Document 7M5v0V8beLjYdJQMB6O6mOa4V

Manheim, Pa. OtVISION and LOCATION Messrs. G. D'Olier J. F. Glennon J. H. Perry J. E. Day R. L. Ramsey J. R. Roe SUBJECT: DATE: May 4, 1977 Attached Data Sheet #18 from the Occupational Health Branch of the Ontario Ministry of Labor concerning asbestos may be of interest. I. H. Weaver grb att. cc: Messrs. J. H. Marsh G. F. Butt E. W. Drislane, FMSI R. H. Mereness, AIA Dr. H. C. Lewinsohn i - - 4 FMSI 05105 <J^) /?////_ /?? 7 Ontario Ministry of Labour Occupational Health Branch Data Sheet Ho. 18 ASBESTOS DESCRIPTION I The name comes from the Greek, "unquenchable" or "indestructible". It is a generic tdrm applied to a number of naturally occurring silicates which can be separated from the parent rock in fibrous form. The common types fall into two categories (a) Serpentine - chrysotile <NS6 (OH), S14010) (b) Amphiboles - amosite (Fe5 Mg2 Si8 022 (0H)2) - actinolite '' (CaSiO^.Kg SiO^.FcSiOj) - anthophyllite ((MgFe)7 Sig022(0H)2) - crocidolite (Na20.Fe203.3Fe0.8Si02.H20) - tremolite (Ca2Mg5Sig022(0H)2) The fibres of the different forms have characteristics - in terms of length, flexibility, texture, tensile strength, resistance to heat.or chemical reaction - which make each suitable for particular purposes. The characteristics which have been moat exploited are flame resistance and tensile strength. OCCUPATIONAL EXPOSURE Exposure to asbestos fibres ih air can occur in mining or milling or in any of hundreds of uses or processes. The fibrous material may be sprayed as insulation, as a fire retardant, or for acoustic effect. Asbestos is combined in cement asbestos board, shingles, brake linings, paints, floor tiling, water and sewer pipes, and ia used in the production of moulded plastic products, in sheet form as lagging for insulation and in flame retardant cloth. FMSI 05106 SOURCE A large proportion of the world's production of chryaotile is from Canada - mainly Quebec, with some from British Columbia, Yukon Territory, Ontario and Newfoundland. The-USSR is also an important producer. The main source of amphiboles is South Africa. Canada imports some crocidolite (blue asbestos) and aiaosite (brown asbestos) for special processes. Altogether the world production consists of chrysotlle 93%, crocidolite 42 and amosite 22 (approximate percentages). Treoollte and actinolite are of little commercial importance but are of health significance as they occur in some talcs. HEALTH ASPECTS When air containing asbestos fibres is inhaled, a number of effects can result. These will vary according to the type, concentration, size of fibre inhaled, the duration of exposure and in some cases to a combined effect with other Inhaled contaminants. The principal forms of asbestos-related disease or disorders ares 1. Asbestosis - a diffuse fibrosis (scarring and shrinking) of lung tissue. ^ 2. Lung cancer. 3. Mesothelioma - a malignant tumour of the pleura or peritoneum (lining of chest and abdomen). 4. Cancer of organs other than the respiratory system e.g. gastro-intestinal tract. 5. Pleural thickening and' calcification. Asbestosis The standard for asbestos control in air mentions fibres greater than 5 micrometers. (5Vtn) in length because it is believed that smaller fibres are removed by the ordinary self-clearing mechanism of the lungs. The smaller fibres are generally engulfed by the so-called scavenger cells. Inside these cells the fibreB become coated with proteinaceous material containing iron to form "ferruginous bodies" which, if coughed up, can be recognized in the sputum. The ferruginous bodies in sputum may be indicative of asbestos exposure but not necessarily of asbestos related disease. Fibres longer than 5pm have a tendency to be retained in the lung tissue where they may cause irritation and eventual scarring. Over years of exposure this may result in increasing shortness of breath. Advanced asbestosis can be diagnosed by chest x-rays but at a much earlier stage, the disease may be recognized by pulmonary function (breathing) tests. Lung Cancer FMSI 05107 The incidence of lung cancer is much higher in asbestos workers than in the general population. Lung cancer in asbestos workers is usually 3 preceded by fibrous changes In the lung tissue as occurs in asbestosls. It is claimed that an asbestos worker who smokes has 92 times greater chance of acquiring lung cancer than a non-smoking, non-asbestos worker. Note - There is evidence in the literature that mineral fibres other than those of asbeBtos e.g. - wollastonite, attapulgite, fibre glass and rock wool, may act on the respiratory organs in a manner similar to asbestos fibres. i Mesothelioma In the general population this is a rare malignant tumour. It has, however, been shown that a large proportion of the people who develop it have had some exposure to asbestos, though not always an occupational exposure. There appears to be a latent period of over 20 years between the first exposure and the appearance of the tumour. There is some evidence that mesothelioma is more likely to be caused by amphlboles (such as crocidolite and amosite) than by chrysotile.(10) Cancer of Organs Other Than the Respiratory System An increased occurrence of cancers in other locations than the respiratory system e.g. the gastro-inte6tinal tract, the breast^*?), has been reported. This is apparently due to the capability of the asbestos fibre to penetrate the lung tissue and to reach other organs outBlde the cheat cavity, or the asbestos fibres swallowed with mucous and food tc penetrate the walls of the stomach and intestines.^) Pleural Thickening and Calcification Asbestos fibres may cause irritation, inflammation and thickening of the pleura (pleural plaques). This is a condition that cm be seen at x-ray examination. It is evidence of prolonged exposure to asbestos but is not necessarily associated with symptoms of ill-health. STANDARD The time-weighted average (TWA) refers to the airborne concentra tion of substances in the workplace air. It is the concentration for a normal 8-hour workday or 40-hour workweek, to which nearly all workers may be repeatedly exposed, day after day, without adverse effect. The maximum allowable concentration (MAC) is the maximum concentration of airborne substances in the workplace air to which a worker may be exposed: such an exposure shall, not exceed more than one 15-minute period per day. For chrysotile - TWA is 2 fibres greater than 5lim in length per cubic centimetre(cc). . MAC is 10 fibres greater than 5v In length per cc. FMSI 05108 For amphiboles TWA ie 0.2 fibres greater than 5pm in length per cc. MAC is 2.0 fibres greater than 5pm in length per cc. The TWA of 2 fibres per cc. is intended to control the risk of asbestosis to less than IT. For sampling technique, see Appendix 1. CONTROLS A. General 1. Personal Hygiene ' (a) Good personal hygiene is to be maintained. This includes washing before eating, drinking, smoking or use of toilet facilities. *(b) Change Rooms - are to be provided by the employer at a fixed place of employment for workers working regularly at the place and who are exposed to airborne concentrations of asbestoB fibres. *(c) Clothes Lockers - two separate lockers for each worker so separated by a shower room that one locker for street clothes is in one room and a second locker for work clothes in another room. Contaminated clothing is to be placed in a covered impermeable container. (d) -Laundry (i) Laundering of asbestos-contaminated clothing is to be done so as to prevent the release of airborne asbestos fibres in excess of the TWA. * (ii) An employer who gives asbestos-contaminated clothing to another person or company for laundering is to inform such person or company of the requirement in subsection (i) of the subdivision. - (iii) Contaminated clothing is to be transported in sealed impermeable bags, or other closed impermeable . container and a caution label affixed (see Section 3(b)(1). 2. Food, Beverages and Tobacco , Food, beverages and tobacco are not to be kept nor consumed in the work area. - FMSI 05109 *Note - For workers engaged in work other than at a fixed-location, e.g, insulators, alternative suitable arrangements are to be provided. Caution Signs and Labels (a) Caution Signs - Posting - Caution signs are to be provided and posted at the approaches to each location where there is significant exposure to asbestos fibres. . The sign is to be printed in English and, as necessary, In other languages so that all workers understand the hazard. A typical asbestos card is shown as Appendix 2. (b) Caution Labels - (i) Labelling - Caution labels are to be affixed to all raw materials, mixtures, scrap, waste, debris and other products containing asbestos fibres, or to their containers except that no label la required where asbestos fibres have been modified by a bonding agent, coating, binder or other material so that during any reasonably foreseeable use, handling, storage, disposal, processing or transportation, no airborne concentrations of asbestos fibres in excess of the TWA will be released. (ii) Labels - All labels are to be in accordance with the Guidelines for the Labelling of Toxic Chemicals for Use in Ontario Industry. Special Clothing . ^' The employer is to provide and to require the use of special clothing head coverings, gloves and foot coverings for workers exposed to airborne concentrations of asbestos fibres. Work Practices Work practices are to be followed, as completely as possible, that prevent the dissemination of duBt. Examples are:- (a) When dumping bags of asbestos: - (i) Local exhaust is to be used. (ii) Excessive shaking of bags is to be avoided. (iii) Bags are to be disposed of carefully using local exhaust, covered containers, plastic bags that can be sealed, etc. (b) When sawing, drilling, grinding, machining asbestos-containing materials, etc. ` (i) Local exhaust is to be used. (ii) Products are to be vacuum cleaned, washed, or otherwise cleaned before further operations, storage or shipping. Compressed air is not to be used for cleaning. 6 6. Housekeeping Housekeeping Includes prompt vacuum cleaning or vet sweeping of all scraps and spills, careful storage and disposal of asbestos-containing materials in plastic bags or other suitable containers, and frequent and regular cleaning of machines, floors, vails and other plant surfaces (by vacuum cleaning) where asbestos dust settles. Dry sweeping, dusting or compressed air are never acceptable. 7. PersoI-n-al "r,Protect' ion,r .' In certain circumstances and for limited periods, of time, the worker may be protected by respiratory equipment. This should, however, never be allowed to take the place of adequate environmental control. . It must be recognized that the wearing of a respirator for long periods of time is often uncomfortable and sometimes of questionable efficiency. Workers may, in some instances, be neglectful in the care and the uae of their respirators, thus decreasing the reliability of this type of protection. . In general, air levels below the standard of 2 fibres per cubic . centimetre are to be maintained. During process changes or while improvements are being made to a plant, workers may be required to wear replaceable filter type respirators of the type approved for control of respirable dust if fibre concentrations do not exceed 10 fibres per cubic centimetre. This is always to be considered a temporary measure and is not to be continued beyond three months. During such periods it is essential that the wearing and the maintenance of respirators be closely supervised by the employer. 15. Engineering Normal procedures for the control of respirable dust apply to the control of the health hazard from asbestos. 1. Substitution When feasible, asbestos should be replaced by less hazardous materials. 2. Segregation It is possible with suitable planning to arrange that work producing high concentrations of asbestos dust is done in isolation. Where segregation is difficult, sb in the case of asbestos spraying, it is possible with suitable scheduling to have the process carried out when other workers are either off the shift or when they are removed from the area. ' Control booths, under positive pressure, may isolate the worker from exposure while allowing him to control a process. In certain circumstances, respirators may hIbo be used (see "Personal Protection"). FMSI 05111 7 3. Ventilation The use of adequate local exhaust is the most important method of du6t control. This method captures the dust at the source end conveys it to a central dust collecting system. Good general exhaust is to be provided so that any fibres not captured by local exhaust are eventually removed from the workroom air. t 4. Dust Suppression The use of water to suppress dust at the source has been universally accepted. Water sprays are perhaps the most common method of applying water to control both visible and respirable dust but it is not possible to control a dust cloud by means of water sprays once the dust has become airborne. C, Medical Surveillance 1. First Aid No acute poisoning due to asbestos has been recorded. 2. Preplacement Medical Examination ^ Preplacement medical examination is to be done within 30 days of the person's employment in an exposure to asbestos. The examination is to include a medical and occupational history including smoking habits; pulmonary, cardiovascular and gaatro-intestinal symptoms; a physical examination; a chest x-ray (posterior - anterior, 14 inches x 17 inches); and pulmonary function tests to include forced vital- capacity (FVC), and forced expiratory volume at one second 3. Periodic Medical Examination Periodic medical examination is to be done at least every two(2) years on each employee engaged in an occupation exposed to asbestos fibres. The examination is to follow the same procedures as outlined for a preplacement medical examination. It is to be noted, however, that those persons who (a) have a history of 10 or more years of employment involving exposure to asbestos, (b) show x-ray findings (such as small opacities, pleural plaques, pleural thickening, pleural calcification) which suggest or indicate pneumoconiosis or other reactions to asbestos, or (c) have changes in pulmonary function which indicates restrictive or obstructive lung disease are to have the chest x-rays and pulmonary function tests at least annually. FMSI 05112 8 4. Termination Medical Examination Within thirty(30) days before or after termination of employment, an employee engaged in an occupation exposed to significant airborne concentrations of asbestos fibre is to have a termination medical examination. The examination is to follow the same procedure as outlined for a preplacement medical examination. 5. Medical Records i Complete and accurate records of medical examinations ore to be maintained. These confidential records are to.be retained~by the employer for a period of forty(40) years or length of employment plus 20 years whichever is longer. ACTIOS LEVELS Depending on the process, it may be difficult for an asbestos-. producing or using industry to meet the standard at all times. It becomes necessary, therefore, to establish criteria on which certain actions are to be taken. It is desirable that work places be monitored- and exposure of personnel be determined regularly. The aim must be to keep both area end personal samples below the TWA. Whenever the standard is exceeded, personnel are to be protected by the wearing of approved respirators until the conditions are corrected. As indicated above (see Personal Protection), this action ie acceptable for limited periods of time, when concentrations are between the TWA and the MAC. At the same time, it is the responsibility of management to advise the representative of the Ministry of Labour that a state of non-compliance exists. When the concentration exceeds the MAC the process is to be shut-down while corrective action is taken. At such times only workers with positive pressure air-supplied respirators are to be engaged in repair or corrective work. The representative of the Ministry of Labour is to be notified and will decide whether or not the work may resume. As a general rule, the mill or plant will not be allowed to maintain production if the asbestos exposure exceeds the MAC. At levels between TWA and MAC production may be allowed to continue on condition that the workers wear appropriate respiratory protection, that positive action is undertaken by the firm to correct the situation and that the ( period of corrective action does not exceed three months. FMSI 05113 9- BIBLIOGRAPHY 1. Enterline, Philip E.: Pitfalls In Epidemiological Research: An Examination of the Asbestos Literature Journal of Occupational Medicine/Vol. 18, No. 3/March 1976. 2. Rajhans, G.S. and G.M. Bragg, Ph.D.: A Statistical Analysis of Asbestos Fiber Counting in the Laboratory and Industrial Environment, American Industrial Hygiene Association Journal, December 1975. 3. Edwards, G.H. and J.R. Lynch: The Method Used by the U.S. Public Health Service for Enumeration of Asbestos Dust on Membrane Filters, Ann. Occup. Hyg. Vol. 11 pp 1-6 (1968). 4. Code of Federal Regulations, Part 1910.1001 Occupational Safety and Health Administration, Department of Labour, U. S. A. Revised as of July 1, 1975. 5. Occupational Exposure to Asbestos ' U. S. Department of Health, Education and Welfare National Institute for Occupational Safety and Health, 1972. 6. Selikoff, I.J., Hammond, E.C., Churg, J.,: Mortality experiences of asbestos insulation workers 1943-1968 in Pneumoconiosis Proceedings of the International Conference, Johannesburg, 1969 Oxford University Press, 1970. 7. I. Doniach, K.V. Swetteriham and M.K.S. Hathom: Prevalence of Asbestos Bodies in a Necropsy Series in East London, Association with Disease Occupations of Domicilliary Address, B. J. of Ind. Med. 32-16-30/75. ' 8. Selikoff, I.J., E.C. Hammond, J. Churg - Asbestos Exposure, Smoking and Neoplasia, JAMA 204/106-112/1968 9. Toll: The Age Distribution of Cancer. Implications for Models of Carcinogenesis, J.R. Stat. Soc. 134/1971 10. McDonald, J.C., McDonald, A.D., Gibbs, G.W., Liemiatiski, J. and Rossiter, C.E.: Mortality in Chrysotile Asbestos Mines and Mills in Quebec, Arch. Envir. Health 22-(677-686) 1971. FMSI 05114 APPENDIX I Sampling Technique for Asbestos In Workplace Air The ideal sampling period to determine a TWA is a full shift* however, this is rarely practicable. It has been demonstrated that a 90 minute sample gives a reasonably accurate estimate of the TWA particularly if the process is a constant1'one. The more continuous the process, the iqore representative a short term sampling will be. It is evident that the skill of the sampler operator and a familiarity with the process are important factors in the assessment of an asbestos exposure. The rate of sampling may vary between 1.8 and 3.0 litres per minute depending on the airborne concentration of asbestos fibres. The counting method used is that of Edwards and Lynch. (3) A segment of the filter is placed on a glass slide and treated to make the filter transparent. Counting of the fibres is done with an optical microscope using phase contrast illumination at 400 to 430 X magnification. The special counting graticule has reference marks which permit a rapid classification of fibres (a "fibre" is described as being over 5pm in length and having an aspect ratio of 3 to 1). The procedure is to count 100 fibres or 20 fields, whichever is less. A simple conversion is then made from the counts per field of view to fibres per cubic centimetre. '_ ' Sampling frequency in to be adapted to meet the needs of the local situation. During a start-up period or during a process change, frequent sampling may be required. The actual time interval will depend on the judgment of the person making the assessment. NOTE: The Ministry of Labour recognizes that there can be considerable variability of fibre counts done by different operators in the same situation. In the hands of a skilled operator the reproduc ibility of the results can be expected to be at least 25Z (plus or minus).(2) However, the actual fibre concentration in the air ; will also vary throughout a workday and the results of a sampling period of less than a full 8-hours requires interpretation by an experienced industrial hygienist. The Ministty recognizes, also, that there is a need for additional research to establish the validity of counting procedures as well as the relative significance of area and personal samples. I FMSI 05115 APPENDIX II ASBESTOS DUST MAY BE HARMFUL TO YOUR HEALTH AVOID BREATHING DUST WEAR ASSIGNED PROTECTIVE EQUIPMENT ' DO NOT REMAIN IN AREA UNLESS YOUR WORK REQUIRES ST FMSI 05116 I c ASBESTOS INFORMATION ASSOCIATION/ North America - 1835 K Street, -N. W. Suite 40? Washington, D. C, 20005 ft This study develops a decision method for evaluating the social acceptability of industrial controls on hazardous materials. Decisions are based on a "multiple criteria approach" thatjointly considers measures such as risk-benefit tradeoff, minimum . reducible health risk, maximum acceptable cost and implicit value of human life. Health risks are calculated by combining separate estimates of production and usage patterns, emissions to air and water, effectiveness of controls, pollutant dispersion and human susceptibility. Economic benefits consider employment, trade and consumer impacts, as well as direct costs ofcontrols. Theanalysis focuses on asbestos as an example hazard. Relative values of hazard reduction alternatives are examined for asbestos manufacturing exhaust filters and for asbestos substitutes in brake linings. Preliminary calculations indicate risk reductions of these alternatives cannot justify their social costs. [f v Risk-benefit analysis for industrial and social needs KENDALL D. MOLL* and DENNIS P. TIHANSKY: Stanford Research Institute; George Washington University The term risk-benefit analysis has become a commonplace expression as both the public and governmental agencies become increasingly concerned with environmental protection and the quality of life. Attendent with this concept is the notion that there is some sort of sheet used in the decision-making process for regulating environmental quality. Informal comparisons of risk* and benefitsdo indeed occur in setting most regulations. The underlying decision structure, however, does not always incorporate three crucial considerations; the necessity of making tradeoffs; the likelihood that tradeoff impacts are noncommensurate (that is not expressed in the same units, as dollar benefits vs. loss of life); and uncertainty about impacts of alternative decisions. Formal riskbenefit analysis is useful here because it systematically applies economic theory and decision analysis to help the policymaker understand the tradeoffs. Risk-benefit analysis has different compari sons for different persons. Some ecologists, for Proem address: Castle and Cooke. Inc.. San Francisco. CA. example, value preservation of nature over costs of control; they might think it meaningless to compare ecological parameters with any economic concept. On the other hand, many industrialists place greater weight on the costs of preserving nature. Uncertainties orperhapseven ignorance of environmental consequences, can significantly affect an individual's value system. Fear of unknown risks often instills greater conservatism in action than is warranted by the objective situation. For instance, a minute but well publicized probability of death from exposure to a hazardous material might make an unknowing person think that he will be affected. Aggregated over all individuals, this risk could thus be exaggerated over its actual level. Publicity about risks and benefits could also be a poor indicator of indiv idual values. Surveys are thus important in assessing variations of public opinion, patterns of risk-taking among members of the general population and perceptions of risks and benefits. A large spread of personal evaluations should be expected given data deficiencies on risks and benefits. Information about ambient concentra- American Industrial Hygiene Association JOURNAL f38) 4177 FMSI 05117 153 if if t* 4S* B S' < u a (E-V- \\ < P*irn**. - i.-ifi V *. .*,i. lions of hazardous materials and the degree of human exposure is not generally available. More fundamental is the lack of information on inherent biological effects of these chemicals on human health. Epidemiological studies have been outstripped by the technology to produce and distribute an ever growing number of hazardous materials. For some new chemicals, the regulatory agency relies on the manufacturer and developer of derived products to furnish health information. But without a character or checklist of needs for conducting risk assessments, the manufacturer cannot really be expected to assume this additional respon sibility. Insights on benefits of product use are typi cally as deficient as judgements on hazard levels. Conditions for optimality in regulation depend on the "welfare function" to be maximized. One criterion frequently employed is maximum value of total output, which in principle should include such non-monetary aspects as safety, aesthetic qualities and reductions in physic tension or anxiety. Economists generally contend that these benefits of controlling hazards can be assessed collectively as the sum of the "willingness-to-pay" of all individuals affected by the action. Analytically, this requires the formulation of a consumer demand curve for feasible controls. Benefit assessment then involves integration of the area under this curve. Determination of the overall curve relating benefits to emission control policies requires a multi-step analysis, in which the economics input occurs only after exposures and health effects have been determined. Consequently, the lack of benefit estimates for regulations can be only partly blamed on economists. The lack of meaningful exposure and dose-response relationships falls in the domain of the life and physical sciences. Without accurate knowledge of these relationships, attachment of economic values is a misleading and perhaps futile exercise. Most regulatory agencies as well as researchers lack an appreciation of the complexities underlying benefit estimation. As a result, they tend to establish controls which minimize risks without explicitly considering product benefits. "Zero-tolerance laws," which Figure 1 -Ftisk/bensfit methodology: Analytical Steps. prohibit any detectable amount of a hazardous material, are inflexible and can quickly become too costly. They may obligate large budgets; they may remain in effect over an extended time span; and they may adversely affect the lives of many individuals. Although risk-benefit analysis should make decision-making more comprehensive, there are a number of limitations to its current usefulness. First, it represents an abstraction of reality and is usually simplistic. It may not be translatable into practice if it does not adequately account for all impacts. Second, the magnitude of uncertainity about both risks and benefits is typically much greater than the magnitude of their most likely values. This difficulty emerges as the most severe constraint of any formal analysis. Proxy measures are often used to measure impacts, and these proxies may be quite crude. Some authors, for example, represent risk in terms of death rates.1 While vital, this measure fails to incorporate non-fatal results that are likely to be far more prevalent and in some cases more damaging in total than the number of deaths. Another limitation is that risks and benefits of various actions do not usually come in single pairs. Rather, each control level implies a diversity of beneficial and costly impacts. The analyst, with limited resources and time, must 154 Am. tod. Hyg. Assoc. J (38) April. 1977 FMSI 05118 STANDARDS: I Figure 2 -Hazardous waste system. select the most important impacts and exclude the above referenced s .;dies. But it goes beyond those of hopefully less importance. As he learns their scope by including uncertainty factors and more, he may change the list of impacts. Still another limitation concerns weights assigned to risks. Individuals seem to waver in their by showing how various decision-making approaches can result in different priorities. The conceptual model to be presented was applied * i i evaluation of generally low-probability events as to a selection of feasible controls on hazardous )S. needs and perspectives fluctuate. materials, namely, cadmium and asbestos usage IUS There have been few attempts to consider such in the United States.5 ne limitations in formal risk-benefit analyses. The methodology follows very closely that ey Because highway data are readily available, recommended in the recent National Academy in; risk-benefit studies are most numerous for of Sciences report on Decision Making for ny vehicular accidents and choices of safety features Regulating Chemicals in the Environment.* The for automobiles, (e.g.).2'3 One extensive most significant aspect of this methodology is its he investigation of risk-benefit ratios for a number reliance on a "multiple criteria" approach. A ire ss. of hazardous events, such as driving and being multiple criteria approach involves the struck by lightning selected "fatalities per consideration of many factors bearing on the li iis person-hour of exposure" as the measure of risk decision in addition to those of "risk" and ito and a benefit index defined as "value to the "benefit," such as "minimum reducible risk," all individual" in the denominator of a risk-benefit "maximum socially acceptable risk," "maximum lity ratio.1 The benefit in most cases was quantified acceptable cost," and "value of human life." ich ely as the monetary investment that consumers are willing to support for a risky venture. As shown in Figure 1, the analysis starts with a 4 ere >xy Another more formal, theoretical model represents benefits of an item as the price a review of present standards and proceeds to parallel examinations of the existing pollution system and alternative controls that might be v*? ind jrs. ath consumer is willing to pay for it.4 As hazards or risks increase with item use, its value declines correspondingly. However, there is an applied to it. The parallel approach continues with second stage investigations of economic benefits on one branch and of population to additional "non-pecuniary" part of risk, he exposures and resultant health hazards on the ) be ore claims, beyond which the consumer or general other. Finally, the economic benefits and the public refuses to buy oruse the item. This implies health risks are combined with other relevant tbs. the existence of a threshold level of risk parameters in a graphical presentation of the sof acceptance. Whether such a level actually exists multiple criteria affecting the decision. This rgle was not demonstrated. presentation is applied under a variety of s a In the remainder of this paper, we present a possible decision rules to derive acceptable The new approach toward risk-benefit analysis. It standards and develop priority research needs HISt . isr7 incorporates risk and benefit comparisons, as in for future improvement of the decisions. American Industrial Hygiene Association JOURNAL (38) 4/77 155 if FMSI 05119 tr / E CONTRC ORIGIN CONTRC EFF NAT 10. Figure 3-U. S. Asbestos flow [metric tons per year). When we examine the top box, Present Standards, in more detail, we discover the kind of control system illustrated in Figure 2. This figure shows how various types of standards impinge on the hazardous waste system and on the monitoring and control mechanisms that accompany it. Standards must be compared with the actual amount of hazardous waste present at a given point in the hazardous waste flow system. For example, production standards must be compared with some measure of the actual production. The difference between the standard and actual production is used as a signal for the monitor and its control system to implement corrective measures if necessary. The same kind of feedback control must operate for usage standards, emission standards, ambient air or water standards, exposure standards and ingestion standards. Note from the figure however, that controls do not have to be applied at the same point that the monitoring signal is observed. Excessive exposures, for example, could be corrected by more restrictive controls over emissions concentrations. These multiple feedback possibilities allow for very complex control systems. The multiple feedback loops do impose two requirements on the monitoring and control systems. Whatever standards are developed must be measureable and they must be consistent with other standards that they may be applied at other points in the control process. For example, the OSHA asbestos standard was reduced from 5 visible fibers to 2 fibers per milliliter on July 1, 1976. However, the visual methods currently used to count asbestos concentrations account for less than 5% of total fibers and no one knows whether the visual identifiable fibers affect health more than the smaller fibers.7 Standards cannot be made very precise until this conversion problem is solved. When we decide that some measure of pollution, say weight, is most meaningful, then the second step is to analyze the industrial pollution of various media (air, water, and land) that occurs at different stages of the system, from extraction to final disposal. These amounts are shown in Figure 3. We have used a materials balance approach to estimate the amounts of asbestos going to each use and each disposal media and reconcile their totals with our best estimates of overall use and disposal. The quantities (not all of which are shown in the figure) therefore are additive both vertically and horizonatally along the different flow paths. This requires some very difficult data search and 156 Am. Ind. Hyg. Assoc. J. (33) April. 1977 Figure recon creati Howt, miss 11 the d plausisystc high I , be rel ; base t In i dispo^ pollut tion i' fabric; exami' also h: the se. emissic brake 1 effectiv pollutii steps, manuf; fabric control 96% of nation; about million Subs brake 1 American I: FMSI 05120 EMISSION SOURCE: MANUFACTURING facilities AUTOMOBILE BRAKES CONTROL METHOD: FABRIC FILTERS SUBSTITUTE MATERIAL ORIGINAL EMISSIONS: 547 METRIC TONS 129 METRIC TONS CONTROL EFFECTIVENESS : 96% 200% NATIONAL COST: LOW KEAN HIGH S2.5 MILLION $3 S3.6 S52 MILLION $65 $81 Figure 4-Asbestos control costs and effectiveness. g reconciliation work, almost to the point of being creative when the data is particularly sparse. However, we feel it is necessary in order not to miss large parts of the problem simply because the data is difficult to obtain. It gives us a plausible base from which to make an overall systems analysis and, at the same time, highlights areas where existing data may need to be refined. It also provides a first-order priority base for examining alternative controls. In our analysis, we observe that although land disposal accounts for the great bulk of asbestos pollution, the largest source of air contamina tion is the 547 metric tons emitted from fabrication operations. We therefore have examined controls for asbestos fabricators. We also have examined the possibility of eliminating the second most significant airborne source: emissions from users of friction products (i.e., brake linings). Economic costs and operational effectiveness of control alternatives for these two pollution sources were examined in the next two steps, as shown in Figure 4. For asbestos manufacturing processes, we ascertained that fabric filters are the most effective method to control emissions.8 They can eliminate about 969c of stack emissions. We estimated the total national cost for such a program would run about S2 million per year, plus another 51 million for monitoring and enforcement. Substitution of other materials for asbestos brake linings is speculative since no satisfactory AREAS EXPOSED - SQUARE KILOMETERS F/gure 5-Asbestos air concentrations. material has yet been found. But if one assumes that a material can be found for a 50% cost increase, the total extra cost would run to about S65 million per year. These economic costs have to be balanced against the control program's reduction in public exposure and consequent hazard, which we also examined. Figure 5 shows the total areas now exposed in the U.S. to various concentration levels from industrial sources and from brake lining emissions, according to our simplified model of emission sources and air dispersion. Note that industrial sources are more concentrated over a smaller area than are brakeshoe sources. Total exposure to the population is about 80 mg/m3 from each source. The big hurdle in our analysis, as in most such analyses, came in trying to develop a doseresponse function to evaluate the effects of these exposures. To do this, we looked at death rate estimates for the three most important causes of death from asbestos: mesothelioma, other cancers and non-cancer respiratory diseases. nretjcan industrial Hygiene Association JOURNAL (38) All7 157 *?vf v s s * % FMS1 05121 Figure 6-Asbestos dose-response: respiratory system disease. Figure 6 shows excess deaths versus accumulated asbestos dosage for non-cancer respiratory diseases* As with most of our data, we had some problems here in converting from one set of measures and dosage assumptions to another. For Figure 6, uncertainties in these conversion factors were the major uncertainties in determining the confidence limits that are shown. Figure 7 illustrates excess death rates of people exposed to asbestos from all cancers except mesothelioma.' The Mae West shape is simply a matter of curve fitting and has no known physiological basis. Excess deaths from mesothelioma only are shown in Figure 8. This diagram presents the widest uncertainty of all because its two references10'11 both attributed their original data to a common source but differed by a factor of five in their resultant calculations. Figure 7-Asbestos dose-response: cancer fexcept mesothelioma). By summing these three causes of excess death and combining their uncertainties as independent random variables, we arrived at the overall dose-response curves of Figure 9. We show these for analytical purposes as straight lines emanating from the zero intercept. 158 Am. Ind. Hyg. Assoc. J. (33) April, 1977 FMSI 05122 i although in fact we have no strong basis for assuming that the mean value is a straight line or that the uncertainties arc fixed fractions of the dosage. Mesothelioma appears to be the most significant contributor to the total excess death rate. Asbestosis produces a significant number of illnesses in addition to the excess deaths, but morbidity effects were not included in our analysis. Hazards from ingestion of food and water were likewise not considered, since most studies of relative hazards indicate that the principal mode of er' - of retained asbestos is via the lungs.INM Finally, we did not consider the effect of population mobility on the hazard, even though people with high exposures are statistically very unlikely to live next to an asbestos factory for their entire lives. . But these neglected effects are all relatively insignificant compared to the order-ofmagnitude uncertainties and the other difficulties of reaching a tradeoff between risks and benefits. An overall method of evaluating and presenting risk/benefit tradeoffs is illustrated in Figure 10. Here, the two main parameters are measured in the separate dimension of the chart. The vertical scale shows risk in terms of lives saved by a control alternative relative to the existing "status quo" situation and the horizontal shows negative economic benefit (or cost) involved. Each alternative, surrounded by an ellipse representing the confidence limits, can be shown in terms of this tradeoff of lives versus non health economic benefits. The dollar tradeoff between health and economic effects is not charted, but any particular valuation of human life can be represented by a diagonal line originating at the status quo position. Alternatives lying above this tradeoff line would be cost-effective in terms of that particular life valuation, whereas alternatives below the line would presumably not be. Other constraints can also be shown on the chart; these act to restrict the feasible domain with which alternative solutions may be sought. At the top a "minimum reducible risk" line represents a limit in the number of lives that MILLIGRAM-YEARS RER CUBIC METER IN AIR Figure 9-Asbestos dose-response: total. COST (non-HEALTH dollar benefits) Figure JO-Multiple criteria comparison method. might be saved by any feasible alternative. The minimum reducible risk might be considered as a background level of contamination below which further reductions are extremely difficult. At the opposite side of the feasible domain, the "maximum acceptable social risk" represents a An-er.can bc,,sttial Hygiene Association JOURNAL (331 AITT 159 FMSI 05123 limits of about a factor of two in each direction. But implementation costs of thetwoalternatives differ by more than 20:1. If one wishes to assign a value to saving lives, the chart shows that the factory filter alternative costs less than S10 million per life saved whereas brake substitution costs about SI 00 million per life. Neither alternative comes close to the 5300,000 valuation that workers in hazardous occupations implicity give to their own lives15'16 (see the upper diagonal line). Figure 11 -Asbestos pollution control alternatives. number of lives lost that will not be readily accepted by society. The maximum risk might be some vague social limit such as the prevailing rate of disease, or it may represent a "here and now" risk limit as defined by existing standards and regulations. Another constraint, not usually mentioned together with risk limits but nevertheless logically comparable, is that of "maximum acceptable social cost." This represents the maximum expenditure that society is willing to obligate at a particular time to solve a particular pollution problem. Equity considerations among the population are also now generally accepted as valid constraints. Equity considerations can be considered by making comparable charts for the analysis of each ethnic, income, geographic, generational, or other identifiable interest group. Together, the risk, benefit, and other constraints make up the "multiple criteria" problem. For this reason, both alternatives appear inefficient as measures for protecting the general public, if economic tradeoffs are considered. As always in such studies, however, we must qualify our conclusions. First, we have not considered potential effects of these measures on the health of industrial workers, which very likely would be more significant than those to the general public. Inhalation hazards to brake shoe installers, for example, would be completely eliminated by substitution of some other material for asbestos in brakes. Second, we have only examined two alternatives. Calculations based on our exposure model indicate that about 37 people per year could potentially be saved by eliminating asbestos from our ambient air (see top horizontal line in Figure 11). This potential life saving is 50 times as great as we get from either of the two alternatives considered, so additional protection possibilities certainly are worth investigating. These and other factors that go beyond the assumptions of a particular analysis almost need to be considered by decision-makers in real life. Therefore, no single chart can give a complete answer. In addition to the formal trade-off charts of the type shown in Figures 10 and 11, a full presentation should include the listing of many supplementary criteria by which decision makers or other interested parties can derive their own values. Such a list is shown in Figure !2.6 ' When we apply this methodology to the asbestos control alternatives we obtain Figure 11, which has been derived specially for this paper. The first feature one notices in this logarithmic scaled chart is that lives saved by the two alternatives are about the same: namely. 0.6 person per year each, give or take confidence Choices among the criteria presented can be made on the basis of many decision procedures. including those of expected value and ordinary old fashioned biases such as optimism, pessimism and probability. So many selection methods are available, in fact, that decision makers in some ways will have greater freedom 160 Am. tod Hyg. Assoc J. (3Si April. 1S7T FMSI 05124 alternatives COMMENTS NO t --t:;n ; option i C:\TP3L : CONTROL REGULATION BAN NOW ' A; B 1. HAZARDS AVOIDED A. HEALTH B. ENVIRONMENTAL II. COSTS OF CONTROL A. DIRECT B. INDIRECT C. MARKET STRUCTURE 111. BENEFITS LOST IV. DISTRIBUTION OF BENEFITS S COST 1/ 'I i :i :i ;i Figure 12-Display of benefits and costs. of action than they ever had before. The choice among alternatives may remain ambiguous. But at least decision makers will have explicit, quantitative means for weighing the practical tradeoffs that in the long run are going to have to be considered. references 1. Staff, C.; Social Benefit Versus Technological Risk. Set. 165:1232(1969) 2. Lave, L.B. sncf W.E. Waver: A Benefit-Cost Analysis of Auto Safety Features. AppL Scon. 2:1 (1970). 3. Calibresi, G.: The Cost of Accidents. Yale University Press. New Haven. Conn. (1970). 4. Muehlhouse, C.O.: Risk-Benefit Analysis in Decision-making. National Bureau of Standards. Washington. D C . unpublished manuscript |1972). 5. Moll, K.D., S. Baum, E. Capener, F.S. Dresch and R.M. Wright: Hazardous Wastes: A Risk-Benefit Framework Applied to Cadmium and Asbestos. Stanford Research Institute for Environmental Protection Agency (September 1975). 6. Davies. J.C. ed: Decision Making for Regulating Chemicals in the Environment. Chapter 5 and Appendix H. National Academy of Sciences. Washington. D C. (1975).7 7. Background Information on the Development of National Emission Standards for Hazardous Air Pollutants: Asbestos. Beryllium, and Mercury. APTD1503. Office of Air and Water Programs, U. S. Environmental Protection Agency, p. 34 (March 1973). 8. Paddock, R.E. et a/:Corr.prehensive Study of Specified Air Pollution Sources to Assess the Economic Impact of Air Quality Standards. Vol. II, Asbestos. Beryllium. Mercury. PB-222 858. prepared for U.S. Environmental Protection Agency by Research Triangle institute (August 1972). 9. Enterline. P., P. DeCoufla and V. Henderson: Mortality in Relation to Occupational Exposure in the Asbestos Industry. J. Occup. Med. 14:897 (1972). 10. Bruckman, L. and R.A. Rubino: Rationale Behind a ProposedAsbestos Air Quality Standard No. 74-222. presented at the 67th Annual Meeting of the Air Pollution Control Association. Denver. Colorado (9-13 June 1974). 11. Selikoff. I.J.: Asbestos Criteria Document Highlights. ASSE J. (3):26 (1974). 12. A Study of the Problem of Asbestos in Water, by the American Water Works Association Research Foundation. AM. Water Works Assoc. Vol. 66. No. 9. Part 2, p. 1 (September 1974). 13. Merliss, R.R.: Talc-Treated Rice and Japanese Stomach Cancer. Set. 173:1141 <1971). 14. Masson. T.J., F.W. McKay and R.W Miller: Asbestos-Like Fibers in Duluth Water SupplyRelation to Cancer Mortality J. Am. Med. Assoc. 228:1019(1974). 15. Thaler. R. and S. Rosen: The Value of Saving a Life: Evidence from the Labor Market, paper presented 30 November 1973. published by University of Rochester. 16. Melinek, S.J.: A Method of Evaluating Human Life for Economic Purposes. Fire Research Note No 950. Herts.. England (November 1972). Accepted 29 1976 Industrial Hygiene Association JOURNAL (33) 4/77 161 FMSI 05125 A method is described for calculating confidence intervals for particle or fiber concentration, and for dust collector penetration. The span of the interval depends upon the value of fiber concentration or collector penetration reported and upon the number of particles or fibres counted. Uncertainty in particle counting and sizing procedures ' DAVID LEITH and MELVIN W. FIRST Harvard School of Public Health, Department of Environmental Health Sciences, Boston, Massachusetts 02115 . 'on Introduction The concentration of particles in a gas can be determined by passing a known gas volume through a filter and counting particles on representative filter portions. Particle concen trations are valuable to determine compliance with legal standards, as for asbestos fibers in workroom air, or to determine the particle size collection efficiency of a dust collection device by making counts of simultaneous upstream and downstream samples. For both applica tions, it is important to estimate the count reliability. Although enough particles must be counted to establish the validity of the result within acceptable limits, it is wasteful to insist upon excessive counting to obtain needlessly high reliability. Particles for microscopic counting are conveniently collected on membrane filters1,2 or electron miscroscope grids1,3,4. The "stratifi cation" particle counting method5'7 illustrated in Table I is often used to reduce counting time. An initial traverse is made by examining a representative number of fields under the microscope. All particles seen are segregated into convenient, continuous size categories. Subsequent traverses note entries for only those size ranges in which particles are present in relatively small numbers. The average num ber of particles in each size range per traverse, x, is then calculated as shown in Table I. Stratified counting is a way to emphasize those particles whose concentrations arc most diffi cult to assess with statistical reliability because of their relative rarity. Nomenclature A -- inverse of the fraction of total filter area examined per traverse G -- inverse of the total volume of gas passed through a filter m -- number of an equal area, counting outward from filter center M -- total number of equal areas into, which a filter is divided David Leith, Assistant Profes sor at the Harvard Uni versity School of Public Health, holds Bachelors and Masters degrees in chem ical engineering from the University of Cincinnati, and a Doctorate in Environ mental Health Sciences from Harvard. His interests lie in industrial hygiene and air pollution control. American Industrial Hygiene Association Journal Melvin W. First, Professor of Environmental Health Engineering at the Harvard University School of Public Health, is a diplomats of the American Academy of Environmental Engineers ' and a Director of the Amer ican Board of Industrial Hygiene. Dr. First was a Di rector of AIHA from 1964 J967. 103 FMSI 05126 TRAVERSE 1 2 3 4 -5 6 Total, N . Mean Count ' par Traverse, X Std. Deviation of Mean, <r~ 95% Confidence Interval TABLE I Data Tabulation for Particle Sizing by Stratified Counting* < 0.44 57 PARTICLE SIZE RANGE, MICROMETERS 0.45- 0.63- 0.89- 1.30- 1.80- 2.50- 3.50- 0.62 0.88 1.29 1.79 2.49 3.49 4.99 87 54 36 21 24 6 12 12 6 57 87 54 36 21 24 18 18 5.007.09 0 3 3 2 2 1 11 7.10- TOTAL PER 10.00 TRAVERSE 3 300 0 21 03 13 13 23 7 333 57 87 54 36 21 24 . 9 9 1.8 1.2 300 . 7.55 9.33 72 105 to to 42 69 7.35 68 to 40 6.00 48 to 24 4.58 30 to 12 4.90 34 to 15 2.12 13 to 4.8 2.12 13 to 4.8 0.55 2.9 to 0.7 0.45 2.1 to CL3 n -- number of traverses performed in a stratified counting procedure N Ndown Nup P Pdon Pup -- total number of particles of a certain size counted through all traverses N for filter downstream of a particle collector --N for filter upstream of a particle collector ' -- number of particles per volume of gas _ -- P for gas downstream of a particle collector -- P for gas upstream of a particle collector Pt -- dust collector penetration (1 -- ef ficiency), Pdown/Pup rm . -- distance from center of filter to point where microscope is to be focused for equal area m it -- radius of filter x -- number of particles in a certain size range present in one traverse x -- mean number of particles in a cer tain size range found per traverse cr -- standard deviation of x ctx -- standard deviation of x ert -- standard deviation of Pt Microscopic field selection When a filter holder has a small diameter in let relative to the filter diameter, the largest particles may concentrate on that part of the filter directly downstream of the gas inlet. Dennis and co-workers5 found that such radial concentration gradients did not occur for particles smaller than 20 micrometers diameter when filter holders were used which had a ten degree included angle between gas inlet and filter surface. However, filter holders of this design are not always practical because of their large size. To avoid bias when using a conventional holder with small inlet, fields are usually selected at random from a pie shaped piece of the circular filter paper. About twenty fields must be examined2 to complete one unbiased estimate of the particle size distribution. Totally random field selection gives an un biased estimate of particle concentration when a sufficiently large number of fields is ex amined. However, the same result can be obtained with fewer fields by an ordered ap proach. Because the pressure drop across a membrane filter is sufficient to assure uniform gas velocity normal to the filter surface, the volume of gas flowing through each equal, concentric area on the filter will be the same, ' unless the central areas become plugged be cause of excessive particle deposition there. The overall dust concentration for the gas sampled will be the average of the concentra tions found in the gas passing through each of the equal areas. To utilize an ordered approach, the microscope should be focused at the center of each equal area ring present in the sector, and a single field examined. After one field in each equal area has been inspected, all data can be combined to make an entry for one traverse, as shown in Table I. 104 February, 1976 FMSI 05127 Additional traverses are made along different radii of the filter sector. This equal area traverse method for locating counting fields is analogous to the method used for positioning a pilot-static tube when determining the average gas velocity in a round duct. Average gas velocity could be determined by measuring the velocity at many random points within the duct cross section and averaging the results found. However, the number of measurements needed to reach the same statistical reliability using this approach is greater than for the equal area method, and the random approach is not used. By analogy, it is as logical to use an ordered approach for locating counting fields on a membrane filter as it is to use it for locating pitot-static meas uring points in a duct. The distance, r,,,, from the center of a filter of radius u to the midpoint of each of M equal areas can be found from I'm _ / 2m -- 1 2M (1) Here, m is the number assigned to an equal area, starting from the filter center and count ing outwards. Alternatively, values can be found in a reference giving the relative distance from the wall at which a pitot probe should be placed in order to have an unbiased estimate of average gas velocity in a circular duct.910 95% confidence intervals - After, stratified counting procedures have been employed and mean concentrations for particles in each size range calculated, it is important to determine confidence intervals for these values. Systematic sources of error such as anisokinetic sampling, inaccuracies of flow measuring de vices, improper microscope calibration and the like can be minimized through careful experi mental technique.11 Assignment of particles to improper size categories is not a significant problem when trained observers use a standard Porton graticule for determining particle diameter.12 A source of non-systematic sampling error that cannot be eliminated by control of experi mental procedures is associated with random variations in the number, x, of particles in a certain size range which are present in each traverse of a stratified counting procedure comprising in traverses. A Poisson distribution describes the variation in these x values ,2 H. The standard deviation of these values, <r, therefore equals the square root of the mean number of particles in that size range per traverse, x, i.e. a = yfl (2) The standard deviation of the mean, crl , is the standard deviation, er, divided by the square root of n, the number of traverses. ** = y[? = >/~7~ (3) From Equations 2 and 3, where N is the total sum of all particles in a certain size range counted through all traverses. Equation 4 is an application to stratified counting of the expression given by Chapman and Ruhf for the relative error associated with repeated counts of particles in liquid suspen sion.15 Appropriate values for the standard de viation of the mean, <r~ , are given for the data shown in Table I. The number concentration of particles in a certain size range, P, is proportional to x, the mean number of particles of this size per traverse, to the inverse, A, of the fraction of total filter area examined per traverse, and to the inverse, G, of the total gas volume passed through the filter. P =7A G (5) The standard deviation of a product can be found from10 ()'- (^)'+ * With careful technique, the standard error associated with A and G can be made small compared to that for x. The substitution of Equation 4 into Equation 6 gives the relative standard error associated with a measurement of concentration, P. m American Industrial Hyglena Association Journal 105 FMSI 05128 Figure 1-Number of particles counted versus relative error of particle concentration. NUMBER OF PARTICLES ON DOWNSTREAM FILTER. Ndow,, until the standard is no longer contained within the confidence interval, as one can then state with 95% certainty whether or not the standard has been met. When the fiber con centration is close to the standard, it will be necessary to count a larger number of particles to establish with 95% confidence whether or not the standard is met, than when the con centration is clearly well above or below the standard ' For example, after counting 50 asbestos fibers on a membrane filter, one might find that the mean concentration in the air passed through the filter was 1.5 fibers/cc. Equation 8 and Figure 1 show that one can state with 95% confidence that the true fiber concentra tion was 1.5 1.96 X 1.5 X (1/50)'^, or from 1.08 to 1.92 fibers/cc. The upper bound of the confidence interval for this example is below the 1976 OSHA standard of 2.0 fibers/ cc. Counting additional fibers would make the 95% confidence interval smaller, but would be unnecessary if 95% confidence that the standard is met is sufficient. Equation 4 can also be used to determine confidence intervals about an experimentally determined value for dust collector penetration or efficiency. Penetration for particles of a certain size is the ratio of particle concentration in the downstream gas to the analogous con centration in the upstream gas. Figure 2-Number of particles of one size counted on downstream filter versus number of particles of that size counted on upstream filter, with relative error of penetration as parameter. A 95% confidence interval about P will ex tend plus and minus 1.96 times the standard deviation for P. P 1.96 P v'l/N (S) Figure 1 is a plot of the relative standard error of particle concentration, o>/P, against the number of particles counted, N. Applications When asbestos fibers are counted to determine compliance with an applicable standard, it is prudent to determine periodically the mean fiber concentration and associated confidence interval. The count should be continued only The standard deviation of this quotient is1* ( --(/ down/Pup \ 1 J \ ^ J^ Pdown^PuD gpt ^ 2 _?up \ ^ \ ^>uo / V *<lown j a The substitution of Equation 7 into Equation 10 yields gpt _ / 1 i 1 Pt \ N., N*.w. (ID Therefore, a 95 % confidence interval about Pt will extend plus and minus 1.96 times the standard deviation of Pt, as shown in Equation 12. Pt 1.96 Pt Vl/N., + 1/N... (12) 106 February, 1S7# FMSI 05129 Equation 11 indicates that the relative error in penetration for particles of a certain size is only a function of the total number of particles of that size counted upstream, N,,p, and downstream, of the collector. This relationship is plotted in Figure 2. Equation 11 and Figure 2 show, for ex ample, that to be 95% confident that pentration of particles in a certain size range is between 40 and 60% (1.96 o>t= 10% with mean penetration of 50%) it will be necessary to count 200 particles in this size range on the upstream filter and 200 particles in the same size range on the downstream filter. Alter natively, 500 particles counted upstream and 125 downstream would give the same result. However, the fewest total particles that must be counted to achieve a given relative error in penetration will always be found when the number of particles counted on the upstream and downstream filters is equal. This can be proven by differentiating Equation 11 with respect to N'up, setting the derivative equal to zero, and proceeding in the usual manner. When the particle deposit is less dense on the downstream filter, it becomes necessary to make more traverses for that filter in order to count about the same number of particles as are counted on the upstream. Or, a larger field size could be used for the less dense filter. When the particles observed on a filter are separated into many size categories to de termine collector particle size efficiency, it is r.ecessary to observe a large total number of particles to generate adequate confidence in the penetration or efficiency results for each size range considered. When it is desirable to maintain a 95% confidence interval of constant size, i.e. a constant value of crPt for all size ranges. Equation 12 shows that fewer particles need be counted in each size range as penetration decreases. The stratified counting technique can be used with good effect to concentrate the microscopist's efforts on the size ranges where penetration is high. Therefore, it is worthwhile to identify these size ranges as soon as possible by making a preliminary estimate of penetration based on an initial traverse of the upstream and downstream filters. This method of calculating confidence intervals applies whenever concentrations are determined by counting. The techniques out lined above can be used for data from auto matic counting devices such as optical instru ments working on light scattering principles, as well as for data from other automatic counting devices. i Summary An equal area traverse method is described for selecting microscopic fields when counting particles or fibers on a membrane filter. This method is analogous to that used to position a pitot-static tube in a duct when determining average gas velocity. The equal area traverse approach is an aid in avoiding inadvertent counting bias due to nonrandom field selection. Although confidence intervals are im portant to establish the significance of particle concentration or collector efficiency data, they are seldom calculated or reported. When particle size data are generated by a stratified counting procedure, the method described can be used to establish with 95% confidence whether or not mean concentrations are below or above a fixed value. Confidence intervals about values of pene tration or efficiency can be calculated in a similar manner. The stratified counting ap proach allows the microscopist's efforts to be concentrated onto those particle size ranges where small confidence intervals for penetra tion are the most difficult to achieve. Charts have been prepared that make it possible to determine easily the number of particles or fibers which must be counted to assure desired confidence intervals. References . 1. SILVERMAN, L., C. E. BILLINGS and M. W. FIRST: Particle Size Analysis in Industrial Hygiene. Academic Press, New York (1971). 2. edwards, G. h. and j. R. lynch: The Method Used by the U. S. Public Health Service for Enumeration of Asbestos Dust on Membrane Filters. Ann. Occnp. Hyg. 11:1 (1968). 3. morrow, p. e. and t. t. sji rcer: A Point to Plane Electrostatic Precipitator for Particle Size Sampling. Am. lnd. Hyg. Assoc. J. 2.5.8 (1964). 4. billings, c. e. and l. silverman: Aerosol Sam pling for Electron Microscopy. J. Air Pollut. Control Assoc. 12:586 (1962). American Industrial Hygiene Association Journal 107 FMSI 05130 5. sichel. ii. s.: On the Size Distribution of Airborne Mine Dust. J. 5. Afr. Inst. Min. Met. 54:171 (1957). . 6. hoel.p. g.: Introduction to Mathematical Statistics. Wiley. New York (1949). 7. Whitby, k. t.: Determination of Particle Size Distribution--Apparatus and Techniques for Plour Mill Dust. Univ. of Minn. Eng. ExpL Station Bull. No. 32 (January, 1950). ' 8. DENNIS. R., L. SILVERMAN, C. E. BILLINGS, E. KRISTAL. D. M. ANDERSON AND P. DRINKER*. Air Cleaning Studies Progress Report for July I, 1955 to June 30,1956. A. E. C. Contract No. AT(30-1 )841 (March 16, 1959). 9. American Conference of Governmental Industrial Hygenists: Industrial Ventilation. 13th ed. P. O. Box 453, Lansing, Michigan (1974). 10. hemeon, w. c.: Plant and Process Ventilation. ' 2nd ed. Industrial Press, New York (1963). 11. hawksley. p. g. w.. s. nADziocii and j. H. b'lackett: Measurement of Solids in Flue Gqses. British Coal Utilization Research Assn., Leatherhead, Surrey, England (1961). 12. fairs, g. L.: XII--Developments in the Technique of Particle-size Analysis by Microscopical Ex amination. J. Roy. Microscop. Soc. 77:209 (1951). 13. corn, m.: Statistical Reliability of Particle Size Distributions Determined by Microscopic Tech niques. Am. Ind. Hyg. Assoc. J. 26:8 (1965). 14. hkrdan. g.: Small Particle Statistics. 2nd ed. ' Academic Press, New York (1960). 15. cHxrMAN, h. M. and r. c. ruhf: Dust Counting Reliability. Am. Ind. Hyg. Assoc. Quart. 76:201 (1955). 16. arkin. h. and r. r. colton: Statistical Methods. 5th ed. Barnes and Noble, New York (1970). Accepted October 15, 1975 submission Of new manuscriDts " . The American Industrial Hygiene Association JOURNAL, as the Official publication of the Association, provides a medium for timely publication of scientific articles and technical reports covering the broad fields of industrial hygiene and occupational health. These relate to the detection, evaluation and control of problems con- cemed with occupational, en vironmental and radiological health, as well as air pollution, human and animal toxicology, product safety and related sub jects. Worthy contributions to the literature are welcomed. Manu scripts will be acknowledged and reviewed for acceptance promptly. When approved, they will be scheduled for publication at the earliest possible date. Only manuscripts submitted exclusively to the AIHA JOUR NAL, not published elsewhere and not being considered by another publisher will be reviewed. Manu scripts not meeting these stipula tions should not be offered. Detailed guidelines covering manuscript preparation to suit the new format requirements of the JOURNAL will be published frequently. Reprints of guidelines also are available to authors. These current guidelines should be re viewed carefully and observed, before the final draft of a manu script is prepared. Requests for reprints of guidelines for authors and submission of manuscripts should be made to the editor, American Industrial Hygiene Asso ciation JOURNAL, 66 S. Miller, Rd., Akron, OH 44313. Galley proofs will be provided for final reading, but not for re writing or revision, just prior to publication. Reprint orders are made available to authors with galley proofs. ' Inquiry regarding current status of a manuscript previously submitted for consideration should be made in writing. Be sure to mention the manuscript's JOUR NAL number, if known, plus the first author's name and the manu script's full title. 108 February, 1976 i * FMSI 05131 . 1975 V. P. 'fealth A nal. I. Hyygiene ionnaMark: Counting Asbestos Fibers by the Most Probable Number Method \ PARKER C. REIST,5c.D. 9' * Department of Environmental Sciences and Engineering, University of North Carolina. Chapel Hill, North Carolina 27514 A procedure for cvnluuliiif: asbestos fiber counts is described which uses the most probable number method of bacteria counting. This technique is faster than conven tional counting methods, with approximately comparable accuracies, although it suffers from a lack of rigor and requires the observer to estimate fiber concentrations to within an order of magnitude before counting. For the routine assessment of a large number of asbestos samples this procedure would seem to be more desirable than conventional counting because of the economy of time as well as being easier on the observer. ' Introduction Recent findings on the toxi city of asbestos have led to increased interest in sampling and analytical proced ures for determining the concentration of asbestos fibers in air. The Occupational Safety and Health Administration has estab lished an interim eight-hour time weighed average airborne allowable concentration of five asbestos fibers greater than 5 microns length per cubic centimeter of air, and this standard will be lowered to 2 fibers/cm3 on July 1, 1976. Evaluation of asbestos fiber concentra tions in air is carried out using samples col lected on membrane filters and viewed with phase contrast illumination at 400X-450X.1-2 This method is the standard field sampling method adopted by the Public Health Service. Fibers arc assumed to be distributed ran domly over the filter surface and at least 20 but no more than 100 fields arc to be viewed. At least 100 fibers arc counted which gives a 95% confidence limit of ziz 20%. Because of the relatively small amount of sample collected, more than 100 fields would< have to be viewed if one were to find 100 fibers on a 10 minute sample collected at a rate of 2 liters per minute from air containing 5 fibers (5 microns length) per milliliter, and if the concentration were only 2 fibers per milliter, 435 fields would have to be assessed.3 Of course, the number of fields necessary could be decreased by in* creasing the sampling time--in:uthe latter case a 90 minute sample would yield 100 fibers in 50 fields--but the flexibility of short-term samples is then lost. Counting fibers is a tedious and time consuming business fraught with a number of subjective decisions for the microscopist to make. For example, in the case of two fibers lying side by side he must decide when a fiber is a fiber or when it is only a large particle. (It is considered that any par ticle having an aspect ratio of three or greater is a fiber). If the sample is relatively light, a great deal of information is lost which actually can be used to determine fiber density. Be sides actually counting the number of fibers, there >s another quite distinct statistical method which could be used for estimating the number of fibers randomly distributed over a surface, the so-called most probable number method. In this paper the method will be applied to asbestos fiber counting FNISI05132 380 May, 1975 anil the advantages and limitations of its use will be discussed. Theory The concept of the most probable number method for csitmuting randomly distributed number densities was first described by McCrady4 and more recently by Cochran* and principally was applied to the problem of estimating bacterial densities. Chapman6 applied the method to dust counting but it was not received with much enthusiasm for reasons which will be discussed later. Most recently the technique has fallen into dis favor even for estimating bacterial concen trations in milk and water samples, mainly because of the advent of more direct membrane filter techniques. For fiber count ing however, the method appears promising because it eliminates the need to resolve individual fibers and with the low densities normally found, gives the same counting accuracy in a much shorter period of time and with much less eyestrain. Consider a filter of area A which is broken up into a fields. If there arc m fibers dis tributed randomly over the filter surface, then the average number of fibers per field, /.is . The probability, P, than n fibers lie in a cer tain field can be expressed using the Poisson relationship P(n> ='n!e't <2> provided that a is a large number. The probability of a field being void of fibers is, from above P(o) = e~t . (3) If the void fields are distributed randomly across the filter and L fields arc sampled, the probability that exactly / of these fields will be void is r>() = /i {L-~t)llP(o)V [1 " P(o)]tL (4) Equation 4 gives a distribution of probabili ties which is very small for small l and rises TABLEI 93% Confidence Interval for a Given Average Number of Particles Per Field Particles/Field True Average 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1-3 . 1.4 1.3 1.6 1.7 1.8 1.9 2 . 95% Confidence Interval Expressed a* a Percentage of the True Average (left column, minus; right column, plus) L 25 100 -- 229.J 76.6 -- 156 69.6 -- 111.3 54.8-- 90.3 56.3 -- 79.4 52.7 -- 76.2 47.5 -- 55.2 45.6 -- 73.8 44.3 -- 75.6 43.2 -- 57.5 L - 50 75.8 -- 138 58 -- 79.7 49.9 -- 58.8 42.1 -- 59.8 41.1 -- 52 37.4 -- 40.8 35.6 -- 47.4 33.6 -- 38.8 32.8 -- 38.9 32.2 -- 39.7 ' *, L = 100 56.7 -- 84.9 41.3 -- 51.7 34.1 -- 43.1 31.5 -- 36 27.8 -- 36.2 27.9 -- 33.6 23.9 -- 32.2 25.1 -- 31.5 22.3 -- 31.4 21.9 -- 31.7 23.9 -- 28.7 21.8 -- 29.6 ' 22 -- 26.8 23.3 -- 28.4 22.6 -- 29.9 20.9 -- 26.8 21.4 -- 28.7 22.1 -- 25.6 20.2 -- 33.1 23.5 -- 30.3 L = 200 43.3 -- 51.1 29.8 -- 35 25.9 -- 28.1 23.1 -- 25.5 21.8 -- 22.6 18.7 -- 22.8 18.8 -- 21.7 18.1 -- 21.2 17.6 -- 19.4 17.4 -- 19.5 17.4 -- 18.1 16.3 -- 18.5 16.5 -- 19.2 15.6 -- 18.1, 15.9 -- 19 ,, 16.3 -- 17.9 16.8 -- 19 16 -- 20 15.2 -- 21.4 15.7 -- 20.2 I I Americai to some small aj that haV Turning! conccnti The a pic proc tion. A determir present 1 the natu fields sc fields co most prj can be i As th creased, { creases.j range foi at varioi Equatior 100 fid 30% rcgardle: provided density i Figure microscop FMSI 05133 American Imliixlrial Hygiene Ax*ocialion Journal to some maximum value before becoming small again! The most likely probability, that having the largest value, occurs at l = Lc-> (5) Turning this around, the most probable fiber concentration is thus / = /(-y-) ' (6) The above development results in a sim ple procedure for assessing fiber concentra tion. A number of fields are scanned to determine only whether fibers arc or are not present in any given field. Then by taking the natural log of the ratio of the number of fields scanned divided by the number of fields containing no fibers (Equation 6) the most probable number of fibers per field can be estimated. ' As the number of fields scanned is in creased, the accuracy of the estimate in creases. Table I shows the 95% probability range for scans of 20, 50, 100 and 200 fields at various fiber densities, as calculated from Equation 6. Thus, for example, assaying 100 fields will yield results within about 30% of the true value 95% of the time, regardless of the total number of fibers seen, provided, of course, that the average fiber density is somewhere around two fibers per Figure I. Various orientations of fibers in the microscope field. 3RI field or less. Assaying 200 fields will in crease the accuracy of the 95% confidence interval to something less than 20% of the true count. For the ease of L = 200, the accuracy ap pears to increase with an increasing average number of particles per field. This--will continue until the point is reached where there is a good chance that every field con tains at least one fiber. This occurs when the average number of fibers per field is slightly in excess of three. Thus for this method it is necessary to estimate ahead of time the fiber concentration to within one order of magnitude so that there will be about 0.2 to 2 fibers per field, or, if there appear to be plenty of fibers in evidence, to then use the direct count method. Of course, the number of fibers per field can be easily varied by changing the field size. Definition of a Field Containing a Fiber Thus far, for the purpose of development of the theory, fibers have been considered as if they were particles. But they are not. A fiber has length and as such may starrin one field, extend through another or several others and finally terminate in yet another field. Figure 1 illustrates such a situation. The fiber originates in the upper left hand corner, continues through the lower left hand corner and then ends in the lower right hand corner. The fiber in the upper right corner represents no problem. There are several ways in which the fiber that passes through several fields can be treated. First, only the lower (or upper) end of the fiber can be considered, and the field that it lies in then is a field not devoid of fibers. In the rare ease where the fiber is perfectly horizontal, some convention, such as choosing the left hand side of the fiber would be appropriate. A field would be considered. blank it the lower end of a fiber were not in it. Thus, in Figure 1, only the fields on the upper right hand side and lower right hand side would be considered, to contain fibers. Since each fiber is asso- FMSI 05134 382 May, 1975 dated with one "lower" end, the estimated whether the fiber is in the field or not. number ends would equal the estimated num ber of fibers. Experimental A second approach would be to call a lit order to determine the efficacy of the field void only if it contained no fiber ends proposed counting procedure a number of at all. In Figure 1 only the lower left hand > asbestos samples were counted using a di field would be considered to be void. The rect counting method and a record was kept most probable number of ends would then of the number of void fields observed. Most be estimated and since each fiber has two probable number data as determined from ends, the most probable number of fibers would be the estimated number of ends di vided by two. An obvious disadvantage of this approach is that the upper limit of density which could be used is half of what otherwise would be used. The most reasonable approach is to assess only one end of the fiber. If the fiber is a bundle with a rough end lying on the edge of a field so that there is some question as to whether it is in or outside of the field, then the same rules as those used in particle counting could be applied to determine Figure 3. Plot of most probable number versus direct count for data from computer simulation for 20 fields. DIRECT COUNT. FIBERS RER FIELD Figure 2. Most probable number data ns deter mined from Hquation A versus direct count. direct count for data from computer simulation for 100 fields. FMSI 05135 American iniluslrial Hygiene Association Journal Equation 6 arc plotted as a function of the direct count which was observed for the samples and the results arc shown on Fig ure 2. Also shown on this figure arc the 95% confidence limits for both the MPN method and direct counting. Although there is some spread in the data, the general repro ducibility is apparent. However, it appears that the most probable number method con sistently gives results which arc lower than those determined by direct count. This ob servation is consistent with a similar one of Chapman's who surmised that the differ ence could be due to failure to sec a single particle in an otherwise void field. For more extensive work a computer sim ulation was developed in which fields of 10,000 bits were assigned particles randomly corresponding to some preset particle den sity. Then fields of various sizes were ran domly chosen and the average number of particles per field determined using the two methods. In addition, the absolute number of particles per field was determined. From this simulation it was possible to carry out non-biased counts using both conventional counting and the most probable number method for samples of 20, 100 and 200 Figure 5. Plot of most probable number versus direct count for data from computer simulation for 200 fields. .. 383 fields. These data tire shown in Figures 3, 4 and 5. Unlike the actual experimental data, however,' there appears to be no bias toward the direct count information, indicating that the higher count averages noted on the actual direct counts results from a bias introduced by the observer rather than by the technique. Similar to Figure 2, error limits for the 95% confidence interval arc shown as dotted lines. Advantages and Disadvantages The advantages of the most probable number method arc threefold. .Sampling times are shorter, lighter samples with less chance of overlap can be used, and counting times arc shorter. Using fairly light samples for asbestos concentration assessment means that shorter sampling times arc needed in the field, often an advantage to the industrial hygienist. For a given number of fibers ob served, the MPN method implies greater ac curacy if this total number is relatively small. There is less eyestrain for the microscopist since he only has to determine whether there is or is not something there, and not rer solve a specific number of fibers. If prob lems of fiber clumping have occurred they will be more evident because of the lighter sample density. Finally, since fields arc be ing counted instead of particles, the counting should proceed at a faster pace. For exam ple, in discussing particle counting by the most probable number method. Chapman pointed out that one observer could deter mine particle concentrations about twice as fast using the most probable number method compared to standard counting methods, while another observer was three times as fast using the MPN method. We have not studied counting times objectively, but sub* jcftivcly the people in this laboratory who have compared the two methods for count ing asbestos fibers also feel that the most probable number method is much faster. The principal disadvantage of the most probable number method is that it is not rigorous. An observer could," in theory, ac curately count the total number of fibers deposited on a filter whereas using the most probable number method, even if the whole filter were assessed, the observer would in the end still only have an estimate of the number of fibers present. In addition, for a given number of particles per field, the 95% confidence intervals for the direct counting method arc slightly narrower than for the most probable number method, the effect becoming increasingly pronounced when the average number of fibers per field exceeds two. The accuracy is sufficient, however, for routine asbestos counting. ' Summary A procedure for evaluating asbestos fiber counts is described which uses the most probable number method of bacteria count ing. This technique is faster than conven tional counting methods, with approximately compariblc accuracies, although it suffers from a lack of rigor and requires the ob server to estimate fibers to within an order of magnitude before counting. However, for the routine assessment of a large number of asbestos samples this procedure would seem to be more desirable because of the economy of time as well as being easier on the observer. References 1. Edwards G. H,, and J. R. Lynch: The Method Used by the Public Health Service for Enu meration of Asbestos Dust on Membrane Filters. Ann. Occup. Hyg. II: (1968). 2. Joint AIHA-ACGIH Aerosol Hazards Evalu ation Committee: Recommended Procedures for Sampling and Counting Fibers. /Imer. Ind. Hyg. Assoc. J. 36:83 (1975). 3. Annon.: Occupational Exposure to Asbestos, p. viii-5, HSM 72-10267. U.S. Dept. H.E.W., NIOSH. Washington (1972). 4. McCrady, M. H.: The Numerical Interpreta tion of Fermentation Tube! Results. J. Infec. Dis. 77:183 (1915). m:!! ; 5. Cochran, W. G.: Estimation of Bacterial Den sities by Means of the Most Probable Num ber. Biometrics <5:105 (1950). 6. Chapman, H. M.: Dust Counting by the Most Probable Number Method. A.M.A. Arch. In- dustr. Hyg. 8:234 (1953). FMS1 05137 JOURNAL OF PAINT TECHNOLOGY. . , . Handling Asbestos th, iSt 5 v > i7- SafeaHfeSliBMK, Jl - ** Chrysotile Asbesto4s -in---P---l-a-s--t-i-c-^- L. Myers ;iv -- i" t*~` ~r` <--* - ir > . ' Union Carbide Corporation** ** . ` ! r 'Jt.-v ' `Jfc * ' 1': ) /* ,: 9 ' ' r ' ^ deal of attention and publicity in recent years, especially after it was designated a "target health hazard" by OSHA and a "hazardous air pollutant" by the EPA. Many of the articles on asbestos by the press have been emotionally oriented or distorted and, in some cases, stories have been sensationalized, based on obvious misinterpretation of facts. The use of half-truths or unsubstantiated statements has led to general con fusion and the unfair castigation of asbestos and products containing asbestos. The purpose of this paper is to put the matter of asbestos use and asbestos hazards into a logical and practical perspective. In this paper, the different types of asbestos and their many uses are discussed, along with government regulations controlling the use of asbestos. The health hazards associated with asbestos, both occupational and environmental, and some industrial ex perience with air sampling and dust control measures are also covered.^ ,. IS KEY WORDS: Asbestos; Plastics; Air pollution; Toxicology..? JM?* flow tile and in phenolic molding compounds. It is also used in other plastics such as polypropylene, poly- . ester, nylon, melamine, epoxy, sili-, cone, and vinyl. Asbestos provides a ' valuable function in such products as brake linings, clutch facings, electrical 9 components, automotive parts,' furni- j hire, boats, sealants,' coatings, adhe sives and mastics. The most important ' functions of asbestos in plastics are re- inforcement, dimensional - stability, heat resistance, flow control, and gen- '. eral-purpose filling. Most of the func tions are supplied by short-fiber chrysotile ' asbestos ' fiber," although What is Asbestos? Asbestos is a commercial or generic term used to describe six naturally occurring "asbestifonn" minerals that are fibrous, hydrated metal silicates. The six varieties are divided into two classes--serpentine and amphibole-- based on their crystal structure. Chrysotile is the only member of the serpentine class, while the aanphiboles include crocidolite, amosite, anthophyilite, tremolite, and actinolite. Chrysotile is by far the most-used variety and accounts for over 95 of U.S. consumption, as noted in Table 1. - . : Crocidolite, also known as blue as bestos, is imported from South Africa. Because of its high mechanical strength and good resistance to adds and alkalis, it is used to reinforce a limited variety of plastics where its pronounced color is not objectionable. Amosite, also imported from South Africa, is used primarily in thermal Presented at the Golden Gate Society's Man agement Seminar held in San Francisco, Calif., June 16. 1975. * Metals Div., Niagara Falls, N.Y. 14302. insulation. Although there are some deposits of anthpphyllite in the U.S., most of it is imported from Finland. longer chrysotile fibers and other as bestos varieties are sometimes re quired for particular properties. "* It is used primarily as a fifier for ' polypropylene and in insulating rrra- Why Use Chrysotile? terials.. A comparison of the four Among the several advantages of varieties of asbestos which are of ^ chrysotile, which set it apart from merdal importance is presented -in 9 ' die amphibole minerals and account Table 2. It should be noted that there" " for its widespread and increasing us are significant differences between age, are world-wide availability, me chrysotile and amphiboles with regard chanical strength, flexibility, positive to chemical composition and certain surface charge, low iron content, soft physical properties. ness, and low refractive index. It is ' V _ f:, ' Where is Asbestos Used and Why? . conservatively estimated that chryso-. tile asbestos is used in over 3,000 ap plications and, in most of these appli cations, it is an essential ingredient Asbestos has served mankind for for which no replacement is readily more than 100 years in a broad vari available. `9 ety of applications. The general areas The information in Table 1 shows in which asbestos is used in die U.S. that the use of chrysotile asbestos and are shown in Table 3. Based on infor its share of the total market are stead mation from asbestos producers and ily increasing. This is due partly to consumption surveys, it is estimated technical advances permitting die that the plastics industry uses about broader use of chrysotile in plastics 33 of the 800,000 tons consumed an and the general decline in die use of nually, which makes it the largest asbestos in certain fireproofing and in single user of asbestos fiber. sulating materials. In addition, there The largest uses of asbestos by die is increasing evidence fliat crocidolite plastics industry are in vinyl-asbestos and amosite are more hazardous to ' '- 9 - _ Vol. 47, No. 611, December 1975 FMSI 05138 V' Year 1967 1968 1969 1970 1971 1972 Table 1--Apparent U.S. Consumption of Asbestos, Tons* Total . Chrysotile Amosite Crocidolite 720,583 817363 784,321 728,131 758371: 808354 - 686,044 (95) 12358(1.7) . 775,711 (95) 20,467(23) 749,708(96) V 695,770(96) Si# ;14461 (2.0) ipi, .14380 (13)`^S 14,917 (2.1) 13,965(1.7) 10,558(1.3) 8,936(1.2) 6,953(0.9) 5374(0.7), (a) Information baled'on import ai Bureau of Mines' Minerals yearbooks. b (b) Percent of total shown in parentheses.* Table 3--Apparent U.S. Consumption of Asbestos By General Use Areas Area of Use Percent of Consumption Construction , Floor tile Felt paper j Friction and packing r, Insulation . ,,t Textiles If 4150!;*v- 15 J 14; 4 fijP*' Otheri Tt . human health than chrysotile.1 Since . mt- iipiMaiis* - or caused solely by the Inhalation of * oers of the International Agency for 1970, the use of crocidolite in Britain asbestos fiber.'^Another important con-" , Research bn'Cancer (a division of the has been restricted after a panel of sideration is the relation between World Health Organization): "There experts "concluded there was suffi cigarette smoking and lung cancer, as is evidence of ah association of meso- cient evidence to suggest other types reported by Dr. E. C. Hammond and thelial tumors with air pollution in of fibre should be substituted for Dr. I. J. SeHkoff.* In this Study, they the neighborhood of crocidolite mines crocidolite wherever possible."*. * - ; ' - ri- - > reported that: "If seems clear, then, that lung cancer is uncommon among and of factories using mixtures of as- bestos fiber types. The evidence*"re- " What is the Asbestos Hazard? If is readily accepted that asbestos, like many other foreign bodies, can cause disabling lung damage (pul monary fibrosis), commonly referred to as asbestosis. This disease and bronchogenic carcinoma (lung can cer) are the two most common as bestos-related diseases. It is important fo note that, based on epidemiological data, these diseases have occurred primarily in workers with high, long term exposures to asbestos dust. It is of further interest that one noted re asbestos insulation workers who have lates to conditions of many years ago. - no history of cigarette smoking and There is evidence of no excess risk that if the risk is increased, such in of mesotheliomas from asbestos; air . crease is not great." pollution which has existed in the . >X- A third disease, mesothelioma, has... more recently been -associated -with t-% persons exposed :to asbestos. Meso-* - ` neighborhood of chrysotile" and amo-t site mines. There are reported differ ences on incidence of mesothelioma thelioma is an extremely rare cancer ; of the lining of the chert (pleura) or " the abdominal cavity (peritoneum). In contrast to the lung diseases, there ` " between urban and rural arras, the ' causes of which have not been estab lished. There is no evidence of a risk to the general public atpresent."`j>f is some evidence that mesothelioma' The same body quoted above has can occur after brief exposures to also concluded that there is, at pres relatively high fiber levels. ent, no evidence of lung damage by . searcher has reported that neither of According to the 33-memfoer Ad asbestos to the general public; and these diseases is peculiarly related to visory Committee on Asbestos Can- such evidence as there is does not in- ' . cheat* any risk of cancer resulting from asbestos fibers present in water, Table 2--Comparative Data for Asbestos Minerals* beverages, food, or in the fluids used *. Formula * Chrysotile Crocidolite Amosite * AnthophyIlit< 3Mg02Si0s'2Hs0 No.O FesO. ?.5Mg05.5Fe0 7MgO 8SiO, HSC 8Si03 H20 8SiOs'HjO * ^ i for die administration of drugE/^jtt^> . * .: " A**,. Z While there seems rto' be general '4s agreement that tbe public is not in n Composition, % SiO. MgO FeO FesO, 37--44 39--44 0--6 0-5 , 49-53 0-3 V 13--20 17-20 ' 49--53 1--7 34--44 _ ' 56-58 * 28--34 ; ' ' '' 3--12 v: any present danger from asbestos, it is also recognized that excessive, long term occupational exposure can cause ~ serious health problems. Abo, if man made emissions are not controlled,., A1.0,, H.O CaO NasO CaO-f-NasO 0-2 . 12-15 0--5 - s_ 2--5 " _ ' 4--8 -. . 2-9 2-3 - " r-.r?!-, 1 __ --3 0-2 1--6 _ '-- then ' environmental contamination could could approach harmful levels. During the past two years, significant legislation has been enacted by the Federal government to reduce and Crystals Fine fibers Brittle fibers Prismatic Prismatic control occupational exposure to as Color Gray green Blue . ; Gray/brown Gray bestos fibers and to minimize fiber Texture Soft silky Harsh Harsh Harsh emissions to the environment. Addi Flexibility Very good Good Good Poor tional standards or regulations have Hardness, mohs 23-4 4 53-6 5.5-6 been proposed or enacted 'by many Fiber diameter, A 180-300 600-900 . 600-900 600-900 states and local governments. : ' Tensile, mpsi Surface charge Resistance to add 800 Positive Poor 600 Negative Good 200 Negative Good . > <4 Negative = Very good ; *. m drtwtert-: h* i* v * Summary of OSHA Regulations Resistance to alkali Good Good Fair , , . Good r<*r The William-Steiger Occupational (a) Sources: Modem Plastics Encyclopedia (I97S), and Encyclopedia of Chem. Tech., Vol. 2. Safety and Health Act of 1970 be came effective on April 28,1971, with FMSI 05139 Journal of Paint Technology the following Congressional puipose: Improper interpretation of die reg EPA Standards "To assure so far as possible every working man and woman in the na tion safe and healthful working con ditions and to preserve our human resources." The Act established the Occupational Safety and Health Ad ministration (OSHA) within the De partment of Labor, which has re sponsibility for administration and enforcement. Research and" related functions are handled by the Depart-, ^' ment of Health, Education and Wel ,w fare (HEW) through the National ' Institute of Occupational Safety and Health (NIOSH). Five million em ployers and 60 million of the nation's 80 million workers are covered by OSHA. Specifically excluded from coverage are government employees and operations which are protected under other federal health and safety laws. In a news release issued January 4, 1972, OSHA announced a Target Health Hazards Program aimed at improving health factors associated with working conditions. The follow ing five substances were designated to be the focus of initial and con certed efforts by OSHA and NIOSH: asbestos, cotton dust, silica, lead, and carbon monoxide.' <= ' ' " At the present time, new standards have been established only for asbes ulations has created many miscon ceptions about equipment and pro cedures needed to properly use asbestos. If exposure limits are not exceeded, there are no further complianee requirements except for med ical examinations. Medical examina- ; tions are required for all employees in any occupation exposed to airborne concentrations of asbestos fioere.^The^* examinations are relatively simple and.3* should cost no more th^n $50 per#.* year, per employee* Respirators and special clothing are required in the construction trade for the spray application of insulation and fireproofing materials, and for tbe removal of such materials. This spe cial protection is not required for anyother use of asbestos unless exposure limits are exceeded. This is also true for other items such as specially equipped tools, change rooms, clothes laundering,* and waste disposal. Res pirators are not a substitute for en gineering controls, but the law allows their use while controls are' "being implemented, in special situations where controls are not feasible or adequate, in emergencies, and for in frequent short-term, job assignments. Caution labels are required on prod ucts containing asbestos except where Whereas OSHA is responsible for the protection of the worker, the En vironmental Protection Agency (EPA) is charged with improving the en- ' vironment to which the general public, - is exposed. On March 31, 1971, as bestos, along with beryllium ^and || mercury, was identified as a "hazard- ous air pollutant" by the Admamstra-.jjt tor. of the EPA. National ,EmissionMl Standards for aAestoswere then pub-`!?` lished by, EPA in thie Federal Regis-- ter,}Vol. 38, No. 66, April 6, 1973. " Although.no numerical emission stan- - dards were established, operating cri teria are prescribed,; to. prevent ' or ,, limit asbestos emissions to the out- ` side air from asbestos mills, roadways, / oertain manufacturing operations, building demolition, and the spray-on application of materials used to in- * sukrte or fireproof equipment and machinery. The law further requires that spray-on materials used to in sulate or fireproof buildings, struc tures, pipes, and conduits shall non- ._ tain less than 1* asbestos on a dry- weight basis. This should significantly reduce emissions., to which the gen eral public may be exposed, especially . in large urban areas. , ' ". . . the Administrator (of EPA) . has determined that, in order to pro tos, although, of the 8,000 toxic sub the fibers have been modified by a vide an ample margin of safety to stances on the NIOSH list, only 500 bonding agent or other material to protect the public health from as are covered by standards and many prevent dusting during any normal bestos, it is necessary to control emis- of those need updating. The new Standard for Exposure to Asbestos subsequent use or handling. Besides raw asbestos fiber, products which skms from major man-made sources of asbestos emissions into the atmos Dust was published in the Federal Register, Vol. 37, No. 110, on June 7, require package labeling could in clude: dry acoustical spray products phere, but that it is.not necessary to prohibit all emissions. , .W , ' ` 1972. The basic exposure standard is and joint cements, unsaturated roof- - "In this determination,-die Admin an 8-hr time weighted average ing felt and textiles, and some , in istrator has relied Ton the. National (TWA) of five fibers, longer than 5 sulating products made without ade Academy of Sciences' report on as- ' micrometers, per cubic centimeter of quate binders. The labeling of a prod bestos, which concludes: "Asbestos is air. The TWA limit is to be reduced to two fibers per cubic centimeter on July 1, 1976. A peak concentration of 10 fibers per cubic centimeter is uct does not prohibit its use. It should be noted here that in at least 90* of the products containing as bestos. the fibers are solidly locked too important in our technology arid. v. economy for its essential use to be \ stopped. But, 'because of the known serious effects of uncontrolled inha not to be exceeded at any time. All into the product thereby presenting lation of asbestos minerals in indus of the fiber concentrations are those to which an employee may be ex little danger of dust generation dur ing normal use and handling of the try, and uncertainty as to the shape and character of the dose-response posed without protective clothing and product.* - . , curve in man, it would be highly im- equipment. The first basic require ment of the new standard is monitor ing to determine whether or not fiber concentrations are in excess of the exposure limits. Some asbestos sup pliers provide a monitoring service to JOHN L. MYERS, Marketing Manager for the Calidria Asbestos Group in Union Carbide's Metals Div.. received customers, and a similar service may be obtained from state health depart ment officials, insurance carriers, or private consultants. The law requires that monitoring be repeated as nec his B.S. Degree in Chemical Engineering from Purdue University in 1951. Joining Union Carbide that same year, he served in the Nuclear Division until 1966, when he be came a Research Engineer in the asbestos group. He was promoted to his present position in 1970. essary to ensure that employees are not exposed to levels in excess of the exposure limits. Vol. 47, No. 611, December 1975 FMSI 05140 prudent to permit additional con Table 4--Typical Air Sampling Dust-Control Measures tamination of the public environment with asbestos. Continued use at min imal risk to the public requires that the major sources of man-made as bestos emission into the atmosphere be defined and controlled.' Tf}*^wsr> _1\ Results' . , Ceiling , Type Plant - Concentration, . Or Operation ; Asbestos Fibers/cc Floor -Xilej Po1?H3m! ... . Asbestos producers and users are spending a considerable amount of time and money on various dust-con- ^ . trol measures. Conventional means to'rj J. achieve minimum dust levels include: , filtered ventilation systems'on process k .Phenolic compounding equipment, local ventilationTor: ww&mM What is Industry Doing? Handling phcnoBCj Caulks and sealants and4_simflittodk',f"convernOT3to^lr| 'wetted" operation^'leakjjproof -pack-' The Asbestos Information Associa Gypsum compounds aging,"vacuum clean-up, "more care in'A* tion/North America reports that, dur ing the past 30 years, the asbestos in (a) Source: Union Carbide Corp.,* from^ air monitoring reports. us - bag disposal and other asbestos waste^, ; handling, and automatic bag'openers?:*, dustry has spent millions of dollars to Unusual innovations include: pel improve mining, milling, and manufac letized asbestos, special packaging, '?!; turing methods.* The establishment of and treated products.'^^^ safer working conditions has been a Only short-fiber chiysotile'is avait:J'#f prime target and this work continues able OSHA level is 10 fibers/oc. In able as pellets, but this product serves ' '*; unabated and in close association with most cases, the TWA exposure level a fair portion of the asbestos market government agencies and independent would be well below OSHA stan Pellets not only reduce dust during - medical researchers." r v - dards. Most of the data were collected ;^/conventional handling but they are The ultimate goals of the asbestos before the installation of any special f also available in bulk hopper oars and" f industry are: Reduction of work-area dust-control measures. Monitoring of-'**.;* can be transferred and used in totally! . dust to minimum levels; Protection ten shows that obviously dusty con- , ' enclosed systems. Barring leaks in the r of workers from asbestos-related dis ditions are caused by materials other ?' 'system, dust in work areas is virtually A- eases; Maintenance of environmental than asbestos. This does not preclude eliminated. Used , in bulk, asbestos . - emissions at levels low enough to pre the need for controls, but it could pellets also reduce shipping - costs,. clude public endangerment.' change their scope and facilitate com eliminate ' warehouse storage and ' pliance with government regulations. handling, facilitate automation, reduce Air Sampling In order to comply with OSHA standards and to determine the need for dust control measures, air monitor ing should be conducted in areas where asbestos is regularly handled or used. OSHA standards require that "all determinations of airborne con centrations of asbestos fibers shall be made by the membrane filter method at 400-450X (magnification) (4 milli meter objective) with phase contrast illumination."10 The equipment for collecting air samples costs less than $400 and is readily available. A phase contrast microscope can be obtained for as little as $600, or an existing microscope can be modified for "counting" the asbestos fibers in com pliance with NIOSH criteria.11 Air samples have been collected and analyzed on a regular basis by the asbestos industry for years. Ex cept for a few applications where dust control is an engineering problem, in dustry is finding that dust levels are Because of "bad press," asbestos is clean-up, and eliminate bag handling , frequently ordered out of use without and disposal. The pellets contain no- *"* regard to whether or not a hazard binder and are friable enough to be . actually exists due to air contamina dispersed in dry .form- or, in aqueous vS^. tion. If acceptable dust -levels are or resinous systems with conventional feasible, there is no need to replace high-shear grinding equipment.1* . ,, ,, asbestos at the expense of product Several types of special packaging quality or economic penalty. Gordon are currently available, and suppliers * s, Everett of EPA points out that in consider customer requests for unus- .. formation on the biological effects of ual requirements. The floor tile indus- * asbestos is very limited and that the try can obtain asbestos in plastic bags - ^ effects of many substitutes have not which can be added directly to the - :;f* been investigated at all. Before as compounding operation. Asbestos .in . bestos is replaced, it should be cer bleached paper. bags assembled ..-with ;, tain that a safer alternative is avail water-soluble glue and printed with * able.1* ' \ - water-dispersible ink: can he added Obviously there are more people . directly to paper-making finishes or exposed to products containing asbes acoustical tile formulations. tos than there are to raw asbestos Although "wetted" asbestos is not ; fibers. As noted previously, more than generally .available, most suppliers 90% of the asbestos used in this coun are working with customers to pro try is in products in which the asbes vide "dustless" products. When justi- tos is "locked in" or bound with ce fled by market demand, asbestos can ment, plastics, or other binders, so be treated with water, mineral spirits, that there is no release, or at least no glycol, or other materials compatible ' significant release, of fibers in work with the application or system. ^ - , areas or to the environment. Materials already within acceptable standards or that minimum changes are nec or products with locked-in fibers would include: floor tile, polyester Conclusion .. - - essary to achieve compliance. Al resins, phenolics, sealants, coatings, Asbestos is one of industry's many though data on many asbestos/plastic brake linings, friction materials, rub raw materials which involves a po applications are not available, the ber, roofing compounds, and rein tential hazard when not used with summary in Table 4 is typical of our forced plastics. Since an abrading ac reasonable respect and care. Although measurements of dust levels during tion on some of these products could all forms of asbestos are recognized asbestos handling in various types of release asbestos fibers, appropriate as hazardous to health when inhaled plants and operations. The dust levels monitoring and/or control measures excessively, there is growing evidence reported are ceiling concentrations, should be instituted if it is thought that crocidolite and amosite are more and it should be noted that the allow- that such action would release fibers. hazardous than chrysotile. Fortunate- FMSI 05141 Journal of Paint Technology ly, the plastics industry uses primarily chrysotile asbestos and, in most prod ucts, the fibers are locked-in to pre vent airborne contamination. Al though asbestos dust levels are gen erally lower than expected, industry continues to expend large amounts of time and money to further improve the quality of tht^' workplace. ^Al though the general puiblic is not coK rently in danger,^occupational controls ate required to prevent future en^ vironmental contamination, `f Chrysotile asbestos is an important and necessary raw material, vital to' the nation's safety and economy; and, with proper control, it can be nsed safely and in compliance with govern ment regulations. Medical, scientific,' government, and industrial personnel must continue to work closely to gether to establish reasonable expo-: sure limits, provide safe work areas, and eliminate any possibility of public -* endangerment. f~| _ , "; > '* # f* ' ** * ^ ^ * -a * j References ' . ' >1 > (1) Enterline, P. E. and Henderaon, V., Arch. Environ. Health, 27, 312 (Nov. 1973). .; . (2) Wagner, J. C.. Ann. Occup. Hyg., IS, 61 (1972). ' , r .. r (3) Wright, G. W., statement before U.S. Dept, of Labor, Occupational Safety and Health hearing on pro posed occupational asbestos stan dard, March 16, 1972. p. 3. Hammond E. C. and Selikoff, I. J., "Relation Of Cigarette Smoking To Risk of Death of Asbestos-Associ ated Disease among Insulation Workers in the U.S.A.," presented at meeting of the Working Group to Assess Biological Effects of As bestos, International Agency for Re search on Cancer. Lyon, France (Oct. 4, 1972). (5) `Report of the Advisory Committee on Asbestos Cancers," Brit. J. In diatr. Med., 30, 180 (1973) . (6) "Asbestos," National Safety Newt (Oct. 9, 1973). v-. . (7) "National Emission Standards for Hazardous Air Pollutants," Federal Register, 38, No. 66. 8820 (April 6, 1973). A - (8) "Protecting the Asbestos Worker," Booklet No. 101D37, The Asbestos Information Association/North America, p 5. (9) Selikoff, I. J., Indnstr. Medicine, 39, No. 4, 21 (April 1970) . (10) "Standard for Exposure to Asbestos Dust," Federal Register, 37, No. 110, 11320 Qnnel, 1972). (11) Bayer, S. G., et al, "Equipment and Procedures for Mounting Millipore Filters and Counting Asbestos Fibres by Phase-Contrast Microscopy," Bureau of Occup. Safety and Health, U.S. Dept, of Health, Edu cation, & Welfare (Feb. 1969). (12) "Asbestos Health Question Per plexes Experts," Chem. Eng. News, 18 (Dec. 10, 1973). (13) Myers, J. L., "Calidria Asbestos Pel lets," ASBESTOS (Oct. 1971). Vol. 47, No. 611, December 1975 *', >' ' ' ' ' ' FMSI 05142 .*- Asbestos Information Association - J North America : 1660 L Street, N. W. Washington, D. C. 20036 Physi.cal Parameters of rAi.rborne Asbestos Fi.bres i.n Various Work Environments--Preliminary Findings G. W. GIBBS and C. Y. HWANG : Department of Epidemiology and Health. AlcUill Universityi. 3775 Uni.ver.sity.Street, Montreal. Province of Quebec. Canada ' The results of a pilot investigation to describe the physical parameters, lencth. aspect . ratio, mass and shape of airhomc fibres in a variety of industries producing process ing'and handling chrssotile, amnsite and crncidolitc arc described. Samples of air- ; borne dust were collected on nucleopore membrane filters and examined by scanning . ' electron microscopy. The diameters and lengths of airborne fibres collected during the dumping of ran amnsite at an asbestos products plant were greater than those of fibres collected during the application of amositc insulation. Chrysolite fibres collected in the carding area of an asbestos textile plant also tended to have smaller diameters * than fibres collected in the dryer and bagging areas of an asbestos mill. The measure- . meats of fibre dimensions indicate that the degree of protection afforded a worker by - optical counts using the memhrarte filter technique is likely to depend on variety of asbestos and stage of processing. Preliminary results arc not in conflict with experi- ' mental data suggesting that ashestosis might he related to the mass of airborne dust - and primary malignant mesothclial tumors to exposure to fibres in a specific range . ' _ of fibre diameter and length. ~ " .' . Introduction ' order to explain the uneven distribution of Although all commercial va rieties OF asbestos have been shown this tumour among occupational groups. The ability of a fibre to initiate a disease to produce mesothclomas when inoculated process in the hmg or on the pleura depends intra-plcurally into rats,1 primary malignant on its penetration, deposition, retention, and mesothclial tumours do not appear to occur hiological activity. These depend mainly on with equal frequency among men working the physical characteristics of the fibre. with asbestos (Table 1). There is some indi First, the settling velocity cf a fibre depends cation that the prevalence of radiological on its actual diameter.11' and explains why changes (asbestosis) and lung cancer mortal fibres as long as 200 /an arc sometimes ity also vary with occupation.-'1'8'11 These found in lung at autopsy. Second, straight differences in health experience might be fibres arc more likely to penetrate deeper explained by differences in fibre type and into the respiratory airways than curly fi the quality and quantity of exposure. Epi demiological investigations in several occu bres.14 Third, diameter and length might both be important in the ability of asbestos pational groups indicate that radiological changes and lung cancer are related to level of exposure.1-"' '111 Smoking appears to be an and other fibres to induce mesothclial tu- mours.,:'-,,` . These physical parameters of airborne fi additional factor which complicates the as bres in the work, environment have.not been bestos lung cancer relationship.1-' The rela previously reported. The possibility that the tionship between primary malignant incso- characteristics of asbestos fibres such as di thelial tumours and levels of exposure are ameter and shape might change during proc not clear and it seems probable that some essing prompted a pilot investigation to other factor or factors must be sought in determine the physical parameters of air- rrt ? v': T .... rrc --rs r Vs FMSl 05143 460 June, 1975 , TABLE I Primary Malignant Mesothelial Tumours Reported . . in Studies of Various Occupational Groups. 4 Study Group Total . Deaths No. Mesothelioma Insulation Workers (USA) (mixed exposures) . Selikoff etal (1969)2 380 22 Chrysotile Miners and Millers (Canada) McDonald et at (I97|)3 t. 3.270 . .. 5 Anthophyllite Mining (Finland) Meurman et al (1972)4 ' 248 . 0* Asbestos Textile Industry (UK) (mixed exposures) Newhouse (1969)* 436 20 . Crocidolite Mining (S. Africa) . Webster (1972) -- ' 88 Amosite Mining (S. Africa) Webster (1972) '-- . 2** . . Amosite Factory (USA) Selikoff (1972)7 One suspected c**e MAmosite minint only 105 5 ' ; ' J j . % - i - i i*0 ' i t' 4 f 1 I f ft . i i . * borae asbestos fibres, to which workers in different occupations are probably exposed. Two samples of airborne dust were ex amined from each location. . Methods Conventionally submicron airborne asbes tos fibres have been examined by transmis sion electron microscopy. This has involved considerable handling of the dust samples and die original fibre diameter distributions in the samples may have been altered in the process. In this study, samples were exam ined by scanning electron microscopy, which required much less laboratory manipulation. Samples of airborne asbestos dust were collected at the following locations using nucleopore membrane filters (G.E. 40. .37 mm diameter, pore size 0.4 fim): (1) at the carding machine in an asbestos textile plant which used chrysotile only. (2) at an asbestos products plant during the emptying of bags of amosite and crocidolite'into hoppers "dumping". (3) at a local oil refinery during the ap plication of insulation materials containing mainly amosite. ' (4) in the dtyer and bagging areas of a chrysotile mill. Preparation of Samples for Analysis A circular portion of the nucleopore filter was coated with 150 A layer of gold-palla dium (60-40) and examined on a stereoscan electron microscope (SEM) type 96113 Mark 2A. Scans were performed across the diameter of the filter beginning at the centre of the original filter which was on circum ference of the mount. High resolution elec tron photomicrographs of the fibres in ran dom fields were taken. . Tfie magnifications on die SEM were identical to those used by Timbrell17 for the measurement of fibres using transmission electron microscopy, i.e. 2,500x for. amosite. and 6.000x for crocidolite. We also used 6,- OOOx for chrysotile. Measurements were made from photographs using a magnified scale and for each fibre we measured: (1) true diameter (dt)--(the width of the central portion of a fibre excluding particles attached to the fibre) - ' (2) coil diameter (d.)--(the maximum fi bre diameter including attached particles or widest diameter of coil) 4 1 i n. .1 A if ' ' J < i .. !. i & I.' f 1 T.' '! 41 FMS1 05144 a . j- i ' I '1 j I} C-t rr. 5. .1 * _, ' | 'S i i j ,i . I | -. - : 'j : 1 '. , i . American Industrial Hygiene Association Journal ' f. (3) coil length (lc)--(the length of the fi bre as it appeared on the filter) (4) true length (I,)--(the length of *a fibre after straightening) (5) coil-aspect ratio (k/dc)--(the ratio of coil length to coil diameter) (6) aspect-ratio (J,/dt)--(the true length to true diameter ratio) (7) mass (m)--calculated from the true fibre diameter and stretched fibre length by assuming that all fibres have a circular cross- section. This assumption was not strictly val id for amosite in which it is rectangular- but was a close approximation for crocidolite in which it is eliptical or circular and for chry- sotiie which in cross-section has a swiss roll appearance. . As the way in which amosite fibres would be orientated on the filter was not known. it was probably reasonable to use a circular cross-sectional area for calculation of aver ' age mass. The densities used in the calcula- tion of mass were 3.45 g/cc for amosite. 3.37 g/cc for crocidolite and 2.55 g/cc for chrysotile. Measurements of the parameters of fibres taken from various regions of the filter showed thatthe distribution was reasonably uniform. Results - . -' . The ranges and median diameters, lengths, mass and aspect ratios of airborne asbestos 461 fibres by fibre type and process are sum marized in Tabic II. True Fibre Diameter (dt) The median diameter of amosite fibres during the emptying of bags (dumping) of raw amosite at the asbestos products plant (0.415 ftm) was greater than during applica tion of amosite insulation (0.370 ftm), but the range was almost the same. The mini mum diameter of the fibres was slightly less than reported by Timbrel! et a/18 (0.098 fun). Cumulative frequency distributions showed a tendency for airborne amosite to be finer during the application of insulation. The median true fibre diameter for crocidolite during dumping was 0.248 yu.m (range 0.011 to 1.347 ytm), which was less than for the amosite fibres (Table II), collected during the application of insulation. The presence of fibres with diameters as small as 0.01 ftm. indicated that a small amount of chrysotile may have contaminated the sam ple; amosite and chrysctile were also used at this plant. The modal diameter flfer croci dolite was 0.25 ftm which was greater than the peak reported for the UICC standard reference samples of crocidolite by Tim brel! et al (0.16 yum).18 This difference may be related to the source of th&_fibre and methods of treatment before measurement. For chrysotile, the median diameter was 0.17 ftm in the dryer. 0.16 ftm in the bag- TABLE 11 " Physical Parameters of Asbestos Fibres d, (pm) lt (pm) dr (pm) lr (pm) Mass (x l(H- gm) Aspect ratio, 1^ fa Amosite median 0.415 3.90 0.510 3.60 2.200 8.00 Dumping range 0.072-2.922 0.57-38.04 0.072-3.00 0.57-38.04 0.011-251.278 3.06-50.16 Amosite Application Crocidolite median 0.370 range 0.071-2.482 median 0.248 2.50 0.47-30.94 2.50 0.428 0.075-3.36 0.292 . 2.93 0.47-33.39 2.49 1.000 0.006-126.033 0.538 7.20 3,10-50.81 9.65 Dumping range 0.011-1.347 0.17-16.18 0.051-1.347 0.17-16.18 0.0001- 38.714 3.22-56.18 Chrysotile median 0.173 1.25 0.317 1.88 0.126 6.30 Drying range 0.011-1.446 0.22-14.88 0.052 2.265 0.22-16.12 0.0012- 25.866 3.07-64.41 Chrysotile median 0.158 1.35 0.264 1.24 0.128 7.25 Bagging range 0.013-1.135 0.34" 9.81 0.013-1.551 0.27-15.01 0.0017- 23.011 3.07-36.76 Chrysotile median 0.150 1.00 0.233 0.93 0.068 6.60 Carding range 0.011 1.251 0.26-12.24 0.011-1.829 0.23-12.16 0.0001- 15.432 3.04-81.95 it i FMSI 05145 SI *4* . - : i i i ,i j i j < 1 1 I ,11 'tv. ii |> tl ila . ..- 462 ging area and 0.15 /un during carding. The percentage of fibres less or equal to 0.5 /un in diameter was 67, 82, and 88 respec tively in the dryer, bagging and carding .areas. - Coil Diameter (dc) There were 3.3 per cent of airborne fibres collected during the dumping of amosite and 0.5 per cent during the application of amo site insulation which were classified as not straight. These deviations from straightness were due mainly to the irregular margins of fibres resulting from attached particles. The coil diameters of amosite were larger than the true diameters (Table 11). Crocidolite fi bres were generally straight and their medi an -coil diameter only very slightly larger than the median true diameter. In contrast, chrysotile fibres had narrower coil diameters at the later stages of processing. True Fibre Length (lt) As there were few "curly" amosite fibres, the median true or stretched lengths were virtuaiy identical to "coil lengths". The true lengths of fibres ranged from 0.57 fim to 38.04 /un (median 3.90 /un) in the manu facturing plant and from 0.47 /un to 30.94 /un (median 2.50 /tm) during the applica tion of insulation. A cumulative frequency distribution showed that insulation workers were exposed to shorter fibres. Crocidolite fibres during dumping ranged in length from 0.17 /un to 16.18 /un with a median length of 2.50 /tm. The median true lengths of chrysotile fi bres collected in the dryer, bagging and card ing areas were 1.25 /tm (range 0.22-14.88), 1.25 /tm (range 0.34-9.81) and 1.0 /tm (range 0.26-12.24) respectively. The distri butions of fibre lengths showed that fibres in the dryer area were longer than elsewhere. The percentages of fibres with lengths great er than 5 /tm in the dryer samples were 11.1 in the bagging area 4.8, and in the carding areas, 2.3. . June, 1975 Coil lengths also decreased from dryer to carding operations. Aspect Ratios (U/dt) Aspect ratios for amosite ranged from 3.1 to 50.2; fibres were defined as having a length to diameter ratio of 3:1. The median 'ratio for airborne amosite collected in the manufacturing plant was 8.0 and for that collected during the application of insula tion was 7.2. The median aspect ratio for crocidolite was 9.7, which was slightly great er than for amosite, although ranges were similar. Although there were differences in the ranges and median aspect ratios (lt/dt) of fibres collected in the dryer, bagging and carding areas (Table II), the median coil as pect ratios in the dryer, bagging area and textile plant were similar, 4.4, 5.0, -4.4 re spectively. Eight per cent of fibres collected at the dryer, 9 per cent of fibres collected during bagging and 12 per cent of all chrysotile fi bres collected during carding were consid ered to be curly (i--. not straight). Mass . The range and median mass of the variour fibre types by process are shown in Table II. Length-diameter Matrices Recently experimental evidence has sug gested that fibres with diameters less than 0.5 /tm and with lengths greater than 5 /tm to 10 /tm might be responsible for the pro duction of mesothelial tumours.15 Length- diameter matrices for. airborne amosite, cro- cidiolitc and chrysotile fibres arc shown in Figure 1. . The percentages , of fibres with diameters less than 0.5 /tm and of length greater than .5 /tm during the handling of raw amosite and during the application of insulation were 17.0 and 18.3 respectively. The results for crocidolite were very similar to the results I! :i ! i, t . H i i4 L: )I *t: 1 w-i * -I* " 1 "I A i * \ * s r-i m I ^4 FMSI 05146 3 1:. . ...CLiiii--'-L'i'J..; American Induilriat Hygiene Association Journal t .1-- Oljncten ^ 0.5 ---------- Slaactcn > 0.50 V" Amite 40^1 0.K <) to.n-o.i -) (0.51-0.75 is) (0.76-1.0 ya) (^1.0 i) Amite *p1ir" InnUtiMl 40 U n-- COuroVclnlStfollt* 40 thrywtllc Orjwr Owysotll* O.1J l~h-- * _ I: o. 40 rn^ n~H-v- ITK ik CKrysotUc Ctrdlng 40 0 5.0 10.0 i.5 7.5 . \k 0 5.0 10.0 2.5 7.5 F=f=V- - r r1 t 11 i- * |" r i r t- - 0 s.o 10.0 0 5.0 10.0 0 5.0 10.0 2.5 7.5 5.5 7.5 2.5 7.5 lmgth ()*} Figure 1.'Length and diameter distributions of fibres collected during various as bestos processes. '- - 463 for amosite with 17.5% of fibres in this size range. The percentages of chrysotile fibres in the dryer, bagging, and carding areas with coil diameters less than 0.5 fim and coil lengths greater than 5 fim, were 1.45, 1.0 and 0.4 respectively. The proportion of fibres with these dimensions decreased from the dryer through to the carding operation, but the number of fibres examined was too small for differences to be reliably assessed. When true lengths and true diameters of chrysotile were considered the percentage of fibres greater than 5 p.m in length and less than 0.5 fitn in diameter measured on two samples of dust from each area was 4.7 and 8.0 per cent at the dryer and 2.8 and 3.4 per cent in the bagging area and 1.1 and 2.0 per cent in the carding area. This trend was similar to.that shown by-thc coil lengths and coil diameters. ' \ The total fibre concentrations in the vari ous areas studied were quite varied, being generally highest in the carding area. Submicron Diameter Fibres If we assume that the optical count of fi bres by membrane filter technique is limited to fibres more than 0.5 fim diameter and more than 5 fim length, the" percentage of the total fibre count that would be observed by optical methods is shown in Table III. Considerably more amosite fibres would be observed optically than chrysotile or crocid- olitc. . Discussion * This .pilot investigation showed that dif ferences in the physical parameters of air borne fibres at various stages of processing could be detected. Although the number of samples examined was small and the results i 'twi : * 1 \ -fits: FMSt 05147 ..'J u t 464 . June, 1975 t TABLEffl Percentage of Fibres Greater than 0.5 Diameter and ................ .f.--------- 5/tin Length .' Percentage of total fibres of diameters >0.5 jun diameter Operation and length >5 ftm . Dumping amosite Applying amosite insulation Dumping crocidolile Drying chrysotile Bagging chrysotile asbestos Carding chrysotile , 27% * 13% 4% 10% '. ' - " 4% ' ' ... *" . .... 5' '. in a textile plant 2.1% ' require confirmation, certain observations were worthy of note . .. (1) The number of fibres visible by op tical microscopy depends on the fibre type. In the control of asbestos exposure and in the investigation of dose-response relation ships these differences could be important. Counts would tend to be underestimated when ftbre diameters were fine, and, hence, more respirable. Existing environmental standards take no account of this. (2) Amosite fibres tended to decrease in diameter from the raw amosite dumping stage to the application of insulation stage while the lengths of fibres were distinctly less in airborne dust collected during the application of amosite insulation. Chrysotilc fibres also tended to be smaller in diameter and length at the later stages of processing. As the settling velocity of a fibre depends on its fibre diameter, the median diameters and the overall diameter distributions sug gested that on diameter criteria only (i.e. excluding factors which, might influence in terception and diffusion), chrysotile encoun tered during carding was potentially the more respirable of the fibres examined. (3) The results suggest that for the same airborne mass concentration the total num ber of fibres to which persons working in a chrysotile textile plant on a carding machine would be exposed would be about 30 times that of a man dumping amosite. Experi mentally, chrysotile. fibres do not readily penetrate the lung because they are inter cepted. However, the chrysotile fibres in these airborne dust samples had coil diam eters less than 3 fim (the limit of respirability). It seems likely, that once airborne, chry sotile fibres such as those encountered in this study are capable of penetrating deep into the respiratory system. The lower prevalence of radiological change among chrysolite asbestos workers than among insulation workers would sug gest, that asbestosis is related to mass rather than fibre concentrations. Experimental evi dence supports this for amphibole fibres.59 For the same concentration, amosite workers would potentially inhale up to 31 times more asbestos in mass than chrysotile workers. Other factors such as solubility, transloca tion dnd fibre length are also likely to play a role, once the fibre, has entered the lung or penetrated to the pleura. (4) Consideration of fibre diameter alone indicates that chrysotile encountered during carding is potentially more respirable than chrysotile or other fibres at different stages of processing. The rarity of mesothelial tu mours among chrysotile miners and millers, when compared to insulation workers (Table I), is probably related to some additional factor such as fibre type or the proportion of long fibre in the airborne dust. The. length-diameter matrices for the vari- . , i Is .c i * t i ! 1 '{ \ i -i i --i err fci f ' te-ye.i < .`Hw; FMSI 05148 i American Industrial Hygiene Association Joumai ous types of asbestos provided some infor mation concerning this. When the fibre types and areas where samples were taken were ranked according to the percentages of fi bres with diameter less than 0.5 pm and length greater than 5 pm present in airborne dust, the following was the result. Percentage of fibres with diameters jo.5 iim and length ^5 Min. Amosite application of insulation Crocidolite Amosite Dumping Chiysotile Dryer . Bagging Carding 18.3% 17.5% 17.0% 1.45% 1.0% 0.4% If. the combination of length and diameter (<0.5 pm diameter >5 pm length) was the only factor responsible for the production of mesothclial tumours, the results presented here would suggest that the hazard for the manufacturer and insulation worker using amosite would be similar, but considerably greater than for chiysotile workers. Taking penetration into account, there might only be a slightly increased risk for insulation workers over factory workers using amosite. This is supported by epidemiological studies of insulation workers and workers in an amosite plant.2'7 Conclusion In conclusion, we recognize that the in ferences made in this report are based on limited data and may not be supported by more extensive measurement. Nevertheless, the findings are sufficient to show the com plexity of interpreting variations in the phys ical characteristics of fibres. The measure ment of physical parameters of fibres in various industries is feasible and observations to test a physical parameter hypothesis arc not in conflict with experimental or epidem iological evidence. Our results so far support the experimental work on animals suggest ing that asbestosis is related to the mass of airborne dust inhaled and that primary ma lignant mesothclial tumours are related to 465 exposure to fibres in a specific range of fibre diameters and lengths. Acknowledgements We wish to thank Mr. G. Scibcl of the Pulp and Paper Research Institute, who per formed the electron microscopy and Miss P. Herapey for valuable technical assistance. This research was supported by a grant from the Institute of Occupational and Environ mental Health. ' References 1. Wagner, J. C., G. Berry, V. Timbrell:' Meso theliomas in Rats. In Pneumoconiosis, Proceed ings of the'International Conference Johannes burg, 1969 (H. A. Shapiro, Ed.}, p. 216, Ox ford University Press, Capetown (1970). 2. Selikoff, 1.1., E. C. Hammond, J. Churg: Mor tality Experiences of Asbestos Insulation .Work ers 1943-1968. In Pneumoconiosis, Proceedings of the International Conference Johannesburg, 1969 (H. A. Shapiro, Ed.), p. 180, Oxford Uni versity Press, Capetown (1970). 3. McDonald, J. C-, A. D. McDonald, G. W. Gibbs, J. Siemiatycki, C. E. Rossiten Mor tality in the Chrysolite Asbestos Mines and Mills of Quebec. Arch. Env. Health 22:677 (1971). . 4. Meurman, L. O., X. Kivilyoto, M. Hakam*. Mortality and Morbidity of Employees of An- thopbyllite Asbestos Mines in Finland. In . Biological Effects of Asbestos, (P. Bogovski, . J. C. Gilson, V. Timbrell, J. C. Wagner, Ed.), p. 199, International Agency for Research on . Cancer, Switzerland (1973). 5. Newhouse, M. L.; The Mortality of Asbestos Factory Workers. In Pneumoconiosis, Proceed ings of the International Conference Johannes- burg, 1969 (H. A. Shapiro, Ed.), p. 158, Ox ford University Press, Capetown (1970). 6. Webstel, I.: Malignancy in Relation to Crocid- oiite and Amosite. In Biological Effects of Asbestos, (P. Bogovski. J.' C. Gilson. V. Tim brell. J. C. Wagner, Ed.), p. 195, International Agency for Research on Cancer, Switzerland (1973). 7. Selikoff, I. J., C. E. Hammond, J. Churg: Car cinogenicity of Amosite Asbestos. Arch. En viron. Health 25:183 (1972). 8. Selikoff, I. J., J. Churg, E. C. Hammond: The Occurrence of Asbestosis among Insulation Workers in the United States. Ann. N.Y. Acad. Sci. 132:139 (1965). 9. Rossiter. C. E., L. J. Bristol, P. H. Cartier. . J. C. Gilson, T. R. Grainger, J. C. McDonald: Radiographic Changes in Chrysolite Asbestos Mine and Mill Workers of Quebec. Arch. En viron. Health 24:388 (1972). I Jt rt iri : f FMSI 05149 466 10. Harries. P. C.. E. A. I . Mit&cnzic. G. Sheers. I. H. Kemp. J. `I. Muigun, T. P. Oliver: A Radiological Survey of Men hxposed to Asireshrs in Naval Dock yards. /IIlit. J. Iinlnsl. Mill. 2V.274 119721 11. Kmcrline, P. E.. H. Weill: Ashcslosis in As bestos Cement Workers. In Biological Effects of Asbestos. IP. Bogovski. J. C. Gilson. V. Timhrell. J. C. Wagner. Eil.). p. 179. Inlcr- . national Agency for Research on Cancer. Switzerland (1973). 12. Newhouse, M. 1..: Cancer among Workers in the Asbestos Textile Industry. In Biological Effects of Asbestos. (P. Bogovski. J. C. Gilson. V. Timbrel). J. C. Wagner. Ed.), p. 203. In ternational Agency for Research on Cancer. Switzerland (1973). - 13. Timhrell. V.: The Inhalation of Fibrous Dusts. Ann. N.r. A nut. Set. 132:255 11965). June. 1975 14. Timbrell. V.. J. W. Skidmore: The Effect of Shape on Particle Penetration and Retention in Animal I.tings. In Inluilctl Particles til (W. H. Walton. Ed.). Unwin Bros. I .Id. Surrey. England 1:49 (1971). 15. Wagner. J. C.. G. Berry., and V. Timbrell: Mesotheliomata in Rats after Inoculation with Asbestos and Other Materials. Hr. J. Cancer 2S.I73 (1973). ' 16. Stanton. M. F. and C. Wrench: Mechanisms of Mesothelioma Induction with Asbestos and l-ibrous Glass. J. Nall. Cancer Inst. 4S:797 (1972). 17. Timbrell. V.: personal communication IR. Timbrell. V.. F. P-ooley. and j. C. Wagner: Characteristics of Respirable Asbestos Fibres. In Pneumoconiosis. Proceetl.ngs of the Intcr. national Conference Johannesburg. 1969 IH. A. Shapiro. Ed.), p. 120. Oxford University Press. Capetown (1970). 19. Skidmore, J. W.: personal communication Seventeenth Graduate Summer Session of Statistics . in the Health Sciences - The seventeenth annual Graduate Summer Session of Statistics in the Health Sciences will be held at Vanderbilt University. Nashville. Tennessee. June 22 to August 1. 1975. ` Tentative course offerings include elementary and intermediate bio statistics. actuarial statistics, demography, sampling methods, research de sign. design of experiments. Bayesian inference, categorical data, and health facility statistics. Several of these subjects will be intensive three-week courses. Instructors for the session will include Helen Abbey, Chin Long Chiang, Jerome Cornfield. Wanzer Dranc. Margaret Drolette. and Norman Johnson. - These courses, at a wide range of academic and experience levels, arc designed to benefit statisticians, epidemiologists, and health science person nel. as well as administrators, health planners, and other health workers who' utilize quantitative data in decision making and problem solving. Graduate students, medical students, and teachers in statistics and in the health, medi cal, and biological sciences will also be interested in the program. The Summer Session is supported by funds from the U.S. Public Health Service. For further information, write to Dr. Charles F. Federspiel. Summer Session of Statistics. Division of Bi<statistics. Department of Preventive Medicine. Vanderbilt University. Nashville, Tennessee 37232. . !- m mw " *' Tif- . > *> FMSI 05150 FMSI 05151