Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
55 ASBESTOS William M. Stigliani Part 1: Environmental Protection Agency' s Assessment of Asbestos-Containing Material in Schools A. BACKGROUND AND CONTEXT - t. Describe the chemical and its uses. Asbestos is a general term for a group of naturally occurring hydrated mineral silicates that separate into fibers. Asbestos minerals used commercially include: chrysotite, amosite9 crocidolite, tremolite, actinolite, and anthophyllite asbestos. Since asbestos in highly resistant to heat, has high tensile strength, and moderate to good chemical resistance, it has many uses ~ These include ~sbestos-cement pipe, asbestos paper, friction products, vinyl asbestos floor tile, paints, coatings and sealants, and gaskets and packings. N : This case study describes assessment procedures and issues and interpretations raised by others, but it is not intended to present independent positions or interpretations on either scientific or policy matters. Me case has been reviewed by individuals outside the study pro Sect who are directly familiar with the federal analyses and decisions described; however, responsibility for the paper rests with the au£J,or, and it does not necessarily reflect the judgment of the Committee on the Institutional Means for Assessment of Risks to Public }search or the National Researcl' Council. It has not been subjected to internal review procedures that apply to reports prepared by NRC commit tees .
56 2. Describe tow the question of risk was elevated to the agency agenda. Warnings about 'near th hazards in schools 'nad been raised from several sources during the 1970s. Dr. William Nicholson and a team of Mt. Sinai scientists in ~ 1978 study did measurements in schools in New Jersey showing high levels of asbestos exposure. At the same t ime 9 the Pub kc Health Service had issued an advisory concerning the hazard of asbes tos in schools . Subsequent to these warnings, the Environmental De fens e Fund (EDF) petitioned the EPA, through section 2L of the Toxic Substances Control Act, to take regulatory action. The petition was denied, and El)F sued the EPA. In an agreement settled out of court, EPA agreed to proceed trite rulemak' ng. The rule became final June 28, 1982. 3. Under what statutes and agency jurisdiction does the chemical fall? What statutory tests governed the decision? The rule falls under the jurisdiction of section 6( a) of the Toxic Substances Control Act 9 which authorizes the Administrator to issue warnings and notification if a hazard exists. The rule requires the governing officials (e. A., superinten- dent, head of school board, headmaster) to inspect ~11 public and private primary and secondary schools in the DO SoA. for friable materials. If such material is found, three samples oust be analyzed for asbestos by polarized light microscopy from an EPA recommended laboratory. If abbes tos is found to be present in these samples, a school must: ( 1) notify the PTA, ~ 2) notify all employees, (3) post notices in administrative areas, (~) give guidance for reducing ashes tos exposure to maintenance people and (5) keep records of all correspondence ~ laboratory information letters to employees, PTA, etcO ~ . 4. What was the decision schedule? Note any statutory or other action deadlines. . . Advance not ice 0 f proposed rulemaking First draft support document assessing risk of asbestos in s choot s Final Rule September 1980 Oct. 1980 Feb. 1981 ~ postponed)
57 Second draf t support document assessing risk of ashes tos in schools Final vers ion of support document assessing risk of a~be~to.s in schoo Is Fina ~ Rule July 1981 January 1982 June 28, 1982 Final date of compliance to rule June 28, 1983 QUANTIFICATION AND CHARACTERIZATION OF RISK 'to HUMANS _ (Sections B and C have been combined. )* I. Cat health endpoints were evaluated? Lung cancer, pleural and peritoneal mesothelioma, cancers of the larynx, oral cavity, esophagus, stomach, colon, and kidney. 2. What were the key data available for review? (What additional data were sought?) Human Data The epidemiological data selec ted to be the basis for making quan- titative estimates of premature death from exposure to asbestos in schools -~as a large study of asbestos insulation workers (12,0S1 men) reported by Hammond et al. ( 1~ and Selikoff ~2~. Various other epidemiological s tudies were considered but EPA decided the insulation workers study was the best one available. Several reasons were c i ted for this pre ference: l) the large sample size; (2) the reasonableness of the estimates of asbestos exposure levels; (3 ~ detailed information on various cancer types; * This section describes the second draf t support document (July 1981~ . The final version of the support documen t (January 1982 ) is not discussed here since it contains no quant' tative estimations of risk; the estimates calculated in the second draft were deleted.
58 (4) verification of death certificates with supple- mental information (e.g., autopsy reports, histo logica ~ s pee imens ~ to de tee t misdiagnosed mesa the 1 imas; (5 ~ appropriateness of the control group; and 6) similarities between the material to which the insulation workers were exposed and the ashes tots present in schools. Data on Exposed School Population EPA gathered information on the presence o f friable asbes tos- containing materials in public schools and the number of people exposed by conducting a survey of the nation's school districts. There was about an 8% response rate. EPA subsequently contacted school districts that did not respond initial ly for further informal ion. Exposure Assessment The prevalent exposure levels in schools containing friable ashes tos materials were estimated based on data from a study by Sebastien et _. (3) of several buildings in Paris. EPA's reasons for choosing this study were the following: t) the areas and materials studied are similar to those in U. S . schools; (2) the measurements were made by transmission electron microscopy ~ the only technique whic h is accurate for environmental sampl ing a ~ low concentrat ion); the measurements were checked by statistical quality control techniques 9 and the samples were taken over relatively long time periods; and (3 ~ comparisons were made with outdoor air and with a significant number of buildings that did not contain ashes tos materials . EPA fen other studies of U.S. schools did not reeet thee. criteria. Ilowever, other studies were used to verify that the results of the Sebastien _ al. study were consistent with data for U.S. buildings .
59 3. To performed the initial analysis? (What was their background? Available analytic resources? ) Two s taf f members of the Health and Environmental Review Division of the Office of Toxic Substances at the EPA performed the initial analysis. One was a Ph.~. epidemiologist; the other had an M.P.H. degree. 4. To what extent were results presented quantitatively? What factors influenced the degree of quantif ication^ lathe results were presented in a precisely quantitative fashion--as lifetime risk and number of premature deaths for students and adul t empl oyees . The EPA staff felt there was sufficient data, with reasonable assumptions, to proceed with a highly quantified assessment. How was uncertainty described in reaching final interprera- tions? Were crucial assumptions made explicit? Uncertainty was described in two important ways . First, utilizing a linearized dose-response curve, EPA calculated a range of risks for school occupants, characterizing this range as low, inter- mediate and high. Me range was based on high, low, and most likely predict ions of three parameters: ( 1) The cumulative exposure of insulation workers in the underlying s tudy; (2 ~ The cumulative exposure of occupants of schools; and (3) Mortality rates among insulation workers (based on observed deaths, and deaths calculated from uppper and lower 95% confidence limits). Second, EPA considered the pass ibility that risk could be described by other dose response models, including the threshold model. The document s tates: EPA's policy is to select curves that cannot 'ce ruled out on the teas is of pharmacokinetics or poor "fit" to available dose-response data and that display the full range of reasonably possible increases in risk....EPA is unaware of information about the pharmacokinetics of ashes tos that would enable the shape of a dose-response curve to be
60 inferred O . ~ .~ should be noted, however, that a curve could be deve loped to yield virtual ly any prediction of risk between zero and the leve 1 of risk predicted by the one-hit model or linear regress ion. All crucial assumptions in describing these uncertainties, were explicitly stated. 6. How were Qualitative factors dealt with Mechanism of action, associated thresholds E ffec ts on population subgroups - Other confounding factors EPA was unaware of information about the pharmacokineeics of asbestos that would provide definitive evidence about the shape of the dose-response curve ~ or the existence of thresholds O EPA did consider biological susceptibility to asbestos as a function of age O After reviewing the literatures EPA determined that there was little confirmatory evidence to assume that children were more susceptible to asbestos exposure than adults. EPA made the assumption that annual incidence rate is not affected by age at first exposure. The longer retraining life expectancy of children compared wi th that o f adults was the only factor that was inco rpora ted in t o the q cant i ~ a ~ ive ri s ~ e s t ima te . EPA did consider the greater risk of lung cancer from asbestos exposure among smokers Data for the smoking-asbestos interaction was incorporated into the risk assessment. The effect of fiber size and type on carcinogenic response was considered. Experimental evidence s trongly sugges ts that fibers of certain sizes that reach the pleura g regardless of chemical composition, are more potent in producing mesothelioma than fibers of other sizes. EPA assumed that its use of data from the study of asbestos insulation workers avoided any major uncertainties that might otherwise have been presented by this finding. Because there were no data indicating that the fiber types or sizes to which the insulation workers were exposed were substantially different from those present in schools, the types and sizes in both settings were assumed to be similar.
61 What qualitative factors affected the weighting of data? Were such criteria explicit and in accord with any general ouide- lines ? - The calculated risk estimates were based on a linear, nonthreshold dose-response curve. SPA acknowledged that these estimates were conservative with respect to other models that could have been used, particularly the threshold model. EPA stated that "the existence of a threshold is theoretically possible but not demonstrated." EPA cited evidence, primarily in the form of mesot'nelioma case reports, that nonoccupational levels of exposure, perhaps as low as those found in schools, are sufficient to elevate risk. Other reasons cited by EPA for use of the linear mode! were: the fact that the authors of two major studies of asbestos workers believe their data for increased respiratory cancer risk are best described by the linear nonthreshold relationship, the fact that reviewers of the asbestos literature have recommended the use of this dose-response curve for lo~dose extrapolations, and that for quantitative risk assessments of carcinogens in general, the EPA Interim Guidelines (EPA 1976, Albert _ al . 1977 ~ Cal is for the use of the 1 inear nonthreshold dose-re s pans e curve . As another issue EPA stated that peak exposures were not included in the risk assessment because their frequency could not be estimated. However, "for some individuals, custodians and maintenance workers in particular, the impac ~ of peak exposures might dwarf the effect of exposure to prevalent asbestos concen- trations....Custodians could easily double their cumulative exposure by spending less than 3 minutes per day dry sweep- ing....Vies~ed in this manner, the inability to incorporate peak exposures into the risk assessment may underestimate custodians' exposure (and therefore risk) by more than an order of magnitude." EPA estimated that 6% of the exposed school children were smokers, and sub ject to the interaction between smoking and asbestos in the elevation of lung cancer risk. This estimate assumes that interaction wit 1 not take place in exposed students who become smokers after leaving school. If the interaction actually does occur, then the risk to school children predicted by EPA could be s ignif~cantly underestimated. Another assumption was that the estimate of premature death when only 16: of the workers have died (observation period of the cohort study) will be the same when all 12,051 have died. This assumption may underestimate the risk. In calculating the cumulative exposure of the insulation workers, EPA assumed an induction period 10. Therefore, only exposures 10 years prior to the beginning of the observation period were considered '~wasted. " There is convincing evidence that the induction period may be longer, perhaps 20. Thus E P.\ may have overes timated the cumulative exposure, thereby unter- estimating the risk at a given cumulative exposure.
62 8. Describe any internal, internal-ad~risory (e.g., EPA' s SAB), and external (e.g., NAS) scientific review of the initial analysisO What, if any, cri ticism was incur ed? In early 1980, a draft risk assessment was reviewed by four outside medical authorities. In the proposed rule ~ September 1980), public comments regarding health risks were solicited. In November 1980, there was a public rulemaking bearing in which comments were invited. In December 1980, there was a meeting with the Toxic Substances Subcommittee of the EPA's Science Advisory Board ~ SAB) to evaluate the risk assessment document . Me SAB raised the issue of the lack of definitive scientific evidence for choosing one dose-response curare in preference to another. Whey suggested that the risk assessment incorporate several different extrapolations. They also urged the authors to incorporate more evidence from qualitative epidemiologic data showing the incidence of mesoehelioma occurring at extremely low levels of asbestos exposure. This data they suggested was the best evidence that a threshold did not exist. The SAB also suggested that more emphasis be given to peak exposures, which they believed caused an inordinately high risk to school maintenance workers. Finally, they strongly urged the El?A to incorporate the separate ef fects of ashes tos exposure on smokers and nonsmokers into the risk assessment. 9. How were the issues raised in the reviewks) accommodated? In the revised draf ~ (July 1981) EPA included a more extensive discussion of different dose-response models. In particular, EPA focused on the question of a threshold response to asbestos exposure. EPA did not dispute its possible existence, but presented information that would argue against such a contention. In particular, the revised risk assessment included more comprehensive documentation of cases of mesothelioma occurring after only short periods of exposure to asbestos. The revised draft also discussed the importance of peak exposures. Although these were not factored into the risk assessment, EPA acknowledged the possible great underes timation of risk due to this omission ~ see Q. C . 7 ~ . Data on ashes tos exposure among smokers and nonsmokers was made available to EPA, and this information was incorporated in the revised risk assessment. The new data lowered the estimated risk since the incidence of smoking among school children is much lower than the incidence among insulation workers.
63 10. What other issues arose concerning scientific data and their use? Briefly describe dissenting opinion. Representatives from the Asbestos Information Association (AIA) also attended the December 1980 meeting. At that time the AIA argued the t: the exposure data was taken from only one French study which was not representative of all the schools in the U. S.; there was no evidence that a threshold d id no ~ exist, and therefore a scientifically valid risk assessment could not be made until the issue was reso lved; and (3) the risk assessment was seriously flawed because no attempt was made to separate out the effect of smoking in the ana ly s i s . L1. Is the substance subject to past or possible future regulatory i_ ordinate W1 th o ther agencies or programs ? The Air and 'hater Offices at EPA each have established standards for asbestos. Also, OSHA, CPSC and FDA have all promulgated rules, pertaining to various aspects of asbestos use. Asbestos is subject to future regulatory actions or revis ion of current s tandards in each o f the four agencies . Mere was an asbestos working group established by the Interagency Regulatory Liaison Group to coordinate activities under the Carter Administration, but this group was disbanded in September 1981. D . INTE RPRETATION 1. fat role did ri sk assessment have in the final agency document where s tandards were established to control the che~nica 1 ? 'the quantitative risk assessment played no identifiable role in the final rule. In fact the calculations and estimation of numbers at risk were removed from the f inal rifle assessment document.
64 Were there variations--over time or across agency programso-in the assumptions used? Were these variations significant to the final risk assessment? , By the spring of 1982 g EPA, CPSC and OSHA all had developed (draft) quantitative risk assesses The author of this case study compared the values of premature deaths quoted in these draft reports ~ standardized to an accumulated exposure of 2 f~bers/cc, 20 years, 40 hours/week) . Specifically, the calculated risks, in premature deaths per 1 9 000 were: OSHA* EPA CPSC 8-260 . 1,000 57-348 1, 000 19-338 1 9 000 It thus appears that there was significant agreement among the agencies in Else calculation of risk from abbes tos. To the extent there were issues/concerns about questions of science, would the outcome have been improved by: A better system of in-house scientific review? Review by an outside scientific organization? Coherent federal guidel ines on carcinogenic risk assessment Better agency guidelines on the performance of risk assessment? - Improved agency decis ion procedures ? Me calculated numbers for school occupants at risk, present in the July 1981 draft, were deleted in the final version of the risk assessment (January 1982~. Apparently, the decision to remove the numbers was made by the Deputy Ass istant Administrator for Toxic Substances without consultation with the Toxic Sub- ~ Lances Subcommittee o ~ the EPA' s Science Advisory Board ~ SAB) . * OSHA's estimates were quoted with no supporting information provided (FR. Jan. 13, 1982, 1807) .
65 (For the role of the SAB, see Q. C. 8) . There were reports of concern among the SAB members, that the deletion and the bypassing of the advisory panel could reduce the credibili ty of EPA risk assessments. This break in communication may have been averted had es tab L ished procedures on the ro le o f the SAB in individual risk assessments been better formulated. PERFORMANCE CONSIDERATIONS . 1. Ability to obtain relevant scientific information. EPA staff probably utilized the best scientific late available to them at the time the assessment was made. The major gap in information was lack o f good comprehensive exposure data for U. S. schoo 1 s . 2. Credibility of assessment; 1 ikel infold that interested parties would accept them as definitive. The draft risk assessment (July 1981) is credible only to the extent that a nonthreshold linear low dose extrapolation is deemed to be credible. EPA s taff was following EPA policy and guidelines in use since the mid-1970s. Since the choice of a model is based in part on a policy decision and not on complete scientific knowledge, some interested parties, quite predictably, attacked the use of the model. ~~ ~ _~ ~ _- the assessment group itself? That was the degree to which interest pressures could be exerted from outside the assessment group? '~hat was the respons iveness of the assessment to these diverse interests? The assessment group was made up of staff from EPA's Health and Environmental Review Division of the Office of Toxic Substances. The SAB, which played a key role in reviewing the document, was made up of academicians and various medical experts. None of these appeared to have any direct association with the asbestos industry. Early in the review process, representatives from the Asbestos Information Association (AIA), an industry trade association, expressed strong opposition to various aspects of EPA's quantitative risk assessment (see Q. C. 10), and voiced their objections on numerous occasions.
66 After completing of the second draf t of the quantitative risk assessment (July, 1981), the Deputy Assistant Administrator for Toxin Subs tances asked the asses sment group to remove the calculated number e of school occupants at risk from the risk an sessment O The fine ~ vers ion o f the s upport document ~ January 1982) reflected this deletion. T'nis change was done apparently without prior consultation with the SAB, and -~as reported to have had the wholehearted support of the AIA. Some interested members of Congress have harshly criticized the deletion, citing evidence float SPA of ficials consulted only the ALl before making the change. See also QO I). 3 and QO E. 7. 4. What were the time and resources necessary to complete the risk assessment? Three to three and one-half person years was needed to complete the risk asses sment . 5. Responsiveness of assessment agenda to public concerns, interest group concerns, professional concerns, and emergence of nest scientific information. Lois question has been answered in Questions A.2, C.8, C.9, and E.3 . 6. Ability of the risk assessment to identify research needs. An exposure s tudy of ashes tos in Ho us ton schools was undertaken. This was done in part to answer the criticism incurred when EPA used data from a Parisian building as the basis for an exposure assessment. The Houston data, in fact, provided numbers that were quite similar to those obtained in Paris. 7. Extent to which risk assessment impeded or facilitated regulation? Initially EPA was planning to require school districts to take corrective actions to protect occupants of schools from asbestos exposure. Such a rule would have been very expens ire to inact . Under the Toxic Substances Control Act (TSCA), the cost of imple- menting a rule must be balanced by the benefits accrued from it. ~us, it would have been necessary under TSCA, for EPA to demon- strate that the health risk was sufficiently great to merit such action. A quantitative risk assessment probably would have been helpful to illustrate the number of lives that could 'he saved by the actions A1SO9 other actions limiting the use of asbestos were being contemplated by SPA. It was under this percept ion, that SPA proceeded with the quantitative risk assessment, believing that it could be useful for a number o f contemplated regulatory ac tions .
~7 The final asbestos rule dealt only with identification and notification of the presence of asbestos in schools. The rule requiring school districts to take corrective action was abandoned by EPA in April 1981. The Deputy Assistant Administrator (DAA) for Toxic Substances, who called for the deletion of the quantitative risk assessment from the July 1981 draft, cited two reasons for doing so. One was that the cost of the identification and notification rule is "very little", and thus a detailed quantitative risk assessment to support the rule was not needed. Secondly, it was important "to get the rule out. " Controversy over the quantitative risk assessment was holding up progress. The asbestos industry, which was not opposed to the identifica- tion and notification rule per se, was quite opposed to the quantitative risk assessment supporting the rule, and its promulga t ion . Some members of Congress disagreed with the DAA's judgment. For example, the chairman of the House subcommittee on Labor standards felt that without the numbers, the "sense of urgency" in implementing the rule is lost for local officials, parents and schoo 1 emp layees . 8. Were related risk assessments consistent? - See Q. D. 2. 9e Extent to which there Is an explicit distinction between weights accorded to scientific factors and policy factors. The July 1981 draft made very explicit distinctions between scientific and policy factors utilized in the quantitative risk assessment. The distinctions were particularly clear regarding the use of dose-response models. See also, Q. C. 5 and Q. C. 7. 10 Mode and frequency of communication between assessors and . regulators . There appears to have been some problem of communication between the regulator and the assessor. See Q. D. 3 and Q. E. 3.
~8 REFERENCES . Albert, Roll. ~ R. Train, and E. Anderson. 1977. Rationale developed by EPA for the assessment of carcinogenic risks. J. National Cancer Into 9 5801537-154~. EPA (Environmental Protection Agency). 1976. Health risk and economic impac t assessments of suspected carcinogens: interim procedures and guidelines. Fed. Reg. 41: 21402-21405. Hammond, EoC. ~ I.J. Selikoff, and H. Seidman. exposure, cigarette smoking and death rates. 330 :473~490. 1979. Asbestos Ann. MY Acad. Scio Selikoff, I.Jo, EoC. Hammond, H. Seidman. 1979. Mortality experience of insulation workers in the United States and Canada, 1943-1976. Ann. NY Acad. Sci 3300 91-1160 Sebastien, P., A. Gaudichet 9 G. Dufour, G. Bonnaud, J . Bignon, J. Tori. 19780 Metrological survey of atmospheric pollution inside buildings insulated by asbestos projection. Paris, France: Paris University Laboratory De par~cmen t ~ EPA ~ rans let ion) .