Modelling the impact of mild food processing conditions on the microbiological safety of food

Abstract

There is significant interest by the food industry in applying milder processing conditions. A major area of research within predictive modelling has been the search for models which accurately predict the effect of combining multiple processes or hurdles. For a mild process, which has temperature as the major microbial injury step, the effect of the other combined hurdles in inhibiting growth of the injured organisms must be understood. The latter means that the inoculum size dependency of the time to growth must also be fully understood. This essentially links injury steps with the potential for growth. Herein, we have been developing the use of optical density (O.D) for obtaining growth rates and lag times using multiple inocula rather than using the traditional methods which use one single inoculum. All analyses were performed in the Bioscreen analyser which measures O.D. The time to detection (TTD) was defined as the time needed for each inoculum to reach an O.D=0.2 and O.D was related to microbial numbers with simple calibration curves. Several primary models were used to predict growth curves from O.D data and it was shown that the classic logistic, the Baranyi and the 3-phase linear model (3-PLM) were the most capable primary models of those examined while the modified Gompertz and modified logistic could not reproduce TTD data. Using the Malthusian approximation of the logistic model the effect of mild temperature shifts was studied. The data obtained showed that for mild temperature shifts, growth rates quickly changed to the new environment without the induction of lags. The growth of Listeria monocytogenes, Salmonella Typhimurium and Escherichia coli was studied at 30⁰C and/or 37⁰C, in different NaCl concentrations, pH and their combinations. The classical 3-parameter logistic with lag model was rearranged to provide the theoretical foundation for the observed TTD and accurate growth rates and lag times could be estimated. As the conditions became more unfavourable, the lag time increased while the growth rate decreased. Also, the growth rate was found to be independent from the inoculum size; the inoculum size affected only the TTD. The Minimum Inhibitory Concentration (MICNaCl and MICpH) was calculated using the Lambert and Pearson model (LPM) and also the Growth/No Growth (G/NG) interface was determined using combinations of NaCl and pH. These data were transformed in rate to detection (RTD) and fitted with a response surface model (RSM) which was subsequently compared with the Extended LPM (ELPM). The LPM and the ELPM could analyse results from individual and combined inhibitors, respectively. Following a mild thermal process a lag due to thermal injury was also induced, the magnitude of which was dependent on the organism and environmental conditions; the observed distribution of the lags appeared, in general, to follow the Log-normal distribution. After the lag period due to injury, growth recommenced at the rate dictated by the growth environment present. Traditional growth curves were constructed and compared with the data obtained from the Bioscreen under the same conditions. From the results obtained, it can be suggested that the increased lag times and growth rates obtained from the traditional plate counts compared with the values obtained from the Bioscreen microbiological analyser, might be an artifact of the plating method or may be due to the use of the modified Gompertz to study the growth. In conclusion, O.D can be used to accurately determine growth parameters, to give a better understanding and quantify the G/NG interface and to examine a wealth of phenomena such as fluctuating temperatures and mild thermal treatments. The comparison between the traditional growth curves against the data obtained from the Bioscreen showed that the TTD method is a rapid, more accurate and cheaper method than the traditional plate count method which in combination with the models developed herein can offer new possibilities both to the research and the food industry.