OEHHA Responses to Public Comments...Cont.

Start | Back | Next

Citizens for a Better Environment
Up | Down
The comment contains two specific concerns:

Comment: Add MTBE to the list of chemicals for evaluation since it is widely used in refineries and for reformulated fuels. The chemical was banned in Alaska because of associated acute health problems.

Response: The health impacts of MTBE have generated concern relatively recently. OEHHA agrees with the comment that acute exposure values protective of public health from MTBE should be developed. Although it is not possible for an addition to the present document, we are working with the Air Resources Board to prioritize substances for future evaluation. Data are not available for OEHHA to develop an acute REL for MTBE at this time.

Comment: Add all chemicals which have been released by Bay Area refineries and chemical plants during the chemical accidents of the last 10 years. Petrochemical facilities have had a poor safety record in recent years, and have been a major source of community acute exposure to chemicals.

Response: The chemicals on the current list include those identified as chemicals of concern by the California Air Pollution Control Officers Association (CAPCOA) in 1993 and many additional compounds that are released in large quantities by facilities in the State. The focus of the document is on predictable controlled releases, rather than on spills or accidental releases. There are still many chemicals that require toxicological evaluation and subsequent health value development, and OEHHA is committed to completing this work as soon as possible. OEHHA will continue to add chemicals to the list based on the needs of local agencies and risk managers, in addition to those chemicals released in high volume.

City Of Los Angeles, Bureau of Sanitation
Up | Down
General Comments
Deterministic versus Probabilistic Models

Comment: Risk assessment criteria and procedure development are under critical review by many organizations at this time. The major criticism has been that deterministic assessments are too subjective and limiting in their scope and do not objectively consider the possibility of deviations from fixed values. The Bureau agrees that deterministic models are too limiting and do not accurately measure potential acute health risk, therefore we strongly recommend that the probabilistic health risk exposure model be used.

We believe that the use of models must be accompanied by a recognition of their inherent limitations and a knowledge of the uncertainties associated with model input and outputs. Models should not be viewed as "black boxes" that yield a single absolute value, but rather must be accompanied by good judgment and knowledge of their role as part of a suite of tools available for policy analysis and decision-making. New and improved tools are needed to assist policy-makers with decisions so that equitable and effective emissions management practices can be applied.

The discussion and analysis presented in Part I and in the Technical Support Document clearly demonstrate that risk/exposure assessment must incorporate a large amount of uncertainty into the evaluation. For example, the uncertainty and modifying factors utilized to determine effective models add additional levels of subjectivity to the values used in deterministic models. As a result of this uncertainty, we question the value of comparing a point value for the estimated exposure/risk to a point value for acceptable exposure/risk as done in the "bright line test". Probabilistic modeling method uses all of the information available about the variability and uncertainty inherent in the assessment, while deterministic assessments discard most of this type of information. Probability distributions thus present the range of exposures or risks not available using deterministic methods.

Response: OEHHA agrees with the comment that probabilistic models are valuable tools when it is possible to use them. The use of distributions to characterize ranges of risk requires adequate data and is therefore more appropriate for assessment of exposure than for determination of toxicological thresholds. For example, for nearly all chemicals, the precise toxicological response distributions of individual thresholds are simply not known, even in laboratory animals. Furthermore, although some specific areas of interspecies differences (metabolism, toxicodynamics, etc.) may be addressed to varying degrees in the literature, it is extremely rare to have enough data to adequately characterize interspecies uncertainty. Uncertainty factors, though somewhat subjective, are used to account for real areas of uncertainty and variability. A considerable body of literature shows that many of the traditional uncertainty factors accurately account for certain areas of variability. For example, the report submitted by Monsanto Company using data from the World Health Organization and other sources (see attached) shows that an uncertainty factor of 10 adequately accounts for interspecies differences in most cases. These factors are therefore not merely artifacts of risk management decisions. Unless considerable data exist to support a range for each chemical, the uncertainty factor method should be used. Recent analyses presented at the Society of Toxicology meetings in 1997 reported that probabilistic analyses are generally consistent with the uncertainty factor approach. Probabilistic methodologies are part of the Hot Spots Technical Support Document for Exposure Assessment and Stochastic Analysis, which is a separate document under review.

Specific Comments

1) Interpretation and Definition of an Acute HI

Comment: A bright line test is not necessarily the best way to present risk/exposure information for another reason. A well defined threshold value should generate a factor that clearly alerts the impacted population that concern is warranted. If the public does not perceive any hazard then a credibility gap could form, and the hazard index may be widely disregarded.

Response: The reference exposure level (REL) is the level at or below which adverse health impacts are not expected for the majority of the population. When exposure concentrations exceed the REL, there exists a gray area where one cannot predict with certainty when adverse impacts will be manifested until a concentration that produces frank or severe effects is reached. RELs are based on the best available scientific data for each compound. The data simply are non-existent to characterize the distributions of individual toxicologic thresholds for each chemical and endpoint. Uncertainty factors are used in many cases since detailed dose-response data in sensitive individuals do not exist. As the comment has pointed out, the onset of toxicant-induced symptoms in the population is probabilistic, and follows a distribution. The purpose of the RELs is to account for those responses at the sensitive tail-end of that distribution. Generating RELs at which large numbers of people reported adverse health impacts would be inappropriate.

The bright line test is a risk management decision and, while it has its limitations, it currently is one way a risk manager can feel certain about protecting public health. Whether the public perceives a hazard is an issue unrelated to setting science-based health-protective RELs.

Comment: The proposed system to determine RELs has been highly diluted with modifying and uncertainty factors such that the hazard indices generated under this system generally do not represent any real acute health threat to the exposed population. The guidelines for the evaluation of the acute impacts, while based on the sound toxicological data, are adjusted to reflect the risk to only the most sensitive portions of the population. Ten of the 53 compounds listed in the document, with reference exposure levels, are flagged as compounds with Level II effects. Level II compounds are those which may cause exposed individuals to seek assistance after an exposure period of 1 hour. One of these ten compounds is perchloroethylene used in the dry cleaning industry. The REL given is more than an order of magnitude less than either the odor threshold and any industrial worker standard.

Customers often observe workers in dry cleaning facilities exposed daily to this "not unpleasant" smelling chemical at very perceptible concentrations and consumers will often smell the perchloroethylene off-gas from freshly cleaned clothing for most of [the] first day after cleaning. By definition, both are thus exposed to an acutely hazardous concentration. This situation establishes a very large credibility gap with the public regarding the hazard associated with (and other) compounds. This is not meant to indicate that there is not a real health threat from exposure to this compound, but we think that exposure at this very low REL should not be used to determine an acute HI. The public may be confused and may be skeptical as to the interpretation of an acute hazard.

Response: For purposes of clarification, OEHHA has defined acute reference exposure levels by duration of exposure (i.e., short-term, 1-hour exposures). The term acute may be used for different purposes in a medical context, and may sometimes imply a severity of the effect (e.g., the term "acute" can describe a short, sharp course). In the TSD, severity of effect is treated separately. It appears that this discrepancy in definitions may have created some confusion in the comment in the above statement.

It is true that the acute REL for perchloroethylene is lower than the occupational standards. However, there are two important issues that determine the REL for this compound. First, the adverse CNS effects used as the basis for the acute REL are from human data, with a specific, clinically significant change in a neurological test. There is, therefore, little uncertainty in the validity of the toxicological endpoint. The principal uncertainty lies in the small sample size of 3; thus an uncertainty factor of 10 was applied to account for sensitive individuals. Since no NOAEL was reported, an uncertainty factor of 10 was used to estimate a NOAEL from the LOAEL. Second, the occupational standard (50 ppm TLV) is based on prevention of irritation, not on neurological effects and includes only a 2-fold margin of safety. The ACGIH short-term exposure limit for perchloroethylene of 200 ppm is based on minimizing anesthetic effects, yet the LOAEL for CNS effects (impaired balance) in Stewart et al. (1970) was 100 ppm for 3 hours. These CNS effects are serious enough to potentially interfere with ability to escape exposure. Furthermore, the odor threshold for perchloroethylene is relatively high (2 - 71 ppm; AIHA, 1989), and is in the same range as the observed CNS effects. The fact that consumers and workers are exposed above the odor threshold does not mean that perchloroethylene is non-toxic at those concentrations.

2) Determination of Interim RELs

Comment: It is unreasonable to place the burden of determining interim RELs on the risk assessor for those compounds that OEHHA’s toxicologists do not have enough data. OEHHA has the expertise and resources to develop RELs, however, most risk assessors just do not have the background in toxicology that OEHHA staff have. This action will completely subvert the acute HI assessment process as the focus will shift from the facilities’ HI to the subjectivity and potentially self-serving REL. We recommend compounds with RELs developed by OEHHA be considered in the health risk assessment process.

Response: OEHHA recognizes the need for consistency and its unique responsibility in this area and has withdrawn this suggestion from the guideline.

3) Inclusion of Background Concentrations of Criteria Pollutants

Comment: The requirement that the background concentration of criteria pollutants should be included in the acute HI evaluation when the HI is greater than 0.5 is tantamount to establishing an acute HI notification level of 0.5 for almost all of the sources regulated by the SCAQMD. The risk/exposure assessment should reflect only the incremental increase as a result of the facility emissions.

Response: The acute REL document no longer includes the procedure of adding the criteria air pollutant effects to those of the non-criteria air pollutants.

4) Guidelines for Evaluation of Acute Risk

Comment: The guidelines for conducting the risk characterization (Section VII.D) are vague and do not comprehensively explain the procedures for evaluation [of] the acute risk. However, experience indicates that a precise risk characterization will represent a substantial amount of work. A number of computer dispersion model runs would be necessary in order to first locate the maximum exposed individual (MEI) resident, worker, and sensitive receptor to the total HI and then to determine the HI at the MEIs by organ or system for both level I and level II compounds. There are a number of approaches a risk assessor could take to complete this evaluation. OEHHA needs to develop detailed guidelines to assure the required data is presented in the risk assessment and that the approaches used by the risk assessors are consistent otherwise the results will not be comparable. These detailed guidelines need to be completed before Part - I guidelines are approved.

Response: OEHHA’s draft exposure guidelines, titled "Exposure Assessment and Stochastic Analysis", was released for public comment in December, 1996. The dispersion modeling runs required to analyze the one-hour maximum are currently used for Hot Spots risk assessments.

Reference
Stewart RD, Baretta ED, Dodd HC, Torkelson TR. Experimental human exposure to tetrachloroethylene. Arch Environ Health 1970;20:224-229.

The Clorox Company
Up | Down

Comment: Requiring action at a hazard index level of less than 1.0 in the risk assessment without safeguards to ensure that risk management identifies a hazard index action level of 1.0 or greater is poor policy.

Response: The acute REL document no longer suggests a specific HI.

Comment: Background pollutant concentrations should be compared to facility emissions, not added to them.

Response: The acute REL document no longer includes the previous procedure of adding the criteria air pollutant effects to those of the non-criteria air pollutants.

TOC | Back | Next

You will need the free program Adobe Acrobat Reader to view or print PDF files

For help with this web site please contact the Web Mistress