Indicators of Food Microbial Quality and Safety
491
In a study of fecal material from 39 diamond-back terrapins from brackish waters along the eastsoutheast U.S. coast, 51% of cloacal swabs were positive for fecal coliforms, and 80% of 10 fecal
specimens were positive.
31
Like many others, the above studies make it clear that coliforms and fecal coliforms exist in a
number of places and in many raw food materials where their presence has little or no relationship
to food safety. The excessive and unwise use of these indicators may, on the one hand, lead to the
rejection of safe products; on the other hand, it may lead to the acceptance of unsafe products because
the incorrect indicators were used.
PREDICTIVE MICROBIOLOGY/MICROBIAL MODELING
The presence/absence of indicator organisms as noted above is used to predict food safety. If a
safety indicator is absent, the product is regarded as being safe relative to the hazard for which the
indicator is used. On the other hand, a product can have extremely low numbers of a safety indicator
and yet not pose a hazard. The latter is true for many foodborne pathogens such as enterotoxigenic
staphylococci. When low numbers of an indicator or pathogen are present, it is important to know how
either will behave in a food product over time. This future behavior calls into question the multiple
parameters that affect the growth and activity of microorganisms in foods, and if predictions are to
be made about the fate of low numbers of pathogens in a given product, how the pathogens and these
parameters interact needs to be handled.
Microbial modeling or predictive microbiology is a rapidly emerging subdiscipline that entails the
use of mathematical models/equations to predict the growth and/or activity of a microorganism in a
food product over time. The predictive or modeling aspect is not new, for it is embodied in heat-process
calculations in the canning of low-acid foods that are described in Chapter 17. What is new is the
interest in broadening the use of this concept to a wider range of food poisoning and food spoilage
organisms by the use of more sophisticated mathematical/computer models that can handle multiple
growth parameters. For more information see references 2, 28, 54, 90, 93, 94.
As noted in Chapter 2 under temperature effects, predicting the growth of an organism for a single
parameter is not too difficult. Difficulty arises when multiple parameters are involved, as relatively
few studies have been conducted to determine their interplay on organisms. One such study is that
of Buchanan and Phillips
8 who studied the interaction of five parameters on the growth of Listeria
monocytogenes: pH, temperature, nitrite, NaCl, and gaseous atmosphere. After fitting data from 709(!)
growth curves using a nonlinear regression analysis in conjunction with the Gompertz function, the
investigators concluded that the model could provide reasonable “first estimates” on the behavior of
L. monocytogenes.
The effective application of predictive microbiology requires the selection of appropriate models to
reflect the effect of growth parameters. Among the many models that have been proposed and tested
are two kinetic models—the nonlinear Arrhenius and Bˇ elehrad´ ek types. The former is applied with
the dependent variable expressed as ln rate, whereas with the latter square-root model, the dependent
variable is expressed as
√
rate. The further development of these models and their utility have been
outlined and discussed by Ratkowsky et al.
72 Computer software packages for predictive microbiology
are available from private and commercial sources.
One of the simplest applications of predictive microbiology is the use of Monte Carlo Simulation.
This method entails probability distributions based on collected data to predict shelf life/safety relative
to changes in environmental parameters such as pH, a w , etc. When Monte Carlo was used to predict the
shelf life of pasteurized milk, the data employed consisted of initial numbers of spoilage organisms in
491
In a study of fecal material from 39 diamond-back terrapins from brackish waters along the eastsoutheast U.S. coast, 51% of cloacal swabs were positive for fecal coliforms, and 80% of 10 fecal
specimens were positive.
31
Like many others, the above studies make it clear that coliforms and fecal coliforms exist in a
number of places and in many raw food materials where their presence has little or no relationship
to food safety. The excessive and unwise use of these indicators may, on the one hand, lead to the
rejection of safe products; on the other hand, it may lead to the acceptance of unsafe products because
the incorrect indicators were used.
PREDICTIVE MICROBIOLOGY/MICROBIAL MODELING
The presence/absence of indicator organisms as noted above is used to predict food safety. If a
safety indicator is absent, the product is regarded as being safe relative to the hazard for which the
indicator is used. On the other hand, a product can have extremely low numbers of a safety indicator
and yet not pose a hazard. The latter is true for many foodborne pathogens such as enterotoxigenic
staphylococci. When low numbers of an indicator or pathogen are present, it is important to know how
either will behave in a food product over time. This future behavior calls into question the multiple
parameters that affect the growth and activity of microorganisms in foods, and if predictions are to
be made about the fate of low numbers of pathogens in a given product, how the pathogens and these
parameters interact needs to be handled.
Microbial modeling or predictive microbiology is a rapidly emerging subdiscipline that entails the
use of mathematical models/equations to predict the growth and/or activity of a microorganism in a
food product over time. The predictive or modeling aspect is not new, for it is embodied in heat-process
calculations in the canning of low-acid foods that are described in Chapter 17. What is new is the
interest in broadening the use of this concept to a wider range of food poisoning and food spoilage
organisms by the use of more sophisticated mathematical/computer models that can handle multiple
growth parameters. For more information see references 2, 28, 54, 90, 93, 94.
As noted in Chapter 2 under temperature effects, predicting the growth of an organism for a single
parameter is not too difficult. Difficulty arises when multiple parameters are involved, as relatively
few studies have been conducted to determine their interplay on organisms. One such study is that
of Buchanan and Phillips
8 who studied the interaction of five parameters on the growth of Listeria
monocytogenes: pH, temperature, nitrite, NaCl, and gaseous atmosphere. After fitting data from 709(!)
growth curves using a nonlinear regression analysis in conjunction with the Gompertz function, the
investigators concluded that the model could provide reasonable “first estimates” on the behavior of
L. monocytogenes.
The effective application of predictive microbiology requires the selection of appropriate models to
reflect the effect of growth parameters. Among the many models that have been proposed and tested
are two kinetic models—the nonlinear Arrhenius and Bˇ elehrad´ ek types. The former is applied with
the dependent variable expressed as ln rate, whereas with the latter square-root model, the dependent
variable is expressed as
√
rate. The further development of these models and their utility have been
outlined and discussed by Ratkowsky et al.
72 Computer software packages for predictive microbiology
are available from private and commercial sources.
One of the simplest applications of predictive microbiology is the use of Monte Carlo Simulation.
This method entails probability distributions based on collected data to predict shelf life/safety relative
to changes in environmental parameters such as pH, a w , etc. When Monte Carlo was used to predict the
shelf life of pasteurized milk, the data employed consisted of initial numbers of spoilage organisms in