Introduction to the Use of Microbial Pathogen Computer Modeling Programs

1
Introduction to the Use of Microbial Pathogen
Computer Modeling Programs

• Robert J. Hasiak, Ph.D.
• Director, Regulatory Affairs
• IEH Center for Food Pharmaceutical
• Process Safety

2
Knowledge

• A little knowledge is good.
• A lot of knowledge with experience is great.
• But, anything in between can be just dangerous!

3
Objectives

• Discuss significant factors and background
  material associated with the various pathogen
  modeling programs
• Compare features of various models
• Apply the Models to field cases
• Answer Questions

4
What are MPCM Programs?

• Estimate microbial growth, lethality, or survival
  in food products based on certain factors or
  assumptions.
• Predict the growth or decline of microbes under
  specific experimental conditions.
• Quantify the effects of two or more factors on
  microbes.

5
Intrinsic and Extrinsic Food Factors Commonly
Addressed in MPCM Programs

• Temperature
• Humidity
• pH
• Water activity
• Atmosphere
• Aerobic, vacuum packaging, MAP
• Fat level
• Additives
• NaCl, nitrites, phosphates, sugars etc.

6
Predicting Behavior

• All bacterial growth and behavior can be
  represented mathematically once the critical
  factors are identified.
• The capability of pathogen modeling is often
  limited by factors not identified or quantified.
  
• Modeling is often limited to certain stages of
  growth or death.

7
Reality of Microbial Models

• Microbial models are mans feeble attempt to
  express the real world of microbial behavior
  through the abstract world of mathematical
  concepts

8
Role of MPCM Programs

• Tools to support Hazard Analyses, estimate CCP
  limits, and evaluate effects of process
  deviations.
• Estimate microbial behavior and provide graphical
  images of microbial behavior.
• Tools to evaluate potential process problems and
  assist with product disposition decisions.

9
Limitations of MPCM Programs

• FSIS Notice 25-05

10
Limitation of MPCM Programs

• Do not replace microbial validation or challenge
  studies.
• Do not consider effects of various food
  components on microbial behavior compared to
  experimental media.
• Do not consider changes in microbial resistance
  due to treatments (e.g., heat, acid tolerance,
  etc.)

11
Limitations of MPCM Programs (Cont.)

• Based on a few specific pathogens and their
  behavior under controlled conditions.
• Do not consider competitive or enrichment growth
  of microbes found normally in foods.
• Do not consider organisms growth phase or
  physiological state.
• No guarantee that predictive values will match
  those in the food system.

12
The Bottom Line for Using MPCM Programs

• Cannot be relied upon as the Sole means of
  assuring food safety or producing an effective
  process system.
• Are support tools that can be used in Hazard
  analyses, identifying CCPs, and validation of
  HACCP plans.
• Serve only as a tool to estimate values and not
  used as a replacement for testing values.

13
Pathogen Modeling Program (PMP) 7.0

• The PMP is a group of models that estimate the
  growth or decline of bacterial pathogens in
  specific environments.
• Estimate the effects of environmental factors on
• growth
• toxin production
• Inactivation (thermal and non-thermal)

14
ARS Pathogen Modeling Program 7.0

• Types of Models
• Growth
• Heat Inactivation
• Survival
• Cooling
• Irradiation
• http//ars.usda.gov/Services/docs.htm?docid6786

15
Four Phases of Microbial Growth

• Lag phase Cells are not multiplying, but are
  synthesizing enzymes appropriate for the
  environment.
• Exponential (or log) growth phase The cell
  population are multiplying by doubling
  (1-2-4-8-16-32-64, etc.). Therefore, logarithm
  values are used for microbial cell numbers and to
  graphically represent them.

16
Four Phases of Microbial Growth

• Stationary phase The rate of growth equals the
  rate of death, resulting in equal numbers of
  cells at any given time.
• Death phase The number of cells dying is
  greater than the number of cells growing. Death
  is due to the exhaustion of nutrients, the
  accumulation of toxic end products and/or other
  changes in the environment (e.g., pH changes).

17
Four Phases of Microbial Growth

18
(No Transcript)
19
Growth Models

• Pathogen Growth Models (Aerobic)
• Aeromonas hydrophila
• Bacillus cereus
• E. coli O157H7
• Listeria monocytogenes
• Salmonella spp.
• Salmonella typhimurium
• Shigella flexneri
• Staphylococcus aureus
• Yersinia enterocolitica

20
Growth Models

• Pathogen Growth Models (Anaerobic)
• Aeromonas hydrophila
• Bacillus cereus
• Clostridium perfringens
• E. coli O157H7
• Listeria monocytogenes
• Shigella flexneri
• Staphylococcus aureus

21
Growth Models

• Scenario 1
• XX Meats produces ground beef and is in the
  process of developing a raw productground HACCP
  plan. When conducting the hazard analysis for
  the process steps, the plant wants to determine
  if bacterial pathogen (e.g., E. coli O157H7 and
  Salmonella spp.) growth is a hazard reasonably
  likely to occur at their processing steps.
• Plant has documentation supporting growth of the
  organism is that in the model.

22
Growth Models

• Scenario 1 (cont)
• As part of their hazard analysis supporting
  documentation, the plant has determined that the
  longest that their ground beef processing takes
  is 2.5 hours. In addition, the plant has
  determined that the product enters processing at
  40? F or less and can warm up to 55? F at the end
  of 2.5 hours in the absence of temperature
  control during the processing steps.

23
Growth Models

• Scenario 1 (cont.)
• The company used the following input values in
  the ARS PMP 7.0 growth models for E. coli
  O157H7 and Salmonella spp.
• Temperature 55.0? F
• pH 6.5
• NaCl 0.5
• Sodium Nitrite 0 ppm
• No Lag phase

24
(No Transcript)
25
Growth Models

• Scenario 1 - Results from PMP 7.0
• E. coli O157H7
• Mean Generation Time 3.2
• LCL Generation Time 2.8
• UCL Generation Time 3.7
• Mean Time to Increase 1.0 logs 10.5
• LCL Time to Increase 1.0 logs 9.2
• UCL Time to Increase 1.0 logs 12.1

26
(No Transcript)
27
Growth Models

• Scenario 1 - Results from PMP 7.0
• Salmonella spp.
• Mean Generation Time 4.3
• LCL Generation Time 3.4
• UCL Generation Time 5.4
• Mean Time to Increase 1.0 logs 14.3
• LCL Time to Increase 1.0 logs 113
• UCL Time to Increase 1.0 logs 18.1

28
Growth Models

• Scenario 1 HA Decision
• Bacterial Pathogen (e.g., E. coli O157H7 and
  Salmonella spp.) growth is not a biological
  hazard reasonably likely to occur during the
  processing steps because of the very limited
  amount of growth predicted by the PMP 7.0 due to
  the short processing time and the limited
  temperature rise (55? F) in product in the
  absence of temperature control.

29
Growth Models

• Scenario 1 HA Decision (cont.)
• The room temperature will be monitored in a
  prerequisite program designed to prevent the
  outgrowth of spoilage bacteria (e.g.,
  Pseudomonas) which is a quality and not a food
  safety issue.

30
Questions

• Why did we choose No Lag over Lag phase?
• What is Generation Time?
• What is the relationship of Generation Time to
  1 Log Growth?

31
Heat Inactivation Models

• Basic inactivation terminology
• D value The time needed to kill 90 of the
  population (reduce cell numbers by 1 log10) of a
  specific microorganism in a specific food at a
  specified temperature.
• z Value The number of degrees necessary to
  change the D value one log cycle.

32
Heat Inactivation Models

• Basic inactivation terminology (cont)
• F Value The process lethality. The equivalent
  time of heating at a specific reference
  temperature.
• Log Reduction (specific microorganism)
• F Value/D Value at reference temperature for the
  specific microorganism

33
Heat Inactivation Models

• Scenario 1
• ZY Meats produces cooked corned beef with a
  product composition of 1.5 salt, 200 ppm of
  sodium nitrite, 0.2 phosphate, and a pH of 6.2.
  The company has a cooking CCP with a critical
  limit of a 6.5 log reduction for Salmonella, E.
  coli O157H7, and Listeria monocytogenes. The
  monitoring procedure is continuously monitoring
  the product core temperature for each batch of
  product which is down loaded to the AMI lethality
  spreadsheet to determine F Value and log
  reductions for the three pathogens of concern.

34
Heat Inactivation Models

• Scenario 1 Determining D Values
• In order to determine the F Value and the log
  reductions for the three pathogens for each batch
  of product, the D Value for each pathogen needs
  to be determined for the reference temperature.
  The ARS PMP 7.0 will be used to determine the D
  Values for Salmonella, E. coli O157H7, and
  Listeria monocytogenes at the reference
  temperature of 145? F which is based on the
  composition (e.g., salt concentration) of the
  corned beef.

35
(No Transcript)
36
Heat Inactivation Models

• Scenario 1 D Values Results
• Salmonella spp.
• Mean D Value 0.3 min
• LCL D Value 0.1 min
• UCL D Value 0.5 min
• E. coli O157H7
• Mean D Value 0.7 min
• LCL D Value 0.6 min
• UCL D Value 0.8 min

37
Heat Inactivation Models

• Scenario 1 D Values Results
• Listeria monocytogenes
• Mean D Value 0.9 min
• LCL D Value 0.8 min
• UCL D Value 1.1 min

38
Heat Inactivation Models

• Scenario 1 Determining F Value
• The plant recorded the following time/temp data
  for this batch of cooked corn beef

39
(No Transcript)
40
Heat Inactivation Models

• Scenario 1 F Value and log reduction results
• F Value 610.73
• Log reductions F Value/D Value at ref T
• Salmonella spp. 610.73/0.5 1221.5 logs
• E. coli O157H7 610.73/0.8 763.4 logs
• Listeria monocytogenes 610.73/1.1 555.2 logs

41
Heat Inactivation Models

• Scenario 1 Monitoring Results
• This batch of cooked corned beef based on its
  cooking profile achieved a log reduction of
  1221.5 logs, 763.4 logs, and 555.2 logs for
  Salmonella spp., E. coli O157H7, and Listeria
  monocytogenes, respectively. Consequently, this
  batch of cooked corned beef met its cooking CCPs
  critical limit of a 6.5 log reduction for
  Salmonella, E. coli O157H7, and Listeria
  monocytogenes.

42
Questions

• Why did we select the UCL values for
• D values for Salmonella, E. coli and Listeria
  monocytogenes?
• What is the real significance of the AMI model
  over the PMP model?

43
Cooling / Growth Models

• Proteolytic Clostridium botulinum in Beef Broth
• Clostridium perfringens in Beef Broth
• Clostridium perfringens Cooling Cured Beef
• Clostridium perfringens Cooling Cured Chicken

44
Cooling/Growth Models

• When entering cooling profile data, you must
  enter time in hours (e.g., 15 minutes 0.25
  hours).
• Temperature data is to be entered in the
  appropriate column as either ?C or ?F
  (conversions are automatic).

45
Cooling/Growth Models

• When applying these predictions to foods, a
  minimum of 5 time-temperature combinations must
  be measured, with 3 or more above 70? F (21? C).
• At least 5 time-temperature combinations are
  needed to sufficiently define the shape of cooked
  product cooling profile.
• The shape of cooked product cooling profile
  impacts on the amount of growth of Clostridium
  perfringens and Clostridium botulinum.

46
Cooling/Growth Chart for Rapid Cooling
47
Cooling/growth Chart forSlow Cooling
48
Cooling/growth Chart forVery Slow Cooling
49
Cooling/Growth Models

• The cooling/growth model can be used for the
  evaluation of cooling deviations if
• The conditions (e.g., food formulation) used to
  produce the PMP model match the food system or
• The model is validated for the specific cooked
  RTE meat/poultry product

50
Cooling/Growth Models

• A validated or applicable model can be used to
  determine appropriate product disposition
  (release, rework, or destroy).
• If no more than 1 log growth of C. perfringens
  and no C. botulinum growth, then the process
  meets the performance standard or Agency policy
  and can be released.

51
Cooling/Growth Models

• If there is more than 1 log growth of C.
  perfringens and no C. botulinum growth, then
  product may be either
• Recooked
• Reworked or
• Destroyed
• If there is gt 1 log growth of C. perfringens and
  0.3 log increase of C. botulinum, then product
  should be
• Destroyed

52
Cooling/Growth Models

• If there are cooling deviations and there is no
  model application data, then product can be
  either
• Destroyed or
• Microbiologically tested (if plant can show that
  no C. botulinum growth could occur-based on
  nitrite and/or salt levels)

53
Cooling/Growth Models

• Scenario 1
• The plant has a cooling CCP CL of products
  maximum internal temperature should not remain
  between 130? F and 80? F for more than 1.5 nor
  between 80? F and 40? F for more than 5 hours.
  For this cooling deviation, the cooked, uncured
  perishable product took 2 hours to reach an
  internal temperature of 80? F and then another 6
  hours to reach an internal temperature of 40? F.
• Plant has documentation supporting growth of the
  organism is that in the model.

54
Cooling/Growth Models

• Scenario 1 (cont.)
• The company recorded the following time/temp data
  as the product cooled down

55
(No Transcript)
56
(No Transcript)
57
Cooling/Growth Models

• Scenario 1 - Results from PMP 7.0
• Clostridium perfringens
• Mean Net Growth 0.18
• LCL Net Growth 0.12
• UCL Net Growth 0.25
• Clostridium botulinum
• Mean Net Growth 0.00
• LCL Net Growth - 0.01
• UCL Net Growth 0.01

58
Cooling/Growth Models

• Scenario 1 - Product Disposition
• Product would be released without any further
  action because
• The UCL net growth for Clostridium perfringens
  is 0.25 which meets the FSIS performance
  standard/policy of no more than 1.0 log increase
  for the pathogen and
• The UCL net growth for Clostridium botulinum is
  0.01 which is not more a 0.3 log increase
  indicating there was no multiplication of the
  pathogen thus meeting the FSIS performance
  standard/policy

59
Cooling/Growth Models

• Scenario 2
• The plant has a cooling CCP CL of products
  maximum internal temperature should not remain
  between 130? F and 80? F for more than 1.5 nor
  between 80? F and 40? F for more than 5 hours.
  For this cooling deviation, the cooked, uncured
  perishable product took approximately 4.5 hours
  to reach an internal temperature of 80? F and
  then another 10.5 hours to reach an internal
  temperature of 45? F.
• Plant has documentation supporting growth of the
  organism is that in the model.

60
Cooling/Growth Models

• Scenario 2 (cont.)
• The company recorded the following time/temp data
  as the product cooled down

61
Cooling/Growth Models

• Scenario 2 - Results from PMP 7.0
• Clostridium perfringens
• Mean Net Growth 1.40
• LCL Net Growth 1.07
• UCL Net Growth 1.73
• Clostridium botulinum
• Mean Net Growth 0.47
• LCL Net Growth - 0.33
• UCL Net Growth 0.61

62
Cooling/Growth Models

• Scenario 2 - Product Disposition
• Product should be destroyed because
• The Mean, LCL and UCL net growth for Clostridium
  perfringens are 1.40, 1.07, and 1.73 which
  exceeds the FSIS performance standard/policy of
  no more than 1.0 log increase for the pathogen
  and
• The Mean, LCL and UCL net growth for Clostridium
  botulinum is 0.47, 0.33, and 0.61, respectively,
  which is more than a 0.3 log increase, indicating
  there was multiplication of the pathogen thus not
  meeting the FSIS performance standard/policy of
  no growth of toxigenic microorganisms.

63
Cooling/Growth Models

• Research has shown that the following compounds
  used to control Listeria can have a significant
  impact on the outgrowth of Clostridium
  perfringens during cooling
• Buffered sodium citrate
• Buffered sodium citrate/sodium diacetate
• Sodium lactate and sodium acetate

64
Cooling/Growth Models

• Intrinsic and extrinsic factors that can impact
  on Clostridium perfringens growth during cooling
  (cont)
• Salt (NaCl)
• Sample bag type (oxygen permeability)
• Growth Medium
• Broth versus ground beef
• Faster growth rates in ground beef

65
THANK YOU!
DO YOU HAVE ANY QUESTIONS?