United States
Environmental Protection
Agency
Office of Research
and Development
Washington, DC 20460
EPA600-R-97-12a
December 1997
www.epa.gov
Research Plan for
Microbial Pathogens and
Disinfection By-Products
in Drinking Water
Cryptosporidium
parvum (small) and
Giardia lamblia (large)
-------
-------
EPA/600/R-97/122
December 1997
Research Plan for Microbiai Pathogens and
Disinfection By-Products In. Drinking Water
Office of Research and Development
Office of Water
U.S. Environmental Protection Agency
Cincinnati, OH 45268
Printed on Recycled Paper
-------
Notice
The U.S. Environmental Protection Agency (EPA) through its Office of Water and
Office of Research and Development generated the research described. It has been
subjected to the Agency's peer and administrative review and has been approved for
publication as an EPA document. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
-------
Contents
Chapter 1. Introduction 1-1
Purpose 1-1
The Problem 1-1
Regulatory Background 1-1
Drinking Water Treatment—Brief Overview 1-3
Policy Questions and Research Goals 1-3
Microbial Pathogens—Research Needs ; 1-6
Disinfection By-Products—Research Needs 1-8
Criteria for Priority Setting 1-10
Contents of the Following Chapters 1-10
Chapter II. Balancing Microbial and DBP Risks: Integrating Research to Support
Rule Development . 2-1
Overview of Approach to Balancing Microbial and DBP Risks 2-1
Integrating Research to Support Rule Development 2-1
Chapter III. Research for Microbial Pathogens 3-1
Background 3-1
Health Effects Research 3-1
Exposure Research 3-4
Risk Assessment Research 3-12
Risk Management Research 3-14
Chapter IV. Research for Disinfection By-Products 4-1
Background 4-1
Health Effects Research 4-1
Exposure Research 4-8
Risk Assessment Research 4-12
Risk Management Research 4-16
-------
-------
Chapter I
Introduction
Purpose
This document describes the research needed to sup-
port EPA's development of drinking water regulations for
community systems concerning disinfectants, disinfec-
tion by-products (DBFs) and microbial pathogens. It
does not specifically address the needs of non-commu-
nity systems; however, much of the information that will
be generated by this research will be relevant to non-
community systems as well. While the research needs
identified in this plan are extensive, the plan is intended
to focus on the key scientific and technical information
needed to develop sound regulations. EPA is committed
to conducting a substantial research program that will
address many of the priority research needs identified in
the plan. EPA also hopes to use this plan as a vehicle for
discussion with outside groups, including those who
participated in the DBP regulatory negotiation, to reach
agreement on the components of the plan. EPA antici-
pates that the plan can then be used to promote coordi-
nation and cooperation among the various agencies,
private organizations, and universities that are involved
in DBP and microbial research. EPA envisions continu-
ing the stakeholder meetings to facilitate coordination of
research.
The Problem
For nearly 100 years, public water supplies have been
treated with a variety of chemicals targeted at the reduc-
tion or elimination of infectious disease risk. But risks
remain, as evidenced by the EPA Science Advisory
Board's 1990 ranking of pollutants in drinking water as
one of the highest health risks meriting EPA's attention.
This ranking was based on the exposure of large popu-
lations to known and unknown contaminants, including
lead, DBPs, and disease-causing microorganisms.
The continued occurrence of waterborne disease out-
breaks demonstrates that contamination of drinking wa-
ter with pathogenic bacteria, viruses, and parasites still
poses a serious health risk when treatment is inad-
equate or when contamination occurs in the distribution
system. Thirty-four outbreaks of waterborne disease
were reported in 1991-1992, and the causative agent
was not identified in most of these outbreaks. A 1993
outbreak of Cryptosporidiosis in Milwaukee, which re-
sulted in an estimated 400,000 cases of acute gastroen-
teritis, represents the largest documented occurrence of
disease associated with contamination of a treated pub-
lic water supply in the U.S. In addition to these known
outbreaks, many others undoubtedly occur each year
but are either unrecognized or unreported. Cases not
associated with an outbreak (endemic or opportunistic
disease) may also be significant, but little is known
about the extent to which this is a problem.
To combat waterborne microbial diseases, public water
systems disinfect drinking water with chlorine or alterna-
tive disinfectants, such as ozone, chloramines, or chlo-
rine dioxide. However, the use of chlorine or other
disinfectants, while reducing microbial risks, creates
new potential risks, because compounds known as dis-
infection by-products are formed during the water treat-
ment process. A wide variety of by-products have been
identified, a number of which have been shown to cause
cancer and other toxic effects in animals under experi-
mental conditions. Additionally, some epidemiology stud-
ies have suggested that consumption of chlorinated
water may be associated with elevated rates of cancer
and adverse reproductive outcomes. In reviewing these
data, however, most experts agree that the scientific
evidence is inconclusive with regard to the significance
of adverse health risks from exposure to disinfected
waters.
The challenge in providing safe drinking water today lies
in adequately characterizing the risks and then reaching
an acceptable balance among competing risks. Increased
disinfection can reduce microbial risks but can increase
the potential risk from disinfection by-products. The
optimal balance will adequately control risks from patho-
gens, simultaneously control DBPs to acceptable levels,
and ensure that costs of water treatment are commen-
surate with public health benefits. To enable EPA to
develop regulations that will achieve this balance, re-
search is needed to obtain a better understanding of the
potential health risks and human exposures to patho-
gens and DBPs. Research is also needed on water
treatment processes and other means of reducing these
risks.
Regulatory Background
The Safe Drinking Water Act (SDWA) mandates that
EPA identify and regulate drinking water contaminants
that may have any adverse human health effects and
which are known or anticipated to occur in public water
1-1
-------
systems. The SDWA also requires the use of filtration
and/or disinfection for public water supplies that serve
most of the U.S. population. Currently, several regula-
tions attempt to control for DBFs and pathogens in
public drinking water supplies:
• Interim total trihalomethane (TTHM) standard (pro-
mulgated 1979)—This standard is applicable to all
community water systems that disinfect and serve at
least 10,000 people. Systems must achieve less
than 0.10 mg/l of total trihalomethanes as an annual
average based on quarterly measurements in the
distribution system.
• Total coliform rule (promulgated 1989)—This stan-
dard is applicable to all public water systems. Sys-
tems must demonstrate that the frequency of total
coliform presence is below acceptable limits; sam-
pling frequency is based on population served. Small
systems that collect fewer than 5 samples per month
must conduct periodic sanitary surveys.
* Surface water treatment rule (SWTR) (promulgated
1989)—This standard is applicable to all public wa-
ter systems that use surface water or groundwater
under the direct influence of surface water. Systems
must achieve at least 3 and 4 log removal and/or
inactivation for Giardia and viruses, respectively,
and if filtration is not part of the treatment process,
the system must meet specific criteria for avoiding
filtration.
A more comprehensive regulatory strategy is needed to
address microbial and DBP contaminants in drinking
water. The TTHM standard addresses only one class of
by-products. While control of TTHM during the treatment
process may also control formation of other by-products,
the extent of such a relationship is unknown. Addition-
ally, the TTHM standard applies only to larger systems,
so the same protection is not afforded to people served
by smaller water systems. While the Total Coliform Rule
and SWTR apply to all system sizes, it is not clear that
achieving these standards will provide adequate protec-
tion from protozoa, such as Giardia and Cryptospo-
ridium, especially when a system uses a poor quality
source water.
In 1992, EPA initiated a negotiated rulemaking to evalu-
ate the need for additional controls for microbial patho-
gens and disinfection by-products. The negotiators in-
cluded representatives from state and local health and
regulatory agencies, public water systems, elected offi-
cials, consumer groups, and environmental groups. The
major goal of the Negotiating Committee was to develop
an approach that would reduce the level of exposure
from disinfectants and DBFs without undermining the
control of microbial pathogens. Early in the regulatory
negotiation process, participants agreed that large
amounts of information necessary to understand how to
optimize the use of disinfectants and concurrently mini-
mize microbial and DBFs' risk were unavailable. Be-
cause of this lack of data the negotiators agreed that
there should be a two-stage DBP rule and Long Term
Enhanced Surface Water Treatment Rule (LTESWTR).
The Stage 1 DBP rule would be proposed, promulgated
and implemented concurrently with the Interim Enhanced
Surface Water Treatment Rule (IESWTR) in order to
ensure that microbial risk was not increased as the
Stage 1 DBP rule is implemented. The Stage 2 DBP rule
would follow after additional information on health risk,
occurrence, treatment technologies, and analytical meth-
ods were developed in order to better understand the
tradeoffs between microbial pathogens risk and risks
from DBFs. Each rule is described below.
The proposed Stage 1 DBP rule included maximum
contaminant levels of 0.08 mg/L for TTHMs, 0.06 mg/L
for five haloacetic acids (HAAS), 0.01 mg/L for bromate,
and 1.0 mg/L for chlorite along with the best available
technologies to control for these DBFs. The proposed
Stage 1 DBP rule also included lower MCLs for TTHMs
(40 jig/L) and HAAS (30 jig/L) as a "placeholder" to
assure participants favoring further DBP controls that
other members would return for Stage 2 DBP negotia-
tions. For the Stage 2 DBP rule, the negotiators agreed
that EPA would collect data on the parameters that
influence DBP formation and occurrence of DBFs in
drinking water through the Information Collection Rule
(ICR). Based on this information and new data gener-
ated through research, EPA would reevaluate the Stage
2 DBP "placeholder" provisions and repropose, as ap-
propriate, depending on the criteria agreed on in a
second regulatory negotiation.
The SDWA was reauthorized in 1996. The 1996 SDWA
amendments required EPA to conduct research in sev-
eral areas including research to better understand the
mechanisms by which chemicals cause adverse effects;
research on new approaches for studying the adverse
effects of contaminant mixtures in drinking water; stud-
ies to identify subpopulations that are at greater risk
than the general public from exposure to contaminants
in drinking water; and pilot waterborne disease occur-
rence studies in at least five major U.S. communities in
collaboration with The Center for Disease Control (CDC).
EPA has provided special emphasis in this plan to
address these statutory requirements. The amendments
also provided an additional $10 million in 1997 for
conducting health effects research on drinking water
contaminants. Of this $10 million, about $9 million has
been targeted toward conducting research for microbial
pathogens and DBFs.
The SDWA amendments also established new dead-
lines for the M/DBP rules. The final Stage 1 DBP rule
and the IESWTR must be finalized by November 1998,
while the Stage 2 DBP rule must be finalized by May
2002 and the LTESWTR must be finalized by November
2000. The deadlines for promulgation of the Groundwa-
ter Disinfection Rule are between August 1999 and May
2002. The ICR was finalized in May 1996, with monitor-
ing starting in July 1997 and ending in December 1998.
1-2
-------
Long-term 1 enhanced surface water treatment rule
(LT1ESWTR) follows. (Long-term 2 enhanced surface
water treatment rule (LT2ESWTR) (final by 5/2002—not
a statutory deadline) will be developed to further im-
prove control for pathogens and will be promulgated in
conjunction with Stage 2 of the D/DBP rule in order to
prevent increased risk for systems complying with the
Stage 2 DBF rule.)
• Information collection rule (ICR) (proposed 2/94;
final 5/96)—Large public systems would be required
to collect approximately $130 million of occurrence
and treatment information concerning pathogens
and DBFs. Information collected under this rule
would be used with research to support the develop-
ment of the interim and long-term enhanced SWTR,
and Stage 2 DBF rule.
• Stage 1 Disinfectant/Disinfection By-Products (D/
DBP) rule (proposed 7/94; final 11/98)—This rule
would be applicable to all community water systems
that disinfect and is intended to reduce risks from
disinfectants and DBFs. Systems would be required
to achieve new limits for TTHMs, the sum concen-
tration for five haloacetic acids, bromate, chlorite,
chlorine, chlorine djoxide, and chloramines. Sys-
tems using conventional treatment (sedimentation
and filtration) would also be required to achieve
percent reductions of DBP precursors (measured as
total organic carbon), depending upon source water
quality, prior to disinfection.
• Interim enhanced surface water treatment rule
(IESWTR) (proposed 7/94; final 11/98)—This rule
would be applicable to all public water systems
using surface water or groundwater under the direct
influence of surface water that serve populations of
10,000 or greater. The purpose of this rule is to
enhance protection from pathogens, including Cryp-
tosporidium, and to prevent increases in microbial
risk while large systems comply with the Stage 1 D/
DBP rule. Several regulatory options were proposed,
including systems being required to achieve a) pro-
portionally higher levels of pathogen removal de-
pending upon pathogen measurements in the source
water, and b) fixed level removal requirements inde-
pendent of pathogen measurements in the source
water.
• Long-term enhanced surface water treatment rule
(LTESWTR) (final by 11/2000)—This rule, which
could include changes to the IESWTR, would ex-
tend applicability to surface water systems serving
less than 10,000 people. The purpose of this rule is
to enhance protection from pathogens, including
Cryptosporidium, and to prevent increases in micro-
bial risk for systems serving less than 10,000 people,
while they comply with the Stage 1 D/DBP rule.
• Stage 2 DBP rule (proposed in part 7/94 with the
Stage 1 D/DBP rule; final by 5/2002)—This rule
would be applicable to all community water systems
that disinfect and is intended to further reduce the
levels of risk achieved under the Stage 1 rule. Only
tentative limits were proposed for TTHMs and the
sum concentration for five haloacetic acids.
• Groundwater disinfection rule (GWDR) (final be-
tween 8/99 and 5/2002)—This rule would be appli-
cable to all public water systems using groundwaters
not under the direct influence of surface water. This
rule would require all vulnerable systems to disin-
fect. This rule is intended to enhance protection
from pathogens as well as to prevent increases in
microbial risk while systems comply with the Stage 1
D/DBP rule.
Figure 1-1 depicts the schedule of rules being developed
and the concurrent ongoing activities of the ICR and
research on microbial pathogens and DBFs.
Drinking Water Treatment—Brief Overview
Figure I-2 shows the conventional treatment process for
surface water supplies along with the possible points of
disinfection. The disinfectants generally used in the U.S.
include chlorine, chloramines, ozone, and chlorine diox-
ide and various combinations. Chlorine is the most
common disinfectant used among water systems. About
80% of large water systems (serving greater than 10,000
people) use chlorine, while almost 100% of smaller
systems use chlorine. Chloramines are used by about
20% of the larger systems, while chlorine dioxide is used
by about 5% of larger systems, and ozone is used by
about 2% of the larger systems (the total does not add to
100% because some systems use combinations of dis-
infectants).
The formation of DBFs depends primarily on where in
the treatment process the disinfectants are added and
the source water quality parameters. Generally, the later
in the treatment process the disinfectants are added, the
fewer DBFs will be formed, all other factors being equal.
The most important water quality parameters that influ-
ence the formation of DBFs include the nature of the
source water, such as the content of precursor materi-
als, water temperature and pH, and conditions under
which the disinfectant is used, such as the concentra-
tion, contact time, point of addition, and residual main-
tained.
Policy Questions and Research Goals
In making decisions on appropriate regulatory levels for
pathogens and DBFs in drinking water, policy-makers
will focus on three major questions:
• What are the health risks caused by exposure to
microbial pathogens?
• What are the health risks caused by exposure to
DBFs from different treatment processes?
• How can these risks be simultaneously controlled?
1-3
-------
CD
OC
g
CD
s
CD
•0=
III"
•
cc
-
O)£Q
S c
TO O
B. WQQtaSdocp:
lll
1-4
-------
Chemical addition
Clearwell
Raw water
(organics,
microbes,
particulates)
Possible disinfection points
(chlorine, chloramines,
ozone, chlorine dioxide)
Distribution
system (DBFs)
Figure 1-2. Conventional treatment (surface supplies).
To provide a specific example, a critical issue will be
whether ozonation should be encouraged in place of
chlorination. The U.S. has many years of experience
with the use of chlorine for disinfection, and most of the
research on DBFs has focused on chlorination by-prod-
ucts. Other disinfectants, such as ozone, chloramines,
and chlorine dioxide, are less widely used than chlorine
and less research has been conducted on their by-
products. Recent research indicates that ozone is much
more effective in inactivating Cryptosporidium than chlo-
rine. This may provide an impetus for many water utili-
ties to switch to ozone as a disinfectant. However,
significant uncertainties remain in assessing the health
risks resulting from chlorination and use of alternative
disinfectants. Before making any major regulatory deci-
sions, policy makers need to know whether the risks
from use of alternative disinfectants are greater than,
less than, or equal to those from chlorination, and whether
pathogens can be adequately controlled with improve-
ments to the chlorination process.
To support the policy-making process, the research
goals are
1. To identify the health effects caused by DBFs and
microbial pathogens in drinking water.
2. To determine the population distribution of exposure
to DBFs and microbial pathogens to which people
are exposed.
3. To assess the risks caused by DBFs and microbial
pathogens in drinking water.
4. To evaluate the effectiveness of options for reducing
risks from DBFs and microbial pathogens.
The first three goals are all part of the risk assessment
process—the research proposed in this plan is intended
to fill data gaps and reduce the major uncertainties in
current estimates of risks from microbial pathogens and
DBFs. The fourth goal addresses reducing the risks.
Research is needed to fill major gaps in current knowl-
edge about the effectiveness of treatment processes in
simultaneously controlling pathogens and DBFs.
The importance of research to support the decision
making process for microbes and DBFs in drinking
water is further highlighted by the strong research provi-
sions of the 1996 SDWA Amendments. This legislation
specifically requires EPA to place a high priority on
studies concerning the health effects of Cryptosporidium
and DBFs, as well as on studies of subpopulations at
greater risk of adverse effects from exposure to drinking
water contaminants. The Agency is required to conduct
research to support the ESWTR, D/DBP rule, and the
GWDR, particularly in the areas of cancer and reproduc-
tive effects (toxicology and epidemiology studies), and
dose-response studies for waterborne pathogens such
as Cryptosporidium and Norwalk virus.
To ensure the credibility and soundness of EPA's regu-
latory decision-making, it is imperative that all research
is of high technical quality and follows EPA quality
assurance guidelines. EPA policy requires peer review
for all major scientific and technical studies that support
regulatory decisions. EPA expects that all research ac-
tivities supporting this research plan will undergo appro-
priate peer review by independent experts.
The following sections provide an overview of research
needs for microbial pathogens and DBFs, in the areas of
health effects, exposure, risk assessment and risk man-
agement. Each section begins with highlights of key
scientific issues and then lists research questions and
research needs.
1-5
-------
Microbial Pathogens—Research Needs
Health Effects Research for Microbial
Pathogens
Current knowledge of waterborne pathogens is inad-
equate for assessing their health risks. While the dis-
ease symptoms caused by pathogens are generally
known, limited information is available on the doses and
conditions that produce effects. For example, EPA has
estimated risks from Cryptosporidiumusing the assump-
tions that ingestion of a single organism can cause an
infection and that the probability of infection from one
organism can be predicted by extrapolation from higher
dose studies. However, there is considerable uncer-
tainty surrounding these assumptions. Additionally, there
are uncertainties with regard to infectivity in susceptible
subpopulations, the significance of the immune response,
and the variation among different strains of Cryptospo-
ridium in infectivity and virulence. Research is needed
on dose-response relationships, and to evaluate the
validity of predicting risks from dose-response curves.
The incidence of waterborne disease in the U.S. is
highly uncertain. Further research is needed to deter-
mine rates of waterborne illness, and, where elevated
rates of illness are identified, to determine whether
illness is caused by inadequately treated water or by
water quality that has deteriorated in the distribution
system.
Research Questions
• What are the waterborne pathogens of public health
concern?
• What is the nature and magnitude of disease asso-
ciated with exposure to these water borne agents?
Research Needs
• Information on the pathobiology of infection and
disease for waterborne pathogens, including data
on dose-response relationships for various patho-
gens and different organism strains, pathogen- and
host-specific factors involved in infection and dis-
ease, and effects on different subpopulations.
• Epidemiology studies to characterize endemic and
epidemic illness rates, to assess magnitude of risk,
and to provide data for use in verifying risk models.
Exposure Research for Microbial
Pathogens
Little information is available on the levels of pathogens
that occur in drinking water. Current techniques lack the
precision and specificity required to measure low levels
of pathogens. Even with improved measurement tech-
niques, it is very difficult to detect very low levels of
organisms in tap water. To estimate exposures from tap
water, EPA expects to rely on measurements of patho-
gens in source water, combined with estimates of the
removal and inactivation of pathogens by different treat-
ment processes. For systems relying on groundwater,
information is needed on the survival and transport of
pathogens in the subsurface. Currently, each of these
estimation steps has significant uncertainties.
Research Questions
• What methods are needed to adequately measure
or estimate occurrence of pathogens in drinking
water?
• What are the frequencies of occurrence and densi-
ties of pathogens in source water, finished water,
and distribution systems, and what is the population
distribution of exposures to pathogens?
• What are the factors affecting microbial contamina-
tion of groundwater?
Research Needs
• Analytical methods to detect and enumerate proto-
zoa, including methods for protozoa that indicate
whether the organism is viable and/or infectious.
• Analytical methods to detect and enumerate viruses
in source and finished waters.
• Occurrence information for pathogens in source water
and finished waters.
• Assessment of sources of pathogens and the impor-
tance of watershed controls.
• Occurrence information for primary and opportunis-
tic pathogens in distribution systems.
• Determination of survival and transport of patho-
gens in groundwater under different conditions.
• Assessment methods for protecting groundwater
sources from pathogens, including methods for de-
termining whether "natural disinfection" is adequate.
Risk Assessment Research for Microbial
Pathogens
Assessing the risks of waterborne pathogens depends
on adequate effects and exposure information. As indi-
cated above, EPA is likely to rely on estimates of source
water pathogen occurrence, and removal and inactiva-
tion by treatment and transport through the distribution
systems to predict concentrations in tap water. In addi-
tion, outbreak data will be assessed to improve our
understanding of real time microbial risks especially as it
applies to actual source water conditions, treatment
techniques, distribution system integrity and health im-
pacts. This information will also be used to improve our
comparative risk modeling. These data will be combined
1-6
-------
with information on dose-response, including sensitive
subpopulations, and water consumption to predict risks.
A comprehensive risk assessment model is needed for
assessing the risks from pathogens in drinking water.
The currently accepted risk assessment models were
developed for assessing chemical risks and do not fully
accommodate issues important to the assessment of
pathogenic risks, such as microbe/host interactions, sen-
sitive subpopulations and consideration of secondary
spread of infection. Dose-response models are needed
that can address threshold and non-threshold assump-
tions within a modified risk assessment paradigm for
microbes.
Research Question
• How can the risks posed by pathogens in drinking
water be characterized?
Research Needs
• Modification of risk assessment paradigm for micro-
bial disease.
• Development and application of dose-response sta-
tistical models and other methods/tools for assess-
ing microbiological disease.
• Methods to characterize risks from mixtures of patho-
gens, and mixtures of pathogens and DBPs.
Risk Management Research for Microbial
Pathogens
Improved methods for analyzing and protecting source
waters are needed to help assure that treated water is of
acceptable quality. Better data on pathogen removal
and inactivation efficiencies of various treatment pro-
cesses are needed to provide better estimates of risk
and to select the treatment processes with the most
potential for risk reduction. Research efforts that evalu-
ate treatment effectiveness should simultaneously evalu-
ate control of DBPs (see discussion below). The great-
est need is to address the uncertainties surrounding
treatment efficiencies for pathogens by focusing re-
search on those organisms that are most resistant to
treatment. For example, current data indicate that Cryp-
tosporidium is much more resistant to disinfection than
most other waterborne pathogens. In surface waters, by
focusing on treatment processes that are effective for
inactivating Cryptosporidium, it is likely that other water-
borne pathogens, including those that may be more
infectious or have more significant health effects, will
also be effectively reduced. However, Cryptosporidium
analysis is expensive and difficult. Because of the ana-
lytical difficulties in measuring Cryptosporidium in drink-
ing water, an important need is to identify surrogate
parameters for evaluating treatment effectiveness. Po-
tential surrogates include indicator organisms and engi-
neering process control parameters such as particle
size counting and disinfection operating conditions. In
addition, as the existence and significance of new, po-
tentially dangerous waterborne pathogens are recog-
nized, there is a need to determine the effectiveness of
various treatment systems to control these new organ-
isms.
In groundwaters, where protozoa are not expected to
occur, it is important to consider viruses when defining
adequacy of treatment. Therefore, there is a need to
develop information on the effectiveness of alternative
treatments for controlling viral contamination in ground-
water sources. Currently, EPA uses disinfection data on
hepatitis A virus (HAV) for estimating disinfection condi-
tions necessary to inactivate viruses in general. HAV
was selected as the target virus because it has been
implicated in waterborne disease, it is among the vi-
ruses with the most significant adverse health effects,
and it is among those most resistant to disinfection.
Norwalk virus may be more resistant to disinfection than
HAV and therefore may be more suitable than HAV for
defining adequacy of disinfection.
While protozoa and viruses might be effectively man-
aged at the treatment plant, bacteria pose a special
problem because of their ability to grow in the water
distribution system. A number of the bacteria species
identified in distribution systems are opportunistic patho-
gens, which can cause illness in certain population
subgroups. Treatment processes at the plant can affect
bacterial growth; for example, use of ozone to more
effectively control protozoa and viruses may, depending
on the organic content of the water, increase nutrient
levels in the treated water and thus potentially increase
bacterial growth in the distribution system. In addition,
there is a need to understand the causes of microbial
intrusion into the distribution system and identify effec-
tive approaches to prevent this from happening.
Research Questions
• How effective are various treatment processes in
removing pathogens?
• How can the quality of treated water be maintained
in distribution systems?
• How can source water be protected to ensure that it
is consistent with finished water quality of accept-
able microbial risk after appropriate treatment?
Research Needs
• Methods to analyze and protect source waters.
• Evaluation of the effectiveness of different treatment
processes in controlling pathogens, with a focus on
Cryptosporidium, including analysis of the effects of
treatment on the viability/infectivity of Cryptospo-
ridium oocysts that pass through the treatment plant,
and on surrogate indicators of treatment effective-
ness. Technologies to evaluate include optimization
1-7
-------
of conventional treatment, filtration, use of sequen-
tial disinfectants, and biological treatment.
• Evaluating the effectiveness of various treatment
technologies to control new, potentially waterborne
pathogens as their existence and significance is
recognized.
• Development and evaluation of technologies appro-
priate for small systems, which face constraints on
cost and operational complexity.
• Evaluating the effectiveness of treatment alterna-
tives to control viral contamination in groundwater
sources.
• Identification and characterization of factors which
influence microbial growth in distribution systems,
and development of strategies to control such growth.
• Understanding the causes of microbial intrusion into
the distribution system and developing cost effective
approaches for preventing such intrusion.
Disinfection By-Products—Research
Needs
Health Effects Research for Disinfection
By-Products
In the 20 years since the discovery of chloroform and
other trihalomethanes in drinking water, a major re-
search issue has been the evaluation of the adverse
health effects of DBFs. However, significant gaps re-
main in our knowledge of the actual risks to human
populations posed by the use of disinfectants, especially
from alternative disinfectants such as ozone and chlo-
rine dioxide. A number of epidemiological studies have
been conducted, but they have generally been inconclu-
sive. Some studies have shown no association between
consumption of chlorinated water and cancer, while
others have suggested weak to moderate association
with cancers of the colon, rectum and bladder. Epidemio-
logical studies of reproductive effects have been simi-
larly inconclusive. Toxicological studies using laboratory
animals have shown that a number of DBFs cause
cancer, reproductive toxicity, and other effects, but at
concentrations higher than typically are found in drinking
water.
Health effects research is needed in three areas. Epide-
miological research should be pursued, with a focus on
avoiding the inadequacies in previous studies: study
design, characterization of exposure, and ascertainment
of health effects. Toxicology studies on individual DBFs
should be conducted to fill key data gaps and to facilitate
extrapolation of animal data to humans. Additionally,
toxicology research using complex mixtures of DBFs
would be valuable, if the technical difficulties in conduct-
ing such mixtures studies can be overcome. For ex-
ample, epidemiological studies on ozonated water will
not be able to evaluate cancer endpoints, because
ozone is a recent technology in the U.S. and long-term
exposures have not occurred for large numbers of people.
EPA will have to rely on individual chemical toxicological
studies and mixtures studies to evaluate long-term health
effects of ozonation by-products.
Research Questions
• What are the health effects associated with expo-
sure to DBFs?
- What are the health effects in communities
served by disinfected drinking water?
- What is the toxicity of individual chemical
contaminants and of mixtures of DBFs?
Research Needs
• Improved methods for epidemiological research, in-
cluding better approaches for assessing exposures
and health effects.
• Epidemiology research to provide qualitative and
quantitative information on the risks of cancer, ad-
verse reproductive outcomes, and possibly other
health effects that may be linked to DBF exposures.
• Toxicology studies on newly-identified or poorly-
characterized DBFs to provide basic information for
traditional hazard identification and dose-response
assessment.
• Pharmacokinetic and mechanistic data to facilitate
the extrapolation of animal data to humans.
• Animal studies on well-defined and complex mix-
tures of DBFs to help bridge the gap between
epidemiology studies and single-chemical toxicity
studies. The goal of these studies would be to
assess DBF interactions and to compare the poten-
tial toxicity of drinking water preparations that are
closer in composition to "real-world" mixtures.
Exposure Research for Disinfection By-
Products
DBFs are formed when disinfectants, for example chlo-
rine or ozone, are added to source and drinking water to
control microbiological organisms. Chlorine, the most
widely used and studied disinfectant, reacts readily with
humic substances from decaying animal and vegetable
matter (precursors) as well as organic contaminants,
i.e., pesticides, and produces a variety of chlorinated
products. When bromide ion is present, as it is in most
source waters to a greater or lesser degree, brominated
and mixed chlorobromo DBFs are also produced. Ozone
adds oxygenated species as does chlorine dioxide. An-
other disinfectant, chloramine, introduces the possibility
of nitrogen containing by-products such as cyanogen
chloride.
1-8
-------
Over 100 different compounds have been detected in
drinking water, most of these resulting from the use of
chlorine. Less is known about by-products from ozone
and chlorine dioxide. For example, it was only recently
recognized that ozone disinfection can produce bro-
mate, an ion which may pose significant health risks.
The characteristics of the source water, the type of
disinfectant used, and how the disinfectant is used are
all factors which influence the types and quantities of by-
products formed.
Current exposure information is inadequate for conduct-
ing a quantitative exposure assessment for the majority
of DBFs. Research is needed to better characterize the
by-products formed from different treatment processes,
especially from alternative disinfectants. Better analyti-
cal methods are needed to assess the frequency and
magnitude of occurrence of DBFs, particularly for polar,
water soluble, and nonvolatile DBFs. Because of limited
methods, there is a paucity of exposure data. We do
know, however, that the wide range of disinfection meth-
ods and background constituents in water supplies will
result in a wide range of exposure differences.
Research Questions
• What methods are needed to adequately measure
occurrence of DBFs in drinking water?
• What levels of DBFs are people exposed to via their
drinking water supplies, and what is the population
distribution of exposures?
Research Needs
• Improved methods of analysis for DBFs of concern,
including practical methods for chemicals likely to
be regulated in the near future, and research meth-
ods to aid in the discovery and characterization of
DBFs
• Identification of new DBFs, particularly from alter-
nate disinfectants
• Data on exposures to DBFs to develop, improve,
and validate exposure models
Risk Assessment Research for
Disinfection By-Products
Determining the health risks caused by DBFs is a critical
part of the DBF research program—current evidence is
inconclusive, and the implications for water treatment
expenditures are significant if lower levels of DBFs are
required. As indicated above, some epidemiological stud-
ies have suggested that elevated rates of cancer may be
linked to consumption of chlorinated water. Estimates of
excess cancer risk vary from 0 to 10,000 or more cancer
cases per year when extrapolated to the U.S. population
as a whole. This wide variation in risk estimates results
from the application of different methods for estimating
risks and different exposure assumptions. Narrowing
the broad range in cancer estimates will depend on
health effects and exposure research and on advances
in risk assessment methodology. In addition to the un-
certainties surrounding cancer risk estimates associated
with these chlorinated water studies, uncertainties exist
for other risks such as reproductive and developmental
effects.
The traditional risk assessment process has been fo-
cused on the evaluation of risks from exposure to an
individual chemical or members of a specific class of
chemicals. But exposure to DBFs in drinking water, as is
true for many other environmental exposures, is really
exposure to a complex mixture of chemicals. Assess-
ment of these compounds or classes of compounds as
components of their source water could have a mean-
ingful impact on the understanding of the actual risks
resulting from exposures to these various mixtures. The
risk assessment should therefore take into account pos-
sible interactions between chemicals and evaluate the
impact on health risks.
Research Questions
• How can we better characterize the risk posed by
exposure to specific DBFs in drinking water?
• How can we characterize the risk posed by expo-
sure to multiple or complex mixtures of DBFs?
• How can the risks from chemicals and microbes be
compared?
Research Needs
• For individual chemicals, refinement and application
of new models for estimating cancer and noncancer
risks, in particular risks associated at exposures
typically found in drinking water.
• Improved estimates of risk using epidemiologic data
from previous and ongoing studies, incorporating
improved estimates of exposure where possible.
« For mixtures of chemicals, application of new meth-
ods for assessing risks that include characterization
of chemical interactions.
• Assessment of the co'mparative toxicity of DBFs
within classes or families of compounds and also
between families.
• For chemicals and microbes, a framework for as-
sessing and comparing risks from different treat-
ment options.
« Better methods for screening and prioritizing poten-
tial risks.
1-9
-------
Risk Management Research for
Disinfection By-Products
Better data are needed on how to control DBFs formed
, by different treatment processes. Simultaneously, can-
didate treatment processes must be evaluated for con-
trol of microbial pathogens. Treatment options include
optimizing conventional treatment; switching to alternate
disinfectants, such as ozone; use of granular activated
carbon (GAG); or use of membrane technology. The
range in treatment costs for the different options is
tremendous. EPA has estimated that the increased
household costs for systems serving 250,000 people
and using conventional treatment would be $5/year
(optimized treatment), $10/year (ozone/ chloramination),
$60/year (GAG), or $120/year (membrane technology)
depending upon which technology might be needed to
comply with new DBF standards. For a typical system
serving 2,500 people and using conventional treatment
the household costs could increase by $10/year (opti-
mized treatment), $60/year (pzone/chloramination), $270/
year (GAG or membrane technology) depending upon
which technology might be needed. Clearly, research
that could lead to improvements in conventional treat-
ment and could demonstrate that acceptable levels of
pathogens and DBFs can be achieved will be highly
cost-effective.
Current regulations only set a limit for total trihalometh-
anes, relying on the assumption that practices that
control formation of these chemicals will also control
other by-products. Research is needed to determine
whether this assumption is valid, or whether other surro-
gate parameters can act as good indicators of DBF
control. Depending on the outcome of the health re-
search, other DBFs may become of greater concern.
Research Question
* How effective are various treatment processes in
minimizing the formation of DBFs?
Research Needs
• Evaluate processes that prevent DBF formation by
reducing precursor compounds in source water.
• Evaluate use of different disinfectants in limiting
DBF formation.
• Evaluate promising innovative technologies for small
systems to control precursors and DBF formation.
Criteria for Priority Setting
Chapters III and IV of this plan identify research projects
intended to meet the research needs described above.
The following criteria were used to select the research
projects that are included in this plan. All projects in-
cluded in the plan are generally considered as high
priorities for research. However, to provide a sense of
relative priority for the projects identified here, priority
rankings of High, Medium or Low were developed. The
criteria used for ranking are listed in descending order of
importance:
• High risk—The research is likely to elucidate the
concentrations, occurrence, or toxicity/pathogenicity
of a contaminant, where preliminary information sug-
gests that it may have a significant impact on public
health.
• High uncertainty—The research is likely to reduce
significant uncertainties associated with current as-
sessments or risk reduction technologies.
• Regulatory relevance—The likelihood is high that
research will lead to regulatory criteria that reduce
the risk to public health in a cost-effective manner.
Additional considerations which were relevant in setting
priorities for some projects:
• Short-term and long-term research needs—An ap-
propriate balance is achieved between research
. that addresses specific short-term needs (many ob-
jectives must be fulfilled with a 3-5 year time frame
to support regulation development) and research
that addresses more strategic, long-term needs (e.g.,
developing new methodologies for more compre-
hensive assessments of exposure to drinking water
contaminants).
• Linkage to other efforts—-The research is comple-
mentary to related research efforts in EPA (e.g., risk
assessment methods research), other federal and
state agencies, and the private sector.
• Anticipatory research—The results of the research
will help anticipate future drinking water problems.
• Wider applicability—The results of the research may
be extended to other drinking water or environmen-
tal issues.
Contents of the Following Chapters
Chapter II describes how the proposed research projects
presented in Chapters III and IV will be used to support
the various drinking water regulations described above.
Chapter II briefly describes how the research will assist
in estimating national/local costs and benefits of the
regulations and how this information can be used by
decision makers to balance the risks from microbial
pathogens and DBFs.
Chapter III and Chapter IV describe research needed for
microbial pathogens and DBFs, respectively. Within
Chapters III and IV, specific research projects have
been organized into the following categories: Health
Effects, Exposure, Risk Assessment, and Risk Manage-
ment, in some cases, the proposed projects cut across
these categories and a coordinated, interdisciplinary
approach is required. An example is project HE.M.7,
1-10
-------
Characterization of Endemic Disease: Health Effects
Associated with Differences in Source Water Quality
and Treatment Process, which would include health
effects, exposure, and risk management research.
The "State of the Science" sections for each major
research question describe past and current research
efforts by EPA and others. In the "Research Topics and
Priorities" section, projects that are listed are primarily
research which is needed but not yet underway. EPA
expects to be able to fund a significant portion, but not
all, of these projects. EPA's ongoing and completed
research has been included in the project listings, as this
research should be considered in the debate about
relative priorities.
The intent of this research plan is to propose research
that will answer the research questions presented gen-
erally in Chapter I and more specifically in Chapters III
and IV. However, the extent to which some of these
questions can be answered may significantly depend on
upon "the results of the research. The plan attempts to
indicate, in general terms, how research directions and
priorities may change depending on outcomes of re-
search.
1-11
-------
-------
Chapter II. Balancing Microbial and DBF Risks:
Integrating Research to Support Rule Development
Chapters III and IV will discuss research needs and
projects to address those needs for each of four re-
search topic areas (health effects, exposure, risk as-
sessment, and risk management) as they relate to achiev-
ing a better understanding of risks and control of those
risks posed by pathogens and DBFs. The purpose of
Chapter II is to indicate how research will be used to
support the development of the DBF Stage 1 rule, DBF
Stage 2 rule, IESWTR, LT1ESWTR, LT2ESWTR, and
GWDR.
Overview of Approach to Balancing
Microbial and DBF Risks
EPA is attempting to develop regulations that will result
in the proper balance between controlling risks from
pathogens and DBFs. Figure 11-1 a shows that in general
as the level of disinfection applied to drinking water
increases, the risk from exposure to pathogens de-
creases while the risk from DBFs increases. Clearly, the
type of source water treated, technology applied, micro-
organism removed and disinfectant used, will influence
the slope of the risk function. Figure 11-1 b shows how
cost/benefit analysis ideally might be used to minimize
social costs associated with the application of technol-
ogy to minimize exposure to pathogens and DBFs si-
multaneously. As the level of treatment to lower expo-
sure from both DBFs and pathogens increases, the cost
of treatment increases while the potential cost of health
damages decreases.
The ability to determine the proper balance between
controlling risks from pathogens and DBFs requires
enormous amounts of data, some of which will be devel-
oped as part of the research described in this document.
Since the occurrence of pathogens and DBFs in drink-
ing waters vary greatly, it is very difficult to estimate total
national risks or total national risk reductions that might
result from different regulatory options. Also, regulatory
decisions which lead to minimum social costs at the
national level may not be justified if they lead to large
adverse changes in social costs at the local level. Rec-
ognizing these problems, EPA is developing an ap-
proach for estimating national and local costs and ben-
efits (regulatory impact analysis [RIA]) that should sup-
port meaningful regulatory decisions. The results from
the research described in this plan, the ICR, and supple-
mental surveys will provide answers to many of the
research questions posed in Chapter I and will provide
input for estimating the costs and benefits of the various
rules and in balancing the risks from pathogens and,
DBFs.
Integrating Research to Support Rule
Development
Following is a discussion, by rule, of key regulatory
issues and how the research plan would be used to help
resolve these issues. Research projects are listed that
are either completed, currently ongoing by EPA, or
planned according to the rule they would most signifi-
cantly support. A key question in the development of the
(a)
•3
be
Source water
and technology
affect slope
DBFs
Regulatory
range
Microbial
Disinfection
Selecting best MCL
(b)
Total
social cost
Cost of
treatment
Cost of
health
damages
incurred
Figure 11-1.
Degree of treatment
Risk tradeoff and cost benefit considerations.
2-1
-------
drinking water regulations is whether the research will
be completed in time to support the development of the
various M/DBP rules. Tables 11-1, II-2, and II-3 provide
the dates by which the research will be completed and
the text under each rule provides a discussion on when
the research is needed. Before discussing the research
projects it is important to first understand how the re-
search will be used in developing regulations.
The overall goal of the SDWA is the establishment of
drinking water standards that will protect public health.
The SDWA specifies that EPA establish Maximum Con-
taminant Level Goals (MCLGs) for contaminants at which
no known or anticipated adverse effects on the health of
persons occur and which allow an adequate margin of
safety. MCLGs are nonenforceable health goals based
predominately on health effects data and estimates of
occurrence/exposure for the contaminant. Establishing
MCLGs for a DBF is the first step in deciding that the
contaminant poses a human health risk that requires
some control. When setting MCLGs for contaminants in
drinking water, EPA must also establish a Maximum
Contaminant Level (MCL) that is as close to the MCLG
as feasible. The SDWA defines feasible as "the use of
best available technology, treatment techniques and
other means which the Administrator finds, after exami-
nation for efficacy under field conditions and not solely
under laboratory conditions, are available (taking cost
into consideration)." To set an MCL, it has to be both
technologically and economically feasible to monitor the
contaminant in the drinking water. In cases where moni-
toring is not feasible, EPA is authorized to develop a
treatment technique requirement instead of an MCL.
The treatment technique requirement must prevent known
or anticipated adverse effects on the health of persons
to the extent feasible. Because of the difficulties in
monitoring for Giardia, viruses, or Cryptosporidium, both
the 1989 SWTR and the 1994 proposed IESWTR are
treatment technique rules. Each rule must also establish
compliance monitoring requirements to ensure the con-
taminant is being accurately characterized.
Under the 1996 amendments to the SDWA, EPA may
set an MCL at a level other than the feasible level (as
defined in the statute) if the technology or treatment
technique would result in an increase in the health risk
from other drinking water contaminants. In cases where
a treatment technique requirement is set at a level other
than the "feasible level," EPA is required to minimize the
overall health effects. This provision is relevant to the
LT2ESWTR and Stage 2 D/DBP rule, because the
control of microbial pathogens may increase the risk to
DBFs and the control of DBPs may increase the risk
from microbial pathogens. It is therefore important that
EPA have adequate knowledge of the extent to which,
under given conditions, reduction of one contaminant
results in an increase in another. In the case of the
LT2ESWTR, this means understanding under what con-
ditions and the extent to which changes in relevant
DBPs result from different microbial treatment technique
requirements, such as the inactivation of Cryptospo-
ridium. Other contaminants or parameters whose levels
could be increased by high disinfectant doses under
certain conditions include arsenic, lead, and assimilable
organic carbon (AOC). AOC provides the nutrients which
promote microbial growth in the distribution system,
possibly resulting in a series of problems that are not
well understood (e.g., growth of opportunistic pathogens
and coliform bacteria, rapid disinfection residual decay,
and corrosion).
As discussed above, the two primary inputs into the
MCLG are health effects data and occurrence/exposure
data. Although some data currently exist on the occur-
rence levels of DBPs and pathogens, the majority of
new occurrence data for DBPs will come from the ICR
and the majority of new occurrence data for pathogens
will come from the ICR, "mini-ICR," and supplemental
surveys. When developing MCLs and/or treatment tech-
niques the main inputs include the availability of analyti-
cal methods; effectiveness of treatment technologies to
control the contaminants; and the cqst of these tech-
nologies. When designing compliance monitoring strate-
gies the main inputs include the costs of monitoring and
the ability of monitoring to accurately characterize the
levels of the DBPs at consumers taps.
Surface Water Treatment Rules
The IESWTR was proposed in July 1994. The final
IESTWR is required by November 1998. The purpose of
the IESWTR is to prevent increases in microbial risk and
to enhance control for Cryptosporidium while systems
serving 10,000 or more people comply with the Stage 1
D/DBP rule. The IESWTR applies to surface water
systems serving at least 10,000 people. The Long-Term
1 ESWTR (LT1ESTWR) will apply to surface water
systems serving less than 10,000 people and is sched-
uled to be proposed in September 1999 and promul-
gated in November 2000. This rule is intended to im-
prove physical removal of Cryptosporidium and prevent
significant increases in microbial risk for smaller sys-
tems while they also comply with the Stage 1 D/DBP
rule. The LT2ESWTR will be developed to further im-
prove control for pathogens and will be promulgated in
conjunction with the Stage 2 D/DBP rule in order to
prevent increased risk for systems complying with the
Stage 2 rule.
Participants in the DBP regulatory negotiations were
concerned about the extent to which microbial risk might
increase if systems were to comply with new DBP
standards. RIA modeling analysis indicated that more
than one percent of the population in some systems
could become infected with Giardia if systems were
required to meet more stringent DBP standards and
comply with the existing SWTR (i.e., providing 3 logs
removal whereas 5 logs might be appropriate). While
this analysis had significant uncertainties concerning
national cyst occurrence and assumptions needed in the
risk assessment, the potential increased microbial risks
that could result from DBP regulation were significant.
2-2
-------
Table 11-1. Research to Support the LT2ESWTR
Expected Completion Date
Priority
A. Health Effects Research
1. Pathobiology of Infection and Disease for Most Important Waterborne Pathogens
HE.M.1 Infectious dose of Cryptosporidium
HE.M.2 Validity of dose-response model for Crypto in animals
HE.M.3 Cryptosporidium virulence study using different strains
HE.M.4 Cryptosporidium infectious dose in normal/immuno-compromised animals
HE.M.5 Infectious dose of Norwalk virus
HE.M.6 Infectious dose of other priority pathogens
AWWARF Related Projects
177 Association between Cryptosporidium in finished water & Cryptosporidiosis
in population
354 Cryptosporidium parvum: Surrogate human pathogenicity animal model using swine
1997
Proposed
1998
1999
2000
Proposed
1997
1999
H
H
H
H
H
M
HE.M.7 Characterization of endemic disease (with AWWARF) 2000
HE.M.8 Immunological assay for assessing exposure in epi studies (with AWWARF) 1998
HE.M.9 Investigations of waterborne disease outbreaks . Ongoing
HE.M.10 Surveillance tool for waterborne disease outbreaks Completed
AWWARF Related Projects
168 Prospective epi study of waterborne microbial disease (with EPA) 1997
268 Fingerprinting techniques for opportunistic pathogens & illness 1998
B. Exposure Research
1a. Microbial Methods—Protozoa
EX.M.1 Immunological techniques.for protozoa 1999
EX.M.2 Gene probes for detection of viable Cryptosporidium oocysts 1998
EX.M.3 Cultural method for Crypto in environmental samples Completed
EX.M.4 PCR methods for Giardia and Cryptosporidium (CRADA) 1997
EX.M.5 Protozoa methodology protocol development workshop Completed
EX.M.6 Comparison of methods for Giardia/Cryptosporidium in water 2000
EX.M.7 New protozoa agents 1998
AWWARF related projects
160 Vital stain for Giardia and Cryptosporidium Completed
162 UV-VIS spectroscopy for rapid on-line detection of protozoa 1998
253 Viability method for Giardia and Cryptosporidium 1998
259 Improve the IFA method for Giardia and Cryptosporidium 1998
283 Detection of Giardia & Cryptosporidium by flow cytometry Completed
351 Cryptosporidium parvum viability assay 1998
358 Cryptosporidium and Giardia antibody protocol 1998
364 New approaches for isolation of Cryptosporidium and Giardia 1998
366 Assessment of molecular epidemiology of waterborne Crypto with respect to origin 1998
395 Compare study of methods for assessing viability/infectivity of Crypto in U.S./U.K. 1999
1b. Microbial Methods—Viruses
EX.M.8 Application of PCR technologies and gene probes 1999
EX.M.9 Norwalk virus 2000
EX.M.10 Methods for emerging viruses Proposed
AWWARF projects
292 Rapid PCR-based monitoring of entero viruses 1998
345 Detection and occurrence of caliciviruses in drinking water 1999
Analysis of viruses by gene probe Completed
Viral and microbial methods for groundwater 1997
Microbial Exposure—Pathogen Occurrence
EX.M.11 Intensive eval. of micro, constituents/treatability in surface source
waters (mini-ICR with Research Council) 2000
EX.M.12 Identification of viruses resistant to disinfection Proposed
2b. Microbial Exposure—Watershed Control
EX.M.13 Distinguish animal versus human sources 1998
AWWARF related projects -
251 Evaluating of sources of pathogens and NOM in watersheds 1998
2c. Microbial Exposure—Opportunistic Pathogens in Distribution System
EX.M.14 Occurrence of Mycobacterium 1998
EX.M.15 Occurrence of heterotrophic bacteria with virulence characteristic 2001
H
H
H
H
H
H
H
M
H
H
H
M
H
H
612
726
2a.
H
M
M
H
H
(Continued)
2-3
-------
Table 11-1. (Continued)
Expected Completion Date
Priority
EX.M.16 PCR method for Legionella
EX.M.17 Pathogenicity of heterotrophic bacteria in drinking water.
EX.M.18 Occurrence of opportunistic pathogens in biofilms
EX.M.19 Opportunistic pathogens assoc. with (POU)/(POE) filter effluents
EX.M. 20 Potential pathogenicity of heterotrophic bacteria eluted from point-of -use
GAG filters
EX.M.21 Occurrence of newly emerging pathogens
EX.M.22 Exposure as a function of population distribution
C. Risk Assessment Research
1. Risks from pathogens
RA.M.1 Dev. comprehensive microbio risk assess paradigm for water
RA.M.2 Evaluation/application of various dose-response models
RA.M.3 Evaluation and application of methods to assess risk associated with
exposures to multiple pathogens, routes, and durations
AWWARF Related Projects
801 Microbial risk assessment for drinking water
D. Risk Management Research
1a. Pathogen Removal—Optimization of Conventional Treatments
RM.M.1 Filtration studies for controlling pathogens
RM.M.2 Filtration removal of protozoa and indicators
RM.M.3 Optimize conventional treatment for removal of oocysts
RM.M.4 Filtration damage viability studies
RM.M.5 Evaluate disinfection and optimization in full-scale 1 plants
RM.M.28 Evaluation of the effectiveness of conventional treatment processes on
removing and inactivating emerging pathogens
AWWARF Related Projects
155 Enhanced/optimized coagulation for removal of microbial contaminants
Treatment options for GiardialCrypto/ and other contaminants in recycled
backwash water
Application of pathogen surrogates measures to improve plant performance
Lime softening processes for Glardia and viruses
Evaluation of roughing filter design variables
Optimization of filtration for cyst removal
Balancing multiple water quality objectives
1b. Pathogen Removal—Effectiveness of Different Filtration Processes
RM.M.6 Biological treatment for control of oocysts
RM.M.7 Filtration techniques other than conventional treatment
RM.M.29 Evaluate filtration processes to remove emerging pathogens
AWWARF Related Projects
181 Biological particle surrogates for filtration performance eval.
Particle Counting as Indicator of Treatment Efficacy •
266 Quantitative particle count method development: standardization, sample stability
291 Particle image velocimetry
423 Joint project on treatment process selection: particle removal optimization
835 Practical guide to on-line particle counting
908 National assess, of particle removal by filtration
Membrane technology for pathogen removal
264 Integrated multi-objective membrane systems microbes/DBP precursors
817 Membrane filtration techniques for microbe removal
Microbial effects of biological filtration
263 Colonization of biologically active filter media with pathogens
917 Microbial effects of biological filtration (Don Reasoner on PAC)
1c. Pathogen Removal—Effectiveness of Disinfection for Inactivating Pathogens
RM.M.8 Control of Norwalk virus by chlorine and ozone
RM.M.9 UV disinfection efficiencies for Norwalk virus
RM.M.10 Inactivation of Giardia & Crypto by sequential disinfectants
RM.M.31 Evaluate disinfection processes to inactivate emerging pathogens
AWWARF Related Projects
Disinfection
262 Booster disinfection for pathogen and DBF control
352
363
608
636
703
834
1998
1998
Proposed
Proposed
Proposed
2000
1999
2000
1998
2000
Completed
1999
1998
1999
1999
1997
Proposed
1997
1998
1998
Completed
Completed
Completed
1997
2000
2000
Proposed
1997
1998
1998
Completed
Completed
Completed
1999
Completed
1998
Completed
Proposed
Proposed
1999
Proposed
1998
L
H
M
M
H
H
H
H
H
M
H
H
H
H
H
M
H
M
M
H
H
H
M
(Continued)
2-4
-------
Table 11-1. (Continued)
Expected Completion Date
Priority
525 Demo scale eval. of PEROXONE evaluation for disinfection, DBFs 1997
702 Devt. and validation of rational design methods of disinfection Completed
Disinfection of Protozoan Cysts—Inactivation Studies
151 Cyst and oocyst survival in watersheds and factors affecting inactivation 1997
273 Synergistic effects of multiple disinfectants (Gene Rice on PAC) 1998
282 Innovative electrotechnologies for Crypto inactivation 1998
375 Utility Giardia and Crypto inactivation study—Southern Nevada Water Authority 1998
731 Ozone disinfection of Giardia and Crypto Completed
906 Effect of various disinfection methods on inactivation of Crypto Completed
Disinfection Evaluation and Design
630 Full scale ozone contactor evaluation Completed
632 Modeling dissolved ozone in contactors Completed
1d. Pathogen Removal—Small Systems Technologies
RM.M.11 Crypiosporidium removal using bag filters 1998
RM.M.12 Cost effectiveness of prefiltration for ultrafiltration unit 1999
RM.M.13 Development/test innovative technologies for small systems 1999
2a. Distribution Systems—Indicators for Control of Pathogens in Biofilms and
Pipe Sediments
RM.M.14 Bacteria interference w/detection of conforms and E. coli Proposed
AWWARF Related Research
429 Male-specific coliphages as indicators of viruses & treatment effectiveness Completed
2b. Distribution Systems—Biofilm Growth and Control of Growth
RM.M.15 Kinetic Models for chlorine decay in distribution systems Completed
RM.M.16 Enhancement of EPANET distribution system model 1998
. RM.M.17 Prelim, studies of biofilm formation rates in pilot-scale distribution systems 1998
RM.M.18 Opportunistic pathogens in biofilms 2000
RM.M.19 Impact of nutrient removal on growth potential for bacteria 1999
RM.M.20 Impact of alternative treatment of biofilm growth Proposed
RM.M.21 Water quality factors in distribution systems 1997
AWWARF Related Projects
Control of Bacteria in Distribution System
270 Occurrence and control of Mycobacterium avium complex 1998
534 Fatty acid profiling for identification of environmental bacteria in distribution system Completed
936 Pathogens in model distribution system biofilms 1997
Minimization of Regrowth Potential
183 Factors affecting microbial growth in distribution systems 1997
704 Factors limiting microbial growth in distribution systems Completed
Disinfectant Residual Decay in Distribution System
261 Guidance manual for installation of booster disinfection 1998
293 Role of pipe-water interface in DBP formation & disinfection 1997
294 Travel times & water quality in dead ends 1997
815 Characterization and modeling of chlorine decay Completed
NWR1 Related Projects
Interactions between pipe materials, corrosion inhibitors, biofilm Completed
Risk reduction in distribution system by on-line monitoring of pathogen ecology Completed
2c. Distribution Systems—Effect of Design and Condition on Bacterial Growth
RM.M.22 Water quality impacts of dead ends 1998
RM.M.23 Mixing in storage facilities • 1998
RM.M.24 Alternative kinetic models for decay and DBP formation Proposed
RM.M.25 Real-time monitoring systems 1999
AWWARF related projects
154 Biological stability of drinking water in treatment plants and distribution systems 1997
254 Managing & operating finished water storage facilities 1998
260 Water quality monitoring of distribution system storage facilities 1998
729 Biofilm reactor BOG measurement Completed
2d. Distrib. Systems—Effect of Treatment on Chemical/Biological Stability of Water
R.M.M.26 Bacterial growth in distribution systems Proposed
R.M.M.27 Integrated approaches for controlling pathogens Proposed
H
H
H
M
H
H
M
M
M
M
M
M
M
H
M
H
M
(Continued)
2-5
-------
Table 11-1. (Continued)
Expected Completion Date
2e. Dlstrib. Systems— Maintaining Distribution System Integrity
RM.M.32 Evaluate primary causes of leaks/failures in distribution systems
RM.M.33 Evaluate devices for determining structural integrity of distribution system
RM.M.34 Evaluate materials for construction of distribution system
3a. Source Water
RM.M.35 Evaluate GIS for defining source water characteristics
RM.M.36 Evaluate water quality models for routing point and non-point source
water discharges
RM.M.37 Evaluate the Impact of sudden increases in source water contamination
on drinking water treatment
Proposed
Proposed
Proposed
Proposed
Proposed
Proposed
Priority
H
H
H
H
• H • .
H
In another RIA analysis, potential impacts for different
Stage 1 DBP rule options were evaluated assuming that
systems would simultaneously have to provide higher
levels of treatment for Giardia (versus the minimum of 3
logs required for all systems under the SWTR) as a
function of higher Giardia concentrations in the source
water. The goal of such a theoretical G/ard/a-based rule,
one of the options proposed in the IESWTR, was for
systems to provide adequate treatment to achieve a risk
level of less than one infection of Giardia per 10,000
people per year. It was assumed that if systems were
required to meet such a risk level, they would not incur
significant increases in microbial risk while making treat-
ment changes to comply with the Stage 1 DBP rule. The
RIA for this Giardia-based rule estimated that 400,000 to
500,000 infections of Giardia could be avoided per year
at a cost of about $400 million per year.
A major shortcoming with the above estimates, in addi-
tion to the uncertainties in the assumptions needed for
the analysis, was the inability to include impacts for
regulating Cryptosporidium since occurrence, treatment
effectiveness data, and dose-response information was
missing. Cryptosporidium may be the preferred target
organism for defining minimum levels of treatment be-
cause it is substantially more resistant to disinfection
than Giardia. A Cryptosporidium-based rule similar to
that described for Giardia would probably be signifi-
cantly more costly. Also, such a rule could trigger signifi-
cant shifts by the industry to greater use of alternative
disinfectants such as ozone (since they appear more
effective than chlorine for inactivating Cryptosporidium),
decreasing by-products from chlorine but increasing
other by-products.
LT2ESTWR
The majority of the research described in this research
plan will be used when designing the LT2ESTWR. The
regulatory needs for the LT2ESTWR include 1) a better
understanding of the magnitude of the risk from Crypto-
sporidium and other pathogens; 2) a more accurate
characterization of the occurrence of Cryptosporidium;
3) an evaluation of the potential for using indicators to
determine pathogen occurrence; and 4) the removal and
inactivation efficiencies of different treatments and the
effects on the distribution system.
EPA is considering at least three different approaches
(and combinations of these approaches) to further regu-
lating Giardia, Cryptosporidium, and viruses (and indi-
rectly regulating other pathogens) in the LT2ESWTR:
fixed treatment approach; proportional treatment ap-
proach; and watershed-based approach. Each of these
approaches has different, but also similar, information
requirements. The fixed treatment approach would re-
quire all systems to provide at least the same minimum
level of treatment (same log reduction of pathogens);
the proportional treatment approach would require sys-
tems to provide a minimum level of treatment at all times
based on pathogen levels in the source water; and the
watershed-based approach would require systems to
provide a minimum level of treatment at all times based
on a combination of factors that indicate the level of the
source water's vulnerability to pathogen contamination.
Hybrids of these regulatory approaches are also under
consideration, e.g., a fixed treatment approach for one
pathogen and a proportional level of treatment for an-
other pathogen. Similarly, different regulatory approaches
might be considered for different water body types (e.g.,
one type of requirement might be appropriate for quies-
cent waters while another type might be appropriate for
running waters).
The advantage of the proportional and the watershed-
based treatment requirements is that the drinking water
treatment required is based on the individual system's
source water quality. The treatment would prevent expo-
sure to unacceptable microbial risk and would avoid
unnecessary disinfection and associated DBPs. By con-
trast, the fixed treatment requirement—the one size fits
all—could require some systems with high quality source
water to provide unnecessary treatment at a high cost to
the consumers, while others with highly contaminated
source waters might provide inadequate treatment.
In its evaluation of a regulatory option under either of
these three structural approaches, EPA plans to use
nationwide system-specific data on source water quality
and water treatment, as well as research results, to
estimate the changes in treatment that would be neces-
sary to meet target microbial risk levels, or conversely to
estimate the risks associated with given treatment re-
quirements. In simple terms, the process of relating
treatment to microbial risk can be divided into three
2-6
-------
steps: step 1) establishing the level of contamination in
the source water (the occurrence of the pathogen),
either using information on pathogen concentrations or
based on pathogen "indicators"; step 2) estimating the
reduction in pathogen occurrence due to treatment of
known efficacy; and step 3) applying an exposure and
dose-response model to the resulting pathogen occur-
rence to estimate the risk of infection. An additional step
in this regulatory process will be to determine the level of
DBFs associated with the different treatment require-
ments. While all of these steps will involve differing
degrees of uncertainties, this analytical approach should
provide a characterization of relative levels of exposure
to microbes (and DBFs) based on source water quality
and treatment provided. EPA believes that results of
ongoing research studies will contribute to reducing
some of the uncertainties in classifying source water
pathogen risk, quantifying treatment effectiveness, and
measuring health effects due to pathogen exposure.
Table 11-1 summarizes research projects completed,
ongoing by EPA, or not yet funded that would support
the development of the LT2ESWTR. These projects are
described more fully in Chapters III and IV. The develop-
ment of cost/benefits as part of the RIA requires infor-
mation on the risk estimates of predicted pathogen
levels in treated waters, occurrence of pathogens in
source waters, and removal and inactivation efficiencies
for different technologies.
To develop a better understanding of the magnitude of
the risk from Cryptosporidium and other pathogens, it
will be necessary to develop a better understanding of
the dose-response to Cryptosporidium and different.
pathogens and how the risk varies according to strain
and immune system response. There are several stud-
ies which are evaluating the dose-response to Crypto-
sporidium (HE.M.1-4 and by the American Water Works
Association Research Foundation (AWWARF) and CDC).
These studies should provide a good understanding of
the range of responses to Cryptosporidium although the
probability of being exposed to one strain versus an-
other is likely to remain uncertain. Several projects are
examining the high and low dose-response, immune
system effects and clinical symptoms associated with
exposure to the caliciviruses, Norwalk virus, and Snow
Mountain Agent (HE.M.5). In addition to dose-response
studies, epidemiology studies may be useful for verify-
ing whether the risk estimates using traditional risk
models are accurate and can provide information to help
evaluate appropriate levels of protection (HE.M 7-10).
They may also identify the factors relevant to reducing
the disease burden from drinking water. AWWARF and
CDC are also sponsoring research in these areas. The
research described above should be completed in time
for the LT2ESTWR.
In order to more accurately characterize the occurrence
of Cryptosporidium and Giardia in raw and finished
water, a more precise, accurate, facile, and low cost
method is needed for the LT2ESTWR. EPA is conduct-
ing a large amount of research to develop a better
method for protozoa (EX.M.1-6). AWWARF and CDC
are also conducting substantial research to develop a
better method for detecting protozoa. EPA believes the
prospects are good for a method to be available in time
for the proposed LT2ESWTR, that can adequately limit
misclassification rates of source water Cryptosporidium
concentration. There is also research into developing
better methods for caliciviruses such as Norwalk (EX.M.8-
9), and studies to improve methods are planned for
adenoviruses, coxsackiviruses, and hepatitis A (EX.M.8).
The ICR will provide 18 months of data on the occur-
rence of Giardia, Cryptosporidium, and viruses from
close to 350 different water treatment plants. Other data
to be collected include data on turbidity, coliforms, water
resource type (river, reservoir/lake) and whether water-
shed control is practiced. EPA will also conduct supple-
mental surveys which will complement the ICR data to
provide source water microbial occurrence data neces-
sary to support the regulatory impact analysis for differ-
ent LT2ESWTR options. The "mini-ICR" will also collect
pathogen occurrence data (EX.D.11) and will provide a
range of data to assess pathogen occurrence variability
and develop a better understanding of differences be-
tween worst case pathogen levels and other statistical
endpoints (e.g., 90th percentile, median). AWWARF
and the USDA are conducting several studies on the
survival of Cryptosporidium in the environment.
Another important aspect of estimating Cryptosporidium
occurrence in source water is its variability and the co-
occurrence of Cryptosporidium with other microorgan-
isms that could possibly serve as indicators. Indicators
could be especially important for smaller systems where
an approach based on indicators of pathogen occur-
rence would be ideal because of the cost and the
complications associated with pathogen monitoring. To
determine which indicators, or combination of indicators,
correlate best with pathogen occurrence, EPA will ana-
lyze the results of the 18 monthly source water samples
collected by systems under the ICR and the results of
the related supplemental surveys. In addition, the "mini-
ICR" will examine the use of indicators (EX.D.11). The
information from the "mini-ICR" should facilitate devel-
opment of source water pathogen categories based on
indicators and watershed characteristics (e.g., the range
of pathogen contamination associated with certain fecal
coliform distributions in the raw water of a system served
by a river source).
From a regulatory standpoint it is important to be able to
determine the level of treatment that is achieved by a
particular treatment process under different water qual-
ity and operating conditions (i.e., performance indicators
are needed to establish pathogen removal and inactiva-
tion). Typically, performance indicators used in the con-
text of microbial treatment technique requirements are
filter effluent turbidity for physical removal'and CT val-
ues (concentration of disinfectant in mg/L multiplied by
the contact time in minutes) for pathogen inactivation.
These and other indicators (and combinations thereof)
need to be evaluated to determine how they can be
2-7
-------
used to demonstrate that the required level of treatment
is being achieved. EPA expects that ongoing research
on treatment optimization to remove pathogens and on
indicators of treatment effectiveness will contribute infor-
mation to establish the level of physical pathogen re-
moval and to develop regulatory performance criteria
and associated monitoring requirements (RM.M.1-7).
Regarding regulatory inactivation requirements of patho-
gens, research on the disinfection oiCryptosporidium\s
a high priority for LT2ESWTR (RM.M 8-10). ICR data
may also be used to validate gross assumptions on
treatment effectiveness if it is determined that this is
possible once the ICR data is received and reviewed. In
addition to EPA's research, AWWARF is conducting a
considerable amount of research in this area (see re-
lated projects). As better design and operating param-
eters become available, they will be used in an existing
water cost model to estimate costs for systems to achieve
the remoyal/inactivation requirements for different regu-
latory options. Data from these projects would also be
used for developing guidelines by which utilities could
estimate removal/inactivation efficiencies. The major work
in this area should be completed in time for the
LT2ESTWR.
There are several projects that will be completed in time
for the LT2ESTWR that would help determine the health
risk significance from bacterial growth in the distribution
system and remedial treatment or operational strategies
(EX.M.14-22 and RM.M.14-27). Results from these ef-
forts would determine what amendments to the SWTR
and IESWTR or changes to related guidance might be
needed to reduce risks from contamination from the
distribution system.
D/DBP rules
During the DBP regulatory negotiations participants were
concerned with the uncertainty in characterizing risks
associated with DBPs from chlorine and alternative dis-
infectants. RIA modeling indicated that the national
baseline incidence of cancer attributed to chlorinated
DBPs in drinking water could range from 1 case per year
(based on central tendency risk factors from animal
toxicology data for TTHMs alone) to over 10,000 cases
per year (based on risk factors suggested by a meta-
analysis of epidemiology studies by Morris et al). The
RIA indicated that under the Stage 1 DBP rule the
national cost for avoiding one case of cancer could
range from several hundred thousand dollars to several
billion dollars per year. Despite these enormous uncer-
tainties, the Negotiating Committee recognized that the
existing risks could be large and therefore should be
reduced. They reached a consensus that the Stage 1
requirements were of sufficient benefit to be proposed
for all system sizes, even though the costs for such a
rule were substantial.
While no Stage 2 DBP rule could be agreed to until the
benefits of such a rule became more apparent, a pos-
sible criterion was proposed, and an RIA was conducted
to understand potential cost impacts. It was estimated
that if all systems in the U.S. were required to achieve a
Stage 2 DBP rule of 40 ng/l for TTHMs and 30 ng/l for
HAAS, without restricted use of alternative disinfectants
to chlorine (which are less expensive than precursor
removal technologies), they would incur annual costs of
$1.5 billion per year above those anticipated for Stage 1
($1 billion per year). If systems were required to use
precursor removal technologies such as GAG and mem-
brane technology to meet a 40/30 ng/l Stage 2 rule, they
were projected to incur annual costs of several more
billions per year than if they could meet such a standard
using alternative disinfectants to chlorine. Table II-2
summarizes ongoing, proposed, and completed research
projects by EPA and other organizations that will provide
valuable inputs to the development of the Stage 1 and
Stage 2 DBP rules.
Stage 1 DBP Rule
As discussed previously, EPA must promulgate the
Stage 1 DBP rule by November 1998. EPA does not
envision making major changes to the 1994 proposal
because it was developed through a negotiation process
and most of the public comments support the proposal.
Research from this Research Plan will be used mainly to
support the development of the Stage 2 DBP rule.
However, there are several research projects that may
have an impact on the Stage 1 DBP rule.
For health effects, a new cancer study for bromate will
be available for the final Stage 1 rule (HE.D.7). This may
be important because the theoretical 1 x 104 cancer risk
level for bromate is 0.005 mg/L while the MCL in the
proposal was 0.010 mg/L. If the cancer risk from bro-
mate is greater than in the proposal; then there may be
concern about the use of ozone as a disinfectant. The
Chemical Manufacturers Association has completed a
two-generation study on the reproductive and develop-
mental effects of chlorite (HE.D.8). The information from
this study will be used to modify, if justified, the MCLG
for chlorite and chlorine dioxide. The International Life
Science Institute (ILSI) convened an expert panel in
1996 to explore the application of the EPA's 1996 Pro-
posed Guidelines for Carcinogen Risk Assessments
the available data on the potential carcinogenicity of
chloroform and dichloroacetic acid (RA.D.1). EPA is
evaluating the' ILSI report and its implications for the
MCLGs for chloroform and DCA. Finally, EPA is re-
evaluating the available cancer epidemiology data using
different meta-analytical techniques to determine the
adequacy and accuracy of the previous meta-analysis
by Morris et al. (RA.D.5-6). The results from this re-
search may be used in developing better estimates of
the potential benefits from the Stage 1 DBP rule.
Additional data on the occurrence of TTHMs and HAAS
has been collected by the American Water Works Asso-
ciation, the American Water Works Systems Company
and several states and has been submitted to EPA for
consideration. In addition, new occurrence information
for chlorite and chlorate have been provided to EPA.
2-8
-------
Table 11-2. Research to Support the DBF Rules
Expected Completion Date
Priority
A. Health Effects Research
1. Epidemiology—Development/Application of Improved Tools for Field Research
HE.D.1 Improving estimates of residential DBF exposure in epi studies
HE.D.2 Improving measures of biologic effect: evaluation of biomarkers
HE.D.3 Improving methods for managing health and exposure data
2. Epidemiology—Feasibility/Full-Scale Studies
HE.D.4 Feasibility studies: Cancer (Research Council)
HE.D.5 Feasibility studies: Reproductive effect (Research Council)
HE.D.6 Full-scale studies; Cancer and reproductive effects
1997
2000
1998
-Cancer
-Reproductive effects
3. Toxicology—Hazard Identification and Dose-Response
HE.D.7 Cancer dose-response studies
HE.D.8 Reproductive/developmental effects screening studies
HE.D.9 Neurotoxicity studies
HE.D.10 Immunotoxicity studies
AWWARF Related Projects
738 Dose-response relationship of DCA & TCA-induced proliferation
CMA: Chlorite/Chlorine Dioxide 2 generation reproductive study
4. Toxicology—Pharmacokinetic and Mechanisms of Action
HE.D.11 Pharmacokinetic and mechanistic research—cancer
HE.D.12 Pharmacokinetic and mechanistic research—reproductive
AWWARF Related Projects
432 Mechanistic basis and relevance of rat kidney tumor formation
617 Carcinogenic mechanisms in rat & mouse hepatocytes
701 Induced hepatic tumors with induction of peroxisomes
5. Toxicology—DBP Mixtures
HE.D.13 Mixtures feasibility study
HE.D.14 Toxicologic evaluation of drinking water mixtures
HE.D.15 Studies of DBP interactions
HE.D.16 Mutagenicity screening studies of drinking water mixtures
B. Exposure Research
1a. DBP Methods—Stage 1 DBP Rule
EX.D.1 Low level bromate measurement
EX.D.2 Improved method for haloacetic acids
EX.D.3 Expand quality control for TOG, evaluate new TOC methods
EX.D.4 Low level CIO2 measurement (depends on health data)
EX.D.5 Real-time monitoring for disinfectant residuals
EX.D.6 PE studies for DBFs and disinfectants
AWWARF Related Projects
159 Improved methods for isolation and characterization of NOM
. 163 Development of improved method for haloacetic acids
417 Development of fiber optic chemical sensors for monitoring organic
compounds (TOC)
830 Bromide-ozone interactions (incl. develop analytical technique for total
organic bromide)
1b. DBP Methods—Stage 2 DBP Rule and longer term
EX.D.7 Methods for peroxides
EX.D.8 Real-time, in-plant monitoring of DBFs
EX.D.9 Improved method for aldehydes
2a. DBP Exposure—DBPs from Different Disinfectant Combinations
EX.D.10 Identify new DBPs from alternate disinfectants
EX.D.11 Methods for nonvolatile DBPs
AWWARF Related Projects
825 Bromide Survey
2b. DBP Exposure—Factors that Affect Exposure Levels and Human Exposures
EX.D.12 DBP changes in distribution system
EX.D.13 DBP interactions w/foods and associations w/dietary intake
1998
1998
Proposed
Proposed
2000
2000
2000
2000
Completed
1997
2000
2000
Completed
1997
1997
2000
Proposed
2000
2001
1999
Completed
1998
Proposed
Proposed
Ongoing
1997
1997
Completed
Completed
2000
1999
1998
2000
2000
Completed
Proposed
2000
H
H
H
H
H
if feasible H
if feasible H
H
H
M
M
H
H
H
if feasible H
M
H
H
H
H
M
L
H
M
M
M
H
H
M
M
(Continued)
2-9
-------
Table 11-2. (Continued)
Expected Completion Date
Priority
EX.D.14 Exposure to DBFs through showering
EX.D.1 5 Markers of DBF exposure
EX.D.1 6 Models of DBF exposure
EX.D.1 7 Exposure as function of population distribution
EX.D.1 8 Tapwater consumption
Proposed
Proposed
2000
Proposed
1997
.. M
M
H
M
H
C. Risk Assessment Research
1. Characterizing Risk of Individual DBPs
RA.D.1 Cancer risk assessments
RA.D.2 Cancer combination study for bromates
RA.D.3 Noncancer risk assessments
RA.D.4 Risk characterization
2. Characterizing Risks from Chlorinated Waters
RA.D.5 Evaluate newer epidemiologic studies
RA.D.6 Assessment of previously conducted studies
RA.D.7 Identify ongoing cancer studies
3. Methods and Models to Characterize Risks from Mixtures
RA.D.8 Characterization of interactions for mixtures of DBPs.
RA.D.9 Threshold studies for D/DBPs
RA.D.10 Use of QSAR model to estimate risk for single cmpnds/classes of
compounds within a mixture
4. Methods and Models to Compare Risks
R.A.D.11 Comparative Risk Analysis
D. Risk Management
1. Effectiveness of Treatment Processes in Reducing DBP Precursors
RM.D.1 Enhanced softening for precursor and pathogen removal
RM.D.2 Effects of ozone & biofiltration for control of precursor and pathogens
a Control precursor, pathogen and pesticide removal
b Effect of pH on ozonation and enhanced coagulation
RM.D.3 Analyze 1CR data from GAG, membrane bench and pilot studies
RM.D.4 Removal of DBP precursors by GAC and membranes
RM.D.5 Membrane scale-up and fouling
AWWARF Related Projects
DBP Precursor Control—GAC
816 Removal of DBP precursors by GAC adsorption
DBP Precursor Control—Coagulation
531 Humic acid removal using ferric chloride
814 Removal of DBP precursors by optimal coag. & and precip. softening
934 Optimizing ozonation for turbidity & organics removal
DBP Precursor Control—Oxidation and Biological Filtration
252 Optimizing filtration in biological filters
289 Advanced oxidation and biodegradation processes
631 Removal of natural organic matter in biofilters
712 Design of biological processes for organics control
DBP Precursor Control—Membrane Technology Issues
170 Reverse osmosis and nanofiltration for organics removal
264 Integrated membrane systems to control microbes/ DBP precursors
601 Ultrafiltration membrane pretreatment and nanofiltration
826 Membrane technology for drinking water—joint report
904 Blofouling In membrane processes
Other Issues
271 Improving clean/veil design for DBP and CT compliance
361 Case studies of impacts of treatment changes on biostability on full-scale distrib.
systems
369 Case studies of modifications of treatment practices to meet the new D/DBP
regulation
2. Effectiveness of Alternative Disinfectants in Limiting DBP Formation
RM.D.6 Ozone by-product formation and control
AWWARF Related Projects
Formation and Control of Ozonation By-Products
156 Strategies to control bromide and bromate ion
2000
1997
2000
2000
1998
1998
1998
1998
1998
1997
2000
2000
1998
2000
1999
Completed
1998
1997
Completed
1997
Completed
-1998
1998
Completed
1997
1997
1999
Completed
1997
1997
1998
1999
1999
1999
H
H
H
H
H
H
H
M
H
H
H
H
H
H
M
(Continued)
2-10
-------
Table 11-2.
(Continued)
Expected Completion Date
Priority
504 Ozone & biological treatment for DBP control and biological stability
525 Evaluation of PEROXONE advanced oxidation process
533 Effect of carbonate/bicarbonate alkalinity on advanced oxidation processes
709 Impacts of ozonation on formation of chlorinated DBPs.
830 Bromide-ozone interaction in water treatment
832 Reaction of ozone and hydroxyl radicals with amino acids
Completed
1997
Completed
1997
Completed
Completed
Formation and Control of Chloramine DBPs
710 Nitrification occurrence and control in chloraminated water systems Completed
803 Factors affecting DBP formation during chloramination 1997
937 Chloramine decomposition kinetics and degradation products. 1997
Formation and Control of Chlorine Dioxide-related DBPs
611 Sources, occurrence, & control of CIO2 by-product residuals in drinking water Completed
833 Minimizing chlorate ion formation in drinking water when hypochlorite used Completed
3. Small Systems Technology for Precursor and DBP Control
RM.D.7 Membranes/advanced oxidation/other technology combinations 1999
This new data will be useful when evaluating the poten-
tial regulatory impacts of the Stage 1 DBP rule.
Improvements in analytical methods will be completed in
time for the Stage 1 DBP rule for bromate (EX.D.1),
haloacetic acids (EX.D.2), and TOG (EX.D.3). Research
related to improving enhanced coagulation and soften-
ing will also be completed in time for the final Stage 1
rule (RM.D.4). Many of these projects will help to better
define the precursor removal requirements in the Stage
1 DBP rule. In addition, there are several AWWARF
projects that will help to improve the analytical methods
and improve DBP precursor removal (see Table H-2).
Stage 2 DBP Rule
The Stage 2 DBP rule is required to be completed by
May 2002. In order to meet this deadline, EPA will need
to initiate a second regulatory negotiation by late 1999 to
complete a proposed rule by late 2000. EPA believes
that the majority of research will be completed in time for
consideration in the final Stage 2 DBP rule. When
developing the Stage 2 DBP rule, there are several
critical issues that need to be addressed in order to
support the development of different regulatory options
including 1) the magnitude of the cancer and noncancer
risks (e.g., reproductive risks) from chlorinated waters;
2) the magnitude and relative cancer and noncancer
risks (e.g., reproductive risks) from the DBPs formed
when using alternative disinfectants (e.g., ozone and
chlorine dioxide); 3) the relative risks from brominated
species versus the chlorinated species; 4) better meth-
ods for DBPs and the collection of additional occurrence
data; 5) the evaluation of the effectiveness of GAG and
membranes to remove DBP precursors; and 6) the
balancing of the risks between pathogens and DBPs.
The magnitude of the health risks from chlorinated wa-
ters and from those DBPs formed using different disin-
fectants is important to understand when trying to deter-
mine "how large is the risk" or "should we be con-
cerned." Understanding the magnitude of the health
risks from the use of different disinfectants is critical
when evaluating the different regulatory options for the
Stage 2 DBP rule and for evaluating the risk-risk tradeoffs
between controlling for microbial pathogens and DBPs.
For example, using alternative disinfectants may be
much more cost effective than using advanced tech-
nologies such as membranes or GAG to comply with
MCLs for chlorinated DBPs. On the other hand, alterna-
tive disinfectants to chlorine create their own DBPs of
concern that need to be considered (e.g., bromate with
the use of ozone and chlorite with the use of chlorine
dioxide). Since ozone and chlorine dioxide are more
effective for inactivating Cryptosporidiumthan chlorine it
becomes especially important to understand the health
risks of DBPs formed using different disinfectants.
Determining the relative risk from alternative disinfec-
tants is important when trying to address "do the by-
products from the use of different disinfectants
present more or less of a health risk than the by-
products from other disinfectants." For example, sys-
tems may be considering switching to ozone for better
control of Cryptosporidium and because it forms fewer
chlorinated DBPs. The apparent major by-product of
ozonation in the presence of bromide is bromate. If
bromate presents a greater health risk than chlorinated
by-products then the switch to ozone may actually in-
crease the risk from DBPs. However, before these is-
sues can be resolved it is critical that the risks from the
by-products formed from alternative disinfectants be
evaluated.
Determining the relative risk from brominated species is
important when trying to address "do brominated spe-
cies pose a greater risk than chlorinated species."
Determining the relative risks between the brominated
and chlorinated species is important because there is
some evidence that in waters with high bromide concen-
trations, technologies to remove DBP precursors could
increase the concentrations of certain brominated DBPs
even though the group concentrations of TTHMs and
HAAS may decrease. This may be of concern if the risks
2-11
-------
from the brominated species are shown to present a
greater health concern than the chlorinated species.
To determine the magnitude and relative risks from the
DBFs formed using different disinfectants, information is
needed on the health effects from individual DBFs or
mixtures of DBFs along with occurrence/exposure infor-
mation. In addition, information is needed on whether
the risks can most cost effectively be controlled by
limiting exposure to individual DBFs and/or surrogate
parameters (e.g., TTHMs, total haloacetic acids, total
organic halides [TOX], or total organic carbons [TOG]). If
the magnitude of national cancer risk attributed to chlori-
nated DBFs after the Stage 1 rule is determined to be
high (e.g., greater than 5,000 cases per year), the costs
for implementing a Stage 2 DBF rule, where many
systems would be required to use GAG or membrane
technology, may be justified. However, if equivalent
protection could be provided allowing use of alternative
disinfectants, then costs may be much lower.
During the negotiations, concern was also raised as to
the significance of reproductive and developmental risks
from disinfected waters. This issue is of particular impor-
tance from a regulatory perspective. If reproductive or
developmental effects are of concern, then standards
might be set to prevent the threshold risk level from
occurring anywhere in the distribution system because a
single excursion above the MCL might induce illness
whereas cancer risk is considered to accrue over time.
This approach to setting the MCL is unlike the existing
TTHM maximum contaminant level (MCL) or newly pro-
posed Stage 1 MCLs for TTHMs and HAAS (i.e., the
sum of concentrations for five haloacetic acids) where
compliance is based on an annual average of concen-
trations measured in the distribution system.
Data from epidemiological and toxicology studies will be
used in a weight of evidence approach to better define
the magnitude and relative of the risks for chlorinated
DBFs (HE.D.1-16 and RA.D.1-11). The epidemiology
research will provide direct human information that will
reduce uncertainties in the cancer and reproductive
risks from chlorinated waters, but uncertainties will re-
main for the Stage 2 DBF rule (HE.D.1-6 and RA.D.5-7).
The majority of cancer epidemiology research will be
completed in time for the Stage 2 rule, but the results
from newly initiated cancer epidemiology studies would
probably not be available in time for consideration for
the Stage 2 DBF rule because it takes several years to
design and carry out such large studies. The toxicology
research will provide substantial new information for
evaluating the cancer and reproductive risks from indi-
vidual chlorinated DBFs and for evaluating the potential
for neurotoxicity and immunotoxicity (HE.D.7-12 and
RA.D.1-4). Although the majority of this research should
be completed in time to be considered for the DBF
Stage 2 rule, it is important to note that the information
from many of the 2-year cancer bioassays described in
Chapter IV will not be available until mid to late 2000.
EPA is hopeful that preliminary results will be available
that could be used in the regulatory negotiations to
provide some direction on the potential health risks from
these DBFs. Mixtures research would also provide new
information to reduce the uncertainties in the risk esti-
mates for chlorinated waters, but this research is com-
plex and many of the studies, if initiated, may not be
available in time for the Stage 2 DBF rule (HE.D.13-16
and RA.D.8-10).
Data from toxicology studies, and to a more limited
extent epidemiology studies, will be used to better de-
fine the magnitude and relative risks for DBFs formed as
a consequence of using alternative disinfectants (HE.D.1-
16 and RA.D.1-11). The toxicology research will provide
information on the major DBFs formed by the use of
chlorine dioxide (chlorite and chlorate), ozone (bromate,
glyoxal, formaldehyde), and possible studies on the
toxicity of cyanogen chloride (a major by-product from
the use of chloramines) (HE.D.7-12). As with the several
of the chlorinated DBFs, many of the 2-year cancer
bioassays described in Chapter IV will not be available
until late 2000. There will be limited information from
mixtures research on the potential risks of DBFs from
using alternative disinfectants (HE.D.13-16 and RA.D.8-
10). Epidemiology studies may provide some informa-
tion on the potential reproductive risks, but it will be
difficult to conduct cancer epidemiology studies for sys-
tems using chlorine dioxide and ozone in the U.S.
because of the limited historical use and small potential
populations exposed.
The EPA/NTP collaborative effort described in HE.D.7-8
will conduct cancer studies, reproductive screening stud-
ies, and mechanistic studies for several brominated
DBFs (e.g., bromodichloromethane, dibromoacetic acid,
dibromoacetonitrile). The reproductive and mechanistic
studies should be completed in time for the Stage 2 DBF
rule, but results from the cancer studies may not be
available until mid to late 2000.
For the Stage 2 DBF rule, there is a need to collect
additional DBF occurrence information and DBF precur-
sor information to more accurately estimate the potential
risks from DBFs. It will be especially important to collect
information on DBFs formed from the use of alternative
disinfectants such as ozone and chlorine dioxide. The
ICR will be collecting information on four disinfectants,
29 DBFs, five general water quality parameters (pH,
alkalinity, calcium hardness, total hardness, and ammo-
nia), and a DBF precursor (bromide) and two surrogates
for DBF precursors (TOC, UV-254). Of the 29 DBFs,
there will be four THMs, six HAAs (3 others are op-
tional), four acetonitriles, two haloketones, seven alde-
hydes, chlorate, chlorite, bromate, chloropicrin, chloral
hydrate, and cyanogen chloride. The first six months of
ICR data should be available by early 1999. In addition
to collecting occurrence data for these DBFs, it is also
critical to identify new DBFs (EX.D.10) and to develop
methods for nonvolatile DBFs (EX.D.11) to ensure the
major DBFs have been identified. EPA believes the
occurrence data from the ICR will provide the needed
information for the critical DBFs for the Stage 2 DBF
rule.
2-12
-------
Another important area for trie Stage 2 DBP rule is the
development or improvement of analytical methods for
DBFs of greatest concern. As discussed above, there
are projects that are attempting to improve the methods
for bromate, HAAs, and TOG. In anticipation of the
Stage 2 rule, there are research projects to improve the
methods for peroxides (EX.D.7) and aldehydes (EX.D.9).
Methods for other DBFs that may be considered for the
Stage 2 rule were included as part of the ICR and are
believed to be adequate for use in the Stage 2 rule (e.g.,
haloacetonitriles and haloketones). Better methods may
need to be developed for MX and cyanogen chloride if
they are included in the Stage 2 rule. Methods for use as
precursor surrogates (TOG, SUVA, and NOM) or DBP
surrogates (TOX and TOBr) need to be developed in
case these prove to be good indicators based on infor-
mation from the ICR. There is research on improving the
TOG method (EX.D.3) and to better characterize NOM
(AWWARF projects). The method for TOX is considered
adequate and there is a project to improve the total
organic bromide method (AWWARF).
Another critical area for the Stage 2 rule is to determine
the most cost-effective precursor removal technologies
and disinfectant application strategies that can be used
to reduce the formation of DBFs. Predicting the removal
of DBP precursors and DBP formation will be based
mostly on data collected from the ICR, but also from
treatment research and studies of processes in the
distribution system that affect DBP formation (RM.D.1-7;
EX.D.16; RM.M.15, 16, 24,-25). In addition to EPA's
research, AWWARF is conducting a considerable amount
of research in this area (see related projects). As better
design and operating parameters become available for
defining precursor removal and DBP formation, they will
be used for estimating costs for systems to achieve
water quality objectives for different regulatory options.
EPA believes the majority of the treatment related re-
search will be completed in time for the Stage 2 DBP
rule.
All the information cited above will be used to assist EPA
in comparing and balancing the risk from DBFs and
microbial pathogens. Current comparative risk models
for drinking water weigh the outcomes of microbial expo-
sures to that of cancer from selected DBFs. Research
needs to be conducted on a variety of models and
methods that will allow for the development of a com-
parative risk assessment model which addresses mul-
tiple outcomes (e.g., cancer, developmental, reproduc-
tive, neurotoxic effects), their impacts and costs. This
research should focus on the risk analysis and risk
reduction benefits derived from minimizing exposures
that would result in adverse outcomes other than can-
cer. A comparative risk framework and strategic model
has been developed for DBFs and pathogens and is
currently being validated and reviewed (RA.D.11). A
related project pertinent to pathogen risk assessment
(RA.M.2) will provide predictive functions concerning the
magnitude of response, severity of effects, and duration
of exposure. The two projects are intended to provide a
prediction of different health risk endpoints resulting
from simultaneous exposure from DBFs and pathogens.
EPA intends to use these models as part of its RIA for
evaluating different LT2ESWTR/Stage 2 DBP combined
regulatory options.
Groundwater Disinfection Rule
The 1996 SDWA amendments require EPA to promul-
gate drinking water standards for groundwater no sooner
than August 1999 and no later than May 2002. EPA
plans to propose a groundwater rule in January 1999
and finalize the rule in November 2000 in conjunction
with the LT1ESWTR.
Under the negotiated rule-making for DBFs, it was as-
sumed that the GWDR would prevent increases in mi-
crobial risk while groundwater systems using disinfec-
tion complied with the Stage 1 DBP rule. Ground water
systems not required to disinfect would avoid risks from
DBFs and not be required to comply with DBP regula-
tions. Also, groundwater systems required to disinfect
but serving only transient populations (most non-com-
munity systems) would not be subject to DBP standards
unless risks from short-term exposure were determined
to be of concern.
While some general problems are common to all drink-
ing water systems and can be adequately addressed by
research developed for the surface water and DBP
rules, a distinct set of conditions and assumptions apply
to contamination and treatment of groundwaters that
must be considered. Key regulatory issues include 1)
whether practical criteria can be determined for when
systems can avoid disinfection while still providing a
safe water; 2) what level(s) of disinfection to require for
systems which are vulnerable to fecal contamination in
the source water; and 3) what control measures to
require to limit contamination to the distribution system.
Approaches being considered for addressing the above
issues include 1) basing regulatory criteria on their
ability to ensure that a system is below a desired risk
level (e.g., less than a 104 annual rate of infection from
most viruses), taking cost into consideration; 2) requir-
ing disinfection to reduce risks from viruses to the extent
that is technically and economically feasible for most
systems, while allowing systems that are clearly not
vulnerable to fecal contamination to avoid disinfection
(e.g., based on occurrence of indicators such as conforms
and/or coliphage, and minimum set-back distances based
on hydro geological features); and 3) same as (2) but
only requiring systems to disinfect that are clearly vul-
nerable to fecal contamination (e.g., based on occur-
rence of conforms and/or coliphage).
Given the potential occurrence of over 100 enteric vi-
ruses in drinking water, the risk-based approach can
only practically be pursued by focusing on those viruses
which occur most frequently and at the highest levels in
drinking waters, are most resistant to ambient stress
and treatment, and are clinically most significant. Ide-
ally, EPA would predict a minimum level of treatment or
2-13
-------
set-back distance (at a particular site) for virus "A" to
ensure that there are less than a fixed rate of infection
(e.g., 1 infection per 10,000 people per year) from the
viruses "A," "B," "C," and "D," etc. The problems of
developing such an approach are numerous including
viruses that behave differently in their die-off and trans-
port through the ground; disinfectants with varying de-
grees of effectiveness for inactivating different viruses;
and viruses that occur at different concentrations with
variability in source waters and upon infection produce a
wide range of different clinical manifestations, ranging
from no symptomatic response or mild gastroenteritis to
hepatitis and possibly death.
Model development of this approach is further ham-
pered by the wide range of uncertainties for aquifer
hydrogeological characteristics and fate and transport
properties of the various pathogens and indicator organ-
isms. These difficulties have led to a two-pronged ap-
proach for research supporting a GWDR. One prong is
to consider the uncertainties and unknowns for the
various elements in these models to determine 1) if they
can be reduced such that meaningful predictions are
fundamentally possible, and 2) if the model predictions
can be validated in the field. The other prong is to focus
on other approaches to estimating groundwater vulner-
ability, primarily considering occurrence data and tradi-
tional well site selection criteria, such as set-back dis-
tances, confining layers, land uses, well depth, etc.
These need to be likewise considered for their ability to
meaningfully predict vulnerability/non-vulnerability.
The GWDR will affect a fundamentally different class of
public water supply systems than the ESWTR. The
10,000 or so surface water systems include a substan-
tial fraction of large community systems with substantial
infrastructure and economic resources. The 160,000
groundwater systems are almost exclusively very small
non-community and community systems with limited
resources and infrastructure. The regulatory realities
require that EPA consider simple, feasible assessment,
treatment and monitoring approaches.
EPA is considering developing the RIA using three
regulatory approaches described above. However, be-
fore these approaches can be used, additional informa-
tion is needed as described below.
Table II-3 summarizes completed, ongoing, and pro-
posed research to help define the significance of the
public health problem in groundwater systems. These
projects include information on known and estimated
disease (outbreaks and endemic disease rates), micro-
bial pathogen and indicator occurrence in source waters
and distribution systems, and exposure risks associated
with various contamination sources (HE.M.5, 6, 9, and
10; EX.M.8, 10 and 14 -18; RA.M.1-3). This information
will indicate the nature and scope of the public health
problem and allow for estimates of the baseline risk level
from which to base the RIA and develop appropriate
regulatory goals. Surveys on disinfection practice by
industry will help characterize types and levels of disin-
fection treatment currently practiced by groundwater
systems.
Strategies to determine system vulnerability to microbial
contamination and criteria to avoid disinfection have to
be developed and field tested. Information on factors
affecting and limiting microbial contamination of ground-
water will help establish criteria for avoiding disinfection.
Such information includes physical and chemical prop-
erties governing fate and transport of microbes in the
subsurface, site-specific factors such as land use pat-
terns, and hydrogeological properties affecting vulner-
ability (EX.M.23-26).
UV disinfection research is targeted because it may be
the least costly technology for many small systems to
adequately treat for viruses, especially non-community
systems (i.e., those not having a distribution system), or
multiple-well systems (RM.M.9 and RM.M.13). Also,
systems using UV disinfection would likely avoid forma-
tion of DBPs of any health risk significance.
Approaches for monitoring to ensure public protection
from waterborne viruses as well as bacteria may include
coliphage and are being investigated (see related
projects).
2-14
-------
Table 11-3. Research to Support the Groundwater Disinfection Rule
Expected Completion Date
Priority
A. Health Effects Research
1. Health Research on Waterborne Pathogens
HE.M.5 Infectious dose of Norwalk virus
HE.M.6 Infectious dose of other priority pathogens
HE.M.9 Investigations of waterborne disease outbreaks
HE.M.10 Surveillance tool for waterborne disease outbreaks
AWWARF Related Projects
268 Fingerprinting techniques for opportunistic pathogens & illness
B. Exposure Research
1b. Microbial Methods—Viruses
EX.M.8 Application of PCR technologies & gene probes to detect viruses in water
EX.M.9 Norwalk virus
EX.M.lOa Methods for emerging viruses
EX.M.1 Ob PCR-based detection of viruses in water
AWWARF Related Projects
292 Rapid PCR-based monitoring of entero viruses
429 Male-specific coliphages as indicators of viruses
612 Analysis of viruses by gene probe
726 Viral and microbial methods for groundwater
916 PCR technologies for virus detection in groundwater
2a. Microbial Exposure—Pathogen Occurrence
EX.M.14 Occurrence of Mycobacterium
EX.M.15 Occurrence of heterotrophic bacteria with virulence characteristic
EX.M.16 PCR method for Legionella
EX.M.17 Pathogenicity of heterotrophic bacteria found in drinking water.
EX.M.18 Occurrence of opportunistic pathogens in biofilms
AWWARF Related Projects
186 Survey viruses in groundwaters
3a. Groundwater—Survival and Transport in the Subsurface
EX.M.23 Virus survival in the subsurface
EX.M.24 Virus transport in the subsurface
AWWARF Related Projects
262 Field study of virus and indicator transport in groundwater
3a. Groundwater—Methods for Protecting Wells and Springs
EX.M.25 Viral transport and fate models
EX.M.26 Vulnerability of groundwater to pathogens
EX.M.27 Delineation of natural protection zones
EX.M.28 Vulnerability & sensitivity analysis
EX.M.29 Occurrence of Clostridium perfringehs
EX.M.30 Aquifer well mapping
EX.M.31 Correlation of water age and microbial viability
EX.M.32 Septic tank siting
EX.M.33 Virus sampling and phage research
C. Risk Assessment Research
1. Risks from pathogens
RA.M.1 Devel. compre. micro risk assess paradigm for water
RA.M.2 Evaluation/application of various dose-response models
RA.M.3 Evaluation and application of methods to assess risk associated with
exposures to multiple pathogens, routes, and durations
D. Risk Management Research
RM.M.9 UV disinfection efficiencies for Norwalk viruses
RM.M.30 Evaluate alternative inactivation processes for controlling viruses
AWWARF Related Projects
180 UV inactivation of viruses in natural waters
353 Inactivation rates of viruses and bacteria in saturated and unsatured subsurface
media
702 Development and validation of rational design methods of disinfection
809 By-products of UV treatment of groundwater
817 Membrane filtration techniques for microbial removal
2000
Proposed
Ongoing
Completed
1998
1999
2000
Proposed
1997
1998
Completed
Completed
1997
1997
1998
2001
Completed
1998
Proposed
Completed
2000
2000
1998
2000
2000
1998
1998
1997
Completed
1999
1997
2000
1998
2000
Proposed
Proposed
1997
1999
Completed
Completed
Completed
H
M
H
H
M
H
H
H
H
H
L
H
M
H
H
M
H
M
H
H
M
H
H
H
H
H
M
H
H
(Continued)
2-15
-------
Table 11-3. (Continued)
Expected Completion Date
Priority
NWRI Projects:
Deposition mechanisms and long-time scale factors influencing virus transport Completed
in porous media
Groundwater transport of viruses Completed
Reid experiment and modeling of vira transport in groundwater Completed
Transport and fate of viruses in the vicinity of pumping wells Completed
2-16
-------
Chapter ill. Research for Microbial Pathogens
Background
Waterborne infectious disease outbreaks have been .
attributed to a variety of pathogenic bacteria, parasites
and viruses. Giardia has been responsible for about half
of the outbreaks of disease where the causative agent
was identified. Cryptosporidium outbreaks have been
reported less frequently, but the number of cases asso-
ciated with the outbreaks have been much larger. Since
1,991, the percent of outbreaks attributable to
Cryptosporidium has doubled. In 1993-1994 reporting
period, about 17% of all reported outbreaks were caused
by Cryptosporidium. Many outbreaks of acute gastrointes-
tinal illness have not been linked to specific pathogens.
Viruses are thought to be the cause of a large portion of
the outbreaks where the agent is not identified. Estab-
lishing the association between water ingestion and
illness is difficult because of our inability to culture many
of the viruses. In addition to reported outbreaks where
the cause is unidentified, it is likely that many outbreaks
occur that are either unrecognized or unreported. Fur-
thermore, if it is determined through further research
that endemic microbial disease is a real occurrence (as
suggested by studies in Canada), public health concern
over waterborhe pathogens would increase significantly.
The- sources of pathogenic microbes in drinking water
associated with disease are usually related to fecal
matter from warm-blooded animals and humans. Hu-
mans are the main source of pathogens, but animals are
frequently implicated as in the outbreaks caused by
Giardia and Cryptosporidium. Another source of poten-
tial pathogens is the water distribution system and its
associated storage facilities. Breaks in the integrity of
the distribution system can allow the introduction of
pathogens. In addition, many bacteria can grow and
persist in the distribution system, and some of them
(e.g., Mycobacteria and Aeromonas) can cause infec-
tions and disease under certain conditions. The condi-
tions are not well defined, but they can include a host
whose natural defense barriers have been compro-
mised, an organism that has the ability to take advan-
tage of the opportunity presented by a compromised
host, and an environment that is conducive to the growth
of the pathogen.
This chapter describes proposed research for microbial
pathogens in the areas of health effects, exposure, risk
assessment, and risk management. The major research
questions in each area are summarized in Table 111-1.
For each question, the state of the science, research
needs, and proposed research projects are described.
Health Effects Research
1. What are the major pathogens of public health
concern?
2. What is the nature and magnitude of disease
associated with exposure to these waterborne
agents?
State of the Science
The continued occurrence of waterborne disease out-
breaks in the U.S. and questions in the scientific com-
munity about the adequacy of conventional water treat-
ment in preventing endemic waterborne disease high-
light the need for research to evaluate the impact of
water quality and type of treatment process on the
occurrence of waterborne disease. A critical public health
issue relates to the need to identify and characterize
pathogens that may pose increased risks of infection
Table 111-1. Major Research Questions for Microbial Pathogens
Health Effects
1. What are the major pathogens of public health concern?
2. , What is the nature and magnitude of disease associated with
exposure to these waterborne agents?
Exposure
1. What methods are needed to adequately measure or estimate
occurrence of pathogens in drinking water?
2. What are the frequencies of occurrence and densities of
pathogens in source water, finished water, and distribution
. systems, and what is the population distribution of exposures to
pathogens?
3. What are the factors affecting microbial contamination of
groundwater? - .
Risk Assessment •..'.,.-
1. How can the risks posed by pathogens in drinking water be.
characterized? , • . .
Risk Management . .
1. How effective are various treatment processes in removing
, pathogens? ' "- : ....
2. How can the quality of treated water be maintained in distribu-
tion systems?
3. How can source water be protected to ensure that it is consis-
tent with finished water quality of acceptable microbial risk after
appropriate treatment? " . '.
3-1
-------
and disease in the general population, as well as in
susceptible populations such as the immuno-compro-
mised, elderly, or infants.
As discussed earlier, a variety of pathogenic bacteria,
viruses and parasites are still considered to pose seri-
ous public health risks when treatment is inadequate. In
general, the health risks associated with exposure to
these agents in drinking water are poorly characterized.
The known source water contaminants of greatest con-
cern include the protozoan parasites Cryptosporidium
and Giardia, and the enteric viruses hepatitis A and
Norwalk virus. "Emerging" pathogens about which rela-
tively little is known but which could be important public
health concerns under certain conditions include
Microsporidium, Cyclosporidium, hepatitis E, and
Helicobacter pylori. Opportunistic or conditional organ-
isms are those that must be amplified in the environment
and then come into contact with a host that is suscep-
tible to disease. Examples of these types of organisms
are Legionella and Mycobacterium. Research is needed
to characterize the magnitude of the risk caused by
these organisms in susceptible populations and to more
carefully describe the risk factors for susceptibility.
Characterizing microbial health risks requires knowl-
edge of a number of important pathogen- and host-
specific factors. Reliable information on the infectious
dose of these agents is necessary for quantitative risk
assessments, yet such data are either limited or unavail-
able for most important waterborne pathogens. A need
also exists to better characterize the virulence of these
agents, the range of responses to infection, and the
influence of host factors (e.g., immune status) on the
course of infection and disease. Understanding the viru-
lence characteristics of pathogenic and opportunistic
microorganisms is important for estimating the likelihood
that infections will lead to disease in healthy and com-
promised individuals. Humans exposed to pathogens
may in some cases be asymptomatic or may experience
mild to severe effects. Although gastrointestinal symp-
toms are most commonly observed, other disease symp-
toms or states (e.g., cardiovascular, respiratory, liver,
and central nervous system effects) may also be in-
duced by certain pathogens. Finally, secondary spread
of disease is an important but poorly quantified factor
that should be investigated.
Cryptosporidium is known to cause an acute, self-limit-
ing disease in immunocompetent persons. Infections in
immuno-compromised individuals generally cause a more
chronic, severe disease for which no safe and effective
form of treatment is available. Additional health research
on Cryptosporidium is needed to permit a more compre-
hensive characterization of the important pathogen- and
host-specific factors described above. Some data on the
infectious dose of Giardia are available, and additional
research on this protozoan is considered to be a lower
priority than Cryptosporidium at the present time since
Cryptosporidium appears to be more resistant to disin-
fection. Very little is known about the pathogenicity of
the enteric viruses, which are believed to be responsible
for a significant portion of the outbreaks for which the
causative agent was not identified. A continuing investi-
gation of the infectious dose of Norwalk virus should
provide important information to help characterize the
risk posed by this agent. Important data gaps exist for
hepatitis A virus, but research in humans is problematic
because of its highly pathogenic nature. As current
studies resolve critical issues for priority pathogens such
as Cryptosporidium and Norwalk virus, efforts will shift
to emerging pathogens. Clearly, the list of known and
emerging pathogens for which major data gaps exist is
extensive and growing, and prioritization of these agents
for further study will be an ongoing effort as new infor-
mation on exposure and health concerns becomes avail-
able.
Epidemiology studies of both endemic waterborne dis-
ease and epidemic events (i.e., outbreaks) can be highly
useful in characterizing the magnitude of microbial risks,
identifying etiologic agents, characterizing the health
effects associated with a variety of different pre- and
post-treatment contamination problems, evaluating dif-
ferences in susceptibilities of different populations to
pathogens, and validating risk models. These studies
can also help to identify important research needs in the
areas of health effects, exposure, risk assessment, and
engineering. Exposure research on known opportunistic
and "emerging" waterborne pathogens should be linked
with health research in the laboratory and field to permit
a determination of the public health importance of find-
ing these agents in water supplies. Better surveillance
tools are needed to enable public health officials and
scientists detect and respond more quickly when water-
borne disease is first identified in a community.
The existing epidemiologic data base is inadequate for
determining the nature and magnitude of endemic and
epidemic waterborne disease in the U.S. A limited num-
ber of follow-up studies of the most significant outbreaks
have been conducted. The results of two recent studies
of endemic waterborne disease in Canada indicated that
there was an excess risk of gastrointestinal illness of
15% to 35% in the study groups that consume water
treated by conventional techniques. As described be-
low, the EPA has initiated epidemiology studies to fur-
ther address the issue in U.S. populations, and addi-
tional studies are being planned through a collaborative
effort between EPA and the Center for Disease Control
(CDC). The EPA and CDC are also working together to
develop a biannual report on waterborne disease out-
breaks. Finally, the EPA is developing improved esti-
mates of annual incidence of endemic waterborne dis-
ease in the U.S., using more recent pathogen occur-
rence data and improving estimates of exposure and
health effects.
3-2
-------
Research Topics and Priorities
a. Pathobiology of infection and disease for the
most important waterborne pathogens
Research on the pathobiology of infection and disease
includes studies to describe dose-response relation-
ships, characterize pathogen virulence and the range of
outcomes of infection, and evaluate the impact of host
immune status on infection and disease. Two approaches
for obtaining this information include highly controlled
clinical studies in humans and laboratory research in
animals. Epidemiology studies, which are described in
section (b), are also very effective tools for characteriz-
ing infections and the factors that influence disease
outcomes.
Nearly all of the health effects research on microbial
pathogens was assigned a high priority because it ad-
dresses high risk, high uncertainty issues of consider-
able regulatory importance. The 1996 SDWA Amend-
ments specifically require EPA to conduct the types of
research described below, including dose-response stud-
ies on pathogens such as Cryptosporidium and Norwalk
virus, and studies of the factors that influence these
relationships in susceptible populations. If individuals
exhibit significant immune response to Cryptosporidium
(project HE.M.1), if the current high-to-low dose extrapo-
lation of risk is not appropriate (project HE.M.2), or if
oocyst strain infectivity (project HE.M.3) in immunocom-
petent individuals varies significantly, then these factors
would greatly affect the overall risk assessment. From a
cost-benefit perspective, research that better defines
the dose-response curve and its significant associated
uncertainties for the most important waterborne patho-
gens is a high priority, because of the high and possibly
exponentially increasing costs of treating water to de-
crease the concentration of viable cysts to lower and
lower levels. Over-predicting the risk could have dra-
matic cost implications as well as cause utilities to
change to alternate disinfectants such as.ozone. Project
HE.M.4 was ranked high because it addresses the
critical risk assessment issue of variations in susceptibil-
ity to Cryptosporidium, and it could affect the EPA/CDC
guidance to severely immuno-compromised persons.
The Norwalk infectious dose study (project HE.M.5) is
also a high priority because of its potential impact on the
risk assessment for groundwater systems. As laboratory
techniques to identify and characterize pathogenic or
toxigenic microorganisms improve, and as current re-
search activities resolve critical uncertainties for
Cryptosporidium and Norwalk virus, priorities for re-
search on other pathogens such as those considered in
project HE.M.6 would increase from medium to high.
HE.M.1—Infectious dose of Cryptosporidium
Infectious dose study in groups of immunocom-
petent volunteers with and without preexisting
antibodies to this protozoan, with rechallenge of
the latter group of volunteers one year after
initial exposure.
Priority: High
HE.M.2—Validity of the dose-response model
for Cryptosporidium using low challenge
doses in test animals Dose-response study in
mice to evaluate the suitability of the low dose
extrapolation approach used in the existing
model.
Priority: High
HE.M.3—Cryptosporidium virulence study
using different strains Identify, collect, and
propagate in calves four isolates of C. parvum
from geographically diverse sites, identify mo-
lecular or biochemical markers of virulence, and
ultimately establish dose-responses in humans
for two of the new isolates.
Priority: High
HE.M.4—Cryptosporidium infectious dose in
normal and immuno-compromised animals
Determine and compare the infectious dose of
oocysts in immunocompetent and immuno-
compromised juvenile and adult guinea pigs.
Priority: High
HE.M.5—Infectious dose of Norwalk virus
Second phase of an EPA-funded infectious dose
study in humans to further characterize the dose-
response and the influence of host-specific fac-
tors (such as immune status) on infection and
disease. The first phase will be complete in
1997, and the second phase will end in 2000.
Priority: High
HE.M.6—Infectious dose of other priority
pathogens (to be determined) Studies of patho-
gens such as rotavirus, opportunistic bacteria,
or emerging pathogens, to be prioritized upon
further consideration of needs.
Priority: Medium
b. Characterization of epidemic and endemic water-
borne disease
Research described in this section includes community-
based epidemiology studies to characterize endemic as
well as epidemic waterborne infectious diseases. A ma-
jor area of interest is the microbial health risks associ-
ated with the use of different drinking water treatment
techniques (particularly filtration vs. no filtration). An
important aspect of this research is to develop better
tools to assess exposure to waterborne pathogens in
epidemiology or surveillance studies, and to facilitate
early detection of outbreaks when they occur. The
projects described below, like those identified above,
are considered high priority because they fulfill the top
three criteria of addressing high risk, high uncertainty
issues that have significant regulatory impact. These
efforts will lead to a better understanding of the magni-
tude of endemic and epidemic waterborne disease, and
of the factors that contribute to their occurrence. Better
3-3
-------
sensitivity in epidemiological studies will allow for im-
proved validation of dose-response predictions from
existing/newly developed dose-response curves. The
projects listed below are considered a high priority be-
cause they are expected to narrow the uncertainty in risk
estimates for Cryptosporidium and for other priority patho-
gens as well. Furthermore, the 1996 SDWA Amend-
ments require EPA and CDC to jointly conduct water-
borne disease occurrence studies for at least five major
U.S. communities or public water systems, and to de-
velop a national estimate of waterborne disease occur-
rence.
HE.M.7—Characterization of endemic dis-
ease: Health effects associated with differ-
ences in source water quality and treatment
process This includes epidemiologic research
to evaluate the impact of different treatment
strategies on reducing endemic waterborne dis-
ease. This effort should be linked to exposure
research involving the identification of new patho-
gens and the development of field detection
techniques, as well as to risk management and
assessment research activities. This research
should also provide opportunities for studying
subpopulations that may be highly susceptible
to waterborne infectious disease. Both surface
and groundwater sources of varying water qual-
ity should be considered for study. In an ongo-
ing, phased effort, potential study sites in sev-
eral communities have been identified where
treatment changes are planned, and the health
status of selected communities before and after
these changes will be determined. The EPA/
CDC waterborne disease occurrence studies
required by the 1996 SDWA Amendments have
also been initiated in 1997. To the extent pos-
sible, these studies will include efforts to identify
the etiology of waterborne illnesses that are
reported.
Priority: High
HE.M.8—Immunological assays for assess-
ing exposure in epidemiology studies Devel-
opment or improvement of immunological tools
to determine the presence of antibodies to im-
portant waterborne pathogens, and application
of these tools in community-based studies. This
research explores the possibility that serological
tools may be useful tools for assessing expo-
sure to important waterborne pathogens. An
ongoing phased study is evaluating the use of
serosurveys as a tool for assessing exposure to
Cryptosporidium. After completing the feasibility
phase to more fully evaluate the usefulness of
this technique, full-scale studies in several cities
may be conducted. Concurrent with this effort is
a study to evaluate the ability of the ELISA
technique to measure Cryptosporidium antibod-
ies in human sera.
Priority: High
HE.M.9—Investigations of waterborne dis-
ease outbreaks Rapid deployment of EPA sci-
entists with appropriate expertise in health, ex-
posure and control technology to conduct epi-
demiologic investigations of outbreaks as they
occur. This would greatly enhance the opportu-
nity to characterize the microorganism(s) re-
sponsible for the outbreak, the host factors that
influence infection and disease (including the
identification of susceptible subpopulations), and
the status of the treatment process prior to and
during the outbreak. This would be a joint initia-
tive with the CDC, and would involve appropri-
ate state and local agency personnel.
Priority: High
HE.M.10—Surveillance tools for waterborne
disease outbreaks Development of tools or
mechanisms of surveillance to improve the rec-
ognition and subsequent investigation of water-
borne disease in communities. An ongoing study
is evaluating the feasibility of using anti-diar-
rheal drug sales as a surveillance tool for water-
borne disease outbreaks.
Priority: High
Exposure Research
1. What methods are needed to adequately mea-
sure or estimate occurrence of pathogens in
drinking water?
State of the Science
Methods for measuring some of the important patho-
gens associated with waterborne infectious disease are
either nonexistent or not effective because they are too
inaccurate and imprecise. In either case, these short-
comings have produced great uncertainty with respect
to exposure assessments, which has hindered the es-
tablishment of sound regulations for drinking water qual-
ity and led to the use of many assumptions in character-
izing risks associated with drinking water.
Giardia and Cryptosporidium usually occur in very low
densities in source water and finished drinking water.
This characteristic necessitates sampling large volumes
of water, sometimes up to 1000 liters, through a filter.
This and subsequent steps in the currently available
method may contribute to its inaccurate and imprecise
nature. A major issue is that only a small fraction of the
total sample volume collected can be feasibly analyzed;
low-percent recoveries and poor precision are difficult to
avoid. The current procedures are being evaluated and
improved by EPA scientists and other researchers. New
methodology, such as the gene probe polymerase chain
reaction method, is being developed and will become
available for evaluation at some time in the future.
Some methodology is available for measuring water-
borne viral pathogens such as the enteric viruses,
3-4
-------
rotavirus, and hepatitis A virus. In the case of the latter
two viruses, the methods for their detection and quantifi-
cation are time-consuming and non-facile. Norwalk virus
has been associated with many waterborne outbreaks
as shown by seroepidemiology. This virus, however,
cannot yet be grown in tissue culture. This lack of a
suitable method for detecting and measuring this viral
pathogen has limited the conduct of exposure assess-
ments and meaningful effects assessments.
Some of the newly emerging pathogens associated with
water cannot be efficiently measured in water samples
because suitable methods are not available. Two ex-,
amples of pathogenic protozoa for which monitoring
methods do not exist are Cyclospora and Microsporidia.
Methods for these parasites should be developed and
evaluated in anticipation of the future need to measure
these pathogens. Similarly, the waterborne pathogen
Mycqbacterium avium cannot be easily measured be-
cause the available methods have been developed for
testing clinical samples. The same can be said for
Helicobacter pylori, another potential waterborne patho-
gen that has been associated with cases of gastric
ulcers and stomach cancer in .humans.
Pathogens may be present in tap water due to inad-
equate disinfection or may enter faulty distribution sys-
tems post treatment. Due to the acute nature of most
microbial diseases, real-time continuous monitoring tech-
niques are needed to detect the possible presence of
pathogens in distribution systems and to provide timely
advice to water consumers. ;
Research Topics and Priorities
a. Methods for detecting and enumerating
Cryptosporidium,and Giardia (including identi-
fying the viability potential) in source and fin-
ished drinking water, and methods for other
emerging protozoa
Cryptosporidium is emphasized because of the great
attention placed on the organism due to recent out-
breaks and its resistance to inactivation by chlorine. This
protozoan may become the focal organism for drinking
water regulations. Research that is expected to result in
improved methods for detecting Cryptosporidium or in.a
method for detecting viable oocysts has generally been
assigned a high priority. These projects are expected to
improve EPA's ability to estimate occurrence and expo-
sure. Project EX.M.4 was assigned as medium priority
because this methodology will not provide means of
determining viability or infectivity of cysts or oocysts.
EX.M.1—Immunological techniques for pro-
tozoa Conduct research to examine and im-
prove immunological techniques for detecting
Cryptosporidium, and other pathogenic proto-
zoa. Practical, improved technology will enhance
the isolation, identification, and quantification of
pathogenic protozoa.
Priority: High
EX.M.2—Gene probes for detection of viable
Cryptosporidium oocysts Develop probes to
mRNA and other nucleic acids, and examine
their potential as viability markers for patho-
genic protozoan cysts arid oocysts.
Priority: High
EX.M.3—Cultural method for Cryptospo-
ridium in environmentarsamples Develop
a cultural method for detecting Cryptosporidium
in environmental water samples. The objective
Is to develop and evaluate a tissue culture as-
say for oocysts.
Priority: High
EX.M.4—PCR methods for Giardia and Cryp-
tosporidium Develop polymerase chain
reaction (PCR) methods for Giardia and
Cryptosporidium in environmental samples.
Methods will be investigated for recovering pro-
tozoa nucleic acids directly from water samples.
Priority: Medium
EX.M.5—Protozoa methodology protocol de-
velopment workshop (Ongoing) This workshop
will produce protocols for developing and evalu-
ating a new method; for comparing methods;
and for determining the equivalency of proce-
dures, reagents, and materials within a method.
It also will develop recommendations on sample
volume for raw waters to be included in a com-
parison survey. ,
Priority: High
EX.M.6—Comparison of methods for Giardia
and Cryptosporidium In water Comparison of
the best available methods for detecting and
quantifying cysts and oocysts through multi-lab
evaluation using water samples from multiple
sites, using criteria developed in project EX.M.5.
Priority: High
EX.M.7—New protozoa agents Develop meth-
ods to identify new potential protozoan patho-
gens in drinking water such as Cyclospora or
Microsporidia. The objective of this research is
to develop procedures, either in vitro or in vivo,
for producing stocks of these potential patho-
gens. Methods for detection and identification
will be developed and evaluated using the stock
organisms.
Priority: High
b. Methods for detecting and enumerating viruses
in source and finished drinking waters
Molecular biology techniques have allowed for the de-
tection of viruses in water samples but they cannot
3-5
-------
indicate whether a virus is or is not infectious. To deter-
mine infectivity, the virus needs to be grown in tissue
culture. Accurate exposure assessments will not be
possible until these research needs are met.
Improving methods for detection and infectivity assess-
ment of Norwalk virus (EX.M.9) was considered a high
priority because it may be used as a target virus for
treatment and/or risk assessment, especially for ground-
water systems (see Chapter II, section on groundwater
disinfection rule). Methods for virus identification which
do not indicate infectivity are generally of medium prior-
ity (EX.M.8). Project EX.M.10 is a high priority because
this data will help determine if viruses other than HAV or
Norwalk may be more appropriate for defining disinfec-
tion conditions necessary to protect from viruses in
ground and surface water supplies.
EX.M.8—Application of PCR technologies
and gene probes for virus detection in water
(Ongoing) Conduct studies to evaluate the ap-
plication of the polymerase chain reaction and
gene probes for virus detection in water. The
viruses being studied are enteric adenovirus,
Coxsackie virus, hepatitis A viruses, and
rotavirus.
Priority: Medium
EX.M.9—Norwalk virus Develop a cultural
method for Norwalk and Norwalk-like viruses.
The objective of this study will be to identify ceil
culture lines that will replicate Norwalk virus
RNA. Develop PCR/gene probe methods for the
In situ detection of the replicated particles.
Priority: High
EX.M.10—Methods for emerging viruses De-
velop methods for identification and quantifica-
tion of emerging viruses.
Priority: High
2. What are the frequencies of occurrence and
densities of pathogens in source water, finished
water, and distribution system water, and what
is the population distribution of exposures to
pathogens? What are the sources of pathogens
in source waters and finished drinking water?
For example, livestock, wildlife, wastewater man-
agement, distribution systems, etc.
State of the Science
The occurrence and densities of waterborne pathogens
in source waters, finished drinking water, and distribu-
tion systems are not well known. Giardia cysts and
Cryptosporidium oocysts have been detected in source
waters and also in drinking water samples. In one study,
fewer than 20% of the drinking water samples contained
Cryptosporidium oocysts, whereas more than 50% of
the surface and spring water samples contained
Cryptosporidium oocysts. Giardia cysts were not found
in drinking water samples, but they were detected in
surface water samples, including pristine river water
samples. In another study, 78% of watersheds were
positive for Giardia on multiple occasions, while 86% of
the sites were multiply-positive for Cryptosporidium. All
of these studies of the occurrence of protozoa in source
and drinking waters have limitations. Surveys of proto-
zoan cysts and oocysts are few in number and, there-
fore, difficult to translate to the national level. The results
of the protozoan surveys, although accomplished with
state-of-the-art methodology, still leave much uncer-
tainty because the method used lacks sensitivity and the
detection of cysts and oocysts does not indicate their
viability or whether or not they are infectious.
Another major issue is the extent to which watershed
control can limit pathogen concentrations in source wa-
ters. While several surveys have indicated lower Giardia
and Cryptosporidium levels in source waters of systems
with protected watersheds, the effectiveness of water-
shed management practices on reducing pathogen lev-
els is not well defined. However, it is clear that various
sources of watershed contamination will play a signifi-
cant role in risk management decisions. EPA recently
awarded a grant that will focus on detecting fecal con-
tamination and its sources in water and watersheds.
AWWARF is also sponsoring two projects in this area;
one project is examining the survival of Cryptosporidium
and Giardia exposed to different environmental condi-
tions including water temperature, cyst/oocyst stage,
and physical stress on the viability and susceptibility to
disinfection. Another study will evaluate the effects of
watershed management practices on pathogen levels in
source waters.
There is also insufficient information about viruses in
surface and groundwaters. Recent studies indicated
that 20% to 25% of all groundwaters were contaminated
by enteric viruses detected by cultural methods and
gene probe methods. Gene probe methods detect im-
portant viruses such as hepatitis A virus, rotavirus and
Norwalk virus, but they cannot determine if the viruses
are infectious and, therefore, whether they pose a defi-
nite risk to exposed individuals. This again points to the
need to develop tissue culture methods for the more
important viruses so that this significant information gap
can be filled. H
Although we have a relatively good understanding of the
linkages between pathogens in drinking water and ill-
ness in water consumers, there are still many questions
to be answered, such as where do pathogens occur,
how many are there and where did they come from.
Some of the questions have been answered for selected
pathogens. Some studies have been conducted to de-
termine the occurrence in nature of Giardia and
Cryptosporidium, but they have been limited in scope.
National surveys have not been conducted and these
are critically needed if the extent of the potential risk
from these organisms is to be defined. Such studies are
in the planning stages. Under the Information Collection
Rule, EPA intends to require that source waters and
3-6
-------
finished waters at water utilities all over the U.S. be
surveyed to determine the occurrence otGiardia, Cryp-
tosporidium, enteric viruses, and indicators. This infor-
mation is sorely needed for developing the Enhanced
Surface Water Treatment Rule. However, since the ICR
will not include intensive monitoring during storm events,
optimal regulatory monitoring strategies to define treat-
ment needs under maximum stress situations will not be
apparent unless additional survey work is conducted.
Also, as discussed above, the lack of effective methods
for enumerating pathogens and for indicating whether
pathogens are viable or infectious will compromise the
risk assessment process.
Research should be considered for pathogens, espe-
cially emerging pathogens, where there is evidence that
they have the potential to become a significant water-
borne problem. The pathogenic protozoa Cyclospora
and Microsporidiaia\\ into this category. These microor-
ganisms are transmitted by the fecal-oral route and
cause severe gastrointestinal illness; there is some evi-
dence that they may be waterbbrne. Anticipatory re-
search should be initiated to determine how frequently
they occur and at what levels in source waters and
drinking water. In the event that they become a signifi-
cant problem, exposure assessment information will be
available to risk assessors and managers. Research
should also be conducted to determine the extent and
densities of emerging bacterial pathogens, such as My-
cobacteria and aeromonads, in drinking water.. These
microorganisms are known to be transmitted via the
waterborne route and studies have shown the linkage
between water and illness in water consumers/Since
the presence of the organisms is not identified by com-
mon measures of drinking water quality, there is no
information available on their occurrence and levels.
Just as the virulence, survival, and infectious dose char-
acteristics in microbial pathogens can vary and affect
their ability to cause waterborne disease, there are
many characteristics of exposed individuals and human
populations that vary and influence the course of infec-
tion and disease. Areas of uncertainty include the pro-
portion of the exposed population at highest risk (such
as the immune-compromised, the aged, and very young);
how much water is consumed (boiled vs. non-boiled
water, tap vs. bottled water); the extent and effect of
protective immunity; and the significance of other expo-
sures, e.g., foods that carry the same microbial patho-
gens as those found in water. Characterizing the nature
of exposed populations and their behavioral patterns will
provide valuable information for the risk assessment
process.
Research Topics and Priorities
a. Surveys to determine pathogen occurrence in
source and finished waters
Research in this area will help in determining optimum
survey strategies, sampling frequencies, and statistical
parameters for adequately-representing or determining
pathogen occurrence in source waters. Watershed char-
acteristics (e.g., land/water use, population densities,
and hydrology) and temporal storm events are important
variables that need to be considered together when
collecting, analyzing and interpreting microbial survey
data. In addition, the research will provide information
on whether indicators can be used for estimating source
water pathogen occurrence (ground or surface) or the
absence of pathogens in finished waters (ground or
surface). Finally, pathogen occurrence data can be used
to evaluate predictions of aquifer vulnerability.
Project EX.M.11 is considered a high priority because it
will supplement ICR information by providing data on
short-term variations in the microbiological quality of raw
and treated water due to storm runoff and pollution
events. Such events and the problems associated with
treating rapidly deteriorating water quality are consid-
ered critical risk factors in, the occurrence of waterborne
disease outbreaks and spikes in eridemic illness (not
detected as an outbreak). Data from this project will
clarify the range of pathogen occurrence, potential indi-
cators, and appropriate targets for defining minimum
treatment under the ESWTR, e.g., whether risk reduc-
tion levels could be set based on the 90th percentile
level of oocysts detected based on some minimum
frequency of monitoring, e.g., monthly sampling for 18
months. Where year-round target treatment levels are
set, i.e., based on source water measurements and with
a monitoring scheme which is intended to prevent big
spikes in endemic levels, has enormous cost/benefit
implications. This study would also validate treatment
effectiveness at full-scale treatment plants by targeting
water systems with high source water pathogen load-
ings. Project EX.M.12 is a medium priority unless new
evidence indicates that emerging viruses are more re-
sistant to disinfection than HAV or Norwalk.
EX.M.11—Intensive evaluation of microbio-
logical constituents and treatability in sur-
face source waters Develop data on short-
term fluctuations in pathogen occurrence re-
lated to meteorological and pollution events.
These data are needed to evaluate the effec-
tiveness of treatment and control measures.
Validate alternative and prototype methods for
indicators and pathogens. Include microbiologi-
cal parameters and methods as in the ICR, but
possibly use more effective analytical techniques,
should they become available.
Priority: High
EX.M.12—Identification of viruses resistant
to disinfection As new virus methods become
available, evaluate new viruses for their poten-
tial occurrence in drinking waters and their re-
sistance to disinfection. Studies should examine
which viruses are most likely to occur in fecally
contaminated source waters receiving different
levels of disinfection.
Priority: Medium
3-7
-------
b. Importance of watershed control (including point
source, non-point source, and septic tank con-
trols) for source water pathogen occurrence
Often it is difficult to determine the origin of non-poitit-
source fecal contamination in surface water source and
therefore difficult to prioritize watershed control mea-
sures. Project EX.M.13 will provide an additional tool for
water utilities to assess whether non-point-source fecal
contamination is of animal or human origin.
EX.M.13—Distinguish animal versus human
sources Distinguish animal from human sources
of pollution using gene probe methodology based
on phylogenetic differences between human and
non-human E. coliorBacteroidesor immunoas-
says for human related chemicals such as caf-
feine. These probes will be used to identify
sources of pollution.
Priority: High
c. Occurrence of and exposure to primary and op-
portunistic pathogens in distribution systems
Project EX.M.14 is a high priority because Mycobacte-
rium is relatively more resistant to disinfection than most
other bacteria, and because a published study showed
an association between hospital-acquired infections and
this organism in drinking water. Projects EX.M.15 and
EX.M.17 were assigned as high priority because they
are expected to provide information which, in conjunc-
tion with the results of other ongoing studies, will estab-
lish whether heterotrophic bacteria in the distribution
system pose a risk to immunocompetent individuals.
Project EX.M.21 is also a high priority, as it is anticipa-
tory in nature and will address potential pathogens that
could affect both immunocompetent and immuno-com-
promised individuals. Depending on the outcome of this
research (i.e., finding of opportunistic pathogens in the
distribution system) the medium or low priority of the
other projects on opportunistic bacteria in this section
and in the distribution system risk management section
would be reassessed.
EX.M.14—Occurrence of Mycobacterium De-
termine the occurrence of Mycobacterium
avium (MAC) in potable water distribution sys-
tems and compare MAC water isolates to MAC
strains isolated from clinical specimens obtained
from immuno-compromised patients.
Priority: High
EX.M.15—Occurrence of heterotrophic bac-
teria with virulence characteristics Determine
the frequency of occurrence of heterotrophic
bacteria isolated from drinking water that dem-
onstrate virulence characteristics.
Priority: High
EX.M.16—PCR method for Legionella Develop
a method for detecting Legionella in water using
the polymerase chain reaction (PCR) proce-
dure. Also determine whether Legionella grow-
ing inside amoebae in potable water can be
detected by PCR.
Priority: Low
EX.M.17—Pathogenicity of heterotrophic bac-
teria found in drinking water
Conduct surveys to identify heterotrophic op-
portunistic bacteria associated with drinking wa-
ter and granular activated charcoal (GAC) filter
effluents. Identify potential pathogens using com-
promised animals. Establish a data base for
distribution systems on a regional and national
scale.
Priority: High
EX.M.18—Occurrence of opportunistic patho-
gens in biofilms Examine the occurrence of
specific opportunistic pathogens in biofilms as-
sociated with treated drinking water distribution
systems and how they are affected by various
factors. Organisms of special concern are
Aeromonas, Pseudomonas, and non-tuberculo-
sis Mycobacteria.
Priority: Medium
EX.M.19—Opportunistic pathogens associ-
ated with point-of-use (POU) and point-of-
entry (POE) filter effluents Determine the po-
tential for colonization and growth of opportunis-
tic pathogens in POU/POE water treatment de-
vices and microbial purifiers under actual use
conditions.
Priority: Medium
EX.M.20—Potential pathogen icity of het-
erotrophic bacteria eluted from point-of-use
GAC filters Initial finding indicate that some
opportunistic pathogens can grow on GAC fil-
ters. Research is needed to determine the po-
tential for heterotrophic bacteria from GAC POU
filters to cause illness in compromised animal
models. Data would be used in developing hu-
man risk estimates.
Priority: Medium
EX.M.21—Occurrence of newly emerging
pathogens Conduct a national survey to deter-
mine the occurrence of newly emerging and
potential pathogens in potable water distribution
systems. Pathogens and opportunistic patho-
gens of special interest are Aeromonas,
Pseudomonas, and Helicobacter.
Priority: High
EX.M.22—Exposure as a function of popula-
tion distribution Conduct a series of studies to
identify exposure as a function of age, other
3-8
-------
individual susceptibility factors such as protec-
tive immunity, behavioral patterns and environ-
mental factors that affect water consumption.
Priority: High
3. What are the factors affecting microbial con-
tamination of groundwater?
State of the Science
Groundwater is the source of drinking water for about
1/3 of the U.S. population. There are approximately
180,000 community and non-community public water
systems utilizing groundwater .About half of the ground-
water community systems disinfect, but a majority of the
non-community systems do not. The drinking water
quality of systems that do not treat for pathogens is
dependent on having source waters at the wellhead that
do not contain pathogens in sufficient numbers to cause
health problems, and on monitoring for indicator organ-
isms to warn of any microbial contamination.
The actual occurrence of pathogens in drinking water
originating from groundwater is uncertain. Preliminary
results of the groundwater survey being carried out by
AWWARF indicate that greater than 20% of the well
waters sampled contain viruses. These results are sur-
prisingly similar to the percentage of systems in which
viruses were detected in an earlier survey of vulnerable
groundwater systems by EPA and AWWARF. Many
state laboratories report that greater than 40% of the
private well waters tested contain coliform bacteria.
Records of waterborne disease outbreaks show that the
primary cause of waterborne disease outbreaks have
included groundwater that had not been disinfected.
This is particularly significant since it is estimated that
these records may underestimate the actual number of
outbreaks by an order of magnitude.
Although the disease agent in most groundwater-related
outbreaks was unknown, the majority of outbreaks were
believed to be caused by viruses. The primary viruses
known to be agents of waterborne disease are enteric
viruses such as polio, Coxsackie, echo, hepatitis A and
E, rotavirus, Norwalk, and Norwalk-like. The bacteria
primarily responsible for waterborne outbreaks are Sal-
monella, Shigella, enteropathogenic Escherichia coli,
and Vibrio cholera. The protozoa, Giardia and Crypto-
sporidium, are also potential groundwater contaminants
although their presence.would likely be due to poor well
location or construction.
The sources of pathogens in groundwater include sur-
face waters, septic tanks, cesspools, leaking sewer
lines, municipal land treatment systems, animal feeding
operations, sludge disposal areas and municipal land-
fills. Since pathogens must be transported by percolat-
ing water from these sources through the unsaturated
zones and then by regional or gradient flow through the
saturated zones to a well or spring, most control mea-
sures are based on placing sources far enough away
from springs or wells that any pathogens will either be
physically removed, die or be inactivated before they
reach a place of withdrawal. The extent of removal and
die off is dependent on numerous factors (not all of
which have been identified) pertaining to the character-
istics of the pathogens, the nature of the soils and
aquifer materials and the distance and time required for
the pathogens to be transported from a source to a well.
As a general rule, viruses are presumed be transported
further in the subsurface than either bacteria or protozoa
due to their much smaller size.
Methods in use or proposed for determining whether a
well or spring is in danger from a potential source of
contamination range from using arbitrary setback dis-
tances to some type of vulnerability assessment method
to predict the relative vulnerability of a water well or
spring to contamination by pathogens. A variety of meth-
ods have been proposed for doing groundwater vulner-
ability assessments including a) methods that give nu-
merical ratings to the physical factors (such as soil
types, geology, and depth to groundwater) that affect
groundwater vulnerability to contamination by patho-
gens, and b) mathematical models. Additionally, deci-
sion trees that combine several methods in an ordered
manner may be the most effective way to make this
determination.
Arbitrary setback distances are very difficult to defend
scientifically but are the easiest to implement from a
regulatory standpoint. The use of vulnerability indexes
or avoidance criteria to assess the vulnerability of ground-
water to pathogens has great potential for screening
purposes, but the use of vulnerability assessment meth-
ods in general has been questioned by the scientific
community if such methods are going to be used for
making site-specific decisions. The use of such methods
will probably require extensive evaluation and testing
before they are accepted as regulatory tools. Efforts are
being initiated by EPA to develop avoidance criteria for
the Groundwater Disinfection Rule (GWDR).
The use of predictive computer models for making regu-
latory decisions regarding disinfection of groundwater
has had difficulty in gaining regulatory acceptance al-
though significant advances have been made in the
development of mathematical models to predict the
transport and survival of pathogens in the subsurface. A
key element of these models is the inactivation rate
chosen for the pathogens (viruses). Survival times for
many of the primary pathogens of concern in the subsur-
face and in groundwater are unknown. Existing informa-
tion about a-few pathogens suggests that the die off of
bacterial pathogens is much faster than viruses, and
that survival times longer than one year for viruses
would be unlikely. However, data on inactivation rates
for pathogens under the various subsurface conditions
are limited and contradictory. A critical need is to better
determine these inactivation rates in the unsaturated
and saturated soils and aquifers. Limited additional work
in this area has recently been funded;by AWWARF, but
a more extensive effort is needed.
3-9
-------
The major hindrance in obtaining final acceptance for
any vulnerability assessment tool is the extensive field
testing needed to confirm the tool's protective capability
for making site-specific decisions. Extensive field valida-
tion is extremely difficult, time-consuming and expen-
sive and the number of field tests that can be performed
is limited by a lack of available resources.
The majority of groundwater-based public water sup-
plies (at least 140,000) are very small, non-community
systems with extremely limited resources. For these
systems, virus fate and transport models will not be
appropriate for predicting vulnerability. Vulnerability as-
sessments might best be made through a sanitary sur-
vey and wellhead protection approaches. Therefore,
general hydrogeological and land use criteria must be
considered and tested for their utility in predicting vul-
nerability. A high priority exists for developing practical
cost-effective approaches. The approach eventually cho-
sen is likely to be a decision tree that combines ele-
ments of several methods in a structured prioritized
manner.
On July 10-11,1996, the EPA's Office of Groundwater
and Drinking Water conducted a Workshop on Predict-
ing Microbial Contamination of Groundwater Systems.
Workshop participants identified the following indicators
as potentially useful for predicting fecal contamination of
groundwater: E. Co//, Enterococci, Clostridium
perfringens, and Coliphages (somatic and male-spe-
cific). The following research on indicators of fecal con-
tamination of groundwater was identified at the work-
shop: Determining how the survival and occurrence of
Clostridium perfringens correlates with the occurrence
of pathogens in contaminated groundwaters; delineating
the range of hosts for somatic coliphages to determine
whether they have enough specificity as an indicator of
fecal contamination; and measuring the occurrence of
male-specific coliphages in small populations to deter-
mine whether they have enough sensitivity to use for
small systems (< 100 people). Where appropriate to do
so, the projects described below incorporate the recom-
mendations of the workshop.
Research Topics and Priorities
a. Survival and transport of pathogens in the sub-
surface
Determination of the survival times of pathogens, espe-
cially viruses, in both the saturated and unsaturated
subsurface is the major research need for any method
for determining whether groundwater does or does not
have to be disinfected. Information on virus inactivation
times in the subsurface for many major viral pathogens
is nearly nonexistent. Survival data must be in a form
that can be incorporated into decision-support systems
for determining whether groundwater should not be
disinfected. Information on the factors determining the
extent of transport of pathogens in different hydrogeologic
settings, especially where preferential flow paths exist
and in the unsaturated zone, is also a major need if
transport models are going to be used.
Project EX.M.23 is considered high priority because it
addresses an area of high risk: the concentration and
occurrence of a major group of contaminants - viruses.
It also is an area in which currently there is a high
degree of uncertainty. For similar reasons, project
EX.M.24 is also a high priority. The assessment of viral
behavior in the subsurface requires knowledge of both
their fate and transport, which are intimately linked.
Furthermore, once viral die-off rates are determined, the
data can be applied to many models for estimating
setback distances. Project EX.M.26 is considered high
priority because this approach has the most promise for
developing practical criteria that can be implemented,
even if such criteria may apply only to a limited number
of hydrogeological situations. Currently, there is a high
degree of uncertainty in assessing the vulnerability of
groundwater to contaminants. Projects EX.M.29 and
EX.M.33 are considered high priority because partici-
pants from the Workshop on Predicting Microbial Con-
tamination of Groundwater Systems considered this re-
search as critical for supporting the Groundwater Disin-
fection Rule. The other projects are considered medium
priority because the feasibility of success in generating
criteria that can be practically implemented is very un-
certain, and research in this area is already underway.
EX.M.23—Virus survival in the subsurface
Inactivation times for the viral pathogens of
regulatory concern are not known for most soil
types and conditions. This work would deter-
mine inactivation times of viral pathogens in
several groundwaters and soils under both un-
saturated and saturated conditions. The project
will focus on the role of subsurface ecology,
including both geochemical and biological fac-
tors such as predation, as a determinant of
subsurface virus survival. Initial work will use
bacteriophages as surrogates for human enteric
viruses. Controlled laboratory experiments with
human enteric viruses will be used to confirm
that the mechanisms that control bacteriophage
survival also control enteric virus survival.
Priority: High
EX.M.24—Virus transport in the subsurface
Methods are needed for predicting the transport
time between wells and springs and sources of
pathogens. This research would evaluate the
factors which govern the transport of viral patho-
gens in both the unsaturated and saturated
zones of the subsurface. Subsequently, meth-
ods would be developed for incorporating this
information into either predictive models or deci-
sion-support systems for determining whether
groundwater should not be disinfected.
Priority: High
3-10
-------
b. Methods for protecting wells and springs from
pathogens
Scientifically defensible methods for determining set-
backs between sources and wells or springs are needed
to aid decision makers in determining whether ground-
water achieves "natural disinfection" or if chemical disin-
fection is required. Additional evaluation/testing of pro-
posed mathematical models such as CANVAS is needed
to address scientific or regulatory concerns. Methods to
assess the vulnerability of groundwater to pathogens by
the development of vulnerability indexes should be criti-
cally evaluated and tested. Both vulnerability rating sys-
tems/models may require additional field tests.
EX.M.25—Viral transport and fate models
Critical evaluation of the use of mathematical
viral fate and transport models for making deci-
sions on whether drinking water from ground-
water sources should or should not be disin-
fected. Emphasis will be placed on the feasibil-
ity and reliability of existing models. Available
information and data will be used when pos-
sible.
Priority: Medium
EX.M.26—Vulnerability of groundwater to
pathogens Critical evaluation, emphasizing fea-
sibility and reliability, of methods being devel-
oped to assess the vulnerability of groundwater
to pathogens by the development of vulnerabil-
ity assessment systems. Available information
and data will be used when possible.
Priority: High
EX.M.27—Delineation of natural protection
zones Examine and test an alternative method
to conventional contaminant transport models
and vulnerability assessment methods for delin-
eating a protective zone around a.well or spring.
The method will integrate the analytical element
method for delineation of time-of-travel capture
zones with information on survival and transport
of pathogens. The capture zone concept will be
tested with data developed under the New En-
gland virus study.
Priority: Medium
EX.M.28—Groundwater vulnerability and sen-
sitivity assessment methods This research
examines the range of physical, biological, insti-
tutional and operational factors that could be
used to determine these vulnerabilities, includ-
ing (but not limited to) hydrogeological factors,
land uses and groundwater/wellhead protection
program determinations. The ultimate product is
to be a pragmatic, field-tested assessment meth-^
odology that, given generally available informa-'
tion, can estimate vulnerability to microbial con-
tamination in small-to-medium groundwater sys-
tems. It may result in separate approaches for
larger and smaller systems, based on economic
feasibility. Initial work has been to develop ap-
propriate criteria. Subsequent work will be to
field validate the criteria against carefully de-
scribed vulnerable and non-vulnerable systems.
Priority: High
EX.M.29—Occurrence of Clostridium per-
fringens'm the subsurface C. perfringens will
be collected from over 100 groundwater sites as
part of a larger virus survey.
Priority: High
EX.M.SOa—National aquifer and well map A
random selection of drinking water systems us-
ing groundwater was mapped onto USGS aqui-
fer maps on a state-by-state basis. Community,
non-community and transient systems were
mapped separately based on their zip codes.
(Completed 1995)
Priority: Medium
EX.M.SOb—Sensitivity mapping A random se-
lection of drinking water systems using ground-
water was mapped onto USGS aquifer maps on
a state-by-state basis. Community, non-com-
munity and transient systems were mapped
separately based on their zip codes. (Com-
pleted 1995)
Priority: Medium
EX.M.31—Correlation of water age and mi-
crobial viability This work would investigate
the relationship between water age and the
presence of human enteric viruses.
Priority: High
EX.M.32—New York City Watershed Protec-
tion Septic Siting Study. The septic system
types to be investigated include continuous use
functioning systems, weekend use systems, sea-
sonal use systems and newly installed systems,
both residential and commercial. Study sites will
be representative of varying hydrological, chemi-
cal, physical and design parameters. A mini-
mum of three sites will be selected for the
Phase 1 investigation, which includes site se-
lection, installation of five monitoring wells at
each site and baseline data collection. Phase II
will include spiking studies using phage, bacte-
rial or other tracers.
Priority: High
EX.M.33—Phase 2 Virus Sampling and Ph-
age Research for the Santa Ana River Water
Quality and Health Study (includes Virus Test-
ing of Groundwater for GWDR and Santa
Ana River Water Quality and Health Study)
Samples of groundwater and river water will be
taken monthly for virus cell culture and PCR
3-11
-------
analysis. Bacteriophage and coliform samples
will also be taken. Phage will be evaluated as
indicator organisms for enteric viruses. All pro-
duction and monitoring wells sampled in the first
project year were negative for enteric viruses.
Priority: High
Risk Assessment Research
1. How can the risks posed by pathogens in drink-
ing water be characterized?
State of the Science
The Agency has not established a formal methodology
to assess the risk for microbial contaminants as it has for
chemicals. In 1991, a joint EPA/AWWARF conference
evaluating drinking water and health in the year 2000
proposed the application of the National Academy of
Science risk paradigm for microbial contaminants. How-
ever, as discussed previously microbial risk assessment
is relatively new and the issues that need to be ad-
dressed are more complex than those currently used for
defining chemical risks. The 1996 Safe Drinking Water
Act Amendments (SDWAA) contains provisions that
focus on the need to ensure that adequate health pro-
tection is conferred to the most sensitive subpopula-
tions, especially children, in all future drinking water
regulations, taking into consideration, cost and benefit of
disease prevention or control. Pathogen interactions
may pose overwhelming health risks to some groups of
Individuals such as infants, young children, the malnour-
ished, the elderly or immuno-compromised individuals.
Under the FY 97 SDWAA initiative additional resources
have been identified for the expansion of research ef-
forts relating to pathogen risk assessment. The focus of
risk assessment studies has been on infectivity and
determinations of the smallest pathogenic dose (num-
bers of infectious microorganisms) able to produce an
infection in healthy individuals. Studies are needed in
selected subpopulations to determine host susceptibili-
ties (lack of or reduced capacity of host resistance and
defense mechanisms) and relationship to pathogen en-
try and establishment of infection or other forms of
parasitism (e.g., transient carrier state, colonization,
pathogenic interaction). Studies are needed to relate
pathogen virulence and host resistance so that risk
assessments can be made about carrier states, coloni-
zation, infection, disease course and outcome (e.g.,
clinically asymptomatic persons, disease state, full re-
covery, chronic debilitate recovery, mortality). Increased
efforts will include research to assess and quantify
effects of immunity of exposed individual and popula-
tion; evaluation and quantification of the disease pro-
cess including secondary spread, magnitude and sever-
ity of effects, in addition to infection and the effects of
multiple exposures and routes of exposure.
Previously, a risk assessment model was developed for
Giardia contamination in drinking water by Rose, et al.,
1991, in which dose-response data from human volun-
teer studies were modeled using a Poisson distribution.
Actual dose-response curves were developed using a
low-dose extrapolation model and a maximum likelihood
estimate. Based on various input variables, such as
number of organisms and mean response rate, the risk
of infection and the risks of disease were estimated,
including risk from exposure to a single organism. Log
order reductions of cysts from source water needed to
achieve an acceptable risk level (e.g., less than one
infection per 10,000 people per year) can be calculated
using this approach. However, as noted before, there
are many variables that need to be taken into account
for pathogen risk assessment that are not currently
being addressed in this approach, such as age, severity
of response, or host immunity. Although, similar models
typically have been used in the quantisation of chemical
carcinogens, risks associated with pathogenic expo-
sures may be biologically more compatible with other
approaches for evaluating dose-response relationships
for pathogens. Therefore, model validation research will
be expanded to include a comprehensive evaluation of
several statistical models. Of particular concern are the
risks posed to children and other potentially susceptible
populations. Research efforts have been expanded to
address issues relating to quantitatively and qualitatively
measuring and characterizing risks to these popula-
tions. These approaches will allow for the development
of risk estimates following acute or chronic exposures
and address issues such as sensitivity, multiple end-
points and severity. The evaluation of these additional
approaches for pathogenic assessment are needed,
therefore, to augment current dose-response models for
quantifying risks.
Research Topics and Priorities
a. Modification of risk assessment paradigm for
characterizing microbial risks
RA.M.1—Development of a comprehensive
microbiology risk assessment model for wa-
ter (Ongoing) This project is a high priority
because EPA needs a suitable and comprehen-
sive microbiological risk assessment model for
water (drinking, recreational, shellfish, waste-
water, reuse) in order to establish or modify
drinking water and water quality standards for
pathogens or their indicators. This project will
identify the proper elements and processes
needed to perform microbial risk assessment.
Existing paradigms such as the National Acad-
emy of Sciences chemical risk assessment ap-
proach and the ecological risk assessment ap-
proach will be used as the starting point and
these will be modified and amended for micro-
bial risk assessment. The selected approach
will be validated with case studies.
Priority: High
3-12
-------
b. Accuracy of dose-response models in predict-
ing waterborne disease
The goal of this effort will be to develop both qualitative
and quantitative microbiological exposure response mod-
els, validate/ confirm input assumptions and character-
ize the uncertainties surrounding those assumptions.
Unlike previous studies, dose-response data will be
analyzed and evaluated using various statistical ap-
proaches such as using a logistic regression model
similar to that developed to characterize the risks from
non-cancer effects. These projects are high priority be-
cause models will address issues relating to the magni-
tude of response and accounts for severity of effects
and duration of exposures. These studies will enhance
current efforts and allow for better comparisons of ef-
fects other than infection, such as respiratory disease,
diabetes and immune deficiency. Of particular concern
are the risks posed to children and other potentially
susceptible populations. Research efforts have been
expanded to address issues relating to quantitatively
and qualitatively measuring and characterizing risks to
these populations.
RA.M.2—Evaluation/application of various dose-re-
sponse relationship models for quantitating
pathogenic risks
2a. Evaluation of various dose-response relation-
ship models
Conduct extensive literature search to evaluate ail avail-
able data relating to adverse effects and dose-response
for Giardia and Gryptosporidiumandi other pathogens of
concern. Research will then be conducted in two phases.
Initial efforts focus on the development and evaluation of
statistical approaches used to characterize the dose-,
response relationship for humans exposed to water-
borne pathogens. Phase 2 will include the analysis of
dose-response data from existing studies comparing the
various statistical approaches including logistic regres-
sion. Model application and severity of effects categori-
zation will be tested using data for bacteria, protozoans
and viruses. As appropriate, quantitative estimates will
be developed for establishing risk levels for specific
pathogens. This research will provide relevant informa-
tion on the virulence of risks associated with a specific
pathogen and individuals or populations at risk and
support development of regulatory guidelines/ criteria.
Priority: High
2b. Virulence factors and host susceptibility: im-
pact of waterborne pathogens on human sub-
populations
Under the FY 97 SDWAA initiative, additional resources.
have been identified for the expansion of research ef-
forts relating to pathogen risk assessment. Increased
efforts will include research to assess and quantify
effects of immunity of the exposed individual and popu-
lation; evaluation and quantification of the disease pro-
cess including secondary spread, magnitude and sever-
ity of effects in addition to infection, and the effects of
multiple exposures and routes of exposures.
c. Characterization of risks posed by exposure to
multiple or complex mixtures of pathogens
Similar to chemical exposures in drinking water, micro-
bial exposures occur in conjunction with chemical or
other pathogenic exposures by multiple routes. While
much has been done in the development of models for
estimating risks for chemical mixtures, very little infor- •
mation is available regarding the risks associated with
mixed or multiple pathogenic exposures. The overall
impact of microbial-environrnental interactions, micro-
,bial-chemical interactions, and microbial-microbial inter-
actions to changes in emergent populations on water
quality has not been thoroughly evaluated. Source wa-
ters containing acids, organochemicals, or heavy metals
can damage or kill biological waste treatment microbial
colonies and again may enhance growth of disinfectant
resistant pathogens. Such microorganisms may be able
to metabolize water disinfectants to reduce their effec-
tiveness. Microorganisms also can metabolize other
chemical contaminants to potentially toxic compounds,
as well as produce toxic compounds effective in patho-
genic interactions (enterotoxins, lysins, and hemolysins
to enhance invasiveness and tissue destruction). As-
sessment risks of waterborne toxins from algal blooms
or bacterial overgrowth can be predicted based on knowl-
edge of the organisms present in ground or surface
waters and in the distribution system, survival and growth
enhancing.mechanisms of nutrient abundance, and waste
runoff. The potential for co-infection between enteric
viruses and bacteria to enhance adverse health effects
will be explored. For most pathogenic agents, exposure
is expected to be extremely low or zero within the
system. The greatest potential for mixed or multiple
contaminant exposure occurs with bacteria due to their
capability for regrowth and biofilm development within
the distribution system. Interactions between microor-
ganisms during mixed infections have yet to be investi-
gated, (e.g., interactions, such as the reported ability of
Coxsackie virus to increase host susceptibility to Shi-
gella). Chemical, environmental, and microbial interac-
tions could contribute to low-level human and animal
population exposures, e.g., liver tumor, promoter toxins
(microcystin-LR). Microcystin-LR produced by
cyanobacteria during water blooms may contribute to
liver tumor promotion in the exposed populations. En-
demic disease of unknown etiology could be the result of
human exposure to low levels of waterborne microbial
toxins. Research is needed to address the impacts of
multiple exposures of pathogens or mixed exposures
(i.e., chemical and microbial exposure) on immune re-
sponse and effects levels. Development of a preliminary
assessment of, the existing available information and
framing of the scope and magnitude of this issue is also
needed.
RA.M.3—Evaluation and application of is-
sues and methods to assess risk associated
3-13
-------
with exposures to multiple pathogens, routes,
disease course and outcomes and duration
3a Feasibility assessment on interactions
Conduct a preliminary assessment and frame
issues related to mixed exposures of pathogens
in drinking water. This preliminary assessment
would include an evaluation of existing data
relating to mixtures and potential interactions,
and identify issues needed to be address and
research needs. Future research projects will
be developed and prioritized based on the re-
search issues and needs identified in the pre-
liminary assessment of mixtures.
Priority: Medium
3b Research would also be conducted to test
the application of dose-response models/ap-
proaches for identifying and estimating risks for
multiple pathogens. The emphasis of this effort
will be to address bacterial agents where the
greatest potential for mixed exposures may oc-
cur. A dose-response plane would be devel-
oped using individual dose-response curves from
various pathogens or chemicals and pathogens.
Validation studies would be conducted using an
appropriate animal for pathogenic response and
multiple exposures scenarios. Study design
would include sequential chemical and patho-
genic exposures as well as continuous expo-
sures at subthreshold doses. This research is
low priority because of the lack of general meth-
ods for assessing mixtures and the need to
develop single pathogen health and exposure
data.
Priority: Low
Risk Management Research
Disinfectants are used by virtually all surface water
systems in the U.S. and by an unknown percentage of
systems that rely on groundwater. Chlorine has been
the most widely used and most cost-effective disinfec-
tant. Disinfection is most efficient when it is applied as
part of the multiple barrier concept; that is, use of the
best available water source, protection of that source
from contamination, and use of water treatment to re-
move and inactivate pathogens. Recently there has
been growing recognition that water quality can deterio-
rate dramatically during distribution. Therefore, another
part of the barrier to infectious disease is a properly
designed and operated distribution system. Each com-
ponent of our research is designed to evaluate the
effectiveness of one of more of these barriers and will
collectively provide information on the combined effec-
tiveness of these barriers. s
While disinfection is an integral part of water treatment,
filtration may be necessary to reduce pathogen levels
and make disinfection more reliable by removing turbid-
ity and other interfering constituents. For example, in the
U. S. the waterborne disease outbreak rate for commu-
nities using surface sources without filtration is eightfold
greater than communities with filtered water systems.
Properly designed and operated water treatment sys-
tems that include filtration and disinfection can greatly
reduce the risk of waterborne disease. There is a need,
however, to find the proper balance between controlling
risks from microbial pathogens and disinfection by-prod-
ucts.
This section deals primarily with the treatment and distri-
bution system research needed to support regulation
development for control of microbial pathogens. This
research must be conducted as an integrated program
with the research described for controlling disinfection
by-products (see Chapter IV). Separate studies not dis-
cussed in this section but which are complementary to
this research deal with development of cost/benefit meth-
odology, which can be utilized to make effective risk
management decisions.
1. How effective are various treatment processes
in removing/inactivating pathogens?
State of the Science
The primary processes for removing pathogens from
drinking water are water treatment filtration systems and
disinfection. For filtration systems the specific sequence
and type of unit processes to be used are dependent on
source water quality. Optimized treatment trains will
lead to a low probability of pathogens entering the
distribution system. Breakdowns, such as in equipment
or lapses in operational control, can lead to significant
problems.
The disinfection process is the final treatment barrier for
minimizing pathogen transmission. In some cases, dis-
infection may be the only treatment barrier (e.g., a
disinfected groundwater supply). In either case, the
disinfection process must also be optimized to aid in
removing pathogens. Water plants must remove as
much particulate material (e.g., clays, microbes, etc.) as
possible in order to increase the effectiveness of disin-
fection.
EPA and others have been conducting research on
pathogen removal and inactivation of microorganisms
for many years, although the types of pathogens studied
and associated removal processes evaluated have var-
ied over this period. However, improvements in treat-
ment technology, pathogen detection and an increased
understanding of waterborne disease etiology have in-
creased the need for research in this area. More sophis-
ticated monitoring requirements have led to rethinking of
which treatment processes should be used and how
these processes should be operated. Most current re-
search on pathogen removal has focused and will con-
tinue to focus on Cryptosporidium oocyst removal and
inactivation because it is an excellent surrogate for
treatment efficiency. However, new pathogens such as
Microsporidia and Cyclospora cayetanensis are appear-
3-14
-------
ing on the horizon and may pose even greater chal-
lenges in the future. Therefore, as we make progress in
answering important questions regarding the treatment
of Cryptosporidium, future efforts will move toward ad-,
dressing the treatability of other waterborne pathogens
as their existence and significance is recognized. The
pathogens of greatest concern in groundwater sources
(especially the, enteric viruses) are different from those
occurring in surface sources and information on treat-
ment alternatives that are effective against viruses are
needed to identify appropriate treatment where needed.
EPA and the American Water Works Association Re-
search Foundation (AWWARF) are conducting the ma-
jority of the treatment research in the U.S.. EPA has
conducted filtration studies with both conventional and
alternative filtration processes for removal of pathogens.
Information gathered from those studies has led to
implementation of the Surface Water Treatment Rule
and has provided guidance to many water utilities.
AWWARF is currently funding research intended to
develop design and operational criteria for the optimiza-
tion of various treatment processes to remove proto-
zoan cysts, specifically Giardia and Cryptosporidium.
Parallel studies are being conducted to determine if a
relationship exists between cyst removal and the re-
moval of cyst-sized particles. AWWARF plans to con-
duct studies at utility-owned pilot- and full- sized facili-
ties. Research is being conducted to assess particle-
size analysis as a means of optimizing filtration and the
impact of sequential disinfection on the survival of
Cryptosporidium oocysts. This information will be used
in workshops and guidance documents to assist drink-
ing water utilities to more effectively operate drinking
water treatment plants. Surveys of water utility perfor-
mance and studies in detection methods have been
funded by AWWARF and are currently ongoing. Mont-
gomery Watson is conducting a disinfection study of
approximately 20 utilities; the study should define disin-
fection conditions necessary tor Cryptosporidium inacti-
vation using multiple sequential disinfectants. The Ameri-
can Water Works Association (AWWA) through its Wa-
ter Industry Technical Action Fund (WITAF), the Water
Research Center in the UK, and many others are also
studying pathogen removal/inactivation processes.
AWWARF has recently initiated studies to begin evalu-
ating the effectiveness of conventional and innovative
treatment processes for removing and inactivating se-
lected emerging waterborne pathogens.
EPA is conducting research on the use of slow sand
filters for removing oocysts, and conducting compara-
tive jar tests for evaluating removal rates of preserved
versus unpreserved oocysts. Pilot conventional and di-
rect filtration systems are being spiked-with fresh oo-
cysts to determine optimal removal conditions. The re-
moval of particles by conventional and enhanced filtra-
tion is being studied. Total particle counts, particle counts
in the oocyst size range and indigenous bacterial spores
are being studied as possible surrogates for oocysts.
There is evidence that oocysts may be damaged through
treatment and therefore susceptible to disinfection. It
has also been observed that 3-5 micron-sized oocysts
may be able to fold or squeeze through 3-micron mem-
branes under pressure. In addition, a,series of studies is
being conducted that evaluate the effectiveness of bio-
logical treatment plus final disinfection on the removal of
oocysts, spores, and DBP precursors. Several small-
system technologies such as diatomaceous earth and
'membrane package plants are being evaluated. EPA is
also conducting studies on cyst viability and on treat-
ment effectiveness under varying conditions, e.g., storm
events with high turbidity :and high pathogen loadings.
There are a number of questions to be answered with
regard to the removal/inactivation of pathogens from
drinking water. The Milwaukee experience reinforces
the need to find answers to these questions which
include a more complete understanding of the optimiza-
tion of filtration for removing pathogens, the role of
disinfection, and the types of technology that might be
useful for small utilities.
Results from research to date indicate that there is much
to be learned regarding the conditions that lead to
optimal operation of drinking water treatment plants for
the removal of oocysts and other pathogens. For ex-
ample, it has become apparent that oocysts have physi-
cal properties that make them both difficult to remove
and very resistant to disinfection. Research has shown
that the organism is flexible and that it can penetrate
porous media. In contrast, it is also sticky so that it
adheres to unit process walls and piping making re-
moval difficult to calculate. It has a surface charge that is
different from particles normally found in water, so that
standard coagulation procedures may not.adequately
remove' oocysts during conventional treatment (coagu-
lation, settling, filtration and disinfection). Results from
EPA studies have shown that preserved oocysts which
are normally used in treatment research and are most
convenient for spiking studies do not behave in the
same way as natural oocysts.; Therefore, more studies
on treatment effectiveness need to be conducted using
natural oocysts. Because oocysts are pathogens and
cannot be spiked into operational treatment plants, there
is need for the development of surrogates which are
nonpathogens but behave in the same manner as oo-
cysts. Small systems present a special problem be-
cause the technologies used must be easy to operate
and have low maintenance requirements. However, as
has been discussed, removal of oocysts requires careful
attention to operational detail. This was one of the major
lessons learned from Milwaukee.
Bench-scale, pilot plant and field-scale studies will be
required to address these research issues. Initially, tech-
nology evaluations will be conducted at bench and pilot
scale. This will be followed by field studies (RM.M.5) of
the most promising technologies which will evaluate
their ability to control both pathogens and • DBPs. In
general/these field studies will be conducted at sites
where these technologies are being utilized by certain
utilities such that evaluations of the important operating
parameters can be conducted and results can be com-
3-15
-------
pared with full-scale operations. New techniques for
monitoring treatment plant effluents and for process
control will also be investigated. Improved electronic
measuring techniques could allow for real-time mea-
surement of treated water quality. All future projects will
incorporate multiple pathogen removal, surrogate analy-
sis and DBF reduction efforts as part of the research *
protocol.
Research Topics and Priorities
a. Optimization of conventional treatment pro-
cesses to remove pathogens
All of the projects listed below except RM.M.28 are
considered high priority and were selected for consider-
ation because they address the major scientific un-
knowns: removal of oocysts via conventional treatment,
development of surrogates for oocyst removal, and vi-
ability of oocysts through multiple treatment steps. Re-
sults of these studies will provide information necessary
for setting feasible risk based treatment levels. They
also provide the data to develop practical performance
monitoring procedures using surrogates and to deter-
mine whether compliance is achieved for meeting level
of treatment requirements. Projects 2, 3 and 4 expand
on current promising research efforts by EPA. RM.M.28
is considered medium priority at this time until the exist-
ence and significance of specific emerging pathogens is
better defined.
RM.M.1—Filtration studies for controlling
pathogens (Ongoing) Bench-scale and pilot-
scale studies have been conducted to evaluate
pathogen removal mechanisms. Current stud-
ies are attempting to find surrogates for patho-
gens or surrogates for treatment evaluation.
Particle counting, zeta potential, microbial count-
ing and other techniques are used to determine
removal of pathogens. Criteria for measuring
damage to oocysts are also being developed.
Priority: High
RM.M.2—Filtration removal of protozoa and
indicators Conduct pilot-scale studies to as-
sess conventional water treatment and direct
filtration for removal of protozoans and indica-
tors.
Priority: High
RM.M.3—Optimize conventional treatment for
removal of oocysts Conduct bench studies to
evaluate and demonstrate that, if treatment (co-
agulation/settling) is optimized for turbidity/spore/
particulate reduction, coincidental removal of
oocysts occurs. Determine if there are some
coagulants that are selective for removing oo-
cysts.
Priority: High
RM.M.4—Filtration damage viability studies
Conduct bench and pilot studies to assess the
capacity of oocysts to remain viable through
various types of treatment. Conduct disinfection
studies to determine if oocysts that are dam-
aged by passing through treatment are more
easily inactivated than undamaged oocysts.
Priority: High
RM.M.5—Evaluate disinfection and optimi-
zation in field-scale treatment plants Con-
duct field-scale studies of pathogen and DBP
control at existing water treatment plants. Use
oocysts, spores, particle and turbidity analysis
as part of the evaluation process. (This is a
component of Project EX.M.11)
Priority: High
RM.M.28—Evaluation of the effectiveness of
conventional treatment processes on remov-
ing and inactivating emerging pathogens As
the existence and significance of new, emerging
pathogens are identified by NERL and NHEERL
researchers, NRMRL will initiate research on
the effectiveness of conventional treatment to
remove and inactivate these pathogens.
Priority: Medium
b. Effectiveness of different filtration processes in
removing pathogens
The research described in these projects repre-
sent variations of conventional treatment. Bio-
logical treatment is rated "high" because it has
the potential for controlling DBPs and patho-
gens simultaneously. Task RM.M.7 is rated "me-
dium" because the technologies are expected to
have similar results for Cryptosporidium as for
Giardia, on which a number of studies have
already been conducted. RM.M.29 is consid-
ered medium priority at this time until the exist-
ence and significance of specific emerging patho-
gens is better defined.
RM.M.6—Biological treatment for control of
oocysts
Conduct studies on the effectiveness of biologi-
cal treatment for control of oocysts while simul-
taneously controlling disinfection by-products.
Priority: High
RM.M.7—Filtration techniques other than
conventional treatment Examine other filtra-
tion techniques such as diatomaceous earth
filtration and slow sand filtration to determine
how Cryptosporidium removals compare with
those of Giardia (already determined from previ-
ous studies).
Priority: Medium
3-16
-------
RM.M.29—Evaluation of different filtration
processes on removing emerging pathogens
As the existence and significance of new, emerg-
ing pathogens are identified by NERL and
NHEERL researchers, NRMRL will initiate re-
search on the effectiveness of innovative filtra-
tion processes to remove these pathogens.
Priority: Medium
(See also project RM.M.14 below)
c. Effectiveness of disinfection processes in inac-
tivating pathogens
Research is needed on the disinfection effectiveness of
various disinfectants for viruses and protozoa. For ex-
ample, some research suggests that reductions in viable
Cryptosporidium oocysts are achieved after sequential
disinfection using different disinfectants (i.e. chlorine/
chloramine, ozone/chloramine). EPA uses disinfection
data on hepatitis A virus (HAV) for estimating conditions
necessary to inactivate viruses because of its relatively
high resistance to disinfection. Norwalk virus, which has
also caused waterborne disease outbreaks, may be
more resistant to disinfection than HAV and therefore
may be more suitable than HAV for defining adequacy of
disinfection.
Establishing whether Norwalk virus is more resistant to
disinfection than HAV is particularly important for ground-
water systems. However, it is also relevant for surface
water systems which use physical means for removal of
protozoa and must provide a minimum disinfectant dose
to assure adequate virus removal/ inactivation. Because
of its potential to reduce health risk, project RM.M.8 was
assigned a high priority. A Norwalk virus project on UV
inactivation was also assigned a high priority because of
its potential usefulness for disinfecting groundwater.
(Norwalk virus was not included in the AWWARF UV
virus' inactivation study.) In addition, systems using
groundwaters will need information on other treatment
alternatives that are effective against viruses, as this is
an important consideration of deciding on groundwater
disinfection methods. Therefore, RM.M.30 is a high
priority project. RM.M.31 is considered medium priority
at this time until the existence and significance of spe-
cific emerging pathogens is better defined.
Project RM.M.10 on sequential disinfection treatment,
which will expand on AWWARF's and Montgomery
Watson's research, was assigned a high priority be-
cause it is expected to develop criteria for more efficient
and cost-effective disinfection strategies. This additional
research is needed to develop a generic model concept
by which utilities with different operating conditions could
estimate inactivation efficiencies.
RM.M.8—Control of Norwalk virus by chlorine and
ozone
Compare kinetics of Norwalk virus inactivation
to that of poliovirus 1 and MS-2 coliphage, and
determine the extent of removal by pretreat-
ment and filtration and the extent of inactivation
with chlorine or ozone.
Priority: High
RM.M.9—UV disinfection efficiencies for
Norwalk viruses
Evaluate efficiency of ultraviolet light for inacti-
vation of Norwalk virus, and determine UV in-
tensity and exposure time to achieve 1,2,3 and
4 Iog10 inactivation.
Priority: High
RM.M.10—Inactivation of Giardia and
Cryptosporidiumby sequential disinfectants
Define the treatment conditions (concentration,
contact time, temperature) needed to inactivate
Giardia and Cryptosporidium with different se-
quential disinfectants.
Priority: High
RM.M.30—Evaluation of alternative inactiva-
tion processes for controlling viruses Evalu-
ate other innovative advanced oxidation pro-
cesses and removal systems such as ultrafiltra-
tion membranes, mixed oxidant generators, TiO2
and UV and combinations of these processes
for removing and inactivating viruses. This project
is closely related to RM.D.7.
Priority: High
RM.M.31—Evaluation of different disinfec-
tion processes in inactivating emerging
pathogens As the existence and significance of
new, emerging pathogens are identified by NERL
and NHEERL researchers, NRMRL will initiate
research on the effectiveness of innovative dis-
infection processes to inactivate these patho-
gens.
Priority: Medium
Other projects that also address this topic:
RM.M.3—Optimize conventional treatment for
removal of oocysts
RM.M.5—Evaluate disinfection and optimization
in full-scale treatment plants
RM.D.2—Effects of ozone and biofiltration for
control of precursors, pathogens and for pesti-
cide removal
RM.D.7—Membranes/advanced oxidation and
other technology combinations
3-17
-------
d. Appropriateness and costs of technologies for
small systems
Any changes in regulations requiring small systems to
upgrade treatment for Giardia or Cryptosporidium would
be a major challenge because of their limited financial
resources for acquiring and operating drinking water
treatment plants. In addition, most utilities have difficulty
in finding and retaining qualified operators. In setting
any new regulatory criteria for small systems, EPA
needs information on the effectiveness, costs and feasi-
bility of appropriate treatment technologies. Therefore,
research projects to develop and demonstrate small-
scale, cost-effective treatment technologies which are
easily installed and automated, and which do not require
highly skilled operators are a high priority. Developing
cost curves will be an important element of all small-
system technology projects.
RM.M.11—Cryptosporidium removal using
bag filters Identify vessel, bag, and basket
design characteristics for optimal removal and
treatment train requirements relative to variable
raw water characteristics. Operational factors
relative to cost and performance should also be
considered.
Priority: High
RM.M.12—Cost-effectiveness of prefiltration
for ultrafiltration unit Determine range of raw
water characteristics relative to various pre-treat-
ment technologies used to increase membrane
life, reduce maintenance costs, and enhance
DBP precursor removal. This is cross-referenced
with RM.M.17, RM.M.18, and RM.M.20.
Priority: High
RM.M.13—Development/testing of innovative
technologies for small systemsConduct stud-
ies on new synthetic resins, pulsed UV, electron
beam irradiation, high-pressure distillation, non-
porous hollow fiber membranes, and cartridge
filtration. In addition, conduct verification testing
of package drinking water systems for use by
small communities. This work should be coordi-
nated with RM.M.4, RM.M.10, RM.M.30 and
RM.D.7.
Priority: High
2. How can the quality of treated water be main-
tained in distribution systems?
State of the Science
While much waterborne disease is associated with con-
taminated water sources and inadequate water treat-
ment, protection of water quality during distribution to
the customer is often neglected. Continuous, adequate
water pressure and residual level of disinfectant are
fmportant, especially in areas lacking adequate sewer-
age and sewage disposal facilities. Sewage and con-
taminated surface or groundwater can enter the water
system through cross-connections and broken or leaky
water pipes in older and poorly maintained water distri-
bution systems. In the U.S., 24% of the waterborne
outbreaks reported in community water systems over
the past decade were caused by contamination entering
the water distribution system, i.e., not originating from
poorly treated water. To lower the risks of waterborne
disease, contamination of water distribution systems
must also be prevented.
Maintenance of a disinfectant residual is required to
provide an additional barrier of protection from patho-
gens that might penetrate the water distribution system;
however, the effectiveness of this residual to protect
against significant contamination through a cross-con-
nection or infiltration is uncertain. The real value of
maintaining a disinfectant residual in the distribution
system is to help identify the occurrence of such con-
tamination by monitoring for the loss of the residual.
Although the behavior of some constituents in the distri-
bution system, namely chlorine, has been well studied,
the fate of others is less known. These include alterna-
tive disinfectants (such as chloramines), biologically as-
similable organic carbon, particulate material (including
microorganisms), and genotoxicity. Better characteriza-
tion of water quality problems is needed particularly at
dead-end branches where stagnant flow conditions pre-
vail.
The survival and growth of bacteria, such as Legionella,
Mycobacterium, and Aeromonas, under limited nutrient
conditions of water distribution systems, is also of con-
cern because of potential risks that may be associated
with inhalation, ingestion, or dermal exposure. The main-
tenance of a disinfectant residual is not always effective
in preventing the survival and growth of bacteria within a
distribution system, and defining criteria for biologically
stable drinking water is important to prevent potential
risks of infection from these or other microorganisms.
Research has shown that Legionella proliferation can be
controlled in potable water by interfering with the
Legionella-protozoan receptor site interaction.
Distribution systems represent the final link in the chain
between raw source water, treatment facilities and the
consumer. However they are generally designed and
operated to satisfy hydraulic reliability objectives—pro-
viding adequate water quantity and pressure for fire
flow, and domestic, commercial and industrial demands—
rather than to maintain water quality. This frequently
results in large service mains, dead-end branches, and
large storage facilities which keep water in the system
for long periods of time, leading to degradation of water
quality.
Individual distribution systems contain hundreds and
even thousands of miles of pipe. As water flows through
this pipe and sits in storage reservoirs, reactions can
take place between constituents within the water itself
and with materials along the wall of the pipe and reser-
voir. The distribution system is thus a giant chemical
3-18
-------
reactor, with residence times far in excess of those seen
within a treatment plant.
The complex, looped nature of the piping network, in
some systems, the presence of multiple sources of
water feeding the system, the fill-and-draw operation of
storage facilities and variable water usage rates at dif-
ferent locations and times of the day create very com-
plex flow patterns. These patterns often defy any intui-
tive sense of where water is traveling at any given point
in time. Such flow patterns, combined with time spent in
storage, contribute to residence times that can exceed
several days. The result, as verified by field surveys, is
that water which left a treatment plant with a very
uniform level of quality, can exhibit a highly variable
pattern of water age and quality in both time and space
throughout a distribution system. Some of the negative
impacts of time spent by water in a distribution system
include loss of disinfectant residual, growth of disinfec-
tion by-products, growth of biofilm colonies along pipe
walls, and the protection and subsequent release of
nuisance and pathogenic organisms from the biofilm
overtime.
External contamination of water flowing in the distribu-
tion system is an ever-present threat to public health
and has been shown to be the most frequent cause of
waterborne disease outbreaks. There is growing evi-
dence that endemic and epidemic waterborne disease is
associated with breaks in the integrity of distribution
lines, cross-connections and other breaches of distribu-
tion system integrity. Research is needed to improve the
design, construction, rehabilitation, operation, and main-
tenance of distribution system integrity. As pipe ages
and corrodes it exerts a higher chlorine demand and
provides more environmental niches for biofilm protec-
tion and proliferation. Many water distribution systems in
this country are approaching 100 years old. An esti-
mated 26% of distribution system pipe is unlined cast
iron and steel and is in poor condition. At current re-
placement rates for distribution system components, a
utility will replace a pipe every 200 years.
The volume available in storage facilities is often several
times that of the pipes in the distribution system/The
degree to which water is well mixed within these facili-
ties is largely unknown, as is the possible impacts of
discharging extremely old water from unmixed zones
under conditions of high demand. Guidelines on how to
design and operate storage facilities to promote better
mixing are needed.
Most medium- and large-scale utilities employ sophisti-
cated data acquisition systems for monitoring and con-
trolling the hydraulic performance of their distribution
systems. The methods and benefits of extending these
systems to include water quality parameters should be
studied. There is currently no consensus in the water
industry on the relative effectiveness of pipe cleaning
versus pipe repair or replacement for enhancing water
quality. The same can be said for identifying effective
institutional programs for implementing cross-connec-
tion and backfiow prevention programs.
For the past decade or more, extensive field studies
have been conducted characterizing the chemical
changes occurring in drinking water in distribution sys-
tems and identifying biofilm growth along pipe walls.
EPA has developed a sophisticated computer model
(EPANET) that tracks the fate of chemical species in
complex pipe networks and is currently used by utilities
and engineers throughout the world. EPA has also
developed methods for determining organic carbon avail-
able to support biofilm growth and has developed guide-
books for controlling such growth in distribution sys-
tems.
Microorganisms are known to colonize the walls of pipe
in the form of thin biofilms. The negative impacts of such
biogrowth, particularly with regard to harboring and breed-
ing pathogens, requires more study. Research is also
needed to determine the proper combination of nutrient
reduction and disinfectant level to control biofilm growth
under site-specific conditions.
The AWWA Research Foundation has supported stud-
ies examining the kinetics of chlorine consumption in
distribution systems and the factors limiting microbial
growth. The National Water Research Institute is fund-
ing research to understand the microbial interactions
within biofilms. French researchers at N.A.N.C.I.E. have
been operating a full-scale distribution system simulator
for several years. They have quantified the effects of
pipe material, biological seed concentration and disin-
.fectant choice on biofilm formation. Researchers at
Lyonnaise des Eau have studied the factors contributing
to chlorine demand from the pipe wall and have devel-
oped a chlorine microsensor that can be placed at the
pipe wall. At least one water authority in the UK is
experimenting with on-line sensing of water quality con-
ditions within distribution systems and is developing a
biofilm monitor.
Research studies indicate the need for more accurate
kinetic models that can simulate the loss of chlorine, the
formation of disinfection by-products.and the growth of
microorganisms in distribution systems. Research is
also needed to identify the factors that affect the growth
of biofilms and the survival of opportunistic pathogens in
distribution systems. There is also strong evidence of
the interactions between treatment and regrowth of bac-
teria in the distribution system. These studies indicate
the need to conduct research on the potential for modify-
ing treatment to control the formation of biofilm and the
survival of pathogens.
Research Topics and Priorities
Priorities were assigned on the following basis: Due to
the growing evidence that endemic and epidemic water-
borne disease is associated with breaks in the integrity
of distribution lines, cross-connections and other
breaches of the distribution system integrity, all projects
3-19
-------
addressing the cause of microbial intrusion into the
distribution system and corrective actions are rated high
priority. Until it has been established that heterotrophic
bacteria pose a health risk and that present disinfection
practices may be inadequate to prevent an unaccept-
able risk, projects proposed which are related to micro-
bial quality are assigned a medium priority. The only
other distribution system projects in this section of the
plan which were considered a high priority are those
which will provide information for disinfection by-product
risk assessment modeling, i.e. the kinetic models for
chlorine decay, DBF formation and the enhancement of
the EPANET model.
a. Use of coliforms (or other surrogates) to indi-
cate adequate control of primary and opportu-
nistic pathogens in biofilms and pipe sediments
Coliforms are used to determine the sanitary quality of
drinking water. If opportunistic pathogens are found to
pose significant risks, additional research in this area
will be needed to assess the use of coliforms as surro-
gates.
RM.M.14—Bacteria interference with detec-
tion of coliforms and E. coll The presence of
high densities of bacteria is known to interfere
with assays for total coliforms, the current basis
for regulating the microbial quality of drinking
water. Research is needed on issues such as
interference with the presence-absence ap-
proach to coliform monitoring.
Priority: Medium
b. Water quality factors affecting biofilm growth
and the effectiveness of disinfectant residuals in
controlling growth
RM.M.15—"Kinetic Models for chlorine de-
cay in distribution systems Using laboratory
and field studies, alternative kinetic models for
chlorine decay in distribution systems are being
evaluated. Reactions of chlorine at the pipe wall
can be very important in some systems and can
be correlated to hydraulic roughness coefficients.
This work is being performed in association with
AWWARF, Montgomery Watson, and Lyonnaise
de Eau.
Priority: Completed
RM.M.16—Enhancement to the EPANET dis-
tribution system water quality model (Ongo-
ing) A new version of EPANET, a program that
models water quality fate and transport in distri-
bution systems, is being developed. It incorpo-
rates new kinetic models for chlorine decay and
THM formation, variable geometry storage tanks
with incomplete mixing, and more flexible op-
erational rules for specifying system operation.
Priority: High
RM.M.17—Preliminary studies of biofilm for-
mation rates in pilot-scale distribution sys-
tems (Ongoing) Study of the factors affecting
the rate of biofilm growth on pipe walls using a
pilot-scale distribution system is beginning. The
pilot system consists of multiple recirculating
loops of 6" pipe with removable biofilm coupons
inserted in the wall. Initial experiments will com-
pare biofilm growth rates under different chlo-
rine levels. This will expand to compare rates
under alternative disinfectants and nutrient con-
trol strategies.
Priority: Medium
RM.M. 18—Opportunistic pathogens in
biofilms Identify the prevalence of opportunistic
pathogen growth in biofilms and develop mitiga-
tion measures.
Priority: Medium
RM.M.19—Impact of nutrient removal on
growth potential for bacteria Pilot studies are
being conducted on removal of nutrients from
treated drinking water. Pilot-scale pipe loop stud-
ies will be conducted to evaluate growth poten-
tial for bacteria in distribution systems.
Priority: Medium
RM.M.20—Impact of alternative treatment on
biofilm growth Assess the potential of alterna-
tive treatment systems to promote biofilm growth
in the distribution system, using a pilot-scale
distribution system.
Priority: Medium
RM.M.21—Water quality factors in distribu-
tion systems (Ongoing) Studies are underway
to investigate the importance of biofilm develop-
ment and microbial interactions on the benefi-
cial and detrimental processes in drinking water
distribution systems. One aspect of this research
addresses the interactions between pipe mate-
rials, corrosion inhibitors, disinfectants, organ-
ics, and distribution biofilms. A second effort
addresses on-line monitoring of pathogen ecol-
ogy for quantitative evaluation of mitigation pro-
cedures.
Priority: Medium
See also RM.M.26—Bacterial growth in distri-
bution systems
c. Effect of design and condition of the distribution
system on bacterial growth
Additional characterization studies are needed to iden-
tify points in the distribution system that are particularly
vulnerable to degradation of water quality. The first
candidates for study are dead-end locations, for which
there is a dearth of long-term data. Studies have shown
3-20
-------
that up to 25% of drinking water consumers drink their
water from dead-end sections.of the distribution system.
Additional knowledge is needed on the role that storage
facilities play in influencing water quality changes. The
long detention times in such facilities can reduce disin-
fectant residual but might also reduce the nutrients that
promote biofilm growth. Further study is needed on how
to design a tank to promote uniform mixing within it.
Guidelines on size and location of influent and effluent
lines to accomplish this goal are needed.
Computer modeling needs to be enhanced to accommo-
date the effect that mixtures of treated source waters
have on kinetic coefficients that model the decay of
chlorine or growth of by-products. More fundamental
kinetic models of by-product formation need to be devel-
oped that take into account the concurrent loss of disin-
fectant and production of multiple species of by-prod-
ucts. The same can be said for modeling the complex
chemistry of chloramines when used as a disinfectant.
Improved information management can help shift the
emphasis in distribution system operation towards a
better blend of public health and hydraulic reliability
objectives. Advances in remote sensing and data acqui-
sition technologies applied to the specific needs of the
water utility industry can make system operators more
aware of current water quality conditions within their
system and help them make short-term decisions to
improve its quality (e.g., disinfectant doses, tank fill/
release rates, and pump control strategies).
RM.M.22—Water quality impacts of dead ends
Examine the water quality impacts of dead ends.
Conduct continuous, long-term monitoring of
water quality conditions such as coliform, chlo-
rine residuals and corrosion by-products at dead-
end branches and loops of a distribution sys-
tem.
Priority: Medium
RM.M.23—Mixing in storage facilities Study
mixing in storage facilities, including the factors
affecting mixing and residence times in distribu-
tion system storage facilities. Develop guide-
lines for maintaining well-mixed conditions.
Priority: Medium
RM.M.24—Alternative kinetic models for de-
cay and DBP formation Develop multi-compo-
nent kinetic models for disinfectant residual de-
cay and formation of selected DBPs in distribu-
tion systems. .
Priority: High
RM.M.25—Real-time monitoring systems
Study real-time data acquisition and control.
Demonstrate monitoring of water quality vari-
ables in real-time and use to modify distribution
system operation to maintain good water qual-
ity.
Priority: Medium
Other projects which also address this topic:
RM.M.30 Opportunistic pathogens in biofiims
d. Effect of filtration/disinfection treatment pro-
cesses on chemical and biological stability of
water in distribution systems
Future research will continue to emphasize the role that
biofilm development along pipe walls plays in posing
risks to public health. Both a better biological and chemi-
cal engineering understanding of the structure and growth
dynamics of biofiims are needed in order to develop
effective control programs, some of which may involve
treatment decisions at the water treatment plant. In
particular, the degree to which treatment technologies
minimize nutrient availability in the treated water will
affect subsequent growth of bacteria in the distribution
system.
RM.M.26—Bacterial growth in distribution
systems Quantify the factors affecting the growth
of biofilm on pipes, and identify effective control
strategies.
Priority: High
RM.M.27—Integrated approaches for control-
ling pathogens Characterize the interactions
among treatments on the microbial quality of
water in distribution systems. Interactions of
concern include effects of corrosion control, loss
of residual disinfectants and GAG for DBP con-
trol on microbial quality.
Priority: Medium
e. Maintaining distribution system integrity
Microbial pathogens and other contaminants can enter
the distribution system in many ways including: piping
failure, construction activities, repair and maintenance
operations, back pressure, sewage cross-connections,
poor system design, and poor operational practices. As
we move toward the 21st century, greater attention will
have to be given to design, operation and maintenance
of distribution systems to ensure water quality as well as
hydraulic reliability. This will include consideration of
advanced materials of construction as well as installa-
tion of sensors for real-time monitoring of important
distribution system quality indicators such as disinfec-
tant residuals, water pressure, flow direction, microbial
densities, total organic halides, and other quality param-
eters.
RM.M.32—Evaluate the primary causes of
leaks and failures in water distribution sys-
tems There is increasing concern over the po-
tential deterioration of drinking water distribution
system components and evidence exists that,
especially in the largest urban areas in the U.S.,
3-21
-------
this deterioration is well underway. Recent boil-
water orders and MCL violations in New York
City and Washington D.C. support this concern.
This task will support a coordinated effort with
organizations such as the American Waterworks
Association Research Foundation, the Ameri-
can Public Works Association, the Civil Engi-
neering Research Foundation, the Water Envi-
ronment Federation Research Foundation, the
Cross Connection Control Association, and the
Association of State Drinking Water Administra-
tors to determine the primary causes of leaks in
distribution systems.
Priority: High
RM.M.33—Evaluate state-of-the-art devices
for determining the structural integrity of
water distribution systems Research will be
conducted on state-of-the-art technology that
can be used to establish the structural integrity
and reliability of drinking water distribution sys-
tem components. For example, sensors that
can be used to provide acoustic leak detection
capability or that can be imbedded into system
components to predict component failure and to
establish structural integrity will be investigated.
Remote monitoring techniques and telemetry
will be investigated to measure real-time water
quality in distribution systems and indicate in-
tegrity and reliability of distribution system com-
ponents.
Priority: High
RM.M.34—Evaluate materials for construct-
ing distribution system components Many
materials in common use in drinking water dis-
tribution systems are subject to corrosion and
failure under stress. These failures can range
from leaks to catastrophic failure. This task will
examine the potential for the use of new con-
struction material that can resist the common
mechanisms of failure in pipelines. For example,
stainless steel is now being evaluated for use in
drinking water systems in Japan. Stainless steel
is resistant to corrosion, is flexible and has a
long structural life.
Priority: High
3. How can source water be protected to ensure
that it is consistent with finished water quality of
acceptable microbial risk after appropriate treat-
ment?
State of the Science
Passage of the 1996 Amendments of the Safe Drinking
Water Act has focused the attention of water utility
mangers and public health and regulatory officials on
source water protection and its role in protecting public
water supplies. There is growing awareness that water
treatment and /or disinfection may not always be ad-
equate to ensure the provision of potable and safe water
to the consumer. The cryptosporidiosis outbreak in Mil-
waukee, Wisconsin, has raised the possibility that even
water suppliers which meet all of the Surface Water
Treatment Rule (SWTR) requirements are vulnerable. In
the Milwaukee case more then 400,000 illnesses and as
many as 100 deaths were associated with the outbreak.
The cause of the outbreak was attributed to a combina-
tion of poor treatment operations and a sudden increase
in source water turbidity resulting from a wet weather
flow event. This and other similar but less dramatic
waterborne outbreaks have spurred an interest in as-
sessing the occurrence of protozoan cysts in source and
treated water. This research activity is closely related to
ongoing research being conducted under NRMRL's Wet
Weather Flow (WWF) Program. The WWF Program is
examining techniques for characterizing WWF events,
and for development of Best Management Practices for
controlling the impact of WWF events on water quality in
receiving streams. For example these studies will char-
acterize the factors that influence the overland flow of
pathogens such as Cryptosporidium, and evaluate the
effectiveness of end-of-the-pipe control technologies and
on-site waste water treatment technologies. This project
will also utilize results from projects currently being
funded, by the AWWARF and WERF.
Research Topics and Priorities
a. The impact of source water protection on the
performance of drinking water treatment sys-
tems
Areas of potentially productive risk management re-
search are as follows: development and calibration of
Geographic Information System (GIS) techniques for
describing source water characteristics; development,
application and calibration of new and existing water
quality models (BASINS); and defining the factors that
influence the performance of conventional treatment
when concentration of contaminants in source water
increases rapidly. Given the current emphasis on source
water protection in the 1996 Amendments to the Safe
Drinking Water Act, all of these activities are categorized
as high-priority research.
Task RM.M.35—Evaluate Geographic Infor-
mation Systems (GIS) techniques for defin-
ing source water characteristics GIS tech-
niques have become widely used for integrating
and displaying land use data for watersheds
and river basins. This tool is very effective for
visually displaying and integrating various types
of data bases that can lend insight into the
factors that influence watershed functions. For
example, GIS have been used in the Ohio River
Basin to show the relative location of water
supply intakes, dams and waste water and in-
dustrial discharges on the Ohio River using
EPA's REACH file as an integrator. However,
little has been done with integrating GIS and
functional models to predict the behavior of the
3-22
-------
watershed and its impact on water quality. For
example GIS, soil moisture vegetative cover
and atmospheric models can be used to predict
the types and severity of wet weather flow events
that may be expected to take place in a given
type of watershed. This information can be used
to understand and characterize the effects of
point and non-point source run off on groundwa-
ter surface water interactions and on surface
and groundwater quality. There are many exist-
ing data bases available through agencies such
as the Corps of Engineers, the USGS and the
USEPA that can be utilized to provide basic
data input for this type of analysis. This task will
evaluate existing data bases and models that
can be integrated with GIS to provide informa-
tion that could be used to define the characteris-
tics of watersheds and show how these charac-
teristics influence water quality in rivers, streams
and the subsurface.
Priority: High
Task RM.M.36—Evaluate water quality mod-
els for routing point and non-point source
water discharges Water quality models can be
used to provide the linkage between wet weather
flow events and water quality in drinking water
sources. The USEPA, the COE, along with other
government agencies, academia and the pri-
vate sector have invested a great many re-
sources in developing and distributing hydraulic
and water quality models for use in river basins
and watersheds. This task will examine, and
evaluate these models and make recommenda-
tion as to which are most appropriate for use in
specific types of circumstances. Issues related
to field calibration, types of input data required,
and ease of use will be investigated. This project
will also address the relationship between best
management practices as evaluated in our Wet
Weather Flow program and their impact on pro-
tecting source waters for drinking water pur-
poses.
Priority: High
Task RM.M.37—Evaluate the impact of sud-
den increases in source water contaminant
concentration on drinking water treatment
There is increasing evidence that drinking water
unit processes are vulnerable to sudden in-
creases in contaminant concentration in raw
water. The Milwaukee cryptosporidiosis outbreak
was in part due to the effect of a wet weather
flow event that resulted in sudden and unex-
pected increases in turbidity in the source wa-
ter. This wet weather flow event in combination
with poor treatment operations resulted in the
largest recorded waterborne disease out break
in the history of the U.S. There is other, al-
though less dramatic, information that, sudden
changes in contaminant concentration will de-
grade the effluent quality of treated water. This
task will investigate the design and operating
variables that influence water treatment sys-
tems. Research will be conducted to define the
conditions that cause water treatment systems
to fail under the influence of rapidly increasing
raw water contaminant concentrations.
Priority: High
Table 111-2. Research Priorities for Health Effects of Microbial
Pathogens
Research Topics ,
a. Pathobiology of infection and disease for the
most important waterborne pathogens
HE.M.1—Infectious dose of Cryptosporidium High
HE.M.2—Validity of d-r model for
Cryptosporidium High
HE.M.3—Cryptosporidium virulence study High
H E.M.4—Cryptosporidium infectious
dose/immunity High
HE.M.5—Infectious dose of Norwalk virus High
HE.M.6—Infectious dose of other pathogens Medium
b. Characterization of epidemic and endemic
waterborne disease
HE.M.7—Characterization of endemic disease High
HE.M.8—Immunological assays for use in
epidemiology studies High .
HE.M.9—Investigation of waterborne outbreaks High
HE.M.10—Surveillance tools for outbreaks High
3-23
-------
Table 111-3. Research Priorities for Exposure to Microbial Pathogens
Research Topics
Proposed Projects
Priorities
Microbial Methods
a. Methods for detecting and enumerating Cryptosporidium and
Glardia (including identifying the viability potential) in source and
finished drinking water, and methods for other emerging protozoa
b. Methods for detecting and enumerating viruses in source and
finished drinking waters
Microbial Exposure
a. Surveys to determine pathogen occurrence in source and finished
waters
b. Importance of watershed control (including point source,
non-point source, and septic tank controls) for source water
pathogen occurrence
c. Occurrence of and exposure to primary and opportunistic
pathogens in distribution systems
Microbes In Groundwater
a. Survival and transport of pathogens in the subsurface
b. Methods for protecting wells and springs from pathogens
EX.M.1—Immunologica) techniques for protozoa High
EX.M.2—Gene probes for detection of viable High
Cryptosporidium oocysts
EX.M.3—Cultural method for Cryptosporidium High
in environmental samples
EX.M.4—PCR methods for Giardia and Medium
Cryptosporidium
EX.M.5—Protozoa methodology protocol High
development workshop
EX.M.6—Comparison of methods for Giardia High
and Cryptosporidium in water
EX.M.7—New protozoa agents High
EX.M.8—Application of PCR technologies Medium
and gene probes for virus detection in water
EX.M.9—Norwalk virus High
EX.M.10—Methods for emerging viruses High
EX.M.11—Intensive evaluation of microbiological High
constituents and treatability in surface source
waters
EX.M.12—Identification of viruses resistant to Medium
disinfection
EX.M.13—Distinguish animal versus human sources High
EX.M.14—Occurrence of Mycobacterium High
EX.M.15—Occurrence of heterotrophic bacteria High
with virulence characteristics
EX.M.16—PCR method for Legionella Low
EX.M.17—Pathogenicity of heterotrophic High
bacteria found in drinking water
EX.M.18—Occurrence of opportunistic Medium
pathogens in biofilms
EX.M.19—Opportunistic pathogens associated Medium
with point-of-use (POU) and point-of-entry (POE)
filter effluents
EX.M.20—Potential pathogenicity of heterotrophic Medium
bacteria from POU filters
EX.M.21—Occurrence of newly emerging High
pathogens
EX.M.22—Exposure as a function of population High
distribution
EX.M.23—Virus survival in the subsurface High
EX.M.24—Virus transport in the subsurface High
EX.M.25—Viral transport and fate models Medium
EX.M.26—Vulnerability of groundwater to pathogens High
EX.M.27—Delineation of natural protection zones Medium
EX.M.28—Vulnerability & sensitivity analysis High
EX.M.29—Occurrence of Clostrldium perfringens High
EX.M.30—Aquifer well mapping Medium
EX.M.31—Correlation of water age and microbial High
viability
EX.M.32—Septic Tank Siting High
EX.M.33—Virus sampling and phage research High
3-24
-------
Table 111-4. Research Priorities Risk Assessment for Microbial Pathogens
Research Topics
Proposed Research
Priority
a. Modifications of risk assessment paradigm for characterizing
microbial risks .
b. Accuracy of dose response models in predicting waterborne
disease
c. Characterization of risks posed by exposure to multiple or complex
mixtures of pathogens
RA.M.1—Development of a comprehensive High
microbiology risk assessment model for water
RA.M.2—Evaluation/application of various dose- High
response relationship models
RA.M.3—Evaluation and application of issues and Medium
methods to assess risk associated with
exposures to multiple pathogens, routes,
disease course and outcomes, and durations
3-25
-------
Table 111-5. Research Priorities for Risk Management of Microbial Pathogens
Research Topics
Proposed Research
Priority
Pathogen Removal
a. Optimization of conventional treatment processes to remove
pathogens
b. Effectiveness of different filtration processes in removing
pathogens
c. Effectiveness of disinfection processes in inactivating
pathogens
d. Appropriateness and costs of technologies for small systems
Distribution systems
a. Use of conforms (or other surrogates) to indicate adequate
control of primary and opportunistic pathogens in biofilms
and pipe sediments
b. Water quality factors affecting biofilm growth and the
effectiveness of disinfectant residuals in controlling growth
RM.M.1—Filtration studies for controlling pathogens High
RM.M.2—Filtration removal of protozoa and indicators High
RM.M.3—Optimize conventional treatment for removal High
of oocysts
RM.M.4—Filtration damage viability studies High
RM.M.5—Evaluate disinfection in field-scale treatment High
plants
RM.M.28—Evaluation of the effectiveness of conven- Medium
tional treatment processes on removing and
emerging pathogens
RM.M.6—Biological treatment for control of oocysts High
RM.M.7—Filtration techniques other than conventional Medium
inactivating treatment
RM.M.29—Evaluation of different filtration processes on Medium
removing emerging pathogens
RM.M.8—Control of Norwalk virus by chlorine and ozone High
RM.M.9—UV disinfection efficiencies for Norwalk viruses High
RM.M.10—Inactivation of Giardia and Cryptosporidium by
sequential disinfectants High
RM.M.30—Evaluation of alternative inactivation High
processes for controlling viruses
RM.M.31—Evaluation of different disinfection processes Medium
in inactivating emerging pathogens
RM.M.11 Cryptosporidium removal using bag filters High
RM.M.12—Cost-effectiveness of prefiltration for High
ultrafiltration unit
RM.M.13—Development/testing of innovative High
technologies for small systems
RM.M.14—Bacteria interference with detection of . Medium
conforms and E. coli
RM.M.15—Kinetic Models for chlorine decay in Completed
distribution systems
RM.M.16—Enhancement to the EPANET distribution High
water quality model
RM.M.17—Preliminary studies of biofilm formation Medium
rates in pilot-scale distribution systems
RM.M.18—Opportunistic pathogens in biofilm Medium
RM.M.19—Impact of nutrient removal on growth Medium
potential for bacteria
RM.M.20—Impact of alternative treatment on biofilm Medium
growth
RM.M.21—Water quality factors in distribution systems Medium
c. Effect of design and condition of the distribution system on
bacterial growth
d. Effect of filtration/disinfection treatment processes on
chemical and biological stability of water in distribution
systems
e. Maintaining Distribution System Integrity
RM.M.22—Water quality impacts of dead ends Medium
RM.M.23—Mixing in storage facilities Medium
RM.M.24—Alternative kinetic models for decay and High
DBP formation
RM.M.25—Real-time monitoring systems Medium
RM.M.26—Bacterial growth in distribution systems High
RM.M.27—Integrated approaches for controlling Medium
pathogens
RM.M.32—Evaluate the primary causes of leaks and High
failures in water distribution systems
RM.M.33—Evaluate state-of-the-art devices for High
determining the structural integrity of water
distribution systems
RM.M.34—Evaluate materials for construction High
distribution system components
(Continued)
3-26
-------
Table IH-5. (Continued)
Research Topics
Proposed Research
Priority
Source water
a. The impact of source water protection on the performance
of drinking water treatment systems
RM.M.35—Evaluate GIS techniques for defining source High
water characteristics . ' ••
RM.M.36—Evaluate water quality models for routing High
point and non-point source water discharges
RM.M.37—Evaluate the impact of sudden increases in High
source water contaminant concentration on
drinking water treatment
3-27
-------
-------
Chapter IV. Research for Disinfection By-Products
Background
Public health concern over the disinfection process was
first raised over 20 years ago with the identification of
chloroform and other trihalomethanes (THMs) in chlori-
nated drinking water. Since that time, over 100 chemical
by-products of the disinfection process have been found
in treated drinking water. Based on several studies, the
most prevalent chlorination by-products by weight are
the total THMs (TTHMs), followed by the haloacetic
acids (HAAs), chloral hydrate, haloacetonitriles,
haloketones, and chloropicrin. Between 20% and 60%
of the halogenated material resulting from chlorination
are accounted for by these compound classes. The
concentrations of the chlorination by-products depend
on several factors, with the total organic carbon (TOC)
level (a surrogate for the amount of precursor material)
in the water being the most important. In addition, the
concentrations of chlorinated by-products are generally
higher in surface waters than ground waters because
the level of TOC is higher in surface waters.
The concentrations of the TTHMs and HAAs likely com-
prise more than 50%, on a weight basis, of the haloge-
nated by-products that have been identified in drinking
water. Based on the model used in the proposed Stage
1 DBF rule to predict occurrence of TTHMs and HAAs in
systems serving greater than 10,000 people and using
surface water, the median concentration of TTHMs na-
tionally is 45 ng/l with a 95th percentile of 104 ng/l (the
proposed Stage 1 Maximum Contaminant Level (MCL)
for TTHMs is 80 jig/I). Of the THMs, chloroform gener-
ally occurs in the highest concentrations, but the bromi-
nated and bromochloro- compounds are found in the
highest concentrations in high bromide waters. For HAAs,
the model predicted that the median concentration for
the sum of five of the HAAs nationally is 27 jig/l with a
95th percentile of 86 jig/l (the proposed Stage 1 MCL for
the HAAs is 60 jig/l). Of the HAAs, dichloroacetic and
trichloroacetic acid generally have the highest concen-
trations, but the brominated and bromochloroacetic ac-
ids occur in the highest concentrations in high bromide
waters. The other chlorination by-products generally
occur at levels less than 10 jj.g/1.
Chloramines react to form chlorine-containing by-prod-
ucts, but generally at significantly lower levels than does
chlorine. When free chlorine is present during the appli-
cation process, the level of chlorine-containing by-prod-
ucts increases. Chloramination also produces nitrogen-
containing by-products, such as cyanogen chloride and
organochloramines. Cyanogen chloride generally oc-
curs at concentrations < 5 p.g/1 (Stage 1 does not pro-
pose an MCL for cyanogen chloride).
The disinfectant chlorine dioxide does not react to form
significant levels of chlorine-containing organic by-prod-
ucts, but chlorine dioxide has been found to produce the
inorganic by-products, chlorate and chlorite. Limited data
suggest that chlorite and chlorate are expected in con-
centrations between 0.5 to 1.5 mg/L and 0.1 to 0.5 mg/L,
respectively (the proposed Stage 1 MCL for chlorite is
1.0 mg/L; an MCL is not proposed for chlorate).
Relative to chlorination, considerably less is known about
by-products of ozonation. The most prevalent ozone by-
products are aldehydes, ketones, carboxylic acids,
ketoacids, and hydrogen peroxide. Bromate, bromo-
form, and dibromoacetic acid are formed in high bro-
mide waters. The available health and exposure data for
these by-products suggest that bromate is of greatest
concern. The nationwide distribution of bromate occur-
rence in surface waters using ozone for pre disinfection
has been roughly estimated as a median of 1 to 2 jig/I
and a 90th to 95th percentile in the range of 5 to 20 ng/
I (the proposed Stage 1 MCL for bromate is 10 ng/l).
This chapter describes proposed research for DBFs in
the areas of health effects, exposure, risk assessment
and risk management. The major research questions in
each area are summarized in Table 1V-1. For each
question, the state of the science, research needs and
proposed research projects are described.
Health Effects Research
1. What are the health effects associated
with exposure to DBFs?
Information on the health effects of DBFs from both
epidemiology and toxicology studies is currently inad-
equate for conducting comparative assessments of the
potential cancer and noncancer risks posed by the use
of chlorine, chloramine, ozone, chlorine dioxide, or com-
binations of these disinfectants. The anticipated increased
use of alternatives to chlorine in the future underscores
the need to assign a high priority to research that will
permit a better characterization of the risks that may be
4-1
-------
Table IV-1. Major Research Questions for Disinfection By-Products
Health Effects
1. What are the health effects associated with exposure to
DBFs?
1.1 What are the health effects in communities served by
disinfected drinking water?
1.2 What is the toxicity of individual chemical contaminants
and of mixtures of DBFs?
Exposure
1. What methods are needed for measuring occurrence of DBFs
in drinking water?
2. What levels of DBFs are people actually exposed to via their
drinking water supplies, and what is the population distribu-
tion of exposures?
Risk Assessment
1. How can we characterize the risk posed by exposure to
specific and multiple or complex mixtures of DBFs in drinking
water?
2. How can the risks from chemicals and microbes be com-
pared?
Risk Management
1. How effective are various treatment processes in minimizing
the formation of DBFs?
associated with exposure to the by-products of these
alternatives. The toxic effects of many DBFs, particu-
larly those associated with the use of disinfectants other
than chlorine, are unknown or poorly characterized.
Beyond the need for basic toxicologic data, an improved
understanding of the chemical, physical and biological
processes involved in the toxic response is critical for
evaluating the human risk associated with exposure to
these contaminants. Improved estimates of exposure
and health must be utilized in future epidemiology stud-
ies so that the public health risks to communities served
by disinfected drinking water may be characterized with
a reasonable degree of confidence.
The criteria listed in Chapter I were used to select and
prioritize research projects. As an iterative research
plan, priorities are likely to change as the results of
ongoing research and information about emerging prob-
lems become available. This is a particularly important
issue with respect to the planning of health research to
assess the risk associated with exposure to DBFs. The
current research plan emphasizes the need for feasibil-
ity studies to determine if full-scale epidemiology studies
are likely to provide better estimates of cancer and
reproductive risks than those derived from previous
studies. In the event that full-scale studies are not
conducted, risk assessors will place a greater reliance
on the use of animal toxicity data for individual DBFs.
This will necessitate an enhanced program to improve
the scientific basis for extrapolating animal toxicity data
to humans. Under this scenario, research on toxic mecha-
nisms and pharmacokinetics will receive considerably
more emphasis. Priorities for health research on indi-
vidual DBFs will also be influenced by the ability of
mixtures research to address the overarching question
of drinking water health risks from consumption of disin-
fected water.
The 1996 SDWA Amendments assign a high priority to
the types of research activities described below, empha-
sizing the need for toxicology and epidemiology re-
search to evaluate the cancer and reproductive risks
that may be associated with exposure to DBFs resulting
from different disinfectants. The Amendments also place
a priority on research to understand the mechanisms by
which DBFs cause their effects (cancer and noncancer),
the risks posed by complex mixtures of contaminants,
and the factors that influence effects in the general
population and in susceptible groups.
1.1 What are the health effects in communities
served by disinfected drinking water?
State of the Science
The identification of potentially carcinogenic DBFs such
as chloroform in drinking water two decades ago
prompted a number of epidemiology studies to evaluate
cancer risks in communities served by chlorinated drink-
ing water supplies. These studies have primarily in-
volved comparisons of populations consuming chlori-
nated surface water or unchlorinated ground water.
Some investigations have shown no associations be-
tween consumption of chlorinated water and cancer,
while others have suggested weak to moderate associa-
tions with cancers of the colon, rectum and bladder
(collectively, up to 10,000 cases annually). In general,
the risk estimates derived from these studies are highly
uncertain due to problems with study design, character-
ization of exposures and ascertainment of health ef-
fects. Recent studies in Iowa and Canada have provided
additional weight-of-evidence for the association be-
tween exposure to chlorinated by-products and the risk
of cancer. Epidemiology studies to investigate the pos-
sible association between DBF exposures and adverse
reproductive outcomes have similarly not provided con-
clusive evidence of causality, but they have been useful
for generating hypotheses and identifying research
needs. The high priority given to epidemiology research
in this plan is responsive to the need to address these
important public health concerns.
Obtaining reliable estimates of exposure in cancer stud-
ies has been particularly problematic due to the need to
reconstruct historical exposures that may span several
decades back in time. Additionally, most epidemiology
studies to date have relied upon either TTHMs or chloro-
form as surrogates for exposure to the complex mixture
of DBFs in water. Recent studies have demonstrated
that THMs as a class or individually may not adequately
predict exposure to other types of DBFs. It is now
recognized that the relative concentrations of the indi-
vidual THMs in a given water supply, and their relative
toxicities in animal tests, vary considerably. Further-
more, other types of by-products may be of greater
concern from a health perspective. For example, recent
toxicologic data suggest that of all the DBFs for which
data exist, dibromoacetic acid is perhaps the most po-
tent male reproductive toxicant in rodents. Levels of this
DBF and other toxicologically important brominated by-
4-2
-------
products, including bromodichloromethane (primarily a
chlorination by-product) and bromate (an ozonation by-
product), are greater in communities where bromide
concentrations in the source water are high. These
concerns highlight the importance of selecting the ap-
propriate marker(s) of DBF toxicity to use in both cancer
and reproductive outcome epidemiology studies.
Health data in cancer epidemiology studies have com-
monly been obtained from cancer mortality rates, death
certificates, or interviews. The variable quality of infor-
mation obtained by these measures and the potential
problems in associating exposures to cases are impor-
tant uncertainties that need to be addressed in future
cancer studies. With respect to adverse reproductive
effects, the few existing studies have focused on a
limited range of health outcomes .(fetal development and
selected congenital malformations). To comprehensively
address this issue, studies need to include an examina-
tion of measures of fecundity and fertility in both males
and females. Future epidemiology studies to address
these health issues clearly need improved assessments
of exposure and health effects, enhanced coordination
with animal toxicology research to ensure emphasis on
the most appropriate endpoints and markers of expo-
sure, and greater emphasis on interdisciplinary ap-
proaches in the field.
EPA has used a three-step strategy to define the critical
issues in human health research for both cancer and
reproductive endpoints: 1) sponsor expert panel work-
shops to discuss the available data, determine if addi-
tional epidemiology research can improve the state of
the science, and if so, obtain guidance on research
approaches and priorities; 2) conduct methods develop-
ment and feasibility studies, with consideration of the
recommendations of the expert panel; and 3) conduct
full-scale studies, depending upon the outcome of the
feasibility studies.
EPA has conducted expert panel workshops (step #1)
for both cancer and reproductive endpoints. The cancer
workshop panel recommended conducting feasibility
studies to identify geographic locations with adequate
exposure data and appropriate cohorts for study (includ-
ing the possibility of using existing cohorts that are being
studied for other potential exposures). Several possible
designs for full-scale studies (i.e., cohort, case-control,
and case-control nested within a cohort) were sug-
gested. The panel recommended research on biomarkers
of exposure, effect, and susceptibility, and strongly en-
couraged research to improve exposure assessment for
epidemiologic studies. The recommendations of the can-
cer workshop panel with respect to assessment activi-
ties are discussed in the Risk Assessment section of this
chapter.
In 1993, EPA and the International Life Sciences Insti-
tute (ILSI) convened an expert panel to review the
published epidemiologic and experimental data on re-
productive and developmental effects, and to develop a
strategy for related short-term and long-term research.
The panel concluded that the current available data on
the effects of chlorination by-products provide an inad-
equate basis for identifying DBPs as a reproductive or
developmental hazard. Recommendations were made
for refining studies using existing data bases, strength-
ening studies designed to collect new data, improving
exposure assessments, investigating selected health
endpoints, and developing a stronger link between ani-
mal research and epidemiology studies (e.g., laboratory
development and field testing of biomarkers). EPA sub-
sequently conducted a one-day workshop in March of
1995 to review ongoing laboratory and field research on
reproductive and developmental effects of DBPs. Par-
ticipants discussed directions for future research in this
area, and reiterated several of the general conclusions
of the 1993 workshop. In July of 1997, EPA convened
an e'xpert panel to review the existing reproductive
epidemiology data base and to, develop recommenda-
tions for future research.
Current Epidemiology Studies Current DBP epidemi-
ology research in the U.S. is focused primarily on evalu-
ations of the of the possible associations between expo-
sure to by-products and the risks of adverse reproduc-
tive outcomes. Two epidemiology studies of adverse
reproductive outcomes are being conducted in New
Jersey. One is using improved methods for estimating
exposures to test the findings of earlier studies in that
state in which neural tube defects were associated with
elevated levels of THMs, nitrates, and solvents. The
second is a cross-sectional study of public drinking
water contamination and birth outcomes in various coun-
ties in New Jersey, using ambient levels of THMs and
selected solvents as markers of exposure. The State of
California is evaluating possible associations between
individual or TTHMs in residential drinking water and
adverse pregnancy outcomes, including spontaneous
abortion, low birth weight, preterm delivery and intrau-
terine growth retardation. Finally, a pilot study using a
geographic information system and water quality model-
ing is evaluating the relationship between birth weight
and exposure to DBPs. This EPA-funded project is listed
below under the section "Development/application of
improved tools for field research." The results of all of
these studies should be available in 1997.
Research Topics and Priorities
The following research addresses many of the recom-
mendations of the participants in the cancer and repro-
ductive effects workshops, as described in the State of
the Science section above. These projects are all highly
ranked because they address potential risks of greatest
concern for DBPs in drinking water (cancer and adverse
reproductive outcomes), and if successful, they will sig-
nificantly reduce uncertainties in the current risk assess-
ments and will lead to more scientifically sound, cost-
effective regulations
4-3
-------
a. Development/application of improved
tools for field research
This research addresses important weaknesses in drink-
ing water epidemiology studies through the develop-
ment and application of better approaches for assessing
exposure and health effects. Methodblogic research
should be conducted concurrently with the feasibility
studies described below, and the improved methods
should be integrated into the design of full-scale studies
as they become available.
HE.D.1—Improving estimates of residential
DBF exposures in epidemiology studies This
includes research described in the Exposure
section that has a direct application to epidemi-
ology studies. Projects EX.D.14 through 16 are
particularly relevant to epidemiology research
and also provide information that can be used in
assessing risks.
Priority: Variable (see Exposure section)
HE.D.2—Improving measures of biologic ef-
fect: Field evaluation of biomarkers Field
evaluation of morphological, biochemical, and/
or molecular alterations that may result from
exposure to DBPs. An initial focus on possible
biochemical indicators of male reproductive tox-
icity is supported by recent progress in this area
in the laboratory. The development of biomarkers
for cancer may provide some important insights
into mechanisms of action. However, the priority
of this line of research for supporting epidemiol-
ogy studies will depend upon the priority given
to new full-scale epidemiology studies in the
future. [See project EX.D. 15 in the Exposure
section on biomarkers of DBP exposure]
Priority: High
HE.D.3—Improving methods for managing
health and exposure data This ongoing project
is evaluating the use of a geographic informa-
tion system (GIS) and water quality modeling in
a study of the potential impacts of chlorination
and chloramination on reproductive health of
populations in Colorado. GIS, which is a data
management and visualization tool that has been
developed for use in other contexts, is assisting
in study design, sample selection, and analysis
of data. If effective, this approach could be
applied in future full-scale studies.
b.
Priority: High
Feasibility/full-scale studies
Plans for future epidemiologic research will be influ-
enced by current assessments of existing data (see
Assessment section below), the outcome of epidemiol-
ogy studies now underway, and ongoing efforts to im-
prove epidemiologic methods. The availability of suit-
able health and exposure data, as well as the amount of
resources available, will be key determinants in the
number, types and locations of studies that will be
conducted. Some elements of study design may be
similar for both cancer and reproductive outcomes, such
as identifying areas with water quality parameters of
interest (e.g., pH, bromide concentration).
Cancer epidemiology studies in communities served by
water treated with disinfectants other than chlorine or
chloramine will not be possible due to the unavailability
of adequate historical exposure data. Assessments of
potential cancer risks associated with exposure to the
by-products of alternative disinfectants will therefore
need to rely upon data from toxicity studies of individual
DBPs and mixtures of DBPs (see toxicology section
below). Epidemiology studies of adverse reproductive
outcomes for the various treatment options may be
possible since these studies are not constrained by the
need for such long-term exposure data.
The general goal of the feasibility studies is to demon-
strate that an acceptable epidemiologic test of the pos-
sible association between DBP exposures and cancer
or noncancer risk can be conducted. More specifically,
the studies must provide reasonable assurances that: 1)
the subjects' exposures to DBPs can be accurately
assessed at least over the latent or critical period of the
development of the adverse effects of concern; 2) study
sites exhibit water quality parameters (e.g., bromide
levels, pH) and treatment processes of greatest interest;
3) the study population exhibits enough variation in DBP
exposure to allow a dose-response function to be estab-
lished; 4) suitable measures or markers of exposure and
effect will be used; 5) the test for DBP effect will not be
unduly confounded by factors such as differences in
demographic characteristics, unmeasured exposure to
environmental contaminants other than DBPs, and resi-
dential mobility; and 6) the study will have sufficient
statistical power across a reasonable range of sample
sizes and odds ratios.
HE.D.4—Feasibility studies: Cancer To be
conducted in communities served by chlorinated
and chloraminated water supplies, with a focus
on cancer of the bladder, colon, rectum, and
other possible cancer sites. Reassessment of
the existing cancer epidemiology data (see As-
sessment Section below) is an essential prereq-
uisite to determining whether new studies should
be carried out.
Priority: High
HE.D.5—Feasibility studies: Reproductive
effects To be conducted in communities served
by chlorinated, chloraminated, and possibly
ozonated water supplies. Studies may focus on
at least one of the following endpoints: human
fertility and fecundity (male and female), growth
retardation, and common malformations. The
few available DBP reproductive epidemiology
4-4
-------
studies have been reviewed in several expert
workshops convened by EPA (described above).
The research recommendations that have been
developed from these workshops should be con-
sidered in the conduct of future feasibility stud-
ies.
Priority: High
HE.D.6—Full-scale studies: Cancer and re-
productive effects Dependent upon the find-
ings of Projects 4 and 5 above. May include new
studies and/or collaboration with planned, re-
cently completed or ongoing case-control stud-
ies.
Priority: Dependent upon outcome of feasibility
studies. (If feasible, high)
1.2 What is the toxicity of the highest priority chemi-
cal contaminants and mixtures of DBFs?
State of the Science
The toxicologic literature on individual DBPs has grown
considerably over the last 20 years, particularly for the
by-products of chlorination. Basic toxicologic informa-
tion is available on many of the THMs, HAAs, alde-
hydes, and miscellaneous organic and inorganic con-
taminants found in treated drinking water. A number of
these by-products have been shown to cause cancer,
reproductive effects, and various target organ toxicities
in experimental animals, though most commonly at ex-
posure levels that are from one to several orders of
magnitude higher than the ambient concentrations to
which people are exposed.
The DBPs that have been studied experimentally repre-
sent only a fraction of the by-products that may be
present in treated drinking water. Consequently, there
are many uncertainties in assessing risks based on such
limited data. While the ambient concentrations (typically
in the jig/L range) of many of the known DBPs are likely
to be greater than the levels of most of the other by-
products in the complex mixture, there may be poorly
characterized or as yet unidentified by-products that are
of concern from a health perspective. For example, the
highly mutagenic by-product MX and related haloge-
nated hydroxyfuranones may be of public health impor-
tance even though they are present in relatively low
concentrations in treated drinking water. The toxicity of
many of the brominated, mixed bromochloro, and inor-
ganic by-products has been inadequately studied. This
includes DBPs such as the brominated THMs and MX-
related compounds (formed with chlorination and
chloramination), bromate and the brominated acids (from
ozonation and chlorination), cyanogen chloride (from
chloramination), and chlorate (from chlorine dioxide). In
addition, nearly all of the available toxicologic data on
DBPs pertain to individual compounds. There is little
information on the toxicity of complex mixtures of DBPs,
or on the toxicologic interactions that may occur be-
tween the individual DBPs in a mixture.
Perhaps the most critical underlying issue for toxicologic
research on DBPs is the need to better understand the
predictive relationship between toxicologic endpoints
and human disease (i.e., animal-to-human extrapola-
tion). The metabolism of DBPs, and the mechanisms by
which they cause their effects in animals, directly influ-
ence the shape of the dose-response curve at low
doses; this has important implications for estimating the
risks posed by DBPs to humans.
Whereas the epidemiology data suggest possible asso-
ciations between DBPs and bladder or cola-rectal can-
cer, the experimental carcinogenicity data indicate that
the liver and kidney are the most common target organs
for the THMs and HAAs. The main exceptions are the
rare intestinal tumors in animals exposed to
bromodichloromethane and bromoform (which corre-
sponds to the epidemiology data), and tumors of the
thyroid and peritoneal mesothelioma in bromate-treated
animals. The mutagenicity of DBPs varies within and
between classes, with the brominated by-products gen-
erally more mutagenic than those that are chlorinated.
Basic cancer dose-response information is inadequate
for a number of potentially important by-products, in-
cluding chlorate, cyanogen chloride, MX, and several of
the brominated acetic acids. In addition, more detailed
biologic information is needed to further characterize-the
carcinogenicity of dichloroacetic acid, bromodichloro-
methane, and bromate, which appear to be the most
important by-products from a risk perspective based on
a consideration of their estimated cancer potencies and
ambient exposure levels.
Several DBPs have been shown to cause reproductive
and developmental toxicity in laboratory animals, al-
though an examination of the existing data base sug-
gests that the studies have not been performed in a
completely systematic manner. In general, more data
are available on the effects of DBPs on the development
of the embryo and fetus than on other parts of the
reproductive process. Recent data from relatively low
dose studies of the brominated acids suggest the need
for more detailed investigations of some of these by-
products. Screening-level data are needed for a number
of halo-methanes, haloacids, halonitriles, ketones and
inorganics (particularly bromate).
The neurotoxicity and immunotoxicity of DBPs in gen-
eral have not been well characterized. There are few
indications in the literature that these endpoints should
be of concern at ambient exposure levels. Nevertheless,
a limited screening effort to evaluate selected DBPs for
these endpoints in a systematic manner is needed to
ensure that this assessment is correct. Preliminary re-
sults from studies in rats have shown that relatively low
doses of dichloroacetic acid (similar to the lowest doses
at which other noncancer effects have been observed
experimentally) can cause reversible neurotoxicity after
exposure periods of three months. These findings sug-
gest that additional research to characterize the poten-
tial neurotoxicity of dichloroacetic acid and related
haloacids is warranted.
4-5
-------
A number of studies have been conducted to address
questions concerning the toxicity of "real world" mixtures
of DBFs. These studies, which used either concentrates
of drinking water or humic acid preparations treated with
various disinfectants, were largely negative or inconclu-
sive, and had a number of technical limitations. In addi-
tion, a few studies have evaluated mixtures of drinking
water contaminants such as pesticides and other organ-
ics, but research on defined mixtures of DBFs is lacking.
Little is known about the possible toxicologic interac-
tions of individual DBFs in a mixture, i.e., whether their
toxicity is additive, as is the current default assumption,
or if their toxicity is either more or less than additive. To
address the many issues relating to toxicology, sample
preparation, chemical analysis and assessment, an inte-
grated, cross-disciplinary research effort is necessary.
With respect to the disinfectants themselves, the toxico-
logic data base on chlorine, chloramine, and chlorine
dioxide is considered variable, although there is gener-
ally little health concern over exposure to the levels of
disinfectant residuals commonly found in finished drink-
ing water. Additional data are needed, however,* to
better characterize the potential immunotoxicity of these
disinfectants.
EPA is currently conducting hazard identification and
dose-response research on a number of DBFs across a
variety of toxic endpoints. Chronic cancer bioassays
have recently been completed or are near completion
for di- and trichloroacetic acid, bromodichloromethane,
chloral hydrate, and bromate. The current focus of mecha-
nistic and pharmacokinetic studies is on dichloroacetic
acid, bromodichloromethane, and bromate. Research is
being conducted at EPA on the male reproductive toxic-
ity and the embryotpxicity of selected HAAs and THMs.
EPA is also evaluating the neurotoxicity of dichloroacetic
acid. These by-products were selected for more detailed
studies because they appear to be of greatest concern
from a risk perspective.
The following listing of research activities outside of EPA
is incomplete, but is nevertheless intended to provide
some useful information on other important ongoing
research. In collaboration with EPA, the National Toxi-
cology Program (NTP) has a screening program to
evaluate the reproductive, developmental, and
immunotoxic effects of selected DBFs. The NTP also
plans to conduct mechanistic research using short-term
models (e.g., transgenic animals) to evaluate selected
DBFs, and will initiate chronic bioassays on several high
priority DBFs, including bromodichloromethane, chlor-
ate, dibromoacetic acid, dibromoacetonitrile and MX.
Mechanistic research on the carcinogenicity of chloro-
form is being conducted by the Chemical Industry Insti-
tute of Toxicology. The Chemical Manufacturers Asso-
ciation is conducting a two-generation rat reproductive
study on chlorite. Research in Finland recently com-
pleted a study that demonstrated the carcinogenicity of
MX in a chronic cancer bioassay. A cancer bioassay on
dibromochloroacetic acid was recently initiated in Ja-
pan.
Research Topics and Priorities
Toxicologic research described in this chapter can be
grouped into three distinct categories: a) hazard identifi-
cation and dose-response~studies to fill data gaps for"
newly identified or priority contaminants; b) developing
more detailed toxicologic information for the most impor-
tant substances, particularly in the areas of pharmacoki-
netics and mechanism(s) of action to facilitate extrapola-
tion of data to humans; and c) evaluating the toxicity of
simple and complex mixtures of DBFs. The preferred
research approach involves initial screening-level stud-
ies of DBFs to fill data gaps, followed by more detailed
research if additional data are needed for quantitative
risk assessment. Toxicology studies of individual DBFs
and possibly drinking water mixtures will be particularly
important for evaluating the relative cancer risks that
may be associated with disinfection strategies other
than chlorination and chloramination, since historical
exposures to the by-products of the alternative treat-
ments are too short to evaluate such risks in epidemiol-
ogy studies.
The data obtained over the, next five years from DBF
screening studies and from trie more focused mechanis-
tic studies on selected priority by-products will greatly
enhance the assessment of risks that may be associ-
ated with exposure to individual by-products of chlorina-
tion and alternative disinfectants. These studies will also
provide insights into the relative risks of different treat-
ments. It is clear, however, that laboratory studies of
DBF mixtures and/or field studies of human populations
represent more relevant approaches for comparing the
risks of the various treatment processes (technical con-
straints and data limitations notwithstanding).
a. Hazard identification and dose-
response
For many newly identified and some known drinking
water contaminants, basic information is needed to de-
scribe their potential toxicity for a variety of endpoints,
including cancer, reproductive and developmental toxic-
ity, and neurotoxicity. An important component of re-
search to address data gaps, potentially applicable to
one or more of the listed toxicity endpoints and closely
aligned with data generation efforts, is the development
of structure-activity relationships (SAR). When sufficient
data on related DBFs exist, SAR models can be derived
and used to provide preliminary estimates of toxicity of
untested DBFs, prioritize newly identified chemicals for
testing, and in some cases, generate insight into mo-
lecular mechanisms of toxicity.
Experimental research to fill data gaps for hazard identi-
fication and dose-response assessment purposes is
described below. The cancer dose-response studies
and the reproductive/developmental effects screening
studies are high priorities because they will help identify
those DBFs with the highest risk, provide dose-response
information that can be used in conducting better risk
assessments, and provide better information to more
4-6
-------
accurately estimate costs and benefits of the rules. The
neurotoxicity and immunotoxicity studies are medium
priorities because of the lack of published studies dem-
onstrating that these endpoints are of great concern.
HE.D.7—Cancer dose-response studies Two-
year exposure studies in laboratory animals to
evaluate the potential carcinogenicity of DBFs,
using study designs that will provide better infor-
mation for use in Agency risk assessments.
Selection of an appropriate range of doses that
will permit an evaluation of the dose-response
in the low-dose range (to the extent possible
experimentally) will be a key consideration in
the study design. These studies also address
selected pharmacokinetic and mechanistic ques-
tions that will provide data for use in more
biologically based risk assessments (see Projects
HE.D.11 and 12). As mentioned above, the NTP
is planning to initiate chronic exposure studies
on several high priority DBFs in 1997 and 1998.
Priority: High
HE.D.8—Reproductive/developmental effects
screening studies Evaluation of the reproduc-
tive and developmental toxicity of priority DBFs
in standardized screening tests, and conducting
specialized developmental toxicity studies on
selected DBFs and disinfectants (e.g., two-gen-
eration studies, research to identify alterations
in pregnancy maintenance and ovarian func-
tion). Current priorities in an ongoing EPA/NTP
collaborative screening program include bro-
mate, chlorodibromomethane, bromodichlorp-
methane, bromoacetonitrile, dibromoacetonitrile,
and bromochloroacetic acid.
Priority: High ,
HE.D.9—Neurotoxicity studies Characteriza-
tion of the neurotoxicity of priority DBFs, with an
initial focus on dichloroacetic acid and related
HAAs.
Priority: Medium
HE.D.10—Immunotoxicity studies Evaluation
of the potential immunotoxic effects of selected
high priority DBFs and disinfectants in labora-
tory animals using a tiered approach that in-
cludes several measurements of immune func-
tion. This research will be conducted in a coordi-
nated effort with NTP.
Priority: Medium
b. Pharmacokinetics and mechanisms of
action
Research on pharmacokinetics (i.e., distribution, me-
tabolism and elimination) and mechanisms of action is
necessary to interpret the biological significance of an
effect and to provide a sound scientific basis for assess-
ing risk. By describing metabolism, tissue dosimetry and
tissue response, physiologically based pharmacokinetic
(PBPK) models and biologically based dose-response
(BBDR) models permit a more accurate prediction of the
shape of the dose-response curve for humans exposed
to a particular contaminant. This research also includes
the development of a structure/ activity framework for
assessing the toxicity and mechanisms of action of
related DBFs, and provides insights on the development
of biomarkers of effect and exposure. Due to the re-
source demands and longer time frame for this type of
research, only the highest priority individual DBFs and
classes of DBFs are selected for study. As mentioned
above, dichloroacetic acid, bromodichloromethane, and
bromate are considered to be three of the most impor-
tant DBFs for which this type of detailed information is
desirable with respect to cancer. Current data suggest
that more detailed studies of the reproductive and devel-
opmental toxicity of certain THMs and HAAs are war-
ranted. These projects are considered to be high priority
because they address major uncertainties in the risk
assessment process, and focus on the toxic endpoints
of greatest concern. This research is also responsive to
the research requirements of the 1996 SDWA Amend-
ments, which direct the Agency to conduct research to
improve the biological basis for drinking water risk as-
sessments.
HE.D.11—Pharmacokinetic and mechanistic
research to improve cancer risk assessment
for priority DBPs This includes research in two
interrelated areas: 1) Studies to evaluate the
pharmacokinetics and toxicity of selected DBFs
to obtain better estimates of metabolism and
relevant target tissue dosimetry; 2) Mechanistic
studies to evaluate the physiological, biochemi-
cal and molecular changes that accompany the
carcinogenic response. These efforts will pro-
vide data to support more biologically based risk
assessments for a small number of the highest
priority DBPs, and will in some cases lead to the
development and subsequent validation of physi-
ologically based pharmacokinetic (PBPK) mod-
els and/or biologically based dose-response
(BBDR) models. Pharmacokinetic and toxicity
research at EPA is currently focused on se-
lected THMs and HAAs, with a primary empha-
sis on the development and validation of PBPK
model for bromodichloromethane. Mechanistic
research is currently being conducted on
dichloroacetic acid and bromate. These by-prod-
ucts were selected for study because the results
of screening studies, combined with estimates
. of exposure, indicated that they could be of
greatest importance from a risk assessment
perspective. This research is linked with the
dose-response studies described in HE.D.7.
Priority: High
HE.D.12—Pharmacokinetic and mechanistic
research to improve assessments of repro-
4-7
-------
c.
ductive toxicity and developmental effects
for priority DBFs Research to further charac-
terize the potential reproductive and develop-
mental toxicity of priority DBFs. Ongoing re-
search is focused on the THMs and HAAs,
using a variety of in vitro and in vivo techniques.
Contaminants are selected for these more de-
tailed investigations on the basis of results of
earlier screening-level studies described in
HE.D.8.
Priority: High
DBF mixtures
As an integral part of a multi-disciplinary effort on drink-
ing water mixtures, toxicology research can provide the
data needed to characterize the relative toxicity of com-
plex mixture preparations and to evaluate the interac-
tions between key components of the mixture. Important
issues in complex mixtures research include the need to
determine the appropriate study design and to over-
come technical problems relating to sample preparation
and chemical analysis. The first two mixtures projects
described below represent a phased approach in which
a feasibility study is conducted before attempting full-
scale studies to evaluate the relative toxicity of complex
mixture drinking water samples. The feasibility study, as
well as the mutagenicity screening study, are consid-
ered high priority because they address a pqtentially
high-risk, high-uncertainty issue, and they could lead to
further research that will provide information on the
relative risks of different disinfected waters. The DBP
interactions study can provide useful insights into the
assumptions currently used in assessing risks for mix-
tures. It is ranked medium because it is a less-direct
approach to assessing the relative risks of complex
drinking water samples.
HE.D.13—Mixtures feasibility study Assess-
ment of the feasibility of studying drinking water
mixtures, and development of a "research strat-
egy. This involves the use of workshop(s), spe-
cial consultations and pilot studies, and must be
linked to similar feasibility assessments in re-
lated areas (e.g., exposure/chemistry).
^Priority: High
HE.D.14—Toxicologic evaluation of drinking
water mixtures Toxicologic testing and devel-
opment of new methods as needed to evaluate
the relative toxicity of drinking water samples
that vary by disinfection technique and selected
water quality parameters. This also includes
research to identify the individual components
or classes of DBFs that contribute most to the
overall toxicity of the mixture.
Priority: Depends upon outcome of Project 13
(High if feasible)
HE.D.15—Mutagenicity screening studies of
drinking water mixtures Use of in vitro assays
to evaluate the relative mutagenicity of complex
mixtures of DBFs in concentrated extracts of
drinking water treated with different disinfec-
tants (ongoing).
Priority: High
HE.D.16—Studies of DBP interactions Re-
search to evaluate DBP interactions in well-
defined mixtures, with an initial focus on the
major classes of DBFs (e.g., THMs and HAAs).
This is linked to mixtures assessment activities
in Section D, Risk Assessment, below.
Priority: Medium
Exposure Research
Approximately 100 different DBFs have been identified
in treated drinking water. For a relatively small number
of DBFs, perhaps 10-15, occurrence data are reason-
ably well established in terms of concentration ranges
typical of source waters and some treatment processes.
Major gaps still exist in our knowledge about what DBFs
are formed from different treatment processes, particu-
larly processes that use alternate disinfectants, such as
ozone. Research in this area is hampered by lack of
analytical methods. Uncertainties also exist regarding
actual exposures to DBFs.
1. What methods are adequate for the
analysis of DBFs?
State of the Science
Analytical methods for specific DBFs are either research
methods, often the methods used in the discovery pro-
cess, or practical methods required for regulatory com-
pliance monitoring or large-scale exposure surveys. Re-
search methods are usually unacceptable for regulatory
use or even large-scale exposure surveys because they
are usually too poorly tested and documented, relatively
complex and costly, and often infested with unexpected
problems. Regulatory methods are desired that are
simple, relatively cheap, free of problems, accurate,
precise, multi-laboratory tested, and have method de-
tection limits (MDLs) no greater than 20% of the pro-
posed MCL The requirement of an MDL no greater than
20% of the MCL is highly desirable because this level is
sufficient to ensure that normal measurement variability
does not produce an out-of-compliance measurement
when the utility is in compliance.
Very good practical methods of analysis are available
for a few of the DBFs, e.g., chloroform, bromo-
dichloromethane, chlorodibromomethane, and bromo-
form (the trihalomethanes, THMs). These successful
practical methods rely on the use of high resolution gas
chromatography (GC) and GC/mass spectrometry, the
development of which had an enormous impact on our
ability to discover, identify, and measure thermally stable,
readily extracted, and reasonably volatile organic com-
pounds in drinking water. Thermally stable volatile com-
4-8
-------
pounds are capable of passing into the "gas phase
without chemical decomposition or other chemical
changes.
Methods for other by-products are useable, but further
improvement is highly desirable. For example, the ana-
lytical methods currently used for the haloacetic acids,
which are currently proposed for regulation, are not
simple. These methods require a very skilled laboratory
technician and are relatively costly compared to the
methods used for the THMs.
A major and very significant analytical limitation exists
for the measurement of polar, water soluble (difficult to
extract), and nonvolatile organic and inorganic sub-
stances in source and treated drinking water. The lack of
research methods for these type of compounds has
impeded the capability to identify and conduct exposure
studies for additional by-products, especially those from
alternate disinfectants or from combinations of disinfec-
tants. A major need is the development of new extrac-
tion techniques and instrumentation for the identification
and measurement of polar, water soluble, nonvolatile
substances in drinking water.
From the exposure point of view, another need is for
real-time methods to continuously monitor finished and
source water for DBFs and precursors. Real-time moni-
toring would provide information on the temporal varia-
tions in the concentrations of DBPs which is needed to
determine the true level of exposure by the consumer.
As a by-product of this development, treatment plants
could use these technologies to provide on-line, in-the-
plant process monitoring and feedback control of DBF
formation by adjusting process parameters as a function
of current conditions. Current AWWARF-sponsored re-
search on real-time monitoring methods addresses TOX
and TOC, but not disinfectant residuals or by-products.
Smaller treatment plants generally need to rely on com-
mercial laboratories for the analysis of DBPs due to the
relatively expensive equipment requiring highly trained
personnel (e.g., high resolution GC/MS). Development
of alternative simpler "kit-like" methods for specific DBPs
would be less expensive, easier to use and may provide
better consistency and quality control in reporting.
Research Topics and Priorities
a. Additional methods or method
improvements needed to implement the
Stage 1 DBP rule
The improvement of the methods for bromate and halo-
acetic acids are considered high priority because they
would help in the implementation of the Stage 1 DBF
rule. In addition, the potential health risks posed by
these chemicals at potentially low levels requires better
analytical methods to ensure the best possible occur-
rence data are available when estimating risks. The
evaluation of the TOC method is a high priority because
of the importance of TOC in the generation of DBPs and
the regulatory requirement to measure TOC in the en-
hanced coagulation requirements. The development of
a more sensitive chlorine dioxide method is a medium
priority because, as discussed in the project description,
the need for a more sensitive method is dependent on
the outcome of the CMA two-generation rat study (de-
scribed in the Health Effects section, above). The real-
time monitoring for disinfectant residuals is a low priority
because compliance with the proposed criteria (MCLs
for TTHMs and HAAs) is based on an annual average of
concentrations measured in the distribution system. How-
ever, if reproductive/developmental effects are of a con-
cern, then standards may be established to prevent the
threshold level from occurring anywhere in the distribu-
tion system. In this case, real-time monitoring may be-
come more important. The performance evaluation (PE)
studies for DBPs are a high priority because it is critical
that there are sufficient and qualified laboratories for
performing analytical work required by the ICR and
Stage 1 DBP rule.
EX.D.1—Low level bromate (BrO3") measure-
ment The need for a practical analytical method
for bromate has been established by the current
proposal to set an MCL of 0.01 mg/L The
current standard analytical method for bromate
has a serious interference from chloride at this
concentration and the method is not sufficiently
sensitive to measure bromate reliably at the
proposed MCL A recently developed ASTM
method has the potential to reduce the method
detection limit (MDL) of bromate ion from 0.02
mg/L to at least 0.002 mg/L which is 20% of the
proposed maximum contaminant level (MCL) of
0.01 mg/L. To be practical, this new method
must be validated by multi-laboratory testing.
Lower detection limits are desired for possible
use in the Stage 2 DBP rule.
Priority: High
EX.D.2—Improved method for haloacetic ac-
ids A recently developed method for the deter-
mination of haloacetic acids reduces some of
the concerns about method deficiencies in ex-
traction, derivatization, and lab safety and it
expands the method to include all nine possible
bromo-, chloro-, and mixed bromochloroacetic
acids. This new method must be validated by
multi-laboratory testing. The long-term need is
to develop a simpler method that has increased
sensitivity. Even more capabilities may be
needed for the Stage 2 rule.
Priority: High
EX.D.3—Expand quality control for TOC,
evaluate new TOC methods
The currently proposed DBP rule specifies total
organic carbon (TOC) as a required test method.
Several different technical approaches are avail-
able and the only documented method has qual-
ity control deficiencies. This project is needed to
4-9
-------
evaluate the published methods and instrumen-
tation for TOG to determine whether they can
achieve the required MDLs and precision. A
documented method with full analytical quality
control directions will be produced. To be practi-
cal, this new method must be validated by multi-
laboratory testing.
Priority: High
EX.D.4—Low-level CIO measurement A more
sensitive laboratory method for chlorine dioxide
will be needed if an ongoing CMA study on the
reproductive effects of chlorite shows risk below
0.8 mg/L. The existing method has a detection
limit deficiency.
Priority: Medium (until risk is clear; could be
elevated to high priority)
EX.D.5—Real-time monitoring for disinfec-
tant residuals Develop performance criteria for
real-time, on-line, in-the-plant continuous moni-
toring technologies for disinfectant residuals to
provide in-plant feedback control of their con-
centrations. This is needed for the proposed
disinfection rule.
Priority: Low (except if developmental or repro-
ductive risks become evident, could be elevated
to high priority)
EX.D.6—PE studies for DBPs and disinfec-
tants A requirement of the proposed rules is
that laboratories participate in performance
evaluation studies to meet the laboratory certifi-
cation requirements of the rules. Some addi-
tional research will be required to develop pro-
cedures for preparing, stabilizing, and distribut-
ing DBP PE samples and statistically analyzing
the results of the PE studies. Conduct perfor-
mance evaluation studies of methods for disin-
fectants and DBPs.
Priority: High
b. Methods needed for Stage 2 DBP rule and
the longer-term
The method for peroxides is a medium priority because
peroxides are relatively unstable and decompose to
form other substances. In addition, peroxides will be
addressed in the Stage 2 DBP rule and therefore there
is not an immediate need to develop improved methods.
The real-time, in-plant monitoring of DBPs is a medium
priority for the same reasons discussed above for moni-
toring of residuals. The improved method for aldehydes
is a medium priority because of the uncertainties with
the results of the ICR and because this information is not
needed for the Stage 1 DBP rule. The priorities of the
projects below may change, or additional projects may
be required, depending on the ICR data, project findings
under the exposure research described in the next
section, health effects, and risk assessment information.
EX.D.7—Methods for peroxides Peroxides are
believed to be formed in water treated with
ozone. Exposure issues are uncertain and can-
not be answered until research indicates whether
these substances are produced in quantities
that would be a concern. Develop research
methods for the determination of peroxides that
may form in water treated with ozone.
Priority: Medium (5 yrs, $100K for multi-labora-
tory validation)
EX.D.8—Real-time, in-plant monitoring of
DBPs Develop real-time, on-line, in-the- plant
or field continuous monitoring technologies for
disinfection by-products to provide data on tem-
poral variations in the concentrations of DBPs
for exposure studies and in-plant feedback con-
trol of their formation.
Priority: Medium
EX.D.9—Improved method for aldehydes Al-
dehydes are formed in water treated with ozone.
A major issue is the reaction of formaldehyde
with chloramine (added to provide a disinfectant
residual) to form cyanogen chloride. The method
for aldehydes being proposed for the Informa-
tion Collection Rule (ICR) is a research method
which may not be a practical method for regula-
tory compliance monitoring. Develop an im-
proved method for aldehydes that form in water
treated with ozone.
Priority: Medium
2. What levels of DBPs are people
exposed to via their drinking water
supplies, and what is the population
distribution of exposures?
State of the Science
The types and concentrations of by-products that are
formed by different disinfectants are not fully known. A
number of factors other than the disinfectant itself are
known to influence the formation and type of DBPs, for
example, variation in the amount of organic material in
the source water, pH and water temperature (see Risk
Management section of this chapter). Because the data
base is incomplete, not all disinfection scenarios have
been thoroughly studied, and there are few analytical
methods for polar, water soluble and nonvolatile DBPs,
this major question cannot be answered at this time.
The complexity of the exposure question is illustrated by
a number of additional questions. When tap water is
used to dilute frozen juices, make ice cubes, take show-
ers, take baths, or make coffee, are consumers exposed
to the same DBPs that are found at the treatment plant
or to different substances? Since 40%-60% of the drink-
ing water ingested by humans is associated with dietary
intake, what effect does food processing and prepara-
tion have on DBPs? What happens to the DBPs found at
4-10
-------
the treatment plant when the water is stored for long
periods in reservoirs or tanks? What happens to these
substances when the water is piped long distances,
trapped for long times in dead-end pipes, or blended
with water from different sources that has been treated
differently? Presently we cannot fully answer any of
these questions.
Although the drinking water regulations were intended to
protect the consumer, except for THMs, very little is
known about changes in DBFs and other substances
that occur in the distribution system and the actual
exposure by people to DBFs. The ICR will provide
extensive DBF occurrence data within treatment plants
and distribution systems, but mainly for chlorinated DBFs.
This data will be used for developing and testing models
for predicting DBF occurrence at consumer taps. In
addition, several long-term research projects are needed
to address these questions.
Research Topics and Priorities
a. Characterize types and concentrations
ofDBPs formed from different
disinfectants and combinations
Research is needed to better characterize the types and
concentrations of DBFs that are formed with chlorine as
a function of water quality and disinfectant conditions.
Similarly, research must address the types and concen-
trations of DBFs formed with ozone plus chlorine and
chloramines as residual disinfectants, with chloramine
alone, and chlorine dioxide alone, as a function of water
quality and disinfectant conditions. The identification of
new DBFs from alternate disinfectants is a high priority
because the information from this project in conjunction
with toxicity data can provide estimates of the risks from
alternate disinfectants and thus help answer questions
relating to the magnitude of the risks from alternative
disinfectants. The methods for nonvolatile DBFs are a
high priority because without these methods it will be
impossible to accurately determine the occurrence of
nonvolatile DBFs and thus estimate the risk from these
chemicals.
EX.D.10—Identify new disinfection by-prod-
ucts from alternate disinfectants Identify pre-
viously unidentified organic DBFs in drinking
water treated with alternative (nontraditional)
disinfectants and determine the effects of oper-
ating conditions on DBF formation. This knowl-
edge will allow more complete exposure as-
sessments and risk assessments and will not be
limited to the examination of XAD resin extracts,
but will explore alternative extraction schemes.
Priority: High
EX.D.11—Methods for nonvolatile DBFs De-
velop extraction methods and advanced instru-
mentation to characterize the nonvolatile-and
difficult to extract organic and inorganic DBFs.
b.
This knowledge will allow the development of
more practical analytical methods for both the
regulatory program and permit more complete
exposure assessments and risk assessments.
Priority: High
Evaluate factors that affect exposure
levels, and assess human exposures to
DBFs
All the projects, except the modeling of DBF exposure,
are a medium priority because the data generated from
these projects, while important, are not absolutely needed
for completing the Stage 1 or 2 DBF rule (the majority of
information on exposure will come from the ICR). The
modeling of DBF exposure is a high priority because this
information can be used in reducing the uncertainty in
risk assessments and in providing more accurate esti-
mates of the costs and benefits of the various rules.
EX.D.12—DBF changes in distribution sys-
tems Characterize the changes that occur in
the distribution system including reservoirs,
tanks, dead-end pipes, or in water blended with
.product from different sources that has been
treated differently. Include the characterization
of DBFs in all exposure field studies that collect
drinking water samples. This work should be
coordinated with distribution system studies in
the Risk Management section. The results will
allow more complete exposure assessments
and risk assessments.
Priority: Medium
EX.D.13—DBF interactions with foods and
associations with dietary intake Determine
what effects disinfectants and DBFs have when
they interact with other foods, e.g., frozen juices,
coffee, etc. Determine what other substances
are formed that impact human exposure and
contribute to the risk assessment. Determine
the influence of food preparation factors such as
temperature and storage on the degradation of
DBFs.
Priority: Medium
EX.D.14—Exposure to DBPs through show-
ering and other household tapwater uses
Determine human exposure to volatile DBPs
through inhalation and dermal exposure during
hot showers and hot tub baths. Expand limited
existing data on trihalomethanes to the poten-
tially more significant haloacetic acids,
haloacetonitriles, haloketones, aldehydes, etc.
Existing data on human activity patterns and
existing household exposure models should be
used. These exposure data are needed to com-
plete the risk assessment.
Priority: Medium
4-11
-------
EX.D.15—Markers of DBP exposure Discover
the underlying biochemical reactions related to
exposure by people to DBFs and use these as
biochemical markers of human exposure. Do
DBFs react with DMA or proteins and can mark-
ers of this exposure be found? This information
could be critical to long-term risk assessment.
Priority: Medium
EX.D.16—Models of DBP exposure Develop
models of human exposure to disinfectants and
DBFs and validate these models with actual
exposure data. The goal of this model develop-
ment is to allow the calculation of actual human
exposure to a variety of DBFs given knowledge
of precursors in the source water, the treatment
technology, consumption patterns, and the model
being developed for the decay of disinfection
residuals in the distribution system (see Risk
Management section, project 3). Various mod-
els will be developed using surveys and ICR
data.
Priority: High
EX.D.17—Exposure as a function of popula-
tion distribution Conduct a series of studies to
identify exposure as a function of age, behav-
ioral patterns and environmental factors that
affect water consumption. These studies.should
be done in conjunction with project EX.M.22.
Priority: Medium
EX.D.18—Tap water consumption and chemi-
cal contaminants The results will consist of
two sets of tables. One set of tables will provide
tap water intake in terms of miililiters/person/
day and the other will be in terms of milliliters/
kilogram bw/day. The tables will provide esti-
mates of the mean with intervals, as well as
estimates of the median with 90th., 95th and 99th
percentiles. In addition, the tables will provide
the sample size and estimated population for
each distribution. Graphics for selected distribu-
tions will also be included. Further, tap water
intake will be analyzed based on economic sta-
tus and age; race and age; residential status
and age; and geographic region and age. The
four geographic regions include: the Northeast,
South, Midwest and West.
Priority: High
Risk Assessment Research
Critical to establishing a regulatory strategy for drinking
water is identifying those contaminants which pose the
greatest risk to human health and, consequently, what
treatments can be used to reduce these risks and at
what cost? To characterize the magnitude and severity
of adverse health effects associated with exposures to
DBFs it is necessary to use the methods and tools of
risk assessment. The central role of the risk assessment
is to evaluate the scientific data, and to provide risk
managers with qualitative and quantitative estimates of
risks posed by specific waterborne agents. With this
information different risk reduction options can be devel-
oped. Through the development and application of con-
sistent methods and tools for integrating and interpreting
the scientific data, risk assessment studies can provide
the framework for comparing chemical and microbiologi-
cal risks. Ideally such comparisons would be made on
information taken from studies involving exposure to
mixtures of DBFs and microbes actually present in the
source and drinking water of various treatment and
distribution systems. However, this is often not possible
because of the lack of information on these whole
mixtures, or alternative approaches for assessing mul-
tiple exposures and effects.
Congressional interest has recently focused on risk
based decision making, with emphasis on better risk
characterizations and on reducing uncertainties in risk
assessment. The risk characterization should incorpo-
rate more realistic exposure scenarios, identify popula-
tions at high risk especially children, provide where
feasible quantitative analyses of uncertainty and sensi-
tivity analyses, and include estimates of risks from expo-
sures to chemical mixtures. Risk characterization and
comparison of DBFs and microbes, the final product of
the risk assessment, will also provide perspective on
assumptions and areas of uncertainty in the risk and
exposure assessments.
1. How can we characterize the risk posed
by exposure to specific and multiple or
complex mixtures of DBFs in drinking
water?
State of the Science
Current approaches for characterizing the risks associ-
ated with exposures to D/DBPS in drinking water con-
tain many assumptions and uncertainties. Many of these
problems relate to deficiencies in the underlying scien-
tific data bases for individual contaminants and mixtures
and the methods and models used in risk assessment.
As mentioned before, much of the uncertainty in con-
ducting risk assessments arises from not knowing what
DBFs are formed and from a lack of toxicity data includ-
ing inadequate human data, insufficient understanding
of the mechanisms of toxicity, extrapolation of animal
data to human and lack of interactions data for mixtures.
Because of this lack of information, risk estimates for
DBFs are only possible for a limited number of individual
contaminants, and only qualitatively for disinfected wa-
ters from epidemiology studies. Consequently, EPA nas
developed a phased approach for characterizing the
risks posed by exposure to DBFs. First, single chemical
assessments will be conducted using improved meth-
ods for both qualitative and quantitative assessments of
human risks associated with exposures to drinking wa-
ter. Particular emphasis will be given to addressing risks
4-12
-------
to children, pregnant women and elderly. In addition,
EPA will evaluate previously conducted epidemiology
studies to provide better estimates of actual risks based
on existing information, and to assist in designing future
epidemiology studies, if warranted. In the second phase,
assessment efforts will use the information on dose and
toxicity for single chemicals to predict risk associated
with definable mixtures of agents, and to develop meth-
ods to assess risk based on generalizations across
classes of DBFs.
Much scientific progress has been made in the develop-,
ment of alternate models for estimating noncancer risks
levels for single chemicals. Traditionally, EPA risk as-
sessments have been based on the most sensitive
endpoint and data set unless it could be demonstrated
that this endpoint was not relevant to human toxicity.
Alternate approaches to analyzing data are being inves-
tigated to calculate risk estimates that use more of the
available data and result in a risk assessment in which
there is greater confidence both statistically and biologi-
cally. Because these approaches use more of the avail-
able data, they provide risk estimates that are closer to
the actual risk. Because the methods incorporate vary-
ing endpoints and severity, some expression of the risk
of adverse health effects from chemical exposures in
excess of the MCLG can be developed. Depending on
the nature of the data available, estimates using these
models can provide incidence rates for any of the effects
occurring in an exposed population. However, these
models have been developed and tested using empirical
data sets only and applied to a few well-studied pesti-
cides. The application of these models for developing
drinking water standards has not been tested nor com-
pared with estimates developed using current ap-
proaches.
Additionally, EPA is revising the cancer risk assessment
guidelines to place greater emphasis on route of expo-
sure and mode of action when assessing the carcinoge-
nicity of chemicals. The revisions will also recommend
chemical-specific determinations of whether the dose
response is linear or nonlinear, thus providing for a
different quantitative approach than the no-threshold
linearized multistage model.
As mentioned in the health effects research, EPA con-
ducted a workshop on scientific considerations for con-
ducting epidemiologic studies for cancer and exposure
to chemical by products of drinking water disinfectants.
The participants concluded that current studies are in-
sufficient to conclude that the reported associations are
causal or provide an accurate estimate of the magnitude
of human risk. In addition to the recommendations dis-
cussed previously, the workshop participants suggested
reanalysis of previously conducted interview-based case
control studies using improved exposure estimates and
analytical methods to determine the validity of these
risks and to address confounding factors and bias not
adequately excluded in previous reports such as Morris,
et al. Reassessment of existing data will also assist in
the design of feasibility studies for cancer and reproduc-
tive/developmental effects and determination if full-scale
studies are warranted by addressing systematic errors
and biases of previous studies.
In addition to applying improving risk assessment meth-
ods for individual DBPs, advancements are being made
in risk assessment methods for mixtures. Typical human
exposures to DBPs are from low doses of multiple
chemicals from multiple routes and can be either con-
tinuous, episodic, chronic, subchronic or acute expo-
sures. However, most of the available laboratory toxicity
data provide information on single chemicals or binary
pairs rather than on the mixture as a whole. Additionally,
animal laboratory studies on whole mixtures have been
hampered by the difficulty of developing representative
concentrates of DBP mixtures. Therefore, established
drinking water standards are based on risk assessments
for individual chemicals and epidemiologic data. The
Agency has not yet characterized risks from interactions
among pollutants and their consequent effects on hu-
man health. The movement toward complete risk char-
acterizations incorporates a need to realistically de-
scribe exposure scenarios at various drinking water
facilities and thus to include the risks from exposures to
chemical mixtures.
However, for any mixture of concern, data are generally
insufficient to characterize the risks: not all components
of the mixture will be known, the proportions of the
known components will be uncertain, interaction effects
data on combinations of the components will be sparse,
and epidemiologic data on human health effects will be
rare. Currently, EPA is developing drinking water feasi-
bility studies that will provide better methods for identify-
ing and quantifying mixtures components, improve
sample preparation and study design for animal testing,
and improve the analytical tools needed to use human
data.
Despite these limitations, significant advances have been
made in theoretical development and application of risk
assessment methods for characterizing the risks from
complex mixtures. These methods use data from single
Chemical components or similar mixtures to generalize
across classes of chemicals or mixtures. If studies can-
not be designed to adequately address the issues de-
scribed above, then the application of these methods
along with improved exposure data and health informa-
tion on simple mixtures can be used as alternative
methods for estimating risks from more complex mix-
tures.
In addition to specific drinking water projects, EPA is
conducting research in other problem areas to develop
scientifically valid approaches for implementing PBPK
models and sensitivity analysis in mixtures risk assess-
ments. This research is aimed at identifying those pa-
rameters which most influence the dose response for a
chemical or mixture. Research is also being conducted
that will test approaches for assessing risks for complex
mixtures using both observational data on the mixture
itself from existing human studies and data on individual
4-13
-------
components of the mixture. These methods may have
application to drinking water mixtures once validated.
The Agency has recently released a risk characteriza-
tion policy that addresses the need for more informative
and consistent approaches for communication of Agency
risk assessments to decisions makers and the public.
The risk characterization provides an evaluation of the
assumptions, uncertainties, and selection of studies and
models used in the risk assessment. Development of a
risk characterization will be required for every Agency
risk assessment.
Research Topics and Priorities
a. Characterizing risks of individual DBFs
In order to characterize the risks associated with expo-
sure to disinfectants and disinfectant by-products, quan-
titative risk assessment models that have been devel-
oped for dose-response modeling need to be applied to
both existing human data and animal laboratory data.
Estimates developed by improved methods need to be
compared against existing values in order to address
uncertainties and assumptions. These projects are high
priority because they will be used to establish MCLGs,
provide information needed for characterization of risks
for each chemical, and provide information for conduct-
ing cost and benefit analysis.
RA.D.1—Cancer risk assessments (Ongoing)
Reanalyze cancer risk assessments for D/DBPs
including chloroform, bromodichloromethane,
dibromochloromethane and bromoform, di- and
trichloroacetic acid, chloral hydrate and bro-
mate using the revised Agency cancer risk as-
sessment guidelines and improved dose-re-
sponse modeling where appropriate. Reassess-
ments of cancer dose-response data and risk
characterizations have been completed for DCA
and chloroform applying the new cancer guide-
lines. A workshop was conducted by I LSI to
discuss these case studies and an independent
expert peer review panel is evaluating case
studies.
Priority: High
RA.D.2—Cancer combination study for bro-
mates Case study examining the validity of
combining bromate data sets prior to modeling
and calculating a pooled slope factor. Also in-
cludes the modification of risk assessment soft-
ware to support computer program for use in
developing other DBP cancer estimates. Sev-
eral papers have been published in peer re-
viewed journals. In addition this method has
been presented in the proposed cancer guide-
lines and as such has undergone extensive
peer review and public comment as an alterna-
tive statistical approach for assessing dose-
response data. Case study for bromate has
b.
been completed and a journal publication is
being developed.
Priority: High
RA.D.3—Noncancer risk assessments (On-
going) Prepare full risk assessment for disinfec-
tants and DBPs, including di- and trichloroacetic
acid, bromo and dibromochloroacetic acid, four
THMs, chlorine, chlorine dioxide, chloramine,
chlorite and chloral hydrate. Incorporate bench-
mark and categorical regression analyses. A
benchmark for TCA has been completed and
presented at the Society of Toxicology meetings
in March 1997. A journal article is in prepara-
tion. Additional assessments will be conducted
on high priority Stage 1 and Stage 2 DBPs as
newer data become available. In addition pre-
liminary assessments will be conducted on newly
identified DBPs to identify data gaps and re-
search needs.
Priority: High
RA.D.4—Risk characterization Develop risk
characterizations for selected disinfectants and
DBPs which implement the revised Agency
policy. Risk assessments for chlorine, chloram-
ine, chlorine dioxide and chlorite, THMs, HAAs
and chloral hydrate, and bromate have been
developed; however, the uncertainties and as-
sumptions associated with those risk estimates
have not been fully characterized. Include the
development of a risk characterization model for
estimating variance in uncertainty factors ap-
plied to cancer and noncancer estimates; par-
ticular emphasis will be given to identifying and
characterizing risks to children and higher-risk
subpopulations. This will be particularly useful
in comparing cancer estimates (i.e. upper-bound
vs maximum likelihood estimates). In prepara-
tion for the final promulgation of Stage I DBP
Rule a risk characterization is being developed
for chlorine and chloramine as well as several
other DBPs.
Priority: High
Characterizing risks from chlorinated
waters
Although many of the epidemiplogic studies have
methodologic problems or systemic biases that limit the
interpretation of results, the application of consistent
statistical tools and improved methods and data on
exposure to chlorinated water or DBPs will provide
greater understanding of the reported risks. Reassess-
ment of existing data will address confounding factors
identified in earlier studies and assist in the design of
feasibility studies for cancer and reproductive/develop-
mental effects and determination if full-scale studies are
warranted. These studies are high priorities because
they are aimed at reducing uncertainties in cancer and
reproductive assessments and providing more realistic
4-14
-------
estimate of actual risks as well as a common methodol-
ogy for evaluating risks associated with chlorinated wa-
ter. In addition, where possible quantitative estimates
will be used for comparative risk analyses with risks
from pathogens.
RA.D.5—Evaluate newer epidemiologic stud-
ies (Ongoing) Four new studies evaluating the
possible association between drinking water and
adverse health outcomes have been conducted
since the publication of the meta-analysis of
chlorinated drinking water studies. A review of
these studies and data using methods/ap-
proaches applied to earlier studies is being con-
ducted to assess their impact on previous find-
ings. The results of this analysis will be included
in the meta-analytical report currently being de-
veloped and included in the peer review work-
shop.
Priority: High
RA.D.6—Assessment of previously con-
ducted studies (Ongoing) The objective of this
effort is to develop and apply consistent statisti-
cal tools and improved methods for analyzing
existing epidemiologic studies and data. Pri-
mary focus is to address confounding factors
and biases identified in earlier studies and to
assist in the design of future human studies.
Reanalysis of a previous meta-anaiytical report
is being conducted using several different meta-
analytical parameters not previously applied. To
date, previously published epidemiologic stud-
ies of total and various site-specific cancers,
including bladder and rectal cancers, have been
reviewed and re-evaluated as to their suitability
for providing a valid summary estimate of rela-
tive risks. A comprehensive re-evaluation of the
application of meta-analytical techniques to these
data has been completed. A combination peer
review/public participation workshop will be con-
ducted to evaluate the technical merit of this
report and implications for future epidemiologic
studies and assessments in this area. The work-
shop will be conducted in the Fall of FY 97.
Priority: High
RA.D.7—Identify ongoing cancer studies(On-
going) Identify cohort studies of dietary and
case control studies of cancer risks recently
planned or compiled or being conducted in ar-
eas with water exposures of interest. This infor-
mation will assist in determining the need for
future studies, targeting geographic areas for
future studies and study design.
Priority: High
c. Methods and models to characterize
risks from mixtures
Scientific advances in the development of innovative
risk assessment tools in the area of study design, statis-
tical methodology and computers models have been
developed to estimate health risks from exposures to
chemical mixtures.
For any mixture of concern, it is not possible to test all
combinations of the components; therefore, risk asses-
sors need methods that can be used to generalize risk
estimates from the data that do exist. With this goal in
mind, several feasibility studies have been designed to
test their application to drinking water. Included in this
research are models that characterize interactions of
chemicals (i.e., synergism, additivity, or antagonism).
These studies will provide estimates based on realistic
exposure scenarios i.e., multiple chemicals, multiple
routes. Another expected result from these studies will
be the development of experimental design methods for
assessing risks of multiple chemicals; currently, much of
this research cannot be conducted using laboratory
studies because of the large number of test animals
needed.
These methodologies, if successfully developed, would
have enormous impact on the evaluation of risk from
exposure to multiple chemicals at drinking water treat-
ment facilities and for source water quality studies.
These studies are high priorities because they can
provide better methods for quantifying the risks from
mixtures, reduce the uncertainty in the current risk esti-
mates for mixtures, and provide a framework for esti-
mating relative risks associated with drinking water.
RA.D.8—Characterization of interactions for
mixtures of DBFs Develop procedures and
conduct studies to extract, evaluate and sum-
marize interaction data from a number of stud-
ies and combine the results across studies and
endpoints to define the type of interaction i.e.,
additivity, synergism or antagonism. This re-
search involves the combining of risk estimates
for mixtures of DBPs. It involves both biological
and statistical evaluations of mixture data. Is-
sues include, but are not limited to, similarity of
mechanism of action across DBPs, incorpora-
tion of interaction data into risk estimates, ap-
propriateness of combining data, and the analy-
sis of variability and uncertainty of risk esti-
mates.
Priority: High
RA.D.9—Threshold studies for D/DBPs Re-
search is being conducted to build a dose-
4-15
-------
response plane using experimental data from a
mixture of trichloromethane, bromodichloro-
methane, dibromochloromethane and trichlo-
romethane under the assumption of additivity
for the single components. For any given point
on this plane, the threshold for an adverse effect
can be estimated for the mixture. Health effects
data for the mixture are being developed by
EPA and will be used to determine if the whole
mixture response is greater or less than the
additive, and also to test study design. Effort
includes development of computer algorithms,
testing of epidemiologic data and the develop-
ment of an optimal study design for laboratory
studies of mixtures. Goals include the detection
of interaction effects, establishment of thresh-
olds for the mixtures tested, comparison of chlo-
rination with ozonation DBF mixtures relative
toxicity by testing proportions of THMs reflecting
actual treatment formation. Work is being con-
ducted in two animal system models thereby
allowing for the investigation into interspecies
scaling factors.
Priority: High
RA.D.10—Use of QSAR model to estimate
risk for single components and classes of
compounds within a mixture Estimate an ad-
verse effect level for DBFs with limited or no
health effects data using QSAR. EPA data on
mixtures of DBPs formed following various treat-
ment trains such as chlorination plus
chloramination could be used to estimate Low-
est Observed Adverse Effect Levels (LOAELs)
or genotoxic potential for single components
and whole mixtures. Compare predicted esti-
mates to measured values to confirm the accu-
racy and precision of the model. Primary focus
for this effort is to evaluate and apply a system
that can be used in conjunction with other short-
term tests for estimating effects and relative risk
levels. Additionally, efforts are being directed to
employing this model in prioritizing and select-
ing newly identified DBPs of potential concern.
Research is focused mainly on Stage 2 and
future DBP assessment efforts. Currently, thirty
DBPs are being evaluated.
Priority: High
d. Methods and models to compare risks
Comparative risk assessment is a relatively new initia-
tive in risk assessment that has not been well defined.
The result is a variety of definitions, approaches and
applications. For drinking water, the focus is to develop
a comparative risk model to measure and compare the
known and potential health risks that might result from
exposure to multiple stressors transmitted from the same
drinking water source. For example, acceptably treated
drinking water serves as a source of chemical (DBPs),
biological (microbial), and physical (distribution system
characteristics) stressors. Exposure to these stressors
may result in a variety of adverse health outcomes
including acute and chronic gastrointestinal illness, can-
cer, liver toxicity, and reproductive and developmental
disorders. Despite scientific advances in identifying and
quantifying the risks associated with any individual stres-
sor, an appropriate methodology has not been devel-
oped that would allow us to assess these different risks
in a similar way in order that they can be compared in a
meaningful way. In addition, there is a paucity of data on
the nature and magnitude of risk from these stressors.
RA.D.11—Comparative risk analysis Current
comparative risks models for drinking water
weigh the outcomes of microbial exposures to
that of cancer from selected DBPs. Research
needs to be conducted on a variety of models
and methods that will allow for the development
of a comparative risk assessment model which
addresses multiple outcomes (e.g., cancer, de-
velopmental, reproductive, neurotoxic effects),
their impacts and costs. This research should
focus on the risk analysis and risk reduction
benefits derived from minimizing exposures that
would result in adverse outcomes other than
cancer. Data will be developed for use in the
regulatory modeling effort described in Chapter
III. A comparative risk framework and strategic
model has been developed for DBPs and is
currently being validated and reviewed. SAB
formal review of this approach is scheduled for
the Spring of FY 98.
.Priority: High
Risk Management Research
How effective are various treatment processes in
minimizing and controlling the formation of DBPs?
State of the Science
Naturally occurring materials, such as humic and fulvic
acids and bromide, which are present in many surface
water sources react with chemical disinfectants to pro-
duce disinfection by-products. Factors affecting the for-
mation of disinfection by-products include the nature of
the source water and content of precursor materials, the
water temperature and pH, and conditions under which
the disinfectant is used, such as the concentration,
contact time, point of addition, and the residual main-
tained. Control strategies must consider these factors
because they affect both the formation of specific by-
products and other potential risks. For example, raising
the water pH to control lead corrosion may decrease the
formation of haloacetic acids, but it increases
trihalomethane formation and results in a species of
chlorine that is a less effective disinfectant.
Several options are available for the control of disinfec-
tion by-products: 1) removal of precursor material before
disinfection, 2) changes in the disinfection process or
4-16
-------
use of a disinfectant that will minimize the formation of
selected by-products, 3) removal of disinfection by-prod-
ucts after they are formed. The various treatment tech-
nologies are very similar with regard to their overall
effect on disinfection by-product formation. Disinfection
by-products tend to increase in time so that moving .the
point-of-disinfection to the end of the water treatment
process is an inexpensive modification that can reduce
the formation of by-products without substantially in-
creasing microbial risks, provided an adequate disinfec-
tion contact time is maintained. Appropriate coagulation
and clarification can effectively remove precursors, and
this treatment can be adjusted for the enhanced removal
of natural organic matter, thereby minimizing by-product
formation potential.
Numerous studies have been conducted on DBF forma-
tion and control. Early bench-scale studies conducted
by EPA, AWWARF and others examined the factors
affecting the formation of trihalomethanes (THMs), the
kinetics of the reaction, and means of assessing precur-
sor, i.e., THM formation potential. Numerous field-scale
studies were conducted examining modifications to the
practice of chlorination, alternatives to chlorine (chloram-
ine, chlorine dioxide and ozone), and removal of precur-
sor, principally by granular activated carbon (GAG).
More recent research conducted by AWWARF and EPA
has included haloacetic acids (HAAs), total organic ha-
lide (TOX), chloral hydrate (CH), haloacetonitriles (HANs)
and other DBPs as targets for removal. Isotherms and
kinetic studies are being conducted to provide inputs to
GAC models and improved GAG models were devel-
oped. Numerous pilot- and full-scale GAC studies are
being conducted in the field, including on-site reactiva-
tion of spent GAC. Other means of precursor control are
currently being studied, including membranes, enhanced
coagulation and, more recently, biological filtration. The
ICR will collect bench- and pilot-scale data characteriz-
ing GAC and membrane performance on high TOG
waters throughout the U.S.
Results from the research to date indicate that en-
hanced softening may be effective in removing Total
Organic Carbon (TOG) which is a measure of the pre-
cursor material for disinfection by-products. Studies have
also shown that biological filtration may be an effective
tool for removing both precursors and protozoa simulta-
neously. It was also learned that control of pH in biologi-
cal and conventional treatment can minimize the forma-
tion of the five haloacetic acids to be regulated. Mem-
branes have shown promise for removing precursors
and pathogens simultaneously and may be especially
useful in small-system applications. Although biological
treatment using ozone is effective for removing precur-
sors and controlling protozoans, it has by-products of its
own that must be evaluated.
Critical questions that must be answered are as follows:
To what extent can surrogates be used to indicate
effectiveness of treatment, e.g., TOX, TOG; and, what
operational concerns pertain to the different technolo-
gies? The projects selected for research in this section
are intended to define the variables that influence the
formation of DBFs. These projects also address con-
cerns of the simultaneous reduction of risks from micro-
bial and chemical contaminants.
Bench- and pilot-plant studies will initially be conducted
to assess different technologies and approaches. This
will be followed by field studies (see RM.M.5) to evalu-
ate the more promising technologies. In general, these
field studies will be conducted at sites where these
technologies are being utilized by certain utilities, such
that we can conduct parametric evaluations of the im-
portant operating parameters and make comparisons
with full-scale operations.
Research Topics and Priorities
a. Effectiveness of different treatment
processes in reducing DBP precursors
(as a function of water quality and
system size)
Research is needed on the effects of enhanced coagu-
lation and enhanced softening on the operation of filtra-
tion processes. Recent EPA studies indicated the pH of
filter operation was critical to aluminum solubility and
filter run time, and that membrane materials are fouled
by different types of organic substances. For utilities
facing control of both precursors and pesticides, assess-
ment of their control by oxidative and biological pro-
cesses is needed. The causes and prevention of mem-
brane fouling, and the reliability of scaling up from
bench-scale membrane studies need to be evaluated.
Projects that develop treatment and operational criteria
for technologies for DBP precursor removal and also
assess their efficacy to reduce microbial or chemical
contaminants were all considered a high priority. The
membrane scale-up and fouling research was assigned
a medium priority because results from a related
AWWARF study and data provided by utilities as the
result of their ICR membrane studies may prove ad-
equate.
RM.D.1—Enhanced softening for precursor
and pathogen removal Assess TOC removal
by lime softening. Study filter operations associ-
ated with enhanced softening. Assess filter run
time, particulate and pathogen control, and cor-
rosion implications downstream of enhanced
softening.
Priority: High
RM.D.2—Effects of ozonation and biofiltra-
tion for control of precursor, pathogens and
for pesticide removal
a.—Control of precursor, pathogen and pes-
ticide removal Evaluate ozonation and biofil-
tration for control of precursors and pathogens.
Assess oxidation and biodegradation of precur-
4-17
-------
sor materials. Include control of pesticides by
ozonation and biodegradation.
Priority: High
b.—Effect of pH on ozonation and enhanced
coagulation Evaluate the effect of pH on
ozonation and enhanced coagulation. Recent
EPA studies indicated levels near the proposed
Stage 2 MCLs for TTHM and HAAs could be
met with these processes. It has been found
that low pH enhances coagulation and mini-
mizes bromate formation. High pH promotes
oxidation by hydroxyl radicals. Study variations
in pH since pH affects aluminum solubility and
filter run time.
Priority: High
RM.D.3—Analyze ICR data from GAC, mem-
brane bench and pilot studies Analyze data
collected from ICR activities involving bench
and pilot studies using GAC and membrane
technology. Cost and performance data will be
developed from this project.
Priority: High
RM.D.4—Removal of DBP precursors by GAC
and membranes ICR bench and pilot studies
will provide performance and cost data for dif-
ferent water quality conditions.
a—Evaluation of membrane reliability Evalu-
ate membrane reliability for multiple contami-
nant removal using bench and pilot studies.
Conduct a pilot study at a full-scale water plant
to evaluate nanofiltration.
Priority: Completed
b—Evaluation of GAC Evaluate GAC for mul-
tiple contaminant removal at bench and pilot
scale.
Priority: Completed
RM.D.5—Membrane scale-up and fouling
Conduct bench- and pilot-scale membranes stud-
ies in parallel to assess the appropriateness of
bench-scale membranes to predict larger-scale
performance. Examine techniques to limit mem-
brane fouling.
Priority: Medium
b. Effectiveness of using different
disinfectants in limiting DBP formation
Research is needed on the use of biological treatment
for the control of ozone DBPs, and the control of Assimi-
lable Organic Carbon (AOC) and Biologically Degrad-
able Organic Carbon (BDOC). AOC and BDOC are
measures of microbial nutrients that could promote dis-
tribution system regrowth. Questions to be answered
include: to what extent can chlorite/chlorate by-products
of CI02 be removed or formation limited; to what extent
can bromate formation be limited while using ozone;
and, for which water qualities are combinations of ozone/
chloramines appropriate for primary residual disinfec-
tions without increasing bacterial growth? Because ozone
promotes regrowth of organisms in distribution systems,
nutrient control is very important. Improving the under-
standing of how, and the extent to which, ozonation by-
products and AOC formation can be controlled is con-
sidered a high priority for assessing the risks and the
benefits associated with the use of ozone.
RM.D.6—Ozone by-product formation and control
Evaluate technologies for control of AOC and
BDOC. Examine factors affecting formation of
ozone DBPs, AOC and BDOC by ozonation and
control by biofiltration. Assess bromate forma-
tion. This project relates to RM.M.14 and
RM.M.26.
c.
Priority: High
Small systems technologies for
precursor and DBP control
Small systems pose a special problem because of the
economic limitations on most small communities. Devel-
oping inexpensive, low-maintenance technologies to con-
trol disinfection by-products is considered a high priority.
RM.D.7—Membranes/advanced oxidation and
other technology combinations Preliminary
results indicate that a combination of ultrafiltra-
tion membranes with advanced oxidation pro-
cesses results in reduction of organics, lower
DBP formation, and less mutagenicity than typi-
cal treatment trains with post-chlorination. Con-
duct additional investigations of these ap-
proaches individually and in combination for
oxidation processes as UV, UV irradiated TiQ,,
iodinated resins, hydrogen peroxide, and on-
site mixed oxidant generators.
Priority: High
4-18
-------
Table IV-2. Research Priorities for Health Effects for Disinfection By-Products
Research Topics Proposed Projects
Priorities
DBP Epidemiology
a. Development/application of improved tools for field research
b. Feasibility/full-scale studies
DBP Toxicology
a. Hazard identification and dose-response
b. Pharmacokinetics and mechanisms of action
c. DBP mixtures
HE.D.1—Improving estimates of residential DBP *
exposures in epidemiology studies
HE.D. 2—Improving measures of biologic effect: High
Field evaluation of biomarkers
HE.D.3—Improving methods for managing health and High
exposure data
HE.D.4—Feasibility studies: Cancer High
HE.D.5—Feasibility studies: Repro effects High
HE.D.6—Full-scale studies: Cancer and repro **
HE.D.7—Cancer dose-response studies High
HE.D.8—Repro/developmental effects screening studies High
HE.D.9—Neurotoxicity studies Medium
HE.D.10—Immunotoxicity studies Medium
HE.D.11—Pharmacokinetic and mechanistic High
research—cancer
HE.D.12—Pharmacokinetic and mechanistic High
research—reproductive effects
HE.D.13—Mixtures feasibility study High
HE.D.14—Toxicologic evaluation of mixtures **
HE.D.15—Mutagenicity screening studies of mixtures High
HE.D.16—DBP interactions Medium
' See projects under part 2.b. of DBP Exposure section.
" Priority depends upon outcome of feasibility studies.
Table IV-3. Research Priorities for Exposure to DBPs
Research Topics
Proposed Projects
Priorities
DBP Methods
a. Additional methods or method improvements needed to
implement the Stage 1 DBP rule
b. Methods needed for Stage 2 DBP rule and the longer term
DBP Exposure
a. Characterize types and concentrations of DBPs formed from
different disinfectants and combinations
b. Evaluate factors that affect exposure levels, and assess
human exposures to DBPs
EX.D.1—Low-level bromate measurement High
EX.D.2—Improve method for haloacetic acids High
EX.D.3—Expand quality control for TOG, evaluate High
new TOG methods
EX.D-4—Low-level CIO2 measurement Medium
EX.D.5—Real-time monitoring for disinfectant residuals Low
EX.D.6—PE studies for DBPs and disinfectants High
EX.D.7—Methods for peroxides Medium
EX.D.8—Real-time, in-plant monitoring of DBPs Medium
EX.D.9—Improved method for aldehydes Medium
EX.D.10—Identify new disinfection by-products from High
alternate disinfectants
EX.D.11—Methods for nonvolatile DBPs High
EX.D.12—DBP changes in distribution systems Medium
EX.D.13—DBP interactions with foods Medium
EX.D.14—Exposure to DBPs through showering Medium
EX.D.15—Markers of DBP exposure Medium
EX.D.16—Models of DBP exposure High
EX.D.17—Exposure as a function of population Medium
distribution
EX.D.18—Tap water consumption High
4-19
-------
Table IV-4. Research Priorities for Risk Assessment for DBFs
Research Topics
Proposed Projects
Priorities
a. Characterizing risks of individual DBPs
b. Characterizing risks from chlorinated waters
c. Methods and models to characterize risks of DBP mixtures
d. Methods and models to compare risks
RA.D.1—Cancer risk assessments High
RA.D.2—Cancer combination study for bromates High
RA.D.3—Noncancer risk assessments High
RA.D.4—Risk characterization High
RA.D.5—Evaluate newer epidemiologic studies High
RA.D.6—Assessment of previously conducted studies High
RA.D.7—Identify ongoing cancer studies High
RA.D.8—Characterization of interactions for mixtures High
of DBPs
RA.D.9—Threshold studies for D/DBPs High
RA.D.10—Use of QSAR model to estimate risk for High
single components and classes of compounds
within a mixture
RA.D.11—Comparative risk analysis High
Table IV-5. Research Priorities for Risk Management of DBPs
Research Topics
Proposed Research
Priorities
a. Effectiveness of different treatment processes in reducing
DBP precursors (as a function of water quality and system size)
b. Effectiveness of using different disinfectants in limiting
DBP formation
c. Small systems technologies for precursor and DBP control
RM.D.1—Enhanced softening for precursor and High
pathogen removal
RM.D.2—Effects of ozonation and biofiltration for
control of precursor, pathogens, and for pesticide
removal
a. Control of precursor, pathogen, pesticide removal High
b. Effect of pH on ozonation and enhanced High
coagulation
RM.D.3—Analyze ICR data from GAC, membrane High
bench and pilot studies
RM.D.4—Removal of DBP precursors by GAC and Completed
membranes
RM.D.5—Membrane scale-up and fouling Medium
RM.D.6—Ozone by-product formation and control High
RM.D.7—Membranes/advanced oxidation & other High
technology combinations
4-20
•&U.S. GOVERNMENT PRINTING OFFICE: 1998 -650-001/80182
------- |