Abstract
The New York State Department of Environmental Conservation promulgates ambient water quality standards to protect sources of potable water from contamination by toxic chemicals and other substances. Ambient water quality standards are a state program with U.S. EPA oversight, including a federal Clean Water Act requirement for “triennial review.” New York's standards are derived according to procedures in state regulation and in conjunction with the New York State Department of Health. Because standards are set at levels much below those that demonstrate effects in laboratory studies, high-to-low dose extrapolations are required. The procedures address both carcinogenic and noncarcinogenic effects. Existing regulations essentially require a linear high-to-low dose extrapolation for carcinogenic effects of a chemical (i.e., there is a finite risk at all doses above zero dose). The regulations also require a nonlinear high-to-low dose extrapolation for the noncarcinogenic effects (uncertainty factor approach) of the chemical (i.e., once below the threshold for the effect, the risk at all doses above zero is zero). New York's ongoing triennial review is addressing both standards and standard-setting procedures. Proposed revisions to the procedures, yet to be formally adopted, would allow greater flexibility and use of a nonlinear uncertainty-factor-based approach for carcinogenic effects of chemicals where warranted. The presentation will focus on the expected revisions to the procedures for carcinogenic effects.
INTRODUCTION
New York State Department of Environmental Conservation (NYSDEC) regulations require that “the discharge of sewage, industrial waste or other wastes shall not cause impairment of the best usages of the receiving water as specified by the water classifications at the location of discharge and at other locations that may be affected by such discharge.” The best usage (best use) of several classes of New York State waters is as a source of drinking water. Thus, NYSDEC promulgates (adopts) ambient water quality standards to protect sources of potable water from contamination by toxic chemicals and other substances. These are referred to as Health (Water Source) standards. New York's existing standards and standard-setting procedures can be found in New York's Codes, Rules and Regulations (at 6 NYCRR Parts 700–706). These are available on the NYSDEC web site at http://www.dec.state.ny.us.website/regs/ch10.htm.
Ambient water quality standards are a state program with U.S. Environmental Protection Agency (USEPA) oversight. The water quality standards program provides the basis for the goal of “Clean and Safe Water” by designating the use or uses to be made of each individual water body within a state (i.e., as a drinking water supply, for recreational uses, and for the protection of aquatic life); setting the applicable water quality criteria that are necessary to protect these uses; and protecting existing water quality from being lowered through antidegradation policies and implementation procedures. The functions of water quality standards include establishing the water quality goals (i.e., the designated uses); serving as the regulatory basis for establishing water-quality-based effluent limitations and total maximum daily loads; and providing the programmatic basis for water quality management and implementation programs. Water quality standards also serve as a basis for tracking water quality improvement or degradation.
New York State's ambient water quality standards apply to the water body itself and are intended to protect the best use(s) of the water body, consistent with the federal approach. Separate standards for finished drinking water (i.e., maximum contaminant levels or MCLs) are established by the New York State Department of Health (NYSDOH).
New York State's ambient water quality standards are derived according to procedures in New York State regulation at 6 NYCRR Part 702. These procedures are promulgated through the State's formal rulemaking process. The procedures to derive the Health (Water Source) standards are developed jointly by NYSDEC and NYSDOH. These procedures are carefully designed to derive standards that are protective of the best uses of the State's waters. New York State's ambient water quality standards predate both the USEPA and the Clean Water Act. NYSDEC first promulgated its standard-setting procedures in 1985 for surface waters and extended them to groundwaters in 1991. Minor modification was made to the procedures in 1998. The Clean Water Act requires that states conduct a review every three years of their water quality standards (“triennial review”).
New York State is currently conducting its triennial review, in which both standards and standard-setting procedures are being addressed (see opening footnote). Procedurally, the proposed revisions must yet undergo both NYSDEC Executive and Governor's Office review before being formally proposed in rulemaking. The formal rulemaking process involves public hearing, public comment, and review by the State Environmental Board before a notice of adoption can be issued. Following adoption, the regulations must be submitted to the USEPA along with certification by the New York State Attorney General. Once the USEPA approves the submitted package, it will be in effect for Clean Water Act purposes.
EXISTING STANDARD-SETTING PROCEDURES
Ambient water quality standards are based on comprehensive review of the toxicity information on a chemical, including, in most cases, primary sources. New York State has always evaluated the potential of a substance to cause both cancer and noncancer effects. Thus, New York State's existing standard-setting procedures allow the derivation (if appropriate) of values for both oncogenic (carcinogenic) and nononcogenic effects of a substance. Because standards are set at levels much below those that demonstrate effects in human populations or in laboratory studies of experimental animals, high-to-low dose extrapolations are required. Existing regulations essentially require a linear high-to-low dose extrapolation for oncogenic effects of a chemical (i.e., there is a finite risk at all doses above zero dose). The lifetime excess cancer risk level prescribed in New York State's regulations is a risk management decision and is set at one in one million. The regulations also require a nonlinear high-to-low dose extrapolation for the nononcogenic effects (uncertainty factor approach) of the chemical (i.e., once below the threshold for the effect, the risk at all doses above zero is zero). The most stringent (numerically lower) of the oncogenic, nononcogenic, and the NYSDOH drinking water standard (MCL) is promulgated as the standard.
One of the strengths of New York State's water quality standards program has been NYSDEC's actions over the years to revise its standards, to be either more stringent or less stringent, as new toxicity data dictates. Such a rigorous, science-based standards program leads to greater public understanding and acceptance of revisions in either direction of stringency. To maintain the strongest scientific basis for the program, NYSDEC must now revise not only its standards but also the procedures by which future standards are derived.
New York State's procedures to derive standards based on oncogenic effects have been changed little since 1985; however, recent years have seen significant advances in understanding the modes of actions of carcinogens and in procedures to derive values protective of human health. Revisions contained in the proposal reflect the latest scientific knowledge and are consistent with USEPA's Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human Health (USEPA, 2000) and with USEPA's latest revision to their Guidelines for Carcinogen Risk Assessment (USEPA, 2003).
In 1985, and in 1986 (when the USEPA issued its first set of cancer risk guidelines), it was believed that any level of exposure to any carcinogenic substance carried some level of risk, and that the level of risk was dependent on the amount of exposure. In other words, all carcinogenic effects are linear at low doses (see Appendix). Thus, the initial 1985 NYSDEC standard-setting procedures, which centered on the health-protective linearized multistage model for extrapolating the carcinogenic effects of chemicals from high to low doses, were consistent with the scientific consensus of the time. Since then, the accumulating scientific evidence has convinced many scientists that some carcinogens act via a nonlinear mode of action (see Appendix). In other words, some chemicals may cause cancer by a mode of action that is nonlinear at low doses (see Appendix). The proposed revisions to the standard-setting procedures allow the use of the uncertainty factor approach for extrapolating the carcinogenic effects of such chemicals from high to low doses. However, such an analysis would be permitted only if the weight of evidence was overwhelming in support of a nonlinear mode of action (see step 5, following).
PROPOSED REVISIONS TO PROCEDURES FOR DERIVING AMBIENT WATER QUALITY STANDARDS BASED ON ONCOGENIC (CARCINOGENIC) EFFECTS
This is a six-step process.
Step 1. Determination of Oncogenicity
First, a determination must be made whether the substance causes an oncogenic effect (see Appendix) as defined in New York State regulations. This definition is unchanged from the existing regulations.
Step 2. Selection of Dose-Response Data
If a substance meets this definition, the dose-response data deemed most appropriate for evaluating potential human health risks at environmental exposures shall be selected. Human data of high quality and adequate statistical power are generally preferred over animal data as the basis of a water quality value. For almost all of existing New York State standards, the human data were judged inadequate and animal data were used to derive the oncogenic value. Factors for data selection include route, duration and timing of exposure, species, strain, tumor types and sites, nature and severity of effects, pharmacokinetics, mode of action, study quality, and statistical significance.
Step 3. Determination of Point of Departure
The starting point for the oncogenic value calculation is the point of departure (POD; see Appendix) for the selected dose-response data, which, depending on the quality of the data, could represent more than one animal species, cancer site, or experiment. The POD is generally the LED10, which is the 95% lower confidence limit on the dose (mg/kg/day) associated with 10% excess risk for oncogenic effects adjusted for background risk.
The POD is derived using the following hierarchy. The first choice is to use a validated biologically based dose-response model (BBDRM; see Appendix). Absent such model, the second choice is to use a mathematical model (multistage, probit, logistic, or Weibull) that best describes the dose-response data within the range of observation. The third option would be to use an alternative POD such as a no-observed-effect level (NOEL) or a lowest-observed-effect level (LOEL).
Step 4. Conversion to Human Dose
If the POD is derived from animal data, the human equivalent dose at the POD is then estimated by multiplying the animal-to-human body-weight ratio raised to the 0.25 power by the animal dose, or by using an alternative trans-species conversion method if deemed more appropriate.
Step 5. Extrapolation to Low Dose
The next major step is to extrapolate to low dose (dose at the level of the ambient standard) from the POD. This is done one of several ways. If the POD was estimated from a validated BBDRM (the first choice in step 3), the ambient standard is derived based on the 95% lower confidence limit on the human dose corresponding to an excess lifetime cancer risk of one in one million.
If a validated BBDRM is not available, a mode-of-action analysis is conducted. The results of this analysis will determine whether the extrapolation from the human dose at the POD to a low dose is via linear or nonlinear approach.
The default approach is linear. If there are no data on mode of action, or if the mode-of-action analysis provides either evidence of linearity at low doses or a lack of unequivocal evidence of nonlinearity at low doses, a linear extrapolation will be performed. Specifically, the ambient standard will be based on the 95% lower confidence limit on the human dose corresponding to an excess lifetime cancer risk of one in one million. To calculate this, the human equivalent dose at the POD is multiplied by a factor equal to the risk level of one in one million divided by the risk level at the POD.
However, if the mode-of-action analysis provides both no evidence for linearity at low doses and unequivocal evidence of nonlinearity at low doses, a nonlinear (uncertainty factor) approach is used to estimate the dose at the level of the ambient standard. The human equivalent dose at the POD will be divided by an uncertainty factor of such magnitude to ensure that the human dose at the standard will be “without appreciable risk to the human population, including children.” The factor will generally range from 10 to 3,000. Factors considered in determining its magnitude will include the nature of the dose-response curve and the point of departure; the relative sensitivities of experimental animals and humans; the nature and extent of human variation, including age-specific differences in sensitivity during a lifetime; and data gaps in the toxicological database.
Step 6. Calculation of Standard
An ambient water quality standard based on oncogenic effects is then calculated from the outcome of step 5. If the human dose at the level of the standard was determined from a BBDRM or a linear approach, that human dose is multiplied by a body weight of 70 kg and divided by a water consumption rate of 2 1 per day to yield the standard. (Alternative body weight and water consumption rates can be used where warranted.)
If the human dose at the level of the standard was determined using a nonlinear approach, that human dose is multiplied by a body weight of 70 kg and a relative source contribution of no more than 20%, and divided by a water consumption rate of 21 per day to yield the standard. (Alternative body weight and water consumption rates can be used where warranted.)
SUMMARY
The proposed revisions to New York State's procedures for deriving ambient water quality standards will provide needed flexibility to address carcinogenic substances that work via a nonlinear mode of action. These revisions will provide this flexibility while retaining a health-protective approach.
This proposed change is a movement toward harmonization of risk assessment methods for evaluating the carcinogenic and noncarcinogenic effects of chemicals. The addition of a procedure for deriving a standard based on carcinogenic effects that are nonlinear at low doses allows New York State to derive a cancer-based standard using the same approach used to derive a standard based on the noncancer effects of a chemical (i.e., identification of NOEL, LOEL, or benchmark, and application of one or more uncertainty factor to compensate for uncertainty and variation). However, the uncertainty factor approach is not the default procedure for cancer effects, as it is for noncancer effects. Moreover, the proposed revisions to the regulations require a substantial amount of data and little or no uncertainty regarding the mode of action before the nonlinear, uncertainty-factor-based approach can be used to derive a standard based on cancer effects.
Inherent in both the linear and nonlinear approach is the assumption that the dose-response for carcinogenic effects is monotonic. In other words, the critical effects decrease as dose decreases across all doses above zero (linear) or decrease as doses decreases across all doses above the threshold (nonlinear). NYSDEC and NYSDOH staff are not at this time planning on further harmonization of the two approaches using recommendations suggested by Calabrese and Baldwin (2003), which incorporate hormesis into risk assessment. However, staff will continue to monitor the developments in the field of hormesis so as to fully understand the implications hormesis may or will have in regulatory risk assessment.