Tobacco Documents Online

National Bureau of Standards Special Publication 506. Proceedings of the Workshop on Asbestos: Definitions and Measurement Methods held at NBS, Gaithersburg, MD, July 18-20, 1977. (Issued November 1978) FDA PROJECTS AND METHODS J. A. Wenninger, I. M. Asher, and P. McGrath Food and Drug Administration 200 "C" Street, S.W. Washington, D. C. 20204 Abstract An overview of FDA projects related to asbestos detection and quantitation is presented. The results of a recent FDA symposium on the availability of suitable techniques are included. We then review the technical and regulatory issues in the food and cosmetics area with regard to asbestos contamination with emphasis on the analysis of parenteral drugs and cosmetic talc. For the present, SEM using Nuclepore filters as a substrate and EDXA for chemical analysis appears to be a reasonable, cost effective method for routine detection of asbestos in foods, drugs, and biologicals, although quantitation and reduction in the number of ambiguous fibers is still a problem. . Key Words: Asbestos; cosmetic talc; ED%A; fibers; food; parenteral drugs; SEM. Part I: An Overview (I. M. Asher) Since the Food and Drug Administration (FDA) programs related to asbestos are spread throughout several Bureaus, my colleagues have asked me to give an overview of FDA research efforts and interests. As you know, asbestos contamination of air and water is largely the domain of the EPA, while contamination of the workplace is of direct importance to OSHA, and products for home use are the responsibility of CPSC. Thus, FDA interest has centered on the asbestos contamination of food, cosmetic talcs, and parenteral drugs. A major problem is developing rapid, reliable methods for the monitoring of asbestos in such products. An FDA symposium to evaluate the current state of electron microscopic methods for microfiber detection and analysis was held last August at Penn State University, with many of the current participants present. Naturally enough, the speakers tended to point out the promise of their methods and the weaknesses of alternate methods, but the consensus was that a single method, simple and accurate enough for embodiment in FDA standards and regulations, has yet to be perfected. (Copies of the_ Proceedings are available, free of charge; from the FDA Office of Science.) In the interim, the FDA has published a regulation banning the use of asbestos filters and other filters releasing mineral contaminants with aspect ratios of >3:1-in the final stages of manufacture of injectable products, unless followed by a nembrane filter (40 FR 11865). It is hoped that this interim regulation can be replaced by appropriate standards, and Phil McGrath's group in the Bureau of Biologics has been trying to perfect and validate an appropriate interim SRM method. An FDA Asbestos Work Group (chaired by Dr. Armand Casola, Bureau of Drugs) meets regularly to discuss these and other issues. The Bureau of Biologics has its own scanning and transmission electron microscope facilities to detect and identify particulate contaminants in biological products and parenteral drugs. The Bureau of Drugs has also initiated additional studies to identify particulate contaminants in commercial samples of parenteral drugs, under contract at the 441 2063105230

University of Kentucky:- The Bureau of Foods has an ongoing program of analyzing cosmetic talcs for tremolite and anthophyllite contamination (by contract). The most ambitious FDA project in this area is an animal study of the carcinogenic effects of subacute, intraveneous injection of chrysotile asbestos in Charles River CD rats and CD-1 mice of both sexes. The project is being conducted by the Bureau of Drugs, under contract at the International Research and Development Corporation, Mattawan, Michigan. There are three negative control groups: saline/single injection, saline/4 weekly injections, and kaolin/10 weekly injections ("inert" particulates)-for each species/sex group of animals. There are also positive control groups receiving methylnitrosourethane once weekly throughout the lifetime of the animal. There are six different asbestos dosages-single or four weekly injections of 0.2, 0.4, 0.8 mg/kg-for each species/sex group. This gives a total of 10 dosages groups for each species/sex group and amounts to 3480 animals in all. The FDA chronic study utilizes 18 grams of asbestos sample, prepared to mimic typical releases of pharmaceutical-grade asbestos- cellulose filters. These are typically short and very narrow. The mean fiber length in the sample was 2.34 pm (ranging from <1 pm to 70 Nm), and the mean diameter was 0.056 pm (ranging from <0.01 pm to 0.25 pm). So far, at 16 months, too little data is available to report definitive results; however, the incidence of lung tumors at necropsy in the male CD-l mouse group at the highest asbestos dosage currently exceeds that for the saline controls (i.e., 9/39 compared 2/23 and 3/24). The co-project officers of this study are Jules Lamar and Stephen Crop of the FDA 8ureau of Drugs. Part II: Food and Cosmetics - An FDA Update on the Asbestos Question (J. A. Wenninger) My discussion will focus on FDA's activities to prevent the contamination of cosmetics and food by asbestos particles. I have been involved only with the problem of asbestos in cosmetics, but I will extend my discussion to cover food. Cosmetics and food share similar regulatory and physical-science characteristics, but there the similarity ends; to a large degree, the problem'with food centers on the ingested fibers, whereas with cosmetics it centers more on inhalation of such fibers. No regulations for either food or cosmetics have yet been established which either prohibit the use of asbestos-containing filters in food processing or limit the amount of asbestos fiber in talc used as a component of food or cosmetics. A proposal for certain restrictions on food only was published in the Federal Register (38 FR 27076-81), September 28, 1973. However, this regulation has not been published as a final order and is still pending. The conments received in response to this proposal clearly indicated that no regulation for food and food processing was warranted until more reliable data could he obtained on methodology for the determination of asbestos and on a more complete evaluation of the health hazard associated with ingested asbestos fibers. FDA's reply to these comments were published in the Federal Register (40 FR 11865-70), March 14, 1975. It should be emphasized that on the basis of information received the agency did conclude that the asbestos content of talc used in the manufacture of food - or drug - contact paper packaging does not represent a potential contaminant of packaged food or drugs as assessed by current methodology. With regard to cosmetics it is unlikely that we will be in a position in the near future to propose a limitation on the asbestos fiber content of talc used for cosmetic talcum powders. However, we do have a modest surveillance program under which we monitor the asbestos fiber content of retail units of cosmetic talcum powder products. To date we have not found any grossly contaminated cosmetic talcum powder products on the market. Although this is somewhat reassuring, our sampling of products was small; for example, we looked at only 28 samples by x-ray powder diffraction during 1975 and 1976. Of these, one sample was found to contain 0.7 percent tremolite and three samples contained traces of tremolite (approximately 0.1%) and anthophyllite. In our laboratories we are now using three basic methods for the evaluation of asbestos contamination of cosmetic taics. We estimate our limits of detection as follows: 442

Differential X-Ray Optical Thermal Diffraction Microscopy Analysis CHRYSOTILE 2%a --- 0.5%a TREMOLITE 0.1% 0.1% ANTHOPHYLLITE 1% --- a In the absence of interference from chlorite. The Cosmetic, Toiletry and Fragrance Association, Inc. (CTFA) has continued to cooperate with FDA's Division of Cosmetics Technology in developing reliable methodology for the determination of asbestos in cosmetic talc. Results from a testing program set up by the CTFA to establish the reliability of analytical methodology are expected to be available in the near future. The CTFA has been active in establishing appropriate specifications for cosmetic talc and developing analytical .ethodology for industry. An article on cosmetic talc powder which appeared in Lancet (Volume 1, pp. 1348-9, June 25, 1977) concluded: . "there is no reason to believe that normal consumer exposure to cosmetic talc in the past led to either cancer at any site or to measprable loss of lung function. It seems unlikely that future exposure to cosmetic talc qf the specifications now agreed to by major manufacturers will present a health hazard." We do not know if this assessment is correct. However, it is the responsibility of all of us to assure that appropriate steps are taken to prevent the use of talc unsuitable for use in food and cosmetics. It now appears that several years may be required to fully clarify some of the scientific questions on this subject. In the meantime it may be prudent to establish by regulation a standard for all to follow. No doubt this approach will be questioned in the absence of widespread contamination. However, we know that efficient enforcement of any specification is very difficult without the assistance of regulation. Part III: Scanning Electron Microscopy for the Detection of Asbestos in Foods. Drugs, and Biologicals (P. P. McGrath and J. 8. Ewell) For the past two days we have heard of the many problems associated with detection, identification, and quantification of asbestos and asbestiform minerals in the environ- ment. We have experienced many of these same problems in an attempt to design methods which could be used for routine electron microscopic analysis to detect asbestos con- taminants in products regulated by the United States Food and Drug Administration. For routine. analysis of these products we feel that scanning electron microscopy (SEM) using energy dispersive x-ray analysis (EDXA) is the most cost effective method. The rationale for choosing SEM-EDXA over Transmission Electron Microscopy, selected area electron diffraction (TEM-SAED) technique is based on many factors. Since most of the products examined contain very low levels of asbestos, the size limitation imposed by an E.M. grid would interfere with detection of these small numbers of fibers. Most samples are prepared for TEM examination through some type of filtration and the filter must be destroyed by chemical or thermal means to allow examination in the TEM. Filter residue left on the E.M. grid consistently interferes with the analysis and production of diffraction patterns. Many fibers do not produce measurable diffraction patterns or are lost during the preparation of the sample [1,2]1. Even those fibers which do produce diffraction patterns must be indexed to identify the fibers. To index these patterns is time consuming and requires sophisticated methods such as that developed by Lee at U.S. Steel [3,4]. 1Figures in brackets indicate the literature references at the end of this paper. 443 2063105232

For analysis in the SEM, the filter surface itself is examined. If one compares a 47 mm or 13 mm diameter filter with a 3.05 mm diameter electron microscope grid, the difference in the area available for examination is obvious, as for example figure 1. Further, less than 70 percent of the surface area of an E.M. grid is available for TEM examination because of the grid bars which are not penetrated by the electron beam as shown in figure 2. Figure 1. Comparison of 47 mm and 13 mm diameter filters with 3.05 mn diameter E.M. grid. Figure 2. Electron microscope grid illustrating surface area of grid bars. Arguments against using SEM for this type of analysis cite the limits of resolution in the SEM, the lack of diffraction capabilities; or that the chemical profiles developed by energy dispersive x-ray analysis are not definitive criteria for classification of these asbestos minerals [5,63. These arguments in our estimation are not valid. The resolution of the majority of scanning electron microscopes is near or below 10 nanometers and most of the newer models guarantee 3 nanometers resolution. The lack of diffraction capabilities is not a major factor because in the TEM a very large percentage of the asbestos fibers do not yield usable diffraction patterns. X-ray analysis of fibers does produce sufficient chemical information to classify fibrous asbestos minerals [7]. One method depends on preparing a clean sample on a Nuclepore filter, enabling the operator to identify the particle of interest in a reasonable length of time [8]. We use Nuclepore filters in preference to Millipore filters because the surface of the Millipore filter Interferes with the detection of small fibers as shown in figure 3. 444

G 0 ~ Figure 3. Asbestos fibers partially obscured by configuration of Miltipore filter surface. The filters are first examined in the 5EM at low magnification to determine if the preparation is usable and to look for large particulates or product residue which might obscure the small asbestos fibers. If the sample is suitable for examination, representa- tive micrographs are taken of fibers found on the filter surface. Fibers or fiber bundles resembling asbestos are subjected to energy dispersive x-ray analysis for 100 to 400 seconds (machine count time) depending on the spectra developed. , Identification of chrysotile asbestos is based on the morphology of the fi4ers or fiber bundles, the x-ray counts for magnesium and silicon, and the absence 'of any appreciable iron or other elemental peaks. We have not established x-ray criteria for amphibole asbestos, but would only record them as a mineral fiber with the chemical profile recorded. At the present time we are able to routinely identify chrysotile asbestos fibers tess than 70 nanometers in diameter on the filter surface using EXDA, but only after long count times, up to 400 seconds. Larger fiber bundles, 1/2 micrometer and above, produce peaks which can be read on the analyzer display CRT in less than a minute, reducing the count time and enabling the operator to go to the next fiber of Interest. To quantitate the number of fibers, we can estimate their size by comparison with a micron marker on the display CRT or by comparing them to the pores in the Nuclepore filter as shown in figure 4. For more accurate counts, SEM micrographs at 5000X or greater are taken and the fiber measured with a ruler and map reader. This is a slow and time-consuming task. We still have problems with uneven filter surfaces and product residue interfering with the analysis, but this is less a problem than it would be in the TEM because the surface examined is so much greater. 445 Figure 4. Small asbestos fiber traversing pores of a Nuclepore filter.

In the future we._plan to incorporate an automated image analysis system similar to that developed at Penn-State University [9]. We also are attempting to develop our own x-ray data reduction system based partially on the work done by Friedman et al., for analysis of neutron activitation spectra here at NBS for the Bureau of Foods, FDA [10]. We feel that the SEM-EDXA, automated image analysis system will enhance our ability to do routine analysis for asbestos and other particulate contaminants. References [1] Beaman, D. R., NBS Workshop on Asbestos, July 18-20, 1977. [2] Beaman, D. R., Difficulties encountered in the identification of asbestos fibers by analytical transmission electron microscopy in Electron Microscopy of Microfibers - Proceedings of the First FDA Office of Science Summer Symposium, Penn State University. August 23-25, 1976 [3] Fisher, R. M. and Lee, R. J. , NBS Workshop on Asbestos, July 18-20, 1977, [4] Lee, R. J. , Computerized SAED and the electron opticai identification of particulates in Electron Microscopy of Microfibers - Proceedings of the First FDA Office of Science Summer Symposium, Penn State University, August 23-25,1976. [5] Stewart, I., Selected thoughts on asbestos quantitation in Electron Microscopy of Microfibers - Proceedings of the First FDA Office of Science Summer Symposium, Penn State University, August 23-2, 7976. [6] Ruud, C. 0., Characteristics of silicate mineral fibers as seen by TEM-SAED and SEM-EDXA in Electron Microscopy of Microfibers - Proceedings of the First FDA Office of Science Sunmer Symposium, Penn State University, August 23-25,1976. [7] Pooley, F. 0., The identification of asbestos dust with an electron microscope microprobe analyzer, Norelco Reporter, 23, No. 2 (October, 1976). [8] McGrath, P. P. and Ewell, J. B., Application of electron microscopy to the problem of particulate contaminants in food, drugs, and biologicals in Proceedings of Electron Microscopy, IIT Research Institute, Chicago, Illinois, Part III, 1976. [9] Johnson, G. G., Image Analysis Techniques in Electron Microscopy of Microfibers - Proceedings of the First FDA Office of Science Summer Symposium, Penn State University, August 23-25, 1976. - [10] Friedman, M. H. and Tanner, J. T., A computer language for reducing activation analysis data, submitted to the Journal of Radioanalytical Chemistry, May, 1977. Discussion H. RHODE: I have a question for Phil McGrath. We are very much interested in the use of the scanning electron microscope for commercial asbestos samples and the problem you illustrated beautifully there with the Millipore is like looking for a needle in a large haystack, except that the NIOSH procedure requires that Nuclepore not be used for collecting air samples. Thus we are kind of on the horns of a dilemma. Have you done anything in the way of trying to mask the structure of the Millipore7 P. McGRATH: No, but Dave Manolin of Millipore told me that the reverse side of a Millipore filter Is smooth. You might reverse the filter. I have not done it. Or, you could ash the Millipore filter, suspend the ash in water, and run it through a Nuclepore filter. RHODE: We tried controlled exposure to acetone vapor with some promising results, but we are not ready to be sure that we've got it yet. I was hoping that someone else had some ideas about collapsing it rather than dissolving it, but that has some problems too. 446

C5 J. LEINEWEBER: Our company seems to be the favorite stopping place for everybody who has developed gadgetry of one sort or another to help in the counting of asbestos fiber. We have encouraged this because we would like to see what is happening in this field, and among the things we have done is to follow the methods of automated analysis and everytime we get into the laboratory with some asbestos fiber samples the fiber sizes are too fine, the samples are too complex; we just have not gotten off the ground in that direction. It's an interesting concept. McGRATH: It is an interesting concept. I have talked to Jerry Johnson of Penn State about the use of image analysis. They probably have the hardware but need to develop programs for fiber analysis. I believe the group at Penn State would be willing to develop the programs if someone would supply the monies. LEINEWEBER: Another comment on the Millipore texture problem and carbon coating Millipore filters for TEM work-many times you carry this texture along and it kind of interferes in the TEN work; we have found that collapsing with acetone vapor prior to coating with carbon does give a much smoother surface and a lot less interference. McGRATH: Sometimes we sputter coat the Nuclepore filters with gold-palladium before we use them. This reduces the pore size slightly but also reduces charging. A. LANGER: The only drama associated with any presentation on asbestos was the drama of Irv Selikoff and myself in 1968 at the Food and Drug Administration, presenting a seminar on the "Contributions of Fibers from Talc to Human Lung Burden." We have given FDA nine years, and I am delighted to hear that you are taking some action. t J. WENNINGER: Let me set the matter straight for the public record of this meeting. The FDA has not taken any action in regard to the possible contamination of cooetic talc by asbestiform minerals. LANGER: I won't be quite as dramatic, but this study in Lancet, is that the study of the Italian deposits? WENNINGER: I don't think it was. It was a general review article summarizing a meeting held in England sometime ago. LANGER: I think this is based on the study of the Italian talcs, the pure talcs. That's the five nought variety that has very, very little mineral contamination, and I think that your assurances are directed to the users in the United Kingdom and not the users of consumer talcums here in the States. WENNINGER: That could be correct, however that was not my understanding. NOTE: The following was a note sent following the meeting and was not part of the verbal discussion at the end of this paper. G. LEE: In his presentation of the FDA regulatory status with respect to cosmetic talc, Mr. Wenninger quoted from an editorial "Cosmetic Talc Powder" which appeared in the June 25, 1977 issue of The Lancet. During the ensuing discussion period, Dr. A. H. Langer speculated that conclusions of the safety of cosmetic talc may have been drawn solely from data restricted to the Val Chisone Italian talc and would therefore bear no relevance to American talc products. To answer this apprehension and to set the records straight, I am including a copy of this precise editorial, which explicitly references data, human and animal, covering cosmetic talcs which are used both in the United Kingdom and the United States. This editorial conclusion clearly applies to American cosmetic talcs as well. 447 2063105236