-------
                            APPENDIX D

                       Description of Derivation
                      of Short-Term Action Levels
SOURCE:  This information is included in this report to illustrate various
           rationales for deriving short-term action levels. The authors do
           not necessarily endorse the specific values given for application
           to Superfund sites.

-------
SB
                UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                      'Office of Air Quality Planning and Standards
       /             Research Triangle Park, North Carolina 27711


                                FEB   9 1993

Mr. Bart Eklund
Radian Corporation
P.O. Box 201088
Austin, Texas  78720-1088

Dear Bart:

     Please find the enclosed draft document, "De minimis Levels for Section
112(b) Hazardous Air Pollutants."    This document contains descriptions of
exposure limits including their intended use and how they were developed.

      As I re-examine this document in the context of our conversation, I
am not convinced that it contains  any information you do not already possess.
However, in an attempt to avoid any duplication of effort, I am sending
you a copy.  Please understand that the document does not represent the
Environmental Protection Agency policy, and that it is a draft which
has not been reviewed for accuracy by the Agency.  Feel free to call me at
(919) 541-2962, if you have  any questions.   I appreciate the opportunity
to be of service and trust this information will be helpful to you.
                                            incere
                                          Kelly Rimer
                                    Pollutant  Assessment Branch
 Enclosure

-------
            DE  MINIMIS LEVELS
                    FOR
SECTION 112 (b)  HAZARDOUS  AIR POLLUTANTS
                Submitted to:

 Office of Air Quality Planning and Standards
     U.S.  Environmental Protection Agency
          Research Triangle Park, NC
                Submitted by:

      Clement International Corporation
               9300 Lee Highway
              Fairfax, VA   22031
              September  30,  1991

-------
                                                                   ** DRAFT **

                              TABLE OP CONTEXTS


                                                                          Page

INTRODUCTION  	    1

PERMISSIBLE EXPOSURE LIMITS	-.    3

RECOMMENDED EXPOSURE LIMITS  	    6

EMERGENCY RESPONSE PLANNING  GUIDELINES  	    8

EMERGENCY EXPOSURE GUIDANCE  LEVEL 	   11

LEVELS OF CONCERN	   14

CAPCOA NONCANCER ACCEPTABLE  ACUTE EXPOSURE LEVELS  	   17

OTHER APPROACHES

      Calabrese  and Kenyon Ambient Air Level Goals   	   19

      Minnesota  Pollution Control Agency Regulatory  Limits for Short-
            Term Exposures	   22

      ACGIH Threshold Limit  Values  	   25

      The State  of Maryland  Screening Levels   	   27

      Potency-based Method for Acute Toxicity  	   30

SUMMARY	    33

TABLES

      Table 1.   Summary of Characteristics of  Short-Term  Exposure  Levels    44

      Table 2.   Hazardous Air Pollutants  --  Exposure Limits  Established
            by Other Organizations  	    ^8

-------
                                                                   ** DRAFT **

                                 INTRODUCTION


      The Clean Air Act Amendments of 1990 requires the promulgation of
 emission standards for major sources and area sources of hazardous air
 pollutants listed in Section 112(b).  Under Section 112(g),  any physical
 change or change in the operation of a major source that results an actual
 emission increase for that source exceeding a de minimis level requires the
 revision of the permit for that facility.  Consequently, EPA must determine
 what constitutes a de minimis level for each of the 190 hazardous air
 pollutants (HAPs) listed in Section 112(b).

      One approach to setting de minimis values is to determine a health
 benchmark for each pollutant, and to identify the de minimis emission rate
 that would ensure that the benchmark is not exceeded.  EPA has determined that
 both long-term and short-term benchmarks may be needed.  Long-term levels are
 protective for chronic health effects including cancer, while short-term
 levels ensure that there is negligible concern for acute effects.  For long-
 term effects, benchmarks can be readily derived using the Agency's cancer
 potency estimates and reference concentrations (RfCs).   At this time, the
 Agency has not established methodologies for evaluating short-term exposures.
 A project has been initiated by the Agency to develop a method for short-term
 RfCs,  but this effort will not be complete for a number of years.
 Consequently, there is a need to determine whether there are interim
 approaches to setting short-term benchmarks that could serve the immediate
 needs of Section 112(g).

      This report focuses on issues to be examined when determining how short-
 term de minimis levels can be developed.  These issues include: what is a
 short-term level in terms of duration and magnitude, what are the acute
 effects they should protect against, what populations should be protected,
 should technical and economic feasibility be considered when determining de
 rainimis levels,  and what have other Federal, state and private sector
 organizations already accomplished.

      This report considered short-term exposure to be 1 hour or less:
however,  a review of the various approaches discussed will show that this
 exposure limit is not commonly used among regulatory and advisory agencies.
Worker health has been the focus of several organizations, such as the
 National Institute for Occupational Safety and Health,  the Occupational Safety
 and Health Administration, and the American Conference of Governmental
 Industrial Hygienists.  These organizations average worker exposure over an  8-
hour day to derive a time-weighted average (TWA) exposure level and assume a
40-hour workweek and a 40 year working lifetime.  It is understood that the
TWA may be exceeded during the workday but that the average concentration  (the
 threshold limit value [TLV],  permissible exposure limit [PEL], or recommended
exposure limit [REL]) should not be.  To protect workers in the event of a
sudden release,  other exposure limits are also used.  These include a 15-
minute short-term exposure limit (STEL), which should not be exceeded for  more

-------
                                                                   ** DRAFT **

than 15 minutes in an 8-hour  day,  a  ceiling value (TLV-C),  which snrr_d never
be exceed for any length of time,  and  an immediately dangerous to life or
health (IDLH) value,  which, if  exceeded  for more than 30 minuces,  may cause
irreparable damage to health  or death  by preventing the  escape of the exposed
worker.

      Although TLVs,  PELs, and  RELs  are  the most commonly  used values for
worker exposure limits,  other organizations have developed their own
methodologies and time limits to protect the general public from toxic
emissions from local  facilities.   California has developed a one-hour	
noncancer acceptable  exposure limit  derived from several other methods.
Maryland uses both a  one-hour and  an eight-hour exposure limit to protttct the
general public, and the  American Industrial Hygiene Association uses three
one-hour exposure levels to protect  against different types of effects in the
community.

      The body of this report consists of a description  of various exposure
limits, their uses,  and how they were  developed.  At the end of each
description there is  also a brief  list of the advantages and disadvantages of
using the method to develop de  minimis levels.   Table 1  presents a summary of
the characteristics of each method for developing short-term exposure levels.
These include such considerations  as whether the method  and/or exposure level
is subjected to peer  review,  whether the method uses primary data or secondary
sources,  whether there is an  opportunity for public comment, and the target
population the value  is  supposed to  protect.  Table 2 lists the actual
exposure value for each of the  methods discussed in the  text.  In addition,
Table 2 also contains other chemicals  that are members of  the chemical groups
listed in the Section 112(b), such as  tetramethyllead, which is considered to
be a lead compound,  but for which  there  are specific exposure levels developed
by other organizations.

-------
                                                                   ** DRAFT **

                 OCCUPATIONAL SAFETY AMD HEALTH ADMINISTRATION
                          PERMISSIBLE EXPOSURE LIMITS


      The Occupational Safety and Health Act of 1970,  requires  the
Occupational Safety and Health Administration to promulgate Permissible
Exposure Limits (FELs) for air contaminants commonly used in the workplace.  A
PEL is the maximum airborne concentration of a substance to which a worker may
be exposed over a specified time period.   At present,  OSHA has  established
PELs for 600 substances, with time weighted averages (TWAs) for 150 of the
Section 112(b) hazardous air pollutants,  including 44 short-term exposure
limits (STELs) and 24 ceiling values (some of the PELs are for  chemicals
classified within the substance groups listed in Section 112(b), e.g.,
hydrogen cyanide).

      PELs are based primarily on Threshold Limit Values (TLVs) that have been
developed by the American Conference of Governmental Industrial Hygienists
(ACGIH).  TLVs may be expressed as a time-weighted average (TWA; usually 8
hours), a 15-minute short-term exposure limit (STEL),  or a threshold limit
value ceiling (TLV-C) which should not be exceeded at any time.

      The most recent list of PELs was published by OSHA in January 1989.
Although many of the PELs are based directly on TLVs,  some are  not.  To
develop the 1989 PELs, OSHA compared the 1987-1988 TLVs with the then current
PEL list.  In cases where the TLVs and PELs were identical, OSHA did not
modify the PELs.  Where the TLV and PEL differed, the PEL was considered for
modification.   If a TLV existed for a substance,  but there was  no PEL, the
substance was also considered for adoption of a PEL.

      For those chemicals considered for modification or adoption of a PEL,
OSHA compared the ACGIH TLV and, if available,  the NIOSH Recommended Exposure
Limit (REL).   The background documentation used to determine the TLV and REL
was reviewed by OSHA and the more appropriate value, in OSHA's  view, was then
adopted as the proposed PEL.  OSHA based its decision on both the scientific
validity of the supporting studies and whether the published documentation for
the TLV or REL would meet OSHA's legal requirements for establishing a PEL.
The legal requirements included a determination that:  (1) there was
substantial evidence of significant risk at the current PEL; (2) significant
risk might exist in the workplace if there was no PEL; (3) the adoption of a
new or revised PEL would substantially reduce the risk; and (4) the new or
revised PEL was technically feasible.  The proposed PELs, with their
supporting documentation and the rationale for each level, were made available
for public comment.  In the Federal Register notice, OSHA summarized all of
the studies reviewed to develop the PELs, listed all references, and presented
justification for the proposed PELs.  After review of the public comments  and
any additional data submitted to OSHA, the PELs were modified as necessary and
adopted as final effective March 1989 (29 CFR 1910).

-------
                                                                   ** DRAFT **

      In establishing  a new or revised PEL, OSHA had  to  conduct a feasibility
analysis to determine  if  the level established by each PEL could actually  be
achieved in the  workplace and at what cost.  OSHA used several  data sources to
make these analyses.   OSHA's Integrated Management  Information  System contains
over 77,000 exposure measurements organized by type of industry and process
and indicates  the  number  of workers included in each  measurement.   OSHA also
used two National  Occupational Hazard Surveys to estimate  the number of
workers potentially exposed to each substance in any  industry segment.   OSHA
had all of these data  reviewed by approximately 20  industrial hygiene and
industrial engineering experts.  In addition, OSHA  conducted a  nationwide
survey of 5,700  firms  asking for information on substances present in the
workplace, processes used, and controls in place.   All of  this  information was
used by OSHA to  ensure that PELs reflected accurate assessments of technical
and economic feasibility  and costs.

      OSHA divided all of the chemicals with proposed PELs into 18 categories
based on the health effect for which there was the  greatest concern, e.g.,
cancer, sensory  irritation, respiratory effects, narcosis, analogy with
another chemical,  etc.  In presenting the final rule, OSHA discussed the
rationale for  each PEL on a chemical-by-chemical basis for each category of
health effect.   The rationale included a brief review of the toxicity data
used to establish  the  PEL, a discussion of various  comments received, and
OSHA's responses to the comments.  For example, a PEL of 50 ppm TWA was
proposed for n-hexane, based on its neurotoxic effects.   This limit is equal
to the ACGIH TLV of 50 ppm but lower than the NIOSH REL  of 100  ppm.  NIOSH
submitted comments to  OSHA agreeing that the lower  value was appropriate for
the PEL.  OSHA presented  a summary of the data used by ACGIH to establish the
TLV, and discussed comments received from several groups,  some  of whom agreed
that the PEL was appropriate (worker unions and NIOSH) and some of whom wanted
che PEL modified (industry trade organizations).  OSHA presented its
interpretation of  the  relevant toxicity data to justify  the value it selected
for each PEL.

      The 600  PELs are based only  on recommended  exposure limits (e.g., TLVs) ;
information on medical surveillance, exposure monitoring,  industrial hygiene
requirements and other ancillary provisions were  not considered when
developing the PELs.

Advantaees of  Using PELs  for Short-term Benchmarks

      •     There are  STELs  or Ceiling Values  for 50 chemicals  that are on the
            Section 112(b)  list  (44  STELs,  24  Ceiling Values,  and  5 with both
            STELs  and  Ceiling Values).  This  is  the largest  data base  other
            than TLVs  to  work  from;  only  60  compounds on  the Section 112(b)
            list do not  have  any  type  of  PEL.

      •     The  data  used to  establish the STEL or Ceiling Value  have  already
            been reviewed by OSHA for  scientific validity.   In addition,  the
            STEL or Ceiling Value is based on either an ACGIH  TLV or a NIOSH
            REL, both of which are widely accepted by industry,  particularly

-------
                                                                   **  DRAFT **

            the  TLVs.   Both  of  these  exposure values  have been determined by a
            review of  the  scientific  literature.

      •     These  values have been  subjected to comment by the public,
            including  affected  industries  and occupational health
            organizations  such  as NIOSH.

Disadvantages  of Using PELs  for Short-term Benchmarks

      •     PELs are designed to protect only workers during a limited
            exposure period  of  an 8-hour workday and  a 40-hour workweek.
            Workers are assumed to  represent healthy  populations;  no attempt
            is made to identify or  consider sensitive populations,  even among
            workers.

      •     Data would have  to  be reviewed to determine appropriate exposure
            limits for general  population.

      •     Secondary  data sources  (e.g.,  ACGIH and NIOSH background
            documentation) were reviewed for the majority of chemicals;
            unpublished data and original  studies were not reviewed in most
            cases  unless submitted  by outside sources e.g.,  industry,  during
            the  comment period.
REFERENCE:

29 CFR Part  1910   Air Contaminants;  Final Rule (Federal Register 54 (12):2232-
2983,  1/19/89)

-------
                                                                  *- DRAFT  **

            MATIONAL INSTITUTE FOR OCCUPATIONAL SAFETY AND HEALTH
                         RECOMMENDED EXPOSURE LIMITS


      As mandated by  the Occupational  Safety and Health Act of  1970, the
National Institute for Occupational  Safety and Health  (NIOSH) develops
Recommended Exposure  Limits  (RELs) for hazardous substances in  the workplace.
Unlike OSHA PELs,  RELs are only recommendations for  occupational  safety and
health standards;  they are not enforceable by law.   However, OSHA and the Mine
Safety and Health Administration review  the RELs when  developing  their
standards for worker  protection.  A  REL  is based on  an 8  or 10-hour workday,
40 hours per week,  for a 40  year exposure  period (working lifetime).  RELs
consist of time-weighted average, short-term (15-minute;  STELs),  and ceiling
values.

      The background  information used  by NIOSH in developing a  REL is
contained in a Criteria Document.  Data  included in  the Criteria  Document must
be publicly available (i.e., unpublished industry studies that  may not be
released are not  included) .  All available scientific  data are  included.
Although most data in the Criteria Document come from  peer-reviewed journals,
NIOSH does not have a policy of quality  checking all studies.   Criteria
Documents are reviewed by internal NIOSH specialists,  as  well as  an outside
panel of peer reviewers,  including representatives from  industry  and other
Federal agencies.   The development of  a  Criteria Document is lengthy and may
take 2 to 3 years prior to publication.

      NIOSH considers a broad variety  of information when developing RELs,
including medical monitoring data and  all  aspects of toxicology,  such  as human
case studies as well  as experimental human and animal  studies,  reproductive
effects, carcinogenicity, and  acute  and  chronic effects.   Uptake, metabolism,
and elimination of the chemical are  also included.  To date, most RELs  have
been established  for  chemicals  for which the endpoint  is  a chronic effect
other than cancer. For example,  the recommended standard for  styrene  is  based
on human nervous  system effects and  on eye and respiratory system irritation.
In addition to presenting exposure effects data, a Criteria Document  includes
information on the physical  and chemical properties of the compound,  uses,
production methods, and potential  for  worker exposure.  The document  also
describes recommended labelling of  the substance in the workplace,  discusses
environmental sampling, analytical methods, biological monitoring of  workers,
medical surveillance, other  applicable occupational health standards  (TLVs,
STELs, ceiling values), and  methods  for worker protection.

      NIOSH prefers to use  toxicity  studies that most closely represent actual
worker exposure conditions,  e.g.,  if the primary concern  for worker exposure
is by inhalation of a vapor, an animal study  (or rarely human study)  in which
the subjects were exposed by inhalation to a vapor  is more significant than a
study in which the subjects  were  dosed orally  by a  liquid form.  RELs may be
developed based on exposures other than airborne concentrations, e.g., medical
monitoring has shown that airborne concentrations of  lead are not the best

-------
                                                                   ** DRAFT **

indicator of levels  of  lead to which workers  have been exposed.   Rather,  blood
levels are a better  indicator, and  the REL  reflects  this by  recommending  that
air concentrations should be such that worker blood  lead levels  do  not exceed
0.060 mg/100 g whole blood.

      NIOSH has adopted a quantitative approach  to establishing  exposure
limits for potential carcinogens.   This  approach is  similar  to EPA's,  in  chat
it includes an assessment of cancer potency and  a subsequent risk assessment.
For reproductive effects for glycol ethers, NIOSH used EPA criteria to
determine the most sensitive species and appropriate endpoints.   Although
NIOSH recommends exposure limits for the workplace,  it feels that by
protecting workers the  health of the general  public  is protected as well.
However, the Criteria Document specifically states that the  standards are not
intended "for the population-at-large, "  and that any such  extrapolation is not
warranted.

Advantages of Using  RELs as Short-term Benchmarks

      •     Data are peer reviewed  by outside and internal experts  and all
            data must be publicly available.

      •     Quantitative RELs exist for  120 chemicals on the Section 112(b)
            Hazardous Air Pollutants List,  with  an addition  29  substances
            having only a carcinogen designation (no quantitative value).

      •     Most RELs are established for chronic effects  although  acute
            effects  are also included.

Disadvantages of Using  RELS as Short-term Benchmarks

      •     RELs are based on worker exposure (i.e., 40-hr/week, 40 year
            lifetime) and workers are assumed to represent healthy
            populations; no effort  is made  to identify or  consider  sensitive
            populations, even among workers.

      •     Although all available  scientific literature on  the substance  is
            reviewed, emphasis is given  to  reports of worker exposure and
            subsequent  health effects.

      •     Airborne concentrations are  not always the  focus of RELs if other
            types of exposure, e.g., dermal,  are more likely.


REFERENCES:

NIOSH.  1990.  NIOSH Pocket Guide  to Chemical Hazards.   Washington, DC:  U.S.
Department of Health and Human Services, Centers for Disease Control, National
Institute for Occupational Safety  and  Health.  June.

Personal communication  with Ralph  Zumwalde of NIOSH.  September  17,  1991.

                                      7

-------
     :                                                              ** DRAFT **

                   AMERICAN INDUSTRIAL HYGIENE ASSOCIATION
                    EMERGENCY  RESPONSE PLANNING GUIDELINES


      Emergency Response Planning  Guidelines  (ERPGs)  are  published by the
American Industrial Hygiene  Association (AIHA).   The  purpose  of the guidelines
is to protect against adverse  health effects  from short-term  exposures to
airborne chemicals  in an emergency situation.   The guidelines are intended for
voluntary use by industry in emergency planning for accidental releases to a
community.   Specifically,  ERPGs, when "used,  together with  other information
such as volatility  and storage volumes,  [can]  provide computerized estimates
of the potential spread and  airborne concentration in case  of a release.   From
chese estimates, action plans  can  be developed."  Thus,  ERPGs are ncc intended
for use as exposure guidelines.

      There are three levels for each ERPG that protect  "nearly all
individuals" following one-hour exposures to  an airborne  chemical:

      •     ERPG-3  is the maximum  concentration in air below  which no "life-
            threatening health effects" would occur;

      •     ERPG-2  is the maximum  concentration in air below  which no
            "irreversible or other serious health effects or  symptoms which
            could impair an  individual's ability to take  protective action"
            would occur; and

      •     ERPG-1  is the maximum  concentration in air below  which no "mild,
            transient adverse  health effects" would occur and no
            "objectionable odor" would be detected.

Thus, industry can  use ERPG  levels as guides  to determine what action should
be taken in the event of an  accidental release of a chemical  at one of the
ERPG levels.

      ERPGs were developed by  chemical companies that agreed to participate
with AIHA in the process.  Each company followed specific guidelines and
methodologies, previously established by the Organization of Resources
Counselors (ORC) ERPG Task Force,  to insure consistency in the development of
each ERPG.  In addition, the ORC  Task Force identified those chemicals for
which ERPGs were to be developed.   Once the ERPGs were developed  and the  data
documented, the AIHA EPRG Committee of scientific experts reviewed, revised
and approved the ERPG documents and exposure levels for publication.

      Each chemical with an ERPG value is reviewed on a case-by-case basis.
All of the data reviewed for establishing an ERPG value and  their references
are included in the documentation, along with the rationale  used  to derive
each ERPG level.

      The procedure for developing ERPGs  emphasizes  the use  of industrial
hygienists, toxicologists, medical experts,  and other health professionals to

                                       8

-------
                                                                   ** DRAFT **

collect and review the data as well as  Co  develop the values and their
supporting documents.   It is also suggested that data be taken from
computerized literature searches of databases (i.e.,  MEDLINE and TOXLINE) and
from industry (i.e.,  unpublished data).  The procedures emphasize the use of
the original sources  of data,  rather than  secondary sources, whenever
possible.

      In general,  ERPGs are based upon  various types  of data.   In order of
preference, these  data are: (1) human data;  (2)  animal data from acute
inhalation toxicity studies that contain clinical and histopathologic testing;
(3) inhalation data from repeated exposure studies that contain clinical and
histopathologic testing; (4) inhalation mortality data; or (5) data from
studies that use other routes  of exposure.   In addition, dose-response data
may be used.  Other data that  are to be considered when developing ERPGs
include: (1) acute and short-term exposure data;  (2)  data for serious adverse
effects from short-term exposures,  such as carcinogenesis or reproductive and
developmental effects; and (3) data for immediate and delayed effects.

      Data from subacute, subchronic, and  chronic toxicity studies, metabolism
and pharmacokinetics  data,  physical/chemical properties, and the currently
existing exposure  limits for each chemical are also summarized in the ERPG
documents.

Advantages of Using the ERPGs  for Short-term Benchmarks

      •     ERPGs  are  intended to be protective of reversible and irreversible
            acute  and serious  health effects from short-term exposures to
            airborne  chemicals.

      •     ERPGs  are  based upon original  data,  especially human data when
            available, thus they are highly likely to insure the protection
            intended  by the limits.

      •     ERPGs  are  based upon data from exposures  to sensitive individuals
            of the population, when available, therefore, they are highly
            likely to  protect  these individuals as well as other members of
            the general population.

      •     ERPGs  are  applicable to l>hour exposures.

      •     The methodology for deriving ERPGs has been reviewed by members  of
            the general scientific community.

      •     The rationale and the supporting data for each ERPG have  been peer
            reviewed.

      •     ERPGs  are established by industry and are  therefore be more  likely
            to be  accepted by industry.

-------
                                                                   ** DRAFT **

Disadvantages of Using the  ERPGs as Shore-term Benchmarks

      •     Very few ERPGs  have been  derived and only  7  chemicals with ERPGs
            are on the CAA  Section  112(b)  list;

      •     Since the maximum  duration  of  exposure  is  one  hour, it is not
            likely that ERPGs  can be  used  to set ambient exposure levels that
            exceed one-hour without some modification;

      •     ERPGs are guidelines for  "emergency planning and response" and
            "are not absolute  levels  demarcating safe  from hazardous;"
            therefore, ERPGs are not  intended for use  as exposure guidelines.


REFERENCE:

American Industrial Hygiene Association (AIHA) .   Concepts  and Procedures for
the Development of Emergency Response Planning Guidelines  (ERPGs). 1989.
                                       10

-------
                                                                   ** DRAFT **

                          NATIOHAL RESEARCH COUNCIL
                       EMERGENCY EXPOSURE GUIDANCE LEVEL


      Emergency Exposure Guidance  Levels (EEGLs)  are  developed by the National
Research Council's Committee on Toxicology  (COT)  for  the  Department of
Defense.  The EEGL is intended for use in emergency  inhalation exposures of 1
to 24 hours by military personnel; these situations  are expected to occur
rarely during the lifetime of an individual.   EEGLs  are expected to protect
against those effects that may impair escape,  judgement,  or performance or
that are irreversible.  Therefore, irritation,  discomfort or other transient
effects may occur from exposures to an EEGL.   In  addition,  the EEGL does not
protect the general population in  an emergency, nor  against effects from
continuous short-term exposures.

      EEGLs are based upon epidemiological,  clinical,  physiological, and acute
toxicity data in animals and humans.   In addition, serious  and chronic effects
that may occur from short-term exposures and immediate and delayed effects are
considered as well.  When data for several  endpoints  exist for a chemical, all
of the data are considered and the most "seriously debilitating, work-
limiting, or sensitive" effect is  used for  deriving  an EEGL.

      The EEGL document contains:  (1) the data used  to derive the EEGL values;
(2) a summary of the physical and  chemical  properties of  the chemical; (3)
data on the pharmacokinetics and metabolism of the chemical; (4) exposure
limits from other sources; (5) the Committee's current and prior
recommendations for an EEGL value; (6) the  rationale for  the EEGL value;  (7)
the rationale for modifying an existing EEGL,  if  applicable; (8)
recommendations for future research;  and (9)  references used for the data.

      EEGLs may also be derived using safety factors to compensate for
inadequate data.  For instance, a  safety factor of 10 is  used when only animal
data are available or if there are no data  for the expected route of exposure.
Extrapolation,  e.g., when oral animal data  are available  and human inhalation
exposure is expected, is done by (1)  assuming that the oral dose in rats  is
equivalent to that in humans and then using the appropriate human breathing
rates, weight,  and exposure durations to calculate the inhaled concentration
by humans or (2) converting the oral dose in rats to an  inhalation dose in
rats and then assuming that dose to be equivalent to the  inhalation dose  in
humans.

      EEGLs have not been derived for complex mixtures, however, it is
recommended by the Committee that  a "proportional reduction in EEGLs for  each
of the constituents of a mixture"  be used,  based  upon the assumption that the
toxic effects of each component in a mixture are  additive.

      When possible, Barber's law (CT - concentration (C) x time  (t)), which
assumes that CT is constant over short periods of time,  is used to develop
EEGLs for different exposure durations.  However, prior  to applying the law,
all data are considered to determine if Harber's  law holds  true.

                                      11

-------
                                                                   ** DRAFT **

      The shore-term public emergency exposure  guidance  Level (SPECIE is
intended for emergency single short-cerm exposures  co  the  general population.
Therefore,  sensitive individuals are protected  from single exposures to the
SPEGL.  SPEGLs are derived from EEGLs and may be  equivalent to 0.1 - 0.5 times
the EEGL.  However,  for the protection of children,  the  elderly or other
sensitive groups,  a safety factor of 2 may be used  and to  protect fetuses or
newborns, a safety factor of 10 may be used.

Advantages of Using EEGLs and SPEGLs for Short-term Benchmarks

      •     EEGLs  and SPEGLs apply to airborne  contaminants.

      •     EEGLs  and SPEGLs are based upon original data,  especially human
            data when available, thus, they are highly likely to insure the
            protection intended by the limits.

      •     The rationale for each EEGL and SPEGL and  the  supporting data for
            each has been peer reviewed.

      •     The maximum duration of exposure  is 24  hours,  thus, EEGLs and
            SPEGLs may be useful for setting  daily  ambient exposure levels.

      •     One chemical may have several EEGLs and SPEGLs for different
            exposure durations (i.e., within  1  to 24 hours) which makes these
            values more applicable to various exposure conditions.

Advantages of Using SPEGLs for Short-term Benchmarks

      •     Four of the five chemicals with SPEGLS  are on the CAA Section
            112(b) list.

      •     SPEGLs are intended  to protect sensitive individuals as well as
            the general population.

Disadvantages of Using EEGLs and SPEGLs  for Short-term Benchmarks

      •     EEGLs  and SPEGLs are intended  to  protect against  some acute,
            reversible health  effects  (i.e.,  those  that would impair emergency
            escape), but not all acute  effects.

      •     EEGLs  and SPEGLs are intended  for rare  or occasional  exposures  and
            not repeated exposures  such as  those likely to be encountered  with
            ambient air pollutants.

      •     EEGLs  and SPEGLs  are guidelines  for emergency planning  in  military
            situations and are not  intended for use as regulatory standards.

Disadvantages of Using EEGLs  for Short-term Benchmarks

      •     Only  18 chemicals  with  EEGLs are on  the CAA Section 112(b) list.

                                      12

-------
                                                                   ** DRAFT **

      •     EEGLs  are  intended  to protect healthy military personnel and not
            the  general  population  or sensitive individuals of the population.

   The disadvantage  of using  SPEGLs  as the basis for short-term ambient
benchmarks  is  that the values have  been developed for only five chemicals on
the Section 112(b) list:   hydrazine,  monomechyl hydrazine, 1,  1'-
dimethylhydrazine, hydrogen chloride,  and nitrogen dioxide.


REFERENCE:

Committee on Toxicology,  National Research Council.  Criteria and Methods for
Preparing Emergency  Exposure  Guidance Level (EEGL),  Short-Term Public
Emergency Guidance Level  (SPEGL), and Continuous Exposure Guidance Level
(CEGL) Documents.  National Academy Press,  Washington, D.C., 1986.
                                      13

-------
                                                                   ** DRAFT **

             SARA TITLE  III  - EXTREMELY  HAZARDOUS SUBSTANCES LIST
                              LEVELS OP CONCERN


      Short-term exposure  limits  for  chemicals  on the  Superfund Amendments and
Reauthorization Act  (SARA) Title  III  Section 302  list  of  Extremely Hazardous
Substance (EHS)  are  called "Levels  of Concern"  (LOCs).  These levels are the
airborne concentrations  of chemicals  on  the EHS list at which no serious,
irreversible health  effects  or death  may occur  following  a single, short-term
exposure.

      LOCs were derived  from the  "Immediately Dangerous to Life and Health"
(IDLH) values developed  by the National  Institute for  Occupational Safety and
Health (NIOSH).   IDLH values are  approximately  one or  two orders of magnitude -
below the median lethal  concentration (LC50) or dose (LD50)  taken from acute
toxicity studies in  mammalian species.   IDLH values may also be based upon the
lowest inhalation exposure concentration that causes death or irreversible
health effects in any species (LCLO) .  Finally, the IDLH  may be equivalent to 5
500 times the Permissible  Exposure  Limit (PEL)  if no acute toxicity data are
available for a chemical.

      The LOG is defined as  one-tenth of the IDLH.  Since the IDLH is designed
to protect workers,  is based upon a 30-minute exposure duration, and protects
against serious and  irreversible  health  effects,  a safety factor of 10 was
used to derive the LOCs:  (1) to  insure protection of the  general population,
including sensitive  individuals;  (2)  to  protect against health effects from
exposures which occur for  more  than 30 minutes; and (3)  to protect against
serious and reversible health effects.

      For chemicals  on the EHS  list that have no IDLH  value, animal acute
toxicity data, i.e., LD50  and LC50  data, have been used  to derive LOG values.
LD50 is the oral or  dermal dose  at  which 502 of the test  animals died
following exposure to a  chemical  and LC50 is the inhalation concentration  at
which 50% of the animals died.   The LC50 data are preferred when available.
These data were taken from the  NIOSH Registry of Toxic Effects of Chemical
Substances (RTECS) data  base.   Estimated IDLH values derived from these  data
are equivalent to one-tenth  (1/10)  of the LC50  or one-one hundredth  (1/100) of
the LD50.  The LOG is then equal  to one-tenth  (1/10) of the estimated IDLH.

      For chemicals  on the EHS  list that have no LD50 or LC50 data available,
LDLO or LCLO data have been  used to derive LOG  values.  The LDLO  and LCLO
values represent the lowest  lethal  dose  or concentration, respectively,  of a
chemical.  LCLO values are preferred when available because they  are based on
inhalation exposures.  Estimated IDLH values derived from these data are
equivalent to the unmodified LCLO or one-tenth of  the LDLO.  The  LOG is  then
equal to one-tenth (1/10)  of the estimated IDLH.

-------
     ;                                 _                            ^ DRAFT •»

Advantages of Using LOG Values  for  Shore-term  Benchmarks

      •     LOG values  apply  co  airborne  contaminants.

      •     LOG values  are  intended co  protect against  reversible and
            irreversible acute  and  serious health  effects  from short-term
            exposures.

      •     LOG values  are  intended to  protect the general population,
            including sensitive  individuals.

      •     LOG values  exist  for many  (55) of  the  chemicals on the CAA Section
            112(b)  list.

Disadvantages of Using  LOG  Values for  Short-term Benchmarks

      •     Most of the LOG values  are  based upon  animal LC50, LD50,  LCLO and
            LDLO data,  which  may not protect against  all health effects in
            humans.

      •     The remaining LOG values are  derived from secondary sources (i.e.,
            the RTECS data  base); thus, data used  to  derived LOG values may
            not be  peer reviewed or may be inaccurately recorded in RTECS.

      •     A safety factor of  10 was  applied  to the  IDLHs to protect
            sensitive individuals of the  population and for protection against
            serious reversible  health  effects;  however, it is not known if
            this safety factor  is sufficient to provide adequate protection in
            these cases.

      •     The methodology for deriving  LOCs  has  not been reviewed by members
            of the  general  scientific  community.

      •     The rationale for each  LOG and some of the  supporting data of each
            have not been peer  reviewed.

      •     It is not known what the maximum duration of  exposure at the LOG
            would be for protection against adverse effects.

      When setting  exposure limits, it is important to  consider serious,
chronic effects, as well as acute effects that may be induced by short-term
exposures to chemicals.  LOG  values and most of the other  exposure levels (see
Table 1) reviewed in this report protect  against chronic  (serious,
irreversible) health effects  from short-term exposures.  Effects such as
chronic respiratory disease,  increased susceptibility to  infection, visual
impairment,  or death may occur  following  short-term inhalation exposures
(i.e.,  as with the  accidental release  of  methyl isocyanate in Bhopal, India).

      In a report previously  prepared  by  Clement  for  the  EPA Office of Toxic
Substances,  several LOG values  were reviewed  to determine  if  the levels were

                                      15

-------
                                                                   ** DRAFT **

protective  againsc chronic effects  that occur from shore-term exposures to
potential neurotoxic agents.   In one  case study of lindane, it was found that
che current LOG did not protect  a worker against convulsions induced by a
single exposure to the pesticide (the route of exposure was not reported).


REFERENCE:

US EPA Technical Guidance for Hazards Analysis, Emergency Planning for
Extremely Hazardous Substances,  1987.
                                       16

-------
                                                                   ** DRAFT **

                                   CAPCOA
                  NONCANCER ACCEPTABLE ACUTE  EXPOSURE LEVELS


      The California Air Pollution Control Officers Association has developed
risk assessment guidelines for  its members to use in the preparation of health
risk assessments required under the California Air Toxics "Hot Spots"
Information and Assessment Act  of 1987.  This law mandates a statewide
inventory of air toxics emissions from individual facilities.  A risk
assessment should include an evaluation of both the potential cancer and
noncancer effects of long-term  and short-term exposures to emissions from a
particular facility.  CAPCOA has developed one-hour maximum concentrations
considered to be acceptable exposure  levels for 13 substances.  Air pollution
control districts and individual facilities should compare their emissions
levels and the potential for acute effects of emissions of these substances to
the CAPCOA levels.   Exposures below the CAPCOA acceptable level are not
expected to result in adverse effects to the  exposed population.  If exposure
is to more than one substance on the  list, a  hazard index approach, which
assumes that multiple subthreshold exposure will result in an adverse health
effect, should be used.  The hazard index approach, established by EPA,
assumes that the effects of each substance are additive for a given adverse
response.

      Of the 13 chemicals for which noncancer acceptable acute exposure levels
have be^tn established by CAPCOA, 12 are on the Section 112(b) list of
hazardous air pollutants.  Several sources were used to develop these exposure
levels.  Scientific data (including original  studies) were used whenever
possible.   Particular emphasis  was given to inhalation studies that used
exposure times as close to one  hour as possible.  An uncertainty factor of 10
was applied to animal data when extrapolating to humans.  A second uncertainty
factor of 10 was also applied to account for  interspecies variation.
Preference was given to studies that  produced a scientifically valid no
observed adverse exposure level or a  lowest observed adverse exposure level.
The resulting levels were considered  to be protective of the general
population,  including sensitive populations.   It should be noted that these
exposure levels were reviewed by members of the California Department of
Health Services ojnjty and were not subject to  outside peer review or general
public comment.

Advantaees of CAPCOA Approach for Short-term  Benchmarks

      •     Exposure levels are for one-hour  and may be used directly by OAQPS
            without further manipulation.

Disadvantages of CAPCOA Approach for  Short-term Benchmarks

      •     Noncancer acute exposure  levels have been developed for only  11  of
            the substances on the Section 112(b) list.
                                      17

-------
                                                                   ** DRAFT **

            Methodology used Co develop  the  exposure  levels  have nee been
            reviewed  by experts outside  of the  California Department of Health
            Services  and  CAPCOA.
REFERENCES:
CAPCOA.   1991.   Air Toxics  "Hoc  Spots"  Program:   Risk Assessmenc Guidelines.
Sacramenco,  CA:   AB 2588 Risk Assessment  CommiCCee  of the California Air
Pollucion Concrol Officers  Association.   January.
                                       18

-------
     :                                                              ** DRAFT **

                             CALABRESE AND KENYON
                           AMBIENT AIR LEVEL GOALS


      Calabrese and Kenyon (L991)  have developed a methodology for developing
Ambienc Air Level Goals.   An  ambient: air level  goal (AALG)  is the level of a
toxicant below which the  risk of an adverse health effect for an exposed
population is less than a predetermined de minimis value such as one in a
million (10*6) or one in one hundred thousand (10*5).   These values are called
goals rather than standards because no technical or economic feasibility
studies are conducted to  determine if the levels are realistic.   The authors
present two methods of deriving  AALGs: the first is a model-based approach for
non-threshold health effects,  such as cancer,  that estimates the risk at a
given dose.  The second is for a threshold effect, which is based on a
modification of the no observed  adverse effect  level (NOAEL) or lowest
observed adverse effect level (LOAEL)  by some uncertainty factors.

      Use of either of these  methods is predicated on a selection scheme that
includes such factors as  production volume, known toxic effects and monitoring
studies.  In addition,  a  preliminary screening  of data using the Registry of
Toxic Effects of Chemical Substances will give  a better indication of the
endpoint of concern and references to the available toxicology literature.
After the pertinent toxicologic  studies have been evaluated, an AALG may be
derived for either a carcinogen  (or mutagen) or a noncarcinogen.  For
chemicals that are preliminarily classified as  carcinogens, Calabrese and
Kenyon present a method for developing an AALG  based on EPA's weight-of-
evidence classification and quantitative risk assessment procedures.

      For chemicals that  are  noncarcinogens, AALGs may be developed by
dividing the NOAEL or LOAEL by applicable uncertainty factors that account for
inter-  and intraspecies differences, extrapolating from a LOAEL to a NOAEL,
and differences between healthy  populations and sensitive populations.  It is
recommended that for chemicals with a Reference Concentration (RfC), that this
level be used as the AALG with or without adjustment for relative source
contribution.  The authors present selection criteria for selecting NOAELs and
LOAELs,  including the use of  occupational limits as their surrogates.
Although the authors recommend that occupational exposure limits  (TLVs and
RELs) not be used as the  starting point in the development of AALGs, they
recognize that these levels are  frequently the most expedient and economically
justified starting point.   Consequently, their methodology includes a scheme
to develop NOAELs and LOAELs  by  applying safety factors to TLVs or RELs.  The
AALGs derived from the NOAEL  should be based on the endpoint of greatest
concern, or if the chemical has  several endpoints of concern, an AALG may be
developed for each endpoint and  then the most conservative value  can be
chosen.

      Calabrese and Kenyon acknowledge that limitations do exist  when
developing AALGs from NOAELs. These limitations  include ignoring the shape  of
the dose-response curve,  selecting an appropriate NOAEL, and incorporating  the
quality of the relevant studies.  Uncertainty factors may be applied  to

                                      19

-------
                                                                   ** DRAFT **

account for a variety of  factors including:  (1)  differences between animals
and humans; (2)  differences between healthy  individuals  and more sensitive
populations;  (3)  extrapolating downward from a.  LOAEL to  a NOAEL;  and (4)
extrapolating from a less-than-lifetime or subchronic study to a lifetime or
chronic value (this may not be a major consideration for developing de minimis
levels for acute  effects).

      Although an occupational exposure level may be used to develop a NOAEL,
Calabrese and Kenyon point out several reasons  as to why this approach is
questionable, with emphasis on the use of TLVs.   First,  the TLVs are not
intended to be used for regulatory purposes  since they are guidelines and are
not thresholds for "safe" exposures.  TLVs are  not necessarily based upon
effective or no-effect doses  from animal or  human studies, and there is no way
to determine the  extent to which each TLV was based upon actual experimental
data.   Secondly,  the TLVs may not be protective of sensitive individuals  in
the population because they are designed to  protect generally healthy workers.
Calabrese and Kenyon emphasize that any TLV  or  REL used to derive an AALG
should be applicable to the general population,  including sensitive
populations,  as  well as to workers.  In addition, there  should be evidence
that the occupational exposure limit may be  used as a human NOAEL or LOAEL.

      Although the issue  of acute effects  in general is  not discussed in this
book,  Calabrese and Kenyon do discuss sensory irritation at some length.   A^
djaclsion table_i£ presented that facilitates extrapolating from  occupational
data^or other sensory irritation A»r* fg_j»ri__AftL5   NOAELs and LOAELs are
preferred, but^in some cases  a mouse RD50  may be used (an RD50 is the
concentration that decreases  the respiratory rate in a test organism by 502) .

      To demonstrate the  applicability of  their method,  Calabrese and Kenyon
derive AALGs for 110 chemicals, 73 of which  are on the Section 112(b) list.
The discussion includes a summary of existing occupational limits and the
basis for derivation, a toxicity profile  that includes all aspects of relevant
toxicity, and the rationale for the selection of the AALG.  For example,
acetonitrile (listed in Section 112(b)) has  a AALG for systemic toxicity of
0.068 mg/m3 for  a 24-hour time weighted average.   This value is based on a
mouse NOAEL of 100 ppm from an NTP study.   This value was adjusted for
continuous exposure and converted to a human equivalent  inhalation exposure
dose.

Advantages of Using the AALG  Approach  for Short-term Benchmarks

      •     Method is well  defined with  specified uncertainty  factors  and may
            be applied to a variety  of  chemicals.

      •     Uses best toxicity  data  available, whether  occupational  exposure
             limits or other  types  of human  or animal data (NOAEL,  LOAEL,  LC50)
             and encourages  review of  all background documentation used in
             support of occupational  exposure limits.
                                      20

-------
                                                                   ** DRAFT **

      •     Because  ic  is based on widely-accepted chemical guidelines (RELs,
            TLVs),  limits derived using  this  method are more likely to be
            acceptable  to affected industries.

      •     Sensitive populations can  be incorporated into the  AALG value.

      •     Either  of several methods  may  be  used to derive AALG  values based
            on amount and type of data available.

Disadvantages  of Using  the AALG Approach for  Short-term Benchmarks

      •     Does not really present new  starting point for developing AALG,
            but uses preestablished limits, such as RELs and NOAELs.

      •     Intended for long-term (chronic or  lifetime) exposures;  however,
            eliminating uncertainty factor in derivation may be sufficient  to
            determine short-term exposure  limits.

      •     Acute effects other than sensory  irritation and LD50  data are not
            incorporated into this method  as  endpoints.  Methods  to protect
            against  other acute effects,  such as nausea, are not  provided.


REFERENCE:

Calabrese,  E.J., and Kenyon. E.M.  1991.   Air Toxics and Risk Assessment.
Chelsea,  MI:   Lewis  Publishers.
                                      21

-------
                                                                   ** DRAFT **

                      MINNESOTA POLLUTION CONTROL AGKffCY
                  REGULATORY LIMITS FOR SHORT-TERN EXPOSURES


      The proposed methodology of the Minnesota Pollution  Control Agency for
setting One-Hour Regulatory Limits for airborne emissions  is  outlined below.

      Data will be taken from scientific reviews, original scientific studies,
and other exposure limit documents.  Searches of the TOXLINE  data base will
also be conducted for irritation and/or inhalation exposure data from the last
10 years.  The Registry of Toxic Effects of Chemical Substances (RTECS) and
the Hazardous Substances Databank (HSDB) will be used  to supplement TOXLINE.
The Agency for Toxic Substances and Disease Registry (ATSDR)  Toxicity
Profiles, the World Health Organization (WHO) Health Criteria,  the National
Institute for Occupational Safety and Health (NIOSH),  and  EPA health documents
will also be used as sources of original data.   All original  data taken from
regulatory documents will be reviewed for "adequacy" prior to use in the
development of any exposure limits.

      Data will be reviewed for effects that occur following  3 minute to 8
hour exposures and that meet the EPA definitions of  (1) adverse effect or (2)
functional impairment.  In addition, any short-term exposure  data for eye,
nose or throat irritation, dizziness, headache, nausea, fatigue, or other
"subjective sensations or reactions (excluding odor)",  changes in
physiological or respiratory function, increased physiological or
psychological stress, and decreased resistance to disease  will be considered.

      Observed adverse effect levels (OAELs), lowest observed adverse effect
levels (LOAELs), and/or no observed adverse effect levels  (NOAELs) will be
identified, using specific criteria, from all significant  data.  Exposure
Limits will be based upon these levels using the safety factors of 10, 5, and
1, respectively.  In addition, another safety factor of 10 may be used to
protect sensitive individuals, however, this safety  factor can be reduced to  3
if the OAELs, LOAELs, or NOAELs are based upon data  taken  from exposures to
sensitive individuals.  When several OAELs, LOAELs,  and/or NOAELs can be
identified from various studies, the "best" study will be  chosen based upon
its ability to "meet the criteria of sound study design,  demonstration of a
graded dose-response relationship, and sensitive measurement of effect".

      Extrapolation of data using a probabilistic approach will be used as  an
alternative to the above strategy when the data are  suitable for applying  the
model (Lewis and Alexeef, 1989).
                                      22

-------
                                                                   ** DRAFT **

      In addition,  RD50 data1 may be  used for setting  exposure limits.
However, prior to the use of such data,  several criteria must be met: (1)
there are no other "adequate" human data that can be used;  (2) sensory
irritation (the endpoint of the RD50 studies) must be  identified as the
critical effect,  since all exposure limits are based upon the critical effect;
and (3) all additional toxicity data have been considered.   If RD50 values
were used to establish exposure limits,  large safety factors would be applied
(i.e., 1,000) and the RD50 data "would be supplemented with supporting
documentation."

      Calabrese and Kenyon (1991) have proposed potential short-term ambient
air levels, called Ambient Air  Limit Goals (AALGs) for several chemicals.
Most of the AALGs are based upon TLV values.   AALGs will be considered as
potential exposure limits by the Minnesota Air Quality Staff; however, all
original data will be reviewed  prior to making any final decisions.

      Once potential short-term exposure limits have been identified by the
Air Quality staff,  according to the methods described above, the proposed
exposure limits will undergo a  review process that consists of several steps:
(1) the proposed exposure limits will be reviewed by the staff toxicologist;
(2) the Minnesota Department of Health's Health Risk Assessment Section will
then review the proposed exposure limits; (3) the Human Health Subcommittee
members will review the proposed exposure limits and all of the supporting
data for comment; and (4) experts in respiratory health that are not
affiliated with the Minnesota Pollution Control Agency or the Minnesota
Department of Health will review the proposed limits.

Advantages of Using the Minnesota Approach for Short-term Benchmarks

      •     These values are protective against reversible and irreversible
            serious and acute effects from short-term exposures to airborne
            chemicals.

      •     These values are based upon original animal and human data,  thus
            they are highly likely to insure the protection intended by  the
            limits.

      •     The limits will be  applicable to 1-hour exposures.

      •     The methodologies,  supporting data, and rationale for these  levels
            are all peer reviewed.

      •     These values are intended to protect all members of the  general
            population, including sensitive individuals and when available.
     1RD50  is  the concentration  at which a SOX decrease in respiration rate is
noted in animals (usually mice)  following a short-term exposure  (usually  4
hours) to a chemical;  this concentration has been correlated with  irritation
levels in humans.

                                      23

-------
                                                                   ** DRAFT **

            use data from exposures Co sensitive populations to derive the
            exposure limits.

      •     These  values  are  intended to be  used as ambient exposure levels
            and,  therefore, no manipulation  of the values will be needed prior
            to their application.

Disadvantages  of Using the Minnesota Approach for Short-term Benchmarks

      •     No Regulatory Limits for airborne chemicals have been developed as
            yet and it may be years before an extensive list of chemicals with
            these  limits  is available.

      •     These  values  will be based upon  one-hour exposures and, therefore,
            may not be applicable to daily exposures or other exposures that
            exceed one-hour.
REFERENCES:

Minnesota Pollution Control  Agency, Air  Quality Division, Air Toxics Program.
Proposed Procedures for Setting Regulatory Limits for Short-Term Exposures.
3/6/91 (First Draft);  6/19/91  (Second  Draft).

Lewis and Alexeef.   Quantitative  risk  assessment of non-cancer health effects
for acute exposure  to  air pollutants.   Presented at the 82nd annual meeting of
the Air and Waste Management Association,  Anaheim, CA. 1989.

Edward J. Calabrese and Elaina M.  Kenyon (Eds).  Air Toxics and Risk
Assessment.  Lewis Publishers,  Inc., Michigan.  1991.
                                      24

-------
                                                                   ** DRAFT **

           AMERICAN CONFERENCE OF GOVERNMENTAL INDUSTRIAL HYGIENISTS
                            THRESHOLD LIMIT VALUES


      The following discussion on the derivation of Threshold Limit Values
 (TLVs) is included because several of the other short-term exposure levels
 discussed in this report are based upon TLVs.   In addition,  TLVs may be useful
 for setting ambient short-term exposure limits for airborne chemicals.   The
 potential uses of these values as protective ambient-air levels and a detailed
 discussion of the advantages and disadvantages for applying TLVs outside of
 the workplace can be found elsewhere (Calabrese and Kenyon,  1991; Chemical
 Manufacturers Association, 1988);  therefore, only a brief discussion of the
 applications of TLVs to setting short-term ambient exposure levels will be
 presented here.

      TLVs are issued by the American Conference of Governmental Industrial
 Hygienists (ACGIH) as guidelines to protect workers from effects of repeated
 daily chemical exposures during the workweek and are referred to as TLV Time-
 Weighted Averages (TWA).   Weekly exposure to TLV-TWAs should not exceed more
 than four 15-minute exposures/day with at least a 60 minute interval between
 exposures.  In addition,  there are two TLV categories that protect workers
 from short-term exposures; TLV-Ceilings (C), which are intended to protect
 workers from single short-term exposures that are not to be exceeded at any
 time during the 8-hour workday,  and TLV-Short-Term Exposure Limits (STELs),
 whl-ch are not to be exceeded for more than 15-minutes during an 8-hour
 workday.

      Each TLV has documented supporting data, much of which is respiratory
 and ocular irritation data from industrial exposures.  In addition, there are
 some acute toxicity data (LCSOs and LDSOs) and other animal data which
 describe  systemic effects, cancer,  and mutagenic effects.  The documentation
 also contains information on the physical properties of each chemical, the
 rationale for choosing each TLV, and the references for the supporting data.

      TLVs are derived by the Chemical Substances TLV Committee on a chemical-
by-chemical basis from available animal and human data and industrial
 experiences.  The kind of data used, and the effects that each TLV protects
 against,  vary for each chemical.  Thus, there is no consistent format for
 choosing  data for derivation of TLVs other than to use the "best available
 data".  In addition, the final decision for setting TLVs is a "judgment" based
upon the  data, but the judgment process used is not always well described in
 the TLV documentation.

Advantages of Using TLVs for Short-term Benchmarks

      •     TLVs apply to airborne contaminants.

      •     TLVs are intended to protect against reversible and  irreversible
            acute and serious health effects from short-term exposures.
                                      25

-------
     '.                                                              ** DRAFT **

      •     TLVs  exist  for many of the  chemicals on the  CAA Section L12(b)
            list.

      •     TLV-Cs  and  TLV-STELs protect  against effects following short-term
            exposures of either (1) one-time  exposures during an 8-hour period
            or (2)  15 minute exposures  in an  8-hour period.   Both of these
            values  may  be applicable  to short-term ambient exposure levels.
            In addition, these exposures  are  allowed on  a daily basis,  thus,
            these levels are suitable for ambient air exposures that may be
            randomly repeated.

Disadvantages of  Using  TLVs for Short-term Benchmarks

      •     TLVs  are not necessarily  directly derived from a complete
            toxicology  data base and, therefore, may not adequately protect
            against all health effects  in humans.

      •     TLVs  only protect a specific, healthy segment of the population
            between the ages of 18-65;  they are not intended to protect the
            general population or sensitive individuals  of the population.

      •     TLVs-TWA protect against  effects  following daily 8-hour exposures
            over  a  40-hour workweek.  These exposure durations may not be
            applicable  to short-term  exposure levels in  ambient air since they
            are lifetime daily exposures.

      •     The methodology for deriving  TLVs has not been reviewed by members
            of the  general scientific community.

      •     The rationale for each TLV  and some of the supporting data of each
            have  not been peer reviewed.


REFERENCES:

ACGIH.  Threshold  Limit  Values for Chemical Substances and Physical Agents,
1990-1991.

Edward J. Calabrese and Elaina M. Kenyon (Eds).  Air Toxics and Risk
Assessment.  Lewis Publishers, Inc., Michigan. 1991.

Chemical Manufacturers  Association.   Chemicals  in  the Community: Methods to
Evaluate Airborne Chemical  Levels.  Washington,  D.C., 1988.
                                      26

-------
                                                                   ** DRAFT **

                    THE STATE OF MARYLAND SCREENING LEVELS


      The State of Maryland uses I-hour and 8-hour Screening levels as guides
for ambient air emission levels that are protective of public health from
noncancer effects.  Thus,  they are not standards.

      A screening level can be equivalent to either the TLV-TWA, TLV-STEL, or
TLV-C divided by a factor of 100.  Screening levels based upon the TLV-TWA are
used for protection from 8-hour exposures, whereas, the other TLV-Based
Screening Levels protect against 1-hour exposures.  A safety factor of 100 is
used with the TLVs to allow protection of the general public.

      The State of Maryland has a second approach to generating screening
levels which uses safety factors with NOEL or LC50 data.  These screening
levels are referred to as  Threshold-Based Screening Levels and are used only
when a substance has no TLV.   The preferred data for these values are human
and long-term studies rather than animal and short-term studies.  The data are
multiplied by factors that convert long-term exposures to short-term
exposures.  The following formulae,  given in order of preference, are used to
develop Threshold-Based Screening Levels:

            (i) Divide a 90-day inhalation NOEL in mg/m3 for rats,  mice,  or
            rabbits by 100;

            (ii) Multiply a 90-day oral NOEL in mg/kg for rats by 2.7 x 10'3,
            for mice by 9.0 x 10"4,  or  for rabbits  by  1.3  x  10"3;

            (iii) Divide a 7-day inhalation NOEL in mg/m3 for rats,  mice,  or
            rabbits by 700;

            (iv) Multiply a 7-day oral NOEL in mg/kg for rats by 3.8 x 10"4,
            for mice by 1.3 x 10'4,  or  for rabbits  by  1.9  x  10'4;

            (v) Divide an LC50 in mg/m3 for rats,  mice,  or rabbits by 10,000;
            or

            (vi) Multiply an oral LD50 in mg/kg for rats by 4.1 x 10"5,  for
            mice by 1.4 x 10'5,  or for  rabbits  by 2.0  x  10'5.

When data from two or more animal tests can be used to develop Threshold-Based
Screening Levels, the data producing the  lowest screening level are preferred.

      A third approach to generating screening levels is used when the other
two approaches do not adequately protect  against health effects or are too
stringent.  The levels generated by this  approach  are called Special Screening
Levels.   No information was provided on the development of  Special Screening
Levels.   However, Special Screening Levels must provide "a.  scientifically
appropriate basis for screening analysis" and must protect  human health prior
to approval by the Department of the Environment.

                                       27

-------
     •                                                             ** DRAFT  **

      Each of these  screening  levels may be  subject  to  scientific  review  and
public comment,  but  the  reviews are not required and are  implemented only upon
request of the Maryland  State  Department of  the  Environment.

      Screening  levels may be  replaced by  Acceptable Ambient Levels (AALs) ,
which are derived by the releasing facility,  if  it can  be shown  that AALs will
protect public health from noncancer effects.  In general, AALs  are used  only
if a source cannot meet  a-screening level.   AALs are based upon  "a'detailed
review of the scientific information" but  there  is no specific methodology for
development of AALs.

Advantages of Using  the  State  of Maryland  Screening  Levels for Short-term
Benchmarks

      •     These values are  intended to protect against  reversible and
            irreversible serious and acute effects from short-term exposures
            to airborne  chemicals.

      •     These values are  applicable  co 1-hour or 8-hour  exposures.

      •     These values are  intended  to protect the general population.

      •     These values are  intended  to be used as  ambient  exposure  levels
            and, therefore,  no manipulation of the values will be needed prior
            to their application.

      •     Approximately 23  chemicals  on  the CAA Section 112(b) list have 8-
            hour Screening Levels  and  3  chemicals have  1-hour  Screening
            Levels.

      •     Additional  Screening Levels  can easily be developed using
            available TLV exposure limits  and with more effort using available
            NOEL or  LC50 data from animal  studies.

Disadvantages of Using the State of Maryland Screening Levels  for Short-term
Benchmarks

      •     These values are based upon 1-hour  or 8-hour exposures and,
            therefore,  may not be applicable to  daily exposures or other
            exposures that exceed eight-hours.

      •     It  is not explicit if these limits  are  intended to  protect
            sensitive individuals of the population.

      •     Application  of a  safety factor  of 100 to the  TLVs may not be
            adequate to  protect the general population.

      •     The methodologies, supporting data,  and rationale for these  levels
            are not always peer reviewed, and the extent to which some  of  the
            values  are  reviewed is not clearly  defined.

                                      28

-------
                                                                   ** DRAFT **
            Some of these values are based upon NOEL and LC50 animal data,  Co
            which safecy factors have been applied.   Thus,  che values based
            upon chese daca may noc adequacely protect against all health
            effects in humans.

            Some of these values are based upon TLV-TWA,  TLV-STEL, and TLV-C
            limits and, therefore,  have all of the disadvantages of the TLVs.
REFERENCES:
Maryland Department of the Environment Overview of Maryland's Proposed Air
Toxics Regulations, June 1988.

Maryland Department of the Environment Toxic Air Pollutants.  July 1990.
                                      29

-------
     ;                                                              ** DRAFT **

                  A POTENCY-BASED METHOD FOR ACUTE TOXIC ITT


      For most of the  chemicals on  Che  Section  112(b)  list,  some toxicity data
for acute exposures  are  available.  For most  chemicals,  there may be toxicity
data to identify an  exposure  concentration  for  a  particular  acute exposure
duration (i.e. 1-hour  or 8-hours) at which  adverse  effects may not be
expected.  However,  the  range of  exposure concentrations over a range of
exposure durations that  is  likely to be without adverse  effects in exposed
populations is not well  characterized.   For most  chemicals,  as either the time
or the dose increases, the  number of animals  exhibiting  a specified effect or
the severity of that effect also  increases.   Also,  in  many cases, chemicals do
not follow Harber's  Law  under acute situations  in which  high exposures may
occur for short durations.  Therefore,  in evaluating acute exposure
situations, the risk assessor/risk  manager  may  have to make  decisions
regarding the impact of  ambient levels  on exposed populations for a exposure
concentration/duration combination  for  which  specific  empirical data do not
exist.

      To address this  issue,  Clement International, for  the  OAQPS/PAB (Clement
1991a)  investigated  the  development of  a dose-duration-response model for
acute/short-term exposures.   Two  dose-duration-response  models, which
represent continuations  of the work proposed  by Hertzberg (1989) and Knauf and
Hertzberg (1989), have been investigated.   Both of these models take into
account the duration of  exposure,  the dose, and the severity of the effect
seen at that dose/duration.   Both models estimate the  probability (P) of an
effect of severity greater than  index  i at  a  dose (d)  for a  duration (T);
expressed mathematically, the P(s>i|d,T).

      To develop these probability  estimates, all adequate  data on adverse
health effects resulting from less  than 24  hour exposures (or other exposure
duration can be specified)  are used.  This  approach to using acute data has a
significant advantage  over other  methods in that  all  acute effects, including
minor,  sublethal, and  lethal  are  used  in this assessment.  The  following steps
are conducted:  (1)  all  the primary literature or competent review sources  for
a chemical are critically reviewed for  dose-response  information;  (2)
administered doses are converted to human equivalent  units according to  the
RfC methodology; (3) for each adequate  study, responses for each  animal  in
each dose group are  assigned a severity index;  and (4) data generated  are
input to the dose-response model.  A severity scale similar to  those developed
by Dourson et al. (1985) and Hertzberg and Miller  (1985) is devised.   It is a
rank-order scale, in that the difference between severity categories  is  not an
arithmetic assignment, i.e.,  an index of 2 is not  twice as severe an  index of
1.

      Rather than a single dose for a single endpoint,  these  models  allow for
the estimation of a family of curves,  each of which represents  the probability
of the occurrence of an  adverse effect of  a  specified level  of severity, or
the no adverse effect curve  if the lowest  level  of severity  is  so defined,  for
                                      30

-------
                                                                   **  DRAFT **

any dose-duration combination.   Both the maximum likelihood estimate and the
corresponding lower confidence  limit (e.g.,  95X or 99Z)  for each curve can be
derived.

      For application in source category ranking schemes,  Clement developed an
extension to these models by allowing estimation of the potency of the
chemical (Clement 1991b).  The  potency is estimated by allowing the model to
solve for d*. which is set equal to the 95X  lower bound on dose for a
specified probability that severity will exceed an index of 2 for variable
time and dose.   Expressed mathematically this is the 95X lower bound on dose
for T-24 [P(s>2|d,T)-cr] .  The choice of T-24 hours is based on the assumption
that, for a given acute  dose, the longer the exposure the worse the acute
effect,  although any time could be used.  Severity exceeding an index  of 2 is
recommended because these effects are in the observable range, and for the
purpose  of de minimis levels, it is desirable to choose an effect associated
with a dose that is higher than threshold for any effect yet still in  the low
dose region.  Then potency, P,  is defined as a/d* - P, where a is synonymous
with risk.   We propose that a - 0.01 is reasonable for comparison across
chemicals for this purpose, although o could be set to any level of interest.
A value  for a of 0.01 has been proposed in alternative methodologies  to
evaluate noncancer health effects (Crump, 1984).  Finally, based on this
potency, any release that would result in an a > 0.01 (or any other level
deemed reasonable) could be regarded as having exceeded the de minimis level.

Advantages  of Using a Potency-Based Approach for Short-term Benchmarks

      •      Data are derived by critical review of the primary
            literature.

      •      All of the data for a chemical can be used rather than just
            one study or one data set within a study and dose-response
            information can be included.

      •      The probability of the occurrence of an adverse effect for
            any dose/duration/severity combination can be estimated.

      •      The potency can be estimated for any risk level (a) deemed
            appropriate  for the chemical or application.

      •      Once the primary literature has been reviewed and the data
            have been compiled on data entry forms and entered  into the
            appropriate  model,  the process could be automated.

Disadvantages of Using a Potency-Based Approach for Short-term  Benchmarks

      •      Although much of the data needed by this scheme exist, a
            review of the primary literature and the running  of the
            suggested models would be labor intensive.
                                      31

-------
                                                                      DRAFT **
            This method has been applied to only 2 chemicals on the
            Section 112(b) list of chemicals.
REFERENCES:
Clement International Corporation (Clement).   1991a.   Health Effects  and Dose-
Response Assessment for Hydrogen Chloride Following Short-term Exposure.
Prepared for Air Risk Information Support Center, Office of Air Quality
Planning and Standards and Office of Health Evaluation Assessment,  U.S.
Environmental Protection Agency,  Research Triangle Park, NC.

Clement International Corporation (Clement).   1991b.   Consideration of
Multiple Health Endpoints in Source Category Ranking.  Prepared for Office of
Air Quality Planning and Standards, Pollution Assessment Branch, U.S.
Environmental Protection Agency,  Research Triangle Park, NC.

Crump, K.  1984.  A new method for determining allowable daily intakes.
Fundamental and Applied Toxicology 4:854-871.

Dourson, M. L. ,  Hertzberg, R. C., Hartung, R.,  and Blackburn, K.  1985.  Novel
methods for the estimation of acceptable daily intake.  Toxicology and
Industrial Health 1(4):23-41.

Hertzberg, R.  1989.  Fitting a model to categorical response data with
application to species extrapolation of toxicity.  Proceedings of 26th Hanford
Life Sciences Symposium. Modeling for Scaling to Man: Biology, Dosimetry, and
Response. Health Phys  (Suppl 1):405-409.

Hertzberg, R. and Miller, M.  1985.  A statistical model for species
extrapolation using categorical response data.  Toxicology and Industrial
Health 1(4):43-57.

Knauf, L. and Hertzberg, R.  1989.  Statistical methods for estimating risk
for exposure above the reference dose.  Internal Report, U. S.  Environmental
Protection Agency, Environmental Criteria and Assessment Office, Research
Triangle Park, NC.
                                       32

-------
                                                                   ** DRAFT *~

                                   SUMMARY


      As can be seen from the  approaches described above,  there is no single,
already established, method that provides an all encompassing scientifically
valid basis for determining short-term ambient benchmarks  for acute effects.
If the method gives a value for short-term exposures,  it is probably designed
to protect only workers at a facility.  In cases where the general population
is protected by the exposure limit,  it is probably for a lifetime exposure.
Any method used by OAQPS to establish de minimis levels will require some
modification; however, the amount of modification will vary for each method.

      Each of the methods described in this document is, to a greater or
lesser extent, flexible with few if any formal mechanisms  for defining the
types of data that are to be reviewed, criteria for deeming a study
scientifically valid, enumerating the uncertainty factors  that must be used,
and applying the method or exposure value to other similar chemicals.  In
part, this is a function of the variability of data available for a specific
substance.  While chemicals have been well characterized in terms of toxicity,
this is not the case for many  of the chemicals on the Section 112(b) list.  In
addition, all of the methods described in this document are resource
intensive, i.e., they require  a review of original data or, for most methods,
secondary data collected by health specialists e.g., supporting documentation
for TLVs.  Some methods, such  as the development of PELs,  also require that
technical and economic feasibility studies be performed.


Other Approaches

      The issue of ambient emissions and community exposure is also being
reviewed by other organizations.  The Chemical Manufacturers Association  (CMA)
has produced a document that presents eight approaches that may be used by
industry "for placing emission levels in context."  Four of the approaches  use
data sets as the starting point, i.e. , TLVs or NOELs, and then modify these
values by dividing them by fixed or variable factors.  A model-driven
approach, or health risk assessment, uses dose-response relationships to
estimate cancer potency and is only applicable to potential carcinogens.
Finally, three integrated methods are discussed in the CMA report including
the NOAEL adjustment method described above for Calabrese and Kenyon, a method
for dividing the TLV by fixed  or variable factors, and a method described by
Lewis et al. that adjusts the  NOEL by various factors depending on the type .
and quality of data.  As is the case with most of the methods described above,
the CMA methods focus on the long-term effects of airborne contaminants and
not short-term, acute exposures.

      A unique aspect of the CMA document is that is presents the advantages
and disadvantages for each method in  terms of resource  (personnel and data)
requirements, the scientific basis for the method, the  flexibility  of the
method, the complexity of communicating the method and  the results  to the


                                      33

-------
                                                                   ~* DRAFT **

general public as  well  as  scientific experts,  and whether or not the method
has been used by public and regulatory agencies.   Because several of the
methods discussed  in the CMA document are variations  of those discussed in
this document, such as  the Calabrese method,  the  advantages  and disadvantages
described for each CMA  approach are applicable to the approaches discussed in
this report.

      The National Research Council (NRC) of the  National Academy of Sciences
is conducting a study to develop methodologies for setting short-term exposure
levels for the EPA Office  of Toxic Substances and the Agency for Toxic
Substances and Disease  Registry (ATSDR) .   It is expected that the exposure
limits will be used by  the EPA as guidelines for  chemicals on the Extremely
Hazardous Substance list and by ATSDR as guidelines for personnel exposed
during evaluations of hazardous waste sites.   The NRC is currently reviewing
various methods used to develop exposure levels,  including methods for the
development of EEGLs, SPEGLS,  and ERPGs.   The methods eventually developed by
the NRC will allow exposure limits to be set on a chemical-by-chemical basis
using currently available  data.

      A dose-response approach for acute exposures and noncancer endpoints has
been developed by Clement  and has been applied to the data for hydrogen
chloride.  This method has undergone internal EPA review and may be useful in
establishing de minimis levels under Section 112(b).   The model takes dose,
duration, and severity  of  the response into consideration and estimates the
probability that a response will occur at a specified level of severity (or
the no observed effect  level if the lowest severity category is so defined)
for any dose-duration combination.


Recommended Approach

      From che methods presented in this document, it would appear  that on the
basis of flexibility and scientific validity, the approach of Calabrese and
Kenyon is most appropriate for setting de minimis levels for short-term acute
exposures as an interim method.  This method  is very well defined and  is
flexible enough to incorporate the available  data for any given chemical.  By
allowing the use of either an occupational exposure limit or a NOAEL or LOAEL
based on animal or human data, an AALG may be derived for any chemical  on the
Section 112(b) list.  Calabrese and Kenyon provide guidance  on  selecting
appropriate uncertainty factors and determining whether  a particular  study
meets the criteria for scientific merit.  ^Although they  assume  that the
ambient air levels are to be protective  for  long-term effects,  their  approach
is easily applicable to short-term, acute exposures.  This may  be  done  by not
extrapolating from acute exposures  (which are often more readily  available)  to
Chronic exposures.In addition, AALGs and  their  rationales  are  provided  for
nTany of the substances on the  Section  112(b)  list, which provides  a convenient
starting point for deriving short-term,  acute exposure  limits.

-------
                                                                   ** DRAFT **

      The following is a brief review of the scheme for deriving AALGs,  as
presented by Calabrese and Kenyon,  and which may be used for che development
of de minimis levels.

      Step 1: Choosing Bndpoints of Concern for Each Chemical for Which an
      AALG will be Derived

      Calabrese and Kenyon use several approaches for deriving AALG values.
One approach is specific for carcinogens,  and others are specific for either
(1) genotoxicity;  (2)  developmental,  reproductive, and systemic toxicity;  or
(3) skin and eye irritation.  Data used to derive an AALG for a given chemical
should be categorized into one or more of these effects.  Subsequently,  the
endpoint of most concern would be used as the basis for the AALG.  If more
than one endpoint applies,  an AALG could be calculated for each endpoint and
the most conservative AALG could be chosen.

      For the initial screening, the RTECS data base is used as a source of
information for determining which of the above categories apply to the
chemical of interest.   From RTECS,  health effects can be identified and the
chemicals can be categorized by the different effects.  However, this is only
a preliminary step.  The studies in RTECS that describe specific health
effects should be retrieved to confirm the cited information.

In addition, these effects should be confirmed using some or all of the
following secondary data sources:


            IARC Monographs
            ACGIH TLV documents
            NIOSH Criteria Documents or Current Intelligence Bulletin (CIBs)
            NTP Carcinogen Technical Report
            EPA Gene Tox data base
            Ambient Water Quality Criteria Documents
            Drinking Water Criteria Documents
            Drinking Water Health Advisories
            ATSDR Health Profiles
            EPA Health Assessment Documents
            EPA Health Effects Documents
            Integrated Risk Information System (IRIS)


Once the crucial endpoint(s) for a chemical are positively identified and
verified, the principal and supporting studies that confirm  the health effects
should be used to derive the AALGs.

      As mentioned previously, Calabrese and Kenyon present  schemes for
deriving AALGs for each of the endpoints indicated above.  However, only  the
scheme for developmental, reproductive or systemic toxicants will be discussed
here as an example of how the schemes may be applied to a specific chemical.


                                      35