ADVANCES IN THE RISK ASSESSMENT GUIDELINES PRESENTED AT THE SESSION ADVANCES IN ASSESSMENT, CCTMJNICATiaN AND MANAGING RISKS 34TH ANNIVERSARY TECHNICAL CONFERENCE MID-ATLANTIC STATES SECTION AIR POLLUTION CONTROL ASSOCIATION OCTOBER 21, 1988 ALAN M. EHRLICH, Ph.D. NOTICE: The views expressed in this paper are those of the author and do not necessarily reflect the views or policies of the U.S. Environmental Protection Agency. OFFICE OF HEALTH AND ENVIRONMENTAL ASSESSMENT OFFICE OF RESEARCH AND DEVELOPMENT U.S. ENVTJOWENTAL PROTECTION AGENCY Washington, D.C. 20460 ------- ABSEWCT In 1986, the Environmental Protection Agency issued five guidelines for conducting risJc assessments. They describe the process used by EFA to evaluate the health effects of exposure to environmental toxicants. When those guidelines were issued, EPA stated its intent to continue the development of new guidelines and revision of existing guidelines. This paper discusses the seven guideline activities under way as of October 1988: female reproductive risk, male reproductive risk, exposure measurements, developmental toxicity, carcinogenicity, non-cancer health effects, and neurotoxicity. ------- ADVANCES IN THE RISK ASSESSMENT GUIDELINES About five yc=ars ago, EPA convened what it called the Toxics Integration Task Force to investigate the quality and consistency of EFA's risk assessments. The task force made two key recoimiendations — (1) to establish mechanisms to better coordinate the risk assessment process, and (2) to develop risk assessment guidelines. The Risk Assessment Forum and the Risk Assessment Council were created to meet the first recommendation. Readers of this current paper may have heard Peter Preuss discuss the first round of guidelines at the 1986 APCA meeting, or may have read the corresponding paper in JAPCA (Preuss and Ehrlich, 1987). This current paper will provide what is, in essence, a progess report on the second round of guidelines. In order to put these most recent developments in context, however, it will first be helpful to review briefly the history of the last five years. The terminology currently used within the Federal government was set out in 1983 in the National Academy of Science's report, Risk Assessment in the Federal Government: Managing the Process (National Research Council, 1983). In that report, the NAS recommended separating risk assessment from risk management; defined risk assessment in terms of its four components of hazard identification, dose-response assessment, exposure assessment, and risk characterization; and recommended that agencies develop "inference guidelines" to describe how risk assessments are done. EPA uses the term "risk assessment guidelines" for its inference guidelines. We can look at the process as a pair of bridges spanning the gaps between available information, risk characterization, and regulatory decision-making. Risk assessment, and the risk assessment guidelines fill the first gap with a bridge constructed of thinking about the data, organizing the data, conducting the assessment, and presenting the findings. Then the second gap is bridged by social, legal, economic, and political considerations. EPA's guidelines set fort' internal Agency procedures that: o Promote consistency across EPA risk assessments by using common approaches to risk assessment o Promote the quality of the science underlying EPA risk assessments by thoughtful development of a consensus approach o Clarify EPA's approach to risk assessment by informing the public and the regulated community about the process by which the EPA will evaluate scientific information Since the guidelines are not regulations, risk assessors outside of EPA are not obligated to use them. The guidelines are also not fixed, rigid rules within the Agency; rather, they are standards to guide EPA in conducting its own risk assessments and in evaluating risk assessments done outside the Agency. That is, they are flexible to encourage the use of all relevant data and the appropriate scientific methods and judgments. The guidelines can, however, influence the risk assessments that underlie EPA regulations by: o Making EPA's risk assessments more consistent and of higher technical quality ------- - 2 - o Familiarizing risk assessors throughout the country with EFA's approach o Making it possible for scientists to plan their experiments to collect the information that EPA scientists would like to have available when conducting a risk assessment In developing the guidelines, EFA has used a process which is typical both of consensus standard development procedures and of administrative procedures. Draft guidelines are first developed by work groups of Agency scientists representing all Agency programs. They are then submitted to invited expert review and internal Agency review that includes the Forum and the Council, and published for comment both by the public at large and the Agency's Science Advisory Board (SAB); the two aspects of the public cement process can be accomplished either in sequence or approxijnately concurrent steps. The guidelines are then revised, submitted for final Agency review and review where appropriate by other Federal agencies, and finally issued. The first five guidelines were proposed in late 1984-early 1985 and issued in September 1986. They are: Guidelines for Carcinogen Risk Assessment, Guidelines for Mutagenicity Risk Assessment, Guidelines for the Health Risk Assessment of Chemical Mixtures, Guidelines for the Health Assessment of Suspect Developmental Toxicants, and Guidelines for Estimating Exposures. For further details about the 1986 guidelines, the reader is referred either to the guidelines (U.S. EPA, 1986a- e; 1987) or to papers describing them (e.g., Preuss and Ehrlich, 1987). When EFA began this Agency-wide guidelines development program, we recognized several important factors: o First, that for any one issue, there was not likely to be a single correct answer, but rather a range of defensible positions the Agency could take o Second, that consensus could be reached on some issues in a relatively short time period, but other issues would take significantly longer, and o Third, that the guidelines would be evolving documents, needing updates as the science base relating to risk assessment leads to new understanding of the effects of toxic substances, or to a reduction of the uncertainty inherent in the risk assessment process. Thus, the five guidelines published in 1986 were planned as the first step in an on-going process of developing new guidelines and amending current ones. That brings us up to today. At this time, seven guideline projects are underway: female reproductive risk, male reproductive risk, exposure measurement, possible developmental toxicity amendments or revisions, neurotoxicity, non-cancer health effects, and possible carcinogenicity amendments. They are listed approximately in order of planned completion through 1990, with the first two already proposed, and the others in various stages of development. Each will be discussed in the following paragraphs, though not necessarily in this implied chronological order. ------- - 3 - The Proposed Guidelines for Assessing Female Reproductive Risk and for Assessing Male Reproductive Risk were published on June 30, 1988 (U.S. EPA, I988a, b). They have been reviewed by the SAB concurrently with the public review. At the public discussion with the SAB on July 14, the review conmittee expressed its concurrence with the general thrust of the guidelines, and the concept of effect- specific guidelines. They made specific suggestions in the areas of the weight- of-evidence categorization and the dose-response discussion. The Agency is currently waiting for the SAB's formal comnent letter. In addition, the public comment period was extended to September 30, 1988 (U.S. EFA, I988c). After all of the comments have been received, the proposed guidelines will be revised as appropriate, and then published as final documents. These risk assessment guidelines are not descriptions of testing protocols as set out in other Agency publications. Rather, the guidelines supplement the testing protocols by giving guidance on the organization and evaluation of the data after test data have been collected. That is, they propose formats for discussing the issues related to reproductive risk based on the four components of risk assessment: hazard identification (the most detailed discussion), dose- response assessment, exposure assessment, and risk characterization. The Hazard Identification section is "the most detailed in these and other health-effect specific guidelines. In the female reproductive risk guidelines, the Hazard Identification section should include a discussion of: o Laboratory Animal Studies; indices, end points, and their interpretation, i.e., protocol considerations such as + use of generational tests and the relatively new continuous breeding tests, + setting appropriate dose ranges, + use of controls, + selection of test animals, + lengths of exposure, and + listing of pathological parameters indices of reproductive toxic ity such as + fertility, + litter size, + alterations during pregnancy or at birth, and + offspring survival ------- - 4 - consideration of other end points, such as alterations of the female reproductive system and reproductive organ toxicity o Human Studies where epidemiological or clinical data exist, especially where precise measures of exposure are available, including epidemiological studies examinations of clusters or case reports human clinical evaluations o FharmacoJcinetics o Comparisons of Molecular Studies o weight-of-Evidence Determinations (No formal scheme has been proposed, but there was a request for Garments on possible development and use of a formal scheme.) These guidelines refer to the current Agency reference dose (RfD) approach for dose-response assessment and to the exposure guidelines for exposure assessment. They also solicited comnents on the applicability of mathematical dose-response models specific to assessing reproductive risk. The fourth step, Risk Characterization is the combination of the first three steps into an estimation of risk, but also including a discussion-of the assumptions, major scientific judgments, and estimates of uncertainty. The proposed male reproductive risk guidelines have a similar format. Again, the Hazard Identification step is the most detailed, and includes consideration of: o Laboratory Testing Protocols, i.e., types of tests, such as + single and multigeneration tests + the relatively new continuous breeding protocols + dominant lethal tests "'••* subchronic toxicity tests * chronic toxicity tests design factors that may influence test results, such as + duration of dosing + length of mating period ------- + number of females mated to each male end points of toxicity, e.g., + routinely evaluated ones such as tody weight, organ weights, pathology, fertility, and pregnancy outcomes + supplemental end points such as sperm evaluation or endocrine evaluations + biochemical markers of toxicity + paternally mediated postnatal outcomes + sexual behavior o Human Studies epidemiologic studies + general design considerations such as power of the study, potential bias in data collection, control of other risk factors—effect modifiers—confounders, and statistical factors + selection of outcomes for study, including semen evaluations and reproductive history studies + community studies and surveillance programs examinations of clusters or case reports o Pharmacokinetic Considerations (which may be a significant issue because of the apparent existence of a blood/testis barrier) o Structure-Activity Relationships o Weight-of-Evidence Determination In the male reproductive risk guidelines, a specific weight-of-evidence categorization and definition of terms has been proposed, as well as sane of the study considerations that go into making that weight-of-evidence determination. The weight-of-evidence categorization was one of the areas of comment by the SAB, and changes, will be considered by the Agency. As proposed, the weight-of-evidence categories are: ------- - 6 - o Known Positive A convincing body of evidence exists that an agent causes an adverse effect on the male reproductive system in humans. o Probable Positive A convincing body of evidence exists that an agent causes an adverse effect on the male reproductive system in nonhuman mammals. o Possible Positive, divided into categories which cause downgrading of the weight of evidence, either A. Evidence frcm human or other mammalian studies show statistically significant adverse effects, but the quality of the studies is questionable; B. Studies with acceptable quality produce inconsistent and conflicting results such that the possibility of adverse effects cannot be discounted; or C. Other data exist from which biologically meaningful adverse effects are plausibly indicated. o Known Negative A convincing body of evidence exists that an agent does not cause an adverse effect on the male reproductive system in humans. o Probable Negative A convincing body of evidence exists that an agent does not cause an adverse effect on the male reproductive system in nonhuman mammals. o Possible Negative Studies with acceptable quality produce no adverse effects, but important aspects of the male reproductive system have not been evaluated. o No Data or Inadequate Data, subdivided into classes A., No data are available; B.- Negative data are available from studies for which the confidence in quality is questionable; or C. Results are available for which the predictive value of the test system or end point has not been established. ------- - 7 - As with the female reproductive risk guidelines, the Dose-Response Assessment step uses the Agency's current RfD approach, the Exposure Assessment step uses the exposure guidelines, and the Risk Characterization step should be a discussion of the estimation of the risk which also includes assumptions, scientific judgments, and an estimation of uncertainties. There are three guidelines that may be amended, or may have major additions: developmental toxicity, exposure measurements, and carcinogenicity. Proposed amendments to the developmental toxicity guidelines are undergoing pre-proposal reviews. The current draft contains a weight-of-evidence categorization, refinements to the RfD discussion, expanded discussions of the relationship between maternal and developmental toxicity and of dose-response modeling techniques, reformatting of sections on human studies and risk characterization as well as updates to many of the other sections. The draft Guidelines for Exposure-Related Measurements are at a similar stage of development—undergoing pre-proposal review. These new guidelines are intended as a companion and supplement to the Guidelines for Estimating Exposures. Those earlier guidelines present the risk assessor with a format or procedural framework for estimating the degree of human contact with a chemical of concern; they recognize the importance of measured data, but recognize that there will be data gaps which can be addressed by use of validated mathematical models. Comments received on the earlier guidelines, especially comments from the SAB, suggested a supplement dealing in more detail with how to make and use exposure assessments. The measurement guidelines are, therefore, intended to assist those who must recommend, conduct, or evaluate an exposure assessment. The draft guidelines contain three major sections: o sources of measurement data for direct measurement, reconstructive approaches (body burden, metabolites, biomarkers) and predictive approaches to exposure assessment o making exposure measurements—including the role of the exposure assessor, sampling plans, uncertainty analysis, quality control, quality assurance and model selection o use of measurements in exposure assessments, including evaluation of uncertainty in'use of measurements; quality assurance and the role of limit of detection values; and the relevance of the various forms of measurement data, including the relationship between measurements and modeling and the combination of measurement data sets from various sources In contrast to the developmental toxicity and exposure measurement projects, the process for amending the carcinogenicity guidelines is just beginning. A recent Federal Register Notice publicized EPA's intent to review the carcinogenicity guidelines and requested that relevant information be submitted by October 11, 1988 (U.S. EPA, 1988d). That notice also states EPA's plans to hold a workshop in early 1989 (date not yet set) and to propose amendments, if any are ------- - 8 - appropriate, by the fall of 1989. The notice sets out several issues that have been designated for consideration, in addition to issues that will be raised as a result of the Federal Register solicitation for information. These designated issues include: o Classification Scheme consideration of options for the weight-of-evidence scheme. Experience with the current scheme indicates some need for change, for instance, either adding more categories (as in European and international chemical industry schemes) or limiting them clarification of terms like "sufficient," "limited," and "inadequate" both in terms of changes in the International Agency for Research on Cancer's (IARC) definitions and in terms of our own experience other classification issues o Decision Logic for Quantitative Risk Estimation developing guidance on when quantification is appropriate for category C chemicals considering additional approaches for characterizing the risk + newer or updated methods for calculating or caning close to the "best" estimates of risk to determine whether any are appropriate for inclusion in the guidelines + fully stating the qualifying considerations that affect the accuracy of risk estimates like the "plausible upper bound" + providing guidance on selection of other extrapolation models, especially those that take into account biological mechanisms and pharmacokinetics expanding guidance on pharmacokinetics and interspecies scaling factors o Other Issues which would not alter basic concepts in the guidelines updating them for current knowledge of mechanisms of carcinogenesis ~ such as site and contact of hormonal carcinogenesis; the role of cellular peroxide formation; the use of promotion studies; data on genotoxicity; information on onoogenes; and considerations of structure-activity relationships, metabolism, and pharmacokinetics additional guidance on combination of benign and malignant tumors, where EPA appears to differ in approach from IARC and the National Toxicology Program ------- - 9 - additional guidance on risk characterization, e.g., types of analyses to be conducted, articulating risk findings such as use of sensitivity analyses, expressing variability in risks, and quantitation when the weight of evidence is low incorporation of Risk Assessment Forum reports Another area of guidelines interest is dose-response estimation for threshold toxicants. This was an area that was of significant interest from the start of the process in 1983-1984. However, guidelines in this area have not been developed because of the difficulty we have had in reaching consensus for such a broad area. This covers many end points and many different target organs that could be considered for any one chemical. In the absence of consensus procedures, each program office at the EFA has historically approached the problem in different ways. Examples include evaluating a specific adverse health effect rather than determining the critical health effect (the adverse health effect occurring at the lowest dose), or assessing risk for less than lifetime exposure rather than determining a lifetime chronic acceptable daily intake (ADI) or reference dose (RfD). In addition, program offices have used different approaches for dealing with uncertainty. Some have estimated a lifetime chronic RfD based on uncertainty factors tied to the available "information, and then established criteria based on that RfD. Some have calculated a margin of exposure (formerly margin of safety) between the highest no-observed-adverse-effect-level (NDML) of the critical effect and the estimated exposure, and then evaluated that margin specifically in terms of the chemical of interest and its expected exposure pattern. Some offices have estimated an appropriate degree of protection on a case-by-case basis, using their best technical and scientific judgment; this is analagous to the process of developing National Ambient Air Quality Standards. Finally, some programs have developed their own quantitative techniques for extrapolating or interpolating across data gaps; few of these have yet gained general acceptance within the Agency.' The guidelines development effort was suspended while an Agency-wide effort was begun to develop consensus RfDs and enter them into the new Integrated Risk Information System (IRIS; U.S. EPA, 1988e). In addition to the specific chemical files, IRES contains a discussion of the approach used for developing these consensus RfDs. IRIS is currently available thru EPA's E-mail, the Public Health Foundation Network, and shortly on TOXNET. It is also updated quarterly for the National Technical Information Service and purchased from them by private data sources. An Agency work group is continuing to consider the development of risk assessment guidelines generically for non-cancer health effects such guidelines. The Agency also has a continuing interest in developing effect-specific guidelines for some non-cancer health effects. The first effect-specific guidelines for threshold effects are the Guidelines for the Health Assessment of Suspect Developmental Toxicants published in 1986 and the Proposed Guidelines for Assessing Female Reproductive Risk and for Assessing Male Reproductive Risk, discussed previously. The next effort is neurotoxicity. A work group has been developing such guidelines. Beyond neurotoxicity, future development of guidelines for immunotoxicity is being considered. Also under consideration for future work are guidelines for more generic issues like ecological risk and pharmacokinetics. These so called future projects are indefinite only in the ------- - 10 - sense of there being no planned date for completion of guidelines. Projects are, in fact, underway in all three areas to develop risk assessment tools and techniques that could become the basis for guidelines in what might be called the third round. When EPA initiated its program of guidelines development almost five years ago with the first five guidelines, the Agency expressed a continued intent to update and revise them, and to add to them as appropriate. Our plate is full for the next few years, and our interest continues beyond this second round. Our intent is to continue to amend or add to the guidelines, as the sophistication of risk assessment increases, as more risJc assessments are performed, and as long as there is a continued need for the assurance of quality and consistency in the Agency's scientific evaluations. ------- REFERENCES National Research Council. (1983) Risk assessment in the federal government: ' managing the process. National Academy Press, Washington, DC. Preuss, P.W. and Ehrlich, A.M. (1987) The Environmental Protection Agency's risk assessment guidelines. J. Air Poll. Control Assoc. 37:784-791 U.S: Environmental Protection Agency. (1986a) Guidelines for carcinogen risk assessment. Federal Register 51:33992-34003. U.S. Environmental Protection Agency. (1986b). Guidelines for mutagenicity risk assessment. Federal Register 51:34006-34012. U.S. Environmental Protection Agency. (1986c). Guidelines for the health risk assessment of chemical mixtures. Federal Register 51:34014-34025. U.S. Environmental Protection Agency. (1986d). Guidelines for the health assessment of suspect developmental toxicants. Federal Register 51:34028-34040. U.S. Environmental Protection Agency. (1986e). Guidelines for estimating exposures. Federal Register 51:34042-34054. U.S. Environmental Protection Agency. (1987, August). Hie risk assessment guidelines of 1986. Office of Health and Environmental Assessment, Washington, D.C. EPA/600/8-87/045. U.S. Environmental Protection Agency.. (I988a, June 30) Proposed guidelines for assessing female reproductive risk. Federal Register 53:24834-24847. U.S. Environmental Protection Agency. (1988b, June 30) Proposed guidelines for assessing male reproductive risk. Federal Register 53:24850-24869. U.S. Environmental Protection Agency. (1988c, August 26) Proposed guidelines for assessing female and male reproductive risk: extension of public cament period. Federal Register 53:32658. U.S. Environmental Protection Agency. (1988d, August 30) Intent to review . guidelines for carcinogen risk assessment. Federal Register 53:32656-32658. U.S. Environmental Protection Agency. (1988e, June 2) Integrated Risk Information System (IRIS); health risk assessment; guidelines, etc. Federal Register 53:20162-20164« ------- Figure 1 TOXICS INTEGRATION TASK FORCE RISK ASSESSMENT SUBCOMMITTEE RECOMMENDATIONS RISK ASSESSMENT GUIDELINES COORDINATE RISK ASSESSMENT PROCESS ------- Figure 2a RISK ASSESSMENT Dose-Response Assessment Risk Characterization Hazard Identification Exposure Assessment ------- figure 2b Risk Characterization ------- Figure 2c Think Organize Conduct Present Data Risk Characterization ------- figure 2d ISK ESSMENT »„.* MANAGEMENT Social Legal Economic Political f Risk Characterization Regulatory Decision ------- Figure 3 RISK ASSESSMENT GUIDELINES USE AND INTENT CONSISTENCY TECHNICAL QUALITY CLARIFY SCIENTIFIC ASSUMPTIONS FLEXIBILITY PROVIDE PUBLIC WITH EPA "KOAD NAP" ------- Figure 4 RISK ASSESSMENT GUIDELINES DEVELOPMENT/REVIEW PROCEDURE INTRA-AGENCY WORK GROUPS INVITED EXPERT REVIEW RISK ASSESSMENT FORUM REVIEW RISK ASSESSMENT COUNCIL REVIEW PROPOSAL FOR PUBLIC COMMENT SAB REVIEW FINAL AGENCY REVIEW OMB REVIEW FINAL PUBLICATION ------- _ Figure 5 RISK ASSESSMENT GUIDELINES ISSUED 1986 CARCINOGENICITY MUTAGENICITY DEVELOPMENTAL TOXICITY ESTIMATING EXPOSURES CHEMICAL MIXTURES ------- __ _ ..Figure 6 RISK ASSESSMENT GUIDELINES CURRENT PROJECTS FEMALE REPRODUCTIVE RISK MALE REPRODUCTIVE RISK EXPOSURE MEASUREMENT DEVELOPMENTAL TOXICITY (AMENDMENTS) NEUROTOXICITY NON-CANCER HEALTH EFFECTS CARCINOGENICITY (AMENDMENTS) AS OF SEPTEMBER, 1988 ------- . _ Figure 7 FEMALE REPRODUCTIVE RISK MALE REPRODUCTIVE RISK CHRONOLOGY PROPOSED JUNE 30, 1988 SCIENCE ADVISORY BOARD REVIEW JULY 14, 1988 COMMENT PERIOD ENDS SEPTEMBER 30, 1988 ------- _ _. - ..Figure 8 PROPOSED REPRODUCTIVE RISK GUIDELINES FORMAT HAZARD IDENTIFICATION O LABORATORY STUDIES O HUMAN STUDIES o PHARMACOKINETICS O MOLECULAR STRUCTURE ANALOGS O WEIGHT OF EVIDENCE DOSE RESPONSE ASSESSMENT O REFERENCE DOSE EXPOSURE ASSESSMENT RISK CHARACTERIZATION ------- Figure 9 MALE REPRODUCTIVE RISK GUIDELINES WEIGHT-OF-EVIDENCE CATEGORIES KNOWN POSITIVE PROBABLE POSITIVE POSSIBLE POSITIVE A. STUDY QUALITY UNCERTAINTY B. CONFLICTING RESULTS C. OTHER DATA KNOWN NEGATIVE PROBABLE NEGATIVE POSSIBLE NEGATIVE NO DATA/INADEQUATE DATA A. NO DATA AVAILABLE B. NEGATIVE, BUT STUDY QUALITY UNCERTAIN C. PREDICTIVE VALUE UNCERTAIN ------- Figure 10 RISK ASSESSMENT GUIDELINES AMENDMENTS/ADDITIONS DEVELOPMENTAL TOXICITY EXPOSURE-RELATED MEASUREMENTS CARCINOGENICITY AS OF SEPTEMBER, 1988 ------- - -- ' "Figure 11 CARCINOGENIC ITY GUIDELINES AMENDMENTS SCHEDULE • INFORMATION SOLICITATION OCTOBER 11, 1988 WORKSHOP EARLY, 1989 PROPOSED GUIDELINES FALL, 1989 AS OF SEPTEMBER, 1988 ------- - — - --Figure 12 CARCINOGENICITY GUIDELINES ISSUES ) CLASSIFICATION SCHEME - WEIGHT-OF-EVIDENCE - DEFINITIONS - OTHER CHANGES > QUANTITATIVE RISK ESTIMATION - CATEGORY C - MLE V. UCL - SCALING FACTORS • OTHER ISSUES - MECHANISMS"OF CARCINOGENESIS - BENIGN AND MALIGNANT TUMORS - RISK CHARACTERIZATION - FORUM REPORTS ------- FIGURE IJ NON-CANCER HEALTH EFFECTS (SYSTEMIC TOXICANTS) DIFFERENT APPROACHES I. END POINT SELECTION EXPOSURE PERIOD DOSE-RESPONSE CALCULATION UNCERTAINTY ADJUSTMENT HANDLING DATA GAPS ------- figure 14 FUTURE GUIDELINES OTHER SPECIFIC HEALTH EFFECTS - NEUROTOXICITY (UNDERWAY) - IMMUNOTOXICITY ECOLOGICAL RISK PHARWACOKINETICS OF SEPTEMBER, 1988 ------- |