Application of Meta-Analysis to Derivation of Points-of-Departure for Short-Term Inhalation Exposure Levels of Hazardous Chemicals

Abstract

BACKGROUND AND AIM: For short-term chemical inhalation exposures to hazardous chemicals, the incidence of a health effect usually conforms to a generalized linear model with a bivariate probit link function dependent on inhalant concentration ({C}) and the duration of exposure ({t}). The National Academy’s Acute Exposure Guideline Levels (AEGLs) Committee relies on this model when deriving AEGLs. Threshold concentrations at AEGL durations are established by the toxic load equation, {C}^{n} x {t} = {constant}, which Toxic Load Exponent {n} (TLE or {n}-value) directly follows from the probit model. When multiple probit datasets are available, often they cannot be directly pooled together because of disparities in experimental design. METHODS: Recently, we reported both a meta-analytical framework for multiple studies and an application of the new methodology to dimethyl sulfide. The new methodology allows accurate TLE estimation even if multiple datasets are heterogenous. The utilized techniques include categorical regression, common-effect and random-effects models. The proposed framework was applied to multiple-study datasets from AEGL technical support documents. RESULTS:Using the framework, both new TLEs and confidence intervals on them were derived. The recalculated TLEs and points-of-departure include ammonia ({n} = 2.13, 95% CI: 1.98–2.27), allyl alcohol ({n} = 0.95, 95% CI: 0.76–1.15), 1,1,1-trichloroethane ({n} = 3.46, 95% CI: 2.85–4.07 ), carbon tetrachloride ({n} = 2.51, 95% CI: 2.27–2.76), and oxygen difluoride ({n} = 1.29, 95% CI: 0.26–2.32). CONCLUSIONS:The new TLEs and their confidence intervals were applied in calculations of short-term inhalation points of departure at five reference AEGL durations. Disclaimer: The findings and conclusions in this presentation have not been formally disseminated by the Centers for Disease Control and Prevention/the Agency for Toxic Substances and Disease Registry and should not be construed to represent any agency determination or policy. KEYWORDS: Risk Assessment, Meta-analysis, Toxic Load Exponent (TLE), n-value, Probit, Categorical Regression