Bacterial thermal death kinetics based on probability distributions: The heat destruction of Clostridium botulinum and Salmonella Bedford

Despite the long history and excellent record of inactivation models used i n thermal processing, there are relatively few approaches that attempt to d escribe the kinetics commonly observed. There are even fewer examples of mo dels that allow the user to deal with the environmental conditions that inf luence these kinetics. We describe an approach that assumes a distribution of inactivation times within a population of bacterial cells. The concept a llows for alternative interpretations of death kinetics and provides excell ent descriptions of data generated with two important foodborne pathogens, Clostridium botulinum and Salmonella Bedford. The Salmonella Bedford data s et used is unusual and perhaps unique in that it provides information where more than 50% of the population survival has been measured. These measurem ents are often overlooked or missed in experimental work but are essential when using a vitalistic approach, enabling calculation of a 50% lethal dose for destruction of bacteria. Use of the normal or Prentice distribution pr ovided better fits to the data than other models commonly used to describe thermal death. There was no obvious bias in the fits even though significan t tailing was evident. In addition, the procedure described allows data fro m all the conditions to be fitted rather than individual independent series . This enables a single equation to be derived that can be judged against t he whole domain of the data. Approaches that provide accurate and unbiased descriptions of thermal death are likely to become increasingly important t o ensure the safety of more marginal heat processes.