A mathematical risk model for Escherichia coli O157:H7 cross-contamination of lettuce during processing

• Published in: Food microbiology Vol. 28; no. 4; pp. 694 - 701
• Main Authors: Pérez Rodríguez, F., Campos, D., Ryser, E.T., Buchholz, A.L., Posada-Izquierdo, G.D., Marks, B.P., Zurera, G., Todd, E.
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Kidlington Elsevier Ltd 01-06-2011; Elsevier
• Subjects: Analysis of variance; Bags; Biological and medical sciences; Chlorination; Computer Simulation; Contamination; Cross-contamination; Escherichia coli; Escherichia coli Infections - prevention & control; Escherichia coli O157 - growth & development; Escherichia coli O157:H7; Food industries; Food microbiology; Food Microbiology - methods; Food-Processing Industry - methods; Fruit and vegetable industries; Fundamental and applied biological sciences. Psychology; Irradiation; Lactuca - microbiology; Lettuce; Lettuces; Mathematical modelling; Mathematical models; Models, Biological; Models, Statistical; Pathogens; Risk; Risk assessment; Sampling plan; Lettuce; Cross-contamination; Chlorination; Risk assessment; Irradiation; Mathematical modelling; Sampling plan; Escherichia coli O157:H7; Vegetables; Escherichia coli; Samplings; Risk; Contamination; Bacteria; Sampling; Mathematical model; Enterobacteriaceae
• ISSN: 0740-0020; 1095-9998
• DOI: 10.1016/j.fm.2010.06.008

A stochastic simulation modelling approach was taken to determine the extent of
Escherichia coli O157:H7 contamination in fresh-cut bagged lettuce leaving the processing plant. A probabilistic model was constructed in Excel to account for
E. coli O157:H7 cross contamination when contaminated lettuce enters the processing line. Simulation of the model was performed using @Risk Palisade© Software, providing an estimate of concentration and prevalence in the final bags of product. Three different scenarios, named S1, S2, and S3, were considered to represent the initial concentration on the contaminated batch entering the processing line which corresponded to 0.01, 1 and 100
cfu/g, respectively. The model was satisfactorily validated based on Standard Error of Prediction (SEP), which ranged from 0.00-35%. ANOVA analysis performed on simulated data revealed that the initial concentration in the contaminated batch (i.e., S1, S2, and S3) did not influence significantly (
p
=
0.4) the
E. coli O157:H7 levels in bags derived from cross contamination. In addition, significantly different (
p
<
0.001) prevalence was observed at the different levels simulated (S1; S2 and S3). At the lowest contamination level (0.01
cfu/g), bags were cross-contaminated sporadically, resulting in very low
E. coli O157:H7 populations (mean: ≤2
cfu/bag) and prevalence levels (<1%). In contrast, higher average prevalence levels were obtained for S2 and S3 corresponding to 3.05 and 13.39%, respectively. Furthermore, the impact of different interventions on
E. coli O157:H7 cross-contamination (e.g., pathogen testing, chlorination, irradiation, and cleaning and disinfection procedures) was evaluated. Model showed that the pathogen was able to survive and be present in the final bags in all simulated interventions scenarios although irradiation (0.5
KGy) was a more effective decontamination step in reducing prevalence than chlorination or pathogen testing under the same simulated conditions.