TY - JOUR
T1 - Analysis of a quantitative risk assessment of listeriosis from pasteurized milk
T2 - The combinations of which factors cause listeriosis in this low-risk food?
AU - Abe, Hiroki
AU - Garre, Alberto
AU - Koseki, Shige
AU - den Besten, Heidy M.W.
AU - Zwietering, Marcel H.
PY - 2023/10
Y1 - 2023/10
N2 - This study proposes a novel methodology for risk assessment of products with extremely low risk. The method is based on the analysis of those iterations that result in illness occurrence. It is demonstrated using a hypothetical scenario on listeriosis from pasteurized milk heated at 72°C–75 °C for 15–20 s and analysed which combinations of factors resulted in illness. Sixty-one cases of listeriosis were predicted from 10 billion simulations, representing a realistically large number of servings for this product. According to the model simulations, the illness cases were caused by extremely high doses resulting from three rare situations occurring concurrently: high initial level, less effective pasteurization (due to high thermotolerance of L. monocytogenes contaminants) and extremely high microbial growth during domestic storage (due to poor storage conditions). However, listeriosis was not always observed even if the three situations occurred. Infectivity (dose–response parameter) had the strongest relevance to illness occurrence. Notably, the results of the sensitivity analysis varied depending on the output variable (microbial concentration at exposure or illness occurrence). Furthermore, the correlation-based sensitivity analysis for the illness occurrence provided unreliable results, discouraging this approach for food products with an extremely low illness probability. Considering that number of illnesses and not exposure is the output variable most relevant for risk management, we propose an innovative method based on graphical representations for sensitivity analysis in low-risk products.
AB - This study proposes a novel methodology for risk assessment of products with extremely low risk. The method is based on the analysis of those iterations that result in illness occurrence. It is demonstrated using a hypothetical scenario on listeriosis from pasteurized milk heated at 72°C–75 °C for 15–20 s and analysed which combinations of factors resulted in illness. Sixty-one cases of listeriosis were predicted from 10 billion simulations, representing a realistically large number of servings for this product. According to the model simulations, the illness cases were caused by extremely high doses resulting from three rare situations occurring concurrently: high initial level, less effective pasteurization (due to high thermotolerance of L. monocytogenes contaminants) and extremely high microbial growth during domestic storage (due to poor storage conditions). However, listeriosis was not always observed even if the three situations occurred. Infectivity (dose–response parameter) had the strongest relevance to illness occurrence. Notably, the results of the sensitivity analysis varied depending on the output variable (microbial concentration at exposure or illness occurrence). Furthermore, the correlation-based sensitivity analysis for the illness occurrence provided unreliable results, discouraging this approach for food products with an extremely low illness probability. Considering that number of illnesses and not exposure is the output variable most relevant for risk management, we propose an innovative method based on graphical representations for sensitivity analysis in low-risk products.
KW - dairy
KW - Listeria monocytogenes
KW - low-risk products
KW - Risk Analysis
KW - sensitivity analysis
U2 - 10.1016/j.foodcont.2023.109831
DO - 10.1016/j.foodcont.2023.109831
M3 - Article
AN - SCOPUS:85158051262
SN - 0956-7135
VL - 152
JO - Food Control
JF - Food Control
M1 - 109831
ER -