Comparing milled fiber, Quebec ore, and textile factory dust: Has another piece of the asbestos puzzle fallen into place?

Abstract

Results of a meta-analysis indicate that the variation in potency factors observed across published epidemiology studies can be substantially reconciled (especially for mesothelioma) by considering the effects of fiber size and mineral type, but that better characterization of historical exposures is needed before improved exposure metrics potentially capable of fully reconciling the disparate potency factors can be evaluated. Therefore, an approach for better characterizing historical exposures, the Modified Elutriator Method (MEM), was evaluated to determine the degree that dusts elutriated using this method adequately mimic dusts generated by processing in a factory. To evaluate this approach, elutriated dusts from Grade 3 milled fiber (the predominant feedstock used at a South Carolina [SC] textile factory) were compared to factory dust collected at the same facility. Elutriated dusts from chrysotile ore were also compared to dusts collected in Quebec mines and mills. Results indicate that despite the substantial variation within each sample set, elutriated dusts from Grade 3 fiber compare favorably to textile dusts and elutriated ore dusts compare to dusts from mines and mills. Given this performance, the MEM was also applied to address the disparity in lung cancer mortality per unit of exposure observed, respectively, among chrysotile miners/millers in Quebec and SC textile workers. Thus, dusts generated by elutriation of stockpiled chrysotile ore (representing mine exposures) and Grade 3 milled fiber (representing textile exposures) were compared. Results indicate that dusts from each sample differ from one another. Despite such variation, however, the dusts are distinct and fibers in Grade 3 dusts are significantly longer than fibers in ore dusts. Moreover, phase-contrast microscopy (PCM) structures in Grade 3 dusts are 100% asbestos and counts of PCM-sized structures are identical, whether viewed by PCM or transmission electron microscope (TEM). In contrast, a third of PCM structures in ore dusts are not asbestos and only a third that are counted by PCM are also counted by TEM. These distinctions also mirror the characteristics of the bulk materials themselves. Perhaps most important, when the differences in size distributions and PCM/TEM distinctions in these dusts are combined, the combined difference is sufficient to completely explain the difference in exposure/response observed between the textile worker and miner/miller cohorts. Importantly, however, evidence that such an explanation is valid can only be derived from a meta-analysis (risk assessment) covering a diverse range of epidemiology study environments, which is beyond the scope of the current study. The above findings suggest that elutriator-generated dusts mimic factory dusts with sufficient reliability to support comparisons between historical exposures experienced by the various cohorts studied by epidemiologists. A simulation was also conducted to evaluate the relative degree that the characteristics of dust are driven by the properties of the bulk material processed versus the nature of the mechanical forces applied. That results indicate it is the properties of bulk materials reinforces the theoretical basis justifying use of the elutriator to reconstruct historical exposures. Thus, the elutriator may be a valuable tool for reconstructing historical exposures suitable for supporting continued refinements of the risk models being developed to predict asbestos-related cancer risk.

Keywords::

• Asbestos
• chrysotile
• elutriator
• fiber size
• lung cancer
• mining
• PCM
• PCME
• TEM
• textile

Acknowledgements

I would like to thank NIOSH, and Everett (Chip) Lehman in particular, for making available the unsummarized data for the cohort of workers from the textile plant at Charleston, South Carolina. I would also like to thank Kenny Crump for sharing his combined master file extracted from the multiple files of the original NIOSH data, which facilitated my summarization and analysis, and John Dement for assisting Kenny Crump and me in correcting our extracts and interpreting the South Carolina data. Acknowledgments are also extended to Owen Crankshaw and Jim Millette for sharing the results from their studies of interlaboratory comparisons and Drew Van Orden for providing the photomicrographs. I gratefully acknowledge the guidance and support provided by Eileen Kuempel, Ralph Zumwalde, David Weissman, and Frank Hearl, and thank Ann Wylie, Martin Rabenhorst, Bruce Case, Pamela Vacek, Brooke Mossman, and Steven Lower for their input. I would also like to thank Anthony Kolk and the staff at EMS Laboratories and Pierre Marois and the staff at Centre de technologie minérale et de plasturgie inc. for their assistance with sample preparation and analysis and Marianna Berman for her assistance with technical editing.

Declaration of interest

I consult for a variety of government and private organizations with competing interests, who (to the best of my knowledge) have no direct financial stake in the outcome of this research. As I have no financial stake in use of the elutriator method or its associated equipment, except for attracting additional research funding, I have no direct financial interest in the outcome of this research. Funding for all early work related to the subject matter of this paper was provided by the US Environmental Protection Agency (EPA). Funding provided for writing this paper and conducting the supporting analyses was provided as a grant by The National Stone, Sand, and Gravel Association (NSSGA). The NSSGA had no input into the preparation of the manuscript. The financial support provided by the EPA and NSSGA should not be construed as an endorsement of the results of my analyses or of this paper. The author has sole responsibility for the analyses, writing, and content of the paper.

Notes

¹An exposure metric is a weighted set of size categories that are defined in the counting rules of an analytical method used to determine asbestos concentrations. The phase-contrast microscopy (PCM) metric, for example, is the set of all structures longer than 5 µm (micrometers) with an aspect (length to width) ratio greater than 3 when viewed using PCM as described in NIOSH Method 7400 (NIOSH, 1994a).

²As used here and throughout, the term “structure” is intended to include not only single fibers (fibrils), but the bundles, clusters, and matrices that make up the set of fibrous particles in an asbestos dust (ISO, 1995).

³Due to the limitations of optical microscopy, when analyzed using other analytical techniques (such as transmission electron microscopy [TEM]), the metric is limited to structures thicker than 0.25 µm (NIOSH, 1994a, 1994b). By limiting widths in this manner, structures respectively included when PCM-sized structures are determined by TEM (the PCM-equivalent or PCME metric) and the PCM metric (determined by PCM) are better matched.

⁴The term “asbestiform” means the particular crystalline form (habit) of a mineral that exhibits the properties of asbestos (composed of high-tensile-strength fibers that are flexibility and resistant to chemical and thermal attack).

⁵In this context (and within this study), the term “character of exposure” is intended to mean the size distribution of structures found in the exposure. This is distinct from (and not to be confused with) the intensity (magnitude) of exposure, which the Modified Elutriator Method does not reproduce. Rather, when used in risk assessment, the magnitude of exposure is derived from the exposure estimates of the published epidemiology studies themselves and then linked to the character of exposure either through the PCME fraction of the size distribution in the manner previously described (Berman and Crump, 2008b) or similar procedures suitable for the data that are developed.

⁶A mineral’s habit is the crystalline form in which it found. Many minerals occur in multiple crystalline habits, including those that occur in the asbestiform habit (as asbestos); nevertheless, even the asbestos-related minerals primarily occur in nonasbestos habits.

⁷A vertical elutriator is simply a circular tube through which air is passed at a controlled velocity equal to the settling velocity of the largest particle of interest (in this case, the largest respirable particle) so that all particles of this size and smaller will rise to the top and be collected on filters, whereas all larger particles will fall to the bottom and be eliminated (Berman and Kolk, 1997, 2000).

⁸Scanning electron microscopy.

⁹Even allowing for additional subdivision of the size categories defined in the stopping rules, numbers were selected to favor a minimum average count of five structures per category. This means that counts in most categories would be determined to better than a factor of 2. In some cases, however, cost considerations meant that stopping rules had to be relaxed for some samples.

¹⁰Traditionally, the PCME metric is defined as structures composed of an asbestos mineral that satisfy the dimensional requirements for PCM, but are identified by TEM. Therefore, when PCME is used in this paper to indicate all structures satisfying the dimensional requirements (no matter their composition), a subscript will be added to clarify the distinction: PCMEall. PCMEall counts are required, for example, when comparing to PCM counts.

¹¹Use of washed play sand is a standard procedure described in the method (Berman and Kolk, 2000). It promotes stable emission of dusts during tumbling and elutriation. Prior to mixing, “sand blanks” are run in the elutriator to assure that it will contribute no asbestos and no more than inconsequential amounts of respirable dust to the filters collected from actual samples.

¹²The published distributions were derived from analysis of air-sample filters that were collected from facilities of interest. Though typically collected over a relatively short time interval, such samples represent dusts actually generated by commercial processing at each particular facility.

¹³A further complication described for this study (Dement et al., 2007; John Dement personal communication) was that the size boundaries for these categories were originally set assuming a target (ideal) magnification, rather than accounting for the actual magnifications under which the analyses were completed. The consequence of this is that the size boundaries for the categories are not precisely “as advertised” so that (at least) a small number of structures have potentially been misclassified by placing them in the wrong category. In such cases, however, such structures would have been placed in categories for sizes contiguous with the category representing the correct size. Except for those categories with relatively few structures, the effect of this consideration on overall size distributions may be relatively limited and has been further ameliorated by at least partially addressing the effect during extraction of summarized information from the original, unsummarized data (John Dement, Kenny Crump, personal communication).

¹⁴Previously defined, see Footnote 10.

¹⁵Importantly, because the elutriated Quebec ore and Grade 3 milled fiber samples were analyzed by the same laboratory using a common method, this source of variation was not relevant to comparisons between those analyses, which is why this issue was not previously considered in this paper.

¹⁶The manner in which the distributions from preparation, twisting, and weaving were combined into a single distribution for this comparison has been previously described (Berman et al., 2008b).

¹⁷Thetford Mines is a mining area in Quebec from which a substantial fraction of the Quebec cohort of miners/millers derives.

¹⁸Interestingly, three of the structures observed among the approximately, 2000 structures counted across the three elutriated Grade 3 milled fiber samples analyzed for this study were found to be tremolite. No tremolite fibers were observed among ore samples. For the reasons previously cited in the Background section, however, a substantially greater number of samples would need to be analyzed before any meaningful inferences might be drawn from such data.

¹⁹As long as the age distributions across the subgroups are not radically different, this assumption is reasonable.