Several models have been developed and are currently available to predict C. perfringens germination and outgrowth in meat and poultry products. The most widely used ''tertiary'' predictive models (software/graphic user interface) are the USDA Agricultural Research Service (USDA-ARS) Pathogen Modeling Program (PMP) and the Perfringens Predictor. These programs have an easy interface, where the user could upload the temperature profile of the product and predict the potential germination and outgrowth of C. perfringens. While the PMP was developed using isothermal growth data for C. perfringens using microbiological media, recent updates to the software include models developed using meat systems. The underlying models used in Perfringens Predictor were collected from literature and was used to develop a dynamic model that allows the user to specify the temperature profile of the product to evaluate potential C. perfringens spore germination and outgrowth.
One of the first approaches to predictive modeling of C. perfringens was by Labbe and Huang (1995) using laboratory media (fluid thioglycollate medium), media supplemented with beef, and autoclaved ground beef matrices. As reported in a previous research (Willardsen, Busta, & Allen, 1979), faster growth rates were observed in autoclaved ground beef compared to laboratory media.
Subsequently, Juneja, Marmer, Phillips, and Palumbo (1996) developed a predictive model for vegetative growth of C. perfringens that incorporated interactive effects of temperature (12-42°C), product pH (5.5-7), sodium chloride (0-3%), and sodium pyrophosphate (0-0.3%) using a model system (tryp-ticase-peptone-glucose-yeast extract broth). The maximal growth rate was observed at 42°C, pH 6.25, with a GT of 12 min and a lag phase duration of 2.27 h. Interactions between the ingredients and product pH on C. perfringens growth were observed. This report only provided information on C. perfringens growth at isothermal temperatures, and a secondary model to explain C. perfringens growth with changes in temperatures (dynamic) over time was not developed.
Further, RTE meat and poultry products are currently formulated to contain salt, phosphates, curing agent (sodium nitrite), a reducing agent (sodium erythorbate), and quite often antimicrobial ingredients such as organic acid salts (sodium or potassium salts of lactic or citric acids). These ingredients have been shown to affect the germination and outgrowth of C. perfringens in meat systems. However, modeling of all these parameters, taking into consideration the variation in concentrations used by RTE meat and poultry processors, may be a daunting task. However, judicious application of predictive models that estimate the potential germination and outgrowth of C. perfringens under worst-case scenario with some information derived from challenge studies evaluating the effects of antimicrobial agents can provide reasonable assurance on the safety of the resulting products.
Limitations of the earlier models include the use of laboratory media, especially since C. perfringens can grow faster in meat systems compared to laboratory media (Willardsen et al., 1979). In a subsequent report, Juneja, Whiting, Marks, and Snyder (1999) described a secondary model for growth of C. perfringens from heat-activated spores during cooling using a meat system (autoclaved ground beef). The limitations of this model have been highlighted by Smith and Schaffner (2004), indicating that the exponential growth rates (EGR) were responsible for the under-prediction of C. perfringens growth rather than the germination, outgrowth, and lag phase (GOL). The authors report that the model performed relatively well (fail safe) when low (<1 log CFU/ml) or high (>3 log CFU/ml) growth was observed (increases) during exponential cooling. However, the model consistently under-predicted growth at intermediate observed increases (1-3 log CFU/ml) as well as in trials using two different rates of exponential cooling.
In an effort to simplify this process, Huang (2002) described the outgrowth of heat-activated spores of C. perfringens in cooked beef and developed a multiple linear model. The growth curves at various temperatures were generated and fitted to the Gompertz equation and a modified multiple linear model. The model consisted of five linear segments to describe the sigmoidal growth of C. perfringens for each temperature. The growth curve was divided into five linear segments described as lag, first transitional, exponential, second
transitional, and stationary phases (Fig. 22.3). This allowed the author to derive lag phase duration parameters as a linear function of the traditional lag phase duration calculated from the Gompertz equation. This in turn permits three-segment linear models to be used to generate five-segment linear growth curves with no need to solve mathematical functions. For the linear models, the mean growth rates observed in the transitional phases were considered the same. The primary models were then fitted to a square root function to determine the effect of temperature on growth parameters. Huang (2002) concluded that the linear method accurately described the sigmoidal shape of growth curves and provided similar parameters for secondary modeling as the Gompertz function.
Recent research includes a model for cured pork ham using the Baranyi model to determine growth kinetic parameters of isothermal growth curves and the square root Ratkowsky model to represent the exponential growth rates as a function of temperature and the Runge-Kutta procedure to solve the numerical functions (Amezquita, Wang, Weller, Thippareddi, & Burson, 2005). Similar methodologies were also used to develop predictive models for cured and non-cured roast beef, cured and non-cured ground pork, and cured and non-cured ground turkey.
Research is needed to determine the interactive effects of other ingredients in the formulation of meat products, as the ingredients such as curing salts, phosphates, and salts of organic acids affect the growth of this organism. Formulating buffered sodium citrate (1.3%), buffered sodium citrate supplemented with sodium diacetate (1.3%), as well as a mixture of sodium lactate and potassium lactate (2.5%), and a mixture of sodium lactate and sodium diacetate (2.5%; 6:4 mixture) were shown to be sufficient to completely inhibit the growth of C. perfringens during extended cooling of injected meat products (Thippareddi, Juneja, Phebus, Marsden, & Kastner, 2003).
Since the issue of germination and outgrowth of C. perfringens spores is time dependent, and during cooling a continuously varying temperature conditions exist, the models should be able to predict the germination and outgrowth of C. perfringens spores under those conditions. As has been reported, the germination process, outgrowth, or both can be affected by antimicrobial ingredients present in the meat and poultry products. Further, RTE meat and poultry processors may encounter cooling deviations, either due to power failure or refrigeration system failure, resulting in non-continuous chilling rates. In such circumstances, the models should be robust enough to be able to accurately predict the potential germination and outgrowth of C. perfringens spores in the particular product of concern. Amezquita et al. (2005) developed a finite element heat diffusion model to predict the temperature of meat product (ham) and integrated it with C. perfringens growth model. Such models allow processors to evaluate the adequacy of their cooling systems or design products with dimensions that allow proper cooling of the products to minimize the risk of C. perfringens germination and outgrowth (Fig. 22.4).
20 I 10
— Predicted growth
— Predicted growth
20 I 10
Fig. 22.4 Comparison of predicted and observed C. perfringens growth during cooling of boneless cooked cured ham, when a deviation from FSIS compliance guidelines occurs at 1.8 h into cooling. The simulated deviation is caused by unexpected equipment failure or electrical outage for a total downtime of 1 h. Cooling time from 54.4 to 26.6°C is 6.6 h and from 26.6 to 7.2°C is 13.2 h (Amezquita et al., 2005)
Was this article helpful?