Asbestos exposure is a recognized risk factor for a variety of serious diseases, including asbestosis, lung cancer and mesothelioma. Research on the mechanisms underlying these diseases, as well as preventive measures and therapies, requires precise studies and the use of appropriate control groups. Asbestos fiber concentration in human lung tissue has proved to be an important biomarker for asbestos-associated lung diseases (1, 2). Since Wagner (3) reported 33 cases of diffuse pleural mesothelioma in the Northwest Province following exposure to crocidolite asbestos, there has been a myriad of analytic epidemiologic studies correlating asbestos and disease in worker cohorts, and case–control studies which have demonstrated the importance of asbestos fiber type, fiber size, fiber biopersistence and latency in the induction of disease. Correlative mineralogic and pathologic studies (4–7) have played a key role in complimenting analytic epidemiology. These mineral analyses have consistently identified that mesothelioma and lung cancer risk, as well as extent of lung fibrosis/asbestosis is correlated with the retained elevated amphibole asbestos fiber content. These studies have also demonstrated that there is no correlation between asbestos related disease and retained chrysotile asbestos, a known biosoluble mineral in human tissue. Mineral analysis has provided complimentary evidence that chrysotile exposure is far less potent than commercial amphibole (crocidolite and amosite) asbestos in inducing asbestos-related diseases. In this analysis, the role of establishing control reference populations is a key tool to contextualizing the significance of an exposure to a mineral.

Control subjects are basically defined by the absence of any known occupational, para-occupational and/or environmental exposure to asbestos, as well as an absence of any asbestos-related disease. In asbestos-related disease risk-assessment, a negative control group serves as a baseline reference to define thresholds reflecting ambient air exposure. There is no significant association between ambient asbestos exposures and the development of asbestos related disease (8–13). In addition, it is recognized that ambient air asbestos fiber concentrations may vary over 10-fold in different geographic locations without any impact on the development of mesothelioma (12). With respect to control reference ranges—these are established within individual analytical laboratories performing fiber burden analyses and the significance of an individual exposure may then be determined by benchmarking the case to the established laboratory control population (14–20). The control population represents, most closely, only subjects with ambient asbestos exposure and no increased risk of asbestos related disease (21). The scientific evidence correlating cumulative asbestos exposure with disease has been extensively established in occupational settings which are many orders of magnitude above background ambient exposure levels. Numerous analytical laboratories have made correlations between retained amphibole asbestos fiber concentrations in lungs of asbestos exposed workers and asbestos related disease (4, 6, 22).

The Helsinki criteria proposed (23) the use of two sets of controls: a “background” control reference population for subjects without known significant exposure or disease and a control named “asbestosis range” reflecting the retained amphibole asbestos fiber count for cases with established histologic asbestosis. This background control can be used in determining causality for mesothelioma. The asbestosis range control can be utilized in claimed asbestos exposed subjects with suspected lung fibrosis, to allow for a determination of disease diagnosis—asbestosis, and for disease causality in asbestos-related lung cancer ex asbestosis. The proposal by the College of American Pathologists-Pulmonary Pathology Asbestosis Guidelines Committee (24) highlighted the use of the total amphibole asbestos count without consideration of chrysotile and the 5th percentile value of an asbestosis cohort as the “lower asbestosis range” value for use in analytic laboratories. These proposals were accepted by the Helsinki update (25).

Studies using lung tissue fiber analysis have shown an increased burden of amphibole asbestos in lungs of patients with asbestosis and mesothelioma compared to non-exposed individuals. Since the ninety’s the use of asbestos has been banned in western countries. Consequently, lung tissue fiber content in patients with mesothelioma has gradually dropped to levels generally observed in lungs of non-exposed individuals (26, 27) suggesting that other mechanisms unrelated to asbestos, such as exposure to other particles or genetic predisposition, have become more relevant in the carcinogenesis of mesothelioma (21, 28–33). This also stresses the need for universal criteria and technical standardization for detection of asbestos fibers to better identify negative control subjects with lung tissue fiber content that reflects ambient air exposure.

We aimed to gather information on asbestos content in lung tissue of non-exposed subjects. There are various terms to describe such a population: background or negative control, exposure to ambient air. Therefore, we performed a PubMed search using the general terms control subjects, asbestos and fiber analysis. Of the 1,349 hits, 26 papers (1, 15–18, 22, 26, 34–51) were selected for this review as they contained comprehensive information on negative control groups including fiber burden summarized in the flow chart in Figure 1. 1,323 publications studying asbestos fiber content only in occupationally exposed subjects and other tissues or air samples were excluded. For the selected studies, information on the selection criteria for control subjects, age, gender, occupational and environmental exposure, methods of fiber analysis, fiber type and fiber dimensions were collected for comparison.

Twenty-six studies conducted in 12 countries in three continents and published over almost five decades (46 yrs) were systematically reviewed. Marked interlaboratory heterogeneity was observed which rendered comparative analysis between reported numeric controls across laboratories problematic. Interlaboratory variation can be attributed to differences in sample preparation, analytical methods and statistical analyses. Differences in the geological setting, the degree of urbanization and industrialization can lead to variations in background exposure of different populations adding to the interlaboratory variations. This finding does not, in any way, diminish the significance or worth of mineral analytic fiber measurement in biologic samples and laboratories but stresses the need for laboratories to establish their own controls and to standardize selection criteria as well as methodology in controlled interlaboratory comparison studies.

Gylseth et al. (52) demonstrated marked interlaboratory variations in results using different techniques and instruments. Nevertheless, a trend to “high” versus “low” exposure could be provided by most institutions, suggesting that laboratories are reasonably consistent in detecting fiber content in asbestos-exposed individuals. De Vuyst et al. (16) concluded that the reference values (higher limit) for electron microscopic fiber counting in individuals without occupational asbestos exposure was roughly 1-2 × 10⁶ of total amphibole fibers and 0.1 × 10⁶ per gram dry lung tissue for amphibole fibers longer than 5 μm. A consensus proposed that subjects with a high probability of occupational asbestos exposure would be likely to have more than 1,000 asbestos bodies per gram of dry tissue, or for numbers of amphibole asbestos fibers longer than 1 μm exceed 1 × 10⁶/g dry tissue (>0.1 × 10⁶ for fibers longer than 5 μm) (25). Ranges for negative control subjects were not presented and each laboratory was advised to establish its own reference values. There can be little question that analytic laboratories should seek to focus on establishing their own internal controls for background non-disease subjects as well as for those with asbestosis.

Using the lower limit for high probability of occupational exposure, as proposed by the Helsinki criteria (23, 25) is intrinsically ambiguous and generates an overlap between control subjects and patients with occupational asbestos exposure being found in 23% of the reviewed studies (6/26). Thus, the use of structured questionnaires and checklists should be applied to identify and exclude subjects selected for control groups having undergone an occupational asbestos exposure (16, 23, 25). Adding to the complexity of separating control subjects from asbestos-exposed individuals, the decline in use and subsequent banning of asbestos in the western world between 1980 and 2010 has led to reduced occupational exposure resulting in lower lung tissue fiber content in patients with mesothelioma (27). Therefore, also ambient outdoor and indoor asbestos fiber concentrations have changed over the prior decades in developed Countries commensurate with the far more limited mining, manufacturing and end-product use of asbestos containing materials (53). Thus, it is plausible that fiber content in lung tissue from negative controls has also decreased over the last four decades. This trend appears when comparing the studies using SEM. And so controls established in analytic laboratories from past decades may not reflect actual environmental exposure in the post-ban period. Therefore, the authors recommend that analytic laboratories should seek to re-establish or update their control reference ranges, to reflect contemporaneous ambient asbestos exposures. The advent of advanced molecular genetics now allows individuals to undergo comprehensive genomic profiling, something which may be incorporated in controls as well as proposed disease cases, including mesothelioma, lung cancer and lung fibrosis. Future case control studies may determine that those subjects with inherited genetic syndromes (heterozygous carriers of pathogenic variant germline mutations), or those who harbor specific genetic polymorphisms in pro-inflammatory genes (TNF-a) or detoxification (GSTM1, GSTT1) have a modulated response to inhaled particulates which is presently not known or accounted for in published control studies.

It is also noted that some laboratories applied more rigorous multi-step inclusion criteria for their control selection, and these will be far less likely to be impacted by the passage of time. The characterization of appropriate controls for background subjects without disease has two clear implications—first in scientific risk-assessment exercises for the consideration of threshold models of exposure which have not been implicated in determining disease; second, in medicolegal settings, to appropriately benchmark an individual case exposure against a relevant control population. Establishing control reference ranges is not without considerable challenge given the difficulty in obtaining appropriate tissue samples. There is a recognized balance from use of surgical resections versus postmortem tissue. The former will inevitably result in more limited tissue availability, although comes with the advantage of potential first-hand characterization of exposures from the patient. The latter allows for widespread “pooled” biologic sample analysis, and more representative mineral content characterization. With respect to optimal tissue selection, the authors recommend the use of “pooled” lung tissue samples which incorporate different anatomic tissue locations and allow for the most representative results of an individual’s prior exposure as already outlined by the guidelines (16, 24). The use of micro-dissected pleural strips, anthracotic lymph nodes, omentum, mesenterial or tumor-containing tissue is not advocated for routine use. The review by Caraballo-Arias and co-authors (54) on pleural tissue showed that case numbers in general are small and controls without asbestos-related disease often missing. The study did not make clear, however, that analytic epidemiology correlates asbestos-related diseases with retained asbestos fibers in pleura. The authors mention several possible complications occurring in analyzing pleura and stress the need for further research. The individual preference is laboratory based although it is clearly preferential to establish controls in tissue most commonly handled by an individual laboratory for test case analyses to maintain a consistency in analytic handling.

Asbestos content of control groups may be expressed as ranges of values if individuals are compared to a group. If two groups (e.g., control vs. exposed group) are compared generally mean and standard deviation are used. Because fiber counts represent not-normally distributed data numbers of fibers should rather be expressed as median and quartile in this context. Reporting range as well as median allows for different usage of data.

It is clear that the observed lack of methodologic standardization across analytic laboratories may be enhanced to establish more robust controls by a number of proposals, which include (1) Establishing an International Task-Force Working Group for the Analysis of Biologic Samples which effectively sets out a consensus methodology for sample preparation, fiber counting and result validation in biologic samples; (2) Establish a robust consensus by determining strict inclusion criteria for control cases. To this end, it is recognized selection bias of cases with remote occupational, para-occupational and environmental/endemic may influence reference cases and upper bound exposure limits. The incorporation of exposure questionnaires is useful in this regard to exclude such cases; (3) establish a network of analytic laboratories engaging in inter-laboratory case control testing, specimen exchange and same sample analysis, with peer-review examination of the variability of results, and proposal and adoption of an external quality audit program. (4) Future plans may incorporate the integration of comprehensive genomic profiling to research for potential individual susceptibility to inhaled particulates, and to correlate this with extent of disease. (5) Implement a longitudinal study in populations with fully characterized exposure data to a wide array of minerals to examine for potential interactions and effects of habits, e.g., tobacco smoking.

In summary, control background groups hold a profound significance in determining the relevance of exposure and disease. They are essential to all analytical laboratories to achieve valid and reliable results. These groups make it possible to verify the causal relationship between asbestos exposure and the subsequent diseases, to avoid distortions and to evaluate the effectiveness of prevention and treatment options. To optimize identification of negative control subjects we propose to use the following criteria: (1) exclusion of occupational, para-occupational, and environmental history of asbestos exposure using structured questionnaires, (2) in autopsy cases death should not be related to asbestos, (3) absence of pleural plaques, and (4) absence of asbestos bodies in sections by light microscopy. Additionally, if possible, controls should be updated periodically, maybe each decade. In addition, control subjects should be matched according to age and gender. To improve comparison of studies, fiber dimensions (e.g., length > 5 μm and width < 0.25 μm) should be specified and fiber counts expressed as median and range.

The analysis of fiber burden in the lung tissue of control subjects, who have been exposed to ambient air, serves as a baseline for understanding environmental exposure to airborne fibers. This baseline is critical for comparing lung fiber burden in occupationally exposed individuals or those with specific environmental exposures. Given the potential long-term consequences of asbestos exposure, it is crucial that future studies in this area continue to use strictly controlled methods and negative controls to better understand and mitigate the health risks of asbestos exposure.

Mineral fiber analysis represents the best arbiter of determining an individual’s prior exposures to respirable asbestos fibers which correlate with disease. An appreciation of controls in analytical laboratories allows researchers to contextualize the significance of fiber burdens and dimensions in individual cases. Due to large differences in geological environments, degree of urbanization and industrialization between populations as well as different methods used in various laboratories control groups should be established for each laboratory within the local setting and using local methods. All analytic laboratories are required to establish their own control reference ranges for subjects without disease or occupational exposure, as well as those with disease – asbestosis, in the setting of the asbestosis range for lung cancer attribution and for the establishment of asbestosis. Ambient asbestos fiber concentrations are subject to change with time, especially in Countries in which asbestos is banned. Accordingly control reference populations should be periodically updated to reflect these lowering exposures so that controls are contemporaneous to periods when testing is conducted in cases. It would be wise to report median and range for asbestos body and fiber content as it allows a range of applications. All analytic laboratories evaluating asbestos fiber content in biologic samples require intra- and interlaboratory standardization to seek to address the issues of quality, validation and accurate mineralogic characterization.