Sensitivity analysis for high quantiles of ochratoxin A exposure distribution

Area of Science:

• Food safety and toxicology
• Risk assessment methodology
• Statistical modeling in environmental health

Background:

• Accurate exposure assessment for food contaminants like ochratoxin A (OA) is crucial for managing foodborne disease risks.
• Risk assessors require reliable high quantiles (95th and 99th) of contaminant exposure distributions, which are challenging to obtain due to data asymmetry.
• Ochratoxin A exposure distribution is complex, derived from the product of consumption and contamination distributions across multiple foodstuffs.

Purpose of the Study:

• To perform a sensitivity analysis (SA) on high quantiles of ochratoxin A (OA) exposure distribution in French children.
• To quantify the influence of consumption and contamination parameters on OA exposure quantiles.
• To develop validated models for predicting OA exposure at high quantiles.

Main Methods:

• Utilized Monte Carlo simulation with available consumption and contamination data for eight foodstuffs.
• Employed a quadratic polynomial regression model with 32 main factors, interactions, and quadratic terms for SA.
• Implemented a resolution-V fractional factorial design (6561 experiments) for parameter estimation and bootstrap sampling for factor range determination.

Main Results:

• Sensitivity analysis identified key parameters influencing OA exposure quantiles.
• Parameters related to "Cereals" consumption and contamination demonstrated a major influence on OA exposure.
• "Pork" consumption parameters showed a lesser, but still notable, influence on OA exposure sensitivity.

Conclusions:

• Validated models were developed for predicting high quantiles of ochratoxin A exposure.
• The study highlights the critical role of "Cereals" in the overall ochratoxin A exposure profile of French children.
• Findings provide valuable insights for targeted risk management strategies concerning ochratoxin A contamination.