Introduction to Human Biomonitoring

LISBETH E. KNUDSEN, NANNA HUNDEBØLL AND DOMENICO FRANCO MERLO

1.1 Definitions

Human biomonitoring (HBM) is a methodology aimed at assessing human exposure to environmental agents that are capable of inducing adverse health effects in exposed subjects. HBM can measure the agents (i.e., chemicals), their metabolites, or reaction products in human tissues or specimens, such as blood, urine, hair, adipose tissue and teeth. It relies on the use of biomarkers, measurable indicators of changes or events in biological systems. The main advantage of using biomarkers is intrinsic to their nature, because they represent an integrative measurement of exposure to a given agent (i.e., the internal dose) that results from complex pathways of human exposure and also incorporates toxicokinetic information and individual characteristics such as genetically based susceptibility.

HBM is a growing discipline used for exposure and risk assessment in environmental and occupational health. For example, measurements of blood lead concentrations, a biomarker of exposure, were used for risk management in industries with high lead exposures. Workers with high blood lead concentrations, above the recommended safety levels, were transferred to less polluted work tasks, and could only return to their original tasks when the blood lead levels had declined. Blood lead levels have also been monitored in children and clear associations were reported with environmental exposure, with increased levels in schoolchildren living in the vicinity of roads during the leaded fuel period and a decline after the removal of lead from petrol. The relevance of biomonitoring in children exposed to lead is supported by the well known association with adverse neurological effects. According to the HBM results, allowed exposure values for lead have steadily declined, and nowadays the recommended maximum exposure level for children of 100 µg/L should be lowered further, based on recent findings of adverse neurobehavioral effects.

Biomarkers were defined by the World Health Organization (WHO) in 1993 in relation to risk assessment, where the term "biomarker" is used in a broad sense to include almost any measurement reflecting an interaction between a biological system and an environmental agent, which may be chemical, physical or biological. Three classes of biomarker are identified:

• biomarker of exposure: an exogenous substance or its metabolite, or the product of an interaction between a xenobiotic agent and some target molecule or cell that is measured in a compartment within an organism;

• biomarker of effect: a measurable biochemical, physiological, behavioral or other alteration within an organism that, depending upon the magnitude, can be recognized as associated with an established or possible health impairment or disease;

• biomarker of susceptibility: an indicator of an inherent or acquired ability of an organism to respond to the challenge of exposure to a specific xenobiotic substance.

Since 1993, when this definition reflected discussions in environmental and occupational health, the biomarker concept has been introduced into a number of other fields such as forensic medicine, clinical surveillance and drug development.

Figure 1.1 depicts the potential usefulness of biomarkers within the investigation of the continuum between human exposure to environmental agents and the occurrence of health effects, including early biological effects, starting from measurements providing information about exposures (biomarkers of exposure) and with biomarkers of early effects (e.g. chromosomal damage) that may predict the occurrence of delayed adverse effects. In Table 1.1, a number of different biomarkers are reported with examples of their possible application.

Human biomonitoring can be performed on an array of human media ranging from urine to semen. Particular attention must be paid to the use of non-invasive sampling especially when the study population includes children. Table 1.2 lists the different media to be considered.

1.2 HBM: Increasingly used as a Tool in Environmental Health and Medicine

Today much attention is paid to biomarkers predictive of diseases, and consequently the development of biomarkers in disease diagnostics and medicine evaluation has increased. Thus, the pharmaceutical industry aims to develop biomarkers alongside the development of new medicines. Figure 1.2 shows comparable levels of biomarkers in environmental health and diagnostics within the exposure–adverse health effect continuum. The different steps in the process of developing a biomarker for human application are comparable for environmental health and drug development, as shown in Table 1.3.

In environmental health, a number of studies have been performed with newborns, children, and adults with classical biomarkers of exposure as well as promising markers of effect and new techniques of 'omics'. Several European Union (EU) financed projects have developed and validated human biomarkers, such as the PHIME (Public Health Impact of long-term, low-level Mixed element Exposure in susceptible population strata), Integrated Project, Newgeneris (Newborns and Genotoxic exposure risks: Development and application of biomarkers of dietary exposure to genotoxic and immunotoxic chemicals and of biomarkers of early effects, using mother–child birth cohorts and biobanks) program, ECNIS (Environmental Cancer Risk, Nutrition and Individual Susceptibility) network of excellence, and the upcoming European pilot program planned in the ESBIO (European Human Biomonitoring effort).

This book includes specific chapters that describe the ongoing HBM activities in Canada, France, Flanders, Germany, India, and Romania, and HBM activities are also ongoing in other countries, for example Austria, the Czech Republic, Poland, Sweden, and the United States. The EU human biomonitoring program of COPHES/DEMOCOPHES is described in a separate chapter. The book includes chapters on measurement of human exposure to phthalates, perfluorinated compounds (PFCs), bisphenol A, brominated flame retardants, lead, polycyclic aromatic hydrocarbons (PAHs), dioxins, mercury, and arsenic with biomarkers of exposure. Chapters describing human biomonitoring of exposures to environmental tobacco smoke (ETS), mycotoxins, physiological stress, hormone activity, oxidative stress, and ionizing radiation are included, as well as chapters on effect biomarkers of hemoglobin adducts, germ cells, micronuclei and of individual susceptibility. A further chapter describes the ethical issues related to human sampling and monitoring.

1.3 Validation

To ensure proper use of biomarkers, validation must be part of the biomarker development process. The EURACHEM guidance Document No. 1 (1993) states that "The validation of standard methods should not be taken for granted – the laboratory should satisfy itself that the degree of validation of a particular method is adequate for its purpose". The methods used in research and development must fulfill at least two of the requirements for routine analysis stated in the guide on accreditation. First, the analytical method should be described clearly and in sufficient detail in order to allow other laboratories to repeat the measurements. Second, precision, range, robustness and other relevant performance parameters are requisite. It is crucial to use methods developed for this purpose, meaning that the method should be validated for the subpopulation studied, in particular with respect to analytical range, limit of detection (LOD), normal biological variation, and other performance parameters.

From a scientific point of view, a valid test (assay) must have acceptable sensitivity, specificity, predictive value, and reliability. In this context, it is relevant to consider laboratory validity and population validity. Laboratory validity depends on the characteristics of the assay (feasibility, reliability, accuracy and precision) and on the biological characteristics of the marker. Laboratory validity implies studies of sensitivity, "traditionally" considered as the minimum level of an analyte that an assay can detect, i.e., the smallest single value that can be distinguished from zero with high confidence. Specificity "traditionally" indicates the ability of an assay to detect a unique analyte from a group of closely related structures. In human biomonitoring studies this implies the identification and statistical analysis of the assay result with due respect to confounding factors, such as sex, age, smoking, medication, and X-ray examinations.

Reproducibility under conditions of routine use must be demonstrated as a prerequisite for the comparability of assay results. This implies repeatability. That is, test of the same specimen must repeatedly give the same result whether performed by several different laboratories (inter laboratory) or by the same laboratory on several occasions (intra laboratory). This requires the strict following of standardized protocols for the processing of material, analysis, scoring and data processing. Along with standardization of techniques laboratory quality assurance and quality control are essential. The feasibility of the assay is another prerequisite, and is provided by scientific and technical skill, housing, technical equipment and data management.

Population validity refers to how well an assay (e.g., a biomarker) depicts an event in a population. The sensitivity of a test is a measure of how accurately the test identifies people with the disease or abnormality when in fact they do have the disease. Those who have the disease or abnormality and are correctly identified by a positive test are classified as true positives. Specificity is a measure of how accurately the test identifies people who do not have the disease or abnormality when they do not have the disease. Those persons correctly identified by a negative test are classified as true negatives. Any test carries a probability of falsely identifying subjects: false positive and false negatives. The predictive value of a test is the test's accuracy in avoiding either false positive or false negative results. The predictive value of a test cannot be estimated unless one knows the frequency of the disease or abnormality in the population to be studied – the prevalence. Some markers, despite high test sensitivity and specificity, will have low predictive value if the prevalence of the condition in the study population is low.

Table 1.4, from ECETOC, illustrates the importance of valid analytical methods combined with knowledge of toxicokinetics and health effects for the use of HBM data in risk assessment. Thus, the need for proper epidemiological studies including biomarkers is stressed by, for example, organizations such as the International Agency for Research on Cancer (IARC) and others in Europe.

Finally, biological plausibility is essential in the process of validation of biomarkers. It is often derived from experimental studies with animals and/or in vitro systems to obtain "proof of evidence" of the association with a given exposure. Such validation is recommended to comply with the modular approach developed by the European Centre for the Validation of Alternative Methods (ECVAM).

Background levels of persistent organic pollutants (POPs) in humans have declined to the point that their measurement has become increasingly difficult, owing to the detection limits being too high. More and more compounds are present at levels below the limits of detection (LODs) established using quality assurance/quality control (QA/QC) criteria. Increasing the sensitivity of the state-of-the-art gas chromatography coupled to isotope dilution magnetic sector high resolution mass spectrometry (GC-IDHRMS) is promising for POP measurements. Some of the challenges related to sample preparation, blank levels, and to the fact that such a limited number of molecules (e.g., fewer than 600 000 TCDD molecules) is being measured would be extremely valuable for consideration in biomonitoring studies in which sample volume is restricted.

1.4 Statistical Power or Sample Size

A critical issue with human biomonitoring research is the low statistical power to detect biologically significant associations/changes given the high degree of variability (individual and intra-laboratory) affecting the measured biomarkers. Any study can miss a real change/difference (i.e., a real effect) because of the high variability of the data or because of a small sample size. This possibility is well known in hypothesis testing in statistics as one of the the two types of error it can produce, and it is called a type II error (or beta error). It leads to a "not significant" result when in fact the effect revealed by the data is a real one and it is simply not statistically significant. There is always a possibility that a study will detect an unreal change or difference (i.e., a false effect) as real. This possibility is called type I error (or alpha error) and leads to the detection of a "significant" effect when in fact the effect revealed by the data is due to chance. Since in hypothesis testing there is no absolute proof, it is evident that a standard (i.e., an error value) has to be set and accepted for rejecting the null hypothesis. The standard is the probability that the observed effect (change, difference) is due to random variability rather than the factor being investigated. The standard is generally set at 5% (p = 0.05) and is the alpha error. This is why an effect is considered statistically significant when the probability associated with a test statistics is <0.05. Figure 1.3 shows the relationships between these two types of error and the null hypothesis that is declared a priori in a study protocol.

As shown in Figure 1.3, the (statistical) power of a study, defined as 1 - β, is the probability of detecting an effect when the effect is real. The sample size required to ensure adequate statistical power (1 - β) at a given level of significance (α) should always be computed to plan research that has a good chance of detecting a difference that is considered biologically relevant for the study hypothesis and the statistical testing method. Since the specification of the effect (also known as effect size) in the study protocol identifies the degree of deviation considered to be important enough (clinically or biologically), it is clear that any detected effect will be considered negligible when it is smaller than that considered of biological relevance. Obviously, such a judgement cannot be subjective and must rely on the accepted probability of falsely accepting the lack of effect. It is clear that power calculation (or the sample size required given a desired power) should account for the known (or assumed) biological and technical variance as well as the variation between subjects (i.e., background level of the biomarkers), and the difference or change that is considered of biological relevance. By computing the study sample size required to test the study hypothesis properly, a researcher accepts the two above-mentioned errors that are known to occur in any research and sets the probability of detecting a difference in the study sample when this is truly present (generally a statistical power >80% or >90% is considered acceptable). The implication behind sample size calculation is that, because studies have low power to find small differences and high power to find large differences, the researcher must define a priori what effect size is of biological relevance in the context of the study and set the level of confidence for believing a negative finding. A practical consideration is that researchers can avoid reaching the conclusion that more subjects are needed to reach a firm conclusion (i.e., the study needs to be repeated with more subjects).