Pseudomonas aeruginosaDose Response Experiments

General overview

Pseudomonas aeruginosa is a gram-negative bacterium capable of causing keratitis, bacteremia, and acute and chronic respiratory infections[1]. P. aeruginosa is more commonly associated with nosocomial, ventilator-associated, and community-acquired pneumonia, however it may also cause lung infections in immunocomponent hosts[2]. The bacteria develop biofilms and has been known to colonize moist, humid environments within engineered water systems.

To better understand the risk of infection associated with the inhalation route of exposure, Dean and Mitchell (2020)[3] fit traditional and nontraditional microbial dose response models to animal data evaluating the course of P. aeruginosa pneumonia. Ojielo et al. (2003)[4] inoculated groups of 10 wild-type, specific pathogen-free B6D2F1/J mice with seven graded doses of P. aeruginosa. The endpoint evaluated was death. Notably, Dean and Mitchell (2020)[3] also identified 19 additional point estimates of the dose associated with a death response from a range of studies. These 19 point estimates were used to benchmark the dose response models fit to the Ojielo et al. (2003)[4] data and support the utility of the mulit-hit model for this pathogen and exposure scenario. A more nuanced discussion of the single-hit versus cooperativity theory for P. aeruginosa can be found in the published article: https://doi.org/10.1016/j.mran.2020.100115

Recommended Model

Neither the exponential or beta-Poisson model provided a good fit to the Ojielo et al. (2003)[4] data. The exponential model faciliates a steeper curve than the beta-Poisson, but the Ojileo et al. (2003)[4] data and the additional dose response point estimates identifed in the literature suggest a more extreme slope may be a feature of P. aeruginosa dose response relationships for the inhalation route of exposure. The multi-hit model is represented by the incomplete gamma function, and has two parameters: the probability that pathogens survives to initiate an infection, k, and the number of organisms needed to initiate an infection, k_min. The multi-hit model did provide a good fit to the Ojielo et al. (2003)[4] data, with median estimates of 4.12 x 10-6 and 11 for k and k_min, respectively. 

Although the multi-hit model provided a good fit to the data, as opposed to the traditionally applied microbial dose response models, it is important to acknowledge that there has historcially been more evidence for the biological plausiblity of the single-hit models and the multi-hit model estimates de minimis risk at low doses. Any application of this dose response model should explore the impact of model selection on risk estimates, and in lieu of additional primary datasets, users are encouraged to conduct uncertainty analyses with the exponential and multi-hit models presented, the distributions explored with the hierarchical analysis, and surrogate models for similar pathogens and the inhalation exposure route. 

References