United States      Science Advisory Board '  EPA-SAB-DWC-S3-Q1S
Environmental      A-101        Augyst 1993
Protection Agency     Washington, DC
REVIEW OF THE
METHODOLOGY FOR
DEVELOPING AMBIENT
WATER QUALITY CRITERIA
FOR THE PROTECTION OF
HUMAN HEALTH

PREPARED BY THE DRINKING
WATER COMMITTEE OF THE
SCIENCE ADVISORY BOARD

-------
              UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                           WASHINGTON, D.C, 20460
                              AugUSt 12, 1993               OFFICE OF THE ADMINISTRATOR
                                                             SCIENCE ADVISORY SOARD
EPA-SAB-DWC-93-Q16

Honorable Carol M, Browner
Administrator
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460

            Subject:     SAB Review of the ongoing revision of the
                        methodology for deriving National Ambient Water
                        Quality Criteria for the protection of human health.

Dear Ms. Browner:

      The Clean Water Act of 1977 required EPA to develop Ambient Water
Quality Criteria (AWQC)  for ambient water contaminants that may adversely
affect human health.  In  1980, the Agency published the methodology for the
development of AWQC and summaries of the criteria for 65 chemicals and
chemical classes. There have been many scientific advances relevant to AWQC
since that time, however, and the Office of Water (OW) is now revising the
methodology. On February 940, 1993,  the Drinking Water Committee of the
Science Advisory Board (SAB) met  to review the ongoing revision of the 1980
methodology for deriving National Ambient Water Quality Criteria (AWQC).
The charge to the Committee focused on the human health aspects of the
AWQC and outlined critical issues in six subject areas, namely Cancer Effects,
Non-Cancer Effects, Bioacetimulation, Exposure, Microbiology, and Minimum
Data Requirements,

      The Committee was pleased to learn of this important activity.  The
Office of Water is clearly engaged  in a systematic effort to develop a state-of-
the-science approach to revising the 1980  methodology.  The' process followed by
the OW to  identify the weaknesses in the methodology included the
development and wide circulation for comment of issue papers, followed by a
workshop of experts which produced the basis for the document reviewed by the
Committee,  We believe that this process was well thought-out and helped to

-------
clearly define many of the critical scientific issues and options concerning the
methodology.

      The Committee is concerned, however, that some of the approaches being
considered for setting AWQC by the Agency do not reflect the necessary
strategy of emphasizing regulation of contaminants in the medium (or media)
where each contaminant is most likely to qause adverse effects.  Instead, the
Agency approach focuses almost exclusively on point source discharges to water
and fails to place the exposures resulting from them in proper perspective. We
are concerned that setting AWQC in this manner could result in the
expenditure of large  sums of money without achieving significant reductions in
human exposure and risks.

      The Committee also wishes to emphasize that we foresee considerable
difficulty in using the concept  of Maximum Contaminant Level Goals (MCLGs)
in the development of AWQC methodology.  Introducing this concept into  the
AWQC is  likely to confuse the public, distort the relative importance of
carcinogens (versus untested contaminants), incorrectly mix carcinogenic and
non-carcinogenic effects in an unscientific manner, and result in  the
misdirection of resources if applied to the permitting process.

      In the remainder of this letter we summarize the major findings of the
review in  each of six subject areas that encompassed the  charge to the
Committee,  The accompanying report discusses .these issues more fully.

Cancer Effects

      Because carcinogenic risk assessment has evolved substantially since 1986,
we do not favor  the interim adoption of the 1986 Cancer Eisk Guidelines for
the revised AWQC methodology. Instead, the Committee urges the Agency to
expedite the revision and publication of updated guidelines, and to seek SAB
review of that revision.

      We urge the Agency to develop and use more accurate indications of the
uncertainty of cancer risk estimates than the currently-used upper and lower
confidence intervals, of the linearized multi-stage model (LMS). The Committee
also suggests  that the Agency routinely provide risk managers with examples of
comparable past  estimates of cancer risk and their corresponding administrative
outcomes, to help ensure that decisions are made with knowledge and

-------
appreciation of past Agency experience, while also making sure to incorporate
advances in scientific knowledge.

      Finally, the Committee recommends that the Agency tackle decisions on
Group C chemicals on a case-by-case basis, based on a clearly-defined process
that adequately defines the weight that will be given to different types of
evidence.

Non-Cancer Effects

      We agree with the Agency that the severity of effects should be
considered in the development of a Reference Dose  (RfD), but we find the
scale(s) that is being applied by  the Agency to this problem inadequate.  We
urge the Agency to consider the  alternative approach discussed in our report.

      The Committee believes that the precision of a given RfD must be based
on specific data sets for specific  RfDs, and suggests  that the Agency consider
reporting  the range from a calculated RfD to the  lowest observed adverse effect
level as a useful measure of such precision.

      The Committee generally  endorses the use  of the benchmark dose by the
Agency and supports the use of  PB-PK models (physiologically-based
pharmokinetie models) for RfD determination. However, we warn against the
dangers of using the results of short-term tests to establish anything other than
interim RfDs.

Bioaccumulation and Exposure Issues

      We strongly urge the Agency to base AWQC on sound experimental
evidence that bioaccumulation  does occur, rather than on hypothetical
assumptions that bioaccumulation might occur.  The Committee believes that
the strategy of setting AWQC by measuring contaminant concentrations  in
certain biota and then applying either a bioconcentration factor (BCF) or a
bioaccumulation factor (BAF) to  calculate water concentrations may not
accurately reflect the complex ways in which the real environment operates.
Although we support the Agency's efforts to develop well-validated BAFs, for
the time being we recommend  that the Agency rely  more heavily on BCF rather
than BAF, because of the higher likelihood of collecting an adequate BCF
database.

-------
      The Committee does not feel that it is appropriate to develop AWQC
geared to ensure that the sum of mil theoretically possible exposures never
exceeds the RfD by even a small amount.  We reject the routine use of the
percentage or subtraction methods for the allocation of the RfD, and the use of
default values, in the absence of reliable exposure data.  Instead, we endorse the
recommendation from the AWQC workshop held by the Agency in 1992 which
calls for bringing together all the appropriate offices or agencies when
significant contributions to txposure are expected from multiple sources, and the
total of those contributions exceeds the RfD.

      Finally, the Committee feels that the best way to protect subpopulations
with high fish consumption is to base health standards on the  levels of chemical
which are found in fish, not in  effluents,

Microbiology Issues

      The Committee found that the Agency was attempting to simultaneously
tackle too many issues related to the regulation of microbiological contaminants.
We strongly urge the Agency to set priorities and to focus efforts on ambient
recreational waters, which are not covered by other regulations or agencies. We
also recommend the formation of a multi-institutional workgroup to help EPA
and other agencies {e.g., Centers for Disease Control and Prevention, Food and
Drug Administration) address scientific and technical issues concerning
microbiological contaminants in ambient waters in a holistic manner.

      The Committee generally favors a risk-based  approach  to criteria for
pathogenic organisms in ambient waters.  We are also  aware  that there are
major gaps in epidemiologic research in microbiology and thus recommend the
coordination  of research and  risk assessments for pathogenic  microbes  in
ambient waters with assessments being done for other waters.  The Committee
supports an approach based on  the likelihood of human exposure to different
types, of ambient water  as the basis for identifying the types of waters  for which
criteria need to be developed.  We believe that AWQC for microbes should
include fecally-fcransmitted diseases other than gastroenteritis, as well as
microbes causing diseases of organs other than  the Gl  tract.

      The Committee also commented on the appropriateness of the currently-
approved indicator organisms for determining the safety of bathing waters, as
                                     4

-------
well as on research needs and their priority in the context of the ambient water
quality criteria initiative.

Minimum. Data Requirement Issues

      The Committee generally found the tiered approach presented by the
Agency to categorize the availability of data to be reasonable.  We have serious
concerns regarding the classification of chemicals in categories (e.g., Group C)
where they may not properly belong, so that a regulatory decision can be made,
or the use by the states of Tier III values to set permanent permit levels.
Finally, the Committee recommended certain criteria that might be applicable to
the proper use  of 28-day study data in developing interim toxicity values.

      Additional detail on the above issues can be found in the complete report
by the Committee, which is attached.  We should also stress that we did not
attempt to comprehensively review all the issues discussed in this report.  We
look forward to revisiting ma&y of these critical issues with your staff in the
future, as they  continue to revise the 1980 AWQC methodology.  The
Committee appreciates the opportunity to conduct this review, and we look
forward to your response to the scientific advice transmitted herein.
                              Sincerely,
Dr, Raymond C, Loehr, Chair        Dr. Verne A. Ray, Chair
Executive Committee                Drinking Water Committee
Science Advisory Board              Science Advisory Board

-------
                                 NOTICE

    This report lias been written as a part of the activities of the Science
Advisory Board, a public advisory group providing extramural scientific
information and advice to the Administrator and other officials of the
Environmental Protection Agency, The Board is structured to provide balanced,
expert assessment of scientific matters  related to problems facing the Agency.
This report has not been reviewed for approval by the Agency and,  hence, the
contents of this report do not necessarily represent the views and policies of the
Environmental Protection Agency, nor  of other  agencies in the Executive
Branch of the Federal government, nor does mention of trade names or
commercial products constitute  a recommendation for  use.

-------
                                 ABSTRACT

       On February 9-10,  1993, the Drinking Water Committee of the Science
 Advisory Board (SAB) reviewed the Agency's revision of the methodology for
 deriving National Ambient Water Quality Criteria (AWQC) for the protection of
 human health.

       The Committee was pleased to learn of the Agency's systematic effort to
 revise this methodology.  They were critical of the emphasis given to point
 source discharges in the ongoing revision. They' commented on the Agency's
 revision of its 1986 Cancer Risk Guidelines, on the need to incorporate
 mechanistic information in them, and on the Agency's treatment of Group C
 chemicals and uncertainty.  They addressed issues of severity scales for non-
 cancer effects, the development and allocation of RfD values, the use of short-
 term study data, Health Advisory Doses, and the  benchmark dose. They
 reviewed the Agency's use of Bioaccumulation Factors (BAFs)  and
 Bioconcentration Factors  (BCFs), the use of MCLGs (Maximum Contaminant
. Level Goals) in AWQC methods, and the use of separate criteria for drinking
 water and fish intake.

       The Committee urged EPA to priority-rank the needs related to
 microbiologic exposures in and supported the use  of new structures to assist
 EPA,  They recommended exposure potential as the basis for microbiological
 criteria,  and supported a risk-based approach to the regulation of microbes.
 They commented on the relationship of indicator  organisms to non-GI illnesses,
 the efficacy of indicators in  tropical waters, and research needs on determinants
 of virulence, injured pathogens, and molecular techniques for pathogen
 identification,   .             '"

       The Committee also reviewed the proposed use of a. tiered approach to
 categorize data availability, and addressed issues concerning the categorization
 of Group C chemicals under this scheme.
 Key Words: Cancer, Risk Assessment, RfD, Ambient Water Quality Criteria,
            Non-Cancer Risk, Bioaecumulation, Exposure, Minimum Data
            Requirements, Point Source Discharges, Microbiological
            Contaminants, Group C Chemicals, Bioconcentration.
                                     ii

-------
            ENVIRONMENTAL PROTECTION AGENCY
                   SCIENCE ADVISORY BOARD
                 DRINKING WATER COMMITTEE
    Review of Revision of Methodology for Ambient Water Quality Criteria

CHAIRMAN
Dr. Verne A. Ray, Medical Research Laboratory, Pfizer Inc., Groton,
      Connecticut
MEMBERS
Dr. Richard J, Bull, College of Pharmacy, Washington State University,
      Pullman, Washington
Dr. Gary P. Carlson, Department of Pharmacology and Toxicology, School of
Pharmacy, Purdue  University, West Lafayette, Indiana
Dr. Keith E.  Cams, East Bay Municipal Utility District, Oakland, California
Dr. Lenore S. Clesceri, Rensselaer Polytechnic Institute, Materials Research
      Center, Troy, New York
Dr. Ramon G. Lee, American Water Works Service Company, Voorhees, New
      Jersey
Dr. Ecb D. Pellizzari, Research Triangle Institute, Research Triangle Park,
      North  Carolina
Dr. Mark D.  Sobsey, Department of Environmental Sciences and Engineering,
      School of Public Health, University of North Carolina, Chapel Hill, North
      Carolina
Dr. James M, Symons, Department of Civil  and Environmental Engineering,
      University of Houston, Houston, Texas
SCIENCE ADVISORY BOARD STAFF
Mr. Manuel R, Gomez, Designated Federal Official, Science Advisory Board
      (A-101F), U.S. EPA, 401 M Street, SW, Washington, DC  20460
Mrs- Marcy Jolly, Staff Secretary, Drinking Water Committee,  Science Advisory
      Board  (A-101F), U.S. EPA, 401 M Street, SW, Washington, DC 20460

-------
                       TABLE OF CONTENTS
1. EXECUTIVE SUMMARY	,	,   1
     1,1   Cancer Risk Issues	,	   1
     1,2   Non-Cancer Risk Issues  	,	   2
     1.3   Bioccumulation and Exposure Issues	   3
     1.4   Microbiology Issues	.	,   5
     1.5   Minimum Data Requirement Issues ,....,..,....,,,....   6

2, INTRODUCTION 	, ,	   7
     2.1   Background	, ,	   7
     2.2   Charge To The Committee	'.,,..   8

3. FINDINGS	,	   9
     3,1   Introduction	   9
     3.2   Cancer Risk Issues	,	 10
     3.3   Non-Cancer Risk Issues  : .	 15
     3.4   Bioaccumulation Issues	 19
     3.5   Exposure Issues	 21
          3.5.1 Allocating the RfD	 21
          3.5.2 Exposure Route, Rate, and Duration Issues . ,	 24
          3.5.3 Fish Consumption Rates	 26
     3.6   Microbiology Issues	 27
     3.7   Minimum Data Requirement Issues .	 ,	 35

4. REFERENCES	 38

-------
                      1. EXECUTIVE SUMMARY
      On February 9-10, 1993, the Drinking Water Committee ("the
Committee") of the Science Advieory Board met to review the Agency's ongoing
revision of the methodology for deriving National Ambient Water Quality
Criteria (AWQC) for the protection of human health,  A document summarizing
the revision process, prepared by the Human Risk Assessment Branch (HRAB)
of the Office of Water (OW) was  presented and discussed  ("Revision of
Methodology for  Deriving National Ambient Water Quality Criteria for the
Protection of Human Health: Report of Workshop and EPA's Preliminary
Recommendations for Revision").   The Committee was pleased to learn that the
Agency is engaged in a systematic effort to develop a  state-of-the-seience
approach to revising the 1980 methodology.   The review focused on key
questions posed to the Committee in six subject areas: Cancer  Risk, Non-Cancer
Risk, Bioaeeumulation, Exposure, Microbiology* and Minimum Data
Requirements.

1.1   Cancer Risk Issues

      a)     The Committee was  pleased to learn that the Agency's 1986 Cancer
            Risk Guidelines are beginning to undergo revision and urged the
            Agency to incorporate in the revised Guidelines, as well as in the
            AWQC methodology, the growing body of scientific knowledge
            regarding carcinogenic mechanisms that has accumulated since the
            Guidelines were adopted.  Because carcinogenic risk assessment has
            evolved  substantially since 1986, the Committee rejected the
            interim  adoption of the 1986 Guidelines in  AWQC methodology,
            and  instead urged the Agency to expedite the review and
            publication  of updated Guidelines.  The Committee also
            recommended that any new Guidelines be appropriately reviewed
            by a multidiscipiinary SAB commmittee.

      b)     With regard to the characterization of risk, the'Committee
            suggested that the Agency routinely provide risk managers with
            risk estimates that are placed in appropriate context with examples
            of comparable past estimates of risk and  their corresponding

-------
            administrative outcomes.  This approach would help ensure that
            decisions are made with knowledge and appreciation of past Agency
            practices, while also making sure to incorporate advances in
            scientific knowledge,

      c)     The Committee rejected the current approach of using the upper
            and lower confidence intervals of the linearized  multi-stage model
            (LMS) as a meaningful estimate of the uncertainty around the
            point  estimate of risk, because these intervals are only a measure
            of the uncertainty of a non-verifiable model, and do not convey the
            uncertainty of interspecies or  low-dose extrapolations.  Instead, the
            Committee suggested that the Agency develop more meaningful
            indications of the uncertainty  of risk estimates,

      d)     Finally, the Committee recommended that the Agency tackle
            decisions on Group C chemicals on a case-by^case basis, based on a
            clearly-defined process that adequately defines the weight that will
            be given to different types of evidence,

1.2   Non-Cancer Risk Issues

      a)     The Committee agreed with the Agency that the severity of effect
            should be  considered in the development of a Reference Dose
            (RfD), but considered that the sealers) that are being applied by
            the Agency to this problem are vague and prior  pronouncements
            and studies conducted  in various offices within the Agency have
            confused rather than clarified  the issue.  Of three scales that may
            be applicable, the Committee considered that the one with  the
            widest applicability for non-cancer risks would be a scale based  on
            whether the effect represents actual pathology (e.g. extensive
            necrosis), a functional deficit (e.g., paralysis, learning deficits,
            impaired biliary secretion), a biochemical change that is clearly
            related to  an adverse effect, or simply a biochemical or behavioral
            change of unknown consequence.

      b)     The Committee believes that defensible statements  about the
            precision of a given RfD must be  based on  specific data sets for
            specific  RfDs, and that reporting the range from a calculated RiD
            to the lowest observed adverse effect level may serve as one useful

-------
            measure of precision of an RfD, because it wouid essentially define
            the uncertainty of an estimated "safe" exposure level.

      e)     The Committee believes that PB-PK modeling can be useful for
            RfD determination.  This utility is greatly enhanced when there is
            reasonably good understanding of the mechanisms by which the
            chemical is acting.

      d)     The Committee concluded that important toxic effects could be
            missed because of the possibility of false negative findings arising
            from studies of less  than  90 days exposure. To guard against this
            danger, they recommend that data from such tests only be used to
            set interim or  temporary  RfDs.

      e)     The Committee rejected the use of site-specific  Health Advisory
            Doses (HADs)  for one-day and even longer term exposures in the
            context of ambient water  quality.

      f)     The Committee generally  endorsed the use of the benchmark dose
            by the Agency, but emphasized several potential pitfalls of this
            method,

1.3    Bioccumulation and Exposure Issues

      a)     The Committee cautions the Agency that the strategy of setting
            AWQC by  measuring contaminant concentrations in certain biota
            and then applying either  a bioconcentration factor (BCF) or a
            bioaeeumulation factor (BAF) to calculate water concentrations may
            not accurately  reflect the  complex ways in which the real
            environment operates.  The report describes a number of instances
            where problems may arise with this approach and urges the
            Agency to  base AWQC on sound experimental evidence that
            bioaeeumulation does occur, rather than on hypothetical
            assumptions that bioaeeumulation might occur. The Committee
            recommended that, for the time being, the Agency focus attention
            on BCF rather than BAF, because of the higher likelihood of
            collecting an adequate  BCF database.

-------
b)    With regard to allocation of the RfD, the Committee did not feel
      that it is appropriate to develop AWQC geared to ensure that the
      sum of all  theoretically possible exposures never exceeds the RfD
      by even a small amount.  They rejected the routine use of the
      percentage or subtraction methods, and the use of default values in
      the absence of reliable exposure data.   Instead, the Committee
      endorsed the recommendation- from the AWQC workshop held by
      the Agency in 1992 which calls for bringing together all the
      appropriate offices or agencies when significant contributions to
      exposure are expected from multiple sources, and the total of those
      contributions exceeds the RfD. Finally, the report recommended
      the use of  separate criteria based on fish intake and water
      consumption (i.e., hypothetieally using up all the RfD in each
      calculation),

c)    The Committee foresees considerable difficulty in using the concept
      of Maximum Contaminant L
-------
      f)     The Committee feels that the best way to protect subpopulations
            with high fish consumption is to base health standards on the
            levels of chemical which are found in fish, not in effluents.

1.4   Microbiology Issues

      a)     The Committee found that th'e Agency was attempting to tackle  too
            many issues related to the regulation of microbiological
            contaminants at the same time, with insufficient resources. They
            strongly urged  the Agency to set priorities and to focus its efforts
            on ambient recreational waters, which are not covered by other
            regulations or agencies, and to leave other waters (e.g,» drinking
            source waters, shellflshing waters) to be addressed through existing
            regulatory programs,  The Committee also supported the formation
            of a multi-organizational workgroup  to help EPA and other
            agencies (e.g., Centers for Disease Control and Prevention, Food
            and Drug Administration) address scientific and technical issues
            concerning microbiological contaminants in ambient waters in a
            holistic manner,

      b)     The Committee generally favored a risk-based approach to criteria
            for pathogenic organisms in ambient waters, but they were also
            aware that there are major gaps in epidemiologic research in
            microbiology that must be addressed, and they recommended the
            coordination of research  and risk assessments for pathogenic
            microbes in ambient waters with assessments being done for other
            waters. The Committee  also supported an approach based on the
            likelihood of huntan exposure to different types of ambient water
            as the  basis for identifying the types of waters for which criteria
            need to be  developed, and the inclusion of microbes causing fecally-
            transmitted diseases other than gastroenteritis, as well as microbes
            causing diseases of organs other than the GI tract. Finally, they
            do not believe that current ambient recreational water quality
            criteria are appropriate or transferable to other ambient waters,

      c)     The Committee believes that the currently-approved indicator
            organisms in beach  waters  are not always appropriate for
            determining the safety  of bathing waters with respect to gastro-
            intestinal (GI) disease,  that these organisms are not likely to be

-------
            adequately predictive of the human health risks from non-GI
            illnesses associated with human or animal fecal contamination, and
            that their validity and usefulness in tropical ambient waters is
            uncertain.  They are also of the opinion that there are candidate
            alternative indicators worthy of consideration and deserving of
            investigation in this initiative, including several examples
            mentioned in the workshop report and EPA's response,

      d)    Finally, the Committee believes  that research efforts in other areas,
            including the virulence determinants of microbial pathogens, the
            public health significance of injured pathogens, and  current
            molecular techniques for pathogen detection and identification are
            important but are beyond the scope of the ambient water quality
            criteria initiative and should be  of low priority in this regulatory
            context,

1.5   Minimum Data Requirement Issues

      a) .   The Committee generally found  the tiered approach presented by
            the Agency to categorize the availability of data to be reasonable.
            The report addresses certain aspects of the  criteria for each tier in
            the proposed system.

      b)    With regard  to the categorization of Group  C chemicals, the
            Committee feels that it would be inappropriate to place a chemical
            in a category where it  does not belong so that a regulatory decision
            can be made, and that allowing  the states to use Tier 111 values to
            set permanent  permit levels would be very risky.

      c)    The Committee discussed certain criteria which might be applicable
            to the proper use of 28-day study data in developing interim
            toxicity values.
                                      6

-------
                          2. INTRODUCTION
2.1   Background

      The Clean Water Act of 1977  (Public Law 95-217) required EPA to
develop Ambient Water Quality Criteria (A'WQC) for ambient water
contaminants that may adversely affect human health,  In 1980, the Agency
published the methodology for the development of AWQC in the Federal
Register  (US EPA, 1980).  They also published summaries of the criteria for 65
chemicals and chemical classes and announced their availability to the public.
The criteria contain recommended maximum permissible pollutant
concentrations consistent with the protection of human health, aquatic
organisms,  and some recreational activities.

      There have been many advances in  scientific disciplines relevant to the
evaluation of ambient water quality and the development of criteria since 1980,
and hence the Human Risk Assessment Branch (HRAB) of the Office of Water
(OW)  is now undertaking a large scale review  and revision of the 1980 criteria
methodology. As  part of the revision process,  the HRAB/OW organized a
workshop of experts in September of 1992 to examine a set of critical issues
concerning the methodology.   The workshop participants were organized into six
subject areas, depending on their expertise. The six areas were Cancer Risk,
Non-Cancer Risk, Bioaccumulation, Exposure, Microbiology, and Minimum Data
Requirements,  Each work group reviewed a set of detailed issue questions and
developed relevant recommendations.

      Based on the outcome  oT the Workshop, the HRAB/OW prepared a draft
report entitled "Revision of Methodology for Deriving National Ambient Water
Quality Criteria for the Protection of Human Health; Report of Workshop and
EPA's Preliminary Recommendations for Revision" (henceforth the "Report").
This draft report defined critical issues in each of the six subject areas listed
above, summarized the recommendations  arising from these areas in the
Workshop (including minority opinions), and reported the Agency's (HRAB/OW
own preliminary recommendations in each area,

-------
2,2   Charge To The Committee

      In its original charge, the  HRAB/OW requested that the Drinking Water
Committee ("the Committee") of the Science Advisory Board (SAB) review the
general direction of the revision process underway for the criteria methodology.
Specifically, the charge was to (a) review each of the critical issues addressed by
subject areas, and (b)  review the workshop recommendations and the
preliminary HRAB/OW recommendations summarized in the Report,

      Based on the discussions during the meeting of February 9-10, 1993, the
HRAB/OW modified the charge to the Committee.  The new charge had the
same broad objectives, but addressed a somewhat reduced number of priority
questions concerning each of the six areas into which the document  is
organized.  The Committee findings are  reported in Chapter Three of this
report, which is organized into the six subject areas described above.  Because
of the wide range of issues tackled by this review, the Committee did not
attempt to  treat them all comprehensively.  Indeed, the Committee expects to
revisit  a number of these issues in more depth in future reviews.

-------
                              3. FINDINGS
3,1   Introduction

      The findings  in this Chapter are organized following the six subject areas
described earlier, namely Cancer Risk, Non-Cancer Risk, Bioaccumuiation,
Exposure, Microbiology, and Minimum Data Requirements. Before describing
the specific findings in this manner, however, the Committee wishes to address
several key general  issues concerning AWQC,  .

      The Committee was pleased to learn of this important activity.  The
Office of Water is clearly engaged in a systematic effort to develop a state-of-
the-science approach to  revising the 1980 methodology. The process followed by
the Office to  identify the weaknesses in  the methodology included the
development  and wide circulation for comment of issue papers, followed by a
workshop of  experts which produced the basis for the document reviewed by the
Committee, We believe that this process was well thought-out and helped to
clearly define many of the critical scientific issues and options concerning the
methodology.

      The Committee is of the opinion,  however,  that the Agency should make
a determined' effort  to begin regulation of contaminants in the medium (or
media) where the contaminant is most likely to cause adverse effects.  This
means that some contaminants would best be regulated by setting standards
based on fish consumption, and others by developing standards based on water,
air, air plus water, etc.  Such an effort requires special coordination among the
various regulations  for which EPA is responsible  as  well as some regulations for
which EPA is not responsible.  In her recent interview in Chemical &
Engineering News Administrator Browner addressed concerns of this type when
she said that "the goal of an agency like EPA should be to look  across [these
various laws] to see how they can be integrated into a more meaningful
regulatory scheme, an environmental protection plan (C&E News,  1993)."  She
went on to indicate  that such integration must cut across  Agency boundaries.

      The Committee feels that some of the approaches being considered for
setting Ambient Water Quality Criteria (AWQC) by the Agency do not  reflect
this important and appropriate strategy.   The document provided by the Agency
for review focuses almost exclusively on  point source discharges to water and

-------
fails to place the exposures resulting from them in proper perspective.  The
Committee is concerned that setting AWQC in this manner could result in the
expenditure of large sums of money without achieving significant reductions in
human exposure and risks.

       Because aquatic systems often receive contaminants from multiple point
and non-point sources  (including fallout from  the air), the concentration of a
given chemical in ambient water may exceed proposed criteria even before the
contribution from any  single  point source is added. Thus a point source might
find itself in the position of being legally unable to discharge water of the same
quality it obtained from ambient water, even if it did not itself contribute to the
presence of those contaminants.  For example, a situation could occur in which
a wastewater treatment plant receiving chloroform from drinking water
disinfection at a level of 50 jig/1 may not be able to discharge this water without
violating AWQC,

3.2    Cancer Risk Issues

       The Committee was asked to comment on a number of critical scientific
issues  concerning the assessment of carcinogenic risk during its  review of the
methodology for developing ambient water quality criteria {AWQC),  Many of
these  issues are also currently being examined by the Agency as part of its
revision of th,e 1986 "Guidelines for Cancer Risk," although this  process is still
at a very early stage, judging from the briefing provided to the Committee.  The
overlap of-these  two revision processes in the  Agency requires that the
Committee place its comments on specific cancer risk assessment issues for the
AWQC methodology in the proper context.

      The  Committee believes that  the update of the  1986 Guidelines is among
the most important activities within the Agency at this time. There have been
substantial advances in our understanding of cancer in the past  few years,
including important insights into mechanisms  of earcinogenesis,  and the Agency
should seek to incorporate these advances into the revised Guidelines.
Moreover, it is unlikely that the Agency will be able to deal effectively and
consistently with  problems  in other water programs, such as disinfection by*
products, without' updating the Guidelines.

      Before new Guidelines  are adopted, however, the Committee feels that
they should be reviewed by the SAB, perhaps utilizing a special  ad hoc

                                     10

-------
 committee constituted from members from several committees.  This is critical
 because the guidelines impact the work of many EPA program offices and thus
 the expertise and concerns of many committees.  Also, the guidelines should be
 revised in. such as way that they can be applied consistently across EPA.

       It is also critical that the Agency clearly define the objectives of the
 revision, and the Committee feels that one such objective must be to develop a
 more scientifically-based rationale for classification of chemicals, with greater
 weight given to mechanistic data than is the case today.  The Agency should
 make sure that the new guidelines improve regulatory policy and practice, and
 that they will facilitate the management of difficult cases, rather than
 maintaining the regulatory boundaries that are in the present scheme.  A
 serious concern to the Committee is whether revised Guidelines would permit
 new concepts and data to be incorporated quickly, as they are validated  in the
 scientific  world, or whether each new scientific development would need to be
. officially recognized through a lengthy process before it can be adopted by the
 Agency's  guidelines,

       These and other critical questions should be part of the EPA's revision of
 the 1986  Guidelines.  Until that revision is complete  and a more comprehensive
 draft of the revised Guidelines is available, however,  our comments in the area
 of cancer risk assessment for AWQC must be  viewed as preliminary.  They will
 need to be re'-examined in context when the revision  of the Guidelines is closer
 to completion,

       a)    Should the 1986 EPA Cancer Assessment Guidelines be adopted
            with allowance for flexibility as the new revised Guidelines  are
            published?  Should emphasis be placed on mechanistic data and
            route and level of exposure to assist in interpretation of relevance
            of tumor occurrence,  as discussed in the draft revised Cancer
            Guidelines?

       The 1986 EPA Cancer Guidelines should not be adopted for the revised
 AWQC methodology.  There has been substantial progress in our understanding
 of cancer since 1986,   The "Working Paper" for the new guidelines includes
 consideration, of many advances in cancer research during this period whose
 implementation would significantly affect risk assessments outcomes.  For the
 Agency to continue to rely on the old Guidelines in the development of water
                                     11

-------
quality criteria has the potential of creating considerable confusion in the
future.

      It is laudable that the EPA "Guidelines for Cancer Risk" should be
reexarained in light of changes in our knowledge.  Clearly, we are beginning to
glimpse the broad mechanistic landmarks of carcinogenesis for some cell and
tumor types. We can begin to understand-some cases of genetic predisposition
to cancer and have identified a number of oncogenes and tumor suppressor
genes which appear to regulate cell growth and differentiation.  Because more
comprehensive  mechanistic insights are likely to emerge in the future, we need
to have the flexibility -of mind and regulatory process to respond to these
advances  as they occur.  For example, mechanistic insights should allow better
understanding of which experimental findings may not be applicable to humans
(e.g., alpha-2n-globulin production and susceptibility to the careinogenieity of
gasoline in male rat kidneys), New methods are allowing us to identify the
biological consequences of exposures as  genetic alterations, and to  measure the
doses to genes that are important to cancer induction.  We have increasing
knowledge about the kinds of lesions that particular carcinogens produce and
the spectrum of sites they affect  in particular genes.  We even have growing
insight about spontaneous genetic events, some of which may represent
oxidative  damage to genes and/or chromosomes in addition to replication errors.
The insights that these advances may bring to the risk  estimation process
should not be ignored.

      Data on genetic activity of chemicals in short  and long term tests in vitro
and even  in vivo should be given greater weight in the future.  However, the
incorporation of this information in risk assessment  should emphasize the
relationships between specific mutation sites, the resulting alterations in specific
proteins, and the development of cancer.  The  simple identification of the ability
of a given compound to produce mutations in various test systems has limited
utility in  risk assessment, beyond the ability of classifying compounds as
putative mutagens or non-rnutagens. Studies  that give insight into the  normal
and abnormal behavior of human cells and tissues should be  used  to help bridge
the gap between studies in animals and effects in human beings.  Powerful new
methods to  assess environmental exposures to chemicals and  internal doses
should be employed-where possible to build the data bases  needed  to link dose
and effect more precisely.
                                     12

-------
      The use of genetic toxicology for classifying the mutagenic potential of
chemicals needs to be more critically evaluated than in the past, particularly
with reference to the magnitude of doses that are required to produce effects.
This is a particularly important issue in the interpretation of data from
chemicals that give rise to clastogenic effects in the absence of an ability to
induce point mutation.

      b)    Should risk be presented in two ways? (1) Point estimate with
            indication of uncertainty; (2) Set of credible alternative estimates
            based on applicable models.

      Estimates of risk are intended to be used as input information by risk
managers.  The estimates of risk are usually derived from a limited body of
data, numerous  extrapolations are necessary, and the resulting risk estimate is
typically limited by the intrinsic error in the data and propagation of these
errors through the  extrapolation processes.  The methods  used for extrapolation
or the concepts  upon which they are based may, with the  advance of scientific
knowledge, be proven Incorrect, The Committee therefore recommends that risk
estimates be presented in both manners: (1)  as point estimates of risk  with an
indication of the uncertainty of the estimate, and (2) as a collection of credible
alternate estimates  based on applicable models (i.e., mechanistic models, not
alternative statistical models which cannot be experimentally verified in any
practical manner).

      Moreover, because it reeogtmes that risk estimates are likely to be
imprecise, the Committee feels that it is important to provide risk managers
with risk estimates that are placed in appropriate context.  The Committee
suggests that the Agency routinely provide assessments to risk managers  with
examples of comparable past estimates of risk and their corresponding
administrative outcomes.  This approach would help ensure that decisions are
made with knowledge and appreciation  of past Agency practices, while  also
making sure to  incorporate advances in scientific knowledge.  It would  permit
the Agency to know, for roughly comparable situations, that risk estimates and
administrative decisions do not differ (weighing costs against risks) by  more
than an  order of magnitude,  without an explicit decision to that effect.  Such
examples of past practice  should not seek to "freeze" the Agency's science, but
rather to ensure that estimates take into account both past Agency practice and
the evolution of scientific knowledge,
                                     13

-------
      The Agency must also be very careful how it expresses uncertainty.  The
conventional use of the upper 95% confidence limit of the LMS (linearized multi-
stage model) estimate of risk has introduced a great deal of confusion Into the
analysis of uncertainty.  The LMS convention was adopted primarily because it
yields a more stable estimate of risk than the maximum likelihood estimate (i.e.
the Maximum Likelihood Estimate or MLE  is too sensitive to the data).
Therefore, the difference between the MLE and upper 95% confidence limit is
more of a measurement of how poorly the data fit the upper 95% confidence
limit of the  extrapolation from the model than any real expression of
uncertainty.  The Committee rejects the current approach of using the upper
and lower confidence intervals of the LMS as a meaningful estimate of the
uncertainty  around the point estimate of risk, because it is an extrapolation
beyond available data or experimental confirmation.  The point estimates
currently in use by the Agency almost always involve extrapolations over orders
of magnitude, based on assumptions that may be incorrect in many cases.  To
employ such figures as meaningful estimates of the uncertainty in the results of
a risk assessment is clearly not appropriate, A more accurate indication of the
uncertainty  would be iimply to  state the extent to which these estimates are
extrapolations (e.g., compare the lowest  dose that has been shown to produce
cancer in a valid study to the model estimate in the same units),

      c)     Should decisions on Group  C Chemicals (whether to use the RfD
            'approach or cancer quantification by LMS) be made on a case-by-
            case  basis after consideration of the appropriate level of concern
            and appropriateness of the  data?

      We recommend that the Agency address these situations on a cmse-by-case
basis, based on a  clearly-defined process, particularly with respect to the weight
that will be  given to different types of evidence.

      The Group C category is  a  reflection of the 1986 Guidelines,  As stated in
item a above, the Committee strongly supports the revision of those Guidelines
to include mechanistic  information and perhaps, as was suggested during the
briefings, elimination of the weight of evidence categorization of carcinogens. In
this context, the comments that follow may soon become outdated by revised
Guidelines,        '

      Group C chemicals are a  very heterogeneous collection. At one extreme,
some chemicals are in this class because there is a tenuous indication of
                                     14

-------
careinogenieity from isolated evidence of an increase in rates of cancer in one
.species or sex, in one organ (often with tumors that commonly occur
spontaneously in that organ), or even because of a structural similarity to an
established carcinogen.  This results in the placement of chemicals into this
category even when they have been adequately studied and found negative in
several other animal species (e,g,, tetrachloroetkylene, trichloroethylenet numerous
pesticides).  Such chemicals also lack human careinogenesis data, since positive
human findings would have resulted in a higher ranking of the chemical Such
chemicals frequently may be best treated by the RfD approach. At the other
extreme, there are chemicals that would have been classified in class B if only
some of the available data were slightly stronger (e.g., with good dose response
information).  In such instances  where scientific judgement would suggest that
the lower classification  is simply the result of limited data* prudence may dictate
the additional conservatism of the linearized multistage model.

       Given such diversity in the chemicals currently in Group C, there is room
to make distinctions between those posing likely carcinogenic risk and  those
with little probability of careinogenieity. At the extremes discussed above, it
would be appropriate to treat chemicals with  the highest probability of risk of
careinogenieity as if they were carcinogens, while the chemicals least likely to be
carcinogens are  treated with the RfD approach and  appropriately conservative
uncertainty factors, as described above.
            /
       The problematic chemicals, however, will be those that fall between these
extremes.  Should they be assessed as carcinogens or not? In these  cases, it
may be appropriate to rely on ancillary data to provide insight into the
mechanism by which the chemical is acting.  For example, is there evidence that
the chemical is mutagenic with appropriate consideration of the doses used in
the test systems? What other types of effects does it have in the target organ that
may be influencing the development of cancer?  Is the chemical related to other
chemicals that are recognized chemical carcinogens?

       Non-Cancer Eisk Issues

       a)    Should the methodology (following the EPA nopcancer  health.
            effects guidelines) expand on the aspects of severity of effect and
            presentation of the RfD?
                                     15

-------
       There is no question that the severity of effect should be considered in
 the development of a RfD, However, the scale(s) that are being applied by the
 Agency to this problem are vague and prior pronouncements and studies
 conducted in various offices within  the Agency have confused rather than
 clarified the issue.  Theoretically, scales of severity could be constructed based
 on: (1) whether the effect is reversible, or irreversible and cumulative; (2)
 whether the effect represents actual pathology (e.g, extensive necrosis), a
 functional deficit  (e.g.,  paralysis, learning deflcits, impaired biliary secretion), a
 biochemical change that is clearly related to an adverse effect, or simply a
 biochemical or behavioral change of unknown consequence; or (3) simply by the
 target organ affected.

       These potential severity scales would have different applications.  Clearly,
 the first is easily  related to assumptions  about chemical carcinogenesis, but also
 selectively applies to some neurotoxic and reproductive effects (e.g., effects of
 acrylamide or n-hexane) and perhaps other effects.

       For most non-cancer risks, however, the second scale will probably have
,the greatest applicability,  Serious consideration should be given to determining
 how dose-response relationships for an effect measured at the biochemical level
 are related to modifications in function and to the development  of overt
 pathology in several target organs.  In the meantime, it is very important that
 the categorization of these effects continue to  be emphasized in  the development
 of a HfD, particularly when it is developed from a  LQAEL,  These
 considerations should also be made  explicit when developing an  BfD from a
 NOAEL,  reducing the standard uncertainty factors when the NOAEL is based
 on biochemical effects at exposures  below those which produce overt toxic
 effects.

       The third option  requires that different values be given to different
 organs or systems, something that is not easily dealt with beyond a  few trivial
 examples. Consequently, the Committee  is skeptical about the value of such a
 scale,

       b)    Should an RfD range estimate be presented,  considering that
            currently the RfD is an estimate with a precision vaiying within an
            order of magnitude?
                                      16

-------
      The notion that the RfD is roughly accurate within an order of
magnitude seems to imply that there is some way to evaluate the precision of a
particular RfD.  There is  little data of which the Committee is  aware that
wculd actually support the application of the term precision to  this effort.  More
defensible statements about the precision of any given RfD would have to be
based on specific data sets for specific RfDs.  The range from the calculated RfD
to the lowest observed effect level may serve as one useful measure of precision
of an RfD, because it would essentially define the uncertainty of an estimated
"safe" exposure level,  It would be very important that the seriousness of the
effect observed at the LOAEL be made clear if such an approach is utilized.

      c)    Should PB-PK and dosimetry modeling be used for RfD
            determination?

      The Committee believes that PB-PK modeling can be useful for RfD
determination.  This utility is greatly enhanced when there  is reasonably good
understanding of the  mechanism by which the chemical is acting,

      d)    Should studies of less than 90 days exposure (e,g., 28-30 days) be
            used with uncertainty factors for RfD determination?

      No justification was provided for the use of data from a 28-30 day
exposure for estimating an RfD.  Moreover,  there is no real definition of what
would be examined in a 28-30 day experiment.  In some cases such data might
be utilized.  For example,  if such an experiment were to include a genuine
attempt to identify target  organs by doing comprehensive histopathological
examination of organs, the results might be used on a  temporary basis. This
could be particularly important if the results provided  clear  indications of
adverse effects.  However, the use of negative data from such short term
experiments would miss effects  that take longer periods of time to develop, and
thus could be quite dangerous.  Finally, there are experiments of this length
that are performed for reasons other than safety evaluation  (e.g., focusing on
whether a particular biochemical effect occurs or not) which may be important
in providing mechanistic data to explain results in other studies.  However, this
type of  data should not be used as the sole basis for  developing an RfD, unless
the pharmacodynam'ie relations between the biochemical interactions and the
toxicologic response' is understood in quantitative terms.
                                     17

-------
      The danger of false negative findings from short-term studies, both 30-
and 90-day, also exists for non-cancer effects  that simply will not be apparent  in
such studies, including reproductive effects, teratogenesis and developmental
toxieities.  Consequently, RfDs based on such data must clearly be labeled as
being interim or temporary.  The establishment of such RfDs should consider
the additional uncertainty created by the fact that such effects have not be
evaluated,

      e)    Should site-specific  health advisory doses (HADs) for one day and
            longer-term exposure situations be derived, rather than ambient
            water criteria based on life-time  exposure?

      The application of a site-specific Health Advisory Doses (HADs) in the
context of ambient water quality criteria is of questionable utility.  The
Committee simply does not see how such figures would be utilized in terms of
the permit program,

      0    Some discussions at the meeting touched on the use of the
            benchmark dose as  a means of dealing with  non-cancer risks. The
            Committee would like to make some opinions known on this issue.

      There are several advantages of using the benchmark  dose (Crump, 1984).
The most important is that it allows use of all the available  data to derive dose
which can be used for developing an RfD.   It  derives a number that is largely
independent of the spacing between doses  and it can be utilized in such a way
as to  reflect the quality of the data that are available.  The Committee generally
endorses the use of the benchmark dose approach for these reasons, but
emphasizing several potential pitfalls that should be avoided.

      With the benchmark dose  there is a tendency to utilize a low benchmark
(e.g.,  an ED01 or even lower) to  which uncertainty factors are applied to  derive
the RfD.  Ordinarily the benchmark is taken to be the upper 95% of the  point
estimate made by the  multistage model. The  Committee  urges the Agency to
resist this temptation  and select  a point that is  within  the range that can be
detected in toxieological studies that are currently in use  (e.g., the ED10),
primarily because the  confidence intervals  of the estimate increase dramatically
as the estimate decreases, The Agency is reminded that the basis of estimating
risks  from non-cancer effects assumes that there is some  dose below which
there  is;no effect.  Application of uncertainty  factors to the benchmark dose  will

                                     18

-------
then drive estimated "safe levels" in any given media to very low levels,
potentially lower than would be obtained if one were to assume that the LMS
applied.  Therefore, the Agency would find that low benchmark doses will be
quite incompatible with MCLs derived by conventional means. However, if the
benchmark approximates  the response that  is detectable in toxieological studies,
the estimates should not vary significantly from those derived by conventional
means,

3.4   Bioaccumulatioii Issues

      The Drinking Water Committee cautions the Agency that the strategy of
setting AWQC by measuring contaminant concentrations in certain biota and
then applying either a bioconcentration factor (BCF) or a bioaccumulation factor
(BAF) to calculate water concentrations may not accurately reflect the complex
ways in which the real environment operates.  This subject was extensively
discussed in a 1993 SAB report which encouraged the Agency ",.,to continue to
explore these approaches"  (US EPA,  1993).  However, this report also noted
several potentially serious problems  in the use of "field" BAFs, and it is
worthwhile to quote at length from this document;

      Field BAFs must be interpreted very carefully, and it should be
      recognized that they may contain substantial errors and variability
      due to 4he following reasons:

      a) Analytical methodologies generally determine total concentrations
      all of which may not be biologically available;

      b) There may be a loss of analyte by sorption or evaporation during
      sampling;

      c) Incomplete extractions may occur, especially if there is a high
      organic carbon  content in the water;

      d) Temporal and spatial variability in water concentration may
      occur due to season, temperature} depth, hydrology, meteorology,
      and microbial and photolytic activity;

      e) There is likely to be variability in fish concentrations due to size,
      age, sex, season, pre- or post-spawning status, migration, the nature
      of and availability of food, the structure of the food chain,
      differences in lipid  content, parasite infestation and general health
      of the organism.
                                     19

-------
        Given these potentials for error, EPA should discuss and quantify
        the variance in field derived BAFs in its guidance, along with FCM
        estimates, and attempt to identify the magnitudes of natural
        variability and analytical errors in each criterion data base, and
        estimate the impacts on the BCFs and FCMs.

        In many cases, the laboratory-generated BCF data are likely to be
        more analytically accurate, but they may be  less representative than
        BAF, in that they do not reflect natural variabilities, especially on
        food uptake. Therefore, field measured BAFs are suitable for the
        calculation of criteria but with the qualifications that the data must
        be interpreted carefully and all' information  should be exploited.
        Specific guidelines need to be developed for the acceptability of
        residue data in tissues and dissolved concentrations in water.   This
        will likely require a research effort to  determine the appropriate
        sampling procedures, such as the number of organisms per stationt
        the sampling frequency, or filtered/ unflltered water.

        To help alleviate the problem, EPA needs to  support a research
        program to develop more sensitive analytical methodologies for
        hydrophobia chemicals in tissues, sediments  and water.
        Consideration should b& given to the establishment of a formalized
        analytical chemistry program which utilizes  the best scientists, the
        b&st instrumentation, adequate support, etc.,  to develop analytical
        methodologies and perform analyses that are not readily achievable
        by "normal" laboratories.  Support to universities and industrial
        support to develop analytical reference materials would help ensure
        the success of the program.

      The same document went on to caution the Agency that the techniques
proposed  "...have not been applied to enough field conditions to judge if the
predictions are realistic," and to note that they were "...particularly concerned that
metabolism is not included [in $ome of the theoretical modeling]."  Finally, the
document concluded that this "is clearly an area in  which more research is
needed."

      The Drinking Water Committee is concerned  that the Agency may have
misinterpreted the above-cited SAB report as a blanket endorsement of these
techniques for  developing AWQC .before a firm scientific basis for doing so has
been developed. For example, we believe that the approximations proposed to
estimate how much chemical is truly dissolved in water (i.e. available for uptake
by the organism) based on organic carbon content, Log KOW, and "total" content of
the water need to be rigorously validated before they are used to establish AWQC.
                                      20

-------
      As another example, studies which indicate a contaminant has a high
oetanol/water partition coefficient (Kow) or BCF (measured when organisms are
exposed to an artificially-maintained constant water concentration) do not
necessarily indicate the contaminant will bioaccumulate in the environment. If the
contaminant is subject to rapid hydrolysis, photodegradation,  absorption,  or
metabolic transformation, the bioaceumulation potential will likely  never be
realized.

      These criticisms should not be taken as a recommendation to relax
standards or to ignore the potential for bioaccumuiation where it is known to play
an important role.  However, it is the Committee's opinion that AWQC criteria
must be based on sound environmental data and good science with a minimum
number of assumptions. AWQC should not be driven by hypothetical assumptions
that bioaccumulation might occur; they must be based on sound experimental
evidence  that bioaccumulation does occur.  In general, the  Committee feels that,
for the time being, the Agency should focus attention on BCF rather than BAF,
because of the higher likelihood of collecting an adequate BCF database,  BCF can
generally be measured in the laboratory with some confidence, while estimations of
BAF involve incorporation of many as yet untested assumptions.  Properly used,
the BCF can provide a suitable basis for evaluating bioaccumulation potential.

3-5    Exposure Issues
             f
3.5.1  Allocating the RfD

      Before discussing procedures for allocating the RfD,  the Committee would
like to note that the RfD ordinarily describes a region of exposures which have
been chosen to provide high confidence that  human populations exposed to such
doses will not develop adverse effects.  Because of the conservative  way in which
RfDs are calculated, it is unlikely that exposure of any populations  to doses
slightly over the RfD (even up to twice the RfD)  would produce significant health
effects.  Consequently the Committee does not feel that it is appropriate to
develop AWQC geared to ensure that the sum  of all theoretically possible
exposures never exceeds the RfD by even a small amount.

      a)     Should fish criteria be derived assuming that fish intake uses «p all
            of the RfD?
                                     21

-------
      The question is ambiguous.  We assume that the Agency means "Should fish
intake and water consumption be combined in developing AWQC?"  The answer to
this question is that the Committee recommends developing separate criteria based
on fish intake and water consumption (i.e., hypothetically using up all the RfD in
each calculation).

      For those few materials which strongly bioaccumulate in fish (BCF > 1,000)
the human exposure from water consumption is negligible.  For materials which
do not bioaccumulate, the exposure from fish consumption is negligible.  Therefore
the Committee recommends developing separate criteria based on either fish intake
or water consumption.   Normally the  most stringent of the two would determine
the AWQC,  Also, the calculation of two separate criteria might allow the
flexibility of basing the AWQC on water intake rather than fish consumption in
those cases where local  populations are cautioned against eating large amounts of
fish.

      On the other hand, if,the Agency means to ask "Should food, inhalation,
dermal exposures, etc. be considered in setting the  AWQC?,1' the answer is more
complicated.  The Committee recognizes that human exposures to substances
present in fish may occur through other routes of exposure (i.e., in other foods or
drinking water).  However, we are concerned that if the Agency attempts to
compensate for other routes of exposure (either by subtracting other inputs or by
allocating only a certain percentage of the RfD for  water exposures) they may
focus  intense regulatory attention on insignificant problems,  thus wasting scarce
resources that should be available for  more significant health risks.

      This is particularly true when reliable data regarding other possible routes
of exposure are not available.  In this case, the Agency should not use "defaults" or
"high  end estimates" of exposure. Use  of defaults or high-end estimates may resvlt
in very conservative criteria being established for water without any real
indication of risk.

      Consequently the Committee  recommends that unless reliable data are
available which indicate that human populations actually encounter total exposures
significantly in excess of the RfD, exposure from other routes should be neglected
in calculations of AWQC.

      b)    How should the RfD be allocated? Should the percentage, subtraction,
            or some other approach be used? Should a 20% floor and 80% ceiling

                                      22

-------
            be used? Should a default of 20% be used when there are inadequate
            data to quantify total background exposure? If not, how could
            anticipated sources of exposure be accounted for when data are
            inadequate to quantify total exposure?

      The Committee feels that apportionment of exposure sources for
development of AWQC can only be attempted when reliable exposure data are
available. In the absence of such data, arbitrary "defaults" or "high end estimates
of consumption" should not be used and the AWQC should be based on water
sources only (as noted in 3.5,la above), Even, when reliable exposure data are
available, the Committee feels that the "percentage" or "subtraction" methodology
can easily result in the misallocation of critical resources to address  insignificant
risks.  Instead of using either the "percentage" or "subtraction" methodology, the
Committee recommends the approach outlined below.

      If the exposure data indicate that total exposures are well below the RfD,
there is no problem and the Agency's goal would be simply to develop  criteria to
ensure that a problem does not develop in the future.  Adequately protective
AWQC could then be developed by assuming that all of the RfD is available for
the water route.  By using this approach,  human populations would only be
exposed to levels slightly above the RfD, even in the unlikely event  that water
input increased to the level permitted by the standard, because the other routes
contribute relatively small amounts to the TOTAL exposure.

      Inherent in  this recommendation is the assumption that exposure data will
be periodically updated to ensure that total 'exposures  do not increase significantly
above the RfD due to sudden increases in other sources of exposure. On the other
hand, if total exposures are at ©r higher than the RfD, then remedial actions may
need to be considered.  When exposure data indicate that each of several sources
of exposure contribute significant fractions of the RfD (e.g., greater than 25%)
then one is faced with a multi-media control  problem.  Even in these cases,
however, it may not be reasonable to develop AWQC if "background sources" which
are beyond human control  are the major contributors  to human exposures,
particularly if the  RfD has been derived with a substantial safety factor (e.g.,
cadmium, mercury, radon).

      Furthermore, when multiple routes  of human exposure exist the  Agency
needs to consider the relative cost effectiveness of control efforts in  other media.
If total human exposures can  be reduced below levels of concern with two or more

                                     23

-------
alternate strategies, it makes sense to choose the least expensive approach.
Consequently, this Committee endorses the recommendation from the AWQC
workshop held by the Agency in  1992:

      "When significant contributions are expected from multiple sources, [and the
      total of those contributions exceeds the RfD] then bring into the discussion
      the offices or agencies responsible for addressing these other sources to
      allow for the development  of an integrated management decision based on
      the relative source contributions, and the ability to control exposures from
      the various sources.  Include in these multimedia exposure discussions both
      noncarcinogenic and carcinogenic endpoints for the chemical of concern."

3.5.2  Exposure Route, Rate, and Duration Issues

      a)    Should there be separate criteria for drinking water and fish intake: a
            water column concentration to protect water consumers and a fish
            tissue concentration to protect fish consumers?

      The Committee suggests that AWQC criteria for drinking water and fish
intake be  established independently (see 3,5.la above). When the two criteria are
not identical, then the lowest (most stringent) criterion should apply.  If the two
criteria  are identical (would result in  equal exposures), then the Agency  might
consider reducing AWQC by a factor of two.

      b)    Should the methodologies for developing MCLGs and AWQC for
            drinking water be consistent?

      The Committee foresees  considerable difficulty in using the concept of
MCLGs in the development of AWQC. There are several reasons for this:

      (1) The MCLG for carcinogens  is arbitrarily set at zero. Introducing this
      concept into the AWQC is  likely to confuse the public, distort the  relative
      importance of carcinogens (versus untested contaminants), and  result in the
      misdirection of resources if applied to the permitting process (e.g., PCBs,
      dioxins  and diben2ofurans, trichloroethylene, tetrachloroethylene),

      (2) The MCLG approach apparently underlies the Agency's decision to
      continue the  practice of adding a safety factor for the presence of  "weak"
      carcinogenic concerns (i.e.,  for certain Category C chemicals).  This use of

                                      24

-------
      the MCLG approach is not scientifically sound, however, because it
      incorrectly mixes two classes of effects; and there are sounder ways to
      achieve the desired end with such chemicals.  For example, carcinogenic and
      non-carcinogenic effects might be considered separately, and the more
      stringent limit from appropriate estimates for each endpoint could be
      selected, or the two effects might be considered jointly, although this would
      involve a major departure from current practices.

      (3) Concerns addressed in setting AWQC are considerably different from
      those applicable to drinking water rules.  The differences include the
      possibility that AWQC will be based on bioaccumulation of chemicals in fish,
      the need to set criteria based on aquatic species, and the need to protect
      certain uses of the water  (i.e., not setting an AWQC for a disinfectant by-
      product which subsequently precludes use of the disinfectant responsible for
      generating it in a downstream drinking water supply),

      On the other hand, the derivation of the RfD (or potency factors for
carcinogens)  used to set the AWQC should be the same as the derivation of the
RfD used to  set the  MCL.  As data are updated in one  program, they should be
updated  in the other.  Nevertheless,  determination of how these numbers are used
to set AWQCs or MCLs involves risk management as well as risk assessment
decisions. For these reasons the Committee considers that it may be reasonable to
develop an AWQC different from a MCLG in certain specific cases.

      c)     Should  "one day" criteria using realistic single meal consumption rates
             be developed to address potential acute effects (including reproductive
             and developmental  effects) from consuming contaminated fish?
                             \
      The Committee notes that current "average fish consumption" rates of 6.5
g/day may not be protective for  potential acute effects associated with a large
single meal.  In the average American diet, we suspect that exposure Is
discontinuous, occurring at weekly or monthly intervals. We believe; therefore,
that criteria for acute effects may need to be developed when it can be reasonably
anticipated that single exposures may induce adverse effects in humans. The
Committee cautions the Agency, however, to be certain  that these criteria are not
applied to adverse health effects which are not reasonably anticipated to result
from single, acute exposures, and notes that public warnings may also be used to
address special circumstances (e.g., pregnant women should avoid consuming large
amounts of fish contaminated with developmental toxicants)

                                      25

-------
      d)    Should there be separate criteria for incidental water intake from
            recreational use for situations where drinking water criteria are not
            used (estuaries) and fish ing«stion or aquatic life criteria do not
            protect recreational users from incidental ingestion.  For example, the
            existing fish ingestion criteria for phenol is 4,600 mg/L, while the one
            day drinking water health advisory is 6 mg/L, To  prevent potential
            acute health risks from incidental recreational ingestion, an  incidental
            intake rate (e.g., 0.01 L/day) could be used with acute toxieify data
            (e.g., one  day drinking water health advisory value multiplied by
            100?).

      The Committee  feels  it is unlikely that AWQC established to protect aquatic
life in an estuary would permit contaminant concentrations high enough to
represent an acute hazard to humans incidentally ingesting such a small quantity
of water.  In the example given (phenol) we feel it is improbable that 4,600 mg/L
would not be deleterious to  aquatic life and hence regulated on  that basis. We feel
that circumstances in which a separate criterion for  incidental ingestion would be
necessary are so rare that they do not .warrant serious consideration by the
Agency.

      e)    Should  water and fish intake assumptions be developed on a per
            kilogram body  weight basis to use the actual body weights of survey
            respondents and to avoid the use of default body weight assumptions?

      In theory it would be better to develop standards on a per kilogram body
weight basis. However, in practice the results are not different  enough to make
much difference in the magnitude of AWQCs, In particular, data should not be
rejected because individual body weights are not available, and funds should not be
allocated for collecting such data since no conceivable benefit would accrue.

3,5.3  Fish Consumption Rates

      a)    Is  it more defensible to include fish consumption in quality criteria
            with or without an assumed body weight? That is, should fish
            consumption rates be normalized to body weight or should we
            continue to assume an average body weight of 70 kg?

      This question appears to be a duplicate of Question 3,5.2e above, and the
same  answer applies,

                                      26

-------
      b)    Is it technically defensible to use upper percentiles of national
            consumption surveys to represent special subpopulations such as
            recreational and subsistence fisherman ethnic groups?

      It is very difficult to account for the differing fish intakes of different
populations through the development of AWQC focused on point source discharges.
A central issue is that fish in a particular body of water are almost always
impacted by a variety of sources. In many cases, the major sources are not subject
to the permitting process  (PCBs, Dioxins, Dibenzofurans),  Consequently, it makes
little sense to try to deal with recreational and subsistence fishermen on the basis
of an AWQC.

      The Committee feels that the best way to protect subpopulations with high
fish consumption is to base health standards on the levels of chemical that are
found in fish, not in effluents.  Under such an approach, state and local
authorities decide upon the safe levels of intake  of particular chemicals and
translate them into guidelines for the consumption of fish types that can be safely
consumed from given bodies of water,  .

3,6   Microbiology Issues

      a)    Would the establishment of a multi-organizational working group to
            provide recommendations to the EPA on technical considerations of
            this effort be appropriate to enhance progress on ambient water
            microbiological quality?

      The Committee believes that  the establishment and implementation of such
a working group with representation from EPA,  The Centers for Disease Control
and Prevention (Cl3C), the Food  and Drug Administration (FDA), academia, the
water and wastewater industry, and the public would be highly beneficial.  This is
because the scientific and  technical  issues for the microbiological contaminants of
human health concern go  beyond the artificial jurisdictional boundaries of the
regulatory and other federal agencies that deal with pollution sources, water
pollution control, treatment of water and wastewater, fisheries resources and
public health.  Indeed, the development of such a multi-organizational working
group would acknowledge  this fact.  Hopefully, such a working group would
address microbiological contamination of ambient waters on the basis of the
pollution types and sources, their transport through (and proliferation in) the
environment, their partitioning among environmental compartments (e.g. water

                                      27

-------
column, sediments, fish), and their potential for human exposure via various
water-related routes.  Only by having expertise to deal with the science and
technology of these phenomena, as well as the health effects and risks from such
exposures, can  microbiological contaminants in ambient waters be addressed in a
rational and effective manner.

      b)    Should ambient water quality criteria for microbes be developed for
            recreational, drinking sources,  shellfish harvesting, reuse, irrigation,
            wetland  and groundwaters?  What should be the relative priority for
            establishing health criteria and monitoring requirements for these
            classes of water?

      There is clear justification and rationale for keeping recreational water
quality criteria as a high priority. No other regulatory program or  legislative
mandate provides a clear basis for developing microbiological criteria for such
waters, and adverse health effects have been documented by disease outbreaks and
in epidemiologie studies.  The Committee believes that despite the desirability of
and need  for a comprehensive and integrated approach to ambient water quality, it
is unrealistic, perhaps inappropriate and in all likelihood  impossible to address all
of these water-related exposure routes of microfaial health effects concern under
this regulatory initiative.   Historically, the ambient water quality criteria for
microbes of human health concern have dealt primarily, if not exclusively, with
recreational bathing waters.

      In contrast to recreational waters, microbiological criteria for  drinking
source waters> ground waters and shellfish harvesting waters are or should be
addressed through other regulatory programs, including the Safe Drinking Water
Act (drinking source waters and ground waters) and the.Food, Drug and Cosmetic
Act (shellfish harvesting waters).  Specifically, microbiological issues and criteria
are being addressed through such drinking water regulatory initiatives as the
Coliform Rule,  the Surface Water  Treatment Rule (SWTR),  its revision as the
Enhanced SWTR, the Ground Water Disinfection Rule, and the Disinfection
By-products Eule, The microbiological quality of shellfish harvesting waters is
addressed through the FDA's National Shellfish Sanitation Program and its
agreements with  the Interstate Shellfish Sanitation Conference, the  EPA and the
National Marine  Fisheries Service. EPA is  or should be a party to these
agreements.  What may be lacking are the communications and interactions
between these various parties and their representatives in the scientific
community. There should be coordination of activities among these  programs and

                                      28

-------
agencies, and the establishment of a multi-agency working group, as proposed in A
above, could greatly facilitate such interaction and coordination.  This might help
to prevent unwitting duplication of efforts, needless and wasteful competition for
scarce resources, conflicting methodology for developing and implementing
microbiological criteria, and unnecessary jurisdictional disputes.

      The question of including AWQC efforts for reuse, irrigation and wetlands is
not only difficult to address but also difficult to justify in the context of this
regulatory effort.  This is because human health effects from microbiological
agents transmitted by these exposure routes are likely to be either  negligible, too
low to be doeumentable, or highly variable,  For example, irrigation water quality
criteria would be extremely difficult to  establish because of the spectrum of
potential quality requirements.  Water  used for irrigation may come from either
ground or surface sources, may be subject to diverse and highly variable sources
and amounts of human or animal fecal contamination, and may be  subject to  other
quality criteria defined by the intended irrigation use.  With respect to the last
point, for example, microbiological criteria may differ for irrigation  of fruits and
vegetables to be eaten raw, fruits and vegetables that are cooked or processed
before eating, non-food crops  (fodder and horticulture plants), greenspace and
forests. Because of the poorly documented, complex and apparently low exposures
from these sources,  the inclusion of irrigation water does not appear to be
adequately justified  at this time.  There are analogous concerns  about reuse and
wetlands waters. Without better justification of potential exposure  sources and
transmission routes, the Committee recommends that these classes  of water not be
included in  the effort.

      c)    Should ambient water criteria for microbes be directed only at
            microbes causing gastrointestinal diseases or be expanded to cover
            other diseases?

      The Committee recommends that the process of developing and evaluating
water quality criteria for microbes should include microbes causing  feeally-
transmitted diseases other than gastroenteritis, and also  include microbes causing
diseases of the skin, respiratory tract, eye, ear, nose, throat and perhaps other
sites of entry and infection.  The criteria development effort should also consider
those animal pathogens that potentially infect humans, such as Cryptospondium
and Giardia.  The reason for inclusion  of these pathogens is that health effects
from such pathogens, including illness from aquatic exposures, have already beer,
documented in the scientific literature (Seyfried et a.L, 1985; Sobsey et al,f 1993),

                                       29

-------
The Committee recommends that hazard identification for these other pathogens
be included in the methodological process for  development of the ambient water
quality criteria.

       d)    Are the ambient recreational water quality criteria appropriate and
            trausferrable to other high priority ambient waters?

       The current ambient recreational water quality criteria are neither
appropriate for nor transferrable to other ambient waters.  The current criteria in
the 1986 guidelines are based on densities of enterococci and E. coli, and the
earlier criteria ("red book")  are based on fecal coliforms.  These indicator criteria
were intended to address only those pathogens causing enteric (gastrointestinal)
illness. They do not address the extra-enteral pathogens and their illnesses.
Furthermore,  the effectiveness or validity of these  indicators to reliably predict
health effects  even of gastrointestinal illness has been questioned as misguided and
has been challenged on the basis of contrary findings in epidemiologic  studies of
recreational water quality (Seyfried et al,, 1085; Cartwright, 1993).

       e)    Should  the stringency for protection of public health be linked to the
            likelihood of human exposure to different types of ambient water?

       The Committee recommends that the likelihood of human exposure to
different types of ambient water be the basis  for identifying the types of ambient
waters for which criteria  need to be developed, as discussed in d) above.  The
potential for human exposure and the risks of exposure  are likely to differ greatly
for the different ambient  waters identified in  f) below.  The case for quality
criteria for recreational waters has been established on the basis of documented
exposures and health effects (illness).  However, the case for exposures and illness
from some of  the "other waters has not  been established.  Furthermore, some  of
the other ambient waters from which there are potential mierobial exposures  are
and should be addressed through other regulatory programs, as noted in the
response to b) above.

       f)    Is it appropriate to develop risk-based health criteria for pathogenic
            microorganisms in ambient waters?

       The Committee believes that a risk-based approach to criteria for pathogenic
microorganisms in ambient  waters  is both appropriate and feasible for at least
some pathogens.  This reflects the  apparent consensus on this approach within the

                                       30

-------
scientific community (Sobsey et al.} 1993), Indeed, this approach has already been
used for drinking water, has been attempted for sludge, and is being explored for
shellfish and land-applied human excreta used in agriculture,

      However, the Committee believes that there are limitations to the
applicability of this approach to the quality criteria for microbial pathogens in
ambient waters.  Specifically, the Committee recommends that, initially, this
approach can and should be directed only to those pathogens that are known to
occur in and cause  health effects from ambient recreational waters and for which
dose-response and epidemiologic data are available,  Examples are rofcaviruses,
adenoviruses and hepatitis A virus,  The  Committee believes, however, that there
are major impediments to the implementation of this approach to the risk
assessment of many pathogens, because of the lack of dose-response data and
             evidence of clear health risks from exposures  via ambient waters*
      g)    Can strain or species differences in virulence be determined by
            monitoring?

      The Committee believes that this question is difficult to answer in
scientifically proven terms because of the lack of knowledge about virulence
determinants, their expression by many of the potential waterborne pathogens, and
their ability to be measured readily.  The dearth of such data is important  but not
unique to the 'ambient water quality methodology initiative.  Such data are  also
needed for other  environmental  routes of pathogen exposure, such as drinking
water.  For this reason, the Committee believes  that further research has to be
done on the identification, characterization and measurement of the virulence
determinants of microbial pathogens and on the factors governing or influencing
the expression of these determinants under different environmental conditions.
Such information is needed before the question about monitoring for microbes on
the basis of virulence can be answered effectively. Furthermore, the role of other
factors in virulence expression, such as host factors, also must be addressed fay
research.  The Committee recommends that this question be given a relatively low
priority for ambient water quality criteria until the necessary research data are
obtained,

      h)    How should EPA deal with injured organisms in risk assessment?

      The Committee believes that this question identifies an important research
need that goes beyond the scope of this initiative on the microbiological  criteria

                                      31

-------
for ambient water quality.  Presently, there are no microbial regulations for any
waters that adequately consider the role of injured organisms.  All of the current
criteria and standards for microbes in water are based on concentrations of
bacteria estimated by cultivation methods.  Some efforts "nave been made to
include "resuscitation" steps during cultivation (e.g., lower incubation temperatures
for brief periods) and to use culture media that are less stressful or  inhibitory to
injured organisms.  However, the health significance of injured  pathogenic
organisms remains uncertain.  Some recent studies in the scientific literature
indicate that injured organisms have  dramatically reduced infectivity in terms of
the dose of cells needed for infection  and the potential to produce disease (as
opposed to sub-clinical infection)  (Jones et aL> 1991; Medema et al.» 1992),  Hence,
the public health significance of injured pathogens remains uncertain.  For this
reason, the Committee recommends that this  issue is presently  beyond the scope of
this methodological initiative for  ambient water quality criteria, and  therefore is a
low priority issue for this effort.  Overall,  it is a high priority question for the
public health aspects  of water microbiology, but the research needs go well beyond
the domain of this initiative.

      i)     Are current molecular techniques for pathogen detection/identification
            promising enough to push for their rapid development for routine
            water monitoring; can such techniques readily be made  quantitative
            and capable of discriminating dead from viable organisms?
             /

      The Committee believes that current molecular techniques for  pathogen
detection/identification in water and other environmental samples are still in their
early stages of development, they lack adequate quantitation, and there is still the
unresolved question of whether or not they can distinguish between live and dead
(infectious versus non-infectious)  organisms (Enriquez et a/., 1993; McCurty and
Atlas, 1993).  The research needs on these issues are substantial and as yet
unmet.  The Committee recommends  that  such research be done because of the
enormous potential of these detection methods. However,  as  with some of the
other issues raised in the workshop report and BPA's response, this  issue is
broadly relevant to microbes in a variety of waters and other environmental
samples.  Hence, it goes well beyond the scope of the ambient water  quality
initiative, and for this reason, the Committee  recommends that  it should be of low
priority in the con-text of this regulatory effort.
                                      32

-------
      j)     Are the currently approved indicator organisms and acceptable levels
             for them in beach waters appropriate for determining the safety of
             waters against gastrointestinal (GI) disease?

      The Committee believes that the currently approved indicator organisms  in
beach waters are probably appropriate for determining the safety of bathing
waters against GI disease in at last some situations or settings. From the
historical record of waterborne disease from bathing waters and the bacteriological
quality  of such waters, it is clear that fecal coliforms, E. coli and enterococci have
the potential to indicate fecal and sewage contamination of bathing waters.
Epidemiological studies at both marine and freshwater bathing beaches in North
America have documented the effectiveness of these indicators in predicting risks
of GI illness from bathing waters.

      However, the Committee  believes there are serious and justifiable concerns
that the GI health risk data predicted by the existing indicators, and the
standards for them can not'be generalized to ALL bathing waters.  Some of the
reasons for this concern is the well-documented variability in the levels of
indicator bacteria  in sources of fecal  contamination, the potential  for some of
these indicators to arise from non-fecal sources, and the variability in the
quantitative relationships between these indicators and various GI pathogens in
the sources of fecal contamination (Cartwright, 1993;  Fleisher, 1991). For
example, there is the variability in indicator bacteria densities and pathogen
densities created   by treatment and disinfection of sewage. In some  situations
treated  effluents contain high levels of indicator bacteria because they are not
disinfected and in other situations they contain low levels  of indicator bacteria due
to disinfection.  Because  of the comparative resistance of  some of the important
GI pathogens, such as enteric viruses and protozoan cysts, to wastewater
treatment and disinfection, there is no consistent relationship between densities  of
these indicators and the densities of these GI pathogens.  Hence, in some cases
ambient bathing waters impacted by effluents will contain  low indicator levels but
relatively high levels of GI pathogens. The Committee believes that  the currently
accepted levels of the bacterial indicators are not uniformly and adequately
protective of health risks from GI pathogens in bathing waters.

      k)    Are the currently approved indicators predictive of other human
            diseases from fecal or other anthropogenic sources and from animal
            sources?
                                      33

-------
      The Committee believes that these indicators are not likely to be adequately
predictive of the human health risks from non-GI illnesses associated with human
or animal fecal contamination, Epidemiologic studies in other countries, including
Canada and the United Kingdom, show that there are documented health risks
from non-G! illnesses in bathing waters subject to fecal contamination, and that
these risks  are not reliably predicted by the fecal indicator bacteria now
recommended by EPA (E. colt and enterococci) (Cartwright, 1993; Fleisher, 1991).
The Committee recommends that in this initiative EPA consider the alternative
and newer epidemiological data on bathing waters from such countries as Canada,
the United Kingdom,  Israel, South Africa, Hong Kong and other studies.

      1)     Are there alternative indicator systems that warrant investigation for
            improving ambient water monitoring?

      The Committee believes there are candidate alternative indicators  worthy of
consideration and deserving of investigation in this initiative.  The Committee
believes that some of the examples mentioned in  the workshop report  and EPA*s
response are indeed worthy candidates-for consideration.  These include
bacteriophages such as eoliphages, Clostridium perfringens and rainfall events.
The Committee recommends that EPA investigate the initiatives on bathing water
quality indicators in other countries and  geographic regions, such as that by the
European Community.

      m)    Is there a justification for developing different indicator organisms to
            monitor microbial pollution of tropical waters?

      Based on the limited studies now available, the Committee believes that the
validity and usefulness of traditional or standard  bacterial indicators of fecal
contamination in tropical ambient waters is uncertain. While  some studies
indicate that coliforms, fecal coliforms and even E, colt and enterococci are
ubiquitous in tropical ambient waters, the amount, quality and representativeness
of these data are limited. The Committee believes that more and better studies
are needed on the microbial ecology or natural history of these and other
candidate indicator  bacteria in tropical ambient waters.  The Committee concludes
that present information is too limited to  draw firm conclusions about  the
usefulness and reliability of current indicators to  monitor fecal contamination in
tropical waters.  The  Committee recommends further  investigation of this issue by
EPA.
                                      34

-------
      n)    Are there any significant issues related to ambient water microbial
            disease not covered by the Workshop report and EPA's response?

      The Committee believes that the situation is quite the contrary: there are
too many issues and too broad an agenda raised by the Workshop report and
EPA's response.  The Committee recommends that EPA prioritize the issues and
focus the agenda on the  most crucial topics that are not covered by other
regulations.  The Committee recommends that EPA focus primarily on ambient
recreational waters, These are the waters that have been historically addressed,
there are clear exposure potentials and documented health effects, and
furthermore, some of the other waters  (drinking sources, shellfish harvesting and
reuse) are covered by other  regulations,  For these reasons, the Committee
recommends that efforts be  made to integrate or coordinate the risk assessment
for pathogenic microbes in ambient waters to the risk assessments being done for
other waters, such as drinking water and shellfishing waters*

3,7   Minimum Data Requirement Issues

      a)    Is the use of a  tiered approach reasonable?

      The use of a tiered approach is reasonable. It must be made absolutely
clear to the  risk assessor and risk manager, however, that the categories are based
on the availability of certain data elements, and that they do not reflect the
quality of the underlying research.  Obviously, the toxicity of the compound is also
independent of the category.

      b)    As presented, do the individual tiers reflect minimum requirements
            necessary for wate^r quality criteria and interim effluent permit limit
            development,

      It  is quite appropriate that Tier I should  include mechanistic,
pharmacokinetic and target organ toxicity data.   The statement that' for chemicals
in Tier II there should be enough data  to generate an  RfD or cancer potency
factor makes sense from the standpoint of having to have some breakpoint, but
this concept is very poorly defined.  The  question is, if there was one reasonably
good 90 day study, which included a NOABL dose level in one species, would a
chemical qualify for Tier II?
                                      35

-------
      Tier III  presents some problems.  The minimum of a 28-day study appears
reasonable.  The Committee agrees with the EPA that the purpose of this category
is to generate some number for regulatory purposes while at the same time
providing an impetus for additional research. The  Committee is concerned with
the possibility that whatever value was derived might de facto become permanent
and strongly suggests that its temporary character  be emphasized by the Agency,
More and better data should be quickly rewarded.  That is, the  results of a
complete and well conducted 90-day study  might very well indicate that  the
chemical is not  as great a concern as suggested by a 30-day study.  In this case it
would be appropriate for EPA to increase the value of the AWQC, on the basis  of
better data and less uncertainty. The question of what is done  with  group C
carcinogens is difficult because it is not clear what  the Agency means by
"insufficient data," If a study were carried out long enough to convincingly
demonstrate  carcinogenicity in a laboratory animal  species, it is  difficult  to imagine
that there would be insufficient data either in that  study or in other  studies  on
that particular compound.  The  Committee recommends that EPA carefully
evaluate whether or not this is a real problem.   The Committee also recognizes
that the new EPA Guidelines for Cancer Risk may  be different and not have the C
category.

      Tier IV guidelines are also reasonable, as is the intention  not to use these
for  standard  setting.  Obviously, if there are data that suggest potential hazards
this information is important to communicate.  Tier V guidelines are  also
reasonable,

      c)    Under what specific circumstances should a Group C chemical be
            categorized under Tier III?  Are there  circumstances in which
            inadequately tested chemicals  (currently categorized as Tier  HI)
            should be placed in Tier  II in  order to  develop water quality criteria?

      There  are two issues involved here.   First, there is no apparent reason that
group C compounds should be handled differently from those in  other categories.
For example, if there  are good data on endpoints other than  cancer, and  perhaps
even a good carcinogenicity study demonstrating positive results in one species,
there is no reason.why the chemical should be a Tier III.  Alternatively,  if the
data base is poor, a Tier III categorization  would be appropriate, although, as
stated in part B above, the likelihood  of this is open to some question.
                                      36

-------
      The purpose of the tier system  is to categorize compounds on the
availability of information in their toxicity data bases, so that regulatory decisions
can be made.  It would be inappropriate, therefore, to move a chemical to a
category where it does not belong so that a regulatory decision can be made.
From the standpoint of public health, the Committee is of the opinion that
allowing the states  to use Tier III values to set permanent permit levels would be
very risky.

      d)    What specific requirements can be place on the use of 28-day study
            data in order to ensure the best quality data is used in developing
            Tier III interim toxcity values?

      This is  a judgment call.   Study design factors such as the number of
animals used (and hence the ability to detect an effect), the number of dose groups
and the spacing of the dose groups are  certainly important. While two species
may not be necessary, they are certainly a plus. To list a whole host of clinical
pathology tests (cell counts, serum enzymes, electrolytes, glucose values, etc.)
would be unwise. If one already knows the target organ, evaluation of parameters
associated with the functioning of that organ can be much  more important  that a
whole host of negative findings in other organs. Consistency of findings, both
within a study and  among studies, also  has merit.   There should be concern if a
study does not establish a NOAEL although this might not be considered
absolutely esential if there is a good dose response study such that a benchmark
dose could still be established.   In other words, the results of a well conducted
study that does not establish a NOAEL, but which  establishes a good dose-
response relationship, utilizes adequate  numbers of animals, etc., may be of higher
quality and engender greater confidence in the  resulting estimates than the  results
of a poorly conducted study which "finds" a NOAEL primarily because it wasn't
truly adequate in these  same parameters.
                                      37

-------
                               4. REFERENCES

Cartwright, R.Y, 1993, Diseases associated with the use of recreational waters;
assessing their relationship to microbiological parameters, Wat. Sci. Tech., 27(3-4);
195-198.

C&E News. Chemical and Engineering News. Vol. 71(9):1, March 1, 1993,

Crump, K.S, 1984, A new method for determining Allowable Daily Intake,
Fundam Apol  Tox 4:854-871,

Enriquez, C.E., M. Abbaszedegan, I.L. Pepper, K.J. Richardson and C.P, Gerba,
1993. Poliovirus detection in water by cellculture  and nupic acid hybridization
Wat. Res. 27(7):1113-1118.

Fleisher, J.M.  1991. A reanalysis of data supporting US federal bacteriological
water quality criteria governing marine  recreational waters.  Research Journal
Water Pollution Control'Federation.  63-;259-265.

Jones, D.M., E.M. Sutcliffe and A. Curry. 1991. Recovery of viable but
nonculturable  Campvlobacter JejunL  Jour. Gen. Microbiol 137(10):2477,

McCurty, S.C/and R.M. Atlas. 1993.  Effect of chlorine disinfection on Legionella
pneumophila detection using polymerase chain reaction with varying size
amplicons.   Abstract Q8 in Abstracts of the 1993 Annual Meeting. Amer. Society
for Microbiology, Washington,  DC.

Medema, G.J., F,M. Schets, A.W. van de Giessen and A.H. Havelaar. 1992. Lack of
colonization of 1 day old chicks by viable non-eulfarable Camphylobacterjejuni. J.
AppL Bacteria: 72:512-516.

Seyfried, P.L., R.S. Tobin, N.E. Brown and P.P. Ness. 1985.  A prospective study
of swimming related illness. I and II.  Am. J. Public Health  75:1068-1075.

Sobsey,  M.D., A.P.. Dufour, C.P. Gerba, M.W. leChevallier and P. Payment 1993.
Using a conceptual framework for assessing risks  to health from microbes in
drinking water, J. Amer. Water Works Assoc.. 85(3):44-48.
                                      38

-------
US EPA. 1980. U,S. Environmental Protection Agency, Fed.Reg.. 45:231.
November 28, 1980.

US EPA. 1993. U.S. Environmental Protection Agency, Science Advisory Board.
1993, "An SAB Report;  Evaluation of the Guidance for the Great Lakes Water
Quality Initiative." EPA-SAB-EPEC/DWC-93-005,  Chapter 5.
                                    39

-------
                            DISTRIBUTION LIST

Administrator
Deputy Administrator
Assistant Administrators
EPA Regional Administrators
EPA Laboratory Directors
Director, Office of Ground Water and Drinking Water
Director, Office of Science and Technology
Office of Modeling, Monitoring Systems, and Quality Assurance
Director, Environmental Monitoring and Assessment Program
Environmental Monitoring Management Council
Director, Risk Assessment Forum
EPA Headquarters Library
EPA Regional Libraries
EPA Laboratory Libraries

-------