# Development of Dose‐Response Models for the Ingestion Exposure Route and Stillbirth Outcome for Listeria monocytogenes

## TL;DR

This paper develops new dose-response models for Listeria monocytogenes that predict infection and stillbirth in pregnant individuals, improving food safety assessments for this vulnerable group.

## Contribution

The study introduces biologically plausible dose-response models for pregnant hosts, specifically for infection and stillbirth outcomes.

## Key findings

- A pooled beta-Poisson model for maternal infection has parameters α = 0.0843 and N50 = 1.01 × 10⁸.
- A pooled beta-Poisson model for stillbirth has parameters α = 0.575 and N50 = 1.31 × 10⁷.

## Abstract

Foodborne listeriosis can cause stillbirth in pregnant individuals, prompting numerous population‐specific food safety guidelines. Dose‐response models are employed in quantitative microbial risk assessments (QMRAs) to develop such guidelines; however, pregnant hosts are under‐considered when developing dose‐response models, and the few published dose‐response models for pregnant hosts do not consider the biological plausibility of the model. Therefore, this study aims to develop biologically plausible dose‐response models for pregnant hosts using infection and stillbirth as health outcomes. The inclusion criteria of dose‐response data for this analysis required reporting of (i) quantified dose, (ii) Listeria monocytogenes as the target pathogen, (iii) infection and/or stillbirth as endpoints, and (iv) specificity to pregnant hosts. The datasets were fit individually and as pooled sets to exponential and beta‐Poisson dose‐response models using Maximum Likelihood Estimation (MLE). To establish uncertainty around the MLE estimates, 10,000 bootstrapped iterations were fit to the models. Recommended dose‐response models for endpoints of stillbirth and infection were generated. The recommended model for infection was based on pooled guinea pig and gerbil maternal infection data fit to the beta‐Poisson model, with parameters of: α = 0.0843 and N50
 = 1.01 × 108. The recommended stillbirth dose‐response model utilized a pooled guinea pig fetal brain infection and stillbirth dataset fit to the beta‐Poisson model parameters of α = 0.575 and N50
 = 1.31 × 107. Integrating such models into QMRAs for pregnant hosts supports a more health‐protective assessment and decision‐making process by separating sensitive populations from the general public when possible.