The reference values in the interpretation of toxicological data.

The worldwide gradual expansion of industrialization has led to a dramatic increase in the production and use of chemical substances. This has resulted in a greater dispersion of these elements in the environment and in an increased exposure of the general population and workers. In this scenario, a thorough knowledge of exposure levels is needed in order to assess chemical risks in environmental and occupational settings. Biological monitoring is among the most useful tools for assessing exposure. However, in order to provide really effective guidance in the application/implementation of risk management measures, biomonitoring results need to be compared with appropriate references. Reference values (RVs) are an excellent resource since useful information for a correct interpretation of toxicological data can be obtained by comparing them with biomonitoring results. In the field of public health, this may enable us to identify potential sources of exposure, define the principal and most frequently exploited routes of exposure, and outline chemical absorption. Similarly, in occupational medicine, RVs can be used to give meaning to biomonitoring findings, especially when a biological limit value is not available for the chemical in question. Furthermore, these values are a valid tool for assessing exposure to chemical carcinogens. Therefore, by integrating reference values in an appropriate and complete system of guide values that also includes action levels and biological limit values, we could obtain both an adequate assessment of exposure and a better understanding of toxicological data.

Introduction

Exposure to chemical substances in workplaces where they are produced, used or handled is one of the most important occupational risk factors that must be properly evaluated and managed within the Occupational Health and Safety systems (59). The evaluation and management of this particular type of occupational risk, both from a qualitative and a quantitative point of view, is extremely important since it has been estimated that more than 80,000 chemicals are currently used, manipulated and/or produced in a variety of industrial sectors and that chemical hazards and toxic substances may cause a wide range of adverse health effects (e.g. irritation, sensitization, carcinogenicity) (35). The “gold standard” for evaluating the risk of exposure to chemical substances is the risk assessment/risk management framework established by the U.S. National Academy of Sciences and described in the U.S. National Research Council report “Risk Assessment in the Federal Government: Managing the Process” (38).

According to this framework, the risk assessment process should be based on four critical steps (figure 1) that include: (i) hazard identification (i.e. the adverse effects a substance has an inherent capacity to cause are defined), (ii) dose-response assessment (i.e. the way potential adverse effects are related to doses are ascertained; critical doses for target organs or cells are determined; possible molecular mechanisms of toxic action are identified), (iii) exposure assessment (i.e. estimation of whether an occupational exposure may or is likely to occur; development and validation of appropriate sampling strategies and analytical methodologies) and (iv) risk characterization (i.e. assessment of the incidence and severity of the adverse effects likely to occur in an exposed human population) (38).

Figure 1

Different steps of the risk assessment and management process proposed by the U.S. National Academy of Sciences (38)

Figura 1 - Le differenti fasi del processo di valutazione e gestione del rischio proposto dalla U.S. National Academy of Sciences (38)

Biological and environmental monitoring are essential tools for the assessment of exposure to toxic agents in both the general and the occupational environment (58). The most complete and thorough definition of biological monitoring was formulated in 1980 during a seminar held in Luxembourg and sponsored jointly by the European Economic Community, the U.S. National Institute for Occupational Safety and Health and the U.S. Occupational Safety and Health Administration (11). On this occasion, biological monitoring was defined as “the measurement and assessment of agents or their metabolites either in tissues, secreta, excreta, expired air or any combination of these to evaluate exposure and health risk compared to an appropriate reference” (11). Therefore, compared to environmental monitoring, in exposure assessment biological monitoring yields important and complementary information since, by measuring the internal dose of a chemical resulting from all sources of exposure, it provides an index of the overall absorption of a xenobiotic (48). Data regarding internal doses, when dose-response relationships have been defined, are extremely important for the prevention of diseases potentially caused by exposure to chemical substances as they are more closely related to the induction of adverse health effects than external concentrations assessed by the environmental monitoring (30).

However, as is clear from the aforementioned definition of biological monitoring a comparison with appropriate guide values must be made in order to obtain a correct interpretation of results. In the field of occupational medicine this comparison is usually carried out using limit values such as for example the biological exposure indices (BEIs) provided by the American Conference of Governmental Industrial Hygienists (ACGIH) (1). However, the application of limit values concerns an area of interest in which exceeding these limits is strongly correlated with the induction of adverse health effects (6). When interpreting the results of biological monitoring, important limitations in the use of limit values should be taken into consideration. First of all, biological limit values are not available for all chemical substances used in workplace (24, 31, 37); secondly, due to scientific progress and the advent of new technologies, work cycles and processes have witnessed the introduction of new chemical substances (e.g. engineered nanomaterials) for which limit values have not yet been formulated (49, 50).

For all these reasons, there is growing interest in the possible use of reference values (RVs) in the field of occupational medicine (2, 4, 6, 8, 12, 51). The term “reference value” refers to the biological indicator levels of occupational and environmental xenobiotics (or of their metabolites) in the general, non-occupationally exposed population (2, 7), and could be defined as the concentration of a xenobiotic (or of its transformation product) measured in biological matrices in reference population groups selected according to predefined criteria (54). Therefore, in principle RVs are useful for assessing whether biomonitoring findings are much higher than would normally be expected (31). RVs might therefore be a valuable tool for identifying new chemical exposures, for studying changes in exposure during periodic workplace measurements and for determining the distribution of a chemical in a number of occupational groups or workplaces (37). In point of fact, RVs are useful not only in the broader public health field where they are frequently used to investigate the relationship between humans and their life environments (especially when identifying sources of exposure, determining principal exposure routes and evaluating xenobiotic absorption), but also in occupational medicine applications (8).

Therefore, considering the importance of RVs and their potentially numerous uses in the field of occupational and environmental medicine, this study aims to provide a comprehensive and thorough analysis regarding the definition and the application of RVs. This information could be valuable for obtaining operative indications into how RVs could be used to correctly interpret the toxicological results of biomonitoring, especially in the absence of other guide values.

Critical issues in defining reference values

In recent years, numerous biomonitoring studies have been performed in several countries in order to improve our knowledge of human exposure to chemical substances and, at the same time, provide lawmakers with useful indications for reducing (or at least controlling) environmental levels of toxicants and for implementing adequate control policies. Indeed, human biomonitoring programmes have been developed in various European countries (46) such as Belgium (31, 47), the Czech Republic (10, 18), France (29), Germany (51, 52), Italy (3, 7-9), Spain (40) and the United Kingdom (12, 37) and similar studies have also been conducted in non-European countries such as Brazil (28), Canada (45), the Republic of Korea (34) and the United States (17). The recent proliferation of such studies arises from the need to establish appropriate RVs nationwide (or in some cases at a regional or local level) as the latter can be strongly influenced by geographic, industrial, dietary, environmental and lifestyle factors (31, 37). Nevertheless, despite local differences of this kind, to obtain adequate RVs, a standardized procedure should be used that takes into account critical issues (table 1) such as sample selection, sample size, time period, exclusion criteria, partitioning criteria, and the analytical quality achieved (45, 60).

Table 1

Critical aspects in defining reference values (45)

Tabella 1 - Aspetti critici nella definizione dei valori di riferimento (45)

Issue	Recommendation
Sample selection	A priori selection is desirable; Apply an adequate system of exclusion and partitioning criteria in the case of a posteriori selection; Exclusion and partitioning criteria can not be generalized and must be established from time to time taking into consideration all the possible confounding factors.
Sample size	The sample size which constitutes the reference population should be sufficiently large to include different subgroups for the main sociodemographic variables: Age; Sex; Race; Ethnicity. IFCC recommends a sample size of at least 120 subjects.
Time period	The reference values should be defined on the analysis of the most recent data available.
Exclusion criteria	Exclusion criteria should be determined before selecting reference population; The variables able to influence biomarker levels and that cannot be ascribed to the general characteristics of the reference population should be classified as exclusion criteria: Smoking habits; Recent surgery; Administration of drugs; Fasting.
Partitioning criteria	When significantly different biomarker levels are observed in different subgroups (related for example to sex, age, race) of the reference population, then specific reference values should be established for each subgroup.
Analytical methods	Highest quality analytical methods should be used: Specificity; Sensitivity. Quality control procedures should be developed and applied to: Sample collection and storage; Preparation of the sample for the analysis; Instrumental analysis.

To this end, the International Federation of Clinical Chemistry and Laboratory Medicine (IFCC) and the International Union of Pure and Applied Chemistry (IUPAC) have introduced the concept of reference intervals and developed statistical methodologies necessary to define chemicals background exposure in a specific reference population (41, 45, 57). In essence, the IFCC and IUPAC documents offer a useful guidance on RVs by providing practical advice and recommendations concerning the definition of the reference population (table 1) (45).

First of all, the sample selection should be an a priori selection although, when this kind of recruitment is not feasible, a posteriori selection can be used. However, in the latter case, strict partitioning and exclusion criteria should be applied so as to take into account specific biomarkers and confounding factors (41, 45, 57). Another important issue concerns sample size which, to be effectively valid, should be wide-ranging enough to include different age, race or ethnicity groups (57). With regard to the time period, the most up-to-date biomonitoring information available should be used since environmental exposure to chemical substances in the general population can vary considerably as a function of time (41). Moreover, exclusion criteria should be adopted in order to avoid including in the sample subjects (e.g. smokers) who present characteristics that do not belong to the general population (45). Similarly, when we expect to find significant differences in various groups of the general population (e.g. in age or race), partitioning criteria should be used to develop specific RVs for these subgroups (45). Lastly, high quality analytical methods (especially in terms of specificity and sensitivity) and control procedures (collection and treatment of the samples) should be adopted (41, 45, 57).

The Italian Society for Reference Values: the path towards reference values in Italy

For more than twenty-five years, the Italian Society for Reference Values (SIVR) has been actively involved in the definition and dissemination of RVs for both environmental and occupational xenobiotics (8). Its main aim is to develop technical guidelines by establishing RVs for various elements and chemical compounds (or related products of transformation). Since 1993, the number of RVs formulated by the SIVR has steadily increased from a dozen in the early years of its activity to 135 currently included in its latest list (55).

Figure 2 schematically illustrates the 8-step strategy adopted by the SIVR to process and establish reliable RVs. Once a chemical substance (or the group of chemical substances) that requires the establishment of RV has been identified, a systematic review of the literature is performed in order to ascertain whether similar RVs for the substance in question have already been produced in other countries (8, 54). The next three phases (from 3 to 5) focus on the choice of the best analytical method for determining analyte levels in the biological matrices. The available analytical techniques are evaluated in order to find a fair compromise between the need to have a high-quality analytical method and adequate instruments in the laboratories that will potentially carry out the analyses (8). After taking into consideration all the pre-analytical factors and selecting the method of analysis, a network of laboratories must be created that will have the practical task to carry out analyses and thereby obtaining concentrations of the chemical (or of its metabolite). During this phase, it is vital that the choice of laboratories contributing to the definition of the RVs be based on appropriate inter- and intra-laboratory data quality control (8).

Figure 2

The Italian Society for Reference Values methodology for defining RVs

Figura 2 - La metodologia SIVR per la definizione dei valori di riferimento

When the biological sample is collected, the subjects enrolled in the study should simultaneously be given an ad hoc questionnaire since the information retrieved by means of this tool is of utmost importance for a correct interpretation of the biomonitoring data. In order to define RVs adequately, the questionnaire must contain items that are able to reveal exposure to the substance under investigation that come from sources not common to the general population, such as smoking or occupational exposure (exclusion criteria) (8, 41, 45, 54, 57). Furthermore, to be sufficiently informative, the questionnaire should also investigate all the possible factors (e.g. age, gender, physical activity, dietary and lifestyle habits, hobbies or “do it yourself” (DIY) activities, urban or rural residence, presence of anthropogenic or environmental emissions in proximity to home and/or work) that could play a role in influencing or altering the internal dose levels of the chemical. Subsequently, in accordance with the main IFCC and IUPAC indications, the reference population is identified a priori, by taking into account the principal socio-demographic and geographical variables, and the sampling protocol, which should ensure the recruitment of a truly representative sample, is established (8). Finally, step 8 involves a statistical analysis of the data. If possible, the sample is stratified into subsamples according to exposure characteristics, in order to minimize the data variability. Subsequently, Kolmogorov-Smirnov (or alternative testing for normal distribution of unaltered and log-transformed data) is used and outliers are identified and evaluated (8). Therefore, the amount of data that is below the limit of detection/limit of quantification is defined and the data is described in terms of min-max, percentiles, index of central tendency and variability of data (8). Finally, to identify possible variables that might affect the data, the analysis of variance (ANOVA) and/or multiple regression analysis is performed.

In addition to RVs produced according to such SIVR procedure, the Society proposes also tentative RVs (TRVs) which are the result of a preliminary phase of definition of SIVR RVs. TRVs may be motivated by the unavailability of a RV for a particular analyte and/or the lack of information regarding possible confounding factors. They are generally defined by specific experimentation between a small number of labs (at least two) in the SIVR circuit using different analytical methods or, alternatively, by a single SIVR lab if it uses a validated analytical method with estimated uncertainty. SIVR includes also RVs based on scientific literature (LRVs), which are defined according to a critical revision of publications, produced in the last 10 years and preferentially investigating Italian or European populations (8).

Reference values in the context of environmental medicine and public health

Over recent decades, the number of chemical substances produced commercially has increased exponentially because of a worldwide expansion in industrialization (13). However, little is known regarding the quantity of xenobiotics produced (and their related diffusion in environmental matrices). A fairly representative estimate, that provides a good example of the huge dimensions of the problem, can be made by consulting the Chemical Abstracts Service (CAS) RegistrySM that currently contains more than 150 million chemicals (16). Commercial chemical production is responsible for the appearance of approximately 200,000 new chemicals per week, but many substances are also emitted into the environment as a result of natural processes (e.g. volcanic eruptions, large forest fires, sea spray aerosols) or as by-products of fossil fuel combustion or industrial processes (39, 43). Furthermore, in some cases, contamination of environmental matrices is of geological origin, e.g. arsenic contamination in groundwater, a well-known public health issue in several countries such as Hungary, Italy, Romania, Argentina, Chile, Mexico, Afghanistan, Bangladesh, Nepal and Pakistan (21, 44, 56).

Regardless of the source of emission, a growing body of evidence suggests that environmental chemical levels (in outdoor and indoor air, soil, water and foods) have increased significantly in recent years, thereby contributing to a greater likelihood of exposure for the general population (13, 33). These substances, which can come into contact with or enter the body via three major routes: ingestion, inhalation and dermal absorption, are potentially hazardous for human health. Everyday sources of exposure include contaminated food and/or drinking water, airborne pollutants, consumer products, soil and household dust (figure 3) (14, 25).

Figure 3

Possible environmental sources of exposure to chemical substances for the general population

Figure 3 - Possibili fonti di esposizione ambientale a sostanze chimiche per la popolazione generale

For a number of chemicals that are essential but also toxic elements, or for various substances that have a physiological role (e.g. chromium, manganese, nickel), RVs can be considered the result of physiological-homeostatic processes and general environmental exposure (6, 8). On the other hand, the presence of measurable biological levels of toxic xenobiotics (which should normally not be present in the biological matrices of the general population) is proof of their absorption following environmental exposure (6, 8). In both cases, the contribution of environmental exposure to chemical concentrations in the human body is extremely important, and for this reason, RVs represent a powerful tool for investigating the complex interplay between people and the environment in which they live.

RVs can provide interesting information for obtaining a deeper insight into the significance of toxicological data reported in human biomonitoring studies. First of all, appropriate RVs are essential for identifying subjects (or subpopulations) that have been seriously exposed to a specific substance and might consequently need targeted prevention or protection measures and intervention (27, 32, 36, 45). However, from a public health point of view, although it is certainly necessary to identify highly-exposed individuals, that alone is not sufficient to protect people from hazardous chemicals. It is also vital to determine the exposure sources and factors that made the biological concentrations in these subjects exceed background levels. An analysis of the findings obtained from questionnaires (that are an integral part of biomonitoring studies aimed at establishing RVs) can provide very useful information on this subject by highlighting the exposure characteristics shared by subjects with higher exposure levels (32, 45). In addition, RVs can also be helpful in defining new statistical associations between a specific health outcome observed in a particular subpopulation and the related (increased) level of exposure (32, 36, 45). As mentioned previously, the presence of biological concentrations of chemical substances above RVs calls for special corrective action to be taken in order to reduce suspected environmental exposure. RVs could also be employed periodically to evaluate the effectiveness of such intervention and provide information on temporal and geographical trends or changes in exposure.

Reference values in the context of occupational medicine

Biological monitoring is a key component in occupational risk assessment and the management of chemical substances, since evaluation of the internal dose of chemicals (or related metabolites) in human biological samples obtained from all potential sources and exposure routes can provide an accurate estimate of the body burden of these substances (30, 48). The integration of information obtained from biological monitoring and data derived from environmental monitoring is the safest and most effective way of adequately assessing workers’ occupational exposure to chemicals and also of implementing, when necessary, appropriate countermeasures to reduce or at least control exposure levels. Therefore, in workplaces, an analysis of these data on the part of occupational health and safety professionals ensures that highly-exposed worker groups will be detected, or that possible differences in exposure in these groups will be assessed. Furthermore, this analysis would make it possible to monitor whether and in what way chemical exposure changes over time, and to verify the effectiveness of the control measures undertaken (36).

However, biomonitoring results alone are of little value since their correct interpretation necessarily requires comparison with appropriate references (11). The need to have guide values for comparison is widely recognized also at a legislative level, since the Chemical Agents EC Directive 98/24 (concerning worker health and safety protection from chemical risks) provides a basis for establishing biological limit values (19). Nevertheless, in the European Union, there is currently only one binding biological limit value that the aforementioned Directive set in reference to blood lead levels. Similarly, there is also a significant lack of biological guide values concerning industrial chemical legislation. The Registration, Evaluation, Authorisation and Restriction of Chemicals (REACH) regulation (42) has developed, for workplace air, the Derived No-Effect Level (DNEL) defined as the level of a substance above which a human should not be exposed, and under which no adverse effects are expected. Interestingly, DNELs may be expressed as internal exposure biomarker values (DNELbiomarker). These apply to substances for which internal exposure data are available and have been reliably associated with effects. However, REACH lacks specific fields that aim to understand the relationship between internal exposure biomarkers of a substance and resulting effects as DNELbiomarkers are not commonly included in Registration dossiers (26, 36). In this context, the recommended biological limit values or biological guidance values proposed for several chemicals by the Scientific Committee on Occupational Exposure Limits (SCOEL) can be used. From 1995 to 2018 this Committee assisted the European Commission in making a scientific evaluation of the relationship between the levels of occupational exposure to hazardous chemical agents and health effects (53). SCOEL biological limit values were health-based limit values that referred to exposure levels below which adverse health effects were unlikely to occur (36). Differently, biological guidance values are the upper concentration of a substance (or a metabolite of the substance) in any appropriate biological medium corresponding to a certain percentile (generally 90 or 95 percentile) in a defined reference population (53). Unfortunately, the number of chemicals for which the SCOEL set biological limit values or biological guidance values is very limited (53). Since 2019, the European Commission has handed over the responsibilities of SCOEL to the Risk Assessment Committee (RAC) of the European Chemicals Agency (ECHA).

Clearly, when biological limit values have not been established by law, it is possible to refer to values proposed by foreign governmental agencies or approved prestigious and reliable non-governmental agencies (e.g. ACGIH, Deutsche Forschungsgemeinschaft’s (DFG), Commission for the Investigation of Health Hazards of Chemical Compounds in the Work Area) (1, 22). However, even in this case, compared to the enormous quantity of chemicals that are commonly produced, used or handled in the workplace, the availability of BEIs or Biological Tolerance Values for Working Materials is still rather limited (in both cases fewer than 100 biological limit values) (1, 22).

Consequently, in the absence of biological limit values, RVs can clearly be an excellent resource since they can be used as guidance for interpreting biological monitoring data (6, 37). In other words, the upper limit (95 percentile) of the range observed in the reference population (non-exposed subjects) can act as a reference to be compared with biomonitoring data so as to gain insight into workers’ exposure (31). The detection of biomarker levels significantly above that upper limit would suggest occupational exposure to a chemical agent. However, it is important to emphasise that RV use should not be limited exclusively to cases of biomonitoring interpretation in which there are no biological limit values. Even if these values are not health-based (as they do not take into account toxicological information on biomarkers), and cannot be exploited to evaluate health risks or considered as threshold levels for undertaking clinical action (45), their combined use with biological action levels and limit values, as part of an integrated guide values system, could contribute significantly to improving the assessment and management of chemical risk (6).

For example, when using biological limit values, it is important to keep in mind that for many chemicals these values also include the background levels of the determinant that can usually be detected in subjects not-occupationally exposed to the substance in question (6). Furthermore, some biomarkers used in biological monitoring are not specific for the chemical under observation and their concentrations in biological matrices may be strongly affected by non-occupational exposure to other agents (6). In these cases, the co-ordinated and simultaneous use of both RVs and limit values guarantees a much more accurate interpretation of the toxicological data, which, in turn, leads to a more thorough assessment of exposure and its related health risks. Therefore, in order to make an assessment of biomonitoring data as complete and detailed as possible, RVs should be included, even when adequate biological limit values are available. In fact, the different types of guide values (i.e. RVs, action levels and limit values) should not be considered separately as they are conceptually related to each other and refer to specific areas of interest that often tend to overlap (figure 4) (6).

Figure 4

Integrated and harmonized guide values system including reference values, action levels and biological limit values

Figure 4 - Sistema integrato di valori guida che comprende i valori di riferimento, i livelli di azione ed i valori limite biologici

In occupational medicine, exposure to carcinogenic chemicals is another interesting field in which RVs are applied. The EC Directive 2004/37 on the protection of workers from risks related to exposure to carcinogens or mutagens at work states (Annex II) that the health surveillance system for workers exposed to these substances should include, where appropriate, biological surveillance (20). However, this Directive fails to indicate biological limit values for any substance. Therefore, in cases of occupational exposure to carcinogens, RVs could be used to interpret the toxicological data in order to identify individuals with exposure levels above RVs and correlate these levels with possible adverse health effects. For example, in epidemiological studies, RVs that express the amount of carcinogenic agents present in the environment and then indicate the exposure levels of the general population could be used to calculate the potential excess of cancer mortality in groups occupationally-exposed to these substances (6). Furthermore, for some carcinogens, whose adverse effects can also be caused by exposure to low doses due to a genotoxic type of action mechanism, defining biological limit values has little meaning in terms of prevention (5, 15, 54). For other chemical substances in which epigenetic mechanisms are principally responsible for carcinogenic action, it is possible to establish a threshold level below which the occurrence of an adverse effect should not occur (15, 23). In both cases, levels of occupational exposure to these substances should be kept as low as possible. In this context, the RVs that refer to the general, non-occupationally exposed population should indicate the lowest exposure levels and would seem to be vital in controlling, monitoring over time and maintaining worker exposure at the lowest possible levels.

Conclusions

Biomonitoring studies are being used more and more frequently in both the fields of public health and occupational medicine to improve our knowledge of human exposure to chemical substances in both general life environments and workplaces. In fact, health risks caused by chemical exposure in workplaces, environmental matrices, diet, lifestyle habits and consumer products are one of the most important issues in the aforementioned medical research areas. However, the detection of measurable levels of chemicals in human biological matrices does not necessarily imply the onset of adverse health effects. A health risk occurs when a quantitative relationship between a health effect and biomarker level is observed (health-based guide values such as action levels or biological limit values). Although RVs cannot be used to predict any adverse health outcomes, they represent a reference point for interpreting the toxicological data provided by human biomonitoring studies.

Therefore, from a public health point of view, it is necessary to define chemical RVs in a specific reference population so as to identify comparatively high levels of exposure, since these play a significant role in public health management. In fact, when a comparison of biological monitoring findings with RVs indicates that high exposure has occurred, it is important to determine the possible exposure sources and factors involved so that appropriate prevention and/or protection measures can be implemented. RVs are also useful for evaluating temporal and spatial changes in exposure to chemicals; for indicating appropriate policy actions to control and limit chemical exposure of the general population; and for verifying the adequacy and validity of such strategies.

In occupational medicine, RVs are a valuable resource that should be used (together with action levels and biological limit values) in an integrated and co-ordinated system of guide values to ensure a thorough assessment of chemical exposure in workplaces. RVs are especially useful for evaluating the toxicological data of substances for which biological limit values are not available or for which a toxicity threshold has not yet been established. They could therefore indicate and guide the eventual adoption of specific prevention and protection measures in the workplace.

No potential conflict of interest relevant to this article was reported by the authors