Document dDaoyz0Y9KoxnbomDmJ0LQawQ

24330 Federal Register / Vol. 49, No. 114 / Tuesday, June 12, 1984 / Proposed Rules ENVIRONMENTAL PROTECTION AGENCY 40 CFR Part 141 (OW-FRL-2514-31 National Primary Drinking Water Regulations; Volatile Synthetic Organic Chemicals AQENCY: Environmental Protection Agency (EPA). action: Proposed rulemaking. Summary: This proposed rule under the Safe Drinking Water Act (42 U.S.C. 300f el seq.) establishes Recommended Maximum Contaminant Levels (RMCLs) for the following volatile synthetic organic chemicals (VOCs) in drinking water: trichloroethylene, tetrachloroethylene, carbon tetrachloride, 1,1.1-trichloroethane, vinyl chloride. 1.2-dichIoroethane, benzene, 1.1-dichloroethylene, and pdichlorobenene. RMCLs (goals) for noncarcinogens are proposed based upon chronic toxicity data, and RMCLs (goals) for carcinogens are proposed at the zero level. VOCs that are not included in this proposal may be considered for subsequent rulemaking as appropriate. RMCLs are non-enforceable health goals which are to be set at levels which would result in no known or anticipated adverse health effects with an adequate margin of safety. This proposal is the initial stage in rulemaking for the establishment of primary drinking water regulations for the VOCs. Following this proposal, Maximum Contaminant Levels (MCLs) and monitoring/reporting requirements will be proposed when the RMCLs are promulgated. MCLs are enforceable standards and are to be set as close to the RMCLs as is feasible and are based upon health, treatment technologies, cost and other factors. Public comments are solicited on the approach to setting RMCLs as proposed in this notice as well as on the alternatives presented. Specifically, comments are requested on the following: Should the RMCLs for carcinogens be zero or a level of exposure considered to constitute a negligible incremental lifetime risk, say one in one million, based upon a conservative risk estimate calculation procedure; or should the RMCLs for carcinogens be established at the limits of analytical detection? dates: Written comments should be submitted by September 10,1084. A public hearing will be held in Washington, D.C. on August 8 and 7, 1884. if needed beginning at 0:00 am.. addressee: Send written comments to Comment Clerk, Criteria and Standards Division, Office of Drinking Water (WH-550). Environmental Protection Agency. 401 M Street, S.W., Washington, D.C. 20460. A copy of the comments and supporting documents will be available for review during normal business hours at the EPA. Room 55EB, 401 M Street, S.W., Washington, D.C. 20460. The public hearing will be held in Room 3906. EPA, 401M. St. S.W,, Washington. D.C. It is requested that anyone planning to attend the public hearing (especially those who plan to make statements) register in advance by calling or writing Ms. Ametta Davie at 202/382-7575, EPA, WH-550, 401 M St, S.W., Washington, D.C. 20460. Persons planning to make statements at the hearings are encouraged to submit written copies of their remarks at the time of the hearing. References cited on section VII will be svailable for inspection at the Drinking Water Supply Branches of EPA't Regional Offices. L JFK Federal Bldg.. Boston, MA 02203, Phone; (617) 223-6486. Jerome Healy n. 26 Federal Plaza, Room 624. New York. NY 10276, Phone: (212) 2B4-1800. Walter Andrews m. 6th k Walnut Sts., Philadelphia, PA 19106, Phone: (215] 507-0873, Bemie Samowski IV. 345 Court!and Street, Atlanta. GA 60365. Phone: (404) 681-3781, Robert Jouidan V. 230 S. Dearborn St.. Chicago, E 60604, Mona: (312) 866-6176, Joseph Harrison VI. 1201 Elm St., Dallas, TX 75270, Phone: (214) 767-2620. James Graham VII. 324 East 11th SL, Kansas City, MO 64106, Phone: (816) 374-6514, Gerald R. Force VH1. i860 Lincoln St, Denver. CO 60295, Phone: (303) 637-2731, Dean Ckaussee DL 215 Fremont St., San Francisco, CA 94105, Phone: (415) 074-8076, Leslie Ragle X. 1200 Sixth Ave., Seattle, WA 96101, Phone: (206) 442-1225, Jerry Opatz Copies of the nine draft health criteria documents will be available for a fee from the National Technical Information Service, U.S. Department of Commerce, 5285 Port Royal Road, Springfield. Virginia 22161. The toll free number it 800/336-4700; local- 703/467-4650. FOR FURTHER INFORMATION CONTACT: Contact Joseph A. Cotruvo, Ph. D,, ~ Director, Criteria and Standards Division, Office of Drinking Water (WH-550), Environmental Protection Agency, 401M Street S.W,, Washington. D.C. 20460, telephone (302) 382-7575. SUPPLEMENTARYJNFORMATION: L Statutory Requirements B. Regulatory Framework m. Background end Summary of Comments IV. VOCs in Drinking Water Occurrence of VOCa in Drinking Water Human Health Considerations --Development of RMCLs for Noncarcinogens --Development.of RMCLs for Carcinogens --Toxicology of VOCs V. RMCL Development Rationale VOCa: Regulatory Approach Regulations for Which VOCs RMCLs: Regulatory Approach VL Other Considerations for Public Comment VU. References VBL Request for Comments L Statutory Requirements The Safe Drinking Water Act (42 USC 300f, etseq.) ("SDWA" or "the Act") requires the EPA to establish primary drinking water regulations which: (1) Apply to public water systems; (2) specify contaminants which in the Judgment of the Administrator, may have any adverse effect on the health of persons: (3) specify for each contaminant either (a) maximum contaminant levels (MCLs) or (b) treatment techniques. See section 1401(1], 42 U.S.C. 300f. A treatment technique requirement would only be set if "it is not economically or technologically feasible" to ascertain the level of a contaminant in drinking water. The SDWA includes provisions for interim and revised regulations. See action 1412.42 U.S.C. 30Qg-l. Interim regulations were to be established within 160 days of enactment of the SDWA. Revised regulations are to be developed in two steps: the Agency is to establish recommended maximum ontaminant levels (RMCLs) and then establish maximum contaminant levels (MCLs) as close to the RMCLs as feasible. MCLs are to be proposed at the time of promulgation of the RMCLs. RMCLs are non-enforceable health goals, RMCLs are to be set at a level which, in the Administrator's Judgment, "no known or anticipated adverse effects on the health of persons occur and which allows an adequate margin of safety". Section 1412(b)(1)(B). The House Report on the 1974 legislation provides congressional guidance on developing RMCLa: * * * the recommended maximum (contaminant) level must be set to prevent the occurrence of any known or anticipated advene effect It must Include,an adequate margin of safety, unlese there is no safe dveehold for a contaminant In auch a case, the recommended maximum contaminant level should be aet at larqieye). C0HFID--3--N---T- ':I All CHA 007S9B subject to Protective Order In Boss v. Conoco, Inc., No. 90--^ < 14th Judicial District Cou-t Calcasieu Parish, Louisi-'"' Federal Register / VoK 49, No. 114 / Tuesday, June 12, 1984 / Proposed Rules 24331 ouse Report No. 93-1185, July 10,1974, 20. MCLs are the enforceable standards. iCLs must be set at dose to RMCLs as feasible. Feasible means "with the use f the best technology, treatment :chnique and other means, which the administrator finds are generally vailable (taking costs into onsideration)." Section 1412(b)(3). RMCLs of themselves have no impact >n public water systems or the public, iy promulgating RMCLs, no system is orced to reduce contaminants to this evei or to take other action regarding jontaminanta. RMCLs serve as goals for he Agency in the course of setting MCLs and are therefore initial steps in the MCL rulemaking that will follow, fn come cases, the MCLs will be set very close to the RMCLs: in other cases control processes or economic considerations may dictate an MCL that is not as dose. Public water systems must comply with the MCL: noncompliance with an RMCL cannot be the basia of an enforcement action tmdar section 1414 of the Safa Drinking Water Act. In addition, the SDWA specifies that primary drinking water regulations contain criteria and procedures to assure a supply of water that complies with the MCLs (i.a- monitoring and reporting requirements). Section 1401(1)(D). Section 1445(a) authorizes GPA to require by regulation any public water supplier to keep records, make reports, conduct monitoring and provide such other information aa may be required to assist in determining compliance with the SDWA. in evaluating health risks of unregulated contaminants, or in sdvising the public of such health risks. The SDWA also requires that tha revised primary drinking water regulations be reviewed every three years and amended whenever change* in technology, treatment techniques or other factors permit greater health protection. In addition to the regulatory mandates, the SDWAprovide# authorities for ensuring the safety of the nation's drinking water in a nan* regulatory contest Section 144Z(a)(Z](B) authorizes EPA to provide technical assistance to States and publicly owned water systems in response to and alleviation of any emergency situation which the Administrator determines to be a substantial danger to public health. In the absence of appropriate Stats or local action, section 1431 authorizes EPA to take such actions as the administrator deems necessary to protect public health from a contaminant that may present as imminent and substantial endangement As the third step, section 1412(b)(1)(B) to the health of persons. provided that EPA must propose and IL Regulatory Fnmeworic promulgate National Revised Primary Drinking Water Regulations (NPDVVR) The issuance of Revised Primary that would include RMCLs, MCLs and Drinking Water Regulations it the third monitoring and reporting requirements step in the evolution of the primary for those contaminants that may have drinking watar regulations mandated by an adverse effect on human health., the SDWA. In the first step, tha National Interim Regulatory Development Approach Primary Drinking Water Regulations (N1PDWR) were promulgated on December 24,1975, with an effective date of June 24,1977. Amendments were issued in 1978. and 1979 and 198a See 40 CFR141. Maximum contaminant levels (MCLs) and monitoring and reporting requirements wen set for numerous microbiological, inorganic, organic, and radionuclide contaminants (40 CFR, Part 141, Subpart B). At tha direction of the Congress, EPA based the NIPDWR in large pert on the 1982 U.S. Public Health Service (PHS) Standards for drinking water which in tun wen derived from previous standards dating as far back as 1915 for tha microbiological standards and tha 1940's for tha MCLs for some of tha inorganic chemcials. Aa the second step, section 1412(e) of ' the SDWA directed EPA to arrange for the National Academy of Science* (NAS) or an equivalent organization to conduct a study to assess the health effects of contaminants in drinking water and to provide proposals for RMCLs at levels at which then were "no known or anticipated effects on the health of parsons * * and a list of contaminants whose levels in drinking water cannot be determined but which may have an advene effect on the health of persona. The NAS submitted its initial report "Drinking Water and Health." to EPA in 1977 which was published in the Federal Register for Sublic comment; four additional reports av* been received. While Congress envisioned that the NAS would provide proposals for RMCLs in tha report the Development of the NPDWR will be accomplished in four phases; Phase 1 Volatile Synthetic Organic Chemicals, Phase 0 Synthetic Organic Chemicals. Inorganic Chemicals and Microbiological Contaminants, Phase ID Radionuclides, Phase IV Disinfectant By-Products including Trihalomethanes. In general the approach for all four phases will be similar. Initially an ANPRM will bo published followed by a comment period and a public meeting. Public technical workshops will also bo held. The workshops provide an opportunity for EPA to present the issues that must be addressed in development of the regulations and to receive information on scientific and technical matter* as well as receive comments on regulatory approaches. RMCLs will then be proposed followed by a public comment period and a public hearing(s). RMCLs will then be promulgated and proposals published for MCLs or treatment techniques, monitoring and reporting, and other requirements followed by a public comment period and a public hearingfs). Technologies will be identified that were used as the basis of determining the MCLs; in addition, generally available treatment technologies (CAT) will be identified for use in compliance with the MCLs and the issuance of variances. The MCLs or treatment techniques, NAS stated assantially that it would do monitoring and reporting, and other toxicological assessments of requirements including GAT will then contaminants in drinking water but that be promulgated. developing proposals for RMCLs was An ANPRM for Phase l(VOC) was not an NAS responsibility but an EPA issued on March 4.1982 (47 FR 9350, et regulatory function. In tha words of the seq.), and a public meeting was held in Academy, "determining safe levels to Washington. D.C- on April 28,1982. In protect the health of persons' drinking addition, four public technical water containing contaminants requires workshops were conducted across the consideration of other factors in country (June-August 1982) on volatils addition to the harmful properties of the synthetic organic chemicals (VOCs) in contaminants" (John 8. Coleman. drinking water. Executive Officer, NAS, Feb. 20,1975). The NAS reports have provided EPA with toxicological assessments of IIL Background and Summary of Comments contaminants in drinking water and - The ANPRM identified the VOCs based upon this information and data listed below as among those most from other scientific sources, EPA is commonly detected in drinking water developing the RMCLs. baaed upon date available at that time. 0 ? e , CK o. u SOKFIDgffTIAL'- ' Subject to Protective Order in 1 Boss v Conoco, Inc., Ho. 90-4833 14th Judicial District Court. ` v Calcasieu Parish, Louisiana 24332 Federal Register / Vol. 49, No. 114 / Tuesday, June 12, 1984 / Proposed Rules trichloroethylene tetrachloroethylene carbon tetrachloride 1.1.1- trichloroethane 1.2-dichloroethane vinyl chloride dichloromethane benzene chlorobenzene dichlorobenzene trichlorobenzene 1.1- dichloroethylene cis-1.2-dichloroethylene trans-1.2-dichloroethylene The purpose of the ANPRM was to solicit comments on the many scientific, technical, legal and economic questions associated with determining the proper approach under the Safe Drinking Water Act (SDWA) to limit human exposure to VOCs, The ANPRM was published to initiate discussions that would assist the Agency in determining the proper approach under the SDWA for minimizing human exposure to VOCs. The public was invited to comment on the following broad issues: What is the significance of contamination of drinking water by VOCs? Should national standards be set for VOCs? If standards are appropriate, how should levels be established? In addition to the above broad questions, comments were requested on specific technical and scientific- questions. Also, available reference materials on occurrence, health effects, analytical methods, and treatment costs of VOCs in drinking water were provided for technical and scientific review. Summary ofPublic Comments A total of 136 public written comments were received with the comment period ending on September 30,1982. The National Drinking Water Advisory Council (NDWAC) met in Washington. D.C. on September 23-24. 1982, to discuss the VOC ANPRM and its related issues. The NDWAC provided its recommendations to the Administrator in a letter dated January 5,1983. Public comments pertinent to this proposal are summarized in this section and in Appendix A. Comments pertinent to proposal of the MCLs and _ monitoring/reporting requirements will be summarized in that proposal. The public workshops conclusions and recommendations and the NDWAC recommendations are briefly summarized below. As representative'of comments received by drinking water industry associations and public interest .implemented for all water supplies, groups, comments submitted by the preferably using the purge and trap American Water Works Association procedure (EPA Method 502.1 or and Natural Resources Defense Council equivalent), but requirements for (NRDC). respectively, are also systems serving less than 10.000 people summarized. would be at the discretion of the State. Summary of Comments From Public Workshops Overall, it was concluded that contamination by VOCs is a national problem warranting action. There was sentiment in favor of establishing MCLs and some sort of monitoring program, provided the health effects data are valid and indicate the need to reduce human exposure. The health effects work groups believed that there are sufficient data to The initial monitoring frequency should be similar to the trihalomethane (THM) rule. In addition, guidance in the form of contamination levels, and action categories for five of the VOCs (i.e,, vinyl chloride, trichloroethylene, tetrachloroethylene, carbon tetrachloride, 1,2-dichloroethane) should be established for all water supplies. Natural Resources Defense Council (NRDC) cause concern. Three groups suggested that MCLs be set. However, every group qualified its recommendation by saying, variously, that the data are limited, more studies are needed, and that the difference between genotoxic and nongenotoxic carcinogens should be addressed by EPA. Aeration and granular activated carbon were identified as generally . available technologies, effective in reducing VOC levels to 10 pg/1 (micrograms per liter or parts per billion (ppb)) or lower. Cost projections presented by EPA were considered to be reasonable but they should be updated. The proposed analytical methods were found to be suitably accurate and the best available at this time. Concerning monitoring, the consensus seemed to be that EPA should provide minimum requirements within which States could develop their own monitoring plans, if data show that VOC contamination can be adequately predicted. EPA would provide criteria and guidance to assist States in predicting which systems wen vulnerable to contamination by VOCs and thus be monitored. American Water Works Association (AWWA) The NRDC recommended comprehensive national standards for volatile organic chemicals (VOCs) saying that the occurrence and health effects data show a significant national problem that warrants action under the SDWA. NRDC stated the EPA should establish RMCLs and MCLs for the 14 VOCs addressed in the ANPRM as well as an RMCL and MCL for total VOCs supported by mandatory national monitoring requirements. Other comments by NRDC included: Recommended Maximum Contaminant Levels (RMCLs) should be set at zero for carcinogens. RMCLs for non-carcinogens may be set at a noobserved-effect-level with an adequate margin of safety because RMCLs are health goals and are not intended to reflect feasibility of attainment. The multi-stage model as modified by the Carcinogen Assessment Group (CAG) should not be used in establishing RMCLs for carcinogens. Mathematical models at best provide crude estimates of the risks resulting from exposure to a carcinogen. Calculations of exposure levels corresponding to lifetime cancer risks of lCT* should provide the upper limit for MCLs. That is, contaminant levels The AWWA recommended that should be set at concentrations contaminants be controlled at their source through EPA's existing statutory authorities. They believed MCLs are not appropriate at this time, since "safe" corresponding to lifetime cancer risks of no greater than 1CT*. MCLs for non carcinogens should be set at correspondingly conservative levels. levels of VOCs cannot be determined from existing health-effects data. However, when the health effects data have been evaluated by a recognized independent scientific organization (l.e.. NDWAC Recommendations The National Drinking Water Advisory Council (NDWAC) provided the following recommendations and National Academy of Sciences (NAS)), the AWWA felt that MCLs should be established if a significant health risk exists. In the interim, AWWA recommended that national monitoring for specific compound identification should be analyses. 1. The occurrence data derived primarily from the random surveys conducted by EPA and selected data produced by the States in conjunction with the health risk data, warrant establishing controls for 5 of the VOCs CONFIDENTIAL Subject to Protective Order in Boss v. Conoco, Inc,, No. 90-4837 14th Judicial District Court CMA 007900 Federal Remitter / Vol. 49. No. 114 / Tuesday, June 12. 1984 / Proposed Rules 24333 d in drinking water. These are: loroethylene, tetrachloroethylene. ion tetrachloride, 1.2-dlchioroethane 1,1,1-trichloroethane. Regulations under the Safe Drinking ter Act should be established for ;e 5 chemicals at this time. '* iitional data would be needed before ecision could be made on other atile organics found in drinking ' ter. Health advisory type guidance mid be provided for these compounds lieu of establishing MCLs. ). Sufficient animal toxicology does ist at this time for establishing RMCLa - those 5 chemicals noted in 1. above, aantitative risk calculations using a learized multi-stage model should be ed for establishing RMCLs for the treinogens. A1 in 100,000 target risk is commended as ths RMCL For 1,1.1- ichloroethane, which the current data idicate is not carcinogenic, the RMCL iiould be calculated horn the No Jbserved Effect Level (NOEL) for eurotoxicity with appropriate safety actors. ** 4. The analytical methodology for ietecting and quantitating VOCa la well stablished (i.e,, EPA Method S02.1 jsing the Purge end Trap technique and similar procedures). No information was provided to the Council on the availability of laboratory services; however, it is assumed that services would be available to meet ultimata demand. The Council believes that monitoring is technically and economically feasible. 5. Sufficient data exist* at this time to determine that granular activated carbon and aeration are "generally available technologies" for central treatment application. Appropriately designed point of use devices, when shown to be effective for VOC control, can also be considered for some small water systems if they are cost/effective and properly managed. IV. Volatile Synthetic Organic Chemicals in Drinking Water Hundreds of chemicals have been detected at one time or another in drinking water in the U.S., but the vaat majority have been detected infrequently and at very low concentrations. Selection of candidate chemicals for revised national primary drinking water regulations is made from an analysis of data on the occurrence frequency, concentrations detected, sire of the exposed populations and the toxicology of the chemicals. This section briefly summarizes the available occurrence data, provides an overview of population exposure estimates, and discusses the health effects data for the VOCs. Additional information can be found in the references listed in section Vff. Occurrence of VOCt in Drinking Water One or more VOCs have been detected in numerous public water systems across die country. Typically, contamination is at low levels (i.e,, less than 1 part per billion, pg/1) but some systems have found higher levels. The VOCs are man-made chemicals, their presence may indicate that a pollution incident hat occurred, and some of them are among the moat frequently detected contaminants around hazardous waste sites. Several of these chemicals are suspected carcinogens, with differing degree* of evidence, while certain of these are mutagens and/or teratogens in some test systems. In 1882. EPA conducted a national sampling (Ground Water Supply Survey (GWSS)) of almost 1000 drinking water systems using ground water, 800 were selected at random and 800 were selected by the States as having high potential for VOC contamination (non random). Table 1 presents results of the random portion of tha GWSS. Approximately 21 percent of the systems in the random set had one or more of the VOCs at detectable levels (mostly in the sub p.g/1 range). The data showed a distinct difference in the frequency of occurrence of VOCa between larger and smaller systems; approximately 28 percent of samples in systems serving over 10,000 detected one or more VOCa in tha drinking water whereas 17 percent of samples in systems serving less than 10,000 detected VOCs. Six tenths percent of all public water systems serving less than 10.000 were sampled in the survey whereas 15 percent of systems greater than 10.000 were sampled. Six national surveys have been conducted by EPA since 1878. These include: National Organics Reconnaissance Survey (NORS) National Organics Monitoring Survey (NOMS) National Screening Program for Organic* in Drinking Water (NSP) Community Water Supply Survey (CWSS) Rural Water Survey (RWS) * Ground Water Supply Survey (GWSS) Based upon the above six surveys, projections of national occurrence and human exposure potential for the VOCs are summarized in Table 2 for levels associated with various risk rates. These surveys were conducted for various purposes over an eight year period which saw a rapidly developing state-of-the-art in water analytical methods. Different analytical procedure* ware used and. consequently, soma surveys were able to detect and measure particular VOCs at lower concentrations than other surveys were able to do. The most-significant portion of the data bass on VOCs. however, is derived from the Ground Water Supply Survey and tha Community Water Supply Survey. In combining the survey data, the national projections of the frequency of occurrence of VOCs at various concentrations can be provided only for those concentrations at or above the level at which all of the surveys were capsble of detecting and measuring them. This level, referred to as the lowest common quantifiable concentration, is generally the highest detection limit or minimum quantifiable concentration from among the surveys that are combined. Table 2 shows the estimated frequency of occurrence of the VOCs at or above the lowest common quantifiable concentration. Individual survey* using detection limits or minimum quantifiable concentration* less than the lowest common quantifiable concentration may report a higher frequency of occurrence of some VOCs. For example, according to Table 2,3.6% of the nation's ground water supplies are projected to have trichloroethlyene at or above the lowest common quantifiable concentration of 0.5 pg/l. whereas the GWSS (random sample), using a minimum quantifiable nominal concentration of 0-2 pg/L reported trichloroethylene to be present in 8.4% of the supplies sampled (Table 1). (Note: The GWSS random sample was found to have 4.1% at or above 0.5 Tasi* 1.--Summary of GWSS Occurrence Data (AaMom aan*iK "-*! ,1 oww ClUff MMwtf hfM |i^1 No. Pmm 04 to 74 04 4 30 0.4 1 27 9.0 4 1| 34 .9 4 10 14 1.1 Como wnamat------------------------------------------------ a IS sa * .4 s 23 70 10 34 2 to C0IT?II3BUTTAT. ' wieV Protectlve r<ler iff .Conoco, Ino.^ffiv 90-4837 14th Judicial District Court Calcasieu Parish, Louisiana 24334 Federal Register / Vol. 49, No, 114 / Tuesday, June 12. 1984 / Proposed Rules -- - -- Table 1--Summary of GWSS Occurrence Data--Continued (Rvidun umcM n-4e] Pram* QwnWinraiA^i NUI No. ftaraam *.MMao* 2 2 2fi 2 .8 2 j i ,8 '1 2 2 .8 2 2 2 2 2 2 2 18 | 17 | 1,7 8 1.9 8 14 f 1,1 2 82 89 9A 12 Q o Q o 0 0 e o o 0 o J 2 9 J 2 ,7 U .1 9 2 1.1 IS 6.9 1.5 9 2.9 21 19 fij 1.1 15 1 11 S.S Tabu 2.--Approximate Percent or Ground Water Systems and Size or Population Projected To Exceed the Nominal Indicated Risk Level Rrabtavra OfMung vatar Corwam twon RUM NMfM 6WW1Q mm UWk Mon tor *BCh Mira raraiabto A^l Tprejrtol PRNl*tort raipcMd ranra) - - -- -- - H.|< NI.Nil 41^ _ -- F 10** to-* 10-* 10** 10`* 10-* 10"* 10"* 10"* 1Q-* 10" 18** 10** 10- 10** 10"* 10** 10-* to-* 10- to-* 10"* 10r* 10"* 26 a 200 1 10 100 A 4 40 47 6.7 67 M a JO s IS t VO 100 217 sir 1170 TS 11J 16 100 027 17 V 0J4 14 24 >0.8 0 0 *0.018 0.18 14 0.5 20 100 4 10 o J 5 40 4 5 70 J so 4 i so 1 10 TOO 9.4 .4 .1 94 J 0 1.6 J 0 1.5 A 0 16 . 0 4 0 0 46 J J01 6420 510 40 4470 440 0 MOO 160 0 1.000 210 0 1410 70 0 1430 0 0 160 270 160 1 RftMrt draft raMNom by E*A' Caramegan ttmnmmm Qr *MbMnary dita; wan cracmoyraac ftnpdpirabi>> Daty mratra (AADQ a 1.6OO yg/tprapaaad RMCL Is 800 pft/L Not coMdval s Mb mosoara m a oaramogan- OfMapra raa tray raranga Ob cbaMMbL * Random Mnratd mm aasunanera* Ml taaiera tB lavra. Estimating the occurrence of VOCs as a class in public water supplies is difficult because not all of the six surveys looked for all the listed VOCs and because the detection limits or minimum quantifiable concentrations for specific VOCs varied from one survey to another. However, some insight to the overall occurrence of VOCs can be gained from analyses of the data from the GWSS and CWSS. As shown in Table 3, in the GWSS, 99 of 466 (21.2%) randomly selected ground water supplies had at least one of the 29 VOCs identified in that survey. In the CWSS, SO of the 330 (15.2%) ground water supplies had at least one of 10 VOCs identified in that survey; 14 of 106 (13.2%) surface water supplies were found to have one or more of the VOCs present. Occurrence of VOCs at levels above S pg/1 appears to be more likely in ground water rather than surface water; however the detection frequencies may be similar. Virtually all persistent occurrences of VOCs above 50 p.g/1 are expected to be in ground water. However, the frequency of specific VOCs occurring above that higher level is expected to bemuch less than 1%. Table 3 also provides data on multiple occurrences of VOCs; 44 of 466 (9.4%) randomly selected sites in the GWSS had measurable levels of two or more VOCs. while 19 of 330 (5.8%) of the ground water supplies in the CWSS had two or more present. CONFIDENTIAL Subject to Protective Order in Ross v. Conoco, Inc,, flo, 90-4837 14th Judicial District Court u\,\. Calcasieu Parish,, Louisiaam CMA 007902 Tabu 3.--Summary of Sing** and Muttpf# OccvTtnca of VOCi at Cits* de to. WtofllMM GWSS 1 Random1 cwss1 OrawnS Mr* Mmamc** A ....... ........................... .... ............................................ ................... . %'t ............................. . ............................................................. ............................ ............................................................ . hA ................. ............................................................................. .. ............................................ vT ... .... ...... ^1 ................... ... ... > 3a*ad <* andyw* tor 29 VOC3. 499 tufipbM vkidiaA 6**j on mwrtm tor 10 VOCa. 313000-w-p--pJt-n-*---' J-- lsr<7i.o%> Ncmi *4 (9.4%) 2t <9.4%) 14 <3,0%) 1 <1 7%) 4 <0.%> 2 (0.4%) 0 e o o M e {| IIIII? 02(98.9%) 14<13%) 9 <.?%> 1 (0.9%) 0 0 0 0 0 Table 4 show* the frequency of occurrence of supplies with total concentrations of the 28 VOCs examined in the GW5S (random sample) above the indicated levels. In addition to the EPA national survey data, numerous incidents of contamination have been reported by States across the country, and contamination in some public water wells has been in the range of 100 pg/l to 1,000 fig/l and higher. Usually when concentrations in that range have been detected, corrective measures have been rapidly taken this could explain the relatively small number detected in the random surveys. Several States, including California. Michigan, New York, and Connecticut, have monitored comprehensively for VOCs while others have generally responded to incidents of contamination. Table S summarizes State data that Were available to EPA. The estimates of population exposed to VOCs in Table 2 are.based only on the data from the EPA surveys; the State daja and miscellaneous information were not included because those data were only from a few States and therefore not geographically representative. Furthermore, since much of the State data were obtained in response to incidents of recognized contamination problems, these data may not be representative of typical conditions existing nationally. However, while these data were not used for computing the national projections, they (including the GWSS non-random data) do provide a valuable and necessary perspective for evaluating those projections. Tasue 4.--Cumuladvc Occurrence of Supplies m the GWSS Random Sample with Total Concentration of 29 VOCs Above the indicated Levels tcm wtttommptoq*dtutotod >ume SH0I >to 14*1 '90 net M00|L*1 49*.------------------- *t <214 pwnS 20 <4.3 PWBWS <2 (ZS pawns 3 <0.4 parawe 0l Table 5.--Summary of State Occurrence Data 1 s I II *** toanbard toffldat ,1 1 .TwNh^nSmm . . . ........................ Misf* <*iNH*N ... ......................................................... 3,639 19' 4.239 19 19 tid 19 2449 f 43 if 2,331 1,799 329 1,000 394 510400 7l 2.290 197 90 333 1JQ0 4 17 177 2,100 12S MO Or Sum sawn pbwWwk sn Th mm NnuN a aiMcmn at MAW mm *w wrtaw aw SMw, m npnws, n--iiiM pccawnoA m M ww W oownnnwcn Pndww. ms mmma, wain dmNa guns mar m not pai wtwvapwmmd idmbmaWa**MWwMwpaimwwwAeMnM * m cm caw Occurrence and Exposure Assessment As part of the basis for determining how to reduce human exposure to VOCs and determine the appropriate regulatory actions, the occurrence data on VOCs are used in two principal areas. As input to the health risk assessment of the VOCa, an estimate is conducted of the number of individuals in the United States exposed to various levels of the VOCs in drinking water from public water supplies. Information on Dietary intake and respiratory intake from ambient air is provided and is used to estimate the relative contributions of the three sources, particularly of drinking water, to the total dose received by individuals. While it is recognized that some individuals may be exposed to the VOCs from other sources, such as occupational settings or the use of particular consumer products, these analyses are limited to drinking water, food and air because these are the major exposure routes common to all individuals. In addition to serving as an input to the health assessment the exposure assessment supports EPA efforts to estimate the economic impact of the regulatory alternatives being considered. To aid in that effort projections are provided to estimate the number of public water supplies of various water source and system size categories likely to have VOCs present and the distribution of the VOCs levels in those water supplies. There are approximately 60.000 public water supplies in the United States. These systems fall into two major categories according to water source (i.e,, surface water and ground water) and for purposes of estimating the potential regulatory impact are divided into eleven size categories according,to the number of individuals served. Probability distributions for computing the expected number of systems with concentrations in specified intervals were examined and tested by statistical significance procedures. Ideally, separate probability distributions should be developed for each water soure and system size category; however, the available data were too limited for this. Therefore, it was necessary to consolidate some of the size categories to have sufficient data for developing the probability distribution. Specifically, for ground water it was necessary to collapse the data into two size categories: less than 10,000 people served and 10,000 or more people served. For surface water, there were insufficient data for statistical analysis even when all size categories were combined. The delta distribution was found to be reasonable for the available data and was used for determining the probability of contamination at various levels within the two ground water size categories. For completing the national estimates for ground water, it was assumed that the probability distribution function established for a given consolidated size CONFIDENTIAL Subject to Protective Order in Boas v. Conoco, Inc., No. 90-4837 ,14th Judicial District Court Calcasieu Parish,: Louisiana CMfit 007903 24336 Federal Register / Vol. 49, No. 114 / Tuesday, June 12, 1984 / Proposed Rules category was directly applicable to each of the systems in a particular source/ size category. Concentrations of VOCs within a given interval were calculated as the product of the probability associated with the interval and the total number of systems in that source/ size category. As noted previously. Table 2 summarizes the estimated population exposures at various levels of contamination. Details of the data base used in these projections for each of the VOCs can be found in the occurrence documents referenced in section VO. Human Health Considerations The underlying principles nsed to assess the potential health risks of exposure to chemicals are discussed in this section. Brief summaries of the toxicology of each selected VOC are also provided. A more detailed evaluation of the health effects of the chemicals is given in the individual health criteria documents referenced in section VH. Development of KMCLs for Noncarcinogens When appropriate data are available from human epidemiology or animal studies, determination of the "no known or anticipated advene effect levels" for RMCL purposes for toxic agents not considered to have carcinogenic potential is a relatively well-accepted procedure. "No effect" levels for chronic or lifetime periods of exposure including a margin of safety are referred to commonly as ADls or Acceptable Daily Intakes. These ADl's are considered to be exposure levels which would be without significant risk to humans when received daily over a lifetime. For noncarcinogenic end-points of toxicity, it is assumed that an organism can tolerate and detoxify some amount of a toxic agent without ill effect up to a certain does or threshold. As the threshold is exceeded, the extent of the response will be a function of the dose applied and the length of time exposed. Hie intent of a toxicological analysis performed as part of the regulatory development procesris to identify the highest no-observed-adverse-effect-level (NOAEL) based upon assessment of human or animal data (usually from animal experiments). To determine the ADI or "no effect" level, the NOAEL is divided by appropriate "uncertainty" or "safety" factors. This process makes accommodations for the extrapolation of animal data to the human, for the existence of weak or insufficient data and far individual differences in human sensitivity toxic agents, among other factors. General guidelines were provided by the NAS Safe Drinking Water Committee {Drinking Water and. Health, Vol. I) which state that an uncertainty factor of 10 is used if there exist valid experimental results via ingestion in humans; an uncertainty factor of 100 is used if there exist valid experimental results on long-term feeding studies on experimental animals; and an uncertainty factor of 1000 is used if there exist inadequate animal data. Additional factors and variations also may be used if the circumstances dictate it. Figure 1 illustrates a process by which an ADI for humans is computed. Figure 1 shows the lower,end of a typical sigmoid-shaped dose-response curve as might be generated experimentally for a non-careinogenic end-point of toxicity believed to have a threshold. The solid line represents the curve as experimentally-determined. Point A represents the highest NOAEL determined during the experiment Point C represents the theoretical threshold dose at or above which an adverse effect might occur in the most sensitive case. CONFIDENTIAL Subject to Protective Order in ss_v. Conoco, Inc., Ho. 90-4837 14th Judicial District Court Calcasieu Parish,, Louisiana CHA 00790A Federal Register / Vol. 49, No, 114 / Tuesday, June 12. 1984 / Proposed Rule# Figure 1 Non-Carcinogenic Effect 24337 Doae A: NOAEL (experimentally derived) B: ADI or "no effect" level Ci Presumed threshold for adverse effect Ci* Another possible presumed threshold for adverse eff ct C2! Non-threshold end point of toxicity To derive the human "no effect" level or ADI baaed upon the experimentallyderived data displayed in Figure 1. the appropriate margin of safety (i.e.. uncertainty factor) is applied to establish an acceptable level of exposure, depicted as Point B. The objective of applying the uncertainty factor is to make Point B fall below Point C. Thus. Point B would represent the ADI or "no effect" level with a margin of safety. It is possible that the actual dose response curve would result in Point Ci not detected in the experiment in which case the calculated ADI (i.e., Point B) might not be below the actual threshold for an adverse effect There is suggestive scientific evidence available to postulate that thresholds do exist for non-carcinogenie end-points of toxicity. In the absence of irrefutable evidence, however, it remains theoretically possible that one or more non-carcinogenic end-points may not have a demonstrable threshold. The dose-response curve for this case is depicted as the dashed line from Point A to the origin or Ct. Ct represents the threshold dose and the "no effect" level in this case would thus be zero. Table 8 summarizes the suggested Adjusted Acceptable Daily Intakes (AADIs) for tlfe VOCs based upon chronic toxicity data without consideration of the potential carcinogenic risk. These values were not used for developing proposed RMCLs 007905 CM* 24338_________ Federal Register / Vol. 49. No. 114 / Tuesday, June 12, 1984 / Proposed Rules for chemicals considered to be potential carcinogens, but are provided to add some perspective on the chemical's total toxicity including potential noncarcinogenic end-points. In addition, these values may have some practical application as guidance on the levels at which no adverse health effects would be expected to occur based upon non-carcinogenic data. This would be especially useful for substances considered to be "weak" carcinogens. Comment is requested on these values. The AADI's were calculated by: Determining the highest NoObserved-Adverse Effect Level (NOAEL). or the lowest observed adverse effect level (LOAEL) in mg/kg body weight/day, Dividing by appropriate safety or uncertainty factorjs) (U.F.), Multiplying by the weight of an adult (70 kg), and Dividing by the amount of water consumed by an adult per day (2 liters/ day). (This allocates the ADI totally to drinking water which would have to be modified to consider other routes of exposure when the RMCL or MCL is computed.) The formula for this calculation is as follows: (NOAEL in mg/kg/day) [70 kg) ------------------ -------------------- -AADI mg/1 UJ'.(s) x 2 liter*/day Table 6.--Suggested Adjusted Acceptable Daily Intake: VOC'8 [Don no* come* ctononmdty M ncAidn lUHiAn from m and food] Commod AADI 0066 *g/L 0J6 mg/l. 0.026 mg/L u-oamwww------------------ 0J6 mg/L 0.06 mg/l. QtSmg/l --.J 026mg/l ITS mg/L The calculated AADIs above assume that the total exposure was from drinking water. Since normally exposure also comes from air and food, in addition to drinking water, and since drinking water is frequently a minor contributor to the total exposure, the RMCL or MCL should be modified to take into account the relative source contributions. The World Health - Organization, in "Guidelines for Drinking Water Quality" (1963), assigned as little as 1 percent of the ADI to drinking water where the chemical was known to bioaccumulate to a high degree, while greater proportions were assigned where the chemical was known to bioaccumulate to a lesser . serve as guidance to EPA when assessing the irreversible effects of long degree. In "Drinking Water and Health" term exposure to non-threshold (1977), the National Academy of substances at low doses: Sciences provided projections of 1 Principle 1: Effects in animals, percent and 20 percent as illustrations of properly qualified, are applicable to drinking water contributions. In the National Interim Primary Drinking Water Regulations for six organic chemicals, drinking water was assumed to contribute 20 percent of the total daily intake. Because of the wide range of environmental exposure distributions that would occur across urban end rural populations as well as because of age and occupationally-related differences, assumption of a 20 percent contribution from drinking water would be reasonably conservative and protective. Thus, in this case, if an AADI value for a non-carcinogen were to be the basis for an RMCL it would be reduced by 80 percent to account for up to 20 percent contribution from drinking water to the total daily burden. Development ofRMCLs for Carcinogens. Evaluations of the texicology of substances which may possess carcinogenic potential is a twophase process. In the first phase, the toxicological data base for noncarcinogenic end-points of toxicity was evaluated in the same mariner as described above for "non-cardnogens" (Table 6). In the second phase, assessment was made of the evidence of the carcinogenic potential (e.g., long term bioassays in rodents and human epidemiology) as well as information which provides indirect evidence [e.g., mutagenicity and other short-term test results). This process is complex since the production of cancer probably is a multi-stage event determined by a . multiplicity of mechanisms, the nature of which remain, for the most pert hypothesized rather than identified. To date, scientists have been unable to demonstrate experimentally a threshold of effect for "carcinogens." acording to the 1977 report of the NAS Safe Drinking Water Committee. This leads to the assumption that sine* no threshold dose can be demonstrated for carcinogens, any exposure might represent eome finite level of risk. Depending upon the potency of the specific carcinogen and the level such e risk could be vanishingly small at very low doses. Human epidemiology data are extremely limited in their ability to identify carcinogenic risks. Thus, animal experiments are conducted from which potential human risk is extrapolated. In the first volume of Drinking Water and Health (1977), the NAS Safe Drinking Water Committee provided principles to man. Principle -Methods do not now exist to establish a threshold for long term effects of toxic agents. Principle 3: The exposure of experimental animals to toxic agents in high doses is a necessary and valid method of discovering possible carcinogenic hazards in man. Principle 4: Material should be assessed in terms of human risk, rather than "safe" or "unsafe". Tumors appear spontaneously in experimental animals, at different rates and different sites depending upon the specie* and strain. It is unlikely that any increased tumor incidence could be detected following exposure of experimental animals to most carcinogen* at dose levels occurring in the ambient environment. Very large numbers of animals would be required to distinguish between treated and control groups. It is possible, as wasahown in the 24,000 animal "mega mouse" study on 2-acetylaminofiuorene at the National Center for Toxicology Research (NCTR), that a definitive answer would not necessarily be forthcoming at the low dose levels. Mathematical extrapolation still would be required to project human risk. Relying on this type of study for individual assessments is impractical because of its great expense and lingering scientific uncertainty. In order to produce quantitative estimates, the assumption has been made that estimated excess cancer risk in humans at low dose levels can be extrapolated using various techniques from results observed in animals at high dose levels. Conventionally, designed carcinogenicity bioatsay studies are conducted using both aexes of two pedes of test animals (usually rat and mouse) with each group of 50 animals exposed at the maximum tolerated dose or one-half the maximum tolerated dose. In addition to the possible existence of threshold*, other sources of uncertainty in high to low doae extrapolation include: (1) heterogeneity of sensitivity in the exposed populations, (2] the pharmacokinetic behavior of the toxic agent in animal* vs. the human and (3) mechanisms of action (i.e.. whether the gent initiates the process or acts at later stage). Classification of carcinogens into genotoxic vs. nongenotoxic carcinogens based on possible mechanisms has also been considered CONFIDENTIAL Subject to Protective Order in Eoss_v. Conoco, Inc., No. 90-4837 14th Judicial District Court Calcasieu Parish,, Louisiana O07906 4. % o Federal Register / Vol. 49. No. 114 / Tuesday. June 12, 1984 / Proposed Rules 24339 but a scientific consensus has not been achieved. Fundamental changes in normal cells are the most probable basis for the conversion of normal cells to cancer cells; however, the nature of these changes and how they are brought about is still a scientific uncertainty. Many scientists believe that the most likely mechanism involves direct alteration of DNA by carcinogens. Many carcinogens are capable of altering DNA; chemically-induced alteration of DNA in germinal cells can also cause heritable changes, or mutations; thus, when a chemical shows a positive response in short-term mutagenicity tests, there is concern that it could also be a carcinogen. Scientists also generally believe that cancer results from a multi-stage process. However, these processes are not well understood and available evidence is insufficient to differentiate between carcinogens on the basis of mechanism (IARC, 1983). Therefore in this proposal EPA did not make a differentiation based upon potential mechanisms. Thus, quantitative risk extrapolation procedures can provide only a rough projection of carcinogenic hazard because of the many unknown factors which enter into these estimates. Models using different assumptions may produce estimatea ranging over several orders of magnitude. Since there is currently no way to demonatrate the accuracy of any model at low doses, this process is a subject of debate In the scientific community. However, in spite of these difficulties, quantitative risk estimation does provide the decision maker one means of setting priorities among pollutants and some gauge of the potential seriousness of environmental hazards (see NCI Subcommittee report referenced in section VII). EPA's Carcinogen Assessment Group employs a multi-stage model among various others to extrapolate potential excess cancer risk expected at doses of the chemical found in the environment from results in high dose animal studies. (U.S. EPA. 1980). Equivalent human doses are established either on a body weight basis (mg/kg) such that the ratio of human to animal body weights is raised to the V4 power; - human body r weight t animal body weight or on a body surface area comparison. The multi-stage model is used for several reasons: (1) it is more systematic than the one-hit model, (2)it invokes fewer arbitrary assumptions, (3) the Shown along with the risk estimates assumption of low dose linearity is not in Table 7 is a qualification of the degree essential in the use of the model and (4) of evidence of carcinogenicity exhibited it incorporates data from all of the by the chemicals. The International dosage groups which are consistent with Agency for Research on Cancer (IARC) the multi-stage model. At the same time, provides guidance for categorizing it is conceptually consistent with the chemicals having sufficient or limited linear, non-threshold concept With this evidence df carcinogenicity. In the IARC model CAG estimated the upper bound Monographs Supplement *1 the excess cancer risk rate at a specific definition for sufficient evidence for exposure level for a 70 kg adult who carcinogenicity indicates that there need consumes 2 liters of drinking water per be an increased incidence of malignant day. every day over a 70 year lifespan. tumors; (a) In mulitple species or strains, These calculated risk rates have or (b) in multiple experiments, or (c) to associated uncertainties. This an unusual degree with regard to uncertainty has many sources, including incidence, site or type of tumor, or age at such uncertainties as the shape of the onset Sufficient evidence of human dose-response relationship at low doses, carcinogenicity indicates a causal differences in responses between association between exposure and humans and laboratory animals, and the . human cancer. Limited evidence of effects of artificial dosing regimens. A carcinogenicity means that the data relatively minor source of uncertainty is suggest a carcinogenic effect but are statistical fluctuation that results from limited because: (a) The studies involve the finite sample size necessarily used a single species, strain, or experiment; of in any experiemental study. This is the (b) the experiments have an inadequate only uncertainty that can be readily period of follow-up, poor survival too quantified: it is expressed in EPA's few animals, or inadequate reporting; or methodology by giving the upper-95% (c) the neoplasms produced often occur confidence limit of the observed spontaneously or are difficult to classify response. Other confidence limits could as malignant by histological criteria also be calculated. (In more technically alone. Limited evidence of human precise terms, the confidence limit is carcinogenicity indicate a possible calculated on the coefficient of the linear term in the multi-stage model assuming that all the statistical carcinogenic effect in humans, although the data are not sufficient to demonstrate a causal association. In uncertainty is loaded on that term.) general although a single study may be Excess cancer risk rates also can be projected using variations within a specific model or other models, such as the one-hit modelJhe Weibull model and logit and probit models. There exists no solid basis in the current understanding of the biological mechanisms involved in cancer to say that one model provides a better estimate of the true risk. The estimates of risk at low doses for these models can differ by several orders of magnitude. However, the linear non-threshold model usually has the best even if limited, scientific biological basis of any of the currently available models for indicative of a cause-effect relationship, confidence in inferring a causal association is increased when several independent studies are concordant in showing the association, when the association is strong, when therer is a dose-response relationship, or when a reduction in exposure is followed by a reduction^ the incidence of cancer. The National Academy of Sciences in their report Drinking Water and Health, Vol. I (1977) classified chemical carcinogens into four categories: human carcinogens, suspected human carcinogens, animal carcinogens and suspected animal carcinogens. giving an upper limit estimate. The Figure 2 presents a typital dose- multi-stage model is presumed to response curve for animal experiments usually give a conservative risk estimate dealing with carcinogens. Usually only (i.e,, less likely to underestimate the two data points are available either actual risk) and thus would usually be from an NTP bioassay or other chronic consistent with a protective regulatory study. Points A, amd A* represent the hilosophy. A similar model was used tumor incidence observed in the animal y the NAS Safe Drinking Water experiment at the high and low dose Committee in the calculations provided levels, respectively. Point B represents to EPA in "Drinking Water and Health". the mathematically extrapolated tumor The NDWAC recommended that the incidence eaimated to occur at an multi-stage model be used in the exposure level below those estimation of cancer risk associated experimentally applied. This exposure with the VOCs. Various calculations level would correspond to a level likely using multi-stage models art presented to exist in the ambient environment in Table 7. (usually far below the experimental Sub-ject to Protective-Order irt Eoss_ v. Conoco, Inc.. tf0. 90-4837 14th Judicial District Court Calcasieu Parish,, Louisiana .. CMA 007907 24340 Federal Register / Vol. 49, No. 114 / Tuesday, June 12, 1984 / Proposed Rules dose). Identification of this point (B) and others along the extrapolated lower end of the curve then allows for the projection of an associated excess human earner risk, Table 7.--Cancer Risk Estimates for VOCs; Projected Upper Limit Lifetime Cancer Risks Figure 2 Carcinogenic Effects Response i iigh dose tuiaor incidence (observed) Aj Low dose tumor incidence (observed) t Tueor Incidence (extrapolated) at dose corresponding to level found in ambient environment Dose Toxicology of VOCt Hie following are short discussions of the toxicity of VOCs for which RMCLs are proposed. Detailed assessments are found in the draft health criteria documents that have been prepared for each VOC and are provided for public comment see section VIL References. Trichloroethylene. Trichloroethylene has been shown to exhibit noncarcinogenlc bioeffects at high (nonenvironmental) doses in humans and several other animal species, including dogs, rabbits, guinea pigs, rats and mice. The major effects demonstrated are liver and kidney damage, central nervous system effects and depression in myocardial contractility. In the calculation of a suggested adjusted ADI for trichloroethylene, liver toxicity was used as the most sensitive end-point with respect to adverse health effects, not including the potential carcinogenic risk that may result from exposure to the chemical. A study in which rats were exposed to trichloroethylene through inhalation with resulting elevation of liver weights was used to calculate a suggested Adjusted ADI of 0.257 mg/1. This value was calculated based upon a minimaleffect-level of 300 mg/m1 (55 ppm), since rats exposed to this dose level (5 days a week for 14 weeks) showed elevation of liver weights. An uncertainty factor of 1000 was applied due to the fact that an animal study, where the no-observedadverse-effect-level was not identified, was used and because the study was only of 14 weeks duration. One hundred percent exposure from drinking water and a 70 kg adult consuming 2 liters of water per day were assumed in the calculations. The NAS has not calculated a chronic non-carcinogenic Suggested NoAdverse-Response Level (equivalent to an Adjusted ADI) for trichloroethylene, because every long-term study, with the exception of the National Cancer Institute (NCI) carcinogenesis investigation, involves trichloroethylene administration by inhalation. The NCI bioassay did not determine a "no-effect level" and thus it was not considered appropriate for use in the deviation of a chronic, noncarcinogenic value. Bacterial mutagenicity studies have shown trichloroethylene to be mutagenic in several systems, including metabolically activated Salmonella typhimurium and coll K12 strain; however, a later study reported trichloroethylene to be non-mutagenic in the Ames test system. Commercial grade trichloroethylene was tested by the National Cancer Institute (NCI) (1976) and was reported to induce hepatocellular carcinomas in male and female mice by oral gavage. A repeat bioassay by the National Toxicology Program (1983) using purified trichloroethylene in com oil found it to cause hepatocellular carcinomas in both sexes of mice, at a dose of 1.000 mg/kg per day, five days per week for 2 years, administered by gavage. Trichloroethylene was not carcinogenic in female rats under the test conditions and the results in male rats were determined to be insufficient to make an CONFIDENTIAL * Subject to Protective Order in Ross v. Conoco, Inc., No. 90-4837 14th Judicial District Court V&r4mh- Louisiana. CMA 007908 Federal Register / Vol 49, No. 114 / Tuesday, lune 12. 1984 / Proposed Rules 24341 adequate evaluation of the carcinogenicity. The dose* administered to the rat were 1.000 and 500 mg/kg/day. The International Agency for Research on Cancer (IARC) has concluded that trichloroethylene has limited evidence of carcinogenicity, based upon experimental animal studies and inadequate evidence from available human data. This means that the data suggest a carcinogenic effect in one species, but lack of confirmation in others. The World Health Organization (1961) has recommended a tentative guideline value of 30 ug/l for trichloroethylene in drinking water. EPA's Carcinogen Assessment Group has usedjthe linearized non-threshold multi-stage model to calculate projected excess cancer risk estimates extrapolated from high dose animal studies. For trichloroethylene, these estimates were based upon the NO bioassay data. Calculated risks correspondingly to various doses are listed in Table 7. Tetrachloroethylene, The principal non-carcinogenic effects of tetrachloroethylene in humans and other animals from both acute and longer-term exposures at relatively high (nonenvironmentai) doses include central nervous system depression and fatty infiltration of the liver and kidney with concomitant changes in serum enzyme activity levels indicative of tissue damage. A suggested adjusted ADI for tetrachloroethylene, considering adverse health effects other than the potential carcinogenic risk, was calculated based upon a series of studies in which rate were exposed by inhalation to tetrachloroethylene with effects observed on the central nervous system, immune system and certain blood components. The value of 0.085 mg/1 was derived from these studies, based upon a no-observed-adverse-effect level of 10 mg/m* [1.5 ppm) and an uncertainty factor of 100. This uncertainty factor was considered appropriate for use with a no-observedadverse effect level from an animal study with no comparable human data. Daily exposure of a 70 kg adult drinking 2 liters of water per day was assumed in the calculations. Tetrachloroethylene in com oil was tested for carcinogenic potential in mice and rats by gavage in the NCI Bioassay Program (1977). In these bioassays, it was shown that tetrachloroethylene increased the incidence of hepatocellular carcinomas in both sexes of mice, but not in rats. A dose rate of 531 mg/kg per day, 5 days/week in male mice and 366 mg/kg in female mice resulted in e tumor incidence rata of 6$ percent and 40 percent respectively. Because of an excessive dose related mortality in the gavage experiment and low dose level in the inhalation study, no conclusion can be made about the carcinogenicity of tetrachloroethylene in rats. Data from the recent gavage study has been withdrawn for the time being pending the results of an indepth audit by the NTP due to unresolved problems with the study as conducted. The majority of mutagenicity studies on tetrachloroethylene were negative. Two positive studies have been reported; however, the purity of the tetrachloroethylene was questioned in these cases. The IARC has concluded that tetrachloroethylene has limited evidence of carcinogenicity in animals and inadequate evidence from available human data: This means that the data suggest a carcinogenic effect in one species, but lack confirmation in others. The World Health Organization has recommended a tentative guideline value of 10 pg/1 for tetrachloroethylene in drinking water. EPA's Carcinogen Assessment Group has used the linearized multi-state model to calculate projected excess cancer riak estimates extrapolated from high-does animal studies. For tetrachloroethylene, these estimates were based upon the 1977 NCI bioassay in mice. Calculated risks corresponding to various doses are listed in Table 7. 1,1,1-TrichIoroethane. The principal toxic effects of 1,1,1-trichloroethane from which (non-environmental) djse exposure in animals and humans are depression of the central nervous system, increase in liver weight and cardiovascular changes. Liver toxicity was used as the most sensitive end-point with respect to adverse health effects, not including the potential carcinogenic risk, in the calculation of an adjusted ADI for 1,1.1trichloroethane. An inhalation study which examined exposure of mice to l.l.l-trichloraethane was used to calculate a suggested Adjusted ADI of 1.0 mg/L This study demonstrated changes in the livers of the mice at various dose levels. Two animal bioassays by the National Cancer Institute (NCI) have been completed in rats and mice [1977; 1983). In the earlier bioassay, rats and mica were treated with 1,1,1-trichloroethane in com oil by gavage. Because only 3 percent of the animals survived to the end of the experiment, due in part to chronic murine pneumonia which was determined to be the most probable cause of the high incidence of natural deaths among the animals, it was concluded that carcinogenicity could not be determined from this study. A repeat carcinogenesis bioassay of l.i.i- trichloroethane was conducted In which doses of 3.000 or 1.500 mg/kg were administered by gavage to both sexes of mice, and rates were given does of 750 or 375 mg/kg. In the preliminary report of this study, 1,1,1-trichloroethane was carcinogenic in both male and female mice showing an increased incidence of hepatocellular carcinomas but not in rats; however, these initial results have been questioned. l.l.l-Trichloroethane has been tested for mutagenicity in several test systems. Both negative and positive results were reported in mutagenicity tests in various Salmonella typhimurium strains, and 1,1,1-trichloroethane was not shown to be mutagenic in studies using yeast as an indicator organism. EPA's Carcinogen Assessment group has used the linearized non-threshold multi-stage model to calculate preliminary excess cancer risk estimates extrapolated from the preliminary reported incidence of hepatocellular carcinomas in female mice in the study cited above. Calculated risks corresponding to various doses are listed in Table 7. Similar calculations were made by the NAS {Drinking Water and Health, Vol. V) except that the average of the results in both male and female mice were used as the basis. The latest bioassay data, on 1,1,1trichloroethane is currently undergoing audit by the NTP and a final report has not been issued. Therefore this proposal will use the noncancer inhalation data as the basis for the proposed RMCL This notice will be amended if the final NTP report determines that 1.1.1trichloroethans was carcinogenic under the conditions of the tests. Carbon Tetrachloride. Carbon tetrachloride (CCIJ has been shown to exhibit non-carcinogenic effects in humans and animals following acute and chronic exposures. The principal effects seen at nigh doses are liver changes such as fatty liver with centrilobular necrosis developing if exposure is continued. A chronic AADl forCCl, of 0.025 mgI 1 was calculated from a recent report of a study (Bruckner, et al,, 1983) which has not yet been published or peer reviewed at this juncture. Rats weighing 200-500 g were randomly divided into groups of 15 to 18 animals each. The animals were given by gavage 0.1.10.33 mg CCl,/kg bw (in com oil). The animals were dosed on a daily basis. 5 times weakly', for a total period of 12 weeks. Blood samples were obtained from alternate animals at the Conoco^Inc^ Ho CMA 007909 24342 Federal Register / Vol, 49, No. 114 / Tuesday, June 12, 1984 / Proposed Rules following intervals: 2, 4, 6, 8.10 and 12 weeks pOBt-treatment. The serum was analyzed for BUN. GPT, SDH and OCT. At 1 mg/kg. there were no significant biochemical/histopathological changes. SDH, the most sensitive index of hepatotoxicity, was elevated (p <0.05) in rats receiving 10 mg/kg for 12 weeks. Also, these rats exhibited mild centrilobular vacuolization. At 33 mg/ kg. levels of GPT, SDH. and OCT were increased (p <0.01) and marked hepatic lesions were apparent. There was no evidence that CC1, was nephrotoxic. Comments on the experimental protocols and interpretations of the data are requested. Carbon tetrachloride has been shown to be carcinogenic in rats, mice and hamsters through oral administration. In the NCI (1976) bioassay for trichloroethylene, carbon tetrachloride was used as the positive control. Carbon tetrachloride was administered in com oil by gavage to rats at two dose levels: 47 and 94 mg/kg for males and 60 and 159 mg/kg for females. In mice, the chemical was administered at 1.250 and 2,550 mg/kg. Carbon tetrachloride was determined to increase carcinomas of the liver in both rats and mice in this bioassay. Carbon tetrachloride has not been shown to be mutagenic in any of the reported Salonjonella (Ames) assays. However, mutagenic activity associated with carbon tetrachloride has been observed in a test system using the yeast Saccharomyces cerevisiae. The 1ARC has concluded that sufficient evidence of carcingenicity in animals exists for carbon tetrachloride. The NCI has also identified carbon tetrachloride as an animal carcinogen and has used it as a positive control in several bioassays. The World Health Organization (1981) has recommended a tentative guideline value of 3 pg/1 for carbon tetrachloride in drinking water. EPA's Carcinogen Assessment Group has used the linearized non-threshold multi-stage model to calculate projected excess cancer risk estimates extrapolated from high dose animal studies. For carbon tetrachloride, the latest draft estimates were based upon the geometric mean of the four cancer studies. Calculated risks corresponding to various doses are listed in Table 7. 1.2-Dichloroethane. Theloxic effects of 1.2-dichioroethane in humans and other animals from both acute and longer-term exposures at relatively high levels include central nervous system depression, liver and kidney damage, gastrointestinal distress, adrenal and pulmonary effects and circulatory disturbances. A series of inhalation studies in which a variety of animal species were exposed for up to 8 months to 1,2dichloroethane were used to calculate a suggested Adjusted ADI for 1,2- dichloroethane. The most sensitive end points. not including the potential carcinogenic risk, identified in these studies were pulmonary congestion, diffused myocarditis, and fatty degeneration of the liver, kidney, adrenal and heart. A value of 0.260 mg/1 was calculated, based upon a noobserved-adverse-effect-level of 405 mg/ m* (100 ppm). A variety of animal species exposed to this dose level for 6 to 7 hours/day. 5 days/week yielded no adverse effects as measured by general appearance, behavior, mortality rates, growth rates, organ function and blood chemistry. An undertainty factor of 1000 was used to account for an animal study with no equivalent human data, and for the use of a study of less than lifetime exposure. One hundred percent exposure from drinking water and a 70 kg adult consuming 2 liters of water per day were assumed in the calculations. 1.2-Dichloroethane has been shown to significantly increase tumor incidences at several sites in both rata and mice when administered by gavage. but not following inhalation exposure. In the NCI bioassay, doses of 47 or 95 mg/kg in com oil administered by gavage to rats and 97 or 195 mg/kg given to male mice and 149 or 299 mg/kg given to female mice were shown to increase the incidence of several types of tumors. 1.2Dichloroethane has also been shown to be mutagenic in a number of biological systems, including Drosophila melonogaster. Salmonella typhimurium and . coli. The 1ARC has concluded that sufficient evidence of carcinogenicity in animals exists for 1.2-dichloroethane. The World Health Organization (1961) has recommended a tentative guideline value of 10 ug/1 for 1.2-dichloroethane in drinking water. EPA's Carcinogen Assessment Group has used the linearized non-threshold multi-stage model to calculate projected excess cancer risk estimates extrapolated from high-dose animal studies. For 1.2-dichloroethane, these estimates were based upon the NCI bioassay data. Calculated risks corresponding to various doses are listed in Table 7. Vinyl chloride. Acute and chroic toxicity studies with vinyl chloride have shown the major non-carcinogenic effects resulting from high dose exposures to be congestion and edema of the lungs and hyperemia of the kidney and liver. Other non-carcinogenic effects have been noted, including disturbances of the central nervous system, pulmonaryInsilffl'ciency. cardiovascular manifestations, gastrointestinal symptoms and acroosteolysis- A suggested Adjusted ADI for vinyl chloride of 0.06 mg/1 considering adverse health effects not including carcinogenic risk, was calculated based upon an oral toxicity study in rats in which a variety of carcinogenic and non-carcinogenic effects were observed at all dose levels. A minimal-effect-levei of 1.7 mg/kg was used in the calculations, as histopathological changes in the liver including clear-cell foci, extensive necrosis, cysts and livercell polymorphism were observed at this dose level. An uncertainty factorof 1000 was applied to account for an animal study where the no-observed-adversceffect level was not identified. One hundred percent exposure from drinking water and a 70 kg adult consuming 2 titers of water per day were assumed in the calculations. Vinyl chloride has been shown to have carcinogenic effects in animals and humans. Animal studies have demonstrated the production of liver angiosarcomas, mammary carcinomas, pulmonary angiosarcomas and other tumor types in rats following oral exposure and carcinogenic effects in mice, rats and hamsters by inhalation exposure have been reported. In humans, studies have linked vinyl chloride with angiosarcoma of the liver and other forms of neoplasm. The 1AKC has concluded that sufficient evidence of carcinogenicity exists for vinyl chloride from animal studies and human studies, and that vinyl chloride should be considered a human carcinogen with target organs of the liver, brain, lungs and haemo-lymphopoietic system. Vinyl chloride was shown to be mutagenic in the test system using metabolically activated Salmonella typhimurium, E. coli K12 strain, in germ cells of Drosophila and Chinese hamsteT V79 cells. EPA's carcinogen assessment Group has used the linearized non-threshold multi-stage model to calculate projected excess cancer estimates extrapolated from high dose animal studies. For vinyl chloride, these estimates were based upon an inhalation study in rats in which vinyl chloride concentrations ranging from 50 to 50,000 ppm resulted in a total tumor incidence rate of 17 percent to 62 percent, respectively. The NAS has also used the multi-stage model to calculate excess cancer risk values. They based their estimates upon the same study as did CAG (Maltoni, et ai, 1975), except ingestion data instead of inhalation data were used. The NAS CONFIDENTIAL. fU^eot to Protective Order in' v. Conooo* Ino,. No. 90-4837 !14tfc Judicial District Court ''asieu Parish, Louisiana CMA 007910 Federal Register / Vol. 49. No. 114 / Tuesday, June 12. 1984 / Proposed Rules 24343 risk estimation used ingestion exposure and thus may be more appropriate for estimating risks from drinking water exposure. Calculated risks corresponding to various doses are listed in Table 7. In addition, data from a recent draft CAG calculation using an ingestion study in rats (EPA, 1984) are also included for comment. Benzene. The toxic effects of benzene in humans and other animals include central nervous system effects, hematological effects as well as immunological effects. The toxicity of benzene to the hematopoietic system of ' humans experiencing chronic exposure to benzene is well documented. Repeated exposure effects include myelocytic anemia, thrombocytopenia and leukemia. In laboratory animals, leukopenia is the most commonly observed effect of chronic benezene exposure. A suggested Adjusted ADI for benzene, considering adverse health effects not including carcinogenic risk, was calculated based upon data from a gavage study in rats in which leucopenia was observed at specific dose levels. A value of 0.025 mg/1 was calculated using a no-observed-adverse-effect level of 1 mg/kg and an uncertainty factor of 1000. This uncertainty factor was used to account for an animal study with no equivalent human data, and for the use of a study of less than lifetime exposure. One hundred percent exposure from drinking water and a 70 kg adult consuming 2 liters of water per day were assumed in the calculations. Benzene has been shown to be carcinogenic in Sprague-Dawley rats, causing tumors at dose levels of 50 mg/ kg and 250 mg/kg. An increase in zymbal gland carcinomas. luekemias and mammary carcinomas in rats has also been observed. Toxic effects on bone marrow cells of rats and other laboratory animals from benzene exposure include changes in chromosome number and chromosome breakage. These types of effects have also been observed in humane. EPA's Carcinogen Assessment group has used the linearized non-threshold multi-stage model to calculate projected excess cancer estimates extrapolated from high-dose animal and human studies. For benzene, these estimates were baaed upon an epidemiologic study of workers exposed to benzene vapors on their jobs. Calculated risks corresponding to various doses are listed in Table 7.. l.l-Dichloroethylene. 1.1* Dichloroethylene has been shown to cause liver and kidney injury in animals from high dose exposures. Liver damage in rats, mice and guinea pigs has been documented, along with renal toxicity, CNS depression and sensitization of the heart. An Adjusted ADI of 350 pg/1 for 1,1dichloroethylene considering adverse health effects not including the potential carcinogenic risk was calculated based upon toxic liver effects using a NOAEL of 10 mg/kg and 100 percent exposure from drinking water. The NAS (1983) has calculated a chronic, suggested-no-adverse-response level (equivalent to an adjusted ADI) of 0.1 mg/1 based upon non-carcinogenic effects only for 1,1-dichloroethylene. from data in the National Toxicology Program bioassay (1982) in rats and mice. A no-observed-adverse-effect level of 2 mg/kg was used and an uncertainty factor of 100, and complete absorption from the GI tract Twenty percent exposure from drinking water and a 70 kg adult consuming 2 liters of water per day were assumed in the calculations, along with conversions from a 5 d/week dosing regime to a 7 d/ week exposure. l.l-Dichloroethylene was found to be mutagenic with microsomal activation in Salmonella typhimurium and E. coli test systems. However, mutagenicity was not observed with V79 Chinese hamster cells or in dominant lethal studies in mice and rats. l.l-Dichloroethylene was shown to produce kidney adenocarcinomas in mice and rats in one study (Maltoni. 1977). However.'most of the other studies have failed to demonstrate significant carcinogenic activity of the chemical. A study by the National Toxicology Program (1982) examined 1.1-dichloroethylene exposures of 1 mg/ kg or 5 mg/kg 5 times per week in rats and 2 mg/kg or 10 mg/kg 5 times per week in mice. In this bioassay, there was no evidence that 1.1dichloroethylene was carcinogenic for either the rats or the mice. However, there was some question as to whether the maximum tolerated dose had been used in this study. The NAS (1983) has concluded that information on 1.1dichloroethylene is not sufficient to reach a definite conclusion on the carcinogenicity of the compound. EPA's Carcinogen Assessment Group found 1,1-dichloroethylene to have limited evidence of carcinogenicity in animals. They have used the linearized, non-threshold, multi-stage model to calculate projected excess cancer estimates extrapolated from high-dose animal studies. For 1,1-dichloroethylene, these estimates were based on results of inhalation studies in mice and rats. Calculated risks corresponding to various doses are listed in Table 7. EPA's SAB has recently questioned validity of this study result. This tentative classification of 1,1-DCE as a carcinogen will be reexamined during the comment period. Comment is ^ solicited in this regard. M . p-Dichlorobenzene. Non-carcinogenicTM adverse effects observed in animal studies include liver and kidney damage, porphyria, pulmonary edema and congestion and splenic weight changes. In humans, exposure to fairly high concentrations of the dichlorobenzenes has been reported to result in anorexia, nausea, yellow atrophy of the liver and blood dyscrasias. A suggested Adjusted ADI of 3.75 mg/ 1 for p-dichlorobenzene considering adverse health effects other than carcinogenic potential was calculated. This value was based upon the rat subchronic gavage study which served as the dose range-finding study for the NTP bioassay. The ADI was based upon a NOAEL of 150 mg/kg/day. ` Uncertainty factors of 100 and 10 were used to account interspecies extrapolation and use of data from an exposure duration significantly less than lifetime. p-Dichlorobenzene has been shown to induce abnormal mitotic division in higher plants. The compound was not seen to be mutagenic when tested in the Salmonella typhimurium or coli WP2 systems, and no evidence of mutagenicity in animals has been reported to date. In June 1980. a carcinogenesis bioassay of p-dichlorobenzene in mice and rats was undertaken by the National Toxicology Program. Doses of 200 mg/kg or 600 mg/kg were administered by gavage to both sexes of mice and to female rats. Male rats were given 150 or 300 mg/kg. The results of this study have not yet been released. V. RMCL Development Rationale The ANPRM requested public comment on the appropriate approach to deal with VOCs in drinking water, specifically requesting consideration of the following; What approach should be followed under the SDWA to reduce human exposure to VOCs? For which VOCs should regulations beset? What approach should be followed in setting RMCLs for suspected carcinogens? Each of these issues is discussed below in regard to the rationale used by the Agency in development of this proposal and the Agency's consideration of the public comments, the cc:Tr~2~"TiAi, Subject to Protective Order In' Ross v. Conoco, Inc., No. 90-4837, 14th Judicial District Court ^ 1 '*~T?aic'a'slmu'^ariiihTrtwii CMA O 1 I 24344 Federal Register / Vol. 49. No. 114 / Tuesday, June 12, 1984 / Proposed Rules requirements of the SDWA. and the available scientific information. VOCs: Regulatory Approach Alternative approaches. The major alternatives considered for limiting human exposure to VOCs in drinking water as discussed in 47 FR 9350 are provided below. (1) No federal regulations. Provision ofhealth advisories for State action as appropriate. Health advisories and advice on treatment and analytical methods are currently being provided to States and public water systems for use in dealing with incidents of VOC contamination. Each State would design its own control strategies to address incidents of contamination on a case-by-case basis or state-wide. Health advisories were developed to deal primarily with isolated incidents of short-term contamination in lieu of standards and not as a substitute for MCLs. Experience has shown that, as would be expected. States have interpreted and applied the health advisories in different ways. Some States have applied the health advisories as if they were standards or considered adopting them as State standards. (2) Set federal monitoring regulations andprovide health advisories for State action as appropriate. This option would set monitoring requirements for VOCs under section 1445 and provide health advisories for State action as needed. This alternative would reault in aU public water systems determining if they have VOCs in their drinking water and could be proposed and promulgated in a shorter period of time than alternative 3. Different States would probably adopt different control options and action levels. (3) Set Primary Drinking Water Regulations far certain ofthe VOCs. This option would set RMCLs, MCLs, monitoring and reporting requirements for a number of VOCs and would reault in consistent nation-wide controls on VOCs. Proposed regulatory approach. The SDWA authorizes EPA to establish RMCLs for "each contaminant which, in [the Administrator's] judgment * * * may have any adverse effect on the health of persons" section 1412(b)[l)(B]. A primary drinking water regulation is to be established for each contaminant for which an RMCL is established. Sectipn 1412(b)(2). fat implementing this broad statutory mandate. EPA is considering the following factors to select contaminants appropriate for regulations. These include: Whether the frequency of occurrence and the concentrations detected in drinking water and the information hecdnfe available will be extent of the population exposed addressed in Phase n and later warrant establishment of national iterations of the National Primary primary drinking water regulations. Drinking Water Regulations along with Whether the available toxicology other contaminants. Specific VOCs data are sufficient to warrant a considered in this proposal are those determination that adverse effects may be known or anticipated at levels found in drinking water. Notwithstanding these factors, EPA feels that primary drinking water regulations may be appropriate in some instances for substances which to date have not been found at high concentrations or frequencies in drinking water, but where in the Administrator's judgement it would be appropriate to anticipate possible future potential for drinking water contamination from spills or improper disposal. Other factors that must be considered as part of the decision on the type of regulation (MCL or treatment requirement) include: Whether monitoring is technically and economically feasible. that have appeared to be the highest priority for regulation based upon occurrence, health risk considerations and available data. Several VOCs have been found across the country in numerous drinking water supplies. In the GWSS, 21 percent of systemshad at least one VOC detected. ETA has concluded that sufficient health effects date are available to cause concern about potential human exposure to certain VOCs via drinking water. Various of the VOCs are suspected or proven mammalian carcinogens, some are known human carcinogens, some are active in certain mutagenic test systems and exposure to certain of the VOCs at high doses has shown other non-carcinogenic toxic effects. EPA recognizes that Whether treatment technologies are interpretation of health risk data raises available to reduce the contaminants to numerous scientific issues. However, appropriate levels. drawing upon the conclusions/ In addition, some guidance was recommendations of the NAS, IARC and provided in the legislative history to the the NDWAC. EPA believes that the data SDWA Senate Report on possible adequately demonstrate concern such candidates for Revised Regulations. that RMCLs and primary drinking water Contaminants listed in the following regulations are warranted. Thus, EPA sources were expected to be considered has determined that human exposure to for regulation, certain VOCs via drinking water may World Health Organization: "Maximum Permissible Concentrations of Harmful Substances in the Water of Water Courses used for Hygienic and Domestic Purposes (1970)," have an "adverse effect upon the health of persons" thereby warranting regulatory action. Selection of VOCs for Regulation World Health Organization: "European Standards for Drinking Water," 2nd edition. Revised, Geneva (1970). National Institute of Occupational Safety and Health annual Hst of toxic substances. Toxic Substances listed under section 307 of the Federal Water Pollution Control Act Information provided by the NAS in the Drinking Water and Health series is an additional source. While numerous contaminants are listed in these sources, this proposal fat Phase 1 of EPA's National Primary Drinking Water Regulations addresses a limited number of contaminants in the VOC category found in drinking water. This section provides a discussion of the factors used to select the specific contaminants for which RMCLs are proposed at this time. VOCs that were not included in this proposal will be reconsidered in Phase 0 of the Revised Regulations as additional data become available. Factors considered. A number of factors were considered in determining which VOCs should be regulated: however, there is no established formula or set criteria for these determinations. The SDWA states that regulations should be set for contaminants that the Administrator determines "may have any adverse effect upon the health of persons" but little additional guidance Because of EPA's desire to avoid delay in developing regulations for certain VOCs that have been detected in ground waters and the need to prioritize the expenditure of limited resources, only nine VOCs are addressed in this initial proposal. Other VOCs for which was provided. Obviously, it is impossible to consider for regulation every chemical that may appear in drinking water and that theoretically may adversely affect health in some remote circumstances. What is needed is some prioritization of contaminants in sufficient occurrence and fceaUh@$rt(tf j Itfrtaktef^ajfr so that a reasonable SuTJJect to Protective Order in Boss v. Conoco, Inc,, No. 90-4837 14th Judicial District Co"'"t Calcasieu Parish, Ton''' CMA 007911 Federal Register / Vol. 49. No. 114 / Tuesday, June 12. 1964 / Proposed Rules 24345 number of contaminants of sufficient concern can be addressed in regulations. To best employ its resources, EPA must select contaminants for regulation based upon considerations that will advance the goals of the Act to assure the safety of drinking water. EPA believes that the most relevant criteria are the: (1) Analytical ability to detect a contaminant in drinking water, (2) the frequency and level of occurrence and population exposed, and (3) potential health aspects of the contaminants. In addition EPA considers regulation when there are sufficient incidents or contamination potential such that national guidance in the form of a Primary Drinking Water Regulation is desirable to assist States and public water systems which must determine appropriate responses. Analytical methods. Analytical methods must be available such that the presence of the chemicals in water can be validly determined. This factor is an important part in determining whether the substance can be regulated and whether an MCL or a treatment technique regulation should be promulgated. National or limitedsignificance. Consideration of occurrence data encompasses both the frequency of occurrence, the level of occurrence and ther'extent of the population exposed. The occurrence data allow EPA to determine whether contamination of drinking water represents isolated or localized incidents of contamination more appropriately dealt with by States, or whether contamination has occurred or has the potential for occurring in numerous locations across the country involving a sufficient number of water supplies and population exposed to warrant action under the Safe Drinking Water Act In the ANPRM for Phase Q of the NPDWR. 48 FR 45502, et seq,, EPA described a categorization system for differentiation between widespread and limited contamination potential. Health effects. Consideration of the potential health effects of a chemical encompasses the: (1) Suitability of the available data for assessing the toxicology of the chemical, and (2] the possibility of human health concern from exposure from drinking water. WheiUt is possible scientifically, section 1412(e)(3) of the SDWA also requires consideration of the impact of the following: (A) The existence of groups or individuals in the populstion which are more susceptible to adverse effects than the normal healthy adult. (B) The exposure to contaminants in other media than drinking water (including exposures in food, in the ambient air and in ' occupational settings) and the resulting body burden of contaminants. (C) Synergistic effects resulting from exposure to or interaction by two or more contaminants. () The contaminant exposure end body burden levels which alter physiological function or structure in a manner reasonably e.g., p-dichlorobenzene and 1,1.1trichloroethane. The MCL would provide guidance that no action was necessary for these systems with less than that level: without regulation!, these types of situations have met widely varying responses by States and public water suspected of incressing the risk of illness. systems. Regulations can provide a These factors were addressed in assessing the potential health effects of each of the VOCs and are discussed in each of the health effects criteria documents as referenced in section VIL However, applicable data are seldom available for any of these factors except B (to a limited extent) which ia addressed in both the occurrence and health effects documents. Other considerations. Additional factors considered in determining which VOCs should be regulated and how ere discussed below. One approach that might be basis for rational and uniform responses to incidents of contamination. Potential impact The potential impact of setting regulations can be conaidered in a general manner however, this factor ia primarily considered during establishment of MCL*. This evaluation conaidera potential burdens including such factors at the affordability of treatment systems, the technical feasibility of meeting MCL*. and other possible impacts such aa monitoring and reporting.- The result* of setting regulations for considered would be to set RMCLs by VOC* will vary widely from no impact category, Le,, the same RMCL for each to installation of treatment systems for VOC or subcategories of VOCs. In effect reduction of VOCs. Recognizing that the this is being proposed for the category great majority of public water systems determined to be non-threshold do not have VOCs in the drinking water, toxicants. However, a categorical RMCL the only burden on these systems would for non-carcinogenic VOCs is not be monitoring and reporting. Theae scientifically supportable due to burdens could be minimized through differing relative toxicities of individual flexible monitoring requirements (see 48 substances (different thresholds) and FR 45502) that would provide states with different toxic endpoints. * authority to determine appropriate * Strength of evidence. Pertaining to requirements beyond the national either the extent of contamination or to minimum. In addition, the VOCs are the potential health risks of exposure, omewhat unique in the sense that the amount of available data of several of them can be analyzed for in a sufficient quality on a certain chemical single analytical procedure. was considered. For example: Other factors. Surrogate parameters --A chemical proven to be a human or aggregate parameters may be needed carcinogen, even though occurring to take into account other potential relatively infrequently in drinking effects not considered in setting RMCLs water supplies might be appropriate and MCLs for individual chemicals, such for regulation, e.g., vinyl chloride and . benzene, --A chemical occurring at a higher as possible additive or synergistic risks of simultaneous exposure to more than one VOC. frequency in drinking water supplies Proposed VOCs. The ANPRM listed but for which the strength of evidence fourteen VOCs being considered for on potential health risks was weaker regulations. Detailed occurrence and could be appropriate for regulation, health effect* information were e.g., trichloroethylene, provided for six of tha fourteen VOCs. tetrachloroethylene. carbon Since the ANPRM was published. EPA tetrachloride, 1,1-dichloroethylene, completed the Ground Water Supply 1.1.1-trichioroethane. 1.2- Survey (GWSS) in which twenty-nine dichloroethane. VOCs were looked for in each sample Nationalguidance to address using the "purge and trap" analytical incidents ofcontamination. Regulation* provide a benchmark for potential action by State and local officials in evaluating incident* of contamination. In certain cases, this factor may bt a major consideration in determining if regulations are appropriate. For example, regulations would be appropriate for a chemical that occurs but at levels normally below those associated with potential health risks. procedure employing gas chromatography (Method 502.1 and Method 503.1. U.S. EPA. Environmental Monitoring and Support Laboratory). As shown in Table 1, not all of the ANPRM list of 14 VOCs were detected in the GWSS. Based upon the above considerations, public comments and recommendations of the NDWAC and other information, EPA has concluded that these chemicals nr>7T PTTvaT'PTar ""v.-v-to JTubJeot to Protective Order lit Boss v.- Conoco, Inc., No. 90-4837 il4tE Judicial District Court (y, 'Calcasieu Parish,, Louisiana CMA 0 0 7 9 1 3 -A 24346 Federal. Register / Vol. 49, No. 114 / Tuesday, June 12, 1984 / Proposed Rules "may have an adverse effect upon the limited evidence. This report is currently 0.8%, max: 0.9 pg/k median: 0.7 pg/1. health of persons" and that RMCLs and being evaluated. > SteWDatafN/A.'* primary thinking water regulations Carbon tetrachloride. Occurrence Health Effects: non-carcionogenic (at under Section 1412 should be proposed (may be a contaminant in chlorine) high doses): kidney and liver damage, at this time. Thejrare: GWSS (Random): 3.2% max: 18pg/k pulmonary edema and congestion, trichloroethylene tetracholoroethylene 1.1.1- trichloroethane carbon tetrachloride 1,2-dichloroethane benzene vinyl chloride p-dichlorobenzene 1.1- dichloroethylene median: 0.4 pg/1. GWSS (Non-random): 3.1%: max: 15 pg/k median: 0.5 pg/1. State Data: 368 positive/2846 sampled, max: 1400 pg/1. Health Effects: Non-cardnogenic effects: liver effects such as fatty liver with centrilobular necrosis. Carcinogenic effects: mutagenic in some test systems: carcinogenic in NCI test: mice, rats, hamsters; suffirient evidence. spleneic weight changes. Carcinogenic effects: NTP test underway. Other VOCs. Several additional VOCs listed in the ANTRM (47 FR 9350) have been found in some drinking water samples but the available data has been fudged to be insufficient to propose RMCLs at this time. Cia-14-dichloroethylene and trans1,2-dichloroethylene As presented previously, the NDWAC lJ-Dichloroethane. Occurrence: These two VOCs have not been tested recommended developing regulations for GWSS (Random); 0.6%, max: 14 pg/k for carcinogenicity by the NTP and the first five of the above nine VOCs. median: 04 pg/1. GWSS (Non-random): adequate atudies on non-carcinogenic Their rationale was based upon an 1.5%; max: 10 pg/k median: 24 pg/L toxicity have not been conducted. evaluation of the available occurrence State Data: 177 positive/1793 sampled, Chlorobenzene and health effects data for each of the max: 2,100 pg/1. While aome occurrence has been VOCs. The NDWAC evaluated the Health Effects: Non-cardnogenic (at reported by a number of States, the information in September 1962. Since high doses): central nervous system GWSS did not detect any chlorobenzene that time additional data have become depression, liver and kidney change, in the random sample: however, it was available and the Agency has concluded gastro-intestinal distress, adrenal and found twice in the non-random sample. I that four additional VOCs warrant regulation. pulmonary effects, circulatory disturbances. Carcinogenic effects: The toxicology evaluation has not been completed. The background occurrence and mutagenic In most test systems: Trichlorobenzene(s) health effects data used as the basis for carcinogenic in NCI test: mice, rets: States have detected trichlorobenzene determining which VOCs warranted sufficient evidence. in a number of water samples; however regulations is summarized below. Vinyl chloride. Occurrence: GWSS the number of drinking water versus Trichloroethylene. Occurrence: GWSS (Random): 0.2%; max: 1.1 pg/1; median: non-drinking water incidences could not (Random): 6.4%; max: 78 pg/k medium: 1.1 pg/1. GWSS (Non-random): 1.3% be determined from the data. In 1.0 pg/L GWSS (Non-random): 12.7%: max: 8 pg/k median: 2.7 pg/1. State addition, analytical difficulties in max: 130 pg/k median: 14 pg/L State Data; 128 positive/1793 sampled, max: analyzing samples in the GWSS Data: 624 positives/4228 sampled, max: 380pg/L precluded obtaining representative 510.000 MlI. Health Effects: Non-cardnogenic (at occurrence data. Health Effects: Noo-cardnogenic high doses): congestion and edema of Dichloromethane effects (at high doses): fiver and kidney the lungs, hyperemia of the kidneys and Because of problems of laboratory damage, central nervous system effects, liver. Carcinogenic effects: mutagenic; contamination and quality asaurance. depression in myocardial contractility. carcinogenic in animal studies: mice, the available occurrence data for Carcinogenic effects: mutagenic hi some rats, hamsters: sufficient evidence for dichloromethane was not considered - test systems; carcinogenic in NCI test human carcinogenicity. reliable. In addition, the NTP initial mice. Limited evidence. Benzene. Occurrence: GWSS report on carcinogenicity has been Tetrachloroethylene. Occurrence: (Random): 04% max: 15 pg/k median: 3 withdrawn and the NTP is currently GWSS (Random): 74%; max: 23 pg/k pg/1. GWSS (non-random): 1.7%; max: 12 conducting an fiwfepth audit of the data. median: 0.5 pg/1. GWSS (Non-random): pg/1; median: 14 pg/1. State Data: 4 These VOCs and several others will 94%; max: 69 pg/k median: 0.7 pg/L positive/845 sampled, max: 17 pg/L ba considered in the Phase 11 portion of State Data: 628 positive/3636 sampled, Health Effects: non-cardnogenic: the Primary Drinking Water Regulations max: 1,000 ug/L central nervous system effects, when sufficient occurrence and Health Effects: Non-carcinogenic hematological and immunological toxicology data become available. effects (at high doses); central nervous affects. Carcinogenic effects: sufficient Among the other compounds being system depression, fatty infiltration of evidence for human carcinogenicity. evaluated are such VOCs as ethylene liver and kidney, tissue damage. 1. l-DichJoroethyiene. Occurrence: dibromide, l.l-dichloroethane, xylenes, Carcinogenic effects: carcinogenic in GWSS (Random): 1.9%. max: 84 pg/k toluene, bromobenzene. NCI test mice; limited evidence. median: 03. pg/1. GWSS (non-random): dibromochloropropana. 14- 1,1,1-Trichioroethane. Occurrence: 3.1%, max: 34 pg/k median: 04 pg/L dichloropropane. and ethylbenzene (see GWSS (Random): S.8%: max: 18 pg/k State Data: NA. ANPRM. October 5.1983.48 FR 45502). median: 04 pg/L GWSS (Non-random): Health Effects: non-cardnogenic Other chemical* in the random GWSS 104%; max: 21 pg/k median: 14 pg/L State Data: 715 poaitive/3330 sampled, max: 2450 pg/L Health Effects: Non-carcinogenic effects (at high dotes): central nervous -, system depression, increase in fiver weight cardiovascular changes. Carcinogenic effects: carcinogenic In preliminary report from NTP test mice;' effecta (at high doses): liver and kidney damage, renal toxicity. CNS depression and sensitization of the heart Carcinogenic effects: mutagenic, carcinogenic in one animal study: mica and rats; limited evidence. pJlichlorobenzene. Occurrence: GWSS (Random): 1.1%, max: 14 pg/L median: OJ pg/L GWSS (Non-random): for which no occurrence information waa obtained but which will receive some consideration in Phase U or other iterations include: 1.1.2-trichloroethane. 1.144-tetrachloroethane, 1.1.14tetrachloroethane, u-propylbenzene, o- chlorotoluens, p-chlorotoluene, ni dichlorobenzene, o-dichlorobenzene, styrene, isopropylbenzene. CONFIDENTIAL Subject to Protective Order Ross v. Conoco, Inc., No. 90 14tfi Judicial District C~ Hal poQ-f on Pa-n-f qK . T.piH CMA 0 0 7 9 1 4 Federal Register / Vol. 49. No. 114 / Tuesday. June 12. 1984 / Proposed Rules 24347 Total VOCs In addition to regulations for individual VOCs, the inclusion of RMCL* and MCLs for total VOCs (TVOC) is being considered. TVOC is not formally proposed in this regulation. Public comments are being solicited on whether it would be proper to include TVOC in drinking water regulations or in supporting guidance. TVOC would represent summation of the levels of the individual VOCs for which RMCLs and MCLs have been set The objective of a TVOC standard is to provide some additional protection from simultaneous exposure to multiple VOCs. As indicated in Table 4, drinking water often contains several VOCs. Generally, toxicology has not yet been able to provide a scientifically based conclusion on possible effects of simultaneous exposure to more than one chemical. Chemicals are normally tested separately and the possible synergistic, antagonistic, or additive health effects are not known. However, the NAS suggestion in this area was that in the absence of any other procedures, exposure to multiple carcinogens could be assessed by adding the risk rates. Comment is requested on the technical validity of this approach. The potential problem that EPA feels must be addressed is a situation where a public water system finds several VOCs in its drinking water at levels slightly below the MCLs. For example, assume that MCLs are set for trichloroethylene, tetrachloroethylene, and carbon tetrachloride: a public water system with the following levels would technically be in compliance with the MCLs: CnwuS . ,, UnwN m/i 'm/ *M/ MCL 10 tiff IS |s^l Sm*1 While technically in compliance with the standards, this condition probably represents an increased risk over any single chemical but the question that cannot be scientifically answered is whether this would be significant EPA feels that multiple exposures could be more significant than indicated from fust consideration of individual substances and requests public comments considering the myriad of possibilities in assessing multiple exposures, the costs and feasibility to reduce all the VOCs by application of one treatment technology, and the unknown aggregate health risk and the SDWA intent to err on the side of safety. If an RMCL and MCL for total VOCs (TVOCs) were appropriate, should EPA adopt the NAS suggestion that risks be considered additive be an appropriate approachT RMCLs: Regulatory Approach EPA is to set RMCLs at levels which, "no known or anticipated adverse effects on the health of persons occur and which allow an adequate margin of safety". Section 1412(b)(1)(B). Recommended MCLs are health goals and not enforceable standards. The proposed RMCLs for non-carcinogens can be determined using the scientific procedures set forth previously by calculating an AADL However, determination of the "no effect" levels for carcinogens is a much more complex decision on what constitutes the safe level for non-threshold toxicants. Guidance on levels for the RMCLs was provided in House Report 93-1189 which stated that "It (The RMCL] must include an adequate margin of safety, unless there is no safe threshold for a contaminant In such a case, the recommended maximum contaminant level should be set at zero leveL" EPA has considered the following approaches for setting RMCLs for carcinogens: 1. Set the RMCLs at zero. 2. Set the RMCLs at the analytical detection limit 3. Set the RMCLs at a non-zero level based upon a calculated negligible contribution to lifetime risk. Although one of these is proposed at this time, EPA requests comments on all three approaches. EPA's analysis of these approaches and the issues they raised are provided below. Alternative 1: Set RMCLs at zero. One approach would be to establish RMCLs at zero for substances considered to be non-threshold toxicants. The existence of a threshold for the action of genotoxic carcinogens cannot be demonstrated by current science; thus, if could be conservatively assumed that no threshold exists, absent evidence to the contrary. Since distinctions between mechanisms of action of most carcinogens also cannot be conclusively made at this time, virtually all substances determined to be "carcinogens" would be assumed to be "non-threshold". Variation of this approach would be to limit the selection of RMCLs at zero only for those substances known to function by genotoxic processes, or perhaps only those determined to be human carcinogens, or only those for which "sufficient" rather than "limited" evidence of mammalian carcinogenicity exists.. Setting RMCLs for carcinogens at zero would follow the guidance provided in House Report 93-1185 and would express a general philosophy that as a goal carcinogens should not be present in drinking water. The Agency believes that the RMCLs (as a goal) should express the ideal concept that drinking water should be free from avoidable contamination and risk and that quality degradation should not be permitted. If RMCLs are set at zero, some explanation may be needed to differentiate an RMCL from an MCL that would not be zero, since MCLs consider factors such as potential health risk, costa of treatment and feasibility of meeting the MCL If these factors changed substantially, MCLs would need to be reexamined. Alternative 2: Set RMCLs at the analytical detection limit Due to limitations in analytical techniques, it will always be impossible to say with certainty that the substance is not present In theory, RMCLS at zero will always be unachievable (or at least not demonstrable). While zero could be the theoretical goal for carcinogens in drinking water, in practice, a goal of achieving the analytical detection limits for specific carcinogens would have to be followed. One possible approach would be for EPA to specify RMCLs for carcinogens based upon defined state-of-the-art analytical detection limits. The verifiable detection limits (i.e,, the RMCLs) would probably fall in the vicinity of 1 pg/1 depending upon the specific VOC. EPA believes this approach is justifiable in that zero is analytically undefinable and the detection limit may be the functional equivalent of zero. Of course, analytical detection limits are also moving targets as the state-of-the-art of analytical chemistry progresses, but at least they do provide a measurable target Alternative A* Set RMCLs at a non zero level based upon a calculated negligible contribution to lifetime risk. Alternative 3 would establish a non-zero level as the RMCL A level could be selected that would present a negligible risk. In practical terms, such a low nominal risk would effectively preclude any discernable adverse effect on the health of the population and, because of the conservative nature of the risk calculation process, may not result in any actual adverse effects on an individual. EPA would have to conclude that this very low risk would result in "no known or anticipated adverse effect on the health of persons and which allows an adequate a margin of safety". This approach would provide some CONFIDENTIAL1 Subject to Protective Order in Ross v. Conoco, Inc., Ho. 90-4837 14th Judicial District' Court Calcasieu Parish.. Louisiana' CMA 00791^ 24348 Federal Register / Vol. 49, No. 114 / Tuesday, June 12. 1984 / Proposed Rules quantitative guidance to public water systems of the ultimate goal which they might wish to use in the operation of water treatment facilities and in the design of future planned facilities. However, it should be recognized that just as with analytical detection limits (Option 2) a calculated risk target would also be moving target because: (1) calculation methods change, and (2} the subjective determination of what is a negligible risk might change. One possible variation of Option 3 would be to set RMCLs as a range of finite risk levels. This alternative would recognize the lack of accuracy and precision of risk calculations and the inherent difficulties in selecting one finite level as the only appropriate health goal in view of the numerous scientific uncertainties of risk estimates. However, this approach has a number of disadvantages including: lack of national uniformity and lack of specific guidance from EPA. If a non-zero level is determined as appropriate for the RMCLs, two questions must be considered. (1) What level should be used as representing the "no effect" level? (2) How can an "adequate margin of safety" be incorporated into the finite risk level? The NAS principles (Drinking Water and Health, Vol. l) state that human exposure to carcinogens should be addressed in terms of risk rather than safe or non-safe. Because zero is not definable in an analytical sense, rather than speaking in terms of zero concentrations for carcinogens RMCLs for carcinogens could be set at levels at which the risks are so small that they are considered virtually nonexistent Determination of RMCLs for carcinogens at a finite level would be based on available science and the only quantitative tools available are cancer risk models. These are based upon animal studies and none of the models is experimentally verifiable as there is no scientifically valid method for determining the actual risks at low environmental exposure levels. Scientific issues surround their use in such areas as the data used, extrapolation techniques, and various factors in the analysis. Risk models are recognized as imperfect but they are the best tool available for estimating toxic potency or risk at low exposure levels. The commonly used risk models are generally conservative in their estimation of human risk of exposure to a contaminant. Selection of a target risk based upon a conservative risk model such as the linearized multi-stage model is arguably in accord with the SDWA, which requires the RMCL to be set at a no effect level "with an adequate margin of safety." The Agency believes that there is no exact or precise way to determine this level. The decision is judgmental---not strictly based upon science but upon a social judgment on what constitutes a negligible risk. Federal regulations for environmental contaminants have generally fallen in the 10-4 to 10-* lifetime risk range, as calculated from a linear multi-stage model. Most of those decisions incorporated consideration of costs and feasibility. The negligible risk concept considered here is based strictly on individual risk rates and exposure. It does not include other economic or technical considerations that are part of setting the enforcement standards (i.e.. the MCLs). The level for the MCLs (not RMCLs) would thus be considered to be the upper limits of risk that are considered to be acceptable based upon our current evaluation of the feasibility and costs of controls. Under this approach to setting drinking water RMCLs, EPA has considered two risk levels as possibly representing an upper limit for a risk: one in 100,000 (10**) probability per 70 years of exposure and one in 1.000,000 (10'1 probability. Ah incremental lifetime risk level of 10**wpuld probably be more representative than 10'* as the "no effect" level for these chemicals in drinking water with a margin of safety as envisioned by Congress. The NDWAC stated that 10~4 would be an appropriate target However, a level of 10~*la the level of concern that commonly has been discussed as the lower limit of concern over the potential health risks of exposure, especially for the generally involuntary risk from exposure to a drinking water contaminant In addition, if RMCLs were to be set at a non-zero level use of the linearized multi-stage model would often appear to be more appropriate than others to meet the Congressional intent The conservative nature of the model could actually mean that the real risk of exposure was probably lower (e.g.. 10'' or 10* ^ if any risk actually exists (assuming a non-threshold mechanism were operative) because the model was structured to be conservative and because of the nature of many of the assumptions in the model. As an example of what 10** would mean in terms of the U.S. population, a total of 20 cases of cancer would result if 10 percent of the population were exposed at a dose level equivalent to a 10~* risk for 70 years. Stated another way, that would be one-third of a cancer case per year as an upper limit in the U.S. population compared to the appropriatelySOOJOOO annual cancer death's that occur. The actual number of cases attributable to that particular substance would probably be less and perhaps none at all would occur unless some additive or synergistic interaction with other substances resulted in enhanced toxicity. Proposed RMCLs: Conclusions. This proposal selects RMCLs for potential carcinogens at zero: the alternatives were carefully considered in view of the intent of the SDWA and public comments. It should be recognized that regardless of which of the three alternatives is ultimately selected for the RMCL it is unlikely that the MCL for a particular substance would be affected, since normally all of the approaches would yield targets that are likely to be below levels that are "technically and economically feasible" using available technologies. MCLs will be set as close to the RMCLs as feasible. Preliminary analyses indicate that the MCLs may fall roughly in the range of 5 to SO pg/1 for most of the VOCs being considered in this proposal. Proposed RMCLs for the following substances considered carcinogenic are "zero": tetrachloroethylene. trichloroethylene, carbon tetrachloride. 1.2-dichoroethane, vinyl chloride, benzene, 1,1-dichloroethylene. The proposed RMCL for 1.1.1trichloroethane is 0.2 mg/1. derived from the calculated AADI of 1.0 mg/1 assuming 20 percent contribution from drinking water to total exposure. If the preliminary NTP report on the carcinogenicity of this compound is affirmed, the RMCL would be zero. EPA would provide formal notice if and when this occurs. The proposed RMCL for pdichlorobenzene (1,4-dichlorobenzene) is 0.7S mg/L derived from the calculated AADI of 3.75 mg/1 assuming 20 percent contribution from drinking water to total exposure. Three of these substances (trichloroethylene, tetrachloroethylene and 1,1-dichloroethylene) have only "limited" animal evidence of carcinogenicity, as this term is used in the 1ARC criteria. Factors which contribute to this classification include lack of replication in multiple experiments or multiple species, as well as defects in particualr studies. In addition, indicators of certain types of tumors, such as in the mouse liver, are considered by some scientists to have less weight than others in predicting carcinogenicity in humans. Data of this type, obtained by com oil gavage, introduces another variable that nOTTFlDEUTlAlt Sub'jec-Tto Proteotivoi?!!? Ross v- Conoco, Ifig.-1 A CMA 007916 Federal Register / Vol. 49. No. 114 / Tuesday.June 12, 1964 / Proposed Rules 24349 complicates interpretation. While determined that one or more of these brief discussions are provided here. evidence for these three substances is of substances should not be treated as These factors are key elements in the a weaker nature than for others that carcinogens, then the AADI calculations determination of the MCL which will be GPA is proposing to regulate as modified by an allocation of 20 percent proposed when the RMCLa are carcinogens, it is nevertheless evidence to drinking water would be the basis for promulgated. that must be weighed by the the promulgated RMCL Methods for removal of these volatile Administrator. The strictly scientific evaluation of such evidence (known as "risk VI. Other Considerations for Public Comment organic chemicals include aeration and granular activated carbon (GAC). The available dAa do not show powdered assessment") can only describe its The next regulatory steps will be activated carbon treatment or strength and weaknesses. EPA's risk ' promulgation of the RMCLa and conventional drinking water treatment assessment is summarized above and proposal of MCLs and monitoring and (i.e., coagulation, sedimentation, and described in detail in the documents reporting requirements. Supporting filtration) to ba sufficiently effective for referenced in Section VIL Health documentation for the MCL proposal long term application. Macroreticular Assessment documents for these three will include: (1) Exposure and risk resins may eventually prove to have substances were reviewed by EPA's assessments, (2) an assessment of value for removing VOCs; questions still Science Advisory Board in April and May of 1984. Those reviews will be considered in this rulemaking action under the SDWA and become part of the record. Decisions about what actions to taka on the basis of the evidence (known aa "risk management"), including decisions about how strong the scientific evidence should be to justify regulating a substance, require policy judgments which must be made by the Administrator, after public comment in the light of the Agency'* etetutory mandate*. EPA strongly believes thet its risk assessments should be consistent among Agency programs. On the other hand, risk management decisions can and ' should vary in the light of differing circumstance* or statutory mandates. It generally available technology. (3) an assessment of available analytical methods and costa of monitoring, and (4) an economic and financial impact analysis. Available information to support several of the assessments is referenced in the next section. The public is requested to review those references and provide comments and other supporting information and data. The public is also requested to comment on the issues and information discussed below on available treatment techniques and coats and current estimates of the potential impact of VOC regulations. Treatment of Control of VOCt Economics, treatment technologies and feasibility are not factors involved in the determination of RMCLa; however exist concerning their use. Data describing actual exhaustive capacity of the resins are not available to define the regeneration frequencies to be expected with the resins. Thus, costs have not yet been estimated for application of resin technology. At this time, substantial operational experience and/or experimental data are available only for aeration and GAC Costs of treatment Preliminary designs and cost estimates have been developed for a hypothetical ground water contamination situation involving trichloroethylene (TCE). Table 8 provides relevant estimated cost information for treatment of TCE at the 90 percent and 99 percent removal levels, respectively, for aeration and GAC technologies. is therefore possible that soma of theat substances might be regulated differently in other Agency programs. TASU 8 --PXEUM1NAAY COSTS SON CONTOQUiNQ TCE M CftlMKINO WATER MNSSsIHt) For example, EPA plana to decide m whether to list several of that* wwd substances as hazardous air pollutants under section 112 of the Clean Air Act TlpsMSMPSM* 100 WO (0.06 mod) vooo ZM0(0.S mgd) 10.000 35.000 (4.0 mgd) The same scientific evidence will b considered along with other factors relevant to that decision. This may or may not lead to a conclusion to list and jwpswMMnwn,. disowns MBs**MCt-SOse* 111| 11 in ttc.000 m 660,000 64 6369.000 6 to regulate them as carcinogens. Public comments are requested on For 00 PW*1 fWMl *, ton 500 MCL 5 mqfi *28.900 6130.000 41 6010,000 11 setting RMCLa for carcinogens at zero, the analytical detection limit and at some finite value based upon risk estimation. Comment is also requested ' on appropriate analytical detection Fpr M ppnm WHOM. .. SOSSMS00 *' MCI M ap* For M pMMM IMIWM. si- SSWBSSOO se1 **ci * S* 626.400 143 636.400 140 664.000 60 14 It KSMOO 10 406,800 22 limits, and on the method for calculating the finite risk value and for determining the risk target Comments are also requested on the RMCLa for non* carcinogenic substances and the assumption of an exposure factor of 20 rw ymm eeewvf i '* <*noji mnuHcwx Ml cun* i RMtOMrw. sor ^ iWSMissa pi . MMMaUA-----------rtd Sfl Mean cants par ocnsmnep tseior* ct SS pmwm ss lying. I and wus-san; iK ns mm pm *** Msmm mm M 1* percent from drinking water, absent quantitative multi-media exposure data. Comment ia also directed to technical Potential Impact ofRegulations the specific chemical, although it determinations. AAD1 calculations, the draft revised CAG risk calculations, and the inclusion of substances with "limited evidence" in the carcinogen category. If, on the basis of the record, tt is The nominal limits at detection attained by the laboratories performing analyses in the GWSS were usually in the (U to OS pg/1 range depending upon appears that precision and accuracy requirements for regulatory compliance determination might require that regulations (MCLs) be set at least one r m tU U VWJ Subject to Protective Order in' PSS, v Conoco, Inc., ff0. 90-4837 14tK Judicial District Court FOJMuuwpu; ny.r'F--1- Calcasleu Parish. ptoii&iifif*. 24350 Federal Register / Vol. 49, No. 114 / Tuesday, June 12, 1984 / Proposed Rales order of magnitude higher. The feasible application of aeration and granular carbon might also lead to MCLs in a similar range, i.e.. on the order of 5-80 pg/1. At this level.'very preliminary projections are that about 1,000 systems would probably need to reduce VOC levels either through treatment technologies or other options such as blending or shutting down wells; most of these would be communities utilizing ground water. Therefore, baaed upon current exposure estimates, risks of most VOCt would not appear likely to represent a high impact nor would regulations result in a significant number of cancer cases avoided based upon total cancer rates and projected risks using the linear multi-stage model. Although VOC contamination is widespread across the country, it is usually at low levels, and the overall population at risk is quite low. Moreover, most VOCs do not appear to be highly potent carcinogens. However, in those communities where exposed levels are relatively high, resulting in correspondingly higher risks per individual control is obviously essential. On the other hand, where that is not the case, the non-quantifiable benefits would probably be of most importance in determining the proper approach. These would include such items as providing federal standards to be used as a benchmark In responding to incidences of contamination, and use in ground water protection and clean-up programs. Economic impact analysis. The proposal of an RMCL is different than proposal of an MCL in that an RMCL is, by law, to be based only on health and safety considerations, while an MCL is to take costs into consideration. Therefore, this RMCL proposal notice does not include an analysis of the economic impacts of various possible RMCLs. However, we intend to fully analyze the probable impacts of the various MCL alternatives, and will report on them at the time an MCL is proposed. Because the economic impact analysis is an import part of the rulemaking decision process, and because some reviewers of this notice may be concerned that insufficient attention is being paid to economic considerations, below is a brief indication of how EPA will conduct the economic analysis of alternative MCLs, and what is considered from die results of the* analysis. Executive Order 12291 and the Regulatory Flexibility Act specify how and when to analyze the probable impacts of a Federal action. In essence. Information on the impacts to industry. consumers and the nation is assembled. Where possible, this information is put in the form of an analysis of the net benefits of the various alternatives. This "regulatory impact" information then becomes a part of the official record in support of whatever action EPA finally takes, and is used by decision-makers when an alternative MCL is selected for proposal, and when final MCL is promulgated. The types of impacts which will be examined for each of the various regulatory and non-regulatory alternatives are of three basic types. The impacts of the alternatives on the water supply industry wil be examined. This will be done by reviewing three elements, the capital cost of technology, the operating and maintenance cost and the feasibility of financing new treatments. The first two elements are derived by the engineering analysis of treatment technologies, and the cost of treatments. The third element, the ability to finance new treatments is derived from an analysis of the water supply industry. A financial model of the industry has been developed by EPA, and this model indicates how likely it is for water systems to be able to finance new treatments, based on the historical financial performance of water systems. The second type of impact is the impact on the consumer. Information on the cost of water to consumers is assembled, based on the cost data prepared during the engineering analysis oftreatment technologies. These costs are compared with the historical cost of water. The third type of impact is the impact on the nation as a whole. The purpose of this analysis is to allow balancing of the cost of a federal action, in this case MCLs, with the benefit to be derived from the action. In some cases, it is not possible to describe the value of the benefits in the same terms as the costs, i.e.. dollars. The benefits which will accrue to the nation .are derived from an analysis of the contamination occurrence, the reduction in human exposure likely to result from an alternative, and the health effect averted by the reduction. The cost of the venous alternative MCLs is more than merely the cost to industry. It also includes the cost to government of implementing the regulation. These national costs are summarized and presented with the national benefits, and this too becomes, a part of the record supporting the proposed and final MCL Because these various analyses are based on estimates, an additional the sensitivity ofanalytical results to the assumptions maded during the analysis. This sensitivity analysis completes the general regulatory and non-regulatory analysis required under E.0.12291. A summary of these analyses will be presented in the preamble of the MCL proposal notice, and full documentation of the underlying analyses will be entered into the formal record of the rulemaking procedure. VIL References The following supporting documentation for this proposal is available on request from die address listed at the beginning of this notice. Beliar. TA. Lichttnberg. ].]. "The Determination of Halogenated Chemical Indicator* of Industrial Contamination in Water by the Purge and Trap Method: Method 9020." US. EPA EMSL #600/4-810S0. Beliar. TA. Lichtenberg. I f "The Analysis of Aromatic Chemicals in Water by the Purge and Trap Method: Method S03.1," U S. EPA. EMSL EPA 600/4-81-057. Love, O. Thomas. Jr. and Richard C. Eilers. "Treatment for tha Control Trichloroethylene end Related Industrial Solvents in Drinking Water." U.S. EPA. Office of Research and Development February 1881. Environmental Science and Engineering. "Treatment for Control of VOCs in Drinking Water," August 1883. EPA Criteria and Standards Division. Draft Criteria Document for Trichloroethylene. EPA Office of Drinking Water. February 188*. EPA Critaria and Standards Division. Draff Criteria Document for Tetrachloroethylene. EPA Office of Drinking Water, February 1884. EPA Criteria and Standards Division. Draft Critaria Document for Carbon Tetrachloride, EPA Office of Drinking Water. February 1884. EPA Criteria and Standards Division. Draft Criteria Document for 1,1,1Trichloroethane. EPA Office of Drinking Water. February 1964. EPA Criteria and Standards Division. Draft Criteria Document for 1.2-Diehloroethane. EPA Office of Drinking Water, February 1864. EPA Criteria and Standards Division. Draff Critaria Document for Vinyl Chloride, EPA. Office of Drinking Water. February 1984. EPA Critaria and Standards Division. Draff Criteria Document for Benzene. EPA Office of Drinking Water, February 1864. EPA Criteria and Standard* Division. Draff Criteria Document for Dichloroethylene, EPA Office of Drinking Water. February ISM* EPA Criteria and Standards Division. Draft Criteria Document for Dichlorobenzene. EPA Office of Drinking Weter, February 188*. EPA Criteria and Standard* Division, Occurrence of Volatile Organic Chemicals vt to Protective Order in Boss v Conoco,' Inc, y Ho. 90-4837 Judicial District Court CMA 007918 Federal Register / Vol, 49. No. 114 / Tuesday. June 12, 1984 / Proposed Rules 24351 in Drinking Water, Benzene. November 1983. EPA, Criteria and Standards Division, Occurrence of Volatile Organic Chemicals in Drinking Water, Dichlorobenzene, December 1983. EPA Criteria and Standards Division, Occurrence of Volatile Organic Chemicals in Drinking Water, Dichloroethylenes, November 1983. EPA, Criteria and Standards Division. Occurrence of Volatile Organic Chemicals in Drinking Water. Trichloroethylene, June 1982. EPA. Criteria and Standards Division. Occurrence of Volatile Organic Chemicals in Drinking Water, Tetrachloroethylenes. June 1982, EPA. Criteria and Standards Division, Occurrence of Volatile Organic Chemicals in Drinking Wetar, Vinyl Chloride, June 1982. EPA. Criteria and Standarda Division. Occurrence of Volatile Organic Chemical* In Drinking Water. 1,1,1-Trichloroethane. June 1982. EPA, Criteria and Standard* Division, Occurrence of Volatile Organic Chemicals in Drinking Water. 1.2-Dichlotoethane, November 1983. EPA, Criteria and Standarda Division, Occurrence of Volatile Organic Chemical* in Drinking Water. Carbon Tetrachloride, November 1983. , EPA. EMSL. 'Total Organic Halida. Method 1 450.1-Interim." EPA 800/4-81-058. EPA. Office of Health and Environmental Assessment. Draft Health Assessment Document for Vinylidene Chloride. Office of Research and Development October 1983. EPA. Office of Health and Environmental Assessment Draft Health Assessment Document for Carbon Tetrachloride. Office of Research and Development August 1983. EPA. Office of Health and Environmental Assessment Draft Health Assessment Document for 1.1.1-Trichloroethane, Office of Research and Development November 1983. EPA. Office of Health and Environmental Assessment Draft Health Assessment for Tetrachloroathyiene. Office of Research and Development December 1983. EPA. Office of Health and Environmental Assessment The Carcinogen Assessment Croup's Evaluation of the Carcinogenicity of Benzene (DRAFT), Office of Research and Development March 1983. EPA. Office of Health and Environmental Aaaeasmant Draft Health Aaaasament Document for Trichloroethylene. Office of Research and Development December 1983. EPA. "Review of a Carcinogenicity Study on Viflyl Chloride." Memo from Robert McCaughy. Office of Research and Development to Joseph A. Cotnivo, Office of Drinking Water. Jan. fl. 1984. National Academy of Sciences. "Drinking Water and Health." Volume I (1977). (1980). IV (1981). V (1983). IARC 1979. IARC Monographs on the evaluation of the carcinogenic risk of chemicals to humans: soma balogenated hydrocarbon*. Vol.2(fc 14^8. IARC, Approaches to Classifying Chemical Carcinogen* According to Mechanism of Action. Technical Report No. 83/001. April 1983. NCL "Policy of Risk Assessment of the Health Effscts of Hazardous Exposures to Populations," Subcommittee on Environmental Carcinogens, National Cancer Advisory Board. 1983. Bruckner, James. Progress Report Coop. Agr. 807449-02. pp. 18-22. Univ. of Texaa Medical Center at Houston. July 11,1983. Oral Toxicity of Carbon Tetrachloride in Rats. Manuscript in preparation. VUL Request foe Comments EPA requests public analyses, comments and information on all aspects of this proposal The questions for which comment is being specifically solicited ere listed below. Comment will be of greet assistance to EPA in formulating protective and practical approach to reducing human exposure to VOCs in drinking water. How strong should the scientific evidence be to Justify regulating a substance, particularly for carcinogenicity? --When positive evidence exists but is sparse or inconclusive, how should it affect decision-making? Should there be e well-defined and uniform minimum level of evidence of carcinogenicity in animals or humans7 If so, what evidence would comprise this minimal level? --When substantial doubt exists ss to whether e substance causes a serious health or environmental risk, bow should EPA balance its mandate to bit on the side of protection against the competing risk of imposing costly regulations on substances which mey later be shown to bs benign? How should evidence of mouse liver tumors be weighed? If evidence is limited to mouse liver tumors, is that sufficient evidence to warrant regulating that substance as a carcinogen? Conversely* what would be the scientific basis for giving mouse liver tumors less weight in the evaluation of the potential for human carcinogenicity? What level should be set for RMCLs that would represent e level such that "no known or anticipated adverse effect would result with an adequate margin of safety"? --For non-carcinogens, is the approach used for computing the AADIs scientifically acceptable? Is providing for an assumed contribution of 20 percent from drinking water appropriate when more predse data is not available. --Should RMCLs for carcinogens be set at zero? If RMCLs are set at zero, wbdt guidance, if any. should be provided on the actually attainable target levels in drinking water? --Should RMCLs for carcinogens be set at the analytical detection limit? What : would this be for each VOC considered in this proposal? --Should setting RMCLs for carcinogens be established at a non-zero level based upon a negligible risk determination? What non-zero level and upon what basis? Which model and which assumptions? Does an incremental lifetime risk level of 10'* represent a virtually non existent or negligible risk? Should higher or lower risk rates be considered? Would another level be more representative yet meet the needs for practical implementation of the SDWA? Would use of the linearized multi-stage model in the non-zero RMCL calculations meet the Congressional intent to incorporate a margin of safety into the RMCLs? --Should a range of finite risk levels for each RMCL be selected such as 10s to 10"* instead of zero or a single value? How should the degree of evidence of potential carcinogenicity be factored into the RMCL determinations? If there is sufficient experimental evidence of human carcinogenicity, should the RMCL be either zero or the one in one million risk equivalent or some other calculated value? Should the RMCL be set at higher concentration and higher nominal risk (to indirectly reflect less concern) as the strength of evidence of carcinogenicity is reduced? For example, if there is only sufficient evidence of animal carcinogenicity, should the RMCL be in the 10"* up to the 10* range, whereas if there is only limited evidence of animal carcinogenicity, should the RMCL be in the 10"* to 10"* risk range? If less than "limited evidence" ia available, should the RMCL be determined based upon an ADI calculation? As another example, could RMCLs for substances such as TCE end PCE with limited, insufficient, or less convincing evidence of carcinogenicity be produced on the basis of chronic toxicity, but with an additional margin of safety or based upon the minimum measured cancer producing dose level such as was suggested by Weil (Toxicology end Applied Pharmacology21454-163 (1972)]? This would differentiate those from substances such as benzene or vinyl choloride which have the moat complete evidence and therefore warrant the most conservative regulatory treatment * Should an RMCL and an MCL be set for total VOCs to address multiple exposure to VOCs? On what basis? n r* * v. i Subject t: Protective Order In" Ross v.. Cor: to,. Ins.., Uo. 90-4833 I4th Judicial District' Court 4,--Calcasieu Parish,,, Louisiana' CMA o m o io 1 24352_________ Federal Register / Vol. 49, No. 114 / Tuesday, June 12, 1964 / Proposed Rules A public hearing will be held in Washington, D.C., for the interested public to comment and provide information and data on the regulatory approach. - EPA recognizes that many significant questions surround the issue of the control of volatile synthetic organic chemicals in drinking water. The Agency has attempted in this proposal to portray current scientific uncertainties in a measured and objective manner. In this way, any data gaps or errors in logic which may exist can be identified and corrected. For that reason, careful review of and thoughtful comment on the information in this proposal is encouraged. Under the Regulatory Flexibility Act, S U.S.C. 601 et seq., I certify that this action will not have a significant impact on a substantial number of small entities. This proposed action will have no economic impact in and of itself because these are non-enforceable health goals. Under Executive Order 12291, EPA must judge whether a regulation is "major" and therefore subject to die requirements of a Regulatory Impact Analysis. This proposed action does not constitute a "major" regulatory because it will not have a major financial or adverse impact on the community and it is a non-enforceable action. This regulation was submitted to the Office of Management and Budget for review as required by Executive Order 12291. List of Subjects In 40 CFR Pari 141 Chemicals, Water supply. 42 U.S.C. 300/SDWA1412. Dated: |une 1,1884. WUlta D. Ruckelshsus, Administrator. For the reasons set out in the preamble, Part 141 of Chapter I of Tide 40 of the Code of Federal Regulations is proposed to be amended as follows: PART 141--NATIONAL PRIMARY DRINKING WATER REGULATIONS L The title of Pari 141 is revised to read as set forth above. 2. In 1141.2, paragraph (u) is added to read as follows: (141.2 IAmended] 4 4 4 *0 (u) "Recommended maximum " contaminant levels" means the maximum level of a contaminant in drinking water at which no known or anticipated advene effect on the health of persons would occur and which includes an adequate margin of safety. 3. A new Subpart F, consisting of $ $ 141.50 and 141.51. la added as follows: Subpart F--Recommended Maximum Contaminant Levels 1141JO Recommended maximum contaminant levels for organic chemicals. The following are Recommended Maximum Contaminant Levels for organic chemicals. They are nonenforceable goals for public water systems. (e) Recommended Maximum Contaminant Levels era zero for the following substances: trichloroethylene, tetrachloroethylene, carbon tetrachloride, lj-dichloroethane, vinyl chloride, 1,1-dichloroethylene, and benzene. (b) Recommended Maximum Contaminant Levels for the following substances are as indicated: 1,1.1'trichloratthan*. p-dichlomtanmni (M-dtoiorabsum*).. krnfrom parin&r 01 071 1141J1 (Reserved) Appendix A--Summary of Public Comments Pertinent to the Propoaed Recommended Maximum Contaminant Levels (RMCLs) for Volatile Synthetic Organic Chemicals (VOCs) in Drinking Water The following is a summary end discussion of the principal public comments to EPA's proposed rule for the establishment of RMCLs for certain VOCs in drinking water. EPA specifically solicited comments on the following three issues in its March 4, 1982, Advance Notice of Proposed Rulemaking: 1. What is the significance of contamination of drinking water by VOCs? 2. What approach should EPA take to deal with VOCs in drinking water? 3. What level should be set for RMCLs such that "no known or anticipated adverse effect" will result? How should the health basis be determined for any MCLs? EPA received 138 written comments during the 210-day public comment period and five oral etatements were presented at the public meeting held in Washington, D.C., on April 28,1982. The '. comments included 28 public interest groups. 14 water utilities, 18 chemical manufacturing companies, 11 state governments and state organizations, 12 focal governments, 40 private citizens and 18 from otherjroups including some members of Congress. The following discussion summarizes comments received on the ANPRM for VOCs. 1. What is the aignificance of contamination of drinking water by VOCs? A total of 88 commenters addressed this issue. A majority of comments (41) feh that VOC contamination in drinking . water is a significant national problem because of the frequency of occurrence and potential health risk warranting action to limit exposure to VOCs. Their reasoning is based on the following: Local problems of sever# VOC contamination, the number of VOCa in drinking water is continually increasing and VOC* have been demonstrated to cause serious carcinogenic and noncarcinogenic toxic effects. Some of the toxic effects are as follows: Some VOCs `an known animal carcinogens and vinyl chloride is both an animal and human carcinogen, causes hepatomas in animals and in some cases in humans, is toxic to the kidneys, has serious effects on the reproductive system, and depresses the central nervous system. One commenter stated that the results of a monitoring study conducted by New Jersey showed 17 percent of the 1.200 wells tested contained VOCs at concentration! above 10 ppb. They felt that "file toxic properties of these chemicals. Including the potential increased risks of cancer and birth defects, warrant federal action". Twenty-five commenters felt that neither the occurrence data, the health effects data, nor the combined data, demonstrate on a national basis the significance of VOC contamination in drinking water: therefore action to limit human exposure to VOCs is not warranted. Reasons cited were: VOC contamination in drinking water ia a localized problem, not a widespread national problem: more information ia needed on occurrence and health effects, especially In order to assess the significance of VOC contamination: state data represented emergency spill situations which an not considered to be statistically representative of national occurrence; when present VOCs usually occur et low part per billion concentrations, whereby significant health risks would not be expected: end the results of the Ground Water Supply 8urvey (GWSS) should be considered questionable because the detection limits that were used (i.e, 02 |Lg/l) are extremely sensitive and can rarely be reached. One comment stated that "positive occurrence data does not present a case for regulation". CONFIDENTIAL o tV ^ q O Subject to ?n 'SC v. Cor.cco. 24th Juii. l ir Calcasi^j ^ 2rder In L -<o. 90-4837 "ict Court Louisiana ' SuBJect to Protective Or'" Ross v. Conoco, Inc., P-' 14tli Judicial Distr: Cl^osieu Parish, T ^ Federal Register / Vol. 49, No. 114 / Tuesday, June 12. 1984 / Proposed Rules 24353 In general, these commenters felt that the major shortcomings of the available health effects data is that it is not scientifically established at this time and subject to debate among the scientific community. One eommenter specifically stated that "until the extent of the threat of human health of VOCs, if any, can be established, federal regulations governing VOCs are not justifiable on mere occurrence data alone". 2. What approach should EPA take to deal with VOCs in drinking water? A total of 118 comments were received that addressed this issue. Comments favored one of the three approaches provided in the ANPRM which are: * Non-federal regulatory approach * Establish montioring requirements and provide Health Advisories (formerly termed SNARLs) for State response as appropriate * Establish national regulations for monitoring and MCLs The non-federal regulatory approach was favored by 22 commenters. Several commenters stated that recent surveys by EPA and their own sampling did not indicate a major VOC contamination problem within certain States, thus, MCLs and monitoring requirements are not warranted. A number of commenters felt that the health effects data were insufficient to show a health risk that would warrant regulation. One eommenter stated that "EPA has not presented any evidence that there is any risk to the population requiring Federal regulation. Furthermore, where the contamination is the result of improper disposal of solvents, guidance is needed by those implementing the RCRA/Superfund cleanup as to an adequate effort". One eommenter summed up this sentiment as "there is a need for a rational!e] and consistent approach to the problem of low levels of carcinogens in drinking water, but the science is not sufficiently developed to guide regulatory and utility actions with any degree of certainty". Some commenters are opposed to national regulations for VOCs because, "shrinking federal and state resources are creating problems for die already existing drinking water programs". They felt that "EPA should focus on source protection and rapid reaction to ground water contamination than attempt to cover all possibilities by regulation". The majority of these commenters favored the continued use of Health Advisories to handle contaminant situations. However, two comments were received that specifically . requested that "once a health advisory is released, it should be published in the Federal Register, detailing EPA's derivation of the health advisory". "Based on the scientific input and 1 *m occurrence information received, EPA must issue an updated health advisory and can then decide if there is a need to establish MCLs." One eommenter stated that "Health Advisories should be the 1 ................ >100 >*00 >500 >500 >ao 10-100 jft jpg 90-400 9&400 29-340 <10 <50 <50 <50 <28 first step in determining whether or not it is necessary to establish an MCL for a particular contaminant". In addition, one eommenter favored the use of Health Advisories as opposed to national monitoring requirements, in that the latter would only gather mory occurrence data. The third option which would require In general, these commenters favored establishing national regulations for EPA continuing to provide research data monitoring and MCLs. was favored by and technical advise (i.e,, Health 82 commenters. Numerous commenters Advisories) when dealing with stated that MCLs and monitoring contaminant situations. In addition, requirements should be set for the "routine, repetitive monitoring VOCs. A number of these commenters requirements must not be put into qualified their statements by saying regulations because monitoring MCLs should be set if it is shown that programs must be flexible and can best the occurrence of VOCs is widespread be developed by States and water . and the health effects data show that utilities". VOCs are a health risk. However, most Option 2, whereby EPA should commenters felt that sufficient data establish monitoring requirements and were available showing VOCs to be a provide Health Advisories for State widespread problem, that data did show response as appropriate was favored by a potential health hazard, and that 13 commenters. The basis for the MCLs were needed. comments which recommended this approach was two-fold: (1) Localized VOC contaminant situations, especially in ground water necessitates monitoring requirements, and (2) the health effects data is unclear and insufficient to establish MCLs since "safe" levels of VOCs cannot be determined at this time. Health Advisories should be used in dealing with contaminant situations. Generally, these commenters felt that gound water contamination is a problem in some places, which must be addressed; however. VOC contamination is not widespread enough to require highly formalized and restrictive requirements. Furthermore, the available data are insufficient to determine the scope of the problem and only monitoring should be done to determine where problems exist. Therefore, instead of setting MCLs, guidance should be provided. These commenters generally supported giving States considerable authority for implementation of the monitoring requirements and for determining appropriate action when contamination is found. A few of the comments received, which favored tire monitoring requirements option, proposed an action-oriented approach in the form of contamination levels and action The following related statements were made: Problems with organic chemicals have been shown in several States and without enforceable standards; the problems will continue to spread; the justification for cleaning contaminated aquifers will be challenged on a case- by-case basis. Consistent nation-wide standards are needed for VOCs (numerous commenters explicitly stated that "uniform, mandatory and enforceable standards" are needed] to provide adequate public health protection in each State. Latitude for stricter rules by the States was suggested by one eommenter. While Health Advisories were noted to have been very useful in addressing incidences of VOC contamination, several public water systems commented that the States have adopted widely varying approaches to dealing with VOCs in drinking water. Some States have adopted the Health Advisories as enforceable standards and consequently public water systems have been forced to make permanent and costly decisions on the basis of health guidance. Alternative involving determination of the acceptable levels of contaminants categories. They felt that guidance for by individual States, based on EPA five of the VOCs should be established^ T -edyieorvoptions. will not be effective, as follows: --- ^ - ^TheirroisoJng is that EPA iidvisory --7- in '10. 90-4837 CCaaVlc. asieu Parish, LouisCiaonuart CMA 007921 24354 Federal Register / Vol. 49, No. 114 / Tuesday, June 12. 1984 / Proposed Rules opinions on health effects will be misinterpreted and misused as demonstrated by many States that have used the Health'Advisories as rigid criteria by which suitability of a water supply is measured. The fact that Health Advisories are developed without consideration for possible carcinogenic properties of a compound and with disregard for econbtnic and technological feasibility of achieving them is forgotten. We are better off with MCLa established in the process involving public participation and intended to be the rigid standards by which quality of drinking water is measured. If left to the States, drinking water guidelines will differ from each other. "This will lead to confusion and poor public image for State agencies which recommend guidelines less stringent than others. In addition, leaving the regulatory activity to States will require more human resources in the area of toxicological evaluation and standard setting. Many States do not have these resources. Thirdly, enforcement will be very difficult, if not impossible, if neighboring States have different drinking water standards. Contamination has no boundaries." One commenter summarized their argument for MCLa in the following manner "Contamination of water supplies in the U.S. with VOCs is indicative of a national trend. Setting legally enforceable national standards will be important in reversing that trend. It will establish a ceiling on how much contamination of drinking water is acceptable and will trigger remedial action in situations where the ceiling is almost reached or exceeded." A number of comments were received that addressed monitoring requirements and treatment costs; however, these comments will be addressed when the proposed rule for the establishment for MCLa is published. 3. How should the potential health risks of exposure to VOCs be assessed? What level should be set for RMCLs such that "no known or anticipated adverse effects" will result? In assessing the potential health risks of exposure to VOCs. numerous questions arise such as: (1) Whether or not a compound can be classified as "genotoxic" or "nongenotoxic"? (2) Should different risk models or approaches be used for carcinogens that are not genotoxic? (3) Would the risk of exposure to two or more VOCs be considered additive, antagonistic, synergistic? (4) Which subgroup of the general population should be addressed? i'--'co::?i3ir?iAL Jub^'ecT tc ?OS5 " (S) How should exposure to VOCs independent peer review before further from other routes (i.e., air or food) be government end industry resources are addressed? spent on discussing approaches to Consideration of the potential health regulate chemicals which msy be non- effects of a chemical encompasses the hazardous or pose an insignificant risk. evaluation of available data and the Three comments were received that potential for human health effects from recommended EPA should continue to exposure via drinking water. A number use the CAG model for both genotoxic of comments addressed the and nongenotoxic carcinogens. Even aforementioned issues. though knowledge of the carcinogenic First the issue as to what criteria mechanism should be a major factor'in could be used to classify a compound as selecting the most appropriate risk "genotoxic" or "non-genotoxic" was model this information is generally not addressed by seven comments. Four of available for environmental the commenters suggested that the carcinogens. One commenter stated that criteria to classify a compound as a distinction between caracinogenic "genotoxic" include: 'mechanisms is arbitrary because there 1. a reliable, positive demonstration of is s lack of experimental data genotoxicity in appropriate prokaryotic establishing a threshold for non- and eukaryotic systems in vitro; 2. positive results in studies on binding to DNA: and 3. evidence of biochemical or biologic consequences of DNA damage. One commenter submitted data and information on recent development# in the understanding of the various mechanisms by which a carcinogenic response can be produced in laboratory animals. These comments can be summarized as follows: mutagenic carcinogens or showing that the dose-response curve is different in the lower range from that for substances that cause gene mutations. In addition, thresholds observed in experiments with an inbred animal populations cannot be estrapolated with any degree of certainty to a diverse human population: therefore, no diatinction between carcinogenic mechanisms should be made at this time. One commenter added that the multi Based on the extent of'e chemical's interaction with DNA it appears that chemicals that have a greater propenity to directly interact with DNA are appropriately classified as genotoxic. Those that do not have this propensity to Interact directly with DNA. but lead to tumors via recurrent tissue injury or other secondary events are classified as non-genotoxic or epigenetic carcinogens. The carcinogenic risk to man stage model as modified by the Carcinogen Assessment Group (CAG) should be used in establishing MCLs for carcinogens regardless of mechanisms of action. Mathematical models at best provide erode estimates of the risks resulting from exposure to a carcinogen. The third issue as to how the risk of exposure to two or more VOCs should posed by such epigenetic carcinogens appears to be substantially less than that posed by purely genetic carcinogens. Whereas, there has been relatively lees disagreement over appropriate measures for the control of those materials categorized as human carcinogens, there has been considerable disagreement emTMg scientists be considered was addressed by five comments. Four commenters felt that two or more chemicals found to be toxic to the seme organ system should be considered to be additive in their cumulative effect on the body. Added margins of wriety should then be regarding appropriate measures for the included in the health basis of each control of the numbers materials categorized as animal carcinogans on the basis of tests in rats, mice or hamsters. MCL The magnitude of the safety factor should reflect where possible, current understanding of synergistic interaction The above commenter* stated that between chemicals and should be different risk models should be used to considered at least additive in account for the differentiation of proportion to the absolute and relative carcinogens recognizing different ' feveie of exposure. In addition, no mechanisms. All of these comments evidence he> been put forth that reject the CAG risk model because it it suggest! that these interactions could too conservative end that both the upper never be e problem. One recommended and lower bound risks must be taken approach wee to set an RMCL and MCL into account. In other words. EPA's for "total VOCs". cancer risk estimation process One commenter believed that the risk overstates the potential risk posed by of exposure to two or more VOCs is not these chemicals in e manner which may additive. The reasoning was based on mislead the public. Furthermore, they believe that EPA has accepted animal two studies in which two chlorinated solvents were administered data at face value without any critical simultaneously for 3 to 8 months, in review. Tbey fwommepcta4$h*X9nT I Amch no synergism was indicated and. health crijjaria documums us iuEjeLl lu. . _ IB ftuEfcihe affects were less than Subject to Protea^TM^^ ROSS V.: rnnooo. EUtH Judicial Inc.. Ho. District 90 Court (Cv.ialoasieu Parish,, Louisiana CHA 007V rt?. Federal Register / Vol. 49. No, 114 / Tuesday, June 12, 1984 / Proposed Rules 24355 additive. Thus, the commenter felt that an increase in the margin of safety is not required. The issue of which subgroup of the general population should be protected received six comments. Three commenters felt that the 10 kg child should be used because adequate protection should be provided to all segments of the general population. Two commenters felt that the 70 kg adult should be the basis for potential MCLs because life-time exposure should be used in the calculation. A10 kg child is not exposed over a 70-year lifetime. One commenter recommended that if MCLs are warranted, the level should be set to protect all signflcant populations groups (i.e.. children, pregnant women, aged adults, etc.). Also, short term exposure risk calculations should be based on a 10 kg child, long term exposures based on a 70 kg adult and the worst case would be controlling. Lastly, the issue as to how exposure to VOCs from other routes should be addressed in the development of RMCLs received eight comments. Five commenters felt that relative source contribution should not be a major factor in determining the acceptable risk. One commenter suggested that EPA state the likely other sources of VOC exposure and average levels. Another commenter put it this way, "The contribution of drinking water to the total exposure to a contaminant should be considered in light of the risk to public health and not in terms of its relative significance to other sources of exposure". Three comments recommended that the total allowable body burden from all media (air, food and water) should be taken into account, based on health effects data. Twenty-six comments were received on what level should be set for RMCLs such that "no known or anticipated adverse effects" will result. Twenty-two commenters recommended that the RMCLs for carcinogens be set at zero. Their reasoning was based on the premise that an RMCL is a health goal, which is not intended to reflect cost and feasibility of treatment and that scientific evfdence to date cannot be used to establish a no adverse health effect threshold for carcinogens. Four commenters recommended that the RMCLs for carcinogens be set at a finite risk level and not zero. Their reasoning was that every water supplfl contains at least some of the chemicaSi listed in the ANPRM. A finite risk level is the only realistic basis. Furthermore, it is impossible to establish with any degree of certainty that the concentration of a contaminant in water is zero, due to limited analytical capability. One commenter stated that. "Tile question of the level of the RMCL for carcinogens is the most fundamental in the ANPRM. RMCLs are confusing and an RMCL set at zero is not useful because it could not be measured." Instead a regulatory target level (RTL), set as a negligible risk level should be established. The level should be HT*. based upon the National Academy of Sciences projections, not CAG's." Another commenter felt that "RMCLs for compounds shown to increase tumors in test animals through nongenetic mechanisms, should be set at a finite number based on the toxicity of the contaminant (i.e., incorporating the threshold concept)". (Fit Doc. FlUd *s Ul| MUJM COM MM-tt-M CMA 007923