Management of Clostridium botulinum in food: A predictive food microbiology approach

Ioulia Koukou

Research output: Book/ReportPh.D. thesis

841 Downloads (Pure)


The current PhD thesis focused on the development of growth and growth boundary models for proteolytic and non-proteolytic Clostridium botulinum. C. botulinum is a pathogenic bacterium that produces very potent neurotoxins (BoNTs). These toxins, if present in food, can cause an intoxication to humans called botulism; a neuroparalytic disease that can cause prolonged physical disability or even death.
Due to the high toxicity of C. botulinum specific codes of practice, premises and trained personnel are required in order to handle this pathogenic bacterium. The use of non-toxigenic surrogates for laboratory experiments is therefore more efficient and less costly. For this reason, non-toxigenic surrogates were used during the present PhD study for model development, calibration and evaluation. Four naturally non-toxigenic Clostridium sp. isolates that showed 99.8 to 99.9% 16S rRNA similarity with non-proteolytic C. botulinum were used as surrogates of non-proteolytic C. botulinum. Similarly, three strains of C. sporogenes were used as surrogates of proteolytic C. botulinum.
Detection of C. botulinum with real-time PCR (qPCR) usually relies on the amplification of sequences from botulinum neurotoxin genes. Within the present PhD study, a qPCR method able to detect and quantify the four non-toxigenic and non-proteolytic Clostridium sp. strains was developed (Paper II). Primers specific to a 170-bp fragment of the 16S rRNA gene were designed and the method was able to quantify a 5-log (CFU/ml) range from 1.95 to 7.0 log (CFU/ml).
Non-proteolytic C. botulinum is a psychrotrophic bacterium which shows a high prevalence in some aquatic environments and has been implicated in numerous botulism outbreaks caused by the consumption of seafood. Management of non-proteolytic C. botulinum in processed seafood usually relies on refrigeration and the addition of salt in amounts higher than 3.5% in the water phase of the product. However, public health organizations recommend a reduction in our daily salt intake due to associations of salt consumption and heart disease.
A new growth and growth boundary model for non-proteolytic Clostridium sp., as a surrogate of non-proteolytic C. botulinum, was developed. This model included the effect of eight environmental factors and their interaction effect (Paper I). The model was calibrated in various seafood products and was successfully validated for seafood and poultry products with time-to-toxin (TTT) bias factors of 1.02 (n = 395) and 0.82 (n = 298), respectively. Furthermore the model can be used to make predictions for vegetable products using a higher safety margin. The model for non-proteolytic Clostridium sp. (Paper I) can be applied for products in modified atmosphere packaging (MAP) and smoked products without taking into account the effect of CO2, and phenol, respectively. Finally, it was shown that lactic acid bacteria (LAB) had no significant inhibiting effect on growth rates (P = 0.49) of non-proteolytic Clostridium sp. (Paper II).
Proteolytic C. botulinum is a mesophilic bacterium which grows well at ambient temperatures and does not grow below 10°C. Spores of proteolytic C. botulinum are widespread in nature and therefore very likely to be found in food. Although outbreaks of botulism due to consumption of processed cheeses are rare, C. botulinum can grow and produce toxins in processed cheeses stored at ambient temperatures. Processed cheeses are usually stored at refrigeration temperatures. However, the raising demand of this type of cheeses in new markets (e.g. Africa, Middle East) has created a challenge in relation to chilled distribution and storage. Furthermore, the available guidelines for management of proteolytic C. botulinum do not provide flexibility for product development of this type of products. These guidelines suggest a minimum water phase salt (WPS) of 10% or a pH below 4.6, if a sufficient heat-treatment step (121°C for 3 min) is not applied during processing.
A new growth and growth boundary model for C. sporogenes, as a surrogate of C. botulinum was developed. The model included the effect of eleven environmental factors, including dairy specific ingredients such as melting salts, and the interaction effect between the eleven factors. The model was calibrated with processed cheese and successfully validated for processed cheese and meat products with μmax-bias factors of 1.19 (n = 20) and 0.95 (n = 72), respectively (Paper III). Furthermore, this extensive model showed a wider range of applicability and similar performance when it was compared to other available proteolytic C. botulinum models for processed cheese (Paper IV).
The two mathematical predictive models developed in the present PhD thesis can be used by food business operators for product development or reformulation of existing recipes and for documentation of food safety. Furthermore, the models can be used by the authorities to support more flexible guidelines that do not rely on the growth limits of the bacterium that may lead to unnecessary over-preservation.
Original languageEnglish
Place of PublicationKgs. Lyngby
PublisherDTU Food
Number of pages199
Publication statusPublished - 2021


Dive into the research topics of 'Management of Clostridium botulinum in food: A predictive food microbiology approach'. Together they form a unique fingerprint.

Cite this