A REComma‘gfi‘fieq \ “>9 ”UBLJC Lin; 3'? UBRARN (Hu;r—',;: . I; LIBRIJY g ; ”NINJA" 1p , \ CALI-”73"”; .\ > L CRITERIA FOR A RECOMMENDED STANDARD.... OCCUPATIONAL EXPOSURE TO RADON PROGENY IN UNDERGROUND NINESJ U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES PUBLIC HEALTH SERVICE CENTERS FOR DISEASE CONTROL NATIONAL INSTITUTE FOR OCCUPATIONAL SAFETY AND HEALTH DIVISION OF STANDARDS DEVELOPMENT AND TECHNOLOGY TRANSFER October 1987 (SH/083m FUEL DISCLAIMER Mention of the name of any company or product does not constitute endorsement by the National Institute for Occupational Safety and Health. DHHS (NIOSH) Publication No. 88-101 I I For sale by the Superintendent of Documenu. U.S. Government Printing Office, Washington, D.C. 20402 ) (\A/2Ll7 FOREWORD F173 3 <5" 3 3/ l9 3"7 [Qt/«8L As Director of the National Institute for Occupational Safety and Health (NIOSH), I am accustomed to making decisions on difficult issues, but few issues have presented the legislative, scientific, and public health dilemmas that accompany recommending criteria to control the exposure of workers to radon progeny in underground mines. The development of this criteria document is subject to the provisions of two legislative mandates. First, the Occupational Safety and Health Act of 1970 [Public Law (PL) 91—596, which established NIOSH] requires safe and healthful working conditions for every working person. The Act further requires NIOSH to preserve our human resources by providing medical and other criteria that will ensure, insofar as practicable, that no worker will suffer diminished health, functional capacity, or life expectancy as a result of work experience [PL 91—596, Sections 6(b)(5)]. The Act also authorizes NlOSH to recommend new criteria to further improve working conditions [PL 91—596, Sections 22(c) and (d)]. In addition, the Federal Coal Mine Health and Safety Act of 1969 [PL 91—173] and the Federal Mine Safety and Health Amendments Act of 1977 [PL 95—164] require NIOSH to develop and revise recommended occupational safety and health standards for mine workers. Specifically, the Secretary of Health, Education, and Welfare (now the Secretary of Health and Human Services) is required to consider, "in addition to the attainment of the highest degree of health protection for the miner . . . the latest available scientific data in the field, the technical feasibility of the standards, and experience gained under this and other health statutes" [PL 91-173, Title 1, Section 101(d)]. These mandates have required NlOSH to weigh its obligation to assure the highest degree of health protection for miners against the technical feasibility of the recommended standard in the development of recommendations for controlling radon progeny exposure in underground mines. The control of exposure to radon progeny presents an unprecedented problem because of the ubiquitous yet variable nature of their presence in mines and the ambient environment. To complicate this matter further, recent reports indicate that an exposure—related health risk may exist at background exposure levels. The full ramifications of this dilemma can easily be appreciated by considering two points. The first is that dilution ventilation (the primary engineering approach to reducing the concentration of radon progeny in mines) is accomplished by the exchange of mine air with air from the outside environment. Obviously, this approach is not a viable option for the total elimination of radon progeny in underground mines because the outside air is also contaminated with radon progeny. In addition, this approach would not be a prudent community environmental public health measure in some situations because it involves releasing an additional burden of radon progeny to the ambient environment and thereby contributing to the background level in the immediate area of the mine. Thus ventilation cannot be used to totally eliminate exposure to radon progeny in mines. The second point to consider in this dilemma is that the variable nature of radon progeny exposure in the ambient environment precludes recommending an annual cumulative exposure limit that includes both occupational and ambient contributions. Because ambient exposure varies, such a recommendation would result in an occupational exposure limit and associated risk that would vary with the locale. This approach is obviously undesirable, for it would lead to a nightmare of confusion and complicated enforcement requirements and w0uld probably result in unequal protection of miners. Data from both human and animal studies clearly demonstrate a direct link between lung cancer and radon exposure. Specific epidemiological studies provide a basis for quantitatively estimating human risk at various exposure levels. Such analyses clearly show that a radon exposure of 4 WLM (4 working level months) per year over a 30—year working lifetime (the current Mine Safety and Health Administration [MSHA] standard) poses a significant and unacceptable risk of lung cancer. This risk must be substantially reduced. In recommending an exposure limit for radon progeny, NIOSH considered not only the results of its own risk assessment and the technical feasibility of the recommended standard, but also the uncertainty of the data available on risk. Uncertainties are inherent in both the risk assessment methods and the scientific data on which the risk assessment is based. This fact must be understood and acknowledged. Some of the factors involved in these uncertainties include the choice of risk assessment method and model, the measurement methods used for data collection, and risk estimates derived from data that are heavily weighted with higher expOSures. The first of these factors in risk uncertainty involves the choice of a risk assessment method and/or model (such as the Cox proportional hazards model used in the NIOSH risk assessment study). NIOSH has attempted to develop a mathematical model that best describes the lung cancer risk in miners exposed to radon progeny. The use of a risk assessment model is merely a practical way to work with a very complex problem. There are modeling approaches other than the one chosen for this study. Each choice would result in a somewhat different description of the relationship between radon progeny exposure and lung cancer risk. NIOSH has attempted to compare the alternatives that are available and applicable. NIOSH scientists have considered the differences that might arise through a review of the available scientific literature and discussions with other scientists who have evaluated this exposure—related lung cancer risk. Although alternative models might yield minimally different quantitative risk estimates, none of them would lead the Institute to a qualitatively different risk assessment (i.e., that exposures to radon progeny at the current standard are associated with excesses of lung cancer). The second factor involved in risk uncertainty is the measurement method used for data collection. This study involves a follow—up period of more than 35 years, more than 3,000 miners, and thousands of measurements. The older data are subject to greater uncertainty than the more recent data because of improvements in the entire measurement process over the course of the study. The third factor involved in risk uncertainty is the process of generating risk estimates at lower exposure levels. One consideration is that such risk estimates are derived from data heavily weighted by higher exposures (note that the annual cumulative exposures of most miners in this study are higher than either the current MSHA standard or the proposed NIOSH recommended exposure limit [REL]). Another consideration is the desirability of placing occupational risk in the context of background exposure risk. However, the latter has not been evaluated and would have to be estimated on the basis of occupational data. We therefore do not believe that it is currently possible to contrast these two types of risks. Nonetheless, EPA has generated some initial information on background exposure risk in A Citizen's Guide to Radon. This document indicates that action should be taken to lower radon progeny levels in homes with measured concentrations of 0.02 WL or greater. NIOSH estimates that this concentration would probably result in a cumulative exposure that is less than 1 WLM but within an order of magnitude of that value. New information is clearly needed on background exposure levels and the hazards associated with such exposure before occupational and nonoccupational risks can be reliably quantified and validly contrasted. Until these data are available, the final target exposure limits cannot be identified for control of this hazard in our total environment. The uncertainties in the data and a recent study commissioned by the Bureau of Mines on the feasibilities of controlling radon progeny levels in mines have been weighed along with the available evidence and the obligations of NIOSH. This process has resulted in an REL of 1 WLM per year. Our own quantitative risk assessment clearly shows that significant health risks are posed by an exposure level of 1 WLM per year over a 30—year working lifetime. NIOSH therefore regards this REL as an upper limit and further recommends that mine operators limit exposure to radon progeny to the lowest levels possible. In addition, NIOSH wishes to emphasize that this recommended standard contains many important provisions in addition to the annual exposure limit. These include recommendations for limited work shift concentrations of radon progeny, sampling and analytical methods, recordkeeping, medical surveillance, posting of hazardous information, respiratory protection, worker education and notification, and sanitation. All of these recommendations help minimize risk. In summary, NIOSH has the legislative, scientific, and public health responsibility to protect the health of miners by developing recommendations *Bloomster CH, Enderlin WI, Young JK, Dirks JA (1984). Cost survey for radon daughter control by ventilation and other control techniques. Volume 1. Richland, WA: Battelle Memorial Institute, Pacific Northwest Laboratories, NTIS PB85-152932. that eliminate or minimize occupational risks. Although I am approving the recommended exposure limit of 1 WLM per year, I do not feel that this part of the recommended standard fully satisfies the lnstitute's commitment to protect the health of all of the Nation's miners. Future research may provide evidence of new and more effective methods for reducing occupational exposures to radon progeny, more reliable risk estimates at low exposure levels, and improved risk assessment methods. If new information demonstrates that a lower exposure limit constitutes both prudent public health and a feasible engineering policy, NlOSH will revise its recommended standard. MW Millar, M.D., D.T.P.H. (Lond.) Assistant Surgeon General Director, National Institute for Occupational Safety and Health Centers for Disease Control vi CONTENTS Foreword. . . Acknowledgments . I. Recommendations for a Radon Progeny Standard. Section 1. Definitions. . . Section 2. Environment (Workplace Air). . Section 3. Monitoring and Recording Exposures . Section 4. Medical Surveillance . . . . Section 5. Posting. . Section 6. Work Practices and Engineering Controls. Section 7. Respirator Selection and Credit for Respirator Use . . Section 8. Informing Workers of the Hazards of Radon Progeny. Section 9. Sanitation . . . Section 10. Recordkeeping Requirements . ll. Introduction . A. Scope . . Current Standard. Uranium Decay Series. Units of Measure. Worker Exposure . Measurement Methods for Airborne Radon Progeny. OW‘II‘HUOW lll. Basis for the Recommended Standard . Assessment of Effects . Risk Assessment . . . Technical Feasibility . Recommendations . UCW> IV. Research Needs . . . A. Epidemiologic Studies . . B. Engineering Controls and Work Practices . C. Respiratory Protection. . D. Environmental (Workplace) Monitoring. V. Public Health Perspective. References. Appendices . Evaluation of Epidemiologic Studies Examining the Lung Cancer Mortality of Underground Miners . . ll. Quantitative Risk Assessment of Lung Cancer in U. S. Uranium Miners . . . . . . . . lll. Engineering Control Methods. Grab Sampling Strategy Requirements for Determination of Radon Progeny Exposures . . V. Medical Aspects of Wearing Respirators . vii Respirator Selection and Credit for Respirator Use. iii viii Chm-bNN—L—L 65 138 177 193 204 ACKNOWLEDGMENTS The document was developed by the Division of Standards Development and Technology Transfer, National Institute for Occupational Safety and Health, Richard A. Lemen, Director, Richard W. Niemeier, Ph.D., Deputy Director, and William D. Wagner, Chief of the Document Development Branch. The criteria document manager was Mary Ballew. David D. Bayse, Ph.D., Robert Bernstein, M.D., Ph.D , Kenneth Busch, Burt Cooper, Kent Hatfield, Ph.D., Richard Hornung, Ph.D., Howard Ludwig, Lawrence Mazzuckelli, Theodore Meinhardt, Ph.D., Mary Newman, Ph.D., Sheldon Rabinovitz, Ph D., Laurence Reed, Randall Smith, Sandra Susten, Ph.D., and Ralph Zumwalde are gratefully acknowledged for their contributions to the writing and critical review of this document. John Bainbridge, Nancy J. Bollinger, Donald L. Campbell, David Groth, M.D., Bryan Hardin, Ph.D., and John Whalen are also gratefully acknowledged for their contributions to the critical review of this document. Special appreciation is extended to Vanessa Becks, Carolyn Browning, Ruth Grubbs, Anne Hamilton, and Eileen Kuempel for their editorial review, to Richard Carlson and Ann Stirnkorb for their preparation of figures, and to Judy Curless, Brenda Ellis, Denise Hill, Susan Marksberry, and Carol A. Ritchey for their support in typing and preparing the document for publication. viii I. RECOMMENDATIONS FOR A RADON PROGENY STANDARD The National Institute for Occupational Safety and Health (NIOSH) recommends that worker exposure to radon progeny in underground mines be controlled by compliance with this recommended standard, which is designed to protect the health of underground miners over a working lifetime of 30 years. Mine operators should regard the recommended exposure limit for radon progeny as the upper boundary for exposure; they should make every effort to limit radon progeny to the lowest possible concentrations. This recommended standard will be reviewed and revised as necessary. Radon progeny (also known as radon daughters) are the short—lived decay products of radon, an inert gas that is one of the natural decay products of uranium. The short—lived radon progeny (i.e., polonium—210, lead—214, bismuth-214, and polonium-214) are solids and exist in air as free ions or as ions attached to dust particles. The NIOSH recommended exposure limit (REL) is based on (1) evidence that a substantial risk of lung cancer is associated with an occupational exposure to radon progeny, and (2) the technical feasibility of reducing exposures. In this document, NIOSH presents recommendations that will protect miners employed year-round at any mine work area for as long as 30 years (the period of time used by MSHA as a miner's working lifetime). The exposure limit contained in this recommended standard is measurable by techniques that are valid, reproducible, and available to industry and government agencies. NIOSH has concluded that current technology is sufficient to achieve compliance with the recommended standard. Because knowledge of the carcinogenic process is incomplete and no data exist to demonstrate a safe level of exposure to carcinogens, NIOSH maintains that occupational exposure to carcinogens such as radon progeny should be reduced to the lowest level technically achievable. Compliance with this standard does not relieve mine operators from complying with other applicable standards. Section 1 - Definitions (a) Miner Miners include all mine personnel who are involved with any underground operation (e.g., drilling, blasting, haulage, and maintenance). (b) Working Level One working level (WL) is any combination of short—lived radon progeny in 1 liter (L) of air that will ultimately release 1.3 x 10 million electron volts (MeV) of alpha energy during decay to lead—210. (c) Working Level Month A working level month (WLM) is the product of the radon progeny concentration in WL and the exposure duration in months. For example, if a miner is exposed at a concentration of 0.083 WL for 1 month 1 (170 hours [hr]),* then the cumulative exposure for the month is 0.083 WLM. If the cumulative exposure of the same miner is 0.083 WLM for each of 12 consecutive months (2,040 hr), then the cumulative exposure for the year is 1 WLM. (d) Work Area A work area is any stope, drift heading, travelway, haulageway, shop, station, lunchroom, or any other underground location where miners work, travel, or congregate. (6) Average Work Shift Concentration The average work shift concentration is the average concentration of radon progeny present during a work shift in a given area. This concentration is used to represent the miner's breathing zone exposure to radon progeny. Section 2 — Environment (Workplace Air) (a) Recommended Exposure Limit (REL) Exposure to radon progeny in underground mines shall not exceed 1 WLM per year, and the average work shift concentration shall not exceed 1/12 of 1 WL (or 0.083 WL). The REL of 1 WLM per year is an upper limit of cumulative exposure, and every effort shall be made to reduce exposures to the lowest levels possible. (b) Sampling and Analysis Grab samples for radon progeny in the workplace shall be taken and analyzed using working level monitors, the Kusnetz method, or any other method at least equivalent in accuracy, precision, and sensitivity. Sampling and analytical methods are described in Chapter II. Details of the recommended sampling strategy are contained in Appendix IV. The recommended sampling strategy allows the use of grab samples for estimating the average work shift concentration of radon progeny. Section 3 - Monitoring and Recording Exposures (a) Exposure Monitoring All operators of underground mines shall perform environmental evaluations in all work areas to determine exposures to radon progeny. (1) An initial environmental evaluation shall be conducted in each work area to determine the average work shift concentration of radon progeny. *Note that Mine Safety and Health Administration (MSHA) regulations are based on 173 hr per month. (2) Periodic environmental evaluations shall be conducted at intervals (as described in Appendix IV) in each work area. An alternative sampling strategy may be used if the mine operator can demonstrate that it effectively monitors exposure to radon progeny. (3) if environmental monitoring in a work area indicates that the average work shift concentration of radon progeny exceeds 1/12 WL (as described in Appendix IV), the mine operator shall prepare an action plan describing the types of engineering controls and work practices that will be implemented to reduce the average work shift concentration in that area. (b) Exposure Monitoring Records The mine operator shall determine and record the exposure to radon progeny. Each miner's exposure shall be calculated using monitoring data obtained for the areas in which the miner worked. These records shall include (1) locations, dates, and times of measurements, (2) sampling and analytical methods used, (3) the number, duration, and results of the samples taken, and (4) all items required by Sections 3(b)(2) and (3). All records shall be retained at the mine site or nearest mine office as described in Section 10. (1) Calculating the Miner's Daily Exposure The average work shift concentration of radon progeny for each work .area shall be used to calculate each miner's daily exposure. If no monitoring has been conducted in a work area on a particular day, the daily average work shift concentration for that area shall be determined by averaging the results obtained on the last day of monitoring with the results from the next day that monitoring is conducted. A miner's exposure (in WLM) for a given area is calculated as follows: _ WL x T ' 170 hr where WL is the average work shift concentration of radon progeny, T is the total time (hours) spent in the area, and 170 is the number of hours worked per month. A miner's total cumulative exposure for the year is the sum of the daily exposures (as calculated above) for all work areas in which time was spent during the work shift. (2) Uranium Mines Exposure to radon progeny shall be recorded daily for each uranium miner. These records shall include the miner's name, social security number, the time spent in each work area, estimated exposure to radon progeny for each work area as determined in 3 Section 3(b)(3), and (if applicable) the type of respiratory protection and duration of its use. (3) Nonuranium Mines Exposure to radon progeny shall be recorded daily for all miners assigned to work in areas where environmental monitoring for radon progeny is required as described in Appendix IV. These exposure monitoring records shall include the miner's name, social security number, time the miner has spent in each work area, estimated exposure to radon progeny for each work area as determined in Section 3(b)(3), and (if applicable) the type of respiratory protection used and the duration of its use. (4) Respirator Credit The type of respirator worn and the credit given for wearing it (see Section 7) shall be recorded for each miner. Mine operators shall record both the average work shift concentration of radon progeny-and the adjusted exposure concentration calculated by using the respirator credit. The adjusted exposure concentration shall be used to determine the miner's cumulative exposure for compliance with the REL of 1.0 WLM/year. Section 4 - Medical Surveillance (a) General (1) The mine operator shall institute a medical surveillance program for all miners. (2) The mine operator shall ensure that all medical examinations and procedures are performed by or under the direction of a licensed physician. (3) The mine operator shall provide the required medical surveillance at a reasonable time and place without loss of pay or cost to the miners. (4) The mine operator shall provide the following information to the physician performing or responsible for the medical surveillance program: a copy of the radon progeny standard, the miner's duration of employment, the miner's cumulative exposure to radon progeny (or an estimate of potential exposure to radon progeny if the miner is a new employee), a description of the miner's duties as they Helate to his exposure, and a description of any protective equipment the miner has used or may be required to use. (5) The mine operator or physician shall counsel tobacco—smoking miners about their increased risk of developing lung cancer from the combined exposure to tobacco smoke and radon progeny. The mine operator or physician shall encourage the miner to participate in a 4 smoking cessation program. The mine operator shall enforce a policy prohibiting smoking at the mine site. (6) The physician shall provide the mine operator and the miner with a written statement describing any medical conditions found during the preplacement or periodic medical examinations that may increase the miner's health risk when exposed to radon progeny. This written statement shall not reveal specific findings, but shall include any recommended limitations on the miner's exposure to radon progeny or ability to use respirators and other personal protective equipment. (b) Preplacement Medical Examination The preplacement medical examination of each miner shall include the following: (1) A comprehensive medical and work history (including smoking history) that emphasizes the identification of existing medical conditions and attempts to elicit information about previous occupational exposure to radon progeny. (2) A thorough examination of the miner's respiratory system, including pulmonary function tests. The initial and subsequent pulmonary function tests shall include determination of forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV1) using the current American Thoracic Society (ATS) recommendations on instrumentation, technician training, and interpretation. A prospective miner with symptomatic, spirometric, or radiographic evidence of pulmonary impairment should be counseled about the risks of continued exposure. (3) A posterio—anterior chest X-ray using the current ATS recommendations on instrumentation, technician training, and interpretation. (4) Other tests deemed appropriate by the physician. (c) Periodic Medical Examination The periodic medical examination for each miner shall include the following: (1) An annual update of medical and work histories (including smoking history). (2) An evaluation of the miner's respiratory system. Because of the potential for chronic respiratory disease, this evaluation shall include spirometry at intervals determined by the physician. Miners that have spirometric or radiographic evidence or symptoms of pulmonary impairment should be counseled by the physician regarding the risks of continued exposure. (3) A posterio—anterior chest X—ray at intervals determined by the physician using the current ATS recommendations on instrumentation, technician training, and interpretation. Periodic chest X—rays are recommended for monitoring miners exposed to fibrogenic respiratory hazards (e.g., quartz). Ordinarily, chest X—rays may be obtained every 5 years for the first 15 years of employment and every 2 years thereafter, depending on the nature and intensity of exposures and their related health risks. A recent X—ray obtained for other purposes (e.g., upon hospitalization) may be substituted for the periodic X—ray if it is of acceptable quality. (4) Other tests deemed appropriate by the physician. Section 5 - Posting All warning signs shall be printed in both English and the predominant language of non—English-reading miners. Miners unable to read the posted signs shall be informed verbally about the hazardous areas of the mine and the instructions printed on the signs. (a) Readily visible signs containing the following information shall be posted at mine entrances or in work areas that require environmental monitoring for radon progeny as described in Appendix IV: AUTHORIZED PERSONNEL ONLY DANGER! POTENTIAL RADIATION HAZARD RADON PROGENY (b) If respiratory protection is required, the following statement shall be added in large letters to the sign required in Section 5(a): RESPIRATORY PROTECTION REQUIRED IN THIS AREA Section 6 — Work Practices and Engineering Controls Effective work practices and engineering controls shall be instituted by the mine operator to reduce the concentration of radon progeny to the lowest technically achievable limit. Since there is no typical mine and each operation has some unique features, the work practices and engineering controls in this section may need to be adapted for use in particular situations. (a) Work Practices (1) Ore Extraction and Handling Examples of effective ore extraction and handling procedures include the following: minimizing the number of ore faces simultaneously exposed, performing retreat mining toward intake air, limiting the underground storage and handling of ore, locating ore transfer points away from ventilation intakes, removing dust 6 spilled from ore cars, minimizing ore spillage by maintaining roadways and carefully loading haulage vehicles, and covering ore until it is moved to the surface. (2) Blasting Blasting should be performed at the end of the work shift whenever possible. Miners shall be evacuated from exhaust drifts until environmental sampling confirms that the average work shift concentration of radon progeny does not exceed 1/12 WL. Refer to Section 7 if respiratory protection is required for subsequent reentry. (3) Worker Rotation The mine operator shall not use the planned rotation of miners to maintain an individual's exposure below the REL of 1.0 WLM per year. NlOSH acknowledges, however, that some miners may inadvert— ently be exposed to short—term high concentrations of radon progeny. For example, such exposures may occur when engineering controls fail. To ensure that the miners' cumulative exposure remains below the REL in such circumstances, it may be necessary to transfer them to other jobs or work areas that have lower concentra— tions of radon progeny. Miners transferred under these circum— stances shall retain their pay as prescribed for coal miners under Section 203(b) of the Federal Coal Mine Safety and Health Act of 1977. (b) Engineering Controls Mechanical exhaust ventilation used alone or in combination with other engineering controls and work practices can effectively reduce exposures to radon progeny. Ventilation systems discharging outside the mine shall conform with applicable local, State, and Federal [40 CFR 61, Subpart B] air pollution regulations and shall not constitute a hazard to miners or to the general population. (1) Ductwork shall be kept in good repair to maintain designed airflows. The effectiveness of mechanical ventilation systems shall be determined periodically and as soon as possible after any significant changes have been made in production or control. A log shall be kept showing designed airflow and the results of all airflow measurements. (2) Fans shall be operated continuously in the work areas of an active mine and before the opening of a previously inactive mine or inactive section until environmental sampling confirms that the average work shift concentrations of radon progeny do not exceed 1/12 WL. Refer to Section 7 if respiratory protection is required. *Code of Federal Regulations. See CFR in references. 7 (3) Fresh air shall be provided to miners in dead end areas near the working faces. (4) Bulkheads, backfill, and sealants shall be used to control exposures as appropriate. Appendix III provides a general discussion of engineering control methods. Section 7 - Respirator Selection and Credit for Respirator Use (a) General Considerations NIOSH has determined that a radon progeny exposure limit of 1.0 WLM per year is technically achievable in mines through the use of effective work practices and engineering controls. Over a 30—year working lifetime, this exposure limit will reduce but not eliminate the risk of lung cancer associated with exposure to radon progeny. NIOSH considers respirators to be one of the last options for worker protection. Work practices and engineering controls are more effective means for limiting exposures and providing a safe environment for all workers. Respirator use in underground mines is not always practical for a number of reasons, including the additional physiological burden and safety hazards they pose. NIOSH therefore recommends that engineering controls and work practices be used where technically achievable to control the exposure of miners to radon progeny. Compliance with an exposure limit of 1.0 WLM per year requires an average exposure of 1/12 WL throughout the year to ensure that the miner can work for an entire year (i.e., 2,040 hr). For average work shift concentrations above 1/12 WL, NIOSH recommends mandatory respirator use as well as the implementation of engineering controls and work practices to reduce exposure to radon progeny. Occupational exposure to radon progeny above background concentrations has been associated with excess lung cancer risk. Therefore, regardless of the exposure concentration, NIOSH advises the use of respirators to further reduce exposure and decrease the risk of lung cancer. Respiratory protection shall be used by miners (1) when work practices and engineering controls are not adequate to limit average work shift concentrations of radon progeny to 1/12 WL, (2) when entering a mine area where concentrations of radon progeny are unknown, or (3) during emergencies. Use only those respirators approved by NIOSH or the Mine Safety and Health Administration (MSHA). (b) Respirator Protection Program Whenever respirators are used, a complete respiratory protection program shall be instituted. This program must follow the recommendations contained in ANSI 288.2—1969 (published by the American National Standards Institute) and the respirator—use criteria in 30 CFR 57.5005. The respiratory protection program described in ANSI 288.2—1969 requires the following: (1) A written program for respiratory protection that contains standard operating procedures governing the selection and use of respirators. (2) Periodic worker training in the proper use and limitations of respirators. (3) Evaluation of working conditions in the mine. (4) An estimate of anticipated exposure. (5) An estimate of the physical stress that will be placed on the miner. A detailed medical examination of each miner shall be conducted according to the guidelines set forth in Appendix V. (6) Routine inspection, maintenance, disinfection, proper storage, and evaluation of respirators. (7) Information concerning the manufacturers' instructions for respirator fit—testing and proper use. (c) Respirator Selection NIOSH makes the following recommendations for respirator selection: (1) A respirator is not required for exposure to average work shift concentrations less than or equal to 1/12 WL. (2) For exposure to average work shift concentrations greater than 1/12 WL, NIOSH recommends those respirators listed in Table l—1. (3) For entry into areas where radon progeny concentrations are unknown or exceed 166 WL, or for emergency entry, NIOSH recommends only the most protective respirators (any full—facepiece, positive—pressure, self-contained breathing apparatus [SCBA] or full—facepiece, positive—pressure, supplied—air respirator and SCBA combination). These recommendations are based on the fact that radon progeny exist as particulates and that miners are not exposed to hazardous concentrations of nonparticulate contaminants. lf protection against nonparticulate contaminants is required, different types of respirators must be selected. (d) Credit for Respirator Use When respirators are worn properly, the miner's average work shift exposure can be reduced by a factor that depends on the class of respirator worn. Table l—1 provides the credit factors for the various classes of respirators. For example, if a miner wears a helmet—type, 9 OI Tabie I—l.——Respirator recommendations for radon progeny Average work shift concentration of Credit factor radon progeny for respirator use (HL) Respirator recommendations 65% utiiization 90% uti1ization 0 to 0.083 (1/12) No respirator required NA+ NA >0.083 to g 0.42 Any disposable respi ator 2.1 3.6 equipped with a HEPA fiiter Any more protective reSpirator # # >0.42 to g 0.83 Any air—purifying haif—mask 2.4 5.3 respirator equipped with a HEPA fiiter Any SAR“ equipped with a half— 2.4 5.3 mask and operated in a demand (negative—pressure) mode Any more protective respirator # # >o.s3 to 5 2.08 Any powered PAPR” equipped 2.7 7.4 with a hood or heimet and a HEPA fiiter Any SAR equipped with a hood 2.7 7.4 or heimet and operated in a continuous fiow mode # # Any more protective respirator See footnotes at end of tabie. TI Tabie I—l (Continued).——Respirator recommendations for radon progeny Average work shift concentration of radon progeny Respirator recommendations Credit factor for respirator use 65% utilization 90% utiiization >2.08 to g 4.15 >4.15 to g 83.0 Any air—purifying, fu11 face— piece respirator equipped with a HEPA fiiter Any PAPR equipped with a tight- fitting facepiece and a HEPA fi1ter Any SAR equipped with a fuii facepiece and operated in a demand (negative—pressure) mode Any SAR equipped with a tight— fitting facepiece and operated in a continuous-f1ow mode Any seif—contained breathing apparatus (SCBA) equipped with a fu11 facepiece and operated in a demand (negative—pressure) mode Any more protective respirator Any SAR equipped with a haif- mask and operated in a pressure— demand or other positive-pressure mode Any more protective respirator See footnotes at end of tabie. 2.8 2.8 2.8 2.8 2.8 2.9 8.5 8.5 8.5 8.5 8.5 9.9 ZI Tab1e I-1 (Continued).——Respirator recommendations for radon progeny Average work shift Credit factor concentration of for respirator use radon progeny Respirator recommendations 65% uti1ization 90% utilization >83.0 to 3 166.0 Any SAR equipped with a fu11 2.9 10.0 facepiece and operated in a pressure demand or other positive pressure mode Any more protective respirator >166.0 or unknown Any SCBA equipped with a fu11 2.9 10.0 concentration or facepiece and operated in a emergency entry pressure demand or other positive pressure mode Any SAR equipped with a fu11 2.9 10.0 facepiece operated in a pressure demand or other positive pressure mode in combination with an auxi1iary se1f—contained breathing apparatus operated in a pressure demand or other positive pressure mode Emergency escape Any se1f—contained se1f—rescuer NA NA (SCSR) itAs estimated using the samp1ing techniques described in Appendix IV. fNAzNot app1icab1e. §HEPA : high-efficiency particu1ate air. nSee appropriated credit factors be1ow. SAR = supp1ied—air respirator. HPAPR : powered air-purifying respirator. powered, air—purifying respirator (PAPR) for 65% of the work shift and the radon progeny concentration in the work area is 0.3 WL, then the miner's exposure can be adjusted by dividing 0.3 WL by 2.7, the credit factor for this class of respirator. This results in an adjusted exposure of 0.11 WL for that miner. Respirator credit is discused in detail in Chapter II. Section 8 - Informing Workers of the Hazards of Radon Progeny (a) Notification of Hazards The mine operator shall provide all miners with information about workplace hazards before job assignment and at least annually thereafter. (b) Training (1) The mine operator shall institute a continuing education program conducted by persons with expertise in occupational safety and health. The purpose of this program is to ensure that all miners have current knowledge of workplace hazards, effective work practices, engineering controls, and the proper use of respirators and other personal protective equipment. This program shall also include a description of the general nature of the environmental and medical surveillance programs and the advantages of participating in them. This information shall be kept on file and be readily available to miners for examination and copying. The mine operator shall maintain a written plan of these training and surveillance programs. (2) Miners shall be instructed about their responsibilities for following proper work practices and sanitation procedures necessary to protect their health and safety. Section 9 — Sanitation (a) Eating and Drinking The preparation, storage, dispensing (including vending machines), or consumption of food shall be prohibited in any area where a toxic material is present. The mine operator shall provide facilities so that miners can wash their hands and faces thoroughly with soap or mild detergent and water before eating or drinking. (b) Smoking Smoking shall be prohibited in underground work areas. (c) Toilet Facilities The mine operator shall provide an adequate number of toilet facilities and encourage the miners to wash their hands thoroughly with soap or mild detergent and water before and after using these facilities. 13 (d) Change Rooms (1) The mine operator shall provide clean change rooms for the miners. (2) The mine operator shall provide storage facilities such as lockers to permit the miners to store street clothing and personal items. (6) Showers The mine operator shall provide showers and encourage the miners to shower at the end of the work shift. (f) Laundering (1) The mine operator shall provide for the cleaning, laundering, or disposal of contaminated work clothing and equipment. (2) The mine operator shall ensure that contaminated work clothing or equipment that is to be cleaned, laundered, or disposed of is placed in a closed container to prevent dispersion of dust. (3) Any person who cleans or launders this contaminated work clothing or equipment must be informed by the operator that it may be contaminated with radioactive materials. Section 10 - Recordkeeping Requirements (a) Record Retention (1) The mine operator shall retain all records of the monitoring required in Section 3(b). (2) All monitoring records shall be retained for at least 40 years after termination of employment. (3) The mine operator shall retain the medical records required by Section 4. These records shall be retained for at least 40 years after termination of employment. (b) Availability of Records The miner shall have access to his medical records and be permitted to obtain copies of them. Records shall also be made available to former miners, or their representative and to the designated representatives of the Secretary of Labor and the Secretary of Health and Human Services. (c) Transfer of Records (1) Upon termination of employment, the mine operator shall provide the miner with a copy of his records specified in Section 10(a). 14 (2) Whenever the mine operator transfers ownership of the mine, all records described in this section shall be transferred to the new operator, who shall maintain them as required by this standard. (3) Whenever a mine operator ceases to do business and there is no successor, the mine operator shall notify the miners of their rights of access to those records at least 3 months before cessation of business. (4) The Director of NIOSH shall be notified in writing before (a) a mine operator ceases to do business and there is no successor to maintain records, and (b) the mine operator intends to dispose of those records. (5) No records shall be destroyed until the Director of NIOSH responds in writing to the mine operator. 15 ll. INTRODUCTION A. Scope Radon is a gas that diffuses continuously from surrounding rock and broken ore into the air of underground mines, where it may accumulate; radon may also be carried into mines through groundwater containing dissolved radon [Snihs 1981]. Radon gas may be inhaled and immediately exhaled without appreciably affecting the respiratory tissues. However, when attached or unattached radon progeny are inhaled, they may be deposited on the epithe— lial tissues of the tracheobronchial airways. Alpha radiation may subse— quently be emitted into those tissues from polonium-218 and polonium—214, thus posing a cancer risk to miners who inhale radon progeny. This document presents the criteria and recommendations for an exposure standard that is intended to decrease the risk of lung cancer in miners occupationally exposed to short—lived, alpha—emitting decay products of radon (radon progeny) in underground mines. The REL for radon progeny applies only to the workplace and is not designed to protect the population at large. The REL is intended to (1) protect miners from the development of lung cancer, (2) be measurable by techniques that are valid, reproducible, and available to industry and government agencies, and (3) be technically achievable. B. Current Standard MSHA has established radiation protection standards for workers in underground metal and nonmetal mines [30 CFR 57.5037 through 57.5047]. This standard limits a miner's radon progeny exposure to a concentration of 1.0 WL and an annual cumulative exposure of 4 WLM. Each WLM is determined as a 173—hr cumulative, time-weighted exposure [30 CFR 57.5040(6)]. Smoking is prohibited in all areas of a mine where radon progeny exposures must be determined; respiratory protection is required in areas where the concen- tration of radon progeny exceeds 1.0 WL. According to current MSHA regulations, the exhaust air of underground mines must be sampled to determine the concentration of radon progeny. 1. Uranium Mines If the concentration of radon progeny in the exhaust air of a uranium mine exceeds 0.1 WL, samples representative of a miner's breathing zone must be taken at random times every 2 weeks in each work area (i.e., stopes, drift headings, travelways, haulageways, shops, stations, lunchrooms, or any other place where miners work, travel, or congregate). lf concentrations of radon progeny exceed 0.3 WL in a work area, sampling must be done weekly until the concentration has been reduced to 0.3 WL or less for 5 consecutive weeks. Uranium mine operators must calculate, record, and report to MSHA the radon progeny exposure of each underground miner. The records must 16 include the miner's time in each work area and the radon progeny concentration measured in each of those areas. 2. Nonuranium Mines If the concentration of radon progeny in the exhaust air of nonuranium mines exceeds 0.1 WL, and if concentrations are between 0.1 and 0.3 WL in an active working area, samples representative of a worker's breathing zone must be taken at least every 3 months at random times until the concentrations of radon progeny are less than 0.1 WL in that area. Samples must be taken annually thereafter. If the concentration of radon progeny exceeds 0.3 WL in a working area, samples must be taken at least weekly until the concentration has been reduced to 0.3 WL or less for 5 consecutive weeks. Operators of nonuranium mines must calculate, record, and report to MSHA the radon progeny exposures of miners assigned to areas with concentrations of radon progeny exceeding 0.3 WL. The records must include the miner's time in each work area and the radon progeny concentration measured in each of those areas. c. Uranium Decay Series Fi ure ll-1 shows the sequence by which the most abundant isotope of uranium (2 8U) decays to a radioactively stable isotope of lead (206Pb). Radon (222Rn) is an inert gas with a radiologic half-life of 3.8 days; it is a product of the natural decay of radium (225Ra). When radon decays, alpha particles and gamma radiation are emitted, and an isotope of polonium (218Po) is formed. Polonium—218 218Po) and its decay products—— lead—214 (214m)), bismuth—214 (21 Bi), and polonium—214 (214Po)--are commonly referred to as short-lived radon progeny because they have half—lives of 27 minutes or less (see Figure ll—1). Both polonium—218 and polonium—214 emit alpha particles as they decay. The short—lived progeny are solids and exist in air as free ions (unattached progeny) or as ions adsorbed to dust particles (attached progeny). Because it is a gas, radon diffuses through rock or soil and into the air of underground mines, where it may accumulate; radon may also be carried into mines through groundwater containing dissolved radon [Snihs 1981]. Radon may be inhaled and immediately exhaled without appreciably affecting the respiratory tissues. However, when the radon progeny (either attached or unattached) are inhaled, they may be deposited in the epithelial tissues of the tracheobronchial airways, where alpha radiation from polonium—218 and polonium—214 may be subsequently emitted. The quantity of mucus in those airways and the efficiency of its clearance (retrograde ciliary action) into the esophagus are important factors that affect the total radiation absorbed at a specific site within the respiratory tract. Alpha particles are energetic helium nuclei. As they pass through tissue, they dissipate energy by the excitation and ionization of atoms in the tissue; it is this process that damages cells. Because alpha particles travel less than 100 micrometers in tissue, intense ionization occurs close to the site of deposition of the inhaled alpha—emitting radon progeny. Beta particles (electrons) and gamma radiation (shortwave electromagnetic radiation) can also cause ionization in tissues, but they travel farther 17 234 (Taken from Radiation Policy Council 1980.) 18 238 92U 92U 5 4.5 x109yr 2.5 x 10 yr ‘7 a 234 B’ an 91Pa ATOMIC WT. * 1.17 min ELEMENT ATOMIC NO. HALF— LIFE 234 230 9011‘ (3,7 90m 24 days 8 x 104yr Abbreviation Element a 7 . ' U Uranium 226R Pa Protactinium 88 3 Th Thorium 1600 yr Ra Radium Rn Radon P0 Polonium 222 01,7 BI Bismuth 86“" Pb Lead 3.8 days Shortlived radon ro en 04,7 / p g y 218 214 210p 84P° 84P° 84 ° -4 3 min 1.6 x 10 sec 138 days 7 a 214 fi’ 210 X 833' 833' “’7 19.7 min 5 days 5,7 6 7 206 1 82Pb 22-3 yr Stable Figure l l—1.—-The uranium (238U) decay series. through tissues and dissipate less energy per unit path length than do alpha particles [Casarett 1968; Wang et al. 1975; Shapiro 1981]. The beta parti- cles and gamma radiation emitted by radon progeny make a negligible contri— bution to the radiation dose in the lung [Evans 1969]. D. Units of Measure The common unit of radioactivity is the curie (Ci), which is the rate at which the atoms of a radioactive substance decay; 1 Ci equals 3.7 x 1010 disintegrations per second (dps). The picocurie (pCi) corresponds to 3.7 x 10‘2 dps. The International System of Units (SI) unit of radio— activity is the becquerel (Bq), which is equivalent to 1 dps. Therefore, 1 pCi is equivalent to 0.037 Bq. When radon gas and radon progeny are inhaled, the radiation exposure is primarily caused by the short—lived radon progeny (polonium—218, lead—214, bismuth-214, and polonium—214, which are deposited in the lung) rather than by the radon gas. Because it was not feasible to routinely measure the individual radon progeny, the U.S. Public Health Service introduced the concept of the working level, or WL [Holaday et al. 1957]. The WL unit represents the amount of alpha radiation emitted from the short—lived radon progeny. One WL is any combination of short-lived radon progeny in 1 liter (L) of air that will ultimately release 1.3 x 105 million electron volts (MeV) of alpha energy during decay to lead-210. The SI unit of measure for potential alpha energy concentration is joules per cubic meter of air (J/m3); 1 WL is equal to 2.08 x 10-5 J/m3 [lCRP 1981]. The equilibrium between radon gas and radon progeny must be known in order to convert units of radioactivity (Ci or Bq) to a potential alpha energy concentration (WL or J/m3). The equilibrium factor (F) is defined as the ratio of the equilibrium—equivalent concentration of the short-lived radon progeny to the actual concentration of radon in air [lCRP 1981]. When the equilibrium factor approaches 1.0, it means that the concentration of radon progeny is increasing relative to the concentration of radon. At complete radioactive equilibrium (F=1.0), the rate of radon progeny decay equals the rate at which the progeny are produced. Thus the radioactivity of the decay products equals the radioactivity of the radon [Shapiro 1981]. In under— ground mines, the equilibrium factor mainly depends on the ventilation rate and the aerosol concentration [Urban et al. 1985]. Values of F ranging from 0.08 to 0.65 are typical in underground mines [Breslin et al. 1969]. Radioactivity and potential alpha energy concentration values at various equilibria are presented in Table ll—1. The common unit of measure for human exposure to radon progeny is the working level month (WLM). One WLM is defined as the exposure of a worker to radon progeny at a concentration of 1.0 WL for a working period of 1 month (170 hr). The SI unit for WLM is jouIe—hour per cubic meter of air (J—h/m3); 1 WLM is equal to 3.6 x 10-3 J—h/m3. *Note that MSHA regulations are based on 173 hr per month. 19 Table ll—1.——Potential alpha energy concentration as a function of the equilibrium factor Radioactivity Potential alpha energy concentration Equilibrium factor (F) pCi Bq WL J/m3 0.30 1 0.037 0.003 6.24x1o-8 0.30 333 12.3 1.000 2.08x10-5 0.50 1 0.037 0.005 1.04x10-7 0.50 200 7.40 1.000 2.08x10-5 1.00 1 0.037 0.010 2.08x10-7 1.00 100 3.70 1.000 2.08x10-5 *F is defined as the quotient of the equilibrium—equivalent radon progeny activity divided by the radon activity. The rad (radiation absorbed dose) is the unit of measure for the absorbed dose of ionizing radiation. One rad corresponds to the energy transfer of 6.24 x 107 MeV per gram of any absorbing material [Shapiro 1981]. The rem (roentgen equivalent man) is the unit of measure for the dose equivalent of any ionizing radiation in man. One rem is equivalent to one rad multiplied by a radiation quality factor (OF). The radiation OF expresses the relative effectiveness of radiation with differing linear—energy—transfer (LET) values to produce a given biological effect. The radiation QFs for beta particles and gamma radiation are each approximately 1; the radiation GP for alpha radiation varies from 10 to 20 [NCRP 1975; ICRP 1977; NCRP 1984a]. For equal doses of absorbed radiation (rads), the dose equivalent (rems) attributed to alpha particles is 10 to 20 times greater than the dose equiva— lent attributed to high—energy beta particles or gamma radiation. The Sl unit of measure for the dose equivalent is the sievert (Sv). One rem is equal to 0.01 Sv [Shapiro 1981]. E. Worker Exposure In 1986, 22,499 workers were employed in 427 metal and nonmetal mines in the United States. In the past few years, the number of underground uranium mines operating in the United States has decreased dramatically from 300 in 1980 [Federal Register 1986] to 16 in 1984 [MSHA 1986]. Accordingly, the number of miners employed in these mines has also decreased from 9,076 in 1979 [Cooper 1981], to 1,405 in 1984 [AlF 1984], and to 448 in 1986 [MSHA 1986]. Table ll—2 shows the range in concentrations of airborne radon progeny measured in U.S. underground metal and nonmetal mines from 1976 through 1985 [MSHA 1986]. As illustrated in Table ll-3, 38 of the 254 operating underground nonuranium mines sampled during fiscal year 1985 contained concentrations of airborne radon progeny equal to or greater than 0.1 WL in 20 Table ||—2.——Radon progeny concentrations (WL) in underground metal and nonmetal mines from 1976 through 1985 Type of mine Range of annual geometric mean concentrations Range of highest annual concentrations (95th percentile) Boron Clay (common) Clay (fire) Copper Ore Fluorspar Gilsonite Gold Gypsum Iron Ore Lead/zinc Lime Limestone (crushed) Marble (crushed) Marble (dimension) Metalf Molybdenum Oil sand Oil shale Perlite Phosphate (rock) Platinum Potash Potash, soda, borate Salt (rock) Sandstone (crushed) Silver Slate (dimension) Talc (pyrophyllite) Tungsten Uranium Uranium/vanadium OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO .01—0 .01—0. .04—0. .02—0 .01—0. .01‘0 .03—0. .01—0 .01—0. .01—0. .01—0. .01—0 .01-0 .01—0 .01—0 .02—0 .01—0 .01—0 .01—0 .12—1 .01—0. .01—0. .01—0 .01-0. .01—0. .02—0 .02—0. .02-0. .02—0 .11-0. .10—0. .05 21 20 .08 29 .02 16 .06 28 13 09 .04 .04 .02 .33 .08 .02 .01 .02 .20 13 02 .02 06 11 .09 25 12 .31 36 25 .00—1.10 .01—0.54 .22—O.83 .04—1.45 .03—2.80 .00—0.23 .18—4.06 .00—O.56 .00—0.04 llllll Medical Screening and Surveillance of Underground Miners Exposed to Short—Lived Alpha Particles Technical Feasibility of Lowering the Standard Conclusion References Appendices A. Mining Industry: Current Workforce and Typical Radiation Exposure B. Current Methods of Control of Radiation Exposures in Underground Mines Glossary 66 I. SUMMARY The National Institute for Occupational Safety and Health (NIOSH) submits this report in response to the Mine Safety and Health Administration's (MSHA) Advanced Notice of Proposed Rulemaking (ANPR) concerning radiation standards for metal and nonmetal mines. In fifteen epidemiologic studies, researchers reported excess lung cancer deaths among underground miners who worked in mines where radon progeny were present. In addition, several studies show a dose—response relationship between radon progeny exposure and lung cancer mortality. In two recent studies, investigators report excess lung cancer deaths due to mean cumulative radon progeny exposures below 100 Working Level Months (WLM) (specifically, at 40-90 WLM and 80 WLM). The health risks from other exposures (i.e , arsenic, diesel exhaust, smoking, chromium, nickel, and radiation) in the mining environment can affect lung cancer risks due to radon progeny exposure. Unfortunately, the literature contains limited information about other exposures found in mines. The available information, concerning whether cigarette smoke and radon progeny exposures act together in an additive or multiplicative fashion is inconclusive; nevertheless, a combined exposure to radon progeny and cigarette smoke results in a higher risk than exposure to either one alone. X—ray surveillance and sputum cytology appear to be ineffective in the prevention of radon progeny—induced lung cancers in individual miners; therefore, these techniques are not recommended. Also, at this point, there is insufficient evidence to conclude that there is an association between one specific lung cancer cell type and radon progeny exposure. According to annual radon progeny exposure records from the Atomic Industrial Forum (AIF) and MSHA, it is technically feasible for the United States mining industry to meet a standard lower than the current annual exposure limit of 4 WLM. Recent engineering research suggests that it is technically feasible for mines to meet a standard as low as 1 WLM. Based upon qualitative analysis of these studies and public health policy, NIOSH recommends that the annual radon progeny permissible exposure limit (PEL) of 4 WLM be lowered. NIOSH wishes to withhold a recommendation for a specific PEL, until completion of a NIOSH quantitative risk assessment, which is now in progress. 67 ll. INTRODUCTION The National Institute for Occupational Safety and Health (NIOSH) submits this report in response to the Mine Safety and Health Administration's (MSHA) Advanced Notice of Proposed Rulemaking (ANPR) concerning radiation standards for metal and nonmetal mines. This report evaluates fifteen epidemiologic studies that examine the lung cancer mortality of underground miners exposed to radon progeny. The fifteen studies are divided into two groups: five primary studies and ten secondary studies. Overall, the ten secondary studies provide additional information about the association between lung cancer mortality and radon progeny exposure, yet have more limitations (in study design, study population size, radon exposure records, thoroughness of follow—up, etc.) than the five primary studies. Recommendations for the medical surveillance of underground miners exposed to radon progeny are included. The United States mining industry's ability to meet a radon progeny exposure standard lower than the present four Working Level Months (WLM), based solely on technical feasibility, is also discussed. A working level (WL) is a standard measure of the alpha radiation energy in air. This energy can result from the radioactive decay of radon (Rn-222) and thoron (Rn—220) gases. A WL is defined as any combination of short—lived radon decay products (polonium—218, lead—214, bismuth—214, polonium—214) per liter of air that will result in the emission of 1.3 X 105 million electron volts (MeV) of alpha energy [1]. NIOSH defines a WLM as an exposure to 1 WL for 170 hours. For the information of the reader, two appendices and a glossary are included. Appendix A contains data from the Atomic Industrial Forum (AlF) an organization representing the interests of the United States uranium mining industry, and MSHA on the numbers and radon progeny exposures of underground miners in the United States. Appendix B lists methods currently in use for controlling radon progeny exposures underground. Finally, there is a glossary containing epidemiologic and health physics terms. 68 III. EVALUATION OF EPIDEMIOLOGIC EVIDENCE A. Introduction This report examines five primary and ten secondary epidemiologic studies of underground miners. It describes the important points, strengths, and limitations of each study. The five primary epidemiologic studies examined lung cancer mortality among uranium miners in the United States, Czechoslovakia, and Ontario; iron miners in Malmberget, Sweden; and fluorspar miners in Newfoundland. The ten secondary epidemiologic studies examined mortality among iron ore miners in Grangesberg, Gallivare, and Kiruna, Sweden; zinc—lead miners in Sweden; metal and Navajo uranium miners in the United States; tin and iron ore miners in Great Britain; uranium miners in France; and tin miners in Yunnan, China. Finally, two recent studies analyze the interaction between radon progeny exposure and smoking. This report focuses on the lung cancer experience of these fifteen underground mining cohorts. In general, the study cohorts did not show excess mortality due to cancers other than lung, except for four studies that reported excess stomach cancers and one report of excess skin cancer among underground miners. Excess stomach cancers were reported among underground tin miners in Cornwall, England (standardized mortality ratio (SMR) = 200, p value unspecified by the authors, however estimated at p<0.05, from the observed deaths and the Poisson frequency distribution) [2]; gold miners in Ontario (SMR=148, p<0.001) [3]; metal miners in the United States (SMR=149, p90) Limestone 2.900 (c) 971,000 0.05—0.16 Open (>90) F1uorspar 1 14 593 0.30-2.20 Underground (>90) Bauxite 11 1 16,600 0.07—1.40 Open (>90) Uranium 36 251 15.900 mean = 0.10 Underground (>50) 3Data on underground exposure to radon decay products come from EPA pub1ications #520/7—79—006 (1979) and #520/4-80—001 (1980) [83,84]. resu1ts of a survey in 1975, 3,344 miners were emp1oyed that year. and nonmeta1 miners' exposures were not sufficient1y detai1ed to permit estimation of weighted mean annua1 exposures. Mine Enforcement and Safety Administration. A sma11, undetermined number are open-pit mines. cA sma11, undetermined number are underground mines. The uranium mining exposure data were taken from the The avai1ab1e data on meta1 The mine production data were taken from a 1978 survey report by the OII Tab1e A—2. Emp1oyment in the U.S. uranium mining industries, 1980-82 Year Underground Open—pit Technicai Other Supervisory Tota1 miners service and miners service and support support 1980 2,760 2,277 2,007 1,407 827 1,408a 1,082 11,768 1981 2,121 1,397 1,117 740 574 788 736 7,473 1982 1,275 875 792 573 503 426 613 5,057b aInc1udes 201 truckers and 371 emp1oyees invo1ved in shaft sinking and construction. May 1ack as many as 140 contract truckers. Taken from Statisticai Data of the Uranium Industry, U.S. Department of Energy, Grand Junction Area Office, Co1orado [85,86,87]. Tab1e A—3. Concentrations of, and exposure to, radon daughters in uranium mines III Average Average annua1 Country Year potentia1 potentia1 a1pha No. of No. of a1pha energy energy exposure miners miners concentration exceeding (WL) (HLM) 4 NLM a/ France 1971 0.18 -—— ——— ——— 1972 0.17 ——- —-- --— 1973 0.18 --- —-— -— 1974 0.13 -—— ——— -—- 1975 0.11 -—— ——— ——- 1976 -—— ——— ——— -- 1977 ——— --- — ——— 1978 ——- 2.0 1,284 Approx. 140 1979 —-- 1.4 1,503 51 United States 1975 0.71 5.68 Approx. 5,000 -—— 1976 0.58 4.64 Approx. 5,000 ——— 1977 0.51 4.08 Approx. 5,000 -—- Ita1y 1975 <1 -— —-- —-— Canada 1978 1 Leaching —-- 0.38 630 -—- 4 Underground ——— 0.74 3,690 ——— 1 Open-pit ——— 0.41 276 ——- 1978 ——— 0.72 4,535 9 1979 ——— 0.74 6,883 1 Argentina Underground 1977-79 -—- 2.4 286-379 ——- 1980 --- 2.4 95 0 Open-pit 1980 ——— 0 12 285 0 3The maximum permissibIe exposure in many countries. Data from the Nationa1 Dose Registry in Canada. -—- = data not avai1ab1e Taken fromI gnizing Radiiati1gn: Sources and EiQIQgi caI Effects. United Nations Scientific Committee on the Effects of Atomic Radiation, N. Y., 1982, p. 199 [82]. Nonuranium Miners a. Miners in the United States In 1984, MSHA reported that 23,721 miners (including 1,127 mining uranium) are employed full—time and 3,063 are employed part—time (includes 177 mining uranium) in metal and non metal underground mines in the United States. (Table A—4). Most of these miners are probably exposed to negligible quantities of radon progeny, although there is insufficient data to prove this. MSHA requires that underground nonuranium mining companies record the individual exposures of all miners who work in areas where radon progeny levels exceed 0.3 ML [88]. Table A—5 lists all of the mines that submitted individual records of exposure to MSHA during 1979—1983, and the number of miners for whom records were submitted. Some mines submitted records for all of their employees, including workers who received no radon progeny exposures; for example, 90 percent of the Climax and Henderson mine exposures were essentially zero during 1983. These mines occasionally have readings above 0.3 WL and, thus, are required to keep exposure records, but an individual miner's annual average exposure may be less than 4 WLM. During 1983, the mining companies were required to keep records on no more than 450 employees (Table A-5). The rest of the approximately 25,000 workers who mine in underground metal and non metal mines (excluding uranium) should receive even lower radon progeny exposures. Miners Outside the United States The figures for the number of hard rock miners are incomplete (see Table A—6). South Africa has a large number of hard rock miners, approximately 320,000, primarily employed in the gold mines. The most recent figures on the number of iron, zinc, lead, copper, or gold miners showed about 1,370 miners in Finland, 2,500 in Italy, 1,380 in Norway, 4,400 in Sweden, and 2,350 in Great Britain [82]. B. Current Exposures in Mining Industries 1. Uranium Miners a. Miners in the United States Most of the information on current underground uranium mining radiation exposure is reliant upon company records. There remains disagreement between the companies' records and the U.S. Mine Safety and Health Administration's (MSHA) inspection records [57]. The average annual cumulative exposure for all underground uranium mine workers is relatively low; members of the Atomic Industrial Forum (AIF) recorded an average exposure of 1.03 WLM in 1978 (see Tables A-7 through A—11). Because of the many temporary workers in the 112 €11 Tab1e A—4. Emp1oyment in United States meta1 and non meta1 underground mines June 26, 1984 Underground Fu11—time Intermittent/ mines personne1 Seasona1 Tota1 ESE—ILRIPT N 99‘er i n Lemmings Loam i Lew 1 Looming; Lemmings Imnom 1 M3 0 0 1 3% Copper are 11 2,316 11 171 22 2,487 Lead/zinc 22 3,093 13 229 35 3,322 Go1d-1ode & PL 29 2,080 181 994 210 3,074 Si1ver ores 24 1,990 58 357 82 2,347 Coba1t 0 0 1 3 1 3 Mo1ybdenum 2 1,297 2 268 4 1,565 Tungsten 2 20 4 110 6 130 Uran-vanad 2 44 9 43 11 87 Uranium 23 1,127 25 177 48 1,304 Metai ores 1 8 0 0 1 8 Antimony 0 0 1 2 1 2 P1atinum GRP 0 0 1 19 1 19 0i1 sha1e 3 174 6 68 9 242 Limestone-0M 1 15 1 22 2 37 Marb1e-DM 1 29 0 0 1 29 S1ate (DM) 1 5 0 0 1 5 Limestone—CB 75 1,789 28 358 103 2,147 Marb1e (CB) 7 88 0 0 7 88 Sandst (CB) 2 29 0 0 2 29 C1ay (Fire) 5 74 1 3 6 77 C1ay (Comm) 1 8 2 10 3 18 F1uorspar 3 63 1 3 4 66 Pot, Soda & Bor 1 397 0 0 1 397 Boron minera1 1 243 0 0 1 243 Potash 4 1,447 2 34 6 1,481 Trona 2 1,426 0 0 2 1,426 Sodium comp 3 1,804 0 0 3 1,804 Phosphate RK 1 117 0 0 1 117 Sa1t rock 13 1,947 1 142 14 2,089 Gypsum 9 371 1 12 10 383 Ta1c—soap & py 5 140 2 2 7 142 Nonmeta] min 2 33 1 4 3 37 Gemstones 0 0 2 6 2 6 Gi1sonite 12 69 5 23 17 92 Per1ite 0 0 1 3 1 3 Sa1t (evap) 1 217 0 0 1 217 Lime 3 958 0 0 3 958 Tota1 273 23,721 360 3,063 633 26,784 Taken from the Mine Safety and Hea1th Administration, June 26, 1984. Table A—5. Nonuranium mines that submitted individual radiation exposure records to MSHA, 1979—1983 Recorded number of employees Years Mine and company name 1979 1980 1981 1982 1983 Climax Molybdenum* 3,196 2,264 1,747 1,889 1,915 Amax Warm Springs Phosphate 24 22 22 23 23 Cominco American Crowell Fluorspar 10 _—— ___ ___ ___ J. l. Crowell Pine Creek Tungsten 299 319 260 ——— 89 Union Carbide Henderson Molybdenum* ——— 1,534 876 1,429 1,462 Amax Emperius ——— 13 15 ——— __- Chevron Bulldog Mt. Project -—— 147 —-- ___ ___ Homestake Ontario ——— ——_ 232 ___ ___ Noranda Stanley ___ ___ 11 ___ ___ Equity Gold Inc. Leadville Unit ——— ___ ___ 95 ___ Asarco Revenue - Virginius ——- ——- ___ 7 ___ Ranchers TOTAL 3,529 4,299 3,163 3,443 3,489 *Climax and Henderson mine exposures ran about 90 percent zeros in 1983. --- = no data submitted Taken from the Mine Safety and Health Administration, August 3, 1984 [97]. 114 SIT Table A—6. Concentrations of, and exposure to, radon daughters in nonuranium minesa Average Annual Country Year potential potential No. of No. of alpha energy alpha energy miners miners concentrations exposure /mines exceeding (WL) (HLM) 4 NLM Finland 1972-1974 0.2-0.4 -- 1,300/23 -—— 1975-1977 -- 0.38 1,370/16 0 Italy 1975 0.01-0.6 -— 2,500/16 Approx. 75 Norway 1972 0.07 0.64 1,870/33 ——- 1980 0.05 0.45 1,380/23 -—- Poland 1970 Copper 1—2 -—— --- ——— Iron 1 --— ——- ——- Pyrite 4 ——— -- -—- Phosphate 0.8 —- —-— -—— Zinc and lead 0.9 ——- --— --— Baryte 0.2 --- ——- —- Coal 0.1 -—- -- --- South Africa 1973 ——— 1.7 320,000 ——— Sweden 1970 ——— 4.8 4,800/5 2000 1974 ——- 2.1 4,600/50 360 1975 --- 1.9 5,300/45 270 1976 --- 1.7 5,300/46 225 1977 --— 1.6 5,200/45 475 1978 ——— 0.9 5,300/47 270 1979 --- 0.7 4,400/35 0 1980 ——— 0.7 4,400/35 0 Uni ted Kingdom 1968 0.01b —— 220,000/420 —— 1976 -- 2—3c 2,000/80 560 National coal 1981 ——- .12 185,200 --- Private coal 1981 --- 0.24 1,500 --- Other than coal 1981 --— 2.60 2,346/108 94 United States 1975 0.31 -—— --— —-— 1976 0.22 —- ——— --- 1977 0.12 -—— /163 —-— aIf not otherwise noted, the mines are iron, zinc, lead, copper, or gold mines. This value is called "typical" for large nationalized coal mines. cBased on measurements in about 80 percent of all noncoal mines. -—— : data not available From Ionizing R§d1§§19n3 ' i . United Nations Scientific Committee on the Effects of Atomic Radiation, N.Y., 1982, p. 198 [82]. Table A-7. Average exposures (WLM) during 1978 to United States uranium miners Job category All Full time Number av WLM Number av WLM Production 3,967 1.20 1,744 1.74 Maintenance 763 0.85 471 0.97 Service 1,759 0.81 626 1.14 Salaried 1,015 0.97 585 1.10 Total 7,504 1.03 3,426 1.45 *The first two columns refer to all miners who worked underground during the year and the last two refer to those who worked underground at least 1,500 hours. Taken from Radon Daughter Exposure to Uranium Miners by B.L. Cohen, pp. 286—291, In: Radiation Hazards in Mining, M. Gomez, ed. 1981 [89]. Table A—8. United States uranium miner exposures Total Average Miners having exposure in employment exposure indicated intervals, percentage 0—1 WLM 1—2 WLM 2—3 WL, 3—4 WLM 4 WLM 3,344 1.07 WLM 56.5 23.5 12.4 6.1 1.4 From Occupational Exposure to Ionizing Radiation in the United States: A Comprehensive Summary for the Year 1975 by J.R. Cook and D.R. Nelson, EPA #520/4—80-001, November 1980, p. D—12 [84]. 116 LII Tab1e A-9. Cumu1ative frequency distribution of annua1 exposures to radon progeny of persons who worked underground 1,500 hours or morea, United States uranium miners Cumulative oercentaae bv years Annua1 1973 1974 1975 1976 1977 1978 1979 7-year Exposure (N:699)b (N:1,216) (N:1,587) (N:2,052) (N:3,158) (N:3,426) (N=3,421) Average $1.0 HLM 39.5 33.8 46.6 41.9 44.1 41.8 37.5 40.7 32.0 WLM 68.5 65.5 75.5 68.7 72.0 74.5 69.6 70.6 53.0 NLM 88.7 88.4 91.4 89.1 92.7 92.1 91.5 90.6 $4.0 WLM 99.0 98.5 98.9 99.8 99.9 99.1 99.8 99.3 35.0 NLM 100.0 99.9 99.6 99.9 100.0 99.6 100.0 99.9 $6.0 WLM — 100.0 100.0 100.0 — 98.8 — — 3Data provided by L.w. Swent (1981). Since this tabu1ation inc1udes on1y those emp1oyees who worked underground 1,500 hours or more, dup1ications are un1ike1y. bN is the number of emp1oyees inc1uded in the report; the number of underground uranium mine operators providing data ranged from 32 in 1974 and 1975 to 71 in 1979. From Radiation Monitoring Priorities for Uranium Miners by K.J. Schiager and J.A. Johnson, p. 738-745, In: Radiation Hazard in Minino, M. Gomez, ed. 1981 [89]. BIT Tab1e A—10. Exposure of United States underground uranium miners to radon daughters in 1979 as reported by 71 underground uranium mine operations, for a11 persons assigned to work underground in 1979a,b A 1 n i n w k n r r n ' 1 7 0—1.0 1.01—2.0 2.01-3.0 3.01—4.0 4.01—5.0 5.01—6.0 Over 6.0 WLM NLM WLM WLM HLM WLM WLM Tgtgl Productionc— No. Persons 2,938 1,082 621 247 3 0 0 4,891 — % 60.0 22.1 12.7 5.1 0.1 0.0 0.0 100.0 Maintenanced- No. Persons 994 187 53 20 3 0 0 1,257 - % 79.1 14.9 4.2 1.6 0.2 0.0 0.0 100.0 Servicee - No. Persons 1,651 330 128 27 0 0 0 2,136 — % 77.3 15.4 6.0 1.3 0.0 0.0 0.0 100.0 Sa1ariedf — No. Persons 1,032 284 98 8 o o 0 1,422 — % 72.5 20.0 6.9 0.6 0.0 0.0 0.0 100.0 Tota1 - No. Persons 6,615 1,883 900 302 6 0 0 9,706 - % 68.1 19.4 9.3 3.1 0.1 0.0 0.0 100.0 aThere is a possibi1ity that persons may have worked for more than one operator in 1979 and, therefore, have been reported more than once in the above tabu1ation. The January 1, 1980 issue of "Statistica1 Data of the Uranium Industry" of the Grand Junction office of the U.S. Department of Energy shows average emp10yment in U.S. underground uranium mines in 1979 to be 5,706 persons. The DOE figures, however, do not inc1ude technica1 or supervisory persons who work underground. bExposures reported in this survey are based on more than 130,000 determinations of radon daughter concentrations. CProduction inc1udes production and deve1opment miners. Maintenance inc1udes mechanics and e1ectricians. eService inc1udes motormen, hau1age crews, drift repairmen, station tenders, skip tenders, etc. Sa1aried inc1udes engineers, supervisors, geo1ogists and venti1ation personne1. In mines where production emp1oyees a1so perform maintenance, service and supervisory duties., such emp1oyees were c1assified as production workers. Taken from A Comparison of Radon Daughter Exposures Ca1cu1ated for U.S. Underground Uranium Miners Based on MSHA and Company Records by H.E. Cooper, pp. 292—295, In: Radiation Hazards in Mining, M. Gomez, ed. 1981 [89]. 611 Tab1e A—11. Exposure of U.S. underground uranium miners to radon daughters in 1979 as reported by 71 underground uranium mine operations, for persons who worked underground 1,500 hours or more in 19793'b Persons who worked underuround 1500 hours orgmpre in 1979 0—1.0 1.01—2.0 2.01-3.0 3.01—4.0 4.01—5.0 5.01—6.0 Over 6.0 WLM WLM WLM NLM HLM HLM WLM Igfigl Productionc -No. Persons 348 609 517 234 3 0 0 1,711 — % 20.3 35.6 30.2 13.7 0.2 0.0 0.0 100.0 Maintenanced -No. Persons 283 135 46 21 3 0 0 488 — % 58.0 27.7 9.4 4.3 0.6 0.0 0.0 100.0 Servicee —No. Persons 401 182 112 23 0 0 0 718 — % 55.9 25.3 15.6 3.2 0.0 0.0 0.0 100.0 Sa1ariedf —No. Persons 253 m 75 5 o o o 504 - % 50.2 33.9 14.9 1.0 0.0 0.0 0.0 100.0 Tota1 —No. Persons 1,285 1,097 750 283 6 0 0 3,421 — % 37.5 32.1 21.9 8.3 0.2 0.0 0.0 100.0 aNo dup1ications of emp1oyees are possib1e in this tabu1ation because no emp1oyee was counted who worked underground 1ess than 1,500 hours (75 percent of a norma1 year of about 2,000 hours). bOperators that reported their data for inc1usion in this survey are: The Anaconda Company, At1as Minera1s, Cobb Nuc1ear Corporation, Cotter Corporation, Exxon Minera1s Company, U.S.A., Gu1f Minera1 Resources Company, Kerr—McGee Corporation, M&M Mining Company, Pathfinder Mines Corporation, Ranchers Exp1oration and Deve1opment Corporation, Ray Hi11iams Mining Company, Reserve 0i1 & Minera1s Corporation, Rio A1gom Corporation, Sohio Natura1 Resources Company, Todi1to Exp1oration & Deve1opment Corporation, Union Carbide Corporation, United Nuc1ear Corporation, United Nuc1ear—Homestake Partners, and Western Nuc1ear, Inc. The Co1orado Bureau of Mines furnished the data for 45 sma11 operators in Co1orado. In cases where corporations had wide1y separated operations under different managers, each was considered a separate operation. CProduction inc1udes production and deve1opment miners. In mines where production emp1oyees a1so perform maintenance, service and supervisory duties, such emp1oyees were c1assified as production workers. dMaintenance inc1udes mechanics and e1ectricians. eService inc1udes motormen, hau1age crews, drift repairmen, station tenders, skip tenders, etc. fSa1aried inc1udes engineers, supervisors, geo1ogists and venti1ation personne1. Taken from A Comparison of Radon Daughter Exposures Ca1cu1ated for U.S. Underground Uranium Miners Based on MSHA and Company Records by N.E. Cooper, pp. 292—295, In: Radiation Hazards in Minina, M. Gomez, ed. 1981 [89]. uranium mines in the United States, this figure is somewhat misleading. These workers can receive high exposures, and because they only work for short periods of time, their annual average exposure is low. The average exposure for those miners working full time, that is over 1,500 hours underground, was higher; 1.45 WLM in 1978. Underground mining exposure records were placed into four general job categories by the AIF, i.e., production, maintenance, service, and salaried. As a grOup, the production workers who worked more than 1,500 hours underground should have higher exposures than the remaining uranium mining work force. In 1978, the average exposure of these workers was 1.74 WLM (see Table A—7) and in 1979 their average exposure was approximately 1.88 WLM [57]. In contrast, in 1979 and 1980, MSHA inspectors recorded average radon progeny WL concentrations for underground uranium mining production workers of 0.30 WL or higher, which means that some of these workers could receive 4 WLM or more per year. Cooper estimated that the average annual exposure of full-time underground production workers was about 2.9 WLM during 1979 [57]. The number of workers that receive these high exposure levels may be small; AIF reported that among full-time underground uranium miners in 1979, only 3 out of 1,711 production workers and 3 out of 488 maintenance workers received more than 4 WLM annually (see Tables A—10 and A—11). Overall, most uranium mine workers' (including those workers who spend only part of their time underground) exposure is well below the standard of 4 WLM and on the average may be about 1 WLM [82], (see Table A—7). A relatively small number of workers, primarily full-time underground production and maintenance workers, have exposures above the 4 WLM standard (see Tables A-9 through A—11). The most recent available data, for 1982, showed that only 2 underground employees (0.1 percent) received radon progeny exposures of 4 0—5.0 WLM and 44 employees (1.6 percent) received exposures of 3.0—4.0 WLM. [58]. It should be possible to lower radon progeny exposure levels for this relatively small number of miners. b. Miners Outside the United States The exposure of underground uranium miners depends on the quality of the uranium ore body and the ventilation rate. In other countries, (excepting Canada) the uranium ore is frequently of a lower grade than the ore in the United States, so with good ventilation techniques, the foreign uranium miners should receive lower exposures than the miners in the United States. Recent figures for radiation exposure in underground uranium mines in Canada, France, India, Argentina, and China have been published in the literature (see Table A—3) [82]. The underground uranium miners of Canada had an average annual exposure to radon progeny of 0.74 WLM in 1978. In 1980, the median exposure for miners in three underground mines in Saskatchewan was below 0.6 WLM and only about three workers in one mine were exposed 120 to 3-4 WLM. In addition, some of these miners had substantial gamma exposure. In the Cluff mine, gamma exposures were as high as 3.5 rem and above, and in the Eldorado and Cluff mines many workers (approximately 60) were exposed to 1—3 rem [90]. The French uranium miners had average annual radon progeny exposures of 2.0 WLM and 1.4 WLM in 1978 and 1979, respectively [82]. In 1975, the median radon progeny exposure was below 0.10 WL, yet as many as 5.35 percent of the workers were exposed to 0.30 to 0.80 WL, potentially receiving more than 4 WLM annually (see Table A—12) [91]. In 1975, there was also a record of gamma exposure in French underground uranium mines. The mean annual dose was 0.49 rem, but some miners received much higher doses; 9.16 percent received 1.0—1.5 rem, 5.3 percent received 1.5-2.5 rem and 0.65 percent received 2.5-3.0 rem [91]. In the underground uranium mines in France, gamma exposure may constitute a major part of the total radiation. There is limited information available concerning typical radon progeny exposures in underground uranium mines in India, Argentina, and China [82]. For the mines in India, figures for potential exposure are given by job category. In 1979, the drilling crew received an estimate of 2.6 WLM of potential alpha energy exposure, the mucking crew about 2.1 WLM, and "others" about 1.7 WLM (see Table A—13) [82]. In Argentina, the average annual radon progeny exposure was about 2.4 WLM during 1980. Nonuranium Miners a. Hard Rock Miners in the United States Some of the highest radon progeny exposures are found in the iron, zinc, fluorspar, and bauxite mines (Table A—1). In 1975, iron miners were exposed to 0 14—0 90 WL, zinc miners to 0.07—1.40 WL, fluorspar miners to 0.30—2.20 WL and bauxite miners to 0.07—1.40 WL [83,84]. If these readings are typical, some hard rock miners in the United States, especially those in fluorspar mines, could have radon progeny exposures much higher than 4 WLM. However, recent data submitted by U.S. metal and non metal mining companies to MSHA suggests that no more than 450 individuals are occasionally exposed to 0.3 WL (Table A—5). During 1983, only 4 companies, 2 molybdenum, 1 phosphate and 1 tungsten, submitted individual exposure records for their employees to MSHA. It is possible that the mining companies failed to report additional employees who received radon expOSures, but this is the only data available. From this data, one concludes that, except for a few molybdeum, phosphate, and tungsten mines, radon progeny exposure is not a problem in U.S. hard rock mines. Thus, in general, hard rock mines should be able to meet an annual radon progeny standard below 4 WLM. 121 ZZT Tab1e A—12. Frequency distribution of radon exposures among French uranium miners (underground workers), 1971-1975 Mean “Exposure range (fraction of MAC)a Annua1 Exposure Year (0.10 0.11-0.20 0.21—0.30 0.31—0.40 0.41—0.50 0.51-0.60 0.61—0.80 0.81—1.00 >1.00 (ML) Percentage of workers 1971 36.08 22.39 19.90 13.12 6.22 2.14 0.15 ——— -—— 0.18 1972 37.30 22.55 21.13 12.27 4.36 2.24 0.15 ——— ——— 0.17 1973 37.70 19.32 19.43 14.40 7.72 1.43 -—— ——- —-— 0.18 1974 43.38 26.89 21.46 6.21 1.35 0.71 ——— ——— —-— 0.13 1975 53.91 24.71 16.03 4.58 0.66 0.11 —-— --— -—— 0.11 aFor each worker the annua1 exposure is represented by the mean annua1 air concentration and is expressed as a fraction of the maximum annua1 concentration (MAC). Given the administrative arrangements and the effective state of \ equi1ibrium between radon and its daughters, the MAC is practica11y equiva1ent to 1 HL. Taken from Source; and Effect; of Ionizing Radiation, United Nations Scientific Committee on the Effects of Atomic Radiation, N.Y., p. 267, 1977 [91]. Table A—13. Estimated potential alpha energy exposure of different categories of mine workers in the Jadugkda underground mines, India Estimated potential alpha energy exposure (WLM) Year Drilling crew Mucking crew Others 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 dowNo—l—LONNAAA hNU’lON-h—KNAQONN ~4_:~Jo>owc>aabwLibstkib> MO—tU'lN—JOO—ANACDNNb b¢bdbbb§§LLbbbb l+|+l+|+|+|+|+|+|+|+l+l+|+l+|+ AOO-P-OOOOO—ANN—A—AN bbb$§bbbbth9hb waammmo—nmmmwmwm QLQGbibbbbbbb; |+|+l+l+l+|+|+|+I+l+|+|+l+|+l+ COO—aggrio—LAAQANA...‘ kabeb$Lbbb§bb l+ I|+|+|+l+|+l+l+l+l+l+|+l+|+ o 004—1049044404 .1 ‘1 OJ -—- data not available Taken from lonizingfiRadiation: Sources and Biological Effects. United Nations Scientific Committee on the Effects of Atomic Radiation, N.Y., 1982, p. 199 [82]. b. ‘Hard Rock Miners Outside the United States Radon progeny exposure levels have been measured in nonuranium mines in Finland, Italy, Norway, South Africa, Sweden, the United Kingdom and Poland (Tables A—14 to A—16) [82,81]. The most recent figures for all of these countries show annual average radon progeny exposures of 2.6 WLM or less. However, in many of these countries the average potential alpha energy concentrations exceed 0.3 WL, suggesting that individual miners may be exposed to more than 4 WLM per year (if they work full time during the year). Nonuranium miners (especially iron, zinc, lead, copper, or gold miners) in Italy, Poland, South Africa, and Great Britain may be exposed to more than 4 WLM annually [82]. in the United Kingdom, 4 percent of the noncoal miners were exposed to 4 WLM or more, however, many of the miners did not work full 8—hour shifts. If the underground noncoal miners in the United Kingdom worked full 8—hour shifts, as many as 20 percent of the workers could be exposed above 4 WLM/yr [81]. Recent reports for five Chinese tin mines showed radon progeny levels of 0.67 to 1.73 WL during 1978 [40]. 123 VZI Tab1e A—14. Distribution of radon-daughter exposure in nonuranium mines in various countries Weighted average Radon—daughter concentration range (NL) annua1 exposurea Country Year <0.1 0.1—0.3 0.3—1.0 >1.0 ALL (NLM) (Number and, in parentheses, percentage of miners or mines) Fin1and Miners 1973 469(35) 246(18) 247(19) 369(28) 1,331 8.8 1974 898(68) 310(23) 119(9) 0 1,327 1.7 Mines 1973 8(36) 4(18) 4(18) 6(28) 22 ——— 1974 13(65) 5(25) 2(10) 0 20 ——- Ita1y Mines 1973 8(50) 4(25) 4(25) 0 16 -- Norway Miners 1972 1.608(86) 264(14) 0 0 1,872 0 9 Mines 1972 20(83) 4(17) 0 0 24 ——— South Africa Miners 1973 227,000(71) 69,000(21) 21,000(7) 3.000(1) 320,000 1.7 Sweden Miners 1970 1,110(22) 1,560(33) 2.000(42) 130(3) 4,800 4.8 1974 1,860(40) 2,390(52) 360(8) 0 4,610 2.1 1976 2,730(51) 2,345(44) 225(4) 0 5,300 1.7 Mines 1970 25(45) 8(15) 18(33) 4(7) 55 ——— 1974 28(56) 14(28) 8(16) 0 50 --- 1976 29(63) 12(26) 5(11) 0 46 —-— United Kingdom Miners 1973 1,073(60) 49(3) 223(12) 443(25) 1,788 4.2 1975 3.4 Mines 1973 25(61) 3(7) 9(22) 4(10) 41 3The weighted annua1 average exposures are ca1cu1ated by mu1tip1ying the number of miners in each group by the mean va1ues of the radon concentration (0.05, 0.2, 0.65 or 2 ML) and by 12 months, obtaining the sum of the products and dividing by the tota1 number of miners. The United Kingdom miners represent 70 percent of a11 noncoa1 miners and the United Kingdom mines represent 41 percent of a11 noncoa1 mines. Taken from Sources and Effects Qf Lgnizjng Radiation, United Nations Scientific Committee on the Effects of Atomic Radiation, N.Y., p. 254, 1977 [91]. Table A-15. Employment and expOSure in British mines Miners WLM in Collective Exposure Type of mine employed underground a year man WLM/y National coal 185,200 0.12 ——— Private coal 1,500 0.24 2.26 104 Other than coal 2,346 2.60 6.10 103 Table A-16. Weighted exposures* of noncoal miners in 1981 and 1976 Exposure Number of men exposed in year % of men exposed in year WLM in a year 1981 1976 1981 1976 0 to 1 938 986 40 49 1 to 4 1,314 454 56 23 4 and more 94 564 4 28 All 2,346 2,004 100 100 *Time—weighted full—shift exposures Tables A—15 and A-16 from Radon in British Mines — a review by M.C. O'Riordan, S. Rae and G.H. Thomas, pp. 74—81, In: Radiation Hazards in Mining, M. Gomez, ed. 1981 [89]. 125 APPENDIX B CURRENT METHODS OF REGULATION AND CONTROL OF RADIATION EXPOSURES IN UNDERGROUND MINES A. Engineering Controls Table 8—1 lists information about mining radiation control methods, including ventilation, sealants, bulkheads, backfilling, wet drilling, air cleaning, and separate air supplies. It may be most effective to combine some of these techniques, e.g., to use positive pressure ventilation in combination with procedures to decrease the volume of the mine air needing ventilation, such as bulkheads or backfilling. Bulkheads could be made more secure against radon gas leaks by maintaining a slight negative pressure behind the bulkhead and painting sealant on nearby exposed rock. Finally, most of the techniques described in Table B—1 and in this chapter will decrease inhalation exposure to alpha radiation from the decay products of radon and thoron gases, but won't affect gamma radiation levels. 1. Mechanical Ventilation Mechanical ventilation is the primary and most successful technique currently in use for reducing exposure to radon decay products. In uranium mines in the United States, during the early 1950's before mechanical ventilation became prevalent, average measurements of 2—200 WL of radon decay products were common [11]. In contrast, during 1979 and 1980, the highest average working level for radon progeny recorded by MSHA was 0.46 WL (Table B—2). Thus, there has been a great decrease in exposure to radon decay products in uranium mines primarily due to improvement in ventilation. Sweden has also successfully reduced radon progeny levels in nonuranium mines with mechanical ventilation. The average annual exposure for the nonuranium miners of Sweden was 4.7 WLM in 1970, due to ventilation improvements, and decreased to 0.7 WLM in 1980 [95]. In the case of uranium miners in the United States, it is not clear whether there could be significant further decreases in exposures to radon decay products with ventilation improvements alone. These few mines may need to use other techniques, besides dilution ventilation, to reduce miners' exposure to radon progeny (Table B—1). 2. Other Dust Control Methods Spraying water and delaying blasting until the end of shifts are two other dust control methods currently in use in most underground uranium mines. Most mines use these methods to control silica dust, but in uranium mines these methods can help control uranium ore dust. Drilling and blasting are two mining activities that generate high levels of uranium ore dust. Exposure to uranium ore dust alone may be carcinogenic, and high dust or smoke levels may modify the respiratory tract distribution of a miner's exposure to radon progeny (by increasing the proportion of radon progeny attached to respirable and nonrespirable size dust particles). In wet drilling, water sprays from the drill onto 126 Table B—1 Mining radiation control methods Type of Control radiation method Description Radon, Thoron Sealants Radon barrier coatings, made from water— Gases and based acrylic latex, water—based epoxies Progeny or other materials, painted on exposed rock surfaces. Coatings can reduce radon flow by 50 to 75 percent [92]. Advantages: particularly useful in limited areas, i.e. intake airways with high radon emanations, lunchrooms, shops, etc. [92]. Disadvantages: Too expensive to use throughout the mine. Radon, Thoron Bulkheads Bulkheads seal off worked—out stopes or Gases and inactive mine areas. Bulkhead effectiveness Progeny increased when used with sealants and a slight negative pressure behind the bulkhead. Bulkheads can be made from brattice cloth, urethane foam, gunite, timber, etc. [93,92]. Advantages: Cost—effective. Disadvantages: Bulkheads can leak if cracked, poorly sealed, or when barometric pressure decreases. (Continued) 127 Table B—1 Mining radiation control methods (Continued) Type of Control radiation method Description Radon, Thoron Gases and Progeny Radon, Thoron Gases and Progeny All Radiation Backfilling Air Cleaning Medical Removal Protection A common uranium mining practice is to fill worked—out areas with mine waste rock and uranium mill tailings. One study showed an approximately 85 percent reduction in radon entering the stope after backfilling [93]. Advantages: Reduces radon emanation, reduces ventilation requirements and provides ground support. Disadvantages: Uranium mill tailings can still release some radiation underground, perhaps including gamma radiation. Radon daughters are removed by an air cleaning apparatus, typically involving a filtering system. Advantages: Useful in limited areas where it is not feasible to install a large ventilation system [92]. Disadvantages: High operating costs, lack of a commercial equipment source and equipment reliability problems [92]. If a person approaches or exceeds the lifetime limit on exposure, they are trans— ferred to another job at a lower exposure level with retention of pay, if available, or are removed from work at full pay if another job is not available. Advantages: Protects individual miners against high cumulative exposures. Disadvantages: Spreads exposure over a larger number of people. This system works best when used with a reliable bioassay for exposure, which is not available in the case of radon gas or progeny. Medical removal may not be effective if intense, short—term exposure to inhaled alpha radiation is more hazardous than cumulative radiation exposure. (Continued) 128 Table B—1 Mining radiation control methods (Continued) Type of Control radiation method Description Uranium Ore Wet Drilling, The drills are equipped with automatic Dust hosing down water valves that turn the water and muck piles, compressed air on simultaneously. Uranium Ore Dust, Radon and Thoron Progeny Radon, Thoron Gases and Progeny Radon, Thoron Gases and Progeny other uses of water to control dust Blasting at the End of Shifts Minimizing Fan Shutdown Ventilation — Blowing/ Positive Pressure (These techniques have been used in mines since the 1930's.) Advantages: The water cuts down on radioactive uranium ore dust. Disadvantages: Difficult to set up in areas where water is scarce. The miners using the drill get wet. Dynamite blasting at the end of shift, instead of throughout the day, reduces exposure to dust and smoke. Also, radon gas levels tend to be high immediately after blasting [93]. Advantages: Most miners have less exposure to dust and smoke particles and thus less radiation exposure. Disadvantages: Extra production schedule planning is necessary. This involves the use of fan maintenance, backup electrical systems, and spare fans to minimize fan shutdowns during working hours. Positive pressure at the rock surface is a barrier to radon flow. One drawback is that high positive pressure in one area may force the radon into nearby low pressure areas [92,93]. (Continued) 129 Table B—1 Mining radiation control methods (continued) Type of Control radiation method Description Radon, Thoron Ventilation Gases and —Exhaust Progeny Ventilation -Push—pu|l Radon Filter Progeny Respirators and Thoron Progeny Exhaust ventilation removes radon, thoron and daughters, as well as diesel fumes, but it also increases the emission of radon from the surrounding rock by creating a negative pressure. Positive pressure ventilation is shut down during times when the mine is inactive, creating a temporary negative pressure. This results in energy savings during the shut down periods. The best ventilation method to use depends on the mine topography and production schedule. Ventilation methods may be most effective when used in combination with techniques that cut down on the area needing ventilation, such as bulkheads and backfilling [92,93]. The filter respirator covers the miner's mouth and nose and filters the mine air through fiber filters [94]. Advantages: As a temporary short—term protective measure, the half—mask respirator affords approximately greater than 90 percent efficiency in reduction of miner's exposure to radon daughters attached to dusts, fumes, and mists. Disadvantages: The respirators may hinder vision, be warm to use under some working conditions, add significant resistance to the miner's breathing and require careful maintenance to assure their continued effectiveness. Filter respirators must be carefully fitted to each wearer, using quantitative respirator fit tests. Only MSHA/NlOSH—certified respirators shall be used. (Continued) 130 Table B—1 Mining radiation control methods (continued) Type of Control radiation method Description Radon, Thoron Supplied— Gases and air Progeny respirators Uranium Ore Robots or other Dusts, Radon, mechanization Thoron Gases and Progeny; maybe gamma The respirator is supplied with respirable breathing air from a central air supply. Advantages: As a temporary short—term protective measure, the supplied—air respirator affords a high degree of pro— tection against all mine air contaminants. Disadvantages: The supplied—air respirator may hinder movement of the miner and the trailing air hose may get caught or tangled up in the mining environment. Respirators require careful maintenance to assure their continued effectiveness. Only MSHA/NlOSH—certified respirators shall be used. The jobs with the highest dust levels could be mechanized further, thus minimizing the time during which the miner receives exposure. High dust exposure jobs include blasting, drilling, filling ore cars, putting in track, dumping waste, etc. 131 the rock while the drill operates, thus decreasing dust levels. Miners also wet down muck piles and the walls of some tunnels to control dust. Since the 1930's these two techniques have been used in some mines. Blasting increases uranium ore dust and radon gas levels remain high for about an hour afterwards [93]. Delaying blasting until the end of the work—shift removes the miner from an area with high dust and radon gas levels, and allows the ventilation system to reduce these levels before the miner returns to work. 3. Additional Control Methods Air cleaning equipment, filter respirators, and separate air supplies are seldom used in the underground mining environment. An air cleaning apparatus can remove dust, but it is expensive compared to traditional ventilation methods and is most useful in circumscribed areas [92]. Filter respirators and supplied-air respirators are difficult to use in the mining environment and their use should be limited to emergency conditions, such as temporary excursions of the radon progeny concentrations above 1 WL. Respirators tend to restrict movement and vision, may be too warm to wear, have significant breathing resistance, and require careful maintenance and fitting to assure their continued effectiveness. Only MSHA/NlOSH—certified respirators shall be used. Another radon progeny control method is robotics or increased automation. Techniques, such as robotics, that minimize the time the miner spends in the high exposure areas of the mine and in activities such as drilling, blasting, or loading ore, will decrease the miner‘s radiation exposure. Although, at present, robotics has a limited place in the mines, it may be possible in the future to further automate the uranium ore mining process. Administrative Controls 1. Medical Removal Protection One type of administrative control is a medical removal protection (MRP) program. Under this program, when an individual's exposure approaches or exceeds a certain limit, the person is reassigned to an area with a lower exposure level. The MRP program has been very effective in reducing exposure in the (noncarcinogenic) lead industries [96]. In this case, blood lead levels could be used as a method to biologically monitor a worker's lead exposure. However, MRP has certain drawbacks when used as an administrative control for exposure to a known human carcinogen such as radon progeny in underground uranium mines. First, according to our current knowledge of radiation carcinogenesis, it is prudent public health policy to presume that there is no threshold for radon—progeny—induced cancer, and thus no exposure can be assumed to be safe. Therefore, the high exposure individuals who are removed from the job are protected against further radon progeny risk, but the radon progeny exposure (and risk) is spread out over a larger population of workers. Second, at this time, there is no good biological monitoring method for radon progeny exposure because the primary health effect is a carcinogenic, rather than a toxicologic, response. Routine, periodic 132 sputum cytological examinations and chest X—rays are not effective screening tests for the detection of early reversible signs of lung cancer, and cancer itself may only appear after years of exposure. Finally, respirators (as they are presently designed) are very difficult to use in the underground mining environment. 2. Alarm Systems Another type of administrative control involves the use of alarm systems. This method has been fairly effective in coal mines where continuous monitors for methane gas have been tied to alarm systems. Reliable continuous monitors for radon progeny are now technically feasible (see [89]) and could be connected to alarm systems, as well as the control center for the ventilation system. The person who controls the ventilation could increase air movement in mine areas with high radon progeny levels. Also, the continuous monitors might be useful for enforcement purposes, because the MSHA inspector would have a record of excessive radon progeny measurements levels since the last inspection. For recordkeeping and enforcement purposes, the use of data from continuous alarm—monitors would depend heavily on the reliability and validity of these devices, as well as their durability and security from tampering in the mine environment. 3. Contract Mining Many underground uranium miners, especially those that drill, blast, and move ore, are given incentive bonuses for the volume of ore removed. Such a system encourages high productivity from the workers, but any time they spend on safety measures means less time to spend mining ore. The contract mining system also encourages miners to work overtime, thus increasing their cumulative internal and external radiation exposures. In addition, some miners, especially before the reduced demand for uranium, went from mine to mine working uranium ore one month and gold the next, getting radon progeny exposures in both locations. This mobility of the work force makes it harder to monitor and track the miners' total radiation exposure, making it more likely that a miner could receive cumulative exposures in excess of current and future standards. One type of administrative control is to modify the contract mining system so that workers would have more incentive to protect their own health on the job. This issue needs further study and discussion, including input from the mining industries, unions, and contract miners. 133 GLOSSARY Absorbed Dose: The amount of energy absorbed by ionizing radiation per unit mass. Absorbed doses are expressed in units of rads or grays, or in prefixed forms of these units such as millirad (mrad, 1O—3 rad), microrad (urad, 10—6 rad), etc. The gray (Gy) is equal to 1 joule per kilogram (1 J/kg). The rad is equal to 6.24 x 106 MeV per gram, or 100 ergs per gram. One gray = 100 53g. Additive Relative Risk Model: The relative risk from the combined exposure to radon progeny and smoking equals the sum of the risks from each exposure considered separately. One example of an additive linear relative risk model is: R = 1+B, WLM+82 PKS where: R = relative risk 81 = excess relative risk per unit of radon progency exposure 32 = excess relative risk per unit of cigarette smoke exposure WLM working level months PKS = cigarettes (in packs) Association: Two variables are*associated if one is more (or less) common in the presence of the second. Attributable (or Absolute) Risk: The rate of disease attributable to exposure+. For radon progeny exposure, it can be expressed as the arithmetic difference in risk between exposed and unexposed groups, in lung cancer deaths per year per WLM. One formula frequently used to calculate the attributable risk from radon progeny is: AR = OBS — EXP x 106 PYR X WLM Where: OBS observed deaths in the cohort EXP = expected deaths in the comparison group PYR = person-years at risk WLM = average working level months of radon progeny exposure 106: 1 million AR = attributable risk Bias: An error in the measure of the association between two variables.** Case—Control Study: Selection of study groups to be compared based on presence or absence of disease. Cohort Study: Selection of study groups to be compared based on presence or absence of exposure. 134 Confounding Bias: A potential attribute of data. In measuring an association between an exposure and a disease, a confounding factor is one that is associated with the exposure and independently is a cause of the disease. Confounding bias can be controlled if information on the confounding factor is present. Coulomb: The charge flowing past a point of a circuit in one second, when there is a current of one ampere in the circuit; also, the aggregate charge carried by 6 x 1018 electrons. Electron Volt: The change in potential energy of a particle having a charge equal to the electronic charge (1.60 x 10‘19 coulombs), moving through a potential difference of 1 volt. Half—Life: The time required for a radioactive substance to decay to one half of its initial activity. Follow-up Period: The length of time between a person entering an epidemiological study cohort and the present report (or the end of the study). Incidence Rate: The number of new cases of disease per unit of population per unit of time, e g., 3/1000/year. Interaction: The association of one factor (occupation) with disease modified by the effect of another factor (smoking). The measure of association can be the rate or odds ratio. This follows a nonmultiplicative model (may be additive). Ionizing Radiation: Any electromagnetic or particulate radiation capable of producing ions, directly or indirectly, in its passage through matter. Lagging Exposures: Lagging of the cumulative exposure assigned to a miner. Some authors consider that radon progeny exposures are "redundant" if they occur after lung cancer is induced. Some authors believe that cumulative exposures should be lagged by a certain number of years (5 or 10), to exclude redundant exposures occurring during these years. For example, Radford and St. Clair Renard [5] discounted the last 5 years of exposures from the cumulative total WLM assigned to each case of lung cancer in their analysis. Biologic Latent Period: The time between an increment of exposure and the increase in risk attributable to it.+ Epidemiologic Latent Period: The time between first exposure and death in those developing the disease during the study interval. Linear Hypothesis: The hypothesis that excess risk is proportional to dose. Matching: A procedure to reduce the biasing effect*of a confounding variable. A feature of selection to study groups. 135 Multiplicative Relative Risk Model: The relative risk from the combined exposure to radon progeny and smoking equals the product of the risks from each exposure considered separately. One example of a multiplicative linear relative risk model is: R = 1+B, WLM+32 PKS where: R = relative risk B1 = excess relative risk per unit of radon progency exposure 82 = excess relative risk per unit of cigarette smoke exposure WLM = working level months PKS = cigarettes (in packs) Person—Years (PY): A standard technique for handling variable follow—up periods; multiply the number of persons by the number of years of follow-up. Person-Years at Risk (PYR): In a lifetable analysis, the number of PY at risk of dying from disease, usually calculated from the time the miner enters the cohort until death or the end of follow—up. Some authors adjust the PYR for an assumed 10—year latent period for lung cancer by subtracting PYR accumulated during the first 10 years after a miner starts to work underground (see above, (Lagging)). Potential Alpha Energy Concentration (PAEC): May cause biological damage during the radioactive decay of radon or thoron gases and their progeny, is measured in units called Working Levels (see below). Proportional Mortality Ratio (PMR): The ratio of two mortality proportions, expressed as a percentage, often adjusted for age or time differences between the two groups being compared. Prospective: A study characteristic. Disease has not occurred in study groups at the start of a study. Units of Radioactivity: Curie and Becquerel 1 curie = 2.22 x 1012 disintegrations/minute 1 becquerel (Bq) = 1 d/sec 1 picocurie (pCi) = 2.22 d/minute Radioactive Decay: Disintegration of the nucleus of an unstable nuclide by spontaneous emission of charged particles, photons, or both. Radon (Rn) or Radon and its Progeny: Specifically refers to the "parent" noble gas (Rn-222), and its short-lived alpha-radiation—emitting radioactive decay products (”progeny" or "daughters”). Radon is a gas, the radon progeny are radioactive solids. Rate: The number of cases per unit of population. Rate Ratio: One rate divided by another rate with the same dimensions. A measure of association without a unit. 136 Relative Risk: The ratio of rates in exposed and nonexposed populations. One formula frequently used to calculate the relative risk for radon progeny exposure is: ERR = Lfisfip x (100 WLM) Where: ERR = excess relative risk OBS = observed deaths in the cohort EXP = expected deaths in the comparison group Rem and Sievert rem = rad x QF x modifying factors sievert = grays x QF x modifying factors 10 mSv = 1 rem Rads and rems are comparable (i.e., the quality factor (QF) = 1) when dealing with beta particles and gamma photons. The CF for alpha particles from inhaled radon progeny are generally considered to be in the range of 10 to 20. Retrospective: A study characteristig; Disease has already occurred in study groups at the start of a study. Standardization: A procedure to reducg*the biasing effect of a confounding variable. A feature of data analysis. Standardized Mortality Ratio (SMR): The ratio of mortality rates, expressed as percentage, usually*adjusted for age or time differences between the two groups being compared. Synergism: The combined action of two factors which is greater than the sum of the actions of each of them. Thoron: A radioactive gas (Rn-220), sometimes found in the presence of radon (Rn—222). Thoron progeny are the solid, short—lived, alpha radiation emitting decay products (progeny or daughters) of thoron gas. Working Level (WL): A standard measure of the alpha radiation energy in air. This energy can come from the radioactive decay of radon (Rn—222) and thoron (Rn—220) gases. The working level is defined as any combination of short—lived radon decay products per liter of air that will result in the emission of 1.3 x 105 million electron volts (MeV) of alpha energy. Working Level Month (WLM): A person exposed to 1 WL for 170 hours is said to have acquired an exposure of one Working Level Month. The Mine Safety and Health Administration defines a Working Level Month as a person's exposure to 1 WL for 173 hours. * Taken from Shapiro (1981) [1]. ** Taken from Monson (1980) [98]. + Taken from Thomas et al., (1985) [99]. 137 APPENDIX ll QUANTITATIVE RISK ASSESSMENT OF LUNG CANCER IN U.S. URANIUM MINERS by Richard W. Hornung, Dr. P.H. Theodore J. Meinhardt. Ph.D. National Institute for Occupational Safety and Health Centers for Disease Control Cincinnati, Ohio 45226 January 16, 1986 138 ACKNOWLEDGMENT The authors would like to acknowledge the advice and input of Dr. Dale Hattis of the Massachusetts Institute of Technology's Center for Policy Technology and Industrial Development, which was especially useful in considering exposure assessment and the characteristics of carcinogenesis. Appreciation is also acknowledged for the assistance of Robert Roscoe of NlOSH in providing updated mortality status for the study group. 139 I. INTRODUCTION A report evaluating epidemiologic studies of lung cancer in underground miners was recently sent to the Mine Safety and Health Administration (MSHA) by the National Institute for Occupational Safety and Health (NIOSH). That report concluded that prolonged exposure to radon progeny at the current standard of 4 WLM/year produced an elevated risk of death from lung cancer. It is the objective of this report to make quantitative risk estimates for various levels of cumulative exposure. ln addition, other factors influencing the exposure—risk relationship will be identified and quantified whenever possible. This report is based upon data collected from a cohort consisting of 3366 white underground uranium miners working in the Colorado Plateau (located within the states of Colorado, Utah, New Mexico and Arizona). The actual risk estimates were computed from data on 3346 members of the cohort. Ten original members were determined to have had no record of underground mining, four were non—white, and six had inadequate cigarette smoking information. Entry into the cohort was defined by race, sex, working at least one month in underground uranium mines, volunteering for at least one medical survey between 1950 and 1960, and providing social and occupational data of sufficient detail [Lundin et al. 1971]. NIOSH has now updated the mortality experience of the cohort through December 31, 1982. Lung Cancer mortality was defined as anyone assigned an International Classification of Disease (ICD) code of 162 or 163 (same designation “1 Sixth through Ninth Revisions). Previous analyses of this cohort reported by Waxweiler et al. [1981] and Whittemore and McMillan [1983] considered follow—up only through 1977. Table 1 presents a comparison of vital status of the cohort at the end of 1977 and 1982. 140 Table 1. Status of Data Base 1977 1982 Number Percent Number Percent Alive 2,388 71.4 2,132 63.7 Deceased 958 28.6 1,214 36.3 Lung Cancer 187 19.5 255 21.0 Other Causes 771 80.5 959 79.0 Total 3,346 100.0 3,346 100.0 141 ll. PROTOCOL FOR STATISTICAL ANALYSIS A. Type of Analysis Used Much of the epidemiologic work in the past regarding the analysis of mortality in occupational cohorts has involved modified life table analysis. This form of analysis has a strong appeal due to its familiarity and ease of interpretation. It is mathematically straight forward .since person—years at risk are simply divided into a number of strata and age-calendar year specific mortality rates from some reference population are applied to each. The U.S. population is often used as the reference population in such life table analyses. This expected Inortality is then compared to the observed mortality via a ratio defined as: 20.. SMRj = I I] EEij i where SMR = standardized mortality ratio for cause j Oij = the observed number of deaths for cause j in stratum i and Eij = the expected number of deaths for cause j in stratum i from reference population rates If the total number of observed deaths in all of the strata of interest is large and if the reference population is the appropriate comparison group, this would be the method of choice. No modeling would be needed in such a situation. However, after stratification by age, race, sex, calendar year, other confounders, and finally the exposure of interest, there are seldom enough observed deaths to make rates in these strata reliable. Another problem frequently encountered is a fundamental difference in certain etiologic characteristics between the study population and the reference population. For example, the study group may smoke at substantially different rates than the reference population. Often the occupational study group is "healthier" than the reference population due to selection criteria for employment (Enterline [1976]). This is usually referred to as the "healthy worker effect.” An alternative to use of the modified life table approach is some form of statistical modeling. Modeling to estimate health risks is necessary when conclusions must be drawn about risk in regions of the exposure-response relationship for which data are too sparse to estimate risk directly. The use of models also permits risk estimates to be simultaneously adjusted for confounders, such as age or co—carcinogenic exposures, as well as interactions between exposure and other risk factors. This flexibility is particularly important in making risk estimates at relatively low cumulative exposures when using the Colorado Plateau data. Most miners in this cohort were exposed to high levels of radon progeny (mean exposure = 834 WLM). Since primary interest in risk estimates is below 120 WLM based on current exposures, some type of statistical model is essential. 142 There have been a number of types of models suggested for examination of cause—specific mortality as a function of various risk factors. The two most popular types of models are the absolute risk model and the relative risk model. The absolute risk model can be written as: R(t;5) = Ro(t) + R(5.B) where R9t;;) is the incidence at age t for someone with risk factors ;, Ro(t) is the baseline or background incidence at age t and R(g, ) is the incremental incidence as a function of the risk factors 5, and coefficients which are estimated from the data. This form of risk model was not used in the risk assessment since it had been rejected due to poor fit to the U.S. uranium miner data by Lundin et al. [1979]. In contrast, the relative risk model generally takes the form: R(t;E) = R0(t) mm). This model assumes that excess risk is proportional to background incidence rates. Relative risk models have become increasingly popular in recent years and were found to provide good fits to the data from earlier follow—ups of the U.S. uranium miners cohort by Lundin et al. [1979] and Whittemore and McMillan [1983]. This type of model has been selected as the basic analytical method for this report. 8. The Proportional Hazards Model A relative risk model which is particularly well—suited to longitudinal mortality studies is one proposed by Cox [1972]. This model is commonly referred u) as the Cox proportional hazards model. A major advantage of this approach over the more common life table method is that it permits the use of internal comparison groups while controlling simultaneously for such confounders as cigarette smoking, age, and year of birth. In addition, time—dependent covariates such as cumulative exposure may be incorporated into the model. This is essential in any longitudinal study where follow—up and the exposure period overlap. Relative risk estimates are based on rate ratios similar to those produced in the modified life table analysis. That is, the Cox model operates in a dynamic framework by considering incidence rates over the entire period of follow—up. The Cox model can be expressed mathematically as: k(t;Z) = ko(t)exp(§5(t)) where k(t;z) for this study is the age—specific lung cancer mortality rate for a miner with exposure and other risk factors represented by a covariate vector 5. The underlying age—specific lung cancer mortality rate for the unexposed is represented byA0(t). The function expflgg) is generally used to model risk of death from the cause of interest which depends upon the risk factors 5 and the coefficients g which are estimated from the data. 143 c. Alternative Forms of the Risk Function Although the exponential or log—linear function exp(B;) is the usual choice of a model for risk, any positive function may be used as long as the risk function is equal to 1.0 when the coefficients are all equal to zero. The most common alternative ris$ functions are the linear (1 + 32) and the power function (exp( fi1nz) = z ). All three forms of risk functions were considered in modeling the U.S. uranium miners data. D. Results of Model Development 1. Identification of Confounders and/or Effect Modifiers Cumulative exposure as measured by total WLM for each miner was the primary exposure variable. Since cigarette smoking is known to have a strong effect upon the risk of lung cancer, cumulative smoking history as measured in pack-years was also included in the model. Another risk factor strongly associated with lung cancer mortality is age. This was tightly controlled by using age as the time dimension t in the model A(t;z). That is, the age at death of each lung cancer victim was recorded and all other miners alive and at risk were compared to him at that age. In this way, the cumulative exposure to radon daughters and pack—years of cigarettes were incorporated as time—dependent covariates by calculating their values at each age of death from lung cancer. This assures that proper age—adjusted comparisons were made throughout the period of follow—up. A number of other variables were examined in developing the appropriate risk model. A list of all potential risk factors considered for inclusion in the model are provided in Table 2. These variables were considered independently as potential confounders in a stepwise fashion (both backward and forward selection procedures) and also as potential effect modifiers by assessing their interaction with cumulative radon daughter exposure. An attempt was made to compare the fit of each of the three models during the model development stage of the analysis. However, it soon became apparent that the linear model did not fit well over the full range of radon daughter exposures and cumulative smoking levels. In fact, the iterative solution to the likelihood equations would not converge when using the linear model when both cumulative exposure and pack—years of smoking were both entered simultaneously (either as linear or linear—quadratic forms). The linear model could only be made to converge when the model was restricted to cumulative exposure below 600 WLM with no other covariates included. The restricted linear model resulted in a non—significant result in this exposure range and was subsequently eliminated from consideration. 0f the remaining two types of relative risk models (log-linear and power function), the covariates found to be most highly associated with lung cancer incidence rates were cumulative exposure (WLM), cumulative smoking (packs), and age at initial exposure (months). Table 3 illustrates the form and degree of fit as measured by the likelihood 144 Table 2. Regression Variables Considered in Development of Model Variable Units Median Range Cumulative Exposure Working Level 0.3—10,000+ Months (WLM) Average Exposure Rate WLM/month 10.3 0.03-998 Cumulative Cigarette Smoking* Packs 10,027 0.0—61,000 Smoking Rate Packs/day 0.64 0.0-3.5 Age at Initial Exposure Months 348.4 101—877 Calendar Year of Initial Year 1954 1908—1963 Exposure Birth Year Calendar year 1921 1877—1948 Height Short (<68 inches) Medium (68-70 inches) Tall (3 70 inches) Duration of employment Months underground 48.0 1—371 Years of Prior Hardrock Years 0.0 0—42 Mining** *20.4 percent never smoked. **62 percent had no prior hardrock mining. 145 Table 3. Comparison of Log—Linear and Power Functions Models Risk Factor Coefficient x2 P—value Log—linear Model Cumulative exposure (WLM) 0.897 125.4 (0.001 Cumulative cigarettes (packs) 0.063 44.6 (0 001 (WLM)2 —0.089 44.5 (0.001 (Packs)2 —0.002 10.5 0.001 Age at Initial Exposure 0.0022 7.9 0.005 (months) LIKELIHOOD RATIO x2 = 205.8 Power Function Model Ln(Cumulative Exposure+BGR) 0.713 135.3 (0.001 Ln(Cumulative Smoking+BGS) 0.295 35.3 (0.001 Age at Initial Exposure 0.0023 8.7 0.003 LIKELIHOOD RATIO X2 = 219.9 1BGR — background radon exposure = 0.2 WLM/year BGS = background cigarette smoking = 0.005 packs/day 146 ratio for these two models. The log—linear model required the addition of quadratic terms in cumulative exposure and cigarette smoking to provide an adequate fit. This was not necessary when developing the power function model. As shown in Table 3, the power function model provided the best fit to the data and will be used hereafter in the risk assessment. Since the power function model involves the natural logarithms of cumulative exposure and cumulative cigarette smoking, zero values of these variables were not permitted. In order to avoid this an estimate of cumulative background exposure was added to each miner's cumulative radon daughter and cigarette totals. Based upon estimates of the NCRP (Report No. 77, 1984), 0.2 WLM per year since birth were added to each miner's exposure totals. This is the estimated background exposure in the U.S. and is also the amount used by Whittemore and McMillan [1983] in an earlier analysis. In a similar fashion 0.005 packs per day were added for each day since birth to the cumulative smoking totals based upon estimates of Hinds and First [1975]. Of particular interest is the joint effect of exposure to radon daughters and cigarette smoking. Therefore, the interaction of radon daughter exposure and cigarette smoking was included in the multiplicative power function model. The results showed a negative, borderline significant result ( B=0.087,p=0.058). When a similar analysis was run with mortality data complete only through 1977, there was no indication of a significant negative effect. Therefore, based on more complete follow—up through 1982, the joint effect of radon daughter exposure and cigarette smoking appears to be slightly less than multiplicative but greater than additive. This is similar to the finding of Thomas and McNeill [1985] in their grouped data analysis of the five major radon daughter cohorts. It is still consistent with a synergistic effect of radon exposure and cigarette smoking which is usually defined as a joint effect exceeding the sum of the individual effects. 2. Weighting Exposure Over Time An important consideration in fitting any of these models was the proper time-weighting of exposure. Since most forms of cancer, including lung cancer, have relatively long latency periods between exposure and manifestation of the disease, some weighting of exposure over time is appropriate. The most common weighting scheme is commonly referred to as lagging. This involves elimination of any exposure accumulated in a specified period of years before death from lung cancer. This provides a way of considering only that exposure that had a reasonable chance of causing death from lung cancer. Obviously exposures received in the few years immediately prior to lung cancer deaths are ineffective in the exposure—response relationship. In order to investigate the appropriate number of years to lag exposure in this cohort, a series of lags ranging from 0 to 12 years was used. Figure 1 illustrates the results of these trials. It is evident from the improved fit, as measured by the log—likelihood of the model, that a 147 UOOEHFNWHL“ owa>pu FIGURE 1 EFFECT OF LAGGING ON LIKELIHOOD OF MODEL 190 180 170 160 o 2 4 6 a 10 12 LAC (YEARS) 148 lag of 6 years for cumulative exposure is the best choice for this analysis. Cumulative cigarette smoking was rather insensitive to the amount of lag in the range of 0 to 12 years. Therefore, for the purpose of consistency cumulative smoking was also lagged 6 years. This contrasts to the lag of 10 years chosen by Whittemore and McMillan [1983] for these data and also by Muller et al. for the Canadian data. Their choices were somewhat arbitrary and largely based on knowledge that most cancers involve relatively long latency periods. The implications of a shorter lag will be discussed in a later section of this report. An issue related to lagging of cumulative exposure and cumulative cigarette smoking is the lack of information on these variables in recent years. Radon daughter exposure was last updated in 1969. However, the absence of current exposure information should have minimal impact upon this analysis since over 90% of the miners in the cohort had retired from uranium mining for more than one year by 1969. Those few who continued mining were exposed at levels considerably less than those experienced in earlier years. Since cigarette smoking status was also unknown after 1969, all miners still smoking at that time were assumed to continue at their last recorded smoking rate. NlOSH is currently conducting a survey of radon daughter exposure and cigarette smoking status subsequent to 1969, but this information will not be available for at least another year. The aim of lagging exposure is the elimination of exposure which is not etiologically responsible for lung cancer mortality. An implicit assumption in the use of this technique is that exposure changes from completely effective to completely ineffective at one instant in time. The actual form of this weighting function is illustrated in Figure 2. Because of the biological implausibility of such a situation, Land [1976] proposed that the effectiveness of cumulative exposure be linearly phased in over a period of several years. An illustration of such a weighting function is provided in Figure 3. Consequently, we tried various combinations of lagging and linear partial weighting with the combination illustrated in Figure 3 providing the best fit, i.e. a lag of 4 years followed by linear partial weighting in the period 4—10 years prior to death from lung cancer. This scheme provided a fit essentially the same as that of a simple lag of six years but was chosen over lagging because of its biological plausibility. 149 09L FIGURE 2 EXAMPLE OF SIX YEAR LAG HEIGHTJNG SCHEME HEIGHT 1.5 ~ 1.0 H , I . I I ~ I I . I I I ' I I ‘ I I I + I — I I - I I I ‘ 3 O. 5 "‘ I I " I I - I I - I I I ‘ I I ‘ I I ~ I - i I J I I ‘ : 0.0 _ WWI merl VT" ffi'VYVVTrVV-VTTYYTV'V‘IYT'TV m “WW"VTTVTTTTVTVTTT‘TTPVT‘IWV HYTfTr 0 l 2 3 4 5 6 7 8 9 10 ll 12 YEARS YEARS IMMEDIATELY PRECEOING LUNG CANCER DEAIH FIGURE 3 EXAMPLE OF LAGGING WITH PARTIAL HEIGHTING VVEKSHT 1.5 1.0 — '0 I I I I I I l l I I 0 1 2 3 4 5 6 7 8 9 10 11 12 Years immediately preceding lung cancer death 151 lll. INFLUENCE OF TEMPORAL FACTORS A. Exposure-Rate Effect Perhaps the most difficult aspect of producing a valid quantitative risk assessment is dealing with the effects of various time—related factors upon the exposure-risk relationship. One very important temporal influence concerns the two components of cumulative exposure itself. In most longitudinal studies the quantitative exposure index is some form of cumulative exposure. However, cumulative exposure is actually the product of duration of exposure and intensity or rate of exposure. When one uses cumulative exposure in assessing risk, the implicit assumption is that high exposure rates for short periods of time are equivalent etiologically to low exposures for long periods of time, all else being equal. A number of investigators have examined the effect of exposure rate in the U.S. uranium miner data. Whittemore and McMillan [198]) found no statistically significant effect of exposure rate. Lundin et al. in the 1971 monograph concluded that there was no significant evidence of an exposure rate effect in the 120—360 WLM cumulative exposure range. These investigators apparently defined exposure rate as the ratio of total cumulative exposure and duration of employment (defined as the period of time between first and last employment in underground uranium mining work histories). For most forms of employment, this is the accepted definition of average exposure rate. However, underground uraniunl mining is a very sporadic form of employment. The actual time spent underground was often a relatively small fraction of the total employment history. Therefore, exposure rate as defined by cumulative exposure divided by the number of months actually spent underground is often a very different measure than that obtained by using duration of employment in the denominator. Consequently, the effect of exposure rate was re—examined using the actual average exposure rate experienced while underground, eliminating any gaps in employment. Although earlier analyses using total duration of employment produced negative but non—significant results, the refined definition showed a statistically significant negative exposure rate effect (B =—0.043, p<0.001) as shown in Table 4. This implies that among groups of miners receiving equivalent cumulative exposures, those exposed to lower levels for longer periods of time are at greater risk of lung cancer. Because the coefficient is relatively small, however, an appreciable effect upon risk of lung cancer would not be expected unless rates were different by an order of magnitude, i.e., a miner with exposure received at a rate ten times lower than a miner of the same age, smoking habits, and cumulative exposure would have (0.1)'-043=1.104 or 10.4% greater risk of lung cancer. Because a negative exposure rate effect is very important and potentially controversial, it was examined in more depth. Of particular interest was the possibility that this effect was different at low versus high cumulative exposure levels. Consequently, the homogeneity of this effect across the full exposure range was examined by forming two sub—cohorts: one below the mean exposure (834 WLM) and one above the mean. The interaction of the exposure rate effect with these two strata was then tested. Results showed a significant interaction ( B=0.157, P=0.019). The direction of the 152 Table 4. Quantitative Relative Risk Model Risk factor Coefficient x2 P—value Ln(cumulative exposure+BGR)(WLM)1 0.731 139.5 (0.001 Ln(cumulative cigarette smoking+BGS) 0.291 34.5 (0.001 (packs)2 Age at initial exposure 0.0023 8.8 0.003 (months) Ln(exposure rate)(WLM/month) —0.043 18.6 (0.001 Exposure Rate Interaction Model Ln(cumulative exposure+BGR) 0.660 101.4 (0.001 Ln(cumulative cigarette smoking+BGS) 0.292 34.8 (0.001 Age at initial exposure 0.0024 9.2 0.002 Ln(exposure rate) —0.198 8.9 0.003 Ln(exposure rate) x exposure category: 0.157 5.5 0.019 Exposure (834 WLM Exposure 3834 WLM ll ll 0 1 1Background for cumulative radon daughter exposure: BGR=0.4 WLM/year 2Background for cumulative cigarette smoking: BGS=0.005 packs/day 153 interaction indicated that the exposure rate effect was stronger in the lower cumulative exposure range (0—834 WLM). Specifically, a miner who received total exposure below 834 WLM at rate one tenth as great as another miner of the same age, smoking status and cumulative exposure would have a 58 percent greater risk of lung cancer. However, the increased risk would only be 10 percent at the lower exposure rate for miners in the 834—10,000 WLM range. Although a statistically significant negative exposure—rate effect had not been found previously in this U.S. cohort, there is considerable evidence of such findings in animal studies of high LET radiation. Raabe et al. [1983] reported a strong low dose—rate effect in beagles exposed to internally deposited isotopes of radium and strontium. Risk of bone cancer was as much as ten times as great per unit dose for low rates as compared to the highest rates used. Cross et al. [1980] found a negative dose—rate effect for risk of lung tumors in rats exposed to airborne radon daughters. Chameaud et al. [1981] found similar results in a French study of Sprague—Dawley rats exposed to inhalation of radon decay products. Hill et al. [1982] found reduced dose rates of fission—spectrum neutrons produced significantly higher neoplastic transformation rates per rad in cell cultures of C3H mouse embryos. Although all of these studies show low dose—rate effects, no study as yet, animal or human, has investigated such effects at the very low dose rates currently found in well-ventilated uranium mines. B. Calendar Time It is well—known that mortality patterns change over time. Such exogenous risk factors as the prevalence of smoking and alcohol consumption, medical care, and various life style characteristics are all influenced by a changing society. Therefore, the effect of calendar time upon risk estimates, often called the cohort effect, must be controlled. The analysis of the U.S. uranium miners cohort was stratified by decade of birth so that miners dying of lung cancer were compared only to those members of the cohort at the same age and who were born within 10 years of the case. The usual assumption in a stratified analysis is that baseline mortality rates may be different from stratum to stratum but the relative risk is the same across all strata for miners with comparable risk factors. In order to check this assumption, the interaction of cumulative radon daughter exposure and birth decade was examined. Results indicated a statistically significant positive interaction ( B=O.173,P=0.002). This implies that miners born in later decades are at a greater risk of lung cancer per unit of exposure when compared to miners of the same age born earlier. Since miners born in later decades were exposed at lower exposure rates, this result could be associated with the negative exposure rate effect described earlier. c. Multistage Theory of Carcinogenesis One of the most popular theories for explaining the temporal patterns in mortality studies of carcinogenesis is the multistage model. Originally proposed by Muller [1951] and Nordling [1953] and later refined by Armitage and Doll [1961], the multistage theory predicts an increase in cancer incidence as a function of time since exposure to some carcinogen. In 154 general, the theory proposes that a malignant tumor arises from a single cell which has undergone a series of heritable changes. The changes may be thought of as distinct stages in the carcinogenic process, each with a low probability of occurrence and a slow progression time in the absence of carcinogenic exposures. A carcinogen may act on any or all of the stages in this process. Carcinogens affecting the first stage are commonly referred to as initiators, while those affecting later stages are called promoters or progressors. Initiators are characterized by long latency periods between initial exposure and death, often exceeding 20 years. Promoters, on the other hand, usually have shorter latent periods since fewer stages must be transgressed before a malignant cell is produced. It is impossible to prove whether or not the mathematical form of the multistage model actually holds in a given situation. However, a number of its predictions have been verified experimentally by Peto et al. [1975]. Therefore, if one subscribes to some form of the nmltistage model, it is possible to predict whether exposure acts at an early or late stage in the carcinogenic process by examining the temporal patterns in the data. Whittemore [1977], Day and Brown [1980], and Brown and Chu [1983] have all reported the effect on excess relative risk of age at initial exposure and time since cessation of exposure. By examining these factors, we may better understand the underlying cancer mechanism operative in this cohort. D. Age at Initial Exposure Whittemore [1977] considered the multistage model using three exposure scenarios: single exposure at one point in time, continuous exposure at a constant rate, and exposure of varying intensity. When considering the latter category (the usual occupational situation) she found that excess relative risk was a decreasing function of age at initial exposure if an early stage was affected. When a late stage is affected by exposure, however, excess relative risk is an increasing function of age at initial exposure. Day and Brown [1980] predicted the functional relationship between excess relative risk and age at initial exposure for the first four stages of a five—stage process when duration was held constant. Figure 4 illustrates their findings which are in qualitative agreement with those of Whittemore. Results of the analysis in our data, as illustrated in Table 4, indicate a positive and statistically significant coefficient for age at initial exposure (3:0.0023, P=0.003). This implies that miners initially exposed at later ages are at greater risk of lung cancer than those exposed at younger ages, all else being equal. Specifically, a miner with the same radon daughter exposure and smoking history who was initially exposed ten years (120 months) later in age than another miner, would have exp(0.0023x120)=1.32 or 32% higher risk of lung cancer. This result is consistent with the effect of radon daughters occurring at a late stage in the carcinogenic process. A similar age effect was reported by Mancuso et al. [1977] in an analysis of cancer risk in the Hanford workers exposed to whole—body radiation. 155 RELATIVE INCREASE IN INCIDENCE FIGURE 4 EFFECT OF AGE AT INITIAL EXPOSURE ON A MULTISTAGE MODEL 100096”- 100% 10% 15 mcflO “4qu 5““ ‘ "com: snag Atrtcrso ‘ ""57 3,46: “ch7 (D I I I 20 25 3O AGE AT INITIAL EXPOSURE 156 An analysis of age at start of smoking among miners resulted in a negative but non—significant coefficient (B=0.0016,p=0.22). This would imply that cigarette smoking in this cohort acted at an early to intermediate stage. It could also be consistent with the hypothesis of Doll and Peto [1978] that smoking acts at both early and late stages, which would tend to obscure predictive ability of age at start of smoking. A plot of the effect of age at initial exposure for both radon daughters and cigarette smoking is given in Figure 5. E. Time Since Cessation of Exposure Day and Brown [1980] predicted the effect upon relative risk of time since cessation of exposure when a multistage model is assumed. They found that when exposure begins some time after infancy, excess relative risk increases, peaks, and then decreases with time since termination of exposure when the first stage is affected. When the penultimate (next to last) stage is affected, relative risk strictly decreases with time after last exposure. Figure 6 illustrates their predictions for the effect of time since cessation of exposure on the first four stages in a five stage model with duration of exposure fixed at five years. In order to investigate the effect of cessation of exposure on this cohort, all miners were identified who had indicated retirement from uranium mining during the course of follow-up. Approximately 95% of the cohort had retired for more than one year prior to 1970. The average time since last exposure was 18.0 years for those miners not dying of lung cancer and 9.9 years for lung cancer cases. The time in months since last exposure was entered as a time—dependent covariable in the original model containing log of exposure, log of smoking, and age at initial exposure. The estimated coefficient of this term was negative and highly significant ( B=-0.0056,p<0.001). Thus a miner's chances of surviving lung cancer increase dramatically with each year outside the mines. Specifically, the model predicts that the risk of lung cancer 10 years after mining uranium is exp(—0.0056x120)=0.511 relative to someone still mining with the same cumulative exposure, smoking history, and age. When a similar analysis of time since cessation of cigarette smoking was run, the results were inconclusive. The coefficient was very small and non—significant (B:0.003,p=0.75). However, since a relatively small number of miners were ex—smokers (7.7%) there is little power for detection of such an effect even if it actually exists. Figure 7 illustrates the effect of time since last exposure for both radon daughters and cigarette smoking. The implication of these results are essentially the same as that obtained by examination of age at initial exposure. The strong negative effect of time since last exposure implies that radon daughters act at a late stage in the carcinogenic process. The effect of stopping cigarette smoking, while based on a small amount of data, still indicates either an intermediate stage effect or a combination of early and late stage effects. 157 FIGURE 5 EFFECT OF AGE AT INITIAL EXPOSURE ON RISK RELATIVE TO MINER BEGINNING AT AGE 15 Exposure to: RELATIVE RISK . Smokmg 3.0 Radon 2.0 1.0 .0 l l l I l l 15 20 25 30 35 4o 45 50 AGE AT INITIAL EXPOSURE 158 FIGURE 6 EFFECT OF CESSATION 0F EXPOSURE ON A MULTISTAGE MODEL Ills! SYAG! AIVICTID / ucono sum nucno 7n n I o 3746! Ange,“ 100% RELATIVE INCREASE IN INCIDENCE 1096 F" A 1% L 1 n 1 1O 20 3O 40 50 YEARS SINCE EXPOSURE STOPPED 159 09L wam m

rmm 24 I lhl L11 1 L l l__.1 1 l_.l l l CS7 PWGURE'7 EFFECT OF TIME SINCE LAST EXPOSURE 0N EXCESS RELATIVE RISK ~~~wm~~~-w~o~n D—‘HN‘ ._ ”HHwMHV~~~ M—n‘wl—Mo—Wuwwwwkgww_awv ,J. .— >~-—--’o—>—--Ho—-.......‘ .4... Q TIME SINCE LAST EXPOSURE (YEARS) RADON “““““ ~ SMOKING LEGEND: EXPOSURE IV. ERRORS IN EXPOSURE DATA AND THEIR EFFECT UPON RISK ASSESSMENT In animal carcinogenesis studies, exposures or doses are usually known with a high degree of accuracy and precision. However, the same cannot be said regarding epidemiologic quantitative risk studies. “1 most epidemiologic studies, the actual dose to target organs can only be estimated by dosimetric modeling. This is seldom attempted in quantitative risk assessments. The dosimetry of radon daughter exposure is very complex, involving such factors as respiration rates, particle size distribution, deposition in the lung, and radon/radon daughter equilibrium. Most risk assessments are modeled as functions of some exposure index, which is the method used in this report. It is the purpose of this section to estimate the magnitude of exposure errors and their effect upon quantitative risk models. According to Lundin et al. [1971], exposures in a given mine and year were estimated in one of four ways: 1. actual measurements 2. interpolation or extrapolation in time 3. geographic area estimation 4. estimates prior to 1950 based upon knowledge of ore bodies, ventilation practices, and earliest measurements. These methods will subsequently be called Methods 1, 2, 3, and 4. In assessing the error associated with individual exposure determinations, it is first necessary to consider the variability introduced by each of the four methods. A. Magnitude of Error in Exposure Data Method 1 Table 5 provides a frequency count of white miners working underground from 1950—68 and the mean number of samples taken in each mine visited in those years. The Kusnetz procedure for measuring radon daughters was most often used during the period of study (Johnson and Schiager 1981). This is an area monitoring method based on alpha counts collected on a filter/pump apparatus. The resulting data were generally thought to be of good quality (Lundin et al., 1971). Data from mines in which 5 or more measurements were taken in a given year were analyzed. These data followed a lognormal distribution with little change over the period 1951—1968. Prior to 1960, samples were taken largely by the U.S. Public Health Service, while post—1960 sampling was conducted by state Inine inspectors. Therefore, data were separated into pre and post 1960 periods and estimates of the coefficient of variation (CV) were made for each period. Results indicated a slight but non—significant increase in CV's after 1960 (106.6% vs 118.3%). Since the measurements were grab samples taken at different times within each mine, the total pooled CV=112.5% over the period 1951—1968 is assumed to include sampling errors, counting errors, and environmental fluctuations over time. This estimate agrees well with the CV of 110% found in an independent study of U.S. mines in the period 1973—79 when exposure levels were much lower (Schiager et al. 1981). In other studies, however, an average CV of 30% 161 Table 5. Number of Miners Exposed and Mean Number of Exposure Measurements Taken by Calendar Year Year Number of Miners Exposed Mean Number of Samples/Mine 1950 534 1.0 1951 668 4.2 1952 748 1.6 1953 1028 8.5 1954 1376 4.3 1955 1383 3.8 1956 1572 14.2 1957 1942 5.6 1958 1798 8.8 1959 1861 6.6 1960 1902 9.9 1961 1588 8.8 1962 1369 12.9 1963 1005 8.4 1964 828 15.6 1965 640 18.1 1966 467 18.5 1967 480 21.4 1968 336 21.9 162 was reported for area samples in Canadian mines (Makepeace and Stocker 1980) while fluctuations of 20—30% around daily means were found for radon measurements in non-uranium Norwegian mines (Berteig and Stranden 1981). Method 2 In order to assess the error in interpolating for gaps in sampling of 1 to 3 years, a simulation procedure was used. Mines having the longest periods of continuous annual measurements were identified. Then the even years' averages wee omitted and the average of the two adjacent years was substituted. In this way it was possible to compare the observed annual average with the expected average had that year been missing. This strategy was repeated by imposing three year gaps in the data and again using the average of adjacent years to estimate the three intervening years. The error variance attributable to Method 2 was then calculated by: 02: E (Iog(0./E.))2 I N—1 where 0i = actual measurements for intervening years Ei = interpolated values estimated by average of adjacent years. The resulting CV was 120.8% for 1 year interpolation and 137.3% for 3 year interpolation. Since these results were not significantly different, they were pooled to yield a CV=131.9%. Method 3 This method used annual mine averages in the same geographic locality to estimate radon daughter levels in mines for which Methods 1 and 2 could not be used. In order to assess the error associated with this method, four of the uranium mining localities with the greatest number of annual measurements were selected. A simulation procedure similar to that used for Method 2 was employed. Annual averages for selected mines in these localities were omitted for 1 to 4 years. The averages for mines in the nearest district were substituted as the expected radon level if the annual average actually had been missing. The error variance was calculated in the same way as Method 2. The resulting CV was 148.6% for this method. Method 4 No measurements were available in the period prior to 1950. Therefore, the estimates made using knowledge of ore bodies, ventilation, and earliest known measurements in these mines could not be verified. These estimates comprised less than 6% of the 34,120 annual averages used in exposure assessment. In addition, since only 8 percent of the total underground exposure time for the cohort occurred prior to 1950, the influence of these measurements should be minimal. However, since the 163 error for this method was probably the greatest of the 4 methods used, we estimated the overall CV for Method 4 to be 25% greater than that for Method 3, i.e. CV=186%. Table 6 shows the number of annual averages for each of the four methods. Actual measurements comprised only about 10% of the data. In order to obtain an overall estimate of the relative error, a weighted average of the CV's for each method was calculated based on the number of determinations for each method. The resulting overall CV=137%. The error associated with each miner's cumulative exposure can then be calculated using our estimate of the error H1 each radon daughter level (WL). The total cumulative exposure (WLM) for each miner is obtained from: WLM = '2‘,- (Mil-Muslim”) I,] where WLi- is the estimated exposure for mine i in year j and UGMONij is the number of months spent underground in mine i during year j. The variance of WLM assuming independence of WLij is then: Var(WLM) 2 .2; (UGMONi j) var (WLi') l,] J 2 .22. (UGMONij)2(CV)2(WLij) l,j where CV is the coefficient of variation for the estimated exposure WLij. If we substitute our estimate of the overall CV=137% and use total cumulative exposure divided by total months underground (WLM/TOTMON) as an estimate of WLi- for each individual miner, the average CV for cumulative exposure (WLM) is 0.97 or a relative standard deviation of 97% of the total WLM for each miner. Since radon daughter measurements were taken in different areas of each mine and often at different times of the day or week, we will assume that the variance in these measurements reflects the variance in exposure levels among individual miners, i.e. 2 Var(WLMij)— oijk where oijk = variance in exposure measurement for miner k ht mine i and year j. 164 Table 6. Exposure Measurement Errors Due to Four Methods of Estimating Annual Radon Daughter Concentration Exposure Assessment Variance of Coefficient of Technique N Natural Log (oz) Variation Actual measurements 3505 0.82 1.13 Interpolation over time 5602 1.01 1.21 Geographic area estimation 23159 1.16 1.49 Estimates prior to 1950 (assumed 1.25 x geographic error) 1.86 165 B. Effect on Relative Risk Estimation of Exposure Measurement Errors There appears to be a general impression that errors in exposure measurements usually cause an underestimation of relative risk. Indeed, Bross [1954] originally demonstrated that if misclassification was equal in two comparison populations, one would tend to underestimate differences in proportions of diseased persons. Keys and Kihlberg [1963] qualified this concept by showing that relative risk is underestimated when misclassifie cation errors are independent of disease and exposure relationships. In general, it has been shown by Copeland et al. [1977] among others, that relative risk estimates are biased too low in the presence of non— differential misclassification (equal misclassification of disease in both exposed and unexposed groups). Little work has been done concerning the effects of errors in continuous measures of exposure upon relative risk estimates obtained from statistical models. It is this situation that is a potential problem to the analysis in this report. Prentice [1982] introduced a method for dealing with errors in individual exposure measures when using the Cox proportional hazards model. Prentice, and more recently Hornung [1985], have shown that the direction of bias in relative risk estimation depends upon the error distribution and the shape of the exposure—response model. In general, when the variability in individual exposure errors increases with the level of exposure and the relative risk model is supra—linear (curving upward), relative risk will actually be overestimated when exposure errors are ignored. The popular log—linear or exponential risk function is an example of a model which may often overestimate relative risk in the presence of errors whose magnitude increases with increasing levels of cumulative exposure. As was reported earlier, the log—linear model did not provide the best fit to the data. Instead, the power function model which involved the logarithms of cumulative exposure and cumulative cigarette smoking provided a better fit. The effect upon risk estimates using this model was investigated when errors in exposure are lognormal as indicated in the previous section. Without presenting the statistical details, it is sufficient to say that under these conditions (power function model and lognormal distribution of exposure errors) the effect upon relative risk estimates is negligible. If the exposure measurements were generally higher than those actually experienced by the miners, as mentioned in the 1971 Monograph, relative risk per WLM would be underestimated regardless of the distribution of exposure measurement errors. In summary, the degree of error in individual exposure measurements was quite high, an estimated CV of 97%. If, however, these individual errors were lognormally distributed about the annual average concentration in each mine, the degree of bias in relative risk estimates generated by the power function model would be minimal. Regardless of the form of the error distribution, the relative risks generated by the exposure-response model would be too low if the exposure measurements were systematically too high. Therefore, examination of the pattern of error in the exposure data would suggest that relative risks produced by the power function model are either unbiased or possibly a bit low. 166 V. QUANTITATIVE RISK ESTIMATES The previous sections have outlined the protocol for the risk model development, the selection of an appropriate quantitative risk model, the temporal factors influencing risk estimation, and the magnitude and effect of exposure measurement errors. These are factors requiring careful study before attempting to make valid quantitative risk estimates. In most risk assessments, results are reported relative to some unexposed population. In animal studies, a control group is generally used for this purpose. In life table, analyses expected mortality is obtained from some standard population, often that of the U.S. The problems inherent with the use of such external referents have been well documented [Enterline 1976]. Although a subcohort of miners unexposed to radon daughters would be ideal for a referent group, there were no unexposed miners in the U.S. cohort. Since the proportional hazards model uses internal comparisons in generating risk estimates, risk projections relative to an unexposed population necessarily involve an extrapolation to zero exposure. In the case of the power function model, a background exposure of 0.2 WLM/year of age was added to every miner's cumulative total. All risk estimates are relative to someone exposed to these background rates. Therefore, quantitative relative risk estimates are somewhat sensitive to the choice of a background exposure rate. One way of checking the appropriateness of the model is to divide cumulative exposure into discrete intervals and calculate lung cancer risks in each interval relative to risks experienced in the lowest interval. In this way, relative risk estimates are free of any exposure—response function. If the risk model then fits the risk estimates in the selected intervals, one would be assured that the model is appropriate for quantitative risk estimation. The cumulative exposure intervals chosen for this analysis were: less than 20 WLM, 20—120, 120—240, 240—480, 480-960, 960—1920, 1920-3720, and greater than 3720 WLM. Risk estimates in each interval are calculated relative to the risk in the interval less than 20 WLM. and are plotted at the mean exposure in each interval: 66.6, 179, 351, 698, 1352, 2579, and 5416 WLM, respectively. Figure 8 illustrates how these interval estimates are uniformly lower than those produced by the risk model when using 0.2 WLM/year as a background rate of exposure. The shape of the risk model, however, shows remarkably good agreement with the pattern of relative risk estimates in the selected intervals. This implies that the quantitative risk model is appropriate exclusive of the intercept. This could be due to either an improper choice of baseline exposure rate or the fact that all interval estimates are relative to exposure in the lowest interval, 0-20 WLM. If there is some level of excess risk in this interval relative to an actual unexposed population, the interval estimates would be too low. The cumulative exposure of 0.2 WLM/year is an estimate is an estimate of the background exposure in the overall U.S. population [NCRP Report No. 77, 1984]. Exposures near ore—bearing lands are known to be considerably higher than average [NCRP Report No. 45, 1975]. Therefore, it is probable that background exposures in the Colorado Plateau area are higher than average U.S. levels. 167 89L xm—x n<——o>r-n:u 100‘ 00* 70- 50‘ 401 FIGURE 8 RELATIVE RISK AS A FUNCTION OF CUMULATIVE RADON DAUGHTER BACKGROUND EXPOSURE-0.2 WLM/YEAR O 1000 2000 3000 4000 UL“ EXPOSURE ‘,_- _,— ,— l‘lll1ilvv IIIIIIVII IIIITIII'I I'V‘IITII TIIV‘I‘I—V‘II IIVTIVIj' I I I l 5000 6000 DDTTED LINES AND VERTICAL BARS REPRESENT 957. CONFIDENCE LIMITS In the interest of using a background more in line with exposures received by persons living in the Colorado Plateau, the background exposure was increased to 0.4 WLM/year. This produced a quantitative risk model that agreed very well with the interval estimates, as can be seen in Figure 9. Using this model, relative risk estimates were calculated for cumulative radon daughter exposures in the range 30 to 120 WLM corresponding to exposure levels of from one to four WLM/year over a 30—year working lifetime. These estimates range from a relative risk of 1.42 at 30 WLM to 2.07 at 120 WLM compared to someone of the same age and smoking habits with a cumulative lifetime background exposure of 24 WLM and a background exposure rate of 0.4 WLM/year. These estimates (0.9 to 1.4 excess relative risk per 100 WLM) are slightly higher than those reported by Muller et al. [1983] for the Ontario miners, but somewhat less than the estimates of Radford and Renard [1984] for the Swedish iron miners. Obviously, these estimates are subject to the usual caveats concerning extrapolation from higher cumulative exposures and exposure rates. Because relatively few data are currently available in this cohort below 120 WLM (10 lung cancer deaths out of 709 miners), there may be some doubt that the model used actually is appropriate at these low levels. However, the pattern of relative risk estimates produced in each of the categorized exposure levels would suggest that this model fits the data well in range of 60 to 6000 WLM. 169 OLL M<—-0>f-MZ xm—a 100' FIGURE 9 RELATIVE RISK AS A FUNCTION OF CUMULATIVE RADON DAUGHTER EXPOSURE BACKGROUND EXPOSURE=0.4 WLN/YEAR '[lllIItIIITTtvIrIIIlivvrvlt'rj'lr11lIIIIIIIIIIVIY'IITVIIITT1 1000 2000 3000 4000 5000 6000 WLM DOTTED LINES AND VERTICAL BARS REPRESENT 95% CONFIDENCE LIMITS VI. SUMMARY AND CONCLUSIONS A valid quantitative risk assessment is much more than simply fitting an exposure—response curve to mortality data. This is especially true when considering an epidemiologic risk assessment. There are a great variety of risk factors and temporal effects that may alter the interpretation of the data analysis. This report is an attempt to address such modifying influences in an effort to better understand the underlying cancer mechanisms operative in the cohort of U.S. uranium miners exposed to radon daughters. There were a number of findings which are important in assessing the risk of lung cancer in the U.S. cohort. 1. Influence of Cigarette Smoking The joint effect of cumulative cigarette smoking and cumulative radon daughter exposure was found to be intermediate between additive and multiplicative. This would imply a synergistic relationship in the usual definition as an effect exceeding the sum 13f the two relative risks. 2. Exposure-Rate Effect Analysis of this data revealed that modeling cumulative exposure alone may not adequately predict the relative risk of lung cancer from chronic exposure to radon daughters. Miners receiving a given amount of cumulative exposure at lower rates for longer periods of time were at greater risk relative to those with the same cumulative exposure received at higher rates for shorter periods of time. This effect is supported by the convex (decelerating) shape of the exposure—response model which indicates lower exposures are more effective per unit WLM than higher exposures. Though this result may seem somewhat counter-intuitive, it is consistent with a variety of animal carcinogenesis and in vitro cellular studies after treatment with alpha radiation. This implies that results extrapolated from historical exposures at high rates may yield conservative results at current lower rates. Indeed, it is possible that lower risk estimates ir1 the U.S. study, when compared to the four other major radon studies, as reported by Thomas et al. [1985] may be due to the higher exposure rates received by U.S. miners. 3. Late-Stage Carcinogenic Effect Careful examination of temporal effects implies that exposure to radon daughters acts at a late stage in the carcinogenic process. All temporal factors agreed in this respect. The appropriate lag to remove redundant exposure was a relatively short six years. Older miners at initial exposure were at greater risk than those exposed at younger ages. The relative risk of lung cancer decreases with the length of time after cessation of exposure. Whether or not the mathematical form of the multistage theory of carcinogenesis applies to this cohort, the temporal patterns are worth noting. 171 4. Magnitude and Effect of Errors in Exposure Measurements Analyses of the errors associated with the four methods of estimating uranium mine exposure levels indicated a lognormal distribution of errors with the relative standard deviation or CV=97 percent. Although errors of this magnitude may cause overestimation of relative risk when using the log—linear risk model, the better-fitting power function model is generally insensitive to errors of this type. In fact, if estimated exposure levels were systematically higher than those actually received by the miners [Lundin et al. 1971], relative risks per unit WLM would be underestimated for this data. 5. Quantitative Risk Estimates Present day radon daughter exposures are considerably less than those experienced in the past by uranium miners. There is also current interest in low~|evel exposure to the general population from indoor radon and its decay products. Consequently, the primary cumulative exposure range of interest in risk assessment appears to be below 120 WLM. Although approximately 20 percent of the cumulative exposures in this study were below this level, there have been only 10 lung cancer deaths among this subgroup as of the end of 1982. Until this cohort is followed to extinction, epidemiologic models such as that produced in this report will be necessary to evaluate the risk of lung cancer mortality at these lower exposures. The model developed for this report provides a very good fit to the data in the range 60 to 6000 WLM. It seems reasonable that predictions based upon this model would be reliable at least for occupational exposure to adult white males. There is little or no mortality data available regarding women and children. The risk estimates provided in Table 7 are presented as an evaluation based upon careful consideration of all factors thought to influence such long—term mortality studies. All of the caveats associated with such evaluations apply to some degree to these results. 172 Table 7. Quantitative Risk Estimates of Lung Cancer at Four Exposure Rates Over 3 Thirty Year Working Lifetime Exposure Cumulative Exposure Relative 95% Confidence Rate (30 Years)1 Risk2 Limits 1 WLM/year 3O WLM 1.42 1.18 — 1.72 2 WLM/year 60 WLM 1.66 1.22 - 2.26 3 WLM/year 90 WLM 1.88 1.28 — 2.76 4 WLM/year 120 WLM 2.07 1.33 — 3.22 1Exclusive of background exposure. 2Risks are calculated using exposure rate interaction model in Table 6 relative to miners of the same age and smoking habits with a cumulative lifetime background exposure of 24 WLM and background exposure rate of 0.4 WLM/year. 173 10. 11. LIST OF REFERENCES Armitage, P., and Doll, R. [1961]. Stochastic Models for Carcinogenesis, Proceedings of Fourth Berkeley Symposium on Mathematical Statistics and Probability 4, University of California Press, 19—38. Berteig, L., and Stranden, E. [1981]. Radon and Radon Daughters in Mine Atmospheres and Influencing Factors, In: M. Gomez, ed. Radiation Hazards in Mining. American Institute of Mining, Metallurgical and Petroleum Engineering, lnc., New York, 89—94. Bross, I. [1954]. Misclassification in 2 x 2 Tables, Biometrics 10, 478—486. Brown, C.C., and Chu, K.C. [1983]. Implications of the Multistage Theory of Carcinogenesis Applied to Occupational Arsenic Exposure JNCI 70, 455—463. Chameaud, J., Perraud, R., Masse, R., and Lafuma, J. [1981]. Contribution of Animal Experimentation to the Interpretation of Human Epidemiological Data, In: M. Gomez, ed. Radiation Hazards in Mining. American Institute of Mining, Metallurgical and Petroleum Engineering, lnc., New York, 222—227. Copeland, K.T., Checkoway, H., McMichael, A.J., and Holbrook, R.H. [1977]. Bias Due to Misclassification in the Estimation of Relative Risk, American Journal of Epidemiology 105, 488—495. Cox, D.R. [1972]. Regression Models and Life Tables, Journal of the Royal Statistical Society, Series B 34, 187—202. Cross, F.T., Palmer, R.F., Dagle, G.E., Busch, R H., and Buschbom, R.L. [1980]. Influence of Radon Daughter Exposure Rate, Unattachment Fraction, and Disequilibrium on Occurrence of Lung Tumours, Radiation Protection Dosimetry 7, 381—384. Day, N.E., and Brown, C.C. [1980]. Multistage Models and Primary Prevention of Cancer, Journal of the National Cancer Institute 64, 977—989. Doll, R. and Peto, R. [1978]. Cigarette Smoking and Bronchial Carcinoma: Dose and Time Relationships Among Regular Smokers and Life—Long Non—Smokers, Journal of Epidemiology and Community Health 32, 303—313. Enterline, P.E. [1976]. Pitfalls in Epidemiologic Research: An Examination of the Asbestos Literature, Journal of Occupational Medicine, 18, 150—156. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. Hill, C.K., Buonaguro, F.M., Myers, D.P., Han, A. and Elkind, M.M. [1982]. Fission—Spectrum Neutrons at Reduced Dose Rates Enhance Neoplastic Transformation, Nature, Vol. 298, 67-69. Hinds, W.C., and First, M.W. [1975]. Concentrations of Nicotine and Tobacco Smoke in Public Places, New England Journal of Medicine 292, 844—845. Hornung, R.W. [1985]. Modeling Occupational Mortality Data with Applications to U.S. Uranium Miners, Doctural Dissertation, Dept. of Biostatistics, School of Public Health, University of North Carolina. Keys, A. and Kihlberg, J.K. [1963]. Effects of Misclassification on Estimated Relative Prevalence of a Characteristic, American Journal of Public Health 53, 1656—1665. Land, E. [1976]. Presentation for 1976 OSHA Hearings on Coke Ovens. Lundin, F.E., Archer, V.E., and Wagoner, J.K. [1979]. An Exposure— Time—Response Model for Lung Cancer Mortality in Uranium Miners: Effects of Radiation Exposure, Age, and Cigarette Smoking, Energy and Health, Breslow and Whittemore (eds.) SIAM, Philadelphia, 243—264. Lundin, F.E., Wagoner, J.K. and Archer, V.E. [1971]. Radon Daughter Exposure and Respiratory Cancer Quantitative and Temporal Aspects, NIOSH-NIEHS Joint Monograph No. 1. Feigl, P. and Zelen, M. [1965]. Estimation of Exponential Survival Probabilities with Concomitant information, Biometrics 21, 826—838. Makepeace, C.E., and Stocker, H. [1980]. Statistical interpretation of a Frequency of Monitoring Program Designed for the Protection of Underground Uranium Miners from Overexposure to Radon Daughters, The Canadian Mining and Metallurgical Bulletin 73, 113—124. Mancuso, T.F., Steward, A., and Kneale, G. [1977]. Radiation Exposures of Hanford Workers Dying from Cancer and Other Causes, Health Physics 33, 369—385. Muller, H.G. [1951]. Radiation Damage to the Genetic Material. Science Program 7, 93—493. Muller, J., Wheeler, W C., Gentleman, J F., Suranyi, G. and Kusiak, P.A. [1983]. Study of Mortality of Ontario Miners, 1955—1977, Part |. Ontario Ministry of Labor. National Council on Radiation Protection and Measurements [1975]. Natural Background Radiation in the United States, NCRP Report No. 45, Washington, D.C. National Council on Radiation Protection and Measurements [1984]. Exposures from the Uranium Series with Emphasis on Radon and its Daughters, NCRP Report No. 77, Washington, D.C. 175 25. 26. 27. 28. 29. 30. 31. 32. Nordling, C.O. [1953]. A New Theory of the Cancer Inducing Mechanism, British Journal of Cancer, 7, 68—72. Prentice, R.L. [1982]. Covariate Measurement Errors and Parameter Estimation in a Failure Time Regression Model, Biometrika 69, 331—342. Raabe, 0.6., Book, 8 A., Parks, N.J. [1983]. Lifetime Bone Cancer Dose—Response Relationships in Beagles and People from Skeletal Burdens of 226 Ra and 90 Sr Health Physics Vol. 44 Supplement No. 1, 33—48. Radford, E.P. and Renard, K.G.S.C. [1984]. Lung Cancer in Swedish Iron Miners Exposed to Low Doses of Radon Daughters, New England Journal of Medicine 310, 1485—1494. Schiager, K.J., Johnson, J.A., and Borak, T.B. [1981]. Radiation Monitoring Priorities for Uranium Miners, in Radiation Hazards in Mining, M. Gomez (ed ) American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc., New York 738-745. Thomas, D.C., McNeill, K.G., and Daugherty, C. [1985]. Estimates of Lifetime Lung Cancer Risks Resulting from Radon Daughter Exposure, Health Physics, in press. Whittemore, A.S. [1977]. The Age Distribution of Human Cancer for Carcinogenic Exposures of Varying Intensity, American Journal of Epidemiology 106, 418—432. Whittemore, A.S. and McMillan, A. [1983]. Lung Cancer Mortality Among U.S. Uranium Miners: A reappraisal, Journal of the National Cancer Institute 71, 489—499. 176 APPENDIX III ENGINEERING CONTROL METHODS A. Introduction This appendix contains examples of engineering control methods that can be used to reduce miners' exposure to radon progeny in underground uranium mines, although the same methods are applicable to other hard rock mines. Many of these control methods have been traditionally used in uranium mines, yet only recently have researchers (primarily from the Bureau of Mines) studied the efficacy of these methods [Bates and Franklin 1977; Bloomster et al. 1984a, 1984b; Franklin et al. 1975a, 1975b, 1977, 1981, 1982; Steinhausler et al. 1981]. B. Mechanical Ventilation Mechanical ventilation is the primary and most successful technique for reducing exposure to radon progeny. Average measurements of 2 to 200 working levels (WL) of radon progeny were common in U.S. uranium mines during the early 1950's before mechanical ventilation became prevalent [Lundin et al. 1971]. In contrast, during 1979 and 1980, the average concentrations of radon progeny recorded by MSHA ranged from 0.30 to 0.46 WL in the production areas of 61 underground uranium mines [Cooper 1981]. Thus the concentration of radon progeny in U.S. uranium mines has been greatly decreased, mainly because of improved ventilation. Sweden has also successfully reduced radon progeny concentrations in mines with improved mechanical ventilation; the average annual exposure for nonuranium miners in Sweden decreased from 4.7 working level months (WLM) in 1970 to 0.7 WLM in 1980 [Snihs 1981]. 1. General Principles Dilution ventilation in large mines consists of primary and secondary ventilation systems. In the primary system, fresh air is brought into the mine either through separate air shafts or through mine entrances used for miner access and equipment transport. The air can be blown in by a fan located at the surface or drawn in by a fan located inside the mine. Once in the mine, the air is blown or drawn through the main active passageways and then is pushed or drawn out of the mine through special ventilation shafts or openings used to remove ore. The secondary or auxiliary ventilation system provides fresh air to miners working in areas that include stopes and faces where access comes from a single shaft or drift and thus the work area is a dead end. For these areas, the air is often removed through the same shaft that was used to bring in the air. The source of fresh air for the secondary system is provided from the primary air system in the main passageway. To prevent mixing fresh air and contaminated air in the shaft or drift leading to the dead end, the secondary system usually consists of ductwork with a fan to blow or exhaust fresh air from the main passageway to the face. The contaminated air then passively returns to 177 the main passageway (because of a pressure gradient) through the shaft without contaminating the supply air. The contaminated air at the stope may also be brought back to the main passageway through a second duct and fan system. Once returned to the main passageway, the contaminated air joins the primary exhaust air stream which is then carried out of the mine. 2. Designing a Dilution Ventilation System Ventilation requirements must be considered when planning and designing the mine. Adding mine ventilation as an afterthought once the mine has been designed or completed is usually more expensive and less efficient. Consideration should be given to the following when designing the ventilation plan for a mine [Ferdinand and Cleveland 1984; Bossard et al. 1983]: 0 Identify the outline of the ore body that will be mined; 0 Determine the rate of emanation of radon from the rock in which the ore occurs; 0 Place as much of the primary ventilation system as possible including entrances and passageways in barren ground (i.e., ground not containing ore); 0 Set up passageways so that a split or parallel system of ventilation can be used; 0 Set up the mine so that working faces ventilated in a series are minimized; I Design the mine so that air inlets are located on one side of the ore body and exhaust airways on the opposite side of the ore body; 0 Design the mine so that the distances ventilation air travels in the mine are minimized (reduce or eliminate reentrainment and short circuits); 0 Design the mine so that adequate volumes of air can be provided without having high pressure drops across air controls in haulage and production areas; 0 Design the ventilation system to account for increasing concentrations of radon gas, and therefore radon progeny, since as the mine ages there will be more surface area for gas exchange into the mine; and 0 Consider control devices, fans, push—pull systems, and minimizing leaks when designing the system. 178 3. Primary Ventilation System The primary ventilation system delivers fresh air for the secondary air system and removes contaminated air from the secondary air system. The design of the primary air system is discussed in the following paragraphs. a. Split or Parallel Ventilation Systems A "split" or "parallel" ventilation system involves providing all or just a few working areas with fresh air that has not been previously used to ventilate other working areas. After the working areas are ventilated, the air is then pushed or drawn back into the primary system where it is moved out of the mine. By contrast, in a "series" ventilation system, all areas are ventilated by a single continuous air circuit. The advantages of a split or parallel system include the reduction of both the residence time and cumulative air contamination [Ferdinand and Cleveland 1984]. A series system, on the other hand, has several disadvantages. In addition to its long residence air times and a cumulative build—up of air contaminants from one area to another, other disadvantages include the following: 0 High air velocities which are often required; 0 Higher power costs associated with moving air at high velocities because of increased static pressures, unless additional ventilation shafts are constructed [Rock et al. 1971]; and O The potential spread of toxic gases to all areas of the mine in the event of a fire. However, to keep residence time down in a split or parallel ventilation system, the air velocities to the multiple drifts must be maintained. This will increase the fan and power requirements; additional ventilation shafts may also be necessary. b. Control Devices Sliding door regulators are used to prevent air from passing where miners and equipment need to pass through periodically. The problem with doors is that to be effective they must be closed after being used. Doors must also be well—constructed to remain secure with repeated usage; steel doors in substantial frames are most commonly used in Canada [Rock and Walker 1970]. c. Pushing Versus Pulling Ventilation Systems The pressure on the air intake side of a mine is always greater than on the exhaust side regardless of whether a pushing or pulling ventilation system is used. The difference is that the intake side 179 4. pressure is greater than atmospheric pressure for a pushing system, whereas in an exhaust or pulling system, the pressure on the intake side is below atmospheric pressure. Exhausting (pulling) offers some advantages over a pushing system. For example, forcing air into haulageways and escape areas often requires air locks and other equipment. Exhaust systems draw air from these locations without the need for air locks and remove air from the mine through special airways to exhaust fans. Secondary (Auxiliary) Ventilation System The secondary (auxiliary) ventilation system brings sufficient fresh air to the working area from the primary air system without mixing it with the returning contaminated air from the face. a. Use of Ducts The use of compressed air from pneumatic equipment is not recommended for ventilating working faces because insufficient air is supplied, the air discharge location cannot be controlled, and excessive dust is often created [Rock at al. 1971]. Thus a duct must be used in the tunnel leading to the active face to separate fresh from contaminated air. Sometimes two ducts are used, one to supply air to the face and another to remove contaminated air. Air can be either pulled or pushed into the work area, or a combination of the two. b. Blowing Duct System (Push System) The most widely used type of secondary ventilation consists of pushing air through a duct in the access tunnel by means of an auxiliary fan located in the primary air system [Rock at al. 1971]. To be effective in ventilating the face, the end of the duct should come within 25 to 30 feet of the face discharging 2000 cubic feet per minute (CFM) [Bossard et al. 1983]. The duct must be properly placed so that the entire work area is swept with the fresh air. The advantage of pushing air is the large contaminant dilution ventilation provided directly in the miners‘ work area. The air stream will blow across the face because approximately 10% of the duct exit velocity will still exist at a distance equal to 30 duct opening diameters from the duct opening [ACGIH 1984]. The disadvantages include the generation of dust due to the high air velocity needed to blow clean air over the entire working face and the return of contaminated air through the access tunnel used by miners on their way to and from the work area. Another advantage of an air—blowing system is that it increases pressure. Measurements of radon gas content in air exhausted from mines have shown that the radon gas emitted into the atmosphere was 20% less with the air—blowing system than with an exhaust system [Franklin 1981]. This indicates that less radon gas diffused into 180 the ventilated areas when an air—blowing system was utilized than when an exhaust system was used. c. Exhaust Duct System (Pull System) In the exhausting or pulling system, contaminated air is drawn from the working face by a duct that runs from the face to the main passageway in the primary air system through the access tunnel. Fresh air is then drawn into the access tunnel toward the work area by the pressure gradient created by removing air. Exhausting (pulling) air from the face instead of blowing (pushing) offers the following advantages: 0 Fresh incoming air is maintained in the tunnel used by miners to access the active stops, and 0 The contaminated air within 10 feet of the duct is very effectively removed from the work area The major disadvantage of using the exhaust system is that only the area within 10 feet of the end of the exhaust duct is effectively ventilated [Rock et al. 1971]. Work areas further away may receive little air movement. Another disadvantage is that if the air must travel through the access tunnel and drifts that contain ore, then the air becomes contaminated as it is drawn toward the working area. Also, when air is drawn through ducts, the ducts are under negative pressure and thus must be reinforced or rigid to prevent collapsing. Finally the static pressure differential across an exhaust (pulling) system is greater than that across an equivalent blowing (pushing) system with comparable total pressure losses [Rock et al. 1971]. Because exhaust systems have higher static pressures, they are also more prone to leakage. d. Push-Pull System A push—pull system contains two ducts in the accessway, one for pushing clean air to the face and the other for exhausting air from the face back to the primary air system. This system has many of the advantages of both the push and the pull systems including the following: 0 The blowing of air that sweeps across and ventilates the active face, thus providing good dilution in work areas; 0 The efficient collection of contaminants near the work face; and 0 Reducing the contamination of air in the access tunnel. The main disadvantages are the cost and that it occupies more drift area [Rock at al. 1971]. 181 5. Overpressurization and Mine Pumping The amount of radon gas diffusing into mine spaces from interstitial rock is dependent on the pressure in the mine space. The lower the atmospheric pressure in the mine space as compared to the pressure in the interstitial rock, the more radon gas will pass from the rock to the mine space. Conversely, the greater the pressure in the mine space as compared to the rock, the less radon gas will seep into the mine space. Overpressurization and mine pumping are two control measures which take advantage of this principle to reduce concentrations of radon gas. In overpressurization, more ventilation air is pushed into mine spaces than is removed. Although Edwards and Bates [1980] have stated "nothing that we have found provides mining companies with sufficient guidelines for applying the overpressurized ventilation system effectively," they conducted a mathematical study of overpressurization and drew the following conclusion: overpressurization does decrease the radon flux. It was estimated that a 2% pressure differential in a sandstone matrix would result in a 50% reduction in radon flux with mine sink lengths of 100 meters or less. A mine sink is an area either in the mine itself or a naturally occurring space or lattice in the matrix where the interstitial air can flow. If the distance between the sink and the mine space approaches 200 meters, the benefit of overpressurization is lost. Because of the dramatic increase in radon gas in the sink area during overpressurization of work areas, no miners should be allowed in those sinks without proper respiratory protection. However, many open spaces that can serve as sinks are filled in and cannot be occupied. The Bureau of Mines is gathering information on the effects of overpressurization in mines. Data from the pressurization of an enclosed chamber in a mine indicated that the radon concentration was 99% lower than the concentration under static conditions and 92% lower than the concentration under controlled ventilation conditions [Bates and Franklin 1977]. In a study by Schroeder et al. [1966] of mine areas that were pressurized by 10 mm of mercury, the radon flux decreased from 5 to 20 fold as compared to normal ventilation conditions. In mine pumping, a negative pressure is created in the mine space by sealing air intake openings and permitting the exhaust fans to operate. This is done during an offshift when no miners are in the mine. Because of the negative pressure created in the mine with respect to the surrounding rock, radon is drawn into the mine space from the inter— stitial rock at a rate higher than would occur under static conditions. The air intakes must be opened well before miners enter the mine to permit the ventilation system to remove the radon gas and radon progeny that have accumulated in the mine spaces. After this accumulation has been removed, the mine spaces should have lower concentrations of radon gas (and therefore radon progeny) when the miners reenter the mine. This is because much of the radon gas in the surrounding interstitial rock has been removed and is not available to diffuse into the mine working areas. However, monitoring of these areas would be required prior to allowing miners to enter. More studies are needed to determine the effectiveness of this control procedure [Bates and Franklin 1977]. 182 6. Fan Operation Continuous fan operation is essential in a mine for maintaining low radon concentrations during working hours. When the main exhaust ventilation system of 1 mine was shut off, radon gas concentrations increased 1,600% in 3 hours and even after 3 hours of fan operation, the radon gas concentrations failed to return to normal [Franklin et al. 1978]. In the first 5 minutes of a fan shutdown, 1 WL may be exceeded [Musulin et al. 1982]. For fan shutdowns of 15 minutes or more, underground miners should be evacuated to areas with natural downcast ventilation [Musulin et al. 1982]. It has been estimated that at least 2 hours of ventilation should be allowed for each hour of fan shutdown [Franklin et al. 1978]. Spare fans, fan maintenance, and backup electrical systems should be used to minimize shutdown. Bulkheads 1. Description The second most important control measure used in underground mines today is the construction of bulkheads across inactive stopes or drifts [Bates and Franklin 1977]. Bulkheads isolate inactive stopes, prevent the mixture of contaminated air from these stopes with fresh air, and help control the direction of air flow to working areas. Maintaining a negative air pressure behind a bulkhead will prevent leaks [Franklin 1981]; this is important because radon progeny concentrations can exceed 1,000 WL behind a bulkhead [Bates and Franklin 1977]. In addition, bulkheads must be strong and flexible enough to maintain an airtight seal during typical mining conditions, such as the ground movement and air shocks from blasting and the impact from accidental contact with mining equipment. A bulkhead consists of three functional parts: (1) the primary structure, (2) the seal between the primary structure and the rock, and (3) a surface seal on the rock within one meter from the plane of the bulkhead [Summers et al. 1982]. The primary bulkhead structure fills most of the opening in the stope and provides resistance to shocks from blasting or contact with machinery. The primary structure consists of timber or an expanded metal lath covered with a continuous non—porous membrane. The membrane may be attached to, or sprayed upon, the timber in the primary structure. The membrane must not crack or develop holes or leaks during mining activities [Franklin 1981; Summers et al. 1982]. The second part of the bulkhead, the seal between the primary structure and the surrounding rock, must resist running water and the air shocks and rock movements due to blasting. The third part of the bulkhead, the seal on the surface of the rock within one meter from the plane of the bulkhead, must be made of a material that adheres to damp rock surfaces and can withstand mining 183 activities. Summers et al. [1982] tested the efficacy of this procedure and found that the amount of radon gas escaping through the surrounding rock was insufficient to warrant the uniform use of a wall sealant, provided that all cracks, fissures, and holes were sealed to prevent major leaks. 2. Membrane Sealants Used on Bulkheads Summers et al. [1982] evaluated 22 different materials for use as bulkhead sealants, looking at the flammability, health and safety hazards, strength, adhesion, flexibility, and radon gas permeability of each material [Summers et al. 1982]. The two best sealants for bulkheads were a preformed ethylene propylenediene monomeric rubber (EPDM) membrane and Aquafas 48—00®, a water—based mastic. A single sheet of the EPDM membrane was laminated (dry) between two layers of plywood; the Aquafas 48-00® was then troweled and sprayed onto a plywood surface. Summers et al. [1982] concluded that a material‘s permeability to radon gas was less important than its ability to prevent air leaks by its adhesive properties and resistance to tearing or brittle fracture. Steinhausler et al. [1981] recommended polyamide foil as the membrane component in a bulkhead because of its low radon permeability, high strength and flexibility, water resistance, and low cost. 3. Negative Air Pressure Behind a Bulkhead A slight negative pressure behind the bulkhead of about 0.03 cm water with respect to active areas will prevent radon gas leaks into fresh ventilation air [Thomas et al. 1981]. To maintain the negative pressure, a bleeder pipe with a small fan is required to vent a bulkhead or series of bulkheads into the exhaust air [Franklin 1981]. A charcoal trap can efficiently adsorb radon from the bleeder pipe. During an experiment by Summers et al. [1982] a charcoal trap adsorbed 99.8% (calculated) of the high concentration of radon gas (34,249 pCi/l) behind a bulkhead; the activated charcoal worked well regardless of the temperature or humidity in the mine. For a daily evacuation rate of 5%, the average age of air behind a bulkhead is 20 days. Because the half—life of radon gas is 3.8 days, radon gas has time to decay behind a bulkhead [Bloomster et al. 1984b]. 4. Efficiency Because mines differ in the air volume from the worked out areas that can be controlled by bulkheads, reports about the overall efficiency of bulkheads in controlling radon gas emissions vary. Based on experiments in two mines, Bloomster et al. [1984b] estimated that the use of the efficient bulkheads designed by Summers et al. [1982] and the use of carbon filters could reduce radon gas emissions into the atmosphere by 14—80%, depending on the percentage of the mine with bulkheads. When 40—45% of a mine was controlled by bulkheads, Thomas et al. [1981] estimated that traditional bulkheads reduced radon emissions by 30—52% from a test mine. Kown et al. [1980] used a hypothetical mine model to 184 estimate that 100 bulkheads sealing 12.5 stopes would reduce the overall radon gas emissions into mine air by 2.25 Ci/day, a reduction of 25%. In summary, bulkheads are very effective in reducing radon gas (and thus radon progeny) in mine air. Especially promising are the new bulkheads designed by Summers et al. [1982] and further tested by Bloomster et al. [1984b]. These bulkheads may eventually replace the leakier and more flammable polyurethane bulkheads presently being used underground. D. Backfilling In the uranium mining process, large quantities of ore are brought to the surface, leaving voids which may collapse if they are not stabilized. The tailings remaining after the uranium is extracted are often used as backfill. There are three benefits of backfilling stopes: (1) ground stabilization, (2) reducing the ventilation requirements by decreasing the mine volume taken up by air, and (3) allowing the removal of the ore in pillars [Franklin et al. 1982]. The process of backfilling involves three steps [Raghavayya and Khan 1973; Franklin 1981]. First, the coarser fraction of the tailings are separated out by hydrocyclones. Next, the coarse tailings "sand" is mixed with water to form a slurry and pumped into worked-out stopes. Sometimes the slurry is mixed with cement before pumping. After the water in the slurry percolates away, the stope is left filled with densely packed sand or cement. The radon progeny hazard can be increased, at least temporarily, by backfilling. Although the sand has considerably less radium than the ore or host rock, the finely divided sand has a larger surface area and many fine interstices between the grains through which radon gas can move. Therefore, the radon gas emanation rate of the sand is much higher than the ore or host rock [Raghavayya and Khan 1973; Thompkins 1982]. During backfilling, agitation of the slurry releases high concentrations of radon gas [Bates and Franklin 1977]. Also, high concentrations of radon gas can collect above the sand in the newly filled stope (possibly reaching 65,000—75,000 pCi/l). Thus the advantage of decreasing the ventilation volume with the backfill must be weighed against the increased emanation rate of the backfill [Bates and Franklin 1977]. Mixing the slurry with cement will not prevent this increase in emanation rate because the radon gas can also travel freely through fine pores in the cement, especially water—filled pores. lndeed, radon gas emanates from porous cement, sand, or ore at a higher rate when it is wet than when it is dry unless the dry material is overlain with a thick layer of water [Thompkins 1982; Bates and Franklin 1977]. Wet, freshly cemented tailings emanate radon at a high rate that gradually decreases to a steady state as the cement dries [Thompkins 1982]. Although backfilling can produce transient increases in radon gas, it can also be efficacious in reducing overall radon progeny emissions from a mine [Bloomster et al. 1984b; Franklin 1981]. As currently practiced, backfilling can reduce the ventilation volume of a stope by 90% or more. During experiments in a mine, Franklin et al. [1981] found that backfilling 90% of a stope reduced the total radon progeny emissions from the stope by 85%. A feasibility study estimated that backfilling can be as effective as 185 bulkheading in reducing overall radon progeny emissions; however, the cost is much higher [Bloomster et al. 1984b]. Two field studies [Franklin et al. 1982; Raghavayya and Khan 1973] reported extremely high emanation rates from cemented or sand backfill; however, in both instances the backfill was wet; these studies did not report the efficacy of the backfill after it dried. Four methods may improve the efficacy of backfilling [Franklin et al. 1982; Bloomster et al. 1984b]: (1) covering the backfill with one meter of clean sand, (2) sealing the surface of the tailings, (3) using a bulkhead to seal the backfilled stope and maintaining a negative pressure behind the bulkhead, and (4) using nonradioactive materials as backfill instead of mill tailings. in summary, backfilling with uranium tailings can be as effective as bulkheading in reducing radon progeny emissions, although it is more costly. Because high radon progeny concentrations are emitted from wet backfill and during the backfilling process, backfilling should not be used in active mine areas and miners should be protected from overexposure during backfilling operations. E. Sealants Used on Mine Walls This section describes sealants used as diffusion barriers against radon gas, including how the sealants are applied and the best materials used as sealants. Also, the effectiveness of sealants for reducing radon emanation and exposure will be discussed. 1. Description a. Sealant Application Methods Sealant application can involve four steps: (1) clearing the area, (2) applying an undercoating, (3) applying the sealant, and (4) applying an overcoating. First, the labor—intensive step in sealant application is clearing the area of loose rock to provide a smooth surface before applying the sealant [Lindsay et al. 1981a]. Shotcrete or Gunite is troweled into any large cracks in the surface [Franklin et al. 1977]. Second, an undercoating of shotcrete or Gunite is troweled or sprayed onto the prepared surface. The undercoating alone will not act as an effective barrier to radon gas, although it does provide a smooth supporting surface for the fragile sealant coating [Lindsay et al. 1981a; Franklin et al. 1977]. Third, a layer of sealant, usually an acrylic polymer, is sprayed or placed on the undercoating. The undercoating and sealant coating should be different colors to ensure complete coverage [Franklin 1981]. Fourth, an overcoating should be applied; the overcoating can be a second layer of sealant, another type of sealant, a layer of shotcrete, or some combination of these materials [Franklin 1981; Steinhausler et al. 1981]. An outer layer of shotcrete protects the sealant surface against mechanical damage in the mine. 186 b. Sealant Materials The ideal sealant material should meet a variety of criteria to resist conditions within a mine [Franklin et al. 1975a; Steinhausler et al. 1981]. The sealant material should: 0 Reduce the radon emission rate by 50% or more; 0 Be easily applied (i.e., sprayable); 0 Lack toxic vapor emissions during application or curing; 0 Resist flame, fire, and water; 0 Tolerate wide changes in temperature; 0 Cure in a mining environment (40—600F, 40—100% relative humidity); 0 Possess mechanical strength and flexibility; and 0 Lack electrical hazards (such as found with metal foils). The sealants that performed well in a variety of laboratory and mine experiments [Franklin et al. 1975a, 1975b; Lindsay et al. 1981a, 1981b; Steinhausler et al. 1981; Summers et al. 1982] included: 0 Hydro Epoxy 156®, a two—component, water-based epoxy; O Hydro Epoxy 300®, a water—based epoxy; O VMX—50 VML®, a water emulsion acrylic latex; I VMX—5O BMT®, a water emulsion acrylic latex; 0 Polyamide foil sheets; 0 Ethylene propylenediene monomer (EPDM) rubber membrane 1.12 mm thick, often used as a roof sealant; and 0 Aquafas 48—00®, a water-based mastic. The estimated lifetime for sealants is approximately 5 to 7 years [Bloomster et al. 1984b]. In addition, the best use of sealants occurs in areas either with high radon emanation rates or where there is little chance of damage from mining activities; these include mined—out areas, lunchrooms, shops, intake airways, and inactive stopes [Kown et al. 1980; Franklin et al. 1980; Bloomster et al. 1984b]. Edwards and Bates [1980] of the Bureau of Mines developed a computer model to evaluate the effect of pinholes in an otherwise impermeable sealant. They concluded that pinholes (2 mm in diameter) were not a 187 problem unless there were several thousand visible pinholes per square meter of sealant. 2. Effectiveness of Sealants in Reducing Radon Emanation Rates The effectiveness of sealants depends, in part, upon the porosity of the rock walls. Sealants produce the greatest decrease in radon emanation when applied to sandstone or other porous rock; sealants applied to granite will appear to be less effective because granite provides a natural barrier to radon emanation [Lindsay et al. 1981a]. Thus results from the tests for the effectiveness of sealants vary greatly depending on the porosity of the rock walls along with the mine ventilation rate, the grade of the uranium ore, and other factors. Five field experiments conducted by the Bureau of Mines between 1975 and 1981 showed that sealants (in this case, water—based epoxy and water—based acrylic latex) reduced radon gas emanation by 50—75% [Franklin 1981]. An acrylic latex sealant and a Gunite (dry—mix concrete) undercoating reduced the radon emanation by 75% when applied to sandstone [Lindsay et al. 1981a, 1981b]. Bloomster et al. [1984b] estimated that the overall decrease in radon emissions would be 56% if the same sealant coating was applied to 80% of the mine surfaces. Although sealants were less effective and more costly than bulkheads, the use of sealants is less disruptive to the mining process than the use of bulkheads [Bloomster et al. 1984b]. In summary, there are at least seven materials available that make effective mine sealants. These materials can reduce radon emanation from mine walls by 50-75%. F. Controlling Radioactive Water Underground Radium—, barium—, or radon-bearing waters cause radon control problems in underground mines. Radon—bearing water releases radon gas until the concentration in air reaches approximately three times the concentration in water [Thompkins 1982]. Open ditches, sumps, or other water accumulations near intake airways or active areas cause unnecessary contamination [Rock and Walker 1970]. Both iron mines in Sweden and coal mines in Poland use ventilation as the primary control for radon released from water. A coal mining company in Poland reduced radon concentrations in air by precipitating out the sulfate salts of barium and radium in the water [Tomza and Lebecka 1981]. A uranium company in Sweden proposes to drill boreholes around the mine and pump up the radon-bearing water into a nearby lake. This method is untested [Snihs 1981]. Thompkins [1982] suggests that placing a ring of wells around a uranium ore body, under a vacuum of 4 psi, and lowering the water table may reduce radon concentrations in a uranium mine by causing air and gas to flow to the wells. 188 G. Automation Another radon progeny control method is increased automation. Techniques such as robotics that minimize the time the miner spends in the high exposure areas of the mine and in activities such as drilling, blasting, or loading ore, will decrease the miner's radiation exposure. Although, at present, robotics has a limited place in the mines, it may be possible in the future to further automate the ore mining process. 189 REFERENCES ACGIH [1984]. Industrial ventilation——A manual of recommended practice. 18th edition. Cincinnati, OH: American Conference of Governmental Industrial Hygienists. Bates RC, Franklin JC [1977]. U.S. Bureau of Mines radiation control research. Proceedings of the Conference on Uranium Mining Technology, Reno NV, April 25—29. Bloomster CH, Enderlin WI, Young JK, Dirks JA [1984a]. Cost survey for radon daughter control by ventilation and other control techniques. Volume I. Richland, WA: Battelle Memorial Institute, Pacific Northwest Laboratories, NTIS PBB5-152932. Bloomster CH, Jackson P0, Dirks JA, Reis JW [1984b]. Cost survey for radon daughter control by ventilation and other control techniques: radon emissions from underground uranium mines. Volume II. Richland, WA: Battelle Memorial Institute, Pacific Northwest Laboratories, NTIS PB85—152940. Bossard FC, LeFever JJ, LeFever JB, Stout KS [1983]. A manual of mine ventilation design practices. Second edition. Butte, MT: Floyd C. Bossard and Associates, Inc. Cooper WE [1981]. A comparison of radon—daughter exposures calculated for U.S. underground uranium miners based on MSHA and company records. In: Gomez M, ed. International Conference: Radiation Hazards in Mining, New York, NY: Society of Mining Engineers of American Institute of Mining, Metallurgical, and Petroleum Engineering, Inc., pp. 292—295. Edwards JC, Bates RC [1980]. Theoretical evaluation of radon emanation under a variety of conditions. Health Phys 89:263—274. Ferdinand FW, Cleveland JE [1984]. Philosophies and practices in radiation monitoring and exposure calculations in modern uranium mines. Unpublished report submitted to the National Institute for Occupational Safety and Health by the Quivira Mining Company. Franklin JC [1981]. Control of radiation hazards in underground uranium mines. In: Gomez M, ed. International Conference: Radiation Hazards in Mining. New York, NY: Society of Mining Engineers of American Institute of Mining, Metallurgical, and Petroleum Engineering, Inc., pp. 441—446. Franklin JC, Bates RC, Habberstad JL [1975a]. Polymeric sealants may provide effective barriers to radon gas in uranium mines. E/MJ, September 1975, pp. 116—118. Franklin JC, Nuzum LT, Hill AL [1975b]. Polymeric materials for sealing radon gas into the walls of uranium mines. Washington, DC: U.S. Department of the Interior, Bureau of Mines, Report of Investigations 8036. 190 Franklin JC, Meyer T0, Bates RC [1977]. Barriers for radon in uranium mines. U.S. Department of the Interior, Bureau of Mines, Report of Investigations 8259. Franklin JC, Meyer TO, McKibbin RW, Kerkering JC [1978]. A continuous radon survey in an active uranium mine. Mining Engineering, pp. 647-649. Franklin JC, Musulin CS, Bates RC [1980]. Monitoring and control of radon hazards. Proceedings of the Second International Mine Ventilation Congress, Reno, NV, pp. 405—411. Franklin JC, Weverstad KD, Black JW, Cleveland JE [1981]. Radiation hazards in backfilling with classified uranium mill tailings. Fifth Annual Cent. New Mexico AIME Uranium Seminar, Albuquerque, NM. Franklin JC, Washington RA, Kerkering JC, Montone H, Regan R [1982]. Radiation emanation from stopes backfilled with cemented uranium mill tailings. Washington, DC: U.S. Department of the Interior, Bureau of Mines, Report of Investigations 8664. Kown BT, Van der Mast VC, Ludwig KL [1980]. Technical assessment of radon—222 control technology for underground uranium mines. Washington, DC: U.S. Environmental Protection Agency, Office of Radiation Programs (ANR—458). Technical Note 0RP/TAD—80—7, Contract No. 68—02—2616, Task No. 9. Lindsay DB, Oberholtzer JE, Summers CH [1981a]. Sealant tests to control radon emanation in a uranium mine. Washington, DC: U.S. Department of the Interior, Bureau of Mines, Report No. 83836. Lindsay DB, Schroeder GL, Summers CH [1981b]. Polymeric wall sealant test for radon control in a uranium mine. In: Gomez M, ed. International Conference: Radiation Hazards in Mining. New York, NY: Society of Mining Engineers of American Institute of Mining, Metallurgical, and Petroleum Engineering, Inc., pp. 790—793. Lundin FE, Wagoner JK, Archer VE [1971]. Radon daughter exposure and respiratory cancer: quantitative and temporal aspects. U.S. Department of Health, Education, and Welfare, Public Health Service, National Institute for Occupational Safety and Health, National Institute of Environmental Health Sciences, Joint Monograph No. 1. Musulin CS, Franklin JC, Roberts FA [1982]. Effects of a fan shutdown on radon concentration in a positive pressure ventilated mine. Washington, DC: U.S. Department of the Interior, Bureau of Mines, Report of Investigations 8738. Ragavayya M, Khan AH [1973]. Radon emanation from uranium mill tailings used as backfill in mines. In: Stanley RE and Moghissi AA, eds. Noble Gases. Las Vegas, NV: U.S. Environmental Protection Agency, National Environmental Resource Center, Office of Research and Development, EPA 600/9—76—026, pp. 269—273. 191 Rock RL, Walker DK [1970]. Controlling employee exposure to alpha radiation in underground uranium mines. Volume 1. Washington, DC: U.S. Department of the Interior, Bureau of Mines. Rock RL, Dalzell RW, Harris EJ [1971]. Controlling employee exposure to alpha radiation in underground uranium mines: appendixes for the use of radon—daughter control specialists. Volume 2. U.S. Department of the Interior, Bureau of Mines. Schroeder GL, Evans RD, Kraner HW [1966]. Effect of applied pressure on the radon characteristics of an underground mine environment. AIME Transaction, p. 235. Snihs J0 [1981]. Radiation protection in Swedish mines: special problems. In: Gomez M, ed. International Conference: Radiation Hazards in Mining. New York, NY: Society of Mining Engineers of American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc., pp. 283—285. Steinhausler F, Pohl—Ruling J, Pohl E [1981]. Control of radon daughter concentration in mine atmospheres with the use of radon diffusion barriers. In: Gomez M, ed. International Conference: Radiation Hazards in Mining. New York, NY: Society of Mining Engineers of American Institute of Mining, Metallurgical, and Petroleum Engineering, Inc., pp. 127—130. Summers CH, Lindsay DB, Lindstrom RS, Alder KL [1982]. Evaluation of bulkheads for radon control. Washington, DC: U.S. Department of the Interior, Bureau of Mines, Minerals Health and Safety Technology. Minerals Research Contract Report No. 86396, pp. 8-12, 17—24, 52—55, 75—76. Thomas VW, Musulin CS, Franklin JC [1981]. Bulkheading effects on radon release from the Twilight Uranium Mine. Richland, WA: Battelle Memorial Institute, Pacific Northwest Laboratory, PNL-3693/UC—11. Thompkins RW [1982]. Radiation in uranium mines: a manual of radon gas emission characteristics and techniques for estimating ventilating air requirements. Part 2. CIM Bulletin Z§(846):62—67. Tomza I, Lebecka J [1981]. Radium—bearing waters in coal mines: occurrence, methods of measurement and radiation hazard. In: Gomez M, ed. International Conference: Radiation Hazards in Mining. New York, NY: Society of Mining Engineers of American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc., pp. 945—948. 192 APPENDIX IV GRAB SAMPLING STRATEGY REQUIREMENTS FOR DETERMINATION OF RADON PROGENY EXPOSURES A. Introduction Airborne concentrations of radon progeny must be monitored regularly to provide the basis for their control. Miners' exposures must be limited to no more than 1.0 WLM per year and the average concentration of radon progeny in any work area must not exceed 1/12 WL during any work shift. The sampling strategy described here was developed after an evaluation of mine sampling data and the typical variability of radon progeny concentrations in underground mines. This strategy will allow the collection of timely and reliable environmental data that can be used as the basis for control of cumulative exposures. This sampling strategy allows for the determination of the arithmetic average of time—varying concentrations of radon progeny during a work shift in a given work area. The determination is based on an unbiased estimate made from grab samples taken at random intervals through— out the work shift. Random sampling of work shifts during a reference period is also included for determination of a long-term arithmetic average work shift concentration. The formulae needed to calculate the statistical quantities used in this sampling strategy are contained in section G of this appendix. The rationale for the critical decision points used in the sampling strategy are contained in section H. B. Definition of Terms and Notations STATION: A sampling location within a work area that represents the radon progeny concentration to which miners are exposed. CLUSTER: Two or more stations at which sampling will be conducted during any work shift. The stations in a cluster should be located at different work areas but must be in close proximity to each other so that alternating grab samples could be taken during the same work shift. BLOCK OF TIME: A period in which two different sampling days are randomly selected. AVERAGE WORK SHIFT CONCENTRATION: The average concentration of radon progeny in working levels (WL) during a work shift at a given station. AVERAGE: The arithmetic mean. The same term can be used for the average of several sample results or for the arithmetic mean of a distribution of concentrations that vary during a continuous period of time. In the latter case, the terms ”average," "arithmetic average," and "time—weighted average" are synonymous. 193 (Xi: A. . 0(1+1- (1.: A n «.11. LCL: UCL : Average work shift concentration for day i, where i = 1,2, ..,12 and day i is the ith day in a time—ordered sequence of the 12 days that were randomly selected from the reference period. Estimate of ai (based on seven grab samples), where i = 1,2,. .,12 and day i is the ith day in a time—ordered sequence of the 12 days that were randomly selected from the reference period. Estimated average work shift concentration for day i+1 (based on seven grab samples), where i = 1,2,. .,11 and day i+1 is the next sampling day following day i in a time—ordered sequence of the 12 days that were randomly selected from the reference period. Estimated average work shift concentration for day A (based on seven grab samples), where "day A" is a term used to designate one of the 12 randomly selected days sampled during the reference period for which the ai value is in a critical range. Estimated average work shift concentration for day B (based on seven grab samples), where day B is the first workday following day A. [Note: Day B may or may not be the next calendar day after day A since a weekend, holiday, or other non-workday may occur between days A and 3.] Long—term average work shift concentration during a reference period from which 12 sampling days were randomly selected. Estimated long—term average work shift concentration (based on seven grab samples per sampling day) during the reference period from which 12 sampling days were randomly selected. Estimated average work shift concentration (based on seven grab samples) on the 11th of the 12 randomly selected days sampled during the reference period. Estimated average work shift concentration (based on seven grab samples) on the last of the 12 randomly selected days sampled during the reference period. 95% one—sided lower confidence limit for a. 95% one—sided upper confidence limit for a. 194 Requirements for Routine Exposure Monitoring 1. Two different sampling days are randomly selected from each 2-week block of time. 2. The stations within a cluster are to be sampled on the same workdays and work shifts. All stations within a cluster are to be alternately sampled, seven times on each sampling day, each time in independent random order. During the work shift, the seven periods for sampling of the entire cluster shall be equally spaced in time. For example, the three stations A, B, and C could be considered a cluster and sampled as ABC, BCA, ACB, CBA, CAB, BAC, and ACB during seven successive intervals of approximately equal durations. If it is not feasible to sample in this manner, then sampling can be conducted along the most efficient path but with a different, randomly determined starting point on each day (e.g., BCA, BCA,..., BCA during one sampling day and ABC, ABC,..., ABC or CAB, CAB, .., CAB during other sampling days). 3. The estimated average work shift concentration (0,) for each sampling day (i = 1,2,...12) is computed from an analysis of the seven grab samples taken on that day. Formulae for this computation are contained in section G. 4. Whenever Qi for a particular station exceeds 0.14 WL, then that station shall be resampled the next workday. [Note: in this case, i = 9A, and sampling on the "next workday" (day B) is in addition to the two randomly selected sampling days required in a 2—week block of time.] a. If $3 (the estimated average work shift concentration on the next workday) is 5 0.14 WL, then exposure monitoring shall continue as described starting at section C,1. b. 1f QB also exceeds 0.14 WL, then: (1) steps shall be taken to reduce the radon progeny concentration in that work area by implementing work practices and engineering controls, (2) respiratory protection shall be required for all miners entering that work area, and (3) grab sampling as described in section C,2 shall be conducted on a consecutive daily basis. Grab sampling shall continue on a consecutive daily basis until the estimated average work shift concentrations on any two consecutive workdays (9A and 03) are both 5 0.10 WL. When QA and QB are both 5 0.10 WL, then the requirements for respiratory protection are waived and exposure monitoring can revert to the schedule described starting at section C,1. [Note: A new reference period shall begin at this time, requiring 12 randomly selected sampling days, the first of which is to be coded as i = 1.] This criterion (as discussed in section H,2) serves to provide early confirmation that the corrective steps taken by the mine operator have been effective in limiting the average work shift concentration of radon progeny to a level not exceeding 1.5 times the recommended exposure limit (REL) of 1/12 WL. 195 D. 5. If &i is g 0.14 WL, then: (a) continue collecting seven grab samples on each of the two randomly selected sampling days in each 2—week block of time, and (b) continue using the criteria given in section C. After 12 weeks of sampling in which no two consecutive sampling days (A; and Qi+1) were in excess of 0.14 WL, use the criteria given in section D for assurance, based on 12 days of sampling, that the average work shift concentration of radon progeny is in compliance with the REL, which, if verified, will result in less frequent exposure monitoring requirements. Criteria for Less Frequent Exposure Monitoring To determine if less frequent exposure monitoring can be conducted at a specific work area, the following statistical decision criteria must be used: 1. Compute Q. (the estimated average work shift concentration using seven grab samples per sampling day) for a work area during the reference period in which 12 samples were taken and no two consecutive sampling days (Qi and Qi+1) were in excess of 0.14 WL. Formulae for this computation are contained in section G. 2. Compute LCL and UCL, the 95% one—sided lower and 95% one—sided upper confidence limits, respectively, for the average work shift concentration during the reference period from which the 12 sampling days were taken. Formulae for these computations are contained in section G; Q. from section 0,1 is a quantity used in the formulae for LCL and UCL. 3. The block length can be increased from 2 weeks to 26 weeks (therefore requiring only 2 randomly selected sampling days per 26—week block of time) if bgth of the following results occur at a station: (a) UCL (for the average work shift concentration during the 12—week period) is 5 1/12 WL, and (b) the estimated average work shift concentrations on any two consecutive randomly selected sampling days (&i and Qi+1) within the same reference period did not exceed 0.14 WL. Criteria for the continuation of less frequent exposure monitoring and for the cessation of exposure monitoring are given in parts E and F, respectively. 4. If LCL exceeds 1/12 WL, then: (a) steps shall be taken to reduce the radon progeny concentration in that work area by implementing work practices and engineering controls, (b) respiratory protection shall be required for all miners entering that work area, and (c) grab sampling as described in section C,2 shall be conducted on a consecutive daily basis. Grab sampling shall continue on a consecutive daily basis until the estimated average work shift concentrations on any two consecutive workdays (AA and Q3) are both 5 0.10 WL. When QA and QB are both 5 0.10 WL, then the requirements for respiratory protection are waived and exposure monitoring can revert to the schedule described starting at section C,1. [Note: A new reference period shall begin at this time, requiring 12 randomly selected sampling days, the first of which is to be coded as i = 1.] 196 E. Criteria for Continuation of Less Frequent Exposure Monitoring After completion of two additional sampling days during the subsequent 26—week period, the data from the last 12 days sampled must be used to compute a new UCL for the period in which the 12 sampling days occurred. F. 1. Sampling may continue under the less frequent sampling schedule (i.e., 2 days per 26—week block of time) if both of the following results occur at a station: (a) UCL for the reference period from which the last 12 sampling days were taken is 5 1/12 WL, gag (b) the estimated average work shift concentrations on the last two of the 12 sampling days (311 and 912) were both 5 0.14 WL. In this case, an updated UCL shall be recomputed after completion of sampling in each subsequent 26—week block of time to determine if less frequent sampling (i.e., on two days during a 26—week period) should be continued according to the criteria of this part. If either of these conditions are ggt met, then LCL must be computed from data obtained from the last 12 days sampled (see section E,2 which follows). 2. If LCL for the reference period from which the 12 sampling days were taken at a station exceeds 1/12 WL, then: (a) steps shall be taken to reduce the radon progeny concentration in that work area by implementing work practices and engineering controls, (b) respiratory protection shall be required for all miners entering that work area, and (c) grab sampling as described in section C,2 shall be conducted on a consecutive daily basis. Grab sampling shall continue on a consecutive daily basis until the estimated average work shift concentrations on any two consecutive workdays (AA and Q3) are both 3 0.10 WL. When QA and Q3 are both 3 0.10 WL, then the requirements for respiratory protection are waived and exposure monitoring can revert to the schedule described starting at section C,1. 3. If LCL for the reference period from which the 12 sampling days were randomly taken is S 1/12 WL, but the estimated average work shift concentration determined for either of the last two of the 12 sampling days (Q11 or Q12) exceeds 0.14 WL, then monitoring at that station shall return to the more frequent sam ling schedule (2 days per 2—week block of time). In this case, Q11 or 812 becomes QA and sampling is required on the next workday to obtain &B: as described starting at section C,4,a. Criteria for Cessation of Exposure Monitoring Sampling can be discontinued at a station if both of the following results occur at that station: (1) UCL for the reference period from which 12 sampling days were taken is g 0.063 WL, aflg (2) the estimated average work shift concentration for the last of the 12 sampling days (Q12) is g 0.033 WL. However, sampling should return to the regular schedule, as described starting at section C,1 if an environmental change or a change in mining operations occurs that may alter radon progeny concentrations in that work area. 197 G. Statistical Considerations and Data Analysis Formulae The following are the statistical notations used in the sampling strategy: cij = measured concentration of radon progeny in the jth grab sample taken on the ith sampling day, where j = 1,2,...,7 for each day and i = 1,2,...,12 (2 workdays selected at random from each of six consecutive blocks of time). cAj = measured concentration of radon progeny in the jth grab sample taken on day A, where j = 1,2,...,7 ch = measured concentration of radon progeny in the jth grab sample taken on the next workday following day A, where j = 1,2,...,7. Xij = natural logarithm of Cij = In Cij XAj = natural logarithm of CAj = ln CAj xBj = natural logarithm of ch = In ch xi = average of the 7 Xij values for the 7 grab samples taken on day i. 7 = (1/7) 'EXij i=1 x. = average of the 12 xi values during the reference period from which 12 sampling days were randomly selected. 12 = (1/12) Xxi i=1 xA = average of the 7 XAj values for the 7 grab samples taken on day A. 7 = (1/7) .ZXAj i=1 x3 = average of the 7 XB' values for the 7 grab samples taken on day B, where day B is the next workday following day A. 7 = (1/7) XXBj i=1 x11 = average of the 7 x11 - values (natural logarithms) for the 7 grab samples taken'on the 11th of the 12 randomly selected sampling days in the reference period. 7 = (1/7) 2X11'j i=1 198 SL (1.. LCL UCL oli+1 average of the 7 x12’- values (natural logarithms) for the 7 grab samples taken on the last of the 12 randomly selected sampling days in the reference period. 7 (1/7) 2X12’j i=1 standard deviation (of daily averages of logarithms) computed from the 12 xi's for the 12 days sampled during the reference period. 12 [(1/11) 2(xi _ x.)2](1/2) i=1 long—term average work shift concentration during the reference period from which 12 sampling days were randomly selected. estimate of a. (based on seven grab samples per sampling day) during the reference period from which 12 sampling days were randomly selected. exp [x. + 0.5 sL ] 95% one—sided lower confidence limit for a. &./[1+1.796 sL (0.1 + 0.05 sL 2)(1/2)] 95% one—sided upper confidence limit for a. &./[1—1.796 sL (0.1 + 0.05 sL 2Wm] average work shift concentration for day i, where i = 1,2,. .,12 and day i is the ith day in a time—ordered sequence of the 12 days that were randomly selected from the reference period. estimate of a; (based on seven grab samples), where i = 1,2,...,12 and day i is the ith day in a time—ordered sequence of the 12 days that were randomly selected from the reference period. exp [xi + 0.5 (1 — 1/7) In2(1.3335)] = 1 036 exp [xi] estimated average work shift concentration for day i+1 (based on seven grab samples), where i = 1,2,.. ,11 and day i+1 is the next sampling day following day i in a time—ordered sequence of the 12 days that were randomly selected from the reference eriod. exp [xi+1 + 0.5 (1 — 1/7) in (1.3335)] = 1.036 exp [xi+1] estimated average work shift concentration for day A (based on seven grab samples), where "day A" is a term used to designate one of the 12 randomly selected days sampled during the reference period for which the a; value is in a critical range. exp [2A + 0.5 (1 — 1/7) |n2(1.3335)] = 1 036 exp [2A] 199 QB = estimated average work shift concentration for day B (based on seven grab samples), where day B is the first workday following day A. [Note: Day B may or may not be the next calendar day after day A since a weekend, holiday, or other non-workday may occur between days A and B.] = exp [>73 + 0.5 (1 — 1/7) In2(1.3335)] = 1.036 exp [x3] Q11 = estimated average work shift concentration (based on seven grab samples) on the 11th of the 12 randomly selected days sampled during a reference Eeriod. = exp [x11 + 0.5 (1 - 1/7) In (1.3335)] = 1.036 exp [x11] Q12 = estimated average work shift concentration (based on seven grab samples) on the last of the 12 randomly selected days sampled during a reference eriod. = exp [212 + 0.5 (1 — 1/7) In (1.3335)] = 1.036 exp [212] H. Rationale for the Critical Points in the Sampling Strategy NIOSH recognizes that the concentration of radon progeny in any work area varies with time. Therefore, exposure estimates based on one or even several grab samples may not provide an accurate measurement of the average work shift concentration. Nevertheless, NIOSH believes that by using estimates of radon progeny concentrations determined from grab sampling measurements, it is possible to determine (with at least 95% confidence) that the long—term average work shift concentration would not exceed 1/12 WL by more than a factor of 3.15, based on exposure data derived from seven grab samples taken during a single work shift (éi). This factor can be reduceg to 1.43 if an estimate of exposure were used based on 12 sampling days ( .). The estimates Qi, 9A, and Q3 for a single work shift's average concentration of radon progeny are based on an assumed log-normal distribution of intraday concentration variations with a geometric standard deviation (GSD) of 1.3335. An assumed log—normal interday distribution with a GSD of 1.3926 was used to calculate critical values of estimates to test hypotheses about the long—term average work shift concentration. The stated GSDs were computed from published historic data on intraday and interday variability of radon progeny concentrations in uranium mines [Johnson 1978]. Other data sets were examined; however, they were not suitable for estimating intraday and interday exposure variabilities that were unaffected by location. The interday and intraday variations in concentrations were modeled as independent log—normal distributions, based on general models for determining occupational exposure concentrations reported elsewhere [Bar—Shalom et al. 1975; Leidel et al. 1975, 1977]. 1. Initial Compliance with the REL Based on analysis of the Johnson [1978] data set, 0.14 WL was calculated to be the 95th percentile of a log-normal model for the distribution of the estimated daily average work shift concentrations ( i's) when the long-term average work shift concentration (a.) was 1/12 WL. Thus when the estimated average work shift concentrations are greater 200 than 0.14 WL on two consecutive workdays, substantial evidence exists that the long—term average work shift concentration exceeds 1/12 WL. Therefore, when QA and AB both exceed 0.14 WL in a work area, NIOSH recommends that radon progeny concentrations be reduced in that work area by implementing work practices and engineering controls, and that the use of respiratory protection be required for all miners entering that work area. These recommendations are also made when the 95% lower confidence limit for the long—term average work shift concentration (LCL) exceeds 1/12 WL (see section 0,4). 2. Return to Compliance with the REL The NIOSH sampling strategy uses criteria with approximately 90% confidence for an initial determination that a work area is tentatively back to compliance with the REL. Specifically, estimated average work shift concentrations from two consecutive workdays (i.e., QA and QB) in which both are 5 0.10 WL was chosen as a criterion that demonstrates reasonable evidence that the average radon progeny concentration is being controlled to 5 0.125 WL (i.e., 1.5 times the REL). Given the levels of intraday and interday variabilities observed in the Johnson [1978] data set, a work area with an average work shift concentration of 0.125 WL (i.e., 50% above the REL of 1/12 WL) has 0.90 probability to have one or both of a pair of consecutive estimated average work shift concentrations above 0.10 WL. This "2—day" decision rule limits the magnitude with which a work area's average work shift concentration may exceed 1/12 WL and be undetected. This rule also has the advantage of permitting an early return to normal operations after a period of corrective actions to reduce exposure concentrations, at the expense of having less than high confidence that the REL is not being exceeded by more than 50%. However, only a small proportion of time passes until the next sampling day (as specified in the sampling strategy relative to the year), so that the 2—day rule limits the contribution of a temporarily excessive exposure in a work area to a miner's cumulative annual exposure. At a later time, the lower confidence limit criterion for noncompliance determined after 12 randomly selected sampling days (i e., LCL > 1/12 WL) would be likely to detect a statistically significant increase above the REL if the long—term average work shift concentration were as high as 0.125 WL. 3. Less Frequent Exposure Monitoring The upper confidence limit criterion (i.e., UCL 5 1/12 WL) gives 95% confidence that the long—term average work shift concentration is not above 1/12 WL, under the assumption that ai's exhibit log—normally distributed random variations. The additional requirement that Qi and Qi+1 do not exceed 0.14 WL is meant to detect a temporarily or periodically high average work shift concentration (i.e., high Qi's that are not sustained for the full block of time from which 12 sampling days were selected). When both of these requirements are met, only 2 randomly selected sampling days are then required per 26—week block of time. 201 4. Cessation of Exposure Monitoring UCLg 0.063 WL gives greater than 95% confidence that the long-term average work shift concentration (a.) is g 0.063 WL (i.e., a. is no larger than 75% of the REL), under the assumption that Qi's exhibit log—normally distributed random variations. Under the additional assumption that geometric standard deviations (GSDs) for intraday and interday (log-normal) variability are similar to those_ reported in Johnson [1978], the criterion that Q12 (the estimated average work shift concentration on the last of the 12 sampling days) be 5 0.033 WL gives 95% confidence that a projected future reference period would have a long—term average work shift concentration 3 0.063 WL. 202 REFERENCES Bar—Shalom Y, Budenaers D, Schainker R, Segall A (1975). Handbook of statistical tests for evaluating employee exposure to air contaminants. Cincinnati, OH: U.S. Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control, National Institute for Occupational Safety and Health, Division of Laboratories and Criteria Development. DHEW (NlOSH) Publication No. 75—147. Johnson JR (1978). Uncertainties in estimating working level months. Chalk River, Ontario: Atomic Energy of Canada Limited, Chalk River Nuclear Laboratories, AECL—6402. Leidel NA, Busch KA, Crouse WE (1975). Exposure measurement action level and occupational environmental variability. Cincinnati, OH: U.S. Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control, National Institute for Occupational Safety and Health, Division of Laboratories and Criteria Development. DHEW (NlOSH) Publication No. 76—131. Leidel NA, Busch KA, Lynch JR (1977). Occupational exposure sampling strategy manual. Cincinnati, OH: U.S. Department of Health and Human Services, Public Health Service, Center for Disease Control, National Institute for Occupational Safety and Health. DHEW (NlOSH) Publication No. 77-173. 203 APPENDIX V MEDICAL ASPECTS OF WEARING RESPIRATORS* In recommending medical evaluation criteria for respirator use, one should apply rigorous decision—making principles [Halperin et al. 1986]; tests used should be chosen for operating characteristics such as sensitivity, specificity, and predictive value. Unfortunately, many knowledge gaps exist in this area. The problem is complicated by the large variety of respirators, their conditions of use, and individual differences in the physiologic and psychologic responses to them. For these reasons, the following guidelines are to be considered as informed suggestions rather than established NIOSH policy recommendations. They are intended primarily to assist the physician in developing medical evaluation criteria for respirator use. A. Background Information Brief descriptions of the health effects associated with wearing respirators are summarized below. More detailed analyses of the data are available in recent reviews by James [1977] and Raven et al. [1979]. 1. Pulmonary Effects In general, the added inspiratory and expiratory resistances and dead space of most respirators cause an increase in tidal volume and a decrease in respiratory rate and ventilation (including a small decrease in alveolar ventilation). These respirator effects have usually been small both among healthy individuals and, in limited studies, among individuals with impaired lung function [Gee et al. 1968; Altose et al. 1977; Raven et al. 1981; Hodous et al. 1983; Hodous et al. 1986]. This generalization is applicable to most respirators when resistances (particularly expiratory resistance) are low [Bentley et al. 1973; Love et al. 1977]. While most studies report minimal physiologic effects during submaximal exercise, the resistances commonly lead to reduced endurance and reduced maximal exercise performance [Craig et al. 1970; Raven et al. 1977; Stemler and Craig 1977; Myhre et al. 1979; Deno et al. 1981]. The dead space of a respirator (reflecting the amount of expired air that must be rebreathed before fresh air is obtained) tends to cause increased ventilation. At least one study has shown substantially increased ventilation with a full—face respirator, a type that can have a large effective dead space [James at al. 1984]. However, the net effect of a respirator's added resistances and dead space is usually a small decrease in ventilation [Craig et al. 1970; Hermansen et al. 1972; Raven et al. 1977; Stemler and Craig 1977; Deno et al. 1981; Hodous et al. 1983]. The potential for adverse effects, particularly decreased cardiac output, from the positive pressure feature of some respirators has been *Adapted from NIOSH Respiratory Decision Logic [NIOSH 1987]. 204 reported [Meyer et al. 1975]. However, several recent studies suggest that this is not a practical concern, at least not in healthy individuals [Bjurstedt et al. 1979; Arborelius et al. 1983; Dahlback and Balldin 1984]. Theoretically, the increased fluctuations in thoracic pressure caused by breathing with a respirator might constitute an increased risk to subjects with a history of spontaneous pneumothorax. Few data are available in this area. While an individual is using a negative-pressure respirator with relatively high resistance during very heavy exercise, the usual maximal—peak negative oral pressure during inhalation is about 15—17 cm of water [Dahlback and Balldin 1984]. Similarly, the usual maximal—peak positive oral pressure during exhalation is about 15—17 cm of water, which might occur with a respirator in a positive—pressure mode, again during very heavy exercise [Dahlback and Balldin 1984]. By comparison, maximal positive pressures such as those during a vigorous cough can generate 200 cm of water pressure [Black and Hyatt 1969]. The normal maximal negative pleural pressure at full inspiration is -40 cm of water [Bates et al. 1971], and normal subjects can generate —80 to —160 cm of negative water pressure [Black and Hyatt 1969]. Thus while vigorous exercise with a respirator does alter pleural pressures, the risk of barotrauma would seem to be substantially less than that of coughing. In some asthmatics, an asthmatic attack may be exacerbated or induced by a variety of factors including exercise, cold air, and stress, all of which may be associated with wearing a respirator. While most asthmatics who are able to control their condition should not have problems with respirators, a physician's judgment and a field trial may be needed in selected cases. 2. Cardiac Effects The added work of breathing from respirators is small and could not be detected in several studies [Gee et al. 1968; Hodous et al. 1983]. A typical respirator might double the work of breathing (from 3% to 6% of the total oxygen consumption), but this is probably not of clinical significance [Gee et al. 1968]. In concordance with this view, several other studies indicated that at the same workloads heart rate does not change with the wearing of a respirator [Raven et al. 1982; Harber et al. 1982; Hodous et al. 1983; Arborelius et al. 1983; Petsonk et al. 1983]. In contrast, the added cardiac stress due to the weight of a heavy respirator may be considerable. A self—contained breathing apparatus (SCBA) may weigh up to 35 pounds. Heavier respirators can reduce maximum external workloads by 20% and similarly increase heart rate at a given submaximal workload [Raven et al. 1977]. In addition, it should be noted that many uses of SCBA (e.g., for firefighting and hazardous waste site work) also necessitate the wearing of 10—25 pounds of protective clothing. 205 Raven et al. [1982] found statistically significant higher systolic and/or diastolic blood pressures during exercise for persons wearing respirators. Arborelius et al. [1983] did not find significant differences for persons wearing respirators during exercise. 3. Body Temperature Effects Proper regulation of body temperature is primarily of concern with the closed circuit SCBA that produces oxygen via an exothermic chemical reaction. Inspired air within these respirators may reach 120°F (49°C), thus depriving the wearer of a minor cooling mechanism and causing discomfort. Obviously this can be more of a problem with heavy exercise and when ambient conditions and/or protective clothing further reduce the body's ability to lose heat. The increase in heart rate because of increasing temperature represents an additional cardiac stress. Closed—circuit breathing units of any type have the potential for causing heat stress since warm expired gases (after exothermic carbon dioxide removal with or without oxygen addition) are rebreathed. Respirators with large dead spaces also have this potential problem, again because of partial rebreathing of warmed expired air [James et al. 1984]. 4. Sensory Effects Respirators may reduce visual fields, decrease voice clarity and loudness, and decrease hearing ability. Besides the potential for reduced productivity, these effects may result in reduced industrial safety. These factors may also contribute to a general feeling of stress [Morgan 19833]. 5. Psychologic Effects This important topic is discussed in recent reviews by Morgan [Morgan 1983a, 1983b]. There is little doubt that virtually everyone suffers some discomfort when wearing a respirator. The large variability and the subjective nature of the psycho—physiologic aspects of wearing a respirator, however, make studies and specific recommendations difficult. Fit testing obviously serves an important additional function by providing a trial to determine if the wearer can psychologically tolerate the respirator. The great majority of workers can tolerate respirators, and experience in wearing them aids in this tolerance [Morgan 1983b]. However, some individuals are likely to remain psychologically unfit for wearing respirators. 6. Local Irritation Effects Allergic skin reactions may occur occasionally from wearing a respirator, and skin occlusion may cause irritation or exacerbation of preexisting conditions such as pseudofolliculitis barbae. Facial discomfort from the pressure of the mask may occur, particularly when the fit is unsatisfactory. 206 7. Miscellaneous Health Effects In addition to the health effects (described above) associated with wearing respirators, specific groups of respirator wearers may be affected by the following factors: a. Perforated Tympanic Membrane While inhalation of toxic materials through a perforated tympanic membrane (ear drum) is possible, recent evidence indicates that the airflow would be minimal and rarely if ever of clinical importance [Cantekin et al. 1979; Ronk and White 1985]. In highly toxic or unknown atmospheres, use of positive pressure respirators should ensure adequate protection [Ronk and White 1985]. b. Contact Lenses Contact lenses are generally not recommended for use with respirators, although little documented evidence exists to support this viewpoint [daRoza and Weaver 1985]. Several possible reasons for this recommendation are noted below: (1) Corneal Irritation or Abrasion Corneal irritation or abrasion might occur with the exposure. This would, of course, be a problem primarily with quarter— and half-face masks, especially with particulate exposures. However, exposures could occur with full—face respirators because of leaks or inadvisable removal of the respirator for any reason. While corneal irritation or abrasion might also occur without contact lenses, their presence is known to substantially increase this risk. .(2) Loss or Misplacement of a Contact Lens The loss or misplacement of a contact lens by an individual wearing a respirator might prompt the wearer to remove the respirator, thereby resulting in exposure to the hazard as well as to the potential problems noted above. (3) Eye Irritation from Respirator Airflow The constant airflow of some respirators, such as powered air—purifying respirators (PAPR's) or continuous flow air-line respirators, might irritate the eyes of a contact lens wearer. 3. Suggested Medical Evaluation and Criteria for Respirator Use The following NIOSH recommendations allow latitude for the physician in determining a medical evaluation for a specific situation. More specific guidelines may become available as knowledge increases regarding human stresses from the complex interactions of worker health status, respirator usage, and job tasks. While some of the following recommendations should be 207 part of any medical evaluation of workers who wear respirators, others are applicable for specific situations. 0 A physician should determine fitness to wear a respirator by considering the worker's health, the type of respirator, and the conditions of respirator use. The recommendation above leaves the final decision of an individual‘s fitness to wear a respirator to the person who is best qualified to evaluate the multiple clinical and other variables. Much of the clinical and other data could be gathered by other personnel. It should be emphasized that the clinical examination alone is only one part of the fitness determination. Collaboration with foremen, industrial hygienists, and others may often be needed to better assess the work conditions and other factors that affect an individual's fitness to wear a respirator. 0 A medical history and at least a limited physical examination are recommended. The medical history and physical examination should emphasize the evaluation of the cardiopulmonary system and should elicit any history of respirator use. The history is an important tool in medical diagnosis and can be used to detect most problems that might require further evaluation. Objectives of the physical examination should be to confirm the clinical impression based on the history and to detect important medical conditions (such as hypertension) that may be essentially asymptomatic. 0 While chest X—ray and/or spirometry may be medically indicated in some fitness determinations, these should not be routinely performed. In most cases, the hazardous situations requiring the wearing of respirators will also mandate periodic chest X—rays and/or spirometry for exposed workers. When such information is available, it should be used in the determination of fitness to wear respirators. Data from routine chest X—rays and spirometry are not recommended solely for determining if a respirator should be worn. In most cases, with an essentially normal clinical examination (history and physical) these data are unlikely to influence the respirator fitness determination; additionally, the X—ray would be an unnecessary source of radiation exposure to the worker. Chest X—rays in general do not accurately reflect a person's cardiopulmonary physiologic status, and limited studies suggest that mild to moderate impairment detected by spirometry would not preclude the wearing of respirators in most cases. Thus it is recommended that chest X—rays and/or spirometry be done only when clinically indicated. 0 The recommended periodicity of medical fitness determinations varies according to several factors but could be as infrequent as every 5 years. Federal or other applicable regulations shall be followed regarding the frequency of respirator fitness determinations. The guidelines for most work conditions for which respirators are required are shown in Table V—1. 208 These guidelines are similar to those recommended by ANSI, which recommends annual determinations after age 45 [ANSI 1984]. The more frequent examinations with advancing age relate to the increased prevalence of most diseases in older people. More frequent examinations are recommended for individuals performing strenuous work involving the use of a SCBA. These guidelines are based on clinical judgment and, like the other recommendations in this section, should be adjusted as clinically indicated. 0 The respirator wearer should be observed during a trial period to evaluate potential physiological problems. In addition to considering the physical effects of wearing respirators, the physician should determine if wearing a given respirator would cause extreme anxiety or claustrophobic reaction in the individual. This could be done during training while the worker is wearing the respirator and is engaged in some exercise that approximates the actual work situation. Present OSHA regulations state that a worker should be provided the opportunity to wear the respirator :in normal air for a long familiarity period..." [29 CFR 1910.134(e)(5)]. This trial period should also be used to evaluate the ability and tolerance of the worker to wear the respirator [Harber 1984]. This trial period need not be associated with respirator fit testing and should not compromise the effectiveness of the vital fit testing procedure. Table V—1.——Suggested frequency of medical fitness determinations* Type of working Worker age (years) conditions <35 35 — 45 >45 Most work conditions Every 5 years Every 2 years 1—2 years requiring respirators Strenuous working Every 3 years Every 18 months Annually conditions with a SCBAT *lnterim testing would be needed if changes in health status occur. TLSCBA = self—contained breathing apparatus. *CFR = Code of Federal Regulations. See CFR in references. 209 0 Examining physicians should realize that the main stress of heavy exercise while using a respirator is usually on the cardiovascular system and that heavy respirators (e.g., SCBA) can substantially increase this stress. Accordingly, physicians may want to consider exercise stress tests with electrocardiographic monitoring when heavy respirators are used, when cardiovascular risk factors are present, or when extremely stressful conditions are expected. Some respirators may weigh up to 35 pounds and may increase workloads by 20 percent. Although a lower activity level could compensate for this added stress [Manning and Griggs 1983], a lower activity level might not always be possible. Physicians should also be aware of other added stresses, such as heavy protective clothing and intense ambient heat, that would increase the worker's cardiac demand. As an extreme example, firefighters who use a SCBA inside burning buildings may work at maximal exercise levels under life—threatening conditions. In such cases, the detection of occult cardiac disease, which might manifest itself during heavy stress, may be important. Some authors have either recommended stress testing [Kilbom 1980] or at least its consideration in the fitness determination [ANSI 1984]. Kilbom [1980] has recommended stress testing at 5—year intervals for firefighters below age 40 who use SCBA and at 2—year intervals for those aged 40—50. He further suggested that firemen over age 50 not be allowed to wear SCBA. Exercise stress testing has not been recommended for medical screening for coronary artery disease in the general population [Weiner et al. 1979; Epstein 1979]. It has an estimated sensitivity and specificity of 78% and 69%, respectively, when the disease is defined by coronary angiography [Weiner et al. 1979; Nicklin and Balaban 1984]. In a recent 6—year prospective study, stress testing to determine the potential for heart attacks indicated a positive predictive value of 27% when the prevalence of disease was 3.5% [Giagnoni et al. 1983; Folli 1984]. While stress testing has limited effectiveness in medical screening, it could detect individuals who may not be able to complete the heavy exercise required in some jobs. A definitive recommendation regarding exercise stress testing cannot be made at this time. Further research may determine whether this is a useful tool in selected circumstances. 0 An important concept is that ”general work limitations and restrictions identified for other work activities also shall apply for respirator use" [ANSI 1984]. In many cases, if a worker is physically able to do an assigned job while not wearing a respirator, the worker will in most situations not be at increased risk when performing the same job while wearing a respirator. 0 Because of the variability in the types of respirators, work conditions, and workers' health status, many employers may wish to designate categories of fitness to wear respirators, thereby excluding some workers from strenuous work situations involving the wearing of respirators. 210 Depending on the various circumstances, several permissible categories of respirator usage are possible. One conceivable scheme would consist of three overall categories: full respirator use, no respirator use, and limited respirator use including ”escape only" respirators. The latter category excludes heavy respirators and strenuous work conditions. Before identifying the conditions that would be used to classify workers into various categories, it is critical that the physician be aware that these conditions have not been validated and are presented only for consideration. The physician should modify the use of these conditions based on actual experience, further research, and individual worker sensitivities. He may also wish to consider the following conditions in selecting or permitting the use of respirators: ——History of spontaneous pneumothorax; ——Claustrophobia/anxiety reaction; —-Use of contact lenses (for some respirators); ——Moderate or severe pulmonary disease; ——Angina pectoris, significant arrhythmias, recent myocardial infarction; ——Symptomatic or uncontrolled hypertension; and —-Advanced age. Wearing a respirator would probably not play a significant role in causing lung damage such as pneumothorax. However, without good evidence that wearing a respirator would not cause such lung damage, the physician would be prudent to prohibit the individual with a history of spontaneous pneumothorax from wearing a respirator. Moderate lung disease is defined by the lntermountain Thoracic Society [Kanner and Morris 1975] as being present when the following conditions exist—-a forced expiratory volume in one second (FEV1) divided by the forced vital capacity (FVC) (i.e., FEV1/FVC) of 0.45 to 0.60, or an FVC of 51% to 65% of the predicted FVC value. Similar arbitrary limits could be set for age and hypertension. It would seem more reasonable, however, to combine several risk factors into an overall estimate of fitness to wear respirators under certain conditions. Here the judgment and clinical experience of the physician are needed. Many impaired workers would even be able to work safely while wearing respirators if they could control their own work pace, including having sufficient time to rest. c. Conclusion Individual judgment is needed to determine the factors affecting an individual's fitness to wear a respirator. While many of the preceding guidelines are based on limited evidence, they should provide a useful starting point for a respirator fitness screening program. Further research 211 is needed to validate these and other recommendations currently in use. Of particular interest would be laboratory studies involving physiologically impaired individuals and field studies conducted under actual day-to—day work conditions. 212 REFERENCES Altose MD, McCauley WC, Kelsen SG, Cherniack NS [1977]. Effects of hypercapnia and inspiratory flow—resistive loading on respiratory activity in chronic airways obstruction. J Clin Invest §gz500—07. ANSI [1984]. American national standard for respirator protection—respirator use—physical qualifications for personnel, ANSI 288.6—1984. New York, NY: American National Standards Institute, Inc., pp. 7—15. Arborelius M, Dahlback G0, Data P—G [1983]. Cardiac output and gas exchange during heavy exercise with a positive pressure respiratory protective apparatus. Scand J Work Environ Health 9:471-477. Bates DV, Macklem PT, Christie RV [1971]. Respiratory function in disease: an introduction to the integrated study of the lung. 2nd ed. Philadelphia, PA: W.B. Saunders Co., p. 43. Bentley RA, Griffin 0G, Love RG, Muir DCF, Sweetland KF [1973]. Acceptable levels for breathing resistance of respiratory apparatus. Arch Environ Health 27:273—80. Bjurstedt H, Rosenhamer G, Lindborg B, Hesser CM [1979]. Respiratory and circulatory responses to sustained positive—pressure breathing and exercise in man. Acta Physiol Scand 19§z204-14. Black LF, Hyatt RE [1969]. Maximal respiratory pressures: normal values and relationship to age and sex. Am Rev Respir Dis 99:696-702. Cantekin El, Bluestone CD, Saez CA, Bern SA [1979]. Airflow through the eustachian tube. Ann Otol 88:603—612. CFR. Code of Federal Regulations. Washington, DC: U.S. Government Printing Office, Office of the Federal Register. Craig FN, Blevins WV, Cummings G [1970]. Exhausting work limited by external resistance and inhalation of carbon dioxide. J Appl Physiol 22(6):847—51. Dahlback GO, Balldin Ul [1984]. Physiological effects of pressure demand masks during heavy exercise. Am Ind Hyg Assoc J fl§(3):177-181. daRoza RA, Weaver C [1985]. Is it safe to wear contact lenses with a full—facepiece respirator? Lawrence Livermore National Laboratory manuscript UCRL—53653, pp. 1—3. Deno NS, Kamon E, Kiser DM [1981]. Physiological responses to resistance breathing during short and prolonged exercise. Am Ind Hyg Assoc J 52(8):616—23. 213 Epstein SE [1979]. Limitations of electrocardiographic exercise testing [Editorial]. N Engl J Med §Ql(5):264—65. Folli G [1984]. Exercise EKG in asymptomatic normotensive subjects [Reply to letter to the editor]. N Engl J Med §19(13):852—53. Gee JBL, Burton G, Vassallo C, Gregg J [1968]. Effects of external airway obstruction on work capacity and pulmonary gas exchange. Am Rev Respir Dis 98:1003—12. Giagnoni E, Secchi MB, Wu SC, Morabito A, Oltrona L, et al. [1983]. Prognostic value of exercise EKG testing in asymptomatic normotensive subjects. N Engl J Med 899(18):1085—89. Halperin WE, Ratcliffe JM, Frazier TM, Becker SP, Schulte PA [1986]. Medical screening in the workplace: proposed principles. J Occup Med 2§(8):547—52. Harber P, Tamimie RJ, Bhattacharya A, Barber M [1982]. Physiologic effects of respirator dead space and resistance loading. J Occup Med 25(9):681—84. Harber P [1984]. Medical evaluation for respirator use. J Occup Med Z§(7):496—502. Hermansen L, Vokac Z, Lereim P [1972]. Respiratory and circulatory response to added air flow resistance during exercise. Ergonomics 1§(1):15—24. Hodous TK, Petsonk L, Boyles C, Hankinson J, Amandus H [1983]. Effects of added resistance to breathing during exercise in obstructive lung disease. Am Rev Respir Dis 128:943-48. Hodous TK, Boyles C, Hankinson J [1986]. Effects of industrial respirator wear during exercise in subjects with restrictive lung disease. Am Ind Hyg Assoc J 51:176—80. James RH [1977]. Breathing resistance and dead space in respiratory protective devices. Cincinnati, OH: U.S. Department of Health, Education, and Welfare, Public Health Service, Center for Disease Control, National Institute for Occupational Safety and Health, DHEW (NIOSH) Publication No. 77—161. James R, Dukes-Dobos F, Smith R [1984]. Effects of respirators under heat/work conditions. Am Ind Hyg Assoc J fl§(6):399—404. Kanner RE, Morris AH, ed. [1975]. Clinical pulmonary function testing: a manual of uniform laboratory procedures for the intermountain area. 1st ed. Salt Lake City, UT: Intermountain Thoracic Society. Kilbom A [1980]. Physical work capacity of firemen. Scand J Work Environ Health §z48—57. 214 Love RG, Muir DCF, Sweetland KF, Bentley RA, Griffin 0G [1977]. Acceptable levels for the breathing resistance of respiratory apparatus: results for men over the age of 45. Br J Ind Med 352126—29. Manning JE, Griggs TR [1983]. Heart rates in fire fighters using light and heavy breathing equipment: similar near—maximal exertion in response to multiple work load conditions. J Occup Med g§(3):215—18. Meyer E, Gurtner HP, Scherrer M [1975]. Physiological appraisal of a new respirator with positive pressure. Pneumonology 1§§:61—72. Morgan WP [1983a]. Psychological problems associated with the wearing of industrial respirators: a review. Am Ind Hyg Assoc J 55(9):671—76. Morgan WP [1983b]. Psychological problems associated with the wear of industrial respirators. J Int Soc Respir Prot 1:67-108. Myhre LG, Holden RD, Baumgardner FW, Tucker D [1979]. Physiological limits of firefighters. Brooks AFB, TX: Air Force School of Aerospace Medicine, ESL-TR—79—06. Nicklin D, Balaban DJ [1984]. Exercise EKG in asymptomatic normotensive subjects [Letter to the editor]. N Engl J Med 319(13):852. Petsonk EL, Hancock J, Boyles C [1983]. Physiologic effects of a self—contained self-rescuer. Am Ind Hyg Assoc J fig(5):368—73. Raven PB, Davis TO, Shafer CL, Linnebur AC [1977]. Maximal stress test performance while wearing a self—contained breathing apparatus. J Occup Med 12(12):802—06. Raven PB, Dodson AT, Davis TO [1979]. The physiological consequences of wearing industrial respirators: a review. Am Ind Hyg Assoc J fig(6):517—34. Raven PB, Jackson AW, Page K, et al. [1981]. The physiological responses of mild pulmonary impaired subjects while using a "demand" respirator during rest and work. Am Ind Hyg Assoc J gg<4):247—57. Raven PB, Bradley 0, Rohm-Young D, McClure FL, Skaggs B [1982]. Physiological response to "pressure—demand" respirator wear. Am Ind Hyg Assoc J 53(10):773—81. Ronk R, White MK [1985]. Hydrogen sulfide and the probabilities of ”inhalation" through a tympanic membrane defect. J Occup Med 21(5):337—40. Stemler FW, Craig FN [1977]. Effects of respiratory equipment on endurance in hard work. J Appl Physiol 52:28—32. Weiner DA, Ryan TJ, McCabe CH, et al. [1979]. Exercise stress testing: correlations among history of angina, ST—segment response and prevalence of coronary—artery disease in the coronary artery surgery study (CASS). N Engl J Med 391(5):230—35. 215 33' U.S.Government Printing Office: 1987—~548-159/60538 JUN e 1888 U.C. BERKELEY LIBRARIES WWI/“WWI“ ' CDDH‘HDDTB J .4 : -‘ 1-? ~ A 5'; fih‘ ’«a Q. 4 , * ' a? .9459 4.3. :EN'T m: HEALTH if) HUMREL semi 9 ,, ‘4 re PUBLIC HEALTH a-wcea i . " ’ - t: 1" ENTEREi-‘OR DISEASE CONTRéL 5,3“ nun-M 0 “WW sew?“ W. \ ‘ - “ , f, .- 1 FT LA 5}“ g? if? Lt. 1- 4 _ ,1 '_ - AY cififi‘mf Ipofifizzq . ‘ PRIVALE us 5300 'L ' ' ,. . i g. I | . _ 'DHHS (NIOSH) Putzhcatlon No. 88; 101 .L’ -' .. . ‘. ‘ ., .3. - - g . 1 . ‘, ‘ o . «_ ' ‘ I 1 I A ‘ ‘ s . ‘0‘ . ‘ 61‘ - . ‘ ‘ '4 ’ » A