes to the achievement of an<br />
FSO.<br />
Performance Criterion: The effect of one or more control<br />
me<strong>as</strong>ure(s) needed to meet or contribute to meet<strong>in</strong>g a PO.<br />
Process Criterion: The conditions pf a process that lead to<br />
a PO.
Food Safety Management: Summary<br />
Gorris, 2004
<strong>Predictive</strong> Microbiology<br />
THE CONCEPT<br />
A detailed knowledge of microbial responses to<br />
environmental conditions, synthesized <strong>in</strong> a<br />
mathematical model, enables objective evaluation<br />
of process<strong>in</strong>g, distribution and storage operations<br />
on the microbiological safety and quality of <strong>food</strong>s,<br />
by monitor<strong>in</strong>g the environment without<br />
recourse to further microbiological analysis
<strong>Predictive</strong> Microbiology<br />
THE CONCEPT<br />
‣Growth, survival and <strong>in</strong>activation of microorganisms <strong>in</strong><br />
<strong>food</strong>s are reproducible responses<br />
‣A limited number of environmental parameters <strong>in</strong> <strong>food</strong>s<br />
determ<strong>in</strong>e the k<strong>in</strong>etic responses of microorganisms<br />
(Temperature, Water activity/water ph<strong>as</strong>e salt, pH, Food<br />
preservatives<br />
‣A mathematical model that quantitatively describes the<br />
comb<strong>in</strong>ed effect of the environmental parameters can be<br />
used to predict growth, survival or <strong>in</strong>activation of a<br />
microorganism and thereby contribute important <strong>in</strong>formation<br />
about product safety and shelf-life<br />
Roberts and Jarvis (1983)
<strong>Predictive</strong> Microbiology<br />
APPLICATIONS<br />
‣Predict the effect of product characteristics and storage<br />
conditions on microbial responses (safety and shelf-life)<br />
‣Predict effect of changes <strong>in</strong> parameters (product<br />
development)<br />
‣HACCP plans – establish limits for CCP<br />
‣Food safety objectives – equivalence of processes<br />
‣Education – e<strong>as</strong>y access to <strong>in</strong>formation<br />
‣Quantitative microbiological <strong>risk</strong> <strong>as</strong>sessment (QMRA)<br />
(The concentration of microbial hazards <strong>in</strong> <strong>food</strong>s may<br />
<strong>in</strong>cre<strong>as</strong>e or decre<strong>as</strong>e substantially dur<strong>in</strong>g process<strong>in</strong>g and<br />
distribution)
<strong>Predictive</strong> Microbiology<br />
TYPES OF MODELS<br />
‣Primary models: describ<strong>in</strong>g the microbial evolution<br />
(growth, <strong>in</strong>activation, survival) <strong>as</strong> a function of<br />
time. Estimate k<strong>in</strong>etic parameters<br />
‣Secondary models: describ<strong>in</strong>g k<strong>in</strong>etic parameters<br />
<strong>as</strong> a function of <strong>in</strong>fluenc<strong>in</strong>g factors like pH,<br />
temperature, water activity, concentration of<br />
preservatives, …<br />
‣Tertiary models: <strong>in</strong>tegrate primary and<br />
secondary models <strong>in</strong> a software <strong>tool</strong>
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Experimental design- data collection<br />
‣Estimation of k<strong>in</strong>etic parameters (primary<br />
models)<br />
‣Mathematical description of the effect of the<br />
environmental factors to the k<strong>in</strong>etic parameters<br />
(secondary models)<br />
‣Validation of the model<br />
‣Integration <strong>in</strong> a software <strong>tool</strong> (tertiary models)
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣data collection<br />
11<br />
9<br />
7<br />
Log cfu/g<br />
4<br />
2<br />
0<br />
0 5 10 15 20 25 30<br />
TIME (h)
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Fitt<strong>in</strong>g data to primary model<br />
11<br />
9<br />
Log cfu/g<br />
7<br />
4<br />
2<br />
0<br />
0 5 10 15 20 25 30<br />
TIME (h)
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Estimation of k<strong>in</strong>etic parameters<br />
9<br />
11<br />
Log N max<br />
Log cfu/g<br />
7<br />
4<br />
Log N max Maximum rate<br />
0 5 10 15 20 25 30<br />
Log N 0<br />
2<br />
0<br />
0 5 10 15 20 25 30<br />
Lag<br />
TIME (h)
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Primary models
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Primary models
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Mathematical description of the effect of the<br />
environmental factors to the k<strong>in</strong>etic parameters<br />
(secondary models)
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Secondary models<br />
‣K<strong>in</strong>etic models<br />
• Polynomial and constra<strong>in</strong>ed l<strong>in</strong>ear polynomial<br />
models<br />
• Square-root-type models<br />
• Arrhenius type models<br />
• Card<strong>in</strong>al parameter models<br />
• Artificial neural networks<br />
‣Growth/no growth <strong>in</strong>terface models<br />
(probabilistic models)
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Secondary models<br />
Card<strong>in</strong>al parameter model
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Secondary models<br />
Square root type model
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Growth no growth boundary models
<strong>Predictive</strong> Microbiology<br />
STEPS IN MODEL DEVELOPMENT<br />
‣Validation of models<br />
Most models are developed <strong>in</strong> laboratory media. There can<br />
be no guarantee that predicted values will match those<br />
that would occur <strong>in</strong> any specific <strong>food</strong> system. Before the<br />
models could be used <strong>in</strong> such a manner, the user would<br />
have to validate the models for each specific <strong>food</strong> of<br />
<strong>in</strong>terest.<br />
Internal validation: Comparison between predicted and<br />
observed values for data used for model development<br />
External validation: Comparison between predicted and<br />
observed values for <strong>in</strong>dependent data
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Software<br />
Pathogen Model<strong>in</strong>g Program (on-l<strong>in</strong>e)<br />
(http://pmp.arserrc.gov/PMPOnl<strong>in</strong>e.<strong>as</strong>px)<br />
ComB<strong>as</strong>e (datab<strong>as</strong>e) (http://www.comb<strong>as</strong>e.cc)<br />
ComB<strong>as</strong>e Predictor (models) (http://www.comb<strong>as</strong>e.cc)<br />
Sea<strong>food</strong> Spoilage Predictor<br />
(http://www.dfu.m<strong>in</strong>.dk/micro/sssp/Home/Home.<strong>as</strong>px)<br />
Refrigeration Index<br />
(ijenson@mla.com.au; http://www.mla.com.au)
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Software<br />
www.comb<strong>as</strong>e.cc<br />
‣large, searchable, datab<strong>as</strong>e of microbiological<br />
raw data<br />
‣still grow<strong>in</strong>g, users can add data<br />
‣web-b<strong>as</strong>ed, free access<br />
‣<strong>in</strong>tegrates “Food Micromodel” and “Pathogen<br />
Model<strong>in</strong>g Program” data, and many more<br />
‣<strong>in</strong>cludes new models <strong>in</strong> “ComB<strong>as</strong>e Predictor”
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Software<br />
www.comb<strong>as</strong>e.cc<br />
35,000 records on growth and survival of<br />
pathogens and spoilage organisms<br />
– ~28,000 records on pathogens<br />
– ~4,000 on spoilage organisms, <strong>in</strong>clud<strong>in</strong>g<br />
– ‘total spoilage bacteria’ (346)<br />
– ‘bacillus spoilage bacteria’ (65)<br />
– Brocothrix thermosphacta (741)<br />
– enterobacteriaceae (338)<br />
– lactic acid bacteria (701)<br />
– Shewenella putref<strong>as</strong>ciens (57)<br />
– “spoilage ye<strong>as</strong>t” (44)
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Software
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Softwares
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Softwares
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Softwares
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Softwares
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Software<br />
sea<strong>food</strong> spoilage and safety predictor-SSSP<br />
www.dfu.m<strong>in</strong>.dk/micro/sssp/<br />
predicts growth of bacteria <strong>in</strong> different fresh and<br />
lightly preserved sea<strong>food</strong>s<br />
allows prediction of:<br />
– rates of spoilage of sea<strong>food</strong><br />
– shelf life of various sea<strong>food</strong>s<br />
– effect of fluctuat<strong>in</strong>g conditions<br />
– simultaneous growth of Listeria monocytogenes (a<br />
pathogen) and spoilage bacteria <strong>in</strong> cold-smoked<br />
salmon
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Softwares<br />
sea<strong>food</strong> spoilage and safety predictor-SSSP<br />
www.dfu.m<strong>in</strong>.dk/micro/sssp/<br />
predicts growth of bacteria <strong>in</strong> different fresh and<br />
lightly preserved sea<strong>food</strong>s<br />
allows prediction of:<br />
– rates of spoilage of sea<strong>food</strong><br />
– shelf life of various sea<strong>food</strong>s<br />
– effect of fluctuat<strong>in</strong>g conditions<br />
– simultaneous growth of Listeria monocytogenes (a<br />
pathogen) and spoilage bacteria <strong>in</strong> cold-smoked<br />
salmon
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Software<br />
sea<strong>food</strong> spoilage and safety predictor-SSSP<br />
www.dfu.m<strong>in</strong>.dk/micro/sssp/
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Software<br />
“refrigeration <strong>in</strong>dex”<br />
www.mla.com.au<br />
Australian product with regulatory approval for use<br />
under Australian Export Meat Orders<br />
• predicts growth of E. coli (<strong>as</strong> an <strong>in</strong>dicator of<br />
safe temperature control) from cont<strong>in</strong>uous<br />
temperature history us<strong>in</strong>g the idea of timetemperature<br />
function<strong>in</strong>tegration
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Softwares<br />
“refrigeration <strong>in</strong>dex”<br />
www.mla.com.au<br />
Australian product with regulatory approval for use<br />
under Australian Export Meat Orders<br />
• predicts growth of E. coli (<strong>as</strong> an <strong>in</strong>dicator of<br />
safe temperature control) from cont<strong>in</strong>uous<br />
temperature history us<strong>in</strong>g the idea of timetemperature<br />
function<strong>in</strong>tegration
<strong>Predictive</strong> Microbiology<br />
Tertiary models-Softwares<br />
“refrigeration <strong>in</strong>dex”<br />
www.mla.com.au<br />
Australian product with regulatory approval for use<br />
under Australian Export Meat Orders<br />
• predicts growth of E. coli (<strong>as</strong> an <strong>in</strong>dicator of<br />
safe temperature control) from cont<strong>in</strong>uous<br />
temperature history us<strong>in</strong>g the idea of timetemperature<br />
function<strong>in</strong>tegration
<strong>Predictive</strong> Microbiology<br />
Risk Assessment Software<br />
“Risk Ranger”<br />
www.<strong>food</strong>safetycentre.com.au/<strong>risk</strong>ranger.htm<br />
‣semi-quantitative <strong>risk</strong> <strong>as</strong>sessment <strong>tool</strong><br />
‣simple, spreadsheet-b<strong>as</strong>ed<br />
‣determ<strong>in</strong>istic, but similar logic to more complex<br />
stoch<strong>as</strong>tic models<br />
‣good for beg<strong>in</strong>n<strong>in</strong>g to understand microbial <strong>food</strong><br />
safety <strong>risk</strong> <strong>as</strong>sessment, contributions/<strong>in</strong>terplay of<br />
factors,differentiation of levels of <strong>risk</strong>
<strong>Predictive</strong> Microbiology<br />
Risk Assessment Software<br />
“Risk Ranger”<br />
www.<strong>food</strong>safetycentre.com.au/<strong>risk</strong>ranger.htm<br />
‣semi-quantitative <strong>risk</strong> <strong>as</strong>sessment <strong>tool</strong><br />
‣simple, spreadsheet-b<strong>as</strong>ed<br />
‣determ<strong>in</strong>istic, but similar logic to more complex<br />
stoch<strong>as</strong>tic models<br />
‣good for beg<strong>in</strong>n<strong>in</strong>g to understand microbial <strong>food</strong><br />
safety <strong>risk</strong> <strong>as</strong>sessment, contributions/<strong>in</strong>terplay of<br />
factors,differentiation of levels of <strong>risk</strong>
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior<br />
‣For many years predictive <strong>microbiology</strong> deals with<br />
the development of determ<strong>in</strong>istic models b<strong>as</strong>ed on<br />
studies with large microbial populations<br />
‣In “real life” however, contam<strong>in</strong>ation of <strong>food</strong>s occurs<br />
at very low levels and determ<strong>in</strong>istic models are not<br />
suitable for such situations<br />
‣Recently, the need for stoch<strong>as</strong>tic models which are<br />
able to predict the effects of more “realistic”<br />
contam<strong>in</strong>ation events (low microbial numbers) <strong>in</strong> <strong>food</strong><br />
safety w<strong>as</strong> stressed (quantitative <strong>risk</strong> <strong>as</strong>sessment)
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior
WHAT IS LIFE?<br />
ERWIN SCHRODINGER<br />
First published 1944<br />
What is life? The Physical Aspect of the<br />
Liv<strong>in</strong>g Cell.
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior<br />
Distribution of NaCl growth limits for Salmonella s<strong>in</strong>gle<br />
cells at different pH values<br />
0.5<br />
0.4<br />
pH:5.0<br />
pH:5.5<br />
pH:7.3<br />
Probability<br />
0.3<br />
0.2<br />
0.1<br />
0<br />
4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0<br />
%NaCl
pH:7.3, NaCl:4.5%, time=0
pH:7.3, NaCl:4.5%, time=2.5 days
pH:7.3, NaCl:6.5%, time=5.5 days
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior<br />
Simulation of Salmonella growth at 30 o C, NaCl=6.0%,<br />
pH=5.5 (with lag)<br />
5<br />
4<br />
Grow<strong>in</strong>g Fraction<br />
Non Grow<strong>in</strong>g Fraction<br />
Population<br />
lag<br />
Lag:<br />
Wn<br />
μ<br />
N G<br />
pseudolag:<br />
NNG<br />
3<br />
Total Population<br />
Ln cfu/cm 2<br />
2<br />
1<br />
Population lag<br />
0<br />
lag<br />
pseudolag<br />
-1<br />
0 10 20 30 40 50 60 70 80<br />
Time (h)
<strong>Predictive</strong> Microbiology<br />
Stoch<strong>as</strong>tic model<strong>in</strong>g-s<strong>in</strong>gle cell behavior<br />
Results from 20 simulations of Salmonella growth at 30 o C,<br />
NaCl=6.0%, pH=5.5 (no lag)<br />
12<br />
9<br />
Ln cfu<br />
6<br />
3<br />
0<br />
0 5 10 15<br />
Time (h)
<strong>Predictive</strong> Microbiology-Risk Assessment<br />
C<strong>as</strong>e Study 1<br />
Exposure <strong>as</strong>sessment of L monocytogenes <strong>in</strong><br />
p<strong>as</strong>teurized milk<br />
Production<br />
Transportation<br />
To Retail<br />
Retail Storage<br />
Domestic Storage<br />
Consumption
<strong>Predictive</strong> Microbiology-Risk Assessment<br />
C<strong>as</strong>e Study 1<br />
Exposure <strong>as</strong>sessment of L monocytogenes <strong>in</strong><br />
p<strong>as</strong>teurized milk<br />
Growth<br />
model<br />
Production<br />
Transportation<br />
To Retail<br />
Retail Storage<br />
Domestic Storage<br />
Consumption<br />
Prevalenceconcentration<br />
Timetemperature<br />
data<br />
Handl<strong>in</strong>g<br />
<strong>in</strong>formation
C<strong>as</strong>e Study 1: Exposure <strong>as</strong>sessment of Listeria<br />
monocytogenes <strong>in</strong> p<strong>as</strong>teurized milk<br />
Development of a model for Listeria monocytogenes<br />
growth <strong>in</strong> p<strong>as</strong>teurized milk
Data collection<br />
10<br />
9<br />
8<br />
1.5 o C<br />
4.0 o C<br />
8.0 o C<br />
12 o C<br />
16 o C<br />
Log10CFU ml -1<br />
7<br />
6<br />
5<br />
4<br />
3<br />
0 500 1000 1500<br />
Time (h)
Fitt<strong>in</strong>g data to primary model (Baranyi and Roberts<br />
1994)-Estimation of k<strong>in</strong>etic parameters<br />
9<br />
8<br />
9<br />
8<br />
9<br />
8<br />
9<br />
8<br />
7<br />
7<br />
7<br />
7<br />
6<br />
6<br />
6<br />
6<br />
5<br />
5<br />
5<br />
5<br />
4<br />
4<br />
4<br />
4<br />
3<br />
3<br />
3<br />
3<br />
2<br />
2<br />
2<br />
2<br />
1<br />
1<br />
1<br />
1<br />
0<br />
0<br />
0<br />
0<br />
0 500 1000 0 500 1000 0 100 200 300 0 100 200 300<br />
9<br />
9<br />
9<br />
9<br />
8<br />
8<br />
8<br />
8<br />
7<br />
7<br />
7<br />
7<br />
6<br />
6<br />
6<br />
6<br />
5<br />
5<br />
5<br />
5<br />
4<br />
4<br />
4<br />
4<br />
3<br />
3<br />
3<br />
3<br />
2<br />
2<br />
2<br />
2<br />
1<br />
1<br />
1<br />
1<br />
0<br />
0 100 200 300 400<br />
0<br />
0 100 200 300 400<br />
0<br />
0 50 100 150 200<br />
0<br />
0 50 100 150 200<br />
9<br />
10<br />
10<br />
10<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
9<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
9<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
9<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
0<br />
0 50 100 150 200<br />
0<br />
0 50 100 150 200<br />
0<br />
0 50 100 150<br />
0<br />
0 50 100 150
Fitt<strong>in</strong>g data to primary model (Baranyi and Roberts<br />
1994)-Estimation of k<strong>in</strong>etic parameters<br />
curve rate lag y0 yEnd se(fit) R^2_stat<br />
1A 0.0046 502.4 3.70 0.225 0.973<br />
1B 0.0035 460.9 3.80 0.133 0.985<br />
4A 0.0121 199.3 3.79 7.74 0.172 0.991<br />
4B 0.0110 164.0 3.60 7.77 0.164 0.992<br />
4C 0.0111 181.2 4.03 8.07 0.202 0.987<br />
4D 0.0083 164.9 4.11 7.68 0.146 0.991<br />
8A 0.0267 16.4 3.50 8.26 0.146 0.992<br />
8B 0.0252 25.2 3.66 8.29 0.106 0.995<br />
8C 0.0250 17.1 3.59 8.46 0.161 0.992<br />
8D 0.0257 19.0 3.80 8.37 0.129 0.995<br />
12A 0.0476 13.3 3.68 8.35 0.126 0.994<br />
12B 0.0476 14.3 3.74 8.25 0.113 0.995<br />
12C 0.0502 9.3 3.60 8.31 0.214 0.984<br />
12D 0.0478 12.8 3.89 8.83 0.160 0.992<br />
16A 0.0773 4.0 3.57 8.51 0.176 0.991<br />
16B 0.0804 5.9 3.71 8.56 0.161 0.992<br />
16C 0.0829 8.8 3.44 8.23 0.173 0.992<br />
16D 0.0832 6.0 3.30 8.15 0.178 0.992<br />
16E 0.0967 10.2 3.60 8.45 0.063 0.999<br />
16F 0.0973 7.8 3.63 8.62 0.100 0.998
Fitt<strong>in</strong>g k<strong>in</strong>etic parameters to a secondary model<br />
0.5<br />
0.4<br />
μ<br />
max<br />
= b −<br />
( T Tm<strong>in</strong><br />
)<br />
(μmax) 0.5<br />
0.3<br />
0.2<br />
0.1<br />
0<br />
0 2 4 6 8 10 12 14 16 18<br />
Temperature ( o C)<br />
Estimated value Lower 95% CI Upper 95% CI r 2<br />
Parameter<br />
b 0.024 0.023 0.025 0.988<br />
T m<strong>in</strong> ( o C) -2.32 -3.02 -1.61
Validation at dynamic temperature conditions<br />
prediction b<strong>as</strong>ed on the square root model for the<br />
estimation of the “momentary” rate and the differential<br />
equations of Baranyi and Roberts model (eq. 2 and 3),<br />
which were numerically <strong>in</strong>tegrated with respect to time:<br />
d<br />
dt<br />
x<br />
=<br />
⎛<br />
q<br />
⎞<br />
q<br />
⎜<br />
⎝ + 1<br />
⎠<br />
⎝<br />
x<br />
x<br />
[ b( T<br />
t<br />
T )] ⎟⎜<br />
⎛ 2<br />
( ) − ⎜<br />
−<br />
⎟<br />
x<br />
m<strong>in</strong> 1<br />
max<br />
⎞<br />
⎟<br />
⎠<br />
m<br />
d q<br />
=<br />
2<br />
m<strong>in</strong><br />
dt<br />
[ b<br />
( T<br />
( t<br />
)<br />
−<br />
T<br />
)<br />
] q
Validation at dynamic temperature conditions<br />
10<br />
20<br />
9<br />
18<br />
Log10 cfu/ml<br />
8<br />
7<br />
6<br />
5<br />
4<br />
16<br />
14<br />
12<br />
10<br />
8<br />
6<br />
4<br />
Temperature ( o C)<br />
3<br />
2<br />
2<br />
0<br />
0 30 60 90 120 150 180<br />
Time (hours)
Validation at dynamic temperature conditions<br />
10<br />
20<br />
9<br />
18<br />
8<br />
16<br />
Log10 cfu/ml<br />
7<br />
6<br />
5<br />
4<br />
3<br />
14<br />
12<br />
10<br />
8<br />
6<br />
Temperature ( o C)<br />
2<br />
4<br />
1<br />
2<br />
0<br />
0<br />
0 30 60 90 120 150 180 210 240<br />
Time (hours)
Validation at dynamic temperature conditions<br />
10<br />
20<br />
Log10 cfu/g<br />
9<br />
8<br />
7<br />
6<br />
5<br />
4<br />
3<br />
2<br />
1<br />
0<br />
4<br />
2<br />
0<br />
0 30 60 90 120 150<br />
Time (hours)<br />
18<br />
16<br />
14<br />
12<br />
10<br />
8<br />
6<br />
Temperature ( o C)
C<strong>as</strong>e Study 1: Exposure <strong>as</strong>sessment of Listeria<br />
monocytogenes <strong>in</strong> p<strong>as</strong>teurized milk<br />
Collection of data on p<strong>as</strong>teurized milk chill cha<strong>in</strong><br />
Wholesaler<br />
Packer/Processor<br />
Pro<br />
duc<br />
er<br />
Consumer<br />
Retailer<br />
Hospital
C<strong>as</strong>e Study 1: Exposure <strong>as</strong>sessment of Listeria<br />
monocytogenes <strong>in</strong> p<strong>as</strong>teurized milk<br />
Collection of data on p<strong>as</strong>teurized milk chill cha<strong>in</strong><br />
Collection of Time-Temperature data from<br />
Transportation trucks<br />
retail refrigerators<br />
domestic refrigerators<br />
Pro<br />
duc<br />
er
C<strong>as</strong>e Study 1: Exposure <strong>as</strong>sessment of Listeria<br />
monocytogenes <strong>in</strong> p<strong>as</strong>teurized milk<br />
Data from trucks dur<strong>in</strong>g transportation to<br />
retail<br />
Transportation<br />
time<br />
Transportation<br />
temperature<br />
20<br />
18<br />
30<br />
%Products<br />
16<br />
14<br />
12<br />
10<br />
8<br />
6<br />
%Trucks<br />
25<br />
20<br />
15<br />
10<br />
4<br />
2<br />
5<br />
0<br />
0 1 2 3 4 5 6 7 8 9 10 11<br />
0<br />
3 4 5 6 7 8 9 10 11 12<br />
Transportation time (hours)<br />
Temperature <strong>in</strong> truck ( o C)
C<strong>as</strong>e Study 1: Exposure <strong>as</strong>sessment of Listeria<br />
monocytogenes <strong>in</strong> p<strong>as</strong>teurized milk<br />
Representative temperature profiles of retail<br />
cab<strong>in</strong>ets<br />
Θερμοκρασία o C<br />
15,0<br />
13,0<br />
11,0<br />
9,0<br />
7,0<br />
5,0<br />
3,0<br />
1,0<br />
-1,0<br />
-3,0<br />
-5,0<br />
15,0<br />
13,0<br />
0 50 100 150 200 250 300 350<br />
Χρόνος (ώρες)<br />
Θερμοκρασία o C<br />
15,0<br />
13,0<br />
11,0<br />
9,0<br />
7,0<br />
5,0<br />
3,0<br />
1,0<br />
-1,0<br />
-3,0<br />
-5,0<br />
15,0<br />
13,0<br />
0 50 100 150 200 250 300 350<br />
Χρόνος (ώρες)<br />
11,0<br />
11,0<br />
Θερμοκρασία o C<br />
9,0<br />
7,0<br />
5,0<br />
3,0<br />
1,0<br />
Θερμοκρασία o C<br />
9,0<br />
7,0<br />
5,0<br />
3,0<br />
1,0<br />
-1,0<br />
-3,0<br />
0 50 100 150 200 250<br />
-1,0<br />
-3,0<br />
0 50 100 150 200 250 300 350<br />
-5,0<br />
-5,0<br />
Χρόνος (ώρες)<br />
Χρόνος (ώρες)
C<strong>as</strong>e Study 1: Exposure <strong>as</strong>sessment of Listeria<br />
monocytogenes <strong>in</strong> p<strong>as</strong>teurized milk<br />
Temperature distribution of retail cab<strong>in</strong>ets<br />
35<br />
30<br />
%Refrigerators<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
-2 0 2 4 6 8 10 12 14<br />
Temperature o C
C<strong>as</strong>e Study 1: Exposure <strong>as</strong>sessment of Listeria<br />
monocytogenes <strong>in</strong> p<strong>as</strong>teurized milk<br />
Storage time at retail cab<strong>in</strong>ets<br />
%Products<br />
45<br />
40<br />
35<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
0 10 20 30 40 50 60 70 80 90 100<br />
Storage time at retail (hours)
Temperature o C<br />
Temperature o C<br />
15<br />
13<br />
11<br />
9<br />
7<br />
5<br />
3<br />
1<br />
-1<br />
-3<br />
-5<br />
Representative temperature profiles of<br />
domestic refrigerators<br />
Upper shelf<br />
Middle shelf<br />
Lower shelf<br />
Door shelf<br />
0 4 8 12 16 20 24<br />
15,0<br />
13,0<br />
11,0<br />
9,0<br />
7,0<br />
5,0<br />
3,0<br />
1,0<br />
Upper shelf<br />
Middle shelf<br />
Lower shelf<br />
Door shelf<br />
Time (hours)<br />
Temperature o C<br />
Temperature o C<br />
15,0<br />
13,0<br />
11,0<br />
9,0<br />
7,0<br />
5,0<br />
3,0<br />
1,0<br />
-1,0<br />
-3,0<br />
-5,0<br />
15,0<br />
13,0<br />
11,0<br />
9,0<br />
7,0<br />
5,0<br />
3,0<br />
1,0<br />
Upper shelf<br />
Middle shelf<br />
Lower shelf<br />
Door shelf<br />
0 4 8 12 16 20 24<br />
Upper shelf<br />
Middle shelf<br />
Lower shelf<br />
Door shelf<br />
Time (hours)<br />
-1,0<br />
-3,0<br />
0 4 8 12 16 20 24<br />
-1,0<br />
-3,0<br />
0 4 8 12 16 20 24<br />
-5,0<br />
-5,0<br />
Time (hours)<br />
Time (hours)
Temperature distribution of domestic<br />
refrigerators<br />
%Refrigerators<br />
35<br />
30<br />
25<br />
20<br />
15<br />
10<br />
Upper shelf<br />
Middle shelf<br />
Lower shelf<br />
Door shelf<br />
5<br />
0<br />
0 2 4 6 8 10 12 14 16 18<br />
Average Temperature o C
Consumer Handl<strong>in</strong>g <strong>in</strong>formation
Consumer Handl<strong>in</strong>g <strong>in</strong>formation<br />
60<br />
Storage time <strong>in</strong> domestic<br />
refrigerator<br />
50<br />
% Consumers<br />
40<br />
30<br />
20<br />
10<br />
0<br />
1 2 3 4 5<br />
Storage time of p<strong>as</strong>teurized milk <strong>in</strong> domestic refrigerator (days)
Consumer Handl<strong>in</strong>g <strong>in</strong>formation<br />
60<br />
50<br />
Position of milk <strong>in</strong> domestic<br />
refrigerator<br />
% Consumers<br />
40<br />
30<br />
20<br />
10<br />
0<br />
Upper shelf Middle shelf Lower shelf Door shelf<br />
Position of p<strong>as</strong>teurized milk <strong>in</strong> domestic refrigerator
0<br />
0<br />
2<br />
1<br />
4<br />
2<br />
3<br />
6<br />
8<br />
0<br />
2<br />
4<br />
0<br />
6<br />
8<br />
1<br />
2<br />
3<br />
0<br />
1<br />
2<br />
3<br />
0,40<br />
0,35<br />
0,30<br />
0,25<br />
Expon(2,5) Shift=-2,5<br />
Exposure Assessment<br />
0,40<br />
Triang(-2,5; 0; 2,5)<br />
0,20<br />
0,35<br />
0,15<br />
0,30<br />
0,10<br />
0,25<br />
0,05<br />
0,20<br />
0,00<br />
-4<br />
-2<br />
90,0%<br />
-2,37 4,99<br />
5,0%<br />
Contam<strong>in</strong>ation<br />
at t=0<br />
10<br />
><br />
0,15<br />
0,10<br />
0,05<br />
0,00<br />
-3<br />
-2<br />
-1<br />
5,0% 90,0%<br />
5,0%<br />
-1,709 1,709<br />
Consumption<br />
data<br />
MODELS (growth<strong>in</strong>activation-survival<br />
Contam<strong>in</strong>ation at<br />
consumption time<br />
Normal(0; 1)<br />
0,40<br />
Triang(-2,5; 0; 2,5)<br />
Monte<br />
Carlo<br />
0,45<br />
0,40<br />
0,35<br />
0,35<br />
0,30<br />
0,30<br />
0,25<br />
0,20<br />
0,15<br />
0,25<br />
0,20<br />
0,15<br />
Exposure<br />
0,10<br />
0,10<br />
0,05<br />
0,00<br />
-3<br />
-2<br />
-1<br />
0,05<br />
0,00<br />
Lognorm(2,5; 2,5) Shift=-2,5<br />
< 5,0% 90,0%<br />
5,0% ><br />
-1,645 1,645<br />
-3<br />
-2<br />
-1<br />
5,0% 90,0%<br />
5,0%<br />
-1,709 1,709<br />
0,40<br />
0,35<br />
Process<strong>in</strong>g,<br />
distribution,<br />
storage<br />
parameters<br />
0,40<br />
0,35<br />
0,30<br />
0,25<br />
0,20<br />
0,15<br />
0,10<br />
0,05<br />
0,00<br />
-4<br />
-2<br />
Lognorm(2,5; 2,5) Shift=-2,5<br />
90,0%<br />
-2,05 4,45<br />
5,0%<br />
10<br />
12<br />
><br />
0,30<br />
0,25<br />
0,20<br />
0,15<br />
0,10<br />
0,05<br />
0,00<br />
-4<br />
-2<br />
0<br />
2<br />
4<br />
90,0%<br />
-2,05 4,45<br />
6<br />
8<br />
5,0%<br />
10<br />
12<br />
>
Concentration of L. monocytogenes <strong>in</strong> contam<strong>in</strong>ated<br />
p<strong>as</strong>teurized milk at the t<strong>in</strong>e of consumption<br />
60<br />
56.57<br />
50<br />
40<br />
%Products<br />
30<br />
20<br />
18.50<br />
10<br />
11.55<br />
6.18<br />
0<br />
3.66 1.77 0.87 0.44 0.27 0.09 0.04 0.03 0.02 0.01<br />
-3 -2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5 3 3.5<br />
L. monocytogenes concentration at consumption time (Log CFU/ml)
Sensitivity Analysis<br />
Domestic Storage Time<br />
0.515<br />
Domestic Temperature-Door Shelf<br />
0.307<br />
Retail Storage Time<br />
Retail Storage Temperature<br />
Domestic Temperature-Upper Shelf<br />
Position <strong>in</strong> Domestic Refrigerator (Shelf)<br />
Domestic Temperature-Middle Shelf<br />
Transportation Temperature<br />
Domestic Temperature-Lower Shelf<br />
Transportation Time<br />
0.209<br />
0.188<br />
0.149<br />
0.114<br />
0.097<br />
0.046<br />
0.038<br />
0.032<br />
0 0.1 0.2 0.3 0.4 0.5 0.6<br />
Corelation Coefficient
C<strong>as</strong>e Study 1: Exposure <strong>as</strong>sessment of Listeria<br />
monocytogenes <strong>in</strong> p<strong>as</strong>teurized milk<br />
Evaluation of “what if” scenarios<br />
Int. I: better control of retail temperature,<br />
Int. II: exclude door shelf from domestic storage,<br />
Int. III: Int. II+ decre<strong>as</strong>e domestic mean<br />
temperature by 2 oC
Evaluation of “what if” scenarios<br />
20<br />
18<br />
Int. III<br />
81.9<br />
16<br />
Int. II<br />
67.6<br />
14<br />
Int. I<br />
61.1<br />
%Products<br />
12<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
Current situation<br />
Current situation<br />
Int. I<br />
Int. II<br />
Int. III<br />
0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0<br />
Total growth of L. monocytogenes (Log CFU/ml) dur<strong>in</strong>g distribution,<br />
retail and domestic storage<br />
56.6<br />
0 20 40 60 80 100<br />
Percent of products <strong>in</strong> which L. monocytogenes will not grow
Evaluation of “what if” scenarios<br />
Incre<strong>as</strong>e shelf life of p<strong>as</strong>teurized milk from 5 to<br />
10 days<br />
40<br />
35<br />
30<br />
%Products<br />
25<br />
20<br />
15<br />
10<br />
Current situation<br />
Int. III<br />
5<br />
0<br />
-3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0<br />
L. monocytogenes concentration at consumption time (Log CFU/ml)
<strong>Predictive</strong> Microbiology-Risk Assessment<br />
C<strong>as</strong>e Study 2<br />
Risk-b<strong>as</strong>ed approach for evaluat<strong>in</strong>g compliance of<br />
RTE <strong>food</strong> to the safety criteria for L.<br />
monocytogenes
The new EC regulation 2073/2005 on<br />
microbiological criteria for <strong>food</strong>stuffs<br />
(safety criteria for L. monocytogenes)<br />
Before January 1 st 2006 the Microbiological criterion<br />
for L. monocytogenes w<strong>as</strong><br />
Zero Tolerance<br />
(EC Directive 92/46)<br />
The problem………<br />
Zero Tolerance w<strong>as</strong> not fe<strong>as</strong>ible <strong>in</strong> practice
Source:Draft Assessment of the Relative Risk to Public Health from Foodborne<br />
Listeria monocytogenes Among Selected Categories of Ready-to-Eat<br />
Foodshttp://www.<strong>food</strong>safety.gov/~dms/lm<strong>risk</strong>.html
The new EC regulation 2073/2005 on<br />
microbiological criteria for <strong>food</strong>stuffs<br />
(safety criteria for L. monocytogenes)
Lm <strong>in</strong> sliced RTE meat products of the Hellenic Market<br />
Mean Prevalence 8,1% (n=209)<br />
Product Product Product<br />
Tested Positive<br />
type name category a Manufactur<strong>in</strong>g company<br />
samples samples<br />
1 Bresaola F F 2 0<br />
2<br />
Turkey<br />
bre<strong>as</strong>t<br />
HT H,I,K,S,U,X,ZA 21 0<br />
3<br />
Smoked<br />
tongue<br />
HT T 1 0<br />
4<br />
Ham<br />
(cooked)<br />
HT E,H,J,K,P,S,X,ZA 27 1<br />
5<br />
Ham<br />
(fermented)<br />
F E,I,N 6 1<br />
6 Copa F E,F 3 0<br />
7<br />
Chicken<br />
bre<strong>as</strong>t<br />
HT Q 4 0<br />
8 Mortadella HT C,D,E,H,I,S,X,ZA 13 0<br />
9 Bacon HT B,C,D,E,H,I,J,M,O,S,U,W,X,Y,Z,ZA 49 12<br />
10 Pork lo<strong>in</strong> HT C,D,H,J,S,ZA 10 0<br />
11 Pariza HT A,C,D,H,X,ZA 14 0<br />
12 P<strong>as</strong>tirma F T 2 0<br />
13 Prosiuto F F,M,R 6 0<br />
14 Salami F D,E,F,G,I,M,P,R,S,U,V,X,ZA 30 3<br />
15<br />
Salami<br />
(cooked)<br />
HT K,L,Q 6 0<br />
16<br />
Pork<br />
shoulder<br />
HT D,H,M,P,S,U,X,ZA 15 0<br />
Total 209 17
The new EC regulation 2073/2005 on<br />
microbiological criteria for <strong>food</strong>stuffs<br />
(safety criteria for L. monocytogenes)
The new EC regulation 2073/2005 on<br />
microbiological criteria for <strong>food</strong>stuffs<br />
Safety Criteria for L. monocytogenes<br />
Food category<br />
Sampl<strong>in</strong>gplan<br />
Limits<br />
n c m M<br />
Stage where the criterion<br />
applies<br />
1.2 Ready-to-eat<br />
<strong>food</strong>s able to support<br />
the growth of L.<br />
monocytogenes<br />
5 0 100 cfu/g 4 placed on the market and<br />
Products ready to be<br />
dur<strong>in</strong>g their shelf-life<br />
5 0 Absence<br />
<strong>in</strong> 25 g 5<br />
Before the <strong>food</strong> h<strong>as</strong> left<br />
the immediate control of<br />
the <strong>food</strong> bus<strong>in</strong>ess operator,<br />
who h<strong>as</strong> produced it<br />
1.3 Ready-to-eat<br />
<strong>food</strong>s unable to<br />
support the growth<br />
of L. monocytogenes<br />
5 0 100 cfu/g<br />
Products ready to be<br />
placed on the market and<br />
dur<strong>in</strong>g their shelf-life
The new EC regulation 2073/2005 on<br />
microbiological criteria for <strong>food</strong>stuffs<br />
(emph<strong>as</strong>is on L. monocytogenes)<br />
the traditional approach b<strong>as</strong>ed on end-product<br />
sampl<strong>in</strong>g and exam<strong>in</strong>ation is no longer sufficient to<br />
<strong>as</strong>sess the compliance of RTE <strong>food</strong>s with the new<br />
criteria<br />
Use of alternative “<strong>tool</strong>s”<br />
<strong>Predictive</strong> Microbiology
Requirements (technical)<br />
for the Food Industry for compliance with the new<br />
safety criteria for LM <strong>in</strong> RTE <strong>food</strong>s<br />
1. Prove that the products support or do not support<br />
growth of L monocytogenes<br />
Probabilistic (Growth/no Growth) models<br />
If the product supports growth<br />
2. Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
K<strong>in</strong>etic models
Requirements (technical)<br />
1.Prove that the products support or do not support<br />
growth of L monocytogenes<br />
Note <strong>in</strong> the regulation<br />
Products are automatically considered to belong to the<br />
category of not support<strong>in</strong>g growth of Lm if<br />
pH
Requirements (technical)<br />
1.Prove that the products support or do not support<br />
growth of L monocytogenes<br />
7,00<br />
pH-a w comb<strong>in</strong>ations of RTE meat<br />
products <strong>in</strong> the Hellenic market<br />
6,50<br />
6,00<br />
pH<br />
5,50<br />
5,00<br />
4,50<br />
4,00<br />
0,890 0,910 0,930 0,950 0,970 0,990<br />
Only 6.1% of the products are automatically considered to belong<br />
to the category of not support<strong>in</strong>g growth of Lm<br />
a w
Requirements (technical)<br />
1.Prove that the products support or do not support<br />
growth of L monocytogenes<br />
Probabilistic (Growth/no Growth) models<br />
7,00<br />
4 o C<br />
6,50<br />
Support Growth<br />
6,00<br />
pH<br />
5,50<br />
5,00<br />
4,50<br />
Do not Support Growth<br />
Growth boundary (P=0.5)<br />
4,00<br />
0,920 0,930 0,940 0,950 0,960 0,970 0,980 0,990<br />
a w<br />
Model:Koutsoumanis et al., 2004, Int. J. Food Micro.
Requirements (technical)<br />
1.Prove that the products support or do not support<br />
growth of L monocytogenes<br />
Probabilistic (Growth/no Growth) models<br />
7,00<br />
15 o C<br />
10 o C<br />
4 o C<br />
6,50<br />
6,00<br />
pH<br />
5,50<br />
5,00<br />
4,50<br />
Growth boundary (P=0.5)<br />
4,00<br />
0,920 0,930 0,940 0,950 0,960 0,970 0,980 0,990<br />
Model:Koutsoumanis et al., 2004<br />
a w
Requirements (technical)<br />
1.Prove that the products support or do not support<br />
growth of L monocytogenes<br />
At which temperature ?
Requirements (technical)<br />
2.Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
Before the <strong>food</strong> h<strong>as</strong> left<br />
the immediate control of the<br />
<strong>food</strong> bus<strong>in</strong>ess operator, who<br />
h<strong>as</strong> produced it<br />
Products ready to be placed<br />
on the market and dur<strong>in</strong>g<br />
their shelf-life<br />
Absence <strong>in</strong> 25 g or<br />
2.Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
Use of available predictive model<strong>in</strong>g software (i.e PMP)
2.Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
Use of available predictive model<strong>in</strong>g software (i.e PMP)<br />
11.0<br />
Listeria monocytogenes<br />
11.0<br />
Listeria monocytogenes<br />
10.0<br />
9.0<br />
4 o C<br />
10.0<br />
9.0<br />
10 o C<br />
8.0<br />
8.0<br />
7.0<br />
7.0<br />
log(CFU/ml)<br />
6.0<br />
5.0<br />
4.0<br />
3.0<br />
2.0<br />
1.0<br />
25 days for 3 logs<br />
<strong>in</strong>cre<strong>as</strong>e<br />
log(CFU/ml)<br />
6.0<br />
5.0<br />
4.0<br />
3.0<br />
2.0<br />
1.0<br />
8 days for 3 logs<br />
<strong>in</strong>cre<strong>as</strong>e<br />
0.0<br />
0 5 10 15 20 25 30 35 40 45 50<br />
Time (Days)<br />
0.0<br />
0 2.5 5 7.5 10 12.5 15 17.5 20 22.5 25<br />
Time (Days)
2.Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
At which temperature ?<br />
General Requirements <strong>in</strong> Reg 2073/2005<br />
the <strong>food</strong> safety criteria applicable<br />
throughout the shelf-life of the products<br />
can be met under re<strong>as</strong>onably,foreseeable<br />
conditions of distribution, storage and use.
Chill Cha<strong>in</strong> Conditions<br />
Temperature distribution <strong>in</strong> retail refrigerators<br />
(survey <strong>in</strong> Greece)<br />
35<br />
30<br />
% Refrigerators<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
0 2 4 6 8 10 12 14<br />
Temperature o C
<strong>Predictive</strong> Microbiology <strong>tool</strong>s at Production level<br />
A probabilistic modell<strong>in</strong>g approach for evaluat<strong>in</strong>g<br />
the compliance of RTE <strong>food</strong>s with the new safety<br />
criteria for Listeria monocytogenes<br />
The approach is b<strong>as</strong>ed on the comb<strong>in</strong>ed use of: a)<br />
growth/no growth boundary models, b) k<strong>in</strong>etic growth<br />
models, c) data on product characteristics (pH, aw, shelflife)<br />
and d) storage temperature data<br />
A probabilistic analysis of the above components us<strong>in</strong>g<br />
Monte Carlo simulation, can lead to a more realistic<br />
estimation of the behavior of L. monocytogenes which can<br />
be further used <strong>as</strong> a decision-mak<strong>in</strong>g <strong>tool</strong> regard<strong>in</strong>g the<br />
design or modification of a product’s formulation or its<br />
“use-by-date” <strong>in</strong> order to ensure its compliance
Requirements (technical)<br />
1.Prove that the products support or do not support<br />
growth of L monocytogenes<br />
Probabilistic approach<br />
G/NG<br />
model<br />
Product<br />
Characteristics<br />
(pH, a w …)<br />
Monte Carlo<br />
simulation<br />
Probability of growth<br />
Chill cha<strong>in</strong><br />
characteristics<br />
1<br />
0,9<br />
Cumulative probability<br />
0,8<br />
0,7<br />
0,6<br />
0,5<br />
0,4<br />
0,3<br />
0,2<br />
0,1<br />
0<br />
0 0,2 0,4 0,6 0,8 1 1,2<br />
Probability of growth for L. monocytogenes<br />
% Refrigerators<br />
35<br />
30<br />
25<br />
20<br />
15<br />
10<br />
5<br />
0<br />
0 2 4 6 8 10 12 14<br />
Temperature o C
Requirements (technical)<br />
1.Prove that the product supports or do not support<br />
growth of L monocytogenes<br />
C<strong>as</strong>e study<br />
Product: cooked ham (pH=5.49, a w =0.943)<br />
1<br />
1<br />
Cumulative probability<br />
0,9<br />
0,8<br />
0,7<br />
0,6<br />
0,5<br />
0,4<br />
0,3<br />
0,2<br />
B<strong>in</strong>omial (1, Pg)<br />
probability<br />
0,9<br />
0,8<br />
0,7<br />
0,6<br />
0,5<br />
0,4<br />
0,3<br />
0,2<br />
No Growth<br />
0,6143<br />
Growth<br />
0,3857<br />
0,1<br />
0,1<br />
0<br />
0 0,2 0,4 0,6 0,8 1 1,2<br />
Probability of growth for L. monocytogenes<br />
0<br />
0 1<br />
Growth/No growth for L. monocytogenes
X
Model used<br />
Comb<strong>in</strong>ation of G/NG and k<strong>in</strong>etic models<br />
0,1<br />
0,08<br />
G/NG boundary model<br />
Koutsoumanis et al., 2004<br />
K<strong>in</strong>etic model<br />
Buchanan., 2001<br />
0,06<br />
mmax<br />
0,04<br />
0,02<br />
Probability of growth<br />
1,2<br />
1<br />
0,8<br />
0,6<br />
0,4<br />
0,2<br />
0<br />
4 6 8<br />
mmax=<br />
Normal (m, sePred)<br />
0<br />
0 2 4 6 8 10 12 14 16 18 20<br />
μ<br />
max<br />
Pg =<br />
=<br />
Temperature ( o C)<br />
fT ( ),Ross et al., 2003<br />
Buchanan, 2001<br />
g( T), Presser et al., 1998<br />
Primary model: Baranyi, μ<br />
a: B<strong>in</strong>omial (1, Pg)<br />
Koutsoumanis et al., 2004<br />
μ: Normal ( ˆ μ , sepred)<br />
max<br />
max<br />
= a×<br />
μ<br />
max
Requirements (technical)<br />
2.Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
C<strong>as</strong>e study<br />
Product: cooked ham (pH=5.49, a w =0.943, N=50ppm, SL=60 days)<br />
Distribution of L. monocytogenes<br />
at the end of shelf life <strong>in</strong> retail<br />
0,25<br />
0,2<br />
In 28% of products Lm will<br />
exceed 100 cfu/g<br />
Probability<br />
0,15<br />
0,1<br />
0,05<br />
0<br />
-2 -1 0 1 2 3 4 5 6 7 8 9 10<br />
L. monocytogenes at the end of retail storage (Log cfu/g)
Requirements (technical)<br />
2.Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
Potential actions for achiev<strong>in</strong>g the desired level of<br />
compliance to the Safety criteria<br />
‣ Adjust the shelf life of the product<br />
‣ Modify the formulation of the product
Requirements (technical)<br />
2.Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
Potential actions for meet<strong>in</strong>g the Safety criteria<br />
Adjust the shelf life of the product<br />
Probability<br />
0,6<br />
0,5<br />
0,4<br />
0,3<br />
0,2<br />
Shelf Life 60 Days Compliance <strong>in</strong> 72th percentile<br />
Shelf Life 43 days:Compliance <strong>in</strong> 90th percentile<br />
Shelf Life 35 Days:Compliance <strong>in</strong> 95th percentile<br />
Shelf Life 25 Days:Compliance <strong>in</strong> 99th percentile<br />
0,1<br />
0<br />
-2 -1 0 1 2 3 4 5 6 7 8 9 10<br />
L. monocytogenes at the end of retail storage (Log cfu/g)
Requirements (technical)<br />
2.Prove that the concentration of the pathogen will<br />
not exceed 100 cfu/g at the end of shelf life <strong>in</strong> retail<br />
Cooked ham<br />
Potential actions for meet<strong>in</strong>g the Safety criteria<br />
Modify the formulation of the product<br />
a w 0.943 0.935<br />
N 50ppm 80ppm<br />
0,45<br />
0,4<br />
0,35<br />
0,3<br />
Compliance <strong>in</strong> 72th percentile<br />
Compliance <strong>in</strong> 95th percentile<br />
a w<br />
-0.848<br />
Tornado graph<br />
Regression Sensitivity for Self-life<br />
Probability<br />
0,25<br />
0,2<br />
0,15<br />
pH<br />
-0.458<br />
0,1<br />
Nitrites<br />
0.269<br />
0,05<br />
-1 -0.5 0 0.5 1<br />
Std b Coefficients<br />
0<br />
-2 -1 0 1 2 3 4 5 6 7 8 9 10<br />
L. monocytogenes at the end of retail storage (Log cfu/g)
Us<strong>in</strong>g the probabilistic approach for both safety<br />
and spoilage<br />
X
Us<strong>in</strong>g the probabilistic approach for both safety<br />
and spoilage<br />
C<strong>as</strong>e study<br />
Product: cooked ham<br />
pH=5.49, a w =0.943, N=50ppm, SL=60 days<br />
SSO: L. sake<br />
Spoilage level 10 7 cfu/g
Us<strong>in</strong>g the probabilistic approach for both safety<br />
and spoilage<br />
C<strong>as</strong>e study<br />
Product: cooked ham SL 60 days<br />
poilage<br />
10<br />
compliant/spoiled<br />
non-compliant/spoiled<br />
Level<br />
LAB at end of SL (log cfu/g)<br />
9<br />
8<br />
7<br />
6<br />
5<br />
4<br />
compliant/unspoiled<br />
3<br />
non-compliant/unspoiled<br />
2<br />
-2 -1 0 1 2 3 4 5 6 7<br />
L. monocytogenes at the end of SL (log cfu/g)<br />
Safety criterion
Us<strong>in</strong>g the probabilistic approach for both safety<br />
and spoilage<br />
C<strong>as</strong>e study<br />
SL 60 days<br />
Product: cooked ham<br />
SL 25 days<br />
10<br />
compliant/spoiled<br />
non-compliant/spoiled<br />
10<br />
compliant/spoiled<br />
non-compliant/spoiled<br />
9<br />
9<br />
LAB at end of SL (log cfu/g)<br />
8<br />
7<br />
6<br />
5<br />
4<br />
compliant/unspoiled<br />
LAB at end of SL (log cfu/g)<br />
8<br />
7<br />
6<br />
5<br />
4<br />
compliant/unspoiled<br />
3<br />
2<br />
-2 -1 0 1 2 3 4 5 6 7<br />
L. monocytogenes at the end of SL (log cfu/g)<br />
72% compliance<br />
non-compliant/unspoiled<br />
17% spoiled before end<br />
of Shelf life<br />
3<br />
2<br />
-2 -1 0 1 2 3 4 5 6 7<br />
L. monocytogenes at the end of SL (log cfu/g)<br />
99% compliance<br />
non-compliant/unspoiled<br />
0.1% spoiled before end<br />
of Shelf life
Acknowledgement<br />
EU Framework VI programme on Food Quality and Safety,<br />
‘<strong>ProSafeBeef</strong>’ Food-CT-2006-36241<br />
EU Framework V programme on “Quality of Life and<br />
Management of Liv<strong>in</strong>g Resources”, Key Action 1-Health<br />
Food and Environment,”SMAS” QLK1-CT2002-02545<br />
“Development and Application of Microbial Time<br />
Temperature Indicators for Monitor<strong>in</strong>g Food Quality &<br />
Safety” Greek Government & Commission of the European<br />
Community, PENED 2003
<strong>Predictive</strong> <strong>food</strong> <strong>microbiology</strong> <strong>as</strong> a <strong>tool</strong> <strong>in</strong><br />
<strong>risk</strong> <strong>as</strong>sessment<br />
Kost<strong>as</strong> Koutsoumanis<br />
Lab of Food Microbiology and Hygiene<br />
Dpt. Of Food Science and Technology<br />
Aristotle University of Thessaloniki,<br />
Contact: kkoutsou@agro.auth.gr