Improving preharvest sampling for pathogens in leafy greens through simulations and development of alternative approaches

Quintanilla Portillo, Jorge Francisco

Permalink

https://hdl.handle.net/2142/124554 Copy

Description

Title

Improving preharvest sampling for pathogens in leafy greens through simulations and development of alternative approaches

Author(s)

Quintanilla Portillo, Jorge Francisco

Issue Date

2024-04-23

Director of Research (if dissertation) or Advisor (if thesis)

Stasiewicz, Matthew J

Doctoral Committee Chair(s)

Miller, Michael J

Committee Member(s)

• Davidson, Paul C
• Varga, Csaba

Department of Study

Food Science & Human Nutrition

Discipline

Food Science & Human Nutrition

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Food Safety
• Romaine Lettuce
• Modeling
• Microtally
• Commercial
• Quality And Safety Indicators

Language

eng

Abstract

The food safety of leafy green products has been increasingly scrutinized given its association with foodborne illness outbreaks. Given that leafy green products usually undergo minimal processing, and lack of kill step, it is critical to manage the risk of pathogen contamination risk before the product enters the market. Preharvest sampling and testing leafy green products for pathogen contamination have been used as a tool to manage this risk. However, the development of powerful preharvest sampling plans is difficult since pathogen contamination usually occurs at low prevalence or low concentrations. In Chapter 2, a sampling simulation model was developed and validated against literature and experimental field trials. This model simulated hazards on simulated 2-dimensional fields. On these, sampling and testing were simulated, and yielded a decision based on test results. The simulation was validated against six field trials by outputting six simulated ranges of positive samples that contained the experimental results (2 – 139 positive samples). The sampling simulation was then used to evaluate the power of sampling plans to recover contamination from four contamination scenarios in generic, one-acre simulated leafy green fields. The first two experiments evaluated the relative performance of sampling patterns to recover point source (0.3% to 1.7% prevalence, 3 Log (CFU/g)) or widespread contamination (-3 Log (CFU/g), 100% prevalence). Results showed that sampling with randomization yielded low variability in the probability of acceptance, hence better sampling performance. The following two experiments evaluated sampling with randomization to recover point source (3 Log(CFU/g) at 0.3% prevalence) or widespread contamination (-1 to -4 Log(CFU/g)). Overall, results showed that collecting more, smaller samples with randomization increased the power to detect point source and low-level widespread contamination. In Chapter 3, the validated sampling simulation was used to assess sampling plans in simulated fresh produce fields in representative regions of the United States. Regions were Arizona, California, New York and Virginia. Each simulated field was parameterized based on commodity-specific production. For each field, simulated hazards were presented as point source, widespread systematic and sporadic contamination. Each simulated field was sampled by collecting 1, 5, and 10 composite produce samples (150, 300, and 1,200 g per sample). Simulations predicted that collecting more composite samples with randomization increased the power of sampling to recover point source and systematic contamination. Simulations predicted that variability in contamination concentration and small contamination spread affects sampling performance to detect point source contamination. Preharvest simulated sampling would not reliably recover sporadic, low-level contamination. In Chapter 4, aggregative sampling is evaluated as an alternative, non-destructive sampling to composite produce samples. One experiment with two replicates collected aggregative and composite samples throughout the preharvest, harvest, and postharvest stages of romaine lettuce lots. Overall, aggregative samples recovered higher Aerobic Plate Counts (APC) than composite produce samples but showed variability in recovery of total coliforms. Particularly, aggregative gloves collected during harvest and postharvest, and aggregative swabs collected from the inside of the romaine heads showed to perform as well as produce samples to recover coliform counts. Aggregative swabs generally underperformed recovering coliforms, as opposed to composite produce samples. Further research was conducted to improve the sampling technique for preharvest aggregative sampling but was not included in this document. In Chapter 5, we evaluated pressure and hydration as factors that affect the recovery of APC, coliforms, and generic E. coli to the aggregative swab during preharvest sampling of commercial romaine lettuce fields. One experiment with two replicated was performed in 2022 and 2023. We evaluated pressure by collecting aggregative swab samples by hand or by using a Manual Sampling Device (MSD), and for hydration, we evaluated dry and pre-hydrated swabs. Results showed that hydration was the main factor driving the recoveries, as they were significantly higher than dry samples. Pressure did not have a significant effect. The improved technique was selected as aggregative sampling using pre-hydrated swabs attached to the MSD, these samples were compared to paired composite produce samples. Overall, results showed that aggregative samples performed as well as composite samples to recover APC, coliforms, and generic E. coli. Furthermore, aggregative sampling can be used as a tool for preharvest sampling but the results should be interpreted carefully, as they provide a different perspective than product testing

Graduation Semester

2024-05

Type of Resource

Text

Handle URL

https://hdl.handle.net/2142/124554

Copyright and License Information

Copyright 2024 Jorge Quintanilla Portillo

Owning Collections