Mineral Characterization

Minerals have fundamental properties that can be defined in terms of physical (e.g., growth habit, hardness, cleavage), chemical (e.g., chemical composition, surface reactivity, solubility), optical (e.g., translucence, refractive indices), and electrical (e.g., conductivity, resistivity) characteristics, among others. A degree of variability in the compositional and structural makeup of specific minerals in nature also exists, reflecting differences in petrogenesis and mineral source (natural conditions of formation). Therefore, minerals exhibit a range of physical and chemical properties that result in varied responses to conditions imposed during extraction, processing, and experimentation. Research on the degree of variability between mineral species and within mineral groups, due to their natural conditions of formation, and the extent to which these varying characteristics influence toxicity is largely absent from the Roadmap and more detail is needed.

Key variables of relevance in studying minerals include surface crystal structure and chemistry, size and shape characteristics, and mineral habit—all of which will vary depending on the composition, environment of growth, and response to physical and chemical processing of mineral samples. Essential to the consistent rigorous characterization of minerals is the consideration of mineralogical properties in a similar way throughout the mineral “life cycle”—from extraction or liberation at the source to processing for use in manufactured products or as research materials. Modern methods of mineral characterization can also provide a statistical representation of the range of minerals and their physical characteristics. This form of minerals characterization is used routinely by the mineral industry to identify and quantify mineral samples in terms of their variability in composition, size, and shape.

Whether establishing a new reference mineral sample for eventual use in health-related research or characterizing an unknown suite of elongate mineral particles from an air sample filter, a basic set of mineral characterization techniques can be employed in a tiered fashion. Additional characterization techniques can then be employed if required. The level of analytical detail will depend upon the goal of the research. Table 3-1 outlines a basic characterization sequence and suggests potential outcomes of the research that can be used to plan further research in a systematic way. Such an approach to mineral characterization is currently absent from the Roadmap.

TABLE 3-1

Tiered Approach to Mineral Characterization.

The committee believes that a fundamental problem with the proposed research in the Roadmap is the reliance on limited and outdated analytical methods such as phase contrast microscopy (PCM). Other methods such as transmission electron microscopy (TEM) or scanning electron microscopy (SEM) are not recommended for use exclusively or as “stand-alone” analytical methods. Rather, TEM or SEM can be used most effectively in conjunction, if possible, with the petrographic techniques listed in Table 3-1. The need to develop new methods based on electron microbeam analysis techniques is critical and should not be limited by existing regulatory constraints or existing policy. The committee strongly believes that the science should drive the policy and regulation, not the reverse.

Prediction and prevention are linked to having well-characterized sample sets for experimental work and analysis. The Roadmap recognizes that workers may be exposed to any number of crushed and ground and/or contaminant particles introduced during mining, milling, manufacturing, and demolition of the materials. In these cases, the size and shape criteria used to describe elongate mineral particles encompass many mineral groups in addition to asbestos and analogous minerals. Mineral source and petrogenetic studies can be used to help characterize mineralogical materials in terms of source (original geological source for the mineral and formation conditions). By using statistically reasonable sample populations from diverse natural sources and rock localities (e.g., mines), including historical data sets, minerals from an air sample filter may be classified not only by mineral composition, optical properties, size, and shape, but also by petrogenetic history. This type of approach also has the potential to aid in the design of tailored toxicological studies linked specifically to a particular mineral source and could provide predictive assessment for materials derived from similar geological settings.

Research is also needed into the potentially toxicological responses of exposure to mixtures of elongate mineral particles. The complexities of this research underscore the need to have well-characterized mineral samples in the context of their natural sources and petrogenesis in order to understand the controlling variables in the toxicological experiments.

Standardized Reference Minerals

Well-characterized reference mineral samples are important for research on the potential health effects of elongate mineral particles, and the need for a well-managed repository should be emphasized. The identification, classification, and characterization of unknown mineral particles from workplace or environmental exposures require comparison to rock-forming minerals that have been characterized mineralogically by conventional petrographic techniques. Similarly, designing and conducting meaningful toxicological experiments require well-characterized reference mineral samples to allow systematic intra- and interlaboratory comparisons of results. The Roadmap notes the need for standardized reference mineral samples but should include more details on an approach to developing a central repository for systematically characterizing and standardizing the samples. This type of repository provides better precursor material for research of the sort proposed in the Roadmap.

Ideally, all minerals studied by laboratory inhalation exposures should either be obtained from the repository or be matched with smaller samples that are well characterized and included in the repository. It should be noted that substantial quantities of minerals would be necessary if the repository intends to support long-term whole-body inhalation studies. (Smaller quantities are needed for nose-only inhalation studies.) For example, tens of kilograms of respirable minerals would be required to conduct a multi-dose inhalation study of sufficient magnitude to test carcinogenic potential. This does not obviate the need for a repository of smaller samples of standardized minerals. A few grams could support comparative in vitro tests that would help place the effects of inhaled minerals into context, even if the study sample did not come from the repository.

Several sets of standardized reference minerals have been developed in laboratories of the National Institute of Environmental Health Sciences (NIEHS) and the former U.S. Bureau of Mines (USBM), as well as other groups, and could be included in a central repository. During the mid-1970s, approximately 500 pounds of respirable crocidolite (from South Africa), chrysotile (from Canada and California), grunerite (from South Africa), and tremolite (from a New York State tremolitic talc deposit) were prepared and mineralogically characterized in conjunction with asbestos-related research conducted by the USBM (Campbell et al., 1980). The existing mineral samples vary in the extent to which they have been characterized, and all existing samples and new mineral samples would eventually have to be examined and represented by the same types of basic analytical data, generated using viable modern techniques (see Table 3-1). Most respirable particle size ranges can be classified or developed from these types of standard minerals. Modest initial studies could, for example, be undertaken using the USBM’s well-characterized asbestos and tremolitic talc samples, as available. The samples originally characterized by the USBM could serve as precursor material. These studies could include petrographic thin sections made from drill cores and/or initial primary crushing samples from the mine, mine run, or crushing runs. Petrographic analysis of thin sections and particle analysis of crushed materials could be augmented by X-ray diffraction, TEM, and SEM. These petrographic analyses could make use of the <2 μm limit of the optics of the petrographic microscope. This type of work could also characterize the minerals needed for future research, including biological testing. Concurrent studies for discriminating asbestos from cleavage fragments and other particle types using samples from personal monitors worn in mines could also be conducted. Once samples are well characterized by these initial studies, the results could be applied to non-mine-related OSHA and Environmental Protection Agency (EPA) monitor samples. This is an example of a tiered approach to the NIOSH research plan (Table 3-1).

Important descriptive and quantitative parameters needed in a unified characterization of standards include the following:

• Mineralogy of all phases of the sample including quantification of all accessory minerals
• Chemical compositional variability of the mineral of interest
• Size, shape, and degree of mineral fabric and texture, including the habit and single-crystal development
• Source (including, if available, the geological and petrogenetic history of the deposit from which the samples are derived and the mineralogy of the source material)
• Industrial or manufacturing history
• Characterization of the aerodynamic properties of the minerals, and the composition of the other materials for toxicological studies
• Surface chemistry, surface structure, and charge

One of the fundamental parameters for reference and testing materials is particle dimension. However, the Roadmap does not deal with the issue of particle dimension, except obliquely as may be inherent to the different source materials. The importance of particle dimension to health cannot be conclusively determined on the basis of comparing results of studies that contrast two different mineral assemblages, each of which varies in particle size and surface area. A few studies have investigated the degree to which particle dimension in samples of the same mineral is a determinant of toxicity (Davis et al., 1991; Berman et al., 1995; Bernstein and Hoskins, 2006; Bernstein et al., 2006). To expand on these studies, researchers will need samples of the same mineral type (whether from the same source or from well-characterized standards) differentiated by size, shape, dimension, and growth habit. A national mineral repository could meet this need. While not trivial to establish, particularly with the need for processing, characterization, storage, and curatorial facilities, such a repository could include suites of minerals with varying size fractions and growth habits as well as suites of minerals with varying chemistry for use as benchmark samples. Without such a repository, much of the proposed Roadmap research will have diminished value. Given likely developments in instrumentation and classification, the standards should be assessed to allow for future enhancements in the data set.

State of Science and Future Directions

As discussed in Chapter 2, the Roadmap needs to explicitly outline a standardized approach to screening for the potential hazards of elongate mineral particles whose biological effects have not yet been characterized. A systematic, tiered structure of increasingly complex in vitro and in vivo assays would be useful not only for screening for toxicity, but also for placing the nature and extent of the hazard of the new material in the context of what is known about the hazards of previously tested mineral particles. The screening process would begin with the assessment of the physical and chemical nature of the mineral particles that would pro vide information to begin to assess where the different types of elongate mineral particles fall along the spectrum of materials for which some information on biological hazard exists. In this and subsequent steps, it is important to ensure that the mineral particles being evaluated represent the physical and chemical characteristics of the respirable material to which humans are, or may be, exposed. If exposures involve mixtures of minerals, the tested samples should represent those mixtures. Separation of mixed exposures into different types would only be appropriate at later stages to assess the components that may impact toxicity.

Careful consideration should be given to selecting the screening tests. The assumption must be that the tests reflect types of responses and perturbations of biological response pathways that are most likely to be caused by elongate mineral particles. It is possible that important responses may not be probed by the selected tests. As knowledge of response mechanisms and critical pathways improves, the selection of tests may evolve (NRC, 2007).

After a series of intermediate steps, the screening may, in certain cases, have to be carried through to long-term inhalation studies of animals. However, an underlying goal would be to establish a screening process and knowledge base where extension to chronic animal studies would rarely be needed to assess the extent to which human exposures should be limited (that is, to establish a regulatory categorization). The following steps (tiers) comprise the proposed framework. This framework is similar to the tiered testing system for natural and synthetic fibers proposed by the International Life Sciences Institute (ILSI, 2005). The intent of the committee’s framework is for research needs to flow from present limitations in addressing each of the steps (Figure 3-1).

FIGURE 3-1

Overview of a possible testing strategy for elongate mineral particles. NOTE: If at any point, human data are available, those data should have priority in assessing toxicity and the potential for health impacts.

Physical and Chemical Characteristics

First the Roadmap should identify the types of studies needed to better understand the biologically relevant physical and chemical characteristics of elongate mineral particles. As noted above, a thorough mineralogical characterization of the mineral particles, particularly their chemical structure, physical dimensions, and biopersistence, is needed for meaningful comparisons of toxicological responses. Regard must be given to the several technical issues related to sampling and characterization, noted in subsequent sections, to ensure that the mineral samples studied accurately reflect the minerals to which humans are exposed. By relating biological responses to standardized mineralogical nomenclature, a knowledge base can be developed that progressively reduces the range of mineralogical particle types that need to be tested de novo in order to classify their hazard potential. The following stepwise approach could be undertaken to examine and categorize the mineral particles. It is possible that having addressed these steps, further investigation of some mineral particles may not be needed.

1.

What are the physical dimensions of the mineral particles to which people are exposed? What is the aerodynamic size distribution of the airborne mineral particles? What is the distribution of the dimensions of the bulk particles?

If the particle is respirable or of a thoracic size fraction, it should be considered potentially hazardous and war rants further evaluation. Concern is heightened if the in haled particles include dimensions that are consistent with other known harmful elongate mineral particles.

2.

What are the mineralogical characteristics of the respirable and thoracic size fractions of particles to which people are exposed? Is this homogeneous or a mixture of particles? Is the particle similar to those for which health hazards are already known?

If the particle is analogous to previously studied parti cles, its toxicity should be considered comparable until proven otherwise. The confidence in the assessment should be proportional to the degree of similarity be tween the particles.

3.

What are the potential exposure concentrations? Specifically, what are the exposure concentrations of the thoracic fraction of the particles? What is the duration of the exposure?

If exposures to the particles meet or exceed accepted ex posure limits for the general class of particle, the expo sure scenario should be considered problematic. If the exposure is below accepted limits for a new elongate mineral particle, some preliminary toxicity testing may still be needed to evaluate the need for a more extensive effort if the particle has characteristics that suggest the potential for toxicity. The magnitude of exposure must be considered in making this decision.

In Vitro Solubility

Elongate mineral particles must reside in the lung for extended times to produce disease beyond the transient inflammation that might be associated with the deposition of any poorly soluble material. Because in vivo solubility is one of the characteristics that determine the persistence of elongate mineral particles, the next step is to assess the solubility of the particle in biological media using state-of-the-art methodologies. A current approach for assessing in vitro solubility uses a lung fluid stimulant, such as Gamble’s solution (de Meringo et al., 1994; Zoitos et al., 1997), to provide a rapid and inexpensive approximation of likely in vivo solubility. Existing data on the relative solubility of several types of elongate mineral particles in this assay (Hesterberg and Hart, 2001; Hesterberg et al., 2002) helps place the results in context; this is particularly valuable when toxicity data may also be available for these other elongate mineral particles. Although solubility assays are reproducible using the same type of elongate mineral particles, the assay should include concurrent control particles with known solubility. Extreme acid, base, or enzymatic solutions should be avoided at this step; solubility in an already-accepted lung fluid stimulant has proven to adequately mirror in vivo biopersistence. It is recommended that solubility studies be conducted at two pHs—one slightly acidic, reflecting the pH inside macrophage phagolysosomes, and the other neutral, reflecting the pH on the surface of the alveoli (Christensen et al., 1994; Sebastian et al., 2002; Maxim et al., 2006).

Elongate mineral particles have a range of solubility and biopersistence, and there is not a well-established cut-off point below which a mineral particle would not be of concern. Developing a quantitative database of how well-characterized minerals respond to dissolution experiments in human and animal tissues and fluids will provide valuable information in deciding what mineral particles should undergo further toxicological testing. Mineral surface structure, chemistry, and electrostatic charge behavior are additional important parameters and require accurate characterization of the test particles used. At present such a description of this type of research is missing from the Roadmap.

Decision point: If the solubility of the elongate mineral particle falls within the range of similar particles known to be biopersistent and to cause disease, it should be considered potentially pathogenic.

In Vitro Toxicity Assays

As noted in the Roadmap, numerous in vitro methods have been developed for exploring the mechanisms of toxicity produced by elongate mineral particles. The appeal of these cellular and tissue assays is that they are more rapid and less costly than in vivo assays, and apart from primary embryo fibroblasts, most do not require live animals. The in vitro studies provide useful information on dose-response toxicity of the mineral particles. The studies also can be used to assess the effects of dimension, chemical composition, and interaction of particles with other environmental toxicants such as chemical carcinogens, viruses, or radiation (radon) in modulating the biological behaviors of the elongate mineral particles. Taking advantage of advances in toxicogenomics, bioinformatics, and computational toxicology as applicable, the goal should be improvement in the utility of in vitro assays for classifying the potential hazard of newly encountered particles and ultimately predicting in vivo health outcomes of exposure without the need for in vivo studies. However, considerable advancement will be necessary before in vivo studies can be discarded; thus, the primary purpose of in vitro tests in the near term will be to screen materials for further investigation. In vitro toxicity studies using single cell types do not reflect complex interactions among tissues and organs, do not include the modifying effects of deposition patterns and clearance by mucociliary actions and macrophages, and cannot be used to ascertain the contribution from biopersistence. In vitro assays have the potential for indicating false positives for causing disease, particularly at the very high concentrations used in some of these studies. For example, vitreous fibers and wools, although positive in many short-term toxicity assays, including cell killing, apoptosis, and generation of reactive radical species (for example, reactive oxygen and reactive nitrogen species), have low potential for causing disease in vivo. At this time, none of the in vitro assays have been validated as predictors of disease in animals or humans to provide the confidence for their use as the sole determinant to predict health hazards of new elongate mineral particles.

In vitro assays have provided rankings of the relative inflammatory and carcinogenic potential of varieties of asbestos having known in vitro toxicity and have also provided insights into the cellular mechanisms of response. Further research and validation may enable in vitro assays to fill a similar role for certain classes of elongate mineral particles. A clearly positive response in vitro would signal the need for further testing and might suggest the longer-term health outcomes to be investigated. A very low or negative response on multiple in vitro assays might be useful for assigning the particle to a low-risk category. Because asbestosis, lung cancer, and mesothelioma likely have different mechanisms of pathogenesis, attention must be given to selecting an array of in vitro assays capable of detecting cellular events thought to be involved in pathways leading to each outcome of concern.

The goal would be to establish an efficient testing approach that uses a small number of assays to estimate in vivo toxicity based on in vitro responses. More than one type of assay would be necessary to explore the range of potential toxicity responses (e.g., genotoxicity, inflammogenicity). The thoracic or respirable fraction of the elongate mineral particles would be tested using a range of doses reflecting those that may occur during human exposure, depending on the choice of the in vitro target cell, in parallel with control particles having extremes of responses in each assay.

Decision points: A tentative assignment of hazard level would result from ranking in vitro toxicity relative to those of elongate mineral particles having known in vivo toxicity. A benign ranking in combination with evidence from existing tiers of low potential toxicity might result in no further testing. Otherwise, the nature of in vitro toxicity would help focus subsequent in vivo assessment on the most likely health outcomes.

In terms of the state of scientific understanding of the toxicity of elongate mineral particles, several opportunities exist to improve the Roadmap. Most carcinogenic asbestos and glass fibers can morphologically transform rodent fibroblast cell lines (Hesterberg and Barrett, 1984; Mikalsen et al., 1988) but not primary human epithelial cells (see Hei et al., 2000, for review). This has been attributed to the extremely low transformation frequency observed in primary cells. As a result, virally immortalized human epithelial cells have been used successfully in morphologic transformation studies with certain types of asbestos (Hei et al., 1997; Wang et al., 2004; Bertino et al., 2007). The results are consistent with the observation that SV40 may act as a co-carcinogen with asbestos in the pathogenesis of mesothelioma (Bocchetta et al., 2000). However, because of the cost and long lag time required for the transformation of human epithelial cells in culture, no dose-response data are available and most of the studies lack adequate asbestos controls. Nevertheless, these studies provide a platform for mechanistic analyses of the neoplastic transforming process (Zhao et al., 2000; Piao et al., 2001; Cacciotti et al., 2005). In the future, a focus on the molecular mechanisms of compensatory cell proliferation and transformation to malignancy will prove to be useful in understanding the disease process.

The description of genetic toxicology studies in the Roadmap needs to be improved. Although various types of asbestos have been shown to induce chromosomal aberrations and sister chromatid exchanges in human mesotheliomas and lung cancers and in cultured human and mammalian cells, mutagenic studies at most mammalian genetic loci have largely been negative (Jaurand, 1996; Hei et al., 2000; Schins and Hei, 2006, for review). This has been attributed to the findings that asbestos induces, either directly or indirectly² through the production of reactive radical species, multilocus deletions that are not easily recoverable at the hprt and oua loci.³ The observation is consistent with data obtained using other mutagenic assays that are proficient in detecting either large deletions, homologous recombinations, or score mutants located on a nonessential gene (Both et al., 1994; Park and Aust, 1998). These findings provide a direct link between chromosomal abnormalities that have frequently been demonstrated in human and rodent cell lines exposed to asbestos and carcinogenicity in vivo. If an elongated mineral particle type is found to be mutagenic in any in vitro genotoxic endpoint, it needs to be further evaluated.

The Roadmap describes in detail the cell-signaling pathways including nuclear factor kappa B, mitogen-activated protein kinase, and c-Jun-N-terminal kinase, that are activated in murine mesothelial cells and immortalized human bronchial epithelial cells treated with either asbestos or nonasbestiform analogs. However, what is missing is a statement on the specificity of these signaling end points for asbestos toxicology. The observation that antioxidants such as vitamin E and catalase can ameliorate the response indicate that this is a general response to oxidative damage. As such, glass fibers, for example, of similar aspect ratio as asbestos that are noncarcinogenic in vivo are likely to be positive in the activation of these signaling end points.

Recent studies on gene expression profile in human mesothelial cells treated with asbestos suggest that the activating transcription factor-3 (ATF3), which modulates the production of inflammatory cytokines (Shukla et al., 2009), is an important marker in asbestos-treated cultures. In addition, there is evidence that extranuclear targets, including mitochondria, are relevant to asbestos toxicology (Xu et al., 2007). These recent findings provide a conceptual link from frustrated phagocytosis of asbestos to reactive radical species, to cytokine production, to tissue inflammation, and ultimately to fibrosis or carcinogenesis.

In Vivo Toxicity Assays

Elongate mineral particles ranked as likely hazardous based on their mineralogical characteristics or having characteristics that prevent close correspondence to particles of known toxicity would be tested in vivo. It should be acknowledged that the three general types of responses of greatest concern, based on experience with asbestos, are fibrosis, mesothelioma, and lung cancer. Moreover, the most important physical and chemical characteristics probably differ among the three outcomes (Lippmann, 1988). The highest standard for in vivo assessment would be repeated inhalation exposures of animals, by which these outcomes could be revealed. Other routes of administration (e.g., intrapleural, intraperitoneal injection) might be useful for certain types of mechanistic studies, but their relevance to human exposure conditions is considerably less certain. Intracavitary injection is not an acceptable substitute for deposi tion via the respiratory tract. Although results of animal studies can rarely be extrapolated to estimates of human risks with absolute confidence (a point that the Roadmap should make more clearly), reliance on animal inhalation studies as a basis for regulatory classification is a well-established practice.

The Roadmap reviews the testing of elongate mineral particles in animals, but does not indicate which animal models have been validated. Although the Roadmap states that “it remains uncertain which species of animal(s) best predict(s) …” (NIOSH, 2009, p. 74), there are publications addressing this issue. For example, the European Union has adopted the in vivo biopersistence protocol ECB/TM/27 rev.7 which is a well-validated 90-day intratracheal test for respirable fibers and particles (European Commission Joint Research Centre, 1999). The Roadmap would benefit from including summary tables, such as those presented by Hesterberg (2009), showing the correlation between disease in animals and humans with various elongate mineral particles known to be nonhazardous. Indeed, the data presented in the Roadmap appear adequate to provide such a correlation. If the correlations were adequately reviewed, it might become clear which types of animal studies would or would not be useful to predict the relative toxicity of an unknown elongate mineral particle with a reasonable degree of certainty. The committee considers that existing data may be sufficient to provide guidance on the design of animal studies. In that context, the following points are presented.

• The particles to be used in toxicological studies need to be well characterized and to the extent possible, the exposure should be designed to achieve target-tissue dosing that simulates dosing in human exposures. A positive control may be warranted if comparative toxicity is an important issue. A negative control may not be necessary if the model has been adequately validated.
• Inhalation appears to be the most relevant route of exposure as proposed by the International Life Sciences Institute (ILSI, 2005) for the study of manufactured vitreous fibers. Other routes can be considered depending on the characteristics of the unknown mineral.
• The rat appears to be the most appropriate rodent species, based on past experience (Mauderly, 1997). Syrian hamsters are known to be less useful in chronic studies because of their susceptibility to infection, and studies of mice may not add significantly to the findings in rats.
• Both genders of rats may not be required because very few studies of asbestos have shown a gender difference (Wagner et al., 1973, 1974, 1982). However, females have demonstrated greater responses than males to some nonelongate particles (Nikula et al., 1995). The potential for gender-based differences in response to elongate mineral particles is unknown.
• Animals should be exposed to multiple concentrations. Exposure levels should be selected based on several considerations, such as maximum tolerated dose, number of thoracic elongate mineral particles in the aerosol having the length range of concern (e.g., >20 μm long), and levels relevant to human exposure. The spacing of dose levels should be chosen to provide the best data for risk assessment. Caution should be exercised to avoid exposures that deposit unrealistically large doses of mineral particles, in order to avoid nonspecific fibrotic and proliferative responses to overloading, such as have occurred with some non-fibrous particles (Oberdörster, 1995).
• Differences in the aerodynamic sizes of particles that are respirable by humans and rats (thoracic size fraction) should be considered. Aerodynamically larger particles can be inhaled and deposited in the thorax of humans in comparison to rats. Accordingly, only a portion of a highly polydisperse population of elongate mineral particles may actually deposit in the lungs of rats. Characterization of the aerodynamic particle size range of the aerosol is necessary to determine whether or not the full range of human exposure is being tested. Dogs and nonhuman primates more accurately model the initial deposition of highly polydisperse elongate particles in humans and could be used for acute studies of initial deposition.
• Subchronic (90-day) exposures may be adequate. ILSI (2005) suggests that a 90-day exposure followed by a 90-day recovery period should be adequate to determine the relative toxicity of elongate mineral particles if subchronic data are available for other materials for comparison. In some cases, lifetime exposures may be warranted. As part of this study, it would be valuable to conduct lung burden evaluations to determine biopersistence, changes in elongate mineral particle morphology, (e.g., splitting and breaking), and changes in the distribution of the elongate mineral particles in the respiratory tract over time.
• Pulmonary fibrosis is a useful predictive end point. ILSI (2005) concluded that pulmonary fibrosis was a useful screening end point for subsequent outcomes such as interstitial fibrosis, COPD (chronic obstructive pulmonary disease), lung cancer, and mesothelioma. In the rat, pulmonary fibrosis has consistently preceded the development of lung cancer and mesothelioma from inhaled elongate mineral particles. However, pulmonary fibrosis is not always followed by the other diseases. In this sense, pulmonary fibrosis in the rat represents a conservative end point for estimating longer-term hazards. In some cases, the longer-term outcomes may have to be explored by lifetime inhalation studies.
  Decision point: If pulmonary fibrosis is observed, the material in question should be considered likely to be hazardous. Because not all fibrosis leads to cancer, the development of fibrosis does not confirm a carcinogenic hazard.

It must be stressed that adequate human data should always take priority over animal data in characterizing the risk from elongate mineral particles. Robust positive data from humans are sufficient for regulatory action. However, negative data from humans may not be adequate to discount long-term risks from exposures to elongate mineral particles because of the relatively long latency period in humans for some diseases. Toxicological testing, both in vitro and in vivo, is appropriate in the absence of data, or the absence of robust data, from humans.

Assessment of the State of the Science

The Roadmap discusses epidemiological studies of workers with mining and/or milling exposures to mineral particles that have been reported to be nonasbestiform from three different regions: the talc mining region of upstate New York, the Homestake gold mine in South Dakota, and the taconite iron ore mines in northeastern Minnesota. The Roadmap section reviews in sufficient detail the published reports of the epidemiological studies of these occupational cohorts through 2007.

The subsection on studies of the New York talc miners and millers concludes with a summary indicating that an excess of pneumoconiosis and pleural plaques is “well recognized to have occurred among workers exposed to talc” (NIOSH, 2009, p. 22). The dust from these talc mines has been reported to contain nonasbestiform tremolite, asbestos anthophyllite altering to talc at the nanometer scale (considered as unregulated by OSHA), and antigorite-lizardite particles (NIOSH, 1980; Kelse, 2005). However, the NIOSH study and others document the presence of asbestos. Mesothelioma rates have been reported to have been elevated in Jefferson County, the site of much of the talc industry in New York (Vianna et al., 1981; Enterline and Henderson, 1987; Hull et al., 2002), but these studies used a nonspecific International Classification of Diseases (ICD) code for malignant neoplasms of the pleura that may have included cancers metastatic to the pleura. Data from 1999 to 2004 using a specific ICD-10 code for malignant mesothelioma show no excess of mesothelioma in Jefferson and St. Lawrence counties (NIOSH, 2009). The most controversial aspect of the published epidemiological studies from the New York talc industry is excess lung cancer mortality. While this finding has been consistently reported, so has the lack of an exposure-response relationship.

The subsection on studies of Homestake gold miners notes that three different groups of investigators have reported on lung cancer risk among these workers (Gilliam et al., 1976; McDonald et al., 1978; Brown et al., 1986, which was updated by Steenland and Brown, 1995). The dust from this mine has been reported to contain nonasbestiform mineral particles that are mostly cummingtonite-grunerite, along with accessory tremoliteactinolite and other amphibole varieties. No excess of mesothelioma has been documented among these gold miners. Although small excesses of lung cancer deaths were found in the most recent study (Steenland and Brown, 1995), there was again a lack of a cumulative dust exposure response. However, the subsection also notes that total dust is likely a poor surrogate for exposure to nonasbestiform mineral particles, making this study largely uninformative about the risk of lung cancer associated with such exposure. This subsection concludes that the studies of the Homestake gold miners provide at best weak evidence of an excess risk of lung cancer, and there are inadequate data on worker exposures to nonasbestiform minerals.

The subsection on studies of taconite iron ore miners and millers in northeastern Minnesota indicates that recent sampling and analysis of the ores from the taconite iron ore mines reported no asbestos, but did find ferroactinolite, ferrian sepiolite, grunerite-ferroactinolite, and actinolite, some of which was fibrous. The subsection concludes with a summary noting that cohort mortality studies (Higgins et al., 1983; Cooper et al., 1988, 1992) have not provided any evidence of an increased risk of respiratory cancer or mesothelioma. In contrast, recent reports from the Minnesota Department of Health (MDH, 2007; Brunner et al., 2008) have documented an excess of mesotheliomas among males, but not females, in this region of Minnesota and many of these cases had been workers in the taconite industry. There is evidence that at least some of the cases may have had exposures to asbestos, but further research to resolve this issue is ongoing.

The summary of the section states that the results from these studies of workers reportedly exposed to nonasbestiform mineral particles “do not provide clear answers regarding the toxicity” (NIOSH, 2009, p. 26) because of a number of limitations. First, all three populations of workers studied were exposed to complex mixtures of particles that included only a relatively small percentage of nonasbestiform mineral particles. Second, data on past exposures to those particles are inadequate, and exposure to total dust is likely a poor surrogate for such exposure. Third, the reliability of death certificate information for mesothelioma diagnosis prior to 1999 may be poor. Fourth, the lack of individual smoking data is another major limitation for interpretation of any lung cancer risk among these populations. Because of these limitations, the section concludes that “the findings from these studies should best be viewed as providing inconclusive as opposed to negative evidence regarding the health hazards associated with exposures to nonasbestiform elongated mineral particles” (NIOSH, 2009, p. 27). This conclusion is judged by the committee to be appropriate. To be useful for causal inference, epidemiological studies require sufficient data on the relevant exposures, health outcomes, and confounding factors. Unfortunately, sufficient data on exposures, outcomes, and confounders do not appear to be available for any of the three U.S. populations. Information or research needs on the potential for synergistic interactions between smoking and elongate mineral particle exposure should be detailed.

The section further concludes that additional studies of these three populations would need improved characterization of exposures to elongate mineral particles, diagnosis of mesothelioma, information on smoking, and information about exposures in other employment; additional studies should be attempted only if “these improvements are deemed feasible” (NIOSH, 2009, p. 27). The committee agrees with this conclusion. Although improved exposure assessment may be possible through reanalysis of archived samples from the three mining sources, it is unclear that improved information on smoking and other employment can be obtained for the New York talc and the Homestake mining populations. Perhaps adequate additional information on mesothelioma diagnosis, smoking, and other employment can be obtained through the ongoing study of the Minnesota taconite-exposed population.

The committee agrees with the Roadmap’s assessment of the results of the epidemiological studies on elongate mineral particles as inconclusive.

Gaps in the Roadmap and Research Directions

The epidemiological studies discussed in the Roadmap reflect NIOSH’s concern with the potential toxicity of elongate mineral particles, and the committee appreciates this concern. The Roadmap however, as currently titled and configured, is intended to outline a research framework for asbestos as well as other elongate mineral particles. In this light, attention to epidemiological studies of workers and nonworkers exposed to vermiculite from the mine in Libby, Montana, is important. NIOSH investigators have been involved in the study of Libby miners and millers for many years (Amandus and Wheeler, 1987; Sullivan, 2007). Because the amphiboles in the Libby ore have been recently reclassified using 1997 IMA amphibole nomenclature to be predominantly winchite and richerite rather than tremolite, it behooves the Roadmap to specify the ongoing research efforts to better understand the exposure-response relationships for various health outcomes (pleural plaques, pulmonary fibrosis, mesothelioma, lung cancer) among both workers and residents. The risk of asbestosis from exposure to elongate mineral particles in nonoccupational settings has not been studied adequately. Significant excess mortality from nonmalignant respiratory disease has been reported by NIOSH among Libby workers with relatively low cumulative lifetime exposures (i.e., at a level allowable by the current OSHA standard over a 45-year working life) (Sullivan, 2007). Such an exposure-response is striking and deserves further investigation. In addition to discussions of occupational exposures, further attention to environmental exposures is needed in keeping with the goal of setting out a research plan that could be used by multiple federal agencies and other organizations to address the range of potential exposures. In examining the impact of environmental exposures it will be important for researchers to explore the effects of childhood exposures as well as exposures in other potentially vulnerable populations.

Further, the document discusses only U.S. epidemiological studies. Multiple international studies of exposure to amphibole particles have not been included. While the committee understands that the Roadmap was not intended to be a comprehensive review of the asbestos epidemiological literature, some attention to these studies seems relevant for planning future research on the health effects of exposure to elongate mineral particles. Several studies have shown risk of pleural plaques and/or mesothelioma to be associated with exposure to amphibole particles in either soil or whitewash derived from soil in some Mediterranean countries and in New Caledonia (Sakellariou et al., 1996; Luce et al., 2000, 2004; Senyiğit et al., 2000; Menvielle et al., 2003). It is important to determine to what extent it is possible to extrapolate from exposures and health outcomes recognized in these studies to populations with occupational exposures.

Scientific Rationale and Research Directions

As noted in discussions in Chapter 2, efforts are needed to clarify the goals of the roadmap and to lay out the research hypotheses. The Roadmap states that “research is needed to assess and quantify potential human health risks associated with occupational exposures to other mineral fibers and elongated mineral particles, as well as to better understand and quantify the epidemiology of asbestos-related diseases using more refined indices of exposure” (NIOSH, 2009, p. 76). The committee agrees with both aspects of this statement.

The Roadmap recommends that “it would be reasonable to thoroughly review, assess, and summarize the available information on asbestiform amphiboles that have not been commercially exploited as asbestos” (NIOSH, 2009, p.77). The committee agrees that such a review could be informative. The Roadmap specifically indicates that it is not meant to be a comprehensive review of the scientific literature. However, such a review (with the literature search methodology thoroughly detailed) would provide NIOSH with a foundation on which to build a focused research program by identifying specific data gaps that are relevant to worker protection as well as potential opportunities for collaboration with partners in the United States (e.g., EPA, Agency for Toxic Substances and Disease Registry [ATSDR]) regarding issues in Libby, Montana, and in other countries where populations have been exposed to amphiboles that have not been commercially exploited.

The Roadmap also notes the need to determine whether elongate amphibole particles pose a risk to human health and recommends that an expert panel be assembled to evaluate whether the existing epidemiological evidence could support development of a likely maximum risk estimate associated with exposure to these elongate mineral particles. Based on the review of the epidemiological literature contained in this document, it does not appear that the epidemiological evidence is sufficiently robust for such an endeavor.

Section 2.3.2 of the Roadmap briefly discusses the possibility of using new tools for the diagnosis of asbestos-related health effects at earlier stages of disease such as “modern medical pulmonary imaging tech niques” (positron emission tomography [PET] scanning is specifically mentioned) or “bioassays of circulating levels of cytokines or other biochemical factors associated with disease processes” (NIOSH, 2009, p. 78). The committee notes that although the development of sensitive bioassays for intermediate end points on the pathways from exposure to asbestos and other elongate mineral particles to pulmonary fibrosis, lung cancer, and mesothelioma is a laudable research goal, the application of such bioassays to medical surveillance of exposed workers would require a major and lengthy validation effort given the long latency of these health outcomes. Research is needed on radiographic changes induced by elongate mineral particles other than asbestos to determine if these differ from asbestos-induced changes. Because chest computed tomography (CT) scanning is a better tool than the standard chest X-ray to detect fiber-induced change, it would play a major role in such research.

Modern chest imaging techniques such as high-resolution CT or PET scanning are not likely to be useful in medical surveillance of asbestosor other related disease in the near term due to the level of radiation exposure, lack of portable equipment, and expense. More details are needed in the Roadmap on NIOSH’s effort to advance the International Labour Organization’s (ILO’s) pneumoconiosis radiographic classification scheme and the NIOSH B-reader certification program into the modern era of digital chest radiographic imaging. This effort is vitally important to the ability to sustain asbestos medical surveillance programs as well as conduct epidemiological research in populations exposed to asbestos or other elongate mineral particles. The committee encourages NIOSH to continue to support its program to develop digital images of the ILO standard films and software to project these images correctly.

The Roadmap discusses the possibility of conducting new studies of (1) worker populations exposed to amphibole cleavage fragments such as the New York talc and Minnesota taconite miners and millers; (2) populations such as Libby workers and residents incidentally exposed to a range of mineral particles from mines; (3) populations exposed to less well-studied elongate mineral particles such as wollastonite and attapulgite; and (4) meta-analyses of data from previous studies of populations exposed to mineral particles with various attributes (NIOSH, 2009). The section specifically lists criteria for selecting and prioritizing such studies, including adequacy of exposure information and work histories, sufficiency of latency and sample size, and availability of data on potential confounding factors such as cigarette smoking. The committee agrees that these criteria should be considered before any new epidemiological study of an exposed population is begun. It is recommended that priority be given to epidemiological studies that will contribute to better understanding of elongate mineral particle characteristics that determine toxicity and to ensuring that rigorous mineralogical characteristics are established. Research should include identifying biomarkers of toxicity that can be studied in human populations as has been discussed at a recent expert panel meeting held by ATSDR (ATSDR, 2008).

Further, the Roadmap states that opportunities for epidemiological studies of workers exposed to asbestos and the other elongate mineral particles are present in other countries. As noted above, the committee urges more discussion of these opportunities in the Roadmap.

Exposure Characterization

The Roadmap contains a thorough discussion of issues associated with sample analytical techniques. There is also recognition of the importance of developing and employing comprehensive sampling strategies to characterize exposures that comprise more than one type of mineral particle (mixed exposures). There is tremendous value in performing comprehensive exposure characterizations in workplaces with mixed exposures regardless of the suitability of these workplaces for health effect studies. Detailed exposure characterizations should be multidimensional and focus on spatial and temporal variability in elongate mineral particle number, size distribution, and other physical and chemical characteristics. The exposure characterization should go beyond analyzing or reanalyzing previously collected samples. Sampling should be task- or activity-based with sufficient power to provide statistically valid estimates of exposure, a detailed description of important physical and chemical properties, and key exposure determinants. In the review section of the Roadmap the need for such an assessment is recognized: “these initial efforts should be supplemented with efforts to systematically identify, sample, and characterize elongated mineral particle exposures throughout U.S. industry” (NIOSH, 2009, p. 76). Roadmap section 2.3.1 recognizes the need to conduct such an assessment. These data will be important for designing, conducting, and interpreting epidemiological and toxicological studies.

Bulk samples of airborne material from selected well-characterized workplaces can be collected as reference samples to be included in the national repository of asbestos and related minerals only after rigorous mineralogical characterization. Attention should be paid to conducting source characterization studies in order to link sources of elongate mineral particles associated with ambient and/or nonoccupational exposure. Similar tools used for bulk-sample characterization in the workplace can also be applied to identifying sources associated with environmental exposures.

Sampling and Analysis

The review of current issues in the Roadmap covers the methods of sampling and analysis for standardized industrial hygiene surveys (e.g., PCM), analytical methods for research, and a short section on differential counting and other proposed methods. One of the strategic goals of the Roadmap is to “develop improved sampling and analytical methods for asbestos fibers and other elongated mineral particles” (NIOSH, 2009, p. 65). Five specific areas of research are listed:

1.

Reduce the inter-operator and inter-laboratory variability of the current analytical methods used for asbestos.

2.

Develop analytical methods with improved sensitivity to visualize thinner elongate mineral particles to ensure a more complete evaluation of airborne exposures.

3.

Develop a practical analytical method for air samples to differentiate between asbestos exposures and exposures to nonasbestiform elongate mineral analogs.

4.

Develop analytical methods to assess the durability of elongate mineral particles.

5.

Develop and validate size-selective sampling methods for elongate mineral particles (NIOSH, 2009, p. 65).

A more comprehensive review of all microscopy methods, including SEM, for counting should be presented in the Roadmap’s review of current issues. Modern SEM holds promise as an analytical tool and, as a result, the discussion of SEM capabilities should be expanded. There is little discussion of the variability associated with asbestos counting in the section of the Roadmap that reviews current issues; yet this topic is one of the priority research areas. Additional discussion of this issue will help to justify its importance as a research priority.

Section 1.7 of the Roadmap discusses the use of analytical methods for industrial hygiene surveys as separate from analytical methods for research. The discussion focuses on integrating the use of optical microscopy methods for surveys and electron microscopy for research. The value of maintaining this distinction needs to be reconsidered. The use of drastically different exposure assessment tools for these two cases will magnify uncertainty about exposure and risk. Phase contrast microscopy methods have been used for more than 40 years. Many of the research gaps identified in the Roadmap and in this report are the result of inherent weaknesses of PCM. Advancement of our understanding of the risk associated with elongate mineral particles will not be achieved until we move aggressively to develop new techniques and employ suitable methods (e.g., PLM) already available. The Roadmap and its proposed research provide an important opportunity to make this case. The limitations of electron microscopy mentioned in the Roadmap (Section 2.4) should not be used as an excuse to maintain the status quo. Electron microscopy methods are used routinely for environmental analysis. In addition, occupational hygienists routinely collect samples that require sophisticated analyses that take days to weeks to turnaround. These limitations seem like a reasonable trade-off. Another way to look at the difference between a routine assessment tool and a research tool is that a routine assessment is less comprehensive, not an inferior method.

The document presents a good case for why PCM is an inadequate tool for elongate mineral particle exposure assessment yet falls short of recommending that it be replaced with PLM and electron microscopy techniques. PCM does not provide the detailed exposure characterization needed to conduct the complex risk assessments described in the Roadmap. Research by its nature drives the development of new technology. The Roadmap focuses to too great an extent on current costs and time delays associated with electron microscopy methods. As noted on page 82, “in some workplace situations, such as in construction, increases in the time needed to analyze samples and identify elongated mineral particles could potentially delay the implementation of appropriate control measures to reduce exposures.” The discussion should rather be focused on research to develop new tools.

The committee believes that the Roadmap should place a high priority on developing new electron microscopy-based methods and a lower priority on improving PCM. The Roadmap should also include a recommendation that research be conducted to relate old methods to new methods to maintain a direct link to information determined using PCM.

The Roadmap also needs to emphasize the importance of integrating the development of new and/or improved sampling and analytical methods in parallel with toxicological and epidemiological research. Since sampling and analytical methods are designed for risk assessment purposes, their development should be driven by hazard assessment information produced in these studies. For example, it does not make sense to develop solubility-based methods until toxicology studies better assess the boundaries of dissolution that are risk-based.

The Roadmap emphasizes the need to characterize thoracic elongate mineral particle deposition. The rationale for this choice must be presented. It is not clear that a thoracic-based metric is appropriate for all disease outcomes. Is a respirable elongate mineral particle metric more appropriate for mesothelioma risk? It is conceivable that different size-based exposure metrics may be appropriate for different health outcomes.

Regardless of method used, more discussion of the statistics of elongate mineral particle counting and counting quality control is needed. The advancement of science with respect to elongate mineral particle exposure assessment requires nationally recognized quality control programs. Expectations should be changed so that each count is presented as an estimate with a range of expected variability (see below).

The Roadmap should include a specific recommendation that automated counting methods be explored and compared to human counting with respect to precision and accuracy.

Counting Strategies

In reviewing the literature on asbestos counting, there appears to be considerable variability in counts from analyst to analyst within a given laboratory as well as between laboratories. As a consequence, a new observed count from a particular analyst from a particular laboratory may deviate considerably from the true value. In an effort to provide a connection between observed and true counts and characterize the uncertainty in the true count, Dulal Bhaumik and colleagues have extended the ideas of Gibbons and Bhaumik (2001) and Bhaumik and Gibbons (2005) to the case of a Poisson random variable, which is the appropriate distribution for rare-event count data. Appendix B provides a brief sketch of one potential methodology for addressing the variability as well as an illustration of the application of this methodology. Statisticians who are also familiar with exposure data analysis should be actively involved in addressing some of the challenging data analysis issues in this area of research.

Additional Statistical Issues

There is considerable discussion in the NIOSH Roadmap on the effect of fiber counts and fiber dimensions on exposure risk. Thus the distribution of these quantities is of obvious interest. For exposure assessment, it is important to characterize the joint distribution of fiber length and width for a given material. Several researchers (see Cheng, 1986; Baron, 2001; Cheng et al., 2006) have investigated this problem and noticed that fiber length and width typically have a bivariate log-normal distribution. Data analysis based on the bivariate lognormal distribution can be complicated, depending on the parameter or parameters for which inference is desired. Likelihood based results for testing or constructing confidence intervals for one or more parameters of the distribution often produce undesirable results (e.g., inflated type 1 error rates or low coverage probability) when sample sizes are small. In order to overcome this problem one can explore procedures based on the novel concepts of generalized p-values (for hypothesis testing) and generalized confidence intervals (for computing confidence intervals) for univariate and bivariate lognormal distributions (see Krishnamoorthy and Mathew, 2003; Krishnamoorthy et al., 2006; Bebu and Mathew, 2008). Major advantages of such procedures are that they are accurate and are applicable to small samples. The concepts of generalized p-values and generalized confidence intervals also provide accurate methodology for comparing two lognormal distributions (for example, to compare the arithmetic means of fiber lengths obtained from two different sites or materials).

The NIOSH Roadmap also addresses the issue of comparing thoracic samplers. To compare thoracic samplers, the OSHA criterion for establishing the equivalence of a sampling device to a reference device requires that “90 percent of the readings of the sampling device should be within plus or minus 25 percent of the readings obtained by the reference device, or within plus or minus 25 percent of the actual airborne chemical concentration.” In this context one can use rigorous statistical tests developed by Krishnamoorthy and Mathew (2002) for comparing two samplers, and Krishnamoorthy et al. (2009) for comparing several samplers. The OSHA criterion, or a suitable version of it, appears to be the right criterion to compare thoracic samplers. Traditional approaches for comparison of samplers based on geometric means using t-tests are generally inadequate.