INTOX Home Page

This report contains the collective views of international groups of experts and does not necessarily represent the decisions or the stated policy of the United Nations Environment Programme, the International Labour Organization, or the World Health Organization.

Environmental Health Criteria 222

Biomarkers In Risk Assessment:
Validity And Validation

Published under the joint sponsorship of the United Nations Environment Programme, the International Labour Organization, and the World Health Organization, and produced within the framework of the Inter-Organization Programme for the Sound Management of Chemicals.

World Health Organization

Geneva, 2001

The International Programme on Chemical Safety (IPCS), established in 1980, is a joint venture of the United Nations Environment Programme (UNEP), the International Labour Organization (ILO), and the World Health Organization (WHO). The overall objectives of the IPCS are to establish the scientific basis for assessment of the risk to human health and the environment from exposure to chemicals, through international peer-review processes, as a prerequisite for the promotion of chemical safety, and to provide technical assistance in strengthening national capacities for the sound management of chemicals.

The Inter-Organization Programme for the Sound Management of Chemicals (IOMC) was established in 1995 by UNEP, ILO, the Food and Agriculture Organization of the United Nations, WHO, the United Nations Industrial Development Organization, the United Nations Institute for Training and Research, and the Organisation for Economic Co-operation and Development (Participating Organizations), following recommendations made by the 1992 UN Conference on Environment and Development to strengthen cooperation and increase coordination in the field of chemical safety. The purpose of the IOMC is to promote coordination of the policies and activities pursued by the Participating Organizations, jointly or separately, to achieve the sound management of chemicals in relation to human health and the environment.

WHO Library Cataloguing-in-Publication Data

Biomarkers in risk assessment: validity and validation.

(Environmental health criteria ; 222)

1.Biological markers

2.Risk assessment - methods

3.Validation studies

4.Reproducibility of results

5.Environmental monitoring

I.International Programme on Chemical Safety II.Series

ISBN 92 4 157222 1

(NLM Classification: QH 438.4.B55)

ISSN 0250-863X

The World Health Organization welcomes requests for permission to reproduce or translate its publications, in part or in full. Applications and enquiries should be addressed to the Office of Publications, World Health Organization, Geneva, Switzerland, which will be glad to provide the latest information on any changes made to the text, plans for new editions, and reprints and translations already available.

©World Health Organization 2001

Publications of the World Health Organization enjoy copyright protection in accordance with the provisions of Protocol 2 of the Universal Copyright Convention. All rights reserved.

The designations employed and the presentation of the material in this publication do not imply the expression of any opinion whatsoever on the part of the Secretariat of the World Health Organization concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.

The mention of specific companies or of certain manufacturers’ products does not imply that they are endorsed or recommended by the World Health Organization in preference to others of a similar nature that are not mentioned. Errors and omissions excepted, the names of proprietary products are distinguished by initial capital letters.







2.1 Hazard identification

2.2 Dose response

2.3 Exposure assessment for risk assessment



4.1 Exposure biomarkers

4.2 Effect biomarkers

4.3 Susceptibility biomarkers






9.1 General recommendations

9.2 Recommendations for future research

9.2.1 Prevalidation stage

9.2.2 Validation stage

9.3 Application











Every effort has been made to present information in the criteria monographs as accurately as possible without unduly delaying their publication. In the interest of all users of the Environmental Health Criteria monographs, readers are requested to communicate any errors that may have occurred to the Director of the International Programme on Chemical Safety, World Health Organization, Geneva, Switzerland, in order that they may be included in corrigenda.

* * *

A detailed data profile and a legal file can be obtained from the International Register of Potentially Toxic Chemicals, Case postale 356, 1219 Châtelaine, Geneva, Switzerland (telephone no. + 41 22 – 9799111, fax no. + 41 22 – 7973460, E-mail

* * *

This publication was made possible by grant number 5 U01 ES02617-15 from the National Institute of Environmental Health Sciences, National Institutes of Health, USA, and by financial support from the European Commission.

Environmental Health Criteria



In 1973 the WHO Environmental Health Criteria Programme was initiated with the following objectives:


to assess information on the relationship between exposure to environmental pollutants and human health, and to provide guidelines for setting exposure limits;


to identify new or potential pollutants;


to identify gaps in knowledge concerning the health effects of pollutants;


to promote the harmonization of toxicological and epidemiological methods in order to have internationally comparable results.

The first Environmental Health Criteria (EHC) monograph, on mercury, was published in 1976 and since that time an ever-increasing number of assessments of chemicals and of physical effects have been produced. In addition, many EHC monographs have been devoted to evaluating toxicological methodology, e.g. for genetic, neurotoxic, teratogenic and nephrotoxic effects. Other publications have been concerned with epidemiological guidelines, evaluation of short-term tests for carcinogens, biomarkers, effects on the elderly and so forth.

Since its inauguration the EHC Programme has widened its scope, and the importance of environmental effects, in addition to health effects, has been increasingly emphasized in the total evaluation of chemicals.

The original impetus for the Programme came from World Health Assembly resolutions and the recommendations of the 1972 UN Conference on the Human Environment. Subsequently the work became an integral part of the International Programme on Chemical Safety (IPCS), a cooperative programme of UNEP, ILO and WHO. In this manner, with the strong support of the new partners, the importance of occupational health and environmental effects was fully recognized. The EHC monographs have become widely established, used and recognized throughout the world.

The recommendations of the 1992 UN Conference on Environment and Development and the subsequent establishment of the Intergovernmental Forum on Chemical Safety with the priorities for action in the six programme areas of Chapter 19, Agenda 21, all lend further weight to the need for EHC assessments of the risks of chemicals.


The criteria monographs are intended to provide critical reviews on the effect on human health and the environment of chemicals and of combinations of chemicals and physical and biological agents. As such, they include and review studies that are of direct relevance for the evaluation. However, they do not describe every study carried out. Worldwide data are used and are quoted from original studies, not from abstracts or reviews. Both published and unpublished reports are considered and it is incumbent on the authors to assess all the articles cited in the references. Preference is always given to published data. Unpublished data are used only when relevant published data are absent or when they are pivotal to the risk assessment. A detailed policy statement is available that describes the procedures used for unpublished proprietary data so that this information can be used in the evaluation without compromising its confidential nature (WHO (1999) Guidelines for the Preparation of Environmental Health Criteria. PCS/99.9, Geneva, World Health Organization).

In the evaluation of human health risks, sound human data, whenever available, are preferred to animal data. Animal and in vitro studies provide support and are used mainly to supply evidence missing from human studies. It is mandatory that research on human subjects is conducted in full accord with ethical principles, including the provisions of the Helsinki Declaration.

The EHC monographs are intended to assist national and international authorities in making risk assessments and subsequent risk management decisions. They represent a thorough evaluation of risks and are not, in any sense, recommendations for regulation or standard setting. These latter are the exclusive purview of national and regional governments.


The layout of EHC monographs for chemicals is outlined


• Summary – a review of the salient facts and the risk evaluation of the chemical

• Identity – physical and chemical properties, analytical methods

• Sources of exposure

• Environmental transport, distribution and transformation

• Environmental levels and human exposure

• Kinetics and metabolism in laboratory animals and humans

• Effects on laboratory mammals and in vitro test systems

• Effects on humans

• Effects on other organisms in the laboratory and field

• Evaluation of human health risks and effects on the environment

• Conclusions and recommendations for protection of human health and the environment

• Further research

• Previous evaluations by international bodies, e.g. IARC, JECFA, JMPR

Selection of chemicals

Since the inception of the EHC Programme, the IPCS has organized meetings of scientists to establish lists of priority chemicals for subsequent evaluation. Such meetings have been held in Ispra, Italy, 1980; Oxford, United Kingdom, 1984; Berlin, Germany, 1987; and North Carolina, USA, 1995. The selection of chemicals has been based on the following criteria: the existence of scientific evidence that the substance presents a hazard to human health and/or the environment; the possible use, persistence, accumulation or degradation of the substance shows that there may be significant human or environmental exposure; the size and nature of populations at risk (both human and other species) and risks for environment; international concern, i.e. the substance is of major interest to several countries; adequate data on the hazards are available.

If an EHC monograph is proposed for a chemical not on the priority list, the IPCS Secretariat consults with the Cooperating Organizations and all the Participating Institutions before embarking on the preparation of the monograph.


The order of procedures that result in the publication of an EHC monograph is shown in the flow chart on p. x. A designated staff member of IPCS, responsible for the scientific quality of the document, serves as Responsible Officer (RO). The IPCS Editor is responsible for layout and language. The first draft, prepared by consultants or, more usually, staff from an IPCS Participating Institution, is based initially on data provided from the International Register of Potentially Toxic Chemicals, and reference data bases such as Medline and Toxline.

The draft document, when received by the RO, may require an initial review by a small panel of experts to determine its scientific quality and objectivity. Once the RO finds the document acceptable as a first draft, it is distributed, in its unedited form, to well over 150 EHC contact points throughout the world who are asked to comment on its completeness and accuracy and, where necessary, provide additional material. The contact points, usually designated by governments, may be Participating Institutions, IPCS Focal Points, or individual scientists known for their particular expertise. Generally some four months are allowed before the comments are considered by the RO and author(s). A second draft incorporating comments received and approved by the Director, IPCS, is then distributed to Task Group members, who carry out the peer review, at least six weeks before their meeting.

The Task Group members serve as individual scientists, not as representatives of any organization, government or industry. Their function is to evaluate the accuracy, significance and relevance of the information in the document and to assess the health and environmental risks from exposure to the chemical. A summary and recommendations for further research and improved safety aspects are also required. The composition of the Task Group is dictated by the range of expertise required for the subject of the meeting and by the need for a balanced geographical distribution.

EHC Preparation Flow Chart

The three cooperating organizations of the IPCS recognize the important role played by nongovernmental organizations. Representatives from relevant national and international associations may be invited to join the Task Group as observers. Although observers may provide a valuable contribution to the process, they can only speak at the invitation of the Chairperson. Observers do not participate in the final evaluation of the chemical; this is the sole responsibility of the Task Group members. When the Task Group considers it to be appropriate, it may meet in camera.

All individuals who as authors, consultants or advisers participate in the preparation of the EHC monograph must, in addition to serving in their personal capacity as scientists, inform the RO if at any time a conflict of interest, whether actual or potential, could be perceived in their work. They are required to sign a conflict of interest statement. Such a procedure ensures the transparency and probity of the process.

When the Task Group has completed its review and the RO is satisfied as to the scientific correctness and completeness of the document, it then goes for language editing, reference checking and preparation of camera-ready copy. After approval by the Director, IPCS, the monograph is submitted to the WHO Office of Publications for printing. At this time a copy of the final draft is sent to the Chairperson and Rapporteur of the Task Group to check for any errors.

It is accepted that the following criteria should initiate the updating of an EHC monograph: new data are available that would substantially change the evaluation; there is public concern for health or environmental effects of the agent because of greater exposure; an appreciable time period has elapsed since the last evaluation.

All Participating Institutions are informed, through the EHC progress report, of the authors and institutions proposed for the drafting of the documents. A comprehensive file of all comments received on drafts of each EHC monograph is maintained and is available on request. The Chairpersons of Task Groups are briefed before each meeting on their role and responsibility in ensuring that these rules are followed.



Dr D. Anderson, TNO BIBRA International Ltd, Carshalton, Surrey, United Kingdom (Rapporteur)

Dr H. Autrup, Department of Environmental Medicine, University of Aarhus, Aarhus, Denmark (Chairman)

Dr S. Bonassi, Department of Environmental Epidemiology, National Institute for Research on Cancer, Genoa, Italy

Dr K. Hemminki, Department of Biosciences at Novum, Karolinska Institute, Huddinge, Sweden

Dr A. Mutti, Laboratory of Industrial Toxicology, Department of Clinical Medicine, Nephrology, and Health Sciences, University of Parma Medical School, Parma, Italy

Dr O. Pelkonen, Department of Pharmacology and Toxicology, University of Oulu, Oulu, Finland

Dr P.A. Schulte, Education and Information Division, National Institute for Occupational Safety and Health, Cincinnati, Ohio, USA


Dr A. Aitio, International Programme on Chemical Safety, World Health Organization, Geneva, Switzerland (Joint Secretary)

Dr Y. Hayashi, International Programme on Chemical Safety, World Health Organization, Geneva, Switzerland (Joint Secretary)


A WHO Task Group on Environmental Health Criteria for Biomarkers in Risk Assessment: Validity and Validation met at TNO BIBRA International, Carshalton, Surrey, United Kingdom from 3 to 6 April 2000. Dr A. Aitio, IPCS, welcomed the participants on behalf of the IPCS and its three cooperating organizations (UNEP/ILO/WHO). The Task Group reviewed and revised the draft monograph.

This Environmental Health Criteria monograph is composed of the main text and four authored papers. The main text was constructed by Dr P.A. Schulte, based on the source documents and was reviewed by the IPCS Contact Points. The comments received were considered by the principal author, and the revisions were discussed and approved by the Task Group. The source documents were similarly subjected to IPCS review and were then revised accordingly by the authors. However, they were not discussed thoroughly during the Task Group meeting and thus represent the views of the authors.

Dr A. Aitio and Mr Y. Hayashi of the IPCS Central Unit were responsible for the overall scientific content of the monograph and Dr P.G. Jenkins of the IPCS Central Unit was responsible for the technical editing of the monograph.

The efforts of all who helped in the preparation of the monograph are gratefully acknowledged.

* * *

The preparation of the draft was financially supported by the US Environmental Agency. Financial support for this Task Group was provided by the UK Department of Health as part of its contribution to the IPCS.



benchmark dose




cytochrome P450


gas chromatography


glutathione S-transferase


International Agency for Research on Cancer


lowest-observed-adverse-effect level


mass spectroscopy




no-observed-adverse-effect level


polycyclic aromatic hydrocarbon


polymerase chain reaction


poor metabolizer


tetrachlorinated dibenzo-p-dioxin


threshold limit value


time-weighted average


xenobiotic metabolizing enzyme


The aim of risk assessments is to provide society with estimates of the likelihood of illnesses and injury as a consequence of exposure to various hazards. Risk assessments are needed when social policy decisions are in dispute, when the health consequences of alternative policies in question are not subject to direct measurement (at least in a timely fashion), and when the scientific analysis of a hazard is not complete (Hattis & Silver, 1993). The assessment procedure involves the development of an exposure-response curve for the target species (e.g., humans), based on animal and human information, followed by the projection of the curves to estimate levels of exposure that may be considered safe (NRC, 1987). For risk assessments to be useful they should lead to projections that are close to the true risks. A strong scientific basis for conducting risk assessments is the best way to assure that projections are close to true risks or at least provide an honest depiction of the state of knowledge and the degree of certainty about risks (Bailar & Bailer, 1999).

Risk assessment has a range of meanings. At the basic level it is an exercise to evaluate the potential of some hazard to induce an adverse human health response. It can be a qualitative or quantitative exercise at the individual or group (population) level. The term quantitative risk assessment (QRA) has been used to describe the response associated with a specific level of exposure (Bailer & Dankovic, 1997). The availability of adequate dose/concentration-response data is a prerequisite to conducting a QRA.

A biomarker is any substance, structure or process that can be measured in the body or its products and influence or predict the incidence of outcome or disease. Biomarkers can be classified into markers of exposure, effect and susceptibility. If biomarkers are to contribute to environmental and occupational health risk assessments, they have to be relevant and valid. Relevance refers to the appropriateness of biomarkers to provide information on questions of interest and importance to public and environmental health authorities and other decision-makers. The use of relevant biomarkers allows decision-makers to answer important public health questions by being used in research or risk assessments in a way that contributes useful information that cannot be obtained better by other approaches, such as questionnaires, environmental measurements or record reviews. For example, chronic exposure to organochlorines is better indicated by serum organochlorine levels than by market-basket studies or industrial hygiene measurements, and early kidney damage may be better indicated by a battery of urinary biomarkers than by morbidity records. Relevance also pertains to whether the questions on which a biomarker can provide information are important questions; not merely ones that can be answered, but ones that should be answered (Muscat, 1996). Thus, the ability to measure a biomarker after exposure to a toxicant may not be as important a question as whether individuals with exposure to the toxicant are at increased risk of disease.

The second characteristic of potentially useful biomarkers is validity. Validity of biomarkers has been widely discussed (Hernberg & Aitio, 1987; Schatzkin et al., 1990; Schulte & Perera, 1993; Boffetta, 1995; Bernard, 1995; Dor et al., 1999). It includes both laboratory and epidemiological aspects. Validity refers to a range of characteristics that is the best approximation of the truth or falsehood of a biomarker. It is a sense of degree rather than an all-or-none state. The validity of a biomarker is a function of intrinsic qualities of the biomarker and characteristics of the analytic procedures (Dor et al., 1999) (see Tables 1 and 2 for an example of this distinction). Additionally, three broad categories of validity can be distinguished: measurement validity, internal study validity and external validity (Schulte & Perera, 1993). Measurement validity (in terms of analytical chemistry, accuracy) is the degree to which a biomarker indicates what it purports to indicate. Internal study validity is the degree to which inferences drawn from a study actually pertain to study subjects and are true. External validity is the extent to which findings of a study can be generalized to apply to other populations. The use of invalid biomarkers can lead to invalid inferences and generalizations and ultimately to erroneous risk assessments.

Although biomarkers have a long history in medicine and public health, the systematic development, validation and application of biomarkers is a relatively new field in environmental health (Shugart et al., 1992; Anderson S et al., 1994), except for biological monitoring in occupational health (Hernberg & Aitio, 1987). However, such efforts are not new in others areas, e.g., in the validation of serum lipid biomarkers in cardiovascular disease. Lessons can be learned from the cardiovascular field that can be applied to the environmental health field; notably that the validation of markers for risk assessment can take a long time and is generally expensive. In the validation efforts, laboratory scientists and epidemiologists, clinicians, exposure assessors and statisticians need to be involved. In addressing societal impediments to the validation, an even broader range of disciplines, such as ethics, laws, and economics also need to participate.

Table 1. Factors affecting the validity and feasibility of biomarker studies: analytical procedures


Sampling constraints (for example, timing requirements)


Number of samples necessary for an acceptable precision


Degree of invasiveness of the sampling procedure


Availability of storage methods after the sample is taken (to avoid the need for immediate analysis)


Controlling or reducing the contamination of the sample when it is taken and when it is manipulated in the laboratory


Simplicity, possibility of routine usage, and speed of the procedure


Trueness, precision and sensitivity


Specificity for the component to be detected: interference must be identified to avoid misinterpretation


Standardization of the procedure

Adapted from: Dor et al. (1999)

Table 2. Factors affecting the validity of biomarkers: intrinsic characteristics of the biomarker


Significance: exposure, effect, individual susceptibility


Specificity in relation to the pollutant or pollutant family


Sensitivity: capacity to distinguish populations with different exposure levels, susceptibilities or degrees of effect


Knowledge of its background in the general population


Existence of dose-response curves between exposure level and marker concentration


Estimation of the inter- and intra-individual variability


Knowledge of confounding factors that can affect marker

Adapted from: Dor et al. (1990)

When used in risk assessment, information from biological markers may replace default assumptions when specific information regarding exposure, absorption and toxicokinetics is unavailable or limited (Table 3) (Ponce et al., 1998). Although examples of how this biomarker information can be used are limited, a general framework can be adduced (see section 2.2). Quantitative evaluations of the utility of biomarker information in risk assessment are rare (Bois et al., 1995; Ponce et al., 1998).

One compelling example of the use of susceptibility markers is the work of El-Masri et al. (1999). They investigated how changing glutathione-S-transferase theta (GSTT1) genotype frequencies would impact cancer risk estimates from dichloromethane by the application of Monte-Carlo simulation methods in combination with physiologically based pharmacokinetic (PBPK) models. They reported that average and median risk estimates were 23% to 30% higher when GSTT1 polymorphism was not included in the models. This analysis was a major factor in the permissible exposure levels promulgated by the US Occupational Safety and Health Administration (OSHA, 1998).

Goldstein (1996) has identified two important impediments to the development of biomarkers of value to risk assessment. The first is the over-reliance on mathematical models to the exclusion of monitoring data. This occurs because regulators have a need to make a decision and, for expedience, use models until better approaches are developed. However, once locked into a regulation, the existence of the model serves as a major inhibition to the development of more reliable methods of indicating exposure and effect, including biomarkers. The second is that ethical review boards may find it difficult to sanction research where participants are exposed in a scientific study to levels they are exposed to in the general environment or at work because the participation in the study is voluntary while the latter is generally involuntary.

Table 3. Use of biomarkers to refine risk assessment information


Use of biomarkers


Establish exposure characteristics


· Route of exposure


· Peak of exposure


· Total exposure


Estimate cumulative exposure


Establish absorption factors


· Inhalation


· Dermal exposure


· Ingestion


Identify factors that influence absorption


Identify interspecies differences


Identify sensitive population characteristics


Establish distribution kinetics


Establish half-life in blood or body


Identify interspecies differences


Identify factors that influence distribution, metabolism or excretion


Estimate cumulative exposure


Estimate peak exposure variables


· Time


· Concentration


Identify sensitive population characteristics


Identify mechanism of toxicity at target organ


Establish target organ potency


Identify sensitive population characteristics


Identify factors that influence target organ toxicity


Identify interspecies differences

(Ponce et al., 1998)

The concepts and principles supporting the use of biomarkers in the assessment of human health risks from exposure to chemicals have been reviewed by the International Programme on Chemical Safety (IPCS, 1993). The IPCS has produced the concise guidelines for the monitoring of genotoxic effects of carcinogens in humans (Albertini et al., 2000). It has also issued monographs on the methodology for the assessment of human health risks in a wider context, which includes the use of biomarkers (IPCS, 1994, 1999).


The widely accepted risk assessment paradigm includes the steps of hazard identification, dose-response assessment, exposure assessment and risk characterization (Fig. 1) (NRC, 1983). Using biomarkers to gauge exposure may contribute in various steps of the risk assessment process. In the hazard identification step, i.e., the determination of whether an agent might pose a threat to human health, there is a need to link an exposure with an adverse outcome. Given the different effects at different levels of exposure, there is need for understanding the specific effects of different exposures, particularly at lower levels of exposure. Then, in the exposure assessment stage, the extent of exposure is highly dependent on the agent and environment and builds on the specific source-path-receiver model utilized during hazard identification. The source-path-receiver model is the common approach to link source chemicals, the pathway of movement in the environment, and the route(s) of exposure of various receptors, in the case of risk assessment, individuals or groups of individuals (Nelson, 1997). Critical issues in exposure assessment include characterization of the magnitude, frequency and duration of exposure, the basis for the assessment and the identification of highly exposed subgroups. The risk characterization step requires consideration of any assumptions and models used, and attendant uncertainties used in developing the risk estimates. These estimates are then the basis for options to be selected in the risk management stage (Schulte & Waters, 1999).

Figure 1

Quantitative estimation of health risks is dependent on both exposure characterization and the nature of the dose-response relationships or toxicity of the agents involved. The greatest uncertainties in risk assessment almost always arise from sparse or inadequate exposure data, inadequate understanding of mechanisms of toxicity, and insufficient understanding of the exposure-dose-response pathway (Becking, 1995; McClellan, 1995). Two additional factors can lead to uncertainties in risk assessments. These include mixed or multiple exposures implicated in the disease pathway, and variability of both exposures and responses within and between individuals.

There are ambiguities in the risk assessment terminology that should be identified. For example, consideration of exposure will occur in two places in the risk assessment model. In the hazard identification stage, exposure is a component of the underlying research. This is distinct from the exposure assessment stage of risk assessment, where the extent of exposure of the population (whose risk is being characterized) to the identified hazard is determined. In a similar sense, dose-response considerations appear in two places. One of the criteria for the identification of a hazard is the finding of a dose-response relationship in the component studies. Additionally in the dose-response stage of risk assessment, the objective is to ascertain if there is a dose-response relationship in all the available data, identifying the shape of the curve and projecting to exposure or dose level, where health effects are reduced or believed absent. Finally, the concept of susceptibility can be in action throughout the risk assessment model. In hazard identification gene-environment interaction or effect modification may be assessed, and similarly in the dose-response stage susceptibility may be taken into account. Finally, in the risk characterization stage, different risk projections could be determined for various population subgroups identified by susceptibility factors.

2.1 Hazard identification

Historically, hazard identification has been the driving force in risk assessment (McClellan, 1999). Various national and international organizations have recognized the role human biological markers of exposure and effect can play in the hazard identification step: both make use of such data in classifying carcinogens.

Like other classic measures of exposure, there are limitations to the use of exposure biomarkers in epidemiological research for hazard identification (Schulte & Perera., 1993; Pearce et al., 1995). The major limitation is the general inability of biomarkers (with some exceptions) to indicate historic exposures. Additionally, their strength in integrating routes of exposure also may be a weakness by introducing confounding due to source, as full use of exposure biomarkers may also require understanding of those inherited and acquired factors that influence the level of exposure biomarkers (Vineis et al., 1990).

The role of biomarkers in hazard identification can be considered in the following examples. In determining whether a xenobiotic is hazardous, biomarkers may yield a more accurate determination than approaches based on less sensitive measures of exposures (e.g., job titles, as exposure proxies). In situations where exposures occurred that were variable or intermittent and the effect of exposure is integrated, biomarkers that represent a cumulation of exposure, such as haemoglobin adducts, might be useful (Perera, 1995). A biomarker approach may allow for clarifying exposure-outcome relationships better than with classical methods, due to reduction in exposure measurement error. For example, the role of aflatoxin exposure and liver cancer was not clear when studied by dietary questionnaire to assess intake of foods potentially contaminated with aflatoxin. However, a strong association was observed based on urinary biomarkers (metabolites and nucleic acid adducts) of aflatoxin exposure (Qian et al., 1994; Howe, 1998) (Table 4). Usually information is not available to permit this kind of comparison. In this example the molecular biomarker is useful because it is quite specific and biologically relevant, and it provides a better indicator of exposure than could be inferred from a questionnaire since respondents are not aware of how much aflatoxin they consume. One might ask if direct measurement of the aflatoxin component of all foodstuffs ingested and measurement of amounts of food intake per day might also lead to a better measure of exposure than the questionnaire surrogate. In the case of aflatoxin this is probably not true due to the difficulty of measuring food intake, the possible variability of aflatoxin levels within food, the difficulty of extractions of aflatoxin from foods, and analytical detection limits for such methods. However, for some other agents, external direct measures of exposures may be both feasible and as cost effective as biological measures, and will also provide improved estimation over such surrogates as questionnaire data.

The basic rationale for using exposure biomarkers is that they could provide, in some cases, a more accurate method for assessing exposure and, ultimately, risk (Fig. 2) (Schulte & Waters, 1999). While use of biomarkers can reduce mis-classification, it is also possible that measurement error in the biomarker may contribute to bias in the measure of association (White, 1997; Saracci, 1997). Such error can be evaluated and its impact adjusted for, but, on balance, it is better to avoid or minimize it with good laboratory and epidemiological practices.

Figure 2

Table 4. Comparison between estimated dietary intake of aflatoxin and biomarker derived exposure data

A. Relative risks based on estimated dietary intake of aflatoxina

Dietary aflatoxin B1 exposure (µg/year)

Relative risk

95% Confidence interval

< 71



71- 113


0.8 – 3.1



0.4 – 1.9

a Estimate of the intake of aflatoxin was based on a dietary questionnaire to assess the intake of foods potentially contaminated by aflatoxins and analysis of related food items for aflatoxin contents.

B. Relative risks based on urinary biomarkers of aflatoxin


Relative risk

95% Confidence interval

Metabolites or adducts


2.1 – 12

Adducts only


3.6 – 73

Source: Study of aflatoxin and risk of liver cancer in Shanghai, China from 1986 to 1992

(No. of cases = 55) (Qian et al., 1994; adapted from Howe, 1998)

The value of biological markers in epidemiological studies that will be useful in the hazard identification step of risk assessment depends on the quality of the design and analysis of the studies. Various reviews of biomonitoring and molecular epidemiological studies have been conducted and in some instances have found that a better design or analysis could have been applied (Bonassi et al., 1994; IARC, 1997). In 1994, Bonassi et al. reviewed three years of biomonitoring studies and found that only 5% of those studies adopted the best statistical techniques available. The major recommendations were to focus more on point estimates and confidence intervals instead of significant tests, utilize appropriate multivariate techniques and pay more attention to adjustment for confounding and evaluation of possible interaction between factors.

2.2 Dose response

One of the most controversial aspects of risk assessment is the extrapolation of higher-level exposure data to lower levels of exposure (Goldstein, 1996). In risk assessment the ascertainment of a dose-response relationship is crucial for ultimately determining the shape of the dose-response curve and for predicting a no-observed- adverse-effect level (NOAEL). Exploring the lower end of the dose-response curve through epidemiological studies is generally impractical, if not impossible, due to extraordinarily large sample sizes (Stayner, 1992). However, biomarkers can contribute to identifying a dose-response relationship at lower levels of exposure. The demand on environmental epidemiology to evaluate increasingly subtle health risks requires more accurate estimation of the quantity and timing of a toxicant reaching target tissue (Kriebel, 1994). Kriebel (1994) described a two-stage approach to derive estimates of dose from exposure data and then linked them to epidemiological models estimating disease risk. Such an approach incorporates physiological processes into epidemiological modeling and is possibly more valid than approaches with less detail. Biomarkers of exposure can be used as indicators of dose, which can then be assessed against classic measures of morbidity or mortality.

Often dose-response determinations are made by use of PBPK models. Examples of how biomarker data could be incorporated in PBPK models include: 1) calibration of the model (empirically determining the population values of kinetic parameters); 2) validation of the model (determining how well the model predicts data of another cohort); 3) prediction (applying the model to new cohorts, and predicting the internal doses associated with given exposure scenarios). These predicted measures of internal (biologically-effective) dose can then be used in dose-response modelling, in lieu of the external exposure measures, to predict disease risk. The internal dose measures may be better predictors of disease risk, especially when exposure-dose is nonlinear (e.g., due to capacity-limited metabolism). A PBPK model could also potentially allow for extrapolation from limited data, such as a short-term laboratory study to predict the biomarker concentrations that might be found in a population. For example, the model for carboxyhaemoglobin formation (Andersen et al., 1991) from exposure to methylene chloride was calibrated against short-term human exposure but could be used to predict carboxyhaemoglobin from long-term exposures.

Another use of biomarkers can be as outcome measures that correlate with exposure. Here exposure markers are not what is needed; rather, there is a need for effect markers. Effect markers are those that relate to or predict disease. A marker validated to predict disease can be used as a surrogate for disease. For example, specific types of chromosomal aberrations that appear to predict cancer risk on a group basis could be used as the outcome variables in a dose-response analysis, as in the case of ionizing radiation exposure (Joksiƒ & Spasojeviƒ -Tišma, 1998).

Persistent microalbuminuria and low molecular weight proteinuria identify incipient diabetic and cadmium-induced nephropathy, respectively (IPCS, 1992; Emancipator, 1999). Proteinuria is also the main factor accelerating progressive renal disease toward end-stage renal failure (for review, see Remuzzi & Bertani, 1998), which confers a prognostic value to its quantitative changes over upper reference limits.

The literature on biomarkers in PBPK models is relatively limited. If there are enough biomarker data there may be no need for the PBPK model at all. This would be particularly true for biomarkers of effect, if the human biomarker data came from the population that was at risk, or one with similar exposures.

2.3 Exposure assessment for risk assessment

Exposure assessment has generally been considered the weak link in hazard evaluation and risk assessment (Dary et al., 1996). The exposure assessment component of risk assessment includes consideration of such issues as representativeness of exposure measurements for a population, differences in exposures within and between individuals, individual differences in uptake and biotransformation, identification of factors that control or modify exposures, exposure estimation methods applicable in the absence of direct measurements, and identification of the most relevant dose metric (the most relevant measure of dose) for the agent under consideration (Schulte & Waters, 1999). The use of biomarkers in assessing exposure for risk assessment increasingly may include consideration of susceptibility factors in conjunction with exposure factors (gene-environment interaction), such as the presence of a specific genetic polymorphism for a metabolic enzyme (Bois et al., 1995; D’Errico et al., 1996). Such genetic differences may account for some interindividual variability in exposure markers.

Quite often epidemiological studies utilize exposure surrogates rather than direct measurement of exposures. For environmental studies, surrogates might include geographical location such as residence for a drinking-water or air pollution study, age of housing in studies of lead-based paint exposures, or proximity of residence to electrical power lines. Occupational studies use surrogates such as job title or job group, years worked at a plant, pounds of pesticide applied per week, and tasks performed when direct measurements are not available or are limited (Stewart et al., 1991; Goldberg & Hémon, 1993). The use of quantified direct measurements of personal exposures can lower uncertainty in the risk assessment process considerably compared to the use of such exposure surrogates (Schulte & Waters, 1999). Biomarkers may serve to evaluate the completeness of exposure assessment information by associating environmental or source information, exposure measurements, and epidemiological and human activity data with internal dose (Dary et al., 1996).

In some cases, biomarkers of exposure may be better than external measurements of exposure for situations where protective equipment has been used or when there is the possibility of dermal (or gastrointestinal) absorption.


The ultimate driving force for whether biomarkers will contribute to environmental health efforts is the validity of the markers. Validity is a complex characteristic that describes the extent to which a biomarker reflects a designated event in a biological system. Generally, these events are exposure, effects of exposure, disease and susceptibility. Validity has meaning according to discipline as well. To the laboratory scientist, validity often refers to the nature of the biomarker and the characteristic of the assay for the biomarker. Thus, the sensitivity of the assay to detect a signal at a given concentration, and the ability of the signal to be specific for a particular event are indications of validity to the laboratory scientist. In addition, the scientist wants to know what factors might influence an assay. The epidemiologist relies on the laboratory definition of validity as the cornerstone of population studies, but then needs to know how likely a person with a positive assay or test is to develop disease (or have been exposed) and how likely a person with a negative test is to be free of disease (or exposure). The epidemiologist also needs to know how feasible the marker is to use in human populations and the reliability of the assay under field conditions. Moreover, the epidemiologist needs to know how the frequency of the marker varies in different population subgroups defined by age, race, gender, pre-existing illness, diet and various behavioral factors. Only when validity at the laboratory and population level has been established is a biomarker ready for the full spectrum of environmental research and uses. As noted, most biomarkers have not had that level of validation. A broad effort is underway but the products of this activity are not available yet.

Validation of candidate biomarkers is an empirical process that can be approached by producing several different, but convergent lines of evidence. There is extensive literature on criteria for validating biological markers (e.g., WHO, 1975; Lucier & Thompson, 1987; Hernberg & Aitio, 1987; Schulte, 1989; Schatzkin et al., 1990; Margetts, 1991; Schulte & Mazzuckelli, 1991; Schulte & Perera, 1993; Schulte & Talaska, 1995; Boffetta, 1995; Ponce et al., 1998). In general, these criteria include understanding the natural history, biological and temporal relevance, pharmacokinetics, background variability, dose response and confounding factors (Schulte & Talaska, 1995). Biomarker validity is also dependent on reliability of the assay to measure the biomarkers. These criteria allow for the assessment of whether a biomarker represents an event that is in a continuum between exposure and resultant disease, and whether the biological specimen containing the biomarkers is appropriate and the marker reflects the time period of concern. Finally, by assessing confounding and effect modifying factors, it is possible to understand what other factors influence a biomarker or its assay.

The careful measurement of strong confounders and effect modifiers should be given as much attention as is given to measurement of the exposure and disease variables or biomarkers. Consideration should be given to mounting validation substudies to quantify measurement error in important covariates (Hatch & Thomas, 1993). Measurements of biological markers are the building blocks of research and mechanistically based risk assessment. If the measurements are inaccurate, the research and risk assessments are likely also to be biased. Controlling measurement validity makes it possible partially to control study validity since measurement errors can produce biased estimates of regression coefficients used in statistical models of exposure and disease (Louis, 1988). Measures of association, such as the odds ratio, can be distorted, depending on the type of error and other characteristics, towards or away from the null hypotheses of no association between the biomarker and disease (or exposure).

The terminology to express measurement error traditionally used in biomarker measurements is different from that applied in analytical chemical laboratories. In the latter, traditionally, variability of the results is considered on an individual basis, and accuracy (trueness and precision) refers to individual analytical results. Trueness refers to how close to the true value the average of the results is, precision to the scatter of the results around their average, while accuracy is the combination of the two characteristics. Defective trueness is bias, defective precision, imprecision. In epidemiological work involving biomarkers, the emphasis is on the biomarker level in the studied population rather than on an individual, and the measurement error includes the intra-individual variation with time (White, 1997). As White (1997) notes: measurement errors for an individual can be defined as the difference between a person’s measured biomarker (the biomarker "test") and the person’s true biomarker. The true biomarker is the biomarker without laboratory or other sources of error and, if the measure can fluctuate over time, the true biomarker is an integration of its concentration over the time period of etiological interest.

Validity in the context of epidemiological research involving biomarkers can be defined as the relation of the biomarker test (the potentially mismeasured biomarker) to true biomarker in the population of interest. Parameters that describe the measurement error in the population are called measures of validity (White, 1997). Two indicators of measurement error are used to describe the validity of an observed measurement compared with the true measurement (Armstrong et al., 1994). The first is systematic error or bias that would occur on average for subjects measured. The second is subject error, which is additional error that varies from subject to subject. The subject error is also called imprecision or the measure of the variation of measurement error in the population. Precision can be assessed by a construct known as the validity coefficient. It ranges from 0 to 1 with the value one indicating that the observed measurement is a perfectly precise indicator of the true measurement (Armstrong et al., 1994). A validity study would be defined here as one in which a sample of individuals is measured twice: once using the biomarker test of interest and once using a perfect measure of the true biomarker (White, 1997). However, for most biomarkers such perfect measures of the true biomarker do not exist, and, in practice, validation of a method must rely on comparison to other (similarly unvalidated) methods. Then the indicator of biomarker measurement error from the validity study can be applied to what is known about the association under study in the parent study to estimate the effects of biomarker error on the association of interest (White, 1997). While the impact of measurement error on exposure-disease associations has been studied extensively, the impact on estimates of interaction of two or more risk factors has been studied less thoroughly (Greenland, 1993). Assessment of interaction of multiple exposures, gene-environment or gene-gene is an important issue in environmental epidemiology and all the more important with biomarkers depicting mechanistic events.

There are numerous sources of measurement error in biomarkers; some of these are shown in Table 5.

Table 5. Examples of sources of error in biomarker measurement in epidemiological studiesa

Errors in the laboratory method as a measure of the exposure of interest

· Method may not measure all sources of the biological true exposure of interest

· Method may measure other exposures that are not the true exposure of interest

· Methods may be influenced by subject characteristics (other than the true exposure) that the researcher cannot manipulate, e.g., by the disease under study or by other diseases

Errors or omissions in the protocol

· Failure to specify the protocol in sufficient detail regarding timing and method of specimen collection, specimen handling, storage and laboratory analytical procedures

· Failure to include standardization of the instrument periodically throughout the data collection

Errors due to variation in execution of the protocol

· Variations in method of specimen collection

· Variations in specimen handling or preparation

· Variations in length of specimen storage

· Variations in specimen analysis between batches (different batches of chemicals, different calibration of instrument)

· Variation in technique between laboratory technicians

· Random error within batch

Adapted from: White (1997)


In addition to measurement error, the uncertainty of the results is affected by biological variability within subjects, i.e., short-term variability (hour to hour, day to day) in biological characteristics due to, for example, diurnal variation, time since last meal, posture (sitting vs lying down); medium-term variability (month to month) due to, for example, seasonal changes in diet; and long-term change (year to year) due to, for example, purposeful dietary changes over time.


Ultimately, validation requires the use of epidemiological study designs to assess at least one of three types of relationships: exposure-dose; biological effects-disease; and susceptibility influencing an exposure-disease relationship. Studies that contribute to these types of validation and bridge the gap between laboratory experimentation and population-based epidemiology have been referred to as "transitional" studies (Hulka, 1991; Schulte et al., 1993; Rothman et al., 1995). They may be designed to evaluate exposures, health effects or susceptibility, and some may have the characteristics of pilot or developmental studies (Hulka & Margolin, 1992).

In this section the kinds of information and approaches to validate specific types of biomarkers are discussed. Characteristics of valid biomarkers are outlined in Table 6.

Table 6. Characteristics of valid biomarkers

Biomarker type

Characteristic of validity


Consistently linked with exposure at relevant levels of exposure with confounding and background exposures assesseda


Consistently linked with increased risk with confounding and effect modifying factors assessed


Can distinguish subgroups at risk given specific exposure


Biomarkers of exposure may also be validated by establishing a constant link to an adverse health effect or to the concentration of the chemical in the target organ.



4.1 Exposure biomarkers

The validation of biomarkers of exposure requires equal attention to assessing both the exposure and the biomarkers so that a fair comparison between them can occur. However, the relationship between the biomarkers and the exposure will vary due to host factors as the biomarkers become further away from the exposure, depending on the number of steps in the absorption, metabolism and clearance pathways between uptake and the specific biomarker (Schulte & Waters, 1999). This applies to any form of exposure. It is due to intervening host factors that vary between individuals such as breathing rate and capacity, activation, detoxification, elimination, DNA repair, etc. Thus a high correlation between exposure and the marker may not always be observed and an exposure-response relationship may vary between people. It is therefore important to identify and adjust for factors that can influence an exposure-response relationship. For example, to validate hydroxy-ethyl haemoglobin adducts as exposure biomarkers for ethylene oxide at low dose, investigators adjusted for age, smoking, and education in a linear regression model (Schulte et al., 1992). Additionally it may be useful to consider effect modifying factors, such as metabolic polymorphisms (Bois et al., 1995).

There are some exceptions to the validation strategy that focuses on the demonstration of a correspondence between a biomarker of exposure and external exposure. Alternative ways to validate biomarkers include the assessment of their relationship with the concentration in the critical organ (e.g., concentration of cadmium in the kidney vs. in blood or urine) or with critical effects (neurotoxic effects of lead vs. blood lead concentration). Indeed, a good biomarker of exposure should be useful to predict adverse effects, rather than exposure levels. This may be especially the case when accurate and valid measurements of the "true" exposure are difficult or impossible to obtain (use of protective devices, multiple pathways of uptake, etc.).

It is possible to apply qualitative tests to determine whether external exposure or an exposure biomarker would be a better predictor for disease (Steenland et al., 1993). One test involves determining if the biomarker is more highly correlated or associated with the disease than external exposure. A second test is whether, given the same level of exposure, those with higher levels of the biomarkers are more likely to develop the disease.

When absorption mainly occurs through the dermal route or when individual protective devices are used, biomarkers of exposure can provide reliable measurements of internal dose, which are useful to assess dose-response relationships. Fig. 3 shows a logistic regression model based on published data (Calleman et al., 1994) showing that 97.5% of subjects with clinical signs of peripheral neuropathy are correctly classified on the basis of acrylamide adducts to N-terminal valine (AAVAL) in haemoglobin (Hb). On the basis of the parameters of the logistic regression, the calculated benchmark dose corresponds to 0.8 nmol AAVAL/g Hb.

In evaluating the role of metabolic polymorphisms, the presence of a range of doses in which the modifying effect of metabolic enzymes could be seen, is a major issue. A pertinent example comes from a study on the urinary excretion of 1-hydroxypyrene in traffic police officers (Merlo et al., 1998). In the study group, subjects carrying the CYP1A1+ polymorphism had higher levels of hydroxypyrene in the urine, but only at low exposures to PAHs.

4.2 Effect biomarkers

Biomarkers of intermediate effects, i.e., between exposure and disease, can be validated in case-control studies and cohort studies (Rothman et al., 1995; Howe, 1998; Hagmar et al., 1998; Muñoz & Gange, 1998). Once validated, these markers can serve as surrogates for disease, albeit with some probability functions since generally not all people with a given biomarker will develop the disease, but the groups with the high levels generally will be at greatest risk. A good example comes from a recent prospective study on the association between cytogenetic biomarkers and cancer risk (Hagmar et al., 1998). This study, which followed five European cohorts has shown that subjects in the group with the highest frequency of chromosomal aberrations experienced an overall cancer risk more than double with respect to the lowest frequency group. In the same study, no association was observed between sister chromatid exchange (SCE) frequency and cancer risk, whereas inconclusive results were found for the micronucleus assay. More recently, a nested case-control study found that the association between chromosomal aberrations and cancer appeared to be independent of host factors like age and sex, and could not be explained by exposure to identified human carcinogens (Bonassi et al., 2000).

Figure 3

The lack of validation of most biomarkers of intermediate effect is probably the most critical impediment to the broad use of biomarkers in risk assessment. Validated biomarkers of effect can be used as disease surrogates and thus will be countable end-points that can fill voids left by the inability to count low frequency adverse morbidity or mortality events (Hattis & Silver, 1993; Goldstein, 1996; McMichael & Hall, 1997). Earlier results may be obtained from epidemiological studies if use of a biomarker increases the statistical power of the study (McMichael & Hall, 1997). The prospective epidemiological study is the gold standard for validation effect biomarkers. The timing and frequency of specimen collection in prospective studies are important and can influence the validation. This type of study provides estimates of the risk of disease of individuals with and without a particular biomarker. These studies are time consuming and costly. To reduce the time and cost variables, efforts are underway to follow prospectively large cohorts and bank biological specimens (Willett, 1998). These will allow for "compressed" evaluations of a biomarker and disease risk at the same time.

A measure of the degree of validation of an intermediate marker of effect is the extent to which the exposure is mediated through a marker, i.e. whether the marker is actually strongly predictive of the clinical disease, and the disease never or rarely occurs without this antecedent marker. This may be assessed by calculating the attributable proportion which has also been referred to in the literature as "population attributable risk" or "etiologic fraction" (Benichou, 1991; Trock, 1995). The attributable proportion associated with a particular biomarker is an estimate of the proportion of diseased cases that must progress through the biomarker, i.e., the cases that would not occur if the event(s) resulting in the biomarker could be prevented (Schatzkin et al., 1990; Trock, 1995). The attributable proportion (AP) includes consideration of the sensitivity (S) of the assay (i.e., the ratio of subjects positive in the assay, who developed the disease, to the whole number of diseased subjects), and the relative risk (RR) of disease for the subjects positive in the assay. It is defined as: AP = S(1- (1/RR)). The sensitivity is the factor with the greatest impact in the attributable proportion (Schatzkin et al., 1990). The attributable proportion takes into account both the strength of an association between a marker and disease and also the prevalence of the marker. Thus, for example, using data from the European studies on chromosomal aberrations and cancer (sensitivity 46/91 = 0.50, RR 1.53/ 0.79 = 1.93), the proportion of cancer cases attributable to the chromosomal damage was calculated to be 24% for the Nordic cohort (Bonassi, 1999). Trock (1995) has described the next step.

Once it has been established that a significant proportion of tumours can be attributed to a particular marker, epidemiological principles concerned with ‘intervening variables’ can be used to examine the extent to which the marker truly represents an event intervening between exposure and cancer. In evaluating the relationship between an exposure and a disease outcome, one typically does not use statistical adjustment methods to adjust for a variable that is an intermediate step between exposure and outcome (Weinberg, 1993). Such an adjustment would sharply reduce or even eliminate the apparent effect of the exposure since the marker’s association with disease is a direct result of its association with exposure (assuming that the marker represents the relevant time period of exposure with respect to onset of disease) (Trock, 1995). One can take advantage of this property to assess the role of a marker as an intervening variable. If one compares the crude (i.e., unadjusted) RR for exposure to the RR for the exposure effect adjusted for the biological marker, the extent to which adjustment for the marker has reduced the apparent exposure effect indicates the degree to which the marker is linked to the exposure-disease relationship (Trock, 1995). If the effect of exposure occurs primarily through a pathway involving the marker, then the marker-adjusted exposure effect will essentially be eliminated, i.e., the adjusted RR will be close to 1.0 (Schatzkin et al., 1990). In some cases a marker will be useful even though it is not on the causal pathway. Those are cases where it correlates to something on the causal pathway (e.g., protein adducts, Ehrenberg et al., 1996) and ultimately to disease risk.

Another measure of validation of a biological marker of effect is the positive predictive value. Predictive value for a marker of disease is the proportion of people studied with a particular disease among all the people who have the marker. Predictive value is not only a property of the marker assay, it is determined by the sensitivity and specificity of the assay and the prevalence of the disease. Thus, for example, a marker that is 90% sensitive and 90% specific will still only have a positive predictive value of 50% when the prevalence of the underlying disease is 10%. Parallel considerations should be extended to the use of the negative predictive value whenever the hypothesis of association is rejected. Field studies that do not incorporate prevalence considerations in planning are likely not to be able to detect an association between a marker of effect and disease, even if one exists (Schulte & Perera, 1993).

Positive predictive value and attributable proportion reflect very different things (Ottman, 1995; Khoury & Wagener, 1995). Positive predictive value, the risk of disease among persons with a specific marker is important from the point of view of the individual. Attributable proportion, on the other hand is the proportion of diseased cases that must progress through the biomarkers and thus could be prevented if that process could be interrupted. This is important from an environmental or public health point of view.

In addition to positive predictive value and attributable proportion, two other concepts are useful in the interpretation of biomarker data: negative predictive value (see above) and sentinel biomarkers. The concept of a sentinel biomarker involves a biomarker that, regardless of predictive value or attributable proportion, may have properties (increased frequency of increased concentration or of occurrence) that might be indicative of exposure to an environmental hazard or onset of a biological effect (see Appendix 2).

4.3 Susceptibility biomarkers

Considerable variability exists in the response of humans to toxic substances. A very large number of genetic conditions potentially enhancing one’s susceptibility to chemicals have been identified (Appendix 3). However, only in a few cases such as the glucose-6-phosphate dehydrogenase deficiency, has a causal relationship been demonstrated, deficient individuals being more susceptible to toxic environmental oxidants (Stokinger & Mountain, 1963).

Polymorphisms may be markers of susceptibility and a long list of genes and their variants (polymorphisms) has been and will be established. Many of these genes are quite general in function, but changes in this function due to a polymorphism may influence the susceptibility of developing disease. In addition to polymorphisms in xenobiotica metabolism enzymes, polymorphisms in genes that influence or control cell differentiation, apoptosis, cell cycle kinetics, signal transduction and DNA repair may influence the health outcome when exposed to an environmental toxicant.

Perhaps the greatest potential contributions of biomarkers to risk assessment and risk management will be the inclusion of inherited susceptibility biomarkers. However, while there has been extensive use of susceptibility biomarkers in the development of pharmaceuticals (Evans & Relling, 1999), the potential contribution to risk assessment in occupational or environmental chemical exposure has rarely been realized. Susceptibility biomarkers may reflect variation in exposure, kinetics and effects, and are therefore important to consider in risk assessments (Bois et al., 1995; Dickey et al., 1997). These biomarkers have both promises and perils for individual and population risk estimation. The promise is for a more refined assessment of risk through the identification of gene-gene and gene-environment interactions and also for the focusing of prevention and control programmes on high-risk individuals. The perils include ethical and social issues including stigmatization, discrimination and the misconception that removing a susceptible person from the exposure scenario without reducing exposure opportunities will reduce risk effectively, when it may not, on a comparative basis (Vineis & Schulte, 1995). There are also issues in using susceptibility markers as effect modifiers in epidemiological studies. These include: whether there is a correspondence between a genotype and phenotype; whether there is a mechanistic reason to consider the marker; and whether the prevalence of the allele is frequent enough to assess in the population in a practical way.

Before applying susceptibility biomarkers in epidemiological studies, there are several important issues to consider. The use of genotype rather than phenotype may lead to misclassification as the actual enzyme activity is influenced by a number of exogenous and endogenous factors, thus diluting the genetic component. Furthermore, susceptibility factors, be they enzymes, receptors or other target molecules, are highly compound-specific, so that even closely related substances may not be substrate to the same polymorphic enzymes. Combination of and interactions between various alternative pathways governed by polymorphic and environmentally regulated enzymes are important as metabolism generally involves several stages. The effect of the genetic polymorphism may also depend on the outcome measure. For example, in a case report, workers exposed to methyl bromide showed different neurotoxic responses, people expressing the gene GSTT1 being more severely affected. In contrast, the least sensitive to the neurotoxic effect had the highest level of alkylation damage (Garnier et al., 1996).

One of the most widely known examples of how susceptibility biomarkers combined with an exposure measure can inform risk assessment is the metabolic polymorphism for N-acetyltransferase in the case of bladder cancer. This enzyme is involved in the detoxification of arylamines, and individuals classified as slow acetylators have an increased risk of bladder cancer when exposed to e.g., beta-naphthylamine (Cartwright et al., 1982). However, in a Chinese study the NAT genotype did not influence the risk of bladder cancer, but in this situation the exposure was to benzidine. Monoacetylation mediated by NAT is an activation, rather than a deactivation pathway of benzidine (Rothman et al., 1996). This stresses the importance of knowledge of the biotransformation of the compound under investigation prior to testing the role of susceptibility markers.

The influence of genetic factors on exposure-disease associations can be large. Calabrese (1997) demonstrated that genetically determined biochemical differences between people for a range of phenotypes could exceed 10-fold. In a Monte-Carlo simulation study, Bois et al. (1995) illustrated, by pharmacokinetic modelling of DNA adducts in the bladder of people exposed to oral bolus doses of 4-aminobiphenyl, that the adduct levels of the most susceptible individuals are 10 000 times higher than for the least susceptible and that the 5th and 95th percentiles differ by a factor of 160. Input parameters for the model were derived from the literature on human in vitro studies, or from dogs in the absence of human data. Therefore, accounting for genetic variability may have important implications for risk assessment (Dickey et al., 1997).

The assessment of gene-environment interaction is important in the validation of biomarkers of susceptibility (and exposure). Gene-environment interaction is defined as "a different effect of an environmental exposure on disease risk in persons with different genotypes," or alternatively, "a different effect of genotype on disease risk in persons with different environmental exposures" (Ottman, 1996). Ottman (1996) has described five biologically plausible models of gene-environment interaction, each of which leads to a different set of predictions about disease risk in individuals classified by the presence or absence of a high-risk genotype or environmental exposure. If a biomarker of susceptibility is to be validated for disease, its relationship to both disease and exposure needs to be determined. Valid evaluation of gene-environment interaction requires the accurate measurement of both genetic and environmental factors (Rothman et al., 1999). Modest exposure assessment errors may produce a biased estimate of the interaction parameter that results in a substantial increase in sample size requirements (Garcia-Closas et al., 1998).


The optimal use of biomarkers in environmental health risk assessments will most likely occur if human studies are linked to studies of laboratory animals and cell lines (Shugart et al., 1992; Anderson S et al., 1994). An additional extension of the use of biomarkers is studying appropriate species of wildlife (Barrett et al., 1997). Biomarkers can serve as a common element in studies of these different groups or materials. Thus a biomarker identified in an exposed laboratory animal or cell line might also be seen in wild or laboratory animals or humans with similar exposures.

A parallelogram type approach (Sobels, 1993; Sutter, 1995) can be used to assess the relationship between markers and risks in those groups (Fig. 4). The parallelogram approach is derived from the work in the 1970s of Sobels (1993) to extrapolate genetic damage from animals to humans. Genetic damage which cannot be measured directly, such as in human germ cells, can be estimated by measuring the same kind of damage in both germ cells and somatic cells of the mouse. With data on the induction of mutations or chromosomal aberrations in both germ cells and somatic cells of the mouse, it is possible to estimate germ cell mutation frequencies in humans on the basis of what can be measured by monitoring genetic damage in human somatic cells (Sobels, 1993). Sutter (1995) has modified this approach to include in vitro-in vivo extrapolation. In the parallelogram experimental approach to knowledge of mechanism, in vitro data are used to test the hypothesis that a specific mechanism of action exists in rodents and humans (Sutter, 1995).

An example where the original parallellogram approach has been used is for the genotoxic compound 1,3-butadiene (Pacchierotti et al., 1998). The purported estimates of heritable damage in man were based on data for heritable translocations in germ cells, and bone marrow micronuclei induced in mice and chromosomal aberrations on lymphocytes of exposed males. The rate of heritable translocation induction per ppm/ h of butadiene exposure was estimated to be approximately 0.8 per million live born compared to spontaneous incidence of balanced translocations in humans of approximately 800 per million born. Other compounds have also been characterized using the parallellogram approach, such as ethylene oxide, cyclophosphamide and acrylamide (Waters & Nolan, 1995).

Figure 4

Clearly, most identified human carcinogens are genotoxic, thus helping to build the case for human germ cell mutagenicity. However, there are 13 putative germ cell non-mutagens adequately tested for carcinogenicity, 11 of which are genotoxic carcinogens and these include vinyl chloride and propylene oxide (Waters et al., 1999).

With regard to endocrine disruptors, those operating through the spindle receptor or the genotoxic estrogens such as fosfesterol could usefully be examined using the parallellogram approach to determine the germ cell risks (DeRosa et al., 1998; Olea et al., 1998).

Biomarkers can be used to reduce high- to low-dose and species extrapolation-related uncertainties by providing information on common mechanisms and the development of mechanistically based mathematical models (Sexton et al., 1995). The incorporation of biomarkers of exposure and susceptibility in physiologically based pharmacokinetic models has allowed for interspecies comparisons and enabled the simulation of different enzyme activities among individuals (Fennell et al., 1996). Biomarkers also may serve as an alternative to the use of PBPK models for determining dose (Rhomberg, 1995). They are particularly useful when they are more easily or accurately measured than the actual exposure. Since some biomarkers have extraordinary sensitivity, they may significantly extend the range of empirical characterization of dose and response in cases where they may be detected and measured at dose levels below those at which other effects are directly observable (Rhomberg, 1995; Ehrenberg et al., 1996). For example, adduct measurements of some alkylating agents may be used to indicate disease risks at levels too low to be detected by epidemiological means (Ehrenberg et al., 1996).


Validation and successful use of biomarkers require a high degree of analytical accuracy and knowledge of what they mean in terms of health and disease.

Recent developments in information technology, molecular biology and instrumentation have provided new tools for use in environmental health research and biologically based risk assessment.

The specificity and sensitivity of many biomarkers will be improved by the introduction of new analytical methodologies, e.g., speciation of metal ions by inductively coupled plasma mass spectrometry (ICP-MS) and mass-spectrometric techniques to detect metabolites and adducts.Chromosomal aberration appears to be one of the most promising biomarkers for its association with cancer risk. Application of new detection methods, i.e., primed in situ labelling (PRINS) and fluorescence in situ hybridization (FISH) will extend the observation from the chromosome level to specific genes relevant for the disease process. Imaging technologies such as magnetic resonance or positron emission tomography (PET) and single photon emission computerized tomography (SPECT) are particularly interesting for studies in human populations as these methods are non-invasive and can measure change at the molecular scale.

Environmental genome projects will identify new single nucleotide polymorphisms (SNP) in genes involved in the disease process, information that can be used to identify new susceptibility factors, e.g., control of cell cycle and apoptosis. A spin-off of the project has been the development of new instrumentation making large-scale epidemiological studies using genetic markers of susceptibility more feasible, including the study of gene-gene interaction. Detection of polymorphisms in almost any protein will result in large amounts of data on their biological and health effects. The data will help detect new susceptible populations. The explosion of polymorphism data requires extension of bioinformatic approaches towards epidemiological databases.

Development of new animal and in vitro models, e.g., expressing specific allelic variant, will enable us to study the role of specific enzymes as risk factors in disease development and on the level of biomarkers. Knowledge of metabolism, product formation and general mechanisms of action are required for the development of biomarkers for environmental agents.

High output technologies, such as DNA microarray, can be used to study gene function and expression. The technologies will allow for evaluation of the temporal and spatial pattern of gene expression under various exposure conditions. Furthermore, they will give additional information on response in different species, thus validating animal models for the study of human disease. The technology could lead to better characterization and understanding of the disease processes, and the development of new relevant biomarkers.

Introduction of new and validated methods based upon new technologies to study biomarkers of exposure, effect and susceptibility at the different levels of the risk management process will be of great assistance to risk assessors.

Advances in biologically based pharmacokinetic modelling and simulations and new approaches for addressing uncertainty will be a major way for incorporating biomarkers into rich assessments. PBPK models are becoming increasingly complex and uncertainty analyses for models of such complexity are difficult to apply by using analytical calculation. Stochastic models such as Monte Carlo simulations and Markovchain Monte Carlo models are useful to address these difficulties. However Monte Carlo simulations have limitations, such as those involving prior distributions for model parameters. Bayesian analysis is a useful approach for addressing these distributions (Bois, 1999).


The Task Group, building on previous categorizations and evaluations of biomarkers for research, considered them for risk assessment. Risk assessment was defined as the set of steps between research and risk management. It provides society with estimates of risk when uncertainty exists about the safety of prevailing or future levels of exposure to environmental and occupational toxicants.

A framework for selecting and validating biomarkers for risk assessment was developed by the Task Group. Examples were cited of how the three types of biomarkers, of exposure, of effect and of susceptibility, could be validated for research and be used in risk assessments. Valid biomarkers can lead to biologically based risk assessments.

There have been few instances where validated biomarkers have been used in quantitative risk assessments. Future work should include scientific, technical, organizational and administrative efforts to coordinate efforts to set an agenda for research on biomarkers that will contribute to conducting important risk assessments. This will require long-term commitments for collaboration and the conduct of prospective studies to link biomarkers to disease risks.


Validated biomarkers are useful in reducing uncertainty in risk assessments. However, biomarkers should be viewed as another set of tools available for researchers and risk assessors, not as a replacement for traditional approaches.

Validation of biomarkers for research and risk assessment requires both laboratory and epidemiological studies.

Successful use of biomarker data implies an understanding of mechanism. The incorporation of mechanistic data in risk assessment is certainly important, but risk assessments and regulations should not wait for the development of mechanistic data nor should uncertainty about mechanism be used to block public health action.

The contribution of biomarkers of susceptibility has great potential but has yet to be realized on a large scale in quantitative risk assessment.

There is a need for a long-term commitment to the assessment of the validity of biomarkers for risk assessment, environmental health research and public health practice.


In making the following recommendations, the Task Group recognized the role given to the IPCS to facilitate and increase coordination of international activities in order to promote the further work needed to define human health effects associated with exposure to chemicals and to provide the basis for priority-setting actions in order to protect health.

9.1 General recommendations

The following recommendations were formulated:


to formulate ethical guidelines to promote biomarker research while guaranteeing individual privacy and integrity;


to incorporate mechanistic information, utilizing biomarkers, in risk assessment;


to develop an international set of principles for the collection, archiving and use of biological specimens and acquired data, notably those from humans;


to establish guidelines for biomarker studies in humans and to critically evaluate whether and under what conditions such experimentation is warranted.

9.2 Recommendations for future research

9.2.1 Prevalidation stage

The Task Group noted that results from validated biomarkers should be used in public health decisions, and recommended:


to develop more incisive biomarkers to fill in the gaps in the continuum of events from environmental exposure to clinical disease expression, taking advantage of new, high-throughput technologies;


to study the genetic basis for different susceptibilities toward environmental exposures and how this exposure influences the phenotype;


to develop and use bioinformatics and advanced statistical methods to fully utilize existing and newly generated data.

9.2.2 Validation stage

The Task Group further recommended:


to better characterize biomarkers with respect to their sensitivity and specificity, and to validate their predictability for adverse health effects;


to validate biomarkers in high-quality analytical epidemiological studies; prospective studies are likely to give the most definitive validation;


to promote validation of biomarkers, researchers should store information and specimens for pooled or subsequent analysis;


to model human and experimental data to link biomarkers with the expression of disease.

9.3 Application

The Task Group recommended:


to study the population distribution and role in disease development of risk factors using validated susceptibility biomarkers;


to use validated biomarkers to study the effects of public health prevention initiatives among populations exposed to toxicants.


Albertini RJ, Anderson D, Douglas GR, Hagmar L, Hemminki K, Merlo F, Natarajan AT, Norppa H, Shuker DE, Tice R, Waters MD, & Aitio A (2000) IPCS guidelines for the monitoring of genotoxic effects of carcinogens in humans. Mutat Res, 463: 111-172.

Andersen ME, Clewell HJ, Gargas ML, MacNaughton MG, Reitz RH, Nolan RJ, & McKenna MJ (1991) Physiologically based pharmacokinetic modeling with dichloromethane, its metabolite, carbon monoxide, and blood carboxyhemoglobin in rats and humans. Toxicol Appl Pharmacol, 108: 14-27.

Anderson D, Sorsa M, & Waters MD (1994) The parallelogram approach in studies of genotoxic effects. Mutat Res, 313: 101-115.

Anderson S, Sadinski W, Shugart L, Brussard P, Depledge M, Ford T, Hose J, Stegemen J, Suk W, Wirgn I, & Wogan G (1994) Genetic and molecular ecotoxicology: a research framework. Environ Health Perspect, 102 (Suppl 12): 3-8.

Armstrong BK, White E, & Saracci R (1994) Principles of exposure measurement in epidemiology. New York, Oxford University Press, pp 49- 114.

Bailar JC III & Bailer AJ (1999) Common themes at the workshop on uncertainty in the risk assessment of environmental and occupational hazards. In: Bailer AJ, Maltoni C, Bailar JC III, Belpoggi F, Braz JV, & Soffritti M ed. Uncertainty in the risk assessment of environmental and occupational hazards. Ann NY Acad Sci, 895: 373-376.

Bailer AJ & Dankovic DA (1997) An introduction to the use of physiologically based pharmacokinetic models in risk assessment. Stat Methods Med Res, 6: 341-358.

Barrett J, Vainio H, Peakall D, & Goldstein BD ed. (1997) 12th Meeting of the scientific group on methodologies for the safety evaluation of chemicals: susceptibility to environmental hazards. Environ Health Perspect, 105: 699-737.

Becking GC (1995) Use of mechanistic information in risk assessment for toxic chemicals. Tox Letters, 77: 15-24.

Benichou J (1991) Methods of adjustment for estimating the attributable risk in case-control studies: a review. Statistics in Medicine, IV: 1753-1773.

Bernard AM (1995) Biokinetics and stability aspects of biomarkers: recommendation for applications in population studies. Toxicology, 101: 65-71.

Boffetta P (1995) Sources of bias, effect of confounding in the application of biomarkers to epidemiological studies. Toxicol Lett, 77: 235-238.

Bois FY (1999) Analysis of PBPK models for risk characterization. Ann. N Y Acad Sci, 895: 317-337.

Bois FY, Krowech G, & Zeise L (1995) Modeling human interindividual variability in metabolism and risk: The example of 4-aminobiphenyl. Risk Analysis, 15: 205-213.

Bonassi S (1999) Combining environmental exposure and genetic effect measurements in health outcome assessment. Mutat Res, 428: 177-185.

Bonassi S, Ceppi M, Fontana V, & Merlo F (1994) Multiple regression analysis of cytogenetic human data. Mutat Res, 313: 69-80.

Bonassi S, Hagmar L, Strömberg U, Montagud AH, Tinnerberg H, Forni A, Heikkilä P, Wanders S, Wilhardt P, Hansteen I-L, Knudsen LE, & Norppa H (2000) Chromosomal aberrations in lymphocytes predict human cancer independently of exposure to carcinogens. Cancer Res, 60: 1619-1625.

Calabrese EJ (1997) Role of genetic factors in environmentally induced toxic response: historical consideration, present status and future directions. University of Massachusetts, Amherst, MA.

Calleman CJ, Wu Y, He F, Tian G, Bergmark E, Zhang S, Deng H, Wang Y, Crofton KM, Fennell T, & Costa LG (1994) Relationship between biomarkers of exposure and neurological effects in a group of workers exposed to biomarkers of exposure and neurological effects in a group of workers exposed to acrylamide. Toxicol Appl Pharmaco, 126: 361-371.

Cartwright RA, Glashan RW, Rogers HJ, Ahmad RA, Barham-Hall D, Higgins E, & Kahn MA (1982) Role of N-acetyltransferase phenotype in bladder carcinogenesis: a pharmacogenetic epidemiological approach to bladder cancer. Lancet, 11: 842-846.

Dary CC, Quackenboss JJ, Nauman CH, & Hern SC (1996) Relationship of biomarkers of exposure to risk assessment and risk management. In: Blancato JN, Brown RN, Dary CC, Saleh MA ed. Biomarkers for agrochemicals and toxic substances. ACS Symposium Series 643, ACS, Washington, DC, pp 2-23.

DeRosa C, Richter P, Pohl H, & Jones DE (1998) Environmental exposures that affect the endocrine system: public health implications. J Toxicol Environ Health B Crit Rev, 1: 3-26.

D’Erricco A, Taioli E, Chen X, & Vineis P (1996) Genetic metabolic polymorphisms and the risk of cancer: a review of the literature. Biomarkers, 1: 149-173.

Dickey C, Santella RM, Hattis D, Tabg D, Hsu Y, Cooper T, Young T-L, & Perera FP (1997) Variability in PAH-DNA adduct measurements in peripheral mononuclear cells: implications for quantitative cancer risk assessment. Risk Anal, 17: 649-654.

Dor F, Dab W, Empereur-Bissonnet P, & Zmirou D (1999) Validity of biomarkers in environmental health studies: The case of PAHs and Benzene. Critical Rev Tox, 29: 129-168.

Ehrenberg L, Granath F, & Törnqvist (1996) Macromolecular adducts as biomarkers of exposure to environmental mutagens in human populations. Environ Health Perspect, 104 (Suppl 3): 423-428.

El-Masri HA, Bell DA, & Portier CJ (1999) Effects of glutathione transferase theta polymorphism on the risk estimates of dichloromethane to humans. Toxicol Appl Pharmacol, 158: 221-230.

Emancipator K (1999) Laboratory diagnosis and monitoring of diabetes mellitus. Am J Clin Pathol, 112: 665-674.

Evans WE & Relling MV (1999) Pharmacogenomics: translating functional genomics into rational therapeutics. Science, 286: 487-491.

Fennell TR, MacNeela JP, Thompson CL, & Bell DA (1996) Hemoglobin adducts from acrylonitrile and ethylene oxide in cigarette smokers: effects of glutathione transferase Tl and Ml genotypes. Toxicologist, 30, 282 [Abstract No. 1443].

Garcia-Closas M, Rothman N, Stewart WF, & Lubin JH (1998) Impact of misclassification in studies of gene-environmental interactions. Proc Am Assoc Cancer Res, 39: 181.

Garnier R, Rambourg-Schepens MO, Muller A, & Hallier E (1996) Glutathione transferase activity and formation of macromolecular adducts in two cases of acute methyl bromide poisoning. Occup Environ Med, 53: 211-215.

Goldberg M & Hémon, D (1993) Occupational epidemiology and assessment of exposure. Int J Epidemiol, 22 (Suppl 2): s5-s9.

Goldstein BD (1996) Biological markers and risk assessment. Drug Metab Rev, 28: 225-233.

Greenblatt M, Bennett W, Hollstein M, & Harris C (1994) Mutations in the p53 tumor suppressor gene: clues to cancer etiology and molecular pathogenesis. Cancer Res, 54: 4855-4878.

Greenland S (1993) Basic problems in interaction assessment. Environ Health Perspect, 101(Suppl 4): 59-66.

Hagmar L, Bonassi S, Strömberg U, Brøgger A, Knudsen LE, Norppa H, Reuterwall C, and the European Study Group on Cytogenetic Biomarkers and Health (1998) Chromosomal aberrations in lymphocytes predict human cancer: a report from the European Study Group on Cytogenic Biomarkers and Health (ESCH). Cancer Res, 58: 4117-4126.

Hatch M & Thomas D (1993) Measurement issues in environmental epidemiology. Environ Health Perspect, 101(Suppl 4): 49-57.

Hattis D & Silver K (1993) Use of biomarkers in risk assessment. In: Schulte PA & Perera FP ed. Molecular epidemiology: principles and practices, San Diego, Academic Press, pp 251-273.

Hernberg S & Aitio A (1987) Validation of biological monitoring tests. In: Foa V, Emmett EA, Maroni M, & Columbi A ed. Occupational and environmental chemical hazards: cellular and biochemical indices for monitoring toxicity. Chichester, England, Ellis Horwood, pp 41-49.

Howe GR (1998) Practical uses of biomarkers in population studies. In: Mendelsohn ML, Mohr LC, & Peeters JP ed. Biomarkers: medical and workplace applications, pp 41-49.

Hulka BS (1991) Epidemiological studies using biological markers: issues for epidemiologists. Cancer Epidemiol Biomarkers Prev, 1: 13-19.

Hulka BS & Margolin BH (1992) Methodologic issues in epidemiological studies using biomarkers. Am J Epidemiol, 135: 200-204.

IARC (1997) Application of biomarkers in cancer epidemiology. Lyon, International Agency for Research on Cancer, IARC Scientific Publications No. 142, p1.

IPCS (1992) Environmental Health Criteria 134: Cadmium. Geneva, World Health Organization, International Programme on Chemical Safety, 280 pp.

IPCS (1993) Environmental Health Criteria 155: Biomarkers and risk assessment: Concepts and principles. Geneva, World Health Organization, International Programme on Chemical Safety, 82 pp.

IPCS (1994) Environmental Health Criteria 170: Assessing human health risks of chemicals: Derivation of guidance values for health-based exposure limits. Geneva, World Health Organization, International Programme on Chemical Safety, 73 pp.

IPCS (1999) Environmental Health Criteria 210: Principles for the assessment of risks to human health from exposure to chemicals. Geneva, World Health Organization, International Programme on Chemical Safety, 110 pp.

Joksiƒ G & Spasojeviƒ -Tišma V (1998) Chromosome analysis of lymphocytes from radiation workers in the tritium-applying industry. Int Arch Occupa Environ Health, 71: 213-220.

Khoury MJ & Wagener DK (1995) Epidemiological evaluation of the use of genetics to improve the predictive value of disease risk factors. Am J Genet, 56: 835-844.

Kriebel D (1994) The dosimetric model in occupational and environmental epidemiology. Occ Hyg, 1: 55-68.

Louis TA (1988) General methods for analyzing repeated measures. Stat Med, 7: 39-45.

Lucier GW & Thompson CL (1987) Issues in biochemical applications to risk assessment. When can lymphocytes be used as surrogate markers. Environ Health Perspect, 76: 187-191.

Margetts BM (1991) Basic issues in designing and interpreting epidemiological research. In: Margetts BM & Nelson M ed. Design concepts in nutritional epidemiology. New York, Oxford University Press, pp 13-51.

McClellan RO (1995) Risk assessment and biological mechanisms, lesson learned, future opportunities. Toxicology, 102: 239-258.

McClellan RO (1999) Human health risk assessment: a historical overview and alternative paths forward. Inhalation Tox, 11: 477-518.

McMichael AJ & Hall AJ (1997) The use of biological markers as predictive early-outcome measures in epidemiological research. In: Toniolo P, Boffetta P, Shuker DEG, Rothman N, Hulka B, & Pearce N ed. Application of biomarkers in cancer epidemiology. IARC Scientific Publications No. 142, Lyon, International Agency for Research on Cancer, pp 281-289.

Merlo F, Andreassen Å, Weston A, Pab C-F, Haugen A, Valerio F, Reggiardo G, Fontana V, Garte S, Puntoni R, & Abbondandolo A (1998) Urinary excretion of 1-hydroxypyrene as a marker for exposure to urban air levels of polycyclic aromatic hydrocarbons. Cancer Epidemiol Biomarkers Prev, 7: 147-155.

Morgenstern H & Thomas D (1993) Principles of study design in environmental epidemiology. Environ Health Perspect, 101(Suppl 4): 23- 38.

Muñoz A & Gange SJ (1998) Methodological issues for biomarkers and intermediate outcomes in cohort studies. Epidemiologic Rev, 20: 29-42.

Muscat JE (1996) Epidemiological reasoning and biological rationale. Biomarkers, 1: 144-145.

Nelson DI (1997) Risk assessment in the workplace. In: DiNardi S ed. The occupational environment – its evaluation and control. American Industrial Hygiene Association Press, 328-359.

NRC (US National Research Council) (1983). Risk assessment in the Federal Government: Managing the process. Washington, DC, National Academy Press.

NRC (US National Research Council) (1987) Pharmacokinetics in Risk Assessment. Drinking Water and Health, Washington, DC, National Academy Press, p 47.

OSHA (US Occupational Safety & Health Administration) (1998) Methylene chloride; final rule. Federal Register, 63: 50711-50732.

Olea N, Pazos P, & Exposito J (1998) Inadvertent exposure to xenoestrogens. Eur J Cancer Prev, 7(Suppl 1): S17-23.

Ottman R (1995) Gene-environment interaction and public health. Am J Hum Genet, 56: 821-823.

Ottman R (1996) Gene-environment interaction: definition and study designs. Prev Med, 25: 764-770.

Pacchierotti F, Adler I-D, Anderson D, Brinkworth M, Demopoulos NA, Lähdetie J, Osterman-Golkar S, Peltonen K, Russo A, Tates A, & Waters R (1998) Genetic effects of 1,3-butadiene and associated risk for heritable damage. Mutat Res, 397: 93-115.

Pearce N, deSanjose S, Boffetta P, Kogevina M, Saracci R, & Savistz D (1995) Limitations of biomarkers of exposure in cancer epidemiology. Epidemiology, 6: 190-194.

Perera FP (1995) Molecular epidemiology and prevention of cancer. Environ Health Perspect, 103(Suppl 8): 233-236.

Perera F, Mayer J, Jaretzki A, Hearne S, Brenner D, Young TL, Fishman H, & Grimes M (1989) Comparison of DNA adducts and sister chromatid exchanges in lung cancer cases and controls. Cancer Res, 49: 4446-4451.

Ponce RA, Bartell SM, Kavanagh TK, Woods JS, Griffith WC, Lee RC, Takaro TK, & Faustman EM (1998) Uncertainty analysis for comparing predictive models of biomarkers: a case study of dietary methyl mercury exposure. Reg Tox Pharm, 28: 96-105.

Qian GS, Ross RK, Yu MC, Yuan JM, Gao YT, Henderson BE, Wogan GN, & Groopman JD (1994) A follow-up study of urinary aflatoxin exposure and liver cancer risk in Shanghai, People’s Republic of China. Cancer Epidemiol Biomarkers Prev, 3: 3-10.

Remuzzi G & Bertani T (1998) Pathophysiology of progressive renal disease. N Engl J Med, 339: 1448-1456.

Rhomberg L (1995) Estimation and evaluation of dose. In: Farland W, Olin S, Park C, Rhomberg L, Scheuplein R, Starr T, & Wilson J ed. Low-dose extrapolation of cancer risks: issues and perspectives. Washington, DC, International Life Sciences Institute Press, pp 61-74.

Rothman N, Stewart WF, & Schulte, PA (1995) Incorporating biomarkers into cancer epidemiology: a matrix of biomarker and study design categories. Cancer Epid Biomarkers Prevention, 4: 301-311.

Rothman N, Bhatnagar VK, Hayes RB, Zenser TV, Kashyap SK, Butler MA, Bell DA, Lakshmi V, Jaeger M, Kashyap R, Hirvonen A, Schulte PA, Dosemeci M, Hsu F, Parikh DJ, Davis BB, & Talaska G (1996) The impact of interindividual variation in NAT2 activity on benzidine urinary metabolites and urothelial DNA adducts in exposed workers. Proc Natl Acad Sci USA, 93: 5084-5089.

Rothman N, Caporaso NE, Wacholder S, Garcia-Closas M, Lubin JH, Marcus P, Hoover RN, & Fraumeni Jr. JF (1999) Evaluation of interactions between environmental and common genetic polymorphisms: a population-based epidemiologic perspective. (Abstract). American Association of Cancer Research Annual Meeting, Philadelphia.

Saracci R (1997) Comparing measurements of biomarkers with other measurements of exposure. In: Toniolo P, Boffetta P, Shuker DEG, Rothman N, Hulka B, & Pearce N ed. Application of biomarkers in cancer epidemiology. IARC Scientific Publications No. 142, Lyon, International Agency for Research on Cancer, pp 303-312.

Schatzkin A, Freedman LS, Schiffman MH, & Dawsey SJ (1990) Validation of intermediate endpoints in cancer research. JNCI, 82: 1746- 1752.

Schulte PA (1989) A conceptual framework for the validation and use of biomarkers. Environ Res, 48: 129-144.

Schulte PA & Mazzuckelli LF (1991) Validation of biological markers for quantitative risk assessment. Environ Health Perspect, 90: 239-246.

Schulte PA & Perera FP (1993) Validation. In: Schulte PA & Perera FP ed. Molecular epidemiology: principles and practices, San Diego, CA, Academic Press, pp 79-107.

Schulte PA & Talaska G (1995) Validity criteria for use of biological markers of exposure to chemical agents in environmental epidemiology. Toxicol, 101: 73-88.

Schulte PA & Waters, M (1999) Using molecular epidemiology in assessing exposure for risk assessment. Ann NY Acad Sci, 895: 101-111.

Schulte PA, Boeniger M, Walker J., Schober SE, Pereira MA, Gulati DK, Wojciechowski JP, Garza A, Froelich R, Strauss G, Halperin WE, Herrick R, & Griffith J (1992) Biological markers in hospital workers exposed to low levels of ethylene oxide. Muta Res, 278: 237-251.

Schulte PA, Rothman N, & Schottenfield D (1993) Design consideration in molecular epidemiology. In: Schulte PA & Perera FP ed. Molecular epidemiology: principles and practices, San Diego, CA, Academic Press, pp 159-198.

Sexton K, Reiter LW, & Zenick H (1995) Research to strengthen the scientific basis for health risk assessment: a survey for the context and rationale for mechanistically based methods and models. Toxicology, 102: 3-20.

Shugart LR, McCarthy JF, & Halbrook RS (1992) Biological markers of environmental and ecological contamination: An overview. Risk Analysis, 12: 353-360.

Sobels FH (1977) Some problems associated with the testing for environmental mutagens and a prospective for studies in comparative mutagenesis. Mutat Res, 46: 245-260.

Sobels FH (1993) Approaches to assessing genetic risks from exposure to chemicals. Environ Health Perspect, 101(Suppl 3): 327-332.

Stayner L (1992) Methodologic issues in using epidemiologic studies for quantitative assessment. In: Clewell HJ ed. Proceedings from Conference of Chemical Risk Assessment in the DOD: Science, Policy, and Practice. ACGIH, Cincinnati, OH, pp 43-51.

Steenland K, Tucker J, & Salvan A (1993) Problems in assessing the relative predictive value of internal markers versus external exposure in chronic disease epidemiology. Cancer Epidemiol Biomarkers Prev, 2: 487- 491.

Stewart PA, Blair A, Dosemeci M, & Gomez M (1991) Collection of exposure data for retrospective occupational epidemiologic studies. Appl Occup Environ Hygiene, 6: 280-289.

Stokinger HE & Mountain JT (1963) Tests for hypersusceptibility to hemolytic chemicals. Arch Environ Health, 6: 495-502.

Sutter JR (1995) Molecular and cellular approaches to extrapolation for risk assessment. Environ Health Perspect, 103: 386-389.

Trock BJ (1995) Application of biological markers in cancer environmental epidemiology. Toxicology, 101: 93-98.

Verberk M (1995) Biomarkers of exposure versus parameters of external exposure; practical applications in estimating health risks. Toxicology, 101: 107-115.

Vineis P & Schulte PA (1995) Scientific and ethical aspects of genetic screening of workers: the case of the N-acetyltranferase phenotype. J Clin Epidemiol, 48: 189-197.

Vineis P, Caporaso N, Tannenbaum SR, Skipper PL, Glogowski J, Bartsch H, Coda M, Talaska G, & Kadlubar F (1990) Acetylation phenotype, carcinogen-hemoglobin adducts, and cigarette smoking. Cancer Res, 50: 3002-3004.

Waters MD & Nolan C (1995) EC/US workshop report: assessment of genetic risks associated with exposure to ethylene oxide, acrylamide, 1,3-butadiene and cyclophosphamide. Mutat Res, 330: 1-11.

Waters MD, Stack HF, & Jackson MA (1999) Genetic toxicology data in the evaluation of potential human environmental carcinogens. Mutat Res, 437: 21-49.

Weinberg (1993) Toward a clearer definition of confounding. Am J Epidemiol, 137: 1-8.

White E (1997) Effects of biomarker measurement error on epidemiological studies. In: Toniolo P, Boffetta P, Shuler DEG, Rothman N, Hulka B, & Pearce N ed. Application of biomarkers in cancer epidemiology. IARC Scientific Publications No. 142, Lyon, International Agency for Research on Cancer, pp 73-93.

WHO (1975) Early detection of health impairment in occupational exposure to health hazards: report of a WHO study group. WHO Technical Report No. 571. Geneva, World Health Organization.

Willett W (1998) Nutritional Epidemiology (2nd Edition) New York, Oxford University Press, pp 484-496.


Kari Hemminki

Department of Biosciences at Novum, Karolinska Institute, 141 57 Huddinge, Sweden





3.1 DNA adduct in humans

3.1.1 Assay of complex polyaromatic adducts

3.1.2 Aromatic DNA adducts in occupational and environmental exposure

3.1.3 Alkenes

3.1.4 UV-induced damage

3.1.5 Adducts excreted in urine

3.1.6 Adducts and metabolic genotypes










Biomarkers are often divided into those measuring exposure, effect and susceptibility (Albertini et al., 1996). The division is somewhat ambiguous but serves in the grouping of biomarkers, as shown in Fig. 5. Biomarkers of exposure and effect for carcinogenicity (genotoxic carcinogens), to be covered in this Appendix, include DNA and proteins adducts, cytogenetic changes and point mutations. As shown in Fig. 5, the chapter covers "Biologically effective dose" and "Early genotoxic effects". The biomarkers were chosen because they are 1) beyond measurement of the chemical or its metabolite (internal dose, biological monitoring), 2) currently applicable to humans, 3) detectable in healthy individuals, unrelated to pre-cancerous conditions (early diagnosis of cancer is outside the scope of this chapter), and 4) mechanistically related to cancer (however for protein adducts the link is indirect). Validity aspects of biomarker applications have been discussed in Appendix 4 and separate publications (IARC, 1997), and this Appendix is mainly limited to technical validity (performance of the assay for human specimens) rather than predictivity for cancer outcome.

Figure 5

Cytogenetic tests involving scoring of microscopic chromosomal aberrations are the oldest of biomarkers used and are still applied in, for example, accidental radiation exposure. Subsequently, sister chromatid changes and micronuclei have been introduced. Recently, in situ fluorescence techniques (FISH) have been used in order to score specific chromosomes and chromosomal loci. Applications of DNA and protein adducts and point mutations in human biomonitoring are also recent, most publications having appeared during the past decade.

There have been substantial technical developments in methods of biomonitoring. Here techniques will be discussed only as far as is necessary for the interpretation of the results. For DNA adducts the main techniques, including the 32P-postlabelling technique, will be described. For protein adducts, haemoglobin adduct techniques will be presented. For point mutations, only two of many systems, i.e., those based on the hypoxanthine-guanine phosphoribosyl transferase (HPRT) gene and on the glycophorin A gene, are discussed here. The modern rather than the conventional cytogenetic techniques will be discussed.

There is an apparent imbalance in the length of text devoted to the different biomarkers. However, the length is probably roughly proportional to t