Probabilistic exposure assessment model to estimate aseptic-UHT product failure rate

Abstract Aseptic-Ultra-High-Temperature (UHT) products are manufactured to be free of microorganisms capable of growing in the food at normal non-refrigerated conditions at which the food is likely to be held during manufacture, distribution and storage. Two important phases within the process are widely recognised as critical in controlling microbial contamination: the sterilisation steps and the following aseptic steps. Of the microbial hazards, the pathogen spore formers Clostridium botulinum and Bacillus cereus are deemed the most pertinent to be controlled. In addition, due to a relatively high thermal resistance, Geobacillus stearothermophilus spores are considered a concern for spoilage of low acid aseptic-UHT products. A probabilistic exposure assessment model has been developed in order to assess the aseptic-UHT product failure rate associated with these three bacteria. It was a Modular Process Risk Model, based on nine modules. They described: i) the microbial contamination introduced by the raw materials, either from the product (i.e. milk, cocoa and dextrose powders and water) or the packaging (i.e. bottle and sealing component), ii) the sterilisation processes, of either the product or the packaging material, iii) the possible recontamination during subsequent processing of both product and packaging. The Sterility Failure Rate ( SFR ) was defined as the sum of bottles contaminated for each batch, divided by the total number of bottles produced per process line run (10 6 batches simulated per process line). The SFR associated with the three bacteria was estimated at the last step of the process (i.e. after Module 9) but also after each module, allowing for the identification of modules, and responsible contamination pathways, with higher or lower intermediate SFR . The model contained 42 controlled settings associated with factory environment, process line or product formulation, and more than 55 probabilistic inputs corresponding to inputs with variability conditional to a mean uncertainty. It was developed in @Risk and run through Monte Carlo simulations. Overall, the highest SFR was associated with G . stearothermophilus (380000 bottles contaminated in 10 11 bottles produced) and the lowest to C. botulinum (3 bottles contaminated in 10 11 bottles produced). Unsurprisingly, SFR due to G. stearothermophilus was due to its ability to survive the UHT treatment. More interestingly, it was identified that SFR due to B. cereus (17000 bottles contaminated in 10 11 bottles produced) was due to an airborne recontamination of the aseptic tank (49%) and a post-sterilisation packaging contamination (33%). A deeper analysis (sensitivity and scenario analyses) was done to investigate how the SFR due to B. cereus could be reduced by changing the process settings related to potential air recontamination source.