Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
l Executive Summary The quality of a public water supply, although quite acceptable when it leaves the treatment plant, may deteriorate before it reaches the user. This may occur as a result of either chemical or biological transformations or by loss of integrity of the system. Serious sources of potential contamina- tion arise when nonpotable water enters the system as a result of cross- connections, backflow, back-siphonage, or during repairs. Although it is not the purpose of this report to review the optimum engineering design, construction, and operation of distribution systems, it is important to recognize that adverse health effects may result from in- adequate attention to any of these areas. Chapters II, III, IV, and V of this report contain reviews of the factors and conditions associated with retention of water quality in the distribu- tion system and the committee's recommendations for control procedures and future research. In Volume 1 of Drinking Water and Health, the Safe Drinking Water Committee examined the health effects associated with microbiological, radioactive, particulate, inorganic, and organic chemical contaminants found in drinking water. In Volume 3, it considered another selected group of chemical contaminants. Chapters VI and VII of this volume con- tain evaluations of the health effects of additional organic and inorganic contaminants. 1
2 DRINKING WATER AND HEALTH DISTRIBUTION SYSTEMS Chemical Effects and Water Quality Changes in the chemical quality of water in the distribution system can result from corrosion, deposition, leaching, and reactions involving water treatment chemicals in the system. The solubility and kinetic factors that determine whether constituents in water will deposit on pipe walls or whether the materials used in the conveyance system will partially dissolve or corrode into the water are examined in this volume. Corrosion A number of water quality indexes have been used to predict whether water will produce corrosion of materials used in the distribution system. These indexes are used not only as criteria for treatment control, but also as a guide to selection of materials. The committee's review and evalua- tion of these indexes can be found in Chapter III. The nature and control of the corrosion process is also described in Chapter III. The chemical and electrochemical reactions involved in cor- rosion cause deterioration of water quality, particularly with respect to such metals as zinc, cadmium, and lead. The economic impacts of corro- sion in water distribution systems are not a part of this study. Nonetheless, they are an important consideration since they provide an incentive for reducing corrosion, which ultimately affects the water to which the con- sumer is exposed. The economic incentive for the use of corrosion in- hibitors may be counterproductive in terms of human health unless the in- hibitors are selected with a full awareness of their possible toxicity. A variety of microorganisms are known to be involved in the corrosion process. They may be heterotrophic or autotrophic and grow under aero- bic or anaerobic conditions. The nature of the biologically mediated cor- rosion process and the organisms involved are described in Chapter IV. Considerable further attention must be directed toward this subject in or- der to elucidate the nature and consequences of microbial action on the . . . pipes in site. Leaching In contrast to corrosion, which is an electrochemical phenomenon, leaching is a process of dissolution governed by solubility and kinetics. The quality of the water influences the leaching of some materials, such as asbestos/cement (A/C) pipe.
Executive Summary 3 Deposits Following water treatment. chemicals can deposit in distribution systems and affect the quality of water distributed to the consumer. Among the elements manifesting this behavior, iron and manganese are receiving most attention, primarily because of esthetic and economic considera- tions. Their effects on human health are not of major importance; however, iron and manganese suspensions are associated with the growth of microorganisms in the distribution system. Furthermore, humans may be exposed to other elements that may be associated with iron and maganese deposits and their resuspension as well as to chemicals added for their control. The control of manganese and iron in water supply systems has been widespread. Three principal methods have been used: oxidation (followed by precipitation and filtration), ion exchange, and stabilization by dis- persing agents. These control measures may alter the concentration of other chemicals that coprecipitate or sorb with them, including organic chemicals, and influence the concentration of bacteria in the system. Of special interest are the effects of stabilizing agents such as silicates and polyphosphates. Polyphosphates have the potential for adversely af- fecting human health, especially when complexed with.manganese. Just as iron and magnesium can be maintained in solution by the addition of these dispersing agents, so can other heavy metals. The extent to which their use increases human exposure and- adverse health effects is not known. Water Treatment Chemicals The data suggest that concentrations of trihalomethanes continue to in- crease in the distribution system when both chlorine and organic precur- sors are present. It is probable that the nonvolatile reaction products of humic material with chlorine also increases in distribution systems, al- though no data have been obtained from systems in operation. Pipes and Linings A/C pipe was originally believed to be resistant to deterioration. Recent findings, both in the laboratory and in field tests, have shown that A/C pipe, like other pipe materials, can be vulnerable to the corrosive action of water. Under certain circumstances, asbestos fibers can be released from the pipe into water supplies. Control measures for limiting the release of these fibers are discussed in Chapter III. The data do not provide an ade
4 DRINKING WATER AND HEALTH quate basis from which to estimate health consequences from ingesting asbestos fibers. Extensive epidemiological studies and laboratory tests now under way at several locations are expected to yield useful infot~a- tion on this subject. Plastic pipes used for conveying potable water are made of polyethylene (PE), polyvinyl chloride (PVC), including chlorinated polyvinyl chloride (CPVC), and acrylonitrile-butadiene-styrene (ABS). Although plastic pipe is composed principally of polymerized organic compounds, there can also be residual unpolymerized monomer present in low concentrations. This factor is especially important in PVC pipe, since vinyl chloride monomer is carcinogenic. In general, water quality is not a major factor in plastic pipe leaching, in contrast to the definite effect of water quality on the cor- rosion of metal or A/C pipe. Nonmetallic linings are commonly used in potable water distribution systems to prevent corrosion of underlying metal or other components. The most common lining materials are coal tar, petroleum asphalt, vinyl, epoxy, or some combination thereof. Laboratory and field studies have shown that some of these materials, especially coal tar compounds, can be released into the water. Of concern is the content of polynuclear aromatic hydrocarbons (PAH's), several of which are known carcinogens. Some lin- ing materials can reach the drinking water through leaching processes or through physical deterioration, which results in the release of small par- ticles from the linings. Higher concentrations appear to leach from coal tar than from asphalt; however, the tests leading to this conclusion were not definitive. PAM's have been found in nanogram and microgram per liter concen- trations in field studies of water exposed to linings in the distribution system. The compounds most commonly found in these studies have been phenanthrene and/or anthracene. Tetrachloroethylene has recently been found in vinyl-lined A/C pipe. Studies are under way to determine con- centrations of this chemical in pipes of different ages receiving varying amounts of usage. There is a need for more research into the effect of linings on water quality. Biological Effects and Water Quality Living organisms may enter the distribution system through raw water receiving insufficient treatment, from in-line reservoirs, or from imperfec- tions or breaks in the pipeline network. The microorganisms found in the distribution system and their effect on water quality are described in Chapter IV.
Executive Summary 5 Microbial processes within the distribution system can significantly in- crease hydraulic roughness and corrosion of pipes. Products of microbial activity include undesirable tastes and odors. e.g., "black" water, which contains sulfide, and "red" water, which con- tains iron. Detachment of biofilms from the pipes provides a continuing inoculum into other locations within the distribution system. Microorganisms may proliferate in portions of the distribution system if the disinfectant residual is low or absent. A disinfectant residual (usually chlorine) provides a relatively effective barrier to growth of microorga- nisms in the distribution system. Distribution networks are dynamic systems. The quality of the water in them can be affected by the quality of the raw water, treatment processes, pipe materials, and the reactions that occur within the pipes. Implications For Human Health Outbreaks of acute diseases associated with contaminated drinking water are reported voluntarily to the Centers for Disease Control (CDC) by state health departments or to the Health Effects Research Laboratoty of the Environmental Protection Agency (EPA) by state agencies having respon- sibility over water supply. In the majority of these reports, the etiologic agent is not identified. However, the clinical and epidemiological evidence suggests that most of the outbreaks are caused by infectious agents. Sources of contamination responsible for outbreaks include untreated surface water or deficient treatment of groundwater, e.g., malfunction of a chlorinator, and deficiencies in the distribution system. Examples in- clude a large outbreak of amebiasis among guests at two Chicago hotels in 1933 and one of infectious hepatitis among members of a football team in 1972. In both outbreaks, potable water was contaminated by wastewater as a result of a defect in the distribution system. Deficiencies in distribution systems were responsible for limo of water- borne outbreaks reported in the United States from 1971 through 1978. The most common defects were indirect cross-connections, which perrnit- ted wastewater or toxic chemicals to gain access to the potable water system by back-siphonage through hoses or defects in water pipes during periods of low pressure. Less frequent defects include direct cross- connections, contamination of the system during construction, leaching of copper from the pipes, and contamination of open reservoirs. Health ef- fects associated with specific constituents of drinking water and the distribution system are provided in tabular form in Chapter V, with references to discussions in this volume and the first three volumes of Drinking Water and Health.
6 DRINKING WATER AND HEALTH The committee recommends intensifying surveillance, investigation, and reporting of outbreaks of waterborne diseases at the local, state, and federal levels. An evaluation of the health effects associated with drinking water depends' both quantitatively and qualitatively, on the completeness and accuracy of the reporting system. The committee also recommends the development of an education program focusing on the preservation of the integrity of the distribution system and emphasizing the necessity of maintaining an adequate chlorine residual throughout the system. TOXICITY OF SELECTED CHEMICAL CONTAMINANTS IN DRINKING WATER The contaminants reviewed in Chapters VI and VII were selected for one or more of the following reasons: · They are contaminants that have been identified in drinking water since the first three volumes of Drinking Water and Health were published. · Sufficient new data have become available to justify further examina- tion of contaminants evaluated in the earlier studies. · Several compounds were judged to be a concern because of potential spill situations. · They are contaminants that have been associated with drinking water distribution systems. · They are structurally related to known toxic chemicals. Thus, the descriptions of some of the contaminants are limited to data generated since the first three volumes of Drinking Water and Health were published, whereas other contaminants and their health effects are evaluated in full for the first time. Several metallic ions were selected for evaluation because of their association with drinking water distribution systems, although contaminants such as lead and strontium pose prob- lems only in certain local areas. The chlorine derivatives are evaluated because of their possible use as alternatives to chlorine in the disinfection of drinking water. A review of asbestos was deferred pending the outcome of ongoing feeding studies in animals. The contaminants reviewed in this volume were selected from a list recommended for evaluation by the U.S. Environmental Protection Agency (EPA). The inclusion of each one was discussed with representatives of the EPA before a final decision was reached. Several were rejected either because the data were insufficient for a new evaluation or because there was not enough new information for a reevaluation.
Executive Summary 7 The committee evaluated the data concerning both acute and chronic exposures to selected chemicals. Information derived from studies of acute exposure provides a basis for judging health effects resulting from accidental spills of chemicals into drinking water supplies. A suggested- no-adverse-response level (SNARL) for acute exposures of 24 hours or 7 days has been calculated for compounds on which sufficient data were available. The acute exposure SNARL's are based on the assumption that lOO~o of the exposure to the chemical was supplied in drinking water dur- ing either the 24-hour or 7-day period. When the chemical was a know n or suspected carcinogen the potential for carcinogenicity after acute ex- posure was ignored. The calculated acute SNARL's should not be used to estimate hazards from exposures exceeding 7 days. They are not a guarantee of absolute safety. Furthermore SNARL's are based on ex- posure to a single agent and do not take into account possible interactions with other contaminants. The safety or uncertainty factors used in the calculations of the SNARL's reflect the degree of confidence in the data as well as the combined judgment of the committee members. SNARL's for chronic exposure were calculated for chemicals that were not known or suspected carcinogens on the basis of data obtained during a major portion of the lifetime of laboratory animals. An arbitrary assump- tion was made that 20~o of the intake of the chemical of concern was derived from drinking water. Therefore, it would be inappropriate to use these values as though they were maximum contaminant levels. At the time of the writing of the first and third volumes of Drinking Water and Health, the use of statistical methods and model systems to ex- trapolate data from studies of animals at high doses in order to estimate risk in humans exposed to low doses was a novel and largely untested ap- proach. When considering various approaches to this problem, the com- mittee agreed on four principles to be used as a basis for estimating risk for nonthreshold effects: · Effects in animals, properly qualified, are applicable to humans. · Methods do not now exist to establish a threshold for long-term ef- fects of toxic agents. · The exposure of laboratory animals to high doses of toxic agents is a necessary and valid method of discovering possible carcinogenic hazards in humans. · Data should be assessed in terms of human risk, rather than as "safe" or "unsafe." The methodology used by the previous Safe Drinking Water Commit- tees in projecting risk estimates have been described in Volumes 1 and 3.
8 DRINKING WATER AND HEALTH That approach stimulated considerable developmental activity in this area that in succeeding years has resulted in the appearance of a variety of statistical models for making such calculations. The regulatory agencies were quick to recognize the utility of the risk estimate approach for deal- ing with carcinogenisis data, and the concept, if not the specific models, has achieved international recognition. When preparing the third volume of Drinking Water and Health, the Subcommittee on Risk Assessment reevaluated six high-dose to low-dose extrapolation models (i.e., the dichotomous response model, the linear no-threshold model, the tolerance distribution model, the logistic model, the "hitness" model, and the time-to-tumor-occurrence model). Because of uncertainties involved in the true shapes of the dose-response curves that must be used for the extrapolation, the committee judged that a multistage model would be most useful for making the risk estimate ex- trapolations for the chemicals reviewed in these volumes. It pointed out that more confidence could be placed in mathematical models for ex- trapolation if they incorporated biological characteristics of the animal studies such as kinetics and time-to-occurrence of tumors. Although these and other concerns have tended to limit somewhat the acceptance of the various extrapolation models by the scientific community, they have led to the development of new and revised protocols for the animal studies that provide the desired information. The current Safe Drinking Water Committee has again evaluated sev- eral of the major extrapolation models in use by the regulatory agencies or by others and has made some effort to anticipate new thrusts in what could be designated as research in applied toxicology. It concluded that since the users of this volume will be likely to favor different varieties of the conventional extrapolation models or will have access to some of the newer developmental methodologies, it is premature at this stage to recommend any single approach by selecting it for calculations in this volume. Since the committee has provided the data from animal toxicity studies for each of the agents discussed in this volume or references to documents containing the data, it should be possible for readers to apply whichever extrapolation model they wish to the data to estimate risk.
