Regulatory Policy Implications of 1,4-Dioxane - Environmental Investigation and Remediation - page 463

10.3 UPDATING THE TOXICOLOGICAL BASIS
FOR 1,4-DIOXANE REGULATION
As detailed in Chapters 5 and 6, the nonlinear dose-response seen when 1,4-dioxane is adminis-
tered to laboratory rats raises questions about the legitimacy of the generic assumption of linear low
dose extrapolation used to determine the drinking water dose that may induce tumors in humans.
A 1996 internal memorandum by the California DTSC recommended PRGs at California cleanup
sites and summarized the nonlinear dose-response conundrum succinctly (DTSC, 1996b). * The
memo explained that because 1,4-dioxane tumorgenicity in animals occurs only at intoxicating
concentrations (10,000 ppm or greater when administered orally), it is thought that the pharmacoki-
netic proi le in animals and humans may explain the disparity in interspecies cancer outcomes. †
Rodents are less efi cient at metabolizing 1,4-dioxane to -hydroxyethoxyacetic acid than humans.

When 1,4-dioxane is supplied to rodents at concentrations that overwhelm their ability to metabo-
lize it, 1,4-dioxane accumulates, causing intoxication followed by systemic cell injury and carcino-
genesis. The tissue death and proliferation mechanism in animals was not accounted for when the
USEPA's CSF for 1,4-dioxane was extrapolated to humans to develop PRGs. The PBPK model by
Reitz et al. (1990) accounts for this mechanism and applies a relationship between the metabolism
of 1,4-dioxane in rats and humans (DTSC, 1996b). In a presentation to the Groundwater Resources
of California's 2001 Symposium on Recalcitrant and Emerging Contaminants , DTSC staff toxi-
cologist Calvin Willhite proi led the 1,4-dioxane nonlinear dose-response issue and USEPA's CSF
extrapolation and famously declared, “These numbers are broken!”
How can environmental regulatory staff effectively manage risks from 1,4-dioxane in the face of
genuine uncertainty regarding the mechanisms and degree of toxicity and possible carcinogenicity?
What improvements could be made to increase coni dence in the scientii c basis for establishing
standards to protect drinking water consumers and other populations exposed to 1,4-dioxane? The
art and science of PBPK modeling has advanced considerably since the Reitz et al. (1990) paper.
Sweeney et al. (2008) made rei nements to estimates and in vitro measurements of intercompart-
mental partition coefi cients. More reliable and independently validated PBPK models will go far to
improve the consensus on a scientii c basis for regulating 1,4-dioxane.
The studies i nding 1,4-dioxane-induced nasal carcinomas upon which the IARC bases its Class
2B “probable human carcinogen” classii cation of 1,4-dioxane have been questioned as unreliable
because of the way in which rats ingested drinking water containing 1,4-dioxane from sipper tubes
(Stickney et al., 2003). Nasal carcinomas observed in rats may have been caused because their nasal
turbinates were probably splashed repeatedly with aspirated water containing 5000-10,000 mg/L
of 1,4-dioxane (Reitz et al., 1990). Rats have a more convoluted nasal turbinate system than humans,
resulting in greater deposition in the upper respiratory tract. The location of the nasal tumors within
the rat nasal turbinate also suggests that water containing 1,4-dioxane entered the nasal cavity dur-
ing drinking (Stickney et al., 2003; see also Chapter 5). Therefore, an additional opportunity to
improve the toxicological basis for regulating 1,4-dioxane in drinking water may be to repeat the
1,4-dioxane rat and mice studies of Argus et al. (1965, 1973), Hoch-Ligeti and Argus (1970), Hoch-
Ligeti et al. (1970) (see Chapter 5), using drinking water administration methods that prevent the
aspiration of 1,4-dioxane-laden drinking water into the rodent nasal cavity. Repeating 1,4-dioxane
carcinogenicity testing could also be important because rodent carcinogenicity assays in general
have been shown to be much less reproducible than might be expected. In a survey estimating the
reliability of 121 replicate rodent carcinogenicity assays from the two parts of the 1997 Carcinogenic
Potency Database (NCI/National Toxicology Program and Literature), researchers found a

β

* The DTSC's analysis of 1,4-dioxane toxicity and appropriate remediation goals was not included in Chapter 6, which
focuses on the drinking water standard-setting process.
† The paucity of human data makes it difi cult to characterize an interspecies disparity in cancer outcomes; this statement
is more likely an opinion than a conclusion based on a survey of available data.