Value-of-Information Approach to Motivate Uncertainty Factors with Mechanistic Data: Chlorine Human Health Inhalation Risk Case Study

Objective/Intended Use

The objective of this research is to develop an homologous data base of endpoints and dosimetry models across rodent test species and in human subjects that will then be used to provide a human health risk assessment of inhaled chlorine and to create a formal statistical approach (value-of-information analysis) to informing the magnitude of the interspecies and intrahuman uncertainty factors (UF) when mechanistic data are available. These uncertainty factors are currently derived empirically and do not take into account mechanistic data on pharmacokinetics or pharmacodynamics. Chlorine is an ideal case study for the project since significant health effects data already exist in rats, mice, monkeys and humans. The statistical approach will generalize to all types of gases and be the first framework for departing from default UF within a formal statistical framework.

Abstract

Data on the health effects of inhaled chlorine gas are available for rats, mice, monkeys, and humans. This makes it an ideal candidate to evaluate the value of quantitative interspecies extrapolation modeling to derive human health risk estimates. This project includes both experimental and computational efforts, aimed at obtaining and integrating mechanistic data to ultimately develop a biologically based non-cancer RA for chlorine. Further, the data base will then be used to create a value-of-information statistical strategy that will provide a formal approach to inform the magnitude of uncertainty factors for interspecies and intra-human extrapolation. The major determinants of airflow delivery, target tissue interactions, or outcome measure sensitivity will be evaluated across species with homologous data to the degree feasible and used to construct the framework for statistical analysis of the value these types of mechanistic data provide to increasing the accuracy and confidence in interspecies extrapolation.

The first phase of the project is to develop a more comprehensive data base on chlorine uptake to support development of computational fluid dynamics (CFD) models of airflow in all of the test species for chlorine. This will include modification of existing CFD models in the rat, human and monkey to be chlorine-specific and generation of the first CFD model for airflow in mice. Experimental studies measuring the uptake of chlorine in humans have been concluded and published by Pennsylvania State University. Results of preliminary uptake studies in rats were consistent with the human data but need to be verified and extended to multiple concentrations and flow rates. Development of a CFD model for simulation of airflow in the mouse upper respiratory tract (URT) has been initiated. This mouse CFD model will be useful to dosimetry modeling of many chemicals and to genomics work in addition to filling the data gap for the chlorine assessment. Mapping of lesion locations from prior studies at CIIT in rodents for correlation with CFD modeling is also underway. Other protocols under development include one to determine the species and sex differences in the RD50 for the rodents and another to measure their ventilatory, neurogenic, and hypothermic responses. Tissue measurements of chloride anion and reaction products in nasal and airway epithelium will be used to extend the CFD models into the tissue compartment for evaluation of additional dose metrics. Gene arrays for oxidative damage may also be incorporated. Another will perform pulmonary function tests in humans that correspond to measures made in the laboratory animals. An expert workshop to interpret, rank, and weight the severity of upper respiratory tract histopathological lesions observed in exposures to rodents for human health risk assessment is planned for the Spring of 2002. The outcome of this workshop is expected to provide a weighting scheme that will allow us to compute an overall response metric that reflects the severity and incidence of all lesions observed, rather than focusing on one particular type. This weighting will be used to inform a strategy for the integration of diverse response data across species in the second phase of the project.

The second phase will utilize the available laboratory animal versus human data in a formal value-of-information strategy to motivate the intrahuman and interspecies uncertainty factors (UFs). These UFs are currently derived empirically and do not take into account mechanistic data on pharmacokinetics or pharmacodynamics. Chlorine is an ideal case study for the project since significant health effects data already exist in rats, mice, monkeys and humans. The approach is envisioned to result in a probabilistic expression of the confidence in a given description of the determinants of mode of action based on the mechanistic pharmacokinetic (e.g., airflow and tissue reaction rates) and pharmacoddynamic (cellular responses and lesion formation) among the species. The statistical approach will generalize to all types of gases and be the first framework for departing from default UF within a formal statistical framewor

Project Status

This project is a collaborative effort between NCEA with NHEERL (ETD, HSD and ECD) and the CIIT Centers for Health Research. The laboratory animal testing will be performed at CIIT Centers for Health Research via collaborative protocols with the Experimental Toxicology Division and the Environmental Carcinogenesis Divsion. The human testing will be performed in the Human Studies Division of NHEERL or at the Pennsylvania Statue University. Experimental studies measuring the uptake of chlorine in humans

(Pennsylvania Statue University) have been completed and published. Results of preliminary uptake studies in rats were consistent with the human data but additonal verification and extension to multiple concentrations and flow rates is underway. Mapping of lesion locations from prior studies at CIIT in rodents for correlation with CFD modeling is also underway. Development of a CFD model for simulation of airflow in the mouse upper respiratory tract (URT) has been initiated and is expected to be completed in fall 2003. Development of an experimental protocol to measure the ventilatory, neurogenic, and hypothermic response in rodents is being developed to evaluate the species and sex differences in observed effects. The human studies protocol for pulmonary function testing is also under development. The health assessment and statistical framework are expected after 2005.

Project Start Date

10/01/0099