EPA Sets First-Time Inhalation Risk Limits For Widespread Toxin Dioxane


Source: InsideEPA
Related Topics: US Environmental Protection Agency

EPA has released for public comment and external peer review an updated version of its year-old draft assessment of the common contaminant 1,4-dioxane, including in the document new studies used to derive first-time risk limits for inhaling the chemical.

The update is unusual given that Integrated Risk Information System (IRIS) assessments often take years to produce; as a result they are supposed to be reviewed and updated every 10 years. But these updates often fall behind the intended schedule.

EPA does not explain the unusually fast update, other than to say that "New studies regarding the toxicity of 1,4-dioxane through the inhalation route of exposure are available that were not included in the 1,4-dioxane assessment that was posted on the IRIS database in 2010," according to its website.

The chemical has long been a concern in drinking water and at Superfund sites, and in recent years environmentalists and public health groups have questioned its appearance in consumer products like soaps and shampoos. In 2007, the Agency for Toxic Substances & Disease Registry listed the chemical among the top contaminants at waste sites. Data on EPA's Web site says the greatest risk stems from low-level inhalation by workers exposed to the chemical.

One Japanese toxicological study, Kasai (2009), forms the basis of the new risk numbers. From this study, EPA calculates a non-cancer risk number, or reference concentration (RfC), of 3x10-2 milligrams per cubic meter (mg/m3). The RfC is the amount of the chemical that EPA believes can be inhaled daily over a lifetime without adverse effects. EPA staff use the same study as the basis for their cancer risk number, or inhalation unit risk, of 5x10-6 (nanograms/m3)-1. Both numbers represent the first time EPA has produced inhalation risk estimates for 1,4-dioxane.

EPA in 2010 published an IRIS assessment of 1,4-dioxane that included risk estimates that only addressed oral exposures. Industry and other federal agencies, including the Defense Department, NASA and the White House Office of Management and Budget (OMB) questioned EPA's strict estimate of cancer risks by oral exposure. The cancer slope factor was 17 times more potent than EPA's 1990 IRIS assessment.

In public comments on the oral risk numbers -- first released in draft form in 2009 -- the consulting firm Arcadis urged EPA to consider three additional studies, Yamamoto (2000), Kasai (2008) and Kasai (2009). "Data from the following three studies should also be obtained and reviewed. Although the [2009 draft IRIS document] focuses on the oral route of exposure, the chronic and short-term inhalation studies listed below should also be incorporated into the existing risk assessment as they provide additional information that is relevant to the [mode of action (MOA)] determination for the cancer assessment and would increase the science underlying the assessment."

Determining MOA
Determining the MOA is critical in EPA's cancer IRIS assessments, because the agency's guidelines require its risk assessors to use more conservative linear modeling when the MOA, or how the chemical causes cancer, is unknown. Linear modeling is considered more health protective because it assumes that there is no safe level of exposure to the chemical under assessment. The guidelines also require agency assessors to use the linear modeling if the MOA is mutagenic. Only if a chemical's MOA is known, and non-mutagenic, will less-conservative non-linear modeling be conducted.

Arcadis argued that the Kasai (2009) study indicated that 1,4-dioxane had a known, non-mutagenic MOA, and therefore, should have undergone the less-stringent non-linear cancer modeling. The firm was hired by Cam-Or Site Extended Group, Seymour Site Trust and the Dioxane Risk Management Consortium to review the 2009 version of the draft assessment. "In conclusion, the results of the recently published 2-year inhalation study (Kasai et al. 2009 [in press]) clearly indicated a cytotoxic-proliferative MOA for 1,4-dioxane-induced liver and nasal cavity tumors."

Arcadis' July 2009 comments add that among EPA's "significant flaws" was that "Low-dose linearity was used to derive the proposed [oral cancer risk estimate (OSF)]. Strong data are available for a [MOA] corresponding to a threshold [nonlinear] effect. Thus, the proposed OSF ignores the available data that indicate that low levels of exposure to 1,4-dioxane would not pose a cancer risk."

Now, other agencies -- DoD and OMB -- are questioning EPA's decision to use linear modeling to estimate the cancer risk for 1,4-dioxane by inhalation exposure as well. In May comments, DoD staff argue that "DoD has found that there is sufficient information to support a nonlinear extrapolation for the carcinogenic potency of 1,4-dioxane.

And in June 16 comments, OMB notes that "it appears that data exist to support a non-linear mode of action as the Kasai studies show accumulation related to saturation at high doses."

Insufficient Data
EPA, however, responds in a Sept. 15 response document that there is insufficient data available to indicate what the MOA of 1,4-dioxane is, and so used linear cancer risk modeling.

"When EPA evaluates whether the available data provide significant biological support for a mode of action for cancer the goal is to identify key events, and to have reasonable confidence in the sequence of events and how they relate to the development of tumors including information on the shape of the dose-response curve at low doses. It is EPA's judgment that there are insufficient data to establish the shape of the dose response curve in the exposure-response curve at low doses based on the mode of action data for cancer effects following exposure to 1,4-dioxane, for both oral and inhalation routes of exposure; thus a default linear extrapolation was used."

In contrast to DOD and OMB, the National Institute of Environmental Health Sciences (NIEHS) and the Center for Disease Control & Prevention praised the draft IRIS assessment. NIEHS, in particular, indicated EPA had performed an "outstanding job." In its June 3 comments, NIEHS recommended EPA delete two tables describing potential MOAs for 1,4-dioxane "since there is no data available for any of the mode-of-action assumptions."

But OMB and NASA also questioned the new proposed RfC, with OMB arguing that it is close to levels that naturally exist in the environment, a frequent concern of industry and some federal agencies. "It seems that EPA is proposing an RfC which is in the range of background level . . . Rough calculations tell us that this is equivalent to about 4x10-3 mg/m3, which is less than an order of magnitude away from the proposed RfC. Considering the closeness of the values, and what may be known about ranges of background exposures, it would be helpful to ensure that the RfC is plausible and that the incidences of nasal lesions expected can be predicted by current exposures. In particular, we recognize that rats are obligate nose breathers while humans are not. It is not clear how EPA has taken this into account when considering the relevance of the RfC to humans. We note that EPA applies an UF of 3x for interspecies comparison but this implies that humans would be more sensitive, not less sensitive to a similar dose."

And NASA questions whether there is sufficient information to produce an RfC. "The lack of studies (only four are mentioned) for the development of an RfC raises significant concerns that the updated draft's proposed RfC is premature and not supported by scientific literature," according to NASA's June 13 comments. "NASA requests EPA re-consider issuing a draft RfC, based on such limited evidence."

EPA is extending its public comment period on the new 1,4-dioxane inhalation risk estimates to November 15. -- Maria Hegstad ( )

Upcoming Events